African Journal of Biotechnology Vol. 10(50), pp. 10218-10231, 5 September, 2011
Available online at http://www.academicjournals.org/AJB
DOI: 10.5897/AJB10.2315
ISSN 1684–5315 © 2011 Academic Journals
Full Length Research Paper
Application of response surface methodology in
process parameters optimization for phenol
mineralization using Fenton’s peroxidation
Diya’uddeen Basheer Hasan, A. R. Abdul Aziz and Wan Mohd Ashri Wan Daud*
Chemical Engineering Department, Faculty of Engineering, Universiti Malaya, Malaysia.
Accepted 21 March, 2011
Adverse effect of the highly biorecalcitrant compound phenol to the environment is well established
and its concentrations in industrial effluents vary greatly from 2.8 to 6,800 mg/l depending on the
source. Fenton process effectively mineralises to CO2 and H2O but reported works consumed more
reagents and require longer reaction times. Due to the strong interaction between the several predictor
variables in the Fenton oxidation, response surface methodology was used to optimise the
mineralization treatment. Efficient, faster and economical operating conditions for phenol removal were
explored by investigating four parameters namely the concentration ratio of hydrogen peroxide to
2+
phenol - ((H2O2):(Phenol)), mass ratio of hydrogen peroxide to ferrous ions - ((H2O2):(Fe )), initial
phenol concentration - ([Phenol]o) and reaction time - (tr). The optimal TOC % reduction obtained were
35 and 88% for initial phenol concentrations of 100 and 5 mg/l, respectively. Reaction conditions
corresponding to this phenol mineralization a reaction time of 20 min at ratios of 6 and 15 for
2+
[H2O2]:[Phenol] and [H2O2]:[Fe ], respectively. For mineralisation at 52.5 mg/l phenol concentration, the
2+
optimal conditions were 20 min, ratios 10 and 15 for the reaction time, [H2O2]:[Phenol] and [H2O2]:[Fe ],
respectively. The soluble iron content of the analysed supernatant was found to be below the 15 mg/l
(the maximum limit allowable for total iron discharges required by common regulative subscribed). This
indicated that, the Fenton reagents were utilised during the peroxidation reaction evident from the
2+
almost near consumption of all Fe introduced in about 85% of the samples, thus, negating the need
2+
for immobilising the Fe catalyst or its removal by post treatment. The work proves that the optimized
Fenton process can be potentially used for treatment of any phenol containing wastewater.
Key words: Fenton process, phenol mineralization, response surface methodology.
INTRODUCTION
Environmental pollution control continues to receive great
attention due to negative impact for ecosystems and
humans from industrial effluents. These discharges are
toxic, carcinogenic and with mutagenic properties (Busca
et al., 2008). Presently, 42% of the global wastewater
generation trace their origin to process industries (Doan
*Corresponding author. E-mail: ashri@um.edu.my. Tel: +603
70675297. Fax: +603 70675319.
Abbreviations: TOC, Total organic carbon; COD,
comparatively values of degradation; CCD, central composite
design; OVAT, one variable at a time.
et al., 2009) and one of the components found in organic
wastewater that attracts significant environmental concern is
phenol. It is most abundant (Iurascu et al., 2009) with a
wide range of application as raw material in many
industries whose operations are associated with generation of large amounts of wastewater, for example,
chemical, petrochemical, pharmaceutical, textile and
agricultural industries (Guo and Al-Dahhan, 2005; Yang
et al., 2008). Global production of phenol is high mainly in
an attempt to augment the ever-high increase in industrial
demand for the raw material and recent figures quoted by
Iurascu et al. (2009) indicate the current annual production to stand now at over three million tons thus, resulting
in its presence in many wastewater (Iurascu et al., 2009).
Diya’uddeen et al.
Apart from the toxic effect of phenol to aquatic life, its
presence in water is a source of unpleasant odor, gives
an off-flavors taste in drinking and food processing
waters at concentrations as low as 1 mg/l (Priya et al.,
2008). Phenol’s acute toxicity is well-established (Iurascu
et al., 2009); this made imposition of a more striker
limitation to disposal of phenol-containing effluents. Its
high oxygen demand, estimated at 2.4 mg O2 /mg of
phenol (Priya et al., 2008) depletes the dissolved oxygen
content of the receiving waters bodies thus, directly
interfering with the natural existence of organisms in the
waters. Currently, the Environmental Protection Agency
(EPA) has set less than 1 part per billion (ppb) of phenol
in surface waters, 0.5 µg/l in potable and mineral waters,
while the limits for wastewater discharges are 0.5 mg/l for
surface waters and 1 mg/l for the sewerage system
(Busca et al., 2008).
The most common wastewater treatment is the
biological treatment (Gernjak et al., 2003). However,
phenol being highly stable, toxic, soluble in water,
refractory coupled with its non-biodegradable nature
makes its treatments less effective due to the inhibition
and even elimination of the abundant bacteriological
populations in municipal wastewater treatment (Guo and
Al-Dahhan, 2005; Hong et al., 2001). Above a concentration of 200 mg/l, the process treatment is not attractive
due to the difficulty in treatment and longer time
requirements (Kavitha and Palanivelu 2004). Such high
decomposition times have been reported for example, 40
and 340 h by Prieto et al. (2002) and Gonzalez et al.
(2001), respectively, in their biological treatment of
phenol at a 200 and 1000 mg/l.
The slow rate of reaction in biological systems has
been attributed to the larger space requirement and
higher sludge formation (Wei et al., 2003). Others are
either phase transfer techniques (adsorption), partially
degrade, (biological or expensive) and membrane
technology. The serious adverse effects and stringent
regulations on discharge of phenol containing wastewater
necessitate the need for cost effective and robust
treatment technologies.
A category of destructive approach is the advanced
oxidation processes (AOP's). A technology well established
as an efficient and effective means of mineralizing wide
range of organic pollutants (RH) and waste streams (Kiril
et al., 2010; Gernjak et al., 2003). AOP’s are gene rally
characterised by the generation of the extremely reactive
hydroxyl radical species that act as an oxidant in the
mineralization of complex organic compounds found in
wastewater (Chakinala et al., 2008). The most cost
efficient and easy to set up process AOP that requires
•
minimal energy input for the hydroxyl radical ( OH)
generation is Fenton’s process (Kiril et al., 2010) . Here,
2+
a mixture of Fe or any other ferrous complex and H2O2
referred to as Fenton’s reagent is introduced into the
wastewater at sufficiently low pH, the H2O2 is catalytically
decomposed to produce the hydroxyl radical (Mahiroglu
10219
et al., 2009; Barros et al., 2006) as follows:
2+
Fe
3+
+ H2O2
•
RH + OH
Fe
−
•
+ OH + OH
•
R + H2O
(1)
(2)
It is well known that the reaction pathways involving •OH
radicals and the Fenton reactions are very complex and
that compounds (and degradation by-products) can play
different roles on the overall reaction rates (Li et al.,
2005; Banerjee and Ghoshal, 2005; Cai et al., 2007;
Chen et al., 2008).
In addition, much work using the different AOPs have
been published on treatment of phenol, bulk of the
research either aimed at providing insight to the reaction
pathways/reaction mechanism (Li et al., 2005; Banerjee
and Ghoshal, 2005; Cai et al., 2007; Chen et al., 2008;
Bach et al., 2010; Pontes et al., 2010) or finding optimum
condition of degrading the phenol. On the latter objective,
literature on optimising Fenton’s process for pheno l is
scares and optimised condition reported were either
based on the one variable at a time (OVAT) approach
(Kavitha and Palanivelu, 2004; Kiril et al., 2010), did not
account for the role of the Fenton reagent ratios (Kiril et
al., 2010; Yalfani et al., 2009) or conducted at a fixed
initial phenol concentration (Yalfani et al., 2009; Zazo et
al., 2009). An implication is that reaction would finish
leaving excess reagent, possible total consumption of the
reagents before reaction time finishes or model
developed might not be precise at some points within the
reactor. Moreover, performance of most chemical
reactions in the reactor could be different at different
levels of process variables such as the concentrations
(Salamatinia et al., 2010). Reported composition ranges
of phenol in chemical processes producing phenol differ
from source to source (Ahmadi et al., 2005; Moussavi et
al., 2010; Busca et al., 2008). However, discharges from
phenol-containing effluents in the lower ranges (5 to 100
mg/l) pose a greater concern and more interest is on their
treatment (Yalfani et al., 2009). Against this backdrop, we
conducted mineralisation of aqueous phenol solution at
concentration ranging between 5 to 100 mg/l.
In this current work, focus was to obtain an empirical
equation that would be employed in determining the most
suitable operating conditions at various phenol concentrations with the objective of maximising its mineralization
efficiency and minimising Fenton’s reagent consumpt ion.
An empirical model equation that adequately described
phenol mineralization by Fenton's process was developed. The suitability of the polynomial model was found
to be satisfactory from the ANOVA performed at 95%
confidence level.
MATERIALS AND METHODS
Stock solutions preparation
All the stock solutions used in this Fenton peroxidation experiment
10220
Afr. J. Biotechnol.
were prepared with ultra pure water from ultra pure water system
(arium 611UF, Sartorius Stedim Biotech GmbH) and consumed
immediately after preparation.
The phenol stock solution was prepared by dissolving ultra pure
phenol grade in ultrapure water and working concentrations of 5,
52.5 and 100 mg/l were obtained by dilution of the stock solution. A
stock solution of FeSO4.7H2O (as the source of Fe2+) was prepared
and appropriately diluted to obtain the desired Fe2+ concentration.
The H2O2 (30%, Merck Chemical Corp.) was used without dilution.
Measured pH of the FeSO4.7H2O solution was determined to be 5.7
and all the solutions were prepared and used immediately.
Materials
The reagents used in this study were all analytical grade and used
as received without any purification. From Merck Chemical
Company, FeSO4.7H2O, H2O2 and phenol were purchased and
from Fisher Scientific, NaOH (50% w/w) and H2SO4 (97% purity)
were obtained.
Experimental procedure
All the Fenton reaction was performed batch wise at an ambient
temperature of 25 °C in 250 ml Erlenmeyer flasks co ntaining 100 ml
of the desired phenol concentration as the total reaction mixture
volume. The range of interest in this study was based on an
extensive survey of published works and solution pH was adjusted
to 3 using the H2SO4 (1 M) solution. The Fenton reagents were
dosed under constant magnetic stirring to avoid concentration
gradient. Desired amounts of Fe2+ in the form of FeSO4.7H2O
solution was added first, the reaction was then, initiated by addition
of hydrogen peroxide (H2O2) in varied concentrations that
maintained the designed mass ratios of the Fenton reagent. To
avoid unnecessary increase in temperature during addition of the
H2O2, the transfer was done slowly as observed by Ahmadi et al.
(2005).
At the end of the reaction, pH of the treated samples were
adjusted to 8 with the aqueous solution of NaOH (1 M, Merck) for
further precipitation and simultaneous coagulation of the Fe3+ in a
form of Fe(OH)3 for two hours. The supernatant was then filtered
through 0.45 m Millipore filter and subjected for analysis.
Analysis
The solution of pH was determined with a pH Laboratory meter 827
(Metrohom) equipped with a pH combination electrode.
Mineralization of the samples were followed by measurement of
total organic carbon (TOC) using a combustion/non-dispersive
infrared gas analysis method (NDIR) in a Shimadzu TOC-VCN 5000
analyser (Shimadzu, Japan) equipped with an ASI-V auto sampler
and a Shimadzu TOC 5050 analyser.
For effluents treated with the Fenton method, the maximum limit
is 15 mg/l for the total iron before direct discharge into the
environment as directed by EU’s regulation (Benatti et al., 2006).
Therefore, concentrations of the total soluble iron ions in the
supernatant were determined in inductively coupled plasma optical
emission spectroscopy 7000 DV (ICP-OES, Perkin Elmer).
investigating several process parameters by keeping one variable
constant and others varied. In this method, interactive effects are
not accounted and optimal conditions obtained are not
representative of the whole process condition. Responses from
processes involving several variables cannot be studied without
taking account of the strong interaction between the variables; an
effective approach to such processes involve conducting a well
planned and predetermined set of chosen experiments based on a
multivariate design of experiments.
A multivariate approach that utilises a second-degree quadratic
polynomial is thus, necessary in order to obtain unbiased
responses that would adequately describe the process with a
minimum number of experiments. We present the degradation of
phenol based on a central composite design (CCD) with two level
factorial design with additional experimental referred to as the star
point; the latter experiments are introduced for determining the
quadratic response surface. Three repetitions at the centre points
were done to allow for better prediction of the reproducibility (Kiril et
al., 2010). In addition, experiments at the centre points facilitate the
determinations of rotatability or orthogonality of the design and
ease of fitting the quadratic polynomials with at least one point
(Tarley et al., 2009).
Four independent variables were investigated namely initial
phenol concentration, the concentration ratio of [H2O2]:[phenol],
mass ratio of [H2O2]:[Fe2+] and reaction time. However, pH was not
considered as a specific variable due to the several reported
literature that established the optimal pH at about 3 (Hermosilla et
al., 2009; Ay et al., 2009; Martins et al., 2005). This was based on
limitation of Fenton reagents performance in different pH media. As
summarised by Kiril et al. (2010), increasing pH value decreases
the oxidation potential of the hydroxyl radicals (•OH) and generally,
in a basic medium, the oxidation yield is observed to decrease as
Fe3+ exists as a precipitate of Fe(OH)3. This again is associated
with hindering the reaction between Fe3+ and H2O2 by the formed
Fe(OH)3, thereby limiting the regeneration step of Fe2+. Moreover,
the Fe(OH)3 is known to catalyze the decomposition of H2O2 to O2
and H2O, thus, decreasing the production of •OH and retarding the
systems performance. On the other hand, at low pH, H2O2 is
stabilized as H3O2+, these results in excess H+ within the reacting
system and favouring the reaction between OH• and H+. This leads
to retarding the regeneration of Fe2+ by reaction of Fe3+ with H2O2
which is necessary for sustenance of the oxidation reaction.
Literature is replete with studies that established optimum pH in
Fenton’s process to range between 2.5 and 3.5 (Bena tti et al.,
2006; Iurascu et al., 2009; Kavitha and Palanivelu, 2004; Zazo et
al., 2009; Kiril et al., 2010). Thus, in this study initial reaction pH
was fixed at 3. The mineralization was monitored at three reaction
times 20, 100 and 180 min for the three different
phenol concentrations.
The choice of CCD out of the numerous available designs was
informed by it is popularity, reliability and its general acceptance as
a standard second-order design (Tarley et al., 2009; Virkutyte et al.,
2010).
Table 1 shows the experimental design matrix with the real
values for the independent variables set at -1, 0 and +1 for low,
medium and high levels, respectively. From the experimental runs
performed, mathematical relationship of the response function ( )
and influence of the independent variables is express as a general
quadratic polynomial model (Equation 3). The regression
coefficients obtained by response surface regression analysis were
achieved by fitting experimental data to the quadratic model.
Experimental design, data analysis and process optimization
Interaction of the variables involved in Fenton’s p rocess is strong,
thus, use of one variable at a time (OVAT) approach to study and
interpret responses in a multivariate system such as the Fenton
process is grossly inadequate. The OVAT refers to the approach of
k
y = βo +
k
β i xi +
i =1
k
β ii xi 2
β ij xi x j +
j
< i=2
(3)
i =1
The regression coefficients for intercept, linear, quadratic and
Diya’uddeen et al.
10221
Table 1. Experimental and predicted responses with corresponding total iron content for each batch.
Run
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
2+
(Phenol)o (mg/l)
(H2O2): (Phenol)
(H2O2):(Fe )
tr (min)
100.0
52.5
100.0
5.0
5.0
52.5
52.5
100.0
100.0
5.0
5.0
100.0
52.5
52.5
52.5
52.5
100.0
5.0
5.0
5.0
100
5.0
52.5
5.0
100.0
100.0
52.5
10
10
6
2
2
10
2
2
6
6
6
6
10
10
2
6
10
2
6
10
2
2
6
6
2
10
10
25
15
15
5
25
5
5
5
25
15
15
5
5
25
25
15
5
5
15
25
25
25
15
15
15
5
25
180
20
20
20
180
100
100
180
20
180
180
20
100
100
100
100
180
20
20
180
180
20
100
100
20
100
180
Responses (%)
yExptl
yPred
41.8
39.5
60.1
60.4
34.6
41.1
88.4
92.1
87.0
91.3
61.8
58.8
54.2
55.4
67.2
14.2
36.1
39.2
87.0
82.8
87.4
79.9
37.5
34.7
61.8
61.9
66.1
61.9
60.6
61.9
60.1
62.2
35.17
28.5
86.6
94.8
88.1
80.0
38.6
46.24
13.6
16.14
85.3
81.7
59.36
61.5
86.5
85.9
14.1
7.9
33.7
32.4
63.0
63.3
Total Fe (mg/l)
10.5
1.3
ND
3.3
0.3
4.4
0.3
ND
0.4
34.7
3.6
ND
ND
14.9
ND
15.7
2.7
30.1
ND
ND
1.7
ND
ND
ND
45.0
0.4
0.4
ND, Not detected; yExptl, experimental responses; yPred, predicted responses.
interaction terms are represented by 0, i, ii and ij, respectively,
while the independent variables that affect the response
are
denoted by xi, xii and xij. Data generated was analysed, coefficients
in the mathematical model predicting the response estimated and
analysis of variance (ANOVA) was used for gaining insight into the
interaction between the variables affecting the process and the
response. Adequacy check was also performed to ascertain the
quality of the fit of polynomial expressed in equation 3 using the
coefficient of determination R2 and R2adj.
RESULTS AND DISCUSSION
Studies were conducted batch wise according to a
statistically designed experimental matrix (Table 1) in
order to map the whole region and obtain reaction
conditions at which Fenton’s reagent consumption is
minimal with corresponding low total iron in the treated
samples.
The dependent variable of the process was the TOC
removal percentage which was approximated using the
CCD and represented by Equation (4).
y = 99.78 - 0.65 x1 - 1.76 x2 + 0.06 x3 - 0.01 x4 + 0.03 x1
-3
x2 + 2.16x10 x1 x3 - 1.60 x1 x4 + 0.03 x2 x3 + 5.91 x2 x4 -4
-4
-3
2.16 x10 x3 x4 + 1.99 x10 x1 x3 x4 + 1.37 x10 x2 x3 x4
(4)
The independent variables namely [Phenol]o, [H2O2]:
2+
[Phenol], ], [H2O2]:[Fe ] and tr, are represented by x1, x2,
x3 and x4, respectively. These variables were studied,
their effects on the response analysed and the
parameters were estimated by the method of least
squares. The response surface was then analysed in
terms of the fitted surface.
To assert the suitability of the proposed model for
navigating the design space in terms of TOC percent
removal, a model adequacy test was conducted. In Table
2, the results of the ANOVA for the proposed model for
percent TOC removals are presented. Proper approximation of the true system can only be achieved with a
valid model. Thus, model adequacy check is necessary
to validate the approximations and ensure that none of
the least square regression assumptions are violated
10222
Afr. J. Biotechnol.
Table 2. ANOVA results of the quadratic polynomial model for Fenton treatment of synthetic phenol wastewater.
Response
TOC removal (%)
Source
Model
Residual
Lack of fit
Pure error
Total
Sum of squares
15994.57
557.14
540.59
16.55
16551.71
Degree of freedom
16
10
8
2
26
Mean square
999.66
55.71
67.57
8.28
F-ratio
17.94
P-value
< 0.0001
8.16
0.1136
R2 = 0.9663; R2adj = 0.9125; Adq. prec. ratio = 14.672.
Residuals vs. Predicted
3.00
1.50
0.00
-1.50
-3.00
7.93
29.65
51.38
73.10
94.83
X: Predicted
Y: Studentized residuals
Figure 1. Studentized residuals and initial phenol concentrations plot of the Fenton
mineralisation.
(Tekin et al., 2006).
From the ANOVA results (Table 2), the model obtained
was significant for the mineralization of the phenol as the
TOC removal percent has an F-value of 17.94, implying
that there is only a 0.01% chance that a model F-value
this large could occur due to noise. P-values less than
0.05 indicate that, the model is significant whereas
greater than 0.10 indicates model not significant; with a
value of 0.001 the model adequacy fit is ascertained. A
value of 14.672 for the adequate precision signal to noise
ratio was derived for the response variable. This value
exceeds the minimum desirable value of 4 signifying
adequate signal for the TOC removal percent.
From the ANOVA analysis in Table 2, the model could
satisfactorily predict the phenol TOC removal percent
well. The value of 0.9663 for the correlation coefficient of
2
the model (R ) indicate that, there is a satisfactory
quadratic fit as it implies that 97% of the total variation in
TOC removal reported was adequately represented by
2
the model. Although, R reasonably predicts model
2
adequately fairly well but having a high value of R does
not necessarily imply a very good prediction. This is
associated with the fact that, there exists the possibility of
having poor predictions of new observations or estimates
2
of the mean response, thus, a modified form of the R
2
2
referred to as R adj is employed. Values of R are always
2
2
higher than or equal to R adj because the R adj adjusts the
number of explanatory terms in a model thus, it can have
a negative value. Merit of the former parameter over the
latter is that, while addition of variables to the model
2
always results in an increase in R value it is significantly
2
not affected in R adj it is often reported to decrease if
unnecessary terms are added. Based on closeness of
2
correlation coefficient values of 0.9663 and 0.9125 for R
2
and R adj, respectively, it appears that the experimental
data fitted the model fairly well and signified non-inclusion
of insignificant terms. The "lack of fit F-value" of 8.16
implies that the lack of fit is not significant relative to the
pure error. There is a 11.36% chance that a "lack of fit Fvalue" this large could occur due to noise.
In Figure 1, studentized residuals against predicted
TOC removal percent is depicted and clearly indicates
Diya’uddeen et al.
10223
Predicted vs . Actual
94. 83
72. 91
51. 00
29. 08
7. 17
7. 17
29. 08
51. 00
72. 91
94. 83
X: Ac t ual
Y : Predic t ed
Figure 2. Plot of actual responses against predicted values of the TOC
removal percent.
Outlier T
3.50
1.72
-0.06
-1.84
-3.62
1
6
11
16
21
26
X: Run number
Y: Outlier T
Figure 3. The outlier plot of the Fenton process for phenol mineralisation.
the lack of need for transformation. The random
scattering of the points instead of funnel-shaped pattern
sometimes observed suggest that, the variance of original
observations to be a constant for all values of the
response (Myers and Montgomery, 2002; Montgomery,
2009).
Experimentally determined responses (actual values)
were plotted against responses obtained from the
developed approximating function (predicted values) the
plot (Figure 2) showed that, there is a good agreement
between the two set of results thus, validating the
model’s reliability under the independent variables
investigated.
Finally, the number of standard deviations of the actual
value deviates from the predicted value was determined
from the outlier plot (Figure 3) for the TOC removal
percent. The outliers (data points that lied far away from
the true regression line) need to be assessed as they can
be used to determine data recording error or region of the
independent factor variable space where the fitted model
is a poor approximation to the true response surface
(Montgomery, 2009).
10224
Afr. J. Biotechnol.
80
70
10
20
30
40
50
60
70
80
60
TOC Removal (%)
50
40
30
20
10
ol]
en
h
P
4
]:[
2
O2
[H 2
6
20
15
10
5
8
10
[H2 O2 ]:[Fe 2+]
Figure 4. 3D response surface graph for phenol mineralisation at pH = 3, [H2O2]:[Phenol] = 2 to
10 and [H2O2]:[Fe2+]= 5 to 25.
Furthermore, the model equation was obtained using the
least squares method, the slope and intercept which are
sensitive to outliers thus, estimates of regression coefficients can be distorted. From the outlier plot, it is shown
that all the standardized residual are within the range of
±3.50 interval. This shows that, the model approximation
of the response surface was satisfactory and not
associated with data recording error.
Effect of independent variables on TOC % removal
The most common wastewater analysis index is COD
reduction. However, it must be mentioned that the COD
test is based on the assumption that all the organic
materials can be oxidized by a strong oxidizing agent
under acidic conditions. However, the COD test has
some restrictions as detailed by Papadopoulos et al.
(2007). Some aromatic compounds are not oxidized
completely within the COD test procedure and thus, the
oxygen demands obtained do not reflect the actual
oxygen requirements for oxidation. Phenol containing
effluents are one of such wastewater as they are
aromatic and yield aromatic by products. Hence, TOC is
the most suitable measure and a direct expression of the
total organic content than COD.
2+
Effect of the ratios [H2O2]:[Phenol] and [H2O2]:[Fe ]
Figure 4, depicts the response surface showing the
interaction between two variables ([H2O2]:[phenol] and
2+
[H2O2]:[Fe ]) at constant pH of 3.0 and over the whole
concentration range considered. A semi spherical
response surface was obtained with maximum TOC
removal at the ratios of 6 and 15 for [H2O2]:[phenol] and
2+
[H2O2]:[Fe ], respectively. Moreover, the significance of
investigating the interaction of both parameters is clearly
demonstrated. It is observed that at a ratio of 15 for
2+
[H2O2]:[Fe ] the optimal TOC reduction was attained.
Here, oxidation reaction performed at a ratio of 2
([H2O2]:[phenol]) results in a sharp decrease in the
system’s performance.
The increase in TOC reduction with the increase in the
ratio is as observed by Chakinala et al. (2008) and that
the ratio increase favours the simultaneous generation of
more ·OH radical as a result of an increase in oxidant
concentration. At the ratio of 15 for the parameter
2+
2+
[H2O2]:[Fe ] sufficient Fe species are available within
the reaction mixture propelling more catalytic degradation
of the oxidant.
High TOC reductions are observed at the edges of the
matrix region, these are represented in experimental
runs. However, when conducting the reaction at these
Diya’uddeen et al.
30
40
50
60
70
80
90
90
80
TOC removal (%)
10225
70
60
50
40
30
40
60
80
100
120
140
160
180
9 8
10
3 2
en
5 4P h
6
[
7
]:
[H
2O
ol
]
2
e ( m in )
Timet i m(min)
Figure 5. 3D response surface graph for phenol mineralisation at pH = 3, [H2O2]:[Phenol] = 2 to 10 and reaction time = 20
to 180 min.
conditions the pH of the reaction mixture should be
monitored as a decrease in the pH can be anticipated.
This is due to the carboxylic intermediate products
generated from the degradation of phenol, which is acidic
in nature (Kavitha and Palanivelu, 2004). This promotes
+
the availability of excess H as described earlier in
materials and methods. Furthermore, reaction between
•
+
OH and the H is enhanced and thus, the system’s
efficiency might decrease (Kiril et al., 2010).
Effect of reaction time and [H2O2]:[Phenol] ratio
The 3D interactive effect of reaction time and
[H2O2]:[Phenol] of the whole concentration is shown in
Figure 5. It is seen that the TOC percent removal
increased with an increase in the ratio [H2O2]:[Phenol]
with the optimal obtained at the maximal oxidation time.
The ratio earlier observed for the ratio was found to be
maintained with respect to time also. The TOC reduction
was found to increase with increase in the ratio up to 6
after which the efficiency decreased. The lower after the
optimum ratio are consequences of the radical reactions
described earlier.
Percent TOC removal for [H2O2]:[phenol] increased
considerably and rapidly for the values investigated. The
exceptions are the ratio of 2 and 10 at reaction times of
approximately 20 min. At these conditions, oxidant
concentration was sufficient from the onset to mineralise
the phenol within the first 20 min. From the contour plot in
Figure 6, it would be seen that ratios lower than 6 for
[H2O2]:[phenol] do not favour the TOC removal likewise
above a ratio of 8. For [H2O2]:[Phenol] ratio, the optimal is
seen to be located around the value of 15. It is thus,
evidently clear that the peroxidation reaction within the
range of 6 to 8 significantly resulted in a TOC removal
increase from a 60 to almost 90% when treated at a ratio
of 15.
The mineralization of the aqueous phenol solution fitted
the experimental and predicted data fairly well. The
quadratic terms in the equation are due to the strong
influence of the regressors on the coefficients and the
response graphically depicted by the surface response
curve in Figure 4.
The significance of the ratio [H2O2]:[phenol] is vividly
seen from the equation as higher values positively
influenced the response, this is also supported by the
high values observed from the interaction coefficient of
the ratio with time.
2+
Oxidant to catalyst mass ratio, [H2O2]:[Fe ]
As observed by Benatti et al. (2006), the ratio of
2+
[H2O2]:[Fe ] is a crucial factor that determines the degradation efficiency. All the variables studied are known to
10226
Afr. J. Biotechnol.
25
60
40
40
50
50
50
50
50
TOC Removal percent
60
20
60
2+
[H2O2]:[Fe ]
40
60
70
60
30
50
70
70
15
20
40
70
60
30
50
60
10
60
40
50
40
50
40
50
40
5
2
3
4
5
6
7
8
9
10
[H2O2]:[Phenol]
Figure 6. Contour plot for phenol mineralisation at pH = 3, [H2O2]:[Phenol] = 2 to 10 and
[H2O2]:[Fe2+]= 5 to 25.
interact strongly and the overall process efficacy relies on
attaining a balance in quantities of both the Fenton’s
reagents used in relation to the strength of the organic
contamination. For the [H2O2], it is of uttermost
importance to introduce just enough amounts that would
sustain the reaction to completion not only because
excess contributes to COD, but also the adverse effect of
the excess to microorganisms when eventually discharged (Chakinala et al., 2008). It is therefore imperative
to determine an optimum ratio at which to attain maximal
TOC removal and consume the reagents.
The required catalyst necessary to catalyses the
decomposition of the oxidant was computed based on the
required oxidant concentration needed to oxidize the
phenol solution. Thereafter, the mass ratio of the oxidant
and catalyst was varied as described in several
literatures. The ranges of values used in this study were
chosen to fall within the reported limits (Benatti et al.,
2006; Martins et al., 2010; Martins et al., 2005).
Several authors have worked with molar ratio as the
basis for establishing the ratios mass ratios
2+
[H2O2]:[phenol] and [H2O2]:[Fe ]. However, recent
studies report use of mass of reagents instead of molar
concentration (Martins et al., 2005; Benatti et al., 2006;
Martins et al., 2010; Bach et al., 2010). We thus, adopted
the mass approach to determine the necessary amounts
of the reagents that would maintain the ratio of the
designed matrix (Table 1).
From the plot of Figure 7, it is evident that the ratio of
15 is the optimal. This condition is furthermore, concurred
with the 3D response surface and contour plots in
Figures 4 and 6, respectively. At ratios above 15,
possible scavenging reactions of the generated radicals
with less reactive hydroperoxyl radicals between the •OH
radicals themselves or the oxidant and •OH radical might
occur (Kavitha and Palanivelu, 2004) as shown in
Equations (5) to (7):
•
•
H2O2 + OH
H2O + HO2
•
H2O + O2
•
HO2 + OH
(6)
•
•
OH + OH
(7)
(5)
H2O2
•
Contribution to the scavenging of the OH radicals have
also been reported by Bouasla et al. (2010) and
represented by equation (8):
2+
Fe
•
+ OH
(8)
3+
Fe
+ HO
-
Initial phenol concentration
The design matrix dictates three initial phenol
concentrations and from Figure 8 of the contour’s p lot,
the process was more effective in TOC removal at a
lower concentration. As seen from the plot, TOC removal
Diya’uddeen et al.
10227
100
TOC removal (%)
80
0
20
40
60
80
100
60
40
20
ti m
0
40
60
80
100
120
140
160
180
e
(m
in
)
25
20
15
10
5
2+
]
[H 2O 2]:[Fe
Figure 7. 3D response surface graph for phenol mineralisation at pH = 3, [H2O2]:[Fe2+] = 2 to 10
and reaction time = 20 to 180 min.
slightly decreased with an increase in reaction time and
concentrations. The maximum phenol TOC removals at
the end of the oxidation reaction time were 88.4 (20 min),
66.1 (100 min) and 41.8% (180 min) for 5, 52.5 and 100
mg/l, respectively. From Table 1, the ratios corresponding
to these were 2 and 5, 10 and 25 and 10 and 25
2+
[H2O2]:[phenol] and [H2O2]:[Fe ], respectively. However,
relatively similar TOC reductions are equally obtained at
a lower reaction time of 20 min for the 52.5 and 100 mg/l
concentrations. These are the runs 3 (34.5%) and 16
(60.1%) at the following conditions: 6 and 15, and 10 and
2+
15 for [H2O2]:[phenol] and [H2O2]:[Fe ] ratios,
respectively.
From Figure 8, it is seen that mineralization is favoured
with the increase in concentration and for all the phenol
concentrations the optimal mineralization could be
attained within the first 20 min. This implies that at the
•
process conditions investigated, OH formation and its
consumption were favoured with a decrease in
concentration as observed by Lopez et al. (2005). This
led to the rapid rate of reaction recorded and clearly
indicated limitations of the H2O2 at a higher
concentration. For effective TOC removal at such
concentrations, the ratio [H2O2]:[phenol] should be higher
than the currently studied limits of 10.
Individual effect of (phenol)o with reaction time
The trend depicted by the treatment at different oxidation
times varied with initial concentration of the phenol
solution (Figure 9). Clearly, a reaction time of 20 min is
sufficient to mineralise approximately 40, 70 and 90% of
the phenol. At a reaction time of 20 min, rapid •OH attack
to the phenolic aromatic ring resulted in the greater TOC
removal percentages recorded. Thus, increase in
reaction time beyond 20 min decreases the efficiency of
the process in treating wastewater containing phenol at
such concentrations and at the operating conditions
reported in this study.
The TOC reduction for the three concentrations with
respect to time is presented in Figure 9. The reduction
ranged from 7.2 to 41.8, 59.3 to 66.1 and 38.6 to 88.4%
for 100, 52.5 and 5 mg/l initial phenol concentrations. The
decrease in TOC reduction is directly related to
competing for and consumption of the generated •OH by
10228
Afr. J. Biotechnol.
180
70
60
50
40
30
160
TOC removal (%)
140
80
70
Time (min)
time (min)
120
60
40
50
100
80
80
70
60
70
60
50
40
60
80
40
50
40
20
20
40
60
80
100
[Phenol]o (mg/l)
Figure 8. Contour plot for phenol mineralisation at pH = 3, [H2O2]:[Phenol] = 2 to 10 and
[H2O2]:[Fe2+]= 5 to 25.
100
TOC reduction (%)
80
60
100 mg/l
52.5 mg/l
5 mg/l
40
20
0
0
20
40
60
80
100
120
140
160
180
200
Time(min)
(min)
time
Figure 9. Plot of phenol % TOC reduction for [Phenol]o = 5, 52.2 and 100 mg/l. Reaction time
= 20 to 180 at pH = 3.
more organic radical associated with higher phenol
concentrations. This strongly indicates that initial
concentration had a significant effect on the degradation
of the phenol.
For the 5 mg/l initial phenol concentrations, results
showed an increase in reaction time did results in
appreciable TOC reduction thus, not time sensitive.
Except for the run at 20 which yielded 38.6% after 180 min
Diya’uddeen et al.
of reaction time; all the other runs ranged between 85.3
and 88.4%. As the highest ratio was employed, the low
efficiency could be associated with excess H2O2 available
in the system which retarded the Fenton process. It has
·
been shown that, the OH radicals are trapped by the
excess H2O2 molecules leading to the earlier described
scavenging reactions (5) to (7) (Kavitha and Palanivelu,
2004).
Generally, the reagents consumption ranged from
0.003 to 0.1802 ml of 30 w/w% H2O2 and 0.008 to 4 ml for
2+
2+
Fe . The concentration of the Fe varied between 0.04
and 12 mg/l. On the reagent utilisation, specifically and at
the optimal conditions the following was recorded: (1)
2+
(Phenol)o = 100 mg/l; H2O2 = 0.1802 ml and Fe = 0.8
ml (4 mg/l), (2) (Phenol)o = 52.5 mg/l; H2O2 = 0.1517 ml
2+
and Fe = 0.7 ml (3.5 mg/l) and (3) (Phenol)o = 52.5
2+
mg/l; H2O2 = 0.1517 ml and Fe = 0.7 ml (3.5 mg/l). It is
evident that less reagents were employed introduced into
the system.
Comparison with
mineralization
similar
works
on
phenol
The main merit of advanced oxidation is its potentials for
mineralizing effluent and the indices determining this
parameter is the TOC values. Comparatively values of
degradation (COD) are higher than mineralization (TOC).
As Bach et al. (2010) highlighted, high COD values do
not necessarily signify completion of the oxidation
reaction or removal efficiencies. Thus, the comparison of
the AOP was limited to works reporting mineralization.
Mostly reported mineralization rates of phenol are less
than 60%, for example, Kavitha and Palanivelu (2004),
achieved 41% mineralization for the unassisted Fenton
method after 2 h. Hermosilla et al. (2009) reported a
reduction of about 80% in TOC in their photo-Fenton
treatment. Lurascu et al. (2009) degraded and
mineralized phenol using an immobilized heterogeneous
photo-Fenton catalyst, a near total conversion of the
phenol at 1 mmol/l concentration was achieved.
However, the process was costly as it involved use of
oxidant, heating at elevated temperatures to prepare the
catalyst and use of UV source for the irradiation. Yalfani
et al. (2009) reported 58% mineralization using the
conventional Fenton process and modified the process
by generating the hydrogen peroxide in situ using oxygen
and formic acid; they obtained almost complete
destruction of phenol at a reaction time of 360 min with
nearly 60% TOC removal.
Structurally, phenol is a cyclic compound consisting of
six and five member rings, thus, the energy requirement
for breaking this cycle is much higher relative to single
chain as the former is less reactive and more stable than
single-chained compounds (Annadurai et al., 2008). In
the open scientific literature, works on combined Fenton’s
reagent treatments with other techniques are abundant.
10229
For example, Chakinala et al. (2008) reported combined
hydrodynamic cavitation and heterogeneous advanced
Fenton’s processing for phenol mineralization in a
multivariate design, the maximum TOC removal after a
reaction time of 105 min using the modified AFP was 50
to 60%.
Other AOP’s conversion of phenol was low when using
low-pressure UV lamps, for example, Hosseini et al.
(2007) reported a maximum of 52.7% conversion with 80
W lamps over a reaction time of 240 min. Though they
achieved higher conversion but employed a 125 W UV
lamp. Similarly, 40 and 80 mg/l phenol was decomposed
by the photocatalytic treatment in 6 h (Augugliaro et al.,
1988) and 8 h (Sivalingam et al., 2004), respectively.
Comparison of researches that considered same reaction
time of 180 min achieved 50% (Hosseini et al., 2007) and
25% (Carpio et al., 2005) TOC removal. Suryaman et al.
(2006) mineralized 50 mg/l phenol, although 98% TOC
removal was achieved, but it required 30 min of an AOP
(photocatalytic process) coupled with 10.5 h of biological
treatment.
From the aforementioned researches compared, it is
evident that mineralizing phenol based on conditions of
our findings is more attractive than most reported TOC
removal percentages. Arguably, some works apparently
achieved a better result than present work. However, as
highlighted earlier their process methodology was capital
intensive, required elaborate design and setting,
maintenance cost, which would offset the marginal
benefit.
Conclusions
Response surface methodology was successfully used in
optimizing the mineralization parameters for a phenol
solution at concentrations ranging from 5 to 100 mg/l by
Fenton oxidization method. An approximating model
equation was developed that adequately described the
mineralization process. Based on residual analysis, the
suitability of the model was found to be satisfactory. Also,
data generated from the quadratic polynomial fitted the
response surface well with correlation coefficient values
2
2
of 0.9663 and 0.9689 for R and R adj, respectively.
The mineralization was effectively achieved within 20
min for the range of concentrations investigated. The
optimum conditions for the process are ratios of 6 and 15
2+
for [H2O2]:[phenol] and [H2O2]:[Fe ], respectively. The
maximum Total Organic Carbon (TOC) reduction was
88.4%.
The findings of the study show fast reaction rate and
low reagents consumption. This is seen in the case of
optimal conditions for: (1) (Phenol)o = 100 mg/l; H2O2 =
2+
0.1802 ml and Fe = 0.8 ml (4 mg/l), (2) (Phenol)o =
2+
52.5 mg/l; H2O2 = 0.1517 ml and Fe = 0.7 ml (3.5 mg/l)
2+
and (3) (Phenol)o = 52.5 mg/l; H2O2 = 0.1517 ml and Fe
= 0.7 ml (3.5 mg/l).
10230
Afr. J. Biotechnol.
The study further asserted effectiveness of the conventional Fenton treatment technique in wastewater
treatment. Furthermore, conducting the phenol mineralization at the experimental conditions negates the need
for post treatment step necessary to reduce total iron
ions, as 85% of the treated samples conformed to the
minimum discharge levels. However, further work using
industrial effluents known to contain myriad of contaminants and compounds with potential for consuming the
radicals generated can be undertaken. Such studies
would give an insight on wastewater containing radical
scavengers such as chlorides and sulfides, for example,
petroleum refinery effluents (PRE).
REFERENCES
Ahmadi M, Vahabzadeh F, Bonakdarpour B, Mofarrah E, Mehranian M
(2005). Application of the central composite design and response
surface methodology to the advanced treatment of olive oil
processing wastewater using Fenton's peroxidation. J. Hazard.
Mater. 123: 187-195.
Annadurai G, Ling LY, Lee JF (2008). Statistical optimization of medium
components and growth conditions by response surface methodology
to enhance phenol degradation by Pseudomonas putida. J. Hazard.
Mater. 151: 171-178.
Augugliaro V, Palmisano L, Sclafani A, Minero C, Pelizzetti E (1988).
Photocatalytic degradation of phenol in aqueous titanium dioxide
dispersions Toxic & Environ Chem. Pp: 89-109.
Ay F, Catalkaya EC, Kargi F (2009). A statistical experiment design
approach for advanced oxidation of Direct Red azo-dye by photoFenton treatment. J. Hazard. Mater. 162: 230-236.
Bach A, Shemer H, Semiat R (2010) Kinetics of phenol mineralization
by Fenton-like oxidation Desalination doi:10.1016/j. desal.
2010.04.011.
Banerjee A, Ghoshal AK (2005). Phenol degradation by Bacillus cereus:
Pathway and kinetic modelling. Bioresour. Technol. 101: 5501-5507.
Barros AL, Pizzolato TM, Carissimi E, Schneider IAH (2006).
Decolorizing dye wastewater from the agate industry with Fenton
oxidation process. Miner. Eng. 19: 87-90.
Benatti CT, Tavares CRG, Guedes TA (2006). Optimization of Fenton's
oxidation of chemical laboratory wastewaters using the response
surface methodology. J. Environ. Manage. 80: 66-74.
Bouasla C, Samar ME-H, Ismail F (2010) Degradation of methyl violet
6B dye by the Fenton process, Desalination 254: 35-41.
Busca G, Berardinelli S, Resini C, Arrighi L (2008). Technologies for the
removal of phenol from fluid streams: A short review of recent
developments. J. Hazard. Mater. 160: 265-288.
Cai W, Li J, Zhang Z (2007). The characteristics and mechanisms of
phenol biodegradation by Fusarium sp. J. Hazard. Mater. 148: 38-42.
Carpio E, Zúñiga P, Ponce S, Solis J, Rodriguez J, Estrada W (2005).
Photocatalytic degradation of phenol using TiO2 nanocrystals
supported on activated carbon. J. Mol. Catal A: Chem. 228: 293-298.
Chakinala AG, Bremner DH, Gogate PR, Namkung K-C, Burgess AE
(2008). Multivariate analysis of phenol mineralisation by combined
hydrodynamic cavitation and heterogeneous advanced Fenton
processing. Appl. Catal B Environ. 78: 11-18.
Chen Y, Chen Z, Xiao S, Liu H (2008). A novel thermal degradation
mechanism of phenol-formaldehyde type resins. Thermochimica Acta
476: 39-43.
Doan HD, Weli A, Wu J (2009). A combined photocatalytic and
electrochemical treatment of wastewater containing propylene glycol
methyl ether and metal ions. Chem. Eng. J. 151: 51-58.
Gernjak W, Krutzler T, Glaser A, Malato S, Caceres J, Bauer R,
Fernandez-Alba AR (2003) Photo-Fenton treatment of water
containing natural phenolic pollutants. Chemosphere. 50: 71-78.
Gonzalez G, Herrera G, Garcia MT, Pena MM (2001). Biodegradation of
phenolic industrial wastewater in a fluidized bed bioreactor with
immobilized cells of Pseudomonas putida. Bioresour. Technol. 80:
137-142.
Guo J, Al-Dahhan M (2005). Catalytic wet air oxidation of phenol in
concurrent downflowand upflow packed-bed reactors over pillared
clay catalyst. Chem. Eng. Sci. 60: 735-746.
Hermosilla D, Cortijo M, Huang CP (2009). The role of iron on the
degradation and mineralization of organic compounds using
conventional Fenton and photo-Fenton processes. Chem. Eng. J.
155: 637-646.
Hong S-S, Ju C-S, Lim C-G, Ahn B-H, Lim K-T, Lee G-D (2001). A
photocatalytic degradation of phenol over TiO2 prepared by sol-gel
method. J. Ind. Eng. Chem. 7: 99-104.
Hosseini SN, Borghei SM, Vossoughi M, Taghavinia N (2007).
Immobilization of TiO2 on perlite granules for photocatalytic
degradation of phenol. Appl. Catal B Environ. 74: 53-62.
Iurascu B, Siminiceanu I, Vione D, Vicente MA, Gil A (2009). Phenol
degradation in water through a heterogeneous photo-Fenton process
catalyzed by Fe-treated laponite. Water Res. 43: 1313-1322.
Kavitha V, Palanivelu K (2004). The role of ferrous ion in Fenton and
photo-Fenton processes for the degradation of phenol. Chemosphere
55: 1235-1243.
Kiril MB, Yonar T, Yalili Kiliç M, Kestioglu K (2010). Pre-treatment
studies on olive oil mill effluent using physicochemical Fenton and
Fenton-like oxidations processes. J. Hazard. Mater. 174: 122-128.
Li X-y, Cui Y-h, Feng Y-j, Xie Z-m, Gu J-D (2005). Reaction pathways
and mechanisms of the electrochemical degradation of phenol on
different electrodes. Water Res. 39: 1972-1981.
Lopez A, Mascolo G, Detomaso A, Lovecchio G, Villani G (2005).
Temperature activated degradation (mineralization) of 4-chloro-3methyl phenol by Fenton's reagent. Chemosphere 59: 397-403.
Mahiroglu A, Tarlan-Yel E, Sevimli MF (2009). Treatment of combined
acid mine drainage (AMD) - Flotation circuit effluents from copper
mine via Fenton's process. J. Hazard. Mater. 166: 782-787.
Martins AF, Vasconcelos TG, Wilde ML (2005). Influence of variables of
the combined coagulation-Fenton-sedimentation process in the
treatment of trifluraline effluent. J. Hazard. Mater. 127: 111-119.
Martins RC, Rossi AF, Quinta-Ferreira RM (2010). Fenton's oxidation
process for phenolic wastewater remediation and biodegradability
enhancement. J. Hazard. Mater. 180:716-721.
Montgomery DC (2009). Design and Analysis of Experiments. 7th
edition, John Wiley and Sons, New York.
Moussavi G, Barikbin B, Mahmoudi M (2010). The removal of high
concentrations of phenol from saline wastewater using aerobic
granular SBR. Chem. Eng. J. 158: 498-504.
Myers RH, Montgomery DC (2002). Response Surface Methodology:
Process and Product Optimization Using Designed Experiments. 2nd
ed., Wiley, New York.
Papadopoulos AE , Fatta D, Loizidou M (2007). Development and
optimization of dark Fenton oxidation for the treatment of textile
wastewaters with high organic load, J. Hazard. Mater. 146: 558-563.
Pontes RFF, Moraes JEF, Machulek A(Jr), Pinto JM (2010). A
mechanistic kinetic model for phenol degradation by the Fenton
process. J. Hazard. Mater. 176: 402-413.
Prieto MB, Hidalgo A, Serra JL, Llama MJ (2002). Degradation of
phenol by Rhodococcus erythropolis UPV-1 immobilized on Biolite in
a packed bed reactor. J. Biotechnol. 97: 1-11.
Priya SS, Premalatha M, Anantharaman N (2008). Solar photocatalytic
treatment of phenolic wastewater-potential challenges and
opportunities. ARPN J. Eng. Appl. Sci. 3: 36-41.
Salamatinia B, Mootabadi H, Bhatia S, Abdullah AZ (2010).
Optimization of ultrasonic-assisted heterogeneous biodiesel
production from palm oil: A response surface methodology approach.
Fuel Proces. Technol. 91: 441-448.
Sivalingam GP, Priya MH, Madras G (2004). Kinetic of the
photodegradation of substituted phenols by solution combustion
synthesized TiO2. Appl. Catal B Environ. 51: 67-76.
Suryaman D, Hasegawa K, Kagaya S (2006). Combined biological and
photocatalytic treatment for the mineralization of phenol in water.
Chemosphere, 65: 2502-2506.
Tarley CRT, Silveira G, dos Santos WNL, Matos GD, da Silva EGP,
Bezerra MA, Miró M, Ferreira SLC (2009). Chemometric tools in
electro analytical chemistry: Methods for optimization based on
Diya’uddeen et al.
factorial design and response surface methodology. Microchem. J.
92: 58-67.
Tekin H, Bilkay O, Ataberk SS, Balta TH, Ceribasi IH, Sanin FD, Dilek
FB, Yetis U (2006). Use of Fenton oxidation to improve the
biodegradability of a pharmaceutical wastewater. J. Hazard. Mater.
136: 258-265.
Virkutyte J, Rokhina E, Jegatheesan V (2010). Optimisation of ElectroFenton denitrification of a model wastewater using a response
surface methodology. Bioresour. Technol. 101: 1440-1446.
Wei Y, Van Houten RT, Borger AR, Eikelboom DH, Fan Y (2003).
Minimization of excess sludge production for biological wastewater
treatment. Water Res. 37: 4453-4467.
10231
Yalfani MS, Contreras S, Medina F, Sueiras J (2009). Phenol
degradation by Fenton's process using catalytic in situ generated
hydrogen peroxide. Appl. Catal B Environ. 89: 519-526.
Yang S, Zhu W, Wang J, Chen Z (2008). Catalytic wet air oxidation of
phenol over CeO2-TiO2 catalyst in the batch reactor and the packedbed reactor. J. Hazard. Mater. 153: 1248-1253.