CH3-ReO3 on γ-Al2O3: Activity, selectivity, active site and deactivation in olefin metathesis

Article abstract of DOI:10.1016/j.jcat.2007.10.007

The active sites (15% of total Re) of CH₃ReO₃ supported on alumina arise solely from the reaction of the C{single bond}H bond of CH₃ReO₃ with reactive Al_S{single bond}O_S sites of a γ-Al₂O₃ surface to yield [Al_SCH₂ReO₃], whereas the major species, inactive, corresponds to MeReO₃ chemisorbed through its oxo ligand(s) on Lewis acid sites of alumina. Monitoring the active sites of CH₃ReO₃ supported on alumina by solid-state NMR spectroscopy shows that no carbenic signals are observed and that the resting states of the catalyst in the metathesis of propene correspond to μ-methylene and μ-ethylidene species, so that the propagating carbenic species are likely formed only in the presence of olefins. Although this system is highly active in the metathesis of propene, with initial rates similar to some of the best catalysts, it undergoes fast deactivation, which is due to the propene metathesis products, most likely ethene.

Full text of DOI:10.1016/j.jcat.2007.10.007

Journal of Catalysis 253 (2008) 180–190
www.elsevier.com/locate/jcat
CH3-ReO3 on γ -Al2O3:
Activity, selectivity, active site and deactivation in olefin metathesis
Alain Salameh a, Anne Baudouin a, Daravong Soulivong a, Volker Boehm b, Michael Roeper b,
Jean-Marie Basset a,∗, Christophe Copéret a,∗
a
Laboratoire de Chimie Organométallique de Surface (UMR 9986 CNRS/ESCPE Lyon), ESCPE Lyon,
F-308, 43 Boulevard du 11 Novembre 1918, F-69616 Villeurbanne Cedex, France
b
BASF AG, Forschung Grundprodukte, GCB/K, M 313, D-67056 Ludwigshafen, Germany
Received 13 June 2007; revised 17 August 2007; accepted 16 October 2007
Available online 19 November 2007
Abstract
The active sites (15% of total Re) of CH ReO supported on alumina arise solely from the reaction of the C–H bond of CH ReO with reactive
3
3
3
3
Al –O sites of a γ -Al O surface to yield [Al CH ReO ], whereas the major species, inactive, corresponds to MeReO chemisorbed through
S
S
2
3
S
2
3
3
its oxo ligand(s) on Lewis acid sites of alumina. Monitoring the active sites of CH ReO supported on alumina by solid-state NMR spectroscopy
3
3
shows that no carbenic signals are observed and that the resting states of the catalyst in the metathesis of propene correspond to μ-methylene and
μ-ethylidene species, so that the propagating carbenic species are likely formed only in the presence of olefins. Although this system is highly
active in the metathesis of propene, with initial rates similar to some of the best catalysts, it undergoes fast deactivation, which is due to the
propene metathesis products, most likely ethene.
© 2007 Elsevier Inc. All rights reserved.
Keywords: Heterogeneous catalysis; Olefin metathesis; Methyl trioxorhenium; Alumina; Active site; Flow kinetic; Deactivation
1. Introduction
catalyst [7]. Activity has been associated with the Lewis acidity
of the support [8], and in fact active systems based on CH3ReO3
use Lewis acid-containing supports [8–12]. In all these sys-
tems, the structure of the active site(s) and the involvement
of the methyl as the source of the carbene have long been a
matter of debate. Using spectroscopic (IR, NMR and EXAFS)
reactivity and molecular modeling studies, we recently showed
that the active sites of CH3ReO3 supported on γ -Al2O3-(500)
arise from the methyl and corresponds to only a minor sur-
face species, [AlSCH2ReO3], present only as ca. 15% of all Re
(Scheme 1) [13]. This species results from the C–H activation
of the methyl ligand of CH3ReO3 onto reactive AlS–O sites
of alumina (where AlS denotes surface aluminum sites) [14],
thus yielding a μ-methylene bridge species [AlSCH2ReO3]
that displays a characteristic 13C NMR signal at 66 ppm [13].
Note that no Re alkylidene has been observed at lower fields
(180–400 ppm) [15], and that the μ-methylene species is ex-
changed on contacting 13C-labeled ethene with CH3ReO3/
γ -Al2O3-(500), demonstrating that this μ-methylene species
Olefin metathesis (OM) has been a very important indus-
trial process since the discovery of the first heterogeneous OM
catalysts [1], and today it is associated with crackers in petro-
chemical plants to improve the yield of propene through the
cross-metathesis of ethene and 2-butenes [2]. Of the various
classical heterogeneous OM catalysts, Re2O7/Al2O3 is the only
one that works efficiently at room temperature [3,4]. Moreover,
this heterogeneous catalyst also can be used to convert oleate
esters [5] when activated by organotin agents, which are, how-
ever, detrimental to the regeneration process [6]. In contrast, in
the early 1990s, Herrmann et al. showed that CH3ReO3 sup-
ported on silica–alumina is a highly active metathesis catalyst
for functionalized olefins, without the need for an additional co-
*
Corresponding authors.
E-mail address: coperet@cpe.fr (C. Copéret).
0021-9517/$ – see front matter © 2007 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2007.10.007

A. Salameh et al. / Journal of Catalysis 253 (2008) 180–190
181
Scheme 1. Grafting of CH ReO /Al O .
3
3
2 3
Scheme 2. Proposed active and inactive sites for the CH ReO /Al O .
3
3
2 3
corresponds to the resting state of the catalyst and that the
alkylidene species is probably formed in situ in the presence
of olefins (Scheme 2).
Here we present a detailed kinetic study of the propene
metathesis in a flow reactor (effect of time on stream and space
time velocities) and an investigation of the active sites and their
life cycle by titration studies and by solid-state NMR spec-
troscopy.
2.2. Preparation of γ -Al2O3-(500)
γ -Al2O3 from Degussa was calcined at 500 ◦C in dry air for
18 h and then treated at 500 ◦C under vacuum (0.13 mPa) for
18 h to give γ -Al2O3-(500) (105 m2/g).
2.3. Reaction of γ -Al2O3-(500) with CH3ReO3
γ -Al2O3-(500) (1 g) was contacted, while stirring, with
100 mg of CH3ReO3 at 25 ◦C. After 3 h, removal of the ex-
cess of molecular complex was performed by reverse subli-
mation, and the resulting solid was then treated under vacuum
(0.13 mPa) for 3 h. Elemental analysis: Re 3.8 wt%.
2. Experimental
2.1. General procedure
All gases were purified on R-3/11 provided by BASF and
MS 4 Å before use. Gas-phase analysis was performed on a
Hewlett–Packard 5890 series II gas chromatograph equipped
with a flame ionization detector and a KCl/Al2O3 on fused sil-
ica column (50 m × 0.32 mm).
2.4. Metathesis of propene catalyzed by CH3ReO3/Al2O3-(500)
CH3ReO3/Al2O3-(500) (50 mg, 10.2 µmol of Re) was intro-
duced under Ar in a 195-mL batch reactor. After evacuation
of Ar, 533 Torr of propene (547 equiv) was added. During the
reaction at 25 ◦C, aliquots were expended in a small volume,
brought to atmospheric pressure, and analyzed by gas chro-
matography (GC).
2.1.1. Solid-state NMR spectroscopy
All solid-state NMR spectra were recorded under MAS
(νR = 10 kHz) on a Brüker Avance 500 spectrometer equipped
with a standard 4-mm double-bearing probe head (125.73 MHz
for 13C). The freshly prepared samples were introduced in the
4-mm zirconia rotor in a glove box and tightly closed. Com-
pressed air was used for both bearing and driving the rotors.
Chemical shifts are reported in ppm downfield from Si(CH3)4
2.5. Metathesis of ethyl oleate catalyzed by
CH3ReO3/Al2O3-(500)
Into a 22-mL batch reactor was introduced 104 mg of
CH3ReO3/Al2O3-(500) (21.4 µmol of Re) and 2.8 mL of a 0.8 M
solution of ethyl oleate in pentane (2.24 mmol). During the re-
action at 25 ◦C, aliquots (1 to 2 drops) were taken, diluted in
pure solvent (0.2 mL), and analyzed by GC.
1
( 0.1 and 1 ppm for H and 13C NMR spectra, respectively).
The typical cross-polarization sequence was used for the 13C
CP MAS NMR spectra: 90◦ proton pulse, cross-polarization
step to carbon spins, and detection of the carbon magnetiza-
tion under proton decoupling TPPM-15 [16]. For the CP step,
a ramp radio frequency (rf) field centered at νCP = 60 kHz was
applied on protons, whereas the carbon rf field was matched to
obtain the optimal signal. The contact time for CP was set to
2 ms. An exponential line broadening of 80 Hz was applied be-
fore Fourier transform. All other details are given in the figure
captions.
2.6. Reaction of CH3ReO3/Al2O3-(500) with Z 2-butene
followed by contact with 13C di-labeled ethene (titration of
active sites)
CH3ReO3/Al2O3-(500) (98 mg, 20 µmol of Re) was intro-
duced under Ar in a 22-mL batch reactor. After evacuation of
the gas phase, Z 2-butene (44 Torr, 2.6 equiv) was added. After

182
A. Salameh et al. / Journal of Catalysis 253 (2008) 180–190
30 min, the gas phase was analyzed by GC and GC/MS and then
evacuated under high dynamic vacuum for 1 h at 25 ◦C before
being replaced by 13C-di-labeled ethene (20 Torr, 1.2 equiv).
After 30 min, the gas phase was analyzed by GC and GC/MS.
ethene line, the pressure was set to 1 bar, and the tubes were
flushed for 2 h before the inlet valve was opened. The temper-
ature was set to 30 ◦C, and the flow rate was set to 5 mL/min.
After 1188 min, the ethene valve was closed and the propene
valve was opened. The propene pressure was set to 1 bar, the
temperature to 30 ◦C, and the flow rate to 77 mL/min. The re-
action was monitored by online GC.
2.7. Reactivity of CH3ReO3/Al2O3 with 13C mono-labeled
propene for 15 min followed by contact with 13C di-labeled
ethene and unlabeled ethane
2.11. Metathesis of Z 2-butene followed by that of propene in
a flow reactor catalyzed by CH3ReO3/Al2O3
Into a 22-mL batch reactor was introduced, under Ar,
CH3ReO3/Al2O3 (262 mg, 53.6 µmol of Re). After evacua-
tion of Ar, 22 Torr of 13C mono-labeled propene (0.48 equiv)
was added. After 15 min, the gas phase was analyzed by GC
and GC/MS, and then evacuated from the reactor under high
dynamic vacuum for 1 h at 25 ◦C. Part of the used catalyst
(86 mg) was used to prepare the NMR sample and the remain-
ing part of the solid (176 mg, 36 µmol of Re) was contacted
with unlabeled ethylene (33 Torr, 1.08 equiv). After 15 h, the
gas phase was analyzed by GC and GC/MS, and the powder
was analyzed by solid-state NMR spectroscopy. The same pro-
cedure was used for contact with 13C di-labeled ethene.
After CH3ReO3/Al2O3 (58 mg, 1.18 µmol of Re) was loaded
into the reactor in a glove box, the reactor was connected to the
Z 2-butene line, the pressure was set to 1 bar, and the tubes
were flushed for 2 h before the inlet valve was opened. The
temperature was set to 30 ◦C; the flow rate, to 5 mL/min. After
1188 min, the ethene valve was closed and the propene valve
was opened. The propene pressure was set to 1 bar, the temper-
ature to 30 ◦C, and the flow rate to 77 mL/min. The reaction
was monitored by online GC.
3. Results and discussion
2.8. Metathesis of propene in a flow reactor catalyzed by
CH3ReO3/Al2O3
3.1. Preliminary tests of olefin metathesis activity of
CH3ReO3 supported on γ -Al2O3-(500) {CH3ReO3/Al2O3-(500)
}
For each reaction, the flow rates were controlled by a
mass-flow controller (Brooks); they were in the range of 6–
500 mL/min. The catalyst was first loaded in the reaction
chamber in a glove box (31 mg, 6.3 µmol of Re). The isolated
reaction chamber was then connected with the reactor system,
the propene pressure was set to 1 bar, and the flow rate was
set to 34.5 mL/min. After the tubes were flushed with propene
(purified on R-3/11 from BASF and MS 4 Å) for 2 h, the re-
action chamber valve was opened (t = 0), and the temperature
was set to 30 ◦C. The outlet flow was analyzed by online GC.
At room temperature, the CH3ReO3/Al2O3-(500) system
rapidly transformed 550 equiv of propene into a thermo-
dynamic mixture of ethene and 2-butene, completing 190
turnovers (TON) in less than an hour, which is ∼100 times
faster compared with Re2O7/Al2O3. Moreover, at room tem-
perature, CH3ReO3/Al2O3-(500) also equilibrates 100 equiv of
ethyl oleate into a thermodynamic mixture (∼1:1) of 9-octa-
decenes and diethyl 9-octadecenedioates in about 11,520 min
(8 days). The initial turnover frequency (TOF) is ca. 0.38 TON/
min (Fig. S1), which is much slower than that obtained for
propene metathesis (ca. 28 TON/min, vide infra). Although
inefficient compared with well-defined homogeneous [17–21]
and silica-supported systems [22–26], it is noteworthy, because
Re2O7/Al2O3 is practically inactive (ca. 1 TON) without the
use of a promoter such as alkyl tin reagents [5,27].
2.9. Simulation of deactivation
Deactivation was modeled by numerical simulation of the
conversion of propene as a function time (A) based on A0 (the
initial conversion of propene, which is proportional to the num-
ber of active sites {A ∝ [cat] in Eqs. (1)–(3) at t = 0 min),
k (the rate constant of deactivation), and [=], the concentration
of ethene measured in the feed (/mol):
3.2. Active site, propagating species, and resting states in
propene metathesis of CH3ReO3/Al2O3-(500)
ꢀA
Although a recent study showed that [AlSCH2ReO3] cor-
responds to an active species, pulses of Z 2-butene and 13C
di-labeled ethene were used to evaluate whether the oxo
ligands are also involved in the generation of the active
sites, as well as to obtain complementary active site titra-
tion experiments [28,29]. When CH3ReO3/Al2O3 is contacted
with 2.6 equiv of Z 2-butene, 0.23 equiv of propene/Re and
0.082 equiv of 2-pentene/Re are formed, besides a fast Z/E
isomerization of 2-butene through metathesis. The 2-pentenes
(0.082 equiv) and a fraction of propene (ca. 0.082 equiv) are
likely formed through the cross-metathesis of 2-butenes with
1-butene, present as traces in commercial Z 2-butenes (0.12%
−
−
−
= k[A],
(1)
(2)
(3)
ꢀt
ꢀA
= k[A][=]measured
,
ꢀt
ꢀA
ꢀ
ꢁ
= [A] × k[=]measured + kꢀ .
ꢀt
2.10. Metathesis of ethene followed by that of propene in a
flow reactor catalyzed by CH3ReO3/Al2O3
After CH3ReO3/Al2O3 (60 mg, 12 µmol of Re) was loaded
in the reactor in a glove box, the reactor was connected to the

A. Salameh et al. / Journal of Catalysis 253 (2008) 180–190
183
Scheme 3. Possible way of decomposition of metallacyclobutanes.
of 1-butene, i.e., 0.003 equiv) and formed in situ by isomeriza-
tion of 2-butenes on acid sites. The remaining part of propene
(0.15 equiv) is formed by cross-metathesis of 2-butenes with
[AlSCH2ReO3]. Moreover, after evacuation of the gas phase,
contacting this catalyst with 99% 13C di-labeled ethene gave
0.11 equiv of 13C mono-labeled propene along with 0.002 equiv
of 13C mono-labeled 1-butene. Whereas 13C mono-labeled
propene is expected from the cross-metathesis of an ethyli-
dene propagating species and Z 2-butene, the presence of small
amount of 13C mono-labeled 1-butene shows that propylidene
propagating species are also present, as expected from the pres-
ence of 1-butene in Z 2-butene. Noteworthy, the number of
active sites determined by this method is close to that obtained
previously with 13C-di-labeled ethene (ca. 15% of total Re)
[13], so that the oxo ligands are not involved in formation of
the active species (in contrast to Re2O7/Al2O3 [29]), so that
the coordinated CH3ReO3 on Lewis acid sites does not lead to
active species.
(a)
For propene metathesis, two propagating carbenic species
(methylidene and ethylidene) are expected. Whereas the former
has been observed as the μ-methylene species, [AlSCH2ReO3],
CH3ReO3/Al2O3 was also contacted with 0.48 equiv of propene
monolabeled on position 2 (2-13C propene) to observe se-
lectively the ethylidene species (or the corresponding rest-
ing state). After 15 min, the reaction mixture was composed
of unlabeled ethene (0.1 equiv), 13C di-labeled E 2-butene
(0.07 equiv), and 13C di-labeled Z 2-butene (0.02 equiv), the
expected olefin metathesis products, along with 13C mono-
labeled E 2-butene (4.2 × 10−3 equiv) and 13C mono-labeled
1-butene (9.4 × 10−4 equiv). These minor products are possi-
bly formed through the decomposition of metallacyclobutane
intermediates, as proposed previously (Scheme 3). The solid-
state 13C CP MAS NMR spectrum of the catalyst displays
a large, intense signal at 85 ppm along with a weaker broad
signal at 30 ppm and a sharp signal at 146 ppm (Fig. 1a),
no signal associated with carbene ligands being detected at
lower fields. The two signals at 146 and 22 ppm also are
observed when γ -Al2O3-(500) is treated with mono-labeled
2-13C propene (Fig. 1b), and thus can be attributed to the
(b)
Fig. 1. (a) C CP-MAS solid state NMR spectrum of CH ReO /Al O
13
3
3
2
3-(500)
13
after contact with mono-labeled 2- C propene, under MAS of 10 kHz. The
number of scan was 26,000 and the recycle delay was set to 2 s. A CP step of
13
2 ms was used. (b) C CP-MAS solid state NMR spectrum of γ -Al O
2
3-(500)
13
after contact with mono-labeled 2- C propene under MAS of 10 kHz. The
number of scan was 5021 and the recycle delay was set to 4 s. A CP step of
2 ms was used.
sociated with a strong Lewis acid aluminum center [13,14].
The signal at 85 ppm, down-field by 19 ppm with respect
to [AlSCH2ReO3], is only present on CH3ReO3/Al2O3-(500)
contacted with mono-labeled 2-13C propene, and is assigned
to [AlSCH(CH3)ReO3], resulting from the exchange of the
μ-methylene in [AlSCH2ReO3] by an ethylidene fragment
through cross-metathesis (Scheme 2, AlSCH(R)–ReO3, R =
CH3 and Rꢀ = H). Note also the absence of formation of
oligomers upon treatment of Al2O3 and CH3ReO3/Al2O3 with
propene, in contrast to what has been observed with ethene.
Further contacting this solid with 1.08 equiv of 13C di-labeled
ethene reveals both peaks at 66 and 85 ppm, which can be
assigned to [AlSCH2ReO3] and [AlSCH(CH3)ReO3], the two
Csp2–H and the Csp3–H carbons of adsorbed olefins [30].
Whereas the former is clearly observed in the 13C NMR
spectrum of CH3ReO3/Al2O3-(500) contacted with monola-
beled 2-13C propene, the latter is hidden under the broad peak
at 30 ppm, which is associated with the major species of
CH3ReO3/Al2O3-(500) (the coordination adduct of CH3ReO3
onto the Lewis acid sites) [13]. The lower amount of adsorbed
olefins on CH3ReO3/Al2O3-(500) as evidenced by solid-state
NMR is consistent with its lower acidity; in fact, this system
has no Brönsted acidity and contains less Lewis acid sites
than Al2O3 (Fig. S2), which in turn is consistent with the
fact that CH3ReO3 has reacted with AlS–OS surface sites as-

184
A. Salameh et al. / Journal of Catalysis 253 (2008) 180–190
Fig. 2. Conversion as a function of time on stream at a flow rate of: (F) 34.5, (") 17, (2) 6.4, (Q) 77, and (×) 165 mL/min.
resting states of the expected propagating Re species associated
with the metathesis of propene (Fig. S3, Scheme 2, R = H or
CH3, respectively).
(<1 × 10−4%) and that isobutene is formed in only minute
amounts (ca. 0.008%).
The change of selectivity at constant flow as a function of
time on stream and deactivation can be explained according to
the following hypotheses:
Finally, when CH3ReO3/Al2O3-(500) is treated first with 13
C
mono-labeled propene for 15 min and then with unlabeled
ethene (1.08 equiv), 0.05 equiv of mono-labeled propene/Re is
detected. Considering the probable presence of a 1:1 mixture
of methylene and ethylidene propagating species, and the titra-
tion of only the ethylidene species (formation of mono-labeled
propene), the number of active sites has decreased from 15 to
10% of Re species, in agreement with partial catalyst deactiva-
tion. During this process, small amounts of unlabeled propene
also are detected (6 × 10−3 equiv), most likely formed by the
decomposition of metallacyclobutanes as previously proposed
(see Scheme 3, and vide infra for further comments on deacti-
vation).
1. The catalyst has two types of two active sites, one more
“E selective” and one more “Z selective,” with different
lifetimes.
2. The nature of the Re-active site changes with time, leading
to a more “Z-selective” catalysts (evolution of the struc-
ture).
3. The active site has an intrinsic selectivity that favors Z
2-butene, but a successive fast Z–E isomerization giv-
ing E 2-butene slows down with time, thus changing the
selectivity as the catalyst ages. In the latter case, Z/E
isomerization can occur through a degenerate metathesis
reaction or can be acid-catalyzed. Therefore, it is impor-
tant to distinguish the intrinsic selectivity of the catalyst
(primary process) from the selectivity resulting from sub-
sequent isomerization processes (secondary processes) or
from a change of structure of the catalyst (Scheme 4).
Overall, the active sites result from the methyl of CH3ReO3
through the formation of a μ-methylene complex, which can
probably interconvert into an alkylidene species in the presence
of olefins.
3.3. Activity, selectivity, and stability of CH3ReO3/Al2O3-(500)
in the metathesis of propene as a function of time on stream
Consequently, we have investigated the influence of the in-
verse space velocity on selectivity (contact time studies) to gain
further insight into the catalytic system. Typically, the inverse
space velocity is defined as a volume of catalyst crossed by a
volume of reagent during a given unit of time:
At 30 ◦C with a flow rate of 34.5 mL/min (227 mol pro-
pene/mol Re min), propene is transformed into a mixture of
ethene and 2-butenes with an initial rate of 45 TON/min af-
ter 6 min (Fig. 2). Considering the amount of active sites at
t = 0 min (ca. 15%), the initial activity is at least 302 mol of
propene transformed/min/active site, comparable to the best
silica-supported well-defined systems [22–26] and much better
than Re2O7/Al2O3 (27 mol of propene transformed/min/active
site considering 2% of active site) [28,29,31,32]. Over 530 min,
conversion decreases from 20 to 5% due to a fast deactivation
process. After 1330 min, a TON of 16,500 was achieved, corre-
sponding to 110,000 TON/Re active site.
Vcat
vvh =
,
(4)
F
where Vcat = volume of catalyst; F = flow rate expressed, e.g.,
in volume unit/h. It also can be expressed in mass or mol num-
bers:
Mcat
mmh =
,
(5)
D
Noteworthy, a clear evolution of the selectivity in 2-butenes
as a function of time (decreasing E/Z ratio) and conversion
(increasing E/Z ratio) also is observed as the catalyst deacti-
vates (Fig. 3). In addition, note that 1-butene is not detected
where Mcat = mass of catalyst; D = flow rate expressed in
mass unit/h, and they are all proportional. However, these de-
finitions never take into account the number of active sites at
a given time. Therefore, as the catalyst deactivates, the inverse

A. Salameh et al. / Journal of Catalysis 253 (2008) 180–190
185
(a)
(b)
Fig. 3. Propene metathesis in a flow reactor: propene (34.5 mL/min; 227 mol P/(mol Re min)), P = 1 atm, and CH ReO /Al O
(31 mg). (a) Selectivity as a
3
3
2 3-(500)
function of time: (F) ethene, (2) E 2-butene, (Q) Z 2-butene. (b) Selectivity as a function of conversion: (F) ethene, (2) E 2-butene, (Q) Z 2-butene.
space velocity per active sites changes. To understand the sys-
tem at a molecular level, it would be informative to express
the inverse space velocity as a function of the number of ac-
tive sites, hereinafter termed intrinsic contact time (ICT):
Moreover, the same behavior has been observed when con-
tact time studies are performed at two different time-on-stream
periods (Figs. S7–S8). The E/Z ratio decreases at shorter con-
tact times (moving away from the thermodynamic ratio of 3)
but obtaining the intrinsic selectivity still requires the use of
much shorter contact times. In addition, when the selectivi-
ties of 2-butenes are plotted as a function of conversion, all
the data points for each butene isomer (including those ob-
tained at various contact times) are aligned on the same curve
(circled data, Figs. 3b and S8). Therefore, the E/Z ratio de-
pends only on conversion, which is consistent with the pres-
ence of only one type of active site, not evolving (structurally)
with time. Thus, the conversion is proportional to the num-
ber of active sites and thereby to the intrinsic contact time,
so that reaching very low conversion should allow the intrin-
sic selectivity to be obtained (kinetic–product ratio). In fact,
when the metathesis of propene is run at a higher flow rate of
77 mL/min (261 mol/(mol Re min)), the conversion decreases
from 12 to 3% within 710 min (∼12 h), and reaches 0.03%
after about 12,000 min (200 h) (Fig. S9). At this conversion
(0.03%), the E/Z ratio of 2-butenes decreases to about 0.42,
a value probably very close to the kinetic ratio of 2-butenes
(Fig. 4).
AS(t)
Fm
ICT =
,
(6)
where AS(t) = number of active sites over time; Fm = molar
flow rate expressed in mol unit/h, but this value is difficult to
obtain for heterogeneous catalysts, because doing so requires
knowledge of the number of active sites at a given time. Note
that if only one type of active site is present and if the reac-
tion is not performed under mass transport limiting conditions,
then conversion will be directly proportional to the number of
active sites and thus can be used to monitor the ICT (vide in-
fra).
Inverse space velocity studies show that the conversion
varies linearly with respect to contact time: mass transport is
therefore not rate limiting so that chemical information can be
inferred (Figs. S4–S5). Conversely, inverse space velocity is
also proportional to the mass of the catalyst and the flow rates
on independent experiment, which further confirms that mass
transport (intragranular diffusion) is not rate-limiting (Fig. S6).

186
A. Salameh et al. / Journal of Catalysis 253 (2008) 180–190
Scheme 4. (a) Productive, (b) degenerated, and (c) isomerization metathesis processes.
The need to use such low contact time to obtain the intrin-
the isomerization rate of Z 2-butene/g of CH3ReO3/Al2O3 is
at least 15 faster than on γ -Al2O3-(500) alone, so that isomeriza-
tion occurs mainly through olefin metathesis (Scheme 4). The
kinetic ratio of 2-butenes (E/Z ratio ∼0.42) must be related
to the structure of the active sites of CH3ReO3/Al2O3. Typi-
cally, terminal olefins give the E products as favored isomers
(Scheme 5) [36], but it has been shown that in certain cases (and
sic selectivity is due to the strong adsorption of olefins on the
alumina support (ꢀH ∼80 kJ/mol) [33]. In fact, it was previ-
ously shown that desorption can be rate-determining even under
chemical regime for Re2O7/Al2O3 [34,35], so that isomeriza-
tion becomes faster than desorption; thus, the difficulty in ob-
taining the intrinsic selectivity at low contact time. Noteworthy,

A. Salameh et al. / Journal of Catalysis 253 (2008) 180–190
187
Fig. 4. Propene metathesis in a flow reactor: propene (77 mL/min; 261 mol P/(mol Re min)), P = 1 atm, and CH ReO /Al O
(60 mg) (see Fig. S3 for
3
3
2 3-(500)
conversion and selectivity as a function of time on stream). Selectivity as a function of conversion: (") E/Z 2-butene ratio, (2) E 2-butene, (Q) Z 2-butene.
Scheme 5. Ways of approach of the olefin towards the metallic center.
typically for alumina-supported catalysts) [37], the Z isomer is
preferred, and this can be associated with a favored pathway in-
volving the approach of propene toward the ethylidene ligands,
whose respective methyl substituents point away from the sur-
face.
nomenon because both would only depend on the concentration
of active sites for a constant flow of propene (of a given purity).
A satisfactory model shows that deactivation is first order in the
concentration in products or in one of the products formed, i.e.
ethene and/or butenes (Eq. (8), Fig. 5a, see Section 2 for more
details on the simulation). Ethene has been used for simulation,
but using 2-butenes leads to the same expression for the deac-
tivation rate law (see Fig. S11). This model can be improved if
one includes a second deactivation pathway, such as a first order
decay of the number of active sites (Eq. (9), Fig. 5b). Overall,
the first order deactivation in the concentration in ethene and/or
butenes requires that they are involved in this process (Eqs. (8)
and (9)).
3.4. Deactivation of CH3ReO3/Al2O3-(500) in the propene
metathesis
The deactivation of CH3ReO3/Al2O3 is relatively fast, and
modeling this phenomenon shows that this process is much
faster than a simple first order decay of the active site (e−ax
,
Eq. (7), Fig. S10) and is faster at higher than at lower conver-
sions. The major deactivation process cannot therefore be due
to either a poison in the feed or to an intrinsic deactivation phe-
dcat
dt
−
= k[cat],
(7)

188
A. Salameh et al. / Journal of Catalysis 253 (2008) 180–190
(a)
(b)
Fig. 5. Modeling deactivation of CH ReO /Al O (F) conversion at a flow rate of 77 mL/min, (") deactivation model with (a) deactivation model of Eq. (5)
3
3
2 3
4
−1
−1
4
−1
−1
ꢀ
−1
(A = 14.3; k = 5.92 × 10 mol min ) and (b) deactivation model of Eq. (6) (A = 13.8; k = 5.26 × 10 mol min ; k = 0.006 min ).
0
0
dcat
dt
In the case of Z 2-butene, the deactivation was less severe
(decrease of ca. 50% of conversion, i.e. active sites), and in
fact, the presence of 1-butene in the feed (1.8% and overall ca.
375 equiv/Re over 1188 min), could in fact be solely responsi-
ble for deactivation because it can form ethene via metathesis.
Overall, all the data show that deactivation of CH3ReO3/
Al2O3-(500) in the metathesis propene is probably mainly due
to its metathesis products, and most likely ethene, but it is not
possible to completely exclude the participation of 2-butenes or
even propene in the deactivation process. Based on these data,
several deactivation pathways can be proposed. First, the for-
mation of stable metallacyclopentane intermediates, as already
proposed for well-defined homogeneous d0 catalysts [38], is
not likely because no signal associated with this species have
been observed by solid state NMR. Second, the formation of
oligomers, as evidenced by NMR, can possibly interfere with
the active sites and be responsible for deactivation; this phe-
nomenon is observed mainly for ethene. Finally, the active site
can convert into an inactive species by reaction with an olefin,
for instance through the insertion of ethene into the Al–C bond
of [AlSCH2ReO3] in place of reacting with the methylidene lig-
and for metathesis (Scheme 6). Again, this process should be
−
−
= k[cat][=],
(8)
(9)
dcat
dt
= k[cat][=] + kꢀ[cat],
where k is the rate constant of deactivation, [cat] is the concen-
tration of active sites, and [=] is the concentration of ethene in
the feed.
To test this hypothesis, we contacted the catalyst in a flow
reactor with either ethene or Z 2-butene, the metathesis of
propene was monitored (77 mL/min), and these experiments
were compared with those from a catalytic test with only
propene (Fig. 6). The flow rates of ethene or Z 2-butene were
set to 5 mL/min (21,000 equiv/Re over 1188 min), which cor-
responds to the amount of ethene or Z 2-butene formed during
the same amount of time for the metathesis of propene carried
out at 77 mL/min (Fig. S9a). A strong deactivation of the cat-
alyst is observed with pure ethene (decrease of ca. 80% of the
conversion in propene metathesis and thereby amount of active
sites). Although ethene has been used previously to titrate the
active sites of CH3ReO3/Al2O3 with no major sign of deacti-
vation, the amount and the pressure of ethene used in titration
studies was always very small (ꢁ1 equiv/Re and ca. 0.04 bar).

A. Salameh et al. / Journal of Catalysis 253 (2008) 180–190
189
Fig. 6. Conversion as a function of time: (a) conversion of propene at a flow rate of 77 mL/min ("), (b) conversion of propene at a flow rate of 77 mL/min, after
contacting the catalyst for 1188 min with ethene at a flow rate of 5 mL/min (F) and (c) conversion of propene at a flow rate of 77 mL/min after contacting the
catalyst for 1188 min with Z 2-butene at a flow rate of 5 mL/min (2).
Scheme 6. Possible deactivation pathways.
more favored for ethene and would lead to a new Re species
that cannot form the propagating carbene species (no tautomer-
ization is possible) and which could probably decompose be-
cause longer alkyl trioxorhenium (CnH2n+1–ReO3) complexes
are usually much less stable than the parent methyl complex
(CH3 ꢂ C2H5 > Me2CH ꢂ tBu) [39].
that are likely present only in small amounts; thus the dif-
ficulty of obtaining spectroscopic evidence of such species
or even to titrate them, which has led to long-standing de-
bates in this area [42]. The presence of the carbenic species
in CH3ReO3/Al2O3, even if not observed directly but formed
in situ from a protected form, is consistent with the activity of
this system in the metathesis of functionalized olefins without
the need for a co-catalyst, in contrast to Re2O7/Al2O3. In fact,
the latter requires to generating the carbenic species in situ from
the olefin, and it is likely that ethyl oleate, a Lewis base, hin-
ders such a step [38]. Moreover, kinetic studies show that this
system is highly active (i.e., having activities similar to those of
some of the best catalysts) but undergoes fast deactivation in the
metathesis of propene, due to the propene metathesis products,
and most likely ethene. Overall, this catalyst behaves more like
a molecular system (i.e., well-defined active site, deactivation
by ethene), and we are currently investigating ways to improve
the catalytic performances of this system.
4. Conclusion
The specificity of CH3ReO3 supported on alumina com-
pared with Re2O7/Al2O3 is that it already contains the car-
benic propagating center as a protected μ-methylene between
Re and AlS in [AlSCH2ReO3]. This corresponds to the rest-
ing state of the catalyst, and the propagating carbenic species
is probably formed in situ in the presence of olefins. The active
sites, [AlSCH2ReO3], arise solely from the reaction of the C–H
bond of CH3ReO3 with reactive AlS–OS sites, whereas the in-
active major species corresponds to MeReO3 chemisorbed on
Lewis acid sites of alumina through its oxo ligand(s). Note-
worthy [AlSCH2ReO3] resembles the Tebbe reagent, one of
the first well-defined homogeneous catalysts, which contains
a carbene protected by an Al Lewis center [40,41]. Although
this study is limited to alumina, our findings suggest that other
heterogeneous systems based on MeReO3 and supported on
silica-alumina, zeolite, or Nb2O5 generate similar active species
Acknowledgments
A.S. thanks BASF for a graduate fellowship. This research
was also sponsored by the ANR for a young investigator fel-
lowship to C.C. (ANR JC05_46372). CNRS and ESCPE Lyon
provided continuous financial support.

190
A. Salameh et al. / Journal of Catalysis 253 (2008) 180–190
Supporting information
[23] M. Chabanas, C. Copéret, J.-M. Basset, Eur. J. Chem. 9 (2003) 971–975.
[24] F. Blanc, C. Copéret, J. Thivolle-Cazat, J.-M. Basset, A. Lesage, L. Em-
sley, A. Sinha, R.R. Schrock, Angew. Chem. Int. Ed. 45 (2006) 1216–
1220.
[25] B. Rhers, A. Salameh, A. Baudouin, E.A. Quadrelli, M. Taoufik, C. Co-
péret, F. Lefebvre, J.-M. Basset, X. Solans-Monfort, O. Eisenstein, W.W.
Lukens, L.P.H. Lopez, A. Sinha, R.R. Schrock, Organometallics 25
(2006) 3554–3557.
[26] (a) F. Blanc, J. Thivolle-Cazat, J.-M. Basset, C. Copéret, A.S. Hock, Z.J.
Tonzetich, R.R. Schrock, J. Am. Chem. Soc. 129 (2007) 1044–1045;
(b) F. Blanc, R. Berthoud, A. Salameh, J.-M. Basset, C. Copéret, R.
Singh, R.R. Schrock, J. Am. Chem. Soc. 129 (2007) 8434–8435.
[27] R.H.A. Bosma, G.C.N. Van den Aardweg, J.C. Mol, J. Organomet.
Chem. 255 (1983) 159–171.
Supporting information for this article may be found on Sci-
enceDirect, in the online version.
Please visit DOI: 10.1016/j.jcat.2007.10.007.
References
[1] R.L. Banks, G.C. Bailey, Ind. Eng. Chem. Prod. Res. Dev. 3 (1964) 170–
173.
[2] J.C. Mol, J. Mol. Catal. A 213 (2004) 39–45.
[3] J.C. Mol, Catal. Today 51 (1999) 289–299.
[4] J.C. Mol, Catal. Today 52 (1999) 377.
[28] Y. Chauvin, D. Commereuc, J. Chem. Soc. Chem. Commun. (1992)
462–464.
[5] J.C. Mol, E.F.G. Woerlee, J. Chem. Soc. Chem. Commun. (1979) 330–
331.
[29] A. Salameh, C. Copéret, J.-M. Basset, V.P.W. Bohm, M. Roper, Adv.
Synth. Catal. 349 (2007) 238–242.
[30] C. Flego, W.O.N. Parker, Appl. Catal. A 185 (1999) 137–152.
[31] F. Kapteijn, H.L.G. Bredt, E. Homburg, J.C. Mol, Ind. Eng. Chem. Prod.
Res. Dev. 20 (1981) 457–466.
[32] J.A. Moulijn, J.C. Mol, J. Mol. Catal. 46 (1988) 1–14.
[33] V.E. Grinev, V.A. Khalif, E.L. Aptekar, O.V. Krylov, Izv. Akad. Nauk
SSSR, Ser. Khim. (1981) 1648–1651.
[6] J.C. Mol, Top. Catal. 27 (2004) 97–104.
[7] W.A. Herrmann, W. Wagner, U.N. Flessner, U. Vokhardt, H. Komber,
Angew. Chem. Int. Ed. 30 (1991) 1636–1638.
[8] A.W. Moses, N.A. Ramsahye, C. Raab, H.D. Leifeste, S. Chattopadhyay,
B.F. Chmelka, J. Eckert, S.L. Scott, Organometallics 25 (2006) 2157–
2165.
[9] R. Buffon, A. Choplin, M. Leconte, J.M. Basset, R. Touroude, W.A. Her-
rmann, J. Mol. Catal. 72 (1992) L7–L10.
[10] A. Malek, G. Ozin, Adv. Mater. 7 (1995) 160–163.
[11] A.M.J. Rost, H. Schneider, J.P. Zoller, W.A. Herrmann, F.E. Kuehn,
J. Organomet. Chem. 690 (2005) 4712–4718.
[12] F.E. Kühn, A.M. Santos, W.A. Herrmann, Dalton Transact. (2005) 2483–
2491.
[13] A. Salameh, J. Joubert, A. Baudouin, W. Lukens, F. Delbecq, P. Sautet,
J.M. Basset, C. Copéret, Angew. Chem. Int. Ed. 46 (2007) 3870–3873.
[14] J. Joubert, A. Salameh, V. Krakoviack, F. Delbecq, P. Sautet, C. Copéret,
J.M. Basset, J. Phys. Chem. B 110 (2006) 23944–23950.
[15] R.R. Schrock, Chem. Rev. 102 (2002) 145–179.
[16] S. Hediger, B.H. Meier, N.D. Kurur, G. Bodenhausen, R.R. Ernst, Chem.
Phys. Lett. 223 (1994) 283–288.
[34] V.M. Mastikhin, I.L. Mudrakovskii, Reac. Kinet. Catal. Lett. 20 (1982)
351–355.
[35] V.M. Mastikhin, I.L. Mudrakovskii, A.G. Pel’menchikov, G.M. Zhido-
mirov, Surf. Sci. 144 (1984) 550–558.
[36] J.L. Bilhou, J.M. Basset, R. Mutin, W.F. Graydon, J. Am. Chem. Soc. 99
(1977) 4083–4090.
[37] J.M. Basset, J.L. Bilhou, R. Mutin, A. Theolier, J. Am. Chem. Soc. 97
(1975) 7376–7377.
[38] W.C.P. Tsang, J.Y. Jamieson, S.L. Aeilts, K.C. Hultzsch, R.R. Schrock,
A.H. Hoveyda, Organometallics 23 (2004) 1997–2007.
[39] W.A. Herrmann, F.E. Kühn, Acc. Chem. Res. 30 (1997) 169–180.
[40] F.N. Tebbe, G.W. Parshall, G.S. Reddy, J. Am. Chem. Soc. 100 (1978)
3611–3613.
[17] T. Weskamp, W.C. Schattenmann, M. Spiegler, W.A. Herrmann, Angew.
Chem. Int. Ed. 37 (1998) 2490–2493.
[41] F.N. Tebbe, G.W. Parshall, D.W. Ovenall, J. Am. Chem. Soc. 101 (1979)
5074–5075.
[18] T.M. Trnka, R.H. Grubbs, Acc. Chem. Res. 34 (2001) 18–29.
[19] M.B. Dinger, J.C. Mol, Adv. Synth. Catal. 344 (2002) 671–677.
[20] R.H. Grubbs, Angew. Chem. Int. Ed. 45 (2006) 3760–3765.
[21] R.R. Schrock, Angew. Chem. Int. Ed. 45 (2006) 3748–3759.
[22] M. Chabanas, A. Baudouin, C. Copéret, J.M. Basset, J. Am. Chem. Soc.
123 (2001) 2062–2063.
[42] The formation of a small amount of cross-metathesis products (5%) in
13
the reaction of ethene and C-labeled CH ReO /Nb O was reported in
3
3
2 5
a footnote of Ref. [9], which states that there are only a small amount of
“Re=CH *” and which is consistent with our current data.
2