Applied Catalysis B: Environmental 36 (2002) 31–43

Enhanced photocatalytic activity of mesoporous and

ordinary TiO2 thin films by sulfuric acid treatment
Jimmy C. Yu a,∗ , Jiaguo Yu a,b , Jincai Zhao c
a Department of Chemistry and Materials Science and Technology Research Centre, The Chinese
University of Hong Kong, Shatin, New Territories, Hong Kong, China
b State Key Laboratory of Advanced Technology for Materials Synthesis and Processing, Wuhan

University of Technology, Wuhan 430070, China

c The Laboratory of Photochemistry, Center for Molecular Science, Institute of Chemistry,

The Chinese Academy of Sciences, Beijing 100080, China

Received 6 June 2001; received in revised form 21 September 2001; accepted 21 September 2001

Abstract
Transparent anatase mesoporous TiO2 (MTiO2 ) and TiO2 nanometer thin films were prepared on soda-lime glass and fused
quartz via the reverse micellar method and sol–gel method, respectively. The as-prepared MTiO2 and TiO2 films were then
treated by dipping them in a H2 SO4 solution. The MTiO2 and TiO2 films before and after surface acid treatment were character-
ized by X-ray photoelectron spectroscopy (XPS), atomic force microscopy (AFM), transmission electron microscopy (TEM),
X-ray diffraction (XRD), BET surface area and UV–VIS spectrophotometry. The photocatalytic activity of the samples was
evaluated by photocatalytic oxidation of acetone in air. It was found that MTiO2 thin films showed higher photocatalytic activity
than that of the TiO2 thin films. This was attributed to the fact that MTiO2 thin films were composed of smaller monodisperse
spherical particles about 15 nm and had higher specific surface areas. Furthermore, the monodispersity of TiO2 particles was
beneficial to transfer and separation of photo-generated electrons and holes in the inner of and on the surface of TiO2 particle
and reduced the recombination of photo-generated electrons and holes. The films deposited on quartz showed the highest
photocatalytic activity because films deposited on quartz exhibited a better crystallization and had no sodium contaminant. The
photocatalytic activity of MTiO2 and TiO2 thin films deposited on different substrates after treated with H2 SO4 solution was sig-
nificantly enhanced. Acid treatment was particularly effective for MTiO2 /glass and TiO2 /glass, which showed activity enhance-
ment of four and over two times, respectively. This increase in activity has been correlated with the reduction of sodium ions and
the increase in the adsorbed hydroxyl content on the surface of TiO2 films. © 2002 Elsevier Science B.V. All rights reserved.
Keywords: Mesoporous TiO2 thin films; Reverse micellar method; Surface acid treatment; Photocatalytic activity; Acetone oxidation

1. Introduction and Honda discovered the photocatalytic splitting of

water on TiO2 electrodes in 1972 [1]. Recently, the
Titanium dioxide has been intensively investigated application of TiO2 photocatalysts has mainly been
as a semiconductor photocatalyst since Fujishima focused on the elimination of toxic and hazardous or-
ganic substances and metals in wastewater, drinking
∗ Corresponding author. Tel.: +852-2609-6268; water, and air, which is important for the protection
fax: +852-2603-5057. of the environment [2–4]. Conventional powder pho-
E-mail address: [email protected] (J.C. Yu). tocatalysts, however, have a serious limitation—the

0926-3373/02/$ – see front matter © 2002 Elsevier Science B.V. All rights reserved.
PII: S 0 9 2 6 - 3 3 7 3 ( 0 1 ) 0 0 2 7 7 - 6
32 J.C. Yu et al. / Applied Catalysis B: Environmental 36 (2002) 31–43

need for post-treatment separation in a slurry system. and shape of TiO2 nanoparticles not only in
This can be overcome by immobilizing TiO2 particles suspensions but also in films [20,21]. Formation of
as thin films on a solid substrate. These materials mesoporous TiO2 films by reverse micelles involves
would have new industrial applications including three stages [20]: (1) hydrolysis of an alkoxide in
antibacterial ceramic tiles and self-cleaning glass a reverse micellar environment and formation of a
[5–12]. gel by inorganic polymerization; (2) deposition of a
The sol–gel method of preparing TiO2 films on the composite inorganic/organic film on a solid substrate;
substrates has many advantages over other methods (3) heating of the film up to 450–500 ◦ C to burn out
such as chemical vapor deposition (CVD), plasma all organic substance and obtain the final mesoporous
spraying, anodization, and thermal oxidation of Ti structure. The resulting mesoporous TiO2 thin films
metal mainly because of the following: (1) no special have high crystallinity and large surface area [20],
apparatuses are required; (2) uniform multicomponent which satisfies the basic requirement for photoactive
films can easily be formed if a homogeneous solution TiO2 films. Therefore, reverse micelles should be a
is available; (3) phase structure of films can be con- most promising method for making highly photoac-
trolled, in this case TiO2 (anatase) may be obtained; tive TiO2 films.
(4) the resultant films may possibly be characterized If a TiO2 film coated on the surface of common
by a porous structure with a large specific area charac- soda-lime glass shows a high photocatalytic activity,
teristic of a gel [13]. TiO2 films of high photocatalytic it would be quite useful as self-cleaning or indoor air
activity have been prepared by the sol–gel method cleaning windows [22]. It is known, however, that the
[14,15]. Recently, TiO2 thin films of high photocat- diffusion of Na+ ions into the nascent TiO2 film from
alytic activity were also obtained from a TiO2 sol the SL-glass substrates during heat-treatment process
prepared by hydrothermal treatment of peroxotitanic significantly deteriorates the photocatalytic activ-
acid sol [16,17] or an alkoxide-derived TiO2 gel [18]. ity [6–9]. To prevent the negative effect of sodium
It is well known that the photocatalytic activity of ions on photocatalytic activity many methods such
TiO2 thin films strongly depends on the preparing as pre-coating soda-lime glass with a SiO2 layer or
method and post-deposition treatments, since they soaking the glass in a HNO3 solution to exchange
give decisive influence on the chemical and physical sodium ions have been reported [5–8]. Hattori et al.
properties of TiO2 included in the films. Therefore, reported that a TiO2 film formed on the SL-glass sub-
it is necessary to choose adequate processing con- strate pre-coated with a F:SiO2 layer showed a higher
ditions to fabricate highly active photocatalytic thin photocatalytic activity than that of the TiO2 film
films [19]. coated on the quartz substrate [22]. Recently, we have
The preparation of mesoporous TiO2 thin film is not reported that the photocatalytic activity of a sol–gel
entirely new. Stathatos et al. [20] and Stathatos and TiO2 film coated on SL-glass could be enhanced by
Lianos [21] have reported the preparation of meso- dipping the TiO2 film in a HCl solution [9]. However,
porous TiO2 thin film by reverse micellar method The enhancement of photocatalytic activity of the
and photocatalytically deposited silver nanoparti- TiO2 thin film was limited due to the adsorption of
cles on titanium dioxide nanoparticle films. How- Cl− ions on the surface of the films.
ever, photocatalytic activity of mesoporous TiO2 This is the first report on the comparative studies
thin films for oxidizing organic contaminants in of photocatalytic activity of mesoporous TiO2 thin
air has never been reported. We found that meso- films and TiO2 thin films prepared by reverse micellar
porous TiO2 thin films prepared by reverse micellar method and the sol–gel processing, respectively. We
method exhibited a higher photocatalytic activity also found that the photocatalytic activity of MTiO2
than TiO2 thin film prepared by sol–gel method and TiO2 thin films on SL-glass substrate could
for the photocatalytic oxidation of acetone in air. be greatly improved by H2 SO4 solution treatment.
Usually, TiO2 colloidal suspensions or films are pre- Explanations are provided based on the influence
pared by hydrolysis of alkoxide precursors in an of surface microstructure of mesoporous TiO2 thin
aqueous environment. However, reverse micelles pro- films and surface acid treatment on photocatalytic
vide a means to control the size, size-polydispersity activity.
 J.C. Yu et al. / Applied Catalysis B: Environmental 36 (2002) 31–43 33

2. Experimental sol which had the same composition as the thin film
samples, was dried at 100 ◦ C in air in order to obtain
2.1. Preparation and surface acid treatment gels, then heat-treated at a rising rate of 3 ◦ C/min up
to and at 500 ◦ C for 1 h in air, and finally ground by
All chemicals used in the present work were from agate mortar to obtain MTiO2 and TiO2 powders.
Aldrich and were used as received. Millipore water To reduce sodium and other impurities on the sur-
was used in all experiments. face of MTiO2 and TiO2 films and enhance the pho-
Mesoporous TiO2 thin films were prepared by the tocatalytic activity of the films, we dipped the freshly
method reported by Stathatos et al. [20] and Stathatos prepared MTiO2 and TiO2 films in H2 SO4 aqueous
and Lianos [21]. Triton X-100 (26 g) and cyclohexane solutions at room temperature. The concentration of
(150 ml) were first mixed to form a reverse micellar so- H2 SO4 was 1 M. After 2 h, the films were withdrawn
lution under stirring vigorously. After 0.5 h, 1.08 g wa- from the H2 SO4 solution. They were carefully rinsed
ter was added to the solution at room temperature. The with de-ionized water and dried at 100 ◦ C for 20 min
solution appeared turbid; however, it clears out upon in an oven in air.
addition of titanium isopropoxide (23 g) [21]. The re-
sultant alkoxide solution was kept stirring at room tem- 2.2. Characterization
perature for hydrolysis reaction for 1 h, resulting in
the TiO2 sol. The concentration of Triton X-100, tita- Surface roughness and morphologies of TiO2 thin
nium isopropoxide and water was 0.2, 0.4 and 0.3 M, films were evaluated by atomic force microscopy
respectively. The mesoporous TiO2 films formed on (AFM: Nano Scope 3a, Digital Instruments Inc.,
the substrates were prepared from the above TiO2 sol Santa Barbara, CA). The X-ray diffraction (XRD)
solution by dipping-withdrawing in an ambient atmo- patterns, obtained on a Philips MPD 18801 X-ray
sphere. The substrates coated with TiO2 gel films were diffractometer using Cu K␣ radiation at a scan rate
heat-treated in air at a rate of 3 ◦ C/min up to 500 ◦ C of 0.05◦ (2θ )/s, were used to determine the iden-
and were left to stay in the furnace at the highest tity of any phase present and their crystallite size.
temperature for about 1 h. The withdrawal speed was The accelerating voltage and the applied current are
4 mm/s. The thickness of the TiO2 films was adjusted 35 kV and 20 mA, respectively. The crystallite size
by repeating the cycle from dipping to heat treatment. was calculated from X-ray line broadening analysis
Precursor solutions for TiO2 films were prepared by Scherrer formula. Phase identification and crys-
by the following method [3,8]. Titanium isopropox- tallite size measurements were also conducted using
ide and triethanolamine were dissolved in ethanol. transmission electron microscopy (TEM) (Jcol Co.,
After stirring vigorously for 1 h at room temperature, 1200EX, Japan) when the thickness of the films de-
the mixture of water and ethanol was added drop- posited on glass is less than 0.2 ␮m. UV–VIS spectra
wise to the solution with a burette under stirring. of films were obtained using a UV–VIS spectropho-
The resultant alkoxide solution was kept standing tometer (Cary 100 Scan Spectrophotometers, Varian,
at room temperature for hydrolysis reaction for 2 h, U.S.A.). X-ray photoelectron spectroscopy (XPS)
resulting in the TiO2 sol. The chemical composition measurements were performed on a PHI Quantun
of the starting alkoxide solution was Ti(OC3 H7 )4 : 2000 XPS System with a monochromatic Al K␣
C2 H5 OH:H2 O:N(C2 H4 OH)3 = 1:26.5:1:1 in molar source and a charge neutralizer; all the binding ener-
ratio. Soda lime glass (83 mm × 98 mm × 2 mm) and gies were referenced to the C 1s peak at 284.8 eV of
quartz (80 mm×30 mm×2 mm) were used as the sub- the surface adventitious carbon. The film thickness
strates for thin films. The substrates coated with TiO2 was measured using cross-sectional observation of
gel films were heat-treated at 500 ◦ C for 1 h in air in scanning electron microscopy (SEM) and surface pro-
a furnace. The withdrawal speed also was 4 mm/s. filer (Alpha-step 500). The Brunauer–Emmett–Teller
Apart from the above-described thin film samples, (BET) surface area (SBET ) and pore parameters of
powder samples were also prepared with the same the powder samples were determined by nitrogen
methods as the thin films. The detailed process was adsorption–desorption isotherm measurements at 77 K
as follows: Firstly, TiO2 reverse micelle sol, or TiO2 on a Micromeritics ASAP 2000 nitrogen adsorption
34 J.C. Yu et al. / Applied Catalysis B: Environmental 36 (2002) 31–43

apparatus. All the samples measured were degassed rate constants. The photocatalytic degradation gen-
at 180 ◦ C before the actual measurements. Pore size erally follows a Langmuir–Hinshelwood mechanism
distributions were calculated from adsorption branch [2,3,11] with the rate r being proportional to the cov-
of the isotherm by the Barrett–Joyner–Halenda (BJH) erage θ
method using the Halsey equation. kKc
r = kθ = (2)
(1 + Kc)
2.3. Photocatalytic activity
where k is the true rate constant which includes various
Acetone (CH3 COCH3 ) is a common chemical that parameters such as the mass of catalyst, the intensity
is used extensively in a variety of industrial and of light, etc. and K is the adsorption constant. Since the
domestic applications. For example, acetone is fre- initial concentration is low (co = 400 ppm = 4.29 ×
quently used as a solvent in the printing industry and 10−3 mol/l), the term Kc in the denominator can be
in analytical laboratories; it is a major constituent neglected with respect to unity and the rate becomes,
of many common household chemicals [10]. There- apparently, first order
fore, we chose it as a model contaminate chemical. dc
Photocatalytic oxidation of acetone is based on the r=− = kKc = ka c (3)
dt
following reaction [10,23]:
where ka is the apparent rate constant of pseudo-first
CH3 COCH3 + 4O2 → 3CO2 + 3H2 O (1) order. The integral form c = f (t) of the rate equation
is
The photocatalytic activity experiments on TiO2 thin co
films for the oxidation of acetone in air were per- ln = ka t (4)
c
formed at ambient temperature using a 7000 ml reac-
tor [23]. The area of the film photocatalyst used for
each experiment was kept at about 140 cm2 . After the 3. Results and discussion
film photocatalysts were placed in the reactor, a small
amount of acetone was injected into the reactor. The 3.1. Thickness, coloration and transmittance of
reactor was connected to a dryer containing CaCl2 MTiO2 and TiO2 thin films
that was used for controlling the initial humidity in
the reactor. The analysis of acetone, carbon dioxide, Table 1 shows that the thickness of both MTiO2 and
and water vapor concentration in the reactor was TiO2 thin films is dependent on the number of coating
conducted with a Photoacoustic IR Multigas Monitor cycles. Coating a glass or quartz substrate three times
(INNOVA Air Tech Instruments Model 1312). The would form a thin film that is slightly more than twice
acetone vapor was allowed to reach adsorption equi- as thick. When other parameters are kept constant,
librium with the catalyst in the reactor prior to an the MTiO2 thin films are always a litter thicker than
experiment. The initial concentration of acetone after that of TiO2 . It is interesting to note that glass and
the adsorption equilibrium was about 400 ± 2 ppm, quartz coatings are of about the same thickness. The
which remained constant until a 15-W 365 nm UV MTiO2 and TiO2 films thus prepared show coloration
lamp (Cole-Parmer Instrument Co.) in the reactor was phenomenon. The color strongly depends on the film
switched on. The initial concentration of water vapor thickness. As the thickness of TiO2 films increases
was 1.20 ± 0.01 vol.%, and the initial temperature was to more than 1 ␮m, the coloration becomes indistinct,
25 ± 1 ◦ C. During the photocatalytic reaction, a near while the films still remain transparent. Therefore, it
3:1 ratio of carbon dioxide products to acetone de- can be concluded from these results that the coloration
stroyed was observed, and the acetone concentration is caused by the interference of light [24].
decreased steadily with increase in UV illumination MTiO2 and TiO2 films prepared by the above
time. Each reaction was followed for 60 min. methods are optically transparent. Fig. 1 shows the
The photocatalytic activity of the films can be quan- UV–VIS absorption spectra of MTiO2 and TiO2 films
titatively evaluated by comparing the apparent reaction deposited on quartz by one coating cycle, together
 J.C. Yu et al. / Applied Catalysis B: Environmental 36 (2002) 31–43 35

Table 1
The thickness and crystallite size of the TiO2 films deposited on different substrates
Samples Coating cycles Film thickness (␮m) Phase structure Crystalline size (nm) Measure method

TiO2 /quartz-1 1 0.18 Anatase 16.7 XRD

TiO2 /quartz-3 3 0.42 Anatase 23.5 XRD
TiO2 /glass-1 1 0.17 Anatase 17 TEM
TiO2 /glass-3 3 0.40 Anatase 23.1 XRD
MTiO2 /quartz-1 1 0.21 Anatase 12.8 XRD
MTiO2 /quartz-3 3 0.46 Anatase 18.9 XRD
MTiO2 /glass-1 1 0.20 Anatase 13 TEM
MTiO2 /glass-3 3 0.44 Anatase 18.5 XRD

with that of the plain quartz and glass substrates. The oscillation of the curve between 800 and 380 nm is
transmittance of quartz is the highest, about 93% over duo to the interference between the TiO2 film and the
the entire 300–800 nm spectral region. The transmitt- substrate. Compared with the transmittance spectra
ance of glass is about 91% over the visible light spec- of TiO2 film, the absorption edge of MTiO2 films is
tral region and its adsorption edge is about 330 nm. observed at a shorter wavelength range than that of
The transmission spectra of MTiO2 and TiO2 films TiO2 film. The shift is considered to occur due to the
on quartz show some different features. For MTiO2 difference in crystallites within the films. MTiO2 film
films transmittance is about 80% at wavelength near consists of relatively small crystallites and shows a
800 nm and gradually drops at short wavelengths pseudo-“blue shift” like fine particle [4,25–28]. The
until it reaches its first minimum value of 70% at difference of transmittance between MTiO2 and TiO2
near 650 nm. At shorter wavelengths, the transmit- films was attributed to the difference in their film
tance rises and reaches its maximum value of 88% at thickness and absorption of light. The thickness of
510 nm. After that, the transmittance decreases and these films ranges between 170 and 210 nm.
rises again, and at about 380 nm it decreases rather Fig. 2 shows UV–VIS absorption spectra of MTiO2
quickly, finally approaches zero at around 330 nm. and TiO2 film deposited on soda lime glass by one
The fast decrease below 380 nm is due to absorption coating cycle before and after treated with 1 M H2 SO4
of light caused by the excitation of electrons from the solution for 2 h. The transmittance of the films after
valence band to the conduction band of TiO2 . The acid solution treatment increases slightly in most of

Fig. 2. UV–VIS spectra of MTiO2 (a and b) and TiO2 (c and d)

Fig. 1. UV–VIS absorption spectra of quartz (a), glass (b), MTiO2 film deposited on soda lime glass by one coating cycle before (a
(c), and TiO2 (d) films deposited on quartz by one coating cycle. and c) and after (b and d) treated with 1 M H2 SO4 solution for 2 h.
36 J.C. Yu et al. / Applied Catalysis B: Environmental 36 (2002) 31–43

the visible light region. This is due to the reduction obviously different. Fig. 3a and b show that MTiO2
of sodium ions and other impurities on the surface of film is composed of monodisperse spherical parti-
the films. Therefore, it can be concluded from these cles of about 15 nm and has a mesoporous structure
experiment results and the changes in transmittance between monodisperse TiO2 particles [20]. Monodis-
that MTiO2 and TiO2 films show good stability to acid persity is the advantage of the reverse micellar route
solution treatment. SEM measurements also show that for the synthesis of TiO2 nanoparticles by hydrolysis
the surface of MTiO2 and TiO2 films before and after of titanium isopropoxide [20,21]. This hydrolysis pro-
surface acid treatment has no obvious changes. cess competes with the hydration of surfactant polar
heads for water molecules. Restructuring of surfactant
3.2. Atomic force microscopy molecules around the polar species formed during hy-
drolysis results in growth limitations and uniform par-
Fig. 3a–d shows 2- and 3D AFM images of the ticles sizes [21]. It is also observed from Fig. 3c and
surface of MTiO2 and TiO2 thin films deposited on d that the TiO2 film also has granular microstructure
quartz by one coating cycle. The surface morpholo- and is composed of about 80 nm sphere particles. But
gies and roughness of MTiO2 and TiO2 films are porous structure between TiO2 particles is not clear.

Fig. 3. AFM 2- and 3D images of the surface of MTiO2 (a and b) and TiO2 (c and d) thin films deposited on quartz by one coating cycle.
 J.C. Yu et al. / Applied Catalysis B: Environmental 36 (2002) 31–43 37

According to the next XRD results, these 80 nm par-

ticles on the surface of TiO2 thin films are aggregates
of many small TiO2 crystallites. In addition to particle
diameter, AFM image analysis also gives the values
of surface roughness. The root mean square rough-
ness values (Rrms ) of MTiO2 and TiO2 are 1.038 and
0.560 nm, respectively. Fig. 3b and d also show that
surface morphology of MTiO2 film is rougher than
that of TiO2 film.

3.3. Crystalline phase and crystallite size of

thin films

Fig. 4 shows XRD patterns of TiO2 films deposited Fig. 5. TEM photograph of TiO2 film deposited on glass with
on glass (a and b) and quartz (c and d) with differ- 0.17 ␮m.
ent thickness. The XRD patterns of the 0.17 ␮m TiO2
film (sample a) deposited on glass show no trace of with an increasing in the film thickness. This indicates
anatase peak. This is not surprising as the presence the formation of greater crystallites in the films. The
of sodium ions in soda lime glass can affect the for- average crystallite size of TiO2 films was detected by
mation of anatase crystals [25]. It can also be seen using line broadening methods and the equation pro-
from Fig. 4 that TiO2 films deposited on quartz ex- posed by Scherrer. The results are shown in Table 1.
hibits a better crystallization than those deposited on Fig. 4 also shows that the anatase TiO2 films exhibit
glass for the same thickness of the films. Moreover, obvious orientation effect in the (1 0 1) peak.
the peaks appear more intense by increasing the film Fig. 5 shows TEM images of the crystallite of the
thickness. However, the width at the half height of the TiO2 film deposited on glass with 0.17 ␮m thick-
main peak of the anatase structure becomes narrower ness. The crystallites are spherical particles and the
crystallite size is about 17 nm. Fig. 6 presents the
electron diffraction ring patterns from the above TiO2
nanocrystalline film that are identified as anatase
[25]. These results indicate that anatase crystals were
indeed present but were of too small an amount to be

Fig. 4. XRD patterns of TiO2 films deposited on glass (a and b)

and quartz (c and d) (the thickness of the films (a–d) is 0.17, 0.40, Fig. 6. Electron diffraction of TiO2 film deposited on glass with
0.18 and 0.42 ␮m, respectively). 0.17 ␮m.
38 J.C. Yu et al. / Applied Catalysis B: Environmental 36 (2002) 31–43

detected by XRD. TEM measurements found that the Table 2

crystallite size was about 13 nm when the thickness BET surface areas and pore parameters of MTiO2 and TiO2 pow-
ders at 500 ◦ C
of MTiO2 film deposited on glass is 0.20 ␮m. These
results are also shown in Table 1. It can be seen from Sample SBET a Porosityb Pore volumec Pore sized
(m2 /g) (ml/g) (nm)
Table 1 that the average crystallite size increases
with increasing in film thickness. This is due to the MTiO2 59.9 23.9 0.085 3.3
TiO2 9.1 9.8 0.029 6.5
growth of the nanocrystals in the interior region of
a BET surface area calculated from the linear part of the BET
the thin films from repeated coatings and longer heat
plot (P /P0 = 0.05–0.3).
treatment. b The porosity is estimated from the pore volume determined

using the adsorption branch of the N2 isotherm curve at the

3.4. BET surface area P /P0 = 0.995 single point.
c Total pore volume, taken from the volume of N adsorbed at
2
P /P0 = 0.995.
The BET surface areas of MTiO2 and TiO2 thin d Average pore diameter, estimated using the adsorption branch
films on quartz or glass substrates could not be of the isotherm and the BJH formula.
measured directly by nitrogen adsorption apparatus
because the amount of the thin films was too small. size distribution calculated from adsorption branch
We measured the BET surface areas of powder sam- of the nitrogen isotherm by the BJH method gives
ples prepared using the same procedure as for the an average pore size of 3.3 nm. The pore diameter
thin films. Fig. 7 shows the pore size distribution distribution is narrow with a range of 2.0–5.8 nm.
curve calculated from adsorption branch of nitro- It is these mesopore structures that allow rapid dif-
gen isotherm by the BJH method using the Halsey fusion of various reactants and products during UV
equation. The inset shows the corresponding nitrogen illumination and enhance the rate of photocatalytic re-
adsorption–desorption isotherms of MTiO2 powders action. The BET surface area and pore parameters of
prepared from the TiO2 reverse micellar solution and MTiO2 and TiO2 powders determined from nitrogen
calcined at 500 ◦ C for 1 h. The sharp decline in des- adsorption–desorption isotherm by the BJH method
orption curve is indicative of mesoporosity, the pore are summarized in Table 2. The pore size distribution

Fig. 7. Pore size distribution curve calculated from adsorption branch of the nitrogen isotherm by the BJH method. Inset: the corresponding
nitrogen adsorption–desorption isotherms of MTiO2 powders prepared from reverse micelle TiO2 solution and calcined at 500 ◦ C for 1 h.
 J.C. Yu et al. / Applied Catalysis B: Environmental 36 (2002) 31–43 39

curve and nitrogen adsorption and desorption Table 3

isotherms of TiO2 powders (not shown here) are Composition (at.%) of MTiO2 and TiO2 thin films before and after
treated with H2 SO2 solution according to XPS analysesa
different from those of MTiO2 powders. The BET
surface area of TiO2 powders is much smaller than Films Ti O C Na Si
that of MTiO2 powders. Moreover, the pore size dis- MTiO2 /quartz 27.4 70.4 2.2
tribution is very wide, ranging from 1.5 to 70 nm. It MTiO2 /quartz + H2 SO4 30.9 67.3 1.8
can also be seen from Table 2 that the porosity and MTiO2 /glass 26.0 56.5 3.4 13.0 1.1
pore volume of MTiO2 powders are larger than those MTiO2 /glass + H2 SO4 30.4 67.7 1.4 0.5
TiO2 /quartz 32.1 66.3 1.6
of TiO2 powders. TiO2 /quartz + H2 SO4 32.0 66.5 1.5
TiO2 /glass 26.2 60.4 2.3 9.9 1.2
3.5. XPS studies TiO2 /glass + H2 SO4 31.2 67.0 1.1 0.7
a The atomic percentage of carbon subtracts the adventitious

Fig. 8 shows the XPS survey spectra for the sur- hydrocarbon.
face of TiO2 films deposited on quartz (a) and glass
before (b) and after (c) treated with 1 M H2 SO4 so-
lution for 2 h. It can be seen from Fig. 8 that TiO2 that some chemical reactions occur at the interface be-
film deposited on quartz only contains Ti, O, and C tween the films and the glass substrates and sodium
elements. The photoelectron peak for Ti 2p appears and silicon ions migrate from the glass substrates into
clearly at a binding energy, Eb , of 458 eV, O 1s at the films [8,9,11]. As shown in Table 3, the amounts of
E b = 530 eV and C 1s at E b = 285 eV. However, sodium on the surface of the film as measured by XPS
TiO2 film deposited on glass before acid treatment not are 9.9 and 13.0 at.% for TiO2 /glass and MTiO2 /glass
only contains Ti, O and C, but also Na and Si. The samples, respectively. This incorporation of Na could
Nakll and Si 2p photoelectron peaks are at binding en- also account for the less crystallized state of titanium
ergies 496 and 102.5 eV, respectively. The XPS peaks dioxide when it is deposited on glass [6–8]. The el-
for Na and Si are observed in the spectra, implying ement C in the Fig. 7 is attributed to the residual
carbon from precursor solution and the adventitious
hydrocarbon from the XPS instrument itself. The C
content in Table 3 only is the residual carbon (such
as unhydrolyzed alkoxide groups, carbonate and so
on) from precursor solution, and the adventitious hy-
drocarbon from specimen handing or pumping oil is
subtracted [29]. It can be seen that the C content of
samples TiO2 /glass and MTiO2 /glass after acid treat-
ment greatly decrease. This is probably due to the
partial dissolution of the residual carbon, such as car-
bonate. On the other hand, the C content of samples
TiO2 /quartz and MTiO2 /quartz is lower than that of
samples TiO2 /glass and MTiO2 /glass, moreover, the
dissolution amount of the residual carbon of the for-
mer is obviously smaller than that of the latter. This
is attributable to the fact that samples TiO2 /glass and
MTiO2 /glass contain sodium contamination, because
the sodium ions usually adsorb CO2 in air to form car-
bonate and cause the increase in the carbon amount,
furthermore, the carbonate dissolves easily in H2 SO4
Fig. 8. XPS survey spectra for the surface of TiO2 films deposited solution. As can be seen in Fig. 8 and Table 3, sam-
on quartz (a), and glass before (b), and after (c) treated with 1 M ples TiO2 /glass and MTiO2 /glass after acid treatment
H2 SO4 solution for 2 h. contain neither S nor Na. We believe that the sodium
40 J.C. Yu et al. / Applied Catalysis B: Environmental 36 (2002) 31–43

Table 4
Results of curve-fitting of the high resolution XPS spectra for the
O 1s regiona
Film O 1s O 1s
(Ti–O) (OH)
MTiO2 /quartz Eb (eV) 530.0 531.9
ri (%) 89.2 10.8
MTiO2 /glass Eb (eV) 529.9 531.9
ri (%) 83.3 16.7
TiO2 /quartz Eb (eV) 529.8 531.9
ri (%) 91.3 8.7
TiO2 /glass Eb (eV) 529.8 531.9
ri (%) 84.9 15.1
MTiO2 /quartz + H2 SO4 Eb (eV) 530.0 531.9
ri (%) 82.4 17.6
MTiO2 /glass + H2 SO4 Eb (eV) 530.0 531.9
ri (%) 75.5 24.5
TiO2 /quartz + H2 SO4 Eb (eV) 529.8 531.9
ri (%) 83.5 16.5
TiO2 /glass + H2 SO4 Eb (eV) 529.9 531.9
ri (%) 76.3 23.7
a r (%) represents the ratio A /


i i Ai (Ai is the area of each
peak).

film samples before and after surface acid treatment,

where ri (%) shows the ratio of each kind contribu-
tion to the total of all the two kinds oxygen contribu-
tions. As can be seen in Table 4 and Fig. 9a and b,
the hydroxyl content of sample MTiO2 /glass is greater
than that of sample MTiO2 /quartz. This is because
MTiO2 /glass contains sodium ions and easily adsorbs
water from air. The hydroxyl content in the all samples
Fig. 9. High resolution XPS spectra of the O 1s region for the after acid treatment is greater than that before treat-
surface of MTiO2 films deposited on quartz (a), and on glass
before (b), and after (c) treated with 1 M H2 SO4 solution.
ment. The hydroxyl content increases after TiO2 films
were treated with acid solution, which can be ascribed
to the fact that some adsorbed H2 O can react with
ions on the surface of TiO2 films are completely ex- TiO2 and forms Ti–OH, such as, H2 O + Ti–O–Ti →
changed with hydrogen ions in the H2 SO4 solution. 2Ti–OH [9]. Hydroxyl group existed in the films is at-
The residual sodium and sulfate ions are then removed tributable to the chemically adsorbed H2 O. Although
by rinsing with de-ionized water. some H2 O is physically adsorbed on the surface of
Fig. 9 shows the high resolution XPS spectra of the TiO2 films, the physically adsorbed H2 O on TiO2 is
O 1s region, taken on the surface of MTiO2 film depo- easily desorbed under the ultra-high vacuum condi-
sited on quartz (a) and glass before (b) and after (c) tion of the XPS system. Therefore, the XPS spectra
surface acid treatment. The O 1s region is composed do not show the physically adsorbed H2 O on samples.
of two peaks. The main contribution is attributed to Usually, the increase of hydroxyl content on the sur-
Ti–O in TiO2 . The other is the hydroxyl group. Table 4 face of TiO2 films is beneficial to the enhancement of
lists the results of curve fitting of XPS spectra for TiO2 photocatalytic activity [2,9].
 J.C. Yu et al. / Applied Catalysis B: Environmental 36 (2002) 31–43 41

Table 5
Photocatalytic activity of MTiO2 and TiO2 thin films deposited on quartz and glass before and after treated with H2 SO2 solutiona
Samples Thickness Mass of Degradation Rate const Specific photo- Specific photo-
(␮m) film (mg) rateb (%) K (min−1 ) activity 1c (mol/g h) activity 2d (mol/m2 h)
MTiO2 /quartz 0.21 10.3 12.2 2.3 × 10−3 4.1 × 10−3 3.0 × 10−3
MTiO2 /glass 0.20 9.8 2.6 4.4 × 10−4 1.0 × 10−3 7.1 × 10−4
TiO2 /quartz 0.18 8.8 4.8 7.6 × 10−4 1.9 × 10−3 1.2 × 10−3
TiO2 /glass 0.17 8.3 1.1 1.8 × 10−4 4.4 × 10−4 2.6 × 10−4
MTiO2 /quartz + H2 SO4 0.21 10.2 14.9 2.9 × 10−3 5.0 × 10−3 3.7 × 10−3
MTiO2 /glass + H2 SO4 0.20 9.3 10.5 2.1 × 10−3 3.9 × 10−3 2.6 × 10−3
TiO2 /quartz + H2 SO4 0.18 8.9 5.9 9.2 × 10−4 2.3 × 10−3 1.5 × 10−3
TiO2 /glass + H2 SO4 0.17 8.0 2.5 4.1 × 10−4 1.0 × 10−3 5.8 × 10−4
a The area of glass and quartz sheets covered by the MTiO2 and TiO2 layer is 140 cm2 . The coating cycle is one.
b Average degradation rate (c/co ) of acetone after 1 h of photocatalytic reaction.
c Acetone degradation amount per unit mass catalyst after 1 h of photocatalytic reaction.
d Acetone degradation amount per unit film surface after 1 h of photocatalytic reaction.

3.6. Photocatalytic activity photo-generated holes and thus prevent electron–hole

recombination [30]. All these factors can enhance the
3.6.1. Effects of preparation method on photocatalytic activity of MTiO2 thin films prepared
photocatalytic activity of thin films by the reverse micellar method.
It can be seen from Table 5 that photocatalytic
activity of MTiO2 thin films prepared by the reverse 3.6.2. Effects of surface acid treatment on
micellar method is obviously higher than that of TiO2 photocatalytic activity of thin films
thin films prepared by the sol–gel method whether the It can be seen from Table 5 that MTiO2 and TiO2
substrate is quartz or glass. The apparent rate constant, thin films deposited on glass show very low pho-
degradation rate and specific photocatalytic activity of tocatalytic activity, compared with these thin films
sample MTiO2 /quartz are greater than those of sample deposited on quartz. However, XPS results showed
TiO2 /quartz by up to more than two times. The same is that the films deposited on glass contain more surface
true for samples MTiO2 films and TiO2 films deposited hydroxyl groups than those coated on quartz. It is clear
on glass. Characterization results of BET, AFM, that the abundance of hydroxyl groups is not a deter-
UV–VIS spectra and XPS can explain the enhance- mining factor in the photocatalytic activity of these
ment in photocatalytic activity. Firstly, BET results thin films. The composition and properties of quartz
show that MTiO2 thin films possess much larger sur- and glass are different. Glass contains a large amount
face area, which results in a higher photodegradation of sodium ions and possesses lower thermal stability.
rate. Secondly, the monodispersity of TiO2 particles When TiO2 /glass and MTiO2 /glass are heat-treated, a
on the surface of MTiO2 films is probably beneficial lot of sodium ions diffuse from soda lime glass into the
to transfer and separation of photo-generated elec- nascent TiO2 films. It has been confirmed that sodium
trons and holes in the inner of and on the surface ions are very detrimental to the photocatalytic activity
of TiO2 particles and reduce the recombination of of the TiO2 films [2,5,6,8]. Paz et al. [7] thought that
photo-generated electrons and holes. Thirdly, UV–VIS high concentration sodium could prevent formation of
spectra show that the absorption edge of MTiO2 films the photoactive anatase phase, and low concentration
is at a shorter wavelength range than that of TiO2 sodium could produce surface and bulk recombina-
film. This indicates crystallites in MTiO2 films with tion centers of photo-generated electron–hole pairs. In
lightly higher band-gap energy and a stronger oxida- order to reduce the deleterious effect of sodium they
tion power [4]. Finally, XPS results show that MTiO2 introduced a thin sodium transport blocking layer at
thin films contain more hydroxyl groups. Of course, the interface of proton-exchanged (“hydrogen”) glass
the increase in surface hydroxyl content will trap more and nascent TiO2 layer. Sodium transport into nascent
42 J.C. Yu et al. / Applied Catalysis B: Environmental 36 (2002) 31–43

TiO2 layers was effectively retarded and the photo- times. This increase in activity can be correlated
catalytic activity of TiO2 films on soda lime glass was with the elimination of sodium ion impurity on the
improved. Other researchers pre-coated glass with a catalyst surface.
SiO2 layer in order to prevent the diffusion of sodium
ions and enhance photocatalytic activity [2,8,22].
This study used a very simple approach to remove Acknowledgements
sodium ions. When thin film samples were treated
with 1 M H2 SO4 aqueous solution, their photocat- The work described in this paper was partially sup-
alytic activity was significantly enhanced as shown in ported by a Grant from the National Natural Science
Table 5. The photocatalytic activity of MTiO2 /quartz Foundation of China and Research Grants Council of
and TiO2 /quartz increased up to about 20%, while the Hong Kong Special Administrative Region, China
the activity of MTiO2 /glass and TiO2 /glass increased (Project No. N CUHK433/00 and 4001161947). This
up to four and more than two times, respectively. The work was also financially supported by the Founda-
small enhancement in the quartz samples may be due tion for University Key Teachers of the Ministry of
to the increase in the surface hydroxyl content. The Education, the National Natural Science Foundation
much more significant activity enhancement in the of China (No. 29725715, 50072016 and 20077027),
glass samples cannot be accounted by the enrichment and the Foundation of Chinese Academy of Sciences.
of hydroxyl groups alone, it has to be due to the
elimination of sodium ions from the thin film sur-
face. It should be noted that thin films on quartz still References
outperform those coated on glass even after sulfuric
acid treatment. This can be explained by the XRD [1] K. Honda, A. Fujishima, Nature 238 (1972) 37.
and TEM results, which show that thin films on glass [2] M.R. Hoffmann, S.T. Martin, W. Choi, D.W. Bahnemann,
Chem. Rev. 95 (1995) 69.
exhibit poor crystallization and contain less anatase [3] A. Fujishima, T.N. Rao, D.A. Tryk, J. Photochem. Photobiol.
phase. C: Photochem. Rev. 1 (2000) 1.
[4] A.L. Linsebigler, G. Lu, J.T. Yates Jr., Chem. Rev. 95
(1995) 735.
4. Conclusions [5] A. Fujishima, R.T. Narasinga, Proc. Indian Acad. Sci. Chem.
Sci. 109 (1997) 471.
[6] Y. Paz, Z. Luo, L. Rabenberg, A. Heller, J. Mater. Res. 10
1. The photocatalytic activity of MTiO2 thin films is (1995) 2842.
significantly higher than that of TiO2 thin films. [7] Y. Paz, A. Heller, J. Mater. Res. 12 (1997) 2759.
The monodispersity of TiO2 particles in MTiO2 [8] J. Yu, X. Zhao, Mater. Res. Bull. 35 (2000) 1293.
thin films is helpful in the transfer and separa- [9] J. Yu, X. Zhao, Mater. Res. Bull. 36 (2001) 97.
tion of photo-generated electrons and holes on [10] M.E. Zorn, D.T. Tompkins, W.A. Zeltner, M.A. Anderson,
Appl. Catal. B: Environ. 23 (1999) 1.
TiO2 particles. Moreover, there are more hydroxyl [11] A. Fernandez, G. Lassaletta, V.M. Jimenez, A. Justo, A.R.
groups on the surface of MTiO2 thin films. The Gonzalez-Elipe, J.M. Herrmann, H. Tahiri, Y. Ait-Ichou, Appl.
abundance of hydroxyl groups and higher surface Catal. B: Environ. 7 (1995) 49.
areas are also beneficial to the enhancement of [12] J.M. Herrmann, H. Tahiri, Y. Ait-Ichou, G. Lassaletta, A.R.
photocatalytic activity. Gonzalez-Elipe, A. Fernandez, Appl. Catal. B: Environ. 13
(1997) 219.
2. The MTiO2 and TiO2 films deposited on quartz [13] T. Yoko, A. Yuasa, K. Kamiya, S. Sakka, J. Electrochem.
show higher photocatalytic activity than films on Soc. 138 (1991) 2279.
glass. This is because films on quartz exhibit better [14] N. Negishi, T. Iyoda, H. Hashimoto, A. Fujishima, Chem.
crystallization and are free of sodium contaminant. Lett. (1995) 841.
3. Sulfuric acid treatment can enhance photocatalytic [15] N. Negishi, K. Takeuchi, T. Ibusuki, J. Mater. Sci. 33 (1998)
5789.
activity of MTiO2 and TiO2 thin films deposited
[16] H. Ichinose, M. Terasaki, H. Katsuki, J. Ceram. Soc. Jpn.
on both glass and quartz. The photocatalytic ac- 104 (1996) 715.
tivity of MTiO2 /glass and TiO2 /glass after treated [17] H. Ichinose, H. Katsuki, J. Ceram. Soc. Jpn. 106 (1998) 344.
with H2 SO4 solution increases by more than two [18] Y. Tanaka, J. Ceram. Soc. Jpn. 107 (1999) 47.
 J.C. Yu et al. / Applied Catalysis B: Environmental 36 (2002) 31–43 43

[19] H. Kominami, H. Kumamoto, Y. Kera, B. Ohtani, Appl. Catal. [26] K. Kato, A. Tsuzuki, H. Taoda, Y. Torii, T. Kato, Y. Butsugan,
B: Environ. 30 (2001) 329. J. Mater. Sci. 29 (1994) 5911.
[20] E. Stathatos, P. Lianos, F. DelMonte, D. Levy, D. Tsiourvas, [27] M. Anpo, T. Shima, S. Kodama, Y. Kubokawa, J. Phys. Chem.
Langmuir 13 (1997) 4295. 91 (1987) 4305.
[21] E. Stathatos, P. Lianos, Langmuir 16 (2000) 2398–2400. [28] D. Duonghong, E. Borgarello, M. Gratzel, J. Am. Chem. Soc.
[22] A. Hattori, K. Shimoda, H. Tada, S. Ito, Langmuir 15 (1999) 103 (1981) 4685.
5422. [29] C. Wagner, G. Muilnberg, Handbook of X-ray Photoelectron
[23] J. Lin, J.C. Yu, D. Lo, S.K. Lam, J. Catal. 183 (1999) 368. Spectroscopy, Physical Electronics Division, Perkin-Elmer
[24] T. Yoko, K. Kamiya, S. Sakka, Yogyo Kyokai Shi 95 Corporation, Eden Prairie, MN, 1979.
(1987) 12. [30] D.F. Ollis, E. Pelizzetti, N. Serpone, Environ. Sci. Technol.
[25] J. Yu, X. Zhao, Q. Zhao, J. Mater. Sci. Lett. 19 (2000) 1015. 25 (9) (1991) 1523.