Performance of Non-stoichiometric Perovskite Catalyst (AxCrO3?δ, A: La, Pr, Nd) for Dehydrogenation of Propane Under Steam Condition

Article abstract of DOI:10.1007/s10562-016-1876-5

Abstract: The aim of this study was to clarify the effect of the non-stoichiometry in the A_xCrO_3?δ (A: La, Pr, Nd, 0.90 ≤ x ≤ 1.05) perovskite-type oxide on the dehydrogenation performance of propane using steam (DHP-S). The slight change in the molar ratio of the A-site (=La, Pr, Nd) to the B-site cation (=Cr) from the stoichiometric composition (A/B = 1) significantly affected the dehydrogenation performance. For the La-based catalyst, the La_xCrO_3?δ catalyst with 0.90 ≤ x ≤ 0.97 displayed a relatively high activity, while the catalyst with 0.97 < x ≤ 1.05 did not show any activity for the DHP-S. The perovskite catalyst having the A-site defect presented a high performance for the DHP-S regardless of the types of A-site cation. The Nd_0.95CrO_3?δ catalyst displayed the highest performance among the A-site defective perovskite catalysts. The Nd_0.95CrO_3?δ catalyst also showed a superior stability compared to the CrO_x/γ-Al₂O₃ catalyst. Based on an XPS measurement, the reduction–oxidation (redox) property was improved by changing the Nd/Cr molar ratio from the stoichiometric composition (Nd/Cr = 1) to a non-stoichiometric composition. Such a good redox property would produce a high performance for the DHP-S. Graphical Abstract: [Figure not available: see fulltext.]

Full text of DOI:10.1007/s10562-016-1876-5

Catal Lett
DOI 10.1007/s10562-016-1876-5
Performance of Non-stoichiometric Perovskite Catalyst
(
A CrO , A: La, Pr, Nd) for Dehydrogenation of Propane Under
x
3−δ
Steam Condition
1
1
1
Ryo Watanabe · Masahiro Tsujioka · Choji Fukuhara
Received: 20 July 2016 / Accepted: 26 September 2016
©
Springer Science+Business Media New York 2016
Abstract The aim of this study was to clarify the effect
of the non-stoichiometry in the A CrO (A: La, Pr, Nd,
Graphical Abstract
x
3−δ
Propane dehydrogenation: C
Propylene sel. 65.5%
3
H
8
C
3
H
6
+ H
2
0
.90≤x≤1.05) perovskite-type oxide on the dehydroge
-
20
15
10
5
0
nation performance of propane using steam (DHP-S). The
slight change in the molar ratio of the A-site ( =La, Pr, Nd)
to the B-site cation ( =Cr) from the stoichiometric compo -
sition (A/B =ꢀ1)ꢀsignificantlyꢀaffectedꢀtheꢀdehydrogenationꢀ
CrO /γ-Al O
x 2 3
performance. For the La-based catalyst, the LaCrO cata-
x
3−δ
Nd0.95CrO3-δ
lyst with 0.90 ≤x≤0.97 displayed a relatively high activ -
ity, while the catalyst with 0.97 <x≤1.05 did not show any
activity for the DHP-S. The perovskite catalyst having the
A-site defect presented a high performance for the DHP-S
Propylene
sel. 77.4%
regardless of the types of A-site cation. The Nd 0.95CrO
catalyst displayed the highest performance among the
3−δ
A-site defective perovskite catalysts. The Nd
catalyst also showed a superior stability compared to the
0.95CrO
3−δ
0
70
140
210
280
350
CrO /γ-Al O catalyst. Based on an XPS measurement,
x
2
3
the reduction–oxidation (redox) property was improved
Time on stream / min
by changing the Nd/Cr molar ratio from the stoichiometric
composition (Nd/Cr =1) to a non-stoichiometric composi -
tion. Such a good redox property would produce a high per-
formance for the DHP-S.
Keywords Non-stoichiometry · Perovskite · Propylene ·
Dehydrogenation
1
Introduction
Propylene is utilized in petrochemistry industry as an impo-r
tant chemical intermediate for the production of polypro
-
pylene, propylene oxide, and acrylonitrile [ 1, 2]. The main
routes for the production of propylene are the thermal crack-
ingꢀofꢀnaphthaꢀandꢀtheꢀfluidꢀcatalyticꢀcrackingꢀ(FCC)ꢀ[3–5].
The naphtha steam cracker and FCC units, respectively,
produce ethylene and gasoline as the main products and
propylene as a byproduct. The growth rate of the propylene
ꢀ
Choji Fukuhara
tcfukuh@ipc.shizuoka.ac.jp
1
Department of Applied Chemistry and Biochemical
Engineering, Graduate school of Engineering, Shizuoka
University, 3-5-1, Johoku, Naka-ku, Hamamatsu, Shizuoka,
Japan
1
3

2
R. Watanabe et al.
demand is estimated at 4–5% per year [2, 6]. Therefore, the
oxygen proceeded via the dissociative chemisorption of O 2
marketꢀ demandꢀ ofꢀ propyleneꢀ willꢀ notꢀ beꢀ easilyꢀ satisfied,ꢀ and the reduced V centers were simultaneously oxidized.
because the operations of the steam cracker and FCC units
are optimized for reduced production of ethylene and gaso-
line. The risk of a propylene supply shortage is expected in
the near future. Recently, a great deal of attention has been
devoted to the on-purpose technologies for the propylene
production, such as the dehydrogenation of propane (DHP,
Eq. 1),ꢀolefinꢀmetathesis,ꢀandꢀtheꢀcatalyticꢀcrackingꢀofꢀC4/
C5 [2, 7, 8].
The rate constants for the re-oxidation of the reduced cen-
ters using O are greater than that for their reduction by the
2
C–H bond activation steps. Several researchers showed that
more reducible oxides could produce high alkane ODH
rates [24–34]. However, for the DHP with O, the propylene
2
selectivityꢀsignificantlyꢀdecreasedꢀbecauseꢀofꢀtheꢀoxidationꢀ
of propane and propylene to form CO and CO 2. In a pre-
vious study, steam was found to regenerate lattice oxygen,
which is used for the dehydrogenation of hydrocarbons
[35]. Namely, steam can act as a mild oxidant of the cata -
lyst. Unlike oxygen, steam would not oxidize the propane
C( 1 H) → C H + H
2
3
8
3
6
Among these processes, the DHP is attractive in terms of the and propylene, which would produce a high selectivity for
direct conversion of an economic feedstock, i.e., propane, to propylene.
the valuable propylene, which can contribute to the future
chemical industry. The CrOx/γ-Al2O3ꢀ catalystꢀ showsꢀ aꢀ A is a rare-earth or an alkali-earth element and B is typically
high activity and good selectivity to propylene for the DHP. a transition metal) have a high mobility of lattice oxygen
However, on the CrOx/γ-Al2O3ꢀ catalyst,ꢀ cokeꢀ depositionꢀ in comparison to some transition metal oxides. The redox
Perovskite-type mixed oxides (general formula is ABO3,
easily occurs and deactivates the performance, which is the
important shortcoming of this catalyst [9–13]. The use of
steam can normally suppress the coke deposition, but steam
ability of the perovskite materials could be controlled by
the incorporation of a variety of elements [36–39]. There-
fore, a high activity for the DHP is expected by regulating
deactivates the CrOx/γ-Al2O3ꢀcatalystꢀdueꢀtoꢀadsorptionꢀbyꢀ redox property of the perovskite-type oxide. In addition, the
the active-site of Cr. Therefore, a regeneration process for
the coke combustion by diluted air is required after short
perovskite-type mixed oxides have a much better thermal
stability than transition metal oxides, and their catalytic
reaction periods [14]. Such a cyclic regeneration process for activities are sometimes competitive with the supported
remaining the deposited coke causes the irreversible deacti-
vation of the catalyst by a phase transformation and sinter -
ing. For such reasons, the development of a novel catalyst
having a stable performance during propylene production
is needed.
The oxidative dehydrogenation (ODH) of propane to
propylene (Eq. 2) over the V-based catalysts has been
widely investigated for the continuous production of pro-
pylene [15–18]. The reaction mechanism over the V-based
catalysts is the Mars van
noble metal catalysts [40]. The catalytic property of the
perovskite-type oxides basically depend on the nature of the
B-site cations and on their valence state, while the A-site
cations are catalytically inactive. The partial replacement
of the A-site or B-site cations with other metals improves
the catalytic properties. The effect of the A-site and B-site
substitution is not only the valence change in the originally
incorporated B-site cation, but also the creation of an oxy-
gen vacancy in order to maintain an electroneutral condition.
Such a valence change in the B-site cation and production
of an oxygen vacancy would improve catalytic performance
due to an enhancement of the lattice oxygen mobility. For
example, La0.8Sr CoO and La Sr FeO showed a bet-
C( 2 H) + 1/ 2O → C H + H O
3
8
2
3
6
2
0.2
3
0.8 0.2
3
Krevelen redox pathways with lattice oxygen involved in
the C–H bond activation steps [19–22]. According to K.
Chen et al., a weak associative adsorption of propane on
the lattice oxygen occurred and the C3H8O* (O*: lattice
oxygen) species as the intermediate was formed. Subse-
quently, C–H cleavage via H-abstraction from the propane
using a neighboring lattice oxygen proceeded on the cata -
lyst, and C3H7O* was produced [23]. The hydride elimi -
nation from the adsorbed alkoxide species then occurred,
followed by desorption of the produced propylene. The two
remaining OH groups were combined to form H 2O and a
lattice vacancy, accompanied by a two-electron reduction
ter performance for C H , CH and CO combustion than
3 8 4
the bare LaCoO and LaFeO catalysts [41–44]. By substi-
3
3
3
+
2+
4+
tuting La with Sr in the A-site, the Co
and oxygen
vacancies were produced in the perovskite to maintain an
electroneutral condition. These effects could produce a high
lattice oxygen mobility, and produce a high activity for the
combustion of hydrocarbons [45]. The substitution in the
B-site also affected the catalytic activity of the perovskite
materials. Gallagher et al. and Yasuda et al. investigated the
catalytic property of the LaCu Mn O (0 ≤x≤1) for the
x
1−x
3
CO oxidation [46, 47]. The LaCu 0.4Mn O catalyst rep-
0.6
3
resented the highest activity, which was dozens of times
n+
(n−2)+
of the high-valent V cations to form one V
or two
higher than the LaMnO catalyst during the initial stage of
3
(n−1)+
V
cations. The replenishment of the consumed lattice
the reaction. They presumed that this synergistic effect is
1
3

Performance of Non-stoichiometric Perovskite Catalyst (A CrO3−δ, A: La, Pr, Nd) for Dehydrogenation...
3
x
due to the bi-functional catalysis of Mn and Cu. From the
TPD (temperature programmed desorption) of oxygen and
obtained powder was next calcined at 1123 K for 10 h. The
CrO /γ-Al O catalyst as a reference was prepared by a
x
2
3
the reaction with CO, the reactivity of oxygen in the catalyst conventional impregnation method using Cr(NO ) ·9H O
3
3
2
was accelerated by the substitution with Mn. In addition, the (Kantoꢀ Kagakuꢀ Co.,ꢀ Ltd.,ꢀ Japan)ꢀ andꢀ γ-Al O support,
2
3
amount of CO adsorption increased by the substitution with
Cu. The authors proposed that the synergistic effect was
produced by the appropriate combination of the activation
which was obtained from the Catalysis Society of Japan
(JRC-ALO-6), with a sequential calcination at 923 K in air.
The loading amount of the Cr component was 15 wt%.
of O by Mn and that of CO by Cu. A similar case was also
2
observed over the LaCu Mn O (0≤x≤1) catalyst for the
2.2 Reaction Conditions
x
1−x
3
NO–CO reaction [48]. There are many other reports about
the effect of the B-site replacement on the catalytic perfor -
mance for various reactions, such as the methane oxidation
The catalytic activity, selectivity and stability were evalu -
atedꢀusingꢀaꢀconventionalꢀfixedꢀbedꢀreactor.ꢀTheꢀreactionꢀ
temperature was 898 K at atmospheric pressure under steam
conditions; the molar ratio of steam to propane was changed
from 0.5 to 4. The propane feed rate was 5 ml min
catalyst weight was 150 mg for the CrO /γ-Al O catalyst
over a La 0.8Sr Mn Co O catalyst, the combustion of
0.2
1−x
x
3
ethanol and propane over a LaCo1 Fe O catalyst and the
−y
y
3±δ
−1
low temperature CO oxidation over a LaFe1
Cu Pd O
3−δ
. The
−x−y
x
y
catalyst [49–51].
x
2
3
To obtain a high performance catalyst, we focused on
and 500 mg for the Cr-based perovskite-type oxide. Gas-
eous products, such as propane, propylene, CO, CO , CH ,
such a perovskite material with a non-stoichiometric com-
position. Ullman et al. [52] reported that the atomic ratio
of the A-site/B-site in the perovskite-type oxide affected
the mobility of the lattice oxygen in the perovskite struc-
ture. Perovskite with an atomic ratio of the A-site/B-site <1
revealed a high mobility of the lattice oxygen due to the
productionꢀofꢀoxygenꢀvacanciesꢀasꢀaꢀresultꢀofꢀtheꢀA-siteꢀdefi-
ciencies. Ge et al. [53] reported that a perovskite with the
A-site/B-site <1 showed a high ionic conductivity although
it had a lower electronic conductivity. By utilizing a non-
stoichiometric perovskite type oxide as the catalyst, a high
activity is expected to be obtained. Therefore, in this study,
for developing a highly-active catalyst for the propane
dehydrogenation with steam (DHP-S), we investigated the
catalytic performance of the Cr-based perovskite catalyst,
and evaluated the effect of the non-stoichiometry in the Cr-
2
4
C H and C H , were analyzed using a gas chromatographic
2
4
2
6
(GC) analyzer with a thermal conductivity detector (TCD)
andꢀwithꢀaꢀflameꢀionizationꢀdetectorꢀ(FID)ꢀ (GC-8A;ꢀ Shi-
madzuꢀCorp.).ꢀTheꢀproductꢀselectivityꢀwasꢀdefinedꢀonꢀtheꢀ
basis of the carbon.
2.3 Characterization of Catalysts
The crystalline structure of the prepared catalyst was ascer -
tained by X-ray powder diffraction with CuK αꢀ radiationꢀ
(λꢀ=1.54 Å, Rint-2200; Rigaku Corp.). For analysis of the
catalytic surface, an X-ray photoelectron spectroscopy (XPS,
ESCA3400; Shimadzu Corp.) measurement was performed
using non-monochromatic MgK αꢀ radiation.ꢀ Theꢀ bindingꢀ
energy was referred to the carbon of C =284.7 eV for con-
1s
based perovskite-type oxide of A CrO (A: La, Pr, Nd,
venience.ꢀTheꢀspecificꢀsurfaceꢀareaꢀofꢀtheꢀperovskiteꢀoxideꢀ
x
3−δ
0
.90≤x≤1.05) on the catalytic property for the DHP-S. In
was measured using N adsorption at 77 K (3Flex Surface
2
this study, the atomic ratio of the A-site/B-site from 0.90
to 1.05 was varied, expecting improvement of the redox
property of the catalyst. The XRD and XPS analyses were
also conducted for clarifying the relationship between the
catalytic performance and the surface condition, such as the
structure and atomic ratio.
Characterization Analyzer, Micromeritics Co.). The sample
was outgassed at 573 K for 2 h before the N
adsorption.
2
The amount of deposited coke over the catalyst was mea-
sured using a thermogravimetric analysis with a TAPS3000S
(Bruker Corp.). The catalyst sample was placed on the cen-
terꢀofꢀtheꢀmuffle.ꢀTheꢀtemperatureꢀofꢀtheꢀcatalystꢀwasꢀthenꢀ
−1
raised by a 10 K min heating rate from room temperature
to 1173 K under a 20 vol% O atmosphere balanced by He.
2
2
Experimental
2
.1 Catalyst Preparation
3 Results and Discussion
The Cr-based perovskite-type oxides, LaCrO , Pr CrO
3.1 Dehydrogenation Performance over A CrO (A:
x
3−δ
x
3−δ
x
3−δ
and Nd CrO (0.90 ≤x≤1.05) as the dehydrogenation
La, Pr, Nd, 0.90≤x≤ 1.05)
x
3−δ
catalysts, were prepared by a complex polymerized method.
First, the metal precursors were dissolved in distilled water.
After citric acid and ethylene glycol were added, solution
was dried to form a gel, then pre-calcined at 673 K. The
The effect of the non-stoichiometry in the Cr-based
perovskite-type oxide on the dehydrogenation performance
wasꢀ firstꢀ examined,ꢀ expectingꢀ theꢀ realizationꢀ ofꢀ aꢀ highꢀ
1
3

4
R. Watanabe et al.
Fig. 1 Effect of the A-site/B-
site ratio on the DHP-S perfor-
mance over the a La CrO
,
x
3−δ
b Pr CrO and c Nd CrO
x
3−δ
x
3−δ
catalysts
performance for the DHP under steam conditions. Fig-
ure 1a, b, c show the propylene yield and propylene selectiv-
produce an oxygen vacancy due to electrical valence com-
pensation in the perovskite structure [36]. Such a production
of oxygen vacancies enhances the lattice oxygen mobility,
which promotes the dehydrogenation performance.
ity at 898 K over the La CrO , Pr CrO and Nd CrO
x
3−δ
x
3−δ
x
3−δ
catalysts, respectively. From the propylene yield shown in
Fig. 1a, the La CrO catalyst with the La/Cr molar ratio
from 0.90 to 0.97 indicated a relatively high value, while the condition produces a positive dehydrogenation over the
Forꢀ confirmingꢀ whetherꢀ orꢀ notꢀ theꢀ non-stoichiometricꢀ
x
3−δ
catalyst with the La/Cr molar ratio from 0.98 to 1.05 showed other perovskite catalysts, the performances of the Mn- and
a low value. The propylene yield over the L aC rO catalyst
Fe-based perovskite-type oxide catalysts were investigated.
Figure 2 shows the propylene yield over the LaBO 3 (B: Cr,
x
3−δ
with 0.98≤x≤1.05 was identical to the yield of the thermal
decomposition of propylene, at 898 K. In other words, the
catalyst of 0.98 ≤x≤1.05 did not show any activity for the
DHP-S. As for the propylene selectivity, almost the same
Mn, Fe) and La 0.95BO catalysts. The improvement of the
3−δ
catalytic activity over the Mn- and Fe-based perovskite cata-
lysts was not obtained by making the A/B ratio change from 1
to 0.95. Namely, the non-stoichiometric Cr-based perovskite
catalyst only showed a high activity for the DHP with steam.
value was obtained over the La xCrO catalyst catalysts
3−δ
with the La/Cr of 0.90–0.97. On the other hand, the catalyst
with La/Cr >0.98 showed a very low propylene selectivity,
which was derived from the gas phase reaction of propane-
3.2 Durability Performance over A0.95CrO (A: La,
3−δ
S. Over the Pr CrO and Nd CrO catalysts, the trends
Pr, Nd)
x
3−δ
x
3−δ
of the activity and selectivity were consistent with the result
of the La CrO catalyst. For the Pr CrO catalyst, a high
activity and a good selectivity were obtained over the cata -
In order to investigate the stability of the A 0.95CrO (A:
La, Pr, Nd) catalyst, the activity test for about 6 h was per -
x
3−δ
x
3−δ
3−δ
lysts with the Pr/Cr molar ratio below 0.98. In case of the
formed. Figure 3 shows the propylene yield with time over
Nd CrO catalyst, the threshold of the Nd/Cr molar ratio
the A0.95CrO (A: La, Pr, Nd) catalyst and the reference
3−δ
x
3−δ
was 0.95. The Nd CrO catalyst with the Nd/Cr molar
CrO /γ-Al O catalyst. The La 0.95CrO , Pr0.95CrO and
x
3−δ
x 2 3 3−δ 3−δ
ratio below 0.95 revealed a high activity and a good selec-
tivity. Regardless of the A-site cation species, the dehydro-
genation activity was observed over the catalyst with the
Nd0.95CrO catalysts produced the propylene yields of 8.8,
3−δ
11.2 and 10.2 % at 30 min, respectively. The Pr 0.95CrO
3−δ
catalyst exhibited a slightly higher activity than the
A-site/B-site molar ratio below 1.0. The ratio of <1.0 would La0.95CrO and Nd 0.95CrO catalysts, however, the
3−δ
3−δ
1
3

Performance of Non-stoichiometric Perovskite Catalyst (A CrO3−δ, A: La, Pr, Nd) for Dehydrogenation...
5
x
In order to investigate the reason for the difference in the
durability among the A0.95CrO (A: La, Pr, Nd) catalysts,
3−δ
thermogravimetric analyses were conducted to measure the
amount of deposited coke on these catalysts. Table 1 shows
the obtained results. A large amount of deposited coke
(
6.76 wt%) was observed over the CrO /γ-Al O catalyst,
x 2 3
while over the Nd0.95CrO catalyst, the amount of depos-
3−δ
ited coke was only 0.03 wt%. Namely, the coke deposition
did not almost occur on the Nd 0.95CrO catalyst. On the
3−δ
other hand, 0.33 and 0.09 wt% cokes were observed over
the La 0.95CrO and Pr 0.95CrO catalysts, respectively.
3−δ
3−δ
The order of the amount of the deposited coke had the
same trend as the deactivation degree. This result implied
that the reason for the deactivation might be caused by the
coke deposition on the catalyst. The high deactivation over
the CrO /γ-Al O catalyst shown in Fig. 3 was presumably
x
2
3
due to covering of the active sites by the coke. On the other
hand, the high stable activity over the Nd0.95CrO catalyst
3−δ
was due to almost no formation of coke on this catalyst.
Fig. 2 Propylene yield at steady-state over the LaBO and La0.95BO
Table 1 also shows the amount of coke on the Nd0.95CrO
3
3−δ
3−δ
(B: Cr, Mn, Fe) catalysts
catalyst after the DHP without steam. The coke was depos-
ited on the catalyst under the DHP without steam, and the
amount of coke was very small. Compared the amount of
coke after the DHP with/without steam, the coke deposi-
tion was suppressed by co-feeding steam with propane. The
deposited coke might be oxidized by lattice oxygen in the
20
15
10
5
0
Propylene sel. 65.5%
x 2 3
CrO O
/γ-Al
La0.95CrO3-δ
Pr0.95CrO3-δ
Nd0.95CrO3-δ
Propylene
sel. 78.6%
Nd0.95CrO catalyst.
3−δ
Based on these results, the Nd 0.95CrO catalyst was a
3−δ
good candidate catalyst for the DHP-S process. For clarify-
Propylene
sel. 77.4%
ing the effect of the partial pressure of steam on the propyl-
ene yield over the Nd0.95CrO and CrO /γ-Al O catalysts,
3−δ
x
2
3
Propylene
sel. 80.0%
the activity tests were carried out under various partial pres-
sures of steam. In order to accurately evaluate the effect of
the steam pressure on the stability, the activity test was con-
ducted under 150 mg of the CrO /γ-Al O catalyst and over
x
2
3
5
00 mg of the Nd0.95CrO catalyst for obtaining the same
3−δ
0
70
140
210
280
350
initial propylene yield. Figure 4a, b show the propylene
yield versus time over the Nd 0.95CrO and CrO /γ-Al O
3
Time on stream / min
3−δ
x
2
catalysts under various steam pressures, respectively. Both
Fig. 3 Propylene yield with time course over the A0.95CrO (A: La,
3−δ
catalysts deactivated at the 3.4 kPa steam partial pressure.
Pr, Nd) and CrO /γ-Al O catalysts
x
2
3
The Nd 0.95CrO catalyst displayed a 15.1 % propylene
3−δ
yield at 30 min, and the propylene yield gradually decreased
differenceꢀinꢀtheꢀactivityꢀwasꢀsmall.ꢀTheꢀsignificantꢀdiffer-
to 10.7% at 290 min. However, over the CrO/γ-Al O cata-
x
2
3
ence, however, appeared in the stability among these cata -
lyst, the propylene yield decreased by 8% during a 190-min-
ute reaction. By increasing the partial pressure of the steam
lysts. The propylene yields at 330 min over the La CrO ,
3−δ
0
.95
Pr0.95CrO and Nd 0.95CrO catalysts were 1.6, 7.2 and
to 6.8 kPa, the Nd 0.95CrO catalyst exhibited a stable
3−δ
3−δ
3−δ
9
.5%, respectively. Especially, the Nd 0.95CrO catalyst
activity during the reaction, while the CrO/γ-Al O catalyst
x 2 3
3−δ
had only a slightly decreased activity during the reaction.
slowly deactivated. A further change in the steam pressure
caused a decrease in the propylene yield over both catalysts,
although the stability was improved. Based on these results,
The reference CrO /γ-Al O catalyst showed a high propyl-
x
2
3
eneꢀyieldꢀofꢀ19.2ꢀ%ꢀatꢀ30ꢀmin,ꢀhowever,ꢀtheꢀactivityꢀsignifi-
cantly decreased to 3.8 % at 330 min. Based on the result,
it was concluded that the Nd0.95CrO catalyst had a supe-
rior stability for the DHP using steam at various partial pres-
3−δ
the Nd0.95CrO catalyst had a superior stability compared
3−δ
to the reference CrO /γ-Al O catalyst.
sures, compared to the CrO /γ-Al O catalyst.
x
2
3
x
2
3
1
3

6
R. Watanabe et al.
Table 1 The amount of carbon deposition over the A 0.95CrO (A:
where “d” is the length of the edge of the cubic crystallites
3−δ
La, Pr, Nd) and CrO /γ-Al O catalysts
x
2
3
ofꢀtheꢀloadedꢀphase,ꢀandꢀ“λ ” is the mean escape depth of
p
the photoelectrons in the loaded phase. The photoelectron
Catalyst
Deposited coke
amount (wt%)
cross sections and mean escape depths of the photoelec
tronsꢀusedꢀinꢀtheseꢀcalculationsꢀwereꢀtakenꢀfromꢀScofieldꢀ
and Penn, respectively [55, 56]. The particle size “d” of the
-
CrO /ꢀγ-Al O
6.76
0.33
0.09
0.03
0.35
x
2
3
La0.95CrO
3−γ
CrO was determined by solving Eq. (3). The surface area
x
Pr0.95CrO
3−γ
ofꢀtheꢀloadedꢀCrꢀonꢀtheꢀγ-Al O was estimated from the par-
2
3
Nd0.95CrO
3−γ
3−γ
ticle size of CrO . As a result of the calculation of the length
x
a
Nd0.95CrO
of the cubic crystallites over the catalyst, the mean particle
size was 1.9 nm. This value was close to that reported in
aDehydartion without steam for 2 h
the literature [57]. The surface area of the loaded CrO was
x
2
−1
In order to compare the turnover number of the propane
dehydrogenation on one Cr site (described as TON) over
estimated to be 75.5 m g-cat by the calculation using
the particle size. Assuming that the CrO structure was the
x
the Nd 0.95CrO catalyst to that over the CrO
/γ-Al O
corundum-type Cr O , the theoretical amount of the surface
2 3
3−δ
x
2
3
2
1
catalyst, the molecular number of Cr per g-cat on both
catalystsꢀwasꢀfirstꢀcalculated.ꢀWeꢀassumedꢀthatꢀtheꢀ(1ꢀ1ꢀ0)ꢀ per g-cat was then 4926 µmolg-cat . Based on the molecu-
face of the Nd 0.95CrO catalyst would be exposed on the lar number of the Cr on the Nd0.95CrO and CrO /γ-Al O
Cr was 2.97 × 10 per g-cat. The molecular number of Cr
−
1
3−δ
3−δ
x
2
3
surface of the Nd0.95CrO catalyst. Based on this assump-
catalyst, the TON was calculated using the propylene yield
at 30 min, which was described as TON (30 min). This
TON (30 min) is not an initial activity, but a slight deac-
tivated value during 30 min reaction. Figure 4a, b show
the TON (30 min). By varying the steam partial pressure
in the feed, various TONs (30 min) were calculated, and
these results are shown in Table 2. The TON (30 min) over
3−δ
−
20
2
tion, one Cr cation was present per 5.74 × 10 m , which
was the surface area for the (1 1 0) face on one unit cell
ofꢀperovskite.ꢀTheꢀspecificꢀsurfaceꢀareaꢀofꢀtheꢀNd0.95CrO
3−δ
2
−1
catalyst was 5.9 m
g-cat . Therefore, the theoretical
−
1
amount of the surface Cr cation was 170.7 µmol g-cat
−
20
6
(
= 5.9/(5.74 × 10 )× 10 ). As for the CrO /γ-Al O cata-
x
2
3
lyst, the surface area of the supported CrO was estimated
the Nd0.95CrO ꢀcatalystꢀwasꢀsignificantlyꢀhigherꢀthanꢀthatꢀ
3−δ
x
by the extension of the Defosse model proposed by Kerkhof over the CrO /γ-Al O catalyst at the various steam partial
x
2
3
and Moulijn [54]. The crystal size of CrO was estimated
pressures. Although the propylene yield is low due to the
low number of Cr sites on the surface, the reaction rate per
x
from the relative XPS intensity of the catalyst sample. First,
0
0
the theoretical XPS intensity ratio of I /I was calculated
Cr site of the Nd0.95CrO catalyst was found to be higher
p
s
3−δ
for the monolayer of CrO ꢀformationꢀonꢀγ-Al O . Here, “p”
than that of the CrO /γ-Al O catalyst. The reason for the
x
2
3
x
2
3
denotes the atomically dispersed loaded phase, i.e., Cr, and
s”ꢀrepresentsꢀtheꢀsupport,ꢀi.e.,ꢀγ-Al O . The experimental
higher TON (30 min) over the Nd 0.95CrO catalyst might
3−δ
“
be derived from the reaction mechanism. Weckhuysen et al.
reported the reaction mechanism of dehydrogenation of an
alkane over the CrO /γ-Al O catalyst, which was the Lang-
2
3
intensity ratio of I /I is given by the following expression:
p
s
x
2
3
0
I
I
I
p
p
muir–Hinshelwood type [10]. The C–H bond of the alkane
is broken, and a OH-bond and a Cr–C bond are formed over
(
3)=
{
1− exp( − d/λp )
}
/(d/λp )
0
I
s
s
Fig. 4 Effect of partial pressure
of steam on propylene yield
over the a Nd0.95CrO and b
3−δ
CrO /γ-Al O catalysts
x
2
3
1
3

Performance of Non-stoichiometric Perovskite Catalyst (A CrO3−δ, A: La, Pr, Nd) for Dehydrogenation...
7
x
Table 2 Turnover number
Catalyst
H2O
Propylene
TON
Selectivity (%)
and selectivity to products
over the Nd0.95CrO3-δꢀandꢀ
CrOx/γ-Al O catalysts
pressure production
(30 min)
−1
C H
C H
C H
CH4 CO
CO2
(kPa)
rate at 30 min
(min
)
3
6
2
4
2
6
2
3
−1
(μmolꢀmin
)
Nd0.95CrO
3.4
30.4
21.0
12.3
10.2
25.2
18.2
9.9
1.85
1.28
0.75
0.62
0.77
0.55
0.30
0.22
81.1
77.4
73.3
63.7
80.0
75.5
68.1
64.0
4.1
6.7
0
6.4 1.4
7.7 0.8
8.5 0.2
11.8 0.2
7.9 0.7
10.5 0.6
13.0 0.5
15.2 0.6
6.9
7.4
6.7
6.6
5.2
5.9
6.2
6.8
3−δ
6
.8
0
1
2
3.5
7.0
3.4
11.0
17.7
5.7
0.3
0
CrO /γ-Al O
3
0.5
0
x
2
6
.8
7.6
1
2
3.5
7.0
11.9
13.4
0.4
0
7.4
the CrO /γ-Al O catalyst. An alkene is then formed, and
the formed alkene is released from the catalyst surface. As
lattice parameter values for the Nd0.95CrO are a= 5.457Å,
x
2
3
3−δ
b=7.693 Å and c=5.443 Å, and the corresponding values for
for the Nd0.95CrO catalyst, propane might be oxidized to
the Nd CrO are a= 5.450Å, b=7.679 Å and c= 5.431Å.
3−δ
0.90
3−δ
propylene using lattice oxygen via the following Eq. (4).
The cell volumes for the Nd
CrO , Nd
CrO and
0.95 3−δ
1.00
3
Nd0.90CrO catalysts are respectively 229.0, 228.5 and
3−δ
C( 4 H) + O2
−
−
227.3 Å . The decreased cell volumes might be due to the
3
→ C H +V + H O 2+ e
3
8
lat
3
6
ox
2
creation of lattice vacancy and/or the increase of valence state
ofꢀ Crꢀ cation.ꢀ Generally,ꢀ theꢀA-siteꢀ deficiencyꢀ couldꢀ beꢀ theꢀ
creation of oxygen vacancies and/or the increase of valence
−
2−
H( 5 )O +V + 2e → O
+ H2
lat
2
ox
The produced oxygen vacancy was immediately regenerated stateꢀofꢀtheꢀB-siteꢀcationsꢀowingꢀtoꢀfulfillꢀtheꢀelectroneutralꢀ
by Eq. (5). Presumably, the perovskite with the A-site/B-site condition.ꢀForꢀtheꢀNd-deficientꢀcatalystꢀsuchꢀasꢀNd CrO
x
3−δ
<
1 had a high lattice oxygen mobility, which would pro-
duce a high activity for the DHP with steam. The different
possible mechanism might produce a superior activity and
(x<1), oxygen vacancy and/or tetravalent Cr species was
produced. The production of free volume of oxygen vacancy
and/or shrink of the size of Cr cation induce a lattice distor -
tion of perovskite, which could decrease a lattice constant and
a stable durability of the Nd 0.95CrO catalyst compared
to the CrO /γ-Al O ꢀcatalyst.ꢀInꢀorderꢀtoꢀconfirmꢀtheꢀredoxꢀ a cell volume. As for the Nd1.05CrO , the lattice parameters
property of the Nd0.95CrO catalyst, the surface electronic
state was investigated and reported in Sects. 3.3 and 3.4.
3−δ
x
2
3
3−δ
and cell volume are a = 5.537 Å, b= 7.803 Å, c= 5.459 Å
3−δ
3
and 235.9 Å . Such result might be caused by the decrease of
oxidation state of Cr cation. The oxidation state of Cr prob-
ably changed from the trivalent state to the divalent state. The
divalent state of Cr cation (0.73 Å) has a larger ionic diameter
that the trivalent state (0.615 Å). Therefore, the lattice expan-
sion proceeded, and the increased cell volume was observed.
3
.3 Effect of Nd/Cr Ratio on Morphology and
Structure of Nd CrO (0.90≤x≤ 1.05)
x
3−δ
Figure 5 shows FE-SEM images for the NdCrO (x: 0.90,
x
3−δ
0
.95, 1.00, 1.05) catalysts. From this Figure, the particle
diameter of the Nd CrO catalysts was about from 100
3.4 Relationship Between Redox Property and
x
3−δ
to 300 nm. There was no difference in the Nd CrO cata-
Dehydrogenation Activity over Nd CrO
x
3−δ
x
3−δ
lysts. In other words, the morphology of each catalyst was
(0.90≤x≤ 1.05)
not affected by the ratio of Nd to Cr.
Figure 6 shows wide and enlarged XRD patterns for
the Nd CrO . Table 3 shows structural parameters of the
To consider the reason for the difference in the activity among
the Nd CrO (0.90≤x≤1.05) catalysts, XPS measurements
x
3−δ
x
3−δ
Nd CrO catalysts. As shown in Fig. 6 , the peak shift to
were conducted. Before the XPS measurements, the N dC rO
x 3−δ
x
3−δ
lower angles was observed with an increase of the ratio
(0.90≤x≤1.05) catalysts were exposed to a H stream and/or
2
of Nd/Cr in the Nd xCrO . The result indicated that the
H with steam for 2 h. The binding energies of XPS results
2
3−δ
structural parameters were changed, depending on the Nd/
for the catalyst samples were referenced with the C
line
1s
Cr ratio. As shown in Table 3 , the calculated lattice param-
(284.7 eV) of the carbon overlayer. The full widths of the peak
at the half-maximum (FWHM) were allowed to change so as
toꢀattainꢀtheꢀbestꢀfitting.ꢀTheꢀFig.ꢀ7 showed a deconvoluted
O XPS spectra for the Nd CrO (0.90≤x≤ 1.05) catalysts
eters of the Nd 1.00CrO are a = 5.481 Å, b=7.709 Å and
3
c=5.421 Å, which are in good agreement with the report by
Hirota et al. [58]. These values of Nd
CrO are attributed
3
1.00
1s
x
3−δ
to the Pnma space group with orthorhombic structure. The
exposed to a H stream and/or H with steam for 2 h. In all the
2 2
1
3

8
R. Watanabe et al.
Fig. 5 FE-SEM image of a Nd0.90CrO , b Nd0.95CrO , c Nd1.00CrO and d Nd1.05CrO
3−δ
3−δ
3−δ
3
(a)
(b)
Table 3 Structural parameter of Nd xCrO catalyst (x: 0.90, 0.95,
1.00, 1.05)
3−δ
Nd1.05CrO3-δ
Catalyst
Unit cell parameter
Unit cell
volume
Nd1.00CrO
3
a (Å)
b (Å)
c (Å)
3
(Å )
Nd0.90CrO
Nd0.95CrO
5.450
5.457
5.481
5.537
7.679
7.693
7.709
7.803
5.431
5.443
5.421
5.459
227.3
228.5
229.0
235.9
Nd0.95CrO3-δ
Nd0.90CrO3-δ
3−δ
3−δ
Perovskite
Nd1.00CrO3
Nd1.05CrO
3−δ
20
30
40
50
60
θ / degree
70
80 32
34
2
was lower over the Nd 0.90CrO and Nd0.95CrO catalysts
3−δ
3−δ
Fig. 6 XRD patterns for a as-made Nd xCrO (x: 0.90, 0.95, 1.00,
compared to that over the Nd1.00CrO and Nd1.05CrO cata-
3−δ
3
3−δ
1
.05) and b enlarged part of the pattern
2−
−
2−
lysts. Compared the intensity ratio of O
to OH /CO
,
lat
3
which was described as I /I , a lower I /I ratio was obtained
1
2
1 2
O XPS spectra, the peaks for the lattice oxygen, hydroxyl
over the Nd 0.90CrO and Nd 0.95CrO catalysts. While,
the Nd1.00CrO and Nd 1.05CrO catalysts showed a higher
3 3−δ
1s
3−δ 3−δ
2−
group or carbonate and adsorbed H O (abbreviated as O lat,
2
−
2−
OH /CO and ads.-H O,ꢀ respectively)ꢀ wereꢀ confirmedꢀ atꢀ I /I ratio. The surface oxygen over the Nd
0.90CrO and
3
2
1
2
3−δ
529.5, 531.2 and 532.8 eV, respectively [59]. Judging from the Nd0.95CrO catalysts was removed by H based on the fol-
3−δ 2
position of each species, we think that the segregated species
lowing Eq. (6). Such a result implied that the reducibility of the
Nd0.90CrO and Nd0.95CrO catalysts might be high com-
is not present in the Nd CrO samples. In, addition, there
x
3−δ
3−δ
3−δ
is no difference in the hydroxyl group or carbonate among
pared to that of the Nd1.00CrO and Nd1.05CrO catalysts.
3
3−δ
2
−
the Nd CrO samples. The peak intensity for O
was
lat
x
3−δ
2
−
→Vox + H O 2+ e−
2
−
H + O
( 62 )
different for each other species. The peak intensity for O
lat
2
lat
1
3

Performance of Non-stoichiometric Perovskite Catalyst (A CrO3−δ, A: La, Pr, Nd) for Dehydrogenation...
9
x
O2-
O2-
(a)
(b)
-
2-
-
2-
OH /CO3
OH /CO3
ads. H O
2
ads. H O
2
I
1
1
/I
2
2
=1.88
I1
/I2
=1.80
Nd1.05CrO3-δ
Nd1.05CrO3-δ
Nd1.00CrO3
I
/I
=1.66
I1
I1
I1
/I2
/I2
/I2
=1.62
=1.15
=1.22
Nd1.00CrO3
Nd0.95CrO3-δ
Nd0.90CrO3-δ
Nd0.95CrO3-
δ
I
1
1
/I
2
2
=2.08
=1.95
Nd0.90CrO3-δ
I
/I
Binding energy / eV
Binding energy / eV
Fig. 7 XPS spectra of O1s for Nd CrO (0.90≤x≤ 1.05) undera H atmosphere and b H with steam atmosphere
x
3−δ
2
2
where, V denotes the oxygen vacancy. In the case of Hwith
property might bring a no-coke deposition (Table 1) in the
ox
2
2
−
steam as shown in Fig. 7b, the peak intensity for O
was
DHP-S over the Nd0.95CrO catalyst.
lat
3−δ
high for all the catalysts. The I /I ratio over the Nd1.00CrO3
1
2
and Nd 1.05CrO catalysts did not change. While, the
3−δ
I /I ratio over the Nd0.90CrO and Nd0.95CrO catalysts
4 Conclusions
1
2
3−δ
3−δ
was increased by feeding H O. The result implied that the
2
reduced Nd0.90CrO and Nd0.95CrO catalysts were oxi-
The non-stoichiometric A 0.95CrO (A: La, Pr, Nd)
perovskite-type oxide showed a high activity for the DHP
3−δ
3−δ
3−δ
dized by H O based on the following Eq. (7):
2
under steam. Especially, the Nd 0.95CrO catalyst had the
3−δ
H( 7 )O + 2e +V → O2− + H2
−
best performance among the A 0.95CrO and exhibited a
2
ox
lat
3−δ
superior stability to the reference CrO /γ-Al O catalyst.
x
2
3
JudgingꢀfromꢀtheꢀobtainedꢀXPSꢀprofiles,ꢀtheꢀNd0.90CrO
The XPS spectra revealed that the redox property was
improved by changing the Nd/Cr molar ratio from the stoi-
chiometric composition (Nd/Cr=1) to a non-stoichiometric
composition (Nd/Cr < 1). The Nd CrO catalyst had a
3−δ
and Nd0.95CrO catalysts might have the superior reduc-
3−δ
tion–oxidation (redox) property compared to the Nd CrO
1
.00
3
and Nd1.05CrO catalysts. Such a difference in the redox
3−δ
0.95
3−δ
property might produce a high performance for the DHP-S
high redox property, which would produce a high activity
over the Nd0.90CrO and Nd0.95CrO catalysts.
for the DHP with steam.
3−δ
3−δ
Perovskite-type oxide was known to the non-stoichiome-t
ric oxide as long as satisfying the electroneutral condition
and the tolerance factor. In other words, oxygen vacancy
could be present in the structure. From Table 3, oxygen
Acknowledgmentsꢀ Thisꢀ studyꢀ wasꢀ financiallyꢀ supportedꢀ byꢀ JSPSꢀ
KAKENHIꢀGrant-in-AidꢀforꢀScientificꢀResearchꢀ(A)ꢀ(26709059).
vacancy seemed to exist in the structure of the Nd xCrO
3−δ
References
(x<1) catalysts, as described in Sect. 3.3. Oxygen vacan -
cies and/or tetravalent B-site cation promotes a diffusion of
lattice oxygen, which enhances a redox property [60–65].
1
. Corma A, Melo FV, Sauvanaud L, Ortega F (2005) Catal Today
07–108:699–706
1
On the other hand, for the Nd xCrO (x >1) catalyst, the
2. Plotkin JS (2005) Catal Today 106:10–14
3. Ren T, Patel M, Blok K (2006) Energy 31:425–451
3−δ
2
+
lower oxidation state of Cr cation (Cr ) caused a decrease
of redox performance. Such a difference in a redox prop-
erty was considered to be resulted in the difference in the
catalytic performance between the Nd CrO (x < 1) and
4
5
. Basu B, Kunzru D (1992) Ind Eng Chem Res 31(1):146–155
. Verstraete J, Coupard V, Thomazeau C, Etienne P (2005) Catal
Today 106:62–71
6. O’Connor P, Verlaan JPJ, Yanik SJ (1998) Catal Today
3:305–313
x
3−δ
4
Nd CrO (x >1) catalysts. In addition, the high redox
x
3−δ
1
3

1
0
R. Watanabe et al.
7
8
. Bhasin MM, McCain JH, Vora BV, Imai T, Pujadó PR (2001)
Appl Catal A 221:397–419.
. Mol JC (2004) J Mol Catal A 213:39–45
38. Teraoka Y, Zhang HM, Okamoto K, Yamazoe N (1988) Mater
Res Bull 23:51–58
39. Zhang HM, Shimizu Y, Teraoka Y, Miura N, Yamazoe N (1990) J
Catal 121:432–440
40. Peña MA, Fierro JLG (2001) Chem Rev 101:1981–2017
41. Yi G, Hayakawa T, Andersen AG, Suzuki K, Hamakawa S, York
APE, Shimizu M, Takehira K (1996) Catal Lett 38:189–195
42. Shimizu T (1992) Catal Rev Sci Eng 34:355–371
43. Zhang HM, Teraoka Y, Yamazoe N (1998) Appl Catal 41:137–146
44. Zhang HM, Teraoka Y, Yamazoe N (1989) Catal Today 6:155–162
45. Nakamura T, Misono M, Yoneda Y, Eguchi K, Seiyama T (1983)
J Catal 83:151–159
ꢀ
9.ꢀ CavaniꢀF,ꢀKoutyrevꢀM,ꢀTrifiròꢀF,ꢀBartoliniꢀA,ꢀGhislettiꢀD,ꢀIezziꢀR,ꢀ
Santucci A, Del Piero G (1996) J Catal 158:236–250
1
0. Weckhuysen BM, Schoonheydt RA (1999) Catal Today
5
1:223–232
1. Puurunen RL, Weckhuysen BM (2002) J Catal 210:418–430
2. De Rossi S, PiaCasaletto M, Ferraris G, Cimino A, Minelli G
1998) Appl Catal A 167:257–270
1
1
1
1
(
3. Rombi E, Cutrufello MG, Solinas V, De Rossi S, Ferraris G, Pis-
tone A (2003) Appl Catal A 251:255–266
4. Bartholomew CH, Ferrauto RJ (2005) Fundamentals of industrial
catalytic processes, 2nd edn
46. Gallagher PK, Johnson DW Jr, Vogel EM (1977) J Am Ceram Soc
60:28–31
1
1
1
1
1
2
2
2
5. Chaar MA, Patel D, Kung HH (1988) J Catal 109:463–467
6. Corma A, Nieto JML, Paredes N (1993) J Catal 144:425–438
7. Pak C, Bell AT, Tilley TD (2002) J Catal 206:49–59
8. Mamedov EA, Cortés-Corberan V (1995) Appl Catal A 127:1–40
9. Andersson SLT (1994) Appl Catal A 112:209–218
0. Creaser D, Andersson B (1996) Appl Catal A 141:131–152
1. Stern DL, Grasselli RK (1997) J Catal 167:560–569
2. Sautel M, Thomas G, Kaddouri A, Mazzocchia C, Anouchinsky
R (1997) Appl Catal A 155:217–228
47. Yasuda H, Fujiwara Y, Mizuno N, Misono M (1994) J Chem Soc
Faraday Trans 90:1183–1189
48. Mizuno N, Fujiwara Y, Misono M (1989) J Chem Soc Chem
Commun 90:316–318
49. Ahn K, Kim H, Chung YC, Kim HR, Son JW, Lee HW, Lee JH
(2010) Appl Catal A 387:203–208
50. Tien-Thao N, Zahedi-Niaki MH, Alamdari H, Kaliaguine S
(2007) J Catal 245 348–357
51. Yang W, Zhang R, Chen B, Bion N, Duprez D, Royer S (2012) J
Catal 295:45–58
52.ꢀ UllmannꢀH,ꢀTrofimenkoꢀN,ꢀTietzꢀF,ꢀStöverꢀD,ꢀAhmad-KhanlouꢀAꢀ
(2000) Solid State Ionics 138:79–90
53. Ge L, Zhou W, Ran R, Liu S, Shao Z, Jin W, Xu N (2007) J
Membr Sci 306:318–328
54. Kerkhof FPJM, Moulijn JA (1979) J Phys Chem 83:1612–1619
55.ꢀ ScofieldꢀJHꢀ(1976)ꢀJꢀElectronꢀSpectroscꢀRelatꢀPhenomꢀ8:129–137
56. Penn DR (1976) J Electron Spectrosc Relat Phenom 9:29–40
57. Park PW, Ledford JS (1997) Langmuir 13:2726–2730
58. Hirota K, Kunifusa Y, Yoshinaka M, Yamaguchi O (2002) Mater
Res Bull 37:2335–2344
2
2
2
2
2
2
2
3. Chen K, Iglesia E, Bell AT (2000) J Catal 192:197–203
4. Blasko T, López Nieto JM (1997) Appl Catal A 157:117–142
5. Kung HH (1994) Adv Catal 40:1–38
6.ꢀ AlbonettiꢀS,ꢀCavaniꢀF,ꢀTrifiroꢀFꢀ(1996)ꢀCatalꢀRevꢀ38:413–438
7. Chen K, Xie S, Bell AT, Iglesia E (2001) J Catal 198:232–242
8. Chen K, Xie S, Bell AT, Iglesia E (2000) J Catal 195:244–252
9. López Nieto LM, Soler J, Concepción P, Herguido J, Menéndez
M, Santamaría J (1999) J Catal 185:324–332
3
3
0. Arene F, Frusteri F, Parmaliana A (1999) Catal Lett 60:59–63
1. Albrecht S, Wendt G, Lippold G, Adamski A, Dyrek K (1997)
Solid State Ionics 101:909–914
3
2.ꢀ Blascoꢀ T,ꢀ Galliꢀ A,ꢀ Lópezꢀ Nietoꢀ JM,ꢀ Trifiróꢀ Fꢀ (1997)ꢀ Jꢀ Catalꢀ
59. Ponce S, Peña MA, Fierro JLG (2000) Appl Catal B 24:193–205
60. Mineshige A, Izutsu J, Nakamura M, Nigaki K, Abe J, Kobune M,
Fujii S, Yazawa T (2005) Solid State Ionics 176:1145–1149
61. Zhou W, Ran R, Shao ZP, Jin WQ, Xu NP (2008) J Power Sources
182:24–31
1
69:203–211
3
3
3. Abello MC, Gomez MF, Cadus LE (1998) Catal Lett 53:185–192
4. Grabowski R, Grzybowska B, Samson K, Sloczynski J, Stoch J,
Wcislo K (1995) Appl Catal A 125:129–144
3
3
3
5. Watanabe R, Hondo Y, Mukawa K, Fukuhara C, Kikuchi E,
Sekine Y (2013) J Mol Catal A 377:74–84
6. Voorhoeve RJH, Remeika JP, Trimble LE (1976) Ann N Y Acad
Sci 272:3–21
62. Hansen KK, Hansen KV (2007) Solid State Ionics 178:1379–1384
63. Yang ZB, Yang CH, Jin C, Han MF, Chen FL (2011) Electrochem
Commun 13:882–885
64. Pang SL, Jiang XN, Li XN, Wang Q, Su ZX, Zhang QY (2012)
Int J Hydrog Energy 37:3998–4001
7. Nitadori T, Ichiki T, Misono M (1988) Bull Chem Soc Jpn
6
1:621–626
65. Konysheva EY, Xu XX, Irvine JTS (2012) Adv Mater 24:528–532
1
3