Aerosol route to nanostructured WO3-SiO2-Al 2O3 metathesis catalysts: Toward higher propene yield

Article abstract of DOI:10.1016/j.apcata.2013.06.041

Aerosol processing is presented as a powerful new method to yield in one step highly efficient olefin metathesis catalysts. Combined with sol-gel chemistry, evaporation induced self-assembly and polyanions, it allows a fine tuning of the dispersion of WO_x species and the generation of acidic matrices which are key parameters controlling the formation of active metathesis sites. These spray dried catalysts are characterized by N₂- physisorption, XRD, TEM, NH₃-TPD, TPO, TPR, ICP-AES and systematically compared to relevant reference catalysts prepared by impregnation methods on different carriers (silica, alumina and silica-alumina). Tested in the industrially relevant cross-metathesis of ethene and 2-butene to propene, the new catalyst outcompetes reference catalysts, reaching 68 mmolpropene g ^-1 h^-1 at 250 °C. By-products formed by isomerization reactions remain very low and high propene yields (39%) are reached.

Full text of DOI:10.1016/j.apcata.2013.06.041

GModel
APCATA-14301; No. of Pages9
ARTICLE IN PRESS
Applied Catalysis A: General xxx (2013) xxx–xxx
Contents lists available at SciVerse ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
Aerosol route to nanostructured WO3-SiO2-Al2O3 metathesis
catalysts: Toward higher propene yield
Damien P. Debeckera,∗, Mariana Stoyanovab, Uwe Rodemerckb, Frédéric Colbeau-Justinc,
Cédric Boissèrec, Alexandra Chaumonnotd, Audrey Bonduelled, Clément Sanchezc,∗
a Institute of Condensed Matter and Nanoscience – MOlecules, Solids and reactiviTy (IMCN/MOST), Université catholique de Louvain, Croix du Sud 2 box
L7.05.17, 1348 Louvain-La-Neuve, Belgium
b Leibniz-Institut für Katalyse e.V. an der Universität Rostock, Albert-Einstein-Str. 29a, 18059 Rostock, Germany
c Laboratoire de la Matière Condensée de Paris (LCMCP), UMR-7574 Collège de France- UPMC-CNRS, Collège de France,11, Place Marcelin Berthelot, 75231
Paris Cedex 05, France
d IFPEN – Lyon, Rond-point de l’échangeur de Solaize, BP3, 69360 Solaize, France
a r t i c l e i n f o
a b s t r a c t
Article history:
Aerosol processing is presented as a powerful new method to yield in one step highly efficient olefin
metathesis catalysts. Combined with sol–gel chemistry, evaporation induced self-assembly and poly-
anions, it allows a fine tuning of the dispersion of WOx species and the generation of acidic matrices
which are key parameters controlling the formation of active metathesis sites. These spray dried cata-
lysts are characterized by N2-physisorption, XRD, TEM, NH3-TPD, TPO, TPR, ICP-AES and systematically
compared to relevant reference catalysts prepared by impregnation methods on different carriers (silica,
alumina and silica–alumina). Tested in the industrially relevant cross-metathesis of ethene and 2-butene
to propene, the new catalyst outcompetes reference catalysts, reaching 68 mmolpropene g−1 h−1 at 250 ◦C.
By-products formed by isomerization reactions remain very low and high propene yields (39%) are
reached.
Received 28 March 2013
Received in revised form 17 June 2013
Accepted 26 June 2013
Available online xxx
Keywords:
Aerosol
Sol–gel
EISA
Mesoporous mixed oxide
Surfactant
© 2013 Elsevier B.V. All rights reserved.
Alkene metathesis
Ethylene
Propylene
1. Introduction
light olefin conversion. It is generally accepted [3,14], that the
driving catalytic step in metathesis is the formation of the met-
The metathesis of light olefins is a widely studied reaction due
not only to its academic interest [1,2] but also to its massive indus-
trial importance [3]. The metathesis reaction technology indeed
affords the adjustment of petrochemical products in the refinery
as a function of market demand. Nowadays, the highly demanded
propene is produced on-site via the cross-metathesis of ethene and
butene in a process called “Olefin Conversion Technology”, licensed
by Lummus Technology.
Supported transition metal oxides, including Mo [4–6], Re
[7–9] and W [10,11] are the typical catalysts for metathesis reac-
tions. Recently, tungsten hydride catalysts prepared by surface
organometallic chemistry was reported as a promising alternative
[12,13]. Silica-supported tungsten oxide is however the catalyst
that is currently applied industrially for heterogeneous gas phase
allocyclobutane intermediate by interaction of an olefin with a
surface metallocarbene complex which was formed by the reaction
between an olefin and dispersed tungsten oxide species. There-
fore, the dispersion of the supported phase and the nature of the
interaction between the carrier and active species are the most
significant parameters that govern the metathesis rate. Usually,
WOx-catalysts are prepared by simple impregnation of ammonium
metatungstate on silica, alumina or mixed Si–Al followed by drying
and calcination. In such two-step preparation methods the chal-
lenge is two-fold.
Firstly, a well-balanced acidity is central for (i) the successful
anchoring of WOx during the catalyst preparation and (ii) the fast
reaction propagation during the catalytic cycle. Alumina-supported
tungsten oxide exhibits strong acid sites promoting high sur-
face dispersion but also inducing side-reactions like isomerization.
Silica-supported tungsten oxide is usually preferred since amor-
phous SiO2 alone is barely acidic. The disadvantage here is that
WOx species sinter easily because their interactions with the sup-
port are weak and this leads to the formation of species that are
∗
Corresponding authors.
E-mail addresses: Damien.debecker@uclouvain.be (D.P. Debecker),
Clement.sanchez@upmc.fr (C. Sanchez).
0926-860X/$ – see front matter © 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.apcata.2013.06.041
Please cite this article in press as: D.P. Debecker, et al., Appl. Catal. A: Gen. (2013), http://dx.doi.org/10.1016/j.apcata.2013.06.041

GModel
APCATA-14301; No. of Pages9
ARTICLE IN PRESS
2
D.P. Debecker et al. / Applied Catalysis A: General xxx (2013) xxx–xxx
too reducible and thus inactive. Thus several authors have studied
mixed Si–Al materials as supports. For example, Huang et al. stud-
ied physically mixed WO3/Al2O3/HY in metathesis of ethene and
2-butene. They stated that during calcination the bulk WO3-phase
becomes better dispersed first into microcrystallites and further,
via chemical transformation thanks to the mixed support struc-
ture, to monomeric surface tungstate species which are identified
as the active species in the metathesis reaction [15]. Another perti-
nent strategy [16] is to modify an alumina support by grafting silica
onto it in order to mitigate the acidity and thus the dispersion and
reducibility of impregnated WOx species.
Secondly, impregnation methods offer limited opportunities in
terms of textures. Whether the support is a commercial one or pre-
pared on purpose with tailored texture, it is difficult to ensure that
the active phase gets truly dispersed on the entire available surface
and down to the bottom of the pores [17]. Instead it is frequent to
observe WO3 aggregates at the outer surface of the support parti-
cles [18]. This being said, it is known that dispersion of the metal site
is critical to high activity, i.e. isolated metal sites have high activity
and good selectivity whereas polymeric MOx species (M = Mo, W or
Re) have either poor activity or poor selectivity [17,19].
Therefore, one-step preparation methods represent desirable
alternatives for the preparation of efficient metathesis catalysts
with adapted texture, composition and dispersion. In this perspec-
tive, methods like flame spray pyrolysis [20] or non-hydrolytic
sol–gel [21,22] constitute attractive options. Recently, an inno-
vative hydrolytic sol–gel method was proposed for the one-step
preparation of nanostructured silica–alumina materials [23] and
MoO3-SiO2-Al2O3 catalyst [24]. The preparation takes advantage
of the aerosol assisted sol–gel process [25]. In the latter, the evap-
oration induced self-assembly of a surfactant is used to generate
a calibrated porosity. Moreover, the kinetics of the sol–gel pro-
cess is quenched so as to obtain a homogeneous dispersion of all
components of the mixed oxide. The obtained aerosol processed
MoO3-SiO2-Al2O3 catalyst exhibited very high metathesis activ-
ity at low reaction temperature (40 ◦C). However, as classically
observed with highly active Mo-based catalysts [26,27], it showed
substantial deactivation with time on stream.
water. The recovered wet cake was dried at 110 ◦C for 12 h and
then calcined stepwise as follow: heating (1 K min−1) up to 250 ◦C,
hold 2 h, then heating (3 K min−1) up to 550 ◦C, hold 8 h in air flow.
At the end, the powder was pressed, crushed and sieved in the
250–600 ␮m fraction. These three reference catalysts are denoted
WOx-imp-Si, WOx-imp-Al and WOx-imp-SiAl.
2.1.2. Aerosol catalysts
The new catalyst was prepared by the aerosol process
(Scheme 1). A surfactant solution is prepared by dissolving 11.5 g
of Pluronic P123 in 77.8 g of water and 137 g of absolute ethanol.
The solution is stirred at room temperature until complete disso-
lution of the surfactant. A tungsten precursor solution is prepared
by dissolving 1.50 g of H3PW12O40 Keggin heteropolyanion in 15 g
of distilled water. An Al–Si precursors solution is obtained by basic
hydrolysis of Si and Al alkoxides as follows: (i) 3.6 g of aluminum
sec-butoxide is dissolved in 17.1 g of aqueous tetrapropylammo-
nium hydroxide (TPAOH) solution (40% in water), (ii) after 15 min
of homogenization 50 g of water and then 36.9 g of tetraethylor-
tosilicate (TEOS) are added, (iii) the solution is stirred at room
temperature for 16 h. All three solutions are then mixed together
and atomized with a Büchi B-290 spray dryer. The input param-
eters are fixed as follows: 220 ◦C for the chamber temperature,
9 cm3 min−1 for the solution flow and 35 m3 h−1 for the suction
flow of heated air. The recovered powder is then calcined under
static air at 130 ◦C (2 K min−1) for 2 h and then 550 ◦C (2 K min−1
for 12 h. This catalyst is denoted WOxSiAl-aer.
)
One silica–alumina powder was prepared by the same aerosol
method, but in the absence of the W Keggin compound. This sam-
ple is denoted SiAl-aer. It was then impregnated with an aqueous
solution of H3PW12O40. The solution was added dropwise while the
powder was mechanically stirred. The obtained material was then
allowed to mature, at ambient temperature at a relative humidity of
100% during 24 h. A drying step was then applied for 24 h at 130 ◦C.
This sample is denoted WOx-impSiAl-aer. Denotations of studied
catalysts together with some textural characteristics are listed in
Table 1.
In this paper, we report on the controlled design and processing
of a new WO3-SiO2-Al2O3 catalyst exhibiting optimized texture,
tungsten dispersion, and acidity and being able to work at higher
temperatures, close to those used in the industry. The catalyst is
prepared in one step via an adapted version of the aerosol process
already reported [24]. Its properties and performance are com-
pared to those of reference catalysts prepared by impregnation on
silica, alumina and silica–alumina supports. The metathesis activ-
ity of these W-catalysts is measured in the industrially relevant
cross-metathesis reactions of ethene and 2-butene.
2.2. Catalyst characterization
2.2.1. Elemental analysis
The WO3 weight loading of all catalysts was measured by induc-
tively coupled plasma-atomic emission spectroscopy (ICP-AES) on
Iris Advantage from Jarrell Ash Co. The samples were dried at 105 ◦C
prior to measurement.
2.2.2. Textural analysis
The specific surface area (SBET) and porosity of the materials
were calculated from nitrogen adsorption–desorption isotherm
collected at −196 ◦C on BELSORB-mini II (BEL Japan, Inc.). Prior
to measurements the samples were outgassed for 2 h at 250 ◦C.
SBET was calculated applying the Brunauer, Emmet and Teller (BET)
equation for N2 relative pressure in range of 0.05 < P/P0 < 0.30. Pore
size distribution was determined by Barrett–Joyner–Halenda (BJH)
method from the desorption branch of the isotherm. Mean pore
2. Experimental
2.1. Catalyst preparation
2.1.1. Reference catalysts
Three commercial supports were used as carrier to prepare
WOx-based catalysts by impregnation. Silica (SiO2, Davisil grade
646), alumina (␥-Al2O3, Evonik) and silica–alumina (Aldrich, grade
135, about 13 wt.% Al2O3). These supports are denoted Si, Al and
SiAl in the following. Prior to impregnation, the supports were cal-
cined at 500 ◦C for 12 h under air flow. Ammonium metatungstate
hydrate (STREM Chemicals, purity 99.9%) was used as a W precur-
sor. The nominal WOx loading was 10 wt.%. 13.8 g of each support
was immersed in precursor solution consisting of 1.594 g ammo-
nium metatungstate hydrate dissolved in 60 mL distilled water. The
resulting suspension was stirred at 80 ◦C until evaporation of the
size, listed in Table 1, was obtained using the formula 4·Vp/SBET
.
The theoretical surface tungsten oxide density (CW) was calculated
using the equation [28]:
ꢀ
ꢁ
ꢀ
ꢁ
LWO3 /100
231.8
N0
S × 1018
CW
=
×
, where
LWO is the loaded amount of WO3 (wt.%); 231.8 is the molecular
weight3of WO3; N0 is the Avogadro number (6.023 × 1023) and S is
the specific surface area (m2 g−1).
Please cite this article in press as: D.P. Debecker, et al., Appl. Catal. A: Gen. (2013), http://dx.doi.org/10.1016/j.apcata.2013.06.041

GModel
APCATA-14301; No. of Pages9
ARTICLE IN PRESS
D.P. Debecker et al. / Applied Catalysis A: General xxx (2013) xxx–xxx
3
Scheme 1. Schematic view of the preparation of a WO3-SiO2-Al2O3 catalyst by the aerosol process. The process has been realized in a commercial Büchi B-290 spray drier
equipped with a cylindrical drying chamber with circulating gases and a cyclone separator. For the sake of clarity, the drying and the particle separation are here schematized
by a simple tubular furnace and a filter. (a) The initial solution (and thus the aerosol droplets produced from it by a spraying nozzle) contains a homogeneous aqueous solution
of ethanol, Pluronic P123, H3PW12O40 (Keggin heteropolyanion), aluminum sec-butoxide, TEOS and TPAOH. (b) During drying, solvent evaporation induces the self-assembly
of the surfactant and the fast polycondensation of the silica–alumina precursors. The Keggin anions are trapped in this matrix. (c) The collected particles have a relatively
spherical shape and, after calcination, a mesoporous texture imposed by the surfactant.
up to 900 ◦C with rate of 10 K min−1 and held at this temperature
for 20 min. Hydrogen consumption was monitored by a quadrupole
mass spectrometer (Pfeiffer Vacuum OmniStar 200).
2.2.3. X-ray diffraction (XRD)
The samples were studied on a STADI P automated transmission
diffractometer from STOE (Darmstadt, Germany) with an incident
beam curved germanium monochromator selecting CuK␣1 radia-
◦
˚
tion (ꢀ = 1.5406 A) operated at 40 kV and 40 mA, with a 6 linear
position sensitive detector (PSD). The alignment was checked with
a silicon standard. The data were collected in the 2ꢁ range from 5
to 60◦ with a step size of 0.5◦ and an acquisition time of 50 seconds
per step. The phase composition of the samples was determined
using the program suite WINXPow by STOE&CIE with inclusion of
the Powder Diffraction File PDF2 of the ICDD (International Center
of Diffraction Data).
2.2.5. Temperature-programmed desorption of ammonia
(NH3-TPD)
The acidity of the catalysts was studied in a CATLAB instrument
from Hiden Analytical. About 100 mg of fresh sample was loaded in
a straight quartz reactor and degased under He (50 mL min−1) up
to 300 ◦C (10 K min−1) maintained for 30 min. Then the sample was
cooled down to 50 ◦C in the same He flow. A NH3/He (5/95) flow was
admitted for 30 min (20 mL min−1) still at 50−◦1C and the sample was
then flushed by a pure He flow (50 mL min ) for 40 min. Finally,
the sample was heated up to 600 ◦C (15 K min−1) and held at this
temperature for 10 min. NH3 desorption was monitored by a mass
spectrometer (QGA model).
2.2.4. Temperature programmed reduction (TPR)
Tungsten oxide reducibility was measured for each sample.
70 mg of each fresh catalyst was filled into fixed-bed continuous
flow quartz reactors. Prior to TPR experiment, all samples were oxi-
dized up to 550 ◦C (10 K min−1) in a flow of 20% O2 in Ar, with a flow
rate of 8 mL min−1. Afterwards the samples were cooled down to
50 ◦C in an Ar flow. The reduction proceeded in a gas mixture of
5% H2 in Ar at a flow rate of 7 mL min−1. The samples were heated
2.2.6. Transmission electron microscopy (TEM)
Samples were dispersed in ethanol and then deposited on a
TEM grid and observed with a TECNAI 12 G2 Spirit Twin from FEI
Table 1
Experimental WO3 weight loading and textural properties of the studied samples.
Catalyst/carrier denotation
ICP- WO3 (wt.%)
SBET (m2 g−1
)
Pore volume (cm3 g−1
)
Mean pore diameter (nm)
Surface densitya
(W atoms nm−2
)
SiO2 (Davisil)
Al2O3(Evonik)
SiAl (Aldrich)
SiAl-aer
WOx-imp-Si
WOx-imp-Al
WOx-imp-SiAl
WOx-impSiAl-aer
WOxSiAl-aer
–
–
–
–
9.8
9.9
9.9
8.7
9.3
301
208
539
680
284
206
325
437
534
1.1
14.2
11.9
5.5
–
–
–
–
0.9
1.2
0.8
0.5
0.5
0.62
0.73
0.62
0.96
0.59
0.52
0.50
0.70
3.7
13.5
11.5
6.4
4.6
5.3
a
Calculated as number of W atoms per square nanometer of the catalyst and considering that all W atoms are at the surface (which is a priori not true for aerosol catalyst).
Please cite this article in press as: D.P. Debecker, et al., Appl. Catal. A: Gen. (2013), http://dx.doi.org/10.1016/j.apcata.2013.06.041

GModel
APCATA-14301; No. of Pages9
ARTICLE IN PRESS
4
D.P. Debecker et al. / Applied Catalysis A: General xxx (2013) xxx–xxx
operated at 120 keV. The camera is an Orius 1000 HR 4 K × 4 K from
number of moles of propene produced per gram of catalyst and per
hour. Selectivity to propene is derived from corresponding yield
and conversion of both ethylene and trans-2-butene.
GATAN, located at the bottom of the column.
2.2.7. Temperature-programmed oxidation
3. Results and discussion
The amount of carbon deposited during the metathesis reaction
was determined from the amount of CO and CO2 evolved during
temperature-programmed oxidation (TPO) of the used catalysts.
After the catalytic experiment (about 90 min on-stream), the sam-
ples were cooled down to room temperature in N2 and 80 mg from
each catalyst was taken for TPO measurements in a setup, which
allows parallel treatment of up to 8 reactors. The samples were
heated in a 7 mL min−1 flow of O2/Ar (5/95) up to 900 ◦C with a
ramp of 10 K min−1 and hold for 20 min at this temperature. Oxygen
consumption and formation of reaction products (COx) were moni-
tored by quadrupole mass spectrometer (Pfeiffer Vacuum OmniStar
200). The following atomic mass units (AMUs) were analyzed: 44
(CO2), 40 (Ar), 32 (O2), 28 (CO, CO2), and 18 (H2O). The concen-
trations of O2, CO and CO2 were determined from the respective
AMUs using standard fragmentation patterns and sensitivity fac-
tors determined by analyzing calibration gas mixture.
3.1. Preparation of the aerosol catalyst and main properties
Scheme 1 describes the method used to prepare the aerosol cat-
alyst. The solution containing the W, Si and Al sources, as well
as the surfactant is sprayed in the form of an aerosol and spray
dried. Each aerosol droplet has initially the same composition as
the precursor solution because all precursors are soluble. Dur-
ing drying, the surfactant forms micelles via evaporation induced
self-assembly and the silica–alumina matrix condenses fast, trap-
ping the W precursor in a homogeneous distribution. Particles
are then separated from the gas flow in a cyclone separator. By
calcination, the surfactant is removed and a true mesoporous
mixed oxide is formed. TEM observations revealed that the aerosol
catalyst consists in roughly spherical particles with a regular
mesoporosity (Fig. 1). Typically, spheres in the 0.5–5 ␮m range
are observed. Larger particles most probably have an empty
core and have broken down during sample handling (calcina-
tion, pressing, sieving). At high magnification, the porosity can
be observed, in the form of lighter spots in the 7–9 nm range.
This porosity is present in all particles. In the following, this
aerosol catalyst will be compared to other reference catalysts
prepared by impregnation on silica, alumina and silica–alumina
supports.
2.2.8. Catalytic test
The cross metathesis of ethylene and trans-2-butene was car-
ried out in a multi-channel apparatus described in detail in [6].
The automated set-up allows catalytic test of up to 15 catalysts,
either in parallel or consecutive mode, with control of gas feeds and
of 3 temperature zones (gas pre-heating, reactor and post reactor
lines) along with reactor switching and product sampling. 0.1 g of
the catalysts – 250–600 ␮m particle size range – was placed into
straight quartz reactors (5 mm i.d.). The pre-treatment of all reac-
tors was done in parallel by heating up to 550 ◦C (temperature ramp
of 5 K min−1) in a nitrogen flow (14 mL min−1 in each reactor) and
holding at this temperature for 3 h. Afterwards, the system was
cooled down to the reaction temperature (250 ◦C) under the same
N2 flow. An 8 mL min−1 reaction feed consisting of ethylene and
trans-2-butene (1:1 molar ratio) together with 10 vol.% N2 (used
as an internal standard), was admitted sequentially in each reac-
tor. The initial metathesis activity of each catalyst was measured for
about 1.8 h on-stream. During the activity measurement of one cat-
alyst, the other reactors were kept under N2 flow. Ethylene (Linde,
purity > 99.95%), trans-2-butene (Linde, purity > 99.0%) and nitro-
gen (Air Liquide, purity > 99.999%) were used for catalytic tests.
Ethylene and trans-2-butene were extra purified using molsieve 3A
(Roth). An additional gas filter cartridge was used to remove oxy-
gen and traces of COx from nitrogen flow (Oxysorb, Linde). The feed
components and reaction products were quantitatively analyzed by
an Agilent 6890 GC equipped with thermal conductivity (TCD) and
flame ionization (FID) detectors. The TCD was used for N2 analysis
(molsieve 5A column), while hydrocarbons were separated on an
HP-AL/M column and detected by FID. Nitrogen was used as inter-
nal inert standard. The conversion of ethylene and trans-2-butene
was calculated on the basis of inlet and outlet concentrations of
these components (Eq. (1)).
3.2. Comparative characterization of aerosol and reference
catalysts
3.2.1. Elementary analysis
The experimental composition of all samples was close to the
nominal one (Table 1). Thus, the aerosol process – like impregna-
tion methods – offers a good control on the final composition. It
should be noted that this is not necessarily the case with other
one-step sol–gel methods in which the different precursors are
allowed to react for prolonged time. Indeed different precursors can
react at different rate or precipitate at the bottom of the flask and
some compounds can be lost during washing and filtration steps.
All this often results in a poor control of the composition. Here the
kinetic quenching that is done during the rapid drying of the aerosol
droplets is the key to control the composition.
3.2.2. N2-physisorption
Textural measurements showed that all samples are meso-
porous (with typical Type IV isotherms [29]) and exhibit relatively
large specific surface area and pore volume (Fig. 2 and Table 1).
For all commercial supports a decrease in specific surface area
and pore volume is observed after impregnation, especially for
the silica–alumina. At the same time, the average pore diame-
ter and the pore size distribution (supplementary materials Fig.
S1) are not markedly affected, which suggests that tungsten oxide
phase preferably gets deposited at the outer surface of the sup-
ports and not in the pores. The SiAl matrix prepared by aerosol
was also mainly mesoporous with high specific surface area and
large pore volume. Additionally some micropores are also present
(microporous volume estimated by t-Plot analysis is 0.23 mL g−1),
as already reported for such matrixes and rationalized by the tex-
turing effect of TPAOH [23]. Like in the case of the commercial
SiAl support, the impregnation results in a marked decrease in
both specific surface area and pore volume together with a slight
modification of the mean pore diameter, suggesting again that part
of the deposit does not enter the pores.
ꢀ
ꢁ
ꢂoi utlet
ꢂiinlet
Xi = 1 −
(1)
The propene yield is calculated on the basis of two feed compo-
nents, i0 and i1 (Eq. (2))
ˇC
· ꢂC
H
H
,outlet
3
6
6
ˇi0 · ꢂi0,inlet +3ˇi1 · ꢂi1,inlet
YC
=
(2)
H
3
6
where ꢂ stands for mole fraction and ˇ is a coefficient equal to
the number of C atoms in C3H6, C2H4 (i0) or C4H8 (i1). Subscripts
“outlet” and “inlet” are used to distinguish between mole fractions
at the reactor outlet and inlet, respectively. Specific activity is the
Please cite this article in press as: D.P. Debecker, et al., Appl. Catal. A: Gen. (2013), http://dx.doi.org/10.1016/j.apcata.2013.06.041

GModel
APCATA-14301; No. of Pages9
ARTICLE IN PRESS
D.P. Debecker et al. / Applied Catalysis A: General xxx (2013) xxx–xxx
5
Fig. 1. TEM micrographs of the WOxSiAl-aer catalyst at various magnification.
The WOx-SiO2-Al2O3 aerosol catalyst prepared in one pot
exhibited an excellent texture, with a large specific surface area
(525 m2 g−1) and large pore volume (0.7 mL g−1). Its texture was
very similar that of SiAl-aer. The sample is clearly mesoporous but
again some microporosity is observed (microporous volume evalu-
ated by t-plot is 0.14 mL g−1). The average pore diameter is 5.3 nm
which is not too far from the estimation made by TEM (∼7 nm)
and which can be correlated to the formation of P123 micelles [30]
during the fast aerosol process.
the tungsten surface density can be estimated. It lays around
1 tungsten (W) atom per nm2 for impregnated catalysts, which
is significantly lower than the theoretical monolayer (5.5 W
atoms per nm2) [28]. In the aerosol catalyst, W atoms can be
located at the surface or in the bulk of the mesopores walls
since they are trapped in a sol–gel matrix during the prepa-
ration. Even assuming that all tungsten atoms are located at
the surface, the coverage (surface density) would be 0.5 W
atoms per nm2 for the aerosol catalyst (Table 1). Thus in
all catalysts, the available surface area is large enough to
accommodate a 10 wt.% WOx loading in the form of a sub-
monolayer.
In impregnated catalysts, tungsten-oxo species are by defi-
nition deposited at the surface of a preformed support. Taking
into account the WOx loading and the specific surface area,
700
b
c
a
600
500
400
300
200
100
0
WOx-imp-Si
Si
SiAl-aer
SiAl
WOx-imp-SiAl
WOxSiAl-aer
Al
WOx-impSiAl-aer
WOx-imp-Al
0.8 1.0
0.0
0.0
0.2
0.4
0.6
P/P0
0.8
1.0 0.0
0.2
0.4
0.6
P/P0
0.2
0.4
0.6
P/P0
0.8
1.0
Fig. 2. N2-physisorption isotherms (adsorption and desorption) measured on (a) the commercial supports, (b) the catalysts prepared by impregnation over commercial
supports and (c) the aerosol-made materials, i.e. the SiAl prepared by aerosol (SiAl-aer), the catalyst prepared by impregnation on this aerosol made SiAl powder (WOx-
impSiAl-aer) and the catalyst made in one pot by the aerosol process (WOxSiAl-aer).
Please cite this article in press as: D.P. Debecker, et al., Appl. Catal. A: Gen. (2013), http://dx.doi.org/10.1016/j.apcata.2013.06.041

GModel
APCATA-14301; No. of Pages9
ARTICLE IN PRESS
6
D.P. Debecker et al. / Applied Catalysis A: General xxx (2013) xxx–xxx
WOx-imp-Al
WOx-impSiAl-aer
*
WOx-imp-Si
WOxSiAl-aer
WOx-imp-Al
WOx-imp-SiAl
WOx-impSiAl-aer
WOxSiAl-aer
10
20
30
40
50
60
WOx-imp-Si
2
130°C
Fig. 4. XRD patterns of supported tungsten oxide catalysts.
WOx-imp-SiAl
[32]. WOx-imp-Si has shown similar reduction profile but shifted to
lower temperatures by about 55 ◦C. This correlates well with XRD
data and further indicates that tungsten oxide forms crystals which
do not interact strongly with the support. All other catalysts showed
different reduction profiles. WOx-imp-Al was characterized by one
low-temperature reduction of weak intensity. Thus some of the
WOx surface species appear relatively reducible. All other cata-
lysts based on SiAl were hardly reducible. The catalysts prepared
by impregnation on silica–alumina (either the commercial or the
aerosol-made) present a reduction peak centered at 820 ◦C. For the
one-pot WOxSiAl-aer catalyst no clear reduction is observed.
The reduction behaviors of the studied catalysts inversely cor-
relate with NH3-TPD results. As acidity increases, reducibility
decreases. In other words, weak interaction between impregnated
tungsten oxide and the support (especially in the case of silica) is
reflected by an easy reduction. Conversely, strong interaction of
tungsten oxo species with the more acidic silica–alumina matrices
results in a marked resistance against reduction. Indeed the reduc-
tion facility of tungsten-oxo species is likely related to the presence
of short W O bonds, therefore acidic matrices that promote the
100
200
300
400
500
600
Temperature (°C)
Fig. 3. NH3-TPD profiles of supported tungsten oxide catalysts. All curves were
obtained with 100 mg of sample.
3.2.3. NH3-TPD
Ammonia thermo-programmed desorption curves of studied
samples are shown in Fig. 3. A main ammonia desorption peak
was observed at 130 ◦C for all catalysts, corresponding to weak acid
sites. In the catalysts made from pure supports (Si or Al) this peak
is small thus indicating a low density of acid sites. For WOx-imp-Al
an additional broad desorption shoulder was observed in the tem-
perature range 350–420 ◦C, which can be attributed to strong Lewis
acidic centers, induced probably from alumina itself [31]. Notice-
ably, the catalyst prepared by impregnation of SiAl shows a larger
desorption peak, indicating that acid sites are more abundant on
this sample. Both the one-pot WOxSiAl-aer catalyst and the catalyst
prepared by impregnation of the aerosol-made silica–alumina sup-
port (WOx-impSiAl-aer) exhibit an intense desorption peak. Also,
desorption continues at higher temperature as compared to the
WOx-imp-SiAl catalyst, thus indicating the presence of stronger
acidic sites.
formation of W OH bonds (W O + H+
W OH) are less prone to
stabilize reduced forms of tungsten oxide. Additionally, in the case
of WOxSiAl-aer one can assume that only part of the tungsten-oxo
species introduced in the preparation are actually present at the
3.2.4. XRD
The X-ray diffraction patterns of the studied catalysts (Fig. 4)
reveal that the samples based on silica–alumina are all amor-
phous. In WOx-imp-Al, the small diffraction peaks correspond to
the diffraction of the support (␥-alumina). WOx-imp-Si is the only
catalyst where nanocrystallite of WO3 were observed (peaks at
24◦). This was expected as it is well-known that W and Si oxides
develop only very weak interaction and that WOx species on silica
tends to sinter during calcination, therefore yielding WO3 [18].
630
680
WO3
795
570
WOx-imp-Al
WOx-imp-Si
515
730
WOx-imp-SiAl
WOx-impSiAl-aer
WOxSiAl-aer
820
3.2.5. H2-TPR
The reducibility of the tungsten catalysts was measured by
temperature-programmed hydrogen reduction (Fig. 5). The reduc-
tion of pure WO3 was also measured for comparison. Assuming
that the active WOx centers for metathesis are generated by par-
tial reduction of the metal oxide once contacted with the reacting
olefins, its reducibility plays an important role for the initiation
of the reaction. All reduction events were recorded at tempera-
tures higher than 500 ◦C. Pure WO3 is reduced in three steps at
630 ◦C, 680 ◦C and 795 ◦C, similar to those reported in the literature
200
400
600
800
Temperature / ºC
Fig. 5. H2-TPR profiles of supported WOx-based metathesis catalysts and of pure
WO3 used as a reference.
Please cite this article in press as: D.P. Debecker, et al., Appl. Catal. A: Gen. (2013), http://dx.doi.org/10.1016/j.apcata.2013.06.041

GModel
APCATA-14301; No. of Pages9
ARTICLE IN PRESS
D.P. Debecker et al. / Applied Catalysis A: General xxx (2013) xxx–xxx
7
80
70
60
50
40
30
20
10
0
interacts strongly with acidic sites perform better. Partial reduc-
tion of WOx species is known to occur at the initial stages of the
metathesis by reaction with the olefin feed [33]. This reduction
is involved in the formation of tungsten-carbene intermediate,
initiator and propagator for metathesis reaction [14]. However,
over-reduction of the WOx sites can cause irreversible catalyst
deactivation. In agreement with Amakawa et al. who describe the
genesis of active sites in MoOx-based metathesis catalysts [33], we
can assume that acidic matrices promote the formation of W OH
bonds with the appropriate reducibility, favoring the formation of
tungsten-carbene intermediate. It is also possible that this catalyst
exhibits more numerous single W sites which are significantly eas-
ier to activate from the oxide form to the active alkylidene species.
WO3/SiO2 are known to be poorly active at relatively low tem-
perature – as observed here at 250 ◦C – and are thus industrially
operated at relatively high temperature (typically 400 ◦C). Being
poorly acidic, the thermally promoted acidity of the catalyst sup-
port is key to generate active sites. Here, in silica–alumina based
catalysts, WOx species in interaction with a stronger acid site
undergo the electron withdrawing effect of the latter, which is
beneficial for the reaction with the olefin and the formation of the
carbene. Thus, the nature of the support has an obvious impact.
Among the silica–alumina based catalysts tested here, the fact
that WOx-impSiAl-aer performs better than WOx-imp-SiAl can ten-
tatively be attributed to the higher acidity exhibited by the former
catalyst, as well as its higher specific surface area. It is remark-
able that the one-pot WOxSiAl-aer catalyst perform even better.
Compared to WOx-impSiAl-aer, it has the same WO3 loading and
acidity and specific surface area in the same range. Moreover, it is
expected that a significant proportion of W atoms is “lost” in the
bulk of the material and not available at the surface. It is well known
[34–36] that only a minor percentage of active centers turn out to
be active metathesis sites in such heterogeneous systems. Thus the
superior activity of WOxSiAl-aer can be explained either by a larger
proportion of the surface species turning into active sites and/or
by a higher intrinsic activity of the active sites formed. A similar
observation was already made for MoO3-SiO2-Al2O3 mixed oxide
catalysts prepared by one-step methods [24,27,37]. It has been
rationalized experimentally by demonstrating that a true mixed
oxide with a homogeneous composition can undergo significant
modifications during calcination [38]. As Mo oxide has a low Tam-
mann temperature and a limited solubility in silica, it tends to
migrate toward the surface of the mixed oxide and generate abun-
dant isolated MoOx surface species. It is put forward that such
phenomenon can take place in the present case too. Therefore, the
actual nature of the WOx surface species in aerosol and impreg-
nated catalysts is different, explaining the difference in activity.
More precisely, highly isolated WOx species are probably stabilized
at the surface of WOxSiAl-aer and these species interact more with
the surrounding silica–alumina matrix.
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
time on stream / h
Fig. 6. Specific activity (mmol of propene produced per g of catalyst and per hour)
with time on stream over (᭹) WOxSiAl-aer, ( ) WOx-impSiAl-aer, ( ) WOx-imp-
SiAl, ( ) WOx-imp-Al, ( ) WOx-imp-Si.
surface of the catalyst and therefore accessible for reduction. This
might explain the fact that reduction signals are barely observed.
3.3. Olefin metathesis activity
The catalytic activity of these tungsten oxide-based catalysts
was tested in the gas phase metathesis of ethene and trans-2-
butene to propene at 250 ◦C. This reaction temperature has been
chosen after a preliminary study that confirmed that catalysts
based on silica–alumina and alumina are already active at rela-
tively low temperature, as compared to silica-based catalysts (see
supplementary materials Fig. S2). It should be noted that run-
ning the industrial reaction at lower temperature would reduce
the environmental impact of the process. Conversion and selec-
tivity varied markedly from one sample to another (vide infra).
Since the product of interest is propene, the specific activity was
defined as the number of mmol of propene produced by 1 g of
catalyst in 1 h. The specific activity of the different catalysts is
plotted in Fig. 6. Clearly, metathesis activity was poor over WOx-
imp-Si. Alumina supports provide better performances reaching
specific activity of 23 mmol of propene per gram of catalyst and
per hour. Moreover this activity remains relatively stable over
the tested period. The catalysts based on silica–alumina sup-
ports performed much better. Indeed, the catalyst prepared by
impregnation over the commercial silica–alumina support reaches
about 45 mmol g−1 h−1 and those prepared by impregnation of the
Selectivity is also highly dependent on the catalyst formulation.
In Fig. 7 the selectivity to propene, cis-2-butene and 1-butene is
plotted against the trans-2-butene conversion. Interestingly, the
most active catalyst was also the most selective to propene. WOx-
imp-Si is poorly active and the molecules of trans-2-butene that
are converted actually yield mainly cis-2-butene. This isomeriza-
tion reaction is accompanied by the isomerization to 1-butene. Note
that only the latter should be considered as a by-product since cis-
2-butene can still react with ethene and yield propene. It is clear
that WOx-imp-Si exhibits a very low amount of metathesis sites.
WO3 crystals are known to be inactive in metathesis but can cat-
alyze some isomerization reactions. Besides, if more dispersed WOx
species are also present on the catalyst surface, these are not acidic
enough to yield active metathesis centers.
aerosol-made silica–alumina powder produces 52 mmol g−1 h−1
.
Strikingly, the one-pot WOxSiAl-aer is the most active, reaching
more than 65 mmol g−1 h−1. In the conditions used here, such level
of activity corresponds to a propene yield of about 39%. Though
some deactivation is observed (as well as a short induction period),
the stability of these catalysts is much better than that exhibited by
similar Mo-based metathesis catalysts tested in the same experi-
mental set-up [24].
In the series of catalysts reported in the present study,
metathesis activity is clearly correlated with matrix acidity and
anticorrelated with reducibility of tungsten-oxo species. In other
words, the catalysts supports in which tungsten-oxo species
Obviously, isomerization can also occur on the other
–
more acidic – catalysts. Both cis-2-butene and 1-butene are
Please cite this article in press as: D.P. Debecker, et al., Appl. Catal. A: Gen. (2013), http://dx.doi.org/10.1016/j.apcata.2013.06.041

GModel
APCATA-14301; No. of Pages9
ARTICLE IN PRESS
8
D.P. Debecker et al. / Applied Catalysis A: General xxx (2013) xxx–xxx
0.8
(a)
0.02
0.6
0.4
0.2
0.0
0.01
0.00
l
r
e
er
-a
l
Si
a
-
l
A
A
mp-
i
i
S
Si
p
3-
WO
-imp-SiA
3
3
im
3-
WO
WO
O
W
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
Fig. 8. Deposited carbon during the metathesis per gram of WOx-based catalysts.
X(trans-2-butene) / -
WOx-impSiAl-aer showed a catalytic behavior similar to that of
WOx-imp-SiAl. As suggested by N2-physisorption, WOx dispersion
is probably not optimal in these catalysts since the impregnated
phase appears to be mainly deposited at the outer surface of the
support particles. Among the studied tungsten catalysts the highest
catalytic performance and selectivity was obtained over WOxSiAl-
aer. WOx surface species in this catalyst seem to benefit from the
most appropriate interaction with the silica–alumina matrix both
in terms of dispersion and acidity.
0.8
0.6
0.4
0.2
0.0
(b)
3.4. Post-test characterization of the carbon deposit
3.4.1. TPO
After 1.75 h on-stream the used catalysts were investigated by
temperature-programmed oxidation in an attempt to quantify the
amount of organics that accumulate on the catalyst surface by
adsorption or coking. The method allows quantitative analysis of
COx and H2O produced during heating up to 900 ◦C. CO and CO2
desorption profiles are shown as supplementary materials (Fig. S3).
The CO2 evolution for the three SiAl catalysts was pronounced with
one peak at 420 ◦C but being much larger for the aerosol made sam-
ples. The total amount of carbon evolved per gram of catalyst is
shown in Fig. 8. The amount of coke was roughly correlated with
the catalytic activity. Interestingly, there is also a correlation with
C5 selectivity (see supplementary materials Fig. S4 and S5). Cok-
ing is tentatively identified as a cause for the slight deactivation
observed in the most active catalysts. In this perspective, it should
be taken into account that the new catalysts might need more fre-
quent regeneration procedures as compared to the usual WO3/SiO2
catalysts.
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
X(trans-2-butene) / -
0.8
0.6
0.4
0.2
0.0
(c)
4. Conclusion
The results reported in this comparative study encourage the
statement that relative low W-atoms density dispersed on well-
balanced acid support surface is of paramount importance in the
genesis of active W-carbene species. It is known that only a small
proportion of the metal oxo species turn out to be active metal-
locarbenes in such heterogeneous catalysts. The stability of this
structure is just as crucial for the reaction as the initial geometric
assembly of metal-oxo species. This latter appeared to be a limiting
factor for transformation of surface WOx-species into the metal
carbene sites upon contact with olefin. In impregnated catalysts,
the nature of the support has a drastic influence. While silica-
supported tungsten oxide is barely active at 250 ◦C, the more acidic
silica–alumina-supported catalysts exhibit significant metathesis
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
X(trans-2-butene) / -
Fig. 7. Selectivity to (a) propene, (b) cis-2-butene and (c) 1-butene vs. trans-2-
butene conversion: (᭹) WOxSiAl-aer, ( ) WOx-impSiAl-aer, ( ) WOx-imp-SiAl,
(
) WOx-imp-Al, ( ) WOx-imp-Si.
systematically detected. However, as tungsten oxo species gets
better dispersed on more acidic supports, more abundant and/or
more active metathesis sites are formed. Obviously, the metathesis
reaction is favored over isomerization reactions when chang-
ing the support from silica to alumina and to silica–alumina.
Please cite this article in press as: D.P. Debecker, et al., Appl. Catal. A: Gen. (2013), http://dx.doi.org/10.1016/j.apcata.2013.06.041

GModel
APCATA-14301; No. of Pages9
ARTICLE IN PRESS
D.P. Debecker et al. / Applied Catalysis A: General xxx (2013) xxx–xxx
9
activity, outcompeting the alumina-supported catalyst too. Thus,
ternary Si-Al-W mixed oxide catalysts were prepared in one-step,
taking advantage of the aerosol process. The latter appears to
be a powerful new method to yield in one step highly efficient
metathesis catalysts. Combined with sol–gel chemistry, evapora-
tion induced self-assembly and polyanions, it allows a fine tuning of
the dispersion of WOx species and the production of acidic matrices
which are key parameters controlling the efficiency of the metathe-
sis active species. These spray dried catalysts present a very high
activity along with favorable propene selectivity, therefore provid-
ing enhanced propene yields.
[9] K. Bouchmella, P.H. Mutin, M. Stoyanova, C. Poleunis, P. Eloy, U. Rodemerck,
E.M. Gaigneaux, D.P. Debecker, J. Catal. 301 (2013) 233–241.
[10] L. Harmse, C. van Schalkwyk, E. van Steen, Catal. Lett. 137 (2010)
123–131.
[11] H.J. Liu, S.J. Huang, L. Zhang, S.L. Liu, W.J. Xin, L.Y. Xu, Catal. Commun. 10 (2009)
544–548.
[12] E. Mazoyer, K.C. Szeto, N. Merle, S. Norsic, O. Boyron, J.-M. Basset, M. Taoufik,
C.P. Nicholas, J. Catal. 301 (2013) 1–7.
[13] M. Taoufik, E. Le Roux, J. Thivolle-Cazat, J.-M. Basset, Angew. Chem., Int. Ed. 46
(2007) 7202–7205.
[14] M. Tlenkopatchev, S. Fomine, J. Organomet. Chem. 630 (2001)
157–168.
[15] S.J. Huang, S.L. Liu, Q.J. Zhu, X.X. Zhu, W.J. Xin, H.J. Liu, Z.C. Feng, C. Li, S.J. Xie,
Q.X. Wang, L.Y. Xu, Appl. Catal., A 323 (2007) 94–103.
[16] N. Liu, S. Ding, Y. Cui, N. Xue, L. Peng, X. Guo, W. Ding, Chem. Eng. Res. Des. 91
(2013) 573–580.
Acknowledgments
[17] D.P. Debecker, M. Stoyanova, U. Rodemerck, A. Leonard, B.-L. Su, E.M.
Gaigneaux, Catal. Today 169 (2011) 60–68.
[18] S. Chaemchuen, S. Phatanasri, F. Verpoort, N. Sae-ma, K. Suriye, Kinet. Catal. 53
(2012) 247–252.
[19] A. Spamer, T.I. Dube, D.J. Moodley, C. Van Schalkwyk, J.M. Botha, Appl. Catal. A
255 (2003) 153–167.
The authors thank Dr. Christophe Poupin for helping on the
NH3-TPD measurements and M. Patrick Legriel for performing the
TEM analyses. D.P. Debecker thanks the FNRS for his Postdoctoral
Researcher position. The authors from UCL are involved in the
“Inanomat” IUAP network sustained by the Service public fédéral
de programmation politique scientifique (Belgium).
[20] B. Schimmoeller, S.E. Pratsinis, A. Baiker, ChemCatChem
1234–1256.
3
(2011)
[21] D.P. Debecker, V. Hulea, P.H. Mutin, Appl. Catal. A 451 (2013) 192–206.
[22] D.P. Debecker, P.H. Mutin, Chem. Soc. Rev. 41 (2012) 3624–3650.
[23] S. Pega, C. Boissiere, D. Grosso, T. Azais, A. Chaumonnot, C. Sanchez, Angew.
Chem., Int. Ed. 48 (2009) 2784–2787.
[24] D.P. Debecker, M. Stoyanova, F. Colbeau-Justin, U. Rodemerck, C. Boissière, E.M.
Gaigneaux, C. Sanchez, Angew. Chem., Int. Ed. 51 (2012) 2129–2131.
[25] C. Boissiere, D. Grosso, A. Chaumonnot, L. Nicole, C. Sanchez, Adv. Mater. 23
(2011) 599–623.
[26] D.P. Debecker, D. Hauwaert, M. Stoyanova, A. Barkschat, U. Rodemerck, E.M.
Gaigneaux, Appl. Catal., A 391 (2011) 78–85.
Appendix A. Supplementary data
Supplementary data associated with this article can be
found, in the online version, at http://dx.doi.org/10.1016/
j.apcata.2013.06.041.
[27] D.P. Debecker, K. Bouchmella, M. Stoyanova, U. Rodemerck, E.M. Gaigneaux,
P.H. Mutin, Catal. Sci. Technol. 2 (2012) 1157–1164.
[28] N. Naito, N. Katada, M. Niwa, J. Phys. Chem. B 103 (1999) 7206–7213.
[29] K.S.W. Sing, Pure Appl. Chem. 54 (1982) 2201–2218.
References
[1] J.L.G. Fierro, J.C. Mol, in: J.L.G. Fierro (Ed.), Metal Oxides: Chemistry and Appli-
cations, Taylor & Francis, Boca Raton, 2006, p. 517.
[30] A. Galarneau, H. Cambon, F. Di Renzo, F. Fajula, Langmuir 17 (2001) 8328–8335.
[31] G. Busca, Phys. Chem. Chem. Phys. 1 (1999) 723–736.
[2] I. Rodriguez-Ramos, A. Guerrero-Ruiz, N. Homs, P. Ramirez De la Piscina, J.L.G.
Fierro, J. Mol. Catal. A 95 (1995) 147–154.
[3] J.C. Mol, J. Mol. Catal. A 213 (2004) 39–45.
[4] D.P. Debecker, M. Stoyanova, U. Rodemerck, P. Eloy, A. Léonard, B.-L. Su, E.M.
Gaigneaux, J. Phys. Chem. C 114 (2010) 18664–18673.
[5] D.P. Debecker, M. Stoyanova, U. Rodemerck, E.M. Gaigneaux, J. Mol. Catal. A 340
(2011) 65–76.
[32] D. Hua, S.-L. Chen, G. Yuan, Y. Wang, Q. Zhao, X. Wang, B. Fu, Microporous
Mesoporous Mater. 143 (2011) 320–325.
[33] K. Amakawa, S. Wrabetz, J. Kröhnert, G. Tzolova-Müller, R. Schlögl, A. Trunschke,
J. Am. Chem. Soc. 134 (2012) 11462–11473.
[34] J. Handzlik, J. Ogonowski, Catal. Lett. 88 (2003) 119–122.
[35] A. Salameh, C. Coperet, J.M. Basset, V.P.W. Bohm, M. Roper, Adv. Synth. Catal.
349 (2007) 238–242.
[6] D.P. Debecker, B. Schimmoeller, M. Stoyanova, C. Poleunis, P. Bertrand, U. Rode-
merck, E.M. Gaigneaux, J. Catal. 277 (2011) 154–163.
[7] A. Behr, U. Schüller, K. Bauer, D. Maschmeyer, K.D. Wiese, F. Nierlich, Appl.
Catal., A 357 (2009) 34–41.
[36] Y. Chauvin, D. Commereuc, J. Chem. Soc. Chem. Commun. 0 (1992) 462–464.
[37] D.P. Debecker, K. Bouchmella, C. Poleunis, P. Eloy, P. Bertrand, E.M. Gaigneaux,
P.H. Mutin, Chem. Mater. 21 (2009) 2817–2824.
[38] D.P. Debecker, K. Bouchmella, R. Delaigle, P. Eloy, C. Poleunis, P. Bertrand, E.M.
Gaigneaux, P.H. Mutin, Appl. Catal., B 94 (2010) 38–45.
[8] J.C. Mol, Catal. Today 51 (1999) 289–299.
Please cite this article in press as: D.P. Debecker, et al., Appl. Catal. A: Gen. (2013), http://dx.doi.org/10.1016/j.apcata.2013.06.041