Support effect in dehydrogenation of propane in the presence of CO2 over supported gallium oxide catalysts

Article abstract of DOI:10.1016/j.jcat.2006.02.017

Dehydrogenation of propane to propene in the absence or presence of CO₂ over different supported gallium oxide catalysts was investigated. Ga₂O₃/TiO₂, Ga₂O₃/Al₂O₃, and Ga₂O₃/ZrO₂ catalysts exhibited high activity, and Ga₂O₃/SiO₂ and Ga₂O₃/MgO were ineffective catalysts for the dehydrogenation of propane. The conversions of propane over Ga₂O₃/TiO₂, Ga₂O₃/Al₂O₃, and Ga₂O₃/ZrO₂ catalysts at 600 °C reached 23, 33, and 39%, respectively. The high dehydrogenation activities are connected with the abundant medium-strong acid sites on the catalyst surface. Carbon dioxide can promote the catalytic dehydrogenation activity of Ga₂O₃/TiO₂ but suppress those of Ga₂O₃/ZrO₂ and Ga₂O₃/Al₂O₃. The conversion of propane becomes 32, 26, and 30% in the presence of CO₂ over Ga₂O₃/TiO₂, Ga₂O₃/Al₂O₃, and Ga₂O₃/ZrO₂ catalysts, respectively. Results of pulse reaction and the H₂ and propane chemisorptions indicate that two contrary roles of CO₂ exist in the reaction: the positive role of removing dissociatively adsorbed H₂ on catalyst surface through the reverse water-gas shift reaction and the negative role of displacing the propane adsorbed on basic sites of catalyst. XPS studies show that the different behavior of the five supported Ga₂O₃ catalysts may be attributed to the different interactions between the support and the Ga₂O₃ species. The dehydrogenation reaction is suggested to proceed through a heterolytic dissociation reaction pathway.

Full text of DOI:10.1016/j.jcat.2006.02.017

Journal of Catalysis 239 (2006) 470–477
www.elsevier.com/locate/jcat
Support effect in dehydrogenation of propane in the presence of CO2
over supported gallium oxide catalysts
Bingjun Xu, Bo Zheng, Weiming Hua ∗, Yinghong Yue ∗, Zi Gao
Department of Chemistry and Shanghai Key Laboratory of Molecular Catalysis and Innovative Materials, Fudan University, Shanghai 200433, PR China
Received 16 January 2006; revised 22 February 2006; accepted 23 February 2006
Abstract
Dehydrogenation of propane to propene in the absence or presence of CO over different supported gallium oxide catalysts was investigated.
2
Ga O /TiO , Ga O /Al O , and Ga O /ZrO catalysts exhibited high activity, and Ga O /SiO and Ga O /MgO were ineffective catalysts
2
3
2
2
3
2
3
2
3
2
2
3
2
2 3
◦
for the dehydrogenation of propane. The conversions of propane over Ga O /TiO , Ga O /Al O , and Ga O /ZrO catalysts at 600 C reached
2
3
2
2
3
2
3
2
3
2
23, 33, and 39%, respectively. The high dehydrogenation activities are connected with the abundant medium-strong acid sites on the catalyst sur-
face. Carbon dioxide can promote the catalytic dehydrogenation activity of Ga O /TiO but suppress those of Ga O /ZrO and Ga O /Al O .
2
3
2
2
3
2
2
3
2 3
The conversion of propane becomes 32, 26, and 30% in the presence of CO over Ga O /TiO , Ga O /Al O , and Ga O /ZrO catalysts,
2
2
3
2
2
3
2
3
2
3
2
respectively. Results of pulse reaction and the H and propane chemisorptions indicate that two contrary roles of CO exist in the reaction: the
2
2
positive role of removing dissociatively adsorbed H on catalyst surface through the reverse water–gas shift reaction and the negative role of
2
displacing the propane adsorbed on basic sites of catalyst. XPS studies show that the different behavior of the five supported Ga O catalysts
2
3
may be attributed to the different interactions between the support and the Ga O species. The dehydrogenation reaction is suggested to proceed
2
3
through a heterolytic dissociation reaction pathway.
 2006 Elsevier Inc. All rights reserved.
Keywords: Supported gallium oxide catalyst; Propane; Dehydrogenation; Carbon dioxide; Reaction mechanism
1. Introduction
Recently, catalytic oxidative dehydrogenation of propane by
carbon dioxide instead of oxygen has been attempted [1,2]. The
promoting effect of carbon dioxide on the reaction has been
observed on many catalysts, including silica-supported Cr2O3
[1,2], rare earth vanadates [3], and Ga2O3 [4]. Because carbon
dioxide is one of the major greenhouse gases, the use of carbon
dioxide is attractive not only economically, but also ecologi-
cally.
Aromatization of light paraffins over Ga-promoted HZSM-
5 catalysts has been studied intensively in the last decade.
The role of gallium oxide in the aromatization reaction is
thought to facilitate the dehydrogenation steps, including the
dehydrogenation of alkanes, higher olefins, and cycloolefins
[5–10]. More recently, it has been found that carbon diox-
ide can markedly promote the dehydrogenation of ethane and
propane over gallium oxide or supported gallium oxide cata-
lysts [4,11–16]. But this promotional effect was not observed
over all of the supported gallium oxide catalysts; in fact, the de-
hydrogenation reaction was sometimes inhibited, such as over
The catalytic dehydrogenation of propane to propene is of
increasing importance due to the growing demand for propene,
which is an important raw material in the production of
polypropene, polyacrylonitrile, acrolein, and acrylic acid. De-
hydrogenation of propane is an endothermic reaction; therefore,
relatively high reaction temperatures are needed to obtain high
propane yields. The high reaction temperatures favor thermal
cracking reactions to form light alkanes and alkenes, leading
to decreased product yield and increased catalyst deactivation.
Catalytic oxidative dehydrogenation of propane by oxygen is
an attractive alternative because it is an exothermic and non-
thermodynamically limited reaction. However, the propene se-
lectivity of the catalytic reaction remains an unsolved problem
owing to the overoxidation of propane.
*
Corresponding authors. Fax: +86 21 65641740.
E-mail address: yhyue@fudan.edu.cn (Y. Yue).
0021-9517/$ – see front matter  2006 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2006.02.017

B. Xu et al. / Journal of Catalysis 239 (2006) 470–477
471
γ -Al2O3-supported Ga2O3 catalyst [12]. The reason for this
unusual support effect remains still unclear. In addition, re-
searchers have not yet reached agreement on the main role of
carbon dioxide in reactions. Nakagawa et al. proposed CO2-
promoted desorption of olefin products from the catalyst sur-
face [12], and Michorczyk et al. considered CO2-consumed
hydrogen via a reverse water–gas shift reaction [13]. In previ-
ous work we found that the reverse water–gas shift reaction and
the Boudouard reaction of CO2 may account for the enhanced
catalytic activity and stability over Ga2O3 catalysts [16].
In the present work, five different supported gallium oxide
catalysts Ga2O3/TiO2, Ga2O3/Al2O3, Ga2O3/ZrO2, Ga2O3/
SiO2, and Ga2O3/MgO were prepared and characterized.
Their catalytic performance for dehydrogenation of propane
to propene in the absence or presence of CO2 was investigated,
as was the support effect as well as the role of CO2. Finally,
a probable mechanism of the reaction over supported gallium
oxide catalysts was proposed.
N2 flow at 300 ◦C for 3 h. The temperature was increased from
50 to 600 ◦C at a ramp rate of 10 ◦C/min. A thermal conductiv-
ity detector (TCD) was used to monitor the hydrogen consumed
during the TPR experiments. X-ray photoelectron spectroscopy
(XPS) was performed using Al-Kα radiation (1486.6 eV) on
a Perkin–Elmer PHI 5000C ESAC system with a base pres-
sure of 1 × 10−9 Torr. The sample was pressed and degassed
in the pretreatment chamber for 2 h before being transferred
to the analysis chamber for XPS measurement. All binding en-
ergy (BE) values were referenced to the C 1s peak at 284.6 eV.
H2 and propane chemical adsorption measurements were car-
ried out on a Micromeritics TPD/TPR 2900 instrument using
500 mg of catalyst under a argon gas flow (40 ml/min). The
temperature was raised from room temperature to 600 ◦C at a
rate of 10 ◦C/min, maintained at 600 ◦C for 2 h, and then cooled
to the adsorption temperature (300 ◦C). For H2 adsorption, sev-
eral pulses of the mixed gas (10% hydrogen and 90% argon)
with fixed volume were introduced until no more H2 was ad-
sorbed on the catalyst. For propane adsorption, when in the
absence of CO2, several pulses of propane with fixed volume
were injected until no more propane was adsorbed on the cat-
alyst. While in the presence of CO2, the experiment conducted
in the gas flow contained 5 vol% CO2 and the balance argon
instead of pure argon.
2. Experimental
2.1. Catalyst preparation
ZrO2 and MgO were prepared by adding an aqueous solution
of ammonia into ZrOCl2 and Mg(NO3)2 solutions, respectively,
followed by aging overnight, filtering, washing, drying, and cal-
cining at 600 ◦C for 6 h. TiO2 (Degussa P25), γ -Al2O3, and
SiO2 were purchased directly and used with no further treat-
ment.
Supported gallium oxide catalysts containing 5 wt% Ga2O3
were prepared by impregnating an aqueous solution of Ga-
(NO3)3·xH2O (Aldrich) on TiO2, γ -Al2O3, ZrO2, SiO2, or
MgO using an incipient wetness method. The impregnated
samples were dried at 100 ◦C and calcined at 600 ◦C for 6 h
in air flow. The catalysts thus obtained were designated as
Ga2O3/TiO2, Ga2O3/Al2O3, Ga2O3/ZrO2, Ga2O3/SiO2, and
Ga2O3/MgO.
2.3. Reaction testing
Catalytic tests were performed in a fixed-bed flow microre-
actor at atmospheric pressure, with nitrogen as the carrier gas
at a flow rate of 20 ml/min. The catalyst load was 200 mg,
and it was activated at the reaction temperature (600 ◦C) for
1 h in nitrogen flow before the reaction. For dehydrogenation
of propane in the absence of carbon dioxide, the gas reactant
contained 2.5 vol% propane and the balance nitrogen. For the
dehydrogenation of propane in the presence of carbon dioxide,
the gas reactant contained 2.5 vol% propane, 5 vol% carbon
dioxide, and the balance nitrogen.
The hydrocarbon reaction products were analyzed using an
on-line gas chromatograph equipped with a 6-m packed column
of Porapak Q and a flame ionization detector. The gas products,
including N2, CO, and CO2, were analyzed on-line by another
chromatograph equipped with a 2-m packed column of carbon
molecular sieve 601 and a TCD. The conversion and selectivity
were calculated as follows:
2.2. Catalyst characterization
The specific surface areas of the catalysts were measured by
nitrogen adsorption at −196 ◦C using a Micromeritics ASAP
2000 instrument and calculated by the BET method. Surface
acidity was measured by NH3 temperature-programmed des-
orption (NH3-TPD) in a flow-type fixed-bed reactor at am-
bient pressure, and surface basicity was measured by CO2
temperature-programmed desorption (CO2-TPD) with a similar
apparatus. A 100-mg sample was preheated at 600 ◦C for 3 h,
and then cooled to 120 ◦C in flowing He. At this temperature,
sufficient pulses of NH3 or CO2 were injected until adsorp-
tion saturation occurred, followed by purging with He for 2 h.
The temperature was then raised from 120 to 600 ◦C at a rate
of 10 ◦C/min, and the NH3 or CO2 desorbed was collected in
a liquid N2 trap and detected by on-line gas chromatography.
Temperature-programmed reduction (TPR) experiments were
carried out on a Micromeritics TPD/TPR 2900 instrument using
25 mg of catalyst under a mixed gas flow (40 ml/min) of hy-
drogen (10%) and argon (90%). The catalyst was pretreated in
C3H8 in − C3H8 out
C3H8 conversion =
CO2 conversion =
× 100%,
C3H8 in
CO2 in − CO2 out
× 100%,
CO2 in
C3H6 out
C3H6 selectivity =
× 100%.
C3H8 in − C3H8 out
2.4. Pulse reaction
Pulse tests were carried out on a fixed-bed pulse microre-
actor. The catalyst load for the test was 200 mg. For CO2 and

472
B. Xu et al. / Journal of Catalysis 239 (2006) 470–477
H2 pulse testing, the catalyst was pretreated with He for 1 h at
reaction temperature, after which a pulse of CO2 and H2 gas
mixture (1:1 molar ratio; 2.0 ml) was injected. For CO2 pulse
reaction, the catalyst was first pretreated at the reaction temper-
ature (600 ◦C) in He for 1 h and then in H2 for 30 min; then He
was used to purge the pulse reaction system for 10 min to re-
move H2. Finally, a pulse of CO2 (2.0 ml) was injected. He, at
a flow rate of 20 ml/min, was used as the carrier gas.
The products of the pulse reaction were analyzed using an
on-line gas chromatograph equipped with a 2-m packed column
of carbon molecular sieve 601 and a TCD. The reaction data in
the work were reproducible with a precision of <5%.
duction of gallium oxide on the supports. Meanwhile, the im-
pregnation of gallium oxide on the supports also significantly
increased the total amount of surface acid sites, especially
those of medium-strong acid strength (expressed as the amount
of NH3 desorbed at 350–600 ◦C), which are suggested to be
active for the dehydrogenation reaction [16]. The amount of
medium to strong acid sites of the catalysts has the following
sequence: Ga2O3/ZrO2 > Ga2O3/Al2O3 > Ga2O3/TiO2 >
Ga2O3/SiO2 > Ga2O3/MgO.
The surface basicity of the supports and the supported gal-
lium oxide catalysts were measured by CO2-TPD. One broad
CO2 desorption peak was observed for all of the supports as
well as the supported gallium oxide catalysts; the peak temper-
ature and amount of desorbed CO2 of all samples are summa-
rized in Table 2. There was a much smaller number of basic
sites than of acidic sites on all of the supported gallium oxide
catalysts except for Ga2O3/MgO, possibly due to the relatively
strong basic property of the MgO support. Therefore, the sup-
ported gallium oxides except Ga2O3/MgO are primarily acid
catalysts. The sequence of increasing amount of basic sites
on the catalyst surface was Ga2O3/MgO > Ga2O3/ZrO2 >
Ga2O3/γ -Al2O3 > Ga2O3/TiO2 ꢀ Ga2O3/SiO2.
3. Results
3.1. Catalyst characterization
TiO2-, γ -Al2O3-, ZrO2-, SiO2-, and MgO-supported gal-
lium oxide catalysts containing 5 wt% Ga2O3 were prepared
by an incipient wetness method. Their powder X-ray diffraction
patterns were measured. No other diffraction peaks but those of
the supports were observed, indicating that gallium oxide was
well dispersed on all of the supports. TPR experiments were
also carried out to study the reducibility of these catalysts. No
reduction peaks could be identified in the TPR profiles, suggest-
ing that the supported gallium oxide catalysts cannot be reduced
by hydrogen below 600 ◦C.
The surface acidity of the supported gallium oxide catalysts
was measured by NH3-TPD; the results are given in Table 1,
together with those of the supports for comparison. All of the
supports exhibited only one broad peak on the TPD profiles,
in the range of 228–327 ◦C, indicating that the acid sites of
the supports are of weak to medium strength. The acidity of
the supports was modified significantly after being impregnated
with gallium oxide. Evidently there were two desorption peaks
on the TPD profiles of all of the supported gallium oxide cat-
alysts. The low temperature peaks at about 218–313 ◦C and
the high peaks at 392–442 ◦C corresponded to the weak and
strong acid sites of the samples. The low temperature peaks
were similar to those of the supports; therefore the emergence
of the high temperature peaks can be attributed to the intro-
The XPS spectra of supported gallium oxide catalysts were
recorded. The deconvoluted spectra in the Ga 3d region are
shown in Fig. 1. Only two peaks appear in the spectra of
the Ga2O3/ZrO2 and Ga2O3/Al2O3 catalysts, at about 24 and
21 eV, respectively, which can be assigned to O 2s and Ga3+
3d bands according to the literature [17]. The Ga 3d band-
ing energy was increased when gallium oxide was supported
on ZrO2 or Al2O3, indicating a strong interaction between the
Ga2O3 species and the ZrO2 or Al2O3 support, in accordance
with previously reported results [18,19]. A low-energy peak
(LBE) at 19.6 eV appeared in the spectrum of Ga2O3/TiO2,
Ga2O3/SiO2, and Ga2O3/MgO catalysts, which can be at-
tributed to the presence of Gaδ+ species (δ < 2) [20], suggest-
ing that Ga2O3 was partially reduced on these supports. Similar
results have been reported previously [21], and the decrease of
Ga 3d binding energy of the gallium species in the Ga2O3/TiO2
catalyst has been tentatively attributed to desorption of some
loosely bound surface oxygen atoms at high calcination tem-
peratures. The relative ratios of the Ga3+ and Gaδ+ species of
the three catalysts were calculated and the percentage of Gaδ+
Table 1
NH -TPD data of the supports and the supported gallium oxide catalysts
3
Table 2
Catalyst
Peak tem-
perature ( C) (mmol/g)
Amount of desorbed NH
3
CO -TPD data of the supports and the supported gallium oxide catalysts
2
◦
Catalyst
Peak tem-
perature ( C)
Amount of desorbed CO (mmol/g)
2
◦
◦
I
II
I (120–350 C)
II (350–600 C)
Total
◦
◦
◦
I (120–350 C)
II (350–600 C)
Total
Ga O /TiO
220
228
313
327
310
320
218
232
310
302
442
–
0.20
0.18
0.29
0.22
0.37
0.16
0.16
0.07
0.12
0.06
0.21
0.12
0.28
0.19
0.34
0.24
0.10
0.06
0.08
0.02
0.41
0.30
0.57
0.41
0.71
0.40
0.26
0.13
0.20
0.08
2
3
2
a
Ga O /TiO
250
222
235
219
361
320
214
211
293
302
0.06
0.05
0.07
0.06
0.07
0.06
0.02
0.01
0.13
0.12
0.02
0.01
0.02
0.01
0.03
0.02
0
0
0.06
0.06
0.08
0.06
0.09
0.07
0.10
0.08
0.02
0.01
0.19
0.18
2
3
2
TiO
2
TiO
2
Ga O /Al O
418
2
3
2 3
a
Ga O /Al O
2
3
2 3
Al O
2
–
3
Al O
2
3
Ga O /ZrO
436
2
3
2
a
Ga O /ZrO
2
3
2
ZrO
–
2
ZrO
2
Ga O /SiO
408
2
3
2
a
Ga O /SiO
2
3
2
SiO
–
2
SiO
2
Ga O /MgO
392
2
3
a
Ga O /MgO
2
3
MgO
–
MgO
a
Not detected.

B. Xu et al. / Journal of Catalysis 239 (2006) 470–477
473
Table 4
Reaction data in the absence of CO
2
Catalyst
S
Propane con- Selectivity (%)
BET
2
a
version (%)
(m /g)
CH
C H
2
C H
2
C H
3
4
4
6
6
Ga O /TiO
2
47
102
19
329
35
23
33
39
7.2
5.3
6.8
3.2
9.1 13.2
2.9 4.8
11 33
7.1
4.2
1.0
0.8
3.6
0.4
22
85
92
74
92
34
2
3
Ga O /Al O
2
3
2 3
Ga O /ZrO
2
3
2
Ga O /SiO
2
3
2
Ga O /MgO
2
3
a
Reaction time: 10 min.
Fig. 2. Conversion of propane over supported gallium oxide catalysts in the
absence of carbon dioxide: (2) Ga O /TiO ; (") Ga O /Al O ; (Q) Ga O /
Fig. 1. XPS spectra of supported gallium oxide catalysts (Ga 3d) on (a) Ga O /
2
3
2
2
3
2
3
2 3
2
3
ZrO ; (a) Ga O /SiO ; (F) Ga O /MgO.
TiO ; (b) Ga O /Al O ; (c) Ga O /ZrO ; (d) Ga O /SiO ; (e) Ga O /
2
2
3
2
2 3
2
2
3
2
3
2
3
2
2
3
2
2
3
MgO.
Table 5
Reaction data in the presence of CO
Table 3
2
XPS data of supported gallium oxide catalysts
a
Catalyst
Conversion (%)
Selectivity (%)
3+
(%)
δ+
(%)
Catalyst
Ga 3d binding
energy (eV)
Ga
Ga
C H
3
CO
CH
C H
2
C H
2
C H
3 6
8
2
4
4
6
a
Ga O /TiO
2
32
30
10
16
1.1
0.4
4.2
0.3
28
73
94
65
92
29
2
3
β-Ga O
20.8 [16]
19.6/20.8
21.6
100
55
100
100
93
0
45
0
0
7
2
3
Ga O /Al O
26
30
5.2
29
2.9
14
3.1
10
3.8
17
4.8
33
2
3
2 3
Ga O /TiO
2
3
2
Ga O /ZrO
2
3
2
Ga O /Al O
2
3
2 3
Ga O /SiO
6.4
4.3
3.1
4.2
2
3
2
Ga O /ZrO
21.6
2
3
2
Ga O /MgO
2
3
Ga O /SiO
19.6/20.8
19.6/20.8
2
3
2
a
Reaction time: 10 min.
Ga O /MgO
92
8
2
3
a
δ < 2.
> Ga2O3/MgO, paralleling the sequence of the amount of
medium to strong acid sites and indicating that surface acid-
ity should play an important role in dehydrogenation. For
all catalysts, propane conversion decreased along with the
reaction time, but the decrease was much slower on the
Ga2O3/Al2O3, Ga2O3/MgO, and Ga2O3/SiO2 catalysts than
on the Ga2O3/TiO2 and Ga2O3/ZrO2 catalysts. Catalyst deac-
tivation is commonly attributed to the formation of coke on the
catalyst surface.
The dehydrogenation of propane was also run over the sup-
ported gallium oxide catalysts in the presence of CO2 at 600 ◦C;
the reaction data are summarized in Table 5 and Fig. 3. CO
was detected in the reaction product. The addition of CO2
in the dehydrogenation reaction has different effects on the
Ga2O3/TiO2, Ga2O3/Al2O3, Ga2O3/ZrO2, Ga2O3/SiO2, and
Ga2O3/MgO catalysts. The initial activity of Ga2O3/TiO2 in
the presence of CO2 was much higher than that in the absence
of CO2; in contrast, that of Ga2O3/Al2O3 and Ga2O3/ZrO2
species was found to decrease in the order Ga2O3/TiO2 ꢀ
Ga2O3/MgO ≈ Ga2O3/SiO2. The XPS results are summarized
in Table 3.
3.2. Catalytic activity
Dehydrogenation of propane over supported gallium oxide
catalysts in the absence of CO2 was carried out at 600 ◦C;
the results, along with the BET surface areas of the cata-
lysts, are given in Table 4 and Fig. 2. The major product
formed in the reaction was propene, and the minor products
were methane, ethane, and ethylene. As Fig. 2 clearly shows,
the catalytic behaviors of the five supported gallium oxide
catalysts in the reaction differ considerably. The initial con-
version of propane on the catalysts decreases in the order
Ga2O3/ZrO2 > Ga2O3/Al2O3 > Ga2O3/TiO2 > Ga2O3/SiO2

474
B. Xu et al. / Journal of Catalysis 239 (2006) 470–477
Fig. 3. Conversion of propane over supported gallium oxide catalysts in
the presence of carbon dioxide: (2) Ga O /TiO ; (") Ga O /Al O ;
2
3
2
2
3
2
3
(Q) Ga O /ZrO ; (a) Ga O /SiO ; (F) Ga O /MgO.
2
3
2
2
3
2
2 3
Fig. 4. Conversion of propane over supported gallium oxide catalysts with dif-
Table 6
The conversion of CO in reactions (1) and (2)
a
ferent CO /C H ratios: (2) Ga O /TiO ; (") Ga O /Al O ; (Q) Ga O /
2
3
8
2
3
2
2
3
2
3
2 3
2
ZrO ; (a) Ga O /SiO ; (F) Ga O /MgO.
2
2
3
2
2 3
Catalyst
Ga O /TiO
2
C
(%)
C
(%)
(2)
(1)
20
2.1
2.9
1.1
1.1
10
3.1
26
2.0
3.1
2
3
of CO2, respectively. The results, given in Table 6, indicate
that the effect of adding CO2 differs among the five cata-
lysts. For Ga2O3/TiO2, CO2 is consumed mainly through
the reverse water–gas shift reaction, which may account for
the enhancement of the dehydrogenation activity; whereas for
Ga2O3/ZrO2, the main role of CO2 becomes the elimination
of coke by the Boudouard reaction, which may account for the
enhancement of stability of Ga2O3/ZrO2 catalyst after the ad-
dition of CO2 (Figs. 2 and 3).
Ga O /Al O
2
3
2 3
Ga O /ZrO
2
3
2
Ga O /SiO
2
3
2
Ga O /MgO
2
3
a
Reaction time: 10 min.
was reduced in the presence of CO2, whereas for Ga2O3/MgO
and Ga2O3/SiO2, CO2 seemed to have little effect on initial
dehydrogenation activity. The initial conversion of propane on
the catalysts in the presence of CO2 decreased in the follow-
ing order: Ga2O3/TiO2 > Ga2O3/ZrO2 > Ga2O3/Al2O3 >
Ga2O3/SiO2 > Ga2O3/MgO.
3.3. Effect of the CO2/C3H8 ratio on the dehydrogenation of
propane
The amount of initial CO2 conversion on the catalysts was
in the order Ga2O3/TiO2 ≈ Ga2O3/ZrO2 ꢀ Ga2O3/Al2O3 ≈
Ga2O3/SiO2 ≈ Ga2O3/MgO, indicating that evidently CO2
participated in dehydrogenation only over Ga2O3/TiO2 and
Ga2O3/ZrO2. It has been reported that the dehydrogenation re-
action can be influenced by CO2 through the reverse water–gas
shift reaction and the Boudouard reaction [16], as follows:
The effect of CO2 partial pressure on the dehydrogenation
of propane over supported gallium oxide catalysts was also
studied; the results are shown in Fig. 4. Propane conversion
over Ga2O3/TiO2 initially increased with the CO2/C3H8 ra-
tio, reaching its peak when the CO2/C3H8 ratio equaled 2, then
decreased with further increases in the CO2/C3H8 ratio. In con-
trast, propane conversion over Ga2O3/Al2O3 and Ga2O3/ZrO2
decreased continuously with increasing CO2/C3H8 ratio, dra-
matically at the beginning and more slowly later on. In
Ga2O3/MgO and Ga2O3/SiO2, the increase in CO2/C3H8 ra-
tio had only a limited negative effect on propane conversion at
first.
CO2 + H2 → CO + H2O
and
(1)
CO2 + C → 2CO.
(2)
The conversion of CO2 in reactions (1) and (2) can be estimated
from the amount of CO and CO2 before and after the reaction,
as follows:
3.4. Pulse reaction
n(CO2)after + n(CO) − n(CO2)before
C(2)
=
× 100%
To obtain more information on the dehydrogenation of
propane over various supported gallium oxide catalysts, CO2
and H2 pulse reaction, as well as CO2 pulse reaction, were
carried out. The results of the CO2 and H2 pulse reaction are
shown in Fig. 5. The process is the reverse water–gas shift re-
action, as shown in reaction (1). As shown in the figure, all of
the supported gallium oxide catalysts except Ga2O3/SiO2 dis-
played high activity for the reverse water–gas shift reaction,
n(CO2)before
and
C(1) = CCO − C(2)
,
2
where n(CO2)before, n(CO2)after, n(CO), and CCO denote
the amount of CO2 before reaction, the amount of CO2 af-
ter reaction, the amount of CO formed, and the conversion
2

B. Xu et al. / Journal of Catalysis 239 (2006) 470–477
475
Table 7
◦
Amount of H chemisorbed on the supported gallium oxide catalysts (300 C)
2
Catalyst
Ga O /TiO
2
Amount of H chemisorbed (µmol/g)
2
2.4
0.2
0.1
0.2
0.2
2
3
Ga O /Al O
2
3
2 3
Ga O /ZrO
2
3
2
Ga O /SiO
2
3
2
Ga O /MgO
2
3
Table 8
◦
Amount of propane chemisorbed in the absence and presence of CO (300 C)
2
Catalyst
Amount in the
absence of CO
(µmol/g)
Amount in the
presence of CO
(µmol/g)
Decreasing
ratio
(%)
2
2
a
Fig. 5. Reaction data of CO and H pulse reaction: (2) Ga O /TiO ;
2
2
2
3
2
Ga O /TiO
2
4.2
5.3
8.1
1.1
2.0
2.5
3.7
5.9
1.1
1.5
40
30
27
0
2
3
(") Ga O /Al O ; (Q) Ga O /ZrO ; (a) Ga O /SiO ; (F) Ga O /MgO.
2
3
2
3
2
3
2
2
3
2
2 3
Ga O /Al O
2
3
2 3
Ga O /ZrO
2
3
2
Ga O /SiO
2
3
2
Ga O /MgO
25
2
3
a
Decreasing ratio = (1 – amount in the presence of CO )/(amount in the
2
absence of CO ) × 100%.
2
with the rate constant of k3:
H− H+
k1
Gax+–O2−–My+ + H2 ꢀ Gax+–O2−–My+
k
2
CO2
−→ Gax+–O2−–My+ + CO + H2O
k
3
(M = Ga, Ti, Al, Zr, Si, or Mg).
(3)
Because Ga2O3/ZrO2, Ga2O3/TiO2, Ga2O3/Al2O3, and
Ga2O3/MgO are good catalysts for the reverse water–gas shift
reaction, and because the amount of CO2 injected is much
larger than that of adsorbed H2, the amount of CO produced
through reaction (3) should connect directly with the amount
of adsorbed H2 remaining after purging with He for 10 min.
The more CO produced, the more adsorbed H2 remaining, and
hence the smaller the k2. Therefore, the CO2 pulse reaction re-
sults suggest that the k2 of Ga2O3/TiO2 is much smaller than
that of Ga2O3/ZrO2, Ga2O3/Al2O3, and Ga2O3/MgO.
Fig. 6. Reaction data of CO pulse reaction: (2) Ga O /TiO ; (") Ga O /
2
2
3
2
2 3
Al O ; (Q) Ga O /ZrO ; (a) Ga O /SiO ; (F) Ga O /MgO.
2
3
2
3
2
2
3
2
2 3
with activity increasing with increasing reaction temperature
for all catalysts. At 600 ◦C, the extent of activity followed
the order Ga2O3/ZrO2 > Ga2O3/TiO2 ≈ Ga2O3/MgO >
Ga2O3/Al2O3 ꢀ Ga2O3/SiO2.
The reverse water–gas shift reaction can proceed only when
both H2 and CO2 are adsorbed on the catalyst surface, as
reported previously [22]. Purportedly, H2 can be dissocia-
tively adsorbed on gallium oxide and supported gallium oxide
[20,23]; meanwhile, basic sites are also needed to adsorb CO2.
Therefore, the low activity of the reverse water–gas shift reac-
tion over Ga2O3/SiO2 can be attributed to its small amount of
basic sites (Table 2).
3.5. H2 and propane chemisorption test
To gain more information on the role of CO2 in the dehy-
drogenation reaction, the H2 and propane chemisorption ca-
pacity of the supported gallium oxide catalysts was measured;
the results are given in Tables 7 and 8. The tables show that
at 300 ◦C, the H2 chemisorption capacity of Ga2O3/TiO2 was
much higher than that of the other four catalysts, indicating
that the chemisorbed H2 over Ga2O3/TiO2 is more stable and
difficult to desorb, in agreement with the results of the pulse
reactions.
The special H2 adsorption property of Ga2O3/TiO2 reported
earlier may be connected to the abundant reduced gallium
atoms on the Ga2O3/TiO2 surface (45%), because reduced gal-
lium ions (Gaδ+ cations, δ < 2) are thought to have higher
dehydrogenation efficiency [25], and a gallium–hydrogen bond
can be formed on these reduced gallium ions by heterolytic
CO2 pulse reaction was also applied for further research.
The results are illustrated in Fig. 6. When the supported gal-
lium oxide catalyst was pretreated with H2 for 30 min, a certain
amount of H2 was heterolytically dissociatively adsorbed on the
catalyst surface [24], with the rate constant of k1 as shown in
reaction (3). Then the catalyst was treated with He for 10 min
to remove H2 in the reaction system. (That a certain amount
of the chemisorbed H2 desorbed simultaneously from the cata-
lyst surface via the reverse reaction with the rate constant of k2
was unavoidable, however.) Finally, a sufficient pulse of CO2
was injected to react with H2 remaining on the catalyst surface,

476
B. Xu et al. / Journal of Catalysis 239 (2006) 470–477
hydrogen dissociation and can be stabilized on the support sur-
face [20]. Only a small fraction or even no reduced gallium
species could be detected on the other four supported gallium
oxide catalysts, leading to the rapid desorption of H2 on the sur-
face.
The results also show that the propane adsorption capac-
ity decreased when CO2 was introduced to all catalyst except
Ga2O3/SiO2, which had a very low adsorption capacity. This
decrease may be due to displacement of adsorbed propane by
the CO2 introduced, because CO2 is obviously much more
acidic than propane and thus much more readily adsorbed on
the basic catalyst sites. But on Ga2O3/SiO2, only a very small
number of basic sites were detected by CO2-TPD, indicating
the catalyst’s weak CO2 adsorption ability; therefore, adding
CO2 in the gas flow has hardly any effect on this catalyst’s
propane adsorption capacity.
The conjugated effect of gallium oxide and proton increases
propane dehydrogenation activity by replacing the slow step (5)
by the fast equilibria (7) and (8). Therefore, the amount of
proton existing on the surface of the supported gallium oxide
catalyst should have a considerable effect on the dehydrogena-
tion activity.
The above reaction mechanism gives a good explanation for
the different activities of the five supported gallium oxide cat-
alysts toward the dehydrogenation reaction in the absence of
CO2, that is, the high activity of Ga2O3/ZrO2, Ga2O3/TiO2,
and Ga2O3/Al2O3 and the low activity of Ga2O3/MgO and
Ga2O3/SiO2. NH3-TPD studies showed that the medium to
strong acid sites are more abundant on the surface of Ga2O3/
ZrO2, Ga2O3/TiO2, and Ga2O3/Al2O3 than on Ga2O3/MgO
and Ga2O3/SiO2, which would facilitate the reactions (7)
and (8) to bypass the slow step (5), thus enhancing the ac-
tivity of propane dehydrogenation. The amount of medium to
strong acid sites has the order Ga2O3/ZrO2 > Ga2O3/Al2O3
> Ga2O3/TiO2 > Ga2O3/SiO2 > Ga2O3/MgO (Table 1), ex-
actly the same order as for the initial activity.
4. Discussion
The dehydrogenation of propane or ethane over reducible
metal oxide catalysts, such as chromium and iron oxides, in
the presence of CO2 has been suggested to follow a redox
mechanism [26,27]. Propane is oxidized to propene with the si-
multaneous reduction of metal oxide (e.g., Fe2O3, Cr2O3), and
subsequently the reduced metal oxide catalyst is reoxidized by
CO2. According to our experimental results, supported gallium
oxide catalysts cannot be reduced under the reaction temper-
ature, and thus the redox mechanism is probably inapplicable
for this reaction. It has been suggested that propane can be het-
erolytically dissociatively adsorbed on gallium oxide, forming
gallium hydride and gallium alkoxide species [24]; the same
process may occur over supported gallium oxide catalysts, as
follows:
When CO2 is introduced into the propane dehydrogenation
reaction, chemisorbed H2 formed from steps (5) and (7) can
also be removed via an alternative route (9) (i.e., the reverse
water–gas shift reaction), besides step (6):
H− H+
Gax+–O2−–My+ + CO2 → Gax+–O2−–My+ + CO + H2O.
(9)
In fact, the major pathway of the removal of the dissociatively
adsorbed H2 depends on the relative rate of steps (6) and (9).
Our pulse reaction tests and H2 chemisorption measurements
indicate that step (6) is relatively slow over Ga2O3/TiO2.
Therefore, the fast step (9) becomes the preferred reaction in-
stead of the slow step (6) after CO2 is introduced, because
Ga2O3/TiO2 has proven to be a good catalyst for the reverse
water–gas shift reaction, which transforms H2 with CO2 into
CO and H2O, shifting the reactions (5) and (7) to the prod-
uct side and thus enhancing the total reaction rate. In contrast,
step (6) is relatively fast over Ga2O3/Al2O3, Ga2O3/ZrO2,
and Ga2O3/MgO and thus will still be the dominant reac-
tion pathway for the removal of chemisorbed H2 after in-
troduction of CO2. Thus, the enhancement effect of CO2 on
the total reaction rate is not evident for these catalysts. This
can be further proven by the fact that the amount of CO2
reacted on Ga2O3/TiO2 via reaction (9) (2 × 20%) is close
to the conversion of propane (32%) in the initial state (be-
cause the molar ratio of CO2/C3H8 equals 2), whereas for
Ga2O3/Al2O3, Ga2O3/ZrO2, and Ga2O3/MgO catalysts, the
amount of CO2 participating in the reverse water–gas shift (9)
is much lower than the conversion of propane (Tables 5 and 6).
For Ga2O3/SiO2, because activity for the reverse water–gas
shift reaction is very low, it would be difficult for CO2 to react
with the chemisorbed H2 over Ga2O3/SiO2, and thus a positive
effect of CO2 would not be observed.
H− C3H7
+
Gax+–O2−–My+ + C3H8 → Gax+–O2−–My+
.
(4)
The alkoxides further decompose to form the dehydrogenation
products,
H− C3H7
H− H+
+
Gax+–O2−–My+ → Gax+–O2−–My+ + C3H6
and
(5)
H− H+
Gax+–O2−–My+ → Gax+–O2−–My+ + H2.
(6)
Step (5) is slow and represents the limiting step in propene
formation [24]. When both Ga2O3 and H+ are present on the
catalyst, the propyl carbenium ion on Ga2O3 will readily ex-
change with a proton via a surface migration reaction [24],
H− C3H7
H− H+
+
Gax+–O2−–My+ + H+S ꢀ Gax+–O2− –My+ + C3H7+S,
(7)
in which S represents catalyst surface. Propene then results
from the equilibrium,
Meanwhile, CO2 has a negative effect on the reaction by
displacing the propane adsorbed on the catalyst surface [4,11].
Propane is dissociatively adsorbed on gallium oxide; basic sites
C3H7+S ꢀ C3H6 + H+S.
(8)

B. Xu et al. / Journal of Catalysis 239 (2006) 470–477
477
are needed for the adsorption of C3H7+, and acidic sites are
needed for the adsorption of H−. Because CO2 is much more
acidic than propane, the adsorption of CO2 on the basic sites
would certainly reduce the possibility of the adsorption of
propane on gallium oxide, leading to reduced propane conver-
sion. This phenomenon is demonstrated by the greatly reduced
adsorption capacity of propane over most supported catalysts
after the introduction of CO2 (Table 8). For those catalysts with
activity enhanced by the addition of CO2, such as Ga2O3/TiO2,
the negative effect of CO2 displacement on the reaction is not
as evident at a low concentration of CO2, and propane conver-
sion increases with increasing CO2/C3H8 ratio. When the CO2
concentration increases, this negative effect becomes dominant,
and the propane conversion begins to decrease. For such cata-
lysts as Ga2O3/Al2O3 and Ga2O3/ZrO2, the addition of CO2
does not have a promoting effect on activity, the negative ef-
fect of CO2 will always prevail. This explains why the propane
conversion over Ga2O3/TiO2 has a maximum with increasing
CO2/C3H8 ratio, whereas the conversion over Ga2O3/Al2O3
and Ga2O3/ZrO2 decreases continuously from the very begin-
ning with increasing CO2/C3H8 ratio (Fig. 4). For Ga2O3/SiO2
and Ga2O3/MgO, the rate-determining step (reaction (5)) is
very slow, +because there are insufficient protons to exchange
with C3H7 via reaction (7) to circumvent this slow reaction.
Therefore, the decreased rate of reaction (4) by the displace-
ment of CO2 against propane would have only a limited nega-
tive effect on the rate-limiting step (5) (Fig. 4).
ative role by displacing propane adsorbed on the catalyst’s
basic sites. Dehydrogenation proceeds on the supported gallium
oxide catalysts probably through a heterolytic dissociation re-
action pathway instead of a redox mechanism, similar to that
on pure gallium oxide.
Acknowledgments
This work was supported by the Chinese Major State
Basic Research Development Program (grant 2000077507),
the National Natural Science Foundation of China (grant
20303004), the Shanghai Major Basic Research Program (grant
03DJ14004), and the Shanghai Natural Science Foundation
(grant 03ZR14013).
References
[1] I. Takahara, M. Saito, Chem. Lett. (1996) 973.
[2] I. Takahara, W.C. Chang, N. Mimura, M. Saito, Catal. Today 45 (1998)
55.
[3] B. Zhaorigetu, R. Kieffer, J.P. Hindermann, Stud. Surf. Sci. Catal. 101
(1996) 1049.
[4] K. Nakagawa, M. Okamura, N. Ikenaga, T. Suzuki, T. Kobayashi, Chem.
Commun. (1998) 1025.
[5] P. Meriaudeau, C. Naccache, J. Mol. Catal. 50 (1989) L7.
[6] P. Schulz, M. Baerns, Appl. Catal. 78 (1991) 15.
[7] M. Guisnet, N.S. Gnep, F. Alario, Appl. Catal. A 89 (1992) 1.
[8] Y. Ono, Catal. Rev. Sci. Eng. 34 (1992) 179.
[9] B.S. Kwak, W.M.H. Sachtler, J. Catal. 145 (1994) 456.
[10] P. Meriaudeau, C. Naccache, Catal. Today 31 (1996) 265.
[11] K. Nakagawa, C. Kajita, Y. Ide, M. Okamura, S. Kato, H. Kasuya, N. Ike-
naga, T. Kobayashi, T. Suzuki, Catal. Lett. 64 (2000) 215.
[12] K. Nakagawa, C. Kajita, K. Okumura, N. Ikenaga, M. Nishitani-Gamo,
T. Ando, T. Kobayashi, T. Suzuki, J. Catal. 203 (2001) 87.
[13] P. Michorczyk, J. Ogonowski, Appl. Catal. A 251 (2003) 425.
[14] P. Michorczyk, J. Ogonowski, React. Kinet. Catal. Lett. 78 (2003) 41.
[15] M. Saito, S. Watanabe, I. Takahara, M. Inaba, K. Murata, Catal. Lett. 89
(2003) 213.
[16] B. Zheng, W.M. Hua, Y.H. Yue, Z. Gao, J. Catal. 232 (2005) 143.
[17] R. Carli, C.L. Bianchi, R. Giannantonio, V. Ragaini, J. Mol. Catal. 83
(1993) 379.
[18] V. Parvulescu, S. Coman, V.I. Parvulescu, P. Grange, G. Poncelet, J. Ca-
tal. 180 (1998) 66.
[19] M. Haneda, Y. Kintaishi, H. Shimada, H. Hamada, Chem. Lett. (1998)
181.
[20] S.E. Collins, M.A. Baltanas, J.L.G. Fierro, A.L. Bonivardi, J. Catal. 211
(2002) 252.
[21] B.M. Reddy, I. Ganesh, E.P. Reddy, A. Fernandez, P.G. Smirniotis,
J. Phys. Chem. B 105 (2001) 6227.
[22] C.S. Chen, W.H. Cheng, S.S. Lin, Appl. Catal. A 238 (2003) 55.
[23] S.E. Collins, M.A. Baltanas, A.L. Bonivardi, Langmuir 21 (2005) 962.
[24] P. Meriaudeau, C. Naccache, J. Mol. Catal. 59 (1990) L31.
[25] I. Nowak, J. Quartararo, E.G. Derouane, J.C. Védrine, Appl. Catal. A 251
(2003) 107.
[26] S.B. Wang, K. Murata, T. Hayakawa, S. Hamakawa, K. Suzuki, Appl.
Catal. A 196 (1998) 1.
[27] P. Michorczyk, P. Kustrowski, L. Chmielarz, J. Ogonowski, React. Kinet.
Catal. Lett. 82 (2004) 121.
5. Conclusion
Dehydrogenation of propane to propene in the absence
or presence of CO2 over five supported gallium oxide cata-
lysts was investigated, and distinct behaviors were observed.
Ga2O3/TiO2, Ga2O3/Al2O3, and Ga2O3/ZrO2 are better cat-
alysts for the dehydrogenation reaction than Ga2O3/SiO2
and Ga2O3/MgO, because they contain many more acid
sites of medium to strong and strong strength on the sur-
face. CO2 has a promoting effect on dehydrogenation activ-
ity over Ga2O3/TiO2 but a negative effect over Ga2O3/ZrO2
and Ga2O3/Al2O3. These different support effects may de-
rive from the catalysts’ differing H2 adsorption capacities and
acid–base properties, probably caused by the differing interac-
tions between the gallium oxide and the support. XPS studies
showed abundant reduced gallium atoms (45%) on the sur-
face of Ga2O3/TiO2, which may account for this catalyst’s
unique catalytic behavior in the dehydrogenation reaction in
the presence of CO2. Results for the pulse reaction and de-
hydrogenation reaction with varying partial pressures of CO2
indicate that CO2 plays two roles in the dehydrogenation reac-
tion: a positive role by removing absorbed H2 on the catalyst
surface through the reverse water–gas shift reaction and a neg-