Mo(II) complexes of 8-aminoquinoline and their immobilization in MCM-41

Article abstract of DOI:10.1016/j.apcata.2013.01.033

Two new Mo(II) complexes [MoBr(η³-C₃H ₅)(CO)₂(8-aq)] (1) and [MoI₂(CO) ₃(8-aq)] (2) containing the bidentate 8-aminoquinoline ligand (8-aq) were synthesized and characterized. They were immobilized in MCM-41. A 3-iodopropyltrimethoxysilane spacer reacted both with the surface, through the silane, and through the other end, with the coordinated 8-aq of complexes 1 and 2, leading to an immobilized form of the complex (MCM-Pr-1,2). In an alternative route, 8-aq reacted with 3-iodopropyltrimethoxysilane to form a new ligand L₁, which could be supported in the MCM-41 and then react with the metal precursors to afford (MCM-L₁-1,2). The complexes and the materials were characterized using FTIR and NMR spectroscopies, and the structure of the materials was checked with powder X-ray diffraction and nitrogen adsorption isotherms. The first synthetic procedure was less efficient in terms of metal load inside the channels of the materials. The complexes and the new materials were tested as catalytic precursors in the epoxidation of cis-cycloctene, styrene, 1-octene, R-(+)limonene, geraniol, cis-3-hexene-1-ol and trans-2-hexene-1-ol, using tert-butylhydroperoxide (TBHP) as oxidant. Although almost all the catalysts were 100% selective toward the epoxide, the conversions were in general poor. The best catalyst was complex 1, but the conversions dropped after immobilization. Conversions could be a bit improved by a careful choice of reaction conditions, the most effective being the absence of added solvent (the substrate acted as solvent).

Full text of DOI:10.1016/j.apcata.2013.01.033

Applied Catalysis A: General 455 (2013) 172–182
Contents lists available at SciVerse ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
Mo(II) complexes of 8-aminoquinoline and their immobilization in
MCM-41
a
a
b
a,∗
Marta S. Saraiva , Carla D. Nunes , Teresa G. Nunes , Maria José Calhorda
a
Departamento de Química e Bioquímica, CQB, Faculdade de Ciências da Universidade de Lisboa, Campo Grande, Ed. C8, 1749-016 Lisboa, Portugal
Centro de Química Estrutural, Instituto Superior Técnico, Universidade Técnica de Lisboa, Av. Rovisco Pais, 1049-001 Lisboa, Portugal
b
a r t i c l e i n f o
a b s t r a c t
3
Article history:
Received 7 August 2012
Received in revised form 18 January 2013
Accepted 31 January 2013
Available online 15 February 2013
3 5 2 2 3
Two new Mo(II) complexes [MoBr(␩ -C H )(CO) (8-aq)] (1) and [MoI (CO) (8-aq)] (2) containing the
bidentate 8-aminoquinoline ligand (8-aq) were synthesized and characterized. They were immobilized
in MCM-41. A 3-iodopropyltrimethoxysilane spacer reacted both with the surface, through the silane, and
through the other end, with the coordinated 8-aq of complexes 1 and 2, leading to an immobilized form
of the complex (MCM-Pr-1,2). In an alternative route, 8-aq reacted with 3-iodopropyltrimethoxysilane to
form a new ligand L1, which could be supported in the MCM-41 and then react with the metal precursors
to afford (MCM-L1-1,2). The complexes and the materials were characterized using FTIR and NMR spec-
troscopies, and the structure of the materials was checked with powder X-ray diffraction and nitrogen
adsorption isotherms. The first synthetic procedure was less efficient in terms of metal load inside the
channels of the materials.
Keywords:
Molybdenum
MCM
Oxidation catalysis
8
-Aminoquinoline
Alkenes
The complexes and the new materials were tested as catalytic precursors in the epoxidation of cis-
cycloctene, styrene, 1-octene, R-(+)limonene, geraniol, cis-3-hexene-1-ol and trans-2-hexene-1-ol, using
tert-butylhydroperoxide (TBHP) as oxidant. Although almost all the catalysts were 100% selective toward
the epoxide, the conversions were in general poor. The best catalyst was complex 1, but the conver-
sions dropped after immobilization. Conversions could be a bit improved by a careful choice of reaction
conditions, the most effective being the absence of added solvent (the substrate acted as solvent).
©
2013 Elsevier B.V. All rights reserved.
1
. Introduction
a covalent bond between the wall of the material and the cata-
lyst or catalyst precursor helps, on one hand, to prevent leaching,
allowing the reutilization of the catalyst; on the other, it is pos-
sible to vary the immobilized catalyst in a relatively easy way, so
that small active homogeneous catalysts can be supported with
minimal change, and their catalytic properties are not likely to
change [5–8]. These functionalized inorganic materials have been
widely applied in several fields, such as catalysis and optoelectron-
ics [1,5,9–11].
The development of heterogeneous catalysts remains a very
important research field, since it has not yet been possible to com-
bine in the same system the advantages of heterogeneous catalysts,
such as the easy catalyst recovery, with those of homogeneous
ones, namely the selectivity. Several routes to prepare new het-
erogeneous systems have been designed in the last decades [1,2].
Using a selected material to act as support where active cata-
lysts can be grafted counts as one of the most successful ones, in
special owing to the family of the Mobil MCM family of meso-
porous solids, reported for the first time in 1992 [3,4]. These
materials containing organized hexagonal channels with a diam-
eter of ∼2–50 nm can be easily functionalized by reaction of the
silanol groups in the walls with molecules carrying Si(OR)3 sub-
stituents. The fourth position around silicon can range from an
organometallic complex, to a ligand which can bind a number of
metal fragments, to a special active molecule. The formation of
3
The Mo(II) complexes [Mo(␩ -C H5)Br(CO) (CH CN) ] [12,13]
3
2
3
2
and [MoI (CO) (CH CN) ] [14] have also proved to be very versa-
2
3
3
2
tile, since the nitrile ligands are easily substituted by other ligands.
They have been used as catalysts [15], in particular in allylic alky-
lations [16], imine aziridation, [17] or phosphine oxidation [18]. It
3
has also been shown that [Mo(␩ -C H5)X(CO) (N N)] (X = Cl, Br)
3
2
species can behave selectively as catalyst precursors in olefin epox-
idation when t-butyl hydroperoxide (TBHP) is the oxidant, [19] a
reactivity previously detected in other Mo(II) complexes, namely
5
[
Mo(␩ -C5H5)(CO) (CH )] [20]. Mo(II) complexes are initially oxi-
3
3
dized to active Mo(VI) species. Catalysts containing Mo(VI) are used
in the ARCO-Halcon process for homogeneous olefin epoxidation
with TBHP as oxidant, thus contributing to the renewed interest
in molybdenum chemistry [21]. Chiral molybdenum complexes in
∗ Corresponding author. Tel.: +351 217500196; fax: +351 217500088.
E-mail addresses: mssaraiva@fc.ul.pt (M.S. Saraiva), cmnunes@fc.ul.pt
C.D. Nunes), teresa.nunes@ist.utl.pt (T.G. Nunes), mjc@fc.ul.pt (M.J. Calhorda).
(
0
926-860X/$ – see front matter © 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.apcata.2013.01.033

M.S. Saraiva et al. / Applied Catalysis A: General 455 (2013) 172–182
173
MCM-41 have also been developed for asymmetric oxidation catal-
Further functionalization of MCM-41with molybdenum(II) frag-
ments was carried out by the reaction of 1.0 g of MCM-Pr
with a solution of complex 1 (1.0 mmol, 0.417 g) or complex
2 (1.0 mmol, 0.578 g) in dimethylformamide and triethylamine
(1.5 mmol, 0.207 g) at 273 K. The mixture was stirred for 72 h at
room temperature. The suspension was filtered, washed with 4×
20 mL dichloromethane, and dried under vacuum during 3 h. MCM-
Pr-1 and MCM-Pr-2 were obtained, respectively.
An alternative functionalization was achieved by grafting the
L1 ligand in MCM-41. A solution of L1 ligand (5.0 mmol, 2.17 g) in
toluene was added to a suspension of 1 g of MCM-41 (1) in 30 mL
of toluene, and allowed to reflux for 24 h. The yellow suspension
of MCM-L1 was filtered, washed with 4 × 20 mL dichloromethane,
and dried under vacuum during 3 h.
ysis [22,23].
3
We
have
used
[Mo(␩ -C H5)Br(CO) (CH CN) ]
and
3
2
3
2
[
MoI (CO) (CH CN) ] complexes as precursors for catalytic
2 3 3 2
olefin epoxidation with TBHP, using a variety of N N (bidentate
nitrogen ligands) in homogeneous conditions or immobilized in
different supports, from MCM-41, to clays, or silsesquioxanes,
but it is not yet totally clear why the activity of some systems
increases with immobilization for some ligands, but not for others
[
7,19,24–30].
In this work, we study the reactivity of another bidentate dini-
trogen ligand, 8-aminoquinoline (8aq), bound to Mo(II) in the two
families of complexes, and explored several routes to support the
heterogeneous counterparts in MCM 41. This ligand has been asso-
ciated with several metal centers to catalyze carbonylation and
other reactions [31].
A
suspension of this immobilized MCM-L1 ligand (1.0 g)
3
in dichloromethane could then react with [MoBr(␩ -
C H5)(CO) (CH CN) ] (1.0 mmol, 0.355 g) in dichloromethane.
3
2
3
2
The mixture was stirred overnight, filtered, washed with
dichloromethane, and dried under vacuum during 3 h, to yield
2
. Experimental
MCM-L -1. MCM-L -2 was obtained by a similar reaction with
1
1
2.1. Catalysts preparation
[MoI (CO) (CH CN) ] (1.0 mmol, 0.488 g), using methanol as
2
3
3
2
solvent.
All reagents were obtained from Aldrich and used as
received. Commercial grade solvents were dried and deoxy-
genated by standard procedures (diethyl ether, tetrahydrofuran
and toluene over sodium/benzophenone ketyl; dichloromethane,
2.2. Catalysts characterization
dimethylformamide over calcium hydride and methanol over
FTIR spectra were obtained as KBr pellets (complexes) and Dif-
fuse Reflectance (DRIFT) measurements (materials) on a Nicolet
magnesium), distilled under nitrogen, and kept over 4 A˚ molec-
3
−1
−1
ular sieves. Complexes [Mo(␩ -C H5)Br(CO) (CH CN) ] [12,13],
6700 in the 400–4000 cm range with 4 cm resolution.
Microanalyses for C, H, N, and Mo quantification were performed
at CACTI, University of Vigo (C, H, and N analyses on a Fisons EA
1108, and Mo on a Perkin Elmer Optima 4300DV using In as internal
standard).
3
2
3
2
3
[
[
Mo(␩ -C H5)Br(CO) (L)] [24], [MoI (CO) (CH CN) ] [14], and
3 2 2 3 3 2
MoI (CO) (L)] [25] were prepared according to literature methods.
The complex [Mo(␩ -C H5)Br(CO) (8-aq)] (1) was prepared
2
3
3
3
2
by adding
a
solution of 8-aminoquinoline (8aq) (1.2 mmol,
3
1
13
0.173 g) in dichloromethane to a stirred solution of [Mo(␩ -
H and C solution NMR spectra were obtained with a Bruker
29
13
C H5)Br(CO) (CH CN) ] (1 mmol, 0.355 g) in dichloromethane in
Avance 400 spectrometer. Solid state Si and C NMR measure-
ments were performed at room temperature on a Bruker MSL 300P
spectrometer operating at 59.60 and 75.47 MHz for the observa-
tion of 29Si and C resonances, respectively. The standard magic
angle spinning (MAS) cross polarization–dipolar decoupling RF
pulse sequence (CP–DD) was used under about 4 kHz spinning rate.
For the acquisition of 29Si spectra, 5 ms contact time was chosen,
6 s recycling delay, and a number of scans always higher than 3000
were selected; the Hartmann–Hahn condition was optimized using
tetrakis-trimethylsilylsilane and tetramethylsilane (TMS) was the
external reference to set the chemical shift scale. 13C spectra were
recorded with 2 ms contact time, 4 s recycling delay and a number
of scans higher than 900. The Hartmann–Hahn condition was opti-
mized using glycine, also the external reference to set the chemical
shift scale (13CO at 176.1 ppm).
3
2
3
2
inert atmosphere and allowed to stir for 2 h. The orange solid was
filtered, washed with diethyl ether, and dried under vacuum during
13
3
h (325 mg, 78% yield).
MoI (CO) (8-aq)] (2) was prepared by adding a solution of
[
2
3
8
-aminoquinoline (8-aq) (1.2 mmol, 0.173 g) in methanol to a
stirred solution of [MoI (CO) (CH CN) ] (1 mmol, 0.488 g) also in
2
3
3
2
methanol, in inert atmosphere, and allowing to stir overnight. The
solution was concentrated by removing part of the solvent, and
diethyl ether was added. The brown solid formed was filtered,
washed with diethyl ether, and dried under vacuum during 3 h
(
480 mg, 83% yield).
The modified 8-aminoquinoline ligand (L ) was obtained from
1
the reaction between 3-iodopropyltrimethoxysilane (1.0 mmol,
.196 mL) and a solution of 8-aq (1.0 mmol, 0.144 g) in 30 mL of
0
tetrahydrofuran, after stirring at 339 K during 72 h. The solution
turned orange without formation of any precipitate, and an oil was
formed after removing the solvent (303 mg, yield 70%).
Powder XRD measurements were taken on a Philips Analytical
PW 3050/60 X’Pert PRO (theta/2 theta) equipped with X’Celerator
detector and with automatic data acquisition (X’Pert Data Collec-
tor (v2.0b) software), using a monochromatized Cu-K␣ radiation as
incident beam, 40 kV − 30 mA.
MCM-41 and derivatized materials were synthesized
by adopting
a
methodology previously described, using
[
(C14H33)N(CH ) ]Br as template [8]. Template extraction from
N sorption measurements were obtained in an automatic appa-
3
3
2
MCM-41 was performed by refluxing the materials with HCl
acidified methanol instead of a calcination procedure. Prior to
the grafting experiment, physisorbed water and methanol were
removed by heating at 453 K in vacuum (10 Pa) for 2 h. Two
batches of MCM-41 were prepared for use in each of the pathways
as will be discussed later.
ratus (ASAP 2010; Micromeritics). BET specific surface areas (SBET,
p/p from 0.03 to 0.13) and specific total pore volume, Vp were esti-
◦
mated from N2 adsorption isotherms measured at 77 K. The pore
size distributions (PSD) were calculated by the BJH method using
the modified Kelvin equation with correction for the statistical film
thickness on the pore walls [32,33]. The statistical film thickness
−
2
◦
(
3-Iodoropropyl)trimethylsilyl-MCM-41 (MCM-Pr) resulted
was calculated using Harkins–Jura equation in the p/p range from
from the addition of 2.0 mL of 3-iodopropyltrimethoxysilane to a
suspension of 1 g of MCM-41 in 30 mL of toluene, and reflux for
0.1 to 0.95. Prior to the measurements, samples were degassed and
physisorbed water was removed by heating at 723 K for MCM and
413 K for the derivatized materials (to minimize the destruction
of the functionalities) in vacuum for 2 h 30 m. TGA studies were
2
4 h. The yellow suspension was filtered, washed with 4× 20 mL
dichloromethane, and dried under vacuum during 3 h.

1
74
M.S. Saraiva et al. / Applied Catalysis A: General 455 (2013) 172–182
characterized by FTIR, H and 13C NMR spectroscopy and elemental
1
analysis.
The ꢀ(C O) bands were observed in the FTIR spectra at 1958,
−
1
3
1
2
928, and 1832 cm
in [MoBr(␩ -C H5)(CO) (8-aq)] (1) and at
3 2
−
1
068, 2016, 1923, and 1869 cm
in [MoI (CO) (8-aq)] (2). The
2
3
number of bands indicates that several isomers were formed
−
34,35]. Bands characteristic of the 8-aq ligand appear at 1624 cm
1
[
Scheme 1. Structure and atom numbering scheme of complexes [MoBr(␩3-
C3H5)(CO)2(8-aq)] (1, left) and [MoI2(CO)3(8-aq)] (2, right).
−
1
in 1 and 2, ꢀ(C=N), and at 1505 or 1508 cm , ꢀC C (1 and 2, respec-
tively).
The spectra 1H and 13C NMR spectra (Table 2) show the peaks
assigned to the protons and carbons of the 8-aq ligand shifted from
their positions in the free 8-aq ligand, and those of the allyl ligand of
performed using a Perkin-Elmer TGA7 thermobalance system at a
◦
heating rate of 10 C min-1 under N .
2
1
, confirming the coordination to Mo in complexes 1 and 2. Elemen-
2
.3. Catalytic tests
tal analysis (Table 1, Section 2.2) confirms the proposed structure
shown in Scheme 1.
The organometallic derivatized materials were obtained by
grafting a spacer (3-iodopropyltrimethoxysilane, Pr) in MCM and
allowing it to react with complexes 1 and 2 containing coordinated
Complexes and materials were tested as catalysts in the
epoxidation of cis-cycloctene, styrene, 1-octene, R-(+)limonene,
geraniol, cis-3-hexene-1-ol and trans-2-hexene-1-ol, using tert-
butylhydroperoxide (TBHP) as oxidant. The catalytic oxidation tests
were carried out at 328 K under normal atmosphere in a reaction
vessel equipped with a magnetic stirrer and a condenser. In a typ-
ical experiment, the vessel was loaded with olefin (100%), internal
standard (DBE), catalyst (1%), oxidant (200%), and 3 mL of solvent.
Addition of the oxidant determines the initial time of the reaction.
The course of the reactions was monitored by quantitative GC-
analysis by collecting samples at 10 min, 30 min, 1 h, 1 h 30 min, 2, 4,
8aq (route A), or by synthesizing the new ligand L , by modifying
1
8aq through reaction with Pr, which then reacts with the pre-
3
cursors [MoBr(␩ -C H5)(CO) (CH CN) ] or [MoI (CO) (CH CN) ]
3
2
3
2
2
3
3
2
(
[
route B).
A
third alternative, based on grafting complexes
3
MoBr(␩ -C H5)(CO) (L )] or [MoI (CO) (L )] led to metal poor
3
2
1
2
3
1
materials. Scheme 2 depicts the steps leading to the desired materi-
als. It should be mentioned that the same Mo containing materials
are obtained from both routes: MCM-Pr-1 is the same as MCM-L1-
6
, 8, and 24 h of reaction. These samples were treated as described
1, and MCM-Pr-2 is the same as MCM-L -2. Pr and L identify the
1 1
previously prior to injection in the GC column [25].
spacer grafted in MCM and thus also the synthetic pathway.
Two batches of the pure siliceous matrix MCM-41 (MCM) par-
ent material, obtained by a template approach, were afterwards
derivatized according to the pathways outlined in Scheme 2 [6]. In
pathway A (Scheme 2, top), 3-iodopropyltrimethoxysilane is first
grafted in the inner walls of pristine MCM, leading to MCM-Pr.
Then it reacts with a solution of [Mo(␩ -C H )Br(CO) (8-aq)] (1)
or [MoI (CO) (8-aq)] (2) in tetrahydrofuran at 339 K during 72 h,
to afford the final materials MCM-Pr-1 and MCM-Pr-2, containing
0.98 (0.1 mmol g ) and 0.68 wt% (0.07 mmol g ) of Mo, respec-
tively.
2
.4. Epoxidation of cis-cyclooctene, styrene, 1-octene,
R(+)-limonene, geraniol, cis-3-hexen-1-ol, and trans-2-hexen-1-ol
Epoxidation of cis-cyclooctene, styrene, 1-octene, R(+)-
limonene, geraniol, cis-3-hexen-1-ol, and trans-2-hexen-1-ol
Substrate (800 mg: 7.3 mmol of cis-cyclooctene, 7.7 mmol
of styrene, 7.1 mmol of 1-octene, 5.87 mmol of R(+)-limonene,
3
3
5
2
2
3
−
1
−1
5
.2 mmol of geraniol; 7.99 mmol of cis-3-hexen-1-ol and trans-2-
hexen-1-ol), 800 mg dibutyl ether (internal standard), 1 mol % of
catalyst, 2.65 mL of TBHP (5.5 M in n-decane) and 3 mL of CH Cl .
2
2
Pathway B (Scheme 2, bottom) starts by grafting the MCM parent
material with the ligand (L , inset in Scheme 2) synthesized by reac-
1
3
. Results and discussion
tion of 3-iodopropyltrimethoxysilane with 8aq in tetrahydrofuran
at 339 K for 72 h. The resulting MCM-L material reacted with com-
1
3
3.1. Characterization of catalysts
plexes [MoBr(␩ -C3H5)(CO)2(CH3CN)2] or [MoI2(CO)3(CH3CN)2]
to yield the organometallic materials MCM-L -1 and MCM-L -
1
1
3
The new complexes (Scheme 1) [MoBr(␩ -C H5)(CO) (8-aq)]
2, respectively. The Mo load in these materials was found to be
3
2
−
1
−
1
(
1; 8-aq = 8-aminoquinoline) and [MoI (CO) (8-aq)] (2) were
4.85 (0.5 mmol g ) for MCM-L1-1 and 2.95 wt% (0.3 mmol g ) for
2
3
Table 1
3
Elemental analysis and relevant vibrational frequencies for [Mo(␩ -C3H5)Br(CO)2(8-aq)] (1), [MoI2(CO)3(8-aq)] (2), L1, and materials MCM-Pr, MCM-Pr-1, MCM-Pr-2,
MCM-L1, MCM-L1-1, and MCM-L1-2.
Elemental analysis (%)
C
IR frequencies
␯C
H
N
Mo
O
␯C
N
␯N
H
a
c
b
b
1
2
39.29 (40.31)
26.00 (25.81)
41.12 (41.48)
4.50e
3.08 (3.14)
2.00 (1.91)
4.90 (5.34)
0.95e
6.62 (6.72)
5.38 (4.92)
6.07 (6.45)
1958, 1928, 1832
2068, 2016, 1923
–
–
1920, 1824
1987, 1891, 1863
–
1942, 1852
1950, 1830
1624
1624
1616
3203,3161
3420, 3220
3374, 3202
d
L1
–
–
MCM-Pr
f
MCM-Pr-1
MCM-Pr-2
MCM-L1
MCM-L1-1
MCM-L1-2
18.02
4.46
6.31
5.90
3.01
2.66
2.30
0.98(0.1)
0.68(0.07)
–
4.85(0.5)
2.95(0.3)
1655, 1570
1655, 1571
1638, 1578
1638, 1566
1635, 1560
3452
3434
3429
3409
3441
f
16.70
21.92
4.07
2.84
g
g
f
17.04
18.00
2.12
2.70
f
a
C14H13MoN2O2Br: found (calc.)
Not determined.
b
c
d
e
f
C12H8MoN2I2: found (calc.)
C15H23O3N2ISi: found (calc.)
−1
Organic loading 0.6 mmol g
−1
Mo loading mmol g
.
g
−1
.
Organic loading 1.0 mmol g

M.S. Saraiva et al. / Applied Catalysis A: General 455 (2013) 172–182
175
Table 2
1
13
3
H and C NMR chemical shifts of free 8-aq, [MoBr(␩ -C3H5)(CO)2(8-aq)] (1) and [MoI2(CO)3(8-aq)] (2) complexes.
H2
H3
H4
H5
H6
H7
NH
Allyla
8
1
2
-aq
-aq
6.93
8.1
8.65
7.31–7.37
7.6
7.79
7.15
8.0
7.71
8.05
7.7
7.71
7.31–7.37
7.9
7.94
8.75
8.7
9.04
4.98
9.8
n.o.
–
1.16/3.57/4.23
–
C1
C2
C3
C4
C5
C6
C7
C8
C9
Allyla
–
54.2/55.8/73.7
–
8
1
2
144.0
141.0
140.2
110.0
127.0
141.5
127.4
128.0
130.9
116.0
122.5
123.8
136.0
131.0
129.3
121.3
124.0
128.3
147.4
140.0
149.9
138.4
146.0
147.0
128.8
129.0
131.6
a
Most abundant isomer.
◦
◦
MCM-L -2.These results indicate that although both pathways lead
reflections in the 2 < 2ꢁ < 10 range, indexed to a hexagonal cell as
1
to the same final materials, pathway B is more efficient as higher
(1 0 0), (1 1 0), (2 0 0), and (2 1 0). The d value of the (100) reflection is
√
amounts of Mo are introduced in the parent MCM.
37.9 A˚ , corresponding to a lattice constant of a = 44.3 A˚ (=2d1 0 0
/
3).
−
1
The C and N contents indicated a loading of 1 mmol g in MCM-
L , based on the C/N ratio of 7.28. The analysis obtained after
The observed peak intensity reduction is common to all mate-
rials and is even more significant in the materials with Mo. This
is not due to a crystallinity loss, but rather to an X-ray scattering
contrast reduction between the silica walls and pore-filling mate-
rial. This has been observed for other types of materials and is well
described in the literature [36–38]. The relevant textural properties
of all materials are collected in Table 3.
Nitrogen adsorption studies at 77 K revealed that the pris-
tine MCM sample exhibits a reversible type IV isotherm (Fig. 2),
characteristic of mesoporous solids (pore width between 2 and
50 nm, according to IUPAC) [39]. The calculated textural parame-
ters (SBET and Vp) of this material agree with literature data (Table 2)
1
−
1
introduction of molybdenum indicate Mo loadings of 0.5 mmol g
−1
for MCM-L -1 and 0.3 mmol g for MCM-L -2, respectively (C/N
1
1
ratios 8.04 for 1 and 7.83 for 2). This means that the metal precur-
sor did not react with all the available L1 ligands. However, more
molybdenum is immobilized in this way than following the alter-
native route leading to MCM-Pr-1 and MCM-Pr-2, containing only
−1
−1
0
.1 mmol g and 0.07 mmol g of Mo.
The powder XRD patterns of MCM-Pr-1 and MCM-L -1 are
1
given in Fig. 1 (those of MCM-Pr-2 and MCM-L -2 are not shown).
1
The pattern of the parent, calcined MCM material, shows four
Scheme 2. Structure of ligand L1 and reaction scheme for preparation of heterogeneous catalysts.

1
76
M.S. Saraiva et al. / Applied Catalysis A: General 455 (2013) 172–182
Fig. 1. XRD powder patterns of mesoporous materials MCM-Pr, MCM-Pr, MCM-Pr-1, MCM-L1, and MCM-L1-1 prepared from pathway A (left) and B (right).
Table 3
Textural parameters for MCM host and derivatized materials from N2 isotherms at 77 K and powder XRD data.
SBET/m2
g
−1
ꢂSBETa/%
Vp/cm3
g
−1
ꢂVpb/%
dBJH (nm)
Sample
d1 0 0/Å
a/Å
MCM
37.9
38.6
38.6
38.6
39.1
40.3
39.4
39.0
44.3
44.0
44.0
44.3
45.0
44.4
43.8
44.4
1040
886
801
887
1036
450
124
115
−
−15
−23
−15
−
0.920
0.729
0.640
0.725
0.914
0.260
0.112
0.095
−
−21
−30
−21
−
3.50
3.36
3.36
3.28
3.45
2.67
2.68
2.67
MCM-Pr
MCM-Pr-1
MCM-Pr-2
MCM
MCM-L1
MCM-L1-1
MCM-L1-2
−57
−88
−89
−72
−88
−90
a
Surface area variation relative to parent MCM.
Total pore volume variation relative to parent MCM.
b
[
0
40,41]. The capillary condensation/evaporation steps appear in the
.20–0.40 relative pressure range, and the sharpness of this step
which is related to the volume of pore space confined by absorbate
film on the pore walls, is smaller in the case of the modified materi-
als (Fig. 2). Furthermore, the maxima of the PSD curves determined
by the BJH method, dBJH, decrease concomitantly (Table 3).
Fig. 3 shows the thermogravimetric analyses (TGA) of the mate-
rials involved in pathway B in (Scheme 2). All materials experience
mass losses corresponding to physisorbed water until 398 K, and
start to decompose at higher temperatures.
reflects the uniform pore size.
For the first set of materials in Table 3 (synthetic pathway A),
the isotherm of the functionalized material MCM-Pr reveals much
lower N2 uptake, accounting for a decrease in Vp (21%) and a con-
comitant variation in SBET (15%) compared to the parent MCM. This
trend is observed after introduction of complex 1, but there is no
change when going from MCM-Pr to MCM-Pr-2 suggesting that
the amount of complex 2 introduced in the material must be very
small, as determined from elemental analysis.
The MCM-L1 material displays a degradation process occurring
between 398 and 1073 K, with loss of 21.1% weight, corresponding
−
1
to a 0.85 mmol g of the ligand inside the MCM pores. This value
−
1
For the second set of materials (pathway B), the isotherm of
the functionalized material MCM-L reveals much lower N uptake
is in agreement with the total amount of L1 (1 mmol g ) obtained
from elemental analysis.
1
2
than in MCM-Pr accounting for a large decrease in Vp (72%) and a
concomitant variation in SBET (57%). Introduction of the precursor
complexes inside the pores induces a further decrease of both Vp
and SBET (88–90%), in agreement with the decrease of the p/p coor-
dinates of the inflection points of the isotherms after post-synthesis
treatments [42]. The height of the capillary condensation steps,
MCM-L -1 and MCM-L -2 undergo loss of the organometal-
1 1
lic moiety between 398 and 1073 K, corresponding to ca. 26.5%,
−
1
−1
(MCM-L -1) and ca. 23.3%, or 0.4 mmolMo g
1
or 0.6 mmolMo g
(MCM-L -2). These values agree quite well with those from
◦
1
−
1
elemental analysis, namely 0.5 mmolMo g
0.3 mmolMo g (MCM-L -2) respectively.
(MCM-L -1) and
1
−
1
1
700
600
500
400
300
200
100
0
0.4
0
.35
.3
.25
.2
0.15
.1
0
0
MCM
MCM-L1
MCM-L1-1
MCM
0
MCM-L1
MCM-L1-1
0
0.05
0
0
0.2
0.4
0.6
0.8
1
2.4
3.4
4.4
5.4
6.4
p/pº
Diameter / nm
Fig. 2. Nitrogen adsorption isotherms of materials MCM, MCM-L1, and MCM-L1-1 at 77 K (left); pore size distribution of MCM, MCM-L1, and MCM-L1-1 (right).

M.S. Saraiva et al. / Applied Catalysis A: General 455 (2013) 172–182
177
1
00
9
5
0
5
0
5
0
5
0
5
9
8
8
7
7
6
6
5
MCM
MCM-L1
MCM-L1-1
MCM-L1-2
300
500
700
T / K
900
1100
Fig. 4. 13C CP MAS-DD spectra of MCM-L1 (bottom) and MCM-L1-1 (top).
Fig. 3. TGA profiles of MCM, MCM-L1, MCM-L1-1 and MCM-L1-2.
[
MoI (CO) (CH CN) ], those proton signals are observed at 9.2
2 3 3 2
(
Si CH ), 22.9 (CH2 CH2 CH ) and 73.4 (HN CH ), and the car-
2
2
2
Similar profiles are obtained for materials MCM-Pr-1 and
MCM-Pr-2. After the release of physisorbed water before 398 K
bons at 123.4, 129.8, 143.0 and 173.8 ppm.
29
The Si CP MAS-DD NMR spectrum of pristine MCM, displays
three resonances at −109.2,−100.6 and −91.4 ppm, characteristic
−
TGA curves not shown), weight losses of 0.11 mmolMo g
1
(
−
1
4
3
2
and 0.08 mmolMo g , respectively, were found, comparable to
of Q , Q and Q species of the silica framework, respectively (Fig. 5).
−1
−1
(MCM-Pr-2)
In MCM-L , the relative size of the Q , Q and Q2 signals
4
3
0
.10 mmolMo g
obtained from elemental analysis.
The presence of the immobilized [MoBr(␩ -C H5)(CO) (L )] and
(MCM-Pr-1) and 0.07 mmolMo g
1
changes, as both Q and Q2 species disappear owing to the reac-
3
3
3
2
1
tion with L . At the same time three broad signals at −65.0, −55.0
1
and −51.5 ppm, assigned to T , T and T1 organosilica species,
3
2
[
MoI (CO) (L )] complexes in the materials was probed by FTIR
2 3 1
spectroscopy.
appear in the spectrum. Introduction of the organometallic frag-
3
In the material MCM-Pr containing grafted 3-iodopropyltri-
methoxysilane (Pr), the stretching vibrations modes of the meso-
ments [MoBr(␩ -C H )(CO) (CH CN) ] and [MoI (CO) (CH CN) ]
3
5
2
3
2
2
3
3
2
29
does not significantly change the Si CP MAS-DD NMR spectra,
besides a slight shift of the signals, as the reaction occurs with the
immobilized L1 and should not involve the material walls.
porous framework (Si
O
Si) are observed at around 1243, 1079,
−1
9
56 cm , as in the parent MCM, while the new bands at ca. 2982
−1
and 2853 cm are assigned to the ␯(C H) stretching modes of the
aliphatic propyl chain. The introduction of complexes 1 and 2 is
confirmed by the presence of ␯(C O) stretching modes observed at
3
3
.2. Catalytic studies
−1
−1
1
920 and 1824 cm for MCM-Pr-1, and 1987, 1891 and 1863 cm
.2.1. Oxidation of substrates
Both complexes, [MoBr(␩ -C H5)(CO) (8-aq)] (1) and
for MCM-Pr-2. These bands are slightly shifted from their position
in the complexes, as the L1 ligand differs from 8aq, and they are
3
3
2
[MoI (CO) (8-aq)] (2), and the final materials, MCM-Pr-1, MCM-
Pr-2, MCM-L -1 and MCM -L -2, were tested as catalyst precursors
2
3
−
1
broader. The observation of ␯(C N) modes at 1654 and 1570 cm
(
1
1
−1
MCM-Pr-1) and at 1655 and 1571 cm (MCM-Pr-2) further con-
firms the coordination of the metal fragments to the material.
When following pathway B, the L1 ligand is grafted in MCM.
−
1
Its presence is attested by the new bands at 2945 and 2850 cm
,
assigned to the ꢀ(C H) stretching of the aliphatic linear chain
and aromatic C H of the ring, as well as ꢀ(C N) modes at 1638
−1
and 1578 cm . The strong stretching vibrational modes of the
mesoporous framework (Si Si) of MCM-L1 are observed at ca.
O
−1
1
236, 1080, and 955 cm , as in the parent MCM. The reaction
with the precursor complexes [MoBr(␩ -C H5)(CO) (CH CN) ] and
3
3
2
3
2
[
1
MoI (CO) (CH CN) ] leads to new ␯(C O) bands at 1942 and
2 3 3 2
− −1
852 cm (MCM-L -1), and at 1950 and 1830 cm (MCM-L -2),
1 1
1
slightly shifted from the values in the free complexes.
The materials were also characterized by 13C CP MAS-DD and
Si CP MAS-DD solid state NMR.
29
As an example, the 13C CP MAS-DD spectrum of MCM-L1
presents the resonances of the propyl chain at 8.9 (Si CH ),
2
2
2.5 (CH2 CH2 CH ) and 73.3 (HN CH ), and four signals at
2
2
1
21.8, 130.7, 146.8 and 176.0 ppm corresponding to aromatic
3
carbons from the L1 ligand. After reaction with [MoBr(␩ -
C H5)(CO) (CH CN) ], the resulting MCM-L -1 shows the three
3
2
3
2
1
resonances of the propyl chain protons slightly shifted, at 8.4
Si CH ), 23.0 (CH -CH -CH ) and 72.6 (HN CH ), as well as
(
2
2
2
2
2
the carbon resonances at 123.1, 129.2, 142.9, 176.2 ppm (Fig. 4).
In the material MCM-L -2, obtained from the reaction with
Fig. 5. 29Si CP MAS-DD spectra of MCM (bottom), MCM-L1 (center) and MCM-L1-1
(top).
1

1
78
M.S. Saraiva et al. / Applied Catalysis A: General 455 (2013) 172–182
Table 4
Olefin oxidation catalyzed by [Mo(␩ -C3H5)Br(CO)2(8-aq)] (1), [MoI2(CO)3(8-aq)] (2), MCM-Pr-1, MCM-Pr-2, MCM-L1-1 and MCM-L1-2.
3
Entry
Epoxidea
Catalyst
Conversionb,c (%)
Selectivityc,d (%)
1
2
3
4
5
6
1
2
74
55
53
45
48
26
100
100
100
100
100
100
MCM-Pr-1
MCM-Pr-2
MCM-L1-1
MCM-L1-2
O
7
8
9
0
1
2
1
2
92 (86)
38 (14)
17 (22)
18 (4)
94 (81)
99 (14)
63 (84)
3 (54)
56 (34)
21 (34)
64 (76)
17 (60)
MCM-Pr-1
MCM-Pr-2
MCM-L1-1
MCM-L1-2
1
1
1
O
13
14
15
16
17
18
1
2
5
11
4
2
5
90
11
46
32
8
MCM-Pr-1
MCM-Pr-2
MCM-L1-1
MCM-L1-2
O
2
53
19
20
21
22
23
24
1
2
69
92
75
34
30
28
100
96
100
80
100
70
HO
HO
MCM-Pr-1
MCM-Pr-2
MCM-L1-1
MCM-L1-2
O
25
26
27
28
29
30
1
2
58
45
14
14
4
100
100
100
100
100
100
HO
MCM-Pr-1
MCM-Pr-2
MCM-L1-1
MCM-L1-2
O
HO
6
31
32
33
34
35
36
1
2
60
17
25
19
25
28
50
73
82
61
83
82
MCM-Pr-1
MCM-Pr-2
MCM-L1-1
MCM-L1-2
O
e
37
38
39
40
41
42
1
2
94
96
82
68
79
51
73
53
37
87
62
73
e
HO
HO
e
MCM-Pr-1
MCM-Pr-2
MCM-L1-1
MCM-L1-2
e
f
O
f
a
All reactions carried out in CH2Cl2 in the presence of 200 mol% oxidant (TBHP) and 1 mol% of molybdenum catalyst at 328 K.
Calculated after 24 h.
Values in parentheses refer to 8 h reaction time.
b
c
d
e
f
Calculated as “Yield of Epoxide”/“Conversion” × 100% after 24 h.
6
6
,7-Epoxide formed as by-product.
,7-Epoxide and geranial formed as by-products.
for olefin epoxidation using a set of substrates (cis-ciclooctene,
styrene, 1-octene, R(+)-limonene, geraniol, trans-hex-2-en-1-ol,
cis-hex-3-en-1ol), and tert-butyl hydroperoxide (TBHP in decane)
as oxygen donor, in dichloromethane, at 328 K (Table 4, see details
in Section 2). No reaction took place in the absence of a catalyst.
All the tested catalysts oxidized cis-cyclooctene 100% selec-
entries 8, 10 and 12). The choice of catalyst and reaction time may
thus determine the nature of the product.
Oxidation of 1-octene was carried out very inefficiently with all
the catalysts (Table 4, entries 13–18), 11% being the highest con-
version observed for complex 2 (entry 14). The highest selectivity
toward the epoxide was observed with complex 1 as catalyst (entry
13), following the trend described for cis-cyclooctene and styrene.
The epoxidation of trans-hex-2-en-1-ol is promoted by all cata-
lysts with fair to good substrate conversion. Catalysts 1, MCM-Pr-1
3
tively to the epoxide. The catalysts containing MoBr(␩ -C H5)(CO)
3
2
(
1, MCM-Pr-1 and MCM-L -1) show the highest conversions and
1
homogeneous catalysts lead to higher conversions than heteroge-
neous ones.
and MCM-L -1 were completely selective to the epoxide (Table 2,
1
The oxidation of styrene leads to the epoxide and benzaldehyde,
this one arising from oxidative clivage of the epoxide [30], and
conversions are relatively low. The catalysts with the allyl ligand
entries 19, 21 and 23), while 2, MCM-Pr-2 and MCM-L -2 yielded
epoxides with lower selectivity and conversions (Table 4, entries
20, 22, 24).
1
(
1, MCM-Pr-1 and MCM-L -1) have higher selectivity toward for-
Cis-hex-3-en-1-ol is selectively oxidized to the epoxide (100%
for all catalysts), but the conversions are very low for the hetero-
geneous catalysts (4–14%). Only the homogeneous catalysts 1 and
2 show fair substrate conversions of 58% and 45%, respectively.
1
mation of the epoxide, while the others lead almost exclusively
to substrate over-oxidation yielding benzaldehyde. Indeed, the
amount of epoxide formed after 8 h is larger than after 24 h (Table 4,

M.S. Saraiva et al. / Applied Catalysis A: General 455 (2013) 172–182
179
Table 5
Olefin oxidation catalyzed by [Mo(␩ -C3H5)Br(CO)2(8-aq)] (1) changing solvent and temperature.
3
Entry
Epoxide
Solvent
Temperature (K)
Conversiona,b (%)
Selectivityb,c (%)
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
cis-Ciclooctene
CH2Cl2
–
–
CH2Cl2
CH3CN
Toluene
–
328
358
393
328
358
393
328
358
393
328
358
393
328
358
393
328
358
393
328
358
393
74
100
100
100
90
87
99
88
99
99
100
99
100 (6 h)
100 (10 min)
5
30
87
1-Octene
13
–
–
30 (2 h)
81 (1.5 h)
69
100 (6 h)
100 (10 min)
58
100 (6 h)
100 (1.5 h)
60
92 (1 h)
99 (10 min)
94
1
1
1
1
1
1
1
1
1
1
2
2
cis-Hex-3-en-1ol
trans-Hex-2-en-1-ol
Geraniol
CH2Cl2
–
–
CH2Cl2
–
–
CH2Cl2
–
–
CH2Cl2
–
–
99
100
100
100
50 (39)
35 (61)d
17 (83)d
73
R(+)-Limonene
61 (2 h)
98 (10 min)
100
100
a
b
c
All reactions carried out in the presence of 200 mol% oxidant and 1 mol% of molybdenum catalyst.
Calculated after 24 h, unless otherwise stated.
Calculated as “yield of epoxide”/”conversion” × 100% after 24 h.
d
Values in parentheses refer to di-epoxide yield.
The oxidation of R(+)-limonene followed the same trend as
trans-hex-2-en-1-ol, with low substrate conversion (below 50%)
for all catalysts, complex 1 being the best (60%). The two materials
with the allyl ligand (MCM-Pr-1 and MCM-L -1) showed higher
1
conversions. All catalysts yielded the endocyclic epoxide as the
major product with selectivity above 50%. Complex 1 shows the
highest conversion and the lowest selectivity (Table 4, entry 31).
Remarkably this catalyst also yields the di-epoxide (in both endo-
and exo-cyclic positions) with 39% selectivity.
Complexes 1 and 2 display very high conversions in the oxi-
dation of geraniol, but they become lower in the heterogeneous
catalyst (Table 4, entries 37 –42). The selectivities, however, are
not very good. 1, MCM-Pr-2, MCM-L -1 and MCM-L -2 yield
1
1
geraniol-2,3-epoxide as the major product (shown in Table 2) with
selectivity ranging from 62% to 87%. The by-product is geranial (all-
trans product). MCM-Pr-1 converts geraniol into the 6,7-epoxide
3
Fig. 6. cis-Cyclooctene epoxidation promoted by [Mo(␩ -C3H5)Br(CO)2(8-aq)] (1)
in CH2Cl2 at 328 K with different oxidants. All reactions carried out in the presence
of oxidant (200 mol%) and 1 mol% of molybdenum catalyst.
(
major product with 64% selectivity and 52% conversion), while
complex 2 results in a ∼1:1 ratio of both epoxides.
The catalysts were not very active, since conversions only
exceeded 90% in very few examples. Cis-cyclooctene and cis-hex-3-
en-1-ol were the only substrates selectively epoxidized (100%) by
all catalysts. Only the catalysts containing 1 have the same perfor-
mance in the oxidation of trans-hex-2-en-1-ol, the others leading
to other products.
Epoxidation of R(+)-limonene and geraniol occurred with fair to
good substrate conversion, and mixtures of 2,3- and 6,7-epoxides
were obtained. The major products were the endocyclic epoxide
limonene) and the 2,3-epoxide (geraniol). The diepoxide was also
obtained in the oxidation of limonene (entry 31) and geranial in
(
Table 7
3
cis-Cyclooctene epoxidation catalyzed by [Mo(␩ -C3H5)Br(CO)2(8-aq)] (1) in
Table 6
CH2Cl2at 328 K with variable oxidant:substrate ratio.
3
cis-Cyclooctene epoxidation catalyzed by [Mo(␩ -C3H5)Br(CO)2(8-aq)] (1) in
CH2Cl2with different oxidants.
Oxidant
Ratio
Conversiona,b (%)
Selectivityb,c (%)
TBHP
TBHP
TBHP
TBHP
TBHP
CHP
1.00
1.25
1.50
1.75
2.00
1.00
1.50
2.00
92
97
99
100
74
77
100
100
100
100
100
100
100
100
Oxidant
Temperature (K)
Conversiona,b (%)
Selectivityb,c (%)
TBHP
TBHP
TBHP
H2O2
CHP
328
333
338
328
328
333
338
74
77
86
11
95
100
100
100
100
100
100
100
d
CHP
CHP
94/88/74
95
CHP
CHP
100
97
a
All reactions carried out in CH2Cl2 in the presence of oxidant and 1 mol% of
molybdenum catalyst at 328 K.
a
All reactions carried out in CH2Cl2 in the presence of oxidant (200 mol%) and
mol% of molybdenum catalyst.
b
1
Calculated after 24 h reaction.
b
c
Calculated after 24 h, unless otherwise stated.
Calculated as “Yield of Epoxide”/“Conversion” × 100% after 24 h.
Calculated as “Yield of Epoxide”/“Conversion” × 100% after 24 h.
c
d
Values refer to 1st, 2nd and 3rd recycling runs.

1
80
M.S. Saraiva et al. / Applied Catalysis A: General 455 (2013) 172–182
3
Fig. 7. cis-Cyclooctene epoxidation promoted by [Mo(␩ -C3H5)Br(CO)2(8-aq)] (1) in CH2Cl2 at 328 K with variable oxidant:substrate ratio using as oxidant TBHP (left) and
CHP (right). All reactions carried out in the presence of 1 mol% of molybdenum catalyst.
the oxidation of geraniol with selectivity up to 30% for both. Oxi-
dation of styrene led to benzaldehyde as major product, arising
from oxidative cleavage of the epoxide, but at shorter (8 h) reac-
tion times, epoxide is generally the major product. The worst results
were found in the oxidation of 1-octene, which was not achieved
by any of the six catalysts.
is raised to 358 K, the substrate conversion reaches 30% after 2 h
reaction; at 393 K, conversion is 81% after 1.5 h.
This is an outstanding result since this is a non activated terminal
olefin in a solvent free, greener process.
The same solventless conditions were tested for the other olefins
(Table 5) and the same trend holds. Indeed, without added solvent
and at 393 K, after 10 min all reactions become almost complete
and epoxide selectivity reaches 100% for the majority of substrates
(only cis-hex-3-en-1-ol requires 1.5 h, entry 15). The only exception
in this seems to be R-(+)limonene for which formation of the di-
epoxide product is favored with increasing temperature. This can
be seen as an important achievement since both limonene epoxide
and di-epoxide products are industrially relevant in the production
of low odor products such as metal coatings, varnishes, printing
inks and UV curing applications.
In general the homogeneous catalysts performed better than
3
the heterogeneous ones and systems containing the MoBr(␩ -
C3H5)(CO)2 fragment (1, MCM -Pr-1, MCM-L -1) better than the
1
others.
3.2.2. The influence of reaction conditions and catalyst recycling
In an attempt to improve the conversion and selectivity in
the oxidation of the substrates studied, the effect of solvent,
3
temperature and oxidant were tested using complex [Mo(␩ -
3
C H5)Br(CO) (8-aq)] (1) as catalyst precursor, since it was the most
Complex [Mo(␩ -C3H5)Br(CO)2(8-aq)] (1) was also bench-
3
2
promising one. Styrene was not considered in this study.
From the results in Table 5, it is obvious that reactions in CH Cl2
at 328 K lead to low substrate conversions. In order to increase the
marked as catalyst precursor in the epoxidation of cis-cyclooctene
using three oxidants, tert-butyl hydroperoxide (TBHP), hydrogen
peroxide (H2O2) and cumene hydroperoxide (CHP). The catalytic
reactions with TBHP and CHP were performed in CH2Cl2 at 328 K,
333 K and 338 K, and with H2O2 in acetonitrile at 328 K. In this case
acetonitrile was chosen since the water formed as by-product of
the reaction of H2O2 would lead to a bi-phasic system with CH2Cl2.
The results are summarized in Table 6.
2
temperature, solvents with higher boiling points, such as CH CN
3
and toluene, were tried. The other possibility was to use the sub-
strate as solvent, which allowed higher temperatures and higher
concentration.
The most spectacular effect is observed in the oxidation of
1
4
-octene. Three different solvents were tested (Table 5, entries
–6) and the substrate conversion raises from 5 to 30 and 87%
H2O2 is the oxidant leading to the worst results, with only 11%
substrate conversion, although epoxide selectivity is complete. In
TBHP, yields range from 74% to 86% and selectivity is kept with
increasing temperature. When the oxidant is CHP, conversions are
frankly better, ∼20% above those observed with TBHP. Indeed con-
version is at least 95% and selectivity is always complete for the
epoxide. It can be seen in Fig. 6 that in CHP reaction kinetics is
faster than in TBHP. For instance, at 328 K, substrate conversion is
when moving from CH Cl2 (328 K) to CH CN (358 K) and toluene
2
3
(
393 K). The selectivity toward the epoxide does not rise so much,
but reaches 99% in toluene. Entries 7–9 show that increasing the
temperature without adding solvent has the same effect both in
conversion and selectivity, but the time to achieve the same con-
version drops significantly. For instance, when the temperature
Table 8
Olefin oxidation catalyzed by MCM-L1-1 with TBHP or CHP.
Entry
1
Olefin
Solvent
CH2Cl2
Oxidant
TBHP
Temperature (K)
328
Conversiona,b (%)
48 (1st run)
Selectivityb,c (%)
cis-Ciclooctene
1
1
6 (2nd run)
4 (3rd run)
1
00
6 (4th run)
100 (1st run)
d
2
cis-Ciclooctene
–
TBHP
393
1
1
00 (2nd run)
00 (3rd run)
100
100 (4th run)
3
4
5
6
cis-ciclooctene
1-Octeno
1-Octeno
CH2Cl2
Toluene
–
–
CHP
328
393
393
393
31
44
53
57
100
97
83
TBHP
TBHP
TBHP
cis-Hex-3-en-1ol
100
a
All reactions carried out in the presence of 200 mol% oxidant and 1 mol% of molybdenum catalyst.
Calculated after 24 h, unless otherwise stated.
b
c
Calculated as “Yield of epoxide”/“Conversion” × 100% after 24 h.
d
Achieved after 2 h.

M.S. Saraiva et al. / Applied Catalysis A: General 455 (2013) 172–182
181
from 48 to 16% (Fig. 8 bottom). Nevertheless, at higher temperature
but no added solvent, though leaching is likely to occur, the cata-
lyst is able to adjust its performance although at the sake of slower
kinetics. Epoxidation of cis-ciclooctene was also tested using CHP
(Table 8, entry 3) as oxidant, but the performance was even lower
than with TBHP under similar conditions (Fig. 8 top).
The same material was also used in the epoxidation of 1-octene
(Table 6, entries 4 and 5) at temperatures higher than 328 K (Table 4,
entry 17). The results are once again remarkable. From an almost
inactive system at 328 K in CH Cl , the conversion increased from
2
2
4
% to 44% in toluene at 393 K, while selectivity increased from 8%
to excellent 97%. In the absence of added solvent, the conversion
remains high at 393 K, but the selectivity drops. Nonetheless, these
achievements demonstrate that 1-octene, a terminal linear olefin,
which is therefore inactivated, requires more extreme conditions
to suffer epoxidation.
Epoxidation of cis-hex-3-en-1ol also benefits from running the
experiments at higher temperature, since conversion increases
from as little as 4% (Table 4, entry 29) to 57% (Table 8, last entry)
keeping complete selectivity toward the epoxide.
4. Conclusions
3
The new Mo(II) complexes [MoBr(␩ -C H5)(CO) (8 -aq)] (1) and
3
2
[
MoI (CO) (8-aq)] (2) were synthesized, characterized and immo-
2 3
bilized in MCM-41, using a 3-iodopropyltrimethoxysilane spacer.
Two procedures were adopted. In the first, the spacer reacted with
Fig. 8. cis-Cyclooctene epoxidation catalyzed by MCM-L1-1 in CH2Cl2 at 328 K with
TBHP or CHP as oxidants (top). Recycling experiments of cis-cyclooctene epoxidation
catalyzed by MCM-L1-1 with TBHP at 328 K in CH2Cl2, and at 393 K without solvent
the 8-aq ligand forming L , which was directly supported in the
1
(
bottom).
material and then allowed to react with the parent complexes.
In the second approach, the spacer was immobilized and reacted
directly with complexes 1 and 2. This last procedure was much less
efficient in terms of metal load inside the material, although the
functionalized materials were the same according to their charac-
terization.
The complexes and the new materials were tested as the cat-
alytic precursors in the epoxidation of several substrates, showing
in most cases a high selectively toward the epoxide, but usually low
conversions, the best results being obtained by complex 1 (homo-
geneous conditions). Careful choice of conditions showed that the
conversions could be improved up to a certain extent, by using the
substrate as solvent and varying the oxidant:substrate ratio. TBHP
seems the best oxidant to use with these catalysts. The materials
exhibit in general a lower catalytic activity than the complexes (see
Tables 4–6).
2
1% for TBHP and 57% for CHP after 1 h, accounting for the benefi-
cial use of CHP. This shows that CHP is a better oxidant to be used
with complex 1 as pre-catalyst for cis-cyclooctene epoxidation.
The oxidant:substrate ratio was also explored for the epoxida-
tion of cis-cyclooctene in the presence of 1 in CH Cl at 328 K, using
2
2
either TBHP or CHP as oxidants. The oxidant:substrate ratio ranged
between 1 and 2. The results are shown in Table 7 and Fig. 7.
With TBHP, substrate conversion increases slightly as the ratio
goes from 100 to 175, but at a ratio of 2 significantly lower con-
version and slower kinetics of the reaction are obtained (Fig. 7),
indicating that possibly a large excess of the TBHP can slow down
the reaction. This may be related to a large quantity of t-butanol
formed as a by-product of TBHP reaction and its competition with
TBHP for the catalyst leading to deactivation.
While we cannot rule out diffusion problems that may prevent
the substrate from reaching active catalyst sites inside the pores, we
believe that deactivation routes may involve the NH groups of the
catalyst. Hydrogen bond networks may be formed with the catalyst,
silanol groups still present in the material walls, and the oxidant
When the oxidant is CHP, the trends are different. Substrate con-
version always increases with the increase of the oxidant:substrate
ratio from 1 to 2 and this increase is rather large from 1 to 1.5.
Recycling experiments conducted for the CHP oxidant with an oxi-
dant:substrate ratio set at 1.5 indicate that the catalytic system
shows recycling capability although it suffers from deactivation, as
substrate conversion drops from 94% to 74% after the third cycle.
A similar study was performed using one heterogeneous catalyst
(an alcohol). This effect has been observed before, but in the homo-
geneous catalyst only the NH groups and the alcohol interact in this
way [27,29]. Interestingly, in the system where the heterogeneized
catalyst is more active than the homogeneous analogue there are no
such groups present [25]. Recent studies, however, have shown that
a number of molybdenum compounds, besides the active species,
may result from the reaction of TBHP on analogues of complexes 1
containing complex 1, MCM-L -1, three substrates (cis-ciclooctene,
1
1
-octene, and cis-hex-3-en-1ol), two oxidants (TBHP and CHP),
and varying the solvent. The reutilization of the catalyst was also
checked. All the results are collected in Table 8.
ꢀ
and 2, with 2,2 -bipyridine ligands. Such decomposition may also
The epoxidation of cis-ciclooctene without added solvent at
take place under milder catalytic conditions, contributing to a deac-
tivation of the catalyst [43].
3
93 K (Table 8, entry 2) results in a surprising performance com-
pared to the standard conditions (Table 8, entry 1). In fact, 100%
conversion is achieved after as little as 2 h in fresh use, while, on
the other hand, the catalyst can be recycled without loss of activity
for four cycles (calculated after 24 h). These results are outstanding
compared to those reported in entry 1 of Table 8, and suggest that
the solvent plays an important role in the catalyst deactivation. In
fact, Mo elemental analysis reflected severe leaching during the first
catalytic cycle, which certainly accounts for the loss of conversion
Acknowledgements
The authors thank FCT, POCI and FEDER (projects PTDC/QUI/
71576/2006, PEst-OE/QUI/UI0612/2011, PEst-OE/QUI/UI0100/
2011) for financial support. MSS (SFRH/BD/48640/2008) thanks
FCT for a research grant.

1
82
M.S. Saraiva et al. / Applied Catalysis A: General 455 (2013) 172–182
References
[23] A.J. Burke, Coord. Chem. Rev. 252 (2008) 170–175.
[
24] M.S. Saraiva, S. Quintal, F.C.M. Portugal, T.A. Lopes, V. Félix, J.M.F. Nogueira,
M. Meireles, M.G.B. Drew, M.J. Calhorda, J. Organomet. Chem. 693 (2008)
[1] M.H. Valkenberg, W.F. Hölderich, Catal. Rev. 44 (2002) 321–374.
[2] A. Taguchi, F. Schüth, Micropor. Mesopor. Mater. 77 (2005) 1–45.
[3] C.T. Kresge, M.E. Leonowicz, W.J. Roth, J.C. Vartuli, J.S. Beck, Nature 359 (1992)
3411–3418.
[
[
[
[
[
[
[
25] M. Vasconcellos-Dias, C.D. Nunes, P.D. Vaz, P. Ferreira, P. Brandão, V. Félix, M.J.
Calhorda, J. Catal. 256 (2008) 301–311.
26] M.S. Saraiva, N.L. Dias Filho, C.D. Nunes, P.D. Vaz, T.G. Nunes, M.J. Calhorda,
Micropor. Mesopor. Mater. 117 (2009) 670–677.
7
10–712.
[
4] 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. 114 (1992) 10834–10843.
27] M.S. Saraiva, C.D. Nunes, T.G. Nunes, M.J. Calhorda, J. Mol. Catal. A 321 (2010)
92–100.
[5] C.D. Nunes, M. Pillinger, A.A. Valente, I.S. Gon c¸ alves, J. Rocha, P. Ferreira, F.E.
Kühn, Eur. J. Inorg. Chem. (2002) 1100–1107.
28] M. Vasconcellos-Dias, S.R.M.M. de Aguiar, C.D. Nunes, P.D. Vaz, T.G. Nunes, M.J.
Calhorda, Curr. Inorg. Chem. 1 (2011) 146–155.
29] N.L.D. Filho, F.C.M. Portugal, J.M.F. Nogueira, P. Brandão, V. Félix, P.D. Vaz, C.D.
Nunes, M.J. Calhorda, Organometallics 31 (2012) 4495–4503.
[
6] T.A. Fernandes, C.D. Nunes, P.D. Vaz, M.J. Calhorda, P. Brandão, J. Rocha, I.S.
Gon c¸ alves, A.A. Valente, L.P. Ferreira, M. Godinho, P. Ferreira, Micropor. Meso-
por. Mater. 112 (2008) 14–25.
30] C.I. Fernandes, N.U. Silva, P.D. Vaz, T.G. Nunes, C.D. Nunes, Appl. Catal. A: Gen.
[
[
[
7] J. Gimenez, C.D. Nunes, P.D. Vaz, A.A. Valente, P. Ferreira, M.J. Calhorda, J. Mol.
Catal. A 256 (2006) 90–98.
8] C.D. Nunes, A.A. Valente, M. Pillinger, A.C. Fernandes, C.C. Romão, J. Rocha, I.S.
Gon c¸ alves, J. Mater. Chem. 12 (2002) 1735–1742.
9] D.M. Ford, E.E. Simanek, D.F. Shantz, Nanotechnology 16 (2005) S458–S475.
384 (2010) 84–93.
31] (a) M.L. Clarke, D.J. Cole-Hamilton, D.F. Foster, A.M.Z. Slawin, J.D. Woollins, J.
Chem. Soc., Dalton Trans. (2002) 1618–1624;
(
b) G. Brancatelli, D. Drommi, G. Feminó, M. Saporita, G. Bottari, F. Faraone, New
J. Chem. 34 (2010) 2853-2860;
c) D. Shabashov, O. Daugulis, J. Am. Chem. Soc. 132 (2010) 3965–3972.
[
[
10] M. Pillinger, C.D. Nunes, P.D. Vaz, A.A. Valente, I.S. Gon c¸ alves, P.J.A. Ribeiro-
Claro, J. Rocha, L.D. Carlos, F.E. Kühn, Phys. Chem. Chem. Phys. 4 (2002)
(
[
32] M. Kruk, M. Jaroniec, Langmuir 13 (1997) 6267–6273.
3
098–3105.
[
33] M. Kruk, V. Antochshuk, M. Jaroniec, J. Phys. Chem. B 103 (1999) 10670–
11] P. Oliveira, A. Machado, A.M. Ramos, I. Fonseca, F.M. Braz Fernandes, A.M.
Botelho do Rego, J. Vital, Micropor. Mesopor. Mater. 120 (2009) 432–440.
12] R.G. Hayter, J. Organomet. Chem. 13 (1968) Pl–P3.
10683.
[
34] J.R. Ascenso, C.G. de Azevedo, M.J. Calhorda, M.A.A.F. de, C.T. Carrondo, P. Costa,
A.R. Dias, M.G.B. Drew, A.M. Galvão, C.C. Romão, V. Félix, J. Organomet. Chem.
[
[
[
13] H. tom Dieck, H. Friedel, J. Organomet. Chem. 14 (1968) 375–385.
14] P.K. Baker, M.B. Hursthouse, A.I. Karaulov, A.J. Lavery, K.M.A. Malik, D.J. Mul-
doon, A. Shawcross, J. Chem. Soc., Dalton Trans. (1994) 3493–3498.
15] (a) Y. Yamaguchi, K. Ogata, K. Kobayashi, T. Ito, Dalton Trans. (2004) 3982–3990;
632 (2001) 197–208.
[
[
[
[
[
[
35] P.M.F.J. Costa, M. Mora, M.J. Calhorda, V. Félix, P. Ferreira, M.G.B. Drew, H.
Wadepohl, J. Organomet. Chem. 687 (2003) 57–68.
36] B. Marler, U. Oberhagemann, S. Voltmann, H. Gies, Micropor. Mater. 6 (1996)
[
[
(
b) P.K. Baker, Adv. Organomet. Chem. 40 (1996) 45–115.
375–383.
16] (a) B.M. Trost, M. Lautens, J. Am. Chem. Soc. 104 (1982) 5543–5545;
37] W. Hammond, E. Prouzet, S.D. Mahanti, T.J. Pinnavaia, Micropor. Mesopor.
Mater. 27 (1999) 19–25.
38] S.J. Gregg, K.S.W. Sing, Adsorption Surface Area and Porosity, 2nd ed., Academic
Press, London, 1982.
(
(
b) B.M. Trost, M. Lautens, J. Am. Chem. Soc. 109 (1987) 1469–1478;
c) B.M. Trost, C.A. Merlic, J. Am. Chem. Soc. 112 (1990) 9590–9600.
[
[
[
17] D. Morales, J. Pérez, L. Riera, V. Riera, R. Corzo-Suárez, S. García-Granda, D.
Miguel, Organometallics 21 (2002) 1540–1545.
18] V.S. Joshi, M. Nandi, H. Zhang, B.S. Haggerty, A. Sarkar, Inorg. Chem. 32 (1993)
39] M.D. Alba, A. Becerro, J. Klinowski, J. Chem. Soc., Faraday Trans. 92 (1996)
8
49–854.
40] A.A. Romero, M.D. Alba, W. Zhou, J. Klinowski, J. Phys. Chem. B 101 (1997)
294–5300.
1
301–1303.
19] J.C. Alonso, P. Neves, M.J. Pires da Silva, S. Quintal, P.D. Vaz, C. Silva, A.A. Valente,
P. Ferreira, M.J. Calhorda, V. Félix, M.G.B. Drew, Organometallics 26 (2007)
5
[
[
41] M. Kruk, M. Jaroniec, Langmuir 15 (1999) 5410–5413.
42] P. Ferreira, C.D. Nunes, P.D. Vaz, N. Bion, P. Brandão, J. Rocha, Prog. Solid State
Chem. 33 (2005) 163–170.
5
548–5556.
[
[
[
20] A. Al-Ajlouni, D. Veljanovski, A. Capapé, J. Zhao, E. Herdtweck, M.J. Calhorda,
F.E. Kühn, Organometallics 28 (2009) 639–645.
21] (a) J.-M. Brégeault, Dalton Trans. (2003) 3289–3302;
[
43] C.A. Gamelas, A.C. Gomes, S.M. Bruno, F.A.A. Paz, A.A. Valente, M. Pillinger, C.C.
Romão, I.S. Gon c¸ alves, Dalton Trans. 41 (2012) 3474–3484.
(
b) K.A. Jørgensen, Chem. Rev. 89 (1989) 431–458.
22] A. Sakthivel, J. Zhao, G. Raudaschl-Sieber, M. Hanzlik, A.S.T. Chiang, F.E. Kühn,
Appl. Catal. A: Gen. 281 (2005) 267–273.