Chiral manganese(III) salen catalysts immobilized on MCM-41 and delaminated zeolites ITQ-2 and ITQ-6 through new axial coordinating linkers

Article abstract of DOI:10.1016/j.jcat.2004.08.021

We report that the catalytic behavior and enantioselectivity of three different chiral Mn(III) salen complexes anchored to traditional supports such as MCM-41 (38-A pore diameter) and delaminated zeolitic materials ITQ-2 and ITQ-6 strongly depend on wheth

Full text of DOI:10.1016/j.jcat.2004.08.021

Journal of Catalysis 228 (2004) 92–99
www.elsevier.com/locate/jcat
Chiral manganese(III) salen catalysts immobilized on MCM-41
and delaminated zeolites ITQ-2 and ITQ-6
through new axial coordinating linkers
Irene Domínguez, Vicente Fornés, María J. Sabater ∗
Instituto de Tecnología Química UPV-CSIC, Universidad Politécnica de Valencia, Avenida Los Naranjos s/n, 46022, Valencia, Spain
Received 17 May 2004; revised 9 August 2004; accepted 16 August 2004
Available online 28 September 2004
Abstract
We report that the catalytic behavior and enantioselectivity of three different chiral Mn(III) salen complexes anchored to traditional
supports such as MCM-41 (38-Å pore diameter) and delaminated zeolitic materials ITQ-2 and ITQ-6 strongly depend on whether the
complexes are attached to the surfaces through the chiral equatorial tetradentate salen ligand or via the apical ligand. As for the case of
unsupported complexes, this experimental observation has been accounted for strong variations in the conformational preference of the
catalyst intermediate toward the approaching olefin, as well as to unfavorable structural changes in the complex. In addition, control of the
hydrophobicity of the surface allows for optimization of selectivity in obtaining chiral epoxides.
2004 Elsevier Inc. All rights reserved.
Keywords: Manganese; Salen; Asymmetric epoxidation; Delaminated zeolites
1. Introduction
In this respect, we report here three new Mn(III) salen
catalysts immobilized through the equatorial and apical po-
sitions of the complexes on the MCM-41 and delaminated
materials ITQ-2 and ITQ-6 [11–13]. In all cases, the prepa-
ration of the supported catalysts implies the design of the
appropriate vinyl monomer bearing the Mn(III) salen func-
tionality and subsequent radical anti-Markonikov addition of
a mercaptoalkoxysilane group that will act as linker or spac-
ing group between the catalyst and the support.
We show that these heterogeneous chiral manganese
complexes do effectively catalyze the epoxidation reaction
of prochiral alkenes no matter the position of anchoring, al-
though only those complexes fixed to the support through the
apical coordinating ligand afford high levels of enantioselec-
tivity. This striking change in chirality has been attributed
to the replacement of bulky substituents around the active
Mn3+ ion and the failure to create a stereogenic environ-
ment, as well as to unfavorable catalyst distortions involv-
ing the planar salen ligand when the complex is anchored
through the robust salen-type ligand [1a,14,15].
Chiral salen based complexes reported by Jacobsen and
Katsuki have emerged as efficient catalysts for enantioselec-
tive epoxidation of simple olefins [1,2].
Soon after these works, intense efforts were made to pre-
pare the heterogeneous version of these chiral metal com-
plexes in an attempt to make possible their recovery and
recycling. Thus, chiral manganese salen complexes have
been encapsulated within the supercages of faujasites Y and
EMT, showing moderate to good stereoselectivity albeit with
strong reduction of activity due to diffusion control of reac-
tants within the pores of the zeolites [3]. Alternatively, they
have been immobilized on a variety of organic and inor-
ganic matrices (some of them with pore diameters ranging
in the mesoporous range) with moderate to excellent re-
sults [4–10].
*
Corresponding author. Fax: +34 96 3877809.
E-mail address: mjsabate@itq.upv.es (M.J. Sabater).
0021-9517/$ – see front matter
doi:10.1016/j.jcat.2004.08.021
2004 Elsevier Inc. All rights reserved.

I. Domínguez et al. / Journal of Catalysis 228 (2004) 92–99
93
On the other hand, we show that this methodology is
reasonably effective for improving the retention of the
monomeric manganese salen complexes to the supports and
allow recycling of the catalyst with only moderate loss of
activity and stereoselectivity.
Characterization data for complex 1: IR(KBr): 3450,
2950, 2904, 2863, 1630, 1600, 1540, 1452, 1395 cm−1
UV–vis (CH2Cl2): 225, 275, 340, 412, 500 nm.
;
2.2. Synthesis of [(R,R)-N,Nꢀ-bis-(3,5-di-tert-
butylsalycilidene)-1,2-cyclohexanediamine] manganese(III)
6-heptenecarboxylate (2)
2. Experimental
2.1. Synthesis of [(R,R)-N,Nꢀ-bis(3-tert-butyl-5-
vinylsalycilidene)-1,2-cyclohexanediamine] manganese(III)
chloride (1)
A 50-ml round-bottom flask fitted with a dropping fun-
nel was charged with 0.172 g (0.830 mmol) of AgClO4 and
15 ml of dry CH3CN. The dropping funnel was charged with
a solution of 0.500 g (0.789 mmol) of Mn(salen)Cl complex
in 25 ml of CH3CN. The solution was added dropwise to the
AgClO4 solution in the flask and a precipitate formed almost
immediately. The suspension was stirred at room tempera-
ture for 48 h and then filtered through Celite.
The chiral catalyst 1 was obtained through the synthetic
sequence designed by Salvadori et al. and given in Scheme 1
as follows: 3-tert-butyl-2-hydroxybenzaldehyde was con-
verted to 3-tert-butyl-5-(chloromethyl)-2-hydroxybenzal-
dehyde by a classic chloromethylacion method [5a,16].
This chloromethylated compound was treated with triph-
enylphosphine in refluxing benzene to give a phosphonium
salt. Subsequent Wittig reaction of this phosphonium salt
with formalin under strongly basic conditions yielded a
vinyl-substituted salicylaldehyde derivative. The final chi-
ral salen ligand was easily formed on refluxing this vinyl
salicylaldehyde derivative in ethanol with (1R, 2R)-1,2-
diaminocyclohexane. Ulterior treatment of the chiral salen
ligand with Mn(OAc)2 afforded the manganese complex 1.
The filtrate was concentrated to a volume of 10 ml
and 0.816 g (5.44 mmol) of sodium 6-heptenecarboxylate
was added. The solution was stirred for 24 h, diluted with
CH3CN, and washed with H2O. The organic phase was dried
with Na2SO4, filtered, and concentrated to give 0.442 g of 2
as a brown powder (R = 77%).
Characterization data for complex 2: IR(KBr): 3452,
2950, 1610, 1545, 1106 cm−1; UV–vis (CH2Cl2): 285,
300, 412, 510 nm; MS(FAB): m/z: 726 [M]+, 599 M–
(CH2=CH–(CH2)4–COO−)]+, 584, 543.
Scheme 1. Schematic representation of the synthetic procedure followed for the immobilization of pure complex 1 on MCM41: (i) HCl, (CH O) ; (ii) PPh ;
n
2
3
(iii) HCHO, NaOH; (iv) (1R,2R)-(−)-C H (NH ) , EtOH, reflux; (v) Mn(OAc) , NaCl, EtOH, reflux; (vi) MCM41, Cl CH, AIBN, N , reflux.
6
10
2 2
2
3
2

94
I. Domínguez et al. / Journal of Catalysis 228 (2004) 92–99
2.3. Synthesis of [(R,R)-N,Nꢀ-bis(3,5-di-tert-
butylsalycilidene)-1,2-cyclohexanediamine] manganese(III)
10-undecenoxide (3)
functionalization and silylation degree achieved in each sam-
ple were: MCM-41: mercapto groups, 1.39 mmolS g−1; tri-
methylsilyl groups 0.65 mmol g−1. ITQ-2: mercapto groups,
0.85 mmol S g−1; trimethylsilyl groups, 0.47 mmol g−1
.
ITQ-6: mercapto groups, 0.66 mmol S g−1; trimethylsilyl
groups, 1.2 mmol g−1. The loadings of bounded chiral
Mn(III) salen catalysts ranged from 0.013 to 0.096 mmol g−1
A 50-ml round-bottom flask fitted with a dropping fun-
nel was charged with 0.172 g (0.830 mmol) of AgClO4 and
15 ml of dry CH3CN. The dropping funnel was charged with
a solution of 0.500 g (0.789 mmol) of Mn(salen)Cl complex
in 25 ml of CH3CN. The solution was added dropwise to the
AgClO4 solution in the flask and a precipitate formed almost
immediately. The suspension was stirred at room tempera-
ture for 48 h and then filtered through Celite.
.
2.6. General procedure for the immobilization of complex 2
and anionic tensioactive molecules
Seven milliliters of dry toluene, 0.72 g of dehydrated pure
The filtrate was concentrated to a volume of 10 ml, and
1.04 g (5.44 mmol) of sodium 10-undecenoxide were added.
The solution was stirred for 24 h, diluted with CH3CN,
and washed with H2O. The organic phase was dried with
Na2SO4, filtered, and concentrated to give 0.48 g of 3 as a
brown powder (R = 77%).
silica MCM-41, and 5.54 mmol of 3-mercaptopropyl-1,1-
dimethoxy-1-methylsilane were added into a round-bottom
flask equipped with a reflux condenser and a magnetical stir-
rer under nitrogen atmosphere. The suspension was heated
overnight at reflux temperature. The solid was recovered
by filtration and washed exhaustively with dichloromethane
(the amount of mercapto groups calculated by elemental
analysis was 0.37 mmol g−1). The resulting solid was sus-
pended in 7 ml chloroform solution containing 0.18 mmol of
pure complex 2. The mixture was purged for 10 min with N2.
Then 15 mg of the radical initiator AIBN were added under
nitrogen and the mixture was heated to 80 ◦C for 16 h while
stirring. After cooling, the pale brown solid was filtered and
Soxhlet extracted with dichloromethane. The resulting solid
was suspended in 8 ml dimethylformamide solution contain-
ing 5 mmol of 12-bromododecanoic acid and the mixture
was heated to 155 ◦C for 24 h. The solid was recovered by
filtration and was washed exhaustively with acetonitrile. The
loaded material was treated with 7 ml of 0.05 M NaOH
ethanolic solution at room temperature while stirring for
15 min. The resulting solid was filtered and washed exhaus-
tively with acetonitrile.
Characterization data for complex 3: IR(KBr): 3463,
2956, 1633, 1544, 1313, 1258, 1091 cm−1; UV–vis (CH2-
Cl2): 212, 305, 420, 500 nm; MS(FAB): m/z: 754 [M]+,
599 [M–(CH2=CH–(CH2)8–CH2–O−)]+, 584, 543.
2.4. Synthesis of MCM-41 and delaminated materials
ITQ-2 and ITQ-6
Pure silica mesoporous MCM-41 (810 m2/g) and pure
silica delaminated materials ITQ-2 (926 m2/g) and ITQ-6
(550 m2/g) were prepared according to previously reported
procedures [11a,12,13].
2.5. General procedure for preparing the catalysts
Fifteen milliliters of dry toluene, 1 g of the dehy-
drated support (MCM-41, ITQ-2, or ITQ-6), and 1.042 g
(5.77 mmol) of 3-mercaptopropyl-1,1-dimethoxy-1-methyl-
silane were added into a round-bottom flask equipped with
a reflux condenser and a magnetical stirrer under nitrogen
atmosphere. The slurries were heated at reflux temperature
for 18 h. The solids were recovered by filtration and washed
exhaustively with dichloromethane.
Seven and one-half millimoles of pure complexes (1, 2,
or 3) were added to 10 ml chloroform solutions containing
1 g of mercapto-functionalized solids (MCM-41, ITQ-2, or
ITQ-6). The resulting slurries were purged for 10 min with
N2. Then, 25 mg of 2,2ꢀ-azobis(isobutyronitrile) (AIBN)
was added under nitrogen, and the mixtures were heated to
80 ◦C for 16 h with stirring. After cooling, the suspensions
were filtered, and the pale brown solids were washed exhaus-
tively with CH2Cl2.
2.7. General procedure for epoxidations
The reactions were carried out at 0 ◦C in a system com-
posed of 2 ml CH2Cl2 and 2 ml of an aqueous buffered
solution of NaOCl as oxidant (pH 11). The reaction solu-
tions contained 1 mmol of olefin, 1 mmol of n-undecane as
internal standard, and 0.5 to 5% of Mn(III) salen catalyst.
The reactions were monitored by chiral GC chromatography
and HPLC.
2.8. Instrumentation
IR spectra of pure complexes were recorded on KBr pel-
lets in a Nicolet 710 FT spectrophotometer. FT-IR spectra
of supported complexes were recorded at room temperature
using a greaseless quartz cell fitted with CaF2 windows in
a Nicolet 710 FT spectrophotometer. Self-supported wafers
(∼ 10 mg) were prepared by pressing the zeolite pow-
der at 1 ton cm−2. The samples were outgassed at 100 ◦C
under 10−2 Pa for 1 h before recording the IR spectra.
Silylation of the samples was accomplished as follows:
1 g of the previously loaded solids was suspended in 15 ml
of fresh toluene with 0.469 g (2.90 mmol) of silylat-
ing reagent, 1,1,1,3,3,3-hexamethyldisilazane (HMDS). The
mixtures were heated again at reflux temperature for 2 h, fil-
tered, and washed exhaustively with dichloromethane. The

I. Domínguez et al. / Journal of Catalysis 228 (2004) 92–99
95
Room-temperature transmission UV–vis spectra of transpar-
ent solutions were recorded in a Shimadzu UV–vis scan-
ning spectrophotometer. DR spectra of the opaque powders
were recorded in a Varian Cary 5G UV–vis NIR spectropho-
tometer adapted with a praying mantis attachment and using
BaSO4 as reference.
Combustion chemical analysis of the samples were
carried out using a Fisons EA 1108-CHNS-O analyzer.
FAB MS spectra of the manganese salen complexes were
recorded using a VG-Autospec.
2.9. Molecular modeling
Molecular simulations were performed using an atomistic
mechanics methodology and Universal Force Field based
calculations [17a,b]. The energetic minimizations were car-
ried out with standard techniques as implemented in the soft-
ware package Cerius 2-4.6 [17c].
Fig. 1. DR–UV spectra of complexes 1, 2, and 3 supported on pure
silica MCM41 and silylated ITQ-6: (a) 1–MCM41; (b) 2–MCM-41;
(c) 3–ITQ-6s.
Scheme 1 and details on this preparation under Section 2).
In this case the FT-IR spectrum of the resulting supported
complex (1–MCM41) exhibited very weak and undefined
bands; herein only the UV–vis spectra in diffuse reflectance
mode showed the typical unfeatureless spectra associated
with these types of Mn(III) salen complexes (Fig. 1).
3. Results and discussion
Three different preformed Mn(III) salen catalysts were
immobilized through a similar anchoring method on the sur-
face of mesoporous MCM-41 (pure silica; diameter pore
38 Å) and pure silica delaminated materials ITQ-2 and
ITQ-6 provided a large external surface and large amount
of available silanol groups on these structured materi-
als [11–13]. Effectively, ITQ-2 consists of very thin silica
layers 2.5 nm in height organized in a “house of cards”-
type structure. This material (prepared by delaminating the
precursor of the pure siliceous MWW zeolite) contains a
very large and well-defined external surface (> 800 m2 g−1),
whereas ITQ-6 (a delaminated material derived from the
lamellar precursor of the ferrierite zeolite) has a total sur-
face area greater than 500 m2 g−1 [12,13]. In this respect,
both type of supports (having an important population of
silanol groups highly accessible) are clear candidates for an-
chor metal precursors or transition metal complexes with
high levels of dispersion on the surfaces.
Inꢀthe first stage we prepared the vinyl monomer [(R, R)-
N,N -bis(3-tert-butyl-5-vinylsalycilidene)-1,2-cyclohex-
anediamine] manganese(III) chloride (1) by following the
synthetic sequence shown in Scheme 1 (see Section 2). Char-
acterization of pure complex 1 by IR spectroscopy showed
the characteristic IR vibration bands of unsupported Mn(III)
salen complexes, i.e., the imine stretching vibration around
1630 cm−1 and the typical band of metallosalen complexes
at 1540 cm−1. On the other hand, the electronic absorption
spectrum in dichloromethane solution showed typical ligand
transitions of the brownish oxidized complex at 225, 275,
and 340 nm and a weak d–d transition around 500 nm. Once
established, the identity of the complex, compound 1, was
covalently linked to the surface of MCM-41 through rad-
ical addition of mercapto groups placed previously in the
solid by using azoisobutyronitrile as radical initiator (see
The efficiency of 1–MCM41 as catalyst was tested in the
epoxidation reaction of 1-phenylcyclohexene and showed
good activity (conversion of 72%, which corresponds to a
TON of 29), moderate levels of selectivity (75%), and very
low levels of stereoinduction (8% ee) at 0 ◦C (see entry 1
in Table 1). As the existence of residual amounts of uncom-
plexed Mn(III) that could be catalyzing the reaction in a non-
selective way is a priori not possible because the complex
was completely built up before the anchoring, the possibility
that the steric and electronic properties of the salen frame-
work could be unfavorably altered by the replacement of
two bulky tert-butyl groups on both aromatic portions is not
unrealistic. In fact, in previous literature references on the
epoxidation reaction mechanism it has been suggested that
the alkene approaches the oxygen at the oxomanganese(V)
intermediate in a “side-on” orientation to avoid steric inter-
actions with the bulky tert-butyl substituents [1a,14]. Hence,
it is very likely that any structural change involving these
substituents at the respective aromatic positions may al-
ter the enantioselectivity values that can be achieved with
tert-butyl substituents.
Taking into account these premises, we devised the prepa-
ration of a new supported chiral manganese complex with
the anchoring point located away from the stereogenic cen-
ter. This complex would maintain the four crucial tert-butyl
substituents at the near-planar salen ligand having simulta-
neously a terminal double bond on the monodentate ligand
at the axial position (see Scheme 2). Moreover, we spec-
ulated on the possibility that the introduction of a longer
spacer between the solid and the metal center (in the form

96
I. Domínguez et al. / Journal of Catalysis 228 (2004) 92–99
Table 1
Results on the catalytic asymmetric epoxidation of alkenes with chiral manganese(III) salen complexes 1, 2, and 3 supported on MCM-41 and delaminated
zeolites ITQ-2 and ITQ-6
a
b
b
b
c
Substrate
1-Phenylcyclohexene
Catalyst
Time (h)
Conversion (%)
Epoxide selectivity (%)
ee (%)
TON
d
d
e
f
1–MCM41
2–MCM41
2–MCM41
2–MCM41
2–MCM41t
3–MCM41s
2–ITQ-2s
2–ITQ-2s
2–ITQ-6s
19
23
29
47
66
48
47
48
47
72
89
66
84
93
99
98
73
98
75
78
66
60
16
82
90
80
90
8
86
76
52
46
70
81
63
76
29
36
d,g
h
37
21
20
h
e,h
h
20
h,i
Indene
3–ITQ-2s
3–ITQ-6s
48
70
56
90
90
90
62
56
112
180
h,i
h,i
1,2-Dihydronaphthalene
a
3–ITQ-2s
3–ITQ-6s
54
54
68
89
85
90
62
56
136
178
h,i
◦
All reactions were performed at 0 C by using solid catalysts containing 5% mmol of chiral Mn(III) complex, 1 mmol of alkene, 1 mmol of undecane as
internal standard, and NaOCl as oxidant.
b
Conversion, epoxide selectivity and enantiomeric excess (ee) were determined by capillary GC (Chiraldex G-PN column) and chiral HPLC (Chiraldex OB
and OD columns).
c
Calculated as millimoles of substrate converted per millimole of catalyst.
2.5% mmol chiral Mn(III) catalyst.
First reuse.
Second reuse.
Tensioactive-containing MCM-41.
Silylated catalysts.
0.5% mmol of chiral Mn(III) complex.
d
e
f
g
h
i
Scheme 2. Schematic representation of the synthetic procedure followed for the immobilization of pure complexes 2 and 3 on MCM41, ITQ-2, and ITQ-6
materials: (i) AgClO , CH CN; (ii) CH =CH–(CH ) COONa (or CH =CH–(CH ) CH ONa); (iii) MCM-41 (or ITQ-2 or ITQ-6), toluene, reflux.
4
3
2
2 4
2
2 8
2

I. Domínguez et al. / Journal of Catalysis 228 (2004) 92–99
97
of a long n-alkyl chain) would be beneficial for both ac-
tivity and enantioselectivity, as the existence of less steric
hindrance would cause the performance of this catalytic sys-
tem to approach that of the homogeneous model in the liquid
phase [18]. Accordingly we chose the 6-heptenecarboxylate
anion as axial coordinating ligand as it had the necessary bi-
functionalization for coordinating the metal (the carboxylate
group) and carrying out the simultaneous attachment to the
solids through the terminal double bond (see Scheme 2).
Thus we prepared the vinyl monomer [(R,R)-N,Nꢀ-
bis(3,5-di-tert-butyl-5-salycilidene)-1,2-cyclohexanediam-
ine] manganese(III) 6-heptenecarboxylate (2) in a straight-
forward fashion by methatesis of the chloride ligand of the
original Jacobsen catalyst by the 6-heptenecarboxylateanion
as outlined in Scheme 2 (details of the procedures, syn-
thetic conditions, and characterization are provided under
Section 2).
ene. As can be followed from results collected in Table 1,
the solid catalyst 2–MCM-41 gave similar yields of chi-
ral epoxide and very good levels of stereoinduction (see
Table 1). The remarkable difference between the chiral in-
ductor characters of the supported catalysts when they are
linked through the stereogenic tetradentate salen ligand or,
alternatively, through the achiral apical ligand is noticeable
and strongly supports the notion that the significant low
enantiomeric excess obtained with the supported catalyst
1–MCM41 is due (at least in part) to the absence of bulky
ligands around the active Mn3+ ion to create a stereogenic
environment (a fact that has been well established for un-
supported complexes) [1a,14]. Nevertheless, the existence of
unfavorable electronic effects and/or ligand distortion modes
that would lead presumably to a loss of the near-planar
geometry of the equatorial salen ligand has been recognized
early on [15].
As in the previous case, the pure complex 2 was cova-
lently linked to the surface of MCM-41 through radical ad-
dition of the mercapto groups placed previously on the solid
to the terminal double bond on the axial coordinating lig-
and, hence affording the heterogeneous catalyst 2–MCM41
as depicted in Scheme 2.
Again the identity of this immobilized complex was
established by UV–vis spectroscopy in diffuse reflectance
mode (see Fig. 1), whereas its catalytic activity was routinely
determined in the epoxidation reaction of 1-phenylcyclohex-
On the other hand, in an attempt to facilitate the inter-
action between the aqueous oxidant ClO−4 and the hetero-
geneous complex, we conceived the preparation of metal
complex-containing particles with amphiphilic character.
These particles contained both the chiral manganese com-
plex 2 and a sort of anionic tensioactive molecule that
was immobilized at the surface of pure silica MCM-41
(2–MCM41t) (see Scheme 3 and details of its preparation
under Section 2). In this case the tensioactive refers to a
molecule with a hydrophilic head group (a carboxylate an-
Scheme 3. Immobilization of the anionic tensioactive and pure complex 2 on MCM41: (i) CH Si(OCH ) CH CH CH SH, toluene, reflux; (ii) 2, Cl CH,
3
3 2
2
2
2
3
AIBN, N , reflux; (iii) BrCH (CH ) COOH, DMF, reflux; (iv) 0.05 M NaOH, EtOH.
2
2
2 10

98
I. Domínguez et al. / Journal of Catalysis 228 (2004) 92–99
ion) and a hydrophobic tail group formed by a long n-alkyl
chain. We expected that the highly hydrophilic character of
the surfactant molecules at the head group would drive the
aqueous oxidant to assemble or accumulate at the surface
of the solid, favoring interaction between the hypochlorite
molecules and the metal. The efficiency of this catalyst was
tested in the epoxidation reaction of 1-phenylcyclohexene,
giving very low values of selectivity (16%) and moderate
levels stereoinduction (46% ee) at 0 ◦C (see Table 1). This
discouraging result can be explained if we take into ac-
count that (as has been demonstrated recently) control of
hydrophobicity of the heterogeneous surfaces can influence
the catalytic activity of a solid catalyst by allowing selective
adsorption of reactants and products and therefore favor-
ing, in some cases, secondary reactions [19]. In our case,
the preferable selective adsorption of water and/or oxidant-
derived species could account for such undesired collateral
reactions (vs epoxide ring opening).
Hereinafter anchoring complex 2 on the surface of ITQ-2
and ITQ-6, we accomplished exhaustive silylation of any re-
maining silanol group (Si–OH) with hexamethyldisilazane
(HMDS) to reduce the hydrophilicity of the surfaces to give
the heterogeneous complexes 2–ITQ-2s and 2–ITQ-6s (see
synthetic sequence described in Section 2.5).
In this case the catalytic activity of these solids was deter-
mined in the epoxidation reaction of 1-phenylcyclohexene,
giving identical TON values, very good yields of chiral
epoxides (around 90%), and good levels of steroinduc-
tion (see Table 1). Fig. 2 is a schematic representation of
the external surface of delaminated zeolite ITQ-2 and the
optimized structural conformation achieved by complex 2
on the surface of this material calculated by using stan-
dard techniques as implemented in the software package
Cerius 2-4.6 [17].
Fig. 2. (a) Molecular simulation of manganese salen complex 2 anchored to
the ITQ-2 surface. (b) Schematic representation of the external structure of
ITQ-2, which is composed of an array of cups containing silanols in spe-
cific positions. (c) Computational simulation of manganese salen complex
3 grafted onto the ITQ-2 surface.
Finally, following with the same synthetic methodology,
a third chiral pentadentate manganese complex was pre-
pared by chloride shifting with 10-undecenoxide anion to
give [(R,R)-N,Nꢀ-bis(3,5-di-tert-butyl-5-salycilidene) 1,2-
cyclohexanediamine] manganese(III) 10-undecenoxide (3)
(see Scheme 2 and Section 2). The resulting pure com-
plex 3 was linked to the surface of mercapto-functionalized
MCM41, ITQ-2, and ITQ-6 matrices through the terminal
vinyl moiety, affording, after subsequent silylation, the het-
erogeneous complexes 3–MCM41s, 3–ITQ-2s, and 3–ITQ-
6s (see Section 2). The DR–UV spectra of these three solid
catalysts were rather similar. For this reason, Fig. 1 shows
the UV spectrum of only one of these supported catalysts
(3–ITQ-6s) recorded in diffuse reflectance mode. In addi-
tion, Fig. 2 shows the optimized structural conformation
achieved by complex 3 on the surface of ITQ-2 according
to the Cerius 2-4.6 program.
selectivity parameters achieved in the epoxidation of 1,2-
dihydronaphthalene and indene were accompanied by the
highest TON values, as the amount of catalyst employed in
these cases was the smallest (see Table 1).
Finally, because the reusability of the complexes an-
chored or fixed on a support is one of the most important
advantages of heterogeneization, the solids 2–MCM41 and
2–ITQ-2 were separated by filtration, washed with fresh
solvent, and recycled. Then, two new epoxidation reac-
tions were repeated with both catalysts by adding 1-phenyl-
cyclohexene and oxidant NaOCl under the same experimen-
tal conditions. Sixty-six percent conversion was achieved
after 29 h with only a slight to moderate loss of activity and
enantioselectivity in the case of 2–MCM41, whereas cata-
lyst 2–ITQ-2 afforded 73% conversion after 48 h, also with
a moderate loss of selectivity and enantioselectivity (see Ta-
ble 1).
These three new supported complexes were used as cata-
lysts in the epoxidation reaction of prochiral 1-phenylcyclo-
hexene, 1,2-dihydronaphthalene, and indene with excellent
selectivities and good levels of stereoinduction (see Ta-
ble 1). It is necessary to point out that the good activity and
In addition, the electronic spectrum of the reused catalyst
2–MCM41 was recorded and compared with the DR spec-
trum of the initially loaded material 2–MCM41, both spectra
being qualitatively very similar (not shown). This fact sup-

I. Domínguez et al. / Journal of Catalysis 228 (2004) 92–99
99
ports the notion that the supported Mn(III) salen complex is
References
not substantially altered under these experimental conditions
as confirmed by the relatively good activity and selectivity
parameters that could be achieved in a first reuse. In a further
experiment the recovered catalyst 2–MCM41 was reused
again in a new catalytic reaction (second reuse). In this case
the activity and stereoselectivity of the catalyst declined, al-
though the epoxide was still formed (see Table 1).
In view of these results it can be concluded that this
methodology is reasonably effective for improving the re-
tention of the monomeric manganese salen complexes to
the supports and allowing the subsequent recycling, whereas
the stability of the complexes under oxidizing conditions is
also reasonably good as the performance of the heteroge-
neous catalysts smoothly decreases with time, being more
pronounced after prolonged reaction times or subsequent
reuses.
[1] (a) W. Zhang, J.L. Loebach, S.R. Wilson, E.N. Jacobsen, J. Am. Chem.
Soc. 112 (1990) 2801–2803;
(b) E.N. Jacobsen, W. Zhang, A.R. Muci, J.R. Ecker, L. Deng, J. Am.
Chem. Soc. 113 (1991) 7063–7064.
[2] (a) T. Irie, K. Noda, N. Matsumoto, T. Katsuki, Tetrahedron Lett. 31
(1990) 7345–7348;
(b) T. Katsuki, Coord. Chem. Rev. 140 (1995) 189–214.
[3] (a) S.B. Ogunwumi, T. Bein, Chem. Commun. (1997) 901–902;
(b) M.J. Sabater, A. Corma, A. Doménech, V. Fornés, H. García,
Chem. Commun. (1997) 1285–1286.
[4] (a) J.M. Fraile, J.I. García, J. Massam, J.A. Mayoral, J. Mol. Catal.
A 136 (1998) 47–57;
(b) P. Sutra, D. Brunel, Chem. Commun. 2485–2486.
[5] (a) F. Minutolo, D. Pini, A. Petri, P. Salvadori, Tetrahedron: Asymme-
try 7 (1996) 2293–2302;
(b) C.E. Song, E.J. Roh, B.M. Yu, D.Y. Chi, S.C. Kim, K.J. Lee, Chem.
Commun. 7 (2000) 615–616;
(c) B.B. De, B.B. Lohray, S. Sivaram, P.K. Dhal, Tetrahedron: Asym-
metry 6 (1995) 2105–2108;
(d) B.B. De, B.B. Lohray, S. Sivaram, P.K. Dhal, Macromolecules 27
(1994) 1291–1296;
4. Conclusion
The preparation of three different Mn(III) salen com-
plexes supported on pure silica mesoporous material MCM-
41 and two pure silica delaminated zeolites ITQ-2 and ITQ-6
is reported. The complexes were structurally modified (with
respect to the original Jacobsen catalyst) to get the neces-
sary bifunctionalization to carry out the simultaneous attach-
ment to the organic supports and coordination or interaction
with metals. These transition metal complexes had a vinyl
functionality placed strategically at the equatorial and axial
positions of the pentacoordinated manganese complex, re-
spectively. This vinyl moiety served to attach the complex
covalently to the supports. In principle, the attachment of the
complex through the equatorial position of the metal com-
plex led to a strong reduction in enantioselectivity. Two new
chiral manganese salen complexes bearing a vinyl group at
the apical coordinating ligand were prepared. These chiral
manganese salen complexes, which were fixed to the re-
spective supports through the axial coordinating ligand, were
active and more enantioselective toward epoxidation of var-
ious prochiral alkenes.
(e) G.-J. Kim, Y.K. Kwon, D.-W. Park, Polym. Prepr. 43 (2002) 539–
543.
[6] K.B.M. Janssen, I. Laquiere, W. Dehaen, R.F. Parton, I.F.J. Vankele-
com, P.A. Jacobs, Tetrahedron: Asymmetry 8 (1997) 3481–3487.
[7] (a) F. Bigi, L. Moroni, R. Magi, G. Sartori, Chem. Commun. 7 (2002)
716–717;
(b) P. Piaggio, C. Langham, P. McMorn, D. Bethell, P.C.B. Page, F.E.
Hancock, C. Sly, G.J. Hutchings, J. Chem. Soc., Perkin Trans. 2 (1)
(2000) 143–148.
[8] B.M. Choudari, N.S. Chowdari, M.L. Kantam, P.L. Santhi, Catal.
Lett. 76 (2001) 213–218.
[9] L. Canali, D.C. Sherrintong, Chem. Soc. Rev. 28 (1999) 85–93.
[10] T.S. Reger, K.D. Janda, J. Am. Chem. Soc. 122 (2000) 6929–6934.
[11] (a) J.S. Beck, J.C. Vartuli, W.J. Roth, M.E. Leonowicz, C.T. Kresge,
K.D. Schmitt, C.T.-H. Chu, O.H. Olson, E.W. Sheppard, S.B. Mc-
Cullen, J.B. Higgins, J.L. Schelenker, J. Am. Chem. Soc. 114 (1992)
10,834;
(b) A. Fuerte, M. Iglesias, F. Sánchez, A. Corma, J. Mol. Catal. A 211
(2004) 227–235.
[12] (a) A. Corma, V. Fornés, S.B.C. Pergher, WO9717290A1, 1998;
(b) A. Corma, V. Fornés, S.B.C. Pergher, Th.L. Maesen, J.G. Buglass,
Nature 396 (1998) 353–356.
On the other hand, when recycling, the performance of
the supported Mn(III) salen catalysts declined, although the
epoxide was still obtained with good values of activity and
enantioselectivity, especially with the first reuses.
Finally, silylation of the surfaces improved the epoxide
selectivity due to reduction of undesirable secondary reac-
tions caused by the selective adsorption of different products
on the hydrophilic surfaces.
[13] A. Corma, U. Díaz, M. Domine, V. Fornés, J. Am. Chem. Soc. 122
(2000) 2804–2809.
[14] N. Hosoya, A. Hatayama, K. Yanai, H. Fujii, R. Irie, T. Katsuki, Syn-
lett 9 (1993) 641–645.
[15] K.B. Lipkowitz, S. Schefzick, Chirality 14 (2002) 677–682.
[16] G. Wulff, A. Akelah, Makromol. Chem. 179 (1978) 2647–2651.
[17] (a) C.R.A. Catlow, A.N. Cormack, Intern. Rev. Phys. Chem. 6 (1987)
227–236;
(b) A.K. Rappe, C.J. Casewit, K.S. Colwell, W.A. Goddard III, W.M.
Skiff, J. Am. Chem. Soc. 114 (1992) 10,024–10,030;
(c) Accelrys Inc., http://www.accelrys.com.
Acknowledgment
[18] A. Corma, H. García, A. Moussaif, M.J. Sabater, R. Zniber, A. Re-
douane, Chem. Commun. (2002) 1058–1059.
Financial support from the Spanish DGICYT (MAT2003-
07945-C02-00) is acknowledged. Irene Domínguez thanks
the CSIC for an I3-P fellowship.
[19] A. Corma, M. Domine, J.A. Gaona, J.L. Jordá, M.T. Navarro, F. Rey,
J. Pérez-Pariente, J. Tsuji, B. McCulloch, T. Nemeth, Chem. Commun.
(1998) 2211–2212.