See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/222613523

Substrate removal kinetics in an upflow anaerobic sludge blanket reactor

decolorising simulated textile wastewater

Article  in  PROCESS BIOCHEMISTRY · March 2005

DOI: 10.1016/j.procbio.2004.04.014

CITATIONS READS

111 168

2 authors:

Mustafa Işik Delia Teresa Sponza

Aksaray Üniversitesi Dokuz Eylul University
34 PUBLICATIONS   1,270 CITATIONS    105 PUBLICATIONS   2,988 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

İleri Kimyasal Oksidasyon Yöntemleri ile İlaç Kalıntılarının Giderimi (Aksaray Üniversitesi Proje No: 2013-16) View project

Yemekhane Atıklarından Biyogaz Eldesi (Aksaray Üniversitesi Proje No: 2017-020) View project

All content following this page was uploaded by Mustafa Işik on 24 December 2018.

The user has requested enhancement of the downloaded file.

 Process Biochemistry 40 (2005) 1189–1198

Substrate removal kinetics in an upflow anaerobic sludge

blanket reactor decolorising simulated textile wastewater
Mustafa Işik, Delia Teresa Sponza∗
Environmental Engineering Department, Engineering Faculty, Dokuz Eylul University, Kaynaklar Campus, 35160 Buca-İzmir, Turkey

Received 26 December 2003; received in revised form 31 March 2004; accepted 10 April 2004

Abstract

A simulated wastewater containing sizing agents, azo dyes, salts and other additives was treated using a lab-scale upflow anaerobic sludge
blanket (UASB) reactor at different hydraulic retention times (HRT) in order to obtain the substrate removal kinetic of the reactor through
decolorization of dyes. COD removal efficiencies decreased from 80 to 29.5% when the HRT decreased from 100 to 6 h. The colour removal
efficiencies were between 90 and 95% for HRTs of 100 and 6. Monod, Contois, Grau second order, modified Stover-Kincannon, and first order
kinetic models were applied to determine the substrate removal kinetic of UASB reactor. The experimental data obtained from the steady-state
conditions showed that Grau second order, modified Stover-Kincannon and Contois substrate removal kinetic models were suitable than the
other applied models for predicting the performance of lab-scale UASB reactor treating the azo dyes.
© 2004 Elsevier Ltd. All rights reserved.

Keywords: Simulated textile wastewater; UASB; Grau; Stover-Kincannon; Contois

1. Introduction stitute about 60–70% of dyes produced. Dyes are synthetic

Textile industries produce high quantities of wastewater
with varying composition depending on the wet processes
employed [1]. Textile effluents are complex waste prod-
ucts containing dyes, sizing agents and dyeing auxiliaries.
They are characterized by their deep blue colour and high
concentration of environmental pollutants. Various physic-
ochemical, advanced oxidation processes, biological pro-
cesses and usually a combination of processes are applied
to treat them to meet regulatory discharge limits [2]. In the
past, mainly chemical coagulation followed by extended ac-
tivated sludge processes were used to treat textile wastew-
aters because colour discharge limits were not applied in
studies of wastewater treatment. It is important to note that
the aforementioned treatment processes met the discharge
limits based on the COD of effluent. Currently, through new
regulations, pressure is being placed on companies to reduce
the amount of colour in industrial wastewater in developed
and developing countries. Azo dyes are characterized by
their typical nitrogen to nitrogen (N=N) bonds. They con-
∗ Corresponding author. Tel.: +90 232 4531008/1144;
fax: +90 232 4531153.
E-mail address: delya.sponza@deu.edu.tr (D.T. Sponza).
organics and generally have a complex chemical structure.
They are xenobiotic in biotic environments and hence show
persistence to biodegradation in nature. Apart from the aes-
thetic deterioration of the natural water bodies, dyes also
cause harm to the flora and fauna in the natural environment
[3]. Due to the electron withdrawing nature of the azo dyes,
they are not susceptible to aerobic oxidative catabolism [4]
and hence are not removed in aerobic processes. However,
under anaerobic conditions, in the presence of co-substrate,
decolorization of azo dyes is achieved with cleavage of the
azo bond, thus rendering the azo dye colourless, with forma-
tion of corresponding aromatic amines [5]. Anaerobic treat-
ment is widely employed for treatment of most of the indus-
trial wastewater containing high concentrations of soluble
organic mater. Studies concerning the anaerobic treatabil-
ity of textile wastewaters have been limited [6]. The upflow
anaerobic sludge blanket (UASB) reactor system developed
by Lettinga et al. [7] in the 1970s has received widespread
acceptance and has been successfully used to treat a variety
of biodegradable industrial wastewaters. Although several
researchers have investigated the anaerobic treatability of
real textile wastewaters [6,8,9], the anaerobic decolorization
of simulated textile wastewaters [10–12] and the colour re-
moval efficiencies in synthetic wastewaters containing pure

0032-9592/$ – see front matter © 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.procbio.2004.04.014
1190 M. Işik, D.T. Sponza / Process Biochemistry 40 (2005) 1189–1198

dyes [13–16], none determined the kinetics of simulated tex- Monod and Contois Eqs. (4) and (5) as follows:
tile wastewater in an UASB reactor.
µmax × S
Therefore, the aims of this paper were to determine the µ= (4)
process kinetics of anaerobic digestion and to compare ki- KS + S
netics among the models applied for describing the substrate µmax × S
removal kinetics of UASB reactor treating/decolorizing a µ= (5)
β×X+S
simulated textile wastewater.
If it is assumed that the concentration of biomass in the in-
fluent can be neglected at steady-state conditions (dX/dt = 0
2. Kinetic approaches and −dS/dt = 0) and the HRT(θ H ) is defined as the volume
of the reactor divided by the flow rate of the influent, fol-
Mathematical models are used in fundamental research lowing equations can be obtained by substituting and rear-
of anaerobic processes to examine the hypotheses, to deter- ranging Eqs. (3)–(5) into Eqs. (1) and (2);
mine the importance of relationships between variables to 1
guide the experimental design and to evaluate the experi- µ= + Kd (6)
θC
mental results. These models also used to control and predict
the treatment plant operation performance and to optimize µmax × S 1
the plant design and the results of scale-up pilot studies. = + Kd (7)
KS + S θC
At present, simplified models, involving only few variables,
are easier to monitor and are needed for industrial applica- The kinetic parameters Y, Kd for Monod and Contois model
tions in order to determine the kinetic constants [17]. Monod can be obtained by rearranging equations as shown below:
type kinetic models have been widely used to describe the

(S0 − S) 1 1 1
process kinetics of anaerobic digesters [18]. Although there = × + × Kd (8)
θH × X Y θC Y
has been some success in applying Monod type kinetics to
the anaerobic process, some research workers found that it The value of µmax and KS could be determined by plotting
is difficult to apply this kinetic. For instance, Grady et al. Eq. (9), which was derived by rearranging Eq. (7).
[19] have shown that the effluent substrate concentration,
θC KS 1 1
expressed as COD, was not dependent on the influent sub- = × + (9)
strate concentration entering the reactor when pure or het- 1 + θ C × Kd µmax S µmax
erogeneous cultures were used. In the equation proposed By substituting Eq. (5) instead of the Monod equation into
by Contois and co-workers [20,21], the specific growth rate Eq. (1), Eq. (10) can be obtained:
was considered as a function of growth limiting nutrient
in both influent and effluent substrate concentrations using µmax × S 1
= + Kd (10)
an empirical constant, which was related to biomass lev- β×X+S θC
els.
Similarly, the values of µmax , and β can be obtained by
plotting Eq. (11), which is obtained by rearranging Eq. (10).
2.1. Application of Monod and Contois kinetic
The value of µmax can be calculated from the intercept of
the straight line and finally, β could be obtained from the
For an UASB reactor without biomass recycle, the rate
slope of the line.
of change of biomass and substrate in the system can be
expressed as Eq. (1): θC β X 1
= × + (11)
dX Q Q 1 + θ C × Kd µmax S µmax
= × X0 − × XE + µ × X − K d × X (1)
dt V V
2.2. Grau second-order multicomponent substrate removal
dS Q Q µ×X model
− = × S0 − ×S− (2)
dt V V Y
The general equation of a second-order kinetic model is
The ratio of the total biomass in the reactor to biomass illustrated in Eq. (12) [22,23].
wasted per given time represent the average time called

2
as mean cell-residence time (θ C ) and calculated from the dS S
Eq. (3) for UASB reactor. − = kS × X × (12)
dt S0
V ×X
θC = (3) If Eq. (12) is integrated and then linearilized, Eq. (13) will
Q × XE be obtained:
The relationship between the specific growth rate and the S0 × θ H S0
= θH − (13)
rate limiting substrate concentration can be expressed by the S0 − S kS × X
 M. Işik, D.T. Sponza / Process Biochemistry 40 (2005) 1189–1198 1191

If the second term of the right part of Eq. (13) is accepted and was taken from the methanogenic reactor treating the
as a constant, the Eq. (14) will be obtained, wastewater of Pakmaya Yeast Baker Factory in Izmir. The
S0 × θ H UASB reactor was operated at 37 ◦ C using an electronic
= b × θH + a (14) heater in the medium part of the system. The suspended
S0 − S
solids (SS) and volatile suspended solids (VSS) in the
(S0 − S)/S0 expresses the substrate removal efficiency and is UASB reactor were 31.7 g l−1 and 22.4 gVSS l−1 , respec-
symbolized as E. Therefore, the last equation can be written tively.
as follows:
θH
= a + b × θH (15) 3.2. Simulation of cotton textile wastewater
E
A simulated textile wastewater was prepared to give
2.3. Modified Stover-Kicannon model
the reproducibility for continuous studies. The processes,
generating wastewater in a cotton textile mill are desizing,
In this model the substrate utilization rate is expressed
scouring, bleaching, dyeing, and printing. The simulated
as function of the organic loading rate by monomolecu-
wastewater used in this study exhibited similar composition
lar kinetic for biofilm reactors such as rotating biological
to the synthetic wastewater used by O’Neill and co-workers
contactors and biological filters. Equation of the modified
[25,26]. The composition and characteristics of simulated
Stover/Kincannon model are as follows [24]:
wastewater used in this study are given in Table 1. The
dS Rmax × (Q × S0 /V ) mixture of azo dyes included 50 mg l−1 of Reactive Black
= (16)
dt KB + (Q × S0 /V ) 5, Direct Red 28, Direct Black 38, Direct Brown 2 and
Direct Yellow 12 as 250 mg l−1 of total dye concentration
where dS/dt is defined in Eq. (17):
in simulated wastewater. Azo dyes were hydrolyzed with
dS Q NaOH at 80 ◦ C for 2 h. The size, CMC and starch, were
= × (S0 − S) (17)
dt V hydrolyzed by mixing with 40 g l−1 of NaOH and leaving
Eq. (18) obtained from linearization of Eq. (17) as follows: it at room temperature overnight.
The UASB reactor was initially started to operate after
V KB V 1 one month of adaptation period with simulated wastewater
= + (18)
Q × (S0 − S) Rmax Q × S0 Rmax since the aforementioned reactor had been previously used
to treat the azo dyes and real textile wastewater at an or-
2.4. First order substrate removal model ganic loading rate as high as 25 kg COD m−3 per day. Once
steady-state conditions achieved at each run at different feed
The rate of change in substrate concentration in the system flow-rates and loading conditions, the daily methane pro-
with assuming the first order model for substrate removal duction, COD, pH, total volatile fatty acids of the effluents
could be expressed as follows: were monitored. Steady state was defined by the constant
dS Q Q daily, effluent COD concentrations, effluent volatile fatty
− = × S0 − S − k1 S (19)
dt V V acid (VFA) concentrations and pH values within 10% vari-
ation for consecutive 2–4 days in each reactor.
Under steady-state conditions, the rate of change in substrate
concentration (−dS/dt) is negligible and the equation given
above can be reduced to the Eq. (20): 3.3. Operating conditions
S0 − S
= k1 S (20) A laboratory scale UASB reactor was operated with
θH
simulated textile wastewater during 46 days in order to
The value of k1 can be obtained by plotting (S0 − S/θ H ) ver- investigate the process kinetics at different hydraulic re-
sus S in Eq. (20), which is obtained by rearranging Eq. (19). tention times (HRT) (6–100 h). HRT decreased from 100
The value of k1 can be obtained from the slope of the line. to 6 h by increasing the organic loading rates (OLR) from
1.0 to 15.8 kg m−3 per day in UASB reactor through 46
days of operation period. The F/M ratios were between
3. Material and methods 0.114–1.718 kg COD kg VSS−1 per day through operation
of UASB reactor These F/M ratios, in this study, exhib-
3.1. Experimental lab-scale reactor and seed ited similar operating conditions to those studied by Zhu
et al. [28] (0.75 kg COD kg VSS−1 per day), An et al. [29]
The anaerobic UASB reactor used for treatment of sim- (0.269 kg COD kg VSS−1 per day), and O’Neill et al. [25]
ulated wastewater was 6 cm in diameter, 100 cm in length (0.15 kg COD kg VSS−1 per day) treating the dye manu-
and had an effective volume of 2.5 l. Partially granulated facturing, dye wastes and simulated textile wastewaters,
anaerobic sludge was used as seed in the UASB reactor respectively in UASB reactors.
1192 M. Işik, D.T. Sponza / Process Biochemistry 40 (2005) 1189–1198

Table 1
Characteristics of simulated wastewater
Materials used Concentration (mg l−1 ) The functions of material

Carboxymethyl cellulose (CMC) 150 Sizing

Starch 1500 Sizing
Acetic acid 500 Suitable pH
Azo dyes 250 Dyeing
Reactive black 5 50
Direct red 28 50
Direct black 38 50
Direct brown 2 50
Direct yellow 12 50
NaOH 660 Hydrolysis
H2 SO4 357 pH neutralization
NaCl 500 For fixing
Na2 CO3 1000 For fixing
NaHCO3 2000 pH buffer in UASB reactor
Glucose 2062.5 Substrate for bacterial growth
Vanderbilt mineral medium [27] For bacterial growth
Parameters

COD 4214 ± 241 (n = 6)

Total alkalinity as mg CaCO3 l−1 3090 ± 260 (n = 9)
pH 8.83 ± 0.11 (n = 9)

3.4. Analytical procedures of the anaerobic stabilization of organic matter, result in a

Gas production was measured by a liquid displacement
method [30]. The methane percentage in biogas was deter-
mined by Dräger Pac® Ex methane gas analyzer. Colour in
the influent and effluent samples of the anaerobic reactor was
measured at maximum wavelengths.The soluble COD was
measured colorimetrically using closed reflux methods [31].
Total alkalinity of samples was determined by titrating a
sample with 0.1N of standard sulphuric acid solution to a pH
of 4.3. Bicarbonate alkalinity (B.Alk.) and total volatile fatty
acid (VFA) concentrations were measured simultaneously
by Anderson and Yang [32] titrimetric method. Biomass
were measured as suspended solids and volatile suspended
solids in UASB reactor following Standard Methods [31].
4. Results and discussion
4.1. Reactor performance
In order to obtain kinetics coefficients for different ki-
netic models the UASB reactor was operated with simulated
wastewater containing desizing agents, dye, salts etc. at six
different HRTs through 46 days of the operation period. The
results obtained for steady state conditions during reactor
operation at six different HRTs are summarized in Table 2.
As the HRT was decreased from 100 to 6 h, the pH and B.
Alk. levels decreased from 7.25 and 3425 mg CaCO3 l−1 to
6.71 and 2273 mg l−1 while the VFA concentration increased
from 709 to 2073 mg l−1 . The sensitivity of the methanogens
to pH, coupled with the VFAs, which are the intermediates
negative response by the anaerobic system [33]. Although
the pH remained in the optimal working range for anaerobic
conditions (6.5–7.5) the VFA concentrations were increased
to 2073 mg CH3 COOH l−1 through the operation at an HRT
of 6 h. The optimal pH could be explained by the neutraliza-
tion of hydrogen ions released from the volatile fatty acid
together with the carbonates dissociated from the carbonic
acid with the bicarbonate alkalinity inside the UASB reac-
tor [27]. The VFA/B.Alk. ratio was greater than 0.4 at an
organic loading rate of 7.16 kg m−3 per day, indicating the
moderate instability of the UASB reactor [34]. Total gas pro-
duction, methane gas production, methane percentage and
methane gas yield were 1176 and 647 ml per day, 54.8%
and 0.318 l methane g COD−1 at the beginning of operation
at an HRT of 100 h. When the HRT was decreased to 6 h,
the aforementioned parameters were found to be 6480 and
1209 ml per day, 18.7% and 0.104 l methane g COD−1 , re-
spectively. The optimum organic loading rate and HRT for
optimum pH, VFA/B.Alk. ratio, and methane fraction was
about 3.2 kg m3 per day and 32 h in the UASB reactor, re-
spectively. When the rate of acid formation exceeds the rate
of break down to methane, a process unbalance results with
decreases in pH and CH4 content of the biogas [35]. Eighty
percent COD removal efficiency was obtained at an HRT
of 100 h while the COD removal efficiencies decreased to
29.4 at an HRT of 6 h. Ninety one percent and 90% colour
removal efficiencies were obtained at HRTs of 100 and 6 h
while the decolorization efficiency was 95% at an HRT of
14 h (see Table 2). This shows that colour removal efficien-
cies were not affected directly by these HRTs in the UASB
reactor.
 M. Işik, D.T. Sponza / Process Biochemistry 40 (2005) 1189–1198 1193

Table 2
Experimental data obtained under steady state conditions at six different HRTs
Parameter HRT (h)

100 46 32 14 9 6

Operation period (days) 0–9 10–17 18–27 28–32 33–38 39–46

Sludge retention time 736 339 232 104 63 47
(days)
pH 7.25 ± 0.04 7.20 ± 0.13 7.20 ± 0.02 6.95 ± 0.06 6.82 ± 0.06 6.71 ± 0.06
VFA(mg CH3 COOH 709 ± 106 840 ± 5 797 ± 64 1314 ± 30 1837 ± 190 2073 ± 157
l−1 )
B.Alk. (mg CaCO3 l−1 ) 3425 ± 30 2993 ± 135 2969 ± 136 2348 ± 4.9 2291 ± 100 2273 ± 90
VFA/B.Alk. 0.21 ± 0.02 0.28 ± 0.01 0.27 ± 0.03 0.56 ± 0.01 0.80 ± 0.05 0.91 ± 0.10
Eflluent COD (mg l−1 ) 835 ± 93 1118 ± 63 1108 ± 54 2233 ± 256 2785 ± 230 2975 ± 61
COD removal 79.9 ± 2.4 73.5 ± 1.5 73.7 ± 1.3 47 ± 6.1 33.9 ± 5.4 29.4 ± 1.5
efficiencies (%)
Colour removal 90.6 ± 12 91.1 ± 9 94.2 ± 5 94.3 ± 3 93.8 ± 3 90.3 ± 3
efficiencies (%)
Total gas production rate 1176 ± 154 2000 ± 211 5130 ± 1275 5680 ± 277 6223 ± 1616 6480 ± 635
(ml per day)
Methane gas production 647 ± 107 997 ± 78 2390 ± 692 1973 ± 175 1756 ± 475 1209 ± 164
rate (ml per day)
Methane percentage (%) 54.8 ± 1.8 50.3 ± 2.2 48.7 ± 1.2 35.8 ± 3.18 28.2 ± 1.6 18.7 ± 2.08
Methane yield 0.318 ± 0.060 0.248 ± 0.015 0.404 ± 0.114 0.238 ± 0.052 0.182 ± 0.070 0.104 ± 0.013
(l methane g COD−1
added)

4.2. Monod kinetic model 140.0

120.0
SRT/(1+SRT*Kd)

Six steady state sets of data were used to determine 100.0

the kinetic parameters required for applying all the kinetic
80.0
models. Fig. 1 was plotted from Eq. (8) for determining
the values of Y and Kd for Monod model. Y and Kd values 60.0 y = 98895x + 9.5641
calculated from the intercept and the slope of the straight 40.0 2
R = 0.9671
line as illustrated in Fig. 1 as 0.125 g VSS g COD−1 and 20.0
0.0065 per day, respectively. The value of µmax and KS 0.0
were determined from Fig. 2 using Eq. (9) as 0.105 per day 0 0.0004 0.0008 0.0012 0.0016
and >4000 mg l−1 , respectively. 1/S

4.3. Contois kinetic model Fig. 2. Determination of maximum specific growth rate (µmax ) and half
saturation constant (KS ) for Monod kinetic model.

For the Contois model the values of specific growth rate

(µmax ) and β were determined from Fig. 3 by plotting
Eq. (11) and the kinetic parameters, µmax and β were cal-
culated from the intercept and the slope of the straight line
150

125
0.250 y = 4.4461x + 9.5641
SRT/(1+SRT*Kd)

2
100 R = 0.9671
0.200
(S0-S)/(HRT*X)

75
0.150
50
0.100
y = 7.9974x + 0.0523
2 25
0.050 R = 0.9193
0
0.000
0 5 10 15 20 25 30
0.0000 0.0050 0.0100 0.0150 0.0200 0.0250
X/S
1/SRT (1/day)
Fig. 3. Determination of maximum specific growth rate (µmax ) and kinetic
Fig. 1. Determination of yield coefficient (Y) and death rate constant (Kd ). constant (β) in Contois model.
1194 M. Işik, D.T. Sponza / Process Biochemistry 40 (2005) 1189–1198

6.0 3500
3000 y = 0.6152x + 116.41
5.0 2
R = 0.932
2500
(S0*HRT)/(S0-S)

4.0
(S0-S)/HRT

2000
3.0
1500
2.0 y = 1.0946x + 0.562
2
1000
R = 0.9948
1.0 500
0
0.0
0.0 1.0 2.0 3.0 4.0 5.0 S
HRT (day)
Fig. 6. The model plot of first order substrate removal.
Fig. 4. Determination of kinetic constants (a, b and ks ) for Grau second
order multicomponent substrate removal model. 1/Rmax and a slope of KB /Rmax results on graph. Saturation
value constant (KB ) and maximum utilization rate (Rmax )
as 0.105 per day and 0.465 mg COD mg VSS−1 , respec- were calculated from the line plotted on graph in Fig. 5 as
tively. 8.2 g (l per day) and 7.5 g (l per day), indicating the substrate
removed by microorganisms during time and the maximum
4.4. Grau second-order multicomponent substrate removal substrate removed by the anaerobic organisms versus time,
model respectively.

In order to determine the kinetic coefficients (a, b and
ks ) Eq. (15) was plotted in Fig. 4. The values of a, and b
calculated from the intercept and slope of the straight line
on the graph. The values of a, and b were found to be
0.562 and 1.095 with high correlation coefficients of (R2 )
0.995. The multicomponent substrate removal rate constant
(ks ) was then calculated from the equation a = S0 /(ks. X) as
0.337 per day indicating substrate removal for each unit of
microorganism depending on second order substrate removal
rate constant (ks ).
4.5. Modified Stover-Kicannon model
Fig. 5 shows the graph plotted between reciprocal of total
organic loading removal rate, [V/(Q(S0 −S)], against to the
reciprocal of total organic loading rate, V/(Q×S0 ). Since the
plot of [V/(Q(S0 −S)] versus V/(Q×S0 ) was linear, linear re-
gressions (least squares method) were used to determine the
intercept and the slope. A straight line portion of intercept,
1.4
1.2
1.0
V/Q*(S0-S)
4.6. First order substrate removal model
The value of k1 was obtained from the slope of the line
by plotting (S0 −S)/HRT versus S in Eq. (20). Fig. 6 shows
the plot between (S0 − S0 )/HRT and S k1 was calculated as
0.615 per day with correlation coefficient of 0.932.
4.7. Evaluation of the kinetic models
All kinetic coefficients calculated from models are sum-
marized in Table 3 with correlation coefficients. The ki-
netic data showed that Stover-Kincannon, Grau second
order multicompenant and Contois substrate removal ki-
netics were more appropriate than the other models for
predicting the performance of the lab scale UASB reactor
when the regression coefficients and kinetic coefficients
were compared. Table 4 summarizes the constants deter-
mined from the applicable models in previous studies. In
this study, the determined β Contois kinetic coefficient
(0.465 mg COD mg VSS−1 ) is very similar to the kinetic
coefficient (0.482 mg COD mg VSS−1 ) obtained by Hu et.
al. [37]. However, the µmax value (0.105 per day) obtained
in this study was lower than the µmax value (0.923 per day)
found by Hu et. al. [37]. The most likely reason for the dif-

0.8 ferences in µmax values is the significant variations in reactor

y = 1.0947x + 0.1333
configurations, wastewater characteristics and microorgan-
0.6
2 isms used in the studies. According to the Contois equation
R = 0.9948
0.4 (Eq. (5)), low µmax values decreases the substrate removal
0.2 rate, therefore the substrate removal rates calculated in this
study was lower than the study mentioned above. Similarly
0.0
the saturation constant (KB ) and maximum utilization rate
0.0 0.2 0.4 0.6 0.8 1.0 1.2
V/(Q*S0)
(Rmax ) values obtained in this study are smaller than those
obtained by Yu et al. [24] and Büyükkamacı and Filibeli
Fig. 5. Stover-Kincannon model plot. [36] in a modified Stover-Kicannon model. According to
 M. Işik, D.T. Sponza / Process Biochemistry 40 (2005) 1189–1198 1195

Table 3
Kinetic parameters of UASB reactor treating simulated cotton textile wastewater
Kinetic models Kinetic parameters Values Regression coefficients (R2 )

Monod Y (mg VSS mg COD−1 ) 0.125 0.912

Kd (per day) 0.0065 0.912
kmax (µmax Y−1 ) (per day) 0.84 0.967
µmax (per day) 0.105 0.967
KS (mg l−1 ) >4000 0.967
Contois Y (mg VSS mg COD−1 ) 0.125 0.912
Kd (per day) 0.0065 0.912
µmax (per day) 0.105 0.967
β (mg COD mg VSS−1 ) 0.465 0.967
Grau second order a (per day) 0.562 0.995
b (dimensionless) 1.095 0.995
ks (per day) 0.337 0.995
Modified Stover-Kincannon KB (g (l per day)) 8.211 0.995
Rmax (g COD (l per day)) 7.501 0.995
First order k1 (per day) 0.615 0.932

Grau second order kinetic models, the multicomponent
substrate removal rate constant (ks ) value obtained in this
study was higher than the ks values determined in other
studies. The ks value will be increased as the substrate re-
moval rate increased depending to initial substrate (So ) and
microorganism concentrations (X) in the reactor. The mod-
ified Stover-Kincannon model suggests that the substrate
removal rate is affected by the organic loading rate entering
to the reactor as described in Eq. (16). Kinetic coefficients
of this model (Rmax , KB ) were lower than the data obtained
in other studies as seen in Table 4.
For anaerobic systems a review of common substrate uti-
lization kinetic models was recently carried out by Kuroda
et al., [38], and also by Mata-Alvarez and Cecchi [39].
The Monod model is most widely used for UASB reactors
and for industrial effluents [40]. In this study, however,
it was found that the Monod model was not appropri-
ate for interpreting the kinetic data of the UASB reactor
treating simulated textile wastewater. Although KS values
estimated from the Monod model are very large, kmax and
Y values were acceptable as 0.125 mg VSS mg COD−1
0.84 per day, respectively. kmax , Y, and KS values were
0.77–6.67 mg COD (mg VSS−1 per day), 0.04–0.11 mg
VSS mg COD−1 , and 105–3180 mg COD l−1 , respec-
tively, for anaerobic oxidation of long-chain fatty acids
[41].

Table 4
Comparison of kinetic constants in the Grau second order, modified Stover Kincannon, and Contois models
Models Substrate Reactor Influent COD HRT (day) Kinetic parameters References
type (mg l−1 )
ks a b

Grau second order Municipal wastewater UASB 230–445 0.25–1.0 0.217 0.002 1.346 Ubay [42]
Grau second order Landfill leachate UASB 9000–25000 1.7–2.8 38.5 0.013 1.066 Öztürk et al.
[23]
Grau second order Molasses AHR 2000–15000 0.5–2 10.81 0.033 1.192 Büyükkamacı
and Filibeli [36]
Grau second order Simulated wastewater UASB 4214 0.25–4.16 0.337 0.562 1.095 This study
Rmax KB

Modified-Stover Molasses AHR 2000–15000 0.5–2 83.3 186.23 Büyükkamacı

Kincannon and Filibeli [36]
Modified-Stover Soybean wastewater AF 7520–11450 1–1.45 83.3 85.5 Yu et al. [24]
Kincannon
Modified-Stover Simulated wastewater UASB 4214 0.25–4.16 7.5 8.2 This study
Kincannon
µmax β Y Kd

Contois Ice-cream wastewater CSTR 5500 2.99–7.45 0.930 0.482 0.212 0.0131 Hu et al. [37]
Contois Simulated wastewater UASB 4214 0.25–4.16 0.105 0.465 0.125 0.0065 This study
AF: anaerobic filter AHR: anaerobic hybrid reactor AFBR: anaerobic fluidized-bed reactor; ACR: anaerobic contact reactor UASB: upflow anaerobic
sludge reactor CSTR: continuous stirred tank reactor.
1196 M. Işik, D.T. Sponza / Process Biochemistry 40 (2005) 1189–1198

4000
Monod y = 1.165x - 239.8 R2 = 0.9153 5
6
3500 Contois y = 1.0872x - 167.87 2
R = 0.9701 1
Predicted COD conc. (mg l )

2
-1

3000 Grau y = 0.9924x - 126.28 R2 = 0.9681 3

4
2500

2000

1500

1000
Stover-Kincannon y = 0.891x + 208.41 R2 = 0.9698
500
First Order y = 0.9716x + 811.16 R2 = 0.9105
0
0 500 1000 1500 2000 2500 3000 3500 4000
-1
Measured COD values of effluent (mg l )
Monod (1) Contois (2) Grau (3)
Stover-Kincannon (4) First order (5) Actual (6)

Fig. 7. Comparison of the predicted and the actual COD values from the lab-scale UASB reactor.

The models applied in this study were also evaluated
by comparing the predicted COD values with the ten-
tative values obtained from this continuous operation of
lab-scale UASB reactor. The COD values predicted with
the Stover-Kincannon, Grau, second order and Contois
models gave a high correlation (97%) with actual COD
concentrations measured from the UASB reactor as shown
in Fig. 7. A first order substrate removal model was not
suitable for predicting the COD values compared to the
other investigated models.
In conclusion, based on the kinetic studies, it appears that
the kinetic coefficients obtained from the anaerobic treat-
ment of simulated cotton textile wastewater, agree with the
Stover-Kincannon, second order multicomponent and Con-
tois substrate kinetic models. Although several studies rele-
vant to anaerobic treatment of some industrial wastewaters
were reported in literature none of the studies contained a
kinetic study relevant to anaerobic reactors treating a simu-
lated cotton textile wastewater.
As seen in Table 3, a Grau second order kinetic model
gives the same correlation as the modified Stover-Kincannon
substrate kinetic model. This result is not surprising because
two models are very similar as mentioned below:
Second order linearized equation (Eq. (13)) could be
transformed to the modified Stover-Kincannon model
(Eq. (18)) by dividing So each part the equation.
V V 1 correlation coefficients of 0.995 were found to be more suit-
= +
Q × (S0 − S) Q × S0 kS × X
If Eq. (21) is simulated to the modified Stover-Kincannon
model, the following equation can be obtained:
KB = Rmax
The calculated kinetic parameters from the plots of both
models are appropriate to the values calculated from the
equations given above. As a results of this similarity, the
two models have the same sensitivity in order to obtain the
kinetic coefficients from the plots when the regression co-
efficient of the graphs are compared. Similar results were
observed between Eqs. (4) and (5) in Monod and Con-
tois kinetic models, respectively, where β × X is equal
to Ks .
5. Conclusions
The result of this study showed that simulated wastew-
ater containing sizing agents, azo dyes, salts and other
additives could be treated effectively by UASB reactor at
different HRTs varying between 6 and 100 h. COD removal
efficiencies decreased from 80.2 to 29.4% with decreas-
ing HRT. Colour removals were between 90 and 95% at
HRTs. The concentration of VFAs and VFA/B.Alk. ratio
increased when the HRT decreased from 100 to 6 h. This
show that there was instability between acidogenesis and
methanogenesis into UASB reactor at low HRTs.
In this study the kinetics of UASB reactors treating sim-
ulated textile wastewater was investigated using different
models such as Monod, Contois, Grau second order, modi-
fied Stover-Kincannon, and first order kinetic. Grau second
order and modified Stover-Kincannon kinetic models with
(21) able than the other applied models for predicting the perfor-
mance of lab-scale UASB reactor together with significant
kinetic coefficients. A Contois type kinetic model with a
correlation coefficient of 0.967 was also found to be suitable
(22) for expressing the microbial kinetics of the UASB reactor.
The results of kinetic studies obtained from the lab scale
UASB reactor can be used to predict the treatment perfor-
mance of full-scale UASB reactors if the textile wastewa-
 M. Işik, D.T. Sponza / Process Biochemistry 40 (2005) 1189–1198 1197

ter was treated at similar loading conditions and wastewater [5] Carliell CM, Barclay SJ, Naidoo N, Buckley CA, Mulholland DA,
composition. Senior E. Microbial decolourasition of a reactive azo dye under
anaerobic conditions. Water SA 1995;21(1):61–9.
[6] Kuai L, De Vreese I, Vandevivere P. GAC-amended UASB reactor
for the stable treatment of toxic textile wastewater. Environ Technol
Acknowledgements 1998;19:1111–7.
[7] Lettinga G, Vanvelsen AFM, Hobina SM, Zee W, Klapswik A. Use
of the upflow sludge blanket (USB) reactor concept for biological
This study was supported by the Turkish Scientific and
wastewater treatment. Biotechnol Bioeng 1980;22(4):699–734.
Technical Research Council (TÜBİTAK) for the project [8] Zaoyan YKS, Guangliang S, Fan Y, Jinshan D, Huanian M.
numbered 199 Y 110, and fund of Niğde University with a Anaerobic-aerobic treatment of a dye wastewater by combination of
project numbered 2002/03, Turkey. The authors gratefully RBC with activated sludge. Water Sci Technol 1992;26(9-11):2093–
acknowledge the financial support from these organizations. 6.
[9] Carliell CM, Barclay SJ, Buckley CA. Treatment of exhausted re-
actived dyebath effluent using anaerobic digestion: Laboratory and
full-scale trials. Water SA 1996;22:225–33.
Appendix A. Nomenclature [10] Basibuyuk M, Forster F. The use of sequential anaerobic/aerobic
processes for the biotreatment of a simulated dyeing wastewater.
a S0 (ks ·X)−1 (per day) Environ Technol 1997;18:843–8.
[11] O’Neill C, Hawkes FR, Esteves SRR, Hawkes DL, Wilcox SJ.
b constant for Grau second-order model Anaerobic and aerobic treatment of a simulated textile effluent. J
(dimensionless) Chem Technol Biotechnol 1999;74:993–9.
k1 first order kinetic constant (per day) [12] Manu B, Chaudhari S. Anaerobic decolorization of simulated textile
KB saturation value constant (g (l per day)) wastewater containing azo dyes. Bioresour Technol 2002;82:225–31.
Kd death rate constant (per day) [13] Tan NCG, Prenafeta-Boldu FX, Opsteeg JL, Lettinga G, Field JA.
Biodegradation of azo dyes in cocultures of anaerobic granular sludge
kmax maximum specific substrate utilization
with aerobic aramatic amine degrading enrichment cultures. Appl
rate (mg COD (mg VSS per day)) Microbiol Biotechnol 1999;51:865–71.
KS half saturation concentration (mg l−1 ) [14] Kalyuzhnyi S, Sklyar V. Biomineralisation of azo dyes and their
ks Grau second-order substrate removal breakdown products in anaerobic–aerobic hybrid and UASB reactors.
rate constant (per day) Water Sci Technol 2000;41(12):23–30.
Q inflow rate (l per day) [15] Sponza DT, Işık M. Decolorization and azo dye degradation by anaer-
obic/aerobic sequential process. Enzyme Microb Technol 2002;31(1–
Rmax maximum substrate removal rate 2):102–10.
(mg COD (l per day)) [16] Işık M, Sponza DT. Aromatic amine degradation in an UASB/CSTR
S0 , S substrate concentration in the feed and sequential system treating Congo Red dye. J Environ Sci Health Part
effluent (g COD l−1 or mg COD l−1 ) A 2003;A38(10):2301–15.
X0 , X, XE concentrations of biomass in the feed, [17] Iza J, Colleran E, Paris JM, Wu WM. International workshop on
anaerobic treatment technology for municipal and industrial wastew-
reactor and reactor effluent (g VSS l−1 ) aters: summary paper. Water Sci Technol 1991;24(8):1–16.
V reactor volume (l) [18] Anderson GK, Kasapgil B, Ince O. Microbial kinetics of a membrane
Y yield coefficient (g VSS g COD−1 ) anaerobic reactor system. Environ Technol 1996;17:449–64.
β kinetic parameter (g COD g VSS−1 for [19] Grady CPL, Harlow LJ, Riesing RR. Effects of growth rate and
Contois model) influent substrate concentration on effluent quality from chemostats
containing bacteria in pure and mixed culture. Biotechnol Bioeng
θ C (SRT) mean cell-residence time (day) 1972;14:391–410.
θ Cmin minumum mean cell-residence time (day) [20] Contois DE. Kinetics of bacterial growth: relationship between pop-
θ H (HRT) hydraulic retention time (day) ulation density and space growth rate of continuous cultures. J Gen
µ, µmax specific growth rate and, maximum Microbiol 1959;21:40–50.
specific growth rate, respectively (per day) [21] Chen YR, Hashimoto AG. Substrate utilisation kinetic model for
biological treatment processes. Biotechnol Bioeng 1980;22:2081–95.
[22] Grau P, Dohanyas M, Chudoba J. Kinetic of multicomponent sub-
strate removal by activated sludge. Water Res 1975;9:337–42.
[23] Öztürk I, Altinbas M, Arikan O, Demir A. Anaerobic UASBR treat-
References ment of young landfill leachate. In: Proceedings of the First In-
ternational Workshop on Environmental Quality and Environmental
[1] Delee W, O’Neill FH, Pinhero HM. Anaerobic treatment of textile Engineering in the Middle East Region. Konya, Turkey; 1998.
effluents: a review. J Chem Technol Biotechnol 1998;73:323–35. [24] Yu H, Wilson F, Tay J. Kinetic analysis of an anaerobic filter treating
[2] Banat IM, Nigam P, Marchant R. Singh, Di, Microbial Decoloriza- soybean wastewater. Water Res 1998;32(11):3341–52.
tion of textile-dye containing effluents: a review. Bioresour Technol [25] O’Neill C, Hawkes FR, Hawkes DL, Esteves SRR, Wilcox SJ.
1996;58:217–27. Anaerobic–aerobic biotreatment of simulated textile effluent contain-
[3] Manu B, Chaudhari S. Decolorization of indigo and azo dyes in ing varied ratios of starch and azo dye. Water Res 2000;34(8):2355–
semicontinuous reactors with long hydraulic retention times. Process 61.
Biochem 2003;38:1213–21. [26] Shaw CB, Carliell CM, Wheatley AD. Anaerobic/aerobic treatment
[4] Knackmuss HJ. Basic knowledge and perspectives of bioelimination of coloured textile wastewater using sequencing batch reactor. Water
of xenobiotic compounds. J Biotechnol 1996;51(4):287–95. Res 2002;36:1993–2002.
 1198 M. Işik, D.T. Sponza / Process Biochemistry 40 (2005) 1189–1198
[27] Speece RE. Anaerobic biotechnology for industrial wastewaters.
Nashville, TN: Archae Press; 1996. p.35.
[28] Zhu J, Yang Y, An H, Qian Y. A study of dye waste treatment using
anerobic–aerobic processes. In: Proceedings of 17th International
Symposium on Anaerobic Digestion. Cape Town, South Africa; 1994.
p. 360–63.
[29] An H, Qian Y, Gu X, Tang WZ. Biological treatment of dye wastew-
ater using anaerobic-oxic system. Chemosphere 1996;33(12):2533–
42.
[30] Beydilli MI, Pavlosathis SG, Tincher WC. Decolorization and tox-
icity screening of selected reactive azo dyes under methanogenic
conditions. Water Sci Technol 1998;38(4–5):225–32.
[31] APHA–AWWA. Standard methods for water and wastewater. 17th
ed. American Public Health Assoc/American Water Works Assoc.
Washington DC, USA; 1992.
[32] Anderson GK, Yang G. Determination of bicarbonate and total
volatile acid concentration in anaerobic digesters using a simple
titration. Water Environ Res 1992;64:53–9.
[33] Grady CPL, Daigger GT, Lim HC. Biological wastewater treatment.
2nd ed. New York: Marcelll Dekker; 1999. p. 261,632.
[34] Behling E, Diaz A, Colina G, Herrera M, Gutierrez E, Chacin E, et
al. Domestic wastewater treatment using a UASB reactor. Bioresour
Technol 1997;61:239–45.
View publication stats
[35] Eckenfelder Jr WW. Industrial water pollution control. 2nd ed. Sig-
napore: McGraw-Hill; 1989.
[36] Büyükkamacı N, Filibeli A. Determination of kinetic constants of
an anaerobic hybrid reactor. Process Biochem 2002;38:73–9.
[37] Hu WC, Thayanith K, Forster CF. A kinetic study of the anaerobic
digestion of ice-cream wastewater. Process Biochem 2002;37:965–
71.
[38] Kuroda M, Sakakibara Y, Jin W, Erjie Y, Baozhen W, Quanda
Y. A therotical consideration of the kinetics of multiple substrate
utilization in anaerobic digestion. Water Sci Technol 1993;28(7):77–
80.
[39] Malta-Alvarez J, Cechi F. Kinetic models applied to anaero-
bic biodegradation of complex organic matter. In: Kamely D,
Chakrabardy A, Omenn GS, editors. Biotecnology and biodegrada-
tion advances in applied biotechnology, vol. 4 (Chapter 17). The
Woodlands, TX: Portfolio Pub; 1990. p. 317–44.
[40] Castillo A, Llabres P, Mata-Alvarez J. Kinetic study of a combined
anerobic—aerobic system for treatment of domestic sewage. Water
Res 1999;33(7):1742–7.
[41] Pavlostathis SG, Giraldo-Gomez E. Kinetics of anaerobic treatment.
Water Sci and Technol 1991;24(8):35–59.
[42] Ubay, G. Anaerobic treatment of municipal wastwaters, Ph.D. Thesis,
Istanbul Technical University, Turkey, 1994.