Design of a chiral mesoporous silica and its application as a host for stereoselective di-π-methane rearrangements

Article abstract of DOI:10.1021/jo047878j

(Chemical Equation Presented) A chiral periodic mesoporous organosilica (ChiMO) was prepared, in which a 1,2-bis-(ureido)-cyclohexyl linker (38 wt %) is introduced into the walls of the hybrid organic-inorganic material. This silica was used as a host for 11-formyl-12-methyldibenzobarrelene (2), and the stereoselectivity of the di-π-methane rearrangement of 2 within this host-guest complex was studied. At low conversions, the only product was the corresponding dibenzosemibullvalene. An enantiomeric excess of 24% at 11% conversion was obtained using the ChiMO as host. These values compare well with those achieved using a system based on conventional faujasites (LiY and NaY) as hosts incorporating a chiral auxiliary. We tested S-phenylglycine, S-proline, S-camphanic acid, and S-mandelic acid as chiral auxiliaries. In contrast to the behavior of the ChiMO material, adsorption of dibenzobarrelene in purely siliceous mesoporous MCM-41 silica (3.2 nm pore size) containing ephedrine failed, a failure that can be explained as arising from the large internal silanol population and high hydrophilicity of the siliceous MCM-41 sample.

Full text of DOI:10.1021/jo047878j

Design of a Chiral Mesoporous Silica and Its Application as a Host
for Stereoselective Di-π-methane Rearrangements
Miriam Benitez,† Gerhard Bringmann,‡ Michael Dreyer,‡ Hermenegildo Garcia,*,†
Heiko Ihmels,*,§ Michael Waidelich,§ and Kathrin Wissel‡
Instituto de Tecnologı´a Quı´mica CSIC-UPV, Av. De los Naranjos, s/n, 46022 Valencia, Spain,
Institut fu¨r Organische Chemie, Universita¨t Wu¨rzburg, Am Hubland, 97074 Wu¨rzburg, Germany, and
Institut fu¨r Organische Chemie, Universita¨t Siegen, Adolf-Reichwein-Strasse 2, 57068 Siegen, Germany
hgarcia@qim.upv.es; ihmels@chemie.uni-siegen
Received November 30, 2004
A chiral periodic mesoporous organosilica (ChiMO) was prepared, in which a 1,2-bis-(ureido)-
cyclohexyl linker (38 wt %) is introduced into the walls of the hybrid organic-inorganic material.
This silica was used as a host for 11-formyl-12-methyldibenzobarrelene (2), and the stereoselectivity
of the di-π-methane rearrangement of 2 within this host-guest complex was studied. At low
conversions, the only product was the corresponding dibenzosemibullvalene. An enantiomeric excess
of 24% at 11% conversion was obtained using the ChiMO as host. These values compare well with
those achieved using a system based on conventional faujasites (LiY and NaY) as hosts incorporating
a chiral auxiliary. We tested S-phenylglycine, S-proline, S-camphanic acid, and S-mandelic acid
as chiral auxiliaries. In contrast to the behavior of the ChiMO material, adsorption of dibenzobar-
relene in purely siliceous mesoporous MCM-41 silica (3.2 nm pore size) containing ephedrine failed,
a failure that can be explained as arising from the large internal silanol population and high
hydrophilicity of the siliceous MCM-41 sample.
Introduction
apparently in most of the cases studied up to now the
interactions between a molecule in its excited state and
a chiral auxiliary are generally too weak to promote high
enantiomeric excesses.
Catalytic enantioselective reactions have been one of
the major achievements in organic chemistry during the
past decades, since the chirality of substoichiometric
amounts of a catalysts may lead to a large amount of
chiral products producing an effective amplification of
chirality present in the catalyst.1,2 In contrast to the
degree of understanding and the remarkable progress in
thermally induced catalytic reactions, the situation in the
field of stereoselective photochemical reactions is still far
from being satisfactory.3,4 The vast majority of known
photochemical reactions do not require any catalyst and
In this regard, a major achievement in the field of
stereoselective photochemical reactions has been the
“ionic-auxiliary concept” developed by Scheffer’s group.5
According to this concept, the achiral chromophore is
connected to a carboxylic acid, and subsequent salt
formation with a chiral amine gives chiral salts that
crystallize in chiral space groups. The solid-state irradia-
tion of these crystals often results in highly stereoselec-
tive transformations. Even though this methodology has
been a significant breakthrough in asymmetric photo-
chemistry and the stereoselectivities achieved are gener-
ally excellent, it is still desirable to expand this meth-
odology in order to avoid the need for the carboxylic acid
† Instituto de Tecnolog´ıa Qu´ımica CSIC-UPV.
‡ Universita¨t Wu¨rzburg.
§ Universita¨t Siegen.
(1) Noyori, R. Angew. Chem., Int. Ed. 2002, 41, 2008-2022.
(2) Sharpless, K. B. Angew. Chem., Int. Ed. 2002, 41, 2024-2032.
(3) Inoue, Y. Chem. Rev. 1992, 92, 741-770.
(4) Rau, H. Chem. Rev. 1983, 83, 535-47.
(5) Gamlin, J. N.; Jones, R.; Leibovitch, M.; Patrick, B.; Scheffer, J.
R.; Trotter, J. Acc. Chem. Res. 1996, 29, 203-9.
10.1021/jo047878j CCC: $30.25 © 2005 American Chemical Society
Published on Web 02/23/2005
J. Org. Chem. 2005, 70, 2315-2321
2315

Benitez et al.
SCHEME 1
and amino functionalities and the rather tedious and
time-consuming crystallization. In this context, host-
guest chemistry has been shown to offer a useful platform
if it is possible to use the host molecule to define a
confined space in which both a chiral auxiliary and the
photoactive system are complexed.6-9 Within such host-
guest systems, the close proximity of the chiral auxiliary
to the reactive center of the chromophore often leads to
asymmetric induction and high stereoselectivities. Zeo-
lites are ideal candidates for host materials because of
their regular porous microstructure,10-12 and because of
remarkable previously reported intrazeolite stereoselec-
tivities using this chiral auxiliary approach.7-9 Neverthe-
less, one limitation of zeolites as hosts for stereoselective
photochemical reactions is that the size of internal voids
in classical zeolites may be too small for the co-inclusion
of auxiliary and the chromophore.13 Even for the so-called
large pore sized zeolites, the diameter of their internal
cavities is about 13 Å, with pore openings of 7.4 Å, and
it would be desirable to explore the use of porous hosts
with larger pore dimensions. Moreover, stepwise co-
inclusion of both the chiral auxiliary and substrate could
lead to an inhomogeneous distribution of the two com-
ponents in the zeolite particle, i.e., not every zeolite cage
contains both a chiral auxiliary and substrate molecule.
In view of the above considerations, we wished to test
mesoporous silicas such as MCM-41, which exhibit
significantly larger pore dimensions than zeolites, as
hosts.14 The MCM-41 structure is formed by an array of
parallel hexagonal channels whose diameter can be
varied by synthesis in the range between 2 and 6 nm.
The MCM-41 channels are so periodically arranged that
MCM-41 silicas exhibit characteristic XRD even though
their walls are constituted by amorphous silica and there
is not a repetitive unit cell. It is worth noting that despite
the extensive work on the use of zeolites as solid matrixes
to perform asymmetric photochemical reactions,7-9 the
same methodology has, to the best of our knowledge, not
yet been applied systematically to mesoporous materials.
Also in this context, an interesting novel type of
material worth examining as host for asymmetric pho-
tochemical reactions is periodic mesoporous organosilicas
(PMOs) in which an organic moiety forms part of the
walls of a strictly regular silica structure with pores in
the mesopore range (>20 Å).15-18 Recently some of us
have reported PMO materials with chiral organic moi-
eties (ChiMOs) and observed that they may be used to
induce chirality (about ∼50% ee) in thermal catalytic
reactions.19 In this paper we compare the conventional
NaY zeolite, MCM-41 silica, and a mesoporous ChiMO
with chiral organic subunits with respect to their ability
to act as host systems for asymmetric photochemical
reactions, namely the di-π-methane rearrangement of the
dibenzobarrelene derivative 2.
Results and Discussion
Synthesis and Photochemistry of Dibenzobar-
relene 2. We chose dibenzobarrelene 2 as the reference
system because it offers the opportunity to investigate
the regioselectivity and stereoselectivity of the di-π-
methane rearrangement. Thus, irradiation may lead in
principle to the dibenzosemibullvalenes 3 and 4, both of
which are chiral (Scheme 1). However, in view of the
precedents in dibenzobarrelene photochemistry, it is
expected that one of the two possible dibenzobullvalenes
(that arising from the most stable biradical) would be
preferentially formed over the other.20,21
Dibenzobarrelene 2 was obtained in 95% yield by acid
catalyzed rearrangement of the known 11,12-bis(hy-
droxymethyl)dibenzobarrelene (1). Irradiation (λ ) 300
nm) of 2 in acetonitrile gave the semibullvalene 3 with
no detectable (1H NMR) traces of 4. Semibullvalene 3 was
isolated in 61% yield by crystallization. The structural
assignment is based on the H and 13C NMR data and
1
especially on the detection of an NOE effect between the
methyl substituent and the proton at C-1 (see Scheme
1). In compound 4 such an NOE would not be expected.
In methanol solution, irradiation of 2 gave the same
result; however, prolonged reaction times led to a reduced
yield of 3 (as determined by gas chromatography) because
of the formation of several unidentified byproducts. In
the solid state, the content of 3 in the reaction mixture
was also significantly lower (68% before isolation) due
to the formation of other unidentified byproducts.
The regiochemistry in the formation of 3 can be
explained as arising from the photochemical generation
of biradical BR1 in the initial step of the photoreaction,
(6) Weiss, R. G.; Ramamurthy, V.; Hammond, G. S. Acc. Chem. Res.
1993, 26, 530-536.
(7) Leibovitch, M.; Olovsson, G.; Scheffer, J. R.; Trotter, J. J. Am.
Chem. Soc. 1998, 120, 12755-12769.
(8) Cheung, E.; Netherton, M. R.; Scheffer, J. R.; Trotter, J. J. Am.
Chem. Soc. 1999, 121, 2919-2920.
(9) Chong, K. C. W.; Sivaguru, J.; Shichi, T.; Yoshimi, Y.; Rama-
murthy, V.; Scheffer, J. R. J. Am. Chem. Soc. 2002, 124, 2858-2859.
(10) Kavarnos, G. J.; Turro, N. J. Chem. Rev. 1986, 86, 401-449.
(11) Scaiano, J. C.; Garc´ıa, H. Acc. Chem. Res. 1999, 32, 783-793.
(12) Thomas, J. K. Chem. Rev. 1993, 93, 301-320.
(13) Meier, W. M.; Olson, D. H.; Baerlocher, C. Zeolites 1996, 17,
1-229.
(14) Garcia, H.; Roth, H. D. Chem. Rev. 2002, 102, 3947-4008
(15) Inagaki, S.; Guan, S.; Ohsuna, T.; Terasaki, O. Nature 2002,
416, 304.
(16) Matos, J. R.; Kruk, M.; Mercuri, L. P.; Jaroniec, M.; Asefa, T.;
Coombs, N.; Ozin, G. A.; Kamiyama, T.; Terasaki, O. Chem. Mater.
2002, 14, 1903.
i.e., the photoinduced vinyl-benzo bridging.20,21 Thus, in
(17) Asefa, T.; MacLachlan, M. J.; Coombs, N.; Ozin, G. A. Nature
1999, 402, 867-871.
(18) Melde, B. J.; Holland, B. T.; Blanford, C. F.; Stein, A. Chem.
Mater. 1999, 11, 3302-3308.
(19) Baleizao, C.; Gigante, B.; Alvaro, M.; Garcia, H.; Corma, A. J.
Catal. 2004, 223, 106-113.
2316 J. Org. Chem., Vol. 70, No. 6, 2005

Stereoselective Photorearrangement
significant decrease in the amount of coadsorbed S-
phenylglycine.
We also tested the stepwise coadsorption of a large
series of chiral auxiliaries prior to attempting the ad-
sorption of dibenzobarrelene 2. In addition to S-phenyl-
glycine, the list included S-proline, S-camphanic acid,
S-phenylalanine, S-mandelic acid, S-aminophenylpro-
panol, S-phenylglicinol, and S-1-phenylethylamine. The
amount of adsorbed chiral auxiliary and 2 was deter-
mined by weight difference between the initial and the
recovered amount. This information was used later to
select the most suitable systems to perform asymmetric
photochemistry in zeolite. Those cases in which the
loadings were too small were not considered further.
In principle the ideal loading of a chiral auxiliary
should be at least one molecule per supercage to ensure
that each substrate molecule is in the close vicinity of a
chiral inductor. Nevertheless, such a high loading cannot
be achieved in conventional microporous zeolites. Usually
the maximum achievable loading by adsorption is about
25-33% of the total supercages.24
FIGURE 1. Two views of the molecular model showing
dibenzobarrelene 2 and S-phenylglycine encapsulated within
an ideal faujasite supercage devoid of charge-balancing cations.
The overlap between the zeolite framework atoms and the
organic molecules is not shown for clarity.
biradical BR1 a formyl-substituted radical center is
formed, whereas the alternative pathway leading to
compound 4 would require the intermediacy of BR2, a
less stabilized biradical in which the corresponding
radical position is substituted with the comparably less
stabilizing methyl substituent.
Mesoporous MCM-41 silicas have much larger pore
dimensions. The structure of MCM-41 is formed by an
array of parallel hexagonal channels.25,26 The pore size
of the MCM-41 silica sample used in this work has a
diameter of 3.2 nm and should be able to accommodate
2 in its interior. Unexpectedly, we were unable to adsorb
significant amounts of dibenzobarrelene 2 in an ephe-
drine-containing MCM-41 sample. Thus, we assume the
low degree of adsorption is caused by the characteristic
large population of silanol groups on the inner surface
of the MCM-41 pores. The resulting hydrophilic internal
surface of the pores is unable to adsorb an organic
hydrophobic molecule such a 2. It may happen that the
chiral auxiliary having functional groups suitable to form
hydrogen bridges could be located at the pore mouth of
the MCM-41 channels interacting with the silanols by
hydrogen bonds and blocking the access of dibenzobar-
relene 2. Again, this situation is far from the ideal
homogeneous distribution of substrate and chiral auxil-
iary within the internal mesopores.
To circumvent this problem, we investigated a novel
type of mesoporous material namely a periodic meso-
porous organosilica (PMO) in which the chiral component
is covalently linked to the silica framework rather than
being coadsorbed in the pores.17,27,28 The synthesis of
these PMOs is accomplished by adding a suitable chiral
disilylalkane to the synthesis gel. Nevertheless, when the
organosilane is a large molecule, the synthesis of a PMO
requires an additional silicon source. Thus, tetraethyl
orthosilicate (TEOS) needs to be added in some propor-
tion to relax the geometrical stress introduced by the
bulky organosilane, thus permitting the formation of the
PMO structure through co-condensation of the organo-
Adsorption of 2 Within Zeolites and Mesoporous
Silica. We next turned to the incorporation of diben-
zobarrelene 2 in zeolites and the investigation of the
photochemistry of the host-guest systems. For this part
of the project, it was necessary to consider that conven-
tional large pore zeolites have a limited void dimension
that may be insufficient to include both the substrate and
the chiral auxiliary.22,23 For instance, NaY faujasite has
a pore system formed by almost spherical cages of 13 Å
diameter interconnected tetrahedrally through four win-
dows of about 0.74 nm; and the organic guest, the chiral
auxiliary, and the charge-balancing cations in order to
impart stereoselectivity have to fit within this space.13
A molecular model of the dibenzobarrelene derivative 2
and S-phenylglycine as chiral inductor encapsulated
inside the supercages of faujasite at the AM1 molecular-
mechanics level is provided in Figure 1. The model
predicts a larger than 20% overlap of the atomic radii
that will make co-inclusion of 2 and S-phenylglycine in
highly crystalline zeolite Y energetically impossible.
Situations different from the one corresponding to the
molecular model provided in Figure 1 can, however, be
envisioned. Thus, it may be that compound 2 and
S-phenylglycine fit inside incomplete cages at lattice
defects or on the external surface or they can interact in
other configurations in which the two molecules are not
completely inside the same zeolite cavity but protruding
throughout the cage windows. In any case, diffusion of
the dibenzobarrelene through the faujasite micropores
may be significantly hindered by the chiral molecules,
which already occupy the binding sites.
The theoretical estimation was confirmed by the
adsorption experiments, because a sample of zeolite NaY
in which S-phenylglycine had previously been adsorbed
did not adsorb significant amounts of dibenzobarrelene
2. Reversing the order of adsorption resulted in a
(24) Corma, A.; Garcia, H. Eur. J. Inorg. Chem. 2004, 1143-1164.
(25) Kresge, C. T.; Leonowicz, M. E.; Roth, W. J.; Vartuli, J. C.; Beck,
J. S. Nature 1992, 359, 710-712.
(26) Beck, J. S.; Vartuli, J. C.; Roth, W. J.; Leonowicz, M. E.; Kresge,
C. T.; Schmitt, K. D.; Chu, C. T.-W.; Olson, D. H.; Sheppard, E. W.;
McCullen, S. B.; Higgins, J. B.; Schlenker, J. L. J. Am. Chem. Soc.
1992, 114, 10834-10843.
(20) Scheffer, J. R.; Trotter, J.; Gudmundsdottir; A. In CRC Hand-
book of Organic Photochemistry and Photobiology; Horspool, W. M.,
Song P.-S., Eds.; CRC Press: Boca Raton, FL, 1995; p 204.
(21) Zimmerman, H. E.; Sulzbach, H. M.; Tollefson, M. B. J. Am.
Chem. Soc. 1993, 115, 6548.
(22) Corma, A.; Garcia, H. Chem. Commun. 2004, 1443-1459.
(23) Davis, M. E. Acc. Chem. Res. 1993, 26, 111-115.
(27) Inagaki, S.; Guan, S.; Fukushima, T.; Oshuna, T.; Terasaki, O.
J. Am. Chem. Soc. 1999, 121, 9611-9614.
(28) MacLachlan, M. J.; Asefa, T.; Ozin, G. A. Chem. Eur. J. 2000,
6, 2507-2511.
J. Org. Chem, Vol. 70, No. 6, 2005 2317

Benitez et al.
SCHEME 2
SCHEME 3
silane and TEOS. Preparation of a homogeneous gel
ensures a homogeneous distribution of the chiral disilyl-
alkane throughout the resulting PMO. This homogeneous
distribution has been previously demonstrated in related
PMOs by XRD and by high-resolution electron micros-
copy and it becomes close to the ideal situation in terms
of a uniform distribution of the chiral auxiliary through
the whole host particle.
FIGURE 2. Powder XRD patterns of as-synthesized CH PMO
solids at two different 6-to-TEOS molar ratios: (a) 5:95 and
(b) 20:80.
Following this synthetic approach, we prepared a
CH PMO sample in which chiral cyclohexadiyl units
(CH) are connected through urea groups to the silica
surface within the cavity.29 The required bis(triethoxysi-
lyl)-substituted cyclohexane derivative 6 was obtained by
reacting enantiopure trans-1R,2S-1,2-diaminocyclohex-
ane (5) with 3-isocyanatopropyltriethoxysilane (Scheme
2). Preparation of CH PMO was accomplished with
cetyltrimethylammonium (CTA+) as the template and
chiral cyclohexadiamine derivative 6 and TEOS as the
silicon sources. Different solids with varying molar ratios
of 6 and TEOS were prepared in order to find the
optimum composition, i.e., the one that allows maximal
loading of the chiral auxiliary with maintained ordering
of a PMO structure. The molar ratios of the components
in the precursor gel were 1.00 Si:0.12 CTABr:8.0 NH3
(20%):114 H2O, where Si corresponds to the total number
of silicon atoms from 6 and TEOS. A model of the
structure of CH PMO material showing the hexagonal
channel structure incorporating the cyclohexadiyl units
is presented in Scheme 3.
FIGURE 3. N2 adsorption (9)-desorption (0) isotherms for
CH PMO (6-to-TEOS molar ratio 20:80). The inset shows the
pore size distribution for this sample.
The periodic structure of the resulting PMO with the
highest CH loading (CH PMO-40 at a weight/weight
ratio of 6 to TEOS of 1 to 1) was studied by powder X-ray
diffraction analysis, which reveals characteristic peaks
of a hexagonal arrangement at 2θ 2° (Figure 2). The
crystallinity of the solid is preserved after removal of
CTA+ by solid-liquid extraction. Isothermal gas adsorp-
tion was found to be of type IV according to the IUPAC
nomenclature and corresponds to a mesoporous material
(Figure 3). Thus, as nitrogen gas adsorption progresses
the gas molecules start to cover the solid surface until
at full coverage a monolayer is formed (first part of the
plot in Figure 3 up to a relative partial pressure of 0.35).
Once a monolayer coverage is reached, a sudden gas
adsorption occurs due to the complete filling of the solid
mesopores by capillarity (jump occurring in the 0.35-
0.4 region of the plot of Figure 3). Beyond this point,
further gas uptake corresponds to the filling of solid
macropores. Thus, from the plot of Figure 3, the partial
pressure at which the sudden gas adsorption occurs is
related to the surface area, the verticality of the sudden
adsorption indicates a narrow pore size distribution, and
the volume of gas uptake in this jump is correlated with
the pore volume. By applying the Brunauer-Teller
(29) De Vos, D. E.; Dams, M.; Sels, B. F.; Jacobs, P. A. Chem. Rev.
2002, 102, 3615-3640.
2318 J. Org. Chem., Vol. 70, No. 6, 2005

Stereoselective Photorearrangement
formalism30,31 to the plot of Figure 3, the surface area
was estimated to be 878 m2 g-1. The pore size distribution
gives a diameter of about 2.7 nm as shown in the inset
of Figure 3. The CH content was determined by combus-
tion analysis (C and N), whereby it was estimated that
38% of the material by weight corresponds to component
6, i.e., 75% of the maximum possible CH loading accord-
ing to the gel composition. To put these characteristics
of CH PMO into context, the BET surface area of NaY
was 410 m2 g-1 and typical organic loadings are below
15 wt %.
TABLE 1. Photoreactivity of Dibenzobarrelene 2 in
Faujasites Containing a Chiral Auxiliary or in CH PMO
T
t
conva eeb
entry
host
[°C] [min]
[%]
[%]
1
2
3
4
5
6
7
8
NaY/S-proline
NaY/S-proline
LiY/S-proline
LiY/S-proline
NaY/1-S-camphanic acid
NaY/1-S-camphanic acid
NaY/S-mandelic acid
CH MCM-41
25
25
25
25
25
0
35
55
15
30
15
30
30
65
20
39
5
>95
5
9
6
11
10
8
8
4
30
8
25
25
<5
24
The integrity of the urea groups was confirmed by the
characteristic IR-absorption band corresponding to the
urea units (1640 cm-1). The covalent linkage between the
terminal silicon atoms of compound 6 and the silicate
walls was also proved by solid-state MAS CP{1H} 29Si
NMR spectroscopy, revealing a peak at -66 ppm, which
was assigned to the Si atoms connected to the methylene
group (tO3Si-CH2-) and covalently grafted to the walls.
Also, the characteristic Q3 and Q4 peaks of MCM-41 (-90
and -100 ppm) corresponding to the Si atoms tri- and
tetrapodally connected to the silica walls, respectively,
were recorded. By contrast, the 29Si NMR signals of the
starting material 6 appeared at ca. -45 ppm, and an
upfield shift upon substitution of the ethoxy groups by
silyloxy bridges had taken place.32 The physical data of
the solid under investigation are consistent with the
proposed structure for CH PMO, i.e., a typical hexagonal
channel system of MCM-41 structure with a large content
of chiral component 6 grafted to the walls.
a Conversion determined by GC analysis relative to an internal
standard (n-dodecane), error (2% of the given data. b Enantio-
meric excess determined by HPLC-CD analysis on the chiral
stationary phase, error (5% of the given data.
HPLC on a chiral phase are indeed enantiomeric to each
other as evidenced by opposite amplitudes in the CD
trace at a given single wavelength and, furthermore,
permits unequivocal assignment of the enantiomers by
recording full CD spectra in the stop-flow mode. A
summary of the conversion and enantiomeric excess of
product 3 obtained for the di-π-methane rearrangement
of 2 associated with chiral auxiliaries within Y faujasites
is given in Table 1. With proline or mandelic acid as the
chiral additive, ee values of 4-10% were detected. With
1-S-camphanic acid as coadsorbate in sodium faujasite,
an ee of 30% was achieved; however, with increasing
conversion (9%), the enantiomeric excess dropped to 8%.
Noteworthy was the result observed for chiral CH PMO,
in which semibullvalene 3 was formed in 24% ee at 11%
conversion. These values compared favorably with the
results obtained using faujasites as hosts. Note that in
the latter case no coadsorbed chiral auxiliary was neces-
sary. Thus, an obvious advantage of CH PMO is that
solid-liquid extraction only led to the recovery of pho-
toproducts along with unreacted starting material.
Photoreactivity of
2
Within Zeolites and
CH PMO. It was investigated in preliminary experi-
ments whether dibenzobarrelene 2 exhibits the same
photoreactivity in zeolites as in solution. Thus, several
zeolites were loaded with 2 and irradiated as solids.
Within the faujasites (M ) Li, Na, K) the same product
was formed as in solution. Within the acidic dealumi-
nated HY zeolite (Si/Al 14) and mesoporous MCM-41 only
very low conversions were observed (<5%). On amor-
phous silica gel, which was used for comparison, the
conversion and the amount of 3 formed were also
significantly lower than those within the faujasites.
We then turned to zeolites containing coadsorbed chiral
auxiliaries as well as CH PMO. Upon irradiation of 2,
which has been co-associated with chiral auxiliaries
within faujasites, only low or moderate enantiomeric
excesses were obtained as determined by circular dichro-
ism (CD) spectroscopy coupled to HPLC33 on a Chiracel
OD-H phase. HPLC-CD coupling is a powerful analytical
tool for the analysis of the enantiomeric ratios of various
naturally occurring34 or synthetically derived35 chiral
compounds. The online coupling to a CD detector facili-
tates determination of whether the peaks resolved by
Summary
In solution the di-π-methane rearrangement of diben-
zobarrelene 2 at room temperature in the presence of
chiral auxiliaries does not lead to a measurable enanti-
oselectivity. Low to moderate ee values (up to 30 %) may
be obtained by performing this photochemical rearrange-
ment inside the cavities of conventional NaY, but the
actual ee depends on the nature of the chiral auxiliary.
Moreover, within zeolites, a uniform distribution of chiral
auxiliary and coadsorbed substrate is difficult to achieve.
In addition, purely siliceous MCM-41, which exhibits
larger pore sizes than zeolites, was found inadequate as
a host, since it fails to adsorb sufficient amounts of
dibenzobarrelene 2. For better adsorption properties, the
use of materials with mesopores having chiral organic
units covalently grafted to the silica walls (ChiMOs)
appears promising. Although the ideal chiral moiety is
still to be found, a simple ChiMO material having chiral
cyclohexadiyl units and urea linkers exhibits simulta-
neously a high adsorption capacity for dibenzobarrelene
2 and the highest ee value at 11% conversion for the
photorearrangement under study.
(30) Brunauer, S.; Deming, L. S.; Deming, W. E.; Teller, E. J. Am.
Chem. Soc. 1940, 62, 1723-32.
(31) Brunauer, S.; Emmett, P. H.; Teller, E. J. Am. Chem. Soc. 1938,
60, 309-19.
(32) Shea, K. J.; Loy, D. A. Acc. Chem. Res. 2001, 34, 707-716.
(33) For method-oriented work on LC-CD coupling, see: Mannschreck,
A. Chirality 1992, 4, 163-169.
(34) For first implementation of LC-CD into phytochemical analysis,
see: Bringmann, G.; Messer, K.; Wohlfarth, M.; Kraus, J.; Dumbuya,
K.; Ru¨ckert, M. Anal. Chem. 1999, 71, 2678-2686.
(35) For LC-CD analysis of synthetic compounds, see: Mino, T.;
Tanaka, Y.; Yabusaki, T.; Okumura, D.; Sakamoto, M.; Fujita, T.
Tetrahedron: Asymmetry 2003, 14, 2503-2506.
J. Org. Chem, Vol. 70, No. 6, 2005 2319

Benitez et al.
(CDCl3, 200 MHz): δ 1.98 (s, 3H, CH3), 3.88 (s, 1H, CH), 5.04
(s, 1H, CH), 7.04-7.21 (m, 8H, Ar-H), 9.59 (s, 1H, CHO).
NOESY NMR (CDCl3, 600 MHz): Cross-peak between 1.98
and aromatic proton at 7.22 (C-1). 13C NMR (CDCl3, 50 MHz):
15.9 (CH3), 50.8 (CH), 51.3 (CH), 52.1 (Cq), 53.6 (Cq), 120.1
(CHar), 120.3 (CHar), 122.6 (CHAr), 123.7 (CHAr), 125.4 (CHAr),
125.5 (CHAr), 126.3 (CHAr), 126.5 (CHAr), 134.8 (Cq), 137.5 (Cq),
147.7 (Cq), 149.1 (Cq), 196.0 (CHO). MS (70 eV): m/z (%) 246
(19) [M+], 217 (100) [M+ - CHO], 202 (77) [M+ - CHO, CH3],
178 (34) [C14H10+]. HRMS calcd for C18H14O (M+) 246.1047,
found 246.1041.
Preparation of N,N′-Bis[3-(trimethoxysilyl)propylami-
nocarbonyl]-trans-(1R,2R)-(-)-cyclohexadiamine (6). Com-
pound 6 was synthesized by slow addition of a solution of
3-isocyanatepropyltrimethoxysilane in 10 mL of acetonitrile
to a solution of 1R,2R-(-)-1,2-diaminocyclohexane in dry
acetonitrile under inert atmosphere. After the addition, the
solution is stirred magnetically for 2 h. Compound 6 is
obtained by recrystallization from methanol of the residue
after acetonitrile removal under reduced pressure. Found
combustion analysis (%): C 57.25, H 10.70, N 9.65. Calcd for
C26H56N4O8 (%): C 56.52, H 10.14, N 10.14.
Experimental Section
General. NMR: 1H NMR chemical shifts refer to δTMS 0.0.
13C NMR chemical shifts refer to solvent signals (CDCl3:
δ77.0). Cq, CH, CH2, and CH3 were determined using the DEPT
pulse sequence. Elemental analyses of organic compounds were
performed at the University of Wu¨rzburg, Institut fu¨r Anor-
ganische Chemie. Melting points are uncorrected. Photoreac-
tions were carried out at room temperature using low-pressure
mercury lamps (λmax ) 300 nm) in sealed quartz glassware (λ
> 210 nm). Samples were placed ca. 10 cm in front of the lamp.
For solution photoreactions, argon gas was bubbled through
the solutions for at least 30 min prior to irradiation to provide
oxygen-free solutions. The solution photoreactions were per-
formed in acetonitrile or methanol solutions (c ) 10-3-10-2
M), and after removal of the solvent the photolyzate was
analyzed by gas chromatography or by 1H NMR spectroscopy.
The mass balance and the conversion were determined relative
to dimethyl isophthalate (NMR) or n-dodecane (GC) as internal
standard. Solid-state photoreactions were carried out by
placing the solid sample (30-60 µmol) between quartz-glass
slides, which were sealed in argon-flushed polyethylene bags.
Each sample was irradiated (λmax ) 300 nm) at room temper-
ature. The photolyzate was dissolved from the quartz plates
with dichloromethane. After evaporation of the solvent the
reaction mixture was analyzed by gas chromatography, HPLC,
or 1H NMR spectroscopy. Gas chromatography was conducted
with a 5% cross-linked phenylmethylsilicone capillary column,
Preparation of CH PMO.37 Compound 6 was used in
combination with tetraethyl ortosilicate (TEOS) in the syn-
thesis of CH PMO and cetyltrimethylammonium bromide
(CTABr) as the structure-directing agent. The molar propor-
tions of the components in the precursor gel were 1.0 Si:0.12
CTABr:8.0 NH3 (20%):114 H2O.
TEOS and compound 6 were used as a source of Si. Several
6:TEOS molar ratios were tested, and highly structured
materials were obtained below 50:50 weight ratio. After mixing
the reactants at 20 °C and stirring the mixture for 2 h the
resulting gel was transferred to a polyethylene container and
heated at 90 °C for 4 days. The solid obtained was washed
with water and dried in air at 60 °C. The structure-directing
agent was removed by extracting the solid with dilute ethanolic
HCl acid solution at 40 °C for 2 h (20 mL of 0.5 M ethanolic
HCl for 0.5 g of solid).
CH PMO was first characterized by powder X-ray diffrac-
tion (XRD) to test for structure periodicity. XRD patterns were
recorded on a Philips X’pert diffractometer using Nickel
filtered Cu KR radiation with λ ) 1.54178 Å operating at 40
KV and 35 mA. Pore size and surface area of the samples were
measured by nitrogen adsorption on a Micromeritics ASAP
2010 analyzer. Samples were outgassed at 150 °C for 6 h.
Infrared spectra were measured on a Jasco 460 Plus spectro-
photometer. Solid-state 29Si MAS NMR spectra were recorded
on a Bruker 400 spectrometer with samples packed in zirconia
rotors spinning at 5.5 kHz.
General Procedure for Zeolite Loading. (a) Suspen-
sion: Prior to the loading of the zeolites, they were heated in
a oven to remove water (MY, MX, HY; 20 h at 300 °C; silica
gel: 3 h at 120 °C; MCM-41: 15 h at 300 °C; CH PMO: 20 h
at 150 °C). To accommodate dibenzobarrelene 4 or chiral
additives within the host systems, 2.40 g of the latter, which
was still hot, was added to a solution of 163 µmol of diben-
zobarrelene 4 in 80 mL of CH2Cl2 or to 163 µmol of the chiral
compound in 80 mL of MeOH. The resulting suspension was
stirred for 20 h under argon-gas atmosphere at room temper-
ature. The complex was filtered off, washed with dichlo-
romethane (guest: 4) or methanol (chiral guest), and dried in
vacuo (25 °C, 10 mbar). The amount of complexed guest was
determined indirectly by gravimetric analysis of the noncom-
plexed substance in the filtrate.
25 m (Tinitial ) 60 °C, tinitial ) 2 min, rate 15 °C/min, Tfinal
)
300 °C, tfinal ) 20 min). HPLC-CD on a chiral phase: The
HPLC system was coupled to a spectropolarimeter with a 5
mm standard flow cell: column, Daicel Chiracel ODH 250 ×
4.6 mm, 5µm; eluent, n-hexane (90%), 2-propanol (10%), each
modified with trifluoroacetic acid (0.05%); flow, 0.5 mL/min;
the CD trace was detected at 233 nm, UV at 280 nm.
11-Formyl-12-methyl-9,10-dihydro-9,10-ethenoanthra-
cene (2). At 0 °C, 3.00 mL of concentrated H2SO4 was slowly
added to a stirred solution of 4.00 g (16.2 mmol) of 136 in 150
mL of THF. The solution containing 1 was stirred for 1 h at 0
°C and for an additional 20 h at 21 °C, and subsequently 150
mL of CH2Cl2 and 150 mL of H2O were added. The organic
layer was separated, and the aqueous phase was extracted
with CH2Cl2. The combined organic layers were washed with
saturated aqueous NaHCO3 solution, until CO2-generation
stopped. The organic phase was dried with anhydrous Na2-
SO4. Removal of the solvent in a vacuum (40 °C, 10 mbar) gave
3.80 mg (15.4 mmol, 95%) of dibenzobarrelene 1 as a white
solid, which was crystallized from MeOH/CH2Cl2 to give white
needles, mp 234-236 °C (MeOH/CH2Cl2). IR (KBr): ν˜ 1652
cm-1 (CdO). UV (CH3CN): λmax (log ꢀ) 210 nm (4.37), 215
1
(4.36), 273 (3.38), 278 (3.42). H NMR (CDCl3, 200 MHz): δ
2.38 (s, 3H, CH3), 4.96 (s, 1H, CH), 5.78 (s, 1H, CH), 6.99-
7.07 (m, 4H, Ar-H), 7.33-7.39 (m, 4H, Ar-H), 9.87 (s, 1H,
CHO). 13C NMR (CDCl3, 50 MHz): δ 46.9 (CH3), 59.0 (CH),
59.1 (CH), 123.4 (CHar), 123.5 (CHar), 124.8(CHar), 125.6 (CHar),
143.3 (Cq), 143.9 (Cq), 144.8 (Cq), 167.8 (Cq), 185.1 (CHO). MS
(70 eV): m/z (%) 246 (37) [M+], 217 (100) [M+ - CHO], 178
(18) [C14H10+]. HRMS calcd for C18H14O (M+) 246.1045, found
246.1046. Anal. for C18H14O (246.30): calcd (%) C 87.78, H
5.73, found (%) C 87.33, H 5.73.
8b-Methyl-8b,8d-dihydro-4bH-dibenzo[a,f]cyclopropa-
[cd]pentalene-8c-carbaldehyde (3). Under argon-gas at-
mosphere, a solution of 70.0 mg (284 µmol) of 2 in 70 mL of
CH3CN was irradiated in a Rayonet photoreactor (λmax ) 300
nm) for 10 min. The solvent was removed in vacuo (40 °C, 10
mbar) and the remaining yellow solid was crystallized from
MeOH/CH2Cl2 to yield 42.3 mg of dibenzosemibullvalene 3 (172
µmol, 61%) as a white powder, mp 62-63 °C (MeOH/CH2Cl2).
(b) Ground solid: In a mortar 1.0 g of the activated zeolite
was ground thoroughly with 350 µmol of the corresponding
guest substance until a homogeneous solid was obtained. The
solid was washed with dichloromethane (guest: 4) or with
methanol (chiral guest) and subsequently dried in vacuo. The
1
UV (CH3CN): λmax (log ꢀ) 269 nm (4.16), 277 (4.14). H NMR
(36) Chen, J.; Pokkuluri, P. R.; Scheffer, J. R.; Trotter, J. Acta
Crystallogr. Sect. C 1993, C49, 2018-19.
(37) Alvaro, M.; Benitez, M.; Debasish D.; Ferrer, B.; Garcia, H.
Chem. Mater. 2004, 16, 2222-2228.
2320 J. Org. Chem., Vol. 70, No. 6, 2005

Stereoselective Photorearrangement
amount of complexed guest was determined indirectly by
gravimetric analysis of the noncomplexed substance in the
filtrate.
Acknowledgment. This work was generously fi-
nanced by the Deutsche Forschungsgemeinschaft and
the Fonds der Chemischen Industrie (Ph.D. fellowship
to K.W. and supplies to G.B.). M.B. thanks to the Ana
y Jose Royo Foundation for a postgraduate scholarship.
H.G. acknowledges financial support by the Spanish
Ministry of Science and Technology (MAT2003-01226)
and Generalidad Valencia (grupos 03-020).
General Procedure for Photolysis of 4 Within Zeolites.
The zeolite complexes were distributed on the inner walls of
quartz glassware, which were previously wetted with dichlo-
romethane. The flask was then flushed with argon gas for 20
min to remove the solvent and oxygen. The flask was then
sealed and irradiated in a Rayonet photoreactor (λ ) 350 nm).
After irradiation, the photolyzate was extracted with CH2Cl2
(Soxhlet apparatus). The solvent was removed in vacuo and
the reaction mixture was analyzed by gas chromatography.
JO047878J
J. Org. Chem, Vol. 70, No. 6, 2005 2321