Polymer-supported (Salen)Mn catalysts for asymmetric epoxidation: A comparison between soluble and insoluble matrices

Article abstract of DOI:10.1021/ja000692r

This paper describes the synthesis of both soluble and insoluble polymer-supported chiral (salen)Mn complexes and their use in asymmetric epoxidation reactions. These studies were undertaken to establish whether high enantioselectivities could be achieved

Full text of DOI:10.1021/ja000692r

J. Am. Chem. Soc. 2000, 122, 6929-6934
6929
Polymer-Supported (Salen)Mn Catalysts for Asymmetric
Epoxidation: A Comparison between Soluble and Insoluble Matrices
Thomas S. Reger and Kim D. Janda*
Contribution from the Department of Chemistry and The Skaggs Institute of Chemical Biology,
The Scripps Research Institute, 10550 North Torrey Pines Road, La Jolla, California 92037
ReceiVed February 25, 2000
Abstract: This paper describes the synthesis of both soluble and insoluble polymer-supported chiral (salen)-
Mn complexes and their use in asymmetric epoxidation reactions. These studies were undertaken to establish
whether high enantioselectivities could be achieved with a polymer-bound catalyst and if the complexes could
be recovered and reused for multiple cycles. Poly(ethylene glycol) monomethyl ether (MeO-PEG) and non-
cross-linked polystyrene (NCPS) were used as soluble supports while JandaJel and Merrifield resins served as
insoluble supports. Each polymer was linked to the salen catalyst through a glutarate spacer. The soluble
catalysts were recovered by precipitation with a suitable solvent while the insoluble catalysts were simply
filtered from the reaction mixture. Three olefins were utilized as epoxidation substrates: styrene, cis-â-
methylstyrene, and dihydronaphthalene. Best results were obtained with cis-â-methylstyrene as the enantio-
selectivity obtained with each polymer-bound catalyst (86-90%) was equivalent to that achieved with the
analogous commercially available, solution-phase (salen)Mn catalyst 1 (88%). The soluble polymer-supported
catalysts 5 and 6 could be used twice before a decline in yield and enantioselectivity was observed and the
JandaJel attached catalyst 7 could be used for three cycles in some cases. The Merrifield-bound catalyst 8 was
found to lose activity with each use. This work presents the most effective (salen)Mn catalyst that has been
attached to gel-type resin to date and may have practical applications in high-throughput organic chemistry.
Introduction
The growth of both polymer-supported chemistry and asym-
1,2
metric catalysis has been tremendous in the past decade. This
is not surprising considering the industrial need for diverse
libraries of enantiomerically pure compounds for lead identifica-
tion in the drug discovery process. Homogeneous catalysis often
provides the best results in achieving high levels of enantiose-
lectivity whereas heterogeneous catalysis offers the advantages
of simplified product purification and the potential for catalyst
recycling. Clearly, heterogeneous catalysts which can provide
high enantioselectivity would be an important development with
potential applications in industrial and high-throughput organic
chemistry.
Figure 1. Typical salen complexes.
Soon after the development of this reaction, a number of
reports appeared describing efforts to incorporate the salen
ligand into a heterogeneous support as a means to recycle the
chiral catalyst. These approaches can be roughly grouped into
four categories: (1) noncovalent immobilization in zeolites,
3
The Jacobsen epoxidation has recently emerged as a powerful
method for the asymmetric oxidation of unfunctionalized olefins.
The best results are usually achieved with cis-disubstituted
alkenes. This reaction is catalyzed by structurally simple Mn-
4
clays, or siloxane membranes, (2) grafting onto inorganic
(III)-salen complexes and has been optimized in terms of the
5
supports such as silica or MCM-41, (3) copolymerization of a
functionalized salen monomer into an organic polymer, and
catalyst structure and choice of stoichiometric oxidants and other
additives.3 The steric and electronic properties of the salen
framework, composed of two salicylaldehyde units and a chiral
diamine, can be tuned by altering the substituents on either
portion. Both enantiomers of the most general catalyst 1 (Figure
6
b
(4) (a) Ogunwumi, S.; Bein, T. Chem. Commun. 1997, 901. (b) Janssen,
K. B. M.; Laquiere, I.; Dehaen, W.; Parton, R. F.; Vankelecom, I. F. J.;
Jacobs, P. A. Tetrahedron: Asymmetry 1997, 8, 3481. (c) Fraile, J. M.;
Garcia, J. I.; Massam, J.; Mayoral, J. A. J. Mol. Catal. A 1998, 136, 47.
(5) (a) Pini, D.; Mandoli, A.; Orlandi, S.; Salvadori, P. Tetrahedron:
Asymmetry 1999, 10, 3883. (b) Kim, G.-J.; Shin, J.-H. Tetrahedron Lett.
1999, 40, 6827. (c) Zhou, X.-G.; Yu, X.-Q.; Huang, J.-S.; Li, S.-G.; Li,
L.-S.; Che, C.-M. Chem. Commun. 1999, 1789. (d) Piaggio, P.; Langham,
C.; McMorn, P.; Bethell, D.; Bulman-Page, P. C.; Hancock, F. E.; Sly, C.;
Hutchings, G. J. J. Chem. Soc., Perkin Trans. 2 2000, 2, 143.
(6) (a) De, B. B.; Lohray, B. B.; Sivaram, S.; Dhal, P. K. Tetrahedron:
Asymmetry 1995, 6, 2105. (b) De, B. B.; Lohray, B. B.; Sivaram, S.; Dhal,
P. K. J. Polym. Sci. A: Polym. Chem. 1997, 35, 1809. (c) Minutolo, F.;
Pini, D.; Salvadori, P. Tetrahedron Lett. 1996, 37, 3775. (d) Minutolo, F.;
Pini, D.; Petri, A.; Salvadori, P. Tetrahedron: Asymmetry 1996, 7, 2293.
1
) are commercially available.
(
1) Chaiken, I. M.; Janda, K. D., Eds. Molecular DiVersity and
Combinatorial Chemistry; American Chemical Society: Washington, DC,
996.
2) Noyori, R. Asymmetric Catalysis in Organic Synthesis; Wiley: New
York, 1994.
1
(
(3) (a) Zhang, W.; Loebach, J. L.; Wilson, S. R.; Jacobsen, E. N. J. Am.
Chem. Soc. 1990, 112, 2801. (b) For a review, see: Jacobsen, E. N. In
Catalytic Asymmetric Synthesis; Ojima, I., Ed.; VCH: New York, 1993;
Chapter 4.2.
1
0.1021/ja000692r CCC: $19.00 © 2000 American Chemical Society
Published on Web 07/06/2000

6930 J. Am. Chem. Soc., Vol. 122, No. 29, 2000
Reger and Janda
(
4) attachment or build-up of a salen structure to a preformed
7
polymer. The inclusion of a salen structure within a rigid
membrane (approach 1) has been moderately successful for
asymmetric epoxidations. Drawbacks of this strategy include
leaching of the complexes from the membrane and lower
conversions to epoxide products. The use of inorganic supports
(approach 2) has produced varied results, giving promising
enantioselectivity for a narrow range of substrates in some cases
and poor selectivity, conversion, and recyclability in others.
Figure 2. Functionalized soluble polymers.
6a,b
6c,d
Dhal and Salvadori used the third approach wherein a salen
monomer containing pendant vinyl groups was copolymerized
with either ethylene glycol dimethacrylate or styrene/divinyl-
benzene, respectively, to yield polymers where the manganese
centers were localized at cross-links. These catalysts provided
adequate yields of epoxide products and could be recycled for
multiple uses but showed very poor enantioselectivity. This
probably resulted from inaccessibility of the substrate to the
sterically crowded catalytic centers at the cross-links. Sherring-
Figure 3. Cross-linker 3 used in the preparation of JandaJel.
7
a,b
7c,d
polymer and should result in more homogeneous catalysts. The
immediate goals of our studies were (1) to evaluate the effect
of a distribution of polymer supports on the enantioselectivity
of alkene epoxidation and (2) to determine the extent to which
any effective polymer-bound catalysts can be recycled for
repeated use. The results of these efforts are described herein.
ton
and Laibinis
have developed a strategy where a
pendant salen structure is built in a stepwise manner onto a
preformed polymer (approach 4). Good results for one substrate
7a
were achieved with a methacrylate resin while styrene/divinyl-
7
b-d
benzene resins
were less successful as supports. The ability
to recycle these supported catalysts, however, was either not
described or found to be severely limited. Pozzi and co-workers
have described an interesting approach to a recoverable, though
not polymer-supported, epoxidation catalyst by taking advantage
of the relatively new phase-separation and immobilization
Results and Discussion
A. The Effect of the Polymer Support on Enantioselec-
tivity. To get a clear picture of the effect the polymer support
has on the enantioselectivity of the Jacobsen epoxidation
reaction, a variety of commercially available or readily prepared
functionalized polymers were examined. Poly(ethylene glycol)
monomethyl ether (MeO-PEG) and non-cross-linked polystyrene
(NCPS) provide examples of soluble supports with broad
8
technique known as FBS (Fluorous Biphase Systems). They
showed that a salen catalyst containing perfluorinated alkyl
chains could be utilized for enantioselective epoxidation reac-
tions carried out under biphasic (fluorous/organic) conditions.
Although the catalyst could be readily recovered from the
fluorous layer by phase separation techniques, reasonable
enantioselectivities were obtained for just one substrate.
In related work, Jacobsen has described the synthesis of
polystyrene-bound chiral Co(salen) complexes and their ap-
plication to the hydrolytic kinetic resolution of racemic ep-
1
1
solubility profiles. They are each soluble in CH Cl , THF,
2
2
EtOAc, and DMF but can be precipitated from ether (MeO-
PEG) or methanol (NCPS). Furthermore, MeO-PEG is com-
mercially available in a wide range of molecular weights while
functionalized NCPS is prepared by the radical polymerization
9
12
oxides. This catalyst proved to be highly effective and excellent
of styrene and hydroxymethylstyrene. We expected that a
enantioselectivities of ring-opened products were observed. In
this study it was desirable to use high-loading supports since
the ring-opening reactions are believed to occur from a
cooperative effect of two separate catalytic units. This stands
in contrast to the Mn(salen)-catalyzed epoxidation where site
isolation of the catalytic centers is believed to be imperative.
This limits the formation of µ-oxo-Mn(IV) dimers, a potential
deactivation pathway resulting from the reaction of a Mn(III)
and a Mn(V)dO species.10 Accordingly, we expected that low
loading polymers would provide the greatest opportunity to
achieve maximum reactivity and selectivity. Therefore, using
approach 4, we attached a salen ligand containing an appropriate
linker to a low loading preformed polymer. This convergent
strategy stands in contrast to the previously described linear,
soluble polymer-supported salen complex would be an effective
catalyst since it is believed that nonbonded interactions between
the catalyst and approaching olefinic substrate are critical in
3
b
determining enantioselectivity. Heterogenization of one of
these components (generally the catalyst) could therefore be
expected to result in diminished enantiomeric excess (ee) of
the product epoxide. The validity of this reasoning has been
demonstrated by us in a closely related catalytic asymmetric
13
transformation.
As a complement to the soluble polymer approach, two
insoluble polymers were also investigated as supports for the
1
4
(salen)Mn catalyst. The JandaJel resins have been recently
developed in our group and are designed specifically for use in
15
synthetic organic chemistry. Well-defined hydroxyl function-
alized beads were prepared from the suspension copolymeri-
7
stepwise construction of the ligand on the polymer as it reduces
the number of transformations that are carried out on the
1
6
zation of styrene, hydroxymethylstyrene, and the flexible
1
7
tetrahydrofuran-based cross-linker 3. This gel-type resin
(7) (a) Canali, L.; Cowan, E.; Deleuze, H.; Gibson, C. L.; Sherrington,
D. C. Chem. Commun. 1998, 2561. (b) Canali, L.; Sherrington, D. C.;
Deleuze, H. React. Funct. Polym. 1999, 40, 155. (c) Angelino, M. D.;
Laibinis, P. E. Macromolecules 1998, 31, 7581. (d) Angelino, M. D.;
Laibinis, P. E. J. Polym. Sci. A: Polym. Chem. 1999, 37, 3888.
exhibits swelling properties which are superior to those of the
(11) Gravert, D. J.; Janda, K. D. Chem. ReV. 1997, 97, 489.
(12) Bamford, C. H.; Lindsay, H. Polymer 1973, 14, 330.
(13) (a) Han, H.; Janda, K. D. J. Am. Chem. Soc. 1996, 118, 7632. (b)
Han, H.; Janda, K. D. Tetrahedron Lett. 1997, 38, 1527. (c) Han, H.; Janda,
K. D. Angew. Chem., Int. Ed. Engl. 1997, 36, 1731.
(14) JandaJel is a registered trademark of Aldrich Chemical Co.
(15) Toy, P. H.; Janda, K. D. Tetrahedron Lett. 1999, 40, 6329.
(16) Wilson, M. E.; Paech, K.; Zhou, W. J.; Kurth, M. J. J. Org. Chem.
1998, 63, 5094.
(
8) (a) Pozzi, G.; Cinato, F.; Montanari, F.; Quici, S. Chem. Commun.
998, 877. (b) Pozzi, G.; Cavazzini, M.; Cinato, F.; Montanari, F.; Quici,
S. Eur. J. Org. Chem. 1999, 1947.
9) (a) Annis, D. A.; Jacobsen, E. N. J. Am. Chem. Soc. 1999, 121, 4147.
b) Peukert, S.; Jacobsen, E. N. Org. Lett. 1999, 1, 1245.
10) Jacobsen, E. N.; Deng, L.; Furukawa, Y.; Martinez, L. E. Tetra-
hedron 1994, 50, 4323.
1
(
(
(

Use of (Salen)Mn Catalysts for Asymmetric Epoxidation
J. Am. Chem. Soc., Vol. 122, No. 29, 2000 6931
Scheme 1. Synthesis of Polymer-Bound (Salen)Mn Complexes
common commercially available Merrifield and Tentagel res-
Scheme 2. Reaction Conditions for the Epoxidation of (a)
Styrene, (b) cis-â-Methylstyrene, and (c) Dihydronaphthalene
1
5
ins. We hoped that the increased swelling of JandaJel would
result in better solvation of the reaction complex and would
therefore minimize any loss of enantioselectivity resulting from
heterogenization of the catalyst. Finally, hydroxymethyl Mer-
rifield resin, which contains a rigid divinylbenzene cross-linker,
was utilized to determine how the nature of the cross-linking
agent influenced the reactivity of the resin-bound catalyst.
For simple access to these polymer-supported catalysts, a
suitable salen precursor that could be easily and efficiently
attached to a hydroxyl functionality was required and salen
derivative 4 containing a glutaric acid linker was chosen. We
anticipated that the five carbon spacer provided by this linking
group would place the catalyst sufficiently away from the
polymer backbone to allow unimpeded access of the olefinic
substrate to the metal center. Compound 4 was prepared by
9
a
reaction of the known unsymmetrical salen ligand 2 with
glutaric anhydride and DMAP. Attachment of each of the four
polymeric supports to 4 was then accomplished using dicyclo-
hexylcarbodiimide (DCC) coupling conditions to provide the
polymer-bound ligands. Insertion of manganese was performed
the product epoxide is shown in Scheme 2. It should be noted
that the unsymmetrical salen ligand 2 was prepared from (R,R)-
trans-1,2-diaminocyclohexane and that each of the polymer-
supported catalysts produced the same enantiomer in excess as
2
0
the commercial (R,R)-salen catalyst 1.
1
8
as previously described by Jacobsen to give catalysts 5-8.
The results from the epoxidation of these substrates with the
polymer-supported catalysts, the commercial catalyst 1, and the
modified catalyst 9 are shown in Table 1. It is noteworthy that
most of the reactions were complete in 15 min or less even at
low temperatures and only reactions involving Merrifield-bound
catalyst 8 required as much as 1 h to proceed to completion. It
is also interesting to note that, in general, the insoluble catalysts
gave equivalent ee as the soluble catalysts. The best substrate
proved to be cis-â-methylstyrene since the enantioselectivities
obtained with all of the catalysts differed by only a few
Interestingly, the color change of the polymers with each
derivatization provided a qualitative means for assessing the
success or failure of each reaction. All of the polymers are white
solids and their color changes to yellow after ligand attachment
and dark brown after metal insertion. The loading level of each
of the polymer-bound catalysts was determined by elemental
analysis of manganese and/or nitrogen content and ranged from
0.08 to 0.75 mmol/g. Finally, the soluble catalyst 9 was prepared
to evaluate the effect of replacing the tert-butyl group found in
commercial Jacobsen catalyst 1 with the glutarate linker.
Jacobsen has described a number of different oxidation
21
percent. In fact, the soluble polymer-bound catalysts were even
slightly better than the commercial catalyst 1. In all cases the
isolated yield of the epoxide products was high.
3
b
systems that are effective for carrying out the epoxidation.
We chose to utilize conditions which employ m-chloroperben-
zoic acid (m-CPBA) as the oxidant and N-methylmorpholine-
N-oxide (NMO) as an additive since this combination has been
shown to produce the highest enantioselectivities.19 The epoxi-
dation of three alkenes, styrene, cis-â-methylstyrene, and
dihydronaphthalene, was studied and the specific conditions for
each as well as the configuration of the major enantiomer of
To determine any matrix effects on the rate of reaction, the
kinetics of the epoxidation of cis-â-methylstyrene catalyzed by
7 and 9 were compared. We chose to examine JandaJel catalyst
7 since it provided the best results of the insoluble polymer
catalysts. Catalyst 9 contains the same glutarate linker and differs
from 7 only in that it is not bound to a polymer support. Thus,
any effects of the polymer on the reaction kinetics should be
(
17) Crivello, J. V.; Ramdas, A. J. Macromol. Sci., Pure Appl. Chem.
992, A29, 753.
18) (a) Zhang, W.; Jacobsen, E. N. J. Org. Chem. 1991, 56, 2296. (b)
(20) This was determined for styrene oxide by comparing the retention
times of the synthetic epoxide enantiomers on a gas chromatography chiral
column with authentic chiral epoxide samples. The absolute configurations
of the major enantiomer for the other two substrates were deduced by
analogy.
1
(
Larrow, J. F.; Jacobsen, E. N.; Gao, Y.; Hong, Y.; Nie, X.; Zepp, C. M. J.
Org. Chem. 1994, 59, 1939.
(
19) Palucki, M.; McCormick, G. J.; Jacobsen, E. N. Tetrahedron Lett.
(21) All ee values for cis-â-methylstyrene refer to the cis-epoxide which
was generally favored by 6-7:1 over the trans diastereomer.
1
995, 36, 5457.

6
932 J. Am. Chem. Soc., Vol. 122, No. 29, 2000
Reger and Janda
Table 2. Effect of Catalyst 7 Loading on Enantioselectivity
Table 1. Epoxidation Results for Different (Salen)Mn Catalysts
a
a
1
The values for catalyst loading were determined by elemental
1
ee determined by H NMR in the presence of the chiral shift
analysis for manganese content. b ee determined by H NMR in the
presence of the chiral shift reagent, Eu(hfc)
determined by GC analysis using a Chiraldex G-TA chiral column.
b
c
3
reagent, Eu(hfc) . Isolated yield. ee of the cis-epoxide as determined
by GC analysis using a Chiraldex G-TA chiral column.
c
3
.
ee of the cis-epoxide as
which can serve to isolate the catalytic centers of even the
highest loading polymers.
These results represent some of the most effective (salen)-
Mn catalysts to date which are attached to readily available
polymeric supports. Previous efforts to attach salen complexes
to styrene/divinylbenzene polymers may have suffered from
lower enantioselectivities as a result of the lack of any bulky
tert-butyl substituents on one-half of the salen framework.7
Our complexes have a glutarate linked polymer in place of one
tert-butyl group and this does not appear to significantly
compromise the enantioselectivity of the catalyst. In fact, the
glutarate linker appears to be critical in positioning the catalyst
away from the polymer to allow a clear approach for the olefin
toward the metal.
b-d
Figure 4. Rates of epoxidation catalyzed by salen complexes 7 and
B. Recyclability of the Polymer-Supported Catalysts. For
a truly effective polymer-supported catalyst, it is critical that
recovery be simple and efficient and that the recovered catalyst
retain its activity through multiple cycles. It has been established
that the commercial catalyst 1 is unstable to prolonged oxidative
9
: 1 equiv m-CPBA added at time ) 0 and 1 equiv added at time )
7
0 s as indicated by the hashed line.
manifested in this comparison. Aliquots of each reaction were
taken at various time intervals and the conversion (as measured
against an internal standard, dodecane) and ee were measured.
In each case the ee was found to remain constant throughout
the course of the reaction. In these experiments it was critical
that the m-CPBA be added in two portions (1 equiv at time )
2
3
conditions and cannot be recycled. Laibinis has reported that
supported (salen)Mn complexes are also susceptible to decom-
position; however, these reactions were carried out at room
7
c,d
temperature for prolonged times (20 h). We hoped that the
short reaction times and low temperatures used in our studies
would prevent or at least slow catalyst degradation. As each of
the tested catalysts displayed adequate selectivity, attention was
next turned to examining their recovery and reuse.
0
s, 1 equiv at time ) 70 s) to ensure maximum conversion.
Addition of all of the m-CPBA at time ) 0 s resulted in a 20%
decrease in conversion. Greater than 50% conversion to products
was achieved after only 30 s for both reactions and maximum
conversion occurred within 5 min. This indicates that the
insoluble nature of catalyst 7 does not adversely affect the rate
of epoxidation.
The PEG-bound catalyst 5 was recovered by dropwise
addition of the reaction mixture into a stirred solution of cold
ether. The precipitate, which included NMO and chlorobenzoic
acid in addition to the polymer-bound catalyst, was filtered and
dried. This mixture could be used in a second epoxidation
reaction with no appreciable loss of enantioselectivity for the
epoxidation of styrene and cis-â-methylstyrene (Table 3). After
the first recycle, however, precipitation resulted in a gummy,
difficult-to-handle solid that gave poor results in further
reactions. The recovery of the NCPS-bound catalyst 6 was
accomplished as follows: (1) evaporation of the reaction solvent
to leave a crude residue that was taken up in a minimum amount
of THF and (2) dropwise addition of this solution to a stirring
solution of cold methanol. This procedure was much cleaner
than that for the PEG-bound catalyst as the solid precipitate
contained only the supported catalyst while NMO and other
reaction byproducts remained in solution. After filtration and
drying, the recycled catalyst was used in a second run with no
As noted previously, it has been postulated that a low
concentration of catalytic sites would provide the best op-
portunity for high conversion and enantioselectivity by limiting
the formation of inactive catalyst dimers. The earlier literature
examples of Merrifield resin-bound (salen)Mn catalysts had
loadings ranging from 0.15 to 0.65 mmol/g and did not provide
7
high ee of the epoxidation products. To evaluate the importance
of the loading level, we prepared a series of JandaJel-bound
catalysts with loading capacities ranging from 0.10 to 0.75
22
mmol/g. As illustrated in Table 2, the loading level of catalyst
had no effect on the enantioselectivity of the product epoxide.
7
This indicates that the extent to which dimer formation occurs
is not dependent on the catalyst loading. This may be a
consequence of the excellent swelling properties of the resin
(22) The highest possible loading for our system is 1.36 mmol/g, which
corresponds to the inverse of the molecular weight of one functionalized
monomer unit.
(23) Zhao, S. H.; Ortiz, P. R.; Keys, B. A.; Davenport, K. G. Tetrahedron
Lett. 1996, 37, 2725.

Use of (Salen)Mn Catalysts for Asymmetric Epoxidation
J. Am. Chem. Soc., Vol. 122, No. 29, 2000 6933
Table 3. Studies on the Recyclability of the Polymer-Bound
Catalysts
nearly equivalent to those for the commercial catalyst 1;
however, only the JandaJel-bound complex 7 could be used
for as many as three cycles. Apparently, oxidative decomposition
over repeated use gradually renders the catalyst inactive, as has
7
c,d
been noted previously. This inability to recycle the catalysts
indefinitely appears to be a fundamental limitation of salen-
based oxidation catalysts which is difficult to overcome without
sacrificing enantioselectivity. Nevertheless, the JandaJel-bound
catalyst described herein provides the highest ee to date for
epoxidation catalysts attached to gel-type resins and may find
utility in high-throughput organic synthesis since they simplify
product purification.
Experimental Section
2
General. Methylene chloride was dried over CaH . Poly(ethylene
glycol) monomethyl ether (MeO-PEG, MW ) 5000) and (R,R)-
Jacobsen catalyst 1 were purchased from Aldrich and hydroxymethyl
Merrifield resin (0.67 mmol/g) was purchased from Novabiochem. The
2 2
m-CPBA was dried and recrystallized from CH Cl prior to use. All
other solvents and chemicals were obtained from commercial sources
and were used without further purification. NMR spectra were recorded
on Bruker AC-250, AMX-400, and DRX-500 spectrometers.
a
ee determined by 1H NMR in the presence of the chiral shift
b
reagent, Eu(hfc)
3
.
ee of the cis-epoxide as determined by GC analysis
using a Chiraldex G-TA chiral column.
Preparation of Hydroxymethyl-NCPS. Argon was bubbled vigor-
ously through a solution containing styrene (39.45 g, 379 mmol),
hydroxymethylstyrene (0.56 g, 4.2 mmol), and toluene (140 mL). After
change in results. Unfortunately, any further attempt to pre-
cipitate and recycle the catalyst again gave poor results. With
each of the soluble polymer-bound catalysts, a color change
from dark brown to a lighter shade of brown was observed with
each successive use. This likely results from leaching of
manganese caused by oxidative degradation of the salen ligand,
a problem that has been documented by others with insoluble
supported catalysts.7c,d
3
0 min, bubbling was discontinued and AIBN (0.29 g, 1.8 mmol) was
added. The mixture was heated at 90 °C for 24 h. The solvent was
removed and the residue was heated at 50 °C under high vacuum for
3
0 min. The crude material was dissolved in 50 mL of THF and added
dropwise to a cold (∼5 °C) vigorously stirring solution of methanol (1
L). The white precipitate was filtered and dried to give 19.0 g (48%)
1
of polymer with a loading of 0.13 mmol/g as determined by H NMR.
Preparation of Hydroxymethyl-Functionalized JandaJel. A solu-
tion of acacia gum (6.0 g) and NaCl (3.75 g) in water (150 mL) was
placed in a 150 mL flanged reaction vessel equipped with a floating
With limited success in recycling the soluble polymer-bound
catalysts, we hoped the insoluble supports would provide a more
robust environment for the salen framework. Recovery of both
the JandaJel- and Merrifield resin-bound catalyst beads was
easily accomplished by filtration of the reaction mixture. The
JandaJel-derived catalyst 7 could be used three times in the
epoxidation of styrene and cis-â-methylstyrene without a sig-
nificant drop in selectivity. Further attempts to use the catalyst,
however, resulted in poor conversion and enantioselectivity of
the epoxide product. The Merrifield-bound catalyst 8 lost activity
with each use and was essentially ineffective by the third recycle.
It is not entirely clear why the JandaJel-bound catalyst 7 was
slightly more robust than 8 but this may be a consequence of
microenvironment effects imparted by the resin. As observed
in the soluble polymer examples, the brown color of the
insoluble catalysts faded with repeated use, possibly resulting
from the loss of manganese. Indeed, elemental analysis of the
JandaJel-bound catalyst after three uses indicated a 35%
decrease in manganese content. Attempts to reload the metal
into the ligand were unsuccessful and this lends further support
2
magnetic stirrer and deoxygenated by purging with N . A solution of
hydroxymethylstyrene (0.20 g, 1.5 mmol), styrene (10.2 mL, 89.2
mmol), cross-linker 317 (0.53 g, 1.81 mmol), and benzoyl peroxide (0.15
g) in chlorobenzene (10 mL) was injected into the rapidly stirred
aqueous solution. This mixture was heated at 85 °C for 16 h. The crude
polymer was collected by filtration and rinsed sequentially with water,
methanol, THF, ether, and hexanes. The beads were dried in vacuo
and a loading of 0.13 mmol/g was determined by UV analysis following
Fmoc derivatization and cleavage. Higher loading resins were prepared
by increasing the proportion of hydroxymethylstyrene in the mixture
of monomers.
Preparation of Salen Glutarate (mono-ester) 4. Glutaric anhydride
(
137 mg, 1.2 mmol) was added to a solution containing unsymmetrical
salen 2 (506 mg, 1 mmol) and DMAP (146 mg, 1.2 mmol) in CH Cl
5 mL). The mixture was stirred for 15 h at room temperature and
2
2
(
then concentrated under reduced pressure. The crude oil was purified
by silica gel chromatography (CH Cl /methanol ) 95/5) to give 4 (390
2
2
1
mg, 63%) as a yellow foam; H NMR (400 MHz, CDCl ) δ 1.27 (s,
3
9H), 1.38 (s, 9H), 1.40 (s, 9H), 1.75-2.06 (m, 10H), 2.49 (t, 2H, J )
7
.3 Hz), 2.60 (t, 2H, J ) 7.3 Hz), 3.33 (br s, 2H), 6.76 (d, 1H, J ) 2.9
to the prospect that the ligand undergoes decomposition under
the reaction conditions.24
Hz), 6.92 (d, 1H, J ) 2.9 Hz), 6.98 (d, 1H, J ) 2.2 Hz), 7.31 (d, 1H,
J ) 2.2 Hz), 8.24 (s, 1H), and 8.31 (s, 1H); 13C NMR (100 MHz,
Conclusion
CDCl ) δ 24.2, 29.1, 29.4, 31.4, 33.1, 33.2, 34.0, 34.8, 34.9, 72.1, 72.4,
3
1
1
+
17.7, 118.2, 121.3, 122.7, 125.9, 126.9, 136.3, 138.6, 139.4, 141.3,
We have described the synthesis and evaluation of soluble
and insoluble polymer-supported (salen)Mn complexes for use
in asymmetric epoxidation reactions. The enantioselectivities
derived from the soluble and JandaJel polymeric catalysts were
57.9, 158.2, 164.6, 165.8, and 171.9; HRMS calcd for [C37
H ] 621.3898, found 621.3881.
H N O
52 2 6
+
General Procedure for the Preparation of the Soluble (Salen)-
Mn Catalysts 5 and 6. Dicyclohexylcarbodiimide (2.1 equiv) was
added to a solution of the soluble polymer (1 equiv), salen 4 (2 equiv),
(24) To further analyze the degradation of this system, a control
experiment was run in which the reaction mixture, after filtration of the
polymeric catalyst, was reused without the addition of fresh catalyst. Further
formation of epoxide products would indicate that the salen catalyst was
cleaved intact from the resin while the absence of additional epoxide
formation would be indicative of manganese leaching and/or ligand
decomposition. It was found that recycling of the crude reaction did not
lead to the formation of additional epoxide product.
2 2
and DMAP (0.5 equiv) in CH Cl (40 mL/mmol polymer). The reaction
mixture was stirred at room temperature for 24 h and then filtered
through a pad of Celite to remove the urea byproduct. The filtrate was
concentrated and then added dropwise to a stirring solution of 150 mL
of cold ether (MeO-PEG-bound ligand) or methanol (NCPS-bound
ligand). The yellow precipitate was filtered and dried in vacuo.

6934 J. Am. Chem. Soc., Vol. 122, No. 29, 2000
Reger and Janda
A solution of the polymer-bound ligand (1 equiv) in DMF (40 mL/
A solution containing the above salen ligand (25 mg, 0.039 mmol)
mmol) was heated to 90 °C. Mn(OAc)
then added at which point the yellow solution turned dark brown. After
0 min, air was bubbled into the solution and after an additional 30
min, LiCl (3 equiv) was added to the mixture which was then cooled
to room temperature over 2 h. The solution was concentrated to dryness,
2
‚4H
2
O (3 equiv) in water was
in DMF (5 mL) was heated to 90 °C. Mn(OAc)
2
2
‚4H O (29 mg, 0.12
mmol) in water (2 mL) was added at which point the reaction turned
dark brown. After 30 min, air was bubbled into the solution and after
an additional 30 min, saturated aqueous NaCl (1 mL) was added and
the mixture was cooled to room temperature over 2 h. The solvent
was removed under reduced pressure and the crude residue was taken
3
taken up in CH
2
Cl
2
, and filtered. The filtrate was concentrated to ∼5
mL and added dropwise to an ether solution to precipitate 5 or a
methanol solution to precipitate 6. The resulting brown solids were
filtered and dried. Elemental analysis indicated a Mn content of 0.86%
for 5 and 0.46% for 6 corresponding to loadings of 0.16 and 0.08 mmol/
g, respectively.
up in CH
with water, dried over Na
2
Cl
2
and washed with brine. The organic layer was washed
SO , concentrated, and dried in vacuo to
2
4
provide 9 (25 mg, 89%) as a black solid.
General Procedure for Epoxidation Reactions.19 A solution con-
taining the alkene (1 mmol), NMO (5 mmol), and the (salen)Mn catalyst
General Procedure for the Preparation of Insoluble (Salen)Mn
Catalysts 7 and 8. DCC (2.1 equiv) was added to a solution of the
insoluble resin (1 equiv), salen 4 (2 equiv), and DMAP (0.5 equiv) in
2 2
(0.04 mmol) in CH Cl (6 mL) was cooled to the desired temperature
(see Scheme 2). m-CPBA (2 mmol) was added in three equal portions
over 2 min and the reaction progress was monitored by TLC. Recovery
of catalyst 5 was accomplished by dropwise addition of the reaction to
a stirring solution of cold ether (150 mL) and filtration. Recovery of 6
was accomplished by evaporation of the reaction solvent and addition
of 4 mL of THF to the crude material followed by its dropwise addition
to a stirring solution of cold methanol (150 mL) and filtration. Recovery
of 7 and 8 was accomplished by filtration of the reaction mixture. In
each case, the filtrate, which contained the product epoxide, was washed
CH
at room temperature for 24 h and filtered through a medium frit. The
collected solid was washed sequentially with water, methanol, CH
Cl , ether, and hexanes to give yellow beads.
A solution containing the polymer-bound ligand (1 equiv) in DMF
20 mL/g of resin) was heated to 90 °C. Mn(OAc) ‚4H O (3 equiv) in
2 2
Cl (15 mL/g of resin). The reaction mixture was vigorously stirred
2
-
2
(
2
2
water was then added at which point the solution turned a dark brown
color. After 30 min, air was bubbled into the solution and after an
additional 30 min, saturated aqueous NaCl (3 mL/g of resin) was added
to the mixture which was then cooled to room temperature over 2 h.
The solution was filtered through a medium frit and the collected solid
with 1 N NaOH and brine. The organic layer was dried over MgSO
4
and concentrated under reduced pressure to give a crude product that
was further purified by passage through a short plug of silica gel. The
ee for epoxides derived from styrene and dihydronaphthalene was
1
was washed with water, methanol, CH
dark brown beads. Elemental analysis indicated a nitrogen content of
.27% for 7 and a Mn content of 1.48% for 8, corresponding to loadings
2
Cl
2
, ether, and hexanes to give
determined by H NMR in the presence of the chiral shift reagent Eu-
(hfc)
determined by GC using a commercial chiral column (Chiraldex G-TA,
100 °C isothermal, t ) 3.3min and 4.7maj min).
3
. The ee for the cis-epoxide derived from cis-â-methylstyrene was
0
of 0.10 and 0.27 mmol/g, respectively.
R
Preparation of Salen Methyl Glutarate 9. 1-(3-Dimethylamino-
propyl)-3-ethylcarbodiimide hydrochloride (EDC) (23 mg, 0.12 mmol)
was added to a solution containing methanol (3 µL, 0.066 mmol), salen
General Procedure for Time-Course Study. The same procedure
as outlined above was followed with the exception that m-CPBA (2
equiv) was added in two equal portions at time ) 0 and 70 s to the
reaction which also contained dodecane (1 equiv) as an internal
standard. Aliquots (25 µL) taken at time intervals of 0, 15, 30, 60,
120, and 300 s were quenched into a solution containing 200 µL of 1
N NaOH and 200 µL of ether. After thorough shaking, 50 µL of the
organic layer was added to 200 µL of an ether solution containing a
4
(36 mg, 0.058 mmol), and DMAP (8 mg, 0.066 mmol) in CH
mL). The mixture was stirred for 16 h at room temperature and then
poured into water (5 mL). The organic layer was dried over MgSO
2 2
Cl (3
4
and concentrated under reduced pressure. The crude residue was purified
by silica gel chromatography (EtOAc/hexane ) 1/4) to provide the
1
expected methyl ester (26 mg, 72%) as a yellow oil; H NMR (500
4
small amount of MgSO . A 50 µL sample of this mixture was then
MHz, CDCl
3
) δ 1.24 (s, 9H), 1.38 (s, 9H), 1.40 (s, 9H), 1.73-1.95
diluted with 100 µL of ether and analyzed by GC for conversion to
epoxide products and ee (cis-epoxide only).
(
m, 8H), 2.02-2.07 (m, 2H), 2.44 (t, 2H, J ) 7.3 Hz), 2.58 (t, 2H, J
)
7.3 Hz), 3.33 (m, 2H), 3.69 (s, 3H), 6.76 (d, 1H, J ) 2.6 Hz), 6.92
Acknowledgment. This work was supported financially by
the National Institutes of Health (GM-56154), The Scripps
Research Institute, and The Skaggs Institute for Chemical
Biology. We thank Dr. Patrick H. Toy for helpful discussions
and assistance in the preparation of some of the resins.
JA000692R
(
d, 1H, J ) 2.6 Hz), 6.98 (d, 1H, J ) 2.0 Hz), 7.31 (d, 1H, J ) 2.0
13
Hz), 8.23 (s, 1H), and 8.30 (s, 1H); C NMR (100 MHz, CDCl
3
) δ
2
5
1
0.0, 24.2, 29.1, 29.4, 31.4, 32.9, 33.0, 33.1, 33.2, 34.0, 34.8, 34.9,
1.7, 72.2, 72.4, 117.7, 118.2, 121.3, 122.8, 126.0, 127.0, 136.4, 138.6,
40.0, 141.4, 157.9, 158.1, 164.7, 165.9, 171.9, and 173.3; HRMS calcd
+
54
for [C35H N
O
2 6
+ H ] 635.4055, found 635.4060.