Photocatalytic degradation of ciprofloxacin by synthesized TiO2 nanoparticles on montmorillonite: Effect of operation parameters and artificial neural network modeling

Article abstract of DOI:10.1016/j.molcata.2015.08.020

TiO₂/MMT nanocomposite was synthesized and characterized by X-ray diffraction (XRD), fourier transform infrared spectroscopy (FTIR), scanning electron microscopy (SEM), Energy-dispersive X-ray spectroscopy (EDX), transmission electron microscopy (TEM), X-ray fluorescence (XRF) and Brunauer-Emmett-Teller (BET) techniques. The average size of TiO₂ nanoparticles was decreased from 60-80 nm to 40-60 nm through the immobilization on MMT. The main influential factors such as the TiO₂/MMT dose, ciprofloxacin (CIP) concentration, pH of the solution, UV light regions, reusability of the catalyst and electrical energy determination were studied. The addition of radical scavengers (e.g. chloride, iodide, sulfate and bicarbonate) and enhancers (e.g. hydrogen peroxide, potassium iodate and peroxydisulfate) on the degradation efficiency was studied. The predicted data from the designed artificial neural network model were found to be in a good agreement with the experimental data (R² = 0.9864). The main intermediates of CIP degradation were determined by GC-Mass spectrometry.

Full text of DOI:10.1016/j.molcata.2015.08.020

Journal of Molecular Catalysis A: Chemical 409 (2015) 149–161
Contents lists available at ScienceDirect
Journal of Molecular Catalysis A: Chemical
journal homepage: www.elsevier.com/locate/molcata
Photocatalytic degradation of ciprofloxacin by synthesized TiO2
nanoparticles on montmorillonite: Effect of operation parameters and
artificial neural network modeling
Aydin Hassania, Alireza Khataeeb,∗, Semra Karacaa,∗∗
a Department of Chemistry, Faculty of Science, Atatürk University, 25240 Erzurum, Turkey
b Research Laboratory of Advanced Water and Wastewater Treatment Processes, Department of Applied Chemistry, Faculty of Chemistry, University of
Tabriz, 51666-16471 Tabriz, Iran
a r t i c l e i n f o
a b s t r a c t
Article history:
TiO2/MMT nanocomposite was synthesized and characterized by X-ray diffraction (XRD), fourier
transform infrared spectroscopy (FTIR), scanning electron microscopy (SEM), Energy-dispersive X-
ray spectroscopy (EDX), transmission electron microscopy (TEM), X-ray fluorescence (XRF) and
Brunauer–Emmett–Teller (BET) techniques. The average size of TiO2 nanoparticles was decreased from
60–80 nm to 40–60 nm through the immobilization on MMT. The main influential factors such as the
TiO2/MMT dose, ciprofloxacin (CIP) concentration, pH of the solution, UV light regions, reusability of
the catalyst and electrical energy determination were studied. The addition of radical scavengers (e.g.
chloride, iodide, sulfate and bicarbonate) and enhancers (e.g. hydrogen peroxide, potassium iodate and
peroxydisulfate) on the degradation efficiency was studied. The predicted data from the designed artifi-
cial neural network model were found to be in a good agreement with the experimental data (R2 = 0.9864).
The main intermediates of CIP degradation were determined by GC–Mass spectrometry.
Received 4 February 2015
Received in revised form 13 August 2015
Accepted 20 August 2015
Available online 24 August 2015
Keywords:
Ciprofloxacin
Nanocatalyst
Pharmaceuticals
Photocatalysis
TiO2/MMT nanocomposite
© 2015 Elsevier B.V. All rights reserved.
1. Introduction
literature, most of these pharmaceuticals, because of their high
resistance to conventional wastewater treatments methods, are
Currently, due to their remarkable utilization and hazardous
biological and ecotoxicological effects, personal care products
(PCPs) and pharmaceuticals can be regarded as one of the major
groups of water contaminants. It is known that pharmaceuticals
and their metabolites get into wastewater treatment plants (WTP);
however, it should be noted that some of them may not be com-
pletely removed or transformed during conventional wastewater
treatment methods. Wastewaters from industries, hospitals and
domestic sites effluents are directly thrown into the surface and
ground waters without any previous decontamination treatment.
This can result in pollution and disturbing effects [1,2]. Recently,
Voogt et al. [3] have proposed a list of the most important 44 phar-
maceuticals related to the water cycle on the basis of their toxicity,
physicochemical properties, occurrence, consumption persistence
and resistance to treatment. As can be observed in the related
persistent in water cycles and therefore, become omnipresent in
the environment. Therefore, the effect of these pharmaceuticals and
their metabolites on the environment, their genotoxicity and acute
toxicity have been the subject of serious scientific research [4].
According to the previous studies in the field, advanced oxidation
processes (AOPs) are well-accepted technologies for the removal
of a wide variety of persistent organic pollutants from water and
wastewater [5–8]. Titanium dioxide (TiO2) is a predominant pho-
tocatalyst for environmental application due to its exceptional
properties such as higher photocatalytic oxidation ability, nonpho-
tocorrosiveness, chemical and biological inertness, nontoxicity and
low cost characteristics [9–11]. However, there are still some inher-
ent drawbacks such as small specific surface area, low adsorption
ability, difficult recovery and the agglomeration phenomenon of
suspended powder TiO2 at high loadings, all limiting its practical
application in the wastewater treatment [12]. This problem can be
tackled through the immobilization of TiO2 on the surface of sup-
ported materials without the loss of activity. Consequently, various
kinds of supports, such as silica, alumina, glass plate, zeolite and
clay, have been widely employed used to immobilize the catalysts
[13–17]. As can be understood from the literature, clay and clay
∗
Corresponding author. Fax: +98 41 33340191.
Corresponding author. Fax: +90 442 2360948.
∗∗
E-mail addresses: a khataee@tabrizu.ac.ir, ar khataee@yahoo.com (A. Khataee),
semra karaca@yahoo.com, skaraca@atauni.edu.tr (S. Karaca).
http://dx.doi.org/10.1016/j.molcata.2015.08.020
1381-1169/© 2015 Elsevier B.V. All rights reserved.

150
A. Hassani et al. / Journal of Molecular Catalysis A: Chemical 409 (2015) 149–161
Table 1
Characteristics of the ciprofloxacin (CIP).
Chemical structure
Molecular formula
Mw (g mol−1
)
ꢀmax(nm)
Therapeutic group
Antibiotic
Solubility in water (mg mL−1
)
C17H18FN3O3
331.346
276
30
based photocatalysts are regarded as promising and appropriate
materials for this purpose because they are chemically inert, resis-
tant to deterioration, inexpensive and commercially available in
larger quantities [14,18]. In recent years, clays such as montmoril-
lonite, rectorite and kaolinite have been of great interest for their
applicability in wastewaters decontamination. It is known that clay
minerals possess layer structures, high porosity, large surface areas,
exchangeable cations and swellable properties [19,20]. The previ-
ous studies in the field have revealed that dispersion TiO2 particles
into layered clays (TiO2 pillared clay) and/or the immobilization
of TiO2 on the surface of MMT can improve catalyst efficiency
as such composite structures are known to stabilize TiO2 parti-
cles and keep most of the surface of TiO2 crystals accessible to
various molecules [14,21]. Ciprofloxacin (CIP), which is a second
generation fluoroquinolone, is an antibacterial agent that can be
very persistent in aquatic environments. The chemical name of CIP
is 1-cyclopropyl-6-fluoro-1, 4-dihydro-4-oxo-7-(1-piperazinyl)-3-
quinolinecarboxylic acid. Because of the poor biodegradability of
CIP, the removal of this pollutant by chemical oxidation has been
attempted. In recent years, degradation of CIP has been investigated
Sigma–Aldrich Co. (USA). All other chemicals and reagents, being
of analytical grade, were purchased from Merck (Germany) and
used without further purification. Ciprofloxacin was supplied by
Farabi pharmaceutical company (Iran). The specifications of CIP
are shown in Table 1.
2.2. Synthesis of TiO2/MMT nanocomposite and characterization
TiO2/MMT nanocomposite was prepared through the synthe-
sis of TiO2 nanoparticles on the surface of the MMT. Fig. 1 shows
the schematic of the preparation method carried out in two stages.
Pure TiO2 nanoparticles were synthesized by the same procedure
without the addition of MMT.
The XRD spectrum of the synthesized TiO2/MMT nanocompos-
ite was obtained using a PANalytical X’Pert PRO diffractometer
(Germany) with Cu–Ka radiation (45 kV, 40 mA, 0.15406 nm). The
chemical composition of the MMT and TiO2/MMT was determined
by X-ray fluorescence (XRF) (Rigaku ZSX Primus II, Japan). Scanning
electron microscope (SEM) model MIRA3 FEG-SEM Tescan (Czech)
was used to detect the morphology and particle sizes of the sam-
ples. The TiO2, MMT and TiO2/MMT samples were analyzed with
a Fourier transform infrared spectroscopy (FTIR) model Tensor 27,
Bruker (Germany), in a wavenumber range of 4000–400 cm−1 using
the KBr pellet technique. The transmission electron microscopy
(TEM) image was conducted using a Cs-corrected high-resolution
TEM (Zeiss-EM10C, Germany) that operated at 100 kV. For this anal-
ysis, the synthesized TiO2/MMT sample was dispersed in ethanol
using ultrasonic vibration (Bandelin Sonorex, Germany) for 15 min;
then a drop of the dispersed sample was placed on a copper
grid coated with a layer of amorphous carbon to record the TEM
image. Textural properties of the MMT, TiO2 and TiO2/MMT sam-
ples were determined from N2 adsorption–desorption isotherms at
77 K on a Gemini 2385 nitrogen adsorption apparatus (Micromerit-
ics Instruments, USA) and their pore structure was analyzed using
Brunauer–Emmett–Teller (BET) and Barrett–Joyner–Halenda (BJH)
methods. Particle size distributions were conducted using Malvern
Mastersizer 2000 (UK).
using photo-Fenton [22], sonolysis [23], UV/S2O2− [24], UV/H2O2
8
[25] and adsorption [26] processes. TiO2 nanoparticles were syn-
thesized and immobilized on the surface of Montmorillonite K10
(MMT) in the present study for photocatalytic degradation of
CIP as the target pollutant. To this aim, X-ray diffraction (XRD),
Fourier transform infrared spectroscopy (FTIR), scanning electron
microscopy (SEM), Energy-dispersive X-ray spectroscopy (EDX),
transmission electron microscopy (TEM), X-ray fluorescence (XRF)
and Brunauer–Emmett–Teller (BET) were employed to character-
ize the as-prepared nanocomposite. Accordingly, CIP was used as
the model organic pollutant to evaluate the photocatalytic perfor-
mance of TiO2/MMT nanocomposite during UV-C radiation. To the
best of our knowledge and on the basis of the literature review,
it can be understood that the use of hydrothermal method for the
preparation of TiO2/MMT nanocomposite and its potential for the
photocatalytic degradation of CIP have not been investigated yet.
More clearly, this study addressed the effect of various parame-
ters such as catalyst dosage, the initial CIP concentration, the effect
of UV light region, the initial pH, the presence of various radical
scavengers and process enhancers on the promoted photocatalytic
process. Additionally, artificial neural network (ANN) was utilized
to model the photocatalytic process.
2.3. Photocatalytic degradation setup and procedure
The experimental set-up for the photocatalytic degradation of
CIP by immobilized TiO2 nanoparticles on MMT is schematically
shown in Fig. 2. Photocatalysis of CIP was performed in a batch
photoreactor with a 500 mL working volume. A 16 W UV-A, UV-
B or UV-C lamp (Sylvania, Japan) was applied as the light source.
Batch studies were performed to evaluate the effect of CIP concen-
tration, TiO2/MMT dose, the initial pH, UV light region, different
scavengers and enhancers on degradation efficiency. For each pho-
tocatalytic experiment, 500 mL of an aqueous solution containing
CIP in the range of 5–25 mg L−1 with 0.025–0.150 g L−1 of the
TiO2/MMT nanocomposite was added in the reaction vessel. The
UV-A, UV-B or UV-C lamp was turned on at the beginning of each
experiment. The pH of the solution was set to the desired value
2. Experimental
2.1. Materials
Montmorillonite
K10
(MMT)
was
purchased
from
Sigma–Aldrich Co. (USA). The cation exchange capacity (CEC)
of the clay (120 meq/100 g) was determined by the ammo-
nium acetate method [27]. Cetyltrimethylammonium bromide
(CTAB) was purchased from Sigma–Aldrich Co. (USA). Tetraethyl
orthotitanate (TEOT) is a precursor of Ti was purchased from

A. Hassani et al. / Journal of Molecular Catalysis A: Chemical 409 (2015) 149–161
151
Ti-precursor
TEOT (aq.)
Reaction medium
Distilled water
Template
CTAB
Distilled water
Pristine clay
Mixing for 24 h
(Swelling clay)
Mixing for 10 min.
TEOT/distilled water: 2.1/80 (volume ratio)
Drop wise addition of CTAB and mixing
for 1h (Adjusted pH=11-12)
Addition of swelling clay and mixing for 1 h
Modified MMT formation
(a)
(b)
Hydrothermal synthesis in a stainless steel autoclave reactor at 160 °C for 24 h
Washing with distilled water and ethanol (3 times)
Drying at 80 °C for 24 h under air flow
TiO2/MMT nanocomposite
Fig. 1. A flow chart used for the preparation of the TiO2/MMT nanocomposite: (a) precursor preparation and (b) nanocomposite forming.
with H2SO4 and NaOH (0.1 M). The solution pH was regulated with
a Mettler Toledo pH meter (China). 5 mL samples were taken from
the reactor at different intervals and residuals concentration of CIP
was measured using Varian Cary 100 UV–vis spectrophotometer
(Australia) at the maximum wavelength of 276 nm. Degradation
efficiency (%) = [(C0–Ct)/C0] × 100 was used to determine the degra-
dation of CIP, where C0 was the initial concentration of CIP solution
and Ct was its concentration after a certain time (t). Adsorption
experiments were conducted using the method applied for photo-
catalytic degradation experiment without UV radiation. The zero
point of charge (pHzpc) of the TiO2/MMT nanocomposite was mea-
sured using the method described by Bessekhouad et al. with some
modifications [28]. In this approach, 500 mL 0.01 M NaCl was pre-
pared and divided into eight solutions with the pH ranging from 3 to
10. 0.2 g of the catalyst was added to each solution. Finally, the final
pH of each solution was measured after 48 h shaking and plotted
against the initial pH to determine the pHzpc of the catalyst. Elec-
trophoretic mobility and the surface charge of the samples were
carried out with the relation of the zeta potential measurements.
Zeta potentials of the samples were measured using a Zeta Meter
3.0+ (Zeta-Meter, Inc., USA). The zeta potential of pure MMT, TiO2
and TiO2/MMT was obtained to be −30.7, +21.06 and −16.4 mV,
respectively.
of ANN in each layer and the transfer functions between the layers
form neural network topology were taken in to account [30]. The
linear transfer function was also utilized in the hidden and output
layers. The input layer was in accordance with four experimental
parameters including the initial CIP concentration (mg L−1), irra-
diation time (min), catalyst dosage (g L−1) and the initial pH. The
output layer, on the other hand, was the degradation efficiency (%).
All data obtained from the experiments (Xi) were scaled in 0.1–0.9
2.4. ANN modeling
It can be inferred from the previous studies that the artificial
neural networks (ANN) have been widely employed to investigate
the relative importance of parameters during photocatalytic degra-
dation process [29]. ANN modeling was employed in this study
using ANN toolbox of Matlab (MathWorks, Ine., USA) mathematical
software. Accordingly, a three-layer feed-forward back propagation
neural network which consisted of input, output and intermediate
hidden layers was used for modeling the photocatalytic degrada-
tion of CIP. The layers, the number of neurons as the processing unit
Fig. 2. Schematic representation of the reactor used for different processes.

152
A. Hassani et al. / Journal of Molecular Catalysis A: Chemical 409 (2015) 149–161
range (Xi) using Eq. (1), and then divided into training, validation
and test sets with a ratio of 72%, 24% and 24%, respectively.
ꢀ
ꢁ
Xi−min(Xi)
xi = 0.6
+ 0.2
(1)
max(Xi) − min(Xi)
where min(Xi) and max(Xi) were the extreme values of the input
variables X. The mean square error (MSE) was used as the error
function [29]. At the same time, the relative importance of each
input variable on the output variable was calculated by Eq. (2).
ꢂ
m=N
ꢃ
h
N
k=1
ꢂ
jh
((|W |/
jm
ih
i
ho
|W |) × |W |)
mn
km
m=1
Ij =
(2)
ꢂ
k=N
k=1
N
ꢃ
i
i
m=N
ho
{
h (|Wikhm|/
|Wih |) × |W |)}
mn
km
m=1
k=1
where Ij was the relative importance of jth input variable on the
output variable and Ni and Nh were the numbers of input and hid-
den neurons, respectively. W is the connection weight, and the
superscripts i, h and o refer to the input, hidden and output layers,
respectively. Also, subscript k, m and n refer to the input, hidden
and output neurons, respectively.
3. Results and discussion
3.1. Structural analysis of TiO2/MMT nanocomposite
Table 2 shows the chemical composition analysis of the pure
MMT and TiO2/MMT nanocomposite. As can be seen, most of the
material in MMT consisted of SiO2 (66.9 wt.%) followed by Al2O3
(13.8 wt.%). Titanium oxide was also observed in low content.
According to the obtained results, the TiO2 content of 58.5 wt.%
in the TiO2/MMT nanocomposite was higher than that in the
pure MMT. It was confirmed by chemical composition analysis
that a considerable amount of TiO2 species was introduced and
hybridized with MMT. Fig. 3(a and b) shows X-ray diffraction
patterns recorded for the raw montmorillonite and TiO2/MMT
nanocomposite, respectively. The MMT clay depicts the diffraction
peak at 2ꢁ of 19.9◦, 35.0◦, 53.9◦ and 61.7◦ and the reflections
at 2ꢁ of 20.9◦, 26.7◦, 36.5◦, 42.4◦, 45.8◦, 50.2◦, 59.9◦ and 68.1◦
which are attributed to the quartz phase (see Fig. 3a) [31–33].
As shown in Fig. 3b, the basal spacing of montmorillonite was
changed by the intercalation of TiO2 nanoparticles. The d00l peaks
were changed and new peaks related to the basal spacing were
observed around at 2ꢁ of 8.5◦ and 17.50◦. This clearly indicated
the enlargement of the basal spacing of the clay as a result of the
interaction between MMT and TiO2 nanoparticles. Additionally,
it was shown that TiO2 nanoparticles were present within the
interlayer space of the MMT clay. On the other hand, as evidenced
in Fig. 3b, the diffraction peaks at 2ꢁ of 8.5◦, 17.5◦, 19.5◦, 35.0◦ and
61.6◦ were characteristic of MMT. The similar characteristic peaks
were observed for quartz according to the XRD pattern of the
raw MMT (see Fig. 3a). Moreover, the XRD pattern of TiO2/MMT
demonstrated a characteristic diffraction peak at 2␪ of 25.4◦, 37.7◦,
48.1◦, 55.1◦ and 75.3◦, thereby indicating the presence of anatase
phase [14,32]. It should be noted that for both samples, peaks
corresponding to quartz were also unequivocally identified; this
has already been shown by the previous studies [33–35]. The spec-
trum totally agrees with the previously reported literature for the
montmorillonite and the modified clay [31,32,36]. SEM analysis
was used to investigate the morphology of the MMT, TiO2 and
TiO2/MMT particles. As can be seen in Fig. 4a, MMT had an uneven
structure with non-uniform size distribution. The pure MMT with
some phase separations and cracks was seen, generally showing a
heterogeneous surface morphology. The SEM image of synthesized
TiO2 particles is presented in Fig. 4b. The particle size distribution
of TiO2 nanoparticles was calculated using Manual Microstructure
Distance Measurement software (Nahamin Pardazan Asia Co.,
Fig. 3. XRD patterns of (a) MMT and (b) TiO2/MMT nanocomposite:
lonite, quartz, and anatase.
montmoril-
Iran). As depicted in Fig. 4e, the particle size distribution plot of
TiO2 nanoparticles showed that most of the nanoparticles range
from 60 to 80 nm with the frequency of 38.09%. Fig. 4c shows
that the SEM image of TiO2/MMT nanocomposite approved the
presence of TiO2 particles on the surface of MMT. Fig. 4f exhibits
the size distribution of TiO2 nanoparticles, indicating that most of
the TiO2 particles in TiO2/MMT nanocomposite are in the range of
40–60 nm with the frequency of 75%. The results revealed that the
average size of TiO2 nanoparticles immobilized on the surface of
MMT was smaller than that of pure TiO2 nanoparticles. The particle
size distributions of TiO2/MMT nanocomposite were measured
using the laser light diffraction method. According to the obtained
results, size rangeability of TiO2/MMT varied from 0.04 to 20 ␮m,
with the majority of the nanocomposite particles being in the size
range of 0.1–0.2 ␮m. The EDX pattern of TiO2/MMT is illustrated
in Fig. 4d. The results indicated that the TiO2/MMT contained C, O,
Si, Ti, Mg, Al, Na and K, providing evidence that TiO2 nanoparticles
were formed and immobilized on the surface of the MMT. Accord-
ing to the SEM analysis, it could be deduced that hydrothermal
method utilized in the present work was a reliable means for the
preparation of TiO2/MMT nanocomposites with predictable results.
TEM analysis was used to investigate the more structural infor-
mation. Fig. 5 illustrates the immobilized state of TiO2 particles
on the MMT. There were various well-distributed small particles
within the MMT layers, thereby indicating the good intercalation
of TiO2 particles into the interlayer of MMT and the destruction
of the ordered structure of MMT clay. Also, the deposited TiO2
nanoparticles were aggregated on the surface of the MMT particles.
FTIR analysis was performed to investigate the variations on
the functional groups of TiO2 and MMT during the preparation
of TiO2/MMT nanocomposite. Fig. 6 shows FTIR spectra of MMT,
TiO2 and TiO2/MMT, respectively. As shown in Fig. 6a, the pure

A. Hassani et al. / Journal of Molecular Catalysis A: Chemical 409 (2015) 149–161
153
Table 2
Chemical composition of MMT and TiO2/MMT samples.
Sample
Chemical compositions (wt.%)
SiO2
Al2O3
Fe2O3
K2O
MgO
CaO
Na2O
TiO2
Others
MMT
TiO2/MMT
66.90
28.30
13.80
7.58
2.75
1.55
1.65
0.883
1.58
0.670
0.29
0.161
0.15
2.27
0.44
58.50
12.44
0.086
Fig. 4. SEM micrographs of (a) MMT, (b) TiO2, (c) TiO2/MMT, (d) EDX elemental microanalysis of TiO2/MMT sample, (e) distribution of TiO2 particles size in pure TiO2 sample
and (f) distribution of TiO2 particles size in TiO2/MMT sample.

154
A. Hassani et al. / Journal of Molecular Catalysis A: Chemical 409 (2015) 149–161
Table 3
Surface area and average pore size characteristics of MMT, TiO2 and TiO2/MMT
nanocomposite samples.
BJH total
volumec
Pore sizeb
(nm)
BET surface
Sample
areaa (m2 g−1
)
(cm3 g−1
)
0.452–0.456
0.035–0.036
0.105–0.106
2.62
4.61
3.87
279.278
8.813
53.058
MMT
TiO2
TiO2/MMT
a
Obtained by BET measurement.
Calculated by the BJH (desorption) method using N2 adsorption isotherm.
Obtained by the BJH method.
b
c
80
UV
TiO
2
TiO /MMT
2
60
40
20
0
UV/TiO
2
Fig. 5. TEM photograph of TiO2/MMT nanocomposite.
UV/TiO /MMT
2
MMT had a weak –OH stretching vibration of crystalline water at
3620 cm−1 and a strong broad band at 3420 cm−1, corresponding
to the stretching vibration of the OH group and the inter-
layer water molecules. The bending vibration of water molecules
caused a peak at 1630 cm−1. The band around 1050 cm−1 was
attributed to asymmetric stretching vibration of Si O in the pure
MMT [36,37]. Several bands around 912 cm−1 and 785 cm−1 were
attributed to the stretching vibration of Al O. The bands around
472 cm−1 and 530 cm−1 corresponded to Si
Al
O Si deformation and
O
Si deformation, respectively [38]. The FTIR spectrum of TiO2
is shown in Fig. 6b. The peaks around 480 cm−1 corresponded to
the stretching vibration of Ti O band [39]. The bands at 1630,
2845, 2924 and 3420 cm−1 were attributed to interlayer water
0
20
40
60
80
100
120
Irradiation time (min)
Fig. 7. Comparison of different processes in the photocatalytic degradation of
ciprofloxacin at different times. Conditions: [Catalyst]0 = 0.1 g L−1, [CIP]0 = 20 mg L−1
and pH 5.
(c)
molecules, symmetric stretching vibration of C H and stretch-
ing vibration of OH, respectively. FTIR spectrum of TiO2/MMT
nanocomposite is presented in Fig. 6c. As it is obvious, the devel-
opment of new absorption peak at 935 cm−1 could be attributed
to Ti
O Si stretching vibration [40]. Moreover, the disappearance
or weak intensities of the bands in the range of 460–530 cm−1 and
1050 cm−1 clearly indicated that the basic structure of the MMT
was destroyed during the preparation of this nanocomposite.
Table 3 shows the BET specific surface area and pore parame-
ters of the synthesized catalysts. The specific surface area of the
pure TiO2 was 8.813 m2 g−1. After the immobilization of TiO2 on
the surface of MMT, an increase in BET value and total volume
was observed for TiO2/MMT. On the other hand, the pore size was
decreased from 4.61 nm to 3.87 nm for TiO2 and TiO2/MMT, respec-
tively. These results indicated that the synthesis of TiO2 on the
surface of MMT by hydrothermal method led to the production of
nanocomposite with high surface area and total volume in com-
parison to those of pure TiO2. The decrease in the specific surface
area of TiO2/MMT nanocomposite, as compared with that of MMT,
could be attributed to the collapse of pores of MMT and the exfo-
liation of MMT in TiO2 matrix as a result of the settlement of TiO2
interlayer galleries of MMT. In addition, the change in surface area
may be due to occupancy of the internal surface of pores by TiO2
nanoparticles. Similar observation with various catalysts reported
in the literature [41–44].
(b)
(a)
4000
3400
2800
2200
1600
1000
400
Wavenumber (cm-1)
Fig. 6. FTIR spectrum of (a) MMT, (b) TiO2 and (c) TiO2/MMT.

A. Hassani et al. / Journal of Molecular Catalysis A: Chemical 409 (2015) 149–161
155
80
60
40
20
0
3.2. Comparison of CIP degradation by different processes
Fig. 7 compares the removal of CIP with the initial concentration
of 20 mg L−1 by different processes. The results indicated 6.57% of
CIP degradation by direct photolysis under UV-C radiation. Hence,
direct photolysis alone could not be used as an impressive proce-
dure for the removal of CIP from aqueous solutions. As can be seen
in Fig. 7, the removal of CIP via adsorption onto TiO2 and TiO2/MMT
was lower in comparison to photocatalysis process. The improved
efficiency in removing CIP by UV/TiO2 process could be ascribed
to the production of hydroxyl radicals, which acted as a powerful
oxidizing agent. Furthermore, the obtained results demonstrated
that the photocatalysis of CIP in the presence of TiO2/MMT was
higher than that of TiO2. It could be attributed to the increase in
the surface area of TiO2/MMT, rather than TiO2, as it led to the
remarkable enhanced adsorption capability of the catalyst. The best
results were achieved in UV/TiO2/MMT system. UV/TiO2/MMT was
employed as the best option for conducting the rest of the experi-
ments. Our findings provided results similar to those of previous
studies with regard to the photocatalysis of different pollutants
[35,45].
-1
5
mg L
-1
10 mg L
-1
15 mg L
-1
20 mg L
-1
25 mg L
0
20
40
60
80
100
120
Irradiation time (min)
Fig. 9. The effect of initial ciprofloxacin concentration on photodegradation effi-
ciency by TiO2/MMT. Conditions: [Catalyst]0 = 0.1 g L−1 and pH 5.
3.3. The effect of TiO2/MMT dosage
3.4. The effect of the initial CIP concentration
Catalyst dosage is an important parameter in heterogeneous
photocatalytic reaction [46]. In order to determine the effect of
catalyst dosage on the photodegradation of CIP, experiments were
carried out by varying the amount of TiO2/MMT from 0.025 g L−1 to
0.150 g L−1 (Fig. 8). The photodegradation efficiency was increased
with an increase in TiO2/MMT dosage up to 0.1 g L−1 and decreased
thereafter. Increasing the amount of photocatalyst from 0.1 g L−1 to
0.150 g L−1 resulted in decreasing the photodegradation efficiency
from 61.70% to 54.30% at the reaction time of 120 min, respectively.
By increasing the catalyst dosage, the surface area was increased,
leading to an increase in the production of reactive species [47,48].
However, more catalyst dosage would also induce greater aggre-
gation of the catalyst and decrease the total active surface area,
thereby leading to a reduction in the degradation efficiency. More-
over, due to an increase in the turbidity of the solution, reduction
in the degree of light penetration through the solution could take
place [7,49]. Therefore, other experiments were carried out with
0.1 g L−1 TiO2/MMT to avoid the excessive usage of the catalyst.
One of the most important things is to study the dependence
of photocatalytic degradation on the concentration of water con-
taminants [50,51]. The degradation of CIP was studied by changing
its initial concentration in the range of 5–25 mg L−1. As can be
seen in Fig. 9, the degradation efficiency was decreased from
71.2% to 57.2% with an increase in the initial CIP concentration
from 5 mg L−1 to 25 mg L−1. When the initial CIP concentration
was decreased, the probability of the reaction between CIP and
oxidizing species was increased, leading to an improvement of
photodegradation efficiency. However, the excess CIP concentra-
tion reduced photodegradation efficiency. On the other hand, at
low CIP concentration, excess active adsorption sites on the pho-
tocatalyst surface could interact with the relatively insufficient CIP
molecules in the solution. Consequently, the degradation efficiency
of CIP was high. However, with increasing the pollutant in the
aqueous solution, there was a surplus. In this case, only parts of
CIP could interact with limited active sites on the photocatalyst
surface, thereby causing lower photodegradation efficiency of CIP
at higher initial concentration. In addition, according to the Beer-
Lambert law, as the initial CIP concentration was increased, the path
length of photons entered the solution decrease, thereby resulting
in lower photon adsorption on catalyst particles and consequently
decreasing the generation of hydroxyl radical [52,53]. Thus, the
photodegradation efficiency was gradually decreased as a result
of increasing CIP concentration.
Literature review revealed that the kinetic model for hetero-
geneous photocatalytic was in accordance with the Langmuir-
Hinshelwood kinetic expression [54,55]. This model for CIP
degradation can be written as follows:
d[C]
dt
Kads[C]
1+Kads[C]0
r = −
= Kc
= Kobs[C]
(3)
(4)
1
Kobs
1
[C]0
=
+
KcKads Kc
where [C]0 is the initial concentration of CIP (mg L−1), Kads is the
Langmuir–Hinshelwood adsorption equilibrium constant (L mg−1),
kc is the rate constant of surface reaction (mg L−1 min−1), and
kobs is the pseudo-first-order rate constant (min−1). When ini-
Fig. 8. The effect of TiO2/MMT amount on the photodegradation efficiency of
ciprofloxacin at the irradiation time of 120 min. Conditions: [CIP]0 = 20 mg L−1 and
pH 5.
1
tial concentrations were plotted versus
, the rate constant of
K
surface reaction and the adsorption equoiblisbrium constant were

156
A. Hassani et al. / Journal of Molecular Catalysis A: Chemical 409 (2015) 149–161
80
60
40
20
0
80
pH=3
pH=4
pH=5
pH=7
pH=9
pH=10
UV-C
UV-B
UV-A
60
40
20
0
0
20
40
60
80
100
120
0
20
40
60
80
100
120
Irradiation time (min)
Irradiation time (min)
Fig. 10. The effect of UV light kind on the photodegradation efficiency of
Fig. 11. The effect of pH on the photodegradation efficiency of ciprofloxacin. Con-
ditions: [Catalyst]0 = 0.1 g L−1, and [CIP]0 = 20 mg L−1
ciprofloxacin. Conditions: [Catalyst]0 = 0.1 g L−1, [CIP]0 = 20 mg L−1and pH 5.
.
determined to be kc = 0.634 mg L−1 min−1 and Kads = 0.0138 L mg−1
,
surface of catalyst and CIP. Furthermore, the low photodegradation
efficiency at more acidic pH (for example, at 3) may be related to the
decomposition and corrosion of catalyst in acidic medium [60]. On
the other hand, both TiO2/MMT and CIP were negatively charged in
basic conditions, leading to electrostatic repulsion between them
and the low photodegradation efficiency of CIP. In the case of pH 5
(the natural pH of CIP), adsorption efficiency was the highest and
the surface of TiO2/MMT was still in the positive form and the CIP
was in its neutral form. In other words, it could be concluded that
adsorption occurred through ion exchange mechanism depend-
ing on the entropy factor. Accordingly, these results, as mentioned
above, showed that the natural pH of CIP facilitated the adsorption
of the CIP molecule and improved photocatalytic degradation.
respectively. The good regression coefficient (R2 = 0.9971) indicated
that the photocatalytic degradation of CIP by the TiO2/MMT fitted
the Langmuir–Hinshelwood kinetic expression well.
It should be noted that the main operational cost of the pho-
tocatalytic degradation process is relevant to the electrical energy
consumption as it is considered an economical factor [56]. Electrical
energy consumption (KWh m−3) of photocatalytic process at opti-
mal experimental conditions is reported in Table 4. As depicted
in Table 4, the lowest EEO consumption could be attributed to
UV/TiO2/MMT process.
3.5. The effect of the type of UV light source
In order to evaluate the impact of the kind of UV light on CIP
removal in the photocatalytic process, different UV light regions
including UV-A, UV-B and UV-C were employed (Fig. 10). The
degradation efficiency of CIP was increased in the order of UV-
A < UV-B < UV-C. The results could be explained in terms of the
energy of photons emitted from UV-C lamp as it was higher than
that of UV-B and UV-A. Hence, photogeneration of e−/h+ pairs under
UV-C radiation was high and consequently, more reactive radicals
were developed for the degradation of the target pollutant [57].
3.7. The effect of the presence of radical scavengers
To assess the effect of the presence of radical scavengers on the
photocatalytic degradation of CIP, bicarbonate, sulfate, iodide and
chloride, which are normally present in water, were employed. As
shown in Fig. 12, at the irradiation time of 120 min, the degradation
efficiency was decreased from 61.70% to 52.87%, 42.94%, 39.86% and
25.83% in the presence of bicarbonate, sulfate, iodide and chloride,
respectively. Accordingly, the order of the inhibitory effect of dif-
ferent radical scavengers was found to be as Cl− > I− > SO42− > HCO3−.
3.6. The effect of the initial pH
80
The solution pH is one of the important parameters that can
markedly affect the photocatalytic process. In the present work, the
effect of the initial pH on the photocatalytic degradation efficiency
of CIP was studied in the pH range of 3–10 (Fig. 11). The results indi-
cated that the degree of photodegradation efficiency was increased
with increasing the solution pH up to 5 and then decreased. Explain-
ing the pH effect on the photocatalytic process is a very laborious
task because of its multiple roles such as electrostatic interactions
between the semiconductor surface, solvent molecules, substrate
and charged radicals formed during the reaction process. The pH
at which the surface of a catalyst is uncharged is defined as the
zero point charge (pHzpc). The pHzpc of TiO2/MMT was obtained to
be 8.4. The CIP has two pKa values (5.9 and 8.89). It can exist as a
cation (CIP0,+), zwitterions (CIP−,+) and anion (CIP−,0) at different
pHs [58,59]. When pH is lower than 5.9, CIP is in cationic species,
while it exists as anionic form when pH value is higher than 8.89.
At the low pH values, the decrease in photodegradation efficiency
could be attributed to the favored hydrogen ions adsorption and/or
scavenger effect of Cl− ions produced from HCl used in adjusting
pH, as well as the repelling effect between the positively charged
Without scavenger
Sodium bicarbonate
Well water
60
40
20
0
Sodium sulfate
Potassium iodide
Sodium chloride
0
20
40
60
80
100
120
Irradiation time (min)
Fig. 12. The effect of the presence of different salts (as radical scavengers) on the
degradation efficiency of ciprofloxacin. Conditions: [Catalyst]0 = 0.1 g L−1, [Radical
scavenger]0 = 20 mg L−1, [CIP]0 = 20 mg L−1, and pH 5.

A. Hassani et al. / Journal of Molecular Catalysis A: Chemical 409 (2015) 149–161
157
Table 4
First order kinetic constants, half-lives and electrical energy consumption for the degradation of 20 mg L−1 CIP by different process.
Process
Degradation
efficiency (%)
(120 min)
Rate constant
R2
t1/2 (min)
EEO (kWh m−3
)
(min−1
)
UV
6.57
56.68
61.70
0.0006
0.0063
0.0069
0.9742
0.9142
0.8965
1155.33
110.03
100.46
2048
195.1
178.1
UV/TiO2
UV/TiO2/MMT
Table 5
Main chemical and physical composition features of well water.
respectively. It is obvious that the addition of hydrogen peroxide
increased the photocatalytic degradation of CIP due to the increased
generation of hydroxyl radicals (Eq. (5)) [29,49].
Parameter
Specification
Appearance
Odor
Colorless liquid
H2O2 + hv → 2OH
(5)
Characteristic
7.82
323
226.1
0.8
9.83
19.84
15.99
0.04
<0.95
6.11
<5
pH
In addition, the generation of hydroxyl radicals from hydrogen
peroxide as electron scavengers was through the following equa-
tion [69]:
Conductivity (␮S cm−1
)
TDS (mg L−1
)
)
)
COD (mg L−1
CL− (mg L−1
H2O2 + e− → OH− + OH•
(6)
SO24− (mg L−1
)
NO−3 (mg L−1
NO−2 (mg L−1
BrO−3 (␮g L−1
)
)
Therefore, hydrogen peroxide can be considered as an effective
chemical for the generation of hydroxyl radicals to enhance the
degradation efficiency [49,70].The addition of potassium iodate can
also enhance the photocatalytic degradation efficiency of CIP. The
reason is due to the formation of highly reactive radicals and also
the capture of the generated electron by this oxidant.
)
Na+ (mg L−1
)
Mg2+ (mg L−1
Fe2+ (␮g L−1
Hardness (mg L−1
HCa (mg L−1
Hmg (mg L−1
)
)
<5
152
80
72
)
)
Finally, the addition of peroxydisulfate anion (S2O2−) produced
the greatest convenient effect on the photocatalytic8degradation
efficiency. As depicted in Fig. 13, in the presence of K2S2O8, the
degradation efficiency was increased from 61.70 to 97.8% at the
reaction time of 120 min. Each peroxydisulfate ion was activated
to generate two SO−4 radicals under UV radiation (Eq. (7)). Then,
hydroxyl radicals could be formed through sulfate radicals as can
be seen in Eq. (8) [24,71]:
)
TDS: total dissolved solids.
COD: chemical oxygen demand.
The presence of sodium chloride in the aqueous solution had the
most inhibiting effect on the photocatalytic activity of TiO2/MMT,
thereby implying that chloride ions scavenged the photo-generated
holes and the produced hydroxyl radicals [61,62]. Bicarbonate ions
reacted with hydroxyl radicals and produced carbonate radicals
[61]. Likewise, bicarbonate ions can scavenge the photo-generated
holes [61]. As depicted in Fig. 12, in the presence of sulfate
ions in the aqueous solution, the photodegradation efficiency was
decreased. The observed retardation effect of sulfate ions was due
to adsorbed sulfate ions on the catalyst surface and also, its reac-
tion with photogenerated holes and hydroxyl radicals [63,64]. The
potassium iodide was added as the scavenger of both hydroxyl rad-
icals and holes [65,66]. In the presence of iodine anion, a consider-
able inhibition in photocatalytic degradation was achieved as both
hydroxyl radicals and holes were suppressed by−iodide ions [67,68].
There are large amounts of ions such as Cl , SO24−, NO3−, NO2−,
etc. in real water samples which affect the photocatalytic
degradation efficiency. In order to investigate the capability of pho-
tocatalytic process in the degradation of CIP, the effect of matrix
was studied with some well water. Table 5 illustrates the main
chemical and physical composition features of the well water. The
observed decrease in the CIP degradation efficiency in the well
water could be a consequence of the presence of several typical
anions (see Table 5). On the basis of these results, it can be con-
cluded that hydroxyl radicals and holes played a determining role
in the photocatalytic degradation of CIP.
S2O2− + h␯ → 2SO•4−
(7)
(8)
8
SO4•− + H2O → H+ + SO42− + OH•
3.9. Reusability of the nanocomposite
Fig. 14 shows the results of the reusability test of TiO2/MMT
nanocomposite for the photocatalytic degradation of CIP within five
successive cycles with the initial CIP concentration of 20 mg L−1
,
catalyst dosage of 0.1 g L−1 and the irradiation time of 120 min. The
photocatalyst in any run was collected, washed with distilled water,
100
80
60
40
Potassium persulfate
Potassium iodate
3.8. The effect of the presence of chemical oxidants
20
Hydrogen peroxide
Without enhancer
In order to investigate the effect of the presence of the chemi-
cal oxidant in the photocatalytic process, the process was carried
out in the presence of various chemical oxidant such as hydrogen
peroxide, potassium iodate and potassium persulfate. As can be
seen from Fig. 13, the presence of hydrogen peroxide, potassium
iodate and potassium persulfate resulted in increasing the photo-
catalytic degradation of CIP from 61.70% to 83.36, 86.43 and 97.88%,
0
0
20
40
60
80 100
120
Irradiation time (min)
Fig. 13. The effect of different enhancers on the photodegradation effi-
ciency of ciprofloxacin. Conditions: [Catalyst]0 = 0.1 g L−1
[CIP]0 = 20 mg L−1 and pH 5.
, [Enhancer]0 = 1 mM,

158
A. Hassani et al. / Journal of Molecular Catalysis A: Chemical 409 (2015) 149–161
Fig. 14. Reusability of the TiO2/MMT nanocomposite within five consecutive
experimental runs. Conditions: [Catalyst]0 = 0.1 g L−1, [CIP]0 = 20 mg L−1 and reaction
time = 120 min.
dried and then used in a new experiment. It could be observed
that the degradation efficiency of CIP was decreased slightly from
61.70% to 55.10% after five repeated experimental runs, thereby
indicating that the as-prepared TiO2/MMT nanocomposite could
be used as a promising photocatalyst for the degradation of organic
pollutants with a significant reusability potential.
3.10. Artificial neural network modeling
Fig. 15. (a) The effect of the number of neurons in the hidden layer on the perfor-
mance of the neural network and (b) schematic illustration of the optimized ANN
structure.
It is known that the number of its layers, the number of nodes
in each layer and the nature of transfer functions determine the
topology of an artificial neural network. In the present work, time
(min), initial CIP concentration (mg L−1), catalyst (g L−1) and the
initial pH were the input variables given to the feed forward three-
layered neural network. The degradation efficiency was selected
as the tentative response or output variable. Out of several data
points generated, 120 experimental sets were employed to feed
the ANN structure. The samples were divided into training, vali-
dation and test subsets, which contained 72, 24 and 24 samples,
respectively. The mean square error (MSE), which was chosen as
the function of error performance during the net training pro-
cess, had the minimum value at 14 neurons among the examined
neurons from 2 to 20. Due to the random initialization of the
weights, each topology was repeated three times to avoid ran-
dom correlation. Fig. 15(a) shows the relationship between the
mean square error (MSE) and the number of neurons in the hid-
den layer. As can be seen, the lowest MSE was obtained with
14 neurons. Therefore, 14 neurons were found to represent the
best performance of the neural network model. Fig. 15(b) dis-
plays the schematic illustration of the optimized ANN structure.
Fig. 16, on the other hand, shows a comparison between the
values of experimental and predicted degradation efficiency for
the test set based on a neural network model. The plot had a
correlation coefficient (R2) of 0.9864 for the output variable of
test data set. The correlation coefficient of the plot showed the
reliability of the model. The results also revealed that neural net-
work model reproduced the degradation in the present system,
within experimental ranges verified in the fitting model. Also, ANN
offered the weights that were coefficients between the artificial
neurons and therefore, could be considered analogous to synapse
strengths between the axons and dendrites in real biological neu-
rons [72].
Table 6 shows the provided weights by ANN in the present
study. The effect of each input variable on the output variable
was calculated using the neural weight data. The obtained results
showed that the relative importance of the initial concentration
of CIP, pH of solution and TiO2/MMT dosage and UV radiation
time on degradation efficiency was 30.71, 24.43, 22.55 and 22.31%,
respectively. The impact of factors on the degradation efficiency
of CIP was also found to be in the following order: CIP concentra-
tion > pH > catalyst > time. The CIP degradation efficiency was found
to be strongly influenced by all independent variables, but the
70
y = 1.0066x - 0.0275
60
50
40
30
20
10
0
R² = 0.9864
0
10
20
30
40
50
60
70
DE (%) [Predicted]
Fig. 16. Comparison of the experimental results with those calculated via neural
network modeling for the test set.

A. Hassani et al. / Journal of Molecular Catalysis A: Chemical 409 (2015) 149–161
159
Table 6
Matrix of weights, W1: weights between input and hidden layers; W2: weights between hidden and output layers.
Neuron
W1 variable
Time
Bias
W2
Concentration
Catalyst
pH
1.6026
Neuron
Weight
1
2
3
4
5
6
7
8
1.1229
0.2667
−1.0565
−1.6795
−0.3288
1.4398
0.2144
0.9911
1.1954
1.8603
−1.3351
1.6084
−2.7032
-2.1967
1.8993
1.5081
1.0871
−0.7822
−0.2667
0.012
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Bias
0.2056
−0.4648
−0.1897
0.175
0.2132
0.6842
0.6893
0.6841
0.0122
0.9178
−0.5899
−0.2002
−0.2641
0.3655
−0.1486
−2.1889
1.4481
1.2432
0.8711
−1.0486
0.0019
−0.0484
1.956
1.4195
2.3491
0.9517
−0.4354
2.1223
−0.772
1.4683
2.2869
−1.0827
1.6607
0.8649
0.3438
−2.6504
0.0514
1.4787
−0.5724
−1.6483
0.4785
−2.1303
−1.7428
1.545
-0.2452
−0.6712
1.126
0.5853
9
−0.3258
−0.4667
−1.2539
1.1937
1.9727
−1.5134
−0.6419
10
11
12
13
14
1.2108
−1.3744
−2.0835
2.3712
−2.7121
1.8825
−1.2518
−0.6554
−0.3296
Table 7
Identified by-products during photocatalytic degradation of 20 mg L−1 CIP at 25 ◦C after 20 min of reaction at pH of 5.
No.
Compound
tR(min)
Main fragments (m/z) (%)
Structure
1
Heptyl octyl phthalate
3.00
149.00 (100.00%), 132.90 (77.44%), 446.10
(27.84%), 67.00 (19.47%), 75.00 (18.63%)
2
1,1,3-Trimethyl-3-phenyl-
2,3-dihydro- 1H-indene
7.41
221.00 (100.00%), 73.00 (43.76%), 222.00
(22.55%), 103.00 (18.25%), 147.00 (13.79%)
3
4
3-Methyl-2-propionyl-benzoic
acid
2.851
2.773
163.00 (100.00%), 133.00 (65.30), 91.00
(37.26%), 59.00 (34.44%), 446.10 (29.25%)
3-(2-Hydroxy-2-
methylpropyl)cyclohex-2-enone
59.00 (100.00%), 60.00 (2.85%), 446.20
(1.88%), 55.00 (0.80%), 57.00 (0.67%)
5
6
Chlorobenzene
Triacontane
3.75
111.90 (100.00%), 77.00 (47.97%), 114.00
(32.79%), 51.00 (13.12%), 75.00 (9.06%)
27.55
57.10 (100.00%), 55.10 (70.22%), 69.10
(64.72%), 71.00 (62.46%), 97.00 (49.02%)
CH3(CH2)28CH3
7
8
Oleic acid
27.852
2.451
57.10 (100.00%), 55.00 (67.92%), 71.00
(66.42%), 69.00 (51.89%), 85.10 (46.31%)
N,N-Dimethylformamide
73.00 (100.00%), 131.00 (29.05%), 446.10
(20.67%), 146.00 (15.23%), 72.10 (12.28%)

160
A. Hassani et al. / Journal of Molecular Catalysis A: Chemical 409 (2015) 149–161
effects of concentration and pH were more noteworthy. Therefore,
each of the input variables could not be ignored in the present study.
[12] Y. Li, S.-J. Kim, J. Phys. Chem. B 109 (2005) 12309–12315.
[13] M.S. Vohra, K. Tanaka, Water Res. 37 (2003) 3992–3996.
[14] D. Chen, Q. Zhu, F. Zhou, X. Deng, F. Li, J. Hazard. Mater. 235–236 (2012)
186–193.
[15] A.R. Khataee, M.N. Pons, O. Zahraa, J. Hazard. Mater. 168 (2009) 451–457.
[16] S. Anandan, M. Yoon, J. Photochem. Photobiol. C: Photochem. Rev. 4 (2003)
5–18.
[17] H. Chen, S.W. Lee, T.H. Kim, B.Y. Hur, J. Eur. Ceram. Soc. 26 (2006) 2231–2239.
[18] L. Bouna, B. Rhouta, M. Amjoud, F. Maury, M.C. Lafont, A. Jada, F. Senocq, L.
Daoudi, Appl. Clay Sci. 52 (2011) 301–311.
[19] M. Kırans¸ an, R. Darvishi Cheshmeh Soltani, A. Hassani, S. Karaca, A.R. Khataee,
J. Taiwan Inst. Chem. Eng. 45 (2014) 2565–2577.
[20] A. Hassani, R. Darvishi Cheshmeh Soltani, S. Karaca, A.R. Khataee, J. Ind. Eng.
Chem. 21 (2015) 1197–1207.
3.11. Analysis of photocatalytic degradation intermediates of CIP
The intermediates during the first 20 min of CIP photocatalytic
degradation were determined using GC-Mass. Eight intermediates
with high match factor of mass spectrum were successfully identi-
fied (see Table 7). Some of the intermediates have not been detected
due to their quick oxidation to the derivatives [53]. Generation
of these intermediates during the degradation of CIP emphasizes
effective photocatalytic degradation of this pollutant in the pres-
ence of TiO2/MMT nanocomposite.
[21] C. Ooka, H. Yoshida, K. Suzuki, T. Hattori, Microporous Mesoporous Mater. 67
(2004) 143–150.
[22] M. Bobu, A. Yediler, I. Siminiceanu, S. Schulte-Hostede, Appl. Catal. B 83
(2008) 15–23.
[23] E. De Bel, J. Dewulf, B.D. Witte, H. Van Langenhove, C. Janssen, Chemosphere
77 (2009) 291–295.
4. Conclusions
[24] C.-C. Lin, M.-S. Wu, J. Photochem. Photobiol. A 285 (2014) 1–6.
[25] H.-G. Guo, N.-Y. Gao, W.-H. Chu, L. Li, Y.-J. Zhang, J.-S. Gu, Y.-L. Gu, Environ.
Sci. Pollut. Res. 20 (2013) 3202–3213.
[26] W.-T. Jiang, P.-H. Chang, Y.-S. Wang, Y. Tsai, J.-S. Jean, Z. Li, K. Krukowski, J.
Hazard. Mater. 250–251 (2013) 362–369.
[27] A. Czímerová, J. Bujdák, R. Dohrmann, Appl. Clay Sci. 34 (2006) 2–13.
[28] Y. Bessekhouad, D. Robert, J.V. Weber, N. Chaoui, J. Photochem. Photobiol. A
167 (2004) 49–57.
[29] A.R. Khataee, M.B. Kasiri, J. Mol. Catal. A: Chem. 331 (2010) 86–100.
[30] A.R. Khataee, M. Zarei, M. Pourhassan, CLEAN—Soil, Air Water 38 (2010)
96–103.
[31] F. Rasouli, S. Aber, D. Salari, A.R. Khataee, Appl. Clay Sci. 87 (2014) 228–234.
[32] S. Miao, Z. Liu, B. Han, J. Zhang, X. Yu, J. Du, Z. Sun, J. Mater. Chem. 16 (2006)
579–584.
[33] I. El Younssi, T. Rhadfi, A. Atlamsani, J.-P. Quisefit, F. Herbst, K. Draoui, J. Mol.
Catal. A: Chem. 363–364 (2012) 437–445.
[34] Y. Kameshima, Y. Tamura, A. Nakajima, K. Okada, Appl. Clay Sci. 45 (2009)
20–23.
[35] J. Liu, M. Dong, S. Zuo, Y. Yu, Appl. Clay Sci. 43 (2009) 156–159.
[36] J. Liu, X. Li, S. Zuo, Y. Yu, Appl. Clay Sci. 37 (2007) 275–280.
[37] P. Vitanov, A. Harizanova, T. Ivanova, T. Dimitrova, Thin Solid Films 517
(2009) 6327–6330.
[38] A.R. Khataee, M. Sheydaei, A. Hassani, M. Taseidifar, S. Karaca, Ultrason.
Sonochem. 22 (2015) 404–411.
[39] Y.L. Pang, A.Z. Abdullah, Chem. Eng. J. 214 (2013) 129–138.
[40] M. Iwasaki, A. Yasumori, S. Shibata, M. Yamane, J. Sol–Gel Sci. Technol. 2
(1994) 387–391.
[41] S. Agarwal, J.N. Ganguli, J. Mol. Catal. A: Chem. 372 (2013) 44–50.
[42] K. Ravi, B. Krishnakumar, M. Swaminathan, Res. Chem. Intermed. 41 (2015)
5353–5364.
[43] S. Miao, Z. Liu, B. Han, H. Yang, Z. Miao, Z. Sun, J. Colloid Interface Sci. 301
(2006) 116–122.
In summary, the photocatalytic degradation of ciprofloxacin
was investigated in the presence of immobilized TiO2 nanoparti-
cles on the surface of MMT. The degradation efficiency of CIP was
decreased in initial CIP concentration and increased with increas-
ing TiO2/MMT nanocomposite dose up to 0.1 g and decreased
thereafter. Photocatalytic performance of TiO2/MMT under UV-C
radiation was higher than that under UV-B and UV-A radiation. The
efficiency of the studied system in terms of degradation rate fol-
lowed the order: UV < TiO2 < TiO2/MMT < UV/TiO2 < UV/TiO2/MMT.
Addition of hydrogen peroxide, peroxydisulfate and potassium
iodate improved the degradation efficiency. The radical scavenger
reduced the degradation efficiency, thereby indicating that the
dominant controlling mechanism of photocatalytic degradation
of CIP was related to the free radicals. The best degradation was
achieved at natural CIP pH (5). Also, after five cycles of repetition
use, the degradation efficiency of CIP still remained 55.10%. The
electrical energy consumption per order of magnitude for photo-
catalytic degradation of CIP was the lowest in the UV/TiO2/MMT
process as compared to those in the UV and UV/TiO2 processes.
Artificial neural network (ANN) modeling was successfully used to
model the photocatalytic process. All studied factors in this work
(time, initial concentration of CIP, catalyst and pH) were found to
have remarkable effects on the degradation efficiency.
[44] F.-F. Wang, J. Liu, H. Li, C.-L. Liu, R.-Z. Yang, W.-S. Dong, Green Chem. 17
(2015) 2455–2463.
Acknowledgments
[45] Y. Li, J.R. Liu, S.Y. Jia, J.W. Guo, J. Zhuo, P. Na, Chem. Eng. J. 191 (2012) 66–74.
[46] M. Pirhashemi, A. Habibi-Yangjeh, J. Alloys Compd. 601 (2014) 1–8.
[47] M. Pelaez, N.T. Nolan, S.C. Pillai, M.K. Seery, P. Falaras, A.G. Kontos, P.S.M.
Dunlop, J.W.J. Hamilton, J.A. Byrne, K. O’Shea, M.H. Entezari, D.D. Dionysiou,
Appl. Catal. B 125 (2012) 331–349.
[48] K. Nakata, A. Fujishima, J. Photochem. Photobiol. C: Photochem. Rev. 13
(2012) 169–189.
[49] N. Daneshvar, D. Salari, A.R. Khataee, J. Photochem. Photobiol. A 162 (2004)
317–322.
The authors thank to TUBITAK for supports by the 2221-
Fellowship Program for Visiting Scientists and Scientists on
Sabbatical Leave. The financial support by the Atatürk University
Scientific Research Project Council (Project No: 2013/92) is grate-
fully acknowledged. The authors also thank the University of Tabriz
(Iran) for all of the support provided.
[50] M.N. Chong, B. Jin, C.W.K. Chow, C. Saint, Water Res. 44 (2010) 2997–3027.
[51] I.J. Ochuma, R.P. Fishwick, J. Wood, J.M. Winterbottom, Appl. Catal. B 73
(2007) 259–268.
References
[52] A.R. Khataee, A. Karimi, S. Arefi-Oskoui, R. Darvishi Cheshmeh Soltani, Y.
Hanifehpour, B. Soltani, S.W. Joo, Ultrason. Sonochem. 22 (2015)
371–381.
[53] A.R. Khataee, S. Arefi-Oskoui, M. Fathinia, A. Esmaeili, Y. Hanifehpour, S.W.
Joo, N. Hamnabard, J. Mol. Catal. A: Chem. 398 (2015) 255–267.
[54] F.J. Beltrán, F.J. Rivas, O. Gimeno, J. Chem. Technol. Biotechnol. 80 (2005)
973–984.
[55] J.-h. Sun, Y.-k. Wang, R.-x. Sun, S.-y. Dong, Mater. Chem. Phys. 115 (2009)
303–308.
[56] James R. Bolton, Keith G. Bircher, William Tumas, C.A. Tolman, Pure Appl.
Chem. 73 (2001) 627–637.
[1] S. Bae, D. Kim, W. Lee, Appl. Catal. B 134–135 (2013) 93–102.
[2] F.J. Beltrán, A. Aguinaco, J.F. García-Araya, Appl. Catal. B 100 (2010) 289–298.
[3] P. de Voogt, M.-L. Janex-Habibi, F. Sacher, L. Puijker, M. Mons, Water Sci.
Technol. 59 (2009) 39–46.
[4] F.J. Beltrán, A. Aguinaco, J.F. García-Araya, A. Oropesa, Water Res. 42 (2008)
3799–3808.
[5] A.R. Khataee, M. Safarpour, A. Naseri, M. Zarei, J. Electroanal. Chem. 672
(2012) 53–62.
[6] A.R. Khataee, M. Fathinia, M. Zarei, B. Izadkhah, S.W. Joo, J. Ind. Eng. Chem. 20
(2014) 1852–1860.
[57] J.A. Cortés, M.T. Alarcón-Herrera, M. Villican˜a-Méndez, J.
González-Hernández, J.F. Pérez-Robles, Environ. Prog. Sustain. Energy 30
(2011) 318–325.
[58] N. Carmosini, L.S. Lee, Chemosphere 77 (2009) 813–820.
[59] T.A. Gad-Allah, M.E.M. Ali, M.I. Badawy, J. Hazard. Mater. 186 (2011) 751–755.
[60] N. Daneshvar, S. Aber, M.S. Seyed Dorraji, A.R. Khataee, M.H. Rasoulifard, Sep.
Purif. Technol. 58 (2007) 91–98.
[7] N. Modirshahla, A. Hassani, M.A. Behnajady, R. Rahbarfam, Desalination 271
(2011) 187–192.
[8] S. Khameneh Asl, S.K. Sadrnezhaad, M. Keyanpour Rad, D. Üner, Turk. J. Chem.
36 (2012) 121–135.
[9] M.R. Hoffmann, S.T. Martin, W. Choi, D.W. Bahnemann, Chem. Rev. 95 (1995)
69–96.
[10] A.R. Khataee, M.B. Kasiri, J. Mol. Catal. A: Chem. 328 (2010) 8–26.
[11] F. Sayilkan, M. Asiltürk, S¸ . S¸ ener, S. Erdemog˘lu, M. Erdemog˘lu, H. Sayilkan,
Turk. J. Chem. 31 (2007) 211–221.

A. Hassani et al. / Journal of Molecular Catalysis A: Chemical 409 (2015) 149–161
161
[61] D.E. Santiago, J. Aran˜a, O. González-Díaz, M.E. Alemán-Dominguez, A.C. [69] B.J.P.A. Cornish, L.A. Lawton, P.K.J. Robertson, Appl. Catal. B 25 (2000) 59–67.
Acosta-Dacal, C. Fernandez-Rodríguez, J. Pérez-Pen˜a, J.M. Don˜a-Rodríguez,
Appl. Catal. B 156–157 (2014) 284–292.
[70] N. Modirshahla, M.A. Behnajady, R. Rahbarfam, A. Hassani, CLEAN—Soil, Air
Water 40 (2012) 298–302.
[62] P.-S. Yap, T.-T. Lim, Appl. Catal. B 101 (2011) 709–717.
[63] M. Sökmen, A. Özkan, J. Photochem. Photobiol. A 147 (2002) 77–81.
[64] A.R. Khataee, Environ. Technol. 30 (2009) 1155–1168.
[65] J. Rabani, K. Yamashita, K. Ushida, J. Stark, A. Kira, J. Phys. Chem. B 102 (1998)
1689–1695.
[71] A.R. Khataee, R. Darvishi Cheshmeh Soltani, Y. Hanifehpour, M. Safarpour, H.
Gholipour Ranjbar, S.W. Joo, Ind. Eng. Chem. Res. 53 (2014) 1924–1932.
[72] A. Hassani, F. Vafaei, S. Karaca, A.R. Khataee, J. Ind. Eng. Chem. 20 (2014)
2615–2624.
[66] L.-S. Zhang, K.-H. Wong, H.-Y. Yip, C. Hu, J.C. Yu, C.-Y. Chan, P.-K. Wong,
Environ. Sci. Technol. 44 (2010) 1392–1398.
[67] R. Palominos, J. Freer, M.A. Mondaca, H.D. Mansilla, J. Photochem. Photobiol. A
193 (2008) 139–145.
[68] K.-i. Ishibashi, A. Fujishima, T. Watanabe, K. Hashimoto, J. Photochem.
Photobiol. A 134 (2000) 139–142.