Electrochemical approach to evaluate the mechanism of photocatalytic water splitting on oxide photocatalysts

Article abstract of DOI:10.1016/j.jssc.2004.08.001

Model of an oxide photocatalyst electrode. Photoelectrochemical measurements of TiO₂, NaTaO₃, and Cr or Sb doped TiO ₂ and SrTiO₃ photocatalysts were carried out in H ₂ and O₂ saturated electrolytes in order to evaluate the reverse reactions during water photolysis. The poor activity of TiO₂ as a result of reverse photoreactions of O₂ reduction and H ₂ oxidation was revealed with the respective high cathodic and anodic photocurrents. The rise in the photocurrents at NaTaO₃ after La doping was in harmony with the doping-induced increase in the photocatalytic activity. NiO loading suppresses the O₂ photoreverse reactions, which declines photocatalytic activity, and/or promotes the photo-oxidation of water, because the O₂ photo-reduction current was scarcely observed near the flatband potential. Photocurrents of O₂ reduction and H ₂ oxidation were observed under visible light for the Cr and Sb doped SrTiO₃ and TiO₂, respectively. These phenomena are in harmony with the previous reports on the photocatalysts examined with sacrificial reagents.

Full text of DOI:10.1016/j.jssc.2004.08.001

ARTICLE IN PRESS
Journal of Solid State Chemistry 177 (2004) 4205–4212
www.elsevier.com/locate/jssc
Electrochemical approach to evaluate the mechanism of
photocatalytic water splitting on oxide photocatalysts
Ã
Yasumichi Matsumotoa, Ugur Unala, , Noriyuki Tanakaa, Akihiko Kudob, Hideki Katob
aFaculty of Engineering, Department of Applied Chemistry, Kumamoto University, Kurokami 2-39-1, Kumamoto-shi, Kumamoto-ken, 860-8555, Japan
bFaculty of Science, Science University of Tokyo, 1-3 Kagurazaka, Shinjuku-ku, Tokyo 162-8601, Japan
Received 12 May 2004; received in revised form 28 July 2004; accepted 2 August 2004
Abstract
Photoelectrochemical measurements of TiO2, NaTaO3, and Cr or Sb doped TiO2 and SrTiO3 photocatalysts were carried out in
H2 and O2 saturated electrolytes in order to evaluate the reverse reactions during water photolysis. The poor activity of TiO2 as a
result of reverse photoreactions of O2 reduction and H2 oxidation was revealed with the respective high cathodic and anodic
photocurrents. The rise in the photocurrents at NaTaO3 after La doping was in harmony with the doping-induced increase in the
photocatalytic activity. NiO loading suppresses the O2 photoreverse reactions, which declines photocatalytic activity, and/or
promotes the photo-oxidation of water, because the O2 photo-reduction current was scarcely observed near the flatband potential.
Photocurrents of O2 reduction and H2 oxidation were observed under visible light for the Cr and Sb doped SrTiO3 and TiO2,
respectively. These phenomena are in harmony with the previous reports on the photocatalysts examined with sacrificial reagents.
r 2004 Elsevier Inc. All rights reserved.
Keywords: Photochemistry; Water splitting; Reverse reactions; TiO2; Oxide photocatalyst; Water photolysis; Oxide films; Semiconductor/solution
interface; Doping
1. Introduction
The mechanism of the water photolysis has been
mainly discussed from the viewpoint of the band
structure and its energy positions. The electrochemical
measurements are very useful for the evaluation of the
photocatalytic mechanism, because the energy positions
of the band edges of the conduction and valence bands
are estimated from the results of the electrochemical
measurements. The evaluation of the reverse reaction
from hydrogen and oxygen to water under illumination
is also important to clear the mechanism of the water
photolysis, since the reverse reaction brings about the
lowconversion efficiency of the water photolysis on the
photocatalyst. However, the electrochemical investiga-
tion of the reverse reactions, i.e., electrochemical
hydrogen oxidation and oxygen reduction under illumi-
nation, has been scarcely made until now.
Many researches on water splitting into hydrogen and
oxygen molecules by semiconductor photocatalyst have
been carried out since Honda-Fujishima principle [1],
where the produced hydrogen is a prospective fuel
because of no production of CO2 after its oxidation. The
water photolysis is a conversion system from solar
energy to chemical fuel of hydrogen, only when
photocatalyst responds to visible light. Currently, the
high conversion efficiency has been discovered for some
photocatalysts, however, as a result of the UV light
response [2–7]. The photocatalysts that can use visible
light with high conversion efficiency must be discovered
and/or developed in the viewpoint of the energy
conversion using solar energy.
The electrochemical research is difficult, if the
photocatalyst is an insulator. Therefore, only semicon-
ductor samples with relatively high conductivity have
Ã
Corresponding author. Fax: +81-96-342-3679.
E-mail address: 016d9573@gsst.stud.kumamoto-u.ac.jp (U. Unal).
0022-4596/$ - see front matter r 2004 Elsevier Inc. All rights reserved.
doi:10.1016/j.jssc.2004.08.001

ARTICLE IN PRESS
Y. Matsumoto et al. / Journal of Solid State Chemistry 177 (2004) 4205–4212
4206
been used in the electrochemical measurements. How-
ever, the electrochemical tests can be performed for the
deposited lowconductive photocatalyst particles on the
metal substrate, when the very fine particles are in
contact with substrate, because the total resistivity of the
photocatalyst is not so large. In this principle, the
measurement of photoelectrochemical properties is
succeeded for some lowconductive layered oxides in
order to measure the electrochemical properties [8,9].
In this paper, the photoelectrochemical properties of
TiO2 with standard properties, NaTaO3 with high
quantum efficiency under UV light illumination [7,10],
and doped TiO2 and SrTiO3 with visible light response
in the water photolysis [11] are demonstrated. The
results give some important evaluation in the photo-
catalytic mechanism. The above principle in preparing
the electrode is applied, because the conductivities of
almost all the samples are very low. Consequently, an
electrophoretic deposition (EPD) technique is used to
prepare the photocatalyst electrodes.
107 O cm (TiO2: 107 O cm [12]). Therefore, the electro-
chemical properties can be measured for the present
photocatalysts, because the particle sizes are smaller
than about few mm and the measured photocurrent are
smaller than about 0.1 mA in general.
The morphology of the deposited films was observed
by SEM. All electrochemical experiments were carried
out in a conventional three-electrode electrochemical
quartz cell with a Pt counter electrode and a saturated
Ag/AgCl reference electrode. 0.1 M K2SO4 solution
(pH=6.4) was used as an electrolyte. The working
electrode potentials were referred to this reference
electrode unless otherwise stated in this paper. A
500 W ultra-high pressure Hg lamp was used as the
light source to measure the photoelectrochemical
properties. Cyclic voltammograms (CV) were measured
under potential sweep rate of 20 mV sÀ1. Electrolyte was
saturated with N2, O2, or H2 prior to the electrochemical
measurements. The apparent porous film thickness of all
the samples were about 3–8 mm, where the maximum
photocurrents were observed.
2. Experimental section
The photocatalyst powders used in this study were
TiO2 (P25, Degussa Co., Ltd., 30 nm average particle
size), TiO2 and SrTiO3 doped with either Cr or Sb,
NaTaO3, NaTaO3 doped with La(2%), and their NiO
loaded samples. The powders except P25, were prepared
by the same methods as described in previous papers
[7,10,11]. The powders were fixed on the Au or ITO
substrate by an electrophoretic deposition (EPD)
technique [8,9] in acetone/I2 solution. After the deposi-
tion, the deposited films were heated at 500 1C on certain
occasions in order to increase in adhesion between the
powder and the substrate. The sample powders used in
this study are listed in Table 1.
The model of the photocatalyst deposited electrode is
illustrated in Fig. 1. The ohmic drop due to the
photocatalyst is neglected for the deposited particle size
finer than about few mm. The voltage difference due to
the ohmic drop is only 10 mV under 0.1 mA for a
particle with few mm, when the specific resistivity is
Fig. 1. Model of the electrode deposited with the photocatalyst.
Table 1
Band gaps and particle sizes of the photocatalysts
Catalyst
Cocatalyst
Band gap/eV
Particle size/mm
TiO2(P25)
NaTaO3
None
None
3.0
4.0
0.03
2–4
NiO (0.05 wt%)
None
NaTaO3, La(2%)
4.1
0.1–0.7
NiO (0.05 wt%)
None
None
TiO2, Sb(3%)+Cr(2%)
SrTiO3, Sb(1.25%)+Cr(0.05%)
2.2
2.4
0.7–0.8
1–1.5

ARTICLE IN PRESS
Y. Matsumoto et al. / Journal of Solid State Chemistry 177 (2004) 4205–4212
4207
3. Results and discussion
of H2O oxidation and reduction, respectively, due to the
larger driving forces for the former reactions. However,
this conclusion may sometimes not fit for some electrode
surfaces, because the V–log i curve depends on the
electrocatalytic activity of the electrode surface. For
example, the electrochemical oxidation rate of H2 and
reduction rate of O2 will be lower than those of H2O, if
the electrocatalytic activities of an electrode surface are
relatively large for the latter reactions rather than the
former reactions. Moreover, Fig. 2 illustrates only the
simple case, where the site for the electron transfer for
H2O reduction to produce H2 is the same as that for the
hole transfer for H2 oxidation to produce H2O, and that
for the hole transfer for H2O oxidation to produce O2 is
the same as that for the electron transfer for O2
reduction to produce H2O. The real mechanism might
be that the paths for both sites mentioned above are
different from each other.
3.1. Model of the simple electron and hole transfers at the
semiconductor (photocatalyst)/solution interface for
water photolysis
Oxygen and hydrogen evolved at a photocatalyst
under illumination partially reacts to water (reverse
reaction) in the water photolysis process. Therefore,
photoelectrochemical reactions of oxygen and hydrogen
should be measured in order to evaluate the above
reverse reaction mechanism. Fig. 2 shows a basic model
of the photoelectrochemical process at the semiconduc-
tor/solution interface, where water, hydrogen, and
oxygen electrochemical reactions are illustrated as the
plots of V vs. log i (based on Tafel equation). RHE,
ROE, RC, and SS denote reversible hydrogen electrode,
reversible oxygen electrode, recombination center in
bulk (and/or in space charge layer), and surface state,
respectively. iH and iO denote exchange current densities
of hydrogen and oxygen electrochemical reactions
respectively, and the former is much larger than the
latter in general. Some excited electrons in the conduc-
tion band (CB) flowto the surface, and then react with
H2O and/or O2 at the CB band-edge and/or the SS
where some electrons recombine. The reacted electrons
bring about the cathodic photocurrents of O2 and H2O
at the CB band-edge (circle symbols) an/or via the SS
(triangle symbols), which correspond to the related log i.
The similar phenomena also occur for the produced hole
in the valence band (VB), where the occurrence of H2
and H2O oxidation brings about the related log i in this
figure (open symbols). In general, it is concluded that
the evolved H2 and O2 easily react with electron and
hole respectively rather than H2O, because the currents
of H2 oxidation and O2 reduction are larger than those
3.2. TiO2 electrodes
Fig. 3(A) shows the voltammograms of the TiO2/Ti
electrode, which was prepared by the heat-treatment of
Ti plate in air at 500 1C. The changing point of
photocurrent (from cathodic photocurrent to anodic
photocurrent or vice versa) in potential approaches to
the flat band potential (Vfb). The Vfb was about
–0.8ꢀÀ0.7 V, although it slightly depended on the
direction of the potential sweep and the saturated gas.
The Vfb measured in the present study was closely in
agreement with those reported already.
The large O2 reduction currents (cathodic photocur-
rent and dark current) were observed in the O2 saturated
electrolyte. The same result has been already reported by
Lindquist et al. [13]. They proposed a mechanism that
the cathodic photocurrent is based on the photoconduc-
tion of TiO2 together with a redox couple (O2/OÀ2 )
species existing on the surface, which collects the
electrons in the conduction band. Evidently, the
formation of cathodic photocurrent takes place under
the presence of oxidants. Surface O2 molecules con-
tribute to the photo-induced current by trapping the
electrons produced under illumination on the surface of
the electrode. In addition, the contribution of the photo-
induced current developed in lowconductive regions to
the observed photocurrent was also reported. Photo-
induced current was assigned to the increase in the
free charge carriers as a result of the excitation in
the sub-band gap states, and subsequent increase in
conductivity, which might, also, be the case for the
current study [13].
Under O2 saturation, a large anodic photocurrent in
the potential region from the Vfb to about 0 V was
observed in the sweep direction from negative to positive
potential, but not in the reverse direction from positive
to negative potential. The above anodic photocurrent is
Fig. 2. Model of the mechanism of electron and hole transfers in the
photocatalytic water splitting.

ARTICLE IN PRESS
Y. Matsumoto et al. / Journal of Solid State Chemistry 177 (2004) 4205–4212
4208
Fig. 3. Cyclic voltammograms of the TiO2/Ti electrode (A) and TiO2(P25)/Au electrode (B) in N2, O2, and H2 saturated electrolytes under UV
illumination. The horizontal arrows show the potential sweep directions.
based on the photo-oxidation of the adsorbed H2O2 (or
O2HÀ) produced by the cathodic photo-reduction of O2,
because the same anodic photocurrent was observed
after the addition of H2O2 into the electrolyte. The
disappearance of the anodic photocurrent in the reverse
sweep from positive to negative is based on the
consumption of the adsorbed H2O2 by the anodic
photo-oxidation in the sweep from negative to positive
potential. In the case of H2 saturated electrolyte, a large
anodic photocurrent was observed in the potential
region more positive than about 0 V compared with
the case of N2 saturation, but no increase in the anodic
photocurrent in the potential region from the Vfb to
about 0 V. Consequently, both large photocurrents in O2
and H2 saturated electrolytes compared with the case of
N2 saturation mean that the reduction of O2 and
oxidation of H2 proceed more preferentially than those
of H2O because of the large driving forces for the former
reactions regardless of the photoconduction and surface
mechanisms.
saturated electrolytes than that in N2 saturated one were
observed as similar to the TiO2/Ti electrode. The
changing point of photocurrent shifts to more positive
in O2 (about À0.1 V). The similar phenomenon has been
reported by Tsujiko et al. for nanometer size TiO2
particle electrode, and explained by the adsorption of O2
[14]. The present positive shift may be explained by the
same mechanism such as O2 and/or H2O2 adsorption.
Oxygen molecules adsorbed on the surface of the TiO2
film captures the surface conduction band electrons
before they reach to the substrate. The phenomenon
brings about decay in the anodic current as the cathodic
current develops and becomes dominant even at a
sufficiently more positive voltage. The shift to the
positive energy may also be explained by the alteration
in the energy levels on the surface as a result of charge
accumulation.
In conclusion, the poor photocatalytic activity of
TiO2 for the water photolysis can be explained by high
reverse reactions of the photo-reduction of the evolved
O2 and H2O2, and the photo-oxidation of the evolved
H2O2 and H2, even if H2 and O2 were evolved under UV
light initially. The observed anodic and cathodic
photocurrent, thus, can be assigned to these dominant
Fig. 3(B) shows the voltammograms of the
TiO2(P25)/Au electrode. This TiO2 electrode has the
following phenomenon different from the above TiO2/Ti
electrode, although higher photocurrents in O2 and H2

ARTICLE IN PRESS
Y. Matsumoto et al. / Journal of Solid State Chemistry 177 (2004) 4205–4212
4209
reverse reactions rather than the occurrence of the
photoconduction under illumination.
in the dark. Some H2O2 produced in this potential
region somewhat will affect the increase in the cathodic
photocurrent, because the relatively large photo-reduc-
tion current was observed in the H2O2 added electrolyte.
Simultaneously, a slight increase in the anodic photo-
current was also observed in the electrolyte with H2O2.
Therefore, sweeping the potential from positive to
negative direction is practically more preferable in order
to decrease the effect of the produced H2O2.
3.3. NaTaO3 electrodes with and without NiO loading
Figs. 4(A) and (B) showthe voltammograms of the
NaTaO3 electrode and that loaded with NiO respec-
tively. Photocurrents were observed for both electrodes
in the electrolytes saturated with N2, O2, and H2. The
Vfb (which sets around the changing potential from
anodic photocurrent to cathodic photocurrent) of the
NaTaO3 was about 0 V, while that of NaTaO3 loaded
with NiO was about 0.4 V. Thus, the Vfb shifted to
positive by the NiO loading. It is estimated from Ref.
[15] that the CB will be positioned at potentials more
negative than RHE. Therefore, the energy difference
between the Fermi level and the CB must be larger than
about 1.0 eV for all the present NaTiO3 samples. In the
O2 saturated electrolyte, cathodic photocurrent larger
than that in the N2 saturated electrolyte was observed
for both electrodes, while the anodic photocurrent in the
H2 saturated electrolyte was almost same as that in the
N2 saturated electrolyte. Figs. 5(A) and (B) showthe
voltammograms of La doped NaTaO3 (2% La) and that
loaded with NiO electrodes, respectively. The photo-
currents larger than those in the non-doped NaTaO3
(Fig. 4) were observed for the La doped samples,
although the same phenomena with the cases of the non-
doped NaTaO3, i.e., positive shift of the Vfb by about
0.4 V by NiO loading, and the increase of the cathodic
photocurrent in the O2 saturated electrolyte were
observed. Especially, the pronounced increase in the
cathodic photocurrent in the O2 saturated electrolyte
occurred in the potential region where the oxygen
reduction current at the Au substrate electrode appeared
The comparisons between the photocurrents of O2,
N2, and H2 near the Vfb are very important to evaluate
the mechanism of the photocatalyst in water photolysis,
because the reaction at the photocatalyst proceeds at a
potential near the Vfb (In fact, the rest potentials of all
the electrodes under the illumination were near the Vfb).
The photocurrent ratios of IO/INC and IH/INA are listed
in Table 2, where IO, INC, INA, and IH denote the
cathodic photocurrent in the O2 saturated electrolyte
that in the N2 saturated electrolyte, the anodic photo-
current in the N2 saturated electrolyte, and the anodic
photocurrent in the H2 saturated electrolyte, respec-
tively. Each photocurrent was calculated from the
voltammograms of the positive to negative potential
scan in order to exclude the H2O2 effect as stated above.
All ratios were approximated to the closest value,
because of the errors in the calculated photocurrent
values. The ratios of IH/INA were lower than about 2–3
at the potential region more positive than the Vfb,
indicating that photo-oxidation of H2 scarcely occurs
for all the NaTaO3 samples. However, the ratios of IO/
INC were very large (4about 3–5) for the samples
without NiO loading, but not large (o about 2–3) for
the NiO loaded samples at potentials near the Vfb. This
is very important to evaluate the NiO loading effect on
the photocatalyst. That is, the NiO loaded into the
Fig. 4. Cyclic voltammograms of the NaTaO3/Au electrode (A) and NaTaO3:NiO/Au electrode (B) in N2, O2, and H2 saturated electrolytes under
UV illumination.

ARTICLE IN PRESS
Y. Matsumoto et al. / Journal of Solid State Chemistry 177 (2004) 4205–4212
4210
Fig. 5. Cyclic voltammograms of the NaTaO3, La/Au electrode (A) and NaTaO3, La:NiO/Au electrode (B) in N2, O2, and H2 saturated electrolytes
under UV illumination.
Table 2
Photocurrent ratios in the N2, O2, and H2 saturated electrolytes for NaTaO3 samples
Samples/substrate
NaTaO3/Au
VCP
Potential/V vs. Ag/AgCl
À0.4
À0.3
À0.2
À0.1
0
0.1
0.2
0.3
—
0.4
—
0.5
—
0.6
—
0.7
0
IO/INC 4 50
50–60
—
30–50 30–50
—
—
—
—
—
0
—
1.2
—
—
1.2
—
—
IH/INA
—
—
4–5
—
3–4
0.1–0.2 0.3–0.5 0.6–0.8 1.1
NaTaO3/ITO
NaTaO3:NiO/Au
0.1
IO/INC 14–15 5–6
IO/INC 8–10 12–13
IH/INA
0.2–0.3 IO/INC 3–4
0–0.3 IO/INC 80–90 50–60
IH/INA
À0.1–0 IO/INC 10–11 10–15
0.4 IO/INC 12–13 12–13
3–4
1–2
—
—
— —
1–2 1–2
—
—
—
—
—
2–3
—
—
—
—
—
—
—
—
1
0.4
15–17 7–8
—
1–2
—
—
—
—
—
—
—
—
—
—
0–1
—
NaTaO3:NiO/ITO
NaTaO3, La/Au
1.9–2.0 ꢀ1.5
ꢀ1
ꢀ0.5 0–0.2
—
—
60–70 30–50
—
—
—
—
—
—
—
20–30
—
—
2–3
—
2–3 2–3
— —
2–3 2–3
2–3
—
2–3 3–4
NaTaO3,La/ITO
NaTaO3,La:NiO/Au
—
—
—
—
—
10–12 3–4
2–3
—
2–3
—
—
—
IH/INA —
NaTaO3,La:NiO/ITO 0.2–0.3 IO/INC 1–2
—
1–2
—
1–2
—
1–2
—
—
—
—
1–2 1–2
—
1–2
1–2
—
—
Note. VCP: the potential where photocurrent changes near Vfb.
samples suppresses the O2 photo-reduction and/or
promotes the H2O photo-reduction.
to a high quantum efficiency for the water photolysis at
the La doped NaTaO3 loaded with NiO. The effects of
the NiO loading were disappeared by heat-treatment at
500 1C in air, suggesting that the oxidized NiO will not
sufficiently act as a promoter in the water photolysis.
In conclusion, O2 photo-reduction rather than H2
photo-oxidation as the reverse reaction in the photolysis
of water occurs on the NaTaO3, leading to the decrease
in the quantum efficiency. La doping will bring about
the suppression of the recombination in the bulk and/or
at surface, and increase the quantum efficiency. This will
be estimated from the increase in the photocurrents by
the La doping. Moreover, the O2 photo-reduction was
relatively suppressed by the NiO loading. This will lead
3.4. Cr, Sb doped TiO2 and SrTiO3 electrodes
Figs. 6(A) and (B) showthe voltammograms of the Cr
and Sb doped TiO2 and SrTiO3 samples measured under
illumination of the visible light (4420 nm), respectively.

ARTICLE IN PRESS
Y. Matsumoto et al. / Journal of Solid State Chemistry 177 (2004) 4205–4212
4211
Fig. 6. Cyclic voltammograms of the TiO2, Sb+Cr/Au electrode (A) and SrTiO3, Sb+Cr/Au electrode (B) in N2, O2, and H2 saturated electrolytes
under visible light illumination. Insets showthe magnified region of interest.
Table 3
Photocurrent ratios in the N2, O2, and H2 saturated electrolytes under illumination of visible light for TiO2,Sb+Cr/Au and SrTiO3,Sb+Cr/Au
samples
Samples/Substrate
VCP
Potential/V vs. Ag/AgCl
À0.4
À0.3
À0.2
À0.1
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
SrTiO3,Sb+Cr/Au
TiO2,Sb+Cr/Au
0.2–0.3
0.5
IO/INC
IH/INA
IO/INC
IH/INA
25
—
—
—
13.3
—
50
—
—
—
—
—
—
—
1
—
0
—
—
—
—
1–2
—
1–2
—
1–2
—
1–2
—
5
—
—
—
2–3
—
0.5–1
—
0.5–1
—
1
—
1
—
—
—
—
10–20
10–20
10–20
Note. VCP: the potential where photocurrent changes near Vfb.
These photocatalysts showed visible light response in
photocurrent. These phenomena are in harmony with
the photocatalytic activity under visible light for these
samples [11]. It was reported that transition metal ions
codoped with Sb have a significant contribution to the
adsorption bands observed in visible light region for
TiO2 and SrTiO3. Photocatalytic activities by band gap
activation was enhanced by Sb–Cr doping. In these
electrodes, Cr3+ created electron donor centers in TiO2
but discrete levels in SrTiO3. In order to prevent the
formation of recombination centers, Sb5+ was codoped
as a charge balance agent [11]. Table 3 lists the ratios of
IO/INC and IH/INA, where IO, INC, INA, and IH denote
the cathodic photocurrent in the O2 saturated electro-
lyte, that in the N2 saturated electrolyte, the anodic
photocurrent in the N2 saturated electrolyte, and that in
the H2 saturated electrolyte under the visible light,
respectively. In the case of the TiO2 sample, the ratios of
IO/INC were lower than about 1, indicating that the O2
photo-reduction scarcely occurs. On the other hand, the
ratios of IH/INA were higher than about 10, indicating
that the H2 photo-oxidation occurs more easily than
H2O. Consequently, water photolysis under visible light
scarcely occurs by the H2 photo-oxidation. At this
sample, O2 evolution has been observed by the Ag+
addition, but not H2 evolution by the methanol addition
[11]. These phenomena are explained by the H2 photo-
oxidation, because the evolved H2 is immediately photo-
oxidized at the catalyst surface.
In the case of the SrTiO3 sample, the ratios of IO/INC
were larger than about 10 while those of IH/INA were
smaller than about 2–3. These results indicate that the
O2 photo-reduction preferentially occurs but not the H2
photo-oxidation under visible light. Consequently,
water photolysis does not occur by the preferential
photo-reduction of O2 under illumination of the visible
light. H2 evolution after the addition of methanol and
very lowamount of O evolved after the addition of
2

ARTICLE IN PRESS
Y. Matsumoto et al. / Journal of Solid State Chemistry 177 (2004) 4205–4212
4212
Ag+ can be explained by the preferential photo-
reduction of the evolved O2 [11].
in harmony with the results of the photocatalysts in the
presence of the sacrificial reagents reported already [11].
4. Conclusions
Acknowledgments
Photoelectrochemical measurements of the photoca-
talysts such as TiO2, NaTaO3, TiO2 and SrTiO3 doped
with Cr and Sb, were carried out in the N2, H2 or O2
saturated electrolytes in order to evaluate the reverse
reactions. Photocatalyst powder deposited electrode
were prepared by the EPD technique at Au or ITO
substrates, where the ohmic drops due to the catalyst
themselves will be negligible. It is concluded that O2
photoreduction and H2 photooxidation preferentially
occurs in respectively O2 and H2 saturated electrolytes
with larger cathodic and anodic photocurrents than
those saturated with N2.
The poor activity of TiO2 is explained by both reverse
photoreactions, because high anodic photocurrent of H2
and high cathodic photocurrent of O2 were observed.
The increase in the photocurrents at NaTaO3 as a result
of La doping was in harmony with the doping-induced
increase in the photocatalytic activity as a result of the
La doping. Relatively high O2 photo-reduction current
was observed, but not for the H2 photo-oxidation at the
NaTaO3 without NiO. In other words, the water
photolysis on these catalysts is suppressed by the O2
photo-reduction. On the other hand, the NiO loading
brought about the suppression of the O2 photo-
reduction and/or the increase in the photo-oxidation
of water, because the O2 photo-reduction current was
scarcely observed for the NiO loaded NaTaO3 samples.
Consequently, high photocatalytic activity of the NiO
loaded NaTaO3 doped with La [7] will be brought about
by the negligible reverse reactions of H2 and O2. The O2
photo-reduction and H2 photo-oxidation currents were
observed under visible light for SrTiO3 and TiO2 doped
with Cr and Sb, respectively. These phenomena are also
This work was supported by the Core Research for
Evolutional Science and Technology (CREST) program
of the Japan Science and Technology Co. (JST), and a
Grant-in-Aid for Scientific Research (No 15350123)
from the Ministry of Education, Culture, Sports,
Science, and Technology.
References
[1] A. Fujishima, K. Honda, Nature 238 (1972) 37.
[2] A. Kudo, K. Sayama, A. Tanaka, K. Asakura, K. Domen, K.
Maruya, T. Onishi, J. Catal. 120 (1989) 337.
[3] T. Tanaka, K. Shinohaya, A. Tanaka, M. Hara, J.N. Kondo, K.
Domen, J. Photochem. Photobiol. A 106 (1997) 45.
[4] Y. Inoue, M. Kohno, T. Kaneko, S. Ogura, K. Sato, J. Chem.
Soc. Faraday Trans. 94 (1998) 89.
[5] T. Ishihara, H. Nishiguchi, K. Fukamachi, Y. Takita, J. Phys.
Chem. B 103 (1999) 1.
[6] M. Machida, J. Yabunaka, T. Kijima, Chem. Mater. 12 (2000)
812.
[7] H. Kato, K. Asakusa, A. Kudo, J. Am. Chem. Soc. 125 (2003)
3082.
[8] Y. Matsumoto, A. Funatsu, D. Matsuo, U. Unal, J. Phys. Chem.
B 105 (2001) 10893.
[9] U. Unal, Y. Matsumoto, N. Tanaka, Y. Kimura, N. Tamoto,
J. Phys. Chem. B 107 (2003) 12680.
[10] H. Kato, A. Kudo, J. Phys. Chem. B 105 (2001) 4285.
[11] H. Kato, A. Kudo, J. Phys. Chem. B 106 (2002) 5029.
[12] R.N. Viswanath, A. Chandra Bose, S. Ramasamy, J. Phys. Chem.
Solids 62 (2001) 1991.
[13] S.E. Lindquist, H. Vidarsson, J. Molecul. Catal. 38 (1996)
131–155.
[14] A. Tsujiko, H. Itoh, T. Kisumi, A. Shiga, K. Murakoshi, Y.
Nakato, J. Phys. Chem. B 106 (2002) 5878.
[15] Y. Matsumoto, J. Solid State. Chem. 126 (1996) 227.