Syntheses, thermoreorganizations, and secondary deuterium iosotope effects of methylene-3-vinylcyclobutanes.

MISSING IMAGE

Material Information

Title:
Syntheses, thermoreorganizations, and secondary deuterium iosotope effects of methylene-3-vinylcyclobutanes.
Uncontrolled:
Methylene-3-vinylcyclobutanes
Physical Description:
viii, 128 leaves. : ill. ; 28 cm.
Language:
English
Creator:
Mancini, Gregory J., 1946-
Publication Date:

Subjects

Subjects / Keywords:
Rearrangements (Chemistry)   ( lcsh )
Deuterium   ( lcsh )
Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis--University of Florida.
Bibliography:
Bibliography: leaves 123-126.
Statement of Responsibility:
Gregory J. Mancini.
General Note:
Manuscript copy.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001027745
notis - AFA9764
oclc - 18066859
System ID:
AA00003938:00001

Full Text












SYNTHESES, TIHERMOREORGANIZATIONS, AND
SECONDAi"Y' DEUTERIUM ISOTOPE EFFECTS
OF METHYLENE-3-VINYLCYCLOBUTANES












By

Gregory J. Mancini


A DISSERTATION PRICil:'I.ID TO TiE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLML'iNT OF THE REQUIREMENTS FOR THE
DE:r.g:; OF DOCTOR OF PHILOSOPHY







UNIVERSITY OF FLORIDA


1975














ACKNOWLEDGE EVENTS


I would like to express my appreciation for the support

and guidance of Dr. William R. Dolbier, Jr., during the

pursuit of this research. I would also like to extend my

thanks to Dr. Roy W. King for his assistance and to the

members of our research group for their spirited friendship

and moral support. Most importantly, I would like to

express my gratitude and appreciation to the support per-

sonnel within the department, being too numerous to name,

without whose cooperation and assistance the execution and

completion of this research would have been nearly

impossible.















TABLE OF CONTENTS


Page


ACKNOWLEDGEMENTS ii

LIST OF TABLES iv

LIST OF FIGURES v

ABSTRACT vii

INTRODUCTION 1

RESULTS 16

DISCUSSION 29

EXPERIMENTAL 54

APPENDIX I 91

APPENDIX II 104

REFERENCES 123

BIOGRAPHICAL SKETCH 128


iii














LIST OF TABLES


Table Page


1 Kinetics for the thermal rearrangement
of (2) to (3) 64

2 Kinetics for the thermal rearrangement
of (2-d ) to (3-d) 70
-4 .- -4
3 Nmr integration in the aliphatic
region of the product mixture of (3a + 3b) 76

4 Nmr integration in the vinylic
region of the product mixture of (3a + 3b) 77

5 Intramolecular isotope effect results for
the rearrangement of (2-d,) to (3a)
and (3b) 78

6 Kinetics for the thermal rearrangement
of (29) to (30) 87

7 Pyrolysis of 1,l-divinyl-3-methylene-
cyclobutane (2) 93

8 Pyrolysis of 1,l-di(vinyl-2,2-d2 -3-
methylenecyclobutane (2-d4) 97

9 Pyrolysis of methylene-3-vinyl-
cyclobutane (29) 98














LIST OF FIGURES


Figure Page


1 Energy diagram for the pyrolysis of (29) 42

2 Energy diagram for the pyrolysis of (2) 43

3 Arrhenius plot for the conversion of
(2) to (3) 65

4 Sample glpc trace of the kinetic
analysis for the thermal conversion
of (2) to (3) 66

5 Arrhenius plot for the conversion of
(29) to (30) 88

6 Sample glpc trace of the kinetic
analysis for the thermal conversion
of (29) to (30) 89

7 Concentration versus time plot comparison
of compounds (2) and (2-d4) at 99.9' 102

8 Concentration versus time plot comparison
of compounds (2) and (2-d4) at 121.10 103

9 Nmr of 1,l-divinyl-3-methylene-
cyclobutane (2) 105

10 Nmr (100 MHz) at 100 cps sweep width of
1,l-divinyl-3-methylenecyclobutane (2) 106

11 Line analysis for the nmr (100 MHz) at
100 cps sweep width of 1,l-divinyl-3-
methylenecyclobutane (2) 107

12 Ir of l,l-divinyl-3-methylenecyclobutane
(2) 108

13 Nmr of 4-methylene-2-vinyl.cyclohexene (3) 109

14 Ir of 4-methylene-2-vinylcyclohexene (3) 110









Figure


15 Nmr of 3-methylene-l,1-di(vinyl-2,2-d2)-
cyclobutane (2-d ) 111

16 Ir of 3-methylene-l,1-di(vinyl-2,2-d2)
cyclobutane (2-d ) 112

17 Nmr of 4-methylene-2-(vinyl-2,-d2-)
cyclohexene-6,6-d2 (3-d ) 113

18 Ir of 4-methylene-2-(vinyl-2,2-d2)-
cyclohexene-6,6-d2 (3--d4) 114

19 Nmr of 3-methylene-1-vinyl-i-
(vinyl-2,2-d )cyclobutane (2-d ) 115

20 Ir of 3-methylene-l-vinyl-1-
(vinyl-2,2-d2) cyclobutane (2-d ) 116

21 Aliphatic region of the nmr (100 MHz) for
the mixture of dideuterated 4-methylene-
2-vinylcyclohexenes (3a) and (3b) 117

22 Vinylic region of the nmr (100 MHz) for
the mixture of dideuterated 4-methylene-
2-vinylcyclohexenes (3a) and (3b) 118

23 Nmr of methylene-3-vinylcyclobutane (29) 119

24 Ir of methylene-3-vinylcyclobutane (29) 120

25 Nmr of 4-methylenecyclohexene (30) 121

26 Ir of 4-methylenecyclohexene (30) 122


Page















Abstract cf Dissertation Presented to the Graduate
Council of the University of Florida in Partial Fulfillment
of the Requirements for the Degree of Doctor of Philosophy

SYNTHESES, TI!E'~ .MOORGANIZATIONS, AND
SECONDARY DEUTErIUM ISOTOPE EFFECTS
OF METHYLE.NE- 3-VINYLCYC]LOI'TAN'ES

By

Gregory J. Mancini

March, 1975

Chairman: W. R. Dolbier, Jr.
Major Department: Chemistry

The thermoreorganizations of methylene-3-vinyl-

cyclobutane to 4-methylenecyclohexene and 1,1-divinyl-3-

methylenecyclobutane to 4-methylene-2-vinylcyclohexene

were investigated. Two mechanistic options were available

for these rearrangements; one which involved the

intervention of diradical intermediates, the other involved

a concerted [3,3]-sigmatropic or Cope rearrangement.

Activation parameters were determined for both of these

pyrolyses. The heats of formation of their transition

states were approximated by the group equivalent method, and

these were consistent with the lower heats of formation,

calculated for the speculated diradical intermediates. The

low entropies of activation for these rearrangements could

very reasonably be accounted for by the loss of internal


vii








rotations of the vinyl groupss, and the subsequent

restriction of a number of atoms to two rigid planar

conjugated radicals.

The indication from the kinetic data that diradical

intermediates intervene in these thermoreorganizations

was corroborated by an intermolecular secondary deuterium

isotope effect study where the rates of reaction of

1, .-divinyl-3-methylenecyclobutane were compared with

1,1-di(vinyl-2,2-d2)-3-methylenecyclobutane. The

preponderant weight of analogy indicates that transformation

of an isotopically substituted carbon from a trigonal to a

tetrahedral orientation in the product should be

associated with an inverse secondary deuterium isotope

effect. However, a normal intermolecular isotope effect

(k /kD = 1.0510.03) was observed at 99.90 which can be

accounted for by a diradical process. From examination

in the nmr of the product ratio, resulting from the

thermolysis of 3-methylene-l-vinyl-l-(vinyl-2,2-d2)-

cyclobutane, an intramolecular isotope effect (kH/kD

1.140.07) was determined. The discrimination process

which gives rise to this kind of secondary deuterium isotope

effect is thought to involve a twisting motion of an

isotopically substituted, terminal methylene group,

resulting in a nonplanar pentadienyl radical with

concomitant destruction of q-bonding at its terminus.


viii













INTRODUCTION


The secondary deuterium isotope effect study for the

thermal reorganization of 1,1-divinylcyclopropane (1) by

Dolbier and Alonsol prompted an examination of the homol-

ogous and novel l,l-divinyl-3-methylenecyclobutane (2).






(2) (3)

The intriguing aspect of the latter, relative to (1), was

that a mechanistic alternative was available for its thermal

conversion to (3), that is, the potential existed that (2)

might thermally convert to (3) by either diradical (4) or

by a concerted [3,3]-sigmatropic rearrangement depicted by

(5).






(4) (5)

The [3,3]-sigmatropic rearrangement was first dis-

covered by Cope and coworkers.2 In a study of substituted

1,5-hexadienes of which (6) was typical, they observed that

(6) was not cleaved into a pair of allyl radicals upon









heating to 100". Dissociation into two allyl radicals

which could recombine in a second step to give cross

products, as (8), was excluded, since only product (7) was

formed. This process demonstrated a lack of response to
Ph


(7)

Ph 0 Ph
1000
Me SMe
Ph
MMe
(6_ ) (8)


Me



catalysts, changes in solvent medium and reaction phase.

Moreover, an additionally significant argument against

cleavage of an acyclic 1,5-hexadiene to a bisallyl system

in the rate-determining step was provided by the character-

istically large negative values for entropies of activation
4
(approximately -12 eu) measured for several Cope reactions.

The Cope rearrangement would thus pass through a cyclic

multicentered transition state in which simultaneously

(or nearly so) old bonds are broken while new bonds are

formed.

One interesting feature of the Cope rearrangement is

that the transition state could have either the four-

centered chair (9) or the six-centered boat (10)

conformation.5 A classic experiment devised by von Doering








and Roth6 was particularly successful in establishing

which conformation was the preferred transition state. It

involved the thermal rearrangement of meso-3,4-dimethyl-

1,5-hexadiene (11) which at 2250 almost exclusively yielded

cis,trans-2,6-octadiene. This could only involve a four-

centered chair-like transition state (9). If a six-centered

boat-like conformation (10) were involved, only

H



H Me




(9) (10) (11)

cis,cis- and trans,trans-2,6-octadiene would have resulted.

They estimated that the difference in free energy in (11)

between the four-centered (9) and the six-centered (10)

conformation amounted to at least 5.7 kcal/mol. This

corresponded nicely to MINDO/2 calculations which correctly

predicted .the chair-like transition state to be more

stable than the boat by 6.6 kcal/mol. Alternative

theoretical treatments for the observed preference of the

four-centered chair-like conformation in the concerted

[3,3]-migrations of acyclic 1,5-hexadienes have been

discussed. However, in cis-1,2-divinylcycloalkanes, as
Q 10
(12 and (14), the boat-like transition state was

energetically more accessible, and the thermal Cope reaction

proceeded quantitatively.










15> 100.



(12) (13) (14) (15)

When neither a chair-like nor a boat-like conformation
could be attained in the transition state, because the
requisite geometry for a concerted 13,3]-sigmatropic
migration would be energetically inaccessible, the chemical
system would of necessity seek a lower energy diradical
pathway if available. This would be specifically true for

a small carbocyclic ring substituted with vinyl groups,
where any incipient radical formed would be allylic and
resonance-stabilized. A case in point is the trans-1,2-
divinylcyclobutane (16) whose thermal rearrangement at 1540
provided three products. The major product 4-vinyl-
cyclohexene (18) could arise either by diradical (17) or
by a concerted [1,3]-sigmatropic migration. The latter was
not supported by kinetic data nor analogy.11' 12 The
secondary product 1,5-cyclooctadiene (15) could also arise





D+ +


70% 20% 4%

(16) (17) (18) (15) (19)








by diradical (17) and/or by cis-trans isomerization from

diradical (17) to cis-1, 2-divinylcyclobutane (14) which

then under the reaction conditions would proceed smoothly

to (15) via a concerted Cope rearrangement. The most

:s Lm.ple mechanistic rationalization would invoke the

diradical intermediate (1.7) as the major pathway with the

concerted [3,3]-sigmatropic rearrangement being preempted

for geometric reasons.

Similarly, Berson and Walsh examined the pyrolytic

behavior of endo- and exo-2-vinylbiclo[22.2.2]oct-5-ene

(20 and 21 respectively).13 The endo-vinyl epimer (20)

could undergo transformation to cis-A '6-hexalin (22)

through either a concerted [3,3j-sigmatropic migration or

through biradical (20a). However, the Cope rearrangement

was geometrically inaccessible to the exo-vinyl epimer (21),




H r-4/



(20) (21)


3240 3240



A (22) A


(21a)


C--~-------


(20a)








so that the biradical (21a) pathway became an energetically

attractive alternative for transformation to (22).

Competition experiments showed that the rates for overall

disappearance between (20) and (21) over an 800 temperature
*
range were nearly identical and gave a AAH of only

1.8+1.5 kcal/mol with the exo-vinyl compound (21) apparently

having the lower AH The authors felt that the data

argued against a concerted Cope rearrangement for either

(20) or (21) and were most simply explained in terms of a

stepwise mechanism passing through diradical intermediates.

Lastly, secondary deuterium isotope effect studies

have been utilized in recent years by physical organic

chemists to test the simultaneity or lack of it in peri-

cyclic reactions, such as the thermal Cope rearrangement.14

What makes isotope effects unique and particularly

attractive as a mechanistic probe is that the potential

energy surfaces, unlike substituent effects, for isotopic

isomers are identical. This would imply that the origin of

isotope effects is of a vibrational, nonelectronic nature.

It was found that for hydrogen isotopes, where vibrational

frequencies and, therefore, zero point energies are large,

that the major contribution to isotope effects stems from

zero point vibrational energy differences. This

simplification arises because all vibrations which involve

hydrogen and its isotopes would essentially remain, at

ordinary temperatures, in their lowest (zero point) energy

levels.1, 15 Indeed, the zero point energy level would






7


always be less for D (or C-D' than for H (or C-H)
,. 15
vibrations.5

Observation of secondary deuterium isotope effects

have been correlated with steric interactions,16

inductive15, 17 and hyperconjugativel8 effects, and
19
rehybridization changes,1 dependent upon which concept

clarified the experimental data best. These points of

view, although somewhat misleading, are interpretations

of vibrational force constant or zero point energy changes.

Yet, they have allowed the practicing chemist to discuss

isotope effects in the same familiar language that is

commonly used to discuss substituent effects, although

the latter have an electronic origin. So too, it allows

reasonable qualitative predictions to be made with regard

to sign and magnitude of vibrational force constant changes

in an isotope effect study as a preliminary to quantitative

measurements.

Because thermal reorganizations of small carbocycles

generally involve little if any charge separation and

usually proceed either through a synchronous or a diradical

pathway, a steric interpretation would appear to offer the

most useful clarification and rationalization of any

deuterium isotope effect study. Indeed, through the
16b, 16c
harmonic approximation employed by Bartell, a one

to one correspondence between steric and vibrational

force constant changes can be made. Thus, a steric approach

might suggest independent predictions of the potential








energy surfaces of hydrogen isotopes in a systematic

fashion.

A steric interpretation of hydrogen isotope effects

derives from the smaller amplitude and effective size of D,
15
so that it has less of a steric requirement than H. This

is purely a vibrational phenomenon, since the more sterical-

ly crowded a C-H bond is, the greater will be the vibra-

tional frequencies, especially the out-of-plane bending
14
frequencies, associated with that bond.4 Thus, on going

from reactant to a transition state which is sterically

more congested, the rate of reaction for the protio isomer

kH should be less than the rate of reaction for the

deuterio isomer kD, and an inverse isotope effect

(kH/kD < 1) should result. If there is less steric

congestion from reactant to transition state, then a

normal isotope effect (kH/kD > 1) ought to result. A

good case in point was the racemization of compound (23).


Br H(D)

HOOC-- )- j -COOH (23)

(D)HI Br


The reaction has a high activation energy with a planar

transition state,20 so that sizeable van der Waals inter-

actions between adjacent H and Br are present. On steric

grounds, an inverse isotope effect would be expected and

in fact was observed (kIH/kD = 0.840.03).









In the Cope rearrangement, both kinetic and thermo-

dynamic deuterium isotope effect studies have been probed

in some detail. When a kinetic deuterium isotope effect

was measured at the bond-making site for the thermal

[3,3]-sigmatropic transformation of the substituted 1,5-

hexadiene (24 vs. 24-d2), an inverse kinetic isotope
21
effect (kH/kD = 0.940.02) resulted. In this case,



Et Me NC Et

Me
NC 930 NC H

=CH2 H

(24) (25)



Et Me NC Et


NC .930 N D

D2 DD


(24-d2) (25-d2)



the sp carbon bearing H has significantly higher C-H
2
bending vibrations associated with it than for an sp
14
carbon. Thus, as the migration in (24) proceeds through
2
a cyclic multicentered transition state, the sp carbon at

C-6 becomes more sp -like, so that the steric requirement

of H attached at the more crowded tetrahedral geometry of
an sp3 position compels (24) to r range more reluctantly
an sp position compels (24) to rearrange more reluctantly









to (25) than (24-d ) to (25-d2) However, this inverse

kinetic isotope effect could also be rationalized by

invoking a two-step mechanism involving the diradical

intermediate (26). To resolve this ambiguity, the


Et
NC

(24-d = = N --M- (25-d2
2 N ( 2
D

(26)

kinetic deuterium isotope effect was also determined at

the bond-breaking site by comparing the rates of rearrange-

ment of (24) with the suitably labeled (24'-d2). At


Et Me NC Et

NC- 930 NCMe


D- == D2C


(24'-d2) (25'-d2)
221

930 a normal isotope effect of 1.190.03 was observed.21

The combination of these two isotope effects precluded the

two-step mechanism as a viable alternative and reinforced

other corroborative experiments, demonstrating the con-

certedness of the thermal rearrangement of acyclic 1,5-

hexadienes.2 These two transformations were practically

irreversible which reflected the greater stability of the

product (AH = -5 kcal/mol), so that the transition state
22
should be more reactant- than product--like. Because









of the proportionately larger kinetic isotope effect at

the bond-breaking site, is bond-breaking farther along

than bond-making? The authors argued that force constant

changes are nonlinear functions along the reaction

coordinate. This is supported by theoretical calcu-

lations.23 Thus, with a relatively small increase in

C-C bond distance, a relatively sharp decrease in the

force constant would ensue. Likewise, as the two atoms

approach one another in the bond-forming process, a

small increase in the force constant should not be

an unlikely expectation.

More frequently thermodynamic secondary deuterium

isotope effects have been encountered in Cope rearrange-

ments.24 The degenerate nature of many of the systems

studied precluded any possible measurement of a kinetic

isotope effect. The thermal equilibrium of labeled

biallyl (27) was observed. Based on previous arguments,


D

C 2 220 2 I D
CDCD
CD., D

(27a) (27b)

where D can better accommodate a more sterically congested

environment, D would be expected to accumulate preferen-

tially at the tetrahedral orientation of the isotopically

substituted carbons. The product ratio (27a/27b) of

0.810.02 was consistent with expectations.25








The thermal rearrangement of 1,1-divinylcyclo-

propane (1) to vinylcyclopentene (28) does not involve

a [3,3--sigmatropic or Cope process. Nevertheless,

the kinetic isotope effects ascertained are enlightening

and suggestive. Both intermolecular and intramolecular

kinetic isotope effects were determined. The former

(k /kD = 1.0810.07) compares the rates of reaction of




A- CD


D D

(1) (28) (1-d ) (28-d )


(1) with (1-d,) and reflects the force constant changes
1
in the rate-determining step. Had the above trans-

formation been a concerted [l,3]-sigmatropic process, an

inverse isotope effect should result, because the overall

transformation involves the conversion of the hydrogen

isotopes from a trigonal to the more crowded tetrahedral

orientation at the carbon of the product. A diradical

mechanism can justify this normal isotope effect, regard-

less of whether the rate-determiring step involves

biradical formation through ring cleavage or biradical

destruction. If the former, ring cleavage creates new
26
torsional modes which favor H over D. Moreover, since

ring cleavage requires the transformation of the
isotopcaly labelled site from sp2 to sp2radical,
xsotopically labelled site from sp to sp -radical,









experimental analogy indicates that such a transformation

should be accompanied by a small but normal isotope

effect.27 For biradical destruction, Dolbier and

Alonso observed a normal intramolecular isotope effect

(kH/kD = 1.120.02).1 It was determined from the product

ratio (28a/28b) in the thermolysis of (l-d2). If an


CD. 2 -CD2 /---CH2

2 __> +
2 H2
H H D D

(1-d ) planar nonplanar (28a) (28b)
--2

inverse intramolecular isotope effect were observed in

biradical destruction, the bulk of the activation energy

would be associated with ring closure in which the
9 3
sp 2-radical becomes transformed to the more crowded sp

position of the product. However, observation of a

small but normal isotope effect demanded that isotopic

discrimination be associated with the twisting of a

terminal methylene group from its planar to its nonplanar

configuration, with concomitant destruction of n-bonding

at the terminus. Considering that the CH2 group has a

relatively higher zero point energy than a CD2 group, it

requires a relatively lower activation energy for this

twisting motion to the nonplanar pentadienyl form, that is,

the CH, group has a head start in surmounting the energy

barrier for rotation because of its relatively smaller









moment of inertia. Thus, the predicted preference for

the CH2 twisting motion is observed upon examination of

the product ratios where (28a) > (28b).

Had the rearrangement of (1) to (28) been concerted,

bond-breaking and bond-making would have a common transition

state, so that an inverse isotope effect would also be

expected. Thus, despite the near identity of the two

isotope effects, the fact that they are normal, demands

that biradical intermediates be involved. With both

isotope effects being the same within experimental error,

it could be construed that biradical destruction is

rate-determining. However, it could also be construed

that the two isotope effects could each be associated

with a different step in the mechanistic pathway

leading to product and are fortuitously the same.

A secondary deuterium isotope effect study on the

pyrolysis of a 1,5-hexadiene involving prior ring cleavage
28
has not been unambiguously examined.8 In conjunction

with activation parameters, a secondary deuterium isotope

effect study on the thermal conversion of (2) to (3)

might allow distinction between mechanisms. This would

entail a determination of both inter- and intramolecular

isotope effects. The former involves comparison of the

rate ratio of (2) with (2-d4). The latter can be

ascertained from the product ratio upon thermolysis of

(2-d ). Determination and comparison of the activation

parameters for methylene-3-vinylcyclobutane (29) to








4-methylenecyclohexene (30) with those determined from
pyrolysis of (2) would also allow insight into the
mechanism for rearrangement.


CD2
ACD


2CD2

2


(2-d 4)


H


(29)


(30)
(30)















RESULTS


Multistep syntheses were employed to construct the

novel isotopic isomers (2), (2-d ), and (2-d ) as well as

the analogous methylene-3-vinylcyclobutane (29). Because

unambiguous, well-documented reaction types were employed,

product identification was considerably simplified.

Efficient construction of a cyclobutane system generally

requires either (a) the alkylation of active methylene
29
compounds by, for example, 1,3-dibromopropane or (b) a

thermal [2 4 2)-cycloaddition under autogenous pressure

in a "bomb" reaction. The synthetic feasibility of the

latter has been demonstrated particularly with the cyclo-

addition of allene to active ethylene compounds.3 This

entry into substituted methylenecyclobutanes prompted the

preparation of diethyl-3-methylene-l,l-cyclobutane-

dicarboxylate (32) by cycloaddition of methylenediethyl-

malonate (31) with allene in an analogous fashion.





S 1 + allene A C32


(31) (32)









The reaction of acrylonitrile with allene is well known,

and from it (29) can be prepared.3

The common precursor to (2), (2-d ), and (2-d2)

was 3-methylenecyclobutane-l,1-diacetic acid (37) which

was prepared according to the scheme depicted below.

CO2Et CH OH CH OTs
k \ C2OH h 2Ts

CO2Et CIOH CH2OTs
22 2

(32) (33) (34)



CH CO H N Br
2 2 CHCN2Br


H 2CO2H CH2CN =H2Br

(37) (36) (35)

The 3-methylenecyclobutane-l,l-dimethanol (33) was con-

veniently prepared by LiAlH4 reduction of (32). Bis-

tosylation of diol (33) produced (34) which was converted

in high yield to l,l-bis(bromomethyl)-3-methylenecyclo-

butane (35) by treatment with LiBr in refluxing anhydrous

acetone. The ditosylate (34) could have been directly

converted to 3-methylene-l,l-diacetonitrile (36); however,

the literature indicated that such a reaction would involve

some decomposition, and the reaction product would be
31
difficult to work up. Reaction of the dibromide (35)

with KCN in DMSO at 850 for approximately 2 days produced









in 87% yield the expected diacetonitrile (36). Iydrolysis

of this product afforded the diacetic acid (37). Nmr

spectroscopy indicated the characteristic triplet (ring-

allyl H) and pentet (exocyclic-methylene 1) splitting

pattern (J = 2.5 Hz) for all these compounds. Confirmation

by ir and mass spectra and by elemental analyses completed

product identifications, and all were consistent with their

proposed product structures.

From the 3-methylenecyclobutane-1,1-diacetic acid (37),

.,l-divinyl-3-methylenecyclobutane (2) was prepared. This


CH2CO2H CHICH2H C OH CH2 OTs


C 2 2 22 CH 2CH2OTs


(37) (38) (39) (2)


was accomplished by LiAlH4 reduction of the diacetic acid

(37), yielding the diol (38), which was easily converted to

the ditosylate (39). Bisdehydrotosylation was effected in

20% yield by treatment of an anhydrous DMSO solution of

ditosylate (39) with solid potassium-t-butoxide at 250 under

full vacuum. The volatile products were condensed in a

coiled trap, cooled by liquid nitrogen. Separation was

effectively accomplished on a 5' (18%) DC-200 column at 550

The structure of (2) was verified by nmr, ir, and mass

spectroscopy and by elemental analysis. The 100 MHz nmr

(see Figure 10) showed a triplet at 6 2.75 (4Ha J ab

2.7 Hz), a pentet at 4.82 (2Hb, Jab = 2.7 Hz), and a










fourteen-line32 ABX pattern for the divinyl region which

has chemical shifts at 5.01 (2Hc, Jcd = 1.4 Hz, Jce

16.6 Hz), 5.06 (2Hd' Jd = 1.4 Hz, Jde = 10.2 Hz), and

6.03 (2H J = 16.6 Hz, J = 10.2 Hz); its ir spectrum

(see Figure 12) showed bands at 3045, 2995, 2960, 1640,
-i
1415, 1220, 994, 880 cm.

The compound (2) was smoothly converted to 4-methylene-

2-vinylcyclohexene (3) at 1150 in a dilute solution of

either n-decane, benzene, or CC14 using sealed tubes. The

conversion appeared to be quantitative and clean. Both

structure and purity of (3) were confirmed by nmr, ir, and

mass spectroscopy as well as by elemental analysis. Its nmr

(see Figure 13) bore similarities to the known 4-methylene-

cyclohexene (30)33 and showed broad singlets at 6 2.30 (4Ha)

and 2.85 (2Hb), multiplets at 4.55-5.20 (2Hc + 21Hd) 5.52-

5.85 (1He), and an AB quartet at 6.28 (1Hf, Jtrans = 17 Hz,

J cis = 10 Hz); its ir (see Figure 14) showed bands at 3090,
CIS
3070, 3000, 2980, 2960, 2910, 2835, 1650, 1638, 1605, 1430,
-1
1335, 1220, 975, 890, 840 cm ; its uv exhibited a X x at

230 nm (c 22,200).
H H
c c
H H Hi
e d b
H^ / Hd

b Hd
HH a
a H Hf
e f









Preparation of (2-d4) was accomplished by a similar

route, where isotopic labelling was introduced in the

reduction of the diacetic acid (37) with LiAID4. Bis-

dehydrotosylation afforded a 17% yield of (2-d4). The


CH CO2H CH2CD20H CH CD OTs
2 2 2 2 22 -CD
=K-> -->_ =0< -> =xV
O-i CA CD 01 2C 2
CH 2 CO2H CH2CD2H CH2CDOTs 2

(37) (38-d ) (39-d ) (2-d )

structure of (2-d ) was verified by nmr, ir, and mass

spectroscopy and by elemental analysis. Its nmr (see

Figure 15) revealed a triplet at 6 2.72 (4Ha Jab = 2.5

Hz), a pentet at 4.89 (2Hb, Jab = 2.5 Hz), and a multiple

at 5.85-6.10 (2H ); its ir (see Figure 16) exhibited bands

at 3090, 3010, 2970, 2935, 2320 (w), 22.30 (w), 1680, 1600,
-1
1420, 1035, 945, 880, 735 cm Mass spectral analysis

(15 ev) was used to establish its isotopic purity at 99.8%.

The thermal reorganization of (2-d ) to (3-d ) was

accomplished in a sealed tube, the sample diluted with

benzene. Gipc analysis on a 5' (18%) DC-200 column at 600

and 1100 indicated that only the expected (3-d4) appeared

to have formed. Both its structure and its isotopic purity

were confirmed by nmr, ir, mass (15 ev) spectroscopy, and

elemental analysis. Its nmr (see Figure 17) showed broad

singlets at 6 2.26 (2Ha), 2.82 (2Hb), 5.71 (1He), as well

as multiplets at 4.67-4.86 (2H ) and at 6.15-6.36 (1Hf);

its ir (see Figure 18) showed hands at 3080, 2990, 2930,








2860, 2190 (w), 2160 (w), 2100 (w), 1680, 1650, 1445,
-I
1070, 1030, 945, 890, 718 cm1.

H H
H c\ a c


D2 CD2
Hb" 115 CD2
H b
a H H
e f

2-) (3-d4)

The synthesis of the novel 3-methylene-l-vinyl-l-

(vinyl-2,2-d2 )cyclobutane (2-d ) was accomplished in six
b 2
steps from the 3-methylene-l,l-diacetic acid (37).

0
CH2CO2H / CH2CO2H


K00 22 2
CH CO H CH CO Et
b0
(37) (40) (41)



CH2CO2H CH2CDOH CH 2 CDOTs
-2 f


CH2CH2OH H CH2OH CHCH 2OTs

(42) (38-d2) (39-d ) (2-d )
-- -2 -2


Dehydration of the diacetic acid (37) with refluxing acetic

anhydride quantitatively produced the cyclic anhydride (40).

eating of (40) for 2 hr in the presence of absolute ethanol

cleanly produced the difunctional half-ester, half-acid

(41) in high yield. The key step, however, was the










Bouveault-Blanc reduction34 of the carboxylate function with

sodium metal in absolute ethanol and liquid NH3. At this

point, (42) could be reduced with LiAlD4 to introduce

deuteriums in the diol (38-d ) which was subsequently con-

verted to the ditosylate (39-d ). The ditosylate (39-d )

when treated with potassium-t-butoxide in a DMSO solvent

resulted in the product (2-d2) whose nmr, ir, and mass

spectra were entirely consistent with its structure. Its

nmr (see Figure 19) showed a triplet at 6 2.87 (4H, J =

2.5 Hz), and multiplets at 4.82-5.18 (3H), 5.21-5.38 (1H),

and at 5.92-6.12 (2H); its ir (see Figure 20) showed strong

bands at 3090, 3010, 2975, 2940, 2280 (w), 2130 (w), 1680,

920, 880 cm-1. Mass spectral analysis (6 ev) was used to

establish its exact isotopic isomer distribution: (2-d2),

0.861; (2-d ), 0.050; (2-d ), 0.089.

In an analogous fashion, methylene-3-vinylcyclobutane

(29) was prepared in an eight-step sequence from 3-methylene-

cyclobutanecarbonitrile (43). Both (43) and (44) are known

in the literature.31 Structures for the other compounds in

this sequence were verified by their nmr, ir, and mass

spectra. All were produced in high yield except the

dehydrotosylation step which afforded (29) in only 20%

yield. The structure of (29) was confirmed by its nmr, ir,

and mass spectra. Elemental analysis indicated its high

purity. The nmr (see Figure 23) showed multiplets at 6

2.33-3.12 (511a), 4.65-4.95 (2Hb + 2Hc), 4.97-5.18 (lHd),








and 5.80-6.40 (1H ); its ir (see Figure 24) had bands at

3090, 2950, 1680, 1645, 1420, 990, 920, 915, 880 cm1.


- > CO2H


(44)


=CH2CO2H<- CH2CN


(47)


CH2OH



(45)




--- CH2OTs


(46)


=- 2CH2 OCi H O -> = C>-CH2CH2OTs --


(49) (50) (29)

In addition, the structure of (29) was confirmed by its

smooth thermoreorganization to the known 4-methylenecyclo-

hexene (30) whose spectral characteristics were nearly


c
2100 a
2100 a
H


a
x


(29) (30)


- 3CN


(43)


(48)









33
identical to those reported in the literature. For

reference, the nmr (see Figure 25) for (30) showed broad

singlets at 6 2.21 (4Ha), 2.72 (211b), 4.68 (2H ), and

5.62 (2Hx).

Activation energy parameters for the thermolysis of

(2) and (29) were determined. The rate of disappearance

of (2) in an n-decane solution with respect to n-nonane

as an internal standard was followed on a Hewlett-Packard

model 5710A gas chromatograph with flame ionization

detector. The gas chromatographic analyses were performed

on a 7.5' (2%) DC-200 column at 560. The chemical

composition of the kinetic samples was determined with a

Vidar Autolab 6300 digital integrator to facilitate and

minimize the error in determining relative peak areas. A

first-order, least squares program was used in conjunction

with a pdp 8/e digital computer to analyze all the data on

a concentration versus time plot as a correlation coefficient

where 1.000 is the optimum fit for this unimolecular

reaction. Rate constants at seven different temperatures

in the range of 86.10 to 121.10 were determined and com-

piled in Table 1. An Arrhenius plot was constructed from

the data in terms of providing Arrhenius parameters. For

(2), the appropriate parameters were determined: energy

of activation, 27.270.30 kcal/mol; log A, 11.87+0.17;

enthalpy of activation, 26.530.30 kcal/mol at 103.20;

entropy of activation, -6.660.78 cal/mol-deg at 103.20;









free energy of activation, 29.0310.78 kcal/mol at 103.2.

The rate of disappearance of (29) in an n-decane

solvent was followed also by the use of a Hewlett-Packard

model 5710A gas chromatograph with flame ionization

detector but without the need for an internal standard.

In every other respect, the analysis, the treatment of the

data, and the determination of the activation parameters

for the thermoreorganization of (29) were identical with

that for the thermolysis of (2). Rate constants at six

different temperatures in the range of 175.0 to 211.00

were determined and compiled in Table 6. An Arrhenius

plot was constructed in Figure 5. Computer evaluation

of the data provided the following activation parameters:

energy of activation, 35.670.41 kcal/mol; log A, 12.70

0.19; enthalpy of activation, 34.740.41 kcal/mol at 192.30;

entropy of activation, -3.290.87 cal/mol-deg at 192.30;

free energy of activation, 36.280.87 kcal/mol at 192.30.

These results indicate that the addition of a second

vinyl group in the ring system lowers the activation

energy necessary for rearrangement by about 8 kcal/mol.

The intermolecular isotope effect (k /kD) was obtained

from the ratio of rate constants of (2) and (2-d4). The
414
analysis and treatment of the data were as previously

described and detailed in the experimental section. At a

particular temperature, the kinetic pyrolysis of (2-d4)

was performed immediately after the kinetic pyrolysis of

(2). At 99.90, the intermolecular isotope effect (k /kD)








was 1.050.03 and, at 121.10, the (kH/kD) ratio was

1.040.03. Within experimental error, the results indicate

that the intermolecular isotope effect was not temperature

dependent. The rate ratio was recalculated at each

temperature by adding the standard error associated with

the rate constant for (2) and by subtracting the standard

error associated with the rate constant for (2-d4). The

deviation of this rate ratio from that determined above

is quoted as the error in the intermolecular isotope effect.

An intramolecular isotope effect (k /kD = 3a/3b =

1.140.07) was obtained by pyrolysis of a dilute benzene

solution of (2-d2) at 121.10 for 5 hr. A sample of (2)

was subjected to the same conditions and used as a control.



SCH22 DD
CD2
D

(2-d2) (3a) (3b)


Determination of the intramolecular isotope effect was based

upon nmr integration ratios of the purified mixture of (3a)

and (3b). Two separate determinations were possible from a

100 MHz nmr spectrum (see Figures 21 and 22). One

involved the continuous-sweep integration of the aliphatic

region with the determination of the ratio of the allylic

H/diallylic H or H /Hb. The other involved the continuous-

sweep integration of the vinylic region with the









determination of the ratio of the terminal-vinyl H/

vinyl-ring H or (H + Hd)/H The paucity of the sample

only allowed the determination of the ratio of (3a/3b)

present from a single nmr sample. As a result, the reported

error was maximized. The nmr integration ratio for each

region aliphaticc or vinylic) was recalculated by adding

the standard error associated with the nmr integration

ratio and also by subtracting the standard error associated

with the same integration ratio. Thus, from the high, low,

and mean values of the nmr integration ratio, three intra-

molecular isotope effect values were determined for each

region. The average and standard deviation of each set of

three values is the reported isotope effect for each region.

Thus, two values for the intramolecular isotope effect,

one for each region, were determined. Their average is

the reported as the final value for the intramolecular

isotope effect.



H H
c c H H H H
c c c c

H H H dc H H
a b a a Hb CD2


x H d D
H H a
a H H H H H
x e f H H


(3-d4)


(30)





28


Assignment of the protons in (3) merits some

attention. Protons Ha at 6 2.30 are monoallylic and

relative to the other hydrogens should be upfield. For

the reported 4-methylenecyclohexene (30),33 the monoallylic

hydrogens are found at 6 2.21 (lit. 2.1). Moreover, the

integral signal at 6 2.26 for the nmr of (3-d4) decreased

by one-half, as expected when half the monoallylic sites

are deuterated. The protons Hb at 6 2.85 are diallylic and

somewhat downfield from H a. When half the monoallylic

positions are deuterated as in (3-d4), the nmr integral

signal for the diallylic protons should and does increase

twofold relative to Ha. Thus, 6 2.85 was assigned to Hb.

The protons H and Hd at 6 4.55-5.20 are not readily

distinguishable from one another in the sense that their

integration values cannot be separated. Intuition tells

us that the chemical shifts for these hydrogens should vary

little from those same kinds of hydrogens in (2). Moreover,

the nmr spectrum for (30) reveals a broad singlet at 6 4.68

(lit. 4.7) for the exocyclic vinylic CH2, and the nmr

spectrum for (3-d4), where no terminal vinylic CH2 is

present, shows an unresolved multiple at 6 4.67-4.86 for

its exocyclic vinylic CH2. Comparative inspection with the

nmr of (2) reveals that the terminal vinylic CH2 was always

slightly downfield from the exocyclic vinylic CH2.

Similarities in the nmr of (2) and (3-d4) suggest that the

AB quartet in (3) be assigned to 6 6.28 as Hf. By default

the signal at 6 5.52-5.85 was assigned to He
e














DISCUSSION


For the thermal reorganization of i,l-divinyl-3-

methylenecyclobutane (2) to 4-methylene-2-vinylcyclo-

hexene (3) and methylene-3-vinylcyclobutane (29) to

4-methylenecyclohexene (30), three distinct mechanistic

pathways can be envisioned: (a) a concerted [l,3]-sigma-

tropic rearrangement, (b) a concerted [3,3]-sigmatropic

or Cope rearrangement, and (c) a multistep process

involving diradical intermediates.

Of the three alternatives, pathway (a) seems least

likely. There are but few examples of the concerted [1,3]-

sigmatropic process. The thermal and stereospecific trans-

formation of endo-5--methylbicyclo[2.1.1]hex-2-ene (and

related derivatives) with inversion of configuration to

exo-6-methylbicyclo[3.1.0]hex-2-ene suggests a concerted

suprafacial 1,3 alkyl shift.35 More often cited is the

work by Berson and coworkers on the thermal conversion of

exo-7-deuterio-endo,cis-bicyclo[3.2.0]hept-2-enyl-6-acetate

to exo-3-deuterio-exo-5-norbornenyl-2-acetate which also

proceeds with inversion of configuration at the migrating

center.36 However, both cases appear to exemplify the need

for two geometric requisites: (a) relief of ring strain,

and (b) a n-bond center held rigidly near to the migrating









a-bond to allow the occurrence of the suprafacial 1,3 shift.

The latter can be achieved by ring fusion, and yet it is

absent in the much studied the'ral reorganization of vinyl-

cyclopropane, despite its considerable ring strain energy

(27.6 kcal/mol),37 to cyclopentene. The observed activation

energy of 49.7 kcal/mol38 can be reasonably accounted for by

a diradical intermediate with a full complement of allylic

stabilization energy. Experimental evidence obtained in a

study of specifically deuterium labelled vinylcyclopropanes

supports the concept of a diradical pathway for this
12b
rearrangement. Rate and stereochemical product studies

on trans-1,2-dialkenylcyclobutanes,1 39 which are

structurally more related to the nature of our research

interests here, also suggested a lack of concert in their

ring expansion to vinylcyclohexenes. The lack of a

geometric lock on the z-bond center adjacent to the

potentially migrating o-bond may be the significant factor

responsible for the lack of concert in these nonfused

systems. The same feature is also lacking in the analogous

methylene-3-vinylcyclobutanes (2) and (29), and with the

preponderant lack of analogy no further comment on the

concerted [1,3]-sigmatropic shift seems appropriate. The

mechanistic picture then for the thermal rearrangement of

(2) and (29) is somewhat clarified, and attention is thereby

focused on the two remaining pathways (b) and (c).

As noted earlier, in the Cope rearrangement either a

four-centered chair (9) or a six-centered boat (10)








conformation is attained in its cyclic transition state. It

appears essential that for the thermally activated concerted

reaction of 1,5-hexadienes, considerable overlap must be

developed between the r-orbitals at C-l and C-6 as the 3,4

o-bond breaks. For those Cope reactions that are irrevers-

ible and exothermic, the transition state should resemble
22
the reactant more than the product geometry.22 Hence,

comparison of ground state models of acyclic 1,5-hexadienes

with the related methylene-3-vinylcyclobutanes (2) and (29)

is instructive. Unquestionably, the ground state boat (52)

and the ground state chair (53) conformations of the acyclic

1,5-hexadienes can easily achieve substantial orbital over-

lap between C-1 and C-6 as the 3,4 a-bond breaks along the

0

< ^ U(53)

(52)




/ (2) R = CH=CH2

(29) R = H



reaction coordinate. However, inspection of the geometric

orientation of the i-bond centers in the methylene-3-vinyl-

cyclobutanes (2) and (29) reveals that an unbearably high

angle strain would be required in the transition state for

development of sufficient 1,6 orbital overlap for concert









in their thermal transformations. Indeed, the alternative

involves strain free diradical intermediates which appear

much more attractive.

One of the well-established features of the Cope

reaction is the characteristically high negative entropy

of activation which reflects the well-ordered cyclic

multicentered transition state for the rearrangement.

Values more or less center around -10 eu to -14 eu for four-

centered reactions2 and appear to be even more negative for

six-centered reactions. Since entropy of activation AS

and frequency factor A are related by the equation, AS =
40
Rln(Ah/ekT),40 Benson and O'Neal have indicated that the

vast majority of four-centered Cope reactions fall within
13.5+1.0 -1
the narrow A factor range of 10 sec and for six-
11.5+1.5 -1
centered reactions a range of 10 sec has been
41 14.0+1.0 -1
ascertained. High A factors (10 sec ) and

positive or nearly zero AS values are generally

anticipated for diradical reactions, since few steric

restrictions are imposed on their transition states.42

Moreover, when diradicals are borne through ring cleavage,

new rotational and torsional modes are generated, and an
*
entropy gain is expected. Albeit rather low AS values

and low A factors were observed in the pyrolysis of

1,l-divinyl-3-methylenecyclobutane (2) to 4-methylene-2-

vinylcyclohexene (3) and methylene-3-vinylcyclobutane (29)

to 4-methylenecyclohexene (30) for the former AS =

-6.7 eu and A = 1011.9 sec and for the latter









12 7 -1
AS = -3.3 eu and A = 10 sec This may seem

surprising. How then can these low values be rationalized

in terms of a multistep process involving diradical

intermediates? It must be noted that in the reactants (2)

and (29) the vinyl groups can undergo internal rotations,

whereas if a biradical resulted the two conjugated groups

cannot. For example, a biradical pathway has been

unambiguously established for the well-known vinylcyclo-
12
propane to cyclopentene rearrangement.1 One of the

steric requirements for allylic stabilization is that all

seven atoms (three carbons and four hydrogens) should be

in the same plane. Compared to the ground state, there has

been considerable reduction in the internal rotations

of the vinyl group. Correspondingly lower values
13.5 -1 38
(AS = -0.3 eu, A = 10 sec ) were observed for the

pyrolysis of vinylcyclopropane over that for either the

isomerization of cis-1,2-dideuteriocyclopropane (AS
16.4 -1 43
12.8 eu, A = 10 4 sec- ) or the decomposition of

1-ethylcyclopropane to 1-pentene (AS = 3.6 eu, A
14.4 -1 44
1044 sec ), depending upon which model serves better

for comparison. Similarly, rearrangement of vinylcyclo-
*
butane to cyclohexene should occur with a lower AS and a

lower A factor over that for either the isomerization of

cis-1,2-dideuteriocyclobutane or the decomposition of

ethylcyclobutane. Unfortunately, data are not available

for the former two pyrolyses. However, the isopropenyl-

cyclobutane to 1-methylcyclohexene rearrangement






34


14 5 -1 45
(AS = 4.6 eu, A = 10 1 sec ) can be conveniently

compared to the decomposition of isopropylcyclobutane to

ethylene and 3-methyl-l-butene (AS = 9.3 eu, A =

1015.6 sec 1),46 provided that the rate-determining step

of the latter involves ring cleavage in diradical
*
formation. The reduction in AS and A factor values with

the formation of allyl radicals has been noted by Frey in

his review on the pyrolyses of small ring compounds from

which other similar correlations can be made.11

Our kinetic results fit rather nicely into the above

picture, if diradical intermediates intervene in these

thermal reorganizations. The low AS (-3.3 eu) and low A
12.7 -1
factor (10127 sec ) for (29) can thus be derived from

the loss of internal rotations of the vinyl group and the

subsequent formation of two planar allyl radicals where each

restricts three carbons and four hydrogens to a rigid plane.

This compares remarkably well with the pyrolyses of endo-

and exo-2-vinylbicyclo[2.2.2]oct-5-ene (20 and 21

respectively). As previously cited, both epimers in all

probability undergo transformation to cis-Al6 -hexalin (22)

through diradical intermediates. In each case, two planar

allyl radicals are formed, and an expectedly lower AS
12.8 -1 13
(-3.3 eu) and A factor (10 sec ) result.

In similar fashion, the formation of a pentadienyl

radical restricts five carbons and six hydrogens to a

single plane, so that the pyrolysis of









12 6
1,1-divinylcyclopropane (1) (AS = -4.1 eu, A = 10

sec-1 solution phase) 47 and of 1,1-divinyl-3-methylene-
011.9 -1
cyclobutane (2) (AS = -6.7 eu, A = 10 sec

solution phase) have expectedly lower values.
*
In conclusion, characteristically large AS values

were not observed for either the pyrolysis of 1,1-divinyl-

3-methylenecyclobutane (2) or methylene-3-vinylcyclobutane

(29). It is, therefore, difficult to rationalize any

element of concert in these reactions when a diradical

pathway can accommodate the data better.

The question also arises as to whether the observed

activation energies for the thermal reorganizations of (2)

and (29) are compatible with diradical pathways. Comparison

of the differences in the activation energies of suitably

chosen models of known diradical processes can be employed

to estimate the activation energies for the pyrolyses in

question. It is noteworthy to point out that since this

activation energy estimate is derived from known diradical

reactions, it follows that the estimate is also one for a

diradical process involving the same type of rate-

determining step. Corroboration of the estimate with

experimental observation would thus support the concept of a

diradical mechanism. For example, a diradical process has

been established for the rearrangement of vinylcyclopropane

to cyclopentene.12b The isomerization of cis-1,2-dideuterio-
43
cyclopropane requires about 65 kcal/mol. If full use is

made of the allylic delocalization energy









(estimated at 12-21 kcal/mol),48 the reaction would require

at most 53 kcal/mol (65 12). Within the uncertainty

limits of the allyl delocalizction energy, the estimated

and observed value (about 50 kcal/mol)38 are in good

agreement. Since other chemical data support the concept of

a diradical pathway for this rearrangement, it follows that

the total contribution of the allyl group to a reduction in

the activation energy amounts here to 15 kcal/mol (65 50).

Adapting this above procedure, an activation energy

estimate for the pyrolysis of methylene-3-vinylcyclobutane

(29) requires a comparison of the isomerization of cis-1,2-

dideuteriocyclobutane with the degenerate rearrangement of

methylenecyclobutane and also with the ring expansion of

vinylcyclobutane. Unfortunately, only the data on the

degenerate rearrangement of methylenecyclobutane (Ea =
49
49.5 kcal/moi) are available.49 The isomerization of

cis-l,2-dimethylcyclobutane requires 61.3 kcal/mol. By

their inspection of the data compiled on the thermal uni-

molecular reactions, Willcott, Cargill, and Sears conclude

that each alkyl substituent reduces the activation energy
42
for the isomerization by about 3 kcal/mol, so that as a

working model we can infer an activation energy for the

isomerization of cis-1,2-dideutericcyclobutane at 67.3

kcal/mol (61.3 + 3 + 3). Thus, incorporation of an

exocyclic double bond into the cyclobutane system imparts

17.8 kcal/mol (67.3 49.5) reduction in the activation

energy. The thermoreorganization of isopropenylcyclobutane









(Ea = 51.0 kcal/mol) offers the closest analogy to the

unknown activation parameters for the pyrolysis of vinyl-

cyclobutane. Comparison of this value with the activation

energy estimate of 67.3 kcal/mol for the isomerization of

cis-1,2-dideuteriocyclobutane might be used to ascertain

the extent of the assistance of the vinyl group in the

transition state, determined here at 16.3 kcal/mol (67.3 -

51.0). Combining the contributions from both moieties in

the transition state, places the total stabilization

assistance at 34.1 kcal/mol (17.8 + 16.3). When this value

is subtracted from the estimated activation energy for the

parent cis-l,2-dideuteriocyclobutane, an estimate of 33.2

kcal/mol (67.3 34.1) results which, when considering the

number of assumptions made, is in remarkable agreement with

experimental observation (35.7 kcal/mol). Since the models

used to arrive at this estimate all involve diradical

processes with rate-determining diradical formation, it

follows that our experimental finding is in very good accord

with a multistep process involving rate-determining ring

cleavage.

An estimate for the pyrolysis of l,l-divinyl-3-

methylenecyclobutane (2) also requires a comparison of the

isomerization of cis-l,2-dideuteriocyclobutane (Ea at about

67.3 kcal/mol) with the degenerate methylenecyclobutane
49
rearrangement (E = 49.5 kcal/mol). However, knowledge

of the contribution by a pentadienyl group to the reduction

of the activation energy for the thermolysis is necessary.









By subtraction of the activation energy for the

isomerization of cis-1,2-dideuteriocyclopropane (Ea = 65.1

kcal/mol)43 with the activation energy for the thermal

rearrangement of 1,1-divinylcyclopropane (1) (Ea = 42.5

kcal/mol),51 22.6 kcal/mol assistance results. The two

moieties contribute a total therefore of 40.4 kcal/mol

(17.8 + 22.6) assistance in lowering the activation energy.

Thus, an estimate of 26.9 kcal/mol (67.3 40.4) is in

excellent agreement with the observed activation energy of

27.3 kcal/mol. Again the estimate was derived from

activation energies of known diradical processes, involving

rate-determining diradical formation and supports the

concept of a multistep process with rate-determining ring

cleavage. Moreover, from our data and by comparison with

the degenerate methylenecyclobutane rearrangement, and

independent estimate for the total contribution of a

pentadienyl group in the transition state can be placed at

22.2 kcal/mol (49.5 27.3) which lends weight to the

heretofore sole determination of this value as ascertained

from the pyrolysis of (1).

In order to construct an energy diagram consistent with

a diradical pathway, it must be demonstrated that the heat

of formation AH for the biradical, formed in the thermo-

reorganization of either (2) or (29), is sufficiently below

the corresponding heat of formation of the transition state

AH By employing the group equivalent method, devised by

Franklin52 and most recently modernized by Benson,37








the AHO for methylene-3-vinylcyclobutane (29) has been

calculated at 48.7 kcal/mol. The pyrolysis of (29) required

a 35.7 kcal/mol activation energy Ea, so that the heat of

formation of the rate-determining transition state AHO
*
(where AH = E + AH ) is 84.4 kcal/mol. The strength of
f a f a
the bond involved in this homolysis is weakened by its

inclusion in the highly strained methylenecyclobutane
53
(EB = 28.2 kcal/mol) and by the involvement of two
strain
resulting allyl resonance-stabilized groups. The strength

of this bond can thus be assessed by subtracting out these

contributions from the bond dissociation energy BDE of a

suitable model. Isopentane may be just such a model in

that it is a trialkyl-substituted ethane as is the weakest

bond in (29). The BDE for isopentane is quoted at 80

kcal/mol.37 Since isopentane already includes the effect of

substituents in reducing the strength of the C-C bond, one

need only employ the allylic delocalization energy ADE
48
(estimated at 12-21 kcal/mol) to correct for the contri-

bution of the allyl groups. Thus, the strength of the bond

involved in this homolysis is at most 27.8 kcal/mol

(80 28.2 2 x 12). In other words, the formation of

this biradical requires 27.8 kcal/mol more than the AH

for (29), so that the AHo for the biradical is at most

76.5 kcal/mol (27.8 + 48.7). An alternative calculation

for the AH1 for the (allyl)2CH2 biradical at 75.8 kcal/mol

by the group equivalent method37 54 compares favorably

with that above. The higher of these two values is still









well below the AHO and very consistent with a biradical

mechanism for the thermolysis of (29). An energy diagram

has thus been constructed in Pigure 1.

A heat of formation AH for 1,l-divinyl-3-methylene-

cyclobutane (2) was evaluated at 67.8 kcal/mol. It follows

that the AHl is 95.1 kcal/mol (67.8 + 27.3). The strength

of the bond involved in this homolysis can be compared to
37
neopentane (BDE = 77 kcal/mol)," provided that corrections

are made for its inclusion in the highly strained

methylenecyclobutane and for the involvement of the result-

ing allyl and pentadienyl resonance-stabilized groups, both

factors which weaken the bond in (2). The ADE has been

previously assessed at a minimum value of 12 kcal/mol.48

From the difference in activation energies of the pyrolysis

of (1) (E = 42.5 kcal/mol) and the decomposition of

1,1-diethylcyclopropane to 3-ethyl-2-pentene (Ea = 63.4

kcal/mol),54 an evaluation of the pentadienyl delocali-

zation energy PDE at 20.9 kcal/mol results. Thus, the

strength of the bond involved in this homolysis is at most

16.1 kcal/mol (77 28.2 12 20.9), and the AH0 for the

allyl-CH2-pentadienyl biradical is at most 83.9 kcal/mol

(16.1 67.8). By employing the group equivalent method,

AH for the allyl-CH2-pentadienyl biradical was calculated

at 89.2 kcal/mol37 54 and compares favorably with the

alternate value above. The higher of these AH estimations

for the biradical is still approximately 6 kcal/mol below





41



the corresponding AH and is quite consistent only with a

biradical pathway for the rearrangement of (2). An

appropriate energy diagram for the pyrolysis of (2) has

been constructed in Figure 2.

Lastly in an effort to promote further understanding of

the mechanism in question here and to emphasize the

effectiveness of secondary deuterium isotope effects as a

tool for probing the nature of the transition states in

thermoreorganizations, a secondary deuterium isotope effect

study was accomplished in the thermal transformation of

l,l-divinyl-3-methylenecyclobutane (2).

For contrast recall that an inverse intermolecular

secondary deuterium isotope effect (k,/kD = 0.940.02)

was measured for the Cope reaction of the substituted

1,5-hexadiene (24).21 As previously stated, with the


Et Me NC t
NC Me
NC H(D) NC H(D)

"- H(D)
H (D)
(24) or (24-d ) (25) or (25-d )

development of a cyclic multicentered transition state, the
2 3
sp carbon at C-6 becomes more sp -like, so that the steric

requirement of H attached at the more congested tetra-

hedral orientation of the sp3 position, compelled (24) to

rearrange to (25) more reluctantly than (24-d ) to (25-d ).




















84.4c
-d

75.8b

35.7a

48.7b

(29)



21.4b

(30)



Reaction coordinate





a Observed E
a
b Calculated AHu by the group equivalent method

c AH = E + AHo
d Speculation
d Speculation


Figure 1: Energy diagram for the pyrolysis of (29)






43









-95. l
$


1 89.2b
E 27.3a
67. 8b/

(2)




-- 34.7b

(3)






Reaction coordinate




a Observed E
a
b Calculated AH by the group equivalent method

f
d AHS = Ela AH
d Speculation


Figure 2: Energy diagram for the pyrolysis of (2)





44


Observation of normal secondary kinetic deuterium

isotope effects in the thermoreorganization of 1,1-divinyl-

3-methylenecyclobutane (2) initially prejudiced our

inclination to envision the mode for rearrangement as one

which involved the intermediacy of diradicals. An inter-

molecular isotope effect (k /kD = 1.050.03) was determined

by a competitive rate study of (2) with (2-d4).





0 CD2 CD
CD 2 D-
D

(2) (3) (2-d ) (3-d )

Moreover, a normal intramolecular isotope effect (kH/kD =

1.140.07) was also observed. It was arrived at by

determining the product ratio (3a/3b) in the thermolysis

of (2-d ).


CD2 CD2 + D
2 D

(2-d ) (3a) (3b)

Insight into the significance of a comparison of inter-

and intramolecular isotope effects was first gained in a

study of allene cycloadditions. Specifically, Dolbier and

Dai examined [2 + 2]-cycloadditions of allene with acrylo-

nitrile (54) and placed it in perspective with other









allene cycloadditions.6 A comparison of rates of cyclo-

addition of allene versus allene-d4 with (54) produced a

normal intermolecular isotope effect at 1.040.02 and

established the nonsynchronous formation of the cycloadduct.

Moreover, if both new sigma bonds were formed in a concerted

reaction, the intramolecular isotope effect would also be

inverse, since in this case the rate-determining step is

also product-forming. Employing 1,l-dideuterioallene

to introduce isotopic discrimination, enabled them to

observe product-forming destruction of the resonance-

stabilized allyl radical. The ratio of (55a/55b) of

1.170.04 defined the magnitude of the intramolecular

isotope effect. The nonidentity of the inter- and intra-

molecular isotope effects thus indicated that rate-

determining and product-forming steps cannot be one and

the same, so that a multistep diradical process must be

operative. The normal intramolecular isotope effect was

initially considered anomalous. For cycliztion there

apparently is a small but significant activation energy,

the bulk of which the authors ascribed to the rotation of a

terminal methylene group from its planar allyl configu-

ration to an orthogonal one. At that time it was thought

that the possible source of this normal intramolecular

isotope effect might be the relief of nonbonded and

torsional interactions which would be found in the planar

allyl but not in the nonplanar allyl system. Thus, the

unlabelled methylene group would be more apt to rotate to








achieve ronplanarity. More recently, the source of this

isotope effect was conceived in a different but supportive

light and will be detailed later. To complete the picture,

CH =CHCN
2 D D
(54) CD2 2

+ D2CN +

CH =C=CD2 H2C CN

planar (55a) (55b)

[2 + 2]-cycloaddition was complicated by a concomitant
49
degenerate methylenecyclobutane rearrangement.4 When

carried to complete equilibration, a predictably inverse

isotope effect (kH/kD = 0.9210.01) was observed and thus

was not responsible for the observed kinetic effect.

In dimerizations of parent allene, 1,2-cyclonona-

diene,56 and 1,2-cyclohexadiene,5 normal intramolecular

isotope effects were also observed in every case. For the

former two dimerizations, normal intermolecular isotope

effects were determined and found to be less than the

intramolecular isotope effects. Again nonplanar radicals

had to be invoked to rationalize the nonidentity of the two

isotope effects, and the fact that they are normal.

The utility of a secondary deuterium isotope effect

study was additionally demonstrated in the thermal

rearrangement of dideuteriobiscyclopropylidene (56) to the

deuterated methylenespiropentanes (57a, 57b, and 57c).26b









Two distinct isotope effects were operative in this system,

defined according to the scheme depicted below, where

(57bb/57c) = k/k) lizat = 1.140.02 and where
H D cyclization
(57a/57b + 57c) = (k/kD)cleavage 1.240.03.

D D




k 2 k'
(57b)





D
D

Although one possible concerted or "pivot" type mechanism

for this rearrangement cannot be rigorously excluded, the

observed isotope effects can be completely rationalized

by invoking a diradical pathway. It was believed that the

large normal isotope effect for ring cleavage stemmed from

the generation of a transition state which is "looser" or

"less rigid" than the reactant and where on proceeding to

the transition state there is a reduction of the rotational

barrier or torsional force constant for the CH2 twisting

motion as the C-C ring bond is weakened. Moreover, the

authors point out that the normal isotope effect for

cyclization is consistent with the intramolecular isotope

effect associated with rotation of a terminal methylene

group to form a nonplanar allyl radical followed by simple

radical combination to product. The magnitude of this





48

isotope effect is nearly identical to other intramolecular

isotope effects reported for allene cycloadditions.56

A normal intramolecular isotope effect (kH/kD =

3a/3b = 1.14+0.07) was also observed in the pyrolysis of

(2-d2). The unique feature of this kind of isotope effect



H2 L >
-OC^~~-- A- ^ J


(2-d2) planar nonplanar




CD 2 + D )
D
(3a) (3b)


is that it allows one to observe isotopic discrimination

after the formation of a planar diradical intermediate.

The main energy event in this discrimination is thought

to involve a twisting motion of an isotopically substituted,

terminal methylene group, resulting in a nonplanar

pentadienyl radical with concomitant destruction of

7-bonding at its terminus. This twisting implies that there

is a drastic lowering of the force constant associated with

w-bonding which, in the limiting case, where the CX2

(X2 = H2 or D2) group twists a full 90, falls to zero.
Conceivably then a single potential energy surface

approximates the discrimination process where the CH2 group,









because of its relatively higher zero point energy, has a

head start over the CD2 group in surmounting this

rotational energy barrier. Hence, H preferentially occurs

within the cyclohexene ring, and significantly more of

(3a) is formed. This recent concept supports an earlier

rationalization where relief of nonbonded and torsional

interactions, which would be found in the planar but not

in the nonplanar conjugated radicals, was used to account

for normal intramolecular isotope effects. Despite the

large error associated with the intramolecular isotope

effect determination, the value is still substantially

greater than and different from the normal intermolecular

isotope effect. This then is significant. The intra-

molecular isotope effect must be associated with product

formation. Ring cleavage or biradical formation must be

associated then with the intermolecular isotope effect,

since by definition it is involved solely with the rate-

determining step. These conclusions are consistent with

the observed activation energy parameters previously

examined. Moreover, the mean values for the intermolecular

(1.08) and the intramolecular (1.12) isotope effects

associated with the 1,1-divinylcyclopropane systems also

indicated, albeit narrowly, that the isotope effects may

indeed be appreciably different enough from one another.

Thus, the same kind of conclusions in the rearrangement of

1,1-divinylcyclopropane (1) to vinylcyclopentene (28) may

be drawn.









It appears then that the normal intermolecular isotope

effect for l,l-divinyl-3-methylenecyclobutane systems must

be derived from a rate-determining ring cleavage step. In

that event, there should result a decrease in the torsional

force constant or in the rotational barrier for the CX2

twisting motion as the C-C ring bond weakens. The same

explanation was used to rationalize the normal ring

cleavage isotope effect in the pyrolysis of dideuterio-
26b
biscyclopropylidene (57).b Moreover, biradical formation

requires the transformation of the isotopically labelled
2 2
CX2 site from sp2 to sp -radical on proceeding toward the

transition state. This bears some analogy to Pryor's

work on the addition of radicals to styrene where the same

kind of transformation occurs in a rate-determining step

and where a small but normal intermolecular isotope
27
effect resulted.27 Thus, both theoretical arguments and

experimental analogy support the concept that (2) could

in fact undergo rate-determining biradical formation at a

faster rate than (2-d ).

With the knowledge that diradicals intervene in these

thermoreorganizations, particularly of (2), where a

secondary deuterium isotope effect study conclusively

demonstrated the mechanistic pathway for rearrangement,

the question arises as to what extent ring closure to

product occurs by way of 1,3 versus 3,3 radical

combination. To begin to investigate this additional








mechanistic complexity would necessitate an examination of

the thermolyses of two novel isotopic isomers (2a) and (2b).

H H DD


D C= H2C


(2a) (2b)

Two factors complicate an evaluation of the partitioning

between these two kinds of ring closure; (a) a competing
49
degenerate methylenecyclobutane type rearrangement,4

favoring the accumulation of D at the more crowded ring

methylene site or (2b), and (b) with the formation of a

planar allyl radical, its terminal CH2 group would rotate

out of conjugation preferentially over its CD2 group in

the product-making step. The former factor emphasizes the

need to study the thermoreorganizations of both isotopic

isomers. Needless to say, the entire evaluation becomes a

very complex one and will not be delved into further here.

Systematically evidence for distinguishing the

mechanistic mode for the thermoreorganizations of the

methylene-3-vinylcyclobutanes (2) and (29) has been

presented. Examination of molecular models suggested that

concert in these reactions required an unbearably high

angle strain for development of cyclic multicentered

transition states, as opposed to strain free diradical
*
intermediates. Very low AS and A factor values

characteristic of the Cope reaction were not observed.

Instead the low AS and A factor values could very









reasonably be accounted for by the loss of internal

rotations of the vinyl groupss, and the subsequent

restriction of a number of atoms to two rigid planar

conjugated radicals. Estimations of the activation

energy for a diradical process from model systems agreed

remarkably well with experimental observations. No

evidence of concert was uncovered from a comparison of

heat of formation calculations. The heats of formation

of the transition state for a diradical process were

approximated by the group equivalent method, and these

were consistent with the lower heats of formation,

calculated for the speculated diradical intermediates.

Moreover, the nonidentity of the two secondary deuterium

isotope effects for the 1,l-divinyl-3-methylenecyclobutane

systems cannot be attributed to a common transition

state for the rate-determining and the product-forming

steps. In addition, the preponderant weight of analogy

indicates that transformation of an isotopically substituted

carbon from a trigonal to a tetrahedral orientation should

be associated with an inverse secondary deuterium effect.

The normal intermolecular and intramolecular deuterium

isotope effects can be better accounted for by invoking

diradical intermediates. The intramolecular isotope effect

has allowed observation of the discrimination process after

formation of a planar diradical intermediate and is thus

associated with the product-forming step. The inter-

molecular isotope effect has allowed observation of the






53


discrimination process in the rate-determining step

and must be associated with ring cleavage resulting in

biradical formation. Thus, Valuable corroborative

insight into the nature of the transition states for

bond-breaking and bond-making has been gained from the

use of secondary deuterium isotope effects as a

mechanistic probe.















EXPERIMENTAL


Both boiling and melting points were uncorrected, the

latter taken on a Thomas-Hoover melting point apparatus.

Infrared spectral data were obtained from either a Perkin-

Elmer model 137 or from a Beckman model IR-10 spectro-
-1
photometer, and all absorption bands were listed in cm .

The ultraviolet spectrum was recorded on a Cary model 15

spectrometer. Nuclear magnetic resonance spectra were

obtained from a Varian model A-60A unless specified as a

XL-100 model. All nmr spectra utilize TMS as an internal

standard. Mass spectral data were obtained from an

Hitachi Perkin-Elmer RMU-6E spectrometer.

Elemental analyses were determined by Atlantic

Microlab, Inc., Atlanta, Georgia.

The glpc analyses and separations were carried out on

a Varian Aerograph model A-90P gas chromatograph, equipped

with the column listed in the text.

Glpc kinetic analyses were performed on a Hewlett-

Packard model 5710A gas chromatograph with flame ionization

detector. The chemical composition of the kinetic samples

was determined with a Vidar Autolab 6300 digital integrator.

First-order, least squares programs were used in conjunction

with a pdp 8/e digital computer.









A Lauda constant tcimcrature circulator type N/S15/12,

equipped with an R-10 electronic relay, an R-20 electronic

controller, and an R-30 temperature and level protection

relay, was used for all pyrolyses.

All reagents which are not referenced were

commercially available.



Diethylmethylenemalonate (31)

This compound was prepared by a procedure modified
58
after that of Bachman and Tanner. In a typical procedure,

a three-necked 2 1. flask, fitted with an overhead stirrer,

a thermometer, and a West condenser, was charged with 400 g

(2.50 mol) of diethylmalonate, 150 g (5.00 mol) of para-

formaldehyde, 25 g of Cu(OAc)2, 25 g of KOAc, and 1000 g

of HOAc. The contents were heated at reflux for 60 min.

Immediately after reflux, the condenser was readjusted and

adapted for distillation at atmospheric pressure in which

all the HOAc was distilled in about 2 hr. The remainder

was cooled at once, treated with an equal volume of water,

and extracted with 2 1. of ether. The ether extract was

washed well with water and dried over Na2SO4 overnight.

After removal of ether by rotary evaporation, the product

was distilled at ca. 105-130/10 mm (lit. 2100/760 mm),

yielding 124 g (29% based on the amount of diethylmalonate

used). Nmr (CC14): 6 1.28 (t, 6H, J = 7 Hz), 4.22 (q,

4H, J = 7 Hz), 6.37 (s, 2H).









Diethyl-3-methylene- 1, 1-cyclobutanedicarboxyla te (32)

A 1.4 1. stainless steel rocker bomb was charged with

353 g (2.05 mol) of diethylmethylenemalonate (31), 41 g

(1.03 mol) of allene, 150 ml of toluene, and 2 g of hydro-

quinone. The bomb was heated to 2200 for 18 hr with

agitation and under autogenous pressure. After cooling

to room temperature, the bomb was opened and the contents

distilled at 148-1530/53 mm, yielding 120 g (55%). Ir

(neat, NaCl): 3000, 2950, 1735, 1690, 1270, 1180, 1100,

1020, 890; nmr (CCl4): 6 1.21 (t, 6H, J = 7 Hz), 3,15

(t, 4H, J = 2.5 Hz), 4.17 (q, 4H, J = 7 Hz), 4.81 (p, 2H,

J = 2.5 Hz); mass spectrum: m/e 212 (M ).

Anal. Calcd for C11H 604: C, 62.22; H, 7.64.

Found: C, 62.19; H, 7.57.


3-Methylene-l,l-cyclobutanedimethanol (33)

In a three-necked 5 1. flask, a solution of 106.0 g

(0.500 mol) of diester (32) in 700 ml of dry ether was

added dropwise over a 6-hr period to a well-stirred slurry

of 47.5 g (1.25 mol) of LiA1H4 in 2800 ml of dry ether.

After addition, the pot contents were refluxed for 3 hr,

followed by hydrolysis with 48 ml of water, 48 ml of 15%

NaOH, and 48 ml of water. The white lithium salts were

subjected to Soxhlet extraction with refluxing ether for

48 hr. The combined ether portions were dried over Na2SO4

overnight giving 47.6 (75%) yield. Bp: 91-930/0.30 mm;

ir (neat, NaCi): 3380, 2900, 1670, 1030, 880; nmr (CCl4):

6 2.41 (t, 411, J = 2.5 Hz), 3.65 (s, 4H), 3.90 (s, 2H),








4.79 (p, 2H, J = 2.5 Hz); mass spectrum: m/e 128 (M+).


3-Methylene-l,1-cyclobutanedimethanol-di-p-toluene-
sulfonate (34)

In a three-necked 1 1. flask, 50 g (0.392 mol) of (33)
59
in 155 g (1.96 mol) of dry pyridine59 was well stirred

and cooled to ca. -100 with an ice-salt bath. To this in

small portions over a 3-hr period, 186.4 g (0.980 mol) of

finely powdered p-toluenesulfonylchloride was added

through a solid addition funnel. The system was contin-

uously subjected to a dry argon flow throughout the

reaction, and the pot temperature was always maintained

below 50. After addition of the tosylchloride was

complete, the reaction mixture was stirred at below 00

for 90 min more and then warming to 250 for only 30 min.

The pasty material that resulted was poured into a

vigorously stirred mixture of cracked ice and water and

stirred, until all the ice had melted and the paste

solidified. The solid was filtered from the ice water,

recrystallized from absolute ethanol, and dried over full

vacuum. A total of 145 g (85%) of a white powdery solid

was obtained. Mp: 108-1100; ir (KBr): 3070, 2970, 1600,

1380, 1180, 1150, 970, 900, 835, 315; nmr (CDC13):

6 2.35-2.54 (m, 10H), 4.04 (s, 4H), 4.85 (p, 2H, J =

2.5 Hz), 7.54 (AB q, 8H); mass spectrum: m/e 436 (M+).


1,1-Bis(bromomethyl)-3-methylenecyclobutane (35)

In a three-necked 5 1. flask, 70.0 g (0.160 mol) of

(34) in ca. 4000 ml of acetone, dried by distillation over





58


molecular sieves 4A, K2CO3, and KMnO4, was well stirred.

To this in one portion, 97.2 g (1.12 mol) of anhydrous

LiBr was added, and the pot contents were heated to reflux

for 20 hr. After reflux, the acetone was removed by

rotary evaporator, and the residual liquid was taken up in

ether. The ether solution was washed well with water and a

saturated NaCI solution, and it was then dried overnight

over Na2SO4. The product was purified by fractional

distillation at 54-55/0.75 mm giving 32.2 g (79%) of a

colorless liquid. Ir (neat, NaCl): 2940, 1675, 1420,

1238, 889; nmr (CC14): 6 2.63 (t, 4H, J = 2.5 Hz), 3.68

(s, 4H), 4.95 (p, 2H, J = 2.5 Hz); mass spectrum: m/e

254 (M+), 256 (M+ + 2), 256 (M+ + 4).


3-Methylene-l,1-cyclobutanediacetonitrile (36)

In a three-necked 1 1. flask, fitted with an overhead

stirrer, a Friedrich condenser with a CaCl2 drying tube,

a thermometer, and a pressure-equalizing dropping funnel,

76.8 g (1.18 mol) of KCN in 100 ml of dry DMSO was heated

to 70-800, and to this a solution of 30.0 g (1.18 mol) of

(35) in 100 ml of dry DMSO was added dropwise over a 3-hr

period. During the addition the temperature was maintained

between 75-850, and, after the addition was complete, the

reaction mixture was heated to 80-900 for 42 hr. The

cooled pot contents were diluted with 800 ml of water.

The aqueous DMSO solution was extracted with ca. 3 1. of

ether. The ether solution was then washed with water to

remove any last traces of DMSO and dried overnight over





59


Na2SO4. The product was purified by fractional distillation

at 93-940/0.25 mm, yielding 15.0 g (87%) of a colorless

liquid. Ir (neat, NaCI): 2910, 2250, 1675, 1415, 890;

nmr (CC14): 6 2.70 (s, 4H), 2.74 (t, 4H, J = 2.5 Hz),

4.99 (p, 2H, J = 2.5 Hz); mass spectrum: m/e 146 (M ).


3-Methylene-, l-cyclobutanediacetic acid (37)

In a one-necked 2 1. flask, 111.6 g of 87% KOH was

dissolved in 1200 ml of 50% ethanol. To this, 28.0 g

(0.192 mol) of (36) was added in one portion, and the

solution was then heated to reflux for 16 hr. The aqueous

ethanol was removed completely by rotary evaporator,

and the residual solid was redissolved in 150 ml, cooled

to 0, and hydrolyzed with cone. HC1 to a pH = 1. The

product was extracted with 1200 ml of ether and dried

overnight over Na2SO4. The product, once freed from ether,

could be recrystallized from hot CC14 giving 34.0 g (96%)

of a white powdery solid. Mp: 122-123; ir (KBr):

3100, 2900, 2750, 1700, 1650, 1420, 1400, 1340, 1250, 1040,

960, 880; nmr (CD COCD3): 6 2.68 (t, 4H, J = 2.5 Hz),

2.72 (s, 4H), 4.80 (p, 2H, J = 2.5 Hz), 8.79 (s, 2H);

mass spectrum: m/e 184 (M ).

Anal. Calcd for C9H1204: C, 58.69; H, 6.57.

Found: C, 58.58; H, 6.66.


3-Methylene-l,1-cyclobutanediethanol (38)

The diacetic acid (37) above was reduced by the

standard method using LiAlH4 in 72% yield.








Bp: 107.5-110.5*/0.15 mm; ir (neat, NaCl): 3300, 2900,

1670, 1090, 1020, 875; nmr (CC14): 6 1.80 (t, 4H, J =

6.5 Hz), 2.45 (t, 4H, J= 2.5 Hz), 3.65 (t, 4H, J = 6.5 Hz),

4.08 (s, 2H), 4.78 (p, 2H, J = 2.5 Hz).

Anal. Calcd for C9H1602: C, 69.19; H, 10.32.

Found: 68.99; H, 10.29.


3-Methylene-l,l-cyclobutanediethanol-di-p-toluene-
sulfonate (39)

In a three-necked 50 ml flask, a solution of 4.0 g

(0.0256 mol) of (38) in 10.0 g (1.28 mol) of dry pyridine59

was stirred by magnetic stirring bar and cooled to -10

with an ice-salt bath under a constant dry nitrogen purge.

To this 12.2 g (0.0642 mol) of p-toluenesulfonylchloride

was added in small portions over a 35-min period. After

addition, the contents were allowed to stir for an addi-

tional 90 min at -100 to 00 and then allowed to warm to

room temperature (taking ca. 20 min). The pot contents

were poured over ca. 400 ml of cracked ice and water and

vigorously stirred with an overhead stirrer, until all

the ice had melted. A peach-colored paste resulted which

could be recrystallized from warm methanol. A white

crystalline solid was obtained and after drying under full

vacuum weighed 10.0 g (84%). Mp: 66-670; ir (KBr):

2980, 2870, 1665, 1600, 1350, 1180, 1170, 1090, 960, 930,

885, 815, 765; nmr (CC14): 6 1.83 (t, 4H, J = 6.5 Hz),

2.40 (t, 4 H, J = 2.5 Hz), 2.48 (s, 6H), 3.95 (t, 4H, J =

6.5 Hz), 4.71 (p, 2H, J = 2.5 Hz), 7.54 (AB q, 8H).









Anal. Calcd for C23H2806S2: C, 59.46; H, 6.08;

S, 13.80. Found: C, 59.34; H, 6.10; S, 13.80.


1,1-Divinyl-3-methylenecyclobutane (2)

A three-necked 250 ml flask was fitted with a pres-

sure-equalizing dropping funnel, a magnetic stirring bar,

and a bent tube connected to two coiled traps in series

through which a vacuum pump was connected. The flask was

charged with 3.5 g (0.035 mol) of potassium-t-butoxide.

To this, a solution of 4.0 g (0.0086 mol) of (39) in 75 ml

of freshly dried DMSO60 was added dropwise over a 45-min

period under 0.25 mm vacuum and with the coiled traps,

cooled by liquid nitrogen. The reaction temperature was

maintained under vigorous stirring at 0.25 mm vacuum for

12 hr more. The volatile compounds were degassed and

transferred to a small tube via vacuum line and sealed

until purification could be accomplished by gas chroma-

tography on a 5' (18%) DC-200 column at 550. Two pure

products were obtained; 0.21 g (20%) of the desired (2)

and 0.10 g (10%) of what was suspected as the isomerized

3,3-divinyl-l-methylcyclobutene. For (2), ir (gas, NaCl):

3045, 2995, 2960, 1680, 1640, 1415, 1220, 994, 915, 880;

nmr (100 MHz, CDC13): 6 2.75 (t, 4H Jab = 2.7 Hz), 4.82

(p, 2Hb' J = 2.7 Hz), and an ABX pattern at 5.01 (2Hc,

Jcd = 1.4 Hz, Jce = 16.6 Hz), 5.06 (2Hd' Jcd = 1.4 Hz,

de = 10.2 Hz), and 6.03 (2H J = 16.6 Hz, Jde = 10.2

Hz); mass spectrum: m/e 121 (9.5), 120 (83.2), 119 (14.6),








117 (8.0), 106 (10.9), 105 (100), 104 (6.9), 103 (14.6),

93 (16.1), 92 (62.8), 91 (93.4), 80 (11.7), 79 (99.3),

78 (25.5), 77 (54.7), 65 (21.9), 63 (11.7), 53 (20.4),

52 (14.6), 51 (27.0), 50 (9.5), 41 (30.7), 40 (20.4),

39 (49.7).

Anal. Calcd for CH 12: C, 89.94; H, 10.06.

Found: C, 89.87; H, 10.10.

For the suspected 3,3-divinyl--l-methylcyclobutene,

ir (gas, NaCI): 3050, 2930, 2900, 2850, 1640, 1480, 1400,

1260, 990, 910; nmr (CC14): 6 1.68-1.88 (m, 3H), 2.30-

2.50 (m, 2H), 4.72-5.16 (m, 4H), 5.72-6.25 (m, 3H).


4-Methylene-2-vinylcyclohexene (3)

A sealed tube containing a dilute solution of (2)

in benzene was heated to 1150 for 7 hr giving a nearly

quantitative conversion to (3). The product was purified

by gas chromatography on a 5' (18%) DC-200 column at 800.

Ir (CC14, KBr liquid cell): 3090, 3070, 3000, 2980, 2960,

2910, 2835, 1650, 1638, 1605, 1430, 1335, 1220, 975, 890,

840; nmr (CDC13): 6 2.30 (s, 411a), 2.85 (s, 2Hb), 4.55-

5.20 (m, 2Hc + 2Hd), 5.52-5.85 (m, 1H ), 6.28 (AB q, 1Hf,

Jtrans = 17 Hz, Jcis = 10 Hz); uv (n-decane): Xmax
230 nm (E 22,200), 225 nm (c 21,300), 238 nm (E 13,700);

mass spectrum: m/e 121 (9.6), 120 (84.3), 119 (14.5),

106 (12.1), 105 (100), 104 (7.2), 103 (15.7), 93 (14.5),

92 (62.7), 91 (91.6), 80 (12.1), 79 (38.6), 78 (22.9),

77 (43.4), 66 (14.6), 65 (20.5), 63 (14.5), 53 (19.3),








52 (13.3), 51 (9.5), 50 (12.0), 41 (43.4), 40 (9.6), 39

(47.0).

Anal. Calcd for C9H12: C, 89.94; H, 10.06.

Found: C, 90.10; H, 9.86.


Kinetic analysis for l,l-divinyl-3-methylenecyclobutane (2)

A solution of ca. 5% of l,l-divinyl-3-methylene-

cyclobutane (2) and ca. 5% of high purity n-nonane (as an

internal standard) in ca. 90% of high purity n-decane61

was prepared in a 30 ml tube. From this homogeneous

solution, 25 pl aliquots were withdrawn and transferred

by syringe to 6" capillary tubes (0. D. = 7 mm; bore,

1.0-1.5 mm). The capillary tubes were sealed under

vacuum with the contents cooled by liquid nitrogen prior

to evacuation. In this manner, seventy samples were

readied for kinetic runs. Each capillary tube, prior to

immersion in hot oil, was protected by a well-fitting wire

gauze sleeve. The tubes were heated in a well-insulated

and well-stirred oil bath, a Lauda constant temperature

circulator type N/S15/12, equipped with an R-10 electronic

relay, an R-20 electronic controller, and a R-30 tempera-

ture and level protection relay, and filled with Ultra-

Therm 330S silicone fluid. Temperatures were monitored

both by an NBS thermometer, calibrated in 0.10, and by

using a calibrated chromel-alumel thermocouple in con-

junction with a Honeywell model 2702 potentiometer. As

few as six, but as many as twelve prepared tubes were









used per temperature run. Each tube was withdrawn from

the oil bath at an appropriate time, cooled immediately

by a dry ice-isopropanol bath, opened, and analyzed by

gas chromatography on a 7.5' (2%) DC-200 column at 560.

The kinetic analyses were performed on a Hewlett-Packard

model 5710A gas chromatograph with a flame ionization

detector. The chemical composition of the kinetic

samples was determined with a Vidar Autolab 6300 digital

integrator by following the disappearance of (2) with

respect to the internal standard, n-nonane (see Appendix I).

A first-order, least squares program was used in con-

junction with a pdp 8/e digital computer to analyze all

the data. Rate constants at seven different temperatures

were determined and compiled in Table 1.



Table 1



Kinetics for the thermal rearrangement of (2) to (3)


Temp (oC)

86.1

93.2

99.9

100.7

107.1

114.6

121.1


Temp (oK)

359.25

366.35

373.05

373.85

380.25

387.75

394.25


-3
10 /Temp (K)

2.7836

2.7296

2.6806

2.67419

2.6298

2.5790

2.5365


k x 105 sec

1.930.04

4.090.06

7.560.09

8.360.16

16.5 0.2

30.9 0.5

57.8 0.5














-7.50 -




-8.00



-8.50




-9.00




-9.50




-10.00




-10.50






2.5 2.6 2.7 2.8 2.9


103/T (OK)


Figure 3: Arrhenius plot for the conversion of (2) to (3)





















-n-decane






















S-(3)


9 6 3


time (min)



Figure 4: Sample glpc trace of the kinetic analysis
for the thermal conversion of (2) to (3)


-n-nonane








A computer program was formulated to produce an Arrhenius

plot of log k = log A Ea/RT for this first-order

reaction,62 and it also calculated the appropriate

parameters. An Arrhenius plot for these data appears in

Figure 3. Energy of activation, 27.270.30 kcal/mol;

log A, 11.870.17; enthalpy of activation, 26.530.30

kcal/mol at 103.20; entropy of activation, -6.660.78

cal/mol-deg at 103.20; free energy of activation,

29.030.78 kcal/mol at 103.20. The correlation coefficient

was 0.9997.


3-Methylene-l,1-cyclobutanedi(ethan-a,a-d2-ol) (38-d4)

The method used for the reduction of 3-methylene-

1,1-diacetic acid (37) in the preparation of (38) was

similarly followed by using LiAlD4 in the preparation of

(38-d ) in 68% yield. Bp: 130-131/0.75 mm; ir (neat,

NaCl): 3240, 3030, 2860, 2800, 2180, 2070, 1660, 965;

nmr (CDCl3): 6 1.78 (s, 4H), 2.42 (t, 4H, J = 2.5 Hz),

3.83 (s, 2H), 4.78 (p, 2H, J = 2.5 Hz).


3-Methylene-l,l-cyclobutanedi(ethan-a,a-d2-ol)-di-p-
toluenesulfonate (39-d )

The procedure used in the bistosylation of (38) was

similarly followed with (38-d4) resulting in a 77% yield.

Mp: 67.5-68.50 (methanol); ir (KBr): 3000, 2880, 2230,

2120, 1660, 1350, 965, 890; nmr (CDC13): 6 1.78 (s, 4H),

2.39 (t, 4H, J = 2.5 Hz), 2.44 (s, 6H), 4.73 (p, 2H,

J = 2.5 Hz), 7.52 (AB q, 8H).








Anal. Calcd for C23 24D 06S2" C, 58.95; H + D,

5.16 + 1.72 = 6.88; S, 13.68. Found: C, 58.98; H + D,

6.87; S, 13.61.


3-Methylene-l, -di(vinyl-2,2-d2)cyclobutane (2-d4)

The preparation of this compound was similar to that

for (2). Gas chromatography on a 5' (18%) DC-200 column

at 550 afforded two hydrocarbons; 0.32 g (17%) of (2-d4),

and 0.21 g (10%) of what was suspected as 1-methyl-3,3-

di(vinyl-2,2-d2)cyclobutene. For (2-d ), ir (gas, NaCI):

3090, 3010, 2970, 2935, 2320 (w), 2230 (w), 1680, 1600,

1420, 1035, 945, 880, 735; nmr (CCl4): 6 2.72 (t, 4Ha,

Jab = 2.5 Hz), 4.89 (p, 2Hb, Jab = 2.5 Hz), 5.85-6.10
(m, 2H); mass spectrum: m/e 124 (M ).

Anal. Calcd for CgIgD : C, 87.02; H + D, 6.49 +

6.49 = 12.98. Found: C, 86.85; H + D, 12.93.

For the suspected l-methyl-3,3-di(vinyl-2,2-d2)-

cyclobutene, ir (CC14, KBr liquid cell): 2900, 2860,

2320 (w), 2210 (w), 1640, 1580, 1430, 930, 860; nmr (CCl4):

6 1.70 (q, 3H, J = 1.5 Hz), 2.31 (q, 2H, J = 1.5 Hz), 5.77

(q, 3H, J = 1.5 Hz), 5.90 (m, 2H).


4-Methylene-2-(vinyl-2,2-d2)cyclohexene-6,6-d (3-d4)

Into four 1 ml tubes, 400 yl each of ca. 10% (2-d ) in

benzene was transferred by syringe and sealed under vacuum.

A protective wire gauze sleeve was placed around each tube,

and the contents of the tube were subsequently heated at

120* for 5 hr. Purification was accomplished by gas








chromatography on a 5' (18%) DC-200 column at 600. The

reaction appeared to be quantitative. Ir (CC14, KBr liquid

cell): 3080, 2990, 2930, 2860, 2190 (w), 2160 (w), 2100

(w), 1680, 1650, 1445, 1070, 1030, 945, 890, 718; nmr (CC14):

6 2.26 (s, 21H ), 2.82 (s, 2Hb), 4.67-4.86 (m, 2H ), 5.71

(1i ), 6.15-6.36 (m, 1Hf); mass spectrum: m/e 124 (M ).

Anal. Calcd for C9 HD4: C, 87.02; H + D, 6.49 +

6.49 = 12.98. Found: C, 87.08; H + D, 12.91.


Kinetic analysis for 3-methylene-l,l-di(vinyl-2,2-d2)-
cyclobutane (2-d.)

A solution of ca. 5% of 3-methylene-l,l-di(vinyl-

2,2-d2)cyclobutane (2-d4) and ca. 5% of high purity n-nonane
61
(as an internal standard) in ca. 90% of n-decane was

prepared in a 10 ml tube. From this homogeneous solution,

25 ul aliquots were extracted and transferred by syringe

to 6" capillary tubes. In every remaining detail, the

kinetics and their analyses were identical to that already

described for (2). The run at each temperature was done

immediately following the kinetic run of the respective (2),

so that an accurate comparison of rate data of (2-d4) with

(2) could be made with a minimum of error. A first-order,

least squares treatment was used in conjunction with a

pdp 8/e digital computer to analyze the data. Rate constants

were determined at two different temperatures (see also

Appendix I) and compiled in Table 2.








Table 2



Kinetics for the thermal rearrangement of (2-d4) to (3-d )



Temp (C) Temp (oK) 10-3/Temp (K) k x 10 sec

99.9 373.05 2.6806 7.170.08

121.1 394.25 2.5365 55.4 0.8



3-Methylene-l,l-cyclobutanediacetic anhydride (40)

In a 100 ml one-necked flask, 10.0 g (0.0473 mol) of

3-methylene-l,l-cyclobutanediacetic acid (37) was dissolved

in 12.0 g (0.118 mol) of acetic anhydride and gently refluxed

for 15 min. The pot contents were cooled and subjected to

full vacuum for 12 hr. The residual material was recrystal-

lized from a 4:1 solution of pet ether (30-800) and benzene,

collecting 7.8 g (100%). In this manner, more anhydride

was made on the same scale until sufficient amounts of

product were obtained. Mp: 55-560; ir (KBr): 2850, 1810,

1760, 1665, 1400, 1150, 950, 890, 810; nmr (CC14): 6 2.63

(t, 4H, J = 2.5 Hz), 2.83 (s, 4H), 4.88 (p, 2H, J = 2.5

Hz); mass spectrum: m/e 166 (M ).


Ethyl, hydrogen-3-methylene-1,1-cyclobutanediacetate (41)

In a 50 ml one-necked flask, 16.0 g (0.096 mol) of (40)

was dissolved and heated with 6.8 g (0.150 mol) of absolute

ethanol over a steam bath for 2 hr. The excess ethanol was

removed by rotary evaporator, and the product distilled at









130-1320/0.45 mm, affording 17.6 g (88%) of a colorless

liquid. Ir (neat, NaCI): 3350, 3130, 2900, 1800, 1720,

1650, 1170, 1050, 965; nmr (CC14): 6 1.25 (t, 3H, J =

2.5 Hz), 2.50-2.80 (m, 8H), 4.09 (q, 2H, J = 7 Hz), 4.75

(p, 2H, J = 2.5 Hz), 11.05 (s, 1H); mass spectrum: m/e

212 (M ).

Anal. Calcd for CH1 1604: C, 62.25; H, 7.60.

Found: C, 62.11; H, 7.59.


1-Acetic acid, l-ethanol-3-methylenecyclobutane (42)

A total of 39.4 g (0.186 mol) of (41) was divided up

into six portions. In a typical run, 6.0 g (0.0351 mol) of

(41) along with 5.5 g (0.120 mol) of absolute ethanol were

dissolved in liquid NH3. A dry ice condenser and a dewar,

cooled with dry ice and isopropanol were used to maintain

NH3 as a liquid in a 1 1. three-necked flask. Precautions

were taken to minimize the condensation of water within

the reaction flask. To this cold solution, 2.4 g (0.105

g-atom) of Na was added in small portions at a rate which

insured that there was no remaining blue color with each

subsequent addition of a Na sliver. After the Na addition

was complete, a small amount of absolute ethanol was added

to hasten the quenching of the blue color. The liquid NH3

was allowed to evaporate overnight. The residual material

was dissolved in water, cooled to 0, and acidified with

conc. HCI to a pH = 1, followed by ether extraction.

The ether extracts were dried over MgSO4 overnight.









After rotary evaporation, the residual liquid was dried

over full vacuum for 12 hr. The crude yield weighed 2.9 g

(49%) and was used as is in the next step. Ir (neat, NaCl):

3500, 3200, 2980, 2900, 2650, 1760, 1670, 1420, 1300, 1260,

1230, 1070, 885; nmr (CCl4): 6 1.92 (t, 2H, J = 6 Hz),

2.44-2.76 (m, 7H), 4.26 (t, 2H, J = 6 Hz), 4.82 (p, 2H,

J = 2.5 Hz), 7.98 (s, 1H).


3-Methylene-l,l-cyclobutanediethanol-a,a-d2 (38-d2)

In a three-necked 2 1. flask, prepurged with dry argon,

4.80 g (0.126 mol) of LiAlD4 was well stirred in 600 ml of

freshly dried ethyl ether. To this over 170 min, a

solution of 17.2 g (0.115 mol) of crude (42) in 500 ml of

freshly dried ethyl ether was added dropwise. After

addition, the pot contents were gently refluxed for 4 hr.

The pot contents were cooled and hydrolyzed in the usual

manner, filtered, and dried over MgSO4. The solid lithium

salts were subjected to Soxhlet extraction with refluxing

ether and also dried. The product distilled at 113-1220/

0.30 mm, affording 5.3 g. Ir (neat, NaCl): 3250, 2870,

2800, 2190, 2080, 1670, 1040, 970, 875; nmr (CDCl3): 6

1.55-2.05 (m, 4H), 2.44 (t, 4H, J = 2.5 Hz), 3.64 (t, 2H,

J = 8 Hz), 3.87 (s, 1H), 3.98 (s, 1H), 4.78 (p, 2H, J =

2.5 Hz).


3-Methylene- ,1-cyclobutanediethanol-a,a-d2-di-p-
toluenesulfonate (39-d2)

The procedure used in the bistosylation of (38) was









similarly followed with (38-d2). Mp: 66-670 (methanol);

ir (KBr): 3000, 2890, 2240, 2180, 1900, 1660, 1590, 1345,

1190, 1170, 1095, 960, 935, 875, 870, 765, 708; nmr (CDC13):

6 1.50-1.96 (m, 4H), 2.20-2.62 (m, 10H), 4.02 (t, 2H, J =

7 Hz), 4.74 (p, 2H, J = 2.5 Hz), 7.55 (AB q, 8H).

Anal. Calcd for C23H26D206S2: C, 59.20; H + D, 5.62 +

0.86 = 6.48; S, 13.74. Found: C, 59.03; H + D, 6.40;

S, 13.65.


3-Methylene-l-vinyl-1-(vinyl-2,2-d2)cyclobutane (2-d2)

The preparation of this compound was similar to that

for (2). Ir (gas, NaC1): 3090, 3010, 2975, 2940, 2280 (w),

2130 (w), 1680, 920, 880; nmr (C6D6): 6 2.87 (t, 4H, J =

2.5 Hz), 4.82-5.18 (m, 3H), 5.21-5.38 (m, 1H), 5.92-6.12

(m, 2H). The exact nmr integration ratio (from a minimum

of 10 integration) of all vinyl H to all allyl H of

1.470.01 (should be 1.50), and the exact nmr integration

ratio (from a minimum of 10 integration) of all terminal-

vinyl H to all internal-vinyl H of 1.950.02 (should be

2.00) both indicate that contaminants (2-d ) and (2-d ) may

be present. By low voltage (6 ev) mass spectral analysis in

the parent peak region, an accurate assessment of the

distribution of isotopic isomers was determined: (2-d2)

0.861; (2-d3), 0.050; (2-d4), 0.089.

Anal. Calcd for C9 H D2: C, 88.45; H + D, 8.25 + 3.30

= 11.55. Found: C, 88.36; H + D, 11.49. Duplicate:

C, 88.39; H + D, 11.47.








For the suspected l-methyl-l-vinyl-l-(vinyl-2,2-d )-

cyclobutene, ir (gas, NaCI): 3080, 3040, 3000, 2960, 2920,

2880, 2300 (w), 2220 (w), 1680, 1585, 1450, 990, 910; nmr

(CC14): 6 1.63-1.84 (m, 3H), 2.30-2.43 (m, 2H), 4.62-5.18

(m, 4H), 5.58-6.24 (m, 3H).


Thermolysis of 3-methylene-l-vinyl-- (vinyl-2,2-d2)-
cyclobutane (2-d 2); Intramolecular isotope effect

A 5% solution of (2-d2) in benzene was prepared,

sealed in a large capillary tube (bore, 5 mm), and pyrolyzed

at 121.50 for 5 hr. A sample of (2) was subjected to the

same conditions and was used as a control. Gas chroma-

tography on a 5' (18%) DC-200 column at 700 was used to

separate solvent from product. The product was then taken

up in CDC13, and a nmr spectrum was taken. Two separate

determinations were possible from a 100 MHz nmr spectrum.

One involved the continuous-sweep integration of the

aliphatic region with the determination of the ratio of the

allylic H/diallylic H or H /Hb. The other involved the

continuous-sweep integration of the vinylic region with

the determination of the ratio of terminal-vinyl H/

vinyl-ring H or (Hc + Hd)/H


Calculation of the intramolecular isotope effect from
the aliphatic region in the nmr of the mixture of the
dideuterated sample of (3a + 3b)



Ha/Hb = 2.00(Fa) + 1.00(F3b) + 1.25(F d3) + 1.00(F3d)
3 4









Ha/Hb = 2.00(F3a) + 1.00(F3b) + 1.25(0.050) + 1.00(0.089)

1.465* = 2.00(F3a) + 1.00(F3b) + 0.1515

1.313 = 2.00(F3a) + 1.00(F3b)

But F3a + F3b = 0.861 or F3 =0.861 F3a

Thus F3a = 0.4525 and F3b = 0.4085

Thus F 3a/F3b = 0.4525/0.4085 = 1.11
3a 3b


* See Table 3.


Calculation of the intramolecular isotope effect from the
vinylic region in the nmr of the mixture of the dideuterated
sample of (3a + 3b)


(He + Hd)/H = 2.00(F3a) + 4.00(F3b) + 2.50(F3d3) +

2.00(F )
3-d
-4

(Hc + Hd)/He = 2.00(F3a) + 4.00(F3b) + 2.50(0.050) +

2.00(0.089)

2.818* = 2.00(F3a) + 4.00(F3b) + 0.303

1.257 = 1.00(F3a) + 2.00(F3b)

But F3a + F = 0.861 or F3a= 0.861 F3b

Thus F3 = 0.3965 and F3 = 0.4645

Thus F 3a/Fb = 0.4645/0.3965 = 1.17
3a 3b


* See Table 4.









Table 3



Nmr integrations in the aliphatic region
of the product mixture of (3a + 3b)


Area of H
-----------a

55.0

56.0

57.0

56.5

56.0

55.5

55.0

54.0

54.0

54.0

55.0

55.0


Area of Hb
---------6
35.0

34.5

36.0

35.0

35.0

34.5

34.5

34.5

34.5

34.5

34.5

34.5


Average:



* For the control sample, Ha/Hb

be 2.000), so that the actual ratio

dideuterated sample is 1.4650.016.


(H /Hb

1.571-0.018

1.623+0.033

1.583+0.007

1.614+0.024

1.600+0.010

1.609+0.019

1.594+0.004

1.565-0.025

1.565-0.025

1.565-0.025

1.594+0.004

1.594-0.004

1.5900.016*


= 2.1700.027 (should

of Ha/Hb in the
a b


Runs

1

2

3

4

5

6

7

8

9

10

11

12









Table 4


Nmr integration
of the product


in the vinylic region
mixture of (3a + 3b)


Area of (H + H)
----------c ---d -i-
55.0

54.5

55.5

56.0

54.5

55.0

57.0

53.5

55.5

55.0

55.0


Area of H
--------------
18.0

17.5

18.5

18.0

17.5

18.0

19.0

17.5

17.5

17.5

18.0


Average:


(H + H )/H-*

3.056-0.024

3.114+0.035

3.000-0.079

3.111+0.032

3.111+0.032

3.056-0.024

3.000-0.079

3.057-0.022

3.171+0.092

3.143+0.063

3.056-0.024

3.0790.046*


* For the control sample, (Hc + Hd)/H = 4.3720.038

(should be 4.000), so that the actual ratio of

(H + Hd)/He in the dideuterated sample is 2.8180.046.


Runs

1

2

3

4

5

6

7

8

9

10

11


__ ____









Table 5



Intramolecular isotope effect results for
the rearrangement of (2-d,) to (3a) and (3b)


Protons

H /H
HaHb
(Hc + Hd)/He


Nmr integration ratio

1.4650.016a

2.8180.046a


Average:


(3a/3b) or (kH/kD)

1.110.06b

1.170.09b

1.140.07


Corrected to the control run.

Corrected for isotopic isomer contamination.








3-Methylenecyclobutanecarbonitrile (43)30

A 1.4 1. stainless steel rocker autoclave was charged

with 230.0 g (4.25 mol) of freshly distilled acrylonitrile,

200 ml dry toluene, 50.0 g (1.25 mol) of allene, and

2.0 g of hydroquinone. The autoclave was heated to 2200

for 16 hr with agitation and under autogenous pressure.

After cooling to room temperature, the autoclave was opened,

and the product mixture was freed from toluene and unreacted

acrylonitrile. The product was fractionally distilled at

64.5-65.0/21 mm (lit.30 64-65"/21 mm), collecting 57.0 g

(59%). Ir (neat, NaCl): 3000, 2350, 1440, 885; nmr (CC14):

6 2.92-3.34 (m, 5H), 4.75-5.05 (m, 2H); mass spectrum: m/e

93 (M+).


3-Methylenecyclobutanecarboxylic acid (44)30

In a 1 1. one-necked flask, 93.5 g of 87% KOH was

dissolved in 670 ml of 50% ethanol. To this, a total of

33.3 g (0.357 mol) of (43) was added in one portion. The

reaction mixture was refluxed on a steam bath for 4 hr.

The carboxylic acid salt was freed from the solvents by

rotary evaporation and then redissolved in 150 ml of water

and 2 ml of benzene. Again the solvents were removed from

the carboxylic acid salt by rotary evaporation. The

carboxylic acid salt was then dissolved in 200 ml of water,

cooled to 00, and hydrolyzed with conc. HC1 to a pH = 1.

The product was extracted with 4 x 100 ml portions of ether

and dried over Na2SO4 overnight. The product was








fractionally distilled at 105-1060/10 mm (lit.30 99-1010/

9 mm), collecting 34.0 g (93%) of a colorless liquid.

Ir (neat, NaCl): 2950, 1700, 1410, 1320, 1290, 1240, 1200,

1040, 935, 885, 785, 690; nmr (CC14): 6 2.72-3.33 (m, 5H),

4.65-4.96 (m, 2H), 0.00 (s, 1H); mass spectrum: m/e

112 (M+).


3-Methylenecyclobutylcarbinol (45)

In a 2 1. three-necked flask, 20.0 g (0.526 mol) of

LiAlH4 was well stirred with 1000 ml of dry ether. To this

over a 3-hr period, 49.6 g (0.442 mol) of (44) was added

dropwise at 250. The pot contents were gently refluxed

for 20 hr and then cooled to 0, prior to hydrolysis with

20 ml of water, 20 ml of 15% of NaOH solution, and 20 ml of

water. The product mixture was filtered, and the ether

solution was dried with Na2SO4 overnight. In addition,

the lithium salts were subjected to Soxhlet extraction with

refluxing ether, and this too was dried over Na2SO4 overnight.

Once freed from ether, the dry product was fractionally

distilled at 101-1030/78 mm, yielding 41.2 g (95%). Ir

(neat, NaC1): 3300, 2900, 1670, 1410, 1040, 1020, 975,

900, 880; nmr (CC14): 6 2.10-3.10 (m, 5H), 3.50 (s, 1H),

3.60 (s, 1H), 4.44 (s, 1H), 4.62-4.82 (m, 2H); mass

spectrum: m/e 98 (M ).


3-Methylenecyclobutylcarbinyl-p-toluenesulfonate (46)

In a 250 ml three-necked flask, 20.0 g (0.204 mol) of
59
(45) in 40.0 g (0.508 mol) of dry pyridine was well








stirred with an overhead stirrer and cooled to ca. -100

with an ice-salt bath. In small portions over a 90-min

period, 48.8 g (0.255 mol) of finely powdered p-toluene-

sulfonylchloride was added through a solid addition tube.

The system was continuously subjected to a dry argon flow

throughout the reaction, and the pot temperature was

always maintained below 5. After addition of the tosyl-

chloride was complete, the reaction mixture was stirred at

below 0 for 45 min more and then warming to 250 gradually

for 30 min. The resulting pasty material was poured into

a vigorously stirred mixture of cracked ice and water. It

was stirred until all the ice had melted and the paste

solidified. The solid was filtered from ice water,

recrystallized from absolute ethanol, and dried over full

vacuum. The product weighed 50.2 g (98%). Mp: 40.5-

41.50; ir (KBr): 3000, 2900, 1910, 1670, 1590, 1490,

1460, 1360, 1190, 1180, 1120, 1100, 1020, 960, 880, 845,

815, 790, 775; nmr (CDC13): 6 2.40 (s, 3H), 2.10-2.90

(m, 5H), 4.00-4.10 (m, 2H), 4.68-4.85 (m, 2H), 7.57

(AB q, 4H); mass spectrum: m/e 252 (M ).


3-Methylenecyclobutaneacetonitrile (47)

In a 500 ml three-necked flask, fitted with an over-

head stirrer, a Friedrich condenser with a CaCl2 during

tube, a thermometer, and a pressure-equalizing dropping

funnel, 51.5 g (0.795 mol) of KCN in 100 ml of dry DMSO

was heated to ca. 65, and a solution of 40.0 g (0.159 mol)








of (46) in 100 ml of DMSO was added dropwise over a

75-min period. During the addition the temperature was

maintained between 65-75, and afterwards the reaction

mixture was heated additionally to 750 for 22 hr. The

cooled pot contents were diluted with 800 ml of water.

The aqueous DMSO solution was extracted with 600 ml of

pentane, and this extract was washed with water and dried

over molecular sieves 3A. The product was purified by

fractional distillation at 104.5-106.50/57 mm, yielding

11.5 g (68%). Ir (neat, NaC1): 3050, 2900, 2250, 1670,

1430, 1410, 1390, 1340, 1200, 885; nmr (CCl4): 6 2.76-

3.87 (m, 7H), 4.73-4.92 (m, 2H); mass spectrum: m/e

107 (M+).


3-Methylenecyclobutylacetic acid (48)

In a 500 ml one-necked flask, 28.9 g of 87% KOH was

dissolved in 250 ml of 50% ethanol. To this, 11.0 g

(0.103 mol) of (47) was added in one portion, and the

solution was then heated to reflux for 4 hr. The aqueous

ethanol was removed completely by rotary evaporation under

30 mm pressure, and the residual solid was redissolved in

100 ml of water, cooled to 00, and hydrolyzed with cone.

HCI to a pH = 1. The product was extracted with 500 ml of

ether and dried overnight over Na2SO4. The product was

fractionally distilled at 118-122/17 mm, yielding 10.5 g

(81%). Ir (neat, NaCl): 3100, 2900, 2650, 1710, 1670,

1430, 1410, 1305, 1280, 1210, 1130, 940, 880;









nmr (CC14): 6 2.00-3.20 (m, 7H), 4.62-4.85 (m, 2H),

12.72 (s, 1H); mass spectrum: m/e 126 (M +).


3-Methylenecyclobutylethanol (49)

In a 500 ml three-necked flask, 3.4 g (0.090 mol) of

LiAlH4 was well stirred with 180 ml of dry ether. To this

over a 75-min period, 9.0 g (0.080 mol) of (48) was added

dropwise at ca. 250. The pot contents were gently refluxed

for 19 hr and then cooled to 00, prior to hydrolysis with

4 ml of water, 4 ml of 15% NaOH solution, and 4 ml of water.

The product mixture was then filtered, and the ether

solution was dried with Na2SO4 overnight. The product was

fractionally distilled at 89.0-91.50/22 mm, giving 6.3 g

(70%) of product. Ir (neat, NaCI): 3300, 3050, 2900, 1670,

1400, 1055, 875; nmr (CC14): 6 1.48-1.93 (m, 2H), 2.00-

2.10 (m, 5H), 3.42-3.67 (m, 3H), 4.51-4.85 (m, 2H);

mass spectrum: m/e 112 (M ).


3-Methylenecyclobutylethanol-p-toluenesulfonate (50)

In a 50 ml three-necked flask, 3.0 g (0.027 mol).of
59
(49) in 5.3 g (0.067 mol) of dry pyridine59 was well stirred

with magnetic stirring bar and cooled to ca. -100 with an

ice-salt bath. In small portions, 6.4 g (0.034 mol) of

p-toluenesulfonylchloride was added through a solid addition

funnel over a 15-min period. The system was continuously

subjected to a dry argon flow throughout the reaction, and

the the pot temperature was always maintained around 0.

When addition was complete, the reaction mixture was








stirred at below 0 for 75 min more and then warming to 250

for only 15 min. The resulting pasty material could not be

solidified when poured into a vigorously stirred cracked

ice and water mixture, so that the resulting oil was taken

up in ether, washed with 200 ml of 3% NaHCO3 solution and

100 ml of water, and then dried over molecular sieves 3A

for 4 hr. Once freed from ether, the product weighed

6.0 g (84%) which was used without further purification

for spectral data and for use in the following reaction.

Ir (neat, NaCI): 3050, 2950, 1670, 1600, 1360, 1190, 1180,

940, 880; nmr (CDC13): 6 1.55-3.00 (m, 7H), 2.45 (s, 3H),

3.90-4.15 (t, 2H, J = 6.5 Hz), 4.60-4.82 (m, 2H), 7.57

(AB q, 4H); mass spectrum: m/e 266 (M ).


Methylene-3-vinylcyclobutane (29)

A 250 ml three-necked flask was fitted with a pressure-

equalizing dropping funnel, a magnetic stirring bar, and a

bent tube connected to a coiled trap through which a vacuum

pump was connected. The flask was charged with 4.6 g

(0.041 mol) of potassium-t-butoxide. To this, a solution

of 6.0 g (0.023 mol) of (50) in 40 ml of freshly dried

DMSO60 was added dropwise over a 35-min period under 25 mm

vacuum and with the coiled trap cooled by liquid nitrogen.

Immediately after addition, the vacuum was increased to

0.75 mm and held there for 8 hr. The volatile compounds,

collected in the coiled trap, were transferred under full

vacuum to a small, more accessible tube via vacuum line








and sealed until purification could be accomplished.

Gas chromatography on a 5' (18%) DC-200 column at 600

afforded 0.43 g (20%) of a pure product. Ir (gas, NaCI):

3090, 2950. 1680, 1645, 1420, 990, 920, 915, 880; nmr

(CC14): 6 2.33-3.12 (m, 5H), 4.65-4.95 (m, 3H), 4.97-

5.18 (m, 1H), 5.80-6.40 (m, IH); mass spectrum: m/e

94 (7.7), 93 (25.6), 91 (17.9), 80 (12.8), 79 (100),

78 (7.7), 77 (41.0), 67 (5.1), 66 (20.5), 65 (10.3), 55

(7.7), 54 (66.7), 53 (23.1), 51 (12.8), 43 (15.4), 42

(7.7), 41 (12.8), 40 (15.4), 39 (56.4).

Anal. Calcd for C7H10: C, 89.29; H, 10.71.

Found: C, 89.13; H, 10.77.


4-Methylenecyclohexene (30)33

In four 1 ml tubes, 400 pl each of 5% of (29) in

n-decane was transferred by syringe and sealed under vacuum.

A protective wire gauze sleeve was placed around each tube

and subsequently heated to 2100 for 15 hr. Purification

was accomplished by gas chromatography on a 5' (18%) DC-200

column at 650. The reaction was quantitative and clean.

Ir (gas, NaCI): 2980, 2890, 1640, 1430, 890; nmr (CC14):

6 2.21 (s, 4H), 2.72 (s, 2H), 4.68 (s, 2H), 5.62 (s, 2H);

mass spectrum: m/e 95 (7.2), 94 (76.7), 93 (20.9), 92 (4.7),

91 (30.5), 80 (9.6), 79 (100), 78 (15.8), 77 (56.8),

67 (5.1), 66 (13.7), 65 (11.0), 55 (6.2), 54 (11.0), 53

(15.8), 52 (6.8), 51 (15.4) 50 (9.6).









Kinetic analysis for methylene-3-vinylcyclobutane (29)

A solution of ca. 5% of methylene-3-vinylcyclo-

butane (29) in ca. 95% of high purity n-decane61 was

prepared in a 10 ml tube. From this homogeneous solution,

25 il aliquots were extracted and transferred by syringe

to 6" capillary tubes (0. D. = 7 mm; bore, 1.0-1.5 mm).

The capillary tubes were sealed under vacuum with the

contents cooled by liquid nitrogen prior to evacuation.

In this manner, forty tubes were readied for kinetic

runs. Each capillary tube, prior to immersion in the

hot oil, was protected by a well-fitting wire gauze sleeve.

The tubes were heated in a well-insulated and well-stirred

oil bath, a Lauda constant temperature circulator type

N/S15/12, equipped with an R-10 electronic relay, an R-20

electronic controller, and a R-30 temperature and level

protection relay, and filled with Ultra-Therm 330S silicone

fluid. Temperatures were monitored both by a thermometer,

calibrated in 0.10, and by using a calibrated chromel-

alumel thermocouple in conjunction with a Honeywell model

2702 potentiometer. Six or seven prepared tubes were used

per temperature run. Each tube was withdrawn from the oil

bath at an appropriate time, cooled immediately by a dry

ice-isopropanol bath, opened, and analyzed by gas

chromatography on a 7.5' (2%) DC-200 column at 50. The

kinetic analyses were performed on a Hewlett-Packard model

5710A gas chromatograph with a flame ionization detector.

The chemical composition of the kinetic samples was









determined with a Vidar Autolab 6300 digital integrator.

A first-order, least squares program2 was used in con-

junction with a pdp 8/e digital computer to analyze all

the data (see Appendix I). Rate constants at six

different temperatures were determined and compiled in

Table 6.



Table 6



Kinetics for the thermal rearrangement of (29) to (30)


Temp (OC)

175.00

182.00

187.75

196.25

202.25

211.00


Temp (K)

448.15

455.15

460.90

469.40

475.40

484.15


10-3/Temp (K)

2.2314

2.1971

2.1697

2.1304

2.1035

2.0655


k x 10 sec

1.970.02

3.800.03

6.230.04

12.6 0.8

20.6 0.3

38.5 0.2


A computer program was formulated to produce an Arrhenius

plot of log k = log A Ea/RT for this first-order

reaction,62 and it also calculated the appropriate

parameters. An Arrhenius plot for these data appears in

Figure 5. Energy of activation, 35.670.41 kcal/mol;

log A, 12.700.19; enthalpy of activation, 34.740.41

kcal/mol at 192.30; entropy of activation, -3.290.87

cal/mol-deg at 192.30; free energy of activation,















-7.50



-8.00




-8.50




-9.00




-9.50




-10.00




-10.50





2.0 2.1 2.2 2.3 2.4


3 K)
i0 /T (0K)


Figure 5: Arrhenius plot for the conversion of (29) to (30)






89
































-n-decane (30) (29)














12 9 6 3

time (min)


Figure 6: Sample glpc trace of the kinetic analysis
for the thermal conversion of (29) to (30)





90


36.280.87 kcal/mol at 192.30. The correlation coefficient

was 0.9998.















APPENDIX I


A detailed compilation of the kinetic data on

concentration versus time is presented in Tables 7 through

9. Table 7 includes the kinetic data for the pyrolysis of

l,l-divinyl-3-methylenacyclobutane (2). Table 8 includes

the kinetic data for the pyrolysis of l,l-di(vinyl-2,2-d2)-

3-methylenecyclobutane (2-d4), and Table 9 includes the

kinetic data on the pyrolysis of methylene-3-vinylcyclo-

butane (29). Under each temperature run, time is listed in

seconds. The ratio is a ratio of l,l-divinyl-3-methylene-

cyclobutane (2) or 1,1-di(vinyl-2,2-d2,)-3-methylenecyclo-

butane (2-d ) to chromatographic quality n-nonane. For the

thermolysis of methylene-3-vinylcyclobutane (29) to

4-methylenecyclohexene (30), the ratio represents a ratio

of (29) to (29)/[(29) + (30)] at time t. The listing,

deviation, is the computer calculation of the deviation

of the observed In ratio (also tabulated) with that which

would be expected if the point lay precisely on a perfectly

straight line. This deviation provides an indication for

the goodness of fit of each point in these first-order

reactions. The rate constant (in sec- ), the intercept,

and the correlation coefficient were also determined by a






92



least squares method computer analysis. A concentration

versus time plot comparison was constructed in Figures 7

and 8 for the kinetic runs of (2) and (2-d ) at 99.90

and 121.10, temperatures at which intermolecular isotope

effects were measured.