Torsional, rotor, and electronic effects in 4-tert-butylmethylenecyclohexane epoxidations and osmylations

Article abstract of DOI:10.1021/ja953040p

The axial epoxidation preference for 2-substituted 4-tert-butylmethylenecyclohexanes is attributed to a combination of small effects, including existing bond torsion and rotor effects. Contributions from developing bond torsion are smaller and may be negl

Full text of DOI:10.1021/ja953040p

3556
J. Am. Chem. Soc. 1996, 118, 3556-3567
Torsional, Rotor, and Electronic Effects in
4-tert-Butylmethylenecyclohexane Epoxidations and
Osmylations
E. Vedejs,* W. H. Dent, III, J. T. Kendall, and P. A. Oliver
Contribution from the Chemistry Department, UniVersity of Wisconsin,
Madison, Wisconsin 53706
ReceiVed September 5, 1995X
Abstract: The axial epoxidation preference for 2-substituted 4-tert-butylmethylenecyclohexanes is attributed to a
combination of small effects, including existing bond torsion and rotor effects. Contributions from developing bond
torsion are smaller and may be negligible. Cieplak (σ-σ*) effects are too small to identify in most of the epoxidations,
but a marginal effect could be present according to comparisons of isosteric systems 11a and 15a or 19a and 19b.
Dimethyldioxirane epoxidations and osmylations are more sensitive to steric factors, resulting in a trend for equatorial
attack.
Introduction
well-understood steric bias and provides a realistic analogy for
more flexible acyclic systems. Third, the methylenecyclohexane
family has been instrumental over the history of attempts to
understand stereoelectronic and torsional factors, and the
literature contains many examples of epoxidation experiments,3,7-9
as well as some osmylations.3 Starting from this information
base, we hoped to clarify the origins of the axial epoxidation
preference of 4-tert-butylmethylenecyclohexane that has stimu-
lated controversy since the report by Carlson and Behn in
1967.7,10
Allylic heteroatom directing effects are useful in synthetic
strategies that rely upon acyclic stereocontrol. The largest
effects usually involve covalent interactions between substrate
heteroatoms and the reagent,1 but other factors may also control
alkene addition reactions.2-6 Earlier work from our laboratory
found no dominant stereoelectronic effects in epoxidations or
osmylations of 4-substituted 2-pentenes,5b but conformational
issues in these flexible substrates could have complicated the
interpretation. Evidence in more rigid analogs was desired.
The goal of the present work was to probe stereoelectronic,
steric, and torsional factors in the epoxidation and osmylation
of 4-tert-butylmethylenecyclohexane derivatives. This system
was chosen for several reasons. First, the bonds at C2 and C6
have a clear stereoelectronic bias. Thus, the axial bonds at C2
and C6 are antiperiplanar to developing axial bonds at C1 while
the ring C2-C3 and C5-C6 bonds are antiperiplanar to develop-
ing equatorial bonds. Second, the system contains reasonably
Background
Selected epoxidation results from prior studies are sum-
marized in Table 1. Cieplak, Tait, and Johnson showed that
the increase in percent axial epoxidation from entry 15 to entry
18 correlates with a decrease in the σ-electron donor ability of
the substituent at C3.3 Because the C2-C3 ring bond is
antiperiplanar to the developing σ* orbitals for equatorial C-O
bond formation at C1, electron-withdrawing substituents at C3
should retard the rate of equatorial attack (Cieplak effect). This
would favor increased axial selectivity from entry 15 to entry
18. Most of the other results in Table 1 were reported earlier
by Sevin and Cense along with an explanation based on torsional
effects.8
X Abstract published in AdVance ACS Abstracts, March 1, 1996.
(1) (a) Henbest, H. B.; Wilson, R. A. J. Chem. Soc. C 1957, 1958. (b)
Review: Hoveyda, A. H.; Evans, D. H., Fu, G. C. Chem. ReV. 1993, 93,
1307.
(2) (a) Raimondi, L.; Wu, Y.-D.; Brown, F. K.; Houk, K. N. Tetrahedron
Lett. 1992, 33, 4409. Houk, K. N.; Duh, H.-Y.; Wu, Y.-D.; Moses, S. R.
J. Am. Chem. Soc. 1986, 108, 2754. (b) Houk, K. N.; Paddon-Row, M.
N.; Rondan, N. G.; Brown, F. K.; Spellmeyer, D. C.; Metz, J. T.; Li, Y.;
Loncharich, R. Science 1986, 231, 1108. (c) Houk, K. N.; Paddon-Row,
M. N.; Rondan, N. G. J. Am. Chem. Soc. 1982, 104, 7162. Houk, K. N.;
Rondan, N. G.; Wu, Y.-D.; Metz, J. T.; Paddon-Row, M. N. Tetrahedron
1984, 40, 2257. (d) Houk, K. N.; Tucker, J. A.; Wu, Y.-D. J. Am. Chem.
Soc. 1991, 113, 5018. (e) Wu, Y.-D.; Houk, K. N.; Florez, J.; Trost, B. M.
J. Org. Chem. 1991, 56, 3656.
(3) Cieplak, A. S.; Tait, B. D.; Johnson, C. R. J. Am. Chem. Soc. 1989,
111, 8447. Cieplak, A. S. J. Am. Chem. Soc. 1981, 103, 4540.
(4) (a) Martinelli, M. J.; Peterson, B. C.; Khau, V. V.; Hutchison, D. R.;
Leanna, M. R.; Audia, J. E.; Drost, J. J.; Wu, Y.-D.; Houk, K. N. J. Org.
Chem. 1994, 59, 2204. (b) Bach, R. D.; Owensby, A. L.; Gonzalez, C.;
Schlegel, H. B.; McDouall, J. J. W. J. Am. Chem. Soc. 1991, 113, 2338.
(5) (a) Vedejs, E.; Gapinski, D. M. J. Am. Chem. Soc. 1983, 105, 5058.
Vedejs, E.; Dent, W. H., III; Gapinski, D. M.; McClure, C. K. J. Am. Chem.
Soc. 1987, 109, 5437. Vedejs, E.; Dolphin, J. M.; Mastalerz, H. J. Am.
Chem. Soc. 1983, 105, 127. (b) Vedejs, E.; McClure, C. K. J. Am. Chem.
Soc. 1986, 108, 1094.
(6) (a) Frenking, G.; Ko¨hler, K. F.; Reetz, M. T. Angew. Chem., Int. Ed.
Engl. 1991, 30, 1146. (b) Frenking, G.; Ko¨hler, K. F.; Reetz, M. T.
Tetrahedron 1993, 49, 3983 and references therein. (c) Nguyen, T. A.;
Maurel, F.; Lefour, J.-M. New J. Chem. 1995, 19, 353. (d) Coxon, J. M.;
Houk, K. N.; Luibrand, R. T. J. Org. Chem. 1985, 60, 418.
(7) Carlson, R. G.; Behn, N. S. J. Org. Chem. 1967, 32, 1363.
(8) Sevin, A.; Cense, J.-M. Bull. Chim. Soc. Fr. 1974, 964.
(9) (a) Reviews: Berti, G. In Top. Stereochem. 1973, 7, 97. Rao, A. S.;
Paknikar, S. K.; Kirtane, J. G. Tetrahedron 1983, 39, 2323. Rebek, J., Jr.
Heterocycles 1981, 15, 517. (b) Ballantine, J. D.; Sykes, P. J. J. Chem.
Soc. C 1970, 731. Della Casa De Marcano, D. P.; Halsall, T. G. J. Chem.
Soc., Chem. Commun. 1970, 1381. (c) Grant, P. K.; Weavers, R. T.
Tetrahedron 1974, 30, 2385. (d) Causa, A. G.; Chen, H. Y.; Tark, S. Y.;
Harwood, H. J. J. Org. Chem. 1973, 38, 1385. Sane, P. P.; Tadwalkar, V.
R.; Rao, A. S. Ind. J. Chem. 1974, 12, 444. Di Maio, G.; Vecchi, E.; Zeuli,
E. Gazz. Chim. Ital. 1983, 113, 823. Sanghvi, Y. S.; Rao, A. S. J.
Heterocycl. Chem. 1984, 21, 317. Danishefsky, S. J.; Mantlo, N. J. Am.
Chem. Soc. 1988, 110, 8129. Vereschagin, A. N.; Bakeleinik, G. A.;
Korylyaeva, G. I. Bull. Acad. Sci. USSR, DiV. Chem. Sci. 1988, 34, 2604.
Paquette, L. A.; Underiner, T. L.; Gallucci, J. C. J. Org. Chem. 1992, 57,
86. (e) Chautemps, P.; Pierre, J.-L. Tetrahedron 1976, 32, 549. (f)
Plamondon, L.; Wuest, J. D. J. Org. Chem. 1991, 56, 2066. (g) Park, T.
K.; Danishefsky, S. J. Tetrahedron Lett. 1994, 35, 2667. (h) Urones, J.
G.; Marcos, I. S.; Perez, B. G.; Diez, D.; Lithgow, A. M.; Gomez, P. M.;
Basabe, P.; Garrido, N. M. Tetrahedron 1994, 50, 10995.
(10) Preliminary communication: Vedejs, E.; Dent, W. H. J. Am. Chem.
Soc. 1989, 111, 6861.
0002-7863/96/1518-3556$12.00/0 © 1996 American Chemical Society

4-tert-Butylmethylenecyclohexane Epoxidations
J. Am. Chem. Soc., Vol. 118, No. 15, 1996 3557
Table 1. Epoxidations with XC6H4CO3H (CH2Cl2)a
tions in the staggered rotamer.12 Similarly, the torsional barrier
in propane is largely hyperconjugative, but in butane a composite
of hyperconjugative and filled orbital (steric) effects is involved.
Torsional barriers involving the sp2-sp3 bond of propene are
smaller than in ethane, and an eclipsed geometry is preferred.13
Both existing and developing bonds may influence torsional
energy as rehybridization occurs from the ground state to the
transition state. Changes in existing bond torsion were proposed
for diimide reduction,14a for the endo vs exo deprotonation of
norbornanone,14b and for 1,6-dimethylcyclohexene epoxidation.14c
The often-cited torsional argument of Felkin et al. (1968) has
a different emphasis, and specifically invokes destabilizing
interactions between partial (developing) bonds and adjacent
C-H bonds.15 Changes in other torsional interactions were not
explicitly treated, but staggered transition state geometries were
emphasized. In the case of ketone addition reactions, torsion
between the developing nucleophile- - -C1 bond and the axial
C-H bonds at cyclohexanone C2 and C6 was given as the reason
why compact reagents prefer to attack from the axial direction.
The developing axial bond was drawn assuming a 90° bond
angle with respect to the carbonyl plane, and a dihedral angle
of ca. 10° was shown between the developing nucleophile- - -
C1 bond and the adjacent axial C2-H and C6-H bonds. This
would be close to an eclipsed geometry between the developing
bond and the existing axial C2-H and C6-H bonds. Klein and
Lichtenberger later suggested that release of torsional strain
between equatorial C2-H and C6-H bonds and the adjacent
CdO bond may also favor axial attack.16 Felkin et al.
commented on methylenecyclohexane epoxidations in a brief
footnote,15b but did not indicate whether the same torsional
effects were intended for epoxidations as for ketone additions.
Sevin and Cense discussed methylenecyclohexane epoxida-
tions in greater detail.8 They noted that an axial C2 methyl
group should affect epoxidation stereochemistry if developing
bond torsion between axial C2-H and C1- - -O is the reason
for axial selectivity in the parent compound as proposed by
Felkin et al.15b Only a small difference was found (Table 1;
compare entries 10 and 11) and the authors concluded
that torsional interactions between the developing equatorial
C1- - -O bond and the axial C2 and C6 hydrogens are unimpor-
tant. Advanced bonding was invoked between peracid oxygen
and the primary (methylene) carbon compared to the tertiary
(C1) carbon (asynchronous transition state).17 Decreased axial
selectivity was expected in a more nearly synchronous transition
state due to increased reagent interactions with axial C3 and C5
hydrogens. Supporting evidence was found in the epoxidation
of 4-tert-butyl-1-isopropylidenecyclohexane (57% axial epoxide
vs 70% axial in the methylene analog).8
a Reaction at 0 °C.
Torsional strain is related to the barrier for rotation around
single bonds.11-13 In the Pitzer interpretation,11 torsional strain
in ethane is attributed to destabilization between filled orbitals
in the eclipsed conformation. This view has been widely
accepted by organic chemists, resulting in the generalization
that all eclipsing interactions along sp3-sp3 bonds are desta-
bilizing. However, increasingly sophisticated theoretical treat-
ments are reaching a rather different consensus. The filled
orbital interactions in the eclipsed rotamer of ethane cannot
cause a barrier to rotation because they haVe been shown to be
stabilizing in nature.12,13d One recent interpretation associates
the ethane barrier with C-C bond lengthening in the eclipsed
rotamer and a corresponding decrease in attractive interactions.13d
An alternative approach (Weinhold et al.)12 uses natural bond
orbital (NBO) analysis of high-level calculations and attributes
the barrier to stabilizing hyperconjugative σ-CH, σ*-CH interac-
Houk et al. have developed a general treatment of addition
reactions at sp2-hybridized carbon that recognizes torsional,
stereoelectronic, and electrostatic factors.2 Nucleophilic addition
reactions of ketones were analyzed using ab initio methods and
MM2 models were developed where the torsional contributions
from developing bonds as well as from existing bonds depend
on the choice of MM2 parameters. In a related study of
epoxidations, Martinelli, Houk et al. correlated transition state
preferences using a developing bond torsional argument.4a They
also referred to the ab initio transition state geometry calculated
(11) Pitzer, R. M. Acc. Chem. Res. 1983, 16, 207.
(12) Weinhold, F.; Brunck, T. K. J. Am. Chem. Soc. 1979, 101, 1700.
Wesenberg, G.; Weinhold, F. Int. J. Quantum Chem. 1982, 21, 487. Reed,
A. E.; Weinhold, F. Isr. J. Chem. 1991, 31, 277.
(13) (a) Allinger, N. L.; Hirsch, J. A.; Miller, M. A.; Tyminski, I. J. J.
Am. Chem. Soc. 1968, 90, 5773. (b) Lowe, J. P. Prog. Phys. Org. Chem.
1968, 6, 1. (c) Wiberg, K. B.; Martin, E. J. Am. Chem. Soc. 1985, 107,
5035. Wiberg, K. B. J. Am. Chem. Soc. 1986, 108, 5817. (d) Bader, R. F.
W.; Cheeseman, J. R.; Laidig, K. E.; Wiberg, K. B.; Breneman, C. J. Am.
Chem. Soc. 1990, 112, 6530. (e) Dorigo, A. E.; Pratt, D. W.; Houk, K. N.
J. Am. Chem. Soc. 1987, 109, 6591.
(14) (a) Garbisch, E. W., Jr.; Schildcrout, S. M.; Patterson, D. B.;
Sprecher, C. M. J. Am. Chem. Soc. 1965, 87, 2932. (b) Schleyer, P. v. R.
J. Am. Chem. Soc. 1967, 89, 699, 701. (c) McCurry, P. M., Jr. Tetrahedron
Lett. 1971, 18, 1841.
(15) (a) Che´rest, M.; Felkin, H.; Prudent, N. Tetrahedron Lett. 1968, 9,
2199. (b) Felkin, H.; Che´rest, M. Tetrahedron Lett. 1968, 9, 2205. Footnote
8 mentions epoxidations.
(16) Klein, J.; Lichtenberger, D. J. Org. Chem. 1970, 35, 2654.
(17) Hanzlik, R. P.; Shearer, G. O. J. Am. Chem. Soc. 1975, 97, 5231.

3558 J. Am. Chem. Soc., Vol. 118, No. 15, 1996
Vedejs et al.
by Bach et al.4b for ethylene + performic acid, but neither group
The 2-methyl-2-methoxy-4-tert-butylcyclohexanones were
obtained from the mixture of 5a + 5b via enol silane 7a.23
Attempts to prepare 7a using the method of Krafft and Holton24a
gave a 3:1 mixture of 7a and 8a. However, the procedure of
Miller and McKean (Me3SiI; [Me3Si]2NH)24b worked well and
afforded 7a with less than 5% 8a (NMR assay). Oxidation of
7a with dimethyldioxirane (DMD)25,26 then gave the separable
acyloins 9a and 9b, 1.0:1.5 ratio (85% yield), and Wittig
methylenation afforded 10a and 10b (>90% yield). The
stereochemistry was established by NOE studies. A 10% NOE
enhancement was observed between the equatorial methyl and
adjacent methylene protons in 10a, but no analogous effect was
seen in the axial methyl isomer 10b. Both isomers 10a and
10b were then methylated (NaH, MeI) to afford the methyl
ethers 11a and 11b. A similar methylation procedure converted
the known 2-hydroxy-4-tert-butylmethylenecyclohexanes 12a
and 12b9e into the methyl ethers 13a and 13b.
An analogous sequence was used to prepare the isomeric
2-ethyl-2-methyl-4-tert-butylmethylenecyclohexanes. Thus, treat-
ment of 7a with methyllithium at 0 °C followed by ethyl iodide
afforded a 3:1 ratio of 14a and 14b. Wittig olefination gave
the corresponding ratio of alkenes 15a and 15b. However,
neither the alkenes 15 nor ketones 14 could be separated into
individual isomers. The sequence was therefore repeated with
a reversal in the order of alkylation steps. Thus, the ethyl ketone
16a,b was converted into enol silane 7b, and enolate generation
followed by methylation gave a mixture of ketones 14a and
14b in a ratio of 1:3. The yield of 14 was poor, but Wittig
methylenation was uneventful and afforded a 1:3 ratio of 15a:
15b, 80% yield. Thus, two different mixtures of the isomers
15a,b were available, and the stereoselectivity of epoxidations
or osmylations could be deduced by comparing product ratios
with the ratio of starting materials in two sets of parallel
experiments.
optimized the geometry for more complicated substrates.
Several groups have reported experiments designed to
separate steric, torsional, and stereoelectronic variables using
substrates with minimal bias close to the reacting sp2 carbon.18
Trends in the epoxidation of the unbiased 4-substituted meth-
yleneadamantanes are consistent with the Cieplak effect.18a
Similarly, changes in the equatorial 3-substituent in methyl-
enecyclohexanes should not affect steric or torsional variables.
Selectivity trends (Table 1; entries 14-18) should therefore
reflect the electronic (not steric or torsional) properties of the
substituents.3
Allylic oxygen-containing systems of interest to synthetic
chemists are inherently biased, but analogs designed within the
methylenecyclohexane skeleton allow bias to be tested in a well-
studied environment.19,20 Methylecyclohexane inversion occurs
via a half-chair TS with nearly coplanar C6, C1, C2, and C3
(inversion barriers, parent:20a ∆G* ) 8.4 kcal/mol; 2-methyl,19a
∆G* ) 9.0 kcal/mol; 2-methoxy,19a ∆G* ) 9.0 kcal/mol; 2,2-
dimethyl,20b ∆G* ) 8.1 kcal/mol). Thus, ground state torsional
contributions from the existing bonds should vary by less than
1 kcal/mol as the 2-H, 2-methyl, and 2-methoxy substituents
are interchanged. However, torsional contributions in the
transition state might depend on the extent of rehybridization.
This is because the torsional energy minimum for an alkene
usually has an eclipsed allylic bond, while the minimum for
sp3-hybridized bonds is likely to be a staggered geometry.13
Replacement of C2-H or C2-alkyl groups by electron-
withdrawing substituents could lead to relatively late transition
states that would be increasingly destabilized by eclipsed bonds
compared to early transition states.
To address the role of torsional variables, a study of the
epoxidations and osmylations of 4-tert-butylmethylenecyclo-
hexane and the 2-methyl-, 2-methoxy-, and 2-methyl-2-methoxy
derivatives was initiated. Rigid analogs that constrain the C2
oxygen and carbon substituents to a spiro-fused tetrahydrofuran
ring were also studied to evaluate rotor effects. Finally, the
2-methyl-2-ethyl (hydrocarbon) analog was investigated. This
alkene is isosteric with the 2-methyl-2-methoxy derivatives but
it contains no allylic oxygen and has minimal electronic bias.
The stereochemistry assigned to 14a and 14b was determined
1
using lanthanide shift reagents. The H NMR chemical shift
of the equatorial 2-methyl signal of 14a was shifted from δ
0.98 ppm [CDCl3] to δ 1.79 ppm [CDCl3 + Eu(fod)3] under
conditions where the axial 2-methyl isomer 14b gave chemical
shifts of δ 1.12 ppm (no shift reagent) and 1.73 ppm [Eu(fod)3
added]. Thus, the relative chemical shifts of the 2-methyl groups
were inverted by the lanthanide reagent. The equatorial methyl
should be influenced more strongly because it is in the plane
of the carbonyl oxygen. The shift reagent-based assignment is
consistent with preferred axial enolate alkylation in both
sequences leading to 14a and 14b, as expected from literature
analogies.22
The same techniques were applied to the synthesis of alkenes
19a and 19b starting from the dimethylhydrazone of 4-tert-
butylcyclohexanone. The mixture of alkylated ketones 17a,b
starting from the dimethylhydrazone of 4-tert-butylcyclohex-
anone. The mixture of alkylated ketones 17a,b was taken
directly to enol silane 7c, and treatment with MCPBA gave a
Preparation of Starting Materials
Alkylation of 4-tert-butylcyclohexanone21 produced a mixture
of the known 2-methyl-4-tert-butylcyclohexanones22 5a and 5b
in a ratio of ca. 1:3 equatorial-axial methyl after hydrolysis.21
Conversion of 5a into the methylene derivative 6a proceeded
smoothly under conventional Wittig conditions, but the axial
methyl isomer 6b was partly isomerized using this procedure.
Lombardo olefination proved more reliable and gave 6b in 68%
yield.
(18) (a) le Noble, W. J.; Srivastava, S. J. Am. Chem. Soc. 1987, 109,
5874. le Noble, W. J.; Li, H. Recl. TraV. Chim. Pays-Bas. 1992, 111, 199.
(b) Ohwada, T. J. Am. Chem. Soc. 1992, 114, 8818. (c) Halterman, R. L.;
McEvoy, M. A. Tetrahedron Lett. 1992, 33, 753. (d) Ohwada, T.; Okamoto,
I.; Haga, N.; Shudo, K. J. Org. Chem. 1994, 59, 3975. (e) Adcock, W.;
Cotton, J.; Trout, N. A. J. Org. Chem. 1994, 59, 1867.
(19) (a) Lessard, J.; Phan Viet, M. T.; Martino, R.; Saunders, J. K. Can.
J. Chem. 1977, 55, 1015. (b) Phan Viet, M. T.; Lessard, J.; Saunders, J.
K. Tetrahedron Lett. 1979, 29, 317.
(23) Fleming, I.; Iqbel, J. Tetrahedron 1983, 39, 841.
(24) (a) Krafft, M. E.; Holton, R. A. Tetrahedron Lett. 1983, 24, 1345.
(b) Miller, R. D.; McKean, D. R. Synthesis 1979, 730.
(25) (a) Murray, R. W.; Jeyaraman, R. J. Org. Chem. 1985, 50, 2847.
(b) Baumstark, A. L.; McCloskey, C. J. Tetrahedron Lett. 1987, 28, 3311.
(c) Adam, W.; Chan, Y.-Y.; Cremer, D.; Gauss, J.; Scheutzow, D.; Schindler,
M. J. Org. Chem. 1987, 52, 2800.
(20) (a) Anet, F. A. L.; Chmurney, G. N.; Krane, J. J. Am. Chem. Soc.
1973, 95, 4423. (b) Bernard, M.; Canuel, L.; St-Jacques, M. J. Am. Chem.
Soc. 1974, 96, 2929.
(21) Corey, E. J.; Enders, D. Tetrahedron Lett. 1976, 17, 3. Corey, E.
J.; Knapp, S. Tetrahedron Lett. 1976, 17, 1687.
(22) (a) House, H.; Tefertiller, B. A.; Olmstead, H. D. J. Org. Chem.
1968, 33, 935. (b) Allinger, N. L.; Blatter, H. M. J. Am. Chem. Soc. 1961,
83, 994.
(26) Previous studies of stereoselective epoxidations with DMD: (a)
Halcomb, R. L.; Danishefsky, S. J. J. Am. Chem. Soc. 1989, 111, 6661. (b)
Adam, W.; Abou-Elzahab, M.; Saha-Moller, C. R. Liebigs Ann. Chem. 1991,
445. (c) Schulz, A. G.; Harrington, R. E.; Tham, F. S. Tetrahedron Lett.
1992, 33, 6097. (d) Kurihara, M.; Ito, S.; Tsutsumi, N.; Miyata, N.
Tetrahedron Lett. 1994, 35, 1577. (e) Adam, W.; Mu¨ller, M.; Prechtl, F.
J. Org. Chem. 1994, 59, 2358. (f) Murray, R. W.; Singh, M.; Williams, B.
L.; Moncrieff, H. M. Tetrahedron Lett. 1995, 36, 2437.

4-tert-Butylmethylenecyclohexane Epoxidations
J. Am. Chem. Soc., Vol. 118, No. 15, 1996 3559
Table 2. Oxidations with MCPBA, Dimethyldioxirane (DMD),
and OsO4
Chart 1
a Initial product ratios through ca. 50% conversion; decomposition
was detected using longer reaction times that allowed higher conversion.
b The ratio could not be determined because of partial product
decomposition.
Isomer ratios and assignments were established using 1H and
13C NMR methods on the initial product mixture, prior to
separation. Epoxides having an axial CH2 group have charac-
teristic long-range coupling involving one of the epoxide CH2
protons and a ring methylene proton (J ) ca. 1.5-2.5 Hz).9c
Furthermore, the same isomers experience characteristic down-
field 13C shifts for the C5 ring methylene carbon, due to
deshielding by equatorial epoxide oxygen at C1.27 The differ-
ence is not large (1-2 ppm), but the 13C chemical shifts for C5
are very consistent (δ 24-25 ppm for the epoxides with axial
C1 oxygen and 26-27 ppm for the isomers having equatorial
mixture of hydroxy ketones 18a:18b. Wittig methylenation,
deprotection (Bu4NF), and Mitsunobu cyclization then produced
19a and 19b. The stereochemical assignment was based on
13C analysis at the stage of 19a,b using the characteristic
deshielding effect of equatorial oxygen at the γ-carbon.27 Thus,
a comparison of 13C chemical shifts for the C4 methine carbon
signals gave values of δ 43.7 for 19a (axial C2-O) and 46.8
for 19b (equatorial C2-O). Similar differences were seen in
the 13C spectra of products obtained from 19a,b by epoxidation
or osmylation (see Experimental Section, Table 4).
1
C2 oxygen; see Table 4). Both the H and 13C criteria gave
self-consistent assignments of stereochemistry.
Methylenecyclohexane Osmylations
Hydroxylation experiments were performed using the catalytic
osmylation method of Van Rheenen et al. (room temperature,
acetone solution).28 As before, product stereochemistry was
assigned by 13C NMR spectroscopy based on the characteristic
deshielding effect of equatorial C1 oxygen on the ring methylene
carbon at C5.27 This signal appears as the highest field
methylene carbon in all of the diols 22 and 23 or the derived
acetonides 24 and 25 (obtained by treatment of the diols with
2,2-dimethoxypropane and TsOH) reported in Table 2. Thus,
23 or 25 having equatorial oxygen were characterized by C5
chemical shifts in the range of δ 23.2-24.5 ppm while the axial
oxygen isomers 22 or 24 were assigned based on C5 δ values
of 21-22 ppm. Several of these assignments were confirmed
by chemical correlation with the epoxides. Thus, epoxides
obtained from Table 2, entries 4, 5, 9, and 11, were converted
stereospecifically into the diols using aqueous NaOH in tert-
butyl alcohol at 70° (entries 4, 5, 9) or aqueous NaHCO3 in
N-methylpyrrolidinone at 130 °C (entry 11). The results were
consistent with the 13C chemical shift assignments.
Methylenecyclohexane Epoxidations
Conversion of the alkenes into epoxides followed one of two
protocols. In the first, the alkene was treated with 1.5 equiv of
m-ClC6H4CO3H (MCPBA) in dichloromethane at 0 °C (3 h)
followed by warming to room temperature. The second
procedure employed dimethyldioxirane as the oxidant.25,26 The
reagent was prepared in acetone (ca. 0.1 M)25 and a large excess
(15 equiv) was added to the alkene in dichloromethane at 0 °C.
The mixture was then allowed to warm to room temperature.
The dioxirane reagent was applied to a limited subset of alkenes
because the reactions proved to be relatively nonselective and
not very sensitive to substituent effects. Also, the presence of
acetone and water in the dioxirane distillate caused problems
with sensitive substrates. Thus, 19a and 19b gave mixtures
upon attempted DMD epoxidation and the diastereoselectivity
could not be determined.
(27) (a) Senda, Y.; Ishiyama, J.; Imaizumi, S. Tetrahedron 1975, 31,
1601. (b) Davis, R.; Kluge, A. F.; Maddox, M. L.; Sparacino, M. L. J.
Org. Chem. 1983, 48, 255.
(28) Van Rheenen, V.; Kelly, R. C.; Cha, D. Y. Tetrahedron Lett. 1976,
23, 1973.

3560 J. Am. Chem. Soc., Vol. 118, No. 15, 1996
Vedejs et al.
Selectivity Patterns
selectivity is virtually unaffected by the presence of a 2-methyl
substituent in the 4-tert-butylmethylenecyclohexane skeleton
(compare MCPBA entries 1, 2, 6; 62-69% axial attack). The
equatorial 2-ethyl, axial 2-methyl example (entry 8; 59% axial
attack) also gives essentially identical results. Only the axial
2-ethyl, equatorial 2-methyl substrate (entry 12) stands out in
epoxidation selectivity (92% axial) among the examples that
contain alkyl (not oxygen) substituents at C2. No significant
difference in σ-donor properties is expected for axial methyl
vs axial ethyl, so the contrast between entries 8 and 12 must
have a different origin. Axial epoxidation in entry 12 (axial
2-ethyl) is probably favored because equatorial attack destabi-
lizes one of the ethyl rotamers, resulting in an entropic penalty.
The rotor effect is less important for equatorial ethyl because
unfavorable rotamer interactions with the peracid occur in both
the equatorial and the axial pathways. The other hydrocarbon
entries in Table 2 are unexceptional, and there is no indication
that the presence of axial C2-H vs C2-alkyl bonds is a factor
in epoxidation selectivity.
The hyperconjugative rationale of Cieplak, Tait, and Johnson
for 3-substituted methylenecyclohexanes (Table 1, entries 15-
18) is independent of the C-H vs C-C σ-donor issue discussed
above. On the other hand, the differences in selectivity for
entries 15-18 are modest, and such small effects are difficult
to evaluate. A more decisive interpretation should be possible
for variable substituents placed at C2, one bond closer to the
reacting center, provided that other factors can be held constant.
An axial C2 σ-bond has the geometry required for interaction
with the σ* orbital for a developing axial C- - -O bond at C1,
and axial acceptor groups at C2 should retard axial epoxidation
if the σ,σ* effect is significant.
Entries 10, 11, and 12 in Table 2 (MCPBA epoxidations)
allow comparisons between rotor groups that are similar in size
and shape, but that differ predictably in acceptor properties
according to Taft σI values: ethyl, σI ) -0.01; acetoxy, σI )
+0.38; methoxy, σI ) + 0.30).29 The difference in epoxidation
selectivity between C2-ethyl (entry 12, 92% axial) and the C2
oxygen substituents (entry 10, OAc, 83% axial; entry 11, OMe,
83% axial) is qualitatively in the direction predicted by the
Cieplak effect, but the difference between the C2-ethyl and
the C2-oxygen substituents is surprisingly modest. The
selectivity in the methoxy and acetoxy vs ethyl examples could
also be attributed to differences in steric effects (filled orbital
interactions) or dipole effects, or to a combination of small,
opposing steric and Cieplak effects. The rigid, spiro-fused
tetrahydrofurans of entry 13 (axial oxygen at C2; 73% axial
attack) and entry 14 (axial CH2CH2 and equatorial oxygen at
C2; 81% axial attack) follow the same trend. Some of the other
oxygen-substituted examples in Table 2 (entries 4, 5, and 9)
cannot be compared with confidence because the isosteric
2-alkyl analogs were not studied. However, there is no clear
indication that σ,σ* effects are important, and the similarity
between entries 9 and 11 suggests that the rotor effect of an
unsymmetrical substituent (methoxy) in the axial position
dominates in both cases.
The oxidations summarized in Table 2 follow a unique pattern
of selectivity for each specific reagent. The well-known Henbest
effect of axial OH is responsible for the observed equatorial
selectivity in the MCPBA reactions of entries 15 and 16.1a The
situation is less clear for the dimethyldioxirane (DMD) and
osmylation experiments because several of the reactions (entries
1-5) proceed with similar equatorial selectivity. However,
some evidence for a Henbest-like effect was obtained from
solvent studies. The osmylation of entry 16 was repeated in
dichloromethane (stoichiometric osmylation conditions; pyridine
as the activating ligand). This gave only 5% axial attack,
corresponding to increasingly favored bonding syn to hydroxyl
oxygen. Since the catalytic osmylation experiments of Table
1 are performed in the presence of hydroxylic species, the
solvent effect suggests a modest contribution from hydrogen
bonding in the dichloromethane experiment. The standard DMD
oxidation conditions are relatively free of hydroxylic agents
(acetone-dichloromethane solution; traces of water), so the
oxidation (entry 16) was repeated in a mixture of acetone and
methanol as the cosolvent to suppress intramolecular hydrogen
bonding. The selectivity changed from 17% axial (acetone-
methylene chloride) to 33% axial (acetone-methanol), consis-
tent with a moderate Henbest effect in the first example. A
similar solvent effect for the dimethyldioxirane epoxidation of
2-cyclohexenol has been attributed to hydroxyl participation.26f
Thus, the axial OH entries for each of the three reagents
probably include at least some component of hydrogen bonding
that favors oxidation syn to the OH group. However, the
remaining data points indicate substantial selectivity differences
among the reagents and will be discussed according to the
reagent. The MCPBA reactions will be considered in detail to
establish a basis for comparing electronic, torsional, and steric
factors.
The most striking feature of the MCPBA results for entries
1-14, Table 2, is their similarity. All of these reactions afford
the axial epoxide as the major product, as do most of the earlier
epoxidation examples reported in Table 1. In the absence of
substantial opposing steric effects, the axial epoxidation pathway
is preferred by ∆∆G* ) 0.8 ( 0.6 kcal/mol. A clear equatorial
preference is seen only in the case of substrates that contain an
axial substituent at C3 or C5 (for example, entry 13, Table 1),
and in the Henbest examples already discussed. Substituents
at C2 (and/or C6) cause relatively modest changes in selectivity.
The most dramatic “effect” is observed with the relatively
mundane C2 ethyl substituent (entry 12; ∆∆G* ) 1.4 kcal/mol;
compare to parent ∆∆G* ) 0.5 kcal/mol, entry 1).
Cieplak’s explanation for the axial epoxidation preference
of the parent 4-tert-butylmethylenecyclohexane (Table 1, entries
2 and 3) is that the axial C-2 and C-6 hydrogens are better σ
donors than are the equatorial C2-C3 and C5-C6 carbon bonds.
If this is correct, then the axial pathway would have a rate
advantage because of better σ-σ* overlap. In addition, the
concept predicts that a molecule containing axial C2-H and
C6-H bonds should react faster than an analog having axial
C2-alkyl substituents. Thus, an axial methyl group at C2 should
decrease the rate of the axial epoxidation pathway. A com-
parison of Table 1, entries 10 and 11, shows a small trend in
the opposite direction.8 Axial C2 methyl decreases the rate of
the equatorial pathway, probably from a small steric effect, but
it slightly accelerates the axial epoxidation. The results are not
consistent with stabilization of the axial epoxidation transition
state by axial C2-H or C6-H bonds.
One of the largest differences among the isosteric examples
in Table 2 involves the equatorial ethyl vs equatorial methoxy
examples (entry 8 vs entry 7; both with axial C2 methyl). Since
the equatorial C2-C or C2-O substituent is roughly orthogonal
to the developing epoxide bond at C1, the σ,σ* effects are not
responsible for this difference. More likely, the difference
between equatorial ethyl and methoxy groups arises because
developing bonds in both the axial and the equatorial epoxida-
tion transition states can feel the steric effect of an unsym-
A similar conclusion follows from facial selectivity com-
parisons for the hydrocarbon entries of Table 2. Epoxidation
(29) Charton, M. Prog. Phys. Org. Chem. 1981, 13, 119.

4-tert-Butylmethylenecyclohexane Epoxidations
J. Am. Chem. Soc., Vol. 118, No. 15, 1996 3561
Table 3. Dihedral Angles (R2eq-C-C-Cexo) in Alkenes and
Epoxides (MM2)
plane. This is the eclipsed geometry that corresponds to the
energy minimum for unconstrained alkenes.13 The epoxide
entries show larger variations and the dihedral angles range from
-15 to -41° for the products of axial epoxidation (equatorial
C
C
exo) and from -4 to +11° for the equatorial epoxides (axial
exo). According to the convention, R2eq is below the plane as
drawn for C2-C1-Cexo when the R2eq-C2-C1-Cexo dihedral
angles are positive, and above the plane when the angles are
negative. In the parent 4-tert-butylmethylenecyclohexane sys-
tem, the epoxide R2eq-C2-C1-Cexo dihedral angle is -21.4°
for the axial epoxidation product 20 and +6.4° for the equatorial
isomer 21 (R2eq ) R2ax ) H). Therefore, intermediate values
between these limits and the -1.5° dihedral angle of the starting
alkene 1b are expected in the competing transition states for
epoxidation. It follows that there must be substantial eclipsing
between existing bonds (R2eq-C2 and C1-Cexo) in the equatorial
epoxidation transition state. On the other hand, the axial
epoxidation pathway involves a decrease in eclipsing as bonding
proceeds. Unfavorable torsional interactions in the alkene
substrate will therefore develop gradually (if at all) in the axial
pathway, and the torsional contribution to ∆G* from existing
bonds will be lower than in the equatorial pathway. The
argument assumes that staggered carbon-carbon bonds will be
preferred in epoxides, as in other sp3-hybridized structures.13 If
this is correct, then the partially rehybridized equatorial transition
state will encounter a larger fraction of the increase in torsional
energy along the reaction coordinate, and the result will be an
advantage for axial epoxidation. In its essential features, this
is the same torsional argument that was advanced by Sevin and
Cense in 1974.8
No other useful correlations emerged from the MM2 com-
parisons. For example, the change in sign from positive dihedral
angles in a few of the alkenes (entries 5, 8, 9, and 10) to negative
angles in the axial epoxides could not be associated with distinct
selectivity behavior. The sign change requires that R2eq must
slip past the C1-Cexo bond, but the resulting eclipsing interaction
apparently occurs early along the reaction coordinate where
eclipsed geometries still correspond to the torsional energy
minimum.
metrical rotor substituent in the equatorial position. The ethyl
and methoxy rotamers will experience similar torsional and
steric interactions, but the equatorial methoxy rotamers will also
encounter dipole and electron pair interactions in the competing
transition states. We do not raise these issues in an attempt to
rationalize entries 7 and 8, but to argue that the σ,σ* effects
are small compared to other variables. This is also clear in the
rigid spiro-fused tetrahydrofurans of entries 13 and 14 where
no rotor effect is possible. As in the axial C2 ethyl vs axial C2
methoxy comparison, the modest increase in axial MCPBA
epoxidation from entry 13 (73%) to entry 14 (81%) probably
reflects the sum of small Cieplak, dipole, and steric effects, all
of which are expected to favor axial attack. It is apparent that
none of these factors is dominant since their sum amounts to a
∆∆G* in the range of 0.3-0.5 kcal/mol.
The magnitude of transition state torsional energy contribu-
tions to ∆∆G* from substrate bonds can be estimated by
comparing ground state CH3-C torsional barriers for methyl-
substituted alkenes with the barriers of the corresponding
epoxides:31 propylene oxide, 2.56 kcal/mol;31c 1-propene, 1.95
kcal/mol;31a trans-2-butene oxide, 2.44 kcal/mol;31d trans-2-
butene 1.95 kcal/mol;31a cis-2-butene oxide, 1.61 kcal/mol;31e
cis-2-butene, 0.75 kcal/mol.31b According to these examples,
the torsional component of ground state energy will increase
by ca. 0.5-1.0 kcal/mol as the alkene undergoes the bonding
and hybridization changes required to form the corresponding
epoxide. Some fraction of this increase in torsional energy will
be felt in the epoxidation transition states, depending on the
degree of rehybridization. The dependence of torsional energies
vs bond angles is not known in the methylenecyclohexane series,
but a similar range in the torsional energy difference between
alkenes and the corresponding epoxides is plausible.
Torsional Effects in Epoxidations
Qualitative torsional energy comparisons require some knowl-
edge of molecular geometry along the reaction coordinate and
of the associated change in hybridization. Several of the
hydrocarbon substrates and derived epoxides were therefore
analyzed using the MM2 force field as implemented in
MACROMODEL.30 Little change was found in the preferred
geometry along the C3-C5 ring segment or in the ring bond
angles, but the dihedral angles calculated between equatorial
The magnitude of the increase in existing bond torsional
interactions is in the range of epoxidation ∆∆G* values.
According to the arguments presented earlier, the axial epoxi-
substituents at C2 (C2-R2eq) and the methylene bond (C1-Cexo
)
were sensitive to the nature of C2 substituents (Table 3). All
of the methylenecyclohexanes with R2ax ) H have the R2eq
substituent within a ca. 4° dihedral angle of the C2-C1-Cexo
(31) (a) Kilpatrick, J. E.; Pitzer, K. S. J. Res. Natl. Bur. Stand. 1946,
37, 163. (b) Sarachman, T. N. J. Chem. Phys. 1968, 49, 3146. Kondo, S.;
Sakurai, Y.; Hirota, E.; Morino, Y. J. Mol. Spectrosc. 1970, 34, 231. (c)
Herschbach, D. R.; Swalen, J. D. J. Chem. Phys. 1958, 29, 761. (d)
Emptage, M. R. J. Chem. Phys. 1967, 47, 1293. (e) Sage, M. L. J. Chem.
Phys. 1962, 35, 142.
(30) Mohamadi, F.; Richards, N. G. J.; Guida, W. C.; Liskamp, R.;
Lipton, M.; Caufield, C.; Chang, G.; Hendrickson, T.; Still, W. C. J. Comput.
Chem. 1990, 11, 440.

3562 J. Am. Chem. Soc., Vol. 118, No. 15, 1996
Vedejs et al.
dation transition state should experience little if any increase
in existing bond torsion compared to the equatorial transition
state. Thus, it is possible that most if not all of the 0.5 kcal/
mol ∆G* advantage for axial epoxidation in the parent 4-tert-
butylmethylenecyclohexane can be obtained from changes in
existing bond torsion from the alkene to the equatorial epoxide.
No torsional contribution to ∆∆G* from the developing C-O
bonds would be necessary according to this analysis and the
question must now be asked whether developing bond torsion
plays any role in epoxidations.
torsion in the product. However, in the general case (especially
for early transition states) it may be best to merge the concepts
of developing bond torsion and traditional steric effects because
neither can be independently measured or evaluated in the
transition state. In the remaining discussion, we will focus on
existing bond torsional effects, extrapolated from the ground
state to the transition state, and in some examples we will also
refer to developing bond steric (not torsional) effects.
Dioxirane Epoxidations, Osmylations
The subsequent discussion is intended to unify and to clarify
the often divergent torsional arguments of different groups where
possible, and to establish criteria that may be useful for
evaluating torsional rationales. First, we note that the evidence
does not require that the magnitude of developing bond torsional
energy is negligible, nor does the argument depend on any
specific range of dihedral angle values. The data of Sevin and
Cense8 as well as our own would be consistent with the scenario
where the ∆G* component due to developing bond torsion is
significant, but happens to be similar for both diastereomeric
transition states (i.e., the developing bond contribution to ∆∆G*
is negligible).
Both the dimethyldioxirane and osmium tetroxide reagents
are sufficiently bulky that the transition state differences in
existing bond torsion are small compared to steric differences
in reagent-substrate interactions that favor equatorial attack.
With the dioxirane, an equatorial C2 methyl group reinforces
the trend for equatorial epoxidation to a small extent (Table 2,
entries 1 vs 2), while an axial C2 methyl (entry 6) is enough to
cancel the equatorial preference, resulting in a nonselective
reaction. The presence of an unsymmetrical rotor in the
equatorial position (C2-methoxy; C2-ethyl) in addition to the
axial C2 methyl group restores the equatorial preference (entries
7 and 8). However, axial methoxy or ethyl rotors (entries 11
and 12) are dominant over the small effect of equatorial C2
methyl, resulting in axial selectivity. This simple picture shows
that dimethyldioxirane has moderately increased steric demand
compared to MCPBA, a result that can be attributed to the gem-
dimethyl groups. Significant sensitivity to substrate steric effects
can also be deduced from other reported diastereoselective
epoxidations using dimethyldioxirane.26
The osmylations show an increased trend toward equatorial
products in the parent alkene 1b (Table 2, entry 1), and the
trend is reinforced by equatorial substituents at C2 (entries 2-5).
An axial methyl group at C2 inverts the normal equatorial
preference and promotes axial osmylation (entry 6), and the axial
ethyl rotor of entry 12 has a similar, but larger effect. The most
striking trend among the osmylations is the tendency for bond
formation to occur away from unconstrained ether or ester
oxygen substituents (Kishi effect).32 This effect dominates all
others, and a combination of lone pair repulsions and the rotor
effect of unsymmetrical substituents appears to be responsible.
The lone pair component is suggested by the large effect from
axial methoxy vs axial acetate (compare entries 4 and 5, or 10
and 11). However, the contrast between entries 11 and 13
suggests that the rotor and lone pair effects are interconnected.
The constrained axial alkoxy group (entry 13) is not nearly as
effective in promoting axial osmylation as is the axial methoxy
analog (entry 11). Since the lone pair density of the tetrahy-
drofuran 19a is directed toward the cyclohexane ring, not toward
the incoming osmium reagent, this is the expected result.
As already mentioned, axial OH appears to exert a Henbest
effect, presumably involving transition state stabilization via
hydrogen bonding. Other examples of hydroxyl directing effects
in osmylations are known.33 However, a comparison of entries
1 and 3 or 1 and 15 shows that hydroxylation syn to axial C2-
OH is less favored than attack syn to axial C2-H. Thus, the
Henbest effect is opposed by the lone pair repulsion or steric
effects of hydroxyl. In other respects, the osmylations resemble
the epoxidations in that all three reagents respond to the
unsymmetrical rotor groups. The selectivity pattern of the axial
acetate and methoxy derivatives could be rationalized by
invoking a small Cieplak effect, but the behavior of the
Since the torsional barrier is the difference between a
composite of steric and steroelectronic factors,12 it is not easy
to evaluate this energy term by intuition nor to dissect it into
its component parts. Nevertheless, the concept has a clear
experimental basis as long as the issue is torsion among existing
bonds. On the other hand, the extension of this concept to
developing bonds encounters formidable difficulties. In the
ground state, the torsional effect is defined by the measurement
of a torsional barrier, an energy that usually represents the
difference in free energies of the eclipsed and the staggered
geometries of the same substance. The torsional contribution
to ground state energy can be represented by parameters in
empirical force fields without having to know exactly what
causes torsional barriers. For simple molecules such as ethane,
the absolute energies can also be computed by ab initio methods
with sufficient accuracy to reproduce the torsional barrier and
to separate these energies into recognizable components.12,13
Is there an analogous way to dissect transition state torsional
energies? Where ground state (existing bond) torsion represents
the difference in energies of eclipsed and staggered rotamers,
the transition state analog would need to evaluate the free energy
of diastereomers and conformers of the transition structures.
To do this by computation, it would be necessary to locate
different saddle points on the energy surface, each of which
would need to be evaluated in terms of total energy and dissected
into components. This does not appear to be a realistic prospect
if transition state torsional effects are to be described by the
same component variables that define the ground state version.
It is already difficult to calculate absolute values for ground
state torsion for non-trivial molecules, and the transition state
analog must deal with additional variations in bond lengths, bond
angles, hyperconjugative contributions, and so on. Only the
filled orbital repulsive component of developing bond torsion
appears to be accessible to intuition. Fundamentally, this would
be no different than to invoke a qualitative overall steric effect.
In the region of the reaction coordinate where internuclear
distances are ca. 2-3 Å, this component of transition state
energy will likely increase with the bulk of interacting groups
and with decreasing distance between them. For reactions with
relatively late transition states and shorter (extensively rehy-
bridized) bonds, “developing bond” torsion might be identified
as a distinct energy term because the new bond would be mostly
formed and would be subject to the factors that contribute to
(32) (a) Kishi, Y.; Christ, W. J.; Cha, J. K. Tetrahedron 1984, 40, 2247.
(b) Review: Cha, J. K.; Kim, N.-S. Chem. ReV. 1995, 95, 1761.
(33) Kon, K.; Isoe, S. Tetrahedron Lett. 1980, 21, 3399. Smith, A. B.;
Boschelli, D. J. Org. Chem. 1983, 48, 1217. Xu, D.; Park, C. Y.; Sharpless,
K. B. Tetrahedron Lett. 1994, 35, 2495.

4-tert-Butylmethylenecyclohexane Epoxidations
J. Am. Chem. Soc., Vol. 118, No. 15, 1996 3563
equatorial analogs (entries 4 and 5) is more compatible with
the lone pair repulsion argument. Entries 9, 10, 11, and 13
show a preference for the non-Cieplak product. Steric interac-
tions between the substrate and the developing bonds, together
with the Kishi effect, play the most important role while
contributions from the Cieplak effect and from existing bond
torsion are too small to clearly identify.
for bond rotation and indicates a relatively nonpolar transition
state. There must be a strong electrostatic interaction or a weak
bond between the partially negative peracid fragment and the
partially positive olefinic carbon, presumably expressed via the
electron pairs at the peroxidic oxygen.
Asynchronous Bonding and Torsion in Cyclohexene
Epoxidations
Our results do not help to clarify the mechanistic picture for
the osmylation reactions.34 Either the 2 + 3 cycloaddition or
the 2 + 2 cycloaddition pathways are consistent with substantial
steric effects, lone pair avoidance by the reagent, and sensitivity
to unsymmetrical rotor groups. In the context of methylene
cyclohexane facial selectivity, the two pathways are not much
different. If the rate-determining event is the ligand-promoted
conversion of a 2 + 2 adduct into the cyclic osmate ester,34a
then the stereochemistry-determining transition state could have
(but need not have) a fully formed C-O bond at the primary
carbon (Cexo), and partially formed C1-Os and C1-O bonds.
If the mechanism involves some variation of the 2 + 3
cycloaddition process,34b then both osmate C-O bonds would
be partially formed. Asynchronous bonding with the unsym-
metrical methylenecyclohexane derivatives is expected in any
case, and the difference in transition state geometries between
the two mechanisms is too subtle for any comment based on
our data.
Jerina et al. have reported that the cyclic styrene 28 reacts
with MCPBA to give a single epoxide 29 (NMR assay; 92%
isolated).37 Related examples have been studied by Martinelli
et al. and the epoxidation selectivities are accurately known.4a
Thus, 30 produces 31 with remarkable 99:1 selectivity (MCPBA
conditions). Several analogous dihydronaphthalene derivatives
afford epoxides with selectivities ranging from 85:15 to 99:1,
but an indene analog of 30 (5-methyl-1,2-benzocyclopentadiene)
gives a 1:1 mixture of epoxides. However, cyclohexenes 32
and 33 react nonselectively with p-nitroperbenzoic acid (45:55
and 50:50, respectively).38 Martinelli, Houk et al. invoked
interactions between developing C- - -O bonds and adjacent axial
C-H bonds to explain these contrasts as follows.4a Formation
of the eclipsed bond in a half-chair 34 might be destabilized by
a developing bond interaction relative to the staggered bond.
Such an effect would be magnified in an asynchronous transition
state with relatively advanced bonding at the styrene â-carbon
because the peracid subunit would be closer to the axial allylic
C-H bond.4a In cyclohexenes 32 and 33, containing two allylic
axial C-H bonds, there is an equal number of interactions for
bonding at either face of the double bond, similar to the model
35, and there would be no selectivity if developing bond
interactions are dominant.4a
The cyclohexene examples differ from the methylene cyclo-
hexane system discussed earlier in at least one important way.
In the exocyclic alkenes, the geometric consequences of an
asynchronous transition state place the most highly developed
C-O bond far from the sp3-hybridized allylic ring carbons (C2
and C6, methylenecyclohexane numbering). The situation is
reversed for the dihydronaphathalenes 28 or 30 where bond
formation occurs near the allylic C-H bonds. Depending on
the extent of rehybridization, interactions between the develop-
ing C- - -O bond and the axial allylic hydrogen could contribute
to the stability difference between the competing transition states
as proposed by Martinelli, Houk, et al.4a However, a transition
state analysis based primarily on existing bond interactions can
also account for the cyclohexene examples.
Simple cyclohexenes (for example, 32) may react with
peracids via nearly synchronous transition states with both
olefinic carbons partially rehybridized. The six-membered ring
resembles a half-chair cyclohexene regardless of which face of
the double bond is attacked and the minimal steric requirement
of the peracid results in a nonselective reaction. In contrast,
structures such as 30 that contain a dihydronaphthalene subunit
prefer highly asynchronous transition states with substantial
benzylic cation character due to the extended delocalization
enforced by ring constraints. Three families of asynchronous
transition structures can be considered, represented by the
limiting cases 36, 37, and 38. Geometries 36 and 37 have the
more developed Câ- - -O σ bond in the plane of the benzylic
p-orbital to maximize the stabilizing interaction between the
peroxidic oxygen electron pairs and the benzylic π-system.
Staggered geometries similar to 38 have a nearly orthogonal
Asynchronous Bonding in Methylenecyclohexane
Epoxidations
Peracid epoxidation mechanisms are well-understood.1,9a,b,35,36
Asynchronous bonding is expected by analogy to the styrene
epoxidations,30 and the Bartlett butterfly arrangement as modi-
fied by Beak et al. appears secure.36a The competing transition
states in the methylenecyclohexane series can therefore be
represented by structures 26 and 27. The preferred geometry
26 is consistent with relatively small developing bond steric
interactions, and also with the observation that epoxidation
selectivity is opposite to the axial/equatorial preference of the
product epoxides.8 There is relatively little rehybridization at
ring carbon C1, and the magnitude of 1,3-diaxial interactions is
therefore small. Sevin and Cense proposed a similar transition
state geometry.8 They also investigated solvent effects on
diastereoselectivity in the reaction of 4-tert-butylmethylenecy-
clohexene with p-nitroperbenzoic acid. Only small differences
in percent axial epoxidation were found: dichloromethane, 70%;
methanol, 73%; ether, 80%. Similar axial selectivities were
observed in the present study with m-chloroperbenzoic acid:
CDCl3, 72%; methanol, 75%; benzene, 78%; acetonitrile, 78%;
THF, 80%. Thus, the competing transition states do not differ
much in terms of charge separation or dipole moment. The
relative insensitivity to solvent effects suggests that the peroxide
oxygen electron pairs provide effective internal stabilization for
the partial positive charge at C1, as also deduced in earlier
mechanistic studies.9a Hanzlik and Shearer found that rehy-
bridization at the benzylic carbon in styrene epoxidations is too
small to detect from the 13C kinetic isotope effect.30 However,
peracid epoxidations of â-substituted styrenes are stereospecific
and the reactions tend to be fastest in nonpolar solvents.9a,30
This rules out ionic intermediates having a lifetime sufficient
(34) (a) Kolb, H. C.; Andersson, P. G.; Sharpless, K. B. J. Am. Chem.
Soc. 1994, 116, 1278. Veldkamp, A.; Frenking, G. J. Am. Chem. Soc. 1994,
116, 4937. (b) Houk, K. N.; Wu, Y.-D.; Wang, Y. J. Org. Chem. 1992,
57, 1362. Corey, E. J.; Noe, M. C. J. Am. Chem. Soc. 1993, 115, 12579.
(35) Review: Schroder, M. Chem. ReV. 1980, 80, 187.
(36) (a) Beak, P.; Woods, K. W. J. Am. Chem. Soc. 1991, 113, 6281.
(b) Rebek, J., Jr.; Marshall, L.; McManis, J.; Wolak, R. J. Org. Chem.
1986, 51, 1649.
(37) Sayer, J. M.; Yagi, H.; Silverton, J. V.; Friedman, S. L.; Whalen,
D. L.; Jerina, D. M. J. Am. Chem. Soc. 1982, 104, 1972.
(38) Casadevall, A.; Casadevall, E.; Mion, M. Bull. Soc. Chim. Fr. 1968,
4498. Berti, G.; Macchia, B.; Macchia, F. Tetrahedron 1968, 24, 1755.

3564 J. Am. Chem. Soc., Vol. 118, No. 15, 1996
Vedejs et al.
σ,p orbital arrangement and can be discounted on stereoelec-
tronic grounds. The transition structure 37 allows better
proximity of oxygen electron pairs and the benzylic orbital, but
this comes at the cost of a geometry that resembles a half-boat
cyclohexene and that also twists the benzylic p-orbital relative
to the benzene orbitals. The alternative envelope structure 36
allows ideal overlap between the aromatic ring and the benzylic
p-orbital and also maintains close proximity between peroxidic
electron pairs and the partially positive benzylic carbon. There
is better staggering of all of the sp3 or partly sp3 bonds (not
only of the developing bonds) compared to the situation in 37.
Some fraction of the 5-6 kcal/mol energy difference between
half-boat and half-chair cyclohexenes39 should stabilize 36
relative to 37.40 The observed ∆∆G* values for the epoxidation
of 28, 30, and related dihydronaphthalenes (1.5-2.7 kcal/mol)
are within the available energy range.
Chart 2
The above discussion is not intended to rule out a contribution
by developing bond steric effects in the epoxidations of 28 or
30. The indicated transition state geometries are not funda-
mentally different from those considered by Martinelli, Houk,
et al., and there are no substantial differences in the interpreta-
tions if the transition state is relatively advanced. Extensive
rehybridization at the styrene â-carbon would surely encounter
a torsional contribution to ∆∆G* because the environment at
Câ and Cγ would have to select between staggered and eclipsed
geometries of sp3 or sp3-like bonds. Differences in the rationales
are more substantial for reactions having early transition states,
and for this scenario it is expected that existing bond torsional
factors will dominate over developing bond steric effects.
discussion of elusive phenomena must await the discovery of
substrates where the effects are dominant over the torsional,
hyperconjugative, or FMO factors that have been featured in
existing explanations.3,4,6,8
Summary
The larger effects in the methylenecyclohexane epoxidations
of Table 2 arise from the steric influence of unsymmetrical rotor
substituents. A similar trend has been observed for Diels-Alder
reactions where allylic methoxy is large compared to methyl,41
and in addition reactions of R-methoxyacetaldehyde.6b In the
absence of unsymmetrical rotors, there is a small, but consistent
trend for axial epoxidation that can be understood by invoking
torsional effects among existing bonds in the asynchronous
equatorial epoxidation transition state.8 We cannot say that these
factors are responsible for all of the axial preference of the parent
4-tert-butylmethylenecyclohexane,6 but they are in the correct
range of energies. Small σ-σ* effects may be responsible for
the trends observed with some of the 3-substituted examples
(Table 1, entries 14-18), but such effects are minimal in the
examples of Table 2 where isosteric substituents can be
compared. The 3-silyl- or 3-stannylmethylenecyclohexanes are
epoxidized nonselectively (Table 1, entries 14 and 15), corre-
sponding to ∆∆G* ) ca. 0.5 kcal/mol toward equatorial
epoxidation compared to the parent alkene. The other (equa-
torially) 3-substituted examples of Table 1 differ so little from
the parent that they require no further rationale. The effect of
silicon or tin substituents at C3 could be hyperconjugative, but
alternative explanations may need to be considered. In view
of the long history of the methylenecyclohexene problem, further
Ground state torsional interactions have complicated origins,
but they are defined by experimental data (rotational barriers)
and the torsional energies of existing bonds can often be
extrapolated to transition structures if the extent of rehybrid-
ization can be estimated. There is currently no way to estimate
the torsional effects for deVeloping bonds, short of including
all possible 1,2-interactions (all permutations of the filled,
unfilled, and partially filled orbital 1,2-interactions) in competing
transition states. This approach appears to be too unwieldy for
the evaluation of transition states. However, qualitative com-
parisons are possible for relatively late transition states where
the developing bonds resemble product bonds in hybridization.
Attempts to separate steric (filled orbital) factors and electronic
(including unfilled and partially filled orbital interactions) factors
have been helpful,3,18 but our data suggest that a distinct
combination of variables is needed for each specific reagent-
substrate family. Disappointingly, the torsional effects, steric
effects, rotor effects, Cieplak effects, FMO effects, electrostatic
effects, etc., do not cleanly “separate” by definition, and they
do not work in the same way in a sufficiently large number of
systems to justify their extrapolation from one system to another
without detailed study.
Note Added in Proof: Yamabe et al. have recently reported
(39) Anet, F. A. L.; Freedberg, D. I.; Storer, J. N.; Houk, K. N. J. Am.
Chem. Soc. 1992, 114, 10969. See also: Conformational Analysis of
Cyclohexenes, Cyclohexadienes, and Related Hydroaromatic Compounds;
VCH, New York, 1989; pp 1-45.
(40) The non-selective epoxidation of 33 can be explained by arguing
that the unconstrained phenyl substituent is not effective in providing
benzylic delocalization due to steric interactions between the benzene ring
and equatorial C-H bonds at C2 and C6 in the necessary rotamer. This
would result in relatively synchronous half-chair epoxidation transition states
for attack from either olefin face.
that an asynchronous transition state for the CH2dCH2
+
HCO3H reaction has lower energy than the symmetrical
transition structure found by Bach et al. using similar ab initio
methods.4b The former authors also found an asynchronous
transition structure for the reaction of propene + HCO3H as
shown below (Yamabe, S.; Kondou, C.; Minato, T. J. Org.
Chem. 1996, 61, 616). The C-C-CH3 geometry resembles
the torsional minimum of propene (methyl C-H eclipsed with
nearly planar C2), and is consistent with the dihedral angles
(41) Datta, S. C.; Franck, R. W.; Tripathy, R.; Quigley, G. J.; Huang,
L.; Chen, S.; Sihaed, A. J. Am. Chem. Soc. 1990, 112, 8472.

4-tert-Butylmethylenecyclohexane Epoxidations
J. Am. Chem. Soc., Vol. 118, No. 15, 1996 3565
Hz). The complementary mixture of diastereomers was prepared as
follows: To a stirred solution or 1-(trimethylsiloxy)-2-ethyl-2-tert-butyl-
1-cyclohexene (17) (2.09 g, 8.23 mmol) in THF at 0 °C was added
MeLi (9.0 mL of a 1.36 M solution, 12.3 mmol) via syringe and the
reaction was stirred for 45 min. The reaction was cooled to -40 °C,
iodomethane (0.62 mL, 9.9 mmol) was added via syringe, and the
reaction was stirred at -40 °C for an additional 7 h and then allowed
to warm to room temperature overnight. The same workup and
purification was employed as was described above. This procedure
yielded 304 mg (19% yield) consisting of a 3:1 ratio of axial methyl
(14b) to equatorial methyl (14a) diastereomers. The same NMR signals
were observed, but the integral ratios were inverted compared to those
described above.
General Procedure for Wittig Methylenation. To a thick suspen-
sion of methyltriphenylphosphonium bromide (6 equiv) and potassium
tert-butoxide (6 equiv) in the minimum amount of THF at room
temperature was added the ketone in minimum THF and the reaction
was stirred at room temperature until TLC analysis indicated complete
conversion to olefin. Following the Fitjer method,42 hindered ketone
reactions were heated (reflux) and the amount of solvent was reduced
in order to increase the ylide concentration. All reactions were
quenched onto aqueous NH4Cl and extracted with Et2O, and the
combined organic layers were washed with brine, dried (MgSO4), and
concentrated (aspirator) to yield crude products. Pure materials were
obtained by flash chromatography or by preparative TLC using the
solvent system recorded for individual examples.
suggested for the peracid atoms relative to the alkene in 36 or
37 and with asynchronous bonding in methylenecyclohexane
epoxidations (see 26 and 27).
Experimental Section
Ketones. The dimethylhydrazone anion technique of Corey et al.21
was used to prepare 2-methyl-4-tert-butylmethylenecyclohexanones 5a
and 5b,22a the 2-ethyl analogs 16,b,22b and the 2-(3-tert-butyldimeth-
ylsiloxy)propyl analogs 17a,b (inseparable mixture).
2-Hydroxy-2-methyl-4-tert-butylcyclohexanones (9a and 9b). To
1-(trimethylsiloxy)-2-methyl-4-tert-butyl-1-cyclohexene23,24b (1.57g, 6.5
mmol) was added dimethyldioxirane solution in acetone (70 mL of a
0.1 M solution, 7.0 mmol)25 at 0 °C and the reaction was stirred at 0
°C for 3 h, dried (MgSO4), and concentrated (aspirator) to yield a yellow
oil. Isomers were separated by flash chromatography on silica gel (80
mm column, 15% EtOAc/hexane). The equatorial alcohol 9b eluted
first. Pure material was obtained by crystallization from 4:1:1 hexane-
Et2O-CH2Cl2. This product 673 mg (56% yield) of white crystals of
9b: mp 47-48 °C; analytical TLC on silica gel, 15% EtOAc/hexane,
Rf ) 0.27. Molecular ion calcd for C11H20C2: 184.14630; found m/e
Alkenes. The following alkenes have been described previously:
4-tert-butylmethylenecyclohexane (1b),7 cis-2-methyl-4-tert-butyl-
methylenecyclohexane (6a),8 and 2-hydroxy-4-tert-butylmethylenecy-
clohexanes 12a and 12b.43
184.1463. Error ) 0 ppm. Base peak ) 127 amu; IR (neat, cm-1
)
trans-2-Methyl-4-tert-butylmethylenecyclohexane (6b). To a stirred
solution of the trans-2-methyl-4-tert-butylcyclohexanone22 (52.4 mg,
0.31 mmol) in CH2Cl2 at 0 °C were added 10-mL aliquots of the
preformed Lombardo reagent44 via a large diameter cannula at 0 °C
until the reaction was complete by TLC analysis. After each addition
the reaction was sonicated for 1 h. The reaction was then quenched
with 2:1 saturated NaHCO3/H2O (2 × 30 mL), dried (Na2SO4), and
filtered through a plug of coarse silica gel and then the solvent was
removed (aspirator). The residue was purified by flash chromatography,
(10 mm column: 10% EtOAc/hexane) to give 35.7 mg (68% yield) of
a clear, colorless oil: analytical TLC on silica gel, 10% EtOAc/hexane,
Rf ) 0.81. 200-MNz NMR (CDCl3, ppm) δ 4.62-4.6 (1 H, m), 4.54
(1 H, t, J ) 2.1 Hz), 2.7-2.5 (1 H, m), 2.31-2.08 (2 H, m), 1.9-1.75
(1 H, m), 1.67-1.55 (1 H, dq, J ) 11.6, 1.8 Hz), 1.48-1.12 (2 H, m),
1.08 (3 H, d, J ) 7.2 Hz), 1.04-0.89 (1 H, m), 0.84 (9 H, s).
2-Hydroxy-2-methyl-4-tert-butylmethylenecyclohexane (10a). The
standard Wittig procedure from 9a (228 mg, 1.24 mmol), 24 h reflux,
gave 221 mg (96% yield) of a clear, colorless oil: analytical TLC on
silica gel, 1:6 EtOAc/hexane, Rf ) 0.22. Molecular ion calcd for
C12H22O: 182.16705; found m/e 182.1671. Error ) 0 ppm. Base peak
) 149 amu. IR (neat, cm-1) 3594, O-H; 3077, dC-H. 200-MHz
NMR (CDCl3, ppm) δ 4.85 (1 H, s), 4.74 (1 H, t, J ) 1.6 Hz), 2.47-
2.39 (1 H, m), 2.27-2.17 (1 H, m), 1.92-1.83 (2 H, m), 1.71-1.55 (1
H, m), 1.41 (3 H, s), 1.28-0.88 (3 H, m), 0.86 (9 H, s). 13C NMR
(500 MHz, CDCl3, ppm) δ 153.0, 106.6, 71.9, 42.8, 42.0, 32.8, 32.2,
29.0, 28.3, 27.8.
2-Hydroxy-2-methyl-4-tert-butylmethylenecyclohexane (10b). The
standard Wittig procedure from 9b (140 mg, 0.76 mmol), 24 h at reflux,
gave 128 mg (93%) of white crystals: analytical TLC on silica gel,
1:6 EtOAc/hexane, Rf ) 0.17. Pure materials (451 mg, 80% yield)
was obtained by crystallization from hexane, mp 78-79 °C. Molecular
ion calcd for C12H22O: 182.16705; found m/e 182.1671. Error ) 0
ppm. Base peak ) 125 amu. IR (neat, cm-1) 3594, O-H; 1646, CdC;
200 MHz NMR (CDCl3, ppm) δ 4.97 (1 H, dd, J ) 1.7, 1.7 Hz), 4.71
(1 H, dd, J ) 1.7, 1.7 Hz), 2.37-2.34 (1 H, m), 2.22-2.02 (1 H, m),
1.93-1.79 (2 H, m), 1.52 (1 H, s), 1.33 (3 H, s), 1.29-0.95 (3 H, m),
0.86 (9 H, s).
1710, CdO; 3497, O-H. 200-MHz NMR (CDCl3, ppm) δ 3.94 (1 H,
s), 2.65-2.48 (2 H, m), 2.22-2.08 (3 H, m), 1.63-1.42 (2 H, m),
1.41 (3 H, s), 0.92 (9 H, s).
Later fractions produced 9a. Pure material was obtained by
crystallization from 4:1:1 hexane/Et2O/CH2Cl2: 350 mg white crystals
(29% yield); mp 70-71 °C; analytical TLC on silica gel, 15% EtOAc/
hexane; Rf ) 0.20. Molecular ion calcd for C11H20O2: 184.14630;
found m/e 184.1463. Error ) 0 ppm. Base peak ) 127.07 amu. IR
(neat, cm-1) 3594, O-H; 1709, CdO. 200-MHz NMR (CDCl3, ppm)
δ 2.91-2.74 (1 H, m), 2.41-2.29 (1 H, m), 2.12-1.99 (2 H, m), 1.81-
1.72 (1 H, m), 1.54-1.38 (2 H, m), 1.31 (3 H, s), 0.90 (9 H, s).
2-Ethyl-2-methyl-4-tert-butylcyclohexanones (14a and 14b). The
enol silanes 7 and 17 were prepared from 2-methyl-4-tert-butylcyclo-
hexanones 5a,b22a and 2-ethyl-4-tert-butylcyclohexanones 16a,b22b using
the procedure of Miller and McKean.24b In the preparation of 17, the
reaction was performed using 3.06 g (16.8 mmol) of ketone 16a,b to
give 3.52 g (38% yield) of product as a pale oil after distillation (bulb-
to-bulb; 70 °C pot temperature; -78 °C receiving flask temperature;
0.5 mmHg). The distilled product 17 was assayed by NMR, which
indicated a >95:<5 ratio of thermodynamic to kinetic isomers.
To a stirred solution of 1-(trimethylsiloxy)-2-methyl-4-tert-butyl-1-
cyclohexene (7) (82 mg, 0.34 mmol) in 3 mL of THF at 0 °C was
added MeLi (0.45 mL of a 1.36 M solution, 0.61 mmol) via syringe
and the mixture was stirred for 45 min. Iodoethane (0.052 mL, 0.65
mmol) was then added via syringe and the reaction was stirred an
additional 7 h at 0 °C before being warmed to room temperature. The
reaction was poured onto 10 mL of 1 M NaHCO3, the layers were
separated, and the aqueous layer was extracted with Et2O (3 × 10 mL).
The combined organic layers were dried (MgSO4) and concentrated
(aspirator) to yield an oil composed of a 15:5:1 ratio of axial alkylated
product to equatorial alkylated product to 2-ethyl-4-tert-butylcyclo-
hexanone according to 1H NMR assay. Also present were polyalkylated
products. The undesired byproducts were removed via preparative plate
chromatography (20 cm × 20 cm × 0.01 cm; 1:19 EtOAc/hexane, two
developments) to yield 16 mg (24% yield) of a clear, colorless oil
composed of an inseparable 3:1 mixture of isomers 14a-14b: analytical
TLC on silica gel, 1:19 EtOAc/hexane, Rf ) 0.21. Molecular ion calcd
for C13H24O: 196.18271; found m/e ) 196.1830. Error ) 1 ppm. IR
(neat, cm-1) 1707, CdO. 200 MHz NMR (CDCl3, ppm) δ 2.60-2.42
(1 H, m), 2.34-2.23 (1 H, m), 2.08-1.95 (1 H, m), 1.89-1.20 (3 H,
m), 1.88-1.26 (3 H, m), 1.12 (0.6 H, s), 0.91 (1.8 H, s), 0.90 (7.2 H,
s), 0.98 (2.4 H, s), 0.84 (0.6 H, t, J ) 7.5 Hz), 0.78 (2.4 H, t, J ) 7.5
2-Methoxy-2-methyl-4-tert-butylmethylenecyclohexane (11b). To
a stirred heterogeneous solution of NaH (744 mg, 31.0 mmol, washed
(42) Fitjer, L.; Quabock, U. Synth. Commun. 1985, 15, 855.
(43) Cross, B.; Whitman, G. H. J. Chem. Soc. 1961, 1650.
(44) Lombardo, L. Org. Synth. 1987, 65, 81.

3566 J. Am. Chem. Soc., Vol. 118, No. 15, 1996
Vedejs et al.
with Et2O and dried by N2) in 35 mL of THF was added 10b (112 mg,
0.61 mmol) in 10 mL of THF and the mixture was refluxed for 3 h.
The mixture was then cooled to room temperature, iodomethane (2.1
mL, 34.0 mmol) was added, and the reaction was stirred for an
additional 2 h. Saturated NH4Cl was added SLOWLY until bubbling
stopped. After separation of layers the aqueous layer was extracted
with Et2O (3 × 60 mL), and the combined organic layers were dried
(MgSO4) and concentrted (aspirator). Bulb-to-bulb distillation (pot
temperature 92 °C; collection vessel temperature -78 °C; 0.5 mmHg)
yielded 110 mg (92% yield) of a clear, colorless chromatographically
pure oil, 11b: analytical TLC on silica gel, 1:19 EtOAc/hexane, Rf )
0.35. Molecular ion calcd for C13H24O: 196.18271; found m/e
EtOAc/hexane, Rf ) 0.30. MS, base peak ) 168. Exact mass calcd
for C13H22O2 210.162; found 210.1627. Error ) 3.6 ppm. IR (neat,
cm-1): C-H, 2970; CdO, 1750; C-O, 1230. 200-MHz NMR (CDCl3)
δ 5.20-5.10 (1H, m), 4.70 (2H, t, J ) 1.6 Hz), 2.42 (1H, dt, J ) 13.4,
2.8 Hz), 2.10 (3H, s), 2.10-2.00 (1H, m), 1.90-1.70 (1H, m), 1.42-
0.95 (4H, m), 0.83 (9H, s).
2-Ethyl-2-methyl-4-tert-butylmethylenecyclohexane (15a,b). The
standard Wittig procedure from a 3:1 mixture of 14b-14a (170 mg,
0.87 mmole) was used, 24-h reflux. This produced 159 mg (94% yield)
of a clear, colorless oil after chromatography, inseparable 3:1 mixture
of isomers 15b:15a: analytical TLC on silica gel, hexane, Rf ) 0.69.
Molecular ion calcd for C14H26 194.20344; found m/e 194.2017. Error
) 9 ppm. Base peak ) 109 amu. IR (neat, cm-1) 3083, dC-H. 200-
MHz NMR (CDCl3, ppm) δ 4.70 (0.7 H, s), 4.60 (0.3 H, s), 4.59-
4.55 (1 H, m), 2.34-2.1 (2 H, m), 1.9-0.89 (7.9 H, m), 1.0 (0.9 H, s),
0.99 (2.1 H, s), 0.84 (9 H, s), 0.70 (2.1 H, t, J ) 7.4 Hz). The
complementary mixture was prepared similarly from the 3:1 mixture
of 14a-14b.
196.1826. Error ) 1 ppm. Base peak ) 149 amu. IR (neat, cm-1
)
2950, C-H; 1150. 200-MHz NMR (CDCl3, ppm) δ 4.9 (1 H, dd, J )
2.1, 2.1 Hz), 4.73 (1 H, dd, J ) 2.1, 2.1 Hz), 3.3 (3 H, s), 2.34 (1 H,
ddd, J ) 14.0, 4.4, 2.8 Hz), 2.2-2.0 (1 H, m), 1.9-1.7 (2 H, m), 1.4-
0.9 (3 H, m), 1.29 (3 H, s), 0.85 (9 H, s). 13C NMR (500 MHz, CDCl3,
ppm) δ 150.95, 105.93, 49.48, 45.42, 37.8, 33.5, 32.0, 29.7, 28.6, 27.6,
24.6, 22.7, 14.1.
Tetrahydrofuran Derivatives 19a and 19b. The silyl enol ether
7c was prepared from a mixture of 16a,b using the TMSCl/NaI/HN-
[SiMe3]2 procedure,24b >9:1 isomer ratio according to NMR assay. To
the crude 7c (200 mg; 0.51 mmol) in hexane (4 mL) at 0 °C was added
mCPBA (0.12 g; 0.571 mmol). After 3 h, the suspension was filtered,
the hexane filtrate was evaporated, and the crude residue was taken up
in 5 mL of CH2Cl2. Triethylamine hydrofluoride (93 mg; 1.02 mmol)
was added dropwise. The reaction was stirred for 2 h and diluted with
CH2Cl2. The mixture was washed successively with aqueous NaHCO3,
1 N HCl, and saturated NaHCO3, dried (MgSO4), and concentrated
under reduced pressure. Separation by flash chromatography (5%
EtOAc/hexane) gave the axial and equatorial hydroxy ketones 18a (59
mg) and 18b (23 mg), 47% yield, as clear oils.
2-Methoxy-2-methyl-4-tert-butylmethylenecyclohexane (11a). The
same procedure as for 11b was used starting with 285 mg (1.43 mmol)
of 10a. The residue was purified using basic Al2O3 (hexane) to yield
239 mg (80% yield) of 11a: colorless oil; analytical TLC on silica
gel, 1:19 EtOAc/hexane, Rf ) 0.4. Molecular ion calcd for C13H24O:
196.18271; found m/e 196.1827. Error ) 8 ppm. Base peak ) 107
amu. IR (neat, cm-1) 1640, CdC; 1070, C-O; 200-MHz NMR
(CDCl3, ppm) δ 4.91 (1 H, s), 4.8 (1 H, s), 3.03 (3 H, s), 2.2-2.1 (2
H, m), 2.0-1.8 (2 H, m), 1.65 (1 h, dddd, J ) 12.3, 12.3, 3.4, 3.4 Hz),
1.2-0.9 (2 H, m), 1.22 (3 H, s), 0.81 (9 H, s).
2-Acetoxy-2-methyl-4-tert-Butylmethylenecyclohexane (11c). The
title compound was prepared by treatment of alcohol 10a with excess
acetic anhydride/(dimethylamino)pyridine; analytical TLC (silica gel
F254), 5% EtOAc/hexane, Rf ) 0.30. MS. Base peak ) 107. Exact
mass calcd for C14H24O2 224.1776; found 224.1773. Error ) 1.4 ppm.
IR (neat, cm-1): C-H, 2970; CdO, 1750; C-O, 1250. 200-MHz
NMR (CDCl3) δ 4.92 (1H, s), 4.86 (1H, s), 2.34 (1H, dt, J ) 13.4, 3.8
Hz), 2.23-2.13 (2H, m), 1.99 (3H, s), 1.90-1.80 (1H, m), 1.63 (3H,
s), 1.47 (1H, dt, J ) 12.4, 3.3 Hz), 1.24-0.86 (2H, m), 0.84 (9H, s).
trans-2-Methoxy-4-tert-butylmethylenecyclohexane (13a). To a
solution containing cis-2-hydroxy-4-tert-butylmethylenecyclohexane
(12a)43 (500 mg; 2.98 mmol) and methyl iodide (0.26 mL; 4.12 mmol)
in 5 mL of anydrous DME was added 78 mg (3.23 mmol) of sodium
hydride in four portions over 15 min. After the heat had evolved an
additional 0.1 mL of methyl iodide was added and the reaction was
stirred at room temperature for 2 h. The solvent was removed under
reduced pressure and the crude residue was taken up in ether. The
sodium iodide was removed by filtration and the filtrate washed with
ether. The combined ether layers were concentrated under reduced
pressure to yield the crude ether (503 mg, 92%): oil, analytical TLC
(silica gel F254), 1% ether/hexane, Rf ) 0.1. MS, base peak ) 93.
Exact mass calcd for C12H22O 182.1671; found 182.1678. Error )
4.1 ppm. IR (neat, cm-1): C-H, 2950; CdC, 1650; C-O, 1000. 200-
MHz NMR (CDCl3) δ 4.85-4.80 (1H, m); 4.80-4.75 (1H, m); 3.71
(1H, t, J ) 2.9 Hz); 3.20 (3H, s); 2.30-2.10 (2H, m); 2.06 (1H, dq, J
) 13.4, 2.9 Hz); 1.90-1.70 (1H, m); 1.58 (1H, tt, J ) 12.5, 3.1 Hz);
1.24 (1H, dt, J ) 13.0, 2.9 Hz); 1.03 (1H, dd, J ) 12.3, 5.7 Hz); 0.83
(9H, s).
cis-2-Methoxy-4-tert-butylmethylenecyclohexane (13b). The same
procedure as described for 13a was used, starting from 12b:43 oil,
analytical TLC (silica gel F254), 1% EtOAc/hexane, Rf ) 0.1. MS,
base peak ) 125. Exact mass calcd for C12H22O 182.1671; found
182.1673. Error ) 1.3 ppm. IR (neat, cm-1): C-H, 2950; CdC,
1650; C-O, 1110. 200-MHz NMR (CDCl3) δ 4.90-4.85 (1H, m),
4.75-4.70 (1H, m), 3.55-3.50 (1H, m), 3.44 (3H, s), 2.50-2.30 (1H,
m), 2.30-2.10 (1H, m), 2.00-1.70 (2H, m), 1.26 (1H, tt, J ) 12.0,
2.9 Hz), 1.10-0.90 (2H, m), 0.85 (9H, s).
trans-2-Acetoxy-4-tert-butylmethylenecyclohexane (13c). The title
compound was prepared from 12a43 (0.48 g), excess acetic anhydride
(1.1 mL), p-(dimethylamino)pyridine (0.1 g), and triethylamine (2.4
mL) in CH2Cl2 (30 mL), 3 h at room temperature. After routine
aqueous workup, the product was purified by flash chromatography to
afford 0.54 g 13c (98%): oil, analytical TLC (silica gel F254), 5%
Each isomer was subjected to a modified Wittig procedure. Thus,
to a stirred solution containing methyltriphenylphosphonium bromide
(0.33 mmol; 0.12 g) in 2 mL of THF was added 0.83 mL (0.33 mmol)
of a 0.4 M solution of potassium tert-butoxide in THF dropwise. The
reaction was stirred for 10 min and a solution containing 18a (56 mg;
0.16 mmol) in 2 mL of THF was added dropwise. The reaction was
heated at 60 °C for 1 h. The cooled reaction was quenched by addition
of saturated NH4Cl and the mixture was extracted with ether. The
ethereal layer was washed with brine, dried over MgSO4, and evaporated
to give the crude allylic alcohol which was purified via flash
chromatography (5% EtOAc/hexane) to given an oil (45 mg; 82%).
To a stirred solution containing the above allylic alcohol (29 mg;
0.09 mmol) in 2 mL of THF was added 0.19 mL (0.19 mmol) of a 1.0
M solution of tetrabutylammonium fluoride in THF. The reaction was
stirred for 10 min, followed by addition of brine and extraction with
ether. The ether layer was washed with water and brine, dried over
MgSO4, and evaporated. Purification via flash chromatography (5%
MeOH/CHCl3) afforded the expected diol (17 mg; 84%) as an oil.
To a stirred solution containing diethyldiazodicarboxylate (0.012 mL;
0.075 mmol) in 1 mL of Et2O was added a solution containing the
above diol (17 mg, 0.075 mmol) and triphenylphosphine (0.075 mmol;
20 mg) in 2 mL of Et2O dropwise over 1 h. The reaction was then
stirred at room temperature for 12 h. The solvent was removed under
reduced pressure and the residue was subjected to flash chromatography
(5% EtOAc/hexane) to afford 19a (13 mg; 82%): oil, analytical TLC
(silica gel F254), 5% EtOAc/hexane, Rf ) 0.35. MS, base peak )
151. Exact mass calcd for C14H24O 208.1827; found 208.1831. Error
) 1.9 ppm. IR (neat, cm-1): C-H, 2920; CdC, 1630; C-O, 1120.
200-MHz NMR (CDCl3) δ 4.76 (1H, s), 4.71 (1H, s), 3.83 (1H, q, J )
7.5 Hz), 3.66 (1H, q, J ) 7.5 Hz), 2.50-2.10 (3H, m), 2.00-1.80
(4H, m), 1.80-1.40 (2H, m), 1.30-1.00 (2H, m), 0.83 (9H, s). 13C
NMR (CDCl3) δ 150.19, 106.10, 83.65, 65.77, 43.68, 39.90, 33.31,
32.99, 32.01, 28.84, 27.44, 24.94.
The same sequence was performed with the minor hydroxy ketone
18b to afford 19b: oil, analytical TLC (silica gel F254), 5% EtOAc/
hexane, Rf ) 0.35. MS, base peak ) 151. Exact mass calcd for
C14H24O 208.1827; found 208.1824. Error ) 1.5 ppm. IR (neat, cm-1):
C-H, 2990; CdC, 1640; C-O, 1000. 200-MHz NMR (CDCl3) δ
4.90-4.85 (1H, m), 4.70-4.60 (1H, m), 4.00-3.80 (2H, m), 2.45-
2.35 (1H, m), 2.10-1.60 (6H, m), 1.25-0.90 (4H, m), 0.83 (9H, s).

4-tert-Butylmethylenecyclohexane Epoxidations
J. Am. Chem. Soc., Vol. 118, No. 15, 1996 3567
Table 4. 13C NMR Chemical Shifts (ppm) for C5; Diastereomeric
Epoxides and Diols or Acetonides
Epoxides 20a and 21a,7 20b and 21b,8,27b 20c and 21c,9e and 20o
and 21o9e are described in the literature. The other epoxides of Table
2 are characterized in the supporting information.
Osmylation Procedure. The procedure of Van Rheenen et al. was
employed.28 To a stirred solution of 4-methylmorpholine N-oxide (1.5
equiv) in acetone (3-6 mL) containing 5 drops of water was added a
small crystal of OsO4 (ca. 17 mg, 0.07 mmol) resulting in a yellow
solution. To this solution was added the olefin (30-50 mg; 1 equiv)
in acetone (2-5 mL) via cannula. The reaction was stirred at room
temperature of 18-36 h after which time 2 drops of 3-mercaptopro-
pionic acid was added and the resulting black solution was stirred at
room temperature for an additional hour. The mixture was partitioned
between 1 M NaHCO3 and EtOAc and the aqueous layer was extracted
with EtOAc. The organic layers were combined, washed with brine
(1 × 15 mL), and dried (MgSO4) and the solvent was removed
(aspirator). All compounds were purified by flash chromatography
using the same solvent system reported for analytic TLC unless
otherwise noted. The diastereomer ratios were established by 1H NMR
assay of the product, and the composition of mixtures was confirmed
after separation where possible, and at the stage of the acetonides in
several cases.
Acetonide Formation. Alcohols that were difficult to purify by
chromatography were converted directly into the acetonides for
characterization. To a stirred solution of the diol (1 equiv) and 2,2-
dimethoxypropane (1 equiv) in 3 mL of anhydrous CH2Cl2 was added
a spatula tip of camphorsulfonic acid and the reaction was stirred for
3 h at room temperature. The reaction was diluted with Et2O and
several drops of NaHCO3 were added. The solution was washed with
brine, dried (MgSO4), and concentrated (aspirator). All acetonides were
purified via flash chromatography.
a Data from ref 27b. b Diastereomer not detected. c Tentative assign-
ment; overlapping signals. d Value given is for the acetonide.
13C NMR (CDCl3) δ 152.33, 104.62, 85.72, 67.54, 46.82, 1.33, 35.73,
33.80, 32.26, 28.58, 27.57, 25.63.
General Procedure for MCPBA Epoxidations. To a stirred
solution of the olefin (1 equiv, approximately 0.1-0.2 mmol) in
methylene chloride (10 mL) was added 80-85% MCPBA (1.5 equiv)
in 5-10 mL of methylene chloride via cannula at 0 °C. The reaction
was stirred at 0 °C for 3 h and then allowed to warm to room
temperature overnight. The reaction was then washed with 10% Na2-
SO3 (1 × 10 mL), 1 M NaHCO3 (1 × 10 mL), and brine (2 × 10 mL)
and dried (MgSO4) and then the solvent was removed (aspirator). Pure
materials were obtained via flash chromatography according to the Still
method employing the solvent system listed for TLC analysis.
Diols/Acetonides. The diastereomeric 4-tert-butylmethylenecyclo-
hexane diols (22a and 23a) are described in the literature.45 The other
osmylation products are characterized in the supporting information.
Acknowledgment. This work was supported by the National
Science Foundation. The authors are also grateful to Prof. K.
Houk for providing copies of manuscripts prior to publication.
Supporting Information Available: Characterization data
for epoxides and diols described in Table 2 (10 pages). This
material is contained in many libraries on microfiche, im-
mediately follows this article in the microfilm version of the
journal, can be ordered from the ACS, and can be downloaded
from the Internet; see any current masthead page for ordering
information and Internet access instructions.
General Procedure for Dimethyldioxirane Epoxidations.25 To a
stirred solution of the olefin (1 equiv, approximately 0.2 mmol) in
methylene chloride (30 mL) at 0 °C was added an excess of the
dimethyldioxirane solution (in acetone 0.01-0.1 M) via syringe
(normally 15-30 equiv to ensure complete conversion). The reaction
was stirred at 0 °C for approximately 1 h and was then warmed to
room temperature overnight, dried (MgSO4), and concentrated (aspira-
tor). Pure materials were obtained via flash chromatography (Still
method) employing the solvent systems listed for TLC analysis.
JA953040P
(45) Patrick, D. W.; Truesdale, L. K.; Biller, S. A.; Sharpless, K. B. J.
Org. Chem. 1978, 43, 2628.