An examination of hyperconjugative and electrostatic effects in the hydride reductions of 2-substituted-4-tert-butylcyclohexanones

Article abstract of DOI:10.1021/jo0011787

To better understand electronic effects on the diastereoselectivity of nucleophilic additions to the carbonyl group, a series of 2-X-4-tert-butylcyclohexanones (X = H, CH₃, OCH₃, F, Cl, Br) were reacted with LiAlH₄. Reduction of ketones with equatorial substituents yields increasing amounts of axial alcohol in the series for X {H < CH₃ < Br < Cl < F ? OCH₃}. These data cannot be explained by steric or chelation ·effects or by the theories of Felkin-Anh or Cieplak. Instead, an electrostatic argument is introduced: due to repulsion between the nucleophile and the X group, axial approach becomes energetically less favorable with an increase in the component of the dipole moment anti to the hydride approach trajectory. The ab initio calculated diastereoselectivities were close to the experimental values but did not reproduce the relative selectivity ordering among substituents. For reduction of ketones with axial substituents, increasing amounts of axial alcohol are seen in the series for X {Cl < Br < CH₃ < OCH₃ < H < F}. After some minor adjustments are made, this ordering is consistent with both the electrostatic model and Felkin-Anh theory. Cieplak theory cannot account for these data regardless of adjustments. Ab initio calculated diastereoselectivities were reasonably accurate for the nonpolar substituents but were poor for the polar substituents.

Full text of DOI:10.1021/jo0011787

1694
J . Org. Chem. 2001, 66, 1694-1700
An Exa m in a tion of Hyp er con ju ga tive a n d Electr osta tic Effects in
th e Hyd r id e Red u ction s of
2-Su bstitu ted -4-ter t-bu tylcycloh exa n on es
Robert E. Rosenberg,*,† Richele L. Abel, Michael D. Drake, Daniel J . Fox, Andrew K. Ignatz,
David M. Kwiat, Kristen M. Schaal, and Peter R. Virkler
Department of Chemistry, SUNY, College at Geneseo, Geneseo, New York 14454
rer3@cisunix.unh.edu
Received August 3, 2000
To better understand electronic effects on the diastereoselectivity of nucleophilic additions to the
carbonyl group, a series of 2-X-4-tert-butylcyclohexanones (X ) H, CH3, OCH3, F, Cl, Br) were reacted
with LiAlH4. Reduction of ketones with equatorial substituents yields increasing amounts of axial
alcohol in the series for X {H < CH3 < Br < Cl < F , OCH3}. These data cannot be explained by
steric or chelation effects or by the theories of Felkin-Anh or Cieplak. Instead, an electrostatic
argument is introduced: due to repulsion between the nucleophile and the X group, axial approach
becomes energetically less favorable with an increase in the component of the dipole moment anti
to the hydride approach trajectory. The ab initio calculated diastereoselectivities were close to the
experimental values but did not reproduce the relative selectivity ordering among substituents.
For reduction of ketones with axial substituents, increasing amounts of axial alcohol are seen in
the series for X {Cl < Br < CH3 < OCH3 < H < F}. After some minor adjustments are made, this
ordering is consistent with both the electrostatic model and Felkin-Anh theory. Cieplak theory
cannot account for these data regardless of adjustments. Ab initio calculated diastereoselectivities
were reasonably accurate for the nonpolar substituents but were poor for the polar substituents.
In tr od u ction
the strongest electron donating group should lie anti-
periplanar to the incipient bond to maximize C2sX σ,
Nu‚‚‚CdO σ* orbital interactions in the TS (Figure 1).1c,6
In an effort to reconcile the two conflicting theories, le
Noble suggested that the ideas of Felkin, Anh, and
Cieplak all compete, with the dominant control element
depending on the particular circumstances of the reac-
tion.7 Rather than examining the TS, some researchers
have looked at the unequal lobes of the π orbital in the
ground state of the ketone.1a,b,d,8 These models have
achieved some predictive successes, probably due to the
early TS of nucleophilic addition reactions.
The diastereoselective addition of nucleophiles to car-
bonyl compounds is a topic of considerable importance.1
While the first theory to address this issue was advanced
by Cram,2 most current researchers cite a hybrid theory
of Felkin and Anh. Felkin proposed that bonds vicinal to
the reacting center would be staggered with respect to
the forming bond in order to minimize torsional interac-
tions in the transition state (TS).3 On the basis of ab initio
calculations, Anh found that the bond vicinal to the
reaction center which contained the strongest electron
withdrawing group, C2-X, should lie antiperiplanar to the
incipient bond in order to maximize Nu‚‚‚CdO σ, C2sX
σ* orbital interactions in the TS.4 These ideas and
experimental work on the angle of nucleophilic attack5
were incorporated into what is commonly known as the
Felkin-Anh model. However, some researchers in this
field subscribe to the theory of Cieplak, which states that
The particular case of hydride reduction of cyclohex-
anones is an important special case in the study of
diastereoselective nucleophilic addition reactions to car-
bonyls. In these systems, there is a solid understanding
of steric effects, but there is no consensus on the origin
of electronic effects. A common strategy used in the study
of electronic effects is to employ systems where the polar
substituent, X, is far from the reaction center (Figure 2).
This approach avoids the complications of competing
steric effects. Houk argued for the Felkin-Anh model
using data from the NaBH4 reduction of 4 substituted
† Currently on leave at the University of New Hampshire. Email:
bob.rosenberg@unh.edu.
(1) A thematic issue of Chem. Rev. was recently devoted to
diastereoselection. (a) Dannenberg, J . J . Chem. Rev. 1999, 99, 1225-
1241. (b) Tomoda, S. Chem. Rev. 1999, 99, 1243-1263. (c) Cieplak, A.
S. Chem. Rev. 1999, 99, 1265-1335. (d) Ohwada, T. Chem. Rev. 1999,
99, 1337-1375. (e) Gung, B. W. Chem. Rev. 1999, 99, 1377-1385. (f)
Kaselj, M.; Chung, W.-S.; le Noble, W. J . Chem. Rev. 1999, 99, 1387-
1413. (g) Adcock, W.; Trout, N. A. Chem. Rev. 1999, 99, 1415-1435.
(h) Mehta, G.; Chandrasekhar, J . Chem. Rev. 1999, 99, 1437-1467.
(i) Wipf, P.; J ung, J .-K. Chem. Rev. 1999, 99, 1469-1480.
(2) (a) Cram, D. J .; Elhafez, F. A. A. J . Am. Chem. Soc. 1952, 74,
5828. (b) Cram, D. J .; Kopecky, K. R. J . Am. Chem. Soc. 1959, 81, 2748.
(3) (a) Cherest, M.; Felkin, H.; Prudent, N. Tetrahedron Lett. 1968,
2199. (b) Cherest, M.; Felkin, H. Tetrahedron Lett. 1968, 2205.
(4) (a) Anh, N. T.; Eisenstein, O. Nouv. J . Chim. 1977, 1, 61. (b)
Anh, N. T. Top. Curr. Chem. 1980, 88, 145.
(6) (a) Cieplak, A. S. J . Am. Chem. Soc. 1981, 103, 4540. (b) Cieplak,
A. S.; Tait, B. D.; J ohnson, C. R. J . Am. Chem. Soc. 1989, 111, 8447.
(c) Cieplak, A. S.; Wiberg, K. B. J . Am. Chem. Soc. 1992, 114, 9226-
9227.
(7) Li, H.; le Noble, W. J . Recl. Trav. Chim. Pays-Bas 1992, 111,
199-210.
(8) (a) Klein, J . Tetrahedron Lett. 1974, 4307. Klein, J . Tetrahedron
1974, 30, 3349. (b) Frenking, G.; Koehler, K. F.; Reeetz, M. T. Angew.
Chem. 1991, 103, 1167. (c) Huang, X. L.; Dannenberg, J . J .; Duran,
M.; Bertran, J . J . Am. Chem. Soc. 1993, 115, 4024-4030. (d) Tomoda,
S.; Senju, T. Tetrahedron 1997, 53, 9057. (e) Tomoda, S.; Senju, T.
Chem. Commun. 1999, 423.
(5) Burgi, H. B.; Dunitz, J . D.; Shefter, E. J . J . Am. Chem. Soc.
1973, 95, 5065.
10.1021/jo0011787 CCC: $20.00 © 2001 American Chemical Society
Published on Web 02/03/2001

Electrostatic Effects in Hydride Reductions
J . Org. Chem., Vol. 66, No. 5, 2001 1695
F igu r e 3. Reduction of 1-X gives 2a -X and 2e-X. Reduction
of 3-X gives 4a -X and 4e-X. X ) H, CH3, OCH3, F, Cl, Br.
F igu r e 1. Molecular orbital illustration of Cieplak versus
Felkin-Anh stabilization.
vicinal to the reaction center and thus maximize electronic/
orbital effects. The steric effects that result from this
proximity are minimized by the small size of the nucleo-
phile. The polar substituents X ) {OCH3, F, Cl, Br} are
bracketed in size between X ) H and X ) CH3, the latter
providing an upper bound for steric effects. The large
conformational bias of the t-Bu group and the well-
defined geometry of the six-membered ring combine to
allow the study of both axial (ax) and equatorial (eq)
substituents. Unlike NaBH4, LiAlH4 is thought to involve
a single reducing species at sufficiently high hydride-to-
ketone ratios.13
Our approach is similar to that of Fraser who looked
at the LiAlH4 reduction of 2-substituted dibenzocyclo-
heptenones.14 Results from Fraser and us will be com-
pared. Paquette examined the addition of allyl indium
reagents to 2-alkoxy and hydroxy-4-tert-butylcyclohex-
anones, though he was concerned primarily with chela-
tion control.15
F igu r e 2. Examples of sterically unbiased ketones.
trans decalones.9 le Noble found that the Cieplak model
best explained the NaBH4 reduction of 5-X-adamantan-
2-ones.1f,10 Adcock also looked at NaBH4 reductions of 5-X-
adamantan-2-ones, but asserted that electrostatic inter-
actions were responsible for the observed product
ratios.1g,11 Using a system similar to le Noble’s, Mehta
and Chandrasekhar argued for both the Cieplak model
and electrostatic effects.1h,12 Unfortunately, these studies
are at odds with each one another, possibly because the
observed selectivities are small.
Resu lts a n d Discu ssion
1. Gen er a l Con sid er a tion s. The results for the
reduction of ketones 1-X and 3-X to the alcohols 2-X and
4-X (Figure 3), respectively, are listed in Table 1. The
eq/ax ratio of alcohols is first converted into ∆G, using
the relationship ∆G ) -RT ln K. K is not strictly an
equilbrium constant but rather equals the following
ratio: [the rate of axial attack by the hydride ) amount
of observed eq alcohol] divided by [the rate of eq attack
by the hydride ) amount of observed ax alcohol]. Then,
∆∆G is calculated using the relation ∆∆GX ) ∆GX - ∆GH.
For example, “∆∆G 1-X > 0.00 kcal/mol” means that less
eq alcohol (alternatively, lower axial selectivity) is formed
in the reduction of 1-X than of 1-H. It is important to
note that ∆∆G contains contributions of the TS energies
from both ax and eq approaches of the hydride.
We hoped to help resolve the Felkin-Anh vs Cieplak
problem by examining the LiAlH4 reduction of 2-X-4-tert-
butylcyclohexanones (1-X, 3-X), where X ) {H, CH3,
OCH3, F, Cl, Br}. Our strategy is to place the X group
(9) (a) Wu, Y.-D.; Tucker, J . A.; Houk, K. N. J . Am. Chem. Soc.
1991, 113, 5018-5027. Other relevant papers by Houk: (b) Wu, Y.-
D.; Houk, K. N. J . Am. Chem. Soc. 1987, 109, 908-910. (c) Wu, Y.-D.;
Houk, K. N. J . Am. Chem. Soc. 1993, 115, 10992. (d) Coxon, J . M.;
Houk, K. N.; Luibrand, R. T. J . Org. Chem. 1995, 60, 418-427.
(10) (a) Cheung, C. K.; Tseng, L. T.; Lin, M.-H.; Srivastava, S.; le
Noble, W. J . J . Am. Chem. Soc. 1986, 108, 1598. (b) Lin, M.-H.; Silver,
J . E.; le Noble, W. J . J . Org. Chem. 1988, 53, 5155. For recent work,
see: (c) Kaselji, M.; le Noble, W. J . J . Org. Chem. 1998, 63, 2758. (d)
Gonikberg, E. M.; Picart, F.; le Noble, W. J . J . Org. Chem. 1998, 63,
4885.
2. Red u ction of Keton es 1-X. For 1-X, ∆∆G increases
(axial selectivity decreases) for X in the order {H < CH3
, Br < Cl < F , OCH3}. As X is relatively close to the
reaction center, both steric and chelation effects could
(13) Wigfield, D. C.; Gowland, F. W. J . Org. Chem. 1980, 45, 653.
(14) Fraser, R. R.; Faibish, N. C.; Kong, F.; Bednarski, K. J . Org.
Chem. 1997, 62, 6164-6176.
(15) (a) Paquette, L. A.; Lobben, P. C. J . Org. Chem. 1998, 63, 5604-
5616. (b) Paquette, L. A.; Lobben, P. C. J . Am. Chem. Soc. 1996, 118,
1917-1930.
(11) Adcock, W.; Cotton, J .; Trout, N. A. J . Org. Chem. 1994, 59,
1867-1876.
(12) Mehta, G.; Ravikrishna, C.; Ganguly, B.; Chandrasekhar, J .
Chem. Commun. 1997, 75-76.

1696 J . Org. Chem., Vol. 66, No. 5, 2001
Rosenberg et al.
Ta ble 1. Red u ction s of 1 a n d 3 w ith LiAlH4 in THF
Ta ble 2. Ab In itio Ca lcu la ted Va lu es of Selected An gles
for Keton es 1 a n d 3a
a t 0 °C
1
3
sum of
deviations
b
b
b
X
C3-C2
C5-C6
O-C1
X
% 2a a
∆∆Gb Fraserc
% 4a a
∆∆Gb Fraserc
H
1
1
1
1
1
3
3
3
3
47.6
2.2
1.5
3.2
3.9
-3.3
2.0
6.9
-0.5
47.6
2.2
0.9
2.7
3.3
-1.7
1.3
5.8
1.0
1.1
0.6
-0.4
1.0
0.7
-0.9
0.2
0.0
5.0
2.0
6.9
7.9
-5.9
3.5
16.4
2.8
H
9(0.6,4)d
11(1)d
0.00
0.12
0.00
9
0.00
0.00
CH3
OCH3
F
CH3
0.82 5(1)d
-0.35
-1.35
OCH3 67(1.7,3)e 1.64
>1.61 8(4.0,3)e
1.11 9(1.4,6)
-0.08 <-1.61
F
40(4.8,6)
34(0.8,5)f 0.90
29(1.9,4) 0.77
1.04
0.00
Cl
Cl
Br
1.15 0.9(0.0,2)g -1.30 <-1.61
1.4(0.04,2)g -1.05
CH3
OCH3
F
a
3.7
2.3
Data are reported in the form m(s,n), where m is the mean
value of n trials and s is the standard deviation. b In kcal/mol. ∆∆G
Cl
is defined in the text. c Reproduced from ref 14. In accord with
d
In degrees using MP2/6-31G* optimized geometries. b For 1-H,
the value shown is the angle made by the designated bond with
the plane defined by C6-C1-C2. For X * H, the value given equals
the angle made by the bond in 1 (or 3)-X minus the angle made
by 1-H.
a
ref 30. e Battioni reports 64% of 2a -OCH3 and 15% of 4a -OCH3
(ref 28). The discrepancy of our data with the latter value is
probably due to spectral congestion in the 3.3-3.9 δ region of the
60 MHz 1H NMR signals for the product mixture. f Cherest reports
significantly higher axial selectivity (19%) with ether as a solvent
(ref 30). The value listed is from integration of a very weak 1H
g
the bonds C3-C2, C5-C6, and O-C1 and the plane formed
by C6-C1-C2. Geometric data for all ketones are pre-
sented in Table 2 and are obtained from MP2/6-31G* ab
initio calculations, a level of theory well-known for giving
highly accurate structural information.17
NMR resonance and thus is an upper limit for the amount of minor
product present.
potentially contribute to the observed selectivity. Steric
repulsion between the substituent and the reagent would
hinder axial but not equatorial attack and therefore lead
to higher values of ∆∆G. The CH3 group serves as a
control for steric effects, since it is both the largest
substituent used in this study {as judged by van der
Waals radii (CH3 2.0 Å, Br 1.85 Å) or by A values (in
kcal/mol, CH3 1.8, OCH3 0.60, Br 0.48)}16 and it is
unlikely to have significant chelation or electronic effects.
Steric effects are seen to be minor as judged by the ∆∆G
of 1-CH3 which is only 0.12 kcal/mol, a value which is
significantly smaller than the ∆∆G’s of 0.77-1.64 kcal/
mol seen for the polar substituents. Another argument
against steric control is the fact that ∆∆G increases as
the halogen atom size decreases (∆∆G for 1-Br is 0.77
kcal/mol, while 1-F has a ∆∆G of 1.04 kcal/mol). In-
creased chelation of the reagent by X would direct the
hydride to the axial face of the ketone and lead to lower
values of ∆∆G. Chelation would be more important in X
in the progression {OCH3 > F > Cl > Br . H ) CH3},
resulting in an ordering in ∆∆G of {OCH3 < F < Cl <
Br , H ) CH3}. This prediction for ∆∆G is exactly the
opposite of what is observed. While not excluded as a
factor, chelation control is clearly not the dominant
control element. Thus, the selectivity seen in the reduc-
tion of 1-X must be governed by torsional strain and/or
electronic effects. The latter includes the diverse ideas
of transition state hyperconjugation and electrostatics.
The torsional strain component of the Felkin-Anh
model (the other component is hyperconjugation, dis-
cussed next) has often been used to explain diastereose-
lectivity in hydride reductions of cyclohexanones.3,4,9ab In
this theory, the large amount of equatorial alcohol seen
in the hydride reduction of 1-H is explained by a TS
which is perfectly staggered during ax attack but has
partial eclipsing interactions during eq attack. Further-
more, Anh has suggested that as the ketone gets flatter,
ax attack becomes even more favored.4,9b The flatness of
the ketones 1-X will be compared to 1-H to see if there
are significant geometric distortions in the former which
may account for the observed selectivities. Here, flatness
is defined as the sum of the three angles made between
For 1-X, the carbonyl becomes more puckered in X in
the order {H < OCH3 < CH3 < F < Cl}. According to the
torsional strain model, the most puckered ketones should
have the least amount of axial selectivity and the largest
values of ∆∆G. The ∆∆G for 1-OCH3 is the highest in
the 1-X series, yet it is just 2° less flat than 1-H, the
ketone with the lowest value of ∆∆G. In contrast, the
two ketones 1-H and 1-CH3 have very similar ∆∆G’s yet
differ in flatness by 5°. From these and other examples,
one can conclude that the small variations in flatness for
the ketones 1-X do not correlate well with their ∆∆G’s
of reduction.
Hyperconjugation in the TS is used to explain diaste-
reoselectivity by both the Cieplak and Felkin-Anh
models. In both theories, stereocontrol arises from the
relative stabilization provided by the orbitals on the
carbon-substituent bond adjacent to the reaction center
which are anti to each of the two approaches of the
reagent. For Felkin-Anh, maximum stabilization occurs
when the σ* orbital is low in energy, while for Cieplak,
maximum stabilization is achieved when the σ orbital is
high in energy. In the case of 1-X, the antiperiplanar
orbitals for ax attack are in the C2-H bond and those
for eq attack are in the C2-C3 bond. Analysis of the
Felkin-Anh and Cieplak models is now reduced to
determining what effect an equatorial and polar X group
has on the relative energies of the C2-H and C2-C3
orbitals. Determining this effect may be quite difficult
given that a seemingly simpler problem, the relative
donor/acceptor abilities of unperturbed C-H vs C-C
bonds, is still being actively debated.18 Even if the relative
orbital perturbations of an X group could be determined,
it is unlikely that subtle differences in the energies of
the C2-H vs C2-C3 orbitals could lead to the sizable
∆∆G’s seen in the 1-X series regardless of which model
is used.19 Another difficulty faced by the orbital theories
is that the ∆∆G for 1-OCH3 is significantly higher than
for 1-F , yet fluorine has a larger effect on orbital energies
(17) Hehre, W. J .; Radom, L.; Schleyer, P. v. R.; Pople, J . A. Ab
Initio Molecular Orbital Theory; Wiley: New York, 1986.
(18) For recent work arguing that C-C is the better donor, see refs
1a and 9a. For recent work claiming that C-H is the better donor, see
refs 1c and 1g.
(16) Carey, F. A.; Sundberg, R. J . Advanced Organic Chemistry Part
A; Plenum Press: New York, 1990; p 135 (A values) and p 120 (van
der Waals radii).

Electrostatic Effects in Hydride Reductions
J . Org. Chem., Vol. 66, No. 5, 2001 1697
Ta ble 3. Ab In itio Ca lcu la ted Va lu es of th e An ti
than methoxy. Thus, the major theories used to describe
electronic effects on nucleophilic additions are incomplete.
At the very least, they are not applicable to the important
case of equatorial substituents which are adjacent to the
reaction center in cyclohexanones. In a broader sense,
the effect of substituents which do not access a geometry
anti to the approach of a nucleophile are poorly described
by the Felkin-Anh and Cieplak models.
Com p on en t for th e C-X Dip olea
ground stateb
transition statec
group moment
X
θ
cosθ
θ
cos θ
for C-Xd
F
Cl
1
1
1
3
3
3
86
89
144
-147
-156
-92
0.08
0.02
0.81
0.84
0.91
0.03
64
74
15
0.43
0.28
0.97
0.98
0.97
0.49
1.41
1.46
1.3
1.41
1.46
1.3
OCH3
F
-170
-164
-119
Cl
Another model used to explain diastereoselective hy-
dride reductions of polar ketones is electrostatic effects.
Wu, Tucker, and Houk invoked electrostatic repulsion
between the hydride and the substituent to account for
the high selectivity for axial attack in the NaBH4 reduc-
tion of axial 4-fluoro trans decalone.9a They also used ab
initio studies to show that when the substituent was OH
or NH2, the selectivity was strongly dependent on the
orientation of the lone pairs. Paddon-Row, Wu, and Houk
found electrostatic interactions between the substituent
and the nucleophile to be the exclusive control element
in a computational study of the addition of LiH to 2,3-
disubstituted 7-norbornanones.20 Adcock argued that
electrostatic interactions could account for the selectivity
seen in the NaBH4 reduction of 5-X-adamantan-2-ones
but did not rule out participation of Felkin-Anh orbital
effects.1g,11 Wipf and Kim invoked electrostatic effects in
explaining the stereoselectivity of methyl Grignard ad-
dition to 4,4-disubstituted cyclohexadienones, though no
selectivity was seen in these systems when hydride
reducing agents were used.1i,21
OCH3
a
For the ground-state compounds 1-X and 3-X, the HF/6-31G*
optimized geometries for 2-X-cyclohexanone were used. For transi-
tion states, the HF/6-31G* optimized geometries of the transition
b
state for LiH addition to 2-X-cyclohexanone were used. θ ) the
angle made between the C-X bond vector and the vector perpen-
dicular to the plane defined by C2-C1-C6. For X ) OCH3, the
C-X bond vector is taken as the vector sum of the C2-O and
O-CH3 bonds. c θ ) the angle made between the C-X bond vector
and C1-H bond vector. For X ) OCH3, the C-X bond vector is
d
defined as in footnote b. Values are in es units × 1018. Taken
from ref 16, p 16.
With this background, it is proposed that axial attack
by the hydride is electrostatically repelled by the partial
negative charge on the X group. This repulsion is lower
for eq attack due to the larger distance between the
hydride and the X group. Thus, as X becomes more
electronegative in the series {Br < Cl < F}, ax attack
becomes more unfavorable relative to eq attack and the
∆∆G will increase as is observed. An explanation of the
very high ∆∆G seen for 1-OCH3 requires a slightly more
elaborate model. Since dipole-charge interactions are
proportional to the cosine of the angle, cos θ, made
between the C-X dipole and the incoming nucleophile
(treated here as a point charge),22 repulsion of axial
attack is greatest when the C-X dipole is anti to the
nucleophile’s trajectory. Ab initio calculated values of cos
θ for the C-X dipole in both the ground state and TS
are listed in Table 3. In 1-F and 1-Cl, only a small
percentage of the C-X dipole lies perpendicular to the
carbonyl group. However, this value increases signifi-
cantly in the TS to 43 and 28%, respectively. For 1-OCH3,
most of the group dipole moment lies perpendicular to
the carbonyl group and this value increases to 97% in
the TS. Taken with the roughly equal values of the C-F,
C-Cl, and C-OCH3 dipole moments, the cos θ values
predict that 1-OCH3 should have significantly greater
electrostatic repulsion of the nucleophile than 1-F
F igu r e 4. Vector illustration of both the bond dipole(s) from
the C-X bond and of the projection of the resultant dipole onto
the y axis.
and 1-Cl. This prediction is consistent with the experi-
mental data (Figure 4).
It is worthwhile to compare and contrast our results
with those of Fraser, who looked at the LiAlH4 reduction
of 2-substituted dibenzocycoheptadienones.14 Fraser’s
data is reproduced in Table 1. A common feature of both
studies is a diminished nucleophilic attack on the face
of the carbonyl that is syn to the electronegative atom,
with the largest effect seen for OCH3. As here, Fraser
argued that his reduction data are best explained by a
combination of electrostatic/dipole forces and argued
against the hyperconjugative theories of the Felkin-Anh
and Cieplak models.
Our system has two minor advantages over Fraser’s.
First, Mehta and Chandrasekhar have suggested that
steric effects of the 2-substituent swamp out the hyper-
conjugative contributions that are present in Fraser’s
system.1h This argument cannot be applied to our system
as the steric effect has been shown to be almost negligible
(0.12 kcal/mol for 1-CH3). In a second example, le Noble
has suggested that complexation of lithium ion to aro-
matic rings could effect stereoselectivity.10b Again, this
concern is not relevant to our system.
(19) This intuitive statement is derived from the following reason-
ing. Assume that differential orbital perturbation can explain the ∆∆G
values for reduction of 1-X. If so, then the effects for the reduction of
3-X must be much larger than for 1-X, as replacement of X for H (3-X)
affects orbital energies substantially more than the perturbation that
X (in 1-X) causes in the energy of the C2-H orbital. Yet the ∆∆G’s for
3-X are comparable in magnitude to those seen for 1-X. Therefore, the
hypothesis is invalid.
(20) Paddon-Row; M. N.; Wu, Y.-D.; Houk, K. N. J . Am. Chem. Soc.
1992, 114, 10638-10639.
(21) Wipf, P.; Kim, Y. J . Am. Chem. Soc. 1994, 116, 11678.
(22) Halliday, D.; Resnick, R.; Walker, J . Fundamentals of Physics,
Vol. 2; J ohn Wiley and Sons: New York, 1997; p 570.
The reduction of 1-X was studied by ab initio calcula-
tions in an attempt at gaining a deeper understanding
of the reaction mechanism. Calculated selectivities for
the reduction of 1-X by LiH, NaH, BH3, and AlH3 are
listed in Table 4. For reduction of 1-X, the agreement

1698 J . Org. Chem., Vol. 66, No. 5, 2001
Rosenberg et al.
Ta ble 4. Ab In itio Ca lcu la ted Dia ster eoselectivity for
th e Red u ction of 1 a n d 3a
data for ketones 3-X are listed in Table 2, with “flatness”
defined the same as for 1-X. For 3-X, the carbonyl
becomes more puckered in X in the order {CH3 < H < Cl
< OCH3 , F}. For most of these ketones, puckering and
∆∆G are not well correlated. For example, while 3-Cl is
just slightly more puckered than 1-H, its ∆∆G is large
and negative. Similarly, 3-Cl and 3-OCH3 have very
similar geometries but significantly different ∆∆G’s.
However, for 3-F there is a very large increase in
puckering and a relatively high amount of eq attack in
accordance with the Felkin-Anh model. The resulting
∆∆G for 3-F is not easily explained by other factors, vide
infra. These data lead to the observation that the
torsional strain contribution to selectivity will not vary
much unless the geometric perturbation between mol-
ecules is relatively large.
X
LiH
NaH
BH3
AlH3
exptb
H
1
1
1
1
1
3
3
3
3
3
-1.12
-0.61
0.40
-1.15
-0.44
-1.12
-1.88
-2.54
-2.84
-5.16
-0.80
-0.27
-0.95
0.28
-1.25
-0.58
-2.18c
-0.09
-0.24
-1.25
-1.68
-1.52c
0.79
-1.18
-0.53
-1.85
-0.07
-0.25
-1.18
-1.46
-1.28
0.54
-1.26
-1.14
0.38
-0.22
-0.36
-1.26
-1.61
-1.34
-1.26
-2.56
CH3
OCH3
F
Cl
H
CH3
OCH3
F
0.19
-0.80
-1.78
-2.77
-2.84
-4.97
Cl
-1.43
-1.82
a
In kcal/mol using MP2/6-311+G**//HF/6-31G*. Entries )
b
∆H(TS for eq alcohol) - ∆H(TS for ax alcohol). From Table 1.
c The HF/3-21G geometry is used.
with experiment by all four reagents is excellent, with
only the value for the reduction of 1-OCH3 by BH3 or
AlH3 deviating from experiment by significantly more
than 1 kcal/mol. While LiH is the reagent that best
reproduces the experimental data, both AlH3 and BH3
would be significantly better than LiH if not for 1-OCH3.
Opposite to expectations, the reasonably accurate ∆G’s
of the reaction did not translate into good predictions of
relative reduction ratios for the various 1-X’s. Experi-
mental reduction of 1-X gives less negative ∆G’s (more
axial alcohol) in X in the order {H, CH3 < F, Cl < OCH3}.
LiH correctly predicts that 1-OCH3 will yield the least
amount of equatorial alcohol but fails to order the
remaining substituents correctly. The other three re-
agents all model the H, CH3 < F, Cl trend well but fail
badly (NaH) to very badly (BH3, AlH3) for 1-OCH3. This
unexpectedly poor ordering of the energetics of this
reaction does not allow one to reliably draw subtle
mechanistic insights on the basis of the ab initio calcu-
lated transition states.
3. Red u ction of Keton es 3-X. For the reduction of
3-X, ∆∆G decreases (axial selectivity increases) for X in
the order {F ) H > OCH3 > CH3 . Br > Cl}. Unlike
1-X, steric and chelation effects cannot be ignored in the
reduction of 3-X. Steric repulsion of the nucleophile by
X retards eq but not ax approach and leads to more
negative values of ∆∆G. Analogous to 1-CH3, the ∆∆G
for the reduction of 3-CH3 (-0.35 kcal/mol) provides an
upper bound for the magnitude of the steric effect in this
system. Though significant, steric effects can account for
only a fraction of the large negative ∆∆G’s seen in the
reduction of 3-Cl (-1.30 kcal/mol) and 3-Br (-1.05 kcal/
mol). Chelation of the nucleophile by the X group would
direct attack to the equatorial face and lead to less
negative values of ∆∆G with the effect decreasing in X
as follows: {OCH3 > F > Cl > Br > H ) CH3}. This
reasoning may help explain the lower axial selectivity
seen for the reduction of OCH3, F as compared to Cl, Br.
Against this, it is not clear why chelation would be
important in the reduction of 3-X but not in 1-X.
Importantly, the high axial selectivity seen in the reduc-
tion of 3-Cl and 3-Br is clearly not due to chelation, which
lowers axial selectivity. Instead, there must be an ad-
ditional effect(s). As with 1-X, torsional effects, transition
state hyperconjugation, and electrostatics will be dis-
cussed.
Unlike that for the reduction of 1-X, the hyperconju-
gation theories of Felkin-Anh and Cieplak have unam-
biguous and opposite predictions for the reduction of 3-X.
Felkin-Anh states that an electron-withdrawing group
anti to the nucleophile will accelerate axial attack leading
to ∆∆G’s with an ordering in X of {CH3 ≈ H > Br > Cl
> OCH3 > F}. Cieplak predicts a maximum acceleration
when the group lying anti to the nucleophile is electron
donating, leading to an exact reversal from the ordering
above. Neither theory fully accounts for the experimental
data. Felkin-Anh correctly predicts lower ∆∆G’s (higher
axial selectivity) for the polar versus nonpolar X’s.
However, this theory incorrectly predicts that the ∆∆G’s
for 3-F and 3-OCH3 should be lower than for 3-Cl and
3-Br , an energy difference between theory and experi-
ment of more than 1 kcal/mol. This discrepancy could be
due to a combination of three factors: (1) the ∆∆G
lowering steric component of X ) Cl, Br, CH3 is larger
than for when X ) F, OCH3, though this factor can
account for at most 0.35 kcal/mol (-∆∆G of 3-CH3), (2)
the ∆∆G for X ) F, OCH3 could be raised by chelation
relative to X ) Cl, Br, though there was no evidence for
effects of similar magnitude in the reduction of 1-X, and
(3) 3-F has a relatively large geometric distortion which
leads to a higher than expected ∆∆G. The Cieplak model
correctly orders the ∆∆G for the polar substituents, F >
OCH3 > Cl > Br. However, this model incorrectly
theorizes that reduction of 3-CH3 should have a more
negative ∆∆G than 3-Br , 3-Cl, an error of over 0.7 kcal/
mol. Unlike the Felkin-Anh model, this discrepancy
cannot be explained by adding steric and chelation
considerations. The 0.7 kcal/mol error mentioned above
is only made larger when one considers the greater steric
bulk of CH3 versus either Cl or Br. Also, any chelation
seen in 3-Br , 3-Cl would raise ∆∆G relative to 3-CH3,
further worsening this already poor prediction.
The electrostatic repulsion argument used for 1-X can
be applied to the reduction of 3-X. In eq attack of 3-X,
there is unfavorable electrostatic repulsion between two
partially negative groups, the hydride and the electrone-
gative X group. For ax attack, there is a favorable
electrostatic interaction as the partially negative hydride
approaches from the positive end of the C-X dipole. As
for 1-X, calculated values of cos θ for the C-X dipole in
both the ground state and TS are listed in Table 3. For
3-F and 3-Cl, over 80% of the C-X dipole lies perpen-
dicular to the carbonyl group. This stabilizing dipole
becomes even more optimally aligned in the TS. In
contrast, the group dipole moment in 3-OCH3 has only
a small component perpendicular to the carbonyl group.
Similar to 1-X, the torsional strain component of the
Felkin-Anh model predicts that increased puckering by
the ketones 3-X will lead to decreasing amounts of ax
attack and thus less negative values of ∆∆G. Geometric

Electrostatic Effects in Hydride Reductions
J . Org. Chem., Vol. 66, No. 5, 2001 1699
This value increases to about 50% in the calculated TS.
The predictions of the electrostatic model would give
decreasing ∆∆G’s with an ordering in X of {H ≈ CH3 >
OCH3 > Br > Cl > F}. Experimentally, the ∆∆G for
3-CH3 is more negative than for both 3-H and 3-OCH3,
a result that can be attributed to steric effects. With this
adjustment, only the value for the reduction of 3-F is
anomalous. As discussed above, there is a severe geo-
metric distortion in 3-F which leads to a higher than
expected amount of eq attack.
trans 2-X-tert-butylcyclohexanones (3-X), either the elec-
trostatic model or the Felkin-Anh theory is consistent
with the data. The Cieplak model fails to describe this
system. Ab initio calculations reproduced experimental
data fairly well, especially with the cis ketones, though
experimental trends across substituents were modeled
less satisfactorily.
Exp er im en ta l Section
While the data of Fraser14 are very similar to ours in
the 1-X series, the corresponding data for ax substituents
are very different. While both studies see large negative
∆∆G’s for Cl, Fraser sees a much more negative ∆∆G
for OCH3 than we do (< -1.61 kcal/mol for OCH3 versus
-0.08 kcal/mol for 3-OCH3). The larger steric effect in
Fraser’s system (∆∆G of -1.35 kcal/mol for CH3 versus
-0.35 kcal/mol for 3-CH3) may contribute to some of this
difference. Another factor could be that the orientation
of the OCH3 group, and hence the direction of the dipole
moment, is different in the two studies.
Gen er a l Meth od s. All reactions were carried out under an
atmosphere of nitrogen in oven-dried glassware. Reagents were
used without further purification unless otherwise indicated.
THF was freshly distilled from potassium. Solutions were
evaporated under reduced pressure with a rotary evaporator.
Column chromatography was performed on a silica gel column
using an ethyl acetate-hexanes mixture as the eluent.
Syn th esis of Keton es. Ketone 1-H was obtained com-
mercially. Other ketones, 1-CH3 and 3-CH3,23 1-F and 3-F ,24
1-Cl and 3-Cl,25 1-Br and 3-Br ,26 and 1-OCH3 and 3-OCH3
15b
were synthesized as mixtures of diastereomers using literature
procedures. Purification into eq and ax isomers was ac-
complished by either preparatory gas chromatography (1-CH3,
3-CH3 and 1-F , 3-F ) or flash chromatography (1-Br , 3-Br 1-Cl,
3-Cl, and 1-OCH3, 3-OCH3). Structural assignments for the
purified ketones were based on comparisons with literature
spectral data (for X ) F, Cl, and Br, see Moreau et al.,27 for X
) OCH3, see Battioni et al.28).
Gen er a l P r oced u r e for R ea ct ion of Ket on es w it h
LiAlH4. To a 0.1 M solution of LiAlH4 in THF (prepared from
either a commercial 1.0 M solution of LiAlH4 in THF or from
solid LiAlH4) at 0 °C was added dropwise an equal volume of
a 0.1 M solution of ketone in THF. Assuming that all the
hydrides in LiAlH4 react, there are 4 equiv of hydride per
equivalent of ketone. The solution was stirred for 30 min at 0
°C, and then an excess of H2O was added cautiously. After
effervescence ceased, an excess of 10% H2SO4 was added in
one portion, followed by extraction with ether. The organic
layer was washed successively with 10% NaHCO3, water, and
brine, dried over Na2SO4, filtered, and concentrated under
reduced pressure. The resulting oil was then subjected to
analysis by either 1H NMR or gas chromatography.
Deter m in a tion of P r od u ct Ra tios. Reduction ratios for
1-H 1-CH3, and 3-CH3 were obtained by gas chromatography
using data from the literature to assign peaks (1-H29 and
1-CH3, 3-CH330).
Reduction products for other alcohols were identified and
quantified by 1H NMR. Peak assignments from Moreau et al.27
were used for 2-X and 4-X when X ) F, Cl, Br. For 2-OCH3
and 4-OCH3, initial assignments were taken from Battioni et
al.28
Unfortunately, the reduction results for 3-X cannot be
explained unambiguously. The electrostatic model is
effective in describing all compounds after a correction
for torsional effects is applied to 3-F . The Felkin-Anh
model is effective after a torsional correction for 3-F and
chelation adjustments to both 3-F and 3-OCH3 are
applied. We prefer the electrostatic model as it is not
dependent on chelation effects for 3-OCH3 and 3-F ,
effects not seen in the reduction of 1-X. Finally, the
Cieplak model fails to explain the high axial selectivity
seen in the reduction of 3-Cl and 3-Br relative to 3-CH3.
This failure cannot be corrected by adjustments made for
steric or chelation effects.
The reduction of 3-X was studied by ab initio calcula-
tions. Specifically, calculated selectivities for reduction
of 3-X by LiH, NaH, BH3, and AlH3 are listed in Table 4.
While all four models correctly reproduced the experi-
mental selectivities for 3-H and 3-CH3, they performed
significantly worse for the polar ketones. In particular,
calculations using LiH and NaH overestimated the
magnitude of the ∆G’s for 3-OCH3, 3-F , and 3-Cl by a
range of 1.20-2.60 kcal/mol. Use of BH3 and AlH3 was
somewhat more successful, correctly modeling 3-OCH3
and 3-Cl but performing very badly in the case of 3-F .
For 3-F , not only was the calculated ∆G off by over 1.5
kcal/mol but the calculated sign of ∆G was also incorrect.
Not surprisingly, the calculated qualitative ordering
for the reduction of 3-X did not mirror the experimental
data. Experimental reduction of 3-X gives less negative
∆G’s in X ) Cl , CH3, OCH3 < H, F. Both LiH and NaH
adequately describe the relative selectivities of 3-Cl,
3-CH3, and 3-H but are not close in ordering 3-F and
3-OCH3. Both BH3 and AlH3 do considerably better, with
AlH3 correctly ordering the whole series. Against this,
the problems of the calculated selectivity of 3-F by AlH3
are quite severe, as described above.
The above NMR peak assignments were confirmed by
isolation of the appropriate alcohols. Thus, reduction of 1-X,
where X ) F, Cl, Br, OCH3, gave a mixture of the alcohols
2a -X and 2e-X. The resulting mixture was separated into two
pure components by column chromatography. Similarly, 4a -
F , 4e-F , 4e-Cl, 4e-Br , and 4e-OCH 3 were synthesized by
(23) (a) Allinger, N. L.; Blatter, H. M.; Freiberg, L. A.; Karkowski,
F. M. J . Am. Chem. Soc. 1966, 88, 3002. (b) House, H. O.; Tefertiller,
B. A.; Olmstead, H. D. J . Org. Chem. 1968, 33, 935.
(24) Umemoto, T.; Tomita, K.; Kawada, K. In Org. Synth. 1990,
69, 129.
(25) Warnhoff, E. W.; Martin, D. G.; J ohnson, W. S. In Organic
Syntheses; Wiley: New York, 1963; Collect. Vol. IV, p 162.
(26) (a) Allinger, J .; Allinger, N. L. Tetrahedron 1958, 2, 64-74.
(b) Allinger, N. L.; Allinger, J . J . Am. Chem. Soc. 1958, 80, 5476-
5480.
Con clu sion s
Selectivity in the reduction of cis 2-X-tert-butylcyclo-
hexanones (1-X) is controlled by electrostatic interactions
between the nucleophile and substituent and not by steric
or chelation effects or hyperconjugation as described by
the theories of Felkin-Anh or Cieplak. For reduction of
(27) Moreau, P.; Casadevall, A.; Casadevall, E. Bull. Soc. Chim.
Fr. 1969, 2013-2020.
(28) Battioni, J .-P.; Chodkiewicz, W. Bull. Soc. Chim. Fr. 1977,
320-328.
(29) Hajos, A. Complex Hydrides; Elsevier: New York, 1979.
(30) Cherest, M. Tetrahedron 1980, 36, 1593-1598.

1700 J . Org. Chem., Vol. 66, No. 5, 2001
Rosenberg et al.
reduction of the appropriate ketone 3-X, followed by purifica-
tion by column chromatography. Alcohol 4a -Cl was synthe-
sized using the method of le Bel and Czaja.31 Alcohol 4a -Br
was prepared using the procedure of Palumbo et al.32 Alcohol
4a -OCH3 was synthesized by the addition of H2SO4/MeOH33
to cis-4-tert-butylcyclohexene oxide.27 The spectra for 2-OCH3
and 4-OCH3 differed from the literature and are reported
below.
sets.36 For ketones, second-order Moeller-Plesset perturbation
theory (MP2)37 optimizations were performed, while TS cal-
culations were done at the Hartree-Fock level. All stationary
points were confirmed with analytical second derivatives.
Single point calculations that employed MP2 theory with the
6-31++G** basis set were used to correct for electron correla-
tion and basis set deficiencies, respectively. Thus, ∆E(MP2/
6-31G*) + ∆E(HF/6-311+G**) - ∆E(HF/6-31G*) is used as an
approximation for a single point calculation at the MP2/6-
311+G** level of theory.
The ketones 3-Cl and 3-Br are reported to be thermally
unstable toward equilibration with the eq isomers at room
temperature, though indefinitely stable at dry ice tempera-
tures.34 We found that storage of these compounds at -20 °C
for periods of several weeks gave only a small percentage of
The computational modeling of LiAlH4 reductions of ketones
is complicated because “LiAlH4” itself does not give an accept-
able TS.38-40 Instead, reagents such as LiH,9ab,41 NaH,9b BH3,
9d,42
and AlH3
are used to model diastereoselectivity. Diaste-
1
eq ketone as judged by H NMR. Fortunately, the undesired
reoselectivities were calculated by subtracting the energy of
the TS that leads to the ax alcohol from the energy of the TS
that leads to the eq alcohol.
1H NMR shifts that resulted from reduction of the eq ketone
at 0 °C did not interfere with the peaks that came from
reduction of the ax ketone.
NMR Da ta . tr a n s-4-ter t-Bu tyl-tr a n s-2-m eth oxycyclo-
h exa n ol (2e-OCH3): 1H NMR (CDCl3, 300 MHz) δ 3.40 (s, 3),
3.35 (m, 1), 2.93 (m, 1), 2.61 (broad s, 1), 2.12 (m, 1), 2.00
(m, 1), 1.71 (m, 1), 1.24 (m, 2), 1.05 (m, 2), 0.86 (s, 9); 13C NMR
(CDCl3, 75 MHz) δ 85.9, 74.6, 57.0, 46.7, 32.9, 32.0, 29.9, 28.1,
25.4.
Ack n ow led gm en t. The authors would like to thank
Professors William le Noble and Peter Wipf for helpful
discussions. This work was supported by the Camille
and Henry Dreyfus Foundation, Research Corporation
(C-3694), the Petroleum Research Fund (32081-B4), and
the Geneseo Foundation.
cis-4-ter t-Bu tyl-cis-2-m eth oxycycloh exa n ol (2a -OCH3):
1H NMR (CDCl3, 300 MHz) δ 4.07 (m, 1), 3.38 (s, 3), 3.13
(m, 1), 2.15 (broad s, 1), 2.05 (m, 1), 2.00 (m, 1), 1.76 (m, 1),
1.31 (m, 3), 0.99 (m, 1), 0.85 (s, 9); 13C NMR (CDCl3, 75 MHz)
δ 82.2, 65.7, 56.3, 46.9, 32.9, 30.6, 28.0, 27.2, 20.4.
tr a n s-4-ter t-Bu tyl-cis-2-m eth oxycycloh exan ol (4e-OCH3):
1H NMR (CDCl3, 300 MHz) δ 3.52 (m, 1), 3.44 (m, 1), 3.34
(s, 3), 2.24 (broad s, 1), 2.12 (m, 1), 1.74 (m, 2), 1.49 (m, 1),
1.24 (m, 1), 0.97 (m, 2), 0.83 (s, 9); 13C NMR (CDCl3, 75 MHz)
δ 79.7, 72.1, 56.7, 40.3, 32.2, 30.9, 28.3, 27.9, 25.6.
cis-4-ter t-Bu tyl-tr a n s-2-m eth oxycycloh exan ol (4a-OCH3):
1H NMR (CDCl3, 300 MHz) δ 3.87 (m, 1), 3.38 (m, 1), 3.32
(s, 3), 1.74 (m, 4.3), 1.47 (m, 2), 1.30 (m, 4.2), 0.83 (s, 9); 13C
NMR (CDCl3, 75 MHz) δ 80.0, 67.3, 56.8, 41.0, 32.5, 29.1, 27.7,
25.4, 20.7.
J O0011787
(35) Gaussian 94 (Revision B. 3). Frisch, M. J .; Trucks, G. W.;
Schlegel, H. B.; Gill, P. M. W.; J ohnson, G.; Robb, M. A.; Cheeseman,
J . R.; Keith, T. A.; Petersson, G. A.; Montgomery, J . A.; Ragavachari,
K.; Al-Laham, M. A.; Zakrzewski, V. G.; Ortiz, J . V.; Foresman, J . B.;
Cioslowski, J .; Stefanov, B. B.; Nanayakkara, A.; Challacombe, M.;
Peng, C. Y.; Ayala, P. Y.; Chen, W.; Wong, M. W.; Andres, J . L.;
Repogle, E. S.; Gomperts, R.; Martin, R. L.; Fox, D. J .; Binkley, J . S.;
Defrees, D. J .; Baker, J .; Stewart, J . P.; Head-Gordon, M.; Gonzalez,
C.; Pople, J . A. Gaussian, Inc. Pittsburgh, Pa 1995.
(36) For a complete description of the basis sets, see: Hehre, W.
J .; Radom, L.; Schleyer, P. v. R.; Pople, J . A. Ab Initio Molecular Orbital
Theory; Wiley-Interscience: New York, 1986.
(37) Moller, C.; Plesset, M. S. Phys. Rev. 1934, 46, 618.
(38) Bonaccorsi, R.; Cimiraglia, R.; Tomasi, J .; Miertus, S. J . Mol.
Struct. 1983, 94, 11. Bonaccorsi, R.; Palla, P.; Tomasi, J . J . Mol. Struct.
1982, 87, 181.
Com p u ta tion a l Meth od ology. Calculations were per-
formed using the Gaussian 94 program suite.35 In short,
geometries are optimized using the 3-21G and 6-31G* basis
(39) Eisenstein, O.; Schlegel, H. B.; Kayser, M. M. J . Org. Chem.
1982, 47, 2886.
(40) Recently, Tomoda (ref 8e) showed that “LiAlH4” did give an
acceptable TS for cyclohexanone itself. However, our preliminary
computational data for 2-substituted cyclohexanones using this method
gave no improvement over the data reported.
(41) (a) Kaufman, E.; Schleyer, P. v. R.; Houk, K. N.; Wu, Y.-D. J .
Am. Chem. Soc. 1985, 107, 5560-5562. (b) Shi, Z.; Boyd, R. J . J . Am.
Chem. Soc. 1993, 115, 9614-9619. (c) Wong, S. S.; Paddon-Row; M.
N. J . Chem. Soc., Chem. Commun. 1991, 327-330.
(42) Coxon, J . M.; Luibrand, R. T. Tetrahedron Lett. 1993, 34, 7097-
7100.
(31) le Bel, N. A.; Czaja, R. F. J . Org. Chem. 1961, 26, 4768-4770.
(32) Palumbo, G.; Ferreri, C.; Caputo, R. Tetrahedron Lett. 1983,
24, 1307-1310.
(33) Winstein, S.; Ingraham, L. L. J . Am. Chem. Soc. 1952, 74, 1160.
(34) For a x 2-Br : Allinger, N. L.; Allinger, J . J . Am. Chem. Soc.
1958, 80, 5476-5480. For a x 2-Cl: Allinger, N. L.; Allinger, J .;
Freiberg, L. A.; Czaja, R. F.; LeBel, N. A. J . Am. Chem. Soc. 1960, 82,
5876.