Dehydrogenation of propane over In2O3-Al 2O3 mixed oxide in the presence of carbon dioxide

Article abstract of DOI:10.1016/j.jcat.2010.03.007

Dehydrogenation of propane (PDH) to propylene in the presence of CO ₂ was performed over binary In₂O₃-Al ₂O₃ mixed oxides prepared by alcoholic coprecipitation. Synergistic effect of different composition on the catalytic performance has been observed. Characterization by X-ray diffraction and X-ray photoelectron spectroscopy revealed that the combination of the two components can result in significant modification in the surface properties and bulk dispersion of the In₂O₃ phase, which in turn leads to a higher reducibility of the In-Al-O composite. A correlation between the temperature-programmed reduction data and the PDH activity for In₂O₃-Al ₂O₃ revealed that the creation of surface metallic indium species during the reaction is crucial for the dehydrogenation performance. The promoting effect of CO₂ on the yield of propylene has been attributed to a facilitation of simple dehydrogenation by its coupling with the reverse water gas shift reaction based on the results from temperature-programmed reaction of CO₂ with H₂.

Full text of DOI:10.1016/j.jcat.2010.03.007

Journal of Catalysis 272 (2010) 101–108
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Dehydrogenation of propane over In O –Al O mixed oxide in the presence
2
3
2 3
of carbon dioxide
Miao Chen, Jie Xu, Yong Cao , He-Yong He, Kang-Nian Fan, Ji-Hua Zhuang **
*
Department of Chemistry, Shanghai Key Laboratory of Molecular Catalysis and Innovative Materials, Fudan University, Shanghai 200433, PR China
a r t i c l e i n f o
a b s t r a c t
Article history:
Dehydrogenation of propane (PDH) to propylene in the presence of CO2 was performed over binary
Received 16 December 2009
Revised 3 March 2010
Accepted 5 March 2010
Available online 3 April 2010
2 3 2 3
In O –Al O mixed oxides prepared by alcoholic coprecipitation. Synergistic effect of different composi-
tion on the catalytic performance has been observed. Characterization by X-ray diffraction and X-ray pho-
toelectron spectroscopy revealed that the combination of the two components can result in significant
modification in the surface properties and bulk dispersion of the In
higher reducibility of the In–Al–O composite. A correlation between the temperature-programmed
reduction data and the PDH activity for In –Al revealed that the creation of surface metallic indium
species during the reaction is crucial for the dehydrogenation performance. The promoting effect of CO
2 3
O phase, which in turn leads to a
Keywords:
Propane
Propylene
Dehydrogenation
2
O
3
2 3
O
2
on the yield of propylene has been attributed to a facilitation of simple dehydrogenation by its coupling
with the reverse water gas shift reaction based on the results from temperature-programmed reaction of
2 3 2 3
In O –Al O
Carbon dioxide
In
2 2
CO with H .
0
Ó 2010 Elsevier Inc. All rights reserved.
1
. Introduction
Dehydrogenation of light alkanes, such as ethane and propane,
the conventional direct dehydrogenation. The ODH gives the follow-
ing benefits: the reaction is irreversible and exothermic; process can
be carried out at lower temperatures without carbon deposition
deactivating the catalyst [12–14]. However, using oxygen as an oxi-
dant brings several problems to the reaction [15], such as flammabil-
ity, necessity of heat removal and deep oxidation, which reduces the
process selectivity due to the production of undesired carbon oxides
[16] and is hard to control. Therefore, it is expected that replacing
into the corresponding alkenes has attracted significant attention
since the alkenes constitute an important and versatile class of
intermediates in the chemical industry [1,2]. At present, propylene
is typically obtained by catalytic dehydrogenation of propane
(
PDH) [3,4] and as a by-product of steam cracking or a fluidized
catalytic cracking (FCC) of naphtha. UOP and AAB Lummus Global,
two of the world leaders of catalytic dehydrogenation technology,
have licensed eight commercial process units [5], which currently
can provide propylene with an aggregate capacity of about 2.1 mil-
lion metric tons per year. Dehydrogenation of propane, however, is
an equilibrium-limited and endothermic reaction [6,7]. For ther-
modynamics reasons, significant yield of alkene requires operation
at relatively high temperatures and thus high-energy costs. More-
over, such process conditions favor undesired thermal cracking
reactions to lighter hydrocarbons and notoriously rapid catalyst
deactivation due to coke formation [8,9].
oxygen in the ODH reactions by soft oxidants, such as N
[17–20], will suppress these disadvantages.
2 2
O and CO
Over the last decade, there has been a growing interest in the
use of CO as a milder and safer oxidant in the ODH of propane
[6,7,16,20–26]. It is worth noting that the dehydrogenation of pro-
pane in the presence of CO can proceed by two different reaction
pathways. The first is a direct process, in which CO promotes the
2
2
2
dehydrogenation through re-oxidizing the catalyst surface as re-
duced by propane. This one-step pathway is in agreement with
the Mars–Van Krevelen mechanism [16,22,24]. The second one,
known as a two-step pathway, is a simple dehydrogenation fol-
lowed by the reverse water gas shift (RWGS) reaction [7,25,26].
An alternative to this process is the oxidative dehydrogenation
(
ODH) of propane by molecular oxygen [10,11], which has attracted
2
In this case, CO removes hydrogen from the reaction system and
increasing attention due to its higher efficiency of using light alkane
resource with minimum energy consumption when compared to
releases the thermodynamic limitation of equilibrium. Many re-
sults indicate that propylene is formed by both presented mecha-
nisms [20]. Moreover, in the dehydrogenation process, carbon
dioxide can act as a diluent [27], which delivers the required heat
and reduces coking of catalyst by coke gasification.
*
*
Corresponding author. Fax: +86 21 65643774.
* Corresponding author. Fax: +86 21 65643774.
To date, the catalysts suggested for propane dehydrogenation
E-mail addresses: yongcao@fudan.edu.cn (Y. Cao), jihuaz@fudan.edu.cn (J.-H.
Zhuang).
2
with CO are mainly based on supported (transition) metal oxides
0
021-9517/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2010.03.007

1
02
M. Chen et al. / Journal of Catalysis 272 (2010) 101–108
that allow effective hydrocarbon activation at temperatures signif-
icantly lower than conventional dehydrogenation process. Among
all the studied catalysts, supported gallium oxides have recently
been recognized as one of the most promising materials due to
their excellent catalytic efficiency as compared to conventional
X-ray photoelectron spectroscopy (XPS) data were recorded on
a Perkin Elmer PHI 5000C ESCA system with Mg K excitation line
(h = 1253.6 eV). The binding energies (BE) were referenced to the
C 1s peak of contaminant carbon at 284.6 eV with an uncertainty of
±0.2 eV. Elemental analysis was performed using ion-coupled plas-
ma (ICP) atomic emission spectroscopy on a Thermo Electron IRIS
Intrepid II XSP spectrometer. Temperature-programmed reduction
(TPR) results were obtained on a homemade apparatus loaded with
20 mg of catalyst. The samples were pretreated in flowing nitrogen
at 773 K for 1 h. After cooled to room temperature, the samples
a
m
Cr
ciated with the Ga
tic deactivation during a few hours on stream [29,30]. Very
recently, we have demonstrated that Ga –Al mixed oxides
are highly active and stable catalysts for PDH in the presence of
CO [25,26]. The superior performance of the Ga –Al system
has been attributed to the formation of gallia-alumina solid solu-
tion between Ga and Al [26], which allows the favorable cre-
ation of a high abundance of coordinatively unsaturated surface
2
O
3
- or Pt-based systems [7,28–33]. One critical limitation asso-
2 3
O
-catalyzed PDH process, however, is the dras-
2
O
3
2 3
O
2
2
O
3
2
O
3
were subsequently contacted with an H
2 2
/Ar mixture (H /Ar molar
ꢁ1
ratio of 5/95 and a total flow of 40 mL min ) and heated at a rate
ꢁ
1
2
O
3
2
O
3
2
of 5 K min , to a final temperature of 1073 K. The H consumption
was monitored by an on-line thermal conductivity detector. Since
water is produced during reduction, the gas exiting from the reac-
tor was passed through a cold trap before entering the thermal
conductivity detector.
3
+
Ga sites with weak Lewis acidity. However, the development of
new improved catalytic system exhibiting desirable stability and
activity still remains a major challenge.
In continuation of our interest in exploring the applications of
mixed oxides containing group III elements (Al, Ga and In) in
propane dehydrogenation [25,26], herein, we report for the first
time the use of In–Al mixed oxide as a new efficient catalyst in
2
.3. Activity measurement
Catalytic tests were performed in a fixed-bed flow microreactor
the dehydrogenation of propane to prepare propylene. In
Al mixed oxides, previously established as an excellent sys-
tem for selective catalytic reduction (SCR) of NO owing to its
unique low-temperature redox properties [34–38], have shown
to be particularly efficient for PDH in the presence of CO . Our
2 3
O –
at atmospheric pressure, and the catalyst amount was 200 mg.
Nitrogen was used as the carrier gas at a flow rate of 10 mL min
The catalysts were pretreated at 873 K for 1 h in nitrogen flow, and
the reaction temperature was 873 K. During reaction, the gas reac-
2
tant contained 2.5 vol% propane, 5–30 vol% CO , and a balance of
ꢁ1
2 3
O
.
x
2
results have shown that the creation and maintenance of the
catalytically active metallic indium species in optimal surface
concentration is indispensable for an efficient dehydrogenation
of propane.
nitrogen. The feed and the reaction products were analyzed using
an on-line gas chromatograph (Type GC-122, Shanghai) equipped
with a 6-m packed column of Porapak Q and a flame ionization
detector (FID). The permanent gas products, including N
2
, CO,
and CO , were analyzed on-line by another GC equipped with a
2
TDX-01 column and a TCD detector. All carbon balances closed
within 92–95%. The conversion and selectivity were calculated as
follows:
2
. Experimental
2.1. Catalyst preparation
C
3
H
8
ꢁ C
3
H
8
out
in
C
3
3
H
H
8
6
conversion ¼
selectivity ¼
ꢂ 100%
ꢂ 100%
A series of mixed In
positions as well as the simple oxide of Al
pared through a previously developed alcoholic coprecipitation
pathway [39]. In a typical synthesis, concentrated aqueous ammo-
nia and ethanol (50:50 in volume) was added dropwise to the eth-
2
O
3
–Al
2
O
3
oxide catalysts with various com-
and In were pre-
C
3 8
H
in
2
O
3
2 3
O
3 6
C H
out
C
C
3
H
8
ꢁ C
3 8
H
out
in
anol solution of indium nitrate hydrate (In(NO
3
)
3
ꢀ5H
ꢀ9H
9.9%) with different In:Al molar ratio until no more precipitation
2
O, Aldrich,
2 2
2.4. Temperature-programmed reaction of CO with H
9
9
9.99%) and aluminum nitrate hydrate (Al(NO
3
)
3
2
O, Fluka,
A temperature-programmed CO –H test was carried out to
2
2
occurred (pH = 8.5). The resulting gel was quickly filtered and
washed by ethanol, dried at 373 K overnight, and finally calcined
at 873 K for 6 h. The final catalyst is denoted as In–Al-n hereinafter
investigate the activity of the In–Al mixed oxide catalysts for the
RWGS reaction. The measurements were performed in a flow mic-
roreactor system. Prior to the experiment, a sample (100 mg) was
outgassed in a flow of pure helium at 873 K for 30 min. Subse-
quently, the microreactor was cooled down to about 443 K. Reac-
2 3
where n represents the mole percentage of In O .
tion was performed using CO
of 1:1:4 and total flow rate of 12 mL min . The experiment was
2 2
/H /He gas mixture with the ratio
2
.2. Catalyst characterization
ꢁ1
ꢁ1
carried out with a linear heating rate of 10 K min up to the final
temperature of 973 K. The reactants and all possible products of
reaction were continuously monitored by the QMS detector (Bal-
zers OmniStar). The signal of the helium line served as the internal
standard to compensate fluctuations of the operating pressure.
The BET-specific surface areas of the samples were determined
by adsorption–desorption of nitrogen at liquid nitrogen tempera-
ture, using a Micromeritics TriStar 3000 equipment. The X-ray
powder diffraction (XRD) of the catalysts was carried out on a Bru-
ker D8Advance X-ray diffractometer using nickel-filtered Cu K
a
radiation at 40 kV and 20 mA. The acidic and basic properties of
each catalyst were characterized by temperature-programmed
3. Results and discussion
3 2
desorption (TPD) of NH and CO , respectively [26]. In a typical
experiment for TPD studies, about 200 mg of the oven-dried sam-
ple (dried at 383 K overnight) was placed in a U-shaped quartz
3.1. Structural characterization and acid–base properties
sample tube. NH
ment at 773 K in a He stream. The NH
determined by gas chromatography at temperatures from 413 to
3
(or CO
2
) was saturated at 413 K after pre-treat-
Fig. 1 illustrates the diffractograms obtained for various In
mixed oxides, along with those for bulk In and Al
references. The as-synthesized alumina was poorly crystallized,
which is a common feature for -alumina [39], and even poorer
crystalline signals of alumina are observed in the mixed oxide sam-
2 3
O –
O as
2 3
3
(or CO ) desorbed was
2
2
Al O
3
O
2 3
ꢁ
1
8
73 K at a ramp rate of 10 K min and subsequently maintained
c
at 873 for 20 min.

M. Chen et al. / Journal of Catalysis 272 (2010) 101–108
103
3.2. Redox behaviors and XPS studies
e
d
The reducibility of indium species in the In
2
O
3
–Al
2
O
3
catalysts
was investigated by TPR experiments. As shown in Fig. 2, the TPR
profile of bulk In exhibits a single and broad peak centered at
ca. 1000 K, attributable to the characteristic reduction of crystal-
2 3
O
0
line In
2 3
O to In [34,35]. The hydrogen consumption is calculated
ꢁ1
to be 10.3 mmolH2
g
, corresponding to ca. 95% of the complete
c
reduction of indium oxide. Distinct profile changes are identified
for the reduction of In–Al mixed oxide samples, where two distinct
reduction features are observed. The main reduction peak (b) as-
2 3
cribed to the reduction of crystalline phase In O is shifted to a
lower temperature of ca. 970 K in the In–Al mixed oxides, indica-
tive of the presence of a In/Al interaction which facilitates the
b
reduction of the dispersed In
observations in their in-depth study on the nature of supported
In catalysts for the SCR of NO and also compared the reduction
of indium on various supports [38]. And it was noted that the dis-
persed InO species on the support are more reducible than the
bulk In . One can also see that with increasing alumina incorpo-
ration, the fraction of low temperature (LT) reduction peak ( ) in
2 3
O . Gervasini et al. reported similar
(
222)
2
O
3
x
x
2 3
O
a
a
(440)
(
400)
the temperature range of 473–773 K, attributable to the reduction
of dispersed indium oxide phase with smaller particles sizes to
metallic In0 species [34,35,38], is observed to be the highest for
(622)
1
2
(211)
×
(431)
sample In–Al-10 when compared to other In
Table 1). This observation further confirms the presence of at least
two types of In phase in the In –Al samples, where crys-
tallized indium oxide (b peak) and In
2 3 2 3
O –Al O samples
(
20
30
40
50
60
70
o
2
O
O
2 3
O
3
2
3
2θ
2
O
3
phase highly dispersed on
Fig. 1. XRD patterns of various In
2
O
3
–Al
2
O
3
mixed oxides: (a) In
2
O
3
; (b) In–Al-40;
alumina (
a
peak) co-exist.
(
c) In–Al-20; (d) In–Al-10; (e) Al O .
2 3
β
ples. In contrast, the bulk In
The peak intensities and their 2h angles (30.6°, 35.6°, 51.3°, and
0.9°) have been identified as characteristic of the cubic structure
of In (c-In , JCPDS 6-0416). For all mixed oxide samples, ex-
cept in the case of the In–Al-10 sample, the diffraction peaks cor-
responding to c-In phase are observed. The diffraction peaks
corresponding to crystalline indium oxide phase of all In–Al–O
samples are much weaker and broader than those of bulk In
pointing to apparently higher component dispersion of the
In –Al composites with higher aluminum content. However,
2 3
O presents a well-crystallized phase.
α
6
e
d
2
O
3
2 3
O
2 3
O
2 3
O ,
2
O
3
2 3
O
the positions of the diffraction lines for the mixed oxide samples
were still the same as for the bulk indium oxide, indicating no Al
substitution with In due to the relatively large difference of the
atomic radii of In (0.79 Å) and Al (0.53 Å) atoms [40]. This fact, tak-
ing together with the significantly decreased crystallinities for
c
2 3 2 3
Al O and In O , suggests that the incorporation of alumina has
not affected the crystalline structure of indium oxide but rendered
better dispersions of both oxides.
Of all the mixed oxides investigated, the In–Al-10 catalyst
shows the highest component dispersion, as reflected from the
totally amorphous nature of this sample. This is also expected
b
a
2
ꢁ1
as a result of the highest specific surface area (195 m g , see Ta-
ble 1) of In–Al-10 when compared to other In –Al samples.
Surface acidity and basicity measurements by NH - and CO -TPD
2
O
3
2 3
O
1
×
3
2
measurements (see Supplementary material, Tables S1 and S2
and Figs. S1 and S2) reveal an abundance of surface basic sites
of medium strength and limited number of weak and medium
strong acid sites in the bulk indium oxide. However, a much high-
er population of acidic sites than that of basic sites can be ob-
served for the In–Al mixed oxide samples. It is worthwhile to
note that the introduction of aluminum significantly modifies
the acid–base distribution, increases both the overall acidity and
basicity of indium oxide.
2
3
73
573
773
Temperature (K)
973
1173
Fig. 2. TPR profiles of various In
In–Al-20; (d) In–Al-10; (e) In–Al-20 pre-reduced by 5 vol% H
by CO exposure at 873 K for 2 h.
2
O
3
–Al
2
O
3
mixed oxides: (a) In
2
O
3
; (b) In–Al-40; (c)
/Ar at 773 K followed
2
2

1
04
M. Chen et al. / Journal of Catalysis 272 (2010) 101–108
Table 1
2 3 2 3
Physicochemical properties and characterization results of the mixed In O –Al O oxides.
BET (m2
g
ꢁ1
)
B.E. (eV) In 3d5/2
In/Al molar ratio
Peak temperature (K)
H
consumption (mmol g
ꢁ1
)
2 3
Percentage of In O (a
) e (%)
Sample
S
2
Bulka
Surfaceb
a
c
bd
a
c
bd
In
2
O
3
23
94
174
195
233
444.7
444.6
444.5
444.6
–
–
–
–
1026
980
976
963
–
0
10.3
6
3.2
1.4
–
0
In–Al-40
In–Al-20
In–Al-10
0.67
0.25
0.11
–
0.72
0.27
0.13
–
566
573
578
–
0.6
1.6
1.2
–
9.1
33
46
–
2 3
Al O
a
b
c
The bulk In/Al molar ratio calculated from the ICP data.
The surface In/Al molar ratio based on XPS analysis.
The hydrogen consumption during 423–773 K calculated from the TPR results.
The hydrogen consumption during 773–1073 K calculated from the TPR results.
d
e
2 3 2 3
The percentage of In O reduced during 423–773 K in TPR from the total amount of In O reduced during the whole temperature range.
XPS experiments make it possible to determine not only the
2). This result agrees well with that (33%, in Table 1) of its likely
precursor (highly dispersed In ) as estimated based on TPR re-
sults, implying all the existing In derive exclusively from the
highly dispersed In rather than bulk ones. TEM experiments
oxidation state, but also the In/Al ratio in the surface. These results
are summarized in Table 1. In all cases, only one oxidation state
was found, with binding energy for In 3d5/2 between 444.5 and
2 3
O
0
2 3
O
4
44.7 eV. These values are characteristic of In
2
O
3
[41]. Considering
have been carried out to observe a possible structure of the metal-
lic In0 particles. However, the identification of the small indium
particles was unsuccessful, possibly due to the poor contrast be-
tween In and metal oxides in the reduced sample. When this par-
2
tially reduced sample is further subjected to an exposure of CO at
that XPS is a surface sensitive technique, the In/Al ratio gives an
idea of the component dispersion. The surface In/Al ratios are in
line with the ICP data (see Table 1), confirming a high component
dispersion of all mixed oxide samples. To gain further insight into
the chemical state of indium species during reduction and the pos-
0
873 K, no spectral variation could be observed (Fig. 3c). This indi-
cates that the preformed metallic In0 cannot be re-oxidized by
2
sible influence of CO exposure upon metallic indium, XPS studies
on In–Al-20 subjected to various atmospheric treatments were car-
ried out (see Fig. 3 and Table 2). In the case of the fresh In–Al-20
sample (Fig. 3a), only bands assigned to In(III) species could be ob-
2
CO under reaction conditions, possibly due to the relatively weak
oxidant potential of this molecule [16].
served. After reduction treatment at 773 K (Fig. 3b), an additional
3.3. Propane dehydrogenation activity
0
feature signaling the formation of metallic In (B.E. In 3d5/2
=
4
43.5 eV) species appeared. The In0 accounted for 39.8 (atom%)
of the total indium species based on deconvolution analysis (Table
The dehydrogenation of propane to propylene over the In-con-
taining materials along with Al O in the absence of CO was inves-
2 3 2
tigated at 873 K. The major product formed in the reaction is
propylene, and the minor products are ethane, ethylene, and meth-
ane. The results, shown in Fig. 4, point to a marked composition ef-
a
b
2 3 2 3
fect on the catalytic performance of the In O –Al O samples. Both
simple oxides of In and Al demonstrated low propane con-
2
O
3
2 3
O
version of <10%, which is fully in line with their inferior dehydro-
genation activities as reported by Nakagawa et al. [21]. It is
apparent moreover that all In-containing samples exhibited a very
low selectivity toward propylene formation in the initial stage of
the reaction, suggesting that an induction period is required to de-
velop the active site. It is of interest to note that this performance
2 3
evolution is quite different from all previously reported Cr O -,
4
55 450 445 440
455
450
445
440
Table 2
Summary of XPS studies.
a
c
d
Sample Sample description
B.E. for In
5/2 (eV)
B.E. for In
3d3/2 (eV)
Percentage
of
3
d
In3+
In0
In3+
In0
reduction
(
%)
A
B
fresh In–Al-20
444.5
–
452.2
–
–
Sample A reduced
at 773 K for 1 h by
444.5 443.5 452.2 451.2 39.8
H
2
/Ar mixtureb
C
Sample B
444.5 443.5 452.2 451.2 39.7
subsequently
treated with
ꢁ
1
4
55
450
445
440
455
B.E. (eV)
450
445
440
10 mL min CO
73 K for 1 h
In–Al-20 reacted for 444.5 443.5 452.2 451.2 39.4
h on stream in the
absence of CO
2
at
8
D
8
2 3 2 3
Fig. 3. XPS results of various In O –Al O catalysts (a) fresh In–Al-20; (b) In–Al-20
2
pre-reduced by H
for 1 h followed by exposure to a CO
2
/Ar at 773 K for 1 h; (c) In–Al-20 pre-reduced by H
2
/Ar at 773 K
stream (10 mL min 1) at 873 K for 1 h; (d) In–
ꢁ
a
Note that the as-analyzed results could be in doubt as the measurement is not
performed in situ.
2
Al-20 reacted for 8 h on stream at 873 K in the presence of CO
conditions: P(C ) = 2.5 kPa; P(N ) = 97.5 kPa; total flow rate: 10 mL min ).
2
(Reaction
ꢁ
1
b
ꢁ1
3
H
8
2
2
H /Ar molar ratio of 5/95 and a total flow of 40 mL min .

M. Chen et al. / Journal of Catalysis 272 (2010) 101–108
105
1
00
a
b
9
0
0
0
0
0
0
0
0
2
1
0
0
8
7
6
5
4
3
2
2
In O3
In-Al-40
In-Al-20
In-Al-10
Al O
2
3
pretreated
In-Al-20
0
10
0
2
4
6
8
10 12
0
2
4
6
8
10 12
Reaction time (h)
Fig. 4. Conversion of propane and selectivity to propylene for In
reaction temperature: 873 K; total flow rate: 10 mL min . For comparison, the activity of a pre-reduced In–Al-20 sample is also included.
2
O
3
–Al
2
O
3
mixed oxides. Reaction conditions: catalyst weight: 200 mg; P(C
3
H
8
) = 2.5 kPa; P(N
2
) = 97.5 kPa;
ꢁ
1
40
35
30
25
20
15
10
5
0
100
1
0
8
6
4
2
0
9
8
7
6
5
4
3
2
0
0
0
0
0
0
0
0
In O3
2
In-Al-40
In-Al-20
In-Al-10
Al O
2
3
10
0
2
4
6
8
10
12
0
2
4
6
8
10
12
0
2
4
6
8
10
12
Reaction time (h)
2 3 2 3 3 8
Fig. 5. Conversion of propane and selectivity to propylene for In O –Al O mixed oxides at 873 K. Reaction conditions: catalyst weight: 200 mg; P(C H ) = 2.5 kPa;
ꢁ
1
2 2
P(CO ) = 10 kPa; P(N ) = 87.5 kPa; total flow rate: 10 mL min .
2
V O
5 2
-, or Ga O
3
-based catalyst systems during propane or ethane
2 3 2 3
seen that all In O –Al O samples led to a significant production
dehydrogenation [4,16,28,42,43]. After the induction period, signif-
icantly higher dehydrogenation activities that lead to selective
conversion of propane with propylene yield in the range of 10–
of propylene with high selectivities (>70%) and appreciable ‘‘ini-
tial” steady conversions of propane ranging from ca. 27–37% at
3 h on stream. The ‘‘initial” steady conversions of propane de-
creased in the order In–Al-20 > In–Al-10 > In–Al-40. Note that the
1
2 3 2 3
5% can be achieved for the In O –Al O samples. The activity for
the mixed oxide samples were shown to be exceptionally stable
from 2 to 12 h, i.e. for In–Al-20 only a conversion loss by 2.2%
was detected.
conversion of CO
that of propane over all catalysts. The positive effect of CO
tration on the propane dehydrogenation over In–Al-20 can be fur-
ther seen from Fig. 6. All data were collected at the reaction time of
3 h. One can see that the rate of propylene formation remarkably
increased together with that of CO formation with increasing the
2
in terms of reaction rate closely resembles to
concen-
2
The dehydrogenation of propane was also run over the In
mixed oxide samples as well as their simple oxide analogs
at 873 K in the presence of CO , as was depicted in Fig. 5. The re-
sults show that the steady activity of the In –Al catalysts in
the presence of CO is twice that in the absence of CO , although
a slightly longer induction period (ca. 3 h) is required. It can be
2 3
O –
2 3
Al O
2
2
O
3
2
O
3
partial pressure of CO
showed no significant increase even with increasing the partial
pressure of CO . All these facts indicate that the essential role of
2 2
. On the other hand, the rate of H formation
2
2
2

1
06
M. Chen et al. / Journal of Catalysis 272 (2010) 101–108
0.04
600
5
4
3
2
1
0
00
00
00
00
00
0
.03
0.02
0.01
0.00
0
5
10
15
20
25
30
P CO2 (kPa)
Fig. 6. Effect of partial pressure of CO
2
on the dehydrogenation of propane over In–Al-20 at 873 K. Rate of formation: (j) C
3
H
6
; (s) H
2
; (h) CO; (d) TOF value for propylene
) = 2.5 kPa; total flow rate:
formation estimated by assuming the In
2
O
3
as quasi-sphere particles exposed with exclusively face (1 1 1) [40]. (Reaction conditions: P(C H
3 8
ꢁ
1
1
0 mL min ; balance gas: nitrogen.)
CO
CH
2
is to promote the dehydrogenation process (CH
= CHCH + H ) by coupling the reaction of CO with H
? CO + H O), which shifts the dehydrogenation equilibrium
by consuming H . It is also interesting to remark that, owing to
3
CH
2
CH
(CO
3
?
its quasi steady activity at 8 h on stream. This is in sharp contrast
to the previously reported supported Ga -based system for PDH
with CO , in which most of the gallia catalysts inclined to deacti-
vate drastically in a few hours [29,30]. With high selectivity to pro-
pylene, such outstanding stability as to the indium-containing
catalysts has never been reported before. This slowing of deactiva-
2
3
2
2
2
2
+
2 3
O
H
2
2
2
2
its intrinsic equilibrium nature, the overall RWGS reaction effi-
ciency is not necessarily 100%, which can explain the evolution
trend of the H
tration as indicated in Fig. 6.
One of the biggest challenges in propane dehydrogenation with
is the deactivation of catalysts over a long period of reaction.
To examine the likely long-term stability of the In–Al mixed oxide
catalysts under the reaction conditions, an extended 30 h on-
stream operation for dehydrogenation of propane in the presence
2
formation rate as a function of the CO
2
feed concen-
2 3
tion, when compared to the conventional Ga O -based catalysts,
could be due to a moderate coke formation as a consequence of
the apparently lower surface acidity of the supported indium-
based materials. It is important, however, to point out that the
In–Al-20 catalyst would deactivate rapidly within 12 h under reac-
CO
2
3 8
tion conditions with higher concentrated C H feeds (for example,
P(C ) = 20 kPa; P(CO ) = 20 kPa; P(N ) = 60 kPa; total flow rate:
3
H
8
2
2
ꢁ1
of CO
2
was carried out on the In–Al-20 catalyst. As shown in
100 mL min ) (data not shown), possibly due to accelerated coke
formation under more demanding conditions. Attempt was also
made to regenerate In–Al-20 after 30 h on stream by re-calcination
of the used catalyst in flowing air at 873 K for 4 h followed by sub-
Fig. 7, even after 30 h on-stream operation, a high conversion of
propane up to 25.5% could still be maintained for In–Al-20. It is
important to remark that this corresponds to only a 2.9% loss of
4
3
3
2
2
1
1
0
5
0
5
0
5
0
5
0
8
7
6
5
4
3
2
0
0
0
0
0
0
0
10
0
0
10
20
30
40
50
60
70
80
90
100
110
120
130
140
Reaction time (h)
Fig. 7. The stability and regeneration of In–Al-20 at 873 K. Reaction conditions: catalyst weight: 200 mg; P (C
3
H
8
) = 2.5 kPa; P (CO
2
2
) = 10 kPa; P (N ) = 87.5 kPa; total flow rate:
ꢁ
1
ꢁ1
1
2
0 mL min ; flow rate of air or purging N : 40 mL min .

M. Chen et al. / Journal of Catalysis 272 (2010) 101–108
107
sequent purge with N
2
for another 2 h (see Fig. 7). It is apparent
period could be the key active species for propane dehydrogena-
tion. This speculation is further corroborated by the excellent cor-
relation as identified for the relationship between the catalytic
dehydrogenation activity and the amount of metallic In0 species
derived from the highly dispersed indium oxide species in the
that the original activity of In–Al-20 could be fully restored, with
no noticeable deactivation being detected even after the third
regeneration.
3
.4. Active sites and structure–activity relationships
2 3 2 3
present In O –Al O catalysts (Fig. 8).
0
Having established that In species are the true active sites for
simple dehydrogenation, it is also highly important to establish
the possible role of indium in RWGS that may contribute to the po-
2 3 2 5
Previous investigations concerning the use of Cr O -, V O -, or
Ga
ence of CO
sites, the reaction may proceed via two different reaction path-
ways. More specifically, it has been demonstrated that CO im-
proves the yield of propylene by two independent pathways. CO
can participate as an oxidizing agent in a Mars–van Krevelen-type
redox cycle, in which the consumption and replenishment of the
lattice surface oxygen species play a key role in the reaction pro-
cess. Such simple redox mechanism has been proposed for the
chromium, and vanadium oxide-based catalysts [16,42,43]. On
2
O
3
-based catalysts for propane dehydrogenation in the pres-
2
have shown that, depending on the nature of the active
sitive effect of CO
grammed reaction of CO
Al samples as well as their simple oxide analogs, and the corre-
sponding signals of CO were collected in Fig. 9a. It is evident that
samples with higher fraction of bulk In exhibit higher yield of
CO at identical temperatures, and the activity decreases in the or-
der of In > In–Al-40 > In–Al-20 > In–Al-10 > Al . This may
suggest that bulk In is likely to be the active species for RWGS.
2
for PDH. To this end, a temperature-pro-
2
with H was conducted over the In
2
2 3
O –
2
2 3
O
2
2 3
O
2
O
3
2 3
O
2 3
O
To further confirm this assumption, In–Al-20 samples after pre-
reduction at 773 and 1073 K were tested for RWGS, and the corre-
sponding results were compared with that of the fresh In–Al-20 as
2
the other hand, CO can play a role in a removal of hydrogen pro-
duced in the dehydrogenation of propane by the reverse water–gas
shift reaction. The reduction of partial pressure of hydrogen shifts
the dehydrogenation equilibrium toward the propylene formation.
shown in Fig. 9b. For the 773 K-reduced In–Al-20 (Fig. 9g), in which
0
only highly dispersed In
2
O
3
was converted to In , analogous activ-
Such promoting role of CO
propylene observed over the gallium oxide-based catalysts [7].
Given the high performance of the In –Al catalyst for PDH,
it is of interest to clarify the likely active species responsible for the
observed dehydrogenation activity. It may be noted that the surface
acidity or basicity (Tables S1 and S2) was not in line with the activ-
ity of different catalysts (Figs. 4 and 5), inferring that the surface
acid–base nature is not the key factor governing the catalytic per-
2
was proposed to explain a high yield of
ity to that of the fresh In–Al-20 is obtained, suggesting that the
metallic In0 may not be the key species for the RWGS reaction.
For the 1073 K-reduced In–Al-20 (Fig. 9h), in which all indium
was thoroughly reduced, dramatically lower CO yield relative to
that of fresh In–Al-20 was observed. All these results clearly sug-
2
O
3
2 3
O
2 3
gest that it is bulk In O rather than metallic indium or highly dis-
persed In performed as the active phase for RWGS.
2 3
O
Finally, it is important to highlight that metallic indium, previ-
ously known to be a useful promoter in a number of formulated
formance of the In
that the highly dispersed In
site for NO removal during the SCR reaction [34,35], one might
2
O
3
–Al
2
O
3
samples. Taking into account the fact
2
O
3
is the widely accepted redox active
Pt–In/Al
[44,45], could become an interesting CO
tion catalyst when finely dispersed on a suitable support. More sig-
2
O
3
systems for commercial propane dehydrogenation
x
2
-mediated dehydrogena-
envisage that the present dehydrogenation reaction can proceed
via a direct redox mechanism involving alternate reduction and oxi-
nificantly, In
2
O
3
2
–Al O
3
was a far better catalyst in terms of catalytic
dation of the surface InO
the In–Al-20 sample was conducted after ‘‘reoxidation” of a hydro-
gen-pretreated In–Al-20 sample by CO . It should be mentioned
here that the H -pre-treatment temperature at 773 K permits the
reduction of the highly dispersed In but not the bulk one (see
Fig. 2). As noted above, if the simple redox pathway is followed,
one might expect that the reduced indium species would be readily
2
re-oxidized by CO atmosphere at 873 K. Nevertheless, the TPR pro-
x
sites. To testify this hypothesis, the TPR of
0
0
0
.04
.03
.02
2
2
2 3
O
file as shown in Fig. 2e does not show any LT reduction feature in
the temperature range of 473–773 K. This fact, together with the
XPS analysis revealing that CO
In (see Fig. 3c), strongly indicates that the direct redox mecha-
nism is not applicable for the present In -catalyzed dehydroge-
nation of propane in the presence of CO
To further elucidate the active sites in the In
the performance of the most active In–Al-20 catalyst for PDH in the
is incapable of re-oxidizing In0 to
2
2 3
O
2 3
O
2
.
2 3 2 3
O –Al O system,
absence of CO
hydrogen (5% H
the H
2
was investigated after the pre-treatment with
/Ar at 773 K for 1 h). It is important to remark that
-pre-treatment has a distinct influence on performance evo-
2
2
0.01
lution of In–Al-20 at the initial stage of the reaction (Fig. 4), with
essentially no influence on the steady-state activity of the catalyst.
2
It is of interest to note that the H -pretreated In–Al-20 shows
immediate high initial activity for PDH without experiencing the
induction period as observed with the fresh In–Al-20 sample.
Moreover, it must be pointed out that it is only after the induction
period (ca. 2 h on stream) a steady conversion of propane with pro-
nounced selectivity to propylene can be attained for fresh In–Al-20.
All these facts, together with the XPS results showing a concomi-
0
.00
5
10
15
20
Propane conversion at 2h (%)
2 3
tant reduction of the dispersed In O to the corresponding metallic
Fig. 8. Relationship between propane conversion at 2 h on stream in the absence of
CO and the corresponding LT hydrogen consumption ( ) of the In –Al
catalysts.
0
In species (see Fig. 5d), suggest that the surface-stabilized metallic
2
a
2
O
3
2 3
O
0
In nanoclusters that are generated in situ during the induction

1
08
M. Chen et al. / Journal of Catalysis 272 (2010) 101–108
the National Basic Research Program of China (2009CB623506)
a
b
the National High Technology Research and Development Program
of China (2006AA03Z336), and Science and Technology Commis-
sion of Shanghai Municipality (08DZ2270500).
a
e
Appendix A. Supplementary material
f
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.jcat.2010.03.007.
h
References
[
[
1] F. Cavani, N. Ballarini, A. Cericola, Catal. Today 127 (2007) 113.
2] G. Caeiro, R.H. Carvalho, X. Wang, M. Lemos, F. Lemos, M. Guisnet, F.R. Ribeiro,
J. Mol. Catal. A 255 (2006) 131.
[
[
3] P. Michorczyk, J. Ogonowski, Appl. Catal. A 251 (2003) 425.
4] M. Saito, S. Watanabe, I. Takahara, M. Inaba, K. Murata, Catal. Lett. 89 (2003)
2
13.
5] M.M. Bhasin, J.H. McCain, B.V. Vora, T. Imai, P.R. Pujado, Appl. Catal. A 221
2001) 397.
6] K. Nakagawa, C. Kajita, Y. Ide, M. Okamura, S. Kato, H. Kasuya, N. Ikenaga, T.
Kobayashi, T. Suzuki, Catal. Lett. 64 (2000) 215.
[
[
[
(
500 600 700 800 900 1000 500 600 700 800 900 1000
7] K. Nakagawa, C. Kajita, K. Okumura, N.-O. Ikenaga, M. Nishitani-Gamo, T. Ando,
T. Kobayashi, T. Suzuki, J. Catal. 203 (2001) 87.
Temperature (K)
[
[
8] J. Gascon, C. Tellez, J. Herguido, M. Menendez, Appl. Catal. A 248 (2003) 105.
9] M. Larsson, M. Hulten, E.A. Blekkan, B. Andersson, J. Catal. 164 (1996) 44.
Fig. 9. Temperature-programmed CO
Al catalysts: (a) In ; (b) In–Al-40; (c) In–Al-20; (d) In–Al-10; (e) Al
In–Al-20; (g) 773 K-H /Ar reduced In–Al-20; (h) 1073 K-H /Ar reduced In–Al-20.
2
–H
2
tests of CO
2
with H
2
for various In
2 3
O –
2
O
3
2
O
3
2
O
3
(f) fresh
[10] M.M. Bettahar, G. Costentin, L. Savary, J.C. Lavalley, Appl. Catal. A 145 (1996) 1.
[11] E.A. Mamedov, V.C. Corberan, Appl. Catal. A 127 (1995) 1.
2
2
[
[
12] Y.M. Liu, Y. Cao, S.R. Yan, W.L. Dai, K.N. Fan, Catal. Lett. 88 (2003) 61.
13] Y.M. Liu, Y. Cao, N. Yi, W.L. Feng, W.L. Dai, S.R. Yan, H.Y. He, K.N. Fan, J. Catal.
224 (2004) 417.
stability for PDH with CO
based catalysts in the current literature [25,26]. Being different
from all previously reported systems (for example, Cr , V
Ga ) [4,16,28,42,43], the results described in the present work
show clearly a unique bifunctional character of the indium compo-
nent in the In –Al materials for propane dehydrogenation
with CO . In this context, the overall reaction pathway can be envis-
aged to involve both metallic In species and the bulk In phase
during the CO -mediated propane dehydrogenation process, i.e.,
2 2 3
compared to the most effective Ga O -
[
[
[
[
14] Y.M. Liu, Y. Cao, K.K. Zhu, S.R. Yan, W.L. Dai, H.Y. He, K.N. Fan, Chem. Commun.
(
2002) 2832.
2
O
3
2 5
O ,
15] K. Nakagawa, C. Kajita, N. Ikenaga, M. Nishitani-Gamo, T. Ando, T. Suzuki, Catal.
Today 84 (2003) 149.
16] K. Takehira, Y. Ohishi, T. Shishido, T. Kawabata, K. Takaki, Q.H. Zhang, Y. Wang,
J. Catal. 224 (2004) 404.
2 3
O
2
O
3
2 3
O
17] P. Kustrowski, M. Zbroja, R. Dziembaj, H. Papp, Catal. Lett. 80 (2002) 1.
2
[18] K. Nowinska, A. Waclaw, A. Izbinska, Appl. Catal. A 243 (2003) 225.
[
19] E.V. Kondratenko, M. Cherian, M. Baerns, D. Su, R. Schlogl, X. Wang, I.E. Wachs,
J. Catal. 234 (2005) 131.
20] S.B. Wang, Z.H. Zhu, Energy Fuels 18 (2004) 1126.
2 3
O
2
[
0
the reaction proceeds via a simple metallic In -catalyzed dehydro-
genation pathway facilitated by the RWGS reaction catalyzed by
[21] K. Nakagawa, M. Okamura, N. Ikenaga, T. Suzuki, T. Kobayashi, Chem.
Commun. (1998) 1025.
[
22] Y. Wang, Y. Ohishi, T. Shishido, Q.H. Zhang, W. Yang, Q. Guo, H.L. Wan, K.
Takehira, J. Catal. 220 (2003) 347.
bulk In
as described above, we can rationalize that a well balance of the
metallic In species and the bulk In phase is indispensable for
an efficient propane dehydrogenation over the In –Al system.
2 3
O . On the basis of the proposed main reaction pathway
[23] I. Takahara, W.C. Chang, N. Mimura, M. Saito, Catal. Today 45 (1998) 55.
[24] H. Shimada, T. Akazawa, N. Ikenaga, T. Suzuki, Appl. Catal. A 168 (1998) 243.
2 3
O
[
[
[
25] M. Chen, J. Xu, Y.M. Liu, Y. Cao, H.Y. He, J.H. Zhuang, K.N. Fan, Catal. Lett. 124
2008) 369.
26] M. Chen, J. Xu, F.Z. Su, Y.M. Liu, Y. Cao, H.Y. He, K.N. Fan, J. Catal. 256 (2008)
93.
2
O
3
2 3
O
(
2
4
. Conclusions
27] D.Y. Hong, J.S. Chang, J.H. Lee, V.P. Vislovskiy, S.H. Jhung, S.E. Park, Y.H. Park,
Catal. Today 112 (2006) 86.
The present work demonstrates the high potential of a new
class of catalytic materials based on binary In–Al–O nanocompos-
ites for the dehydrogenation of propane with CO . An optimum in
the catalytic behavior was achieved for the catalyst containing 20%
atom.) In , which can afford a propylene yield of 27.3% and
maintain an exceptionally steady propane conversion of >24% dur-
ing 30 h on stream. XRD and H -TPR characterizations suggest that
the combination of indium and aluminum can result in significant
modification in the bulk dispersion and surface redox properties of
2 3
the In O phase. The correlation of the catalytic activity and the
[28] H.Y. Li, Y.H. Yue, C.K. Miao, Z.K. Xie, W.M. Hua, Z. Gao, Catal. Commun. 8 (2007)
317.
1
[
[
29] B.J. Xu, B. Zheng, W.M. Hua, Y.H. Yue, Z. Gao, J. Catal. 239 (2006) 470.
30] B. Zheng, W.M. Hua, Y.H. Yue, Z. Gao, J. Catal. 232 (2005) 143.
2
[31] M.P. Lobera, C. Téllez, J. Herguido, M. Menéndez, Appl. Catal. A 349 (2008) 156.
[
32] M.S. Kumar, N. Hammer, M. Ronning, A. Holmen, D. Chen, J.C. Walmsley, G.
Oye, J. Catal. 261 (2009) 116.
(
2 3
O
[
33] M.S. Kumar, D. Chen, A. Holmen, J.C. Walmsley, Catal. Today 142 (2009) 17.
2
[34] J.A. Perdigon-Melon, A. Gervasini, A. Auroux, J. Catal. 234 (2005) 421.
[35] P.W. Park, C.S. Ragle, C.L. Boyer, M.L. Balmer, M. Engelhard, D. McCready, J.
Catal. 210 (2002) 97.
[
[
36] T. Maunula, Y. Kintaichi, M. Haneda, H. Hamada, Catal. Lett. 61 (1999) 121.
37] M. Haneda, Y. Kintaichi, H. Hamada, Catal. Lett. 55 (1998) 47.
TPR data unambiguously clarify that the creation and maintenance
of surface metallic indium species in optimal surface concentration
at the initial stage of the reaction are essential for achieving pro-
[38] A. Gervasini, J.A. Perdigon-Melon, C. Guimon, A. Auroux, J. Phys. Chem. B 110
2006) 240.
39] C.O. Arean, M.R. Delgado, V. Montouillout, D. Massiot, Z. Anorg. Allg. Chem.
31 (2005) 2121.
[40] C.T. Prewitt, R.D. Shannon, D.B. Rogers, A.W. Sleight, Inorg. Chem. 8 (1969)
985.
(
[
6
nounced dehydrogenation activity. The promoting effect of CO
2
1
on the yield of propylene over the mixed oxide catalysts has been
attributed to a facilitation of simple dehydrogenation by coupling
[
[
41] A.W.C. Lin, N.R. Armstrong, T. Kuwana, Anal. Chem. 49 (1977) 1228.
42] Y. Ohishi, T. Kawabata, T. Shishido, K. Takaki, Q.H. Zhang, Y. Wang, K. Takehira,
J. Mol. Catal. A 230 (2005) 49.
2 3
with the bulk In O -catalyzed RWGS reaction.
[
[
[
43] P. Michorczyk, K. Gora-Marek, J. Ogonowski, Catal. Lett. 109 (2006) 195.
44] F.A. Marchesini, S. Irusta, C. Querini, E. Mir, Catal. Commun. 9 (2008) 1021.
45] F.B. Passos, M. Schmal, M.A. Vannice, J. Catal. 160 (1996) 106.
Acknowledgments
This work was financially supported by the National Natural Sci-
ence Foundation of China (20633030, 20721063, and 20873026),