Chiral fluoro ketones for catalytic asymmetric epoxidation of alkenes with oxone

Article abstract of DOI:10.1021/jo020050h

Two structurally dissimilar, chiral fluoro ketones have been prepared and their potential as enantioselective catalysts for asymmetric epoxidation with Oxone has been evaluated. The tropinone-based ketone (-)-5 was easily prepared and showed excellent reactivity but only modest enantioselectivity. The biphenyl-based ketone (-)-6 was prepared in a somewhat lengthy synthesis (along with its monofluoro and geminal fluoro analogues). This ketone exhibited only modest reactivity; 30 mol % of (-)-6 was needed to bring about complete conversion in a reasonable time. The enantioselectivity of this catalyst was generally much higher, but again very substrate dependent.

Full text of DOI:10.1021/jo020050h

J . Org. Chem. 2002, 67, 3479-3486
3479
Ch ir a l F lu or o Keton es for Ca ta lytic Asym m etr ic Ep oxid a tion of
Alk en es w ith Oxon e
Scott E. Denmark* and Hayao Matsuhashi
Roger Adams Laboratory, Department of Chemistry, University of Illinois, Urbana, Illinois 61801
denmark@scs.uiuc.edu
Received J anuary 23, 2002
Two structurally dissimilar, chiral fluoro ketones have been prepared and their potential as
enantioselective catalysts for asymmetric epoxidation with Oxone has been evaluated. The tropinone-
based ketone (-)-5 was easily prepared and showed excellent reactivity but only modest
enantioselectivity. The biphenyl-based ketone (-)-6 was prepared in a somewhat lengthy synthesis
(along with its monofluoro and geminal fluoro analogues). This ketone exhibited only modest
reactivity; 30 mol % of (-)-6 was needed to bring about complete conversion in a reasonable time.
The enantioselectivity of this catalyst was generally much higher, but again very substrate
dependent.
In tr od u ction a n d Ba ck gr ou n d
classes of alkenes still remained a challenging problem.
Moreover, the interest in non-metal-based “organocataly-
sis” has achieved a heightened level of importance in
recent years.3 In this context, chiral dioxirane-mediated
epoxidation represents an attractive alternative method.4
The chemistry of dimethyldioxirane 1 has recently re-
ceived much attention in both mechanistic and synthetic
applications. Dioxiranes can be generated from ketones
and Oxone (peroxymonosulfate) in situ, or in some
instances they can be isolated5 (Scheme 1). Because
dioxiranes are powerful oxidants, they are quite useful
in organic synthesis.5
The asymmetric epoxidation mediated by a chiral
dioxirane is conceptually simple as shown in Scheme 2.
A chiral ketone can be transformed to the corresponding
dioxirane by the action of persulfate, and the chiral
dioxirane may react selectively on one enantiotopic face
of the substrate alkene. After oxygen atom transfer from
dioxirane to alkene, the chiral ketone is regenerated, thus
rendering the reaction system catalytic.
In 1984, Curci et al. reported first example of a chiral
dioxirane mediated epoxidation. These workers employed
the readily available ketone 2 (Chart 1) as the chiral
ketone and observed a 12% ee for the reaction with
1-methyl-1-cyclohexene.6 Significant progress has been
made in the design and application of chiral ketones for
asymmetric epoxidation in the intervening 17 years. The
first critical advance was reported by Yang et al., who
described a highly enantioselective epoxidation protocol
using C2-symmetric ketone 3.7 Subsequent reports de-
The epoxidation of alkenes is the most fundamental
oxygen functionalization of carbon-carbon double bonds.
The importance of this transformation is a direct conse-
quence of the utility of the product epoxides as synthetic
intermediates. Accordingly, the ability to carry out enan-
tioselective oxygen atom transfer to alkenes has served
as the vanguard for advances in asymmetric catalysis
over the past two decades. The first practical asymmetric
epoxidation reported by Sharpless et al. for allylic alco-
hols was a watershed event in the exploration for more
selective and general methods of epoxidation.1 In this
reaction, the allylic hydroxyl group plays a fundamental
role, in coordination of the substrate to the chiral
titanium complex. As a result, a favorable asymmetric
environment can be crafted between reactant and sub-
strate by virtue of the strong coordination of the alcohol
to the titanium metal center.
The next important advance in enantioselective oxygen
atom transfer was the asymmetric epoxidation of un-
functionalized alkenes. In this case, the enantioselectivity
must be controlled by nonbonding, which are much
weaker and less directional than coordination bonds
between reactant and substrate. The J acobsen epoxida-
tion system employing the now-familiar manganese-
salen complexes is the archetypal example of this reaction
and one that has been extensively developed and applied
over the past 10 years.2
Despite the enormous success with Mn-salen complex
mediated asymmetric epoxidation, the reaction of certain
(1) (a) Katsuki, T.; Sharpless, K. B. J . Am. Chem. Soc. 1980, 102,
5974. (b) Katsuki, T.; Martin, V. S. Org. React. 1996, 48, 1. (c) J ohnson,
R. A.; Sharpless, K. B. Comprehensive Organic Synthesis, Vol. 7;
Oxidation; Ley, S. V., Ed.; Pergamon Press: Oxford, 1991; Chapter
3.2. (d) Katsuki, T. In Comprehensive Asymmetric Catalysis, Vol. II;
J acobsen, E. N., Pfaltz, A., Yamamoto, H., Eds.; Springer-Verlag:
Heidelberg, 1999; Chapter 18.1.
(2) (a) Zhang, W.; Loebach, J . L.; Wilson, S. R.; J acobsen, E. N. J .
Am. Chem. Soc. 1990, 112, 2801. (b) Zhang, W.; J acobsen, E. N. J .
Org. Chem. 1991, 56, 2296. (c) J acobsen, E. N.; Wu, M. F. In
Comprehensive Asymmetric Catalysis; J acobsen, E. N., Pfaltz, A.,
Yamamoto, H., Eds.; Springer-Verlag: Heidelberg, 1999; Vol. II,
Chapter 18.2. See also: (d) Hosoya, N.; Hatayama, A.; Irie, R.; Sasaki,
H.; Katsuki, T. Tetrahedron 1994, 50, 4311.
(3) For a recent review, see: Dalko, P. Angew. Chem., Int. Ed. 2001,
40, 3726.
(4) (a) Adam, W.; Curci, R.; Edwards, J . O. Acc. Chem. Res. 1989,
22, 205. (b) Curci, R.; Dinoi, A.; Rubino, M. F. Pure Appl. Chem. 1995,
67, 811. (c) Murray, R. W. Chem. Rev. 1989, 89, 1187.
(5) (a) Murray, R. W.; Singh, M. Org. Synth. 1996, 74, 91. (b)
Murray, R. W.; J eyaraman, R. J . Org. Chem. 1985, 50, 2847.
(6) (a) Curci, R.; Fiorentino, M.; Serio, M. R. J . Chem. Soc., Chem.
Commun. 1984, 155. (b) Curci, R.; D’Accolti, L.; Fiorentino, M.; Rosa,
A. Tetrahedron Lett. 1995, 36, 5831.
(7) (a) Yang, D.; Yip, Y.-C.; Tang, M.-W.; Wong, M.-K.; Zheng, J .-
H.; Cheung, K.-K. J . Am. Chem. Soc. 1996, 118, 491. (b) Yang, D.;
Wang, X.-C.; Wong, M.-K.; Yip, Y.-C.; Tang, M.-W. J . Am. Chem. Soc.
1996, 118, 11311.
10.1021/jo020050h CCC: $22.00 © 2002 American Chemical Society
Published on Web 04/17/2002

3480 J . Org. Chem., Vol. 67, No. 10, 2002
Denmark and Matsuhashi
Ch a r t 2
Sch em e 1
Sch em e 2
On the basis of these observations, the 3-fluorotro-
panone ammonium salt 5 (Chart 2) was designed as the
first target.12 To enhance reactivity, both the R-fluorine
substituent and ammonium moiety should increase the
electrophilicity of the carbonyl group to facilitate the
formation of the dioxirane intermediate as well as oxygen
atom transfer to alkene. Of course, another important
issue is the chiral environment. In the case of 5, it is
necessary to control the facial selectivity of dioxirane
intermediate due to the diastereotopic relationship of the
two oxygen atoms in the dioxirane. We assumed that the
top-face approach of an alkene to dioxiranes prepared
from 5 would be shielded (and thus disfavored) by the
methyl group of the ammonium bridge. Therefore, the
alkene should approach more favorably from the bottom
face. A different class of chiral ketones was inspired by
our studies of epoxidation with fluorocyclohexanones.11
To maintain an efficient asymmetric environment, we
choose to incorporate the axial chirality of a biaryl by
analogy to cyclic ketone 3. However, 3 contains an 11-
membered ketone, and the distance of the binaphthyl
moiety seems to be far from the carbonyl group (and the
corresponding dioxirane). Indeed, ketone 3 works well
only for stilbene-type trans-diarylalkenes.7 Accordingly,
the seven-membered cyclic difluoro ketone 6 emerged as
our second target. The expected dioxirane intermediate
corresponding to 6 is also C2 symmetric, so the oxygens
of the dioxiranes are now homotopic. Although these
candidates did not satisfy criterion 4 well, we were willing
to make the synthetic investment to learn more about
the success of the other elements of design. If the
structural requirements for high catalytic activity, se-
lectivity, and generality could be learned, then it was
expected that a more synthetically accessible target could
be envisioned.
Ch a r t 1
scribing the use of derivatives of 3 also showed promise,
but mostly with a limited set of aromatic substrates. The
second advance, due to Shi, was the successful catalytic
epoxidation using fructose-derived ketone 4 as the cata-
lyst.8 Ketone 4 has a pseudo-C2-symmetric structure, and
it was found that a wide variety of (E)-alkenes can be
oxidized with high enantioselectivity with substoichio-
metric amounts of 4 and Oxone in the presence of potas-
sium carbonate.8b A detailed account of the evolution and
the state of the art in catalytic enantioselective epoxi-
dations with dioxiranes can be found in recent reviews.2c,9
Ca ta lyst Design . The most important criteria for
chiral, dioxirane-mediated catalytic asymmetric epoxi-
dation are as follows: (1) high catalytic activity of the
ketone, (2) high level of stereodifferentiation, (3) broad
generality of substrate, and (4) ease of synthesis of the
ketone. How to structurally engineer these characteristics
is a not-insignificant challenge. At the outset, we recog-
nized that the reactivity of the ketone was the most
important and critical issue. Indeed, dioxirane-mediated
epoxidation had been carried out with stoichiometric or
excess amounts of ketone in many cases, and there are
only a few example of catalytic epoxidation. In 1995, we
published the first examples of catalytic, dioxirane-
mediated epoxidation, which were accomplished with a
1-oxo-4-piperidinium triflate.10 The success of this ketone
was based on the ability of nearby electron-withdrawing
substituents to increase the reactivity of a carbonyl group
toward nucleophiles. In this case, the ammonium moiety
plays a key role as an electron withdrawing function and
as a phase-transfer mediator. In addition, we have also
reported that R-fluorine substituents tremendously in-
crease the reactivity of the carbonyl group in dioxirane-
mediated epoxidation and that these effects are strongly
stereoelectronically controlled. Thus, cis-4-tert-butyl-2-
fluorocyclohexanone shows good catalytic activity for the
epoxidation whereas the trans isomer does not.11
Resu lts a n d Discu ssion
F lu or otr op a n on e Am m on iu m Sa lt 5. P r ep a r a tion .
The preparation of 5 has been carried out following
Davies’ 2-tropanone synthesis as shown in Scheme 3.13
Thus, 2-ethoxycarbonylpyrrole 7 was converted to diester
9 by formylation and Horner-Emmons reaction in 63%
and 92% yield, respectively. The aromatic diester 9 was
then transformed to the saturated diester 10 by stepwise
hydrogenation of the double bond with Raney Ni catalyst
followed by reduction of the aromatic ring with Rh-Al2O3
in high yield. The N-methylpyrrolidine (obtained by
reductive methylation of 10) was cyclized under basic
conditions to give bicyclic â-ketoester 11 in 79% yield.
Finally, acidic saponification and decarboxylation of 11
afforded 2-tropanone 12 in 93% yield.
(8) (a) Tu, Y.; Wang, Z.-X.; Shi, Y. J . Am. Chem. Soc. 1996, 118,
9806. (b) Wang, Z.-X.; Tu, Y.; Frohn, M.; Shi, Y. J . Org. Chem. 1997,
62, 2328. (c) Wang, Z.-X.; Tu, Y.; Frohn, M.; Zhang, J .-R.; Shi, Y. J .
Am. Chem. Soc. 1997, 119, 11224. (d) Wang, Z.-X.; Shi, Y. J . Org. Chem.
1998, 63, 3099.
(11) Denmark, S. E.; Wu, Z.; Crudden, C. M.; Matsuhashi H. J . Org.
Chem. 1997, 62, 8288.
(9) (a) Denmark, S. E.; Wu, Z. Synlett 1999, 847. (b) Frohn, M.; Shi,
Y. Synthesis 2000, 1979.
(10) Denmark, S. E.; Forbes, D. C.; Hay, D. S.; DePue, J . S.; Wilde,
R. G. J . Org. Chem. 1995, 60, 1391.
(12) For a structurally similar fluorotropanone catalyst, see: Arm-
strong, A.; Hayter, B. Chem. Commun. 1998, 621.
(13) Davies, W. A. M.; Pinder, A. R.; Morris, I. G. Tetrahedron 1962,
18, 405.

Catalytic Asymmetric Epoxidation of Alkenes with Oxone
J . Org. Chem., Vol. 67, No. 10, 2002 3481
Sch em e 3
Sch em e 5
Sch em e 6
Sch em e 4
Sch em e 7
The introduction of a fluorine atom was first attempted
with electrophilic fluorination of 12 via its TMS enol
ether using the Selectfluor reagent,14 Scheme 4. The TMS
enol ether was prepared quantitatively by the usual
procedure, however, no fluorinated product was obtained
after treatment with Selectfluor. In addition, the starting
material seemed to decompose under the reaction condi-
tions. This result may arise from the low reactivity of
the TMS enol ether and competitive oxidation of nitrogen
atom.
salt with Ba(OH)2) were dissolved in boiling water, and
then the salt was recrystallized three times to provide
(+)-13, which was >98% ee by CSP-GC analysis. The
absolute configuration of the tropanone was determined
by single-crystal X-ray analysis of the salt, thus allowing
correlation to the configuration of (-)-15.15a After the salt
was neutralized with K2CO3, (+)-13 was obtained in 28%
yield.
Treatment of 2-tropanones 12 and (+)-13 with methyl
triflate in acetonitrile provided the target ammonium
salts (()-16 and (-)-5 in 62% and 73% yields, respectively
(Scheme 7).
Ep oxid a tion . With the tropanone ammonium triflates
(-)-5 and (()-16 in hand, their potential as catalysts for
the epoxidation was evaluated. For the initial survey of
conditions, we selected a convenient monophasic protocol
for alkene epoxidation that was described by Yang et al.16
The outlined general procedure involves the following
steps: (1) the substrate and ammonium salt are dissolved
A simple solution was found by introducing the fluorine
substituent on the more nucleophilic â-ketoester anion.13
The previous substrate 11 was deprotonated to generate
the stabilized anion, which was then treated with the
Selectfluor to give R-fluoro-â-ketoester 14 as a mixture
of diastereomers in low yield. After a survey of reaction
conditions, we found that the reaction of the sodium
enolate, generated from 11 with sodium hydride, and
Selectfluor took place in DMF to provide 14 in 26-39%
yield after distillation, Scheme 5. Saponification and
decarboxylation of 14 took place readily under acidic
conditions to afford the target 3-fluoro-2-tropanone in
high yield. The equatorial orientation of the fluorine atom
1
was easily established by the H NMR coupling pattern
and was later confirmed by X-ray crystallography.
Enantiomer resolution was performed by conversion
of racemic 13 to its diastereomeric salt with (-)-bro-
mocamphorsulfonic acid (15), as was reported for the
resolution of 2-tropanone, Scheme 6. Equimolar amounts
of racemate 13 and (-)-15 (generated from its ammonium
(15) (a) The crystallographic coordinates of 13‚15 salt have been
deposited with the Cambridge Crystallographic Data Centre, deposition
no. 176803. (b) The crystallographic coordinates of (-)-6 salt have been
deposited with the Cambridge Crystallographic Data Centre, deposition
no. 176802.
(14) Lal, G. S. J . Org. Chem. 1993, 58, 2791.
(16) Yang, D.; Wong, M.-K.; Yip, Y.-C. J . Org. Chem. 1995, 60, 3887.

3482 J . Org. Chem., Vol. 67, No. 10, 2002
Denmark and Matsuhashi
Ta ble 1. Ep oxid a tion w ith F lu or otr op in iu m Sa lt (-)-5
F igu r e 1. Hypothetical transition structures for approach of
alkenes to dioxirane from (-)-5.
the two most selective cases ((E)-stilbene and â-methyl-
styrene) was established by comparison of the optical
rotation to literature values.17
These results showed the significant catalytic potential
for fluoro ammonium ketones such as 5, but clearly the
enantiotopic discrimination was not sufficient. To for-
mulate a transition-structure model requires a detailed
picture of the preferred approach of alkenes to dioxiranes
in the oxygen atom transfer step. At present, there is no
strong experimental evidence for either of the limiting
transition structures proposed (spiro or butterfly), al-
though most authors tend to rationalize the stereochem-
ical course of epoxidation with chiral catalysts in terms
of the spiro model.7,8,18 Moreover, computational analyses
have also led to the prediction of a preferred spiro
transition structure.19 Indeed, a recent computational
study, specifically addressing the preferred approach of
alkenes to fluorocyclohexanones, concluded that for the
equatorial fluoro ketone, the preferred approach is to-
ward the equatorially oriented oxygen of the dioxirane.19b
Thus, incorporating these features into a model for
reaction with (-)-5 produces the hypothetical arrange-
ments for the dioxirane intermediate shown in Figure 1.
On the basis of both steric shielding by the N-methyl
groups and the stereoelectronic factors noted above, we
conclude that approaches c and d are more favorable
than approaches a and b. From approach c, the (S,S)-
epoxide is formed, whereas the (R,R) epoxide is produced
from approach d . Inspection of molecular models reveals
that the most distinguishing feature that differentiates
these arrangements is the interaction of the alkene
substituents with the two-carbon bridge and fluorine
atom in c, which are absent in d . Although this analysis
does rationalize the major enantiomer formed, the dif-
ference is not very dramatic, and it is thus not surprising
that the enantioselectivities were only modest and highly
substrate dependent.
a
Oxone (2.5 equiv) and NaHCO3 (7.5 equiv) were added.
Oxone (2 equiv) and NaHCO3 (6 equiv) were added. c Yield based
on reacted starting material. (+)-5 was employed as catalyst.
b
d
in acetonitrile/aqueous Na2EDTA (0.09%) (3/2) mixed
solvent system; (2) Oxone (1 equiv) and NaHCO3 (2.6
equiv) are added portionwise every 2 h. Under these
conditions using 10 mol % of the parent salt 16, only 21%
conversion of â-methylstyrene to the epoxide was ob-
served after 5 h.
Fortunately, R-fluorinated ketone (-)-5 showed excel-
lent catalytic activity. The epoxidation of various kinds
of alkenes with 10 mol % catalyst are summarized in
Table 1. The reactions were complete within 12 h in most
cases, and there were no undesired side reactions, e.g.,
epoxide ring opening. Reactive alkenes such as â-meth-
ylstyrene, 1-phenyl-1-cyclohexene, and indene readily
oxidized under the reaction conditions to give the corre-
sponding epoxides in high yields (Table 1, entries 2, 4,
and 5). The reaction took place even with relatively
unreactive alkenes with high conversion after 9-30 h
(Table 1, entries 1 and 6). In the case of (E)-stilbene,
however, the reaction was not complete even after 72 h,
and a significant amount of starting material was
recovered (Table 1, entry 3). This may be due to the
limited solubility of the substrate in the reaction medium.
Indeed, before addition of Oxone and NaHCO3, the
substrate did not visibly dissolve at all.
Although (-)-5 showed good general reactivity, its
enantioselectivity depended upon the substrate structure.
For E-alkenes, at least one aromatic substituent was
necessary for modest selectivity; (E)-stilbene provided the
highest ee (58%), and â-methylstyrene gave the epoxide
product in 35% ee. On the other hand, this catalyst could
not effectively differentiate the faces of a simple alkene
such as 1-benzyloxy-4-hexene. With other alkenes, the
product selectivity varied with each substrate with no
obvious trend. The absolute stereochemical pathway for
(17) (a) Witkop, B.; Foltz, C. M. J . Am. Chem. Soc. 1957, 79, 197.
(b) Chang, H.-T.; Sharpless, K. B. J . Org. Chem. 1996, 61, 6456.
(18) Baumstark, A. L.; McCloskey, C. J . Tetrahedron Lett. 1987, 28,
3311.
(19) (a) Bach, R. D.; Andres, J . L.; Owensby, A. L.; Schlegel, H. B.;
McDouall, J . J . W. J . Am. Chem. Soc. 1992, 114, 7207. (b) Armstrong,
A.; Washington, I.; Houk, K. N. J . Am. Chem. Soc. 2000, 122, 6297.

Catalytic Asymmetric Epoxidation of Alkenes with Oxone
J . Org. Chem., Vol. 67, No. 10, 2002 3483
Sch em e 8
Sch em e 10
Sch em e 9
Diflu or od iben zocycloh ep ta n on e ((-)-6). P r ep a r a -
tion . The cyclic difluoro ketone (-)-6 was prepared
following the literature procedures.20 The synthesis began
with a copper(I)-mediated oxidative homocoupling of the
diazonium salt generated from commercially available
2-amino-3-methylbenzoic acid 18 to afford biphenic di-
carboxylic acid 19,20a Scheme 8.
Sch em e 11
The optical resolution of 19 has been described using
morphine.20b In view of the poor availability of this
reagent, we examined other chiral bases for this resolu-
tion. After a brief survey, we found that quinine also
worked well for this purpose. Thus, equimolar amounts
of (()-19 and quinine were dissolved in 90% aqueous
EtOH, and the resulting salt was recrystallized twice to
give diastereomerically pure salt. After decomposition of
Sch em e 12
the salt under acidic conditions, (+)-19 ([R]25 +19.2 (c
D
) 1.00, MeOH)) was obtained in 70% yield, Scheme 9.
The enantiomeric purity of 19 was determined by HPLC
analysis of the corresponding dimethyl ester, prepared
by treatment of 19 with excess diazomethane. The
absolute configuration of (+)-19 was determined to be S
by comparison with reported rotation value (lit.20c (R)-
19 [R]D -20 (dioxane)).
With chiral acid (+)-19 in hand, the remaining steps
in the synthesis of ketone could be undertaken as
outlined in Scheme 10. Dicarboxylic acid 19 was trans-
formed to dibromide 21 by the two-step sequence of
reduction with lithium aluminum hydride and bromida-
tion of the corresponding alcohol in 87% and 86% yields,
respectively.20d Treatment of 21 with potassium cyanide
in aqueous EtOH under reflux gave the dinitrile.20e 1H
NMR analysis of the crude product showed that it
contained small amount (ca. 5-10%) of the cyclized
product 23. Therefore, the crude material was treated
with base directly to afford cyclic imine 23 in 75% yield
over two steps. Finally, hydrolysis of the cyano group and
imine, followed by decarboxylation under acidic condi-
tions, provided target ketone 24 in 92% yield.20c
to give monofluoro ketone 25 as a mixture of diastereo-
mers in a 1/1 to 2/1 ratio. Treatment of the mixture of
diastereomers with triethylamine at room temperature
resulted in the epimerization of the fluorine-bearing
center to produce a single diastereomer 25, where the
fluorine atom is in an equatorial position. In this way,
the desired compound 25 was obtained as a single
diastereomer in 65% yield over three steps.
Introduction of the second fluorine atom by the same
protocol was, expectedly, plagued by formation of regio-
isomeric silyl enol ethers from 25. To facilitate separation
of the isomers, the triethylsilyl (TES) enol ether was
chosen. Monofluoro ketone 25 was transformed to its TES
enol ether by deprotonation with LDA in the presence of
TESCl at low temperature. Separation of the mixture of
enol ethers by silica gel chromatography afforded the
desired isomer 26a (54%) and more substituted isomer
26b (14%), Scheme 12.
The fluorination of 26a with Selectfluor took place
under the usual conditions to afford target difluoroketone
(-)-6 in moderate yield, Scheme 13. The configuration
of (-)-6 was determined by single-crystal X-ray analysis.15b
gem-Difluoro ketone 27 was also prepared from the
regioisomer 26b to test the effect of geminal fluorine
substitution in the epoxidation reaction.
Op tim iza tion of Ep oxid a tion . Orienting experi-
ments with difluoro ketone (-)-6 under monophasic
reaction conditions were initially disappointing. For
example, the epoxidation of â-methylstyrene with 18 mol
Fluorination of ketone 24 followed the previous se-
quence, Scheme 11. Thus, ketone 24 was converted to
its TMS enol ether under the usual conditions, and the
crude enol silane was treated with Selectfluor in DMF
(20) (a) Atkinson, E. R.; Lawler, H. J . In Organic Syntheses; Wiley:
New York, 1941; Collect. Vol. I, p 222. (b) Bell, F. J . Chem. Soc. 1934,
835. (c) Mislow, K.; Glass, M. A. W.; O’Brien, R. E.; Rutkin, P.;
Steinberg, D. H.; Weiss, J .; Djerassi, C. J . Am. Chem. Soc. 1962, 84,
1455. (d) Bergmann, E. D.; Pelchowicz, Z. J . Am. Chem. Soc. 1953, 75,
2663. (e) Newman, P.; Rutkin, P.; Mislow, K. J . Am. Chem. Soc. 1958,
80, 465.

3484 J . Org. Chem., Vol. 67, No. 10, 2002
Denmark and Matsuhashi
F igu r e 2. Epoxidation rates at (E)-2-methylstyrene with ketones.
Sch em e 13
asymmetric epoxidation using a stoichiometric amount
of chiral ketone. These initial studies employed â-meth-
ylstyrene as a substrate, and the results are graphically
depicted in Figure 2. Difluoro ketone (-)-6 promoted the
reaction much more efficiently than the corresponding
monofluoro ketone 25 and parent ketone 24. The reaction
promoted by (-)-6 was complete within 8 h, whereas the
reaction promoted with 25 was quite slow and only 33%
conversion was observed after 11 h. In the case of 24, it
did not promote the reaction at all, as the reaction
conversion was almost the same as the background
reaction (8 h, 6% conversion). Interestingly, gem-difluoro
ketone 27 was the most effective promoter of the reaction
affording complete consumption of the alkene in 6 h. This
intriguing result contrasts our previous observation of
lower epoxidation rates with a gem-difluorocyclohexanone
skeleton.11
Sch em e 14
Although the reactivity of (-)-6 was lower than ex-
pected, it should be noted that R-fluorine substitution still
strongly activated the carbonyl group. In previous stud-
ies, we have employed the rate and degree of hemiketal
formation as a measure of the propensity to form the
intermediate dioxirane. Whereas this does not necessarily
correlate with epoxidation efficiency, it does reflect the
electrophilicity of the ketone.11,22 The graph in Figure 3
illustrates significant enhancement of hemiketal forma-
tion for 6 compared to 25.
Sch em e 15
Ca ta lytic Ep oxid a tion . Encouraged by the high
asymmetric induction observed with (-)-6, we next
investigated the optimal reaction conditions for catalytic
epoxidation. The standard system chosen for evaluation
was â-methylstyrene together with 4 equiv of Oxone and
10 mol % of (-)-6. The first variable tested was the
reaction medium.7 Of the many aqueous solvent blends
tested (propionitrile, DME, DMF, DMAC, dioxane, DMPU)
only CH3CN/H2O (3/2) gave appreciable conversion (28%)
to the epoxide. Increasing the proportion of acetonitrile
to 5/1 decreased the conversion to 10%. The reason for
this solvent effect is not apparent. One possibility is that
the cyano group of CH3CN may contribute to the reaction
in some way.23 In the case of propionitrile, the reaction
did not take place because of phase separation.
% of (+)-6 proceeded quite slowly, and only 28% of the
alkene was converted to the epoxide after 8 h, Scheme
14. This result suggested that two R-fluorine substituents
may not provide sufficient activation for the carbonyl
group of 6. Alternatively, the low reactivity may be
partially caused by the conformation of the cyclic ketone
moiety. It is known that the reaction of cyclohexanone is
more favorable than of cycloheptanone with nucleo-
philes.21 Indeed, there was a significant difference in rate
of epoxidation between cyclohexanone and cyclohep-
tanone when the reactions were tested using 1-benzyloxy-
4-hexene as a substrate with 1 equiv of the ketone,
Scheme 15.
The next variable we tested that is known to influence
the reaction rate was pH. Shi had already pointed out
(22) (a) Stewart, R.; Van Dyke, J . D. Can. J . Chem. 1970, 48, 3962.
(b) Yang has reported that for a series of R,R′-diacyloxy ketones, there
is no correlation between hydration equilibrium and oxidation rate:
Yang, D.; Yip, Y.-C.; Tang, M.-W.; Wong, M.-K.; Cheung, K.-K. J . Org.
Chem. 1998, 63, 9888. Although we have clearly stated that there needs
to be a balance of effects to produce a good catalyst, we are skeptical
of conclusions based on time to complete reaction under heterogeneous
conditions.
Despite the low reactivity of 6, it still exhibited good
asymmetric induction as shown in Scheme 14. Accord-
ingly, it was deemed profitable to survey the various
fluoro ketones for their rate and enantioselectivity in the
(21) Forbes, D. C. Ph.D. Thesis, University of Illinois, Urbana-
Champaign, 1996.
(23) Bose, D. S.; Baquer, S. M. Synth. Commun. 1997, 27, 3119.

Catalytic Asymmetric Epoxidation of Alkenes with Oxone
J . Org. Chem., Vol. 67, No. 10, 2002 3485
F igu r e 3. Hemiketalization of ketones 6 and 25.
Ta ble 2. Asym m etr ic Ep oxid a tion s of Va r iou s Alk en es
Sch em e 16
w ith (-)-6
Sch em e 17
that simply changing the buffer solution to potassium
carbonate increases the reaction rate.8d In our system,
the reaction conversion was 55% after 8 h with 10 mol %
of 6. We also compared the reactivity of gem-difluoro
ketone 27 under the same conditions and were pleased
to find that again 27 showed slightly higher reactivity
than 6. This result reflects the behavior observed with
stoichiometric amounts of catalysts. To consume all of
the test alkene required the use of 30 mol % of (-)-6;
under these conditions, the reaction was complete after
11 h, Scheme 16. The improved reactivity at higher pH
can be understood in terms of the increased rate of
formation of the dioxirane. As mentioned above, the low
reactivity of 6 may have its origin in the slow formation
of the tetrahedral intermediate from nucleophilic attack
of peroxymonosulfate to the carbonyl group. At higher
pH (∼10) the concentration of the peroxymonosulfate
dianion (KOOS(O)2OK) is higher and the concentration
of the conjugate base of the adduct will also be higher,
thus facilitating the formation of the dioxirane, Scheme
17.
a
b
The reaction was carried out in CH3CN/H2O (3/1). Starting
material (53%) was recovered. c The reaction was carried out with
50 mol % of 6. Starting material (23%) was recovered.
d
established by comparison of the optical rotations to
literature values.17,24
Analysis of the stereochemical outcome of epoxidation
with (-)-6 (via the dioxirane intermediate) is simplified
because of the overall C2 symmetry of the molecule, which
renders the two oxygens homotopic. Proceeding from the
reasonable assumption that the reaction occurs via a
spiro transition state, two limiting arrangements of the
alkene with respect to the dioxirane can be formulated,
Figure 4. These two approaches, a and b, differ only in
the proximity of the alkene substituents to the biaryl
skeleton and the methyl groups in particular. Clearly,
approach b is disfavored compared to approach a because
of this steric interaction. It is not clear if, in the case of
aromatic substituted alkenes, there is an additional
electronic interaction between substrate and catalyst,
such as π-π stacking of the aromatic rings, enforced by
a hydrophobic effect of the aqueous medium. From this
analysis, the absolute configuration of the major enan-
tiomer formed from (-)-6 can be correctly predicted. This
Having established suitable conditions for catalytic
epoxidation, we surveyed the generality of the asym-
metric epoxidation with various alkenes using 30 mol %
of (-)-6. The results are summarized in Table 2. The
reactions were carried out at 0 °C with portionwise
addition of Oxone (1 equiv every 2 h) and K2CO3 (3 equiv
every 2 h) into CH3CN/H2O (3/2) mixed solvent. The
reaction with trans alkenes proceeded in excellent ee
(Table 2, entries 1-4), albeit with lesser selectivity for
the aliphatic (E)-alkene. No undesired byproducts were
observed due, in part, to the mild reaction conditions.
Moderate selectivities were observed in case of trisub-
stituted and monosubstituted alkenes (Table 2, entries
5 and 6). The absolute configuration of the products was
(24) (a) Brandes, B. D.; J acobsen, E. N. J . Org. Chem. 1994, 59, 4378.
(b) Pedragosa-Moreau, S.; Morisseau, C.; Zylber, J .; Archelas, A.;
Baratti, J .; Furstoss, R. J . Org. Chem. 1996, 61, 7402.

3486 J . Org. Chem., Vol. 67, No. 10, 2002
Denmark and Matsuhashi
Con clu sion
The two fluoro ketones prepared and examined herein
provided useful lessons that can be applied to the design
of newer more synthetically useful catalysts. The solubil-
ity and strong electronic activation provided by the
ammonium group afforded excellent reactivity charac-
teristics to (-)-5. For this class of catalyst to be useful,
there needs to be a greater differentiation of the two
competing equatorial approaches to the dioxirane. In
biphenyl-derived ketone (-)-6, the need to for at least
two fluorine atoms was clearly established to render it
sufficiently electrophilic to promote oxygen atom transfer.
Even so, the reaction rates were modest with as much
as 30 mol % of ketone. The enantioselectivity was quite
high in some cases, but structurally dependent in such
a way as to suggest the need for a more sterically defined
environment. Finally, the need to experimentally estab-
lish the preferred alignment of alkene and dioxirane is
apparent to aid the design of more selective catalyst
structures.
F igu r e 4. Hypothetical transition structures for approach of
alkenes to dioxirane from (-)-6.
model also can explain the less than stellar selectivities
observed with trisubstituted and monosubstituted al-
kenes.
If indeed the origin of enantioselectivity is as portrayed
in Figure 4, then one possible improvement would be to
increase the size of the substituents responsible for the
stereodifferentiating effect, e.g., resorting to a structur-
ally analgous binaphthyl ketone. Of course, the problems
associated with catalyst synthesis and solubility are not
addressed by this modification. Moreover, it is likely that
the rather simplistic picture of the spiro transition state
is much more maleable than is currently viewed. Small
angular deviations from the spiro arrangement may not
be very energetically costly and can thus provide low
energy access to diastereomeric transition structures in
sterically nondemanding substrates.
Ack n ow led gm en t. We are grateful to the Pharma-
cia Corp. for generous support of this research.
Su p p or tin g In for m a tion Ava ila ble: Complete experi-
mental details with full spectroscopic and analytical data for
all new compounds and general procedures for epoxidation is
provided. This material is available free of charge via the
Internet at http://pubs.acs.org.
J O020050H