Available online at www.sciencedirect.

com

Bioresource Technology 99 (2008) 5867–5872

Simultaneous nitrification–denitrification achieved by an

innovative internal-loop airlift MBR: Comparative study
a,*
Y.Z. Li , Y.L. He b, D.G. Ohandja a, J. Ji a, J.F. Li a, T. Zhou a

a
Institute of Environmental Science and Engineering, Nanyang Technological University, 18 Nanyang Drive, Singapore 637723, Singapore
b
Department of Environmental Science and Engineering, Xi’an Jiaotong University, Xi’an 710049, PR China

Received 1 August 2007; received in revised form 28 September 2007; accepted 2 October 2007
Available online 19 November 2007

Abstract

This study assessed the performance of different single-stage continuous aerated submerged membrane bioreactors (MBR) for nitro-
gen removal. Almost complete nitrification was achieved in each MBR irrespective of operating mode and biomass system. Denitrifica-
tion was found to be the rate-limiting step for total nitrogen (T-N) removal. The MBR with internal-loop airlift reactor (ALR)
configuration performed better as regards T-N removal compared with continuous stirred-tank reactor (CSTR). It was demonstrated
that simultaneous nitrification and denitrification (SND) is the mechanism leading to nitrogen removal and the contribution of micro-
environment on SND is more remarkable for the MBRs with hybrid biomass. Macroenvironment analyses showed that gradient distri-
bution of dissolved oxygen (DO) level in airlift MBRs imposed a significant effect on SND. Higher mixed liquor suspended solid (MLSS)
concentration led to the improvement in T-N removal by enhancing anoxic microenvironment. Apparent nitrite accumulation coupled
with higher nitrogen reduction was accomplished at MLSS concentration exceeded 12.6 g/L.
Ó 2007 Elsevier Ltd. All rights reserved.

Keywords: Airlift bioreactor; Dissolved oxygen; Membrane bioreactor; Powdered activated carbon; Simultaneous nitrification–denitrification

1. Introduction of recalcitrant compounds and microbial metabolic prod-

ucts (Jun et al., 2003), which may eventually lead to the dif-
The application of membrane bioreactor (MBR) tech- ferences in nitrogen removal characteristics between
nology has been gaining increased acceptance in wastewa- membrane bioreactor and conventional activated sludge
ter treatment in the past few years due to a significant cost process.
reduction in membrane production. By integrating mem- Currently, total nitrogen (T-N) removal in MBRs is
brane filtration and biological process, MBR possess sev- mostly achieved either in a two-stage continuous flow
eral distinctive characteristics, such as long sludge ages, anoxic/oxic system or in an alternating anoxic/oxic system
low organic loading rates and high biomass concentration. produced by intermittent aeration (Kraume et al., 2005).
However, nitrogen removal in most MBRs is generally not This is generally based on the principle that nitrification
satisfactory as compared with COD removal. In conven- requires an aerobic condition for oxidation of ammonia,
tional activated sludge process, the loss of nitrifier is always whereas denitrification occurs under anoxic condition in
unavoidable due to periodical sludge discharge under the presence of electron donors (Tchobanoglous et al.,
shorter sludge retention time (SRT). In comparison, the 2004). The differences in environmental requirements indi-
presence of membrane in bioreactor not only prevents the cate the necessity of keeping these two processes separated
washout of nitrifiers, but also results in the accumulation spatially or temporally. The spatial separation process,
which requires two independent tanks for nitrification
and denitrification respectively, result in a larger footprint,
*
Corresponding author. Tel.: +65 67905366; fax: +65 67921291. installation of recirculating pipeline and complicated
E-mail address: [email protected] (Y.Z. Li). operating procedures. The temporal separation process is

0960-8524/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.biortech.2007.10.001
5868 Y.Z. Li et al. / Bioresource Technology 99 (2008) 5867–5872

compact, reducing the consumption for alkalinity and tor. Constant flux was maintained by frequently adjusting
energy. However, membrane filtration conducted only the rotation rate of the suction pump.
within the aeration period in order to take advantage of Synthetic wastewater was used as the influent with glu-
air bubbles for membrane fouling control (Nah et al., cose, starch, ammonium chloride and sodium bicarbonate
2000). being the macro nutrients while peptone, KH2PO4,
Recently, studies have shown that nitrification and deni- MgSO4 Æ 7H2O, MnSO4 Æ 7H2O, CaCl2 and FeSO4 were
trification can occur concurrently in the same reactor used as the trace nutrients. Seed activated sludge, collected
through a process commonly known as simultaneous nitri- from a municipal wastewater treatment plant, was incu-
fication and denitrification (SND) (Helmer and Kunst, bated and acclimated in synthetic wastewater for three
1998; Collivignarelli and Bertanza, 1999; Guo et al., weeks before the experiments. Two MBR were set to oper-
2005). SND is achieved by allowing nitrifiers and denitrifi- ate in parallel for a period of 331 days. The first MBR,
ers (two species of bacteria with different growth environ- called AS-MBR, contained conventional activated sludge;
ment) to coexist in a single bioreactor. Compared to the second MBR contained biological powdered activated
nitrogen removal through conventional nitrification and carbon (BPAC) sludge referred to as BPAC-MBR. The
denitrification, SND offers several advantages: (1) it elimi- BPAC sludge, which was mainly used to mitigate mem-
nates the need for either two separate tanks operated in brane fouling, was developed by intermittent addition of
serial or intermittent aeration operation in a single tank, powdered activated carbon (PAC) in the activated sludge
thus continuous effluent output can be achieved with a mixed liquor in the bioreactor. Fresh PAC with an amount
smaller footprint (Guo et al., 2005); (2) it utilizes 22–40% equal to that lost from the discharged excess sludge was
less carbon source and reduces sludge yield by 30% (Turk supplied immediately to maintain a constant 1.2 g/L of
and Mavinic, 1987); (3) neutral pH level and less demand PAC concentration in the bioreactor.
for alkalinity can be accomplished in the reactor because The operating modes and biomass systems of the MBRs
alkalinity is consumed during nitrification but produced investigated during the entire operation are shown in Table
during denitrification (Chang and Tseng, 1999) and (4) it 1.Three kinds of operating modes were used in the AS-
consumes less energy due to the reduction in aeration MBR at different stages respectively to afford comparison.
requirement (Collivignarelli and Bertanza, 1999). After 231 days, fibrous carrier with the composite material
Up till now, few studies have been conducted towards of polythene and polyamide was introduced in the down-
integration of nitrification and denitrification in MBR. In comer compartment of AS-MBR for development of the
this study, an innovative single-stage continuously aerated hybrid biomass composed of biofilm and suspended acti-
MBR with internal-loop airlift reactor (ALR) configura- vated sludge.
tion was developed for simultaneous nitrification and deni-
trification of synthetic domestic wastewaters. More than 2.2. Analytical method
90% COD removal efficiency was obtained throughout this
study. Comparison was made on various systems in which Dissolved oxygen (DO) concentration was measured by
the presence of biomass was in different forms. It was dem- a DO meter (OxiCal-SL 197, WTW, Germany). Particle
onstrated that the MBRs could be effectively designed to size analysis was conducted using the Mastersizer 2000
achieve nitrogen removal utilizing SND. (Malvern, UK). A scanning electron microscope (JSM-
840, JEOL, Japan) was used to visualize the morphology
2. Methods of biofilm. Total nitrogen, ammonia, nitrate, nitrite and
mixed liquor suspended solid (MLSS) were analyzed
2.1. Experimental set-up and operating conditions according to standard methods (APHA, 1998).

Two identical submerged MBRs with a configuration

similar to internal-loop airlift reactor were used in this
study. Each reactor was made of plexiglass with a working
volume of 17.2 L. A baffle separated the bioreactor into Table 1
two parts: riser compartment and down-comer compart- Operating modes and biomass systems of the MBRs investigated
ment, which are connected by a bottom flow channel and System Reactor Operating mode Biomass Operating
an overflow channel. The membrane module, located in time (d)
the riser, is made of polyvinylidene fluoride (PVDF) hol- I AS- Internal-loop Suspended 180
low-fiber membranes with pore size of 0.2 lm and filtration MBR airlift activated sludge
II AS- Continuous Suspended 51
area of 0.4 m2. Air was continuously supplied from an aer-
MBR stirred-tank activated sludge
ation pipe located directly below the membrane module for reactor
biomass growth and scouring the membrane surface. The III AS- Internal-loop Hybrid biomass 100
filtrate was obtained by using a vacuum pump with a suc- MBR airlift
tion mode of 8-min-on and 2-min-off. A water level sensor IV BPAC- Internal-loop Suspended BPAC 331
MBR airlift activated sludge
was used to maintain a constant water level in the bioreac-
 Y.Z. Li et al. / Bioresource Technology 99 (2008) 5867–5872 5869

3. Results and discussion 200 100

180
3.1. Nitrogen removal
160 80

Total nitrogen (mg/L)

Fig. 1 shows the profiles of the influent concentration, 140

Removal (%)
effluent concentrations and removal efficiencies of ammo- 120 60
nia nitrogen in AS-MBR and BPAC-MBR during the per- 100
iod of the experiments.
80 40
As shown in Fig. 1, low ammonia removal efficiencies
occurred in both reactors during the initial stage. Low pop- 60
ulation of nitrifier, resulting from the slow specific growth 40 20
rate of nitrifier, accounted for the low nitrification effi-
20
ciency. However, after 23 days, both reactors achieved
about 77% ammonia removal and from the 32nd day 0 0
0 50 100 150 200 250 300 350
onwards, about 84–96% ammonia nitrogen removal were Time (day)
observed in both reactors regardless of the operating mode. Influent Effluent (AS-MBR)
Introduction of membrane in the bioreactors promoted Effluent (BPAC-MBR) Removal (AS-MBR)
Removal (BPAC-MBR)
complete retention and enrichment of nitrifier (Silva
et al., 1998), which ultimately led to sufficient nitrification. Fig. 2. Profiles of total nitrogen concentrations and the corresponding
Fig. 2 shows the variations of the influent concentration, removal efficiencies in AS-MBR and BPAC-MBR.
effluent concentrations and removal efficiencies of total
nitrogen in AS-MBR and BPAC-MBR. Moderate but fluc-
tuating total nitrogen removals were observed in both reac- nitrogen removal and this resulted from a deficiency in
tors during the period of the experiments. To evaluate the anoxic condition. The lowest T-N removal in the MBR
total nitrogen removal of the different systems as shown in with CSTR configuration was due to a significant inhibi-
Table 1, the average total nitrogen removal efficiencies of tion of denitrification by completely and uniformly distrib-
every system were further compared under the approxi- uted aerobic condition. For the airlift MBRs, the highest
mate operating conditions and significant difference in total T-N removal was achieved in the hybrid biomass system.
nitrogen removal was found in the four MBR systems. The In the case of the four MBR systems, the removals of
average total nitrogen removals of the MBRs with ALR total nitrogen were accomplished in each continuously aer-
configuration (39.4–63.1%) were always higher than that ated single tank and this indicates the occurrence of simul-
with CSTR configuration (18.7%). Since complete nitrifica- taneous nitrification and denitrification in the aerobic
tion was accomplished in all systems after the initial stage, MBRs. Generally DO level is regarded as the essential fac-
denitrification is determined to be the rate-limiting step for tor for the occurrence of SND (Priyali and Dental, 1998)
and two hypotheses are currently considered to determine
its mechanism: (1) SND via stratification of the reactor—
180 100 nitrification and denitrification occurring in different
160
locations in the reactor; (2) SND via stratification of the
activated sludge flocs or biofilm—nitrification and denitri-
140 80
fication zones exists on a microscopic scale. Hence analyses
of the environment conditions from aspects of the reactor
Ammonia (mg/L)

120
Removal (%)

60 and flocs/biofilm are necessary for better understanding

100
the mechanism of SND in the MBRs.
80
40
60 3.2. Microscopic environment perspective
40 20 Pochana et al. (1999) showed that simultaneous nitrifi-
20 cation and denitrification could occur as a consequence
0 0 of DO concentration gradients within microbial flocs or
0 50 100 150 200 250 300 350 biofilm. Based on this, anoxic microenvironments develop
inside the flocs/biofilm, whereas aerobic condition prevails
Time (day)
on the outer surface layer. Nitrifiers will then develop on
Influent Effluent(AS-MBR) the surface of flocs/biofilm with fairly high DO levels, while
Effluent(BPAC-MBR) Removal(AS-MBR) denitrifiers exist inside the flocs/biofilm due to absence of
Removal(BPAC-MBR) oxygen. As a result, the effectiveness of SND is dependant
Fig. 1. Profiles of ammonia nitrogen concentrations and the correspond- on the size, density and distribution of flocs or the thick-
ing removal efficiencies in AS-MBR and BPAC-MBR. ness of biofilm (Pochana et al., 1999).
5870 Y.Z. Li et al. / Bioresource Technology 99 (2008) 5867–5872

10 fibrous carriers in the down-comer for attached growth

of microorganism, considerable amount of biomass is
Distribution frequency (Volume %)

immobilized in the bubble free anoxic compartment. Due

8
to the absence of shear stress, a thick and rough biofilm
layer developed gradually and this led to an increase in
6 oxygen transfer resistance. As a result, a stable anoxic zone
developed within the biofilm allowing sufficient growth and
enrichment of denitrifiers. Therefore, the MBR with hybrid
4 biomass can uncouple the restrictive factors affecting SND
and encourage optimal growth of different microorganism.
Coexistence of various microorganisms in a single reactor
2
in turn reduces the footprint of wastewater treatment pro-
cess while increasing nitrogen removal rate.
0
1 10 100 1000
3.3. Macroscopic environment perspective
Floc size (µm)
AS-MBR The effect of macroenvironment of dissolved oxygen on
BPAC-MBR
inoculated activated sludge
SND includes two factors: DO level and distribution. Bliss
and Barnes (1986) reported 0.2 mg/L as the critical DO
Fig. 3. Floc size distribution of inoculation sludge and activated sludge value, below which nitrification does not occur, whereas,
from AS-MBR and BPAC-MBR.
Pochana and Keller (1999) found that complete SND
occurred at DO = 0.3 mg/L, but below this, denitrification
To understand the effect of microenvironments on SND will prevail. In order to access the anoxic volumetric pro-
in the airlift MBRs, identifications of floc size distribution portion of the airlift MBRs, a scale factor f, defined as
were performed simultaneously on the suspended activated the volumetric ratio between the anoxic zone and the entire
sludge from AS-MBR and BPAC-MBR. As shown in reactor, was adopted using 0.3 mg/L DO as the critical
Fig. 3, the average floc size in AS-MBR and BPAC- level. DO data were monitored at six equally divided loca-
MBR were 47 lm and 56 lm respectively. In comparison, tions along both sides of the baffle.
the average floc size of the seed activated sludge was As shown in Table 2, gradient distribution of DO levels
83 lm. The reduction of floc size in the two MBRs was occurred in all airlift MBRs, artificially producing an alter-
caused by the shear stress continuously imposed on the nating oxic and anoxic conditions in a single tank. In this
sludge. Pochana et al. (1999) found that small flocs (up study, the design of the MBRs was based on internal-loop
to about 60 lm) show only limited decrease in DO gradi- airlift reactor configuration with aeration fully conducted
ents and the internal DO concentrations remain close to in the riser, a local complete mixed regime of solid, liquid
the external. In another experiment, they further demon- and gas was attained in this compartment. In contrast,
strated that the floc size range of 50–110 lm could create due to absence of bubble, plug flow regime can be consid-
an anoxic zone inside the flocs leading to denitrification ered to describe the hydraulic properties in the down-
in this region (Pochana and Keller, 1999). In other words, comer. As a result, DO concentration gradients were pro-
large flocs of activated sludge have much more potential to duced within the whole reactor and low DO levels were
develop significant anoxic zones internally than small flocs. observed in the down-comer and the bottom zone of the
According to these findings, the floc sizes in both AS-MBR reactor.
and BPAC-MBR fall into the lower limit of the range Besides, it was observed that the f values of the hybrid
where anoxic microenvironments was formed. Therefore, biomass MBR (System III) were always higher than those
the contribution of microenvironments on SND was not of the suspended biomass MBRs (System I and System IV).
remarkable for the MBRs with suspended growth biomass. The higher f values in the hybrid biomass MBR were due
In contrast, higher total nitrogen removal was achieved to sufficient depletion of oxygen by a large amount of
in the airlift MBR with hybrid biomass (63.1%) than that attached growth biomass accumulated in the bubble free
with suspended growth biomass (39.4%). By introducing down-comer. Therefore, the extent of denitrification was

Table 2
Effect of aeration rates and the overflow liquid levels on the anoxic volumetric ratio f in airlift MBRs
Overflow liquid level H (cm) 0.5 2 4
Aeration rate UG (m3/m2 h) 38 62 85 38 62 85 38 62 85
Anoxic volumetric ratio f (%) AS-MBR (System I) 38 32 30 31 23 20 24 19 16
AS-MBR (System III) 47 43 41 40 35 30 33 28 23
BPAC-MBR (System IV) 41 36 33 33 25 22 26 21 17
 Y.Z. Li et al. / Bioresource Technology 99 (2008) 5867–5872 5871

determined by both the ratio of the anoxic zone and the Table 3
amount of biomass in the corresponding region. This also The average ammonia and total nitrogen removal in AS-MBR (System I)
and BPAC-MBR (System IV) under different MLSS levels
accounts for the fact that the MBRs with ALR configura-
tion exceed their CSTR counterpart in T-N removal. MLSS (g/L) 4.1– 8.5– 12.6–
5.5 9.6 13.5
Table 2 also shows that both aeration rates and the over-
flow liquid levels affected DO levels in the reactors. In the air- Corresponding SRT (d) 15 30 40
lift reactors, the driving force leading to recirculation of the AS-MBR (System I) f 17 29 38
mixed liquor is generated by the density difference of the NHþ 4 -N removal 98.4 95.9 91.7
(%)
fluids between the riser and down-comer compartment. T-N removal (%) 30.1 50.3 63.6
Increasing aeration rate can improve gas hold-up and NO 2 /NOx

4.2 11.2 67.1
decrease fluid density. Since aeration restrictedly conducted
BPAC-MBR (System f 19 27 39
in the riser, the contribution to gas hold-up from the increase IV) NHþ 4 -N removal 98.8 97.0 94.2
of aeration rate is different for the two compartments. Con- (%)
sequently the gap in fluid density between the two compart- T-N removal (%) 34.3 55.5 64.5
ments can be enhanced under higher aeration rate, which will NO 2 /NOx

3.4 8.3 53.7
subsequently increase the fluid recirculation rate and
improve the comprehensive DO level in the reactor. As a IV), comparisons were made on average removal efficien-
result, a decline of the anoxic volumetric ratio f was induced. cies of ammonia and total nitrogen under different MLSS
The recirculation flow rate in the airlift reactor can be deter- levels and approximate operating conditions.
mined by Chisti equation derived from an energy balance As shown in Table 3, with the increase of MLSS concen-
analysis (Yusuf and Murray, 1993; Chisti et al., 1995). trations, NHþ 4 -N removals reduced slightly whereas T-N
2 31=2 removals improved remarkably in both systems. The
decline in ammonia removal was attributed to the reduc-
6 2ghðer  ed Þ 7
U Lr ¼ 4  2 5 ð1Þ tion of the availability of dissolved oxygen for nitrifier
K B AAdr ð1e1 Þ2 (Guo et al., 2005). When MLSS was within 4.1–5.5 g/L,
d
low anoxic volumetric ratio f was observed, which led to
where ULr is the superficial liquid velocity in the riser, g is the strong oxygen penetrating capability and small anoxic
gravitational acceleration, h is the height of the gas–liquid micro zones inside the flocs. As a result, high nitrification
dispersion, Ar and Ad are the cross-sectional areas of the riser together with low denitrification occurred and less than
and down-comer, respectively, er and ed are the gas hold-up 35% T-N removals were observed in both systems. With
in the riser and down-comer, respectively, and KB is the fric- the increase of MLSS level, a concurrent improvement of
tional loss coefficient of the bottom zone of reactor. As the anoxic volumetric ratio f occurred. The reduction of
shown in Eq. (1), under the given fluid characteristics and DO level in the bulk solution was originated from the
geometry of the reactor, fluid recirculation rate increases increase of DO consumption, which was caused by the
mainly with the driving force (er  ed). If fewer bubbles were overgrowth of biomass. Hence, oxygen penetrability was
entrained into the down-comer in the MBRs, the gas holdup deteriorated and obvious DO gradient was developed
in the down-comer can be neglected (ed  0). inside the activated sludge flocs, which produced adequate
In addition, increasing the overflow liquid level can anoxic micro zone inside flocs for the growth of denitrifier.
enlarge the cross section surface of the recirculation fluid Although nitrification decreased slightly, denitrification
on the top of the baffle. This would consequently raise was significantly enhanced and an increase in T-N removal
DO levels in the down-comer through increasing fluid recir- was achieved ultimately.
culation ratio and lead to the reduction in the proportion Table 3 also shows a concurrent increase in both total
of anoxic zone. Therefore, a favorable range of fluid recir- nitrogen removal and the ratio of NO  
2 /NOx (NOx is the
culation rate can be easily maintained in the airlift MBR by  
sum of NO2 and NO3 ) in the filtrate when MLSS reached
manipulating aeration rate and overflow liquid level. the highest concentration range of 12.6–13.5 g/L. This
Under these optimal recirculation rates, not only oxygen implies that the reduction of total nitrogen might partially
supply, blending and membrane surface scouring can be stem from short-cut nitrification denitrification pathway
well accomplished in the riser, but also enough anoxic zone and sustained nitrite accumulation is generally proved to
can be obtained in the down-comer. Thus, integration of be a crucial step (Wyffels et al., 2003). Partial nitrification
nitrification and denitrification can be achieved via success- is obtained through disequilibrating the populations of
fully stratification of the reactor. ammonia-oxidizing bacteria (AOB) and nitrite-oxidizing
bacteria (NOB) by selectively inhibiting the growth of the
3.4. Effect of mixed liquor suspended solids (MLSS) latter (Kim and Seo, 2006; Wang and Yang, 2004). Advan-
concentrations tages of short-cut nitrification denitrification includes: (1)
25% reduction of the oxygen demand for nitrification (Peng
In order to evaluate the effect of MLSS on nitrogen et al., 2003; Bernet et al., 2001); (2) up to 40% reduction of
removal in AS-MBR (System I) and BPAC-MBR (System the carbon source for denitrification (Wyffels et al., 2003;
5872 Y.Z. Li et al. / Bioresource Technology 99 (2008) 5867–5872

Peng et al., 2003; Bernet et al., 2001; Garrido et al., 1997); Bernet, N., Peng, D.C., Delgenes, J.P., Moletta, R., 2001. Nitrification at
(3) lower biomass yield (Kim and Seo, 2006; Peng et al., low oxygen concentration in biofilm reactor. J. Env. Eng. 127 (3), 266–
271.
2003) and (4) 1.5–2 times higher denitrification rate as com- Bliss, P., Barnes, D., 1986. Modeling nitrification in plant scale activated
pared with that via nitrate pathway (Wang and Yang, sludge. Wat. Sci. Tech. 18 (6), 139–148.
2004; Peng et al., 2003; Garrido et al., 1997). It is proved Chang, Y.J., Tseng, S.K., 1999. A novel double-membrane system for
that nitrite accumulation can be achieved by maintaining simultaneous nitrification and denitrification in a single tank. Lett.
low DO levels since the oxygen saturation coefficients of Appl. Microbiol. 28 (6), 453–456.
Chisti, Y., Wenge, F., Murray, M.Y., 1995. Relationship between riser
Monod kinetics for nitritation and nitratation are known and downcomer gas hold up in internal loop airlift reactors without
to be 0.3 and 1.1 mg/L, respectively (Wang and Yang, gas–liquid separators. Chem. Eng. J. Biochem. Eng. J. 57 (1), 7–13.
2004). In other words, ammonium-oxidizing bacteria cope Collivignarelli, C., Bertanza, G., 1999. Simultaneous nitrification denitri-
with low DO level better than nitrite-oxidizing bacteria. fication processes in activated sludge plants: performance and appli-
With MLSS exceeding 12.6 g/L, over 38% anoxic volumet- cability. Wat. Sci. Tech. 40 (4–5), 187–194.
Garrido, J.M., Van Benthum, W.A.J., Van Loosdrecht, M.C.M., Heijnen,
ric ratio f was monitored in both reactors, which ultimately J.J., 1997. Influence of dissolved oxygen concentration on nitrite
led to NOB inhibition from oxygen limitation. accumulation in a biofilm airlift suspension reactor. Biotechnol.
However, since the MBRs adopted in our study were Bioeng. 53 (2), 168–178.
open and continuous system, it was not feasible to main- Guo, H.Y., Zhou, J.T., Su, J., Zhang, Z.Y., 2005. Integration of
tain a stable nitrogen removal via nitrite pathway com- nitrification and denitrification in airlift bioreactor. Biochem. Eng. J.
23, 57–62.
pletely. In fact, two nitrogen removal pathways should Helmer, C., Kunst, S., 1998. Simultaneous nitrification/denitrification in
present simultaneously in the MBRs, i.e. SND via both an aerobic biofilm system. Wat. Sci. Tech. 37 (4–5), 183–187.
conventional and short-cut nitrification and denitrification. Jun, B.H., Miyanaga, K., Tanji, Y., Unno, H., 2003. Removal of
nitrogenous and carbonaceous substances by a porous carrier-mem-
4. Conclusions brane hybrid process for wastewater treatment. Biochem. Eng. J. 14,
37–44.
Kim, D.J., Seo, D., 2006. Selective enrichment and granulation of
A novel process, internal-loop airlift submerged mem- ammonia oxidizers in a sequencing batch airlift reactor. Process
brane bioreactor, was developed and nitrogen removal per- Biochem. 41 (5), 1055–1062.
formances utilizing SND were investigated for the Kraume, M., Bracklow, U., Vocks, M., Drews, A., 2005. Nutrients
treatment of synthetic wastewater under various modes. removal in MBRs for municipal wastewater treatment. Wat. Sci. Tech.
51 (6–7), 391–402.
After the initial phase of nitrifier accumulation, almost Nah, Y.M., Ahn, K.H., Yeom, I.T., 2000. Nitrogen removal in household
complete ammonia removal was achieved in all the MBRs. wastewater treatment using an intermittently aerated membrane
Deficiency in anoxic condition led to insufficient denitrifi- bioreactor. Environ. Technol. 21 (1), 107–114.
cation and unsatisfactory total nitrogen removal. By appli- Peng, Y.Z., Chen, T., Tian, W.J., 2003. Nitrogen removal via nitrite at
cation of the reactor with ALR configuration, T-N removal normal temperature in A/O process. J. Environ. Sci. Health A 38 (6),
1007–1015.
was improved via formation of oxic/anoxic zones. Intro- Pochana, K., Keller, J., 1999. Study of factors affecting simultaneous
duction of biofilm in the bubbleless down-comer was nitrification and denitrification (SND). Wat. Sci. Tech. 39 (6), 61–68.
shown to further reduce T-N level by developing a stable Pochana, K., Keller, J., Lant, P., 1999. Model development for simulta-
anoxic microenvironment within the biofilm. The results neous nitrification and denitrification. Wat. Sci. Tech. 39 (1), 235–243.
of floc size distribution showed that the effect of microenvi- Priyali, S., Dental, K.S., 1998. Simultaneous nitrification–denitrification in
a fluidized bed reactor. Wat. Sci. Tech. 38 (1), 247–254.
ronment on SND was not significant for the MBRs with Silva, D.G.V., Urbain, V., Abeysinghe, D.H., Rittmann, B.E., 1998.
suspended biomass. DO stratification in the airlift reactors Advanced analysis of membrane-bioreactor performance with aerobic-
improved total nitrogen removal remarkably by developing anoxic cycling. Wat. Sci. Tech. 38 (4–5), 505–512.
considerable anoxic zone spatially. Increase of oxygen con- Tchobanoglous, G., Burton, F.L., Stensel, H.D., 2004. Wastewater
sumption under high MLSS level promoted DO gradient engineering treatment and reuse, fourth ed. Metcalf & Eddy Inc.,
McGraw-Hill, New York, pp. 750.
distribution within the activated sludge flocs and led to Turk, O., Mavinic, D.S., 1987. Benefits of using selective-inhibition to
higher total nitrogen removal through SND. Two total remove nitrogen from highly nitrogenous wastes. Environ. Technol.
nitrogen removal pathways via SND were proposed to Lett. 8 (9), 419–426.
explain the reduction of total nitrogen at MLSS concentra- Wang, J.L., Yang, N., 2004. Partial nitrification under limited dissolved
tion of 12.6–13.5 g/L. oxygen conditions. Process Biochem. 39 (10), 1223–1229.
Wyffels, S., Boeckx, P., Pynaert, K., Verstraete, W., Cleemput, O.V., 2003.
Sustained nitrite accumulation in a membrane-assisted bioreactor
References (MBR) for the treatment of ammonium-rich wastewater. J. Chem.
Technol. Biotechnol. 78 (4), 412–419.
APHA, 1998. Standard methods for the examination of water and Yusuf, C., Murray, M.-Y., 1993. Improve the performance of airlift
wastewater, 20th ed. Washington, DC. reactors. Chem. Eng. Prog. 89 (6), 38–45.