Optimization of photocatalytic performance of TiO2 coated glass microspheres using response surface methodology and the application for degradation of dimethyl phthalate

Article abstract of DOI:10.1016/j.jphotochem.2013.04.008

Hollow glass microspheres coated with photocatalytic TiO₂ (HGM-TiO₂), recently became commercially available and have the distinct advantages of easy separation and recovery after treatment. With this in mind, we determined the optim

Full text of DOI:10.1016/j.jphotochem.2013.04.008

Journal of Photochemistry and Photobiology A: Chemistry 262 (2013) 7–13
Contents lists available at SciVerse ScienceDirect
Journal of Photochemistry and Photobiology A:
Chemistry
journal homepage: www.elsevier.com/locate/jphotochem
Optimization of photocatalytic performance of TiO2 coated glass microspheres
using response surface methodology and the application for degradation of
dimethyl phthalate
Wenjun Jianga, Jeffrey A. Joensa, Dionysios D. Dionysioub, Kevin E. O’Sheaa,∗
a Department of Chemistry and Biochemistry, Florida International University, Miami, FL 33199, USA
b Environmental Engineering and Science Program, University of Cincinnati, Cincinnati, OH 45221-0012, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Hollow glass microspheres coated with photocatalytic TiO2 (HGM-TiO2), recently became commercially
available and have the distinct advantages of easy separation and recovery after treatment. With this in
mind, we determined the optimum conditions for hydroxyl radical generation from HGM-TiO2 photocata-
lysis using response surface methodology (RSM). The hydroxyl radical yield and its average generation
rate are critical parameters for practical applications of TiO2 photocatalysis. In this study, terephthalic
acid was used as a hydroxyl radical trap because of the selective formation of the readily detectable
hydroxyl radical adduct, 2-hydroxy terephthalic acid. Three independent variables, including loading of
HGM-TiO2, concentration of terephthalic acid and irradiation time, were investigated. The 3D response
surface graphs of hydroxyl radical yield and average hydroxyl radical generation rate indicated that opti-
mum conditions of loading of HGM-TiO2, concentration of terephthalate acid and irradiation time were
8.0 g/L, 4.0 mM, and 20 min, respectively. Under these optimized conditions, we measured the photo-
catalysis employing HGM-TiO2 for the remediation of dimethyl phthalate (DMP), as a representative
compound for problematic phthalate acid esters. HGM-TiO2 photocatalysis leads to the rapid destruction
of DMP and there is a linear correlation between the DMP destruction and hydroxyl radical production.
The results of our study demonstrate RSM can be used to readily determine the optimal conditions for
hydroxyl radical production and the subsequent treatment of target compounds may be correlated to
the hydroxyl radical production during HGM-TiO2 photocatalysis.
Received 21 January 2013
Received in revised form 22 March 2013
Accepted 14 April 2013
Available online 23 April 2013
Keywords:
Advanced oxidation
Photocatalysis
FTIR
Response surface methodology
Dimethyl phthalate remediation
© 2013 Elsevier B.V. All rights reserved.
1. Introduction
oxygen species (ROS), including hydroxyl radical, hydrogen perox-
ide, singlet oxygen, and superoxide anion radical. Among the ROS,
Access to clean water is a global problem and one of the primary
causes of human health problems worldwide [1]. TiO2 photocata-
lysis is an attractive method for the purification of water, due to
its abilities to effectively degrade a tremendous variety of tox-
ins and pollutants [2]. The degradation processes are initiated at
the surface of TiO2 [3]. When TiO2 is photoexcited by photons
with energy equal to or greater than the band gap, an electron is
promoted from the valence band to the empty conduction band,
resulting in an electron–hole pair. Electron–hole pairs can recom-
bine or migrate to the surface and react with the adsorbed species
on TiO2 surface. The process can generate a variety of reactive
hydroxyl radical is generally responsible for the degradation dur-
ing UV TiO2 photocatalysis in aqueous solution [4,5]. The processes
initiated during TiO2 photocatalysis are represented in Eqs. (1)–(8).
Hydroxyl radical is capable of reacting with most organic com-
pounds and many inorganic compounds often at nearly diffusion
controlled rates. Hydroxyl radical is a powerful electrophile and
reacts with organic substances mainly by the addition to double
and triple bonds, and aromatic rings, hydrogen-atom abstraction
from C(sp3) H bonds, and electron transfer pathways [6]. Since the
performance of TiO2 photocatalysis for water treatment processes
is highly dependent on hydroxyl radical, it is critical to evaluate and
optimize experimental conditions for maximizing hydroxyl radical
generation during TiO2 photocatalysis.
Abbreviations: 2-HTA, 2-hydroxy terephthalic acid; ANOVA, analysis of variance;
DMP, dimethyl phthalate; FTIR, Fourier transform infrared spectroscopy; HGM-TiO2,
hollow glass microspheres coated with photocatalytic TiO2; ROS, reactive oxygen
species; RSM, response surface methodology; TA, terephthalic acid.
TiO2 + hꢀ → h+VB + eC−B
(1)
∗
•
Corresponding author. Tel.: +1 305 348 3968; fax: +1 305 348 3772.
E-mail address: osheak@fiu.edu (K.E. O’Shea).
−
e−CB + O2 → O2
(2)
1010-6030/$ – see front matter © 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.jphotochem.2013.04.008

8
W. Jiang et al. / Journal of Photochemistry and Photobiology A: Chemistry 262 (2013) 7–13
•
2O2 − + 2H2O → H2O2 + 2OH− + O2
(3)
(4)
(5)
(6)
HGM-TiO2, concentration of TA, and reaction time coded at five
levels.
The RSM results were used to guide the application of HGM-TiO2
H2O + h+
→
→
•OH + H+
VB
photocatalysis in the photocatalytic degradation of DMP, as a model
for problematic phthalate acid esters which have wide spread use
and an annual production of approximately 4 million tons [19].
The US Environmental Protection Agency and European Union have
classified these compounds as priority pollutants [20,21] because
of the significant threat they pose on reproductive and behavioral
health of humans and wildlife at low concentrations [22,23]. HGM-
TiO2 photocatalysis leads to the rapid destruction of DMP and there
is a linear correlation between the DMP destruction and hydroxyl
radical production. The results of our study demonstrate RSM can
be used to readily determine the optimal conditions for hydroxyl
radical production and the subsequent treatment of target com-
pounds can be correlated to the hydroxyl radical production during
HGM-TiO2 photocatalysis.
OH− + h+VB
•OH
H2O2 + e− → OH− + •OH
CB
TiO2 + hꢀ + 3O2 → TiO2
+
1O2
(7)
(8)
•OH + toxin → oxidation product
In an attempt to improve the performance of UV TiO2 photocata-
lysis, a variety of TiO2 materials have been developed, including
surface modification [7], TiO2 films [8], doped TiO2 [9], TiO2 nano-
tubes [10], porous TiO2 microspheres [11] and microspheric cores
covered with TiO2 shell/film [12]. Among these means, micro-
spheric cores covered with nano- or micro-sized shells are novel
fabricated composite materials and have received significant atten-
tion [13]. As an important composite material, HGM-TiO2, recently
became commercially available and has the major advantages of
easy separation and recovery. Thus, HGM-TiO2 is promising for
use in industrial wastewater treatment plants and gaseous pol-
lutants reduction, due to unique properties, such as low density
(0.22 g/cm3), buoyancy, and transparency to visible light. However,
detailed studies on the photocatalytic performance and applica-
tions of HGM-TiO2 have received limited attention.
2. Materials and methods
2.1. Chemicals
HGM-TiO2 material was obtained from Microsphere Tech-
nology Limited (Limerick, Ireland). The characterization
information of this material (including median diame-
ter, particle size) is available from the company website
http://www.microspheretechnology.com/photospheres.php.
TA (disodium salt) and DMP were purchased from Aldrich. HPLC
grade methanol was obtained from Fisher. 2-Hydroxy terephthalic
acid (2-HTA) was synthesized for calibration by using a published
method [24]. All the chemicals were used without further purifi-
cation and all solutions were made with Millipore filtered water
(18 Mꢁ cm).
An excellent study on the optimization of the experimental con-
ditions to produce the highest •OH yield during photocatalysis
using suspensions of Degussa P25 TiO2 has been reported [14], but
to the best of our knowledge there are no reports on optimized
conditions for HGM-TiO2 photocatalysis. The •OH generation rate is
also an important parameter, since a high generation rate results in
rapid degradation and short reaction times to achieve specific treat-
ment objectives. The specific experimental conditions are critical
to generation rate and yield of •OH during HGM-TiO2 photocata-
lysis. For example, as the concentration of catalyst increases, the
photodegradation efficiency can increase to a maximum at a spe-
cific catalyst loading above which light scattering and screening
may reduce the TiO2 photocatalytic efficiency. Terephthalic acid
(TA) was employed to trap •OH effectively and selectively. Although
quenching and inter filter effects may occur at high TA concentra-
tions, these issues are not significant under dilute concentrations
[15]. Another important factor is irradiation time. Since there are
significant costs associated with generation of UV light, it is impor-
tant to evaluate the treatment time required to achieve desired
levels of degradation. From our study, •OH yield increases, whereas
•OH generation rate decreases under extended irradiation time.
Photocatalytic deactivation can occur when the intermediate prod-
ucts compete for radical species (•OH) leading to inhibition of the
photocatalytic performance [16]. Therefore, it is critical to deter-
mine the optimal yield and rate of •OH generation, as a function of
loading of HGM-TiO2, concentration of TA and irradiation time.
The classical one-variable-at-a-time methodology does not
enable the study of combined effects of two or more variables
on a measured response. Probing each variable independently is
also labor intensive and time consuming. Thus, RSM was originally
developed by Box and Wilson [17], to access the interactions of
various variables simultaneously and provide an empirical descrip-
tion of effects of variables and their interactions on a measured
response. The RSM has successfully applied to determine the opti-
mal conditions for a variety of processes [18]. Herein, RSM is used
to optimize yield and average generation rate of •OH by HGM-TiO2
photocatalysis. A central composite design was used to investi-
gate the effects of three independent variables, namely loading of
2.2. Fourier transform infrared spectroscopy (FTIR)
The TA and DMP loaded HGM-TiO2 were prepared by adding
1.0 g HGM-TiO2 into 100 mL solution with 1.0 g TA or DMP. The sus-
pension was put on an orbit shaker at 300 RPM for half an hour. Solid
samples for FTIR were separated, and dried in a vacuum oven at
room temperature. FTIR was collected using Perkin Elmer Spectrum
100 FTIR spectrometer.
2.3. Photocatalytic and analytical methods
HGM-TiO2 suspension was prepared by suspending HGM-TiO2
into 100 mL TA aqueous solution in a Pyrex cylindrical reactor
(12 × 1 in., ∼150 mL capacity, with a vented Teflon screw top).
The suspension was magnetically stirred and purged with oxygen
gently for 15 min prior to radiation and during the reaction, in order
to maintain the adsorption/desorption equilibrium. The suspen-
sions were irradiated in a Rayonet photochemical reactor (Southern
New England Ultra Violet Company, http://www.rayonet.org,
model RPR-100), equipped with a cooling fan on the bottom and
four phosphor-coated low-pressure mercury lamps (RPR 350 nm,
8.34 × 10−9 Einstein mL−1 s−1). Samples (3 mL) were taken from
the suspension at given time intervals and immediately filtered
through a 0.45 ␮m PTFE filter to remove suspended particles prior
to analysis.
TA is used to selectively trap •OH and to produce 2-HTA with a
percent yield of 35% [25] (Scheme 1). The yield of •OH is quantified
by fluorescent measurement of the generated 2-HTA. 2-HTA was
excited at 315 nm to emit fluorescence at 425 nm [26], which was
measured on a Horiba FluoroMax 3 spectrofluorometer.

W. Jiang et al. / Journal of Photochemistry and Photobiology A: Chemistry 262 (2013) 7–13
9
100
90
80
70
60
50
Scheme 1. Reaction of TA with •OH.
40
30
20
10
HGM-TiO2
HGM-TiO2 + DMP
HGM-TiO2 + TA
Table 1
Central composite design for RSM.
Experiment entry
Variable in coded levels
3600
3100
2600
wavenumber (cm-1)
2100
1600
1100
600
x1
x2
x3
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
−1
−1
−1
−1
1
−1
−1
1
1
1
−1
1
Fig. 1. FTIR spectra of HGM-TiO2, TA- and DMP-loaded onto HGM-TiO2.
−1
1
1
defined by the following equation:
1
1
−1
1
1
0
−1
−1
0
1
Y = b0 = b1X1 = b2X2 + b3X3 + b12X1X2 + b13X1X3 + b23X2X3
+ b11X12 + b22X22 + b33X32
−1
˛
(10)
0
0
−˛
0
˛
0
where Y is the predicted response, b0 is the intercept, b1, b2 and b3
are linear coefficients, b12, b13 and b23 are squared coefficients, b11
b22 and b33 are quadratic coefficients.
0
˛
−˛
0
0
0
0
0
0
0
0
,
0
˛
0
0
0
0
0
2.5. Degradation and analysis of DMP
0
0
0
0
The optimal conditions for the generation of •OH were employed
for degradation of DMP. The concentration of DMP residual was
analyzed using a Varian ProStar HPLC system equipped with a
ProStar 410 autosampler and a ProStar 335 photodiode array
detector with the stationary phase being a Luna RP C18 column
of 250 mm × 4.6 mm I.D. (5 ␮m particle size). The mobile phase
consisted of a mixture of water (50%) and methanol (50%), and
detection wavelength was 230 nm. The flow rate was 1 mL min−1
and the injection volume is 30 ␮L at room temperature [28]. The
retention time of DMP is ∼12.2 min under these HPLC conditions.
0
0
0
0
0
2.4. Experimental design and data analysis
To identify best combination of experimental parameters for
optimal •OH generation, RSM was applied. Central composite
design was employed for the optimal conditions. A full factorial
5-level experimental design with 3 variables needs to run 53 = 125
experiments. However, as shown in Table 1, application of the cen-
tral composite design reduces the number of required experiments
to 20 (8 factor points, 6 axial points and 6 replications at the cen-
ter points) [27]. Alpha (˛) is the coded level of axial point from the
center. The coded level is (xi) defined by the following equation:
3. Results and discussion
3.1. FTIR
The FTIR spectra of were recorded in the range 600–4000 cm−1
to identify the functional groups of HGM-TiO2 and access the
adsorption of TA and DMP onto HGM-TiO2 (Fig. 1). For the neat
HGM-TiO2 sample, the infrared band at 3374 cm−1 shows the pres-
ence of OH stretching vibration [29], while the IR band at 1631 cm−1
is ascribed to Si H2O adsorption [30,31]. The peak at 1404 cm−1
corresponds to O H bending mode [32], the band at 1019 cm−1
Xi − X0
xi =
(9)
ꢂXi
where X0 is the real value of the independent variable at the center
point, Xi is the real value of the independent variable, and ꢂXi is
the step changing value.
In this study, the loading of HGM-TiO2 (X1), TA concentration
(X2) and irradiation time (X3) were varied. The real levels and
coded levels are showed in Table 2. The equation can quantitatively
describe the predicted response as a function of three variables
and the optima of three variables are obtained by surface response.
Herein, Y is the yield of •OH or its average generation rate, which is
corresponds to Si OH and Si
O Ti vibration modes [33] while
a Ti
O
Ti band appears at 787 cm−1. These assignments clearly
demonstrate the HGM-TiO2 contains characteristic bands associ-
ated with TiO2 and glass [34]. The DMP loaded HGM-TiO2 materials
exhibit a band at 2954 cm−1 corresponding to C(sp3) H stretch-
ing [35], a band at 1727 cm−1 is assigned to C O stretching [36],
a C O stretching band at 1289 cm−1, a band at 743 cm−1 indicative
of ortho-disubstituted benzene ring [37] and a band at 1434 cm−1
is assigned to C H group [38]. For TA loaded HGM-TiO2, the band
at 1555 cm−1 corresponds to antisymmetric CO2 stretching [39],
Table 2
Real and coded levels of three variables.
Variable
Symbol coded
Coded level
the band at 823 cm−1 is ascribed to
C H bending of an aromatic
−˛ (−2)
−1
0
1
˛ (2)
[HGM-TiO2] (g/L)
[TA] (mM)
Irradiation time (min)
X1
X2
X3
2
1
5
4
2
10
6
3
15
8
4
20
10
5
25
ring [40]. The remaining bands are ascribed to HGM-TiO2. The
results demonstrate adsorption of DMP and TA onto the HGM-TiO2
materials.

10
W. Jiang et al. / Journal of Photochemistry and Photobiology A: Chemistry 262 (2013) 7–13
Table 3
term influenced the predicted response at a significant confidence
level. For Eqs. (11) and (12), X1, X2, X3, X1X1, X2X2 and X3X3 were sig-
nificant terms for the predicted response, and X1X2, X1X3 and X2X3
were insignificant terms. The determination coefficient (R2) for •OH
yield was calculated as 0.982, which can explain the variability of
response at a 0.982 confidence level. Moreover, the adjusted R2
value of 0.966 was also close to 1. For Y2, R2 was 0.974 and adjusted
R2 was 0.951. Therefore, as well as the model for •OH yield, the
predicted values had a good agreement with experimental data.
The response surface models were further analyzed by ANOVA
and the output results are summarized in Table 5. The P value
of regression model was less than 0.05, so the models were ade-
quate to describe the variability of Y1 and Y2 as a function of X1,
X2 and X3. The linear and square effects were highly significant
for the predicted responses, whereas the interaction effects were
insignificant. In conclusion, both Eqs. (11) and (12) are good approx-
imations.
RSM central composite design and experimental and expected responses.
X1 X2 X3 Experimental Y1 Predicted Y1 Experimental Y2 Predicted Y2
6
6
6
2
10
6
6
4
4
8
8
4
4
8
8
6
6
6
6
6
3
3
3
3
3
1
5
2
2
2
2
4
4
4
4
3
3
3
3
3
5
95.8
99.4
176.5
221.6
114.1
192.4
114.7
171.1
103.1
152.1
133.3
188.3
116.3
183.5
158.3
231.5
176.5
176.5
176.5
176.5
176.5
19.17
11.78
9.13
7.20
13.47
7.67
11.63
10.70
7.60
13.15
8.90
18.28
11.80
9.73
7.39
13.02
7.63
11.34
11.11
7.36
14.08
9.25
15 176.7
25 228.2
15 108.0
15 202.1
15 115.0
15 174.4
10 107.0
20 152.0
10 131.5
20 178.0
10 120.9
20 179.5
10 152.6
20 221.9
15 175.9
15 177.4
15 176.9
15 176.0
15 177.5
12.09
8.97
12.57
8.85
15.26
11.09
11.72
11.83
11.79
11.73
11.83
16.30
11.50
11.80
11.80
11.80
11.80
11.80
To obtain the optimal conditions for •OH generation, the 3D
response surface technique was employed to evaluate the effects
of independent variables on Y1 and Y2. Fig. 2 shows the effect of
any two independent variables on the •OH yield and its average
•OH generation rate, keeping the coded level of the third one at its
central level (0).
3.2. Model fitting and 3D response surface
Fig. 2a, c, d and f demonstrates the effect of concentration of TA
on Y1 and Y2. Y1 and Y2 increase with increase in concentration of
TA until Y1 and Y2 reach a highest value. The optimal value of con-
centration of TA for Y1 and Y2 is 4.0 mM. The effect of loading of
HGM-TiO2 on Y1 and Y2 is shown in Fig. 2b, c, e and f. The optimal
value of concentration of HGM-TiO2 on Y1 and Y2 is 8.0 g/L. Fig. 2a,
b, d and e demonstrates the effect of irradiation time on Y1 and
Y2. •OH yield increases whereas the average •OH generation rate
decreases with increase in irradiation time. The competition for
hydroxyl radical between TA and the oxidation products and TA
consumption are factors leading to the observed decrease in •OH
generation rate. Although the overall •OH yields increase with irra-
diation time, the rate of •OH yield slowed significantly after 20 min.
Thus, the optimal time for Y1 and Y2 is 20 min. In summary, the opti-
mal conditions for the yield of •OH were 8.0 g/L HGM-TiO2, 4.0 mM
TA and 20 min for irradiation time.
The data of •OH yield and average rate of •OH generation are
shown in Table 3. The experimental data were fitted to empiri-
cal second-order polynomial models by a regression function in
Microsoft Excel 2007. The results indicate that the experimen-
tal and expected responses match well, and the experimental
responses fit the second-order polynomials well. The empirical
second-order polynomials were obtained as the following two
equations:
Y1 = −98.78 + 20.49X1 + 41.91X2 + 7.28X3 + 1.48X1X2
+ 0.15X1X3 + 0.91X2X3 = 1.45X12 − 8.39X22 − 0.16X32
(11)
Y2 = 6.71 + 2.02X1 + 3.81X2 − 0.93X3 + 0.096X1X2 − 0.027X1X3
+ 0.0015X2X3 − 0.10X12 − 0.58X22 + 0.022X32
(12)
where Y1 is the predicted response of yield of generated •OH (␮M),
Y2 is the predicted average generation rate of •OH (␮M/min), the
variables of X1, X2 and X3 are the loading of HGM-TiO2 (g/L), con-
centration of TA (mM) and irradiation time (min), respectively.
In order to check the adequacy of the second-order models, the
significance test and analysis of variance (ANOVA) were employed
for the second-order models for •OH yield and the average •OH
generation rate. The significance tests of estimated regression
coefficients for •OH yield and average •OH generation rate are
shown in Table 4. The probability value (P) of coefficients was
greater than 0.05, indicating the term did not have a significant
effect on the predicted response. Otherwise, it was rejected and this
3.3. The degradation of DMP
The optimal conditions for the •OH generation were applied for
the degradation of DMP. The initial concentration of DMP is 50 ␮M.
The loss of DMP follows the pseudo-first order kinetic model. The
pseudo-first order kinetic model can be expressed as [41]:
d[C]
dt
= −k[C]
(13)
After definite integration, the equation becomes
ꢀ
ꢁ
C0
C
ln
= kt
(14)
where k is the rate constant of pseudo-first order model (min−1),
t is time (min), C0 is the initial concentration of DMP (␮M) and C
is the concentration of DMP (␮M) at the specific time. Since R2 of
plot of ln C0/C versus t is greater than 0.95, the experimental data
fit the pseudo-first-order model nicely.
Table 4
Estimated regression coefficients for •OH yield and the average •OH generation rate.
Term
•OH yield
Average •OH generation rate
Coefficient
Std. error
P
Coefficient
Std. error
P
As shown in Fig. 3, the rate constants for degradation of DMP
increased with increase in loading of HGM-TiO2, and reached the
highest at 8.0 g/L HGM-TiO2. No significant increase is observed at
loading levels above 8.0 g/L HGM-TiO2. At high loadings of HGM-
TiO2 (>8.0 g/L), the rate does not change but the standard deviation
becomes larger due to the light scattering and screening effects.
Therefore, the loading of HGM-TiO2 = 8.0 g/L was considered opti-
mal. In order to assess the relationship between •OH yield and
X1
X2
X3
X1X2
X1X3
X2X3
X1X1
X2X2
X3X3
20.49
41.91
7.28
1.48
0.15
6.71
13.43
2.69
1.23
0.24
0.49
0.35
1.38
0.055
0.012
0.011
0.022
0.256
0.550
0.093
0.002
0.000
0.015
2.02
3.81
0.60
1.19
0.24
0.11
0.022
0.044
0.031
0.12
0.007
0.010
0.003
0.398
0.236
0.973
0.008
0.001
0.001
−0.93
0.096
−0.027
0.0015
−0.10
−0.58
0.022
0.91
−1.45
−8.39
−0.16
0.0049

W. Jiang et al. / Journal of Photochemistry and Photobiology A: Chemistry 262 (2013) 7–13
11
Table 5
ANOVA for the second-order models of •OH yield and average •OH generation rate.
Source
•OH yield
DFa
Average •OH generation rate
Sum of squares
F
P
DF
Sum of squares
F
P
Regression
Linear
Square
Interaction
Residual error
Total
9
3
3
3
10
19
26446.4
23932.0
2260.3
254.0
481.8
26928.2
60.99
5.02
15.64
1.76
0.000
0.022
0.000
0.219
9
3
3
3
10
19
144.03
118.47
24.65
0.90
3.80
147.83
42.08
19.02
21.61
0.79
0.000
0.000
0.000
0.526
a
DF is degree of freedom.
degradation of DMP, the plot of degraded DMP against •OH yield
is demonstrated in Fig. 4. The •OH yield derived from Eq. (11) with
the coded level of X2 at the central level (0) and X3 being 20 min,
showing increased •OH yield as the increasing of loading of HGM-
TiO2 (2–10 g/L). After 20 min, DMP was eliminated in presence of
≥8.0 g/L HGM-TiO2, in agreement with the optimal conditions of
•OH generation. The linear relationship between •OH yield and
DMP degradation indicates the •OH is responsible for degradation
of DMP. The bimolecular reaction rate constants for the reaction of
•OH with TA is 3.3 × 109 M−1 s−1 [42], while it is 4 × 109 M−1 s−1 for
DMP [43]. The rate constants are similar; however the trend-line in
Fig. 4 demonstrates the destruction of DMP correlates to the •OH
yield, but not in a 1:1 ratio. The yield of •OH and TA adduct versus
destruction of DMP may be related to the different adsorption prop-
erties of DMP versus TA on HGM-TiO2. Since •OH is a relatively
nonselective radical and is able to react with substance at nearly
diffusion controlled rate, the oxidation products of DMP may be
degraded further before they are released back to aqueous phase.
Fig. 2. Response surface showing of any two independent variables on the •OH yield (a–c) and its average •OH generation rate (d–f).

12
W. Jiang et al. / Journal of Photochemistry and Photobiology A: Chemistry 262 (2013) 7–13
Acknowledgement
0.24
0.2
K.E.O. and D.D.D. gratefully acknowledge NSF support (CBET,
award 1033317).
y = 0.018x + 0.023
R² = 0.98
References
0.16
0.12
0.08
0.04
[1] E.K. O’Shea, D.D. Dionysiou, Advanced oxidation processes for water treatment,
Journal of Physical Chemistry Letters 3 (2012) 2112–2113.
[2] M.R. Hoffmann, S.T. Martin, W. Choi, D.W. Bahnemann, Environmental appli-
cation of semiconductor photocatalysis, Chemical Reviews 95 (1995) 69–96.
[3] M.A. Fox, M.T. Dulay, Heterogeneous photocatalysis, Chemical Reviews 93
(1993) 341–357.
[4] A.L. Linsebigler, G. Lu, J.T. Yates, Photocatalysis on TiO2 surfaces: principles,
mechanisms, and selected results, Chemical Reviews 95 (1995) 735–758.
[5] U. Khan, N. Benabderrazik, A.J. Bourdelais, D.G. Baden, K. Rein, P.R. Gardinali,
L. Arroyo, K.E. O’Shea, UV and solar TiO2 photocatalysis of brevetoxins (PbTxs),
Toxicon 55 (2010) 1008–1016.
[6] G.V. Buxton, C.L. Greenstock, W.P. Helman, A.B. Ross, Critical review of rate
constants for reactions of hydrated electrons, hydrogen atoms and hydroxyl
radicals (•OH/•O−) in aqueous solution, Journal of Physical and Chemical Ref-
erence Data 17 (1988) 513–886.
[7] Y. Kado, R. Hahn, P. Schmuki, Surface modification of TiO2 nanotubes by low
temperature thermal treatment in C2H2 atmosphere, Journal of Electroanalyt-
ical Chemistry 662 (2011) 25–29.
[8] C. Yogi, K. Kojima, N. Wada, H. Tokumoto, T. Takai, T. Mizoguchi, H. Tamiaki,
Photocatalytic degradation of methylene blue by TiO2 film and Au particles-
TiO2 composite film, Thin Solid Flims 516 (2008) 5881–5884.
[9] D. Graham, H. Kisch, L.A. Lawson, P.K.J. Robertson, The degradation of
microcystin-LR using doped visible light absorbing photocatalysts, Chemo-
sphere 78 (2010) 1182–1185.
0
4
8
12
16
[HGM-TiO2] (g/L)
Fig. 3. The rate constants of pseudo-first order kinetic model for degradation of
DMP as a function of HGM-TiO2 loading. The data at 10, 12 and 15 g/L were not used
for the trend-line since the rate did not increase above 8.0 g/L. Error bars represent
standard deviation of triplicate experiments. The insert is the molecular structure
of DMP.
55
[10] Z. Liu, X. Zhang, S. Nishimoto, M. Jin, D.A. Tryk, T. Murakami, A. Fujishima,
Highly ordered TiO2 nanotube arrays with controllable length for photoelec-
trocatalytic degradation of phenol, Journal of Physical Chemistry C 112 (2008)
253–259.
[11] C. Wang, X. Zhang, H. Liu, X. Li, W. Li, H. Xu, Reaction kinetics of photocat-
alytic degradation of sulfosalicylic acid using TiO2 microspheres, Journal of
Hazardous Materials 163 (2009) 1101–1106.
[12] G. Li, R. Bai, X.S. Zhao, Coating of TiO2 thin films on the surface of SiO2
microspheres: toward industrial photocatalysis, Industrial and Engineering
Chemistry Research 47 (2008) 8228–8232.
[13] N.B. Jackson, C.M. Wang, Z. Luo, J. Schwitzgebel, J.G. Ekerdt, J.R. Brock, A. Heller,
Attachment of TiO2 powders to hollow glass microbeads: activity of the TiO2-
coated beads in the photoassisted oxidation of ethanol to acetaldehyde, Journal
of the Electrochemical Society 138 (1991) 3660–3664.
[14] S.A.V. Eremia, D. Chevalier-Lucia, G. Radu, J. Marty, Optimization of hydroxyl
radical formation using TiO2 as photocatalyst by response surface methodol-
ogy, Talanta 77 (2008) 858–862.
[15] K. Ishibashi, A. Fujishima, T. Watanabe, K. Hashimoto, Detection of active
oxidative species in TiO2 photocatalysis using the fluorescence technique, Elec-
trochemistry Communications 2 (2000) 207–210.
[16] L. Cao, Z. Gao, S.L. Suib, T.N. Obee, S.O. Hay, J.D. Freihaut, Photocatalytic oxi-
dation of toluene on nanoscale TiO2 catalysts: studies of deactivation and
regeneration, Journal of Catalysis 196 (2000) 253–261.
[DMP]0 = 50 µM
20 min reaction time
50
45
40
35
135
155
175
195
215
·OH yield (µM)
Fig. 4. Plot of degraded DMP against •OH yield. Error bars represent standard devi-
ation of triplicate experiments.
[17] G.E.P. Box, K.B. Wilson, On the experimental attainment of optimum conditions,
Journal of the Royal Statistical Society 13 (1951) 1–45.
[18] I. Arslan-Alaton, G. Tureli, T. Olmez-Hanci, Treatment of azo dye production
wastewaters using Photo-Fenton-like advancedoxidation processes: opti-
mization by response surface methodology, Journal of Photochemistry and
Photobiology A 202 (2009) 142–153.
[19] Z.-P. Lin, M.G. Ikonomou, H. Jing, C. Mackintosh, F.A.P.C. Gobas, Determination
of phthalate ester congeners and mixtures by LC/ESI-MS in sediments and biota
of an urbanized marine inlet, Environmental Science and Technology 37 (2003)
2100–2108.
[20] US Environmental Protection Agency (US EPA), National Primary Drinking
Water Regulation: Federal Register, 40 CFR, Part 141, Washington, DC, July 1,
1991 (Chapter I).
[21] B.G. Hansen, A.G. van Haelst, K. van Leeuwen, P. van der Zandt, Priority setting
for existing chemicals: European Union risk ranking method, Environmental
Toxicology and Chemistry 18 (1999) 772–779.
[22] M. Matsumoto, M. Hirata-Koizumi, M. Ema, Potential adverse effects of phthalic
acid esters on human health: a review of recent studies on reproduction, Reg-
ulatory Toxicology and Pharmacology 50 (2008) 37–49.
[23] J.A. Dean, Lange’s Handbook of Chemistry (eBook), 15th ed., McGraw-Hill, New
York, 1999.
[24] T.J. Mason, J.P. Lorimer, D.M. Bates, Y. Zhao, Dosimetry in sonochemistry: the
use of aqueous terephthalate ion as a fluorescence monitor, Ultrasonics Sono-
chemistry 1 (1994) S91–S95.
4. Conclusions
The FTIR spectra show that the HGM-TiO2 exhibits adsorp-
tion of TA (•OH trap) and DMP (target compound). The loading of
HGM-TiO2, the concentration of TA and irradiation time affect •OH
generation significantly. The optimum conditions of •OH genera-
tion by HGM-TiO2 were measured using RSM. Statistical analyses
indicate the empirical second-order polynomials can accurately
describe the •OH yield and the average •OH generation rate. The
3D response surface graphs showed that the optimum conditions of
loading of HGM-TiO2, concentration of TA and irradiation time were
8.0 g/L, 4.0 mM, and 20 min, respectively. The optimal condition for
•OH generation by HGM-TiO2 has applied to the degradation of
DMP. The degradation of DMP follows the pseudo-first order kinetic
model nicely, and rate constant increased linearly as increasing of
loading of HGM-TiO2 up to 8.0 g/L. We have demonstrated RSM can
be used to determine the optimal conditions for •OH generation.
Employing the optimal conditions, the problematic pollutant, DMP,
is readily degraded. The effective application of RSM for determin-
ing optimal conditions for hydroxyl radical production and rapid
destruction of DMP show HGM-TiO2 photocatalysis is a promising
material for water treatment.
[25] X. Fang, G. Mark, C. von Sonntag, OH radical formation by ultrasound in aque-
ous solutions. Part I: the chemistry underlying the terephthalate dosimeter,
Ultrasonics Sonochemistry 3 (1996) 57–63.
[26] K. Ishibashi, A. Fujishima, T. Watanabe, K. Hashimoto, Quantum yields of active
oxidative species formed on TiO2 photocatalyst, Journal of Photochemistry and
Photobiology A 134 (2000) 139–142.

W. Jiang et al. / Journal of Photochemistry and Photobiology A: Chemistry 262 (2013) 7–13
13
[27] N.B. Parilti, C.S.U. Demirel, M. Bekbolet, Response surface methodological
approach for the assessment of the photocatalytic degradation of NOM, Journal
of Photochemistry and Photobiology A 225 (2011) 26–35.
[36] B.N. Khare, M. Meyyappan, A.M. Cassell, C.V. Nguyen, J. Han, Functionaliza-
tion of carbon nanotubes using atomic hydrogen from a glow discharge, Nano
Letters 2 (2002) 73–77.
[28] Y. Chen, L. Chen, N. Shang, Photocatalytic degradation of dimethyl phthalate in
an aqueous solution with Pt-doped TiO2 coated magnetic PMMA microspheres,
Journal of Hazardous Materials 172 (2009) 20–29.
[37] C.F. Liu, R.C. Sun, J. Ye, Structural and thermal characterization of sugarcane
Bagasse phthalates prepared with ultrasound irradiation, Polymer Degradation
and Stability 91 (2006) 280–288.
[29] M. Burgos, M. Langlet, The sol–gel transformation of TIPT coating: a FTIR study,
Thin Solid Films 349 (1999) 19–23.
[30] J. Wang, T. Ma, Z. Zhang, X. Zhang, Y. Jiang, W. Sun, R. Li, P. Zhang, Investigation
on the transition crystal of ordinary rutile TiO2 powder by microwave irradi-
ation in hydrogen peroxide solution and its sonocatalytic activity, Ultrasonics
Sonochemistry 14 (2007) 575–582.
[31] B. Ding, H. Kim, C. Kim, M. Khil, S. Park, Morphology and crystalline phase study
of electrospun TiO2–SiO2 nanofibres, Nanotechnology 14 (2003) 532–537.
[32] B.P. Singh, S. Nayak, S. Samal, S. Bhattacharjee, L. Besra, The role of
poly(methacrylic acid) conformation on dispersion behavior of nano TiO2 pow-
der, Applied Surface Science 258 (2012) 3524–3531.
[33] J. Zhai, T. Yang, L. Zhang, X. Yao, The optical waveguiding properties of
TiO2–SiO2 composite films prepared by the sol–gel process, Ceramics Inter-
national 25 (1999) 667–670.
[38] L.M. Matuana, D.P. Kamdem, J. Zhang, Photoaging and stabilization of rigid
PVC/wood–fiber composites, Journal of Applied Polymer Science 80 (2001)
1943–1950.
[39] L. Liao, C. Lien, D. Shieh, M. Chen, J. Lin, FTIR study of adsorption and pho-
toassisted oxygen isotopic exchange of carbon monoxide, carbon dioxide,
carbonate, and formate on TiO2, Journal of Physical Chemistry B 106 (2002)
11240–11245.
[40] D.L. Pavia, G.M. Lampman, G.S. Kriz, J.A. Vyvyan, Introduction to Spectroscopy,
4th ed., Brooks/Cole, Belmont, CA, 2009.
[41] S. Ye, Q. Tian, X. Song, S. Luo, Photoelectrocatalytic degradation of ethylene by a
combination of TiO2 and activated carbon felts, Journal of Photochemistry and
Photobiology A 208 (2009) 27–35.
[42] W. Song, S. Yan, W.J. Cooper, D.D. Dionysiou, K.E. O’Shea, Hydroxyl radical
oxidation of cylindrospermopsin (cyanobacterial toxin) and its role in the pho-
tochemical transformation, Environmental Science and Technology 46 (2012)
12608–12615.
[43] W.R. Haag, C.C.D. Yao, Rate constants for reaction of hydroxyl radical with sev-
eral drinking water contaminants, Environmental Science and Technology 26
(1992) 1005–1013.
[34] M. Minella, M.G. Faga, V. Maurino, C. Minero, E. Pelizzetti, S. Coluccia, G. Martra,
Effect of fluorination on the surface properties of titania P25 powder: an FTIR
study, Langmuir 26 (2010) 2521–2527.
[35] J. Ibarra, E. Mun˜oz, R. Moliner, FTIR study of the evolution of coal structure
during the coalification process, Organic Geochemistry 24 (1996) 725–735.