Journal of Cleaner Production 414 (2023) 137453 Contents lists available at ScienceDirect Journal of Cleaner Production journal homepage: www.elsevier.com/locate/jclepro Nitrite pathway in A2/O WWTPs: Modelling organic matter reduction, operational cost and N2O emissions Àlex Gaona , Borja Solís , Javier Guerrero , Albert Guisasola , Juan A. Baeza * GENOCOV, Departament d’Enginyeria Química, Biològica i Ambiental, Escola d’Enginyeria, Universitat Autònoma de Barcelona, 08193, Bellaterra, Barcelona, Spain A R T I C L E I N F O A B S T R A C T Handling Editor: Jin-Kuk Kim Deciding whether the implementation of the nitrite pathway is useful for the treatment of a given wastewater depends on several criteria such as organic matter availability, N and P removal, N2O emissions and operational Keywords: costs. This work is a simulation-based study with a conventional Anaerobic/Anoxic/Oxic (A2/O) WWTP where Denitritation the nitrite pathway can be implemented. The outcomes are general correlations to calculate the minimum COD Enhanced biological phosphorus removal requirements for a certain influent and a practical decision tree on the opportunities of nitrite pathway in A2/O (EBPR) WWTPs as a function of the influent wastewater composition (for P and N concentrations in the ranges 3–11 Nitritation Nitrite pathway mgPO−4 3-P⋅L− 1 and 20–60 mg NH+ − 1 4 -N⋅L ). This study shows that the implementation of the nitrite pathway N2O emissions reduces the COD requirements (depending on the influent, between 9 and 68%). It also can lead to a reduction in Wastewater aeration costs. For an equivalent COD in the influent, the implementation of the nitrite pathway compared to the nitrate pathway results in a reduction of 41-47% in aeration costs and similar sludge production, leading to a reduction in the operational cost index of 10-16%. However, it is essential to note that this strategy can also lead to increased N2O emissions, with an emission factor for the nitrite pathway in the range of 2.5–17 times that of the nitrate pathway. 1. Introduction among them, the nitrite pathway or shortcut biological N removal has attracted much attention, particularly for wastewaters with a low Nutrient removal (N and P) is mandatory for many WWTPs. Con­ chemical oxygen demand (COD) to N ratio. ventional biological nitrogen removal (BNR) involves two main pro­ Nitrite pathway, i.e. autotrophic ammonium oxidation to nitrite cesses called nitrification and denitrification. Nitrification is the (nitritation) and posterior heterotrophic nitrite denitrification (deni­ oxidation of ammonium (NH+ 4 ) to nitrate (NO3 ), while denitrification is − tritation), results in 40% lower COD requirements for denitrification, the reduction of nitrate to nitrogen gas. The commonly used biological savings in aeration costs and lower sludge production (Turk and process for removing P is the well-established enhanced biological Mavinic, 1986) when compared to the conventional nitrification/deni­ phosphorus removal (EBPR), which is based on the enrichment of acti­ trification processes based on the complete oxidation of ammonium to vated sludge in phosphorus accumulating organisms (PAO) (Oehmen nitrate and nitrate denitrification (nitrate pathway). It is therefore not et al., 2007). The simplest WWTP configuration combining the processes surprising that numerous studies are looking for the best strategies to required to achieve simultaneous biological C/N/P removal is the implement it (Duan et al., 2022; Liu et al., 2020; Regmi et al., 2014; Xu widely used anaerobic-anoxic-aerobic (A2/O) continuous plant (Tcho­ et al., 2021). Thus, a configuration involving nitrite-reducing deni­ banoglous et al., 2014). Various operational strategies have been studied trifying PAO (DPAO) and N removal via nitrite appears to be a good under the A2/O configuration with the aim of improving the quality of alternative to treat wastewaters with a low COD/N influent ratio with treated effluents and significantly reducing its energetic and economic lower aeration needs and less sludge production. burden. Aeration requirements and excess sludge treatment represent Two major issues arise against the nitrite shortcut process: the po­ the highest operating costs of a WWTP (Changqing et al., 2011). Many tential nitrite inhibition on biological processes (Zhou et al., 2011) and strategies have been put forward to reduce the aeration needs and, the potential nitrous oxide (N2O) production and emissions (Kirim et al., * Corresponding author. E-mail addresses: Alex.Gaona@uab.cat (À. Gaona), Borja.Solis@uab.cat (B. Solís), FranciscoJavier.Guerrero@uab.cat (J. Guerrero), Albert.Guisasola@uab.cat (A. Guisasola), juanantonio.baeza@uab.cat (J.A. Baeza). https://doi.org/10.1016/j.jclepro.2023.137453 Received 4 February 2023; Received in revised form 8 May 2023; Accepted 10 May 2023 Available online 17 May 2023 0959-6526/© 2023 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). À. Gaona et al. Journal of Cleaner Production 414 (2023) 137453 2022). Regarding the first issue, although inhibition of DPAO by nitrite implementation regarding COD requirements to meet N and P discharge has been reported (Zhou et al., 2010), the possibility of long-term stable limits, N2O emissions, economic cost and type of carbon source are also maintenance of simultaneous denitritation and P-removal has also been discussed. demonstrated experimentally (Guisasola et al., 2009; Tayà et al., 2013; Vargas et al., 2011). With respect to N2O emissions, different production 2. Materials and methods pathways have been described in parallel to the nitrification and deni­ trification processes (Massara et al., 2017; Spérandio et al., 2016), which 2.1. Description of WWTP configuration and influent data gives some ability to predict its emissions for different scenarios. Since an electron donor is required for both denitrification and EBPR, This work studies a conventional A2/O WWTP, designed for the low loaded influents may result in a failure of one of these processes biological removal of N, P and organic matter (Fig. 1). Table 1 sum­ depending on the plant configuration, the influent C/N/P and the con­ marizes the volumes of the reactors and the operational parameters trol strategies applied (Guerrero et al., 2014). This challenge can be chosen for the study. The volumes were initially based on general design overcome by adding external carbon or metal salts for chemical P pre­ parameters (Tchobanoglous et al., 2014) but the reactors were increased cipitation, but with a resulting increase in operational costs. in size to ensure good plant performance for the full range of influent There are also several operational strategies that can minimise the values tested. Two different dissolved oxygen (DO) set points in the adverse effect of low COD, but designing and/or retrofitting existing aerobic reactor were used to achieve ammonium oxidation to nitrate WWTPs is not a straightforward issue due to the complex interactions (DO = 3.0 mg/L) or nitrite (DO = 0.8 mg/L). between the different operational variables. For example, the choice of The influent characteristics are summarized in Table 2. The COD was an internal recycle is a trade-off decision: a high internal ratio can fractionated into soluble inert organics (SI), particulate inert organics decrease the effluent nitrate/nitrite at expenses of a higher operational (XI), biomass (XH, XPAO, XAOB and XNOB), soluble fermentable substrate cost. Besides that, an excessive ratio increases the oxygen load to the (SF), volatile fatty acids (SA) and slowly biodegradable substrate (XS), anoxic reactor and reduces the available carbon source for denitrifica­ using different values depending on the scenario. The concentrations of tion. Then, the value of the optimal internal recycling ratio should be PO3− 4 studied were for 3, 5, 7, 9, 11 mg P⋅L − 1 and the influent NH+ 4 carefully selected to optimise a WWTP performance (Baeza et al., 2004). concentration ranged from 20 to 60 mg N⋅L− 1. Considering this complex scenario, it is not surprising that mathe­ matical modelling has become a very useful tool for the prediction of WWTP performance under different operational conditions and when 2.2. Model description different control strategies are applied (Henze et al., 2008; Jeppsson et al., 2013). The activated sludge models (ASM) developed by the In­ The kinetic model used was the novel ASM2d-N2O developed by ternational Water Association (IWA) (Henze et al., 2000) describe the Massara et al. (2018). The ASM2d-N2O model can predict COD, P and N biological and chemical transformations occurring in activated sludge removal and N2O production in full-scale WWTPs (Solís et al., 2022c). systems. Besides the default models (ASM1, ASM2 and ASM3), several The ASM2d-N2O model structure is based on the ASM2d model (Henze extended ASM-based models have been put forward to predict the et al., 2000) and can account for the three known N2O production bio­ WWTP performance under different scenarios. For instance, the logical pathways. On the one hand, it was extended with the 2-pathway ASM2d-N2O model (Massara et al., 2018; Solís et al., 2022a) has recently model for N2O emissions by AOB, developed by Pocquet et al. (2016). been proposed to account for N2O emissions in WWTPs with C/N/P Moreover, it includes the denitrification processes described in the removal and has proven to be a good prediction tool for full-scale WWTP conditions (Solís et al., 2022c). Table 1 The main objective of this work is to provide, for the first time, easy Operational conditions. manageable tools/correlations for A2/O WWTPs to estimate the organic Parameter Description Value matter requirements to meet legal discharge limits (TN < 10 mgN⋅L− 1 V1 Anaerobic reactor volume 2800 m3 and P < 1 mgP⋅L− 1) when treating different wastewaters and using ni­ V2 Anoxic reactor volume 2800 m3 trite or nitrate pathway for N-removal. The obtained correlations could V3 Aerobic reactor volume 9000 m3 be used as a preliminary rule of thumb to decide the most suitable QIN Influent flowrate 20000 m3 d− 1 RI Ratio of internal recycle to influent flowrate 5 pathway for each wastewater. For this purpose, an A2/O WWTP is RE Ratio of external recycle to influent flowrate 1/3 simulated in numerous scenarios, determining the COD requirements for QW Purge flowrate 300 m3 d− 1 different N and P influent content, both for nitrite and nitrate pathways. DONO3P DO setpoint aerobic reactor (nitrate pathway) 3.0 mg L− 1 The competitive advantages and drawbacks of nitrite pathway DONO2P DO setpoint aerobic reactor (nitrite pathway) 0.8 mg L− 1 Fig. 1. Schematics of the A2/O WWTP configuration. 2 À. Gaona et al. Journal of Cleaner Production 414 (2023) 137453 Table 2 discharge limits. The same procedure was applied to determine XS-N and Influent characteristics. XS-P but modifying XS in the influent. Composition Concentration (mg⋅L− 1) 2.5. Evaluation criteria PO3- 4 -P 3–11 NH+4 -N 20–60 NO−3 -N 2.6 Various criteria have been used to assess the performance of a WWTP TSS 58 (Gernaey et al., 2014; Stare et al., 2007). In this study, the operation cost XI 45.6 index (OCI [kWh⋅d− 1]) that combines costs for sludge disposal, aeration SI 30 and pumping was calculated with equation (3). OCI = AE + PE + fSP ⋅SP (3) Activated Sludge Model for Nitrogen (ASMN), developed by Hiatt and Grady (2008). ASMN describes the N2O production by OHO and PAO as Where, AE is aeration energy, PE is pumping energy, and SP is sludge an obligate intermediate of the denitrification process. The complete production for disposal. The weight fSP has a value of 5 to account for the description of the model can be found in the supplementary information, sludge production processing being carried out in that system and the including model components (Table S1), stoichiometry (Tables S2 and expected sludge disposal cost savings as a result. S3), kinetic equations (Table S4), kinetic parameters (Table S5) and Average aeration energy (AE [kWh⋅d− 1]) was based on equation (4). stoichiometric parameters (Table S6). ∫ To evaluate the plant performance, each scenario was simulated with 24 tend ( ) AE = 3.0248 ⋅ 10− 4 ⋅ V ⋅ KL a(t)2 + 0.5882 ⋅ 10− 2 ⋅ V ⋅ KL a(t) dt (4) constant flow and wastewater composition until achieving steady state tobs tstart conditions. Simulations of 200 days were required to ensure steady state for nitrate pathway and 300 days for nitrite pathway. Where, KL a, is the oxygen transfer rate value expressed in h− 1, V is the volume of the aerobic reactor (m3) and tobs represents the total length of the evaluation period in days, i.e., tend − tstart . KL a(t) values were ob­ 2.3. Selection of influent variables and effluent discharge variables tained from the model simulation of a proportional integral control loop implemented to maintain the DO setpoint in the aerobic reactor. Both cases (nitrite pathway and nitrate pathway) were simulated for Pumping energy (PE [kWh⋅d− 1]) was calculated as a weighted different wastewater compositions in the range reported in Table 2. The average sum of the various pumped flows in the system under study, minimum amount of organic matter that was necessary to meet the using equation (5). discharge limits for N and P, considering only soluble nitrogen (SNH4, ∫ tend ( ) SNO3, SNO2) and phosphate (SPO4) was obtained for each pair of N and P 1 PE = fPE Qint ⋅ Qint + fPE Qr ⋅ Qr + fPE Qw ⋅ Qw + fPE Qpu ⋅ Qpu ⋅ dt concentrations in the wastewater under two different scenarios. In the tobs tstart first scenario, all organic matter was considered to be like volatile fatty (5) acids (SA) to simulate the most favourable organic matter composition for PAO, whereas in the second scenario only slowly biodegradable Where, Qint is the internal recycle flow, Qr is the external recycle flow, organic matter (XS) was used to have a less favourable organic matter Qw purge flow and Qpu is the underflow from secondary clarifier, all fractionation. For the latter case, the stoichiometric parameters iNXS and flows expressed in m3⋅d− 1. Table 3 shows the values used for the PE iPXS were assumed to be zero, to avoid an unrealistic increase of P and N factors. load when increasing XS concentration. Sludge production for disposal (SP [kg⋅d− 1]) was calculated as the The P and N content of the particulate compounds in the effluent was mass of solids accumulated in the plant or removed from the process not considered, because a change in the amount of biomass in the with the equation (6): effluent could interfere with the results provided by the model, resulting ( ∫ tend ) 1 in too high and unrealistic organic matter requirements. Then, the SP = ⋅ MTSS (tend ) − MTSS (tstart ) + TSSx (t) ⋅ Qx (t) ⋅ dt (6) tobs ⋅1000 biomass concentration in the effluent was considered to be negligible, e. tstart g. by installing a sand filter removing the non-settled biomass in the Where, Qx (t) (m3⋅d− 1) is the sludge flow and TSSx is the total solids sedimentation tank. concentration in the sludge flow stream. MTSS is defined as the sum of the total suspended solids mass (in kg) present in the plant, including re­ 2.4. Methodology for determining organic matter requirements actors and secondary settler (equation (7)). (7) The effluent legal discharge limits considered were PO3− MTSS (t) = MTSS,an (t) + MTSS,anox (t) + MTSS,aer (t) + MTSS,s (t) 4 < 1 mg P⋅L− 1, TN < 10 mgN⋅L− 1 and COD <125 mg O2⋅L− 1 (EEC Council, 1991). With, Effluent organic matter never exceeded the discharge limits because the model was run with the minimum amount of organic matter in the MTSS,x (t) = TSS,x (t)⋅Vx (8) influent. The minimum required concentrations of SA and XS were estimated using the MATLAB optimisation function fminbnd and mini­ Subscripts ‘an’, ‘anox, ‘aer’ and ‘s’ refer to the reactors and the secondary mising independently equation (1) for N and (22) for P. settler. Vx is the volume of each unit process in m3. fobjN = abs(SNH4 + SNO2 + SNO3 − 10) (1) fobjP = abs(SPO4 − 1) (2) Table 3 Where SNH4, SNO2, SNO3 and SPO4 are the effluent steady state nitrogen Pumping energy factors. concentrations of ammonium, nitrite, nitrate and phosphorus concen­ Factor Value Units tration of orthophosphate obtained for a given influent composition. 3 0.004 kWh⋅m− Then, the optimisation of equation (1) by modifying SA in the influent f PE Qint 3 0.008 kWh⋅m− led to the obtention of SA-N, i.e. the minimum amount of SA for satisfying f PE Qr 3 f PE 0.050 kWh⋅m− N discharge limits, while the optimisation of equation (2) by modifying Qw 3 f PE 0.075 kWh⋅m− SA in the influent gave SA-P, i.e. the minimum amount of SA to get P Qpu 3 À. Gaona et al. Journal of Cleaner Production 414 (2023) 137453 3. Results and discussion setpoints was performed to find proper setpoints for nitrite pathway and nitrate pathway. Fig. S2 shows the steady-state AOB and NOB concen­ 3.1. Selection of internal recycle ratio and DO setpoint trations obtained for each DO setpoint. Nitrite pathway obtainment was considered successful when the ratio AOB/NOB was higher than 4.5 and − 1 The first step of this work was to set a suitable internal recycle ratio ammonium oxidation was complete (NH+ 4 < 1 mgN⋅L ). This ratio (RI) for all the simulations, since it was detected (in a preliminary set of prevented nitrate formation and enabled nitrogen removal mainly via simulations) that the results of minimum amount of readily biodegrad­ nitrite. As can be observed in Fig. S2, the range where AOB are retained able organic matter (SA) required to meet the nutrient discharge limits while NOB are washed out is very narrow. Considering these results, the depended on the recycle ratio. Fig. S1 displays the SA requirements for DO setpoint selected to reach nitrite pathway was 0.8 mg L− 1 while a DO different RI (i.e. 2, 4, 5, 6, 8) for an influent ammonium range of 10–60 set-point of 3 mg L− 1 was chosen for nitrate pathway in all subsequent mgN⋅L− 1 and an influent P of = 7 mgP⋅L− 1. For RI = 2, the SA re­ simulations. Stable biological nitrogen removal was obtained with both quirements rose severely for influent ammonium values higher than 35 strategies. mgN⋅L− 1. A similar effect was observed at RI = 4 for influent ammonium values higher than 50 mgN⋅L− 1. This sudden increase is caused by ni­ 3.2. Minimum SA requirements under different scenarios trate/nitrite limitations due to insufficient internal recycle. Therefore, it was decided to use an RI of 5 in all subsequent simulations, as it is the The minimum amount of SA required to meet legal discharge limits minimum RI value that did not show any limitation in the range of (TN < 10 mgN⋅L− 1 and P < 1 mgP⋅L− 1) was evaluated for influent NH+ 4 influent ammonium values used. concentrations in a range of 20–60 mgN⋅L− 1 and phosphate concentra­ Another required preliminary step was the choice of two proper DO tions in the range of 3–11 mgP⋅L− 1, using RI = 5, and under two different setpoints in the aerobic reactor that would allow either nitrate or nitrite DO setpoints (0.8 mg⋅L− 1and 3 mg L− 1) to achieve both nitrite and ni­ pathway. The nitrite pathway is achieved by creating conditions under trate pathway. Fig. 2 shows the SA requirements for nitrogen (solid lines) which nitrite-oxidizing bacteria (NOB) are eliminated from the system and phosphorus removal (dashed lines). Fig. 2-left (and Fig. S3) sum­ while ammonia-oxidizing bacteria (AOB) are retained. Mathematically, marizes the results for nitrate pathway and Fig. 2-right (and Fig. S4) for equation (9) should be satisfied to eliminate NOB while retaining AOB: nitrite pathway. SA requirements increase proportionally to the increase of ammonium in the influent. Tables S7 and S8 report the AOB/NOB μNOB - bNOB < 1/SRT < μAOB - bAOB (9) ratios obtained for each specific scenario to verify that NOB washout where μNOB and μAOB are the apparent growth rates of NOB and AOB, was really accomplished in the nitrite pathway scenario. respectively, bNOB and bAOB the apparent decay rates of these pop­ Regarding the results for nitrate pathway, the SA requirements for N ulations, and SRT the sludge retention time. removal (SA-N, solid lines) start at around 140 mgCOD⋅L− 1 at the lowest The aim was to select a DO setpoint of the aerobic reactor that influent N load (20 mgN⋅L− 1) and increase up to 570 mgCOD⋅L− 1 for an satisfied equation (9), also considering that DO affects the apparent influent with 60 mgN⋅L− 1 and 11 mgP⋅L− 1. The SA-N is also affected by growth rate values. Then, a simulation study under different DO the influent P (i.e., the slope of SA-N vs influent ammonium increases with influent P) showing an additive effect of increasing both N and P Fig. 2. Minimum SA requirements to meet the legal P and TN effluent limits (PO3− 4 < 1 mgP⋅L − 1 and TN < 10 mg N⋅L− 1) when operating under nitrate pathway (left) and nitrite pathway (right) conditions. 4 À. Gaona et al. Journal of Cleaner Production 414 (2023) 137453 concentrations. The rationale is that the influent COD is preferably hand, the SA-P lines have the same trend as in the nitrite pathway, but consumed by PAO in the anaerobic reactor and, thus, higher EBPR ac­ the absolute values are slightly lower. In this case, both N or P removal tivity results in higher COD requirements. Besides, the electrons for ni­ can be the limiting step depending on the specific case, and only for trate reduction in the anoxic reactor are both provided by the organic influent ammonium concentrations above 50 mgN⋅L− 1 N, SA-N becomes matter not degraded under anaerobic conditions and by the stored PHA the only limiting factor. involved in the DPAO metabolism. Comparing both scenarios, the SA requirements are always lower The SA requirements for P removal (SA-P, dashed lines) increase when using nitrite pathway either for meeting P or TN discharge limits. almost linearly with the influent P in the range tested (3–11 mgP⋅L− 1) Depending on the specific influent, reductions in the overall need for SA for a given ammonium concentration. Moreover, SA-P increases when N between 9% and 68% can be achieved (see Supplementary Information in the influent increases and parallel SA-P lines are observed for different 3). For instance, for an influent of 9 mgP⋅L− 1 and 40 mgN⋅L− 1, SA-P and influent P values. Considering both SA-N and SA-P, the limiting step SA-N under nitrate pathway conditions are 171 mg COD L− 1 and 359 mg under nitrate pathway conditions is usually related to N removal (except COD⋅L− 1, respectively. When the nitrite pathway is implemented, both for influent ammonium values lower than 25 mg N⋅L− 1). That means requirements are reduced to 145 and 136 mg COD⋅L− 1, respectively, that, in most cases, SA limitations would result in an effluent that does which would represent a reduction of 60% in the overall SA needs. not meet TN requirements. Reduction of organic matter requirements is, obviously, one of the major Concerning the results for the nitrite pathway (Fig. 2-right), the SA-N advantages of using the nitrite pathway. requirements also increase when the influent ammonium concentration increases, but the observed values are less than 50% of those of the ni­ trate pathway. Moreover, SA-N requirements are not significantly 3.3. Development of general correlations (SA) affected by the increased P concentration and therefore the lines ob­ tained for different P levels in the influent mostly overlap. On the other To improve the usability of the modelling results described in the previous section, different correlations were proposed, their parameters were fitted by least squares and their performance was evaluated. The Table 4 best correlations in terms of ability to correctly describe the above data Correlations of SA and XS requirements for nitrate and nitrite pathway. while maintaining low complexity are shown in Table 4. These corre­ Scenario Correlation lations can be useful for a simple and reliable estimation of the SA re­ SA requirements quirements for WWTPs with A2/O configuration treating wastewater Nitrate pathway SA-P = 2.51 + 14.03 P + 1.046 N (10) with influent nutrient concentrations in the ranges of 3–11 mgP⋅L− 1 and SA-N = 66.51–14.67 P + 4.72 N + 0.600 P N (11) 20–60 mgN⋅L− 1. Nitrite pathway SA-P = − 4.51 + 14.12 P + 0.517 N (12) Thus, equations (10) to (13) can describe SA requirements based only SA-N = − 32.92 – 3.169 P + 4.144 N + 0.066 P N (13) on the influent N and P. Correlations (10) and (12) for SA-P were XS requirements expressed as linear equations (y = b0+b1⋅P + b2⋅N), while correlations Nitrate pathway XS-P = − 7.21 + 31.45 P + 3.22 N (14) (11) and (13) for predicting SA-N were extended with a non-linear XS-N = − 18.73 – 7.67 P + 8.80 N + 0.305 P N (15) summand (y = b0+b1⋅P + b2⋅N + b3⋅P⋅N) for a better description of Nitrite pathway XS-P = − 29.52 + 27.67 P + 1.68 N (16) the observed N-P dependence. These equations correlated well with the XS-N = − 78.00 – 0.367 P + 5.69 N + 0.008 P N (17) original model data, obtaining R2 values of 0.9993, 0.9923, 0.9997 and 1 Fig. 3. Comparison of the predicted SA requirements using the proposed correlations and the model runs for an influent of 5 mg P⋅L− under (left) nitrate pathway and (right) nitrite pathway conditions. 5 À. Gaona et al. Journal of Cleaner Production 414 (2023) 137453 2. The maximum SA value obtained with equations (10) and (11) for a given influent determines the minimum SA required to meet the discharge limits when the nitrate pathway is applied. Equations (12) and (13) allow the same determination but for the nitrite pathway scenario. Metcalf and Eddy (Tchobanoglous et al., 2014), a widely used book for WWTP design, also recommends certain threshold ratios for the design of WWTP configurations aiming at simultaneous C, N and P removal (A2/O and UCT): a readily biodegradable COD (rbCOD) content of 6.6 g rbCOD⋅g− 1NO−3 -N and 10 g rbCOD⋅g− 1P for nitrate pathway. Fig. 4 compares the correlations obtained in this work to those in Metcalf and Eddy (Tchobanoglous et al., 2014) in three different case studies, Case 1: 3 mgP⋅L− 1 and 20 mgN⋅L− 1, Case 2: 7 mgP⋅L− 1 and 40 mgN⋅L− 1, Case 3: 11 mgP⋅L− 1 and 60 mgN⋅L− 1. The correlations obtained in this study predict approximately the same SA requirements under the nitrate pathway. The SA requirements for nitrite pathway (not provided in the reference book) result 50%–72% lower. Fig. 4. Comparison of the SA predictions with equations (10-13) versus Metcalf Finally, Fig. 5 shows a useful decision tree based on correlations (10) and Eddy’s recommendations on readily biodegradable COD requirements. to (13) to select the best operational conditions for the proper treatment Three different influent compositions (low, medium and high) are evaluated. of a given wastewater under an A2/O configuration. If the SA content is not enough to meet the desired effluent concentrations, alternatives 0.9925 for equations (10) to (13). For example, Fig. 3 compares the should be considered: the addition of an external carbon source, the results of the model run under different conditions for a given influent addition of chemicals for phosphate precipitation, or the implementa­ with 5 mg P⋅L− 1 with the values predicted by equations (10-13), tion of the nitrite short-cut pathway. For example, given an influent showing the goodness of the fit. The equations in Table 4 are imple­ water with 350 mgCOD⋅L− 1 as SA, 5 mgP⋅L− 1, and 30 mgN⋅L− 1, the SA-P mented in an Excel file that can be found in supplementary information and SA-N are 104 and 225 mgCOD⋅L− 1 respectively. Hence, SA-N is Fig. 5. Decision tree for a continuous A2/O plant to decide whether the implementation of the nitrate or nitrite pathway is required depending on the SA, P and N content of the influent wastewater. SA-nitrate and SA-nitrite stand for the SA requirements when either nitrate or nitrite pathway is selected. 6 À. Gaona et al. Journal of Cleaner Production 414 (2023) 137453 Fig. 6. N2O emissions (top) and N2O-EF (bottom) for different runs under nitrate (left) and nitrite (right) pathway. than that available in the wastewater and N and P removal could not Table 5 meet legal discharge limits. In this case, the SA-P and SA-N under the Influent conditions, OCI results and N2O emissions under nitrate and nitrite nitrite pathway are 154 and 223 mgCOD⋅L− 1, respectively. Since the pathway for different scenarios. required SA is lower than that in the wastewater (350 mgCOD⋅L− 1), Scenario Low Medium High nutrient removal thorough the nitrite pathway is possible. Influent concentrations PO3− 4 (mgP⋅L ) − 1 3 7 11 − 1 NH+4 (mgN⋅L ) 20 40 60 3.4. Evaluating the effect of nitrite pathway on N2O emissions and WWTP SA (mg⋅L− 1) 155 347 590 economy Nitrate Pathway kLa (d− 1) 68.0 130.0 202.5 AE (kWh⋅d− 1) 4123 8800 15371 Textbook knowledge and the previous sections show that the organic PE (kWh⋅d− 1) 7968 7968 7968 matter requirements are much lower using the nitrite pathway. How­ SP (kg⋅d− 1) 1884 3295 5021 ever, the nitrite pathway can also entail other detrimental aspects that OCI 17742 26653 38401 should also be considered for a comprehensive evaluation of both pro­ N2O-EF (%) 1.35 1.37 1.25 Nitrite Pathway cesses. For example, it has been extensively reported that nitrite accu­ kLa (d− 1) 38.9 81.8 133.2 mulation triggers off N2O production and, thus, N2O emissions: high AE (kWh⋅d− 1) 2200 5087 9062 nitrite and low DO promote the NirK and Nor enzymes activity and, thus, PE (kWh⋅d− 1) 7968 7968 7968 the N2O production through the ND pathway (Massara et al., 2017; SP (kg⋅d− 1) 1916 3322 5034 OCI 15915 23021 32132 Pocquet et al., 2016). Hence, a plant operating under nitrite shortcut N2O-EF (%) 3.38 6.40 7.81 conditions may lead to higher GHG emissions which may not compen­ sate the carbon savings from avoiding the full nitrate pathway. Fig. 6 shows the N2O emissions and N2O emission factor (N2O-EF), i.e., the limiting but the value is lower than the actual SA content in the influent percentage of N that is emitted as N2O compared to the influent nitrogen and thus, nutrient removal through the nitrate pathway is possible. On ammonium load, for different runs under a wide range of influent the other hand, for a wastewater with 350 mgCOD⋅L− 1 as SA, 9 mgP⋅L− 1, ammonium and P under nitrate and nitrite pathway and using the and 60 mgN⋅L− 1, SA-P and SA-N are 192 and 542 mgCOD⋅L− 1, respec­ minimum SA required to remove N and P. tively. Thus, the organic matter necessary for these conditions is higher The results show that the N2O-EF predicted via nitrate were in the 7 À. Gaona et al. Journal of Cleaner Production 414 (2023) 137453 range 1.0–1.5, considerably lower than the range 2.8–24.0 obtained via short-chain VFAs and, thus, a preliminary anaerobic hydrolysis pro­ nitrite. Analysing the ratio between both emissions for each simulation, cess would be required to favour PAO growth in the case of an influent it was found that via nitrite represents between 2.8 and 17 times the with complex COD. Thus, the presence of slowly biodegradable organic emission via nitrate. The highest N2O emissions were obtained at the matter would benefit the denitrification process and would alter the highest influent N, because of the high nitrite accumulation. It was also organic matter requirements for complete nutrient removal (Daigger observed that with an increase in influent P, the emission factor and Littleton, 2014). For instance, an influent may contain a sufficient decreased. Higher SA requirements are needed to remove high influent P amount of COD in terms of slowly biodegradable COD but the hydrolysis concentrations and, therefore, N2O denitrification is promoted. could be limiting step and P removal not achieved. This issue was Although the current values in a full-scale WWTP are likely to differ studied with an additional set of simulations where the influent carbon from the emissions predicted by the model, these results demonstrated a source was set as particulate or slowly biodegradable carbon source (XS), clear trend of increasing N2O emissions when using the nitrite pathway. i.e., organic matter that is hydrolysed to easily biodegradable ferment­ Therefore, there is a trade-off between the operational costs and the able organic matter (SF) which then ferments to produce SA. overall sustainability of the WWTP. In particular, the results showed that Fig. 7 shows the amount of XS required to meet the TN and P legal aiming at the reduction of costs related to aeration costs and external discharge limits (XS-N and XS-P, respectively) for the same N and P carbon dosage with the implementation of the nitrite pathway could ranges and scenarios used in section 3.2. For these simulations, XS was lead to a considerable increase in N2O emissions and consequently a free of N and P (model parameters iNXS = iPXS = 0), to avoid increasing higher carbon footprint. In any case, operation strategies based on phase the N and P load with increasing XS, which could mislead the interpre­ alternation (Zou et al., 2022) and intermittent aeration (Solís et al., tation of the results. The values of Figs. 7 and 2 are very different, 2022b) could also be applied to reduce these emissions. reflecting the importance of the wastewater COD fractionation in these Besides N2O emissions, the implementation of nitrite pathway also simulations. For the case of nitrate pathway, XS-P becomes the limiting influences the WWTP economy since it reduces aeration needs. Table 5 step for a much wider range of concentrations than that for SA-P. The summarizes the results of the evaluation criteria described in Section increase in XS-P with respect to SA-P is very important, up to 290% (11 2.4, obtained under the nitrate and nitrite pathway in different scenarios mgP⋅L− 1, 60 mgN⋅L− 1), indicating that the processes of XS hydrolysis to with enough SA to perform either pathway. The OCI values for nitrite SF or SF fermentation to SA are competing with denitrification processes pathway were reduced in the range 10-16% compared to those for ni­ and reduce the available SA for PAO. However, if both XS-P and XS-N are trate pathway. The main difference appeared in the aeration costs (AE considered, total XS requirements do not differ much from SA re­ was reduced in the range 41-47%), while similar sludge production was quirements for some high N scenarios. Supplementary information 3 observed (SP increased 0.3-2% for nitrite pathway). For example, the includes a comparison of XS and SA requirements for a wide range of aeration energy requirements were 8800 kWh⋅d− 1 in the case of an influent concentrations. The use of XS for the nitrate pathway can in­ influent with 7 mgP⋅L− 1 and 40 mgN⋅L− 1 under the nitrate pathway crease the COD requirements from 0 to 127%, while for the nitrite configuration, while the energy requirements were 5087 kWh⋅d− 1 for pathway the increase ranges from 14 to 103%. In any case, these results the nitrite pathway scenario. Although the most important factor in the are obtained with the default parameters used in the ASM2d model for OCI reduction are the aeration costs, this is not the result of reduced hydrolysis and fermentation processes and specific parameter calibra­ oxygen requirements compared to full nitrification. Stoichiometrically, tion might be needed for each particular WWTP. the oxidation of ammonium to nitrite represents 25% less oxygen than Regarding the results for nitrite pathway, XS-P and XS-N also increase the oxidation of ammonium to nitrate, but considered globally, the ni­ with respect those of SA, in this case up to 208% (11 mgP⋅L− 1, 60 trate produced can be reused as an electron acceptor during the deni­ mgN⋅L− 1). XS-N does not become a limiting factor in most of the trification process, recovering the excess oxygen used for the oxidation wastewater composition scenarios. Comparing the overall XS needs in of nitrite to nitrate. Then, the main reason for the decrease in aeration both pathways, the nitrite pathway also systematically requires less COD costs is the lower DO setpoint required to decrease NOB activity for the as in the case of SA. The reduction in XS requirements ranges from 23 to nitrite pathway, which increases the driving force of oxygen transfer 58% depending on the specific influent. (DO saturation – DO setpoint) and hence reduces the required kLa General correlations and a decision tree can be obtained for XS re­ (Table 5). Full-scale studies have already demonstrated the significant quirements likewise the case of SA. These results would ease the decision savings of power demand when implementing a moderate decrease in of A2/O configuration needed when treating wastewater with nutrient DO setpoint while ensuring a good effluent quality (Pasini et al., 2021). concentrations in the ranges 3–11 mgP⋅L− 1 and 20–60 mgN⋅L− 1 and N2O emissions were also simulated for these three scenarios. The slowly biodegradable organic matter. Table 4 also shows the correla­ emissions for the nitrite pathway were lower than those shown in Fig. 6. tions for estimating XS requirements: equations (14) and (16) for XS-P For all three cases, the emission via nitrite is between 2.5 and 6.2 times and (15) and (17) for XS-N, following the same type of equations used for that of the nitrate pathway. These results can be understood because the SA estimations. These equations also highly correlate with the original simulations via nitrite are run with the amount of organic matter needed model data, giving R2 values of 0.9877, 0.9966, 0.9969 and 0.9993 for for the nitrate pathway, which represents a higher amount than strictly equations (14) to (17). Fig. 8 shows the goodness of fit of the correlations necessary. The high values obtained in Fig. 6 can be related to the fact for a given influent with 5 mg P⋅L− 1. that these are all scenarios in which the minimum amount of organic Fig. 9 shows a decision tree based on equations (14) to (17) that can matter necessary to meet the discharge limits is used. In fact, one of the be used to select the best operational conditions for the proper treatment proposed strategies to decrease emissions is using denitrification as an of a given wastewater under an A2/O configuration with slowly biode­ N2O sink (Chen et al., 2020), which would be clearly favoured by an gradable organic matter. The same individual cases as in section 3.2 are excess of organic matter. discussed to ease the comparison. For a wastewater of 350 mgCOD⋅L− 1 as XS, 5 mgP⋅L− 1 and 30 mgN⋅L− 1, the XS-P and XS-N were 247 and 253 3.5. Analysis of XS requirements: development of general correlations mgCOD⋅L− 1. These values were lower than those in the wastewater (350 mgCOD⋅L− 1) and nutrient removal via the nitrate pathways was All the previous simulations have been conducted assuming that the possible. On the other hand, for a wastewater with 350 mgCOD⋅L− 1 as influent organic matter was readily biodegradable organic matter or XS, 9 mgP⋅L− 1, and 60 mgN⋅L− 1, the XS-P and XS-N were 469 and 605 fermentation products, i.e, VFA. However, the COD fractionation in real mgCOD⋅L− 1, respectively. As observed, the organic matter necessary wastewaters is much more complex and the degree of COD biodegrad­ was higher than that available in the wastewater and nitrite pathway ability is another interesting factor to consider in the previous simula­ implementation would be needed for nutrient removal. The XS-P and XS- tions (Vollertsen and Hvitved-Jacobsen, 2002). PAO can only use N requirements using nitrite pathway were 320 and 264 mgCOD⋅L− 1. 8 À. Gaona et al. Journal of Cleaner Production 414 (2023) 137453 Fig. 7. Minimum XS requirements to meet the legal P and TN discharge limits (PO3− 4 < 1 mgP⋅L − 1 and TN < 10 mg N⋅L− 1) operating under nitrate pathway (left) and nitrite pathway (right) conditions. 1 Fig. 8. Comparison of the predicted XS requirements using the proposed correlations and the model runs for an influent of 5 mg P⋅L− under (left) nitrate pathway and (right) nitrite pathway conditions. 9 À. Gaona et al. Journal of Cleaner Production 414 (2023) 137453 Fig. 9. Decision tree for a continuous A2/O plant to decide whether the implementation of the nitrate or nitrite pathway is required depending on the XS, P and N content of the influent wastewater. XS-nitrate and XS-nitrite stand for the XS requirements when either nitrate or nitrite pathway is selected. These values are lower than that in the wastewater (350 mgCOD⋅L− 1) The operational mode leading to NOB washout was based on DO and nutrient removal through the nitrite pathway was possible. reduction. The model predicts that NOB washout can be stably main­ tained by setting a DO value, although in practice the inherent vari­ ability of operational conditions, as well as the existence of different 3.6. Practical implications NOB microorganisms (with different kinetic characteristics), would make it more difficult. However, there are alternative methods to ach­ The main objective of this study was to obtain general correlations ieve NOB washout, such as sludge treatment by free nitrous acid (FNA) describing the amount of organic matter (SA or XS) that would be and free ammonia (FA) (Duan et al., 2019a, 2019b, 2022). In this case, required, both via nitrate and via nitrite, to remove N and P down to the given that the conditions in the aerobic reactor may be with a higher European default discharge limits. The plant was oversized to avoid oxygen concentration, organic matter requirements can also be reduced kinetic limitations and its operational parameters were selected to be the in the case via nitrite, but the economic results may be less favourable, same over the whole range of influent values evaluated, thus avoiding since, as we have seen, the main saving of the nitrite pathway is due to noise due to additional changes. For example, the internal recirculation the reduction in aeration requirements. ratio was selected at 500% because no additional COD increase at higher Finally, the implementation of the nitrite pathway reduces opera­ RI values was observed for high ammonium concentrations. Also, the tional costs compared to the traditional nitrate pathway, but on the oxygen concentration via nitrate or via nitrite remained the same for all downside N2O emissions, which contribute to global warming, can also simulations. Thus, the results should be considered as the minimum COD be higher. The model consistently predicts higher N2O emissions for the requirements for each scenario, i.e. the lower limit if the plant is well nitrite pathway and, currently, research efforts are set to develop dimensioned and operating correctly. A higher COD concentration operational modes and control strategies to decrease this undesirable might be necessary if, for example, a lower recycle ratio or a higher DO effect (Bellandi et al., 2020; Chen et al., 2020; Santín et al., 2017; Solís concentration is used. et al., 2022a, 2022b). For example, implementing mitigation strategies The findings are specific to the studied configuration (A2/O system), such as optimizing the DO concentration in the aeration tank to reduce and hence extrapolation to other systems such as sequenced batch re­ N2O production and using anoxic periods to enhance N2O denitrification actors, University of Cape Town (UCT), reverse A2/O or other configu­ can decrease N2O emissions. Another solution is to integrate an appro­ rations should be done carefully and would preferably require priate N2O abatement technology for off-gas treatment, such as chemical additional simulations with these configurations. Nevertheless, the ni­ scrubbing or biological N2O reduction. Therefore, it is necessary to trite pathway requires consistently less organic matter than the nitrate develop and implement comprehensive strategies that consider both the pathway, which can help to achieve P or TN discharge limits. 10 À. Gaona et al. Journal of Cleaner Production 414 (2023) 137453 cost savings and environmental impact of the nitrite pathway to ensure Appendix A. Supplementary data sustainable and efficient wastewater treatment. Supplementary data to this article can be found online at https://doi. 4. Conclusions org/10.1016/j.jclepro.2023.137453. The decrease of COD requirements when implementing nitrite References pathway (nitritation/denitritation) for N removal was quantified for a conventional A2/O WWTP aiming at simultaneous C/N/P removal. The Baeza, J.A., Gabriel, D., Lafuente, F.J., 2004. Effect of internal recycle on the nitrogen removal efficiency of an anaerobic/anoxic/oxic (A2/O) wastewater treatment plant organic matter requirements were clearly reduced (depending on the (WWTP). Process Biochem. 39, 1615–1624. https://doi.org/10.1016/S0032-9592 influent, between 9 and 68% less SA and between 23 and 58% less XS). It (03)00300-5. was shown that a higher amount of organic matter as XS was required Bellandi, G., Weijers, S., Gori, R., Nopens, I., 2020. Towards an online mitigation strategy for N2O emissions through principal components analysis and clustering techniques. when compared to SA since the limiting step may be hydrolysis of XS or J. Environ. Manag. 261 https://doi.org/10.1016/j.jenvman.2020.110219. fermentation of SF, which limits PAO activity. General correlations were Changqing, L., Shuai, L., Feng, Z., 2011. The oxygen transfer efficiency and economic developed to help estimate SA and XS requirements in nitrite or nitrate cost analysis of aeration system in municipal wastewater treatment plant. Energy Proc. 5, 2437–2443. https://doi.org/10.1016/j.egypro.2011.03.419. pathway scenarios for the design of future WWTPs. Chen, H., Zeng, L., Wang, D., Zhou, Y., Yang, X., 2020. Recent advances in nitrous oxide Nitrite pathway route has several advantages compared to the nitrate production and mitigation in wastewater treatment. Water Res. 184, 116168 pathway (conventional nitrification-denitrification), such as less aera­ https://doi.org/10.1016/j.watres.2020.116168. Daigger, G.T., Littleton, H.X., 2014. Simultaneous biological nutrient removal: a state-of- tion needs (41-47%), less COD demand, and lower costs (10-16%). the-art review. Water Environ. Res. 86, 245–257. https://doi.org/10.2175/ Therefore, the implementation of the nitrite pathway reduces energy 106143013x13736496908555. requirements and external COD needs, allowing the removal of N in Duan, H., Watts, S., Zheng, M., Wang, Z., Zhao, J., Li, H., Liu, P., Dwyer, J., McPhee, P., wastewaters with low COD/N ratio. Rattier, M., Larsen, E., Yuan, Z., Hu, S., 2022. Achieving robust mainstream nitrite shunt at pilot-scale with integrated sidestream sludge treatment and step-feed. Water On the downside, the model predicts a considerable increase of N2O Res. 223, 119034 https://doi.org/10.1016/j.watres.2022.119034. emissions with nitrite pathway, 2.8–17 times the emission via nitrate for Duan, H., Ye, L., Lu, X., Yuan, Z., 2019a. Overcoming nitrite oxidizing bacteria the scenarios using the minimum COD required and 2.5–6.2 times in adaptation through alternating sludge treatment with free nitrous acid and free ammonia. Environ. Sci. Technol. 53, 1937–1946. https://doi.org/10.1021/acs. scenarios with excess of COD. However, several N2O mitigation strate­ est.8b06148. gies can be proposed to improve this scenario, which can help reduce Duan, H., Ye, L., Wang, Q., Zheng, M., Lu, X., Wang, Z., Yuan, Z., 2019b. Nitrite oxidizing emissions while maintaining the cost savings achieved through the ni­ bacteria (NOB) contained in influent deteriorate mainstream NOB suppression by sidestream inactivation. Water Res. 162, 331–338. https://doi.org/10.1016/j. trite pathway. Then, it is essential to design WWTPs that consider both watres.2019.07.002. the environmental impact and cost savings associated with the use of the EEC Council, 1991. Council Directive of 21 May 1991 Concerning Urban Waste Water nitrite pathway. Treatment (91/271/EEC). Off. J. Eur. Communities. Gernaey, K.V., Jeppsson, U., Vanrolleghem, P.A., Copp, J.B., 2014. Benchmarking of Control Strategies for Wastewater Treatment Plants. IWA Publishing. https://doi. CRediT authorship contribution statement org/10.2166/9781780401171. Guerrero, J., Guisasola, A., Baeza, J.A., 2014. A novel control strategy for efficient biological phosphorus removal with carbon-limited wastewaters. Water Sci. Technol. Àlex Gaona: Methodology, Software, Validation, Formal analysis, 70, 691. https://doi.org/10.2166/wst.2014.280. Investigation, Writing – original draft, Writing – review & editing, Guisasola, A., Qurie, M., Vargas, M., del, M., Casas, C., Baeza, J.A., 2009. Failure of an Visualization. Borja Solís: Methodology, Software, Data curation, enriched nitrite-DPAO population to use nitrate as an electron acceptor. Process Visualization. Javier Guerrero: Methodology, Software, Formal anal­ Biochem. 44, 689–695. https://doi.org/10.1016/j.procbio.2009.02.017. Henze, M., Gujer, W., Mino, T., van Loosdrecht, M.C.M., 2000. Activated Sludge Models ysis, Investigation. Albert Guisasola: Conceptualization, Methodology, ASM1, ASM2, ASM2d and ASM3: Scientific and Technical Report No. 9. IWA Validation, Formal analysis, Resources, Data curation, Writing – review Publishing, London. https://doi.org/10.2166/9781780402369. & editing, Supervision, Project administration, Funding acquisition. Henze, M., van Loosdrecht, M.C., Ekama, G.A., Brdjanovic, D., 2008. Biological Wastewater Treatment : Principles, Modelling and Design. IWA Pub. Juan A. Baeza: Conceptualization, Methodology, Software, Validation, Hiatt, W.C., Grady, C.P.L., 2008. An updated process model for carbon oxidation, Formal analysis, Investigation, Resources, Data curation, Writing – nitrification, and denitrification. Water Environ. Res. 80, 2145–2156. https://doi. original draft, Writing – review & editing, Visualization, Supervision, org/10.2175/106143008X304776. Jeppsson, U., Alex, J., Batstone, D.J., Benedetti, L., Comas, J., Copp, J.B., Corominas, L., Project administration, Funding acquisition. Flores-Alsina, X., Gernaey, K.V., Nopens, I., Pons, M.-N., Rodríguez-Roda, I., Rosen, C., Steyer, J.-P., Vanrolleghem, P.A., Volcke, E.I.P., Vrecko, D., 2013. Declaration of competing interest Benchmark simulation models, quo vadis? Water Sci. Technol. 68, 1–15. https://doi. org/10.2166/wst.2013.246. Kirim, G., McCullough, K., Bressani-Ribeiro, T., Domingo-Félez, C., Duan, H., Al- The authors declare that they have no known competing financial Omari, A., De Clippeleir, H., Jimenez, J., Klaus, S., Ladipo-Obasa, M., Mehrani, M., interests or personal relationships that could have appeared to influence Regmi, P., Torfs, E., Volcke, E.I.P., Vanrolleghem, P.A., 2022. Mainstream short-cut N removal modelling: current status and perspectives. Water Sci. Technol. 85, the work reported in this paper. 2539–2564. https://doi.org/10.2166/wst.2022.131. Liu, X., Kim, M., Nakhla, G., Andalib, M., Fang, Y., 2020. Partial nitrification-reactor Data availability configurations, and operational conditions: performance analysis. J. Environ. Chem. Eng. 8, 103984 https://doi.org/10.1016/j.jece.2020.103984. Massara, T.M., Malamis, S., Guisasola, A., Baeza, J.A., Noutsopoulos, C., Katsou, E., Data will be made available on request. 2017. A review on nitrous oxide (N2O) emissions during biological nutrient removal from municipal wastewater and sludge reject water. Sci. Total Environ. 596 (597), Acknowledgements 106–123. https://doi.org/10.1016/j.scitotenv.2017.03.191. Massara, T.M., Solís, B., Guisasola, A., Katsou, E., Baeza, J.A., 2018. Development of an ASM2d-N2O model to describe nitrous oxide emissions in municipal WWTPs under This work was supported by Grant CTQ2017-82404-R funded by dynamic conditions. Chem. Eng. J. 335, 185–196. https://doi.org/10.1016/j. MCIN/AEI/10.13039/501100011033 and by ERDF A way of making cej.2017.10.119. Oehmen, A., Lemos, P.C., Carvalho, G., Yuan, Z., Keller, J., Blackall, L.L., Reis, M.A.M., Europe, and Grant PID2020-119018RB-I00 funded by MCIN/AEI/ 2007. Advances in enhanced biological phosphorus removal: from micro to macro 10.13039/501100011033 and by European Union NextGenerationEU/ scale. Water Res. 41, 2271–2300. https://doi.org/10.1016/j.watres.2007.02.030. PRTR. The authors are members of the GENOCOV research group (Grup Pasini, F., Garrido-Baserba, M., Sprague, T., Cambiaso, P., Rosso, D., 2021. Quantification of energy and cost reduction from decreasing dissolved oxygen levels de Recerca Consolidat de la Generalitat de Catalunya, 2021 SGR 00515, in full-scale water resource recovery facilities. Water Environ. Res. 1–13 https://doi. www.genocov.com). org/10.1002/wer.1660. 11 À. Gaona et al. Journal of Cleaner Production 414 (2023) 137453 Pocquet, M., Wu, Z., Queinnec, I., Spérandio, M., 2016. A two pathway model for N2O Tayà, C., Garlapati, V.K., Guisasola, A., Baeza, J.A., 2013. The selective role of nitrite in emissions by ammonium oxidizing bacteria supported by the NO/N2O variation. the PAO/GAO competition. Chemosphere 93, 612–618. https://doi.org/10.1016/j. Water Res. 88, 948–959. https://doi.org/10.1016/j.watres.2015.11.029. chemosphere.2013.06.006. Regmi, P., Miller, M.W., Holgate, B., Bunce, R., Park, H., Chandran, K., Wett, B., Tchobanoglous, G., Stensel, H.D., Tsuchihashi, R., Burton, F., Abu-Orf, M., Bowden, G., Murthy, S., Bott, C.B., 2014. Control of aeration, aerobic SRT and COD input for Pfrang, W., 2014. Metcalf & Eddy. Wastewater Engineering. Treatment and Resource mainstream nitritation/denitritation. Water Res. 57, 162–171. https://doi.org/ Recovery, fifth ed. McGray-Hill Education. 10.1016/j.watres.2014.03.035. Turk, O., Mavinic, D.S., 1986. Preliminary assessment of a shortcut in nitrogen removal Santín, I., Barbu, M., Pedret, C., Vilanova, R., 2017. Control strategies for nitrous oxide from wastewater. Can. J. Civ. Eng. 13, 600–605. https://doi.org/10.1139/l86-094. emissions reduction on wastewater treatment plants operation. Water Res. 125, Vargas, M., Guisasola, A., Artigues, A., Casas, C., Baeza, J.A., 2011. Comparison of a 466–477. https://doi.org/10.1016/j.watres.2017.08.056. nitrite-based anaerobic–anoxic EBPR system with propionate or acetate as electron Solís, B., Guisasola, A., Flores-Alsina, X., Jeppsson, U., Baeza, J.A., 2022a. A plant-wide donors. Process Biochem. 46, 714–720. https://doi.org/10.1016/j. model describing GHG emissions and nutrient recovery options for water resource procbio.2010.11.018. recovery facilities. Water Res. 215, 118223 https://doi.org/10.1016/j. Vollertsen, J., Hvitved-Jacobsen, T., 2002. Biodegradability of wastewater - a method for watres.2022.118223. COD-fractionation. Water Sci. Technol. 45, 25–34. https://doi.org/10.2166/ Solís, B., Guisasola, A., Pijuan, M., Baeza, J.A., 2022b. Exploring GHG emissions in the wst.2002.0046. mainstream SCEPPHAR configuration during wastewater resource recovery. Sci. Xu, G., Zhang, Z., Gao, F., 2021. Effect of COD/N ratios and DO concentrations on the Total Environ. 849, 157626 https://doi.org/10.1016/j.scitotenv.2022.157626. NOB suppression in a multi-cycle SBR. J. Environ. Chem. Eng. 9, 105735 https://doi. Solís, B., Guisasola, A., Pijuan, M., Corominas, L., Baeza, J.A., 2022c. Systematic org/10.1016/j.jece.2021.105735. calibration of N2O emissions from a full-scale WWTP including a tracer test and a Zhou, Y., Ganda, L., Lim, M., Yuan, Z., Kjelleberg, S., Ng, W.J., 2010. Free nitrous acid global sensitivity approach. Chem. Eng. J. 134733 https://doi.org/10.1016/j. (FNA) inhibition on denitrifying poly-phosphate accumulating organisms (DPAOs). cej.2022.134733. Appl. Microbiol. Biotechnol. 88, 359–369. https://doi.org/10.1007/s00253-010- Spérandio, M., Pocquet, M., Guo, L., Ni, B.J., Vanrolleghem, P.A., Yuan, Z., 2016. 2780-3. Evaluation of different nitrous oxide production models with four continuous long- Zhou, Y., Oehmen, A., Lim, M., Vadivelu, V., Ng, W.J., 2011. The role of nitrite and free term wastewater treatment process data series. Bioproc. Biosyst. Eng. 39, 493–510. nitrous acid (FNA) in wastewater treatment plants. Water Res. 45, 4672–4682. https://doi.org/10.1007/s00449-015-1532-2. https://doi.org/10.1016/j.watres.2011.06.025. Stare, A., Vrečko, D., Hvala, N., Strmčnik, S., 2007. Comparison of control strategies for Zou, X., Zhou, Y., Gao, M., Yang, S., Mohammed, A., Liu, Y., 2022. Effective N2O nitrogen removal in an activated sludge process in terms of operating costs: a emission control during the nitritation/denitritation treatment of ammonia rich simulation study. Water Res. 41, 2004–2014. https://doi.org/10.1016/j. wastewater. J. Environ. Chem. Eng. 10, 107234 https://doi.org/10.1016/j. watres.2007.01.029. jece.2022.107234. 12