Full-scale implementation of coagulant dosing for recalcitrant nitrogen and orthophosphate control during dewatering of thermal hydrolysis pretreatment-enhanced anaerobic digester sludge

Introduction In 2024, WSSC Water, the largest water/wastewater utility in Maryland, commissioned a centralized biosolids processing facility (Bioenergy Facility), which utilizes a thermal hydrolysis pretreatment (THP)-enhanced anaerobic digester (AD) process. The Bioenergy Facility will receive biosolids from six regional Water Resource Recovery Facilities (WRRFs), which have a combined treatment capacity of 100 MGD. The filtrate from the Bioenergy Facility's dewatering process is treated using the AnitaMox deammonification process before being returned to the headworks of the 30 MGD Piscataway WRRF. Bench-scale evaluations estimated that the total nitrogen (TN) in Piscataway's effluent could increase by up to 1 mg/L due to the production of recalcitrant dissolved organic nitrogen (rDON) via the Maillard reaction resulting from the THP process as detailed in our previous study [1]. The anticipated increase in rDON loading will place the Piscataway WRRF at risk of exceeding the stringent TN effluent limit of 3.25 mg/L. In addition, operational changes at two of the biosolids source WRRFs, which plan to shift to biological phosphorous removal, are expected to result in higher orthophosphate (ortho-P) presence is in the Bioenergy Facility's AD effluent and dewatering filtrate, potentially causing struvite formation in downstream equipment. To address these issues, a temporary, easy-to-implement approach, that can be applied at this and other WRRFs operating a THP-AD process, was tested in pilot-scale to control rDON and ortho-P. A previous study reported that rDON mainly contains negatively-charged molecules, which have a remarkable coagulation capacity with many metallic ions, such as ferric chloride (FeCl3) [2, 3]. Phase 1 of this study evaluated the influence of FeCl3 addition during dewatering of AD effluent on filtrate rDON and ortho-P concentrations at lab scale. Results from these lab-scale experiments reported herein demonstrate the strong potential of FeCl3 for reducing rDON and ortho-P concentrations in AD filtrate. Phase 2 of the study will focus on validating these lab-scale results by conducting a second round of testing at full-scale. A comparison of the lab-scale and full-scale results will be reported, alongside an assessment of the practicality and effectiveness of FeCl3 addition for rDON and ortho-P control at this centralized biosolids processing facility. Materials and Methods The process flow diagram of the study is shown in Figure 1. Samples of dewatered mixed sludge cake with an approximately 22% total solids (TS) content were collected from WSSC Water's four largest WRRFs (Piscataway, Western Branch, Seneca, and Parkway) with a blending ratio (wet mass) of 4.3 : 3.3 : 2.4 : 1 that is based on an estimated future biosolids production. The blended cake samples were diluted to around 16% TS followed by processing through a pilot-scale Cambi THP unit (Cambi, Asker, Norway) operated at 6 bar steam pressure (equivalent to 165 °C) for 30 minutes. The pretreated sludge was then mixed with AD inoculum collected from a full-scale THP-AD system at a local WRRF and anaerobically incubated at 36.5 °C until the cumulative biogas production reached a plateau. After digestion, the optimal polymer dose (OPD) of the digested sludge was determined based on capillary suction time (CST) [4]. The lab-scale dewatering protocol, developed by VT-CAWRI, includes shearing of the sludge-polymer mixture and piston compression of the solids, which has been calibrated to a full-scale belt filter press installation according to methods detailed in a previous study [5]. Two variables were investigated in this study: FeCl3 dose and polymer dose. Briefly, FeCl3 solutions were added to the digested sludge during dewatering at doses of 0.5%, 1.0%, 2.0%, 2.7%, and 5.5% Fe (g Fe3+/g DS) along with polymer dosing at either OPD or 1/2 OPD. The intent for varying polymer dose is to understand whether FeCl3 can compensate part of the role played by polymer during dewatering, ultimately saving the cost of polymer consumption. Samples of dewatering filtrate were collected for characterization and recalcitrant species tests. The recalcitrant species test, developed by VT-CAWRI, consisted of ammonia (TAN) stripping to minimize background noise when calculating DON = sTN - NOx - TAN and subsequent aerobic incubation to eliminate biodegradable DON, with the remainder treated as rDON which ultimately contributes to Piscataway WRRF's effluent. Results and Discussion Effects on rDON removal Without FeCl3 addition, around 24.1% and 9.4% of rDON were removed as part of the biosolids during dewatering with OPD and 1/2 OPD after accounting for polymer dilution of the sludge, leaving nearly 76% and 91% of rDON in the filtrate, respectively (Figure 2a). As the FeCl3 dose increases, a linear reduction in filtrate rDON was observed, with a nearly 60% removal at a 5.5% FeCl3 dose alongside 1/2 OPD (Figure 2b). Previous studies have reported that rDON produced from the THP process mainly consists of melanoidin molecules, which are negatively charged nitrogen-containing byproducts of the Maillard reaction that have shown remarkable binding capacity with metallic ions [6]. Therefore, the removal mechanism of rDON in this tested approach was believed to be precipitation or coagulation by ferric ions and cationic polymer. Although both polymer and FeCl3 demonstrated rDON removal capability according to Figure 2a, the contribution from FeCl3 was greater since polymer alone could only achieve a maximum removal of 24.1% at the OPD. Effect on ortho-P removal Unlike rDON removal, the removal of ortho-P without any FeCl3 addition was minimal. Only less than 5% of ortho-P was removed from the liquid phase at OPD, and even no change was observed with 1/2 OPD of polymer during dewatering (Figure 3a). However, it is notable that even the lowest dose of FeCl3 (i.e., 0.5% Fe) combined with 1/2 OPD of polymer achieved an over 85% removal of ortho-P from the filtrate, while dosing 2% Fe resulted in nearly complete elimination (Figure 3a). This exponential reduction of ortho-P in Figure 3b indicates that FeCl3 dosing during dewatering is more efficient for ortho-P removal than for rDON removal, making rDON levels in the filtrate likely the determining factor for appropriate FeCl3 dose selection. Moreover, the effective removal of ortho-P from the dewatering filtrate essentially eliminated the risk of struvite formation in downstream equipment. Effect on sludge dewatering and cation overdosing It is our hypothesis that the addition of metallic coagulants during dewatering along with the optimal dose of polymer may lead to cation overdosing, which disrupts charge neutralization and ultimately results in worse cake dryness. To test this, cake dryness at full and half OPD, with and without coagulant addition, was analyzed. Without any FeCl3 addition, results in Figure 4 showed exceptional cake dryness of 33.1% TS at OPD and a deteriorated cake dryness of 25.3% TS at 1/2 OPD due to underdosing of cations. Adding 5.5% Fe on top of OPD substantially aggravated cake dryness to less than 25% TS (Figure 4). However, in addition to 1/2 OPD, dosing 1% Fe could recover the cake dryness to as high as 31.6% TS from the 25.3% TS at 1/2 OPD only (Figure 4). The results suggest that the deterioration in cake dryness as a result of overdosing of cations during dewatering can be solved by cutting back polymer usage while not sacrificing the superiority in the removal of rDON and ortho-P by FeCl3. Effect on pH and alkalinity It is our concern that the drop in pH and alkalinity at high FeCl3 doses may negatively impact the anammox-based sidestream deammonification process in full-scale operation. As shown in Figure 5, consistent decreases in filtrate pH and alkalinity as a result of FeCl3 dosing during dewatering was observed due to hydrogen ion release from the formation of metal-hydroxide complexes [7]. Without extra alkalinity supplementation, the tradeoff between effective rDON removal and decreases in pH and alkalinity becomes a crucial consideration for selecting FeCl3 doses in full-scale operations. It is important to ensure a sufficient alkalinity for deammonification and a suitable pH range of 6.5 to 7.5 for AOB and anammox bacterial activities [8, 9]. Conclusions and Ongoing Work The results from this pilot study revealed that FeCl3 dosing along with polymer during digestate dewatering can (1) reduce rDON by nearly 60% at 5.5% Fe, (2) achieve nearly 100% ortho-P at 2% Fe, and (3) maintain excellent cake dryness while reducing polymer dosing by 50%. These promising findings provide strong proof of concept, which is currently being validated in full-scale operations at the Bioenergy Facility through early 2025. Potential side effects, such as decreased pH and alkalinity will also be verified and addressed.