The nature of multilayered TiO2-based photocatalytic films prepared by a sol-gel process

Article abstract of DOI:10.1016/S0025-5408(98)00019-1

Transparent anatase TiO₂-based multilayered photocatalytic films on porous alumina and glass substrates were synthesized via a sol-gel process. All films had a sponge-like microstructure and a mean crystallite dimension of ca. 8 nm. The simple anaerobic decomposition of acetic acid reaction was used as the standard test system for evaluating the photocatalytic activity of the films. Doping with iron(III) impeded the photocatalytic activity. 1998 Elsevier Science Ltd.

Full text of DOI:10.1016/S0025-5408(98)00019-1

Materials Research Bulletin, Vol. 33, No. 5, pp. 795–801, 1998
Copyright © 1998 Elsevier Science Ltd
Printed in the USA. All rights reserved
0
025-5408/98 $19.00 ϩ .00
PII S0025-5408(98)00019-1
THE NATURE OF MULTILAYERED TiO -BASED PHOTOCATALYTIC FILMS
2
PREPARED BY A SOL-GEL PROCESS
P. Sawunyama, A. Yasumori*, and K. Okada
Department of Inorganic Materials, Tokyo Institute of Technology, 2–12-1 O-okayama,
Meguro-ku, Tokyo 152, Japan
(Refereed)
(Received August 18, 1997; Accepted September 2, 1997)
ABSTRACT
Transparent anatase TiO -based multilayered photocatalytic films on po-
2
rous alumina and glass substrates were synthesized via a sol-gel process.
All films had a sponge-like microstructure and a mean crystallite dimen-
sion of ca. 8 nm. The simple anaerobic decomposition of acetic acid
reaction was used as the standard test system for evaluating the photocat-
alytic activity of the films. Doping with iron(III) impeded the photocata-
lytic activity. © 1998 Elsevier Science Ltd
KEYWORDS: A. thin films, B. sol-gel chemistry, C. electron microscopy, D.
microstructure, D. catalytic properties
INTRODUCTION
TiO -based photocatalytic technology is making inroads in such diverse applications as
2
environmental remediation, organic compound transformation [1–4], and self-cleaning ma-
terials, e.g., TiO -coated windows and bathroom tiles [5,6]. Except for the latter, TiO is
2
2
normally used in the form of particulate suspensions. In recent years, however, interest has
shifted toward nanometer-sized particles (crystallite dimension Յ 10 nm) in the form of
colloids or film networks on supports [7].
An attractive property of nanometer-sized photocatalytic films or systems is the ability to
alter the electrochemical potentials of the photogenerated charge carriers by simply modi-
fying the size of the particles. This desirable feature, coupled with negligible band bending,
*
To whom correspondence should be addressed.
7
95

7
96
P. SAWUNYAMA et al.
Vol. 33, No. 5
means that the major deactivation event in photocatalysis, that of electron-hole recombina-
tion, is kept at an absolute minimum. Moreover, film networks do not suffer from one of the
major drawbacks of particulate materials: that of recovery of the photocatalyst at the end of
the reaction.
Nonetheless, the photocatalytic activity of TiO films is often governed by factors such as
2
film preparation procedure, density or microstructure of films, crystalline phase, and the
presence of impurities or dopants, as well as the nature of the support. Although numerous
reports [7–10] have been made on the preparation and photocatalytic activity of TiO films
2
on glass, research into TiO films on other supports, for example, porous ceramic materials,
2
has been largely lacking. Porous supports offer, among other features, increased reaction
surfaces, the ability to support a multifunctional catalytic system comprising, for example, a
dispersed metal and metal oxide and possible sieve properties for selective or controlled
synthetic work. In this paper, we report on the nature and photocatalytic activities of
multilayered TiO -based films supported on porous alumina substrates. The photocatalytic
2
activities of the films were evaluated using the anaerobic decarboxylation of acetic acid, also
known as the photo-Kolbe reaction [11] as the test system:
TiO2h␯
Ϫ1
CH COOH O3 CO ϩ CH ⌬G ϭ 52.3 kJ mol
(1)
3
2
4
EXPERIMENTAL
Materials. TiO sols were prepared via a sol-gel process [12]. Briefly, this involved the
2
Ϫ3
Ϫ3
acetic acid (0.033 mol dm ) catalyzed hydrolysis of acetylacetone (0.028 mol dm
modified titanium(IV) tetraisopropoxide (0.028 mol dm ) in isopropanol (140 mL) at 80°C
under an inert atmosphere. The resulting monodispersed sol was stable for several months.
)
Ϫ3
Similarly, iron(III)-doped TiO sols were prepared by adding the requisite amount of dopant,
2
iron(III) acetylacetonate (concentration; 0.5, 1.0, 1.5, and 2.0% atomic weight of TiO2,
respectively), to the precursor solutions. Films of uniform thickness were deposited on clean
glass slides (20 mm ϫ 20 mm) and porous alumina membranes (diameter: 25 mm; Whatman
filter discs, pore size 0.02 mm), respectively, by spin coating. A standard multilayered film
on glass was fabricated by repeating the coating cycle eight times, i.e., spin coating, drying
at room temperature, and drying at 300°C for 10 min. For films on alumina substrates, a
porous SiO film (from ethanolic SiO sol) was deposited by spin coating. This was followed
2
2
by the sputter coating of platinized photocatalysts (Pt) and/or the deposition of the desired
number of layers of TiO by spin coating. The resulting multilayered films were calcined at
2
4
00°C for 4 h. Figure 1 is an illustration of a typical platinized TiO multilayered film on an
2
alumina substrate.
All other materials (Wako Pure Chemicals) used were of the highest grade available.
Doubly distilled deionized water was used throughout.
Characterization. UV-vis spectra of films supported on glass were recorded with a double
beam spectrophotometer (Jasco UVIDEC-610). Film thickness was estimated by profilom-
etry (Surfcom-Tokyo Seimitsu). The texture of films on porous alumina was observed by
field emission scanning electron microscope [(FE-SEM), Jeol JSM-890S]. X-ray diffraction
(XRD) patterns were recorded using monochromated Cu K␣ radiation with a Rigaku
Geigerflex diffractometer.

Vol. 33, No. 5
TITANIA FILMS
797
FIG. 1
An illustration of a typical platinized multilayered photocatalyst.
Photocatalytic Activity Evaluation. A 300 W Xe arc lamp, operated at maximum power,
was used as the light source. A thermostated Pyrex vessel (volume: 1000 mL) with a
provision for circulating the head space gas was used as the photochemical reaction vessel.
2
In a typical experiment, the photocatalyst, supported on a sample holder (area: 1 cm ), was
placed in the vessel; the coated surface faced the irradiation source, and 700 mL of a 1429
ppm (ca., 0.14%) acetic acid solution was added. The solution was purged with He for ca.,
3
0 min before irradiation. The gas phase was circulated continuously, and gas phase samples
were taken at appropriate intervals throughout and analyzed for carbon dioxide by gas
chromatography (Shimadzu GC-6AM with a gaskuropak 54 column, length, 2.0 m; o.d., 5
mm; i.d., 3 mm; and a thermal conductivity detector interfaced to an integrator, Shimadzu
Chromatopac CR6A). Helium was used as the carrier gas.
RESULTS
XRD. XRD patterns of TiO -based films supported on glass and porous alumina substrates
2
were recorded in the range 2␪ ϭ 70–10°. All XRD patterns were identical for the same
number of layers regardless of the nature of the substrate. The XRD data revealed that the
multilayered films were composed of highly oriented anatase TiO crystallites, i.e., oriented
2
along the (101) crystallographic plane. Peak intensity of the (101) plane increased with an
increasing number of coating applications. No crystallographic planes of iron oxides were
3
ϩ
3ϩ
identified for Fe -doped films. This could be due to the low concentrations of the Fe
3
ϩ
dopant and also to the possible incorporation of Fe within the anatase lattice. The mean
crystallite dimension of the films perpendicular to the (101) plane was estimated using the
Scherrer equation [13]. The mean crystallite dimension was ca. 8 nm.
UV-vis and Film Texture. UV-vis absorption spectra of films on glass are shown in Figure
2
. All films were transparent and exhibited interference patterns. An increase in the number

7
98
P. SAWUNYAMA et al.
Vol. 33, No. 5
FIG. 2
UV-vis absorption spectra of TiO films on glass. The number of layers are (from left to
2
right) 4, 6, 8, 12, 16, and 20.
3
ϩ
of layers and doping with Fe resulted in a red-shift of absorption edges of the films.
3
ϩ
However, Fe absorption peaks were not evident. It is possible that doping of TiO with
2
3
ϩ
3ϩ
Fe resulted in the incorporation of Fe into the TiO lattice as suggested above.
2
It was possible to estimate film thickness using the UV-vis transmission spectra data by
invoking the following relationship [14]:
d ϭ m␭ ␭ /(2[n(␭ ␭ Ϫ n(␭ )␭ ])
(2)
1
2
1
2
2
1
where d is film thickness (in nm), m is the number of oscillations between the two minima
or maxima occurring at wavelengths ␭ and ␭ (in nm), and n(␭ ) and n(␭ ) are the refractive
1
2
1
2
indices of the film at ␭ and ␭ , respectively. The refractive index of anatase in the visible
1
2
region is ca. 2.49–2.55 [15]. Calculated film depths of the multilayered films are presented
in Table 1.
Analyses of film surfaces by profilometry revealed that the surfaces of all films were very
jagged. Due to the low resolution of this technique, however, detailed analyses of the surfaces
were not performed. Nevertheless, film thickness of films deposited on glass substrates was
estimated using this technique, and the results are shown in Table 1. From these results, it was
estimated that each coating application deposited a TiO layer approximately 30-nm thick.
2
This result was confirmed by FE-SEM (vide infra).
FE-SEM images of a typical platinized multilayered film (i.e., 20 TiO coatings, Pt film,
2
and SiO film) are shown in Figure 3. Figure 3A reveals that the surface morphology of the
2
films consists of a granular and rough microtexture, more or less sponge-like. Figure 3B
TABLE 1
Film Thickness As Determined By Profilometry
and UV-Visible Spectra Data
Number of layers
4
8
12
16
20
Thickness d/nm (UV-vis)
Thickness d/nm (profilometry)
158
150
302
250
416
-
464
400
596
600

Vol. 33, No. 5
TITANIA FILMS
799
FIG. 3
FE-SEM images of a multilayered photocatalyst on porous alumina substrate after heat
treatment at 400°C for 4 h. (A) surface morphology, (B) fractured surface, (C) a higher
magnification of the image shown in B, and (D) a backscattered electron image of panel A
showing the Pt layer (white).
shows the cross section of the fractured multilayered film. The thickness was determined to
be ca. 600 nm. Thus, each TiO coating application deposited a film ca. 30-nm thick. This
2
result is in close agreement with those obtained by profilometry and from UV-vis spectra
(vide supra). A higher magnification of the edge profile in Figure 3B is depicted in Figure 3C.
The micropores are clearly visible. A backscattered electron image of the fractured cross
section showing the Pt film (white) beneath the TiO layers is presented in Figure 3D. The
2
thickness of the Pt film was estimated to be ca. 20 nm. Thus the TiO film network was
2
composed of agglomerates of smaller crystallites with channels ranging in size from a few
nanometers to perhaps a few tens of nanometers, giving rise to a fairly porous photocatalyst
film.
Photocatalytic Activity. Irradiation of an aqueous acetic acid solution under anaerobic
conditions in the presence of a TiO -based photocatalytic film resulted in the evolution of
2
CO and CH as the main gaseous products. The activities of the different TiO -based
2
4
2
photocatalysts on porous alumina substrates were evaluated by monitoring the evolution of
CO quantitatively. In a typical experiment, the desired photocatalyst (8 layers of TiO ) was
2
2
Ϫ3
immersed in a 0.14% (0.02 mol dm ) aqueous acetic acid solution, and irradiation was
carried out as outlined in the “Experimental” section. The temperature of the reaction

8
00
P. SAWUNYAMA et al.
Vol. 33, No. 5
TABLE 2
Amount of CO Evolved in the Photodecomposition
2
of Acetic Acid Mediated By a TiO -Based Photocatalytic Film
2
on Porous Alumina
Rate
(␮mol cm
Amount of CO evolved in
2
Ϫ2 Ϫ1
h
Ϫ2
Photocatalyst
TiO2
)
5 h (␮mol cm
)
3.0
1.2
2.7
2.2
4.0
1.5
1.8
1.6
15.0
6.2
13.7
11.2
20.0
7.7
Fe/TiO2
Fe/TiO (1.0%)
2
Fe/TiO (2.0%)
2
Pt/TiO2
Pt/Fe/TiO (0.5%)
2
Pt/Fe/TiO2
9.1
7.8
Pt/Fe/TiO (2.0%)
2
Experimental conditions: 0.14% acetic acid solution, initial pH ϭ ca.3.19,
Ϫ4
2
surface area of photocatalyst film ϭ 1.0 ϫ 10 m , temperature ϭ 40°C.
solution was maintained at 40°C, and irradiation was carried out for 5 h. Table 2 summarizes
the results of this investigation.
DISCUSSION
Films prepared in this work were transparent and had a porous microstructure; however, the
exact dimensions of the film pores were uncertain. The rough surface texture and the porous
microstructure of the film networks were formed from the agglomeration of elementary units
during calcination. The colloidal particles of the TiO sol had an estimated average particle
2
size of ca. 5 nm. Thus, the porosity of the films was largely controlled by the elementary
particle sizes, although sintering during calcination could have resulted in the production of
larger particles. Although a highly dense film is desirable for efficient charge diffusion, an
open network is crucial for a good photocatalyst.
All photocatalytic films prepared in this work were of the anatase phase and were oriented
along the (101) plane. In contrast, Weinberger and Garber [9] have reported the synthesis of
porous anatase films oriented along the (220) crystallographic plane by a reactive magnetron
sputtering technique. It is not clear how preferred crystallite orientation affects photocatalytic
activity, and thus, this subject warrants further investigation.
The illumination of a photocatalyst with light energy, h␯ Ն E , results in the generation
bg
of electrons and holes. The photogenerated charge carriers rapidly diffuse to the surface of
the photocatalyst, and, in the case of platinized photocatalysts, the driving force for charge
carrier separation is believed to be the potential gradient between the Pt electrode and the
solution. Therefore, the separation of reductive sites from oxidative sites was achieved
through a judicious arrangement of the multilayered photocatalytic film. The separation of
reaction centers with a concomitant minimization of charge carrier recombination would thus
improve the efficiency of the overall photocatalytic process.
In this study, the platinized-undoped photocatalyst was the most active, and the platinized-
3
ϩ
ϩ
doped films were the least active. For Fe -doped films, although the nature of the Fe3

Vol. 33, No. 5
TITANIA FILMS
801
3
ϩ
centers in anatase is unknown, work by Gratzel and Howe [16] suggests that the Fe ions
4
ϩ
are substitutionally located in Ti lattice sites. Of course, the substitutional procedure is
3
ϩ
expected to influence the placement of the Fe ions within the TiO lattice. It is possible that
2
3
ϩ
3ϩ
the low reactivity of Fe -doped films may have been due to the fact that Fe ions acted as
charge carrier recombination sites.
Nevertheless, the photocatalytic materials reported in this work hold promise for such
applications as environmental remediation, self-cleaning materials, and controlled synthesis.
Further work into other TiO -mediated photocatalytic reactions is necessary before some of
2
the properties of the TiO -based multilayered photocatalysts reported in this study can be
2
conclusively demonstrated.
CONCLUSIONS
Porous films synthesized in this work had several unique features: (1) high surface area to
volume ratios, (2) high optical transparency, (3) a novel way of separating the reductive sites
from the oxidative sites, and (4) possible molecular sieve properties. The separation of
reductive sites from oxidative sites was achieved through a prudent arrangement of TiO
2
3
ϩ
layers, Pt electrode, and the porous silica–alumina base. Doping with Fe
detrimental to the overall rate of acetic acid decomposition.
ions was
ACKNOWLEDGMENTS
P.S. thanks UNESCO/MONBUSHO for the award of a research fellowship.
REFERENCES
1
.
D.F. Ollis, E. Pelizzetti, and N. Serpone, in Photocatalysis: Fundamentals and Applications, ed.
N. Serpone and E. Pelizzetti, Chap. 18, Wiley-Interscience, New York (1989).
M.A. Fox and M.T. Dulay, Chem. Rev. 93, 341, (1993).
A. Mills, R.H. Davies, and D. Worsley, Chem. Soc. Rev. 22, 417 (1993).
M.R. Hoffmann, S.T. Martin, W. Choi, and D.W. Bahnemann, Chem. Rev. 95, 69 (1995).
Y. Paz, Z. Luo, L. Rabernberg, and A. Heller, J. Mater. Res. 10, 2842, (1995).
A. Fujishima, Chem. J. Chinese Univ. 16, 7 (1995).
A. Hagfeldt and M. Gratzel, Chem. Rev. 95, 49 (1995), and references therein.
F. Cao, G. Oskam, P.C. Searson, J.M. Stipkala, T.A. Heimer, F. Farzad, and G.J. Meyer, J. Phys.
Chem. 99, 11974 (1995).
2
3
4
5
6
7
8
.
.
.
.
.
.
.
9
.
B.R. Weinberger and R.B. Garber, Appl. Phys. Lett. 66, 2409 (1995).
1
1
1
1
1
0. N. Negishi, T. Iyoda, K. Hashimoto, and A. Fujishima, Chem. Lett. 841 (1995).
1. B. Kraeutler and A.J. Bard, J. Am. Chem. Soc. 100, 5985 (1978).
2. F. Babonneau, A. L e´ austic, and J. Livage, Mater. Res. Soc. Symp. Proc. 121, 317 (1988).
3. H.P. Klug and L.E. Alexander, X-ray Diffraction Procedures, Chap. 9, Wiley, New York (1954).
4. A.H. Clark, in Polycrystalline and Amorphous Thin Films and Devices, ed. L.L. Kazmerski,
Chap. 4, Academic Press, New York (1980).
1
1
5. N.N. Greenwood and A. Earnshaw, Chemistry of the Elements, Chap. 21, Pergamon, Oxford
(1984).
6. M. Gratzel and R.F. Howe, J. Phys. Chem. 94, 2566 (1990).