You are on page 1of 11

Applied Catalysis A: General 289 (2005) 186–196

www.elsevier.com/locate/apcata

Enhancement of photocatalytic activity of mesoporous TiO2


by using carbon nanotubes
Ying Yu a,c, Jimmy C. Yu b,d, Jia-Guo Yu e, Yuk-Chun Kwok a, Yan-Ke Che f,
Jin-Cai Zhao f, Lu Ding g, Wei-Kun Ge g, Po-Keung Wong a,d,*
a
Department of Biology, The Chinese University of Hong Kong, Shatin, NT, Hong Kong SAR, China
b
Department of Chemistry, The Chinese University of Hong Kong, Shatin, NT, Hong Kong SAR, China
c
College of Physical Science and Technology, Central China Normal University, Wuhan 430079, China
d
Environmental Science Programme, The Chinese University of Hong Kong, Shatin, NT, Hong Kong SAR, China
e
State Key Laboratory of Advanced Technology for Materials Synthesis and Processing, Wuhan University of Technology, Wuhan 430070, China
f
Center for Molecular Science, Institute of Chemistry, The Chinese Academy of Science, Beijing 100080, China
g
Department of Physics, Hong Kong University of Science and Technology, Clear Bay, Kowloon, Hong Kong SAR, China
Received 25 November 2004; received in revised form 18 April 2005; accepted 27 April 2005
Available online 13 June 2005

Abstract

Titanium dioxide/carbon nanotubes (TiO2/CNTs) composites were prepared with the aid of ultrasonic irradiation. Products of different
TiO2:CNTs molar ratio were characterized by X-ray diffraction (XRD), Raman spectroscopy, X-ray photoelectron spectroscopy (XPS),
scanning electron microscopy (SEM), Brunauer–Emmett–Teller (BET) adsorption analysis, thermogravimetric and differential thermal
analysis (TGA–DSC), photoluminescence (PL) and UV–vis spectroscopy measurements (UV–vis). The photocatalytic activity was
evaluated by the degradation of acetone and by the detection of the hydroxyl radical (OH) signals using electron paramagnetic resonance
(EPR). It is found that the crystalline TiO2 is composed of both anatase and brookite phases. The agglomerated morphology and the particle
size of TiO2 in the composites change in the presence of CNTs. The CNTs in the composites are virtually all covered by TiO2. Other than an
increase of the surface area, the addition of CNTs does not affect the mesoporous nature of the TiO2. Meanwhile, more hydroxyl groups are
available on the surface of the composite than in the case of the pure TiO2. The higher the content of CNTs, there is more effective in the
suppression of the recombination of photo-generated e/h+ pairs. However, excessive CNTs also shield the TiO2 from absorbing UV light.
The optimal amount of TiO2 and CNTs is in the range of 1:0.1 and 1:0.2 (feedstock molar ratio). These samples have much more highly
photocatalytic activity than P25 and TiO2/activated carbon (AC) composite. The mechanism for the enhanced photocatalytic activity of
TiO2 by CNTs is proposed.
# 2005 Elsevier B.V. All rights reserved.

Keywords: Carbon nanotubes; Photocatalytic activity; Organic pollutant; TiO2; Sonochemistry; Enhancement; Hydroxyl radical

1. Introduction toxicity and long-term photostability [2]. Unfortunately, the


photocatalytic activity of pure titania is not high enough for
The use of heterogeneous photocatalytic oxidation for industrial purposes [3]. Several methods have been reported
water and air purification and remediation is the subject of a to improve the photocatalytic efficiency. These include
wide range of investigations [1]. Among various oxide increasing the surface area of TiO2, the generation of defect
semiconductor photocatalysts, titania is the most common structures to induce space-charge separation, and the
one because of its strong oxidizing power, absence of modification of TiO2 with metal(s) or other semiconduc-
tor(s) [4–6]. Another method that might possibly increase
* Corresponding author. Tel.: +852 2609 6383; fax: +852 2603 5767. the photocatalytic efficiency of TiO2 is to add a co-sorbent
E-mail address: pkwong@cuhk.edu.hk (P.-K. Wong). such as silica, alumina, zeolites or clay, but no apparent

0926-860X/$ – see front matter # 2005 Elsevier B.V. All rights reserved.
doi:10.1016/j.apcata.2005.04.057
Y. Yu et al. / Applied Catalysis A: General 289 (2005) 186–196 187

improvement of photocatalytic efficiency is observed [7–9]. were filtered, washed with 600 mL distilled water for five
Therefore, the development of new materials for modifying times to remove acid, and finally dried at 105 8C in an oven
TiO2 is urgently needed to increase the photocatalytic (Shel-Lab 1350 FX, Sheldon Manufacturing Inc., USA).
activity of TiO2 for organic pollutant treatment. After treatment, most of the impurities disappear and the
Carbon nanotubes (CNTs) attract considerable attention surface of CNTs becomes cleaner. Moreover, the strong acid
since their discovery [10] due to their special structure, their treatment generates several types of functional groups,
extraordinary mechanical and unique electronic properties including hydroxyl groups (3416 cm1), carboxyl groups
and their potential applications. Their high mechanical (1721 cm1) and carbonyl groups (1583 cm1) [24], on the
strength makes them to be good candidates for advanced surface of the CNTs. The diameters of the treated CNTs
composites [11]. They can either be semiconducting, semi- were about 10–40 nm and the lengths were several
metallic or metallic, depending on the helicity and the micrometers [24].
diameter of the tube [12]. Such variety opens a promising Activated carbon (AC) (F-300) was purchased from
field in nanoscale electrodevice applications. Their large Calgon Carbon Corporation, its surface area is 825 m2/g.
specific surface area, hollow and layered structures indicate Before use, AC was ground to fine powders with a mesh size
that they can be ideal hydrogen storage materials [13]. of about 350 mm.
Recently, researchers found that CNTs are efficient
adsorbents for dioxin, fluoride, lead and cadmium [14– 2.2. Preparation and characteristic of TiO2 and TiO2/
17]. Thus, CNTs can be used as a promising material in CNTs composites
environmental cleaning. CNTs can conduct electrons
[18,19] and have a high adsorption capacity. Anatase CNTs as a support were combined with TiO2 through
TiO2 is known for its superior photocatalytic ability sonochemical and calcination methods. The treated CNTs
compared with other photocatalysts. Therefore, the applica- were used because their adsorption ability is higher than that
tion of CNTs to enhance the photocatalytic activity of TiO2 of raw CNTs [15–17]. To synthesize the TiO2/CNTs
is proposed. composite, titanium tetraisopropoxide (TTIP, Aldrich,
Mesoporous TiO2 is a highly photocatalytically active USA) was used as a titanium source. Firstly, CNTs were
photocatalyst because it has a high surface-to-volume ratio sonicated in 20 mL ethanol for 15 min to disperse them well.
and offers more active sites for carrying out catalytic Then, each mixture of TTIP and CNTs, with a different mole
reactions [20]. Recently, sonochemical processing has been ratio of titanium and carbon (1:0.1, 1:0.2, 1:0.5, 1:1 and 1:2,
proven to be a useful technique in the synthesis of respectively), was sonicated for 30 min to improve the
mesoporous TiO2 with high photocatalytic activity interaction between the two chemicals. After 80 mL Milli-Q
[21,22]. In this study, mesoporous TiO2 and TiO2/CNTs water was added, the sol samples formed by the hydrolysis
composites were prepared by the ultrasound technique. We process were treated with ultrasonic irradiation in an
investigated the effect of CNTs on the microstructure, ultrasonic cleaning bath (Branson B2510E-MTH, USA) for
physicochemical properties, and photocatalytic activity of another 30 min, followed by ageing at 25 8C for 20 h in
TiO2 for the degradation of air pollutant acetone by X-ray order to further hydrolyze the TTIP and form monodispersed
diffraction (XRD), Raman spectroscopy, X-ray photoelec- TiO2 particles [21]. These samples were then dried in a
tron spectroscopy (XPS), scanning electron microscopy 100 8C oven for 8 h. Finally, the dried mixtures were
(SEM), Brunauer–Emmett–Teller (BET) adsorption analy- calcinated in a 400 8C furnace (Ney VULCANTM 3-1750,
sis, thermogravimetric analysis and differential thermal Germany) for 1 h, and the TiO2/CNTs composites, with the
analysis (TGA–DSC), photoluminescence (PL), UV–vis color from white grey to black when increasing the amount
spectroscopy (UV–vis) and electron paramagnetic reso- of CNTs, were obtained. The abbreviations of the TiO2/
nance (EPR). To the best of our knowledge, this is the first CNTs composites with different molar ratios are shown in
report about utilizing TiO2/CNTs composites in photo- Table 1. The actual loading amount of TiO2 in the
catalysis. composites was estimated from TGA–DSC detection and
the data corrected from the content of pure TiO2 in the
sample are also present in Table 1. The preparation process
2. Experimental of TiO2 was similar to that of the composites, except that the
first 15 min sonication for the CNTs dispersion was skipped.
2.1. Materials For comparison, the TiO2/AC composite was also synthe-
sized by a process similar to that used for TiO2/CNTs
CNTs (multiwall CNTs) were synthesized by the composites. The molar ratio of Ti and AC was 1:0.2.
catalytic decomposition of acetylene at 780 8C using Co XRD patterns of the prepared TiO2 and TiO2/CNTs
and Fe catalyst [23,24]. For the purification, 8 g of raw composites were analyzed by a Bruker D8 Advance X-ray
CNTs were boiled in 400 mL of concentrated nitric acid diffractometer (Germany) using Cu Ka radiation at a scan
(65% HNO3, Merck, Germany) for 40 min to get oxygenated rate of 0.058. Raman spectra were detected by a Renishaw
functionalities on the nanotube surface. Then, the CNTs RM3000 Micro-Raman system (UK). XPS spectra were
188 Y. Yu et al. / Applied Catalysis A: General 289 (2005) 186–196

Table 1
Summary of the physicochemical properties of TiO2 and TiO2/CNTs composites
Material TiO2 ratio in Material Anatase Brookite SBET Pore Pore volumef
composites (wt.%)a abbreviations Crystalline Content (%) c content (%) c (m2/g) d sizee (nm) (mL/g)
sizeb (nm)
TiO2 100 TiO2 8.3 83.7 16.3 127 7.05 0.32
TiO2/CNTs (1:0.1)g 97 Ti/C-1 – – – 137 8.46 0.33
TiO2/CNTs (1:0.2)g 95 Ti/C-2 8.3 85.0 15.0 142 9.14 0.35
TiO2/CNTs (1:0.5)g 91 Ti/C-3 8.2 88.0 12.0 147 8.23 0.37
TiO2/CNTs (1:1)g 85 Ti/C-4 – – – 161 9.68 0.39
TiO2/CNTs (1:2)g 75 Ti/C-5 7.6 89.9 10.1 168 8.04 0.42
a
Estimated from the data of TGA–DSC measurement and corrected from the content of pure TiO2 in the sample.
b
Calculated by the Scherrer formula.
c
Calculated using the formula in Ref. [28].
d
BET surface area calculated from the linear part of the BET plot.
e
BJH desorption average pore diameter (4V/A).
f
Single point total pore volume of the pores at P/P0 = 0.97.
g
The mole ratio of TiO2 and CNTs.

recorded on a PHI Quantum 2000 XPS system (USA) with a contaminate chemical. Photocatalytic oxidation of acetone
monochromatic Al Ka source and a charge neutralizer. BET is based on the following reaction [25]:
surface areas (SBET) and porous structure were determined
CH3 COCH3 þ 4O2 ! 3CO2 þ 3H2 O (1)
by using a Micromeritics ASAP 2010 nitrogen adsorption
apparatus (USA). All samples were degassed at 180 8C The photocatalytic activity experiments on TiO2 and TiO2/
before the measurements. The microstructures of the CNTs powders for the oxidation of acetone in air were
samples were determined by using a field emission SEM performed at ambient temperature using a 7000 mL reactor.
(JEOL JSM-6700F) (Japan). TGA–DSC was recorded on a The photocatalysts were prepared by coating an aqueous
NETZSCH STA 449C (Germany) thermal analysis instru- suspension of sample powders onto two dishes with dia-
ment and the scan rate was 108/s at air atmosphere. The PL meters of 7.0 cm. The weight of the photocatalyst used for
emission spectra of the samples were measured using the each experiment was kept at about 0.1 g. For TiO2/CNTs
following procedure: firstly, at 261 8C, a 325 nm He–Cd composites, 0.1 g is the total amount of TiO2 and CNTs. The
laser was used as an excitation light source; then the light photocatalysts were pretreated in an oven at 100 8C for about
from the sample was focused into a spectrometer (Spex500) 2 h and then cooled to 25 8C before use. After the dishes
and detected with a photo-multiplier tube (PMT); and coated with these powder photocatalysts were placed in the
finally, the signal from the PMT was input into a photon reactor, a small amount of acetone was injected into the
counter (SR400) before being recorded by a computer. A reactor. The reactor was connected to a dryer containing
Varian Cary 100 Scan UV–vis system (USA) equipped with CaCl2 that was used for controlling the initial humidity in
a Labsphere diffuse reflectance accessory (USA) was used to the reactor. The analysis of acetone, carbon dioxide, and
obtain the reflectance spectra of the samples over a range of water vapor concentration in the reactor was conducted with
200–800 nm. A Labsphere USRS-99-010 was employed as a a Photoacoustic IR Multigas Monitor (Model 1312,
reflectance standard. EPR signals of radicals spin-trapped INNOVA Air Tech Instruments, Denmark). The acetone
with 5,5-dimethyl-pyrroline-N-oxide (DMPO) were vapor was allowed to reach adsorption equilibrium with
recorded with a Bruker EPR 500 E spectrometer (Germany). the catalyst in the reactor prior to each experiment. The
The irradiation wavelength of the laser exciter is 355 nm. initial concentration of acetone after the adsorption equili-
The settings for the EPR spectrometer were: center brium was about 400 mg/L. This value remained constant
field = 3486.70 G; sweep width = 100.0 G; microwave fre- until a 15 W 365 nm UV lamp (Cole-Parmer Instrument,
quency = 9.82 GHz; power = 5.05 mW. USA) in the reactor was switched on. The initial concentra-
tion of water vapor was 1.20  0.01 vol.%, and the initial
2.3. Measurement of photocatalytic degradation of temperature was 25  1 8C. During the photocatalytic reac-
acetone tion, we observed about a 3:1 ratio of carbon dioxide
products to acetone destroyed, and the acetone concentration
Acetone (CH3COCH3) is a common chemical that is used decreased steadily with increase in UV illumination time.
extensively in a variety of industrial and domestic Each reaction was followed for 60 min.
applications. For example, acetone is frequently used as a The photocatalytic activity of the samples can be
solvent in the printing industry and in analytical labora- quantitatively evaluated by comparing the apparent reaction
tories; it is a major constituent of many common household rate constants. The photocatalytic degradation generally
chemicals [25]. Therefore, we selected it as a model follows a Langmuir–Hinshelwood mechanism [26] with the
Y. Yu et al. / Applied Catalysis A: General 289 (2005) 186–196 189

rate r being proportional to the coverage u: content of TiO2 in a sample can be calculated from the
integrated intensities of anatase (1 0 1), rutile (1 1 0) and
kKc
r ¼ ku ¼ (2) brookite (1 2 1) peaks with the following formulas [28]:
1 þ Kc
kA AA
where k is the true rate constant which includes various WA ¼ (5)
k A AA þ A R þ k B AB
parameters such as the mass of catalyst and the intensity of
light, and K is the adsorption constant. Since the initial AR
WR ¼ (6)
concentration is low (c0 = 400 mg/L = 4.29  103 mol/L), k A AA þ AR þ k B AB
the term Kc in the denominator can be neglected with respect kB AB
to unity and the rate becomes, apparently, first order: WB ¼ (7)
k A AA þ AR þ k B AB
dc WA, WR and WB represent the weight fractions of anatase,
r¼ ¼ kKc ¼ ka c (3)
dt rutile and brookite, respectively. The other symbols AA, AR
where ka is the apparent pseudo-first order rate constant. The and AB are the integrated intensities of anatase (1 0 1), rutile
integral form c = f(t) of the rate equation is (1 1 0) and brookite (1 2 1) peaks, respectively. The vari-
ables kA and kB are two coefficients; their values are 0.886
c0
ln ¼ ka t (4) and 2.721, respectively [28]. The calculated phase contents
c of anatase and brookite in the TiO2 of composites are shown
in Table 1. It can be seen that with the increase of the amount
of CNTs, the size for anatase crystalline becomes smaller,
3. Results and discussion the anatase content in the composites increases and corre-
spondingly the brookite content becomes detrimental gra-
3.1. Characterization of the prepared TiO2 and TiO2/ dually. It is well known that only anatase phase in TiO2 has a
CNTs composites highly photocatalytic activity. Thus, the composites with
high CNTs level may have a good photocatalytic activity.
To characterize the crystalline structure of the samples, Raman spectroscopy was applied to unambiguously
the XRD patterns of TiO2 and TiO2/CNTs composites discriminate the local order characteristics of TiO2 [29,30].
(Fig. 1) are obtained. If one compares these with the pattern The technique is capable of elucidating the photocatalyst
of pure CNTs (Supplementary data), one finds that there is structural complexity as peaks from each material are clearly
no apparent peak in these composites at the positions of 26.0 separated in frequency, and therefore the phases are easily
and 43.4, which are the characteristic peaks for CNTs [27]. distinguishable. Fig. 2 shows the comparison of the Raman
Only anatase (JCPDS no. 21-1272) and brookite (JCPDS no. spectra of TiO2 and the TiO2/CNTs composites. A well-
29-1360) phases attributed to TiO2 are present. The reason resolved TiO2 Raman peak is observed at about 145 cm1
may be that the main peak of CNTs at 26.0 is overlapped for all the samples. This peak is attributed to the main Eg
with the main peak of anatase TiO2 at 25.4 since their anatase vibration mode. Furthermore, vibration peaks at
positions are so close. Moreover, the crystalline extent of 400 cm1 (B1g), 517 cm1 (A1g) and 640 cm1 (Eg) are
CNTs is much lower than that of TiO2, leading to the present in the spectra, indicating that the anatase crystallites
shielding of the peaks of CNTs by those of TiO2. The phase in the materials are the major species [29]. Meanwhile, there
are two weak peaks at 246 and 322 cm1, which are
attributed to A1g and B1g vibration modes for the brookite
phase of TiO2 [31]. No peak assigned to rutile phase is
observed in Fig. 2. Thus, the prepared TiO2 is composed of
anatase phase and brookite phase, which corresponds to the
results shown in Fig. 1. For the spectra of Ti/C-2 and Ti/C-4
in the wave number range between 1200 and 1700 cm1,
there is a broad and weak peak shown at 1600 cm1 in Ti/C-
2. However, the intensity of the peaks at 1332 cm1 and
about 1600 cm1 in the sample of Ti/C-4 is high. The peak at
1332 cm1 is assigned to the ill-organized graphite, the so-
called D-line in CNTs [32]. The broad peak at around
1600 cm1 is an unresolved Raman triplet identified with
tangential carbon atom displacement modes, which is
related to the E2g mode in graphite at 1582 cm1 [33]. With
the increase of the content of CNTs in the composites, it is
Fig. 1. XRD patterns of TiO2 and TiO2/CNTs composites: A denotes reasonable that the peaks attributed to CNTs become strong.
anatase; B denotes brookite. Thus, the peak at about 1340 cm1 in the sample Ti/C-2 is so
190 Y. Yu et al. / Applied Catalysis A: General 289 (2005) 186–196

Fig. 2. Raman spectra for TiO2, Ti/C-2 and Ti/C-4. Inset is the fitting curve of E2g mode of graphite with Lorentzian equation. The superscript A denotes anatase
phase of TiO2; B denotes brookite phase of TiO2; and C denotes CNTs.

weak that it can hardly be observed. By fitting the inset TiO2. The effect of BET surface areas, average pore
figure with a Lorentzian equation, we find that the peak at diameter and pore volumes is listed in Table 1. For the N2
around 1600 cm1 is split into two peaks at 1582 and adsorption–desorption isotherm for TiO2, it is type IV with
1612 cm1, respectively. Although E2g mode in graphite can H2 hysteresis-loop, characteristic of mesoporous materials
be split into A1g, E1g and E2g three modes [34], Raman [36]. The pore size distribution of the sample is narrow and
spectrum of pure CNTs has no peak at 1612 cm1. Thus, the the pore size changes in the range 2.5–12.5 nm. When CNTs
appearance of the new peak at 1612 cm1 indicate that there are composited with TiO2, the N2 adsorption–desorption
is a strong interaction between TiO2 and CNTs through isotherm and pore size distribution of TiO2 do not change
chemical adsorption of carboxyl acid group on the surface of greatly. The composites still have mesoporous structures
the oxidized CNTs onto the surface of TiO2 by the formation although the shape of these hysteresis-loops changes slightly
of ester-like linkages [35]. with the increase of CNTs content. However, the pore size
Fig. 3 shows the effect of CNTs amount on the N2 distribution of the sample Ti/C-5 becomes broader. The pore
adsorption–desorption isotherm and pore size distribution of size is in the range 2.5–17.5 nm. Moreover, with the increase

Fig. 3. BJH pore size distribution plot and N2 adsorption–desorption isotherm (inset) of TiO2 and the TiO2/CNTs composites with different mole ratios.
Y. Yu et al. / Applied Catalysis A: General 289 (2005) 186–196 191

of CNTs content in the composites, the SBET and pore randomly and lack discernible long-range order in the pore
volume of the composites increase gradually. CNTs have a arrangement. It is also found that, with the increase of the
larger surface area of 292 m2/g and their pore size amount of CNTs, the morphology of TiO2 changes from
distribution is broader (Supplementary data). It is reasonable spherical particles (Fig. 4a), big agglomerated particles
that the presence of CNTs in TiO2 leads to the increase of (Fig. 4b) to bulk (Fig. 4c) and there are more and more CNTs
SBET and the presence of a large amount of CNTs results in observed in the composites. Thus, CNTs have an effect on
the broader pore size distribution of this composite Ti/C-5. the agglomerated morphology of TiO2. Meanwhile, the
The mesoporous structure of these samples is formed by agglomeration of monodispersed TiO2 particles can be
the agglomeration of TiO2 nanoparticles under high-intensity clearly observed in the SEM image with a higher resolution
by ultrasound irradiation. Slow hydrolysis under high- (Supplementary data). The spherical TiO2 particles with the
intensity ultrasound irradiation promotes the formation of size of about 40 nm (Fig. 4a) composed of the nanoparticles
monodispersed TiO2 sol particles. Thus, the mesoporous TiO2 with the diameter of about 10 nm, which is basically in
with a relatively narrow pore size distribution is produced by accordance with the crystalline size calculated from their
the ultrasound induced agglomeration of the monodispersed XRD patterns. A similar phenomenon is found in the
TiO2 sol particles [22]. The reason for the fact that the composite Ti/C-4. The bulk TiO2 in the composite is
presence of CNTs does not destroy these mesoporous composed of smaller particles with the size of several
structures may be the presence of ethanol in the sol samples nanometers (Supplementary data). Thus, CNTs also affect
before the strong hydrolysis. The evaporation of ethanol in the size of TiO2 crystallites. The CNTs in the composites are
high temperature leads to the formation of pores, in which well dispersed and almost covered by TiO2 (Fig. 4b and c) so
ethanol once was present. Here, the functions of ethanol are that it can be speculated that under ultrasound irradiation,
not only as a dispersing solvent but also as a soft template. CNTs and TTIP contact with each other. When the
The field emission SEM images of the samples TiO2, the hydrolysis of the sol samples starts, CNTs gradually interact
composites Ti/C-2 and Ti/C-4 are shown in Fig. 4. All with TiO2 through strong chemisorption [35] (Fig. 2). Due to
samples have pores in their structures, which are connected the interaction between TiO2 and CNTs and with the

Fig. 4. SEM images of TiO2 and the TiO2/CNTs composites: (a) TiO2, (b) Ti/C-2, and (c) Ti/C-4.
192 Y. Yu et al. / Applied Catalysis A: General 289 (2005) 186–196

very stable since there is only a slight weight loss (about


3.6%) appears in the range of 100–200 8C, which
corresponds to the removal of the adsorbed water and
may include the evaporation of a small amount of ethanol
(Fig. 5a). The composite Ti/C-2 is not as stable as TiO2
(Fig. 5b). In the range of the measured temperatures, the
total weight loss is 8.5%, larger than that of TiO2.
Meanwhile, there are two observed weight losses: the first
one, in the range of 100–300 8C, corresponds to the removal
of adsorbed water and ethanol, measuring to about 3.5%, the
second one corresponds to a larger endothermic peak at
545 8C, which is resulted from the removal of CNTs in the
composites since the materials start to evaporate at 500 8C
(Supplementary data). Meanwhile, there are no exothermic
peaks at around 450 8C appearing in either of the curves
(Fig. 5a and b) so that there is no transformation of brookite
to anatase phase in these prepared samples [22]. One
possible reason is that the samples have been calcinated in
400 8C for 1 h and the brookite phase has thus become more
stable. The TiO2/CNTs composite is thermally stable, which
is good for the practical applications.
Fig. 6 shows the high-resolution XPS spectra for the O 1s
region for TiO2 and the composite Ti/C-2. The O 1s region is
composed of two peaks. One peak is attributed to the Ti–O in
TiO2 and the composite, while the other one is assigned to
the hydroxyl group. Table 2 lists the results of curve fitting of
XPS spectra for the two samples. The contribution of each

Table 2
Results of curve fitting of high-resolution XPS spectra with Gaussian
equations for the O 1s region of the sample TiO2 and the composite Ti/
C-2
Fig. 5. TGA–DSC curves of TiO2 (a) and Ti/C-2 (b). Materials O 1s (Ti–O) O 1s (OH)
TiO2
Eb (eV) 530.7 531.8
limitation of CNTs, TiO2 crystalline particles may not grow Area 3960 3131
larger and it is not easy for the TiO2 small particles to be ri (%) 55.85 44.15
assembled in an orderly way to form monodisperse spherical Ti/C-2
big particles [37]. Eb (eV) 530.8 531.5
The TGA–DSC curves of the samples of TiO2 and Ti/C-2 Area 2402 2148
are presented in Fig. 5. It is found that the obtained TiO2 is ri (%) 52.79 47.21

Fig. 6. High-resolution XPS spectra of O 1s region of TiO2 and the composite Ti/C-2.
Y. Yu et al. / Applied Catalysis A: General 289 (2005) 186–196 193

type of oxygen is expressed by ri (%). It is noted that the


hydroxyl content in the composite is higher than that in
TiO2, which indicates that the amount of the hydroxyl group
from the chemisorption of water molecules on the surface
TiO2 [38] is enhanced by the presence of CNTs with large
surface area. Although some H2O is easily adsorbed on the
surface of the samples, the physically adsorbed water on the
samples is desorbed under the ultra-high vacuum condition
of the XPS system. Therefore, XPS spectra only show
chemisorbed H2O. The increase in hydroxyl content on the
surface of the composite can enhance the photo-induced
super-hydrophilicity and photocatalytic activity [39,40].

3.2. Photocatalytic activity on the oxidation of acetone


of TiO2 and TiO2/CNTs composites
Fig. 8. UV–vis spectra of the samples of TiO2, Ti/C-1, Ti/C-2 and Ti/C-4.

The photocatalytic activities of TiO2 and TiO2/CNTs


composites were evaluated by using the oxidation of acetone The results in Table 1 show that the anatase phase content
in air. For comparison, the activity of commercial in TiO2/CNTs composites increases with the increase of the
photocatalyst, P25, and synthesized TiO2/AC composite amount of CNTs. Here, the photocatalytic activity of the
were tested under the same conditions. Fig. 7 shows the composites with high CNTs content is much lower,
comparison of the apparent rate constants ka for the different especially for the samples Ti/C-4 and Ti/C-5. Are these
photocatalysts. The value of ka for TiO2 is slightly larger two aspects contradictory? The reason that more CNTs
than that for P25, which is attributed to the porous structure inhibit the photocatalytic activity of TiO2 may be explained
and large surface area of TiO2 prepared by the sonochemical by the UV–vis absorbance spectra of the samples of TiO2,
method. The photocatalytic activity of TiO2 is larger than Ti/C-1, Ti/C-2 and Ti/C-4 (Fig. 8). It is found that, with the
that of TiO2/AC composite, which indicates that AC cannot increase of the CNTs content, the absorption of UV light by
enhance the photocatalytic activity of TiO2 for the the composites decreases greatly. When the molar ratio of
degradation of acetone under the experimental conditions. carbon and titania in the composite is 1:1 (sample Ti/C-4),
When a small amount of CNTs are present in the there is no absorption at all for the composite. The color of
composites, the value of ka increases greatly and rapidly. CNTs is black so that they can shield the UV light for the
However, with the increase of the CNTs content in the absorption by TiO2. If there is no UV absorption for TiO2,
composites, the photocatalytic activity decreases gradually. the sample cannot show any photocatalytic activity. Thus,
The optimum content of CNTs in the composites is 0.1– when the CNTs content in 1 mol of TiO2 is more than
0.2 mol/1 mol of TiO2 (feedstock ratios). The photocatalytic 0.5 mol (feedstock ratio), the photocatalytic activity of the
activities of the samples Ti/C-1 and Ti/C-2 are much higher composites will be lower than that of the pure TiO2, i.e. no
than those of P25 and the TiO2/AC composite. Therefore, enhancement effect of the photocatalytic activity of TiO2 by
CNTs can greatly enhance the photocatalytic activity of CNTs. For a photocatalyst, light absorption is the most
TiO2. critical factor.

3.3. Mechanism for the enhancement of the


photocatalytic activity of TiO2 by CNTs

Activated carbon (AC) has been reported to facilitate the


photocatalytic efficiency of TiO2 for the removal of phenolic
series organic pollutants as a co-adsorbent due to its very
high surface area (more than 800 m2/g) [41–44]. The surface
area of CNTs is much less than that of AC. Thus, adsorption
will not be the only factor for the enhancement of the
photocatalytic activity of TiO2 by CNTs. CNTs have 1D
carbon-based ideal molecule with the nanocylinders, which
can conduct electricity at room temperature with essentially
no resistance. This phenomenon is known as ballistic
Fig. 7. Photocatalytic activity comparison of the samples of TiO2, Ti/C-1, transport [45–47], where the electrons can be considered as
Ti/C-2, Ti/C-3, Ti/C-4, Ti/C-5, P25 and TiO2/AC. The molar ratio of Ti and moving freely through the structure, without any scattering
AC for the TiO2/AC composite is 1:0.2. from atoms or defects. While the electrons formed by the
194 Y. Yu et al. / Applied Catalysis A: General 289 (2005) 186–196

recombination, the sample Ti/C-4 cannot absorb UV light


efficiently, leading to less e and h+ production. Thus, the
photocatalytic activity of this sample is still lower.
Meanwhile, when CNTs are composited with TiO2, the
absorption edge shifts to shorter wavelength, which is a blue
shift (Fig. 8). The band gaps determined by the linear
extrapolation of the steep part of the UV–vis light absorption
toward the baseline [53] for TiO2, Ti/C-1 and Ti/C-2 are
3.17, 3.35 and 3.54 eV, respectively. The band gap of TiO2 is
clearly less than those of the composites. It is commonly
accepted that a larger band gap corresponds to more
powerful redox ability [22]. Because the photocatalytic
process system can be considered to be similar to that of an
electrochemical cell, the increase in band gap results in an
enhanced oxidation–reduction potential [54], the effect of
Fig. 9. Photoluminescence (PL) spectra: comparison of the samples of
which is the same as the suppress reaction of the
TiO2, Ti/C-1, Ti/C-4 and CNTs.
recombination of electron and hole pairs [22]. Thus, the
result of UV–vis spectra is in accordance with that of PL.
UV irradiation migrate to the surface of TiO2, it is easy for Moreover, XPS results (Fig. 6) show that there are more
the electrons to transport in CNTs which are bound with hydroxyl groups provided by chemisorbed water on the
TiO2. Thus, the high rate of e/h+ pair recombination, which surface of TiO2/CNTs composite due to the higher
reduces the quantum yield of the TiO2 process, can decrease. adsorption ability (Table 1), which is also advantageous
This can be confirmed by photoluminescence (PL) spectra of to the increase of photocatalytic activity. The more hydroxyl
the composites. group on the surface of photocatalyst, the more hydroxyl
The PL emission spectra have been widely used to radicals (OH) will be produced by the oxidation of h+ [55].
investigate the efficiency of charge carrier trapping, Fig. 10 shows EPR spectra of the DMPO-OH spin
immigration and transfer, and to understand the fate of adducts produced in the systems of TiO2 and the composite
e/h+ pairs in semiconductor particles [2]. In order to study Ti/C-1 as easily identified by the classical 1:2:2:1 spectral
the effect of CNTs on the recombination of e/h+ produced signature of spin-trapped OH [56]. It can be seen that, under
by TiO2, the PL spectra are detected for the composites the same conditions, the intensity of the 1:2:2:1 spectral
(Fig. 9). TiO2 powder shows a broad PL emission band, signature of the DMPO-OH spin adducts for the composite
which is similar to the results of a previous study [48]. The Ti/C-1 as a photocatalyst is obviously stronger than that with
emission band corresponding to the peak position of about TiO2. Therefore, Ti/C-1 is photocatalytically more active
525 nm is for anatase powder [49]. Compared with the than TiO2 because there are more OH radicals produced by
spectrum of TiO2, it is found that the PL spectra of the
composites Ti/C-1 and Ti/C-4 in the same range of
wavelength are much lower than that of pure TiO2,
especially for sample Ti/C-4, little photoluminescence is
observed. Since the observed PL spectrum in TiO2 can be
attributed to the radiative recombination process of self-
trapped excitations [2,48,50], the reduction of PL intensity
in the sample Ti/C-1 and Ti/C-4 indicates the decrease of the
radiative recombination process. Thus, with the assistance of
CNTs, the recombination of e/h+ excited by TiO2 under UV
light can be detrimental. The capacity of e transport for
CNTs has been widely used for the composition with p-
conjugated polymers for the application of light-emitting
diodes, field effect transistors and photovoltaic devices [51].
It has been confirmed that photo-induced charge transfer
occurs in the electronic interaction between the polymer
chains and CNTs [51]. Since there is a strong interaction
between TiO2 and CNTs (Fig. 2) and they connect well with Fig. 10. The hydroxyl radical (OH) EPR spectra of TiO2 and the composite
each other (Fig. 4), we propose that e transfer also happens Ti/C-1 formed in aqueous dispersions in a Pyrex vessel containing about
20 mg sample after 162 s irradiation by laser light (l = 355 nm). In the
in the TiO2/CNTs composites [52], leading to the reduction systems, the concentration of DMPO is 0.4 mol/L and the added support
of the e/h+ recombination and the increase of the photon Rhodamine B is 2  105 mol/L. The control spectrum for the samples is
efficiency. Although more CNTs can greatly inhibit e/h+ obtained without the laser irradiation.
Y. Yu et al. / Applied Catalysis A: General 289 (2005) 186–196 195

the TiO2/CNTs composite under UV irradiation, which Yu. The authors thank Dr. Z. Jia and Mr. X.X. Ding, Institute
further support the results of PL, UV–vis and XPS. of Nano Science and Technology, Central China Normal
Based on all results mentioned above, CNTs do not only University, for the kind provision of carbon nanotubes and
play the role to prevent e/h+ pairs from recombination but for SEM measurements, respectively.
also may provide a large surface area to lead to more
hydroxyl group on the surface of the TiO2/CNTs. Therefore,
the addition of appropriate amount of CNTs can greatly
Appendix A. Supplementary data
improve the photocatalytic activity of TiO2.
Supplementary data associated with this article can be
found, in the online version, at doi:10.1016/
4. Conclusions
j.apcata.2005.04.057.
The effect of oxidized CNTs on the crystalline phase and
composition, the materials morphology and the physico-
chemical properties of TiO2 is investigated using XRD, References
Raman spectroscopy, BET, SEM, TGA–DSC, XPS, PL,
EPR and UV–vis spectroscopy. The results show that the [1] M.I. Litter, Appl. Catal. B: Environ. 23 (1999) 89.
crystalline part of the obtained TiO2/CNTs composites is [2] J.G. Yu, H.G. Yu, B. Cheng, X.J. Zhao, J.C. Yu, W.K. Ho, J. Phys.
Chem. B 107 (2003) 13871.
composed of anatase and brookite phases, similar to that of [3] Y.T. Kwon, K.Y. Song, W.I. Lee, G.J. Choi, Y.R. Do, J. Catal. 191
the TiO2. CNTs can affect the agglomerated morphology (2000) 192.
and the particle size of the TiO2. Meanwhile, CNTs in the [4] J.C. Yu, L.Z. Zhang, J.G. Yu, New J. Chem. 26 (2002) 416.
composites are all covered by the TiO2 through the [5] C.N. Rusu, J.T. Yates, Langmuir 13 (1997) 4311.
[6] J.C. Yu, J. Lin, R.W. Kowk, J. Photochem. Photobiol. A: Chem. 111
formation of ester-like linkage and there is more hydroxyl
(1997) 1999.
group on the surface of the TiO2/CNTs composite. [7] C. Minero, F. Catozzo, E. Pelizzetti, Langmuir 8 (1992) 489.
Moreover, with the increase of the amount of CNTs, the [8] N. Takeda, T. Torimoto, S. Sampath, S. Kuwabata, H. Yoneyama, J.
surface area of the prepared TiO2/CNTs composites Phys. Chem. 99 (1995) 9986.
becomes larger although the pore size distribution almost [9] J.F. Tanguay, S.L. Suib, R.W. Coughlin, J. Catal. 117 (1989) 335.
remains constant. Both the prepared TiO2 and the TiO2/ [10] S. Iijima, Nature 354 (1991) 56.
[11] L.S. Schadler, S.C. Giannaris, P.M. Ajayan, Appl. Phys. Lett. 73
CNTs composite are thermally stable. (1998) 3842.
The comparison of the photocatalytic activity of TiO2 [12] T.W. Ebbesen, H.J. Lezee, H. Hiura, J.W. Neentt, H.F. Ghaemi, T.
and TiO2/CNTs composites for acetone degradation in air Thio, Nature 382 (1996) 54.
indicates that the presence of a small amount of CNTs can [13] A.C. Dillon, K.M. Jones, T.A. Bekkedahl, C.H. Kiang, D.S. Heben,
enhance photocatalytic activity of TiO2 greatly. However, M.J. Heben, Nature 386 (1997) 377.
[14] R.Q. Long, R.T. Yang, J. Am. Chem. Soc. 123 (2001) 2058.
excess amount of CNTs also shield TiO2 from absorbing [15] Y.H. Li, S. Wang, A. Cao, D. Zhao, X. Zhang, C. Xu, Z. Luan, D. Ruan,
UV. CNTs can enhance the photocatalytic activity of TiO2 J. Liang, D. Wu, B. Wei, Chem. Phys. Lett. 350 (2001) 412.
because the presence of CNTs can prevent the e/h+ pairs [16] Y.H. Li, S. Wang, J. Wei, X. Zhang, C. Xu, Z. Luan, D. Wu, B. Wei,
produced by TiO2 under UV light from recombination. Chem. Phys. Lett. 357 (2002) 263.
[17] Y.H. Li, S. Wang, Z. Luan, J. Ding, C. Xu, D. Wu, Carbon 41 (2003)
That is, electrons excited by TiO2 may easily migrate to
1057.
the nanostructure of the CNTs because of the strong [18] P.J. Brittl, K.S.V. Santhanam, A. Rubio, J.A. Alonso, P.M. Ajayan,
interaction between TiO2 and CNTs. Meanwhile, CNTs Adv. Mater. 11 (1999) 154.
raise the band gap of TiO2, which can make the e/h+ pairs [19] C. Yang, M. Wohlgenannt, Z.V. Vardeny, W.J. Blau, A.B. Dalton, R.
recombination less likely to occur. Moreover, the abundant Baughman, A.A. Zakhidov, Physica B 338 (2003) 366.
hydroxyl groups adsorbed on the large surface of the [20] Y.J. Ying, AIChE J. 46 (2000) 1902.
[21] J.C. Yu, J. Yu, W. Ho, L. Zhang, Chem. Commun. (2001) 1942.
composites can lead to the formation of more OH [22] J.C. Yu, L. Zhang, J. Yu, Chem. Mater. 14 (2002) 4647.
radicals. Therefore, through the two aspects, i.e. the [23] Z. Jia, Z. Wang, L. Liang, B. Wei, D. Wu, Carbon 37 (1999) 903.
decrease of the e/h+ pairs recombination and more [24] Y. Yu, L.L. Ma, W.Y. Huang, F.P. Du, J.C. Yu, J.G. Yu, J.B. Wang, P.K.
hydroxyl group on the surface of the composites, the Wong, Carbon 43 (2005) 670.
[25] M.E. Zorn, D.T. Tompkins, W.A. Zeltner, M.A. Anderson, Appl.
photocatalytic activity of TiO2 can be greatly enhanced by
Catal. B: Environ. 23 (1999) 1.
the presence of CNTs. [26] A. Fernandez, G. Lassaletta, V.M. Jimenez, A. Justo, A.R. Gonzalez-
Elipe, J.M. Herrmann, H. Tahiri, Y. Ait-Ichou, Appl. Catal. B: Environ.
7 (1995) 49.
Acknowledgements [27] S.W. Lee, W.M. Sigmund, Chem. Commun. (2003) 780.
[28] H. Zhang, J.F. Banfield, J. Phys. Chem. 104 (2000) 3481.
[29] P. Falaras, A. Hugot-le Goff, M.C. Bernard, A. Xagas, Sol. Energy
This work was supported by a RGC research grant (No. Mater. Sol. Cells 64 (2000) 167.
CUHK4325/03M) of the Hong Kong SAR Government to [30] I.M. Arabatzis, T. Stergiopoulos, D. Andreeva, S.G. Neophytides, P.
P.K. Wong and a NSFC research grant (No. 20207002) to Y. Falaras, J. Catal. 220 (2003) 127.
196 Y. Yu et al. / Applied Catalysis A: General 289 (2005) 186–196

[31] S.F. Yang, W. Luo, Y.C. Zhu, Y.H. Liu, J.Z. Zhao, Z.C. Wang, G.T. [44] J. Matos, J. Laineb, J.M. Herrmanna, Appl. Catal. B: Environ. 18
Zou, Chem. J. Chin. Univ. 24 (2003) 1933. (1998) 281.
[32] H. Hiura, T.W. Ebbesen, K. Tanigaki, H. Takahashi, Chem. Phys. Lett. [45] J.C. Charlier, Acc. Chem. Res. 35 (2002) 1063.
202 (1993) 509. [46] M. Endo, S. Iijima, M.S. Dresselhaus (Eds.), Carbon Nanotubes,
[33] L. Grigorian, K.A. Williams, S. Fang, G.U. Sumanasekera, A.L. Pergamon Press, Oxford, 1996.
Loper, E.C. Dichey, S.J. Pennycook, P.C. Eklund, Phys. Rev. Lett. [47] X.Z. Li, F.B. Li, C.L. Yang, W.K. Ge, J. Photochem. Photobiol. A:
80 (1998) 5560. Chem. 141 (2001) 209.
[34] M.A. López-Manchado, J. Biagiotti, L. Valentina, J.M. Kenney, J. [48] K. Fujihara, S. Izumi, T. Ohno, M. Matsumura, J. Photochem.
Appl. Polym. Sci. 92 (2004) 3394. Photobiol. A: Chem. 132 (2000) 99.
[35] X. Li, J. Liu, J. Zhang, H. Li, Z. Liu, J. Phys. Chem. B 107 (2003) 2453. [49] H. Tang, K. Prasad, R. Sanjines, P.E. Schmid, F. Levy, J. Appl. Phys.
[36] K.S. Sing, D.H. Evertt, R.A.W. Haul, L. Moscou, R.A. Pierotti, J. 75 (1994) 2042.
Rouquerol, T. Siemieniewska, Pure Appl. Chem. 57 (1985) 603. [50] C. Yang, M. Wohlgenannt, Z.V. Vardeny, W.J. Blau, A.B. Dalton, R.
[37] Y. Yu, L.L. Ma, W.Y. Huang, J.L. Li, P.K. Wong, J.C. Yu, J. Solid State Baughman, A.A. Zakhidov, Physica B 338 (2003) 366.
Chem. 78 (2005) 1488. [51] D.W. Bahnemann, C. Kormannn, M.R. Hoffmann, J. Phys. Chem. 91
[38] J. Yu, J.C. Yu, W. Ho, Z. Zhao, New J. Chem. 26 (2002) 607. (1987) 3789.
[39] A. Fujishima, T.N. Rao, D.A. Tryk, J. Photochem. Photobiol. C: [52] Y. Yu, J.C. Yu, C.Y. Chan, Y.K. Che, J.C. Zhao, L. Ding, W.K. Ge, P.K.
Photochem. Rev. 1 (2001) 1. Wong, Appl. Catal. B: Environ. 61 (2005) 1.
[40] H. Kominami, H. Kumamoto, Y. Kera, B. Ohtani, Appl. Catal. B: [53] J. Lin, J.C. Yu, D. Lo, S.K. Lam, J. Catal. 183 (1999) 368.
Environ. 30 (2001) 29. [54] M.R. Hoffmann, S.T. Martin, W. Choi, W. De Bahnemannt, Chem.
[41] C.G. Silva,J.L. Faria, J.Photochem. Photobiol. A: Chem. 155 (2003)133. Rev. 95 (1995) 69.
[42] T. Torimoto, S. Ito, S. Kuwabata, H. Yoneyama, Environ. Sci. Technol. [55] J. Li, C. Chen, J. Zhao, H. Zhu, J. Orthman, Appl. Catal. B: Environ. 37
30 (1996) 1275. (2002) 331.
[43] J. Arana, J.M. Dona-Rodriguez, E.T. Rendon, C.G. Cabo, O. Gonza- [56] S. Horikoshi, H. Hidaka, N. Serpone, Chem. Phys. Lett. 376 (2003)
lez-Diaz, J.A. Herrera-Melian, J. Perez-Pena, G. Colon, J.A. Navio, 475.
Appl. Catal. B: Environ. 44 (2003) 153.

You might also like