Mechanisms of the Base-Induced Isomerizations of Cyclopentene and Cyclohexene Oxides: Influence of Structure and Solvent on α and β Proton Removal

Full text of DOI:10.1021/jo951981e

820
J . Org. Chem. 1996, 61, 820-821
Communications
Sch em e 1
Mech a n ism s of th e Ba se-In d u ced
Isom er iza tion s of Cyclop en ten e a n d
Cycloh exen e Oxid es: In flu en ce of
Str u ctu r e a n d Solven t on r a n d â P r oton
Rem ova l
Kathleen M. Morgan*,† and J oseph J . Gajewski*
Department of Chemistry, Indiana University,
Bloomington, Indiana 47405
Received November 8, 1995
kinetic deuterium isotope effects of 3.3 ( 0.5 (benzene)
and 2.4 ( 0.4 (ether) were calculated for rate-determining
removal of the R-proton. Under these reaction conditions,
cyclopentanone is also formed (15 ( 2% in benzene, 25
( 4% in ether), but its deuterium content could not be
determined since H/D exchange occurs with workup.5
Hexamethylphosphoric triamide (HMPA) is known to
promote allyl alcohol formation for compounds which
exhibit carbenoid behavior in nonpolar solvents,1 and, in
fact, when treated with LDA in HMPA, cyclopentene
oxide gives only allyl alcohol with a deuterium distribu-
tion consistent with â-elimination (Scheme 1).6 In all
three solvents, the half-life of the reaction is 15-30 min
at 0 °C.
For comparison purposes, 1-deuteriocyclohexene oxide
was subjected to an identical series of reaction conditions,
and the allyl alcohol produced had a deuterium distribu-
tion consistent with exclusive â-elimination in all three
solvents. Cyclopentene oxide reacts 2-10 times faster
than cyclohexene oxide during in situ competition, yet
other epoxides thought to rearrange via the R-elimination
mechanism react much more slowly.1 This might be due
to greater acidity of the bridgehead proton of cyclopen-
tene oxide or to relief of ring strain at the transition state.
It is also possible that cyclohexene oxide undergoes a
rapid, unproductive reversible R-deprotonation, followed
by productive â-elimination. However, the GC/MS spec-
tra of ether and benzene reaction mixtures quenched
after partial isomerization showed no evidence of a
preequilibrium. For the reaction of 1-deuteriocyclopen-
tene oxide in HMPA, GC/MS data showed that at long
reaction times (>50% reaction) the recovered epoxide was
a mixture of non-, mono-, and dideuterated material,
though not in a statistical mixture. The 1-deuteriocyclo-
hexene oxide recovered from the reaction conducted in
HMPA shows minimal equilibration.
Because of the synthetic value of base-induced rear-
rangements of alkyl-substituted epoxides to allyl alcohols,
the reaction has been investigated in some detail.1
Frequently, allyl alcohols can be formed with high
regioselectivity, and preparation of optically active prod-
ucts using chiral lithium amide bases has shown some
success.2 In acyclic systems, the reaction proceeds via a
â-elimination (i.e., 1,2- or E2-type).1 Cyclohexene oxide
has been shown through deuterium-labeling studies to
undergo exclusive syn â-elimination in ether.3 Complex-
ation of Li+ to the epoxide oxygen apparently directs the
facial selectivity via a six-atom cyclic transition state.
Some epoxides, generally those of medium-sized cyclic
and bicyclic olefins, appear to react via a carbenoid
formed by an R-elimination followed by epoxide opening.
This pathway is supported by the observation of ketones
and transannular C-H bond insertion products.1 Allyl
alcohols are sometimes formed competitively, but whether
they result from R- or â-elimination is unclear. We report
here that cyclopentene oxide gives 2-cyclopentenol via an
R-elimination in benzene or in ether solvents but via
â-elimination in HMPA.
Cyclopentene oxide gives a mixture of three productss2-
cyclopentenol, cyclopentanone, and the product of nu-
cleophilic additionswhose proportions depend strongly
on the base used.4 The steric bulk and poor nucleophi-
licity of lithium diisopropyl amide (LDA) prevent its
addition to cyclopentene oxide; thus, LDA was employed
to study the mechanism(s) of allyl alcohol formation.
First, it was ascertained that cyclopentanone is a primary
reaction product and not due to rearrangement of allyl
alcoholate to enolate under the highly basic reaction
conditions.
When 1-deuteriocyclopentene oxide was treated with
2.5 equiv of LDA in benzene or in ether,1 2-cyclopentenol
and cyclopentanone were formed after quenching. NMR
analysis of the GC-purified alcohol showed some deute-
rium loss, and the remaining deuterium (0.77D in
benzene and 0.71D in ether) was exclusively on the
carbinol carbon (i.e., no vinyl deuterium), consistent with
the carbenoid mechanism and not â-elimination (Scheme
1). From the amount of deuterium present, primary
The importance of Li+ complexation in the nonpolar
solvents was investigated by treating epoxides with LDA
in ether containing 12-crown-4 (1 equiv with respect to
LDA). Neither cyclopentene nor cyclohexene oxide un-
derwent reaction yet, in HMPA, a strong cation-complex-
ing solvent, cyclopentene oxide reacts as rapidly as in
the less polar solvents without crown. HMPA, no doubt,
favors the dipolar transition state, and complexation of
Li+ to the epoxide oxygen is apparently not necessary for
â-elimination in this solvent.
† Current address: Department of Chemistry, College of William
and Mary, Williamsburg, VA 23187.
(1) Crandall, J . K.; Apparu, M. Org. React. 1983, 29, 345-443 and
references therein.
(2) Cox, P. J .; Simplans, N. S. Tetrahedron: Asymmetry 1991, 2,
1-26. Bhuniya, D.; Datta Gupta, A.; Singh, V. K. Tetrahedron Lett.
1995, 36, 2847-2850.
(5) It is interesting to note that here the carbenoid gives more of
the thermodynamically less stable allyl alcoholate. The origins of this
effect are under investigation.
(6) Note that at longer reaction times the deuterium content of the
allyl alcohol is diminished, indicative of R-deprotonation that does not
lead to product. The data have been corrected to take this into account.
(3) Thummel, R. P.; Rickborn, B. J . Am. Chem. Soc. 1970, 92, 2064-
2067.
(4) Crandall, J . K.; Chang, L.-H. J . Org. Chem. 1967, 32, 435-439.
Kissel, C. L.; Rickborn, B. J . Org. Chem. 1972, 37, 2060-2063.
0022-3263/96/1961-0820$12.00/0 © 1996 American Chemical Society

Communications
J . Org. Chem., Vol. 61, No. 3, 1996 821
Sch em e 2
Sch em e 4
Sch em e 3
monodeuterated compounds was treated with LDA in
HMPA, and the epoxide was recovered after 40% reac-
tion. Significant enhancement of deuterium at the anti
â position is observed along with diminution at all other
positions; thus, anti â-elimination occurs here. The
distribution of deuterium in the tertiary allyl alcohol is
consistent with these observations, and the product ratio
suggests a primary deuterium isotope effect of ap-
proximately 3 for rate-determining deprotonation at the
anti â position.
AM1,8 ab initio,9 and density functional theory10 cal-
culations agree that, in simple E2 reactions of primary
derivatives with strong base, the transition state is quite
reactant-like, with minimal dihedral angle distortion.
The gas-phase structures of cyclopentene and cyclohexene
oxides have been determined by electron diffraction,11 but
the appropriate H-C-C-O torsional angles were not
determined. However, the heavy-atom geometries were
reproduced using PCMODEL,12 which then provided the
desired torsional angle. These calculations suggest that
cyclohexene oxide can adopt a favorable syn-gauche
conformation, but the structure of cyclopentene oxide will
not easily lead to syn â-elimination. The geometry of both
epoxides is reasonably favorable for anti-eliminations,
which occur for cyclopentene oxide in HMPA.
To determine the isotope effect on the carbene insertion
into the â-CH bond, a mixture of syn- and anti-3,4-
dideuteriocyclopentene oxide was treated with LDA in
benzene (Scheme 2).
The NMR spectrum of the allyl alcohol showed deute-
rium at all positions except the carbinol carbon. No
deuterium was lost, and the distribution is interpreted
in terms of the products shown in Scheme 2. Thus, the
carbenoid mechanism derives further support. The
observed distribution of deuterium indicates a primary
kinetic isotope effect of approximately 2.3 for insertion
of the carbenoid into the adjacent carbon-hydrogen
(deuterium) bond.
A bulky R-tert-butyl group on cyclopentene oxide was
found to sterically inhibit R-elimination by a rate factor
of 1/200 relative to the parent so that â-elimination is
more favorable in both ether and benzene. As expected,
HMPA solvent promotes exclusive â-elimination (Scheme
3: note that the 9:1 ratio of the first two products
represents the ratio of D and H at C-2 of the starting
material), and here the reaction is 10 times slower than
the parent.
HMPA has been known to reverse the facial selectivity
of E2 eliminations from that which occurs in nonpolar
solvents.7 This was also found to be the case in the
cyclopentene oxide system (Scheme 4). A mixture of four
Ack n ow led gm en t. We thank Professors J ack Cran-
dall and Scott Gronert for helpful discussions. This
work was supported by the National Science Foundation.
J O951981E
(8) Dewar, M. J . S; Yuan, Y.-C. J . Am. Chem. Soc. 1990, 112, 2088-
2094. Dewar, M. J . S; Yuan, Y.-C. J . Am. Chem. Soc. 1990, 112, 2095-
2105.
(9) Gronert, S. J . Am. Chem. Soc. 1993, 115, 652-659. Gronert, S.
J . Am. Chem. Soc. 1992, 114, 2349-2354. Gronert, S.; Lee, J . M. J .
Org. Chem. 1995, 60, 4488-4497.
(10) Bickelhaupt, F. M.; Baerends, E. J .; Nibbering, N. M. M.;
Ziegler, T. J . Am. Chem. Soc. 1993, 115, 9160-9173.
(11) Landolt-Bo¨rnstein; Springer-Verlag: Berlin, Vol. II/7, 1976; Vol.
II/15, 1987.
(12) PCMODEL, Version 5.1 for Windows, Serena Software, Box
3076, Bloomington, IN 47402.
(7) There is precedent for this type of mechanistic switch: Reichardt,
C. Solvents and Solvent Effects in Organic Chemistry, 2nd ed.; VCH:
New York, 1988; Chapter 5, p 255.