Why Are Silyl Ethers Conformationally Different from Alkyl Ethers? Chair-Chair Conformational Equilibria in Silyloxycyclohexanes and Their Dependence on the Substituents on Silicon. The Wider Roles of Eclipsing, of 1,3-Repulsive Steric Interactions, and o

Article abstract of DOI:10.1021/ja035936x

An NMR study of the diaxial/diequatorial chair equilibrium in a range of silylated derivatives of trans-1,4- and trans-1,2-dihydroxycyclohexane is reported and discussed with a view to explaining unusually large populations of chair conformations with axi

Full text of DOI:10.1021/ja035936x

Published on Web 11/13/2003
Why Are Silyl Ethers Conformationally Different from Alkyl
Ethers? Chair-Chair Conformational Equilibria in
Silyloxycyclohexanes and Their Dependence on the
Substituents on Silicon. The Wider Roles of Eclipsing, of
1,3-Repulsive Steric Interactions, and of Attractive Steric
Interactions
,‡
Cecilia H. Marzabadi,*,† J. Edgar Anderson,*,‡ Jorge Gonzalez-Outeirino,
Piers R. J. Gaffney,‡ Christopher G. H. White,‡ Derek A. Tocher,‡ and
Louis J. Todaro§
Contribution from the Department of Chemistry and Biochemistry, Seton Hall UniVersity,
400 South Orange AVenue, South Orange, New Jersey 07079, Chemistry Department, UniVersity
College, Gower Street, London WC1E 6BT United Kingdom, and Chemistry Department, Hunter
College, 695 Park AVenue, New York, New York 10021
Received May 3, 2003; E-mail: marzabce@shu.edu; j.e.anderson@ucl.ac.uk
Abstract: An NMR study of the diaxial/diequatorial chair equilibrium in a range of silylated derivatives of
trans-1,4- and trans-1,2-dihydroxycyclohexane is reported and discussed with a view to explaining unusually
large populations of chair conformations with axial substituents, noted previously for some monosilyloxy-
cyclohexanes and in some silylated sugars. X-ray diffraction studies of three bis-triphenylsilyloxycyclohex-
anes are reported and show both axial and equatorial silyloxy groups with the exocyclic bonds eclipsed.
Eclipsing is also suggested by molecular mechanics (MM3) calculations on such derivatives. Both axial
and equatorial tertiary silyl groups have 1,3-repulsive interactions with whatever substituents or hydrogen
atoms are at the two adjacent equatorial positions, and these are relieved by rotation toward the eclipsed
conformation of the exocyclic C-O bond. The three substituents on silicon interact attractively with the
nine atoms at the 3, 4, and 5-positions of the cyclohexane ring and calculations suggest that these stabilizing
interactions are significantly greater in the axial than in the equatorial conformation. An equatorial C-OSiR3
bond with one or two equatorial neighbors has a restricted potential energy well that becomes much broader
when the bond is axial without any equatorial neighbors in the alternative chair. Adjacent silyl groups in the
1,2-disubstituted series interact in a stabilizing way overall in all conformations, this being particularly marked
in the diaxial conformation of the more complex ethers. These factors lead to unusually large axial
populations.
Introduction
A fundamental set of conformational results, which still awaits
explanation, comes from the work of Eliel and Satici2 on
Trialkyl(or aryl)silyl [R1R2R3Si-] derivatives of alcohol
functions, henceforth called silyl derivatives, are of great
practical use in protection strategies in organic synthesis.1
Several aspects of the conformations of such derivatives of six-
membered ring alcohols including pyranose sugars are strikingly
unusual,2-8 and have been usefully exploited in synthesis.5,8
silyloxycyclohexanes 1 and their axial-equatorial equilibria
(Scheme 1). The greater the bulk of the substituents on silicon,
the greater the population of the chair conformation with the
single silyloxy substituent axial. The authors reported that
preliminary molecular mechanics calculations confirmed the
trend of the results but gave no details.
The second conformational oddity is that of O-silylated sugar
derivatives,4-8 where it is found that the chair conformation
preferred is the one with several -O-SiR1R2R3 and other
groups axial, while the same sugar with hydroxy groups
† Seton Hall University.
‡ University College.
§ Hunter College.
Current address: CCRCsUniversity of Georgia, 220 Riversbend Road,
Athens, GA 30602.
(1) Fleming, I.; Barbero, A.; Walter, D. Chem. ReV. 1997, 97, 2063-2192.
(2) Eliel, E. L.; Satici, H. J. Org. Chem. 1994, 59, 688-9.
(3) (a) Nagao, Y.; Goto, M.; Ochiai, M.; Shiro, M. Chem. Lett. 1990, 1503-
1506. (b) Romero, J. A. C.; Tabacco, S. A.; Woerpel, K. A. J. Am. Chem.
Soc. 2000, 122, 168-169.
(6) Walford, C.; Jackson, R. F. W.; Rees, N. H.; Clegg, W.; Heath, S. L. J.
Chem. Soc. Chem. Commun. 1997, 1855-56.
(7) Roush, W. R.; Sebesta, D. P.; Bennett, C. E. Tetrahedron 1997, 53, 8825-
8836.
(4) Hosoya, T.; Ohashi, Y.; Matsumoto, T.; Suzuki, K. Tetrahedron Lett. 1996,
37, 663-666.
(8) (a) Ichikawa, S.; Shuto, S.; Matsuda, A. J. Am. Chem. Soc. 1999, 121,
10 270-10 280. (b) Shuto, S.; Terauchi, M.; Yahiro, Y.; Abe, H.; Ichikawa,
S.; Matsuda, A. Tetrahedron Lett. 2000, 41, 4151-4155. (c) Matsuda, A.
J. Am. Chem. Soc. 2001, 123, 11 870-11 882.
(5) Koslowski, J.; Marzabadi, C.; Rath, N.; Spilling, C. Carbohydr. Res. 1997,
300, 301-313.
9
10.1021/ja035936x CCC: $25.00 © 2003 American Chemical Society
J. AM. CHEM. SOC. 2003, 125, 15163-15173
15163

A R T I C L E S
Marzabadi et al.
Scheme 1. Chair-Chair Interconversion of Silyloxycyclohexanesa
a Conformational equilibrium in CD2Cl2 solution at about 200 K. The same trend is shown for toluene solutions but with a greater population of the axial
conformation.2
Scheme 2. Preferred Ring Conformation of Different O-Protected
Olivosides 2 and 3, with Protecting Groups R in Plausible
Locations4
Scheme 3. trans-1,4- and trans-1,2-Bis-silyloxycyclohexanes
Derivatives
protected by alkyl substituents adopts the other, ring-inverted,
chair conformation with equatorial substituents. The contrasting
equilibria of compounds 2 and 3 (Scheme 2), show an early
example.4 The axial preference in 2 is attributed to the extra
steric interaction of the three substituents on adjacent tertiary
silyloxy groups when these are equatorial (as compared with
the benzyl protecting group, which is primary). Clearly,
Scheme 4. Chair-Chair Equilibria in trans-1,4- and
trans-1,2-Bis-silyloxycyclohexanes
4
however, the substituents on silicon in the C1 chair need not
be near each other, and the 1C4 chair has a methyl/oxygen and
other 1,3-diaxial interactions; the essence of Eliel and Satici’s
results2 is that even when there is a lone silyloxy group,
increasing substituent bulk leads to more of the axial conforma-
tion.
There is thus an unusual steric effect, and to examine it
further, we now want to report chair-chair conformational
equilibria in a range of bis-silyloxycyclohexanes, with the
substituents either trans-1,4-, 4, or trans-1,2-, 5 (Scheme 3).
The trans-1,4-disubstituted set 4a-4f allows us to extend
Eliel and Satici’s work without concerns for steric interactions
of the substituents, while the trans-1,2-disubstituted set 5a-5f
allows us to investigate interactions of adjacent silyloxy groups.9
In both cases the chair-chair equilibrium is between the
diequatorial conformation (6 or 8) and the diaxial conformation
(7 or 9) (Scheme 4).
We report dynamic NMR measurements of the series 4 and
5 in deuterodichloromethane solution, MM3 calculations of
these and series 1, and crystal structure data for three bis-silyloxy
compounds.
We hope to clarify the nature of the steric effect by a
systematic study of these two series, but clearly, disubstitution
introduces perhaps quite subtle stereoelectronic effects. This is
(9) Calculations for 5 are for the (R,R) configuration shown in Scheme 3, and
drawings, torsion angles, and discussion will be in terms of this enantiomer,
but experimental results are in fact for racemic 5a-f.
9
15164 J. AM. CHEM. SOC. VOL. 125, NO. 49, 2003

Conformational Equilibria in Silyloxycyclohexanes
A R T I C L E S
Table 1. Chair-Chair Conformational Equilibria in Mono- and trans-1,4-Bis-silyloxycyclohexanes 4
trans-1,4-bis-silyloxycyclohexane
our calculation
monosilyloxycyclohexane (Eliel and Satici’s experiments)
our experiments
d
silyl ether
number of atoms in R
% axial
A-valuea,d
% diaxialb
difference in final steric energyc
% diaxiala
∆G
3
4a OSiMe3
4b OSiEt3
4c OSiMe2-t-Bu
4d OSi(i-Pr)3
4e OSiPh3
12
21
21
30
33
35
3.6
4.0
6.5
8.6
14.3
19.6
1.31
1.26
1.06
0.94
0.71
0.56
8
12
23
22
15
23
0.87
0.70
0.44
0.46
0.62
0.43
2.0
0.8
0.7
19
17
23
1.39
1.72
1.77
0.52
0.57
0.43
4f OSiPh2-t-Bu
a kcal/mol, free energy differences were determined experimentally at various low temperatures and have been standardized to give population values at
200 K by assuming ∆S° ) 0. Solvent was CD2Cl2. b At 180 K, based on differences in the final steric energy, calculated assuming entropies are zero.
c kcal/mol, for the most stable calculated axial and equatorial conformations, see text discussion on effect of conformations within the SiR3 group. d kcal/mol
at 180 K, axial less stable.
illustrated by the quite marked solvent dependence of the
equilibrium in the obvious alkoxy analogues of the series 4 and
5, the dimethoxycyclohexanes 10 and 11, where nonpolar
solvents favor diaxial conformations.10-12
than the H-methyl eclipsing interactions (see 14, for example,
an alternative view of 12).
More interestingly for the present work, compound 15 with
a tert-butoxy group, but only hydrogen atoms adjacent, also
adopts an eclipsed conformation, (O-tert-butyl eclipsing C-H;
see 16), plausibly for the same reasons.15
Results such as these are reminders that the equilibrium for
vicinal disubstituted cyclohexanes 5 involves several factors that
may differ between the two chair conformations.10-13 The
inherent equatorial preference of a substituent and the stabilizing
gauche effect in an O-C-C-O fragment both favor the
diequatorial conformation. On the other hand, that conformation
is disfavored by the steric repulsion of adjacent methoxy groups
and the destabilizing interaction of carbon-oxygen bond
dipoles. By observing our sets of compounds under uniform
conditions, we hope to elucidate the particular, unusual steric
factor in silyl ethers that is associated with the size of the
substituents on silicon.
Two other relevant features of the alkoxycyclohexane study14,16
are worth noting. When the alkoxy group has two equatorial
neighboring substituents of different size, the rotational mini-
mum has the alkoxy group somewhat displaced from eclipsed
on the less-hindered side. Also, there is an intermediate level
of neighboring interactions where the calculated potential energy
diagram for the exocyclic C-O bond shows either inflections
or even shallow minima on either side of eclipsed.16 Here, in
the wider view, there is a broad potential well, centered on
eclipsed.
There are earlier results on substituted mono-alkoxycyclo-
hexanes14 that particularly help an understanding of the results
we now present. NMR coupling constant measurements show
that methoxycyclohexanes 12 and 13 with equatorial substitu-
ents on either side of the methoxy group adopt an eclipsed
conformation for the exocyclic C-O bond, with O-Me
eclipsing C-H. Molecular mechanics calculations agree with
this minimum energy bond conformation, and it has been
suggested14 that it arises because O-methyl-C-methyl repulsions
in the conventional staggered conformation are more substantial
Results
NMR Spectra. Each of the compounds 4a-4f and 5a-5f
shows two sets of signals of different intensity at low temper-
atures when chair-chair interconversion is slow, both in the
proton and carbon-13 NMR spectra. All population measure-
ments for these series are from integration of proton NMR
spectra recorded at ∼-95 °C in deuterodichloromethane solu-
tion. Carbon-13 NMR spectra were not used to determine the
position of equilibria but showed changes in agreement with
proton NMR spectra. No signals or other spectral changes were
seen that might suggest that nonchair conformations contribute
to the equilibria.
(10) (a)Hammarstro¨m, L.-G.; Berg, U.; Liljefors, T. J. Chem. Res. (S), 1990,
152. (b) Eliel, E. L.; Wilen, S. H. Stereochemistry of Organic Compounds;
Wiley: New York, 1994; Chapter 11.
(11) Rockwell, G. D.; Grindley, T. B. Aust. J. Chem. 1996, 49, 379.
(12) (a) Zefirov, N. S.; Samoshin, V. V.; Subbotin, O. A.; Baranenkow, V. I.;
Wolfe, S. Tetrahedron 1978, 34, 2953. (b) Zefirov, N. S.; Samoshin, V.
V.; Palyulin, V. A. Zh. Org. Khim. 1983, 19, 1888. (c) Zefirov, N. S.;
Zelenkina, O. A.; Subbotin, O. A.; Samoshin, V. V. Zh. Org. Khim. 1987,
23, 1122.
(13) (a) Kirby, A. J. The Anomeric Effect and Related Stereoelectronic Effects
at Oxygen; Springer-Verlag: Berlin, 1983. (b) Deslongchamps, P. Stereo-
electronic Effects in Organic Chemistry; Pergamon: Oxford, 1983. (c)
Juaristi, E.; Cuevas, G. Tetrahedron 1992, 48, 5019. (d) Thatcher, G. R. J.
The Anomeric Effect and Associated Stereoelectronic Effects; ACS
Symposium Series; American Chemical Society: Washington, DC, 1993.
(e) Graczyk, P. P.; Mikolajczyk, M. Topics Stereochem. 1994, 21, 159.
(14) Anderson, J. E.; Ijeh, A. I. J. Chem. Soc., Perkin Trans. 2 1994, 1965.
The 1,4-Disubstituted Cyclohexanes (Table 1). The low-
temperature NMR data for the trans-1,4-bis-silyloxycyclohex-
anes 4a-4f show the position of the equilibrium between the
chair conformations 6 and 7 (Table 1).
(15) Anderson, J. E.; Khan, S., unpublished results cited elsewhere.13
(16) Anderson, J. E. J. Org. Chem. 1996, 61, 3511-3513.
9
J. AM. CHEM. SOC. VOL. 125, NO. 49, 2003 15165

A R T I C L E S
Marzabadi et al.
Table 2. Chair-chair Conformational Equilibria in trans-1,2-Bis-silyloxycyclohexanes 5
trans-1,2-bis-silyloxycyclohexanes
our calculations
H−C−O−Si,e deg
our experients
excess SiR /OSiR
difference in final
steric energyc
3
3
f
d
silyl ether
no. of atoms in R
% diaxialb
in eq cf ax
axial
equatorial
% diaxiala
∆G
3
5a OSiMe3
5b OSiEt3
5c OSiMe2-t-Bu
5d OSi(i-Pr)3
5e OSiPh3
12
21
21
30
33
35
0.03
0.04
0.03
28.0
6.4
0.87
1.21
1.44
0.73
-0.49
-2.49
2.87
2.25
2.61
0.33
0.96
0.7, 1.1
19.4, 21.7
0.2, 0.1
6.4, 5.3
35.0, 34.8
31.9, 34.8
12.2, 12.2
20.1, 15.7
18.4, 18.4
32.1, 32.1
6.1, 31.9
1
5
11
95
32
87
1.64
1.05
0.75
-1.05
0.27
5f OSiPh2-t-Bu
99.4
-1.84
29.7, 35.3
-0.66
a kcal/mol, free energy differences were determined experimentally at various low temperatures and have been standardized to give population values at
200 K by assuming ∆S° ) 0. Solvent was CD2Cl2. b At 180 K, based on differences in the final steric energy, calculated assuming entropies are zero.
c Diaxial and diequatorial, kcal/mol, the negative value means diaxial is more stable. See text discussion on effect of conformations within the SiR3 group.
d kcal/mol at 180 K, axial less stable. e Torsion angles for structures 8 (diequatorial) and 9 (diaxial).9 f These refer to net attractive interactions between silyl
groups.
The chair conformation with equatorial substituents is pre-
ferred, but for the first three of the compounds, 4a-4c, the
tendency to populate the diaxial chair conformation is dimin-
ished compared with series 1. For the remaining three com-
pounds 4d-4f, however, the diaxial tendency is enhanced. With
the larger substituents on silicon, the population of the axially
substituted chair conformation is uncommonly high, in keeping
with the earlier work for monosilyloxycyclohexanes reported
in the Introduction.
The 1,2-Disubstituted Cyclohexanes (Table 2). There is a
surprising range of results for the equilibrium between chair
conformations 8 and 9 for the trans-1,2-bis-silyloxycyclohexanes
5a-5f.
The trimethylsilyl derivative 5a is the only compound to show
a decrease in the proportion of the diaxial conformation
compared with monosubstituted. The remaining compounds 5
Figure 1. ORTEP diagram of the diequatorial conformation of trans-1,4-
bis-triphenylsilyloxycyclohexane (4e).
accentuate the results for the monosubstituted series 1, showing
an increased population of the diaxial conformation that is
particularly striking for compounds 5d and 5f.
Eliel and Satici reported2 that the proportion of the axial
conformation in series 1 is greater in toluene solution than in
methylene chloride. We examined several compounds in the
series 4 and 5 and observed similar relatively small increases.
The proportion of diaxial conformation of compound 4d, for
example, increases from 19% in CD2Cl2 to 29% in toluene.
and equatorial silyloxy groups, and two different molecules are
found in the unit cell. For the trans-1,2-isomer 5e whose CP
MAS NMR is discussed above, the unit cell contains 12
molecules, and both diaxial and diequatorial chair conformations
are found (see Supporting Information).9
Since compound 5e is crystalline at room temperature,
CPMAS solid-state proton decoupled carbon-13 NMR spectra
were recorded and, surprisingly, fiVe signals of intensity ratio
∼1:1:2:1:1 were seen at room temperature at around 68-78
ppm, the region of the silyloxygenated C1 and C2 carbon atoms.
Less readily assigned multiple signals were seen in the remainder
of the spectrum. This suggests that more than one conformation
may be present in the crystal, so an X-ray diffraction study of
the compound was carried out, and the results obtained confirm
this.
When triphenylsilyloxy groups have no neighbors, as with
the axial groups in 4e, the exocyclic C-O bonds are nearly
eclipsed with H-C-O-Si torsion angles of 13°. In the two
different molecules of compound 17, the torsion angles are 21°
and 13° for axial silyl groups and 21° and 13° for equatorial
silyl groups.
One other crystalline triphenylsilyloxy cyclohexane derivative
with no adjacent substituents has been the subject of an X-ray
diffraction study. In O-triphenylsilyl-17R-mercaptocholestan-
3R-ol,17 the triphenylsilyloxy group is axial and the unit cell
contains two different molecules, both of which show a marked
eclipsing tendency with H-C-O-Si torsion angles of 4° and
12°. A search of the Cambridge Crystallographic Data Base18
yielded 36 further examples of other kinds of silyloxy groups
Crystal Structures. The X-ray diffraction crystal structures
of three bis-triphenylsilyloxycyclohexanes have been deter-
mined, namely, trans-1,4-bis-triphenylsilyloxycyclohexane 4e,
its cis epimer 17, and the trans-1,2-isomer 5e. Structures for
the 1,4-disubstituted isomers are shown in Figures 1 and 2 and
are described in detail in the Supporting Information. The trans-
1,4-isomer 4e adopts the chair conformation with two axial
silyloxy substituents (i.e., the minor populated conformation in
solution). The cis-1,4-disubstituted compound 17 has both axial
(17) Wiedenfeld, D. J. Chem. Soc. Perkin Trans. 2 1997, 339.
9
15166 J. AM. CHEM. SOC. VOL. 125, NO. 49, 2003

Conformational Equilibria in Silyloxycyclohexanes
A R T I C L E S
Figure 2. ORTEP diagram of cis-1,4-bis-triphenylsilyloxycyclohexane (17).
on a saturated six-membered ring with no adjacent equatorial
substituents. Along with the seven examples from the present
work, such silyloxycyclohexanes with a range of silicon
substituents but no adjacent equatorial substituents on the ring
have a mean torsion angle H-C-O-Si of 15.6°, with an
average value almost 3° lower for axial silyloxy groups than
for equatorial. The sample is large enough to suggest that such
near-eclipsing of exocyclic C-O bonds is not an artifact of the
crystalline state.
Trans-1,2-disubstituted compounds 5 are chiral,9 and the
racemic form of compound 5e crystallizes in the monoclinic
system. The unit cell contains 12 molecules, four each of three
kinds (Supporting Information). The first of these three is in a
diaxial conformation with two molecules of each enantiomer.
The H-C-O-Si torsion angles found for axial OSiPh3 groups
are 0° and 10°. The other two kinds of molecules have
diequatorial conformations of opposite configuration, there being
four of each in the unit cell.
torsion angle averages 31.3° with the silyl group turned away
from the substituent, agreeing well with the four values reported
for equatorial 5e. The absence of one adjacent substituent thus
produces torsion angles that are 15° further from eclipsing on
average.
Molecular Mechanics Calculations. For molecular mechan-
ics calculations of the mono- and bis-silyloxy cyclohexane series
1, 4, and 5, we used Allinger’s MM3(94) program whose
parametrization for ethers and for silanes but not for silyloxy
compounds has been discussed.19,20 Such calculations should
be broadly reliable for indicating trends in the axial/equatorial
population ratio along a series of compounds. The anti
conformation for the exocyclic C-O bond in the equatorial
conformation of compound 1a (i.e., H-C-O-Si ∼ 180°) is
more than 3 kcal/mol less stable than the minimum-energy
eclipsed conformation, while the anti, axial conformation is even
more unstable than eclipsed, axial. Similar results are calculated
for all silyl groups. Nonchair conformations of the cyclohexane
ring are even less stable by calculation, so neither those
conformations nor anti conformations about the C-O bond will
be further discussed.
Finding the minimum energy structure for the mono- and
bis-silyoxycyclohexanes requires that the conformation about
each of the three or six Si-R bonds be properly investigated,
so that a total energy for the molecule does not reflect a less-
than-optimum conformation within an SiR3 group. This is
especially a problem in compounds with multiple ethyl, iso-
propyl, or phenyl groups. In 1d for example, there are three
staggered conformations for each isopropyl-silicon bond, so
33 ) 27 combinations of conformation, within a single tris-
isopropylsilyl group. Such conformational possibilities have
been discussed in detail previously.21a,b Investigation of the tert-
The two diequatorial structures in the crystal of 5e have bond
lengths, bond angles, and torsion angles which are similar but
different numerically.9 Equatorial triphenylsilyl groups have very
different substituents at the two neighboring equatorial positions,
hydrogen and triphenylsilyloxy, respectively, so none of the four
different equatorial SiPh3 groups are eclipsed but rather turned
away from its neighboring OSiPh3 group toward the unsubsti-
tuted ring position, with H-C-O-Si torsion angles all in the
range |26|° - |31|°.
The Crystallographic Data Base yielded a further 22 examples
of silyloxycyclohexanes or analogues with a single, adjacent,
equatorial substituent,18 and in these cases the H-C-O-Si
(18) The Development of Versions 3 and 4 of the Cambridge Structural Database
System. Allen, F. H.; Davies, J. E.; Galloy, J. J.; Johnson, O.; Kennard,
O.; Maccrae, C. F.; Mitchell, E. M.; Mitchell, G. F.; Smith, J. M.; Watson,
D. G. J. Chem. Inf. Comput. Sci. 1991, 31, 187-204. July 1991 CSD
Release.
(19) Allinger, N. L.; Rahman, M.; Lii, J.-H. J. Am. Chem. Soc. 1990, 112, 8293.
(20) Chen, K.; Allinger, N. L. J. Phys. Org. Chem. 1997, 110, 697.
9
J. AM. CHEM. SOC. VOL. 125, NO. 49, 2003 15167

A R T I C L E S
Marzabadi et al.
Table 3. Calculated Relative Energy ∆EAx (kcal/mol) of the Axial
Chair Conformation of Silyloxycyclohexanes 1, Equatorial ) 0.0,
monosubstituted series, and in fact, in the trisisopropyl case the
difference is slightly less. It is notable that the exocyclic torsion
angle H-C-O-Si tends toward eclipsed, varying irregularly
in size with the silyl group, and that it is similar in size in the
axial and equatorial chair conformations of the same compound.
The phenylsilyl compounds 1e, 1f, 4e, and 4f are calculated to
have the least tendency toward eclipsing, although they are
hardly staggered (τ ) 20.3-28.5°). Crystal structures reported
above show torsion angles H-C-O-SiPh3 much nearer
eclipsed.
and the Exocyclic Torsion Angles H-C-O-Si τ
and τ Ax, deg
Eq
∆EAx
substituent SiR
τ
eq(H−C−O−Si)
MM3 calcd
experimental
τax(H−C−O−Si)
3
a, SiMe3
3.8
12.4
5.8
0.61
0.51
0.41
0.48
0.46
0.42
1.31
1.26
1.06
0.94
0.71
0.56
1.4
9.8
b, SiEt3
c, Si-t-BuMe2
d, Si(i-Pr)3
e, SiPh3
3.7
6.5
7.5
23.2
26.7
20.3
25.9
f, SiPh2-t-Bu
Molecular mechanics energy-minimized structures do not
necessarily show the strain present in the idealized staggered
chair conformations but rather show how the molecule has
deformed to accommodate such strain. Nonetheless, calculations
for series 1, 4, and 5 reproduce the trends of the experimental
results; so to see if an explanation of the trends appears, we
looked at the complete set of pairwise atom-atom interactions,
repulsive or attractive, that MM3 provides. We report on each
series in turn.
Calculations for Monosilyloxycyclohexanes 1a-f, Table
5. The parent silyloxycyclohexane 1, SiR3 ) SiH3, has been
included, since it extends the trend in the series 1a-f insofar
as the overall axial/equatorial energy difference, 0.66 kcal/mol,
is greater than that calculated for all the more substituted
compounds.
The conventional explanation of the excess energy of the axial
conformation over the equatorial is the interaction of the axial
oxygen atom with C3 and C5 and the axial hydrogen atoms at
these positions. These four interactions are indeed repulsive but
are almost the same in all compounds 1a-f, totalling in the
narrow range 0.98-1.02 kcal/mol. This suggests that we
examine whether differences along the series reflect the interac-
tion of the ring with what is beyond the oxygen, i.e., with the
atoms of the SiR3 group. The interactions of C1 and H1 with
SiR3 are predominantly repulsive and are remarkably similar
in both chair conformations. Their total is always more than 1
kcal/mol, but the difference between the two axial and equatorial
chair conformations is always less than 0.1 kcal/mol and usually
much less. Notably, the repulsions with C1 and H1 are least
for 1e and 1f, but this plausibly reflects the fact that their
minimum-energy conformations are quite far from eclipsed
(Table 3).
Table 4. Calculated Relative Energy ∆EAx (kcal/mol) of the Diaxial
Chair Conformation of trans-1,4-Bis-silyloxycyclohexanes 4,
Diequatorial ) 0.0, and the Exocyclic Torsion Angles H-C-O-Si,
τeq and τax, deg
∆E
ax
substituent SiR
τ
eq(H−C−O−Si)
MM3 calcd
experimental
τax(H−C−O−Si)
3
a, SiMe3
4.4
14.1
3.6
0.87
0.70
0.44
0.46
0.62
0.43
1.55
1.91
1.97
0.49
0.63
0.48
1.9
11.4
2.1
b, SiEt3
c, Si-t-BuMe2
d, Si(i-Pr)3
e, SiPh3
3.8
7.3
25.2
28.5
27.1
27.8
f, SiPh2-t-Bu
butyldimethylsilyl group and the tert-butyldiphenylsilyl group,
as in 1c and 1e, consistently showed that the conformation with
the tert-butyl group antiperiplanar to the O-cyclohexyl bond is
most stable. Interestingly, the conformation within SiR3 groups
is essentially the same by calculation in the axial and equatorial
minima. However, as the viewpoint used in 18 and 19 shows,
the geometry of the interaction of SiR3 with C2 and C6 and
their attached equatorial hydrogens is similar in the two chair
conformations.
There are remarkably few repulsive interactions of SiR3 with
the flanking CH2 groups at positions 2 and 6, in the minimum
energy conformations with near-to-eclipsing C-O bonds, il-
lustrating how effective eclipsing is in relieving these. Once
again the total of such interactions is very similar in the two
chair conformations of any compound, differing by 0.00-0.07
kcal/mol in the series.
Thus overall, interactions of SiR3 with C1, C2, and C6 and
their attached hydrogens are repulsive, but as the second to last
column of Table 5 shows, there is little difference in the total
of these interactions in axial and equatorial conformations.
Idealized structures 18 and 19 emphasize the geometric similar-
ity of conformations, and while there are small differences in
calculated bond lengths, bond angles, torsion angles, and
interatomic distances, the net interaction of SiR3 with this part
of the molecule is very similar in the two conformations.
The remaining interactions of atoms of the SiR3 group, which
might produce differences between axial and equatorial con-
Tables 3 and 4 show calculated overall energies for minimum
energy axial and equatorial conformations for the series 1 and
4, and the exocyclic H-C-O-Si torsion angles associated with
these.
The axially substituted chair conformation is always calcu-
lated to be the less stable but the axial/equatorial energy
difference is always underestimated compared with experimen-
tally determined values. Along the series, however, the calcu-
lated and experimental trends match. In particular, the relative
stability of the axial conformation is greater when the silicon
substituents R contain a large number of atoms. The calculated
diaxial/diequatorial energy differences for the trans-1,4-disub-
stituted series 4 are only a little greater than those of the
(21) (a) For discussion of the conformations within a trisisopropylsilyl group,
for example: Anderson, J. E.; Casarini, D.; Lunazzi, L.; Mazzanti, A. J.
Org. Chem. 2000, 65, 1729-1737. (b) Anderson, J. E.; Koon, K. H.; Parkin,
J. E. Tetrahedron 1985, 41, 561-567.
9
15168 J. AM. CHEM. SOC. VOL. 125, NO. 49, 2003

Conformational Equilibria in Silyloxycyclohexanes
A R T I C L E S
Table 5. Difference in MM3-Calculated Interactions (kcal/mol) of SiR3 with Parts of the Cyclohexane Ring in Minimum Energy Equatorial
and Axial Conformations of Monosubstituted Compounds 1a
monosilyloxycyclohexane (1a-f)
difference between total energy
difference between total energy of axial and
equatorial conformaions as to interactions of
difference between total energy of axial and
equatorial conformations as to interactions of SiR
of axial and equatorial conformations
no. of
atoms in R
3
silyl ether
OSiH3
OSiMe3
OSiEt3
OSiMe2-t-Bu
OSi(i-Pr)3
OSiPh3
experimental2
MM3 calcd
SiR atoms with C6H2C1HC2H2 atoms
3
atoms with C3H2C4HC5H2 atoms
3
3
12
21
21
30
33
35
0.66
0.61
0.51
0.41
0.48
0.46
0.42
0.04
0.00
0.09
-0.05
0.06
0.15
-0.04
-0.18
-0.23
-0.43
-0.49
-0.48
-0.35
-0.37
1.31
1.26
1.06
0.94
0.71
0.56
OSiPh2-t-Bu
a A negative sign means that interactions are more stabilizing in the axial conformation.
formations, are with the atoms of the three methylene groups
at the far end of the cyclohexane ring. Unsurprisingly, these
pairwise interactions are all attractive though small. Even in
the simplest trimethylsilyl compound 1a, there are already 117
such interactions between the nine atoms of CH2 CH2CH2 and
the thirteen atoms of Si(CH3 )3, so their total may be substantial.
Viewing the diagrams 18 and 19, the relative arrangements
of these two fragments is different in the two chair conforma-
tions, and the calculation that such stabilization is greater in
the more compact axial conformation than in the extended
equatorial conformation confirms this. Over the six compounds
1a-f, the average size of the attraction between one atom from
each fragment is 4.4 and 2.7 cal/mol, respectively, in axial and
equatorial conformations.
There are hundreds of such pairwise interactions, however,
in each compound 1, and the net results are shown in the right-
hand column in Table 5. It is striking that these long-range
attractive interactions contribute several hundred calories per
mole toward stabilizing the axial conformation and the contribu-
tion increases erratically down the series. The corresponding
calculations for series 4 and 5 yield similar results.
values found in series 1; this minimizes repulsive interactions.
The diaxial conformations, 5a and 5c are eclipsed, but for the
other compounds (5b, 5e, and 5f), the silyl groups are turned
toward each other, presumably to maximize attractive interac-
tions.
We do not show the numbers, but the sum of attractive inter-
actions of silyl groups with the far end of the cyclohexane ring
is once again greater in the diaxial conformation, is more marked
than in the monosubstituted series, and makes an increasing
contribution to the equilibrium along the series 5a-f.
What is more inconsistently variable is the interaction of a
silyl group with its neighboring silyloxy group in the two
conformations. Calculations suggest that the sum of these
interactions is always substantially attractive in the optimized,
diequatorial, and diaxial conformations, although there is at least
one repulsive interaction in every conformation of each member
of the series. Summing all pairwise interactions of this type
shows that the net silyl/silyloxy attraction is greater in the
diequatorial conformation for 5a-d but in the diaxial conforma-
tion for 5e and 5f.
The sum of the attractive interactions of silyl atoms with
silyloxy atoms ranges up to 3.4 kcal/mol for trialkylsilyl groups
and up to 7.2 kcal/mol for arylsilyl groups. Repulsive interac-
tions are never more than 0.7 kcal/mol in total. With such
conformational energies in play, it is little wonder that the
conventional equilibrium of cyclohexane chairs with equatorial
or axial substitution (with A values of 0.56-1.31 kcal/mol in
series 1) can be overwhelmed by two, many-atomed, silyl
substituents arranging their relative positions to best effect.
Calculations for 1,4-Bis-silyloxycyclohexanes 4a-f, Table
4. When the silyl group is SiMe3 4a, SiEt3 4b, and SiPh3 4e,
the excess energy of the diaxial conformation is calculated to
be markedly higher than that of the corresponding monosilyl
compound 1 shown in Table 3, but it is certainly not twice as
large. In 4c, 4d, and 4f, the excess energy in the diaxial
conformation is much the same as that found for the corre-
sponding monosilyl compounds.
Attractive interactions between SiR3 groups and the far end
of the ring favor the diaxial conformation even more markedly
than in the series 1, but details are not shown. Conformations
with both SiR3 groups on the same side of the H1, C1, C4, H4
plane are more stable by over 100 cal/mol than those with groups
on opposite sides of the plane This shows that there is some
residual attractive steric interaction between silyl groups, even
when they are separated by the ring and two oxygen atoms.
Calculations for 1,2-Bis-silyloxycyclohexanes 5a-f. The
calculations, shown in Table 2, reproduce the trend of the
experimental results without closely matching absolute values.
It is particularly encouraging that the significant diaxial popula-
tions found in the NMR spectra of the three compounds 5d-f
are also predicted by the calculations. The torsion angles
calculated for the now-adjacent exocyclic C-OSiR3 bonds are
interesting. In diequatorial conformations, SiR3 groups are turned
away from each other and thus away from the near-to-eclipsing
The calculations reported were carried out assuming a
dielectric constant of 1.5, the MM3 default value. The value
used might affect results, particularly in the series 4 and 5, with
two silyloxy groups. Changing the constant to 5 produced a
fairly uniform increase of 0.2 kcal/mol in the calculated axial/
equatorial energy difference (and insignificant changes in
geometric parameters), thus improving agreement with experi-
ment. Further increasing the constant to 10 or 20 produced much
smaller additional changes for all compounds.22
Calculations of Rotational Potentials. Molecular mechanics
allows the calculation of the molecular energy as the H-C-
O-Si torsion angle is changed through a range of values. This
should approximate the rotational potential for the bond, if
(22) A referee has pointed out that the dielectric constant of dichloromethane
is given as 9 at 293 K and 16 at 180 K in the CRC Handbook, so the
improved agreement with experiment may be valid, but it is not clear
whether solvent can dominate the dielectrics within a single molecule.
9
J. AM. CHEM. SOC. VOL. 125, NO. 49, 2003 15169

A R T I C L E S
Marzabadi et al.
Figure 3. Potential energy diagrams of compounds 1a (equatorial and axial), 5a (diequatorial), and 20.
Discussion
rotation is a simple process as it probably is for at least
trimethylsilyloxycyclohexane, 1a. Rotation about the C-O bond
is complex in most other molecules with significant changes
within silyl groups, so we seldom tried to calculate these.
Conformation of the Exocyclic Bond and the Axial-
Substituted Chair Conformation. The tendency of a silyloxy
group on a six-membered ring to adopt eclipsed conformations
for the exocyclic C-O bond is so marked that such an eclipsed
conformation should be the point of departure for a discussion
of more complicated molecules. Rotation away from a staggered
conformation minimizes marked 1,3-repulsive steric interactions
of the tertiary silyl group with the adjacent equatorial positions,
even when the latter are not substituted. Eclipsing is equally
likely for axial and equatorial silyloxy groups, since both have
a similar geometric relationship to the adjacent equatorial posi-
tions and calculations of the repulsions involved confirm this.
Neighboring equatorial substituents other than hydrogen
interact more strongly, so that in the minimum energy diequa-
torial conformation of the 1,2-disubstituted compounds 5, silyl
groups rotate away from each other somewhat (i.e., away from
eclipsing).
For both equatorial and axial forms of 1a (see Figure 3), a
similar potential curve is obtained for H-C-O-Si torsion
angles between -60° and 60°. The energy minimum is close
to 0°, but the striking consequence of a preferred eclipsed
conformation is that the potential minimum is flat-bottomed and
240° wide. Maxima occur about 120° on either of the minimum
when the OsSi bond eclipses a ring C-C bond. Whether the
trimethylsilyloxy group is axial or equatorial, rotation from the
minimum to a torsion angle of 60° increases the energy by less
than 1kcal/mol.
A much steeper sided potential well results when an equatorial
siloxy group has two equatorial neighbors; see Figure 3. This
shows a plot for compound 20, which has three equatorial
trimethylsilyloxy groups and is thus a model for a silylated
sugar. For an alternative view, see 20a.
When there are equatorial substituents on both sides of a
silyloxy group, as in many silylated sugars, there is a rather
more marked steric compression with a reduced possibility of
relief by such a rotation. However, ring inversion to give the
alternative chair, having axial silyloxy groups with hydrogen
atoms as the adjacent equatorial substituents, is still a possible
relief, although with a conventional axial-substituent penalty.
In fact, in the diaxial conformation, the simpler alkylsilyl groups
remain eclipsed but otherwise, particularly in 5b, 5e, and 5f,
they rotate toward each other to maximize attractive interactions
to the point where the first repulsive interaction is encountered.
Whenever a trimethylsilyloxy substituent has no equatorial
neighbors, a broad potential energy well is calculated, and many
slightly different conformational states of similar energy can
presumably be populated. When an equatorial silyloxy group
has one or two equatorial neighbors, the potential energy well
has steeper sides, and fewer states are accessible. The entropy
of such a conformation is reduced. A different broad minimum
with a wide choice of states can be reached, however, by ring
inversion to give axial groups with no equatorial neighbors.
There is again an enthalpic penalty in changing to an axially
substituted chair, but there is a counterbalancing entropic gain.
Role of Attractive Interactions in the Extended Equatorial
Conformation and the Folded Axial Conformation. Since the
When a trimethylsilyloxy group has only one equatorial
neighbor (e.g., diequatorial 5a), the potential has a narrower,
skewed minimum which is a combination of those for 1a and
20.
Calculations with full matrix minimization quantified the
vibrational entropy of both diaxial and diequatorial conforma-
tions in series 4 and 5. The results obtained vary greatly, but
on average, such entropy is lowest in the diequatorial conforma-
tion of the series 5 when silyloxy groups are most able to
constrain each other. Thus, both aspects of the calculations
suggest that reduced entropy of a chair conformation with
adjacent equatorial substituents of the type OSiR3 may favor
the axial chair conformation.
It is significant that for 5a and 20 in the ring-inverted form
when all substituents are axial and thus not interfering, the
overall enthalpy is greater, but the potential well is broad again.
9
15170 J. AM. CHEM. SOC. VOL. 125, NO. 49, 2003

Conformational Equilibria in Silyloxycyclohexanes
A R T I C L E S
three R groups are distant from the ring, due to long C-Si and
O-Si bonds, steric repulsion is significant for only two of the
conformational points at issue. The first is the tendency toward
eclipsed exocyclic bonds discussed above. The second is the
consequence of an oxygen substituent being axial. Within a
given series, this is always the same, being the conventional
repulsion of that oxygen atom by two axial hydrogen atoms.
Nonetheless, the position of chair-chair equilibria ranges widely
down each series, so interactions less obvious than these
repulsions need to be considered.
both when axial and equatorial. NMR observations bear out this
last point. Finally, the same calculations show that the tert-
butyl fragment has attractive interactions with the far end of
the ring that are 0.21 kcal/mol greater in the axial chair than in
the equatorial.
tert-Butoxycyclohexane thus behaves like the bulkier sily-
loxycyclohexanes despite having significantly shorter bonds and
thus suffering relatively more steric compression. We expect
that the odd conformational behavior of silylated sugars will
also be observed for any poly-O-tert-alkylated sugars that can
be prepared.
The silyl group locates antiperiplanar to the cyclohexane ring
along the O-C bond, so the axial silyloxycyclohexane is a more
compact molecule than the equatorial alternative. In particular,
SiR3 when axial is nearer to the nine atoms making up the
methylene groups at positions 3, 4, and 5 in the cyclohexane
chair, than when it is equatorial.24 This results in greater
attractive stabilization of the molecule in the axial conformation
by several hundred calories per mole. The relative stabilization
increases with the number of atoms in the SiR3 group, but not
in a linear fashion, since not all SiR3 atoms can locate at the
positions that maximize attractive stabilization (see Table 5).
The fit would be better or worse depending on whether
molecular mechanics under- or overestimates attractive interac-
tions. None of the model conformational results used in the
parametrization of MM3 specifically focused attention on
attractive interactions.
The importance of 1,3-interactions in cyclohexane confor-
mational analysis thus extends beyond that of the axial sub-
stituent with axial hydrogen atoms at positions 3 and 5 on the
ring. Our results show that, for OX both axial and equatorial,
the attached X-group when it is quaternary, such as SiR3 or
tert-butyl, has significant 1,3-interactions with the adjacent
equatorial hydrogen atoms.24 The consequence is that rotation
about the exocyclic C-O bond minimizes this particular 1,3-
interaction, which seems thus to be more important than the
bond eclipsing which results.
We do not conclude that the present work totally explains
the “silyl effect”, but it sheds light on two important aspects
that can be taken forward to a discussion of polysilyloxy
compounds.
With the trans-1,2-disubstituted series 5, two further points
have to be considered. First, there are interactions between the
side chains which are substantially attractive in total in all
conformations, which increase with the bulk of the side chain,
and which while they favor the diequatorial conformation for
simpler silyl groups, favor the diaxial conformation for the
bulkier. Second, the diaxial conformations are favored by higher
entropy since equatorial substituents reduce the surface of each
other’s potential energy well.
Care must be exercised in interpreting small sets of interac-
tions extracted from molecular mechanics calculations. Thus,
the exocyclic C-O bond is calculated to be eclipsed in both
conformations of compound 1a (and in many others). Inspection
of the detail of these minimum energy conformations of 1a
shows that the interactions between the atoms of SiMe3 and
the equatorial hydrogen atoms at C2 and C6 are all attractiVe.
We can aver confidently, however,14 that such a bond is eclipsed
because of repulsion between SiMe3 and these equatorial
hydrogens in the conVentional staggered conformation. Calcula-
tions show only what the molecule has done to relieve what
plausibly are the determining repulsive interactions in the
molecule.
In polysilyloxy compounds such as protected sugars, poly-
equatorial conformations are disfavored since a third equatorial
substituent prevents two vicinal substituents from optimizing
their interactions by moving apart. Polyaxial conformations still
have their entropic advantage and neighboring substituents retain
their freedom to optimize attractive interactions.
This can be confirmed by examining the calculations for the
equatorial conformation of compound 5a, where the second
trimethylsilyloxy group forces the first to rotate toward the C6
position with an H-C-O-Si torsion angle of 12°. Here the
repulsive methyl-CH2 pairwise interactions total 0.13 kcal/mol
whereas in the equatorial conformation of 1a, with a torsion
angle of 4°, there are no such repulsive interactions.
Contrast with Alkyl Ethers. The conformational behavior
of the silyloxycyclohexanes is not unusual compared with that
of alkoxy analogues, when it is realized that comparisons up to
now have involved primary alkoxy derivatives. A more realistic
comparison is with tert-alkoxycyclohexanes. A long-known
NMR study of tert-butoxycyclohexane and associated molecular
mechanics calculations both suggest an axial/equatorial equi-
librium that places the steric effect of the tert-butoxy group
between that of the methoxy and the ethoxy groups.23 The
experimental A-values under standard conditions are OMe, 0.60,
OEt, 0.90, and O-t-Bu, 0.75. It has also been noted from
calculations of minimum energy conformations14,15 that while
methoxycyclohexane has an H-C-O-Me torsion angle of
about 42°, the C-O bond of tert-butoxycyclohexane is eclipsed
Trans-1,2-disubstituted Cyclohexanes as Models for More
Complex Molecules. The series 5 presents a simple model for
the interactions of two complex silyloxy groups, in the light of
the conformational features that have emerged from the series
1 and 4. There would be repulsive interactions between SiR3
groups in an undistorted diequatorial conformation, and to that
extent, the simple rationalization is correct, but, in the minimum
energy diequatorial conformation, the two silyl groups rotate
away from each other to minimize repulsions. In a polysilylated
sugar, for example, a â-D-glucose derivative with potentially
four equatorial silyloxy groups, two adjacent silyloxy-groups
rotating away from each other encounter further equatorial
substituents opposing this movement. In the alternative chair
(23) Senderowitz, H.; Abramson, S.; Aped, P.; Schleifer, L.; Fuchs, B.
Tetrahedron Lett. 1989, 30, 6765-6768.
(24) In compound 1a, calculations suggest that the maximum and minimum
separation of two hydrogen atoms H-C-Si‚‚‚C4-H is 9.04 and 6.25Å in
the equatorial conformation, whereas these limits are only 8.00 and 5.92
Å in the axial conformation.
9
J. AM. CHEM. SOC. VOL. 125, NO. 49, 2003 15171

A R T I C L E S
Marzabadi et al.
conformation, in contrast, each axial silyloxy group is quite free
of vicinal restrictions on obtaining its optimum local conforma-
tion.
To this discussion of repulsive interactions in such compounds
must be added the attractive interactions that we have demon-
strated in the model series. Axial SiR3 groups have enhanced
attractive interactions with the far end of the sugar ring
compared with equatorial groups, while the mutual attraction
of neighboring, polyatomic, SiR3 groups and entropy favor
diaxial conformations.
The overall picture that emerges from this work is thus of a
different kind of conformational analysis from all-carbon
systems. Oxygen-silicon and silicon-carbon bonds are quite
long compared with the carbon-for-silicon analoguessthe
standard values used in MM3 for example are O-Si 1.64 Å,
O-C 1.42 Å; C-Si 1.87, C-C 1.53 Å, so repulsive interactions
play a lesser role. On the other hand, readily available silyl
chlorides allow the introduction of quite complex trialkylsilyl
groups with many atoms, which if they are too far from the
ring to repel, must attract. Much that is observed in conforma-
tional terms can be attributed to enhanced attractive interactions
of one or more silyl groups in the axial conformation. This is
certainly the rationalization that applies to Eliel and Satici’s
work on monosilyloxycyclohexanes.
Figure 4. Methine proton NMR signals for compound 5e at 178 K.
SiR3 atoms attract each other overall, and for bulkier silyl
substituents, this attraction is greater in the diaxial. In addition,
in the axial conformation, the two silyloxy groups enjoy a broad
rotational potential minimum with a greater range of states than
in the diequatorial conformation.
The conventional axial/equatorial equilibrium of simple
compounds depends on repulsive interactions of a few atoms
at positions 1, 3, and 5 on the ring. Polysilylated six-membered
rings have many hundreds of attractive interactions totalling
several kilocalories per mole, so it not surprising that in many
cases, these dominate the overall molecular equilibrium ob-
served.
We have tried not to place too much emphasis on the exact
sums of attraction and repulsion between groups. Molecular
mechanics calculations do not address the possible influence
of undoubted differences in the electronic nature of the
substituents on silicon,1 but we doubt if the correct balance
between steric and electronic factors can be obtained and related
to the single experimental number we have obtained for each
compound. We believe that the broader indications as to the
relative importance of steric attraction and repulsion and
conformational entropy are nonetheless quite credible.
Experimental Section
Crystal Structures. Hydrogen atoms were not located directly in
the determination. For the purpose of determining the H-C-O-Si
torsion angles reported, the methine hydrogen atoms were assumed to
be located on the external bisector of the ring C-C-C bond angles.
For the purpose of determining closest approach of silyl groups in the
crystal of 1e, hydrogen atoms were located conventionally on the
observed benzene rings.
NMR Spectra. Spectra were recorded for solutions ∼5 mM in CD2-
Cl2 using a Bruker DRX-500 MHz spectrometer with a gradient unit
system with variable temperature control or using a Bruker AMX 300-
MHz spectrometer. Solid-state spectra were recorded on a Bruker MSL
300-MHz solid-state spectrometer. To measure populations at low
temperatures, the methine proton signals were electronically integrated.
These signals were assigned on the basis of their width at half-height,
since coupling was not always clearly resolved (Figure 4). The
diequatorial conformation has a much broader signal (∼20 vs ∼7 Hz)
since the axial methine proton has large couplings to adjacent axial
protons. In 5e and 5f, in contrast to 5a-d and 4a-f, axial protons are
downfield from equatorial protons, since they are deshielded by an
anisotropic, equatorial substituent at C2.
Conclusions
A range of silylated derivatives of trans-1,2- and trans-1,4-
dihydroxycyclohexane have been studied by NMR spectroscopy,
X-ray diffraction studies, and molecular mechanics (MM3)
calculations, to explain unusually large populations of chair
conformations with axial substituents. For both axial and
equatorial, tertiary, silyoxy groups, eclipsed conformations for
the exocyclic CsO bond relieve parallel-1,3-interactions be-
tween the SiR3 group and the adjacent equatorial hydrogen
atoms. The axial silyloxycyclohexane has a more compact shape
than the equatorial alternative. While its oxygen has a conven-
tional repulsive interaction with C3 and C5, the atoms of the
SiR3 group have only attractive interactions with the methylene
groups at positions 3, 4, and 5 in the ring, and these are greater
than in the extended equatorial conformation and increase down
each series 1, 4, and 5, as the number of atoms in SiR3 increases.
When two silyloxy groups are next to each other, the
exocyclic bond conformation can adjust so that the two sets of
For compounds 4, methine proton signals are not first order (i.e.,
not triplets of triplets) due to a large coupling between isochronous
axial protons at C2 and C3 and at C5 and C6.
Barriers to ring inversion appear to be about 11.6 kcal/mol in all
compounds. For several compounds of series 4 and 5, populations were
also measured for toluene-d6 solution, and in agreement with Eliel and
Satici’s observations for series 1, the population of the diaxial
conformation was found to be somewhat higher than in dichlo-
romethane-d2 solution.
Molecular Mechanics Calculations. Molecular mechanics calcula-
tions were carried out using Allinger’s MM3 (94) program.19,20 Both
the dihedral drive option and the stochastic search option were used to
supplement chemical intuition in locating overall minima for each ring
conformation. The rest of the parameters were used as default
parameters in the program and run on a Silicon Graphics work station.
Interaction of groups, e.g., the trimethylsilyl groups of 5a, was
determined by summing the 169 atom-atom pairwise interactions listed
in the MM3 (94) output.
9
15172 J. AM. CHEM. SOC. VOL. 125, NO. 49, 2003

Conformational Equilibria in Silyloxycyclohexanes
A R T I C L E S
Acknowledgment. C.M. would like to gratefully acknowledge
the Clare Boothe Luce Fund for a Professorship and the ACS
Petroleum Research Fund for partial support of this research.
J.G.O. is grateful for a Bogue Research Fellowship (UCL) for
his partial support during this project. We would also like to
thank to Dr. Abil Aliev from UCL NMR services for his help
in the acquisition of the spectra and his very useful comments
regarding the interpretation of the NMR data.
Supporting Information Available: Experimental procedures
and spectral data for all new compounds (5a-f, 4a-f, 17).
Tables of crystallographic data for compounds 4e, 5e, and 17.
Unit cell diagram for compound 5e. This material is available
free of charge via the Internet at http://pubs.acs.org. See any
current masthead page for ordering information and Web access
instructions.
JA035936X
9
J. AM. CHEM. SOC. VOL. 125, NO. 49, 2003 15173