Response surface methodology for oxidative degradation of the basic yellow 28 dye by temperature and ferrous ion activated persulfate

Article abstract of DOI:10.14233/ajchem.2013.14941

The decomposition of the persulfate by Fe²⁺ ions results in the production of oxidative sulfate radical anions. In present study, the oxidative degradation of the basic yellow 28 dye solution was studied by using persulfate and ferrous ions. To

Full text of DOI:10.14233/ajchem.2013.14941

Asian Journal of Chemistry; Vol. 25, No. 12 (2013), 6831-6839
http://dx.doi.org/10.14233/ajchem.2013.14941
Response Surface Methodology for Oxidative Degradation of the
Basic Yellow 28 Dye by Temperature and Ferrous Ion Activated Persulfate
1,*
1
2
BELGIN GÖZMEN , ÖZGÜR SÖNMEZ and MERAL TURABIK
1Department of Chemistry, Faculty of Science & Arts, Mersin University, Yenisehir, Mersin 33343, Turkey
2Chemical Prog., Technical Science Vocational School, Mersin University, Yenisehir, Mersin 33343, Turkey
*Corresponding author: Fax: +90 324 3610046; Tel: +90 324 3610001; E-mail: bgozmen@mersin.edu.tr
(Received: 8 December 2012;
Accepted: 5 June 2013)
AJC-13582
The decomposition of the persulfate by Fe2+ ions results in the production of oxidative sulfate radical anions. In present study, the
oxidative degradation of the basic yellow 28 dye solution was studied by using persulfate and ferrous ions. To obtain maximum mineralization
efficiency for aqueous solution of basic yellow 28, Box-Behnken design combined with response surface modeling (RSM). Four independent
variables, namely temperature (40-70 ºC), initial concentration of persulfate (4-12 mM), ferrous ion (1-3 mM) and time (2-8 h) were
transformed to coded values. Subsequently, it was determined that quadratic model to be the most suitable models to estimate the responses.
Analysis of variance (ANOVA) was carried out for quadratic model and was observed the significance of the independent variables and
their interactions. The predicted values of the oxidative degradation efficiency were verified to be in good agreement with the experimental
values (R2 = 0.9902 and R2adj = 0.9804). Maximum mineralization efficiency 93 % was obtained under following conditions: persulfate
concentration 9.87 mM, ferrous ion concentration 1.95 mM, temperature at 65 ºC and time at 8 h for 40 mg/L basic yellow 28 aqueous
solution.
Key Words: Oxidative degradation, Green chemistry, Basic yellow 28, Persulfate, Response surface methodology, Box-Behnken design.
oxidant can be decomposed by heat, UV light or transition
metal ions (Men+), such as Fe2+. Subsequently, this leads to the
formation of SO4•−, which exhibits a redox potential of 2.6 V8,9.
INTRODUCTION
During the dye production and textile manufacturing
processes, large volumes of dyeing wastewater which has the
potential of toxicity and intense colour has been left into aquatic
systems. Even if dye concentrations were low values (e.g.,
10-20 mg L-1), it affects the water transparency and solubility
of gases in water. Besides, the cationic dyes are generally
considered to be more toxic in comparison to the anionic dyes
toward the aquatic biota1. This is because cationic dyes easily
interact with the membrane surfaces of the negatively charged
cells, enters into the cells and concentrates in the cytoplasm2.
Subsequently, the symbiotic process is affected as the natural
equilibrium is disturbed by the reduced photosynthetic activity.
The above mentioned reduction in the photosynthetic activity
is due to the colouration of the water in the streams.
In recent years, there has been growing interest in the use
of persulfate (S2O82-), which is one of the strongest oxidants
known in aqueous solutions3-5. It offers certain advantages over
other oxidants, such as high stability at room temperature, high
aqueous solubility, a solid fact that can be easily stored and
transferred and relatively low cost6,7. Persulfate anion with
a redox potential of 2.01 V is a non-selective oxidant. This
S2O82- + heat/UV → 2SO4•−
(1)
(2)
S2O82- + Men+ → SO4•− + Me(n+1)+ + SO42-
If a ferrous ion is used as the transition metal:
S2O82- + Fe2+ → SO4•− + Fe3+ + SO42- (k = 27 M-1 s-1) (3)
In a persulfate-water system, hydroxyl radicals can also
be formed, as shown in eqns. 4-5. The hydroxyl radicals may
participate in the oxidation of the contaminants10.
SO4•− + H2O → OH• + HSO4– (k = 500 s-1)
SO4•− + OH− → SO42- + OH•
(4)
(5)
Response surface methods (RSM) offer statistical design
of experiment tools that lead to peak process performance.
The RSM-based modeling studies have been widely used in
many wastewater treatment systems such as photocatalytic
oxidation11, Fenton/photo-Fenton12,13, electro-Fenton/photo-
electron-Fenton14,15, electrochemical oxidation16, ozonation17,
wet air oxidation18 and adsorption19,20 for various pollutants.
To our best of knowledge, optimization of oxidative degradation

6832 Gözmen et al.
Asian J. Chem.
process parameters of persulfate/iron/temperature system to
obtain maximum mineralization of basic yellow 28 by using
with the Box-Behnken design has not been study conducted
until now. Many researchers had investigated the alone oxidative
effect of the heat, persulfate dosage and/or metal ions amount
on activated persulfate by temperature or metal ions for various
types of organics21-25. Anipsitakis and Dionysiou5 investigated
the activation of persulfate by 9 different transition metals
including Ag+, Ce3+, Co2+, Fe2+, Fe3+, Mn2+, Ni2+, Ru3+ and V3+
and reported thatAg+ is the best activating agent for persulfate.
Despite this result, in environmental applications iron is often
used due to environmentally friendly nature, cost effectiveness.
However, Liang et al.4 found that Fe2+ used in small incre-
ment In previous studies have shown that an increase in pH3,5,7,25
and ionic strength3,25 resulted in a decrease in the rate constant
of organic pollutants. Xu and Li7 reported that the inhibiting
effects of some ions ranged from low to high in an order of
NO3– < Cl– < H2PO4– < HCO3–. It has been known that in the
neutral and alkaline media, the amounts of soluble Fe2+ and
Fe3+ could decrease due to the formation of complexes and
precipitations. For this reason, the chelating agents were
employed to stabilize the iron in solution at near neutral pH or
pH of the solution was adjusted under pH 4 value. Some of
previous studies have been performed to determine the opti-
mum conditions, but in most of them one of the three variables
(temperature, persulfate concentration, Fe2+ concentration)
were kept constant7,24,26. Also optimum values obtained from
these studies differ from each other.
In the present study, basic yellow 28 (Cationic Gold
Yellow X-GL) which is a synthetic basic cationic dye, imparts
a yellow colour in the aqueous solution, was selected as the
target dye.A series of experiments were conducted at ambient
temperatures (between 40-70 ºC) to evaluate the effectiveness
of the persulfate oxidation that activated by ferrous ion for the
destruction of basic yellow 28. The Box-Behnken design of
Design Expert 8 program was used for the response surface
methodology in the experimental design during the first
series of experiments. The experiments were designed to
determine the influence of factors, such as temperature, initial
ferrous ion and persulfate concentrations and time on the degra-
dation as well as studying the optimum working state. The
second series of experiments were investigated for the oxida-
tive degradation pathway of basic yellow 28 at optimum
conditions.
designs, especially for a large number of variables28. The
independent variables of temperature, processing time, initial
K2S2O8 and initial FeSO4 concentration were coded with the
low and high levels in the Box-Behnken design, as shown in
Table-1. The mineralization per cent (TOC removal%) of the
basic yellow 28 solution was the obtained response. In the
optimization process, the responses can be simply related to
the chosen factors by the linear or quadratic models. A
quadratic model is shown as follows:
k
k
k
Y = β + β x + β x2 +
β x x + e
ij i j i
∑
∑
∑ ∑
0
j
j
jj
j
(6)
j=1
j=1
i
< j=2
where, Y is the response, k is the number of factors, xi and xj
are the coded variables, β0 is the offset term, βj, βjj and βij are
the first-order, quadratic and interaction effects respectively, i
and j are the index numbers for factor and ei is the residual
error29.
Model fitting and graphical analyses were carried out
using the Design-Expert software. Soundness of the fit was
studied using F-test, p-value, chi-square (X2), the root mean
square error of prediction (RMSEP) and the relative error of
prediction (RSEP), correlation coefficient (R2) and adjusted
correlation coefficient (R2adj).
General procedures: Basic yellow 28 (BY28) was
supplied by the DyStar textile firm with the commercial name
Astrazone Goldgelb GL-E. Potassium persulfate (K2S2O8) and
ferrous sulfate (FeSO4·7H2O) were obtained from Acros and
Merck, respectively. The aqueous solutions were prepared in
high purity de-ionized water (resistivity 18.2 mΩ cm), which
was obtained using a Milli-Q water purification system
(Millipore). The basic yellow 28 stock solution was prepared
at 1000 mg/L. The desired concentration (40 mg L-1) of basic
yellow 28 was obtained through the appropriate dilutions of
the stock solution. The K2S2O8 and FeSO4 solutions (100 mM)
were prepared prior to each batch experiment. However, the
FeSO4 solution was prepared in acidic de-ionized water (pH
3.5).
The batch experiments for the decomposition of the basic
yellow 28 solution were carried out in plugged borosilicate
glass bottles. The initial concentration of basic yellow 28 was
prepared as 40 mg L-1 from the stock solution. The pH was
adjusted to 3.5 with 0.5 M H2SO4. The reaction solutions were
obtained by adding the FeSO4 solution and the K2S2O8 solution
into the aqueous basic yellow 28 solutions with a total volume
of 60 mL. The mineralization experiments were conducted in
a thermo-controlled ( 1 ºC) water bath shaker at 120 rpm at
different temperatures (40-70 ºC) in the range of 2 to 8 h. The
experiments were carried out in parallel. At the end of the
reaction time, sodium azide was added to quench any further
oxidation reactions. Subsequently, the aliquots of the samples
were withdrawn from the reactor and filtered through a 0.45
µm membrane filter before the analysis.
EXPERIMENTAL
Response surface modeling: Response surface method-
ology (RSM) uses mathematical and statistical techniques to
fit of a polynominal equation to the experimental data, which
should be the objective of making the statistical previsions27.
Determine the optimum set of operational variables for a process
is the main goal of the RSM. The statistical experimental
designing with RSM can reduce the process variability,
experimentation time as well as the overall cost with an improved
process output28. The Box-Behnken design (BBD) of Design
Expert 8 program was used for the RSM in the experimental
design where each factor takes only three levels. This design
is more efficient and economical according to the other 3k
Detection methods: The total organic carbon (TOC)
contents of the aliquots were analyzed by using the Merck
TOC cell test (1.14878.0001 / 5.0 - 80.0 mg L-1 TOC). At the
end of the reaction time, the sample aliquots were immediately
filtered through a 0.45 µm membrane. The TOC contents of
the cell tests were measured by using the Merck Spectroquant
NOVA 30 model photometer.

Vol. 25, No. 12 (2013)
Response Surface Methodology for Oxidative Degradation of the Basic Yellow 28 Dye 6833
TABLE-1
EXPERIMENTAL RESULTS OF BOX-BEHNKEN DESIGN EXPERIMENTS
Observed TOC
removal (%) (Y)
Predicted TOC
removal (%) (Y)
Run
X1: [S2O82-]0
X2: [Fe2+]0
X3: T (ºC)
X4: t (h)
1
2
3
4
5
6
7
8
4 (-1)
12 (+1)
4 (-1)
12 (+1)
8 (0)
8 (0)
8 (0)
8 (0)
4 (-1)
12 (+1)
4 (-1)
12 (+1)
8 (0)
8 (0)
8 (0)
8 (0)
4 (-1)
12 (+1)
4 (-1)
12 (+1)
8 (0)
8 (0)
8 (0)
8 (0)
8 (0)
8 (0)
8 (0)
8 (0)
8 (0)
1 (-1)
1 (-1)
3 (+1)
3 (+1)
2 (0)
2 (0)
2 (0)
2 (0)
2 (0)
2 (0)
2 (0)
2 (0)
1 (-1)
3 (+1)
1 (-1)
3 (+1)
2 (0)
2 (0)
2 (0)
2 (0)
1 (-1)
3 (+1)
1 (-1)
3 (+1)
2 (0)
2 (0)
2 (0)
2 (0)
2 (0)
55 (0)
55 (0)
55 (0)
55 (0)
40 (-1)
70 (+1)
40 (-1)
70 (+1)
55 (0)
55 (0)
55 (0)
55 (0)
40 (-1)
40 (-1)
70 (+1)
70 (+1)
40 (-1)
40 (-1)
70 (+1)
70 (+1)
55 (0)
55 (0)
55 (0)
55 (0)
55 (0)
55 (0)
55 (0)
55 (0)
55 (0)
5 (0)
5 (0)
5 (0)
5 (0)
2 (-1)
2 (-1)
8 (+1)
8 (+1)
2 (-1)
2 (-1)
8 (+1)
8 (+1)
5 (0)
5 (0)
5 (0)
5 (0)
5 (0)
5 (0)
5 (0)
5 (0)
2 (-1)
2 (-1)
8 (+1)
8 (+1)
5 (0)
5 (0)
5 (0)
5 (0)
5 (0)
26
64
48
43
4
21
65
47
47
3
57
62
91
16
43
65
88
23
28
61
69
22
42
54
81
18
21
72
79
71
75
72
77
76
54
65
91
16
40
68
88
25
27
62
67
22
40
57
82
22
24
70
76
74
74
74
74
74
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
The TOC removal % was calculated according to the equa-
tion given below:
temperature was fixed at 70 ºC for 5 min. Subsequently, the
column temperature was programmed from 70 to 150 ºC at
4 ºC/min and then from 150 to 280 ºC with the 5 ºC/min rate.
The MS detector was operated in the EI mode (70 eV).






TOC0 − TOCt
TOCremoval (%)
=
×100
TOC0
RESULTS AND DISCUSSION
where, TOC0 and TOCt are the initial and final TOC (mg L-1)
content values of the solution before and after the oxidative
treatment, respectevily.
Evaluation of experimental results with design-expert:
A set of four process variables, the operating temperature (40-
70 ºC), the K2S2O8 initial concentration (4-12 mM), the FeSO4
initial concentration (1-3 mM) and processing time (2-8 h)
were identified to investigate their influences on the TOC
removal % of basic yellow 28 at initial pH 3.5. These intervals
were determined as a result of preliminary experiments. When
the dye solutions were mixed with various ion concentrations
of S2O82- and the Fe2+, the dye immediately started to degrade.
The reaction intermediates can occur during the oxidation of
the target compounds. It is possible that some of the interme-
diate products are more toxic and long-lived than starting com-
pounds during decomposition. Therefore, it is important to
monitor the contents of the TOC in various stages of treatment.
Mineralization of the treated basic yellow 28 solutions during
the S2O82-/Fe2+ oxidative treatment was followed by measuring
the total organic carbon (TOC). For the above mentioned four
variables, a set of 29 experiments is required (Table-1). Box-
Behnken design is a spherical, revolving design. It consists of
five replicates central point (Cp) and the edge points are located on
a hypersphere equidistant from Cp. Calculate the number of nece-
ssary experiments number is made according to the following
For the ammonium, nitrate and nitrite analysis, 2 mL
samples were taken from the reaction bottles at the different
time intervals. Next, methanol (0.5 mL) was added to quench
any further oxidation reactions. The DIONEX ICS 3000 Dual
model ion chromatograph equipped with a conductivity detector,
an AS9-HC, 4 mm × 250 mm anion-exchange and CS12A 4
mm × 250 mm cation-exchange columns were used to measure
the inorganic ions. The mobile phase was a solution of: sodium
carbonate (10 mM) and a total flow rate of 1.0 mL/min for the
anion analysis, meta-sulfonic acid (0.02 N) and a total flow
rate of 1.0 mL/ min for the cation analysis.
The GC-MS analysis was performed with the 5890A
Agilent model gas chromatograph, interfaced with the ECD,
NPD and 5975C mass selective detector. The aqueous solutions
were extracted three times with 50 mL dichloromethane. A
3 mL sample was analyzed on GC-MS. The analytical column,
which was connected to the system, was an HP5-MS capillary
column (30 m × 0.25 mm × 0.25 µm). Helium was used as the
carrier gas with a flow rate of 2 mL/min. The GC injection
port temperature was 250 ºC (split mode = 1/5) and the column

6834 Gözmen et al.
Asian J. Chem.
formula: N = 2k (k-1) + Cp, where k is the number of factors
and Cp is the number of the central points. The experimental
results obtained in the Box-Behnken design with the real and coded
values for the four variables studied are shown in Table-1.
In an aqueous medium, the individual and interactive
effects of the setted four variables on the mineralization of the
basic yellow 28 dye in an aqueous medium were investigated
using the Box-Behnken design application. The experimental
results were evaluated with RSM of Design-Expert® 8. The
approximating functions of the TOC removal percent (Y) in
terms of the coded variables obtained are shown as follows:
experimental results, the computed F-value should be greater
than the tabulated F-value at a level of significance α. There-
fore, the calculated F-value (Fmodel = 101.16) was compared
with the tabulated F-value (F0.05, df,(n-(df+1))) at significance level
of 0.05, when the df for model is 14 and n = 29. The Tabular
F-value (F0.05,14,14 = 2.48) is clearly less than calculated F-
value20.
To test the fit between the model responses and the
observed data is performed chi-square (X2) test. This value
was computed by the following equations20:
2
N
(Ymeas,i-Ypred,i
)
Chi-Square=
Y = 74.20 +10.83X + 2.00X +19.33X + 24.83X −10.75X X
+
(9)
∑
1
2
3
4
1 2
Ypred,i
i=1
1.75X X −1.00X X + 0.75X X + 1.00X X − 6.00X X
−
1 3
2
1 4
2
2 3
2
4
3 4
where Ypred,i and Ymeas,i are the model predicted and measured
values of the variable, Y and N is the number of run.
2
2
4
12.14X −17.14X −11.89X − 9.14X
(7)
1
2
3
The calculated chi-square value of the model (X2 = 4.60)
was found to be less than the tabulated value (X20.05 = 41.34).
This result revealed that there is not a significant difference
between the observed data and model response. The quality
of the fit of the polynomial model was also expressed by the
determination of correlation coefficient, R2 and the adjusted
coefficient R2adj. The correlation between the experimental data
and predicted responses is explained by the value of R2. The
Fig. 1 provides Pareto graphic analysis, respectively. These
analyses introduce single or synenergistic positive or inimical
effects of variables on the studied response30. Pareto graphic
analysis based on the following formula gives the percentage
effects of each factor on the response.
Time (h) β4
31.87
obtained values of R2 and R2 were 0.9902 and 0.9804,
Temperature (ºC) β3
19.31
adj
respectively. The adjusted R2 (R2adj) value is more suitable to
compare the models with different numbers of the independent
variables. R squared estimates the proportion of the dependent
variable variance that can be attributed to the predictors, but
unfortunately this statistic exhibits an overestimate bias. The
smaller adjusted R squared attempts to eliminate this bias32.
As shown Fig. 2(a), the correlation coefficient (R2) between
the experimental and model predicted values of the response
variable showed the goodness of fit of the model.A plot of the
normal probability of the residuals is shown in Fig. 2(b). The
residuals from the analysis should be normally distributed.
The above mentioned plot is an important diagnostic tool to
detect and explain the systematic departures from the assum-
ption33. The trend of the residual to a normal distribution, where
the errors are normally distributed and independent of each
other is shown in Fig. 2(b) In addition, the error variance is
homogeneous. Signal to noise ratio is measured by “Adeq
Precision” which value should be grater than 4. The ratio of
35.60 indicates an adequate signal. Fig. 2(c) displays a plot of
residuals versus predicted response that provides a handy
diagnostatic for nonconstant variance. In this case, the pattern
exhibits the hoped-for random scatter, suggesting the variance
of original observations is constant for all values of the
response.
An effective tool for checking lurking variables that may
have influenced the response during the experiment is the
normal plot of the standardized residuals versus the experi-
mental run number as shown in Fig. 2(d). The plot should
show a random scatter around the center line and within the
interval of 3.50. Trends indicate a time-related variable
lurking in the background. Blocking and randomization
provide insurance against trends ruining the analysis34. As a
result, Fig. 2(d) represents that there was no apparent deviation
with the observation order ( 3.00).
Fe(II)*Fe(II) β22
15.18
Per*Per β11
7.62
Temp*Temp β33
7.31
Persulfate (mM) β1
6.06
Time*time β44
4.32
Temp*time β34
Fe(II)*time β24
Fe(II) (mM) β2
1.86
0.05
0.05
Fig. 1. Graphical pareto analysis
βi2


Pi =
×100 (i ≠ 0)




(8)
β 2
∑
i


It is demonstated that the processing time (β4), tempe-
rature (β3) and persulfate ion initial concentration (β1) are the
crucial factors for the mineralization of basic yellow 28. How-
ever, the ferrous ion initial concentration (β2) has a slight
influence on the oxidation.Analysis of variance and regression
takes place in almost all statistical studies. The relationships
between an unlimited number of independent variables and a
response or dependent variable are investigated by regression.
Regression also allows values on one variable to be predicted
from the values recorded on one or more other variables. In
the same way, ANOVA places no restriction on the number of
groups or conditions that may be compared, while factorial
ANOVA allows examination of the influence of two or more
independent variables or factors on a dependent variable with
t-test statistics in the program for eqn. 6. As a result of the
experimental data quadratic model was determined as statis-
tically significant. The quadratic regression model was highly
significant as the F-test (F-value) was found to be 101.16
with a very low probability value (P-value < 0.0001). This
indicated that only 0.01 % of the model was due to noise31. If
the model has a high degree of adequacy to predict of the

Vol. 25, No. 12 (2013)
Response Surface Methodology for Oxidative Degradation of the Basic Yellow 28 Dye 6835
100.00
(a)
(b)
99
80.00
60.00
40.00
20.00
0.00
95
90
80
70
50
30
20
10
5
1
0.00
20.00
40.00
60.00
80.00
100.00
-3.00
-2.00
-1.00
0.00
1.00
2.00
Eperimental value (%)
Standardized residuals
(c)
(d)
3.00
2.00
1.00
0.00
-1.00
-2.00
-3.00
3.00
2.00
1.00
0.00
-1.00
-2.00
-3.00
10
50
9
13
17
21
25
29
0.00
20.00
40.00
60.00
80.00
100.00
Run number
Fitted value (%)
Fig. 2. (a) The actual and predicted plot of TOC removal percent (R2 = 0.9902 and R2adj = 0.9804), (b) the standardized residual and normal
% probability plot of TOC removal per cent, (c) the predicted TOC removal per cent and standardized residuals plot, (d) the standardized
residuals plot for TOC removal efficiency
Effects of variables in the three-dimensional response
surface plots: The contour plot and the corresponding 3D
view for the initial persulfate ion concentration versus the
initial ferrous ion concentration, while holding temperature at
65 ºC and time 8 h is shown in Fig. 3(a). It is well established
that persulfate plays an important role as a source of SO4•−
production in the S2O82-/Fe2+ oxidative treatment. In the present
study, various initial concentrations of S2O82- in the range of
4 and 12 mM were used for the mineralization of the basic
yellow 28 dye in an aqueous solution to obtain its optimal
concentration. The TOC removal per cent values showed fast
increase due to an increase of the initial persulfate concen-
tration to 10 mM, especially Fe2+ concentration in the range
of 1.5-2.25 mM. The TOC removal percentage value reached
96 % at the end of the 8 h experiments, while the persulfate
concentration was 9.7 mM at [Fe2+]0 = 2 mM. However, an
increase of the persulfate concentration beyond 10 mM
resulted in a decrease in the mineralization efficiency. Regard-
ing the effect of iron appears to have been achieved the best
results around 2 mM. If the concentration of Fe(II) is less than
1.5 and higher than 2.5 mM, the TOC removal percents may
decrease.
Temperature is one of the main parameters which influence
the production of SO4•− in the oxidative processes based on
persulfate. In Fig. 3(b), the response surface and contour plots
show that obtained TOC removal (%) values as a function of
the initial persulfate concentration and reaction temperature,
keeping the initial concentration of Fe2+ at 2 mM and time at 8
h. It can clearly be observed that a rise in the temperature is
significantly effective on the removal of basic yellow 28. When

6836 Gözmen et al.
Asian J. Chem.
the effects of the temperature and persulfate concentration on
TOC removal were investigated, the effective temperature and
persulfate concentration intervals were obtained as 55-65 ºC
and 8-10 mM, respectively. The effect of time and concen-
tration of Fe(II) on the TOC removal is shown in Fig. 3(c).
The highest TOC removal data was obtained when the
concentration of Fe(II) was hold around 2 mM. Fig. 3(d) shows
the TOC removal per cent at the mentioned range of time and
temperature, keeping the initial concentration of Fe2+ at 2 mM
and persulfate at 8 mM. An increase in TOC conversion with
both time and temperature was observed. This tendency can
be explained as follows: based on a previously described free-
radical involved reaction mechanism, the concentration of free
radicals increases with temperature and then enhances the
organic compounds degradation. Fig. 3(d) is examined, when
the amount of Fe(II) is kept around 2 mM, the results show
that the temperature should be adjusted over 60 ºC and the
time should be over 6.5 h.
alcohol), at these temperatures degradation of poly(vinyl
alcohol) was completed 120 and 30 min, respectively. But,
TOC removal of poly(vinyl alcohol) in the experiments has
not been followed. Li et al. also investigated the effect of pH,
persulfate concentration, ionic strength, temperature and cata-
lytic Fe3+ and Ag+ on the degradation efficiency of diphenyl-
amine (DPA) by persulfate in batch experiments25. They
observed that Ag+ ion is more efficient than Fe3+ ion and the
increase of either the pH value or ionic strength caused those
inhibitive effects on the degradation of diphenylamine (20 µg
mL-1). Their results showed that the rate of diphenylamine
degradation increased as the initial persulfate concentration
increased in the range of 4.4 to 33 mM. Nfodzo and Choi26
has also investigated degradation of triclosan by using
persulfate/Ag+ ion. They found that an oxidant: metal molar
ratio of 1:1 (stoichiometric amount) was not an optimum for
degradation of this substance. As a result, it is reported that
the reactivity of oxidant/metal systems to be depending on
the target compound. In another study7, an azo dye orange G
(OG) was degraded by the persulfate/ferrous system at pH
3.5. Batch experiments were performed in 50 mL test tubes.
In the above mentioned study it was observed that at [Fe2+] =
1 mM, increasing the dosage of persulfate from 1 to 4 mM
Oh et al. obtained that the rate of poly(vinyl alcohol) (PVA)
(50 mg L-1) oxidation was maximized when the persulfate to
Fe2+ molar ratio was 1:1 at 20 ºC (250 mg L-1 K2S2O8 = 0.92
mM K2S2O8)24.They also observed that increasing the temperature
from 20 to 80 ºC accelerated the oxidation rate of poly(vinyl
(a)
(b)
12.00
12.00
70.00
10.00
1.00
10.00
1.50
62.50
8.00
2.00
8.00
Persulfate (mM)
55.00
Persulfate (mM)
6.00
2.50
6.00
47.50
Fe(II) (mM)
Temperature (ºC)
4.00
3.00
4.00
40.00
(c)
(d)
70.00
62.50
55.00
Temperature (ºC)
8.00
6.50
3.00
8.00
6.50
2.50
5.00
2.00
5.00
47.50
40.00
3.50
Time (h)
3.50
1.50
Fe(II) (mM)
Time (h)
2.00
2.00
1.00
Fig. 3. The effect of (a) initial persulfate and ferrous ion concentrations (T = 65 ºC, t = 8 h), (b) temperature and persulfate concentration
([Fe2+]0 = 2 mM, t = 8 h), (c) time and ferrous ion concentrations (T = 65 ºC, [S2O82-]0 = 8 mM), (d) time and temperature ([Fe2+]0 = 2
mM, [S2O82-]0 = 8 mM) on TOC removal

Vol. 25, No. 12 (2013)
Response Surface Methodology for Oxidative Degradation of the Basic Yellow 28 Dye 6837
([Persulfate]/[Fe2+] = from 1:1 to 4:1) resulted in an increase
orange G degradation. When Fe2+ concentration varied from
0.5 to 4 mM (persulfate concentration is kept constant as 4
mM), the efficiency of orange G degradation was increased
from 54 to 99 % within 0.5 h. In this study, the optimum condi-
tions for orange G degradation was found as persulfate at 4
mM and ferrous ion at 4 mM at pH 3.5 and 20 ºC. However,
with a further increase of the Fe2+ concentration up to 8 mM,
orange G degradation reduced7. The above mentioned tendency
explained that the ferrous ion can also act as a sulfate radical
scavenger at its high concentration, according to eqn. 10.
(i) response and the number of responses being optimized is
n34.
The numerical optimization finds a point that maximizes
the desirability function. The characteristics of a goal may be
altered by adjusting the weight or importance. For several
responses and factors, all goals get combined into one desir-
ability function. The possible goals are: maximize, minimize,
target, within range, none (for responses only) and set to an
exact value (factors only)40. In this study, all factors was in the
range of experimental design value, whereas removal of basic
yellow 28 were maximized. Fig. 4 shows the TOC removal %
values in the end of the maximization study for the basic yellow
28 removal per cent (between 85 and 100 %) with the desir-
ability value of 1 depending on the selected goal for the variables.
Optimal conditions were chosen as persulfate at 9.87 mM and
ferrous ion at 1.95 mM, time at 8 h and temperature at 65 ºC
to avoid spending excess persulfate and ferrous ions for 40
mg L-1 basic yellow 28 at pH 3.5. The optimal parameter values
were validated through persulfate/ferrous oxidation experi-
ments. As a result of confirmatory experiments are performed
by using the optimal conditions and closely results (93 %)
was obtained with the data from optimization analysis using
desirability function.
Fe2+ + SO4•− → Fe3+ + SO42- (k = 3.0 × 108 M-1 s-1) (10)
The optimum conditions obtained in the present study
differs from others may be due to differences both target comp-
ound and experimental conditions. In addition to the previous
studies the effects of four variables alone or interaction on
basic yellow 28 degradation has been examined with the RSM
in the present study. The negative effect of the ferrous ion was
determined over the 2.5 mM Fe(II) value. The sulfate radicals
were also consumed by some other reactions35-37
SO4•− + H2O → OH• + HSO4– (k = 500 s-1)
SO4•− + SO4•− → 2 SO42- (k = 8.9 × 108 M-1 s-1) (11)
:
(4)
SO4•− + OH• → HSO4– + ½ O2
(12)
(13)
TOC removal (%)
3.00
OH• + OH• → H2O2 (k = 5.2 × 109 M-1 s-1)
SO4•− + H2O2 → HO2• + H+ + SO42- (k = 1.2 × 107 M-1 s-1) (14)
SO4•− + S2O82- → S2O8– + SO42- (k = 6.1 × 105 M-1 s-1) (15)
2.50
2.00
1.50
1.00
In work of Gozmen and Turabik, the degradation of basic
yellow 28 at the same concentration was performed by UV/
TiO2/IO4– system and was obtained 85 % of mineralization
after 3 h treatment at pH 5.2 ([IO3–] = 5 mM)38. However, the
obtained mineralization values of this study are lower than in
the previous study. It could be said that persulfate/Fe2+ system
is not very effective on the degradation of basic yellow 28
dye, but it can be thought of as environmentally friendly and
more economical.
It is also established that the TOC removal decreases with
the Fe3+ and/or Fe2+ ions in the solution because of the formation
of Fe ions complexes with carboxylic acids and the degradation
products39. To understand the combined effect of S2O82-/Fe2+,
the experiments were conducted with the use of Fe2+ and S2O82-
alone. However, the TOC removal was not achieved in the
above mentioned experiments.
Optimization of the removal of dye: The experimental
results were optimized by Design-Expert software using the
approximating function of basic yellow 28 removal percent in
eqn. 7. Desirability is an objective function that ranges from
zero outside of the limits to one at the goal. The simultaneous
objective function is a geometric mean of all transformed
responses:
4.00
6.00
8.00
10.00
12.00
Persulfate (mM)
Fig. 4. The maximize TOC removal % values with desirability value of 1
(between 85 and 100 %)
Identification of intermediates: A solution 40 mg L-1 of
basic yellow 28 was treated for 20, 60, 180 and 300 min at
65 ºC. The remaining organics were extracted for the analysis
of GC-MS. The hydroxyl radical and sulfate radical anion are
powerful oxidants, which can degrade the organic molecules
at a faster rate41. The degradation reaction of a dye can be
written as follows:
Dye + OH• → degradation product
(12)
1
1
n
n
Dye + SO4•− → Dye•− + SO42-
(13)



n
D = d
1
×d
2
×…×d
n
=
di
(11)
(
)



∏
The direct degradation of basic yellow 28 using S2O82-
may be ignored, because of no detectable degradation was
observed after 1 h treatment at room temperature. The aromatic
i=1

The di, which ranges from 0 to 1 (the least to most desir-
able, respectively), conceives the desirability of each individual

6838 Gözmen et al.
Asian J. Chem.
SO4-., OH.
N-N
B
A
N+
O
H2N
N
SO4-., OH.
H
=O
H
O
O
N
I
HO
II
III
O
OH
O
O
O
H
N
HO
OH
N
OH
malonic acid
2-pentanone
2-propanone, methylhydrazone
Propanoic acid,
2-methyl-2-hydroxy
Fig. 5. Proposed reaction scheme for the initial degradation of basic yellow 28 by persulfate/ferrous oxidation
oxidation reaction intermediates identified at the early stage
sulfate ions. The evolution of these inorganic ions monitored
by the ion chromatography analyses at 65 ºC during the
treatments is shown in Fig. 6.
2-
(after 20 min for 65 ºC) of the S2O8 /Fe2+ treatment may
indicate that the degradation pathway was initiated by two
possible paths. This was due to the oxidative attack of the
radicals. The probable pathway can be described by Fig. 5. As
the observed structures of I and III, it can be said that the main
attack of the radicals materialized to regions ofA and B, which
contain the double bond. On the one hand, 2-indolinone 1,3,3,-
trimethyl (I) comes from the hydroxyl radical attack on A re-
gion, while on the other hand benzenamine-4-methoxy (II)
comes from the cleavage of the N-N (-C=N-N) bond (B
region). The supinidine, 1-β, 2-β,-epoxy (III) can form from
the hydroxylation of (I). Then, oxidative ring opening reactions
lead to formation of small structures such as 2-pentanone (M+:
86), 2-propanone, methylhydrazone (M+: 86), 2-pentanone
(M+: 86), malonic acid (M+: 104) and propanoic acid, 2-
methyl-2-hydroxy (M+: 104) (detected at before 5 min). After
the attack on the B region by the radicals, the nitrogen element
departed from the aqueous phase as NO3–, NO2– or NH4+. In
the basic yellow 28 dye structure, basic yellow 28 absorbance
value at 438 nm was not observed due to the conjugated
chromophore group is broken in the oxidative degradation
process. This signifies the removal of the dye.
5
4
+
NH+
NH4
4
-
N
N
OO3–
3
3
2
1
0
-
N
N
OO2–
2
1
2
3
4
5
time (h)
Fig. 6. Time course of inorganic ions produced during mineralization of
40 mg/L basic yellow 28 aqueous solution ([S2O82-]0 = 9.87 mM,
([Fe2+]0 = 2 mM, T = 65 ºC)
At the beginning of the oxidation process nitrite ion
concentration is seen to be more than the nitrate. However,
the nitrite ion concentration dropped quickly and reached 0.2
mg L-1 after 5 h. The nitrate ions showed a slow rise during the
5 h up to 1 mg L-1 concentration. The nitrate concentration
Mineralization of basic yellow 28 leads to the conversion
of the nitrogen and sulfur heteroatoms present in the molecule
into inorganic ions, such as nitrate, nitrite, ammonium and

Vol. 25, No. 12 (2013)
Response Surface Methodology for Oxidative Degradation of the Basic Yellow 28 Dye 6839
remained stable at 4.2 mg L-1 after 2 h. Hence, the sulfate ions
formed from sources other than the dye molecule were not
measured.
12. Y. Wu, S. Zhou, F. Qin, X. Ye and K. Zheng, J. Hazard. Mater., 180,
456 (2010).
13. I. Arslan-Alaton, G. Tureli and T. Olmez-Hanci, J. Photochem.
Photobiol. A, 202, 142 (2009).
14. S. Garcia-Segura, L.C. Almeida, N. Bocchi and E. Brillas, J. Hazard.
Mater., 194, 109 (2011).
15. A.R. Khataee, M. Zarei and S.K. Asl, J. Electroanal. Chem., 648, 143
(2010).
16. I. Yahiaoui, F. Aissani-Benissad, F. Fourcade and A. Amrane, Environ.
Prog. Sustain. Energy, 31, 515 (2012).
17. M.J. Lee, Y.S. Kim, C.K. Yoo, J.H. Song and S.J. Hwang, Environ.
Technol., 31, 7 (2010).
18. L. Lei, Q. Dai, M. Zhou and X. Zhang, Chemosphere, 68, 1135 (2007).
19. A. Kumar, B. Prasad and I.M. Mishra, J. Hazard. Mater., 150, 174
(2008).
20. K.P. Singh, S. Gupta, A.K. Singh and S. Sinha, J. Hazard. Mater., 186,
1462 (2011).
21. C.J. Liang. C.J. Bruell, M.C. Marley and K.L. Sperry, Soil Sediment
Contam., 12, 207 (2003).
22. K.C. Huang, Z. Zhao, G.E. Hoag, A. Dahmani and P.A. Block, Chemosphere,
61, 551 (2005).
Conclusion
Degradation and mineralization of 40 mg L-1 of the basic
yellow 28 aqueous solution by the S2O82-/Fe2+ system at pH
3.5 was studied as a function oxidation time, temperature and
initial concentration of Fe2+ and S2O82-. The application of the
Box-Benkhen design combined with the response surface
modeling and optimization helped in attaining the optimal
solution of reaching the maximum basic yellow 28 degradation
in an aqueous solution by the S2O82-/Fe2+ system. The optimum
conditions were satisfied as 8 h treatment at 65 ºC reaction
temperature; 9.87 mM persulfate initial concentration and 1.95
mM ferrous ion concentration; thereby realizing 93 % TOC
removal per cent. It was demonstrated that the basic yellow
28 structure is broken down into two main aromatic interme-
diates early in the process. To reach a TOC removal percent of
93 %, it was observed that the above mentioned aromatic struc-
tures underwent a ring opening as well as the occurrence of
smaller structures, such as organic carboxylic acids.
23. R.H. Waldemer, P.G. Tratnyek, R.L. Johnson and J.T. Nurmi, Environ.
Sci. Technol., 41, 1010 (2007).
24. S.Y. Oh, H.W. Kim, J.M. Park, H.S. Park and C.Yoon, J. Hazard. Mater.,
168, 346 (2009).
25. S.X. Li, D. Wei, N.K. Mak, Z.W. Cai, X.R. Xu, H.B. Li andY. Jiang, J.
Hazard. Mater., 164, 26 (2009).
26. P. Nfodzo and H. Choi, Chem. Eng. J., 174, 629 (2011).
27. B. Jancic-Stojanovic,A. Malenovic, D. Ivanovic, T. Rakic and M. Medenica,
J. Chromatogr. A, 1216, 1263 (2009).
ACKNOWLEDGEMENTS
The authors are grateful to the Mersin University Research
Found (contract no: BAB-FEF KB (BG) 2010-4A) for the
financial support and Mersin University Research and Appli-
cation Center (MEITAM) for providing the facilities at the
GC-MS and ion chromatography analysis.
28. M.A. Bezerra, R.E. Santelli, E.P. Oliveria, L.S. Villar and L.A. Escaleria,
Talanta, 76, 965 (2008).
29. R.H. Myers and D.C. Montgomery, Response Surface Methodology:
Process and Product Optimization using Designed Experiments, John
Wiley & Sons, USA, edn. 2 (2002).
30. S. Hammami, N. Oturan, N. Bellakhal, M. Dachraoui and M.A. Oturan,
J. Electroanal. Chem., 610, 75 (2007).
31. S. Ghasempur, S.F. Torabi, S.O. Ranaei-Siadat, M. Jalali-Heravi, N.
Ghaemi and K. Khajeh, Environ. Sci. Technol., 41, 7073 (2007).
32. H.L. Liu and Y.R. Chiou, Chem. Eng. J., 112, 173 (2005).
33. A.R. Khataee, Environ. Technol., 31, 73 (2010).
34. M.J. Anderson and P.J. Whitcomb, RSM Simplified, Optimizing Pro-
cesses using Response Surface Methods for Design of Experiments,
CRC Press Taylor & Francis Group, New York (2005).
35. W.J. Mcelroy and S.J. Waygood, J. Chem. Soc. Faraday Trans., 86,
2557 (1990).
REFERENCES
1. O.J. Hao, H. Kim and P.C. Chang, Crit. Rev. Environ. Sci. Technol.,
30, 449 (2000).
2. S.F. Li, Bioresour. Technol., 101, 2197 (2010).
3. K.C. Huang, R.A. Couttenye and G.E. Hoag, Chemosphere, 49, 413
(2002).
4. C. Liang, C.J. Bruell, M.C. Marley and K.L. Sperry, Chemosphere,
55, 1213 (2004).
5. G.P. Anipsitakis and D.D. Dionysiou, Appl. Catal. B, 54, 155 (2004).
6. T.K. Lau, W. Chu and N.J.D. Graham, Environ. Sci. Technol., 41, 613
(2007).
7. X.R. Xu and X.Z. Li, Sep. Purif. Technol., 72, 105 (2010).
8. L. Eberson, Electron Transfer Reactions in Organic Chemistry,
Springer-Verlag, Berlin (1987).
36. J.D. Laat and T.G. Le, Environ. Sci. Technol., 39, 1811 (2005).
37. M.S. Tsao and W.K. Wilmarth, J. Phys. Chem., 63, 346 (1959).
38. B. Gozmen, M. Turabik and A. Hesenov, J Hazard. Mater., 164, 1487
(2009).
39. E. Brillas, M.A. Banos and J.A. Garrido, Electrochim. Acta, 48, 1697
(2003).
9. I.M. Kolthoff, A.I. Medalia and H.P. Raaen, J. Am. Chem. Soc., 73,
1733 (1951).
10. C. Liang, Z.S. Wang and C.J. Bruell, Chemosphere, 66, 106 (2007).
11. I.H. Cho and K.D. Zoh, Dyes Pigments, 75, 533 (2007).
40. Design-Expert Software version 8.00 User's Guide (2010).
41. B. Neppolian, H. Jung, H. Choi, J.H. Lee and J.W. Kang, Water Res.,
36, 4699 (2002).