An innovative granular sludge deammonification system was incorporated into a conventional-activated sludge process. The process incorporated an internal baffle in the bioreactor for continuous separation of granular biomass from flocculent biomass, which allowed for controlling the solids retention time of flocculent sludge. The process was evaluated for ammonium removal from municipal digested sludge dewatering centrate under various operating conditions lasting over 450 days. The process successfully removed, on average, 90% of the ammonium from centrate at various ammonium loading reaching 1.4 kg/m3d at 20 hours hydraulic retention time. Controlling the retention time of the flocculent biomass and maintaining low nitrite concentration were both found to be effective for nitrite oxidizing bacteria management, resulting in a low nitrate concentration (below 50 mg/L) over a wide range of flocculent biomass concentration in the bioreactor.

INTRODUCTION

The wastewater treatment industry is facing challenges that are being driven by more stringent government legislation on discharge regulations and public support of sustainable activities to reduce greenhouse gas emissions. The removal of nutrients, particularly nitrogen, from many municipal and industrial wastewater facilities has become imperative.

Within the context of biological nitrogen removal, side-stream deammonification is being established as a cost-effective solution, due to reduced specific nitrogen removal costs compared to conventional nitrification/denitrification (Wett 2007). Deammonification is a fully autotrophic two-stage reaction that shortcuts conventional nitrification/denitrification by synergy of ammonium-oxidizing bacteria (AOB) and anaerobic AOB (anammox) where ammonium converts to nitrogen gas (Sliekers et al. 2003). First, ammonium is partly oxidized to nitrite by AOB (Reaction (1)) follow by the conversion of the leftover ammonia and produced nitrite by anammox bacteria (Reaction (2)). Deammonification (Reaction (3)) offers a more sustainable ammonium removal solution, substantially lowering the energy demand and eliminating the addition of external carbon typically required for conventional side-steam treatment systems (Kaui & Verstraete 1998). 
formula
Reaction 1
 
formula
Reaction 2
 
formula
Reaction 3
Anammox-based processes stemmed from discovery of anammox biological pathway in a fluidized bed reactor (Mulder et al. 1995) and are classified into one or two reactor configurations. Two-stage systems are comprised of two sequential reactors, where partial nitrification takes place in the first tank, followed by anaerobic ammonium oxidation (Anammox) in the second tank (Abma et al. 2007).

Deammonification combines anammox and partial nitrification in a single tank, wherein partial nitrification and anammox occur simultaneously under controlled and low dissolved oxygen conditions. Various reactor configurations including sequencing batch reactors (Kaui & Verstraete 1998; Wett 2007), moving bed biofilm reactors (Szatkowska et al. 2007; De Clippeleir et al. 2013), Integrated Fixed Film Activated Sludge (IFAS) (Veuillet et al. 2014) and air lift ranular biomass reactors (Sliekers et al. 2003) have been used to support deammonification.

To date, anammox granular cultures have shown substantial removal rates reaching over 70 kg/m3d in the anammox part of the two-stage reactors (Tang et al. 2011). However, there is very limited research on deammonification utilizing granular biomass in continuous flow, particularly in scalable conventional reactors. In addition, there is a merging gap between the side-stream deammonification and its effect for enhanced mainstream nitrogen management, wherein controlling the side-stream reactor process parameters (such as solids retention time (SRT)) to maximize sludge seed production becomes desirable.

This study presents, for the first time, how a deammonification process is incorporated into a continuous flow-activated sludge process. The main objective of this research was to study the long-term performance of the process under field conditions in removing ammonium from high ammonium sludge dewatering centrate. At the same time, the process design addresses the inherent operational challenges associated with the successful operation of anammox-based nitrogen removal systems, which mainly include preventing nitrite build-up in the bioreactor and the management of nitrite-oxidizing bacteria (NOB) to minimize nitrate production. An innovative approach that uses the difference in settling velocities of anammox granules and AOBs to control the SRT of AOBs within the process was tested and studied.

MATERIALS AND METHODS

Feed

Digested sludge dewatering centrate from the Annacis Island Wastewater Treatment Plant in Delta, BC, Canada was used as a feed throughout the experiment. The Annacis Wastewater Treatment Plant is a 480,000 m3/d plant and uses trickling filter-solids contact process for secondary treatment. The thermophilic anaerobic digesters at Annacis digest the combination of primary sludge and biological sludge at 55 °C and operate at a retention time of approximately 20 days. Table 1 summarizes the centrate quality that was used throughout the experiment.

Table 1

Summary of the centrate quality used throughout the experiment

Parameter Abbreviation Unit Minimum Maximum Average 
Ammonia-nitrogen NH4-N mg/L 868 1,585 1,149 
Nitrite-nitrogen NO2-N mg/L 10.3 3.2 
Nitrate-nitrogen NO3-N mg/L 0.7 0.1 
Ortho-phosphate PO4-P mg/L 124 176 154 
pH pH NA 7.8 8.4 8.1 
Total suspended solids TSS mg/L 60 258 142 
Volatile suspended solids VSS mg/L 48 258 125 
Chemical oxygen demand COD mg/L 737 2,150 1,371 
Soluble biological oxygen demand BOD5 mg/L 230 300 265 
Alkalinity as CaCO3 ALK mg/L 2,960 4,040 3,513 
Parameter Abbreviation Unit Minimum Maximum Average 
Ammonia-nitrogen NH4-N mg/L 868 1,585 1,149 
Nitrite-nitrogen NO2-N mg/L 10.3 3.2 
Nitrate-nitrogen NO3-N mg/L 0.7 0.1 
Ortho-phosphate PO4-P mg/L 124 176 154 
pH pH NA 7.8 8.4 8.1 
Total suspended solids TSS mg/L 60 258 142 
Volatile suspended solids VSS mg/L 48 258 125 
Chemical oxygen demand COD mg/L 737 2,150 1,371 
Soluble biological oxygen demand BOD5 mg/L 230 300 265 
Alkalinity as CaCO3 ALK mg/L 2,960 4,040 3,513 

Experimental setup

Deammonification was incorporated into a pilot-scale, activated sludge process and evaluated for treating high ammonia centrate. The suspended growth granular-activated sludge process (Figure 1) was composed of a 2,000 L centrate storage tank, a feed pump, a bioreactor with a working volume of 380 L, an external clarifier with a working volume of 6.5 L, and a return-activated sludge (RAS) pump. The bioreactor was equipped with a propeller mixer which was operated constantly at 150 rpm, a disc aeration system and a baffle that separates granules from flocculent biomass before entering the external clarifier. The process components were hosted in a 6 m long, enclosed trailer.

Figure 1

Process schematics of the experimental setup.

Figure 1

Process schematics of the experimental setup.

The main control strategy relied on using a pH controller for controlling the feed flow/nitrogen load to the bioreactor and separate aeration control. The air was provided by an air blower connected directly to an on/off timer, which allowed for continuous or intermittent aeration. The dissolved oxygen during the aeration ON time was controlled using a DO controller and a solenoid valve. The airflow was recorded and adjusted manually using a rotameter in each stage of study. The temperature was controlled at 29–32 °C by means of an immersion heater and a thermostat.

The bioreactor was initially seeded with a mixture of waste biological sludge (1,200 mg/L TSS) and anammox biomass. The waste biological sludge was taken from a Membrane Enhanced UCT Process operated at the University of British Columbia Pilot Plant treating municipal wastewater. The seed biomass included 8 L of the anammox biomass, containing 3,200 mg/L of VSS (a mixture of AOB and anaerobic AOB (anammox)). The anammox biomass was collected over time from a bench scale anammox reactor treating high ammonia centrate as described by Kosari et al. (2014).

Centrate from the Annacis Wastewater Treatment Plant (Table 1) was fed to the bioreactor by a feed pump. The process was operated in three different stages. Table 2 summarizes the operational conditions at each stage.

Table 2

Summary of process operational conditions

Parameters Stage (I) start up Stage (II) high loading Stage (III) flocculent biomass wasting 
Operating length (days) 1–130 131–295 296–475 
Operating temperature (°C) 29–31 29–31 29–31 
Aeration (L/min) 8–14 14–41 35 
Dissolved oxygen (mg/L) 0.2–0.6 0–0.09 0.3–0.5 
Nitrogen loading (feed flow) Controlled at pH (6.7) Controlled at higher pH (7.7) Controlled at higher pH (7.7) 
Feed flow (mL/min) 75–110 110–305 Reduced back to 150 
Return-activated sludge (RAS) flow (%) 20–30% of influent flow 50–100% of influent flow RAS pump was off 
Parameters Stage (I) start up Stage (II) high loading Stage (III) flocculent biomass wasting 
Operating length (days) 1–130 131–295 296–475 
Operating temperature (°C) 29–31 29–31 29–31 
Aeration (L/min) 8–14 14–41 35 
Dissolved oxygen (mg/L) 0.2–0.6 0–0.09 0.3–0.5 
Nitrogen loading (feed flow) Controlled at pH (6.7) Controlled at higher pH (7.7) Controlled at higher pH (7.7) 
Feed flow (mL/min) 75–110 110–305 Reduced back to 150 
Return-activated sludge (RAS) flow (%) 20–30% of influent flow 50–100% of influent flow RAS pump was off 

In the first stage, the ammonium loading to the bioreactor was controlled by a pH controller at 6.7. The pH values shown in Table 2 represent the pH controller set point, not the actual pH in the bioreactor. The nitrogen loading was controlled by connecting the feed pump to the pH controller. The feed pump was always ON if the actual pH in the reactor was below the pH controller set point. The aeration was provided intermittently during the start-up phase (10 min on, 10 min off). Nitrite was monitored on a daily basis and airflow was adjusted accordingly to keep nitrite below 5 mg N/L.

In the second stage, once the higher quantity of anammox biomass was established, RAS flow and ammonium loading was gradually increased by increasing the flow, along with the aeration. The aeration timer was operated at continuous aeration from this stage forward.

In the last stage of the experiment, the process operated under maximum AOB wasting conditions by turning off the RAS pump. This resulted in raising the sludge blanket in the clarifier, which allowed for continuous wasting of flocculent biomass to the clarifier effluent, while keeping the granular biomass in the bioreactor.

Analysis

NH4+, NO2 and NO3 were determined by flow injection analysis of spectrophotometry (Quikchem 8000, Lachat). Alkalinity, biological oxygen demand (BOD), mixed liquor suspended solids (MLSS) and mixed liquor volatile suspended solids (MLVSS) were determined according to Standard Methods for the Examination of Water and Wastewater (APHA 1995). Chemical oxygen demand (COD) was measured using the closed reflux calorimetric method. Particle size distributions in the bioreactor were analyzed (using the Malvern Mastersizer 2000). The biomass settling velocity was measured in a settling tube by tracking granular and flocculent biomass in the settling tube.

RESULTS AND DISCUSSION

Granular biomass retention and baffle performance

The current state of the art concerning microbial community of granular deammonification systems suggests anammox bacteria tend to gather more in the granules or larger aggregates versus AOB and NOB that are found in more concentrations in the flocs (Nielsen et al. 2005; Sabine Marie et al. 2015). This biomass characteristics is now being utilized in over 30 full-scale SBR deammonification plants where hydrocyclones are used to separate and waste the flocculent biomass from granules for NOB management (Lacknera et al. 2014).

In this study, under different hydrodynamic conditions, the biomass was composed of two distinguished types of biomass; flocculent and granular (Figure 2(d1) and 2(d2)). Anammox granules were, on average, 1,000 micron in diameter versus the flocculent biomass averaging 100 micron.

Separating SRT of anammox granules from flocculent biomass was possible due to the difference in settling velocities of anammox and flocculent biomass. As it is shown by settling velocity profile in the bioreactor (Figure 2(c)), granules showed higher settling velocity ranging from 8 to 25 m/h, as opposed to flocculent biomass with a settling velocity of 2–8 m/h. Therefore, a simple baffle/internal clarifier allowed for retention of granules and separating the SRT of the anammox and flocculent sludge. RAS flow rate was used to control the flocculent sludge concentration/SRT in the bioreactor by changing the up-flow velocity within the baffle/internal clarifier zone. As expected, large granules were mainly retained in the bioreactor, confirmed by particle size distribution measurement in the baffle effluent (Figure 2(b)). Small quantities of granules occasionally passed to the external clarifier. However, these granules were captured and returned to the bioreactor by RAS.

Figure 2

(a) Particle size distribution of biomass (mixture of granular and flocculent biomass) in the bioreactor. (b) Particle size distribution in the bioreactor effluent. (c) Settling velocity profile of biomass in the bioreactor. (d) Photo of anammox granules (1) and flocculent biomass (2) in the bioreactor effluent.

Figure 2

(a) Particle size distribution of biomass (mixture of granular and flocculent biomass) in the bioreactor. (b) Particle size distribution in the bioreactor effluent. (c) Settling velocity profile of biomass in the bioreactor. (d) Photo of anammox granules (1) and flocculent biomass (2) in the bioreactor effluent.

Ammonium and total inorganic nitrogen removal

Figure 3 shows the long-term performance of the system for ammonium removal from centrate. The performance curve can be divided into three sections of controlled ammonium loading (Stage I), high ammonium loading (Stage II), and flocculent biomass wasting (Stage III). In the first stage, the feed pump was controlled by the pH to prevent high pH and high ammonia condition to occur in the reactor. This ammonium loading control strategy was aimed at providing the optimum growth conditions for anammox bacteria based on the established criteria by other researchers, where both free ammonia (Jung et al. 2007; Aktan et al. 2012; Jaroszynski et al. 2012; Jin et al. 2012) and high pH (Puyol et al. 2014) have been reported to slow down the anammox reaction.

Figure 3

Ammonium, nitrite, nitrate and dissolved oxygen in the bioreactor over time at different ammonium loading rates.

Figure 3

Ammonium, nitrite, nitrate and dissolved oxygen in the bioreactor over time at different ammonium loading rates.

In the first 3 months, ammonium was reduced from 1,140 to 170 mg/L on average, at an ammonium loading of 0.4 kg N/m3 d. The increased aeration from day 90 onwards (to 14 L/min) resulted in elevated nitrite concentration, and short-term nitrate build-up in the bioreactor (days 90–130). The elevated nitrate concentration is expected to be the main cause of decline in ammonium removal rate during this period. This can be explained by competition of NOB and anammox for nitrite. As shown by Reaction (2), 1 mole of ammonium reacts with 1.32 mole of nitrite and any portion of nitrite flow to NOB takes away its equivalent as ammonium from anammox bacteria.

Increasing the RAS flow from day 130 onwards increased the biomass concentration in the bioreactor, leading to the reduction of nitrate concentration in the bioreactor. This may be attributed to the increased concentration of small anammox granules that passed the baffle and accumulated in the clarifier sludge blanket. Increased concentration of flocculent biomass in the bioreactor (day 130) dropped the dissolved oxygen concentration to below 0.1 mg/L. The increase in flocculent biomass allowed for process operation under increased loading while operating under low dissolved oxygen conditions. The increased loading was achieved by proportional increase of feed and airflow reaching 305 mL/min and 41 L/min on day 300. Operating under higher feed flow (hydraulic retention time of 20 hours) resulted in 90% ammonium removal at a volumetric ammonium loading of 1.42 kg N/m3d.

In the last stage of operation (after day 300), the RAS pump was turned off. This led to continuous flocculent biomass accumulation and overflow from the external clarifier to the effluent. Increased sludge wasting is particularly desirable for mainstream bioaugmentation, where short retention time and maximizing the seed is required for process efficiency. The ammonium removal at the beginning of sludge wasting was lowered, due to NOB activity; however, NOBs were washed out due to continuous flocculent sludge wasting.

The granular biomass was still retained in the bioreactor in high concentrations by means of the internal baffle. Ammonium removal during this period was 90% at an average ammonium loading of 0.65 kg N/m3d. The total inorganic nitrogen removal rate was 1–3% lower than the ammonium removal, with the exception of short-term nitrate peaks observed at transition periods.

Nitrate production and NOB management

Figure 4 shows nitrate, nitrite and biomass concentration in the bioreactor. Nitrate production in the early stages of operations occurred when residual nitrite in the bioreactor started to increase (day 125). The combination of elevated nitrite concentration and low anammox concentration, in the early stages, allowed for NOB activity, where nitrate concentration reached 250 mg/L. The elevated nitrite in the bioreactor, ultimately relates to the non-synchronization of nitrite production rate by AOBs and nitrite consumption by the anammox bacteria. This suggested that the anammox bacteria biomass concentration was not adequate enough to out-compete the NOBs at the operating dissolved oxygen levels.

Figure 4

Flocculent VSS, nitrate and nitrite concentration over time.

Figure 4

Flocculent VSS, nitrate and nitrite concentration over time.

A low anammox bacteria biomass proved to be the case, as increasing the RAS flow suppressed nitrate production by increasing the anammox bacteria and AOB in the bioreactor. Although high AOB SRT in deammonification systems have led to increased NOB activities (Wett et al. 2010), the current study suggests that it can be managed by low dissolved oxygen and controlling the nitrite at very low concentrations, even at higher flocculent SRT in the bioreactor.

For the last stage of the work – short AOB SRT, nitrite production was not a concern, even at elevated dissolved oxygen and residual nitrite concentrations. The nitrate production was suppressed by continuous wasting of flocculent sludge resulting in NOB washout and retaining the granular sludge in the bioreactor.

CONCLUSIONS

A single-stage, granular sludge deammonification was successfully incorporated into a continuous flow, activated sludge process and was used for nitrogen removal from centrate.

Operation under different SRT for anammox and flocculent biomass was possible due to differentiated settling velocities of anammox and flocculent AOBs, where RAS pumping can successfully be used for SRT management.

The nitrogen loading control using pH, along with dissolved oxygen, allowed for minimizing the unstable condition of the bioreactor. Ammonium removal, on average, was 90% at ammonium loading rates between 0.65 and 1.42 kg/m3d. Higher concentrations of flocculent biomass allowed for operation under higher ammonium loading rates.

NOB activity was managed within a wide range of flocculent biomass concentration/SRT in the bioreactor, resulting in a nitrate production rate equal to 1–3% of ammonium loading; this was significantly lower than the 11% expected by anammox stoichiometry.

Under short flocculent SRT, NOB activity is suppressed within a wide range of dissolved oxygen and residual nitrite concentration. Increasing the flocculent biomass SRT narrows this range to low nitrite concentration and very low dissolved oxygen, which is below 0.1 mg/L. The increased SRT would allow for operation under high loading conditions as long as there is a high concentration of anammox biomass retained in the bioreactor.

ACKNOWLEDGEMENTS

The authors would like to thank Paul Kadota, Darlene Reigh, Ron Howell and Jeff Carmichael of Metro Vancouver for their support during operation of the pilot at Annacis Wastewater Center.

REFERENCES

REFERENCES
Abma
W. R.
Schultz
C. E.
Mulder
J. W.
van der Star
W. R. L.
Strous
M.
Tokutomi
T.
van Loosdrecht
M. C. M.
2007
Full-scale granular sludge Anammox process
.
Water Science and Technology
55
(
8–9
),
27
33
.
Aktan
C. K.
Yapsakli
K.
Mertoglu
B.
2012
Inhibitory effects of free ammonia on Anammox bacteria
.
Biodegradation
23
(
5
),
751
762
.
APHA
1995
Standard Methods for the Examination of Water and Wastewater
19th edn
.
American Public Health Association/American Water Works Association/Water Environment Federation
,
Washington, DC
,
USA
.
De Clippeleir
H.
Vlaeminck
S. E.
De Wilde
F.
Daeninck
K.
Mosquera
M.
Boeckx
P.
Verstraete
W.
Boon
N.
2013
One-stage partial nitritation/anammox at 15 °C on pretreated sewage: feasibility demonstration at lab-scale
.
Applied Microbiology and Biotechnology
97
(
23
),
10199
10210
.
Jaroszynski
L. W.
Cicek
N.
Sparling
R.
Oleszkiewicz
J. A.
2012
Impact of free ammonia on anammox rates (anoxic ammonium oxidation) in a moving bed biofilm reactor
.
Chemosphere
88
(
2
),
188
195
.
Jin
R.
Yang
G.
Yu
J.
Zheng
P.
2012
The inhibition of the Anammox process: A review
.
Chemical Engineering Journal
197
,
67
79
.
Jung
J. Y.
Kang
S. H.
Chung
Y. C.
Ahn
D. H.
2007
Factors affecting the activity of anammox bacteria during start up in the continuous culture reactor
.
Water Science and Technology
55
(
1–2
),
459
468
.
Kaui
L.
Verstraete
W.
1998
Ammonium removal by the oxygen-limited autotrophic nitrification-denitrification system
.
Applied Environmental Microbiology
64
(
11
),
4500
4506
.
Kosari
S. F.
Rezania
B.
Lo
K. V.
Mavinic
D. S.
2014
Operational strategy for nitrogen removal from centrate in a two-stage partial nitrification–Anammox process
.
Environmental Technology
35
(
9–12
),
1110
1120
.
Lacknera
S.
Gilberta
E. M.
Vlaeminckb
S. E.
Jossc
A.
Horna
H.
van Loosdrechtd
M. C. M.
2014
Full-scale partial nitritation/anammox experiences – an application survey
.
Water Research
55
,
292
303
.
Mulder
J. W.
van de Graafb
A. A.
Robertson
L. A.
Kuene
J. G.
1995
Anaerobic ammonium oxidation discovered in a denitrifying fluidized bed reactor
.
FEMS Microbiology Ecology
16
,
177
.
Nielsen
M.
Bollmann
A.
Sliekers
O.
Jetten
M.
Schmid
M.
Strous
M.
Schmidt
I.
Larsen
L. H.
Nielsen
L. P.
Revsbech
N. P.
2005
Kinetics, diffusional limitation and microscale distribution of chemistry and organisms in a CANON reactor
.
FEMS Microbiology Ecology
51
,
247
256
.
Puyol
D.
Carvajal-Arroyo
J. M.
Li
G. B.
Dougless
A.
Fuentes-Velasco
M.
Sierra-Alvarez
R.
Field
J. A.
2014
High pH (and not free ammonia) is responsible for Anammox inhibition in mildly alkaline solutions with excess of ammonium
.
Biotechnology Letters
36
,
1981
1986
.
Sabine Marie
P.
Pmpel
T.
Markt
R.
Murthy
S.
Bott
C.
Wett
B.
2015
Comparative evaluation of multiple methods to quantify and characterise granular anammox biomass
.
Water Research
68
,
194
205
.
Sliekers
A. O.
Third
K. A.
Abma
W.
Kuenen
J. G.
Jetten
M. S.
2003
CANON And Anammox in a gas-lift reactor
.
FEMS Microbiology Letters
218
(
2
),
339
344
.
Szatkowska
B.
Cema
G.
Plaza
E.
Trela
J.
Hultman
B.
2007
A one-stage system with partial nitritation and Anammox processes in the moving-bed biofilm reactor
.
Water Science and Technology
55
(
8–9
),
19
26
.
Tang
C.
Zheng
P.
Wang
C.
Mahmood
Q.
Zhang
J.
Chen
X.
Zhang
L.
Chen
J.
2011
Performance of high-loaded ANAMMOX UASB reactors containing granular sludge
.
Water Research
45
(
1
),
135
144
.
Veuillet
F.
Lacroix
S.
Bausseron
A.
Gonidec
E.
Ochoa
J.
Christensson
M.
Lemaire
R.
2014
Integrated fixed-film activated sludge ANITAMox process – a new perspective for advanced nitrogen removal
.
Water Science and Technology
69
(
5
),
915
922
.
Wett
B.
Hell
M.
Nyhuis
G.
Puempel
T.
Takacs
I.
Murthy
S.
2010
Syntrophy of aerobic and anaerobic ammonia oxidizers
.
Water Science and Technology
61
(
8
),
1915
1922
.