Bioresource Technology 143 (2013) 178–186

Contents lists available at SciVerse ScienceDirect

Bioresource Technology
journal homepage: www.elsevier.com/locate/biortech

Denitrification of simulated municipal wastewater treatment plant

effluent using a three-dimensional biofilm-electrode reactor: Operating
performance and bacterial community
Ruixia Hao a,⇑, Sumei Li a, Jianbing Li b,⇑, Chengcheng Meng a
a
Key Laboratory of Water Quality Science and Water Environment Recovery Engineering, Beijing University of Technology, Beijing 100124, China
b
Environmental Engineering Program, University of Northern British Columbia, Prince George, British Columbia V2N 4Z9, Canada

h i g h l i g h t s

 A 3-D biofilm-electrode reactor was used for wastewater denitrification.

 A nitrate removal of 98.3% was obtained with C/N ratio of 3.0 and HRT of 7 h.
 Both heterotrophic and autotrophic bacteria were responsible for nitrate removal.
 A phylogenetic tree of gene sequences in biofilm was established.
 The biofilm was abundant with Thauera-like and Enterobacter-like bacteria.

a r t i c l e i n f o a b s t r a c t

Article history: A three-dimensional biofilm-electrode reactor (3D-BER) was applied for nitrate removal from simulated
Received 12 April 2013 municipal wastewater treatment plant (WWTP) effluent. It was found that when the influent C/N ratio
Received in revised form 30 May 2013 ranged from 1.0 to 2.0, both heterotrophic and autotrophic denitrifying microorganisms played impor-
Accepted 1 June 2013
tant roles in nitrate removal. The extension of hydraulic retention time (HRT) could enhance nitrate
Available online 6 June 2013
removal, but too long HRT was not necessary. A phylogenetic tree of gene sequences in biofilm was estab-
lished, and the biofilm was abundant with Thauera-like and Enterobacter-like bacteria. The results illus-
Keywords:
trated that 3D-BER is a feasible and effective technology for the denitrification of WWTP effluent with
16S rDNA
Bacterial community
poor organic carbon source. A nitrate removal of 98.3% was obtained with C/N ratio of 3.0 and HRT of
Biofilm-electrode reactor 7 h. About 85.0–90.0% of nitrate removal was found at a C/N ratio of 1.5 and HRT of 10 h due to cooper-
Carbon source ative heterotrophic and autotrophic denitrification.
Wastewater denitrification Ó 2013 Elsevier Ltd. All rights reserved.

1. Introduction
The shortage of water resources has become an exacerbating is-
sue around the world as a result of accelerated industrialization
and urban growth as well as changed climatic conditions. Among
various measures of addressing this challenging problem, waste-
water reclamation has received tremendous attention (Yi et al.,
2011). The reclaimed water from municipal wastewater treatment
plant (WWTP) can serve for many purposes, such as urban scenic
and landscape greening, urban lake and river water replenishment,
car washing, construction site water use, urban road sprinkling,
industrial circulating cooling water use, irrigation water use, and
⇑ Corresponding authors. Address: College of Architecture and Civil Engineering,
Beijing University of Technology, Beijing 100124, China. Tel.: +86 10
67391726x8005; fax: +86 10 67391648 (R. Hao), tel.: +1 250 9606397; fax: +1
250 9605845. (J. Li).
E-mail addresses: haoruixia@bjut.edu.cn (R. Hao), Jianbing.Li@unbc.ca (J. Li).
the recharge of groundwater (Hao et al., 2013). As a result, the re-
claimed water represents an indispensable part of water resources.
For example, the utilization of reclaimed water in Beijing in North-
ern China (i.e. a region with serious water shortage problem)
reached 0.65 and 0.68 billion m3 in 2009 and 2011, accounting
for about 18% and 20% of its total annual water consumption vol-
ume, respectively (Wei et al., 2011). In general, the quality of re-
claimed water is affected by the available treatment processes
whereas many pollutants could not be thoroughly removed. In par-
ticular, WWTP effluent may still contain a relatively high concen-
tration of inorganic nitrogen such as nitrate and nitrite. For
example, a total nitrogen (TN) concentration (mostly nitrate) of
15–40 mg/L in effluent was observed in many WWTPs in Beijing
(Li et al., 2011). The discharge of nitrogen components into the
environment through wastewater reclamation can lead to a num-
ber of public health and environmental concerns. These include
the eutrophication of lakes and rivers, excessive vegetative growth
and reduced fruit set of crops, decreased groundwater quality, and

0960-8524/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.biortech.2013.06.001
 R. Hao et al. / Bioresource Technology 143 (2013) 178–186
other hazards (e.g., disease of methemoglobinemia in the fetus) for
human and animal health (Ghafari et al., 2008). As a result, the
effective denitrification of WWTP effluent is of fundamental
importance.
Biological denitrification has received the greatest interests
among various processes since the bacteria can decompose nitrate
into harmless nitrogen gas with the presence of an electron donor
(Karanasios et al., 2010). However, WWTP effluent usually contains
low concentration of chemical oxygen demand (COD) with little
organic carbon source or internal electron donor left for bacteria
to utilize, and it is thus necessary to use an external source as elec-
tron donor for denitrification. This problem can be addressed by
using a biofilm-electrode reactor (BER) which produces hydrogen
gas (H2) through water electrolysis, while the denitrifying microor-
ganisms immobilized on the cathode surface directly contact the
hydrogen gas and use it as an electron donor (Zhao et al., 2012).
If the anode of BER is made of graphite, CO2 can also be generated
to serve as an inorganic carbon source for microorganisms and a
pH buffer in the denitrification system (Mousavi et al., 2012). Many
studies of using BERs have been reported. For example, Park et al.
(2005) applied a BER with a dimensionally stable anode (DSA) and
a graphite felt cathode to treat an aqueous solution containing high
concentrations of nitrate, and they observed a maximum nitrate
removal efficiency of 98% at an applied current of 200 mA, corre-
sponding to a nitrate removal of 0.17 mg NO3–N/(cm2 of biofilm
surface area day). Li et al. (2010) used Fe as a cathode and Ti/
IrO2–Pt as an anode in an undivided cell to treat nitrate contami-
nated wastewater, and they found that the nitrate concentration
decreased from 100.0 to 7.2 mg/L within 180 min with the pres-
ence of NaCl at 0.50 g/L, without the formation of ammonia and ni-
trite by-products. Feng et al. (2013) and Huang et al. (2013) used
rectangular graphite electrodes in a cylindroid reactor to investi-
gate the effects of different carbon sources and C/N ratio on the
denitrification efficiency, and they found that starch could be uti-
lized by microbes to generate electrons, while the increase of C/N
ratio from 2.0 to 3.5 resulted in an increase of nitrate removal from
0.69 ± 0.02 to 1.09 ± 0.16 mg/(L h).
In general, the BER denitrification process can be affected by
many operating parameters, such as temperature, nitrate concen-
tration, carbon source, salinity, pH, hydraulic retention time
(HRT), electrode materials, reactor configuration, and current
intensity (Zhao et al., 2011). In conventional BER (with two elec-
trodes), the cathode not only provides hydrogen for autotrophic
denitrifying bacteria as their electron donor, but also is used as
the carrier for microorganisms. As a result, its nitrate removal
capacity can be restricted by the surface area of cathode. Moreover,
the denitrification can be affected by limited inorganic carbon
source in BER (Zhao et al., 2011). Consequently, the BER is usually
associated with a relatively low denitrification rate and requires a
long hydraulic retention time. To address this problem, a number
of efficient and economical alternatives of BER configuration have
been proposed. These include the combined electrochemical reac-
tor and biological reactor, and multi-electrode reactor or three-
dimensional biofilm-electrode reactor (3D-BER) (Mousavi et al.,
2012). In 3D-BER, activated carbon (AC) or AC mixed with another
substance is filled in the cathode part as a third bipolar electrode.
This third electrode (i.e. AC filler) provides a high surface area for
biomass growth and attachment as well as increases hydrogen
gas yield. It also produces more CO2 to provide a desirable anoxic
condition where less organic carbon source is required to remove
nitrate (Zhou et al., 2007). For example, Zhou et al. (2009) used
PbO2 as an anode and activated carbon fiber (ACF) as a cathode
to treat nitrate-polluted groundwater, while the anode and cath-
ode were set in the center and surrounding the inner wall of the
reactor, respectively. The conventional BER (or 2D-BER) was con-
verted to a 3D-BER when activated carbon (AC) was added and
179
packed in the reactor, and they found that the 3D-BER’s nitrate re-
moval was increased from 57.93% to 78.10% as compared to 2D-
BER at a HRT of 4 h, while the nitrate removal was further in-
creased to 99.80% at a HRT of 8 h (Zhou et al., 2009). The 3D-BER
was also applied to oxidize pollutants from industrial solid waste
landfill leachate (Rao et al., 2009), remove acid orange 7 in simu-
lated wastewater (Zhao et al., 2010), and remove NOx from flue
gas (Zhou et al., 2012).
The novel 3D-BER technology is still in its early development
stage with only a few denitrification studies being reported in lit-
eratures, and even very few studies were found for nitrate re-
moval from WWTP effluent. In addition, the denitrification
performance of BER can be greatly affected by the involved
microorganisms. However, previous studies on both conventional
BER and 3D-BER technologies have mainly focused on reactor de-
sign (e.g., electrode materials and reactor configurations) and the
effects of various operating factors. The information of involved
denitrifying bacterial community as well as the interaction be-
tween bacteria community and reactor operating conditions are
still unclear (Mousavi et al., 2012; Feng et al., 2013). In fact, the
understanding of microbial structure and diversity during denitri-
fication process is of great importance for optimizing BER perfor-
mance. The development of new molecular biology techniques
such as the 16S rDNA-based methods and other methods like
scanning electron microscopy (SEM) can make it possible to char-
acterize and evaluate the bacterial community in BER (Park et al.,
2006; Cho et al., 2008). Such techniques have been applied to
characterize the microbial communities in many biological treat-
ment studies of wastewater (Osaka et al., 2008; Srinandan et al.,
2012). As described above, the 3D-BER can be a promising choice
for achieving a higher denitrification rate and a shorter HRT due
to the reactor’s unique configurations. It is thus of importance
to extend its application to WWTP effluent treatment. The objec-
tive of this study was then to investigate the performance of a
3D-BER for removing nitrate from simulated WWTP effluent un-
der various C/N ratio and HRT conditions. The graphite cathode
and anode were used, and the activated carbon (AC) was filled
in the reactor to form a third electrode. The bacterial community
in the biofilm developed on AC filler was characterized through
the analysis of 16S rDNA. Such results can provide valuable infor-
mation for designing more cost-effective denitrifying process for
wastewater reclamation.
2. Methods
2.1. Experimental apparatus and simulated wastewater
Fig. 1 shows the schematic diagram of experimental setup and
the 3D-BER configuration. The reactor made of plexiglass was
cylindrical with a diameter of 15 cm and a height of 56 cm. The
cathode consisting of six graphite rods (diameter of 2 cm and
height of 42 cm) was set surrounding the inner wall of the reactor,
while the graphite rods were linked using insulated electrical cop-
per wires. The anode was a graphite rod (diameter of 3 cm and
height of 42 cm) and was set in the center of the reactor. The
remaining space in the reactor was filled with granular activated
carbon (AC) with particle size of 3–5 mm and a total volume of
6 L. The reactor had an effective volume of 3.4 L. Prior to use, the
AC was washed with H2SO4 solution (0.02 M) and deionized water
for several times and was then dried at 105 °C for 24 h (Zhou et al.,
2009). The AC filler can serve not only as a biofilm carrier but also
as the third electrode. A DC regulated power supply (model HSPY-
60-2 provided by Beijing Hansheng Puyuan Technology Co., Ltd. in
China, 0–60 V, 0–2 A) was used to provide constant current for
water electrolysis.
180 R. Hao et al. / Bioresource Technology 143 (2013) 178–186
4 CODCr concentration of 334–373 mg/L. Such nutrient-containing
2 1 2 60-L influent tank, and its pH was adjusted once everyday to about
3 isms adapt to the environment, electric current in the 3D-BER
6 tored. After this phase, more than 90% of nitrate was steadily re-
5 After biofilm immobilization and acclimatization, the 3D-BER
Fig. 1. Schematic diagram of 3D-BER apparatus (1) graphite rod anode (1); (2)
graphite rod cathode (6); (3) activated carbon particles; (4) DC regulated power
supply; (5) peristaltic pump; 6 influent tank.
Other reactor components consisted of an influent plastic tank
(60 L) and a peristaltic pump. The influent water in this study
was a simulated WWTP effluent prepared by tap water amended
with a stock solution of KNO3, CH3COONa, and phosphorus. The
nitrogen was provided by KNO3, with a NO3–N concentration of
30 mg/L being maintained. The phosphorus was provided by KH2-
PO4, with a PO43–P concentration of 0.3 mg/L being obtained. The
CH3COONa was added to provide carbon source according to de-
sired C/N ratios, and no microelements were added since they were
available in tap water. About 20–30 L of synthetic wastewater was
prepared every 2 days, and its pH was adjusted to about 7.0 using
HCl and NaOH every morning on each day. Such simulated WWTP
effluent was used for the normal operation of 3D-BER after its bio-
film development. Using a peristaltic pump, wastewater was
pumped from the influent tank into the 3D-BER through its bottom
inlet, and flew through the biofilm on AC filler, and then flew out
from the upper of the reactor.
2.2. Biofilm development and experimental procedure
The anaerobic sludge taken from an anaerobic–anoxic–oxic (A/
A/O) process-based WWTP in Beijing was used to enrich the ni-
trate-reducing microorganisms. The immobilization of denitrifying
bacteria to form biofilm on the surface of AC filler was prepared
through adaptation, enrichment, immobilization and acclimatiza-
tion (Zhou et al., 2009). About 10 L of seed sludge (i.e. a mixture
of wastewater and sludge) was added to a plastic tank (12 L). A
nutrient solution of KNO3, CH3COONa, and KH2PO4 was added into
the tank according to C:N:P = 100:5:1, with a NO3–N concentra-
tion of 100 mg/L and pH of 7. By using a peristaltic pump, the
sludge-water was driven to continuously circulate between the
3D-BER and plastic tank through the reactor inlet and outlet. Nutri-
ent solution was supplemented to the tank once everyday based on
desired C:N:P ratio. Such circulation was conducted for the first
2 days until the yellowish-brown sludge was evenly distributed
on AC filler.
After the first step, the immobilization and acclimatization of
biofilm were conducted to enrich the denitrifying bacteria. The
tap water was amended with a stock solution of KNO3, CH3COONa,
and KH2PO4 to form a NO3–N concentration of 155–178 mg/L and a
water (about 40 L, amended once every 3 days) was added to the
7.0 with KH2PO4 and K2HPO4. The wastewater was pumped from
the influent tank to flow through the 3D-BER, and was then
drained to the laboratory drainage system. To make microorgan-
was gradually changing from 0 to 20, 20 to 40, and 40 to 60 mA,
with each current level lasting for about 1 week. The influent and
effluent were run continuously, and the hydraulic retention time
(HRT) was set at 7 h. This phase lasted for about 1 month, and
the effluent and influent concentration of nitrate were also moni-
moved. A thick layer of dark-brown biofilm was visibly seen to
cover the AC filler, and the effluent of 3D-BER was also seen with
no suspended sludge, indicating that the biofilm had been well
formed. All bacterial inoculation and acclimation were carried
out during winter time at a room temperature of 10–15 °C.
system was operated under different experimental conditions of
C/N ratio and HRT using simulated WWTP effluent as described
in Section 2.1. The simulated WWTP effluent was put in the 60-L
influent tank, and was pumped to flow through the 3D-BER, and
then flew out of the reactor and was discharged to the laboratory
drainage system. A constant current of 40 mA was applied for all
the normal operation experiments. The C/N ratio was tested for
0.5, 1.0, 1.5, 2.0, and 3.0, and the HRT was tested for 5, 7, 10, and
12 h, respectively. The operational conditions (C/N and HRT) for
each run were maintained until steady nitrate removal for at least
6 days. At every 24 h interval, 50 mL of sample was taken from the
reactor effluent for nitrate and nitrite concentration analysis, and
each sample was taken for at least two successive HRTs.
2.3. Analytical methods
On a UV–Visible spectrophotometer (Shimadzu, Japan), the ni-
trate (NO3–N), nitrite (NO2–N), and total nitrogen (TN) in water
samples were measured using ultraviolet spectrophotometry,
spectrophotometry based on N-(1-naphthyl)ethylenediamine
dihydrochloride, and ultraviolet spectrophotometry based on alka-
line potassium persulfate digestion, respectively (SEPA, 2002).
CODCr was measured using a COD analyzer (model 5B-6C provided
by Lianhua Technology Co., Ltd., Beijing, China). pH was measured
using a pH meter (model E-201-9 provided by Shanghai Yoke
Instrument Co., Ltd., China). The temperature of water sample
was measured using a thermometer that was fixed on the wall of
the 3D-BER reactor. The removal of nitrate or total nitrogen (g)
was calculated using the following formula:
C0  Ct
g¼  100% ð1Þ
C0
where C0 is the initial concentration of nitrate or total nitrogen (mg/
L), and Ct (mg/L) is the remaining concentration of nitrate or total
nitrogen at time t.
2.4. Microbiological analysis
Bacterial community in the biofilm formed on the AC filler in
3D-BER was analyzed by DNA-based molecular techniques. About
20 mL of AC filler sample was taken from the 3D-BER after
2 months of biofilm immobilization and acclimatization as well
as normal operation when its operating condition was C/N = 1.0,
 R. Hao et al. / Bioresource Technology 143 (2013) 178–186
I = 40 mA, and HRT = 7 h. The sample was put in a 50-mL beaker to
be stirred by using a glass rod, and the biofilm was then stripped
off from the AC filler. The stripped biofilm sample (about 10 mL)
was collected in a 10-mL centrifuge tube and stored at 80 °C be-
fore analysis. The sample was sent to Sangon Biotech (Shanghai)
Co., Ltd., in China (a professional biotechnology company) for the
characterization of bacterial community by means of DNA extrac-
tion, PCR amplification, cloning, sequencing and phylogenetic
analysis.
2.4.1. DNA extraction and PCR amplification
A detailed description of DNA extraction and PCR amplification
of 16S rDNA can be found in various literatures (Park et al., 2006;
Brons and van Elsas, 2008). Total DNA was extracted from about
0.2 mL of wet biofilm sample with the SK1201-UNIQ-10 column-
type DNA extraction kit (Sangon Biotech Shanghai Co., Ltd., China)
according to the manufacturer’s instructions. The DNA was precip-
itated by adding ethanol before using as a template in polymerase
chain reaction (PCR) amplification (Park et al., 2006). A forward
primer (50f; 50 -AACACATGCAAGTCGAACG-30 ) and a reverse primer
(1492r; 50 -GGTTACCTTGTTACGACTT-30 ) were used for PCR amplifi-
cation. The PCR mixture contained 1 lL of extracted DNA, 10 lM
concentration of each primer (0.5 lL for each), 10 mM concentra-
tions of deoxyribonucleotide triphosphate (dNTP) (0.5 lL), 2.5 lL
of 10 Taq reaction buffer, and 5 U/lL concentration of Taq DNA
Polymerase (0.2 lL) (Sangon biotech. Co., Ltd., Shanghai in China).
The PCR mixture was added with sterile deionized water to obtain
a final volume of 25 lL. The PCR amplifications were performed in
0.2-mL reaction tubes on a model 2720 Thermal Cycler (BBI, Can-
ada) using the following programs: initial denaturation for 5 min
at 94 °C, 35 cycles of denaturation (30 s at 94 °C, 30 s at 55 °C,
60 s at 72 °C), and a final extension of 72 °C for 8 min. All of the
PCR amplification products were purified with the Agarose Gel
DNA purification kit (Sangon Biotech Shanghai Co., Ltd., China)
according to the manufacturer’s instructions, and analyzed by elec-
trophoresis in 1.0% (wt/vol) agarose gels followed by ethidium bro-
mide staining (Brons and van Elsas, 2008).
2.4.2. 16S rDNA sequencing and phylogenetic analysis
The purified PCR amplicons were cloned using the T-carrier
cloning kit (Sangon Biotech Shanghai Co., Ltd., China) following
the manufacturer’s instructions. After blue and white screening,
102 positive (i.e. white) clones were randomly selected and then
sequenced by a 3730 type DNA sequencing system (ABI, USA) using
a BigDye Terminator v1.1 sequencing kit (ABI, USA). The sequences
were compared with those in the sequence database from Ribo-
somal Database Project (RDP) (http://rdp.cme.msu.edu/seqmatch/
seqmatch_intro.jsp). Sequences determined in this study and those
retrieved from the database were aligned using a DNAStar SeqMan
software (Brons and van Elsas, 2008). Phylogenetic tree was con-
structed by a neighbor-joining algorithm using the MegAlign soft-
ware (DNASTAR, USA) to obtain broader groupings to reveal the
compositions of bacterial community (Park et al., 2006).
2.4.3. Scanning electron microscopy (SEM) analysis
The surface morphology for microorganism in biofilm was
investigated by scanning electron microscopy (SEM). A series of
sample treatment procedures (e.g., fixation, washing, dehydration,
drying, and coating) were applied before SEM analysis (Park et al.,
2005; Cho et al., 2008). About 0.6 g of biofilm sample was put in a
10-mL centrifuge tube (CT) and was centrifuged at 6000 rpm for
5 min (model HC-2518 provided by Anhui Zhongke Zhongjia Scien-
tific Instrument Co., Ltd., China). The supernatant was removed and
7 mL of 2.5% glutaraldehyde was then added to mix with the set-
tled sample in the bottom of CT. The CT was then placed in a refrig-
erator at 4 °C for 4 h to allow for sample fixation. After that, the CT
181
was put in the centrifuge for centrifugation at 6000 rpm for 5 min,
and the supernatant was removed after centrifugation. The left bio-
film sample in the bottom of CT was added with about 7 mL of
ultrapure water and was centrifuged again at 6000 rpm for
5 min, and the supernatant was removed. Such washing through
centrifugation was conducted for three times. The washed biofilm
was then dehydrated with 7 mL of 30%, 50%, 70%, 80%, and 90% eth-
anol (static dehydration for 10 min for each concentration of etha-
nol), respectively. The biofilm was finally dehydrated three times
with about 7 mL of 100% ethanol (15 min for each time). At last,
7 mL of ethanol and isoamyl acetate mixture (100% ethanol:iso-
amyl acetate = 1:1) was poured into the CT staying for 15 min. Pure
isoamyl acetate was then added for substitution for 15 min. The
settled sample was put on a filter paper (10–15 lm) and was dried
with a vacuum dryer (Alpha 1–2 LD plus, CHRIST) for 24 h at a pre-
cooling temperature of 10 °C. The dried sample was then sput-
tered with a 10–20 nm gold layer (Puig et al., 2011). The coated
sample was examined with a SEM (HITACHIS-4300, Hitachi), and
the images were captured digitally.
3. Results and discussions
3.1. Effect of C/N ratio on denitrification performance
Fig. 2a and b present the variations of average nitrogen concen-
tration and nitrogen removal under different C/N ratio conditions
during the 3D-BER treatment process (with HRT = 7 h), while the
nitrate (NO3–N) concentration in the influent was kept as a rela-
tively constant value (i.e. 30 mg/L). The experiment for each C/N
ratio was conducted for at least 6 days. It was observed that the
C/N ratio greatly affected the accumulation of nitrite and the re-
moval of nitrogen, with low C/N ratio beneficial for nitrite forma-
tion. It was observed that the concentration of NO3–N in the
effluent of 3D-BER decreased from 16.69 to 12.15, 12.15 to 8.55,
and 8.55 to 0.51 mg/L when the C/N ratio increased from 0.5 to
1.0, 1.0 to 2.0, and 2.0 to 3.0, respectively. Similarly, the concentra-
tion of nitrite (NO2–N) in the effluent decreased from 2.35 to 1.78,
1.78 to 1.39, and 1.39 to 0.09 mg/L when the C/N ratio increased
from 0.5 to 1.0, 1.0 to 2.0, and 2.0 to 3.0, respectively. In general,
denitrification efficiency increased with C/N ratio, and a high C/N
ratio of 3.0 significantly enhanced the removal of nitrogen and
the restriction of nitrite accumulation. The removal efficiency of ni-
trate increased from 48.9% to 62.2% when the C/N ratio increased
from 0.5 to 1.0, and a relatively stable nitrate removal efficiency
of around 71.7% was maintained for C/N ratio ranging from 1.5 to
2.0. However, a very high nitrate removal of 98.3% was obtained
when the C/N ratio increased from 2.0 to 3.0. The removal effi-
ciency of total nitrogen was lower than that of nitrate due to the
formation of nitrite during the treatment process. The results ob-
tained in this study were in agreement with results reported in
previous studies of BER treatment which indicated that high ni-
trate removal efficiency (>95%) could be achieved when organic
carbon was abundant such as C/N > 2 (Zhao et al., 2012).
The denitrification effect can also be described using the ratio of
DC/DN, where DC and DN represent the difference of CODCr and
total nitrogen (TN) between the 3D-BER influent and effluent,
respectively. As a result, DC/DN ratio can be used to represent
the approximate COD consumption per unit mass of nitrate re-
moval. Fig. 2c presents the variation of DC/DN ratio under different
influent C/N conditions. The heterotrophic denitrification theoreti-
cally requires 2.86 g of COD to remove 1 g of NO3–N, although the
actual DC/DN ratio for complete nitrate removal is usually above
3.0 due to the presence of oxygen and cell synthesis of microorgan-
isms (Lee et al., 2001). In this study, sodium acetate (CH3COONa)
was added into the simulated WWTP effluent as organic carbon
182 R. Hao et al. / Bioresource Technology 143 (2013) 178–186

the 3D-BER process. However, when the organic carbon content

was gradually decreasing in the reactor, the heterotrophic denitri-
fying bacteria could be gradually domesticated into autotrophic
denitrifying bacteria, and the original autotrophic denitrifying bac-
teria could gradually become the main microorganisms for nitrate
removal. The autotrophic microorganisms usually need a long
adaptive phase with slow growth, and thus may lead to a lower
denitrification rate (Zhao et al., 2011). For example, when the C/
N ratio was 0.5, a DC/DN ratio of 0.77 was achieved, and the nitrate
removal of only 48.9% was obtained, with a considerable amount of
nitrite being accumulated. Thus, the autotrophic microorganisms
may have played a major role at a very low C/N ratio (i.e. 0.5).
When the influent C/N ratio ranged from 1.0 to 2.0, both heterotro-
phic and autotrophic denitrification microorganisms may play
important roles in nitrate removal when using the 3D-BER, with
a less considerable amount of nitrite being generated (Zhao et al.,
2012). For example, when the C/N ratio was 1.0, 1.5, and 2.0, a
DC/DN ratio of 0.99, 1.49, and 2.21 was achieved, and the average
nitrate removal of 62.2%, 70.4%, and 72.9% was obtained, respec-
tively. The nitrate removal was nearly stable especially for C/N ra-
tio of 1.5 and 2.0. Such cooperative heterotrophic and autotrophic
denitrification in 3D-BER system can lead to significant advantages
(e.g., high nitrogen removal efficiency and thus larger treatment
capacity) as compared with only autotrophic denitrification (Zhao
et al., 2011).

Fig. 2. Impact of C/N ratio on the denitrification effect of 3D-BER (experimental

condition: flow rate = 504 mL/h; HRT = 7 h; T = 19 ± 2 °C; influent pH  7.0;
I = 40 mA).

source for heterotrophic bacteria. It was observed from Fig. 2c that

DC/DN increased from 0.77 to 2.72 when the influent C/N in-
creased from 0.5 to 3.0. All of these DC/DN ratios are below the
theoretical value of 2.86 for complete nitrate removal by heterotro-
phic denitrifying microorganisms. This may indicate that both het-
erotrophic and autotrophic denitrification microorganisms existed
during the 3D-BER treatment process. A high C/N ratio can acceler-
ate the growth of heterotrophic denitrifying bacteria and thus pro-
mote the total denitrification rate (Zhao et al., 2012). A study by Pei
et al. (2010) found that a wetland based biological denitrification
removed 70% of NO3–N at a C/N ratio of 7.0, but a complete re-
moval of NO3–N and NO2–N was achieved at a C/N ratio of 8. In this
study, when the C/N ratio was 3.0, a DC/DN ratio of 2.72 was
achieved, and the nitrate removal and total nitrogen removal
reached 98.3% and 98.0%, respectively, while the nitrite production
was minimal. As a result, the heterotrophic denitrification may Fig. 3. Impact of hydraulic retention time (HRT) on nitrogen removal in 3D-BER
have played a major role at a relatively high C/N ratio when using (experimental conditions: C/N = 1.5; T = 24 ± 2 °C; influent pH  7.0; I = 40 mA.
 R. Hao et al. / Bioresource Technology 143 (2013) 178–186
3.2. Effect of HRT on denitrification performance
It was found from Fig. 2a that an influent C/N ratio of 1.5 was
associated with a relatively satisfactory denitrification perfor-
mance (i.e. nitrate and total nitrogen removal of 70.4% and
65.2%). This C/N ratio was then used to examine the effect of
HRT on the denitrification performance of 3D-BER, and Fig. 3 pre-
sents the results. The experiment for each HRT was conducted for
at least 8 days. It was observed that the nitrate and total nitrogen
removal reached 75.9% and 70.5% at a HRT of 7 h and a temper-
ature of 24 ± 2 °C, respectively. However, the corresponding re-
moval was 70.4% and 65.22% (Fig. 2a) at a HRT of 7 h and a
temperature of 19 ± 2 °C, respectively. Such difference might
mainly be caused by temperature variation during different
experimental period, with higher temperature promoting nitrate
removal. It was found from Fig. 3 that HRT could greatly affect
the removal of nitrogen and the accumulation of nitrite. At low
HRT condition such as HRT = 5 h, a relatively lower denitrification
rate (e.g., nitrate and total nitrogen removal of 66.0% and 54.6%)
was achieved, and a NO3–N and NO2–N concentration of above
10 and 3.5–4 mg/L was observed in the effluent of 3D-BER,
respectively. The denitrification rate increased with HRT, but it
reached a maximum value (e.g., average nitrate removal of
88.4%, ranging from 85.0% to 90.0%) when the HRT was 10 h. Fur-
ther increase of HRT beyond 10 h resulted in the decrease of deni-
trification rate. Longer HRT (i.e. 10–12 h) also led to less
accumulation of nitrite in the 3D-BER effluent (e.g., less than
1 mg/L). In the environment with relatively poor organic carbon
source (i.e. C/N = 1.5), both heterotrophic and autotrophic bacte-
rial activities may be important, but the autotrophic bacteria need
more time to degrade nitrogen due to their relatively slow
growth. As a result, a longer HRT could facilitate the nitrogen re-
moval by such denitrifying bacteria and result in a lower nitrite
accumulation. However, in the continuous running mode of 3D-
BER, too long HRT under relatively poor organic carbon source
condition could result in the endogenous respiration phase of het-
erotrophic bacteria (i.e. a phase when bacteria live with nutrition
in their bacterial body) for a long time. This would result in the
reduction of heterotrophic bacteria and thus the poor denitrifica-
tion effect, such as the reduced nitrate and total nitrogen removal
of 79.0% and 76.0% at HRT = 12 h as shown in Fig.3. Consequently,
when using 3D-BER for treating WWTP effluent with poor organic
carbon source, the extension of HRT could enhance the nitrogen
removal and restrict the nitrite accumulation, but too long HRT
is not necessary due to the cooperative heterotrophic and auto-
trophic denitrification mechanisms.
3.3. Bacterial community in 3D-BER
The analysis of 16S rDNA gene clones was conducted to assess
the bacterial community in the biofilm of 3D-BER. A total of 102
Table 1
Phylogenetic groups in the biofilm of 3D-BER.
Phylogenetic group Number of clones Percentage (%)
Proteobacteria
a-proteobacteria 1 0.98
b-proteobacteria 63 61.77
c-proteobacteria 31 30.39
e-proteobacteria 2 1.96
Firmicutes
Clostridia 3 2.94
Bacilli 1 0.98
Actinobacteria
Actinobacteria 1 0.98
Total 102 100
183
clones were obtained and sequenced without dividing operational
taxonomic units (OUTs). Table 1 lists the summarized phylogenetic
groups of all of the analyzed bacterial clones. It can be found that
the 3D-BER biofilm was dominated by Proteobacteria phylum in
which the b- and c-Proteobacteria were the most abundant,
accounting for 61.77% and 30.39% of the total clones, respectively.
Only a small percentage (4.90%) of the total clones were related to
other Gram-positive phyla of Firmicutes and Actinobacteria which
were usually detected in anaerobic sludge digester and methano-
genic inoculums (Ndez et al., 2008).
Proteobacteria was found to be dominant in many sulfur-
based chemolithotrophic denitrification bioreactor studies (Ndez
et al., 2008). Park et al. (2006) reported that Proteobacteria and
flavobacteria were the dominant bacteria in a conventional bio-
film-electrode reactor where the b-Proteobacteria played key
roles in the denitrification performance of biofilm reactor. How-
ever, the flavobacteria was not detected in the biofilm of 3D-BER
in this study. It was reported that b-Proteobacteria bacteria tend
to survive in the anaerobic environment utilizing organic nutri-
ents, while some of them may also utilize hydrogen, ammonia,
methane and volatile fatty acids (Xia et al., 2010). Thus, the
anaerobic operating environment in 3D-BER may facilitate the
survival of b-Proteobacteria in the biofilm (i.e. 61.77% of the total
clones). The b-Proteobacteria detected in the biofilm sample was
further subdivided into classes and genera as shown in Fig. 4. A
significant microbial diversity in the biofilm was observed.
Fig. 4a shows that 57.15% of the total clones related to b-Prote-
obacteria were pertaining to the family Rhodocyclaceae class (i.e.
the genera of Thauera, Sulfuritalea and Azospira). Among them,
83.0% of the total clones belonged to genus Thauera, which have
been reported in many denitrifying bioreactor studies (Osaka
et al., 2008; Daniel et al., 2009). Mao, 2009 examined 8 Thauera
bacteria (Thauera aminoaromatica, Thauera linaloolentis, Thauera
phenylacetica, Thauera terpenica, Thauera sp. DNT-1, Thauera sp.
27, Thauera sp. 28, and Thauera sp. 63), and found that all of
them had denitrification function. It was found from Fig. 4a that
33.33% of the b-Proteobacteria consisted of the family class of
Comamonadaceae (Acidovorax, Albidiferax, Aquincola, Giesbergeria,
Hydrogenophaga, Pelomonas, Polaromonas, Simplicispira, unclassi-
fied Comamonadaceae), which had been found in denitrifying
reactors with acetate as the organic carbon source for microor-
ganism (Ginige et al., 2005; Osaka et al., 2008). In this study,
the organic carbon source in the 3D-BER was supplemented with
sodium acetate, and this would lead to a relatively high percent-
age of Comamonadaceae bacteria in the biofilm. The Burkholderi-
ales-like bacteria (Aquabacterium, Ideonella, Rubrivivax,
unclassified Burkholderiales) accounted for 9.52% of the b-Proteo-
bacteria in the biofilm of 3D-BER, while this family class of bac-
teria have been reported in the studies of anoxic/aerobic
digestion bioreactors (Zheng, 2012).
Unlike the prominent diversity of b-Proteobacteria in the 3D-
BER biofilm, c-Proteobacteria contained three genera (Enterobacter,
Citrobacter and Erwinia) (Fig. 4c), where the Enterobacter-like bac-
teria of Enterobacteriaceae took up 88.89% of the clones belonging
to c-Proteobacteria. Enterobacteriaceae-like bacteria were found to
have the physiological function of denitrification and polyphos-
phate accumulation even under anoxic condition, although their
denitrification capacity and aerobic phosphorus accumulation
capacity were excellent (Bao et al., 2007). This can explain the re-
moval of both nitrate and phosphorous in 3D-BER. Fig. 5 presents
the phylogenetic tree of gene sequences in 3D-BER biofilm, where
M represents a clone. For example, clone M89 represents the Enter-
obacter-like bacteria. Such significant diversity of microbial com-
munity in the biofilm as shown in Fig. 5 could enhance the
denitrification performance of 3D-BER for wastewater with low
C/N ratio.
184 R. Hao et al. / Bioresource Technology 143 (2013) 178–186

Fig. 4. Distribution of b-and c-Proteobacteria in the biofilm of 3D-BER (a) percentage of b-proteobacteria classes; (b) percentage of genera in b-proteobacteria; (c) percentage
of genera in c-proteobacteria (experimental condition: C/N = 1, HRT = 7 h, T = 19 ± 2 °C, I = 40 mA).

Fig. 5. Phylogenetic tree of gene sequences in 3D-BER (note: M represents a clone).

 R. Hao et al. / Bioresource Technology 143 (2013) 178–186
3.4. Surface morphology of biofilm in 3D-BER
The surface morphology and structure of biofilm in the 3D-BER
under different influent C/N ratio conditions were obtained (see
Supplementary Material). It was found that the microorganisms
in the 3D-BER biofilm at relatively low C/N ratio (i.e. 1.0) were
mainly short rod (1–2 lm) shaped bacteria, while those at higher
C/N ratio consisted of both short rod (1–2 lm) and axiolitic shaped
(0.5–1 lm) bacteria. This is in agreement with the results of 16S
rDNA analysis which indicated that the biofilm was abundant with
Thauera-like and Enterobacter-like bacteria, while such bacteria
were mostly short rod shaped (Buchanan and Gibbone, 1984;
Mao, 2009). It can also be found that the biofilm contained a much
less amount of bacteria at relatively low influent C/N ratio than
that at a higher C/N ratio (see Supplementary Material). The reason
for this can be explained that the growth of heterotrophic bacteria
depends on organic carbon source. The normal operation experi-
ments of 3D-BER in this study started from a high C/N ratio (i.e.
3.0) to a low C/N ratio (i.e. 0.5). The low C/N ratio could lead to
the slow growth of bacteria and thus decrease the amount and type
of heterotrophic bacteria (e.g., mainly short rod bacteria, 1–2 lm)
as compared to environment condition with higher influent C/N
(e.g., both short rod and axiolitic bacteria, 1–2 and 0.5–1 lm). In
fact, it was visibly seen during the experiments that the C/N ratio
directly affected the amount of biofilms, and the biofilms at C/N ra-
tio of 0.5 and 1.0 were thinner and less intensive as compared to
those at a C/N ratio of 3.0. It was also found that an extensive pore
structure existed in the biofilm, and this could facilitate the suffi-
cient contact of bacteria with nitrogen components and the easy
flow of nitrogen gas during the denitrification process, leading to
enhanced denitrification rate. However, for the biofilm with low
C/N ratio, such pore structure was not obvious.
4. Conclusion
The 3D-BER technology was effectively used for removing ni-
trate from simulated WWTP effluent. An average nitrate removal
of 98.3% and 88.4% was obtained at C/N ratio of 3.0 and HRT of
7 h as well as at C/N ratio of 1.5 and HRT of 10 h, respectively.
The biofilm on the AC filler of reactor was dominated by Proteobac-
teria in which the b- and c-Proteobacteria were the most abundant.
The diverse and dynamic bacterial community composition during
treatment could greatly improve the performance of 3D-BER and
make it an attractive denitrification technology for wastewater
with poor organic carbon source.
Acknowledgements
This research was supported by the Natural Science Foundation
of China (No. 51178005) and Beijing Natural Science Foundation
(No. 8132017). The authors thank the journal editor and the anon-
ymous reviewers for their comments and suggestions that helped
in improving the manuscript.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.biortech.2013.
06.001.
References
Bao, L.L., Li, D., Li, X.K., Huang, R.X., Zhang, J., Lv, Y., Xia, G.Q., 2007. Phosphorus
accumulation by bacteria isolated from a continuous-flow two-sludge system. J.
Environ. Sci. 19 (4), 391–395.
185
Brons, J.K., van Elsas, J.D., 2008. Analysis of bacterial communities in soil by use of
denaturing gradient gel electrophoresis and clone libraries, as influenced by
different reverse primers. Appl. Environ. Microbiol. 74 (9), 2717–2727.
Buchanan, R.E., Gibbone, E.N., 1984. Bergey’s Manual of Determinative Bacteriology.
Science Press, Beijing.
Cho, J.S., Park, J.Y., Yoo, Y.J., 2008. Novel 3-dimensional bioelectrode for
mediatorless bioelectrochemical denitrification. Biotechnol. Lett. 30 (9),
1617–1620.
Daniel, L.M.C., Pozzi, E., Foresti, E., Chinalia, F.A., 2009. Removal of ammonium via
simultaneous nitrification–denitrification nitrite-shortcut in a single packed-
bed batch reactor. Bioresour. Technol. 100 (3), 1100–1107.
Feng, H., Huang, B., Zou, Y., Li, N., Wang, M., Yin, J., Cong, Y., Shen, D., 2013. The effect
of carbon sources on nitrogen removal performance in bioelectrochemical
systems. Bioresour. Technol. 128, 565–570.
Ghafari, S., Hasan, M., Aroua, M.K., 2008. Bio-electrochemical removal of nitrate
from water and wastewater – a review. Bioresour. Technol. 99 (10), 3965–3974.
Ginige, M.P., Keller, J., Blackall, L.L., 2005. Investigation of an acetate-fed denitrifying
microbial community by stable isotope probing full-cycle rRNA analysis and
fluorescent in situ hybridization–microautoradiography. Appl. Environ.
Microbiol. 71 (12), 8683–8691.
Hao, R.X., Li, S.M., Li, J.B., Zhang, Q.K., Liu, F., 2013. Water quality assessment for
wastewater reclamation using principal component analysis. J. Environ. Inf. 21
(1), 45–54.
Huang, B., Feng, H., Wang, M., Li, N., Cong, Y., Shen, D., 2013. The effect of C/N ratio
on nitrogen removal in a bioelectrochemical system. Bioresour. Technol. 132,
91–98.
Karanasios, K.A., Vasiliadou, I.A., Pavlou, S., Vayenas, D.V., 2010. Hydrogenotrophic
denitrification of potable water: a review. J. Hazard. Mater. 180 (1–3), 20–37.
Lee, D.U., Lee, I.S., Choi, Y.D., Bae, J.H., 2001. Effects of external carbon source and
empty bed contact time on simultaneous heterotrophic and sulfur-utilizing
autotrophic denitrification. Process Biochem. 36 (12), 1215–1224.
Li, M., Feng, C., Zhang, Z., Yang, S., Sugiura, N., 2010. Treatment of nitrate
contaminated water using an electrochemical method. Bioresour. Technol.
101 (16), 6553–6557.
Li, T., Hao, R., Liu, F., Liu, J., 2011. Analysis on running effect for biological nitrogen
and phosphorus removal of A/A/O technology. Chin. J. Environ. Eng. 5 (8), 1729–
1734 (in Chinese).
Mao, Y., 2009. Structural and Functional Analysis of Thauera Genus in Wastewater
Treatment Plants (Ph.D. thesis). Shanghai Jiao Tong University, China (in
Chinese).
Mousavi, S., Ibrahim, S., Aroua, M.K., Ghafari, S., 2012. Development of nitrate
elimination by autohydrogenotrophic bacteria in bio-electrochemical reactors –
a review. Biochem. Eng. J. 67, 251–264.
Ndez, N.F., Sierra-Alvarez, R., Field, J.A., Amils, R., Sanz, J.L., 2008. Microbial
community dynamics in a chemolithotrophic denitrification reactor inoculated
with methanogenic granular sludge. Chemosphere 70 (3), 462–474.
Osaka, T., Shirotani, K., Yoshie, S., Tsuneda, S., 2008. Effects of carbon source on
denitrification efficiency and microbial community structure in a saline
wastewater treatment process. Water Res. 42 (14), 3709–3718.
Park, H.I., Choi, Y., Pak, D., 2005. Nitrate reduction using an electrode as direct
electron donor in a biofilm-electrode reactor. Process Biochem. 40 (10), 3383–
3388.
Park, H.I., Kim, J.S., Kim, D.K., Choi, Y.J., Pak, D., 2006. Nitrate-reducing bacterial
community in a biofilm-electrode reactor. Enzyme Microb. Technol. 39 (3),
453–458.
Pei, Y., Yang, Z., Tian, B., 2010. Nitrate removal by microbial enhancement in a
riparian wetland. Bioresour. Technol. 101 (14), 5712–5718.
Puig, S., Serra, M., Vilar-Sanz, A., Cabré, M., Bañeras, L., Colprim, J., Balaguer, M.D.,
2011. Autotrophic nitrite removal in the cathode of microbial fuel cells.
Bioresour. Technol. 102 (6), 4462–4467.
Rao, N.N., Rohit, M., Nitin, G., Parameswaran, P.N., Astik, J.K., 2009. Kinetics of
electrooxidation of landfill leachate in a three-dimensional carbon bed
electrochemical reactor. Chemosphere 76 (9), 1206–1212.
SEPA (State Environmental Protection Administration), 2002. Monitoring and
Analytical Methods for Water and Wastewater, fourth ed. China
Environmental Science Press, Beijing (in Chinese).
Srinandan, C.S., D’Souza, G., Srivastava, N., Nayak, B.B., Nerurkar, A.S., 2012. Carbon
sources influence the nitrate removal activity, community structure and biofilm
architecture. Bioresour. Technol. 117, 292–299.
Wei, D., Tan, Z., Du, Y., 2011. A biological safety evaluation on reclaimed water
reused as scenic water using a bioassay battery. J. Environ. Sci. 23 (10), 1611–
1618.
Xia, S., Li, J., Wang, R., Li, J., Zhang, Z., 2010. Tracking composition and dynamics of
nitrification and denitrification microbial community in a biofilm reactor by
PCR–DGGE and combining FISH with flow cytometry. Biochem. Eng. J. 49 (3),
370–378.
Yi, L., Jiao, W., Chen, X., Chen, W., 2011. An overview of reclaimed water reuse in
China. J. Environ. Sci. 23 (10), 1585–1593.
Zhao, H., Sun, Y., Xu, L., Ni, J., 2010. Removal of acid orange 7 in simulated
wastewater using a three-dimensional electrode reactor: removal mechanisms
and dye degradation pathway. Chemosphere 78 (1), 46–51.
Zhao, Y., Feng, C., Wang, Q., Yang, Y., Zhang, Z., Sugiura, N., 2011. Nitrate removal
from groundwater by cooperating heterotrophic with autotrophic
denitrification in a biofilm-electrode reactor. J. Hazard. Mater. 192 (3), 1033–
1039.
186 R. Hao et al. / Bioresource Technology 143 (2013) 178–186

Zhao, Y., Zhang, B., Feng, C., Huang, F., Zhang, P., Zhang, Z., Yang, Y., Sugiura, N., 2012.
Behavior of autotrophic denitrification and heterotrophic denitrification in an
intensified biofilm-electrode reactor for nitrate-contaminated drinking water
treatment. Bioresour. Technol. 107, 159–165.
Zheng, L., 2012. A Research on the Mechanism and Application of Ultrasonic-
Anoxic/Aerobic Digestion for Sludge Reduction (Ph.D. thesis), Guangdong
University of Technology, China (in Chinese).
Zhou, M., Fu, W., Gu, H., Lei, L., 2007. Nitrate removal from groundwater by a novel
three-dimensional electrode biofilm reactor. Electrochim. Acta 52 (19), 6052–
6059.
Zhou, M., Wang, W., Chi, M., 2009. Enhancement on the simultaneous removal of
nitrate and organic pollutants from groundwater by a three-dimensional bio-
electrochemical reactor. Bioresour. Technol. 100 (20), 4662–4668.
Zhou, Y., Gao, L., Xia, Y.F., Li, W., 2012. Enhanced reduction of Fe(II)EDTA–NO/
Fe(III)EDTA in NOx scrubber solution using a three-dimensional biofilm-
electrode reactor. Environ. Sci. Technol. 46 (22), 12640–12647.