Title: Reactivity in the tricyclo 4 heptan-5-ylidene and the tricyclo 5 octan-2-ylidene series
Full Citation
Permanent Link: http://ufdc.ufl.edu/UF00098311/00001
 Material Information
Title: Reactivity in the tricyclo 4 heptan-5-ylidene and the tricyclo 5 octan-2-ylidene series
Alternate Title: Tricyclo 4 heptan-5-ylidene and the tricyclo 5 octan-2-ylidene series, Reactivity in the
Tricyclo 5 octan-2-ylidene series, Reactivity in the tricyclo 4 heptan-5-ylidene and the
Physical Description: x, 176 leaves : ill. ; 28cm.
Language: English
Creator: Garza, Oscar Trinidad, 1947-
Copyright Date: 1975
Subject: Corbenes   ( lcsh )
Chemistry thesis Ph. D
Dissertations, Academic -- Chemistry -- UF
Genre: bibliography   ( marcgt )
non-fiction   ( marcgt )
Statement of Responsibility: by Oscar Trinidad Garza.
Thesis: Thesis--University of Florida.
Bibliography: Bibliography: leaves 171-175.
General Note: Typescript.
General Note: Vita.
 Record Information
Bibliographic ID: UF00098311
Volume ID: VID00001
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: alephbibnum - 000162477
oclc - 02707498
notis - AAS8825


This item has the following downloads:

reactivityintric00garzrich ( PDF )

Full Text






To Mawy Jane



author wishes to express his deepest grati-

WtR. Dolbier, for the patience, encour-

d dance which were afforded the author

years of doctoral research. Without Dr. Dolbier's

and suggestions, this work would not have

i she d

auth *shes to express his appreciation to

s of pervisory committee, in particular

research accomplished with the

gro'up. Further gratitude is ex-

r his invaluable varied services

arch goals on many occasions.

ially owed to Pat and

ng~h tim d, did


ar uthor'



ACKNOWLEDGEMENTS ................................... iii
LIST OF TABLES ................... ................. v
LIST OF FIGURES .................... ............... vi
ABSTRACT ...................... ................... ix

I. INTRODUCT ION ............. .................. 1
II. SYNTHETIC METHODS ........................... 13
SODIUM SALTS .............................. 39
IV. DISCUSSION AND CONCLUSIONS .................. 59
V. EXPERIMENTAL ........... ...................... 88

APPENDIX .................. ............. ............ 138
REFERENCES ......................................... 17
BIOGRAPHICAL SKETCH ................................ 176

L' 'v
/ 1


26-s + Eu(fod)3. 100

26-a + Eu(fod)3. 100

Proton Assignments for 26-s and 26-a in
the 100 MHz Spectra. 101

Reaction Product Yields from syn-Tricyclo
[,4] heptan-5-one Tosylhydrazone
Sodium Salt. 119

Reaction Product Yields from Flow Pyroly-
sis of syn-Tricyclo [,4] heptan-
5-one Tosylhydrazone Sodium Salt. 121

Reaction Product Yields from anti-Tricyclo
[,4] heptan-5-one Tosylhydrazone
odium Salt. 122

Reaction Product Yields from syn- and
nti-Tricyclb [,5] octan-2-one
sylhydrazone Sodium Salts. 123
latile Product Yields from Static Py-
lysis of syn- and anti-Tricyclo
.1.0.02,4] heptan-5-one Tosylhydrazone
dium Salts: 1400-2250. 137


I i A


Figure Page

1 'H 60 MHz nmr of syn-tricyclo [4.1.02,4]
heptan-5-one (26-s). 29
2 Nmr of Eu(fod)3 / 26-s: Molar ratio, 0.13. 30

3 Nmr of Eu(fod)3 / 26-s: Molar ratio, 0.25. 31

4 Nmr of Eu(fod)3 / 26-s: Molar ratio, 0.38. 32

5 Nmr of Eu(fod)3 / 26-s: Molar ratio, 0.50. 33

6 60 MHz nmr of anti-tricyclo [,4]
heptan-5-one (26-a).

7 Nmr of Eu(fod)3 / 26-a: Molar ratio, 0.10.

8 Nmr of Eu(fod)3 / 26-a: Molar ratio, 0.20.

9 Nmr of Eu(fod)3 / 26-a: Molar ratio, 0.30.

10 Nmr of Eu(fod)3 / 26.f : Molar ratio, 0.40.

11 Hilh-vacuum pyrolysis system.

A-i 'H 100 Mhz nmr of syn-tricyclo [4.1 .02,4
heptan-0-one (26-s).
A-2 100 MHz nmr of anti-tri yclo [4.1 2,
heptan-5-one (26-a).

A-3 Nmr of syn -Iicyclo [
one (49 -s)w

A-4 Nmr anti-
one -a).

A-5 Nmr of -tri
0 d one t

F Page
Nmr of anti-tricyclo [,5] octan-
2-one tos hlhydrazone (50-a). 146

Nmr of dimer 52 obtained from 33-a sodium
salt. 147
)mr of dimer 52 obtained from 33-s sodium
alt. 148

Nmr of 4-ethynyl-cyclopentene (55). 149

100 MHz nmr of trans-1-ethynyl-2-vinyl cy-
clopropane (54-t). 150

00 MHz nmr of cis-l-ethynyl-2-vinyl-
clopropane (54-c). 151

0 MHz nmr of cyclopropyl region of 54-c. 152

of cis-l-allyl-2-ethynyl-cyclopropane
). 153

cf octa-l,2,5,7-tetraene (66). 154

o n-tric clq [, 4] heptan-5-
6-s) 24155

yclo [,4] heptan-5-

yclo [,5] octan-2-

cyc [,5] octan-2-

.1.0.02,4] heptan-5-
s). 159

.0.02'4] heptan-5-
5] octan-2-one

,5 octan-2-
S .162

Figure Page *
A-25 Ir of dimer 52 from 33-s sodium salt. 163
A-26 Ir of dimer 52 from 33-a sodium salt. 164
A-27 Ir of dimer 52 from cis-l-ethynyl-2-vinyl
cyclopropane (54-c). 165
A-28 Ir of trans-l-ethynyl-2-vinyl-cyclopropane
(54-t). 166
A-29 Ir of cis-1-ethynyl-2-vinyl-cyclopropane
(54-c). 167
A-30 Ir of 4-ethynyl-cyclopentene (55). 168
A-31 Ir of cis-1-allyl-2-vinyl-cyclopropane (65). 169
A-32 Ir of octa-1,2,5,7-tetraene (66). 170

v( i

A vii i

Wf Dissertation Presented to the Graduate Council
versity of Florida in Partial Fulfillment of the
events for the Degree of Doctor of Philosophy


Oscar Trinidad Garza

August, 1975

William R. Dolbier, Jr.
artment: Chemistry

e thermal conversions of the p-toluenesulfonylhy-

ium salts of (a) the syn-and anti-tricyclo

] heptan-5-ones as well as those of (b) the syn-
icyclo [,5] octan-2-ones were investi-

try resulting from the generation and sub-
o W f the respective carbenes for each series
-for ty casr of the tricyclic hep-

wer-loeveloped for the required syn-

nti-tricyclo [,4] heptan-5-
affording nearly equimolar mix-

tone while the third method af-
ure i e ratio 2:98. The novel
gly labile in

Carbene product distributions for both isomers in

a given series proved similar thereby demonstrating that the

cyclopropane orientation differences within an isomeric set

afforded no readily identifiable differences in reactivity.

Although the observed products for each series studied

could be explained largely by the established ethylenic-

acetylenic fragmentation process common to cyclopropyl car-

benes, the case for the isomeric tricyclic heptanylidenes

was subjected to further scrutiny because of the plausible

operation of a novel six-electron pericyclic process in-

volving the highly-strained species 1,2,5-cycloheptatriene

as the primary product derived from either tricyclic hep-


Three C7H8 monomers and one C7H8 dimer were obtained

from the tricyclic heptanylidenes while two C8H10 monomer

were isolated from the tricyclic octanylidenes. The novel

C7H8 monomer cis-l-ethynyl-2-vinyl cyclopropane afforded

unusual reactivity, undergoing ostensibly a Cope rearra

ment with an approximate half-life, assuming a unimolecu

rate-determining step, of two to three hours max

250C. It remains the subject of ongoing studies

proved to be the most interesting carbene produ

from the studies of both tricyclic carbene sys



The question of a quasiaromatic structure as a

picturee of the electronic makeup of certain suitably-

id carbenes has received considerable attention in

;t decade.1 Although cyclopropenylidene (1) has

dy thus far, cyclopentadienylidene (2) and cyclo-

ylidene (3) have been studied in detail.1,2


s 4 aAd 5 to 2 and 3, re-

bstantial experimental verifica-
k by two research groups34 sup-
el calculations performed by

4 (p2) 5 ( 2)

The EHT calculations indicate that the nucleophilic or

electrophilic character ought to alternate in the series

1 3. The nucleophilic character of 1 and 3, with the
2 2
a2 singlet of each lower than the p singlet on the poten-

tial energy profile, was predicted from calculations of

the total charge residing on the respective methylene car-

bons of 1 and 3: -0.68 and -0.86. Experimental evidence
was in turn provided by W.M. Jones and coworkers in the

case of 3. A Hammett study was performed in which cycle
heptatrienylidene was generated in the presence of an
eight-to-teftkold excess of an equimolar mixture styr

and 3- or 4-substitutid styrenei he p value .05
0.05 obtain d from tit study pr vided the first anti
tative asse ment of e nucleoph city of a

aromati carbene.

lopentadie ne, on
readily lend itself o

it jae o e x

d a


+ :C*


imattc carbene. In the case of 6 the vigorous

resultingg in extrusion of carbon atoms may pre-

existence of a bishomoaromatic carbene (9).

immett studies and suitable trapping reactions

Roth electron-rich and electron-poor olefins,

Lt afford information concerning the

electrophilicity or nucleophilicity of 6 and 7, no evidence

for bishomoaromaticity in 6 and 7 would appear to be forth-

coming. Even the isolation of stable forms, dimers and/o 9

oligomers, of 10 and 11, derived from 6 and 7, respectively,

while perhaps suggesting a special intermediate as their

precursor, would not demonstrate bishomoaromaticity as an

operative force in these systems.

10 11

The discussion of bilhomoromati

of 7, fortunately, can be extended du
work performed y Bergman and Rajadhyak

the same pap Treatment of 3-br

* octa-2,6-di 0 ota

.^ 250 afford%

rate determining step of the reaction, and (b) rapid, re-
versible deprotonation-reprotonation of 12 occurred at a
rate greater than that of rearrangement. The scheme postu-
lated by Bergman envisioned a homoconjugated anion (13)

Br 0Br

12 13

S+ Br


y o a homoconjugated neutral intermediate (14)
equently underwent cleavage in the manner typi-
propylcarbenes. Since 7 and 13 both afforded
d t (@, it was logical that a common inter-
ould be invoke, the final result being
ep nation (15) which corresponds to
atio the singlet carbene. In summary

O *A6m

it can be stated that the isolation of 8 via generation of

7 from the tosylhydrazone sodium salt does not appear to be


merely a simple case ofacyclopropylcarbene undergoing th

well-documented ethylene-acetylene fragmentation process

but is, in fact, a case of a stabilized carbene givin

product (8). Alt ugh further argument concern ing t

mechanistic ii ff 8 may prove ethereal, it is

able that 8 cou arise from all&ne (11) b


S Consideration of the 1,3-bishomocyclopentadienide

ecies (16), of which 15 would be the a cation, allows

rather insight into the postulated existence of 15. HMO

culations predict appreciable bonding interactions be-

the allylic anionic and olefinic systems in 16.1

sonance integral B27 (or B46) is approximately equal

0B where 0 is assigned a value of 18 kcal/mole,
normally used for benzenoid systems.3 Charge



17 18

st at C-2 and C-4 (0.426), considerably
C-7 (0.064), and least at C-3 (0.021).

y enough, the tetracyclic hydrocarbon (18)
ne n' ree products upon equilibration

user's catalyst-solvent system,

er test the hypothesis of bishomo-
bene chemistry, it was


deemed necessary to investigate this question from the
viewpoint of bishomocyclopentadienylidene, a species
which could possess two isomeric forms; i.e., syn and
anti configurations of the fused cyclopropane rings should
be possible. The bishomocyclopentadienylidenes, syn- and
anti-tricyclo [,4] heptan-5-ylidenes, 19-s and
19-a devoid of the molecular constraints as encountered in

19-s 19-a

6 and 7, would be capable of generating the novel cyclic
allene 1,2,5-cycloheptatriene (20), whosf structural de-
formation would place it in the same class with the known
1,2-cycloheptadiene (21) and 1,2-cyclohexadiene (22) s
cies. 14ab Monomers 21 and 22 have been sh n to be

L I *


intermediates as evidenced by the isolation of

dimeric and tetrameric products. In the case of 22
OM l o been ,trapped employing both styrene

,3-dipheny rbenzo [c]-furan.5ab The observation of
ssible d @r and/or tetrameric products resulting from

Vould lend credence to the concept of a bishomoaromatic
e4 49thiah bch an observation, by itself, constitutes

ique multiple-bond fragmentation process for a new car-

ather than the actual generation of a particular

stabilized carbene.

their int sting features of the bishomocyclo-

igation would lie in the isolation

['403,5] heptane (24).
ts an interesting problem
e from 20 via a Cope process

ry product from the ethy-
ocess common to cyclopropyl-
echanistic alternatives

inherent stability and reac-
tracyclic heptane (24), syn-

and Bri.ntrup,16 converted

a reaction halflife of
of approximately #


suggesting that suitably-generated 19-s or 19-a might give
rise to 24 via C-H insertion.

23 (=54c)

Further speculation about the reactivity of 19-s
and 19-a only served to increase the demand for the actual
experimental work. The work described in this disser
was performed principally on the syn-and anti-tricyc
[,4] heptan-5-ylidenes in order to determi
basic reactivity of these systems and thereby add
discussion of homoaromaticity in alicyclic carbe
their elaboratTon of the working hypothesis was
by inspection of the syn-and anti-tricy
octanA2JylidenIs, 28 a and 9 h
test for operational homoaromhti r y
possessing favorable ge entries.



1% 1




SSnthe try into both the tricyclo [,4]
an-5- lidene and the tricyclo [,5] octan-2-

ne ser w s I'rovided by the synthesis of the corre-

nding syWnd ti ketones for each series. Conversion

the ketonel to the corresponding p-toluenesulfonylhydra-

)hydrazones) was desirous since tosylhydrazone
iunPor potassium salts are normally stable pre-
ch can be subsequently pyrolyzed or photolyzed
1,2or pyrolyzed in the solid state a la Schecter.17

synthetic methods were developed for the syn-
ti-tricyclo [,4] heptan-5-ones,
ly. The first two methods (methods
farly even mixture of the syn
third method (method C) afforded
ratio of 2:98.

I) began with the known conversion
responding acid chloride fol-

mine to afford the dimethyl-
obutanedione (27).18

ner,19 with allyl Grignard

afforded the C14 keto-alcohol, 5-allyl-5-hydroxy-2,4,4-
trimethyl-7-octen-3-one (28), which was subsequently con-
verted to diallyl ketone (29) by base-catalyzed cleavage
with barium hydroxide.19 Diallyl ketone (29) was converted
to its ethylene ketal (30) which was treated with iodoben-
zene dichloride. 20ab The crude mixture of cis-and trans-

7,8-bis (chloromethyl)-1,4-dioxaspiro [4.4] nonanes (31),
the product of addition of one mole of chlorine concomitant I
with radical cyclization to generate the five-membered carbo-
cyclic structure, was subjected to acid-catalyzed hydroly-
sis resulting in restoration of the ketone moiety. The
resulting mixture of cis-and trans-3,4-bis (chloromethyl)-
cyclopentanones (32) was treated with 50% sodium hydroxide
whereby an a, loss of two moles of HC1 was effected re-
sulting in the isolation of two ketonic products which nm
ir, uv, mass spectral, and elemental analyses indicated to
be C7H80 isomers.
The isolation of the two ketonic products, which
proved to be the desired tricyclic heptanones, 26-s an
26-a, proved initially troublesome when attempt'
separation work employing typical glpc methods. An
tion of the isomeric mixture of ketones onto various ar
wax 20 M columns (column temperatures: 130C-1600C
ically resulted in the isolation of t e some
(a) anti-tricyclo [4.1.0. 4] hetan-5-o

cyclohen and (c) -cy heptadie
&^^B& 211 ^^^^^^

SOC12 NEt3
----h--t --s-



1) CH = CH-CH MgBr

2) NH4C1




Cm (fP3


H20, EtOH

4 1 0


n I


Scheme I continued


N-NI -Tos






Tos -N -N =


N N Tos




icyclic ketone was identified by comparison of its

0 MHz) with that reported by Gajewski and Shih
.21 The outstanding feature of the nmr of 26-a is

Symmetrical quartet (J = 3.5 Hz) with a two-proton
ation at 6 0.85. The remainder of the nmr spectrum
d three multiplets of two-proton integration each
at 6 1.25, 1.56, and 2.08. The ir (1720 cm-1),

ax 287 nm, E 28), and mass spectrum (M+ 108) were
confirmative of the anti isomer; however, elemental
s id not prove satisfactory, a situation which
until pc method were rectified.
The 2,4 a 3,5-cyclcheptadienones, recognizable
e fo 'otPn multiple in the region 6 2.17-

ee r and (b)the four-proton doublet
entered at 6 3.00 for the 3,5 isomer,22 were

ca--promoted rearrangement which de-
ction of the syn-trifyclic ketone
rtion of the anti PIction (26-a).
Borg and Kloosterzil had shown that the
terconvertible in the temperature
1,5 hydrogen shift, resulting
dominated by the 2,4 isomer.22
d th rvation that the cyclo-

h~ g umn as a mixture
26-a elu d as a



0 0 0


Isolation of analytically pure syn-and anti-tric
heptanones was accomplished by the use of an alkaline co
umn (10% Carbowax 20 M) employing 3.5% potassium hydroxid
to effectively remove active sites from the inert support
typically Chromasorb P-Regular. Whereas liberal inje ns
of ammonia vapor had not prevented destructiogfo
tones in the initial glpc work, the potassium hydroxi
coated column allowed for almost quantitati
* cyslohept nofR as iltiernal standard) se ion
lion Of 26-s anti in tht temphr-atur rng 30
It" rt &td t 4 W1.itkt column tSmo t
range k850-19 0 led gain to the a t complete
f 4 (4% recovery) s-s esli t
some s y le.
T~he novel y.1 2
forde fout com
1 ti

The multiplets at 1.50 and 1.78 were overlapping, a featui'

whi readily distinguishes 26-s from 26-a. The ir (1700

uv (Xmax 283, c 70), and mass spectrum (M+ 108) also
firmed the structure of this C7H80 ketone isomer. Ele-

tal analysis of 26-s proved satisfactory.

Silica gel chromatography conveniently afforded

paration of 26-s and 26-a of sufficient purity to allow

r subsequent conversion to the corresponding tosylhydra-

nes, 33-s and 33-a, respectively. Whereas 33-s was ob-

erved to have been formed in 83% yield, 33-a, unfortunately,

as obtained in only 31% yield. Conversion of 33-s and

33-a to the corresponding sodium salts, 34-s and 34-a, was

anti ively achieved employing sodium hydride (1.2

nts) with tetrahydrofuran solvent. Potassium salts

and 33-a were made by treatment of the tosylhydra-

otassium tert-butoxide, again using tetrahydro-


od B (ScIeme II) employed a synthetic sequence

ed to bt'a mort convenient synthetic route than

Drawing on the analogy provided by workers such

d Gutsche24 in performing intramolecular

ediates of the copper-ketocarbene type

e bond, a sequence of steps was devised

ii ze cis-ethyl-2-vinyl-cyclopropanecarboxy-

he point of entry into the synthetic se-
no stereospecific synthesis of 35-c has

Scheme II






CO Et t
35 9






e reported to date, a cc. venient synthesis of cis-and

trans-ethyl-2-vinyl-cyclopropanecarboxylates (35) was

supplied by Vogel and coworkers.25 The cis-trans ratio of
5was determined to be 40:60. Saponification of cis-and

trans-35 afforded the corresponding mixture of cis-and

trans-cyclopropanecarboxylic acids (36). Treatment of 36

with thionyl chloride afforded a 38:62 mixture of the cis-

nd trans-cyclopropanecarbonyl chlorides (37). Up to this

int the synthetic work was essentially a duplication of

gel's work which had been concerned with the synthesis

d reactivity cis-and trans-2-vinyl-cyclopropylisocya-
atel Treatment of 37 with ethereal diazomethane af-

a c'ude mixture of cis- and trans-l-diazomethylketo-
-c loflopanes (38) as evidenced by the strong ir
-1 ad the diazomethyl singlet at 6 5.31

nmr. pper-catalyzed.decomposition of 38 in re-
0 *
exane afforded the desired isomeric tricyclic
nd 26-a, in the syn-anti ratio of 47:53;
ed upon the cia acid chloride (37).

ot providing a satisfactory syn-
id pr ide an interesting route to 26-a.
r the synthetic sequence, as outlined

clo [3.1.0] hexan-2-one (39)

ly synthesized from either 4-

Scheme III

I 0





3% H2SO4

(CH3 3SO I



syn :anti

Initially employing the basic procedure for the
synthesis of bicycle [3.1.0] hex-3-en-2-one (42) outline
by Russel and Stevenson,28 conversion of 39 to the cru
bromoketal (40) was Accomplished only after allowing
bromination to proceed at 250-380C in ethyle l9 gl
stead of at 00. Further modification of the basi
dure was found necessary in the s quent step
reverse addition (pinchwise additi odiu
P to a DMSO solution of 40) app be a
edition for obtaining a r

hhthe 6Lylene l1 (41

SH4O affording 42 in 58% yield. Subsequent treatment of

42 with trimethylsulfoxonium ylide afforded 26-a and 26-s

in 66% yield; however, the syn-anti ratio proved to be

2:98, a result which was not totally unexpected.

The synthesis of 42, a monohomocyclopentadienone,

catalyzed an attempt at the preparation of the correspond-

ng tosylhydrazone from which the carbene of 42 could ul-

tely be generated. The attempt met with failure because

sy1hydrazine destroyed the carbon-carbon double bond in

1 addition fashion. This result agreed with similar

s reported by Kirmse who observed that various cyclo-

nones and cyclohexenones underwent addition as well

ndensation in reactions with p-tosylhydrazine.29

The synthesis of syn and anti tricyclo [,5

ones, 49-s and 49-a, outlined in Scheme IV, repre-

ynthetic Sequence developed by Lambert, Koeng,

a.30 Although Lambert had removed a substantial

the ci-, cik-tricyclo ['5] octyl-2-

acid (46) from the mixture of cis, cis and

ns acids (46) by fractional crystallization, no

w made in this work to separate isomeric tricyclic

until arrival at the ketone stage. Birch

o of benz c id afforded 1,4-dihydrobenzoic acid

hi s with ethereal diazomethane. Methyl-

te product of methylation with

Scheme IV

1 2


CH2 12

Zn Cu
(3 runs)

10% NaOH









V continued















diazomethane, was cyclopropanated twice by three-fold treat- 1
ment with methylene iodide and a zinc-copper couple. The
product mixture, largely cis, cis and cis, trans-methyl-
tricyclo [,5] octyl-2-carboxylates (45), was con-
verted to the corresponding mixture of acids (46) with 10%
sodium hydroxide solution. Decarboxylation of 46 with lead
tetraacetate afforded the isomeric tricyclic octyl acetat
(47) which were subsequently reduced with lithium aluminum
hydride to a crude mixture of the alcohols (48). Oxidati
of 48 with the Jones reagent provided the tricyclic octa
49-s and 49-a, which were separated and purified by fou
consecutive short-path distillations. The syn-anti
of the ketones was determined to be 77:23 from t
Conversion of 49-s and 49-a to the corres
tosylhydrazones was accomplished although the

not good. %O-tricyclo [,5] ocfn-2-
zon -( ) wa iin 42% yield le t
hydraZone T i n 'Ty
quest conversion 50 ot
51-s and 51-a ished e
that *ffLh o


eptanones 26-s and 26-a. Employing Eu (fod)3,

(III), lanthanide-induced shifts produced some
resting spectral changes, the most interesting

demonstrated that the two protons adjacent to the

ion (a-methine cyclopropyl protons) in both
ere t located farthest downfield in the

Ir s ctra. It appeared that the protons most de-

in stems were the two protons located at

ine cyclopropyl protons. Further,

tions of the agreement factor R31a-c for the

afforded values of 0.16 and

2 'respectively. It was also

erdo protons at C-3 and C-7 in

upfield than the endo protons of

the opposing banana bonds
>Wrting a shielding effect
Oi4 the geometry of 26-s 4x-

of 26-s and 26-a brought to

hex-3-en-2-on (42) where

t ~ clopropy

he yclopropyl
s might


1 H
H exo


42 42-r

of the suitably disposed double bond. Since (42) has a

resonance form (42-r) which places a positive charge adja-

cent to the B-methine proton, it may be unfair to use (42)

as a model for analogy with regard to 26-s and 26-a. Never-

theless, the observed effects in the nmr of 26-s and 26-a
serve to create speculation about the possible contribution

00 0 0 -

6 N4 "
o o
1 2 6

26-aa 26 26-r

of a species which possesses finite partial charge separ

tion as in 26-r. The extreme case, of course, would be

contribution from the anti-homoaromatic species 26-aa.
The nmr spectra on the following pages demonst

thele eap-frog" effects ich occurred upon the addi

to eit&r 26-AI

- -ta






















L - ....--- -- .. .. ----

r l'q












-- --


8 1 ...: .- -.-- ....

l. I












rO - I~- --r- -- --* 1- - l


; ,;~ I.

S 0

I ,. I
i--I- *r .



II '. I

.. i


1 -I

I -




_____ __?___ _









--. i- -

o Ca


I, 0

: i



Simple speculation concerning the probable products

from both the tricyclo ['4] heptan-5-

dthe tricyclo ['5] octan-2-ylidenes led

on a moderately volatile products were

ibilities. With this in mind it Jas concluded

S4I0iues which avoided solution work and

on lems would be preferred. -The method

eared to,b J:igh-vacuum pyrolysis of the neat

Idium alts wQAe, for the most part, pyrolyzed

three elected procedures. The fore-

xtr, tion of the static technique
works ch as Schecter,17'33 was

op-stati t(_-S) tec niqut oin which con-
f the sodium s a aed g ss

v um (7 x 1 mm) was ac-
vir 11y a cuum unit

for ing, and

f opera-

ttech n i q u e ,




.C 0r
U V)


E Ta
3 C

E "-
3 u
U *r-
r, C
> f0


*r- I-

- 0

Q- 0


r- 0)
0 C.
o m
0 E

1--- >



comT ly referred to as the "hot tube" technique. This

technique, in comparison with the D-S technique, allowed

fo longer contact times thereby increasing the chances for

other rearrangement of initially-formed carbene products.

e f al technique was simply the infrequently-used static

hnique which did find one important application re-

in the isolation of an elusive compound (54-c).

y solution rk performed involved several trapping

ions involving the syn-tricyclo [,4] heptan-5-

-cies apstyrene or dimethyl maleate.

of pyrolysis work, involving the sodium

51-s and l-a, was performed on 34-s and

t was hoped that the different disposition

ring serve to alter the product

in g e syn to the anti carbene in

prove to be the case. In fact,

anti reactivity for each series

IO simple static (S) pyrolyses

ge' O0-500C afforded a white,

rply at 46o-470, which was conve-

cold finger (0-100) of the modi-

e and S pyrolyses. The waxy

odor whicP always served to

pyrolysa ,The solid

discolored upon exposure to air, assuming a yellow-brown

coloration within one hour. The thermal stability of the

solid appeared suspect since attempted sublimation under

high vacuum of impure material led to tarring. Heating
the solid in benzene or toluene at temperatures > 500 even

under argon or nitrogen also had a deleterious effect.

The structure of the solid was established by spec-
troscopic evidence and analysis. Infrared bands at 3020,

1650,and 650 cm-1 indicated the necessary double bond
feature. The nmr spectrum, perhaps the most confirmative

piece of spectroscopic evidence, showed a four-proton

multiple at 6 2.33, a two-proton multiple at 2.35-2.90, a
four-proton multiple at 3.02, and a six-proton multiple

at 5.55. The two-proton multiple was overlapped on each

end by the adjacent four-proton multiplets. The uv spec-

trum showed a Xmax at 254 (E 8120). The uv absorption maxi-

mum corresponded rather well with that reported for tricyclo

[,10] octadeca-5,9,11,15-tetraene (53).34 Mass

spectral analysis showed the parent peak to be m/e 184.

Elemental analysis proved satisfactory for a C14H16 ole-
finic hydorcarbon. The structure therefore assigned to

this unique product was tricyclo [,8] tetradeIa
4,7,9,12-tetraene (52).

Xmax = 254 max = 252
c = 8120 c = 8600


52 53

Hydrogenation of 52 resulted in the uptake of four

moles of hydrogen, the hydrogenated product clearly possess-

ing the parent peak of m/e 192. Attempts to form Diels-

Adler adducts with maleic anhydride, dimethyl acetylene-

dicarboxylate, and tetracyanoethylene failed, resulting

in the tarring of dimeric 52. This failure to form a

suitable Diels-Alder adduct was disappointing due to the

effort spent in trying to achieve the desired result; never-

thcleS, it may very well be distortions of the molecular

rework in the vicinity of the butadiene moiety are pro-

ting the 2 + 4 cycloaddition.

An interesting facet of the isolation and charac-

ization work of 52, formally the dimer of 1,2,5-cyclo-

atriene (20), is that the dimer is the major product

all pyrolytic conditions, irregardless of technique

mperature. In light of this fact, the monomeric

20 would deserve consideration as the primary prod-

ng from collapse of both tricyclic heptany-
nes via a unique six-electron pericyclic reaction

inf 6

44 V

Volatile products isolated from the D-S and S py-

rolyses of 34-s and 34-a included toluene (57), cyclohep-

tatriene (56), and two much more interesting products,

trans-ethynyl-2-vinyl-cyclopropane (54-t) and 4-ethynyl-

cyclopentene (55). Glpc analysis empqbying a column

perature of 68-70C provided analytical samplMs of 54-t

54-t 55

and 55. The cyclopropane

in the temperate range

appearing only
nmr spectrum
wi t hAi

as being destroyed. In fact, until the actual isolation

f 54-c, the assignment of trans sterochemistry to the

glpc-purified cyclopropane hinged upon this observation as

well as the known instability of both cis-1,2-divinyl-

cyclopropane (68)35 and cis-l,2-diethynyl-cyclopropane (69).36
The ir spectrum of 54-t showed the necessary mono-

substituted-acetylene bands at 3320 and 2120 cm1 with the

vinyl absorptions displayed at 1635, 980, and 905 cm

The 100 MHz nmr spectrum at high field showed a three-

proton mulitplet at 6 0.80-1.40 (H6, H7, and H8), an ace-

tylenic doublet at 1.85 (H5), and an allylic multiple at

1.58-1.94 (H4). The olefinic regions displayed an ABX pat-

tern with multiplets centered at 4.96 (H3), 5.12 (H2), and

(H1). Coupling constants afforded by first order analy-
are given below:

1 H J = 16.8 Hz
H 1,2

7 H 31,3 = 9.5
2 ,4 = 7.5

4 J2,3 = 2.5

J = 2.0


Mass spectral analysis' of 54-t demonstrated the

parent peak to be m/e 92, the parent peak also being the

base peak. Elemental analysis in turn proved satisfactory
for the C7H8 olefinic hydrocarbon.
The assignment of the 4-ethynyl-cyclopentene (55)

structure resulted from comparisons with the reported spec-

troscopic data of 55 which was isolated by Cristl and
Harrington as the major product from the pyrolysis of n

tricyclenone p-tosylhydrazone sodium salt (58).37 The ir
spectrum of 55 displayed bands at 3325 and 2125 cm" ,

Na ( 1600

N Tos

58 55*

dicative of a monosubstituted acetylene. The c
the carbon-carbon double bond was supplied by p
and 690 cm-1. The 60 MHz nmr spectrum sh d

doublet at 6 2.04, a five-proton multi e r

3.20, and an olefinic singlet of two ro
5.68. Mass spectral aiaysis

the base peak, to be m/e 92. Elemental analysis proved
satisfactory thereby confirming the structure of 55. It
should be noted that flow (F) pyrolytic work led to in-
creased yields (6.9-25.9%) of 55 in comparison with D-S
The fourth and final true carbene product, discount-
ing cycloheptatriene and toluene as bona fide carbene prod-

* ucts, to be isolated from pyrolyses of 34-s and 34-a was
cis-l-ethynyl-2-vinyl-cyclopropane (54-c) in 17-23% yield.
This elusive compound was successfully generated by a simple
modification of the basic static technique. Maintaining the
sublimator cold finger at -72 to -650C, twice the normal
Amount of given sodium salt usually employed was heated
from an initial temperature of 1400 to 2250 over a period
thirty minutes. For the first time, the diazo precursor,

or anti-5-diazo-tricyclo [,4] heptane (59) was
Served to have bE5n conveniently generated and trapped
(on thqocold finger). Coinciding with the observation of
e red diazo compound was the successful isolation of 54-c,
ecies of unusual reactivity. Whereas 54-t proved ther-
ly stable at 1590C in deuterochloroform solution, 54-c
rwent complete rearrangement within thirty-six hours in
rochloroform solution at 250 to afford 52. The impli-
ns of this rearrangement upon the nature of the primary
t obtained from collapse of the carbene to product
nable. At this point, although no hard kinetic
n able. At this point, although no hard ki-netic


evidence is yet available, it appears that 54-c has a half-

life in solution of perhaps two to three hours maximum,

assuming the unimolecular process, 54-c--+20, to be rate


The syn-and anti-5-diazo-tricyclo [,4]

heptanes (59) proved extremely unstable, appearing to Pery

slowly lose nitrogen even at the low temperatures employee

This was not surprising since secondary aliphatic diazo '

compounds had generally been demonstrated to not survive

the pyrolysis conditions wrien generated from the correspond-

ing p-tosylhydrazone salts. Moreover, the instability f of

59 was in accordance with the reported instability of di

cyclopropyldiazomethane (60) which must be kept below -3

in order to sufficiently retard decomposition.38



The ir spectrum of 54-c showed diagnostic absorp-

at 3310, 3085, 2120, 1635, and 985 cm-1 although in-

ence from cycloheptatriene and dimer 52 tended to ob-

absorptions elsewhere in the spectrum. The 100 MHz

ectrum at high field showed two one-proton multiplets

78 (H8) and 1.20 (H7). A two-proton multiple region

observed at 1.48-1.84 (H4 and H6) for the allylic and

1 protons followed by an acetylenic doublet (H5) at
olo finic regions displayed an ABX pattern with

centered at 5.10 (H3), 5.24 (H2), and 5.64 (H1).

stants are given below:

62= 15.8 Hz

J = 8.5 Hz

Ol 4 = 7.5 Hz

go J2,3 = 2.5 Hz

2 = 2.0 Hz
HH 5,6

r nmr analy 'of both
vided d ethy-

In an initial attempt to trap the key reactive in-
termediate, 1,2,5-cycloheptatriene (20), pyrolysis of 34-s
via the D-S technique was effected under high vacuum (9 x
10-5 mm: maximum pressure reading) employing a matrix iso-
lation unit, with the matrix window maintained at 60K as
the pyrolysate trap.* Inspection of the uv-visible range
200-500 nm showed the only readily discernible absorption
to be due to 52, Amax 254. Ir inspection showed principally
the presence of the cyclopropane 54 and dimer 52. A peak at
2040 cm- with a shoulder at 2035 cm-1 was of primary inter-
est although without considerably more work in this area it
would be extremely premature to assign the absorption to a
high-energy allenic stretching mode which would probably be
characteristic of 20.
Several attempts to trap the monomer 20 with sty-
rene and dimethyl maleate failed. The products obtained
from these reactions were the spiro adducts resulting from
trapping of the carbene itself. Heating 34-s in the pres
ence of either styrene or dimethyl maleate in tetraglyme
solution afforded the respective adducts 2-phenyl-s iro'
[cyclopropane-1,5'-syn-tricyclo [,4] heptane] (6
and trans-2,3-dicarbomethoxy-spiro [cyclopropane-1, sy
tricyclo ['4] heptane] (62). o h

re matrix 1
se of mat
es of Dr. o
i versit

SMe02C CO2Me

61 62

S(34%) and 62 (39%) were not spectacular, they were

eless sufficient so as to allow characterization of


The mass spectrum of 61 showed the parent peak to

196 with elemental analysis proving satisfactory for

enoid hydrocarbon. Peaks at 3070, 3040, 1605,

indicated the presence of a monosubstituted

ear atic structure while an absorption at 3010

the tyclopropane structural feature. The

informative since a set of overlap-

tIt at 6 2.15 and 2.30 provided evidence

1n isomeric mixture consisting of cis,
1. The 60 MHz nmr of isomeric 61


showed two four-proton multiple regions at 6 -0.05 0.730
and 0.75 1.93. The spiro cyclopropane hydrogens gave rise
to (a) one distinguishable doublet at 1.25 and (b) the set
of overlapping triplets previously mentioned. The high-
field triplet had a coupling constant of 7.5 Hz while the
low-field triplet possessed a coupling constant close to
7.0 Hz in magnitude. The uv spectrum displayed B-bands at
262, 267, and 276 nm with a much more intense K-band at 22
nm (E 7490). The K-band, while 22 nm higher than that of
benzene was reasonable since cyclopropylbenzene itself s
a K-band at 219 nm (E 8900).40
The assignment of trans stereochemistry to the
spiro dicarboxylate 62 was based on the size of the coup
constant obtained from the AB quartet generated by the
cyclopropane hydrogens in the 60 MHz nmr spectrum. T
Hz splitting observed corresponded well with the 5
splitting reported for dimethyl-3,3-dimethy c
tranl-1,2-dicarboxylate (6). 41 Further th*,Qcs
cyclopropane dicarboxylates 64-c and 64-t have re


02CH3 H CO Me MeO2

S63 g6 .


f 8.8 and 6.6 Hz, again confirming the trans stereo-
stry of spiro adduct 62.4 The 60 MHz nmr of 62 dis-
yed two four-proton multiplets at 6 0.52 and 1.70, the
roton AB quartet with'doublets located at 2.42 and 2.65,
two three-prpton singlets at 3.72 and 3.76. The mass
tum showed a very weak parent peak (< 1%) but the base
*1 +
t m/e 117, corresponding to C9H9 demonstrated the
loss of formally two carbomethoxy radicals and one
atom. *Th ,r spectrum showed particularly strong
1 -1
ons 't 1740 1440, 1340, 1290, 1230, and 1165 cm 1.
yrolyses ~ yn- andanti-tricyclo [,5] octan-
y1hydrazon odium salts, employing-the D-S tech-
ded lro products in variable total yields rang-
79.9% to a low of 26.5%. The range of
s from 2600 t6 4000C with product
ur vicinity of 4000. The primary
o b' cis-l-allyl-2-ethynyl-cyclo-
it5 hy rogen-shifted derivative,
rowing in yield with corre
lysis temperature. 0
65 owed a one-proton multiple
t-proton multiple at 1.02
otb l tiplet ai 1.38
laye t 1.81 (H 5
V 1 I

The vinyl protons gave rise to an ABX pattern with multiplets
centered at 5.02 (H3), 5.10 (H2), and 5.95 (H1). Simple

first-order analysis supplied the relevant coupling constants:

H6 J = 17.6 Hz

8 HC 7 H
1 J = 9.7 Hz

H3 l ,4 = 6.4 Hz



The ir spectrum of 65 showed the necessary acetylenic a
sorptions at 3320 and 2120 cm l with the vinyl contrib
displayed at 3080, 1645, 990 and 910 cm 1. The ma
trum showed the parent peak to be m/e 106 with
located at m/e 91, C7H7 Elemental analysis pr
factory for the C8H10 compound thereby establ"
identity of the hitherto unknown 65.
The acyclic octatetraene 6, predo
in the pyrolysate, was established

ived from a 1,5 hv rogen shift in c
&.sol oduct derived frQm

dent flow (F) pyrolysis of glpc-purified 65 at 410.
e octatetraene was obtained in 72% yield as the only

eadily-identifiable product with no trace of 65 in the py-

The 100 MHz nmr of 66 showed a pseudo heptuplet

f two-proton magnitude at 6 2.84 (H3). The olefinic re-

ion was more complicated giving rise to a two-proton pseudo

entuplet at 4.74 (H1), a three-proton multiple region at
4.84-5.32 (H2, H7, and H8), and another three-proton multi-

plet region at 5.50-6.60 (H4, H5, and H6). The proton as-

signments appeared justified on the basis of comparison with

HH1 H H 7
H 5


H3 H3

66 67

um of e reasonable model compound 1,2,5-

atr 6 .43 The ir of 66 possessed an absorption at

Srm f tha allene moiety. Other important

absorptions attributed to carbon-carbon double bond group-
ings were noted at 3090, 3005, 1645, 1605, 995, 910, 845
and 740 cm-1. The uv spectrum provided an absorption maxi-
mum at 225 nm, e 27,100. Since no compound possessing a
cis-butadiene structural feature has been demonstrated to
have an extinction coefficient greater than 10,000,4 it
appears that 66, if a cis-trans mixture, contains a sizable
trans component. Mass spectral analysis placed the pare
peak at m/e 106 while elemental analysis proved satisfact
for the C8H10 isomer.
A summary of the basic reactivity of the syn add
anti tricyclic heptanylidenes and octanylidenes shows th
four C7H8 monomeric and dimeric species and two C8H10
meric species were isolated. As yet unanswered are
stereochemistry of the ring fusion in 52 and (b)
metry of the butadiene moiety in 66. Probably th

5 5
52 54-c 54-t
b 6w9
rP l


ing products in terms of continuing interest are the

ic cyclopropanes 54-c and 54-t. Kinetic studies of

ompounds will extend the body of knowledge surrounding

Cope rearrangement. Although not mentioned earlier, 54-t,

h having proved stable at 1690C in solution, did rear-

q quantitatively to give 52 at 2080 in solution as evi-

ed by uv and mass spectral fingerprints. Moreover, it

appear that 54-c, even without further kinetic exami-

will een cis, ,2-divinyl-cyclopropane (68)

s-1,2- i hyn -cyclopropane (69) in proclivity towards

ype meant .35 6L6


ser of cycloheptatriene and

te ob ined from the tricyclic heptan-

sonable to say that the cyclohepta-

tonated 5-diazo tricyclic heptane

wed by cationic re-

This hypothesis

ium or


4 4

carbonium ion intermediates (or both) obtained from p-

tosylhydrazone salts when decomposition fails to exclude
a protic environment. The formation of toluene, which

could derive from cycloheptatriene at high temperatures as
outlined by Woods, is also due in part to the breakdown

of the p-toluenesulfinate anion since toluene was also o

served in the pyrolysate afforded by the tricyclic octan

ylidenes. The end result is that cycloheptatriene and t

uene are not carbene-derived and warrant no further di

s ion.

2 H







d The question of paramount importance with regard to
e isolation and characterization of the various products

ived from the syn and anti tricyclic heptanylidenes is

intriguing one: Do the observed products stem from the

collapse of the carbene to 1,2,5,-cycloheptatriene (20)

owed by rearrangement and/or dimerization, or does cis

hynyl-2-vinyl-cyclopropane (54-c) merit designation as

le primary carbene product undergoing subsequent re-

ment to afford the other observed products?

S? ?

0 54-c

T answer to such a mechanistic question must be

(a ervations thus far obtained (Chapter

(bpea'ing in the chemical literature.

Analogy or precedent was afforded by the remarkable insight
of Sasaki and coworkers who proffered a novel explanation
for the considerable difference of the carbene product dis-
tributions between the aldehyde and ketone p-tosylhydrazones
71 and 72.46 It was concluded that the chrysanthemyl carbenes

--- + H C=CH
2 2 2
71 (cis and trans) 68-73%

K __ -- 1r
72 (cis and trans) 70-92%
derived from cis and trans-71 underwent largely fragment
as opposed to ring expansion due to a favorable electronic
effect imparted by the isobutenyl moiety. The electronic
effect would result in the lowering of the transition-sta
energy for the fragmentation process by virtue of format
of a conjugated diene, assuming the carbene reacti r
ceeded via either a concerted process or via a s
process involving either an ion-pair or radical

S* 61


0 O,

H* _

radical pair ion pair

e of the chrysanthemyl carbenes derived

-72, the ring expansion process predomi-

sive migration of the isobutenyl-substituted

d), again suggesting that the electronic
isobutenyl substituent'plays an important role

ion state. However, another piece of rationale

der to explain the drastic difference in

uct s ibutions despite obvious electronic

th yl -substituent. It was Sasaki's con-
t i y of cyclopropylcarbene reactions,
0 I11ti dr ring expansion, had been explained

o tipn of strain and electronic fac-
eration of conformational effects.

hat the co-formational effect played

n Qptermining the cyclopropylcar-

y. Maximum interaction of the

nd (the ring-expansion process)
1 uent R ca mud-an s-trans-like

pane i while an

s-cis-like conformation afforded poor interaction, presum-
ably facilitating fragmentation as the process of choice.


R3 R3


S-trans s-cis

1 1

R3 RCR ------ Rc

R2 R2 R3

The observed ring-expansion of the chrysanthemyl
carbenes of 72 was thus rationalized on the basis of the
s-trans conformational effect while the chrysanthemyl car-
benes of 71, minus the methyl substituent, were free to C
assume either conformation, the s-cis being probably no
more or no less favored than the s-trans with the elect
effect exerted by the isobutenyl moiety being the dec si
,,a _

The development of the conformational effect as a
n determining the cyclopropylcarbene rearrangement

ude was further extended by Sasaki to geometrically
rained species, some of which are locked into either
-trans or an s-cis conformation. Whereas 73, an s-trans
:ies, underwent ring expansion exclusively, 74, 75, and
underwent the fragmentation reaction due to the s-cis
picture presumably.46


73 n = 2,3

r^ n

75 76

s precedent as provided by Sasaki would re-
inyl cyclopropane (54-c) as the pri-
y to be obtained upon collapse of
anyl i~ species to product. Cursory
e t s outlined in Chapter III would

tend to argue in behalf of 54-c as the sole primary product
since 54-c, once formed, pauved extremely labile (t1/2
2-3 hours maximum) at 25 rearranging smoothly to dimer 52.
Further, at present, there is no evidence to argue for re-
versibility in the formation of monomer 20 with the pre-
dominant formation of dimer 52 under all pyrolysis condi-
tions arguing for an essentially irreversible reaction lead-
ing from 54-c to 52. The simple kinetic picture suggested
by the observed reactivity of 54-c is supplied by the follow-
ing reaction sequence treated by the steady state approximaj

k k3 I
2C 2M --- D i

rate = d[D]/dt = k3 [M]2

d[M]/dt = k1 [C]2 k2 [M]2 k3 [M]2 C = cyclopropane 54-c

= 0 M = monomer 20

[M] = kl/2 [C]/(k + k3)1/2 D = dimer 52
1 2 3)

d[D]/dt = klk3[C]2/(k2 + k3)

Assuming k2 << k3, the rate expression w Id
to d[D]/dt = kl[C]2, the rate of formation of t
to 0 Am Ahk[C

The intermediate, once formed, goes on to product rapidly.

In this situation no equilibrium is established between re-

actant and intermediate, the intermediate (M = 20) being

dubbed a van't Hoff intermediate: Mechanism A.




Reaction Coordinate


L 1

spite the foregoing discussion of precedent and

g qualitative kinetic analysis, there remain some

ionstied into the exact identity of the

I &





primary product obtained from the tricyclic heptanylidenes.

Assuming 54-c to be the primary product, how does one ac-

count for the formation of trans-l-ethynyl-2-vinyl cyclo-

propane (54-t) and 4-ethynyl-cyclopentene (55), products,

which if they arise from 54-c, must do so by a nonconcerted

process presumably involving biradical intermediates; i.e.,

54-c would be forced to undergo epimerization via cleavage

of the C1-C2 bond to afford 54-c at lower temperatures

(1600-400C), while a 1,3 alkyl shift involving migration of B

the C1-C2 bond would afford 55 at higher temperatures

(> 5000C). In reality such reactivity for 54-c does not A

nstrates the facile nature of the six-electron peri-
lic reactions for such substituted cyclopropane systems.35,36

ever, in the case of 69 surface-catalyzed reactions in the
s-phase kinetics cast doubt on the exact nature of the re-
angement of 69. In contrast to the foregoing purportedly
certed pericyclic reactions, the cis-trans isomerizations
simple cyclopropanes typically afforded Ea's in the range

).4-65.5 kcal/mole while 1,3 alkyl shifts of simple vinyl-

propanes gave rise to Ea 's in the range 44.5-54.6 kcal/


(S Ea = 60.5 kcal/mole


-- Ea = 49.6 kcal/mole

o be made is at the heart of the discussion of
le anisms: if 54-c is the only true primary
54g. and musIt derive from a species other
e s ime, the low yields of 54-t (1,0-
.8-254 s est that the importance of
illm+C wvver deserving of mechanistic scrutiny,


should not be overly weighted so as to create havoc with
the mechanistic pathway proving, in the final analysis, true
for the tricyclic heptanylidene systems.
The suggestion to be made at this point is to also
consider an alternative mechanism which could account for
the observed products. In this alternative mechanism, 1,2,
5-cycloheptatriene (20) would arise via a novel six-electron
reaction involving the carbene site and the C1-C6 and C2-C4
cyclopropane bonds, a postulate only briefly mentioned ear-
lier in Chapter I. The application of the Mobius-HUckel
concept as applied to organic molecules and reactions by
Zimmerman49 to the cyclic arrays involved in bond-breakin
and bond-forming processes in the transition states of th
tricyclic heptanylidenes shows that both systems posse
HUckel orbital arrays with the syn isomer (19-s) having
zero sign inversions while th anti isomer (19-a) po
four inversions. This is schematically outlined
The result of this simple analysis is the tendency
the fact that either carbene could afford t
species 20 outright without invoking an interme
as 54-c. It should be noted tha e foreg

arrays employ tAe p2 configurati the
sumptio which tter alomm the dev
nco ed in n
it 0 .ene

/ Bond breaking

Bond forming

Bonds unchanged

S anti dark lobes: + of basis set
light lobes: of basis set

ular models. Another examination of the possible
ness" of this carbene arrangement would be afforded
olching the hypothetical arrangement from the view-
gta a tetro-cycloaddition involving a Tt4s + T2s proc-
r 1-s and a T4a + r2a process for 19-a wherein the
tive T groupings have been tagged with the HOMO-LUMO
nations derived from the Fukui Frontier Orbital ap-
to pericyclic reactions. The connecting lines
slof the same phase serve to demonstrate the
facial-antarafacial interactions which af-
ene from the respective carbenes. Again it
notd tha he p2 configuration of the carbene
q P6,

2 + 4 retro-cycloaddition:


T4S + I2s Homo

4rr4a + Tr2a

anti -

has been employed in describing the electronic interactions

leading to the allene as product. The exact configuration

of the carbene carbon in the tricyclic heptanylidenes as

well as the electronic makeup of the allene 20 will be dis-

cussed further videe infra), but first a hypothetical pic-

ture must be drawn based upon the primary formation and sub-

sequent reactivity of 20: Mechanism B.




K 54-c





Since it hasbeen experimentally established that

arranges smoothly to 52 presumably via 20, then 20,

irst formed as specu ted, would be in equilibrium with

ble 54-c he rate'determining step being that leading

at dimer 52. In this sense 20 would be an

ius 9 Bediate3 and the energy profile would quali-

ly appeagas follows. Further, an understanding of

rigin of p ct 54-t and 55 would be more easily

M = 20

C = 54-c

D = 52

oordi nate

My formed species 20 could be
I and triplet states, the triplet

en-chain biradical species 77

4-t or depending upon ,the

Si"ce both singlgj
1 1 2
d for c ns general, 12

it is reasonable to assume that a triplet tricyclic heptan-
ylidene species would give rise to a triplet allene species

which could then fragment to give biradical 77. A biradical
such as 77 is a potential intermediate as evidenced by the
assumed formation of such intermediates in the pyrolyses
of (a) the salt of bicycle [6.1.0] nona-2,4-6-triene carbox- 0
aldehyde tosylhydrazone (73)51 and (b) 7-cycloheptatrienyl
diazomethane (79).52

Na I
78 CH N-N-Tos -CH

79 CH = N2 C-


Contrast of the preceding discussion of Mechanism B

st Mechanism A points to the fact that triplet 20,

ined in Mechanism B by collapse of a small triplet com-

onent of the tricyclic heptanylidene to triplet 20, would

not be allowed under Mechanism A since, under the concerted

process assumed taking place in the conversion of 54-c-20-
2, singlet cyclopropane (54-c) would give rise to singlet

which would dimerize as fast as it is formed.
Two alternative mechanisms have thus far been pro-

sed. A clea choice between the mechanistic alternatives
be left to the individual reader at this point. A

r chanism, which.would invoke both 54-c and 20 as
ry products via competitive formation from the carbene,
tO beg the question of mechanism, and

s and electronic character of 20, the
f tricyclic heptanylidene chemistry, de-
I o eration. The smallest cyclic allents
rf 1,2-cyclononadiene53 (80) and 1,2,6-
81). Thg smallest cyclic allene yet ob-
exadiene (22) with 1,2-cycloheptadiene
terms of framework distortion. As
Chat the observation of 21 and
ion of their dimers and tetramers
trapping experiments.

h&IA I

0 4d
80 81

In the smaller ring systems such as 21 and 22 the various
*authors have presented the vierwoint that these systems
must be something less than a fully-bonded allenic structu
i.e., research groups such as Ball and Landor14a and Moo
and Moser 4b have proposed that the C6 and C7 cyclic all
possess a planar structure in which carbon atoms 1,2, an
are sp2 hybridized with unbonded electrons for the sing
species (S) residing in the hybrid orbital at C-2, w
in the triplet species.(T), one electron has been p
from the nonbonding hybrid at C-2 to the nonbonding
Mol cent Dil nd UndOrwo
MO calcula I imber is
allene as and mal 1
the lowes
tr u t i n a

basis set: dark orbitals: S T
in plane
light orbitals:

-2. On the other hand, 1,2-cyclohexadiene was shown by the

Iculations to most likely possess a triplet ground state

st approximated by an allyl radical with a second unpaired

ectron occupying the in-plane orbital at C-2.

Returning to the question of the initial carbene

configuration of the tricyclic heptanylidene, it can be

at the singlet p2 configuration assumed earlier (note

-Mobius and 2 + 4 retro-cycloaddition schemes) would

ise to a-specifs formally containing an allyl anion

sitiwl charge residing in the in-plane orbital at C-2.

angement of the singlet species disagrees with pre-

n.54 On the other hand the lower-energy 2 configura-

f the carbene, because of its disposition away from

clopropane bonds (Cl-C6 and C2-C4) would not

high energy singlet


allow for the necessary overlaps in the transition state 0
leading directly to monomer 20. In fact such an arrangement

( D

a2 low energy singlet

would most likely afford 54-c, the primary product of Mech-
anism A, since no other reaction appears possible. Thus
determination of the singlet carbene configuration (a2
p ) for the tricyclic heptanylidene might aid in the ch
of mechanistic alternatives. On the other hand, it
remembered that the demonstration of p2 character i
benes such as the syn and anti tricyclic heptanyll
not guarantee subsA4ljnt Inique reactivity as would b
served in asi of Mechanism B where the
tion, asru shocm l iati sup pose'dl o
six-elect ric(i- c' t n leading
tion arises f. gman>l s nmen f1 o
to 15, discuss in Ch ter which under
Slashed f K^ pro asd t
cycl ic pr
b VP


W The point to be made in this discussion about the
electronic structure of 20 is that 20, if the primary
Orbene product of tricyclic heptanylidenes, must arise by
what appears to formally be a bishomoaromatic carbene in a
singlet process and would possess at least initially singlet
character reverse that of prediction.
As a matter of final interest, it would be of value
to attempt to place the key C7H8 species on an energy sur-
face although it would be only a crude approximation in the
case of both 20 and the tricyclic heptanylidenes themselves.
__ 55
employing Benson's additivity rules55 for the estimation of


102 1 89-90

65 66

"strain free"

thermochemical properties, 54-c, 54-t, and 55 are easily
calculated allowing one less kcal/mole for the heat of
formation of 54-t due to the missing cis interaction as
would be observed in 54-c. The straightforward calcula-
tions afford values of 90, 89, and 66 kcal/mole for 54-c,
54-t, and 55, respectively. Employing Wentrup's extra-

polation56 of Benson rules to carbenes, a rough calcula-
tion of the energy of the tricyclic heptanylidene species
itself employing the heat of formation of triplet methylene
(92 + 1 kcal/mole) for the carbene "group" affords a value
of ca. 135 kcal/mole. Ring corrections used in obtaining
this value are derived from bicycle [3.1.0] hexane (32.7 kcal/

mole) and cyclopropane (27.6 kcal/mole). The value doe
not appear altogether unreasonable since 7-cyclohepta-
trienyl methylene (82) has been calculated to possess a
of formation of ca. 145 kcal/mole,56 a value which s v
to indicate that the heat of formation for certain s
constructed carbene speciqj Lin be rather high.



d a speci


The C7H8 species most difficult to assess on the

Energy surface is the monomer 20. Perhaps the best way to

view this species energetically is to assume the view of
Dolbier57 who approximated a value for diradical 20 by first

assuming a i-bond energy in strain-free 20 of 60 kcal/mole

and subsequently subtracting out (a) 8 kcal/mole for the

allene destabilization energy and (b) 15 kcal/mole for the

allyl radical stabilization energy. The resulting value

(37 kcal/mole) was added to the heat of formation of "strain-

free" 20, 65 kcal/mole, to afford a final figure of 102

kcal/mole. This calculation, while only a rough approxi-

ation, probably represents a minimum value since Underwood's

INDO-MO calculations of 1,2-cycloheptadiene (21) under-
score the fact the 21 may indeed be a "distorted"allene as

opposed to a planar species, thus by analogy raising the

energy of 20 above the 102 kcal/mole figure approximated.

il the energy diagram shows that the carbene will cer-

i y convert to productss, it is also obvious that acti-
20 should convert to some or all of the species below

n the profile. In passing it should be noted that the

52 afforded a calculated heat of formation of ca. 88

/mole, a value which is probably high due to lack of a

able model for ring corrections.

A discussion of the chemistry of the tricyclic

ocAnylidenes is essen-tially a matter of the fragmentation

process occurring for cyclopropylcarbenes. The idea that
these systems could possibly give rise to the novel cyclic
allene monomer bicycle [5.1.0] octa-3,4-diene (84) was
developed early in the work with the tricyclic heptanylidenes.

< > dimer


2 1
Whereas the p singlet of both tricyclic heptanylidenes
would possess to a certain degree parallel geometry with
regard to the adjacent cyclopropane rings, the tricyclic
octanylidenes in the p2 singlet state would possess th
bisected geometry which plays such a vital role }n the
carbonium ion chemistry of th(se systems.30 It was tho
that this favorable added factor might help overcome


R bise R
bisected APw aral


1 I-

ssary enefget i expiredd for the trishomo interac-
n leading to monome( 84. Molecular models showed
hat the syn species was especially well set for this
e of interaction.
It appears from the pyrolysis results, though,
t any trishomo interaction in these systems will have
eestablished.y ,trapping reactions with olefins and
mmett studies in order to disclose the elctrophilic-
or nucleophilicify of these systems.
The obse ce of the 1,5-hydrogen shift in the
is-l-allyl ethynyl cyclopropane (65), the
primary c'arben oduct from the syn and arnti tri-
c octanyliden s nit without precedent since
d Hop d afacile 1,5-hydrogen shift
1 cis-1-ethynyl-2-methyl cyclo-

Swhih 1l converted to hexa-1,2,5-triene



(67) at the above specified temperatures.58 Analysis

of conformational effects favoring a ready transfer of

hydrogen in 65 shows that the situation most favored

would have an orientation where the vinyl group of the

allyl moiety is oriented away from the neighboring

triple bonded substituent as pictured below. This

situation would lead to a predomi

tetraene (66) having trans geomet

t formation o

for the bu'd

is supported by the observed uv extinction coefficient
27,100 for 66 as previously noted in Chapter III.

'In retrospect, the question of operational homo-
-omaticity in the tricyclic heptanylidenes 19-s and 19-a
in the tricyclic octanylidenes 25-s and 25-a has been
d. To fully answer the question, further work will be
red in order to demonstrate the electrophilicity or

ophilicity of said carbenes. It can be stated, however,
t such homoaromatic reactivity appears a definite possi-
y for 19-s and 19-a due to the nature of the observed
cts. The case for 25-s and 25-a is not as promising
rings to mind a mechanistic consideration which should
orted the fragmentation process of cyclopropylcarbenes
s mechanistic pathway may prove, in the final ana-
erative for both the tricyclic heptanylidenes and
idenes. The consideration afforded to the fragmen-
ts is one whereby the process may be visualized

d process employing th6 cheletropic designation;
on in which two a-bonds which terminate at
m re made or broken in a concerted fashion.47

Bsengagement of acetylene from ethylene
rba4 e as reactant has been treated by
y Lia to what is now known as
wing method.59 Recently Jones and
gantI~ implified pro the
^*ified Prob

MO following method in general which allows for better visu-
alization of the overall process leading from reactant to
transition state to product. The following are the "rules
of thumb," developed by Zimmerman and expanded upon by Jones
and Brinker,60 which are put to use in constructing the
schematic correlation diagram for the cyclopropylcarbene
fragmentation process:
(1) For the MO following, treat the carbene
as the sum and difference of two "p"
orbitals. "s" character is added after
compatibility of reactant, transition
state, and product orbitals is deter-
mined. Thus, during the "following,"
the orbitals of the lowest singlet state
of methylene would be represented by LS-1
and LS-2. Subsequent addition of "s"
character to the orbital used in the
reaction ((pl + pn) or (pI pn)) leads *
to the familiar filled sp2 orbital of the
carbene and depending on the orbital used
in the reaction, permits prediction of
the orientation of the carbene during
the reaction.

LS-1 LS-2

C C and = C

(P1 + Pn) (P1 Pn)

Add "s" Add "s"
o C to -
!Pi + Pn (Pl-Pn

(2) Number the transition state linear
array beginning with one carbene
orbital and ending with the other.

(3) Assign the familiar linear polyene
orbital signs to the transition state

(4) Follow each occupied reactant MO to
product, using the carbene MO that is
compatible with the highest occupied
transition state MO.
(pp. 7-9)
Thus, in conclusion of this chapter, the application

the Jones-Brinker extrapolation of Zimmerman's rules to

cyclopropylcarbene cheletropic reaction affords the

owing correlation diagram (Scheme V). The most impor-

t ob ervation52 to be obtained from the diagram is that,

"s" character is finally added to the proper carbene

ital used in the reaction (LS-1 vs. LS-2) under the

pices of rule 1, it will be added to pl + P6 (LS-1),

vertical p orbital, which correlates in symmetric

ashon with thl bonding product orbital T 34. The quali-

ve t, an important one in light of preceding dis-

n concerning the reactivity of p vs. a singlet con-

ions of cyclopropylcarbenes 19-s, 19-a, 25-s, and

i that the electron pair of the carbene must become

te n an sp2 hybrid in the vertical plane (o). Once

assumed only one mechanism is apt to derive,

tion. Closing the discussion on an opti-

Sstructure such as 19-a through the

parted by homoaromaticity, were able

Scheme V


+* *
2 -A

023 -45

c\r C

Symmetry element: vertical plane (o)
Transition State Products

* *



P s ---- sp2 (vertical) = o2

F the pi configuration. Could one reasonably assume
f e"of reaction for such a carbene? The con-
could very well be yes.





z =



Melting points were taken on a Thomas-Hoover melting

point apparatus and were uncorrected. Infrared spectra of

synthetic intermediates were recorded on either a Perkin-

Elmer Model 137 or a Model 437 spectrophotometer while the

spectra of pyrolysis products were recorded on a Beckiian IR

10 spectrophotometer. Nuclear magnetic resonance (nmr)

spectra were obtained from a Varian Model A-60-A spectrom-

eter and, less frequently, from a Varian Model XL-100 spe

trometer. Mass spectral data were obtained from both an

Hitachi Perkin-Elmer RMU-6E mass spectrometer and an AEI

MS 30 high-resolution mass spectrometer. Ultraviolet

tra were recorded on a Cary 15 double-beam spectrop

Elemental analyses were carried out by Atlan
Microlab, Inc., Atlanta, Georgia. Glpc work, analyst

and preparative, was performed on a Varian Aet

A-90-P3 gas chromatograph employing the cut-an

method of analysis in analytica ork. Four c

d in glpc work and are reference s foll

1) column A 3% FFAP o omaso

5 ft x 0.25 alumi

(2) column B 10% Carbowax 20 M + 3.5% KOH on
Chromasorb P-Regular, 10 ft x 0.25 in -


S (3) column C 6% SE-30 on Chromasorb P-Regular,
5 ft x 0.25 in aluminum.

(4) column D 18% DC-200 on Chromasorb P-Regular,
15 ft x 0.25 in copper.

All compounds which were not referenced were com-

cially available.

oration of diallyl ketone (29).

Tetramethyl -1,3-cyclobutanedione (27) was pre-

ccording to the procedure of Miller and Johnson:18

30-1140 (lit18 mp 1150-1160). Reaction of allyl

rd (1.25 moles) with 27 (0.50 mole) using the pro-

e of Dreyfussl9 afforded 5-allyl-5-hydroxy--2,4,4-
thyl-7-octen-3-one (28) in 55% yield (lit19 61%).
analyzed cleavage of 28 (22.5 g, 0.105 mole) em-
barium hydroxide19 followed by distillation of
lysate (bp 450-570/20 mm) on a 23 mm, 36 in
aust semi-works spinning band column afforded four
boiling (20 mm) at 320-380, 40-48, 490-550,

-560. Redistillation of the third fraction and

tion with the fourth fraction gave 7.5 g (0.068
f ketone (bp 550-56/20 mm, lit.19 bp 51-
inated with 5-10% allyl propenyl ketone.

Net yield of 29 was typically 62-64% (6.8-7.1 g, 0.062-

0.065 mole). Ir (film): 3025, 2960, 2920, 2870, 1730,

1640, 1420, 1385, 1360, 1325, 1285, 1215, 1135, 1105,

1070, 1050, 995, 920; nmr (CDC13): 6 3.23 (d with split-

ting, J = 7 Hz, 4 H), 4.90-5.32 (AB m, terminal vinyl H,

4 H), 5.58-6.33 (m, 2 H); mass spectrum (m/e): 110 (M ),
69 (M+ C3H5, major peak).

Preparation of 2,2-diallyl-1,3-dioxolane (01).

A solution of 6.95 g (0.0631 mole) of 29 in 70 ml

of benzene was mixed with 5.15 g (0.0830 mole) of ethylene

glycol and 0.05 g of p-toluenesulfonic acid monohydrate

in a 200-ml round bottom flask. The flask was fitted with

a Dean-Stark trap and a condenser (equipped with a drying

tube). The mixture was refluxed until 1.1 ml of water had
been collected (97% of the theoretical amount). The cooled
reaction mixture was washed with 20 ml of 10% sodium hy-
droxide solution followed by five 10-ml washes with watAr.
The benzene extract was dried over anhydrous K2CO3 and

the benzene removed by rotary evaporation. The residual
liquid was distilled at 20 mm affording 4.88 g (0.0316
mole) of 30 (bp 75-76, 50%). Preparative glpc on colu
A (column 820, He flow 60 ml/min) afforded an analytic 1

sample of 30. Ir (film): 3040, 2930, 2875, 1645,

1320, 1300, 1285, 1265, 1240, 1200, 1175, 1140, 1115,

n fi~

1040, 1000, 990, 920; nmr (CDC 3): 6 2.39 (d with

hitting, J = 7 Hz, 4 H), 3.90 (s, 4 H), 4.84-5.28 (AB m,

minal vinyl H, 4 H) 5.47-6.10 (m, 2 H); mass spectrum

in/e): 154 (M+), 113 (M+ C3H5, major peak).

0 Anal. Calcd for CgH1402: C, 70.10; H, 9.15.
found: C, 70.35; H, 9.11.

paration of cis- and trans-3,4-bis (chloromethyl)-
lopentanone (32

A mixture of 6.75 g (0.0437 mole) of 30 and 12.1

.0440 mole) of iodobenzene dichloride 20a in 75 ml of

oroform was heated at reflux for two hours under a slow

*m of nitrogen. The chloroform solvent was removed by

vaporation affording a slightly colored mixture

ng iodobenzene and crude cis- and trans-7,8-bis

thyl)-1,4-dioxaspiro [4.4] nonane (31). The
atio of 31 was not established although the

reeilinfpmulti'plet centered at 3.65 6 (-CH2C1)

r P ~~ e of a ois-trans mixture. The iodo-
s r moved at 0.25 mm while heating the flask to

istillation of 31 (bp 88-95/0.25 mm)

02 mole, 69%) of 31 possessing a

e use of hydroquinone stabilizer in

m rovedqut did not eliminate decom-

75 g .0300ole) was dissolved
ate ng 200 mg of

University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - - mvs