Complexation of crystal-like mesoporous phenylene-silica with Cr(CO) 3 and catalytic performance in the oxidation of cyclooctene

Article abstract of DOI:10.1016/j.molcata.2010.08.010

Chromium carbonyl complexes were incorporated into a molecularly ordered mesoporous phenylene-silica by reaction with Cr(CO)₆ in refluxing 1,2-dichloroethane (DCE). The modified material with a chromium loading of 0.07 mmol g^-1 was characterised by powder X-ray diffraction, N₂ adsorption, thermogravimetric analysis (TGA), FT-IR, ¹³C and ²⁹Si MAS NMR spectroscopies. TGA under air indicated that thermal decarbonylation between 250 and 350 °C preceded the Cr-promoted decomposition and release of organic fragments at a lower temperature (onset ca. 380 °C) than that registered for the phenylene-silica precursor (onset ca. 550 °C). The catalytic performance of the modified phenylene-silica was evaluated in the oxidation of cis-cyclooctene at 55 °C with tert-butylhydroperoxide (TBHP) as the oxidant and DCE as the co-solvent. The main product was the epoxide, formed in 83% yield after 24 h, with 2-cycloocten-1-one being the main by-product. Although the activity of the catalyst decreased after the first recycle, possibly due to partial chromium leaching, it remained nearly the same thereafter, and leaching tests indicated that the catalytic reaction was essentially heterogeneous in nature. Tests carried out using a radical scavenger indicated that the reaction of the olefin with TBHP may involve both free radical and non-radical reaction mechanisms.

Full text of DOI:10.1016/j.molcata.2010.08.010

Journal of Molecular Catalysis A: Chemical 332 (2010) 13–18
Contents lists available at ScienceDirect
Journal of Molecular Catalysis A: Chemical
journal homepage: www.elsevier.com/locate/molcata
Complexation of crystal-like mesoporous phenylene-silica with Cr(CO)3 and
catalytic performance in the oxidation of cyclooctene
Ana C. Coelhoa, Salete S. Balulaa, Margarida M. Antunesa, Tsvetelina I. Gerganovab, Nicolas Biona,c
,
Paula Ferreirab, Martyn Pillingera, Anabela A. Valentea, João Rochaa, Isabel S. Gonc¸ alvesa,∗
a Department of Chemistry, CICECO, University of Aveiro, Campus de Santiago, 3810-193 Aveiro, Portugal
b Department of Ceramics and Glass Engineering, CICECO, University of Aveiro, 3810-193 Aveiro, Portugal
c Laboratoire de Catalyse en Chimie Organique, CNRS/University of Poitiers, 40 avenue du Recteur Pineau, 86022 Poitiers Cedex, France
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 6 April 2010
Received in revised form 16 July 2010
Accepted 6 August 2010
Available online 14 August 2010
Chromium carbonyl complexes were incorporated into a molecularly ordered mesoporous phenylene-
silica by reaction with Cr(CO)6 in refluxing 1,2-dichloroethane (DCE). The modified material with a
chromium loading of 0.07 mmol g−1 was characterised by powder X-ray diffraction, N2 adsorption, ther-
mogravimetric analysis (TGA), FT-IR, 13C and 29Si MAS NMR spectroscopies. TGA under air indicated that
thermal decarbonylation between 250 and 350 ◦C preceded the Cr-promoted decomposition and release
of organic fragments at a lower temperature (onset ca. 380 ◦C) than that registered for the phenylene-silica
precursor (onset ca. 550 ◦C). The catalytic performance of the modified phenylene-silica was evaluated
in the oxidation of cis-cyclooctene at 55 ◦C with tert-butylhydroperoxide (TBHP) as the oxidant and DCE
as the co-solvent. The main product was the epoxide, formed in 83% yield after 24 h, with 2-cycloocten-
1-one being the main by-product. Although the activity of the catalyst decreased after the first recycle,
possibly due to partial chromium leaching, it remained nearly the same thereafter, and leaching tests
indicated that the catalytic reaction was essentially heterogeneous in nature. Tests carried out using a
radical scavenger indicated that the reaction of the olefin with TBHP may involve both free radical and
non-radical reaction mechanisms.
Keywords:
Periodic mesoporous organosilicas
Supported catalysts
Chromium
Decarbonylation
Epoxidation
© 2010 Elsevier B.V. All rights reserved.
an alternative approach, {␩6-(organosilyl)arene}Cr(CO)3 species
1. Introduction
were covalently linked to silica supports by condensation reac-
Arene chromium tricarbonyl complexes have been widely used
as selective catalysts for numerous organic transformations [1],
such as the 1,4-hydrogenation of conjugated dienes to monoenes
[2], the hydrogenation of alkynes to (Z)-alkenes [3], the hydro-
genation of ␣,␤-unsaturated carbonyl compounds to saturated
analogues [3], and room temperature cycloaddition reactions [4].
The immobilisation of these complexes on inorganic oxides or
polymeric supports is of special interest, not just to prepare more
practical catalysts for the above reactions, but also to prepare het-
erogeneous catalysts for alkene polymerisation [5]. For example,
the adsorption of (␩6-arene)Cr(CO)3 on alumina or silica-alumina
gives sub-carbonyl species which are active as ethene polymerisa-
tion catalysts [6]. Various arene chromium tricarbonyl complexes
have been encapsulated within the supercages of faujasite zeo-
lites, either directly by adsorption of the preformed complex or
indirectly by adsorption of Cr(CO)6 (to give surface tricarbonyl
species, Cr(CO)3(Oz)3) followed by adsorption of the arene [7]. In
tions of alkoxysilane-functionalised complexes with the reactive
SiOH groups on the silica gel surfaces [8]. Preliminary catalytic
tests for the polymerisation of phenylacetylene indicated that
only a ladder intermediate was formed, the formation of the final
poly(phenylacetylene) not being detected.
Several research groups have sought the immobilisation of
arene chromium tricarbonyl complexes where the aromatic lig-
and is integrated into a silica or metal–organic network [5c,9].
The organosilica materials are generally prepared by a sol–gel
process involving the hydrolysis and polycondensation of sub-
stituted alkoxysilanes containing nonhydrolysable Si–C bonds.
There are two routes for the incorporation of the organometal-
lic complexes: route A, one-pot reaction using (␩6-arene)Cr(CO)3
containing one, two or three Si(OR)3 groups [5c,9a,9b]; route B,
hydrolysis of the organoalkoxysilane precursor followed by com-
plexation [9a]. Route A generally ensures a better homogeneity
of the complexed metal throughout the polymer, and also allows
the co-polymerisation of metal-bound and metal-free molecular
precursors [5c]. For route B, the hydrolysis-polycondensation of
C6H5Si(OR)3 leads to a solid in which most of the organic groups
are located at the surface and readily react with Cr(CO)6 to give
∗
Corresponding author. Fax: +351 234 370084.
E-mail address: igoncalves@ua.pt (I.S. Gonc¸ alves).
1381-1169/$ – see front matter © 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcata.2010.08.010

14
A.C. Coelho et al. / Journal of Molecular Catalysis A: Chemical 332 (2010) 13–18
anchored (␩6-arene)Cr(CO)3 complexes [9a]. In contrast, the use
of 1,4-bis(trialkoxysilyl)benzene gives a phenylene-bridged hybrid
silisesquioxane in which most of the organic groups are located
inside the solid rather than at the surface, and are therefore not
accessible for chemical reaction [9a]. An important development in
this field was the surfactant-mediated synthesis of periodic meso-
porous phenylene-silica materials with crystal-like pore walls [10].
These materials possess hexagonal arrays of uniform mesopores
and surface structures with alternating hydrophobic phenylene and
hydrophilic silica layers. Kamegawa et al. reported the incorpora-
tion of arenetricarbonyl complexes [–C6H4M(CO)3–, M = Cr, Mo]
by treatment of a crystal-like phenylene-silica with Cr(CO)6 or
Mo(CO)6 vapour, achieving complexation of 15% of the phenylene
groups [11].
In recent work we found that the liquid phase deposition
of Mo(CO)6 onto crystal-like mesoporous phenylene-silica gave
immobilised arene molybdenum tricarbonyl complexes with metal
loadings between 1.5 and 5.9 wt.% depending on the reaction con-
ditions [12]. These materials exhibit promising performance in
liquid phase catalytic olefin epoxidation. We have now extended
this study to chromium and report on the performance of a
Cr-containing material as a pre-catalyst for the oxidation of cis-
cyclooctene at 55 ◦C, using tert-butylhydroperoxide (TBHP) as the
oxidant.
with 4.5 ␮s 1H 90◦ pulses, a contact time of 2 ms, spinning rates of
7.0 and 9.0 kHz, and 4 s recycle delays.
2.2. Preparation of PMO-phCr(CO)3
Physisorbed water was removed from PMO-ph (1.0 g) by heat-
ing the material at 120 ◦C under reduced pressure for 2 h. A
suspension of Cr(CO)6 (0.50 g, 2.27 mmol) in 1,2-dichloroethane
(15 mL) was then added to a suspension of PMO-ph in 1,2-
dichloroethane (15 mL). After refluxing the mixture for 12 h, the
solution was filtered off and the solid washed with CH2Cl2 (3×
20 mL), and vacuum-dried (first cycle). A further two cycles were
performed using freshly prepared suspensions of Cr(CO)6 in 1,2-
dichloroethane. Elemental analysis found: (after first cycle) C 37.82,
H 3.08, Cr 0.14%; (after second cycle) C 37.15, H 3.19, Cr 0.26%; (after
third cycle) C 37.61, H 3.15, Cr 0.35%. Data from type IV N2 sorption
isotherm: BET specific surface area (p/p0 = 0.01–0.1) = 639 m2 g−1
,
specific total pore volume = 0.6 cm3 g−1, pore width (taken from
the maximum of the pore size distribution curve calculated using
the BJH method) = 3.7 nm. IR (KBr, cm−1): 3434s, 3062m, 3015m,
1978w, 1935m, 1830w, 1636m, 1385m, 1160vs, 1101sh, 1077vs,
1020sh, 922m, 809m, 756m, 559sv, 520 s, 432m, 388m.
2.3. Catalytic studies
The liquid phase catalytic epoxidation reactions were carried
out with magnetic stirring under air (atmospheric pressure) in
closed borosilicate reaction vessels (5 mL capacity), which were
immersed in a thermostatted oil bath (55 ◦C). The reaction vessel
was loaded with Cr(CO)6 (3.3 ␮mol) or PMO-phCr(CO)3 (50 mg, cor-
responding to 3.3 ␮mol of chromium), cis-cyclooctene (1.7 mmol),
TBHP (2.7 mmol, 5.5 M in decane) and 1,2-dichloroethane (2 mL). In
order to avoid the photodissociation of the carbonyl ligands (which
could occur by interaction with reagents or solvent [14]), the cat-
alytic tests were carried out in the dark. The reaction courses were
monitored using a Varian 3800 GC equipped with a flame ionisa-
tion detector and a DB-5 column (30 m × 0.53 mm), with hydrogen
as the carrier gas.
2. Experimental
2.1. Materials and methods
All air sensitive operations were carried out using standard
Schlenk techniques under
a nitrogen atmosphere. Octade-
cyltrimethylammonium bromide (ODTMA) and Cr(CO)6 were
obtained from Aldrich and used as received. Solvents were dried
by standard procedures, distilled under nitrogen and stored over
˚
4 A molecular sieves. 1,4-Bis(triethoxysilyl)benzene (BTEB) was
prepared by a Barbier–Grignard reaction and purified by distil-
lation under reduced pressure [10b,13]. The benzene-silica PMO
with a crystal-like wall structure (PMO-ph) was prepared exactly
as described previously, under basic conditions using ODTMA as
the structure directing agent and BTEB as the silica source [10b].
The surfactant was removed from the as-synthesised material by
solvent extraction with an acidic solution. Powder XRD (hkl in
3. Results and discussion
3.1. Catalyst preparation and characterisation
Mesoporous phenylene-silica with crystal-like pore walls
parentheses): 2ꢀ (◦) = 1.88 (1 0 0 overlapping with 1 1 0 and 2 0 0),
11.60, 23.40, 35.40; d1 0 0 = 47.0 A; a = 2d1 0 0
√
˚
˚
/
3 = 54.3 A.
(PMO-ph) was treated with a mixture of Cr(CO)6 and 1,2-
Microanalyses for C and H, and ICP-AES determination of Cr were
performed at the University of Aveiro. The N2 adsorption isotherm
was measured at −196 ◦C using a Gemini 2375 Micromeritics
Instrument (Eng. M.C. Costa, University of Aveiro). A pre-treatment
at 150 ◦C was applied prior to analysis. Thermogravimetric anal-
yses (TGA) were carried out under air using a Shimadzu TGA-50
system, with a heating rate of 5 ◦C min−1. Powder XRD data were
collected on a Philips X’Pert MPD diffractometer (Cu K␣ X-radiation,
dichloroethane (DCE) under reflux for 12 h (Scheme 1). The solid
was then filtered, washed with DCE, and vacuum-dried, giving
a pale yellow material with a chromium content of 0.14 wt.%
(0.03 mmol g−1). A similar treatment of PMO-ph with Mo(CO)6
gave a material with a much higher metal loading of 5.9 wt.%
(0.6 mmol g−1) [12]. In an attempt to boost the chromium load-
ing, the recovered solid was subjected to two more treatments
with Cr(CO)6 and DCE, which resulted in the final derivatised PMO
(PMO-phCr(CO)3) with a chromium content of 0.35 wt.%.
˚
ꢁ = 1.54060 A) fitted with a graphite monochromator and a flat
plate sample holder, in a Bragg-Brentano para-focusing optics con-
figuration. Samples were step-scanned in 0.02◦ 2ꢀ steps with a
counting time of 2 s per step. IR spectra were obtained using a
FT-IR Mattson-7000 infrared spectrophotometer with 2 cm−1 res-
olution. Attenuated total reflectance (ATR) spectra were measured
on the same instrument equipped with a Specac Golden Gate Mk
II ATR accessory having a diamond top-plate and KRS-5 focusing
lenses. Solid-state magic-angle spinning (MAS) NMR spectra were
recorded at 79.49 MHz for 29Si and 125.76 MHz for 13C on Bruker
Avance 400 and 500 spectrometers, respectively. 29Si MAS NMR
spectra were recorded with 40◦ pulses, a spinning rate of 5.0 kHz,
and 60 s recycle delays. 13C CP MAS NMR spectra were recorded
Fig. 1 shows the powder XRD patterns of PMO-ph and PMO-
phCr(CO)3. The pattern of PMO-ph contains one strong low-angle
peak at 47.0 A, which is assigned to the (1 0 0) reflection of a two-
˚
dimensional hexagonal symmetry (p6mm) lattice with a lattice
˚
constant a = 54.8 A. The broad, weak shoulder on this peak toward
higher angles (3–4◦ 2ꢀ) comprises the overlapping (1 1 0) and (2 0 0)
reflections. In addition to the low-angle peaks, three medium-
˚
range reflections at 7.6, 3.8 and 2.5 A are observed, which arise
from the molecular-scale periodicity in the PMO-ph pore walls
along the channel direction [10]. The pattern for PMO-phCr(CO)3 is
practically identical, which indicates retention of the mesoporous
structure as well as the molecular-scale periodicity.

A.C. Coelho et al. / Journal of Molecular Catalysis A: Chemical 332 (2010) 13–18
15
Fig. 2. Solid-state 29Si MAS NMR spectra of PMO-ph (a) and PMO-phCr(CO)3 (b).
ical modification. The 13C CP MAS NMR spectrum (Fig. 3) displays
a single peak at 133.5 ppm for the benzene ring carbon atoms,
in agreement with that observed for PMO-ph. The absence of
resonances for aromatic carbon atoms complexed with Cr(CO)3
(expected ca. 95 ppm) or CO ligands (expected ca. 230 ppm) can
be attributed to the low chromium loading. PMO-ph and PMO-
phCr(CO)3 exhibit IR bands in the range of 400–1700 cm−1 very
similar to those described by Hoffmann et al. for a benzene-bridged
PMO prepared using BTEB as the precursor [15]. Toward higher
frequencies the spectra display absorption bands assigned to com-
bination modes originating from the aromatic groups (1830 and
1935 cm−1) [16], ꢂCH (3015 and 3062 cm−1) and ꢂOH of hydrogen-
bonded silanol groups (3200–3800 cm−1). A very weak band at
1978 cm−1 (absent for PMO-ph) is assigned to the A1 vibra-
tional mode involving carbonyl groups of Cr(CO)3 units (Fig. 4)
[5b,8,9a,9b,11].
Scheme 1. Preparation of PMO-phCr(CO)3.
The material PMO-phCr(CO)3 was further characterised by FT-
IR, solid-state 29Si and 13C MAS NMR spectroscopies. The 29Si
MAS NMR spectrum (Fig. 2) is similar to that for PMO-ph, exhibit-
ing peaks at about −79.5 and −70.1 ppm attributed to T3 and T2
organosilica species [Tm = RSi(OSi)m(OH)3−m], respectively [10b].
The relative intensity of the peaks shows that the dominant silicon
species is T3, which is consistent with a high degree of condensa-
tion. Accordingly, virtually no signal around −60 ppm is present,
indicating that the proportion of T1 sites is no higher than about
5%. The absence of SiO4 species in the range of −110 to −90 ppm
confirms that no Si–C bond cleavage took place during the chem-
Fig. 3. Solid-state 13C CP MAS NMR spectra of PMO-ph (a) and PMO-phCr(CO)3 (b).
Spinning sidebands are labeled with asterisks.
Fig. 1. Powder XRD patterns of PMO-ph (a) and PMO-phCr(CO)3 (b).

16
A.C. Coelho et al. / Journal of Molecular Catalysis A: Chemical 332 (2010) 13–18
Table 1
Reaction of cis-cyclooctene with TBHP at 55 ◦C catalysed by PMO-phCr(CO)3.
Solvent
Conversion at
6 h/24 h (%)
Epoxide selectivity
at 6 h/24 h (%)
DCE
n-Hexane
Toluene
50/88
44/70
43/64
71/85
100/94
100/93
100/100
91/90
Without co-solvent
3.2. Catalytic oxidation of cis-cyclooctene
The reaction of cis-cyclooctene (Cy8) with TBHP at 55 ◦C,
using DCE as the co-solvent and PMO-phCr(CO)3 as the pre-
catalyst (Cr:Cy8:TBHP molar ratio = 0.2:100:159), gave conversions
of 50%/88% at 6 h/24 h, and 1,2-epoxycyclooctane was formed with
100%/94% selectivity (Table 1, Fig. 6). 2-Cycloocten-1-one was the
main by-product. When the same reaction was carried out using
PMO-phMo(CO)3 containing 1.5 wt.% Mo (Mo:Cy8:TBHP molar
ratio = 1:100:154), the epoxide was obtained in quantitative yield
within 1 h [12]. The formation of epoxides and enones as main prod-
ucts was previously reported for chromium-catalysed oxidation of
ꢃ5-steroids with TBHP using CH2Cl2 as the solvent [17]. The reac-
tion in the absence of PMO-phCr(CO)3 or in the presence of PMO-ph,
under similar reaction conditions, was very sluggish, giving conver-
sions at 6 h/24 h of 2%/4% and 2%/7%, respectively (Fig. 6). Hence,
for the reaction carried out using PMO-phCr(CO)3, the catalytically
active species contain chromium. When the material possessing
0.26 wt.% Cr was used instead of PMO-phCr(CO)3 (0.35 wt.% Cr),
under similar reaction conditions (identical Cr:Cy8:TBHP molar
ratios), conversion and epoxide selectivity at 24 h were 86% and
97%, respectively, which are similar to the results obtained for
PMO-phCr(CO)3 with the higher metal loading (88% and 94%,
respectively). Hence, it may be beneficial to increase the amount of
supported metal species to reduce costs associated with the catalyst
preparation (by using a lower amount of the PMO support).
Fig. 4. FT-IR spectra in the range of 1500–2100 cm−1 for PMO-ph (a) and PMO-
phCr(CO)3 (b).
Thermogravimetric analysis of PMO-ph under air shows a
weight loss of 8.0% below 100 ◦C due to desorption of physisorbed
water (Fig. 5). A second weight loss takes place above 550 ◦C and
corresponds to the decomposition and release of organic fragments
from the pore wall. The TGA curve for PMO-phCr(CO)3 shows a
small weight loss of 2.7% extending from room temperature to
150 ◦C, a second weight loss of 1.4% between 250 and 350 ◦C, and
full decomposition above 380 ◦C. The second step with a peak tem-
perature of 325 ◦C in the derivative thermogravimetric curve can
be assigned to the escape of carbon monoxide, and is in agreement
with that reported for hybrid organic-inorganic silica gels prepared
by the hydrolytic sol–gel polymerisation of pure arene Cr(CO)3 pre-
cursors containing one, two, or three Si(OMe)3 groups [9b]. Taking
into account a “background” loss of at least 0.5% in the range of
250–350 ◦C, the actual weight loss due to the removal of CO is less
than 0.9%, which is close to the calculated value of 0.6% based on
the Cr content of 0.35 wt.%. Despite the low metal loading, the sup-
ported Cr species generated by thermal decarbonylation appear to
promote the early decomposition and release of organic fragments
from the pore wall.
When the reaction of cis-cyclooctene was carried out under
nitrogen instead of atmospheric air, no decrease in reaction rate
was observed, suggesting that TBHP is the oxidant involved in the
(ep)oxidation. Iodometric titration carried out for the reaction of
TBHP (without Cy8), under similar reaction conditions, gave 83%
Fig. 6. Reaction of cis-cyclooctene with TBHP at 55 ◦C using PMO-phCr(CO)3 as the
pre-catalyst (ꢀ run 1, + run 2, ꢁ run 3), PMO-ph (ꢀ) and Cr(CO)6 (×). The results of
the leaching tests LT-run1 (ꢁ) and LT-run3 (ꢂ) are also shown. DCE was used as the
co-solvent for all reactions. The dotted lines are visual guides.
Fig. 5. TGA curves of PMO-ph (– – –) and PMO-phCr(CO)3 (—).

A.C. Coelho et al. / Journal of Molecular Catalysis A: Chemical 332 (2010) 13–18
17
conversion at 5 h, suggesting that the “non-productive” decomposi-
tion of TBHP during the catalytic oxidation of Cy8 may be significant
and therefore affect the overall reaction rate. It may be benefi-
cial to carry out the reaction under semi-batch mode to avoid the
“non-productive” decomposition of TBHP.
When the precursor Cr(CO)6 was used instead of PMO-
phCr(CO)3 (with the same equivalent amount of chromium), the
reaction of cis-cyclooctene was rather sluggish, giving 4%/18% con-
version at 6 h/24 h. Possibly, the dissolved precursor decomposed
into less active/inactive metal species, accounting for the slower
reaction in comparison to that observed for PMO-phCr(CO)3 (under
the applied oxidising conditions, the metal species on the PMO sup-
port may possess enhanced stability in comparison to those present
in homogeneous phase).
to the corresponding enone, which was not detected when using
RS), and (B) oxometal pathways responsible for the epoxidation
of cis-cyclooctene to the corresponding epoxide (the only prod-
uct formed when using RS). Taking into consideration that epoxide
selectivity is at least 94% until 88% conversion, it seems that the het-
erolytic pathways are preferred for the reaction of cis-cyclooctene
using PMO-phCr(CO)3 as the pre-catalyst. Similar mechanistic con-
siderations have been reported in the literature for the reaction
of olefins with TBHP in the presence of CrO3 and mono/bidentate
pyridines: the allylic oxidation of the olefin and the decomposition
of TBHP were reported to be free radical reactions, while the epox-
idation reaction was considered to be dependent on the formation
of a TBHP–CrO3 complex [18c]. In the case of (A) the chain reaction
may take place in the liquid bulk, which may contribute to the very
small increase in conversion between 30 min and 6 h for LT-run3
(free radicals formed prior to catalyst filtration).
The reaction of cis-cyclooctene with TBHP using PMO-phCr(CO)3
as the pre-catalyst was further investigated using toluene or n-
hexane as the co-solvent instead of DCE, or using no co-solvent
(Table 1). After 6 h the latter system gave the highest conversion of
71% (cf. 43–50% for the other systems), which rose to 85% after 24 h
(cf. 88% for DCE, 64% for toluene and 70% for n-hexane). The absence
of dilution effects when no co-solvent is added may account for the
higher conversion at 6 h. Of the co-solvents studied, DCE gives the
best results, which may be partly due to the higher solvent polarity:
a balance is required taking into consideration the surface polarity
of the PMO supported catalyst and the quite different polarities of
the reagents. No significant effects on product distribution were
observed with or without a co-solvent: the epoxide was always the
only product at 6 h reaction and after 24 h epoxide selectivity was
at least 93%.
After the first run using PMO-phCr(CO)3, the catalyst was sep-
arated from the reaction mixture by centrifugation, washed three
times with n-hexane, dried at room temperature, and used in a
second run. The reaction rate (based on conversions at 6 h/24 h)
decreased significantly from the first to the second run, possibly
due to partial metal leaching during the reaction (Fig. 6); ICP-AES
analysis indicated 20% metal leaching. When the solid was recov-
ered and used in a third run, only a minor decrease in the reaction
rate was observed (compared with the second run, Fig. 6); conver-
sions after 24 h were 69% for run 2 and 68% for run 3. In all three
consecutive batch runs the epoxide was the main product, and 2-
cycloocten-1-one was a minor by-product formed in less than 5%
yield at 24 h. To assess the homo/heterogeneous nature of the cat-
alytic reaction, leaching tests (LT) were performed for the fresh
catalyst (denoted LT-run1) and the solid recovered after the sec-
ond run (LT-run3). The LT were performed by separating the solid
phases from the reaction mixtures (runs 1 and 3) after 30 min by fil-
tration at the reaction temperature of 55 ◦C through a 0.2 ␮m PVDF
w/GMF Whatman membrane. The filtrates were then allowed to
react at the same temperature for a further 5.5 h. The increases in
conversion between 30 min and 6 h were 19% for LT-run1 and 4%
for LT-run3, compared with 45% for run 1 and 27% for run 3 carried
out without filtration, and 2% for the reaction carried out without
catalyst. These results suggest that the catalytic reaction in run 1
takes place in heterogeneous and homogeneous phases (possibly
due to partial metal leaching), and in subsequent runs the catalytic
reaction is essentially heterogeneous in nature.
The ATR FT-IR spectrum of the solid recovered after the third
reaction run was similar to that for PMO-phCr(CO)3 except that the
ꢂCO band at 1978 cm−1 was absent, indicating that decarbonylation
of the supported metal species took place. An induction period of
at least 10 min is observed for runs 1 and 2, which may be partly
due to the relatively slow formation of active oxidising species
(expected to involve primary decarbonylation, at least during run
1). Possibly, TBHP interacts with zero valent chromium centres to
give high valent oxometal intermediates that are responsible for
the transfer of the oxygen atom to the olefin. The oxygen trans-
fer may be effected by electron transfer, radical, carbocation or
metallaoxetane formation, or a combination of these mechanisms
[18]. GC–MS analysis of the reaction mixture for PMO-phCr(CO)3
showed the presence of di-tert-butylperoxide, which is formed by
the decomposition of TBHP into chain initiating alkoxy and alkyl
peroxy radicals in one-electron transfer processes. When the reac-
tion of cis-cyclooctene with TBHP was carried out in the presence
of PMO-phCr(CO)3 using equimolar amounts of TBHP and a radi-
cal scavenger (2,6-di-tert-butyl-4-methylphenol, denoted RS), the
cis-cyclooctene reaction was slower than without RS (suggesting
the formation of free radical intermediates), although it still took
place to a significant extent, giving 11%/47% conversion at 6 h/24 h,
with the epoxide being the only product. Two reaction mechanisms
for cis-cyclooctene oxidation may be involved: (A) free radical chain
process (possibly responsible for allylic oxidation of cis-cyclooctene
4. Conclusions
Mesoporous phenylene-silica with molecule-scale periodicity
in the pore walls can be modified with chromium carbonyl species
by the liquid phase deposition of Cr(CO)6. Elemental analysis indi-
cated a metal loading of about 0.07 mmol g−1, which is about one
order of magnitude lower than that previously achieved using
Mo(CO)6 under similar conditions. The reason(s) for the different
degrees of functionalisation are unclear at present. Attempts to
prepare PMO-phCr(CO)3 materials with higher loadings by one-
pot reactions using alkoxysilane-functionalised complexes have so
far been unsuccessful, mainly due to phase separation or to the
disintegration of incorporated (␩6-arene)Cr(CO)3 complexes dur-
ing surfactant removal. The characterisation of the supported Cr
species in PMO-phCr(CO)3 is complicated by the low loading. Nev-
ertheless, the FT-IR and TGA data are consistent with the presence
of chromium tricarbonyl complexes. The modified material pro-
motes the epoxidation of cis-cyclooctene with TBHP, leading to a
much faster reaction than that achieved using the hexacarbonyl
precursor in homogeneous phase. Recycling and leaching tests
for fresh and used catalysts suggest that active chromium species
retained on the PMO support are largely responsible for the epox-
idation reaction in recycling runs. Decarbonylation of supported
complexes takes place and it is postulated (on the basis of tests
using a radical scavenger) that the reaction of the olefin with TBHP
may involve a combination of free radical and non-radical reac-
tion mechanisms, with the latter possibly being responsible for the
formation of the epoxide.
Acknowledgements
The authors are grateful to the Fundac¸ ão para
a Ciên-
cia e a Tecnologia (FCT), POCI 2010, OE and FEDER (Projects

18
A.C. Coelho et al. / Journal of Molecular Catalysis A: Chemical 332 (2010) 13–18
POCI/CTM/55648/2004 and PPCDT/CTM/55648/2004), and to
RNME (Project FCT: REDE/1509/RME/2005) for financial support.
The FCT is acknowledged for PhD grants to A.C.C. and M.M.A.
[8] M. Morán, I. Cuadrado, M.C. Pascual, C.M. Casado, J. Losada, Organometallics 11
(1992) 1210.
[9] (a) G. Cerveau, R.J.P. Corriu, C. Lepeytre, J. Mater. Chem. 5 (1995) 793;
(b) G. Cerveau, R.J.P. Corriu, C. Lepeytre, Chem. Mater. 9 (1997) 2561;
(c) S.S. Kaye, J.R. Long, J. Am. Chem. Soc. 130 (2008) 806.
[10] (a) S. Inagaki, S. Guan, T. Ohsuna, O. Terasaki, Nature 416 (2002) 304;
(b) N. Bion, P. Ferreira, A. Valente, I.S. Gonc¸ alves, J. Rocha, J. Mater. Chem. 13
(2003) 1910;
References
[1] (a) M. Sodeoka, M. Shibasaki, Synthesis (1993) 643;
(b) J.H. Rigby, M.A. Kondratenko, Top. Organomet. Chem. 7 (2004) 181.
[2] (a) M. Cais, D. Fraenkel, K. Weidenbaum, Coord. Chem. Rev. 16 (1975) 27;
(b) P. Le Maux, J.Y. Saillard, D. Grandjean, G. Jaouen, J. Org. Chem. 45 (1980)
4524;
(c) S. Fujita, S. Inagaki, Chem. Mater. 20 (2008) 891;
(d) G. Smeulders, V. Meynen, G. Van Baelen, M. Mertens, O.I. Lebedev, G. Van
Tendeloo, B.U.W. Maes, P. Cool, J. Mater. Chem. 19 (2009) 3042.
[11] (a) T. Kamegawa, T. Sakai, M. Matsuoka, M. Anpo, J. Am. Chem. Soc. 127 (2005)
16784;
(c) P. Le Maux, R. Dabard, G. Simonneaux, J. Organomet. Chem. 241 (1983) 37;
(d) J.R. Tucker, D.P. Riley, J. Organomet. Chem. 279 (1985) 49;
(e) M. Sodeoka, Y. Ogawa, Y. Kirio, M. Shibasaki, Chem. Pharm. Bull. 39 (1991)
309;
(b) M. Matsuoka, T. Kamegawa, T.-H. Kim, T. Sakai, M. Anpo, J. Nanosci. Nan-
otechnol. 10 (2010) 314.
[12] A.C. Coelho, S.S. Balula, S.M. Bruno, J.C. Alonso, N. Bion, P. Ferreira, M. Pillinger,
A.A. Valente, J. Rocha, I.S. Gonc¸ alves, Adv. Synth. Catal. 352 (2010) 1759.
[13] K.J. Shea, D.A. Loy, O. Webster, J. Am. Chem. Soc. 114 (1992) 6700.
[14] (a) T. Szyman´ ska-Buzar, Coord. Chem. Rev. 250 (2006) 976;
(b) T.T. To, E.J. Heilweil, C.B. Duke III, T.J. Burkey, J. Phys. Chem. A 111 (2007)
6933.
[15] F. Hoffmann, M. Güngerich, P.J. Klar, M. Fröba, J. Phys. Chem. C 111 (2007) 5648.
[16] B. Onida, L. Borello, C. Busco, P. Ugliengo, Y. Goto, S. Inagaki, E. Garrone, J. Phys.
Chem. B 109 (2005) 11961.
[17] M.A. Fousteris, A.I. Koutsourea, S.S. Nikolaropoulos, A. Riahi, J. Muzart, J. Mol.
Catal. A: Chem. 250 (2006) 70.
[18] (a) B. Rihter, S. SriHari, S. Hunter, J. Masnovi, J. Am. Chem. Soc. 115 (1993) 3918,
references cited therein;
(f) T. Chandiran, S. Vancheesan, J. Mol. Catal. 88 (1994) 31.
[3] M. Sodeoka, M. Shibasaki, J. Org. Chem. 50 (1985) 1147.
[4] J.H. Rigby, L.W. Mann, B.J. Myers, Tetrahedron Lett. 42 (2001) 8773.
[5] (a) C.U. Pittman Jr., B.T. Kim, W.M. Douglas, J. Org. Chem. 40 (1975) 590;
(b) R.A. Awl, E.N. Frankel, J. Am. Oil Chem. Soc. (1981) 557;
(c) R. Ophir, Y. Shvo, J. Mol. Catal. A: Chem. 140 (1999) 259.
[6] (a) G. Hunter, C.H. Rochester, A.G. Wilkinson, J. Paton, J. Chem. Soc., Faraday
Trans. 93 (1997) 1205;
(b) G. Hunter, C.H. Rochester, A.G. Wilkinson, J. Paton, J. Chem. Soc., Faraday
Trans. 92 (1996) 4763;
(c) G. Hunter, C.H. Rochester, A.G. Wilkinson, J. Paton, J. Chem. Soc., Faraday
Trans. 92 (1996) 5093.
(b) I.W.C.E. Arends, R.A. Sheldon, Appl. Catal. A: Gen. 212 (2001) 175;
(c) G. Rothenberg, H. Wiener, Y. Sasson, J. Mol. Catal. A: Chem. 136 (1998) 253;
(d) S. Jarupinthusophon, U. Thong-In, W. Chavasiri, J. Mol. Catal. A: Chem. 270
(2007) 289.
[7] (a) S. Özkar, G.A. Ozin, K. Moller, T. Bein, J. Am. Chem. Soc. 112 (1990) 9575;
(b) C. Brémard, Coord. Chem. Rev. 178–180 (1998) 1647;
(c) W.M. Shirley, S.P. Scoville, Inorg. Chem. 42 (2003) 358;
(d) Y. Huang, R.R. Poissant, Langmuir 18 (2002) 5487.