Mechanistic Understanding of the Kinetic Difference Between the Methanol and Glycerol-driven Partial Denitrification Anammox in Low Nitrogen Polishing Moving Bed Biofilm Reactors

Introduction Partial denitrification coupled with anaerobic ammonia oxidation (PdNA) process has been identified as a promising approach to intensify biological nutrient removal systems to increase total nitrogen removal capacity while reducing chemical and energy demand (Campolong, 2019; Le et al., 2019; Zhang et al., 2020). Understanding how operational factors impact PdNA is necessary to extract optimal performance at water resource recovery facilities (WRRFs) regardless of the configuration employed. For example, work performed previously (Campolong, 2019; Le et al., 2019) has shown that PdNA is impacted by multiple factors including the type of carbon utilized e.g., methanol, glycerol, acetate, etc.). This prior work hypothesized that partial denitrification results due to two main mechanisms, i) rate differential between heterotrophic nitrate and nitrite reduction due to different substrate availability and carbon sources, ii) nitrite sink due to anammox activity (Al-Omari et al., 2021). This current study provides insights into understanding how these different mechanisms may dictate partial denitrification in a low nitrogen polishing environment. Data from long term operation (417 days) of two pilot-scale MBBR treatment trains treating secondary effluent from the Noman M. Cole Jr. Pollution Control Plant (NCPCP) was used to develop a mechanistic understanding of methanol versus glycerol driven PdNA performance with a view to informing how to scale PdNA into other WRRFs. This work also provides a framework for informing design practices as related to stimulating stable PdNA. Materials and methods Reactor setup and operation: Two parallel MBBR treatment trains dosed with methanol and glycerol, respectively, were operated on site at NCPCP. In each treatment train, two 14 L working volume anoxic MBBRs filled with K1 media at a volumetric fraction of 45% were connected in series, followed by an 8 L working volume reaeration reactor filled with K1 media at a volumetric fraction of 32% (configuration simulated full-scale MBBR trains at NCPCP). K1 media with established biofilms of heterotrophs and nitrifiers from the full-scale MBBR train were used to seed the pilots. The pilots were operated for 417 days with a focus on demonstrating the ability to achieve effluent nutrient limits while reducing supplemental carbon needs. A description of performance is provided in (Sun, 2022; WANG, 2022). Model development: Kinetic analysis of PdNA was investigated by developing expressions describe denitratation (NO3--N → NO2--N) and denitritation (NO2--N → N2) (Glass and Silverstein, 1998). The specific nitrate uptake rate of denitratation culture (qDenitra.) can be expressed as Equation 1 according to Activated Sludge Model NO. 1 (Henze et al., 2000). Similarly, the specific nitrite uptake rate of denitritation culture (qDenitri.) can also be expressed as Equation 2. Equation 3 shows the specific nitrite uptake rate of anammox (AMX) bacteria (Bi et al., 2015). qAMX/qDenitri., qAMX/qDenitra. and qDenitra/qDenitri. as shown in Equations 4, 5 and 6 were used to compare effects of conditions on specific substrate uptake rates with a view to identifying which process might govern performance. All the kinetic parameters used in the investigation are listed in Table 1. Results and discussion Pilot results: Methanol-driven PdNA achieved an effluent TIN concentration of 2.6 +/- 0.6 mg/L and observed an average of 1.3 mg/L anoxic NH4+-N removal. Glycerol-driven PdNA achieved an effluent TIN concentration of 2.7 +/- 1.0 mg/L and observed an average of 1.3 mg/L anoxic NH4+-N removal. Results confirmed it was possible to achieve nutrient targets while also reducing supplemental carbon by 20 to 30% relative to conventional BNR. Nitrite sink by AMX governed PdNA performance in methanol fed train: Data from pilot operation were overlaid on heat maps plotting specific utilization rate ratios of denitratation, denitritation and AMX (Figures 1, 2 and 3). Results indicated that the in situ anammox substrate utilization rate (qAMX) was likely greater than the denitritation utilization rate (qDenitri) for the entire 372 days of operation period. Additionally, it was observed that anammox substrate utilization and denitratation utilization rates were similar i.e., qAMX/qDenitra ~ 1. Collectively, these results suggest that heterotrophic denitritation was outcompeted by anammox for nitrite in the methanol system. This implies that nitrite sink from anammox governed PdNA performance. Rate differential for heterotrophic denitratation versus denitritation governed PdNA performance in glycerol fed train: In contrast to the methanol train, pilot data from the glyercol train suggested that qAMX was less than qDenitri for the majority of pilot operation. Further, it was observed that qAMX was less than qDenitra while qDenitra. was greater than qDenitri. This data implies that the rate differential for heterotrophic nitrate and nitrite reduction dictated PdN performance in the glycerol train. Significance of Results The approach described in this paper can be used to help improve our understanding of existing PdNA systems to help clarify how different conditions may impact which mechanisms dictate performance.