Spectroscopically distinct sites present in methyltrioxorhenium grafted onto silica-alumina, and their abilities to initiate olefin metathesis

Article abstract of DOI:10.1021/ja072707s

Deposition of CH₃ReO₃ onto the surface of dehydrated, amorphous silica-alumina generates a highly active, supported catalyst for the metathesis of olefins. However, silica-alumina with a high (10 wt %) Re loading is no more active than silica-alumina with low (1 wt %)Joading, while CH ₃ReO₃ on silica is completely inactive. Catalysts prepared by grafting CH₃ReO₃ on silica-alumina contain two types of spectroscopically distinct sites. The more strongly bound sites are responsible for olefin metathesis activity and are formed preferentially at low Re loadings (≤1 wt %). They are created by two Lewis acid/base interactions: (1) the coordination of an oxo ligand to an Al center of the support and (2) interaction of one of the adjacent bridging oxygens (AlOSi) with the Re center. At higher Re loadings (1-10 wt%), CH₃ReO₃ also interacts with surface silanols by H-bonding. This gives rise to highly mobile sites, most of which can be observed by ¹³C solid-state NMR even without magic-angle spinning. Their formation can be prevented by capping the surface hydroxyl groups with hexamethyldisilazane prior to grafting CH ₃ReO₃, resulting in a metathesis catalyst that is more selective, more robust, and more efficient in terms of Re use.

Full text of DOI:10.1021/ja072707s

Published on Web 06/27/2007
Spectroscopically Distinct Sites Present in
Methyltrioxorhenium Grafted onto Silica-Alumina, and Their
Abilities to Initiate Olefin Metathesis
†
†
†
†
Anthony W. Moses, Christina Raab, Ryan C. Nelson, Heather D. Leifeste,
†
†
‡,§
Naseem A. Ramsahye, Swarup Chattopadhyay, Juergen Eckert,
,
†
,†,|
Bradley F. Chmelka,* and Susannah L. Scott*
Contribution from the Department of Chemical Engineering, Department of Chemistry, and
Materials Research Laboratory, UniVersity of California, Santa Barbara, California 93106 and
the Manual Lujan, Jr., Neutron Scattering Center, Los Alamos National Laboratory,
Los Alamos, New Mexico 87545
Received April 18, 2007; E-mail: bradc@engineering.ucsb.edu; sscott@engineering.ucsb.edu
Abstract: Deposition of CH
highly active, supported catalyst for the metathesis of olefins. However, silica-alumina with a high (10 wt
) Re loading is no more active than silica-alumina with low (1 wt %) loading, while CH ReO on silica
is completely inactive. Catalysts prepared by grafting CH ReO on silica-alumina contain two types of
3 3
ReO onto the surface of dehydrated, amorphous silica-alumina generates a
%
3
3
3
3
spectroscopically distinct sites. The more strongly bound sites are responsible for olefin metathesis activity
and are formed preferentially at low Re loadings (e1 wt %). They are created by two Lewis acid/base
interactions: (1) the coordination of an oxo ligand to an Al center of the support and (2) interaction of one
of the adjacent bridging oxygens (AlOSi) with the Re center. At higher Re loadings (1-10 wt %), CH
3 3
ReO
also interacts with surface silanols by H-bonding. This gives rise to highly mobile sites, most of which can
13
be observed by C solid-state NMR even without magic-angle spinning. Their formation can be prevented
by capping the surface hydroxyl groups with hexamethyldisilazane prior to grafting CH ReO , resulting in
a metathesis catalyst that is more selective, more robust, and more efficient in terms of Re use.
3
3
Introduction
while condensation with the surface hydroxyls was suggested
for CH3ReO3 on silica and silica-alumina. However, solid
oxides possess a variety of Lewis and Brønsted acidic and basic
centers that can serve as grafting sites for organometallic
catalysts. In the presence of such heterogeneity, positive
identification of the structures of the sites that become activated
in the catalytic reaction is challenging.
Solid-state NMR spectroscopy is a powerful tool for inves-
tigating supported organometallic catalysts, because the chemical
shift can be highly sensitive to variations in the ligand
environment, while the line width provides information about
site mobility. Prior to grafting, organometallic complexes can
be selectively modified by introduction of NMR-active isotope
labels to provide further insights into the nature of the surface
species. In this work, we combine solid-state NMR and IR
measurements with computational analyses to explore the
structures of the major sites formed by grafting CH3ReO3 onto
amorphous silica-alumina, and we establish some of their
relationships to the olefin metathesis activity of this material.
These insights point to simple modifications in catalyst prepara-
tion that lead to higher selectivity, less leaching, and more
efficient use of the expensive Re complex.
6
Methyltrioxorhenium (CH3ReO3) deposited on silica-alumina
is a catalyst for the metathesis of olefins.1,2 The silica-alumina
must serve as an activator as well as a catalyst support, because
solutions of CH3ReO3 do not catalyze homogeneous olefin
metathesis. The amorphous nature of the silica-alumina also
appears to be crucial, since crystalline aluminosilicates are much
3
less effective in activating CH3ReO3. Recently, we reported
experimental evidence for an early suggestion that CH3ReO3
1
is formed upon treatment of silica- and silica-alumina-supported
perrhenates with the promoter (CH3)4Sn.4
The nature of the interaction between CH3ReO3 and silica-
alumina or aluminosilicates, as well as the mechanism of
activation of the resulting supported CH3ReO3 for olefin
metathesis, has been the object of several inquiries. A hydrogen-
3
,5
bonding interaction was proposed for CH3ReO3 in zeolites,
†
Department of Chemical Engineering.
Materials Research Laboratory.
Los Alamos National Laboratory.
Department of Chemistry.
‡
§
|
(
(
(
(
(
1) Herrmann, W. A.; Kuchler, J. G.; Wagner, W.; Felixberger, J. K.;
Herdtweck, E. Angew. Chem., Int. Ed. Engl. 1988, 27, 394-396.
2) Herrmann, W. A.; Wagner, W.; Flessner, U. N.; Volkhardt, U.; Komber,
H. Angew. Chem., Int. Ed. Engl. 1991, 30, 1636-1638.
3) Bein, T.; Huber, C.; Moller, K.; Wu, C.-G.; Xu, L. Chem. Mater. 1997, 9,
Experimental Methods
2
252-2254.
4) Moses, A. W.; Leifeste, H. D.; Ramsahye, N. A.; Eckert, J.; Scott, S. L.
Catal. Org. React. 2006, 115, 13-22.
Materials. The silica-alumina, Davicat 3113 (7.6 wt % Al,
2
3
5) Malek, A.; Ozin, G. A. AdV. Mater. 1995, 7, 160-163.
BET surface area 573 m /g, pore volume 0.76 cm /g), was
8912
9
J. AM. CHEM. SOC. 2007, 129, 8912-8920
10.1021/ja072707s CCC: $37.00 © 2007 American Chemical Society

Methyltrioxorhenium Grafted onto Silica
−Alumina
A R T I C L E S
provided by Grace-Davison (Columbia, MD). The nonporous,
fumed silica Aerosil 380 (BET surface area 340 m /g, average
determined at 224 nm, using a calibration curve prepared with
NH4ReO4 (Aldrich). Based on three trials, the relatiVe standard
deviation of the Re analysis was determined to be 4.6%.
2
particle size 7 nm) was provided by Degussa (Piscataway, NY).
For each experiment, the appropriate amount of the oxide
support was first heated in a Pyrex reactor at 450 °C at 10 °C/
Kinetics and Selectivity of Propene Metathesis. A precisely
weighed amount (10.0 mg) of each catalyst was loaded into a
glass batch reactor (volume ca. 120 mL) under Ar, after which
the reactor was removed from the glovebox and evacuated. The
section of the reactor containing the catalyst was immersed in
an ice bath at 0 °C in order to control the rate of the reaction
on a readily monitored time scale, as well as to maintain
isothermal reaction conditions. Propene was introduced at the
desired pressure via a high vacuum manifold. Aliquots of 1.9
mL were expanded at timed intervals into an evacuated septum
port separated from the reactor by a stopcock. 50 µL samples
of the aliquot were removed with a gastight syringe via a
septum. Gases were analyzed by FID on a Shimadzu GC 2010
equipped with a 30 m Supelco Alumina Sulfate PLOT capillary
column (0.32 mm i.d.), using the peak area of the small propane
contaminant present in the propene as an internal standard.
-
4
min and then held at 450 °C under dynamic vacuum (e10
Torr) for a minimum of 4 h. The solid was subsequently calcined
for 12 h under 350 Torr of O2 at the same temperature to remove
hydrocarbon impurities and surface carbonates and then cooled
to room temperature under dynamic vacuum. This treatment is
sufficient to completely remove adsorbed water even from
7
highly hygroscopic silica-alumina; therefore these supports
are referred to as “dehydrated”. They were subsequently handled
under an inert atmosphere or vacuum.
Hexamethyldisilazane (HMDS, Aldrich, >99.5%) was sub-
jected to multiple freeze-pump-thaw cycles to remove dis-
solved gases and then stored under vacuum over P2O5 in a glass
reactor. To prepare trimethylsilyl-capped silica-alumina, HMDS
was transferred via the vapor phase under reduced pressure onto
the dehydrated support until there was no further uptake (as
judged by stabilization of the pressure). The reactor was then
evacuated, and the solid was heated to 350 °C for 4 h under
dynamic vacuum to remove unreacted HMDS, as well as
ammonia produced during the silanol reaction. This procedure
results in the capping of ca. 80% of the silanol sites, evaluated
Infrared Spectroscopy. The IR experiment was performed
in a Pyrex cell equipped with KCl windows affixed with
TorrSeal (Varian). Its high-vacuum ground-glass stopcock and
joints were lubricated with Apiezon H grease (Varian). A self-
supporting pellet of silica-alumina was prepared in air by
2
pressing ca. 25 mg of the solid at 40 kg/cm in a 16 mm stainless
1
by quantitative comparison of the H NMR and IR spectra of
steel die. It was then mounted in a Pyrex pellet holder.
Sublimation of CH3ReO3 via a vacuum manifold directly onto
the silica-alumina pellet in the presence of TorrSeal led to
unwanted side reactions of the latter with the organometallic
complex. Instead, the pellet was calcined and dehydrated in an
all-glass Schlenk tube, exposed to CH3ReO3 by sublimation,
and then transferred to the IR cell in a glovebox. IR spectra of
the self-supporting pellet were recorded in transmission mode
on a Shimadzu PrestigeIR spectrophotometer equipped with a
DTGS detector and purged with CO2-free dry air from a Balston
capped and uncapped silica-aluminas.
CH3ReO3 (Aldrich, >98.0%) was used as received. To
1
resolve signals in solid-state H NMR experiments and to
enhance sensitivity in solid-state 13C NMR experiments, iso-
13
topically labeled CD3ReO3 (>99% D) and CH3ReO3 (>99%
1
3
13
C) were also prepared from CD3SnBu3 and CH3SnBu3,
respectively, according to literature procedures.6 In all cases,
,8
volatile CH3ReO3/CD3ReO3 was sublimed at room temperature
-4
and 10 Torr into an all-glass reactor containing the dehydrated
support. Physisorbed material was recovered by subsequent
desorption (ca. 12 h) to a liquid N2 trap at room temperature
under vacuum. However, the desorption step was generally
omitted for samples with low Re loadings (far below the
maximum uptake by chemisorption, ca. 10 wt.% Re on Davicat
7
5-52 Purge Gas Generator. Background and sample IR spectra
-1
were recorded by co-adding 64 scans at a resolution of 4 cm .
NMR Spectroscopy. Solution-state NMR experiments were
performed at room temperature on a Bruker AVANCE 200
spectrometer with a 10-mm broadband probehead, operating at
3
113) and for the silica support (since CH3ReO3 desorbs
1
13
2
00.1012 MHz for H and 50.3202 MHz for C. Chloroform-d
completely from silica under these conditions). Where noted,
desorption from silica-alumina was performed at 80 °C, by
heating the material under vacuum and collecting volatile CH3-
ReO3 in a liquid N2-cooled receiving flask. CH3ReO3-modified
silica-alumina is yellow-brown, while CH3ReO3-modified silica
is pale yellow.
Rhenium Analysis. Re loadings were determined by quan-
titative extraction, followed by UV spectrophotometric analysis.
Approximately 15 mg of solid material was first weighed
precisely in a dry argon atmosphere. Re was extracted quanti-
tatively as perrhenate by stirring overnight in air with 5 mL of
(
99.8% D) and benzene-d6 (99.5% D) were purchased from
Cambridge Isotopes Laboratories, Inc. (Andover, MA).
Solid-state NMR experiments were performed on a Bruker
AVANCE 300 NMR spectrometer with a 4-mm broadband
MAS probehead operating at 300.1010 MHz for H and 75.4577
MHz for C. This instrument is equipped with a variable
temperature unit. The highly air-sensitive samples were packed
into zirconia MAS rotors with tightly fitting caps sealed with
Viton R O-rings (Wilmad) under an argon atmosphere in a
glovebox equipped with O2 and moisture sensors. Room-
temperature MAS spectra were acquired at a spinning rate of
1
1
3
3
M NaOH. Samples were diluted to 25 mL with 3 M H2SO4
1
2 kHz. Low-temperature MAS experiments were recorded
and filtered, and their UV spectra were recorded on a Shimadzu
UV2401PC spectrophotometer. The Re concentration was
while the sample was spinning at 8 kHz.
1H MAS experiments were performed with a 90° pulse length
(
6) Morris, L. J.; Downs, A. J.; Greene, T. M.; McGrady, G. S.; Herrmann,
of 3.7 µs, an acquisition time of 25 ms, and a recycle delay of
W. A.; Sirsch, P.; Scherer, W.; Gropen, O. Organometallics 2001, 20,
13
13
3
s. C CP-MAS spectra were recorded using a C 90° pulse
length of 3.5 µs, a contact time of 3 ms, an acquisition time of
1 ms, a recycle delay of 2 s, and two-pulse phase modulation
2
344-2352.
(7) Hunger, M.; Freude, D.; Pfeifer, H. J. Chem. Soc., Faraday Trans. 1991,
8
7, 657-662.
2
(8) Herrmann, W. A.; Kuehn, F. E.; Fischer, R. W.; Thiel, W. R.; Romao, C.
1
C. Inorg. Chem. 1992, 31, 4431-4432.
(TPPM) H decoupling using a pulse length of 6.60 µs during
J. AM. CHEM. SOC.
9
VOL. 129, NO. 28, 2007 8913

A R T I C L E S
Moses et al.
1
3
C signal detection. For single-pulse 13C MAS experiments,
9
a 13C 90° pulse length of 3.5 µs, an acquisition time of 25 ms,
and a recycle delay of 15 s were used. H and C chemical
shifts were referenced to tetramethylsilane. The same experi-
1
13
13
1
mental conditions were used in two-dimensional (2-D) C{ H}
HETeronuclear chemical shift CORrelation (HETCOR) experi-
ments. For each of the 256 t1 increments, 128 data acquisitions
were collected with a 2 s recycle delay. During the detection
period t2, 1024 points were collected with a dwell time of 20
µs. In Figure 3 and Supplementary Figures S2 and S3, the outer
contour levels are drawn at 45% of the maximum intensity. No
additional correlations were found using lower 2-D threshold
levels. The same experimental settings were used for NMR
spectra recorded sequentially under MAS/static conditions or
at room temperature/-100 °C. However, in low-temperature
experiments, dry N2 was used as both the spinning and bearing
gas, instead of air.
Figure 1. Kinetics of propene metathesis (top: disappearance of C3H6;
bottom: appearance of C2H4) in a constant-volume batch reactor containing
Computational Methods
1
0.0 mg of CH3ReO3/silica-alumina with 1.0 wt % Re (b) and 10 wt %
Computations were performed on the VRANA-5 and VRANA-8
clusters at the Center for Molecular Modeling of the National Institute
of Chemistry (Ljubljana, Slovenia), as well as Linux workstations with
dual Intel Xeon processors using the DFT implementation in the
Gaussian03 code, revision C.02. The B3PW91 density functional was
used, with a mixed basis set for the orbitals. A fully uncontracted basis
set from LANL2DZ was used for the valence electrons of Re,11
augmented by two f functions (ú ) 1.14 and 0.40) in the full
optimization. Re core electrons were treated by the Hay-Wadt
relativistic effective core potential (ECP) given by the standard LANL2
parameter set (electron-electron and nucleus-electron). For geometry
optimization of the hydrogen-bonded models in Scheme 1, the
Re (O), at 0 °C. Solid lines are nonlinear least-squares curve fits to the
integrated first-order rate equation.
shifts were included using the polarizable continuum model (PCM).
Values are summarized in Tables S1 and S2.
10
The correlations between calculated NMR shieldings ( H, 13C) and
1
observed isotropic chemical shifts are linear (Figure S1). The maximum
1
H NMR chemical shift deviation is 0.54 ppm, with an average deviation
of 0.35 ppm; the maximum 13C NMR chemical shift deviation is 4.4
ppm, with an average deviation of 2.1 ppm.
Results and Discussion
6
-31+G** basis set was used to describe the rest of the system.
Calculation of NMR Chemical Shifts. Proton and 13C NMR
The maximum amount of CH ReO that is strongly adsorbed
3
3
by dehydrated Davicat 3113 silica-alumina (i.e., the amount
-4
shieldings for computational models of the grafted CH
3
ReO
using the 6-311+G(2d,
p) basis set for all atoms other than Re and the gauge-independent
3
sites were
that resists desorption during several hours of evacuation at 10
1
2,13
calculated following a literature method,
Torr and room temperature) corresponds to (10 ( 1) wt % Re.
Materials with higher Re loadings are unstable, and the excess
CH3ReO3 desorbs rapidly upon evacuation of the reactor. These
results suggest that specific grafting sites are responsible for
the adsorption of CH3ReO3 at loadings up to 10 wt % Re.
Olefin Metathesis Activity and Selectivity. Dehydrated
silica-aluminas with a wide range of CH3ReO3 loadings (0.1-
10 wt % Re) catalyze the self-metathesis of propene. The
catalytic activities of two samples, one containing 10 wt % Re
and the other containing 1.0 wt % Re, were directly compared
under identical reaction conditions (0 °C, 45 Torr of C3H6, 10.0
mg of catalyst). Their kinetic profiles and their corresponding
nonlinear first-order curve fits are shown in Figure 1. After 1
h, the partial pressure of ethene and 2-butene isomers formed
1
13
atomic orbital (GIAO) method. Calculated H and C shielding values
were averaged for chemically equivalent atoms and then converted to
NMR chemical shifts via calibration curves constructed from a select
set of molecular organorhenium(VII) compounds. In addition to CH
ReO , [(Me CCH PhReO ] and [Me CCH ReO (dCHC(CH )] were
chosen to calibrate the calculated shieldings, based on their Re(VII)
formal oxidation state, the availability of crystallographically determined
atomic coordinates, and well-documented H and C NMR chemical
shift data. The structures of the calibration compounds were first
subjected to geometry optimization, using the 6-31G* basis set for all
atoms other than Re.
3
-
3
3
)
2 2
2
3
2
2
3 3
)
1
13
1
4
1
5,16
NMR chemical shielding values for the
calibration compounds were calculated using the method and basis set
described above, with one exception: solvent effects on NMR chemical
(
13 Torr in total) is slightly lower than the observed decrease
(
9) Bennett, A. E.; Rienstra, C. M.; Auger, M.; Lakshmi, K. V.; Griffin, R. G.
in propene partial pressure (16 Torr), due to the well-known
adsorption of propene on the silica-alumina surface.
J. Chem. Phys. 1995, 103, 6951-6958.
(
10) Frisch, M. J., et al. Gaussian 03, revision C.02; Gaussian, Inc.: Wallingford,
CT, 2004.
17
Unexpectedly, the catalyst containing 10 wt % Re was found
(
11) Pietsch, M. A.; Russo, T. V.; Murphy, R. B.; Martin, R. L.; Rappe, A. K.
Organometallics 1998, 17, 2716-2719.
to be slightly less active for the metathesis of propene (kobs )
0.11 min ) than the catalyst containing 10 times less Re (kobs
)
alumina, no olefins other than the expected metathesis products
(
equilibrium was attained, in ca. 1 h. However, in the same
reaction time, the catalyst containing only 1.0 wt % Re on
(
(
(
12) Forsyth, D. A.; Sebag, A. B. J. Am. Chem. Soc. 1997, 119, 9483-9494.
13) White, R. E.; Hanusa, T. P. Organometallics 2006, 25, 5621-5630.
14) Cai, S.; Hoffman, D. M.; Wierda, D. A. Organometallics 1996, 15, 1023-
-
1
-1
0.18 s ). For the catalyst containing 10 wt % Re on silica-
1
032.
(
15) Unreasonable NMR chemical shifts were obtained when the crystallo-
graphically determined atomic coordinates were used without geometry
optimization. Comparison of the X-ray crystallographically characterized
structures with the geometry-optimized structures revealed that most of
the structural differences are in the C-H bond distances. The positions of
the heavy atoms are affected to a very small extent by the geometry
optimization.
ethene, trans- and cis-2-butene) were detected by GC after
(
16) Schmidt, J.; Hoffmann, A.; Spiess, H. W.; Sebastiani, D. J. Phys. Chem. B
(17) MacIver, D. S.; Zabor, R. C.; Emmett, P. H. J. Phys. Chem. 1959, 63,
484-489.
2
006, 110, 23204-23210.
8
914 J. AM. CHEM. SOC. VOL. 129, NO. 28, 2007
9

Methyltrioxorhenium Grafted onto Silica
−Alumina
A R T I C L E S
Silica-alumina containing a low loading of grafted CH3ReO3
e1 wt % Re) exhibits a single, broad C CP-MAS peak with
1
3
(
a chemical shift of 29 ppm (fwhm 6.10 ppm/460 Hz), Figure
1
3
2
b (bottom). For comparison, the C chemical shift of CH3-
1
3
ReO3 in CDCl3 is 19.6 ppm, while the C CP-MAS spectrum
of polycrystalline CH3ReO3 shows a complex line shape with
maximum intensity at 34 ppm. The simple line shape in Figure
2b (bottom) suggests that the organorhenium complex is indeed
molecularly dispersed on the silica-alumina support.
According to our recent DFT study, the most favorable
grafting process involves two synergistic Lewis acid-base
1
9
interactions. One oxo ligand of CH3ReO3 binds to a Lewis
acidic Al site of the silica-alumina surface, while a neighboring
framework oxygen (AlOSi) binds to Re. This is depicted in
structure I, with an aluminosilsesquioxane monosilanol cube
Figure 2. Solid-state NMR spectra of 13CH3ReO3 adsorbed onto dehydrated
silica-alumina, at two different loadings, 0.4 wt % Re (black) and 10 wt
1
13
%
Re (red): (a) single-pulse H MAS NMR spectra; (b) C CP-MAS NMR
spectra, all acquired at room temperature using an MAS spinning rate of
2 kHz.
1
silica-alumina did produce trace amounts of side products,
including 1-butene and higher olefins. These products are typical
of the olefin isomerization and oligomerization reactions that
18
are catalyzed by the acidic hydroxyl groups of silica-alumina.
At longer reaction times, their formation eventually perturbs
the equilibrium distribution of olefins arising from propene self-
metathesis.
These results suggest that (1) different grafted CH3ReO3 sites
exist with different activities and (2) the first-deposited CH3-
ReO3 sites, formed preferentially at low Re loadings on silica-
alumina, are responsible for metathesis activity, while later-
deposited CH3ReO3 influences selectivity (by blocking the acidic
silanol sites on the silica-alumina surface responsible for side
reactions, as shown below).
representing the functional group (Lewis acid and base, Brønsted
acid) sites present on the surface of amorphous silica-alumina.
The aluminum corner is capped with a neutral siloxane ligand
to model the dominant coordination environment of the dis-
torted, four-coordinate Al sites in dehydrated silica-alumina.22
Interaction of CH3ReO3 with the cube model is accompanied
by an increase in the coordination number of the aluminum
center.
The calculated 13C and H NMR isotropic chemical shifts
for structure I, 30 and 2.2 ppm, respectively, agree reasonably
well with the observed chemical shifts, at 29 and 2.8 ppm.
Evidence for five-coordinate Al in the silica-alumina sites
Deposition of CH3ReO3 at Low Loading on Silica-
Alumina. Recently, we reported spectroscopic and computa-
tional evidence for the most energetically favored interaction
of CH3ReO3 with the surface of an amorphous, dehydrated
1
1
9
silica-alumina. For samples with low Re loadings (e1 wt
1
3 3
interacting with CH ReO was found by curve-fitting EXAFS
%
), the H MAS NMR spectrum is dominated by a signal from
1
9
13
data recorded at the Re LIII-edge. The broad C CP-MAS
NMR signal at 29 ppm that appears at low CH ReO loadings
the surface silanol protons of the silica-alumina support at 1.6
ppm,2
0,21
as shown here for a sample with 0.4 wt % Re in Figure
3
3
1
on silica-alumina is therefore assigned to surface analogues
of structure I.
2a (bottom). A quantitative H MAS NMR experiment dem-
onstrated that silanols are not consumed in this grafting process.
Deposition of Higher Loadings of CH3ReO3 on Silica-
Alumina. The H MAS spectrum of silica-alumina containing
10 wt % Re as CH ReO is shown in Figure 2a (top). The peak
3 3
At low CH3ReO3 loadings (<1 wt % Re), the methyl signal
1
1
generally cannot be resolved in the 1-D H NMR spectrum.
1
However, at higher loadings (e.g., 10 wt % Re), the H MAS
for the methyl protons, at 2.8 ppm, is clearly resolved and more
intense relative to the signal at 1.6 ppm for the surface hydroxyl
protons compared to the sample with only 0.4 wt % Re, and
consistent with the higher Re loading. In contrast, the 13C CP-
MAS spectrum is qualitatively different. Whereas at low loading
spectrum yields a partly resolved signal at 2.8 ppm (Figure 2a
(
top)), which has been assigned to the methyl protons of grafted
13
CH3ReO3 according to its correlation to the C signal of the
same site in a 2-D 1 C{ H} HETCOR experiment. For
comparison, the isotropic H chemical shift of CH3ReO3 in
3
1
19
1
(0.4 wt % Re), only one relatively broad signal is present at 29
CDCl3 is 2.64 ppm.
ppm, a narrower signal is also evident at 20 ppm at higher
loading (10 wt % Re), Figure 2b (top). The latter spectrum
demonstrates the existence of at least two chemically distinct
(
18) Pater, J. P. G.; Jacobs, P. A.; Martens, J. A. J. Catal. 1999, 184, 262-267.
19) Moses, A. W.; Ramsahye, N. A.; Raab, C.; Leifeste, H. D.; Chattopadhyay,
S.; Chmelka, B. F.; Eckert, J.; Scott, S. L. Organometallics 2006, 25, 2157-
(
2
165.
13
C environments for grafted CH3ReO3 at the higher Re loading.
(
20) Hunger, M.; Freude, D.; Pfeifer, H.; Bremer, H.; Jank, M.; Wendlandt,
K.-P. Chem. Phys. Lett. 1983, 100, 29-33.
21) Dor e´ mieux-Morin, C.; Heeribout, L.; Dumousseaux, C.; Fraissard, J.;
Hommel, H.; Legrand, A. P. J. Am. Chem. Soc. 1996, 118, 13040-13045.
(
(22) Omegna, A.; van Bokhoven, J. A.; Prins, R. J. Phys. Chem. B 2003, 107,
8854-8860.
J. AM. CHEM. SOC.
9
VOL. 129, NO. 28, 2007 8915

A R T I C L E S
Moses et al.
Figure 4. Room-temperature 13C NMR spectra for 13CH3ReO3 adsorbed
onto dehydrated silica-alumina. (a) Single-pulse 13C spectra were acquired
consecutively for the same sample (6.5 wt % Re) with the same number of
1
3
scans, under both MAS (12 kHz, red) and static conditions (black); (b)
C
CP-MAS spectra (12 kHz) acquired for the same material before (10 wt %
Re, red) and after (6 wt % Re, black) desorption of volatiles at 80 °C for
Figure 3. Two-dimensional 13C{ H} HETCOR MAS NMR spectrum for
1
6 h.
1
3
silica-alumina modified with CH3ReO3 (10 wt % Re), recorded at room
temperature with a mixing time of 3 ms and an MAS spinning rate of 12
kHz. The corresponding 1-D C CP-MAS and single-pulse H MAS NMR
spectra are plotted along their respective axes.
(Figure S2). Consequently, both grafted sites are relatively
distant from these silanols (>1 nm).
The dynamics of the more mobile of the two types of grafted
13
1
The chemical shift of the 13C CP-MAS signal at 20 ppm
species were probed with static (non-MAS) solid-state NMR
13
experiments. The room-temperature single-pulse C NMR
resembles that of CH3ReO3 in CDCl3 at 19.6 ppm. The relatively
narrow line width (fwhm 1.26 ppm/95 Hz) indicates that these
grafted sites experience more uniform environments (due to
either higher mobility or greater local order) than the sites
represented by the broad peak centered at 29 ppm (fwhm 6.10
ppm/460 Hz). Moreover, the narrow signal persists even after
room-temperature evacuation of the solid for several hours at
spectra of a sample of 3CH3ReO3 on silica-alumina (6.5 wt
1
%
Re), acquired under both MAS and static conditions, are
compared in Figure 4a. Whereas chemical shift anisotropy and
dipole-dipole interactions are averaged by rapid molecular
motion or under fast MAS conditions, they generally lead to
broad spectral lines in solid samples under static conditions.
1
3
-
4
Consequently, without MAS the C signals associated with
grafted organometallic species are often undetectable. As
expected, the peak at 29 ppm is broadened into the spectral
baseline without MAS, confirming that the species responsible
for this signal possess low mobility. The line width of the peak
at 20 ppm increases slightly to 2.65 ppm fwhm (200 Hz),
compared to 2.26 ppm (171 Hz) with 12 kHz MAS, but remains
surprisingly narrow.
e10 Torr, establishing that it cannot be attributed to phys-
13
isorbed CH3ReO3. Narrow C CP-MAS NMR lines are unusual
for supported organometallic complexes, which typically experi-
ence greatly reduced mobility when covalently bound to
heterogeneous oxide surfaces, and thus exhibit broad solid-state
2
3
NMR peaks. For example, the sites that give rise to the broad
1
3
C signal at ca. 29 ppm, attributed to CH3ReO3 immobilized
by two-point grafting onto adjacent Lewis acid/base sites, appear
to be just such heterogeneous, low-mobility sites.
This finding of high mobility was confirmed with low-
temperature NMR measurements, in which reduced molecular
mobility results in diminished motional averaging of chemical
shift anisotropy and dipole-dipole interactions. Lower mobility
is typically manifested by greatly increased line widths.
The integrated intensities of the correlated signals observed
13
1
in solid-state 2D NMR C{ H} HETCOR experiments depend
1
13
on the populations of H- C spin pairs and the strengths of
their associated dipole-dipole couplings. As shown in Figure
1
3
Nevertheless, the C signal at 20 ppm is still visible, albeit
1
3
1
3
, the 2-D C{ H} HETCOR spectrum of CH3ReO3/silica-
broadened (fwhm 13 ppm/1000 Hz), at -100 °C under static
alumina (10 wt % Re) shows intensity correlations for both of
(non-MAS) conditions, Figure 5. Thus some mobility of grafted
1
3
1
the C methyl signals at 20 and 28 ppm with the broad H
CH3ReO3 persists even at -100 °C.
24
signals centered at 2.8 ppm. Thus the methyl ligand is intact
at both grafted sites, and the mobility of each site is low enough
to produce efficient dipole-dipole couplings. No correlations
were detected between either of the C signals and the terminal,
non-hydrogen-bonded silanol H signal at 1.6 ppm, even when
Evacuation at elevated temperatures results in partial desorp-
13
tion of CH3ReO3 from the silica-alumina surface. C CP-MAS
NMR spectra recorded before and after heating a sample initially
13
-
4
containing 10 wt % Re at 80 °C and 10 Torr for 6 h are
compared in Figure 4b. Prior to heating, the relative areas of
the peaks at 29 and 20 ppm were ca. 2:1 (Figure 4b, top). Upon
1
the temperature was lowered to -100 °C and the contact time
was increased from 3 to 12 ms to probe weaker couplings
(
24) Only one 1H signal was resolved, even at high loading (10 wt% Re),
1
(
23) Fischer, H. E.; King, S. A.; Bronnimann, C. E.; Schwartz, J. Langmuir
presumably because of the smaller chemical shift range for H compared
1
3
1
993, 9, 391-393.
to C.
8916 J. AM. CHEM. SOC.
9
VOL. 129, NO. 28, 2007

Methyltrioxorhenium Grafted onto Silica
−Alumina
A R T I C L E S
Scheme 1. DFT-Calculated Structures and Energy Differences for CH3ReO3 Interacting by H-Bonding with an Aluminosilsesquioxane
Monosilanol Cube (Representing the Silica-Alumina Surface)a
a
Calculated chemical shifts for the grafted CH3ReO3 fragment are shown. Color scheme: Re (purple), Al (gray), Si (blue), O (red), C (green), H (white).
CH3ReO3 has been suggested to interact with strongly acidic
bridging hydroxyl groups in zeolites via H-bonding, on the basis
of the close resemblance in the EXAFS spectra of polycrystalline
CH3ReO3 before and after its adsorption on zeolite HY, as well
as a red shift in the frequency of a mode assigned to υ(Red
3
,5
28
O). Since silica-alumina is also strongly Brønsted acidic,
2
2,28,29
although lacking bridging hydroxyls,
we considered
hydrogen-bonding interactions between CH3ReO3 and the
terminal surface hydroxyl groups as a means of creating highly
mobile sites. It has been proposed that some of the terminal
silanols of the silica-alumina surface acquire strong Brønsted
acidity by virtue of possessing one or more Al neighbors.28
Three model structures were calculated for the interaction of
CH3ReO3 with the silanol of the aluminosilsesquioxane mono-
silanol cube, Scheme 1. Structure III has a single H-bond
between the terminal silanol and an oxo ligand of CH3ReO3,
with a calculated H-bond energy of 17 kJ/mol. In structure II,
there is a second H-bond between a methyl proton and the
silanol oxygen, similar to those calculated for a dimer of CH3-
Figure 5. 13C CP NMR spectra for CH3ReO3 grafted onto dehydrated
silica-alumina (10 wt % Re). Both spectra were recorded at -100 °C, (a)
with 8 kHz MAS and (b) without spinning.
30
ReO3, as well as for a CH3ReO3-methanol adduct. H-Bonding
involving the methyl protons is also implied in the differing
C-H bond lengths observed in the structure of polycrystalline
CH3ReO3 obtained by single-crystal neutron diffraction.31 In
structure IV, there is an additional Lewis acid/base interaction
between Re and the silanol oxygen, analogous to that proposed
thermal treatment, the 13C signal at 20 ppm disappeared,
consistent with weaker binding of the more mobile grafted CH3-
ReO3 sites (Figure 4b, bottom). At the same time, the Re content
of the solid decreased to 6 wt % Re, implying desorption of
some of the organometallic complex from the silica-alumina.
Computational Investigation of Mobile, Grafted CH3ReO3
on Silica-Alumina. Highly mobile supported organic and
organometallic molecules with narrow MAS NMR spectral lines
have been reported in cases where adsorption occurs via weak
1
7
on the basis of the O NMR spectrum of CH3ReO3 in
32
methanol and calculated for a related methanol adduct of CH3-
3
0
ReO3. The small energy differences between the model
structures in Scheme 1 and the inherent accuracy of the DFT
1
1
calculations (( several kJ/mol) make it likely that all three
structures coexist on the silica-alumina surface and that they
interconvert readily.
2
5-27
electrostatic interactions or by ion-pair formation.
The
possible protonation of CH3ReO3 was investigated by using DFT
and the same aluminosilsesquixoane monosilanol cube model
used in structure I to represent the silica-alumina surface. The
The calculated isotropic 13C chemical shifts associated with
the grafted CH3ReO3 sites in each of the structures II, III, and
IV (17.3, 16.3, and 18.5 ppm, respectively) all lie close to the
+
cation [CH3Re(O)2(OH)] was found to be highly energetic,
and the proton moved spontaneously back to the aluminosil-
sesquioxane monosilanolate anion, thus ruling out the formation
of stable ion pairs.
(
28) Cr e´ peau, G.; Montouillet, V.; Vimont, A.; Mariey, L.; Cseri, T.; Maug e´ ,
F. J. Phys. Chem. B 2006, 110, 15172-15185.
(29) Dor e´ mieux-Morin, C.; Martin, C.; Br e´ gault, J.-M.; Fraissard, J. Appl. Catal.
1991, 77, 149-161.
(
(
(
25) Scott, S. L.; Dufour, P.; Santini, C. C.; Basset, J.-M. J. Chem. Soc., Chem.
Commun. 1994, 2011-2012.
(30) Vaz, P. D.; Ribeiro-Claro, P. J. A. Eur. J. Inorg. Chem. 2005, 1836-
26) Scott, S. L.; Spakowicz, M.; Mills, A.; Santini, C. C. J. Am. Chem. Soc.
1840.
1
998, 120, 1883-1890.
(31) Herrmann, W. A.; Scherer, W.; Fischer, R. W.; Bl u¨ mel, J.; Kleine, M.;
Gruehn, R.; Mink, J.; Boysen, H.; Wilson, C. C.; Ibberson, R. M.;
Bachmann, L.; Mattner, M. J. Am. Chem. Soc. 1995, 117, 3231-3242.
27) Pan, V. H.; Tao, T.; Zhou, J.-W.; Maciel, G. E. J. Phys. Chem. B 1999,
03, 6930-6943.
1
J. AM. CHEM. SOC.
9
VOL. 129, NO. 28, 2007 8917

A R T I C L E S
Moses et al.
Figure 7. Room-temperature 13C CP-MAS NMR spectrum (12 kHz MAS)
of dehydrated silica-alumina modified first with hexamethyldisilazane,
followed by 13CH3ReO3 (1 wt % Re).
Figure 6. IR spectra of dehydrated silica-alumina before (black) and after
(
red) adsorbing CH3ReO3 (ca. 10 wt % Re), and the difference spectrum
inset).
(
methyl protons and the surface silanol protons of silica-
alumina.
experimentally observed 13C chemical shift for the relatively
mobile CH3ReO3 sites (20 ppm), as well as CH3ReO3 in CDCl3
34
The formation of H-bonded CH ReO sites on silica-alumina
3
3
(19.6 ppm). High mobility in structure III can arise by rotation
can be prevented by converting the accessible surface silanol
of the [CH3ReO2] fragment about the Re-O-H-O axis, as
well as rotation of the [CH3ReO3H] fragment about the Si-O
axis. In addition, CH3ReO3 may interchange H-bonded and non-
H-bonded oxo ligands via a hopping mechanism. Rapid inter-
conversion between structures II, III, and IV, as well as
migration between silanol sites, may also contribute to the
narrowness of the 13C NMR signal.
groups to siloxanes prior to deposition of CH ReO . Treatment
of dehydrated silica-alumina with an excess of hexamethyl-
3
3
disilazane (HMDS) at room temperature effects this transforma-
3
5
tion with concomitant generation of ammonia, eq 1:
2
tSiOH + (CH ) SiNHSi(CH ) f
3
3
3 3
2 tSiOSi(CH ) + NH (1)
3 3
3
Evidence for H-Bonded CH3ReO3 on Silica-Alumina. The
lower rate of side reactions over the catalyst with the higher
Re loading is consistent with blocking of the acidic hydroxyl
sites by adsorption of CH3ReO3, as modeled by structures II-
IV. The IR spectra of a self-supporting pellet of dehydrated
silica-alumina, recorded before and after deposition of CH3-
ReO3, provide direct evidence for this hydrogen-bonding
interaction. Grafting results in a decrease in the intensity of the
υ(SiO-H) mode at 3747 cm attributed to noninteracting
surface silanols and the appearance of a broad, red-shifted peak
characteristic of the υ(SiO-H) modes of the H-bonded silanols,
Figure 6. The changes in the υ(SiO-H) region were reversed
when the sample was heated to 80 °C for 5 h to desorb the less
strongly bound H-bonded CH3ReO3 sites; i.e., the peak assigned
to H-bonded silanols disappeared, and the peak due to non-
interacting silanols was restored to its initial intensity.
After reaction, the trimethylsilyl-capped silica-alumina was
heated to 350 °C under dynamic vacuum to completely desorb
NH3 from the Lewis acid sites, as judged by the disappearance
13
of the υ(N-H) modes by IR. Deposition of CH3ReO3 onto
13
this capped silica-alumina produced a material with two
C
NMR signals, Figure 7. The peak at -2 ppm is assigned to
CH3)3Si, while that at 30 ppm is assigned, as before, to CH3-
(
ReO3 grafted onto Lewis acid sites. No signal at 20 ppm was
observed, consistent with the absence of silanols available to
hydrogen-bond to CH3ReO3.
-
1
A catalyst prepared with trimethylsilyl-capped silica-alumina
0.5 wt % Re) and exposed to the same reaction conditions used
(
to study the activity without hexamethyldisilazane treatment (i.e.,
0
°C, 45 Torr of C3H6, 10.0 mg of catalyst) behaved similarly
-1
toward propene (kobs ) 0.13 s ). This result indicates that Lewis
acidity is both necessary and sufficient for the activation of CH3-
Hydrogen-bonding is also expected to cause a significant
ReO ; hydroxyl sites are not required. Furthermore, the catalyst
prepared with the trimethylsilyl-capped support is a much less
3
1
downfield shift in the H MAS NMR signal of the interacting
33
silanol protons. The strongly correlated signal intensity in the
active catalyst for olefin isomerization and oligomerization,
despite its low Re loading: only propene metathesis products
were detected after 60 min of reaction time at 0 °C. In contrast,
catalysts with low loadings (<1 wt % Re) that are not
trimethylsilyl-capped yield trace amounts of side products such
as 1-butene and pentenes under the same reaction condi-
tions.
3
2
2
-D HETCOR spectrum in Figure 3 between the 1 C peak at
1
0 ppm and the H peak between 2 and 3 ppm is consistent
with this interaction. Unfortunately, the chemical shifts of the
signals from H-bonded silanols overlap with those of the methyl
protons of grafted CH3ReO3. Furthermore, the use of deuterium-
labeled CD3ReO3 to eliminate the signal due to the methyl
protons is precluded by extensive H/D exchange between the
(
34) Scott, S.; Moses, A.; Leifeste, H.; Chattopadhyay, S.; Raab, C.; Chmelka,
B. F.; Ramsahye, N.; Eckert, J. Abstracts of Papers; 229th ACS National
Meeting (American Chemical Society), San Diego, CA, 2005; Abstract
No. PETR-022.
(
32) Herrmann, W. A.; K u¨ hn, F. E.; Roesky, P. W. J. Organomet. Chem. 1995,
85, 243-251.
33) Nedelec, J. M.; Hench, L. L. J. Non-Cryst. Solids 2000, 277, 106-113.
4
(
(35) Rosenthal, D. J.; White, M. G.; Park, G. D. AIChE J. 1987, 33, 336-340.
8
918 J. AM. CHEM. SOC. VOL. 129, NO. 28, 2007
9

Methyltrioxorhenium Grafted onto Silica
−Alumina
A R T I C L E S
Figure 8. Room-temperature solid-state NMR spectra of 13CH3ReO3
Figure 9. 1H MAS NMR spectra for (a) dehydrated silica and (b) CD3-
ReO3 (ca. 1 wt % Re) on dehydrated silica, both recorded at room
temperature with 12 kHz MAS.
adsorbed onto dehydrated silica (recorded after 10 min of exposure to
1
dynamic vacuum, to yield a sample with 4 wt % Re): (a) H MAS NMR
1
3
spectrum; (b) C CP-MAS NMR spectrum (12 kHz MAS).
Thus the surface hydroxyl groups are implicated in lowering
catalyst selectivity for olefin metathesis. The side reactions can
be dramatically slowed either by engaging the hydroxyls in
H-bonding interactions with CH3ReO3 or by trimethylsilyl-
capping.
1
To observe the H-bonded silanols more clearly, the H MAS
NMR spectra of silica before and after adsorption of CD3ReO3
was recorded, Figure 9. Upon H-bond formation, silanol protons
are expected to show broadened, downfield-shifted signals by
1
37
H MAS NMR, with the magnitude of the change in chemical
Deposition of CH3ReO3 onto Silica. In order to explore
hydrogen-bonding interactions between CH3ReO3 and surface
silanols in the absence of competing adsorption on Lewis acid
sites, we deposited CH3ReO3 onto amorphous silica. CH3ReO3
does not adsorb strongly on silica and can be completely
removed by simple evacuation at room temperature (10- Torr).
Furthermore, CH3ReO3 dispersed on amorphous silica (4 wt %
Re) shows no activity toward propene self-metathesis under the
reaction conditions described above.
38
shift a function of the strength of the hydrogen bond. Figure
9
a shows a sharp signal at 1.7 ppm attributed to noninteracting
silanols on silica, with a broad signal at lower field due to the
small fraction of silanols that experience H-bonding. In Figure
b, the presence of CD3ReO3 results in increased intensity in
the broad, downfield signal from 2 to 5 ppm, consistent with
H-bonding to the silanols, as well as a small signal for
incompletely deuterated methyl groups, at 2.5 ppm. The strong
correlation of the C signal for H-bonded CH3ReO3 at 20 ppm
with the broad H signal at 2.8 ppm in Figure 3 (silica-alumina)
and Figure S3 (silica) reflects the dipole-dipole coupling of
the methyl carbon with both the methyl protons and the silanol
protons in the H-bonded sites. The correlation between the
signal at 28 ppm for CH3ReO3 engaged in Lewis acid-base
bonding to the silica-alumina (Figure 3) is weaker, consistent
with dipole-dipole coupling principally between the methyl
carbon and the methyl protons.
39
9
4
13
The H MAS and 13C CP-MAS NMR spectra of CH3ReO3
dispersed by vapor deposition onto dehydrated silica are shown
in Figure 8. A brief (10 min) exposure to dynamic vacuum
1
1
1
3
yielded a material with a loading of 4 wt % Re. Its C CP-
13
C
13
MAS spectrum in Figure 8b consists of a single, sharp C peak
at 19 ppm (fwhm 1.13 ppm/85 Hz), confirming that CH3ReO3
is again molecularly dispersed. The chemical shift and line width
13
of the C NMR signal observed for CH3ReO3/SiO2 are
remarkably similar to those of the mobile CH3ReO3 species
present at high Re loading on dehydrated silica-alumina. Since
silica possesses no Lewis acidity and only weak Brønsted
Origin of Metathesis Activity. In summary, four types of
materials containing dispersed CH3ReO3 (0.1-10 wt % Re) have
been prepared and characterized, and their metathesis activities
compared, for which the spectra and kinetics presented here
are representative. One type displays C NMR signals at both
0 and 29 ppm (silica-alumina with >1 wt% Re), two types
show only the C signal at ca. 29 ppm (silica-alumina with a
e1 wt % Re or capped silica-alumina with up to 6 wt % Re),
and one type has only a C signal at 19 ppm (silica). Only
1
3
acidity, these results support the assignment of the C NMR
signal at 20 ppm for CH3ReO3 on silica-alumina to mobile
species that are hydrogen-bonded to surface silanol groups.
1
3
2
1
The H MAS spectrum in Figure 8a contains, in addition to
1
3
a signal at 1.7 ppm due to unperturbed silanol protons, a doublet
1
attributed to the methyl protons (2.6 ppm, JH-C ) 135 Hz)
1
3
36
superimposed on a broad signal. The doublet has the same
13
materials that display the C CP-MAS peak at 29 ppm are active
for propene metathesis. From this result, as well as our structural
assignment for these sites, we infer that metathesis activity
coupling constant as 13CH3ReO3 in CDCl3, and a 2-D C{ H}
13
1
13
HETCOR experiment confirmed that the C signal correlates
1
with the center of the H doublet (Figure S3). The broad signal
centered at ca. 3 ppm is attributed to H-bonded silanols.
(
37) Hu, J. Z.; Kwak, J. H.; Herrera, J. E.; Wang, Y.; Peden, C. H. F. Solid
State Nucl. Magn. Reson. 2005, 27, 200-205.
(
36) The 1H signals of grafted CH
3
ReO
3
are presumably broader on silica-
(38) Gardiennet, C.; Marica, F.; Fyfe, C. A.; Tekely, P. J. Chem. Phys. 2005,
122, 054705.
(39) Morrow, B. A. Stud. Surf. Sci. Catal. 1990, 57A, 161.
1
13
alumina than on silica since H- C coupling was not detected on silica-
1
13
alumina, even in the H MAS spectra of C-enriched samples.
J. AM. CHEM. SOC.
9
VOL. 129, NO. 28, 2007 8919

A R T I C L E S
Moses et al.
correlates with the presence of CH3ReO3 associated with Lewis
acidic Al centers; conversely, CH3ReO3 that is hydrogen-bonded
to hydroxyl groups on either silica or silica-alumina is inactive.
This explanation also accounts for the low activating ability of
silazane). Furthermore, the weaker hydrogen-bonding interac-
tions in the catalytically inactive CH3ReO3 sites are likely the
origin of the severe leaching that has been observed to
41
contaminate metathesis products. The detailed loading depen-
dence of the reaction rate and the mechanism of solid Lewis
acid activation of CH ReO are under investigation.
3
the HY zeolite, which possesses strong Brønsted acid sites but
little Lewis acidity. Reports of little or no changes in the
structure of CH3ReO3 upon grafting (as assessed by EXAFS)
are likely to be a consequence of preparing materials in which
the dominant site is the inactive, H-bonded form of CH3ReO3,
whose structure is hardly perturbed by this interaction with the
support. Furthermore, models for CH3ReO3 activation based on
3
3
Acknowledgment. This work was funded by the U.S.
Department of Energy, Basic Energy Sciences, Catalysis Science
Grant No. DE-FG02-03ER15467. The authors thank Marissa
Jaffee for Re analysis and Drs. J. D. Epping and J. Hu for helpful
discussions. We also thank Dr. Du sˇ anka Jane zˇ i cˇ and The Center
for Molecular Modeling at the National Institute of Chemistry
in Ljubljana, Slovenia for use of their computing facility. This
work made use of MRL Central Facilities supported by the
MRSEC Program of the National Science Foundation, under
Award No. DMR05-20415.
6,41
its proposed condensation with the surface hydroxyls of silica,
or mechanisms without any organometallic-support interac-
4
2
tions, are unable to account for the observed metathesis
reactivity. A correlation between metathesis activity and Lewis
4
3
acidity has been noted for CH3ReO3 supported on niobia and
on Zn-modified aluminas;40 however, no structures for the
grafted sites were proposed.
From this study, it is also clear that the preparation of
supported metathesis catalysts with high CH3ReO3 loadings is
an inefficient use of Re, despite the improvement in selectivity.
The same effect can be achieved by other methods of blocking
the Brønsted acid sites (e.g., by their reaction with hexamethyldi-
Supporting Information Available: Complete ref 10; ob-
1
13
served and calculated H and C NMR shieldings and chemical
shifts for calibration compounds and models for grafted
13
1
structures; 2-D C{ H} HETCOR spectra for CH3ReO3/silica-
alumina, recorded at -100 °C, and for CH3ReO3/silica, recorded
at room temperature; Cartesian coordinates and calculated
energies (in hartrees) for DFT model structures II, III, and IV.
This material is available free of charge via the Internet at
http://pubs.acs.org.
(
40) Rost, A. M. J.; Schneider, H.; Zoller, J. P.; Herrmann, W. A.; K u¨ hn, F. E.
J. Organomet. Chem. 2005, 690, 4712-4718.
(
41) Zhang, X.; Narancic, S.; Chen, P. Organometallics 2005, 24, 3040-3042.
42) Basset, J. M.; Buffon, R.; Choplin, A.; Leconte, M.; Tourode, R.; Herrmann,
W. A. J. Mol. Catal. 1992, 72, L7-L10.
(
(43) Oikawa, T.; Masui, Y.; Tanaka, T.; Chujo, Y.; Onaka, M. J. Organomet.
Chem. 2007, 692, 554-561.
JA072707S
8920 J. AM. CHEM. SOC.
9
VOL. 129, NO. 28, 2007