Molecular structure-activity relationships for the oxidation of organic compounds using mesoporous silica catalysts derivatised with bis(halogeno)dioxomolybdenum(VI) complexes

Article abstract of DOI:10.1002/chem.200304887

Mo K-edge XAFS spectra have been measured for ordered mesoporous silica MCM-41 grafted with the complexes [MoO₂X₂(thf) ₂] (X = Cl, Br). For grafting reactions in the absence of triethylamine, materials with 1 wt.% Mo are obtained; the Mo K-edge EXAFS results indicate the co-existence of isolated surface-fixed monomeric species {MoO₂[(-O)₃SiO]₂(thf)_n} and {MoO₂[(-O)₃SiO]X(thf)_n}. When Et₃N is used in the grafting reactions, materials with 4 wt. % Mo are obtained. The EXAFS data for the material prepared using [MoO₂Cl ₂(thf)₂] and Et₃N indicate the presence of dinuclear species with two Mo^VI centres, each with two Mo=O groups and each linked by one or two oxo bridges (Mo...Mo 3.27 A). The molybdenum centres in the material prepared using the dibromo complex comprise mainly isolated four-coordinate dioxomolybdenum(VI) and trioxomolybdenum(VI) monomeric species, with a small contribution from dimeric acterised in the solid state by powder X-ray diffraction, N₂ adsorption analysis, MAS NMR (¹³C, ²⁹Si) and FTIR spectroscopy. The derivatised MCMs perform differently as catalysts in the liquid-phase oxidation of various olefins and alcohols with tert-butyl hydroperoxide. The highest alkene epoxidation activity was recorded for the catalysts with low metal loading, whereas the material containing oxo-bridged dimers had the highest activity for oxidation of alcohols. The recyclability of all the catalysts was tested: the catalytic activity of the derivatised materials tended to stabilize with ageing.

Full text of DOI:10.1002/chem.200304887

FULL PAPER
Molecular Structure Activity Relationships for the Oxidation of
Organic Compounds Using Mesoporous Silica Catalysts Derivatised
with Bis(halogeno)dioxomolybdenum(vi) Complexes
Carla D. Nunes, Anabela A. Valente, Martyn Pillinger, Joao Rocha, and
ƒ
Isabel S. GonÁalves*[a]
ˆ
Abstract: Mo K-edge XAFS spectra
have been measured for ordered meso-
porous silica MCM-41 grafted with the
complexes [MoO2X2(thf)2] (X ˆ Cl, Br).
For grafting reactions in the absence of
triethylamine, materials with 1 wt.%
Mo are obtained; the Mo K-edge EX-
AFS results indicate the co-existence of
isolated surface-fixed monomeric spe-
À
two MoVI centres, each with two Mo O
acterised in the solid state by powder
X-ray diffraction, N2 adsorption analy-
sis, MAS NMR (13C, 29Si) and FTIR
spectroscopy. The derivatised MCMs
perform differently as catalysts in the
liquid-phase oxidation of various olefins
and alcohols with tert-butyl hydroperox-
ide. The highest alkene epoxidation
activity was recorded for the catalysts
with low metal loading, whereas the
material containing oxo-bridged dimers
had the highest activity for oxidation of
alcohols. The recyclability of all the
catalysts was tested: the catalytic activ-
ity of the derivatised materials tended to
stabilize with ageing.
groups and each linked by one or two
oxo bridges (Mo ¥¥¥ Mo 3.27 ä). The
molybdenum centres in the material
prepared using the dibromo complex
comprise mainly isolated four-coordi-
nate dioxomolybdenum(vi) and trioxo-
molybdenum(vi) monomeric species,
with a small contribution from dimeric
species. All materials were further char-
cies
{MoO2[( O)3SiO]2(thf)n}
À
and
{MoO2[( O)3SiO]X(thf)n}. When Et3N
is used in the grafting reactions, materi-
als with 4 wt.% Mo are obtained. The
EXAFS data for the material prepared
using [MoO2Cl2(thf)2] and Et3N indicate
the presence of dinuclear species with
Keywords: EXAFS spectroscopy ¥
heterogeneous catalysis
¥
meso-
porous materials ¥ molybdenum ¥
oxidation
surfaces of these materials contain reactive silanol groups that
can be used to prepare derivatised materials.[8] Functionalisa-
tion of ordered mesoporous silicas with transition-metal
oxides has been of particular interest. One approach is to
graft organometallic or metal-organic complexes onto the
support; this is followed by calcination to produce isolated
oxo metal species. Suitable precursors include metal alk-
oxides[9] or metallocenes.[10] The preparation of isolated MoVI
active sites on ordered mesoporous silica is especially
interesting, given that silica-supported molybdenum oxides
are widely used for a variety of selective oxidative dehydro-
genations such as the oxidation of methane to methanol and
formaldehyde.[11] Some of us recently reported a convenient
one-step procedure for the preparation of MoVI active sites on
MCM-41 and MCM-48,[12] involving direct grafting of com-
plexes [MoO2X2(thf)2] (X ˆ Cl, Br) onto the purely siliceous
supports using dichloromethane as solvent. A subsequent
calcination step was not necessary. Preliminary tests showed
that the grafted materials were active as catalysts for the
epoxidation of cyclooctene with tert-butyl hydroperoxide
(TBHP) as the oxidant. It was assumed that the active sites
Introduction
Silica and alumina substrates have been used for decades to
support transition-metal oxides for catalytic applications.[1]
These supports are attractive for their high specific surface
area, mechanical stability and promotion of well-dispersed
active sites.[2] In the last ten years or so, micelle-templated
inorganic oxides have emerged as promising alternatives[3]
mainly because, in addition to the characteristics mentioned
above, they exhibit an ordered array of uniform mesopores
that permit catalytic reactions in constrained environments
(for example, those involving bulky substrate and/or product
molecules).[4] Most of the work has been with silica-based
materials, in particular hexagonal mesoporous silicas such as
MCM-41,[5] SBA-3[6] and FSM-16,[7] and cubic mesoporous
silicas such as MCM-48[5] and SBA-1.[6] Like those of the
commonly used silica and alumina supports, the internal
[a] Dr. I. S. GonÁalves, C. D. Nunes, Dr. A. A. Valente,
Dr. M. Pillinger, Prof. Dr. J. Rocha
Department of Chemistry, CICECO, University of Aveiro
Campus de Santiago, 3810-193 Aveiro (Portugal)
Fax : (‡351)234-370084
À
were surface-fixed monomeric species {MoO2[( O)3-
SiO]X(thf)n}, although we had no direct spectroscopic evi-
E-mail: igoncalves@dq.ua.pt
4380
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/chem.200304887
Chem. Eur. J. 2003, 9, 4380 4390

4380 4390
dence to support this conclusion (the materials were charac-
terised by elemental analysis, powder X-ray diffraction
(XRD), N2 adsorption, and magic-angle spinning (MAS)
NMR spectroscopy). In the present work, Mo K-edge X-ray
absorption fine structure (XAFS) spectroscopy and IR
spectroscopy have been used to probe the local structure of
supported Mo species in MCM-41 grafted with the bis(halo-
geno)dioxomolybdenum(vi) complexes. We have also studied
the effect of using triethylamine in the grafting reactions to
activate the surface silanols and to sequester the HCl
liberated, and we report in detail on the activities of the
derivatised materials as catalysts for liquid-phase epoxidation
of cyclooctene with TBHP as the oxygen source, and the
oxidation of alcohols. The differences in the observed
catalytic behaviour are correlated with the variability in the
structures of the supported oxomolybdenum catalysts.
Figure 1. Powder XRD patterns of a) MCM-41, b) MCM-41/1 and
c) MCM-41/1/Et3N. Similar results were obtained for MCM-41/2 and
MCM-41/2/Et3N.
Results and Discussion
Textural characterisation: The MCM-41 sample was of higher
quality than the one used previously for grafting reactions
with complexes [MoO2X2(thf)2] (X ˆ Cl (1), Br (2)).[12] Five
reflections that were observed in the powder X-ray diffraction
pattern (Figure 1) were indexed, assuming a hexagonal cell, as
(100), (110), (200), (210) and (300). The d value of the (100)
reflection is 35.7 ä, giving a lattice constant of a ˆ 41.2 ä.
Treatment of calcined and dehydrated MCM-41 with a
dichloromethane solution of 1 or 2 (in excess) gave deriva-
tised materials with a metal loading of about 1 wt.%
(0.1 mmolgÀ1). This is the maximum that can be achieved in
the absence of agents to activate the surface silanol groups.
The loading corresponds to a surface coverage of 0.1 atom-
snmÀ2 (assuming the metal atom layer thickness to be
negligible and an average support surface area of
800 m2 gÀ1). For comparison, the approximate abundance of
compared with those of MCM-41/1 or MCM-41/2, presumably
because of the higher metal loading.
The N2 adsorption desorption isotherms for the pristine
and modified MCM-41 samples were type IV as defined by
IUPAC, characteristic of mesoporous solids. All the isotherms
were reversible, with clear capillary condensation steps
indicating that the textural characteristics of the silica support
are preserved during the grafting experiments and that the
channels remain accessible. Figures 2 and 3 show isotherms
for MCM-41 before and after derivatisation of the silica
surface, with and without triethylamine treatment. The
modifications led to a decrease in the limiting N2 uptake at
high p/p0, and both the surface area and pore volume
decreased (Table 1). The plot of differential volume as a
function of pore width shows a relatively narrow pore size
distribution (PSD), with median pore diameters ranging from
2.9 to 3.2 nm for pristine MCM-41 and 2.6 to 3.2 nm for the
modified samples. The grafting procedure without Et3N
caused an increase in the width at half-height, indicating an
increase in pore heterogeneity (Figure 2). For the catalysts
prepared using Et3N, a shift of the maximum of the PSD curve
from 2.9 to 2.6 nm was observed upon surface modification
(Figure 3).
silanol groups on calcined MCM-41 has been reported to be in
[13]
the range 1 3 nmÀ2
;
of these about 25% are free silanols
(involving single and geminal groups), and the rest are
hydrogen-bonded.[13a] Zhao et al. showed that only the free
silanol groups on MCM-41 are readily accessible to the
silylating agent, chlorotrimethylsilane.[13a] It can be assumed
therefore that the molybdenum complexes react only with
free silanol groups. The powder XRD patterns for the grafted
materials MCM-41/1 and MCM-41/2 were quite similar to that
of pristine calcined MCM-41 (Figure 1), which indicates
retention of the long-range hexagonal symmetry. There was
a slight attenuation of XRD peak intensities. This is not
interpreted as a loss of crystallinity. Instead, it is likely that
there is a reduction in the X-ray scattering contrast between
the silica wall and pore-filling material.[14] When triethylamine
(Et3N) was used in the grafting reactions, materials with a
much higher metal loading were obtained (approximately
4 wt.% Mo, 0.4 mmolgÀ1), corresponding to a surface cover-
age of 0.3 atomsnmÀ2 (assuming an average support surface
area of 800 m2 gÀ1). At least four reflections were observed in
the powder XRD patterns of MCM-41/1/Et3N and MCM-41/
2/Et3N. The peak intensities were reduced considerably
Spectroscopic characterisation: Figure 4 shows the 29Si MAS
and CP-MAS NMR spectra of pristine MCM-41, MCM-41/1
and MCM-41/1/Et3N. The 29Si MAS spectrum of the starting
material exhibits two broad overlapping peaks at d ˆ À102.0
and À109.0 ppm assigned to, respectively, Q3 and Q4 units of
the silica framework (Qn ˆ Si(OSi)n(OH)4Àn). A small pro-
portion of Q2 environments is also present (there is a faint
peak at d ˆ À92.5 ppm). The 29Si CP-MAS NMR spectrum
shows a marked increase in the relative intensity of the Q2 and
Q3 lines in comparison with the 29Si MAS spectrum, confirm-
ing that these silicons are attached to hydroxyl groups. In
agreement with those previously reported,[12] both the 29Si CP-
MAS and MAS NMR spectra show that upon functionalisa-
tion with 1 in the absence of Et3N, the relative intensity of the
Chem. Eur. J. 2003, 9, 4380 4390
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4381

I. S. GonÁalves et al.
FULL PAPER
*
Figure 3. Nitrogen adsorption (
)
desorption (Â) isotherms at 77 K of
*
Figure 2. Nitrogen adsorption (
)
desorption (Â) isotherms at 77 K of
a) the parent MCM-41 and b) MCM-41/1/Et3N, and the corresponding pore
a) the parent MCM-41 and b) MCM-41/1, and the corresponding pore size
size distribution (PSD) profiles.
distribution (PSD) profiles.
Q3 (and Q2) peak(s) changes only slightly, consistent with the
low metal loading (approximately 1 wt.% Mo). For the
material prepared in the presence of triethylamine, the 29Si
CP-MAS NMR spectrum clearly shows a reduction in the Q3
and Q2 resonances, and a concomitant increase in the Q4
resonance. This indicates an extensive interaction of surface
silanols with triethylamine and/or the grafted molybdenum
species (consistent with the higher metal loading). The
materials MCM-41/2 and MCM-41/2/Et3N exhibited similar
29Si NMR spectra to those of MCM-41/1 and MCM-41/1/Et3N,
respectively (not shown). In the 13C CP-MAS NMR spectra of
MCM-41/1 and MCM-41/2 (not shown), lines observed at d ˆ
24.7 and 68.0 ppm are assigned to coordinated THF mole-
cules. By contrast, 13C CP-MAS NMR showed that the grafted
materials MCM-41/1/Et3N and MCM-41/2/Et3N do not con-
tain THF, but do contain triethylammonium ions (lines at 7.6
and 46.6 ppm).
Table 1. Textural parameters of MCM-41 samples determined from N2 isotherms
at 77 K.
[b]
Sample
SBET [m2 gÀ1
]
DSBET[a] [%] VP [cm3 gÀ1
]
DVP [%] dP [nm]
MCM-41-1
MCM-41/1
MCM-41/2
MCM-41-2
MCM-41/1/Et3N 760
MCM-41/2/Et3N 652
928
734
855
863
0.69
3.2
À 21
À 8
0.64
0.67
0.58
0.55
0.46
À 7
À 3
3.2
3.2
2.9
2.6
2.6
À 12
À 24
À 5
À 21
[a] Variation of surface area in relation to parent MCM-41. [b] Variation of total
pore volume in relation to parent MCM-41.
1087 cmÀ1, with a weak peak at 1180 cmÀ1. A strong peak at
800 cmÀ1 is attributed to nsym of the Si O Si bonds while a
broad band at 570 cmÀ1 is ascribed to the ™crystallinity peak∫
in zeolite synthesis, indicative of the formation of an ordered
network solid.[9f] An absorption at 963 cmÀ1 is assigned to the
Si
O
dÀ of a Si OH stretch. When MCM-41 was grafted with
The precursor MCM-41 samples and derivatised materials
were also characterised by FTIR spectroscopy. Pristine
calcined MCM-41 presented a spectrum comparable with
that of MCM-48.[9f] Two framework bands for the asymmetric
stretch (nas) of the Si O Si bonds are observed at 1235 and
1 or 2 in the absence of Et3N, these bands due to the support
did not change significantly in appearance or frequency.
Several new bands were observed; they are assigned to the
grafted species and/or their interaction with the host. The
4382
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 4380 4390

Oxidation of Organic Compounds
4380 4390
Figure 5. Mo K-edge XANES spectra of a) [MoO2Cl2(thf)2] (1),
b) [MoO2Br2(thf)2] (2), c) MCM-41/1, d) MCM-41/2, e) MCM-41/1/Et3N,
f) MCM-41/2/Et3N and g) Na2MoO4 ¥ 2H2O.
Figure 4. 29Si MAS and CP-MAS NMR spectra of a) MCM-41, b) MCM-
41/1 and c) MCM-41/1/Et3N.
is usually strongest for molybdenum coordination systems
with non-inversion symmetry, such as tetrahedral geometry.[15]
Sodium molybdate, for example, exhibits an intense well-
resolved pre-edge peak (Figure 5). Any distortion of the
regular tetrahedral symmetry leads to a decrease in the peak
intensity. The XANES spectra of complexes 1 and 2 are
therefore consistent with the expected molybdenum coordi-
nation, that is, distorted octahedral geometry. The clear
differences in both amplitude and frequency between the
XANES oscillations of the precursor complexes and deriva-
tised materials demonstrate that the grafting process has
imposed changes in the local structural order around the
molybdenum centres. The increase in the intensity of the pre-
edge peak for the grafted materials indicates a change in the
molybdenum coordination geometry. Enhancement of this
transition has also been associated with an increasing number
presence of THF is indicated by bands at 2890 and 2980 cmÀ1.
A weak shoulder at 918 cmÀ1 may be due to the asymmetric
ˆ
stretch of the C O C bonds of THF or to a Mo Ostretching
vibration (the latter usually occurs at ꢀ900 cmÀ1 for cis-
dioxomolybdenum(vi) complexes). A second shoulder at
about 890 cmÀ1 is assigned to a Mo Ostretching vibration.
ˆ
ˆ
For comparison, the Mo Ostretching vibrations for the
precursor complex 2 appear at 904 cmÀ1 (asymmetric) and
942 cmÀ1 (symmetric). The FTIR spectra of MCM-41/1/Et3N
and MCM-41/2/Et3N show bands due to triethylammonium
ions (1398, 1475, 2492, 2690, 2726, 2790 and 2990 cmÀ1). The
absorption at 965 cmÀ1 due to the Si
O
dÀ of a Si OH stretch
is reduced to a weak intensity and a new band appears at
945 cmÀ1. In agreement with the 29Si NMR results, this
indicates a strong interaction of a large fraction of the surface
silanols with triethylamine and/or oxomolybdenum species.
Both samples also exhibit a shoulder at 918 cmÀ1 that is
[16]
ˆ
of Mo Obonds.
Good quality EXAFS spectra were obtained up to 16
20äÀ1 for all the compounds examined. Analysis of the
EXAFS for 1 and 2 gave the expected first coordination
spheres of two oxygen atoms at 1.69 1.70 ä (Mo Ot),
approximately two oxygen atoms at 2.24 2.26 ä (Mo
assigned to a Mo Ostretching vibration (the 13C CP-MAS
ˆ
NMR spectra showed that the samples do not contain THF).
The grafted material MCM-41/1/Et3N (but not MCM-41/2/
Et3N) exhibits a prominent shoulder at 885 cmÀ1. Considering
the Mo K-edge EXAFS data (described below), this band is
tentatively assigned to a Mo O Mo stretching vibration
(usually expected in the range 700 900 cmÀ1).
OTHF) and two halogen atoms at either 2.36 ä (Cl) or 2.52 ä
(Br) (Figures 6 and 7; Table 2). The Mo Ot and Mo
X
distances are in agreement with those obtained by single-
crystal X-ray diffraction for Lewis base adducts [MoO2X2L2]
(L2 ˆ 2,2'-bipyridine, 2,2'-bipyrimidine).[18] If 1 was grafted
into MCM-41 (MCM-41/1), a substantially different model
was required for a reasonable fit to the Mo K-edge EXAFS,
comprising oxygen shells at 1.70, 1.93 and 2.24 ä (Figure 6b).
The first shell is extremely well defined and confirms the
presence of dioxo species. The appearance of the new shell at
1.93 ä, coupled with the fact that no acceptable fit could be
obtained for a chlorine shell at approximately 2.35 ä, is
consistent with the formation of Mo O Si linkages (by
reaction of the complex with nucleophilic surface silanol
groups) (Figure 8). The refined coordination number for the
EXAFS spectroscopy is perhaps the best technique to
probe the molecular structure of the molybdenum centres in
the grafted materials. However, where structural diversity of
metal species arises, the traditional approach to modelling the
resultant EXAFS spectrum can give only an indication of the
average structural environment. Mo K-edge X-ray absorption
spectra were collected at room temperature in the solid state
for the precursor complexes 1 and 2, and the four derivatised
materials. The XANES data are shown in Figure 5. The
spectra for the precursor complexes are quite similar, both
exhibiting a rather weak pre-edge feature that is attributed to
a bound state transition (1s ! 4d). The intensity of this peak
Chem. Eur. J. 2003, 9, 4380 4390
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4383

I. S. GonÁalves et al.
FULL PAPER
Figure 6. Room-temperature Mo K-edge EXAFS and Fourier transforms
of a) 1, b) MCM-41/1 and c) MCM-41/1/Et3N. The solid lines represent the
experimental data and the broken lines show the best fits using the
parameters given in Table 2.
Figure 7. Room-temperature Mo K-edge EXAFS and Fourier transforms
of a) 2, b) MCM-41/2 and c) MCM-41/2/Et3N. The solid lines represent the
experimental data and the broken lines show the best fits using the
parameters given in Table 2.
Table 2. Mo K-edge EXAFS-derived structural parameters for 1, 2 and deriva-
tised MCM-41 materials.
second oxygen shell is close to 1, lower than the expected
value of 2 for a bipodally anchored species. Static disorder in
the Mo Ob bonds may account for this discrepancy. Refine-
ment of coordination numbers and Debye Waller factors
simultaneously does not always give reliable results, owing to
the high correlation of the two parameters. Attempts to model
a Mo ¥¥¥ Si distance (in the range 3 4 ä) corresponding to a
Mo O Si linkage were unsuccessful. This may again be due
to disorder in the Mo O Si interactions. There are still THF
molecules coordinated to the MoVI centre, as evidenced by the
shell at 2.24 ä (in agreement with the 13C CP-MAS NMR
results). In the case of MCM-41 grafted with the dibromo
complex 2 (MCM-41/2), three oxygen shells were again in
evidence, corresponding to Mo Ot, Mo Ob and Mo OTHF
bonds. A fourth shell for bromine atoms was also included in
the final fit (coordination numberˆ 0.6). The EXAFS results
for the grafted materials therefore indicate the coexistence of
Compound
Atom CN[a]
r [ä]
2s2[b] [ä2] Ef[c] [eV] R[d] [%]
1
O2.0(1) 1.692(2)
0.0023(2)
À 1.2(7) 26.2
O2.5(7) 2.242(13) 0.0257(51)
Cl
1.8(2) 2.355(2)
0.0053(3)
0.0022(2)
0.0114(16)
1.8(1) 2.515(2) 0.0055(2)
2
O2.1(1) 1.696(2)
O2.2(3) 2.262(7)
À 3.4(6) 18.0
Br
MCM-41/1
MCM-41/2
O2.0(1) 1.695(2)
O1.0(2) 1.927(7)
O1.4(4) 2.240(15) 0.0210(49)
O2.0(1) 1.689(2)
O1.0(3) 1.929(6)
0.0044(3)
4.3(7) 31.2
0.0076(14)
0.0036(2)
0.0061(10)
1.5(7) 28.6
5.4(4) 25.9
4.1(4) 25.8
O2.0(5) 2.275(13) 0.0262(47)
Br 0.6(1) 2.526(5) 0.0084(7)
0.0048(2)
0.0101(10)
0.0083(3)
O0.8(2) 2.310(14) 0.0127(36)
MCM-41/1/Et3N O2.0(1) 1.695(2)
O1.5(2) 1.924(4)
Mo
1.4(2) 3.274(3)
MCM-41/2/Et3N O2.5(1) 1.714(2)
O1.0(1) 1.940(4)
Mo
0.0043(2)
0.0043(7)
isolated
surface-fixed
monomeric
species
{MoO2-
0.3(2) 3.255(14) 0.0093(23)
À
À
[( O)3SiO]2(thf)n} (site A) and {MoO2[( O)3SiO]X(thf)n}
(site B) (Figure 8). The relative population of monopodally
anchored species (B) seems to be higher for the grafted
material MCM-41/2, consistent with the greater lability of the
Cl ligands than the Br ligands in the precursor complexes 1
and 2.
[a] CN ˆ Coordination number. Values in parentheses are statistical errors
generated in EXCURVE. The true errors in coordination numbers are likely to
be of the order of 20%, and those for the interatomic distances ꢁ1.5%.[17]
[b] Debye Waller factor; s ˆ root-mean-square internuclear separation.
[c] Ef ˆ edge position (Fermi energy), relative to calculated vacuum zero.
„
[d] R ˆ ( [Stheory À Sexptl]k3dk/[Sexptl]k3dk)  100%.
4384
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 4380 4390

Oxidation of Organic Compounds
The inclusion of triethyla-
4380 4390
the Mo ¥¥¥ Mo shell is clearly higher than the ideal value of 1
for a dimeric species. It was thought that this might be due to
multiple scattering involving the Mo Ob Mo unit, despite
the rather acute Mo O Mo angle (approximately 117 8
assuming equal Mo Ob bond lengths of 1.92 ä). However,
inclusion of multiple scattering in the calculations for the
Mo Ob Mo unit (with the coordination numbers fixed at
1.0) resulted in a worse fit to the EXAFS. The initial single
scattering approach was therefore retained. The FT of the
room-temperature Mo K-EXAFS of MCM-41/1/Et3N con-
tains a weak peak above 4.5 ä. This was best fitted as
about one molybdenum atom at 4.70 ä. The existence of this
shell would suggest a contribution from polymeric species
in addition to the dimeric species described above (possibly
explaining the calculated coordination number of 1.4
for the Mo ¥¥¥ Mo shell). However, the shell is not included
in the final four-shell model because the improvement
in the fit was modest (R ˆ 25.2%) and there were un-
acceptably large statistical errors in the refined struc-
tural parameters (90% in the coordination number, for
example).
The Mo K-edge EXAFS results for the grafted material
MCM-41/2/Et3N show that the surface chemistry of the
dibromo complex is very different from that of the dichloro
complex when triethylamine is present to activate the surface
silanols. These differences are very evident in both the
EXAFS spectra and Fourier transforms (Figure 7). For
example, the FT of the Mo K-edge EXAFS of MCM-41/2/
Et3N contains only a weak peak at about 3.3 ä, compared
with the very intense peak observed for the grafted material
MCM-41/1/Et3N. The EXAFS data for MCM-41/2/Et3N were
initially modelled by oxygen shells at 1.71 ä and 1.94 ä (R
factorˆ 27.1%). No acceptable fit could be obtained for a
bromine shell at approximately 2.5 ä. A modest improve-
ment in the fit was obtained by addition of a shell for
molybdenum atoms (CN ˆ 0.3) at 3.26 ä (R factorˆ 25.8%;
Table 2). The contribution of this shell to the overall fit is
clearly much less than that observed for the corresponding
shell in the fit to the EXAFS of MCM-41/1/Et3N. In fact, a
reasonable fit could also be obtained by modelling about one
silicon atom at 3.33 ä instead of the molybdenum shell. These
results indicate that the surface species in MCM-41/2/Et3N
comprise mainly isolated monomers, with perhaps a small
contribution from dimers. The EXAFS-derived bond distance
and coordination number for the terminal oxo ligands (1.71 ä,
2.5) are significant, since the other grafted materials and
precursor complexes 1 and 2 yielded consistent values of
1.695 Æ 0.005 ä and 2.0 Æ 0.1. The average coordination
number of 2.5 suggests the presence of some trioxomolybde-
num(vi) species. Anions of the type [XMoO3]À are known to
exist (in aqueous solution, for example), but very few
complexes have been isolated and characterised.[21 23] Arzou-
manian et al. obtained crystals of the complex [PPh4][Mo-
(h1-1,3,5-C6H2Me3)(O3)] suitable for structure determination
by X-ray diffraction.[24] The Mo Obond lengths were found
to be 1.720 Æ 0.003 ä. Figure 10 illustrates a possible mech-
anism for the formation of trioxomolybdenum(vi) species on
the surface of MCM-41, starting from an immobilised species
mine in the grafting reaction
of [MoO2Cl2(thf)2] with MCM-
41 results in an EXAFS spec-
trum that is very different from
that of MCM-41/1 (Figure 6c).
An intense new peak is present
in the Fourier transform (FT)
above 3 ä. This was best mod-
elled as 1.4 molybdenum atoms
at 3.27 ä. In the final fit, 1.69,
1.92 and 2.31 ä Mo Odistan-
ces were also included (Ta-
ble 2). The goodness-of-fit R
factor decreased from 27.6 to
25.9% upon addition of the
oxygen shell at 2.31 ä. A sim-
ilar improvement in the fit was
obtained by fitting one nitrogen
at 2.34 ä instead of the oxygen
shell at 2.31 ä. Therefore the
presence of coordinated trie-
thylamine molecules cannot be
ruled out. No evidence was
found for a shell at approxi-
Figure 8. Representation
of
the predominant species pres-
ent in MCM-41 derivatised with
the complexes 1 and 2 in the
absence of Et3N, as evidenced
by Mo K-edge XAFS spectro-
scopy.
mately 2.25 ä for coordinated THF molecules. The results
indicate that the Mo centres formed in MCM-41/1/Et3N are
largely dinuclear species with two MoVI centres, each with two
ˆ
Mo Ogroups and each linked by one or two oxo bridges
(Figure 9). Mono-oxo-bridged [Mo2O5]2‡ entities, for exam-
ple, oxalate complexes M2[Mo2O5(C2O4)2(H2O)2] (M ˆ Na, K,
Rb, Cs), are quite common in oxomolybdenum(vi) chemis-
try.[19] Dioxo-bridged fragments [Mo2O6] with Mo ¥¥¥ Mo
3.27 ä have been structurally characterised in catalysts
prepared by immobilisation of species derived from [MoO2-
(acac)2] on a polystyrene N-hydroxypropyl aminomethylpyr-
idine resin.[20] The EXAFS-derived coordination number for
Figure 9. Representation of the predominant species present in MCM-41
derivatised with the complexes 1 and 2 in the presence of Et3N, as
evidenced by Mo K-edge XAFS spectroscopy.
À
{MoO2[( O)3SiO]2}.
Chem. Eur. J. 2003, 9, 4380 4390
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4385

I. S. GonÁalves et al.
FULL PAPER
Table 3. Cyclooctene epoxidation using TBHP in the presence of MCM-41
materials.
Sample
TOF[a]
[mmolgÀ1
Conv.[b] [%] Selectivity to
h ] cyclooctene
À1
cat
oxide [%]
MCM-41/1 (1st run)
MCM-41/1 (2nd run)
MCM-41/1 (3rd run)
MCM-41/2 (1st run)
MCM-41/2 (2nd run)
MCM-41/2 (3rd run)
MCM-41/1/Et3N (1st run)
MCM-41/1/Et3N (2nd run)
MCM-41/1/Et3N (3rd run)
MCM-41/2/Et3N (1st run)
MCM-41/2/Et3N (2nd run)
MCM-41/2/Et3N (3rd run)
38
6
5
32
6
5
5
1
1
1
100
86
76
100
84
72
56
34
29
49
27
23
100
100
100
100
100
100
80[c]
94[c]
93[c]
100
100
100
Figure 10. Possible mechanism for the formation of surface-fixed trioxo-
molybdenum(vi) species in the grafted material MCM-41/2/Et3N.
Catalytic epoxidation of olefins: The behaviour of the
derivatised MCM-41 materials as catalysts or catalyst pre-
cursors for the liquid-phase epoxidation of olefins was
investigated using cyclooctene as a model substrate and tert-
butyl hydroperoxide (TBHP) as the mono-oxygen source, at
558C under air, without additional solvent. All materials
catalysed the oxidation of cyclooctene to the corresponding
epoxide, and no reaction took place without catalyst. In the
first catalytic runs, the systems containing materials MCM-41/
1 and MCM-41/2 exhibited excellent catalytic activity and
selectivity, with conversion of substrate reaching more than
90% within a few hours (Figure 11; Table 3). This perform-
1
1
[a] Turnover frequency calculated after 1h of reaction. [b] Cyclooctene
conversion after 24 h. [c] 1,2-Cyclooctanediol was formed.
improved long-range structural order (as evidenced by the
presence of five reflections in the powder XRD pattern)
resulting from the different method by which the host
material used here was prepared. The higher catalytic activity
may have resulted from improved adsorption characteristics.
Oyama et al., for example, reported that the silica surface of
silica-supported molybdenum materials participated in the
oxidation of methanol by holding reaction intermediates,
indicating that the storage capacity of the support can affect
the overall activity.[11a]
The recyclability of MCM-41/1 and MCM-41/2 for cyclo-
octene epoxidation was studied by using the catalysts in
second and third reaction runs (Figure 11; Table 3). In both
cases a loss of activity was observed from the first to the
second runs, while catalytic activity was almost completely
retained from the second to the third runs. Selectivity to the
epoxide remained very high (Table 3). It is possible that, in the
first run, the epoxidation reaction took place within the
mesopores of the silica support, involving a relatively small
number of highly active oxomolybdenum centres. The active
species may have been deactivated (and/or leached into
solution) during the course of the reaction, leading to a
decrease in activity for the recycled catalyst. As noted above,
the catalytic activity and kinetic profiles for the first runs were
very different from those usually associated with complexes of
the type [MoO2X2(thf)2]; this suggests that the ordered
mesoporous support still plays a part in the reaction. For the
second and third runs the reactions appeared to be truly
heterogeneous, involving oxomolybdenum species covalently
anchored to the support. For a given catalytic run, the
activities of the catalysts derived from either complex 1 or 2
were similar. In accordance with the EXAFS results, this
suggests that the halogeno ligands had been at least partially
removed during the grafting reactions, since dichloro com-
plexes [MoO2Cl2S2] usually exhibit higher activity (because of
the increased electrophilicity of the metal centre) than the
corresponding dibromo complexes.[26]
Figure 11. Conversion versus time curves for the oxidation of cyclooctene
with TBHP in the presence of a) MCM-41/1, b) MCM-41/2, c) MCM-41/1/
~
*
Et3N and d) MCM-41/2/Et3N ( , 1st run; Â, 2nd run; , 3rd run).
ance was much better than that usually observed in homoge-
neous-phase reaction for solvent adducts [MoO2X2S2] (X ˆ
Cl, Br).[25] For example, under reaction conditions similar to
those used in the present work, conversion of cyclooctene in
the presence of the dibromo complex [MoO2Br2(CH3CN)2]
quickly reached about 65% but did not proceed significantly
further.[25] This was attributed to the pronounced water
sensitivity of the complex.
The MCM-41/1 and MCM-41/2 samples were prepared
according to the procedure described in ref. [12] and had
similar molybdenum loadings (approximately 1 wt.%). How-
ever, in the present study the catalysts were considerably
more active in cyclooctene epoxidation than those prepared in
the previous work. This can only be attributed to the much
The catalytic performance of MCM-41/1 was also tested in
the epoxidation of other olefins, since cyclooctene has high
reactivity and a low tendency towards epoxide opening. The
4386
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 4380 4390

Oxidation of Organic Compounds
4380 4390
values of turnover frequency (TOF) and olefin conversion are
shown in Figure 12. This catalyst oxidises 1-octene, trans-2-
octene and cyclododecene with excellent product selectivity,
the corresponding epoxide being the only observed product.
In the oxidation of a-pinene, pinene oxide and campholenic
aldehyde are produced each with 50% selectivity, at 20%
conversion. MCM-41/1 is less reactive towards the terminal
carbon carbon double bond of 1-octene than to a more
substituted olefin (with a more nucleophilic double bond)
such as trans-2-octene. For bulkier substrates such as cyclo-
dodecene, the reaction is slower than for cyclooctene.
standing of this mechanism would require more detailed
spectroscopic studies in situ, which are probably not worth-
while because of the relatively poor catalytic performance.
Catalytic oxidation of alcohols: The oxidation of alcohols to
carbonyl compounds and carboxylic acids is a fundamental
synthetic transformation. Molybdenum oxide-based catalysts
have been used successfully as catalysts for the gas-phase
oxidative dehydrogenation (ODH) of alcohols.[10d, 11a] Ruthe-
nium complexes can oxidise alcohols effectively in the
presence of basic TBHP; the reaction gives active intermedi-
ate tert-butylperoxoruthenium species, RuOO(tBu), which
are proton acceptors.[27] Species of this nature are also formed
by the reaction of TBHP with molybdenum complexes
[MoO2X2L2] (X ˆ Cl, Br, CH3), where L2 is a Lewis base
ligand.[18] Hence, we studied the catalytic potential of the
derivatised MCM-41 materials in the liquid-phase oxidation
of a range of primary and secondary benzylic and aliphatic
alcohols, namely benzyl alcohol, sec-phenethyl alcohol, cyclo-
hexanol, cyclooctanol, cyclohexylmethanol, 1-cyclohexyletha-
nol, 1-octanol and (1R,2S,5R)-(À)-menthol. Control experi-
ments were performed without a catalyst and, where neces-
sary, the net conversion with catalyst was calculated by
subtracting the non-catalytic contribution from the observed
conversion (Table 4). Without TBHP no alcohol oxidation
was observed. Pristine MCM-41 had no significant catalytic
activity. All derivatised materials exhibited poor activity in
the oxidation of primary aliphatic alcohols such as 1-octanol
and cyclohexylmethanol, giving less than 3% conversion to
the corresponding aldehyde. The most outstanding results
were those obtained for the oxidation of secondary aliphatic
and benzylic alcohols in the presence of MCM-41/1/Et3N.
Conversions at 24 h ranged from 43% to 59% and the
corresponding carbonyl products were mainly formed (Ta-
ble 4). A maximum TOF of 5.9 mmol(gÀ1 cat)hÀ1 was ob-
served for the oxidation of benzyl alcohol to benzaldehyde
and benzoic acid (Figure 13). The latter was a secondary
Figure 12. Oxidation of olefins using TBHP at 558C in the presence of
MCM-41/1. The turnover frequency (TOF) was calculated at 1 h after the
start of reaction and the substrate conversion at 24 h for cyclooctene
ˆ
ˆ
ˆ
(Cy8 ); cyclododecene (Cy12 ), 1-octene (1nC8 ); trans-2-octene
ˆ
(2nC8 ); a-pinene (Pinene).
The materials prepared in the presence of triethylamine,
MCM-41/1/Et3N and MCM-41/2/Et3N gave worse results for
the catalytic epoxidation of cyclooctene than those prepared
in the absence of the amine. For MCM-41/1/Et3N, the activity
for the first run was higher than that for MCM-41/2/Et3N, but
for the second run there was a decrease in activity and both
systems exhibited similar kinetic curves (Figure 11). The poor
results were probably a consequence of the high metal
loadings (approximately 4 wt.%). As discussed above, the
EXAFS data point to the presence of oxo-bridged dinuclear
species in as-made MCM-41/1/Et3N. This may only represent
the resting structure of the catalyst; that is, transformation to
active monometallic species may take place in the presence of
TBHP.[20] Leaching of active mononuclear species into
solution would account for the drop in activity between the
second and third runs. The system containing MCM-41/1/Et3N
is also unusual in that 1,2-cyclooctanediol is always formed as
a secondary reaction product. Epoxide ring opening occurs
when the Lewis acidity, provided by the MoVI metal centre
and required for peroxide activation, suffices for converting
the epoxide into the diol. If the interaction of Et3N with the
surface results in an increase in protonic acid strength of some
silanol groups able to decompose the epoxide, then ring
opening should also be observed in the presence of MCM-41/
2/Et3N, which is not the case. The production of 1,2-cyclo-
octanediol must instead be due to the formation of certain
oxomolybdenum species during the reaction. A better under-
Figure 13. Oxidation of alcohols using TBHP at 558C in the presence of
MCM-41/1/Et3N. The TOF was calculated at 1 h after the start of reaction
of 1-octanol (1-nC8OH), cyclooctanol (Cy8OH), cyclohexanol (Cy6OH),
1-cyclohexylethanol (Cy6EtOH), cyclohexylmethanol (Cy6MeOH), ben-
zyl alcohol (BzOH), sec-phenethyl alcohol (PhEtOH) and (1R,2S,5R)-(À)-
menthol (menthol).
Chem. Eur. J. 2003, 9, 4380 4390
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4387

I. S. GonÁalves et al.
FULL PAPER
Table 4. Alcohol oxidations using TBHP at 558C in the presence of MCM-41
materials.
Substrate Catalyst
Conversion[a] [%] Selectivity
to carbonyl[b] [%]
Cy8OH
Cy6OH
MCM-41/1
MCM-41/2
MCM-41/1/Et3N (1st run) 53
MCM-41/1/Et3N (2nd run) 48
MCM-41/1/Et3N (3rd run) 38
MCM-41/2/Et3N
7
7
100
100
100
100
100
no catalyst
MCM-41/1
29
100
MCM-41/2
2
48
100
100
MCM-41/1/Et3N
MCM-41/2/Et3N
no catalyst
11
6
5
43
2
12
4
6
59
8
10
2
12
45
19
11
10
100
100
100
100
100
100
100
100
100
100
100
89[c]
85[c]
67[c]
100
87[c]
100
Cy6EtOH MCM-41/1
MCM-41/2
MCM-41/1/Et3N
MCM-41/2/Et3N
no catalyst
Figure 14. Conversion versus time curves for the oxidation of cyclooctanol
~
,
*
using TBHP in the presence of MCM-41/1/Et3N ( , 1st run; Â, 2nd run;
PhEtOH MCM-41/1
&
3rd run); [MoO2Cl2(thf)2] ( ) and without a catalyst ( ).
MCM-41/2
MCM-41/1/Et3N
MCM-41/2/Et3N
no catalyst
MCM-41/1
MCM-41/2
MCM-41/1/Et3N
MCM-41/2/Et3N
no catalyst
catalyst is the presence of oxo-bridged Mo dimers. Since the
presence of other Mo species cannot be ruled out, it cannot be
assumed that those species are totally responsible for the
observed ODH activity. As mentioned above, dinuclear
species may be partly transformed into monometallic species
in the presence of TBHP. Marchi et al. studied the surface
species of silica-supported molybdenum catalysts,[28] and
reported that doping of the silica surface with sodium
modified the surface properties, generating some species with
molybdenum in tetrahedral coordination, which may be more
reducible at lower temperatures than those formed on non-
doped silica. Kikutani, in detailed EXAFS and UV/Vis
spectroscopic studies on silica-supported molybdenum cata-
lysts for the conversion of ethanol, showed that the reduction
of the catalysts by the alcohol is a complicated process in
which at least two distinctive states may be involved.[29] These
findings indicate that a full understanding of the mechanisms
of alcohol oxidation on supported catalysts is complicated and
requires in-situ characterisation techniques and possibly
labelled compounds.
BzOH
menthol
MCM-41/1
MCM-41/2
MCM-41/1/Et3N
MCM-41/2/Et3N
no catalyst
10
2
100
100
[a] Substrate conversion after 24 h. In the presence of a catalyst the net
conversion was corrected by subtracting the non-catalytic contribution from the
observed conversion. [b] Calculated as: (number of moles of carbonyl product
formed)/(total number of moles of products formed) Â 100. [c] Benzoic acid was
formed.
product resulting from the consecutive oxidation of benzal-
dehyde. This stepwise oxidation of benzyl alcohol was
confirmed by carrying out the oxidation of benzaldehyde in
the presence of MCM-41/1/Et3N, which yielded 94% benzoic
acid after 24 h.
The recyclability of MCM-41/1/Et3N for cyclooctanol
oxidation is comparable with that observed for cyclooctene
epoxidation (Figure 14). Metal leaching was less than 5%.
Water is a possible by-product formed in ODH, which may
contribute to solvation of some metal species. Also shown in
Figure 14 are the results obtained for cyclooctanol oxidation
in the presence of the precursor complex [MoO2Cl2(thf)2],
with a 1% molar ratio of catalyst/substrate. The kinetic curve
is almost identical to that obtained without catalyst. This is
another indication that the grafted metal species in MCM-41/
1/Et3N have a structure quite different from that of the
precursor dioxomolybdenum(vi) complex.
For the reactions studied here, the formation of the
aldehyde or ketone results from a nucleophilic attack on the
alcohol group. Hence, the nucleophilic oxygen species should
show high enough basicity to abstract hydrogen from the
alcohol molecule. A distinctive feature of the MCM-41/1/Et3N
Concluding Remarks
Mesoporous silica MCM-41 derivatised with oxomolybde-
num(vi) species can be prepared in a straightforward manner
by grafting calcined MCM-41 with the complexes
[MoO2X2(thf)2] (X ˆ Cl, Br). EXAFS spectroscopy shows
that these species are composed of isolated dioxomolybde-
num(vi) groups at low loadings. The high dispersion of active
sites appears to promote good catalytic activity for the
epoxidation of olefins by TBHP. At higher loading, achieved
using triethylamine to activate the surface silanol groups, the
material formed with the dichloro complex contains dimeric
or polymeric species. At the same high loading, the material
formed with the dibromo complex contains more mononu-
clear than dinuclear sites, revealing that the complex has a
quite different reactivity and surface chemistry. The materials
4388
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 4380 4390

Oxidation of Organic Compounds
4380 4390
desorption branch of the experimental isotherms by a method based on the
area of the pore walls.[32, 33] Assuming open cylindrical pores with radius rp
and zero contact angle, and correcting for the thickness of the layer already
adsorbed, rp can be calculated by summing the Kelvin radius and the
statistical average thickness (t, calculated using the Halsey equation) of the
adsorbed layer.
with the high metal loadings were not very active as catalysts
for olefin epoxidation, but good results were obtained for the
oxidation of secondary aliphatic and benzylic alcohols using
the material containing oxo-bridged dimers. The actual
catalytic species is likely to be formed by cleavage of the
Mo O Mo bond by TBHP.
Mo K-edge X-ray absorption spectra were measured at room temperature
in transmission mode on beamline BM29 at the ESRF (Grenoble),[34]
2
operating at 6 GeV in ³3 filling mode with typical currents of 170
200 mA. One scan was performed for each sample and set up to record
the pre-edge at 5 eV steps and the post-edge region in 0.025 0.05äÀ1 steps,
giving a total acquisition time of approximately 45 min per scan. The order-
sorting double Si(311) crystal monochromator was detuned by 40% to
ensure harmonic rejection. Solid samples were diluted with BN and pressed
into 13 mm pellets. Ionisation chamber detectors were filled with Kr to give
30% absorbing I0 (incidence) and 70% absorbing It (transmission). The
programs EXCALIB and EXBACK (SRS Daresbury Laboratory, UK)
were used in the usual manner for calibration and background subtraction
of the raw data. EXAFS curve-fitting analyses, by least-squares refinement
of the non-Fourier filtered k3-weighted EXAFS data, were carried out
using the program EXCURVE (version EXCURV98[35]), applying fast
curved wave theory.[36] Phase shifts were obtained within this program using
ab-initio calculations based on the Hedin Lundqvist/von Barth scheme.
The calculations were performed with single scattering only.
Experimental Section
Catalyst preparation: All preparations and manipulations were carried out
using standard Schlenk techniques under nitrogen. Solvents were dried by
standard procedures (THF, hexane and diethyl ether with Na/benzophe-
none; CH2Cl2 with CaH2), distilled under nitrogen and kept over 4 ä
molecular sieves. MoO2Cl2 and MoO2Br2 were obtained from Aldrich and
used as received. The solvent adducts 1 and 2 were prepared as described
previously.[30] Purely siliceous MCM-41 was synthesised using
[(C14H29)N(CH3)3]Br as the templating agent.[31] Before the grafting
experiments, physisorbed water was removed from calcined MCM-41 by
heating it at 1808C in vacuo (10À2 Pa) for 2 h.
In a typical grafting experiment, a solution of [MoO2X2(thf)2] (1 mmol) in
CH2Cl2 (10 mL) was added to a stirred suspension of dry MCM-41 (1.0 g) in
CH2Cl2 (10 mL). For MCM-41/1 and MCM-41/2, the reaction mixture was
stirred at ambient temperature overnight. For MCM-41/1/Et3N and MCM-
41/2/Et3N, freshly distilled triethylamine (0.3 mL, 1.8 mmol) was added
dropwise 30 min after the addition of the [MoO2X2(thf)2] solution. The
reaction mixture was then stirred at ambient temperature overnight. The
derivatised silicas were isolated by filtration, washed repeatedly with
CH2Cl2, and dried in vacuo at ambient temperature for several hours.
Catalytic oxidation reactions: The liquid-phase oxidation of cyclooctene
was carried out at 558C under an air atmosphere, in a micro reaction vessel
equipped with a magnetic stirrer. The reactor was loaded with catalyst
(175 mg), cyclooctene (7.3 mmol) and tert-butyl hydroperoxide (11 mmol)
as oxidant (5.5m in decane). Samples were withdrawn periodically and
analysed using
a gas chromatograph (Varian 3800) equipped with a
capillary column (DB-5, 30 m  0.25 mm for olefin epoxidation and CP
WAX 52CB, 30 m  0.53 mm for alcohol oxidation) and a flame ionisation
detector. The products were quantified using calibration curves and n-
nonane or undecane as internal standard (added after the reaction).
Catalysts were recycled by recovering the solids by filtration, washing them
thoroughly with dichloromethane and drying them at room temperature
overnight.
MCM-41/1: elemental analysis found (%) for: Mo 1.1; IR (KBr):
nÄ ˆ 2982(w), 2886(w), 1230(s), 1077(vs), 965(m), 917(sh), 889(sh), 800(m),
573(m), 459(s).
MCM-41/2: elemental analysis found (%) for: Mo 0.5; IR (KBr):
nÄ ˆ 2979(w), 2894(w), 1226(s), 1071(vs), 956(m), 918(sh), 891(sh), 802(m),
567(m), 453(s).
MCM-41/1/Et3N: elemental analysis found (%) for: Mo 4.0; IR (KBr):
nÄ ˆ 2990(m), 2793(m), 2730(m), 2693(m), 2494(m), 1474(m), 1398(m),
1385(w), 1233(s), 1080(vs), 949(m), 918(sh), 885(sh), 804(m), 577(m),
459(s).
Acknowledgements
This work was partly funded by the FCT, POCTI and FEDER (Project
PCTI/1999/QUI/32889). A.A.V. and M.P. thank the FCT for post-doctoral
grants, and C.N. thanks the University of Aveiro and the FCT for doctoral
grants. We acknowledge the European Synchrotron Radiation Facility for
provision of synchrotron radiation facilities and we thank Stuart Ansell for
assistance in using beamline BM29.
MCM-41/2/Et3N: elemental analysis found (%) for: Mo 3.9; IR (KBr):
nÄ ˆ 2986(m), 2787(m), 2726(m), 2689(m), 2490(m), 1475(m), 1398(m),
1385(w), 1079(vs), 965(w), 945 (m), 919(sh), 803(s), 566(m), 449(s).
Physical measurements: The molybdenum content in the samples was
determined by ICP-AES. Powder XRD data were collected on a Philips
X×pert diffractometer using CuKa radiation filtered by Ni. Data were
collected at room temperature between 28 ꢂ 2q ꢂ 128 with a step size of
0.028 and a count time of 5 s/step. IR spectra in transmission mode were
measured with a Mattson Mod 7000 FTIR spectrometer using KBr pellets.
13C and 29Si NMR spectra were recorded at 100.62 and 79.49 MHz,
respectively, on a (9.4 T) Bruker Avance 400 spectrometer. 29Si MAS NMR
spectra were recorded with 408 pulses, spinning rates of 5.0 5.5 kHz, and
60 s recycle delays. 29Si CP-MAS NMR spectra were recorded with 5.5 ms
1H 908 pulses, 8 ms contact time, a spinning rate of 4.5 kHz and 4 s recycle
delays. 13C CP-MAS NMR spectra were recorded with a 4.5 ms 1H 908 pulse,
2 ms contact time, a spinning rate of 8 kHz and 4 s recycle delays. Chemical
shifts are stated in parts per million from tetramethylsilane (TMS).
[1] a) M. A. Banares, I. E. Wachs, J. Raman Spectrosc. 2002, 33, 359;
b) M. A. Banares, Catal. Today 1999, 51, 319; c) N. C. Ramani, D. L.
Sullivan, J. G. Ekerdt, J. M. Jehng, I. E. Wachs, J. Catal. 1998, 176, 143;
d) M. A. Banares, L. J. Alemany, M. L. Granados, M. Faraldos, J. L. G.
Fierro, Catal. Today 1997, 33, 73; e) I. E. Wachs, Catal. Today 1996, 27,
437;f) I. E. Wachs, G. Deo, M. A. Vuurman, H. C. Hu, D. S. Kim, J. M.
Jehng, J. Mol. Catal. 1993, 82, 443.
[2] Y. Iwasawa, in Tailored Metal Catalysts (Ed.: Y. Iwasawa), D. Riedel,
Dordrecht, 1986, pp. 1 85.
Nitrogen adsorption desorption isotherms were measured at 77 K in a
gravimetric adsorption apparatus equipped with a CI electronic MK2-M5
microbalance and an Edwards Barocel pressure sensor. The out-gassing
temperature was raised slowly (1 KminÀ1) to 723 K for the pristine MCM-
41 samples and to 413 K for the derivatised materials (to minimise
destruction of the functionalities), and maintained at that temperature
overnight at a residual pressure of approximately 10À4 mbar. The specific
surface areas (SBET) were calculated by applying the BET equation. The
specific total pore volume (micropore plus mesopore, VP) was estimated
from the N2 uptake at p/p0 ꢁ 0.95. The pore size distribution (PSD) curves
(the differential volume adsorbed with respect to the differential pore size
per unit mass, plotted against pore width) were computed from the
[3] a) J. Y. Ying, C. P. Mehnert, M. S. Wong, Angew. Chem. 1999, 111, 58;
Angew. Chem. Int. Ed. 1999, 38, 56; b) U. Ciesla, F. Sch¸th, Micropor.
Mesopor. Mater. 1999, 27, 131; c) X. He, D. Antonelli, Angew. Chem.
2002, 114, 222; Angew. Chem. Int. Ed. 2002, 41, 214.
[4] a) A. Corma, Chem. Rev. 1997, 97, 2373; b) A. Sayari, Chem. Mater.
1996, 8, 1840.
[5] a) C. T. Kresge, M. E. Leonowicz, W. J. Roth, J. C. Vartuli, J. S. Beck,
Nature 1992, 359, 710; b) J. S. Beck, J. C. Vartuli, W. J. Roth, M. E.
Leonowicz, C. T. Kresge, K. D. Schmitt, C. T.-W. Chu, D. H. Olson,
E. W. Sheppard, S. B. McCullen, J. B. Higgins, J. L. Schlenker, J. Am.
Chem. Soc. 1992, 114, 10834.
[6] Q. Huo, D. I. Margolese, G. D. Stucky, Chem. Mater. 1996, 8, 1147.
Chem. Eur. J. 2003, 9, 4380 4390
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4389

I. S. GonÁalves et al.
FULL PAPER
¬
[18] F. E. K¸hn, M. Groarke, E. Bencze, E. Herdtweck, A. Prazeres, A. M.
[7] S. Inagaki, Y. Fukushima, K. Kuroda, J. Chem. Soc. Chem. Commun.
1993, 680.
ƒ
Santos, M. J. Calhorda, C. C. Romao, I. S. GonÁalves, A. D. Lopes, M.
[8] a) K. Moller, T. Bein, Chem. Mater. 1998, 10, 2950; b) T. Maschmeyer,
Curr. Opin. Solid State Mater. Sci. 1998, 3, 71; c) D. E. De Vos, M.
Dams, B. F. Sels, P. A. Jacobs, Chem. Rev. 2002, 102, 3615.
Pillinger, Chem. Eur. J. 2002, 8, 2370.
¬
œ ¬
[19] M. Cindric, N. Strukan, V. Vrdoljak, M. Devcic, Z. Veksli, B.
Kamenar, Inorg. Chim. Acta 2000, 304, 260.
[9] a) M. S. Morey, J. D. Bryan, S. Schwarz, G. D. Stucky, Chem. Mater.
2000, 12, 3435; b) M. S. Morey, A. Davidson, H. Eckert, G. D. Stucky,
Chem. Mater. 1996, 8, 486; c) P. VanDerVoort, M. S. Morey, G. D.
Stucky, M. Mathieu, E. F. Vansant, J. Phys. Chem. B 1998, 102, 585;
d) M. S. Morey, S. Schwarz, M. Frˆba, G. D. Stucky, J. Phys. Chem. B
1999, 103, 2037; e) M. S. Morey, A. Davidson, G. D. Stucky, Micropor.
Mater. 1996, 6, 99; f) M. S. Morey, S. O×Brien, S. Schwarz, G. D. Stucky,
Chem. Mater. 2000, 12, 898.
[10] a) T. Maschmeyer, F. Rey, G. Sankar, J. M. Thomas, Nature 1995, 378,
159; b) L. Marchese, E. Gianotti, V. Dellarocca, T. Maschmeyer, F.
Rey, S. Coluccia, J. M. Thomas, Phys. Chem. Chem. Phys. 1999, 1, 585;
c) P. Ferreira, I. S. GonÁalves, F. E. K¸hn, M. Pillinger, J. Rocha, A. M.
Santos, A. Thursfield, Eur. J. Inorg. Chem. 2000, 551; d) I. J. Shannon,
T. Maschmeyer, R. D. Oldroyd, G. Sankar, J. M. Thomas, H. Pernot,
J.-P. Balikdjian, M. Che, J. Chem. Soc. Faraday Trans. 1998, 94, 1495.
[11] a) M. Seman, J. N. Kondo, K. Domen, R. Radhakrishnan, S. T. Oyama,
J. Phys. Chem. B 2002, 106, 12965; b) S. Takenaka, T. Tanaka, T.
Funabiki, S. Yoshida, J. Phys. Chem. B 1998, 102, 2960; c) G. Mestl,
T. K. K. Srinivasan, Catal. Rev.–Sci. Eng. 1998, 40, 451; d) F. Arena,
A. Parmaliana, J. Phys. Chem. 1996, 100, 19994; e) M. Faraldos, M. A.
Banares, J. A. Anderson, H. C. Hu, I. E. Wachs, J. L. G. Fierro, J.
Catal. 1996, 160, 214; f) H. C. Hu, I. E. Wachs, J. Phys. Chem. 1995, 99,
10911; g) M. A. Banares, H. C. Hu, I. E. Wachs, J. Catal. 1994, 150,
407; h) A. Parmaliana, V. Sokolovskii, D. Miceli, F. Arena, N.
Giordano, J. Catal. 1994, 148, 514.
[20] S. Leinonen, D. C. Sherrington, A. Sneddon, D. McLoughlin, J.
Corker, C. Canevali, F. Morazzoni, J. Reedijk, S. B. D. Spratt, J. Catal.
1999, 183, 251.
[21] G. N. Schrauzer, L. A. Hughes, N. Strampach, Z. Naturforsch. B 1982,
37, 380.
[22] M. S. Rau, C. M. Kretz, G. L. Geoffroy, A. L. Rheingold, Organo-
metallics 1993, 12, 3447.
[23] a) D. Saurenz, F. Demirhan, P. Richard, R. Poli, H. Sitzmann, Eur. J.
Inorg. Chem. 2002, 1415; b) E. Collange, J. A. Garcia, R. Poli, New J.
Chem. 2002, 26, 1249.
[24] H. Arzoumanian, R. Lai, S. Le Bot, M. Pierrot, F. Ridouane, Inorg.
Chim. Acta 1991, 182, 173.
[25] F. E. K¸hn, E. Herdtweck, J. J. Haider, W. A. Herrmann, I. S.
GonÁalves, A. D. Lopes, C. C. Romaƒo, J. Organomet. Chem. 1999,
583, 3.
[26] a) F. E. Kuhn, A. D. Lopes, A. M. Santos, E. Herdtweck, J. J. Haider,
ƒ
C. C. Romao, A. G. Santos, J. Mol. Catal. A: Chem. 2000, 151, 147;
b) F. E. Kuhn, A. M. Santos, A. D. Lopes, I. S. GonÁalves, E.
ƒ
Herdtweck, C. C. Romao, J. Mol. Catal. A: Chem. 2000, 164, 25.
[27] W.-H. Fung, W.-Y. Yu, C.-M. Che, J. Org. Chem. 1998, 63, 2873.
[28] A. J. Marchi, E. J. Lede, F. G. Requejo, M. RenterÌa, S. Irusta, E. A.
Lombardo, E. E. Miro, Catal. Lett. 1997, 48, 47.
[29] Y. Kikutani, J. Mol. Catal. A: Chem. 1999, 142, 247.
[30] W. M. Carmichael, D. A. Edwards, G. W. A. Fowles, P. R. Marshall,
Inorg. Chim. Acta 1967, 1, 93.
[12] P. Ferreira, I. S. GonÁalves, F. E. K¸hn, A. D. Lopes, M. A. Martins,
[31] C. D. Nunes, A. A. Valente, M. Pillinger, A. C. Fernandes, C. C.
ƒ
ƒ
M. Pillinger, A. Pina, J. Rocha, C. C. Romao, A. M. Santos, T. M.
Romao, J. Rocha, I. S. GonÁalves, J. Mater. Chem. 2002, 12, 1735.
Santos, A. A. Valente, Eur. J. Inorg. Chem. 2000, 2263.
[13] a) X. S. Zhao, G. Q. Lu, A. K. Whittaker, G. J. Miller, H. Y. Zhu, J.
Phys. Chem. B 1997, 101, 6525; b) P. L. Llewellyn, F. Sch¸th, Y. Grillet,
F. Rouquerol, J. Rouquerol, K. K. Unger, Langmuir 1995, 11, 574.
[14] a) B. Marler, U. Oberhagemann, S. Voltmann, H. Gies, Micropor.
Mater. 1996, 6, 375; b) W. Hammond, E. Prouzet, S. D. Mahanti, T. J.
Pinnavaia, Micropor. Mesopor. Mater. 1999, 27, 19.
[15] G. Sankar, M. A. Roberts, J. M. Thomas, G. U. Kulkarni, N. Ranga-
vittal, C. N. R. Rao, J. Solid State Chem. 1995, 119, 210.
[16] S. P. Cramer, K. O. Hodgson, W. O. Gillum, L. E. Mortenson, J. Am.
Chem. Soc. 1978, 100, 3398.
[32] S. J. Gregg, K. S. W. Sing, Adsorption, Surface Area and Porosity, 2nd
ed., Academic Press, London, 1982.
[33] J. M. Thomas, W. J. Thomas, Principles and Practice of Heterogeneous
Catalysis, Wiley VCH, Weinheim, 1997.
[34] A. Filipponi, M. Borowski, D. T. Bowron, S. Ansell, A. D. Cicco, S. D.
¡
Panfilis, J.-P. Itie, Rev. Sci. Instrum. 2000, 71, 2422.
[35] N. Binsted, EXCURV98, CCLRC Daresbury Laboratory, Warrington
(UK), 1998.
[36] a) S. J. Gurman, N. Binsted, I. Ross, J. Phys. C 1984, 17, 143; b) S. J.
Gurman, N. Binsted, I. Ross, J. Phys. C 1986, 19, 1845.
[17] J. Evans, J. T. Gauntlett, J. F. W. Mosselmans, Faraday Discuss. Chem.
Soc. 1990, 107.
Received: February 20, 2003 [F4887]
4390
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 4380 4390