Cultivation of Nitrifying and Nitrifying-Denitrifying Aerobic Granular Sludge for Sidestream Treatment of Anaerobically Digested Sludge Centrate

Abstract

In this study, three 1.2-L aerobic granular sludge sequencing batch reactors (AGS-SBRs) were used to cultivate nitrifying and nitrifying-denitrifying granules (w/supplemental carbon) and investigate sidestream treatment of synthetic-centrate and real-centrate samples from Ashbridges Bay Treatment Plant (ABTP) in Toronto, Ontario, Canada. Results showed that although the cultivation of distinct granules was not observed in the nitrifying reactors, sludge volume index (SVI₃₀) values achieved while treating real and synthetic centrate were 72 ± 12 mL/g and 59 ± 11 mL/g (after day 14), respectively. Ammonia-nitrogen (NH₃-N) removal in the nitrifying SBRs were 93 ± 19% and 94 ± 16% for real and synthetic centrate, respectively. Granules with a distinct round structure were successfully formed in the nitrifying-denitrifying SBR, resulting in an SVI₃₀ of 52 ± 23 mL/g. NH₃-N, chemical oxygen demand (COD) and phosphorus (P) removal in the nitrifying-denitrifying SBR were 92 ± 9%, 94 ± 5%, and 81 ± 14% (7th to 114th day), respectively with a low nitrite (NO₂-N) and nitrate (NO₃-N) concentration in the effluent indicating simultaneous nitrification-denitrification (SND) activity. High nutrient removal efficiencies via the nitrification and SND pathways shows that AGS technology is a viable process for treating sidestreams generated in a WWTP.

1. Introduction

Wastewater treatment plants (WWTPs) face the challenge of implementing sustainable unit processes that enhance nutrient removal and contribute to the recovery of energy and value-added products [1]. In recent years, technologies that permit sustainable wastewater management and solids handling are being increasingly employed worldwide [2]. Anaerobic digestion (AD) is a widely used technology for solids handling that utilizes biochemical processes to facilitate the degradation of organic matter (OM) by a consortium of bacteria under anaerobic conditions. The process of anaerobic digestion also leaves behind digestate (or anaerobically digested sludge) which is a liquid slurry of microorganisms, digested substrates, and the carrier medium [3]. Dewatering anaerobically digested sludge generates a liquid sidestream–commonly referred to as centrate or reject water–with elevated concentrations of N (predominantly NH₃-N) and P. In some installations, AD is preceded by thermal pre-treatment which increases the concentration of NH₃-N up to 2500 mg/L in centrate [4,5]. Centrate can consists of 0.5–1% of mainstream flow, 10–30% of the plant’s influent nitrogen load and 20–40% of the plant’s phosphorus load [6]. WWTPs predominantly return centrate to the mainstream process for treatment, but the additional nutrient loading negatively impacts removal efficiency. Returned centrate can result in shock loads, process instability and increased operation costs [7]. The inclusion of separate sidestream treatment can improve the performance and reliability of a WWTP [8].

Centrate can be adequately treated in a separate sidestream process using various wastewater technologies that belong to mainly two categories, (1) biological, and (2) physical/chemical [9]. Ammonia stripping is a physical/chemical treatment technology which has been used to remove 80–90% of nitrogen from digester supernatant using air at full scale [10]. Alternatively, ammonia can be recovered as an ammonia salt solution through acid stripping. NH₃-N concentration in a feed stream was drastically reduced from 21,006 to 10.6 mg/L (99.95% NH₃-N removal) at a normal temperature of 25 °C using acid stripping [11]. Ion exchange is another physical/chemical treatment technology that can achieve high ammonia removal from a liquid waste stream [12]. Target ions present in the liquid/aqueous phase such as ammonium ion (NH₄⁺) are capable of displacing similar cations on the exchange resin; resulting in nutrient removal [7]. Digested sludge liquor with an ammonia concentration of >600 mg/L was successfully treated using a clay-based ion exchange medium (MesoLite), resulting in >95% ammonia removal [13].

Biological treatment is an inexpensive alternative compared to physical/chemical processes. Conventional activated sludge (CAS) is the most widely applied biological treatment technology. In CAS systems, nitrogen is removed from wastewater via biological nitrification and denitrification [14]. The first step, nitrification, occurs under aerobic conditions, where NH₃-N is oxidized to NO₂-N by ammonia oxidizing bacteria (AOB) and further to NO₃-N by nitrite oxidizing bacteria (NOB) [15]. In the denitrification step, organic loading in the influent or supplemental organic carbon (e.g., methanol) acts as an electron donor for the conversion of NO₃-N to dinitrogen gas (N₂) by heterotrophic denitrifying bacteria [9,16].

A submerged activated growth bioreactor (SAGB) that was fed centrate and operated with a nitrogen loading between 0.54 to 1.51 kg-N/m³·d, achieved stable NH₃-N oxidation and a nitrification efficiency of 98% according to Pedros et al. [17]. Furthermore, treatment of co-digested piggery/poultry manure and agro-waste sludge liquor from a full-scale anaerobic digester was investigated by Scaglione et al. [18] in a pilot-scale SBR for partial-nitritation (PARNIT). In conjunction with a bench-scale SBR that was inoculated with anaerobic ammonium oxidation (anammox) granular sludge and fed with a combination of PARNIT effluent and synthetic wastewater, an average nitrogen removal of 91 ± 10% was achieved Scaglione et al. [18]. The anammox step allows for the conversion of NH₃-N to N₂ using NO₂-N as the electron acceptor instead of supplemental carbon source, leading to considerable cost savings [19]. Typically, processes that utilize a shortcut route to nitrogen removal require less oxygen and biodegradable carbon [8,20]. Nutrient-rich centrate typically comprises of low biodegradable COD and insufficient alkalinity-both of which are required for biological treatment [9].

Similar to the above-mentioned technologies, aerobic granular is a fitting technology when attempting to treat a sidestream such as centrate with fluctuations in the NH₃-N concentration and limited biodegradable COD [21]. An earlier study on formation of nitrifying granules was investigated by Liu et al. [22] using acclimated sludge and synthetic inorganic wastewater in an SBR. Nitrifying granules were reportedly formed within 21 days with a size of 240 μm and the biomass displayed a sludge volume index (SVI) of 40 mL/g [22]. Both nitrification and partial-nitrification was reported, with the authors stating that the system was negatively impacted by changing influent NH₃-N concentrations (100 to 450 mg/L) [22]. López-Palau et al. [19] also worked with anaerobic digestion supernatant to form nitrifying (NIT) granules in an SBR. A conventional SBR with suspended growth biomass was operated for comparison with the characteristics of aerobic granules. Granules with a distinct and slightly spherical shape were formed with the diameter ranging from 2 to 5 mm [19]. In terms of performance, partial nitrification was achieved in both systems with the NO₂-N/NH₄⁺ ratio equating to approximately 1 in the effluent–mainly due to insufficient alkalinity in the system which induced pH suppression, thus resulting in partial nitrification [19,23].

Aerobic granular sludge (AGS) has also gained interest for its robustness and ability to maintain performance in adverse conditions [24]. The impact of COD/N ratio is a topic that is being continuously investigated by researchers worldwide. After investigating different COD/N ratios using ammonia-rich synthetic wastewater, Wei et al. [25] reportedly achieved the highest nitrogen removal at a ratio of 9. Kocaturk and Erguder [26] recommended a COD/N ratio of 7.5–30 for stable granules that favor the growth of heterotrophs and 2–5 for slow growing nitrifying granules. A recent investigation by Figdore et al. [27] pursued the growth of nitrification-denitrification (NDN) and phosphorus accumulating organisms (PAOs) granules from sidestream centrate. Growth of NDN-PAO granules were successful as reported by the author and removal efficiencies were 95% and 88% for NH₃-N and total nitrogen (TN), respectively [27]. The application of AGS to centrate–containing high nutrient load and low biodegradable COD—is practical through the well-known nitrification-denitrification reaction pathway. Additionally, the microbial orientation of AGS offers the potential for multiple redox reactions across the layers of a granules in a single system [28].

This research aims to evaluate the application of AGS–to separate sidestream treatment of centrate. The experimental work explored formation of nitrifying granules with and without an organic carbon source and assessed the treatment of centrate at a low COD/N ratio. The objectives of this study were to: (1) Compare the cultivation aerobic granular sludge and treatment of diluted centrate via nitrification only and nitrification-denitrification reaction pathways, (2) determine the impact of increasing nitrogen (N) loading and a low COD/N (i.e., <5) ratio on nitrification and/or nitrification-denitrification, and (3) determine the rate of nitrification with respect to aeration time and volatile suspended solids (VSS) concentration and the efficiency of nitrification-denitrification. First, the cultivation of nitrifying granules without the addition of external carbon was investigated using real centrate and synthetic wastewater. Secondly, the cultivation nitrifying-denitrifying granules were investigated using synthetic wastewater with the addition of an external carbon source follow by the treatment of real centrate with supplemental carbon. Pollutant degradation was monitored during a designated cycle to determine nitrification rate (NR) and specific nitrification rate (SNR). Simultaneous nitrification-denitrification efficiency was computed using influent and effluent data. The effect of a low COD/N was also investigated during the treatment of centrate (i.e., after granule cultivation using synthetic organic wastewater).

2. Materials and Methods

2.1. Reaction Configuration and Operation

The experimental study was performed in three column-type SCH 40 PVC reactors—AGS-SBR-1, AGS-SBR-2, and AGS-SBR-3—with a 5-cm internal diameter, height of 70 cm and a total volume of 1.38-L. The reactors were operated with a working volume of 1.2-L, equivalent to a working height of 61 cm and height-to-diameter (H/D) ratio of 12.2. Aeration was provided by fine air bubble diffusers located at the bottom of the reactors and flow meters were used to control the airflow at 3 L/min, resulting in a superficial air velocity (SAV) of 2.55 cm/s. Air flow in AGS-SBR-1 was reduced to 1 L/min (SAV = 0.8 cm/s) during the final period (i.e., days 162–183) of the reactor’s operation. The AGS-SBRs were designed to accommodate static feeding from the bottom of each reactor and upwards through the bed of settled biomass. The decant ports were conveniently located at the midpoint of each reactor column (i.e., at the 0.6-L mark), providing a volume exchange ratio (VER) of 50%. Peristaltic pumps were used to feed the reactors and discharge treated effluent. The reactors were operated sequentially on a 6-h cycle followed by a 4-h cycle to allow for a reduction in hydraulic retention time (HRT) as displayed in

Table 1. Cycle time and sequential phases for the investigated AGS-SBRs.

	AGS-SBR-1		AGS-SBR-2		AGS-SBR-3
Phase	6-h	4-h	6-h	4-h	6-h	4-h
Influent Filling (min)	15	15	15	15	60	60
Aeration (min)	265–315	160–195	265–315	190–195	220–270	145–150
Settling (min)	10–60	10–45	10–60	10–15	10–60	10–15
Effluent discharge (min)	15	15	15	15	15	15
Idle (min)	5	5	5	5	5	5

. The duration of influent filling, effluent discharge, and idle remained constant throughout the experimental study. AGS-SBR-3 was operated with a 60-min anoxic filling phase to stimulate denitrification. Settling times were decreased during the granule cultivation stage with the excess time added to the aeration phase.

The operating parameters for each reactor are shown in

Table 2. Summary of operating conditions for aerobic granular sludge SBR systems.

Parameter	Unit	AGS-SBR-1	AGS-SBR-2	AGS-SBR-3
Operating Duration	days	183	72	212
pH	-	7.8 ± 0.5	8 ± 0.5	7.8 ± 0.3
Temperature	°C	22–25	22–25	22–25
VER	%	50	50	50
DO	mg/L	>2 (non-limiting)	>2 (non-limiting)	>2 (non-limiting)

. The experiments were conducted at ambient temperature (22–25 °C). A common peristaltic pump was used to dose an 80 g/L solution of sodium bicarbonate (NaHCO₃) to AGS-SBR-1 and AGS-SBR-2 for pH control. In AGS-SBR-3, pH was controlled using a 1 M hydrochloric acid (HCl) solution and an Etatron DLX series metering pump attached to a probe. Dissolved oxygen (DO) was considered >2 mg/L or non-limiting since air flow was controlled by a flow meter.

Operational periods for each reactor are displayed in

Table 3. Summary of reactor operational parameters for AGS-SBR-1 & 2.

Reactor	Period	Operation Days	Influent NH3-N Concentration (mg/L)	HRT (h)
AGS-SBR-1	I	Days 1–61	123 ± 5	12
	II	Days 71–99	180 ± 17	12
	III	Days 106–120	182 ± 9	8
	IV	Days 134–155	255 ± 7	8
	V	Days 162–183	292 ± 11	8
AGS-SBR-2	I	Days 1–14	84 ± 2	12
	II	Days 24–51	187 ± 7	12
	III	Days 58–72	192 ± 13	8

and

Table 4. Summary of reactor operational parameters for AGS-SBR-3.

	Period	Operational Days	Influent NH3-N Concentration (mg/L)	Supplemental COD Concentration (g/L)	HRT (h)	COD:N:P
AGS-SBR-3	I	Days 1–38	77 ± 12	2.09 ± 0.15	12	105:4:1
		Days 38–72	88 ± 3	2.20 ± 0.07	8	110:4:1
	II	Days 79–93	84 ± 2	1.52 ± 0.01	8	76:4:1
	III	Days 100–114	74 ± 11	0.95 ± 0.06	8	47:4:0.4
	IV	Days 121–135	149 ± 7	0.72 ± 0.05	8	36:8:2
		Days 149–156	131 ± 0	0.54 ± 0.06	8	27:7:1
	V	Days 163–212	121 ± 5	1.39 ± 0.05	8	70:6:2

. The experimental study consisted of several changes that were carried out in different periods. The experimental schematic is shown in Figure 1.

2.2. Seed Sludge

The AGS-SBRs were inoculated with return activated sludge (RAS) from the Ashbridges Bay Treatment plant (ABTP) in Toronto, Ontario, Canada, resulting in an initial sludge concentration of 4.3–5.2 g/L (65–80% volatile). Biomass in AGS-SBR-1 and AGS-SBR-2 was augmented with RAS during periods of low sludge concentration or following a process upset. Additionally, biomass augmentation with RAS was performed to maintain mixed liquor suspended solids (MLSS) and mixed liquor volatile suspended solids (MLVSS) concentrations above 2 g/L and 1.5 g/L, respectively. Biomass augmentation points are indicated on Figure 2a,b,d,e. Seed sludge sampled from ABTP had an average suspended solids (SS) concentration 10 ± 1 g/L (70% volatile) and sludge volume index (SVI₃₀) of 115 ± 9 mL/g.

2.3. Wastewater Media Characteristics

Several feed streams were used to provide substrate to the AGS-SBRs in this experimental study. The characteristics of centrate collected from ABTP and composition of synthetic wastewater is described in

Table 5. Characteristics of centrate and synthetic centrate used for AGS-SBR-1, 2 & 3.

AGS-SBR-1		AGS-SBR-2		AGS-SBR-3 §
Parameter	Average Concentration (Per Litre)	Component	Per Litre	Component	Per Litre
NH3-N	484.83 ± 46.96 mg	NaHCO3	1.2 g	CH3COONa	2.56 g
NO2-N	0.074 ± 0.05 mg	NH4Cl	0.35 g	NH4Cl	0.35 g
NO3-N	0.91 ± 0.13 mg	K2HPO4	0.05 g	K2HPO4	0.03 g
TSS	0.089 ± 0.01 mg	KH2PO4	0.045 g	KH2PO4	0.025 g
VSS	0.075 ± 0.01 mg	CaCl2·2H2O	0.03 g	CaCl2·2H2O	0.03 g
tCOD	674.90 ± 139.62 mg	MgSO4·7H2O	0.025 g	MgSO4·7H2O	0.025 g
sCOD	547.68 ± 136.34 mg	FeSO4·7H2O	0.02 g	FeSO4·7H2O	0.02 g
CaCO3	1922.58 ± 263.84 mg	Micronutrients *	1 mL	Micronutrients *	1 mL
TN	524.75 ± 29.29 mg
PO43−	103.10 ± 11.27 mg

^§ AGS-SBR-3 feed solution changed to centrate with CH₃COONa as supplemental carbon source after day 79. * The micronutrient solution consisted of (in g/L): H₃BO₃, 0.05; ZnCl₂, 0.05; CuCl₂, 0.30; MnSO₄·H₂O, 0.05; (NH₄)₆Mo₇O₂₄·4H₂O, 0.05; AlCl₃, 0.05; CoCl₂·6H₂O, 0.05 and CoCl₂·6H₂O, 0.05.

. ABTP utilizes the conventional activated sludge process for biological treatment of wastewater and anaerobic digestion for sludge stabilization followed by the addition of a polymer compound prior to dewatering.

• AGS-SBR-1: Centrate collected from ABTP was diluted to provide NH₃-N feed concentrations between 80–300 mg/L. NaHCO₃ was added as inorganic carbon source and pH control.
• AGS-SBR-2: Synthetic centrate with NaHCO₃ as inorganic carbon source and pH control, NH₄Cl as nitrogen source and K₂HPO₄ & KH₂PO₄ as phosphorus sources.
• AGS-SBR-3: Synthetic centrate with acetate as organic carbon source, NH₄Cl as nitrogen source and K₂HPO₄ & KH₂PO₄ as phosphorus sources. The feed stream was switched to diluted centrate after day 79 with sodium acetate (CH₃COONa) as supplemental carbon source.

2.4. Analytical Methods

2.4.1. Biomass (Solid Phase) Analysis

MLSS and MLVSS tests were performed according to Standard Methods 2540 D, 2540 E, and 2710 D, respectively [29]. Sludge was allowed to settle in the reactor and timed for 30 min to determine SVI₃₀; similarly, to standard method 2710 D [29]. SVI₃₀ tests were carried out in the reactor to prevent process disruptions associated with dismantling the system to access mixed liquor.

2.4.2. Wastewater (Liquid Phase) Analysis

Nutrient concentrations for influent and effluent were measured using a HACH DR 3900 (HACH company) portable spectrophotometer for the entire study period. NH₃-N was analysed using the Salicylate method (10031). NO₂-N, NO₃-N and alkalinity analysis were performed using HCH test kits TNT plusTM839 & 840, TNT plusTM835 and TNT plusTM870, respectively. TN was determined by the persulfate digestion method (10071). Reactive phosphorus (PO₄³⁻) was determined using the Molybdovanadate method (8114). Total chemical oxygen demand (tCOD) and soluble COD (sCOD) (0.45 µm-filtered) were determined by using the reactor digestion method. Digestion was performed using a Labnet Accublock Digital dry bath (Labnet International Inc., Mayfield, NJ, USA) followed by spectrophotometric measurement. Total suspended solids (TSS) and Volatile suspended solids (VSS) were analysed in accordance with standard methods 2540 D, and 2540 E, respectively [29]. Fisher Scientific Accumet basic pH meter was used to measure pH in AGS-SBR-1 and AGS-SBR-2.

2.4.3. Reactor Performance and Removal Analysis

Removal efficiencies for NH₃-N, sCOD, and PO₄³⁻ were determined by Equation (1) [30]. The accumulation of nitrite in the reactors was determined by Equation (2) [31]. In AGS-SBR-3, the efficiency of simultaneous nitrification-denitrification was determined by Equation (3) [31].

$$\mathrm{Efficiency}(\%)=\frac{{\mathrm{C}}_{\mathrm{influent}}-{\mathrm{C}}_{\mathrm{effluent}}}{{\mathrm{C}}_{\mathrm{influent}}}\times 100$$

(1)

$$\mathrm{Where},\mathrm{C}=\mathrm{Concentration}\mathrm{in}\mathrm{mg}/\mathrm{L}$$

$$\mathrm{Nitrite}\mathrm{Accumulation}\mathrm{Rate}\left(\mathrm{NAR}\right)(\%)=\frac{{\mathrm{NO}}_2-\mathrm{N}}{{\mathrm{NO}}_2-\mathrm{N}+{\mathrm{NO}}_3-\mathrm{N}}\times 100$$

(2)

$$\mathrm{Simultaneous}\mathrm{Nitrification}-\mathrm{Denitrification}\left(\mathrm{SND}\right)\mathrm{Efficiency}(\%)=\left(1-\frac{{[{\mathrm{NO}}_{\mathrm{x}}-\mathrm{N}]}_{\mathrm{accumulated}}}{{[{\mathrm{NH}}_3-\mathrm{N}]}_{\mathrm{removed}}}\right)\times 100$$

(3)

Sludge production based on wastewater characteristics is determined by Equation (4). Parts A, B & C were computed to determine sludge production in AGS-SBR-3. Sludge production in AGS-SB-1 and AGS-SBR-2 were determined with only Part C. The concentration of nitrogen oxidized as NH₃-N was determined with Equation (5). Nitrification rate was calculated using Equation (6) and specific nitrification rate was determined by taking into the consideration the VSS concentration as shown in Equation (7) [30].

$$\begin{gathered} {\mathrm{P}}_{\mathrm{x},\mathrm{bio}}=\mathrm{Part}\mathrm{A}\left(\mathrm{Heterotrophic}\mathrm{Biomass}\right)+\mathrm{Part}\mathrm{B}\left(\mathrm{Cell}\mathrm{debris}\right)+\mathrm{Part}\mathrm{C}\left(\mathrm{Nitrifying}\mathrm{bacteria}\mathrm{biomass}\right) \\ {\mathrm{P}}_{\mathrm{x},\mathrm{bio}}=\frac{{\mathrm{QY}}_{\mathrm{H}}\left({\mathrm{S}}_0-\mathrm{S}\right)\left(\frac{1\mathrm{kg}}{{10}^3\mathrm{g}}\right)}{1+{\mathrm{b}}_{\mathrm{H}}\left(\mathrm{SRT}\right)}+\frac{\left({\mathrm{f}}_{\mathrm{d}}\right)\left({\mathrm{b}}_{\mathrm{H}}\right){\mathrm{QY}}_{\mathrm{H}}\left({\mathrm{S}}_0-\mathrm{S}\right)\mathrm{SRT}\left(\frac{1\mathrm{kg}}{{10}^3\mathrm{g}}\right)}{1+{\mathrm{b}}_{\mathrm{H}}\left(\mathrm{SRT}\right)}+\frac{{\mathrm{QY}}_{\mathrm{n}}\left({\mathrm{NO}}_{\mathrm{x}}\right)\left(1\mathrm{kg}/{10}^3\mathrm{g}\right)}{1+{\mathrm{b}}_{\mathrm{n}}\left(\mathrm{SRT}\right)} \end{gathered}$$

(4)

$$\begin{gathered} {\mathrm{NO}}_{\mathrm{x}}={\mathrm{concentration}\mathrm{of}\mathrm{NH}}_3-\mathrm{N}\mathrm{in}\mathrm{the}\mathrm{influent}\mathrm{flow}\mathrm{that}\mathrm{is}\mathrm{nitrified},\mathrm{mg}/\mathrm{L} \\ {\mathrm{Y}}_{\mathrm{H}}=\mathrm{heterotrophic}\mathrm{bacteria}\mathrm{synthesis}\mathrm{yield}\mathrm{coefficient},0.4\mathrm{g}\mathrm{VSS}/\mathrm{g}\mathrm{COD} \\ {\mathrm{b}}_{\mathrm{H}}=\mathrm{endogenous}\mathrm{decay}\mathrm{coefficient}\mathrm{for}\mathrm{heterotrophic}\mathrm{organisms},0.16\mathrm{g}\mathrm{VSS}/\mathrm{g}\mathrm{VSS}\mathrm{d} \\ {\mathrm{b}}_{\mathrm{n}}=\mathrm{endogenous}\mathrm{decay}\mathrm{coefficient}\mathrm{for}\mathrm{nitrifying}\mathrm{organisms},0.2\mathrm{g}\mathrm{VSS}/\mathrm{g}\mathrm{VSS}\mathrm{d} \\ {\mathrm{Y}}_{\mathrm{n}}=\mathrm{nitrifying}\mathrm{bacteria}\mathrm{synthesis}\mathrm{yield}\mathrm{coefficient},0.12\mathrm{g}\mathrm{VSS}/{\mathrm{g}\mathrm{NH}}_4-\mathrm{N} \\ {\mathrm{f}}_{\mathrm{d}}=\mathrm{fraction}\mathrm{of}\mathrm{cell}\mathrm{mass}\mathrm{that}\mathrm{remains}\mathrm{as}\mathrm{cell}\mathrm{debris},0.10-0.15\mathrm{g}\mathrm{VSS}/\mathrm{g}\mathrm{biomass}\mathrm{VSS}\mathrm{depleted}\mathrm{by}\mathrm{decay} \end{gathered}$$

$$\begin{gathered} \mathrm{Nitrogen}\mathrm{oxidized}=\mathrm{nitrogen}\mathrm{in}\mathrm{influent}-\mathrm{nitrogen}\mathrm{in}\mathrm{effluent}-\mathrm{nitrogen}\mathrm{in}\mathrm{cell}\mathrm{mass} \\ \left({\mathrm{NO}}_{\mathrm{x}}\right)=\left({\mathrm{TKN}}_{\mathrm{o}}\right)*-{\mathrm{N}}_{\mathrm{e}}-0.12\left(\frac{{\mathrm{P}}_{\mathrm{x},\mathrm{bio}}}{\mathrm{Q}}\right) \end{gathered}$$

(5)

$${*\mathrm{Influent}\mathrm{NH}}_3-{\mathrm{N}\mathrm{used}\mathrm{in}\mathrm{place}\mathrm{of}\mathrm{TKN}}_{\mathrm{o}}$$

$$\mathrm{Nitrification}\mathrm{Rate}=\frac{{\mathrm{NO}}_{\mathrm{x}}}{{\mathrm{t}}_{\mathrm{aeration}}}$$

(6)

$$\mathrm{Specific}\mathrm{nitrification}\mathrm{rate}\left(\mathrm{SNR}\right)=\frac{\mathrm{Nitrification}\mathrm{rate}}{{\mathrm{MLVSS}}_{\mathrm{reactor}}}$$

(7)

2.4.4. Granule Structure and Morphology

Sludge was routinely collected from the lowest sample port of each reactor and poured into a Petri dish. Photographic images were captured with a Samsung Galaxy A50 smart phone camera and microscopic images were taken with a Leica DM1000 LED microscope. Following installation and start-up of the microscope, the Leica representative returned to calibrate the device using precise measurement bars.

3. Results and Discussion

3.1. Characteristics of Granular Sludge

3.1.1. Centrate Nitrifying Sludge (AGS-SBR-1)

In AGS-SBR-1, SVI₃₀ averaged 72 ± 12 mL/g; displaying good sludge settleability. The SVI₃₀/SVI₅ ratio averaged 0.5 ± 0.08, equating to 50% granulation as shown in Figure 2a. Settleability of the sludge was observed to be influenced by the reactor design and operation: (1) H/D ratio of 12.2 and, (2) air flow rate of 3 L/min. A high H/D ratio represents a slender column, which influences the settling time [24]. Non-limiting DO air flow during the aeration phase resulted in adequate oxidation of ammonia.

Sludge concentration for reactor start-up was approximately 4.37 g MLSS/L and 2.85 g MLVSS/L. After 11 days of operation, sludge concentration decreased to approx. 2.34 g MLSS/L and 1.50 g MLVSS/L. To maintain the sludge concentration, RAS samples from ABTP were added to AGS-SBR-1 throughout the operation. On average, a sludge concentration of 2.68 ± 1 g MLSS/L and 1.87 ± 0.63 g MLVSS/L was achieved for the entire duration of the study. This includes fluctuations in the MLSS and MLVSS concentration between 1.16–5.33 g/L and 0.79–2.97 g/L, respectively. In Figure 2d, instances in which RAS was used to augment the sludge concentration are indicated on the MLSS (g/L) concentration curve. The ratio of MLVSS/MLSS fluctuated in response to the sludge concentration, but an average vale of 0.72 ± 0.09 was determined for AGS-SBR-1. According to Czarnota et al. [32], MLVSS/MLSS ratio signifies sludge activity or organic content, and it is typically in the range 0.7–0.8 [30]. In Figure 2d, reactor re-seed on day 34 in response to a substantial decrease in sludge concentration resulted in a lower than usual MLVSS/MLSS ratio. Evidently, a lower MLVSS/MLSS concentration ratio indicated a change in the biomass component such as a decrease in the concentration of viable sludge [33].

3.1.2. Synthetic Centrate Nitrifying Sludge (AGS-SBR-2)

Sludge settleability for AGS-SBR-2 in terms of SVI₃₀ averaged 111 ± 4 mL/g for days 4–14 and 59 ± 11 mL/g for the remainder of the reactor operation. The average granulation percentage for the entire research study based on the ratio SVI₃₀/SVI₅ was 53 ± 9%. Visual inspection of the sludge in AGS-SBR-2 did not show obvious evidence of distinct granule formation, but the settling behaviour was like that of a typical granular sludge reactor. Activated sludge systems are known to have an average SVI₃₀ between 110 and 160 [34], but AGS-SBR-2 averages for settling was substantially lower, especially after day 14 of operation as shown in Figure 2b.

Sludge concentration for AGS-SBR-2 at start-up was approx. 4.78 g MLSS/L and 3.35 g MLVSS/L. The sludge concentration decreased to 2.43 g MLSS/L and 1.76 g MLVSS/L by day 14 as shown in Figure 2e. RAS from ABTP was added to AGS-SBR-2 to maintain sludge concentration and the MLVSS/MLSS ratio. In comparison to AGS-SBR-1, MLSS and MLVSS concentration experienced higher stability in AGS-SBR-2. The average MLSS and MLVSS concentration for AGS-SBR-2 were 3.03 ± 0.75 g/L and 1.99 ± 0.54 g/L, respectively. The average MLVSS/MLSS ratio for AGS-SBR-2 was 0.66 ± 0.05, which was lower than that of AGS-SBR-1 at 0.72 ± 0.09. The average MLVSS/MLSS ratio of 0.66 ± 0.05 was also lower than the typical ratio of 0.7–0.8 [30]. The ratio between MLVSS and MLSS highlighted sludge inactivity or subpar organic content [32].

3.1.3. Nitrifying-Denitrifying Sludge (AGS-SBR-3)

The transition from activated to predominantly granular sludge occurred in parallel with decreasing values for SVI₃₀ as shown in Figure 2c, highlighting the progression of sludge settleability. The curve for SVI₃₀ undergoes a downward trend, which begins with SVI testing performed on day 7 and onwards. In Figure 2c, peaks on SVI₃₀ curve on days 14 and 52 were associated with excess sludge production and washout, respectively. On day 14, increased sludge production in AGS-SBR-3 resulted in a MLSS concentration of 7.37 g/L which increased settling time due to excess biomass present in the system. A longer sludge settling time on day 52 was recorded after a reactor upset two days prior. The system is estimated to have completed 2 cycles without receiving influent. Aeration of settled sludge with one-half of the reactor working volume led to disruption of biological aggregates that previously enhanced the settling time. The average SVI₃₀ was 52 ± 23 mL/g for the entire duration of the study and 45 ± 15 mL/g for all tests performed after day 14. The SVI₃₀/SVI₅ ratio for the entire duration of the study was 0.59 ± 0.21, which quantifies the percent granulation in the system. In Figure 2c, SVI₃₀/SVI₅ ratio is shown to have gradually increased after day 59 and reached a maximum of 0.96 on day 149. Granules were occasionally collected to monitor sludge morphology through visual inspection and microscopic imaging. As the research objectives were completed by day 163, the reactor operation was maintained to accommodate sludge sampling and storage. On day 212, final sampling and testing were completed and the SVI₃₀ narrowly increased in response to a power outage and a routine campus lockout.

The potential for aerobic granular sludge reactors to operate at high sludge concentrations compared to activated sludge was witnessed in AGS-SBR-3. Irrespective of the fluctuations in MLSS concentration displayed in Figure 2f, consistent biomass growth allowed for concentrations of >7 g/L in AGS-SBR-3. Furthermore, MLSS concentration appeared to trend upwards throughout the experimental study. Sludge concentration increased from 5.20 to 7.37 g MLSS/L and 4.21 to 6.14 g MLVSS/L between days 7 and 14. A high sludge concentration was maintained until day 45 when MLSS and MLVSS decreased to 5.41 and 4.86 g/L, respectively. On day 51, the lowest MLSS (4.49 g/L) and MLVSS (3.99 g/L) concentrations for the entire experimental were recorded. A reduction in sludge concentration was caused by an improper timer setting for the decant pump on day 43, which led to a discharge of approximately 50% of the reactor’s volume as mixed liquor. The robustness of the AGS-SBR-3 was confirmed as the sludge concentration increased to 9.62 g MLSS/L and 8.66 g MLVSS/L by the 72nd day of the experiment study. Numerous process modifications were applied to AGS-SBR-3 between days 72 and 156. Process modifications included altering the concentration of COD and NH₃-N in the feed solution to lower the COD/N ratio. On day 159, both COD and NH₃-N were adjusted to 1500 mg/L and 120 mg/L, respectively. Sludge analysis on day 163 confirmed a reduction in the sludge concentration to 10.76 g MLSS/L and 9.95 g MLVSS/L. Similarly, a reduction in sludge concentration was also recorded on day 170. AGS-SBR-3 rebounded prior to the end of the experiment with the sludge concentration on day 212 remaining at >8 g MLSS/L and >7 g MLVSS/L. The MLVSS/MLSS ratio curve in Figure 2f, displays a slightly horizontal curve that gradually trend upwards. The average MLVSS/MLSS ratio of 0.89 ± 0.03 proves that MLVSS concentration was also high and continued to increase with the progression of the research study. A high MLVSS concentration signifies good microbial activity within AGS-SBR-3 [32]. The presence of various organisms was captured with the assistance of a microscope.

3.2. Morphology and Structure of Aerobic Granules

Images are organized according to the timeline of the study, which shows the following: (1) Transition from activated to granular sludge, and (2) Shape of granules. In Figure 3a,b,e,f, distinct nitrifying granules were not found even though the system displayed good settleability. In Figure 3e, several small aggregates in the sample were observed to be >100 μm. Similarly, the formation of small aggregates in AGS-SBR-2 was also observed in Figure 3f. In a study by Liu et al. [22], nitrifying granules with a size of 240 μm were formed using synthetic wastewater with an inorganic carbon source as feed. Figure 3c,g provide adequate evidence of granule formation as early as day 32 in AGS-SBR-3. In Figure 3h a granule with diameter 1130 μm (1.13 mm) was captured using the microscope. The diameter could have been slightly less since the cover slide crushed the granules during microscopic analysis. Nonetheless, large-sized granules were sampled as shown in Figure 3d.

3.3. Pollutants Removal Efficiencies

3.3.1. Centrate Nitrifying SBR (AGS-SBR-1)

AGS-SBR-1 displayed evidence of nitrification through the production of NO_x-N [NO₂-N + NO₃-N]. During start-up, RAS was acclimated with undiluted centrate for three days to ensure adequate oxidation of NH₃-N could be achieved during the subsequent 6-h (days 0–99) and 4-h (days 99–183) cycles. Following three days of acclimation, the reactor was fed with diluted centrate. Experimental analysis of samples collected on day 4 confirmed a 96% NH₃-N removal and the production of NO₃-N which signifies the occurrence of nitrification. NH₃-N removal >96% was maintained throughout days 4–27 in AGS-SBR-1. Sludge sample collected and analyzed on the 27th day resulted in concentrations of 1.16 g MLSS/L and 0.96 g MLVSS/L. Though NH₃-N removal was approx. 100%, a weekend shutdown was completed between days 34 to 36. With a reactor working volume of 1.2 L, half volume was re-seeded with RAS and the remainder was fed with undiluted centrate. On day 36, AGS-SBR-1 regained operation with a 6-h cycle time and a diluted centrate feed targeting 100 mg/L of NH₃-N. Between days 36–61, it can be seen in Figure 4a that NH₃-N removal was low on day 40, but it increased with continued operation. During this period, the production of NO₃-N was hampered and NO₂-N in the effluent was 109.8 mg/L. Secondly, on day 40 the pH reached 8.7 compared to a typical value within the range 7.3–7.9. The presence of AOBs in the mixed liquor allowed for efficient oxidation of NH₃-N to NO₂-N, resulting in partial nitrification. Further oxidation of NO₂-N to NO₃-N was impacted due to the absence of NOBs coupled with the increased pH resulting from the addition of NaHCO₃ to supplement the alkalinity in centrate. A high pH would have impacted the growth of NOBs, which are considered to be more sensitive to increasing pH [35]. By day 61, experimental data obtained from analyzed sample confirmed complete recovery, i.e., low NO₂-N and high NO₃-N concentration in the effluent.

Increasing the concentration of NH₃-Ncentrate feed to approx. 200 mg/L after day 61 did not impact performance. Between days 71 and 120, influent NH₃-N averaged 180 ± 14 mg/L, equating to an NH₃-N loading of 0.36 ± 0.03 g/L.d and the removal efficiency was 98 ± 3 %. A second increase in NH₃-N influent concentration was interrupted due to a malfunction with the ancillary equipment. Prior to operational disruption, NH₃-N removal was 88% on day 134, effluent NO₃-N and NO₂-N were 200 mg/L and 15 mg/L, respectively. The malfunctioning feed pump failure was discerned on day 137 and with unsuccessful attempts to guarantee adequate operation overnight, the remaining sludge in the reactor was augmented with 200 mL of RAS from ABTP on day 140. Reactor performance was severely impacted as shown by the low removal of NH₃-N on day 155 at 11% in Figure 4a. The concentration of NO₃-N slightly decreased to 198 mg/L, but NO₂-N increased to a value of 60 mg/L. In addition to the effluent NH₃-N of 232 mg/L, the nitrogen balance for day 155 was affected by the operational disruption–effluent concentrations of NH₃-N, NO₃-N and NO₂-N exceeded the influent NH₃-N concentration. The sludge concentration also decreased during this period, therefore, to prevent further sludge loss through resuspension while allowing for ammonia oxidation, SAV was lowered to 1 L/min after day 155. The impact of operating at a lower SAV was observed with samples collected between days 162 and 183. Excellent NH₃-N removal was achieved, averaging 97 ± 2% but partial nitrification was dominant with NO₂-N concentration remaining consistently high between 208–268 mg/L. Over the course of the entire research study (i.e., 183 days), average NH₃-N removal was determined to be 93 ± 19%; inclusive of the reactor disruption periods. In Figure 4a, the downward peaks are indicative of the reactor disruptions periods. Additional information on the reactor’s performance can be found in the electronic Supplementary Material (Table S1).

3.3.2. Synthetic Centrate Nitrifying SBR (AGS-SBR-2)

The occurrence of nitrification in AGS-SBR-2 was confirmed after 7 days of operation, inclusive of an acclimation phase. An average NH₃-N concentration of 84 ± 2 mg/L was fed to AGS-SBR-2 during acclimatization until day 14. Experimental analyses for days 7 and 14 confirmed 100% NH₃-N removal, conversion of NO₂-N during nitrification and an average effluent concentration of 83 ± 3 mg/L for NO₃-N. Increasing the concentration of influent NH₃-N after day 14 negatively impacted the performance of the system as shown in Figure 4b; NH₃-N removal decreased to 52% and NO₃-N concentration was 90 mg/L on day 24. On day 31, NH₃-N returned to approximately 100% removal and NO₃-N present in the effluent reflected the influent NH₃-N concentration. Between days 31 and 72, influent NH₃-N of 190 ± 8 mg/L was completely removed while average effluent concentration of NO₃-N was 185 ± 3 mg/L. NO₂-N effluent concentration was undetectable for majority of the period (i.e., <0.2 mg/L) until day 72 where a concentration of 2.19 mg/L was detected. Overall, NO₂-N was consistently converted to NO₃-N. The growth of distinct nitrifying granules was not detected in AGS-SBR-2, but the reactor’s performance in oxidizing NH₃-N is an indication of the technology’s capability. The difficulty in cultivating nitrifying granules is related to the slow growth rate of autotrophic bacteria, which are responsible for nitrification [36]. Previously, researchers have confirmed the presence of nitrifying granules [22,37,38], but the cultivation process is understood to be time-consuming. Recent research have shifted to shortening the cultivation time through inoculation with mature granules or supplementing the carbon source [36,39].

The relationship between influent NH₃-N and effluent NO₃-N in AGS-SBR-2 between days 31 and 72, proved that the use of synthetic wastewater positively impacted nitrogen balance in the system. Synthetic centrate prepared in the laboratory contained considerably less impurities and was void of polymer compounds in comparison to real centrate collected from ABTP. In comparison to AGS-SBR-1, NO₃-N in the effluent of AGS-SBR-2 was lower than the influent NH₃-N concentration expect for days 31 and 58. The difference in concentration for days 31 and 58 were 2% and 3%, respectively. Similarly, day 7 shows NO₃-N slightly higher than influent NH₃-N with the difference being 3%. The average pH in AGS-SBR-2 for the duration of the research study was 8 ± 0.5. Minor fluctuation in pH was observed as depicted in Figure 4b, but the system displayed adequate stability. Over the course of the entire research study (i.e., 72 days), average NH₃-N removal was determined to be 94 ± 16%. This takes into consideration a disruption in performance captured on day 24 in Figure 4b. Additional information on the reactor’s performance can be found in the electronic Supplementary Material (Table S2).

3.3.3. Nitrifying-Denitrifying SBR (AGS-SBR-3)

In AGS-SBR-3, the marginal difference between influent and effluent NH₃-N represented an average removal of 92 ± 9% for the entire duration of the research study. Complete nitrogen removal was achieved in AGS-SBR-3 prior to day 100 with low NO_x-N concentration in the effluent. Therefore, it was evident that simultaneous nitrification and denitrification (SND) is possibly the main removal mechanism involved. Aerobic granules are capable of SND due to its layered structure which can provide anoxic/anaerobic conditions deep within the granule for conversion of NO_x-N to dinitrogen gas (N₂) [40]. An increase in NH₃-N concentration in the feed stock with reduced COD resulted in excess NO₃-N in the effluent as shown in Figure 4c. NO₃-N accumulation is shown to moderately increase between days 100 and 114 with an average COD/N ratio of 13 ± 1. A tremendous increase in effluent NO₃-N is seen between days 114 and 135; the latter representing the peak value of 134.8 mg/L. During this period, the average COD/N ratio was 5 ± 0.4. Effluent NO₃-N concentrations of 105.2 mg/L and 102.4 mg/L were recorded on days 145 and 156, respectively. NO₃-N concentration > 100 mg/L is considered high in comparison to recorded values prior to the first substantial spike on day 114. The COD/N ratio averaged 4 ± 0.5 for day 145 and 156. Overall, the average COD/N ratio for days 114 to 156 was 5 ± 0.5. NO_x-N accumulation at a low COD/N highlights the dependence on readily biodegradable in the feed for adequate SND. In a study by Ren et al. [31], NO_x-N accumulation and low SND efficiencies occurred in an aerobic granular sludge system when biodegradable COD in the feed decreased. NO₃-N continued to decrease after day 156 through to the 212th day with an average COD/N ratio of 12 ± 0.5 recorded during this period. The intermediate product, NO₂-N, generated during nitrification was continuously converted to NO₃-N. The maximum concentration of NO₂-N recorded in the effluent was 10.28 mg/L on day 38, while 42% of samples had concentration of <0.2 mg/L. SND efficiency for the period with low NO_x-N accumulation was 87 ± 11%. pH readings were recorded at various instances during a cycle and automatic adjustments was performed with the help of a pump used to dose HCl. An average pH of 7.86 ± 0.29 was tabulated for the duration of the experimental study. Additional information on the reactor’s performance can be found in the electronic Supplementary Material (Table S3).

COD concentration between days 7 and 72 averaged 2143 ± 139 mg/L as shown in Figure 4d. After confirming the presence of granules in AGS-SBR-3, COD concentration in the feed was gradually lowered in stages until day 159. COD concentration in the feed was increased to 1445 mg/L on day 159 and an average of 1390 ± 54 was maintained for the remainder of the study as shown in Figure 4d. CH₃COONa was included in both the synthetic wastewater and diluted centrate feed solution as the main source of COD; with a switch in feed streams occurring after day 79. Supplemental acetate-COD played two important roles in AGS-SBR-3: (1) Nitrogen removal—Heterotrophic organisms utilize acetate-COD as electron donor for energy generation [30,41], and (2) Phosphorus removal—Acetate-COD is hydrolyzed into organic acids to aid in phosphorus release and uptake [42]. Excellent COD removal is supported by the fact that it is the limiting substrate in nitrogen and phosphorus removal [43].

Effluent monitoring of AGS-SBR-3 captured the removal of phosphorus as shown in Figure 4e, where an average removal of 81 ± 14% was achieved between days 1–114 of operation. During this period, COD in the feed was gradually decreased after day 72 and the feed stream was switched to centrate instead of synthetic wastewater after day 79. Synthetic wastewater consisted of K₂HPO₄, KH₂PO₄ as a phosphorus source. Phosphorus removal drastically decreased on day 121, at which point the COD was <800 mg/L and NH₃-N was increased to >120 mg/L in the feed. Average phosphorus removal determined for the remainder of the study was 23 ± 10%. Sequential operation of AGS-SBR-3 did not change; therefore the presence of anoxic-aerobic conditions would have influenced the growth of phosphate-accumulating organisms (PAOs)—widely known their role in biological phosphorus removal [43]. A reduction in phosphorus removal on day 121 indicates that a low COD concentration negatively impacts the function of PAOs. According to Zeng et al. [43], COD is a limiting substrate in nitrogen and phosphorus removal processes. Furthermore, the connection between nitrogen and phosphorus removal (i.e., the denitrification step) led researchers to recognizing the function of denitrifying PAOs [44]. Pronk et al. [45] expressed that dPAOs can operate with double functionality in an AGS system: (1) an electron donor in the denitrification process, and (2) store phosphate as poly-P. In this study, NH₃-N was removed without NO₂-N and NO₃-N accumulation, indicating simultaneous nitrification, denitrification, and phosphorus removal. In a study by Figdore et al. [27], a cycle test using NDN-PAO granules resulted in concurrent phosphate uptake and NH₃-N removal without accumulation of NO₂-N and NO₃-N. Furthermore, it was concluded that NO₂-N and NO₃-N occurred during the aeration phase—thus confirming SND mechanism in a granular system [27].

3.4. Pollutants Degradation in SBR Cycle

3.4.1. Cycle Test: AGS-SBR-1 (Day 120)

The cycle test completed for AGS-SBR-1 captured the concentrations of NH₃-N, NO₃-N and NO₂-N during a 4-h cycle period in Figure 5a. The pH curve depicts the effect of nitrification reaction on alkalinity during the cycle. It is evident in Figure 5a that the pH decreases from 7.68 to 5.99 within the first hour of the cycle. NaHCO₃ was dosed to the reactor at two different points during the 4-h cycle; the 1-h and 1.5-h mark. Nitrification reduces alkalinity and subsequently the pH of the system through oxidation of NH₃-N [46]. The H⁺ ions are freed from the nitrogen bond as NH₃-N is converted to NO₂-N and inevitably NO₃-N, thus creating acidic conditions and altering the pH buffer capability.

Ammonia removal in parallel with the production of NO₃-N via conversion of NO₂-N occurred with 1.5 h of the cycle as shown in Figure 5a. Concentrations of NH₃-N, NO₃-N and NO₂-N measured at time zero were 72.5 mg/L, 109.5 mg/L and 7.85 mg/L, respectively. At the 1.5-h mark, both NH₃-N and NO₂-N were depleted, and the concentration of NO₃-N was 197.0 mg/L, approximately 7 mg/L more than the sum of concentrations for all three nitrogenous compounds at time zero. While this confirmed ammonia oxidation and complete nitrification within 1.5 h, the additional production of NO₃-N is presumably linked to the presence of organic nitrogen in the centrate. Organic nitrogen can form NH₃-N and undergo nitrification. The rate of NH₃-N uptake was determined by performing linear regression on NH₃-N depletion over time. In Figure 5b, NH₃-N is shown to have been depleted within 1.5 h. Linear regression of the curve resulted in a nitrification rate (NR) of 45.3 mg NH₃-N/L·h. Furthermore, the average MLVSS concentration was 2.36 ± 0.02 g/L, resulting in a specific nitrification rate (SNR) of 19.19 mg NH₃-N/g VSS·h. Calculated values for NR and SNR using Equations (4)–(7) were 64.16 mg NH₃-N/L·h and 26.96 mg NH₃-N/g VSS·h, respectively.

3.4.2. Cycle Test: AGS-SBR-2 (Day 72)

In Figure 5c, the depletion of NH₃-N is monitored for one complete cycle. In addition to NH₃-N, Figure 5c shows how NO₂-N is produced and convert to NO₃-N. The concentration of NH₃-N after 1.5 h is 1.6 mg/L, followed by 0.8 mg/L at the 2-h mark. NH₃-N is completed depleted after the 2 h of operation. Interestingly, the concentration of NH₃-N at the 0.5-h mark was 49 mg/L, which is high compared to AGS-SBR-1 where a lower concentration of 22 mg/L was achieved within the same time span. Between 0.5 to 1 h, majority of the NH₃-N in AGS-SBR-1 was oxidized. The production and conversion of NO₂-N to NO₃-N occurred for the entire duration of the cycle, but majority of NO₂-N was converted after the 2.5 h. The concentration of NO₂-N at the 2.5-h mark was 5.37 mg/L. The slow-paced conversion of NO₂-N resulted in a gradual increase in NO₃-N. The NO₃-N curve in Figure 5c increased to a concentration of 184 mg/L. Reactor pH was dependent on the occurrence of nitrification and the addition of NaHCO₃ to supplement alkalinity. AGS-SBR-1 and AGS-SBR-2 were fed with NaHCO₃ from a common pump which operated on a timer. NaHCO₃ would be dosed at the 1-h and 1.5-h mark during a cycle. It can be seen in Figure 5c that pH increased after the alkalinity dosing points for AGS-SBR-2. The pH in AGS-SBR-2 decreased to 7.01 before the first dose of NaHCO₃. In AGS-SBR-1, Figure 5a shows that pH decreased to 5.99 before the first dose of NaHCO₃. It is evident through the consumption of alkalinity–ability of the system to buffer pH decreases—that AGS-SBR-2 initially experienced a slower nitrification rate compared to AGS-SBR-1, but the rate increased after the 0.5 and 1-h mark. Computational determination of NR and SNR were performed by analyzing the NH₃-N curve in Figure 5d. Irrespective if the slow start to NH₃-N removal in AGS-SBR-2, linear regression of the curve resulted in a NR of 40.36 mg NH₃-N/L·h. The average concentration of MLVSS between the start and end of the cycle was 1.475 ± 0.26 g/L, therefore SNR was 27.36 mg NH₃-N/g VSS·h. Calculated values for NR and SNR using Equations (4)–(7) were 63.31 mg NH₃-N/L·h and 38.14 mg NH₃-N/g VSS·h, respectively.

3.4.3. Cycle Test: AGS-SBR-3 (Day 128)

The cycle test completed on day 128 as shown in Figure 5e, accurately displays the impact of a low COD/N ratio on NH₃-N removal via the nitrification-denitrification reaction pathway. During the anoxic feeding period, NO₃-N is biologically used as an oxidizer to remove organics–representing denitrification. In Figure 5e, NO₃-N concentration is reduced by 58% during the anoxic period from 111.50 to 46.70 mg/L. NH₃-N concentration continues to increase in the system as the reactor is fed during the anoxic period. NO₂-N in the feed is typically <0.2 mg/L, yet it’s generation during the anoxic period is linked to residual oxygen within the system. An extended aeration period from the previous cycle introduces a substantial amount of oxygen in the system that may promote minimal nitrification during the anoxic period. At the 1-h mark, air bubbles are introduced to develop an aerobic environment within the system. NO₂-N is produced and converted to NO₃-N—which continues to increase for the remainder of the cycle. NO₂-N concentration peaked at after 1 h while NH₃-N is decreased to a concentration of 3.59 mg/L within 1.5 h of aeration. The final concentration for NH₃-N recorded at the end of the cycle was 1.23 mg/L. Accumulation of NO₃-N discussed previously was evident in the cycle test conducted on day 128. The COD/N ratio was determined to be 4.77 on the 128th day of operation. Linear regression of the curve resulted in a NR of 58.48 mg NH₃-N/L·h as shown in Figure 5f. The average concentration of MLVSS between the start and end of the cycle was 8.52 ± 0.26 g/L, therefore SNR was 6.86 mg NH₃-N/g VSS·h. Calculated values for NR and SNR using Equations (4)–(7) were 79.86 mg NH₃-N/L·h and 9.18 mg NH₃-N/g VSS·h, respectively. Nitrification and/or specific nitrification rates for various studies are presented in

Table 6. Summary of nitrification and specific nitrification rates reported in the literature.

Reactor Type	Parameter	References
Aerobic Granular Sludge Sequencing Batch Reactor (AGS-SBR)	Maximum specific ammonia oxidizing rate = 29.8 mg N/g VSS/h	[23]
Aerobic Granular Sludge Sequencing Batch Reactor (AGS-SBR)	Maximum specific ammonia oxidizing rate NIT granules = 31.25 mg N/g VSS/h NDN-PAO granules = 5.83 mg N/g VSS/h	[39]
Activated Sludge Sequencing batch reactor (AS-SBR)	Specific Nitrification rate = 30–45 g NH4-N/kg MLVSS/h	[49,50]
CAS	Specific nitrification rates in the membrane and conventional mixed liquor: between 3 and 5 mg N/g VSS/h	[51]
CAS	Specific nitrification rates: between 4 and 7 mg N/g VSS/h were measured in both membrane and conventional systems.	[51,52]
Activated Sludge (AS)	Maximum Nitrification rate = 1.0–4.5 mg N/g MLVSS/h	[53]
SBR	Ammonia utilization rate (AUR) = 2.95 +/− 0.26 mg NH4+-N/g VSS/h Ammonia utilization rate (AUR) was faster = 6.16 +/− 0.34 mg NH4+-N/g VSS/h	[54]
AGS-SBRs	Nitrifying AGS-SBR with diluted centrate: Nitrification rate = 45.30 mg N/L·h Specific nitrification rate = 19.19 mg N/g VSS·h Nitrifying AGS-SBR with synthetic centrate: Nitrification rate = 40.36 mg N/L·h Specific nitrification rate = 27.36 mg N/g VSS·h Nitrifying/Denitrifying AGS-SBR: Nitrification rate = 54.48 NH3-N/L·h Specific nitrification rate = 6.86 mg NH3-N/g VSS·h. SND efficiency during periods with negligible NOx-N accumulation was 87 ± 11%.	(This Study)

.

The concentration profiles of COD and phosphorus for day 128 are shown in Figure 5g. COD increases during the anaerobic feed phase before it is assimilated under aerobic conditions. A fraction of the COD is assimilated under anaerobic conditions for denitrification and phosphorus release [47,48]. Low COD concentrations on day 128 resulted in subpar phosphorus removal as shown in Figure 5g.

4. Conclusions

This study shows that aerobic granular sludge systems are a viable treatment technology for liquid sidestreams generated in a WWTP. Cultivation of nitrifying-denitrifying granules <1000 μm (<1 mm) was observed after 32 days, whereas nitrifying sludge consisted of small aggregates with diameter >100 μm for majority of the study period. Good settleability of nitrifying sludge was observed with SVI₃₀ values of 72 ± 12 mL/g (AGS-SBR with real centrate) and 59 ± 11 mL/g (AGS-SBR-2 with synthetic centrate). Nitrifying-denitrifying displayed an average SVI₃₀ of 52 ± 23 mL/g for the entire duration of the study. Excellent NH₃-N removal, >90%, was achieved in the nitrifying AGS-SBRs. The nitrifying-denitrifying AGS-SBR displayed good COD (>90%) and P (>80%) removal without NO_x-N accumulation COD/N ratio >11 mg sCOD/mg NH₃-N. Simultaneous nitrification-denitrification (SND) and phosphorus removal was dominant at higher COD/N ratios. NO_x-N accumulation was observed, and P removal was affected when decreasing the COD/N (4–5 mg sCOD/mg NH₃-N). The nitrifying-denitrifying AGS-SBR displayed produced the highest NR 58.48 mg NH₃-N/L·h, while the overall SND during periods with negligible NO_x-N accumulation was 87 ± 11%. With the formation of small aggregates (i.e., microgranules) in the nitrifying systems fed with real and synthetic centrate, the NR was 45.3 mg NH₃-N/L·h and 40.36 mg NH₃-N/L·h, respectively. The AGS-SBRs in this study produced higher nitrification rates compared to that reported in the literature. Future research should focus on enhancing the mechanism of formation for nitrifying granules without an organic carbon source within a short period and implementing SND using nitrifying-denitrifying at pilot scale.