Studies of the stereochemistry of [2+2]-photocycloaddition reactions of 2-cyclohexenones with olefins

Article abstract of DOI:10.1021/ol063092r

(Chemical Equation Presented) Stereochemical studies of the photocycloaddition of a chiral α,β-enone and isobutylene lead to a working hypothesis for the geometry of the ³(π-π*) excited state of the enone.

Full text of DOI:10.1021/ol063092r

ORGANIC
LETTERS
2007
Vol. 9, No. 6
1057-1059
Studies of the Stereochemistry of
[2 2]-Photocycloaddition Reactions of
2-Cyclohexenones with Olefins
+
Ruichao Shen and E. J. Corey*
Department of Chemistry and Chemical Biology, HarVard UniVersity,
12 Oxford Street, Cambridge, Massachusetts 02138
corey@chemistry.harVard.edu
Received December 20, 2006
ABSTRACT
Stereochemical studies of the photocycloaddition of a chiral
3 π−π*) excited state of the enone.
r,
â
-enone and isobutylene lead to a working hypothesis for the geometry of the
(
Ultraviolet irradiation of 2-cyclohexenone (1) at 300-320
nm to effect an n f π* transition generates an electronically
excited species that is sufficiently long lived to be trapped
by isobutylene to form as major products the cis- and trans-
fused [4.2.0]-bicyclooctanones 2 and 3, with the trans isomer
predominating.1,2 It is now generally accepted that this
it is now clear that the photoreaction involves intersystem
crossing of the initially formed n f π* singlet to an n f
π* triplet which then relaxes to a less energetic π-π* triplet,
3(π-π*). Reaction of this triplet with isobutylene would be
expected to provide the triplet diradical 4 which could be
rapidly transformed via the singlet diradical 4 into the
products 2, 3, and 5.4Assuming this pathway is actually
reaction proceeds via the diradical 4, for which there is much
experimental evidence, including not only the formation of
byproduct 5 (which can result from 4 by 1,5-H migration)
but also direct trapping experiments with H2Se.3 In addition,
followed, it becomes of some interest to inquire as to the
three-dimensional geometry of the (π-π*) state. All the
evidence available suggests that this state involves twisting
3
about the R,â-bond of the R,â-enone chromophore. First of
(1) (a) Corey, E. J.; Mitra, R. B.; Uda, H. J. Am. Chem. Soc. 1963, 85,
362-363. (b) Corey, E. J.; Mitra, R. B.; Uda, H. J. Am. Chem. Soc. 1964,
86, 485-492. (c) Corey, E. J.; Bass, J. D.; LeMahieu, R.; Mitra, R. B. J.
Am. Chem. Soc. 1964, 86, 5570-5583.
(2) For related studies, see: (a) Bu¨chi, G.; Goldman, I. M. J. Am. Chem.
Soc. 1957, 79, 4741-4748. (b) Cookson, R. C.; Crundwell, E. Chem. Ind.
1958, 1004-1004. (c) deMayo, P.; Takeshita, H.; Sattar, A. B. Proc. Chem.
Soc. 1962, 119-119. (d) Eaton, P. E. J. Am. Chem. Soc. 1962, 84, 2344-
2348, 2454-2455.
(3) Maradyn, D. J.; Weedon, A. C. Tetrahedron Lett. 1994, 35, 8107-
8110.
(4) (a) Schuster, D. I.; Heibel, G. E.; Caldwell, R. A.; Tang, W.
Photochem. Photobiol. 1990, 52, 645-648. (b) Schuster, D. I.; Dunn, D.
A.; Heibel, G. E.; Brown, P. B.; Rao, J. M.; Woning, J.; Bonneau, R. J.
Am. Chem. Soc. 1991, 113, 6245-6255. (c) Yamanchi, S.; Hirota, N.;
Higuchi, J. J. Phys. Chem. 1988, 92, 2129-2133. (d) Wilsey, S.; Gonza´lez,
L.; Robb, M. A.; Houk, K. N. J. Am. Chem. Soc. 2000, 122, 5866-5876.
10.1021/ol063092r CCC: $37.00
© 2007 American Chemical Society
Published on Web 02/10/2007

all, it is widely accepted that the lowest triplet of ethylene
is the structure in which the two CH2 planes are at an angle
of 90°.5 Second, the well-known E to Z photoisomerization
of acyclic R,â-enones is most readily understood on the basis
of the twisted geometry of the electronically excited state.
In the case of acyclic R,â-enones, this R,â-twisting and
rotation provides a mechanism for rapid dissipation of energy
that explains the fact that these photoexcited substrates do
not live long enough to be trapped by olefinic partners.
2-Cycloheptenone and 2-cyclooctenone undergo Z f E
isomerization and cannot be trapped by simple olefins,6
whereas 2-cyclopentenone and 2-cyclohexenone, which are
unlikely to give E-isomers, do undergo photocycloaddition
with olefins, paralleling 1 f 2 + 3. The formation of the
highly strained and thermodynamically less stable trans
adduct 2 from 1 and isobutylene provides another argument
was only 74%. Nonetheless, the product was nicely crystal-
line and could be recrystallized from hexane to give 6 in
98% ee.
The second synthetic pathway to 6, shown in Scheme 1,
was more practical for the synthesis of gram amounts.
3
for a twisted (π-π*) precursor.7 Various theoretical and
Scheme 1. Synthesis of 6-(S)-Triphenylsilyl-2-cyclohexenone
physical studies also lend support to the R,â-twisted structure
as the minimum-energy geometry for the 2-cyclohexenone
3(π-π*) state.4 However, the three-dimensional geometry
of that state is still unclear. The present study was initiated
with two major objectives: (1) to learn more about the
preferred minimum-energy conformation(s) of the 3(π-π*)
excited state and (2) to find a methodology for exerting
effective enantiocontrol of the [2+2]-photocycloaddition
reactions of cyclic R,â-enones and olefins, such as the
reaction 1 f 2 + 3. Complimentary studies in this laboratory
have explored nonphotochemical approaches to the en-
antioselective synthesis of compounds such as 3 from R,â-
enones with some success.8
The particular 2-cyclohexenone on which we have focused
is the chiral 6-(S)-triphenylsilyl-2-cyclohexenone (6). Two
syntheses of this substrate were developed. The choice of
the triphenylsilyl group at the chiral center was based on a
number of considerations including: (1) subsequent ease of
removal from the attached carbon, (2) large effective steric
size, and (3) the unusually high crystallinity of triphenylsilyl-
containing molecules (apparently not previously appreciated).
The latter effect would seem to be a consequence of the
availability of the 6-fold phenyl embrace binding motif for
tight crystal packing first noticed with triphenylphosphino
compounds.9 The excellent crystallinity of 6 was confirmed
by an initial synthesis using the enantioselective methodology
recently developed in these laboratories,10 specifically, the
reaction of triphenylsilane and 6-diazo-2-cyclohexenone10 in
the presence of H. M. L. Davies’ chiral rhodium(II) catalyst.10
Even though this process provided access to 6, the yield could
not be raised above 28% and the enantiopurity of the product
Reaction of 1,4-cyclohexadiene oxide with triphenylsilyl-
lithium (from triphenylchlorosilane and lithium in THF at
23° for 3 h)11 proceeded smoothly to form trans-(()-2-
triphenylsilyl-4-cyclohexen-1-ol (8) which was converted to
a diastereomeric mixture of esters with the monooxalic ester
of levorotatory menthol. As we had hoped, as single
crystalline diastereomer (9) could be obtained by simple
recrystallization (from hexane-isopropyl alcohol). Base-
catalyzed hydrolysis of 9 to 10 followed by oxidation with
the Dess-Martin periodinane reagent afforded the â,γ-enone
11 in excellent yield. The isomerization of 11 to the chiral
R,â-enone 6 was attempted under a wide variety of basic
(5) El-Taher, S.; Hilal, R. H.; Albright, T. A. Int. J. Quantum Chem.
2001, 82, 242-254 and refs cited therein.
(6) Corey, E. J.; Tada, M.; LaMahieu, R.; Libit, L. J. Am. Chem. Soc.
1965, 87, 2051-2052. Although E-2-cycloheptenone does not add to simple
olefins because it rapidly reverts to the Z-isomer, it can be trapped by
cyclopentadiene to give a (nonphotochemical) Diels-Alder adduct.
(7) In the case of the photocycloaddition of 2-cyclohexenone to tetra-
methylethylene, the trans-fused adduct predominates heavily. See: Nelson,
P. J.; Ostrem, D.; Lassila, J. D.; Chapman, O. L. J. Org. Chem. 1969, 34,
811-813.
(8) Liu, D.; Hong, S.; Corey, E. J. J. Am. Chem. Soc. 2006, 128, 8160-
8161.
(9) Dance, I.; Scudder, M. J. Chem. Soc., Chem. Commun. 1995, 1039-
1040.
(10) Ge, M.; Corey, E. J. Tetrahedron Lett. 2006, 47, 2319-2321.
(11) Gilman, H.; Aoki, D.; Wittenberg, D. J. Am. Chem. Soc. 1959, 81,
1107-1110.
1058
Org. Lett., Vol. 9, No. 6, 2007

conditions but was unsatisfactory due to concomitant race-
mization. However, chiral 6 of 98% ee could be obtained
under neutral conditions using 5 mol % of RuClH(CO)-
(PPh3)3 in CH2Cl2 at 23 °C under an atmosphere of H2.12
The isomeric byproduct 7 was also formed in the isomer-
ization of 11 to 6 but was readily separable by chromatog-
raphy.
The structure and absolute configuration of 6, mp 109-
111 °C, were fully supported by X-ray single-crystal dif-
fraction which revealed the structure shown above and also
confirmed the close packing of phenyl groups both with one
another and with the 2-cyclohexenone ring, as shown in the
Supporting Information. It is interesting that the 2-cyclo-
hexenone subunits also stack face to face along the crystal
a-axis (see Supporting Information).
of the bond between the carbonyl group and the attached
silylated R carbon.
Treatment of the trans-fused bicyclic ketone 12 of 92%
ee with K2CO3-MeOH gave the cis-fused isomer 14 of 92%
ee, as expected. The absolute configuration of 14 follows
not only from the method of synthesis but also from
comparison with a sample of the same chiral compound that
was recently prepared in these laboratories.8
We believe that the photochemical results shown in
Scheme 2 can be used to illuminate the three-dimensional
3
geometry of the (π-π*) excited state of the R,â-enone 6.
To this end, we advance the following working hypothesis
which is outlined in Scheme 3. If we make the reasonable
The photocycloaddition reaction of 6 and isobutylene was
complicated by the occurrence of photoracemization of the
[2+2]-cycloaddition products at a rate competitive with their
formation (see below). It was best conducted in acetone as
solvent for a minimal reaction time, as summarized in
Scheme 2. The two principal products of the photoreaction
Scheme 3. Working Hypothesis for the Photocycloaddition of
6 and Isobutylene
Scheme 2
assumption that spin density of the π-π* triplet enone is
mainly concentrated at C(R) and C(â) with the spin-paired
electrons in orthogonal orbitals and that the C(R)- and C(â)-
hydrogens are trans to one another, two structures emerge
as possibilities. These are shown in stereoformulas 15 and
16 as two different conformers of this kind of 3(π-π*) state
in which the triphenylsilyl substituent is equatorial. Attack
of isobutylene at C(R) then would generate the 1,4-diradicals
17 from 15 and 18 from 16. Rapid ring closure of 17 would
give 12, whereas the direct cyclization of 18 would give 13.
It is clear from the inspection of the stereostructures 17 and
18 that the equatorial triphenylsilyl group would have little
influence on the face of C(R) which is attacked by iso-
butylene. The preference for initial attachment to C(R) may
arise from the inductive effect of the attached carbonyl that
renders C(R) more electron deficient and reactive than C(â).
If the chairlike conformer 15 were to predominate somewhat
over the twist-boat-like conformer 16, the 2:1 predominance
of adduct 12 over 13 would be easy to understand.
We believe that this mechanistic model, the proposed
3
conformations of the (π-π*) state of a 2-cyclohexenone,
and the question of whether the unpaired electrons are
significantly localized in spatially orthogonal orbitals deserve
further experimental and theoretical investigation.
were the trans-fused bicyclo[4.2.0]octanones 12 and 13
which were formed in ca. 2:1 ratio. Further irradiation of 12
converted it to ent-13, whereas irradiation of 13 produced
ent-12. These isomerization reactions can be interpreted as
Norrish type-1 processes13 involving the reversible cleavage
Supporting Information Available: Experimental pro-
cedures for the reactions described herein and characteriza-
tion data. Details for the X-ray crystallographic analysis of
6. This material is available free of charge via the Internet
at http://pubs.acs.org.
(12) Bingham, D.; Webster, D. E.; Wells, P. B. J. Chem. Soc., Dalton
1974, 1519-1521.
(13) Hwu, J. R.; Chen, B.-L.; Liu, C.-F.; Murr, B. L. J. Organomet.
Chem. 2003, 686, 198-201.
OL063092R
Org. Lett., Vol. 9, No. 6, 2007
1059