Citation
Attempts to generate a homoaromatic carbene and bisnorcaradiene

Material Information

Title:
Attempts to generate a homoaromatic carbene and bisnorcaradiene
Creator:
Parker, Richard Henry, 1950-
Copyright Date:
1979
Language:
English
Physical Description:
ix, 86 leaves : ill. ; 28 cm.

Subjects

Subjects / Keywords:
Carbenes ( jstor )
Chlorides ( jstor )
Ethers ( jstor )
Isomers ( jstor )
Mass spectroscopy ( jstor )
Molecular spectra ( jstor )
Potassium ( jstor )
Protons ( jstor )
Signals ( jstor )
Sodium ( jstor )
Aromatic compounds ( lcsh )
Chemistry thesis Ph. D
Dissertations, Academic -- Chemistry -- UF
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis--University of Florida.
Bibliography:
Bibliography: leaves 82-85.
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Richard H. Parker.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright [name of dissertation author]. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Resource Identifier:
023052765 ( AlephBibNum )
05691729 ( OCLC )
AAK4380 ( NOTIS )

Downloads

This item has the following downloads:


Full Text











ATTEMPTS TO GENERATE A HOMOAROMATIC CARBENE
AND BISNORCARADIENE




















By

Richard H. Parker













A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE DOCTOR OF PHILOSOPHY






UNIVERSITY OF FLORIDA


1979






































The great tragedy of Science [is] the slaying of a
beautiful hypothesis by an ugly fact.

Thomas Henry Huxley


i
















ACKNOWLEDGMENTS


The author would like to sincerely acknowledge the

guidance and patience of his research director,

Dr. W. M. Jones. His ability to allow the author to

continually reassimilate organic chemistry, without

frustration, is truly unique and genuinely appreciated.

The author would also like to thank Dr. U. H. Brinker

for introducing him to the subtleties and the true art of

organic synthesis.

It must be stated that the identities of some of the

compounds contained herein might have remained unknown

without access to time and the talents of Dr. Roy King and

Tom Baugh.

The author would also like to acknowledge the financial

support of the National Science Foundation.

Finally, not to acknowledge the never ending line of

chemists who have passed through 413 CRB with explanations

and suggestions (both sought and unsought) would be unfair

and possibly plagiaristic.

















TABLE OF CONTENTS

Page
ACKNOWLEDGEMENTS ....................................... .. iii

ABSTRACT.............. ......... ................ ........ .... vii

ATTEMPTS TO GENERATE A HOMOAROMATIC CARBENE................ 1

Introduction......................... ........... ..... 1

Results and Discussion................................ 10

AN ATTEMPT TO GENERATE BISNORCARADIENE.................... 46

Introduction......... ... .. .......................... 46

Results and Discussion........ ....................... 49

EXPERIMENTAL............ .......... ........................ 54

General ......................................... 54

Preparation of tropone........................... 55

Preparation of 1-, 2-, and 3-chloro-
cycloheptatriene................................... 55

Preparation of zinc copper couple............... 56

Attempted reaction of 1-, 2-, and 3-
chlorocycloheptatriene with Simmons-
Smith reagent.................................... 57

Preparation of 7,7-dibromobicyclo[4.1.0]-
2-heptene ....................................... 58

Preparation of 2- and 3-bromocyclohep-
tatriene .......................... ............. 58

Preparation of 2- and 3-cycloheptatrienyl
formaldehyde ................... ............... 59

Preparation of cycloheptatrienyl formal-
dehyde tosylhydrazones.......................... 60









Preparation of the sodium salts of
cycloheptatrienyl formaldehyde
tosylhydrazones.................................. 61

Pyrolysis of the sodium salts of the
cycloheptatrienyl formaldehyde
tosy1hydrazones................................. 61

Photolysis of the sodium salts of the
cycloheptatrienyl formaldehyde
tosylhydrazones................ .................. 61

Preparation of 8-chlorobicyclo[5.1.0]-
2,4-octadiene (10).............................. 62

Reaction of 8-chlorobicyclo[5.1.0]2,4-
octadiene (10) with maleic anhydride............ 63

Preparation of 8,8-dibromobicyclo-
[5.1.0]2,4-octadiene (17)........................ 64

Reduction of 8,8-dibromobicyclo[5.1.0]-
2,4-octadiene (17) with tri-n-butyltin
hydride.......................................... 65

Reaction of 8-chlorobicyclo[5.1.0]2,4-
octadiene (10) with potassium t-butoxide,
general procedure .......... ..................... 66

Preparation of 8-chlorobicyclo[5.1.0]-
2,4-octadiene (10), enriched in 11 .............. 67

Reaction of 8-chlorobicyclo[5.1.0]2,4-
octadiene (10) with potassium t-butoxide
at 90 /1 mm........................................ 67

Reaction of 8,8-dibromobicyclo[5.1.0]-
2,4-octadiene (17) with potassium
t-butoxide at 00 ................................ 69

Preparation of 3-phenylphthalide................ 70

Preparation of 1,3-diphenylisobenzofuran......... 70

Reaction of 8,8-dibromobicyclo[5.1.0]-
2,4-octadiene (17) with potassium
t-butoxide and 1,3-diphenylisobenzo-
furan, general procedure......................... 71

Reaction of 8-chlorobicyclo[5. .0]2,4-
octadiene (10) with potassium t-butoxide
and ],3-diphenylisobenzofuran at -780........... 72









Reaction of 8,8-dichlorobicyclo[5.1.0]-
2,4-octadiene (17) with potassium
t-butoxide and 1,3-diphenylisobenzo-
furan at -78 ................................... 73

Preparation of 8,8-dichlorobicyclo-
[5.1.0]2,4-octadiene............................ 73

Reaction of 8,8-dichlorobicyclo[5.1.0]-
2,4-octadiene with potassium t-butoxide
and 1,3-diphenylisobenzofuran at -78............ 74

Reaction of 8,8-dibromobicyclo[5.1.0]-
2,4-octadiene (17) with potassium
t-butoxide and 1,3-diphenylisobenzo-
furan at 00...................................... 74

Preparation of 1,2,4-triphenyl-1,4-pro-
pandione.. ...................................... 75

Preparation of 2,3,5-triphenylfuran............. 76

Reaction of triphenylfuran with benzyne.......... 76

Reaction of 8,8-dibromobicyclo[5.1.0]-
2,4-octadiene (17) with potassium
t-butoxide with subsequent addition of
1,3-diphenylisobenzofuran ........................ 77

Preparation of coumalic acid.................... 78

Preparation of 2-pyrone ......................... 78

Preparation of the sodium salt of
tropone tosylhydrazone.......................... 79

Preparation of 6-chloro-2-pyrone................ 79

Preparation of 6-d-2-pyrone....................... 79

Pyrolysis of the sodium salt of tropone
tosylhydrazone in the presence of
2-pyrone......................................... 80

Pyrolysis of the sodium salt of tropone
tosylhydrazone in the presence of
6-d-2-pyrone ............. ....................... 80

Pyronlysis of the sodium salt of tropone
tosylhydrazone in the presence of 2-pyrone...... 81

REFERENCES ...................................... 82

BTOGRAPII I CAl, SKE'TCH .................. ............ 86

VI

















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

ATTEMPTS TO GENERATE A HOMOAROMATIC CARBENE
AND BISNORCARADIENE

by

Richard II. Parker

June 1979

Chairman: W. M. Jones
Major Department: Chemistry

A neutral homoaromatic species has not been reported in

the chemical literature. The similarity of the charge sepa-

rated resonance structure of 2,5,7-cyclooctatrienylidene, 1,

to the structure of the homotropylium ion implies that this

carbene might be a good candidate for a neutral homoaromatic

species.

The chemistry of the parent aromatic system, cyclo-

heptatrienylidene, suggests two methods of generating 1.

Base induced dehydrohalogenation of 5-chlorobicyclo[5.1.0]-

2,4-octadiene or 5-chlorobicyclo[5.1.0]2,5-octadiene should

yield an allene that could rearrange to the carbene, 1.

This carbene could also result from the carbene-carbene

rearrangement of 2- and 3-cycloheptatrienylcarbene. In this

study, the chlorides could not be generated and attempts to

generate the cycloheptatrienylcarbenes resulted in pyrazoles.


vi I


I










The cyclopropene, 2,4,8-bicyclo[5.1.0]octatriene, 2,

is the intermediate that would be expected in the carbene-

carbene rearrangement of the cycloheptatrionylcarbenes men-

tioned above. This species was generated from the reaction

of trans-8-chlorobicyclo[5.1.0]2,4-octadiene with potassium

t-butoxide and trapped with diphenylisobenzofuran. In the

absence of trapping materials, this reaction yielded cyclo-

octatetraene, heptafulvene and a trace of styrene.

It is postulated that the cyclopropene, 2, rearranges

to 1, which inserts into one of the transannular methylene

C-H bonds to yield 2,4,7-bicyclo[4.2.0]octatriene. This

species opens in its usual manner to yield cyclooctatetraene.

It is proposed that the other two products are the result of

a base induced isomerization of 2 to 1,3,5-bicyclo[5.1.0]oct-

atriene. This either undergoes electrocyclic ring opening

to heptafulvene or a norcaradiene rearrangement to a bicyclo-

butene, which opens to styrene.

Reaction of 8,8-dibromobicyclo[5.1.0]2,4-octadiene in

the same manner seems to follow a similar pathway. The

only products isolated from this reaction were 8-bromohepta-

fulvene, trapped with diphenylisobenzofuran, and 8-t-butoxy-

styrene. No bromocyclooctatetraene was noted, although the

cyclopropene precursor is definitely formed since it was

trapped with diphenylisobenzofuran. The styrene ethers

appear to have been formed from a small amount of starting

material that was able to doubly dohydrohaloqenato.


viii










An attempt was also made to synthesize bisnorcaradiene,

2,4,8,10-tricyclo[5.4.0.01'6]undecatetraene, 3. Cyclohep-

tatetraene was allowed to react with 2-pyrone to yield,

ultimately, 3,4-benzocycloheptatriene, 4. This reaction is

postulated to proceed through an intermediate pentaene, 1,3,

5,8,10-bicyclo[5.4.0]undecapentaene, 5. These pentaenes,

when substituted, are believed to undergo electrocyclic ring

closures to bisnorcaradienes. The ability of these bisnor-

caradienes to open to pentaenes with different substitution

patterns has led to the postulation of the intermediacy of

bisnorcaradienes in many chemical reactions in which substi-

tuent scrambling of this type was noted. When a deuterated

2-pyrone was allowed to react with cycloheptatetraene, sub-

stituent scrambling of the type mentioned above did not

occur. This reaction implies a competition between the

ability of 5 to close to 3 and the ability of 5 to aromatize

to 4 previously not noted.


_
















ATTEMPTS TO GENERATE A HOMOAROMATIC CARBENE

Introduction

Cycloheptatrienylidenel is an example of an aromatic

carbene. It appears that the vacant p orbital of the sin-

glet carbene has electron density from the six pi electron

system delocalized into it, in much the same way that elec-

tron density is delocalized into the vacant orbital in the

tropylium ion. The effect is evident in the reduced elec-

trophilicity of the system compared to that of other

carbenes lacking this conjugated pi system. The Hammett
9
rho value of typical electrophilic carbene, such as

dichlorocarbene (-0.619 at 800, -0.378 at 0), differs dras-

tically from that noted for this carbene (+1.05).2 The

increased stability of this carbene is also displayed in its

lack of reactivity with alkenes not containing electron

withdrawing substituents. In the absence of groups of this

kind, the carbene merely dimerizes to heptafulvalene.



1 -I O=


One of the more interesting properties of this carbene

is its participation in carbene-carbene rearrangements.3

This reaction of carbenes is one in which a carbene at a

vinylogous position can migrate to the adjacent carbon.










This interconversion of vinylcarbenes takes place through

an intermediate cyclopropene. Closure of the vinyl carbene

to the cyclopropene is a thermally allowed (2 +2 ) cyclo-
s a
C C
c-c c=c C-C

addition of the singlet carbene. Thermal opening of the

ring along similar lines leads to the formation of the new

vinylcarbene.

When phenylcarbene is generated at 2500 in the gas
4
phase, heptafulvalene is produced. This compound is postu-

lated to arise from rearrangement of the phenylcarbene to

cycloheptatrienylidene, through the bicyclic triene. The

cycloheptatrienylidene then dimerizes as described above.



\-'/ \- =


Rearrangements of substituted cycloheptatrienylidenes

to yield substituted pheny]carbenes have been noted also.

Diphenylcarbene5 generated in the gas phase yields fluorine,

presumably through a phenyl substituted cycloheptatrienyl-

idene.



C




While the interconversion of phenylcarbene and cyclo-

heptatrienylidene is a gas phase phenomenon, substitution of

this parent system can result in carbenc-carbene rearrange-










ments in solution. Decreasing the aromaticity of the

aromatic carbene lessens the impact on the delocalization

energy of the carbene when the cyclopropene intermediate is

formed, thus facilitating the reaction. An example is the

4,5-benzocycloheptatrienylidene system in which an annelated

ring is incorporated into the 4,5 bond of the cycloheptatri-

ene moiety. When formed by either photolysis or thermolysis,

it rearranges in solution to yield beta-naphthylcarbene.6

The annelated ring decreases the effect of the loss of

aromaticity when the cyclopropene is formed. The cyclopro-

pene intermediate in this reaction has been trapped in a

Diels-Alder reaction with butadiene, cyclopentadiene, and

furan.


S\ /7 q















Another interesting feature of this carbene is its

ability to equilibrate to its valence isomer, cyclohepta-

tetraene. Untch has found that dehydrohalogenation of a

mixture of 1-, 2-, and 3-ch1orocycl1ohoptariones yields

hptau l 9 his is oos o ro l throu an
heptafulvalene. This is proposudq to proceed throuI h an






4

initial dehydrohalogenation to the cyclic allene that

isomerizes to the carbene, which dimerizes. The intercon-

version of these isomeric intermediates has been shown to

be independent of their precursors. Regardless of whether

these intermediates are generated from thermolysis of the

sodium salt of tropone tosylhydrazone or the reaction of




Cl
S\ \KOC4Hq9




the above mentioned chloride mixture with potassium t-but-
i0
oxide, the allene can be trapped with isobenzofuran and

the carbene can be trapped with styrene.


Cl N-N-To

0C4H9 4H














Homoaromaticity is a term coined by Winstein to title

the effect noted in some carbonium ions in which unusual

stabilization is retained in ions in which conjugation has

been interrupted in a pi system that would otherwise be

aromatic. This effect, seen in many cationic species, has

not been noted in neutral species.









The interaction of the vacant p orbital with the pi

system in cycloheptatrienylidene mimics the interaction of

the vacant p orbital in the tropylium ion. Indeed, a reso-

nance structure of this carbene can be drawn in which the

tropylium ion is incorporated.


0





Homotropylium ions have been prepared and extensively
12
studied. The apparent structure of these ions seems to

be that of a non-classical monocyclic system, although some

arguments have been made toward a classical bicyclic system.

This ion suggests that it may be possible to generate a

carbene that would show this type of homoaromatic behavior.






This neutral species, which would have a resonance structure

that mimics the homotropylium ion, might be an exception to

the generality that homoaromaticity is restricted to charged

species.








There are two ways of studying the possibility of homo-

aromaticity in a cycloheptatrienylidene system. Replacing









one of the olefinic positions with a fused cyclopropane

ring would yield a carbene that would be analogous to the

bicyclic homotropylium ion above. This reaction has been

carried out by Oda et al. When the sodium salt of bicyclo-

[5.1.0]octa-3,5-dien-2-one tosylhydrazone was photolyzed or

pyrolyzed, the ensuing carbene immediately underwent ring

opening of the cyclopropane to yield 1,2,4,6-cyclooctatet-










raene, which dimerizes under the reaction conditions.

The other possibility is to introduce a methylene

insulating group into the pi system. This can be done in

one of two positions to yield either 2 or 3. It would be

expected that in isomer 2, the rapid 1,2 hydrogen migra-

tion to yield cyclooctatetraene would predominate over any







2 3

other carbene reaction. Isomer 3 cannot undergo this reac-

tion.

The rearrangement of the carbene generated by Kitahara,

to the allene, along with the known ability of cyclohepta-

tetraene, to isomerize to cycloheptatetraene, suggests that

isomerizaLion of 3 to its al ene form, 4, might he a dominant















3 4

reaction of this carbene. If this is so, then there would

be no opportunity to observe the reaction of the carbene.

Arguments can be made that suggest that rearrangement

of 3 to 4 may be less likely than the rearrangement of

Kitahara's carbene to the corresponding allene. Calculation

of the differences in the enthalpies of formation between

these allenes and the corresponding carbenes were made
14
using the group equivalents of Benson.4 Without correction

for allene strain or any aromatic or homoaromatic stabili-

zation, the difference in enthalpies (LAH ) between 3 and

4 was shown to be 10 Kcal greater than the difference

between Kitahara's carbene and allene. Intuition would lead

to the notion that the strain in the seven membered ring

of 4 would be greater than that in the eight membered ring

of Kitahara's allene. This would imply a greater AAlf than





S 4 3
dali0 f









the calculations indicate. This !'tHf between 3 and 4 is

5 Kcal greater than the corresponding difference in the

enthalpies of formation of cyclohce[.t trienylidene and cyclo-

heptatetraene, a system shown to be in equilibrium videe

supra). This would indicate that the rearrangement of the

carbene, 3, to the allene, 4, would be more endothermic, or

less exothermic, than the rearrangement noted by Kitahara,

and less likely to occur.

The chemistry of cycloheptatrienylidene suggests three

ways to generate the homoaromatic carbene, 3. The most

straightforward would be from thermolysis or photolysis of

the sodium salt of the tosylhydrazone of 2,5,7-cycloocta-

trienone, 5.

If homocycloheptatrienylidene exists in equilibrium



N---Tcs





5


with an allene isomer, a second method might be possible to

generate the carbene, through a dehydrohalogenation reaction

similar to the one used to generate cycloheptatrienylidene.

The allene isomer, 4, that would be the valence isomer of 3

could be generated from the dehydrohalogenation of the chlo-

rides above.

CI Cl

~ ) I\\ \\









The third possible pathway to the homoaromatic carbene

is through a carbene-carbene rearrangement. The cyclohep-

tatrienylcarbenes below have the opportunity to rearrange

in a similar manner to that of phenylcarbene, to give 3. In

principle, these carbenes could be generated by thermolysis

or photolysis of the sodium salts of the tosylhydrazones of

the corresponding aldehydes.

An evaluation of homoaromaticity in this system could

















be carried out along the same lines as those used to study

the aromaticity of cycloheptatrienylidene. If full or par-

tial retention of the ability to delocalize the pi electron

system into the vacant orbital of the carbene occurs, the

system can be studied by the nucleophilicity of the carbene.

Its ability to react with olefins reactive to normal elec-

trophilic carbenes can be tested. The ideal study would be

one in which the carbene is allowed to react with substituted

styrenes and its Hammett rho value determined. This should

give a quantitative measure of the nucleophilicity, i.e., the

aromaticity, of the the carbhne.









Results and Discussion

Because of the lack of a feasible synthetic route to

the sodium salt, 5, attempts were made toward the synthesis

of 3 via the rearrangement of isomeric materials. Isomeri-

zation of the allene, 4, would be the simplest route. The

required chlorides are simply the addition products of

methylene to l-chlorocycloheptatriene. Although this com-

pound, free of other isomeric chlorides, has been synthesized

by Untch,15 the synthesis requires many steps and results

in a low yield. A more economical method of preparation


Ci Ci
I





might be through separation of the isomers of 1-, 2-, and

3-chlorocycloheptatrienes, synthesized by Folisch6 from

tropone. Reaction of this material with Simmons-Smith

reagent should yield the desired compounds.

O Ci
C2 02C LArh C


CI


Chlorotropylium chloride was synthesized from tropone

by reaction with oxalyl chloride in methylene chloride.

This was reduced to the above mixture of chlorocyclohepta-

trienes with lithium aluminum hydride. Before an attempt

was made to separate the isomers, the mixture was tested









with Simmons-Smith reagent to insure its reaction. Copper-

zinc couple was prepared by the method of LeGoffl7 and its

reactivity with methylene iodide confirmed in the presence

of cyclohexene. Repeated reaction of the reactive couple

C1

C CHI


with methylene iodide in the presence of the chlorocyclo-

heptatriene mixture resulted only in the recovery of the

starting chlorides.

Attention was then diverted to preparation of 3 via



Br +r
0 C Br3, +- __ B
Br






S + CO .N-N-Tos

'C:O H C-N-N-Tos H
H H No


rearrangement of cycloheptatrienylcarbene. Lindsay and
18
Reese8 have synthesized 2- and 3-bromocycloheptatriene by

heating 7,7-dibromobicyclo[4.1.0]-2-heptene in quinoline.

Dibromocarbene was generated from bromoform by reaction with

potassium 5-butoxide in pentane, and allowed to react with

1,3-cyclohexadiene, yielding this bicyclic dihromide. This

was heated in quinoline, under vacuum, distilling off the









mixture of bromides as they were formed. This mixture was

then allowed to react with methyllithium in THF at -78.














c: \ "


Addition of dimethylformamide to this lithium reagent at

this temperature and hydrolysis with aqueous ammonium chlo-

ride yielded a mixture of 2-, and 3-formylcycloheptatrienes.

The tosylhydrazones were generated from reaction of this

mixture with tosylhydrazide in ethanol and the sodium salts

by reaction with sodium hydride in THF.

It can be seen from the scheme above that isolation of

pure isomers of these sodium salts is not necessary; both

carbenes generated from these salts should yield the same

homoaromatic carbene upon rearrangement.

This mixture of salts was thermally decomposed by re-

fluxing in diglyme. The resulting material was taken up in

water to dissolve the sodium toluenesulfinate formed and

the aqueous phase was extracted with pentane. No products

could be found in the pentane fraction. Extraction of the

aqueous phase with ether yielded an oil containing two com-

ponents. These were identified as C811 N2 isomers by high

rtesolut ion m slis spj cf1 1-0omn0 t I .










Assignment of the structures as pyrazoles 6 and 7 was

accomplished by nmr. Two signals appeared in the region

expected for resonances corresponding to the methylene pro-

tons. The triplet at 2.68* was assigned to the protons on

C4 of isomer 7 flanked by the protons on C3 and C5; while

the doublet at 3.47 corresponded to the hydrogens on C6 of

isomer 6, adjacent to the proton on C5. Two singlets at 7.5

and 7.6,corresponding to a single proton each, were assigned



N, N

H



N
_N- /\_N / '\ 6




r- 'NH 8









H
H NN HN q
H VN H 'N






N ,



to isomers 6 and 7, respectively, by comparison of the area

of the integrated signals with those already assigned to the




*All proton shifts are reported in units of .










methylene protons. These were determined to be the protons

on C8 by comparison with the spectra of substituted pyra-
19
zoles. These protons, in other pyrazoles, center around

7.6. The vinyl signals could not be assigned to their par-

ticular isomers.

Formation of these pyrazoles is explained in the scheme

above. Loss of sodium toluenesulfinate from the salt mix-

ture yields the corresponding diazo compounds. These can

then close to yield the indicated 3-H pyrazoles. Tautero-

merism via migration of the indicated proton yields 6 and 7

directly.

The formation of pyrazoles by this scheme is a well

documented barrier to the formation of carbenes from tosyl-
20
hydrazones. Bartlett and Stevens0 have noted the tendency

of diazo compounds generated from tosylhydrazones of

a,B-unsaturated aldehydes to close in this way to yield the

corresponding 3-H pyrazoles. While there seems to be no way

to overcome this difficulty in the thermal decomposition of

the sodium salts, low temperature photolysis21 has been

utilized successfully in some cases to yield carbenes from

systems that otherwise give only these pyrazoles.

When our salts were photolyzed through Pyrex in THF and

worked up in a manner similar to that described above, a

mixture of two new compounds was isolated. Mass spectra of

this material confirmed that these were also CH N2 isomers.

Again no evidence of compounds not containing nitrogen was

seen.









These new pyrazoles, 8 and 9, were shown not to be the

intermediate 3-H pyrazoles postulated above by the presence

of a N-H stretch in the infrared at 3200 cm1. Neither did

they rearrange to 6 and 7 by heating in refluxing diglyme.

The nmr did show a superimposed doublet and triplet between

2.7 and 3.0 that appear to be the protons on the methylene

carbons, and signals at 7.3 and 7.5 that appear to be the

protons on C The other signals, again, could not be

assigned to any isomer with any certainty.

These pyrazoles might result from a photoinduced

tauteromerism of 6 and 7. Tauteromerism of pyrazoles of
22
this type is well known.

Attempts to alleviate the cyclization problem by low

temperature photolysis failed.

Because of the difficulties involved in the low tem-

perature photolysis, a new approach to the problem was taken.
CI CI CI













IfI
HC




H









Billups and Reed23 has shown that it is possible to generate

both carbene participants of a carbene-carbene rearrangement

from the intermediate cyclopropene, formed by the dehydro-








KOR



10 /


halogenation of a di-halocyclopropane videe supra).

This suggests the possibility that it might be possible

to generate the homoaromatic carbene by dehydrohalogenation

of an 8-halobicyclo[5.1.0]2,4-octadiene.

Synthesis of 10 was attempted by the addition of chloro-

carbene to cycloheptatriene. This reaction yielded a yellow

oil. Elemental analysis and high resolution mass spectrom-

etry showed that this material was a C8H Cl isomer, as

expected. The nmr implied the presence of two isomers.

A multiple at 1.3-1.8 was assigned to the two bridgehead

protons and a multiple centered at 2.5 was assigned to the

methylene protons on C6. The vinyl protons were represented

by a multiple centered at 5.8. Two apparent triplets at

3.6 (J = 3 11z) and 3.0 (J = 8 Hz) appeared in the area of the

spectrum in which the proton on the carbon bearing the

chlorine should lie.






























L
____~=;

-----5
--~-5
~-L
-"----,
--


--

-L

r


cJ


C)
ra
V
4J
u
0





0





0


>1
u

0

0


u






,4


Ul
44











0

E
:3









4J
uu



C-)
CL
Lo





'4
a



'-

'-4

C)

'4
1









It was disturbing that the signals that correspond to

the protons geminal to the chlorine in these compounds ap-

peared to be triplets. Simple first order effects would lead

to the conclusion that these protons would be represented

as doublets of doublets if the carbene had inserted, as





CI


expected, into the 1,2 bond of the cycloheptatriene. If in-

sertion had occurred into the 3,4 bond of the cycloheptatri-

ene, the symmetry of the resulting chlorides would be expec-

ted to result in triplets corresponding to these protons.

Two proofs of structure for this molecule were attempt-

ed. If insertion of chlorocarbene in the 1,2 bond had

occurred to yield 10, the resulting molecule would contain a

conjugated diene. This diene would be capable of undergoing

a Diels-Alder reaction with a dienophile. When the chloride

was allowed to react with maleic anhydride, a solid material

was isolated that was shown by mass spectrometry and nmr to

be the expected adduct. Fragmentation of the molecule in

the mass spectrometer was indicative of its structure. The

molecular ion (m/e 238) lost CO and CO2 to yield a new ion

of mass 166. This fragmented to C 45 Cl+ and benzene (parent


01 0



;CI C CH C1










ion), implying the structure above. The nmr supported this

structure and the infrared showed the expected anhydride
-l
doublet at 1775 and 1850 cm1

Further support of 10 as the structure of this oil was

obtained by synthesis of a model compound. The known

isomer, 8,8-dibromobicyclo[5.1.0]2,4-octadiene, was synthe-
24
sized by the method of Van Vooren et al. This was al-

lowed to react with freshly prepared tri-n-butyltin hydride,


sBr r Br



to yield a mixture of monobromides. The nmr of this mixture

was identical to that of the mixture of chlorides.
25
Closs et al.25 have shown in studies of the 7-chloro-

bicyclo[4.1.0]heptane system that the coupling constant of

the H7 signal in the isomer, in which HI, H6, and the chlo-

rine atom are cis, has a coupling constant of 3 Hz. From

this information, the material isolated from the chlorocar-

bene reaction was assigned the structures 11 and 12. Isomer


Ci



11, in which the protons on the cyclopropane are assumed to

be cis (J = 8 Hz), is the major constituent, comprising 75%

of the isomer mixture. This mixture could not be separated

by chromatographic methods or spinning band distillation.












C0 CI


11 12

The mixture of 11 and 12 was subjected to five equiva-

lents of potassium t-butoxide in THF at reflux. Work up

with water and extraction into pentane yielded a small

amount of cyclooctatetraene and styrene as the only isolable

products. The cyclooctatetraene was determined not to re-

arrange to styrene under these conditions. The identity of

these compounds was confirmed by the nmr spectra and compar-

ison of their retention times to those of authentic samples

upon gas chromatography. The balance of the material re-

mained as an intractable polymer.


Cy CI KO r- CH H




Because of the low yield of products from this reaction,

an experiment was undertaken to determine if preferential

reaction of one of the isomers in the mixture was taking

place. To a solution of the above chlorides in THF at room

temperature was added only a half equivalent of potassium

t-butoxide. This was worked up with water and pentane to

ultimately yield an oil containing cyclooctatetraene and a

mixture of the above chlorides, heavily enriched (92%) in

isomer 11. The cyc]ooctatet racne was removed by careful

bulb to bulb distillation. This enriched mixture was then


















oD





0



0
0

-A



N
(U
0








OL




0 J
-0














r 0O
I






Hc



U
0-
C-
0 a

o


0.0



u)



OH


0


.,



CI








L-









allowed to react with five equivalents of potassium t-but-

oxide in the original manner. Only a trace of cycloocta-

tetraene was found after work up, an amount that could

easily be explained by reaction of the residual amount of

12 in the starting mixture. No residual 11 was isolated





11 12 11

from this reaction.

Attempts were made to trap either the cyclopropene or

one of the possible carbenes formed in this reaction. No

trapped products were seen using styrene, tetramethylethyl-

ene, t-butyl alcohol, butadiene or cyclopentadiene.

Cyclooctatetracne and styrene can be explained by a

scheme similar to that derived by Billups26 in his synthesis

of benzocyclopropene. He has shown that 13, when allowed to

react with potassium t-butoxide in DMSO, undergoes a dehydro-

halogenation to the cyclopropene, which rearranges to the

methylenecyclopropane moiety shown. Repetition of this

C >l

13

sequence yields benzocyclopropene.

A scheme in which an allyl hydrogen is removed by base,

with concerted loss of chloride ion to yield an intermediate

chlorocycloheptatriene has been discounted by labeling stud-
27
ics of Prestien and GunLther and the absence of hepta fu va-

ene,28 which would have been formed under these conditions.
lene, which would have been formed under these conditions.










RO H


Ci







Loss of HCI in 12 can occur in one of two ways. If

loss of the proton on C7 occurs first, the resulting C7-C8

cyclopropene has the choice of two possible pathways. The

homocycloheptatrienylidene, 2, would result from the ring

expansion of the cyclopropene to the carbene. This carbene

would be expected to rapidly rearrange to yield cycloocta-

tetraene. Base assisted isomerization of the cyclopropene
29
to the methylenecyclopropane29 yields the bicyclic triene,

14. This moiety can then undergo a norcaradiene rearrange-

ment to the bicyclobutane shown. This, when it opens in its




/" 2
/-





S14
12


'\ I Ig









usual manner, yields styrene.

Loss of the hydrogen on C1 to ultimately yield a C1-C8

cyclopropene, has the same options as the cyclopropene

postulated above. Base assisted isomerization again yields

14 and ultimately styrene. This time, ring opening to the

carbene yields homocycloheptatrienylidene, 3. This carbene

does not possess the ability to migrate hydrogen as 2. How-

ever, it is apparent, upon study of molecular models, that

the empty orbital of the carbene and the hydrogen on the

methylene carbon are in excellent alignment for insertion of

the carbene into this C-H bond. Intramolecular insertions

of carbenes into carbon-hydrogen bonds in saturated and un-

saturated rings are not only well known, but this reaction

is competitive with possible 1,2 hydrogen migrations.30

This insertion gives the bicyclic octatriene shown. This

species is the well known valence isomer of cycloocta-

tetraene. Simple electrocyclic opening of the cyclobutene

moiety results in cyclooctatetraene.

It is not unexpected that the above carbenes were not

trapped by any olefin or alcohol. The inability to trap

carbenes of this sort has become a generality.

With few exceptions, it can be stated that if a
singlet carbene has an internal insertion or cyclo-
addition available, it will not be possible to trap
it with an external reagent.31

A direct opening of the chloride by reaction with the

base appears extremely unlikely in light of the labeling

studies in the benzocyclopropene system mentioned above.









However, in the case at hand, it is not possible, due to the

unique symmetry of the cyclooctatetraene molecule, to carry

out labeling studies of the type performed by Prestien and


RO





Gunther to disprove this pathway.

It is apparent from the low yield of cyclooctatetraene

and styrene that some other route is available for reaction

of the chloride. An interesting similarity exists in

intermediate 14 and intermediate 15, below. Washburn et
32
al.2 has studied the dehydrohalogenation of the dibromide,

16, and determined, by studies using labeled dibromides, that


R-O H
BBr ROH OR
B- -"- < ^h
16 15

the reaction proceeds via the route shown below. Rearrange-

ment of 15 to 8-t-butoxyfulvene is a simple electrocyclic

ring opening of the methylenecyclopropano. If 14 does open,

or could be induced to open in an analogous manner, it would




14



yield heptafulvene. This highly reactive material would

polymerize rapidly under the reaction conditions.









An attempt was then made to induce reaction of the

chlorides 11 and 12 to heptafulvene while minimizing de-

struction of any products formed. The reaction was carried

out in a solution of potassium t-butoxide in tetraglyme,

heated to 900 under vacuum to distill off volatile products

into a liquid nitrogen cooled trap. Immediately upon injec-

tion of some of the mixture of chlorides into this solution

via syringe, a scarlet material was noted in the liquid

nitrogen trap. Upon warming, the scarlet material polymer-

ized. Gas chromatography showed the presence of cycloocta-

tetraene and styrene in the trapped material.

An attempt was then made to trap this material by the

method of Doering and Wiley.33 The above reaction was re-

peated, and the contents of the cold trap were washed into

cold THF containing dimethyl acetylenedicarboxylate. After

standing at 00 for 72 hours, the red color had disappeared.

Exchange of the solvent with benzene and air oxidation using

Pd/C catalyst yielded a blue solid. This was identified,

by comparison of its uv with the literature, 33 as dimethyl

azulene-1,2-dicarboxylate.

When the red material was trapped in an nmr tube

.- .Es
-s-CE Es Es

Es = COOMe

containing THF and a trace of triethylamine, an nmr was

taken that showed signals corresponding to those reported

fr heptafulve 34
for heptafulvene.









When this reaction was repeated with pure 11 under the

same conditions, a very small amount of colored material was

obtained, not enough to identify by the methods outlined

above.

It is obvious from the preceding reactions that only

one isomer, 12, is yielding isolable products. Preferential

reaction of this isomer in which the hydrogen and chloride

to be eliminated are coplanar might be expected. It is

unusual that reaction of isomer 11 does occur, albeit slowly,

but does not give the same products.

No evidence can be found in the literature on the

relative rates of elimination between the isomers of mono-

halocyclopropanes. While the number of reactions is

small,35 one set of experiments has shown that elimination,

when the hydrogen vicinal to the halide is trans, can occur

and yields the cyclopropene.36 The reaction pathway of

isomer 11 cannot be suggested, since the only material iso-

lated from its addition to base is an intractable polymer.

In order to increase the yield so that the reaction

Br -e



could be more readily studied, a system was then sought in

which the configuration of the eliminated moieties could be

assured to be cis. The simplest system of this kind would

be an 8,8-dibromobicyclo[5.1.0]2,4-octadiene, 17. In this

molecule, one of the halogens will always be cis to the

hydrogen to be eliminated.










When this molecule was subjected to potassium t-butoxide

in THF at 0 and worked up with water and pentane as outlined

above, a very small amount of two compounds could be

isolated. These were separated by TLC and their structures

identified as cis and trans-B-t-butoxystyrene by nmr and

mass spectrometry. No bromocyclooctatetraene was noted in

this reaction.

Since the mechanisms of these reactions were merely



'Sr sAt
S OR OR
17

speculative at this time, it was then set about to confirm

intermediates along these suspected pathways. The first

intermediate, one common to both systems, is the cyclopro-

pene formed from the initial dehydrohalogenation.

The first attempt was made to trap this cyclopropene

by a Diels-Alder addition to 1,3-diphenylisobenzofuran,

34.37 It has been shown that 34 is a potent dine in

Diels-Alder reactions. Its large molecular weight would

make isolation of the trapped material possible by crystal-

lization.






34

The mixture of chlorides, 11 and 12 (1 equivalent of

12), was allowed to react with five equivalents of potassium










t-butoxide and one equivalent of 34 at -780 in TIF. Along

with some cyclooctatetraene and a great deal of 11, a crys-

talline material was isolated after work up with ether.

Chromatography and recrystallization allowed isolation of

the crystalline material which was identified as the 34

adduct of the cyclopropene, 18. Elemental analysis and high

resolution mass spectrometry confirm the C28H220 formula.

The proton spectrum showed the expected shifts for this

molecule, a complex multiple for the aromatics from 7.0-8.0

(14H), a vinyl multiple at 5.8-6.2 (4H), a doublet at 1.7

(1H) and a complex multiple from 2.1-3.2 that corresponds

to three protons. Confirmation of the exact structure could

not be made from this spectrum alone. As mentioned above,

dehydrohalogenation could occur across either C1-C8 or C7-C8

to yield the cyclopropenes below. Either of these materials,






1 18






when trapped with 34, could give signals of this type in

the proton spectrum. The complex coupling precluded a sim-

ple assignment of structure.

Spin decoupling allowed the assignment of this molecule

as structure 18 to be carried out. The complex multiple

from 2.1 to 3.2 was suspected to be formed from the signals

















0






24



0

_0




rn

I m
0
N
'C)












c Q




.0 4
u r
0












Q) 4
u



).h







ma

S.
a-,

'H



0
0C.


E


I


1















J 2
12

of the methylene protons (H6) and the bridgehead proton (H7)

When the doublet at 1.7, assigned to H11, was irradiated, the

far downfield portion of the multiple simplified to a doub-

let of doublets (J=4.5, 9 Hz.). This would denote that this

proton (H7) while being adjacent to H was also being split

by two non equivalent protons (H6a,b). Irradiation of the

downfield portion of the vinyl signal (H5) resulted in two

overlapping doublets of doublets (J=4.5, 14 Hz.; 8.5, 14 Hz.)

in the upfield portion of the 2.1-3.2 multiple. These two

non equivalent protons (H6a,b) were being split, not only by

an adjacent proton (17 ), but also by each other. This is the

reason for the identical coupling constant (J-14 Hz.) in

each of these patterns. This type of coupling cannot be ex-

plained by a product resulting from the addition of 34 to

the C -C cyclopropene.

Similar reaction of 8,8-dichlorobicyclo[5.1.0]2,4-octa-

diene with five equivalents of potassium t-butoxide and iBF

results in a crystalline material after the same type of

work up. An almost identical material is isolated from the

reaction of 17 in a similar manner. Each of these compounds

is a C28I2 OX isomer with almost identical proton spectra.

These isomers were assigned structures 19 and 20. The

molecular formulas were confirmed by elemental analyses.











I.
0
U
H
.0
0
0
i 0
U

cn

VH
o0

U
u4-
0





,.




rlJ


,.0

0ri
4-,
K
U~








I-,~
) 0









The proton spectrum of the bromide, 19, as in the case of 18,

could not unequivocally establish the structure. Each of the

protons could be assigned to a signal but the splitting was

not simple enough to define the structure exactly. A two

proton multiple centered at 2.6 was assigned to the methyl-

ene protons on C6 and a multiple at 3.2 was assigned to H7,

each of these signals appearing at a similar shift to the





S/ 19 X=Br
6 20 X=C1




corresponding protons on 18. The aromatics again appear in

a multiple from 5.9-6.5. A glaring deviation from the

vinyl signals of 18 existed in the appearance of a new doub-

let at 5.5. This signal could be assigned to H2. A more

complete discussion will appear later.

The ability to work with a greater amount of the reac-

tive isomer videe supra) in the dibromide system, resulting

in a greater yield of trapped material, allowed studies of

the carbon-13 spectrum. The distinct dissimilarity of the

proton vinyl signals raised a real question as to the simi-

larity of structures assigned to this bromide and the

hydrocarbon, 18. The absence of bromocyclooctatetraene in

any reaction of the dibromide, 17, with base suggests that

rearrangement of the cyclopropene to the methylenecyclopro-









pane must be a rapid reaction. If this is true, then this

trapped bromide might be formed from the methylenecyclopro-

pane rather than the cyclopropene.

The bromide expected from the trapping of a mcthylene-

cyclopropano with 34 would be 21. This could be easily

differentiated from the trapped cyclopropene by the fact that

this molecule contains three methyne carbons (Ci, C3, C4) and

one quaternary carbon in what would be expected to be the up-

field portion of the carbon-13 spectrum. Structure 19 con-





o

i 21





tains one methylene carbon (C6), one methyne carbon (C7), and

two quaternary carbons (C1, C ) in the corresponding portion

of the carbon spectrum. The carbon-13 spectrum of this bro-

mide showed a triplet at 27.873 (ppm downfield from tns), a

doublet at 35.401 and two singlets at 41.372 and 48.461. The

triplet was assigned to the methylene carbon (C6) on 19, the

doublet to the methyne carbon (C7) and the singlets to the

quaternary carbons (C C ).

Final confirming evidence for 19 as the structure of

this solid was found in the spin decoupled proton spectrum.

When the signal at 2.6 (I6 ) was irradiated, the vinyl multi-

plet split into a doublet and a triplet, denoting that the









protons on C6 are the only protons adjacent to the vinyl

system.

The dissimilarity in the vinyl signals of the proton

spectra of 18 and 19 can be explained when models of these

two systems are constructed. In structure 18, there are

two conformations with respect to ring flip of the seven-

membered ring. In one of these conformations, one of the

protons on C6 is almost coplanar with H In this conforma-

tion, the proton on C2 is removed from the proximity of the

neighboring phenyl ring, eliminating any steric interaction.

When the same conformation is attempted in compound 19, the

proton on C6 is now almost coplanar with the bromine. In-

deed, this exact conformation cannot exist because of the

steric bulk of this halide. A ring flip to the other con-

formation eliminates this steric problem, but in doing so,

forces the proton on C2 into the shielding area of the neigh-

boring phenyl ring. This, in turn, shifts this proton

signal upfield, generating the doublet at 5.5 that was noted

in the spectrum of 18.

The next step, following the proposed mechanism, would

be to trap the methylene cyclopropane, if possible. This

was attempted in the dibromide system because of the afore-

mentioned reactivity of this material. At 00, one equiva-

lent of 17 was allowed to react with five equivalents of

potassium t-butoxide and one equivalent of 34. Again a

crystalline material was isolated in the same manner as be-

fore. The structure of this material was determined to be


















4i

a


a


0
:4
a
I 'a
'a
a
a
ao
a
0
N,
a



* a
0



a
4,!









a

a ma


421
a
a
0.


a,






(NI


~ U';
a
ar
i ca










22. High resolution mass spectrometry confirmed a C28H21 OBr

formula. The proton spectrum showed no upfield protons,

implying that the cyclopropane ring had been broken. The

vinyl signals consisted of a six proton multiple between

5.2-6.5, the aromatics, a fourteen proton multiple between

7.1-8.0 and the remaining signal was a singlet, corresponding

to one proton at 4.5.

The carbon-13 spectrum supported this structure. The

upfield portion of this spectrum indicated four sp carbons.

Two of these signals could be assigned to the bridgehead

carbons of bicyclic ether by their similarity to the respec-

tive carbons in structure 19. The only sp signal that de-






22




noted a proton bearing carbon was a doublet at 66.101 that

was assigned to C The remaining quaternary signal (56.898)

was assigned to C2.

This molecule is a labile material and readily aroma-

tizes with loss of HBr to yield a more stable new material.

This material, a C282H 0 isomer, 23, is isolated after

chromatography on silica gel of the crude material from the

reaction of 17 with potassium t-butoxide and 34 at 0,

although it cannot be detected in the nmr of the crude

material before chromatography. This halogen-free substance







33

is the only material isolated from a solution of 22 in aceto-

nitrile allowed to stand at room temperature for two days.

Spectral data allowed structure 23 to be assigned.

Elemental analysis and high resolution mass spectrometry

confirm the C 2811200 formula. The infrared spectrum indicated
-i
a carbonyl stretch at 1670 cm corroborated by a large loss

of Ph-Co(+) from the molecular ion in the mass spectrum. The

proton spectrum was simple, containing a multiple between

6.9 and 7.7, implying a high degree of aromatization and no

indication of protons on vinyl positions. The only other

proton signal was a sing3et at 5.5 corresponding to one pro-

ton. The only assignable signals in the carbon-13 spectrum

were the carbonyl carbon at 198.97 and a proton bearing

carbon at 63.44.

These data can be fit to a 1,2,3 trisubstituted indene

in which the substituents are two phenyls and a benzoyl.

Synthesis of l-carbomethoxy-2,3-diphenylindene has been

reported38 and the proton on C1 is a singlet at 4.8. This

molecule was used as a basis for calculation of chemical
39
shifts by the additivity rules of Shoolery.3 The additivity

factor was taken as the difference in the shifts of the



23 /



methyl signals of acetophenone and methyl acetate (+.6 ppm).

Using this information, the calculated shift for this proton

in l-benzoyl-2,3-diphenylindene would be 5.4.



















































C)
'U




H










a)


'Hl







N
U

S







C)
N,










The isolation of 23 from 22 in acetonitrile shows that

a solvolytic dehydrohalogenation is one possible mechanism.

Loss of the bromide would yield the carbonium ion below.

Isomerization of the cycloheptatriene ring to a norcaradiene

yields a cyclopropylcarbinyl cation. Interconversion of this

carbon to the allylcarbinyl cation results in a species that

can lose a proton to aromatize. This new compound, 24,

could protonate on the oxygen and rearrange, as outlined

below, to 23.

Compound 24 should be stable in the absence of acid.

If solvolysis of 22 occurs under the basic conditions of the

reaction in which 22 is generated, 24 may be an isolable

product.

Benzyne was allowed to react with 2,3,5-triphenylfuran

to yield a colorless solid whose nmr and mass spectra are

compatible with the structure of 24. The proton spectrum of


















'i it 23
24

OH



q, 1
^2










this material consists of a multiple between 6.9-7.9. The

carbon spectrum shows the bridgehead carbons (92.416 and

94.950) at positions comparable with those on 19 and 22.

The rest of the spectrum cannot be assigned.

The intermediacy of this material in the solvolysis of

22 was refuted by its stability to silica gel chromatography

and its inertness in acetonitrile solution. Refluxing this

material in acidic methanol produced no 23.

With this in mind, it is more likely that the actual

mechanism of the solvolysis of 22 resembles that below. Loss

of bromide ion yields a non classical carbonium ion that




o 24



opens, as indicated, and rearranges to the ketone. The cy-

clopropylcarbinyl cation formed from rearrangement of the

cycloheptatriene to norcaradiene, rearranges to the allyl-

carbinyl cation, which aromatizes to 23.

Dehydrohalogenation in the systems of 12 and 17 seem to

follow a similar pathway. The first step is dehydrohalogen-

ation across the C1-C8 bond to yield the cyclopropene. This,

in the case of the halogen free cyclopropene, 25, has the

ability to open to the homoaromatic carbene 3, which can

intramolecularly insert into the C-H bond of the methylene

across the ring to form the bicyclic triene, which can then

open to cyclooctatetraene. Competing with this reaction is













+ Q~&
/1


o-0


4<
C ;O
r1
-I


S C
/ -


23


the base induced isomerization of the cyclopropene to the

methylenecyclopropane. This moiety can apparently undergo

electrocyclic ring opening to yield a heptafulvene. Both of

these heptafulvenes have been trapped, 28 with the acetylene

diester yielding the substituted azulene, and 29 with jBF

to yield 22. The styrene formed in the reaction of 12 with

base in the absence of trapping material appears to arise

from a bicyclobutane formed by norcaradiene rearrangement




12 X=C1 Y=II 25 Y=H
17 X=Y=Br 26 Y=Br


14 Y= H -
27 Y=Br '


Y
If

S-k2 Y=i2
\-/ 29 Y-Br


K1 I()1


9r
o










of 14. This would be consistent with the increase in yield

of styrene noted by increasing the temperature of the reac-

tion.

The absence of bromocyclooctatetraene in the reaction

of 17 with base must be recalled. It is unlikely that the

difference in the eliminated halide is changing the course

of the reaction, since trapping of the halocyclopropene was

accomplished regardless of the halide eliminated. The dif-

ference in these reactions is most likely caused by the rate

at which the cyclopropene isomerizes to the methylenecyclo-

propane. The intermediate anion in this isomerization can

be stabilized by resonance through the pi system and onto the

carbon bearing the halide. The halogen can stabilize the

anion at this position by induction. The overall effect is

to enhance the ability of the cyclopropene to isomerize

through this intermediate. This allows the cyclopropene to

isomerize before opening to the carbene can occur.

The question of the genesis of the styrene ethers has

still not been addressed. These ethers are only apparent in

the reaction of 17 with no trapping material. At no time

during the isolation of 19 or 22 were either of these com-

pounds noted. It is apparent that to isolate these compounds,

neglecting nucleophilic substitution of the t-butoxy group



x









for bromide, that two dehydrohalogenations must have

occurred. Two plausible mechanistic pathways can be pos-

tulated.

The cyclopropane previously shown to have been found

in this reaction can either be trapped with t-butyl alcohol

to yield 30 or rearrange to 27. A second dehydrohalogenation

of 30 with accompanying rearrangement to the methylenecyclo-

propane yields 31. This intermediate can also be formed

from a dehydrohalogenation of 27 to yield the tetraene, 32,

which can be trapped by t-butyl alcohol in a manner remini-

scent of that of the trapping of m-benzyne to yield 15.

Compound 31 can then aromatize, in the manner postulated

earlier for the formation of styrene from the monochloride





17 27 32

I I










system, to yield the styrene ethers.

To trap the progenitor of the ethers would require de-

lay of inclusion of the trapping material until the bromo-

cyclopropene could react further and the second dehydrohalo-

genation to either 32 or 33 could occur. The reaction must










be done at a low temperature to prevent the ring opening to

bromoheptafulvalene known to occur at 0.

At -780, the dibromide 17 and five equivalents of po-

tassium 5-butoxide were stirred for one hour, then, while

still at this temperature, one equivalent of 34 was intro-

duced. When the reaction had warmed and was worked up with

water and ether in the usual way, a large amount of polymer

was noted. The only isolable product was a small amount of

22.

It is apparent that trapping of the doubly dehydro-

halogenated material is not taking place. This is unfortu-

nate in that identification of this material, which must be

a precursor of the styrene ethers, cannot be made.
















AN ATTEMPT TO GENERATE BISNORCARADIENE

Introduction

Bisnorcaradienes, 2,4,8,10-tricyclo[5.4.0.0 1'6lundeca-

tetraenes, 35, have been postulated as intermediates to

explain the scrambling of substituents in substituted ben-

zocycloheptatrienes. These interesting intermediates were

first postulated in the thermal isomerization of di-deutero-

3,4-benzocycloheptatetraene by Gruber and Pomerantz,41 where

it was noted that deuterium substituted on the cyclohepta-




:J I X
36 35 37

triene ring of 3,4-benzocycloheptatriene (37; R1, R = D;

R2, R4 = H) would rearrange to the benzene portion upon

heating at 4050 for 10 min. They postulated that the reac-

tion proceeded through the scheme above.

The next system thought to involve bisnorcaradienes was
42
that of Durr et al. in which a cyclopentadienylidene was

allowed to react with benzene to ultimately yield a spiro

compound that was thought to rearrange to the bisnorcara-

diene by photolysis.

This photolytic reaction of the spiro compound was

thought to proceed through the di-pi-methane rearrangement

to yield the bisnorcaradiene directly.

4 1











H/ H, / ,

N /35


43
In 1974, Durr et al. were able to isolate a material

from the photolysis of the spiro compound, below, at -700

This material was identified by its nmr at -600 as the



r k (3

bisnorcaradiene. When this material was warmed to -100, it

isomerized to the benzocycloheptatrienes.

Also during that year, Mitsuhashi and Jones44 isolated

substituted benzocycloheptatrienes that were postulated to

have been formed via an intermediate bisnorcaradiene, from

the addition of cycloheptatrienylidene to tetracyclones.




/ -- 35 (R R4 = R )

454

In 1976, Pauly and Durr45 reported the isolation of a

material that he identified by proton nmr as the first
46
stable bisnorcaradiene. This material, which melts at

1040, has a nmr spectrum similar to that of the material46

isolated in the low temperature work mentioned above.



0 R =

F "-)(RI R 4







48

A possible property of this intermediate that has not

been explored is the potential of 35 to undergo a degenerate

[5,5] antara, antara, sigmatropic rearrangement. This

reaction would be a symmetry allowed, concerted reaction.

It was with this possibility in mind that the system





R-- 35 (R2 R. =benzo)



47
outlined below was devised.4

This system would allow studies to be made of the

possibility of bisnorcaradiene formation in the parent

system at modest temperatures, and determination of the







ability of this bisnorcaradiene to undergo this degenerate,

semibulvalene type of rearrangement. The known first order






Er/ D
39b
39a






D / 7D


39c 39b










proton spectrum of 3,4-benzocycloheptatriene would allow

simple assignment of the positions of the deuterium atoms in

the rearranged molecules.

Entrance to this system would be made through compound

38. This 1,3,5,8,10-bicyclo[5.4.0]undecapentaene would have

the potential to rearrange to bisnorcaradiene in the manner

of the substituted analogues noted above. Compound 38 could

be formed by the addition of cycloheptatrienylidene-cyclo-

heptatetraene to 3,6-di-deutero-2-pyrone. The resulting

lactones could be decarboxylated to yield 38.

If the bisnorcaradiene were formed in this system,

the resulting benzocycloheptatrienes should show deuterium

in the cycloheptatriene ring (39a,39b). This effect would

be evident in the changes in the splitting patterns of the

vicinal hydrogens. If the [5,5] rearrangement should take


0




/ 38


place, a new splitting pattern for H -If6 should become

evident due to the introduction of a new molecule with

deuterium at C7 (39).


Results and Discussion

Coumalic acid, prepared by the method of Wiley and

Smith from malic acid, was decarboxylated at 6500 through
copper turnings to yield -pyon49
copper turnings to yield 2-pyrono.








50

Attempts to react 2-pyrone with cycloheptatetraene

generated from chlorocycloheptatrienes and potassium t-bu-

toxide50 were unsuccessful due to the instability of the

2-pyrone in base. The 2-pyrone was reactive to cyclohep-



0 0
l0 0
On O- 0--o 0
HOOC-C2 CCH-COO /
Coo)

tatraene generated from the thermolysis of the sodium salt
51
of tropone tosylhydrazone. When this salt was added to a

solution of ten equivalents of 2-pyrone in diglyme at 110,

3,4-henzocycloheptatriene was isolated, after work up with

water and pentane, in 30% yield.

Similar results were seen after photolysis of the

sodium salt through Pyrex in THF in the presence of ten

equivalents of 2-pyrone. Again, 3,4-benzocycloheptatriene

was isolated as the major component, but 1,2-benzocyclo-







heptatriene was also formed. These results were noted in

photolyses at temperatures as low as -780. Photolyses were

not as clean as the thermolyses, and polymeric materials

were also isolated.

The mechanism of th is reaction should be that proposed

above, addition of the allene to the pyrone to yiuld the

lactone and decarboxylation to the bicyclic pentaune.













N



Hydrogen can migrate to yield the benzocyclohcptatriene, or

the ring can close to the bisnorcaradiene.

The presence of the 1,2-benzocycloheptatriene appar-

ently results from a photoinitiated 1,3 hydrogen shift in

the bicyclic pentaene.


10 ni











It was necessary to use an excess of 2-pyrone in the

photolysis reaction because of the photolability of this
52
molecule. It can undergo [2+2] cycloaddition intramolecu-

\/







larly to yield a lactone that decarboxylates to cyclobuta-

diene.









52
Chapman et al. have been able to synthesize di-deu-

tero-2-pyrones by sequential deuteration of the pyrone, from

methods developed by Pirkle and Dines.5



(OU1) I C)1 (I




Synthesis of 6-d-2-pyrone, 40 was accomplished from

glutaconic acid by the method shown above. Before the

second deuteration was attempted, this material was allowed

to react with cycloheptatetraene to determine if rearrange-

ment through a bisnorcaradiene was taking place. Formation

of the bisnorcaradiene would cause some of the deuterium to

become incorporated into the cycloheptatriene portion of

the product, after rearrangement.

Reaction of an equivalent amount of sodium salt with 40

in diglyme at 1100 yielded the benzocycloheptatriene, as

expected. There was no change in the splitting patterns of

the proton spectrum, although the mass spectrum showed that

the deuterium had been retained. Integration of the proton

spectrum indicated that all of the deuterium was incorpor-

ated in the benzene portion.

This implied that the 1,5 hydrogen migration, resulting

in the aromatization of the pentaene was much faster than

the closure to a bisnorcaradiene. Since this migration is

thermally initiated, an attempt was made to generate the

cycloheptatetraene by photolysis of the sodium salt at -78,

as before. The scarcity of the deuterated pyrone precluded










the ability to work with the excess of starting material

enjoyed in the reaction of the non-deuterated pyrone.

Photolysis of an equivalent amount of the sodium salt and

40 at -780 did not yield enough benzocycloheptatriene to

identify. No pyrone was recovered after the reaction.

An attempt was then made to determine if the closure

to bisnorcaradiene could become competitive with aromatiza-

tion by increasing the temperature of the reaction. At

temperatures as high as 2000, aromatization remained as the

only observable pathway of the pentaene.

It appears that the energy required to drive the pen-

taene toward aromatization and the energy required for

closure to a bisnorcaradiene must not differ greatly. The

effect of aryl substitution in this system is apparently

enough to induce bisnorcaradiene formation by either retar-

dation of aromatization or acceleration of closure. The

nature of this is not obvious.
















EXPERIMENTAL

General. All melting points were obtained on a Thomas

Hoover melting point apparatus and are uncorrected. Nuclear

magnetic resonance spectra (60 MHz) were recorded on either

a Varian A-60A or JEOL PMX 60 spectrometer and are reported

in ppm (3) downfield from tetramethylsilane, the internal

standard. Carbon-13 and 100 MHz proton spectra were recorded

on a JEOL FX 100 spectrometer and are reported in ppm down-

field from tetramethylsilane. Infrared data were recorded on

a Beckman IR 10 spectrophotometer. Combustion analyses were

performed by Atlantic Microlab, Inc., Atlanta, Georgia.

Mass spectra were obtained on an AEI MS 30 spectrometer.

Cycloheptatriene was purified by distillation through

a column of glass beads. All tetrahydrofuran (THF) was

distilled from lithium aluminum hydride immediately before

use. Diglyme and tetraglyme were distilled from calcium

hydride and stored over sieves until used. Dimethylformamide

(DMF) was stirred for 24 hrs over 3A molecular sieves and

calcium hydride, decanted onto barium oxide and distilled at

20 mm into septum bottles containing molecular sieves.

Other commercially obtained reagents were used as received.

Chromatography was carried out on Baker 60-200 mesh

silica gel or Fisher basic alumina. The progress of column

chromatography was followed by the addition of 1F type








55

609 phosphor (DuPont Photoproducts Department) to the

adsorbent, employing a quartz column, and observing with a

hand-held UV unit (Fisher Scientific).

Materials noted as vacuum dried were placed in a vacuum

oven at 130 for at least 12 hours.

Preparation of tropone. In a 2 liter round-bottomed

Morton Flask equipped with a mechanical stirrer and a reflux

condenser was stirred 13.5 g (0.10 mol) of KH2PO4 in 33 ml

of water and 350 ml of dioxane. To this was added in one

portion, 46 g (0.50 mole) cycloheptatriene and 54 g

(0.49 mol) of selenium dioxide. This suspension was then

heated to 900 in an oil bath and stirred for 16 hours.

After the mixture had cooled to room temperature, the

residual selenium was filtered and the supernatant liquid

added to 700 ml of water. This was then extracted four times

with 250 ml of methylene chloride, which was washed with

250 ml of 10% sodium carbonate solution and dried with MgSO4.

The methylene chloride was removed by rotary evaporation, and

the residual material was distilled under reduced pressure.

A yellow liquid was obtained with a boiling point of 90-2
54
(3 mm) (22 q/41%) [lit.5 91-20 (4 mm)].

Preparation of 1-, 2-, and 3-chlorocycloheptatriene.1

In a 100 ml three-necked round-bottomed flask equipped with

a drying tube addition funnel and a nitrogen inlet was

stirred, in an ice bath, 2.0 g of tropone (19 mmol) in

12 ml of CH2CI2 under nitrogen. To this was added dropwise,

over 15 minutes, 2.5 g (20 mmol) of oxallyl chloride in 4 ml









of CH2Cl2. The resulting slurry was stirred for an addi-

tional 35 minutes. With minimal exposure to air, the solvent

was removed by rotary evaporation.

In the same flask, the yellow residue was dissolved in

45 ml of dry THF and the flask was equipped with a drying

tube, a nitrogen inlet, and a solid addition tube containing

1.49 g of lithium aluminum hydride. The system was purged

with nitrogen and cooled in an ice bath. The lithium

aluminum hydride was added in small portions, with stirring

over 15 minutes. The resulting suspension was stirred for

an additional hour and allowed to come to room temperature.

The excess hydride was destroyed with water and the precip-

itated material filtered. The mother liquor was extracted

with 250 ml of CH2C12 and dried with MgSO4. The solvent was

removed by rotary evaporation and the resulting liquid puri-

fied by column chromatography on silica gel using pentane as

the elutant. The resulting yellow oil (1.5 g, 60% yield)

was a mixture of 1-, 2-, and 3-chlorocycloheptatriene, that

had the nmr spectrum reported in the literature.16

Preparation of zinc copper couple.7 To 2.0 g of cupric

acetate dissolved in 50 ml of glacial acetic acid in a 125 ml

Erlenmeyer flask heated to 1000 and stirred magnetically,

was added in one portion, 35 g of zinc dust. The slurry was

shaken well and an additional 50 ml portion of acetic acid

was added. When the material had cooled, the acetic acid was

decanted from the resulting red-grey solid. This solid was

washed three times with 100 ml of ether and stored under ether.







57

The reactivity of the above couple was tested by reac-

ting it with methylene iodide in the presence of cyclohexene.

In a 50 ml round-bottomed flask equipped with a reflux con-

denser and magnetic stirrer, was placed ca. 7 g of the wet

couple and 20 ml of ether. To this was added ten drops of

CH2I2 which resulted in evolution of bubbles from the sol-

vent. A mixture of 5 g of cyclohexene and 13 g of CH2I2

was then added dropwise over 15 minutes. When the initial

reaction subsided, the material was refluxed overnight. The

slurry was then filtered and the solid washed with 25 ml of

IN HC1, which was added to the ether. The layers were

separated and the ether was washed three times with water

and dried with MgSO4. The nmr spectrum of the residual

material after removal of the ether by distillation showed

the presence of norcarane.

Attempted reaction of 1-, 2-, and 3-chlorocyclohepta-

triene. In a 50 ml round-bottomed flask equipped with a

reflux condenser and a magnetic stirrer was placed 320 mg of

the above couple in 20 ml of ether. To this was added 0.5 g

(4 mmol) of the chloride mixture in 1.1 g (4.1 mmol) of

CH212, and the resulting slurry refluxed overnight. The

precipitated material was filtered off and washed with 10 ml

of IN 11Cl which was used to wash the supernatant ether

solution. The ether was washed with an additional 10 ml

portion of IN llC1 solution and then three times with 10 ml

of water. The ether was then dried with MgSO Evaporation










of the ether yielded a yellow oil shown by nmr and tic to be

only the starting mixture of chlorocycloheptatrienes.

Repetition of the above reaction with extended reflux

times yielded the same results.

Preparation of 7,7-dibromobicyclo[4.1.0]-2-heptene.18

In a 500 ml three-necked round-bottomed flask equipped with

a mechanical stirrer addition funnel and nitrogen purge,

cooled in an ice bath, were stirred 200 ml of pentane, 24 g

(0.29 mol) 1,3-cyclohexadiene and 33.7 g (0.30 mol) potassium

t-butoxide. To this slurry was added with efficient stir-

ring, 75.8 g (0.30 mol) of bromoform in 90 ml of pentane,

dropwise over 2 hours. The slurry was allowed to stir an

additional 4 hours at 00, and allowed to warm to room

temperature overnight. To this was then added 400 ml of

water and the layers were separated. The aqueous phase was

then extracted twice with 200 ml of pentane. The combined

pentane extracts were then washed two times with 200 ml of

water and dried over MgSO4. Removal of solvent by rotary

evaporation leaves an oil that is purified by reduced

pressure distillation. Thirty-seven grams of liquid[bp 80-20

(5 mm), 68-700 (.8 mm)] (49% yield) was collected. The

identity of this compound was confirmed by comparison of its
18
nmr with that recorded in the literature. This material

was stable over a period of weeks if kept refrigerated.
18
Preparation of 2- and 3-bromocycloheptatriene. 1 In a

250 ml three-necked round-bottomed flask equipped for dis-

tillation through a short Vigreux tube into a receiver









cooled in an ice bath, 72 g of quinoline was heated to

135-450. The system was then evacuated via a water aspir-

ator, and 24.3 g (96.4 mmol) of the above dibromide was

added dropwise. The system was adjusted by an nitrogen

bleed so that a rapid distillation of the resultant bromide

occurred with only limited co-distillation of the quinoline.

The distillate was then chromatographed on silica gel,

eluting with pentane, to yield 9.32 g (57%) of yellow oil,

identified as a mixture of 2- and 3-bromocycloheptatriene by

nmr. The isomer composition of the 2- and 3- isomers deter-

mined by integration of the methylene signals of the nmr
18
spectrum is 3 to 1, respectively [bp 740 (15 mm), lit.1

76-8 (16 mm)].

Preparation of 2- and 3-cycloheptatrienyl formaldehyde.

In a 500 ml round-bottomed flask equipped with a low temper-

ature thermometer, an argon inlet and septum, 6.8 g

(0.40 mol) of the mixture of 2- and 3-bromocycloheptatriene

was dissolved in 150 ml of dry THF. Under an argon blanket,

this solution was cooled to -70" by immersion in a dry ice

bath while stirring magnetically. To this was added drop-

wise at a rate that allowed the temperature to remain below

-65, 31 ml of n-butyllithium (1.6 M in hexane) (0.05 mol).

Stirring was continued for an additional 20 minutes at -70.

Then a solution of 7.3 g (0.10 mol) of dry dimethylformamide

in 50 ml of THF was added dropwise at a rate that allowed the

temperature to remain below -65". Again the solution was

stirred for 90 minutes at -700; then it was allowed to warm










to -200. One hundred milliliters of saturated NH C1 solution

was added and the mixture was filtered, separated, and the

aqueous phase was extracted three times with 100 ml of ether.

The combined organic phases were washed five times with ice

in water and dried with NaSO4. The solvent was removed by

rotary evaporation and the residue was purified by chroma-

tography on silica gel, utilizing pentane to elute a fast-

moving impurity, ultimately eluting the aldehyde with 25%

ether in pentane. A yellow oil (2.6 g, 55%) was obtained

that was identified by nmr as the aldehydes named above, in

the same ratio as the starting bromides.

H NMR (CDC13) 2.5 (t), 2.4 (t), 5.2-7.4 (m), two sing-

lets 569 and 574 Hz downfield from TMS.

Preparation of cycloheptatrienyl formaldehyde tosylhy-

drazones. In around-bottomed flask under nitrogen, 2 g of

the above aldehyde was stirred with 3.1 g of tosylhydrazide

in 30 ml of absolute ethanol for 1.5 hours. On silica gel

using CH2Cl2 as the eluting solvent, TLC showed almost

complete conversion to tosylhydrazone. Removal of the etha-

nol by rotary evaporation yielded a glass that was recrystal-

lized twice from ether, yielding 3.6 g (72%) of crystals

that tended to retain ether. The ether was removed by chro-

matography on silica gel with methylene chloride.

IH NMR (CDCl3) 2.3 (m), 2.35 (s), 5.2-7.5 (m), 7.5

(A2X2), two singlets 513 and 520 Hz downfield from TMS.










Preparation of the sodium salts of the cycloheptatrienyl

formaldehyde tosylhydrazones. Under a nitrogen atmosphere in

a dry box, 576 mg (12 mmol) of NaH (50% in oil) was added to

3.5 g of the above tosylhydrazone mixture in 40 ml of dry

THF. After stirring for 15 minutes, 40 ml of hexane was ad-

ded to precipitate the salt. This was filtered and washed

twice with 10 ml of hexane. After drying under nitrogen,

2.53 g of material that decomposed between 165-750 was

obtained.

Pyrolysis of the sodium salts of the cycloheptatrienyl

formaldehyde tosylhydrazones. In a 25 ml round-bottomed

flask equipped with a solid addition funnel and a reflux con-

denser, 15 ml of diglyme was refluxed under nitrogen. To

this was added in portions 310 mg of the sodium salt. The

resulting slurry was refluxed for 30 minutes and then poured

into 100 ml of water. This was extracted three times with

75 ml of pentane and then three times with 75 ml of ether.

The ether was washed with water (10 x 100 ml) and dried over

MgSO4. Evaporation of the ether yielded a brown oil (40 mg,

30%) that was shown to contain the pyrazoles 6 and 7.

H NMR (CDC13) 7.5 (s), 7.6 (s), 6.5-6.8 (m), 3.5 (d,

J = 6 Hz), 2.7 (t, J = 6 Hz); mass spectrum m/e (rel inten-

sity) 132(21.1), 131(28.0), 105(15.3), 104(9.0); high reso-

lution mass spectra m/e 132.06785, theoretical: 132.06870.

Photolysis of the sodium salts of the cycloheptatrienyl

formaldehyde tosylhydrazones. To 125 ml of dry THIF in a

Pyrex immersion well, irradiated (550 W, Hlanovia medium







62

pressure lamp) under nitrogen, was added with efficient

stirring, 310 mg of the sodium salt over 20 minutes. Photo-

lysis was continued for 2.5 hours yielding a yellow mixture.

The solvent was removed by rotary evaporation, and the

residue was washed three times with 10 ml of methylene

chloride. The remaining solid weighed 115 mg. (A quanti-

tative yield of sodium toluene sulfinate would have been

178 mg.) The solvent was evaporated from the methylene

chloride fraction, and the resulting material was chromato-

graphed on silica gel with chloroform/methanol 3:2. A

yellow band was eluded yielding an oil (87 mg, 66%) that

was shown by mass spectroscopy to be isomeric with the

pyrazoles isolated from the pyrolysis of the salt. The

proton spectrum shows them to be different compounds.

IH NMR (CDC13) 7.5 (s), 7.3 (s), 5.5-6.5 (m), 2.8

(d, J = 5 Hz), 2.8 (t, J = 4 Hz); mass spectrum m/e (rel

intensity) 132(99.7), 131(100), 105(21.7), 104(23.1);

IR (major peaks) 3200, 3000, 2900, 1700, 1600, 1180.

Preparation of 8-chlorobicyclo[5.1.0]2,4-octadiene (10).

In a 1 liter three-necked round-bottomed flask equipped with

an addition funnel and a dry ice condenser 175 g (1.90 mol)

of freshly distilled cycloheptatriene and 55 g of methylene

chloride (0.65 mol) were stirred in 200 ml of ether under

nitrogen. To this was added dropwise 200 ml (1.77 M,

0.35 mol) of methyllithium in ether over 1.5 hours. The

ether refluxed. The material was allowed to cool to room

temperature while stirring was continued for an additional









two hours. The resulting slurry was poured into 250 ml of

water to dissolve the precipitated material. The layers

were separated and the aqueous layer was extracted three

times with 200 ml of ether. The organic phases were com-

bined and dried over MgSO4. The ether was removed by rotary

evaporation, and the residual cycloheptatriene was removed

by bulb to bulb distillation at 10 mm. The remaining oil

was distilled through a Vigreux column at 5 mm. The fraction

boiling at 52-600 was collected yielding 11 g of yellow oil

(22%). This oil was distilled through a spinning band

column with no apparent change in the nmr [bp 57-8 (3 mm)].

H NMR (CDC13) 5.9 (m), 5.7 (m), 3.6 (t, J = 3 Hz),

3.0 (t, J = 8 Hz), 2.5 (m), 1.6 (m); mass spectra m/e (rel

intensity) 140(10.1), 105(100), 91(29.2); high resolution

mass spectrum m/e 140.0392, theoretical: 140.0392; IR (major

peaks) 3050, 3000, 1620, 1440, 1380, 1280, 1240, 1050, 840,

760.

Anal. Calcd for C8 9 Cl: C, 68.34; H, 6.45. Found: C,

68.23; H, 6.46.

Reaction of 8-chlorobicyclo[5.1.0]2,4-octadiene (10)

with maleic anhydride. In a 200 ml round-bottomed flask,

1.000 g (7.1 mmol) of 10 and 0.7 g (7 mmol) of maleic anhy-

dride were refluxed in 100 ml of carbon tetrachloride for

2 days. The solvent was evaporated and the residue was

dissolved in a minimum amount (30 ml) of hot carbon tetra-

chloride. Upon cooling, crystals formed. These were

separated and the mother liquor was concentrated to half its









original volume by heating on a steam cone. A second crop

of crystals was taken. Combination of these two crops

yielded 513 mg of white crystals that decomposed at 1640

(sealed tube). The mother liquor was chromatographed on

silica gel, eluting with pentane. The fast running fraction

corresponding to 10 was collected. Evaporation of the sol-

vent resulted in the recovery of 246 mg of 10. Correcting

for the recovered starting material, the yield of adduct was

45%.

H NMR (CDC13) 6.5 (m, 2H), 3.8 (m, 1H), 3.4 (d, 2H),

3.0 (m, 2H), 2.8 (m, 2H), 2.3 (m, 2H); mass spectrum m/e

(rel intensity) 238(1.4), 166(3.5), 150(2.7), 131(3.4),

78(100); IR (major peaks) 3500, 3000, 1850, 1775, 1220,

1080, 960, 920, 900, 785.

24
Preparation of 8,8-Dibromobicyclo[5.1.0)2,4-octadiene2

(17). In a 2 liter Morton flask equipped with a mechanical

stirrer and cooled in an ice bath, 96 g (1.0 mol) of cyclo-

heptatriene and 151 g (0.60 mol) of bromoform were stirred

in 200 ml of pentane, under nitrogen. Over a period of

1.5 hours, 56 g (0.50 mol) of potassium t-butoxide was added

in small portions. The resulting slurry was allowed to come

to room temperature and was stirred overnight. This material

was added to 300 ml of water and the layers were separated.

The aqueous layer was extracted two times with 250 ml of

methylene chloride. This was added to the pentane fraction

and the combined organics were washed with 200 ml of water

and dried over MySO4. The solvent was removed by rotary









evaporation and the remaining cycloheptatriene and bromoform

were removed by bulb to bulb distillation at 5 mm. The

pressure was lowered to 0.05 mm and the remainder of the

material was distilled from a steam bath into a vessel

cooled in dry ice. This material was quickly distilled

again at 0.05 mm, collecting a fraction boiling above 50.

This was chromatographed on silica gel, eluting with

pentane, to yield 33 g (24%) of yellow liquid whose nmr

matches that reported in the literature for 1724 [lit. bp
24
42-60 (0.02 mm)]24 This material darkened at storage temper-

atures below 0, but its properties were not affected.

IH NMR (CDCl3) 5.1-6.3 (m, 4H), 2.3-2.6 (m, 211),

2.1-2.4 (m, 2H).

Reduction of 8,8-dibromobicyclo[5.1.0]2,4-octadiene (17)

with tri-n-butyltin hydride. The tri-n-butyltin hydride was

prepared by the method of Kuivila55 and used as soon as

possible after preparation. In a 25 ml pear-shaped flask

with a septum top, 3.0 g (11 mmol) of the dibromide was

stirred magnetically under nitrogen. To this was added

dropwise, via syringe, 3.3 g (11 mmol) of the fresh hydride.

This was stirred for 24 hours and then distilled under

reduced pressure through a Vigreux column, collecting a

fraction boiling between 55 and 60' (1 mm). This material

was contaminated with tri-n-butyltin bromide, and so was

chromatographed on silica gel, eluting with pentane. A

fast running band was taken that yielded 500 mg of an oil,









still contaminated with the tin bromide, with an nmr similar

to that of 10.

H NMR (CDC 3) 5.3-6.2 (m), 3.5 (t, J = 4 Hz), 3.0 (t,

J = 8 Hz), 2.0-2.6 (m), 1.4-1.8 (m); mass spectrum m/e (rel

intensity) 186(4.2), 184(10.8), 182(7.8), 105(100), 91(51.2);

IR (major peaks) 2970, 2870, 1800, 1590, 1480, 1440, 1420,

1360, 1300, 1240, 1200, 1060, 1030, 870, 840, 800.

Reaction of 8-chlorobicyclo [5.1.0]2,4-octadiene (10) with

potassium t-butoxide, general procedure. In a 100 ml three-

necked flask, 5.0 g (45 mmol) of potassium t-butoxide was

stirred under nitrogen in 50 ml of dry THF. If a trap was

used, it was added to this solution. This solution was then

allowed to come to the desired temperature, and a solution of

700 mg of the chloride, 10, in 20 ml of THF was added dropwise.

Stirring was continued for one hour and the resulting material

was poured into 100 ml of water. This was extracted three

times with 75 ml of pentane, which was washed three times with

100 ml of water to remove any remaining THF. The pentane so-

lution was dried over MgSO4. The solvent was evaporated and

the material chromatographed on silica gel, eluting with pen-

tane. A fast running band was taken, which was shown to con-

tain cyclooctatetraene and styrene, by nmr and gas chromato-

graphy (11.5' 10% SE-52, 900)

Yields [corrected to reactive isomer,(12)

T COT Styrene

250 56 trace
650 36% 421
900 8% 8% (solvent: tetraglyme)









Yields were determined by gas chromatography using p-methyl-

styrene as an internal standard. Styrene was not generated

from the reaction of cyclooctatetraene with potassium

t-butoxide under the above conditions.

Preparation of 8-chlorobicyclo[5.1.0]2,4-octadiene (10)

enriched in 11. In a 50 ml round-bottomed flask, under

nitrogen, 1.4 g (10 mmol) of the mixture of chlorides was

stirred in 20 ml of dry THF. To this was added dropwise, at

room temperature, 270 mg (2.5 mmol) of potassium t-butoxide

in 20 ml of THF. The solution was stirred for 45 minutes

and then poured into 100 ml of water. The phases were sepa-

rated, and the aqueous layer was extracted three times with

50 ml of pentane. The combined extracts were washed three

times with water (300 ml) and dried over MgSO4. Evaporation

of the solvent and chromatography on silica gel, eluting

with pentane, yielded 818 mg of yellow oil. The oil was

pumped overnight at 1 mm to remove cyclooctatetraene,

yielding 410 mg of almost pure 11 (92% isomer purity by nmr

integration of the triplets at 3.0 and 3.6) containing no

cyclooctatetraene.

Reaction of 8 chlorobicyclo[5.1.0]2,4-octadiene (10)

with potassium t-butoxide at 900/1 mm. In a 100 ml round-

bottomed flask equipped with a septum and connected to a

vacuum system through a Vigreux column and a trap cooled in

liquid nitrogen, was stirred 5.0 g (45 mmol) of potassium

t-butoxide in 10 ml of tetraglyme. The cooled trap had

previously been charged with 0.1 ml of dimethyl acetylene










dicarboxylate in 10 ml of THF. With stirring, the flask

was evacuated (0.5-1 mm) and heated in an oil bath (90-5').

When the system had come to temperature, a solution of

1.4 g (10 mmol) of 10 in 20 ml of tetraglyme was injected

dropwise through the septum via syringe. With the first

addition of chloride a red material began to collect in the

cold trap. After the addition was complete the black solution

was allowed to react for an additional hour, although no addi-

tional reaction was apparent. Nitrogen was bled into the

system, and the red material was washed into the trap

with a solution of 0.2 ml of the diester in 10 ml of THF.

This solution was placed in the refrigerator for 24 hours,

during which the color changed from red to orange. After

48 hours, the solution was almost yellow. After 72 hours in

the refrigerator the solution was reduced to one-fourth its

original volume on the steam cone. Benzene (25 ml) was

added, and this solution was reduced in volume in the same

manner. Benzene was then added to make the solution to

25 ml, and 0.5 g of Pd on charcoal was added. Air was bub-

bled through this mixture overnight, after which the catalyst

was removed by filtration through a small column of alumina,

washing well with ether. The resultant material was chroma-

tographed on deactivated alumina, eluting with ether to

bring down a yellow band. Elution with methanol yielded a

purple material which was purified further by TLC (alumi-

na/ether), to yield a blue material.










UV/VIS (C6H12) 575, 365, 347, 305, 290, 235; lit.33

580, 370, 350, 300, 290, 240; mass spectrum m/e (rel inten-

sity) 244(50.3), 213(100), 183(32.6).

Reaction of 8,8-dibromobicyclo[5.1.0]2,4-octadiene (17)

with potassium t-butoxide at 0'. In a 100 ml round-bottomed

flask, equipped with an addition funnel and immersed in an

ice bath, was stirred a solution of 5.0 g (45 mmol) of

potassium t-butoxide in 25 ml of THF. To this was added

dropwise a solution of 1.32 g (5.0 mmol) of 17 in 20 ml

of THF. This was stirred at 00 for 30 minutes and allowed

to come to room temperature. This was poured into 300 ml of

water, and the resulting solution was extracted three times

with 100 ml of pentane. Subsequent extraction of the

aqueous phase with ether yielded nothing. The combined

pentane extractions were combined and washed three times

100 ml of water and dried over MgSO4. The residue was

chromatographed on silica gel plates (1.5 mm, activated at

1400 for 30 minutes). Two bands (Rf .3 and .5 were isolated

after elution with pentane for three cycles. These two

fractions (Rf = .5, 45 mg; Rf = .3, 35 mg) had nmr and mass

spectral characteristics of t-butyl styryl ethers.

For R = .5 (trans isomer) IH NMR (CDC 3) 7.2 (s),

7.1 (d, J = 12 Hz), 6.0 (d, J = 12 Hz), 1.4 (s); mass

spectrum m/e (rel intensity) 176(3.0), 120(100), 91(37.4),

77(3.4).







70

For Rf = .3 (cis isomer) 1H NMR (CDC13) 7.0-7.8 (m),

6.5 (d, J = 7 Hz), 5.3 (d, J = 7 Hz), 1.4 (s); mass spec-

trum m/e (rel intensity) 176(3.8), 120(100), 91(40.3),

77(3.5).

Preparation of 3-phenylphthalide.56 In a 2 liter round-

bottomed flask equipped for reflux, 200 g of zinc dust,

100 g (0.44 mole) o-benzoylbenzoic acid, 200 ml of water,and

800 ml of glacial acetic acid were refluxed vigorously for

2 hours. The resultant yellow solution was decanted from

the residue, which was washed with 200 ml of hot glacial

acetic acid. The solutions were combined and allowed to

stand overnight. Beautiful white needles were separated

by filtration. Addition of 500 ml of water to the mother

liquor generated a second crop, and addition of another

500 ml of water, followed by cooling in ice, allowed a third

crop to be collected. These combined crystals were added

carefully in portions to 600 ml of 5% NaHCO3 solution with

stirring. Stirring was continued for an additional hour to

allow the flocculent material to digest. This was filtered

and recrystallized from 800 ml of hot ethanol. Hot filtra-

tion was necessary to remove an insoluble impurity. After

two crops, 51 g of material was isolated (mp 114-5, lit.56

115-6) (55%). Additional material of questionable purity

could be isolated (10 g) but was not retained.

Preparation of 1,3-diphenylisobenzofuran.3 In a 250 ml

addition funnel capped at the bottom with a 50 ml round-

bottomed flask, and equipped with a reflux condenser,









phenylmagnesium bromide was generated under nitrogen from

40 g of bromobenzene and 6 g of magnesium turnings.

The funnel was transferred to a 1 liter three-necked

round-bottomed flask and the Grignard was dripped into 42 g

of 3-phenylphthalide in 250 ml of dry THF over a period of

45 minutes. After an additional hour of stirring, 250 ml

of benzene was added, and the resulting solution was washed

two times with 250 ml of IN HC1. The organic phase was

then washed with 100 ml of 10% K2CO3. The aqueous phases

were then combined and extracted with 100 ml of benzene.

The combined organic were filtered through a frit contain-

ing MgSO4. The solvent was evaporated and the residue was

recrystallized from 250 ml of hot benzene. A second crop

was precipitated by addition of 200 ml of ethanol. The

combined solids were again recrystallized from benzene to
37
yield 38 g (70%, lit 87%) of yellow needles melting at

128-30 (lit.37 130-1o).

Reaction of 8,8-dibromobicyclo[5.1.0]2,4-octadiene (17)

with potassium t-butoxide and 1,3-diphenylisobenzofuran,

general procedure. In a 100 ml round-bottomed flask

equipped with an addition funnel and nitrogen inlet, was

stirred 5.0 g (45 mmol) of potassium t-butoxide and 1.35 g

of 34 (5.0 mmol) in 40 ml of dry TI1F. Under a blanket of

nitrogen, the solution was brought to the desired tempera-

ture. A solution of 1.32 g (5.0 mmol) of 17 in 20 ml of THF

was added dropwise,and after addition was complete, the

solution was allowed to stir for an hour, Ether (100 ml)









was added and the phases separated. The aqueous phase was

extracted three time with 100 ml of ether. The organic

phases were combined and washed four times with 100 ml of

water and dried over MgSO4. The solvent was removed by

rotary evaporation and the resulting solid pumped overnight

at 1 mm to remove any traces of TIIF. This material was

chromatographed on silica gel, eluting with carbon tetra-

chloride, taking the adduct as a freely moving band.

Reaction of 8-chlorobicyclo[5.1.0]2,4-octadiene (10)

with potassium t-butoxide and 1,3-diphenylisobenzofuran at

-780. A solution of 2.8 g of 10 (20 mmol, 5 mmol of 12) in

20 ml of TIIF was allowed to react with 5.0 g (45 mmol) of

potassium t-butoxide and 1.35 g (5.0 mmol) of 34 in the

manner described for the general reaction of 17. After the

addition of 10 was complete and the resulting solution was

stirred for one hour, a solution of maleic anhydride in THF

was added until tlc (silica gel/pentane)showed that the 34

was consumed. Work up was then resumed in the general

manner. Chromatography on silica gel, eluting with CC14,

yielded a crystalline material as the second material

eluted from the column. (The first material eluted was

unreacted 11 and some cyclooctatetraene.) Recrystallization

from benzene/methanol yielded 197 mg (11%) of white crys-

tals, melting at 168-90. Two more recrystallizations

yielded an analytically pure material, mp 169-169.5.

11 NMR (CDC13) 7.0-8.0 (m, 1411), 5.8-6.2 (m, 411),

2.1-3.2 (in, 311), 1.7 (d, J = 4 11z, 1il); mass spectrum m/e









(rel intensity) 374(5.7), 270(100), 91(6.9), 77(16.8); high

resolution mass spectrum m/e 374.16865, theoretical

374.16706; ir (major peaks) 3050, 2970, 2900, 1590, 1480,

1430, 1330, 1280, 1040, 1020, 970, 910, 900.

Anal. calcd for C28H220: C, 89.81; H, 5.92. Found:

C, 89.71; H, 6.01.

Reaction of 8,8-dibromobicyclo[5.1.0]2,4-octadiene (27)

with potassium t-butoxide and 1,3-diphenylisobenzofuran at

-78. Silica gel chromatography (CC14) of the material

formed at -780 by the above procedure allowed a moderately

slow moving material to be separated from a small amount

of rapidly eluted 34. After evaporation of the solvent,

1.157 g of material was isolated that was shown to be 19.

Three recrystallizations from THF/pentane gave an analyti-

cally pure sample (mp 156-70) whose nmr did not differ from

the material separated by chromatography (51% yield).

1H NMR (CDC 31 ) 7.2-8.0 (m, 14H), 5.8-6.5 (m, 3H),

5.5 (d, J = 10 Hz, 1I1), 3.0-3.4 (m, 1H), 2.5-2.9 (m, 211);

mass spectra m/e (abs intensity) 20 eV 454(32), 452(47),

373(50457); (rel intensity) 70 eV 373(100); ir (major

peaks) 3080, 3050, 2960, 2940, 1510, 1460, 1350, 1310, 1150,

1120, 1000.

Anal. Calcd for C28H 21BrO: C, 74.18; H, 4.67. Found:

C, 73.99; H, 4.72.

Preparation of 8,8-dichlorobicyclo[5.1.0]2,4-octadiene.

In the same manner as the preparation of 17, 120 g (1 mole)

of chloroform and 100 g (1.1 mole) of cycloheptatriene were









allowed to react with 52 g of potassium t-butoxide (.5 mol)

Work up in the same way yielded 12.4 g (14%) of oil boiling

at at 450 (0.05 mm) [lit.57 36-90 (0.05 mm)].

H NMR (CDC13) 5.5-6.4 (m, 4H), 2.4-2.6 (m, 2H),

2.0-2.4 (m, 2H); mass spectrum m/e (rel intensity) 178(1.5),

176(5.7), 174(7.3), 139(62.3), 103(100), 91(21.4).

Reaction of 8,8-dichlorobicyclo[5.1.0]2,4-octadiene

with potassium t-butoxide and 1,3-diphenylisobenzofuran at

-780. Silica gel chromatography (CC14) of the material

formed at -780 by the general procedure allowed a material

very similar to the 34 adduct of the cyclopropene formed

from 17 to be isolated. This solid (0.792 g, 39% yield) was

recrysta]lized from THF/pentane to yield a material with a

melting point of 148.5-150.

IH NMR (CDCl3) 7.1-7.9 (m, 14H), 5.8-6.5 (m, 3H),

5.5 (d, J = 11 Hz, 1H), 3.0-3.5 (m, 1H), 2.5-2.9 (m, 2H);

mass spectra m/e (abs intensity) 20 eV 408(131), 410(46),

373(39475); (rel intensity) 70 eV 373(100).

Anal. Calcd for C28H21C10: C, 82.24; H, 5.18. Found:

C, 80.73; H, 5.36.

Reaction of 8,8-dibromobicyclo[5.1.0]2,4-octadiene (1_)

with potassium t-butoxide and 1,3-diphenylisobenzofuran at

00. The reaction was repeated via the general procedure at

00. Chromatography yielded a material (22) that was eluted

with CCl4 in the usual manner. Another material (23) was

eluted with ether. This material appeared to have been

spread throughout the length of the column during the










elution with CCl4. This compound, 23, was not apparent in

the nmr of the crude material before chromatography. This

crude material (1.99 g) appeared to be almost totally com-

posed of 22. Determination of the composition of this crude

material by nmr, using 1,3,5-trinitrobenzene as the internal

standard, showed it to contain 2.57 mmol (1.16 g, 51%) of

22. Two recrystallizations of the material eluted with CC14

in benzene/methanol yielded 335 mg of a labile white mate-

rial that darkened at 950 and decomposed at 1100. Recrystal-

lization of 23 from chloroform/methanol yielded crystals

melting at 200-1 (340 mg).

For 22, 1i NMR (CDC13) 7.7-8.0 (m), 7.3-7.5 (m),

7.2 (s), (1411); 5.2-6.5 (m, 611); 4.5 (s, 1H). Mass spectrum

m/e (rel intensity) 454(.5), 452(.4), 267(100). High reso-

lution mass spectrum m/e 454.07642, theoretical, 454.07540.

IR (major peaks) 2950, 1550, 1460, 1420, 1410, 1310, 1270,

980.

For 23 1H NMR (CDC 3 ) 7.0-7.7 (m, 19H), 5.5 (s, 10);

mass spectrum m/e (rel intensity) 372(1.7), 267(100); high

resolution mass spectrum m/e 372.15138, theoretical

372.15130; ir (major peaks) 3100, 1670, 1605, 1595, 1500,

1450, 1000.

Anal. calcd for C28H 20: C, 90.29; H, 5.41. Found:

C, 90.43; 11, 5.46.
40
Preparation of 1,2,4-triphenyl-l,4-butandione. In a

500 ml round-bottomed flask, 18 g of acetophenone and 31 q

of benzoin were refluxed with 4 g of KCN in 75 g of water









and 75 g of ethanol. After boiling for a half hour, an

oil began to form. After 1.25 hours, the oil was separated

and the mother liquor was allowed to stand. More oil was

collected and combined with that previously obtained. After

this material had stood for 1.5 hours and had begun to

solidify, it was transferred to porous tiles, where it

immediately crystallized. This material was boiled in

300 ml of ethanol and filtered hot. The mother liquor was

allowed to stand, and crystals formed. The residue was

added to 300 ml of ethanol and refluxed for a half hour.

This was filtered while hot and allowed to stand until cry-

stals formed. The two sets of crystals were combined to

yield 22 g (40%) of yellow crystals that were used without

further purification.

40
Preparation of 2,3,5-triphenylfuran.40 In a 250 ml

Erlenmeyer flask, the above crystals were stirred in 150 ml

of coned H21SO4, for two hours. This was poured onto

1000 ml of ice. The sticky material was separated and

dissolved in 200 ml of boiling ethanol. When cooled, this

precipitated white needles (16.2 g, 78%), that melted at

90-1.50 (lit.40 92).

Reaction of triphenylfuran with benzyne. In a 250 ml

Erlenmeyer flask, 6.8 g (.05 mol) of anthranilic acid was

dissolved in 75 ml of ethanol and 5 ml of coned HC1 and

cooled in an ice bath to 0-50. With stirring, 12.5 ml

(.093 mol) of isoamyl nitrite was added dropwise at a rate

keeping the temperature below 10. After ten minutes of









stirring, a precipitate formed. Ether (75 ml) was added

and the solid material filtered and washed with ether.

Five grams (.27 mol) of this material, along with 8 g

of 2,3,5-triphenylfuran and 3 g of propylene oxide (acid

scavenger) was refluxed in 50 ml of dichloroethane for

3 hours. After cooling the solvent was removed by rotary

evaporation. The residual material was chromatographed on

silica gel, eluting with pentane. The residual triphenyl-

furan was removed as a fast running band. Elution with

ether allowed collection of a solid material after evapora-

tion of the solvent. This material was recrystallized

from THF/pentane to yield 3 g of crystals melting at 170-5.

Additional recrystallizations from CCI4 yields a colorless

material melting at 172-40

H NMR (CDC3) 7.0-8.0 (m); mass spectrum m/e (rel

intensity) 372(36.7), 270(100), 261(60.0); ir (major peaks)

3050, 2950, 1590, 1480, 1430, 1330, 1290, 1040, 995, 980.

This material was inert in acetonitrile solution, and

does not show any evidence of being converted to 23 after

overnight reflux in acidic methanol overnight.

Reaction of 8,8-dibromobicyclo[5.1.0]2,4-octadiene (11)

with potassium t-butoxide with subsequent addition of 1,3-

diphenylisobenzofuran. As in the general procedure, 1.32 g

of 17 was allowed to react with 5.0 g of base, without 34,

at -78. After an hour of stirring at this temperature,

].35 g of 34 in 20 ml of Till was added dropwise over

10 minutes, and the resulting solution was allowed to warm









to room temperature. After the usual work up with ether and

water, a large amount of polymer was noted. The nmr spec-

trum of the crude material showed some 22 and no bromocyclo-

octatetraene. There was also no evidence of the styrene

ethers.

48
Preparation of coumalic acid. In a 2 liter round-

bottomed flask at room temperature 200 g (1.5 mol) of

DL-malic acid was stirred magnetically in 170 ml of coned

H2SO4. To this was added carefully, in one portion, 50 ml

of fuming H2SO4, and the mixture was allowed to stir for

45 minutes. This procedure was repeated twice (total,

150 ml fuming H2SO4). After the last addition of acid the

material was allowed to stir for one hour and was trans-

ferred to the steam cone, where it was heated for 2 hours.

After a slight cooling period, the warm solution was poured

carefully onto 800 g of ice. This was allowed to stand

overnight. The solid was filtered and washed three times

with 300 ml of ice cold water. The solid was dried on a

sand bath at 650, and divided into two portions. One of

these portions was recrystallized from 170 g of methanol,

decolorizing with Norite. The solid was separated and the

mother liquor was used to recrystallize the other portion.

This resulted in 45 g of sulfur yellow crystals (22%)

melting at 206-90 (lit.48 206-9).
49
Preparation of 2-pyrone. Coumalic acid (30 g,

.21 mol) was sublimed at 0.5 mm through a Vycor tube packed

loosely with copper turnings and heated to 6500. The









resulting yellow oil was collected in an ice-cooled trap.

Distillation of this oil at 26 mm yielded 12.3 g of material

49
boiling between 1060 and 1090 [lit.4, 1100 (26 mm)]. This

material was stored under nitrogen below 0' until used.

Preparation of the sodium salt of tropone tosylhydra-

zone. The sodium salt of tropone tosylhydrazone was pre-

51
pared by the method of Jones and Ennis,51 by the reaction

of tropone tosylhydrazone with sodium hydride in THF. The

resulting precipitate was recrystallized from 1% DMSO in

THF to yield purple needles.

Preparation of 6-chloro-2-pyrone53 To a 100 ml round-

bottomed flask containing 10 g (0.077 mol) of glutaconic

acid and cooled in an ice bath, was added in portions 30 g

(.14 mol) of phosphorous pentachloride. During the course

of the addition, gas was evolved and the solids melted.

When spontaneous reaction ceased, the solution was allowed

to warm to room temperature. Again, reaction commenced and

when this moderated, the material was heated for 15 minutes

on a steam cone. The resulting red liquid was distilled

yielding 7.1 g of material [bp 64-5' (.5 mm), lit. 67'

(9 mm)] identified by comparison with the nmr spectrum

noted in the literature.5
53
Preparation of 6-d-2-pyrone. In a 50 ml round-bot-

tomed flask containing 5.8 g (44 mmol) of 6-chloro-2-pyrone

in 30 ml of THF, under nitrogen, was added 11.6 g (77 mmol)

of vacuum-dried Nal. This was allowed to stand, with

occasional swirling, for 4 hours. The resulting material










was poured into 20 ml of deuterioacetic acid and 30 g of

zinc dust (vacuum dried at 130). This was allowed to stand

under nitrogen overnight, and then was poured into 100 ml

of CH2C12, filtered, and the resulting solid washed three

times with 100 ml of water and two times with 10% Na2CO3

solution and dried over Na2SO4. After chromatography on

silica gel, eluting with CH2Cl2 and distillation from a

bath at 1150 (25 mm), 607 mg of material was isolated that
58
was identified by comparison of the nmr and mass spec-

trum53 with the literature.

Pyrolysis of the sodium salt of tropone tosylhydrazone

in the presence of 2-pyrone. Under nitrogen, 1.4 g

(14 mmol) of 2-pyrone in 15 ml of dry diglyme was heated

to 1100. In portions, 0.5 g (1.4 mmol) of the sodium salt

was added through a solid addition funnel. Stirring was

continued for 30 minutes at this temperature. The resulting

material was poured into 150 ml of water and was then

extracted three times with 100 ml of pentane. The pentane

extracts were washed six times with 50 ml of water and

dried over MgSO4. Chromatography (Fisher basic alumina,

activity grade III/pentane) yielded 58.9 mg (29%) of mate-

rial identified as 3,4-benzocycloheptatriene by comparison
4]
with the nmr reported in the literature.4

Pyrolysis of the sodium salt of tropone tosylhydrazone

in the presence of 6-d-2-pyrone. In the same manner as

the pyrolysis reported above, 100 mg (1.04 mmol) of 6-d-2-

pyrone was allowed to react with the pyrolysis product of










the sodium salt of tropone tosylhydrazone (500 mg, 1.4 mmol).

After the same type of work up, 33 mg of the 3,4-benzocy-

cloheptatriene was isolated. The nmr spectrum showed that

the deuterium was retained in the phenyl portion of the

molecule.

H NMR (CCl4) 2.50 (t, 1.8H, J = 6 Hz), 5.82 (d of t,

2.1H, J = 6, 10 Hz), 6.59 (d, 2.011, J = 10 Hz), 7.25 (s,

3.3H1). Mass spectrum m/e (rel intensity) 144 (9.6),

143(90.0), 142(100), 116(26.9); for 3,4-benzocycloheptatri-

ene 143(8.6), 142(79.4), 141(100), 115(29.8). IR (major

peaks) 3010, 2950, 2820, 2220, 1480, 1450, 1430, 895, 820,

800.

Photolysis of the sodium salt of tropone tosylhydrazone

in the presence of 2-pyrone. In a Pyrex photochemical

immersion apparatus was stirred under nitrogen, 1.4 g

(14 mmol) of 2-pyrone in 100 ml of dry TIF. While irradi-

ating with a 550 W Hanovia medium pressure lamp, 0.50 g of

the sodium salt (1.4 mmol) was added in portions over

30 minutes. The mixture was worked up in the same manner

as the pyrolyses. The average yield for the photolyses was

25 mg. The ratio of the 3,4-benzocycloheptatriene to 1,2-

benzocycloheptatriene was 2:1.

















REFERENCES


1. W. M. Jones and C. L. Ennis, J. Amer. Chem. Soc., 91,
6391 (1969).

2. L. W. Christensen, E. E. Waali and W. M. Jones, J. Amer.
Chem. Soc., 94, 2118 (1972).

3. For a review of this type of rearrangement see Some
Pericyclic Reactions of Carbenes, W. M. Jones and
U. H. Brinker in "Pericyclic Reactions," Vol 1, Academic
Press, New York, 1977; and W. M. Jones, Ace. Chem.
Res., 10, 353 (1977).

4. R. C. Joines, A. B. Turner and W. M. Jones, J. Amer.
Chem. Soc., 91, 7754 (1969).

5. J. A. Meyers, R. C. Joines and W. M. Jones, J. Amer.
Chem. Soc., 92, 4740 (1970).

6. K. E. Krajca, T. Mitsuhashi and W. M. Jones, J. Amer.
Chem. Soc., 94, 3661 (1972).

7. T. T. Coburn and W. M. Jones, J. Amer. Chem. Soc., 96,
5218 (1974).

8. C. Mayor and W. M. Jones, Tetrahedron Lett., 3855
(1977).

9. K. Untch, International Symposium on the Chemistry of
Nonbenzenoid Aromatic Compounds, Sendai, Japan,
August, 1970; personal communication.

10. B. Duell and W. M. Jones, private communication.

11. S. Winstein and J. Sonnenburg, J. Amer. Chem. Soc., 83,
3244 (1961).

12. For leading references see P. R. Storey and B. C. Clark,
Jr., "Carbonium Ions," Vol 3, Wiley Interscience,
New York, 1972, p 1084.

13. M. Oda, Y. Ito and Y. Kitahara, Tetrahedron Lett.,
2587 (1975).

14. S. W. Benson, "Thermochemical Kinetics," Wiley,
New York, L968.







83

15. K. Untch, personal communication.

16. B. Folisch, Chem. Ber., 101, 2717 (1968); ibid., 104,
2324 (1971).

17. E. LeGoff, J. Org. Chem., 29, 2048 (1964).

18. P. G. Lindsay and C. B. Reese, Tetrahedron, 21, 1673
(1965).

19. T. J. Batterham, "NMR Spectra of Simple Ieterocycles,"
Wiley, New York, 1973.

20. R. K. Bartlett and T. S. Stevens, J. Chem. Soc. C,
1964 (1967).

21. U. II. Brinker and W. M. Jones, Tetrahedron Lett., 577
(1976).

22. J. Elguero, C. Marzin, A. R. Katritzky and P. Linda,
"The Tauteromerism of Heterocycles," Academic Press,
New York, 1976.

23. W. E. Billups and L. E. Reed, Tetrahedron Lett., 2239
(1977).

24. P. J. Van Vooren, R. J. Fletterick, J. Meinwald and
R. E. Hughes, J. Amer. Chem. Soc., 93, 4394 (1971).

25. G. L. Closs and L. E. Closs, J. Amer. Chem. Soc., 82,
5723 (1960); G. L. Closs, R. A. Moss and J. J. Coyle,
ibid., 84, 4985 (1962).

26. For a review of reactions of this type see
W. E. Billups, Acc. Chem. Res., 11, 245 (1978).

27. J. Prestien and H1. Gunther, Anqew. Chem. Tnt. Ed., 13,
276 (1974).

28. R. iH. Parker, unpublished results.

29. T. C. Shields and P. D. Gardner, J. Amer. Chem. Soc.,
89, 5425 (1967).

30. L. Friedman and H. Shechter, J. Amer. Chem. Soc., 83,
3159 (1961); W. Kirmse and G. Munschen, Annalen, 726,
42 (1969).

31. W. J. Baron, M. R. DeCamp, M. E. Hendrick, M. Jones,
Jr., R. H. Levin and M. B. Sohn in "Carbenes," Vol 1,
M. Jones and R. A. Moss ed. Wiley, New York, 1973,
p 20.






84


32. W. N. Washburn, R. Zahler and I. Chen, J. Amer. Chem.
Soc., 100, 5863 (1978).

33. W. von E. Doering and E. Wiley, Tetrahedron, 11, 183
(1960).

34. H. E. Zimmerman and L. R. Sousa, J. Amer. Chem. Soc.,
94, 834 (1972).

35. P. Binger, Synthesis, 190 (1974); W. Eymann and
M. Ilanack, Tetrahedron Lett., 4213 (1972); Z. Yoshida,
H. Miyahara, Chem. Lett., 335 (1972).

36. C. L. Osborn, T. C. Shields, B. A. Shoulders,
J. F. Krause, H. V. Cortez and P. D. Gardner, J. Amer.
Chem. Soc., 87, 3158 (1965).

37. M. S. Newman, J. Org. Chem., 26, 2630 (1961).

38. D. H. Williams and I. Fleming, "Spectroscopic Methods
in Organic Chemistry," second edition, McGraw- Hill,
London, 1973.

39. V. V. Razin and E. M. Rud', Z. Org. Khim., 12(3), 689
(1976).

40. A. Smith, J. Chem. Soc., 57, 643 (1890).

41. G. W. Gruber and M. Pomerantz, Tetrahedron Lett., 3755
(1970).

42. H1. Durr, H. Kober, I. Halberstadt, V. Neu,
T. T. Coburn, T. Mitsuhashi and W. M. Jones, J. Amer.
Chem. Soc., 95, 3818 (1973).

43. H. Durr, M. Kausch and H. Kober, Angew. Chem. Int. Ed.
Engl., 13, 670 (1974).

44. T. Mitsuhashi and W. M. Jones, Chem. Commun., 103
(1974).

45. K. 11. Pauly and H. Durr, Tetrahedron Lett., 3649 (1976).

46. Recent evidence indicates that the assignment of these
structures as bisnorcaradienos may be incorrect.
H. Durr, private communication.

47. R. H. Parker and W. M. Jones, J. Org. Chem., 43, 2548
(1978).

48. R. HI. Wiley and N. R. Smith, "Organic Syntheses,"
Collect. Vol IV, Wiley, New York, 1963, p 201.


~







85


49. H. E. Zimmerman, G. L. Grunewald and R. M. Paulfer,
"Organic Syntheses," Collect. Vol V, Wiley, New York,
1973, p 982.

50. K. Untch, International Symposium on the Chemistry of
Nonbenzenoid Aromatic Compounds, Sendai, Japan, August
1970.

51. W. M. Jones and C. L. Ennis, J. Amer. Chem. Soc., 91,
6391 (1969).

52. 0. L. Chapman, D. de la Cruz, R. Roth, and J. Pacansky,
J. Amer. Chem. Soc., 95, 1338 (1973).

53. W. H. Pirkle and M. Dines, ibid., 90, 2318 (1968).

54. P. Radlick, J. Org. Chem., 29, 960 (1964).

55. H. G. Kuivila, Synthesis, 2, 499 (1970).

56. C. R. Hauser, M. T. Tetenbaum and D. S. Hoffonberg,
J. Org. Chem., 23, 861 (1958).

57. W. M. Wagner, H. Kloosterziel, S. van der Ven, Rec.
Tray. Chem., 80, 740 (1961).

58. W. H. Pirkle and M. Dines, J. Heterocycl. Chem., 6, 1
(1969).















BIOGRAPHICAL SKETCH


Richard H. Parker was born on the sixth of February,

1950, to Homer and Eris (Tucker) Parker. He was educated in

the LaGrange (Ga.) public school system and graduated from

LaGrange High School in 1968. He entered Auburn University

during the same year and received the Bachelor of Science

degree in 1972. He remained at Auburn University and re-

ceived the Master of Science degree in 1974, under the

direction of Dr. P. B. Shevlin. He entered the University

of Florida in September of that year. He is a member of the

American Chemical Society.










I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, c
a dissertation for the degree of Doctor of Philosophy.


William M. Jones, Chairman
Professor of Chemistry






I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.






l/-I
William R. Dolbier
Professor of Chemistry






I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.





/

John F. Helling
Professor of Chemistry









1 certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality,
as a dissertation of the degree of Doctor of Philosophy.








John R. Eyler
Associate Professor-of Chemistry






I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy.


Charles A. Burnap
Associate Professor


) ") ) 1,4h 1-


of Mathematics


This dissertation was submitted to the Graduate
Faculty of the Department of Chemistry in the College of
Liberal Arts and Sciences and to the Graduate Council, and
was accepted as partial fulfillment of the requirements for
the degree of Doctor of Philosophy.






June 1979
Dean, Graduate School




Full Text

PAGE 1

ATTEMPTS TO GENERATE A HOMOAROMATIC CARBENE AND BISNORCARADIEME By Richard H . Parker A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 1979

PAGE 2

The great tragedy of Science [is] the slaying of a beautiful hypothesis by an ugly fact. Thomas Henry Huxley

PAGE 3

ACKNOWLEDGMENTS The author would like to sincerely acknowledge the guidance and patience of his research director, Dr. W. M. Jones. His ability to allow the author to continually reassimilate organic chemistry, without frustration, is truly unique and genuinely appreciated. The author would also like to thank Dr. U. H. Brinker for introducing him to the subtleties and the true art of organic synthesis. It must be stated that the identities of some of the compounds contained herein might have remained unknown without access to time and the talents of Dr. Roy King and Tom Baugh. The author would also like to acknowledge the financial support of the National Science Foundation. Finally, not to acknowledge the never ending line of chemists who have passed through 413 CRB with explanations and suggestions (both sought and unsought) would be unfair and possibly plagiaristic .

PAGE 4

TABLE OF CONTENTS Page ACKNOWLEDGEMENTS m ABSTRACT v ii ATTEMPTS TO GENERATE A HOMOAROMATIC CARBENE 1 Introduction 1 Results and Discussion 10 AN ATTEMPT TO GENERATE BISNORCARADIENE 4 6 Introduction 46 Results and Discussion 49 EXPERIMENTAL 54 General 54 Preparation of tropone 55 Preparation of 1-, 2-, and 3-chlorocycloheptatriene 55 Preparation of zinc copper couple 56 Attempted reaction of 1-, 2-, and 3chlorocycloheptatriene with SimmonsSmith reagent 57 Preparation of 7 , 7-dibromobicyclo [ 4 . 1 . ] 2-hoptene 5 8 Preparation of 2and 3-bromocycloheptatriene 5 8 Preparation of 2and 3-cyc loheptatrieny 1 formaldehyde 5 9 Preparation of cycloheptatr ienyl formaldehyde tosyl hydra zones 6

PAGE 5

Preparation of the sodium salts of cycloheptatrienyl formaldehyde tosylhydrazones 61 Pyrolysis of the sodium salts of the cycloheptatrienyl formaldehyde tosylhydrazones 61 Photolysis of the sodium salts of the cycloheptatrienyl formaldehyde tosylhydrazones 61 Preparation of 8-chlorobicyclo [ 5 . 1 . ] 2,4-octadiene (10) 62 Reaction of 8-chlorobicyclo [ 5 . 1 . ] 2 , 4octadiene (1_0) with maleic anhydride 63 Preparation of 8 , 8-dibromobicyclo[5.1.0] 2,4-octadiene ( 17 ) 64 Reduction of 8 , 8-dibromobicyclo [ 5 . 1 . ] 2,4-octadiene (1/7) with tri-n-butyltin hydride 6 5 Reaction of 8-chlorobicyclo [ 5 . 1 . ] 2 , 4octadiene (1_0) with potassium t-butoxide, general procedure 66 Preparation of 8-chlorobicyclo [ 5 . 1 . ] 2,4-octadiene (ljO) , enriched in 11_ 67 Reaction of 8-chlorobicyclo [ 5 . 1 . ] 2 , 4octadiene (1_0) with potassium t-butoxide at 90°/l mm. 67 Reaction of 8, 8-dibromobicyclo [5 . 1 . ] 2,4-octadiene (i/7) with potassium t-butoxide at 0°T 69 Preparation of 3-phenylphthalide 70 Preparation of 1 , 3-diphenylisobenzof uran 70 Reaction of 8 , 8-dibromobicyclo [ 5 . 1 . ] 2,4-octadiene (17) with potassium t-butoxide and I, 3-dipheny li sobenzofuran, general, procedure 71 Reaction of 8-chlorobi cyclo [ 5 . 1 . ] 2 , 4octadiene (10) with potassium t-butoxide and 1 , 3-dipheny 1 i sobenzof uran at -78° 72

PAGE 6

Reaction of 8 , 8-dichlorobicyclo [ 5 . 1 . 0] 2,4-octadiene ( 17 ) with potassium t-butoxide and 1 , 3-diphenyl isobenzofuran at -78° 73 Preparation of 8 , 8-dichlorobicyclo[5.1.0)2, 4-octadiene 73 Reaction of 8 , 8-dichlorobicyclo [ 5 . 1 . ] 2,4-octadiene with potassium t-butoxide and 1 , 3-diphenylisobenzofuran at -78° 74 Reaction of 8 , 8-dibromobicyclo [ 5 . 1 . ] 2,4-octadiene (1_7) with potassium t-butoxide and 1 , 3-diphenylisobenzofuran at ° 74 Preparation of 1 , 2 , 4triphenyl-1 , 4-propandione 75 Preparation of 2 , 3 , 5triphenylf uran 76 Reaction of triphenylf uran with benzyne 76 Reaction of 8 , 8-dibromobicyclo [ 5 . 1 . ] 2,4-octadiene (17) with potassium t-butoxide with subsequent addition of 1 , 3-diphenylisobenzofuran 77 Preparation of coumalic acid 78 Preparation of 2-pyrone 78 Preparation of the sodium salt of tropone tosyl hydra zone 79 Preparation of 6-chloro-2-pyrone 79 Preparation of 6-d-2-pyrone 79 Pyrolysis of the sodium salt of tropone tosylhydrazone in the presence of 2-pyrone 80 Pyrolysis of the sodium salt of tropone tosylhydrazone in the presence of 6-d-2-pyrone 80 Pyronlysis of the sodium salt of tropone tosylhydrazone in the presence of 2-pyrone 81 REFERENCES 8 2 BIOGRAPHICAL SKETCH 8 6

PAGE 7

Abstract of Dissertation Presented to the Graduate Council of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy ATTEMPTS TO GENERATE A HOMOAROMATIC CARBENE AND BISNORCARADIENE by Richard II. Parker June 19 79 Chairman: W. M. Jones Major Department: Chemistry A neutral homoaromatic species has not been reported in the chemical literature. The similarity of the charge separated resonance structure of 2 , 5 , 7-cyclooctatrienyl idene , 1, to the structure of the homotropylium ion implies that this carbene might be a good candidate for a neutral homoaromatic species . The chemistry of the parent aromatic system, cycloheptatrienylidene , suggests two methods of generating 1. Base induced dehydrohalogenation of 5-chlorobicyclo [ 5 . 1 . ] 2 , 4-octadiene or 5-chlorobicyclo [ 5 . 1 . 0] 2, 5-octadiene should yield an allene that could rearrange to the carbene, 1. This carbene could also result from the carbene-carbene rearrangement of 2and 3-cyclohepta tr ienylcarbene . In this study, the chlorides could not be generated and attempts to generate the cycloheptatrienylcarbenes resulted in pyrazoles.

PAGE 8

The cyclopropene , 2 , 4 , 8-bicyclo [ 5 . 1 . 0] octatriene , 2, is the intermediate that would be expected in the carbenecarbene rearrangement of the cycloheptatrienylcarbenes mentioned above. This species was generated from the reaction of trans8-chlorobicyclo [5. 1. 0] 2 , 4-octadiene with potassium t-butoxide and trapped with diphenylisobenzof uran. In the absence of trapping materials, this reaction yielded cyclooctatetraene , heptafulvene and a trace of styrene. It is postulated that the cyclopropene, 2, rearranges to 1, which inserts into one of the transannular methylene C-H bonds to yield 2 , 4 , 7-bicyclo [4. 2 . 0] octatriene. This species opens in its usual manner to yield cyclooctatetraene. It is proposed that the other two products are the result of a base induced isomerization of 2 to 1 , 3 , 5-bicyclo [5 . 1 . ] octatriene. This either undergoes electrocyclic ring opening to heptafulvene or a norcaradiene rearrangement to a bicyclobutene, which opens to styrene. Reaction of 8 , 8-dibromobicyclo [5 . 1. ] 2 , 4-octadiene in the same manner seems to follow a similar pathway. The only products isolated from this reaction were 8-bromoheptafulvene, trapped with di phenylisobenzofuran , and 3-t-butoxystyrene. No bromocyclooctatetraene was noted, although the cyclopropene precursor is definitely formed since it was trapped with diphenylisobenzof uran. The styrene ethers appear to have been formed from a small amount of starting material that was able to doubly dehydrohal oqenate .

PAGE 9

An attempt was also made to synthesize bisnorcaradiene , 2,4,8,10-tricyclo [5. 4. 0. ' ] undecatetraene , 3. Cycloheptatetraene was allowed to react with 2-pyrone to yield, ultimately, 3 , 4-benzocycloheptatriene , 4. This reaction is postulated to proceed through an intermediate pentaene, 1,3, 5, 8 , 10-bicyclo [5. 4. 0] undecapentaene, 5. These pentaenes , when substituted, are believed to undergo electrocyclic ring closures to bisnorcaradienes . The ability of these bisnorcaradienes to open to pentaenes with different substitution patterns has led to the postulation of the intermediacy of bisnorcaradienes in many chemical reactions in which substituent scrambling of this type was noted. When a deuterated 2-pyrone was allowed to react with cycloheptatetraene , substituent scrambling of the type mentioned above did not occur. This reaction implies a competition between the ability of 5 to close to 3 and the ability of ^5 to aromatize to 4 previously not noted.

PAGE 10

ATTEMPTS TO GENERATE A HOMOAROMATIC CAR3ENE Introduction Cycloheptatrienylidene is an example of an aromatic carbene. It appears that the vacant p orbital of the singlet carbene has electron density from the six pi electron system delocalized into it, in much the same way that electron density is delocalized into the vacant orbital in the tropylium ion. The effect is evident in the reduced electrophilicity of the system compared to that of other carbenes lacking this conjugated pi system. The Hammett rho value of typical electrophilic carbene, such as dichlorocarbene (-0.619 at 80°, -0.378 at 0°), differs drastically from that noted for this carbene (+1.05). The increased stability of this carbene is also displayed in its lack of reactivity with alkenes not containing electron withdrawing substituents . In the absence of groups of this kind, the carbene merely dimerizes to heptaf ulvalene . // w -o One of the more interesting properties of this carbene . . . 3 is its participation in carbene-carbene rearrangements. This reaction o f carbenes is one in which a carbene at a vinylogous position can migrate to the adjacent carbon.

PAGE 11

This interconversion of vinylcarbenes takes place through an intermediate cyclopropene. Closure of the vinyl carbene to the cyclopropene is a thermally allowed (2+2 ) cyclos a c-c c = c c-c addition of the singlet carbene. Thermal opening of the ring along similar lines leads to the formation of the new vinylcarbene. When phenylcarbene is generated at 250° in the gas 4 phase, heptafulvalene is produced. This compound is postulated to arise from rearrangement of the phenylcarbene to cycloheptatrienylidene, throuqh the bicyclic triene. The cycloheptatrieny lidene then dimerizes as described above. / \ I Rearrangements of substituted cycloheptatri enyl idenes to yield substituted phenyl carbenes have been noted also. 5 Diphenylcarbene generated in the gas phase yields fluorene, presumably through a phenyl substituted cycloheptatrienylidene. While the interconversion of phenylcarbene and cycloheptatrienylidene is a gas phase phenomenon, substitution of this parent system can result in carbene-carbene rearrange-

PAGE 12

ments in solution. Decreasing the aromaticity of the aromatic carbene lessens the impact on the derealization energy of the carbene when the cyclopropene intermediate is formed, thus facilitating the reaction. An example is the 4, 5-benzocycloheptatrienylidene system in which an annelated ring is incorporated into the 4,5 bond of the cycloheptatriene moiety. When formed by either photolysis or thermolysis, it rearranges in solution to yield beta-naphthylcarbene. The annelated ring decreases the effect of the loss of aromaticity when the cyclopropene is formed. The cyclopropene intermediate in this reaction has been trapped in a Diels-Alder reaction with butadiene, cyclopentadiene , and 7 .C f uran. Another interesting feature of this carbene is its ability to equilibrate to its valence isomer, cycloheptao tetraene. Untch has found that dehydroha 1 ogenation of a mixture of 1-, 2-, and 3-chl orocycloheptatrienes yields heptaf ul valene. " This is proposed to proceed through an

PAGE 13

initial dehydrohalogenation to the cyclic allene that isomerizes to the carbene, which dimerizes. The interconversion of these isomeric intermediates has been shown to be independent of their precursors. Regardless of whether these intermediates are generated from thermolysis of the sodium salt of tropone tosylhydrazone or the reaction of K0C 4 H g Ci o a.t // \ the above mentioned chloride mixture with potassium t-butoxide, the allene can be trapped with isobenzof uran and the carbene can be trapped with styrene. Homoaromaticity is a term coined by Winstein to title the effect noted in some carbonium ions in which unusual stabilization is retained in ions in which conjugation has been interrupted in a pi system that would otherwise be aromatic. This effect, seen in many cationic species, has not been noted in neutral species.

PAGE 14

The interaction of the vacant p orbital with the pi system in cycloheptatrienylidene mimics the interaction of the vacant p orbital in the tropylium ion. Indeed, a resonance structure of this carbene can be drawn in which the tropylium ion is incorporated. Homotropylium ions have been prepared and extensively 12 studied. The apparent structure of these ions seems to be that of a non-classical monocyclic system, although some arguments have been made toward a classical bicyclic system. This ion suggests that it may be possible to generate a carbene that would show this type of homoaromatic behavior. This neutral species, which would have a resonance structure that mimics the homotropy lium ion, might be an exception to the generality that homoaromaticity is restricted to charqed species . There are two ways of studying the possibility of homoaromaticity in a cycloheptatrienylidene system. Replacinq

PAGE 15

one of the olefinic positions with a fused cyclopropane ring would yield a carbene that would be analogous to the bicyclic homotropylium ion above. This reaction has been 13 carried out by Oda et al. When the sodium salt of bicyclo[ 5 . 1 . ] octa-3 , 5-dien-2-one tosy lhydrazone was photolyzed or pyrolyzed, the ensuing carbene immediately underwent ring opening of the cyclopropane to yield 1 , 2 , 4 , 6-cyclooctatetN-N-Tos raene, which dimerizes under the reaction conditions. The other possibility is to introduce a methylene insulating group into the pi system. This can be done in one of two positions to yield either 2 or 3. It would be expected that in isomer 2, the rapid 1,2 hydrogen migration to yield cyclooctatetraene would predominate over any ^ 2 3 other carbene reaction. Isomer 3 cannot undergo this reaction . The rearrangement of the carbene generated by Kitahara, to the allene, along with the known ability of cycloheptatetraene, to isomerize to cyclohepta tetraene , sugqests that isomeri zation of 3 to its allene form, 4, might be a dominant

PAGE 16

3 4 reaction of this carbene. If this is so, then there would be no opportunity to observe the reaction of the carbene. Arguments can be made that suggest that rearrangement of 3 to 4 may be less likely than the rearrangement of Kitahara's carbene to the corresponding allene. Calculation of the differences in the enthalpies of formation between these allenes and the corresponding carbenes were made 14 using the group equivalents of Benson. Without correction for allene strain or any aromatic or homoaromatic stabilization, the difference in enthalpies (AAH,.) between 3 and 4 was shown to be 10 Kcal greater than the difference between Kitahara's carbene and allene. Intuition would lead to the notion that the strain in the seven membered ring of 4^ would be greater than that in the eight membered ring of Kitahara's allene. This would imply a greater AAII f than ./.I I 3

PAGE 17

the calculations indicate. This A£H, between 3 and 4 is 5 Kcal greater than the corresponding difference in the enthalpies of formation of cyclohept itrienylidene and cycloheptatetraene, a system shown to be in equilibrium (vide supra) . This would indicate that the rearrangement of the carbene, _3, to the allene, 4, would be more endothermic, or less exothermic, than the rearrangement noted by Kitahara, and less likely to occur. The chemistry of cycloheptatrienylidene suggests three ways to generate the homoaromatic carbene, 3^. The most straightforward would be from thermolysis or photolysis of the sodium salt of the tosylhydrazone of 2 , 5 , 7-cyclooctatrienone, 5. If homocycloheptatrienylidene exists in equilibrium N-N-Tcs 'I \\ with an allene isomer, a second method might be possible to generate the carbene, through a dehydrohalogenation reaction similar to the one used to generate cycloheptatrienylidene. The allene isomer, 4, that would be the valence isomer of 3 could be generated from the dehydrohalogenation of the chlorides above. Ci
PAGE 18

The third possible pathway to the homoaromatic carbene is through a carbene-carbene rearrangement . The cycloheptatrienylcarbenes below have the opportunity to rearrange in a similar manner to that of pheny lcarbene , to give 3. In principle, these carbenes could be generated by thermolysis or photolysis of the sodium salts of the tosylhydrazones of the corresponding aldehydes. An evaluation of homoaromaticity in this system could "-%, 1 w ^ ^> V \\ be carried out along the same lines as those used to study the aromaticity of cycloheptatrienylidene . If full or partial retention of the ability to delocalize the pi electron system into the vacant orbital of the carbene occurs, the system can be studied by the nucleophi 1 ici ty of the carbene. Its ability to react with olefins reactive to normal electrophilic carbenes can be tested. The ideal study would be one in which the carbene is allowed to react with substituted styrenes and its Hammett rho value determined. This should give a quantitative measure of the nucleophilici ty , i.e., the aromaticity, of the the carbene.

PAGE 19

10 Results and Discussion Because of the lack of a feasible synthetic route to the sodium salt, 5, attempts were made toward the synthesis of _3 via the rearrangement of isomeric materials. Isomerization of the allene, A_, would be the simplest route. The required chlorides are simply the addition products of methylene to 1-chlorocycloheptatriene . Although this compound, free of other isomeric chlorides, has been synthesized 15 by Untch, the synthesis requires many steps and results in a low yield. A more economical method of preparation Ci f^\ Q might be through separation of the isomers of 1-, 2-, and 3-chlorocycloheptatrienes , synthesized by Folisch from tropone. Reaction of this material with Simmons-Smith reagent should yield the desired compounds. Ci C-jOp cu LA H II \ . CI Chlorotropyl ium chloride was synthesized from tropone by reaction with oxalyl chloride in methylene chloride. This was reduced to the above mixture of chlorocycloheptatrienes with lithium aluminum hydride. Before an attempt was made to separate the isomers, the mixture was tested

PAGE 20

1 1 with Simmons-Smith reagent to insure its reaction. Copper1 7 zinc couple was prepared by the method of LeGoff and its reactivity with methylene iodide confirmed in the presence of cyclohexene. Repeated reaction of the reactive couple / I CI Z"Cu — XCH 2 , 2 with methylene iodide in the presence of the chlorocycloheptatriene mixture resulted only in the recovery of the starting chlorides. Attention was then diverted to preparation of 3 via P 7 \ H 7 \ -,0 C'N-N-Tos h' Na W ,N-N-Toj C ,' Nq H rearrangement of cycloheptatr ienylcarbene . Lindsay and 1 8 Reese have synthesized 2and 3-bromocycloheptatr iene by heating 7 , 7-dibromobicyclo [4 . 1 . 0] -2-heptene in quinoline. Dibromocarbene was generated from bromoforrn by reaction with potassium 5-butoxide in pentane, and allowed to react with 1 , 3-cyclohexadiene , yielding this bicyclic di bromide. This was heated in quinoline, under vacuum, distilling off the

PAGE 21

12 mixture of bromides as they were formed. This mixture was then allowed to react with methyllithium in THF at -78°. \L \ — / / V — V ^ 4> / ^ Addition of dimethylf ormamide to this lithium reagent at this temperature and hydrolysis with aqueous ammonium chloride yielded a mixture of 2~, and 3-f ormy lcycloheptatr ienes The tosylhydrazones were generated from reaction of this mixture with tosylhydrazi.de in ethanol and the sodium salts by reaction with sodium hydride in THF. It can be seen from the scheme above that isolation of pure isomers of these sodium salts is not necessary; both carbenes generated from these salts should yield the same homoaromatic carbene upon rearrangement. This mixture of salts was thermally decomposed by refluxing in diglyme. The resulting material was taken up in water to dissolve the sodium toluenesulf inate formed and the aqueous phase was extracted with pentane. No products could be found in the pentane fraction. Extraction of the aqueous phase with ether yielded an oil containing two components. These were identified as C II N isomers by high o o Z resolution mass spectrometry.

PAGE 22

13 Assignment of the structures as pyrazoles 6 and 7 was accomplished by nmr . Two signals appeared in the region expected for resonances corresponding to the methylene protons. The triplet at 2.68* was assigned to the protons on C 4 of isomer 1_ flanked by the protons on C. and C r ; while the doublet at 3.4 7 corresponded to the hydrogens on C, of isomer 6, adjacent to the proton on C . Two singlets at 7.5 and 7 . 6, corresponding to a single proton each, were assigned 7.N, 'N N-H H * to isomers 6 and 7, respectively, by comparison of the area of the integrated signals with those already assigned to the C A11 proton shifts are reported in units of 6.

PAGE 23

14 methylene protons. These were determined to be the protons on Cg by comparison with the spectra of substituted pyra19 zoles . These protons, in orther pyrazoles, center around 7.6. The vinyl signals could not be assigned to their particular isomers. Formation of these pyrazoles is explained in the scheme above. Loss of sodium toluenesulf inate from the salt mixture yields the corresponding diazo compounds. These can then close to yield the indicated 3-H pyrazoles. Tauteromerism via migration of the indicated proton yields 6 and 7 directly. The formation of pyrazoles by this scheme is a well documented barrier to the formation of carbenes from tosyl20 hydrazones. Bartlett and Stevens have noted the tendency of diazo compounds generated from tosylhydrazones of a , B-unsaturated aldehydes to close in this way to yield the corresponding 3-H pyrazoles. While there seems to be no way to overcome this difficulty in the thermal decomposition of 21 the sodium salts, low temperature photolysis has been utilized successfully in some cases to yield carbenes from systems that otherwise give only these pyrazoles. When our salts were photolyzed through Pyrex in THF and worked up in a manner similar to that described above, a mixture of two new compounds was isolated. Mass spectra of this material confirmed that these were also C„H N isomers. o o Z Again no evidence of compounds not containing nitrogen was seen .

PAGE 24

15 These new pyrazoles, 8 and 9, were shown not to be the intermediate 3-H pyrazoles postulated above by the presence of a N-H stretch in the infrared at 3200 cm . Neither did they rearrange to 6 and 7 by heating in refluxing diglyme. The nmr did show a superimposed doublet and triplet between 2.7 and 3.0 that appear to be the protons on the methylene carbons, and signals at 7.3 and 7.5 that appear to be the protons on C g . The other signals, again, could not be assigned to any isomer with any certainty. These pyrazoles might result from a photoinduced tauteromerism of 6 and 1_. Tauteromerism of pyrazoles of 22 this type is well known. Attempts to alleviate the cyclization problem by low temperature photolysis failed. Because of the difficulties involved in the low temperature photolysis, a new approach to the problem was taken. Ci Cl Cl CI O ® O o (drS 1 pro]

PAGE 25

16 23 Billups and Reed has shown that it is possible to generate both carbene participants of a carbene-carbene rearrangement from the intermediate cyclopropene, formed by the dehydroa KOR 10 halogenation of a di-halocyclopropane (vide supra). This suggests the possibility that it might be possible to generate the homoaromatic carbene by dehydrohalogenation of an 8-halobicyclo [ 5 . 1 . ] 2 , 4-octadiene . Synthesis of 10^ was attempted by the addition of chlorocarbene to cycloheptatriene . This reaction yielded a yellow oil. Elemental analysis and high resolution mass spectrometry showed that this material was a C„H CI isomer, as o y expected. The nmr implied the presence of two isomers. A multiplet at 1.3-1.8 was assigned to the two bridgehead protons and a multiplet centered at 2.5 was assigned to the methylene protons on C,. The vinyl protons were represented by a multiplet centered at 5.8. Two apparent triplets at 3.6 (J = 3 Hz) and 3.0 (J = 8 II z ) appeared in the area of the spectrum in which the proton on the carbon bearing the chlorine should lie.

PAGE 26

17 E a u -p o 0) a 10 u E C

PAGE 27

18 It was disturbing that the signals that correspond to the protons geminal to the chlorine in these compounds appeared to be triplets. Simple first order effects would lead to the conclusion that these protons would be represented as doublets of doublets if the carbene had inserted, as O cl expected, into the 1,2 bond of the cycloheptatriene . If insertion had occurred into the 3,4 bond of the cycloheptatriene, the symmetry of the resulting chlorides would be expected to result in triplets corresponding to these protons. Two proofs of structure for this molecule were attempted. If insertion of chlorocarbene in the 1,2 bond had occurred to yield 3^0, the resulting molecule would contain a conjugated diene. This diene would be capable of undergoing a Diels-Alder reaction with a dienophile. When the chloride was allowed to react with maleic anhydride, a solid material was isolated that was shown by mass spectrometry and nmr to be the expected adduct. Fragmentation of the molecule in the mass spectrometer was indicative of its structure. The molecular ion (m/e 238) lost CO and CO,, to yield a new ion of mass 166. This fragmented to C.H CI and benzene (parent Ck^O C i © ._.,,. ( I

PAGE 28

19 ion), implying the structure above. The nmr supported this structure and the infrared showed the expected anhydride doublet at 1775 and 1850 cm" 1 . Further support of 10 as the structure of this oil was obtained by synthesis of a model compound. The known isomer, 8 , 8-dibromobicyclo [ 5 . 1 . ] 2 , 4-octadiene , was synthe24 sized by the method of Van Vooren et al. This was allowed to react with freshly prepared tri-n-butyltin hydride, to yield a mixture of monobromides . The nmr of this mixture was identical to that of the mixture of chlorides. 25 Closs et al. have shown in studies of the 7-chlorobicyclo [ 4 . 1 . ] heptane system that the coupling constant of the H_ signal in the isomer, in which H, , H , , and the chlo/ lb rine atom are cis, has a coupling constant of 3 Hz. From this information, the material isolated from the chlorocarbene reaction was assigned the structures 11 and 12. Isomer 11 , in which the protons on the cyclopropane are assumed to be cis (J = 8 Hz), is the major constituent, comprising 75% of the isomer mixture. This mixture could not be separated by chromatographic methods or spinning band distillation.

PAGE 29

2 11 Ci / _ V 12 The mixture of JL1 and 12 was subjected to five equivalents of potassium t-butoxide in THF at reflux. Work up with water and extraction into pentane yielded a small amount of cyclooctatetraene and styrene as the only isolable products. The cyclooctatetraene was determined not to rearrange to styrene under these conditions. The identity of these compounds was confirmed by the nmr spectra and comparison of their retention times to those of authentic samples upon gas chromatoqraphy . The balance of the material remained as an intractable polymer. Cl koi-C,h c I + Because of the low yield of products from this reaction an experiment was undertaken to determine if preferential reaction of one of the isomers in the mixture was taking place. To a solution of the above chlorides in THF at room temperature was added only a half equivalent of potassium t-butoxide. This was worked up with water and pentane to ultimately yield an oil containing cyclooctatetraene and a mixture of the above chlorides, heavily enriched (92%) in isomer 11 .. The cyclooctatetraene was removed by careful bulb to bulb distillation. This enriched mixture was then

PAGE 30

21 LO

PAGE 31

22 allowed to react with five equivalents of potassium t-butoxide in the original manner. Only a trace of cyclooctatetraene was found after work up, an amount that could easily be explained by reaction of the residual amount of 12 in the starting mixture. No residual 11 was isolated 11 12 11 from this reaction. Attempts were made to trap either the cyclopropene or one of the possible carbenes formed in this reaction. No trapped products were seen using styrene, tetramethylethylene, t-butyl alcohol, butadiene or cyclopentadiene . Cyclooctatetraene and styrene can be explained by a scheme similar to that derived by Billups in his synthesis of benzocyclopropene . He has shown that 1J3, when allowed to react with potassium t-butoxide in DMSO, undergoes a dehydrohalogenation to the cyclopropene, which rearranges to the methylenecyclopropane moiety shown. Repetition of this 0> c 'Q> CI ~ sequence yields benzocyclopropene. A scheme in which an allyl hydrogen is removed by base, with concerted loss of chloride ion to yield an intermediate chlorocycloheptatriene has been discounted by labeling stud27 les of Prestien and Gunther aru | the absence of heptafulva2 8 lene, which would have been formed under these conditions.

PAGE 32

23 HO h '<£ C I CI Cp Loss of HC1 in _12 can occur in one of two ways. If loss of the proton on C occurs first, the resulting C -C / 7 8 cyclopropene has the choice of two possible pathways. The homocycloheptatrienylidene , 2, would result from the ring expansion of the cyclopropene to the carbene. This carbene would be expected to rapidly rearrange to yield cyclooctatetraene. Base assisted isomerization of the cyclopropene 2g to the methylenecyclopropane yields the bicyclic triene, 14 . This moiety can then undergo a norcaradiene rearrangement to the bicyclobutane shown. This, when it opens in its [' > 12 V (i v V_:/ ^ \~y ] 4 /,,A

PAGE 33

2 4 usual manner, yields styrene. Loss of the hydrogen on C, to ultimately yield a C,~C -L 18 cyclopropene , has the same options as the cyclopropene postulated above. Base assisted isomerization again yields 14 and ultimately styrene. This time, ring opening to the carbene yields homocycloheptatrienylidene , 3. This carbene does not possess the ability to migrate hydrogen as 2. However, it is apparent, upon study of molecular models, that the empty orbital of the carbene and the hydrogen on the methylene carbon are in excellent alignment for insertion of the carbene into this C-H bond. Intramolecular insertions of carbenes into carbon-hydrogen bonds in saturated and unsaturated rings are not only well known, but this reaction is competitive with possible 1,2 hydrogen migrations. This insertion gives the bicyclic octatriene shown. This species is the well known valence isomer of cyclooctatetraene. Simple electrocyclic opening of the cyclobutene moiety results in cyclooctatetraene . It is not unexpected that the above carbenes were not trapped by any olefin or alcohol. The inability to trap carbenes of this sort has become a generality. With few exceptions, it can be stated that if a singlet carbene has an internal insertion or cycloaddition available, it. will not be possible to trap it with an external reagent.-^-'A direct opening of the chloride by reaction with the base appears extremely unlikely in light of the labeling studies in the benzocyclopropene system mentioned above.

PAGE 34

25 However, in the case at hand, it is not possible, due to the unique symmetry of the cyclooctatetraene molecule, to carry out labeling studies of the type performed by Prestien and Gunther to disprove this pathway. It is apparent from the low yield of cyclooctatetraene and styrene that some other route is available for reaction of the chloride. An interesting similarity exists in intermediate 14 and intermediate 15, below. Washburn et 3 2 al. has studied the dehydrohalogena tion of the dibromide, 16 , and determined, by studies using labeled dibromides, that f>*ery 1 ROH OR R-0 H 'C 16 15 the reaction proceeds via the route shown below. Rearrangement of 15 to 8-t-butoxyf ulvene is a simple electrocyclic ring opening of the methyl enecyclopropane . If 14 does open, or could be induced to open in an analogous manner, it would / \ ,p yield heptaf ulvene . This highly reactive material would polymerize rapidly under the reaction conditions.

PAGE 35

2 6 An attempt was then made to induce reaction of the chlorides 11 and 1_2 to heptafulvene while minimizing destruction of any products formed. The reaction was carried out in a solution of potassium t-butoxide in tetraglyme, heated to 90° under vacuum to distill off volatile products into a liquid nitrogen cooled trap. Immediately upon injection of some of the mixture of chlorides into this solution via syringe, a scarlet material was noted in the liquid nitrogen trap. Upon warming, the scarlet material polymerized. Gas chromatography showed the presence of cyclooctatetraene and styrene in the trapped material. An attempt was then made to trap this material by the 3 3 method of Doering and Wiley. The above reaction was repeated, and the contents of the cold trap were washed into cold THF containing dimethyl acety lenedicarboxylate . After standing at 0° for 72 hours, the red color had disappeared. Exchange of the solvent with benzene and air oxidation using Pd/C catalyst yielded a blue solid. This was identified, 3 3 by comparison of its uv with the literature, as dimethyl azulene-1 , 2-dicarboxylate . When the red material was trapped in an nmr tube Li W E5-C5C-Es Es = COOMe containing THF and a trace of triethylamine , an nmr was taken that showed signals corresponding to those reported for heptafulvene 3 4

PAGE 36

27 When this reaction was repeated with pure 11 under the same conditions, a very small amount of colored material was obtained, not enough to identify by the methods outlined above . It is obvious from the preceding reactions that only one isomer, 12 , is yielding isolable products. Preferential reaction of this isomer in which the hydrogen and chloride to be eliminated are coplanar might be expected. It is unusual that reaction of isomer LI does occur, albeit slowly, but does not give the same products. No evidence can be found in the literature on the relative rates of elimination between the isomers of monohalocyclopropanes . While the number of reactions is 35 small, one set of experiments has shown that elimination, when the hydrogen vicinal to the halide is trans, can occur and yields the cyclopropene . The reaction pathway of isomer 11^ cannot be suggested, since the only material isolated from its addition to base is an intractable polymer. In order to increase the yield so that the reaction could be more readily studied, a system was then sought in which the configuration of the eliminated moieties could be assured to be cis. The simplest system of this kind would be an 8,8-dibromobicyclo [ 5 . 1 . ] 2 , 4-octadiene, 17 . In this molecule, one of the halogens will always be cis to the hydrogen to be eliminated.

PAGE 37

28 When this molecule was subjected to potassium t-butoxide in THF at 0° and worked up with water and pentane as outlined above, a very small amount of two compounds could be isolated. These were separated by TLC and their structures identified as cis and trans fi-t-bu to xystyrene by nmr and mass spectrometry. No bromocyclooctatetraene was noted in this reaction. Since the mechanisms of these reactions were merely R R speculative at this time, it was then set about to confirm intermediates along these suspected pathways. The first intermediate, one common to both systems, is the cyclopropene formed from the initial dehydrohalogenation . The first attempt was made to trap this cyclopropene by a Diels-Alder addition to 1 , 3-diphenyli sobenzof uran, 3 7 34 . It has been shown that 34 is a potent diene in Diels-Alder reactions. Its large molecular weight would make isolation of the trapped material possible by crystallization. The mixture of chlorides, 11 and 12 (1 equivalent of 12), was allowed to react with five equivalents of potassium

PAGE 38

29 t-butoxide and one equivalent of 3_4 at -78° in THF. Along with some cyclooctatetraene and a great deal of 11, a crystalline material was isolated after work up with ether. Chromatography and recrystallization allowed isolation of the crystalline material which was identified as the 34_ adduct of the cyclopropene , 1_8. Elemental analysis and high resolution mass spectrometry confirm the C„ o H~„0 formula. 2 o 2 2 The proton spectrum showed the expected shifts for this molecule, a complex multiplet for the aromatics from 7.0-8.0 (14H) , a vinyl multiplet at 5.8-6.2 (4H), a doublet at 1.7 (1H) and a complex multiplet from 2.1-3.2 that corresponds to three protons. Confirmation of the exact structure could not be made from this spectrum alone. As mentioned above, dehydrohalogenation could occur across either C,-C Q or C n -C to / 8 to yield the cyclopropenes below. Either of these materials, when trapped with 3j4, could give signals of this type in the proton spectrum. The complex coupling precluded a simple assignment of structure. Spin decoupling allowed the assignment of this molecule as structure 1_8_ to be carried out. The complex multiplet from 2.1 to 3.2 was suspected to be formed from the signals

PAGE 39

3 o

PAGE 40

31 of the methylene protons (H ) and the bridgehead proton (H_) . When the doublet at 1.7, assigned to H„ , was irradiated, the far downfield portion of the multiplet simplified to a doublet of doublets (J=4.5, 9 Hz.). This would denote that this proton (H 7 ) while being adjacent to H , was also being split by two non equivalent protons (H ) . Irradiation of the oa , d downfield portion of the vinyl signal (H r ) resulted in two overlapping doublets of doublets (J=4.5, 14 Hz.; 8.5, 14 Hz.) in the upfieid portion of the 2.1-3.2 multiplet. These two non equivalent protons (H ) were being split, not only by an adjacent proton (H ) , but also by each other. This is the reason for the identical coupling constant (J-14 Hz.) in each of these patterns. This type of coupling cannot be explained by a product resulting from the addition of 34 to the C -C cyclopropene. Similar reaction of 8 , 8-dichlorobicyclo [ 5 . 1 . ] 2 , 4-octadiene with five equivalents of potassium t-butoxide and iBF results in a crystalline material after the same type of work up. An almost identical material is isolated from the reaction of 1]_ in a similar manner. Each of these compounds is a C 28 I, 21 OX isomer with almost identical proton spectra. These isomers were assigned structures 19 and 20. The molecular formulas were confirmed by elemental analyses.

PAGE 41

32 u >i u •H XI O e o J-l eq c

PAGE 42

33 The proton spectrum of the bromide, 19, as in the case of 18, could not unequivocally establish the structure. Each of the protons could be assigned to a signal but the splitting was not simple enough to define the structure exactly. A two proton multiplet centered at 2.6 was assigned to the methylene protons on C g and a multiplet at 3.2 was assigned to H each of these signals appearing at a similar shift to the correspondinq protons on 1_8. The aromatics again appear in a multiplet from 5.9-6.5. A glaring deviation from the vinyl signals of 18 existed in the appearance of a new doublet at 5.5. This signal could be assigned to H . A more complete discussion will appear later. The ability to work with a greater amount of the reactive isomer (vide supra) in the dibromide system, resulting in a greater yield of trapped material, allowed studies of the carbon-13 spectrum. The distinct dissimilarity of the proton vinyl signals raised a real question as to the similarity of structures assigned to this bromide and the hydrocarbon, 18. The absence of bromocyclooctatetraene in any reaction of the dibromide, 17 , with base suggests that rearrangement of the cyclopropene to the methylenecyclopro-

PAGE 43

34 pane must be a rapid reaction. If this is true, then this trapped bromide might be formed from the methylenecyclopropane rather than the cyclopropene . The bromide expected from the trapping of a methylenecyclopropane with 34^ would be 21. This could be easily differentiated from the trapped cyclopropene by the fact that this molecule contains three methyne carbons (C, , C , C ) and one quaternary carbon in what would be expected to be the upfield portion of the carbon-13 spectrum. Structure 19 contains one methylene carbon (C ) , one methyne carbon (C ) , and two quaternary carbons (C., C ) in the corresponding portion of the carbon spectrum. The carbon-13 spectrum of this bromide showed a triplet at 27.873 (ppm downfield from tms) , a doublet at 35.401 and two singlets at 41.372 and 48.461. The triplet was assigned to the methylene carbon (C fi ) on 19, the doublet to the methyne carbon (C_) and the singlets to the quaternary carbons (C, , C„) . 1 a Final confirming evidence for 1_9 as the structure of this solid was found in the spin decoupled proton spectrum. When the signal at 2.6 (H,) was irradiated, the vinyl multiplet split into a doublet and a triplet, denoting that the

PAGE 44

35 protons on C g are the only protons adjacent to the vinyl system. The dissimilarity in the vinyl signals of the proton spectra of 18_ and 19 can be explained when models of these two systems are constructed. In structure 1_8, there are two conformations with respect to ring flip of the sevenmembered ring. In one of these conformations, one of the protons on Cg is almost coplanar with H . In this conformation, the proton on C„ is removed from the proximity of the neighboring phenyl ring, eliminating any steric interaction. When the same conformation is attempted in compound 19, the proton on C g is now almost coplanar with the bromine. Indeed, this exact conformation cannot exist because of the steric bulk of this halide. A ring flip to the other conformation eliminates this steric problem, but in doing so, forces the proton on C 2 into the shielding area of the neighboring phenyl ring. This, in turn, shifts this proton signal upfield, generating the doublet at 5.5 that was noted in the spectrum of 18. The next step, following the proposed mechanism, would be to trap the methylene cyclopropane, if possible. This was attempted in the dibromide system because of the aforementioned reactivity of this material. At 0°, one equivalent of 1/7 was allowed to react with five equivalents of potassium t-butoxide and one equivalent of 34. Again a crystalline material was isolated in the same manner as before. The structure of this material was determined to be

PAGE 45

3 6 a ,c o £ o u a

PAGE 46

37 22 . High resolution mass spectrometry confirmed a C H OBr 2 8 21 formula. The proton spectrum showed no upfield protons, implying that the cyclopropane ring had been broken. The vinyl signals consisted of a six proton multiplet between 5.2-6.5, the aromatics, a fourteen proton multiplet between 7.1-3.0 and the remaining signal was a singlet, corresponding to one proton at 4.5. The carbon-13 spectrum supported this structure. The 3 upfield portion of this spectrum indicated four sp carbons. Two of these signals could be assigned to the bridgehead carbons of bicyclic ether by their similarity to the respec3 tive carbons in structure 19. The only sp signal that de[f" 2 2 noted a proton bearing carbon was a doublet at 66.101 that was assigned to C-, . The remaining quaternary signal (56.898 was assigned to C ? . This molecule is a labile material and readily aromatizes with loss of HBr to yield a more stable new material . This material, a C no H„ n O isomer, 23, is isolated after Zo Zu chromatography on silica gel of the crude material from the reaction of 17 with potassium t-butoxide and 34 at 0°, although it cannot be detected in the nmr of the crude material before chromatography. This halogen-free substance

PAGE 47

33 is the only material isolated from a solution of 2J2 in acetonitrile allowed to stand at room temperature for two days. Spectral data allowed structure 22_ to be assigned. Elemental analysis and high resolution mass spectrometry confirm the C„ o H„ A formula. The infrared spectrum indicated 2 o Z (J a carbonyl stretch at 1670 cm , corroborated by a large loss of Ph-Co(+) from the molecular ion in the mass spectrum. The proton spectrum was simple, containing a multiplet between 6.9 and 7.7, implying a high degree of aromatization and no indication of protons on vinyl positions. The only other proton signal was a singlet at 5.5 corresponding to one proton. The only assignable signals in the carbon-13 spectrum were the carbonyl carbon at 198.97 and a proton bearing carbon at 6 3.44. These data can be fit to a 1,2,3 trisubs ti tuted indene in which the substituents are two phenyls and a benzoyl. Synthesis of l-carbomethoxy-2 , 3-diphenylindene has been 3 ft reported and the proton on C, is a singlet at 4.8. This molecule was used as a basis for calculation of chemical 39 shifts by the additivity rules of Shoolery. " The additivity factor was taken as the difference in the shifts of the V* methyl signals of acetophenone and methyl acetate (+.6 ppm) . Using this information, the calculated shift for this proton in l-benzoyl-2 , 3-diphenyli ndene would be 5.4.

PAGE 48

39 o N c 0) CQ V u a a.

PAGE 49

1 The isolation of 2_3 from 2_2 in acetonitrile shows that a solvolytic dehyclrohaloqenat ion is one possible mechanism. Loss of the bromide would yield the carbonium ion below. Isomerization of the cycloheptatriene ring to a norcaradiene yields a cyclopropylcarbinyl cation. Interconversion of this carbon to the allylcarbinyl cation results in a species that can lose a proton to aromatize. This new compound, 24, could protonate on the oxygen and rearrange, as outlined below, to 23. Compound 2/4_ should be stable in the absence of acid. If solvolysis of 2^ occurs under the basic conditions of the reaction in which 22 is generated, 24 may be an isolable product . Benzyne was allowed to react with 2 , 3 , 5-triphenylf uran to yield a colorless solid whose nmr and mass spectra are compatible with the structure of 24. The proton spectrum of 'TO 4< w r * ^ H 24 2 3

PAGE 50

41 this material consists of a multiplet between 6.9-7.9. The carbon spectrum shows the bridgehead carbons (92.416 and 94.950) at positions comparable with those on 19 and 22. The rest of the spectrum cannot be assigned. The intermediacy of this material in the solvolysis of 22^ was refuted by its stability to silica gel chromatography and its inertness in acetonitrile solution. Refluxing this material in acidic methanol produced no 23. With this in mind, it is more likely that the actual mechanism of the solvolysis of 2/2 resembles that below. Loss of bromide ion yields a non classical carbonium ion that $ p opens, as indicated, and rearranges to the ketone. The cyclopropylcarbinyl cation formed from rearrangement of the cycloheptatriene to norcaradiene , rearranges to the ailylcarbinyl cation, which aromatizes to 23. Dehydrohalogenation in the systems of 1_2 and 17 seem to follow a similar pathway. The first step is dehydrohalogenation across the C -C bond to yield the cyclopropene . This, in the case of the halogen free cyclopropene, 25, has the ability to open to the homoaromatic carbene 3, which can intramolecularly insert into the C-H bond of the methylene across the ring to form the bicyclic triene, which can then open to cyclooctatetraene. Competing with this reaction is

PAGE 51

42 \-0 C-0 (fXM 'DO * 23 the base induced isomerization of the cycloprooene to the nethylenecyclopropane. This moiety can apparently undergo electrocyclic ring opening to yield a heptaf ulvene . Both of these heptaf ulvenes have been trapped, 2j3 with the acetylene diester yielding the substituted azulene, and 29 with iBF to yield 2_2. The styrene formed in the reaction of 12 with base in the absence of trapping material appears to arise from a bicyclobutane formed by norcaradiene rearrangement o* -

PAGE 52

43 of 14 . This would be consistent with the increase in yield of styrene noted by increasing the temperature of the reaction. The absence of bromocyclooctatetraene in the reaction of 1/7 with base must be recalled. It is unlikely that the difference in the eliminated halide is changing the course of the reaction, since trapping of the halocyclopropene was accomplished regardless of the halide eliminated. The difference in these reactions is most likely caused by the rate at which the cyclopropene isomerizes to the methylenecyclopropane. The intermediate anion in this isomerization can be stabilized by resonance through the pi system and onto the carbon bearing the halide. The halogen can stabilize the anion at this position by induction. The overall effect is to enhance the ability of the cyclopropene to isomerize through this intermediate. This allows the cyclopropene to isomerize before opening to the carbene can occur. The question of the genesis of the styrene ethers has still not been addressed. These ethers are only apparent in the reaction of 17 with no trapping material. At no time during the isolation of 1_9 or 22 v/ere either of these compounds noted. It is apparent that to isolate these compounds, neglecting nucleophilic substitution of the t-butoxy group

PAGE 53

4 4 for bromide, that two dehydrohaloqenations must have occurred. Two plausible mechanistic pathways can be postulated. The cyclopropane previously shown to have been found in this reaction can either be trapped with t-butyl alcohol to yield 3_0 or rearrange to 21_. A second dehydrohalogenation of 3Tj with accompanying rearrangement to the methylenecyclopropane yields 3_1. This intermediate can also be formed from a dehydrohalogenation of 2J7 to yield the tetraene, 32, which can be trapped by t-butyl alcohol in a manner reminiscent of that of the trappinq of m-benzyne to yield 15. Compound 31_ can then aromatize, in the manner postulated earlier for the formation of styrene from the monochloride fy. _ rf 17 Br /or 30 27 / 33 system, to yield the styrene ethers. OH \ /. OH 31 To trap the progenitor of the ethers would require delay of inclusion of the trapping material until the bromocyclopropone could react further and the second dehydrohalogenation to either 32 or 33 could occur. The reaction must

PAGE 54

4 5 be done at a low temperature to prevent the ring opening to bromoheptaf ulvalene known to occur at 0°. At -78°, the dibromide 1/7 and five equivalents of potassium 5-butoxide were stirred for one hour, then, while still at this temperature, one equivalent of 34 was introduced. When the reaction had warmed and was worked up with water and ether in the usual way, a large amount of polymer was noted. The only isolable product was a small amount of 22. It is apparent that trapping of the doubly dehydrohalogenated material is not taking place. This is unfortunate in that identification of this material, which must be a precursor of the styrene ethers, cannot be made.

PAGE 55

AN ATTEMPT TO GENERATE BISNORCARADIENE Introduction Bisnorcaradienes, 2 , 4 , 8 , 10tricyclo [5 . 4 . . ' ] undecatetraenes, 3J5, have been postulated as intermediates to explain the scrambling of substituents in substituted benzocycloheptatrienes . These interesting intermediates were first postulated in the thermal isomerization of di-deutero3 , 4-benzocycloheptatetraene by Gruber and Pomerantz, where it was noted that deuterium substituted on the cyclohepta3 7 33 triene ring of 3 , 4-benzocycloheptatriene (37; R, , R = D; Rp/ R^ = H) would rearrange to the benzene portion upon heating at 405° for 10 rain. They postulated that the reaction proceeded through the scheme above. The next system thought to involve bisnorcaradienes was 4 2 that of Durr et al. in which a cyclopentadienylidene was allowed to react with benzene to ultimately yield a spiro compound that was thought to rearrange to the bisnorcaradiene by photolysis. This photolytic reaction of the spiro compound was thought to proceed through the di-pi-me thane rearrangement to yield the bisnorcaradieno directly. 4

PAGE 56

4 7 N, f -35 4 3 In 1974, Durr et al. were able to isolate a material from the photolysis of the spiro compound, below, at -70°. This material was identified by its nmr at -60° as the J .LAJJ — CX bisnorcaradiene. When this material was warmed to -10°, it isomerized to the benzocyclohepta trienes . Also during that year, Mitsuhashi and Jones isolated substituted benzocycloheptatrienes that were postulated to have been formed via an intermediate bisnorcaradiene, from the addition of cycloheptatrienylidene to tetracyclones . 35 (R x , R 4 = 45 In 1976, Pauly and Durr reported the isolation of material that he identified by proton nmr as the first stable bisnorcaradiene This material, which melts at 104° , has a nmr spectrum similar to that of the material isolated in the low temperature work mentioned above. 46 f *!>/* — i o •// \ -3 5 I' R 4

PAGE 57

48 A possible property of this intermediate that has not been explored is the potential of 3_5 to undergo a degenerate [5,5] antara, antara, sigmatropic rearrangement. This reaction would be a symmetry allowed, concerted reaction. It was with this possibility in mind that the system 35 (R, ,R =benzo) 47 outlined below was devised. This system would allow studies to be made of the possibility of bisnorcaradiene formation in the parent system at modest temperatures, and determination of the O 'A ability of this bisnorcaradiene to undergo this degenerate, semibulvalene type of rearrangement. The known first order / \ \ I 39b 7 '~A D Pi 3 9c V / \ I I 3 9b

PAGE 58

49 proton spectrum of 3 , 4-benzocycloheptatriene would allow simple assignment of the positions of the deuterium atoms in the rearranged molecules. Entrance to this system would be made through compound ^3 8_. This 1 , 3 , 5 , 8 , 10-bicyclo [5 . 4 . ] undecapentaene would have the potential to rearrange to bisnorcaradiene in the manner of the substituted analogues noted above. Compound 38 could be formed by the addition of cycloheptatrienyl idene-cycloheptatetraene to 3 , 6-di-deutero-2-pyrone . The resulting lactones could be decarboxylated to yield 38. If the bisnorcaradiene were formed in this system, the resulting benzocycloheptatrienes should show deuterium in the cycloheptatr iene ring (39a, 39b) . This effect would be evident in the changes in the splitting patterns of the vicinal hydrogens. If the [5,5] rearrangement should take \ / ,0 \ / ./ -J 'I \ place, a new splitting pattern for H,-H, should become 1 o evident due to the introduction of a new molecule with deuterium at C 7 ( 39 ) . Results and Discussion Cournalic acid, prepared by the method of Wiley and Smith from malic acid, was decarboxylated at 650° through 49 copper turnings to yield 2-pyrone.

PAGE 59

so Attempts to react 2-pyrone with cycloheptate traene generated from chlorocycloheptatrienes and potassium t-butoxide were unsuccessful due to the instability of the 2-pyrone in base. The 2-pyrone was reactive to cyclohepo Oil HOOC-CH^CH-COOn coon o \ tatraene generated from the thermolysis of the sodium salt of tropone tosylhydrazone . ' When this salt was added to a solution of ten equivalents of 2-pyrone in diglyme at 110°, 3,4-benzocycloheptatriene was isolated, after work up with water and pentane, in 30 & yield. Similar results were seen after photolysis of the sodium salt through Pyrex in THF in the presence of ton equivalents of 2-pyrone. Again, 3,4-benzocycloheptatriene was isolated as the major component, but 1 , 2-benzocyclo\ /. \ / heptatriene was also formed. These results were noted in photolyses at temperatures as low as -78°. Photolyses were not as clean as the thermol y ses, and polymeric materials were also isolated. The mechanism of this reaction should be that proposed above, addition of the allene to the pyrorie to yield the lactone and decarboxylation to the bicyclic pentaene.

PAGE 60

51 A // o O—ii \ / exT) . CO Hydrogen can migrate to yield the benzocyclohepta triene , or the ring can close to the bisnorcaradiene . The presence of the 1 , 2-benzocycloheptatriene apparently results from a photoinitiated 1,3 hydrogen shift in the bicyclic pentaene. o (i \ It was necessary to use an excess of 2-pyrone in the photolysis reaction because of the photolabili ty of this molecule 52 It can undergo [2 + 2] cycloaddi tion intrarnolecu43 ,, p., larly to yield a lactone that decarboxylates to cyclobutadiene .

PAGE 61

52 52 Chapman et al . have been able to synthesize di-deutero-2-pyrones by sequential deuteration of the pyrone, frc methods developed by Pirkle and Dines. 53 coon .0 o o o o \\ 'Cl \\ /) ;---[) \\ /> -'-* D {\ /) Br -"-d/ \o Synthesis of 6-d-2-pyrone , 4_0 was accomplished from glutaconic acid by the method shown above. Before the second deuteration was attempted, this material was allowed to react with cycloheptatetraene to determine if rearrangement through a bisnorcaradiene was taking place. Formation of the bisnorcaradiene would cause some of the deuterium to become incorporated into the cycloheptatri one portion of the product, after rearrangement. Reaction of an equivalent amount of sodium salt with 40 in diglyme at 110° yielded the benzocycloheptatriene , as expected. There was no change in the splitting patterns of the proton spectrum, although the mass spectrum showed that the deuterium had been retained. Integration of the proton spectrum indicated that all of the deuterium was incorporated in the benzene portion. This implied that the 1,5 hydrogen migration, resulting in the aromati zation of the pentaene was much faster than the closure to a bisnorcaradiene. Since this migration is thermally initiated, an attempt was made to generate the cycloheptatetraene by photolysis of the sodium salt at -78°, as before. The scarcity of the deuterated pyrone precluded

PAGE 62

53 the ability to work with the excess of starting material enjoyed in the reaction of the non-deuterated pyrone . Photolysis of an equivalent amount of the sodium salt and 4_0 at -78° did not yield enough benzocycloheptatriene to identify. No pyrone was recovered after the reaction. An attempt was then made to determine if the closure to bisnorcaradiene could become competitive with aromatization by increasing the temperature of the reaction. At temperatures as high as 200°, aromatization remained as the only observable pathway of the pentaene. It appears that the energy required to drive the pentaene toward aromatization and the energy required for closure to a bisnorcaradiene must not differ greatly. The effect of aryl substitution in this system is apparently enough to induce bisnorcaradiene formation by either retardation of aromatization or acceleration of closure. The nature of this is not obvious.

PAGE 63

EXPERIMENTAL General . All melting points were obtained on a Thomas Hoover melting point apparatus and are uncorrected. Nuclear magnetic resonance spectra (60 MHz) were recorded on either a Varian A-60A or JEOL PMX 60 spectrometer and are reported in ppm (6) downfield from tetramethylsilane, the internal standard. Carbon-13 and 100 MHz proton spectra were recorded on a JEOL FX 100 spectrometer and are reported in ppm downfield from tetramethylsilane. Infrared data were recorded on a Beckman IR 10 spectrophotometer. Combustion analyses were performed by Atlantic Microlab, Inc., Atlanta, Georgia. Mass spectra were obtained on an AEI MS 30 spectrometer. Cycloheptatriene was purified by distillation through a column of glass beads. All tetrahydrof uran (THF) was distilled from lithium aluminum hydride immediately before use. Diglyme and tetraglyme were distilled from calcium hydride and stored over sieves until used. Dimethylf ormamide (DMF) was stirred for 24 hrs over 3A molecular sieves and calcium hydride, decanted onto barium oxide and distilled at 20 mm into septum bottles containing molecular sieves. Other commercially obtained reagents were used as received. Chromatography was carried out on Baker 60-200 mesh silica gel or Fisher basic alumina. The progress of column chromatography was followed by the addition of 1% type 54

PAGE 64

55 609 phosphor (DuPont Photoproducts Department) to the adsorbent, employing a quartz column, and observing with a hand-held UV unit (Fisher Scientific) . Materials noted as vacuum dried were placed in a vacuum oven at 130° for at least 12 hours. P reparation of tropone . In a 2 liter round-bottomed Morton Flask equipped with a mechanical stirrer and a reflux condenser was stirred 13.5 g (0.10 mol) of KH PO in 33 ml of water and 350 ml of dioxane. To this was added in one portion, 46 g (0.50 mole) cycloheptatriene and 54 g (0.49 mol) of selenium dioxide. This suspension was then heated to 90° in an oil bath and stirred for 16 hours. After the mixture had cooled to room temperature, the residual selenium was filtered and the supernatant liquid added to 700 ml of water. This was then extracted four times with 250 ml of methylene chloride, which was washed with 250 ml of 10% sodium carbonate solution and dried with MgSO The methylene chloride was removed by rotary evaporation, and the residual material was distilled under reduced pressure. A yellow liquid was obtained with a boiling point of 90-2° (3 mm) (22 q/41%) [lit. 54 91-2° (4 mm)]. Preparation of 1, 2-, and 3-chloro c ycloheptatriene . 16 In a 100 ml three-necked round-bottomed flask equipped with a drying tube addition funnel and a nitrogen inlet was stirred, in an ice bath, 2.0 g of tropone (19 mmol) in 12 ml of CH 2 C1 2 under nitrogen. To this was added dropwise, over 15 minutes, 2.5 g (20 mmol) of oxallyl chloride in 4 ml

PAGE 65

56 of CH 2 C1 2 . The resulting slurry was stirred for an additional 35 minutes. With minimal exposure to air, the solvent was removed by rotary evaporation. In the same flask, the yellow residue was dissolved in 45 ml of dry THF and the flask was equipped with a drying tube, a nitrogen inlet, and a solid addition tube containing 1.49 g of lithium aluminum hydride. The system was purged with nitrogen and cooled in an ice bath. The lithium aluminum hydride was added in small portions, with stirring over 15 minutes. The resulting suspension was stirred for an additional hour and allowed to come to room temperature. The excess hydride was destroyed with water and the precipitated material filtered. The mother liquor was extracted with 250 ml of CH 9 C1 and dried with MgSO . The solvent was removed by rotary evaporation and the resulting liquid purified by column chromatography on silica gel using pentane as the elutant. The resulting yellow oil (1.5 g, 60% yield) was a mixture of 1, 2-, and 3-chlorocycloheptatriene, that had the nmr spectrum reported in the literature. 17 Preparation of zinc copper couple. To 2.0 g of cupric acetate dissolved in 50 ml of glacial acetic acid in a 125 ml Erlenmeyer flask heated to 100° and stirred magnetically, was added in one portion, 35 g of zinc dust. The slurry was shaken well and an additional 50 ml portion of acetic acid was added. When the material had cooled, the acetic acid was decanted from the resulting red-grey solid. This solid was washed three times with 100 ml of ether and stored under ether.

PAGE 66

57 The reactivity of the above couple was tested by reacting it with methylene iodide in the presence of cyclohexene In a 50 ml round-bottomed flask equipped with a reflux condenser and magnetic stirrer, was placed ca . 7 g of the wet couple and 20 ml of ether. To this was added ten drops of CH 2 I 2 which resulted in evolution of bubbles from the solvent. A mixture of 5 g of cyclohexene and 13 g of CH I was then added dropwise over 15 minutes. When the initial reaction subsided, the material was refluxed overnight. The slurry was then filtered and the solid washed with 25 ml of IN HC1, which was added to the ether. The layers were separated and the ether was washed three times with water and dried with MgSO^ . The nmr spectrum of the residual material after removal of the ether by distillation showed the presence of norcarane. Attempted reaction of 1-, 2-, and 3-chlorocyclohepta triene . In a 50 ml round-bottomed flask equipped with a reflux condenser and a magnetic stirrer was placed 320 mg of the above couple in 20 ml of ether. To this was added 0.5 g (4 mmol) of the chloride mixture in 1.1 g (4.1 mmol) of CH 2 I 2 , and the resulting slurry refluxed overnight. The precipitated material was filtered off and washed with 10 ml of IN HC1 which was used to wash the supernatant ether solution. The ether was washed with an additional 10 ml portion of IN IIC1 solution and then three times with 10 ml of water. The ether was then dried with MqSO . . Evaporation

PAGE 67

5 8 of the ether yielded a yellow oil shown by nmr and tic to be only the starting mixture of chlorocycloheptatrienes . Repetition of the above reaction with extended reflux times yielded the same results. Preparation of 7 , 7-dibromobicyclo [4 . 1 . J -2-heptene. 18 In a 500 ml three-necked round-bottomed flask equipped with a mechanical stirrer addition funnel and nitrogen purge, cooled in an ice bath, were stirred 200 ml of pentane, 24 g (0.29 mol) 1, 3-cyclohexadiene and 33.7 g (0.30 mol) potassium t-butoxide. To this slurry was added with efficient stirring, 75.8 g (0.30 mol) of bromoform in 90 ml of pentane, dropwise over 2 hours. The slurry was allowed to stir an additional 4 hours at 0°, and allowed to warm to room temperature overnight. To this was then added 400 ml of water and the layers were separated. The aqueous phase was then extracted twice with 200 ml of pentane. The combined pentane extracts were then washed two times with 200 ml of water and dried over MgSO . Removal of solvent by rotary evaporation leaves an oil that is purified by reduced pressure distillation. Thirty-seven grams of liquid [bp 80-2° (5 mm), 68-70° (.8 mm)] (49% yield) was collected. The identity of this compound was confirmed by comparison of its 1 8 nmr with that recorded in the literature. This material was stable over a period of weeks if kept refrigerated. 1 8 Prepara tion o f 2 and 3-bromocyclo hepta triene . In a 250 ml three-necked round-bottomed flask equipped for distillation through a short Vigreux tube into a receiver

PAGE 68

59 cooled in an ice bath, 72 g of quinoline was heated to 135-45°. The system was then evacuated via a water aspirator, and 24.3 g (96.4 raraol) of the above dibromide was added dropwise. The system was adjusted by an nitrogen bleed so that a rapid distillation of the resultant bromide occurred with only limited co-distillation of the quinoline. The distillate was then chromatographed on silica gel, eluting with pentane, to yield 9.32 g (57%) of yellow oil, ikentified as a mixture of 2and 3-bromocycloheptatriene by nmr. The isomer composition of the 2and 3isomers determined by integration of the methylene signals of the nmr spectrum is 3 to 1, respectively [bp 74° (15 mm), lit. 18 76-8° (16 mm) ] . Preparation of 2and 3-cycloheptatrienyl formaldehyde . In a 500 ml round-bottomed flask equipped with a low temperature thermometer, an argon inlet and septum, 6.8 g (0.40 mol) of the mixture of 2and 3-bromocycloheptatriene was dissolved in 150 ml of dry THF . Under an argon blanket, this solution was cooled to -70° by immersion in a dry ice bath while stirring magnetically. To this was added dropwise at a rate that allowed the temperature to remain below -65°, 31 ml of n-butyllithium (1.6 M in hexane) (0.05 mol). Stirring was continued for an additional 20 minutes at -70°. Then a solution of 7.3 g (0.10 mol) of dry dimethyl formamide in 50 ml of THF was added dropwise at a rate that allowed the temperature to remain below -65°. Again the solution was stirred for 90 minutes at -70°; then it was allowed to warm

PAGE 69

6 to -20°. One hundred milliliters of saturated NH CI solution was added and the mixture was filtered, separated, and the aqueous phase was extracted three times with 100 ml of ether. The combined organic phases were washed five times with ice in water and dried with NaS0 4 . The solvent was removed by rotary evaporation and the residue was purified by chromatography on silica gel, utilizing pentane to elute a fastmoving impurity, ultimately eluting the aldehyde with 25% ether in pentane. A yellow oil (2.6 g, 55%) was obtained that was identified by nmr as the aldehydes named above, in the same ratio as the starting bromides. H NMR (CDC1 3 ) 2.5 (t), 2.4 (t), 5.2-7.4 (m) , two singlets 569 and 574 Hz downfield from TMS . Preparation of cycloheptatrienyl formaldehyde tosylhy drazones . In a round-bottomed flask under nitrogen, 2 g of the above aldehyde was stirred with 3.1 g of tosylhydrazide in 30 ml of absolute ethanol for 1.5 hours. On silica gel using CH 2 C1 2 as the eluting solvent, TLC showed almost complete conversion to tosylhydrazone . Removal of the ethanol by rotary evaporation yielded a glass that was recrystallized twice from ether, yielding 3.6 g (72%) of crystals that tended to retain ether. The ether was removed by chromatography on silica gel with methylene chloride. H NMR (CDC1 3 ) 2.3 (m) , 2.35 (s), 5.2-7.5 (m) , 7.5 (A 2 X 2 ), two singlets 513 and 520 Hz downfield from TMS.

PAGE 70

61 Preparation of the sodium salts of the cycloheptatrieny l formaldehyde tosylhydrazones . Under a nitrogen atmosphere in a dry box, 576 mg (12 mmol) of NaH (50% in oil) was added to 3.5 g of the above tosylhydrazone mixture in 40 ml of dry THF. After stirring for 15 minutes, 40 ml of hexane was added to precipitate the salt. This was filtered and washed twice with 10 ml of hexane. After drying under nitrogen, 2.53 g of material that decomposed between 165-75° was obtained. Pyrolysis of the sodium salts of the cycloheptatrienyl formaldehyde tosylhydrazones . In a 25 ml round-bottomed flask equipped with a solid addition funnel and a reflux condenser, 15 ml of diglyme was refluxed under nitrogen. To this was added in portions 310 mg of the sodium salt. The resulting slurry was refluxed for 30 minutes and then poured into 100 ml of water. This was extracted three times with 75 ml of pentane and then three times with 75 ml of ether. The ether was washed with water (10 x 100 ml) and dried over MgS0 4 . Evaporation of the ether yielded a brown oil (40 mg , 30%) that was shown to contain the pyrazoles 6 and 7. H NMR (CDC1 3 ) 7.5 (s), 7.6 (s), 6.5-6.8 (m) , 3.5 (d, J = 6 Hz), 2.7 (t, J = 6 Hz); mass spectrum m/e (rel intensity) 132(21.1), 131(28.0), 105(15.3), 104(9.0); high resolution mass spectra m/e 132.06785, theoretical: 132.06870. Photolysis of the sodium salts of the cyc l oheptatrie nyl f ormaldeh y de tosylhydr azones. To 125 ml of dry THF in a Pyrex immersion well, irradiated (550 W, Ilanovia medium

PAGE 71

62 pressure lamp) under nitrogen, was added with efficient stirring, 310 mg of the sodium salt over 20 minutes. Photolysis was continued for 2.5 hours yielding a yellow mixture. The solvent was removed by rotary evaporation, and the residue was washed three times with 10 ml of methylene chloride. The remaining solid weighed 115 mg . (A quantitative yield of sodium toluene sulfinate would have been 178 mg . ) The solvent was evaporated from the methylene chloride fraction, and the resulting material was chromatographed on silica gel with chloroform/methanol 3:2. A yellow band was eluded yielding an oil (87 mg, 66%) that was shown by mass spectroscopy to be isomeric with the pyrazoles isolated from the pyrolysis of the salt. The proton spectrum shows them to be different compounds. H NMR (CDC1 3 ) 7.5 (s), 7.3 (s), 5.5-6.5 (m) , 2.8 (d, J = 5 Hz), 2.8 (t, J = 4 Hz); mass spectrum m/e (rel intensity) 132(99.7), 131(100), 105(21.7), 104(23.1); IR (major peaks) 3200, 3000, 2900, 1700, 1600, 1180. Preparation of 8-chlorobicyclo [ 5 . 1 . 0] 2 , 4-octad iene (10). In a 1 liter three-necked round-bottomed flask equipped with an addition funnel and a dry ice condenser 175 g (1.90 mol ) of freshly distilled cycloheptatriene and 55 g of methylene chloride (0.65 mol) were stirred in 200 ml of ether under nitrogen. To this was added dropwise 200 ml (1.77 M, 0.35 mol) of methyllithium in ether over 1.5 hours. The ether refluxed. The material was allowed to cool to room temperature while stirring was continued for an additional

PAGE 72

6 3 two hours. The resulting slurry was poured into 250 ml of water to dissolve the precipitated material. The layers were separated and the aqueous layer was extracted three times with 200 ml of ether. The organic phases were combined and dried over MgS0 4 . The ether was removed by rotary evaporation, and the residual cycloheptatriene was removed by bulb to bulb distillation at 10 mm. The remaining oil was distilled through a Vigreux column at 5 mm. The fraction boiling at 52-60° was collected yielding 11 g of yellow oil (22%). This oil was distilled through a spinning band column with no apparent change in the nmr [bp 57-8° (3 mm)]. H NMR (CDC1 3 ) 5.9 (m) , 5.7 (m) , 3.6 (t, J = 3 Hz), 3.0 (t, J = 8 Hz), 2.5 (m) , 1.6 (m) ; mass spectra m/e (rel intensity) 140(10.1), 105(100), 91(29.2); high resolution mass spectrum m/e 140.0392, theoretical: 140.0392; IR (major peaks) 3050, 3000, 1620, 1440, 1380, 1280, 1240, 1050, 840, 760. Anal . Calcd for C H CI: C, 68.34; H, 6.45. Found: C, 6 8.23; H , 6.46. R eaction of 8-chlorobicyclo [ 5 . 1 . ] 2 , 4-octadiene (10) with maleic anhydride . In a 200 ml roundbottomed flask, 1.0 00 g (7.1 mmol) of 10 and 0.7 g (7 mmol) of maleic anhydride were refluxed in 100 ml of carbon tetrachloride for 2 days. The solvent was evaporated and the residue was dissolved in a minimum amount (30 ml) of hot carbon tetrachloride. Upon cooling, crystals formed. Those were separated and the mother liguor was concentrated to half its

PAGE 73

64 original volume by heating on a steam cone. A second crop of crystals was taken. Combination of these two crops yielded 513 mg of white crystals that decomposed at 164° (sealed tube) . The mother liquor was chromatographed on silica gel, eluting with pentane. The fast running fraction corresponding to 10 was collected. Evaporation of the solvent resulted in the recovery of 246 mg of 10. Correcting for the recovered starting material, the yield of adduct was 45%. H NMR (CDC1 3 ) 6.5 (m, 2H) , 3.8 (m, 1H), 3.4 (d, 2H), 3.0 (m, 2H), 2.8 (m, 2H), 2.3 (m, 2H) ; mass spectrum m/e (rel intensity) 238(1.4), 166(3.5), 150(2.7), 131(3.4), 78(100); IR (major peaks) 3500, 3000, 1850, 1775, 1220, 1080, 960, 920, 900, 785. Preparation of 8 , 8-Dibromobicyclo [ 5 . 1 . 0] 2 , 4-octadiene 24 LLZJ In a 2 liter Morton flask equipped with a mechanical stirrer and cooled in an ice bath, 96 g (1.0 mol) of cycloheptatriene and 151 g (0.60 mol) of bromoform were stirred in 200 ml of pentane, under nitrogen. Over a period of 1.5 hours, 56 g (0.50 mol) of potassium t-butoxide was added in small portions. The resulting slurry was allowed to come to room temperature and was stirred overnight. This material was added to 300 ml of water and the layers were separated. The aqueous layer was extracted two times with 250 ml of methylene chloride. This was added to the pentane fraction and the combined organics were washed with 200 ml of water and dried over MgSC> 4 . The solvent was removed by rotary

PAGE 74

65 evaporation and the remaining cycloheptatriene and bromoform were removed by bulb to bulb distillation at 5 mm. The pressure was lowered to 0.05 mm and the remainder of the material was distilled from a steam bath into a vessel cooled in dry ice. This material was quickly distilled again at 0.05 mm, collecting a fraction boiling above 50°. This was chromatographed on silica gel, eluting with pentane, to yield 33 g (24%) of yellow liquid whose nmr matches that reported in the literature for 17 24 [lit. bp 42-6° (0.02 mm)] This material darkened at storage temperatures below 0°, but its properties were not affected. H NMR (CDC1 3 ) 5.1-6.3 (m, 4H), 2.3-2.6 (m, 2H), 2.1-2.4 (m, 2H) . Reduction of 8 , 8-dibromobicyclo [ 5 . 1 . ] 2 , 4-octadiene (12) with tri-n-butyltin hydride . The tri-n-butyltm hydride was prepared by the method of Kuivila and used as soon as possible after preparation. In a 25 ml pear-shaped flask with a septum top, 3.0 g (11 mmol) of the dibromide was stirred magnetically under nitrogen. To this was added dropwise, via syringe, 3.3 g (11 mmol) of the fresh hydride. This was stirred for 24 hours and then distilled under reduced pressure through a Vigreux column, collecting a fraction boiling between 55 and 60° (1 mm) . This material was contaminated with tri-n-butyltin bromide, and so was chromatographed on silica gel, eluting with pentane. A fast running band was taken that yielded 50 mg of an oil,

PAGE 75

66 still contaminated with the tin bromide, with an nmr similar to that of 10. X H NMR (CDCI3) 5 3 " 6 2 («i)» 3.5 (t, J = 4 Hz), 3.0 (t, J = 8 Hz), 2.0-2.6 (m) , 1.4-1.8 (m) ; mass spectrum m/e (rel intensity) 186(4.2), 184(10.8), 182(7.8), 105(100), 91(51.2); IR (major peaks) 2970, 2870, 1800, 1590, 1480, 1440, 1420, 1360, 1300, 1240, 1200, 1060, 1030, 870, 340, 800. Reaction of 8chlorobicyclo [ 5 . 1 . ] 2 , 4-octadiene (10) wit h potassiu m t-butoxide, general proced ure. In a 100 ml threenecked flask, 5.0 g (45 mmol) of potassium t-butoxide was stirred under nitrogen in 50 ml of dry THF. if a trap was used, it was added to this solution. This solution was then allowed to come to the desired temperature, and a solution of 700 mg of the chloride, 10, in 20 ml of THF was added dropwise. Stirring was continued for one hour and the resulting material was poured into 100 ml of water. This was extracted three times with 75 ml of pentane, which was washed three times with 100 ml of water to remove any remaining THF. The pentane solution was dried over MgS0 4 . The solvent was evaporated and the material chroma tographed on silica gel, eluting with pentane. A fast running band was taken, which was shown to contain cyclooctatetraene and styrene, by nmr and gas chromatography (11. 5' 10% SE-52, 90°). Yields [corrected to reactive isomer, (12) ] T ° COT Styrene 25° 5 6'o trace 65° 36% 4% '"' , 8t 8% (solvent: tetraglyme)

PAGE 76

67 Yields were determined by gas chromatography using p-methylstyrene as an internal standard. Styrene was not generated from the reaction of cyclooctatetraene with potassium t-butoxide under the above conditions. Preparation of 8-chlorobicyclo [ 5 . 1 . ] 2 , 4-octadiene (10) enriched in 1JL . In a 50 ml round-bottomed flask, under nitrogen, 1.4 g (10 mmol) of the mixture of chlorides was stirred in 20 ml of dry THF . To this was added dropwise, at room temperature, 270 mg (2.5 mmol) of potassium t-butoxide in 20 ml of THF. The solution was stirred for 45 minutes and then poured into 100 ml of water. The phases were separated, and the aqueous layer was extracted three times with 50 ml of pentane. The combined extracts were washed three times with water (300 ml) and dried over MgSO. . Evaporation of the solvent and chromatography on silica gel, eluting with pentane, yielded 818 mg of yellow oil. The oil was pumped overnight at 1 mm to remove cyclooctatetraene, yielding 410 mg of almost pure 11 (92% isomer purity by nmr integration of the triplets at 3.0 and 3.6) containing no cyclooctatetraene . Reaction o f 8 chlorobicyclo [ 5 . 1 . ] 2 , 4-octadi ene ( 1_0 ) with potassium t-butoxide at 90°/l mm . In a 100 ml roundbottomed flask equipped with a septum and connected to a vacuum system through a Vigreux column and a trap cooled in liquid nitrogen, was stirred 5.0 g (45 mmol) of potassium t-butoxide in 40 ml of tetraglyme. The cooled trap had previously been charged with 0.1 ml of dimethyl acetylene

PAGE 77

dicarboxylate in 1.0 ml of THF. With stirring, the flask was evacuated (0.5-1 mm) and heated in an oil bath (90-5°). When the system had come to temperature, a solution of 1.4 g (10 mmol) of 10 in 20 ml of tetraglyme was injected dropwise through the septum via syringe. With the first addition of chloride a red material began to collect in the cold trap. After the addition was complete the black solution was allowed to react for an additional hour, although no additional reaction was apparent. Nitrogen was bled into the system, and the red material was washed into the trap with a solution of 0.2 ml of the diester in 10 ml of THF. This solution was placed in the refrigerator for 24 hours, during which the color changed from red to orange. After 48 hours, the solution was almost yellow. After 72 hours in the refrigerator the solution was reduced to one-fourth its original volume on the steam cone. Benzene (25 ml) was added, and this solution was reduced in volume in the same manner. Benzene was then added to make the solution to 25 ml, and 0.5 g of Pd on charcoal was added. Air was bubbled through this mixture overnight, after which the catalyst was removed by filtration through a small column of alumina, washing well with ether. The resultant material was chromatographed on deactivated alumina, eluting with ether to bring down a yellow band. Elution with methanol yielded a purple material which was purified further by TLC (alumina/ether) , to yield a blue material.

PAGE 78

69 UV/VIS (CgH 12 ) 575, 365, 347, 305, 290, 235; lit. 33 580, 370, 350, 300, 290, 240; mass spectrum m/e (rel intensity) 244(50.3), 213(100), 183(32.6) . Reaction of 8 , 8-dibr omobicyclo [5. 1.0] 2, 4-octadiene (17) with potassium t-butoxide at 0° . In a 100 ml round-bottomed flask, equipped with an addition funnel and immersed in an ice bath, was stirred a solution of 5.0 g (45 mmol) of potassium t-butoxide in 25 ml of THF. To this was added dropwise a solution of 1.32 g (5.0 mmol) of 17 in 20 ml of THF. This was stirred at 0° for 30 minutes and allowed to come to room temperature. This was poured into 300 ml of water, and the resulting solution was extracted three times with 100 ml of pentane. Subsequent extraction of the aqueous phase with ether yielded nothing. The combined pentane extractions were combined and washed three times 100 ml of water and dried over MgSO.. The residue was 4 chromatographed on silica gel plates (1.5 mm, activated at 140° for 30 minutes). Two bands (R f .3 and .5 were isolated after elution with pentane for three cycles. These two fractions (R f = .5, 45 mg; R f = .3, 35 rng) had nmr and mass spectral characteristics of t-butyl styryl ethers. For R f = .5 (trans isomer) H NMR (CDC1 ) 7.2 (s), 7.1 (d, J = 12 Hz), 6.0 (d, J = 12 Hz), 1.4 (s); mass spectrum m/e (rel intensity) 176(3.0), 120(100), 91(37.4), 77 (3.4) .

PAGE 79

70 For R f = .3 (cis isomer) H NMR (CDCIO 7.0-7.8 (m) , 6.5 (d, J = 7 Hz), 5.3 (d, J = 7 Hz), 1.4 (s); mass spectrum m/e (rel intensity) 176(3.8), 120(100), 91(40.3), 77 (3.5) . Preparation of 3-phenylphthalide 5 6 In a 2 liter roundbottomed flask equipped for reflux, 200 g of zinc dust, 100 g (0.44 mole) o-benzoylbenzoic acid, 200 ml of water, and 800 ml of glacial acetic acid were refluxed vigorously for 2 hours. The resultant yellow solution was decanted from the residue, which was washed with 200 ml of hot glacial acetic acid. The solutions were combined and allowed to stand overnight. Beautiful white needles were separated by filtration. Addition of 500 ml of water to the mother liquor generated a second crop, and addition of another 500 ml of water, followed by cooling in ice, allowed a third crop to be collected. These combined crystals were added carefully in portions to 600 ml of 5% NaHCCu solution with stirring. Stirring was continued for an additional hour to allow the flocculent material to digest. This was filtered and recrystallized from 800 ml of hot ethanol. Hot filtration was necessary to remove an insoluble impurity. After two crops, 51 g of material was isolated (mp 114-5°, lit. 115-6°) (55%). Additional material of questionable purity could be isolated (10 g) but was not retained. 37 Preparation of 1 , 3-diphenylisobenzo f ura n . In a 25 ml addition funnel capped at the bottom with a 50 ml roundbottomed flask, and equipped with a reflux condenser,

PAGE 80

71 phenylmagnesium bromide was generated under nitrogen from 40 g of bromobenzene and 6 g of magnesium turnings. The funnel was transferred to a 1 liter three-necked round-bottomed flask and the Grignard was dripped into 42 g of 3-phenylphthalide in 250 ml of dry THF over a period of 45 minutes. After an additional hour of stirring, 250 ml of benzene was added, and the resulting solution was washed two times with 250 ml of IN HC1. The organic phase was then washed with 100 ml of 10% K„CO,. The aqueous phases were then combined and extracted with 100 ml of benzene. The combined organics were filtered through a frit containing MgSO. . The solvent was evaporated and the residue was recrystallized from 250 ml of hot benzene. A second crop was precipitated by addition of 200 ml of ethanol. The combined solids were again recrystallized from benzene to 37 yield 38 g (70%, lit . 87%) of yellow needles melting at 128-30° (lit. 37 130-1°) . R eaction of 8 , 8-dibromobicyclo [ 5 . 1 . ] 2 , 4-octadiene (1 _7) with potassium tbutoxide and 1 , 3-diphenylisobenzof ura n, ge neral procedure . In a 100 ml round-bottomed flask equipped with an addition funnel and nitrogen inlet, was stirred 5.0 g (45 mmol) of potassium t-butoxide and 1.35 g of 34 (5.0 mmol) in 40 ml of dry THF. Under a blanket of nitrogen, the solution was brought to the desired temperature. A solution of 1.32 g (5.0 mmol) of 17 in 20 ml of THF was added dropwise, and after addition was complete, the solution was allowed to stir for an hour. Ether (100 ml)

PAGE 81

72 was added and the phases separated. The aqueous phase was extracted three time with 100 ml of ether. The organic phases were combined and washed four times with 100 ml of water and dried over MgSO. . The solvent was removed by rotary evaporation and the resulting solid pumped overnight at 1 mm to remove any traces of THF. This material was chromatographed on silica gel, eluting with carbon tetrachloride, taking the adduct as a freely moving band. Reaction of 8-chlorobicyclo [ 5 . 1 . ] 2 , 4-octadiene (10) with potassium t-butoxide and 1 , 3-diphenylisobenzof uran at -78° . A solution of 2.8 g of 10 (20 mmol, 5 mmol of 12) in 20 ml of THF was allowed to react with 5.0 g (45 mmol) of potassium t-butoxide and 1.35 g (5.0 mmol) of 34_ in the manner described for the general reaction of 17. After the addition of .10 was complete and the resulting solution was stirred for one hour, a solution of maleic anhydride in THF was added until tic (silica gel/pentane) showed that the 34_ was consumed. Work up was then resumed in the general manner. Chromatography on silica gel, eluting with CC1., yielded a crystalline material as the second material eluted from the column. (The first material eluted war. unreacted 11 and some cyclooctatetraene . ) Recrys talliza tion from benzene/methanol yielded 197 mg (11%) of white crystals, melting at 168-9°. Two more recrystalli zations yielded an analytically pure material, mp 169-169.5°. i! NMR (CDC1 3 ) 7.0-8.0 (m, 1411), 5.8-6.2 (m, 411), 2.1-3.2 (m, 311), 1.7 (d, J 4 Hz, 111); mass spectrum m/e

PAGE 82

73 (rel intensity) 374(5.7), 270(100), 91(6.9), 77(16.8); high resolution mass spectrum m/e 374.16865, theoretical 374.16706; ir (major peaks) 3050, 2970, 2900, 1590, 1480, 1430, 1330, 1280, 1040, 1020, 970, 910, 900. Anal, calcd for C^H O: C, 89.81; H, 5.92. Found: C, 89.71; H, 6.01. Reaction of 8 , 8-dibromobicyclo [5 . 1 . ] 2 , 4-octadiene (1 7) with potassium t-butoxide and 1 , 3-diphenylisobenzof uran at ~ 78 ° • Silica gel chromatography (CC1 4 ) of the material formed at -78° by the above procedure allowed a moderately slow moving material to be separated from a small amount of rapidly eluted 34_. After evaporation of the solvent, 1.157 g of material was isolated that was shown to be 19. Three recrystallizations from THF/pentane gave an analytically pure sample (mp 156-7°) whose nmr did not differ from the material separated by chromatography (51% yield) . 1 H NMR (CDC1 3 ) 7.2-8.0 (m, 14H), 5.8-6.5 (m, 3H) , 5.5 (d, J = 10 Hz, 1H), 3.0-3.4 (m, 1H), 2.5-2.9 (m, 2H) ; mass spectra m/e (abs intensity) 20 eV 454(32), 452(47), 373(50457); (rel intensity) 70 eV 373(100); ir (major peaks) 3080, 3050, 2960, 2940, 1510, 1460, 1350, 1310, 1150, 1120, 1000. Anal . Calcd for C^H^BrO: C, 74.18; H, 4.67. Found: C, 73.99; H, 4.72. P reparation of 8 , 8-dichlorobicyclo [ 5 . 1 . ] 2 , 4-oct adiene . In the same manner as the preparation of 1_7, 120 g (1 mole) of chloroform and 100 g (1.1 mole) of cyclohepta triene were

PAGE 83

74 allowed to react with 52 g of potassium t-butoxide (.5 mol) Work up in the same way yielded 12.4 g (14%) of oil boiling at at 45° (0.05 mm) [lit. 57 36-9° (0.05 mm)]. H NMR (CDC1 3 ) 5.5-6.4 (m, 4H) , 2.4-2.6 (m, 2H) , 2.0-2.4 (m, 2H) ; mass spectrum m/e (rel intensity) 178(1.5), 176(5.7), 174(7.3), 139(62.3), 103(100), 91(21.4). Reaction of 8 , 8-dichlorobicyclo [5 . 1 . ] 2 , 4-octadiene with potassium t-butoxide and 1 , 3-diphenylisobenzof uran at -78_°_. Silica gel chromatography (CC1.) of the material formed at -78° by the general procedure allowed a material very similar to the ^4 adduct of the cyclopropene formed from 17 to be isolated. This solid (0.792 g, 39% yield) was recrystallized from THF/pentane to yield a material with a melting point of 148.5-150°. l H NMR (CDC1 3 ) 7.1-7.9 (m, 14H), 5.8-6.5 (m, 3H), 5.5 (d, J = 11 Hz, 1H) , 3.0-3.5 (m, 1H), 2.5-2.9 (m, 2H); mass spectra m/e (abs intensity) 20 eV 408(131), 410(46), 373(39475); (rel intensity) 70 eV 373(100). Anal . Calcd for C^H CIO: C, 82.24; H, 5.18. Found: C, 80.73; H, 5.36. Reaction of 8 , 8 -dibromobicyclo [5 . 1 . 0] 2, 4-octadiene (17) with potassium t-butoxide and 1 , 3-dipheny l isobenzof uran a t 0° • The reaction was repeated via the general procedure at 0°. Chromatography yielded a material ( 22 ) that was eluted with CC1 4 in the usual manner. Another material (2_3) was eluted with ether. This material appeared to have been spread throughout the length of the column during the

PAGE 84

75 elution with CC1 4 . This compound, 23, was not apparent in the nmr of the crude material before chromatography. This crude material (1.99 g) appeared to be almost totally composed of 22. Determination of the composition of this crude material by nmr, using 1 , 3 , 5-trinitrobenzene as the internal standard, showed it to contain 2.57 mmol (1.16 g, 51%) of 22 Two recrystallizations of the material eluted with CCl in benzene/methanol yielded 335 mg of a labile white material that darkened at 95° and decomposed at 110°. Recrystallization of 2_3 from chlorof orm/methanol yielded crystals melting at 200-1° (340 mg) . For 22, H NMR (CDC1 3 ) 7.7-8.0 (m) , 7.3-7.5 (m) , 7.2 (s), (14H) ; 5.2-6.5 (rn, 611); 4.5 (s, 111). Mass spectrum m/e (rel intensity) 454(.5), 452(.4), 267(100). High resolution mass spectrum m/e 454.07642, theoretical, 454.07540. IR (major peaks) 2950, 1550, 1460, 1420, 1410, 1310, 1270, 980. For 23 l H NMR (CDC1 3 ) 7.0-7.7 (m, 19H), 5.5 (s, 1H) ; mass spectrum m/e (rel intensity) 372(1.7), 267(100); high resolution mass spectrum m/e 372.15138, theoretical 372.15130; ir (major peaks) 3100, 1670, 1605, 1595, 1500, 1450, 1000. Anal, calcd for C^II 0: C, 90.29; H, 5.41. Found: C, 90.4 3; H, 5.46. Preparatio n of 1 , 2 , 4-t riphenyl -l , 4-butan dione . 4 ° I n a 500 ml round-bottomed flask, 18 g of acetophenone and 31 g of benzoin wore refluxed with 4 g of KCN in 75 g of water

PAGE 85

76 and 75 g of ethanol. After boiling for a half hour, an oil began to form. After 1.25 hours, the oil was separated and the mother liquor was allowed to stand. More oil was collected and combined with that previously obtained. After this material had stood for 1.5 hours and had begun to solidify, it was transferred to porous tiles, where it immediately crystallized. This material was boiled in 300 ml of ethanol and filtered hot. The mother liquor was allowed to stand, and crystals formed. The residue was added to 300 ml of ethanol and refluxed for a half hour. This was filtered while hot and allowed to stand until crystals formed. The two sets of crystals were combined to yield 22 g (40%) of yellow crystals that were used without further purification. Preparation of 2 , 3, 5triphenyl f uran . 4 ° In a 25 ml Erlenmeyer flask, the above crystals were stirred in 150 ml of coned H 2 S0 4 , for two hours. This was poured onto 1000 ml of ice. The sticky material was separated and dissolved in 200 ml of boiling ethanol. When cooled, this precipitated white needles (16.2 g, 78%), that melted at 9 0-1.5° (lit. 40 92°). Reaction of triphenylf uran with benzyne . In a 250 ml Erlenmeyer flask, 6.8 g (.05 mol) of anthranilic acid was dissolved in 75 ml of ethanol and 5 ml of coned HC1 and cooled in an ice bath to 0-5°. With stirring, 12.5 ml (.093 mol) of isoamyl nitrite was added dropwise at a rate keeping the temperature below 10°. After ten minutes of

PAGE 86

77 stirring, a precipitate formed. Ether (75 ml) was added and the solid material filtered and washed with ether. Five grams (.27 mol) of this material, along with 8 g of 2, 3, 5-triphenylfuran and 3 g of propylene oxide (acid scavenger) was refluxed in 50 ml of dichloroethane for 3 hours. After cooling the solvent was removed by rotary evaporation. The residual material was chroma tographed on silica gel, eluting with pentane. The residual triphenylfuran was removed as a fast running band. Elution with ether allowed collection of a solid material after evaporation of the solvent. This material was recrystallized from THF/pentane to yield 3 g of crystals melting at 170-5° Additional recrystallizations from CC1. yields a colorless material melting at 172-4°. H NMR (CDC1 3 ) 7.0-8.0 (m) ; mass spectrum m/e (rel intensity) 372(36.7), 270(100), 261(60.0); ir (major peaks) 3050, 2950, 1590, 1480, 1430, 1330, 1290, 1040, 995, 980. This material was inert in acetonitrile solution, and does not show any evidence of being converted to 23 after overnight reflux in acidic methanol overnight. Reaction of 8 , 8-dibromobicyclo [ 5 . 1 . ] 2 , 4-octadiene (17) with potassium t-butoxide with subsequent addition o f 1,3diphenylisobenzofuran . As in the general procedure, 1.32 g of 1/7 was allowed to react with 5.0 g of base, without 34, at -78°. After an hour of stirring at this temperature, 1.35 g of 34 in 20 ml. of TIIF was added dropwise over 10 minutes, and the resulting solution was allowed to warm

PAGE 87

7 8 to room temperature. After the usual work up with ether and water, a large amount of polymer was noted. The nmr spectrum of the crude material showed some 2!2 and no bromocyclooctatetraene . There was also no evidence of the styrene ethers . 4 8 Preparation of coumalic acid. In a 2 liter roundbottomed flask at room temperature 200 g (1.5 mol) of DL-malic acid was stirred magnetically in 170 ml of coned H 2 S0 4 . To this was added carefully, in one portion, 50 ml of fuming H 2 S0 4 , and the mixture was allowed to stir for 45 minutes. This procedure was repeated twice (total, 150 ml fuming H 2 SC> 4 ). After the last addition of acid the material was allowed to stir for one hour and was transferred to the steam cone, where it was heated for 2 hours. After a slight cooling period, the warm solution was poured carefully onto 800 g of ice. This was allowed to stand overnight. The solid was filtered and washed three times with 300 ml of ice cold water. The solid was dried on a sand bath at 65° , and divided into two portions. One of these portions was recrystallized from 170 g of methanol, decolorizing with Norite. The solid was separated and the mother liquor was used to recrystallize the other portion. This resulted in 45 g of sulfur yellow crystals (22%) melting at 206-9° (lit. 48 206-9°). 4 9 Preparation of 2-pyrone. _ Coumalic acid (30 g , .21 mol) was sublimed at 0.5 mm through a Vycor tube packed loosely with copper turnings and heated to 650°. The

PAGE 88

79 resulting yellow oil was collected in an ice-cooled trap. Distillation of this oil at 26 mm yielded 12.3 g of material boiling between 106° and 109° [lit. 49 , 110° (26 mm)]. This material was stored under nitrogen below 0° until used. Preparation of the sodium salt of tropone tosylhydra zone. The sodium salt of tropone tosylhydrazone was prepared by the method of Jones and Ennis, 51 by the reaction of tropone tosylhydrazone with sodium hydride in THF . The resulting precipitate was recrystalli zed from 1% DMSO in THF to yield purple needles. Preparation of 6-chloro-2-pyrone. 5 3 To a 100 ml roundbottomed flask containing 10 g (0.077 mol) of glutaconic acid and cooled in an ice bath, was added in portions 30 g (.14 mol) of phosphorous pentachloride . During the course of the addition, gas was evolved and the solids melted. When spontaneous reaction ceased, the solution was allowed to warm to room temperature. Again, reaction commenced and when this moderated, the material was heated for 15 minutes on a steam cone. The resulting red liquid was distilled yielding 7.1 g of material [bp 64-5° (.5 mm), lit. 53 67° (9 mm) ] identified by comparison with the nmr spectrum c o noted in the literature. Pre paration of 6-d2-pyrone . 53 In a 50 ml round-bottomed flask containing 5.8 g (44 mmol) of 6-chloro-2-pyrone in 30 ml of THF, under nitrogen, was added 11.6 g (77 mmol) of vacuum-dried Nal. This was allowed to stand, with occasional swirling, for 4 hours. The resulting material

PAGE 89

80 was poured into 20 ml of deuterioacetic acid and 30 g of zinc dust (vacuum dried at 130°). This was allowed to stand under nitrogen overnight, and then was poured into 100 ml of CH 2 C1 2 , filtered, and the resulting solid washed three times with 100 ml of water and two times with 10% Na CO, solution and dried over Na 2 SC> 4 . After chromatography on silica gel, eluting with CH CI and distillation from a bath at 115° (25 mm) , 607 mg of material was isolated that 5 8 was identified by comparison of the nmr and mass spec5 3 trum " with the literature. Pyrolysis of the sodium salt of tropone tosylh y d r a zone in the p resence of 2-pyrone . Under nitrogen, 1.4 g (14 mmol) of 2-pyrone in 15 ml of dry diglyme was heated to 110°. In portions, 0.5 g (1.4 mmol) of the sodium salt was added through a solid addition funnel. Stirring was continued for 30 minutes at this temperature. The resulting material was poured into 150 ml of water and was then extracted three times with 100 ml of pentane. The pentane extracts were washed six times with 50 ml of water and dried over MgSO. . Chromatography (Fisher basic alumina, activity grade III/pentane) yielded 58.9 mg (29%) of material identified as 3 , 4-benzocycloheptatriene by comparison with the nmr reported in the literature. Pyrolysis of the sodium s alt of tropone tosylhydrazone in the presenc e of 6-d-2-py rone . In the same manner as the pyrolysis reported above, ]00 mg (1.04 mmol) of 6-d-2pyrone was allowed to react with the pyrolysis product of

PAGE 90

81 the sodium salt of tropone tosylhydrazone (500 my, 1.4 rnmol) After the same type of work up, 33 mg of the 3,4-benzocycloheptatriene was isolated. The nmr spectrum showed that the deuterium was retained in the phenyl portion of the molecule . H NMR (CC1 4 ) 2.50 (t, 1.8H, J = 6 Hz), 5.82 (d of t, 2.1H, J = 6, 10 Hz), 6.59 (d, 2. OH, J = 10 Hz), 7.25 (s, 3.3H). Mass spectrum m/e (rel intensity) 144 (9.6), 143(90.0), 142(100), 116(26.9); for 3 , 4-benzocycloheptatriene 143(8.6), 142(79.4), 141(100), 115(29.8). IR (major peaks) 3010, 2950, 2820, 2220, 1480, 1450, 1430, 895, 820, 800. P hotolysis of the sodium salt of tropone tosylhydrazone in the p resence of 2-py rone. In a Pyrex photochemical immersion apparatus was stirred under nitrogen, 1.4 g (14 rnmol) of 2-pyrone in 100 ml of dry THF . While irradiating with a 550 W Hanovia medium pressure lamp, 0.50 g of the sodium salt (1.4 rnmol) was added in portions over 30 minutes. The mixture was worked up in the same manner as the pyrolyses. The average yield for the photolyses was 25 mg. The ratio of the 3 , 4-benzocycloheptatriene to 1,2benzocycloheptatriene was 2:1.

PAGE 91

REFERENCES 1. W. M. Jones and C. L. Ennis, J. Amer . Chem. S oc. , 91, 6391 (1969) . — 2. L. W. Christensen, E. E. Waali and W. M. Jones, J. Amer Chem. Soc. , 94, 2118 (1972). 3. For a review of this type of rearrangement see Some Pericyclic Reactions of Carbenes, W. M. Jones and U. H. Brinker in "Pericyclic Reactions," Vol 1, Academic Press, New York, 1977; and W. M. Jones, Ace. Chem Res. , 10, 353 (1977). 4. R. C. Joines, A. B. Turner and W. M. Jones, J. Amer. Chem. Soc. , 91 , 7754 (1969). 5. J. A. Meyers, R. C. Joines and W. M. Jones, J. Amer. Chem. Soc , 9^2, 4740 (1970). ; 6. K. E. Krajca, T. Mitsuhashi and W. M. Jones, J. Amer. Chem. Soc , 94 , 3661 (19 72). 7. T. T. Coburn and W. M. Jones, J. Amer. Chem. Soc., 96, 5218 (1974). ~~ — 8. C. Mayor and W. M. Jones, Tetrahedro n Lett., 3855 (1977) . 9. K. Untch, International Symposium on the Chemistry of Nonbenzenoid Aromatic Compounds, Sendai, Japan, August, 1970; personal communication. 10. B. Duell and W. M. Jones, private communication. 11. S. Winstein and J. Sonnenburg, J. Amer. Chem. Soc, 83, 3244 (1961) . 12. For leading references see P. R. Storey and B. C. Clark, Jr., "Carbonium Ions," Vol 3, Wiley Interscience , New York, 1972, p 1084. 13. M. Oda, Y. I to and Y. Kitahara, Tetrah edron Lett., 2587 (1975). 14. S. W. Benson, "Thermochemical Kinetics," Wiley, New York, 1968. 8 2

PAGE 92

8 3 15. K. Untch, personal communication. 16. B. Folisch, Chem. Ber. , 101, 2717 (1968); ibid., 104, 2324 (1971). 17. E. LeGoff, J. Org. Ch em., 29, 2048 (1964). 18. P. G. Lindsay and C. B. Reese, Tetrahedron, 21, 1673 (1965). — 19. T. J. Batterham, "NMR Spectra of Simple Heterocycles , " Wiley, New York, 1973. 20. R. K. Bartlett and T. S. Stevens, J. Chem. Soc . C, 1964 (1967). 21. U. H. Brinker and W. M. Jones, Tetr ahedron Lett., 577 (1976) . 22. J. Elguero, C. Marzin, A. R. Katritzky and P. Linda, "The Tauteromerism of Heterocycles," Academic Press, New York, 1976. 23. W. E. Billups and L. E. Reed, Tetrahedron Lett. , 22 39 (1977). ~~ 24. P. J. Van Vooren, R. J. Fletterick, J. Meinwald and R. E. Hughes, J . Amer . Chem. S oc. , 93, 4394 (1971). 25. G. L. Closs and L. E. Closs, J. A mer. Ch em. Soc, 82, 5723 (1960); G. L. Closs, R. A. Moss and J. J. Coyle, ibid/ 84, 4 98 5 (1962) . 26 27 For a review of reactions of this type see W. E. Billups, Ace. Chem. Res. , 11 , 245 (1978). J. Prestien and H. Gunther, Angew. Chem. Int. Ed., 13, 276 (1974) . 28. R. H. Parker, unpublished results. 29. T. C. Shields and P. D. Gardner, J. Amer. Chem. So c, 89, 54 2 5 (1967) . 30. L. Friedman and H. Shechter, J. Amer . Che m. S oc. , 83, 3159 (1961); W. Kirmse and G. Munschen, Annalen , 726, 4 2 (1969) . ~ 31. W. J. Baron, M. R. DeCamp, M. E. Hendrick, M. Jones, Jr., R. H. Levin and M. B. Sohn in "Carbenes," Vol 1, M. Jones and R. A. Moss ed . , Wiley, Now York, 197 3, P 20.

PAGE 93

84 32. W. N. Washburn, R. Zahler and I. Chen, J. Ainer. Chem . Soc. , 100, 5863 (1978). 33 W. von E. Doering and E. Wiley, Tetrahedron , 11, 183 (1960). 34. H. E. Zimmerman and L. R. Sousa , J. Amer. Ch em. Soc, 94, 334 (1972) . ' 35. P. Dinger, Synt hesis, 190 (1974); W. Eymann and M. Hanack, Tetrahedron Lett., 4 213 (1972); Z. Yoshida, H. Miyahara, Chem. Lett. , 335 (1972) . 36. C. L. Osborn, T. C. Shields, B. A. Shoulders, J. E. Krause, H. V. Cortez and P. D. Gardner, J. Am er. Chem. S oc. , 87, 3158 (19 65). 37 38, M. S. Newman, J. Org. Chem. , 26 , 2630 (1961). D. H. Williams and I. Fleming, "Spectroscopic Methods in Organic Chemistry," second edition, McGraw-Hill, London, 1973. 39 V. V. Razin and E. M. Rud' (1976) . Org. Khim. , 12 (3) , 689 40. 41. A. Smith, J. Chem. Soc. , 57, 643 (1890). G. W. Gruber and M. Pomerantz, Tetrahedron Lett., 3755 (1970). 42. 43 44 H. Durr, H. Kober, I. Halberstadt, V. Neu, T. T. Coburn, T. Mitsuhashi and W. M. Jones, J. Amer Chem. Soc, 95 , 3818 ( 19 73 ) . H. Durr, M. Kausch and II. Kober, Angew. Chem. Int. Ed Engl., 13, 670 (1974). T. Mitsuhashi and W. M. Jones, Chem. Commun., 103 (19 7 4) . 45. 46. 47. K. H. Pauly and H. Durr, Tetrahedron Lett., 3649 (1976). Recent evidence indicates that the assignment of these structures as bisnorcaradienes may be incorrect. H. Durr, private communication. R. H. Parker and W. M. Jones, J. Org. Chem., 43, 2548 (1978). — 48. R. II. Wiley and N. R. Smith, "Organic Syntheses, Collect. Vo] IV, Wiley, Now York, 1963, p 201.

PAGE 94

85 49. H. E. Zimmerman, G. L. Grunewald and R. M. Paulfer, "Organic Syntheses," Collect. Vol V, Wiley, New York, 1973, p 982. 50. K. Untch, International Symposium on the Chemistry of Nonbenzenoid Aromatic Compounds, Sendai, Japan, August 19 70. 51. W. M. Jones and C. L. Ennis, J . Am er. Chem. Soc, 91, 6391 (1969). 52. 0. L. Chapman, D. de la Cruz, R. Roth, and J. Pacansky, J. Amer. Ch em. Soc, 95, 1338 (1973) . 53. W. H. Pirkle and M. Dines, i±>id. , 90, 2318 (1968). 54. P. Radlick, J. Org. Ch em., 29, 960 (1964). 55. H. G. Kuivila, Synthesis , 2, 499 (1970). 56. C. R. Hauser, M. T. Tetenbaum and D. S. Hoffenberg, J . Org. Chem ., 23, 861 (1958). 57. W. M. Wagner, H. Kloosterziel , S. van der Ven, Rec. Trav. Chem. , 80 , 740 (1961). 58. W. H. Pirkle and M. Dines, J. Heterocycl. Chem. , 6, 1 (1969) .

PAGE 95

BIOGRAPHICAL SKETCH Richard H. Parker was born on the sixth of February, 1950, to Homer and Eris (Tucker) Parker. He was educated in the LaGrange (Ga.) public school system and graduated from LaGrange High School in 1968. He entered Auburn University during the same year and received the Bachelor of Science degree in 1972. He remained at Auburn University and received the Master of Science degree in 1974, under the direction of Dr. P. B. Shevlin. He entered the University of Florida in September of that year. He is a member of the American Chemical Society. Hd

PAGE 96

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. William M. Jones, Chairman Professor of Chemistry I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Ax<-u u. ( q William R. Dolbier Professor of Chemistry I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. John F. Helling Professor of Chemistry

PAGE 97

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation of the degree of Doctor of Philosophy. -{,-*• v. v A r. c John! R. Eyler ^LA Associate Professor-'of Chemistry I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Ox UMj tr Bu 1/lSi/OCharles A. Burnap Associate Professor of Mathematics This dissertation was submitted to the Graduate Faculty of the Department of Chemistry in the College of Liberal Arts and Sciences and to the Graduate Council, and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy. June 1979 Dean, Graduate School

PAGE 98

Iliijii