Relationship between cation arrangement and photocatalytic activity for Sr-Al-Nb-O double perovskite

Article abstract of DOI:10.1021/ic101208q

The relationship between the photocatalytic activity and the arrangement of metal cations was investigated with Sr-Al-Nb-O double perovskite (SAN) synthesized at 1400 °C for various calcination times using a solid state reaction. Transmission electron microscopic observation revealed that SAN particles had a domain structure of completely B-site ordered (Sr ₂AlNbO₆) and disordered (SrAl_0.5Nb _0.5O₃) phases. The ordered phase fraction was determined using a newly proposed mixed-phase model for the Rietveld refinement and a method using the relative intensity of the superlattice line of powder X-ray diffraction. It turned out that the mass fraction of the ordered phase in SAN calcined at 1400 °C could be controlled by the calcination time as 33% (10 h), 37% (20 h), 44% (30 h), and 48% (50 h). Photocatalytic activities of SAN for the evolution of H₂ and O₂ respectively from aqueous solutions of methanol and AgF decreased with increasing the calcination time, that is, with increasing the fraction of the ordered phase. These results suggested that the photocatalytic activity of ordered Sr₂AlNbO ₆ should be lower than that of disordered SrAl_0.5Nb _0.5O₃. This is practically the first report to reveal the photocatalytic activity of SAN as well as the effect of cation ordering in oxides on the photocatalytic activity.

Full text of DOI:10.1021/ic101208q

11362 Inorg. Chem. 2010, 49, 11362–11369
DOI: 10.1021/ic101208q
Relationship between Cation Arrangement and Photocatalytic Activity
for Sr-Al-Nb-O Double Perovskite
Hironori Iwakura,† Hisahiro Einaga,‡ and Yasutake Teraoka*,‡
†Department of Molecular and Material Sciences, Interdisciplinary Graduate School of Engineering Sciences,
‡
Kyushu University, Kasuga, Fukuoka 816-8580, Japan, and Department of Energy and Material Sciences,
Faculty of Engineering Sciences, Kyushu University, Kasuga, Fukuoka 816-8580, Japan
Received June 17, 2010
The relationship between the photocatalytic activity and the arrangement of metal cations was investigated with
Sr-Al-Nb-O double perovskite (SAN) synthesized at 1400 °C for various calcination times using a solid state
reaction. Transmission electron microscopic observation revealed that SAN particles had a domain structure of
completely B-site ordered (Sr2AlNbO6) and disordered (SrAl0.5Nb0.5O3) phases. The ordered phase fraction was
determined using a newly proposed mixed-phase model for the Rietveld refinement and a method using the relative
intensity of the superlattice line of powder X-ray diffraction. It turned out that the mass fraction of the ordered phase in
SAN calcined at 1400 °C could be controlled by the calcination time as 33% (10 h), 37% (20 h), 44% (30 h), and 48%
(50 h). Photocatalytic activities of SAN for the evolution of H2 and O2 respectively from aqueous solutions of methanol
and AgF decreased with increasing the calcination time, that is, with increasing the fraction of the ordered phase.
These results suggested that the photocatalytic activity of ordered Sr2AlNbO6 should be lower than that of disordered
SrAl0.5Nb0.5O3. This is practically the first report to reveal the photocatalytic activity of SAN as well as the effect of
cation ordering in oxides on the photocatalytic activity.
1. Introduction
in many cases, the photocatalytic activity has depended on
the bulk structure of the catalytic materials. For example, the
photocatalytic activity has decreased with increasing the
Photocatalysis over semiconductors has been widely studied
in order to apply the reaction to efficient utilization of solar
energy for water splitting, and many kinds of semiconductors
have been developed.1,2 The mechanism of heterogeneous
photocatalysis by a semiconductor is described below.3
Electron-hole pairs are initially generated in the semicon-
ductor particles using light absorption with an energy equal
to or greater than the band gap of the semiconductor. After
the separation of the photogenerated electrons and holes,
they migrate to the surface of the semiconductor, and the
redox reactions between photogenerated carriers and ad-
sorbed reactants proceed. According to the reaction mecha-
nism, the photogeneration of electrons and holes and their
migration in a semiconductor is the most significant process
in governing its photocatalytic properties. Therefore, the
bulk structure is one of the important factors for a semi-
conductor photocatalyst. So far, much research has been
performed to investigate the relationship between the bulk
structure and photocatalytic activity of semiconductors, and
amount of lattice defects in TiO2 and BaTi4O9.5 In other
4
research, the effects of the crystal systems and of the crystal
forms on the photocatalytic activity have been investigated
for AþTaO3 (Aþ=Liþ, Naþ, Kþ),6 A2þIn0.5Nb0.5O3 (A2þ
Ba2þ, Sr2þ, Ca2þ),7 TiO2,8 BaTa2O6,9 and La3B5þO7 (B5þ
=
=
Nb5þ, Ta5þ).10 On the other hand, the effect of the cation
arrangement ina semiconductoron its photocatalytic activity
has been rarely investigated. The effect of cation arrangement
on the photocatalytic activity of La3B5þO7 (B5þ=Nb5þ, Ta5þ
)
was suggested,10 but both the cation arrangement and the
crystal form changed in these materials. Therefore, the effect
of cation arrangement on photocatalytic activity has not been
clarified in a strict sense.
(4) Ikeda, S.; Sugiyama, N.; Murakami, S.; Kominami, H.; Kera, Y.;
Noguchi, H.; Uosaki, K.; Torimoto, T.; Ohtani, B. Phys. Chem. Chem. Phys.
2003, 5, 778–783.
(5) Yamashita, Y.; Yoshida, K.; Kakihana, M.; Uchida, S.; Sato, T.
Chem. Mater. 1999, 11, 61–66.
(6) Kato, H.; Kudo, A. J. Phys. Chem. B 2001, 105, 4285–4292.
(7) Yin, J.; Zou, Z.; Ye, J. J. Phys. Chem. B 2003, 107, 61–65.
(8) Torimoto, T.; Nakamura, N.; Ikeda, S.; Ohtani, B. Phys. Chem. Chem.
Phys. 2002, 4, 5910–5914.
(9) Kato, H.; Kudo, A. Chem. Phys. Lett. 1998, 295, 487–492.
(10) Abe, R.; Higashi, M.; Sayama, K.; Abe, Y.; Sugihara, H. J. Phys.
Chem. B 2006, 110, 2219–2226.
*To whom correspondence should be addressed. Tel.: þ81-92-583-7526.
Fax: þ81-92-583-8853. E-mail: teraoka.yasutake.329@m.kyushu-u.ac.jp.
(1) Osterloh, E. F. Chem. Mater. 2008, 20, 35–54.
(2) Kudo, A.; Miseki, Y. Chem. Soc. Rev. 2009, 38, 253–278.
(3) Linsebigler, L. A.; Lu, G., Jr.; Yates, T. J. Chem. Rev. 1995, 95,
735–758.
r
pubs.acs.org/IC
Published on Web 11/15/2010
2010 American Chemical Society

Article
Inorganic Chemistry, Vol. 49, No. 24, 2010 11363
Aþ4Nb6O17 (Aþ=Kþ, Rbþ),23,24 Bi2M3þNbO7 (M3þ=Al3þ
,
Ga3þ, In3þ),25 NaB5þO3 (B5þ=Nb5þ, Ta5þ),26,27 K2Ta2O6,27
AmBmO3mþ2 (m = 4, 5; A = Ca2þ, Sr2þ, La3þ; B = Nb5þ
,
Ti4þ),28 SrmTanO(mþ5n/2) (m = 1, 4, 5; n = 2, 4),29 SrTiO3,30
and BaTi4O9.31 Therefore, a double perovskite containing
one of these cations must be a good candidate material for the
present purpose of investigating the effect of cation arrange-
ment on photocatalytic activity.
In this paper, Sr-Al-Nb-O double perovskites (SAN),
in which Al3þ and Nb5þ cations occupy B sites, with various
degrees of cation-ordering have been synthesized, and their
photocatalytic properties have been investigated in relation
with cation arrangement.
Figure 1. Cation arrangements at the B site in double perovskite.
(a) Random arrangement, (b) rock-salt ordered arrangement.
Perovskite-type oxides have a general formula of ABO3.
In this structure, A and B cations are 12-fold and 6-fold
coordinated with oxygen anions, respectively. Important aspects
of perovskite-type oxide include the facts that most of the
metal cations are known to be stably incorporated into
the structure and the metal cations that occupy the A and
B site can be partially substituted by other metal cations. The
variety in material design for perovskite-type oxides results in
a variety of functional properties.11 Double perovskite, in
which two kinds of metal cationsoccupy theB site at the same
molar ratio, is well-known as a subclass of perovskite-type
oxide. It is known that two kinds of arrangements of B cations
are possible in double perovskite, random (AB0.5B00.5O3) and
rock-salt ordering (A2BB0O6), as shown in Figure 1.12 The
differences in charge and size of the B cations are important
factors for the prediction of the arrangement, and the large
differences in charge and size favor the ordered arrangement
of B cations. When the difference in charge of B cations is
equal to or larger than three, the formation of the ordered
phase has priority. In the case where the difference is two, a
partially ordered arrangement of B cations has occurred,13-18
and the degrees of cation-ordering have been changed by
calcination conditions without changing the crystal structure
in some kinds of oxides.17,18 So far, several kinds of perovskite-
type oxides including double perovskites have been reported
as photocatalysts.6,7,19-22 However, the relationship between
cation arrangement and photocatalytic activity has not been
investigated.
2. Experimental Section
2.1. Catalyst Preparation. Polycrystalline powder of Sr-Al-
Nb-O double perovskite (SAN) was synthesized using a con-
ventional solid-state reaction method. SrCO3 (purity: > 99.9%,
Kanto Chem. Co., Inc.), Al2O3 (>99.9%, Aldrich Chem. Co.),
and Nb2O5 (>99.9%, Kishida Chem. Co., Ltd.) were used as
starting materials. SrCO3 (2.955 g), Al2O3 (0.510 g), Nb2O5
(1.330 g), and ethanol (5 mL) were added into a zirconia pot
(45 mL), and the mixture was then ground by a planetary ball
mill at 500 rpm for 5 h (Fritsch, pulverizette-7). After drying the
mixture, the powder was calcined in the air under appropriate
conditions with intermediate grindings every 10 h. TiO2 powder
(P25) was supplied from the Catalysis Society of Japan as a
reference catalyst, JRC-TIO-4.
RuO2 was supported on SAN by using an impregnation method
according to a procedure reported earlier.32 SAN (1.0 g) was
suspended in tetrahydrofuran (THF, 100 mL) using magnetic
stirring. Separately, Ru3(CO)12 (purity >95%, Wako Pure Chem.
Industries, Ltd.) was dissolved in THF with a concentration of
2.8 ꢀ 10-5 mol Ru mL-1, and the solution (2.7 mL) was added
into the suspension. After evaporation to dryness at 80 °C, the
obtained powder was calcined at 400 °C for 5 h in the air to form
RuO2 supported on SAN.
2.2. Characterization of Catalyst. The crystal structure of the
catalyst was examined using powder X-ray diffraction (PXRD)
measurement using Cu KR radiation (Rigaku RINT-2200, 40 kV,
20 mA) and transmission electron microscopy (TEM; JEOL
JEM-2000EX/T, accelerating voltage: E = 200 kV). Rietveld
analysis using a PXRD pattern was performed to refine the
crystal structure. The RIETAN-2000 program33 was used for
the structural refinement, and a pseudo-Voigt peak-shape function
was used as a profile function. For the refinement of a crystal
structure, the PXRD pattern was measured at room tempera-
ture using the fixed time mode over an angular range of 10 e
2θ/° e 140 with a 0.02° (2θ) step and a counting time of 1.5-3.0 s.
Space groups reported in a previous paper12 were used for the
refinement.
Photocatalytic activities of mixed metal oxides containing
Nb5þ, Ta5þ, or Ti4þ cations have been widely investigated,
and many of the oxides have beenrevealed to be active photo-
catalysts. In addition to the oxides listed above,5-10 photo-
catalytic properties of the following oxides were reported:
(11) Pea, A. M.; Fierro, G. L. J. Chem. Rev. 2001, 101, 1981–2017.
(12) Anderson, T. M.; Greenwood, B. K.; Taylor, A. G.; Poeppelmeier,
R. K. Prog. Solid State Chem. 1993, 22, 197–233.
(13) Choy, J. H.; Hong, S. T.; Choi, K. S. J. Chem. Soc., Faraday Trans.
1996, 92, 1051–1059.
(14) Kawano, T.; Takahashi, J.; Yamada, T.; Yamane, H. J. Ceram. Soc.
Jpn. 2007, 115, 792–796.
The UV-vis diffuse reflectance spectrum (UV-vis DRS) of
the sample was measured using a Shimadzu UV-3100 equipped
(15) Faik, A.; Gateshki, M.; Igartua, M. J.; Pizarro, L. J.; Insausti, M.;
Kaindl, R.; Grzechnik, A. J. Solid State Chem. 2008, 181, 1759–1766.
(16) Mandal, K. T.; Poltavets, V. V.; Croft, M.; Greenblatt, M. J. Solid
State Chem. 2008, 181, 2325–2331.
(17) Woodward, P.; Hoffmann, R.-D.; Sleight, W. A. J. Mater. Res. 1994,
9, 2118–2127.
(18) Setter, N.; Cross, E. L. J. Mater. Sci. 1980, 15, 2478–2482.
(19) Wrighton, S. M.; Ellis, B. A.; Wolczanski, T. P.; Morse, L. D.;
Abrahamson, B. H.; Ginley, S. D. J. Am. Chem. Soc. 1976, 98, 2774–2779.
(20) Kato, H.; Kobayashi, H.; Kudo, A. J. Phys. Chem. B 2002, 106,
12441–12447.
(21) Li, D.; Zheng, J.; Zou, Z. J. Phys. Chem. Solids 2006, 67, 801–806.
(22) Hatakeyama, T.; Takeda, S.; Ishikawa, F.; Ohmura, A.; Nakayama, A.;
Yamada, Y.; Mathusima, A.; Yea, J. J. Ceram. Soc. Jpn. 2010, 118, 91–95.
(23) Kudo, A.; Tanaka, A.; Domen, K.; Maruya, K.; Aika, K.; Onishi, T.
J. Catal. 1988, 111, 67–76.
(24) Sayama, K.; Arakawa, H.; Domen, K. Catal. Today 1996, 28,
175–182.
(25) Zou, Z.; Ye, J.; Arakawa, H. Chem. Phys. Lett. 2001, 333, 57–62.
(26) Saito, K.; Kudo, A. Inorg. Chem. 2010, 49, 2017–2019.
(27) Ishihara, T.; Baik, S. N.; Ono, N.; Nishiguchi, H.; Takita, Y.
J. Photochem. Photobiol. A 2004, 167, 149–157.
(28) Kim, G. H.; Hwang, W. D.; Kim, J.; Kim, G. Y.; Lee, S. J. Chem.
Commun. 1999, 1077–1078.
(29) Yoshioka, K.; Petrykin, V.; Kakihana, M.; Kato, H.; Kudo, A.
J. Catal. 2005, 232, 102–107.
(30) Domen, K.; Kudo, A.; Onishi, T. J. Catal. 1986, 102, 92–98.
(31) Inoue, Y.; Asai, Y.; Sato, K. J. Chem. Soc., Faraday Trans. 1994, 90,
797.
(32) Ogura, S.; Sato, K.; Inoue, Y. Phys. Chem. Chem. Phys. 2000, 2,
2449–2454.
(33) Izumi, F.; Ikeda, T. Mater. Sci. Forum 2000, 198, 321–324.

11364 Inorganic Chemistry, Vol. 49, No. 24, 2010
Iwakura et al.
phases eventually disappeared, and the PXRD pattern was
ascribable to the cubic perovskite-type structure. At a calci-
nation temperature of 1500 °C, additional diffraction lines
from impurities, such as SrAl2O4, were observed, and the
amount of impurity phases depended on the calcination
procedure. When the single-phase SAN, which was obtained
at 1400 °C, was additionally calcined at 1500 °C (Figure 2e),
weak but clear diffraction lines of impurity phases were
observed. This indicates that the decomposition of SAN
perovskite proceeds at a temperature of 1500 °C. When
the solid-state mixture was directly calcined at 1500 °C
(Figure 2d), the amounts of impurity phases were larger than
those of the product after two-step calcination (Figure 2e).
According to the phase diagrams of SrO-Nb2O5,
SrO-Al2O3, and Al2O3-Nb2O5 systems,34,35 the liquid
phase appears over 1410 ( 10 °C for the Al2O3-Nb2O5
system around the stoichiometry. When the solid-state mix-
ture of the starting materials is directly heated to 1500 °C, as
in the case shown in Figure 2d, impurity phases of mixed
oxides and unreacted single oxides will be formed or present
at lower temperatures. If they survive above ca. 1400 °C, the
Al-Nb-O liquid phase will be formed. The appearance of
the liquid phase may cause volatilization of the components
and lead to the formation of a large amount of impurities. It
is noted that the calcination temperatures used in this study
were lower than the melting temperature of SAN, 1790 °C.36
The results mentioned above clearly indicate that a calcina-
tion temperature of 1400 °C is suitable for the formation
of single-phase SAN perovskite because a temperature of
1300 °C is too low to form the single phase under the present
synthesis conditions, and the decomposition of SAN
proceeds at 1500 °C.
Figure 2. PXRD patterns of Sr-Al-Nb-O double perovskites calcined
at various temperatures. Starting solid-state mixture was directly calcined
for 10 h at (a) 1200 °C, (b) 1300 °C, (c) 1400 °C, and (d) 1500 °C. (e) After
calcination of sample c at 1500 °C for 10 h.
with an integrating sphere attachment. R-Al2O3 powder was used as
a reference sample. The mixture of SAN and Al2O3 (10 wt %
SAN/Al2O3) powders was used to obtain a reflectance spectrum.
A reflectance spectrum was collected over wavelengths from 800
to 220 nm with a 0.5 nm step. The reflectance was transformed
into the Kubelka-Munk function. Photon energy at the absorp-
tion edge was estimated by extrapolating the linearly ascending
portion of the spectrum to the x axis, and the value was defined
as an optical band gap.
The specific surface areas of prepared samples and TiO2 were
determined using the BET method respectively from krypton
and nitrogen adsorption isotherms measured at 77 K (BEL JAPAN,
Inc., BELSORP 18PLUS-SP, BELSORP-mini).
It has been reported that a cubic ordered double
perovskite with rock-salt ordering of B cations has a
superlattice structure with unit cell dimensions of 2ap ꢀ
2apꢀ2ap: ap is the lattice constant of cubic perovskite of a
primitive ABO3 type.12 According to indexing based on
the superlattice structure, the appearance of 111, 311, 331,
and 511 diffraction lines, namely, superlattice diffraction
lines, is evidence of the rock-salt ordering of B cations.12
As shown in Figure 2c, all of the diffraction lines were
indexed on the basis of the superlattice structure, indicat-
ing that SAN prepared at 1400 °C for 10 h contained the
ordered phase with an alternate arrangement of Al3þ
2.3. Photocatalytic Activity Measurement. The photocatalytic
activity of the catalyst was examined using a gas-closed circulation
system equipped with a GC-TCD (Shimadzu, GC-8A). Powdered
1wt%RuO2/SAN (0.5 g), SAN (0.5 g), or TiO2 (0.5 g) was dispersed
into a 10 vol % methanol aqueous solution (500 mL) in an inner-
irradiation-type reaction cell made of Pyrex glass. The reaction cell
was connected to the gas-closed circulation system. After deaera-
tion of the dispersion with stirring using a magnetic stirrer, Ar gas
was introduced into the reactor (1.3ꢀ104 Pa). Then, the dispersion
was irradiated with a 100 W high-pressure mercury lamp, and the
amount of H2 evolved was measured by the GC-TCD. According
to the same procedure, the amount of O2 evolved from a 0.01 M
AgF aqueous solution (500 mL) was measured using the powdered
catalysts (0.3 g) in a He atmosphere (1.3 ꢀ 104 Pa).
and Nb5þ
.
The effect of the calcination time was investigated for
the synthesis of SAN at 1400 °C (Figure 3). Although all
of the samples calcined for 10 to 50 h were single-phase
cubic perovskites, relative intensities (I/I0) of the super-
lattice diffraction lines tended to increase with increasing
calcination time. The superlattice diffraction lines arise
from the ordered arrangement of B cations as stated
above, and the strongest line with an intensity of I0 is
contributed by both the 220 and 110 diffractions of the
ordered and disordered phases, respectively. This means
that the value of I/I0 of the superlattice diffraction lines
includes information about the degree of cation-ordering
at the B site. Therefore, the results in Figure 3 showed that
3. Results and Discussion
3.1. Cation Arrangement at the B Site in Sr-Al-Nb-O
Double Perovskite. Figure 2 shows PXRD patterns of SAN
perovskites prepared using the solid-state reaction with
various calcination temperatures for 10 h. After calcination
at 1200 and 1300 °C, Sr5Nb4O15 was clearly observed as an
impurity phase in addition to the perovskite phase, and the
intensity of the diffraction lines from Sr5Nb4O15 decreased
with increasing calcination temperature. Considering the
stoichiometry, impurity phases other than Sr5Nb4O15 might
be present below 1300 °C, though they were not detected
by PXRD. At 1400 °C, diffraction lines from any impurity
(34) Levin, M. E.; McMurdie, F. H. Phase Diagrams for Ceramists Vol. 3;
The American Ceramic Society: Westerville, OH, 1975.
(35) Roth, S. R.; Dennis, R. J.; McMurdie, F. H. Phase Diagrams for
Ceramists Vol. 5; The American Ceramic Society: Westerville, OH, 1987.
(36) Brandle, D. C.; Fratello, J. V. J. Mater. Res. 1990, 5, 2160.

Article
Inorganic Chemistry, Vol. 49, No. 24, 2010 11365
Rietveld analysis was performed to determine the
degree of cation-ordering in SAN. So far, the degree of
cation-ordering in A2BB0O6 has been determined by
Rietveld analysis using the structural model A2(BxB01-x)-
(B0xB1-x)O6 (0.5 e x e 1.0), where x means the occu-
pancy of B and B0 cations at their original sites. The
degree of cation-ordering has been defined by eq 1 using a
refined x value;13-17 x = 1.0 and 0.5 correspond respec-
tively to completely ordered and disordered phases.
The degree of cation- ordering at the B site
¼ 2 ꢀ j100x - 50j=%
ð1Þ
The calculation model was based on the single-phase
A2(BxB01-x)(B0xB1-x)O6 and set one value of full width
at half-maximum (fwhm) for all of the diffraction lines.
TEM observation clearly indicated that the SAN pre-
pared in this study was a mixture of Sr2AlNbO6 (space
group: Fm3m) and SrAl0.5Nb0.5O3 (space group: Pm3m).
In addition, fwhm values of superlattice lines were differ-
ent from those of the other lines, as shown in Figure 5
for the 111 and 220 diffraction lines of SAN prepared at
1400 °C for 30 h. The larger fwhm values of the super-
lattice 111 diffraction mean that the crystallite (domain)
size of the Sr2AlNbO6 ordered phase is smaller than that
of SrAl0.5Nb0.5O3, and this is consistent with the TEM
observation. The above discussion clearly suggests that
not the conventional single-phase model but the new
mixed-phase model should be applied for Rietveld anal-
ysis in the present case. The refinement with the mixed-
phase model is done for the mixture of y(Sr2AlNbO6) and
(1-y)(SrAl0.5Nb0.5O3), where y is a mass fraction of the
ordered phase and therefore corresponds to the degree of
ordering in SAN. In the mixed-phase model, different
fwhm values can be set for different phases, and cell
parameters for each phase can be refined.
The representative results of Rietveld analysis by the
mixed-phase model are shown in Figure 6 and Table 1 for
SAN calcined for 30 h at 1400 °C. The results showed that
Sr2AlNbO6 and SrAl0.5Nb0.5O3 phases both had a cubic
unit cell, and the lattice constant of the ordered phase was
nearly double that of the disordered phase, as expected. In
order to confirm the validity of the mixed-phase model,
the Rietveld refinement using the conventional single-
phase model was performed, and the result was compared
with that using the mixed-phase model. The refinement
using the conventional model was possible with a good-
ness-of-fit indicator (S) of 2.2. The S value in the mixed-
phase model refinement was 1.9 (Table 1), and therefore
the new model was more valid than the conventional
model. In the present study, the new model was used to
refine the crystal structure of SAN.
Figure 3. PXRD patterns of Sr-Al-Nb-O double perovskites calcined
at various calcination times at 1400 °C. (a) 10 h, (b) 20 h, (c) 30 h, (d) 40 h,
(e) 50 h, (f) the enlarged 111 superlattice diffraction lines of a-e.
the degree of cation-ordering at the B site increased with
increasing calcination time.
TEM observations were performed to investigate the
structure of SAN in detail. Figure 4 shows the TEM
images of SAN calcined at 1400 °C for 30 h. The selected
area of the particle in Figure 4a gave electron diffraction
(ED) spots (Figure 4b), and therefore the single crystal
was formed in the observed area. The distances of the
crystal plane (d{hkl}) calculated from the ED pattern
(d{111} = 0.46 nm, d{200} = 0.40 nm, d{113} = 0.25 nm,
d{220} = 0.29 nm) agreed with those of the crystal plane
calculated from the PXRD pattern (d{111} = 0.450 nm,
d{200} =0.389 nm, d{113} =0.234 nm, d{220} =0.275 nm),
indicating that the particle was made of a single-phase
perovskite-type oxide. The appearance of 111 and 113
superlattice ED spots proved the existence of the ordered
phase in the particle, as in the PXRD results. Dark field
observation is a useful method for determining the do-
main which gives an ED spot. For the B-site ordered
double perovskite, dark field observations using the 111
superlattice diffraction spot were performed to distin-
guish the domain of the completely B-site ordered
structure.37 The dark field image of the particle is shown
in Figure 4c, and the observation clarified that the
completely B-site ordered domain existed in the near-
surface region of the particle (bright area in Figure 4c).
On the basis of PXRD measurements and TEM observa-
tions, it has been confirmed that the SAN crystallite
consisted of two domains of completely B-site ordered
Sr2AlNbO6 and disordered SrAl0.5Nb0.5O3, and the
Sr2AlNbO6 domain in a crystallite expands with increas-
ing calcination time. Randall et al. have reported a similar
trend of expansion of the ordered-phase domain for the
Pb2ScTaO6 single crystal.37
Figure 7 shows the dependence of the fraction of
Sr2AlNbO6 on calcination time at 1400 °C. With increas-
ing the calcination time, the fraction increased linearly to
44% up to 30 h and then leveled off to reach 48% after
50 h, indicating that the growth of the ordered phase is
almost completed after calcination for 30 h under the
present synthesis conditions. Woodward et al. reported
that the degree of cation ordering depended on the
calcination temperature, and it was 94% for SAN synthe-
sized by a flux method with calcination at 1400 °C.17 The
difference in the fraction of the ordered phase might arise
(37) Randall, A. C.; Barber, J. D.; Whatmore, W. R.; Groves, P. J. Mater.
Sci. 1986, 21, 4456–4462.

11366 Inorganic Chemistry, Vol. 49, No. 24, 2010
Iwakura et al.
Figure 4. TEM images of Sr-Al-Nb-O double perovskite calcined for 30 h at 1400 °C. (a) Bright field image, (b) electron diffraction pattern, (c) dark
field image using 111 diffraction.
Table 1. Result of Rietveld Analysis for Sr-Al-Nb-O Double Perovskite
Calcined for 30 h at 1400 °Ca
SrAl0.5Nb0.5O3 (SG: Pm3m)
a = 0.389575(5) nm
2 b
˚
B (A )
site
g
x
y
z
O
3d
1b
1a
1a
1
1
1
1
1/2
1/2
0
0
1/2
0
0
1/2
0
0.83(7)
0.73(2)
0.40(2)
0.40(2)
Sr
Al
Nb
0
0
0
RF = 1.24%, RI = 1.78%, mass fraction = 0.5553
Sr2AlNbO6 (SG: Fm3m)
a = 0.77955(17) nm
2 b
˚
B (A )
Figure 5. Dependence on calcination time at 1400 °C of fwhm of (111)
and (220) diffraction lines from Sr-Al-Nb-O double perovskites. Open
circle, 111 diffraction line; filled circle, 220 diffraction line.
site
g
x
y
z
O
24e
8c
4b
4a
1
1
1
1
0.2489(5)
0
0
0.83
0.73
0.40
0.40
Sr
Al
Nb
1/4
1/2
0
1/4
1/2
0
1/4
1/2
0
RF = 3.34%, RI = 5.14%, mass fraction = 0.4447
a Rwp = 12.10%, Rp = 7.75%, RR = 10.29%, Re = 6.34%, S = 1.9.
b B values were refined using Rietveld analysis for SrAl0.5Nb0.5O3, and
the obtained values were used for Sr2AlNbO6.
Figure 6. Result of Rietveld analysis for Sr-Al-Nb-O double perov-
skite calcined for 30 h at 1400 °C.
from the different preparation methods used. Two B-site
cations of Al3þ and Nb5þ ions in the ordered SAN phase
are arranged alternately, and therefore it is reasonably
accepted that a preparation method with more homo-
geneous mixing of Al and Nb precursors gives the ordered
phase more easily. The flux method is well-known to
facilitate the formation of mixed metal oxides. In the
present solid state reaction method, Al2O3, Nb2O5, and
SrCO3 were used as starting materials. The diffusion of
cations in the solid is required to form the mixed metal
oxide phase, and the presence of Al and Nb in the 1:1
ration is indispensable in forming the ordered phase. The
Figure 7. Relationship between the calcination time and the mass frac-
tion of the ordered Sr2AlNbO6 in Sr-Al-Nb-O double perovskites
refined by Rietveld analysis. Sr-Al-Nb-O perovskites were calcined at
1400 °C.
calcination time dependence of the fraction (Figure 7)
implies that the disordered phase is first formed
(kinetically favored) and that it is gradually transformed

Article
Inorganic Chemistry, Vol. 49, No. 24, 2010 11367
reported that the optical band gap of perovskite-type
oxide is mainly determined by the kind of B cation and the
-B-O-B- bond angle.9,38 The effects of these factors,
however, are not so important for the optical band gap of
SAN because Sr2AlNbO6 and SrAl0.5Nb0.5O3 include
Al3þ and Nb5þ cations at the same molar ratio, and the
-B-O-B- angles in the both structures are 180°. Re-
garding the relationship between cation arrangement and
band structure of the perovskite-type oxide, Eng et al.
have investigated the effect of the isolation of the TaO6
octahedron on the optical band gap by comparing KTaO3
and Sr2AlTaO6. The band gap of Sr2AlTaO6 was larger
than that of KTaO3, and they suggested that the isolation
of the TaO6 octahedron by the AlO6 octahedron is the
origin of the increase of the optical band gap.38 To the
best of our knowledge, however, the relationship between
the arrangement of two B cations (Al and Nb in the
present SAN) and the optical band gap has not been
investigated. Considering the cation isolation discussion,
ordered Sr2AlNbO6 with a complete isolation of the
NbO6 octahedron by the AlO6 octahedron might have a
larger band gap than disordered SrAl0.5Nb0.5O3, in which
the linkage of the NbO6 octahedra is present. This
inferred difference of optical band gap might be true if
one considers that disordered SrAl0.5Nb0.5O3, which is
the major phase in SAN perovskites, dominates the
absorption spectrum at longer wavelengths. The relation-
ship between the cation arrangement and the band gap
should be determined experimentally by using, for exam-
ple, a completely ordered material or theoretically by the
calculation of the band structure.
3.3. Photocatalytic Property. The results of the photo-
catalytic reactions performed in this study are summar-
ized in Table 2. The photocatalytic H2 and O2 evolutions
from an aqueous methanol solution and an aqueous AgF
solution, respectively, were observed with SAN perov-
skites with and without a RuO2 cocatalyst, and no change
of the crystal structure of SAN perovskites was noticed
from PXRD measurements of fresh and used catalysts.
These results indicate that SAN works as a photocatalyst,
and this is the first report to reveal the photocatalytic
activity of the SAN perovskite.
Time courses of photocatalytic H2 evolution from a
10 vol % methanol aqueous solution over SAN and 1 wt
% RuO2/SAN are shown in Figure 10. The amount of H2
evolution increased with increasing reaction time over
SAN perovskites both with and without the cocatalyst.
The amount of H2 evolved up to 8 h decreased with an
increase in the calcination time, and this trend was held
after normalization via specific surface area (Sa;
Table 2a). In most of the powdered photocatalysts, a
cocatalyst has been used to improve the photocatalytic
activity.2 In this study, RuO2, which has been reported to
show the functionality of a cocatalyst,39,40 was supported
on SAN. As shown in Figure 10, the amount of photo-
catalytic H2 evolution was greatly increased over 1 wt %
RuO2/SAN as compared with the bare SAN, confirming
Figure 8. Relationship between the mass fraction of the ordered
Sr2AlNbO6 phase and the relative intensity of the 111 diffraction line
for Sr-Al-Nb-O double perovskites prepared at 1400 °C for 10 h (open
circle), 20 h (open triangle), and 30 h (open square). Solid line, theoretical
line based on the mixed-phase model (see text for detail).
to the thermodynamically favored ordered phase until reach-
ing the equilibrium value under the conditions given.
As stated above, the I/I0 value of the superlattice
diffraction lines can be used to determine the fraction of
the ordered phase. The mass fraction of Sr2AlNbO6 in the
mixed-phase model principally has a proportional rela-
tion to a value of I/I0. This relation is show in Figure 8 for
the 111 diffraction line: the I(111)/I0 value at the 100%
ordering (15.8%) was calculated on the basis of the crystal
data of Sr2AlNbO6 and drew a straight line passing
through the origin and the calculated value of the 100%
ordering. The experimental results of SAN perovskites
calcined at 1400 °C are also plotted in Figure 8, and they
were practically on the theoretical line, indicating that the
observed I(111)/I0 value can be directly transformed to
the fraction of Sr2AlNbO6 by using the relation. SAN
perovskites were synthesized at 1400 °C several times with
the same method and under the same conditions. As
shown in Table 2, the fractions of the ordered phase in
prepared SAN perovskites were 30% according to the
Rietveld refinement and 31 and 30% according to the
estimation from Figure 8 after a calcination time of 10 h,
37% (refinement) and 36% (estimation) after 20 h of
calcination, and 44% (refinement) and 42 and 47%
(estimation) after 30 h of calcination. These reasonably
close values for each case clearly indicate the good
reproducibility of the preparation as well as the validity
of the estimation from the linear relationship between the
relative intensity of the superlattice diffraction and the
fraction of the ordered phase.
3.2. Optical Absorption Property. Figure 9 shows the
UV-vis DRS of SAN perovskites with various frac-
tions of Sr2AlNbO6. The observed photon energy at the
absorption edges, namely optical band gap, for SAN
perovskites was 4.1 eV for all of the samples. The identical
optical band gap of SAN perovskites with different
ordered phase fractions does not necessarily mean that
ordered Sr2AlNbO6 and disordered SrAl0.5Nb0.5O3 have
the same optical band gap. Since SAN perovskites pre-
pared in this study are a mixture of ordered and dis-
ordered phases, the absorption spectrum at longer wave-
lengths and the estimated band gap from the portion are
governed by the phase with smaller band gap. It has been
(38) Eng, W. H.; Barnes, W. P.; Auer, M. B.; Woodward, M. P. J. Solid
State Chem. 2003, 175, 94–109.
(39) Sakata, T.; Hashimoto, K.; Kawai, T. J. Phys. Chem. 1984, 88, 5214–
5221.
(40) Inoue, Y.; Kubokawa, T.; Sato, K. J. Phys. Chem. 1991, 95, 4059–
4063.

11368 Inorganic Chemistry, Vol. 49, No. 24, 2010
Iwakura et al.
Table 2. Results of the Photocatalytic Reactions over Sr-Al-Nb-O Double Perovskite
(A) H2 evolution from an aqueous methanol solution over catalysts (0.5 g)
H2 evolutiona
catalyst
Sr-Al-Nb-O
calcination time/h
I(111)/I0b/%
Fopc/%
Sa/m2 g-1
μmol
μmol (m2 cat)-1
10
20
30
10
20
30
5.1
5.7
7.5
5.1
5.9
7.0
30d
37d
44d
31e
36e
42e
3.1
3.0
2.9
3.2
2.9
2.7
55.3
6.0
2.5
1.6
26.7
19.3
16.6
24.9
3.9
1.7
1.1
16.8
13.4
12.4
0.9
1 wt % RuO2/ Sr-Al-Nb-O
TiO2
(B) O2 evolution from an aqueous AgF solution over catalysts (0.3 g)
O2 evolutiona
catalyst
calcination time/h
I(111)/I0b/%
Fopc/%
Sa/m2 g-1
μmol
μmol (m2 cat)-1
Sr-Al-Nb-O
10
20
30
10
20
30
4.9
5.7
7.8
5.1
5.9
7.0
30e
37d
47e
31e
36e
42 e
3.4
2.9
3.6
3.2
2.9
2.7
55.3
20.5
12.0
9.1
21.6
12.0
9.0
20.2
13.8
8.5
22.7
13.9
11.1
69.5
1 wt % RuO2/ Sr-Al-Nb-O
TiO2
1153.8
a Gross and surface-area normalized amount of gas evolved after 8 h of photoirradiation. b Relative intensity of the 111 diffraction line. c Mass fraction
of the ordered Sr2AlNbO6 phase. d Refined by Rietveld analysis. e Estimated from the I(111)/I0 - Fop relation (Figure 8).
Figure 10. Time course of the H2 evolution from a 10 vol % methanol
Figure 9. Absorption spectra of Sr-Al-Nb-O double perovskites
aqueous solution over Sr-Al-Nb-O double perovskites calcined at
1400 °C. Calcination time: 10 h (triangle), 20 h (circle), 30 h (square).
Open symbols, bare SAN; filled symbols, 1 wt % RuO2/SAN.
calcined for 10, 20, and 30 h at 1400 °C.
the cocatalytic function of RuO2 on the SAN photocata-
lyst. As in the case of bare SAN perovskites, the amount
of H2 evolved over 1 wt % RuO2/SAN also decreased
when the calcination time of SAN was prolonged. The
agreement of the activity order over the catalysts with and
without RuO2 implies that the dependence of the activity
on the calcination time is governed mostly by the bulk
property.
A photocatalytic reaction with an aqueous AgF solu-
tion was also performed. The evolution of O2 was ob-
served (Table 2b), and the formation of Ag metal was
confirmed by PXRD after the reaction. The amount of O2
evolved increased in proportion to the photoirradiation
time, up to 8 h over SAN and 1 wt % RuO2/SAN.
However, the cocatalytic function of RuO2 was negligible
for the reaction. The amounts of O2 evolved on both gross
and Sa-normalized bases decreased with increasing calci-
nation time. Since the same dependence was observed in
both H2 and O2 evolution reactions, the dependence of the
activity on the calcination time originates from the bulk
property of SAN irrespective of the kind of reactions.
As described in section 3.1, the mass fraction of ordered
Sr2AlNbO6 in SAN increased with increasing calcination
time. The amounts of gases evolved up to 8 h of photo-
irradiation are plotted as a function of the mass fraction
of ordered Sr2AlNbO6 in Figure 11. The photocatalytic
activity of SAN decreased with an increase in the fraction
of Sr2AlNbO6, indicating that the photocatalytic activity
of disordered SrAl0.5Nb0.5O3 should be higher than that
of ordered Sr2AlNbO6. The relation was not straight, and
the decrease in the lower fraction range seemed to be
larger than that at the higher fraction range. This might be
reasonable from the TEM observation that the less active
ordered phase is formed from the near-surface region of
particles of the active disordered phase.

Article
Inorganic Chemistry, Vol. 49, No. 24, 2010 11369
ordered phase, although the reason for that will be clari-
fied in future studies.
4. Conclusion
In this study, Sr-Al-Nb-O double perovskite (SAN)
was synthesized using the conventional solid-state reaction
method, and its structure and photocatalytic properties were
investigated from a viewpoint of cation-ordering at the
octahedral B site. TEM observations confirmed that SAN
particles have a domain structure with disordered SrAl0.5N-
b0.5O3 and ordered Sr2AlNbO6 phases and that the latter
domain was preferentially present at the surface region of the
particles of the former phase. The fraction of the ordered
phase was estimated from the newly proposed mixed-phase
model for Rietveld analysis and the relative intensity of the
superlattice PXRD line. The photocatalytic activity of SAN
for both H2 evolution from an aqueous methanol solution
and O2 evolution from an aqueous AgF solution were found
to increase with the decreasing fraction of the ordered phase.
The present study revealed the effect of the cation ordering on
the photocatalytic activity of mixed metal oxides, and the
origin of the relation between the cation arrangement and
catalytic activity will be made clear in future works.
Figure 11. The relationship between the mass fraction of Sr2AlNbO6 in
Sr-Al-Nb-O double perovskite and the amounts of the gas evolved
after photoirradiation for 8 h. A total of 0.5 g of SAN was used for the
photocatalytic H2 evolution from an aqueous methanol solution, and
0.3 g of SAN was used for the photocatalytic O2 evolution from an
aqueous AgF solution.
In order to evaluate the activity level of the newly
developed catalyst, the photocatalytic activity of bare
SAN was compared with that of TiO2. If the surface
area-normalized activity of the most active SAN after
calcination for 10 h was used, the activity of SAN was
4.3 times higher and 3.4 times lower than TiO2 for H2 and
O2 evolution, respectively. Roughly speaking, it can be
said that the activity of the newly developed SAN is on the
same level as TiO2.
The present study revealed, for the first time, the effect
of the cation ordering on the catalytic activity with the
catalysts of the same constituent metal cations. The dis-
ordered phase was inferred to be more active than the
Acknowledgment. This work was supported by a Grant-
in-Aid for Scientific Research on Priority Areas (Nos.
19028050 and 20037053, “Chemistry of Concerto Cat-
alysis”) from the Ministry of Education, Culture, Sports,
Science and Technology, Japan and a grant from the
Global-Centre of Excellence in Novel Carbon Resource
Sciences, Kyushu University. The authors thank Dr. K.
Yoshida for valuable discussions.