Asymmetric ring-opening of epoxides on chiral Co(Salen) catalyst synthesized in SBA-16 through the "ship in a bottle" strategy

Article abstract of DOI:10.1016/j.jcat.2007.03.006

Chiral Co(Salen) complex was synthesized in the mesoporous cage of SBA-16 through the "ship in a bottle" method. The pore entrance size of SBA-16 was precisely tailored by varying the autoclaving time and silylation with phenyltrimethoxysilane to trap Co(Salen) complex in the cage of SBA-16. Chiral Co(Salen) trapped in SBA-16 shows enantioselectivity (up to 87-96% ee) as high as that of the homogeneous catalyst for the asymmetric ring opening of terminal epoxides and can be recycled at least 10 times with no apparent loss of activity. The activity for the catalyst trapped inside SBA-16 can be significantly increased when the surface is modified with organic groups. This work extends the "ship in a bottle" synthesis from microporous materials to mesoporous cage-like materials and develops an effective strategy to trap metal complex catalyst with large molecular size into the nanopores or cavities of mesoporous materials.

Full text of DOI:10.1016/j.jcat.2007.03.006

Journal of Catalysis 248 (2007) 204–212
www.elsevier.com/locate/jcat
Asymmetric ring-opening of epoxides on chiral Co(Salen) catalyst
synthesized in SBA-16 through the “ship in a bottle” strategy
∗
∗
Hengquan Yang, Lei Zhang, Weiguang Su, Qihua Yang , Can Li
State Key Laboratory of Catalysis, Dalian Institute of Chemical Physics, Chinese Academy of Sciences, 457 Zhongshan Road, Dalian 116023, China
Received 4 January 2007; revised 7 March 2007; accepted 7 March 2007
Available online 24 April 2007
Abstract
Chiral Co(Salen) complex was synthesized in the mesoporous cage of SBA-16 through the “ship in a bottle” method. The pore entrance size of
SBA-16 was precisely tailored by varying the autoclaving time and silylation with phenyltrimethoxysilane to trap Co(Salen) complex in the cage
of SBA-16. Chiral Co(Salen) trapped in SBA-16 shows enantioselectivity (up to 87–96% ee) as high as that of the homogeneous catalyst for the
asymmetric ring opening of terminal epoxides and can be recycled at least 10 times with no apparent loss of activity. The activity for the catalyst
trapped inside SBA-16 can be significantly increased when the surface is modified with organic groups. This work extends the “ship in a bottle”
synthesis from microporous materials to mesoporous cage-like materials and develops an effective strategy to trap metal complex catalyst with
large molecular size into the nanopores or cavities of mesoporous materials.
©
2007 Elsevier Inc. All rights reserved.
Keywords: Ship in a bottle; Mesoporous cage-like material; Salen catalyst; Asymmetric ring-opening
1
. Introduction
olite [8]. Compared with microporous zeolites, ordered meso-
porous silicas have larger pore size and pore volume, which
provides possibilities for trapping larger molecules and more
“comfortable” microenvironment for guest molecules. Previ-
ously, Algarra and Tanamura [21,22] tried to synthesize copper
phthalocyanine and porphyrin in MCM-41 through the “ship
in a bottle” method, but reported no catalytic properties of the
prepared catalysts. So far, the leaching of metal complexes re-
mains a problem when metal complexes are accommodated in
the mesoporous materials without modifying the pore structure.
Ordered mesoporous silicas with cage-like structures, such
as SBA-1 (cubic, Pm3n) [23,24], SBA-16 (cubic, Im3m)
[25,26], FDU-12 (cubic, Fm3m) [27], and FDU-1 (cubic,
Fm3m) [28], have been well synthesized. These mesoporous
cage-like silicas have tunable cage sizes (4–8 nm for SBA-16;
10–22 nm for FDU-12) and their cages are interconnected
three-dimensionally by tunable pore entrances. The large cages
of these mesoporous materials can accommodate metal com-
plexes of large molecular size, whereas the smaller pore en-
trances may prevent leaching of the metal complex confined
in the mesoporous cage. In addition, the existence of plentiful
hydroxyl groups in the mesoporous silicas provides the possi-
The “ship in a bottle” synthesis has become an efficient
method for immobilizing homogeneous catalysts within the
solid matrix, particularly in microporous materials like zeolites
[1,2]. Over the last two decades, metal complexes with lig-
ands like Salen, porphyrin, phthalocyanine, bipyridine, triaza-
cyclononane, and CO have been successfully synthesized inside
the supercages of X, Y, EMT, MCM-22 zeolites, and mont-
morillonite for various catalytic reactions [3–19]. The pore en-
trance size and cage dimension of the microporous zeolites are
strictly limited by their crystalline topology (usually <1.5 nm).
It is difficult to synthesize the metal complex with relatively
larger molecular size (like Jacobsen’s catalyst with the ligand
ꢀ ꢀ
,5,3 ,5 -tetra-tert-butyl-substituted SalenH ) [6,7,20] by the
2
3
“
ship in a bottle” method in the cage of a zeolite, except for
the cage-enlarged faujasites [20], EMT [19] and MCM-22 ze-
*
Corresponding authors. Fax: +86 411 84694447.
E-mail addresses: yangqh@dicp.ac.cn (Q. Yang), canli@dicp.ac.cn
(C. Li).
URL: http://www.canli.dicp.ac.cn (C. Li).
0021-9517/$ – see front matter © 2007 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2007.03.006

H. Yang et al. / Journal of Catalysis 248 (2007) 204–212
205
bility of tailoring the pore entrance size by a simple silylation
2.3. Synthesis of chiral Co(Salen) in the cage of SBA-16
through the “ship in a bottle” method
reaction [25]. Simultaneously, the surface properties, such as
hydrophobicity and hydrophilicity, can be modified by silyla-
tion using silane precursors with different organic groups.
Although mesoporous cage-like silicas are good candidates
for accommodation of the metal complexes, the synthesis of
metal complexes in the mesoporous cage-like materials through
the “ship in a bottle” method has been rarely reported. An un-
derlying reason for this is the difficulty in exactly tailoring and
determining the pore entrance size, which is a key factor in
the successful encapsulation of a metal complex through the
1
.0 g of SBA-16(4.9) or SBA-16(n)-Ph (n = 4.9, 5.4, 5.9)
and 0.15 g of SalenH2 ligand were mixed together and kept at
28 K for 24 h under vacuum. After cooling to 368 K, 0.113 g
4
of Co(OAc)2·4H2O, 15 mL of ethanol, and 7.5 mL of toluene
were added. The mixture was stirred at 368 K for 24 h un-
der Ar atmosphere. The isolated solid was thoroughly washed
in a Soxhlet apparatus consecutively with CH2Cl2, THF, and
ethanol. Co(Salen)/SBA-16(4.9) and Co(Salen)/SBA-16(n)-Ph
“
ship in a bottle” strategy. Our primary results indicated that ho-
(
1
n = 4.9, 5.4, 5.9) were prepared from SBA-16(4.9) and SBA-
6(n)-Ph (n = 4.9, 5.4, 5.9), respectively. The chiral Co(Salen)
synthesized in SBA-16 is schematically presented in Scheme 1.
mogeneous catalysts with large molecular size can be directly
encapsulated in the nanocavities of mesoporous materials [29].
Herein, we investigated the “ship in a bottle” synthesis of chi-
ral Co(Salen) inside the mesoporous cage of SBA-16 using the
2
.4. Catalytic reaction
ꢀ
ꢀ
molecular fragments of 3,5,3 ,5 -tetra-tert-butyl-SalenH2 and
Co(OAc)2. Tailoring the pore entrance size via silylation is
crucial to confine the metal complex. The Co(Salen) catalyst
trapped inside SBA-16 (silylated with phenyltrimethoxysilane)
exhibited comparable enantioselectivity to its homogeneous
counterpart in the asymmetric ring-opening of the terminal
epoxides. No apparent loss of activity and enantioselectivity
was observed for the heterogeneous catalyst even after 10 re-
action cycles.
The catalyst was oxidized in a mixture of toluene and acetic
acid by air for 2.5 h before reaction. The solid catalyst was ob-
tained by centrifugation and was evacuated under vacuum to
remove toluene and acetic acid. For asymmetric ring-opening
of epichlorohydrin, when THF was used as solvent, 2 mmol
of epichlorohydrin, 0.2 g of THF, and 1.3 mol of water were
mixed with the catalyst (Co content 0.02 mmol). The reaction
was performed at 298 K. In the case of the reaction without
solvent, the reaction was performed at 298 K with a molar ra-
tio of epichlorohydrin:H2O:Co of 1:0.75:0.01. For asymmetric
ring-opening of propylene oxide, the reaction was performed
at 283 K with a molar ratio of propylene oxide:H2O:Co of
2
. Experimental
2
.1. Reagents and materials
1
:0.8:0.005. After reaction, 0.35 mL of THF was added to the
Pluronic P123 copolymer (EO20PO70EO20), phenyltrimeth-
reaction mixture, and nonane was added as an internal standard.
The diol thus-obtained was derived with dimethoxypropane in
the presence of p-toluenesulfonic acid. The derivatives were
purified with short gel column and then analyzed by gas chro-
matography (Agilent 6890 with HP-Chiral19091G-B213 capil-
lary column). The catalyst was isolated from the reaction mix-
ture by centrifugation. After being washed with THF, oxidized
in toluene/acetic acid by air for 2.5 h, isolated by centrifuga-
tion, and dried under vacuum, the recovered catalyst was used
in the recycling experiments.
oxysilane (98%), and 2,2-dimethoxypropane (>99%) were
purchased from Aldrich. Pluronic F127 (EO106PO70EO106)
was obtained from Sigma. Tetraethylorthosilicate (TEOS, AR),
epichlorohydrin (>99%), and propylene oxide (>99%) were
purchased from the Shanghai Chemical Reagent Company
of the Chinese Medicine Group. (R,R)-N,N -bis(3,5-di-tert-
butylsalicylidene)-1,2-cyclohexanediamine (denoted SalenH2)
was synthesized as described previously [30].
Mesoporous cage-like material SBA-16 was synthesized ac-
cording to a modified method [25]. SBA-16(4.9), SBA-16(5.4),
and SBA-16(5.9) were synthesized by autoclaving at 373 K
for 5.5, 9, and 12 h, respectively, where the number in parenthe-
ses is the cage size (in nanometers) of the mesoporous material
analyzed from N2 adsorption branch based on the BJH method.
ꢀ
2
.5. Characterization
Powder X-ray diffraction patterns of SBA-16 were recorded
on a Rigaku D/Max3400 powder diffraction system (CuKα,
40 kV, 30 mA). N2 physical adsorption analysis was carried out
on Micromeritics ASAP 2020 volumetric adsorption analyzer.
Before the adsorption measurements, the samples were out-
gassed at 393 K for 6 h. UV–vis spectra and diffuse-reflectance
UV–vis spectra were recorded on a JASCOV-550 UV–vis spec-
trophotometer using dichloromethane and BaSO4 as the refer-
ence, respectively. FT-IR spectra were collected on a Thermo-
Nicolet Nexus 470 infrared spectrometer, and Co content was
analyzed on a Plasma-spec-II (Leeman). TEM micrographs
were taken using a JEM-2010 transmission electron microscopy
at an acceleration voltage of 120 kV.
2
.2. Modification of SBA-16 by silylation
.0 mL of dry toluene was added to 1.0 g of SBA-16(n)
1
(
evacuated at 398 K for 6 h; n = 4.9, 5.4, 5.9), followed by the
addition of 1.25 mL of anhydrous n-butylamine and 5 mmol
of phenyltrimethoxysilane. After refluxing at 384 K for 24 h
under Ar atmosphere, the resulting solid was isolated by rapid
filtration and thoroughly washed with toluene and the mixture
of CH2Cl2 and diethyl ether. The resultant materials were des-
ignated SBA-16(n)-Ph, where n is the cage size of SBA-16 and
Ph denotes the phenyl group.

206
H. Yang et al. / Journal of Catalysis 248 (2007) 204–212
Scheme 1. Schematic description of chiral Co(Salen) synthesized in the cages of phenyl-modified SBA-16 through the “ship in a bottle” method.
3
. Results and discussion
3
.1. Tuning the pore entrance size of SBA-16
To obtain mesoporous cage-like silicas with different pore
entrance sizes, SBA-16(4.9), SBA-16(5.4), and SBA-16(5.9)
with 4.9, 5.4, and 5.9 designating cage sizes (in nm) of these
(
three materials) were synthesized as described previously [25].
Fig. 1 shows the XRD patterns of these materials. All materi-
als exhibit two diffraction (110) and (200) peaks, characteristic
of mesoporous materials with the cubic Im3m structure. The
2
θ value of the (110) diffraction peak gradually shifts to lower
angle from SBA-16(4.9) to SBA-16(5.9), suggesting that the
materials with larger unit cells can be obtained with longer au-
toclaving time.
Fig. 1. XRD patterns of mesoporous SBA-16 materials available for trapping
metal complex catalysts.
The N sorption isotherms of SBA-16(4.9), SBA-16(5.4),
2
and SBA-16(5.9) are displayed in Fig. 2. The sorption plots
show a typical H2 hysteresis loop characteristic of materials
with a good-quality cage-like porous structure. The cage size
also increases with increasing autoclaving time (Table 1).
The pore entrance size is very important for the successful
synthesis of metal complexes in the cage of SBA-16 through the
The chiral ligand (SalenH ) with flexible structure can eas-
ily twist its way to diffuse through the pore entrance into the
2
cage (at 428 K). Once SalenH in the cage coordinates with
2
Co(OAc) ·4H O, Co(II)(Salen) with a rigid structure can be
2
2
formed, and thus it cannot diffuse out from the mesoporous
cage if the pore entrance size is smaller than the molecular size
of Co(II)(Salen).
“
ship in a bottle” method. In principle, the pore entrance size
should be larger than the molecular size of the fragments used
for constructing the metal complex and smaller than the metal
complex catalyst formed inside the cage. In this work, we chose
chiral Co(Salen) as a model metal complex, and synthesized
Co(Salen) complex in the cage of SBA-16 through the flexible
ligand method [9]. The process of synthesizing Co(Salen) in
the cage of SBA-16 is schematically displayed in Scheme 1.
However, it is difficult to estimate the pore entrance size
(<4 nm) only from the desorption branch of N2 sorption
isotherms, due to the intrinsic N capillary evaporation [31].
2
To date, two methods for determining the pore entrance size
of mesoporous cage-like materials have been reported [32,33].
One of these methods is extensive high-resolution transmis-

H. Yang et al. / Journal of Catalysis 248 (2007) 204–212
207
Fig. 2. N sorption isotherms of (a) SBA-16(4.9); (b) Co(Salen)/SBA-16(4.9); (c) SBA-16(4.9)-Ph; (d) Co(Salen)/SBA-16(4.9)-Ph; (e) SBA-16(5.4);
2
(f) SBA-16(5.4)-Ph; (g) Co(Salen)/SBA-16(5.4)-Ph; (h) SBA-16(5.9); (i) SBA-16(5.9)-Ph; and (j) Co(Salen)/SBA-16(5.9)-Ph.
Table 1
Surface area, pore volume and cage diameter of the cage-like materials from
N sorption analysis
2
a
b
Mesoporous materials
Surface area Pore volume
Cage diameter
(nm)
2
3
(m /g)
(cm /g)
SBA-16(4.9)
Co(Salen)/SBA-16(4.9)
SBA-16(4.9)-Ph
Co(Salen)/SBA-16(4.9)-Ph 238
SBA-16(5.4)
SBA-16(5.4)-Ph
Co(Salen)/SBA-16(5.4)-Ph 207
SBA-16(5.9)
SBA-16(5.9)-Ph
755
287
257
0.65
0.35
0.24
0.24
0.70
0.35
0.24
0.61
0.37
0.37
4.9
4.9
3.3
3.4
5.4
4.4
4.3
5.9
5.2
5.2
748
377
892
432
Co(Salen)/SBA-16(5.9)-Ph 433
a
Single point pore volume calculated at relative pressure P/P of 0.99.
BJH method from adsorption branch.
0
b
Fig. 3. UV–vis spectra of Co(II)(Salen) solution in dichloromethane absorbed
with cage-like materials. Absorption experiment protocol is as followed:
0
.08 g of SBA-16 materials were dispersed in 3.8 mL of CH Cl containing
2 2
sion electron microscopy imaging from different directions.
However, this method is currently applicable only for three-
dimensional ordered materials with sufficiently large ordered
domains. The other method is an indirect method based on
monitoring pore accessibility after modification of the pore
surface with different-sized molecules. This method can give
a range of pore entrance sizes rather than an exact value.
−6
3
.3 × 10 mol of (R,R)-Co(Salen)(II). The mixture was stirred in a sealed
tube for 5 h. After centrifugation, the solution was measured with UV–vis spec-
troscopy.
creased only slightly. This result shows that the dominating
entrance size of SBA-16(4.9) is smaller than the molecular
size of Co(II)(Salen). This is in good agreement with the
conclusion of Jaroniec et al. [25] that the pore entrance size
increases with autoclaving time. The results of the UV–vis
experiments show that only the pore entrance size of SBA-
16(4.9) is suitable for encapsulation of Co(II)(Salen) within the
mesoporous cage. Therefore, the pore entrance sizes of SBA-
16(5.4) and SBA-16(5.9) were further tuned by silylation with
phenyltrimethoxysilane.
Herein we propose an easy method for estimating the pore
entrance size whether it is larger or smaller than a probe mole-
cule, based on an absorption method using Co(II)(Salen) as
a probe molecule. When Co(II)(Salen) in CH Cl solution is
2
2
absorbed into the cage of SBA-16 material, the concentra-
tion of Co(II)(Salen) in the filtrate should be decreased, and
the concentration change can be monitored by UV–vis spec-
troscopy. UV–vis spectra of the filtrates after absorption with
SBA-16(4.9), SBA-16(5.4), and SBA-16(5.9) are shown in
The N sorption isotherms of SBA-16(5.4)-Ph and SBA-
2
16(5.9)-Ph (silylated with phenyltrimethoxysilane) exhibit type
Fig. 3. The UV–vis spectrum of the Co(II)(Salen) in CH Cl2
IV isotherm patterns with H2 hysteresis loops, confirming that
the cage-like structure of SBA-16(5.4) and SBA-16(5.9) was
maintained after the silylation (also shown in Figs. 2B and 2C,
respectively). For comparison, SBA-16(4.9) was also sily-
2
exhibits bands at 360 and 420 nm. After absorption with
SBA-16(5.4) and SBA-16(5.9), the intensities of the bands are
dramatically decreased, indicating that the pore entrance size
of SBA-16(5.4) and SBA-16(5.9) is larger than the molecu-
lar size of Co(II)(Salen). After absorption with SBA-16(4.9),
the intensities of the characteristic bands of Co(II)(Salen) de-
lated with phenyltrimethoxysilane. The N sorption isotherm
2
of SBA-16(4.9)-Ph is of type I (Fig. 2A). After silylation, the
surface area, pore volume, and pore size of SBA-16(n) appar-

208
H. Yang et al. / Journal of Catalysis 248 (2007) 204–212
ently decreased (Table 1). This confirms that the phenyl group
was grafted on the surface of SBA-16. FT-IR spectra of SBA-
with a bulkier group is needed to reduce the pore entrance size
of SBA-16(5.9) to less than the molecular size of Co(Salen).
The foregoing results indicate that SBA-16(4.9), SBA-16(4.9)-
Ph, and SBA-16(5.4)-Ph may be suitable host materials for
trapping Co(Salen) in the mesoporous cage.
1
6(5.4) and SBA-16(5.4)-Ph are presented in Fig. 4. Compared
with SBA-16(5.4), SBA-16(5.4)-Ph exhibits additional peaks at
−
1
−1
3
078, 3058, and 1434 cm . The peaks at 3078 and 3058 cm
are attributed to the C–H stretching vibrations of phenyl group.
−
1
The peak at 1434 cm can be assigned to C=C stretching
vibration of phenyl group. The IR results further confirm suc-
cessful grafting of a phenyl group in SBA-16 by silylation.
UV–vis spectra of the filtrate after absorbing Co(II)(Salen)
with SBA-16(4.9)-Ph, SBA-16(5.4)-Ph, SBA-16(5.9)-Ph are
shown in Fig. 5. After absorption with SBA-16(4.9)-Ph and
SBA-16(5.4)-Ph, the intensities of the bands at 360 and 420 nm
remain almost the same as that of the initial solution of
Co(II)(Salen) in CH2Cl2. This indicates that the pore entrance
size of these materials is small enough to prevent diffusion of
the probe molecule through the pore entrance. These results
show that silylation is an efficient method of reducing the pore
entrance size of SBA-16. When SBA-16(5.9)-Ph is used as the
absorbent, the intensity is dramatically reduced. This suggests
that the pore entrance size of SBA-16(5.9)-Ph is still larger than
the molecular size of Co(II)(Salen). Thus, a silane precursor
3
.2. Synthesis of Co(Salen) in the cage of SBA-16 through the
“
ship in a bottle” method
Co(Salen) was synthesized in the mesoporous cage of SBA-
1
6 using SalenH2 and Co(OAc)2·4H2O through the “ship in
a bottle” method, as illustrated in Scheme 1. The formation
of Co(Salen) in SBA-16 was verified using FT-IR and UV–
vis spectroscopy. The FT-IR spectra of SalenH2, Co(II)(Salen),
Co(Salen)/SBA-16(4.9), and Co(Salen)/SBA-16(5.4)-Ph are
shown in Fig. 6. The FT-IR spectrum of Co(II)(Salen) shows
−
1
the C–H stretching vibrations at 2956 and 2866 cm and
C–H bending vibrations at 1384 and 1360 cm . The char-
acteristic vibration of Co(II)(Salen) was found at 1530 cm
SalenH2 gives the C–H stretching vibrations only at 3000–
−
1
−1
.
−
1
−1
2
800 cm and bending vibrations at 1400–1300 cm . In the
FT-IR spectra of Co(Salen)/SBA-16(4.9) and Co(Salen)/SBA-
1
2
6(5.4)-Ph, the stretching vibrations of C–H at 2956 and
866 cm and the characteristic C=N vibration at 1530 cm
−
1
−1
are clearly observed, similar to Co(II)(Salen). The bending vi-
−1
brations of C–H were also found at 1400–1300 cm with
a slight blue shift. The FT-IR results demonstrate the forma-
tion of Co(II)(Salen) inside SBA-16(4.9) and SBA-16(5.4)-Ph
through the “ship in a bottle” method. The FT-IR spectrum of
Co(Salen)/SBA-16(5.4)-Ph clearly shows the C–H vibration
of Ph group at 3078 and 3058 cm , further confirming the
existence of phenyl group on SBA-16 after the “ship in a bot-
tle” synthesis. No characteristic vibration of Co(Salen) was
found in the FT-IR spectra of Co(Salen)/SBA-16(4.9)-Ph and
Co(Salen)/SBA-16(5.9)-Ph (data not shown); this indicates that
almost no Co(II)(Salen) exists in SBA-16(4.9)-Ph and SBA-
−
1
1
6(5.9)-Ph.
Fig. 4. FT-IR spectra of (a) SBA-16(5.4) and (b) SBA-16(5.4)-Ph (modified
with phenyl group).
Fig. 6. FT-IR spectra of (a) SalenH ; (b) Co(II)(Salen); (c) Co(Salen)/
2
SBA-16(4.9); and (d) Co(Salen)/SBA-16(5.4)-Ph. (The spectra were recorded
after heating the samples at 523 K for 3 h under vacuum.) (A) The spectra
Fig. 5. UV–vis spectra of Co(II)(Salen) in dichloromethane solution absorbed
with cage-like materials. Absorption experiment is the same as that in Fig. 3.
−1
between 2500 and 3200 cm ; (B) the magnified spectra between 1350 and
−1
1700 cm
.

H. Yang et al. / Journal of Catalysis 248 (2007) 204–212
209
Co(Salen)/SBA-16(4.9) and Co(Salen)/SBA-16(5.4)-Ph
were further characterized by UV–vis spectroscopy (Fig. 7).
The UV–vis spectrum of Co(II)(Salen) in CH2Cl2 shows three
SBA-16(4.9), SBA-16(5.4)-Ph can provide a microenvironment
for Co(Salen) more akin to that in solution.
The mesostructure of Co(Salen)/SBA-16(5.4)-Ph was char-
acterized by XRD and TEM. The small angle X-ray diffraction
pattern of Co(Salen)/SBA-16(5.4)-Ph still exhibits (110) and
∗
bands at 360 (the π–π transition of C=N), 420, and 495 nm
∗
(
d–π transition) (Fig. 7). Similar to Co(II)(Salen), three bands
(
200) diffractions of a cubic Im3m structure (Fig. 8A). TEM
image of Co(Salen)/SBA-16(5.4)-Ph is shown in Fig. 8B. The
110) projection corresponding to cubic Im3m structure indi-
at 345, 420, and 495 nm were also found in the UV–vis spec-
tra of Co(Salen)/SBA-16(5.4)-Ph. (SBA-16(5.4)-Ph shows only
one band at 262 due to the phenyl group; b of Fig. 7). This fur-
ther confirms that Co(II)(Salen) was formed inside the cage
of SBA-16(5.4)-Ph. For Co(Salen)/SBA-16(4.9), only a broad
band at around 380 nm was observed instead of the three
well-defined bands, similar to the Co(Salen) loaded on MCM-
(
cates that the mesoporous structure of SBA-16(5.4) was main-
tained after the “ship in a bottle” synthesis.
The pore volume and surface area of Co(Salen)/SBA-16(4.9)
and Co(Salen)/SBA-16(5.4)-Ph are sharply decreased (Table 1)
compared with the parent materials SBA-16(4.9) and SBA-
4
1 by the adsorption method [34]. This may be due to the
1
6(5.4)-Ph. This could be attributed to the decreased cage
strong interaction between Co(Salen) and the surface hydroxyl
group of unmodified SBA-16(4.9) [35]. Compared with the
free Co(II)(Salen) in solution, only a slight blue shift of the
space due to the existence of Co(Salen) in the mesoporous
cage. The N sorption isotherms of Co(Salen)/SBA-16(4.9)-
2
Ph and Co(Salen)/SBA-16(5.9)-Ph are almost overlapped with
their parent materials SBA-16(4.9)-Ph and SBA-16(5.9)-Ph
∗
band due to the π–π transition of C=N (from 360 to 345 nm)
was observed for Co(Salen)/SBA-16(5.4)-Ph. The foregoing re-
sults probably indicate that surface modification by a Ph group
can greatly reduce the strong interaction of Co(Salen) with
the hydrophilic microenvironment of SBA-16. Compared with
(Figs. 2A and 2C). In addition, the surface area and pore vol-
ume of SBA-16(4.9)-Ph and SBA-16(5.9)-Ph remain almost
the same before and after the “ship in a bottle” synthesis (Ta-
ble 1), indicating that few Co(Salen) molecules exist inside the
cages, consistent with the results of FT-IR and UV–vis char-
acterization. Moreover, the adsorption and desorption branches
of N2 sorption isotherm of Co(Salen)/SBA-16(5.4)-Ph do not
close at low relative pressure (P/P0 < 0.4). That is a common
phenomenon often observed in the case of mesoporous silicas
containing large amounts of organic species [25].
The results of FT-IR, UV–vis, and N2 sorption analy-
sis show that the chiral Co(Salen) exists in the mesoporous
cages of SBA-16(4.9) and SBA-16(5.4)-Ph. The pore entrance
size of SBA-16(4.9)-Ph is smaller than the molecular size of
Co(Salen). Therefore, the fact that no Co(Salen) was formed
in the cage of SBA-16(4.9)-Ph probably indicates that the pore
entrance size of SBA-16(4.9)-Ph is also smaller than the mole-
cular size of SalenH2. For SBA-16(5.9)-Ph, the metal complex
formed in the cage of SBA-16(5.9)-Ph is almost washed off
during the synthesis process, because the pore entrance size
of SBA-16(5.9)-Ph is too large to confine Co(Salen) within
the mesoporous cage, as evidenced by the absorption results.
Fig. 7. UV–vis spectra of Co(II)(Salen) in CH Cl (DCM), Co(Salen)/SBA-
2
2
16(4.9), SBA-16(5.4)-Ph (diffusion reflection), and Co(Salen)/SBA-16(5.4)
(diffusion reflection).
Fig. 8. XRD pattern and TEM image of SBA-16 after synthesis of Co(Salen) in SBA-16. (A) XRD pattern of Co(Salen)/SBA-16(5.4)-Ph; (B) TEM image of
Co(Salen)/SBA-16(5.4)-Ph.

210
H. Yang et al. / Journal of Catalysis 248 (2007) 204–212
The pore entrance size of mesoporous cage-like material deter-
mines the synthesis of metal complex in the mesoporous cage
successfully or not through the “ship in a bottle” method. The
pore entrance size should be varied depending on the molecu-
lar size of the metal complex and the fragment for constructing
the metal complex.
dicating that nearly no metal complex was trapped in SBA-
16(4.9)-Ph and SBA-16(5.9)-Ph.
The activity and enantioselectivity of Co(Salen)/SBA-
16(4.9) are much lower than those of Co(Salen)/SBA-16(5.4)-
Ph. During the catalytic reaction, the Co(Salen)/SBA-16(4.9)
catalyst aggregated together and could not be well dispersed
in the reaction system even under vigorous stirring. In con-
trast, Co(Salen)/SBA-16(5.4)-Ph can be well dispersed in the
reaction system under the reaction conditions. This differ-
ence probably results from the different surface properties of
Co(Salen)/SBA-16(4.9) and Co(Salen)/SBA-16(5.4)-Ph. The
surface of Co(Salen)/SBA-16(4.9) without organic modifi-
cation is relatively hydrophilic, hindering its dispersion in
lipophilic propylene oxide. In contrast, the surface hydropho-
bicity of Co(Salen)/SBA-16(5.4)-Ph is greatly increased after
modification with phenyl group, which makes them “dissolve”
into propylene oxide. The enhanced hydrophobicity may ac-
count in part for the significantly improved catalytic activity
of Co(Salen)/SBA-16(5.4)-Ph. Therefore, we can see that the
catalytic performance of Co(Salen) trapped in the mesoporous
cage through the “ship in a bottle” method is greatly dependent
on the surface properties of the host materials. Of course, the
different host–guest interaction imposed by SBA-16 and modi-
fied SBA-16 also may affect the reactivity.
3
.3. Asymmetric ring-opening of epoxides catalyzed by
Co(Salen)/SBA-16 synthesized through the “ship in a bottle”
method
The performances of catalysts prepared through the “ship in
a bottle” strategy were tested in the asymmetric ring-opening of
epoxides [36,37]. Before the reaction, the Co(II)(Salen) accom-
modated in SBA-16 was oxidized to Co(III)(Salen)(OAc) in the
presence of acetic acid under air (as described in Section 2). The
activities and enantioselectivities were summarized in Table 2.
The asymmetric ring-opening of propylene oxide was car-
ried out under solvent-free conditions. For Co(Salen)/SBA-
1
6(4.9), only 3% diol yield with 86% ee was obtained. The
activity and enantioselectivity are lower than that of the ho-
mogeneous Co(Salen) (Jacobsen catalyst) (Table 2, entry 1).
However, for Co(Salen)/SBA-16(5.4)-Ph, 50% diol yield with
9
6% ee is obtained under identical reaction conditions. The
isolated yield can reach 45% (Table 2, entry 4). Although the
activity of Co(Salen)/SBA-16(5.4)-Ph is slightly lower than
that of its homogeneous counterpart, the enantioselectivities
of the two catalysts are similar. Very low activity is observed
on Co(Salen)/SBA-16(4.9)-Ph, while Co(Salen)/SBA-16(5.9)-
Ph shows a diol yield of 10% with 96% ee due to the Co(Salen)
residue remaining in SBA-16(5.9)-Ph. These catalytic results
are consistent with the foregoing characterization results, in-
The catalytic properties of the catalysts also were examined
in the asymmetric ring-opening of epichlorohydrin. In solvent-
free conditions, Co(Salen)/SBA-16(4.9) gave higher activity
(28% diol yield with 88% ee) in the ring-opening of epichloro-
hydrin. Co(Salen)/SBA-16(4.9) could be well dispersed in
epichlorohydrin. This differs from the case of propylene oxide,
because epichlorohydrin is relatively hydrophilic in comparison
with propylene oxide. As expected, Co(Salen)/SBA-16(5.4)-
Table 2
Asymmetric ring-opening of epoxides catalyzed with Co(Salen) and Co(Salen) trapped in SBA-16
a
b
Entry
R
Catalysts
Solvent
Time (h)
Diol (yield%)
Diol (% ee)
c
1
2
3
4
5
CH
Co(Salen) (Jacobsen catalyst)
Co(Salen)/SBA-16(4.9)
Co(Salen)/SBA-16(4.9)-Ph
–
–
–
–
–
12–14
20
20
20
20
45
3
Trace
50 (45 )
99
86
3
d
e
Co(Salen)/SBA-16(5.4)-Ph
Co(Salen)/SBA-16(5.9)-Ph
96
96
d
10
6
7
8
9
0
1
2
3
CH Cl
2
Co(Salen)/SBA-16(4.9)
Co(Salen)/SBA-16(5.4)-Ph
Co(Salen) (Jacobsen catalyst)
Co(Salen)/SBA-16(4.9)
–
–
20
15
12–14
20
20
20
20
23
28
45
40
23
Trace
46
7
88
87
95
89
f
THF
THF
THF
THF
THF
THF
d
1
1
1
1
Co(Salen)/SBA-16(4.9)-Ph
Co(Salen)/SBA-16(5.4)-Ph
Co(Salen)/SBA-16(5.9)-Ph
87
87
92
d
g
Co(Salen)/SBA-16(5.4)-Ph (5th)
40
a
b
c
d
e
f
Diol yield, GC analysis.
From GC analysis, derived with dimethoxypropane in the presence of p-toluenesulfonic acid.
Data obtained from Ref. [37], 0.2 mol% equiv of catalyst, 0.45 mol equiv of H O, room temperature.
The weight of solid catalyst is the same to Co(Salen)/SBA-16(5.4)-Ph.
The isolated yield.
2
Data obtained from Ref. [37], 0.5 mol% cat, 0.45 mol equiv of H O, THF as a solvent.
The fifth run of Co(Salen)/SBA-16(5.4)-Ph.
2
g

H. Yang et al. / Journal of Catalysis 248 (2007) 204–212
211
Fig. 9. Recycle test of Co(Salen)/SBA-16(5.4)-Ph in asymmetric ring-opening of propylene oxide.
Ph could also “dissolve” in solvent-free conditions and gave
a 45% diol yield with 87% ee after 15 h in the asymmetric
ring-opening of epichlorohydrin. The catalytic results indicate
that the degree of dispersion of the solid catalyst in the reaction
mixture has a direct effect on catalytic activity.
alyst, we added extra epichlorohydrin to the filtrate after the
removal of Co(Salen)/SBA-16(5.4)-Ph and found that no more
diol is produced under identical conditions.
4. Conclusion
THF is usually a good solvent for the homogeneous asym-
metric ring-opening of epichlorohydrin [37]. With THF as
the solvent, the asymmetric ring-opening of epichlorohydrin
with Co(Salen)/SBA-16(4.9) gives a 23% diol yield with
The chiral metal complex Co(Salen) was accommodated in
the cage-like mesoporous material SBA-16, using the “ship in
a bottle” strategy. Fine-tuning of the pore entrance size of the
mesoporous cage-like silica was found to be the key factor for
successful trapping of Co(Salen) in the cage of the mesoporous
materials. The pore entrance size of SBA-16 can be modified
by a silylation reaction according to the molecular sizes of
Co(Salen) catalyst, reactants, and products. Chiral Co(Salen)
trapped in SBA-16 exhibits a 50% diol yield with 96% ee
and a 46% diol yield with 87% ee for the asymmetric ring-
opening of propylene oxide and epichlorohydrin, respectively.
The enantioselectivity for the solid catalyst is comparable to
that for the homogeneous counterpart. Meanwhile, the silyla-
tion reaction also can modify the inner surface properties of
SBA-16. The activity for the catalyst trapped inside SBA-16
can be increased significantly when the surface of SBA-16 is
modified with organic groups. No significant decrease in activ-
ity or enantioselectivity was observed even after ten cycles of
the solid catalyst in the asymmetric ring-opening of propylene
oxide. The “ship in a bottle” strategy, combined with tailor-
ing the pore entrance size and modifying the surface properties
of mesoporous cage-like material, is a promising method for
preparing efficient and stable heterogeneous catalysts.
8
9% ee after 20 h. Co(Salen)/SBA-16(5.4)-Ph shows a 46%
diol yield with 87% ee under similar conditions. The ac-
tivity of Co(Salen)/SBA-16(5.4)-Ph is higher than that of
Co(Salen)/SBA-16(4.9). The increased activity may be at-
tributed to the improved surface hydrophobicity of Co(Salen)/
SBA-16(5.4)-Ph and the reduced interaction of Co(Salen) with
host materials, SBA-16(5.4)-Ph. Co(Salen)/SBA-16(4.9)-Ph
and Co(Salen)/SBA-16(5.9)-Ph still exhibits low activity, as in
the case of propylene oxide. The catalytic activity of the solid
catalysts is lower with THF as solvent than without THF. This
may be explained by the diffusion competition of the solvent
and the substrate.
The stability of Co(Salen)/SBA-16(5.4)-Ph was investigated
in the asymmetric ring-opening of propylene oxide (Fig. 9). The
enantioselectivity of the recovered Co(Salen)/SBA-16(5.4)-Ph
remains almost constant (96% ee) even after the ninth recy-
cling. The activity of the recovered Co(Salen)/SBA-16(5.4)-Ph
is slightly decreased compared with that of the fresh one. The
deceased activity is due in part to the loss of solid catalysts dur-
ing the recycling process.
Co(Salen)/SBA-16(5.4)-Ph also can be recovered efficiently
for the ring-opening of epichlorohydrin in THF (Table 2). For
the fifth run of the catalyst, a 40% diol yield with 92% ee still
can be obtained. No apparent loss of activity and enantioselec-
tivity was observed, confirming that Co(Salen)/SBA-16(5.4)-Ph
has high stability during the catalytic process. To further con-
firm that the reaction was catalyzed by the heterogeneous cat-
Acknowledgments
This work was supported by the National Natural Science
Foundation of China (grants 20321303 and 20673113), the Na-
tional Basic Research Program of China (grant 2003CB615803),
Programme for Strategic Scientific Alliances between the China

212
H. Yang et al. / Journal of Catalysis 248 (2007) 204–212
and the Netherlands (2004CB720607) and the Knowledge In-
novation Program of the Chinese Academy of Science (grant
DICP K2006B2).
[18] P. Gelin, Y.B. Taarit, C. Naccache, J. Catal. 59 (1979) 357.
[19] S. Ernst, H. Disteldorf, X. Yang, Microporous Mesoporous Mater. 22
(1998) 457.
[20] C. Heinrichs, W.F. Hölderich, Catal. Lett. 58 (1999) 75.
[
21] F. Algarra, M.A. Esteves, V. Fornés, H. García, J. Primo, New J. Chem.
References
22 (1998) 333.
[22] Y. Tanamura, T. Uchida, N. Teramae, M. Kikuchi, K. Kusaba, Y. Onodera,
[
[
[
1] N. Herron, Inorg. Chem. 25 (1986) 4714.
2] A. Corma, H. García, Eur. J. Inorg. Chem. 6 (2004) 1143.
3] P.P. Knops-Gerrits, D.D. Vos, F.T. Thibault-Starzyk, P.A. Jacobs, Nature
Nano Lett. 1 (2001) 387.
[23] Q. Huo, D.I. Margolese, U. Ciesla, P. Feng, T.E. Gier, P. Sieger, R. Leon,
P.M. Petroff, F. Schüth, G.D. Stucky, Nature 368 (1994) 317.
[24] M.J. Kim, R. Ryoo, Chem. Mater. 11 (1999) 487.
[25] T.W. Kim, R. Ryoo, M. Kruk, K.P. Gierszal, M. Jaroniec, S. Kamiya,
O. Terasaki, J. Phys. Chem. B 108 (2004) 11480.
3
69 (1994) 543.
4] D.E. Devos, J.L. Meinershagen, T. Bein, Angew. Chem. Int. Ed. 35 (1996)
211.
[
2
[5] W. Kahlen, H.H. Wagner, W.F. Hölderich, Catal. Lett. 54 (1998) 85.
[6] S.B. Ogunwumi, T. Bein, Chem. Commun. (1997) 901.
[7] M. Sabater, A. Corma, A. Domenech, V. Fornés, H. García, Chem. Com-
mun. (1997) 1285.
[26] J.M. Kim, Y. Sakamoto, Y.K. Hwang, Y.U. Kwon, O. Terasaki, S.E. Park,
G.D. Stucky, J. Phys. Chem. B 106 (2002) 2552.
[27] J. Fan, C.Z. Yu, J. Lei, Q. Zhang, T.C. Li, B. Tu, W.Z. Zhou, D.Y. Zhao,
J. Am. Chem. Soc. 127 (2005) 10794.
[
[
8] G. Gbery, A. Zsigmond, K.J. Balkus Jr., Catal. Lett. 74 (2001) 77.
9] R.A. Sheldon, I.W.C.E. Arends, H.E.B. Lempers, Catal. Today 41 (1998)
[28] C. Yu, Y. Yu, D.Y. Zhao, Chem. Commun. (2000) 575.
[29] H.Q. Yang, J. Li, J. Yang, Z.M. Liu, Q.H. Yang, C. Li, Chem. Commun.
(2007) 1086.
387.
[
10] C. Schuster, W.F. Hölderich, Catal. Today 60 (2000) 193.
[30] J.F. Larrow, E.N. Jacobsen, J. Org. Chem. 59 (1994) 2939.
[31] M. Kruk, J.R. Matos, M. Jaroniec, Colloids Surf. A Physicochem. Eng.
Aspects 241 (2004) 27.
[11] R.F. Parton, I.F.J. Vankelecom, M.J.A. Casselman, J. Am. Chem. Soc. 116
(1994) 4746.
[12] R.I. Kureshy, N.H. Khan, S.H.R. Abdi, I. Ahmad, S. Singh, R.V. Jasra,
J. Catal. 221 (2004) 234.
[32] Y. Sakamoto, M. Kaneda, O. Terasaki, D.Y. Zhao, J.M. Kim, G.D. Stucky,
H.J. Shin, R. Ryoo, Nature 408 (2000) 449.
[
[
13] K.J. Balkus Jr., M. Eissa, R. Levado, J. Am. Chem. Soc. 117 (1995) 10753.
14] K.J. Balkus Jr., A.K. Khanmamedova, K.M. Dixon, F. Bedioui, Appl.
Catal. A Gen. 143 (1996) 159.
[33] M. Kruk, V. Antochshuk, J.R. Matos, L.P. Mercuri, M. Jaroniec, J. Am.
Chem. Soc. 124 (2002) 124.
[34] G.J. Kim, D.W. Park, Catal. Today 63 (2000) 537.
[35] L. Frunza, H. Kosslick, H. Landmesser, E. Höft, R. Fricke, J. Mol. Catal.
A Chem. 123 (1997) 179.
[
[
15] C. Baleiz aˇ o, H. García, Chem. Rev. 106 (2006) 3987.
16] M. Alvaro, E. Carbonell, M. Esplá, H. García, Appl. Catal. B Environ. 57
(
2005) 37.
17] K.O. Xavier, J. Chacko, K.K. Mohammed Yusuff, Appl. Catal. B Gen. 258
2004) 251.
[36] D.A. Annis, E.N. Jacobsen, J. Am. Chem. Soc. 121 (1999) 4147.
[37] S.E. Schaus, B.D. Brandes, J.F. Larrow, M. Tokunaga, K.B. Hansen, A.E.
Gould, M.E. Furrow, E.N. Jacobsen, J. Am. Chem. Soc. 124 (2002) 1307.
[
(