Experimental evidence for negative hyperconjugation as a component of the polar effect: Variation of the ease of α-sulfonyl carbanion formation with the orientation of a β-alkoxy substituent

Article abstract of DOI:10.1021/ja001320l

We report evidence for a strongly geometry-dependent substituent effect. The rate constants for H-D exchange of the α-hydrogens, (k(exch))OR, in a set of 19 β-alkoxy sulfones of known, fixed (or strongly preferred) H-C(α)-C(β)-OR torsion angles have been

Full text of DOI:10.1021/ja001320l

10308
J. Am. Chem. Soc. 2000, 122, 10308-10324
Experimental Evidence for Negative Hyperconjugation as a
Component of the Polar Effect: Variation of the Ease of R-Sulfonyl
Carbanion Formation with the Orientation of a â-Alkoxy Substituent
James F. King,* Rajendra Rathore, Zhenrong Guo, Manqing Li, and Nicholas C. Payne*
Contribution from the Department of Chemistry, The UniVersity of Western Ontario, London,
Ontario, Canada N6A 5B7
ReceiVed April 14, 2000. ReVised Manuscript ReceiVed July 28, 2000
Abstract: We report evidence for a strongly geometry-dependent substituent effect. The rate constants for
H-D exchange of the R-hydrogens, (k
preferred) H-CR-Câ-OR torsion angles have been measured. Those of corresponding model compounds
lacking the alkoxy group, (kexch model, were also measured, thereby providing the ratio, kN ) (kexch OR/(kexch model;
exch)OR, in a set of 19 â-alkoxy sulfones of known, fixed (or strongly
)
)
)
the kN values so obtained range over more than 4 orders of magnitude. We show that when due allowance is
made for steric and other influences, our observations are consistent with an equation of the form log kN ) a
+ b cos2 θ (where a ) 1.70 ( 0.17 and b ) 2.62 ( 0.20). It is further shown that the observed rate constant
ratios are not consistent with a substituent effect consisting only of the inductive effect and the field effect,
and that they are fully consonant with the additional presence of a third effect, namely negative hyperconjugation
or the generalized anomeric effect, specifically a torsion-angle-dependent donation of the partial negative charge
of the incipient carbanion into the σ/C-O orbital. This effect is largest with torsion angles of 0 and 180° and at
these torsion angles constitutes the major source of stabilization of the incipient carbanion by the â-alkoxy
group. The present observation of a torsion-angle-dependent substituent effect may be combined with the
known adherence of the specific rate of R-sulfonyl carbanion formation to the Taft equation, to provide an
equation yielding torsion-angle-dependent Taft σ* constants for the alkoxy group: (σ/)OR ) 0.35 + 0.55
θ
cos2θ. The idea of torsion angle dependence is usefully applied to the long-standing problem of the mechanisms
of base-promoted elimination with 2-tosyloxycyclohexyl p-tolyl sulfones.
Introduction
overall picture has been accepted as sufficient for at least four
decades by a great many chemists. It is in the full knowledge
of this that we wish to present in this paper experimental
evidence that the above explanation for substituent effects is
not adequate, and that the major mechanism by which one
substituent influences the reaction center in at least one reaction
is by way of negatiVe hyperconjugation or the generalized
anomeric effect.6-9
In our study we have examined the rate of exchange of
hydrogen atoms for deuterium atoms R to a sulfonyl group, a
process known to take place by way of the R-sulfonyl carbanion,
and the substituent examined is the alkoxy group placed on the
â-carbon, i.e., on the carbon next to that undergoing H-D
exchange. The choice of this system was dictated by two critical
A key procedure in the study of reaction mechanisms is the
examination of the influence of a small perturbation in
structuresa change of substituentson properties and reactivity.
For reactions of nonconjugated systems, the electronic effect
of a substituent is customarily described as a polar effect,
consisting of (a) an inductive (or electronegativity) effect in
which the action of the substituent is propagated along the
σ-bond array, and (b) a field (or Coulombic) effect in which
the substituent acts directly through space; another effect,
polarizability, may also be included.1 Though a number of
features of this description, including the independence2 (or
otherwise)3 of these factors, their relative importance,4 and
methods of quantifying them,5 have received attention, the
(6) (a) Lemieux, R. U. Pure Appl. Chem. 1971, 25, 527-548. (b) Eliel,
E. L. Angew. Chem., Int. Ed. Engl. 1972, 11, 739-750. See also: (c)
Thatcher, G. R. J. In The Anomeric Effect and Associated Stereoelectronic
Effects; Thatcher, G. R. J., Ed.; ACS Symposium Series 539; American
Chemical Society: Washington, DC, 1993; Chapter 2.
* Correspondence concerning the organic chemistry should be directed
to J.F.K. (E-mail scijfk@julian.uwo.ca, telephone (519) 661 2111 ext. 86348,
fax (519) 661 3022) and that related to the X-ray crystallography to N.C.P.
(E-mail noggin@julian.uwo.ca, telephone (519) 661 3793, fax (519) 661
3022).
(1) (a) Hine, J. Physical Organic Chemistry, 2nd ed.; McGraw-Hill: New
York, 1962; p 83. (b) Lowry, T. H.; Richardson, K. S. Mechanism and
Theory in Organic Chemistry, 3rd ed.; Harper and Row: New York, 1987;
p 152n. (c) Topsom, R. D. Prog. Phys. Org. Chem. 1976, 12, 1-20. (d)
Reynolds, W. F. Prog. Phys. Org. Chem. 1983, 14, 165-203. (e) Taft, R.
W.; Topsom, R. D. Prog. Phys. Org. Chem. 1987, 16, 1-83.
(2) Reynolds, W. F. J. Chem. Soc., Perkin Trans. 2 1980, 985-992.
(3) Hammett, L. P. Physical Organic Chemistry, 2nd ed.; McGraw-
Hill: New York, 1970; p 376.
(7) The “anomeric effect” is a phenomenological effect which refers to
the tendency for axial tetrahydropyranyl carbohydrates to be more stable
than their equatorial anomers (contrary to the usual dictates of conforma-
tional analysis). The term “generalized anomeric effect” recognizes that
the phenomenon found in pyranose sugars shows up in a wide variety of
structures and has “been coined to describe the phenomenon in all its
aspects”.8a The theoretical basis of the “anomeric effect” has been subjected
to much discussion but currently is regarded as arising (primarily) from n
f σ* (or analogous) delocalization;9 i.e., it is identified as an expression
of the theoretical concept of “negative (or anionic) hyperconjugation”.
(8) Eliel, E. L.; Wilen, S. H. Stereochemistry of Organic Compounds;
John Wiley & Sons: New York, 1994; (a) p 753; (b) pp 696-697; (c) pp
720-722; (d) p 707.
(4) Stock, L. M. J. Chem. Educ. 1972, 49, 400-404.
(5) Marriott, S.; Reynolds, W. F.; Taft, R. W.; Topsom, R. D. J. Org.
Chem. 1984, 49, 959-965.
10.1021/ja001320l CCC: $19.00 © 2000 American Chemical Society
Published on Web 10/10/2000

NegatiVe Hyperconjugation in the Polar Effect
J. Am. Chem. Soc., Vol. 122, No. 42, 2000 10309
Scheme 1a
pieces of information. (a) First, the H-D exchange R to the
sulfonyl group is known to have a well-defined stereochemistry,
specifically, that the preferred orientation of the hydrogen atom
in removal or addition has the hydrogen simultaneously gauche
to both sulfonyl oxygens or, equivalently, an antiperiplanar
arrangement of the H-CR-S-CR′ bonds as shown in A.10
(b) Second, the exchange of hydrogen isotopes R to a sulfonyl
group was known from the work of Thomas and Stirling11a on
the detritiation of PhSO2CHTCH2X to be very strongly de-
pendent on X, showing a good correlation with σ* (of CH2X,
F* ) 4.89); the effect of changing X from CH3 to OPh, for
example, was to increase the rate by 50 000-fold. This extremely
high sensitivity to substituents seemed, in our view, too large
to arise solely from the conventional polar effect, and it seemed
possible that it might involve an additional phenomenon, e.g.,
negative hyperconjugation or the generalized anomeric effect.7
Such an effect could readily arise given the arrangement in A,
if the electrons of the incipient negative charge on CR were to
donate into the antibonding (σ/C-X) orbital which is anti-
periplanar to the C-H bond. This would occur in addition to,
and be entirely analogous to, the usual generalized anomeric
stabilization of the R-sulfonyl carbanion involving H-CR-S-
CR′, as already mentioned above. The test of this hypothesis
would be to make compounds with different, fixed H-CR-
Câ-X torsion angles and to see if, and how, the rate constant
for H-D exchange varied with the torsion angle. Initially, we
looked at a limited set of â-alkoxy sulfones (A, X ) OR); our
results, which have been published in preliminary form,12 proved
extremely promising and encouraged us to carry out the sizable
task of studying the wider array of compounds described in this
study.
Ideally, in this test one would find how the rate of R-hydrogen
exchange varies in a series of â-alkoxy-substituted sulfones in
which the only change in structure is the H-C-C-O torsion
angle, and in which all steric effects are unchanged. In practice,
this is impossible, and what we did was to synthesize a series
of 19 â-alkoxy sulfones with (more or less) fixed H-C-C-O
torsion angles, with the idea that the effect of the alkoxy group
could be estimated by dividing the specific rate of the alkoxy-
substituted compound by that of an ideal model compound, i.e.,
one which is identical in all respects with the â-substituted
material except that the polar effect of the alkoxy group is
replaced by that of a hydrogen atom. For those compounds in
a Reagents: (a) PhSH, NaOH; (b) H2O2, HOAc; (c) MeI, KOH,
DMSO; (d) Swern oxidation; (e) LiAlH4; (f) KOH, EtOH, ∆ (60%
exo:40% endo).
which steric factors are unimportant, a reasonable approximation
for the above “ideal” model is one in which the alkoxy group
is replaced by a hydrogen (the “hydrogen model”), or, where
the oxygen is in a ring, by a methylene group, or in other
instances, the methyl group (the “carbon model”). As will be
apparent from what follows, some of the alkoxy-substituted and
model compound pairs are acceptably close to the ideal, whereas
others are not, usually because the introduction of the alkoxy
group leads to steric effects not present in the model.
Results and Discussion
Materials and Methods. Table 1 shows (a) the â-alkoxy-
substituted sulfones, (b) the H-C-C-O torsion angle in each,
as determined from PCModel calculations or from X-ray single-
crystal structure determination, (c) the corresponding model
compound(s), (d) kexch, the rate constant for the H-D exchange
with the specified medium and temperature for both the alkoxy
compound ((kexch OR
)
) and the model ((kexch model
)
), and (e) finally,
the key parameter, log kN, where kN ) (kexch OR
)
/(kexch model
)
.
The syntheses of the previously unknown [2.2.1]bicycloheptyl
alkoxy sulfones 2, 3, 9, and 10 are summarized in Scheme 1, a
route patterned on the earlier synthesis of the p-tolyl analogues
by Kleinfelter et al.;13 the [2.2.2]bicyclooctyl sulfones 1 and
11 were prepared similarly. trans-1,4-Oxathiadecalin 4,4-dioxide
(15) was originally prepared as described by Rooney and
Evans,14a but a more convenient procedure for obtaining both
14 and 15 from some of the same intermediates was developed
in this work (see Scheme 2). Note that formation of 14, which
may be regarded as a Michael-type addition of the alkoxy anion
onto the R,â-unsaturated sulfone, takes place with stereoselective
formation of 14 with no sign of 15; this is to be expected of an
axial addition to the double bond with formation of the
comparatively stable cis-R-sulfonyl carbanion with the favored
orientation of the sulfonyl and alkoxy groups as in A (X ) O).
Other preparations are recorded in the Experimental Section.
(9) (a) David, S.; Eisenstein, O.; Hehre, W. J.; Salem, L.; Hoffmann, R.
J. Am. Chem. Soc. 1973, 95, 3806-3807. (b) Schleyer, P.v. R.; Kos, A. J.
Tetrahedron 1983, 39, 1141-1150. (c) Wolfe, S.; Stolow, A.; LaJohn, L.
A. Tetrahedron Lett. 1983, 24, 4071-4074. (d) Bors, D. A.; Streitwieser,
A., Jr. J. Am. Chem. Soc. 1986, 108, 1397-1404. In this paper Fourier
analysis of the rotational potential surface yielded V(θ) ) V1/2(1 - cos θ)
+ V2/2(1 - cos 2θ) + V3/2(1 - cos 3θ), where V1, V2, and V3 are,
respectively, -0.170, 14.150, and -0.130, i.e., the V2 term (due to
“conjugative overlap effects”) dominates. (e) Apeloig, Y.; Rappoport, Z.
J. Am. Chem. Soc. 1979, 101, 5095-5098. Also: Rappoport, Z. Acc. Chem.
Res. 1981, 14, 7-15. Apeloig, Y.; Karni, M.; Rappoport, Z. J. Am. Chem.
Soc. 1983, 105, 2784-2793.
(10) For reviews on sulfonyl carbanion chemistry, see: (a) Boche, G.
Angew. Chem., Int. Ed. Engl. 1989, 28, 277-297. (b) Grossert, J. S.; Hoyle,
J.; Cameron, T. S.; Roe, S. P.; Vincent, B. R. Can. J. Chem. 1987, 65,
1407-1415.
(11) (a) Thomas, P. J.; Stirling, C. J. M. J. Chem. Soc., Perkin Trans. 2
1977, 1909-1913. (b) Marshall, D. R.; Thomas, P. J.; Stirling, C. J. M. J.
Chem. Soc., Perkin Trans. 2 1977, 1914-1919. (c) Marshall, D. R.; Thomas,
P. J.; Stirling, C. J. M. J. Chem. Soc., Perkin Trans. 2 1977, 1898-1909.
(12) King, J. F.; Rathore, R. J. Am. Chem. Soc. 1990, 112, 2001-2002.
(13) Kleinfelter, D. C.; Squires, T. G.; Mashburn, J. H.; Watsky, R. P.;
Brown, S. B. J. Org. Chem. 1977, 42, 1149-1153.
(14) (a) Rooney, R. P.; Evans, S. A., Jr. J. Org. Chem. 1980, 45, 180-
183. (b) Frieze, D. M.; Hughes, P. F.; Merrill, R. L.; Evans, S. A., Jr. J.
Org. Chem. 1977, 42, 2206-2211. (c) Barry, C. N.; Baumrucker, S. J.;
Andrews, R. C.; Evans, S. A., Jr. J. Org. Chem. 1982, 47, 3980-3983. (d)
Murray, W. T.; Kelly, J. W.; Evans, S. A., Jr. J. Org. Chem. 1987, 52,
525-529..

10310 J. Am. Chem. Soc., Vol. 122, No. 42, 2000
King et al.
Table 1. Rate Constants for H-D Exchange in â-Alkoxy Sulfones with “Fixed” HR-CR-Câ-O Torsion Angles (Plus Model Compounds)
θa (deg)
kexch (M-1 s-1) (conditions)b
4.9 × 10-4 (A)
model(s)
(kexch) )
model (M-1 s-1
4.2 × 10-8 (A)
log kN
c
alkoxy sulfone
1
26.2 (6.2)
18
4.06
2
3
4
7.1 (6.4)
13.9 (9.5)
59.3
1.04 × 10-4 (A)
1.14 × 10-3 (A)
19
1.5 × 10-8 (A)
3.1 × 10-4 (B1)
3.84
20
(4.5 × 10-8)d (A)
9.2 × 10-4 (B1)
(4.40)
2.57
C-2He: 4.5 × 10-4 (C)
C-2Ha: <1.8 × 10-5 (C)
21
1.2 × 10-6 (C)
(22)
(∼2 × 10-6)e (C)
(∼2.35)
5
60.0
61.3
1.14 × 10-3 (B3)
3.67 × 10-4 (B2)
23b
23a
24a
23b
24b
18
1.75 × 10-6 (B3)
3.3 × 10-6 (B2)
1.1 × 10-6 (B2)
2.81
2.04
2.52
1.81
(2.33)
0.67
6a
6b
61.9
1.12 × 10-4 (B3)
1.75 × 10-6 (B3)
1.5 × 10-4 (B2)
(5.25 × 10-7)f (B3)
4.4 × 10-5 (B2)
7
8
78.4 (63.9)
74.5
1.04 × 10-3 (B2)
5.07 × 10-4 (B3)
2.24 × 10-4 (B2)
20
(2.07 × 10-7)g (B3)
9.2 × 10-4 (B1)
(3.42)
9
133.0 (128.4)
133.1 (130.3)
2.00 × 10-4 (A)
4.13 × 10-4 (A)
19
20
1.5 × 10-8 (A)
4.12
10
(4.5 × 10-8)d (A)
(3.96)
11
12
145.0 (147.6)
152.4
1.36 × 10-3 (A)
1.07 × 10-3 (B3)
(4.30 × 10-2)e (C)
18
19
4.2 × 10-8 (A)
4.51
4.24
6.96 × 10-8 (B3)
3.1 × 10-4 (B1)
13
14
172.9
173.5
21
(22)
1.2 × 10-6 (C)
4.55
(∼4.33)
(∼2 × 10-6)e (C)
C-3He: 4.8 × 10-2 (C)
C-3Ha: 1.6 × 10-4 (C)
C-10He: 3.3 × 10-5 (C)
21
21
1.2 × 10-6 (C)
1.2 × 10-6 (C)
4.60
4.43
15
173.5
C-3He: 3.2 × 10-2 (C)
C-3Ha: 1.6 × 10-4 (C)
C-10Ha: <10-8 (C)

NegatiVe Hyperconjugation in the Polar Effect
Table 1. (continued)
J. Am. Chem. Soc., Vol. 122, No. 42, 2000 10311
θa (deg)
kexch (M-1 s-1) (conditions)b
model(s)
(kexch
)
model (M-1 s-1
)
log kN
c
alkoxy sulfone
16
174.0
C-3He: 1.6 × 10-2 (C)
C-3Ha: ∼2 × 10-4 (C)
C-5He: 2.3 × 10-2 (C)
C-5Ha: ∼2 × 10-4 (C)
3.90 × 10-4 (B3)
21
1.2 × 10-6 (C)
4.12
173.3
178.0
4.28
2.73
17b
23b
25b
1.75 × 10-6 (B3)
1.47 × 10-4 (B2)
(8.96 × 10-7)h (B3)
7.5 × 10-5 (B2)
(3.02)
17a
179.4
2.35 × 10-4 (B3)
23a
25a
4.2 × 10-8 (B3)
3.3 × 10-6 (B2)
(1.67 × 10-8)i (B3)
3.75
(4.16)
1.3 × 10-6 (B2)
a The torsion angles (θ) were estimated by calculation using PCMODEL (PCM4), except those in parentheses, which were obtained from X-ray
b
structure determination. Reaction conditions: (A) CD3CN-D2O (1:1), 21 °C; (B1) dioxane-d8-D2O (1:1), 77 °C; (B2) dioxane-d8-D2O (1:1), 64
°C; (B3) dioxane-d8-D2O (1:1), 25 °C; (C) D2O, 20 °C. Rate constants in parentheses are not directly measured values but are calculated as
described in the accompanying footnote. c kN ) kexch/(kexch
)
.
d Value estimated at 21 °C by multiplying the rate constant, kexch, for 19 at 21 °C
model
by the rate constant ratio of 20:19 at 77 °C. e This rate constant is that for H-D exchange per H, and was obtained by multiplying the experimental
value by 2 (see discussion in the text). f Value estimated at 25 °C by multiplying the rate constant, kexch, for 23b at 25 °C by the rate constant ratio
of 24b:23b at 64 °C. g Value estimated at 25 °C by multiplying the rate constant, kexch, for 19 at 25 °C by the rate constant ratio of 20:19 at 77 °C.
h Value estimated at 25 °C by multiplying the rate constant, kexch, for 23a at 25 °C by the rate constant ratio of 25a:23a at 64 °C. i Value estimated
at 25 °C by multiplying the rate constant, kexch, for 23b at 25 °C by the rate constant ratio of 25b:23b at 64 °C.
Scheme 2a
Chart 1
a Reagents: (a) HCl, DME; (b) H2O2, HOAc; (c) 5% aqueous KOH,
DME; (d) KOH, H2O-EtOH (3:1), 100 h reflux (9:1 trans:cis).
The structures of all new compounds are assigned by (a) the
mode of synthesis, (b) full concurrence of the 1H and 13C NMR
spectra and exact mass values with expectation, and (c) in the
(15) Diffraction data for compounds 1, 2, 3, 9, 10, and 11 were collected
from single crystals on either a Siemens P4 or an Enraf-Nonius CAD4
diffractometer. Structure solution and refinement was done using the
SHELXTL suite of programs. Hydrogen atom parameters were refined for
1 and included in idealized positions in the other models. Torsion angles
involving H atoms are probably accurate to 1°. Crystallographic data: 1,
C15H20SO3, monoclinic, space group P21/c, Mo radiation λ ) 0.71069 Å,
a ) 8.483(2), b ) 10.359(1), and c ) 16.091(2) Å, â ) 101.53(1)° with
Z ) 4. Least-squares refinement of 253 variables on F2 using 4249 unique
data gave agreement factors R1 ) 0.040 and R2 ) 0.038. 2, C14H18SO3,
monoclinic, space group P21/n, Mo radiation λ ) 0.71073 Å, a ) 5.850-
(1), b ) 16.624(3), and c ) 13.500(3) Å, â ) 95.13(3)° with Z ) 4. Least-
squares refinement of 164 variables on F2 using 2870 unique data gave
agreement factors R1 ) 0.043 and R2 ) 0.047. 3, C14H18SO3, monoclinic,
space group P21/n, Mo radiation λ ) 0.71073 Å, a ) 11.056(2), b ) 9.560-
(2), and c ) 12.653(3) Å, â ) 90.84(3)° with Z ) 4. Least-squares
refinement of 165 variables on F2 using 2408 unique data gave agreement
factors R1 ) 0.039 and R2 ) 0.054. 9, C14H18SO3, monoclinic, space group
P21/c, Mo radiation λ ) 0.71073 Å, a ) 11.555(2), b ) 15.188(1), c )
16.218(3) Å, â ) 108.20(1)° with Z ) 8. Least-squares refinement of 186
variables on F2 using 2403 unique data gave agreement factors R1 ) 0.083
and R2 ) 0.076. 10, C14H18SO3, monoclinic, space group P21/n, Mo
radiation λ ) 0.71073 Å, a ) 21.771(15), b ) 8.322(3), and c ) 30.856-
(23) Å, â ) 106.56(6)° with Z ) 16. Least-squares refinement of 322
variables on F2 using 3083 unique data gave agreement factors R1 ) 0.075
and R2 ) 0.073. 11, C15H20SO3, monoclinic, space group P21/n, Mo
radiation λ ) 0.71073 Å, a ) 8.822(2), b ) 15.746(3), and c ) 10.920(2)
Å, â ) 110.17(3)° with Z ) 4. Least-squares refinement of 174 variables
on F2 using 2486 unique data gave agreement factors R1 ) 0.043 and R2 )
0.058. These analyses will be reported in detail elsewhere.
case of the [2.2.1]bicycloheptyl and [2.2.2]bicyclooctyl alkoxy
sulfones, by single-crystal X-ray crystallography.15 The H-C-
C-O torsion angles so obtained (estimated error (1°) are given
in Table 1 in parentheses.
The rate constants shown in Table 1 were determined by
observing the 1H NMR spectrum under the specified conditions
of medium and temperature as a function of time; see the
1
Experimental Section. Normally the H and 13C NMR spectra
taken during and upon completion of the exchange reaction
showed only clean exchange, with no sign of products of
elimination or isomerization; the significant exceptions are
discussed below.
Geometric Effects. Relationship to the Sulfonyl Group.
For the majority of the compounds in Table 1, there is only
one hydrogen R to the sulfonyl group, and hence there is no
ambiguity about which hydrogen is being exchanged. With the
compounds carrying a methylsulfonyl group (6a, 17a, 23a-
25a, Chart 1) the exchange of the methyl hydrogens was found
to be very much faster than that of the methyne hydrogen on
the other side of the sulfonyl group; this is in accord with the

10312 J. Am. Chem. Soc., Vol. 122, No. 42, 2000
King et al.
kN
a
Table 2. H-D Exchange in â-Alkoxy Sulfones of Unfixed Conformation and γ-, δ-, ꢀ-Alkoxy Sulfones
alkoxy sulfone
kexch (M-1 s-1
1.83 × 10-3
)
model(s)
(kexch) )
model (M-1 s-1
1.47 × 10-4
28
29
23b
23b
12.4
1.35 × 10-3
1.47 × 10-4
9.2
b
30b
30c
30d
30e
32
PhSO2(CH2)2OCH3
5.12 × 10-1
1.68 × 10-3
31a
31b
31b
31c
31c
31d
31d
31e
23b
PhSO2CH2CH3
PhSO2(CH2)3CH3
4.14 × 10-4
2.3 × 10-4
2.3 × 10-4
2.03 × 10-4
2.03 × 10-4
1.80 × 10-4
1.80 × 10-4
1.80 × 10-4
1.5 × 10-4
1.23 x 103
2.27 x 103
7.5
8.3
2.0
2.2
2.2
2.2
49
PhSO2(CH2)3OCH3
PhSO2(CH2)4OCH3
PhSO2(CH2)5OCH3
PhSO2(CH2)4CH3
PhSO2(CH2)5CH3
PhSO2(CH2)5CH3
8.6 × 10-3
4.2 × 10-1
2.4 × 10-5
1.79 x 104
b,c
CH3SO2(CH2)2OCH3
CH3SO2CH2CH3
a In D2O-dioxane-d8; cyclohexyl sulfones at 64 °C; alkyl sulfones at 25 °C. b For comparison: PhSO2CH3, kexch ) 2.26 × 10-2 M-1 s-1; CH3SO2CH3,
kexch ) 9.95 × 10-4 M-1 s-1
. )
c Rates refer to the exchange of the methylene hydrogens; for the R-methyl group of ethyl methyl sulfone: kexch
9.9 × 10-4 M-1 s-1
.
general observation that (when otherwise equivalent) methyl
groups form a carbanion more readily than a methylene group,
which in turn reacts more readily than a methyne. With the
substrates and conditions in this study, the results in Tables 1
and 2 indicate that the relative rates of hydrogen removal from
methyl, methylene, and methyne groups are roughly 200:50:1.
Seven of the 19 alkoxy compounds of “fixed” geometry in
Table 1 have the sulfonyl group in a six-membered ring and
have at least one methylene group directly attached to the
sulfonyl function. In most of these (4, 14, 15, 16, and 21) the
methylene hydrogens are diastereotopic, but in two of them (13
and 22) the hydrogens of the methylene groups become
equivalent by simple chair H chair interconversion; the members
of this latter group present no problems in interpreting their
spectra, but they do require inclusion of a statistical correction
factor for comparison with other compounds in this study. The
exchange reaction that is being followed is the replacement of
all four R-hydrogens; at any time, however, only half of these
hydrogens are in the (more reactive) equatorial position, and
hence the measured rate constant must be multiplied by 2 for
comparison with compounds in which the R-hydrogen is at all
times in the favorable orientation.
preference for carbanion formation of the R-hydrogen with the
antiperiplanar H-C-S-C torsion angle and would have
required little comment had it not been that Katritzky and co-
workers16 reported that the equatorial:axial rate ratio in a model
system of theirs (cis-3,5-diphenyl-trans-4-hydroxy-3,4,5-trideu-
teriotetrahydrothian 1,1-dioxide) was only 1.6. The origin of
the difference between our results and Katritzky’s is not obvious
to us, but it is our contention that the comparative simplicity of
our compounds indicates that our systems typify the general
case and it is the Katritzky example which is the exception,
perhaps because of perturbations due to the hydroxy and phenyl
groups. This is also in agreement with the observation by Fuji
et al.17 that the equatorial R-hydrogens in 6-methyl-1,3-oxathiane
3,3-dioxide exchange 15-25 times more rapidly than the axial
R-hydrogens (in NaOCD3-CD3OD at 20 °C). It might be argued
that most of our examples are also perturbed by the presence
of an ether oxygen, and it is the basic point of this paper that
this oxygen and its orientation strongly influence the rate of
H-D exchange. In the two simple examples (21 and 4) noted
above, however, one (21) has no ether oxygen at all and the
other has the methoxy group (more or less) symmetrically
disposed with respect to the two R-methylene hydrogen atoms,
and hence it is most unlikely that the oxygen in 4 is directly
responsible for the difference between the reactivities of the
two hydrogens. In the other three compounds (14, 15, and 16)
the oxygen accelerates the exchange reaction, but in our view
it is the orientation of the hydrogen atom with respect to the
CR′-S bond that is the primary factor which determines which
hydrogen is exchanged.
In the group in which the individual hydrogens of the
methylene group are diastereotopic (i.e., 4, 14, 15, 16, and 21),
the two hydrogens can be expected to exchange at different rates.
The hydrogens are readily assigned as axial or equatorial (in
1
the most stable conformation) by their H NMR signals; the
axial hydrogens show the typical large axial-axial coupling
(10-15 Hz) with each vicinal axial hydrogen and a much
smaller axial-equatorial coupling (2-4 Hz) with their vicinal
equatorial neighbors, whereas the equatorial hydrogens show
small couplings with both axial and equatorial neighbors. On
observing the exchange reactions by 1H NMR, it became evident
that in all of the six-membered cyclic sulfones the equatorial
hydrogen(s) exchange(s) faster than the epimeric axial hydrogen
atom(s). With 21, for example, the equatorial R-hydrogen (C-
2e) exchanges about 100 times faster than the axial hydrogen
of the methylene group (C-2a), and this in turn exchanges much
faster than the axial methyne hydrogen (C-10a). Similarly, in 4
it is the equatorial hydrogen which is exchanged more rapidly
than the axial, by a factor of more than 25. These observations
are, indeed, what was expected on the basis of the known
The evidence in the present study is fully consistent with the
picture that, with the compounds and the conditions used in
this investigation, at least, exchange of the R-hydrogens takes
place only when the hydrogen has the antiperiplanar torsion
angle around the H-CR-S-CR′ bond. The exchange of
hydrogens lacking the favorable H-CR-S-CR′ torsion angle
in the major conformation requires a conformational flip either
to the alternative chair conformation (i.e., 4 with the methoxy
group axial, 14 with the sulfonyl group axial to the cyclohexane
(16) Brown, M. D.; Cook, M. J.; Hutchinson, B. J.; Katritzky, A. R.
Tetrahedron 1971, 27, 593-600.
(17) Fuji, K.; Usami, Y.; Sumi, K.; Ueda, M.; Kajiwara, K. Chem. Lett.
1986, 1655-1658.

NegatiVe Hyperconjugation in the Polar Effect
J. Am. Chem. Soc., Vol. 122, No. 42, 2000 10313
3.33 (with 24b). It is customary to assume that part of these
rate reductions may arise from a small electronic effect of the
methyl group (σ* of CH2X for X ) H is 0.00 and for X ) CH3
is -0.1),18a and hence the steric effect of the methyl group
(which must be only a part of the total effect) cannot be large.
With 16 the effect of the 2-methyl group is to reduce the specific
rate of H-D exchange from 3 × 10-2 to 1.6 × 10-2 M-1 s-1
.
Again, in this more reactive system, the steric effect of the
methyl group must be small; note that unless the steric effect
of the methyl group is to accelerate the reaction (which seems
unlikely in these cases), the maximum electronic effect of the
methyl group is to slow the reaction by a factor of 2. Methyl
and methoxy groups can often be regarded as of comparable
“size”, with a slightly larger Taft Es value18b for OCH3 vs CH3
(respectively, for CH2X, -0.19 and -0.07) and a distinctly
lower A value8b for methoxy as compared with methyl (respec-
tively, a range from 0.55 to 0.75 vs 1.74 kcal mol-1). It is within
this framework that the points in Figure 1 given as bullets or
black diamonds are regarded as unlikely to be strongly
influenced by steric factors; i.e., these points are the least
perturbed by effects other than the electronic effect of the
â-alkoxy group. The bullets and black diamonds clearly fall
into three clusters, one at low torsion angles (0 to 20°), one
around 60°, and finally, the largest, in the 170-180° range. The
values of log kN are around 4 for the first and last of these groups
and are close to 2 for the 60° angle compounds. These results
in themselves clearly show log kN to be angle dependent. They
are also not readily accounted for on the basis of the field effect;
calculation of field effects using the Kirkwood-Westheimer
approach (see below) cannot yield the variation shown by the
bullets and black diamonds in Figure 1. Negative hyperconju-
gation or the generalized anomeric effect, which would lead to
a torsion angle effect roughly corresponding to a cos 2θ (or
cos2 θ) dependence (as is discussed more fully below), presents
itself as a most attractive possibility. We show in the next three
subsections that when due allowance is made for either steric
assistance or steric hindrance or the “γ-effect”, we may, with
the aid of arguments totally independent of the generalized
anomeric effect, factor out the steric (and γ-effect) contributions
to a number of these kN values, leading to the “corrected” log
kN points shown as heavily outlined circles (“bulls-eyes”) and
leading to the cosine curve in Figure 1.
(b) Correction for Large Steric Assistance. In contrast to
the examples considered in the previous section, in which steric
factors are regarded as fairly small, a substantial steric effect
was immediately evident from the observation in the reactions
of the cis-methoxy sulfones, 9, 10, and 11, not only of H-D
exchange, but also of concomitant isomerization to the equili-
bration mixture; e.g., starting with 9 led to an equilibrium
mixture of the R-D isotopomers of 9 and 3, with the latter
predominating. In deuterated media, the free energy changes
accompanying equilibration (∆G°C ) G°cis - G°trans) of the
respective isotopomers were 2.5 ( 0.2 for 3 h 9, 4 ( 1 for 2
h 10, and 2.4 ( 0.2 kcal mol-1 for 1 h 11. The likely source
of the clear preference for the trans isomer is a strong
nonbonding repulsion (probably both steric and Coulombic)
between the eclipsed cis methoxy and arylsulfonyl groups. In
addition to this, there is a difference between the ∆G° values
for 3 h 9 vs 2 h 10 which is readily assigned to the well-
known lower energy of exo- vs endo-[2.2.1]bicycloheptane
systems. This was confirmed by treating 19 and 20 individually
with strong base (at 77 °C) to give mixtures of 19 and 20,
indicating an equilibrium mixture with 86 ( 5% of 19 and 14
Figure 1. Plot of kN ) (kexch)OR/(kexch)model for 18 of the 19 compounds
listed in Table 1. Circles (filled or open) refer to kN values determined
using a hydrogen model, and the turned squares (“diamonds”) refer to
kN values obtained using a carbon (methyl or methylene) model. The
filled circles (“bullets”) and filled turned squares (“black diamonds”)
are believed to represent the effect of the alkoxy group without requiring
major correction for steric or other effects. The heavily outlined circles
(“bulls-eyes”) are values corrected for steric and γ effects as noted in
the text. Compounds 1 and 7 are each represented by two horizontally
yoked circles, reflecting the uncertainty resulting from discrepancies
between the estimated (PCModel) and X-ray crystallographic torsion
angles (see Table 1). Compounds 6a, 6b, 17a, and 17b are shown with
vertically yoked points (circles and diamonds, from the hydrogen and
carbon models, respectively). The curve was calculated from log kN )
(3.01 ( 0.10) + (1.31 ( 0.11)cos 2θ ) (1.70 ( 0.17) + (2.62 (
0.20)cos2 θ, which was obtained by a weighted nonlinear squares fit
of the bullets and black diamonds with each of the yoked points given
a weighting of 0.5; inclusion of the bulls-eyes with weightings
corresponding to 1/(uncertainty)2 did not affect the parameters in the
above equations.
ring, or 16 with the methyl axial) or to the twist conformation
of the heterocyclic ring in 15 or 21. The very slow exchange of
the (axial) methyne hydrogen in 15 could conceivably be an
exception to this picture and could be taking place by way of
a conformation in which the H-CR-S-CR′ torsion angle is not
antiperiplanar; it should be noted, however, that even this
hydrogen can come close to achieving the antiperiplanar
arrangement when both rings are in twist conformations.
Geometric Effects. Relation to the Alkoxy Group. (a)
Small Steric Effects. The experimental values of log kN are
shown in Figure 1 as either black filled or fully open circles (b
or O) or turned squares (“diamonds”). With those points shown
as black filled circles (“bullets”, b) or black turned squares
(“black diamonds”, [) it is believed that the principal influence
on log kN is the electronic effect of the alkoxy group, i.e., steric
factors have only a relatively small effect on log kN; within the
context of a range of kN values spanning more than 4 orders of
magnitude, “small” may be taken as a factor of 3 or (more
usually) less. Some notion of the result of a small perturbation
may be gained by examining the effect of replacing a â-hy-
drogen by a methyl group, i.e., by comparing the specific rates
of H-D exchange either of (a) the R-hydrogens of 23 with those
in 24 and 25 or (b) the C-3 and C-5 equatorial hydrogens in
16. In the case of 24b and 25b vs 23b, the effect of the methyl
group is to slow the reaction by factors of 2.0 (with 25b) and
(18) Taft, R. W. In Steric Effects in Organic Chemistry; Newman, M.
S., Ed.; John Wiley & Sons: New York, 1956, (a) p 595. (b) p 598.

10314 J. Am. Chem. Soc., Vol. 122, No. 42, 2000
King et al.
( 5% of 20, and corresponding to ∆G°endo - ∆G°exo ) ∆G°NX
) 1.26 ( 0.3 kcal mol-1. If we label the cis interaction energy
difference in 9 vs 3 as (∆G°C)X and that in 10 vs 2 as (∆G°C)N,
and if we then assign the observed energy differences primarily
to the cis and endo interaction energies, we may write the
following:
G°20 - G°19 ) ∆G°NX ) 1.3 ( 0.3 kcal mol-1
G°9 - G°3 ) (∆G°C)X - ∆G°NX ) 2.5 ( 0.2 kcal mol-1
G°10 - G°2 ) (∆G C)N + ∆G°NX ) 4 ( 1 kcal mol-1
From this we readily obtain (∆G°C)X ) 3.8 ( 0.5 kcal mol-1
and (∆G°C)N ) 2.7 ( 1.3 kcal mol-1
.
The exo-cis and endo-cis isomer interaction energies appear
to be somewhat different, though the accumulated error in these
estimates is large and precludes precise comparison. Inspection
of the S- - -OMe interaction distances and the O-C-C-S
torsion angles,19 which indicate slightly shorter distances in the
exo isomer, are in accord with an apparently higher energy for
(∆G°C)X. The energy difference between the cis- and trans-
[2.2.2]bicyclooctyl sulfones, G°11 - G°1, was found to be 2.4
( 0.2 kcal mol-1, i.e., a value apparently smaller than the other
cis interactions, in accord with the larger O-C-C-S torsion
angle and S- - -OMe internuclear distance in 11 vs 10 and 9.19
The abstraction of a proton to form the carbanion from 9,
10, or 11 must be accompanied by a change in the O-C-C-S
torsion angle, and this in turn leads to a measure of relief of
the ground-state strain noted above; this well-known phenom-
enon is often referred to8c as “steric assistance”. The question
that now arises is what proportion of the observed energies of
the cis and endo interactions is expressed in the transition states
leading to the carbanion; i.e., to what extent does ground-state
strain affect the rate of carbanion formation (and hence of H-D
exchange)?
To approach this problem, we have looked at the relationship
between rates and equilibria in two simple sulfonyl systems,
19 h 20 and 26 h 27. An energy diagram for 19 h 20 is
shown in Figure 2. Note that ∆Gq is 26.22 kcal mol-1 for 20
and 25.46 kcal mol-1 for 19; the difference in free energy of
activation, ∆∆Gq ) ∆Gq20 - ∆Gq19 ) 0.76 kcal mol-1 is clearly
ascribable to steric acceleration in the reaction of 19. Of the
original 1.26 kcal mol-1 of strain energy (in 19 vs 20), 0.76
kcal mol-1, or 60%, is released as steric assistance in the
formation of the R-sulfonyl carbanion from the sulfone. In the
other system 26 h 27 (Chart 1), the axial isomer was found to
be higher in energy by 2.66 kcal mol-1, and the difference in
free energies of activation was 1.91 kcal mol-1; the release of
strain energy is therefore 1.91/2.66, or 72% of the total. The
results from the two reactions are in reasonable accord (average
66%) and consistent with other work on steric assistance.20
Returning to the alkoxy systems, for example the reaction of
9, we take 66% of 3.7 ( 1 ) 2.44 ( 0.6 kcal mol-1 as the
strain energy released in making the carbanion. To find the
electronic effect of the methoxy group, we must correct the value
of log kN accordingly; i.e., we must reduce log kN by the effect
of the 2.44 kcal mol-1, which is to say, by 2440/2.303RT )
1.52 ( 0.2 (T ) 77 °C), and the value of log kN for 9 corrected
for steric acceleration becomes 2.60 ( 0.2, the value given by
Figure 2. Free energy vs reaction coordinate diagram for carbanion
formation and epimerization reactions of 20 and 19 (with sodium
deuterioxide in dioxane-d8-D2O, 1:1, at 77 °C). The ∆G° value for
19 h 20 was obtained from the estimated equilibrium ratio (86:14);
the ∆Gq values are calculated from the kexch values in Table 1 and from
kinv ) 3.8 × 10-6 M-1 s-1 for 20 f 19 determined by Guo44b under
these conditions.
the bulls-eye (and error bars) in Figure 1. In the same way, the
corrected (bulls-eye) values for log kN for 10 and 11 (2.81 and
3.52) were also obtained.
(c) Correction for the γ-Effect. In light of our observation
that an alkoxy substituent in the â position may accelerate the
H-D exchange by more than 104-fold, it would appear prudent
to take due account of an alkoxy interaction at the γ position
and perhaps more distant locales as well. Table 2 includes the
results of a brief look into γ and more distant effects. From
this we see that an alkoxy group connected by a single three-
carbon chain to the sulfonyl group increases the rate of H-D
exchange by 7.5-12.4 times. With one oxygen connected to
the R-carbon by two identical three-carbon chains, as in 32
(Chart 1), the effect of the oxygen is increased to 49-fold. By
the same token, we must conclude that in the reactions of the
two 7-oxabicyclo[2.2.1]heptyl sulfones (8 and 12) in which the
oxygen is â to the sulfur by one route and γ by another, the
observed rate must be influenced not only by the â interaction
but also by the γ effect, and hence the magnitude of the γ effect
must be accounted for if we are to estimate the magnitude of
the â effect by itself in these compounds. An upper limit of
roughly 12-fold for the γ effect may be set by the value for 28
(Chart 1) in Table 2. The lower limit is less clear because it
would appear reasonable to expect that when one oxygen is
(20) Ru¨chardt and Beckhaus, for example, find slopes for plots of ∆Gq
vs strain enthalpy for thermal cleavage of hydrocarbons which point to about
40-67% of strain energy being released in the reaction.20a The effects of
strain energy release are, however, known to be highly variable,20b,20c and
estimates of these effects must be made with great caution. We emphasize
that our approach in the present study is simply empirical: we have (a)
taken two examples of the reactions under study (i.e., carbanion-mediated
H-D exchange), (b) found that they show similar strain release factors,
and (c) applied the average of these factors to obtain our estimate of the
strain release correction. (a) Ru¨chardt, C.; Beckhaus, H. D. Angew. Chem.,
Int. Ed. Engl. 1980, 19, 429-440. (b) Tonachini, G.; Bernardi, F.; Schlegel,
H. B.; Stirling, C. J. M. J. Chem. Soc., Perkin Trans. 2 1988, 705-709. (c)
Sella, A.; Basch, H.; Hoz, S. J. Am. Chem. Soc. 1996, 118, 416-420.
(19) The O-C-C-S torsion angles (as determined by X-ray crystal-
lography) in 9 are 11.2 and 0.7° (two conformations), those in 10 are 13.9,
17.3, 6.2, and 10.9° (four conformations), and that in 11 is 30.8°; the S- - -
OCH3 internuclear distances are, respectively, in 9, 2.830 and 2.900 Å; in
10, 2.888, 2.971, 3.009, and 2.887 Å; and in 11, 2.930 Å.

NegatiVe Hyperconjugation in the Polar Effect
J. Am. Chem. Soc., Vol. 122, No. 42, 2000 10315
interacting by two paths, the total effect will not be as large as
if there were two separate oxygen atoms acting independently;
i.e., to the extent that an oxygen atom acting by, for example,
a â effect withdraws charge from the R-carbon in the transition
state, then that same oxygen atom acquires an increased negative
charge and can therefore be expected to be less electron-
withdrawing by the other pathway than if this involved another,
quite different, oxygen atom. In accord with this, we find that
the “double γ effect” via the two pathways (in 32, 49-fold) is
less than that expected from two independent γ effects. In the
absence of any straightforward basis for assigning the γ effect
when the same oxygen also shows a â effect (as in 8 and 12),
we suggest that the value 5 ( 3 very likely covers the γ effect
range in these substrates. The corrected point for 12 in Figure
1 has been obtained using this number; the large error bars
reflect the uncertainty in the γ effect assignment. For 8, an
additional (steric) correction is required because the strain energy
difference between 8 and 12 is 2.15 kcal mol-1 (or 0.9 kcal
mol-1 greater than that observed for 19 vs 20), and hence is
not adequately accounted for with 20 as the model. Accordingly,
a correction corresponding to 66% of the difference in strain
energies (i.e., 0.66 × 0.9 kcal mol-1) is applied to 8 in addition
to the γ effect correction; the final corrected values of log kN
for 8 and 12 are, respectively, 2.25 and 3.54.
Table 2 also records examples of the effect of more distant
alkoxy groups, i.e., those in the δ and ꢀ positions from the
sulfonyl group. It is interesting to note that even at the ꢀ position
the effect of the alkoxy group is detectably different (by a factor
of 2) from that of the hydrogen; note that there is no effect on
introducing a methyl group (31e vs 31d). It is tempting to
postulate direct donation from the R center into the σ* orbital
of the alkoxy-bearing carbon by way of a cyclic structure (three-
membered for γ, four-membered for δ, and five-membered for
ꢀ).
(d) Other Deviant Points. Even the most cursory examina-
tion of Figure 1 shows two points (unfilled circles) to be clearly
removed from the others, those for (a) cis-2-methoxycyclohexyl
phenyl sulfone (17b), and (b) the bridgehead methoxy sulfone
7; the kN value of the latter, 4.6 (log kN ) 0.66), is very low,
being only about half that of the typical γ effect kN values shown
in Table 2! This, in itself, suggests a rate suppression factor of
some kind. Each case is examined in footnote 21, and it is
concluded that it is possible to assign a likely origin for the
deviation from the pattern shown by the other points, but that
we cannot derive quantitative corrections such as those discussed
above. At this stage we regard the results with 7 and 17b as
neither contributing to nor detracting from the conclusions drawn
from the bullets, black diamonds, and bulls-eyes in Figure 1.
Variation of log kN with the H-C-C-O Torsion Angle.
(a) Negative Hyperconjugation and the Generalized Ano-
meric Effect. Inspection of Figure 1 reveals a pattern consistent
with variation of log kN with cos 2θ (or cos2 θ), where θ is the
H-C-C-O torsion angle; the line corresponds to eq 1 derived
from a nonlinear least-squares best fit.
Chart 2
of imprecision is made worse by the fact noted already that kN
cannot, by its very nature, always be uniquely and precisely
determined.22 Within this framework we suggest that the results
in Figure 1 constitute a strong case for eq 1. The variation of
experimental log kN values with θ ranges from log kN of about
2 to about 4.6; i.e., there is a torsion-angle-dependent component
(21) Examination of molecular models suggests that two factors combine
to slow the H-D exchange in 7 relative to its model (18). As was noted in
the Introduction, the preferred (and in the present examples, probably the
only) orientation of the R-sulfonyl system for carbanion formation is that
shown in A. For 7 this would require the conformation shown in 33b (Chart
2) with very strong nonbonding repulsions between the phenyl and the
nearest methylene groups. In the model (18 ) 33a) this can be alleviated
in considerable measure by having the phenyl group move as shown by
the arrow with concomitant twisting of the [2.2.2]bicyclooctyl system (which
presumably also helps to minimize the eclipsing interactions in this array).
With 7 ()33b), however, such motion leads to a shrinkage of the O-C-
C-S torsion angle from its original value of about 60° in the untwisted
bicyclo system to a value approaching 0°. This extreme would correspond
to the eclipsed arrangement of the phenylsulfonyl and methoxy groups,
which, as has already been seen with 9, 10, and 11, above, would add extra
energy of the order of ∼2.5 kcal mol-1 or more. It would appear that the
methoxy group in 7 introduces a steric-cum-electronic effect not adequately
factored out by dividing by the rate constant for the model system (18),
and hence a correction must be applied to increase kN to find the purely
electronic effect of the methoxy group in 7. It should also be noted that
any tendency to involve hyperconjugation in 7 (which would be expected
to be small because of the unfavorable torsion angle) would probably be
accompanied by at least a small change in hybridization (from sp3 to sp2)
at C-1 and C-2. Such a change would generate angle strain in the bridged
structure associated with Bredt’s rule. The net result could be either (a) no
hyperconjugation or (b) some hyperconjugation but with little or no energy
release because of increased strain. The other unfilled circle well removed
from the others is that due to 17b. Examination of molecular models shows
that the appropriate conformation of the H-D exchange is that shown in
34 (Chart 2). One immediately notes what appears to be a substantial “1,3-
synaxial-like” interaction between the methoxy and phenyl groups. Interest-
ingly (as has been noted earlier), however, when cis-2-methylcyclohexyl
phenyl sulfone (27a) was examined, the effect of the “1,3-synaxial-like”
methyl-phenyl interaction was found to be comparatively minor, with the
rate constant for H-D exchange in 25b about one-half of that of cyclohexyl
phenyl sulfone (23b). One possibility raised by these results is that there is
a substantial MeO- - -Ph interaction not mimicked by the methyl analogue,
i.e., one presumably not primarily steric in origin. That there is something
unusual about the “1,3-synaxial-like” MeO- - -Ph interaction is indicated
by the following evidence which is quite apart from anything related to
Figure 1. Note the effect on relative rates of H-D exchange of changing
from a methyl sulfone to the corresponding phenyl sulfone: 23-fold from
CH3SO2CH3 to CH3SO2Ph, 45-fold from 23a to 23b, 26-fold from 6a to
6b, 58-fold from 25a to 25b, but only 4-fold from 17a to 17b. These results
are consistent with a repulsive nonbonding MeO- - -Ph interaction in 17b
()34) (and analogous compounds) not compensated for by models such as
23b or 25b. Without a clear picture of its origin, we are not in a position
to make any estimate of its effect on rates of H-D exchange magnitude
other than the roughly 1 order of magnitude change noted in the methyl vs
phenyl sulfone series above, and we merely indicate, by an upward-pointing
arrow in Figure 1, that it is likely that the points for both 7 and 17b, when
properly corrected, will likely have distinctly higher log kN values.
(22) A similar problem of precise determination appears with some well-
known chemical parameters, e.g. resonance (or delocalization) energy and
effective molarity (alias effective concentration). Like these, log kN can be
useful, and absence of high precision is not in itself grounds for not making
use of the idea.
log kN ) (3.01 ( 0.10) + (1.31 ( 0.11) cos 2θ )
(1.70 ( 0.17) + (2.62 ( 0.20) cos2 θ (1)
Of the three log kN values not included in the least-squares
treatment (those for 5, 7, and 17b), the last two are discussed
above and in footnote 21, while 5 is taken up in a later section.
Experimental free energy relationships in chemistry are always
subject to “noise”, i.e., to minor deviations from perfect fit to
the line. In addition, in the present case of log kN, the problem

10316 J. Am. Chem. Soc., Vol. 122, No. 42, 2000
King et al.
of log kN which appears to constitute more than half of the
maximal value, and also a component (presumably the inductive
and field effects) which appears to be (more or less) independent
of θ (see below).
1.2 × 103). Thomas and Stirling11a report 0.44 and 3.7 × 10-4
M-1 s-1 (ratio 1.2 × 103) for their detritiations of PHSO2-
CHTCH2X (in EtO-/EtOH at 25 °C), and it is clear that the
two systems are very much alike. If we write eq 1 (log kN
)
3.01 + 1.31 cos 2θ ) 1.70 + 2.62 cos2 θ) from our work and
take log kN ) 4.89σ/θ from Thomas and Stirling’s results, we
Such a cosine relationship is, of course, what is expected, as
a first approximation, for electron donation of the n f σ* type
yielding a (full or partial) π-bond. In an ab initio study, Bors
and Streitwieser9d describe the variation of the energy with
rotation about the CR-S bond in the ChH2SO2CH3 anion; the
variation is given as a function of cos θ, cos 2θ, and cos 3θ,
but the actual coefficients are such as to make it primarily a
function of cos 2θ.9d Schleyer and Kos9b similarly find a roughly
sinusoidal dependence of relative energy with torsion angle
accompanying rotation of the C-N bond in FCH2NH2. Of
especial relevance to the present study are their results of
calculations with ChH2CH2OH. The total interaction energy was
get eq 2, which gives the σ/ value of the CH2OR group as a
θ
function of the torsion angle, θ: In this formulation, σ/θ for the
(σ/ )OR ) 0.62 + 0.27 cos 2θ ) 0.35 + 0.54 cos2 θ
(2)
θ
CH2OR group may vary from σ/ ) 0.35 through σ6/0 ) 0.49
90
to σ/0 ) σ1/80 ) 0.89. A plot of the “corrected” log kN values
(bullets, black diamonds, and bulls-eyes) vs σ/θ gives an
approximately straight line with slope (F*) of 4.88. Note that
σ/180 ) 0.89 is distinctly larger than the value of σ* for CH2-
OMe (0.64) used by Thomas and Stirling. This reflects the fact
that the maximum value of log kN in our study is taken as 4.3,
whereas in Thomas and Stirling’s work log(kOMe/kH) ) 3.08;
i.e., in these experiments we are seeing larger kN values with
alkoxy groups in fixed conformations than with the conforma-
tionally mobile species (30b). Such a situation arises in a
conformationally mobile system whenever there is a mixture
of conformations of differing reactivity and the system is subject
to Winstein-Holness kinetics.27 In the simplest case of two
equimolar conformations, one of which is reactive and the other
totally unreactive, for example, the measured kobs is one-half
of the specific rate of the reacting conformation. In the present
case of PhSO2CH2CH2OMe (30b), the reacting conformer is
arranged as in A (X ) OMe, CR′ ) Ph). From the discussion
already given21 in connection with the relatively low reactivity
of 17b, the arrangement as in A is expected to have high energy.
The low value of log(kOMe/kH) is presumably a reflection of a
relatively low concentration of the reacting conformer (A, X )
OMe, CR′ ) Ph).
-
calculated as the energy of the following reaction: CH3CH2
+ HOCH2CH3 f CH3CH3 + HOCH2CH2-. When the free
electron pair and the C-O bond were antiperiplanar, the total
interaction energy was found to be 23.5 kcal mol-1, whereas
when the torsion angle of the electron pair and the C-O bond
was 90°, the interaction energy was 10.3 kcal mol-1. They
assigned the difference (13.2 kcal mol-1) to negative hyper-
conjugation, i.e., the anomeric effect, with the residual 10.3 kcal
mol-1 due to the inductive effect. Putting our experimental
observations and arguments together with the calculations of
Schleyer and Kos, the inescapable conclusion is that the effect
of a â-alkoxy group on the ease of R-carbanion formation
displays negative hyperconjugation or, alternatively, is an
illustration of the generalized anomeric effect.
(b) The Synperiplanar Lone-Pair Stereoelectronic Effect.
One notable feature of Figure 1 which warrants special mention
is the powerful effect of synperiplanar alkoxy groups, as shown
by the large log kN values for 1, 2, and 3, in which the effect
on kN is close to or as large as that of the antiperiplanar alkoxy
groups (e.g., compounds 14, 15, and 16). Padwa and Wanna-
maker23 have proposed synperiplanar n f σ* overlap in
connection with carbanion formation from a substituted 2-meth-
oxycyclopropyl p-tolyl sulfone. Experimental evidence for the
synperiplanar lone-pair effect in the chemistry of acetals has
been put forward by Deslongchamps and Kirby and co-workers24
(who concluded that the synperiplanar effect is weaker than the
antiperiplanar) and supported by calculations.25 Notwithstanding
a suggestion that “synperiplanar...overlap would be disfa-
vored”,26 the case for a strong synperiplanar lone-pair effect
appears established.
This suggests that a rigorous treatment of a reaction of
conformationally mobile species by the Taft equation would
require a full conformational analysis with individual rate
constants for conformers taken with a set of σ/θ values.
Presumably, the success of simpler treatments suggests that (a)
some systems are not complex, and (b) we often deal with
averaging of conformational populations, rates, and σ* values.
Observation of experimental “noise” is not surprising.
(d) The Angle-Independent Term. As has been noted
already, nearly half of the effect of alkoxy groups on log kN
appears to be more or less independent of the torsion angle, θ.
This is readily reconciled with the presence of the inductive
effect which operates by successive bond polarizations and is
therefore uninfluenced by the torsion angle. The field effect,
however, is expected to be controlled to some extent by
geometry. The Coulombic interaction of a dipolar bond (the
C-O bond in this instance) and charge (or partial charge in a
transition state) may be calculated from the Kirkwood-
Westheimer equation;28 for a particular reaction and substituent
(e.g., the alkoxy group) this equation may be written as ∆E )
C(cos ú)/r2, where with the system at hand, RO-Câ-CR-H, r
is the distance from the hydrogen nucleus to the center of C-O
bond, ú is the angle that the hydrogen-to-mid-bond line makes
with the C-O bond, and C is a constant in this system.29 We
(c) Angle Dependence of σ*. A New Parameter, σ/θ. The
variation of log kN with the torsion angle may be described in
terms of the Taft parameter σ* provided one adds a new feature,
namely angle dependence; we shall symbolize the angle-
dependent σ* as σ/θ, where θ implies an unspecified angle and
σ/30, for example, refers to the value at θ ) 30°. Under the
conditions of our experiments (NaOD in D2O-dioxane-d8 at 25
°C), the rate constants for H-D exchange of PhSO2CH2CH2-
OMe (30b) and PhSO2CH2CH3 (31a) (Chart 1) were found to
be, respectively, 0.51 and 4.14 × 10-4 M-1 s-1 (ratio ) kN )
(23) Padwa, A.; Wannamaker, M. W. Tetrahedron Lett. 1986, 27, 2555-
2558.
(24) Li, S.; Kirby, A. J.; Deslongchamps, P. Tetrahedron Lett. 1993,
34, 7757-7758. Li, S.; Deslongchamps, P. Tetrahedron Lett. 1993, 34,
7759-7762. Deslongchamps, P.; Jones, P. G.; Li, S.; Kirby, A. J.; Kuusela,
S.; Ma. Y. J. Chem. Soc., Perkin Trans. 2 1997, 2621-2626.
(25) Grein, F.; Deslongchamps, P. Can. J. Chem. 1992, 70, 604-611;
1562-1572.
(27) Winstein, S.; Holness, N. J. J. Am. Chem. Soc. 1955, 77, 5562-
5578.
(28) Kirkwood, J. G.; Westheimer, F. H. J. Chem. Phys. 1938, 6, 506-
512. Westheimer F. H.; Kirkwood, J. G. J. Chem. Phys. 1938, 6, 513-
517.
(26) Laszlo, P. Organic Reactions, Simplicity and Logic; John Wiley &
Sons: Chichester, 1995; p 457 (transl. by H. R. Meier).

NegatiVe Hyperconjugation in the Polar Effect
J. Am. Chem. Soc., Vol. 122, No. 42, 2000 10317
obtained atomic coordinates with PCModel and then calculated
(cos ú)/r2 as a function of θ, the O-C-C-H torsion angle,
and obtained a smooth sigmoid curve varying from a low of
0.033 Å-2 with θ ) 0° to a maximum of 0.11 Å-2 with θ )
180°; i.e., the energy of the field effect is smallest with θ ) 0°
and is greatest (by about 3.3-fold) with θ ) 180°.
1 to see if any of these might react to a substantial extent via
a conformation different from that shown. One likely example
was, in fact, found.
Compound 5 in the most stable (equatorial) conformation (5e)
clearly has the PhSO2 group equatorial, with the H-C-C-O
torsion angle (θ) close to 60°. The alternative axial conformation
(5a) can be expected to have higher energy than 5e, but θ in 5a
is close to 180°, and from eq 1 the rate constant for 5a must be
greater than that for 5e by about 114 ( 5 times because of the
angular dependence of the stereoelectronic effect. Using the
Note that the overall shape shown by the bullets, black
diamonds, and bulls-eyes in Figure 1, and that for the field effect
as described by Kirkwood and Westheimer, are qualitatively
totally different. The field effect therefore cannot be responsible
for the major features of Figure 1, but it is still possible that it
could be making a detectable contribution. Inspection of Figure
1 suggests that the points around θ ) 180° appear to be slightly
higher than those around θ ) 0°, perhaps by as much as 0.5
log kN units. One conceivable source of this differencesif it is
realscould be the field effect.30 More complete evaluation of
the Kirkwood-Westheimer equation29 requires the use of
somewhat arbitrary parameters. Our calculations, which suggest
that the field effect may make a contribution to log kN of 0.22
at θ ) 0° and 0.74 at θ ) 180°, were calculated by using DE
) 4 and a developing charge in the region originally occupied
by the hydrogen of 0.75 of a negative charge; note that either
a larger value for DE or a smaller magnitude for the charge
lowers the log kN contributions due to the field effect estimated
in this way.
The results and discussion to this point indicate that the polar
effect of an alkoxy group may be qualitatively analyzed as
follows: at θ ) 0° and 180°, the polar effect comprises a
negative hyperconjugative effect plus a smaller inductive effect
plus a still smaller field effect. At other values of θ the
hyperconjugative effect diminishes, until at θ ) 90° it is zero
and the polar effect consists of only the inductive effect and
the field effect.
Winstein-Holness equation,27 we may write kobs ) keqneq
+
kaxnax, where n is the number of mole fractions of the subscripted
conformer and k is its rate constant. To estimate n for each of
the conformers we note from an earlier section that ∆G° for 26
h 27 is 2.66 kcal mol-1 (at 64 °C) (a number in accord with
the reported conformational energy 2.50 kcal mol-1 at -90 °C
of the methylsulfonyl group attached to cyclohexane).32 We also
note, on the other hand, that Garc´ıa Ruano et al.33 have estimated
that ∆Gax - ∆Geq for 3-methylsulfonyltetrahydropyran (35,
Chart 2) is “in the 1.4 to 1.5 kcal mol-1 range”, and have
suggested that the most favorable axial arrangement is that with
the methyl group pointing into the ring to avoid oxygen-oxygen
repulsions (but unable to avoid a methyl-1,3-axial hydrogen
interaction). The phenylsulfonyl group may be expected to show
a stronger interaction with the 1,3-axial hydrogen, and hence
∆Gax -∆Geq for 5 may be expected to be >1.5 kcal mol-1
.
We suggest that 2.0 ( 0.5 kcal mol-1 probably encompasses
the likely range, and from this we estimate nax ≈ 0.03 (range
0.01-0.09), from which, if we set kax ≈ 114 keq, then it is
evident from the Winstein-Holness equation that about 80%
(range 63-90%) of the reaction of 5 will proceed by way of
the axial conformer (5a). We also find log kN ) 4.2 ( 0.3 for
5a and 2.14 ( 0.3 for 5e, values of log kN in good accord with
expectation from eq 1 on the basis of torsion angles of,
respectively, 180 and 60°; these values are not included in the
array used for the best fit for eq 1, however, because they are
not estimated independently of eq 1.
Reaction via a Less Stable Conformer: The Case of
Compound 5. Although many reactions take place by way of
the most stable conformer(s), one must keep in mind that
exceptions to this generalization frequently occur when, as in
the present study, stereoelectronic requirements are important.31
It therefore seemed prudent to examine the substrates in Table
An Application. We now apply the results of this study to a
question of mechanism dating back to the middle 1950s, when
Bordwell and Pearson and co-workers34 looked at syn and anti
eliminations in a series of sulfones including 36 and 37 (Chart
2). These authors concluded initially that the anti elimination
from 36 and the syn elimination from 37 (to form 1-cyclohex-
enyl p-tolyl sulfone in each case) both proceeded by way of
concerted (E2) reactions. In 1962, Hine and Ramsay35 ques-
tioned this conclusion and provided evidence that the syn
elimination, contrary to the contention of Bordwell and Pearson,
was not faster than could be expected for a carbanion process.
Subsequently, Bordwell, Weinstock, and Sullivan36 marshalled
arguments and evidence to conclude that the syn and anti
eliminations both take place by the carbanion mechanism. Their
argument invoked two ideas, (a) internal return of the carbanion
to account for the apparent difference between rates of H-D
exchange and elimination reactions, and (b) a “retardation effect”
in forming carbanions from 1,2-diequatorially disubstituted
(29) According to the Kirkwood-Westheimer equation ∆E ) qµ(cos
ú)/Dr2, where r is the distance from the hydrogen nucleus to the center of
the C-O bond, q is the magnitude of the charge (which we took initially
as that of the electron, 1.60 × 10-19 C, see below), µ is the bond moment
of the dipolar bond, taken as 0.7 D where 1 D ) 3.34 × 10-30 C‚m, and
D ) DE(4πꢀ0), where DE is the “effective dielectric constant” (see below)
and ꢀ0 is the permittivity of free space (8.85 × 10-12 C2 N-1 m-2). From
this (with Avogadro’s number) we obtain ∆E ) 2.03 × 10-15(cos ú)/DEr2
J mol-1. As is discussed in the text, when θ is 0° then (cos ú)/r2 ) 0.033
Å-2 ()3.3 × 1018 m-2), and when θ is 180° then (cos ú)/r2 ) 0.11 Å-2
.
At this point we must make a choice of two parameters, DE and the
magnitude of the developing negative charge. Kirkwood and Westheimer28
have suggested that the effective dielectric constant, DE, for the space
between the charge and a dipole should be in the range 3-10; we find that
DE ) 4 gives a range of log kN values consistent with Figure 1. The
developing anionic charge must be less than 1, but, from the considerable
sensitivity of the reaction to substituents (cf. F* ) 4.89) we find it difficult
to imagine a charge less than 0.5; we used 0.75 in our calculations while
noting that the results are not very sensitive to the precise number. These
parameters (with 4.184 J/cal) gave ∆E values of 0.30 and 1.00 kcal mol-1
(for θ ) 0° and 180°, respectively); division by 2.303RT gives log kN field
effect contributions of 0.22 (minimum) and 0.74 (maximum), respectively.
(30) Another possible source could be the choice of model compounds
for 13, 14, 15, and 16; in all of these the alkoxy oxygen is modeled by a
methylene group, which could (by its small polar effect) be slowing the
reactions of the models, leading to slightly higher kN values for these
substrates.
(32) Eliel, E. L.; Kandasamy, D. J. Org. Chem. 1976, 41, 3899-3904.
(33) Garc´ıa Ruano, J. L.; Rodr´ıguez, J.; Alcudia, F.; Llera, J. M.;
Olefirowicz, E. M.; Eliel, E. L. J. Org. Chem. 1987, 52, 4099-4107.
(34) Bordwell, F. G.; Kern, R. J. J. Am. Chem. Soc. 1955, 77, 1141-
1144. Weinstock, J.; Pearson, R. G.; Bordwell, F. G. J. Am. Chem. Soc.
1956, 78, 3468-3472. Bordwell, F. G.; Landis, P. S. J. Am. Chem. Soc.
1957, 79, 1593-1597. Weinstock, J.; Bernardi, J. L.; Pearson, R. G. J.
Am. Chem. Soc. 1958, 80, 4961-4964.
(31) For purposes of illustration we note two examples: (a) the
antiperiplanar preference over a synclinal arrangement in the E2 reaction
which leads to reaction by way of the axially oriented leaving group on a
cyclohexane ring, and (b) the formation (and opening) of cyclohexene oxides
by way of the diaxial conformation of the chlorohydrin.
(35) Hine, J.; Ramsay, O. B. J. Am. Chem. Soc. 1962, 84, 973-976.
(36) Bordwell, F. G.; Weinstock, J.; Sullivan, T. F. J. Am. Chem. Soc.
1971, 93, 4728-4735.

10318 J. Am. Chem. Soc., Vol. 122, No. 42, 2000
King et al.
cyclohexyl derivatives; in our view, both of these effects need
at least a little modification.
Returning to Hine and Ramsay’s paper,35 the key piece of
evidence was a plot of log k vs σ* in which the reactions were
(a) the H-D exchange reactions of cyclohexyl p-tolyl sulfone
(23c) and 2-cis-methoxycyclohexyl p-tolyl sulfone (17c), and
(b) the elimination reactions of 36, 37, and 2-cis-chloro- and
2-cis-fluorocyclohexyl p-tolyl sulfones. They commented as
follows: “From the best straight line through the unsubstituted,
cis-2-methoxy and trans-2-tosyloxy sulfones [i.e. 23c, 17c and
37], it is seen that the rate of cis elimination of the trans-tosyloxy
compound [37] is not faster but instead rather slower than would
be predicted from the σ*-constants of the groups used.” The
agreement between the points and “the best straight line” in
the above quote is not very good, and the authors discussed
possible steric factors for the observation that the point for 37
was below that expected from the σ* value. Hine and Ramsay’s
plot included 36 and two cis-halo analogues; the points for the
three were well above the best straight line through the other
three points, consistent with the idea that these compounds react
by an E2 reaction. Note that they used the same σ* value for
both the cis- and trans-tosylates (36 and 37).
We wish to point out that the notion of angle-dependent
substituent effects automatically puts all of the points on the
same straight line (modified) Taft plot. If we recall from the
above discussion that σ6/0 ) 0.49, i.e., roughly half of σ/0 )
σ/180 ) 0.88, it is tempting, indeed, to simply assign to the
trans-tosylate a σ* value about half that of the cis isomer;
immediately one sees all six points arrayed in a pattern
remarkably close to a straight line! Obviously, a more rigorous
treatment of these data in the light of the idea of angle-dependent
σ* values requires a set of independently determined σ1/80
values for the CH3, CH2F, CH2Cl, CH2OTs, and CH2OMe
groups and a σ/60 value for the CH2OTs function. We do not
have such a set of experimental values, but a curve obtained
using roughly estimated values gave an acceptable straight line.37
Since two of the Hine-Ramsay reactions are carbanion forma-
tions, it is reasonable to infer that the eliminations also involve
carbanions or, if not, that the transition states are extremely
similar to those of carbanions. In other words, the reactions are
either E1cB or very E1cB-like E2 processes.
Figure 3. log-log plot of Hine and Ramsay’s rate constants for H-D
exchange in 23c and for elimination in 37 and 36 vs our rate constants
for H-D exchange in 23b, 6b, and 17b (Table 1).
could also arise from the factor responsible for the relatively
low reactivity of 17b noted21 in connection with the deviation
of this point from eq 1.
Bordwell et al.36 had already drawn the conclusion that the
anti and syn elimination reactions are probably both carbanion
processes, and it is now appropriate to see how the present
results relate to their arguments, namely internal return and the
retardation effect. First we note that Thomas and Stirling’s
results show the isotopic exchange reactions to be very sensitive
to substituents (F* ) 4.89), and hence there is no longer any
discrepancy between exchange and elimination to require an
explanation in terms of internal return. In addition, the question
of internal return had been considered by Thomas and Stirling,11
who concluded from their observation of a sizable tritium isotope
effect (kH/kT ) 7.1) that internal return is probably not an
important feature of this hydrogen isotope exchange reaction.
The second effect postulated by Bordwell et al.36 was the
“retardation effect” in carbanion formation in six-membered
cyclic compounds to account for the observation that syn
eliminations in cyclohexyl compounds were apparently slow
relative to either (a) the analogous anti eliminations in cyclo-
hexyl or cyclopentyl substrates or (b) the corresponding syn
eliminations in cyclopentyl species. If we now look at these
results in the light of the present work, we note that the H-C-
C-O torsion angle in the cyclohexyl substrates undergoing syn
elimination (e.g., 37) is around 60°, and these substrates are
expected to form the carbanion more slowly than either the
trans-cyclopentyl compounds (reacting by syn eliminations), in
which the torsion angle can readily be close to 0°, or the cis
substrates in either the cyclopentyl or cyclohexyl systems, in
which the H-C-C-O torsion angle is close to 180°. The results
of the present study predict that the syn-cyclohexyl (θ ) 60°)
reactions will be slower than any of the other reactions (with 0
or 180° torsion angles). No further “retarding effect” is required
to account for the results, though it is, of course, always possible
that such an additional effect is operating in some measure.38
Finally, we note that the stereospecificity of the reactions of
threo- and erythro-3-p-toluenesulfonyl-2-butyl brosylates,
p-TolSO2(Me)CHCH(Me)OBs,36 also readily follows from the
present work. The threo and erythro diastereomers each have
three conformations about the C(2)-C(3) bond, but the kinetic
In view of the complexity of some of the discussion in this
paper to this point, it may be helpful to view this topic from a
perspective requiring specifically neither σ/ nor log kN values.
θ
A simple plot of log k for the reactions of 23c (H-D exchange),
36, and 37 (both eliminations) (data as presented by Hine and
Ramsay)35 vs our results for H-D exchange in 23b, 6b, and
17b is shown in Figure 3. The three points fit a straight line
with a slope of 2.7, and we are immediately led to the same
conclusions as those drawn in the previous paragraph. Note that
the slope is expected to reflect the ratio of the σ* values (1.31/
0.52 ) 2.5 if we take Hine and Ramsay’s values). The point
for the cis-tosylate (36) is slightly above the (projected) straight
line joining 23c and 37 (which has a slope of 2.5); this could
conceivably permit an E1cB mechanism for 37 and an E1cB-
like E2 reaction for 36; it is also possible that the deviation
(37) Assuming that the ratio of σ* to σ/180 is probably much the same
for all substituents. we readily get σ/180 values of for the respective groups
by multiplying the σ* values (CH3, 0.0; CH2F, 1.16; CH2Cl, 1.04; CH2-
(38) We wish to point out that we fully concur with the basis of the
skepticism of Bordwell et al., i.e., that “it does not seem possible to account
for more than ca. 103 acceleration [in hydrogen exchange] on the basis of
an inductive effect.” It was precisely this difficulty that we experienced
when faced with Thomas and Stirling’s results and that led us to the present
study, i.e., to look for another component of the polar effect in addition to
the inductive and field effects.
OMe, 0.56; and CH2OTs, 1.31) by (σ/
)OMe/(σ*)OMe ) 0.89/0.64 ) 1.41;
180
the σ/180 values so obtained were, respectively, 0.0, 1.64, 1.47, 0.79, and
1.85. The (σ/ )OTs ) 1.00 was estimated from (σ/ )OMe ) 0.49 (see above)
60
60
by multiplying by the scaling factor 1.31/0.64. The linear least-squares line
corresponded to the equation log k ) -6.1 + 4.1σ/θ, with r ) 0.985.

NegatiVe Hyperconjugation in the Polar Effect
J. Am. Chem. Soc., Vol. 122, No. 42, 2000 10319
acidity of the R-hydrogen in the one with the H-C-C-O
torsion angle of 180° must be much greater than that in either
of the other conformers (with θ ) 60°), and in each case that
conformer reacts much faster than the others to give the olefin
stereospecifically. This is perhaps the best place to point out
explicitly that the stereospecificity (whether syn or anti) of E1cB
and E2 reactions may be regarded as a consequence of the
torsion angle dependence of the acidifying effect of the
substituent-cum-nucleofuge, taken, of course, with any relevant
accompanying factors such as conformational populations.
Relation to Other Work. (a) By Ahlberg and Thibblin. It
is perhaps appropriate to point out here that hyperconjugative
stabilization of transition states for carbanion-forming elimina-
tion reactions has been proposed by Ahlberg and Thibblin.39
They have focused particularly on the borderline between E1cB
and E2 processes, and though their picture and ours would
appear to have a common feature at a fundamental level, their
work is distinctly different from ours; they have proposed and
applied a modified linear free energy relationship with two new
parameters related to leaving group activity, whereas our studies
have been primarily concerned with the experimental proof of
torsion angle dependence and examining its consequences.
(b) By Rappoport and Apeloig. A somewhat different
approach, involving substitution and addition reactions of
electrophilic olefins, has led Apeloig and Rappoport9e to propose
an ingenious rationalization involving negative hyperconjugative
stabilization of an intermediate carbanion as an important factor
influencing the stereochemistry of their reactions. In our view,
their conclusions and ours are fully consistent and mutually
supportive.
We have noted in our preliminary communication12 that our
“results constitute a good case for a kinetic anomeric effect not
readily accounted for by PLNM”. In our view the further work
reported in the present paper, notably the observation of both
synperiplanar and antiperiplanar effects, adds further weight to
our contention that the generalized anomeric effect (the peripla-
nar lone-pair hypothesis, PLPH?) does, and PLNM does not,
account for our results.
Generality. The present study, in our view, establishes that
the effect of â-alkoxy substituents on the rate of H-D exchange
via R-sulfonyl carbanions is subject to variation in rate cor-
relating with the H-C-C-O torsion angle. The work of
Thomas and Stirling also clearly showed that the reaction rate
conforms well to the Taft equation (with F* ) 4.89 for their
conditions) with a varied range of substituents. Since the
anomeric effect and negative hyperconjugation are not confined
to alkoxy groups, the only reasonable construction that we can
make from these observations is that we must expect the torsion
angle dependence to be a general phenomenon, capable of
appearing to a greater or lesser measure with all substituents.
The effect can be expected to be small or negligible with some
substituents (e.g., alkyl groups), but with most heteroatomic
substituents it must be significant. Obviously further experiments
are required to determine the general importance of torsion angle
dependence, and work on this problem is underway in our
laboratory.42
Experimental Part
General. Melting points were determined on a Kofler Hot Stage or
a Fisher Johns apparatus and are uncorrected. IR spectra were obtained
1
with a Bruker IFS 32 FTIR spectrometer and H NMR spectra with
(c) By Lambert. A study related and in some ways
complementary to ours of the effect of substituent torsion angles
has been carried out by Lambert and co-workers40 on the
â-trimethylsilyl substituent and its effect on the ease of
carbocation formation. In this admirable investigation, they find
a very large angle dependence of the substituent fully compatible
with (positive) hyperconjugation leading to stabilization of the
carbocation. It is probably not coincidental that the influence
of substituent hyperconjugation has first been most clearly
shown in the reactions of carbocations or carbanions. For one
thing, these are the reactions showing the greatest electron
supply or demand, and which can be expected to show the
greatest substituent dependency (highest absolute F* values) and
hence the most clearly evident hyperconjugative effect. In
addition, the carbocation- and carbanion-forming reactions are
processes strongly dependent on conjugative or hyperconjugative
influences. With aromatic systems these are the reactions which
lead to correlations with σ+ or σ-, greater through conjugation
than in the reference reaction; similarly, with the analogous
nonaromatic carbocation- or carbanion-forming systems positive
or negative hyperconjugation may be expected to be more in
evidence than in the reference process.
Varian XL-200 or XL-300 and Gemini 200 or 300 instruments; the
Gemini 300 instrument was used in all kinetic determinations; 13C NMR
spectra were determined (at 75 Hz) with the 300-MHz instruments.
Where not otherwise specified, spectra were run on CDCl3 solutions
referenced with TMS or the CHCl3 signal; TMS was used as the
reference for solutions in acetone-d6, and sodium trimethylsilylpro-
panesulfonate (DSS) was used for D2O solutions. Mass spectra were
run on a Finnigan MAT311A spectrometer. A Waters model 510 and
a Waters 490 programmable multiwavelength detector were used in
HPLC analyses (columns, C18, λ ) 260 nm; solvents, CH3OH:H2O
1:1). Most materials used for kinetic measurements and HRMS
molecular weight determinations showed 1H and 13C NMR spectra
appropriate to pure compounds; the two exceptions, 24a and 24b, were
obtained as specimens contaminated with up to 25% of the cis isomer
(25). The structural assignments for these latter compounds follow
unambiguously from their their mode of synthesis and spectra; the
signals due to the R-hydrogens were clearly visible, and their disap-
pearance was readily followed.
Reagent grade chemicals and solvents were used without purification
except where otherwise noted. Et3N was distilled from CaH2. CH3CN,
absolute EtOH, THF, and DMSO were dried over CaH2 and distilled;
CH2Cl2 was dried over P2O5 and Et2O over LiAlH4 before distillation.
Workup, where not described, involves, minimally, extraction with CH2-
Cl2 (at least 3 times), washing of the CH2Cl2 layers with water and, as
appropriate, with dilute aqueous acid (HCl or H2SO4) or base (NaOH
or NaHCO3), drying of the organic layer over anhydrous MgSO4 or
Na2SO4, and evaporation of the CH2Cl2 using a Bu¨chi Rotovap attached
to a water aspirator.
(d) By Sinnott. A vigorous criticism of the kinetic anomeric
effect (the antiperiplanar lone-pair hypothesis, ALPH), has been
put forward by Sinnott,41 who has argued that the principle of
least nuclear motion (PLNM) accounts well for observations
interpreted by others in terms of the kinetic anomeric effect.
Materials and Spectra. 1,4-Oxathiane 4,4-dioxide (13) (Ald-
1
rich): mp 131-133 °C; H NMR δ 3.13 (t, 4 H), 4.14 (t, 4 H); 13C
(39) Ahlberg, P. Chem. Scr. 1973, 3, 183-189. Thibblin, A.; Ahlberg,
P. J. Am. Chem. Soc. 1977, 99, 7926-7930. Thibblin, A.; Ahlberg, P. J.
Am. Chem. Soc. 1979, 101, 7311-7318. Thibblin, A. Chem. Scr. 1980, 15,
121-127. Thibblin, A. J. Am. Chem. Soc. 1988, 110, 4582-4586.
(40) See, for example: Lambert, J. B.; Liu, X. J. Organomet. Chem.
1996, 521, 203-210. Lambert, J. B.; Zhao, Y.; Emblidge, R. W.; Salvador,
L. A.; Liu, X.; So, J.-H.; Chelius, E. C. Acc. Chem. Res. 1999, 32, 183-
190.
NMR (CDCl3) δ 52.8, 66.1.
Thiane 1,1-dioxide (22) from oxidation of thiane (Aldrich) with
MCPBA: mp 98-99 °C (reported43 mp 98.5-99 °C); IR (KBr) νmax
(42) Preliminary experiments in this laboratory point to large torsion angle
effects with the alkylthio, alkylsulfonyl, dialkylamino, and trialkylammonio
groups; see: King, J. F.; Li. M. Proceedings of the 82nd Canadian Society
for Chemistry Conference, Toronto, Ontario, June 1999; Abstract 622.
(41) Sinnott, M. L. AdV. Phys. Org. Chem. 1988, 24, 113-204.

10320 J. Am. Chem. Soc., Vol. 122, No. 42, 2000
King et al.
2940 (s), 2956 (m), 1446 (w), 1319 (m), 1281 (vs), 1196 (m), 1130
(vs) 1067 (w), 963 (w), 849 (m) cm-1; 1H NMR δ 1.58 (sym m, 2 H),
2.03 (m, 4 H), 2.97 (t, 4 H); 13C NMR δ 23.8, 24.2, 52.1.
trans-1-Methoxy-2-(phenylsulfonyl)cyclohexane (6b). trans-1-
Chloro-2-(phenylthio)cyclohexane48 (4.5 g, 20 mmol) was refluxed in
methanol for 10 h to give, after workup, trans-1-methoxy-2-(phenyl-
thio)cyclohexane. Oxidation by hydrogen peroxide (30%) in acetic acid
trans-1,4-Oxathiadecalin 4,4-dioxide (15):14 mp 120-121 °C; IR
(KBr) νmax 2942 (s), 2870 (s), 1456 (m), 1308 (vs), 1269 (vs), 1227
(w), 1183 (m), 1130 (vs), 1024 (m), 939 (w), 853 (w), 729 (m), 538
1
gave 6b in good yield: mp 77-79 °C; H NMR δ 1.0-2.4 (m, 8 H),
3.00-3.12 (m, 1 H), 3.09 (s, 3 H), 3.46 (m, 1 H), 7.4-7.9 (m, 5 H);
13C NMR δ 23.4, 24.0, 24.6, 30.0, 55.6, 67.4, 127.8, 128.5, 132.9, 141.6;
calcd exact mass for C13H18O3S 254.0977, found 254.0975.
1
(s) cm-1; H NMR δ 0.97-2.45 (m, 8 H), 2.86 (dddd, 1 H, C-10Ha),
3.07 (d of t, Jgem ) 13.6, Jea ) 2.83, Jee ) 2.6, 1 H, C-3He), 3.28 (d of
q, Jgem ) 13.6, Jaa ) 10.9, Jae ) 5.3 Hz, 1 H, C-3Ha), 3.67 (quintet, 1
H, C-9Ha), 4.12 (d of t, Jgem ) 12.8, Jaa ) 10.9, Jae ) 2.83 Hz, 1 H,
C-2Ha), 4.28 (d of q, Jgem ) 12.8, Jee ) 2.6, Jea ) 5.3 Hz, 1 H, C-2He)
(the chemical shifts of 15 were dependent on the solvent); 13C NMR,
see ref 14a; HRMS calcd for C8H14O3S 190.0664, found 190.0663.
The three R-(R′-)monodeuterated, three R-(R′-)dideuterated, and the
single R,R′-trideuterated isotopomers of 15 have been prepared by
successive deuteration and dedeuteration; their 1H and 13C NMR spectra
are recorded elsewhere.44a
cis-1-Methoxy-2-(methylsulfonyl)cyclohexane (17a). trans-1-
Bromo-2-methoxycyclohexane49 (10 g, 0.0518 mol) and NaSCH3 (4.0
g, 0.0571 mol) were dissolved in 2-butanol and refluxed for 24 h.
Workup as usual gave cis-1-methoxy-2-(methylthio)cyclohexane (4.6
g, 55%), which was oxidized with H2O2 (30%) in acetic acid to give
17a (80%). Recrystallization from ethyl acetate-hexane gave 17a as
1
colorless crystals: mp 44-47 °C; H NMR δ 1.1-2.2 (m, 8 H), 2.73
(dt, J ) 12.7 and 2.9 Hz, 1 H), 2.82 (s, 3 H), 3.30 (s, 3 H), 4.0 (m, w1/2
) 6 Hz, 1 H); 13C NMR δ 18.7, 22.0, 25.1, 27.1, 39.4, 55.9, 67.1,
73.4; calcd exact mass for C8H16O3S 192.0820, found 192.0823.
3-Methoxythiane 1,1-dioxide (4):45 mp 66-68 °C (lit.45 mp 66-
1
66.5 °C); H NMR δ 1.27-1.48 (m, 1 H), 1.8-2.3 (m, 3 H), 2.76-
cis-1-Methoxy-2-(phenylsulfonyl)cyclohexane (17b). trans-1-
Bromo-2-methoxycyclohexane49 (15 g, 0.0777 mol) and NaSPh (9.4
g, 0.0855 mol) were dissolved in 2-butanol and refluxed for 48 h.
Workup as usual gave cis-1-methoxy-2-(phenylthio)cyclohexane (16
g, 93%). Oxidation of the sulfide by hydrogen peroxide (30%) in acetic
2.93 (m, 2 H), 2.96-3.09 (m, 1 H), 3.39 (s, 3 H), 3.42 (sym m, 1 H),
3.69 (t of t, J ) 10.7 and 3.8 Hz, 1 H); 13C NMR (CDCl3) δ 19.4,
30.4, 50.9, 56.0, 56.6, 75.7.
trans-1-Thiadecalin 1,1-dioxide (21):14a mp 114-115 °C (lit.14a mp
114-115.8 °C); 1H NMR δ 1.0-2.25 (m, 13 H), 2.59 (ddd, J ) 12.5,
10.0, 3.5 Hz, 1 H, C-9Ha), 2.91 (m, 1 H, C-2Ha), 3.10 (dddd, J ) 13.5,
4.0, 3.5, 1.4 Hz, 1 H, C-2He); 13C NMR, see ref 14a.
1
acid gave 17b: mp 98-98.5 °C; H NMR δ 1.0-2.2 (m, 8 H), 2.95
(m, 1 H), 3.17 (s, 3 H), 3.95 (m, 1 H), 7.4-7.9 (m, 5 H); 13C NMR δ
18.5, 21.6, 24.9, 27.3, 55.7, 67.9, 73.1, 128.3, 129.1, 133.1, 138.7; calcd
exact mass for C13H18O3S 254.0977, found 254.0977.
cis-1,4-Oxathiadecalin 4,4-Dioxide (14). trans-2-(2-Hydroxyethyl-
thio)cyclohexanol14c (5 g, 28.4 mmol) in DME (10 mL) with concen-
trated HCl (25 mL), on workup after standing overnight, gave the
dichlorosulfide (6 g), which on oxidation with 30% H2O2 (10 mL) in
HOAc (30 mL) and Ac2O (10 mL) gave, after workup, the dichloro-
sulfone (6.3 g); this in turn was mixed with 5% KOH (50 mL) and
DME (10 mL), and the mixture was refluxed for 3 h. Workup gave 14
as white crystals (4.5 g, 85%): mp (after recrystallization from ether-
cyclohexane) 130-131 °C; IR (KBr) νmax 2944 (s), 2869 (m), 1449
(w) 1393 (m), 1291 (vs), 1240 (s), 1136 (s), 1121 (s), 1094 (vs), 1021
(m), 959 (w), 855 (w), 527 (m); 1H NMR δ 1.16-2.20 (m, 8 H), 2.79
Cyclohexyl methyl sulfone (23a):32,50 NMR, see ref 32.
Cyclohexyl phenyl sulfone (23b):51 NMR, see ref 52.
cis-2-Methyl-1-(phenylsulfonyl)cyclohexane (25b):53 mp 73-74 °C
(lit.53 mp 74-74.5 °C); 1H NMR (CDCl3) δ 1.00-2.00 (m, 8 H), 1.20
(d, J ) 7.1 Hz, 3 H), 2.45 (symmetric m, 1 H), 2.99 (dt, J ) 12, 3.8
Hz, 1 H), 7.48-7.93 (m, 5 H); 13C NMR (CDCl3) δ 13.3, 19.3, 20.2,
25.5, 27.7, 33.4, 66.3, 128.4, 129.0, 133.3, 138.8.
trans-2-Methyl-1-(phenylsulfonyl)cyclohexane (24b).53 cis-2-Meth-
yl-1-(phenylsulfonyl)cyclohexane (25b, 1.0 g) was dissolved in dioxane
(20 mL), and sodium hydroxide solution (1 M, 20 mL) was added.
The resulting solution was refluxed for 5 days. Workup as usual gave
a cis/trans mixture of the 2-methyl-1-(phenylsulfonyl)cyclohexanes.
Several recrystallizations using different solvents gave a roughly 4:1
mixture of the trans/cis epimers. Thin-layer chromatography using
different solvents did not effect separation; mp of the mixture was 86-
87 °C. The mixture was used directly in the determination of the H-D
exchange rate. Data for the trans epimer (24b): 1H NMR (CDCl3)
0.80-2.20 (m, 9 H), 1.22 (d, J ) 7.1 Hz, 3 H), 2.71 (ddd, J ) 11.3,
9.5, 3.4 Hz, 1 H), 7.45-7.95 (m, 5 H); 13C NMR (CDCl3) δ 21.4,
24.8, 25.0, 27.4, 32.3, 35.4, 69.1 (the 13C NMR signals for the carbons
of benzene ring for the trans epimer (24b) are not clear because of
overlap with the signals of the cis epimer (25b)); calcd exact mass for
C13H19O2S (M + 1) (the mixture) 239.1106, found 239.1100.
cis-2-Methyl-1-(methylsulfonyl)cyclohexane (25a). Methyl iodide
(1.1 g, 7 mmol) was added to a solution of 2-methylcyclohexanethiol53
(0.30 g, 2.3 mmol) and NaOH (0.09 g) in aqueous ethanol (1:1, 2 mL).
After the resulting mixture was stirred for 30 min, it was poured into
water, and the sulfide was obtained by extraction with petroleum ether.
Oxidation of the crude sulfide with 30% of H2O2 in acetic acid gave
the crude sulfone (95% from the thiol) as a cis/trans mixture.
Recrystallization from methanol gave pure 25a (0.21 g): mp 86-88
°C; 1H NMR (CDCl3) δ 1.16 (d, J ) 7.0 Hz, 3 H), 1.2-2.0 (m, 8 H),
2.57 (symmetric m, 1 H), 2.81 (3 H, s), 2.96 (dt, J ) 11.9, 3.8 Hz, 1
H); 13C NMR (CDCl3) δ 13.3, 19.3, 20.8, 25.5, 27.8, 33.3, 39.4, 65.5;
calcd exact mass for C8H17O2S (M + 1) 177.0949, found 177.0943.
(sym m, 2 H, C-3He, and C-10He), 3.25 (d of q, Jgem ) 12.2, Jaa
11.1, and Jae ) 4.9 Hz, 1 H, C-3Ha), 4.00 (d of t, Jgem ) 12.4, Jaa
)
)
11.1, and Jae ) 2.2 Hz, 1 H, C-2Ha), 4.20 (br m, 1 H, C-9Ha), 4.25 (d
of q, Jgem ) 12.5, Jee ) 4.9, Jea ) 2.5 Hz, 1 H, C-2He) (the chemical
shifts were dependent on solvent); 13C NMR δ 19.1, 22.5, 24.7, 31.7,
47.5, 62.4, 65.2, 74.0; HRMS calcd for C8H14O3S 190.0664, found
190.0660.
Equilibration. 15 (500 mg) in 10% KOH in H2O-EtOH (3:1) (25
mL) was heated at 100 °C for ∼100 h. The cooled reaction mixture
was acidified with 5% HCl solution and extracted with CH2Cl2.
Evaporation of the solvent gave, in quantitative yield, a 9:1 mixture of
15 and 14, as determined by 1H and 13C NMR spectra; similar treatment
of 14 gave a mixture with the same NMR spectra as that from 15.
3-(Phenylsulfonyl)tetrahydropyran (5). 3-(Phenylthio)tetrahydro-
pyran46 was oxidized by H2O2 (30%) in acetic acid and gave a
quantitative yield of 5 as a colorless liquid: 1H NMR δ 1.45-2.25
(m, 4 H), 3.05-3.23 (m, 1 H), 3.27 (t, J ) 11.2 Hz, 1 H), 3.47 (t, J )
11.2 Hz, 1 H), 3.84 (d, J ) 11.2 Hz, 1 H), 4.07 (d, J ) 11.2 Hz, 1 H),
7.50-7.90 (m, 5 H); 13C NMR δ 22.8, 24.6, 60.3, 66.0, 67.6, 128.7,
129.2, 133.9, 137.2; calcd exact mass for C11H15O3S (M + 1) 227.0742,
found 227.0737.
trans-1-Methoxy-2-(methylsulfonyl)cyclohexane (6a):47 NMR, see
ref 47.
(43) Grishkevich-Trokhimovskii, E.; Tzuikina, O. J. Russ. Phys. Chem.
Soc. 1916, 48, 928-943; Chem. Abs. 1917, 11, 786.
(44) (a) Rathore, R. Ph.D. Thesis, University of Western Ontario, 1991.
(b) Guo, Z. R. Ph.D. Thesis, University of Western Ontario, 1994. (c) Li,
M. Ph.D. Thesis, University of Western Ontario, 1999.
(45) Fehnel, E. A.; Lackey, P. A. J. Am. Chem. Soc. 1951, 73, 2473-
2479. Also: Fehnel, E. A. J. Am. Chem. Soc. 1952, 74, 1569-1574.
(46) Martin, M.; Bassery, L.; Leroy, C. Bull. Soc. Chim. Fr. 1972, 4763-
4766.
(48) Wittig, G.; Vidal, F.; Chem. Ber. 1948, 81, 368-371.
(49) Grady, G. L.; Chokshi, S. K. Synthesis 1972, 483-484.
(50) Truce, W. E.; Milionis, J. P. J. Org. Chem. 1952, 17, 1529-1533.
(51) Cunneen, J. I. J. Chem. Soc. 1947, 36-40.
(52) Curran, D. P.; Totleben, M. J. J. Am. Chem. Soc. 1992, 114, 6050-
6058.
(47) Carren˜o, M. C.; Carretero J. C.; Garcia Ruano, J. L.; Rodriguez, J.
H.; Tetrahedron 1990, 46, 5649-5664.
(53) Bordwell, F. G.; Hewett, W. A. J. Am. Chem. Soc. 1957, 79, 3493-
3496.

NegatiVe Hyperconjugation in the Polar Effect
J. Am. Chem. Soc., Vol. 122, No. 42, 2000 10321
trans-2-Methyl-1-(methylsulfonyl)cyclohexane (24a). 25a (100 mg)
in methanol containing sodium methoxide (5%) was refluxed for 2 days.
Workup as usual gave a 3:1 mixture of 24a and 25a as a semisolid
(100 mg); this mixture was used as such to obtain the rate constant for
24a. Data for 24a: 13C NMR δ 21.1, 25.0, 25.1, 27.1, 32.8, 35.4, 39.1,
68.7; calcd exact mass for C8H17O2S (M + 1) 177.0949, found
177.0946.
(Phenylthio)-exo-2-norborneol (3.50 g, 15.9 mmol) in dichloromethane
(10 mL) was added dropwise over about 5 min, and the mixture was
stirred for 30 min at -78 °C. Triethylamine (7.26 g, 71.7 mmol) was
added dropwise over 5 min, and stirring was continued at -78 °C for
30 min. The cooling bath was removed, and water (30 mL) was added
at room temperature. The mixture was stirred for ca. 10 min, and the
organic layer was separated. Workup gave endo-3-(phenylthio)-2-
norbornanone (3.47 g, 15.9 mmol, 99% yield), which was distilled at
245 °C (oil bath, 8 Torr) to give a slightly yellowish oil: IR (neat)
trans-4-tert-Butylcyclohexyl phenyl sulfone (26) and cis-4-tert-
butylcyclohexyl phenyl sulfone (27):54 NMR, see ref 55.
Phenyl ethyl sulfone (31a):56 1H NMR, see ref 57; 13C NMR, see
ref 58.
ν
max 3478 (w), 3058 (m), 2967 (s), 2876 (m), 1746 (s), 1583 (m), 1480
(s), 1439 (m), 1316 (m), 1293 (m), 1157 (s), 1073 (s), 1026 (m), 941
(m), 742 (s), 693 (s) cm-1; H NMR δ 1.62-2.06 (m, 6 H), 2.73 (m,
1
Phenyl propyl sulfone (31b):56 13C NMR δ 12.9, 16.5, 57.9, 128.0,
2 H), 3.71 (dd, 1 H), 7.24 (m, 3 H), 7.43 (m, 2 H); 13C NMR δ 21.8,
25.2, 36.2, 40.7, 49.8, 59.7, 126.9, 128.9, 131.0, 134.9, 213.4; HRMS
calcd for C13H14OS 218.0765, found 218.0769.
1
129.2, 133.6, 139.1; H NMR, see ref 59.
Phenyl butyl sulfone (31c):56 1H NMR, see ref 60.
Phenyl pentyl sulfone (31d):60 13C NMR δ 13.6, 22.0, 22.2, 30.3,
1
endo-3-(Phenylthio)-endo-2-norbornanol. endo-3-Phenylthio-2-
norbornanone (1.50 g, 6.87 mmol) in anhydrous diethyl ether (10 mL)
was added dropwise over 6 min to a slurry of lithium aluminum hydride
(0.508 g, 6.87 mmol) in anhydrous diethyl ether (20 mL) at 0 °C. The
mixture was stirred for 0.5 h at 0 °C. Ethylene glycol (1.706 g, 27.48
mmol) was then added slowly at 0 °C. Stirring was continued for 5
min at 0 °C, and the mixture was filtered. The solid was washed with
dichloromethane (3 × 30 mL). The combined filtrate was washed with
water (3 × 30 mL), and the organic layer was dried over anhydrous
magnesium sulfate. The solvent was evaporated to give the product
(1.42 g, 6.41 mmol, 94.2% yield) as a colorless liquid: IR (neat) νmax
3443 (s), 3058 (w), 2957 (s), 2876 (s), 1586 (s), 1480 (s), 1452 (m),
1383 (m), 1320 (m), 1296 (m), 1125 (m), 1090 (s), 1063 (s), 1026 (s),
56.2, 127.9, 129.2, 133.5, 139.1. H NMR, see ref 61.
Phenyl hexyl sulfone (31e):62 13C NMR δ 13.6, 22.2, 22.5, 27.9,
1
31.1, 56.3, 128.0, 129.2, 133.5, 139.1; H NMR, see ref 59.
2-Methyl-1,4-oxathiane 4,4-Dioxide (16). Concentrated HCl (25
14d
mL) was added slowly to a solution of HOCH2CH2SCH2CH(OH)CH3
(5 g, 37 mmol) in DME (10 mL), and the mixture was left at room
temperature overnight. Workup gave ClCH2CH2SCH2CHClCH3 as a
clear oil (6.2 g, ∼100%): 1H NMR δ 1.56 (d, 3 H), 2.74-3.00 (m, 4
H), 3.62 (t, 2 H), 4.09 (sextet, 1 H); 13C NMR δ 23.8, 35.0, 41.9, 42.9,
56.7. The dichlorosulfide (2 g, 11.8 mmol) on oxidation with MCPBA
(4 g, 24 mmol), after workup, gave ClCH2CH2SO2CH2CHClCH3 as a
viscous liquid: IR (neat) νmax 2984 (s), 2934 (m), 1451 (m), 1395 (s),
1318 (vs), 1125 (vs), 1034 (s), 868 (s), 735 (m) cm-1; 1H NMR δ 1.71
(d, 3 H), 3.34-3.79 (m, 4 H), 3.93 (t, 2 H), 4.56 (sym m, 1 H); 13C
NMR δ 25.2, 35.6, 49.7, 56.5, 62.9. The sulfone (2.5 g) in DME (5
mL) was refluxed for 5 h with 10% aqueous KOH (15 mL). Workup
gave 16 (mp 110-111 °C from ether-petroleum ether): IR (KBr) νmax
2974 (m), 2936 (m), 1375 (s), 1314 (s), 1287 (vs), 1256 (s), 1196 (s),
1127 (vs), 1082 (vs 509 (s) cm-1; 1H NMR δ 1.34 (d, 3 H), 2.83-3.23
(m, 4 H), 3.95-4.13 (m, 2 H), 4.30 (ddq, J ) 12.7, 2.3, 4.7 Hz); 13C
NMR δ 21.2, 51.5, 58.4, 65.0, 72.7; HRMS calcd for C5H10O3S
150.0351, found 150.0352.
1
738 (s), 691 (s) cm-1; H NMR δ 1.47 (m, 4 H), 1.65 (m, 1 H), 1.87
(m, 1 H), 2.55 (d, 2 H), 2.96 (d, 1 H), 3.62 (dq, 1 H), 4.18 (m, 1 H),
7.27 (m, 5 H); 13C NMR δ 19.49, 23.53, 36.09, 42.63, 42.68, 55.64,
69.68, 126.36, 129.02, 129.46, 136.38; HRMS calcd for C13H16OS
220.0922, found 220.0920.
endo-3-(Phenylthio)-endo-2-methoxynorbornane. endo-3-(Phe-
nylthio)-endo-2-norbornanol (1.200 g, 5.45 mmol) in DMSO (3 mL)
was added to a solution of potassium hydroxide (1.220 g, 21.8 mmol)
in DMSO (7 mL) at room temperature. The mixture was stirred at room
temperature for 1-2 min. Iodomethane (3.094 g, 21.8 mmol) was
added, and stirring was continued for 15 min. Water (30 mL) and
dichloromethane (40 mL) were added. Workup gave the product (1.263,
99% yield) as a slightly yellowish liquid: IR (neat) νmax 3058 (w),
2955 (s), 2872 (s), 2824 (m), 1585 (m), 1480 (s), 1439 (s), 1360 (m),
endo-3-(Phenylthio)-exo-2-norbornanol. Thiophenol (13.00 g,
0.118 mol) and exo-norbornene oxide63 (8.50 g, 0.077 mol) were added
to a solution of sodium hydroxide (6.00 g, 0.15 mol) in 50% aqueous
EtOH solution (v/v, 75 mL), and the mixture was refluxed for 100 h
and then cooled to room temperature. Workup gave a brown oil which
solidified on standing; recrystallization from toluene and petroleum ether
gave endo-3-(phenylthio)-exo-2-norbornanol as white crystals (15.21
g, 0.069 mol, 89.7% yield): mp 64-65 °C; IR (KBr) νmax 3291 (s),
3055 (s), 2940 (s), 2961 (s), 1582 (m), 1482 (m), 1437 (s), 1318 (m),
1
1294 (w), 1190 (m), 1129 (s), 1109 (s), 957 (m), 909 (w) cm-1; H
NMR δ 1.34 (m, 4 H), 1.78 (m, 1 H), 2.05 (m, 1 H), 2.30 (s, 1 H),
2.53 (s, 1 H), 3.33 (s, 3 H), 3.69 (dd, 1 H), 3.81 (dd, 1 H), 7.24 (m, 3
H), 7.35 (m, 2 H); 13C NMR δ 19.9, 22.8, 35.8, 40.0, 41.1, 52.6, 57.9,
80.4, 125.8, 128.7, 130.2, 137.1; HRMS calcd for C14H18OS 234.1078,
found 234.1077.
1
1117 (m), 1053 (s), 1005 (s), 814 (m), 737 (s), 688 (s); H NMR δ
1.22 (m, 1 H), 1.39 (m, 2 H), 1.58 (m, 1 H), 1.78 (m, 3 H), 2.20 (d, 1
H), 2.42 (d, 1 H), 3.29 (m, 1 H), 3.56 (t, 1 H), 7.21-7.28 (m, 3 H),
7.42-7.46 (m, 2 H); 13C NMR δ 22.3, 24.7, 35.6, 40.6, 44.9, 59.1,
81.9, 126.1, 128.9, 130.0, 136.6; HRMS calcd for C13H16OS 220.0922,
found 220.0924.
endo-3-(Phenylthio)-2-norbornanone. Dimethyl sulfoxide (3.438
g, 44 mmol) was added dropwise over ca. 5 min to a solution of oxalyl
chloride (2.539 g, 20 mmol) in dichloromethane (30 mL) at -78 °C.
The reaction mixture was stirred for 15 min at -78 °C. endo-3-
endo-3-(Phenylsulfonyl)-endo-2-methoxynorbornane (10). Hy-
drogen peroxide (30%, 8.5 mL) was added dropwise to a solution of
endo-3-(phenylthio)-endo-2-methoxynorbornane (1.03 g, 4.4 mmol) in
acetic acid (3 mL) at room temperature. The mixture was heated on a
steam bath for 30 min. Workup gave the product 10 (1.12 g, 95% yield)
as white crystals, which were recrystallized from 85% methanol: mp
96-98 °C; IR (KBr) νmax 3058 (w), 2946 (s), 2874 (s), 2830 (m), 1447
(s), 1327 (s), 1310 (s), 1279 (s), 1248 (m), 1144 (s), 1113 (s), 1084
(s), 1073 (s), 720 (s), 693 (s) cm-1; 1H NMR δ 1.43 (m, 4 H), 2.06 (m,
1 H), 2.46 (m, 1 H), 2.60 (m, 2 H), 3.29 (s, 3 H), 3.45 (ddd, 1 H), 3.84
(ddd, 1 H), 7.56 (m, 3 H), 7.91 (m, 2 H); 13C NMR δ 19.4, 22.6, 36.8,
40.1, 40.5, 58.1, 67.0, 80.0, 128.4, 128.8, 133.0, 141.4; HRMS calcd
for C14H19O3S (M + H) 267.1055, found 267.1060.
(54) Eliel, E. L.; Ro, R. S. J. Am. Chem. Soc. 1957, 79, 5995-6000.
(55) Duddeck, H.; Korek, U.; Rosenbaum, D.; Drabowicz, J. Magn.
Reson. Chem. 1986, 24, 792-797.
(56) Baldwin, W. A.; Robinson, R. J. Chem. Soc. 1932, 1445-1451.
(57) Ishida, M.; Minami, T.; Agawa, T. J. Org. Chem. 1979, 44, 2067-
2073.
endo-3-(Phenylsulfonyl)-exo-2-norbornanol. Hydrogen peroxide
(30%, 34 mL) was added dropwise to a solution of endo-3-(phenylthio)-
exo-2-norbornanol (3.93 g, 17.8 mmol) in acetic acid (13 mL) as above
to give the product (4.08 g, 91.0% yield) as a white solid, which was
recrystallized from 60% ethanol: mp 114-115 °C; IR (KBr) νmax 3453
(s), 3067 (w), 2940 (s), 2872 (m), 1304 (m), 1284 (s), 1269 (s), 1134
(s) 1086 (s), 1067 (s), 720 (s), 604 (s); 1H NMR δ 1.39 (m, 2 H), 1.70
(m, 2 H), 2.13 (m, 2 H), 2.33 (d, 1 H), 2.51 (d, 1 H), 3.19 (dt, 1 H),
4.29 (m, 1 H), 7.61 (m, 3 H), 7.91 (m, 2 H); 13C NMR δ 22.2, 24.3,
(58) Birchall, J. D.; Glidewell, C. J. Chem. Soc., Dalton Trans. 1978,
604-607.
(59) Manescalchi, F.; Orena, M.; Savoia, D. Synthesis 1979, 445-446.
(60) Babler, J. H. J. Org. Chem. 1987, 52, 4614-4616.
(61) Pine, S. H.; Shen, G.; Bautista, J.; Sutton, C., Jr; Yamada, W.;
Apodaca, L. J. Org. Chem. 1990, 55, 2234-2237.
(62) Katritzky, A. R.; Saba, A.; Patel, R. C. J. Chem. Soc., Perkin Trans.
1 1981, 1492-1494.
(63) Walborsky, H. M.; Loncrini, D. F. J. Am. Chem. Soc. 1954, 76,
5396-5399.

10322 J. Am. Chem. Soc., Vol. 122, No. 42, 2000
King et al.
37.0, 39.3, 45.2, 74.7, 76.2, 127.9, 129.3, 133.6, 140.2; HRMS calcd
for C13H16O3S 252.0820, found 252.0822.
72.9, 81.9, 128.4, 129.1, 133.5, 138.8; HRMS calcd for C14H19O3S (M
+ H) 267.1055, found 267.1057.
exo-3-(Phenylthio)-exo-2-norbornanol was partially separated from
the above mixture of the exo-exo and exo-endo isomers by column
chromatography on silica gel: 1H NMR δ 1.15-1.77 (m, 4 H), 2.35
(m, 2 H), 2.85 (m, 2 H), 3.35 (dd, 1 H), 3.92 (dd, 1 H), 7.21-7.38 (m,
5 H); 13C NMR δ 25.2, 30.0, 34.5, 44.4, 45.3, 59.8, 76.0, 127.2, 130.1,
130.4, 137.7.
exo-3-(Phenylthio)-exo-2-methoxynorbornane. Powdered potas-
sium hydroxide (0.20 g, 3.6 mmol) was added to a solution of a mixture
of exo-3-(phenylthio)-exo-2-norbornanol (80%) and exo-3-(phenylthio)-
endo-2-norbornanol (20%) (0.115 g, 0.52 mmol) in DMSO (2 mL).
This mixture was stirred for 2 min and treated with iodomethane (0.74
g, 5.2 mmol) as above, giving a mixture of exo-exo (80%) and exo-
endo ethers (20%) as a yellowish oil (0.114 g, 0.49 mmol, 94% yield).
Spectral data of exo-3-(phenylthio)-exo-2-methoxynorbornane: 1H
NMR δ 1.40-1.70 (m, 4 H), 1.92 (m, 2 H), 2.22 (m, 1 H), 2.57 (m, 1
H), 3.39 (s, 3 H), 3.41 (m, 1 H), 3.45 (m, 1 H), 7.29 (m, 3 H), 7.51 (m,
2 H); 13C NMR δ 24.8, 34.0, 39.8, 41.5, 43.5, 56.6, 59.9, 86.7, 126.4,
129.8, 130.2, 138.9.
exo-3-(Phenylsulfonyl)-exo-2-methoxynorbornane (9). Hydrogen
peroxide (30%, 2 mL) was added dropwise to a solution of a mixture
of exo-3-(phenylthio)-exo-2-methoxynorbornane (80%) and the exo-
endo isomer (20%) (0.114 g, 0.49 mmol) in acetic acid (1 mL) as above
to give a mixture (0.125 g, 0.47 mmol, 95.9% yield) of exo-3-
(phenylsulfonyl)-exo-2-methoxynorbornane (9) (major) and exo-3-
(phenylsulfonyl)-endo-2-methoxynorbornane (2) as a colorless oil which
solidified slowly on standing. 9 was separated from the mixture by
column chromatography using ethyl acetate (20%) and petroleum ether
(80%) as the eluate and recrystallized from ethyl acetate and petroleum
ether: mp 98-99 °C; IR (KBr) νmax 2980 (s), 2877 (s), 1450 (m), 1303
(s), 1124 (s), 1103 (s), 719 (m), 691 (m) cm-1; 1H NMR δ 0.97 (m, 1
H), 1.14-1.1.20 (m, 2 H), 1.58 (m, 2 H), 1.95 (m, 1 H), 2.38 (s, 1 H),
2.78 (s, 1 H), 3.23 (s, 3 H), 3.29 (m, 1 H), 3.47 (m, 1 H), 7.52 (m, 3
H), 7.93 (m, 2 H); 13C NMR δ 23.4, 29.6, 34.2, 38.5, 40.1, 58.4, 72.0,
85.2, 128.6, 128.7, 132.9, 141.4; HRMS calcd for C14H19O3S (M + H)
267.1055, found 267.1057.
endo-3-(Phenylsulfonyl)-exo-2-methoxynorbornane (3). endo-3-
(Phenylsulfonyl)-exo-2-norborneol (0.25 g, 1.0 mmol) was added to a
solution of potassium hydroxide (0.224 g, 4.0 mmol) in DMSO (2 mL)
and treated with iodomethane (0.57 g, 4.0 mmol) as above. Workup
gave the product (3) (0.238 g, 0.89 mmol, 89% yield) as a white solid,
which was recrystallized from 60% methanol: mp 125-126 °C; IR
(KBr) νmax 3069 (w), 2976 (m), 2940 (m), 2880 (m), 1447 (s), 1363
(m), 1298 (s), 1269 (s), 1150 (s), 1130 (s), 1084 (s), 756 (s), 720 (s),
689 (s), 606 (s) cm-1; 1H NMR δ 1.27-1.41 (m, 3 H), 1.58-1.67 (m,
2 H), 2.20 (m, 1 H), 2.50 (m, 2 H), 3.16 (s, 3 H), 3.18 (m, 1 H), 3.67
(m, 1 H), 7.29 (m, 3 H), 7.63 (m, 2 H); 13C NMR δ 22.9, 23.7, 37.0,
38.8, 41.3, 56.9, 74.0, 83.5, 127.8, 129.0, 133.3, 140.2; HRMS calcd
for C14H19O3S (M + H) 267.1055, found 267.1059.
exo-3-(Phenylthio)-2-norbornanone. endo-3-(Phenylthio)-2-nor-
bornanone (3.80 g, 17.4 mmol) was added to a solution of potassium
hydroxide (1.95 g, 34.8 mmol) in methanol (70 mL) and water (10
mL). The mixture was refluxed for 2 h, and water (20 mL) was added.
Most of the methanol was removed on a rotary evaporator, and the
residue was worked up as usual to give a yellowish liquid (3.59 g,
16.5 mmol, 95% yield). The 1H NMR spectrum showed the product to
consist of endo-3-(phenylthio)-2-norbornanone (40%) and exo-3-
phenylthio-2-norbornanone (60%), which were separated by column
chromatography on silica gel using diethyl ether-petroleum ether, 20:
80. The exo isomer eluted first; this was distilled in a coldfinger
apparatus at 235-237 °C (oil bath, 8 Torr) to give a colorless oil: IR
(neat) ν
3476 (w), 3058 (w), 2963 (s), 2876 (s), 1748 (s), 1581
max
(m), 1480 (s), 1439 (s), 1304 (s), 1172 (s), 1073 (s), 1026 (m), 937
1
(m), 911 (m), 747 (s), 691 (s) cm-1; H NMR δ 1.55 (m, 3 H), 1.86
(m, 2 H), 2.17 (d, 1 H), 2.61 (s, 1 H), 2.69 (s, 1 H), 3.31 (d, 1 H), 7.30
(m, 3 H), 7.45 (m, 2 H); 13C NMR δ 24.3, 27.2, 35.1, 42.0, 49.3, 56.5,
127.2, 129.0, 131.3, 134.4, 213.3; HRMS calcd for C13H14OS 218.0765,
found 218.0765.
exo-3-(Phenylthio)-endo-2-norbornanol. exo-3-(Phenylthio)-2-nor-
bornanone (1.330 g, 6.09 mmol) in THF (10 mL) was added dropwise
to a slurry of lithium aluminum hydride (0.300 g, 4.06 mmol) in THF
(15 mL) at 0 °C, and after the mixture was stirred at 0 °C for 1.5 h, it
was worked up to give a mixture (24:76 by 1H NMR) of exo-3-
(phenylthio)-exo-2-norbornanol and exo-3-(phenylthhio)-endo-norbor-
nanol (1.070 g, 4.86 mmol, 79.7% yield) as a colorless oil which gave
white crystals on standing. The solid was recrystallized from petroleum
ether (60-80 °C) to give pure exo-3-(phenylthio)-endo-2-norbornanol
as white crystals: mp 69-70 °C; IR (KBr) νmax 3237 (s), 2957 (s),
2943 (s), 2869 (s), 1584 (m), 1480 (s), 1437 (m), 1337 (m), 1318 (w),
1157 (m), 1067 (s), 1053 (s), 740 (s), 700 (m), 689 (s) cm-1; 1H NMR
δ 1.31-1.46 (m, 3 H), 1.61-1.80 (m, 4 H), 2.23 (m, 1 H), 2.36 (m, 1
H), 2.85 (m, 1 H), 4.05 (m, 1 H), 7.18-7.34 (m, 5 H); 13C NMR δ
19.5, 29.4, 35.3, 43.0, 43.7, 57.4, 80.2, 126.1, 129.0, 129.6, 136.6;
HRMS calcd for C13H16OS 220.0922, found 220.0924.
exo-2-(Phenylsulfonyl)norbornane (19). Thiophenol (3.51 g, 31.9
mmol) was added in small portions to norbornene (3.0 g, 31.9 mmol)
at room temperature. The temperature rose to 60 °C about 5 min after
the first portion was added. The remainder of the thiophenol was added
at such a rate that the temperature remained between 60 and 70 °C.
After all of the thiophenol was added, the mixture was heated at 75-
80 °C for 30 min with stirring. This was then subjected to vacuum
distillation. The product was collected at 128-130 °C (5 Torr) to give
exo-(phenylthio)norbornane as a colorless oil (5.50 g, 84.4% yield):
IR (neat) νmax 3059 (w), 2955 (s), 2869 (s), 1586 (s), 1480 (s), 1451
(s), 1439 (s), 1313 (s), 1239 (m), 1269 (m), 1138 (m), 1068 (m), 1091
1
(s), 1026 (s), 953 (m), 736 (s), 690 (s) cm-1; H NMR δ 1.18-1.84
(m, 8 H), 2.27 (m, 2 H), 3.17 (m, 1 H), 7.11-7.32 (m, 5 H); 13C NMR
δ 31.3, 31.5, 38.2, 39.1, 41.2, 44.9, 50.7, 128.1, 131.4, 131.5, 140.4.
Hydrogen peroxide (30%, 20 mL) was added dropwise to a solution
of the exo-thioether (3.0 g, 14.7 mmol) in acetic acid (10 mL) as above
to give white crystals of 19 (3.456 g, 99.5% yield), which were
recrystallized from methanol: mp 80-81 °C; IR (KBr) νmax 3063 (w),
2965 (s), 2872 (m), 1445 (m), 1325 (m), 1298 (s), 1271 (s), 1242 (m),
1146 (s), 1086 (s), 721 (s), 696 (s), 612 (s) cm-1; 1H NMR δ 1.18 (m,
3 H), 1.80 (m, 1 H), 2.01 (m, 1 H), 2.40 (dd, 1 H), 2.68 (dd, 1 H), 3.00
(ddd, 1 H), 7.60 (m, 3 H), 7.88 (m, 2 H); 13C NMR δ 28.1, 29.8, 32.6,
36.1, 38.8, 66.7, 128.4, 129.1, 133.4, 139.1; HRMS calcd for C13H16O2S
(M + 1) 237.0949, found 237.0949.
exo-3-(Phenylthio)-endo-2-methoxynorbornane. A mixture of
DMSO (6.5 mL) and powdered potassium hydroxide (0.759 g, 13.5
mmol) was stirred for 5 min at room temperature. exo-3-(Phenylthio)-
endo-2-norbornanol (0.745 g, 3.38 mmol) was treated with iodomethane
(2.399 g, 16.9 mmol) as above; workup gave the product (0.750 g, 3.2
mmol, 94.7% yield) as a slightly yellowish liquid: 1H NMR δ 1.37
(m, 3 H), 1.79 (m, 3 H), 2.22 (m, 1 H), 2.51 (m, 1 H), 2.89 (t, 1 H),
3.30 (s, 3 H), 3.55 (m, 1 H), 7.30 (m, 5 H); 13C NMR δ 20.7, 30.1,
35.8, 40.9, 44.0, 55.6, 58.2, 89.3, 126.7, 129.7, 130.2, 138.0.
exo-3-(Phenylsulfonyl)-endo-2-methoxynorbornane (2). Hydrogen
peroxide (30%, 8.0 mL) was added dropwise to a solution of exo-3-
(phenylthio)-endo-2-methoxynorbornane (0.75 g, 3.20 mmol) in acetic
acid (3 mL) as above to give exo-3-(phenylsulfonyl)-endo-methox-
ynorbornane (2) (0.814 g, 95.5% yield) as white crystals, which were
recrystallized from 50% MeOH: mp 87-88 °C; IR (KBr) νmax 2979
(m), 2959 (s), 2876 (s), 2836 (m), 1451 (m), 1304 (s), 1291 (s), 1215
(m), 1148 (s), 1111 (s), 1084 (s), 903 (m), 760 (m), 725 (s), 693 (s),
endo-2-(Phenylsulfonyl)norbornane (20). A mixture of formic acid
(3.05 g, 66.3 mmol), 10% palladium-on-charcoal catalyst (Pd-C) (0.35
g), and endo-2-(phenylsulfonyl)-5-norbornene64 (0.49 g, 2.1 mmol) in
absolute ethanol (80 mL) was refluxed for 2 h. The Pd-C was filtered
off, and the solvent was removed to give a colorless liquid (0.40 g,
81% yield), which solidified on standing; 20 was recrystallized from
50% aqueous methanol: mp 61-62 °C; IR (KBr) νmax 2975 (m), 2883
1
606 (s), 550 (s) cm-1; H NMR δ 1.30 (m, 3 H), 1.63 (m, 2 H), 1.94
(d, 1 H), 2.59 (d, 2 H), 2.73 (dd, 1 H), 3.17 (s, 3 H), 4.05 (t, 1 H), 7.59
(m, 3 H), 7.91 (d, 2 H); 13C NMR δ 19.5, 29.8, 34.9, 39.0, 39.1, 57.4,
(64) Maccagnani, G.; Montanari, F.; Taddei, F. J. Chem. Soc. B 1968,
453-458.

NegatiVe Hyperconjugation in the Polar Effect
J. Am. Chem. Soc., Vol. 122, No. 42, 2000 10323
(w), 1445 (m), 1276 (s), 1152 (s), 1086 (s), 722 (s), 703 (s), 603 (s);
1H NMR δ 1.37 (m, 5 H), 1.77 (m, 2 H), 2.38 (m, 2 H), 2.52 (m, 1 H),
3.32 (m, 1 H), 7.52 (m, 3 H), 7.86 (m, 2 H); 13C NMR δ 23.6, 28.9,
31.2, 37.3, 39.8, 40.4, 65.8, 127.9, 129.2, 133.3, 140.6; HRMS calcd
for C13H16O2S 236.0871, found 236.0874.
trans-3-(Phenylthio)-2-methoxybicyclo[2.2.2]octane. A mixture of
DMSO (15 mL) and powdered potassium hydroxide (1.50 g, 26.7
mmol) was stirred for 5 min at room temperature. trans-3-(Phenylthio)-
2-bicyclo[2.2.2]octanol (1.50 g, 6.4 mmol) (the minor product of
reduction of the ketone, above) was added, followed immediately by
iodomethane (4.54 g, 32 mmol), as above. Workup gave the product
(1.45 g, 90.6% yield) as a colorless oil: 1H NMR δ 1.03-1.40 (m, 4
H), 1.50-1.70 (m, 4 H), 1.80-1.85 (m, 2 H), 2.97 (m, 2 H), 3.06 (s,
3 H), 7.15-7.26 (m, 3 H), 7.35-7.45 (m, 2 H); 13C NMR δ 19.3,
21.0, 24.1, 27.2, 29.4, 30.5, 54.6, 57.7, 85.9, 127.4, 129.9, 131.8, 137.0.
trans-3-(Phenylsulfonyl)-2-methoxybicyclo[2.2.2]octane (1). Hy-
drogen peroxide (30%, 12 mL) was added dropwise to a solution of
trans-3-(phenylthio)-2-methoxybicyclo[2.2.2]octane (1.20 g, 4.8 mmol)
in acetic acid (5.0 mL), as above. Workup gave the product (1) (1.15
g, 85.4% yield) as a white solid which was recrystallized from 60%
methanol to give white crystals: mp 84-86 °C; IR (KBr) Vmax 3061
(w), 2948 (s), 2863 (s), 2824 (m), 1447 (s), 1337 (m), 1306 (s), 1289
(s), 1246 (w), 1148 (s), 1107 (m), 1086 (s), 758 (m), 723 (s) cm-1; 1H
NMR δ 1.37-1.69 (m, 7 H), 1.97 (m, 1 H), 2.23 (m, 2 H), 3.03 (s, 3
H), 3.05 (m, 1 H), 3.77 (m, 1 H), 7.60 (m, 3 H), 7.91 (m, 2 H); 13C
NMR δ 17.9, 20.6, 22.6, 25.6, 26.9, 27.3, 56.3, 70.7, 77.8,128.4, 129.0,
133.4, 139.2; HRMS calcd for C15H21O3S (M + 1) 281.1211, found
218.1212.
2-(Phenylthio)bicyclo[2.2.2]octane. Thiophenol (1.22 g, 11.1 mmol)
was added in small portions to bicyclo[2.2.2]oct-2-ene (1.20 g, 11.1
mmol) as in the preparation of 2-(phenylthio)bicyclo[2.2.1]heptane,
above. The product was vacuum distilled at 185-190 °C (oil bath, 8
Torr) to give the product (2.07 g, 85.6% yield) as a colorless oil: IR
(neat) Vmax 3073 (w), 2934 (s), 2863 (s), 1586 (m), 1479 (s), 1439 (s),
1267 (w), 1092 (m), 1026 (m), 739 (s), 691 (s) cm-1; 1H NMR δ 1.62
(m, 10 H), 2.10 (m, 2 H), 3.50 (m, 1 H), 7.26 (m, 5 H); 13C NMR δ
20.5, 24.7, 24.9, 25.5, 26.4, 28.3, 34.4, 45.3, 126.0, 128.8, 130.3, 136.7;
HRMS calcd for C14H18S 218.1129, found 218.1125.
2-(Phenylsulfonyl)bicyclo[2.2.2]octane (18). Hydrogen peroxide
(30%, 16 mL) was added dropwise with stirring to a solution of
2-(phenylthio)bicyclo[2.2.2]octane (2.00 g, 9.17 mmol) in glacial acetic
acid (6 mL) at room temperature, as above. Workup gave the product
(2.15 g, 93.7% yield) as a colorless oil which solidified on standing.
This was recrystallized from 60% methanol: mp 61-63 °C (lit.65 mp
58-59 °C for 2-(phenylsulfonyl)bicyclo[2.2.2]octane-5,6-d2); IR (KBr)
Vmax 3073 (w), 2942 (s), 2867 (s), 1451 (m), 1306 (s), 1275 (s), 1148
(s), 1086 (m), 767 (m), 723 (s), 702 (m) cm-1; 1H NMR δ 1.39-1.78
(m, 9 H), 2.10 (m, 2 H), 2.27 (m, 1 H), 3.16 (m, 1 H), 7.59 (m, 3 H),
7.88 (m, 2 H); 13C NMR δ 20.9, 24.1, 24.3, 24.6, 24.9, 26.8, 27.0,
62.5, 128.4, 129.1, 133.3, 139.1.
1-Methoxy-2-(phenylsulfonyl)bicyclo[2.2.2]octane (7). Hydrogena-
tion of 1-methoxy-6-endo-(phenylsulfonyl)bicyclo[2.2.2]oct-2-ene66
(0.20 g, 0.72 mmol) in chloroform and Pt/C (10%, 0.015 g) under 1
atm of H2 gave a quantitative yield of 7: mp 91.5-94.0 °C; 1H NMR
δ 1.3-2.5 (m, 11 H), 2.63 (s, 3 H), 3.47 (m, 1 H), 7.3-8.0 (m, 5 H);
13C NMR δ 24.2, 25.1, 26.0, 26.3, 28.5, 30.0, 48.4, 63.6, 75.0, 128.4,
128.8, 132.8, 141.6; calcd exact mass for C15H20O3S 280.1133, found
280.1137.
trans-3-(Phenylthio)-2-bicyclo[2.2.2]octanol. Thiophenol (3.42 g,
31.0 mmol) and 2,3-epoxybicyclo[2.2.2]octane63 (2.80 g, 25.5 mmol)
were added to a solution of potassium hydroxide (2.5 g, 44.6 mmol)
in 50% aqueous ethanol (25 mL), and the solution was refluxed for
120 h with stirring. Workup gave the product (4.736 g, 20.2 mmol,
89.6% yield) as a slightly yellowish oil, which was distilled from a
coldfinger apparatus at 235-240 °C (oil bath, 8 Torr) to give a colorless
oil: IR (neat) Vmax 3393 (s), 3057 (w), 2940 (s), 2864 (s), 1583 (m),
1479 (s), 1454 (m), 1061 (m), 1026 (s), 738 (s), 691 (s) cm-1; 1H NMR
δ 1.38 (m, 2 H), 1.69 (m, 6 H), 1.84 (m, 2 H), 2.28 (s, 1 H), 3.24 (m,
1 H), 3.73 (m, 1 H), 7.27 (m, 3 H), 7.40 (m, 2 H); 13C NMR δ 18.1,
19.7, 23.4, 26.1, 29.9, 32.6, 56.3, 76.0, 126.5, 128.9, 130.9, 135.7;
HRMS calcd for C14H18OS 234.1078, found 234.1077.
3-(Phenylthio)-2-bicyclo[2.2.2]octanone. Swern oxidation (DMSO,
oxalyl chloride, Et3N, as above) of trans-3-(phenylthio)-2-bicyclo[2.2.2]-
octanol (4.00 g, 17.2 mmol) gave the product (3.950 g, 99.5% yield)
as a yellowish oil: IR (neat) Vmax 3060 (w), 2946 (s), 2870 (s), 1742
(s), 1581 (m), 1480 (s), 1455 (s), 1325 (w), 1111 (m), 1062 (s), 1026
(m), 741 (s), 691 (s), 666 (s) cm-1; 1H NMR δ 1.58 (m, 3 H), 1.84 (m,
4 H), 2.15 (m, 2 H), 2.38 (m, 1 H), 3.74 (t, 1 H), 7.27 (m, 3 H), 7.47
(m, 2 H); 13C NMR δ 20.0, 22.7, 23.5, 25.3, 33.1, 42.5, 57.6, 127.2,
129.0, 131.8, 134.4, 213.5; HRMS calcd for C14H16OS 232.0922, found
232.0923.
cis-3-(Phenylthio)-2-bicyclo[2.2.2]octanol. 3-(Phenylthio)bicyclo-
[2.2.2]octan-2-one (4.50 g, 19.4 mmol) in 2-propanol (15 mL) was
added dropwise to a solution of sodium borohydride (0.733 g, 19.4
mmol) in 2-propanol (95 mL). The mixture was stirred overnight at
room temperature. Concentrated hydrochloric acid (6 mL) was added,
and the 2-propanol was removed on a rotary evaporator. Workup gave
a mixture of the cis and trans alcohols as an oil (4.31 g, 18.4 mmol,
1
94.8% yield). The H NMR spectrum showed the mixture to contain
the cis isomer (78%) and the trans isomer (22%), which were separated
by column chromatography on silica gel using a mixture of diethyl
ether (20%) and petroleum ether (80%) as the eluent, with the cis isomer
eluting first from the column; this was distilled in a coldfinger apparatus
at 225-228 °C (oil bath, 8 Torr): IR (neat) Vmax 3440 (s), 3058 (w),
2936 (s), 2865 (s), 1584 (m), 1480 (s), 1456 (m), 1439 (s), 1067 (s),
1
1044 (s), 1026 (m), 738 (s), 691 (s) cm-1; H NMR δ 1.36 (m, 2 H),
1.60 (m, 4 H), 1.88 (m, 4 H), 3.07 (s, 1 H), 3.62 (dt, 1 H), 4.01 (dt, 1
H), 7.27 (m, 3 H), 7.37 (m, 3 H); 13C NMR δ 18.1, 20.7, 23.0, 25.9,
31.3, 31.4, 55.0, 68.2, 126.4, 129.0, 129.9, 136.3; HRMS calcd for
C14H18OS 234.1078, found 234.1074.
cis-3-(Phenylthio)-2-methoxybicyclo[2.2.2]octane. cis-3-(Phenylthio)-
2-bicyclo[2.2.2]octanol (4.10 g, 17.5 mmol) was methylated with MeI
in KOH/DMSO as above to give a yellow oil (4.169 g, 95.9% yield).
The 1H NMR spectrum showed that the product contained the cis isomer
(86%) and the trans isomer (14%). Chromatography on silica gel gave
the pure cis product as a colorless liquid: IR (neat) Vmax 3057 (w),
2938 (s), 2865 (s), 1584 (m), 1482 (m), 1439 (m), 1102 (s), 739 (s),
2-exo-(Phenylsulfonyl)-7-oxabicyclo[2.2.1]heptane (12). Hydro-
genation of 2-exo-(phenylsulfonyl)-7-oxabicyclo[2.2.1]hept-5-ene67 (mp
63.5-65 °C, reported67 to be a liquid) in CHCl3 and Pt/C (10%) under
1
691 (s); H NMR δ 1.27-1.42 (m, 4 H), 1.51-1.68 (m, 4 H), 1.85-
1.94 (m, 2 H), 3.37 (s, 3 H), 3.65 (s, 2 H), 7.20-7.28 (m, 3 H), 7.33-
7.40 (m, 2 H); 13C NMR δ 18.7, 20.3, 22.6, 25.6, 27.8, 29.0, 51.5,
58.1, 79.2, 125.9, 128.8, 130.6, 136.5.
1
H2 (1 atm) gave a quantitative yield of 12: mp 100-101.5 °C; H
NMR δ 1.0-2.3 (m, 6 H), 3.29 (m, 1 H), 4.63 (t, J ) 4.8 Hz, 1 H),
4.94 (d, J ) 5.0 Hz, 1 H), 7.45-8.00; 13C NMR (CDCl3) δ 29.0, 29.9,
33.7, 67.8, 76.4, 76.6, 128.8, 129.2, 133.7, 138.0; calcd exact mass for
C12H14O3S 238.0666, found 238.0664.
cis-3-(Phenylsulfonyl)-2-methoxybicyclo[2.2.2]octane (11). Hy-
drogen peroxide (30%, 20 mL) was added to a solution of cis-3-
(phenylthio)-2-methoxybicyclo[2.2.2]octane (2.0 g, 8.0 mmol) in acetic
acid (8.5 mL) as above. Workup gave 11 (2.15 g, 96.3% yield) as a
colorless oil which solidified on standing; recrystallization from ethyl
acetate (10%) and petroleum ether (90%) gave 11: mp 127-128 °C;
IR (KBr) Vmax 3065 (w), 2984 (s), 2818 (s), 2924 (s), 2864 (s), 1448
(s), 1316 (s), 1302 (s), 1269 (s), 1284 (s), 1234 (m), 1144 (s), 1103
(s), 1086 (s), 995 (w), 762 (s), 721 (s), 692 (s) cm-1; 1H NMR δ 3.19
(s, 3 H), 3.39 (dt, 1 H), 3.53 (dd, 1 H), 7.57 (m, 3 H), 7.93 (m, 2 H);
13C NMR δ 18.4, 20.9, 21.6, 25.4, 26.6, 27.0, 57.4, 66.4, 76.9, 128.6,
128.8, 133.0, 141.0; HRMS calcd for C15H21O3S (M + 1) 281.1211,
found 281.1207.
2-endo-(Phenylsulfonyl)-7-oxabicyclo[2.2.1]heptane (8). Separation
by TLC (CH2Cl2) of the mixture of 2-endo- and 2-exo-(phenylsulfonyl)-
7-oxabicyclo[2.2.1]hept-5-enes67 gave 2-endo-(phenylsulfonyl)-7-
1
oxabicyclo[2.2.1]hept-5-ene: mp 52-53 °C; H NMR δ 1.6-2.3 (m,
(65) Williams, R. V.; Kelley, G. W.; Loebel, J.; van der Helm, D.;
Bulman Page, P. C. J. Org. Chem. 1990, 55, 3840-3846.
(66) Carr, R. V. C.; Williams, R. V.; Paquette, L. A. J. Org. Chem. 1983,
48, 4976-4986.
(67) De Lucchi, O.; Lucchini, V.; Pasquato, L.; Modena, G. J. Org. Chem.
1984, 49, 596-604.

10324 J. Am. Chem. Soc., Vol. 122, No. 42, 2000
King et al.
2 H), 3.70 (m, 1 H), 5.04 (m, 2 H), 6.50 (m, 2 H), 7.4-8.1 (m, 5 H);
13C NMR δ 28.7, 62.6, 78.2, 79.6, 127.6, 129.4, 131.5, 133.7, 137.2,
140.3; calcd exact mass for C12H13O3S (M + 1) 237.0585, found
237.0588. Hydrogenation of the alkene in CHCl3 and Pt/C (10%) under
H), 3.16 (t, J ) 5.2 Hz, 2 H), 3.34 (s, 3 H), 3.77 (t, 2 H); 13C NMR δ
43.0, 55.1, 58.9, 66.1.
4-(Phenylsulfonyl)tetrahydropyran (32). Tetrahydro-4H-pyran-4-
ol (Aldrich) was converted into 32 by the method of Eliel and Ro for
preparing 27:54 mp 87-89 °C; 1H NMR δ 1.6-1.9 (m, 4 H), 3.11 (tt,
J ) 11.8 and 4.1 Hz, 1 H), 3.30 (ddd, J ) 11.8, 11.8, and 2.6 Hz, 2
1
H2 (1 atm) gave a quantitative yield of 8: mp 75.7-77 °C; H NMR
δ 1.6-2.7 (m, 6 H), 3.5-3.63 (m, 1 H), 4.60-4.70 (m, 2 H), 7.5-7.9
(m, 5 H); 13C NMR δ 26.0, 29.9, 32.9, 65.8, 77.2, 78.3, 127.6, 129.4,
133.7, 140.5; calcd exact mass for C12H14O3S 238.0664, found
238.0668.
H), 4.02 (ddd, J ) 11.8, 4.8, and 1.2 Hz, 2 H), 7.5-7.9 (m, 5 H); 13
C
NMR δ 25.6, 60.6, 66.5, 129.1, 129.2, 133.9; calcd exact mass for
C11H14O3S 226.0664, found 226.0666.
Determination of the Rates of H-D Exchange and Epimerization
Equilibria Compositions. Solutions of NaOD in D2O (0.10-0.50 M)
were prepared by dissolving Na metal in D2O under N2. The solutions
were titrated using 0.01 M HCl. Buffer solutions were prepared from
(a) solid Na2CO3 in D2O adjusted to the appropriate pD (pH meter
reading + 0.37) with 37% DCl in D2O, or (b) Na3PO4‚12H2O previously
dried overnight in an oven at 200 °C, cooled in a desiccator, and
dissolved in D2O, and the pD was adjusted using 37% DCl. A typical
run was carried out by dissolving a 9 mg sample of the substrate in
CD3CN (0.35 mL) in an NMR tube. NaOH solution (0.35 mL) was
added using a 0.5-mL syringe, and the NMR tube was flame-sealed
immediately. The concentration of the starting material was measured
by determining the integral of the changing R-hydrogen signal relative
to that of a convenient unchanging signal; the time for each measure-
ment was taken as the average of the start and stop of data collection.
Faster reactions were followed by taking a series of NMR spectra at
the temperature of the NMR probe (21 ( 2 °C); with slower reactions
the sample tube was placed in a water or oil bath maintained at the
cis-3-Methoxy-1-(phenylsulfonyl)cyclohexane (28) and trans-3-
Methoxy-1-(phenylsulfonyl)cyclohexane (29) (Chart 1). A mixture
of 3-(phenylthio)cyclohexanone68,69 (6.9 g, 0.0334 mol) and sodium
borohydride (4.88 g, 12.9 mmol) in methanol (200 mL) was stirred for
2 h at 0 °C, and then hydrolyzed with water and acidified with sulfuric
acid (10%). Workup gave 3-(phenylthio)cyclohexanol (6.0 g, 86%, cis-
trans mixture). This alcohol (4.84 g, 23.2 mmol) in DMSO (10 mL)
was added to KOH (5.2 g, 93 mmol) in DMSO (40 mL) at room
temperature, and the resulting mixture was stirred for 2 min. Io-
domethane (13.2 g, 0.093 mol) was added and stirred 30 min. Workup
gave 3-methoxy-1-(phenylthio)cyclohexane (4.8 g, 93%, cis-trans
mixture). The sulfide was oxidized by H2O2 (30%) in acetic acid, giving
a quantitative yield of 3-methoxy-1-(phenylsulfonyl)cyclohexane. Thick
layer chromatography (ether-hexane, 3:2) gave 28 and 29 separately.
Compound 28: mp 95.5-97.5 °C; 1H NMR δ 1.0-2.5 (m, 8 H), 2.90
(tt, J ) 12.4 and 3.4 Hz, 1 H), 3.07 (tt, J ) 11.0 and 4.1 Hz, 1 H),
3.30 (s, 3 H), 7.5-7.9 (m, 5 H); 13C NMR (CDCl3) δ 22.6, 25.1, 31.0
(double intensity), 56.0, 62.0, 78.0, 129.1 (double intensity), 133.7,
136.8; calcd exact mass for C13H19O3S (M + 1) 255.1055, found
255.1056. Compound 29: mp 63.5-64.5 °C; 1H NMR δ 1.0-2.4 (m,
8 H), 3.17 (s, 3 H), 3.22 (tt, J ) 12.2 and 3.5 Hz, 1 H), 3.58 (m, 1 H),
7.40-8.0 (m, 5 H); 13C NMR δ 18.8, 24.8, 28.1, 28.4, 55.7, 58.4, 73.8,
128.8, 128.9, 133.5, 137.2; calcd exact mass for C13H19O3S (M + 1)
255.1055, found 255.1053.
1
specified temperatures ((0.5 °C or better), with H NMR spectra run
at appropriate intervals. The value of kobs was obtained from the slope
of the plot of ln(% concentration of starting material at time t) vs t (in
seconds); normally a plot of kobs vs [OD-] for reactions at two to four
different NaOD concentrations gave kexch. Standard deviations in kobs
were normally much less than (5%, but in some instances, either
because of overlapping NMR signals, or with very slow reactions (kobs
< 2 × 10-6 s-1 at 25 °C) apparently accompanied by slow decomposi-
tion processes, the errors may approach (10%; for these and some
compounds determined at only one concentration of OD- (4, 13, 14,
15, and 16 and their models) the kexch values are expressed by two
significant figures. The log kN points are drawn in Figure 1 with heights
of 0.1 log unit corresponding to errors in kN of (10%. Equilibrium
constants for 1 h 11, 2 h 10, and 3 h 9 were determined by heating
a sample of 1, 2, or 3 (9 mg) in NaOD (0.30 M) in methanol-d4-D2O
(1:1) (0.7 mL) for 15 days at 77 °C in a sealed NMR tube. The solutions
were neutralized with concentrated HCl, and the mixture was analyzed
directly by HPLC; the equilibrium ratios were, respectively, 1 h 11,
97.1:2.9; 2 h 10, 99.7:0.3; and 3 h 9, 97.6:2.4. Samples of 19 and 20
(10 mg of each in 0.70 mL of 0.30 M NaOD in dioxane-d8-D2O, 1:1)
were heated in a sealed tube for 70 days at 77 °C (H-D exchange was
Phenyl 2-methoxyethyl sulfone (30b):11c,70 1H NMR, see ref 11c;
13C NMR δ 55.7, 58.3, 65.3, 127.6, 128.8, 133.5, 139.3.
Phenyl 3-Methoxypropyl Sulfone (30c). Phenyl 3-hydroxypropyl
sulfide71 (5.0 g, 30 mmol) was added to KOH (6.7 g, 120 mmol) in
DMSO (50 mL) and stirred for 2 min at room temperature. CH3I (16.7
g, 118 mmol) was added, and the resulting solution was stirred for 15
min. Workup gave a quantitative yield of phenyl 3-methoxypropyl
sulfide, which was then oxidized by H2O2 (30%) in acetic acid to give
30c (100%) as a colorless oil: 1H NMR δ 1.70-2.00 (m, 2 H), 3.05-
3.15 (m, 2 H), 3.16 (s, 3 H), 3.32 (t, J ) 6.0 Hz, 2 H), 7.30-8.00 (m,
5 H); 13C NMR δ 22.9, 53.2, 58.3, 69.8, 127.7, 129.1, 133.5, 138.9;
calcd exact mass for C10H14O3S: 215.0742, found 215.0748.
Phenyl 4-Methoxybutyl Sulfone (30d). Phenyl 4-methoxybutyl
sulfide71 was oxidized with H2O2 (30%) in acetic acid and gave a
quantitative yield of 30d as a colorless oil: 1H NMR δ 1.50-1.85 (m,
4 H), 3.05-3.15 (m, 2 H), 3.23 (s, 3 H), 3.30 (t, J ) 6.0 Hz, 2 H),
7.45-7.95 (m, 5 H); 13C NMR δ 19.9, 28.1, 56.0, 58.5, 71.7, 128.0,
129.2, 133.6, 139.0; calcd exact mass for C11H16O3S 229.0899, found
229.0898.
1
complete in 1 day). The H NMR spectrum of the product from 19
showed it to consist of the R-deuterated isotopomers of 19 and 20 in
the ratio 57.2:42.8. In the product from 20 the ratio was 93.2:6.8; the
equilibrium mixture ratio was estimated to be 86:14. Equilibrium
constants for 28 h 27 and 12 h 8 were determined on solutions in
dioxane-d8-D2O at 64 °C for at least 45 days until the 1H NMR spectra
from each direction indicated that equilibrium had been reached; careful
integration gave the respective equilibrium constants, 0.019 and 0.040
(est. ( 0.01). Further details are available elsewhere.44
Phenyl 5-Methoxypentyl Sulfone (30e). Phenyl 5-methoxypentyl
sulfide71 was oxidized with H2O2 (30%) in acetic acid to give a
quantitative yield of 30e as a colorless oil: 1H NMR δ 1.30-1.75 (m,
6 H), 2.98-3.1 0 (m, 2 H), 3.24 (s, 3 H), 3.28 (t, J ) 6.0 Hz, 2 H),
7.45-7.95 (m, 5 H); 13C NMR δ 22.4, 24.9, 28.9, 56.1, 58.5, 72.0,
127.9, 129.2, 133.6, 139.0; calcd exact mass for C12H19O3S (M + 1)
243.1055, found 243.1056.
Acknowledgment. Much of the first draft of this paper was
prepared during a sabbatical leave at Cambridge University by
J.F.K., who is indebted to Professor A. J. Kirby, F.R.S., for his
kind hospitality and critical comments. We thank NSERC of
Canada for financial support of this work.
Methyl 2-Methoxyethyl Sulfone. Oxidation of MeOCH2CH2SMe72
with H2O2-HOAc as above gave the sulfone: 1H NMR δ 2.93 (s, 3
(68) Bakuzis, P.; Bakuzis, M. L. F. J. Org. Chem. 1981, 46, 235-239.
(69) Chamberlain, P.; Whitham, G. H. J. Chem. Soc., Perkin Trans. 2
1972, 130-135.
(70) Pillot, J.-P.; Dunogues, J.; Calas, R. Synthesis 1977, 469-472.
(71) Cere`, V.; Pollicino, S.; Fava, A. Tetrahedron 1996, 52, 5989-5998.
JA001320L
(72) Doering, W. v. E.; Schreiber, K. C. J. Am. Chem. Soc. 1955, 77,
514-520.