VOx/SiO2 Catalyst Prepared by Grafting VOCl3 on Silica for Oxidative Dehydrogenation of Propane

Article abstract of DOI:10.1002/cctc.201500607

The VO_x/SiO₂ catalysts for oxidative dehydrogenation of propane were synthesized by a simple grafting method. The VOCl₃ was first grafted at the surface of SiO₂, which was dehydrated at different temperature (from 200 to 1000°C). The formed grafted complexes were then calcined in air, leading to the formation of VO_x/SiO₂ catalysts. The synthesized catalysts were characterized by nitrogen adsorption, SEM, Raman spectroscopy, temperature-programmed reduction, and extended X-ray absorption fine structure analysis. The SiO₂ pretreatment temperature has an evident effect on the loading and dispersion of VO_x on SiO₂, which finally affects their catalytic performance. High SiO₂ treatment temperature is beneficial to dispersing the vanadium oxide species at the SiO₂ surface. These materials are efficient catalysts for the catalytic oxidative dehydrogenation of propane to propylene. The best selectivity to propylene is achieved on the VO_x/SiO_2-(1000) catalyst. The high selectivity and activity are well maintained for three days catalytic reaction. Silica likes it hot: The SiO₂ pretreatment temperature of grafted VO_x/SiO₂ catalysts has an evident effect on the loading and dispersion of VO_x on SiO₂, which finally affects the catalytic performance. The high selectivity and activity in propane oxidative dehydrogenation obtained on the VO_x/SiO_2-(1000) catalyst persists over three days reaction time.

Full text of DOI:10.1002/cctc.201500607

DOI: 10.1002/cctc.201500607
Full Papers
VOx/SiO2 Catalyst Prepared by Grafting VOCl3 on Silica for
Oxidative Dehydrogenation of Propane
Haibo Zhu,[a] Samy Ould-Chikh,[a] Hailin Dong,[a] Isabelle Llorens,[b] Youssef Saih,[a]
Dalaver H. Anjum,[c] Jean-Louis Hazemann,[d] and Jean-Marie Basset*[a]
The VOx/SiO2 catalysts for oxidative dehydrogenation of pro-
pane were synthesized by a simple grafting method. The
VOCl3 was first grafted at the surface of SiO2, which was dehy-
drated at different temperature (from 200 to 10008C). The
formed grafted complexes were then calcined in air, leading to
the formation of VOx/SiO2 catalysts. The synthesized catalysts
were characterized by nitrogen adsorption, SEM, Raman spec-
troscopy, temperature-programmed reduction, and extended
X-ray absorption fine structure analysis. The SiO2 pretreatment
temperature has an evident effect on the loading and disper-
sion of VOx on SiO2, which finally affects their catalytic per-
formance. High SiO2 treatment temperature is beneficial to dis-
persing the vanadium oxide species at the SiO2 surface. These
materials are efficient catalysts for the catalytic oxidative dehy-
drogenation of propane to propylene. The best selectivity to
propylene is achieved on the VOx/SiO2-(1000) catalyst. The high
selectivity and activity are well maintained for three days cata-
lytic reaction.
Introduction
Propylene is a major industrial chemical intermediate. It is used
as the basic building blocks for an array of chemicals and plas-
tics production, such as polypropylene, acrylic acid, and ace-
tone. The global propylene market increases each year owing
to a large consumption of the downstream products derived
from propylene. Therefore, much research efforts have been
dedicated to the development of new technologies or process-
es for the efficient production of propylene all over the
world.[1]
process for the existing propylene production.[1b] A great varie-
ty of catalysts with various compositions have been used in
the production of propylene from propane by ODH.[3] Among
these catalysts, the vanadia dispersed on different “classical”
supports, such as SiO2, Al2O3, TiO2, and ZrO2, has been found to
be the most active catalyst for the propane ODH reaction.[4]
Indeed, this kind of supported vanadia catalyst is proven to be
a structure-sensitive catalyst, and its catalytic performance is
closely related to the dispersion of vanadium oxide at the sur-
face of support.[4a]
Nowadays propylene is produced from the feedstock of
crude oil and natural gas by steam cracking, fluid catalytic-
cracking, and catalytic dehydrogenation processes.[1a,2] The de-
velopment of a less energy-intensive process for the propylene
production would be advantageous compared with these tra-
ditional processes. The thermodynamically favored oxidative
dehydrogenation (ODH) of propane is an attractive alternative
It has been demonstrated that the preparation method has
a significant influence on the nature of VOx species at the sur-
face of the support. To develop effective vanadia catalyst for
propane ODH, a variety of synthetic methods, such as impreg-
nation,[5] grafting,[6] thermolytic molecular precursor,[7] atomic
layer deposition,[8] and flame spray pyrolysis,[9] have been ap-
plied. Herein, we use a grafting method to anchor VOCl3 at the
surface of SiO2. The density of grafting sites (OH groups) at the
silica surface is strongly dependent on its dehydroxylation tem-
perature, and thus the final VOx loading and dispersion at SiO2
surface can be controlled systematically by the variation of
SiO2 pretreatment temperature. The produced VOx/SiO2 cata-
lysts exhibit different level of activity and selectivity in propane
ODH reaction.
[a] Dr. H. Zhu, Dr. S. Ould-Chikh, Dr. H. Dong, Dr. Y. Saih, Prof. J.-M. Basset
KAUST Catalysis Center
King Abdullah University of Science and Technology
Thuwal 23955-6900 (Kingdom of Saudi Arabia)
E-mail: Jeanmarie.basset@kaust.edu.sa
[b] Dr. I. Llorens
Institut de Recherches sur la Catalyse et l’Environnement de Lyon
IRCELYON, UMR 5256
CNRS—UniversitØ Lyon 1
2 av A. Einstein, 69626 Villeurbanne Cedex (France)
[c] Dr. D. H. Anjum
Results and Discussion
KAUST Core Lab
King Abdullah University of Science and Technology
Thuwal 23955-6900 (Kingdom of Saudi Arabia)
Characterization of catalysts
[d] Dr. J.-L. Hazemann
It is well known that the density and structure of OH group at
the surface of SiO2 is sensitive to its treatment temperature.[10]
If the SiO2 is calcined at different temperatures, dehydration,
dehydroxylation, and rehydroxylation processes will occur at
Institut Neel, CNRS
25, avenue des Martyrs, F-38042 Grenoble Cedex 9 (France)
Supporting Information for this article is available on the WWW under
http://dx.doi.org/10.1002/cctc.201500607.
ChemCatChem 2015, 7, 3332 – 3339
3332
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
the surface of SiO2, which leads to the reconstruction of the
surface and, in particular, the nature of silanol groups and
ꢀSiÀOÀSiꢀ moieties. Our synthesis strategy relies on the
grafting of the VOCl3 molecular precursor onto various silica
materials having different level of silanol density. To tune its
surface chemistry, silica was treated at different temperatures
(from 200 to 10008C) under high vacuum system (10À5 mbar)
prior to the VOCl3 grafting stage. Two kinds of typical absorp-
tion bands arising from n(OH) stretching vibrations can be ob-
served from the IR spectra of the SiO2 treated at different tem-
peratures (Figure 1a). The sharp classical band at around
3740 cmÀ1 is ascribed to isolated silanols, whereas the broad
band at 3500–3750 cmÀ1 is attributed to silanols interacting
between each other.[10] The intensity of these two bands sys-
tematically decreases with increasing pretreatment tempera-
ture. Furthermore, when the SiO2 treatment temperature was
increased to 7008C, the broad absorption band corresponding
to the silanols interaction totally disappears, and only the
sharp band from isolated silanols remains. Thus, the total
amount of hydroxyl groups at the surface SiO2 reduces gradu-
ally with the SiO2 treatment temperature, and only one single
type of silanol group is kept at the SiO2 surface if it was cal-
cined above 7008C. Notably, the calcination temperature has
a limited effect on the specific surface area of SiO2 as shown in
Table 1 (from 191 m2 gÀ1 at 2008C down to 178 m2 gÀ1 at
10008C).
The SiO2 treated at different temperatures subsequently
reacts with the VOCl3 in the H2O-free system, which results in
producing the VOClx/SiO2 intermediate. The reaction between
SiO2 and VOCl3 leads to a rapid and remarkable decrease of
the sharp band at 3740 cmÀ1 of the isolated silanol group in
the IR spectra (Figure 1b). In most cases, this silanol group is
completely consumed during the grafting procedure. On the
contrary, the broad band attributed to the interacting silanols
was kept nearly unchanged. Therefore, the isolated OH at the
surface of SiO2 can be considered as the active site for bonding
the VOCl3 species.
The VOClx/SiO2 intermediate was then submitted to the cal-
cination process at 6008C in air, leading to the formation of
VOx/SiO2 catalysts. Element analysis on the obtained VOx/SiO2
shows that the V content depends closely on the SiO2 pretreat-
ment temperature (Table 1). The VOCl3 grafted onto the SiO2
treated at higher temperature has lower V content. As revealed
above, the amount of surface isolated OH groups, which are
presumably the active center for grafting the VOCl3 species,
decreases with the treatment temperature. As a consequence,
less VOx is grafted at the surface of SiO2 that was treated at
higher temperature. Therefore, the V loading on the SiO2 sur-
face could be varied by changing the SiO2 pretreatment tem-
perature.
Figure 1. IR spectra of a) SiO2 dehydroxylated at different temperatures
under 10À5 mbar, b) VOCl3 grafted onto various thermally treated SiO2.
formation of new SiÀOÀV linkage at the surface of SiO2.[12]
Therefore, the high loading of V on SiO2 likely results in a de-
crease of the surface area by modification of the surface of
silica.
The BET surface area of the synthesized VOx/SiO2 sample de-
creases with the vanadium oxide loading (Table 1), which is
consistent with the results of VOx supported on MCM-41,
MCM-48, and SBA-15 materials obtained by impregnation
method.[5a,11] The vanadium oxide species resulting from the air
treatment, reacts with the OH group at the SiO2 surface, thus
a contraction of the SiO2 structure may occur because of the
The SEM images in Figure 2 show the morphology of the
SiO2 before and after grafting vanadium oxide at its surface.
The SiO2 is of spherical shape with a relatively uniform size of
approximately 300 nm. Its morphology is well kept after vana-
dia loading at its surface. Except SiO2, no other particle is ob-
served on the prepared VOx/SiO2 sample, indicating that the
VOx exists as a high dispersed species at the surface of SiO2.
ChemCatChem 2015, 7, 3332 – 3339
www.chemcatchem.org
3333
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
SiO2-(700) sample. Only amorphous SiO2 is observed in the high-
resolution (HR) TEM image in Figure 3c, d, and there is no evi-
dent V2O5 crystal at the surface of SiO2. High-angle annular
dark-field (HAADF) STEM imaging technique in Figure 3e was
further used to characterize this sample. This advanced tech-
nique is quite sensitive to the surface structure, and even small
surface clusters can be detected.[13] However, in Figure 3e no
clear clusters were observable at the surface of VOx/SiO2-(700)
sample. Therefore, from TEM and HAADF STEM characteriza-
tions, it can be concluded that the VOx is highly dispersed at
the SiO2 surface without formation of crystalline structure.
The Raman spectroscopy is recognized to be a useful tech-
nique to study the dispersion and oligomeric state of surface
VOx species.[14] It is well known that the structure of surface
VOx is quite sensitive to dehydroxylation conditions. The pres-
ence of H2O can result in different levels of hydration process
of the surface VOx species, leading to the formation of
a V2O5·nH2O-like gel.[5a,15] An in situ Raman spectroscopic mea-
surement was thus performed in He gas flow at 5008C. This
testing condition is relatively close to the catalytic condition
for propane ODH, and therefore the structure information of
VOx inferred from Raman spectroscopy will reflect to some
extent its real structure in the catalytic process. As shown in
Figure 4, a sharp peak located at 1040 cmÀ1 is observed in all
the samples, and this peak is generally attributed to the sym-
metric stretching mode of the V=O bond of the isolated tetra-
hedral VO4 species at the surface of SiO2. The shoulder peak at
around 1060 cmÀ1 is ascribed to ꢀSiÀOÀV bond.[11c,16] A broad
band from 400 to 480 cmÀ1 is attributed to the SiO2 feature
absorption.[11c,16] Most importantly, no detectable band around
970 cmÀ1 is observed in all the samples. This absorption band
is usually ascribed to the formation of crystalline V2O5.[5,11b,c]
We can thus infer that the vanadium is highly dispersed at the
surface of SiO2 without formation of bulk V2O5 phase.
Table 1. Elemental analysis and surface area of the SiO2 support vanadi-
um oxide samples.
Entry
V loading
[%]
Surface area
Surface area
[a]
[b]
[m2 gÀ1
]
[m2 gÀ1
]
SiO2-(200)
SiO2-(300)
SiO2-(500)
SiO2-(700)
SiO2-(800)
SiO2-(1000)
3.8
3.5
3.0
1.9
1.5
1.0
191
189
187
183
179
178
128
135
145
148
152
155
[a] Surface area measured before grafting of VOCl3. [b] Surface area of
VOx/SiO2 catalysts.
Figure 2. SEM images of SiO2 and a typical VOx/SiO2-(700) catalyst. Scale
bars=2 mm.
Figure 3. TEM and HAADF STEM of the VOx/SiO2-(700) sample: a) Low-magnifi-
cation TEM image and selected-area electron diffraction pattern; b) energy-
dispersive spectrometry (EDS) spectrum from the area in (a); c,d) high-mag-
nification TEM images; and e) HAADF STEM image.
Further characterizations including TEM and STEM tech-
niques (Figure 3) were also used to study the VOx/SiO2-(700)
sample to provide more detailed information about its micro-
structure. The big particles observed in the SEM turned out to
be the agglomerates of nanosized units having a size below
50 nm. The EDS spectrum from the area shown in Figure 3a re-
veals the existence of V, Si, and O elements in the sample,
along with the Ni and C elements from the TEM grid. The se-
lected-area electron diffraction (SAED) pattern in Figure 3a
(inset) indicates that there is no crystalline structure in VOx/
Figure 4. Raman spectra of VOx/SiO2 samples synthesized from grafting of
VOClx onto the SiO2 treated at different temperature.
ChemCatChem 2015, 7, 3332 – 3339
www.chemcatchem.org
3334
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
a vanadyl moiety and VÀO single bonds. The coordination
number of the short V=O double bond is in the range of 1.0–
1.2Æ0.1 at a distance of 1.58–1.62Æ0.01 Š, and the coordina-
tion number of VÀO single bond is in the range of 2.9–3.1Æ
0.1 at a distance of 1.84–1.89Æ0.02 Š. Furthermore, the X-ray
absorption near-edge spectroscopy (XANES) spectrum for each
Figure 5. TPR profiles of VOx/SiO2 samples synthesized from grafting of
VOClx onto the SiO2 treated at different temperature.
The temperature-programmed reduction (TPR) profiles of
VOx/SiO2 catalysts analyzed from room temperature to 7008C
are shown in Figure 5. Only one sharp peak located from 530
to 5708C is detected in all the samples. This peak is attributed
to the reduction of monomeric or low-oligomeric VOx surface
species.[5a,11a,17] The intensities of these peaks decrease system-
atically with the SiO2 pretreatment temperature, which is at-
tributed to a lower V loading in the SiO2 treated at higher tem-
perature. Moreover, an increase of vanadium loading leads to
a gradual shift of the reduction peak to high temperature. This
observation is in agreement with TPR investigation on meso-
porous-silica-supported VOx catalysts.[5b,11b] The progressive
shift of the reduction peak to high temperature with the V
loading is relevant to a gradual formation of a highly polymeric
structure, which is less reducible in the TPR test. It should be
noted that no reduction peak at approximately 5908C associat-
ed with the reduction of polymeric V5+ species in V2O5-like
structure is found in all the sam-
Figure 6. EXAFS k3 c(k) functions (top) and Fourier transforms (bottom) for
selected VOx/SiO2 samples with different V loadings. The solid lines are the
experimental data and the dash lines are the fit results with a k range of
À1
[2.3, 11.7] Š and R range of [0.5, 3.3] Š.
ples.[5,11a,b] Therefore, the TPR
Table 2. Parameters extracted from EXAFS fit of VOx/SiO2 samples. Italicized characters denote a fixed parame-
ter.
characterization is consistent
with Raman data, further con-
firming the high-level dispersion
of VOx at the SiO2 surface.
Sample
Shell
N[a]
d [Š]
s2 [Š2]
DE0 [eV]
R-factor
VOx/SiO2À(200)
VÀO1
VÀO2
VÀV[b]
VÀO1
VÀO2
VÀV[b]
VÀO1
VÀO2
VÀV[b]
VÀO1
VÀO2
VÀV[b]
1.1Æ0.1
2.9Æ0.1
0.9Æ0.2
0.9Æ0.1
3.1Æ0.1
0.6Æ0.2
1.0Æ0.1
3.0Æ0.1
0.7Æ0.1
1.2Æ0.1
2.8Æ0.1
0.5Æ0.2
1.58Æ0.01
1.85Æ0.02
3.02Æ0.04
1.61Æ0.01
1.88Æ0.02
3.02Æ0.04
1.62Æ0.01
1.89Æ0.02
3.03Æ0.04
1.59Æ0.01
1.84Æ0.02
3.00Æ0.06
0.0009
Vanadium K-edge k3-weighted
EXAFS signals for selected VOx/
SiO2 samples with different V
loadings and the corresponding
Fourier transforms are shown in
Figure 6 and extracted parame-
ters given in Table 2. For all the
samples, the first neighbor con-
tribution is fitted with a total of
four oxygen atoms separated in
two scattering paths modeling
0.010Æ0.002
À3.9Æ1.5
0.065
0.002
VOx/SiO2À(500)
VOx/SiO2À(700)
VOx/SiO2À(1000)
0.0009
0.012Æ0.002
0.002
0.0009
0.011Æ0.001
0.002
0.0009
0.0085Æ0.0016
0.002
À0.7Æ1.4
À0.6Æ1.3
À5.3Æ2.0
0.066
0.057
0.071
[a] N refers to coordination number. [b] VÀV scattering path is caused by a VÀOÀV arrangement.
ChemCatChem 2015, 7, 3332 – 3339
www.chemcatchem.org
3335
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
compound depicted a very intense pre-edge located at
5470.5 eV (Figure S1). Intense transitions in the pre-edge
region of K absorption edge for transition metals are known to
occur if the local symmetry around the metal center does not
include an inversion center.[18] In this case, the hybridization of
vanadium 3d with 4p states allows a large enhancement of the
dipolar transitions in the pre-edge region. Thus, considering
the presence of a vanadyl group and a total of four oxygen
atoms in the first coordination sphere, the local symmetry
around vanadium center would be consistent with a C3V punc-
tual group. It is also noteworthy to mention that the use of
a VÀCl scattering path with a distance of 2.1 Š (Figure S2 and
Table S1) did not improve any of the statistical agreement fac-
tors (R-factor and cv2). Therefore, the chloride in VOClx/SiO2 is
totally removed after calcination in air. However, the fit was im-
proved by including an additional VÀV scattering path at 3.00–
3.03Æ0.04 Š. The latter VÀV scattering path is ascribed to the
condensation of two [VO4] units to create small oligomers at
the surface of silica. Correlations between mean-square relative
displacement and amplitude parameters prevented an accu-
rate determination of the VÀV coordination number. To com-
pare materials at least relatively, the mean-square relative dis-
decrease of the proportion of dimer from approximately 90%
down to 50%. This is an important result as it proves that the
tuning of the initial surface chemistry of silica has a strong
impact on the final structure of grafted vanadium oxide.
Catalytic reaction
The catalytic performance of the SiO2-supported vanadium
oxide catalyst was evaluated in the ODH reaction of propane
in the temperature range 500 to 6008C at atmospheric pres-
sure. Propylene, carbon oxides including CO and CO2, meth-
ane, and ethylene are the gaseous products in the reaction.
The liquid products are acrolein and acetaldehyde. The blank
test on pure SiO2 did not show propane conversion, whereas
the VOx/SiO2 catalysts exhibits high performance in the pro-
pane ODH reaction.
The six VOx/SiO2 catalysts show different level of activity for
the propane ODH reaction (Figure 8a). High activities are ob-
served in these catalysts at 6008C, and the corresponding pro-
pane conversions vary from 25% to 55%. Among these cata-
lysts, the VOx supported on SiO2-(200) exhibits the highest pro-
pane conversion. With the increase of the silica treatment tem-
perature, the catalyst activity slowly decreases. The lowest pro-
pane conversion is observed with VOx/SiO2-(1000) catalyst. As
described above, the VOx loading is related to the SiO2 pre-
2
placement was fixed to 0.2”10À3 Š for all VOx/SiO2 samples.[19]
Doing so, the coordination number of VÀV shell decreases
from 0.9Æ0.2 to 0.5Æ0.2 as the pretreatment temperature of
silica is increased from 2008C to 10008C. The result is in ac-
cordance with Raman spectroscopy, which suggests that there
is a certain amount of low-polymeric VOx species at the surface
of SiO2 and the degree of oligomerization increases with V
loading.
Based on the structural characterization results above, a con-
figuration of the supported vanadium oxides is suggested and
schematically depicted in Figure 7. As revealed from TEM,
Raman spectroscopy, and H2 TPR tests, no VxOy nanocrystallite
is present in any of the samples. Further EXAFS analysis indi-
cates that the vanadium oxide exists as low-polymeric VOx spe-
cies. The VÀV coordination numbers for [VO4] monomer and
[V2O7] dimer are 0 and 1, respectively. Considering the experi-
mental value obtained for the materials in the present study
(between 0 and 1), we can describe the structure of the graft-
ed vanadium oxide as a mixture of monomer and dimer. By
using this simplified assumption, a general structure of vanadi-
um oxide at the SiO2 surface is shown in Figure 7. The configu-
rations of vanadium oxides supported at SiO2, which is pre-
treated at three typical conditions: low (2008C), moderate
(7008C), and high (10008C) temperatures, show a systematic
Figure 7. Schematic representation of molecular structure of vanadium
oxides supported at SiO2 pretreated at different temperature. The vanadium
oxide groups at the SiO2 surface are proposed to exist as [VO4] monomer
and [V2O7] dimer, and the proportion of dimers decreases with increasing
SiO2 calcination temperature.
Figure 8. a) Propane conversion and b) propylene selectivity on the different
VOx/SiO2 catalysts in propane ODH reaction.
ChemCatChem 2015, 7, 3332 – 3339
www.chemcatchem.org
3336
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
treatment temperature, and the SiO2 calcined at lower temper-
ature results in higher content of VOx at its surface. Therefore,
more active species is located at the surface of SiO2 treated at
low temperature. Accordingly, high propane conversion is ach-
ieved on the VOx supported on the SiO2 treated at low temper-
ature.
In contrast to propane conversion, the propylene selectivity
is reversely correlated with the V loading in SiO2 (Figure 8b).
Overall, the propylene selectivity increases systematically with
the decrease of V loading, and the VOx/SiO2-(1000) catalyst shows
the highest propylene selectivity. In addition, it is evident that
the propylene selectivity decreases dramatically with the reac-
tion temperature, which is a result of the deep oxidation of
propane at high temperature. Appreciable propylene selectivi-
ty is achieved in the VOx/SiO2-(1000) catalyst. In this catalyst, high
propylene selectivity of approximately 90% at 5008C is ob-
tained, and a selectivity of 68% still remains at 6008C.
Figure 10. Effect of O2/C3H8 ratio on the propylene selectivity in propane
ODH reaction over VOx/SiO2-(1000) catalyst at 5508C.
The propylene selectivity is related not only to the reaction
temperature but also to the propane conversion. To obtain
a better understanding of the reaction network, the variation
of the propylene selectivity with the propane conversion at
5508C was analyzed by changing the contact time
(0.025 gsmLÀ1 <W/F<0.15 gsmLÀ1) in the reaction (Figure 9).
Clearly, the selectivity to propylene decreases significantly with
the propane conversion in all the vanadium oxide catalysts,
which suggests the overoxidation of propylene into COx.[5,11b,20]
The selectivity to propylene is close to 100% if the propane
conversion is extrapolated to zero. Therefore, all the byprod-
ucts are mainly produced directly from the overoxidation of
propylene, indicating a consecutive reaction network in pro-
pane ODH reaction. On the other hand, the propylene selectiv-
tion in the feed was kept constant (10%), and the O2 concen-
tration in the feed was varied from 5% to 15% to obtain differ-
ent O2/C3H8 ratios. The O2/C3H8 ratio has a limited effect on the
propylene selectivity in the propane ODH reaction (Figure 10).
With the same level of propane conversion, the propylene se-
lectivity does not change much with the O2/C3H8 ratios. There-
fore, the propylene formation rate is independent on the O2
partial pressure, and the variation of O2/C3H8 ratio does not
change the propylene selectivity. As confirmed in a previous
study, the reaction order in oxygen is zero in the propene
combustion reaction,[21] and the change of O2/C3H8 ratio thus
has no substantial effect on the propylene selectivity.
As shown above, the VOx/SiO2-(1000) catalyst exhibits the high-
est selectivity to propylene, which could be a reasonably good
catalyst for the production of propylene from propane. The
stability of the catalyst with time is an important parameter for
its practical application. The VOx/SiO2-(1000) catalyst was further
investigated at 5808C with time-on-stream of 72 h. As shown
in Figure 11, the initial propane conversion (ꢁ31%) is quite
stable in the long-term test, and does not change with time
on stream. The calculated TOFpropane value at this reaction con-
dition is 2.59 minÀ1, which is close to the results in the previ-
ous reports.[4–9] More importantly, the selectivities to all the
products including propylene, ethylene, CO, CO2, acetaldehyde,
and acrolein are substantially maintained in 72 h catalytic reac-
tion. The propylene selectivity of 60% is well kept, which cor-
responds to the propylene yield of 18%. No activity loss in
VOx/SiO2-(1000) catalyst indicates that the intrinsic active site
(highly dispersed vanadium oxide) in the catalyst is very stable
at such high-temperature reaction, and it is resistant to this
harsh reaction condition. Therefore, the VOx/SiO2-(1000) material
could be a fairly good candidate for propane ODH in the prac-
tical industrial application thanks to its high stability in the re-
action.
ity is negatively relevant to the
V loading, and the
VOx/SiO2-(1000) catalyst shows the highest selectivity. Low density
of VOx at the surface SiO2-(1000) results in highly dispersed vana-
dia species at SiO2 surface. Low polymeric VOx species is there-
fore the predominant structure at the surface SiO2-(1000). The
highly dispersed VOx species, such as isolated VOx is believed
to be more selective for the formation of propylene.
The influence of O2/C3H8 ratio on the selectivity was studied
at 5508C on the VOx/SiO2-(1000) catalyst. The propane concentra-
The chemistry of the silanol group at the SiO2 surface
strongly relies on its treatment temperature under the high-
vacuum technique. Isolated, geminal, and vicinal silanols are
the typical OH group at the SiO2 surface. SiO2 treated at low
temperatures from 200 to 5008C, has high density of OH
group with various kinds of silanol group. By contrast, if the
Figure 9. Propane conversion versus propylene selectivity on different VOx/
SiO2 in propane ODH reaction.
ChemCatChem 2015, 7, 3332 – 3339
www.chemcatchem.org
3337
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
stable in 72 h time-on-stream reaction. Thus, these VOx/SiO2
materials prepared by the grafting method could be the po-
tential catalyst for the production of propylene from propane
oxidative dehydrogenation process.
Experimental Section
Catalyst preparation
The silica used in the study was nonporous Degussa Aerosil-200,
with a BET surface area ꢁ200 m2 gÀ1. The commercial available
VOCl3 (99%, Sigma Aldrich) was distilled under vacuum at RT
before grafting. The silica was dehydroxylated by calcination under
vacuum (10À5 mbar) for 8 h at 200, 300, 500, 700, 800, and 10008C,
respectively. The dehydroxylated SiO2 was then transferred into
a hexane solution of VOCl3 in the Schlenk tube. The resulting mix-
ture was kept on stirring for 4 h at RT. The yellow powder was sep-
arated from solution by filtration in the glove box, and afterwards
the isolated powder washed with hexane several time to remove
any physical adsorbed VOCl3. Finally, the obtained powder was cal-
cined in the air at 6008C for 4 h to obtain VOx/SiO2 catalyst. The
synthesized catalysts are denoted as VOx/SiO2-(T), where T is the
silica dehydrogenation temperature.
Characterization
N2 adsorption/desorption isotherms were acquired from a Micro-
meritics ASAP2420, and the samples were degassed for 2 h at
5008C before the measurement. The surface areas of the samples
were calculated by multipoint BET analysis method. The SEM of the
catalyst was taken from an FEI Quanta 600 FEG environmental
scanning electron microscope (ESEM). The TEM and HAADF STEM
analyses were conducted on aberration-corrected FEI Titan instru-
ment. Samples were dispersed in ethanol with the help of ultra-
sound treatment, and then supported onto a 200 mesh carbon-
coated nickel grid before TEM measurement. In situ Raman spectra
were recorded on a Horiba Yvon LabRAM Aramis with a CCD-
camera as a detector using a 100”objective, a 600 grmmÀ1 gra-
ting, a 100 mm hole, a 100 mm slit, a D1 filter, and a 473 nm cobalt
laser. The sample was pressed into thin pellet and then loaded in
an in situ chamber. All spectra were recorded at 5008C in He gas
flow. TPR measurements were performed on an Altamira Instru-
ment equipped with a TCD detector. The catalyst powder (10 mg)
was placed in a U-shaped quartz reactor and pretreated in flowing
Ar (30 mLminÀ1) for 0.5 h at 3508C, followed by cooling to RT. The
temperature was then raised from RT to 7008C at a rate of
108CminÀ1 in a 5% H2/Ar flow (30 mLminÀ1).
Figure 11. Propane conversion (top) and product selectivities (bottom) with
time on stream over VOx/SiO2-(1000) catalyst for propane ODH reaction at
5808C.
SiO2 was treated above 7008C, only isolated OH groups exist
at the surface of SiO2. The VOCl3 reacting with these isolated
OH groups is beneficial for getting highly dispersed vanadium
oxide species at silica surface, because the low density of OH
group can prevent the formation of highly polymeric vanadia
species. In such a case, highly dispersed vanadium oxide forms
at the surface of SiO2 that was treated at high temperature.
The catalyst with low V loading, such as VOx/SiO2-(1000) has
shown high performance in propane oxidative dehydrogena-
tion reaction, because the highly dispersed VOx species is ben-
eficial for the formation of propylene.
Conclusions
X-ray absorption spectroscopy (XAS) experiments were performed
on the CRG-FAME beamline (BM30B), at the European Synchrotron
Radiation Facility in Grenoble. Spectra were recorded either in fluo-
rescence (VOx/SiO2) or in transmission (V(=O)(OiPr)3 reference)
modes at the V/K-edge (5.465 keV). XAS data were analyzed using
the HORAE package, a graphical interface to the AUTOBK and IFEF-
FIT code and Cherokee software (for the extraction of the EXAFS
signal).[22] XANES and EXAFS spectra were obtained after perform-
ing standard procedures for pre-edge subtraction, normalization,
polynomial removal and wave vector conversion. The extracted
EXAFS signal was fitted in k3 c(k) without any additional filtering.
A series of VOx/SiO2 catalysts have been prepared by grafting
VOCl3 at the surface of SiO2, which was dehydrated before at
different temperatures under vacuum. The SiO2 pretreatment
temperature affects the density and structure of silanol group
at the surface of SiO2. The nature of silanol group has a signifi-
cant impact on the dispersion of VOx at the surface of SiO2.
The SiO2 treated at high temperature results in low loading
and high dispersion of VOx at the SiO2 surface, which produces
efficient VOx/SiO2 catalysts for the propane oxidative dehydro-
genation reaction. The prepared VOx/SiO2-(1000) catalyst exhibits
the highest selectivity to propylene in the reaction and ach-
ieves 18% propylene yield in this reaction. Moreover, the activ-
ity and selectivity achieved on VOx/SiO2-(1000) catalyst are quite
À1
Fitting was conducted in k-space betweeÀn1 2.1 and 12.3 Š for
V(=O)(OiPr)3 and, between 2.0 and 12.5 Š for VOx/SiO2. During
the fitting procedure, a restriction on the number of fitted parame-
ters was taken into account: Dk and DR allowed at least 15 param-
ChemCatChem 2015, 7, 3332 – 3339
www.chemcatchem.org
3338
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
[6] A. Comite, A. Sorrentino, G. Capannelli, M. Di Serio, R. Tesser, E. Santace-
saria, J. Mol. Catal. A 2003, 198, 151–165.
[7] a) J. L. Male, H. G. Niessen, A. T. Bell, T. Don Tilley, J. Catal. 2000, 194,
431–444; b) C. Pak, A. T. Bell, T. D. Tilley, J. Catal. 2002, 206, 49–59.
[8] J. Keränen, P. Carniti, A. Gervasini, E. Iiskola, A. Auroux, L. Niinistç, Catal.
Today 2004, 91–92, 67–71.
eters and 12 parameters were used at most to describe different
paths. Determination of the energy level (E0) was performed at the
first inflection point of the edge. For each atomic shell, the follow-
ing structural parameters were adjusted: coordination number (N),
bond length (R), and the so-called Debye–Waller factor through
the mean-square relative displacement (s2) of the considered bond
[9] B. Schimmoeller, S. E. Pratsinis, A. Baiker, ChemCatChem 2011, 3, 1234–
1256.
2
length. The amplitude factor (S0 ) was fitted to the EXAFS spectrum
obtained for V(=O)(OiPr)3. To do so,
a molecular model of
[10] a) L. T. Zhuravlev, Colloids Surf. A 2000, 173, 1–38; b) X. Rozanska, F. Del-
becq, P. Sautet, Phys. Chem. Chem. Phys. 2010, 12, 14930–14940; c) L.
Lefort, M. Chabanas, O. Maury, D. Meunier, C. CopØret, J. Thivolle-Cazat,
J.-M. Basset, J. Organomet. Chem. 2000, 593–594, 96–100.
[11] a) H. Berndt, A. Martin, A. Brückner, E. Schreier, D. Müller, H. Kosslick,
G. U. Wolf, B. Lücke, J. Catal. 2000, 191, 384–400; b) B. Solsona, T.
Blasco, J. M. López Nieto, M. L. PeÇa, F. Rey, A. Vidal-Moya, J. Catal.
2001, 203, 443–452; c) C. Hess, G. Tzolova-Müller, R. Herbert, J. Phys.
Chem. C 2007, 111, 9471–9479.
[12] M. Morey, A. Davidson, H. Eckert, G. Stucky, Chem. Mater. 1996, 8, 486–
492.
[13] a) W. Zhou, E. I. Ross-Medgaarden, W. V. Knowles, M. S. Wong, I. E.
Wachs, C. J. Kiely, Nat. Chem. 2009, 1, 722–728; b) B. Qiao, A. Wang, X.
Yang, L. F. Allard, Z. Jiang, Y. Cui, J. Liu, J. Li, T. Zhang, Nat. Chem. 2011,
3, 634–641.
V(=O)(OiPr)3 was built and optimized at the UFF[23] level with the
Avogadro molecular builder and visualization tool.[24] Then the ge-
ometry of the whole structure was fully optimized at the DFT level
from Gaussian09[25] code using the MO6[26] functional and TZVP[27]
basis set for all the atoms except for the vanadium atom to which
we described the electronic structure of s and p core electrons
from the Stuttgart/Dresden[28] effective core potential. The resulting
atomic coordinates were used as an input to calculate the theoreti-
cal phases, amplitudes and electron free mean path with the
FEFF6 code[29] implemented in IFEFFIT programs suite. S0 for
V(ÀO)(OiPr)3 was then found equal to 0.87.
2
[14] Y. T. Chua, P. C. Stair, I. E. Wachs, J. Phys. Chem. B 2001, 105, 8600–8606.
[15] S. Xie, E. Iglesia, A. T. Bell, Langmuir 2000, 16, 7162–7167.
[16] X. Gao, S. R. Bare, B. M. Weckhuysen, I. E. Wachs, J. Phys. Chem. B 1998,
102, 10842–10852.
[17] F. Arena, N. Giordano, A. Parmaliana, J. Catal. 1997, 167, 66–76.
[18] a) F. de Groot, G. Vankó, P. Glatzel, J. Phys. Condens. Matter 2009, 21,
104207; b) D. Cabaret, A. Bordage, A. Juhin, M. Arfaoui, E. Gaudry, Phys.
Chem. Chem. Phys. 2010, 12, 5619–5633.
[19] A. L. Hector, W. Levason, A. J. Middleton, G. Reid, M. Webster, Eur. J.
Inorg. Chem. 2007, 3655–3662.
[20] a) B. Frank, M. Morassutto, R. Schomäcker, R. Schlçgl, D. S. Su, Chem-
CatChem 2010, 2, 644–648; b) I. Rossetti, L. Fabbrini, N. Ballarini, C.
Oliva, F. Cavani, A. Cericola, B. Bonelli, M. Piumetti, E. Garrone, H. Dyr-
beck, E. A. Blekkan, L. Forni, J. Catal. 2008, 256, 45–61.
[21] A. Dinse, S. Khennache, B. Frank, C. Hess, R. Herbert, S. Wrabetz, R.
Schlçgl, R. Schomäcker, J. Mol. Catal. A 2009, 307, 43–50.
[22] a) B. Ravel, M. Newville, J. Synchrotron Radiat. 2005, 12, 537–541; b) M.
Alain, M. Jacques, M.-B. Diane, P. Karine, J. Phys. Conf. Ser. 2009, 190,
012034–012035.
[23] A. K. Rappe, C. J. Casewit, K. S. Colwell, W. A. Goddard, W. M. Skiff, J. Am.
Chem. Soc. 1992, 114, 10024–10035.
[24] M. Hanwell, D. Curtis, D. Lonie, T. Vandermeersch, E. Zurek, G. Hutchi-
son, J. Cheminf. 2012, 4, 17.
Catalytic tests
Catalytic propane ODH was performed on a micropilot (from PID
Eng&Tech) equipped with a quartz glass reactor at atmospheric
pressure. The catalyst powder (25 mg) was placed in the reactor
supported by quartz glass wool. The effect of temperature on the
activity and selectivity was measured from 500 to 6008C, and the
gas feed composed of 10% C3H8 and 10% O2 in He was introduced
through the catalytic bed at a total flow rate of 30 mL/minute. The
propane conversion versus propylene selectivity was determined
at 5508C, and the total flow rate (10% C3H8 and 10% O2) was
changed from 10 mL/min to 60 mL/min. The stability test was eval-
uated at 5808C with total flow rate of 10 mL/min (10% C3H8 and
10% O2). The reaction products were analyzed with on-line Varian
490 micro-GC equipped with a TCD detector and four columns:
a MolSieve 5 Š column to quantify O2 and CO, a poraPLOT Q
column to quantify CO2, C2H4 as well as C2H6, an alumina column
to quantify C3H8 as well as C3H6. Propane conversion and selectivity
to propylene were calculated on a carbon basis. Carbon mass bal-
ance was superior to 98% in all cases.
[25] Gaussian 09, Revision D.01, M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E.
Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, B. Men-
nucci, G. A. Petersson, H. Nakatsuji, M. Caricato, X. Li, H. P. Hratchian,
A. F. Izmaylov, J. Bloino, G. Zheng, J. L. Sonnenberg, M. Hada, M. Ehara,
K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O.
Kitao, H. Nakai, T. Vreven, J. A. Montgomery, Jr., J. E. Peralta, F. Ogliaro,
M. Bearpark, J. J. Heyd, E. Brothers, K. N. Kudin, V. N. Staroverov, R. Ko-
bayashi, J. Normand, K. Raghavachari, A. Rendell, J. C. Burant, S. S. Iyen-
gar, J. Tomasi, M. Cossi, N. Rega, J. M. Millam, M. Klene, J. E. Knox, J. B.
Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann,
O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, R. L. Martin,
K. Morokuma, V. G. Zakrzewski, G. A. Voth, P. Salvador, J. J. Dannenberg,
S. Dapprich, A. D. Daniels, Ö. Farkas, J. B. Foresman, J. V. Ortiz, J. Cio-
slowski, D. J. Fox, Gaussian, Inc., Wallingford CT, 2009.
Keywords: alkanes · alkenes · dehydrogenation · silicates ·
vanadium
[1] a) M. M. Bhasin, J. H. McCain, B. V. Vora, T. Imai, P. R. Pujadó, Appl. Catal.
A 2001, 221, 397–419; b) F. Cavani, N. Ballarini, A. Cericola, Catal. Today
2007, 127, 113–131.
[2] a) A. Corma, F. Melo, L. Sauvanaud, F. J. Ortega, Appl. Catal. A 2004, 265,
195–206; b) Q. Shao, P. Wang, H. Tian, R. Yao, Y. Sun, J. Long, Catal.
Today 2009, 147, S347–S351.
[3] a) Z. Bai, P. Li, L. Liu, G. Xiong, ChemCatChem 2012, 4, 260–264; b) Y. V.
Kolen’ko, W. Zhang, R. N. d’Alnoncourt, F. Girgsdies, T. W. Hansen, T.
Wolfram, R. Schlçgl, A. Trunschke, ChemCatChem 2011, 3, 1597–1606;
c) X. Sun, R. Wang, B. Zhang, R. Huang, X. Huang, D. S. Su, T. Chen, C.
Miao, W. Yang, ChemCatChem 2014, 6, 2270–2275; d) S. R. G. Carrazµn,
C. Peres, J. P. Bernard, M. Ruwet, R. Ruiz, B. Delmon, J. Catal. 1996, 158,
452–476; e) K. Chalupka, C. Thomas, Y. Millot, F. Averseng, S. Dzwigaj, J.
Catal. 2013, 305, 46–55.
[26] Y. Zhao, D. Truhlar, Theor. Chem. Acc. 2008, 120, 215–241.
[27] A. Schäfer, C. Huber, R. Ahlrichs, J. Chem. Phys. 1994, 100, 5829–5835.
[28] D. Andrae, U. Häußermann, M. Dolg, H. Stoll, H. Preuß, Theor. Chim.
Acta 1990, 77, 123–141.
[29] S. I. Zabinsky, J. J. Rehr, A. Ankudinov, R. C. Albers, M. J. Eller, Phys. Rev. B
1995, 52, 2995–300.
[4] a) C. A. Carrero, R. Schloegl, I. E. Wachs, R. Schomaecker, ACS Catal.
2014, 4, 3357–3380; b) R. Grabowski, Catal. Rev. Sci. Eng. 2006, 48,
199–268; c) H. H. Kung, Adv. Catal. 1994, 40, 1–38.
[5] a) Y.-M. Liu, Y. Cao, N. Yi, W.-L. Feng, W.-L. Dai, S.-R. Yan, H.-Y. He, K.-N.
Fan, J. Catal. 2004, 224, 417–428; b) Y.-M. Liu, W.-L. Feng, T.-C. Li, H.-Y.
He, W.-L. Dai, W. Huang, Y. Cao, K.-N. Fan, J. Catal. 2006, 239, 125–136.
Received: May 30, 2015
Published online on September 7, 2015
ChemCatChem 2015, 7, 3332 – 3339
www.chemcatchem.org
3339
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim