This article was downloaded by: [Selcuk Universitesi]

On: 10 February 2015, At: 22:28

Publisher: Taylor & Francis
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House,
37-41 Mortimer Street, London W1T 3JH, UK

Environmental Technology
Publication details, including instructions for authors and subscription information:
[Link]

Performance and influence factors of completely

autotrophic nitrogen removal over nitrite (CANON)
process in a biofilter packed with volcanic rocks
a a b a ab
Yuhai Liang , Dong Li , Xiaojing Zhang , Huiping Zeng & Jie Zhang
a
Key Laboratory of Water Quality Science and Water Environment Recovery Engineering,
Beijing University of Technology, Beijing 100124, People's Republic of China
b
State Key Laboratory of Urban Water Resource and Environment, Harbin Institute of
Technology, Harbin 150090, People's Republic of China
Accepted author version posted online: 25 Sep [Link] online: 17 Oct 2014.

Click for updates

To cite this article: Yuhai Liang, Dong Li, Xiaojing Zhang, Huiping Zeng & Jie Zhang (2015) Performance and influence
factors of completely autotrophic nitrogen removal over nitrite (CANON) process in a biofilter packed with volcanic rocks,
Environmental Technology, 36:8, 946-952, DOI: 10.1080/09593330.2014.969327

To link to this article: [Link]

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained
in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no
representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the
Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and
are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and
should be independently verified with primary sources of information. Taylor and Francis shall not be liable for
any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever
or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of
the Content.

This article may be used for research, teaching, and private study purposes. Any substantial or systematic
reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any
form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://
[Link]/page/terms-and-conditions
 Environmental Technology, 2015
Vol. 36, No. 8, 946–952, [Link]

Performance and influence factors of completely autotrophic nitrogen removal over nitrite
(CANON) process in a biofilter packed with volcanic rocks
Yuhai Lianga , Dong Lia∗ , Xiaojing Zhangb , Huiping Zenga and Jie Zhanga,b
a Key Laboratory of Water Quality Science and Water Environment Recovery Engineering, Beijing University of Technology, Beijing
100124, People’s Republic of China; b State Key Laboratory of Urban Water Resource and Environment, Harbin Institute of Technology,
Harbin 150090, People’s Republic of China
(Received 11 June 2014; final version received 21 September 2014 )

Completely autotrophic nitrogen removal over nitrite (CANON) process was considered as one of the most efficient and eco-
nomical nitrogen removal processes, which was suitable for treating wastewater with low ratio of carbon to nitrogen. In this
study, an enlarging start-up strategy for CANON process was proposed, and a 40-L CANON reactor was successfully started
by seeding 2-L mature biofilm containing both aerobic ammonia-oxidizing bacteria (AerAOB) and anaerobic ammonia-
Downloaded by [Selcuk Universitesi] at 22:28 10 February 2015

oxidizing bacteria (AnAOB). The effects of dissolved oxygen (DO), ammonia loading rate and the ratio of air inflow to
water inflow (Qair /Qwater ) on nitrogen removal performance were investigated. The distribution of AerAOB and AnAOB
was analysed using fluorescence in situ hybridization (FISH) technique. The system reached a maximum NRR of 3.11 kg N
m−3 d−1 with a removal efficiency of 89.5%, and the average value in steady state was 2.42 ± 0.26 and (83.07 ± 6.89)%,
respectively. Analysis of influence factors showed the important role of high DO (around 5 mg L−1 ), for the high-rate nitro-
gen removal, and the Qair /Qwater should be controlled at 28–40 for stable operation. FISH results suggested that AerAOB
and AnAOB predominated in the reactor, with proportions of 46.8% and 39.3%, respectively. This study demonstrated that
the biofilter operated with high effluent DO was a feasible setup for CANON process.
Keywords: CANON; start-up; nitrogen removal; DO; FISH

1. Introduction nitrogen removal and biogas production but also decrease

In conventional nitrification–denitrification nitrogen the power consumption.
removal processes, massive organic carbon was consumed CANON process has been successfully used to treat
in denitrification stage. Moreover, more and more organic industrial sewage like sludge digestion and landfill leachate
compounds are converted to biogas by the nearly devel- with high temperature or high ammonia.[1,8,9] However,
oped anaerobic treatment processes, which is feasible for there are still some challenges for the adoption of CANON
the energy utilization from the wastewater.[1] The bio- to application, such as long start-up period and low nitro-
gas conversion leads to the shortage of organic carbon for gen removal rate (NRR). Moreover, almost all the research
denitrification. Thus, autotrophic nitrogen removal process about CANON process were carried out in small-scale
could be a better choice. It could remove nitrogen without reactors. However, large-scale system is needed for the
organic carbon consumption and also less oxygen con- actual application. Thus, study about the rapid start-up of
sumption, such as partial nitrification and Anammox.[2– CANON reactor with large volume, as well as the real-
5] Completely autotrophic nitrogen removal over nitrite ization of high NRR, were both essential for the CANON
(CANON) process that combined partial nitrification and development.
Anammox in one single reactor has been developed in A high-rate CANON process depends on a coopera-
recent years.[6] In CANON process, partial ammonia is tive relationship between AerAOB and AnAOB, as well as
oxidized to nitrite by aerobic ammonia-oxidizing bacteria the effective inhibition on nitrite-oxidizing bacteria (NOB),
(AerAOB), then the remaining ammonia and the pro- which would result in an almost complete conversion of
duced nitrite are converted to dinitrogen gas by anaerobic ammonia to dinitrogen gas alone with small amounts of
ammonia-oxidizing bacteria (AnAOB) in oxygen-limited nitrate.[10] However, the three functional bacteria had dif-
condition.[7] Thus, CANON process could be a good ferent demands of dissolved oxygen (DO). To be specific,
choice for nitrogen removal from wastewater with low car- limited air supply would inhibit the activity of AerAOB,
bon. Moreover, the adoption of CANON process could and further constrain the improvement of NRR. In con-
not only eliminate the organic carbon competition between trast, supplying more oxygen than is necessary could either

*Corresponding author. Emails: lidong2006@[Link]; bjut_lidong@[Link]

This article was originally published with errors. This version has been corrected. Please see Erratum [Link]
09593330.2014.981104.
© 2014 Taylor & Francis
 Environmental Technology 947

cause nitrite accumulation or complete nitrification, both of Phase II (day 92–108): Activity inducement of AerAOB
which are harmful for the CANON process.[11,12] More- in oxygen-limited condition. In this phase, limited oxygen
over, the Anammox rate would sharply decrease with the was supplied to the system for start-up of CANON, and
increase of DO,[13] and the nitrite concentration in the nitrite concentration in influent was decreased.
reactor due to the more oxygen supply could totally inhibit Phase III (day 109–184): High-rate nitrogen removal by
the activity of AnAOB if it is over 100 mg L−1 .[14,15] In CANON process in aerobic condition. In this phase, only
addition, ammonia loading rate (ALR) could also impact ammonia nitrogen was supplied for the system, under high
the NRR; without cautious regulation of DO and associ- DO and ALR condition.
ated ALR adjustments, nitrogen removal may be impaired The main operational conditions in each phase are
due to large amount of unconverted ammonia or exces- summarized in Table 1.
sive accumulation of nitrite in the reactor. It is necessary
to adjust the oxygen supply levels in the reactor to an
optimal value associated with the variable ALR for the 2.2. Seed sludge and influent
maximum nitrogen removal. Consequently, when the reac- The enlarging start-up strategy was adopted for the start-
tor set-up and influent were constant, the ratio of air supply up of CANON in this study. A 2-L volcanic rock with
to water supply (Qair /Qwater ), which both integrated the mature biofilm was taken from a CANON reactor, the
DO and ALR, would be a key parameter for a high-rate biofilm contained microorganisms of both AerAOB and
Downloaded by [Selcuk Universitesi] at 22:28 10 February 2015

CANON. But unfortunately, there are currently no reports AnAOB. Then, the 2-L volcanic rock with biofilm was
of their success as investigating the effect of Qair /Qwater on mixed completely with another 38-L free volcanic rock
CANON performance. carrier, and the mixture was used as the seeding to start-
In this study, CANON was started-up in a volcanic up the experimental reactor (40-L), with an enlarging ratio
rock-biofilter. The feasibility of enlarging start-up strat- of 1:20.
egy was investigated, and a high-rate nitrogen removal The synthetic wastewater used in this study contained
was achieved. The effects of DO, ALR and Qair /Qwater on (NH4 )2 SO4 and NaHCO3 as main substrates, together with
nitrogen removal performance were also examined. The (in g L−1 ) KH2 PO4 (0.136), CaCl2 (0.136), MgSO4 ·7H2 O
spatial and quantity distribution of the functional microor- (0.3) and trace element solution of 1 mL L−1 .[16] No
ganisms was detected by fluorescence in situ hybridization organic carbon was fed into the reactor. The nitrogen com-
(FISH), to acquire more information for a better reactor ponents in the influent of different phases are summarized
performance. in Table 1.

2.3. Analytical methods

2. Materials and methods
The NH4 + , NO2 − and NO3 − concentrations of both the
2.1. Reactor and experimental setup influent and effluent were daily analysed by UV/VS spec-
An upflow biofilter with an effective volume of 40-L trophotometers (UV755b, Youke, China) according to
(H × D = 130 × 25 cm) was adopted in this study, which Standard Methods.[17] The temperature and pH of the
was made of Plexiglas. Both the wastewater and air were reactor were measured with online pH instruments (pH296,
fed into the bottom of the reactor consecutively, while the WTW, Germany). The DO was detected using online DO
effluent was outflowed from the top of the reactor via the instruments in effluent (Oxi296, WTW, Germany).
biofilm. The biofilm carrier used in this study was volcanic
rock, with the porosity and particle sizes of 80% and 4.0–
6.0 mm, respectively. The temperature during the experi- 2.4. FISH analysis
ment was controlled at 30 ± 1°C by the water bath. The pH The distribution and semi-quantification of AerAOB and
and DO were detected using the online instruments. AnAOB in biofilm samples were examined by FISH. The
The whole experiment was conducted in three phases. biofilm was obtained from the reactor on day 90 and
Phase I (day 1–91): Activity inducement of AnAOB in 180, respectively. FISH analysis was carried out accord-
anaerobic condition. The influent containing both ammonia ing to the method described in our previous study.[18,19]
and nitrite was fed into the reactor for Anammox reaction. Semi-quantitative analysis of AerAOB and AnAOB was

Table 1. Nitrogen forms concentration of influent substrate and operational conditions during the experiment.

Phase Time (d) NH4 + (mg L−1 ) NO2 − (mg L−1 ) NO3 − (mg L−1 ) Aeration (L min−1 ) HRT (h) T (°C) DO (mg L−1 )

I I-1 1–55 300 400 15 0 12.2 30 0

I-2 56–91 200 300 8 0 8.4 30 0
II 92–108 300 100 6 0.3–0.6 6.6 30 1.05–6.15
III 109–184 400 3 5 0.8–1.65 4.5 30 0.50–5.65
 948 Y. Liang et al.

conducted by calculating the ratio of the specific colour decrease tendency in the later phase I-1, which perhaps was
area to the total biofilm area. due to the bioactivity inhibition of AnAOB by ammonia or
nitrite.[21–23] Then, the influent ammonia and nitrite were
decreased to 200 and 300 mg L−1 , respectively, in phase
3. Results and discussion
I-2. As a result, the TN removal efficiency rose to 50%
3.1. Start-up of CANON process from day 59, indicating the feasibility of reducing nitrogen
In phase I-1 (day 1–55), the influent total nitrogen concentration in influent.
(TN) was about 700 mg·L−1 with a nitrite to ammonia In phase II (day 92–108), limited aeration was sup-
ratio of 1.3, to satisfy the theoretical reaction ratio of plied to the system for the inducement of the AerAOB.
Anammox.[20] The hydraulic retention time (HRT) was The aeration rate was 0.3–0.6 L min−1 , with the ammo-
set as 14 h for acclimation of the microorganisms and nia and nitrite concentration in influent of 300 and 100 mg
the biofilm formation. As a result, the TN removal effi- L−1 , respectively. Given that the growth rate of AerAOB
ciency reached 40% on day 9, indicating the favourable is faster than AnAOB, the HRT was reduced to 7 h in
activity of AnAOB when seeded to a new reactor. How- this phase. As shown in Figure 1, the ammonia removal
ever, nitrogen was not totally removed in the reactor and efficiency and TN removal efficiency rapidly achieved
lots of ammonia and nitrite nitrogen remained in effluent more than 99% and 70%, respectively, from day 97. This
(Figure 1). Moreover, some sludge were observed in efflu- result proved the existence of AerAOB throughout the
Downloaded by [Selcuk Universitesi] at 22:28 10 February 2015

ent, which also indicated the excess inflow. Thus, the HRT experiment, even in the anaerobic condition, which was
was increased to 23 h from day 10 to 15 to avoid the sludge consistent with the FISH result. The results in this phase
loss, to facilitate the biofilm formation. As a result, TN suggested the preferability of seeding CANON biofilm
removal efficiency rose to more than 70%, and no sludge for start-up, which could significantly shorten the start-up
was observed in the effluent. However, it was not profitable period. Since AerAOB and AnAOB both existed in the
for the material exchange between the bacteria and influ- CANON sludge, the start-up period could be remarkably
ent with long HRT, which would lead to the shortage of shortened.
substrate at the top of the reactor. Hence, the HRT was Finally, in phase III (day 109 to 184), more oxygen
decreased to 10 h from day 16, which led to the sudden was supplied to the system with the aeration rate increas-
increase of both ammonia and nitrite in effluent (Figure 1). ing to 0.8 L·min−1 , and the ammonia nitrogen was the
Subsequently, the ammonia and nitrite in the effluent were only substrate in influent. As shown in Figure 1, after 2
both gradually decreased with the operation, together with days’ acclimation, the ammonia removal efficiency and
the increase of nitrate produced by AnAOB. However, as TN removal efficiency reached 98.6% and 81.4%, respec-
shown in Figure 1(d), TN removal efficiency showed a tively. NRR achieved to 1.10 kg N m−3 d−1 , indicating

influent effluent removal efficiency

(a) (b) 600
I II III I II III
500 100
ammonia concentration (mg·L–1)

ammonia removal efficiency(%)

nitrite concentration (mg·L–1)

90
400
80 400

300 70
60
200 50 200
40
100
30

0 20 0
0 20 40 60 80 100 120 140 160 180 0 20 40 60 80 100 120 140 160 180
Time (d) Time (d)
(c) (d) 900 100
70
800
nitratea concentration (mg·L–1)

60 80
TN concentration (mg·L–1)

700
TN removal efficiency(%)

50 600
60
40 500
400
30 40
300
20
200 20
10 100
0 0 0
0 20 40 60 80 100 120 140 160 180 0 20 40 60 80 100 120 140 160 180
Time (d) Time (d)

Figure 1. Variations of the nitrogen component during the experiment.

 Environmental Technology 949

Table 2. Some CANON reactors in previous reports.

Reactor Volume (L) Seed sludge Start-up period (d) NRR (kg m−3 d−1 ) Reference

Gas-lift 1.8 Anammox 107 1.5 [6]

SBR 1.5 Nitrification/anammox 295 0.36 [9]
SBBR 6.5 Activated sludge 118 0.072 [15]
biofilter 1.25 Anammox 270 0.35 ± 0.19 [25]
biofilter 5.0 Anammox/conventional 190 0.92 [8]
activated sludge
SBBR 18 CANON sludge 164 0.791 [26]
Biofilter 40 CANON biofilm 114 2.42 ± 0.26 This study

the successful start-up of CANON. Moreover, no nitrite finally reached a maximum of 3.11 kg N m−3 d−1 with the
accumulation was observed in the effluent, indicating the efficiency of 89.5% (Figure 2), both of which were higher
co-operative work of AerAOB and AnAOB. Simultane- than other CANON systems in previous reports (Table 2).
ously, the biofilm in the reactor turned to be red, which For more information about the reactor performance,
was the signature colour of AnAOB aggregation.[16,24] liner fitting between NRR and ARR was adopted, respec-
Finally, the maximum ammonia removal efficiency, maxi- tively, from day 114 to 140 and day 141 to 184 (Figure 3).
Downloaded by [Selcuk Universitesi] at 22:28 10 February 2015

mum TN removal efficiency and NRR reached 99%, 89.5% Two liner equations were obtained, as shown in Equation
and 3.11 kg N m−3 d−1 , respectively. And those mean val- (2) (day 114 to 140) and Equation (3) (day 141 to 184). The
ues after day 114 were (95.00 ± 6.81)%, (83.07 ± 6.89)% two liner fitting equations both showed perfect correlation,
and 2.42 ± 0.26 kg N m−3 d−1 , respectively. with the R2 of 0.998 and 0.999, respectively, indicating the
The nitrogen removal performance observed in this good representativeness of the data
study combined both high NRR and excellent nitrogen
removal efficiency, while the NRR in other CANON sys- 1NH3 + 0.85O2 → 0.11NO−
3 + 0.44N2
tems was much lower (Table 2). Results suggested that + 0.14H+ + 1.43H2 O (1)
enriching AnAOB first and then inducing the activity of ∗
NRR(day 114−140) = 0.846 ARR, R = 0.998
2
(2)
AerAOB with limited oxygen supply was feasible to real-
ize CANON process. The start-up period was remarkably NRR(day141−184) = 0.913∗ ARR, R2 = 0.999 (3)
shortened using this method when compared with other
systems (Table 2). In addition, the reactor used in this study According to the reaction formula of CANON pro-
was 40-L, while the reactors used in previous reports were cess (Equation (1)), the theoretical ratio of TN removal to
almost in the range of 1–10 L (Table 2). Results in this ammonia removal was 0.89. However, the ratio in Equation
phase indicated the effectiveness of the enlarging start-up (2) was 0.846, which was lower than the theoretical ratio.
strategy for the expansion of CANON reactor, as well as This result indicated that the TN removal was lower than
the rapid start-up and high-rate nitrogen removal. Thus, the maximum rate in CANON process, which perhaps was
in actual application, a small reactor could be started-up due to the nitrite accumulation or excess nitrate produc-
first, then the mature biofilm could be used as seeding for tion by NOB. Figure 1(b) shows that a small amount of
a larger reactor. Using this method, a CANON plant could nitrite remained in effluent, proving the speculation above.
be gradually implemented. This suggested that oxygen supply in this period was more
than demand and thus led to the excess production of nitrite
by AerAOB and less consumption by AnAOB. During day
3.2. High-rate nitrogen removal performance 141 to 184, the ratio of NRR to ARR was 0.913, a little
higher than 0.89. This ratio indicated that part of nitrogen
The nitrogen removal performance is shown in Figure 2.
was removed in other way except for CANON pathway.
In phase I, no oxygen was supplied for biofilm formation
Other nitrogen removal existed in the system such as
and recovery of the AnAOB activity. Both ammonia and
endogenous denitrification, through which, the denitrifying
nitrite were removed efficiently, NRR achieved to 0.56 kg
bacteria removed the nitrogen using the dead microor-
N m−3 d−1 . Then in phase II, together with the increase of
ganism as the organic carbon. Thus, both the CANON
ammonia and decrease of nitrite in influent, limited oxygen
and endogenous denitrification contributed to the high-rate
was supplied to the system for the inducement of AerAOB.
nitrogen removal, while denitrification only amounted a
NRR achieved to 0.91 kg N m−3 d−1 , which suggested the
little in the whole.
simultaneous bioactivity of AerAOB and AnAOB. Thus,
only ammonia with more oxygen was supplied as the sub-
strate for the system in phase III. As shown in Figure 3.3. Effect of Qair /Qwater
2, ammonia removal rate (ARR) and NRR both rose up, CANON process requires the simultaneous bioactivity of
correspondingly with the increase of aeration supply. NRR AerAOB and AnAOB, as well as effective inhibition on
 950 Y. Liang et al.

4 26
I II III
24
ammonia removal rate TN removal rate
22

ammonia removal rate, TN removal rate (kg·m–3·d–1)

HRT
20
3
18

16

HRT(h)
14
2
12

10

8
1
6
Downloaded by [Selcuk Universitesi] at 22:28 10 February 2015

0 0
0 20 40 60 80 100 120 140 160 180
Time (d)

Figure 2. Nitrogen removal performance of the reactor during the experiment.

3.5
NRR
AnAOB by the high DO. Moreover, NRR showed wholly
Liner fit of NRR with ARR (day 114 to 140) increase trend with the rose of ALR (Figure 4(b)). How-
Liner fit of NRR with ARR (day 141 to 184) ever, sufficient DO supply was essential for a high removal
3.0
efficiency when ALR was high. In addition, the DO in
the biofilter with volcanic rock as carrier possibly could
2.5
not exactly indicate the oxygen in the inner biofilm. Thus,
NRR (kg·m–3·d–1)

y2 = 0.913x the Qair /Qwater ratio that combines both the oxygen and the
2.0 R2 = 0.999 influent load was adopted to this study, as a control parame-
y1 = 0.846x ter for high-rate CANON process. As shown in Figure 4(c),
R2 = 0.998 with the increase of Qair /Qwater , NRR rose up initially and
1.5
then declined. When Qair /Qwater was increased to 30.4,
NRR achieved to the maximum. When the Qair /Qwater was
1.0 kept in the range of 28–40, NRR showed small fluctua-
tion with the variation of Qair /Qwater , which could be kept
1.0 1.5 2.0 2.5 3.0 3.5 around 2.5 kg m−3 d−1 . Thus, to achieve the simultaneous
ARR (kg·m–3·d–1) bioactivity of both AerAOB and AnAOB, the Qair /Qwater
Figure 3. Liner fitting result of NRR with ARR during the stable should be controlled at 28–40, since lower ratio would
operation period. inhibit the bioactivity of AerAOB and higher one would
limit that of AnAOB. Certainly, the figuration of the reac-
tor and influent ammonia concentration would significantly
NOB. However, the three functional bacteria presented dif- affect the best Qair /Qwater ; however, the Qair /Qwater could be
ferent demands for oxygen. The bulk DO concentration a feasible parameter for stable CANON, since the setup of
in the reactor reached 5.9 mg L−1 in phase III, which was the system and the influent would both be constant in actual
much higher than the DO levels reported in other CANON operation.
systems.[25,27,28] As shown in Figure 4(a), NRR in In addition, the volcanic rock carrier used in this study
this study rose up correspondingly with the increase of may also be contributed to the efficient NRR. The vol-
DO. Thus, the high-rate nitrogen removal in this sys- canic rock had lots of pore, in which AnAOB could be
tem was possibly attributed to the higher DO. Whereas, survived. On the other hand, both the high ALR and the
when DO was increased to more than 5 mg·L−1 , NRR active microorganisms enabled the consumption of DO by
presented a decline tendency, indicating the inhibition on AerAOB, which protected the AnAOB from the oxygen
 Environmental Technology 951
Downloaded by [Selcuk Universitesi] at 22:28 10 February 2015

Figure 4. Effect of Qair /Qwater on the nitrogen removal performance.

inhibition. Thus, the biofilter installed with volcanic rock from the densely packed sludge floc in phase I, the targeted
was an effective experimental setup for CANON pro- microorganisms became scattered with the cell distances
cess. However, the exact mechanism of DO consumption increased, AerAOB and AnAOB trended to be separate
and nutrient transfer inside the biofilm still need further clusters in the FISH micrograph of phase III. Meanwhile,
investigation. the organisms hybridized with specific probes became
to be larger and brighter ones in phase III. One puta-
tive explanation for this phenomenon was proposed as
3.4. FISH results
the lower ammonia remained in the reactor in phase III
The distribution characterization of functional microorgan- might lead to the increased cell distances.[30] Besides,
isms on days 90 and 180 was detected by FISH (Supple- the nitrogen gas produced in the biofilter by AnAOB per-
mentary material). AerAOB (orange) and AnAOB (blue) haps also contributed to the incompact biofilm, which
both existed in the two biofilm samples, which was con- would be dispersed easily when suffering the pretreatment
sistent with the speculation in Section 3.1. The proportion of ultrasonication and then led to the scattered distribu-
of AerAOB and AnAOB were detected by calculating the tion of the targeted microorganisms. The high-rate nitrogen
ratio of orange colour area and blue colour area, respec- removal performance maybe was attributed to the larger
tively, to the total biofilm area. The proportion of the and brighter microorganisms in phase III, which indicated
two kinds of bacteria showed direct relationship with the the favourable bioactivity of AerAOB and AnAOB.
performance of the reactor. On day 90, the reactor was
conducted as Anammox process in anaerobic condition, 4. Conclusions
the abundance of AerAOB and AnAOB was 22.3% and
The enlarging start-up strategy was feasible for rapid start-
67.5%, respectively, and the proportions changed to 46.8%
up and high-rate nitrogen removal of CANON process.
and 39.3% on day 180 in phase III when the system showed
The start-up period was remarkably shortened (114 d);
a high-rate performance of CANON process. Surprisingly,
both high NRR and excellent nitrogen removal efficiency
AerAOB existed in the biofilm when the reactor was oper-
were achieved. Both the high DO and ALR contributed
ated as Anammox process. There was one explanation for
to the high-rate nitrogen removal. The Qair /Qwater should
this phenomenon that the observed bacteria contained both be controlled at 28–40, to achieve a stable and effective
active and non-active cells,[10] which were difficult to dis- CANON operation. FISH results proved the predominance
criminate by FISH. The AerAOB existed in phase I was of AerAOB and AnAOB, both of which performed a har-
inhibited since no oxygen was supplied to the system, so monious work for the autotrophic nitrogen removal in the
AerAOB could not perform the bioactivity and existed as system.
non-active organisms on day 90.
Some previous reports suggested that AerAOB stay Acknowledgements
in the outerlayer of AnAOB in the biofilm of CANON
This work was supported by Trans-Century Training Programme
process. However, similar to the previous report,[29] the Foundation for the Talents by the State Education Commis-
microorganisms in the present study were distributed side sion (NCET-10-0008) and water project of National Science and
by side with each other in the FISH micrographs. Different Technology Major Project (Grant No. 2012ZX07202-005).
 952 Y. Liang et al.

Supplemental data sequencing batch biofilm reactor (SBBR). Bioresour Tech-

Supplemental data for this article can be accessed at nol. 2010;101:6309–6322.
10.1080/09593330.2014.969327 [16] van de Graaf AA, de Bruijn P, Robertson LA, Jetten MSM,
Kuenen JG. Autotrophic growth of anaerobic ammonium-
oxidizing micro-organisms in a fluidized bed reactor. Micro-
biology 1996;142:2187–2196.
References [17] APHA. Standard methods for the examination of water and
[1] Kartal B, Kuenen JG, van Loosdrecht MCM. Sewage treat- wastewater. 19th ed. Washington, DC: American Public
ment with anammox. Science. 2010;328:702–703. Health Association; 1995.
[2] An P, Xu X, Yang F, Liu L, Liu S. A pilot-scale study on [18] Zhang X, Li D, Liang Y, He Y, Zhang Y, Zhang J.
nitrogen removal from dry-spun acrylic fiber wastewater Autotrophic nitrogen removal from domestic sewage in
using anammox process. Chem Eng J. 2013;222:32–40. MBR-CANON system and the biodiversity of functional
[3] Xu G, Xu X, Yang F, Liu S, Gao Y. Partial nitrification microbes. Bioresour Technol. 2013;150:113–120.
adjusted by hydroxylamine in aerobic granules under high [19] Zhang X, Li D, Liang Y, Zhang Y, Fan D, Zhang J. Appli-
DO and ambient temperature and subsequent anammox for cation of membrane bioreactor for completely autotrophic
low C/N wastewater treatment. Chem Eng J. 2012;213:338– nitrogen removal over nitrite (CANON) process. Chemo-
345. sphere. 2013;93:2832–2839.
[4] Zhang L, Yang J, Furukawa K. Stable and high-rate nitro- [20] Strous M, Heijnen JJ, Kuenen JG, Jetten MSM. The
gen removal from reject water by partial nitrification and sequencing batch reactor as a powerful tool for the study
subsequent anammox. J Biosci Bioeng. 2010;110:441–448. of slowly growing anaerobic ammonium-oxidizing microor-
Downloaded by [Selcuk Universitesi] at 22:28 10 February 2015

[5] Zhang L, Yang J, Hira D, Fujii T, Zhang W, Furukawa ganisms. Appl Microbiol Biotechnol. 1998;50:589–596.
K. High-rate nitrogen removal from anaerobic digester [21] Chen T, Zheng P, Shen L, Ding S, Mahmood Q. Kinetic
liquor using an up-flow anammox reactor with polyethy- characteristics and microbial community of Anammox-
lene sponge as a biomass carrier. J Biosci Bioeng. 2011; EGSB reactor. J Hazard Mater. 2011;190:28–35.
111:306–311. [22] Fernandez I, Dosta J, Fajardo C, Campos JL, Mosquera-
[6] Sliekers AO, Third KA, Abma W, Kuenen JG, Jetten MSM. Corral A, Mendez R. Short- and long-term effects of
CANON and anammox in a gas-lift reactor. FEMS Micro- ammonium and nitrite on the anammox process. J Environ
biol Lett. 2003;218:339–344. Manage. 2012;95S:170–173.
[7] Third KA, Sliekers AO, Kuenen JG, Jetten MSM. The [23] Tang CJ, Zheng P, Hu BL, Chen JW, Wang CH. Influence
CANON system (completely autotrophic nitrogen-removal of substrates on nitrogen removal performance and microbi-
over nitrite) under ammonium limitation: interaction and ology of anaerobic ammonium oxidation by operating two
competition between three groups of bacteria. Syst Appl UASB reactors fed with different substrate levels. J Hazard
Microbiol. 2001;24:588–596. Mater. 2010;181:19–26.
[8] Qiao S, Nishiyama T, Fujii T, Bhatti Z, Furukawa K. [24] Jetten MSM, Strous M, van de Pas-Schoonen KT, Schalk J,
Rapid startup and high rate nitrogen removal from anaerobic van Dongen U, van de Graaf AA, Logemann S, Muyzer G,
sludge digester liquor using a SNAP process. Biodegrada- van Loosdrecht MCM, Kuenen JG. The anaerobic oxidation
tion. 2012;23:157–164. of ammonium. FEMS Microbiol Rev. 1998;22:421–437.
[9] Vazquez-Padin JR, Pozo MJ, Jarpa M, Figueroa M, Franco [25] Egli K, Fanger U, Alvarez PJJ, Siegrist H, van der Meer
A, Mosquera-Corral A, Campos JL, Mendez R. Treatment JR, Zehnder AJB. Enrichment and characterization of an
of anaerobic sludge digester effluents by the CANON pro- anammox bacterium from a rotating biological contactor
cess in an air pulsing SBR. J Hazard Mater. 2009;166:336– treating ammonium-rich leachate. Arch Microbiol. 2001;
341. 175:198–207.
[10] Strous M, Fuerst JA, Kramer EHM, Logemann S, Muyzer [26] Daverey A, Su SH, Huang YT, Chen SS, Sung S, Lin
G, van de Pas-Schoonen KT, Webb R, Kuenen JG, Jetten JG. Partial nitrification and anammox process: a method
MSM. Missing lithotroph identified as new planctomycete. for high strength optoelectronic industrial wastewater treat-
Nature. 1999;400:446–449. ment. Water Res. 2013;47:2929–2937.
[11] Helmer C, Tromm C, Hippen A, Rosenwinkel KH, Seyfried [27] Strous M, Gerven EV, Kuenen JG, Jetten M. Effects of aer-
CF, Kunst S. Single stage biological nitrogen removal by obic and microaerobic conditions on anaerobic ammonium-
nitritation and anaerobic ammonium oxidation in biofilm oxidizing (anammox) sludge. Appl Environ Microbiol.
systems. Water Sci Technol. 2001;43:311–320. 1997;63:2446–2448.
[12] Lv Y, Wang L, Sun T, Wang X, Yang Y, Wang Z. [28] Helmer-Madhok C, Schmid M, Filipov E, Gaul T, Hippen
Autotrophic nitrogen removal discovered in suspended A, Rosenwinkel K-H, Seyfried CF, Wagner M, Kunst S.
nitritation system. Chemosphere. 2010;79:180–184. Deammonification in biofilm systems: population structure
[13] Zhang D, Cai Q, Cong L. Enhancing completely autotrophic and function. Water Sci Technol. 2002;46:223–231.
nitrogen removal over nitrite by trace NO2 addition to an [29] Pynaert K, Smets BF, Wyffels S, Beheydt D, Siciliano
AUSB reactor. J Chem Technol Biotechnol. 2010;85:204– SD, Verstraete W. Characterization of an autotrophic
208. nitrogen-removing biofilm from a highly loaded lab-scale
[14] Cho S, Fujii N, Lee T, Okabe S. Development of a simul- rotating biological contactor. Appl Environ Microbiol.
taneous partial nitrification and anaerobic ammonia oxida- 2003;69:3626–3635.
tion process in a single reactor. Bioresour Technol. 2011; [30] Okabe S, Kindaichi T, Ito T, Satoh H. Analysis of size
102:652–660. distribution and areal cell density of ammonia-oxidizing
[15] Zhang Z, Chen S, Wu P, Lin L, Luo H. Start-up of the bacterial microcolonies in relation to substrate microprofiles
Canon process from activated sludge under salt stress in a in biofilms. Biotechnol Bioeng. 2004;85:86–95.