Nitrite accumulation from simultaneous free-ammonia and free-nitrousacid inhibition and oxygen limitation in a continuous-flow biofilm reactor

Seongjun Parka,d, Jinwook Chungb,*, Bruce E. Rittmanna and Wookeun Baec,*
5

a. Swette Center for Environmental Biotechnology, Biodesign Institute at Arizona State
University,

1001

South

McAllister

Avenue,

Tempe,

AZ

85287-5701

U.S.A,

Seongjun.Park@asu.edu, Rittmann@asu.edu

10

b. R&D Center, Samsung Engineering Co. Ltd., 415-10 Woncheon-Dong, Youngting-Gu,
Suwon, Gyeonggi-Do, Republic of Korea, jin-wook.chung@samsung.com

c. Department of Civil & Environmental Engineering, Hanyang University, Sa 1-Dong,
Ansan, Gyunggi-Do, Republic of Korea, wkbae@hanyang.ac.kr
15

d. Construction Technology Center, Samsung Construction and Trading, Yeoksam-Dong,
Gangnam-Gu, Seoul, Republic of Korea, sj93.park@samsung.com

* Author for correspondence: wkbae@hanyang.ac.kr
20

8139 (fax)

* Submission to:

* Running title: Nitrite accumulation in a CFBR
25

+82-31-400-5148 (tel) +82-31-417-

ABSTRACT
To achieve nitrite accumulation for shortcut biological nitrogen removal (SBNR) in a
biofilm process, we explored the simultaneous effects of oxygen limitation and free
ammonia (FA) and free nitrous acid (FNA) inhibition in the nitrifying biofilm.
5

We used

the multi-species nitrifying biofilm model (MSNBM) to identify conditions that should or
should not lead to nitrite accumulation, and evaluated the effectiveness of those conditions
with experiments in continuous flow biofilm reactors (CFBRs).

CFBR experiments were

organized into four sets with these expected outcomes based on the MSNBM: 1. Control,
giving full nitrification; 2. oxygen limitation, giving modest long-term nitrite build up; 3.
10

FA inhibition, giving no long-term nitrite accumulation; and 4. FA inhibition plus oxygen
limitation, giving major long-term nitrite accumulation.

Consistent with MSNBM

predictions, the experimental results showed that nitrite accumulated in sets 2 – 4 in the
short term, but long-term nitrite accumulation was maintained only in sets 2 and 4, which
involved oxygen limitation.
15

Furthermore, nitrite accumulation was substantially greater in

set 4, which also included FA inhibition.

However, FA inhibition (and accompanying

FNA inhibition) alone in set 3 did not maintained long-term nitrite accumulation.

NOB-

activity batch tests confirmed that little NOB or only a small fraction of NOB were present
in the biofilms for sets 4 and 2, respectively. The experimental data supported the previous
modeling results that nitrite accumulation could be achieved with a lower ammonium
20

concentration than had been required for a suspended-growth process. Additional findings
were that the biofilm exposed to DO limitation and FA inhibition was substantially denser
and probably had a lower detachment rate.

25

Keywords: ammonium oxidation, biofilm, free ammonia inhibition, oxygen limitation,
nitrite accumulation

2

1. Introduction

In recent years, nitrite accumulation has been spotlighted for its role in shortcut
biological nitrogen removal (SBNR) and anaerobic ammonium oxidation (Anammox)
5

(Chung et al. 2007; Strous et al. 1997).

By using nitrite as a primary electron accepter, the

SBNR process uses 40% less organic electron donor. The Anammox process uses nitrite
as an electron acceptor and ammonium as an electron donor to bring about total-N removal
without any organic donor. To ensure the practicality of both processes, the key is stable
nitrite accumulation in nitrification, which is achieved by securing ammonium-oxidizing
10

bacteria (AOB), but suppressing nitrite-oxidizing bacteria (NOB).
Nitrite accumulation has been associated with inhibition from high or low pH, free
ammonia (FA), free nitrous acid (FNA), low dissolved oxygen (DO), and combinations
(Jiang et al. 2011, Park et al. 2007; Park and Bae 2009; Park et al. 2010a, b; Hanaki et al.
1990; Bernet et al. 2001).

15

The pH can affect nitrification in two ways:

1) directly by

changing the enzyme’s reaction mechanism (Van Hulle et al. 2007; Park et al. 2007; Boon
and Laudelout 1962; Quinlan 1984), and 2) indirectly by changing the speciation of total
ammonium and total nitrite to the inhibitor forms, FA and FNA (Anthonisen e al. 1976;
Hellinga et al. 1999; Van Hulle et al. 2007; Carrera et al. 2004; Lee et al. 2004; Park et al.
2010c; Jiang et al. 2011). The FA concentration increases in a basic condition, but the

20

FNA concentration increases in an acidic condition.
Because a biofilm process can be advantageous to secure the accumulation of slowgrowing bacteria, such as nitrifiers (Bishop and Zhang 1995; Okabe et al. 1999; Rittmann
and Manem 1992), it can provide an advantage or a disadvantage when nitrite accumulation
is the goal.

25

The superior retention of slow-growing biomass in a biofilm is good for AOB,

but works in the wrong direction for NOB.

In addition, a decreased pH through the

biofilm caused by acid generation by AOB could help NOB survive by decreasing the FA
concentration, which is disadvantageous for nitrite accumulation. On the other hand,
depletion of DO inside of the biofilm may limit the activity of NOB, since the NOB are
more sensitive to low DO than are the AOB, which is advantageous for nitrite accumulation
3

(Bernet et al. 2001; Park et al. 2010c). Perez et al. (2009) and Bartrolí et al. (2010)
concluded that an oxygen affinity for AOB was the key parameter and stable complete
nitrite accumulation was maintained by a constant ratio of DO/TAN in the bulk liquid of
the biofilm reactor, respectively.
5

Recently, Park et al. (2010c) suggested that FNA

inhibition to accentuate nitrite accumulation can be increased by allowing the pH to
decrease in the biofilm.
Despite some ambiguity of what mechanisms are at work, researchers have reported
evidence that biofilm processes can accomplish nitrite accumulation (Fux et al. 2004;
Chung et al. 2007; Yamato et al. 2008; Perez et al. 2009; Park et al., 2010c; Brockmann and

10

Morgenroth, 2010).

In particular, Park et al. (2010c) developed the multi-species

nitrifying biofilm model (MSNBM), which has three biomass types -- AOB, NOB, and
inert biomass -- and can track the effects of DO, FA, and FNA inhibition on the growth of
the two groups of nitrifiers in the biofilm.

MSNBM simulation results explain that a

biofilm can be advantageous for accumulating nitrite while simultaneously maintaining a
15

low ammonium concentration, because FA inhibition can occur at the surface of the biofilm,
while FNA inhibition and oxygen limitation occur inside the biofilm. These factors can
be simultaneously regulated by aeration intensity, influent ammonium concentration, and
buffer concentration (Flora et al. 1999; Perez et al. 2009; Park et al. 2010c).
We explore the simultaneous effects of oxygen limitation and FA and FNA inhibition

20

in a nitrifying biofilm by operating a continuous-flow biofilm reactor (CFBR) with
different oxygen and influent ammonium concentrations. The experimental results are
compared with the simulated results of the MSNBM to ascertain whether the proposed
benefits of carrying out SBNR can be achieved in practice or not.

25

4

2. Material and Methods

2.1. Seeding microorganisms and mineral medium
The biomass used in all CFBRs originated from a 50-L sequencing batch reactor (SBR)
5

operated as a typical nitrification process:

2 cycles per day, with aeration time = 10 hr and

decant and filling times = 1 hr during one cycle.

In the 50-L reactor, the DO was

maintained over 4 mg/L, the temperature was 27  C  2 C , and the pH was held near 8
(  0.2 ) by injecting sodium bicarbonate to balance the alkalinity consumed by ammonium
oxidation.
10

After the mixture of seeding sludge and feeding solution was introduced into a CFBR,
it was recirculated for 5 days to allow the microorganism to attach onto the biofilm surface
of the reactor. After 5 days, CFBR operation began, and the original suspended sludge
from the SBR gradually washed out.

The feed solution to the SBR and CFBR contained

100 mgN/L of (NH4)2SO4 in a mineral medium that contained (in mg/L): K2HPO4 390,
15

MgSO4•7H2O 100, FeSO4•7H2O 4, CaCl2 8, MnSO4•H2O 10, NaHCO3 1020, and KCl 14.
This medium composition also was used for the activity batch tests of NOB.

2.2 Reactor configuration and operation
Each 1-L (working volume, 10cm×10cm×10cm) CFBR was made of polyacrylic
20

plastic (Fig. 1) and contained an air diffuser, a mixing paddle, and a detachable substratum
divided into four sections (each 5cm×5cm×1.5cm deep). The detachable substrata were
used to obtain a biofilm sample easily. The total biofilm surface area of the reactor,
including all geometric structures onto which the biofilm could accumulate, was 750 cm 2.
The CFBR was submerged in a water bath for temperature control. The reactor’s liquid

25

contents were completely mixed, and Fig. S1 in the Supporting Information (SI) shows
results of tracer tests that demonstrate complete mixing.

5

During a 60-day initiation period, four CFBRs were operated with identical conditions
that minimized inhibition.

After each CFBR was seeded with a one-liter mixture of

sludge from the SBR and feed medium containing 60 mgN/L of total ammonium nitrogen
(TAN), the CFBR was operated at a flow rate of 2 L/day and hydraulic retention time (HRT)
5

of 12hr with 6.5 mgDO/L, 30°C, and pH 7±0.5 in the bulk liquid.

The initial biomass

concentration was approximately 4,000 mgVSS/L.
Table 1 lists the operating conditions for each reactor. Reactor 1 was maintained with
the same operating condition throughout the full 200-day experimental period, and it served
as a control with minimal inhibition.
10

Reactor 2 was operated with a condition of potential

DO limitation: The DO concentration was reduced to 3.8 mg/L at day 61 and then further
reduced to 2.5 mg/L at day 161.

In Reactor 3, the influent TAN concentration was

increased from 50 mg/L to 150 mg/L to cause potential FA inhibition at a sufficient DO
concentration of 7.6 mg/L. Reactor 4 combined the conditions of reactors 2 and 3 to
impose both types of limitation/inhibition together. Table S1 of SI explains why these
15

conditions could have achieved the desired types of inhibition.

2.3 Analytical measurements
For all liquid samples, the concentrations of TAN, TNiN, and total nitrate nitrogen
(TNaN) were measured according to Standard Methods (APHA, AWWA, WEF, 1998).
20

Temperature, pH (Orion, 720A), and DO (Orion, 850) were detected using selective
electrodes. In case of mixed liquor volatile suspended solid (MLVSS) leaving the CFBR
in the early phase of seeding, it also was analyzed by the volatile suspended solid (VSS)
method in Standard Methods (APHA, AWWA, WEF, 1998).

25

2.4 Thickness of the biofilm
The thickness of the biofilm was measured by using a stereomicroscope linked to a
closed-circuit television (CCTV) (Olympus, SZ-CTV). Near the end of the experiments
(~ day 200), a detachable substratum from the bottom of each CFBR was carefully moved

6

to the microscopic stage to capture images perpendicular to the top of the detachable
substratum, as depicted in Fig. 1. The images of the biofilm captured by CCTV were
transferred to OriginPro 7.5 (OriginLab), which has a digital ruler.

The average thickness

of the biofilm was taken from 100 points having an arithmetic interval located along the
5

edge of substratum. As soon as an image was taken, the substratum piece was replaced
into the CFBR for further experiments.

2.5 NOB activity in the biofilm
NOB-activity assays were executed through batch tests for the specific nitrite10

utilization rate.

When all CFBR operations were finished, day 200, the biofilm was

removed by a soft brush and suspended by using a stirrer. A 15-mL volume of the
suspended-biofilm sample was taken for MLVSS measurement, and the rest of biomass
was used for the batch test. For the batch test to obtain the specific nitrite utilization rate,
NaNO2 was introduced to give a 30-mgTNiN/L initial concentration in a 1-L volumetric
15

flask that contained mineral medium, with pH 7 and temperature of 30°C.

The decline of

the nitrite concentration over time gave the specific nitrite-utilization rate.

2.6 Simulations with MSNBM
We used the MSNBM by Park et al. (2010c) to simulate the conditions of the
20

experiments. The MSNBM model is developed to simulate a shortcut nitrogen removal
(SBNR) in biofilm with different operational parameters: buffer, pH, loading, DO, etc.
The feature of model is including the pH model. One modification was that we included
suspended reactions by the seeding bacteria between days 0 and 5.

Table 2 shows the

kinetic parameter values used. The maximum substrate utilization rate ( q̂ ) and the half25

maximum-rate concentration (KS) were adjusted for temperature, as needed, by using 25°C
kinetic values from Rittmann and McCarty (2001) and temperature correction according to
Novak (1974).

Most kinetic parameters for AOB and NOB are from Rittmann and

McCarty (2001). Diffusion coefficients for nitrogen species and oxygen were taken from

7

Picioreanu et al. (1997).

The values related to direct pH, FA, and FNA inhibitions were

from Park et al. (2007) and Park and Bae (2009). Operating parameters, i.e., flow rate
(Q), volume of reactor (V), and biofilm surface area (A), are from the experiment of this
study.
5

The biofilm density was an estimated value that is presented in Results and

Discussion.
The simulation conditions to evaluate the experiments followed the same conditions in
Table 1.

In the case of Reactor 1, the influent TAN concentration, DO, and pH were

maintained with 60 mg/L, 8 mg/L and 7, respectively. When the operational parameters
were changed to those of Reactors 2 and 4 at day 61, the input values to begin the
10

simulation were the simulated results at day 61, i.e. biomass distribution and biofilm depth.
However, simulation results for Reactor 3 after day 61 used the original initial values, since
severe biofilm sloughing occurred and required a new seeding.

8

3. Results and discussion

3.1. Biofilm-development period of all reactors: days 0 to 60
Fig. 2 presents the experimental and simulated results from all reactors during the
5

biofilm-development period, days 0 to 60.

The four CFBRs gave similar trends of

nitrogen concentrations. Nitrification due only to the suspended seed occurred during the
initial 5 days, after which the biofilm started to take over nitrification as the VSS washed
out to below 10 mg/L.

Experimental results between days 10 and 20 fluctuated because of

intermittent plugging of the feeding tube in this period, but the reactors were stable after
10

day 20, when nitrification to NO3- was fully stabilized.
The lines in Fig. 2 show the results of simulation by the MSNBM.

Because the

tubing problem was not simulated, the results have the poor correlation between days 10
and 20, but match well with the experimental results thereafter. The biofilm thickness
(not shown in the figure) predicted by the MSNBM at day 5 was 17 μm, and it increased to
15

54 μm at day 10. The predicted FA inhibition value can be calculated with the form
1/(1+FA/KFA,AOB or NOB). Rapid biofilm growth was possible, because FA inhibition in the
reactor (calculated by 1/(1+FA/KFA)) was less than 20% of the maximum potential given in
Table S1 of supported information, which was insignificant for preventing biofilm growth.
As TAN decreased out to 10 days, FA inhibition upon of NOB further weakened, and full

20

nitrification was achieved.

3.2 Minimal limitation and inhibition in Reactor 1
The results for Reactor 1 after day 60 are shown in Fig. 3a. An unexpected problem
of DO control caused nitrite accumulation during days 60 to 100. However, reactor 1
25

returned to stable full nitrification after day 100, and the model simulations are nearly
identical with the experimental results.

We operated the system with a high surface

loading (1.6 kgN/1000 m2-d at 60 mgN/L), compared to a standard trickling filter (0.5-0.8
kgN/1000 m2-d), rotating biological contactor (0.2-0.6 kgN/1000 m2-d), and circulating bed

9

biofilm reactor (< 1 kgN/1000 m2-d) (Rittmann and McCarty 2001).

Despite the high

loading, the system was able to maintain full nitrification; one reason was that it did not
have any organic donor in the influent, which kept it relatively free from DO and space
competition with heterotrophs.
5

3.3 DO limitation in Reactor 2
Fig. 3b shows the experimental results of Reactor 2, which had DO limitation.

As

the DO concentration was decreased (from 3.8 mg/L from day 61 and then to 2.5 mg/L
from day 161), the effluent concentrations of ammonium and nitrite increased, while nitrate
10

decreased.

The simulated results showed that DO limitation should have caused prompt

decreases of TNaN and increases of TNiN and TAN, but the experimental results
responded more slowly and with considerable fluctuations.

Under more severe DO

limitation (2.5 mg/L), the full nitrification efficiency fell significantly to ~60%, since both
TAN and TNiN increased to approximately 10 mgN/L.
15

3.4. FA inhibition in Reactor 3
Reactor 3 was reseeded at day 61, since severe biofilm sloughing occurred at that time.
It was restarted with an influent TAN concentration of 150 mgN/L. Shown in Fig. 4a, the
experimental results in Reactor 3 demonstrate nitrite accumulation for only about 20 days.
20

Eventually, full nitrification set in and was stable.

The early responses in TAN and TNiN

concentrations resemble those in Fig. 2, but the peaks were higher due to increased TAN
loading.
The simulated curves in Fig. 4a capture the trend of experimental data well, although
ammonium and nitrate sometimes fluctuated in the experiments.
25

The higher TAN

concentration (150 mgN/L) and pH (= 8) in the influent might have caused temporary NOB
suppression at first.

In this case, the maximum potential inhibition values (that is,

1/(1+FA/KFA,AOB or NOB)) are 1/(1+15/10) = 0.4 for AOB and 1/(1+15/0.75) = 0.05 for NOB.
(See Table S1 in SI for details.)

Thus, NOB were more severely inhibited at the early

10

phase of elevated TAN concentration in the reactor (as high as ~80mg/L), and this was
responsible for nitrite accumulation at first.

However, inhibition of NOB was insufficient

after day 20, because the TAN concentration was decreased by AOB, which caused a
decrease of FA inhibition at the surface of the biofilm; furthermore, a pH decline inside the
5

biofilm minimized the impact of FA inhibition, while FNA inhibition always was
insufficient to prohibit NOB growth.

The maximum FNA inhibition value to NOB that

was shown around day 70 was calculated as 0.92 (that is, 92% of the non-inhibitory

10

reaction was secured), which was insufficient to prohibit net NOB growth.

These factors

eventually led to full nitrification in the experiments and the model output.

Bartrolí et al.

(2010) demonstrated the feasibility of maintaining stable complete nitrite accumulation by
maintaining the DO/TAN ratio in the bulk liquid of the biofilm reactor below 0.25. The
complete nitrification in the reactor 3 indirectly supported their suggestion as the DO/TAN
ratio in the bulk liquid was over 10.

15

3.5 FA inhibition and DO limitation in Reactor 4
Reactor 4 was operated under DO limitation and potential FA inhibition after day 60.
Fig. 4b shows that the TAN concentration in the reactor rapidly increased after elevation of
the influent concentration (from 60 to 150 mg/L), pH (from 7 to 8), and DO limitation
(from 6.5 to 3.8 mg/L). Though it slowly decreased with time, full nitrification did not

20

recover throughout the period of 3.8 mgDO/L, where the DO/TAN ratio was below 0.2 in
most cases. When the DO concentration was decreased further to 2.5 mg/L (day 160), the
TAN concentration further increased accordingly, resulting in almost complete suppression
of NOB activity. The AOB activity was also affected as shown with the decreased TNiN
accumulation in Fig. 4b. Comparing to the results in Fig. 4a (in which the reaction

25

conditions were identical except for the DO concentration that was two times higher), it
was clear that the DO concentration was one of the critical parameters to control nitritation
as indicated in the literature (Perez et al. 2009; Bartrolí et al., 2010).

Important to note

was that FA inhibition had been another critical parameter for nitritation in Fig.4b since Fig.

11

3b did not accumulate nitrite much in which the DO concentrations were identical but the
average FA concentration in the reactor was much lower.
The model curves are similar to the trends in the experimental data when DO was 3.8
mg/L (Fig. 4b). However, the profile of the TNaN was not an exact match, because the
5

DO concentration was not well controlled, having a standard deviation of ±1.2 mg/L.
Similarly, Park et al. (2010b) showed unstable TNiN accumulation when the actual DO
concentration was close to a minimum oxygen concentration (Park et al., 2010a) for AOB
survival or over a minimum oxygen concentration for NOB survival. When the DO
concentration was 2.5 mg/L (day 161-200), again the model simulation correctly captured

10

the overall trends in TAN and TNiN concentrations, although the data and model gave a
significant discrepancy.

One possibility for the discrepancy might be the complexity

delivered by the growth of heterotrophic bacteria on the soluble microbial products (SMP)
produced by the nitrifiers (Furumai and Rittmann, 1992; de Silva et al., 2000a,b; Merkey et
al. 2009). Another possibility is an increase in the biofilm density and thickness that
15

results in diffusion resistance of TAN into biofilm; we discuss biofilm density and
thickness in next section.
Park et al. (2010c) reported that a biofilm system might be able to maintain a stable
nitrification at a lower TAN concentration than a suspended system, since FNA inhibition
and DO limitation can deepen NOB suppression inside the biofilm.

20

A suspended system

needed approximately 80 mgTAN/L to suppress nitrite oxidation (Park et al., 2010b). while
we observed nitrite accumulation at 50±30 mgTAN/L from day 61 to day 160 under similar
DO and pH.

12

3.6. Measurement of biofilm depth and estimation of biofilm density
Fig. 5 shows MSNBM-simulated profiles of biofilm composition, as well as the
measured and simulated biofilm depths at the end of operation of each reactor. The
predicted results show that higher DO and substrate concentrations led to thicker biofilms,
5

and the simulated and experimental values for Reactors 1, 2, and 3 are very close to each
other. However, the experimental value is much larger for Reactor 4. The experimental
result for Reactor 4 appears to be inconsistent with the experimental result for the other
reactors. For example, the biomass accumulation for Reactor 4 ought to be considerably
less than for Reactor 3, since Reactor 3 had a substantially higher rate of nitrification. The

10

apparent inconsistency can be explained if the biofilm detachment rate for Reactor 4 were
considerably smaller than for the other reactors, which we discuss below.
The table that is part of Fig. 5 provides the amounts of each type of biomass computed
from the simulated results of day 200.

One key finding is that the mass of AOB and NOB

in Reactor 1 and 3 are almost the same. This shows that active biomass in the biofilm is
15

similar when the degree of nitrification is the same: full nitrification for both reactors.
In the model, we assume that the daily detached biofilm length is 6% of total biofilm length
(i.e., 0.06/day specific detachment rate) and the detachment phenomenon occurs at the
surface of biofilm.

Since 6% of biofilm length in Reactor 3 is 31.1 μm and that in Reactor

1 is 14.4 μm, the biomass detached from Reactor 3 is more active.
20

Since Reactors 1 and 3

have active-biomass fractions of 0.72 and 0.83, respectively, at the surface of the biofilm,
the ratio of active biomass detached from the surface is 1 : 2.5 (i.e., 0.72*14.4 μm :
0.83*31.1 μm).

This ratio is the same as the influent-loading ratio (1.6 kg/1000 m2-d : 4

kg/1000 m2-d = 1 : 2.5).

The correspondence of ratios indicates that the higher biomass

synthesis with higher loading was balanced by higher detachment of active biomass, even
25

though the total active biomass was similar in the two reactors. However, the higher
loading gave a substantially larger accumulation of inert biomass, which was located away
from the outer surface and comparatively free from detachment loss.
The AOB mass with oxygen limitation was similar in Reactor 2 and 4, but the NOB
mass was different.

Similar to the comparison of Reactors 1 and 3, loss from the surface

13

of the biofilm was the reason why the AOB mass was similar in Reactor 2 and 4.
However, Reactor 4 had complete suppression of NOB, while Reactor 2 had only partial
suppression of NOB; thus, Reactor 4 had no NOB.
The biofilm density was calculated by dividing the total biomass in the whole reactor
5

with the biofilm volume (volume = average depth × total surface area).

Table 3 shows

that the measured biofilm densities of Reactor 1, 2, and 3 were similar, giving an average
value of 20±2.5 mg/cm3, which is close to the value used in the simulations (18 mg/cm3,
Table 1). However, the density of Reactor 4 was much higher (38 mg/cm 3). While these
variations are in the range of biofilm density reported in the literature (e.g., Rittmann and
10

McCarty, 2001), the high density for Reactor 4 strongly affects the biofilm thickness.
Biofilm density can be affected by microbial species and physical forces (e.g., Christensen
and Characklis, 1990; Vieira et al. 1993; Trinet et al., 2001; Sharma et al., 2005; Garny et
al. 2008), and Laspidou and Rittmann (2004) presented a consolidation model to describe
the increase of density over time.

15

The results in Table 3 suggest that low DO

concentration and high TAN made the biofilm denser in Reactor 4, although the cause-andeffect relationship cannot be determined from these results alone.

Additional simulations

3

with the higher density (38 mg/cm ) produced a similar biofilm thickness (587 µm), but the
detachment rate had to be much smaller, 0.015/day.

20

3.7. Biofilm activity of each reactor
NOB activity in the biofilm was observed through batch tests, as shown in Fig. 6.
Since the slope of each line represents the specific nitrite utilization rate (q), the slope
should be smaller as the NOB of the biomass decreased. The slopes from the batch tests
(inset in Fig. 6) match well with the simulated results in Fig. 5 (the ratio of NOB/total

25

biomass). The concentration of NOB in the biofilm increased at higher DO concentration,
thus, giving a steeper slope. With DO limitation (Reactor 2 and 4), Reactor 4 had a nearly
zero slope, which indicated that Reactor 4 had little NOB due to DO limitation and FA
inhibition.

14

4. Conclusions
We explored the simultaneous effects of oxygen limitation and FA and FNA inhibition
in the nitrifying biofilm, giving special focus to testing the conditions that the MSNBM
identifies for nitrite accumulation in biofilms. CFBR experiments were organized into
5

four sets with these expected outcomes based on the MSNBM: 1. Control, giving full
nitrification; 2. oxygen limitation, giving modest long-term nitrite build up; 3. FA inhibition,
giving no long-term nitrite accumulation; and 4. FA inhibition plus oxygen limitation,
giving major long-term nitrite accumulation.

Consistent with MSNBM predictions, the

experimental results showed that nitrite accumulated in sets 2 – 4 in the short term, but
10

long-term nitrite accumulation was maintained only in sets 2 and 4, which involved oxygen
limitation.

Furthermore, nitrite accumulation was substantially greater in set 4, which also

included FA inhibition.

However, FA inhibition (and accompanying FNA inhibition)

alone in set 3 did not maintained long-term nitrite accumulation.

NOB-activity batch tests

confirmed that little NOB or only a small fraction of NOB were present in the biofilms for
15

sets 4 and 2, respectively. The experimental data supported the previous modeling results
that nitrite accumulation could be achieved with a lower ammonium concentration than had
been required for a suspended-growth process.

Additional findings were that the biofilm

exposed to DO limitation and FA inhibition was substantially denser and probably had a
lower detachment rate.
20

Acknowledgement
This work was partially supported by grant No R01-2001-00437 from the Korea Science &
Engineering Foundation (KOSEF), by the Korea Research Foundation Grant (KRF-2006352-D00122) funded by the Korean Government (MOEHRD) and by the MYRG project
25

from the University of Macau Research Committee.
Eunyoung Hong for their experimental helps.

15

We thank Mr. Gihyun Bae and Ms.

References
APHA, AWWA, WEF. 1998. Standard methods for the examination of water and
wastewater. 20th ed. Washington D.C.
5

Bartrolí A, Pérez J, Carrera J. 2010. Applying ratio control in a continuous granular reactor
to achieve full nitritation under stable operating conditions. Environ. Sci. Technol.
44:8930–8935.
Bernet N, Dangcong P, Delgenes JP, Moletta R. 2001. Nitrification at low oxygen
concentration in biofilm reactor. J. Environ. Eng.-ASCE. 127(3), 266-271

10

Bishop PL, Zhang TC, Fu YC. 1995. Effects of biofilm structure, microbial distributions
and mass-transport on biodegradation processes. Water Sci. Techol. 31(1), 143-152.
Boon B, Laudelout H. 1962. Kinetics of nitrite oxidation by Nitrobacter winogradskyi.
Biochem. J. 85(3), 440-447

15

Brockmann D, Morgenroth E. 2010. Evaluating operating conditions for outcompeting
nitrite oxidizers and maintaining partial nitrification in biofilm systems using biofilm
modeling and Monte Carlo filtering. Water Res. 44:1995–2009.
Carrera J, Jubany I, Carvallo L, Chamy R, Lafuente J. 2004. Kinetic models for nitrification
inhibition by ammonium and nitrite in a suspended and an immobilized biomass systems.
Process Biochem. 39(9), 1159-1165

20

Christensen BE, Characklis WG., 1990. Physical and chemical properties of biofilm. In:
Biofilms. New York , John Wiley and Sons
Chung J, Bae W, Lee YW, Rittmann BE., 2007. Shortcut biological nitrogen removal in
hybrid biofilm/suspended growth reactors. Process Biochem. 42(3), 320-328.
Flora EMCV, Suidan MT, Flora JRV, Kim BJ. Speciation and chemical interactions in
nitrifying biofilms. I: Model development. J Environ Eng, 1999, 125(9):871–877.

25

30

Furumai H., Rittmann BE. 1992. Advanced modeling of mixed populations of heterotrophs
and nitrifiers considering the formation and exchange of soluble microbial product. Wat.
Sci. Tech. 26 (3): 493-502
Fux C, Huang D, Monti A, Siegrist H. 2004. Difficulties in maintaining long-term partial
nitritation of ammonium-rich sludge digester liquids in a moving-bed biofilm reactor
(MBBR). Water Sci. Technol. 49(11-12), 53-60
Garny K, Horn H, Neu TR, 2008. Interaction between biofilm development, structure and
detachment in rotating annular reactors, Bioprocess. Biosyst. Eng. 31(6), 619–629.
Hanaki K, Wantawin C, Ohgaki S. 1990. Nitrification at low-levels of dissolved-oxygen
16

with and without organic loading in a suspended-growth reactor. Water Res. 24(3), 297-302.
Hellinga C, Van Loosdrecht MCM, Heijnen JJ. Model based design of a novel process for
nitrogen removal from concentrated flows. Math. Comp. Model Dyn. 5(4), 351-371
5

Jiang, G., Gutierrez, O., Yuan, Z., 2011. The strong biocidal effect of free nitrous acid on
anaerobic sewer biofilms. Water Res. 45(12), 3735-3743
Laspidou CS, Rittmann BE, 2004. Modeling the development of biofilm density including
active bactera, inert biomass, and extracellular polymeric substances. Water Res. 38, 33493361

10

Lee WK, Chung J, Bae W, Park S, Kim Y, Lee YW, Park DW, 2004. Operational factor for
nitrite accumulation from a mixed culture by cell-immobilization. J. Ind. Eng. Chem. 10(6),
959-966
Merkey BV, Rittmann BE, Chopp DL, 2009. Modeling how soluble microbial products
(SMP) support heterotorpic bacteria in autotroph-based biofilms. J. Theor. Biol. 259: 670683

15

20

Okabe, S., Satoh H., Watanabe Y., 1999. In situ analysis of nitrifying biofilms as
determined by in situ hybridization and the use of microelectrodes. Appl. Environ.
Microbiol. 65, 3182-3191
Park S, Bae W. 2009. Modeling kinetics of ammonium oxidation and nitrite oxidation under
simultaneous inhibition by free ammonia and free nitrous acid. Process Biochem. 44(6),
631-640
Park S, Bae W, Chung J, Baek S-C. 2007. Empirical model of the pH dependence of the
maximum specific nitrification rate. Process Biochem. 42, 1671-1676

25

Park S, Bae W, Rittmann BE. 2010a. Operational boundaries for nitrite accumulation in
nitrification based on minimum/maximum substrate concentrations that include effects of
oxygen limitation, pH, and free ammonia and free nitrous acid inhibition, Environ. Sci.
Technol. 44(1), 334-342
Park S, Bae W, Rittmann BE, Kim SJ, Chung JW, 2010b. Operation of suspended-growth
shortcut biological nitrogen removal (SSBNR) based the minimum substrate concentration,
Water Res. 44(5), 1419-1428

30

Park S, Bae W, Rittmann BE, 2010c. Multi-species nitrifying biofilm model (MSNBM)
including free ammonia and free nitrous acid inhibition and oxygen limitation, Biotech.
Bioeng. 105(6), 1115-1130
Perez J. Costa E. Kreft J-U. 2009. Conditions for partial nitrification in biofilm reactors and
a kinetic explanation, Biotech. Bioeng. 103(2), 282-295.

17

Picioreanu C, vanLoosdrecht MCM, Heijnen JJ. 1997. Modelling the effect of oxygen
concentration on nitrite accumulation in a biofilm airlift suspension reactor. Water Sci
Technol. 36(1), 147–156.
5

Quinlan AV. 1984. Prediction of the Optimum-pH for ammonia-N oxidation by
Nitrosomonas-europaea in Well-Aerated Natural and Domestic-Waste Waters. Water Res.
18(5), 561-566
Rittmann BE, Manem JA. 1992. Development and experimental evaluation of a steadystate, multispecies biofilm model. Biotech. Bioeng, 39(9), 914-922.

10

Rittmann BE, McCarty PL. 2001. Environmental biotechnology: principles and
applications. Boston: McGraw-Hill.
de Silva DGV., Rittmann BE, 2000a. Nonsteady-state modeling of multispecies activatedsludge processes. Water Environ. Res. 72, 554–565
de Silva DGV., Rittmann BE, 2000b. Interpreting the response to loading changes in a
mixed-culture completely stirred tank reactor. Water Environ. Res. 72, 566–573

15

Sharma PK, Gibcus MJ, van der Mei HC, Busscher HJ. 2005, Influence of fluid shear and
microbubbles on bacterial detachment from a surface. Appl Environ Microbiol 71, 3668–
3673
Snoeyink VL, Jenkins D. 1980. Water Chemistry. Canada: John Wiley and Sons. p 463.

20

Strous M, VanGerven E, Zheng P, Kuenen JG, Jetten MSM. 1997. Ammonium removal
from concentrated waste streams with the anaerobic ammonium oxidation (anammox)
process in different reactor configurations. Water Res. 31(8), 1955-1962.
Trinet, F., Heim R., Amar D., Chang HT, Rittmann BE. 1991. Study of biofilm and
fluidization of bioparticles in a three-phase liquid-fluidized-bed reactor. Water Sci. Technol.
23, 1347-1354.

25

Van Hulle SWH, Volcke EIP, Teruel JL, Donckels B, van Loosdrecht MCM, Vanrolleghem
PA. 2007. Influence of temperature and pH on the kinetics of the Sharon nitritation process.
J. Chem. Technol. Biot. 82(5), 471-480
Vieira MJ, Melo LF, Pinheiro MM, 1993. Biofilm formation: hydrodynamic effects on
internal diffusion and structure. Biofouling 7, 67–80

30

Yamamoto T, Takaki K, Koyama T, Furukawa K. 2008. Long-term stability of partial
nitritation of swine wastewater digester liquor and its subsequent treatment by Anammox.
Bioresource Technol. 99(14), 6419-6425

18