Dehydrogenation of ethylbenzene over vanadium oxide-loaded MgO catalyst: Promoting effect of carbon dioxide

Article abstract of DOI:10.1006/jcat.2002.3593

The promoting effect of CO₂ on the dehydrogenation of ethylbenzene over vanadium oxide-loaded MgO catalyst was studied. The styrene yield in the presence of CO₂ was 2.5 times higher than that in the absence of CO₂ (argon a

Full text of DOI:10.1006/jcat.2002.3593

Journal of Catalysis 209, 16–24 (2002)
doi:10.1006/jcat.2002.3593
Dehydrogenation of Ethylbenzene over Vanadium Oxide-Loaded MgO
Catalyst: Promoting Effect of Carbon Dioxide
∗
∗
∗,
∗
Yoshihiro Sakurai, Takamasa Suzaki, Kiyoharu Nakagawa, ‡ Na-oki Ikenaga,
∗
, ,1
Hiroyuki Aota,† and Toshimitsu Suzuki ‡
∗
Department of Chemical Engineering, †Department of Applied Chemistry, Faculty of Engineering, and ‡High Technology Research Center,
Kansai University, Suita, Osaka, 564-8680, Japan
Received July 19, 2001; revised November 26, 2001; accepted February 25, 2002
tion; with this process, the reaction becomes exothermic
The dehydrogenation of ethylbenzene over a vanadium oxide- and can be conducted at a low temperature. Several investi-
loaded MgO catalyst was investigated using carbon dioxide. The gators have reported that a magnesium vanadate catalyst is
styrene yield in the presence of carbon dioxide was 2.5 times higher
than that in the absence of carbon dioxide (argon atmosphere) at
the best catalyst for this reaction. Various physicochemical
measurements, including X-ray diffraction (2–6), EXAFS
8
23 K, indicating that carbon dioxide markedly promoted the dehy-
(
5), ESR (3, 12), UV–visible (9), Raman (10), IR (6–8), and
drogenation of ethylbenzene. At 873 K, the same catalyst afforded
the highest styrene yield, 73.8% with a selectivity of 90.1%, in the
presence of carbon dioxide. In order to elucidate the role of car-
bon dioxide in this reaction, characterization of the catalyst was
carried out via methods such as temperature-programmed reduc-
tion, temperature-programmed reaction with carbon dioxide, and
temperature-programmed reduction (2), have been used to
investigate the active species of the catalyst and the reac-
tion mechanism, i.e., redox properties. Lopez-Cartes and
coworkers reported that the oxidative dehydrogenation of
propane proceeded with a redox cycle between Mg3V2O8
UV–visible, FT-IR, and XRD spectroscopies. Carbon dioxide be- and MgV2O4 (5), while on the other hand, Chang et al. re-
haved as an oxidant for the vanadium species, and the surface ported a redox cycle between Mg3V2O8 and Mg2VO4 dur-
vanadium species were kept in a high oxidation state with car- ing the dehydrogenation of ethylbenzene (2).
bon dioxide during the dehydrogenation reaction. Active phases
In this process, gas-phase oxygen is indispensable in
maintaining the catalytic activity because the vanadium ac-
of vanadium in the dehydrogenation reaction were believed to
5+
be V species in V2O5 or Mg3V2O8 on highly dispersed MgO.
tive species is reduced by the loss of a lattice oxygen of vana-
dium oxide with abstracting a hydrogen atom from alkanes.
In the absence of oxygen, the catalytic activity sharply de-
creases with a decrease in the amount of available lattice
oxygen (2, 11). However, in the presence of oxygen, the
deep oxidation of alkanes to COx occurs more frequently,
decreasing the desired olefin selectivity. Thus, this new de-
hydrogenation process cannot replace the process currently
used without improving the olefin selectivity.
4
+
3+
The reduced species, V and V , were less reactive sites for the
dehydrogenation. ꢀc 2002 Elsevier Science (USA)
INTRODUCTION
Recently, increasing demand for styrene as one of the
important monomers for synthetic polymers prompted sev-
eral investigators to exploit a new process which reduces
the cost of styrene production. Styrene monomer is com-
mercially produced by means of the dehydrogenation of
ethylbenzene using iron oxide as a catalyst, with an excess
amount of superheated steam as a diluent and heat carrier
On the other hand, the use of carbon dioxide as a dilu-
ent instead of steam makes it possible to reduce the en-
ergy of the process and the cost of styrene production with
high selectivity. If the present process using steam is re-
placed with the new process using carbon dioxide, the en-
ergy required for styrene production is calculated to be
(
1). Since the product and effluent are rapidly cooled to
avoid the polymerization of styrene, this process wastes a
large amount of latent heat of vaporizing water. Currently,
the development of a new energy-saving and more econom-
ical dehydrogenation process is of great importance.
The oxidative dehydrogenation of ethylbenzene (2–4)
and short-chain alkanes (5–12) has received much atten-
6
6
from 6.3 × 10 to 2.6 × 10 kJ/1000 kg of styrene (13). It is
of additional interest that the gas that causes global warm-
ing, carbon dioxide, is effectively utilized in this chemical
reaction.
One of the authors of the present study first reported that
the dehydrogenation of ethylbenzene with an iron-loaded
activated-carbon catalyst is promoted with carbon dioxide,
and it was concluded that carbon dioxide acts as an oxi-
dant in the redox cycle of iron oxide (14). Following this
1
To whom correspondence should be addressed. Fax: +81-6-6388-8869.
E-mail: tsuzuki@ipcku.kansai-u.ac.jp.
16
0
ꢀ
021-9517/02 $35.00
c 2002 Elsevier Science (USA)
All rights reserved.

DEHYDROGENATION OF ETHYLBENZENE OVER MgO CATALYST
17
discovery, several investigators reported the use of carbon with activated carbon.
dioxide in the dehydrogenation of ethylbenzene with an
iron-based catalyst (13, 15–17). However, to our knowl-
edge, except for our previous report there has been no re-
port about catalysts, other than those that are iron based,
which are applicable to this new dehydrogenation process.
ꢀ
ꢁ
EB recovered in traps in mol
weight loss in EB saturator/106
Conversion (%) = 1 −
× 100 (%),
[1]
In our previous study, the styrene yield in a carbon dioxide where EB is ethylbenzene.
atmosphere was 1.27 times higher than that in an argon at-
Catalyst Characterization
mosphereat823Koveravanadiumoxide-loadedactivated-
carbon catalyst. Among several supports of vanadium
oxide, initial activity was not as high as that of activated
carbon, but MgO exhibited the highest promoting effect
of CO2 (18). In this paper, we report on the remarkable
promoting effect of carbon dioxide on the dehydrogena-
tion of ethylbenzene over vanadium oxide-loaded MgO
catalyst, and in order to elucidate the role of carbon diox-
ide in this reaction, characterizations of the catalyst were
performed.
The surface area of the catalyst was measured by the
BET method at 77 K using nitrogen as the adsorbate, with
an automatic Micromeritics Gemini model 2375.
The temperature-programmed reduction was carried out
with an automatic temperature-programmed desorption
apparatus equipped with a quadrupole mass spectrometer
as a detector (TPD-1-AT, BEL JAPAN INC.). Fresh cata-
lyst (70 mg) was first treated with an O2 (10 ml/min)/He
(
40 ml/min) mixture at 823 K for 1 h and cooled to 323 K
in the same atmosphere. After replacing the O2/He mix-
ture with He, the catalyst was heated from 323 to 1073 K
at a heating rate of 10 K/min under a stream of a H2
EXPERIMENTAL
(
5 ml/min)/He (45 ml/min) mixture.
The catalysts treated under different conditions were an-
Materials
All chemicals were used without further purification.
Four kinds of magnesium oxides (Ube Materials Industries
Ltd.), which have different surface areas, were used as
supports. The catalyst was prepared by impregnating MgO
with NH4VO3 according to following procedure. A certain
amount of NH4VO3 was dissolved in an aqueous solution
of oxalic acid (0.25 mol/L). To this solution MgO was
added and it was allowed to stand overnight. Excess water
was evaporated in vacuo, and the catalyst was calcined in
air at 873 K for 4 h.
alyzed by a Shimadzu UV-310PC UV–visible spectropho-
tometer with a diffuse reflectance mode. Fourier transform
infrared (FT-IR) spectra were obtained by transmission
mode with a JEOL JIR 7000. The catalyst was mixed with
KBr and pressed into a thin pellet.
Powder X-ray diffraction was obtained with a Shimadzu
model XRD-6000 diffractometer with monochromated Cu
Kα radiation.
The temperature-programmed reaction with carbon di-
oxide was carried out by means of an online quadrupole
mass spectrometer (HAL 201, Hiden Analytical Ltd.), and
carbon monoxide (m/Z = 28) was continuously monitored.
The fresh catalyst was first reduced with hydrogen at 823 K
for 1 h. After cooling to room temperature, the catalyst was
heated from room temperature to 1073 K at a heating rate
of 20 K/min under a stream of carbon dioxide.
Catalytic Tests
The reaction was carried out with a fixed-bed flow-type
quartz reactor operated at atmospheric pressure. A catalyst
(
50–150 mg) was placed at the center of the reactor using
quartz glass wool plugs. The reaction was carried out at 773–
73 K for 1–6 h. Ethylbenzene was fed (ca. 1.5 mmol/h) to
8
RESULTS
the reactor by passing carbon dioxide or argon (30 ml/min)
through the ethylbenzene saturator thermostated at 316 K.
The effluent from the reactor was condensed in two traps
containing heptane connected in a series. The condensed
Catalytic Tests
Table 1 shows the effect of the reaction temperature on
material was cooled externally in an ice water bath. The the dehydrogenation of ethylbenzene supported on MgO
gaseous products were collected in a gas bag. and bulk vanadium oxide catalysts in the presence and ab-
Reaction products (ethylbenzene, styrene, toluene, and sence of carbon dioxide. The numerals of YCO2/YAr in-
benzene) were analyzed with a gas chromatograph equip- dicate the ratio of a styrene yield (reaction for 1 h) in the
ped with a FID (Shimadzu GC14A), using a 3 mm × 3 m presence of carbon dioxide (CO2/ethylbenezene = 45) to
glass column packed with silicon SE-30. Carbon mate- that in an argon atmosphere. For a vanadium oxide-loaded
rial balances were above 95% in all the cases. Analyses MgO (V/MgO) catalyst, styrene yield increased with an in-
of gaseous products (CO, CO2, and H2) were carried out crease in reaction temperature and styrene selectivity was
with a gas chromatograph equipped with a TCD (Shimadzu kept constant above 90% at these temperatures in both
GC8A), using a 3 mm×2.5 m stainless steel column packed atmospheres. The styrene yield in the presence of carbon

1
8
SAKURAI ET AL.
TABLE 1
Effect of Reaction Temperature on Dehydrogenation of Ethylbenzene over Various Vanadium Oxide Catalystsa
Under CO2
ST yield
Under Ar
ST yield
Temp
(K)
ST sel
(%)
CO
H2
ST sel
(%)
CO
H2
Entry Catalyst
EB Conv
%
mmol
mmol (mmol) EB Conv
%
mmol
(mmol) (mmol) YCO2/YAr
1
2
3
4
5
V/MgO 773
17.2
59.1
50.2
81.9
21.5
15.8
53.8
46.3
73.8
13.9
0.26
0.74
0.68
1.17
0.19
92.2
91.1
92.4
90.1
64.7
0.32
1.09
1.08
1.95
0.15
0.00
0.05
0.04
0.14
0.20
8.8
22.7
26.3
67.1
21.8
8.3
21.9
24.3
62.2
15.3
0.13
0.32
0.34
0.85
0.23
95.2
96.5
92.3
92.6
70.2
0
0.1
0
0.1
0.1
0.11
0.43
0.39
1.20
0.28
1.90
2.46
1.91
1.19
0.91
V/MgO 823
V/MgO 823b
V/MgO 873
c
V2O5
823
a
2
Catalyst, 100 mg; V , 1.0 mmol/g of MgO; surface area, 144 m /g; reaction time, 1 h; W/F, 70 g of cat · h/mol.
b
c
Catalyst was reduced with H2 at 823 K for 6 h.
2
Bulk vanadium oxide surface area less than 1 m /g.
dioxide was 2.5 times higher than that in an argon atmo- the styrene yield increased, but the styrene selectivity grad-
sphere at 823 K, showing the remarkable promoting effect ually decreased even in the presence of carbon dioxide.
of carbon dioxide. At 873 K, the V/MgO catalyst afforded This was due to an increase in the amount of by-products,
the highest styrene yield (73.8%) in the presence of carbon especially benzene. However, the amounts of benzene and
dioxide. These results indicate that the V/MgO catalyst is an toluene did not exceed about 3% of styrene and carbon
excellent catalyst for the dehydrogenation of ethylbenzene mass balance was over 95%. In the argon atmosphere, the
in the presence of carbon dioxide. Prior to the dehydro- styrene yield was low and it increased slightly with an in-
genation of ethylbenzene, the catalyst was treated with hy- crease in W/F.
◦
drogen at 550 C for 6 h. The styrene yield in the presence
Figure 3 shows the effect of specific surface areas of the
of carbon dioxide decreased greatly, but that in the pres- catalyst on the dehydrogenation of ethylbenzene. The cata-
ence of argon increased slightly. When bulk vanadium(V) lyst having a small surface area yielded a small amount of
oxide was tested in the dehydrogenation reaction, only a styrene in both atmospheres, and the yield did not depend
low catalytic activity was observed in both atmospheres as on the atmosphere of the reaction. Styrene yield markedly
compared to that with the V/MgO catalyst, and the promot- increased with an increase in the surface area of the cata-
ing effect of carbon dioxide was not observed. Its effect was lyst in the presence of carbon dioxide. However, in an argon
only observed on the V/MgO catalyst, in which vanadium atmosphere, the styrene yield did not increase with an in-
species were highly dispersed throughout the MgO support. crease in the surface area of the catalyst. The significant
In Table 1 the amounts of CO and H2 during the reaction
are also shown. Within experimental errors, the amount of
CO produced agreed with the amount of ethylbenzne con-
verted in the runs under CO2. However, in runs under Ar
only H2 was produced, with very small amounts of CO.
Figure 1 shows the effect of the vanadium-loading level
on the styrene yield and selectivity in the presence and ab-
sence of carbon dioxide. A bare MgO exhibited very low
styrene yield and selectivity in both atmospheres. In the
presence of carbon dioxide, the styrene yield increased with
anincreaseintheloadinglevelfrom0to1.0mmol/gofMgO,
butfurtherincreasesintheloadingleveldidnotincreasethe
styrene yield. On the other hand, in the argon atmosphere,
the styrene yield monotonously increased with increases in
the loading level. The styrene selectivity decreased above
1
.5 mmol/g of MgO of the vanadium loading level in both
atmospheres.
Figure 2 shows the effect of the contact time factor (W/F)
on the dehydrogenation of ethylbenzene. In these experi-
ments, W/F was changed by varying the amount of the
catalyst charged under a constant ethylbenzene feed rate
FIG. 1. Effect of loading level of vanadium on the styrene yield and
the selectivity. Reaction temperature, 823 K; reaction time, 1 h; W/F,
70 g of cat · h/mol; catalyst, 100 mg. Ethylbenzene/CO2 (mol/mol) = 45.
Open symbols, under carbon dioxide flow; solid symbols, under argon flow;
(
1.5 mmol/h). With an increase in the contact time factor, triangles, styrene yield; boxes, styrene selectivity.

DEHYDROGENATION OF ETHYLBENZENE OVER MgO CATALYST
19
FIG. 4. Effect of partial pressure of carbon dioxide on the styrene
yield and the selectivity. Reaction temperature, 823 K; reaction time,
FIG. 2. Effect of W/F on the styrene yield and the selectivity. Reac-
tion temperature, 823 K; reaction time, 1 h; V , 1.0 mmol/g of MgO. Keys
are the same as shown in caption to Fig. 1.
1
h; W/F, 70 g of cat · h/mol; catalyst, 100 mg; V , 1.0 mmol/g of MgO.
Ethylbenzene/CO2 (mol/mol) = 45 at PCO = 98 kPa; ethylbenzene/CO2
2
(
mol/mol) = 9at PCO = 17kPa. (ꢀ)Ethylbenzeneconversion, (ꢁ)styrene
2
yield, (ꢂ) styrene selectivity.
promoting effect of carbon dioxide was only observed on
the catalyst with a large surface area, in which vanadium
species were highly dispersed on the MgO support.
an increase in styrene yield to give benzene and toluene as
by-products.
The effect of partial pressure of carbon dioxide on the de-
hydrogenation of ethylbenzene was also investigated. The
results are shown in Fig. 4. The partial pressure of car-
bon dioxide was changed by diluting carbon dioxide with
argon, and the total flow rate was kept at 30 ml/min. In
the absence of carbon dioxide, a low styrene yield was ob-
served. However, by introducing a small amount of carbon
dioxide into the reaction, the styrene yield markedly in-
creased, and above a carbon dioxide partial pressure of
Figure 5 shows the styrene yield and selectivity as a func-
tion of the reaction time at 823 K in the presence or ab-
sence of carbon dioxide. In the initial stage of the reaction,
the styrene yield was very high in the presence of carbon
dioxide (54.3%). However, when the reaction period in-
creased, the styrene yield decreased gradually, and after
6
h, the styrene yield decreased to 37.3%. On the other
hand, under an argon atmosphere, the initial styrene yield
was low compared to the run under a carbon dioxide at-
mosphere, and the catalytic activity gradually increased to
a small extent during the run for 6 h.
5
0 kPa (partial pressure of ethylbenzene = 3.3 kPa) the
yield reached a level 2.5 times higher than that in the ab-
sence of carbon dioxide. Styrene selectivity decreased with
FIG. 3. Effect of specific surface area of catalyst on the styrene yield
and the selectivity. Reaction temperature, 823 K; reaction time, 1 h; W/F,
FIG. 5. Effect of time on stream on the styrene yield and the selectiv-
ity. Reaction temperature, 823 K; W/F, 70 g of cat · h/mol; catalyst, 100 mg;
V , 1.0 mmol/g of MgO. Keys are the same as shown in caption to Fig. 1.
7
0 g of cat · h/mol; catalyst, 100 mg; V , 1.0 mmol/g of MgO. Keys are the
same as shown in caption to Fig. 1.

2
0
SAKURAI ET AL.
The styrene yield increased by air oxidation of the used
catalystafterthereactionfor6hincarbondioxide, although
the activity could not completely return to its initial level.
On the other hand, the catalyst used under the argon atmo-
sphere exhibited a lower activity after the oxidation of the
catalyst.
During the treatment with oxygen (O2–Ar, 5–15 ml/min),
carbon dioxide and carbon monoxide were produced due to
the oxidation of deposited carbon over the catalyst. Judging
from the amount of produced COx , the coke deposition
seemed to proceed more significantly in the reaction under
a carbon dioxide atmosphere compared to that under an
argon atmosphere.
Catalyst Characterization
Temperature-programmed reduction with hydrogen.
The reducibility of the vanadium species over the V/MgO
catalyst was measured by means of a temperature-
programmed reduction with hydrogen. Figure 6a shows the
effect of the vanadium-loading level, and Figure 6b shows
FIG. 7. UV–visible diffuse reflectance spectra of V(1.0)/MgO cata-
lysts treated under various conditions: (a) fresh catalyst, (b) dehydro-
genated under carbon dioxide flow at 823 K for 6 h, (c) dehydrogenated
under argon flow at 823 K for 6 h, (d) fresh catalyst reduced with hydrogen
at 823 K for 6 h, and (e) treated with carbon dioxide at 823 K for 6 h after
the reduction with hydrogen.
the effect of the surface area of the catalyst on the reduc-
tion temperature of the vanadium species. Figure 6 (bot-
tom) shows that the support MgO itself afforded a small
H2O peak at 525 K. This seems to be the absorbed water
or hydroxy groups on MgO. On the other hand, in the
V/MgO catalysts, the reduction peak appeared at 908 K,
and the amount of water produced by the reduction in-
creased with an increase in the loading level of vanadium
from 0.5 to 1.5 mmol/g of MgO. However, at the higher
loading level of 3.0 mmol/g of MgO, the reduction peak
shifted to a high-temperature region compared to that of
low-vanadium-loading catalysts, and multiple overlapping
reduction peaks were observed. The reduction tempera-
ture shifted to a higher temperature region (1050 K) with a
smaller amount of water when a smaller surface area cata-
2
lyst (91 m /g) was examined (Fig. 6b).
UV–visible spectra. Figure 7a shows the UV–visible dif-
fuse reflectance spectrum of the fresh V/MgO catalyst. The
absorption band at 270 nm indicates the presence of tetra-
5
+
hedral V species (9). After the dehydrogenation of ethyl-
benzene in the presence of carbon dioxide (Fig. 7b), this ab-
sorption was broadened and weakened. However, a weak
and broad band with an absorption maximum between 500
and 700 nm appeared in the catalyst after the reaction under
the argon atmosphere (Fig. 7c). Based on the fact that this
broad band was observed in the hydrogen-reduced V/MgO
catalyst (Fig. 7d), it is obvious that the reduction of the
vanadium species proceeded during the dehydrogenation
reaction in the argon atmosphere. On the other hand, the
FIG. 6. Temperature-programmed reduction profiles for vanadium
oxide-loaded MgO catalyst. (a) variations in the vanadium-loading level.
Dotted line shows peak temperature of 910 K. (b) variation in the surface
area. Catalyst, 70 mg; heating rate, 10 K/min; flow rate, He-45 ml/min,
H2-5 ml/min.

DEHYDROGENATION OF ETHYLBENZENE OVER MgO CATALYST
21
vanadium species was kept at a higher oxidation state in
the carbon dioxide atmosphere. Figure 7e shows the spec-
trum of the catalyst treated with carbon dioxide for 6 h at
8
23 K after the reduction with hydrogen. The broad band
assigned to the reduced vanadium species disappeared af-
ter being treated with carbon dioxide at 823 K, and distinct
absorptions at 270 and 245 nm were observed.
IR spectroscopy. The FT-IR spectra of various V/MgO
catalysts under different conditions are shown in Fig. 8.
In these spectra, the contribution from MgO has been
subtracted.
Figure 8a shows the FT-IR spectra of a fresh catalyst.
−
1
This catalyst exhibited bands at 860 and 690 cm due to
Mg3V2O8 (6–8). These bands became weak after the de-
hydrogenation reaction in the presence of carbon dioxide
(
Fig. 8b), and they disappeared after a dehydrogenation
reaction in an argon atmosphere (Fig. 8c) and after the
reduction with hydrogen (Fig. 8d). However, these bands
appeared again after the treatment with carbon dioxide
(
Fig. 8e).
ESR spectroscopy. The ESR spectra of various V/MgO
catalysts under different conditions were measured in order
to investigate the changes in the oxidation state of the vana-
dium species. Details of the results have been described in
FIG. 9. XRD patterns of V(3.0)/MgO catalysts treated under various
a previous report (19). The ESR spectra demonstrated the conditions: (a) fresh catalyst, (b) fresh catalyst reduced with hydrogen at
oxidation capability of carbon dioxide for the vanadium 823 K for 6 h, and (c) treated with carbon dioxide at 823 K for 6 h after
the reduction with hydrogen.
species, and these results were consistent with the results of
UV–visible and FT-IR spectra.
X-ray powder diffraction. Figure 9 shows the XRD pat-
terns of the high-vanadium-loading catalyst. The fresh
catalyst with the high-vanadium loading, 3.0 mmol of vana-
dium/g of MgO, showed weak diffraction peaks assigned to
Mg3V2O8 (Fig. 9a), although for the low-vanadium-loading
catalyst only diffraction peaks assigned to MgO were ob-
served. Figure 9b shows the diffraction patterns of the
hydrogen-reduced catalyst. The diffraction peaks due to
Mg3V2O8 disappeared and new peaks assigned to a spinel-
type phase, Mg2VO4 and/or MgV2O4, appeared. After the
treatment with carbon dioxide, no significant changes in the
diffraction patterns of the catalyst were observed (Fig. 9c).
Temperature-programmed reaction with carbon dioxide.
In order to confirm the oxidation of the reduced vanadium
species by carbon dioxide, a temperature-programmed re-
action with carbon dioxide was carried out on the hydrogen-
reduced V/MgO catalyst. The result is shown in Fig. 10. The
catalyst afforded a small amount of carbon monoxide at a
temperature between 700 and 900 K. The results clearly
show that carbon dioxide oxidized the reduced vanadium
species to a high-oxidation-state vanadium species and that
carbon dioxide was reduced to carbon monoxide during the
TPR study.
FIG. 8. FT-IR spectra of V(1.0)/MgO catalysts treated under various
conditions. Conditions are the same as shown in caption to Fig. 7.

2
2
SAKURAI ET AL.
However, it seems difficult to keep the vanadium species
at the initial high oxidation state throughout the reaction,
since the oxidation capability of carbon dioxide is weaker
than that of oxygen. This was supported by the ESR and
XRD analyses. Based on the fact that no significant changes
in the XRD patterns were observed between the reduced
catalyst and the carbon dioxide-treated catalyst and that
only a small amount of carbon monoxide was formed in the
temperature-programmed reaction with carbon dioxide, it
can safely be said that carbon dioxide can oxidize only the
surface species of the reduced catalyst, while oxidation of
the bulk of the reduced vanadium species cannot proceed
with carbon dioxide. As shown in Table 1, formation of CO
in the dehydrogenation under CO2 is strong evidence of
a redox cycle, although the possibility of hydrogenation of
CO2 with H2 produced by the simple dehydrogenation can-
not be ruled out. However, by such CO2 hydrogenation the
promoting effect of CO2 in the dehydrogenation cannot
easily be interpreted. Small amounts of formation of CO in
Ar seem to be due to a reaction of deposited carbon with
vanadium oxide.
FIG. 10. Temperature-programmed reaction of the reduced vana-
dium oxide with carbon dioxide on V/MgO catalyst. Catalyst, 100 mg;
V , 1.0 mmol/g of MgO; heating rate, 20 K/min; flow rate, CO2-20 ml/min.
H2 reduction: 823 K, 2 h → Ar → CO2 TPR.
DISCUSSION
With a decrease in W/F, the promoting effect of CO2
measured by yield in CO2/Ar decreased from 2.95 at W/F
of 110 to 2.4 at W/F of 35, even in the presence of carbon
dioxide, as seen in Fig. 2. On the other hand, conversion
in Ar did not significantly change with a decrease in W/F.
Such results can be interpreted as follows: The reoxidation
rate of a reduced vanadium species with carbon dioxide is
small. Thus, the reduction of the vanadium species proceeds
faster than the reoxidation of the reduced vanadium species
at a higher space velocity, resulting in a decrease in the
number of active sites and the apparent promoting effect
of carbon dioxide.
Table 2 shows the results of thermochemical calculations
of the oxidation of the lower-valency-state vanadium oxides
with carbon dioxide at 800 and 900 K (20). From the val-
ues of ꢀG in the respective reactions, it seems it would be
difficult for carbon dioxide to oxidize the bulk vanadium
oxide of a lower valency state, although it may be possible
for some of this reaction to occur. However, loading it onto
the support may alter the thermochemical parameters of
vanadium oxide in a mixed oxide. Especially over a MgO
In the previous paper, we showed that vanadium oxide
loaded on activated carbon exhibited a high catalytic activ-
ity in the dehydrogenation of ethylbenzene under carbon
dioxide (18). In the dehydrogenation of ethylbenzene over
a V/MgO catalyst at 823 K, the styrene yield in the pres-
ence of carbon dioxide was 2.5 times higher than that in
an argon atmosphere, indicating that the dehydrogenation
reaction was markedly promoted by carbon dioxide. The re-
sults of various characterization techniques, including UV–
visible, ESR, FT-IR, and CO2-TPR spectra clearly indicate
that lattice oxygen of the magnesium vanadate abstracts
the hydrogen from ethylbenzene, that water and styrene
are produced, and, consequently, that the vanadium species
is reduced. In an argon atmosphere, the vanadium species
is continuously reduced during the dehydrogenation reac-
tion. On the other hand, in a carbon dioxide atmosphere,
carbon dioxide can oxidize the reduced vanadium species,
and the vanadium species is kept in a higher oxidation state
(
Scheme 1), resulting in a high styrene yield compared to
that in an argon atmosphere.
TABLE 2
Oxidation of Lower Valent Vanadium Oxide with Carbon Dioxide
ꢁ
G (kJ/mol of V)
Reaction
800 K
900 K
VO2 + 1/2 CO2 → 1/2 V2O5 + 1/2 CO
78.27
114.35
73.35
78.72
113.81
72.27
1
/2 V2O3 + CO2 → 1/2 V2O5 + CO
VO + 3/2 CO2 → 1/2 V2O5 + 3/2 CO
1
/2 V2O3 + 1/2 CO2 → VO2 + 1/2 CO
36.08
35.09
VO + CO2 → VO2 + CO
−4.92
−6.45
SCHEME 1. Catalytic cycle of the dehydrogenation of ethylbenzene
in the presence of carbon dioxide.
VO + 1/2 CO2 → 1/2 V2O3 + CO
−41.00
−41.54

DEHYDROGENATION OF ETHYLBENZENE OVER MgO CATALYST
23
support, the strong interaction between acidic vanadia and MgO), a single and low-temperature peak was observed,
basic magnesia results in the formation of mixed oxides. but for the high-vanadium-loading catalysts (V = 3.0
UV–visible, ESR, and CO2-TPR experiments strongly sug- mmol/g of MgO), overlapping peaks at higher tempera-
gest that carbon dioxide can oxidize the lower-valency-state tures were detected. On the low-loading catalyst, an iso-
vanadium oxide to higher valency oxides.
lated VO4 species forming a Mg3V2O8 structure exists with
Considering that the promoting effect of carbon diox- a high dispersion over the MgO support. On the higher
ide was observed on the V/MgO catalyst with a large sur- loading case, in addition to this species, polymeric V5+
face area but not observed on bulk V2O5, the oxidation of species with bridging V–O–V oxide ions forming MgV2O6
lower-valency-state vanadium species with carbon dioxide and Mg2V2O7 structures coexist (10, 22), resulting in a de-
may occur only in mixed oxides highly dispersed on magne- crease in the styrene selectivity over a V(3.0)/MgO catalyst.
sium oxide. Therefore, even if the bulk lower-valency-state It is believed that the styrene yield did not increase above
vanadium oxides could not be oxidized, complex oxides of the loading level of 1.0 mmol/g of MgO since the isolated
magnesium vanadate could be oxidized to a higher valency VO4 species did not increase above this loading level in the
state with carbon dioxide.
presence of carbon dioxide.
According to the results shown in Fig. 5, in a carbon
This interpretation can be applied to the behavior shown
dioxide atmosphere, the styrene yield decreased with an in- in Fig. 3. For the catalyst with a small surface area, the
crease in the reaction period due to the coke deposition and dispersion of vanadium species lowered, and MgV2O6
increases in the lower-valency-state vanadium species. On and/or Mg2V2O7 were formed. Therefore, this catalyst did
the other hand, in an argon atmosphere, the initial styrene not afford the promoting effect of carbon dioxide.
yield was very low and the styrene yield slightly increased
Only vanadium species on a high-surface-area MgO ex-
with an increase in the reaction period. Presumably, under hibited high catalytic activity and large promoting effect of
an argon atmosphere, the dehydrogenation reaction pro- CO2. Bulk Mg3V2O8, and MgV2O6, and Mg2V2O7 com-
ceeds on the site different from that of the redox cycle pound oxides did not show high catalytic activity and pro-
with higher-valency-state oxides and lower-valency oxides. moting effect of CO2 in the dehydrogenation of ethyl-
Therefore, after the available active oxygen in the lattice of benzene and 3-ethylbipheny, although the surface areas of
the catalyst was consumed in the initial period, the catalytic these oxides were small (23). Catalytically active phase of
activity was not lost and the catalyst continued to produce V2O5 loaded on a high-surface-area MgO, in the dehydro-
a low yield of styrene.
genation of ethylbenzene under CO2, is difficult to deter-
Ishida et al. reported that the reduction of a vanadium mine whether Mg3V2O8 or highly dispersed V2O5 on MgO.
oxide-loaded Al2O3 catalyst creates a new Lewis acid site Results from IR spectra show the possibility of redox cy-
assigned to the reduced vanadium species (21). The fact cles of Mg3V2O8 and Mg2VO4, but in the XRD, it is ob-
that the styrene yield increased and styrene selectivity de- served that Mg2VO4 was not oxidized into Mg3V2O8. UV
5+
4+
creased with the hydrogen reduction of the V/MgO catalyst and ESR results only afforded redox cycles of V and V
.
in an argon atmosphere (Table 1) leads to the conclusion Thus at present, we could safely conclude that the active
that the active site changed and a new acidic site was created phase of CO2-promoted dehydrogenation of ethylbenzene
5
+
by the reduction during the dehydrogenation of ethylben- would be V species in V2O5 or Mg3V2O8 on highly dis-
zene in the argon atmosphere, resulting in a slight increase persed MgO.
in the styrene yield with an increase in the reaction period.
The active site in a carbon dioxide atmosphere would seem
to be the redox site and that in an argon atmosphere a weak
acid site of magnesium vanadate. The active sites differ ac-
ACKNOWLEDGMENTS
cording to the atmosphere.
This work was partially supported by a Grant-in-Aid for Scientific Re-
It is well-known that various magnesium vanadates search (No. 11650811) from the Japan Society for the Promotion of Science
can be formed and that among them meta-MgV2O6, (JSPS). K. Nakagawa is grateful for his research fellowship from JSPS for
Young Scientists.
pyro-Mg2V2O7, and ortho-Mg3V2O8 are the stable com-
pounds (10). According to Chang et al., the reducibility
of these vanadates differs and the order is reported to
be Mg3V2O8 > MgV O6 > Mg V2O7. Furthermore, in the
dehydrogenation of ethylbenzene in the presence of oxy-
gen, Mg3V2O8 is active and selective; on the other hand,
MgV2O6 and Mg2V2O7 are active but nonselective toward
styrene formation (2).
2
2
REFERENCES
1. Lee, E. H., Catal. Rev. Eng. Sci. 8, 285 (1973).
2. Chang, W. S., Chen, Y. Z., and Yang, B. L., Appl. Catal. A 124, 221
(
1995).
3
4
. Oganowski, W., Hanuza, W., Drulis, H., Mista, W., and Macalik, L.,
Appl. Catal. A 136, 143 (1996).
. Oganowski, W., Hanuza, J., and Kepinski, L., Appl. Catal. A 171, 145
(1998).
In our study, differences in the reduction temperature of
the vanadium species with hydrogen were observed (Fig. 6).
<
For the low-vanadium-loading catalyst (V 1.5 mmol/g of
=

2
4
SAKURAI ET AL.
5
. Burrows, A., Kiely, C. J., Perregaard, J., Hojlund-Nielson, P. E.,
Vorbeck, G., Calvino, J. J., and Lopez-Cartes, C., Catal. Lett. 57, 121
1999).
15. Chang, J. S., Park, S. E., and Park, M. S., Chem. Lett. 1123
(1997).
16. Badstube, T., Papp, H., Kustrowski, P., and Dziembaj, R., Catal. Lett.
55, 169 (1998); Appl. Catal. A 204, 153 (2000).
17. Mimura, N., and Saito, M., Catal. Lett. 58, 59 (1999); Catal. Today 55,
173 (2000).
18. Sakurai, Y., Suzaki, T., Ikenaga, N., and Suzuki, T., Appl. Catal. A 192,
281 (2000).
19. Sakurai, Y., Suzaki, T., Nakagawa, K., Ikenaga, N., Aota, H., and
Suzuki, T., Chem. Lett. 2000, 526.
(
6
7
. Siew, D., Hew, S., Soenen, V., and Volta, J. C., J. Catal. 123, 417 (1990).
. Chaar, M. A., Patel, D., Kung, M. C., and Kung, H. H., J. Catal. 105,
4
83 (1987).
. Corma, A., Lopez Nieto, J. M., and Paredes, N., J. Catal. 144, 425
1993).
. Soenen,V., Herrman, J. M., and Volta, J. C., J. Catal. 159, 410 (1996).
8
(
9
1
1
0. Kung, H. H., and Kung, M. C., Appl. Catal. A. 157, 105 (1997).
1. Lopez Nieto, J. M., Soler, J., Concepcion, P., Herguido, J., Menendez,
M., and Santamaria, J., J. Catal. 185, 324 (1999).
20. Knacke, O., Kubaschenski, O., and Hesselmann, K., “Thermochemi-
cal Properties of Inorganic Substances.” Springer-Verlag, Berlin/New
York, 1991.
1
1
1
2. Sharma, V., Wokaun, A., and Baiker, A., J. Phys. Chem. 90, 2715
(
1986).
21. Ishida, S., Imamura, S., and Fujimura, Y., React. Kinet. Catal. Lett. 43,
453 (1991).
22. Blasco, T., and Lopez Nieto, J. M., Appl. Catal. A 157, 117
(1997).
3. Mimura, N., Takahara, I., Saito, M., Hattori, T., Ohkuma, K., and
Ando, M., Catal. Today 45, 61 (1998).
4. Sugino, M., Shimada, H., Tsuruda, T., Miura, H., Ikenaga, N., and
Suzuki, T., Appl., Catal. A 121, 125 (1995).
23. Suzuki, T., et al., manuscript in preparation.