Available online at www.sciencedirect.

com

CERAMICS
INTERNATIONAL
Ceramics International 40 (2014) 4009–4018
www.elsevier.com/locate/ceramint

The dependence of the photocatalytic activity of TiO2/carbon

nanotubes nanocomposites on the modification of the
carbon nanotubes
Veljko R. Djokića,n, Aleksandar D. Marinkovića, Ovidiu Ersenb, Petar S. Uskokovića,
Rada D. Petrovića, Velimir R. Radmilovića,c, Djordje T. Janaćkovića
a
Faculty of Technology and Metallurgy, University of Belgrade, Karnegijeva 4, 11120 Belgrade, Serbia
b
Institut de Physique et Chimie des Matériaux de Strasbourg (IPCMS), UMR CNRS–Université de Strasbourg 7504, 23, rue du Loess, BP 43,
F-67034 Strasbourg Cedex 2, France
c
Nanotechnology and Functional Materials Center, Faculty of Technology and Metallurgy, University of Belgrade, Karnegijeva 4, 11120 Belgrade, Serbia
Received 13 June 2013; accepted 10 August 2013
Available online 19 August 2013

Abstract

Nanostructural TiO2/modified multi-wall carbon nanotubes photocatalysts were prepared by hydrolysis of Ti(iso-OC3H7)4 providing chemical
bonding of anatase TiO2 nanoparticles onto oxidized- or amino-functionalized multi-wall carbon nanotubes (MWCNT). The processes of
functionalization of the MWCNT and the deposition of TiO2 influence the photocatalytic activity of the synthesized nanocomposites. The phase
composition, crystallite size, and the structural and surface properties of the obtained TiO2/modified-MWCNT nanocomposite were analyzed
from XRD, FEG-SEM, TEM/HRTEM and FTIR data, as well low temperature N2 adsorption. In the photocatalytic study, the TiO2/oxidized-
MWCNT catalyst showed the highest and the TiO2/amino functionalized-MWCNT catalysts somewhat lower degradation rates, indicating that
the enhancement of photocatalysis was supported by the more effective electron transfer properties of the oxygen- than amino-containing
functional groups, which support the efficient charge transportation and separation of the photogenerated electron-hole pairs.
& 2013 Elsevier Ltd and Techna Group S.r.l. All rights reserved.

Keywords: A. Sol–gel processes; B. Nanocomposites; D. TiO2; Carbon nanotubes; Photocatalysis

1. Introduction has an exceptional ability for more effective charge separation, i.e.,
suppression of electron/hole recombination [1–5].
Due to its outstanding physical and chemical properties, and Pristine nanomaterials, i.e., carbon nanotubes (CNT), are
especially photocatalytic activity, titanium dioxide (TiO2) is one of suitable preferentially for fundamental studies, but the vast
the most studied semiconductor materials in fields of environ- number of their applications require their integration into
mental protection and renewable energy sources. In the last four composite structures. Improvement of the photocatalytic activity
decades, much effort has been invested in research related to of TiO2 was achieved by synthesis of composite materials based
improving the photocatalytic activity of TiO2 and its possible on the coupling of titania and functional nanostructured carbon
practical applications. Generally, in order to ensure a high photo- materials. Special attention was devoted to CNT due to their
catalytic activity, it is necessary to synthesize TiO2 that, in chemical stability, unique structures, remarkable electron-
addition to a large specific surface area, porosity and crystallinity, transfer and adsorption, as well as their mechanical and thermal
properties. In this way, CNT as photocatalysts carriers increased
the separation and migration of photogenerated electrons and
n holes, prompting the creation of excess holes in the valence
Corresponding author. Tel.: þ381 64 270 1090; fax: þ 381 11 3370 387.
E-mail addresses: [email protected], band, while the excited electrons moved from the conduction
[email protected] (V.R. Djokić). band of TiO2 to the electron-accepting surface of the carbon

0272-8842/$ - see front matter & 2013 Elsevier Ltd and Techna Group S.r.l. All rights reserved.
http://dx.doi.org/10.1016/j.ceramint.2013.08.052
4010 V.R. Djokić et al. / Ceramics International 40 (2014) 4009–4018

nanotubes. In addition, the relatively high specific surface area The synthesis of high performance photocatalysts demands the
of CNT leads to an increase in the number of active sites for design of such processes that could provide a MWCNT product
adsorption on the surface of TiO2/CNT nanocomposites. Thus, with a surface containing chemical functionalities that possess
CNT in a composite with TiO2 have, at least, the dual effect of high affinity and surface bonding specificity towards TiO2. With
improving the photocatalytic activity by: (i) reducing the this in mind, the synthesis of TiO2/modified MWCNT at different
recombination of photogenerated electron/hole pairs by increas- ratios MWCNT/Ti(IV) isopropoxide precursor was performed to
ing their separation and (ii) increasing the specific surface area, study the effect of the functionalization method on the crystal-
i.e., the number of active sites for the adsorption of the linity, morphology and photocatalytic activity in the photo-
pollutants to be photocatalytically degraded [4,6,7]. activated degradation reaction of the textile dye C.I. Reactive
However, the application of CNT was greatly hampered by Orange 16 that has a detrimental impact when discharged
their poor dispersion in solvents due to strong interactions untreated into the environment (Fig. 1).
between the nanotubes, which lead to the formation of
sparingly soluble aggregates. CNT functionalization was found
to be an effective method to overcome the solubility limita- 2. Materials and methods
tions, reactivity with respect to different modification agents,
and to provide homogeneous distributions in a variety of 2.1. Preparation of the photocatalysts
mediums. One of the most common functionalization methods
is the oxidative treatment of carbon nanotubes. After oxidative MWCNT (95 þ %, Aldrich), prepared by a chemical vapor
treatment, the CNT are shortened and mostly opened at the deposition method, were used as received without purification.
end, and oxygen-containing functional groups are unevenly Titanium(IV) isopropoxide (98%, Ti(iso-OC3H7)4, Aldrich),
distributed on surface, changing their surface properties and propan-2-ol, absolute ethanol and ethylenediamine (EDA,
providing numerous sites for TiO2 bonding [8]. A covalent or 99%, Fluka), 1,4-phenylene-diamine (PDA, 99%, Fluka), N,
non-covalent modification of carbon nanotubes usually follows N-di-isopropylethylamine (DIEA, 99%, Sigma), O-(7-aza-
the oxidation step. The covalent modifications of the side-wall benzotriazol-1-yl)-N,N,N′,N′-tetra-methyluroniumhexa-fluoro-
of CNT have been described in detail in several review articles phosphate (N-HATU, 98%, Fluka), concentrated sulfuric
[9,10]. In principle, major approaches comprise: (i) amidation (H2SO4, 98%, Fluka) and nitric (HNO3, 99%, Fluka) acid
or esterification of carboxylated CNT and (ii) side-wall and methanol (Fluka, p.a.) were used as received. Millipore
covalent attachment of functional groups directly to pristine deionized water (DW; 18 MΩ cm resistivity) was used for
CNT. Among them, amino-functionalized CNT have high sample washing and solution preparation. Dimethylformamide
reactivity, and can react with many functionalities, such as (DMF, 99.8%, Fluka) was purified and dried to remove any
different Ti-precursors [11]. The introduction of functional trace impurities and water [14,16].
groups cause changes in the physical and electronic properties
of nanotubes, providing a crucial prerequisite for improvement 2.1.1. Oxidation of MWCNT
of the electron transport properties at titania/CNT interface, Raw-MWCNT (480 mg) were dispersed in 520 ml of a
i.e., photocatalytic performances, of the obtained hybrid (3:1 v/v) mixture of concentrated H2SO4 and HNO3 acid,
materials [4,6,11–13]. respectively, and sonicated for 3 h at 40 1C. After cooling to
The electronic properties of MWCNT depend on their room temperature, the dispersion of the oxidized MWCNT (o-
structure and covalent functionalization, which affects their MWCNT) were added drop-wise to 1300 ml of cold DW
semiconducting and conducting properties. Despite the theore- water, and vacuum-filtered through a 0.05 μm pore size poly
tical assumption that electron transport may be expected through (tetrafluoroethylene) (PTFE) membrane filter. The filtrant was
both the outer and inner shells of multi-wall carbon nanotube washed with DW water until pH was neutral. The sample was
(MWCNT), it was found that electron transport at low energies then dried in a vacuum oven at 80 1C for 8 h [16].
was mainly achieved through the outer surface of MWCNT.
Improved electronic properties are highly desirable in the
2.1.2. Functionalization of o-MWCNT with amines
fabrication of supercapacitor, photovoltaic, nanoelectronic
o-MWCNT (160 mg) were dispersed in 95 ml of the liquid
devices, transistors, high frequency nanoelectronics, storage
amine (EDA) for 10 min. The coupling agent N-HATU (12.8 mg)
media, field emission and nanoscale array fabrications [11,12].
was then added, and dispersion was sonicated for 4 h at 40 1C
A comparative examination of the electrochemical behavior
of oxidized and ethylenediamine-modified MWCNT by cyclic
voltammetry and electrochemical impedance spectroscopy con-
firmed that the existing oxygen-containing functionalities were
responsible for the rapid electron transfer on the o-MWCNT
surface. Amino-modification of the o-MWCNT, via oxygen
containing groups, led to a slight decrease in the conductivity of
the layer, the charge transfer properties, i.e., reduced rate of
electron transfer between the MWCNT and solution species, and
the apparent capacitance [14,15]. Fig. 1. The structure of C.I. Reactive Orange 16.
 V.R. Djokić et al. / Ceramics International 40 (2014) 4009–4018 4011

[16]. In the case of the solid amine (PDA), a mixture of 160 mg o- [18] was applied for the determination of type and quantity of
MWCNT, 12.8 mg N-HATU, 16 μl DIEA (the molar ratio N- functional groups present on the surface of the oxidized and
HATU:DIEA was 1:3) and 4.8 ml anhydrous DMF was prepared amino-functionalized MWCNT. Fourier-transform infrared (FTIR)
in order to enhance activation of carboxyl groups by DIEA [14]. spectra were recorded in the transmission mode using a BOMEM
After activation for 15 min, PDA dissolved in anhydrous DMF (Hartmann&Braun) spectrometer.
(32 g PDA/160 ml DMF) was added into the reaction mixture and
sonicated for 4 h at 40 1C. The product was diluted with 480 ml 2.3. Photocatalytic activity
of methanol, vacuum-filtered (0.05 μm PTFE membrane) and
washed extensively with excess methanol. The amino-func- The photocatalytic decomposition of the textile dye C.I.
tionalized MWCNT (EDA: e-MWCNT, and PDA: p-MWCNT) Reactive Orange 16 (CI RO 16; Bezema) (Fig. 1) was followed
were dried in a vacuum oven at 60 1C for 8 h, and stored in a by measuring the decrease of the absorbance of an aqueous dye
desiccator. solution (50 mg dm  3) in the presence of the synthesized
photocatalysts (50 mg in 25.0 cm3 of solution). The dispersions
2.1.3. Synthesis of the TiO2/MWCNT nanocomposites were mixed in the dark for 30 min in order to establish the
In a three-necked flask (50 cm3) equipped with a condenser adsorption/desorption equilibrium and then illuminated with UV-
and thermometer, 160 mg o-MWCNT, 13.6 cm3 2-propanol A light using a Philips 125 W lamp with an intensity of 12.7 mW/
and 2, 3, 4, 5, 6 or 8 cm3 Ti(iso-OC3H7)4 were added under cm2. Aliquots (3.0 cm3) were taken and separated from the
magnetic stirring. Reaction mixture was heated at 80 1C for 7 h photocatalysts by filtration through a 0.2 μm poly(vinylidene
under moderate mixing (150 rpm), and oil-bath was removed, fluoride) (PVDF) syringe filter (Whatman), and used for the
and mixing was continued, for additional 60 min providing determination of time-dependent change of dye concentration
reaction mixture to attain room temperature. After completion using a Shimadzu 1800 UV–vis spectrophotometer. The light
of the reaction, the obtained grayish/black TiO2/o-MWCNT irradiation intensity was measured by a UVX Digital Ultraviolet
dispersion was vacuum-filtered (0.05 μm PTFE membrane) Intensity Meter (Cole-Parmer, USA; Model: 97-0015-02(UVX))
and washed with ample deionized water. After drying at 70 1C using a 365 nm sensor (97-0016-02 (UVX36)). The durability of
for 24 h, the material was ground to a fine powder and calcined all nanocomposite catalysts was checked in five consecutive
at 400 1C for 2 h. The TiO2/raw-, TiO2/e- and TiO2/p- cycles. After each cycle, the catalyst was recovered by centrifuga-
MWCNT photocatalysts were prepared using the optimal ratio tion, washed with ample DI water, dried at 80 1C, and reused in
found for the preparation of the TiO2/o-MWCNT photocata- the next degradation cycle.
lyst. Additionally, pure TiO2 (TiO2-blank) was prepared by an
analogous procedure but in the absence of MWCNT, and used
in a comparative photocatalytic study.
3. Results and discussion
2.2. Characterization
The nanocomposite photocatalysts were prepared by a two-
The specific surface area and pore volume of the samples were step process, functionalization of the MWCNT (consecutive
measured using an accelerated surface area and porosimetry system processes of oxidation and amino modification) and precipita-
– ASAP 2020 (Micromeritics, USA). The X-ray diffraction (XRD) tion of TiO2, providing chemical bonding of TiO2 deposit on
data were obtained using a BRUKER D8 ADVANCE with a the MWCNT support. To perform crystallization of precipi-
Vario 1 focusing primary monochromator (Cukα1 radiation, tated Ti(OH)4, a low temperature heat treatment in air was
λ¼ 1.54059 Å). The crystallite size of the anatase, D, was performed to prevent the decomposition of the nanotube
determined according to the Scherrer Equation: support and anatase/rutile phase transformation.
According to the results of the photocatalytic study of the
D ¼ Kλ=ðβ cos θÞ synthesized nanocomposites, the highest activity was obtained
where λ is the wavelength of the X-ray radiation (λ¼ 1.54059 Å), for TiO2/o-MWCNT. The optimal ratio used for the synthesis
K is the Scherrer constant (K¼ 0.9), θ is the angle of the of TiO2/o-MWCNT was also applied for the preparation of all
characteristic X-ray diffraction peak (θ¼ 12.651) and β is the other nanocomposites (160 mg MWCNT, 13.6 cm3 2-propanol
full-width-at-half-maximum-height of the anatase (101) plane (in and 4 cm3 Ti(iso-OC3H7)4: [MWCNTs]loading ¼ 12.9 wt%).
radians). Scanning electron microscopy (FEG-SEM) was per- Higher multi-wall carbon nanotubes (MWCNTs) loadings led
formed with a field emission gun TESCAN MIRA3 XMU electron to increased shielding of the active photocatalytic sites and
microscope. Conventional transmission electron microscopy increased absorbance and scattering of photons, which caused
(CTEM) was performed using an AEM JEOL 200CX microscope a decrease in the photocatalytic activity.
at 200 kV, and high resolution transmission electron microscopy
(HRTEM) using a TEAM0.5 aberration corrected electron micro- 3.1. Boehm analysis
scope operated at 80 kV, below the threshold of knock-on damage
of the CNT [17]. The atomic compositions of the obtained In order to optimize the properties of the surface of carbon
materials were determined by electron energy-loss spectroscopy nanotubes in a nanocomposite, i.e., charge transfer at the
(EELS) in the scanning TEM (STEM) mode. The Boehm method nanomaterial interface, it was necessary to study the influence
4012 V.R. Djokić et al. / Ceramics International 40 (2014) 4009–4018

Table 1
Results of the Boehm titrationa of the raw and functionalized MWCNT.

Carboxyls Lactones Phenols Total acidic sites Total basic sites

MWCNT (raw-MWCNT) 0.07 0.28 0.24 0.58 0.19

o-MWCNT 0.87 1.78 1.43 4.09 0.42
e-MWCNT 0.15 1.24 1.83 3.22 1.15
p-MWCNT 0.18 1.32 1.70 3.15 1.04
a
In mmol/g.

of the introduction of different types of functional groups on Table 2

the photocatalytic activity. The functionalization method of Specific surface area (Sp) and pore volume (Vp) of the raw-, o-, e-, and p-
the raw-MWCNT caused the introduction of different types, MWCNT, as well as the synthesized nanocomposites and pure TiO2.
quantities and distribution of the functional groups, which Sample Sp (m2 g  1) Vp (cm3 g  1)
contribute significantly to the type and strength of chemical
bonding of TiO2 onto the CNT surface and thus the TiO2–Blank 17.44 0.048
MWCNT (raw-MWCNT)a 187.58 0.755
effectiveness of the charge transfer properties through the o-MWCNTa 78.49 0.328
TiO2/modified MWCNT interface. Oxidative treatment of e-MWCNTa 101.24 0.538
MWCNT caused degradation of the graphitic structure, p-MWCNT 108.42 0.581
mainly at the nanotube ends, and graphitic plane opening TiO2/o-MWCNT 66.25 0.081
created sidewall defect sites thereby introducing oxygen TiO2/e-MWCNT 59.15 0.072
TiO2/p-MWCNT 57.02 0.069
functionalities. Subsequent reaction with EDA or PDA
introduced amino-terminated functional groups which were a
Ref. [16].
quantified from the results of Boehm titration data (Table 1).
The applied mild and efficient coupling method for amino
functionalization provided a high extent of transformation of
carboxylic groups to amide groups containing free amino the interaction of the MWCNT supports and the deposited TiO2
terminal functionalities. photoactive material (Fig. 2). FTIR analysis of the synthesized
nanocomposites was performed to obtain qualitative estimations
3.2. Textural analysis of the presence of residual functional groups after TiO2
precipitation. The differences in the FTIR spectra of the modified
Specific surface area (Sp) and pore volume (Vp) of the raw-, MWCNT and the nanocomposite materials were analyzed by the
o- and amino-functionalized MWCNT, as well as correspond- differences in the peak intensity, peak shifting and peak
ing nanocomposites and pure TiO2 are given in Table 2. appearance or disappearance, which is an indication of the
Largest surface area, Sp, was found for the raw-MWCNT, degree of interaction of the functional groups and the TiO2
while the value was significantly decreased for o-MWCNT. deposits. Changes in the interaction of the groups resulted in
The decrease in the surface area and pore volumes were bond strength changes, which resulted in changes in the vibration
probably due to the smaller confined space between the frequencies i.e., changes in the wavelength values. Band shifts
o-MWCNT [19], and the introduction of oxygen-containing to lower or higher frequencies indicates bond weakening or
functional groups [8], which contributed to the decrease of the strengthening, respectively.
effective MWCNT surface. Modification with EDA and PDA The bands observed at E3400 cm  1 and E1650 cm  1 were
resulted in increases in Sp and Vp. This could be explained if assigned to the stretching and deformation vibrations of the OH
interparticle repulsions between e-MWCNT and p-MWCNT groups present at the TiO2 surface and from adsorbed water,
resulted in smaller-sized “globs” of MWCNT and/or if respectively. The former arises from partial hydrolysis during
the additional ultrasound treatment used during amino-func- synthesis and moisture adsorption from the air. Additionally, the
tionalization resulted in smaller aggregates of e-MWCNT and absorption bands in the region from 1000 to 500 cm  1 correspond
p-MWCNT [16]. Additionally, the surface area and pore to Ti–O bond vibrations in the TiO2. Generally, it could be stated
volume of pure TiO2 was low (Table 2), which is additional that synthesis of TiO2 nanocomposite provided a method for TiO2
evidence that TiO2 precipitation caused the decreases in Sp and deposition covering most of the functional groups present at the o-,
Vp of all the nanocomposites. e- and p-MWCNT surfaces. The most intimate contact of two
phases was achieved in the TiO2/o-MWCNT nanocomposite,
3.3. FTIR analysis while analysis of FTIR spectra of TiO2/e- and TiO2/p-MWCNT
photocatalysts indicated that differences in the reactivity of the
Analysis of the FTIR spectra of the investigated samples alkyl (higher) and aromatic (lower) amino groups governed the
proved itself to be a useful tool for obtaining information about contribution of either TiO2 nanoparticles formation bonded
 V.R. Djokić et al. / Ceramics International 40 (2014) 4009–4018 4013

Fig. 2. FTIR spectra of the raw-, o-, e- and p-MWCNT (a); and synthesized nanocomposite photocatalysts (b).

to MWCNT surface or homogeneous nucleation in solution. The

FTIR spectra of the TiO2/e- and TiO2/p-MWCNT photocatalysts
clearly showed that the higher reactivity of the amino-alkyl group
resulted in a more homogeneous deposition of TiO2,
i.e., drive preferentially the reaction at the MWCNT surface and
thus more nucleation center were formed and the subsequent
crystallization (crystal growth) produced a more homogeneous
covering. Lower reactivity and probably lower compatibility of
aromatic moieties with the TiO2 deposit contributed to the creation
of larger non-bonded TiO2 grains and clusters with also an
observable TiO2 deposit at the p-MWCNT surface (Fig. 4).

3.4. XRD analysis

Phase composition, crystallite size, structural and surface

properties of the obtained TiO2/MWCNT nanocomposites
were analyzed according to the XRD (Fig. 3) and TEM/
HRTEM results (Fig. 5). Powder X-ray diffraction (XRD) Fig. 3. XRD patterns of the synthesized photocatalysts.
patterns of the synthesized nanocomposite photocatalysts are
presented in Fig. 3. Studies of the crystalline structure showed Table 3
that pure anatase TiO2 phase was present in the nanocompo- The mean crystallite size of the anatase nanoparticles.
sites and the pure TiO2. Sample Dmean (nm)
The mean crystallite size of TiO2 nanoparticles (Table 3)
decorating the raw-, oxidized and amino functionalized TiO2–Blank 16.270.9
MWCNT surfaces were calculated using the Scherrer Equation TiO2/MWCNT 14.570.8
TiO2/o-MWCNT 12.170.3
based on the full-width-at-half-maximum-height of the char-
TiO2/e-MWCNT 13.370.5
acteristic peak in the corresponding XRD patterns (Fig. 3). TiO2/p-MWCNT 13.570.7
A somewhat smaller crystallite size was found for TiO2/o-
MWCNT, indicating that more reactive centers were accessible
for coordination, while the introduction of aminoethyl and
aminophenyl bridging groups diminished the numbers of the 3.5. SEM analysis
available reactive centers.
The results given in Table 3 indicate a slight dependence of FEG-SEM images of the samples obtained by precipitation
the mean crystallite size on the method of modification of of TiO2 onto raw-, o-, e- and p-MWCNT supports are shown
the MWCNT. in Fig. 4.
4014 V.R. Djokić et al. / Ceramics International 40 (2014) 4009–4018

Fig. 4. FEG-SEM images of (a) TiO2/MWCNT; (b) TiO2/o-MWCNT; (c) TiO2/e-MWCNT, and (d) TiO2/p-MWCNT photocatalysts.

The treatment of o-MWCNT with Ti(iso-OC3H7)4 resulted be assumed that the titania nanoparticles at the MWCNT surface
in a more uniform precipitation of titania, leading to tubes with were bonded by Ti–O(O)C/Ti–OC (carboxylic/phenolic) and
a population of small observable prominences more or less Ti–NHRNH(O)C (R is ethyl or phenyl) bonds in TiO2/o-
uniformly distributed on the surface of the nanotubes. Com- MWCNT and TiO2/amino modified-MWCNT nanocomposite,
pared with those of TiO2/o-MWCNT, the sidewalls of TiO2/e- respectively.
and p-MWCNT are more corrugated and larger aggregates of
bonded TiO2 are detectable with increased diameter at domains
where a significant numbers of functional groups were located. 3.6. TEM/HRTEM analysis
The largest noticeable distinct phase of titania, with embedded
raw-MWCNT, was found for TiO2/MWCNT (Fig. 4a), indi- TEM/HRTEM analysis was applied in order to study the
cating very low bonding specificity of the non-modified morphologies of the obtained nanocomposites, i.e., the inter-
MWCNT surface, which is mostly a graphitic, non-polar action of the TiO2 deposits and the functional groups at the
surface, toward the titania precursor. surface of the modified MWCNT. As an example, the obtained
The FEG-SEM micrographs revealed that TiO2/o-MWCNT results are given for TiO2/o-MWCNT in Fig. 5. The high
(Fig. 4b), compared to TiO2/amino-modified MWCNT (Fig. 4c resolution images (Fig. 5b) shows clearly the presence of TiO2
and d), had a more homogeneous and thicker TiO2 coating. nanocrystals on the surfaces of the o-MWCNT. The HRTEM
The larger diameter of the observable pure TiO2 nanoparticles images (Fig. 5b and c) indicate intimate contact between the
and clusters obtained for TiO2/e- and TiO2/p-MWCNT indi- MWCNT surface and randomly orientated TiO2 nanocrystals.
cated lower affinities of the titania precursor for the surface of It could be also observed that, except for the nanoparticle
the amino-modified nanotube supports. A study of the different directly bonded to the MWCNT surface, randomly distributed
reaction condition, e.g., the presence of water, even at larger aggregates of TiO2 exist in Fig. 5a. The lateral binding
ppm quantity in reaction mixture, caused the formation of of the TiO2 aggregates and nanoparticles present at the
larger titania clusters, bonded and non-bonded to the MWCNT MWCNT surface create weak connections to the MWCNT
support. After sonication, the intimate coatings on the support (Fig. 5a). The electron energy-loss spectrum-EELS
MWCNT surfaces were stable after prolonged time of expo- (insets in Fig. 5b and c) clearly show the presence of Ti L23
sure, but the non-bonded aggregates were easily broken away edges at 450 eV and O K edges at 530 eV, indicating pure
becoming free standing particulates. From these results, it could TiO2 nanocrystals.
 V.R. Djokić et al. / Ceramics International 40 (2014) 4009–4018 4015

Fig. 5. TEM/HRTEM images of TiO2/o-MWCNT nanocomposite: (a) TEM image of nanocomposite; (b and c) HRTEM images with TiO2 nanoparticles, marked
by arrows, present on surface of o-MWCNT; EELS of TiO2 crystallites – inset in (b and c).

The structural changes of the MWCNT surface due to moisture is an important factor which governs the hydrolysis
oxidation, i.e., the introduction of oxygen containing function- of the titania precursor, providing conditions for either homo-
alities, create plane structures and MWCNT end opening, geneous nucleation in solution or heterogeneous nucleation on
which then becomes highly vulnerable sites for additional the MWCNT surface. In this study, the mildest hydrolytic
oxidative attack. Further oxidative attack at the formed defects method was applied to direct nano-TiO2 precipitation onto the
favors the formation of cluster defects rather than a random modified MWCNT surfaces by favoring competitive reactions
distribution along the MWCNT. The defect sites behave as between either moisture (water hydroxyl group) or hydroxyl
adhesive centers that result in concentrations of degradation and carboxyl groups (o-MWCNT) or amino functionalities
processes within limited domains and hence, the oxidative (e- and p-MWCNT) at surface support. According to the
process causes penetration into the inner layers of the presented results the most convenient competitiveness between
MWCNT [20]. The primary binding sites, i.e., the oxygen- the reactivity of moisture and groups on the modified
containing functionalities, are nucleation centers for crystal MWCNT surface was achieved in the case of o-MWCNT.
growth of TiO2 at the o-MWCNT surface, and crystallization, Besides the optimization of the synthesis procedure presented
grain growth and crystal growth inside mesopores (Fig. 5b). in this work, it was shown that the higher reactivity of titania
Homogenous nucleation in solution provides crystallization precursor TiBr4, in comparison to Ti(iso-OC3H7)4, with the
centers that are subjected to crystal growth and coarsening, i.e., oxygen containing surface functionalities provided higher
lateral binding, resulting in the formation of large aggregates photocatalytic activities of the synthesized nanocomposite
of TiO2 nanoparticles. The reactivity of Ti(iso-OC3H7)4 with photocatalysts [13].
4016 V.R. Djokić et al. / Ceramics International 40 (2014) 4009–4018

3.7. Photocatalytic study electrons, and these processes depends on the total acidity (the
acidic functional groups present in o-MWCNT) and the type of
The efficiencies of all the synthesized photocatalysts in the amino functionalities present at the e- and p-MWCNT surfaces
photocatalytic degradation of the chosen dye are shown in (Table 1).
Fig. 6. Photolysis of the dye in the absence of a photocatalyst The dye degradation rate obtained using TiO2/o-MWCNT
showed negligible time-dependent changes in the dye concen- was higher than for all other nanocomposites and pure TiO2,
tration and the amount of degraded dye was 1% after 6 h which indicates the great significance of oxygen-containing
(Fig. 6). Low activity was found in the presence of pure TiO2 functionalities at the MWCNT surface on the photocatalytic
(TiO2–Blank), i.e., 3% of initial dye still remained in solution activity, i.e., their electron transporting properties. Addition-
after 6 h of irradiation. However, the TiO2/MWCNT hybrid ally, the low photocatalytic activity of pure TiO2 indicates that
samples exhibits higher photocatalytic degradation efficiency; the amino modified-MWCNT in TiO2/e- and p-MWCNT also
the highest value was obtained for TiO2/o-MWCNT (47 min), contributed to enhanced photocatalytic activities. The
while TiO2/e-MWCNT (84 min) and TiO2/p-MWCNT (98 min) MWCNT could act as electron sinks, i.e., the photogenerated
exhibited somewhat lower activities. electrons are transferred into the MWCNT providing a low
Efficient application of titania nanocomposite demands well- probability of the recombination process. The large electron-
dispersed photocatalyst nanoparticles, i.e., the availability of storage capacity of the MWCNT is the driving force for the
surface active sites at the particle/solution interface. Such
behavior depends on catalyst morphology: particle size and
surface area, as well as on inter-particle interaction. It is worth
mentioning that the TiO2/o-, TiO2/e- and TiO2/p-MWCNT
samples displayed very good photocatalytic efficiency for CI
RO 16 degradation, which implies that MWCNT support
contributed to an enhancement of the photocatalysis. Three
contributing factors are important: adsorption of the dye
molecules, light absorption, and an effective charge separation
and transportation process [13]. Many studies revealed that dye
adsorption on MWCNT arise from two contributing factors:
physical adsorption and π–π interaction of the hydrophobic
part of the sorbate and the intact graphene plane present at the
surface of MWCNT [21]. It was found that the adsorptive
capabilities of the nanocrystalline TiO2 was not a main factor
controlling the photocatalytic activity [22] as although surface
hybridization of TiO2 with graphite-like carbon improved the
photocatalytic efficiency, only a small improvement in the
adsorption properties was registered [23]. The photocatalytic
Fig. 6. Relative dye concentration change vs. irradiation time in the presence
activity of graphene-supported TiO2 was almost independent
of the synthesized photocatalysts.
of the specific surface area, while the life-time of the
photogenerated electron/hole pairs was found to be the main
factor influencing the photocatalytic efficiency [24].
Similarly, the results of specific surface area and pore
volume (Table 2) and photocatalytic studies (Fig. 6) of the
synthesized photocatalysts indicated that the textural para-
meters are a contributing factor that improved the photocata-
lytic efficiencies. A comparison of the photocatalytic activities
of all the nanocomposites and pure TiO2 indicated that
efficient separation of the photogenerated electron/hole is the
main factor contributing to an increased photocatalytic activity
(Fig. 7). According to a previous study, the MWCNT support
contributes to more effective charge separation and suppres-
sion of electron/hole pair recombination [13]. Hybridizing both
raw and modified MWCNT with a grafted inorganic nanos-
tructure, TiO2, resulted in photocatalysts of improved charge
transfer properties, which helps in reducing electron-hole
recombination (Fig. 7). The oxygen-containing and amino-
terminated functionalities at the TiO2/o-, e- and p-MWCNT
interface, respectively, supports the electron transporting Fig. 7. The simplified mechanism for the enhanced photocatalytic activity of
properties, i.e., the migration efficiency of the photogenerated TiO2/CNT nanocomposite catalyst.
 V.R. Djokić et al. / Ceramics International 40 (2014) 4009–4018 4017

capture of photon-excited electrons from the conduction band photogenerated electron-hole pairs. The amino functional
of the TiO2. Otherwise, the MWCNT could act as impurities groups were somewhat less effective.
through Ti–O(O)C/Ti–OC and Ti–NHRNH(O)C bonds, which
contribute to efficient electron/hole separation. Simultaneously, Acknowledgments
MWCNT could provide fast conduction of the captured
photon-excited electrons, hindering thereby the recombination Financial support through the Ministry of Education,
of electron-hole pairs. Moreover, the TiO2/MWCNT systems Science and Technological Development of the Republic of
may be more complex as the contribution of two distinct Serbia, Project no. III 45019, is gratefully acknowledged. V.R.
effects was recognized: the first, extended light absorption due R. acknowledges the support of the Nanotechnology and
to the presence of C–O–Ti bonds, similar to those in carbon- Functional Materials Center, funded by the European FP7
doped TiO2, and the second addresses the electronic config- Project no. 245916, from the Ministry of Education, Science
uration of the CNT that consists of large numbers of mid band- and Technological Development of the Republic of Serbia,
gap states introduced by the existing defects, which aids Project no. 172054. Quantitative high resolution electron
electronic excitation [4,6]. microscopy analysis was performed at the National Center
Moreover, recently a DFT study on the electronic structure of for Electron Microscopy, Lawrence Berkeley National Labora-
semiconducting and metallic tubes in TiO2/CNT heterojunctions tory, funded by the U.S. Department of Energy under Contract
showed that the photogenerated electron, injected into the DE-AC02-05CH11231.
conduction band of TiO2, effectively reduced the transition
energy, which contributed to an enhancement of the phototac- References
tivity. Additionally, the DFT simulations evidenced that the
higher visible light, compared to ultraviolent irradiation, induced [1] A. Fujishima, X. Zhang, D.A. Tryk, TiO2 photocatalysis and related
photocatalytic activity of CNT/TiO2 hybrid materials of semi- surface phenomena, Surface Science Reports 63 (2008) 515–582.
conducting, contrasted with metallic CNT/TiO2 systems, was [2] A. Fujishima, K. Honda, Electrochemical photolysis of water at a
due to efficient charge separation across the interface [25]. semiconductor electrode, Nature 238 (1972) 37–38.
[3] N. Ruzycki, G.S. Herman, L.A. Boatner, U. Diebold, Scanning tunneling
The advantageous application of the selected catalysts is microscopy study of the anatase (100) surface, Surface Science Letters
justified by their prolonged use, i.e., long-term reusability, in the 529 (2003) 239–244.
course of consecutive cycles of catalyst application for dye [4] K. Woan, G. Pyrgiotakis, W. Sigmund, Photocatalytic carbon-nanotube–
photodegradation. The reusability study showed that the photo- TiO2 composites, Advanced Materials 21 (2009) 2233–2239.
degradation efficiency of all the synthesized photocatalysts were [5] G. Socol, Yu. Gnatyuk, N. Stefan, N. Smirnova, V. Djokić, C. Sutan,
V. Malinovschi, A. Stanculescu, O. Korduban, I.N. Mihailescu, Photo-
almost preserved in five consecutive experiments, i.e., the overall catalytic activity of pulsed laser deposited TiO2 thin films in N2, O2 and
decreases in efficiency were in the range 3–9% of the initial CH4, Thin Solid Films 518 (2010) 4648–4653.
values, indicating good catalyst recovery in the purification [6] R. Leary, A. Westwood, Carbonaceous nanomaterials for the enhancement
process applied in a previous study [13]. From the practical point of TiO2 photocatalysis, Carbon 49 (2011) 741–772.
[7] K. Ouyang, S. Xie, X. Ma, Photocatalytic activity of TiO2 supported on
of view, such results indicate the feasibility of efficient application
multi-walled carbon nanotubes, Ceramics International 39 (2013) 7531–
and reusability of the catalysts, which could be a good basis for 7536.
the development and practical application of the obtained photo- [8] K.A. Wepasnick, B.A. Smith, K.E. Schrote, H.K. Wilson, S.
catalysts in environmental clean-up. R. Diegelman, D.H. Fairbrother, Surface and structural characterization
of multi-walled carbon nanotubes following different oxidative treat-
4. Conclusions ments, Carbon 49 (2011) 24–36.
[9] S. Banerjee, T. Hemraj-Benny, S.S. Wong, Covalent surface chemistry of
single-walled carbon nanotubes, Advanced Materials 17 (2005) 17–29.
Improved photocatalytic performance of the synthesized [10] D. Tasis, N. Tagmatarchis, A. Bianco, M. Prato, Chemistry of carbon
nanocomposite photocatalysts was obtained by a sol–gel pro- nanotubes, Chemical Reviews 106 (2006) 1105–1136.
cess, i.e., the hydrolysis of Ti(iso-OC3H7)4 to provide TiO2 [11] D. Eder, Carbon nanotube-inorganic hybrids, Chemical Reviews 110
(2010) 1348–1385.
growth on the surface of differently functionalized MWCNT.
[12] G. Jiang, X. Zheng, Y. Wang, T. Li, X. Sun, Photo-degradation of
Oxidation and subsequent amino-functionalization of MWCNT Methylene Blue by multi-walled carbon nanotubes/TiO2 composites,
introduced significant amounts of functional groups, which play Powder Technology 207 (2011) 465–469.
an important role in the nucleation and binding of TiO2, as well [13] V.R. Djokić, A.D. Marinković, M. Mitrić, P.S. Uskoković, R.D. Petrović,
as contributing to differences in photocatalytic activities of the V.R. Radmilović, Dj.T. Janaćković, Preparation of TiO2/carbon nano-
synthesized nanocomposites. Additionally, the presented pro- tubes photocatalysts: The influence of the method of oxidation of the
carbon nanotubes on the photocatalytic activity of the nanocomposites,
cesses resulted in pure anatase phase TiO2 nanoparticles Ceramics International 38 (2012) 6123–6129.
chemically bonded to the surface of TiO2/modified-MWCNT. [14] G. Vuković, A. Marinković, M. Obradović, V. Radmilović, M. Čolić,
The highest UV-A induced activity was found for TiO2/o- R. Aleksić, P.S. Uskoković, Synthesis, characterization and cytotoxicity
MWCNT and somewhat lower values for TiO2/e-MWCNT and of surface amino-functionalized water-dispersible multi-walled carbon
TiO2/p-MWCNT photocatalysts, indicating that the enhance- nanotubes, Applied Surface Science 255 (2009) 8067–8075.
[15] M.D. Obradović, G.D. Vuković, S.I. Stevanović, V.V. Panić,
ment of photocatalysis was supported by the effective electron P.S. Uskoković, A. Kowal, S.Lj. Gojković, A comparative study of the
transfer properties of oxygen-containing functional groups that electrochemical properties of carbon nanotubes and carbon black, Journal
support efficient charge transportation and separation of the of Electroanalytical Chemistry 634 (2009) 22–30.
4018 V.R. Djokić et al. / Ceramics International 40 (2014) 4009–4018

[16] G.D. Vuković, A.D. Marinković, M. Čolić, M.Ð. Ristić, R. Aleksić, A. coupled with gas chromatography–mass spectrometry, Journal of Chro-
A. Perić-Grujić, P.S. Uskoković, Removal of cadmium from aqueous matography A 1217 (2010) 5462–5469.
solutions by oxidized and ethylenediamine-functionalized multi-walled [22] M. Yan, F. Chen, J. Zhang, M. Anpo, Preparation of controllable
carbon nanotubes, Chemical Engineering Journal 157 (2010) 238–248. crystalline titania and study on the photocatalytic properties, Journal of
[17] B.W. Smith, D.E. Luzzi, Electron irradiation effects in single wall carbon Physical Chemistry B 109 (2005) 8673–8678.
nanotubes, Journal of Applied Physics 90 (2001) 3509–3606. [23] L.W. Zhang, H.B. Fu, Y.F. Zhu, Efficient TiO2 photocatalysts from
[18] H.P. Boehm, in: D.D. Eley, H. Pines, P.B. Weisz (Eds.), Advances in surface hybridization of TiO2 particles with graphite-like carbon,
Catalysis, Vol. 16, Academic Press, New York, 1966, pp. 179–274. Advanced Functional Materials 18 (2008) 2180–2189.
[19] C. Lu, C. Chiu, Adsorption of zinc(II) from water with purified carbon [24] Y. Zhang, Z.-R. Tang, X. Fu, Y.-J. Xu, Engineering the unique 2D mat of
nanotubes, Chemical Engineering Science 61 (2006) 1138–1145. graphene to achieve graphene–TiO2 nanocomposite for photocatalytic
[20] N. Lachman, X. Sui, T. Bendikov, H. Cohen, H. Daniel Wagner, selective transformation: what advantage does graphene have over its
Electronic and mechanical degradation of oxidized CNTs, Carbon 50 forebear carbon nanotube?, ACS Nano 5 (2011) 7426–7435.
(2012) 1734–1739. [25] R. Long, Electronic structure of semiconducting and metallic tubes in
[21] J. Ma, R. Xiao, J. Li, J. Yu, Y. Zhang, L. Chen, Determination of 16 TiO2/Carbon nanotube heterojunctions: density functional theory calcula-
polycyclic aromatic hydrocarbons in environmental water samples by tions, Journal of Physical Chemistry Letters 4 (2013) 1340–1346.
solid-phase extraction using multi-walled carbon nanotubes as adsorbent