Thermal Stereomutations and Stereochemically Elucidated [1,3]-Carbon Sigmatropic Shifts of 1-(E)-Propenyl-2-methylcyclobutanes Giving 3,4-Dimethylcyclohexenes

Article abstract of DOI:10.1021/ja030516t

The thermal stereomutations and [1,3] carbon sigmatropic shifts shown by (+)-(1S,2S)-trans-1-(E)-propenyl-2-methylcyclobutane and by (-)-(1S,2R)-cis-1-(E)-propenyl-2-methylcyclobutane in the gas phase at 275 °C leading to 3,4-dimethylcyclohexenes have been followed. The reaction-time-dependent data for concentrations and enantiomeric excess values for substrates and [1,3] shift products have been deconvoluted to afford rate constants for the discrete isomerization processes. Both trans and cis substrates react through four stereochemically distinct [1,3] carbon shift paths. For one enantiomer of the trans reactant the relative rate constants are k_si = 58%, k_ar = 5%, k_sr = 33%, and k _al = 4%. For a single enantiomer of the cis reactant, k′ _si = 18%, k′_ar = 11%, k′_sr = 51%, and k′_ai = 20%. A trans starting material reacts through orbital symmetry allowed suprafacial, inversion and antarafacial, retention paths to give trans-3,4-dimethylcyclohexenes 63% of the time. A cis isomer reacts to give the more stable trans-3,4-dimethylcyclohexenes through orbital symmetry-forbidden suprafacial, retention and antarafacial, inversion paths 71% of the time. The [1,3] carbon sigmatropic shifts are not controlled by orbital symmetry constraints. They seem more plausible rationalized as proceeding through diradical intermediates having some conformational flexibility after formation and before encountering an exit channel. The distribution of stereochemical outcomes may well be conditioned by dynamic effects. The thermal stereomutations of the 1-(E)-propenyl-2-methylcyclobutanes take place primarily through one-center epimerizations. For the trans substrate, the relative importance of the three distinction rate constants are k₂ = 48%, k₁ = 34%, and k₁₂ = 18%. For the cis isomer, k′ ₂ = 44%, k′₁ = 32%, and k′₁₂ = 24%. These patterns are reminiscent of ones determined for stereomutations in 1,2-disubstitued cyclopropanes.

Full text of DOI:10.1021/ja030516t

Published on Web 11/22/2003
Thermal Stereomutations and Stereochemically Elucidated
[1,3]-Carbon Sigmatropic Shifts of
1-(E)-Propenyl-2-methylcyclobutanes Giving
3,4-Dimethylcyclohexenes
John E. Baldwin* and Richard C. Burrell
Contribution from the Department of Chemistry, Syracuse UniVersity, Syracuse, New York 13244
Received August 29, 2003; E-mail: jbaldwin@syr.edu
Abstract: The thermal stereomutations and [1,3] carbon sigmatropic shifts shown by (+)-(1S,2S)-trans-
1-(E)-propenyl-2-methylcyclobutane and by (-)-(1S,2R)-cis-1-(E)-propenyl-2-methylcyclobutane in the gas
phase at 275 °C leading to 3,4-dimethylcyclohexenes have been followed. The reaction-time-dependent
data for concentrations and enantiomeric excess values for substrates and [1,3] shift products have been
deconvoluted to afford rate constants for the discrete isomerization processes. Both trans and cis substrates
react through four stereochemically distinct [1,3] carbon shift paths. For one enantiomer of the trans reactant
the relative rate constants are ksi ) 58%, kar ) 5%, ksr ) 33%, and kai ) 4%. For a single enantiomer of
the cis reactant, k′si ) 18%, k′ar ) 11%, k′sr ) 51%, and k′ai ) 20%. A trans starting material reacts through
orbital symmetry allowed suprafacial,inversion and antarafacial,retention paths to give trans-3,4-
dimethylcyclohexenes 63% of the time. A cis isomer reacts to give the more stable trans-3,4-
dimethylcyclohexenes through orbital symmetry-forbidden suprafacial,retention and antarafacial,inversion
paths 71% of the time. The [1,3] carbon sigmatropic shifts are not controlled by orbital symmetry constraints.
They seem more plausible rationalized as proceeding through diradical intermediates having some
conformational flexibility after formation and before encountering an exit channel. The distribution of
stereochemical outcomes may well be conditioned by dynamic effects. The thermal stereomutations of the
1-(E)-propenyl-2-methylcyclobutanes take place primarily through one-center epimerizations. For the trans
substrate, the relative importance of the three distinction rate constants are k2 ) 48%, k1 ) 34%, and k12
) 18%. For the cis isomer, k′2 ) 44%, k′1 ) 32%, and k′12 ) 24%. These patterns are reminiscent of ones
determined for stereomutations in 1,2-disubstitued cyclopropanes.
Scheme 1
Introduction
The first study to define the full stereochemistry of a thermal
vinylcyclopropane-to-cyclopentene rearrangement followed the
reactions of a nonracemic sample of trans-1-(E)-propenyl-2-
methylcyclopropane.1 The same utilization of methyl substitu-
ents as stereochemical markers was employed in this our first
definition of the full stereochemistry for a vinylcyclobutane-
to-cyclohexene rearrangement using (+)-(1S,2S)-trans-1-(E)-
propenyl-2-methylcyclobutane (SS-1) and (-)-(1S,2R)-cis-1-(E)-
propenyl-2-methylcyclobutane (SR-2) as reactants.
and infer the kinetically controlled stereochemistry of the
reaction. Fragmentations to give olefins, pentadienes, and
hexadienes as well as isomerizations to dimethylcyclohexenes
and octadienes lead to complex mixtures of products. Scheme
1, in which 1 and 2 represent diastereomers without enantiomeric
distinctions, introduces the kinetic situation in an abbreviated
and quite simplified form.
The first-order fragmentations and isomerizations included
in Scheme 1 consume vinylcyclobutane reactants. They may
be accommodated experimentally and within a kinetic model,
but all of them that compete with [1,3] shifts diminish the
amount of dimethylcyclohexene products 3 and 4 available for
stereochemical scrutiny (Scheme 2).
The experimental challenges posed by this or any program
seeking to define the stereochemical characteristics of [1,3]
carbon shifts leading from vinylcyclobutanes to cyclohexenes
are complicated by other kinetically competitive reactions. One
cannot simply heat a stereochemically well-defined substituted
vinylcyclobutane, isolate the mixture of cyclohexene products,
The thermal stereomutations of substituted vinylcyclobutane
reactants present more of a problem (Scheme 3). Here subscripts
(1) Andrews, G. D.; Baldwin, J. E. J. Am. Chem. Soc. 1976, 98, 6705-6706.
9
10.1021/ja030516t CCC: $25.00 © 2003 American Chemical Society
J. AM. CHEM. SOC. 2003, 125, 15869-15877
15869

A R T I C L E S
Baldwin and Burrell
Scheme 2
of better than 99% ee by capillary GC analyses of synthetic
precursors, were prepared from 2-methylcyclobutanecarbox-
aldehydes through a Wittig-like coupling reaction employing
CH3CHI2 and CrCl2.3,4 The olefinic products were obtained in
high yield and with ∼14:1 E:Z stereoselectivity. The E isomers
were then purified by preparative GC on a â,â′-oxydipropioni-
trile (ODPN) column.
In trial runs samples of racemic 1, the internal standard
cyclooctane, and pentane to serve as a bath gas were heated for
given times in a 300-mL quartz bulb encased in a well-insulated
thermostated aluminum block. At the end of a thermal reaction
the samples were transferred to a vacuum line, cooled and
condensed, and analyzed by capillary GC.
Scheme 3
Some trial runs gave complex product mixtures in complete
accord with expectations, while others led to mixtures containing
additional, unexpected products and irreproducible isomer
distributions. This random oscillation between well-behaved
thermal reactions and totally useless kinetic runs was eventually
traced to a minor impurity. The capillary GC analyses run to
measure the relative proportions at t ) 0 of 1 and cyclooctane
(retention times 8.0 and 21.8 min) showed a small peak (0.05%)
at 24.1 min in some of the solutions, the ones that gave rise in
thermal reactions to the irregular kinetic trends and additional
products. Solutions that did not contain that slight impurity
displayed well-behaved kinetic profiles and the anticipated array
of isomers. The trace component at 24.1 min was determined
to be CH3CHI2, a treacherous contaminant to have in a gas-
phase static reactor at 275 °C, for it could well decompose and
initiate unwanted reactions.
1,1-Diiodoethane “bleeds” slowly from the ODPN column.
A first collection of 1 after long GC column conditioning gave
1 free of any 1,1-diiodoethane, while samples of 1 collected
through repeated preparative GC runs, one after the other, were
often contaminated with small amounts of this impurity. Samples
of 1 subjected to a second preparative GC purification using
an SE-30 column to remove all traces of 1,1-diiodoethane
exhibited good kinetic behavior and the anticipated products
reliably. Hence, SS-1 and SR-2 samples for kinetic work were
purified through preparative GC separations first on a ODPN
column and then on an SE-30 column.
Concentration Versus Time Kinetic Profiles. Kinetic work
on thermal rearrangements of hydrocarbons usually depends on
capillary gas chromatographic analyses. Relative concentration
versus time profiles and individual rate constants for various
reactions of trans- and cis-1-(E)-propenyl-2-methylcyclobutanes
SS-1 and SR-2 dependent upon capillary GC columns having
achiral liquid phases have been published.2e The eight rate
constants defined in Schemes 1 and 2 and summarized in Table
1 were measured independently by following thermal reactions
of racemic samples of 1 and 2.2e These rate constants were
reproduced to within an average deviation of 2.0% (from 0.7
to 3.1%).
label rate constants to distinguish one-center (k1, k2) and
simultaneous two-center (k12) epimerizations shown by one trans
isomer, SS-1 or RR-1. The epimerizations of a cis isomer have
rate constants k′1, k′2, and k′12. One must follow the time-
dependent evolution of all four possible stereoisomers of a
stereochemically well-defined reactant and then take account
of each reacting to afford all four possible 3,4-dimethylcyclo-
hexene products. Appropriate experimental data must be secured,
and the deconvolutions of rate data to afford specific rate
constants must be managed carefully.
Other challenges associated with the required syntheses of
substrates and assignments of absolute stereochemistry for all
substituted vinylcyclobutanes and cyclohexenes involved are of
a different sort and can be dealt with through well-known
methods.
Parts of our extended effort to unravel the thermal reactions
of the 1-(E)-propenyl-2-methylcyclobutanes have been reported.2
This culminating report provides a full account of kinetic
experiments with nonracemic trans- and cis-1-(E)-propenyl-2-
methylcyclobutanes SS-1 and SR-2 as starting materials. The
steps and simple equations required for rigorous deconvolutions
of raw rate data to obtain the eight independent rate constants
leading from 1-(E)-propenyl-2-methylcyclobutanes to the 3,4-
dimethylcyclohexenes SR-3, RS-3, SS-4, and RR-4 are detailed.
The thermal stereomutation rate constants defined in Scheme 3
are also obtained and reported.
Results and Discussion
The kinetic situation outlined in Scheme 1 dictates the form
of 1(t) and 2(t) equations. Each is equal to B1 exp(-λ1t) + B2
exp(-λ2t); the constants B1 and B2 depend on the starting
material, with an initial relative concentration equal to (B1 +
B2). The parameters λ1 and λ2 (λ1 ) 1.84 × 10-5 s-1 and λ2 )
Purified Reactants and Reproducible Kinetic Behavior.
The 1-(E)-propenyl-2-methylcyclobutanes SS-1 and SR-2, each
(2) (a) Alexander, J. S.; Baldwin, J. E.; Burrell, R. C.; Ruhlandt-Senge, K.
Chem. Commun. 2000, 2201-2202. (b) Baldwin, J. E.; Burrell, R. C. J.
Org. Chem. 2000, 65, 7139-7144. (c) Baldwin, J. E.; Burrell, R. C. J.
Org. Chem. 2000, 65, 7145-7150. (d) Baldwin, J. E.; Burrell, R. C. J.
Am. Chem. Soc. 2001, 123, 6718-6719. (e) Baldwin, J. E.; Burrell, R. C.
J. Org. Chem. 2002, 67, 3249-3256. (f) Baldwin, J. E.; Burrell, R. C. J.
Phys. Chem. A 2003, 107, 10069-10073.
(3) Okazoe, T.; Takai, K.; Utimoto, K. J. Am. Chem. Soc. 1987, 109, 951-
953.
(4) Getty, S. J.; Berson, J. A. J. Am. Chem. Soc. 1991, 113, 4607-4621.
9
15870 J. AM. CHEM. SOC. VOL. 125, NO. 51, 2003

1-(E)-Propenyl-2-methylcyclobutanes
A R T I C L E S
Table 1. Rate Constants in Schemes 1 and 2, for Reactions of
SS-1 and SR-2 at 275 °C
rate constant
value2e
k
k′
1.70 × 10-5 s-1
6.81 × 10-5 s-1
6.55 × 10-6 s-1
2.24 × 10-5 s-1
2.44 × 10-6 s-1
1.45 × 10-6 s-1
5.34 × 10-6 s-1
2.19 × 10-6 s-1
k(1f2) ) k1 + k2
k(2f1) ) k′1 + k′2
k(1f3) ) ksi + kar
k(1f4) ) ksr + kai
k(2f3) ) k′sr + k′ai
k(2f4) ) k′si + k′ar
Figure 2. Mole percent concentrations of [SS-1 + RR-1] (9) and [SR-2
+ RS-2] (b) for reactions at 275 °C starting SR-2, and least-squares-fit
functions of the theoretically required form given in eqs 3 and 4.
Table 2. Mole Percent Concentrations Calculated for 1, 2, 3, and
4 from Thermal Reactions of SS-1 at 275 °C
time (s)
1
2
3
4
0
7200
99.8
83.7
71.3
53.3
40.5
31.0
23.7
18.2
0.2
2.4
3.3
3.4
2.8
2.2
1.7
1.3
0.0
1.7
3.1
5.6
7.4
8.9
10.0
10.8
0.0
1.0
1.8
3.2
4.3
5.1
5.7
6.2
14400
28800
43200
57600
72000
86400
Figure 1. Mole percent concentrations of [SS-1 + RR-1] (9) and [SR-2
+ RS-2] (b) for reactions at 275 °C starting from SS-1, and least-squares-
fit functions of the theoretically required form given in eqs 1 and 2.
9.02 × 10-5 s-1) are the same in all four equations. For reactions
starting with SS-1 eqs 1 and 2 apply.2e
Table 3. Mole Percent Concentrations Calculated for 1, 2, 3, and
4 from Thermal Reaction of SR-2 at 275 °C
time (s)
1
2
3
4
1(t) ) (SS-1 + RR-1)(t) )
0
1800
3600
0.0
3.6
6.6
10.9
13.6
15.2
16.0
16.3
100.0
85.1
72.4
52.4
37.9
27.5
19.9
14.5
0.0
0.9
1.7
2.9
3.9
4.7
5.2
5.7
0.0
0.4
0.7
1.2
1.6
2.0
2.2
2.5
89.2 exp(-λ1t) + 10.6 exp(-λ2t) (1)
7200
2(t) ) (SR-2 + RS-2)(t) )
10800
14400
18000
21600
6.5 exp(-λ1t) - 6.3 exp(-λ2t) (2)
For reactions using SR-2 as substrate, eqs 3 and 4 correlate
experimental points with theory-consistent functions.2e Figures
1 and 2 display GC-derived data and the functions given in eqs
1-4.
GC columns able to achieve similar resolutions of the enantio-
mers of 1 and of 2 proved unsuccessful.
This temporary impasse stimulated a search for some alterna-
tive approach. By combining chemical and chromatographic
methods a satisfactory solution to the analytical challenge was
found. Following a kinetic run the product mixture was analyzed
by GC, and integrated peak intensities gave mole percent
concentration data. The mixture was then separated into four
fractions by preparative GC on a ODPN column. The first
fraction contained both enantiomers of 1; the second contained
trans-1,5-octadiene, cis-3,6-dimethylcyclohexene, and both
enantiomers of 2; the third contained cis-1,5-octadiene, trans-
1,6-octadiene, and the trans-3,4-dimethylcyclohexenes (3); the
fourth contained cis-3,4-dimethylcyclohexenes (4). Analysis of
fraction 3 by capillary GC using a CycloSil B column gave the
enantiomeric excess of trans-3,4-dimethylcyclohexene (3).
Analysis of fraction 4 by GC using a Cyclodex B column
provided the enantiomeric excess value for cis-3,4-dimethyl-
cyclohexene (4).
1(t) ) (SS-1 + RR-1)(t) )
30.7 exp(-λ1t) - 30.7 exp(-λ2t) (3)
2(t) ) (SR-2 + RS-2)(t) )
0.4 exp(-λ1t) + 99.6 exp(-λ2t) (4)
From functions 1-4 and the rate constants for [1,3] shifts
quoted in Table 1 it is easy to calculate mole percent concentra-
tions for 1, 2, 3, and 4 at the given reaction times. These values
are summarized in Tables 2 and 3.
Enantiomeric Excess Values Versus Time Profiles. The
stereochemical ambitions of the present study depended on
gaining information on enantiomeric excess values for 1, 2, 3,
and 4 at various reaction times. For the 3,4-dimethylcyclo-
hexenes sound methods based on capillary GC analyses using
columns with chiral stationary phases were worked out and
validated early on.2c Unfortunately, all attempts to find “chiral”
9
J. AM. CHEM. SOC. VOL. 125, NO. 51, 2003 15871

A R T I C L E S
Baldwin and Burrell
were of better than 99% ee.2b The ee data for RS-2 in Table 4
thus reflects 2 from a composite of sources, with some formed
from nonracemic SS-1 through k1 and k2 paths, some associated
with the original RS-2, and some contingent on k′12. Fortunately,
these complications turned out to have no significant bearing
on the stereochemical concerns at hand.
Scheme 4
The SR-2 used in the kinetic runs summarized in Table 3
was free of the trans isomer, and the trends in ee values of Table
5 presented no seeming anomalies.
Differences in Mole Percent Concentrations of Enanti-
omers Versus Time. The differential equations governing time-
dependent changes of mole percent concentrations of each
isomer of 1-(E)-propenyl-2-methylcyclobutane are given in eqs
5-8.
Table 4. Enantiomeric Excess Data for 1, 2, 3, and 4 in Reaction
Mixtures Derived from SS-1 at 275 °C
time (s)
SS-1
RS-2
SR-3
SS-4
-d[SS-1]/dt ) (k + k1 + k2 + k12 )[SS-1] - k12 [RR-1] -
k′2[SR-2] - k′1[RS-2] (5)
0
7200
99.0
96.6
94.1
89.5
85.1
82.4
79.0
71.1
99.0
12.3
7.3
2.2
4.8
7.4
4.3
-4.2
85.1
80.2
75.6
72.1
70.3
69.5
67.6
80.7
82.5
69.9
70.1
67.5
63.4
61.7
14400
28800
43200
57600
72000
86400
-d[RR-1]/dt ) -k12 [SS-1] +
(k + k1 + k2 + k12 )[RR-1] - k′1[SR-2] - k′2[RS-2] (6)
-d[SR-2]/dt ) -k2 [SS-1] - k1 [RR-1] +
(k′ + k′1 + k′2 + k′12 )[SR-2] - k′12[RS-2] (7)
-d[RS-2]/dt ) -k1 [SS-1] - k2 [RR-1] - k′12[SR-2] +
(k′ + k′1 + k′2 + k′12 )[RS-2] (8)
Table 5. Enantiomeric Excess Data for 1, 2, 3, and 4 in Reaction
Mixtures Derived from SR-2 at 275 °C
time (s)
SS-1
SR-2
SR-3
SS-4
Subtracting expression 6 from 5 and expression 8 from 7 gives
expressions 9 and 10.
0
1800
3600
99.0
96.4
93.6
89.0
85.8
81.6
79.1
71.2
9.6
11.5
14.3
16.1
15.5
15.2
13.4
44.0
42.3
41.2
40.7
38.4
36.4
36.2
24.4
27.6
22.9
24.5
22.9
21.2
18.6
-d[SS-1 - RR-1]/dt )
7200
10800
14400
18000
21600
(k + k1 + k2 + 2k12)[SS-1 - RR-1] +
(k′1 - k′2)[SR-2 - RS-2] (9)
-d[SR-2 - RS-2]/dt ) (k1 - k2 )[SS-1 - RR-1] +
(k′ + k′1 + k′2 + 2k′12)[SR-2 - RS-2] (10)
The first and second fractions were combined, dissolved in
aqueous dioxane, and oxidized with OsO4/NaIO4.5 Following
a normal workup the crude mixture of aldehydes obtained was
reduced with LiAlH4 to give a mixture of 2-methylcyclo-
butanemethanol isomers (Scheme 4).
This pair of differential equations will have solutions of the
form shown in expressions 11 and 12.
[SS-1 - RR-1](t) ) C1 exp(-λ3t) + C2 exp(-λ4t)
[SR-2 - RS-2](t) ) D1 exp(-λ3t) + D2 exp(-λ4t)
(11)
(12)
The enantiomeric excess values for trans-methylcyclo-
butanemethanols (SS-5 and RR-5) were found using a capillary
CycloSil B column. Analyses with the same column at a slightly
higher temperature gave ee values for samples of cis-2-
methylcyclobutanemethanols (SR-6 and RS-6). Authentic ra-
cemic samples of the trans and cis alcohols and of (+)-(1S,2S)-
trans-2-methylcyclobutanemethanol (SS-5) and (-)-(1S,2R)-cis-
2-methylcyclobutanemethanol (SR-6) were prepared and used
to assign the separated GC peaks to specific enantiomers.
Analyses of thermal reaction mixtures derived from SS-1 and
from SR-2 gave the ee data summarized in Tables 4 and 5.
From these data and the rate constants of Table 1, and eqs 1-4,
the eight independent rate constants for [1,3] carbon shifts
linking 1-(E)-propenyl-2-methylcyclobutanes with 3,4-dimeth-
ylcyclohexenes were found.
Thus, according to theory, the two functions should depend
on the same two exponential terms whatever the initial condi-
tions, while C1, C2, D1, and D2 will depend on initial concentra-
tions. The differences in mole percent concentrations of
enantiomers at any time are readily calculated, knowing the sum
of the two mole percent concentrations from eqs 1-4 and the
enantiomeric excess values summarized in Tables 4 and 5. The
Table 6. Mole Percent Differences of Enantiomers of 1, 2, 3, and
4 in Mixtures from SS-1
time (s)
RR-1
SR-2 − RS-2
SR-3 − RS-3
SS-4 − RR-4
0
7200
98.8
80.9
67.1
47.7
34.5
25.5
18.7
12.9
-0.2
-0.3
-0.2
-0.1
-0.1
-0.2
-0.1
0.1
0.0
1.4
2.5
4.2
5.3
6.3
7.0
7.3
0.0
0.8
1.5
2.2
3.0
3.4
3.6
3.8
14400
28800
43200
57600
72000
86400
The SS-1 material used in the kinetic runs summarized in
Table 2 was 99. 8% 1 and 0.2% 2. Both components were 2S
isomers as a result of common synthetic precursors, and both
(5) Pappo, R.; Allen, D. S. Jr.; Lemieux, R. U.; Johnson, W. S. J. Org. Chem.
1956, 21, 478-479.
9
15872 J. AM. CHEM. SOC. VOL. 125, NO. 51, 2003

1-(E)-Propenyl-2-methylcyclobutanes
A R T I C L E S
Table 7. Mole Percent Differences of Enantiomers of 1, 2, 3, and
4 from SR-2 at 275 °C
Scheme 5
time (s)
SS-1 − RR-1
SR-2 − RS-2
SR-3 − RS-3
SS-4 − RR-4
0
1800
3600
0.0
0.3
0.8
1.6
2.2
2.4
2.4
2.2
99.0
82.0
67.8
46.6
32.5
22.4
15.7
10.3
0.0
0.4
0.7
1.2
1.6
1.8
1.9
2.1
0.0
0.1
0.2
0.3
0.4
0.5
0.5
0.5
7200
10800
14400
18000
21600
Scheme 6
2]dt values at any time t.
[SS-1 - RR-1](t) ) 89.3 exp(-λ3t) + 9.5 exp(-λ4t) (13)
[SS-1 - RR-1](t) ) 4.4 exp(-λ3t) - 4.4 exp(-λ4t) (14)
Figure 3. Mole percent concentrations of [SS-1 - RR-1] (9) and [SR-2
- RS-2] (b) for reactions at 275 °C starting from SS-1, and least-squares-
fit function of the theoretically required form shown in eq 13 for the [SS-1
- RR-1] data.
[SR-2 - RS-2](t) ) 1.3 exp(-λ3t) + 97.7 exp(-λ4t) (15)
Enantioselective Rate Constants for [1,3] Carbon Sigma-
tropic Shifts. At any given time during a thermal reaction the
processes leading to the two trans enantiomers of 3,4-dimeth-
ylcyclohexene, SR-3 and RS-3, are as portrayed in Scheme 5.
The corresponding reactions leading to cis enantiomers SS-4
and RR-4 are shown in Scheme 6.
The relationships of Schemes 5 and 6 may be expressed
through expressions 16 and 17 or, in integrated form, 18 and
19.
d[SR-3 - RS-3]/dt ) (ksi - kar)[SS-1 - RR-1] +
(k′sr - k′ai)[SR-2 - RS-2] (16)
d[SS-4 - RR-4]/dt ) (ksr - kai)[SS-1 - RR-1] +
(k′si - k′ar)[SR-2 - RS-2] (17)
[SR-3 - RS-3](t) ) (k - k ) t [SS-1 - RR-1]dt +
∫
Figure 4. Mole percent concentrations of [SS-1 - RR-1] (9) and [SR-2
- RS-2] (b) for reactions at 275 °C starting from SR-2, and least-squares
fits of the theoretically required form shown in eqs 14 and 15.
si
ar
0
(k′ - k′ ) t [SR-2 - RS-2]dt (18)
∫
sr
ai
0
[SS-4 - RR-4](t) ) (k - k ) t [SS-1 - RR-1]dt +
∫
sr
ai
0
differences in mole percent concentrations for the enantiomers
of 1 and 2, starting from either SS-1 or SR-2 (Tables 6 and 7),
follow well the functional forms required (Figures 3 and 4).
(k′ - k′ ) t [SR-2 - RS-2]dt (19)
∫
si
ar
0
When SS-1 is the starting material, eq 13 applies. The [SR-2
- RS-2](t) difference is essentially zero at all times (Table 6,
Figure 3). When SR-2 is the substrate, eqs 14 and 15 are the
appropriate functions. In eqs 13-15 λ3 ) 2.20 × 10-5 s-1 and
λ4 ) 10.6 × 10-5 s-1. The functions 13-15 are easily
differentiated to provide d[SS-1 - RR-1]dt and d[SR-2 - RS-
When SS-1 is the starting material, very little 2 builds up: it
reacts further much faster than it is formed. The mole percent
difference [SR-2 - RS-2] is essentially zero throughout the
reaction (Table 6), and hence the simple linear relationships of
20 and 21 may be used to determine (ksr - kai) and (ksi - kar).
The corresponding plots are shown in Figure 5. The values found
9
J. AM. CHEM. SOC. VOL. 125, NO. 51, 2003 15873

A R T I C L E S
Baldwin and Burrell
Figure 5. Mole percent concentrations of [SR-3 - RS-3] (9) and [SS-4
t
t
- RR-4] (b) as functions of ∫0 [SS-1-RR-1]dt as in eqs 20 and 21, for
Figure 6. Known functions [SR-3 - RS-3](t) - (ksi - kar) ∫0 [SS-1 -
reactions at 275 °C starting from SS-1.
t
RR-1]dt (g) and [SS-4 - RR-4](t) - (ksr - kai) ∫0 [SS-1 - RR-1]dt (n)
t
versus ∫0 [SR-2 - RS-2]dt as in eqs 22 and 23, for reactions at 275 °C
Table 8. Rate Constants for [1,3] Carbon Shifts from SS-1 and
SR-2 in Schemes 5 and 6
starting from SR-2.
constant
value
value (%)a
ksi
kar
ksr
2.26 × 10-6 s-1
0.18 × 10-6 s-1
1.28 × 10-6 s-1
0.17 × 10-6 s-1
1.37 × 10-6 s-1
0.82 × 10-6 s-1
3.85 × 10-6 s-1
1.50 × 10-6 s-1
58
5
33
4
18
11
51
20
kai
k′si
k′ar
k′sr
k′ai
a Of (ksi + kar + ksr + kai) or (k′si + k′ar + k′sr + k′ai).
are (ksi - kar) ) 2.08 × 10-6 s-1 and (ksr - kai) ) 1.12 × 10-6
s-1
.
[SR-3 - RS-3](t) ) (k - k ) t [SS-1 - RR-1]dt
(20)
(21)
∫
si
ar
0
[SS-4 - RR-4](t) ) (k - k ) t [SS-1 - RR-1]dt
∫
Figure 7. Function -d[SS-1 - RR-1]/dt calculated from eq 13 versus
[SS-1 - RR-1] as in eq 24 for reactions at 275 °C starting from SS-1.
sr
ai
0
When SR-2 is the starting material, some of its diastereomer
1 builds up as it decays, but the 1 formed is of relative low ee
Rate Constants for Thermal Stereomutations of SS-1 and
SR-2. The derivations of rate constants for the [1,3] carbon shifts
given above did not depend on explicit knowledge of stereo-
mutation rate constants (Scheme 3), although these equilibrations
played a role in determining time-dependent variations in ee
values for 1 and 2. They may be obtained with the aid of
expressions 9 and 10 and 13-15. For runs starting from SS-1,
-d[SS-1-RR-1]/dt is easily calculated for any time t from 13
and [SS-1 -RR-1] is known at any t (Table 6); [SR-2-RS-2] is
quite negligible (Table 6). Equation 9 simplifies to (24), and
the corresponding plot gives a slope equal to (k + k1 + k2 +
t
(Table 5). Nevertheless, the values of ∫0 [SS-1 - RR-1]dt at
various reaction times do make contributions to expressions 18
and 19. These equations may be rearranged to give 22 and 23,
expressions expressing a simple linear relationship, for all terms
on the left-hand sides of these equations are known.
[SR-3 - RS-3](t) - (k - k ) t [SS-1 - RR-1]dt )
∫
si
ar
0
(k′ - k′ ) t [SR-2 - RS-2]dt (22)
∫
sr
ai
0
2k12), as shown in Figure 7. That slope is 2.63 × 10-5 s-1
.
[SS-4 - RR-4](t) - (k - k ) t [SS-1 - RR-1]dt )
Since (k + k1 + k2) is 23.55 × 10-6 s-1 (Table 1), k12 ) 1.4 ×
∫
sr
ai
0
10-6 s-1
.
(k′ - k′ ) t [SR-2 - RS-2]dt (23)
∫
si
ar
0
-d[SS-1 - RR-1]/dt )
(k + k1 + k2 + 2k12)[SS-1 - RR-1] (24)
The linear plots based on these equations are shown in Figure
6. The slopes of these plots provide (k′sr - k′ai) ) 2.35 × 10-6
s-1 and (k′si - k′ar) ) 0.55 × 10-6 s-1
.
For runs starting from SR-2, expressions 14 and 15 may be
used to calculate -d[SS-1 - RR-1]/dt and -d[SR-2-RS-2]/dt
at any time t. Equation 9 may be rearranged to give a simple
linear relationship 25 with both terms on the lefthand side
These four differences of rate constants and the sums of the
constants from Table 1 provide all eight of the rate constants
for [1,3] carbon sigmatropic shifts (Table 8).
9
15874 J. AM. CHEM. SOC. VOL. 125, NO. 51, 2003

1-(E)-Propenyl-2-methylcyclobutanes
A R T I C L E S
The stereomutation rate constants for SS-1 fall in the order
k2 ) 0.38 × 10-5 s-1, k1 ) 0.27 × 10-5 s-1, and k12 ) 0.14 ×
10-5 s-1. In percentage terms, the ordering is 48% k2, 34% k1,
and 18% k12. For stereomutations of SR-2, the order is k′2 )
1.30 × 10-5 s-1, k′1 ) 0.94 × 10-5 s-1, k′12 ) 0.71 × 10-5
s-1, or, in percentage terms, 44% k′2, 32% k′1, and 24% k′12.
The sum of one-center epimerizations shown by each diaste-
reomer is 3 to 4 times more significant than the competitive
two-center stereomutation process, a pattern totally consistent
with relative rate constants found for many examples of thermal
stereomutations shown by 1,2-disubstituted cyclopropanes.6
Summary
The complexities inherent in the multiple sorts of thermal
reactions shown by vinylcyclobutanes such as SS-1 and SR-2
can be sorted out through careful experimental and analytical
studies, as the present work focused on the stereochemistry of
[1,3] carbon sigmatropic shifts amply demonstrates. The results
obtained bear directly on long contested mechanistic issues and
allow a very clear primary inference to be drawn: The [1,3]
shifts are not controlled by the dictates of orbital symmetry
theory.
Figure 8. Known function -d[SS-1 - RR-1]/dt - (k + k1 + k2 + 2k12)-
[SS-1 - RR-1] versus [SR-2 - RS-2] as in eq 25 for reactions at 275 °C
starting from SR-2.
Full stereochemical studies of [1,3] carbon sigmatropic shifts
leading from substituted vinylcyclopropanes to isomeric cyclo-
pentenes have been reported for a range of systems.7 The results
obtained constitute a consistent pattern of stereochemical
preferences best interpreted in terms of short-lived diradical
intermediates capable of some conformational flexibility before
closing to an isomeric vinylcyclopropane or one of four isomeric
cyclopentenes. Orbital symmetry constraints are not in evidence.
Dynamic factors appear to be instrumental determinants of
stereochemical outcomes.8
A direct comparison between SS-1 and the [1,3] shifts shown
by the analogous cyclopropane system 7 reveals considerable
similarities. The distribution of reaction paths favors the
“allowed” options and formation of trans-dimethyl-substituted
products, but some cis products are evident as well. The
proportions are si:ar:sr:ai ) 58:5:33:4 for SS-1, and 65: 8:22:5
for 7.1
Figure 9. Known function -d[SR-2 - RS-2]/dt - (k1 - k2 )[SS-1 - RR-
1] versus [SR-2 - RS-2] as in eq 26 for reactions at 275 °C starting from
SR-2.
known. The corresponding plot (Figure 8) provides (k′1 - k′2)
) -3.54 × 10-6 s-1
.
-d[SS-1 - RR-1]/dt -
(k + k1 + k2 + 2k12)[SS-1 - RR-1] )
(k′1 - k′2)[SR-2 - RS-2] (25)
Two other substituted vinylcyclobutanes, the deuterium-
labeled racemic trans and cis isomers of 1,2-dicyanocyclo-
butanes (8 and 9), have been subjected to careful experimental
studies leading to full stereochemical details for the [1,3] carbon
shifts leading to substituted cyclohexenes.9,10 The stereochemical
findings for the isomerizations shown by 8 and 9 are remarkably
similar to those found here for SS-1 and SR-2. The trans systems
Since (k′1 + k′2) ) 2.24 × 10-5 s-1 (Table 1), then k′1 )
0.94 × 10-5 s-1, k′2 ) 1.30 × 10-5 s-1, and k′1/k′2 ) 0.72.
Here microscopic reversibility comes into play, for k1k′2 ) k′1k2
(Scheme 3), or k1/k2 ) k′1/k′2. From (k1 + k2) ) 6.55 × 10-6
s-1 and k1/k2 ) 0.72, k1 ) 0.27 × 10-5 s-1, and k2 ) 0.38 ×
10-5 s-1, and (k1 - k2) ) -1.1 × 10-6 s-1. Thus, both left-
hand side terms in the rearranged form of eq 10, that is, 26, are
known at any t, and the equation defines a simple linear function.
Plotting this function (Figure 9) gives slope ) (k′ + k′1 + k′2
+ 2k′12) ) 10.48 × 10-5 s-1. With (k′ + k′1 + k′2 ) ) 9.05 ×
10-5 s-1 (Table 1), the two-center epimerization rate constant
(6) Baldwin, J. E. In The Chemistry of the Cyclopropyl Group; Rappoport, Z.,
Ed.; Wiley: Chichester, 1995; Vol. 2, pp 469-494.
(7) Baldwin, J. E. Chem. ReV. 2003, 103, 1197-1212.
(8) (a) Baldwin, J. E.; Villarica, K. A.; Freedberg, D. I.; Anet, F. A. L. J. Am.
Chem. Soc. 1994, 116, 10845-10846. (b) Davidson, E. R.; Gajewski, J. J.
J. Am. Chem. Soc. 1997, 119, 10543-10544. (c) Houk, K. N.; Nendel,
M.; Wiest, O.; Storer, J. W. J. Am. Chem. Soc. 1997, 119, 10545-10546.
(d) Baldwin, J. E. J. Comput. Chem. 1998, 19, 222. (e) Doubleday: C. J.
Phys. Chem. A 2001, 105, 6333-6341.
k′12 ) 0.71 × 10-5 s-1
.
(9) Cheng, X. Ph.D. Dissertation, Harvard University, 1989; Diss. Abstr. Int.
B 1990, 50, 3472.
(10) Doering, W. von E.; Cheng, X.; Lee, K.; Lin, Z. J. Am. Chem. Soc. 2002,
124, 11642-11652.
-d[SR-2 - RS-2]/dt - (k1 - k2 )[SS-1 - RR-1] )
(k′ + k′1 + k′2 + 2k′12)[SR-2 - RS-2] (26)
9
J. AM. CHEM. SOC. VOL. 125, NO. 51, 2003 15875

A R T I C L E S
Baldwin and Burrell
concentrated by distillation. The ethereal solution (200 mL) was cooled
to 0 °C, and LiAlH4 (0.62 g, 16.4 mmol) was added.14,15 A concentrated
crude product was secured through a conventional workup. It was found
to contain trans-2-methylcyclobutanemethanol (SS-5) and a small
amount of cis-2-methylcyclobutanemethanol (RS-6) (0.68 g, 83% for
the two steps, trans/cis, 84/16, Ultra 2 column) as an 8% solution. A
small sample of SS-5 was purified by preparative GC (1-m, 17%
Carbowax, 110 °C) and had [R]D +91 (c 0.085, CHCl3). Analysis by
GC using a CycloSil B capillary column (70 °C, 15 psi) gave one peak
with a retention time of 17.1 min. A racemic sample of 52b under these
conditions gave two peaks with retention times (relative peak areas)
16.6 min (48.2%) and 17.1 min (51.8%). Samples of 5 and SS-5 were
combined and analyzed as described above to give two peaks with
retention times (relative peak areas) of 16.7 min (17.4%) and 17.1 min
(82.6%).
favor trans products, formed through orbital symmetry “allowed”
paths: si:ar:sr:ai ) 58:5:33:4 for SS-1 and 47:20:27:6 for 8.10
The cis systems also favor formation of trans cyclohexene
products, which can be reached only through orbital symmetry
“forbidden” paths: si:ar:sr:ai ) 18:11:51:20 for SR-2 and 13:
5:66:16 for 9.10
The experimental evidence leaves little room for imagining
that these sigmatropic shifts may be influenced by orbital
symmetry considerations. The stereochemical outcomes strongly
suggest the intervention of transient diradical intermediates
which may either revert to starting material or an isomeric
version of starting material, giving a stereomutation product,
or proceed to form one or another [1,3] carbon shift product, a
substituted cyclohexene. The stereochemical outcome may
depend on the conformational details of the diradical as the new
C-C bond begins to form, which may in turn hinge on dynamic
effects as well as on the potential energy surface.
Other vinylcyclobutane systems within bicyclic structures
which rearrange through [1,3] carbon sigmatropic shifts, includ-
ing some once thought to be controlled by orbital symmetry
rules, now seem on closer consideration to probably involve
diradical intermediates.11 There are instances of high selectivity
in favor of orbital symmetry allowed stereochemical outcomes,
others showing strong preferences for forbidden paths, and some
following allowed and forbidden paths to comparable extents.
A diradical mechanistic formulation rationalizes them all in a
consistent fashion.11
One may anticipate that theoretical work to afford good
potential energy surfaces and reaction dynamics calculations
for vinylcyclobutane-to-cyclohexene isomerizations will be
forthcoming, probably after experimental information on reac-
tion stereochemistry becomes available for a system labeled so
minimally that all four [1,3] carbon sigmatropic shift paths have
equal enthalpic barriers.12 Suitable deuterium-labeled versions
of (E)-propenylcyclobutane or of vinylcyclobutane could be
prepared and studied to provide full stereochemical information.
In such studies, however, the stereochemical characteristics of
reactants and products could not be secured as functions of
reaction time through the sorts of GC analyses on which the
present work relied. Analytical separations by capillary GC and
by “chiral” GC methods would be of no avail!
(-)-(1S,2R)-cis-2-Methylcyclobutanemethanol (SR-6). A solution
of (-)-(1S,2R)-cis-2-methylcyclobutanecarboxylic acid2b of better than
99% ee (0.88 g, 7.7 mmol) in ether (10 mL) was added dropwise to a
suspension of LiAlH4 (0.29 g, 7.7 mmol) in ether (60 mL) at 0 °C.
The reaction and workup proceeded to give 0.68 g (87% yield) of (-)-
(1S,2R)-cis-2-methylcyclobutanemethanol (SR-6). A small sample was
purified by preparative GC (1-m, 17% Carbowax, 100 °C) and had
[R]D -15 (c 0.39, CHCl3). Analysis by GC using a CycloSil B column
(85 °C, 15 psi) gave one peak with a retention time of 13.0 min. A
sample of racemic 62b on this column under the same conditions gave
two peaks with retention times (relative peak areas) of 12.7 min (48.4%)
and 13.1 min (51.6%). Samples of 6 and SR-6 were combined and
analyzed as described above to give two peaks with retention times
(relative peak areas) of 12.8 min (25.1%) and 13.1 min (74.9%).
Thermal Reactions of SS-1 and SR-2. A sample of SS-1 of better
than 99% ee was purified by preparative GC (2.3-m, 20% ODPN on
Chromosorb P AW-DMCS, 50 °C) and again on a second column (1-
m, 10% SE-30, 60 °C) to provide 123 mg of SS-1 free of any trace
of 1,1-diiodoethane. This SS-1 material was dissolved in pentane (3.9
mL), and cyclooctane (62 mg) was added as an internal standard. Eight
500-µL samples of this stock solution were injected into a 300-mL
quartz bulb encased in an aluminum block and heated at 275.1 ( 0.1
°C for various times.2e At the end of each kinetic run the thermal
reaction mixture was removed from the bulb and analyzed by GC (DB-
1301), then separated into four fractions by preparative GC (2.3-m,
20% ODPN on Chromosorb P AW-DMCS, 50 °C). Fraction 1 contained
1; fraction 2 contained trans-1,5-octadiene, cis-3,6-dimethylcyclohexene
(4), and 2; fraction 3 contained cis-1,5-octadiene, trans-1,6-octadiene,
and trans-3,4-dimethylcyclohexene (3); fraction 4 contained cis-3,4-
dimethylcyclohexene (4). Analysis of fraction 3 by GC using a CycloSil
B column (50 °C, 15 psi) gave the enantiomeric excess of 3.2c Analysis
of fraction 4 by GC using a column with a Cyclodex B column (30 °C
for 38 min, inc 10 °C/min to 100 °C, 12 psi) gave the enantiomeric
excess of 4.2c Fractions 1 and 2 were combined and dissolved in 2 mL
of dioxane. This mixture was placed in a round-bottomed flask along
with water (0.6 mL) and OsO4 (∼10 mg, 2.5% solution in tert-butyl
alcohol).5 After 5 min at room temperature, NaIO4 (20 mg) was added.
The brown solution was then stirred for 1 h at room temperature. At
that time the solution was quenched with water (5 mL), and the aqueous
layer was extracted with ether (4 × 5 mL). The organic layers were
combined, washed with water (3 × 5 mL), dried (Na2SO4), and filtered.
The mixture of aldehydes in ether was added to a round-bottomed flask.
The solution was cooled to 0 °C, and LiAlH4 (∼50 mg) was added.
The gray suspension was warmed to room temperature and stirred for
15 min. At that time the mixture was cooled to 0 °C and quenched
with water (5 mL). The aqueous layer was acidified with 2 M HCl (5
mL) and extracted with ether (4 × 5 mL). The organic layers were
combined, dried (Na2SO4), filtered, and concentrated by distillation to
Experimental Section
(+)-(1S,2S)-trans-2-Methylcyclobutanemethanol (SS-5). To a 25-
mL flask were added an ethereal solution of (+)-(1R,2S)-cis-1-
methoxycarbonyl-2-methylcyclobutane2b (better than 99% ee by cap-
illary GC on a G-TA γ-cyclodextrin column at 45 °C, 15 psi; [R]D
+58 (c 0.25, CHCl3); 1.05 g, 8.2 mmol) and a 1 M solution of potassium
tert-butoxide (16.4 mL, 16.4 mmol) in THF.13 The yellow solution was
stirred for 10 min at room temperature under argon. At that time the
mixture was quenched with water (75 mL), and the aqueous layer was
extracted with ether (5 × 25 mL). The organic layers were combined
and washed with water (2 × 25 mL). The aqueous solution was acidified
with 2 M HCl and extracted with ether (3 × 25 mL). The organic
solutions were combined, dried (Na2SO4), filtered, and partially
(11) Leber, P. A.; Baldwin, J. E. Acc. Chem. Res. 2002, 35, 279-287.
(12) See articles cited in ref 8, and (a) Baldwin, J. E.; Keliher, E. J. J. Am.
Chem. Soc. 2002, 124, 380-381. (b) Suhrada, C. P.; Houk, K. N. J. Am.
Chem. Soc. 2002, 124, 8796-8797. (b) Doubleday: C.; Suhrada, C. P.;
Houk, K. N. Manuscript in preparation.
(13) Ito, Y. K.; Ariza, X.; Beck, A. K.; Boha´c, A.; Ganter, C.; Gawley, R. E.;
Ku¨hnle, F. N. M.; Tuleja, J.; Wang, Y. M.; Seebach, D. HelV. Chim. Acta
1994, 77, 2071-2110.
(14) To¨ro¨k, B.; Molna´r, A. J. Chem. Soc., Perkin Trans. 1 1993, 801-804.
(15) Hill, E. A.; Chen, A. T.; Doughty, A. J. Am. Chem. Soc. 1976, 98, 167-
170.
9
15876 J. AM. CHEM. SOC. VOL. 125, NO. 51, 2003

1-(E)-Propenyl-2-methylcyclobutanes
A R T I C L E S
give a mixture of alcohols 5 and 6. Analysis of the concentrate by GC
using a CycloSil B column (75 °C, 15 psi) gave an ee value for 5.
Analysis of the concentrate by GC on the same column at a moderately
higher temperature (85 °C, 15 psi) gave the ee of 6. The ee
measurements for 3, 4, 5, and 6 were performed in duplicate, and the
data were tabulated and averaged (Table 4, where data for 5 and 6 are
ascribed to 1 and 2).
Similarly, a sample of (-)-(1S,2R)-cis-1-(E)-propenyl-2-methyl-
cyclobutane (SR-2) of better than 99% ee was purified by preparative
GC (2.3-m, 20% ODPN on Chromosorb P AW-DMCS, 55 °C) and
then using a second GC column (1-m, 10% SE-30, 60 °C) to give
122 mg of SR-2 completely free of any trace of 1,1-diiodoethane. The
sample was dissolved in pentane (3.9 mL), and cyclooctane (61 mg)
was added as an internal standard. Eight 500-µL samples of this stock
solution were injected into the bulb and heated at 275.1 ( 0.1 °C over
a range of times. When the time period planned for a given kinetic run
was over, the thermal reaction mixture was removed from the bulb
and analyzed by GC (DB-1301). It was processed as described above,
and ee values for 3, 4, 5, and 6 were measured in duplicate, and the
data were tabulated and averaged (Table 5, where data for 5 and 6 are
ascribed to 1 and 2).
Acknowledgment. We thank the National Science Foundation
for supporting this work through Grants CHE-9902184 and
CHE-0211120.
JA030516T
9
J. AM. CHEM. SOC. VOL. 125, NO. 51, 2003 15877