A Case for Using Internally Stored Carbon for Intensifying Biological Nutrient Removal – Mechanism, Evidence in Full-Scale Facilities, and Process Design

Introduction
Total nitrogen removal has historically been limited by the amount of reducing electrons in wastewater to perform denitrification. Most wastewater resource recovery facilities (WRRFs) use an external carbon source such as methanol or acetic acid to provide the electron donor required for denitrification. However, this can be expensive especially with stringent nitrogen limits. For this reason, there has been an increased interest in using internally stored carbon for denitrification. Cycling of internally stored carbon is a key mechanism that has been successfully used for removing phosphorus in the enhanced biological phosphorus removal (EBPR) process. However, a complete understanding of production and usage of internally stored carbon could provide benefits beyond EBPR to also perform total nitrogen removal and biological nutrient removal (BNR).
Mechanism
It has been recognized that bacteria adopt a survival strategy to tackle feast-famine cycles. Figure 1 shows the conceptual metabolic pathway for PHA accumulation. The storage of PHA is largely triggered by the presence of excess (feast - beyond required for growth) soluble substrate such as VFAs and energy (ATP) (Beun et al., 2000; Third et al., 2003). The energy required can be produced through aerobic/anoxic respiration (PHA-accumulating OHOs) or through hydrolysis of polyphosphate (in the case of polyphosphate accumulating organisms - PAOs). Subsequently, when there is a lack of excess VFAs (famine), the PHA produced can either be used for P-uptake under the presence of oxygen (in the case of PAOs) and/or be used for denitrification under carbon-limited conditions by other PHA-accumulating heterotrophs (Basset et al., 2016; Vocks et al., 2005; Zhou et al., 2012). Extended aerobic conditions can result in a depletion of PHA since PHA will also be used for aerobic respiration and growth under famine. However, with the advent of low DO operating strategies, there is potential to minimize PHA utilization for growth and maximize utilization for denitrification. The other carbon storage polymer that has been implicated in providing denitrification potential is glycogen (Coats et al., 2011). In a BNR process, glycogen is typically used up under anaerobic or feast conditions to generate reducing power to uptake volatile fatty acids (VFAs) and form PHA while under aerobic famine conditions, glycogen is regenerated as a by-product of bacterial growth through the use of PHA. However, unlike PHA which is degraded under famine/aerobic conditions, longer famine/aerobic periods will increase glycogen storage. So a key difference between using PHA versus glycogen for denitrification in post-anoxic or microaerobic conditions is that extended aerobic periods prior to denitrification can deteriorate denitrification potential with PHA as an electron donor versus maximizing denitrification potential with glycogen as an electron donor.
Evidence in Full-Scale Facilities
There have been several recent observations of post-anoxic denitrification without any external carbon addition. Bauhs (2021) showed that approximately 33% of NOx removal in the post-anoxic zone of the VIP+2 process at the Virginia Initiative Plant (VIP) could be attributed to a 'mysterious' denitrification after accounting for denitrification due to methanol and endogenous decay. Furthermore, ex-situ denitrification tests showed that extending the aeration period before the anoxic phase (without addition of carbon) reduced the specific denitrification rates (SDNR) which would implicate PHA as the primary carbon store for post-anoxic denitrification. The Rochester WRP, MN consists of a two-stage high-purity oxygen plant and an activated sludge plant in parallel. A demonstration was conducted at the activated sludge plant which has the ability to be operated in an AO or A2O mode. Low DO operation was achieved by setting DO setpoints for Zone 2 and adjusting airflows proportionally to the other aerated zones (Figure 2). During the course of the demonstration study several strategies for improving total inorganic nitrogen (TIN) removal were tested including the addition of a mixed liquor recycle (MLR) at 1Q and the implementation of a post-anoxic zone (Zone 3) by switching off air and only turning it on twice a day to enable mixing. Without the post-anoxic zone, the process was able to achieve an effluent TIN of ~15 mg-N/L at a DO setpoint of 0.2 mg/L (Figure 3). However, when the post-anoxic zone was implemented, the TIN concentrations at the end of the BNR process decreased from 15 mg-N/L to <10 mg-N/L suggesting the availability of unused carbon (slowly biodegradable COD or internally stored carbon) for denitrification. Profiling for PHA and glycogen and batch activity tests are being conducted to elucidate the source of electrons for denitrification. At the Pueblo CO WRF, comprising a modified Johannesburg configuration with low DO controls and physical selector for sludge densification, a 30±6% increase in specific denitrification rates from mixed liquor entering the aerobic zone has been observed when acetate is fed to the anaerobic zone for feast-famine control (Figure 4). Carbon profiling through the anaerobic and anoxic zone indicates complete utilization of the acetate prior to the aerobic zone suggesting an increase in internally stored carbon by storage-product accumulating heterotrophic bacteria as a possible pathway for denitrification (Figure 4).
Process Design
An ideal process design for taking advantage of internal carbon stores (particularly PHA) will take into account the metabolic and regulatory processes that influent PHA accumulation and utilization while accounting for wastewater engineering considerations. Studies have shown that PHA storage is possible in PHA-accumulating OHOs under anoxic and aerobic conditions and the PHA yield is higher under low O2/NO3- supply rates compared to higher supply rates (Beun et al., 2000; Çiǧgin et al., 2009; Third et al., 2003). A conceptual framework for process design for PHA-driven denitrification under low DO operation is shown in Figure 5. Figure 6 shows a conceptual framework for why high DO operation is counter-productive to utilization of PHA for denitrification. The full presentation will discuss important features of the mechanism of internal carbon storage and subsequent use for denitrification along with discussions of the microbial ecology of internal carbon storing bacteria. We will also provide evidence of utilization of carbon stores for denitrification using data from multiple facilities that are part of an ongoing Water Research Foundation study (WRF 5083) which will include profiling of glycogen and PHA along with results of batch activity tests. Finally, we will discuss potential design and operational considerations for carbon management and optimal usage of internal carbon storage for BNR.