Bioresource Technology 96 (2005) 633–643
Effects of alkalinity and co-substrate on the performanceof an upflow anaerobic sludge blanket (UASB) reactor
through decolorization of Congo Red azo dye
Mustafa I+ık, Delia Teresa Sponza *
Environmental Engineering Department, Engineering Faculty, Dokuz Eylul University, Buca Kaynaklar Campus, 35160, Izmir, Turkey
Received 23 July 2003; received in revised form 26 May 2004; accepted 1 June 2004
Available online 7 August 2004
Abstract
The effect of substrate (glucose) concentrations and alkalinitiy (NaHCO3) on the decolorization of a synthetic wastewater
containing Congo Red (CR) azo dye was performed in an upflow anaerobic sludge blanket (UASB). Color removal efficiencies
approaching 100% were obtained at glucose-COD concentrations varying between 0 and 3000 mg/l. The methane production
rate and total aromatic amine (TAA) removal efficiencies were found to be 120 ml per day and 43%, respectively, while the
color was completely removed during glucose-COD free operation of the UASB reactor. The complete decolorization of CR
dye under co-substrate free operation could be attributed to TAA metabolism which may provide the electrons required for
the cleavage of azo bond in CR dye exist in the UASB reactor. No significant differences in pH levels (6.6–7.4), methane pro-
duction rates (2000–2700 ml/day) and COD removal efficiencies (82–90%) were obtained for NAHCO3 concentrations ranging
between 550 and 3000 mg/l. However, decolorization efficiency remained at 100% with decreasing NaHCO3 concentrations as
low as 250 mg/l in the feed. An alkalinity/COD ratio of 0.163 in the feed was suggested for simultaneous optimum COD and
color removal.
� 2004 Elsevier Ltd. All rights reserved.
Keywords: Azo dye; UASB; Glucose-COD; Alkalinity; Decolorization; Aromatic amines; Color
1. Introduction
Azo colorants are substances that have a coloring ef-
fect of one or more azo groups (N‚N double bonds) in
their chemical structure. They can be used in synthetic
and natural textile fibers, plastics, leather, paper, min-
eral oils, waxes, and even (with selected types) foodstuffs
0960-8524/$ - see front matter � 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.biortech.2004.06.004
* Corresponding author. Tel.: +90 232 4531008/1119; fax: +90 232
4531153.
E-mail addresses: [email protected] (M. I+ık), delya.
[email protected] (D.T. Sponza).
and cosmetics (Geisberger, 1997). Azo dyes are the class
of dyes most widely used in textile finishing, having aworld market share of 60–70%. They have become of
concern in wastewater treatment because of their color,
bio-recalcitrance, and potential toxicity to animals and
humans (Yoo et al., 2000).
The azo dyes are not readily degradable under natu-
ral conditions and are typically not removed from
wastewater by conventional wastewater-treatment sys-
tems. Several chemical and physical decolorizationmethods such as coagulation/flocculation and precipita-
tion, oxidation, adsorption, electrolysis, and membrane
extraction are available (Cooper, 1993). These methods
are costly. In recent years there has been a tendency to
634 M. Is�ık, D.T. Sponza / Bioresource Technology 96 (2005) 633–643
use biological treatment systems to treat wastewaters
containing azo dyes. The recalcitrant nature of azo dyes
together with their toxicity to microorganisms makes
aerobic treatment difficult (Knackmuss, 1996). Under
anaerobic conditions, azo dyes are readily cleaved via
a four electron reduction through the linkage of azobonds generating aromatic intermediates (Wuhrmann
et al., 1980; Weber and Wolfe, 1987). Various azo dyes
were shown to be decolorized by anaerobic sludge, sedi-
ments and pure cultures in batch reactors (Rajaguru
et al., 2000; I+ık and Sponza, 2003). Currently, in the
biological treatment of textile wastewater, studies have
been focused on anaerobic biotechnology since anaero-
bic treatment can achieve color removal. Furthermorethe anaerobic treatment has classical advantages which
includes low sludge formation, no aeration requirement
and methane gas formation when compared to aerobic
treatment systems (Delee et al., 1998). Generally during
the anaerobic process a reduction in COD of up to 60–
70% can also be achieved (Zaoyan et al., 1992). Hence it
seems that an anaerobic processes hold promises for the
effective treatment of textile wastewaters on the basis ofCOD and color. Upflow anaerobic sludge blanket
(UASB) reactors are the most commonly used high rate
anaerobic systems. They are generally used for waste-
waters that have a low suspended solid concentration
and can be used for the treatment of dye wastes (An
et al., 1996; Razo-Flores et al., 1997; Delee et al.,
1998; O�Neill et al., 2000a).
Several studies have reported that suitable co-sub-strate and alkalinity in a bioreactor are necessary for
color removal. The electrons required for decoloriza-
tion are provided by an electron donating carbon
source such as glucose, VFA or starch (Banat et al.,
1996; Brown and Laboureur, 1983; Chinwetkitvanich
et al., 2000; Bras et al., 2001). Gingell and Walker
(1971) and Carliell et al. (1995) reported that decolori-
zation take places under anaerobic conditions if co-subsrate is present to donate electrons. Co-substrates
were used in these studies for effective decolorization
of azo dyes. Bicarbonate alkalinity was used to provide
favorable conditions for conversion of substrate to
methane and enough electrons from electron transport
chain for concurrently decolorizing of azo dyes (Carliell
et al., 1996). However, studies performed by Razo-
Flores et al. (1997) and Beydilli et al. (1998) showedthat decolorization of dyes did not depend on COD
since the dye itself can be used as carbon source by
the anaerobic microorganisms. It is generally assumed
that the aromatic amines generated from the linkage
of azo bound compounds are not further degraded
under anaerobic conditions (Brown and Hamburger,
1987; Haug et al., 1991). However, several studies dem-
onstrated that the aromatic amines could potentially befully or partially biodegradable in anaerobic environ-
ments (Razo-Flores et al., 1997; O�Connor and Young,
1993; Battersby and Wilson, 1989; Kalyuzhnyi et al.,
2000).
One of the major problems is the inadequate influent
alkalinity/influent COD ratio in the operation of the
UASB system treating textile industry wastewaters con-
taining azo dye. This may cause the minimum pH in thebed to fall below about 6.2–6.6, which can lead to failure
of the system. Carbon dioxide produced via micro-
organisms often exceeds the weak acids in aqueous
anaerobic systems. Therefore a sufficient bicarbonate
alkalinity must be present to neutralize it and is, there-
fore, of prime importance. If the acid concentrations
(H2CO3 and total volatile fatty acid––TVFA) exceed
the available alkalinity, the reactor will sour, severelyinhibiting microbial activity, especially the methanogens
(Speece, 1996). Souza et al. (1992) and Moosbruger et al.
(1993) found that an alkalinity/COD ratio of 0.5 in the
effluent decreased the pH to 6.6 which is considered as
the lower limit recommended for the anaerobic digestion
processes. With carbohydrate wastes the alkalinity
requirement is 1.2–1.6 g alkalinity as CaCO3/g influent
COD which is sufficient to maintain the pH above 6.6(Speece, 1996).
Limited studies have been performed investigating the
effect of co-substrate and alkalinity on the treatment effi-
ciencies using continuous plug flow systems. The effect of
different alkalinity and substrate concentrations on the
treatment efficiencies using continuous anaerobic reac-
tors treating azo dyes has not been extensively reported
in previous studies. For instance, Chinwetkitvanich etal. (2000) reported that the supplementation of tapioca
starch as a co-substrate apparently gave a better color re-
moval performance. However, excessively high concen-
trations of tapioca did not enhance the process
capability in terms of color removal efficiencies in an
UASB system. O�Neill et al. (2000b) suggested that if
the color removal efficiency decreases, carbonhydrate
should be added to UASB reactor operating atsludge loading rates varying between 0.11 and 0.15 kg
COD/kg TVSday. However, azodisalicylate azo dye (75
mg/l) could be continuously used as the sole carbon and
energy source with a removal efficiency of 89% in an
UASB reactor at a HRT of 26 (Razo-Flores et al., 1997).
The alkalinity required to buffer a wastewater con-
taining azo dyes and the supplementation of external
carbon source adversely affect the economical feasibil-ity of the anaerobic treatment. The cost for alkalinity
and sometimes for additional co-substrate supplemen-
tation to wastewater can easily exceed the value of
methane produced in UASB reactors. This study was
designed to investigate the effects of glucose-COD
and NaHCO3 alkalinity concentrations on the reactor
performances (COD, color, TAA removal efficiencies,
VFA and methane gas productions) in a lab-scale up-flow sludge blanket (UASB) reactor treating Congo
Red azo dye.
M. Is�ık, D.T. Sponza / Bioresource Technology 96 (2005) 633–643 635
2. Methods
2.1. Experimental design and operating conditions
In the first step of the study (Runs 1–5) the glucose-
COD was reduced from 3000 to 0 mg/l in order to deter-mine the effect of substrate on the color, COD removals
and methane gas productions. The operating conditions
during the operation in UASB reactor are summarized
in Table 1. The NaHCO3 concentrations were adjusted
by taking into consideration the alkalinity required for
each COD level under anaerobic conditions. Although
it was aimed to study at constant HRTs, they varied be-
tween 15.7 and 19 h due to some difficulties in adjustingthe dosage pump. The flow rate taking into considera-
tions are the mean values.
In the second step of the study (Runs 6–11) the COD
concentrations were kept constant at 2000 mg/l and the
effect of decreasing NaHCO3 concentrations (from
3000 to 250 mg/l) on the COD, color removal efficiencies,
TVFA accumulation and methane gas productions were
investigated. Although it was aimed to study at constantflowrates the HRTs varied between 15.54 and 21.83 h due
to the weekly variations in dosage pump (see Table 1).
The UASB reactor used in this study was previously
used for removal of synthetic wastewaters containing
Reactive Black 5 and Congo Red azo dyes (Sponza
and I+ık, 2002; I+ık and Sponza, 2002). After that, the
reactor was operated under dye-free conditions at an or-
ganic loading rate of 4.54 kg COD/m3day, with an HRTof 16 h in the beginning of these studies (Run 1). This
volumetric organic loading rate used in this study is near
the optimum organic loading rate range (2–4 kg COD/
m3day) proposed for UASB reactors treating textile
wastewater (Manu and Chaudhari, 2003). The COD re-
moval efficiency, gas production rate, methane percent-
age, pH and TVFA concentrations in effluent of UASB
reactor were 88%, 4280 ml/day, 80%, 7.3 and 78 mg
Table 1
Operation parameters and loading conditions of UASB reactor
Stage Period
(days)
Glucose-COD
(mg/l)
NaHCO3
(mg/l)
HRT
(h)
Effect of glucose-COD on UASB reactor performance
Run 1 0–16 3000 5000 15.7
Run 2 17–23 500 667 19
Run 3 24–30 250 333 18
Run 4 31–37 100 333 18.3
Run 5 38–50 0 333 18.9
Effect of alkalinity on UASB reactor performance
Run 6 51–56 2000 3000 20.4
Run 7 57–63 2000 3000 15.54
Run 8 64–70 2000 1500 15.54
Run 9 71–77 2000 750 18.19
Run 10 78–91 2000 550 21.83
Run 11 92–103 2000 250 17.6
CH3COOH/l, respectively, before starting the aforemen-
tioned runs, indicating the steady-state conditions for
effective continuous operation of UASB reactor. For
each run, steady-state conditions were assumed to be
achieved if the effluent COD concentrations and gas
productions values did not differ by more than ±5%for at least four consecutive days.
2.2. Experimental lab-scale reactors and seed
The anaerobic UASB reactor used for Congo red
decolorization was 6 cm in diameter, 100 cm in length
and had an effective volume of 2.5 l. The schematic con-
figuration of the sequential UASB reactor is illustratedin Fig. 1. Partially granulated anaerobic sludge was used
as seed in the UASB reactor and was taken from a meth-
anogenic reactor treating industrial effluent from the
Pakmaya Yeast Baker Factory in Izmir. The UASB
reactor was operated at 37 �C using an electronic heater
in the medium part of the system.
2.3. Azo dyes and composition of synthetic wastewater
Congo Red (Direct Red 28) which is a banned azo
dye including the carcinogenic aromatic amines (benzi-
dine) was used in dissolved form. The color index num-
ber, the maximum wavelength of absorbance, and the
COD value of 100 mg/l of CR dye solution are 22120,
497 nm, and 74.6 mg/l, respectively. The structure of this
dye is depicted in Fig. 2. Vanderbilt mineral medium(Speece, 1996) was used as feed.
2.4. Analytical procedures
Total suspended solid (TSS) in granulated sludge were
measured by filtration technique using membrane filters
with pore sized 0.45 lm (APHA, 1992). The COD in
influent and effluent samples were detected by using a
Org. load
(kg COD/m3day)
Dye load rate
(g dye/ m3h)
Upflow vel.
(m/h)
4.54 0 0.064
0.845 5.3 0.053
0.437 5.6 0.056
0.286 5.5 0.055
0.195 5.4 0.054
2.3 0 0.049
3.27 6.43 0.064
3.27 6.43 0.064
2.86 5.62 0.056
2.42 4.76 0.048
2.99 5.88 0.056
6 cm
10 c
m5
cm15
cm
15 c
m20
cm
20 c
m15
cm
100
cmTotal Gas Measurement2% (w/v) H SO 10% (w/v) NaCl solution2 43% NaOH (w/v) solution
Methane Measurement
Gas
Out
let
Time ControlledPerilstaltic Pump
UASB FeedingTank+4 C24 L
Sam
lplin
g Po
rts
Heater
EffluentTank
o
Fig. 1. The schematic configuration of the UASB reactor.
Fig. 2. The structure of Congo Red azo dye.
636 M. Is�ık, D.T. Sponza / Bioresource Technology 96 (2005) 633–643
closed reflux colorimetric method following standard
methods (APHA, 1992). Total volatile fatty acid (TVFA)
and bicarbonate alkalinity (B.Alk.) concentrations in the
effluent of UASB reactor were measured using the titri-
metric method described by Anderson and Yang
(1992). pH was determined immediately after sampling
to avoid any change due to the CO2 evolution using a
pH meter (WWT pH 330). Gas production was meas-ured with a liquid displacement method. Total gas was
measured by passing it through a liquid containing 2%
(v/v) H2SO4 and 10% (w/v) NaCl (Beydilli et al., 1998).
Methane gas was detected using a liquid solution con-
taining 3% NaOH (w/v) (Razo-Flores et al., 1997).
Color in the influent and effluent samples of the
UASB reactor was measured by an Optima Photomech
301-D UV–VIS spectrophotometer (Nova, UK) at max-imum wavelength. The samples were centrifuged at 7000
rpm for 10 min and the absorbance values of superna-
tants were recorded for color measurements. Abiotic
tests were performed with autoclaved anaerobic par-
tially granulated sludge and a mineral medium in batch
serum bottles at 37 �C to determine the biological dye
degradation.
The total aromatic amines were determined colori-
metrically at 440 nm after reacting with 4-dimethylami-nobenzaldehyde–HCl according to the method
described by Oren et al. (1991). Total aromatic amines
(TAA) released from anaerobic, and chemical reduction
were quantified using benzidine as a standard at absorb-
ance maxima of 440 nm. The chemical reduction of azo
dyes was carried out with sodium dithionite. The condi-
tions of the reduction process are as follows: 0.06 g of dye
sample heated in boiling temperature with 1 M NaOHfor 1 h. After 30 min 0.6 g sodium dithionite was added.
A 100 mg/l sample of dye was reduced according to
method mentioned by Pielesz (1999). The aromatic
amine recoveries were calculated from the TAA values
of effluents to expected TAA in influent samples.
2.5. Statistical analysis
The values given in figures are mean ± standard devi-
ation (SD) and n is the numbers of samples. Error bars
represent the SD.
M. Is�ık, D.T. Sponza / Bioresource Technology 96 (2005) 633–643 637
Regression analysis between y and x variables was
performed using the EXCELL in Microsoft Windows�(HP, USA). The linear correlation was assessed with r2.
The r2 value is the correlation coefficient and reflects sta-
tistical significance between dependent and independent
variables.
3. Results and discussion
3.1. Effect of glucose-COD on reactor performance
The effect of glucose-COD on the color and the COD
removal efficiencies are shown in Fig. 3. Color was re-moved very effectively (approximately 100%), at all glu-
cose-COD concentrations varying between 100 and 500
mg/l. This shows that the electrons required for the
reductive cleavage of azo dyes was not depend to glu-
cose-COD through decolorization under anaerobic con-
ditions. A slight linear relationship between color
removal and COD concentrations was observed for
the r2 values of 0.42 but it was not statistically significant(P = 0.34 and F = 1.47). The electrons providing the re-
duced environment appear to be sufficient for the cleav-
age of the azo bond at COD concentrations as low as
100 mg/l for complete decolorization. Both COD levels
were able to perform more than 98–99% color removal
in CR dye. In Run 5, the co-substrate was excluded in
the influent whether CR azo dye could be degraded as
the sole carbon and energy source by the anaerobicmicroorganisms in the UASB reactor. As shown in
Fig. 3, almost complete (99%) decolorization was
achieved even at substrate-free operation. This shows
that the dye and the degradation products could be used
as a carbon source by methanogens as reported by
Manu and Chaudhari (2002). The abiotic decolorization
studies with nonviable granulated sludge indicated that
no significant color removals (E = 1–2%) were observedin co-substrate containing and co-substrate free samples.
0
10
20
30
40
50
60
70
80
90
100
500 250 100 0Glucose-COD concentration in feed (mg/l)
CO
D a
nd c
olor
rem
. eff
. (%
)
0
2
4
6
8
10
12
TA
A c
onc.
(m
g be
nzid
ine/
l)
COD rem. eff. Color rem. eff. TAA conc.
Fig. 3. Effect of glucose-COD on color, COD removals, and TAA
concentrations in UASB reactor (mean ± SD, n = 3).
This indicates that the dye decolorization occurs only
with biodegradation.
The COD removal efficiencies decreased from 78 to
68% as the glucose-COD concentrations decreased from
500 to 100 mg/l. This could be attributed to a low co-
substrate concentration which was a limiting factor sincethe methanogens have very low growth rates, and their
metabolism is usually considered rate limiting in the
anaerobic processes (Bras et al., 2001; Beydilli et al.,
1998; Lun et al., 1995). When the co-substrate was not
a limiting factor (3000 mg/l of glucose-COD), high
COD removal efficiencies (88%) were obtained (I+ıkand Sponza, 2002). In Run 5 without co-substrate oper-
ation low COD removal efficiencies (28%) could bemainly attributed to the inert COD originated from
the CR dye together with the anaerobic extracellular
products as reported by Germirli et al. (1991). Regres-
sion analysis indicated that the linear relationship be-
tween glucose-COD and COD removal efficiency was
not statistically significant (r2 = 0.69, P = 0.15 and
F = 5.14).
As shown in Fig. 4, the methane production rates de-creased from 445 to 125 ml/day while the glucose COD
in the feed was reduced to 100 from 500 mg/l. The meth-
ane production was recorded as 73 ml/day at substrate-
free operation. The COD dependence of methane gas
production was statistically significant (r2 = 0.99,
P = 0.004 and F = 236).
The concentrations of TAA varied between 8.2 and 9
mg benzidine/l in the effluent samples at glucose-CODconcentrations of 100 and 500 mg/l, respectively, in the
feed. No significant differences in TAA concentrations
(7.6 mg benzidine/l) were obtained through decoloriza-
tion of CR dye while the UASB reactor was fed without
co-substrate. Regression analysis indicated that there
was no linear relationship between TAA and COD con-
centrations (r2 = 0.03, P = 0.82 and F = 0.03). TAA and
COD removal efficiencies were observed to be 43% and28%, respectively, in substrate-free operation (see Fig.
4). The CR dye was degraded to aromatic amines
0
20
40
60
80
100
500 250 100 0Glucose-COD concentrations (mg/l)
TA
A r
emov
al e
ffic
ienc
y (%
)
0
100
200
300
400
500
600
Met
hane
pro
duct
ion
rate
(m
l/day
)
TAA Methane production rate
Fig. 4. Effect of glucose-COD on the TAA removal efficiencies and
methane production rates in UASB reactor (mean ± SD, n = 3).
638 M. Is�ık, D.T. Sponza / Bioresource Technology 96 (2005) 633–643
through the cleavage of dye chromophore. The color re-
moval was provided by the electrons produced from the
TAAs, which can be used as carbon and energy sources
through anaerobic decolorization. These findings agree
with the studies performed by Razo-Flores et al.
(1997). They indicated that the reducing environmentwas provided by the electrons released from the aro-
matic amines through the cleavage of azo dye azodisal-
icylate (ADS). This dye was used as carbon and energy
sources by the methanogens in the UASB reactor feed-
ing with glucose-free wastewater. As aforementioned,
the aromatic amines provides electrons supporting the
reduction of the dye resulting in complete decolorization
even during substrate-free operation. The TAA removalefficiencies were 40% and 38% at COD concentrations
500 and 100 mg/l, respectively, indicating the partial
degradation of benzidine under anaerobic conditions
(see Fig. 4). Supporting these findings, Kalyuzhnyi
et al. (2000) and Razo-Flores et al. (1997) also reported
that the aromatic amines could be further mineralized
under anaerobic conditions in an UASB reactor. O�Con-
nor and Young (1993) found that some aromatic amines(2-aminophenol, 4-aminophenol) can be mineralized by
methanogenic consortia. However, Brown and Labour-
eur (1983), Haug et al. (1991) and Bras et al. (2001)
showed that aromatic amines are not degraded anaero-
bically in co-substrate containing and co-substrate-free
reactors.
3.2. Differences in absorbance spectra in influent and
effluent samples at various co-substrate levels
The absorbance spectra at varying wavelengths in the
samples taken from the UASB reactor effluent in the dif-
ferent runs and in the feed containing 100 mg/l of CR
dye with mineral medium is illustrated in Fig. 5. These
spectra exhibited marked alterations as exemplified in
the aforementioned figure. These changes may be ex-plained by structural modifications of the CR dye mol-
0.00
0.05
0.10
0.15
0.20
0.25
0.30
0.35
0.40
300 400 500 600
Wavelengt
Abs
orba
nce
of e
fflu
ents
3000 mg/l Glucose-COD and dye-free
glucose-COD free and 100 mg/l CR
Fig. 5. Visible absorbance spectra of CR in influent and efflue
ecule due to azo bond reduction under anaerobic
conditions. The absorbance peak in the influent (feed)
samples of the UASB reactor was obtained at a wave-
length of 497 nm while the absorbency peak was at a
wavelength of 340 nm in the effluents of UASB reactor
through Runs 2 and 5. In other words, the absorbancepeak at 497 nm completely disappeared after the feed
decolorized in the UASB reactor. This suggests that a
transformation of CR dye to dye-intermediates had
taken place under anaerobic conditions. The absorbance
levels in the effluent samples of the UASB reactor were
higher (0.37) compared to that of dye-free samples
(0.26). This shows that the aromatic amines produced
through the cleavage of the azo bond were not ulti-mately degraded. The partial degradation of TAAs in
the anaerobic UASB reactor resulted in higher absorb-
encies in the effluent samples containing dye compared
to dye free samples as reported by Rajaguru et al.
(2000). This suggests a biological mineralization of
amines. Since the the aromatic amine removal efficien-
cies varied between 25% and 42% (see Fig. 4) it can be
conluded that most of the aromatic amines released weremetabolized in UASB reactor. Razo-Flores et al. (1997),
Rajaguru et al. (2000), Kalyuzhnyi et al. (2000) and I+ıkand Sponza (2002) demonstrated the degradability of
azo dye metabolites under anaerobic conditions.
3.3. Effect of NaHCO3 alkalinity on reactor performance
Traditionally, the total alkalinity in an anaerobic di-gester includes all the bicarbonate alkalinity and
approximately 80% of the volatile fatty acids. Because
only bicarbonate alkalinity is usable for neutralizing vol-
atile acids, total alkalinity does not always represent the
available buffering capacity in a digester (Anderson and
Yang, 1992). A stable anaerobic treatment system re-
quires a balance among all organisms. The maintenance
of this balance is normally indicated by a low VFA con-centration and a stable pH. When the anaerobic system
700 800 900
h (nm)
0.00.30.50.81.01.31.51.82.02.32.52.8
Abs
orba
nce
of f
eed
500 mg/l glucose-COD and 100 mg/l CR
Feed
nts of reactor at different glucose-COD concentrations.
0
500
1000
1500
2000
2500
3000
3000 1500 750 550 250
NaHCO3 concentration (mg/l)
Met
hane
pro
duct
ion
rate
(m
l/day
)
0
2
4
6
8
10
12
14
16
18
20
TV
FA
/B.A
lk *
10 a
nd p
H
Methane production VFA/B.Alk*10 pH
Fig. 7. Effect of decreasing concentrations of NaHCO3 alkalinity on
TVFA/Alk ratio, methane productions and pH levels (mean ± SD,
n = 3).
M. Is�ık, D.T. Sponza / Bioresource Technology 96 (2005) 633–643 639
is in balance, the methanogens could be inactivated by
unfavorable environmental conditions, e.g., pH decrease
due to insufficient alkalinity, accumulation of VFA and
toxicity of intermetabolites (Kuai et al., 1998).
The effect of bicarbonate alkalinity on the COD, col-
or removal efficiencies and VFA productions in theUASB system is illustrated in Fig. 6. Color was effec-
tively removed (greater than 99%) at all NaHCO3 con-
centrations. No significant differences in COD removal
efficiencies (E = 84–88%) were obtained while the alka-
linity decreased to 550 from 3000 mg/l. The NaHCO3
was not a significant parameter influencing the COD re-
moval. Regression analysis indicated that the linear rela-
tionship was not statistically significant (r2 = 0.14, P =0.52 and F = 0.51). However, the COD removals de-
creased to 68% while the alkalinity decreased to 250
mg/l in the feed since the bicarbonate alkalinity could
not buffer the pH in the UASB reactor. Similarly, the
VFA concentrations increased from 81 to 403 mg/l as
the NaHCO3 concentrations in feed decreased to 250
mg/l from 3000 mg/l. No statistically significant linear
relationship was observed between NaHCO3 and VFAconcentrations (r2 = 0.26, P = 0.37 and F = 1.07).
If a UASB reactor is stable, the TVFA/B.Alk ratio of
reactor effluent is lower than 0.4 (Behling et al., 1997).
The TVFA/B.Alk. ratios were between 0.03 and 0.28
as the NaHCO3 concentrations varied between 550
and 300 mg/l in the feed (see Fig. 7). However, this ratio
increased to 0.68 when the NaHCO3 concentrations de-
creased to 250 mg/l in the feed, indicating the instabilityof the UASB reactor. The regression analysis indicated
that the linear relationship between NaHCO3 concentra-
tion and the TVFA/B.Alk. ratio was not statistically sig-
nificant (r2 = 0.32, P = 0.31 and F = 1.43). The pH
ranges decreased from 7.3 to 6.2 as the bicarbonate alka-
linity decreased from 3000 to 250 mg/l (see Fig. 7). The
regression analysis indicated that the linear relationship
between pH and B.Alk. was statistically significant(r2 = 0.74, P = 0.06 and F = 8.66). The well being of a
0102030405060708090
100
3000 1500 750 550 250
NaHCO3 concentration (mg/l)
CO
D a
nd c
olor
rem
oval
eff
. (%
)
0
100
200
300
400
500
TV
FA
(m
g C
H3C
OO
H)
COD removal eff. Color removal eff. TVFA
Fig. 6. Effect of decreasing concentrations of NaHCO3 alkalinity on
COD, color removal efficiencies, and VFA concentrations in UASB
reactor (mean ± SD, n = 3).
UASB reactor can often be judged by the effluent pH.
The pH in a normally functioning reactor is close to
neutral and controlled by a system that consist of bicar-
bonate and volatile fatty acids. Although acidogenic
phase was not preferred as dominant phase, the color
could be removed together with acidogenic and metha-
nogenic consortia in the buffered anaerobic reactors
(Behling et al., 1997; Chinwetkitvanich et al., 2000).TAAs released through the cleavage of CR dye and
corresponding removal efficiencies and recoveries are
shown in Fig. 8. The TAA concentrations increased to
10.1 from 5.83 mg/l as the bicarbonate concentration
in influent was increased from 250 to 3000 mg/l. CR
breakdown products such as benzidine were greatly re-
moved and the corresponding aromatic amine was
recovered by less than 44%. This indicates that the aro-matic amines released from azo dye cleavage was being
partly metabolized.
As shown in Fig. 8, no significant difference in meth-
ane production rates (2000–2400 mg/l) was observed as
the NaHCO3 concentrations decreased from 3000 to
0
2
4
6
8
10
12
3000 1500 750 550 250
NaHCO3 concentration (mg/l)
TA
A c
once
ntra
tion
(
mg
benz
idin
e/l)
0
20
40
60
80
100
TA
A r
ecov
erie
s (%
)
TAA concentration TAA recoveries
Fig. 8. Effect of decreasing concentrations of NaHCO3 alkalinity on
the TAA productions and recoveries (mean ± SD, n = 3).
Table 2
Variations of alkalinity in influent and effluent samples in UASB reactor (mean values, n = 3)
NaHCO3 mg/l in feed As mgCaCO3/l equivalent Effluent bicarbonate alkalinity (mg CaCO3/l) Effluent total alkalinity (mg CaCO3/l)
3000 dye-free 1786 1857 1870
3000 1786 2387 2406
1500 893 1510 1556
750 446 1127 1133
550 327 682 720
250 149 378 595
640 M. Is�ık, D.T. Sponza / Bioresource Technology 96 (2005) 633–643
550 mg/l in the feed. The regression analysis indicated
that the linear relationship between methane production
and NaHCO3 concentrations was not statistically signif-
icant (r2 = 0.24, P = 0.46 and F = 2.66). Since 550 mg/l
of NaHCO3 concentration (as 327 mg CaCO3/l) pro-
vided the optimum buffering capacity to effectively con-
vert glucose-COD to methane and VFA, it could be
suggested as the optimum bicarbonate alkalinity con-centration in influent to maintain the pH above 6.6. This
value corresponds to an Alk./COD ratio of 0.163 for
optimum operation, which is lower than the ratios pro-
posed by Gonzalez et al. (1998) (0.4), Souza et al. (1992)
and Moosbruger et al. (1993) (0.5), and Speece (1996),
(1.2). The total and bicarbonate alkalinity levels in the
influent and effluent samples of the UASB reactor are
summarized in Table 2. Increases in alkalinity concen-tration was observed in the effluent of the UASB reactor
compared to influent samples. For example when 1500
mg/l NaHCO3 (93 mg CaCO3/l) was added to the feed
the bicarbonate and the total alkalinity concentrations
were 1510 and 1556 mg CaCO3/l, respectively. This
shows that the alkalinity was regenerated in the upper
layer (effluent) of UASB reactor. The VFA will be con-
sumed more rapidly than they can be produced by themethanogens and subsequently the VFA were converted
to methane. The VFA alkalinity will be regenerated to
bicarbonate alkalinity. Furthermore, since the NH4-N
concentration (NH4Cl = 400 mg/l) in the feed is high,
00.1
0.2
0.30.40.50.6
0.70.80.9
300 400 500 600
Wavelength
Abs
orba
nce
of e
fflu
ents
3000 mg/l NaHCO3-CR free1500 mg/l NaHCO3-100 mg/l CR250 mg/l NaHCO3-100 mg/l CR
Fig. 9. Visible absorbance spectra of CR in influent and effl
the generated ammonium bicarbonate alkalinity con-
tributes to total alkalinity in effluent samples. However,
since the reserve alkalinity is the bicarbonate alkalinity
that maintains the pH above 6.2–6.6 in the UASB reac-
tor, it can be assumed that this alkalinity was provided
in our system since the lowest pH level recorded was 6.3.
3.4. Differences in absorbance spectra in influent and
effluent samples at various bicarbonate alkalinity levels
The change of UV–visible spectra of the influent
(feed) and effluent samples containing different bicarbo-
nate alkalinity levels in the UASB reactor are displayed
in Fig. 9. The feed spectrum has a peak in the visible re-
gion at 560 nm, which accounts for the color and a peak
at 390 nm. In the anaerobic treated samples containingdifferent bicarbonate alkalinity levels and in the dye-free
samples, there is a peak in the visible region at spectra
380 nm. The absorbance peak at 560 nm completely dis-
appeared after anaerobic decolorization. This indicates
that CR dye was degraded and exhibited maxima
absorbance peak at 380 nm. Similarly the absorbance
spectrum in the dye-free samples was obtained at a
wavelength of 380–390 nm. Although different bicarbo-nate alkalinity and dye containing UASB reactor efflu-
ent samples exhibited similar maxima spectra the
observed absorbance levels was found to be quite differ-
ent. For example the absorbance levels increased to 0.88
700 800 900
(nm)
0.00.30.50.81.01.31.51.82.02.32.52.8
Abs
orba
nce
of f
eed
3000 mg/l NaHCO3-100 mg/l CR550 mg/l NaHCO3-100 mg/l CRFeed, 100 mg/l CR
uents of reactor at different NaHCO3 concentrations.
M. Is�ık, D.T. Sponza / Bioresource Technology 96 (2005) 633–643 641
from 0.4 as the bicarbonate alkalinity levels increased to
1500 mg/l from 250 mg/l in the effluent of UASB reactor
with the exception of 3000 mg/l of bicarbonate alkalinity
containing effluents (A = 0.65). This findings demon-
strated that the alkalinity could change the dominated
acidogen and methanogen bacteria resulting in modifi-cations in chemical structure of the intermetabolite
products released through cleavage of CR. The modifi-
cations in the bacterial genus and species from methano-
genic and acidogenic bacteria probably affect their
metabolic pathway through degradation of CR dye
resulting in different metabolic products.
4. General discussion
The findings of this study showed that the decoloriza-
tion of CR dye in an UASB reactor did not correlate
with COD concentrations, COD removal or bicarbonate
alkalinity. It seems that the reducing environment pre-
vailing in the UASB reactor by the electrons releasing
from the CR dye and its inter-metabolite products,mainly from the aromatic amines provides the color
removal.
Although some studies showed that the addition of
an external carbon source to the anaerobic reactor im-
proved the color removal which helps ascertain a reduc-
ing environment and possible increase in the
concentration of enzyme cofactors (Carliell et al.,
1995; Gingell and Walker, 1971; Banat et al., 1996;Brown and Laboureur, 1983; Chinwetkitvanich et al.,
2000), other studies have indicated that the dyes can
be used as sole carbon source (Chang et al., 2001; Chung
and Stevens, 1993; Razo-Flores et al., 1997; Kalyuzhnyi
et al., 2000; Blumel, 1998; Coughlin et al., 1997) which
are in agreement with findings in this study.
Several researchers have shown that the aromatic
amines could be decolorized under anaerobic conditions(Razo-Flores et al., 1996, 1997; Battersby and Wilson,
1989; O�Connor and Young, 1993). The results of this
study showed that partial mineralization of benzidine re-
leased from the cleavage of CR azo dye provides elec-
trons to support continued reduction of CR dye,
supporting the published references given above. How-
ever, these results conflict with the studies performed
by Haug et al. (1991), Brown and Hamburger (1987),and Field et al. (1995) indicating the aromatic amines
could not be degraded further in anaerobic reactors.
In a UASB system treating textile wastewaters con-
taining azo dye, alkalinity is lost through NH4-N con-
version to organic nitrogen, the buffering of H2CO3
acidity and the VFA generation by the granules. The
VFA that was generated increased the acidity allowing
the partial phase separation of acidogenesis from meth-anogenesis in the lower active zone in UASB reactor. In
the upper active zone of the granular bed the VFA con-
verted to methane, regenerating the alkalinity and
increasing the pH.
The textile industry in Turkey is one of the most
important industrial sectors both in terms of its contri-
bution to economy and environmental emissions. There-
fore, besides cost-effective anaerobic treatment process,additional external carbon sources and alkalinity
requirements should be kept in mind for economical
costs. If alkalinity is required to buffer the textile indus-
try wastewaters containing azo dyes, this could ad-
versely affect the economy of the anaerobic treatment
(Speece, 1996). However, this study showed that 100
mg/l of CR azo dye could be used as substrate and a
bicarbonate alkalinity as low as 500 mg/l is enough forsimultaneous decolorization (E = 100%) and COD re-
moval (E = 56).
5. Conclusions
Congo Red could be completely decolorized under
glucose-COD concentrations as low as 100 mg/l. In thecase of co-substrate-free operation, greater than 99%
color removal was achieved. In the COD-free operation
the CR dye was used as a carbon and energy source
by the anaerobic microorganisms. Meanwhile, the azo
bond was cleaved initially, aromatic amines was pro-
duced second and then the color removal could be main-
tained by TAAs which are used as carbon end energy
source. Fifty-eight percent COD, 100% color, 39%TAA removal efficiencies and 380 ml/day methane pro-
duction rates were obtained in 100 mg/l COD concentra-
tion as external co-substrate while 25% COD, 99%
color, 40% TAA removal efficiencies and 320 ml/l meth-
ane production rates were monitored in co-substrate free
operation.
The results of this study showed that the alkalinity
requirement is 0.16 HCO3Alk/g influent COD providing84% COD and 99% color removals at a pH 6.5 through
decolorization of 100 mg/l CR dye with 2000 mg/l glu-
cose-COD in the UASB reactor. Moreover, the methane
production rate was 2000 ml/day and the VFA was neu-
tralized resulting in a VFA concentration of 250 mg/l in
the effluent. Although no significant differences in pH
and methane production rates were observed up to a
NaHCO3 concentration of 550 mg/l in the influent ofUASB reactor, the color was effectively (greater than
99%) removed at all NaHCO3 concentrations.
Acknowledgments
The Turkish Scientific and Technical Research Coun-
cil (TUBITAK) and Turkish Government PlanningInstitution (DPT) and Dokuz Eylul University Research
Foundation funded this project. The authors would like
642 M. Is�ık, D.T. Sponza / Bioresource Technology 96 (2005) 633–643
to thank them for the financial support to project with
grant numbers 199 Y 110 and 0 908 20 0001.
References
An, H., Qian, Y., Gu, X., Tang, W.Z., 1996. Biological treatment of
dye wastewater using anaerobic–oxic system. Chemosphere. 33
(12), 2533–2542.
Anderson, G.K., Yang, G., 1992. Determination of bicarbonate and
total volatile acid concentration in anaerobic digesters using a
simple titration. Water Environ. Res. 64, 53–59.
APHA, 1992. Standard Methods for the Examination of Water and
Wastwater, 18th ed. American Public Health Association, Wash-
ington, DC, USA.
Banat, I.M., Nigam, P., Singh, D., Marchant, R., 1996. Microbia
decolorization of textile-dye containing effluents: a review. Biores.
Technol. 58, 217–227.
Battersby, N.S., Wilson, V., 1989. Survey of the anaerobic bidegra-
dation potential of organic chemicals in digesting sludge. Appl.
Environ. Microbiol. 55, 433–439.
Behling, E., Diaz, A., Colina, G., Herrera, M., Gutierrez, E., Chacin,
E., Fernandez, E., Forster, C.F., 1997. Domestic wastewater
treatment using a UASB reactor. Biores. Technol. 61, 239–245.
Beydilli, M.I., Pavlosathis, S.G., Tincher, W.C., 1998. Decolorization
and toxicity screening of selected reactive azo dyes under metha-
nogenic conditions. Water Sci. Technol. 38, 225–232.
Blumel, S., Contzen, M., Lutz, M., Stolz, A., Knackmuss, H.J., 1998.
Isolation of a bacterial strain with the ability to utilize the
sulfonated azo compound 4-crboxy-40-sulfoazobenzene as the sole
source of carbon and energy. Appl. Environ. Microbiol. 64, 2315–
2323.
Bras, R., Ferra, M.I.A., Pinheiro, H.M., Goncalves, I.C., 2001. Batch
test for assessing decolorization of azo dyes by methanogenic and
mixed cultures. J. Biotechnol. 89, 155–162.
Brown, D., Hamburger, B., 1987. The degradation of dye stuffs. Part
III. Investigations of their ultimate degradability. Chemosphere.
16, 1539–1553.
Brown, D., Laboureur, P., 1983. The degradation of dyestuff. Part I.
Primary degradation under anaerobic conditions. Chemosphere.
12, 397–404.
Carliell, C.M., Barclay, S.J., Naidoo, N., Buckley, C.A., Mulholland,
D.A., Senior, E., 1995. Microbial decolorization of a reactive azo
dye under anaerobic conditions. Water SA 21 (1), 61–69.
Carliell, C.M., Barclay, S.J., Buckley, C.A., 1996. Treatment of
exhausted reactived dyebath effluent using anaerobic digestion:
laboratory and full-scale trials. Water SA 22, 225–233.
Chang, J.-S., Chou, C., Lin, Y.-C., Lin, P.-J., Ho, H.-Y., Hu, T.L.,
2001. Kinetic characteristics of bacterial azo-dye decolorization by
Pseudomonas luteola. Water Research 35, 2841–2850.
Chinwetkitvanich, S., Tuntoolvest, M., Panswad, T., 2000. Anaerobic
decolorization of reactive dyebath effluents by a two stage UASB
system with Tapioca as a co-substrate. Water Res. 34, 2223–
2232.
Chung, K.T., Stevens, S.E., 1993. Degradation of azo dyes by
environmental microorganisms and helminits. Environ. Toxicol.
Chem. 12, 2121–2132.
Cooper, P., 1993. Removing color from dye house wastewaters: a
critical review of technology available. J. Soc. Dyers Col. 109, 97–
100.
Coughlin, M.F., Kinkle, B.K., Tepper, A., Bishop, P.L., 1997.
Characterization of aerobic azo dye degrading bacteria and their
activity in biofilm. Water Sci. Technol. 36 (1), 215–225.
Delee, W., O�Neill, F.H., Pinhero, H.M., 1998. Anaerobic Treatment
of textile effluents: a review. J. Chem. Technol. Biotechnol. 73, 323–
335.
Field, J.A., Stams, A.J.M., Kato, M., Schraa, G., 1995. Enhanced
degradation of aromatic pollutant in coculture of anaerobic and
aerobic bacterial consortia. Anton. Leeuw. 67, 47–77.
Geisberger, A., 1997. Azo dyes and the law––open debate. J. Soc.
Dyers Col. 113, 197–200.
Germirli, B.F., Orhon, D., Artan, N., 1991. Assessment of the initial
inert soluble COD in industrial wastewaters. Water Sci. Technol.
23, 1077–1086.
Gingell, R., Walker, R., 1971. Mechanisms of azo reduction by
Steptococcus feacalis––II The role of soluble flavins. Xenobiota 1,
231–239.
Gonzalez, J.S., Rivera, A., Borja, R., Sanchez, E., 1998. Influence of
organic volumetric loading rate, nutrient balance and alkalinity/
COD ratio on the anaerobic sludge granulation of an UASB
reactor treating sugar cane molasses. Int. Biodet. Biodeg. 41, 127–
131.
Haug, W., Schmidt, A., Nortemann, B., Hempel, D.C., Stolz, A., 1991.
Mineralization of the sulfonated azo dye Mordant Yellow 3 by a 6-
aminonaphthalene-2-sulfonate-degrading bacterial consortium.
Appl. Environ. Microbiol. 57, 3144–3149.
I+ık, M., Sponza, D.T., 2002. Aromatic amine degradation in an
UASB/CSTR sequential system treating Congo Red dye. In:
Proceedings of Fifth Specialised International Conference on Small
Water and Wastewater Treatment Systems, 24–26 September
Istanbul, vol. 1, pp. 461–470.
I+ık, M., Sponza, D.T., 2003. Effect of different oxygen conditions on
decolorization of azo dyes by Escherichia coli, Pseudomonas sp. and
fate of aromatic amines. Process Biochem. 38, 1183–1192.
Kalyuzhnyi, S., Sklyar, V., Mosolova, T., Kucherenko, I., Russkova,
J.A., Degtyaryova, N., 2000. Methanogenic biodegradation of
aromatic amines. Water Sci. Technol. 42, 363–370.
Knackmuss, H.J., 1996. Basic knowledge and perspectives of bioelim-
ination of xenobiotic compounds. J. Biotechnol. 51, 287–295.
Kuai, L., De Vreese, I., Vandevivere, P., 1998. GAC-amended UASB
reactor for the stable treatment of toxic textile wastewater.
Environ. Technol. 19, 1111–1117.
Lun, S.-Y., Wu, J., Chien, J., 1995. Anaerobic waste treatment:
efficient separation of the acid and methane forming phases using
two UASB reactors. Process Biochem. 30, 523–529.
Manu, B., Chaudhari, S., 2002. Anaerobic decolorization of simulated
textile wastewater containing azo dyes. Biores. Technol. 82, 225–
231.
Manu, B., Chaudhari, S., 2003. Decolorization of indigo and azo dyes
in semicontinuous reactors with long hydraulic retention time.
Process Biochemistry Technology 38, 1213–1221.
Moosbruger, R.E., Wentzel, M.C., Ekama, G.A., Marais, G., 1993.
Lauterntun waste in UASB systems, feasibility, alkalinity require-
ments and pH control. Water SA. 19, 41–52.
O�Connor, O.A., Young, L.Y., 1993. Effect of nitrogen limitation on
the biodegradability and toxicity of nitro and aminophenol isomers
to methanogenesis. Arch. Environ. Contam. Toxicol. 25, 285–
291.
O�Neill, C., Lopez, A., Esteves, S., Hawkes, F.R., Hawkes, D.L.,
Wilcox, S.J., 2000a. Azo-dye degradation in an anaerobic–aerobic
treatment system operating on simulated textile effluents. Appl.
Microbiol. Biotechnol. 53, 249–254.
O�Neill, C., Hawkes, F.R., Hawkes, D.L., Esteves, S., Wilcox, S.J.,
2000b. Anaerobic–aerobic biotreatment of simulated textile effluent
containing varied ratios of starch and azo dye. Water Res. 34,
2355–2361.
Oren, A., Gurevich, P., Henis, Y., 1991. Reduction of nitro substituted
aromatic compound by the Eubacteria Haloanaerobium praevalens
and Sporohalobacter marismortiu. Appl. Environ. Microbiol. 57,
3367–3370.
Pielesz, A., 1999. The process of the reduction of azo dyes used in
dyeing textiles on the basis of infrared spectroscopy analysis. J.
Molecu. Struct. 511–512, 337–344.
M. Is�ık, D.T. Sponza / Bioresource Technology 96 (2005) 633–643 643
Rajaguru, P., Kalaiselvi, M., Palanivel, M., Subburam, V., 2000.
Biodegradation of azo dyes in a sequential anaerobic–aerobic
system. Appl. Microbiol. Biotechnol. 54, 268–273.
Razo-Flores, E., Svitelskaya, A., Donlon, B., Field, J., Lettinga, G.,
1996. The effect of granular sludge source on the anaerobic
biodegradability of aromatic compounds. Biores. Technol. 56, 215–
220.
Razo-Flores, E., Luijten, M., Donlon, B.A., Lettinga, G., Field, J.A.,
1997. Complete biodegradation azo dye Azodisalicylate under
anaerobic conditions. Environ. Sci. Technol. 31, 2098–2103.
Souza, M.E., Fuzaro, G., Polegato, A.R., 1992. Thermophilic anaer-
obic digestion of vinasse in plot plant UASB reactor. Water Sci.
Technol. 25, 213–222.
Speece, R.E., 1996. Anaerobic Biotechnology for Industrial Wastewa-
ters, first ed. Archae Press, Nashville, TN, USA, p. 115, 184.
Sponza, D.T., I+ık, M., 2002. Decolorization and azo dye degradation
by anaerobic/aerobic sequential process. Enzyme Microb. Technol.
31, 102–110.
Weber, A.J., Wolfe, N.L., 1987. Kinetic studies of the reduction of
aromatic azo compounds in anaerobic sediments/water systems.
Environ. Toxicol. Chem. 6, 911.
Wuhrmann, K., Mechsner, K., Kappeler, T., 1980. Investigation on
rate-determining factors in the microbial reduction of azo dyes.
Eur. J. Appl. Microbiol. 9, 325.
Yoo, E.S., Libra, J., Wiesman, U., 2000. Reduction of azo dyes by
desulfovibrio desulfiricans. Water Sci. Technol. 14, 15–22.
Zaoyan, Y., Ke, S., Guangliang, S., Fan, Y., Jinshan, D., Huanian,
M., 1992. Anaerobic–aerobic treatment of a dye wastewater by
combination of RBC with activated sludge. Water Sci. Technol. 26,
2093–2096.