Isotopic and kinetic assessment of photocatalytic water splitting on Zn-added Ga2O3 photocatalyst loaded with Rh2-yCryO3 cocatalyst

Article abstract of DOI:10.1016/j.cplett.2010.01.006

Photocatalytic overall water splitting using Zn-added Ga₂O₃ modified with Rh_2-yCr_yO₃ showed a weaker hydrogen-deuterium isotope effect and a lower apparent activation energy than previously reported electrochemical H₂ evolution reactions. The photocatalytic reaction rates were dominated by bulk processes, rather than surface redox reactions. The isotope effect varied with the amounts of photocatalyst and cocatalyst, and with light intensity, due to varying contributions from surface reactions.

Full text of DOI:10.1016/j.cplett.2010.01.006

Chemical Physics Letters 486 (2010) 144–146
Contents lists available at ScienceDirect
Chemical Physics Letters
journal homepage: www.elsevier.com/locate/cplett
Isotopic and kinetic assessment of photocatalytic water splitting on Zn-added
Ga2O3 photocatalyst loaded with Rh2ÀyCryO3 cocatalyst
Takashi Hisatomi a, Kazuo Miyazaki a, Kazuhiro Takanabe a, Kazuhiko Maeda a, Jun Kubota a,
Yoshihisa Sakata b, Kazunari Domen a,
*
a Department of Chemical System Engineering, School of Engineering, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8656, Japan
b Graduate School of Science and Engineering, Yamaguchi University, 2-16-1 Tokiwadai, Ube, Yamaguchi 755-8611, Japan
a r t i c l e i n f o
a b s t r a c t
Article history:
Photocatalytic overall water splitting using Zn-added Ga2O3 modified with Rh2ÀyCryO3 showed a weaker
hydrogen–deuterium isotope effect and a lower apparent activation energy than previously reported
electrochemical H2 evolution reactions. The photocatalytic reaction rates were dominated by bulk pro-
cesses, rather than surface redox reactions. The isotope effect varied with the amounts of photocatalyst
and cocatalyst, and with light intensity, due to varying contributions from surface reactions.
Ó 2010 Elsevier B.V. All rights reserved.
Received 14 October 2009
In final form 7 January 2010
Available online 11 January 2010
1. Introduction
Ga2O3 (band gap: ca. 4.6 eV) modified with an appropriate
cocatalyst, such as NiO and Ni, has photocatalytic activity for over-
Photocatalytic water splitting has been studied over the past
three decades as a promising means of solar energy conversion
[1–3]. The reaction proceeds via bulk processes, such as photoexci-
tation, charge separation, migration, and recombination, and also
via surface redox reactions [1–3]. It is necessary to understand
which reaction steps have a significant impact on the overall pho-
tocatalytic reaction rate in order to design an efficient photocata-
lytic system.
A useful knowledge of the mechanistic bottlenecks in photocat-
alytic reactions can be elicited from isotopic and kinetic studies
[4–6]. Baba et al. measured isotopic separation factors in photocat-
alytic water decomposition using TiO2 modified with noble metals
to identify the active sites for photocatalytic H2 evolution [4]. Hus-
sein et al. suggested that the apparent activation energy for photo-
catalytic dehydrogenation of propan-2-ol on TiO2 was associated
with the trap depth of TiO2, since the apparent activation energy
was constant among TiO2 samples modified with various noble
metals [5]. Recently, we observed a weak isotope effect and a
low apparent activation energy of photocatalytic water splitting
on visible-light-active Rh2ÀyCryO3-loaded (Ga1ÀxZnx)(N1ÀxOx) un-
der steady light irradiation. We attributed this to a shortage of
photoexcited carriers contributing to surface redox reactions, be-
cause of rapid charge recombination in the bulk [6]. Finally, we
proposed that improving the bulk properties of (Ga1ÀxZnx)(N1ÀxOx)
by reducing the recombination constant and moderating the
n-type semiconductor property would boost photocatalytic
performance.
all water splitting under ultraviolet irradiation [7,8]. Recently, Sak-
ata et al. found that addition of Zn to Ga2O3 [8] and modification
with Rh2ÀyCryO3 as a cocatalyst [9] drastically improved the photo-
catalytic activity of Ga2O3. In our preliminary experiments, the
apparent quantum efficiency for overall water splitting on Zn-
added Ga2O3 modified with Rh2ÀyCryO3 (designated as Rh2ÀyCryO3/
Zn:Ga2O3 hereafter) was estimated to be 10% at a wavelength of
254 nm. The study of Rh2ÀyCryO3/Zn:Ga2O3 photocatalyst can elu-
cidate the rate-limiting step during overall water splitting in a
more generalized way because undesirable charge recombination
is eliminated as much as possible on this highly active photocata-
lyst. In the present study, the hydrogen–deuterium isotope effect
and the apparent activation energy for photocatalytic overall water
splitting were investigated to examine contribution of surface
reaction rates to the entire photocatalytic reaction rate and reac-
tion paths determining overall photocatalytic activity.
2. Experimental
Zn:Ga2O3 was prepared by a previously reported method [8].
The amount of Zn was set to 1 mol% relative to the Ga ions in
Ga2O3. b-Ga2O3 powder (Kojundo Chemical Laboratories Co., Ltd.,
99.9%) was suspended in distilled water containing an appropriate
amount of Zn(NO3)2. The suspension was evaporated to dryness
with continuous stirring. The powder was collected and calcined
at 1273 K for 6 h to produce Zn:Ga2O3. The Zn:Ga2O3 photocatalyst
was loaded with Rh2ÀyCryO3 cocatalysts (metallic content: Rh
0.5 wt.%–Cr 0.75 wt.% and Rh 0.05 wt.%–Cr 0.075 wt.%) using a pre-
viously reported impregnation method [9,10]. Alternative samples
* Corresponding author. Fax: +81 3 5841 8838.
E-mail address: domen@chemsys.t.u-tokyo.ac.jp (K. Domen).
0009-2614/$ - see front matter Ó 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.cplett.2010.01.006

T. Hisatomi et al. / Chemical Physics Letters 486 (2010) 144–146
145
were prepared by instead depositing Ni (1 wt.%) cocatalyst in order
to compare cocatalyst performance [8].
for H2O splitting and 9.4 mmol hÀ1 of D2 and 4.6 mmol hÀ1 of O2
for D2O splitting, yielding an isotope effect of 1.2. This was slightly
higher than that observed under irradiation by a Xe lamp, but
smaller than the reported separation factors for electrochemical
and photocatalytic H2 evolution (3–8) [4], in which bond cleavage
and formation processes involving H atoms known as Volmer
(electrochemical discharge), Heyrovsky (electrochemical desorp-
tion), and Tafel (catalytic desorption) reactions were rate-deter-
mining. Note that the number of photons absorbed by the
photocatalyst was estimated to be at least 100 times higher under
irradiation by the Hg lamp than under the Xe lamp, from compar-
ison of the photocatalytic water splitting rates.
Photocatalytic reactions were performed in a reaction vessel
connected to a closed circulation system. The photocatalyst pow-
der was suspended in H2O (distilled water) or D2O (Cambridge Iso-
tope Laboratories Inc., isotopic purity 99%). After evacuation, the
suspension was irradiated using either a 300 W Xe lamp equipped
with a dichroic mirror through a quartz window of a top-irradia-
tion cell, or a 450 W high-pressure Hg lamp through a quartz com-
ponent of an inner-irradiation cell. The temperature of the
suspension was maintained at approximately 288 K using a chiller.
Gaseous products were analyzed by gas chromatography (TCD,
MS-5A column, Ar carrier).
Table 1 summarizes the isotope effect on photocatalytic overall
water splitting by Rh2ÀyCryO3/Zn:Ga2O3 photocatalysts. The iso-
tope effect became larger when the light intensity increased (from
Xe lamp to Hg lamp) and when the loading amount of cocatalyst
and the amount of photocatalyst were reduced. Decreasing the
amount of photocatalyst and increasing the light intensity in-
creased the average number of photons absorbed by each photo-
catalyst particle. Reduced cocatalyst loading created a kinetic
bottleneck on the H2 evolution process consuming photoexcited
electrons. This resulted in the accumulation of photoexcited elec-
trons in the photocatalyst particles. Thus, a stronger isotope effect
resulted from the higher steady-state concentration of photoexcit-
ed carriers available for surface redox reactions under irradiation.
The overall water splitting rates under irradiation of a Hg lamp de-
creased with the increase in concentration of photoexcited carriers,
because charge recombination was not proportional to the concen-
tration of photoexcited carriers. The absence of an isotope effect
under Xe lamp irradiation was due to light-intensity-limited pho-
tocatalysis, in which the isotopes had little influence on the photo-
catalytic reaction rate because the overall rates were insensitive to
adsorption, desorption, or bond cleavage steps associated with H2O
and D2O on the surface. Note that the isotope effect was as small as
1.2, although the present Rh2ÀyCryO3/Zn:Ga2O3 had an excellent
water splitting rate, as shown in Fig. 1b. This was because approx-
imately 90% of the photoexcited carriers recombined, limiting the
number of photoexcited carriers contributing to the water splitting
reaction. The deficiency of active carriers because of inevitable
charge recombination should be a common feature of photocata-
lytic overall water splitting reactions with apparent quantum effi-
ciencies below 10%. In contrast, active electrons can be supplied
sufficiently to electrodes by external voltage in typical electro-
chemical reactions, which guarantees that surface reaction steps
are rate-determining. Thus, the strong isotope effects reflecting
the surface-limited kinetics can be observed for the electrochemi-
cal H2 evolution reactions.
To investigate the temperature dependence of the photocata-
lytic activity, the temperature of the suspension was controlled
at each target temperature using a chiller and a thermostat under
relatively weak irradiation using a Xe lamp, and the light irradia-
tion was stopped after 1 h. Gaseous products were analyzed after
the suspension was cooled to ambient temperature to avoid any
uneven distribution of gas components in the reaction system
due to vapor pressure. To evaluate the stability of the photocatalyst
at elevated temperatures, an open flow system was employed for
photocatalytic overall water splitting. The suspension was irradi-
ated under a continuous flow of Ar and N2 at 101 kPa, with the
N2 serving as an internal standard for gas analysis. This system en-
abled us to examine the dependence of photocatalytic water split-
ting rate on reaction time at various temperatures.
3. Results and discussion
Fig. 1 shows time courses of H2O and D2O splitting reactions on
Rh2ÀyCryO3/Zn:Ga2O3 (Rh 0.5 wt.%–Cr 0.75 wt.%). Stoichiometric
water splitting was observed. At steady state, the photocatalytic
reaction rates were 117
H2O splitting and 114
l
l
mol hÀ1 of H2 and 53
l
mol hÀ1 of O2 for
mol hÀ1 of D2 and 53
l
mol hÀ1 of O2 for
D2O splitting under irradiation by a Xe lamp. The isotope effect, de-
fined as the ratio of the formation rates of H2 and D2, was unity.
Under irradiation by a high-pressure Hg lamp, the photocatalytic
reaction rates were 11.1 mmol hÀ1 of H2 and 5.7 mmol hÀ1 of O2
0.30
0.25
0.20
0.15
0.10
0.05
0.00
15
10
5
a
b
Fig. 2 shows a time course of the overall water splitting reaction
on Rh2ÀyCryO3/Zn:Ga2O3 (Rh 0.5 wt.%–Cr 0.75 wt.%) at various
temperatures. The photocatalyst showed higher activity at ele-
vated temperatures than at room temperature, with no notable
Table 1
Dependence of water splitting rate and isotope effect on reaction conditions.
Zn-added
Ga2O3 (g)
Rh2ÀyCryO3
(wt.%)
Light
H2O splitting Isotope
sourcea
rate
effectb
(mmol hÀ1
)
0
0.4
0.4
0.4
Rh 0.5–Cr 0.75 Xe
Rh 0.5–Cr 0.75 Hg
0.11
11.2
7.2
1.0
1.2
1.4
0.0 0.5 1.0 1.5 2.0 2.5
Reaction time / hour
0.0 0.5 1.0 1.5 2.0 2.5 3.0
Reaction time / hour
Rh 0.05–Cr
0.075
Hg
0.04
Rh 0.05–Cr
0.075
Hg
1.0
1.9
Fig. 1. Time courses of H2O (closed symbols) and D2O (open symbols) splitting
reactions using Zn-added Ga2O3 loaded with Rh2ÀyCryO3 (Rh 0.5 wt.%–Cr 0.75 wt.%).
Circles and triangles indicate H2 (D2) and O2 evolution, respectively. Reaction
conditions: catalyst, 0.40 g; solution, 140 mL; light source, (a) 300 W Xe lamp
(200 nm < k < 500 nm), (b) 450 W high-pressure Hg lamp (k > 200 nm).
a
Xe: 300 W Xe lamp (200 nm < k < 500 nm), Hg: 450 W high-pressure Hg lamp
(k > 200 nm).
b
Ratio of decomposition rates of H2O and D2O.

146
T. Hisatomi et al. / Chemical Physics Letters 486 (2010) 144–146
tions proceed after a sequence of photoexcitation and charge
dark
migration in photocatalytic water splitting. Under relatively weak
irradiation using a Xe lamp, the reaction rates were significantly
limited by the frequency of photon absorption by photocatalyst
particles rather than rates of surface redox reactions. Rh2ÀyCryO3
cocatalyst, with its smaller apparent activation energy, had higher
photocatalytic activity than Ni in addition to having a negligible
isotope effect. This suggests that electron injection from the photo-
catalyst to the cocatalyst was responsible for the water splitting
rates and apparent activation energies of the present photocata-
lysts. In fact, there exists a potential barrier for electron injection
to a cocatalyst at the interface of an n-type semiconductive photo-
electrode and a deposited cocatalyst [13], while the observed acti-
vation energies did not directly represent the activation barrier for
electron injection, because they were subject to the influence of
light absorption and carrier migration inside the photocatalyst
bulk.
301 K
324 K
345 K
150
100
50
0
0
5
10
15
20
25
30
4. Conclusion
Reaction time / hour
The weak hydrogen–deuterium isotope effect and small appar-
ent activation energy for photocatalytic overall water splitting on
Rh2ÀyCryO3/Zn:Ga2O3 resulted from a limited number of photoex-
cited carriers participating in surface redox reactions. Although
Rh2ÀyCryO3/Zn:Ga2O3 had an excellent water splitting rate, the
majority of photoexcited carriers recombined before migrating to
active surface sites where they could contribute to the water split-
ting reaction. The isotopic and kinetic approach successfully dem-
onstrates that the enhancement of bulk processes, such as carrier
migration to the photocatalyst surface, electron injection to cocat-
alyst particles, and carrier lifetime, rather than the improvement of
surface redox reactions, can improve the photocatalytic activity of
Rh2ÀyCryO3/Zn:Ga2O3. The method and concept presented in this
study will be applicable to other photocatalysts and provide useful
information on relative rates of bulk processes and surface
reactions.
Fig. 2. Time course of photocatalytic water splitting rate on Zn-added Ga2O3 loaded
with Rh2ÀyCryO3 (Rh 0.5 wt.%–Cr 0.75 wt.%) at various temperatures. Circles and
triangles indicate H2 and O2, respectively. Reaction conditions: catalyst, 0.40 g;
distilled water, 140 mL; light source, 300 W Xe lamp (200 nm < k < 500 nm).
5.5
5.0
4.5
4.0
n
3.5
3.0
2.5
Acknowledgements
This work was supported by the Research and Development in a
New Interdisciplinary Field Based on Nanotechnology and Materi-
als Science Program of the Ministry of Education, Culture, Sports,
Science and Technology of Japan (MEXT), the Global Centre of
Excellence (GCOE) Program for Chemistry Innovation, and the To-
kyo Metropolitan Collaboration of Regional Entities for the
Advancement of Technological Excellence, Japan Science and Tech-
nology Agency (JST).
3.1
3.2
3.3
3.4
3.5
3.6
1000 T -1/ K-1
Fig. 3. Temperature dependences of photocatalytic water splitting rate on Zn-
added Ga2O3 loaded with Rh2ÀyCryO3 (Rh 0.5 wt.%–Cr 0.75 wt.%, open symbols) or
Ni (1 wt.%, closed symbols). Circles and triangles indicate H2 and O2, respectively.
Reaction conditions: catalyst, 0.40 g; distilled water, 140 mL; light source, 300 W Xe
lamp (200 nm < k < 500 nm).
References
[1] K. Maeda, K. Domen, J. Phys. Chem. C 111 (2007) 7851.
[2] A. Kudo, Y. Miseki, Chem. Soc. Rev. 38 (2009) 253.
[3] Y. Inoue, Energy Environ. Sci. 2 (2009) 364.
[4] R. Baba, S. Nakabayashi, A. Fujishima, K. Honda, J. Phys. Chem. 89 (1985) 1902.
[5] F.H. Hussein, R. Rudham, J. Chem. Soc. Faraday Trans. 1 (80) (1984) 2817.
[6] T. Hisatomi, K. Maeda, K. Takanabe, J. Kubota, K. Domen, J. Phys. Chem. C 113
(2009) 21458.
[7] T. Yanagida, Y. Sakata, H. Imamura, Chem. Lett. 33 (2004) 726.
[8] Y. Sakata, Y. Matsuda, T. Yanagida, K. Hirata, H. Imamura, K. Teramura, Catal.
Lett. 125 (2008) 22.
[9] Y. Matsuda, T. Yanagida, Y. Sakata, H. Imamura, K. Teramura, in: 14th
International Congress on Catalysis, Seoul, Korea, July 2008 (Abstract PIII-65-
37).
[10] K. Maeda, K. Teramura, D. Lu, T. Takata, N. Saito, Y. Inoue, K. Domen, J. Phys.
Chem. B 110 (2006) 13753.
[11] A.N. Correia, S.A.S. Machado, Electrochim. Acta 43 (1998) 367.
[12] P.K. Wrona, A. Lasia, M. Lessard, H. Ménard, Electrochim. Acta 37 (1992) 1283.
[13] M. Yoshida, A. Yamakata, K. Takanabe, J. Kubota, M. Osawa, K. Domen, J. Am.
Chem. Soc. 131 (2009) 13218.
deactivation. There was no significant difference in X-ray diffrac-
tion patterns or UV–Visible diffuse reflectance spectra of the phot-
ocatalysts before and after the reaction. These results demonstrate
the excellent durability of the photocatalyst.
Fig. 3 shows Arrhenius plots for the photocatalytic water
splitting reaction on Zn:Ga2O3 modified with Rh2ÀyCryO3 (Rh
0.5 wt.%–Cr 0.75 wt.%) or Ni (1 wt.%). The water splitting rates
monotonically increased with increasing reaction temperature in
the range from 278 to 323 K. The apparent activation energies for
the photocatalytic reaction were 8 or 15 kJ molÀ1 for photocata-
lysts loaded with Rh2ÀyCryO3 or Ni, respectively. These were com-
parable to the apparent activation energies of some photocatalytic
reactions [5,6], but were considerably smaller than those of elec-
trochemical H2 evolution on metal electrodes (31 kJ molÀ1 for Rh
and 56 kJ molÀ1 for Ni) [11,12]. This is because surface redox reac-