Carbon-containing nano-titania prepared by chemical vapor deposition and its visible-light-responsive photocatalytic activity

Article abstract of DOI:10.1016/j.molcata.2007.01.031

Ultraviolet and visible-light-responsive titania was synthesized and employed in the NO_x photomineralization. A thermal decomposition reaction of titanium isopropoxide was carried out with a metal-organic chemical vapor deposition (MOCVD), enabling continuous production of TiO₂ nanoparticles. Carbon-containing titanium dioxide with the anatase phase prepared at 500 °C under nitrogen atmosphere exhibited high photocatalytic activity for NO oxidation under visible-light illumination. Experimental results indicate that up to 48% removal of NO_x can be achieved in a continuous flow type of reaction system under visible-light illumination (green LED). The chamber temperature in this MOCVD process plays an important role in lattice structure formation, and also affected TiO₂ carbon content. The carbonaceous species on the TiO₂ surface, shown by X-ray diffractometry (XRD), and Raman, UV-vis, and X-ray photoelectron spectroscopies (XPS), is important to the visible-light absorption and visible-light-catalytic mineralization of NO_x.

Full text of DOI:10.1016/j.molcata.2007.01.031

Journal of Molecular Catalysis A: Chemical 270 (2007) 93–100
Carbon-containing nano-titania prepared by chemical vapor deposition
and its visible-light-responsive photocatalytic activity
Chien-Sheng Kuoa, Yao-Hsuan Tsengb, Chia-Hung Huangb, Yuan-Yao Lia,∗
a Department of Chemical Engineering, National Chung-Cheng University, Chia-Yi 621, Taiwan
b Energy and Environment Research Laboratories, Industrial Technology Research Institute, Hsinchu 310, Taiwan
Received 19 July 2006; received in revised form 11 January 2007; accepted 14 January 2007
Available online 24 January 2007
Abstract
Ultraviolet and visible-light-responsive titania was synthesized and employed in the NOx photomineralization. A thermal decomposition reac-
tion of titanium isopropoxide was carried out with a metal-organic chemical vapor deposition (MOCVD), enabling continuous production of TiO2
nanoparticles. Carbon-containing titanium dioxide with the anatase phase prepared at 500 ◦C under nitrogen atmosphere exhibited high photocat-
alytic activity for NO oxidation under visible-light illumination. Experimental results indicate that up to 48% removal of NOx can be achieved in
a continuous flow type of reaction system under visible-light illumination (green LED). The chamber temperature in this MOCVD process plays
an important role in lattice structure formation, and also affected TiO2 carbon content. The carbonaceous species on the TiO2 surface, shown by
X-ray diffractometry (XRD), and Raman, UV–vis, and X-ray photoelectron spectroscopies (XPS), is important to the visible-light absorption and
visible-light-catalytic mineralization of NOx.
© 2007 Elsevier B.V. All rights reserved.
Keywords: Visible-light-responsive photocatalyst; Carbon-containing; Chemical vapor deposition
1. Introduction
Approximately 4% of the energy in the solar light spectrum on
theearth’ssurfacecomesfromthe280to400 nmwavelengthUV
region. Meanwhile, the 400–700 nm wavelength visible-light
region accounts for 45% of the energy. Further, UV light inten-
sity, 0.1–5 ␮W/cm2, is very low in indoor illumination. This
has limited TiO2 applications, giving visible-light-responsive
photocatalysts much more attention for increasing solar and
indoor lighting energy utilization. Previous studies considered
the sol–gel method [6], sputtering plasma [7,8], and metal ion-
implantation [9,10] to enhance the visible-light response of
TiO2. However, the equipment required for these methods is
expensive, and mass production is difficult. Materials such as
CdS, TaN5, TaON, and InVO4 [11,12] also function as visible-
light-responsive photocatalysts. Photocatalytic water splitting
researchers have studied and developed these methods and mate-
rials. However, environmental purification requires considering
the cost and stability of a material for any practical application.
Few feasibility studies exist for these methods and materials
for industrial-scale production due to their higher cost, environ-
ment pollution issues, and material instability. A metal-organic
chemical vapor deposition (MOCVD) is a practical way to pro-
duce visible-light-responsive TiO2 material because it enables
Since the discovery of photoelectrochemical water split-
ting using titanium dioxide electrodes [1], researchers have
extensively studied semiconductor-based photocatalysis. Today,
titanium dioxide is the most common photocatalyst material
due to its strong redox ability and wide use in air purification,
water treatment, deodorization, self-cleaning, sterilization coat-
ing [1–4], and other applications. In environmental applications,
TiO2 exhibits excellent photocatalytic abilities for mineralizing
NOx, SOx, and VOCs under solar irradiation. TiO2 absorbs pho-
ton−s, and evolves active oxygen species, such as OH radicals and
O2 ions, by reacting with H2O and O2 adsorbed on the TiO2
surface. The high oxidation potential of active oxygen species
decomposes many kinds of pollutants [3–5].
Many photocatalytic applications use nano-sized anatase
phase TiO2. A light source shorter than 388 nm in wavelength
can activate a TiO2 photocatalyst with an energy gap of 3.2 eV.
∗
Corresponding author.
E-mail address: chmyyl@ccu.edu.tw (Y.-Y. Li).
1381-1169/$ – see front matter © 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcata.2007.01.031

94
C.-S. Kuo et al. / Journal of Molecular Catalysis A: Chemical 270 (2007) 93–100
the mass production of high quality nanoparticles as well as thin
films [13–15]. This process also controls photocatalyst influen-
tial factors such as composition, crystal structure, purity, and
particle size [15–17].
at X ◦C. If the photocatalysts were prepared in the presence of
oxygen gas, the sample was named TiO2 X(O).
The photocatalyst crystal phase was identified using X-ray
diffractometry with Cu K␣ radiation (Shimadzu XRD 6000).
Material compositions were determined by X-ray photoelec-
tron spectroscopy (Perkin Elmer SSI-M probe XPS system).
Measuring average particle size and morphology was per-
formed by transmission electron microscopy (TEM, Hitachi
S600-100KV). The surface area of the powders was measured
by nitrogen adsorption using the BET equation (Micrometrics
Model ASAP 2400). Ultraviolet–visible light absorption spec-
tra were obtained with a powder spectrophotometer (Shimadzu
UV-2450). Photocatalyst surface structures were identified by a
Raman spectrometer (Renishaw 1000B), and a scattering exper-
iment was conducted with a low-power green laser (30 mW) at
532 nm for 10 s.
This study aims to prepare TiO2 nanoparticles with large
surface areas, narrow particle size distribution, and high visible-
light-responsive activity. Specifically, this paper reports the
thermaldecompositionreactionoftitaniumisopropoxide(TTIP)
to synthesize photocatalyst particles by the MOCVD method
at various growth temperatures. Synthesized nanoparticles were
collected at the reactor outlet via thermophoretic collection [18].
The synthesized photocatalysts were then evaluated for NOx
degradation under visible-light illumination. This study also
carried out various characterization analyses of the visible-light-
responsive TiO2 to determine the crystal phase, particle size,
UV–vis absorption, and surface composition.
2.2. NOx degradation over TiO2 photocatalyst
2. Experimental
Fig. 2 depicts the continuous flow system where NOx degra-
dation was performed. A round-shaped Pyrex glass vessel was
used to conduct the NOx, and a sample dish was located inside
the vessel containing the TiO2 powder. LEDs (visible light) or a
black lamp (UVA light) provided a light source with an intensity
of 1 mW/cm2. The light source spectra, shown in Fig. 3, were
measured using a spectrophotometer (Ocean Optics, USB2000).
The NOx degradation was carried out at room temperature using
an air stream containing 1.0 ppm NO as feedstock. Two mass
flow controllers (MFCs) (Brooks 5850E) manipulated relative
humidity in the feeding stream. The reaction gas in the feed-
ing stream passed through the vessel containing TiO2 powder
(0.2 g) at a flow rate of 1 L/min. An on-line chemilumines-
cent NOx analyzer (Eco Physics, CLD 700 AL) continuously
monitored NO and NO2 concentrations for gas analysis in the
outlet.
2.1. Apparatus, experimental conditions and
characterization
Fig. 1 shows a diagram of the experimental setup, which was
reported in a previous study [15]. This setup consists of stain-
less steel gas flow lines, mass flow controllers, a syringe pump,
and a quartz tube (25 mm o.d., 23 mm i.d., and 1000 mm long)
in an electric furnace functioning as the reactor. The hot zone
was 300 mm long. A mass flow controller and a syringe pump
controlled the gas and titanium isopropoxide flow rates, respec-
tively. A mechanical pump enabled a low pressure operation,
to a few Torr, and a convectron gauge measured the chamber
pressure. A stainless steel net trapped the product in a collection
system assembled with a 4 ◦C cooling jacket, which reduced
particle aggregation.
The quartz tube was purged by nitrogen during the initial
heating step. Upon reaching the desired reaction temperature
(between 400 and 1000 ◦C), the furnace reactor was filled with
a mixture of TTIP (1.0 mL/min) as the catalyst precursor and
nitrogen (500 sccm) as the carrier gas. Oxygen gas was also
introduced at 100 sccm, as necessary. The collection system
gathered the resulting TiO2 powder. After a 10 min reaction, the
reactor returned to room temperature under a 100 sccm nitrogen
flow. The synthesized TiO2 materials was named TiO2 X when
prepared in the MOCVD process under a nitrogen atmosphere
3. Results and discussion
3.1. Photocatalyst activity under UV- and visible-light
irradiation
Many researchers use NO oxidation to determine photo-
catalytic reactivity in various TiO2 photocatalytic applications
[19–24]. The electron–hole pair (e−–h+) generated upon light
Fig. 1. Schematic diagram of the TiO2 synthesis reactor.

C.-S. Kuo et al. / Journal of Molecular Catalysis A: Chemical 270 (2007) 93–100
95
Fig. 2. Schematic diagram of continuous flow system for photocatalytic NOx removal.
NO2 + OH∗ → HNO3
NO + OH∗ → HNO2
(5)
(6)
excitation is trapped at the TiO2 surface as spatially separated
redox active sites. Some studies report the formation of reactive
oxygen species, such as superoxide ion (O2•−), atomic oxygen
(O), O−, OH and HO2 radicals on the surface of TiO2 irradiated
with UV light [25–27]. The general mechanism of NOx oxida-
tion by photocatalyst is as follows. Hydrogen ions and hydroxide
ions are dissociated from water. The active oxygen species are
produced on the TiO2 surface.
Finally, the nitric acid forms on the catalyst. Photocatalyst
activity lessens as acid accumulates. To illustrate the reaction
behavior of NO photocatalytic oxidation, 0.2 g of TiO2 500 and
1 mW/cm2 of light intensity were used to conduct the experi-
ment. Fig. 4 shows the NOx oxidation results under UVA and
visible light performed in a continuous flow system. At first,
a 1 ppm NO gas stream was introduced into the reactor under
dark conditions. The NO concentration in the gas phase dropped
due to NO adsorption on the TiO2. A few minutes later, the
NO was saturated on the TiO2, and the NO concentration was
returned to 1 ppm. As Fig. 4 shows, the steady state of this pho-
tocatalytic reaction was achieved as soon as the photocatalyst
was illuminated. The NO concentration decreased from 1 to
0.46 ppm, and the NO2 concentration increased to 0.06 ppm.
The NOx concentration, which includes NO and NO2 concen-
trations, was maintained at 0.52 ppm for 30 min under green
LED illumination. TiO2 500 activity did not obviously decrease
−
O2 + e− → O2
(1)
(2)
(3)
OH− + h+ → OH∗
H+ + O2− → HO2
∗
The nitric monoxide is oxidized to nitric acid or nitrous acid by
active oxygen species. Based on the gas-phase chemistry of NOx
[28], NO is converted to HNO3 as a consecutive photooxidation
via a NO2 intermediate.
NO + HO2∗ → NO2 + OH∗
(4)
Fig. 4. Reaction profiles of NO over the TiO2 500 photocatalyst during 1.5 h
on stream; catalyst loading: 0.2 g; intensity of irradiation: 1 mW/cm2; inlet con-
centration of NO: 1 ppm; inlet flow rate: 1 L/min; reaction temperature: 27 ◦C.
Fig. 3. UVA lamp and LED photon energy distribution profiles (the visible-light
region of UVA lamp is small and not shown in this figure).

96
C.-S. Kuo et al. / Journal of Molecular Catalysis A: Chemical 270 (2007) 93–100
Fig. 5. Average removal rate of NOx (␮mol/h) on various photocatalysts
vs. wavelength of irradiation; catalyst loading: 0.2 g; irradiation intensity:
1 mW/cm2; inlet concentration of NO: 1 ppm; inlet flow rate: 1 L/min; reaction
temperature: 27 ◦C.
Fig. 6. XRD patterns of prepared photocatalysts.
over time, probably due to a large surface area of 109 m2/g.
Under a blue LED, NO, NO2, and NOx concentrations decreased
to 0.38, 0.04, and 0.42 ppm, respectively. The amount of NOx
removal increased as the illumination wavelength decreased, but
theamountofNO2 generatedincreasedastheilluminationwave-
length increased. This is because the number of induced active
sites on the TiO2 surface decreased with the increase in the wave-
length of illumination, and NO2 is more difficult to oxidize than
NO. Hence, the decreasing reaction rate under visible light is
primarily caused by the decrease in the NO2 oxidation rate.
The comparison above used commercially available photo-
catalysts (Degussa P25). Fig. 5 shows NOx removal rates using
various photocatalysts under UV- and visible-light irradiation.
Under UV illumination, these photocatalysts exhibit a similar
activity in NOx mineralization, except for TiO2 900 and
TiO2 1000. TiO2 400 was inactive. Although the photoactivi-
ties of P25 and TiO2 500(O) were good under UV illumination,
their visible-light activities were low. These results demonstrate
that the activity of prepared TiO2 decreases as the synthesis
temperature increases. Under visible light, the activity levels of
these photocatalysts fell in the order of TiO2 500 > TiO2 600 >
TiO2 700 > TiO2 800 > TiO2 900 > TiO2 1000–TiO2 500(O) >
P25. This order of activity is the same under different visible-
light sources (blue and green lights). Adding oxygen gas to
the synthesis reaction obviously decreased the visible-light
response of prepared TiO2. Therefore, the optimal CVD
reaction temperature is 500 ◦C, without adding oxygen gas, for
the highest visible-light-responsive activity.
Thepeaksforthe(1 0 1), (0 0 4), (2 0 0), (1 0 5), and(2 1 1)reflec-
tions characteristic of the TiO2 anatase phase appeared on the
material at a synthesis temperature above 500 ◦C. Process tem-
peratures ranging from 500 to 900 ◦C produce the anatase phase.
The anatase phase was partially transformed to the rutile phase
at 1000 ◦C [29]. The TiO2 anatase phase is more photoactive
than the rutile and brookite phases [30]. Sclafani and Hermann
report that titania samples with the rutile phase were consistently
much less photoactive than anatase samples [31]. Therefore, it
was strongly suggested the synthesis temperature should not
exceed 1000 ◦C due to the appearance of the rutile phase. The
crystallite sizes of the prepared photocatalysts were determined
from a half width of (1 0 1) peak using the Scherrer formula
(d = 0.9 λ/β cos θ). As shown in Table 1, vapor molecule accre-
tion caused the crystallite size of TiO2 to gradually increase as
the temperature increased. Cluster–cluster and particle–particle
collided rapidly, accelerating crystallite growth size (primary
particle) and cluster (secondary particle) [32]. Porter et al.
reports that crystallite size of sol–gel-synthesized TiO2 and P25
rapidly increased over 700 ◦C [33]. However, the crystallite size
of MOCVD-synthesized TiO2 did not obviously increase with
the temperature, and the rutile phase began to appear at a very
Table 1
Characterization of prepared TiO2
Sample
C content
(at.%)
Surface area
(m2/g)
Particle
size (nm)a
Crystallite
size (nm)b
TiO2 500
TiO2 600
TiO2 700
TiO2 800
TiO2 900
TiO2 1000
TiO2 500(O)
22.2
20.4
17.7
16.5
14.2
67.3
5.63
109.4
61.0
58.9
41.6
29.0
27.0
138.3
14.1
25.3
26.2
37.1
53.1
57.1
11.2
9.8
10.6
11.2
13.5
15.9
16.2
7.7
3.2. TiO2 photocatalyst characterization
3.2.1. XRD
Fig. 6 shows the X-ray diffraction patterns of TiO2 sam-
ples. At a synthesis temperature of 400 ◦C, no detectable peaks
indicate the presence of an amorphous structure. This lack is
responsible for the very low photocatalytic activity of TiO2 400.
a
Calculated by BET formula, dBET = 6/ρA.
b
Calculated by Scherrer formula, d = 0.9λ/β cos θ.

C.-S. Kuo et al. / Journal of Molecular Catalysis A: Chemical 270 (2007) 93–100
97
to 900 ◦C. The broad and low blue-shift peaks of TiO2 500 and
TiO2 600 were caused by a low degree of crystallization on the
surface and small particles. The broad and weak intense peaks
of TiO2 1000 included both anatase and rutile phases, 440 and
608 cm−1, due to the mixed and incomplete lattice structure
(i.e., amorphous-like structure) on TiO2 surface, which was not
observed in the XRD pattern. The amorphous-like structure was
probably caused by impurities such as carbonaceous species,
covered or doped on the TiO2 surface.
3.2.3. TEM
The morphology of the prepared photocatalysts was observed
by TEM. Fig. 8 shows TEM images of TiO2 500 and TiO2 1000,
where the mean particle sizes are close to 15 and 65 nm,
respectively. The size distribution of TiO2 500 particles was
narrow and non-aggregated. In contrast, the sol–gel nanoparti-
cle preparation method easily produced particle agglomeration
after calcination at high temperature. The MOCVD process can
produce a good visible-light-responsive TiO2 without agglomer-
ation under proper conditions. The particle sizes ranged from 10
to 70 nm, which increased as the reaction temperature increased.
Morphological observation by TEM and XRD calculations
reveal that the large particle size of TiO2 1000 was mainly
caused by agglomeration, not the growth of crystallite size.
Fig. 7. Raman spectra of prepared photocatalysts.
high temperature (1000 ◦C). This is due to a short retention time
(0.25 s) of TTIP in the heating zone during the synthesis process.
The heat effect on crystallite size growth was not as obvious as
in previous studies.
3.2.4. BET
3.2.2. Raman spectroscopy
Table 1 shows the specific surface area of samples prepared
by the BET method. Agglomeration caused the surface area to
decrease as the synthesis temperature increased. Furthermore,
TEM observation confirmed that the prepared TiO2 powders
had an almost spherical and nonporous form. Therefore, this
relationship can be expressed by a simple equation, assuming a
spherical and nonporous particle:
Raman spectroscopy can be used to clearly characterize the
TiO2 surface structure [20,34]. Raman spectroscopy wave vec-
tor selection rules, which limit the wave vector of detectable
photons to the Brillouin zone center, break down at a very small
size (>10 nm). As a result, all the spectra peaks broadened as
the particle size decreased. Iida et al. [34] and Li et al. [35]
both claim that this method is more sensitive for nanometer-
sized crystals than X-ray diffraction. This study used lower laser
power to avoid laser heating or laser damage effect. Fig. 7 shows
TiO2 Raman spectra at different synthesis temperatures. The
Raman spectra were consistent with XRD patterns for TiO2 500
to TiO2 900. The anatase phase, Raman shift = 396, 517, and
638 cm−1 [36], existed in all TiO2 samples, and the anatase
phase peaks were narrowed with increasing synthesis tempera-
ture except for TiO2 1000. This shows that the anatase crystallite
size increased as the synthesis temperature increased from 500
6
ρA
dBET
=
(7)
where dBET is the calculated particle size (nm), ρ the density of
TiO2 (3.84 and 4.20 g/cm3 for anatase and rutile, respectively),
and A is the specific surface area (m2/g) [37]. Table 1 shows
that the calculated diameter of the particles is about 14–57 nm.
The results of specific surface area agree well with the TEM
observation, but were different from the XRD calculations. Yue
Fig. 8. TEM photograph of (a) TiO2 500 and (b) TiO2 1000.

98
C.-S. Kuo et al. / Journal of Molecular Catalysis A: Chemical 270 (2007) 93–100
to completely convert to C3H6 and CO2 gas. This caused the
carbonaceous species, existing in the condensed and coke-like
structure, to remain on the photocatalyst. Table 1 shows the car-
bon content in the TiO2 at different synthesis temperatures. The
amount of carbon on the TiO2 500(O) is relatively low because
the alkoxide group was completely oxidized to CO2 gas in the
presence of sufficient oxygen gas. The amount of carbon on the
TiO2 surface decreased as the synthesis temperature increased,
except for the TiO2 1000. The information from Fig. 5 and
Table 1 indicates that the visible-light-responsive activity of pre-
pared TiO2 increased as the amount of carbon increased on the
TiO2, until 900 ◦C. The amount of carbon on the TiO2 surface is
responsible for good visible-light absorbance and visible-light-
responsive activity. This finding agrees well with previous works
by the same authors, and with Maier’s experiment [19–21,45].
However, these visible-light-responsive TiO2 samples were pre-
pared in the CVD process under an anaerobic atmosphere, so a
few oxygen atoms on the TiO2 surface were substituted by car-
bon to form a TiO2−xCx structure. The oxygen defects of TiO2
possibly act as a recombination center, reducing photoactivity.
XPS analysis of Ti 2p3/2 did not immediately reveal the Ti–C
bond, so the TiO2−xCx structure of these samples is probably
scarce. In this case, the small number of oxygen defects in the
TiO2 samples does not enhance the recombination rate between
excited electrons and holes, where the similar UV-responsive
photoactivities of TiO2 500(O) and TiO2 500 samples obvi-
ously demonstrate this ratiocination, as shown in Fig. 5. The
amount of carbon in the TiO2 1000 material increased to 67%,
but its UV- and visible-light-responsive activities were smaller
than that of TiO2 500 to TiO2 900. TTIP thermal decomposi-
tion at an extremely high temperature caused this high amount of
carbonaceous species. Many TTIP alkoxide groups transformed
to highly condensed species, like coke, and were deposited and
doped onto the TiO2 surface. Thus, some photoactive sites were
covered by coke species, causing an incomplete lattice structure
(amorphous-like surface structure) on the TiO2 1000 sample.
This is shown in the Raman spectra. Therefore, the visible-light-
responsive activity of TiO2 1000 was less than TiO2 500–900 in
spite of its higher amount of carbonaceous species. In addition,
after the Ar+ ion etching process of TiO2 500, the peak intensi-
ties of carbonaceous species decreased. After 120 s, carbon was
not detected on the TiO2. The etching experiment shows that the
carbonaceous species mainly exist on the surface.
Fig. 9. C 1s XPS spectra of prepared photocatalysts.
and coworkers claim that the specific surface area of sol–gel-
synthesized TiO2 depends only on the size of crystals (primary
particle) and was unaffected by aggregation [38]. This is prob-
ably because aggregation in the sol–gel process is loose, so the
N2 gas can penetrate the gap between aggregated particles. In
this study, TiO2 aggregation occurs mostly at high tempera-
tures (>500 ◦C); therefore, the particles are tightly aggregated
with high surface energy. Okuyama and coworkers report this
phenomenon as incomplete sintering in the CVD process [16].
Hence, the specific surface area is more related to secondary
particles than primary particles in this MOCVD process. More-
over, the specific surface area of TiO2 500(O) was larger than
all prepared samples due to a rapid nucleation rate in an oxygen
atmosphere [39]. Nanoparticle size in the MOCVD process is
determined by the relative magnitudes of nucleation and growth
rate. The TTIP gas was totally decomposed to the monomer
(TiO2(g)), after which all monomers were converted into TiO2(s)
in a very short time. This prevented further particle growth.
Therefore, it is reasonable to assume that the grain size decreased
after oxygen gas was added to the reaction atmosphere.
3.2.5. XPS
Some studies discuss carbon-doped and carbon-covered
titania materials [19,20,40–46]. These researchers report that
carbonaceous species play a sensitizer role, inducing visible-
light absorption and response. To investigate the carbon states on
the prepared TiO2, the C 1s core levels were measured by XPS,
as shown in Fig. 9. All samples were cleaned with irradiation
of 2 mW/cm2 UV light for 12 h before XPS analysis, removing
the originally adsorbed organic compounds. Three XPS peaks
occurred at 284.8, 286.7, and 288.8 eV for the prepared TiO2.
The first value arises from adventitious elemental carbon (C–C),
and the other two small peaks indicate the existence of C–O and
C O, corresponding to a carbonate species [40,41]. The C–O
and C O were not chromophores, so the C–C structure on the
TiO2 surface is most likely responsible for visible-light absorp-
tion. The carbonaceous species covered the TiO2 by the thermal
decomposition reaction of TTIP. Without oxygen gas added to
the reaction atmosphere, the TTIP alkoxide group was unable
3.2.6. UV–vis diffuse reflectance spectra
Fig. 10 shows the UV–vis diffuse reflectance spectra of
prepared TiO2 materials and P25. This reflectance data was
converted by instrument software to absorbance values, F(R),
based on the Kubelka–Munk theory [19,20,23,42,45]. The
visible-light absorbance of the prepared TiO2 depends on the
synthesis temperature and the presence of oxygen gas. The
visible-light absorption of these photocatalysts was in the order
of TiO2 1000 > TiO2 500 > TiO2 600 > TiO2 700 > TiO2 800 >
TiO2 900 > TiO2 500(O)–P25. Except for TiO2 1000, the order
of visible-light-responsive activities was the same as the order
of visible-light absorption. In this study, all photocatalysts
prepared in an oxygen-free atmosphere exhibited much bet-

C.-S. Kuo et al. / Journal of Molecular Catalysis A: Chemical 270 (2007) 93–100
99
visible-light-responsive photocatalysts. These materials can be
practically applied in the purification process in both water and
air without UV illumination.
Acknowledgment
We thank the Ministry of Economic Affairs of the Republic
of China, Taiwan, for financially supporting this research.
References
[1] A. Fujishima, K. Honda, Nature 238 (1972) 37.
[2] T. Ibusuki, K. Takeuchi, J. Mol. Catal. 88 (1994) 93.
[3] M.R. Hoffmann, S.T. Martin, W. Choi, D.W. Bahnemann, Chem. Rev. 95
(1995) 69.
[4] A. Fujishima, N.R. Tata, A.T. Donald, J. Photochem. Photobiol. C 1 (2000)
1.
[5] A. Mills, S.L. Hunte, J. Photochem. Photobiol. A 108 (1997) 1.
[6] T. Ohno, M. Akiyoshi, T. Umebayashi, K. Asai, T. Mitsui, M. Matsumura,
Appl. Catal. A: Gen. 265 (2004) 115.
Fig. 10. UV–vis absorption spectra of various photocatalysts.
[7] R. Asahi, T. Morikawa, T. Ohwaki, K. Aoki, Y. Taga, Science 293 (2001)
269.
[8] I. Nakamura, N. Negishi, S. Kutsuna, T. Ihara, S. Sugihara, K. Takeuchi,
J. Mol. Catal. A 161 (2000) 205.
[9] H. Yamashita, M. Harada, J. Misaka, M. Takeuchi, K. Ikeue, M. Anpo, J.
Photochem. Photobiol. A 148 (2002) 257.
[10] M. Anpo, M. Takeuchi, K. Ikeue, S. Dohshi, Curr. Opin. Solid State Mater.
6 (2002) 381.
[11] G. Hitoki, A. Ishikawa, T. Takata, J.N. Kondo, M. Hara, K. Domen, Chem.
Lett. 7 (2002) 736.
[12] G. Hitoki, T. Takata, J.N. Kondo, M. Hara, H. Kobayashi, K. Domen, Chem.
Commun. 16 (2002) 1698.
[13] W. Li, S. Ismat Shah, M. Sung, C.P. Huang, J. Vac. Sci. Technol. B 20
(2002) 2302.
[14] S.C. Jung, S.J. Kim, N. Imaishi, Y.I. Cho, Appl. Catal. B: Environ. 55
(2005) 253.
[15] C.S. Kuo, Y.H. Tseng, Y.Y. Li, Chem. Lett. 35 (2006) 356.
[16] K. Nakaso, K. Okuyama, M. Shimada, S.E. Pratsinis, Chem. Eng. Sci. 58
(2003) 3327.
[17] O.J. Jung, S.H. Kim, K.H. Cheong, W. Li, S. Ismat Saha, Bull. Korean
Chem. Soc. 24 (2003) 49.
[18] Y. Sun, T. Egawa, L. Zhang, X. Yao, Jpn. J. Appl. Phys. 41 (2002)
1389.
[19] Y.M. Lin, Y.H. Tseng, J.H. Huang, C.C. Chao, C.C. Chen, I. Wang, Environ.
Sci. Technol. 40 (2006) 1616.
ter visible-light-responsive activities than commercial powder
(P25). Furthermore, the TiO2 500(O) sample, prepared in an
oxygen atmosphere, and P25, did not perform well in visible-
light absorption due to the absence of carbonaceous species.
TiO2 1000 possessed the highest visible-light absorption due
to its huge amount of highly dense coke and amorphous-like
surface structure. The visible-light absorptions of TiO2 500 and
TiO2 600 were higher than others, except for TiO2 1000. The
amount of carbonaceous species on the surface of TiO2 500
and TiO2 600 was higher than on TiO2 700, TiO2 800, and
TiO2 900, but not very different. Raman analysis reveals a low
crystallization on the surface of the TiO2 500 and TiO2 600,
so the loose and amorphous-like structures of TiO2 500 and
TiO2 600 were probably responsible for their better visible-
light absorbance. In this experiment, TiO2 700, TiO2 800, and
TiO2 900 exhibited well-crystallized surface structures, so their
visible-light absorbance was only due to the existence of car-
bonaceous species. Therefore, carbon in the samples may play
two roles; it is the sensitizer for visible-light absorption due to its
coke-like structure; and it also exists as an impurity, the lattice
defect, of TiO2 to form interface states that effectively lower the
band gap.
[20] Y.H. Tseng, C.S. Kuo, C.H. Huang, Y.Y. Li, P.W. Chou, C.L. Cheng, M.S.
Wong, Nanotechnology 17 (2006) 2490.
[21] P.W. Chou, S. Treschev, P.H. Chung, C.L. Cheng, Y.H. Tseng, Y.J. Chen,
M.S. Wong, Appl. Phys. Lett. 89 (2006) 131919.
[22] N. Negish, K. Takeuchi, T. Ibusuki, J. Mater. Sci. 33 (1998) 5789.
[23] T. Sano, N. Negishi, K. Koike, K. Takeuchi, S. Matsuzawa, J. Mater. Chem.
14 (2004) 380.
[24] S. Matsuda, H. Hatano, A. Tsutsumi, Chem. Eng. J. 82 (2001) 183.
[25] A. Walker, M. Formenti, P. Meriaudeau, S.J. Teichner, J. Catal. 50 (1977)
237.
[26] W.F. Behnke, F. Nolting, C. Zetzsch, J. Aerosol Sci. 18 (1987) 65.
[27] T. Ibusuki, K. Takeuchi, Atmos. Environ. 20 (1986) 1711.
[28] R. Atkinson, D.L. Baulch, R.A. Cox, R.F. Hampson Jr., J.A. Kerr, J. Troe,
J. Phys. Chem. Ref. Data 18 (1989) 887.
[29] Y. Sun, T. Egawa, L. Zhang, X. Yao, Jpn. J. Appl. Phys. 41 (2002)
945.
4. Conclusions
Carbon-containing TiO2 photocatalysts were successfully
prepared by an MOCVD process using titanium isopropoxide
as the precursor. Carbon-containing TiO2 showed a relatively
high photocatalytic activity in NOx oxidation under visible-
light illumination. The visible-light activity of prepared TiO2
decreased as the synthesis temperature increased. The optimal
synthesistemperaturewas500 ◦Cinanoxygen-freeatmosphere.
The coke-like carbonaceous species on the TiO2 surface, as evi-
denced by XPS and Raman spectroscopy, are responsible for
good visible-light absorption and photoactivity. The amount of
TiO2 carbonaceous species and visible-light response decreased
as synthesis temperature increased, up to 900 ◦C. In conclusion,
a simple and continuous process was achieved for preparing
[30] A. Masakazu, Y. Hiromi, I. Yuichi, E. Shaw, J. Electroanal. Chem. 396
(1995) 21.
[31] A. Sclafani, J.M. Herrmann, J. Phys. Chem. 100 (1996) 13655.
[32] Y. Sun, A. Li, M. Qi, L. Zhang, X. Yao, Mater. Sci. Eng. 86 (2001) 185.
[33] J.F. Porter, Y.G. Li, C.K. Chan, J. Mater. Sci. 34 (1999) 1523.

100
C.-S. Kuo et al. / Journal of Molecular Catalysis A: Chemical 270 (2007) 93–100
[34] Y. Iida, M. Furukawa, T. Aoki, T. Sakai, Appl. Spectrosc. 52 (1998) 673.
[35] M. Li, Z. Feng, G. Xiong, P. Ying, Q. Xin, C. Li, J. Phys. Chem. B 105
[40] M. Janus, B. Tryba, M. Inagaki, A.W. Morawaski, Appl. Catal. B: Environ.
52 (2004) 61.
(2001) 8107.
[41] M. Inagaki, F. Kojin, B. Tryba, M. Toyoda, Carbon 43 (2005) 1652.
[42] S. Sakthivel, H. Kisch, Angew. Chem. Int. Ed. 42 (2003) 4908.
[43] Z. Jiang, C. Lu, H. Wu, Ind. Eng. Chem. Res. 44 (2005) 4165.
[44] T. Ohno, M. Akiyoshi, T. Umebayashi, K. Asai, T. Mitsui, M. Matsumura,
Catal. Lett. 98 (2004) 255.
[36] M.P. Moreta, R. Zallena, D.P. Vijayb, S.B. Desu, Thin Solid Films 366
(2000) 8.
[37] Y.H. Tseng, H.Y. Lin, C.S. Kuo, Y.Y. Li, C.P. Huang, React. Kinet. Catal.
L 89 (2006) 63.
[38] A.J. Maira, K.L. Yeung, C.Y. Lee, P.L. Yue, C.K. Chan, J. Catal. 192 (2000)
185.
[45] C. Lettmann, K. Hildenbrand, H. Kisch, W. Macyk, W.F. Maier, Appl.
Catal. B: Environ. 32 (2001) 215.
[39] Y. Sun, A. Li, M. Qi, L. Zhang, X. Yao, J. Mater. Sci. 37 (2002) 1343.
[46] S.U.M. Khan, M.W.B.I. Al-Shahry Jr., Science 297 (2002) 2243.