Dehydrogenation of ethylbenzene over zirconium-based perovskite-type catalysts of AZrO3 (A: Ca, Sr, Ba)

Article abstract of DOI:10.1016/j.apcata.2014.06.012

The aim of this work was to investigate the catalytic performance of the AZrO₃ (A: Ca, Sr or Ba) catalysts for the dehydrogenation of ethylbenzene (EBDH) to produce styrene and to clarify an important factor for the high dehydrogenation activit

Full text of DOI:10.1016/j.apcata.2014.06.012

Applied Catalysis A: General 482 (2014) 344–351
Contents lists available at ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
Dehydrogenation of ethylbenzene over zirconium-based
perovskite-type catalysts of AZrO (A: Ca, Sr, Ba)
3
a
b
a,∗
Ryo Watanabe , Yoshinori Saito , Choji Fukuhara
a
Department of Applied Chemistry and Biochemical Engineering, Graduate school of Engineering, Shizuoka University, 3-5-1, Johoku, Naka-ku,
Hamamatsu, Shizuoka, Japan
Murata Manufacturing Co., Ltd., Nagaokakyo-shi, Japan
b
a r t i c l e i n f o
a b s t r a c t
Article history:
The aim of this work was to investigate the catalytic performance of the AZrO3 (A: Ca, Sr or Ba) cata-
lysts for the dehydrogenation of ethylbenzene (EBDH) to produce styrene and to clarify an important
factor for the high dehydrogenation activity. Among the AZrO3 catalysts, only the BaZrO3 (BZO) catalyst
showed a significantly high activity for EBDH and the activity increased with time, while the CaZrO3 and
SrZrO3 catalysts almost did not provide any activity at 823 K. Comparing the styrene yield over the BZO
catalyst with that over the industrial potassium-promoted iron oxide (Fe–K) catalyst, the BZO catalyst
showed a lower styrene yield than the Fe–K catalyst at the initial stage of the reaction. However, after
Received 10 December 2013
Received in revised form 7 June 2014
Accepted 11 June 2014
Available online 19 June 2014
Keywords:
Dehydrogenation of ethylbenzene
Perovskite oxide catalyst
BaZrO3
Oxygen vacancy
ESR
4
0 min of EBDH, the BZO catalyst exhibited a higher styrene yield than the Fe–K catalyst. Based on an ESR
measurement, a sharp signal at g = 2.004, which was identified as an unpaired electron trapped in oxy-
gen vacancies, was detected in the BZO catalyst after dehydrogenation. The number of oxygen vacancies
increased with change in the dehydrogenation activity. In addition, the BZO catalyst with a pretreat-
ment by H2 reduction presented a high activity without an induction period. Comparing the profiles of
the temperature desorption of ethylbenzene (EB) over the prereduced catalyst to that of the untreated
BZO catalyst, a chemisorbed species of EB was detected over the prereduced BZO catalyst, although a
physisorbed species was present on the surface of the untreated catalyst. Hence, the production of oxy-
gen vacancies opened the adsorption channel of EB and created the reactive site, which produced a high
EBDH activity.
©
2014 Elsevier B.V. All rights reserved.
1
. Introduction
stability than transition metal oxides and their catalytic activities
are sometimes competitive with supported noble metal catalysts
[13].
Zirconium (Zr)-based materials, such as a yttria-stabilized ZrO2
and a CexZr1 O , have been applied to a variety of applications in
In a preliminary study, the catalytic activity for the dehydro-
genation of ethylbenzene (EBDH) to produce styrene (Eq. 1) has
been examined over Mn- and Fe-based perovskite-type oxides [14].
−x
2
ceramics, solid oxide fuel cells, oxygen sensors, and catalysts [1–8].
The success of Zr-based materials for widespread applications is
mainly derived from its oxygen mobility coupled with the redox
ability of these oxides. Despite its diverse applications, the investi-
gation of the catalysis of the Zr-based perovskite materials has not
been performed until recently. Perovskite-type mixed oxides (gen-
C8H → C H + H
(1)
10
8
8
2
As a result, a La0.8Ba0.2Fe0.4Mn0.6
O
(denoted as LBFMO) cata-
3−ı
lyst had a high performance and revealed an activity superior to the
industrial Fe–K catalyst at the low temperature of 813 K. In addi-
tion, the catalytic activity and stability were found to depend on
the redox characteristics of the perovskite material [15,16].
The La element, which was contained in the LBFMO catalyst,
is an expensive rare-earth metal and unevenly distributed around
the world. The price of such a rare earth metal element (denoted
as REE) has gradually increased since 1995. By 2010, the REE was
traded at a very high price; La oxide was 6.2–6.5 US$/kg, Ce oxide
was 4.7–5.2 US$/kg, Pr oxide was 30–30.5 US$/kg and Nd oxide
eral formula ABO , A is a rare-earth or an alkali-earth element and
3
B is typically a transition metal) have a high lattice oxygen mobility
in comparison to common transition metal oxides, and in addition,
the redox ability could be controlled by changing the incorpo-
rated element [9–12]. In addition, they have a much better thermal
∗ Corresponding author. Tel.: +81 53 478 1171; fax: +81 53 478 1172.
E-mail address: tcfukuh@ipc.shizuoka.ac.jp (C. Fukuhara).
http://dx.doi.org/10.1016/j.apcata.2014.06.012
0
926-860X/© 2014 Elsevier B.V. All rights reserved.

R. Watanabe et al. / Applied Catalysis A: General 482 (2014) 344–351
345
was 30.5–31 US$/kg [17]. Therefore, the lanthanoid element as the
[Sty]
[Sty] + [Bz] + [Tol]
Styrne yield =
× EB conversion
× 100
(3)
A-site cation cannot be included in the catalyst for industrial appli-
cations due to its cost. However, the price of a divalent metal, such
as Ba, was low compared to the rare earth metal. Barite, which is a
raw material of Ba, is trading at 30 US$ per ton. The advantage is not
only its availability, but also the adequate catalytic performance of
perovskite catalysts containing a divalent cation. The incorporation
[
Sty]
Styrne selectivity =
(4)
[
Sty] + [Bz] + [Tol]
[EB], [Sty], [Bz], and [Tol] denote the concentration of EB, styrene,
benzene and toluene in the effluent gas, respectively. The carbon
balances in this study were over 95% for all the experimentally
obtained results.
2+
2+
2+
of an alkaline earth metal, such as Ca , Sr and Ba , into the A-
site of the perovskite structure would generate electron holes and
oxygen vacancies as the charge compensation. Such an incorpo-
ration of the divalent cation could induce a high oxygen mobility
derived from the mixed conduction by an electron and oxygen ion
2
.3. Characterization of the catalyst
[
18–20]. Additionally, the incorporation of elements with a large
The crystalline structure of the prepared catalyst was ascer-
ionic radii, such as Ba, into the A-site of the perovskite structure
produces a high free volume in the lattice, which decreases the
activation energy of the oxygen ion migration [21,22]. These effects
promoted the releasing rate of the lattice oxygen, thus expecting a
high performance for the EBDH.
As already mentioned, the A-site substitution by the divalent
cation could provide a high redox flexibility of the perovskite-
type oxide. Thus in the present study, the catalytic performances
of the Zr-based perovskite catalysts containing a divalent cation,
such as Ca, Sr and Ba, were investigated, and compared to that
over the industrial potassium-promoted iron oxide catalyst (Nis-
san Girdler Catalyst; G-84C). In addition, an important factor for
the dehydrogenation activity over the BaZrO3 perovskite catalyst
was investigated in terms of the structural state measured by X-
ray diffraction, the catalyst morphology observed by field emission
scanning electron microscopy and its oxidation state measured by
electron spin resonance.
tained using an X-ray powder diffraction with CuK␣ radiation
(
ꢀ = 1.54 A, Rint-2000; Rigaku Co. Ltd., Japan). The specific surface
˚
area of the zirconium-based perovskite oxide was analyzed by the
N2 absorption method (Model 4200; Nikkiso Co. Ltd., Japan). The
sample was outgassed at 423 K for 30 min before absorbing N2.
X-ray photoelectron spectroscopy (XPS, Quantum2000; Physi-
cal Electronics Co. Ltd., USA) measurements were performed using
non-monochromatic AlK␣ radiation. The pass energy of the ana-
lyzer was 23.5 eV. The binding energy of C1s at 284.7 eV was used
for calibration.
In order to study the morphology of the BZO catalyst, a field
emission scanning electron microscopy (FE-SEM) measurement
was performed using a SU8040 (Hitachi High-Technologies Cor-
poration, Japan). The ESR spectra were obtained by an EMX
spectrometer (Bruker BioSpin Corp., USA) at the field modulation
of 100 kHz, an amplitude modulation of 0.8 mT and a microwave
power of 0.02 mW. The measurement was performed at room tem-
perature and in ambient air without vacuum-pumping. A sharp
signal at g = 2.004, which was identified as one electron trapped
in Vox, was detected, and the number of Vox was calculated from
the sharp signal using the following equation.
2
. Experimental
2.1. Catalyst preparation
Ssample
Nel,sample = Nel,standard × S
(5)
(6)
The zirconium-based perovskite oxides of AZrO3 (A: Ca, Sr,
standard
Ba) were prepared by the solid state reactions of ZrO2 and an
alkali carbonate. These powders were thoroughly mixed in an agi-
tate mortar, pelletized and then calcined at 1373 K for 12 h. The
industrial potassium-promoted iron oxide (G-84C, made by Nissan
Girdler K.K.; composition = 77 wt%-Fe O , 10 wt%-K O, 5.0 wt%-
Nvox = Nel,sample
Nel, sample and Nel, standard are the number of electron spins in the
catalyst and in the standard sample of CuSO4·5H2O, respectively.
The number of electron spins which is included in the standard
2
3
2
21
Ce O , 2.5 wt%-MoO , 2.2 wt%-CaO, 2.2 wt%-MgO, and less than
sample of 1 g is 2.41 × 10 . Ssample and Sstandard are the signal areas
of the catalyst and the standard sample, respectively. NVox is the
number of oxygen vacancies with an unpaired electron in the cat-
alyst.
2
3
3
0
.1 wt%-of Cr O ) catalyst was used as the reference catalyst.
2 3
2
.2. Activity test
Temperature-programmed desorption (TPD) was performed in
order to investigate the surface active site, and adsorption and
desorption properties with increasing temperature. Before the TPD
measurement, EB (4.6 kPa) was supplied to the catalyst (0.50 g) at
323 K for 30 min as a pretreatment. After the EB adsorption on
the catalyst, EB in gas phase was completely purged by helium
(He) for 2 h. The catalyst was then heated in He (100 ml min ) at
10 K min–1 from 323 to 1073 K. The effluent gas was monitored by
an on-line Omnistar GSD301 quadrupole mass spectrometer (Pfeif-
fer Vacuum, Germany). The parent peak of EB was scanned by the
mass spectrometer: m/z; 91 (EB).
The catalytic activity, selectivity and stability of the prepared
catalyst for the dehydrogenation of ethylbenzene (EBDH) were
examined using a conventional fixed bed reactor. The reactor used
in this study consisted of a quartz tube (10-mm o.d.) containing
a catalyst bed, which was fixed by quartz wool. A type-K ther-
mocouple for controlling the temperature was positioned outside
the quartz tube. The reactions were conducted at 823 K under
atmospheric pressure. The weight hourly space velocity (WHSV,
EB based) was 1.2–24 h–1 and the EB supplied to the catalyst bed
was diluted by helium (PEB: 6.3 kPa). The catalyst weight was
–1
0
.005–1.0 g. Liquid products such as EB, benzene, toluene, and
3
. Results and discussion
styrene were analyzed using an off-line thermal conductivity detec-
tion (TCD) gas chromatograph (GC-8A; Shimadzu Co. Ltd., Japan).
The conversion (Eq. 2), styrene yield (Eq. 3) and styrene selectivity
3.1. Characterization of the AZrO (A: Ca, Sr, Ba) catalysts
3
(
Eq. 4) were determined using the following equations.
Fig. 1 shows the structural states of the AZrO3 (A: Ca, Sr, Ba)
and ZrO2 catalysts measured by XRD. Hereafter, the CaZrO3 cata-
lyst was abbreviated as CZO, the SrZrO3 catalyst as SZO, and the
BaZrO3 as BZO. The SZO and BZO catalysts showed the typical
[
Sty] + [Bz] + [Tol]
EB conversion = [
× 100
(2)
EB] + [Sty] + [Bz] + [Tol]

3
46
R. Watanabe et al. / Applied Catalysis A: General 482 (2014) 344–351
70
P
Equilibrium
CaZrO
SrZrO
BaZrO
ZrO
Fe-K
3
BaZrO
3
60
50
3
P
P
P
3
P
Sty sel.
97.4%
P
P
2
P
SrZrO
3
P
P
4
3
0
0
P
P
P
P
P
CaZrO
3
P P
m
P
P
P
P
P
PP
20
m
1
0
m
ZrO
2
m
mm
50
m
m
m
mm
m
m
0
0
20
40
60
80
100
120
20
30
40
60
70
Time on stream / min
2
/ degree
Fig. 2. Styrene yield over AZrO3 (A: Ca, Sr, Ba), ZrO2 and potassium-promoted iron
oxide catalysts for EBDH under the following condition: the feed rate of EB was
Fig. 1. XRD patterns of AZrO3 (A: Ca, Sr, Ba) and ZrO2 materials. The symbol “P” was
for perovskite; and “m” was for monoclinic.
–
1
1.2 g h ; the catalyst weight was 1.0 g; and the reaction temperature was 823 K.
showed a higher styrene yield than the Fe–K catalyst after 40 min
of EBDH.
perovskite structure [23]. The CZO catalyst also showed a perov-
skite structure, although the XRD pattern of the CZO was different
from those of the SZO and BZO catalysts [24]. The SZO and BZO have
an ideal perovskite structure, while the CZO structure is composed
In order to examine the effect of the reaction temperature on
the dehydrogenation performance, activity tests over the AZO (A:
Ca, Sr or Ba) and ZrO2 catalysts were performed within the tem-
of a distorted ZrO octahedra. Due to this distortion, the coordina-
6
–1
2
+
perature range of 823–923 K. The feed rate of EB was 1.2 g h , and
the catalyst weight was 0.05 g for the BZO, 0.2–0.4 g for the SZO,
tion of the Ca ion is reduced from 12 in ideal perovskite to 8 in
the CZO. The difference in the XRD pattern might come from such
a distortion of the perovskite material, while the ZrO2 has only a
monoclinic phase in the material. Table 1 presents the specific sur-
0
.1–0.8 g for the CZO and 0.8 g for ZrO2 catalysts, respectively. The
styrene yield versus time is shown in Fig. 3. In addition, Fig. 4(a)
denotes the maximum formation rate of styrene over these cata-
lysts at each reaction temperature, which was obtained from Fig. 3.
From Fig. 3(a) and (b), an induction period was observed over the
CZO and SZO catalysts. The induction period became shorter at the
higher temperatures. The BZO catalyst showed a high dehydro-
genation activity at a lower reaction temperature than the other
catalysts, while the ZrO2 catalyst showed a significantly low dehy-
drogenation activity and did not have an induction period. Fig. 4(b)
shows the styrene selectivity over the AZO (A: Ca, Sr, Ba) catalysts
and the ZrO2 catalyst at each reaction temperature. Among these
catalysts, the BZO catalyst showed the highest selectivity to styrene.
The CZO and SZO catalysts also had a high selectivity to styrene, but
face area of the perovskite oxides and ZrO . The specific surface
2
area increased with an increase in the atomic radius of the A-site
cation in the zirconium-based perovskite-type oxides.
3
.2. Dehydrogenation performance over the AZrO (A: Ca, Sr, Ba)
3
catalysts
The EBDH properties over the AZO (A: Ca, Sr, Ba) perovskite-type
oxide catalysts and ZrO2 were investigated for clarifying their abil-
ity to produce styrene at the low temperature of 823 K. The feed
–
1
rate of EB was 1.2 g h , and the catalyst weight was 1.0 g. Fig. 2
shows the styrene yield with time over these catalysts. The Fe–K
industrial catalyst (G-84C, made by Nissan Girdler K.K.) was utilized
as the reference. As shown in Fig. 2, the CZO, SZO and ZrO2 cata-
lysts did not show any dehydrogenation activity at 823 K. On the
other hand, the BZO catalyst showed an induction period of about
the ZrO catalyst showed a relatively low selectivity. The reason for
2
the induction period and the higher dehydrogenation activity over
the BZO catalyst is considered in Sections 3.3–3.6.
3.3. Bulk property of the BZO catalyst
1
5 min, and the activity passed through a maximum with time, fol-
lowed by a slow deactivation. The maximum styrene yield of 64.9%
with a 97.4% styrene selectivity was obtained, which was relatively
close to the equilibrium value. The styrene yield then decreased
to 36.0% at 100 min of reaction. The Fe–K catalyst represented the
high initial styrene yield of 67.0% and styrene selectivity of 98.7%
at 20 min. The styrene yield then decreased to 15.0% at 100 min.
Comparing the styrene yield over the BZO catalyst with that over
the industrial Fe–K catalyst, the BZO catalyst had a slightly lower
dehydrogenation activity at 20 min. However, the BZO catalyst
For investigating the relation between the gradual increase in
the dehydrogenation activity and the structural change in the BZO
catalyst, XRD measurements of the used BZO catalysts were con-
ducted. The used catalyst was defined as the BZO catalyst after
EBDH, and the styrene yield over the used catalyst is shown in
parenthesis. Four kinds of used BZO catalysts (3.7), (35.5), (55.4) and
(
57.3 → 47.2) were characterized. The used catalyst (57.3 → 47.2)
means the degraded value (47.2% styrene yield) after the maximum
yield (57.3%). These catalysts are the same lot, which was prepared
by the same method. Each catalyst had a different time courses of
the EBDH reaction. For example of the BZO (3.7), the reaction was
purposely stopped at the time when the BZO catalyst showed the
styrene yield of 3.7%. Other catalysts was obtained in the same way.
Fig. 5 shows the XRD patterns of the used BZO catalysts. A structural
change and a peak shift were not observed in these XRD patterns.
Table 1
Specific surface area of AZrO3 (A: Ca, Sr, Ba) and ZrO2 catalysts.
Catalyst
Specific surface area (m2 g-cat–1
)
CaZrO3
SrZrO3
BaZrO3
ZrO2
2.1
4.7
10.0
0.2
The lattice constant of all the used catalysts was 4.193 A. The bulk
property of the catalyst was not affected by the reaction gas. The
˚

R. Watanabe et al. / Applied Catalysis A: General 482 (2014) 344–351
347
3
2
1
.0 x 10-6
.0 x 10-6
.0 x 10-6
6.0 x 10-6
(
a) CaZrO3
(b)SrZrO 3
848 K
848 K
873 K
898 K
8
73 K
8
9
98 K
23 K
9
23 K
4.0 x 10-6
2.0 x 10-6
0
0
0
30
60
90
120
0
60
120
180
240
Time on stream / min
Time on stream / min
2
1
1
5
.0 x 10-5
.5 x 10-5
.0 x 10-5
.0 x 10-6
1.0 x 10-6
(
d) BaZrO3
(d) ZrO2
8
23 K
823 K
833 K
848 K
873 K
8
33 K
48 K
.0 x 10-7
.0 x 10-7
.0 x 10-7
.0 x 10-7
8
8
6
4
2
873 K
0
0
0
70
140
210
0
10
20
30
40
Time on stream / min
Time on stream / min
Fig. 3. Styrene yield versus time over (a) CaZrO3, (b) SrZrO3, (c) BaZrO3 and (d) ZrO2 catalysts under the following condition: the feed rate of EB was 1.2 g h–1; the catalyst
weight was 0.05 g for the BZO, 0.2–0.4 g for the SZO, 0.1–0.8 g for the CZO and 0.8 g for ZrO2 catalysts; and the reaction temperature was from 823 to 923 K.
morphology of the BZO catalyst would not change with time. For
confirming the shape of the catalyst, FE-SEM images of the used
catalysts were obtained as shown in Fig. 6. From Fig. 6(b) and (d),
the cubic BZO crystal was observed over the BZO (3.7) and BZO
almost the same of about 50 nm. Comparing to the morphology and
the cubic diameter of both catalysts, these structural property did
not change, in spite of a different in the catalytic activity. Hence,
we postulated that the induction period was not related to the bulk
structural property and the morphology of the BZO catalyst.
(
57.3 → 47.2) catalysts. The crystal diameter of these catalysts was
-
-
5
5
2
1
1
5
.0 x 10
100
CaZrO3
SrZrO3
BaZrO3
ZrO2
(
a)
(
b)
95
.5 x 10
9
8
8
0
5
0
CaZrO3
SrZrO3
BaZrO3
ZrO2
.0 x 10-5
-
6
.0 x 10
0
75
800
8
00
830
860
890
920
950
830
860
890
920
950
Reaction temperature / K
Reaction temperature / K
Fig. 4. Effect of reaction temperature on (a) maximum formation rate of styrene and (b) styrene selectivity over AZrO3 (A: Ca, Sr, Ba) catalyst for EBDH.

3
48
R. Watanabe et al. / Applied Catalysis A: General 482 (2014) 344–351
C-C/C=C
-
COO
BZO
57.3 47.2)
C=O
(
(
d)
c)
BZO (55.4)
BZO (35.5)
(
(b)
BZO (3.7)
(a)
2
92
290
288
286
284
282
Binding energy / eV
20
30
40
50
60
70
degree / 2
Fig. 7. C1s XPS spectra of the used BZO catalyst.
Fig. 5. XRD patterns of (a) BZO (3.7), (b) BZO (35.5), (c) BZO (55.4) and (d) BZO
57.3 → 47.2) catalysts.
(
courses. Hence, the carbonyl group on the used BZO would not
affect the dehydrogenation activity of the BZO catalyst.
To obtain information on the reduced state of the BZO catalyst,
ESR spectra for the used catalysts were recorded at room tempera-
ture as shown in Fig. 8. The used BZO catalyst gave the ESR signal at
g = 2.004, whose g-value corresponded to oxygen vacancies in the
catalyst [27]. The intensity of the signal at g = 2.004 was found to
increase with an increase in the styrene yield. Therefore, the num-
ber of oxygen vacancies was calculated for relating the activity and
oxygen vacancy in the catalyst.
3
.4. Surface property for the BZO catalyst
Osswald et al. mentioned that unsaturated ketone-/diketone-
type carbonyl groups (C O) have a substantial electron density
around the oxygen atom, and thus can serve as a Lewis bases to
activate saturated hydrocarbons [25]. For clarifying whether or
not carbonyl groups were confirmed on the used catalysts, X-ray
photoelectron spectroscopy (XPS) analysis of the used catalysts
was performed. The peak intensity of carbonyl group (C O) would
increase if the carbonyl group were accumulated on the catalyst
with time on stream. The XPS spectra of C1s for these catalysts
are shown in Fig. 7. The peaks at 284.4, 288.1 and 290.2 eV were
assigned to C–C/C C, C O and O C–O, respectively [26]. Compar-
ing to the peaks of the carbonyl group (C O) at 288.1 eV over the
four used BZO (3.7), (35.5), (55.4) and (57.3 → 47.2) catalysts, the
peak intensities were almost the same. Namely, the amount of
the carbonyl group might not increase on the catalyst with time
Table 2 shows the area of each peak and the number of oxygen
vacancies (denoted as Nvox). These areas of the signal at g = 2.004
6
6
increased as follows: 9.2 × 10 for the BZO (3.7%), 9.8 × 10 for
6
the BZO (35.5%), and 13.5 × 10 for the BZO (55.4%). Comparing
to the calculated Nvox from these areas, the Nvox increased with
18
–1
the increase in the styrene yield: 4.8 × 10 mol-cat for the BZO
18
–1
18
(3.7%), 5.2 × 10 mol-cat for the BZO (35.5%), and 7.5 × 10 mol-
–1
cat for the BZO (55.4%). There was a tendency that the number of
oxygen vacancies increased with the increase in the catalytic activ-
ity. The increase in the number of oxygen vacancies suggested that
Fig. 6. FE-SEM images of (a), (b) BZO (3.7) catalyst and (c), (d) BZO (57.3 → 47.2) catalyst. Magnification: (a), (c) ×50,000 and (b), (d) ×200,000.

R. Watanabe et al. / Applied Catalysis A: General 482 (2014) 344–351
349
the BZO catalyst was gradually reduced during the dehydrogena-
tion reaction. Hence, the reduced state of the BZO catalyst might
have an impact on the dehydrogenation activity. So far, there have
been some reports about the effect of the Fe component reduction
in the Fe–K catalyst on its EBDH performance [28,29]. It was men-
tioned that a deactivation occurred by reduction of the hematite
catalyst to the magnetite-type. Zhu et al. also insisted that the par-
tial reduction of the Fe3+-containing iron oxide phases caused the
fast initial deactivation [30]. On the other hand, Weiss et al. investi-
gated the role of the iron oxide stoichiometry and the defect-site on
the catalytic activity [31]. They concluded that the styrene forma-
tion rate increased with the increasing defect-site concentration,
and the defect-site was the active site for the EB dehydrogenation
over Fe O . In this study, the reduced state of the catalyst may con-
2
3
tribute to the enhancement of the dehydrogenation activity. These
results were considered to be the important results, and should be
further verified in the next section.
Fig. 8. ESR spectra of the used BZO catalyst.
3
.5. Effect of reduced state in the BZO catalyst on the
Table 2
dehydrogenation performance
The number of oxygen vacancies (NVox) in the used BZO catalyst.
Catalysta
Area (×10 )
6
NVox (×10 mol-cat–1
18
)
The EBDH performance of the BZO catalyst, which was prere-
duced and/or prereduced followed by oxidation, was investigated
to clarify the effect of the reduced state of the BZO catalyst. Fig. 9(a)
shows the styrene yield with time over the BZO catalysts with-
BZO (3.7)
BZO (35.5)
BZO (55.4)
BZO (57.3 → 47.2)
9.2
9.8
13.5
15.8
4.8
5.2
7.5
8.5
out pretreatment, with prereduction by 10 vol%-H at 798 or 823 K,
2
a
The value in parenthesis is the styrene yield.
with prereduction by 100 vol%-H2 at 923 K and with prereduction
60
50
40
30
20
10
0
60
Without pretreatment
(a)
(b)
Without pretreatment
0 vol%-H , 823 K
00 vol%-H , 923 K
1
0 vol%-H
2
2
2
, 798 K
, 823 K
, 823 K
1
1
2
10 vol%-H
50
40
30
20
10
0
2
1
0 vol%-H
2
2
0 vol%-O , 1073 K
0
5
10
15
20
25
30
0
5
10
15
20
25
30
Time on stream / min
Time on stream / min
6
5
4
3
2
1
0
0
0
0
0
0
0
Without pretreatment
00 vol%-H , 923 K
(
c)
1
2
0
5
10
15
20
25
30
Time on stream / min
Fig. 9. Effect of prereduction by H2 on the catalytic performance of (a) BZO, (b) SZO and (c) ZrO2 at 823 K. The feed rate of EB was 1.2 g h–1 and the catalyst weight was 0.2 g
for BZO, 1.0 g for SZO and ZrO2.

3
50
R. Watanabe et al. / Applied Catalysis A: General 482 (2014) 344–351
by 10 vol%-H2 at 823 K followed by oxidation by 20 vol%-O at
1
was 0.2 g. As clearly seen from Fig. 9(a), the BZO catalyst with pre-
treatment by H2 reduction represented a high activity without the
induction period. On the other hand, the induction period was
observed again over the BZO catalyst with prereduction by H2 at
Region (I)
Region (II)
94 K
2
073 K. The feed rate of EB was 1.2 g h–1, and the catalyst weight
5
With pretreatment
6
30 K
Chemisorbed
species
8
23 K followed by oxidation with O2 at 1073 K. These phenomena
9
47 K
were considered to be based on the following Eqs. (7) and (8), where
O
2–
and Vox denote a lattice oxygen and an oxygen vacancy in the
lat
catalyst, respectively.
Physisorbed species
2
–
–
H + O
→ H O + Vox + 2e
(7)
(8)
2
lat
2
O + Vox + 4e → 2O2–
–
2
lat
Without pretreatment
When the BZO catalyst was reduced by H , the oxygen vacancy
2
was created based on Eq. (7). Thus, the BZO catalyst showed a
high activity during the initial stage of the dehydrogenation. By
reduction followed by oxidation, the lattice oxygen was replen-
ished based on Eq. (8). Hence, the induction behavior was again
confirmed over the BZO catalyst. Fig. 9(b) shows the effect of pre-
reduction at 823 K by 10 vol%-H2 or 923 K by 100 vol%-H2 on the
dehydrogenation performance over the SZO catalyst. The SZO cata-
lyst without pretreatment did not show a dehydrogenation activity
at 823 K. When the reduction was performed over the SZO catalyst,
the representation of the dehydrogenation activity was confirmed.
Such a result was considered to be derived from the progression
of the reduction due to Eq. (7). A further increase in the H2 con-
centration and reduction temperature from 10 vol%-H2 at 823 K to
3
00
500
700
900
1100
Temperature / K
Fig. 10. Profiles of temperature programmed desorption of EB over the BZO catalyst
with and without prereduction by H2.
the BZO catalyst with or without prereduction could be explained
by the different adsorption sites on the catalyst. Since the outer-
most surface layer of the as-made BZO catalyst consisted of lattice
oxygen, no chemical interaction of EB occurred on the oxygen ter-
minated BZO surface. Joseph et al. investigated the adsorption of
EB on a well-ordered epitaxial iron oxide model catalyst by near-
edge X-ray absorption fine structure spectroscopy [32]. The soft
basic aromatic molecules of EB were chemisorbed on the acidic iron
sites of the iron oxide catalyst surfaces. Therefore, in this study, the
almost absence of chemisorbed species on the bare BZO catalyst
might come from a repulsive electrostatic interaction between the
surface lattice oxygen and the ␲ orbital in EB. As for the BZO cat-
alyst with prereduction, the surface lattice oxygen was removed
1
00 vol%-H at 923 K produced a higher activity. Fig. 9(c) shows the
2
catalytic behavior over the ZrO catalyst with or without reduction.
2
Unlike the BZO and SZO catalysts, the ZrO catalyst did not produce
2
any dehydrogenation activity even after prereduction by 100 vol%-
H at 923 K. The ZrO catalyst might not be reduced even under the
2
2
reductive atmosphere.
The BZO and SZO catalysts showed a dehydrogenation activity
by reduction of the catalysts, while the ZrO2 did not reveal any
activity even if the prereduction was performed. By inserting Zr in
the perovskite-type structure, a reducibility of the catalyst might
be enhanced. Especially, the incorporation of Ba as the A-site cation
would be the most effective for lowering the reduction tempera-
ture. The incorporation of Ba with a large ionic radii in the A-site
of the perovskite structure produces a large free volume in the lat-
tice, which provides a high mobility for the lattice oxygen. Such an
effect could create oxygen vacancies in the BZO structure under a
mild reducing atmosphere.
by H , and the Zr cation as the acidic site was then exposed on
2
the catalyst surface. Namely, the production of the oxygen vacancy
opened the EB adsorption channel, and chemisorption species were
observed as shown by the higher desorption temperature. There-
fore, the increase in the oxygen vacancy created the reactive site
for EBDH, which produced a high activity over the BZO catalyst.
4. Conclusions
The BZO catalyst produced a high activity for EBDH at the low
temperature of 823 K, while the CZO, SZO and ZrO2 catalysts did
not show a EBDH activity. The reason for the high activity over the
BZO catalyst was derived from the production of oxygen vacancies
during the EBDH reaction, as confirmed by ESR measurements and
some activity tests. Comparing the EB-TPD property of the as-made
and the reduced BZO catalysts, the physisorbed species was mainly
present on the as-made catalyst, whereas the chemisorbed species
was on the reduced catalyst. The key point of the high activity over
the Zr-based perovskite-type catalyst was the reduced state under
the stated reaction conditions.
3
.6. Effect of reducing state of the BZO catalyst on EB
adsorption/desorption properties
In order to examine the effect of the reducing state of the
catalyst on EB adsorption/desorption properties, a temperature-
programmed desorption (TPD) was performed. Fig. 10 shows the
EB-TPD profiles of the BZO catalyst with or without prereduction by
H2 at 823 K. From the TPD profile of the bare BZO catalyst, the pro-
file of EB (m/z = 91) desorption showed one sharp signal in region
(
(
I). The signal in region (II) was almost not observed. The region
I) signal below 423 K came from the physical adsorption of the EB
species, whereas the region (II) signal above 423 K was derived from
the chemical adsorption of the EB species. The desorption profile
indicated that the EB species with a weak interaction was present
on the BZO catalyst without prereduction. For the BZO catalyst with
prereduction, the some signals were observed in regions (I) and (II).
As clearly be seen from the profile, the signal intensity for region (I)
was significantly lower than that for region (II). In other words, the
catalyst surface was covered by the EB species with a strong inter-
action on the catalyst surface. The differences in the TPD profiles of
References
[
[
[
[
1] J.-H. Lee, H. Moon, H.-W. Lee, J. Kim, J.-D. Kim, K.-H. Yoon, Solid State Ionics 148
2002) 15–26.
2] E.S. Hech, G.K. Gupta, H. Zhu, A.M. Dean, R.J. Kee, L. Maier, O. Deutschmann,
Appl. Catal. A: Gen. 295 (2005) 40–51.
(
3] H. Koide, Y. Someya, T. Yoshida, T. Maruyama, Solid State Ionics 132 (2000)
253–260.
4] S. Yu, Q. Wu, M. Tabib-Azar, C.-C. Liu, Sens. Actuators B 85 (2002) 212–218.
[5] N. Rajabbeigi, B. Elyassi, A. Khodadadi, S.S. Mohajerzadeh, M. Sahimi, Sens.
Actuators B 100 (2004) 139–142.

R. Watanabe et al. / Applied Catalysis A: General 482 (2014) 344–351
351
[
[
[
[
6] S.M. de Lima, I.O. da Cruz, G. Jacobs, B.H. Davis, L.V. Mattos, F.B. Noronha, J.
Catal. 257 (2008) 356–368.
7] S.M. de Lima, A.M. Silva, I.O. da Cruz, G. Jacobs, B.H. Davis, L.V. Mattos, F.B.
Noronha, Catal. Today 138 (2008) 162–168.
8] P.P. Silva, F.A. Silva, H.P. Souza, A.G. Lobo, L.V. Mattos, F.B. Noronha, C.E. Hori,
Catal. Today 101 (2005) 31–37.
9] R.J.H. Voorhoeve, J.P. Remeika, L.E. Trimble, Ann. N. Y. Acad. Sci. 272 (1976)
[19] Y. Teraoka, T. Nobunaga, N. Yamazoe, Chem. Lett. 88 (1988) 503–506.
[20] J.W. Stevenson, T.R. Armstrong, R.D. Carmeim, L.R. Pederson, L.R. Weber, J.
Electrochem. Soc. 143 (1996) 2722–2729.
[21] R.L. Cook, A.F. Sammells, Solid Sate Ionics 45 (1991) 311–321.
[22] J. Hu, H. Hao, C. Chen, D. Yang, X. Hua, J. Membr. Sci. 280 (2006) 809–814.
[23] J. Huang, L. Zhou, Z. Wang, Y. Lan, Z. Tong, F. Gong, J. Sun, L. Li, J. Alloys Compd.
487 (2009) L5–L7.
3
–21.
[24] P. Stoch, J. Szczerba, J. Lis, D. Madej, Z. P e˛ dzich, J. Eur. Ceram. Soc. 32 (2012)
[
[
10] T. Nitadori, T. Ichiki, M. Misono, Bull. Chem. Soc. Jpn. 61 (1988) 621–626.
11] Y. Teraoka, H.M. Zhang, K. Okamoto, N. Yamazoe, Mater. Res. Bull. 23 (1988)
665–670.
[25] S. Osswald, G. Yushin, V. Mochalin, S.O. Kucheyev, Y. Gogotsi, J. Am. Chem. Soc.
128 (2006) 11635–11642.
[26] A. Achour, A.A. El Mel, N. Bouts, E. Gautron, E. Grigore, B. Angleraud, L. Le
Brizoual, P.Y. Tessier, M.A. Djouadi, Diamond Related Mater. 34 (2013) 76–83.
[27] Y. Nosaka, S. Takahashi, Y. Mitani, X. Qiu, M. Miyauchi, Appl. Catal. B: Environ.
111–112 (2012) 636–640.
5
1–58.
12] H.M. Zhang, Y. Shimizu, Y. Teraoka, N. Miura, N. Yamazoe, J. Catal. 121 (1990)
32–440.
[
4
[
[
13] M.A. Pe n˜ a, J.L.G. Fierro, Chem. Rev. 101 (2001) 1981–2017.
14] R. Watanabe, Y. Sekine, M. Matsukata, E. Kikuchi, Catal. Lett. 131 (2009)
5
4–58.
15] R. Watanabe, Y. Sekine, J. Kojima, M. Matsukata, E. Kikuchi, Appl. Catal. A: Gen.
98 (2011) 66–72.
16] R. Watanabe, K. Mukawa, J. Kojima, E. Kikuchi, Y. Sekine, Appl. Catal. A: Gen.
62–463 (2013) 168–177.
[28] G.R. Meima, P.G. Menon, Appl. Catal. A: Gen. 212 (2001) 239–245.
[29] O. Shekhah, W. Ranke, R. Schlögl, J. Catal. 225 (2004) 56–68.
[30] X.M. Zhu, M. Schön, U. Bartmann, A.C. van Veen, M. Muhler, Appl. Catal. A: Gen.
266 (2004) 99–108.
[31] W. Weiss, D. Zscherpel, R. Schlögl, Catal. Lett. 52 (1998) 215–220.
[32] Y. Joseph, M. Wühn, A. Niklewski, W. Ranke, W. Weiss, C. Wöll, R. Schlögl, Phys.
Chem. Chem. Phys. 2 (2000) 5314–5319.
[
[
3
4
[
[
17] http://www.bgs.ac.uk/mineralsuk/home.html
18] S. Li, W. Jin, P. Huang, N. Xu, J. Shi, Ind. Eng. Chem. Res. 38 (1999) 2963–2972.