Waste Management 22 (2002) 491–499

www.elsevier.com/locate/wasman

Electrocoagulation of some reactive dyes: a statistical

investigation of some electrochemical variables
Ahmet Gürsesa,*, Mehmet Yalçina, Cetin Doğarb
a
Atatürk University, K.K. Eğitim Fakültesi, Department of Chemistry, 25240 Erzurum, Turkey
b
Atatürk University, Erzincan Eğitim Fakültesi, Department of Chemistry, 24030 Erzincan, Turkey

Received 9 October 2000; accepted 20 November 2001

Abstract
In this study, the decolorization of strong colored solutions containing the reactive textile dyes by electrocoagulation was inves-
tigated. The experiments were statistically designed and carried out according to a 24 full factorial design with two replicate and
four center points. This design was extended with eight additional axial points. Then, the non-linear regression was applied on the
data by using MINITAB software. The variables chosen for this work are mixing rate, cell voltage, electrolysis time and current
density. Results showed that the effective variables on decolorization process are cell voltage, electrolysis time and current density.
In addition, to determine the effect of dye and electrode type on decolorization, the additional experiments were carried out. It was
determined the dye and electrode type are important on the decolorization process. # 2002 Elsevier Science Ltd. All rights reserved.

1. Introduction and their various combinations (Zoltek and Melear,

The textile industry is one of those industries that
consume considerable amounts of water in the manu-
facturing process. The water employed in the dyeing
and finishing processes eventually ends up as waste-
water that needs to be treated before final discharge.
Frequent changes of dyestuff employed in the dyeing
process cause considerable variation in the wastewater
characteristics, particularly pH, color and wastewater
COD value. Large pH fluctuation is especially trouble-
some because the pH tolerance of conventional biological
and chemical treatment systems is very limited. Hence
without proper pH adjustment, normal operation of the
treatment processes is essentially impossible.
Strong color is another important component of the
textile wastewater that is very difficult to deal with
combination of strong color and high dissolved solid
content resulting in high turbidity of the wastewater.
The strong color of wastewater if not removed would
cause disturbance to the ecological system of the
receiving waters. Traditional methods for dealing with
the wastewater consist of biological, physical, chemical
* Corresponding author. Tel.: +90-442-231-4007; fax: +90-442-
218-4172.
E-mail address: ahmetgu@yahoo.com (A. Gürses).
0956-053X/02/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved.
PII: S0956-053X(02)00015-6
1981; Peters et al., 1985; Rebhun and Lurie, 1993; Do
and Chen, 1994; Lin and Peng, 1994, 1996; Pols and
Harmsen, 1994; Szpyrkowicz et al., 1995; Jiang and
Graham, 1996; Morass et al., 1999). Although the cost
of biological treatment of wastewater is less than that of
other methods, the smaller efficiency of decolorization
using this treatment is due to the toxicity of the dis-
solved or suspended matter. The coloring matter of
dyeing wastewater can be effectively removed by
method of chemical coagulation, adsorption, ultra-
filtration and ozonization (Thebault et al., 1981; Ben-
schoten and Edzwald, 1990; Churchley, 1994; Gahr et
al., 1994). In recent years, electrocoagulation has been
successfully tested to treat various industrial wastewater
(Vik et al., 1984; Do and Chen, 1994; Lin and Peng,
1994, 1996; Pouet and Grasmick, 1995; Tsai et al., 1997;
Lin et al., 1998).
Electrocoagulation is a process consisting of creating
a floc of metallic hydroxides within the effluent to be
cleaned, by electrodissolution of soluble anodes. Com-
pared with traditional flocculation and coagulation,
electrocoagulation has in theory, the advantage of
removing the smallest colloidal particles: the smallest
charged particles have a greater probability of being
coagulated because of the electric field that sets them in
motion. It has also the advantage of producing a relatively
492 A. Gürses et al. / Waste Management 22 (2002) 491–499

2. Materials and methods

Nomenclature
2.1. Experimental
d.a.c. (cm) distance between anode and
cathode In the experiments, as electrodes, iron and aluminum
X1, X2, X3 and X4 original values for electrolysis plates 4.0
4.0 cm dimensions and cylindrical graphite
time, cell voltage, current density (8.05 cm2) were used. They were treated with the solu-
and mixing rate, respectively tion of HCl (15%wt.) for cleaning prior to use. The
x1, x2, x3 and x4 coded values for electrolysis experimental apparatus is given in Fig. 1. The solutions
time, cell voltage, current density of 25 and 12.5 ppm of textile dyes of Remazol Brilliant
and mixing rate, respectively Orange 3R (C.I.:17757), Remazol Brilliant Yellow GL,
S.D. standard deviation Remazol Red RB and Remazol Blue 3R New were pre-
pared and then their calibration curves were obtained.
At all experiments, except for the experiments in which
the effect of dye and electrode types were investigated,
low amount of sludge (Pouet and Grasmick, 1995). In Remazol Red RB and Fe electrode were used. Stirring
addition, it has been reported that the adsorption of the was done by using a magnetic stirrer. Electrolyte (NaCI)
hydroxide on the mineral surface was 100 times more on concentration was kept constant at 8.5 mM at the all
‘‘in situ’’ than on pre-precipitated hydroxides when runs. Cell voltage and current were measured as digital.
metal hydroxides were used as coagulants (Rebhun and At the experiments in which cell voltages were kept
Lurie, 1993). Secondary pollution may be caused by constant, it was observed a small variation in steady
chemical substance added at a high concentration when current. At the end of experiments, the treated solution
chemical coagulation is applied to treat dyeing waste- was filtered and then 5 mL filtrate was centrifuged at
water. Excessively added coagulants can be avoided by 2000 rpm for 5 min. The concentration measurements of
electrocoagulation, due to the generation of the coagu- the dyes before and after electrolysis were carried out
lants by electrooxidation of a sacrificial anode. The spectrophotometrically. In the measurements, since the
characteristics of electrocoagulation are simple equip- steady current density increased by shortening the dis-
ment and easy operation, brief reactive retention period, tance between the electrodes at a given voltage (Güden
decreased or negligible equipment for adding chemical and Karakaya, 1994; Brillas et al., 1997), current density
and decreased amount of sludge.
Therefore, electrocoagulation has been widely used to
treat waters containing food and protein wastes, oil
wastes, synthetic detergent effluents mine wastes and
heavy metal-containing solutions. This method can also
be used to remove phosphate, for defluoridation of
water, to treat potable water (Vik et al., 1984; Do and
Chen, 1994).
In recent years, although there has been a lot of works
on the wastewater treatment by electrochemical meth-
ods, there are so far no reports related to the statistical
evaluation of experimental results and also no study
that compare the dye and electrode types for the
electrocoagulation process. In this study, our aim was to
investigate the decolorization by electrocoagulation
from the solutions containing some textile dyes and to
determine the influence of the variables such as current
density, cell voltage, mixing rate and electrolysis time on
decolorization process. The experiments was carried out
according to 24 full factorial experimental design with
two replicate and four center points. Statistical calcula-
tions were done by using MINITAB software. More-
over, to order to determine the influence of electrode
and dye type on the decolorization process, additional
experiments were carried out the electrolysis conditions
in which decolorization is high. Finally, the fit of the
experimental data to a statistical model was investigated. Fig. 1. Experimental apparatus.
 A. Gürses et al. / Waste Management 22 (2002) 491–499 493

can be adjusted changing the anode–cathode distance. Table 1

Therefore, at the various anode–cathode distances (2.0, Original and coded variables
2.5 and 3.0 cm) and at a constant cell potential (7.5 V), Main Original Coded variable (x)
electrolysis experiments were performed. The con- variable variable (X)
ductivities were measured by using KARL KOLB D- 2 1 0 1 2
6072 conductometer for the various electrolysis times at Electrolysis time (min) X1 0 5 10 15 20
the given anode–cathode distances. The zeta potentials Cell voltage (V) X2 5 7.5 10 12.5 15
of the particles formed after each electrolysis period Current density (mAcm2) X3 1 2 3 4 5
were measured by using micro-electrophoresis cell Mixing rate (rpm) X4 200 400 600 800 1000
(Zeta-Meter 3.0+). Decolorization percentage was cal-
culated using the difference between dye concentrations
after and before electrolysis. Table 2
24 Full factorial experiment design with two replicate and four center
2.2. Statistical analysis and modeling points

Decolorization%
The variables chosen for this work are cell voltage,
I Rep. II Rep. Run x1 x2 x3 x4
current density, electrolysis time and mixing rate. From
experimental observations, it was assumed that the 20.16 15.36 1 1 1 1 1
higher order interactions were small relative to the low 34.56 29.76 2 1 1 1 1
24.96 20.16 3 1 1 1 1
order interactions, because a system with several process
46.52 46.52 4 1 1 1 1
variables is conducted primarily by some of the main 32.16 32.16 5 1 1 1 1
effects and low order interactions. Therefore, the pre- 100.00 100.00 6 1 1 1 1
sent work considers only two-way interactions. The 29.76 24.96 7 1 1 1 1
other higher order interactions have been added in the 100.00 100.00 8 1 1 1 1
22.56 22.56 9 1 1 1 1
error. The center points could be used to check the cur-
34.56 34.56 10 1 1 1 1
vature in the model to be fitted (Bennett and Franklin, 29.76 22.56 11 1 1 1 1
1967; Hines and Montgomery, 1990; Strange, 1990). 34.56 39.32 12 1 1 1 1
Before statistical analysis, the main variables were mod- 24.96 24.96 13 1 1 1 1
ified to their variables coded in the following manner: 100.00 100.00 14 1 1 1 1
36.92 24.96 15 1 1 1 1
X1, X2, X3 and X4 are original variables, where X1, X2,
100.00 100.00 16 1 1 1 1
X3 and X4 represent electrolysis time (min), cell voltage 29.76 17 0 0 0 0
(V), current density (mA cm2) and mixing rate (rpm), 36.96 18 0 0 0 0
respectively. Accordingly, x1, x2, x3 and x4 are the coded 34.56 19 0 0 0 0
variables, where 27.36 20 0 0 0 0

x1 ¼ ðX1  10Þ=5
x2 ¼ ðX2  110Þ=2:5
x3 ¼ ðX3  3Þ=1
x4 ¼ ðX4  600Þ=200
The levels for the four main variables were chosen
from our previous experiences. The original and coded
levels are given in Table 1. The full factorial design is
given in Table 2. In order to determine the influence of
the main variables on the decolorization, the data was
statistically analysed by using MINITAB software and
the results are given in Table 3. As can be seen from
Table 3, the P value (0.000 for curvature) clearly shows
existence of curvature. The P values for main effect
(0.000) and two-way interaction (0.000) are very small.
This indicates that at least one of both the main effects
and the interactions is important. The main effect plots
of the four variables are given in Fig. 2. This figure
shows that the exception for x4,as the factor levels are
increased, decolorization percentage increases. Fig. 3
shows the plots of interactions for four main variables.
The significant interactions are indicated by non-paral-
lel lines on the plots. As can be seen from this figure,
there are two important interactions: the first one is
between electrolysis time (x1) and current density (x3)
and the second one is between cell voltage (x2) and cur-
rent density (x3). Since the existence of curvature masks
the importance of the main effects, a reduced model in
which center points are eliminated was used. The var-
iance analysis table of reduced model is given in Table 4.
From Table 4, it can be seen that the main effects and
two-way interactions which are significant at 0.05 level
are x1, x2, x3, x1*x3 and x2*x3, respectively. As can be
seen from Table 3, a curvature exists. Therefore, a
quadratic or higher order model will be convenient to fit
to the obtained data. However, since the data was not
enough for the higher order model, eight axial points
were added to experimental design. The results obtained
494 A. Gürses et al. / Waste Management 22 (2002) 491–499

Table 3
Analysis of variance for decolorization percentage

Source DF Seq SS Adj SS Adj MS Fa P

Main effects 4 24 562.2 24 562.2 6140.55 128.03 0.000

Two-way interactions 6 6342.7 6342.7 1057.11 22.04 0.000
Residual error 25 1199.0 1199.0 47.96
Curvature 1 800.2 800.2 800.20 48.16 0.000
Lack of fit 5 186.3 186.3 37.27 3.3 0.025
Pure error 19 212.5 212.5 11.18
Total 35 32 103.8
a
These values are calculated using residual error.

Fig. 2. Main effect plots for four main variables.

Fig. 3. The interactions plots for main variables.

from additional experiments are given in Table 5. The
final model obtained for Remazol Red RB dye and Fe
electrode is as follows:
Decolorization % ¼ 29:9 þ 22:1x1 þ 2:52x2 þ 17:8x3
þ 14:0x1 x3  1:49x2 x3 þ 4:47ðx1 Þ2
þ 7:61ðx3 Þ2 þ 5:15ðx1 Þ2 ðx3 Þ2
For this equation, s (root MSE) and R2 values (the
multiply coefficient of determination) are 4.592 and
98.2%, respectively.
In addition, to order to determine the influence of
electrode type and dye types on decolorization, the
additional experiments with two replicates were carried
out under the conditions for run 6 given in Table 2. The
obtained data have been statistically analysed. Results
of two-way variance analysis (ANOVA) of these
experiments have been given in Table 6. From this table,
it can be seen that at least one of dyes and electrodes
chosen are important (P=0.000) and also there is
strong interaction between them (P=0.000). this inter-
action are graphically shown in Fig. 4. Because of the
interaction between the variables, the dyes and electro-
des can be compared with each other by performing
 A. Gürses et al. / Waste Management 22 (2002) 491–499 495

Table 4
Full factorial fit for the reduced model (without center points)Estimated effects and coefficients for decolorization percentage

Term Effect Coef Std Coef t-value P

Constant 47.162 0.7126 66.19 0.000

X1 43.209 21.604 0.7126 30.32 0.000
X2 3.284 1.642 0.7126 2.30 0.032
X3 34.531 17.266 0.7126 24.23 0.000
X4 0.294 0.147 0.7126 0.21 0.839
X1*x2 0.889 0.444 0.7126 0.62 0.540
X1*x3 27.936 13.968 0.7126 19.60 0.000
X1*x4 1.489 0.744 0.7126 1.04 0.308
X2*x3 2.989 1.494 0.7126 2.10 0.048
X2*x4 0.294 0.147 0.7126 0.210 0.839
X3*x4 0.611 0.306 0.7126 0.43 0.672

Analysis of variance for decolorization percentage

Source DF Seq SS Adj SS Adj MS F P

Main effects 4 24 562.2 24 562.2 6140.55 377.93 0.000

Two-way interactions 6 6342.7 6342.7 1057.11 65.06 0.000
Residual error 21 341.2 341.2 16.25
Lack of fit 5 186.3 186.3 37.27 3.85 0.018
Pure error 16 154.9 154.9 9.68
Total 31 31 246.0

Table 5 Table 6
Additional runs for decolorization process Two way variance analysis for influence of electrode and dye types

Decolorization% Run x1 x2 x3 x4 Source DF SS MS F P

0.00 21 2 0 0 0 Dye 3 2813.8 937.9 79.26 0.000

95.69 22 2 0 0 0 Electrode 2 22 797.3 11 398.6 963.26 0.000
16.70 23 0 2 0 0 Dye * electrode 6 4385.4 730.9 61.77 0.000
40.80 24 0 2 0 0 Error 12 142.0 11.8
20.76 25 0 0 2 0 Total 23 30 138.5
100.0 26 0 0 2 0
20.80 27 0 0 0 2
32.50 28 0 0 0 2

one-way variance analysis. In other words, in the presence

of interactions, treatment combinations are compared
according to the differences between them (Hines and
Montgomery, 1990; Milton and Arnold, 1990; Menden-
hall and Sincich, 1995). The results of one-way variance
analysis for this comparison are given in Table 7.

3. Results and discussions

Fig. 2 shows the main effect plots of four variables on

the decolorization percentage. From this figure, It can
be seen that the exception for mixing rate, as the levels
of cell voltage,current density and electrolysis time are
increased, the decolorization percentage increases.
As can be seen in Fig. 3, significant interactions are
indicated by non-parallel lines on the plots. Thus, it can
be concluded that there are two important interactions:
the first one is between electrolysis time (x1) and current
density (x3) and the second one is between cell voltage
(x2) and current density (x3). However, as can be seen in
Fig. 4. The lines of interactions for dye and electrode type.
Fig. 3, the latter interaction has relatively a minor effect
on the electrocoagulation. On the other hand, the inter-
action between current density and electrolysis time
could be resulted from the increasing rates of all reac-
tions. In other words, the interaction could be resulted
from shortening of the time required for both coagulation
and electrode dissolution processes. The investigation of
496 A. Gürses et al. / Waste Management 22 (2002) 491–499

Table 7
One way variance analysis for influence of electrode and dye types and individual 95% confine intervals for mean based on pooled standard deviation

Source DF SS MS F P

Cell (E/D) 11 29 996.5 2727.0 230.45 0.000

Error 12 142.0 11.8
Total 23 30 138.5

Level E/D N Mean S.D. - - - - - - - - - - +- - - - - - - - - - + - - - - - - - - - -+- - - - - -

1 Fe/R 2 100.00 0.00 ( -)
2 Fe/O 2 100.00 0.00 ( -)
3 Fe/Y 2 92.50 0.71 (-  -)
4 Fe/B 2 99.00 0.00 (-  -)
5 Al/R 2 100.00 0.00 ( -)
6 Al/O 2 100.00 0.00 ( -)
7 Al/Y 2 51.00 2.83 (-  -)
8 Al/B 2 73.50 2.12 (- )
9 Gr/R 2 19.00 1.41 ( -)
10 Gr/O 2 15.00 9.90 (-  -)
11 Gr/Y 2 6.50 4.95 (-  -)
12 Gr/B 2 62.50 2.12 (-  -)
- - - - - - - - - - +- - - - - - - - - - + - - - - - - - - - -+- - - - - -
Pooled S.D.=3.44 30 60 90

E, electrode; Fe, iron; Al, aluminium; Gr, graphite; D, dye; R, red RB; O, orange 3R; Y, yellow GL; B, blue 3R new.

variation of zeta potentials and conductivities with
electrolysis time could give the more information for the
interaction. The results related to variation of the values
of both the zeta potentials and the conductivities with
electrolysis time for Fe, Al and Graphite electrode are
given in Figs. 5–7, respectively.
Ferric ions electrogenerated may form monomeric
ions, ferric hydroxo complexes with hydroxide ions and
polymeric species, namely, Fe(H2O)+3 6 , Fe(H2O)5
OH+2, Fe(H2O)4(OH)+1 2 , Fe2(H2O)8(OH)2
+4
and Fe2
+2
(H2O)6(OH)4 , depending on pH range (Johnson and
Amirtharajah, 1983). The complexes (i.e. hydrolysis
products) have a pronounced tendency to polymerize at
pH 3.5–7.0 (Pykhteer et al., 1999). When Fe electrode is
used, the amount and the variety of hydrolysis products
formed by anodic dissolution depends significantly on
electrolysis time. Figs. 5a and b show that at the begin-
ning of electrolysis, monomeric hydrolysis products
with low concentration is main products in the relative
acidic medium. Thus, the colloidal particles having zero
and negative zeta potential value are produced by
adsorption of the negatively charged dye molecules on
the monomeric hydrolysis products (Gregory, 1993;
Rebhun and Lurie, 1993; Oxspring et al., 1995). This
causes somewhat to increase in the decolorization per-
centage. However, after the electrolysis of 7.5 min, the
zeta potential values of the colloidal particles increase
negatively with the increasing medium pH up to 7. As
can be seen from Fig. 5a and b, after this time (7.5 min),
the decolorization percentage has been reduced. This is
may be due to the restabilization. On the other hand,
the different trends in Fig. 5b can be attributed to

Fig. 5. The plots for the variation of the decolorization percentages, zeta potential values and conductivities with the electrolysis time at the several
d.a.c. in the case of Fe electrode.
 A. Gürses et al. / Waste Management 22 (2002) 491–499 497

Fig. 6. The plots for the variation of the decolorization percentages, zeta potential values and conductivities with the electrolysis time at the several
d.a.c. in the case of Al electrode.

Fig. 7. The plots for the variation of conductivities with the electrolysis time at the several d.a.c. in the case of graphite electrode (a) with the dye (b)
without the dye.

change in the formation rate of hydrolysis products, as low pH range is explained as precipitation while the
a result of the increasing current density (i.e. as a result higher pH range ( > 6.5) as adsorption (Rebhun and
of the shortening of the anode–cathode distance). Lurie, 1993).
In the experiments in which Al electrode is used, the
electrogenerated aluminium ions may form monomeric Precipitation : DYE þ Monomeric AI !
and polymeric hydroxo complexes, depending on the ½DYE Monomeric AI
ðsÞ pH ¼ 4:05:0
pH range (Johnson and Amirtharajah, 1983). These ð1Þ
complexes have tendency to polymerize in the pH 4–7
ranges. Oversaturation and formation of amorphous DYE þ Polymeric AI !
hydroxide precipitate AI(OH)3(H2O)3s enmeshes colloi-
dal particles in a ‘‘sweep floc’’. Many monomeric and ½DYE PolymericAI
ðsÞ pH ¼ 5:06:0
polymeric species such as AI(OH)+2, AI(OH)+ 2 , AI2 ð2Þ
(OH)+42 , AI(OH) 4, AI6(OH)+3
15 , AI7(OH)+4 17 , AI8 Adsorption : DYE þ AIðOHÞ3 ðsÞ ! ! ½particle

(OH)+4 +7 +5
20 , AI13O4(OH)24 , AI13(OH)34 have been repor-
ted (Rebhun and Lurie, 1993). Formation rates of the ð3Þ
different species play an important role in decolorization
process. Several interaction mechanisms are possible ½DYE PolymericAI
ðsÞ þ AIðOHÞ3 ðsÞ
simultaneously between dye molecules and hydrolysis
! ! ! ½particle

products and the rates of these depend on pH of the

medium and types of ions present. Two major interac- ð4Þ
tion mechanisms are being considered in the recent For the low and middle anode–cathode distance (2.0
years: precipitation and adsorption, each one being and 2.5 cm), during electrolysis of 2.5 min, monomeric
proposed for a separate pH range. Flocculation in the hydrolysis products (AI (OH)+2,AI(OH)+ +4
2 ,AI2(OH)2 )
498 A. Gürses et al. / Waste Management 22 (2002) 491–499
mainly exist in the medium. Under these conditions,
electrostatic charge neutralization between these hydro-
lysis products and hydrolyzed negatively charged dye
molecules occurs [Eqs. (1) and (2)]. As a result of this
process, aggregation tendency of the monomeric hydro-
lysis products increases. As can be seen from Fig. 6a, at
the end of this time (2.5 min), the small amount of the
dye removes by adsorbing on the aggregates. Also, the
results (Table 7) indicate that the formation of coagu-
late depend on the structure of the dyes chosen and so
on the adsorption ability of dye molecules on hydrolysis
products or flocs. The negative zeta potentials values
also support this mechanism. On the other hand, at
longer electrolysis time than 5 min, precipitation of Al
(OH)3 forms because the pH is high (> 7). As can be
seen from Fig. 6b, there are positively charged aggre-
gates at the shorter anode–cathode distance, but the
chargeless aggregates at the middle and high anode–
cathode distance. The positive charge of the aggregates
may be attributed to the adsorption of dye-monomeric
and dye-polymeric colloidal particles on the Al (OH)3
precipitates, because the sweep floc mechanism pre-
dominates when Al(OH)3 begin to precipitate. In the
cases of both Fe and AI electrodes, the decreasing of
conductivity with electrolysis time at relatively higher
current density (Figs. 5c and 6c) indicates the effective-
ness of the mechanism of sweep floc for the certain
electrolysis periods.
In the case of Fe electrode, the conductivities and the
zeta potential values (0.0 mV) of the particles decrease.
The decreasing decolorization percentage, especially
after electrolysis time of 7.5 min, observed at the anode-
cathode distance of 2.0 cm shows the relative adsorption
inability of the dye molecules on the precipitate of
Fe(OH)3 formed in the sweep flocculation. At experi-
ments performed as with dye and without dye by using
graphite electrode, the variation of solution con-
ductivities with electrolysis time exhibites similar trends.
Because the predominant decolorization mechanism in
the case of graphite electrode is electrooxidation, more-
over, the decolorization has not been observed, it may
be said that the main decolorization mechanism in this
process is electrocoagulation.
4. Conclusions
In the present work, the electrocoagulation of the four
reactive dyes has been statistically and systematically
investigated. It was determined the influence of vari-
ables such as cell voltage, current density, mixing rate
and electrolysis time on decolorization. In order to
determine the effect of the important variables, addi-
tional experiments were systematically carried out. As
discussed in the text, the importance of the interaction
both between current density and electrolysis time and
the effect of the dye types on electrocoagulation process
can be explained with the pronounced dependence of
the hydrolysis species to anode-cathode distance and
electrolysis time. This can also be attributed to the dif-
ferent adsorptive properties of dyes. In addition, from
the regression analysis of all 44 runs, a mathematical
equation showing the relation between decolorization
percentage and the effective variables was obtained. The
high R2 square value of 98.2% for this mathematical
model indicates the presence of a high correlation
between observed and predicted values.
References
Bennett, C.A., Franklin, N.L., 1967. Statistical Analysis in Chemistry and
the Chemical Industry, second ed.. John Wiley & Sons, New York.
Brillas, E., Maestro, A., Moratalla, M., 1997. Electrochemical extrac-
tion of oxygen from air via hydroperoxide ion. J. Applied Electro-
chemistry 27, 83–92.
Churchley, J.H., 1994. Removal of dyewaste color from sewage efflu-
ent-the use of a full-scale ozone plant. Wat. Sci. Tech. 30, 275–284.
Do, J.S., Chen, M.L., 1994. Decolorization of dye-containing solutions
by electrocoagulation. J. Applied Electrochemistry 24, 785–790.
Gahr, F., Hermanutz, F., Oppermann, W., 1994. Ozonation-an
important technique to comply with new German laws for textile
wastewater treatment. Wat. Sci. Tech. 30, 255–263.
Gregory, J., 1993. The role of colloid interactions in solid–liquid
separation. Wat. Sci. Tech. 27, 1–17.
Güden, M., Karakaya, I., 1994. Electrolysis of MgCl2 with a top
inserted anode and an Mg–Pb cathode. J. Applied Electrochemistry
24, 791–797.
Hines, W.W., Montgomery, D.C., 1990. Probability and Statistics in
Engineering and Management Science, third ed.. John Wiley &
Sons, New York.
Jiang, J.Q., Graham, J.D., 1996. Enhanced coagulation using Al/
Fe(lll) coagulants: effect of coagulant chemistry on the removal of
color-causing NOM. Environmental Technology 17, 937–950.
Johnson, P.N., Amirtharajah, A., 1983. Ferric chloride and alum as
single and dual coagulants. J. AWWA 5, 232–239.
Lin, S.H., Peng, C.F., 1996. Continuous treatment of textile waste-
water by combined coagulation, electrochemical oxidation and
activated sludge. Wat.Res. 30, 587–592.
Lin, S.H., Peng, C.F., 1994. Treatment of textile wastewater by elec-
trochemical method. Wat. Res. 28, 277–282.
Lin, S.H., Shyu, C.T., Sun, M.C., 1998. Saline wastewater treatment
by electrochemical method. Wat. Res. 32, 1059–1066.
Mendenhall, W., Sincich, T., 1995. Statistics for Engineering and the
Sciences, forth ed.. Prentice Hall, NJ.
Milton, J.S., Arnold, J.C., 1990. Introduction to Probability and Sta-
tistics, second ed.. McGraw-Hill International Editions, Singapore.
Morais, L.C., Freitas, O.M., Gonçalves, E.P., Vasconcelos, L.T.,
Beça, C.G.G., 1999. Reactive dyes removal from wastewaters by
adsorption on eucalyptus bark: variables that define the process.
Wat. Res. 33, 979–988.
Oxspring, D.A., Smyth, W.F., Marchant, R., 1995. Comparison of
reversed-polarity capillary electrophoresis and adsorptive stripping
voltammetry for the detection and determination of reactive textile
dyes. Analyst 120, 1995–2000.
Peters, R.W., Walker, T.J., Eriksen, E., Peterson, J.E., Chang, T.-K.,
Ku, Y., Lee, W.-M., 1985. Wastewater treatment—physical and
chemical methods. Journal WPCF 57, 503–517.
Pols, H.B., Harmsen, G.H., 1994. Industrial wastewater treatment
today and tomorrow. Wat. Sci. Tech. 30, 109–117.
 A. Gürses et al. / Waste Management 22 (2002) 491–499
Pouet, M.F., Grasmick, A., 1995. Urban wastewater treatment by
electrocoagulation and flotation. Wat. Sci.Tech. 31, 275–283.
Pykhteev, O.Y., Efimov, A.A., Moskvin, L.N., 1999. Hydrolysis of
iron (lll)aqua complexes. Russian Journal of Applied Chemistry 72
(1), 9–20.
Rebhun, M., Lurie, M., 1993. Control of organic matter by coagula-
tion and floc separation. Wat. Sci. Tech. 27, 1–20.
Strange, R.S., 1990. Introduction to experiment design for chemists. J.
Chem. Edu. 67, 113–115.
Szpyrkowicz, L., Naumczyk, J., Zilio-Grandi, F., 1995. Electro-
chemical treatment of tannery wastewater using Ti/Pt and Ti/Pt/Ir
electrodes. Wat. Res. 29, 517–524.
499
Thebault, P., Cases, J.M., Fiessinger, F., 1981. Mechanism underlying
the removal of organic micropollutants during flocculation by an
aluminum or iron salt. Wat. Res. 15, 183–189.
Tsai, C.T., Lin, S.T., Shue, Y.C., Su, P.L., 1997. Electrolysis of soluble
organic matter in leachate from landfills. Wat. Res. 31, 3073–3081.
van Benschoten, J.E., Edzwald, J.K., 1990. Chemical aspects of coa-
gulation using aluminum salts II. Coagulation of fulvic acid alum
and polyaluminum chloride. Wat. Res. 24, 1527–1535.
Vik, E.A., Carlson, D.A., Eikum, A.S., Gjessing, E.T., 1984. Electro-
coagulation of potable water. Wat. Res. 18, 1355–1360.
Zoltek, J., Melear, E.L., 1981. Physical and chemical methods. Journal
WPCF 53, 679–687.