Role of iron ion electron mediator on photocatalytic Overall water splitting under visible light irradiation using Z-scheme systems

Article abstract of DOI:10.1246/bcsj.80.2457

A combination system consisting of a H₂ production photocatalyst, Pt/SrTiO₃:Rh, and an O₂ production photo-catalyst, BiVO₄ or WO₃, decomposed water into H ₂ and O₂ under visible light irradiation in the presence of an Fe³⁺/Fe²⁺ redox couple as an electron mediator. O₂ evolution on the BiVO₄ photocatalyst was inhibited by Fe²⁺ ions, because of the oxidation of Fe²⁺ instead of water. In contrast, H₂ evolution on the Pt/SrTiO₃:Rh photocatalyst was enhanced when Fe³⁺ ions co-existed. It is due to the back-reactions between H₂ and O₂ to form water, and the reduction of Fe³⁺ by H₂, which easily proceeded on the bare Pt cocatalyst surface, being efficiently suppressed by adsorption of [Fe(SO₄)(H₂O)₅]⁺ and/or [Fe(OH)(H₂O)₅]²⁺ on the Pt surface. Overall water splitting was achieved with the suppression of the back-reactions even using a Pt cocatalyst. Thus, it clears that iron ions contributed to the present Z-scheme systems not only as an electron mediator but also as an inhibitor of the back-reactions.

Full text of DOI:10.1246/bcsj.80.2457

ꢀ 2007 The Chemical Society of Japan
Bull. Chem. Soc. Jpn. Vol. 80, No. 12, 2457–2464 (2007) 2457
Role of Iron Ion Electron Mediator on Photocatalytic Overall Water
Splitting under Visible Light Irradiation Using Z-Scheme Systems
ꢀ1;2
Hideki Kato,1 Yasuyoshi Sasaki,1 Akihide Iwase,1 and Akihiko Kudo
1Department of Applied Chemistry, Faculty of Science, Tokyo University of Science,
1-3 Kagurazaka, Shinjuku-ku, Tokyo 162-8601
2Core Research for Evolutional Science and Technology, Japan Science Technology Agency (CREST/JST)
Received June 6, 2007; E-mail: a-kudo@rs.kagu.tus.ac.jp
A combination system consisting of a H2 production photocatalyst, Pt/SrTiO3:Rh, and an O2 production photo-
catalyst, BiVO4 or WO3, decomposed water into H2 and O2 under visible light irradiation in the presence of an Fe3þ
/
Fe2þ redox couple as an electron mediator. O2 evolution on the BiVO4 photocatalyst was inhibited by Fe2þ ions, because
of the oxidation of Fe2þ instead of water. In contrast, H2 evolution on the Pt/SrTiO3:Rh photocatalyst was enhanced
when Fe3þ ions co-existed. It is due to the back-reactions between H2 and O2 to form water, and the reduction of
Fe3þ by H2, which easily proceeded on the bare Pt cocatalyst surface, being efficiently suppressed by adsorption of
[Fe(SO4)(H2O)5]þ and/or [Fe(OH)(H2O)5]2þ on the Pt surface. Overall water splitting was achieved with the suppres-
sion of the back-reactions even using a Pt cocatalyst. Thus, it clears that iron ions contributed to the present Z-scheme
systems not only as an electron mediator but also as an inhibitor of the back-reactions.
Since the first report of photoelectrochemical splitting of
water using a semiconductor electrode,1 the water splitting
reaction using photoelectrodes2–10 and photocatalysts11–18 has
been extensively studied for a photon energy conversion.
The research has been especially focused on the visible light
response with the hopes of utilizing the solar energy. There
are some reports on visible-light-driven photoelectrodes.4–10
Among them, the GaInP2/GaAs and dye-sensitized TiO2/
WO3 tandem electrodes are able to split water without any ex-
ternal voltage.6,8 On the other hand, it has been reported that
some metal oxide powders with wide band gaps function as
active photocatalysts for water splitting.11–16 Moreover, it
has been recently reported that Ge3N4 is the first non-oxide
photocatalyst for overall water splitting.17 Unfortunately, such
photocatalysts cannot use sunlight efficiently, because their
band gaps are wider than 3 eV. Domen et al. have recently re-
ported overall water splitting under visible light irradiation
using a GaN–ZnO solid solution photocatalyst, and its appar-
ent quantum yield is 2.5% at 420–440 nm.18 On the other hand,
there are many reports on H2 production from water containing
an electron donor, such as alcohols and sulfide ions, and O2
production from water in the presence of an electron acceptor,
such as Fe3þ and Agþ ions, using metal oxides,14,19,20 (oxy)-
nitrides,21 and sulfides14,22–24 photocatalysts with visible light
responses. The reaction of H2 or O2 production from water
containing sacrificial reagents is a half reaction of overall
water splitting. Therefore, it can be considered that overall
water splitting is achieved by systems in which two photo-
catalysts are active for H2 and O2 evolution. The combined
systems involving two photoexcitation processes are similar
to the photosynthesis system in the green plants, and called
‘‘Z-scheme systems.’’ There are a few reports on water split-
ting reaction using the Z-scheme system with powdered photo-
catalysts. Fujihara et al. have reported overall water splitting
using a two-compartment system.25 In this system, the H2 evo-
lution compartment involving a Pt/TiO2 photocatalyst and
Brꢁ ions as electron donors is connected to the O2 evolution
compartment involving TiO2 and Fe3þ ions as electron accep-
tors by a Pt wire and an ion-exchange membrane. Sayama et
al. have reported the production of H2 and O2 from water using
the system combined the photochemical H2 production process
using Fe2þ under UV light irradiation (ꢀ < 280 nm) with the
photocatalytic O2 production system of RuO2/WO3 with Fe3þ
ions.26 One-compartment Z-scheme systems constructed of
two kinds of photocatalysts, Pt/SrTiO3:Cr,Ta or Pt/TaON
ꢁ
for H2 evolution and Pt/WO3 for O2 evolution, and an IO3
/
Iꢁ redox couple of the electron mediator have been recently
reported.27–29 In these systems, the overall water splitting reac-
tion proceeds under visible light irradiation. The highest quan-
tum yield of the IO3ꢁ/Iꢁ system is 1% at 420 nm. Thus, the
Z-scheme system can be regarded as a candidate for photo-
catalysis systems aimed at water splitting using sunlight. How-
ever, there are significant barriers against its utilization of
the solar energy at the present stage. The efficiency of the Z-
scheme is still low. Moreover, the useful light is limited with
the wavelength shorter than 460 nm, which corresponds to the
band gap of the WO3 photocatalyst.
The authors have recently reported the overall water split-
ting using some Z-scheme photocatalysis systems, which con-
sist of two kinds of authors’ original photocatalysts and an
Fe3þ/Fe2þ redox couple as an electron mediator.29 Their quan-
tum yields are 0.4–0.5% at 420 nm. The schematic diagram of
the reaction is shown in Fig. 1. On the photocatalyst for O2
evolution (denoted as O2-photocatalyst), photogenerated elec-
trons and holes react with Fe3þ ions and water to produce Fe2þ
and O2, respectively (reactions 1 and 2). On the other hand, on
Published on the web December 13, 2007; doi:10.1246/bcsj.80.2457

2458 Bull. Chem. Soc. Jpn. Vol. 80, No. 12 (2007)
Photocatalytic Water Splitting under Vis
connected to a gas-closed circulation system with on-line gas
chromatograph. Photocatalyst powders (50–100 mg for each com-
ponent) were dispersed into a pH-adjusted aqueous solution con-
taining iron ions. Sulfuric acid or perchloric acid was used to ad-
just pH of the reactant solution. The reactant solution was stirred
with a magnetic stirrer. The temperature of reactant solution was
Visible light
Visible light
H2
H+
Pt
CB
e-
CB
e-
Fe2+
Fe3+
Electron relay
h+
Donor level
formed by Rh
VB
H2O
h+
VB
kept at 293 K. Some iron salts, FeCl2 4H2O (Wako, purity:
ꢂ
O2, H+
99.9%), FeCl3 6H2O (Kanto, purity: 99.0%), Fe(NO3)3 9H2O
(Kanto, purity: 99.0%), and Fe2(SO4)3 xH2O (Wako, purity: 60–
ꢂ
ꢂ
H2-Photocatalyst
Rh-doped SrTiO3
O2-Photocatalyst
BiVO4
ꢂ
80% as Fe2(SO4)3), were used as sources of iron ions. An argon
gas (40 Torr) was introduced into the system after deaeration. A
300 W Xe-arc lamp (Parkin-Elmer: Cermax-PE300BF) attached
with a sharp cut-off filter (Hoya: L42) was employed for visible
light irradiation (ꢀ > 420 nm). Gaseous products were analyzed
by using on-line gas chromatograph.
(Band gap: 2.4 eV)
(Energy gap: 2.4 eV)
Fig. 1. Overview of overall water splitting on Z-scheme
photocatalysis system with an iron ion redox couple.
the photocatalyst for H2 evolution (denoted as H2-photo-
catalyst), photogenerated electrons reduce protons to H2, and
photogenerated holes regenerate Fe3þ ions (reactions 3 and 4).
The concentration of an Fe2þ ion, C(Fe2þ), was determined
by using colorimetric analysis based on a 1,10-phenanthroline–
iron(II) complex.32 For the determination of concentration of
Fe3þ ions, C(Fe3þ), Fe3þ ions were firstly reduced to Fe2þ with
hydroxylammonium chloride to determine the whole concen-
tration of iron ions, C(Fe). Then, C(Fe3þ) was estimated from
difference between C(Fe) and C(Fe2þ).
Back-reactions between H2 and O2 to form water and reduction
of Fe3þ to Fe2þ by H2 (reactions 5 and 6) were investigated. H2
(20 Torr) and O2 (10 Torr) were introduced into the reaction
system, in which photocatalyst powders and a reactant solution
were set as usual photocatalytic reactions except for illumination.
Pressure changes were measured in the dark.
Fe3þ þ eꢁ ! Fe2þ
VB
;
ð1Þ
ð2Þ
ð3Þ
ð4Þ
CB
2H2O þ 4hþ ! O2 þ 4Hþ;
CB
2Hþ þ 2eꢁ ! H2;
VB
Fe2þ þ hþ ! Fe3þ
:
Thus, the Fe3þ/Fe2þ redox couple relays electrons from the
O2-photocatalyst to the H2-photocatalyst in the present Z-
scheme system. The Pt cocatalyst is generally regarded as
not effective cocatalyst for water splitting in the powdered
photocatalysis system, because it is active for the rapid back-
reaction to form water. Moreover, Fe3þ ions may inhibit the
H2 production due to capturing of photogenerated electrons
in a conduction band of SrTiO3:Rh and/or to a reaction be-
tween H2 and Fe3þ on Pt. However, the present Z-scheme sys-
tems were active in overall water splitting in spite of the sus-
pension system using a Pt cocatalyst, implying suppression of
the back-reactions on the Pt cocatalyst. Certainly, in other
photocatalysis systems involving Pt cocatalysts, overall water
splitting has been achieved with efficient suppression of the
back reaction by coating with NaOH,11 carbonate,30 and io-
dine.31 Therefore, clarification of the mechanism of suppres-
sion of back-reactions on Pt is important for the present
systems.
2H2 þ O2 ! 2H2O;
ð5Þ
ð6Þ
H2 þ 2Fe3þ ! 2Hþ þ 2Fe2þ
:
Results
Overall Water Splitting on Z-Scheme Photocatalysis
Systems with an Fe3þ/Fe2þ Electron Mediator. Overall
water splitting reactions under visible light irradiation using
a (Pt/SrTiO3:Rh)–(BiVO4) system with FeCl3 or FeCl2 are
shown in Fig. 2. When an aqueous FeCl3 solution was used,
O2 was dominantly produced at the initial stage of the first
run as shown in Fig. 2a. However, the rate of O2 production
decreased gradually. Eventually, H2 and O2 were produced
in a 2:1 stoichiometric ratio. In a second run after evacuation
of gaseous products, H2 and O2 were produced in a stoichio-
metric ratio even at the initial stage. Moreover, the activity,
which decreased in the first run, recovered in the initial period
of the second run. On the other hand, no gases evolved in the
dark. When an aqueous FeCl2 solution was used, O2 evolved
after 3 h of the reaction time, although the ratio H2/O2 (2.7)
deviated from stoichiometry. However, H2 and O2 were pro-
duced in a stoichiometric ratio in the second run. In addition,
the rate of H2 evolution in the second run (20.7 mmol hꢁ1) was
higher than that in the first run (12.1 mmol hꢁ1). Thus, overall
water splitting proceeded in the both cases of FeCl3 and FeCl2.
When the reaction proceeded steadily, the concentrations of
Fe3þ and Fe2þ ions were 1.5 and 0.5 mmol Lꢁ1, respectively,
in the both cases.
In the present research, effects of pH and the Fe3þ/Fe2þ
ratio in a reactant solution on photocatalytic performance of
the Z-scheme system and the back-reactions were investigated
to clear the role of iron ions. A mechanism of overall water
splitting using the Z-scheme system with an iron ion electron
mediator is proposed.
Experimental
Powders of rhodium-doped SrTiO3 (denoted as SrTiO3:Rh) and
BiVO4 were prepared by solid-state and liquid–solid-state reac-
tions, respectively, according to previous literature.20a,c A WO3
powder was purchased from a chemical company (Nacalai tesque,
purity: 99.5%). Deposition of a Pt cocatalyst on SrTiO3:Rh was
carried out using a photodeposition method from H2PtCl6. A pow-
der of SrTiO3:Rh was dispersed in an aqueous methanol solution
(10 vol %) containing the appropriate amount of H2PtCl6, and then
the suspension was irradiated. After photodeposition of the Pt co-
catalyst, the Pt/SrTiO3:Rh powder was filtered, and then washed
with water.
Figure 3 shows overall water splitting using a (Pt/SrTiO3:
Rh)–(BiVO4) system with various iron(III) salts, FeCl3,
Fe(NO3)3, and Fe2(SO4)3. Overall water splitting steadily pro-
ceeded after an induction period in all cases. No remarkable
Photocatalytic reactions were conducted in a top-window cell

H. Kato et al.
Bull. Chem. Soc. Jpn. Vol. 80, No. 12 (2007) 2459
(a)
H2SO4
HClO4
HCl
500
400
300
200
100
0
Dark
λ > 420 nm
500
400
300
200
100
0
Evac.
H2
O2
0
10
20
30
40
50
60
1.3 2.0 2.4 2.5 1.0 2.4 2.4
pH
Time / h
(b)
400
300
200
100
0
Fig. 4. Dependence of photocatalytic activity of the (Pt
(0.5 wt %)/SrTiO3:Rh (1%))–(WO3) system on pH under
visible light irradiation. Catalyst: 0.1 g for each compo-
nent, reactant solution: 2 mmol Lꢁ1 of FeCl3, 120 mL, light
source: 300 W Xe-arc lamp with a cut-off filter (ꢀ > 420
nm), cell: top-irradiation cell with a Pyrex window.
Evac.
differences in gas production profiles and photocatalytic
activities were observed among three cases.
Figure 4 shows the amounts of gases produced for 22 h of
irradiation time over the (Pt/SrTiO3:Rh)–(WO3)–(FeCl3) sys-
tem with various pH values. Here, the pH was adjusted by sul-
furic acid or perchloric acid. The pH value did not change dur-
ing each reaction. The activity of the (Pt/SrTiO3:Rh)–(WO3)–
(FeCl3) system was similar to that of the (Pt/SrTiO3:Rh)–
(BiVO4)–(FeCl3) system as shown in Figs. 2a and 4. The
photocatalytic activity of the present Z-scheme system was
sensitive to pH and the kind of acid used for pH control. When
sulfuric acid was used for the pH adjustment, the Z-scheme
system showed activity in the pH range 1.3–2.5, although
the activity depended on the pH. As the pH was increased
up to 2.4, the activity increased. However, the activity decreas-
ed when the pH was increased to 2.5. On the other hand, when
the pH was adjusted with perchloric acid, the Z-scheme system
showed activity at pH 2.4, although the activity was only one
sixth of that when sulfuric acid was used. When the pH was
decreased to 1, the activity was negligible. A similar activity to
pH-2.4–FeCl3–HClO4 was obtained when the pH was adjusted
with hydrochloric acid. The highest activity was obtained in
the aqueous FeCl3–H2SO4 solution with pH 2.4.
Absorption spectra of 2 mmol Lꢁ1 aqueous FeCl3 solutions
with various pH values, which were adjusted by addition of
sulfuric acid or perchloric acid, are shown in Fig. 5. Two ab-
sorption bands around 225 and 305 nm were observed for all
aqueous FeCl3–H2SO4 solutions. Both absorption bands be-
came small as the pH increased. On the other hand, remarkable
change in the absorption profile was observed between pH 1.3
and 2.4 in the aqueous FeCl3–HClO4 solution. An absorption
maximum was observed at 240 nm in pH-1.3–FeCl3–HClO4,
whereas an absorption maximum at 210 nm and shoulder peak
around 300 nm were observed in pH-2.4–FeCl3–HClO4.
Characteristic Properties of the Z-Scheme Systems with
an Fe3þ/Fe2þ Electron Mediator. Table 1 lists photocata-
lytic activity of Pt/SrTiO3:Rh for H2 evolution under visible
light irradiation from an aqueous FeCl2 solution with variation
0
10
20
Time / h
30
40
Fig. 2. Photocatalytic overall water splitting using (Pt
(0.5 wt %)/SrTiO3:Rh (1%)–(BiVO4) system under visible
light irradiation in an aqueous solution of (a) FeCl3
(2 mmol Lꢁ1) and (b) FeCl2 (2 mmol Lꢁ1). Open marks:
H2, closed marks: O2. Catalyst: 0.1 g for each component,
reactant solution: 120 mL, pH: 2.4, light source: 300 W
Xe-arc lamp with a cut-off filter (ꢀ > 420 nm), cell: top-
irradiation cell with a Pyrex window.
(a)
(b)
(c)
300
250
200
150
100
50
0
0
10
0
10
Time / h
0
10
Fig. 3. Photocatalytic overall water splitting over (Pt
(0.5 wt %)/SrTiO3:Rh (1%))–(BiVO4) system under visi-
ble light irradiation in aqueous solutions of (a) FeCl3,
(b) Fe(NO3)3, and (c) Fe2(SO4)3. Open marks: H2, closed
marks: O2. Catalyst: 0.1 g for each component, reactant
solution: 2 mmol Lꢁ1 of iron, 120 mL, pH: 2.4, light
source: 300 W Xe-arc lamp with a cut-off filter (ꢀ >
420 nm), cell: top-irradiation cell with a Pyrex window.

2460 Bull. Chem. Soc. Jpn. Vol. 80, No. 12 (2007)
Photocatalytic Water Splitting under Vis
60
50
40
30
20
10
0
(a)
2
(a)
(b)
1.5
(c)
(d)
pH 1.1
1
pH 2.2
pH 2.4
0.5
0
0
1
2
3
4
200
250
300
350
400
Time / h
Wavelength / nm
(b)
Fig. 6. Photocatalytic O2 evolution on the BiVO4 photo-
catalyst under visible light irradiation in 2 mmol Lꢁ1 aque-
ous FeCl3 solutions with (a) 0, (b) 1, (c) 2, and (d) 5
mmol Lꢁ1 of FeCl2. Catalyst: 0.1 g, reactant solution:
120 mL, pH: 2.4, light source: 300 W Xe-arc lamp with
a cut-off filter (ꢀ > 420 nm), cell: top-irradiation cell with
a Pyrex window.
2
pH 1.3
1.5
1
In this case, the activity enhancement was significant when
the concentration of FeCl3 was equal to and higher than
2 mmol Lꢁ1. However, when the concentration of FeCl2 was
20 mmol Lꢁ1, the activity was not so high even in the presence
of FeCl3 with 2 mmol Lꢁ1 of the concentration (Run 8). In
contrast, relatively high activity was obtained when the ratio
of [Fe3þ]/[Fe2þ] was at unity even in the low concentration
(Run 7). Thus, the effect of co-presence of FeCl3 on en-
hancement of H2 evolution was significant in the ratio
pH 2.4
0.5
0
200
250
300
Wavelength / nm
350
400
Fig. 5. Absorption spectra of aqueous FeCl3 solutions
(2 mmol Lꢁ1) with various pH values; (a) for the solution
of FeCl3–H2SO4 and (b) for the solution of FeCl3–HClO4.
Optical path: 2 mm.
½Fe3þꢃ=½Fe2þꢃ ꢄ 1 rather than the concentration of Fe3þ
.
In contrast to H2 evolution, the O2 production on the BiVO4
photocatalyst was remarkably inhibited by the co-presence of
Fe2þ ions as shown in Fig. 6. It clearly indicates that oxidation
of water to form O2 competes with oxidation of Fe2þ on the
BiVO4 photocatalyst. However, it is noteworthy that BiVO4
produced O2 even in the presence of a high concentration, such
as 5 mmol Lꢁ1, of Fe2þ ions. On the other hand, the same ac-
tivity was obtained when hydrochloric acid was used for pH
adjustment instead of sulfuric acid. Thus, the Clꢁ ion does
not function as a hole scavenger for the BiVO4 photocatalysts.
Table 2 lists rates of back-reaction (reactions 5 and 6) in the
(Pt (0.5 wt %)/SrTiO3:Rh)–(WO3) system, which was moni-
tored by observing the decrease in the pressure of H2 and O2
in the dark. When iron ions were absent from the solution
(Run 1), H2 and O2 were consumed in the ratio 2:1 due to
the formation of water. In contrast, in the cases of 2 mmol Lꢁ1
of FeCl3, and 1.5 mmol Lꢁ1 of FeCl3 with 0.5 mmol Lꢁ1 of
FeCl2 solutions, of which pH was adjusted to be 2.4 by sulfuric
acid (Runs 2 and 4), H2 was slightly consumed, and the con-
sumption of O2 was negligible. On the other hand, when the
composition of iron ions in the solutions was ½Fe2þꢃ >
½Fe3þꢃ (Runs 7 and 8), H2 and O2 were consumed in a 2:1 ratio
with the comparable rates to that without iron ions. When the
concentration of FeCl3 was as low as 0.5 mmol Lꢁ1 (Run 3),
the rates of consumption of H2 and O2 became higher than
those in the case of 2 mmol Lꢁ1, but were still lower than those
in the absence of iron ions. Interestingly, in this case, the con-
sumption ratio of H2:O2 was 3:1. The significant effect of
Table 1. Photocatalytic Activity of Pt/SrTiO3:Rh for H2
Evolution from an Aqueous Iron Chloride Solution under
Visible Light Irradiationa)
Run
Concentration of iron
chlorides
Rate of H2 evolution
/mmol hꢁ1
/mmol Lꢁ1
FeCl2
FeCl3
1
2
3
4
5
6
7
8
0
2
2
2
2
2
0.4
20
2
0
0.5
1
2
5
0.4
2
ngb)
8.6
9.8
11.1
22.3
23.0
15.3
11.0
a) Catalyst: 50 mg, reactant solution: 120 mL, pH was adjusted
with H2SO4 to be 2.4, light source: 300 W Xe lamp with a
cut-off filter (ꢀ > 420 nm). b) Negligible.
of the FeCl3/FeCl2 composition. It is a half reaction of the Z-
scheme. In an aqueous FeCl3 solution (Run 1), H2 evolution
was negligible, and no O2 was produced. In contrast, the Pt/
SrTiO3:Rh photocatalyst produced H2 in the presence of FeCl2
(Runs 2–8). H2 evolution was surprisingly enhanced with an
increase in the concentration of coexisting FeCl3 when the
concentration of FeCl2 was fixed at 2 mmol Lꢁ1 (Runs 2–6).

H. Kato et al.
Bull. Chem. Soc. Jpn. Vol. 80, No. 12 (2007) 2461
Table 2. Backward Reaction Rates in the (Pt (0.5 wt %)/
SrTiO3:Rh (1%))–(WO3) System in the Darka)
440
22
20
18
16
(b)
Run
Concentration of
iron chloride
/mmol Lꢁ1
Acid pH
Rate of decrease in
pressureb)/Torr hꢁ1
400
FeCl2
FeCl3
H2
O2
360
1
2
3
4
5
6
7
8
9
0
0
0
0.5
0.5
0.5
1.5
2
0
2
H2SO4 2.4
H2SO4 2.4
H2SO4 2.4
H2SO4 2.4
H2SO4 2.0
H2SO4 1.3
H2SO4 2.4
H2SO4 2.4
HClO4 2.4
HClO4 1.0
2.8
trace
1.1
trace
0.36
0.66
3.3
4.1
0.38
2.6
1.4
(a)
ngc)
0.40
ngc)
0.01
0.12
1.6
0.5
1.5
1.5
1.5
0.5
0
2
2
2
320
0
1
2
3
Time / h
2.0
Fig. 7. Decrease in H2 pressure due to the reaction between
H2 and Fe3þ on the Pt (0.3 wt %)/SrTiO3:Rh photocatalyst
in the dark in aqueous solutions; (a) 1 mmol Lꢁ1 FeCl3,
2 mmol Lꢁ1 FeCl2 and (b) 5 mmol Lꢁ1 FeCl3, 2 mmol Lꢁ1
FeCl2. Catalyst: 0.1 g, solution: 120 mL, pH: 2.4, gas phase
volume: 350 mL.
0
0
0
0.11
0.09
0.03
10
11
HCl
2.4
0.47
a) Catalyst: 0.1 g for each component, solution: 120 mL, gas
phase volume: 350 mL. b) Initial pressure: 20 and 10 Torr
for H2 and O2, respectively. c) Negligible.
addition of FeCl3 on the suppression of back-reactions was
also observed for the (Pt/SrTiO3:Rh)–(BiVO4) system.
When the pH value of the solution of 1.5 mmol Lꢁ1 FeCl3
with 0.5 mmol Lꢁ1 FeCl2 became low with addition of sulfuric
acid (Runs 4–6), the consumption rates of both H2 and O2
became high. Although consumption of O2 was observed for
pH 2.0 and 1.3, the consumption ratio of H2:O2 significantly
deviated from 2:1, indicating that reduction of Fe3þ proceeded
in addition to the formation of water.
When perchloric or hydrochloric acid was used to adjust
the pH to 2.4 instead of sulfuric acid (Runs 9 and 11), the con-
sumption of H2 and O2 was also remarkably suppressed in
comparison to the case of the absence of iron ions. However,
the consumption of O2 was still observed, although it was neg-
ligible in the case of sulfuric acid. When the pH was decreased
to 1 with perchloric acid (Run 10), H2 was consumed. The rate
of consumption was similar to that in the absence of iron ions,
whereas O2 consumption was remarkably suppressed.
The reduction of Fe3þ with H2 (reaction 6) on the Pt/
SrTiO3:Rh was examined in the dark by monitoring the de-
crease in pressure of H2 as shown in Fig. 7. The concentration
of Fe2þ was fixed at 2 mmol Lꢁ1, and that of Fe3þ was 1 or 5
mmol Lꢁ1. The reduction of Fe3þ was observed in both cases.
The reaction rate based on the consumption of H2 in the solu-
tion with 5 mmol Lꢁ1 of Fe3þ (0.37 Torr hꢁ1) was one fifth of
that with 1 mmol Lꢁ1 of Fe3þ (1.9 Torr hꢁ1), in spite of the fact
that the former contained five times more substrate than the
latter. The reaction stopped at 2 h of reaction time in the latter
case. It was due to the consumption of Fe3þ (120 mmol) con-
tained in the reactant solution. Thus, stoichiometric consump-
tion of H2 (60 mmol) was confirmed.
irradiation, (Pt/SrTiO3:Rh)–(BiVO4), produced H2 and O2 in a
stoichiometric ratio after an induction period in the presence of
iron ions. The activity of the Z-scheme system at steady state
was independent of iron salts in the starting solution. The Pt/
SrTiO3:Rh photocatalyst did not produce O2 from aqueous
solutions of FeCl3 and FeCl2, whereas H2 evolution was not
observed on the BiVO4 photocatalyst. Therefore, the produc-
tion of O2 or H2 from the starting solution of FeCl2 or FeCl3,
respectively, clearly indicated that iron ions functioned as an
electron mediator without consumption, according to the Z-
scheme as shown in Fig. 1. The concentrations of Fe3þ and
Fe2þ in the steady state were 1.5 and 0.5 mmol Lꢁ1, respective-
ly, in both cases of the starting solution of FeCl3 and FeCl2. An
excess amount of O2 or H2 evolved in the induction period was
observed when either FeCl3 or FeCl2 was used in the starting
solution. The (Pt/SrTiO3:Rh)–(BiVO4)–(FeCl3) system pro-
duced 890 and 450 mmol of H2 and O2, respectively, for 48 h
of irradiation time. The number of reacted electron/hole pairs
in each photocatalyst was determined to be 1.8 mmol from the
amounts of products. It was larger than mole quantities of Pt/
SrTiO3:Rh (545 mmol) and BiVO4 (275 mmol) photocatalysts
used for the reaction. Moreover, it was also larger by 7 times
than Fe3þ in the reactant solution (240 mmol), proving that iron
ions turned over as an electron mediator in the present system.
Thus, the results indicate that the overall water splitting
proceeds as a Z-scheme photocatalysis process.
Relationship between Photocatalytic Activity and Iron
Ion Species in the Solutions.
In an aqueous perchloric
acid solution, FeIII ions are coordinated by six water mole-
cules, [Fe(H2O)6]3þ, if the pH is low enough. However,
[Fe(H2O)6]3þ is sequentially hydrolyzed to [Fe(OH)(H2O)5]2þ
and [Fe(OH)2(H2O)4]þ as the pH increases. Absorption spec-
tra of these FeIII species in an aqueous perchloric acid solution
has been reported by Moulik and Gupta.33 According to the
report, [Fe(H2O)6]3þ and [Fe(OH)(H2O)5]2þ possess absorp-
tion maxima at 240 and 300 nm, respectively. Therefore, the
main iron species in the pH-1.3–FeCl3–HClO4 solution, which
showed an absorption maximum at 240 nm as shown in
Discussion
Overall Water Splitting Using Z-Scheme Systems with
Fe3þ/Fe2þ Electron Mediator under Visible Light Irra-
diation. The system combined with two kinds of photo-
catalysts, which were active for either H2 or O2 formation in
the presence of electron donors or acceptors under visible light

2462 Bull. Chem. Soc. Jpn. Vol. 80, No. 12 (2007)
Photocatalytic Water Splitting under Vis
Fig. 5b, was assigned to [Fe(H2O)6]3þ, and that in the pH-2.4–
FeCl3–HClO4 solution, which showed a shoulder peak at
300 nm, was assigned to [Fe(OH)(H2O)5]2þ. An intense peak
at 210 nm observed in the pH-2.4–FeCl3–HClO4 solution
would also be due to [Fe(OH)(H2O)5]2þ. The pH-2.4–FeCl3–
HCl solution gave a quite similar spectrum to pH-2.4–FeCl3–
HClO4, indicating [Fe(OH)(H2O)5]2þ of the main FeIII species
(see Supporting Information). On the other hand, [Fe(SO4)-
(H2O)5]þ and [Fe(SO4)2(H2O)4]ꢁ, which are formed in an
[Fe(OH)(H2O)5]2+ or
Unsuppressed path
Suppressed path
H2O
[Fe(SO4)(H2O)5]+
[Fe(H2O)6]2+
H2, O2
H2O
H2
, H+
H2, O2
Pt
aqueous sulfuric acid solution, also possess an absorption max-
imum at 300 nm as well as [Fe(OH)(H2O)5]2þ
34
.
The absorp-
SrTiO3:Rh
tion peaks observed at 225 and 300 nm in the pH-1.1– and
pH-2.2–FeCl3–H2SO4 solutions should be due to [Fe(SO4)-
(H2O)5]þ and/or [Fe(SO4)2(H2O)4]ꢁ, because these absorp-
FeCl3-H2SO4 (pH 1-2.4),
FeCl3-HClO4 (pH 2.4)
Without iron(III) ions
2ꢁ
tion grew stronger as the concentration of SO4 was higher.
In the pH-2.4–FeCl3–H2SO4 solution, absorption in the short
wavelength region shifted to 215 nm. It implies that [Fe(OH)-
(H2O)5]2þ, which showed absorption at 210 nm, has a higher
concentration in the pH-2.4–FeCl3–H2SO4 solution than the
solutions with pH 1.3 and 2.2. On the other hand, it is thought
that [Fe(H2O)6]3þ existed in all solutions according to the
equilibrium. The concentration of [Fe(H2O)6]3þ should be
larger as the pH decreases.
The pH dependence of activity of the (Pt/SrTiO3:Rh)–
(WO3)–(FeCl3) system shown in Fig. 4 can be related to FeIII
species in the reactant solution as follows. In the FeCl3–H2SO4
solutions in which [Fe(SO4)(H2O)5]þ, [Fe(OH)(H2O)5]2þ, and
[Fe(H2O)6]3þ were contained, overall water splitting proceed-
ed in all solutions with 1.3–2.5 of the pH range, although
the activity decreased as the pH became low. On the other
hand, in the case of FeCl3–HClO4 system, water splitting pro-
ceeded in the solution with pH 2.4 which contained both of
[Fe(OH)(H2O)5]2þ and [Fe(H2O)6]3þ, whereas there was no
activity in the solution with pH 1, which contained only
[Fe(H2O)6]3þ. Therefore, it is thought that [Fe(SO4)(H2O)5]þ
or [Fe(OH)(H2O)5]2þ species are indispensable for the present
Z-scheme systems to achieve overall water splitting.
Effects of the Ratio of Fe3þ/Fe2þ in the Solution on the
Photocatalytic Activities of Half Reactions. O2 production
from water containing Fe3þ ions on the BiVO4 photocatalyst
was inhibited by the coexistence of Fe2þ ions, as mentioned
previously (Fig. 6). Competition of oxidation reactions be-
tween water and Fe2þ ions is reasonably understood consider-
ing the electron donor ability of Fe2þ. However, it is note-
worthy that the BiVO4 photocatalyst can produce O2 even in
the presence of Fe2þ with 2.5 times higher concentration than
Fe3þ. Ohno et al. have reported a unique effect of iron ions that
adsorption of Fe2þ becomes negligible on a TiO2 photocatalyst
when Fe3þ coexists.35 It is thought that Fe3þ causes the sup-
pression of Fe2þ adsorption also for the BiVO4 photocatalyst,
resulting in the suppression of the inhibition from O2 produc-
tion by Fe2þ. It is preferable property for O2-photocatalysts in
the Z-scheme systems.
Fig. 8. Suppression of backward reactions by iron(III) ions.
of the back-reaction is probably due to adsorption of FeIII
species on the Pt surface. The effect of FeIII species on the
back-reactions is further discussed later. The results also indi-
cate that the reduction of Fe3þ ions by photogenerated elec-
trons is not dominant on the Pt/SrTiO3:Rh photocatalyst. In
the case of Z-scheme systems with an IO3ꢁ/Iꢁ electron medi-
ator, H2 production is eliminated in the presence of small
ꢁ
27,28
amount of IO3
.
However, the fact that H2 production is
not inhibited by Fe3þ ions is the characteristic feature of the
Pt/SrTiO3:Rh photocatalyst. This feature must contribute to
achievement of overall water splitting with the present Z-
scheme system using an Fe3þ/Fe2þ electron mediator.
Suppression of Back-Reactions by Iron Ions. The pro-
posed mechanism for the suppression of back-reactions by
Fe3þ ions from the results (Table 2 and Fig. 7) is illustrated
in Fig. 8. From the results listed in Table 2, the rate of water
formation can be estimated from the decrease in the O2 pres-
sure. In addition, the rate of reduction of Fe3þ ions by H2
can also be estimated from the deviation of the consumption
ratio from 2:1. Water formation from H2 and O2 rapidly pro-
ceeds on the surface of Pt when any Fe3þ species are absent.
In contrast, water formation is significantly suppressed in
the presence of Fe3þ species if the iron ion composition in
the reactant solution is ½Fe3þꢃ ꢄ ½Fe2þꢃ. Such Fe3þ species
would be adsorbed by not only the SrTiO3:Rh support but
also the Pt surface, resulting in the inhibition of water forma-
tion. Thus, Fe3þ species adsorbed on the Pt surface suppress
water formation as well as NaOH,11 carbonate,30 and iodine.31
However, the suppression of water formation is much more
efficient in the pH-2.4–FeCl3–H2SO4 solution than in the
pH-2.4–FeCl3–HClO4 and pH-2.4–FeCl3–HCl solutions. It is
due to the difference in strength of chemisorpion between
[Fe(SO4)(H2O)5]þ and [Fe(OH)(H2O)5]2þ. The [Fe(SO4)-
(H2O)5]þ ions may bind to Pt via oxygen atoms of the sulfate
ligand more strongly than [Fe(OH)(H2O)5]2þ ions. In contrast
to the Fe3þ species, [Fe(H2O)6]2þ does not suppress water for-
mation. On the other hand, the reduction of Fe3þ ions by H2 of
another back-reaction is also remarkably suppressed when the
iron ion composition is [Fe3þ] ꢄ [Fe2þ], except for the pH-1–
FeCl3–HClO4 solution, in which [Fe(H2O)6]3þ is the main
Fe3þ species. Thus, when [Fe(SO4)(H2O)5]þ and/or [Fe(OH)-
(H2O)5]2þ are the main iron species in the reactant solution,
On the other hand, H2 production from water containing
Fe2þ ions increased remarkably when Fe3þ ions were present
in higher concentration than Fe2þ (½Fe3þꢃ ꢄ ½Fe2þꢃ). It is
due to that the back-reaction between H2 and Fe3þ (reaction 6)
is significantly suppressed when the composition of iron ions
in the reactant solution is ½Fe3þꢃ ꢄ ½Fe2þꢃ. The suppression

H. Kato et al.
Bull. Chem. Soc. Jpn. Vol. 80, No. 12 (2007) 2463
both back-reactions are remarkably suppressed. However, the
efficiency for suppression of reduction of Fe3þ by H2 is dif-
ferent between [Fe(SO4)(H2O)5]þ and [Fe(OH)(H2O)5]2þ. It
is thought that [Fe(SO4)(H2O)5]þ suppresses the reduction
of Fe3þ more efficiently than [Fe(OH)(H2O)5]2þ (Runs 2 and
9 in Table 2). On the other hand, the reduction of Fe3þ oc-
curred rapidly in the pH-1–FeCl3–HClO4 solution, in which
[Fe(H2O)6]3þ is the main Fe3þ species. It indicates that
[Fe(H2O)6]3þ does not inhibit the adsorption of H2 on Pt.
Then, [Fe(H2O)6]3þ is reduced instead of O2. The increase
in the rates of back-reactions by decreasing the pH, observed
in the FeCl3–H2SO4 solutions, is due to the increase in the
concentration of [Fe(H2O)6]3þ (Runs 4–6 in Table 2). As de-
scribed above, it has been found that back-reactions on the
Pt surface are efficiently suppressed by [Fe(OH)(H2O)5]2þ
and/or [Fe(SO4)(H2O)5]þ. It is a characteristic feature of the
present Z-scheme systems that iron ions contribute to water
splitting not only as the electron mediator but also as an
inhibitor of the back-reactions.
The back-reactions should be suppressed by a decrease in
sites for the dissociative adsorption of H2 and in the mobility
of the dissociated hydrogen on the Pt surface due to the
adsorption of Fe3þ species, such as [Fe(SO4)(H2O)5]þ and
[Fe(OH)(H2O)5]2þ. Therefore, it is concluded that the suppres-
sion of reduction of Fe3þ with H2 by adsorption of Fe3þ is the
main reason for the promotion of H2 production with increas-
ing in the concentration of Fe3þ ions as listed in Table 1. The
promotion of H2 production accompanied by the increase in
the concentration of Fe3þ ions is somewhat strange from
the viewpoint of electron acceptability of Fe3þ. However,
the positive effect of Fe3þ on suppression of the back-reaction
should overcome the negative effect of Fe3þ on the trapping of
photogenerated electrons.
and iron ions of an electron mediator. It has been experimen-
tally demonstrated that the iron ions turned over as an electron
mediator. The presence of Fe2þ inhibited O2 production on
the BiVO4 photocatalyst. On the other hand, H2 production
on the Pt/SrTiO3:Rh photocatalyst increased when Fe3þ was
presented in a higher concentration than Fe2þ
.
The back-reactions of water formation and reduction of
Fe3þ on the Pt surface were remarkably suppressed in the pres-
ence of Fe3þ ions. The suppression of the back-reactions de-
pended on the kind of iron species in the reactant solution. It
is clear from analysis of the back-reactions in the dark that
the most effective iron species for suppression of back-
reactions is [Fe(SO4)(H2O)5]þ. In addition, the suppression
of back-reactions was significant when the concentration of
Fe3þ was higher than that of Fe2þ
.
The composition of iron ions in the reactant solution was
½Fe3þꢃ > ½Fe2þꢃ in the steady state of the overall water split-
ting reaction. In addition, the highest activity was obtained
when [Fe(SO4)(H2O)5]þ was the main Fe3þ species. Thus,
the conditions needed to suppress the back-reactions efficiently
were determined, resulting in overall water splitting, in spite of
using the Pt cocatalyst.
This work was supported by the Core Research Evolutional
Science and Technology (CREST), and Grant-in-Aids for
the Priority Area (No. 417) and for Young Scientists
(No. 17750135) from the MEXT of the Japanese Government.
Supporting Information
Absorption spectrum of a pH-2.4–FeCl3–HCl solution. This
material is available free charge on the Web at: http://www.csj.jp/
journals/bcsj.
The pH dependence of photocatalytic activity of the present
Z-scheme system for water splitting (Fig. 4) can be explained
in terms of the back-reactions as follows. The back-reactions
proceed faster as the pH of the reactant solution becomes
lower (Table 2). Therefore, the decrease in the suppression
of back-reactions due to an increase in the concentration of
[Fe(H2O)6]3þ is the main reason for the pH dependence of ac-
tivity. On the other hand, the decrease in the activity in the
pH 2.5 solution is due to the collapse of active site by adsorp-
tion of the FeO(OH) colloid, which is caused by the pH
increase.
As described above, the rates of back-reactions are strongly
dependent upon the iron species in the reactant solution in
the present Z-scheme system using an Fe3þ/Fe2þ electron
mediator. [Fe(SO4)(H2O)5]þ and/or [Fe(OH)(H2O)5]2þ effec-
tively suppress such back-reactions in comparison with [Fe-
(H2O)6]3þ. It is concluded that the suppression of back-reac-
tions significantly contributes to achievement of overall water
splitting by construction of the Z-scheme system in spite of
using a Pt cocatalyst.
References
1
2
A. Fujishima, K. Honda, Nature 1972, 238, 37.
H. Yoneyama, H. Sakamoto, H. Tamura, Electrochem.
Acta 1975, 20, 341.
3 A. B. Ellis, S. W. Kaiser, M. S. Wrighton, J. Phys. Chem.
1976, 80, 1325.
4 G. Campet, M. P. Dare-Edwards, A. Hamnett, J. B.
Goodenough, Nouv. J. Chim. 1980, 4, 501.
5 M. Matsumura, M. Hiramoto, H. Tsubomura, J. Electro-
chem. Soc. 1983, 130, 326.
6
O. Khaselev, J. A. Turner, Science 1998, 280, 425.
S. U. M. Khan, J. Akikusa, J. Phys. Chem. B 1999, 103,
7
7184.
8
9
2001, 105, 936.
M. Gratzel, Nature 2001, 414, 338.
¨
C. Santato, M. Ulmann, J. Augustynski, J. Phys. Chem. B
10 S. U. M. Khan, M. Al-Shahry, W. B. Ingler, Jr., Science
2002, 297, 2243.
11 S. Sato, J. M. White, Chem. Phys. Lett. 1980, 72, 83.
12 J. M. Lehn, J. P. Sauvage, R. Ziessel, Nouv. J. Chim. 1980,
4, 623.
13 K. Domen, J. N. Kondo, M. Hara, T. Takata, Bull. Chem.
Soc. Jpn. 2000, 73, 1307.
Conclusion
Photocatalytic water splitting into H2 and O2 in a stoichio-
metric ratio under visible light irradiation was possible by
using a Z-scheme system composed of two kinds of visible-
light driven photocatalysts, Pt/SrTiO3:Rh and BiVO4 or WO3,
14 A. Kudo, H. Kato, I. Tsuji, Chem. Lett. 2004, 33, 1534.
15 a) S. Ogura, M. Kohno, K. Sato, Y. Inoue, Appl. Surf. Sci.
1997, 121–122, 521. b) M. Kohno, T. Kaneko, S. Ogura, K. Sato,
Y. Inoue, J. Chem. Soc., Faraday Trans. 1998, 94, 89. c) J. Sato,

2464 Bull. Chem. Soc. Jpn. Vol. 80, No. 12 (2007)
Photocatalytic Water Splitting under Vis
N. Saito, H. Nishiyama, Y. Inoue, J. Phys. Chem. B 2003,
107, 7965.
16 H. Kato, K. Asakura, A. Kudo, J. Am. Chem. Soc. 2003,
125, 3082.
17 J. Sato, N. Saito, Y. Yamada, K. Maeda, T. Takata, J. N.
Kondo, M. Hara, H. Kobayashi, K. Domen, Y. Inoue, J. Am.
Chem. Soc. 2005, 127, 4150.
18 a) K. Maeda, T. Takata, M. Hara, N. Saito, Y. Inoue,
H. Kobayashi, K. Domen, J. Am. Chem. Soc. 2005, 127, 8286.
b) K. Maeda, K. Teramura, D. Lu, T. Takata, N. Saito, Y. Inoue,
K. Domen, Nature 2006, 440, 295.
Chem., Int. Ed. 2005, 44, 3565. c) I. Tsuji, H. Kato, H. Kobayashi,
A. Kudo, J. Phys. Chem. B 2005, 109, 7323.
24 A. Ishikawa, T. Takata, J. N. Kondo, M. Hara, H.
Kobayashi, K. Domen, J. Am. Chem. Soc. 2002, 124, 13547.
25 K. Fujihara, T. Ohno, M. Matsumura, J. Chem. Soc.,
Faraday. Trans. 1998, 94, 3705.
26 K. Sayama, R. Yoshida, H. Kusama, K. Okabe, Y. Abe,
H. Arakawa, Chem. Phys. Lett. 1997, 277, 387.
27 a) K. Sayama, K. Mukasa, R. Abe, Y. Abe, H. Arakawa,
Chem. Commun. 2001, 2416. b) K. Sayama, K. Mukasa, R.
Abe, Y. Abe, H. Arakawa, J. Photochem. Photobiol., A 2002,
148, 71. c) R. Abe, K. Sayama, H. Sugihara, J. Phys. Chem. B
2005, 109, 16052.
19 J. R. Darwent, A. Mills, J. Chem. Soc., Faraday. Trans. 2
1982, 78, 359.
20 a) A. Kudo, K. Omori, H. Kato, J. Am. Chem. Soc. 1999,
121, 11459. b) T. Ishii, H. Kato, A. Kudo, J. Photochem. Photo-
biol., A 2004, 163, 181. c) R. Konta, T. Ishii, H. Kato, A. Kudo,
J. Phys. Chem. B 2004, 108, 8992. d) R. Niishiro, H. Kato, A.
Kudo, Phys. Chem. Chem. Phys. 2005, 7, 2241.
21 a) G. Hitoki, A. Ishikawa, T. Takata, J. N. Kondo, M. Hara,
K. Domen, Chem. Lett. 2002, 736. b) G. Hitoki, T. Takata, J. N.
Kondo, M. Hara, H. Kobayashi, K. Domen, Chem. Commun.
2002, 1698. c) A. Kasahara, K. Nukumizu, G. Hitoki, T. Takata,
J. N. Kondo, M. Hara, H. Kobayashi, K. Domen, J. Phys. Chem.
A 2002, 106, 6750.
28 R. Abe, T. Takata, H. Sugihara, K. Domen, Chem.
Commun. 2005, 3829.
29 H. Kato, M. Hori, R. Konta, Y. Shimodaira, A. Kudo,
Chem. Lett. 2004, 33, 1348.
30 K. Sayama, H. Arakawa, J. Chem. Soc., Faraday Trans.
1997, 93, 1647.
31 R. Abe, K. Sayama, H. Arakawa, Chem. Phys. Lett. 2003,
371, 360.
32 W. B. Fortune, M. G. Mellon, Ind. Eng. Chem., Anal. Ed.
1938, 10, 60.
33 S. P. Moulik, K. K. S. Gupta, J. Indian Chem. Soc. 1972,
49, 447.
34 R. A. Whiteker, N. Davidson, J. Am. Chem. Soc. 1953, 75,
3081.
22 T. Sakata, in Heterogeneous Photocatalysis at Liquid–
Solid Interfaces in Photocatalysis, ed. by N. Serpone, E.
Pelizzetti, Wiley, New York, 1984, p. 311.
23 a) I. Tsuji, H. Kato, H. Kobayashi, A. Kudo, J. Am. Chem.
Soc. 2004, 126, 13406. b) I. Tsuji, H. Kato, A. Kudo, Angew.
35 T. Ohno, D. Haga, K. Fujihira, M. Matsumura, J. Phys.
Chem. B 1997, 101, 6415.