Citation
Attempts to generate cycloheptatrienylidene

Material Information

Title:
Attempts to generate cycloheptatrienylidene
Creator:
Ennis, Cecil Lawrence, 1943-
Publication Date:
Language:
English
Physical Description:
vi, 92 leaves. : illus. ; 28 cm.

Subjects

Subjects / Keywords:
Cycloheptatrienylidene ( lcsh )
Chemistry thesis Ph. D
Dissertations, Academic -- Chemistry -- UF
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis--University of Florida, 1968.
Bibliography:
Bibliography: leaves 87-91.
General Note:
Manuscript copy.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
The University of Florida George A. Smathers Libraries respect the intellectual property rights of others and do not claim any copyright interest in this item. This item may be protected by copyright but is made available here under a claim of fair use (17 U.S.C. §107) for non-profit research and educational purposes. Users of this work have responsibility for determining copyright status prior to reusing, publishing or reproducing this item for purposes other than what is allowed by fair use or other copyright exemptions. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder. The Smathers Libraries would like to learn more about this item and invite individuals or organizations to contact the RDS coordinator (ufdissertations@uflib.ufl.edu) with any additional information they can provide.
Resource Identifier:
030437441 ( ALEPH )
17018460 ( OCLC )

UFDC Membership

Aggregations:
University of Florida Theses & Dissertations

Downloads

This item has the following downloads:


Full Text














ATTEMPTS TO GENERATE CYCLOHEPTATRIENYLIDENE












By
CECIL LAWRENCE ENNIS













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










UNIVERSITY OF FLORIDA
1968















ACKN01IOWIEDG!NENT


The author would like to express his deepest gratitude to Professor W. M. Jones. Perhaps some of the many virtues and scholarly ideals to which he adheres have rubbed off

on this student during the past years. His frequent counsel, not only as a teacher, but as a friend, has left little to be desired. Because of him and the various members of his research group, this period of time has been educational and richly rewarding.

There are no words to express the feeling the author bears for his wife. An egiual amount of her blood, sweat, and tears have gone into not only the typing, but the making of this dissertation. For providing the steadying influence, this author is deeply grateful.

Finally, the author wishes to thank the College of Arts and Sciences of the University of Florida and the Gulf oil Company for providing fellowships during this period, and The U.S. Army Research Office, Durham, for further financial assistance.


11j












TABLE OF CONTENTS


Page

ACKNOWLEDGMENT. ii

CHAPTER

I. INTRODUCTION AND THEORY. 1

II. DECOMPOSITION OF 2,4 ,6-CYCLOBEPTATRIENONE
p-TOLUENESULFONYLHYDRAZONE AS A ROUTE TO
CYCLOHEPTATRIENYLIDENE

A. Synthesis of 2,4,6-Cycioheptatrienone
p-Toluenesulfonylhydrazone. 13

B. Decomposition of 2,4,6-Cycloheptatrienone p-Toluenesulfonylhydrazone Sodium
salt. 16

C. Decomposition of 2,4 ,6-Cycloheptatrienone p-Toluenesulfonylhydrazone Sodium
Salt in the Presence of Dimethyl Fuma-rate, Fumaronitrile, and Maleinitrile.
Formation of SpiroL2.6Jnonatrienes .20

D. Mechanism of the Formation of Heptafulvalene and the Spiro2.6lnonatrieies 30

E. Other Olef ins as Cycloheptatrienylidene Acceptors. 38

F. Decomposition of 2,4,6-Cycloheptatrienone p-Toluenesulfonylhydrazone Sodium
Sal~t under High Vacuum. 42

III. THERMAL REARRANGEMENT OF 8,9-DICARBOMETHOXYSPIROfh.6QONA-2,4,6-TRIENE TO 1,2DICARBOMETHOXYINDANE. 46

IV. ADDITIONAL SCHEMES EXAMINED AS ROUTES TO
CYCLOHEPTATRI ENYLIDENE

A. Base-promoted Decomposition of Methyl
N-Ni troso-N-cyc loheptatrienylcarbama te
and N ,N-Dimethyl-N '-nitroso-N'-cycloheptatrienylurea. 52


iii










B. at-Elimination Schemes 57.


V. SUMMARY. 6i

VI. EXPERIMENTAL

A. General. 63

B. Chiorocycloheptatrienylium Chloride 64

C. 2,4,6--Cycloheptatrienone p-Toluenesulfonyihydra zone Hydrochloride. 65

D. 2,4 ,6-Cycloheptatrienone p-Tcluenesulfonylhydrazone.*65

E. Photolysis of 2,4,6-Cycloheptatrienone
p-Toluenesulfonylhydrazone Sodium Salt.
Formation of Keptafulvalene. 66

F. Determination of the Yield of Heptafulvalene. 67

G. Photolysis of 2,4,6-Cycloheptatrienone
p-Toluenesulfonylhydrazone Sodium Salt
in the Presence of Dimethyl Fumarate.
Formation of 8 ,9-Dicarbomethoxyspiro[2.(Jnona-2,4,6-triene. 68

H. Isornerization of 8,9-Dicarbomethoxyspiro[2.(jnona-2,4,6--triene to 8-Carbomethoxy-8-carbomethoxymethylmethy-lenecyclohepta-2,4,6-triene on Acid
Alumina. 70

I. Photolysis of 2,4,6Cycloheptatrienone
p-Toluenesulfonylhydrazone Sodium Salt
in the Presence of Fumaronitrile.'
Formation of trans-B ,9-Dicyanospiro
[2.61nona-2,4,6-triene. 70

J. Isomerization of trans-8,9-Dicyanospiro [2.67nona-2 ,4, 6-triene t~o 8-Cyano8-cyanomethylmethyleriecyclohepta2,4,6-triene on Acid Alumina. 72

K. Photolysis of 2,4,6-Cycloheptatrienone
p-Toluenesulfonylhydrazone Sodium Salt
in the Presence of Maleinitrile.
Formation of cis-8,9-Dicyanospiro[2.67nona-2,4,6--triele. 73


iv


57










L.Photolysis of 2,4,6-Cycloheptatrieione
p-Toluenesulfonylhydrazone Sodium Salt
in the Presence of Other Acceptors. 74

14. Thermal Decomposition of 2,4,6-Cycloheptatrierione p-Toluenesulfonylhydrazone Sodium Salt with and without
Dimethyl Fumarate. 74

N. Photolysis of Heptafulvalene in the
Presence of Dimethyl Fumarate. 75

0. Photolysis of 2,4,6-Cycloheptatrienone
p-Toluenesulfonylhydrazone Sodium Salt
in the Presence of Dimethyl Fumarate
and Triethylamine. 75

P. Thermal Decomposition of 2,4,6-Cycloheptatrienone p-Toluenesulfonylhydrazone Sodium Salt under Vacuum. 76

Q. Thermal Decomposition of 2,4,6-Cycloheptatrienone p-Toluenesulfonylhydrazone Sodium Salt under Vacuum with
Dimethyl Fumarate. 76

R. Thermal Rearrangement of 8,9-Dicarbomethoxyspiro ~2. Gnona-2, 4'6-triene
to 1,2-Dicarbomethoxyindane. 77

S. Thermal Rearrangement of 8,9-Dicarbomethoxyspiro [2 .63nona-2, 4 ,6-triene
to 1, 2-Dicarbomethoxyindane in n-Decane and Triglyme. 78

T. Methyl N-Cycloheptatrienylcarbamate .78

U. Methyl N-Nitroso-N-cycloheptatrienylcarhamate. 79

V. Decomposition of Methyl N-Nitroso-Ncycloheptatrienylcarbamate with
Sodium Methoxide.S80

W. Decomposition of Methyl N-Nitroso-Ncycloheptatrienylcarbamate with
Sodium Methoxide in the Presence of
Dimethyl Fumarate. 80

X. Decomposition of Methyl N-Nitroso-Ncycloheptatrienylcarbamate with
Potassium t-Eutoxide. 81


V










Y. Decomposition of Methyl N-Nitroso-Ncycloheptatrienyicarbamate with
Potassium t-Butoxide in the Presence
of Dimethyl Fumarate. 82

Z. N ,N-Dimethyl-N'-cycloheptatrienylurea. 82

EA. N ,N-Dimethyi-N' -nitroso-N' -cycloheptatrienylurea. 83

BB. Decomposition of N,N-Direthyl-N'nitroso-N' -cycloheptatrienylurea
with Sodium Methoxide. 84

CC. Decomposition of N,N-Dimethyi-N'nitroso-N' -cycloheptatrienylurea
with Sodium Methoxide in the
Presence of. Dimethyl Fumarate. 85

DD. Carboxycycloheptatrienylium Fluoroborate. 85

LIST OF REFERENCES. 87

BIOGRAPHICAL SKETCH. 92


vi














CHAPTER I



INTRODUCTION AND THEORY



Although carbenes are normally electrophilic species (1,2) by virtue of their sextet of electrons, a ground state singlet carbene in which the nonbonding pair of electrons is localized in the sp orbital and the empty p orbital is stabilized by strong electron donation is potentially nucleophilic. Conjugation of the p orbital with nonbondi4ng electrons on adjacent heteroatoms adds electron density to the carbene carbon and profoundly influences its reactivity. There are several examples of stable divalent carbon compounds in which the adjacent heteroatom is doubly bonded t'-o the electron-deficient carbon. Carbon monoxide, I, isonitriles, II, and fulminic acid derivatives, III, behave as nucleophiles toward Lewis acids and metal carbonyls (1,2).


1---


I II



III



In the recent literature, there have been extensive,


1







2


accounts of research on divalent carbon singly bonded to heteroatoms. Tetraaminoethylenes have been investigated by a number of workers (3,4,5,6,7,8,9,10,1_1,12,13). The. reports have centered around the purported dissociation of dimers such as IV to carbene V in a reversible fashion. The carbene is stabilized by electron donation from the nonbonding electrons on the nitrogen atoms into the vacant p orbital of the carbene carbon. Evidence for the dissociation includes consistently low molecular weights for the olefin, IV, by the Rast method.


R R
I I
R'-NN-R'

C=C

R'-N N-R'


R

R'- N

C:

R'- N

R


IV


Dimers of this type do not react with olefins with the exception of tetracyanoethylene. With this highly electrondeficient olefin, a charge-transfer species, VI, is formed rather than a normal carbene adduct (4) The dimers, however, do react with oxygen, sulfur, diazo compounds, and Lewis acids to give products expected of carbenes.


NC CN



NC CN


FR'RN NR'RN I

+ C-- C "

ViV


R R '- N +
\\C



R


V


IV +








3

There has been considerable doubt cast on the validity of these dissociations by crossover experiments in which mixtures of dimers with different R groups failed to qivre crossed products (6,7) An alternate reaction scheme Epostu-lated to incorporate the results of the crossover experiments involves reaction of the dimer directly with the electrophile. This scheme still involves the carbene as a bypro-duct which can dimnerize to regenerate the olefin or react with the trapping agents in a normal carbene process. R'R NRR'RN E NR'R R'RN-i E N'


R'RN R'R R'RN IRK;



R'RN R R'N NR'R R'R



R'RN R'RN NR'R





Similar carbenes have been proposed to explain the

acidity of tetrazolium cations, VII (14,15) and thiazolium~ cations, VIII (16,17,18,19,20). Early workers actually claimed to have isolated the latter carbene as a stable, crystalline solid but later retracted their statements (17,20). The thiazolium carbene has also been generated by ca-elimination (21) and decomposition of N--azides(22).













H




~N -N N-N

R R R R


H


N-P



R Ri


VII


C
.S: :NRP


Rk R/I

VIII


Another similar species is IX which was generated by a decarboxylation scheme (23). A report has been made on the possible intermediacy of Xa, but it reacts as its isomer, Xb

(24).


ii N

R


LiO-C- (2t (Et) 2


Xa


Xb


Ix


There have been several reports of attempts to generate dioxocarbenes. Thermal decomposition of norbornadienone ketal, XI, gives tetraalkoxyethyl1ene, XII, as a major product under aprotic conditions. In alcohol solution, decomnposition gives little or no ethylene but substantial yields of orthoformates, XIII (25,26,27,28) The ethylene, XII, was found by one group to undergo dissociation-type reactions similar to the tetraaminoethylenes (29) An approach involv-


4










ing base-catalyzed decomposition of the p-toluenesulfonylhydrazone, XIV, gave only radical products under aprotic conditions (30).


RO OR




ROH

XI


ZO I'NOR


XII

~OR H --C -OR
NOR

XIII


CHO3




XIV


An alternate approach to enhancing the nuceophilicity of carbenes involves conjugation of the p orbital with an olefinic system. For maximum effect, this p orbital can be incorporated in a. Tr system which will be aromatic if the orbital contains neither of the nonbondingx electrons of the carbene. This may allow the singlet, in which the two electrons are in the sp orbital, to become the ground state. The simplest example of this system is cyclopropenylidene, XV, in which the carbene p orbital is involved in cyclopropenylium-type resonance.


R





R


R





R


XV






6


Diphenylcyclopropenylidene, XV (R = phenyl), has been generated by n-elimination of carbamnic acid from the carbamate, a-el-im-ination of carboxy'Lic acids from the corresponding esters, and base-catalyzed decomposition of N-nitrosoN-(2,3-diphenyl--2cyclopropenyl)carbamoate (31,32,33). It was found to be nucleophilic, reacting with electron-deficient double bonds such as dimethyl fumarate, N,N,N',N'tetramethylfumaramide, and fumaronitrile to give the corresponding spiropentenes, XVI, which underwent facile rearrangement to the triafulvenes, XVII.


Ph R Ph R Ph

/> >CH 2
Ph> R Ph :< R Ph

XVI XVI

R. Breslow and L. F. Altman have pointed out that perhaps one factor contributing to the acidity of the ring hydrogen of n-propylcyclopropenone 4S contribution of a cyclopropenylidene resonance form (34).





>4- 0 0

Pr Pr

The work which will be described here concerns the next higher case in which the Tr system will be aromatic, cycloheptatrienylidene, XVIII. Here the p orbital of the







7


methylene is involved in tropylium--type resonance.


CC)


XVIII


The di-- and tribenzoderivatives, XIX and XX, Of Cycloheptatrienylidene have been generated, and their low-temperature esr observed. The carbenes were reported to have triplet ground states and to react with butene-2 (35,36,37). Because of the involvement of the aromatic rings, these systems are apparently representative of aryl-substituted methylenes. Comparison of their properties with diphenyl-methylene, XXI, and 2,3 ,6,7-dibenzocycloheptadienylidene, XXII, makes this analogy even more obvious. Both XXI and XXII are reported by the same group to have triplet ground


00


J. XX




00


XXI XXII


xxi









states and to react with nucleophilic olefinis such as butene2.

The theoretical foundation for the foregoing qualitative discussion has recently been set forth by R. Gleiter and R. Hoffmann (38). 1 A linear methylene has two degenerate orbital~s in which to distribute the two carbene elec-trons. The lowest energy state (ground state) of such a species is a triplet with the unpaired electrons in separate orbitals. To obtain a singlet ground state for a methylenie, the degeneracy of these two orbitals must be destroyed, and, according to Gleiter and Hoffmann, the resultant splitting must be greater than 2 ev. They have postulated three factors which influence the multiplicity of the ground state and the relative nucleophilicity of the methylene. The carbene can be modified by 1) bending the R-C-R angle, 2) approaching or connecting the methylene to a system with lowlying unoccupied levels, or 3) approaching or connecting the methylene to a system with high-lying occupied levels.

The first way, involving bending of the R-C-R angle, distorts the p orbital in the plane of the bending (p ) while the second orbital (p x is virtually unaffected. As the bond angle is decreased from 180 0, the distorted orbital acquires s character and is therefore stabilized. P will be



(1) Since the ensuing material has not yet appeared in print, it is discussed in much greater detail than would normally be the case.






9


referred to as 0 in the following discussion. The splitting developed between G and p x is not great at the higher angles and becomes important in cases such as cyclopropylidene where the splitting is calculated to be 1.13 cv. The optimum case is vinylidene where the angle is essentially 0 0, and the splitting is calculated as 1.92 cv.

The second case concerns conjugation, of p, with lowlying unoccupied 'Levels. In this case, the methylene would transfer electron density to the wr system and thus become more electron deficient itself. The unoccupied orbital of the ir system must have the proper symmetry to mix with






px





METHYLENE HI SYSTEM


of the methylene to form a cyclic polyene. Thus, in all systems with 4n ff electrons, the lowest unoccupied molecular orbital has the correct symmetry to mix with p x Conversely, in all ir systems with 4n + 2 7r electrons, the highest occupied molecular orbital has the proper symmetry. In the 4n case, stabilization of p x does not increase splitting because a

(p ) has also been stabilized by bending. In cyclopentadienylidene, the splitting was calculated as 0.13 cv. The symmretr-.y of the orbitals is indicated by S (symmetric) and









A (antLisymmetric)


S -- S









In the third case, the methylene is conjugated with a

high--lying occupied level. As mentioned above, in a 'ir syster with 4n + 2 7iT electrons, the highest occupied molecular orbital has the correct symmetry to mix with p x. This results i.n a new 7 system in which the lowest unoccupied orbital. is higher than p x. Furthermore, if a is stabilized by bending, the splitting is reinforced. The optimum case should be cyclopropenylidene where n = 1. The interaction diagram for the 7i system is shown below. The Gj-p X splitting amounts to

3.17 ev which leads to a prediction of a singlet ground

state.



A


S --1 -S






C ==C Px







I I


In the general case, there appears to be a possibility of stabilizing a singlet ground state when the miethylene

interacts with a -,-i system which already contains 4n -f 2 IT electrons. A number of systems which fit this requirement are shown below.



I 9



N


3.17 cv 3.45 ev 1.77 ev 1.00 ev

XXIII XXIV XXV XVIII



Derivatives of the carbenes XXIII, XXIV, and XXV have been discussed and were found to react with electrophiles and electron-deficient olefins. There is nothing known about their ciround states, however. The splitting for carbene XVIII is somewhat low and by the above standards would be predicted to have a triplet ground state. experimental evidence will be presented, however, which will argue for a singlet ground state.

The effect of the Tr system on the nucleophilicity of the

carbene must also be considered. Gleiter and H1offmann chose to examine the total charge on the methylene carbon. There are two distinct cases:

1. p2 lower than G The two electrons of the carbene
are now delocalized through the 7 system so that the
methylene loses e-lectron density. The carbene should
be more electrophilic.





12


HC---CHO ("lIinear) +0. 33

HC -NO02 (13)0,) +0. 41 XXVI 1

2. a lower than p 2 Electron density is donated to the
methylene center by the polyene which should make the
carbene more nucleophilic.



C 1 (120 0) -0.50 fQ.:-0.43

HH
C H
6 5,C: (120') -0.56 -0.65
H K


-0.68 (bent 200) -0.43



1-0.86 XXVIII



Therefore, by the above arguments, cycloheptatrienylidene should be nucleophilic, reacting preferentially with electron-deficient double bonds.

The purpose of this paper is to report several at-pt to generate cycloheptatrienylidene (39). The Bamford-Stevens reaction which involves base-promoted decomposition of p-toluenesulfonylhydrazones, and base-catalyzed decomposition of N ,N-dirnethyl-N '-nitroso-N '-cyclohept'atrienylurea and methyl NI'-nitroso-N-cycloheptatrienylcarbamate was examined.















CHAPTER II


DECOMPOSITION OF 2,4, 6-CYCLOHEPTATRIENONE
p--TOLUEIN ESUI.FONY 11LHYDRAZ OIN E AS A ROUTE TO CYCLOT{,'EPTIATPJ "EN\YIDfENE


A. Synth esis of 2,4,6-CyclohepLatrienone p-Toluenesulfonyl-hydrazone.

A popular method ofgenerating diazo compounds and carbenes is the Bamford-Stevens reaction in which p-toluenesulfonyihydrazones, XXIX, are decomposed in the presence of base

(40). The reaction is particularly useful since the decomposition can be effected thermally or photochemically under relatively mild conditions. Since the hvdrazones are ketone derivatives, availability of the properly substituted ketone




R H 1) Base R R
NCNN-SO 2Ar >. /c- C-N
R 2) A or liv RR


XXIX

is important. Tropone,XXX, is easily synthesized by selenium dioxide oxidation of cycloheptatriene (41) but because of a highly polarized resonance form, it does not alw-ays undergo normal ketone reactions. Treatment of tropone with hydrazine does not produce the expected hydrazone, but 2aminotronone instead (42) Recently, the imi ne and several


13







14


substituted imines have been, synthesi*zedt although somewhat indirec'-ly (4-3,44) rThe~ phenyihycdrazone (45) and azine (46) of tropone, however, have been made by classical methods; and fortunately, the p-toluenesulfonylhydrazone conformed to the latter cases.

By stirring a solution of tropone and p-toluenesulfonylhyvdrazine in 1% HCl in absolute ethanol, a yellow solid was obtained which was shown to be the hydrochloride salt of tropone p-toluenesulfonylhydrazone, XXXIII. The yield could be improved from 50% to above 90% by first converting the tropone to chlorotropylium chloride, XXXII (47), with thionyl chloride followed by treating an absolute ethanol solution of the chloride with a solution of p-toluenesulfonylhydrazine




1% HCl
-0 + 12N NHSO 2 Ar >- -N-HSOArHCl


XXX XXXI XXXIII

S1 ~ Cl H NNSO2x

Cl


XXXII



The nmr spectrum of the salt in D20 shows in addition

to the methyl7 absorption, a multiplet partially superimposed on the downfield portion of the aromatic A 2 B2 absorption. The position and structure of this multi~plet is characteristic of trcopone, tropylium salts, and monosubstituted




is


tropylium salts and indicates a highf degree of tropylium ion character confirming the structure assigned to XXXIV.




2 EII~ CH 3
GCl Si i

XXXIV


When an aqjueous solution of XXXIV was titrated with dilute base, the light yellow solution became cloudy, then rapidly cleared to form a dark red solution. The pBI Was adjusted until the cloudiness remained, and the orange mixture was extracted with dichlorcmethane. After drying the deep red solution, pentane was added, and the neutral ptoluenesulfonylhydrazone, XXXV, was obtained as bright red crystals. Two crystalline forms were found; rapid crystallization produced deep red plates, mp 142.5-143.5 while slow crystallization formed clusters of needles, mp 144-145. The ir and nmr spectra showed both forms to be the same compound.

From the behavior of the p-toluenesulfonylhydrazone in the presence of acid and base, it was apparent that there was yet another species present in strongly basic solution. The acidity of the nitrogen proton in p-toluenesulfonylhydrazones is well known, and is in fact, the basis of the Bamford-Steveis decomposition reaction. Therefore, it was not unexpected to find this third species, XXXVI, which is the conjugate base of XXX"'V. All three species are quite





16


stable and are readily int-.ercoiverteT ,,ith ac.-d or base.


H 11

4 --N- so 2Ar

Yellow

HCO 3/ + XXXIV

H Na+
=N-N--so Ar 011N-N--SO Ar

Red H+ Brown





Several attempts were made to isolate XXXVI in a pure form, but none were successful. The salt typically precipitated from solution as a muddy brown solid which clung tenaciously to 'the last traces of solvent and decomposed before drying completely. To circumvent this problem 4Ln reactions where the dry salt was needed, an aprotic solvent, tetrahydrofuran, was used to generate the species. Almost any base was found to be satisfactory for removing the proton of the p-toluenesulfonylhydrazone, but sodium hydride was chosen because of the lack of any contaminating byproducts. Despite reports to the contrary (48) this base was found to be entirely satisfactory. B. Decomposition of 2,4,6-Cycloheptatrienone p-toluenesulfonyihydrazone Sodium Salt.

Decomposition of the sodium salt of a p-toluenesulfonylhydrazone is believed to proceed through the corresp onding diazo comrnound which is frequently an isolable intermediate.





17


The stability of the diazo compound is largely determined by the groups attached directly to the c-carbon ato (49).

Groups which can conjugate directly with the diazo

carbon tend to stabilize the molecule, and strongly electronwithdrawing subs tituents, such as trifluoromethyl, can give rise to highly stable diazo compounds. Bis-(trifluoromethyl)diazomethane, XXXVII, and diazocyclopentadiene, XXXVIII, are two well-known examples.



CF 3\ + + ~

CF3/ 2 10 4
CF_'


XXXVII XXXVIII XXXIX


If the groups are electron-donat'-ing, however, the aiazo compound may be unstable (50). -Diazocycloheptatriene, XXXIX, would fall in this category, and relative to diazocyclopenta-diene, should be quite unstable.

D. G. Farnum has reported a variation of the BamfordStevens reaction in which good yields of diazo compounds were obtained (51) A pyridine solution of the p-tolee sulfonylhydrazone and sodium methoxide is heated gently, then poured into water, and extracted with pentane. When this reaction was applied to XXXV, only starting material was recovered from the reaction mixture. In another attempt to generate XXXIX, a sample of the sodium salt, XXXVI, was placed on a column of basic alumina. Elution with aprotic solvents gave no results. Elution w-ith ethanol caused pro-tonation of XXXVI, and a quantitative yield of XXXV was





18


obLained as a red band. NumerouLs attempts were made to trap a diazo intermediate on a liquid nitrogen cold finger from thermal decomposition of XXXVI under high vacuum without results. Also, attempts to find evidence for a pyrazoline intermediate when the decomposition was carried out in the presence of olefins were unsuccessful. These results, how-ever, do not rule out the possibility of an inermediate diazo

compound.

When a slurry of the sodium salt, XXXVI, in tetrahydr-o-furan was photolyzed, about. 90% of the theoretical. gas evolution was observed. Filtration of the dark red reaction mixture gave up 95% of sodium p-toluenesulfinate as a light brown powder. The filtrate was poured into water and extracted with pentane. After drying and concentrating the extracts, a black, oily solid remained. The uv spectrum showed absorptions at 234 and 362 mo which were unaffected upon short exposure to acid. After several minutes with acid, the absortions gradually disappeared. A tlc of the residue from the reaction mixture showed a fast moving red spot in 25% ether in hexane and a slow moving red spot, attributable to unreacted p-toluenesulfonvlhydrazone. Upon standing on the dry plate, the fast moving red spot decolorized within five minutes. Since the compound was apparently decomposing, deactivated basic alumina was used for column chromatography. Elution with 5% ether in hexane gave a sharp red band. Concentration of the deep red. clutions under a stream of dry nitrogen





19


gave permanganate-colored plates (~mp J.1114-3.16 0) which were identified as heptafulvalene, XL. Recrystallization could be effected from methanol-water, and a pure sample was obtained with mp 119-121O Solutions of the dark crystalline solid were deep red in color but decolorized within several days, while the pure crystalline solid decomposed within one hour when exposed to air. The nmr spectrum of the solid showed only a multiplet in the olefinic region, and the mass spectrum showed a molecular ion of m/e 180. The properties of the compound were identical with the properties as reported by Doering (52) This olefin is the dimer expected from dimerization of cycloheptatrienylidene.





XXXVI h -+ N 2 + NaScO 2Ar


XL


Molecular orbital calculations on heptafulvalene show two electrons in a low-lying antibonding 'iO (53). Consequently, the compound should be rather unstable and highly susceptible to oxidation. This was found to be true since several attempts to obtain an analysis of the diner were unsuccessful. Although it could be purified easily by recrystallization or sublimation, heptafulvalene rapidly decomposed in the crystalline state. A sample kept in the dark, under nitrogen, and at -10 0 still decomposed within two weeks.






20


An attempt to obtain an analysis in this laboratory on a freshly purified sample was also unsuccessful. The olefin could be k1ept for extended periods in pentane solution at dry ice temperature and under nitrogen.

Because of this instability, acceptable yields of the

dimer could not be obtained by weighing the dry, crystalline solid. Since the 362 mvi peak of heptafulvalene was free of interfering absorptions, it was used to determine the concentrations of accurately diluted solutions of the dimer and thus the yields. The extinction coefficient of this absorb tion was reported as 21000 (52) but values obtained in this laboratory on -freshly purified samples w.ere 21400, 25000, and 26000. The lowest value was probably the least accurate because of the length of time between purification and determination of S. The value selected for determining yields was 25000. The value of the yields for comparison purposes is questionable at any rate, because the dimer was also found to be photolytically unstable, and the yields were very much dependent on the length of the photolyses. The dimer yield does give a minimum for the extent of reaction. C. Decomposition of 2,4,6-Cycloheptatrienone r_)-Toluenesuj-f onylhy ,dra zone Sodium Salt in the Presence of Dimetfhyl Fumarate, Fumaronitrile, and MaleinitL rile. Formation of SoiroL2 .6 nonatrienes.

When nucleo-hilic olefins such as cyclohexene and

butene-2 were added to the reaction mixture- in five- to tenfold excess in an attempt to trap the carbene or the diazo compound, there was no noticeable change. Gas evolution weas






21


still within 90%6 of theoretical. l, and after the same workup, the dim~er yield was between 40 and 7096. No evidence of an adduct between the carbene and these olefins was found. This was not unexpected, for if the carbene was actually nucleophilic in nature, it would not be expected to react with nucleophilic olefins but with those which contained electron-deficient double bonds.

When the electr-on-deficient olefin, dimethyl fumarate, was added to the reaction mixture, gas evolution was still normal; but at the end of the photolysis, the normally dark red solution was almost colorless. Filtration of the mixture gave a high yield of sodium p-tcluenesulfinate which showed that the decomposition had proceeded as expected. The reaction mixture was again poured into water and extracted with pentane. After drying and concentrating the solution, tlc showed a small quantity of the dimer and one other colorless spot which upon standing on the dry plate turned bright yellow. Since this color change was suspected to be due to an acid catalyzed rearrangement, basic alumina was used for chromatography; and to shorten the elution time, the adsorbent was deactivated to activity grade III. Careful cooling of 'the pentane solution before chromatography allowed removal of most of the unreacted dimethyl fumarate by fractional crystallization. Elution with 5% ether in hexane gave the dimer as a dark red band in lowa yield (0-10%) By gradually increasing the ether content to 10%, the compound observed on tlc was eluted immediately before dimethyl





22


fumarate. Evaporation of the solveriL gave white needles with mp 65-680 and mixture mp with dimethyl fumarate 63-97. The comp-ound was identified as the spirononatriene, XLI,' which formally, is the adduct of cycloheptatrienylidene and dimethyl fumarate. The ir spectrum showed a carbonyl absorption at 1720 cm 1; and the uv spectrum showed an absorption at 261 mpi which became 341 mp with acid and 341 and 253 miJ with subsequent treatment with triet-hylamine. The nmr spectrum showed a multiplet from T3.20 to 3.83, a doublet centered at T4.55, a singlet at T6.3l, and a singlet at T7.69 with area ratios 4:2:6:2. The mass spectrum showed a parent peak at in/c 234, and analysis indicated C13 H 14 04 as the empirical formula.



COOCH 3



COOCH 3

XLI



As mentioned above, the spot dlue to the spiro compound

on a tlc plate turned bright yellow upon standing. Also, the behavior of the uv spectrum of XLI upon treatment with acid arid then base (vide supra) indicated a rearrangement rather than a simple protonation. In the analogous diphenylcyclopropenyl system- the carbene-dimethyl fumarate adduct, XLII, underwent a facile, acid-catalyzed rear-rangement to the triafulvene, x=~ (31,32,33). It was therefore suspected that a





23


similar reaction was taking place with XLI .


Ph COOCII 3+Ph / O C

I > C OH

Ph < COOCHI3 PhC27'C3
XLI I XLIII



XIH+ -- COOCH3
\L HH ) z C OCH
OC

XLIV


Accordingly, a sample of XLI in anhydrous ether was placed on a column of dry acid alumina, activity grade I (alumina of lower activity gave poor conversion if at all), and eluted onto the column w ith a small quantity of ether. Upon contact with the alumina, the colorless solution turned bright yellow. Elution of the yellow band with acetone and concentration of.L- the solution gave a dark oil. By cooling a pentane solution of the oil in a dry-ice-acetone bath, a fluffy, yellow solid was obtained which became an oil agaiLn at room temperature. Upon slowly cooling the solution to

-20O yellow needles of the heptafulvene, XLIV, were formed which after several recrystallizations from pentane gave mp 49.5-51O The ir spectrum showed a pair of carbonyl peaks at 1750 and 1705 cm- and the nmr spectrum showed a doublet at T2.57, and a multiplet at ca. -r3.60 for the six cycloheptatrienyl protons. There wecre two peaks for the nonequivalent






24


methyl esters at T6.29 and 6.34 and a two-proton singjlet. at

-6.65. The mass spectrum-, w-as almost identical with that of XLI.

The degree of stereoselectivity of the addition of a carbene to an olefin is frequently used as evidence of the multiplicity of the carbene as it adds (1) Consequently, it was highly desirable -to determine the steric course of the photolysis of the salt, XXXVI, with dimethyl fumarat[-e. If the spiro compound from the reaction with the trans olef in has the trans configuration., as in XLIa, then perhaps the cis olefin, dimethyl maleate, would produce the cis spirononatriene, X~lb.


H1 COOCH 3 H C00CH 3




H,000H H COOCH3
33
XLI a XLIb

Photolysis of the sodium salt, XXAXVI, in the presence of dimethyl maleate also produced the spiro compound in comparable yield. This sample was identical in every way with that obtained from the reaction with dimethyl fumarate. If this were one isomer of a cis-trans pair, some differences in the nmr spectrum would be expected. For example, the posltions of the cyclopropyl protons in the two isomers below differ by 0.30 ppm (54).


H H H C00''

CH 0000CH 300 \C > KC103






25


A possible explanation for the failure to find a second stereoisoner arose when it was found that dimethyl maleate rapidly isomerized to the trans isomer tinder the reaction conditions. Either base or uv light caused the isornerization to a final equilibriumn mixture consisting of ca. 10% cis and 90% tra-ns olefin.



H>1 H- h H COOCH3

CHO30C 3 uOOCI13Ci3 H


-B-B


H3 0CB



If the addition of the carbene to cis and trans olefin

proceeded at the same rate, then a small amount of cis spironionatriene would be expected. R. Buisgen has shown, however, that the addition of 1,3-dipoles to dimethyl fumarate and dimethyl maleate is much faster with the trans isomer (55). The difference in reactivity is attributed to steric hinderance to resonance in the cis isomer, which weakens the reactivity of the olef in with nucleophiles, and to increased steric repulsion of the cis ester functions as they chance from sp 2to sp3.

The use of other electrophilic olefins was limited

primarily by their instability to basic conditions. Tetracyanoethylene, a strongly electron-deficient,_ olefin, is






26


quite unstable to bases and decomposes rapidly in the presence of sodium hydride. Fumaronitrile, XLV, is also unstable to bases, but the decomposition is much slowecr. When the photolytic decomposition of the salt, XXXVI, was carried out with a fivefold excess of this olefin present, the

nitrogen evolution was normal. The reaction mixture was dark red,indcating the presence of heptafulva.ene as, at least one of the reaction products. A tlc of the mixture confirmed the presence of the dimer and showed one other major spot. Upon standing for several minutes on the dry plate, the latter spot turned bright yellow as did the carbene adduct with dimethyl1 fumarate. After removal of the tetrahydrofLuran, nmr of the residue showed a series of multiplets in the olefinic region quite similar to XLI. A singlet at T8.02 completed the pattern expected of the carbene adduct with the olefin. The dark residue was chromatographed on deactivated basic alumina to give the dimer (18.5%) with 5% ether in hexane. Upon increasing the solvent polarity to 30% ether in hexane, the spot observed earlier on t1-lc was eluded. Concentration of the eluted fractions gave the new compound as a mixture with fLumaronitrile. The olefin was easily sublimed from the mixture, leaving the crude spiro-nonatriene, XLVIa, with mp 110-115 Rersalzainfo

benzene-pentane to constant melting point gave white needles with mp 121-123. An nmr spectrum of the cr aial showed only those absorptions described previously. The uv spectrum showed an absorption atL 262 mpi, and the ir spectrum





27


showed a nitrile peak at 2250 cm-i The mass spect-rum contamned a molecular ion at rn/c 160, and analysis indicated C 1 1 8 N 2as the empirical florrau~a. This compound was later shown conclusively t~o b-e the trans isoxer.

CN
NC\
XXXVI + C C \ hv 1

NC

XLV XLVIa


As mentioned previously, this spirononatriene also turned bright yellow on a dry tlc plate. Since this was indicative of an acid--catalyzed isomerization analogous to that of XLI, a sample of XLVIa was placed on a dry column of acid alumina (activity grade I) and eluted onto the column with ether. The solution turned orange upon contact with the column. Elution of the rearranged product with acetone gave a red solution which was concentrated to ca. 1 ml under a stream of nitrogen. Pentane was added, and red crystals of the heptafulvene, XLVII, were obtained which were filtered and dried, mp 97-99O Recrystallization from benzene--hexane gave red plates, mp 99.5-100.5. Analysis and mass spectral data indicated that this compound was isomeric with XLVIa. The uv spectrum showed absorptions at 247 and 346 mP and the ir spectrum showed nonequivalent nitriles at 2255 and 2195 cm- confirming the assignment.





28


XLVIa N
~CliL- CN


XLVII


At this timne there was still very little known about

the stereo-chemistry of the addition process; and before any real conclusions could be drawn, the cis spiro compound had to be isolated and characterized. E-Fforts in this direction, in the case of the earlier spiro compound, had been to no avail because of the equilibration of the cis and trans olefin pair under the reaction conditions. It was hoped, therefore, that this problem would not arise with the new olefin. Accordingly, the cis isomer of fumaronitrile, maleiLrnitrile, was tried in the photolysis reaction. Again, a. tlc of the crude reaction mixture showed the presence of a spiro compound. Examinati on of the nmr spectrum of the oil after removing the solvent showed, that indeed, the cis olefin had not equilibrated with its trans isomer during the reaction. The position of the cyclopropyl protons of the spirononatriene from this reaction appeared to be 2-3 cps downfield from the analogous absorption in the trans case, but the conclusive evidence for the cis isomer lay in the absorption=clue to the 2 and 7 protons of the cycloheptatriene ring. In the trans isomner, XLVIa, these protons are equivalent, each in the vicinity of a proton and a nitrile groutp on the






29


cyclopropyl ring. Consequently, they appear as a broad doublet split by the 3 arnd 6 protons and further by the 4 and 5 protons. In the cis isomer, however, the 2 and 7 protons are nonequivalent. One is on the side of the spiro cyclopropyl ring with two protons, and the other is on the side with two nitrile groups. Therefore, in the nmr spectrum of the cis isomer, there are two distinct multiplets for these prot4-ons, each integrating to a relative area of one.





C P4





XLVI a XLVIb

The ir spectrum of the cis isomer, XLVIb, is quite

similar to that of XLVIa, but at the same time, it is obviously nonsuperimposable. The uv and mass spectra of the two isomers are identical. Further proof of the isomeric nature of XLVIb is found in its isomerization to a heptafulvene which was identical by ir and glc with XLVII, the heptafulvene obtained from XLV~a.

The maleinitrile which was used in this reaction was

contaminated with 13% of the trans isomer. The nmr spectrum of the crude reaction mixture described above showed that this ratio had changed little if any during the course of the reaction. 2A close examination of this spectrum showed a






30


small but sharp singlet at the positioni of the cyclopropyl protons in the trans isomer, XLVIa. Although the region around this absorption was much too noisy to allow integration for comparison purposes, the peak area could be estimated as ca. 10% of the peak area for the cis isomer. Thus, it appears that the carbene addition to this olefin is highly stereoselective if not stereospecific.



xxXV i



H ,CN NC CN
h\) hv
NC H H =H

H CN NC CN

NC H H H





H+H

XLVIa ;/ XLVIb


I /AN
I O___ \CHf--CN


D. Mechanism of theFormation of Heptafulvalene and the Spir nonatrienes.

This brings up the question of how heptafulvalene and

the spirononatrienes are formed. Formation of the latter can be envisioned as proceeding by several different routes. Since the olefins used are electron deficient, they are





31


quite susceptible to base attack. A addition of the anion of the p-toluenesulfonyl1hydrazone across the double bond, followed by cyclization to the pyrazoline, XLVIII, and decomposition to XLI by loss of nitrogen is certainly a rational mechanism.



CH1000 H
XXXVI + COOCII d.~N4 0 Ar
3 2
CH--0 CH00




XLI -N

XI
COOCU i3

XLVIII


Participation of the olef in before or during the ratCedetermining step can be excluded by a simple kinetic scheme. The salt, XXXVI, was found to undergo smooth thermal decomposition above 100 0to give the same products as the photolytic decomposition. Thus, two reactions were carried out under identical conditions in which the salt was decomposed thermally at 101.5 0 (ref luxing dioxane) in dry diglyme. One reaction contained a twofold excess of dimethyl fumarate and the other contained none. Nitrogen evolution was followed as a measure of the rates of the two reactions. If the olefin was involved in a step before or during 'the rate-determining step in the decomposition, its presence should then cause a noticeable increase in the rate of the reaction if





32


XLI is a major product w,,hich forms at the expense of the dimer. The rates of the tw%-o reactions were virtually identical with half.-lives of approximately 11 min. In the reaction without olefin, the yield of heptafulvalene was 33%; and in the reaction with dimethyl fumarate, the yield of XLI was 38%, while the dimer yield dropped to 9%.

Reaction of heptafulvalene with dimethyl fum.arate in a reaction analogous to that proposed by Lemal (6) to explain the products of reaction of electrophiles with tetraamino-ethylenes vicee sunra) is also a possibility as the source of XLI. This path was excluded by photolysis of a solution of heptafulvalene and the olefin in tetrahydrofuran. Although the dimer decomposed under the photolysis conditions and a





hv Decornoosition
Dimethyl Fumarate




colorless solution remained, no trace of the adduct, XLI, could be found. The only isolable product was dimethyl fumarate.

Another possible route involves diazo addition to the olef in1 followed by rapid decomposition of the resulting pyrazoline. This reaction sequence is quite common in carbene chemistry, and, in a number of cases, both the diazo compound and the pyrazoline are isolable intermediates (1,2).





33


hvXV '2-: XL I
2 C- __C -COOCH.
-) COOCH 3




Unsuccessful attempts to isolate diazocycloheptatriene have already been described, and an attempt to intercept a 1pyrazoline intermediate was likewise unsuccessful. It has been shown (56) that unstable 1--pyrazolines can be intercepted by conversion to the more stable 2-pyrazolines in the presence of added base. Isolation of the 2-pyrazoline at the expense of the cyclopropane is evidence that the 1--pyra-, zoline is an intermediate in formation of the cyclopropane.




R C/ N N -N R

R/ R R> Y


base //N _N-H

/ R
x

Accordingly, the salt, XXXVI, was photolyzed in the pre-sence of dimethylfumarate and a tenfold excess of triethylamine. Nitrogen evolution was in excess of 80% and was evidence, in itself, that little or no pyrazoline was being trapped. The reaction mixture was worked up in the usual manner to give a 16'0 yield of heptafulvalene and a 49%0 yield






34


of the spiro compound. No evidence of a 2-pyrazoline Was f ound. This does not el-iminatCe the possibility of the diazo-pyrazoline route to XLI; but if the pyrazoline is actually an intermediate, it is apparently rapidly decomposed under the reaction conditions.

Another likely route involves decomposition of the sodium salt to an unstable diazocycloheptatriene which rapidly loses nitrogen to form the carbene, or alternatively, concerted decomposition of the salt to give cycloheptatrienylidene directly, which in either case, adds to the electron-deficient double bond to form XLI.




-N2-7:, XLI
=N--N -.SO 2Ar -NaSO 2Ar'






-NaSO Ar- I 2 -N2
2 (



There has been a variety of debate on the assignment

of multiplicities to carbenes on the basis of sterospecificity or nonsterospecificity of their addition to olefins (1). Skell has pointed out that a singlet carbene can undergo concerted (and therefore, stereospecific) addition Without violating spin conservation principles (57). A triplet, however, cannot undergo addition without a spin-inversion pro-cess and may proceed at first to a diradical species such as





35


XLIX. If the spin-inversion process is slow co-mpared to rotation about the carbon-carbon single bond, then the end





Rzzz R/C c 1) S. R
RI1 1 2) ring R

XLIX closure

result would be a nonstereospecific addition. Consequently, nonstereospecificity is generally considered to indicate triplet carbene, and stereospecificity is assigned to singlet carbene. This argument is not without weaknesses, however, for Gaspar and Hammond (58) have warned that since the relative rates of spin inversion and carbon-carbon bond rotation are not known, it is quite conceivable that the latter might be the slower process. Conversely, there is no assur-ance that a singlet carbene will add in a concerted fashion just because there is spin conservation. It is possible, therefore, to envision a stereospecific triplet additio-n anid a nonstereospecific singlet addition.

In the case of cycloheptatrienylidene, there is a good possibility that singlet addition to an olefin could proceed through a zwitterion such as L to give nonstereospecific addition products. The addition process has been shown to be stereospecific, however; so if L is indeed an intermediate, then ring closure must be faster than rotation about. the single bond.





36


R R R:+ I=:- ~ ~ + -R- Produc ts



L


It is generally conceded that carbenes are generated in a singlet spin state, although this is not necessarily true (59). The singlet carbene then either reacts as a singlet or undergoes intersystem crossing if there is a more stable triplet state available. In cases of this type, added inert solvent should decrease the overall degree of stereo-specificity by providing more chances for collisional deactivation to the more stable triplet. M4. Jones and K. R. Rettig have studied a case where this apparently occurs (60). They irradiated 9-diazofluorene in cis-2--butene and observed the cis/trans ratio of the spiro products as the reaction was diluted with hexafluorobenzene. They observed a rapid decrease in this ratio with dilution, which the y attributed to an increase in intersystem crossing to produce the triplet which then added nonstereospecifically to the cis-2-butene. Addition of 50 mole% of hexafluorobenzene was sufficient to decrease the cis/trans ratio to 0.6. In an analogous case, Jones and coworkers irradiated dimethyl diazomalonate in the presence of an inert solvent and observed the degree of stereospecificity of addition of the carbene to cis-4-methyl-2pentcne (61) Above 95 moleO of hexafluorobenzene, they





37


observed a sharp decreaSe in the amount of cis product,w ,hch they again attributed to triplet production. Therefore, from the two examples above, it might be expected that carbenes with triplet ground states can undergo intersystemn crossing from the initially produced singlet under conditions of high dilution.

Considering tetrahydrofuran as an inert solvent in the photolytic decomposition of the salt with the cis olefin, maleinitrile, the mole'O of the inert solvent can be calculated as 99.3 %. Since the reaction is still highly stereoselective at this point, it can be assumed that the carbene is reacting in its ground state or that the transition to the ground state is strongly forbidden. Since the addition process is stereospecific, any transition state for the addition must not allow rotation before ring closure. At the same time, a species such as the zwitterion L, should be stable and could easily arise in the addition process. Perhaps a combination of these can produce a logical intermediate. A transition state involving much greater bonding on one side of the forming three-membered ring, such as LI, would still incorporate stereospecific addition.




I S~R


R
LI





38


This "nonsynchronous concerted addition" would correspond to addition of a 1,1-dipole to an olefin.

Low-temperature mr studies of cycloheptatriene have shown that the ring inversion process can be stopped at

-150O0 indicating an activation energy of 6.3 Kcal/mole (62). The spiro compound, XLI, could have a lower E because of the spiro three-meinbered ring. The C. IC 1 C2 angle in the sevenmembered ring of XLI would be greater than the corresponding angle in cyc"loheptatriene. In the limiting case, where E a =0, the seven-memtbered ring of XLI would be planar. The low-temperature spectrum was examined from -30O0 to -95O0 and as expected, there was no change. Unfortunately, lower temperatures could not be obtained.











The spirononatriene, XLI, did not undergo Diels-Alder reactions with common dienophiles as do some derivatives of cycloheptatriene (63). A potential route to a spiropentene involving a Diels-Alder addition to XLI. was therefore precluded (64).

E. Other Olefins as Cycloheptatrienylidene Acceptors.

The list of electrophilic olefins tried as trapping

agents is quite short. The presence of strongly electronwithdrawing goups on the olefin makes it gjuite unstable to







39


basic conditions which are necessary for the generation of the carbene. In a number of cases where a tenfold excess of the electron-deficient olefin was used, not a trace could be found at the end of the photolysis. Perhaps a number of the olefins listed would easily form adducts with the carbene if it could be generated under less basic conditions. Maleic anhydride, fumarainaide, mnethyl cinnamate, dimethyl acetylenedicarboxylate, methyl -acetylacrylate, and the mixed methyl ester-dimethylamide of furnaric acid were all tried, but none gave evidence of a Spiro compound. It should be mentioned, however, that in each case except fumararide, methylcinnamate, and dimethyl acetylenedicarboxylate, there was no olef in at the end of the photolysis. This could mean that the olefin completely decomposed before the photolysis was well under way or alternatively, the spiro adduct, like the olefin, was unstable to the reaction conditions.

The addition of cycloheptatrienylidene to olef ins suggests a potential source of Em,63 spirarenes (65,66). The spirarenes consist of two orthogonal pi systems joined by a common Spiro carbon in which m and n represent the p orbitals in each ring. There is no easy synthetic route to these compounds, and only a few have been made. Cycloheptatrienylidene would form one of the pi systems, and it would become






40


attached to an olef in by a spiro carbon. If suitable acceptors could be found, then [26 f,46 and [6.6] spirareries could be synthesized.

The [2.6] spirarene system involves perhaps the easiest synthesis to visualize. Addition of the carbene to an acetylene to give the spiro [2.6] nonatetraene, LII, would be a direct route. Dirnethyl acetylenedicarboxylate was tried as a trapping agent, but no spiro compound was isolated. There





COOCH


+

COOCH 3 C O 3

LII


Br Br


Br Br ~acetoneCD



LIII LIV



Cl C1base


COOCH 3 COOCH 3 COC3


LV LVI


LV





41


was no heptafulvalene produced in this reaction which may indicate formation of the spirarene did occur, but the compound subsequently decomposed. An indirect synthesis could involve addition to a halogen-containing olefin follov,ed by dehalogenation or dehydrohalogenation. 1, 2-Dibromoethylene was tried with the hope that the adduct, L=T, could be debrominated to LIV, but again no spiro compound was obtained. lieptafulvalene was the only product in a yield of 44%. A third attempt at formation of a C26 spirarene involved methyl trans- -chloroacrylate. Dehydrochiorination of the adduct, LV, would give the desired system, LVI, but here also, the adduct was not detected, and the dimer (39%) was the only product.

Synthesis of the remaining two spirarenes is not as

easy to imagine. A possible route to the [4.6] system lies in a 1,4 cycloaddition reaction between the carbene and a diene followed by conversion of -the resulting cyclopentene ring to a cyclopentadiene ring to give LVII. The addition process would be the reverse of several symmetry-allowed fragmentation reactions studied by Lemal (67,68) Work is still in progress on this system.


R R R





R R R


LVI I






42


An interesting entry )-nto the tf6.61 spiracene system could be provided by benzoquinone. Conversion of the ene-dione, LVEI, to the di'ene would be followed by conversion to the triene, LIX, by analogy with the norcaradiene-cycloheptatriene system. Unfortunately, the instability of benzocjuinone to base precluded this reactLion.


0 0






LVIII LIX

F. Decomp-osition of_2,4,6-Cyclohentatrienone p-Toluenesulfonyihydrazone Sodium Salt under Hicrh Vacuum.

A convenient method of preparing some diazo compounds involves pyrolysis of the dry sodium salt of the p--toluenesulfonylhydrazone of the corresponding ketone under high vacuum (69). As the diazo compound is formed1 it distills out of the reaction mixture into a cold trap in yields up to 95%. If the analogous pyrolysis of XXXVI would give diazo-cycloheptatriene or possibly even cycloheptatrienylidene, perhaps it could be trapped at low temperature. The reaction vessel employed was a 500 ml boiling flask fitted with a cold finger.

The salt, XXXVI, was generated from the p-toluenesul-,

fonylhydrazone, XXXV, with sodium hydride in tetrahydrofuran.

'Removal of the solvent in vacuo gave the salt as a brow, n, muddyv solid which still contained traces of tetrahydrofuran.






43


After heating t1-o 50 0overniuht at ca. 10- 4 a o un hr still remained appreciable quantities of solvent. Since higher temperatures led to slow decomposition of the salt, it was used without further purification. Best results were obtained when the solid was distributed as a thin, uniform film over the inner surface of the reaction flask.

The decomposition could be carried out photochemically or thermally to give the same results. Phtolysis of the solid was generally much slower, however, anda only the outer Layer of salt decomposed. In contrast, the thermal

reaction was much smoother. When the temperature was gradually increased to 150 0, the decomposition proceeded slowly as evidenced by a gradual increase in pressure as nitrogen was evolved. If the reaction vessel. was quickly lowered into a bath at 1500, a rapid increase in pressure was noted, but here also, the reaction pro-ceded aut smotl wt little or no "bumping". Decomposition was complete within 2-3 sec in this case.

When a dry--ice-acetone cold finger was used, either rapid or slow decomposition of the salt gave at first a white coating of tetrahydrofuran. As the decomposition proceeded, the cold finger became discolored so that at the end of the reaction, a smooth, deep red coating was observed. After admitting an atmosphere of dry nitrogen, the finger was allowed to warm to room temperature. There was *no color change during the entire warming period. Had the diazo copound been trapped, considerable distortion of the smooth~






44


surface would have been expected as nitrogen was evolved. Examination of the material on the fincer showed only hepta-fulvalene and tetrahyd.-rofuran. Sodium p-toluenesulfinate accounted for most of the solid residue in the reaction f lask.

The formation of hepatfulvalene probably proceeds

through the intermediacy of diazocycloheptatriene or cycloheptatrienylidene rather than by dimerization of the salt, XXXVI (vide supra) The heptafulvalene must be formed in one of three places; the dimerization process can occur 1) in the dry salt as the reactive intermediate is generated, 2) in the space between the residue and the cold finger, and 3) on the cold finger either before or after the "cooling process" occurs. The first possibility involves a heterogeneous reaction arid the second, a gas phase reaction. In the third case, if the carbene or the diazo compound condenses on the finger, there is the possibility that it may react while still "hot" or vibrationally excited, or that it may "cool" somewhat before reaction. If the latter case actually occurs, then trapping the reactive intermediate at lower temperatures may prevent subsequent reaction. With this in mind, the reaction was repeated using liquid nitrogen in the cold finger.

Slow decomposition produced a smooth, light brown coat on the finger. After nitrogen was bled into the system, it was allowed to slowly warm. Soon after the liquid nitrogen in the finger evaporated (ca. -150 0) no-ticeable darkening






45


of the smooth coat was observed, and again only heptafulvalene and solvent were found. iYqo gas- evolution could be detected on the cold finger which might indicate decomposition of a diazo intermediate, but this does not preclude this possibility. A reasonable explanation, however, is that the

carbene actually con-densed on the finger and did not dimerize until warmed somewhat. The dimerization Drocess would cause the darkening observed.

Several attempts were made to trap the reactive intermediate with dimethyl fumarate. The olefin was sublimed onto the -finger through small orifices during the entire decomposition to provide intimate contact with the intermediate. Although darkening of the finger was again observed as the system was warmed, no spiro compound could be found. This was not entirely unexpected, for at the low temperature, the dimerization process may well be much faster than addition to the olefin. A possible alternative involves the use of an olefin which is a liquid at dry-ice temperature, so that as the intermediate comes in contact with the finger, it can dissolve and undergo addition in the solutiLon phase rather than the solid phase as in the case of dimethyl fu.Tarate. 1,2-Difluoroethylene and 1,l,l,4,4,4-hexafluoro-2butene are electron-de-ficient olefins which are being considered. Further work in this direction is anticipated.














CHAPTER III


THERMAL REARR~ANGEMENIT OF
8, 9-DICARLOMETHOXYSPIR~O 2.6j NONA-2 ,4, 6-TRIENE TO
1, 2-DICARDOMETHOXYINDANE



During the search for a second isomer of the spirononatriene, a sample was injected into a gic. on a 1/8" X 5' 20% SL30 on 60/80 Gas ChromZ column, a sample of XLI gave two peaks with approximate retention times of 8 and 10 min. The ratio of the peak areas in order of their elution was 0.052:1.00, and repeated recrystallization of the Spiro compound produced no change in this ratio. If these two peaks were due to cis and trans XLI, some change in this ratio would then be expected.

When a solution of XLI was treated with acid, rearrangement to the heptafulvene occurred. A sample of this solution showed only a single peak due to the rearrangement product. Both of the above two peaks had disappeared indicating that both were unstable to acid. A possible explanation for the two peaks lay in the isomerization of the pure isomer of XLI to a mixture of cis and trans isomers of XLI in the injection port of the g1c. This would explain why recrystallization failed to change the area ratio. Isolation of the two comnponents by preparative scale glc was the next obvious step.


46





47


Reasonable separat-ion of the two peaks was obtained

on a 1/4" X 5' columi-n of the material described previously at 150 0 and 50 psi. From the larger of the two peaks, a white crystalline solid was obtained with mp 70-72 0. Characteristics of the compound were completely different from XLI, arid on the basis of the physical properties given below, it was assigned the structure of trans-l,2-dicarb-omnethoxyindanie, M~a.

The nmr spectrum showed an aromatic multiplet of four protons at -T2.70, a doublet comprising one proton at T5S.48, and a multiplet of three protons between 15.85 and 6.85 partially obscured by nonequivalent methyl ester peaks at T6.17 and 6.25. The mass spectrum was identical to that of XLI, and it is very likely that XLI underwent the same thermal rearrangement in the inlet of the mass spectrometer. The structure of LXa was confirmed by comparison with an authentic sample synthesized according to Cook and Stephenson

(70).

COOCHI3 COOCH3
if COOCH 3

COOCH 3I



LXa LXb

A small quantity of the minor peak was obtained as a 50:50 mixture with LXa and appeared to be similar since ir and uv of the mixture were identical to those of the trans





48


indane. Mass spectra of the two peaks o.'o-tained ais they were elutea directly from the column into the mass spectrometer were also superimposable, so the minor peak was assigned the structure of the cis indane, LXb.

This provides a logical explanation for the behavior

of XLUI on gic analysis. Apparently, pure trans XLI rearranges thermally in the injector to a 95:5 mixture of the trans and cis indane. This isomerization can also be followed conveniently by high-temperature nmr. At 95 0, XLI is apparently stable indefinitely in heptane, but at 131 0 in n-decane, the reaction is 90% complete in 3.5 hr.

The mechanism of this reaction is a source for interesting conjecture. A logical sequence of events would involve rupture of the three-ruembered ring and reclosure to form a new ring at a different site either concerted or in a stepwise fashion with concomitant or subsequent aromatization.




-~COOCH 3 COOCH 3

XLI +k> COOC-13 ()CO OCH 3


L



CO C OO3 CHO H


LX -<- COOCHI

HC


LXI






49


Such a sequence could involve an intLermediate such as LX1 which could aromatize by a concerted hydride shift-bond migration step.

The formation of LXI would involve bond migration from C 1 to C7 and at first appears to he an interesting example of a 1,7-sigmatropic rearrangement. However, a suprafacial 1,7 shift is symmetry forbidden by a concerted thermal process involving a symmetrical orbital on the migrating atom

(71). J. A. Berson (72) has shown that an antisymnmetric orbital can be involved in a thermal, suprafacial 1,3 migration, and the same should be true for a 1,7 shift. This results in inversion of configuration at the migrating atom.




EXE+






If the thermal conversion of XLI to LXI could occur via a sigmatropic shift, it must be either antarafacial or an inversion process at the migrating atom. Examination of a molecular model of XLI makes it quite obvious that an antarafacial process would be essentially impossible. The second alternative would correspond physically to a 1,2 shift with inversion of configuration,,which is highly unlikely.

An alternative route involves the zwitterion, L, which would form by a heterolytic cleavage of a cyclopropyl bond.






50


This would be a relatively stable species since the positive charge i s. stabilized by trop--yl1ium- type resonance and -Ehe negative charge by resonance with the adjacent carbonyl. Attack by the anion on C 2of the tropylium system would lead directly to LXI. This intermediate would also be expected to give nonstereospecific products which could explain the mixture of cis and trans indane from the apparently pure isomer of XLI. It should be recalled, however, thatL this zwitterion was also mentioned as a possible intermediate in the addition of cycloheptatrienylidene to dimetzhyl fumarate. Although it cannot be ruled out in either case, L cannot be an intermediate in both schemes.

Although a 1,7 sigmatropic shift in the transition from XLI to LXI is forbidden in a concerted thermal process, it is symmetry allowed in a concerted photochemical process (71). Thus, it appeared possible that photolysis of the spiro compound at a wave length lower than that used to generate it might result in the formation of LXI which at. room temperature might xiot undergo further thermal reaction to LX. Accordingly, XLI was photolyzed in a quartz vessel in benzene solution. After 3.5 hr of photolysis with an immersion lamp, the solvent was removed to give an oily residue. The nmr spectrum of the oil showed decomposition of most of the. spiro compound, but no significant new absorptions attributable to a species such as LXI were found nor was the spectrum of LX observed.

If L is indeed an intermediate in the conversion, then

its ionic character might cause a rat-e increase in more polar






51


sol vents. Samples of XLI in triglyme and n-decane were placed in an oil bath at 150 0 and .removed at five-minute intervals to be examined by nmr. The approximate half-lives for formation of indane in n-decane and triglyme respectively were 20 min and 8 min This corresponds to a rate increase of 2.5 in the more polar solvent and is not inconsistent with the above mechanism. Careful examination of the spectra from these two reactions revealed no transient species which might indicate a preequilibrium of XL. with a species such as LXI.

The dicyanospirononatrienes, XLVIa and XLVIb, are pre-sently being examined for an analogous thermal rearrangement.














CHAPTER IV


ADDITIONAL SCHEMES EXAIMNIED AS
ROUTES TO CYCLOHEPTATRIEDIYLIDENE



A. Base-promr.oted Decomposition of M~ethyl N--Nitroso--N-cycloheptatri eny carbamate and_N,N-- Dimethyl --N '-nitroso --N '--cyciloheptatrienlra

Another route to diazocycloheptatriene or cycloheptatr ienylidene which w.as explored was base-induced decomposition of nitrosoureas and nitrosocarbamates. Previous work in this laboratory (31,33) had shown that treatment of N,,,Ndirnethyl-N'-nitroso--N' (2,3--diphenyl-2-cyclopro-penyl)urea, LXII, with potassium t-butoxide in the presence of diimethylfumarate gave the spiropentene, LXIV. M~ethyi N-nitroso-N(2,3-diphenyl-2-cyiclopro-enyl)carbamate, LXIII, also proved tc be a good source of LXIV. By analogy, the corresponding[ N-cycloheptatrienyl compounds should provide a route to XLI.




Ph VO1,Ph COOCH 3

N-C -B
baseI


Ph Ph COOCH 3LXII, B =N(CH 3) 2 LXIV

LXIII, B =OCH 3


52





53


The 'areas and carbamates are readily synthesized by

treatment of the corresponding isocyanatCe with an amine or an alcohol. Accordingly, cycloh-epta-2,4,6-trie-n-lyiSoDCYanat'-e was synthesized by addition of potassiumn cyana-te to an aqjueous solution of tCropyliun bromnide (73) and by rearrangement of the acid azide under anhydrous conditions (74). Addition of diethylanine or methanol. to a solution of the iso-cyanate gave the urea, LXV, and the carbamate, LXVI, res-pectively.



N C __B-H


LXV, B N N(CH 3) 2

LXVI, B 0CH13


Nitrosation of the carbamate w,.as carried out-- at -20 0 by adding a cold ether solution of dinitI-.rogen tetroxide until the green color persisted for 5 ini. Addition of' a large excess of the dinitrocosa tetroxi-de solution led to markedly decreased yields. After washing tihe reaction mixture with base and drying, a dark yellow oil remained. This o-4l was passed through a column of deactivated silica cfe! to Cr4 ve a pale yellow solution. Concentration under a stream of dry nitrogen afforded relatively pure nitrosocarbamate, LXVII, as a yellow oil. Akt room teinperat ure, the oil- decomposed
0
slowly with bi'bbling, but at 0 under nitrogen it was apparently stable indefinitely.






54


LXVI N2 C -4-OCI



LXVI I




Stirring the nitrosourea in hex--ane with sodium methoxide and several drops of methanol led to decomposition with quantitative gas evolution. Chromatography of the reaction mixture on deactivated basic alumina gave heptafulvalene

(7%), but no other major product. When dimethyl fumarate was added to the reaction mixture, the Spiro adduct, XLI, was obtained in 4% yield.




spirononatriene

NaOCH3

LXVII


NaOCF
heptafulvalene


The major product of this reaction was shown to be methoxysuccinate, LXVIII, which results from addition of methoxide across the double bond of the olef in. This provides confirmation of the mode of isomerization of the cis olefin to the trans in the presence of a base. A stirred solution of dimethylfurnarate and sodium methoxide in mnethanol gave rise to the samec products in good yield.






55)


CHO OO ,COOCOCI CO G~
CH HOC /CC\H 3


LXVIII


The use of potassium t-butoxide as the base in the nitr,-socarbamate decomposition gave slightly higher yields. Heptafulvalene was obtained in 13% yield without dirnethyl furna-. rate, and with the olef in, the yield of spiro adduct was

4.5%. The product of base addition across the olefinic double bond, t-butoxysuccinate, was not observed presumably because of the bulky size of the base.

The N,N-dimethylurea, LXV, was nitrosated as in the case of the carbamate, but the product was quite unstable and rapidly darkened during the workup. W 7hen the conditions of the workiir- were modified so that the nitrosourea, LXIX,


NO9

LXV NO 2NL4'N (CH3) 2


LMIX


was never allowed to warm to 0 0 reasonable yields were obtained. Low temperature nmr ofl the unstable compound confirmed its structure by analogy with the nitrosocarbamate, but as the sample was warmed to 10 0, the spectrum disappeared. Among the peaks which appeared in the spectrum above 100,





56


were those of2 N,N-dimethylcycloheptatrienylamine. This was confiJrmed by synthesizing a sample with dimethyl amine and tropyliun bromide.

In the case of the diphenylcyclopropenylnitrosourea,

LXX, decomposition without base led to the carbamate, LXXI, which also gave rise to the spiropentene via an a-eliminaLion when treated with potassium t-butoxide.

An attempt was made to convert LXUX into the corresponding carbam ate, but dimethylcycloheptatricnylamine was the only identifiable product. Apparently carbon dioxide as well as nitrogen is easily lost from the cycloheptatrienyl nitrosourea.





LXX PhA> 0 -N (CH 3)2
2 Ph'


LXXI




MXIX N(H3
-N CO
2 2






Treatment of a cold solution of the nitrosourea with sodium methoxide and a few drops of methanol gave besides the amine, a 2% yield of heptafulvalene. In a similar reaction with dimethyl fumarate present, a 0.5% yield of the






57


spirononatriene was found. These yields are based on starting urea and not on the nitrosourea. Since neither of the above routes appeared promising, they were not pursued further.

B. a-Eliminat'ion Schemes.

A nu-mber of other methods were examined as possiLble

routes to cycloheptatrienylidene. Since it was highly desirable to find a route which did not involve the possible intermediacy of diazocycicheptatriene, several u-eliminations were considered. u-Dehalogenations can frequently be effected by strong bases or metals (1,2) Accordingly, chiorot1-ropylium chloride was treated with electrolytic copper dust in tetrahydrofuran. After stirring overnight, the mixture had turned cloudy, but examination by uv and tic showed the dimer, heptafulvalene, had not formed. Similar treatment with sodium metal followed by aqueous workup gave no trorpone indicating a reaction with the chloride had taken place. once again however, rno evidence of heptafulvalene was found. The chloride was also treated with n-butyllithium and tbutyllithium with similar results.

Previous work in this laboratory had shown that treatment of diphenylcyclocpropenyl acetates, benzoates, and --nitrobenzoates with strong base in the presence of olefins gave products which indicated generation of the carbene (33). Cycloheptatrienyl acetate was synthesized (75) and treated with potassium t-butoxide in.the presence of dimethyl fumarate. After refluxing-- for two days in heptane, the mixture









was examined by gic ancd tic. No evidence was found to iridi-cate that either heptafulvalene or the spiro compound, XLI, was formed. Similar results were observed when a solution of the acetate and dirrethyl funmarate were treated with t-butyllithium at -50 0. Since this ai-elimination route did not appear promising, it was not pursued further.

Another method concerned decarboxylation of a carboxy-tropylium salt to give the carbene directly. Carboxytropy~hium bromide has been reported and was found to be quite unstable (76). Since fluoroborate salts are in general much more stable than the corresponding bromides, a synthesis of carboxytropylium f luoroborate was undertaken. A solution of 2,4,6-cycloheptatriene carboxylic acid in dichloromethane was added dropwise to a solution of trityl fluoroborate in dichlorornethane at room temperature. A light brown solid formed, and after thirty minutes of stirring, it was filtered and dried. The fluoroborate charred at 150 0 and melted with gas evolution at 165 0. Recrystallization from nitromethane-benzene gave white plates which charred at 153 0 and melted with gas evolution at 1650 The ir spectrum showed a carbonyl peak at 1722 cm 1, and an absorption at 272 mp was found in the uv spectrum. The nmr spectrum in nitromethane showed a three-proton multiplet at T2.0-2.6 and a broad singlet at T2.80. The analysis indicated the empirical formula as C H7 0 2EF V This salt was quite stable at room temperature under nitrogen as opposed to the behavior of the corresponding bromide.













CO 2 H Ph3 4.F


BF 4


+ Ph 3-C


LXXII


Further confirmation of the structure of LXXII was

obtained when the salt was reduced with lithium aluminumtri(t-butoxy)hydride. A mixture of three isomneric cycloheptatriene carboxylic acids was obtained in 50% yield. 2,4,6-Cyclohentatriene carboxylic acid was not found and was not an expected product since it would involve hydride attack at the most hindered position of the tropyliun ring. 1,3,5-Cvcloheptatr'iene 1-carboxylic acid, LXXIII, composed 35% while the other two isomeric acids made up the remaining 65%.


H COH


35%


LXXIII


LiAl (Ot1Bu) 3H


H H H01


- 65%


59





60


If the carboxylic acid proton could be removed from

LXXII, decarboxylation should lead directly to cyclohLeptatrienylidene. The acid-salt was treated with a number of bases including lithium carbonate, sodium hydride, triethylamine, and lithium metal. In each case, reaction occurred





Q C0011- + o COO 0




but the product was normally a gum-my oil which resisted all efforts at characterization. Further work is anticipated in this system.














CHAPTER V


SUMIRY



The base-promoted decomposition of 2,4,6-cycloheptatrienone p-toluenesulfonylhiydrazone has been investigated

as a source of cycloheptatrienyliderie. Either photolytically or thermally, the decoinposi tion gives up to 70% of the un-stable dimer, heptafulvalene, when no acceptors are present. In the presence of dimethyl fumarate, an electron-deficient olefLin, 8,9-dicarbometh-ox,.yspir- oC2.6J nona-2,4,6-trienc can be isolated in 50% yield. This spirononatriene is the adduct expected if the carbene adds to the olefin. With fumaronitrule, an analogous spirononatriene is formed. The cis isomer of furnaronitr-ile, maleinitrile, gives the cis adduct indicating that the addition process is highly stereoselective if not stereospecific. This stereospecificity at high dilution is given as evidence that the carbene is adding in a singlet state which may be the ground state of the species. Evidence is presented against the diazo compound as the reactive intermediate in the formation of the spirononatrienes.

The spirononat-rieies were found to undergo a facile,

acid--catalyzed rearrangement to the corresponding hept.aful-


61







62


venes. Above 130 0, the adducts rearrange to 1,2-disubs Li-tuted indanes, anid an interesting mechanism is postulated to explain this thermal isomerization.

When electron-rich double bonds are used as acceptors,

only the dimer is formed. Thus, it appears that the reactive intermediate is nucleophilic as well as stereospecific in its reaction with olef ins.

Base-catalyzed decomposition of methyl N-cycloheptatrienyl-N-nitrosocarbamate and N ,N-dimethyl-N' -nitroso-N cycloheptatrienylurea was also investigated as a carbene source. The spirononatriene is formed when dimethyl fumarate is present, and heptafulvalene is a product in the absence of the olefin; but in each case, the yields are quite low, and side reactions predominate.














CHAP TLR VI


EXPE RIIENTAL



A. General. LMelting points were taken in a Thomas Hoover Unimelt apparatus and are uncorrected. Elementa. analyses were performed by Galbraith Laboratories, Inc., Knoxville, Tennessee. Ultraviolet spectra were recorded on a Cary 14, or Gary 15 double-beam spectrophotometer using 1 cm. silica cells. Infrared spectra were recorded with a Beckmann Model IR10 spectrophotometer. In all cases where the K13r technique was not used, sodium chloride plates were substituted. Nuclear magnetic resonance spectra were determined on a Varian A-60A hich resolution spectrometer with a variable temperature probe. Chemical shifts are reported in tau values from internal tetramethy lsilane standard. Mass spectra were determined on a Hitachi model R:'Um-6E Mass Spectrometer.

Analytical thin layer chromatography (tlc) was accomplished on 2" X 8'; plates coated in these laboratories with 0.25 wu'm layers of E. Merck HF 254 silica gel; compo-nents were visualized by their quenching of fluorescence under ultraviolet light. Gas-liquid chromatography (glc) was conducted on a Wilkens Aerocjraph Model. 600-B, or 600-D

Hi-Fi instruments, with flaiae ionization detectors using


63






64


nitrogen carrier gas; preparative scale work was accomplished on a Vvilkens 1Aer-ographn Dual Column Temperature Programmer Gas Chromatograph, using helium carrier gas.

All photolvses were carried out with a H1anovia, 450 watt, high-pressure mercury immersion lamp, and unless otherwise noted, a pyrex vessel was used. Tetrahydrofuran for the photolysis reactions was dried by distillation from lithium aluminum hydride. The sodium hydride was obtained from Alfa Inorganics, Inc. It was weighed as a dispersion, washed three times with pentane, and used as a powder for each reaction. All technical hydrocarbon solveths were distilled before use.

The cycloheptatriene used in the following reactions was generously donated by Shell Chemical Co. and contained 9% toluene.

B. Chlorocycloheptatrienylium Chloride. This compound has been reported (47) and was prepared as described with the following exception: thionyl chloride was used as solvent and halogenating agent. Tropone was added dropwise to an excess of thionyl chloride at 0 0 (reverse addition caused vigorous decomposition of the sample with no production of the desired product). At the end of the addition, the solution was refluxed gently for 5 ruin on a steam bath. Removal of the excess thionyl chloride on a rotary evaporator gave the chloride as a yellow crystalline solid which was used without further purification.






65


C. 2,4,6-Cycioheptatrienone p-Toluenesulfonylhydrazone Hydrochloride. A solution of 450 mg (2.42 mmoles) of ptoluenesulfonylhydrazine in 5 ml of absolute ethanol was added rapidly with stirring to a solution of 304 mg (2.45 rnmoles) of chiorotropylium chloride in 5 ml of absolute ethanol. The dark red solution was rapidly stirred at room temperature for 30 min. The yellow solid which had formed was filtered, washed with ether, and dried in vacuo to give 600 mg (90%) of the salt. Three recrystallizations from methanol-ether gave an analytical sample, mp 170 0 d (st), liquefaction and gas evolution at 210 0 Vma K B318, 2930, 2700 (broad) 1638, 1595, 1522, 1498, 1450, 1348, 1254, 1233, 1190, 1165, 1090, 1030, 905, S33, 812, 766, 710, 643, 583, 555, 546, 442, and 37-3 cm ;1 Et, 229 (El13200) and 315 m Q~ 15700); TD02.56 (complex pattern, 10 H, aromatic and cycloheptatriene ring protons) and 7.73 (singlet,

3 H, methyl protons); rne ( 70 ev) 274, 156, 155, 139, 119, (base peak), 107, 91, 90, 89, 78, 77, 65, 53, 51, 50, 39, 38, and 36; metastable peaks, 46.5 and 60.5.

Anal. Calcd for C14H15N 202 SC1 C, 54.10; H, 4.83; N, 9.03. Found: C, 54.37; H, 5.03i N, 9.26. D. 2 4,6-Cycloheptatrienone p-To luenesulfonyihydra zone. To a vigorously stirred mixture of 50 ml of 10% aqueous NaHCO 3 solution and 25 ml of dichioromethane was added 1.00 g (3.22 mmoles) of the hydrochloride salt. After 15.mm, gas evolution had ceased and the two layers were separated. The aqueous layer was washed twice with 10 ml portions of






66


dichiorome thane. The organic fractions were combined, dried over magnesium sulfa Le, and concentrated on a rotary evapo-rator to give the p-toluenesulfonyihydrazone as a dark red solid. Recrystallization from benzene-pentane afforded 850 mg (96%) of XXXV. The p-toluenesulfonylhydrazone has two distinct crystalline forms: dark red plates, mp 142.5-143.5 0 and light red needles, mp 144-1.45. Both forms gave identical infrared spectra in nujol mulls. Several recrystallizations from benzene--pentane gave an analytical sample; V max 3240, 2038, 1603, 1567, 1471, 1393, 1304, 1310, 1228, 12.87, 1170, 1092, 1042, 1003, 920, 864, 813, 743, 700, 660, 565, 547 C--1 X EtOH 245 (c7900) and 315 mpQ13100); 1CDC!
max TI
1.84-2.70 (A 2 B2, 4 H, aromatic protons), 3.33-3.83 (complex pattern, 6 H, cycloheptatriene ring protons) and 7.55 (singlet, 3 H, methyl protons); m/e (70 ev) 274 (molecular ion), 180, 179, 156, 155, 139, 119 (base peak), 107, 91, 90, 89, 78, 77, 65, 63, 61, 50, and 39.

Anal. Calcd for C1H14N 2 2S; C, 61.30; H, 5.11; N, 10.22. Found: C, 61.50, H, 5.13; N, 10.06. E. Photolysis of 2,4,6-Cycloheptatrienone o--Toluenesu2.fohvlhydrazone Sodium Salt. Formation of Hentafulvalene. In a typical reaction 0.500 g (1.82 mmoles) of the p-toluenesulfonyihydrazone was dissolved in 120 ml of dry tetrahydrofuran. The sodium salt was formed by addition of 50 mg (2.08 mmoles) of sodium hydride powder to the dark red solution. When gas evolution had ceased (15 min), the photolysis vessel was swept with a stream of dry nitrogen for 10 minutes to






67


remove hydrogen gas. Photolysis of the dark slurry was fol-lowed by nitrogen evolution which stopped at 35 ml (85%) after 3 hr. The cold finger was coated with a light brown solid which ir showed to be sodium p-toluenesulfinate. The red solution was poured into 300 ml of water and extracted with three 50 ml portions of pentane. The pentane extracts were combined, dried over magnesium sulfate, and concentrated to a crude black solid. This solid was taken up in 10 ml of 5% ether in hexane and chromatographed on 20 g of basic alumina (Woelrn, activity grade III) using 5% ether in hexane as eluent. The heptafulvalene was fluted rapidly as a sharp, dark red band. Uv determination of the yield as described below gave 107 mg (65%) of the dimer. Concentration of the solution gave a crude black crystalline solid which after several recrystallizations from methanol-water (pH 7-9) or sublimation gave permanganate-colored plates with mp 1191210 [lit. (52) rnp 12203 V maxr 290C, 1540, 1440, 1265, 975, 935, 898, 888, 835, 805, 790, 781, and 712 cm-1* .XEtOH max
234 (E 24000) and 362 my (c 25000) [lit. (52) X max 234. (c 22000) and 362 my K~ 21000)]; TM3409(tucue singlet); rn/c (70 ev) 180 (molecular ion and base peak), 165, 152, 139, 115, 102, 90, 89, 76, and 63; metastable peaks, 130, 68.8, and 56.8.

A satisfactory analysis was not obtained for this compound.

F. Determination of the Yield of Heotaful-valene-. Due to the losses incurred in handling heptafulvalene, the yields.






68

were determined by a spectroscopic method. The reaction mixtures were first chiromaatographed on basic alumina (ola activity grade III) using 5'," ether in hexane as eluent. Thisallowed rapid elution of the diner which decomposes slowly on the column, while still obtaining excellent separatCion from all other components of the reaction mixtures. The fractions which contained he-ptafulvalene were then comnbined and diluted accurately to give a solution estimated to be 10- 5 The exact concentration was then calculated from the uv absorption of the 362 mp peak. The extinction coefficient for this absorption is 21000 as reported by Doering (52). The e-xtinction coefficients calculated in these laboratories on three samples of heptafulvalene which were recrystallized three times from methanol-water (pHi 7-9) and triply sublimed, were 26000, 25000, and 21400. The last sample was left overnight under nitrogen and may have partially decomposed, so the value selected was 25000. G. Photolysis of 2,4,6-Cvcloheptatrienone p-Toluenesulfonylhydrazone Sodium Salt in the Presence ot Dimethyl Furnarate. Formation of 8,9-Dicarbomethoxyvsp~i.ro[2.63nona-2,4,6-triene. In a typical reaction 1.00g (3.65 rnmoles) of the p-toluenesulfonylhydrazone was dissolved in 120 ml of dry tetrahydrofuran. The sodium salt was formed by addition of 100 mg (4.15 nnrioles) of sodium hydride powder to the dark red solution with vigorous stirring. W, hen gas evolution had ceased (ca. 15 min), the photolysis vessel weas

swept with a stream of dr-y nitrogen to remove hydrogen aas.





69


To this slurry was added 2.50 g (17.4 mmoles) of dimethyl fumarate. After 4 hr of photolysis, the nitrogen evolution had stopped at 70 ml (85%). The light red solution was poured into 500 ml of water and extracted with three 200 ml portions of pentane. The pentane extracts were combined, dried over magnesium sulfate, and concentrated to a dark solid. This residue was taken up in 20 ml of 5% ether in hexane and chromatographed on 50 g of basic alumina (Woelm, activity grade MII) Heptafulvalene came off rapidly with 5% ether in hexane and the yield, as determined by uv, was 38.7 mg (10.5%). Gradually increasing the ether corncentration to 10% gave a mixture of dimethyl fumarate and the

spirononatriene compound. Fractional crystallization gave 814 mug of recovered dimethyl fumarate and 421 mug (49.6%) of XLI with mp 72-74 0. Two recrystallizations from methanolwater (pH 7-9) gave an analytical sample with rup 76.5-77.5; Vmax 3080, 3020, 3002, 2970, 1720, 1450, 1440, 1340, 1270, 1200, 1158, 1078, 1018, 895, 885, 872, 745, 705, 645, 590, 502, and 330 cm- 1 X EtOH 262 mp( 2800); T CDC13 3.203.83 (complex pattern, 4 H, 3,4,5, and 6 protons on cycloheptatriene ring), 4.55 (doublet, 2 H, 2 and 7 protons on cycloheptatriene ring), 6.31 (singlet, 6 H, ester methyls), and 7.69 (singlet, 2 H, cyclopropyl protons); in/e (70 ev) 234 (molecular ion) 203, 175, 174, 1.16, 115 (base peak) 91, 90, and 89; metastable peaks, 131 and 76.

Anal. Calcd for C1H1404 C, 66.66; H, 5.98. Found: C, 66.50; H, 5.93.





70


ii. Isomerization of S;,9--Dicarbomethoxyspiro L2.G1nona-2,4,6-tri-ene to 8 -- Carbome thoxy- c arb ome Ll-oxymethylme thyl1 enecy clo0hpta-2 ,4,6-triene on AcidAlumina. A solution of 42 -mg (0.18 iaxoles) of the spirononatriene in 5 ml of anhydrous ether was placed on a dry column of 5 g of acid alumina (Woeim, activity grade I). Upon contact, the column imtmediately turned bright yellow. The yellow. barnd was fluted with ace-tone and concentrated to a dark yellow oil on a rotary evaporator. This oil was was taken up in pentane and cooled in dry ice to give a fluffy yellow solid. Recrystallization from pentane gave 38 mag (91%) of the heptafulvene as yellow needles with rap 48-50 0. Two further recrystallizations gave an analytical sample with rap 49.5-510 v r 3020, 2960, 2910, 1750, 1705, 1642, 1570, 1442, 1355, max
1278, 1203, 1184, 1118, 1030, 822, 783, 744, 660, and 545 cm -1 X EO 245 (c 10000) and 340 mit (E 13000): r Cl3
max TMS
2.57 (poorly resolved doublet, 2 H, 2 and 7 protons on cycloheptatriene ring), 3.40-3.85 (multiplet, 4 H, 3,4,5, and 6 protons on cycloheptatrienie ring), 6.29 and 6.34 (singlets, 6 H1, ester methyls) and 6.65 (singlet, 2 H, methylene protons); m/e (70 ev) 234 (molecular ion), 203, 175 (base peak), 115, 91, 89, and 59; metastable peaks, 131 and 76.

Anal. Calcd for C 13 H, 4 0 A C, 66.66; H, 6.02. Found: C, 66.43; H, 5.88.

I. Photolysis of 2,4,6-Cycloheptatrienone p-Toluenesulffonyl.hydrazone Sodi-umSalt in the Presence of Fumaronitri-le. Formation of trans-8 9-Di-cyanospiroL2 .'- nona-2 4, 6-- triene.





71


To a solution oil 548 mug (2.0 rumoles) of the p-toluenesulfonyihydrazone in 120 ml of dry tetrahvdrofuran was added 62 mug (2.58 mmoles) of sodium hydride powder. Vigorous gas evolution occurred, and the sodium salt precipitated to form a dark brown slurry. When the reaction was complete (ca. 15 ruin) the reaction fLlask was flushed well with dry nitrogen to remove all hydrogen gas, and 780 mg (10.0 mmoles) of fumaronitrile was added. After 2.5 hr of photolysis, the nitrogen evolution had stopped at 42.5 ml (89%). The dark red solution was poured into 300 ml of water and extracted with three 100 ml portions of pentane. The pentane extracts were combined, dried over magnesium sulfate, and concentrated to a dark oily solid on a rotary evaporator. Tlc of the crude residue showed heptafulvalene as a dark spot aind a second colorless spot in 33% ether in pentane. The second

spot turned bright yellow after standing on the dry plate for 15 ruin, consistent with the behavior of the adduct of cycloheptat-1rienylidene and dimethyl fumarate. The nmr spectrum of the crude residue showed dinner, fumaronitrile, and a set of absorptions consistent with a spiro compound. The residue was taken up in 10% ether in hexane and chromatographed on 25 g of basic alumina (Woelm, activity grade III) using 10% ether in hexane as eluent. The dimer was eluted rapidly as a characteristic red band, and the yield was found to be 33 rug (18.5%). When the solvent polarity was increased to 20% ether in hexane, a white solid was obtained which was shown to be the adduct, XL~,97 rrug (29%) with




72


0
mp 110-114 Three recrystallizations from benzene-hexane gave an analytical sample with mp 121-123 0 v r 3030, max
2250, 1528, 1440, 1400, 1238, 1212, 1265, 1030, 970, 875, 830, 760, 705, 660, 625, 622, 485, and 450 cm- 1 X tI
max
262 mp (c 2760); TCDC13 3.10-3.60 (complex pattern, 4 H, TiAS
3,4,5, and 6 protons of cycloheotatriene ring), 4.53 (doublet,

2 H, 2 and 7 protons on cycloheptatriene ring) and 8.01 (singlet, 2 H, cyclopropyl protons) ; m/e (70 ev) 168 (molecular ion) 167, 141, 140, 128, 115, 90 (base peak), 89, 77, and 63; metastable peaks, 117.7 and 48.3.

Anal. Calcd for C 11H 8N 2:C; 78.55; H, 4.79; N, 16.65. Found: C, 78.82; H, 4.81; N, 16.70. J. Isornerization of trans-3,9-DicyanospiroL2.Gnona-2,4,6triene to 8-Cyano-8-cyanomethylmethylenecycloepta-2,4,6-triene on Acid Alumina. A solution of 20 mg (0.12 mnmoles) of the Spiro compound in 2 ml of anhydrous ether was added to a dry column of 5 g of acid alumina (Woelm, activity grade I) An additional 2 ml of anhydrous ether was added to wash the solution onto the column. The bright orange band which rapidly developed was then eluted with acetone to give a red solution. This solution was concentrated to 1 ml under a stream of dry nitrogen, and pentane was added to the cloud point. After cooling in the refrigerator for 1 hr, the red crystals of the heptafulvene were filtered and 00
from benzene-hexane gave red plates with mp 99.5- 100.50 V Kax 2255, 2195, 1640, 1543, 1520, 1474, 1408, 1272, 1160,






73


912, 062, 821, 760, 71, 580, 523, and 500 cm-' 1 AEtOH max

247 Q~ 9500) and 345 mo K~ 17000); CDC13 2.83-3.30 (unresolved multiplet, 2 H, 2 and 7 protons on cycloheptatriene ring), 3.50 (broad singlet, 4 H, 3,4,5, and 6 protons on cycloheptatrierie ring) and 6.72 (singlet, 2 H, methy:lene protons); rn/e (70 ev) 168 (molecular ion) 167, 141, 140, 128, 115, 90 (base peak) 89, 77, and 63; metastable peaks, 117.7, 92.5, 80.0, and 48.3.

Anal. Calcd for C 11H 8 H C, 78.55; H, 4.79; N, 16.65. Found: C, 78.43; H, 4.72; N, 16.51. K. Photolysis of 2,4,6-Cycloheptatrienone p-Toluenesulfonylhydrazone Sodium Salt in the Presence of Maleinitrile. Formation of cis-8,9-Dicyanospiro[2.6jnona-2,4,6-triene. This

reaction was carried out as described in section I, with one exception; maleinitrile was substituted for fumaronitrile. The yield of heptafulvalene was 17.6%. The spirononatriene was obtained in 30% yield and was assigned the cis structure on the basis of the physical characteristics given below. An analytical sample was obtained after several recrystallizations from ether-pentane with mp 11-1. ; ~ 3020,
max
2245, 1532, 1444, 1399, 1378, 1353, 1315, 1274, 1208, 1169, 1058, 1009, 988, 940, 888, 811, 755, 702, 668, 619, 558, 516, 444, and 372 cm- 1 X EtOH 262 p (c 2720); T CDC 13 3.10-3.35 max TMS
(complex pattern, 2 H, 4 and 5 protons on cycloheptatriene ring), 3.40-3.80 (complex pattern, 2 H, 3 and 6 protons on cycloheptatriene ring), 4.25-4.55 (broad doublet, 1 H, 2 proton on cycloheptatriene ring), 4.55-4.85 (broad doublet,





74


1 H, 7 proton on cycloheptatriene ring), and 7.97 (singlet, 2 H, cyclopropyl protons) ; m/e (70 ev) 168 (molecular ion), 167, 141, 140, 128, 115, 90 (base peak), 89, 77, and 63; metastable peaks, 117.7 and 48.3.

Anal. Caicd for C 11H 8N 2 C, 78.55; H, 4.79; N, 16.65. Found: C, 78.35; H, 4.77; N, 16.60. L. Photolysis of 2,4 ,6-Cvcloheptatrienone p-Tolueneaulfonyl-hydrazane Sodium Salt in the Presence of Other Acceptors.

In a typical reaction, 0.500 g (1.82 mmoles) of the p-toluenesulfonyihydrazone, XXXV, in 120 ml of dry tetrahydrofuran was converted to the sodium salt with 53 mg (2.20 inmoles) of sodium hydride powder. After gas evolution ceased, the reaction flask was flushed with nitrogen and a tenfold excess of the acceptor was added. The reaction was followed by nitrogen evolution and was normally complete within 3 hr. The crude reaction mixture was examined with tic and nmr for evidence of an adduct. The yield of heptafulvalene was determined as in the previous reactions and varied from 30-* 70%. No evidence of an adduct was found in the reaction with cis--2-butene, cyclohexene, N--piperidylcyclohexene, trans-pdinitrostilbene, 1,2-dibromoethylene, methyl trans-$-chloroacrylate, maleic anhydride, fumaramide, methyl cinnamate, dimethyl acetylenedicarboxylate, methyl P-acetylacrylate, and the mixed methyl ester-dimethylamide of fumaric acid. M. Thermal Decomposition of 2 ,4,6-Cycloheptatrienone p-Tolu-enesulfonylhydrazone SodiumSalt with and withoutDimethyl Furarate. Two 10 ml portions of dry diglyme (distilled from





75


lithium aluminum hydride) w~ere thermostated at 101.5 0 under nitrogen in a constant temperature bath containing refluxing dioxane. To one sample was added 288 mng (2.00 rumoles) of dimethyl fumarate. Employing a solid addition tube, 368 mug (1.00 mmoles) of the salt was added in each case. The rates

of both reactions were followed by nitrogen evolution and were virtually identical, giving half--lives of ca. 11 muin. Both solutions were evaporated to dryness in vacuo and worked up as usual. The yield of heptafulvalene in the reaction without dimethyl fumarate was 29.5 rmg (33%).- In the reaction with dimethyl fumarate, the yield of the spiro compound was found to be 90 mug (38%) while the heptafulvaleie had dropped to 8.0 rug (9%0).

N. Photolysis o-f HeT:tafulvalene in the presence of DiimCethy. Fumarate. A solution of 17.3 mug (0.096 runoles) of heptafulvalene and 69, mg (0.48 rumoles) of dimethyl fumarate in 50 ml of hexane was photolyzed with two Kenmore Sunlaimrps through a pyrex filter for 15 hr. At the end of this time, the colorless solution was concentrated and examined for products. Tlc, uv, and gio of the reaction mixture showed that no spirononatriene had been formed. Dimethyl fumarate was the only identifiable compound.

0. Photolysis of 2,4,6-Cycloheptatrienone p-Toluenesulfonvlhydrazone Sodium Salt in the Presence of Dimethyl Furnarate and Trie thj'lamirie. This reaction was identical to the decomposition described in section G ex-cept for the presence of

3.7 g (36.5 rumoles) of triethylamine. The nmr spectrum was









examined carefully for the presence of a 2-pyrazoline in the

reaction mixture, but no evidence for one was found. The yields of heptafuivalene and spirononatriene were 53 mg (16%) and 420 rag (49%) respectively. P. Thermal Decomposition of 2,4 ,6-Cycloheptatrienone p- Toluenesulfonylhvdrazone Sodium Salt under Vacuum. In a typical reaction, 250 mg (0.91 noles) of the p-toluenesulfonylhydrazone, XXXV, in 5 ml of tetrahydrofuran was converted to its sodium salt with 26 mg (1.1 mnioles) of sodium hydridae powder. As the solvent was removed on a rotary evaporator, the salt was distributed uniformly on the inside of the 500 ml flask. The flask was then fitted with a cold finger. The salt was dried at 50 0 and ca. 10- mm. of Hg overnight. Liquid nitrogen was added to the cold finger, and the reaction flask was
0
heated slowly with an oil bath to 150 .The finger was coated with a light red material at the end of this time. An atmosphere of dry nitrogen was bled into the system, and the liquid nitrogen was removed. As the finger began to warm (ca. -150 0), a noticeable darkening of the coating occurred. After warming to room temperature, the coating was examined arid found to consist of tetrahydrofuran and heptafulvalene, 16.4 rmg (20%).

Q. Thermal Decomposition of 2,4,6-Cycloheptatrienone o-Toluenesulfonylhydrazonie Sodium Salt under Vacuum with Dimethy l Fumarate. A uniform coating of the sodium salt was prepared from 250 mg (0.91 mrnoles) of the p-toluenesulfonylhydrazone and 26 mg (1.1 rumoles) of sodium hydride as described above.







-7

The reaction flask had been modi fi-ed by adding two small tubes pointed at the cold finger through whi-ch dimethyl fumarate could be sublimed. The decomposition was also carried out as described above with the fuma-rate subliming onto the finger during the entire reaction. Again a light red coat was cobserved and again a noticeable darkening occurred. When the material on the finger was examined after the reaction, with nmr, glc, and tlc, only heptafulvalene, tetrahydrofuran, and

dimethyl fumarate were foundI. No trace of a spirononatriene was found.
R. Thermal Rearrange-men-t of 8, 9-D-icarho-metho xvsio[,J nona-2,4,6-triene to 1,2-Dicarbomethoxy;indane. A solution of 100 mg (0.43 rumoles) of the spirononatriene in 1 ml of ether was injected into the preparative gic on a 3/8" X 5' 20% SE30 on 60/80 Gas ChromiZ column at 150 0 and 50 psi4 in

0.1 ml portions. The two peaks which were eluted were collected at room temperature. The larger peak (retention time ca. 8mrin) gave white crystals, r-p 68-70. Recrystaliato from benzene--hexane gave 30 rug (30%) of trans--l,2-dicarbomiethoxy0
indane with mrp 70-72 .This sample was identical with a sample synthesized by the method of Cook and Stephenson (70).

Anal. Calcd for C 13H 1404 C, 66.66; H, 6.02. Found: C, 66.51; H, 5.99.

The minor peak (retention time ca. 6 ruin) also gave a
0
white solid, 2 rug, with r 65-69 Gle showed this to be a 50:50 mixture of the two peaks. Ir and uv of the solid were identical with those of the trans isomer above, so the minor





7 "


peak was assigned thle structure of cis-1,2-dicarbomet-hoxindane.

S. Thermal Rearrangement of 8,9-]Di~carbomethox. ysiro[2, 67L nona-2,4,6-triene to 1,2-Dicarbomethoxyindane in n-Decane and Triglyme. Two solutions were made consisting of 32.7 mg (0.14 mpmoles) in 0.5 ml of n-decane and 33.3 mg (0.14 inioles) in 0.5 ml of triglyme. The solutions were placed in nmr tubes and suspended in an oil bath at 1500 The reaction was followed by removing the tubes at 5 min intervals and examining their nmr spectra. The approximate half-lives as determined by the appearance of the aromatic protons of the indane were 20 min and 8 min for the n-decane solution and the tricglyme solution, respectively. T. Mehl ,N-'cloheotatrienylcarbarniate. To a solution of 13.3 g (0.10 moles) of 2,4,6-cyclohentatri.enyl--l-isocyanate in 150 ml of dry benzene at 10 0 was added 5.4 g (0.10 moles', of sodium me-thoxide in 30 ml of methanol. Reaction was vigorous and exothermal. An ir spectrum of the crude mixture showed no isocyanate peak and a new carbonyl at 1680 cm After filtration, the filtrate was concentrated on a rotary evaporator to a dark oil. After dissolving the oil in a small amount of dichioromethane, pentane was added slowly. The solid which formed was filtered and dried to give 7.3 g (44%) of the crude carbamate with mp 69-72. No attempt was made to improve this yield. Three recrystallizations from hexane gave a pure sample with mp 78.5-80.5 0 V r max
3270, 3080, 2945, 1680, 1555, 1310, 1255, 1050, 700, and





79


440 cm'l X tl 253 my~ (c 4400); ICDC13- 3.34 (complex
max T:4S

pattern, 2 H, 3 and 6 protons on cycloheptatriene ring),

4.53 (complex pattern, 2 H, 2 and 7 protons on cyclohepta-triene ring), 4.83 (broad peak, 1 H, nitrogen proton), 5.97 (quartet, 1 H, methyne proton), and 6.34 (singlet, 3 F, methyl protons) ; rn/c (70 ev) 165 (molecular ion) 150, 133, 106 (base peak) 91, 79, 77, 65, arnd 59.

Anal. Calcd for C 9H 11NO C,: 65.44; H, 6.71; 0f 19.37. Found: C, 65.63; H, 6.73; 0, 19.3S. U. Methyl N-Nitroso-N-cycloheptatrienvl.carbamate. A stirred slurry of 250 mg (1.51 mmoles) of Pethyl N-cycloheptatrie nylcarbamate, 1.23 g (15.0 mrnoles) of anhydrous sodium acetate, and 0.5 g of anhydrous sodium sulfate in. 15 ml of dry dichloromethane was cooled to -25O A 0.48 M ether solution of dinitrogen tetroxide (prepared by bubbling the gas into a tared volumetric flask of ether at -50 0 and noting the increase in weight) was added in 3 ml portions until the green color persisted for at least 10 min. The excess dinitrogen tetroxide and approximately half of the dichloromethane were removed in vacuo and 10 ml of anhydrous ether was added. The reaction mixture was then washed three times with ice-cold acqueous sodium bicarbonate solution saturated with salt and twice with ice-cold salt solution. The aqueous layers were combined and washed twice with dichloromethane. All organic fractions were combined and dried over sodi~um sulfate, then filtered through magnesium sulfate to remove the last traces of water. After removing the solvent in vacuo, the dark oil was taken





so


up in 1:1. ether-penLane solution and passed through a short

silica gel. column which had been deactivated by packing in water-saturated ether. The yellow eluent was collected

until it was colorless and concentrated under a stream o.f nitrogen. The yellow oil (237 mg, 80%) was reasonably pure by nmr; lts 3020, 2960, 1.755, 1510, 1440, 1360, 1310,

1.200, 1150, 1070, 770, and 705 cm* 1 T-CC4 3.38 (broad triplet, 2 H, 4 and 5 protons on cycloheptatrienc ring),

3.75-4.10 (complex pattern, 2 H, 3 arid 6 protons on cycloheptatriene ring), 4.70-5.10 (complex pattern, 2 H, 2 and 7 protons on cycloheptatriene ring) 5.20-5.50 (broad triplet,

1 11, methyne proton), and 5.89 singletsf. 3 H, methyl protons). V. Decompsitio of ____v _____s-Ncv lhptatri envlcarbamate with Sodium Methcoxide. To a stirred solution of 240 mg (1.23 irnoles) of the nitrosocarbamate in 30 ml of anhydrous ether were added 850 rig (15.7 rmoles) of sodium methoxide and 2-3 drops of methanol. Quantitative nitrogen evolution occurred within 30 min. The solid was removed bv vacuum filtration and the filtrate was concentrated to a dark oil on a rotary7 evaporator. This oil was taken up in 10 ml of 5% ether in hexane and chromatographed on 3-0 a of basic alumina (Woelm, Activity grade III) using 5% ether in hexane as eluent. The heptafulvalene came off rapidly as a dark red band, and the yield, as determined by uv, was 8 mg

(7%).

W. Decomposition of Methyl N-Nitroso-'N-cyc oh e a tri'e -vl carbamate with Sodium. 2ethoxide in the Presence of Dimethyl





81


Fumarate. To a stirred solution of 237 mng (1.22 romles) of the ritrosocarbamate in 30 ml of anhydrous ether were added 1.75 g (12.2 mmolces) of dimethyl fumarate and 2-3 drops of methanol. After the olefin had completely dissolved, 600 mg (11.1. mmoles) of sodium methoxide was added. Quantitative nitrogen evolution occurred within 1 hr and 15 Min. The solid was filtered and discarded, and the filtrate was concentrated under a stream of dry nitrogen to 10 ml. The remaining solution was eva porated to dryness on a. rotary evap

orator and taken up in 10 ml of 5% ether in hexane. Careful chromatography of this solution on 30 g of basic alumina (Woelm, activity grade III) using 5% ether in hexane and gradually proceeding to 8% ether in hexane gave 7 mg (4%) of the spirononatriene which was pure by glc. X. Decomposition of Methyl N-Nitroso-N-ccloheptatrienylcarbamate with Potassium t-Butoxide. A solution of 210 mg (1.08 mrnoles) of the nitrosocarbamate in 30 ml of anhydrous ether was cooled to 0 0 with stirring. To this was added 650 mg (5.8 mmoles) of potassium t-butoxide, and the dark MIXture was allowed to warm to room temperature. Gas evolution. was measured during the next 2.5 hr to be 20 ml (83%) The mixture was filtered, and the filtrate evaporated to dryness on a rotary evaporator. The oily residue was taken up in 10 ml of 5% ether in hexane and chromatographed on 20 g of basic alumtina (Woelm, activity grade III) using 5% ether in hexane as eluent to give the dimer as a dark red band. Determination of the concentration of a solution of the dimer showed









13 incj (13')

Y. Decomposition ofMethyl N-NitroEo-N-cclohetatr-ien-y1carbamate with Potassium t-ButoXide in the Presence of

Dime thylurte A solution of 200 mg (1.03 rnmoles) of

the nitrosocarbamate and 1.50 g (10.4 mmoles) of dimrethyl fumarate in 30 m'L of anhydrous ether was cooled to 0C0 wit~h stirring. To this was added 700 mg (6.2 mmoles) of potassium t-butoxide, and the mixture was allowed to warm to room temperature. Gas evolution over the next 4 hr was 20 ml (87%). The orange slurry was poured into 50 ml of water and extracted with three 25 ml portions of dichloromethane. The organic extracts were combined, dried over sodium sulfate, and concentrated to an oily solid on a rotary evaporator. The residue was taken up in 10 ml of 5% ether in hexane and chroInatographed on 30 g of basic alumina (Woeim, activity grade III) using 5% ether in hexane as eluent to give 108 mg of a white crystalline solid which was shown by nmr to consist of 11 mg (4.5%) of the spirononatriene and 99 mg of dimethyl fumarate. Glc and tic show.,ed these to be the only components. Z. N,N-Dimethyl-H'l-cvcloheptatLrienylurea_(77). To a solution of 2.5 g (19.0 nrn'oles) of 2,4,6-cycloheptLatrien-l--ylisocyanate in 20 ml of anhydrous ether at 0 0, was added 3 g (66 mmoles) of dimethylamine. After stirring the solution for 10 min, the ether and excess dimethylamine were removed on a rotary evaporator to give 1.3 g (38%) of the urea as.a crude solid with mp 110-.115. Three recrystallizations from benzenehexane gave an analytical sample with mp 124-1.27 0 V r max





83


3320, 3010, 2930, 1625, 1525, 1380, 1302, 1235, 1198, 1127, 1069, 1058, 1010, 923, 864, 857, 772, 748, 730, 703, 625, 595, 461, 418, 393, and 262 cm- ; 1mX E 258 my1~ Q 4200); T Cs13 3.35 (complex pattern, 2 Hi, 4 and 5 protons on cyclo-heptatriene ring) 3.80 (complex pattern, 2 H, 3 and 6 protons on cycloheptatriene ring), 4.48 (complex pattern, 2 H, 2 and 7 protons on cycloheptatriene ring) 5.0 (broad doublet, 1 H, nitrogen proton) 5.83 (quartet, 1 H, methyne proton), and 7.18 (singlet, 6 H, methyl protons); in/e (70 ev) 178 (molecular ion) 163, 133, 132, 106, 104, 91, 72 (base peak), 65, 46, 45, 44, and 15; metastable peaks, 149 and

99.5.

Anal. Calcd for C 10 H14 N2O0: C 67.39; H, 7.93; 0, 8.98. Found: C, 67.23; H, 7.94; 0, 9.11. AA. N,N-Dimethyl-N'-nitroso-N'-cyc oheptatrienylurea. A mixture of 500 mng (2.81 moles) of the urea, 2.30 g (28.1 inmoles) of sodium acetate, and 0.5 g of sodium sulfate in 20 ml of dry dichioromethane was cooled to -25 0 with stirring. A solution of 0.29 M dinitrogen tetroxide in anhydrous ether was added in 3 ml portions until the green color persisted for at least 5 min. The excess dinitrogen tetroxide and ca. 10 ml of dichloromethane were removed in vacuo while maintaining the temperature between -25 and -30 0. The nitrosourea decomposed rapidly above 0 0, so the following steps were carried out as quickly as possible. The reaction mixture was added to 10 ml of cold anhydrous ether then washed twice with ice-cold 10% aqrueous sodium bicarbonate solution




84


and twidce with Ice-cold aciueous salt solution. The orcjanic'' layer was quickly cooled to -20 0, and sodium sulfate was added. After 15 min, the mixture was filtered to give an orange solution which was kept below 400 and used without further purification; T DI3 (-200) 2.94-3.30 (poorly resolved triplet, 2 H1, 4 and 5 protons on cycloheptatriene ring), 3.44-3.84 (comp]ex pattern, 2 11, 3 and 6 protons on cycloheptatriene ring), 4.36-4.76 (complex pattern, 2 11, 2 and 7 protons on cycloheptatriene ring) 5.29 (triplet, 1 1, methyne proton), and 6.71 and 6.80 (singlets, 6 11, methyl groups). Upon warming to 100, the cycloheptatriene ring pattern broadened and lost all resolution. The methyl peaks disappeared, and two new singlets appeared at T 6.96 and 7.61. BB. Decomnposition of N,N-Dimethyl-N'-nit-roso-N'-cycloheot-trienylurc .a with Sodi;um- 'NJthoxide. To0 a solution of the nitrosourea prepared from 200 iig (1.15 mrnoles) of the urea in 30 ml of anhydrous ether at -20 0 were added 300 mg (5.55 inmoles) of sodium methoxide and 2-3 drops of methanol. The dark mixture was allowed to warm slowly to room-, temperature with stirring. After filtration, the filtrate was taken to dryness on a rotary evaporator. The oily residue was taken up in 5 ml of 5% ether in hexane and chromatographed on 12 g of basic alumina (Woelm, activity grade !II) using 5% ether in hexane as eluent. Heptafulvalene was eluted as a dark red band., and uv determination of the yield showed 2.1 mg (2.1%,

based on starting urea).

CC. Decom~position of N.,',N-Dimet-hyl-N'-ni,.roso-,.7'-cyclohepta-





85


trinylrcawit SoiumIleho-de in thePresence of Dimethyl Fumarate. To a solution of the n j4t-rosourea prepared from 125 mg (0.70 moles) of the urea in 25 ml of anhydrous ether at -20 0 was addedl a cold (0 0) mixture of 40 nil of anhydrous ether,.1.0 g (7.65 mnoles) of dimethyl furmarate, 300 mrg (5.55 mrnoles) of sodium rnethoxide, and 2-3 Crops of methanol. The dark reaction mixture was allowed to warm slowly to room temperature with stirring. After filtration, the filtrate was taken to dryness on a rotary evaporator. The crude solid was taken up in 5% ether in hexane and chromatographed on 20 g of basic alumina (Woelm, activity grade III) using 5% ether in hexane. Elution of the spirononatriene was followed by glo on a 5', 20% SE30 column at 160 0 and an inlet pressure of 20 psi. Using -methoxynaphthalene as an internal standard, the yield was determined by 91c to be 0.80 mg (0.5%, based on starting urea). DD. Carboxvcycloheptatrien2yL_.ium uoroborate. A solution of 0.500 g (3.68 rnroles) of 2,4,6-cycloheptatrienecarboxylic acid in 10 ml of dichloromethane was added dropwise over a 30 min period to a stirred solution of 1.50 a (4.55 mmoles) of trityl fLluorob~orate (78) The dark mixture was then refluxed for 1.5 hr under nitrogen. The mixture was filtered to give a light brown solid which, after drying in vacuo, weighed 665 mg (82%). The crude carboxytropylium filuorobo-rate charred at 1500, and melted with gas evolution at 1650 After several recrystallizations from nitromethane--benzene, an analytical sample was obtained which charred at 153 0 and





86


melted with gas evolution at 16max 3010, 2760, 2563,

2420, 1722, 1472, 1439, 1334, 1214, 1070, 350, 792, 734, 636,

53,522, and 438 cm- 1 AE 272 my (s 5700); TA2'
.max M

-0.1 (broad singlet, 1 H, carboxylic acid proton), 0.0-0.4 (complex pattern, 2 Hi, 2 an6 7 protons on cycloheptatriene ring), and 0.5 (broad singlet, 4 H, 3,4,5, and 6 protons on cycloheptatriere ring).

Anal. Calcd for C 8H 7 0 2BF 4 C, 43.25; H, 3.16. Found: C, 43.52; H, 2.99.















.LIST OF REFERENCES


1. J. Hine, "Divalent Carbon," Ronald Press, New York,
1964.

2. W. Kirmse, "Carbene Chemistry," Academnic Press, New
York, 1964.

3. H. W. Wqanzlick, An ew. Chem., Internat._Ed. Engl., 1,
75 (1964).

4. N. Wiberg and J. W. Buchier, Chem. Ber., 96, 3223
(1963)

5. N. Wiberg and J. W. Buchier, Chem. Ber., 96, 3000 (1963)

6. D. M. Lemal, R. A. Lovala, and K. I. Kawano, *J. Am.
Chemn. Soc., 86-, 2518 (1964).

7. H. E. Winberg, J. E. Carnahan, D. D. Coffman, and M.
Brown, J. Am. Chemn. Soc., 87, 2055 (1965).

8. H. W. Wanzlick and H. Ahrens, Chem. Ber., 99, 1580 (1966).

9. N. Wiberg and J. W. Buchier, Ancjew. Chem., Internat.
Ed. Efngl. 1, 49 6 (19 62).

10. D. M. Lemal and K. I. Kawano, J. Am. Chem. Soc., 84,
1761 (1962).

11. N. Wiberg and J. W. Buchier, J. Am. Chem. Soc., 85,
243 (1963).

12. A. Piskala, Tetrahedron Lett., 1964, 2587. 13. H. W. Wanzlick, B. Lachmann, and E. Schikora, Chem.
Ber., 98, 3170 (1965).

14. W. P. Norris and R. A. Henry, Tetrahedrci Lett., 1965,
1213.

15. R. A. Olofson, W. R. Thompson, and j. S. Michelman,
J. Am. Chem. Soc., 86, 1865 (1964).


87





88


16. Rk. Breslow, JT. AIm. Chemn. Soc., 80, 3719 (1953). 17. H1. W. Warnzlick and 1-. JT. Kiciner,_Ancrew. Chemn., Intornat. Ed. Engi., 3, 65 (1964).

18. H. Quast and S. H-unig, Ancrew. Chem. Internat. Ed. Engl. ,
3, 800 (1964).

19. JT. Metzger, et al., BuLll. Soc. Chima. Fr., 1964 2857. 20. H. W. Wanzlic'K, et al., Ancrew. Chem., Interniat._Ed.
Engjl., 5, 126 (196-6)

21. H. Wahl and JT. CT. Vorsanger, Bull. Soc. Chim. Fr.,
1965, 3359.

22. H. Balli, AneChe m. Internat. Ed. Eng1. 3, 809
(1964).

23. H. Quast and E. Frankenfeld, Ancrew._Chem.,_Internat._Ed.
Engl. 4, 691 (1965).

24. U. Scholikoof arid F. Gerhart, Anqewi. Chem., Internat.
Ed. Eng 6 805 (1967).

25. D. M4. Lernal, E. P. Gosselink, and A. Ault, Tetrahedron
Lett., 1964, 579.

26. Rk. W. Hoffmann and H. Hlauser, Tetrahedron Lett., 1964,
197.

27. Rk. W. Hoffmnann and H. Hlauser, Tetrahedron, 21, 391
(1965).

28. D. M. Lemal, E. P. Gosselink, and S. D. McGregor, CT. Am.
Chem. Soc., 88, 582 (1966).

29. R. W. Hoffmann and H. Hauser, Tetrahedron Lett., 1964,
1365.

30. R. JT. Crawford and Rk. Raap, Proc. Chem. Soc., 1963, 370. 31. W. M. Jones and jT. M4. Denham, JT. Am. Chemn. Soc., 86, 944
(1964).

32. W. M. Jones and M. E. Stowe, Tetrahedron Lett., 1964,
3459.

33. 1. E. Stowe, Ph.D. Dissertation, University of Florida,
August, 1967.





89



34. R. Breslow and L. J. Altman, J. Am. Chem._ Soc., 38,
504 (1966).

35. 1. I'Moritani, et al,, Tetrahedron Lett~., 1966, 373. 36. 1. Moritani, et al., J. Am. Chem. Soc., 89, 1259 (1967). 37. S.-I. Murahashi, I. 14oritani, and M. Nishino, J. Am.
Chem. Soc., 89, 1257 (1967).

38. Private communication with R. Hloffmarnn. 39. A preliminary account of this worh has already appeared.
W. M. Jones and C. L. Ennis, J. Am. Chem. Soc., 89,
3069 (1967).

40. W. R. Bamford and T. S. Stevens, J. Chem. Soc., 1952,
4735.

41. P Radlick, J. Orcj. Chem., 29, 960 (1964). 42. Cf. D. M. G. Lloyd, "Carbocyclic Non-benzenoid Arormatic
Compounds", Elsevier Publishing Co., New York, N. Y.,
1966, p. 135.

43. N. L. Bauld and Y. S. Rim, J. Am. Chem. Soc., 89, 6763,
(1967).

44. H. J. Dauben and D. F. Rhoades, J. Am. Chem._Soc.,89
6764 (1967).

45. T. Mukaki, Bull. Chem. Soc., Jayan, 33, 238 (1960). 46a. G. Sunnagawa and N. Soma, Japanese Patent 12674 (1962);
Chem. Aflstr., 60, P4064h (1964).

46b. Y. Kitihara, T. Asao, and M. Funamizu, Japanese Patent
11629 (1964); Chem. Abstr., 61, P16021e (1964).

47. Y. Kitihara, T. Asao, M. Funamizu, Japanese Patent
11628 (1964); Chem. Abstr., 61, P16021e (1964).

48. W. Kirmse, B.-G. Von Bulow, and H. Schepp, Anal._Chem.,
691, 41 (1966).

49. C. G. Overberger, J.-P. Anselme, and J. G. Lombardino,
"Organic Compounds with Nitrogen-Nitrogen Bonds,"'
Ronald Press, New York, 1966.

50. R. Baltzly, et al., J. Org._Chem., 26, 3669 (1961).





90


51. D. G. Farnum, j. Org, Chem., 2a, 870 (1963).

52. W. von E. Doering in "Theoretical Organic Chemistry.
The Kekule Symposium," Academic Press, Inc., New York,
N. Y., 1959.

53. A. Streitwieser, Jr., "Molecular Orbital Theory for
Organic Chemists," John Wiley and Sons, Inc., New York,
N. Y., 1961.

54. D. j. Patel, M4. E. H. Howden, and J. D. Roberts, J. Am.
Chem. Soc., 85, 3218 (1963).

55. R. Huisgen, Angew. Chem., Internat. Ed._Engi., 2, 633
(1963).

56. W. M. Jones, T. H. Glenn, and D. G. Baarda, J. Org.
Chem., 28, 2887 (1963).

57. P. S. Skell and R. C. Woodworth, J. Am. Chem. Soc., 78,
4496 (1956).

58. P. P. Gaspar and G. S. Hammond in "Carbene Chemistry,''
W. Kirmse, Ed., Academic Press, New York, N. Y., 1964,
Chapter 12.

59. M4. Jones, Jr. and K. R. Rettig, j. Am. Chem. Soc., 87.,
4015 (1965).

60. M4. Jones, Jr. and K. R. Rettig, J. Am. Chem. Soc., 87,
4013 (1965).

61. 14. Jones, Jr., A. Kulczycki, Jr., and K. F. Huunel,
Tetrahedron Lett., 1967, 183.

62. F. A. L. Anet, J. Am. Chemn. Soc., 86, 458 (1964).

63. Cf. K. Alder and C. Jacobs, Chem. Ber., 86, 1528 (1953),
K. Alder, R. Muders, W. Krane, and P. Wirtz, Ann., 627,
59 (1960) and earlier references cited therein.

64. This reaction was suggested by Prof. J. A. Berson at
the Univ. of Wisconsin.

65. H. E. Simmons and T. Fukunaga, J. Am. Chem. Soc., 89,
5208 (1967).

66. R. Hoffmann, A. Imamura, and G. D. Zeiss, J. Am. Chem.
Soc., 89, 5215 (1967).

67. D. M. Lemal and S. D. McGregor, J. Am. Chem. Soc., 83
1335 (1966).






91


68. S. D. M~cGregor and D. M. Lemal, J. Am. Chem. Soc., 88,
2858 (1966).

69. C. M. Kaiifman, J. A. Smith, G. G. Vander Stouw, and
H. Shechter, J. Am. Chem. Soc., 87, 935 (1965).

70. J. W. Cook arnd E. F. M. Stephenson, J-. Chem. Soc. ,
1949, 842.

71. R. B. Woodward and R. Hoff nann, J. Am. Chem. Soc. 87,
2511 (1965).

72. JT. A. Berson and G. L. Nelson, J. Am._Chem. Soc. 89,
5503 (1967).

73. W. von E. Doering arnd L. E. Helgen, J-. Chem._Soc.,
1961, 482.

74. D. N. Kursanov, M. E. Vol'pin, I. S. Akhrem, and I. Ya.
Kachkurova, Izvest. Akad, Nauk S.S.-S.R., Otdel. Khim.
Nauk, 1957, 1371; Chem. Abstr., 52, 726 (1958).

75. C. M. Orl rtdo, Jr. and R. Weiss, J. Org. Chem., 27,
4715 (1962).

76. A. W. Johnson, A. Langemain, anid M. Tisler, j. Chem.
Soc., 1.955-, 1622.

77. This compound was first prepared by VT. N. Jones.

78. H. J. Dauben, Jr., L. R. Honnen, and K. N. Harmon, J.
Ora._Chem., 25, 1442 (1960).














BIOGRAPHICAL SKETCH


Cecil Lawrence Ennis was born August 14, 1943, in

Auburn, Alabama. Ile was graduated from Auburn High School in June, 1960 and entered Auburn University the following September. While there he was a member of Phi Lambda Upsiion chemi cal honorary society. Hie obtained his Bachelor of Science degree in chemistry in June, 1964 and enrolled in the Graduate School. of the University of Florida the following September. He was an Arts and Sciences Fellow and Gul.f Oil Fellow during his graduate study. He is presently a member of The American Chemical Society and The Chemical Society of London.

Mr. Ennis is -married to the former Linda R~ae Hayes. He is a First Lieutenant in the U.S. Air Force and will enter active duty in April, 1968.


92











This dissertation was prepared under the directions of the chairman of the candidate' s supervisory committee. and has been approved by all members of that committee. It was submitted~ to the Dean of the College of Arts and Sciences and to the Graduate Council, and was approved as partial fulfillment of the requirements for the degree of Doctor of Philosophy.



March 19682


Dean, Coller-/of Ar s and Sciences




Dean, Graduate School

S up e r v i sory Committee:


Chairman




Full Text

PAGE 1

ATTEMPTS TO GENERATE CYCLOHEPTATRIENYLIDENE By CECIL LA WREN CE ENNIS A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 1968

PAGE 2

ACKlWWLEDG.MENT The author would like to express his deepest gratitude to Professor W. M. Jones. Perhaps some of the many virtues and scholarly ideals to which he adheres have rubbed off on this student during the past years. His frequent coun sel, not only as a teacher, but as a friend, has left lit tle to be desired. Because of him and the various members of his research group, this period of time has been educa tional and richly rewarding. There are no words to express the feeling the author bears for his wife. An equal amount of her blood, sweat, and tears have gone into not only the typing, but the mak ing of this dissertation. For providing the steadying influence, this author is deeply grateful. Finally, the author wishes to thank the College of Arts and Sciences of the University of Florida and the Gulf Oil Company for providing fellowships during this period, and '1 1 he U.S. Army Research Off ice, Durham, for further financial assistance. ii

PAGE 3

TABLE OF CONTENTS ACKNOWLEDGMEN'l' ... ..... ................. ........ Page ii CHAP'fER I. INTRODUC'I'ION AND THEORY 1 II. DECOMPOSITION OF 2,4,6-CYCLOHEPTATRIENONE p-TOLUENESULFONYLHYDRAZONE AS A ROUTE TO CYCLOHEP'l'A'l'RIENYLIDENE A. Synthesis of 2,4,6-Cycloheptatrienone p-Toluenesulfonylhydrazone ............. 13 B. Decomposition of 2,4,6-Cycloheptatri enone p-Toluenesulfonylhydrazone Sodium Salt . . . . . . . . 16 C. Decomposition of 2,4,6-Cycloheptatrienone p-Toluenesulfonylhydrazone Sodium Salt in the Presence of Dimethyl Fumarate, Fumaronitrile, and Maleinitrile. Formation of Spiro[2.6]nonatrienes ..... 20 D. Mechanism of the Formation of Heptaful valene and the Spiro (2. 6] nonatrienes .... 30 E. Other Olefins as Cycloheptatrienylidene Acceptors . . . . . 3 8 F. Decomposition of 2,4,6-Cycloheptatri enone p-Toluenesulfonylhydrazone Sodium Salt under High Vacuum . . . 4 2 III. THERM..:l'\L REARRANGEMENT OF 8, 9-DICJ\.RBOMETH OXYSPIRO [2. 6] NONA-2, 4, 6-TRIENE TO 1, 2DICARBOMETHOXYINDANE 46 IV. ADDITIONAL SCHEMES EXA.i.''iINED AS ROUTES '1 1 0 CYCLOHEPTATRIENYLIDENE A. Base-promoted Decomposition of Methyl N-Ni troso=-N-cyc lohepta trienylcarba!na te and N, NDimethyl-N' -ni troso-N' cyclo hepta.trienylurea . . . . . 52 iii

PAGE 4

B. a-Elimination Schemes ... ........... 57 V S UMMJ\ .. R Y e61 VI. EXPERIMENTAL "fl...... General . . . . . . . 63 B. Chlorocycloheptatrienylium Chloride ... 64 C. 2,4,6-Cycloheptatrienone p-Toluenesulfonylhydrazone Hydrochloride .......... 65 D. 2,4,6-Cycloheptatrienone p-Toluenesulfonylhydrazone . . . . . . 65 E. Photolysis of 2,4,6-Cycloheptatrienone p-Toluenesulfonylhydrazone Sodium Salt. Formation of lleptafulvalene .. ........ 66 F. Determination of the Yield of Heptafulvalene . . . . . . . . 67 G. Photolysis of 2,4,6-Cycloheptatrienone p-Toluenesulfonylhydrazone Sodium Salt in the Presence of Dimethyl Fumarate. Formation of 8,9-Dicarbomethoxyspi~o[2. 6] nona-2, 4, 6-triene . . . . 68 H. Isomerization of 8,9-Dicarbomethoxy spiro (2. 6] nona-2, 4, 6-triene to 8-Car bomethoxy-8-carbomethoxymethylrnethy lenecyclohepta-2,4,6-triene on Acid Alumina . . . . . . . 7 0 I. Photolysis of 2,4,6Cycloheptatrienone p-Toluenesulfonylhydrazone Sodium Salt in the Presence of Fumaronitrile. Formation of trans-8,9-Dicyanospiro [2.6]nona-2,4,6-triene .. .... .. .. .. 70 J. Isomerization of trans-8,9-Dicyanospiro [2. 6] nona-2, 4, 6-triene to 8-Cyano8-cyanomethylmethylenecyclohepta2,4,6-triene on Acid Alumina ......... 72 K. Photolysis of 2,4,6-Cycloheptatrienone p-Toluenesulfonylhydrazone Sodium Salt in the Presence of Maleinitrile. Formation of ~~i~-8,9-Dicyanospiro[2.6]nona-2,4,6-triene .... ...... .......... 73 iv

PAGE 5

L. Photolysis of 2,4,6-Cycloheptatrienone p-Toluenesulfonylhydrazone Sodium Salt in the Presence of Other Acceptors. . 7 4 M. Thermal Decomposition of 2,4,6-Cyclo heptatrienone p-Toluenesulfonylhydra zone Sodium Salt with and without Dimethyl Fumarate .... ............... 74 N. Photolysis of Heptafulvalene in the Presence of Dimethyl Fumarate ......... 75 O. Photolysis of 2,4,6-Cycloheptatrienone p-Toluenesulfonylhydrazone Sodium Salt in the Presence of Dimethyl Fumarate and Triethylamine .................. 75 P. Thermal Decomposition of 2,4,6-Cyclo heptatrienone p-Toluenesulfonylhydraione Sodium Salt under Vacuum......... 76 Q. Thermal Decomposition of 2,4,6-Cyclo heptatrienone p-Toluenesulfonylhydra zone Sodium Salt under Vacuum with Dimethyl Fumarate ...... ... ......... .. 76 R. Thermal Rearrangement of 8,9-Dicarbo methoxyspiro [2. 6] nona-2, 4, 6-triene to 1,2-Dicarbomethoxyindane ........... 77 S. Thermal Rearrangement of 8,9-Dicar bornethoxyspiro [2. 6J nona-2, 4, 6-triene to 1,2-Dicarbomethoxyindane inn-Decane and Triglyme ... ...... ............ 78 T. Methyl N-Cycloheptatrienylcarbamate.... 78 U. Methyl N-Nitroso-N-cycloheptatrienylca.rbarna te . . . . . . 7 9 V. Decomposition of Methyl N-Nitroso-N cycloheptatrienylcarbamate with Sodium Methoxide ................... 80 W. Decomposition of Methyl N-Nitroso-N cycloheptatrienylcarbamate with Sodium Methoxide in the Presence of Dimethyl Fumarate .................... 80 X. Decomposition of Methyl N-Nitroso-N cycloheptatrienylcarbamate with Potassium t-Eutoxide ................... 81 V

PAGE 6

Y. Decomposition of Methyl N-Nit.roso-N cycloheptatrienylcarbamate with Potassium t-Butoxide in the rresence of Dimethyl Fumarate ................. 82 z. N,N-Dimethyl-N'-cycloheptatrienylurea ................................... 82 AA. N,N-Dimethyl-N'-nitroso-N'-cycloheptatrienylurea ......................... 83 BB. Decomposition of N,N-Dirnethyl-N' nitroso-N'-cycloheptatrienylurea with Sodium Methoxide . . . . 84 CC. Decomposition of N,N-Dimethyl-N' nitroso-N'-cycloheptatrienylurea with Sodium Methoxide in the Presence of Dimethyl Fumarate ......... 85 DD. Carboxycycloheptatrienylium Fluoroborate .................................. 85 LIS'l' OF REFERENCES . . . . . . . . 87 BIOGRAPHICAL SKETCH. 9 2 vi

PAGE 7

CHAPTER I INTRODUCTION A N D TH E ORY Although carb e n e s ar e normally electrophilic species ( 1, 2) by virtue of their sextet of electrons, a ground state singlet carbene in which the nonbonding pair of electrans is localiz e d in the 2 orbital and the empty orbi~ sp p tal is stabiliz e d by strong electron donation is potentially nucleophilic. Conjugation of the p orbital with nonbonding electrons on adjacent heteroatoms adds electron density to the carbene carbon and profoundly influences its reactivity. There are several examples of stable divalent carbon com pounds in which the adjacent heteroatom is doubly bonded to the electron-deficient carbon. Carbon m onoxide, I, isoni triles, II, and fulminic acid derivatives, III, behave as nucleophiles tow ar d Lewis acids and metal c~rbonyls (1,2). + Q=C: -<.:-->-t10=-C: R N = C: "'"-<-=---::;,. RN =C: I II + R-0-N=C:-~--~ R-0-N~C: III In the r e cent lit e ratur e th e re h ave be e n e x tensi ve 1

PAGE 8

2 accounts of research on divalent c arbon singly bonded to heteroatoms. Tetraaminoethylenes have been investigated by a number of workers (3,4,5,6,7,8,9,lO,J.l,12,13). The reports have centered around the purported dissociation of dimers such as IV to carbene Vin a reversible fashion. The carbene is stabilized by electron donation from the nonbond ing electrons on ~he nitrogen atoms into the vacant p orbi tal of the carbene carbon. Evidence for the dissociation includes consistently low molecular weights for the olefin, IV, by the Rast method. R R I I R'-N N-R' .. \ I .. c==c .. / \ .. R'-N N-R' I I R R -----<.. -,R I R'-N ..\ C: ../ R'-N I R R I R'-N+ c:.. / RLN I R IV V Dimers of this type do not react with olefins with the exception of tetracyanoethylene. With this highly electron deficient olefin, a charge-transfer species, VI, is formed rather than a normal carbene adduct (4). The dimers, how ever, do react with oxygen, sulfur, diazo compounds, and Lewis acids to give products expected of carbenes. IV+ NC CN \ I c=c I \ NC CN ----~R'RN NR'R \\ /,: ..1..' C I+ I J --C, ,,' I \ ._ R'RN NR'R NC CN '-\ // -:c-c~'/ \'N C C F VI

PAGE 9

3 There has been considerable doubt cast on the validity of these dissociations by crossover experiments in which mixtures of dimers with different R groups failed to give crossed products (6,7). An alternate reaction scheme postu lated to incorporate the results of the crossover experi ments involves reaction of the dimer directly with the elec trophile. This scheme still involves th~ carbene as a bypro duct which can dimerize to regenerate the olefin or react with the trapping agents in a normal carbene process. R'RN NR'R "\ /" E-tC=C --~ .. I \ .. R'RN NR'R R'RN E NR'R R'RN+ E /NR'R .. \ \ \ \ .. +C--C ~ C-C .,/ \.. ., / \ .. R'RN NR'R R'RN NR:R R'RN .. \ + C=E .. / R'RN =-<.. R'RN .. \ C : .. / R'RN + NR'R + / .. E ==C \ .. NR'R Similar carbenes have been proposed to explain the acidity of tetrazolium cations, VII (14,15), and thiazolium cations, VIII (16,17,18,19,20). Early workers actually claimed to have isolated the latter carbene as a stable, crystalline solid but later retracted their statements (17,20). The thiazolium carbene has also been generated by a-elimina tion (21) and decomposition of N-azides(22).

PAGE 10

4 H "' I /c, .N :N! /c~ + : S: N-R \\ I N-N: I+\ R R H --;>, ,, R R VII VIII Another similar species is IX whi.ch was generated by a decarboxylation scheme (23). A report has been made on the possible intermediacy of Xa, but it reacts as its isomer, ib ( 24 ). 00 ~-)-. + N N# LiO-C-N (Et) 2 Li-g-N (Et) 2 I I R R Xa Xb IX There have been several reports of attempts to generate dioxocarbenes. Thermal decomposition of norbornadienone ketal, XI, gives tetraalkoxyethylene, XII, as a major product under aprotic conditions. In alcohol solution, decomposi tion gives little or no ethylene but substantial yields of orthoformates, XIII (25,26,27,28). The ethylene, XII, was found by one group to undergo dissociation-type reactions similar to the tetraaminoethylenes (29). An approach involv

PAGE 11

ing base-catalyzed decomposition of the p-tolu e nesulfonyl hydrazone, XIV, gave only radical products under aprotic concli tions ( 3 0) RO'/OR C=c RO OR RO/ "--oR x__ y \)--1"'~ ROH XI XII ..,...OR H--C -OR ....._OR XIII XIV 5 An alternate approach to enhancing the nuceophilicity of carbenes involves conjugation of the p orbital with an olefinic system. For maximum effect, this p orbital c~n be incorporated in a TT system which will be aromatic if the orbital contains neither of the nonbonding electrons of the carbene. This may allow the singlet, in which the two elec2 trans are in the sp orbital, to beco m e the ground state. The simplest exam p le of this system is cyclopropenylidene, XV, in which the carbe~e p orbital is involved in cyclopro penylium-type resonance. R R R xv

PAGE 12

6 Diphenylcyclopropenylidene, XV (R =ph e nyl), has been generated by a--el.imination of carbamic acid from the carba mate, a-elimin a tion of carboxylic acids from the corresp?nd ing esters, and base-c a talyzed deco m position of N-nitrosoN(2, 3-diphenyl -2cyclopropenyl) carbama te (31,32,33). It was found to be nucleophilic, reacting with electron-defi cient double bonds such as dimethyl fumarate, N,N,N',N' tetramethylfumaramide, and fumaronitrile to give the corre sponding spiropentenes, XVI, which underwent facile rear rangement to the triafulvenes, XVII. Ph R ---+ Phx.R--'>XVI XVII R. Breslow and L. F. Altman have pointed out that per haps one factor contributing to the acidity of the ring hydrogen of n-propylcyclopropenone is contribution of a cyclopropenylidene resonance for m (34). J>=o Pr The work which will be described here concerns the next higher case in which the TI system will be aromatic, cycloheptatrienylidene, XVIII. Here the p orbital of the

PAGE 13

7 methylene is involved in tropylium-type resonance. XVIII The diand tribenzoderivatives, XIX and XX, of cyclo heptatrienylidene have been generated, and their low-tempera ture esr observed. The carbenes were reported to have trip let ground states and to react with butene-2 (35,36,37). Because of the involvement of the aromatic rings, these systems are apparently representative of aryl-substituted methylenes. Comparison of their properties with diphenyl methyl2ne, XXI, and 2,3,6,7-dibenzocycloheptadienylidene, XXII, makes this analogy even more obvious. Both XXI and XXII are reported by the same group to have triplet ground XIX xx XXI XXII

PAGE 14

8 states and to react with nucleophilic olefins such as butene2 The theoretical foundation for the foregoing qualita tive discussion has recently been set forth by R. Gleiter and R. Hoffmann (38) 1 A linear methylene has two degenerate orbitals in which to distribute the two carbene elec trons. The lowest energy state (ground state) of such a species is a triplet with the unpaired electrons in separate orbitals. To obtain a singlet ground state for a methylene, the degeneracy of these two orbitals must be destroyed, and, according to Gleiter and Hoffmann, the resultant splitting must be greater than 2 ev. They have postulated three factors which influence the multiplicity of the ground state and the relative nucleophilicity of the methylene. The carbene can be modified by 1) bending the R-C-R angle, 2) approaching or connecting the methylene to a system with low lying unoccupied levels, or 3) appro a ching or connecting the methylene to a system with high-lying occupied levels. The first wa involving bending of the R-C-R angle, distorts the p orbital in the plane of the bending (p) y while the second orbital (p) is virtually unaffected. As X the bond angle is decreased from 180, the distorted orbital acquires s character and is therefore stabilized. P will be y (1) Since the ensuing material has not yet appeared in print, it is discussed in much greater detail than would normally be the case.

PAGE 15

9 referred to as o in the following discussion. The splitting developed between a and p is not great at the higher angles X and becomes important in cases such as cycJ.opropylidene where the splitting is calculated to be 1.13 ev. The optimum case 0 is vinylidene where the angle is essentially O and the splitting is calculated as 1.92 ev. The second case concerns conjugation of p with lowx lying unoccupied levels. In this case, the methylene would transfer electron density to the TI system and thus become more electron deficient itself. The unoccupied orbital of the TI system must have the proper symmetry to mix with px METHYLENE II SYSTEM of the methylene to form a cyclic polyene. Thus, in all systems with 4n TI electrons, the lowest unoccupied molecular orbital has the correct symmetry to mix with p. Conversely, X in all TI systems with 4n + 2 TI electrons, the highest occupied nLolecular orbital has the proper symmetry. In the 4n case, stabilization of p does not increase splitting because a X {py) has also been stabilized by bending. In cyclopentadienylidene, the splitting was calculated as 0.13 ev. The symmetry of the orbitals is indicated by S (symmetric) and

PAGE 16

10 A (antisymmetric). A s A -I+--s --H---+rV Px In the third case, the methylene is conjugated with a high-lying occupied level. As mentioned above, in a TT sys tem with 4n + 2 TT electrons, the highest occupied molecular orbital has the correct symmetry to mix with p. This results X in a new TT system in which the lowest unoccupied orbital is higher than Px Furthermore, if a is stabilized by bending, the splitting is reinforced. The optimum case should be cyclopropenylidene where n = 1. The interaction diagram for the TT system is shown below. The o--p_, splitting amounts to ,I\. 3.17 ev which leads to a prediction of a singlet ground state. A s -~ s \ I C =:c I \ ~-H--Px

PAGE 17

J. 1 In the general case, there appears to be a possibility of stabilizing a singlet ground state when the methylene interacts with a 11 system which already contains 4n + 2 TI' electrons. A number of systems which fit this requirement are shown below. .. (N> H C>= I 0 N .. 3.17 ev 3.45 ev 1. 77 ev 1. 00 ev XXIII XXIV XXV XVIII Derivatives of the carbenes XXIII, XXIV, a11d XXV have been discussed and were found to react with electrophiles and electron-deficient olefins. There is nothing known about their ground states, however. The splitting for car bene XVIII is somewhat low and by the above standards would be predicted to have a triplet ground state. ~xperimental evidence will be presented, however, which will argue for a singlet ground state. The effect of the TI system on the nucleophilicity of the carbene must also be considered. Gleiter and Hoffmann chose to examine the total charge on the methylene carbon. There are two distinct cases: 2 2 1. p lower than cr : The two electrons of thecarbene are now delocalized through the TI system so that the methylene loses electron density. The carbene should be more electrophilic.

PAGE 18

12 .. HC CHO ( "1 inear") +0,33 XXVI +0.41 XXVII 2. a 2 lower than p 2 Electron density is donated to the methylen e center by the polyene which should make the carbene more nucleophilic. -"" (120) -0.50 (): -0.43 C--! / H C 6 Hc: H I (120) N :.i,c -0.56 -0.65 ,,, H t>: 0 -0.68 (h (bent 20 ) -0.43 0: -0.86 XXVIII Therefore, by the above arg-uments, cycloheptatrienyli dene should be nucleophilic, reacting preferentially with electron-deficient double bonds. The purpose of this paper is to report several attempts to generate cycloheptatrienylidene (39). The Bamford-Ste vens reaction which involves base-promoted decomposition of p-toluenesulf0nylhydrazones, and base-catalyzed decomposition of N,N-dimethyl-N'-nitroso-N'-cycloheptatrienylurea and methyl N-nitroso-N-cycloheptatrienylcarbamate was examined.

PAGE 19

CHAPTER II DECOMPOSI'l 1 ION OF 2, 4, 6-CYCLOHEPTNrRIENONE p-TOLUENESULFONYLHYDRAZONE AS A ROUTE TO CYCLOHEPTATRIENYLIDENE A. ~nthesis of 2, 4, 6-Cycloheptatrienone p-Toluenesulfonylhydrazone. A popular method of generating diazo compounds and car benes is the Bamford-Stevens reaction in which p-toluenesul fonylhydrazones, XXIX, are decomposed in the presence of base (40). The reaction is particularly useful since the decomposi tion can be effected thermally or photochemically under rela tively mild conditions. Since the hydrazones are ketone derivatives, availability of the properly substituted ketone R H ~C=N-Nso 2 Ii.r R XXIX 1) Base -----> 2) A ,:,r hv --':)R ,,.,, /"-. R is important. Tropone,XXX, is easily synthesized by sele nium dioxide oxidation of cycloheptatriene (41}, but because of a highly polarized resonance form, it does not always undergo normal ketone reactions. Treatment of tropone with hydrazine does not produce the expected hydrazone, but 2aminotropone instead (42). Recently, the imine and several 13

PAGE 20

14 substituted imines ha,ve been synthesized although somewhat indirectly (43,44). The phenylhydrazone (45) and azine (46) of tropone, however, have been made by classical methods; and fortunately, the p-toluenesulfonylhydrazone conformed to the latter cases. By stirring a solution of tropone and p-toluenesulfonyl hydrazine in 1% HCl in absolute ethanol, a yellow solid was obtained which was shown to be the hydrochloride salt of tropone p-toluenesulfonylhydrazone, XXXIII. The yield could be improved from 50% to above 90% by first converting the tropone to chlorotropylium chloride, XXXII (47), with thionyl chloride followed by treating an absolute ethanol solution of the chloride with a solution of p-toluenesulfonylhydrazine 0=0 + H 2 NNHS0 2 Ar XXX XXXI SOC12~ OCl Cl XXXII 1% HCl O=N-NHS02ArHCl ;>in Et.OH XXXIII t The nmr spectrum of the salt in n 2 o shows in addition to the methyl absorption, a multiplet partially superimposed on the downfield portion of the aromatic A 2 B 2 absorption. The position and structure of this multiplet is characteris tic of tropone, tropylium salts, and monosubstituted

PAGE 21

15 tropylium salts and indicates a high d e gr e e of tropylium ion character confirming the structure assigned to XXXIV. XXXIV When an aqueous solution of XXXIV was titrated with dilute base, the light yellow solution became cloud then rapidly cleared to form a d~rk red solution. The pH was adjusted until the cloudiness remained, and the orange mix ture was extracted with dichloromethane. After drying the deep red solution, pentane was added, and the neutral toluenesulfonylhydrazone, XXXV, was obtained as bright red crystals. Two crystalline forms were found; rapid crystal 0 lization produced deep red plates, mp 142.5-143.5 while 0 slow crystallization formed clusters of needles, mp 144-145 Their and nmr spectra showed both forms to be the same compound. From the behavior of the p-toluenesulfonylhydrazone in the presence of acid and base, it was apparent that there was yet another species present in strongly basic solution. The acidity of the nitrogen proton in p-toluenesulfonylhy drazones is well known, and is in fact, the basis of the Bamford-Stevens decomposition reaction. Therefore, it was not unexpected to find this third species, XXXVI, which is the conjugate base of XXXV. All three species are quite

PAGE 22

16 stable and are re a dily interconverted with acid or base. Cl H H -N-1'1-S0 2 Ar Yellow HCO 3 # H+ XXXIV O=N-~--S02Ar ~:o-N-~a+S02Ar Red Brown XXXV XXXVI Several attempts were made to isolate XXXVI in a pure form, but none were successful. The salt typically preci pitated from solution as a muddy brown solid which clung tenaciously to the last traces of solvent and decomposed before drying completely. To circumvent this problem in reactions where the dry salt was needed, an aprotic solvent, tetrahydrofuran, was used to generate the species. Almost any base was found to be satisfactory for removing the pro ton of the p-toluenesulfonylhydrazcne, but sodium hydride was chosen because of the lack of any contaminating bypro ducts. Despite reports to the contrary (48), this base was found to be entirely satisfactory. B. Decomposition of 2,4,6-Cycloheptatrienone p-toluenesul fonylhydrazone Sodium Salt. Decomposition of the sodium salt of a p-toluenesulfonyl hydrazone is believed to proceed through the corresponding diazo compound which is frequ e ntly an isol a ble interm e diate.

PAGE 23

17 The stability of the diazo compound i s largely dete r mined by the groups attached directly to the a-carbon atom (49). Groups which can conjugate directly with the diazo carbon tend to stabilize the molecule, and strongly electron withdrawing substituents, such as trifluoromethyl, can give rise to highly stable diazo compounds. Bis-(trifluoromethyl) diazomethane, XXXVII, and diazocyclopentadiene, XXXVIII, are two well-known examples. CF @-Ni O=N2 3,c_ ... ,+ / -1~2 CF 3 XXXVII XXXVIII XXXIX If the groups are electron-donating, however, the diazo compound may be unstable (50). Diazocycloheptatriene, XXXIX, would fall in this category, and relative to diazocyclopenta diene, should be quite unstable. D. G. Farnum has reported a variation of the Bamford Stevens reaction in which good yields of diazo compounds were obtained (51). A pyridine solution of the p-toluerie sulfonylhydrazone and sodium methoxide is heated gently, then poured into water, and extracted with pentane. When this reaction was applied to XXXV, only starting material was recovered from the reaction mixture. In another attempt to generate XXXIX, a sample of the sodium salt, XXXVI, was placed on a column of basic alumina. Elution with aprotic solvents gave no r e sults. Elution with ethanol caused pro tonation of XXXVI, and a quantitative yield of XXXV was

PAGE 24

18 obtain e d as a red band. Num e rous att e m p ts w e re m a de to trap a diazo int e rm e diate on a liquid nitrogen cold finger from thermal decomposition of X X XVI under high vacuum without results. Also, attempts to find evidence for a pyrazoline intermediate when the d e composition was carried out in the presence of olefins were unsucc e ss f ul. These results, how ever, do not rule out the possibility of an inermediate diazo compound. When a slurry of the sodium salt, XXXVI, in tetrahydro furan was photolyzed, about 90 % of the theoretical gas evolu tion was observed. Filtration of the dark red reaction mix ture gave up 95% of sodium p-toluenesulf inate as a light brown powder. The filtrate was poured into water and extracted with pentane. After drying and concentrating the extracts, a black, oily solid remained. The uv spectrum showed absorp tions at 234 and 362 m which were unaffected upon short expo sure to acid. After several minutes with acid, the absorp tions gradually disappeared. A tlc of the residue from the reaction mixture showed a fast moving red spot in 25% ether in hexane and a slow moving red spot attributable to unreacted p-toluenesulfonylhydrazone. Upon standing on the dry plate, the fast moving red spot decolorized within five minutes. Since the compound was apparently decom p osing, deactivated basic alumina was used for column chromatography. Elution with 5% ether in hexane gave a sharp red band. Concentration of the deep red elutions under a stream of dry nitro g en

PAGE 25

19 gave permanga.nate -colored plates (mp 114 116) which wer e identified as heptafulvalene, XL. Recrystalliz a tion could be effected f r om methanol water, and a pure sample was obtained with mp 119-121. Solutions of the dark crystalline solid were de e p red in color but decolorized within several days, while the pure crystallin e solid decomposed within one hour when exposed to air. The nmr spectrum of the solid showed only a multiplet in the olefinic region, and the mass spectrum sho w ed a molecular ion of m/e 180. The properties of the compound were identical with the properties as reported by Doering (52). This olefin is the dimer expected from dimerization of cycloheptatrienylidene. XXXVI hv -->XL + + Molecular orbital calculations on heptafulvalene show two electrons in a low-lying antibonding 1':IO (53). Conse quently, the compound should be rather unstable and highly susceptible to oxidation. This was found to be true since several attempts to obtain an analysis of the dimer were unsuccessful. Although it could be purified easily by recrys tallization or sublimation, heptafulvalene rapidly decomposed in the crystalline state. A sample kept in the dark, under nitrogen, and at --10 still decomposed within two weeks.

PAGE 26

20 An attempt to obtain an analysis in this labor a tory on a freshly purified sample was also unsuccessful. The olefin could be kept for extended period s in pentane solution at dry ice temperature and under nitrog e n. Because of this instability, acceptable yields of the dimer could not be obtained by weighing the dry, crystalline solid. Since the 362 m peak of heptafulvalene was free of interfering absorptions, it was used to determine the concentrations of accurately diluted solutions of the dimer and thus the yields. The extinction coefficient of this absor~ tion was reported as 21000 (52), but values obtained in this laboratory on freshly purifie~ samples were 21400, 25000, and 26000. The lowest value was probably the least accurate because of the length of time between purification and deter mination of s. The value selected for determining yields was 25000. The value of the yields for corn?arison purposes is questionable at any rate, because the dimer was also found to be photolytically unstable, and the yields were very much dependent on the length of the photolyses. The dimer yield does give a minimum for the extent of reaction. C. Decomposition of 2,4,6-Cycloheptatrienone p-Toluenesul fonylhvdrazone Sodium Salt in the Pre$ence of Dimethyl F'-1.ma rate, Fumaronitrile, a.nd Maleinitrile. Formation of Spiro[:2.6] nonatrienes. When nucleophilic olefins such as cyclohexene and butene-2 were added to the reaction mixture in fiveto ten fold excess in an attempt to trap the carbene or the diazo compound, there was no noticeable change. Gas evolution was

PAGE 27

21 still within 90 % of theoretical, and after the same workup, the dimer yield was between 40 and 70%, No evidence of an adduct between the carbene and these olefins was found. This was not unexpected,for if the carbene was actually nucleophilic in nature, it would not be expected to react with nucleophilic olefins but with those which contained electron-deficient double bonds. When the electron-deficient olefin, dimethyl fumarate, was added to the reaction mixture, gas evolution was still normal; but at the end of the photolysis, the normally dar~ red solution was almost colorless. Filtration of the mix ture gave a high yield of sodium p-toluenesulfinate which showed that the decomposition had proceeded as expected. The reaction mixture was again poured into water and extracted with pentane. After drying and concentrating the solution, tlc showed a small quantity of the dimer and one other colorless spot which upon standing on the dry plate turned bright yellow. Since this color change was suspected to be due to an acid catalyzed rearrangement, basic alumina was used for chromatography; and to shorten the elution time, the adsorbent was deactivated to activity grade III. Careful cooling of the pentane solution before chromatogra p hy allowed removal of most of the unreacted dimethyl fumarate by frac tional crystallization. Elution with 5% ether in hexane gave the dimer as a dark red band in low yield (0-10%). By gradually increasing the ether content to 10 % the compound observed on tlc was eluted immediately before dimethyl

PAGE 28

22 fumarate. Ev a poration of the solvent gave white needles with mp 65-68 and mixture mp with dimethyl fumarate 63-97. The compound was identified as the spirononatriene, XLI, which formally, is the adduct of cycloheptatrienylidene and dimethyl fumarate. Their spectrum showed a carbonyl absorp tion at 1720 cm -l and the uv spectrum showed an absorption at 261 m which became 341 m with acid and 341 and 253 m with subsequent treatment with triethylamine. The nmr spec trum showed a multiplet from T3,20 to 3.83, a doublet cen tered at T4.55, a singlet at T6.31, and a singlet at T7.69 with area ratios 4:2:6:2. The mass spectrum showed a parent peak at m/e 234, and analysis indicated c 13 H 14 o 4 as the empirical formula. XLI As mentioned above, the spot due to the spiro compound on a tlc plate turned bright yellow upon standing. Also, the behavior of the uv spectrum of XLI upon treatment with acid and then base (vide supra) indicated a rearrangement rather than a simple protonation. In the analogous diphenylcyclo propenyl system the carbene dimethyl fumarate adduct, XLII, underwent a facile, acid-catalyzed rearrangement to the tri afulvene, XLill (31,32,33). It was therefore suspected that a

PAGE 29

similar reaction was taking place with XLI. Ph COOCH 3 Ph~COOCH 3 XLII XLI ---->XLIII XLIV 23 Accordingly, a sample of XLI in anhydrous ether was placed on a column of dry acid alumina, activity grade I (alumina of lower activity gave poor conversion if at all), and eluted onto the column with a small quantity of ether. Upon contact with the alumina, the colorless solution turned bright yellow. Elution of the yellow band with acetone and concentration of the solution gave a dark oil. By cooling a pentane solution of the oil in a dry-ice-acetone bath, a fluffy, yellow solid was obtained which became an oil again at room temperature. Upon slowly cooling the solution to -20, yellow needles of the heptafulvene, XLIV, were for m ed which after several recrystallizations from pentane gave mp 0 49.5-51. Their spectrum showed a pair of carbonyl peaks -1 at 1750 and 1705 cm and the nrnr spectrum showed a doublet at T2.57, and a multiplet at ca. T3.60 for the six cycloheptatrienyl protons. There were two peaks for the nonequivalent

PAGE 30

24 methyl esters at T6. 29 and 6. 34 and a two proton sin g let at T6.65. of XLI. The mass spectrum was aJ.most identical with that The degree of stereoselectivity of the addition of a carbene to an olefin is frequently used as evidence of the multiplicity of the carbene as it adds (1). Consequently, it was highly desirable to det e rmine the steric course of the photolysis of the salt, XXXVI, with dimethyl fumarate. If the spire compound from the reaction with the trans olefin has the trans configuration, as in XLia, then perhaps the cis olefin, dimethyl maleate, would produce the cis spirononatriene, XLib. H COOCH 3 H COOCH 3 XLia XLib Photolysis of the sodium salt, XXXVI, in the presence of dimethyl maleate also produced the spiro compound in comparable yield. This sample was identical in every way with that obtained from the reaction with dimethyl fumarate. If this were one isomer of a cis-trans pair, some differences in the nmr spectrum would be expected. For example, the posi tions of the cyclopropyl protons in the two isomers below differ by 0.30 ppm (54).

PAGE 31

25 A possible explanation for the failure to find a second stereoisomer arose when it was found that dimethyl maleate rapidly isornerized to the trans isomer under the reactio~ conditions. Either base or uv light caused the isomeriza tion to a final equilibrium mixture consisting of ca. 10% cis and 90% trans olefin. hv <.. If the addition of the carbene to cis and trans olefin ___,_ __ proceeded at the same rate, then a small amount of cis spiro nonatriene would be expected. R. Huisgen has shown, however, that the addition of 1,3-dipoles to dimethyl fumarate and dimethyl maleate is much faster with the trans isomer (55). The difference in reactivity is attributed to steric hinder ance to resonance in the cis isomer, which weakens the reac tivity of the olefin with nucleophiles, and to increased steric repulsion of the cis ester functions as they change 2 3 from sp to sp The use of other electrophilic olefins was limited primarily by their instability to basic conditions. Tetra cyanoethylene, a strongly electron-deficient olefin, is

PAGE 32

26 quite unstable to bases and decomposes rapidly in the pre sence of sodium hydride. Fumaronitrile, XLV, is also unsta ble to bases, but the decomposition is much slower. When the photolytic decomposition of the salt, XXXVI, was carried out with a fivefold excess of this olefin present, the nitrogen evolution was normal. The reaction mixture was dark red, indicating the presence of heptafulvalene as at least one of the reaction products. A tlc of the mixture confirmed the presence of the dimer and showed one other major spot. Upon standing for several minutes on the dry plate, the latter spot turned bright yellow as did the car bene adduct with dimethyl fumarate. After removal of the tetrahydrofuran, nmr of the residue showed a series of multi plets in the olefinic region quite similar to XLI. A sing let at l8,02 completed the pattern expected of the carbene adduct with the olefin. The dark residue was chromatographed on deactivated basic alumina to give the dimer (18.5 % ) with 5% ether in hexane. Upon increasing the solvent polarity to 30% ether in hexane, the spot observed earlier on tlc was eluted. Concentration of the eluted fractions gave the new compound as a mixture with fumaronitrile. The olefin was easily sublimed from the mixtur~ leaving the crude spiro nonatriene, XLVIa, with mp 110-115. Recrystallization from benzene-pentane to constant melting point gave white needles with mp 121-123. An nmr sp~ctrum of the pure material showed only those absorptions described previously. The uv spectrum showed an absorption at 262 m~ and their spectrum

PAGE 33

showed a nitril e peak at 2250 -1 cm 27 The mass spectrum contained a molecular ion at m/e 168, and analysis indicated c 11 H 8 N 2 as the empj .r ical formula. This compound was later shown conclusively to he the t rarrs isom e r. NC XXXVI + "c=c "en XLV CN hv ----;,, XLVIa As mentioned previously, this spirononatriene also turned bright yellow on a dry tlc plate. Since this was indicative of an acid-catalyzed isomerization analogous to that of XLI, a sa~ple of XLVIa was placed on a dry column of acid alumina (activity grade I) and eluted onto the column with ether. The solution turned orange upon con tact with the column. Elution of the rearranged product with acetone gave a red solution which was concentrated to ca. 1 ml under a stream of nitrogen. Pentane was added, and red crystals of the heptafulvene, XLVII, were obtained which were filtered and dried, mp 97-99. Recrystallization from 0 benzene-hexane gave red plates, mp 99.5-100.5 Analysis and mass spectral data indicated that this compound was iso meric with XLVIa. The uv spectrum showed absorptions at 247 and 346 mi-t, and their spectrum showed nonequivalent nitriles at 2255 and 2195 crn-l confirming the assignment.

PAGE 34

28 XLVIa XLVII At this tim e there was still very little known about the stereo-chemistry of the addition proc e ss; and before any real conclusions could be drawn, the cis spiro compound had to be isolated and characterized. Efforts in this direction, in the case of the earlier spire compound, had been to no avail because of the equilibration of the cis and trans olefin pair under the reaction conditions. It was hoped, therefore, that this problem would not arise with the new olefin. Accordingly, the cis isomer of fumaronitrile, ma.Leinitrile, was tried in the photolysis reaction. Again, a tlc of the crude reaction mixture showed the presence of a spiro compound. F.xamination of the nrnr spectrum of the oil after removing the solvent showed, that indeed, the cis olefin had not equilibrated with its trans isomer during the reaction. The position of the cyclopropyl protons of the spironona triene from this reaction appeared to be 2-3 cps downfield from the analogous absorption in the trans case, but the conclusive evidence for the cis isomer lay in the absorptions due to the 2 and 7 protons of the cycloheptatriene ring. In the trans isciner, XLVIa, these protons are equivalent, each in the vicinity of a proton and a nitrile grou p on the

PAGE 35

29 cycloprop y l rin g Consequ e ntly, they appear as a broad dou blet split by the 3 and 6 p rotons and further by the 4 and 5 protons. In the cis isomer, howev e r, the 2 and 7 protons are nonequivalent. One is on the side of the spiro cyclo propyl ring with two protons, and th e other is on the side with two nitrile grou p s. Th e refore, in the nmr spectrum of the cis isomer, there are two distinct multiplets for these protons, each integrating to a relative area of one. cti ~ ,: ., r-~ fiN 1 H H XLVIa XLVIb Their spectrum of the cis isomer, XLVIb, is quite similar to that of XLVIa, but at the same time, it is obvi ously nonsuperimposable. The uv and mass spectra of the t w o isomers are identical. Further proof of the iso m eric nature of XLVIb is found in its isomerization to a heptafulvene which was identical by ir and glc with XLVII, the heptaful vene obtained from XLVIa. The maleinitrile which was used in this reaction was contaminated with 13% of the trans isomer. The nmr spectrum of the crude reaction mixture described above sho w ed that this ratio had changed little if any during the course of the reaction. A close examination of this s pe ctrum sho w ed a

PAGE 36

30 small but sharp si n glet at the position of the cycloprcpyl protons in the trans iso m er, XLVIa. Although the region around this absorption was much too noisy to allow integra tion for comparison purposes, the peak area could be esti mated as ca. 10% of the peak area for the cis isomer. Thus, it appears that the carbene addition to this olefin is highly stereoselective if not stereospecific. H NC XLVIa XXXVI CN H ~+ XLVIb 0 CN =C~CH-CN 2 D. Mechanism of the Formation of Heptafulvalene and the Spiro]?: ____] nonatrienes. This brings up the question of how heptafulvalene and the spirononatrienes are formed. Formation of the latter can be envisioned as proceeding by se v eral different routes. Since the olefins used are electron d e ficient, they are

PAGE 37

31 quite susceptible to base attack. Addition of the anion of the p-toluenesulfonylhydrazone across the double bond, fol lowed by cyclization to the pyrazoline, XLVIII, and decompo sition to XLI by loss of nitrogen is certainly a rational mechanism. XXXVI + XLI -N 2 -<.. XLVIII Participation of the olefin before or during the rate determining step can be excluded by a simple kinetic scheme. The salt, XXXVI, was found to undergo smooth thermal decompo sition above 100 to give the same products as the photolytic decomposition. Thus, two reactions were carried out under identical conditions in which the salt was decomposed ther mally at 101.5 (refluxing dioxane) in dry diglyme. One reaction contained a twofold excess of dimethyl fumarate and the other contained none. Nitrogen evolution was followed as a measure of the rates of the two reactions. If the ole fin was involved in a step before or during the rate-deter mining step in the decomposition, its presence should then cause a noticeable increase in the rate of the reaction if

PAGE 38

32 XLI is a major product which forms at the expense of the dimer. The rates of the two reactions were virtually iden tical with half-lives of approximately 11 min. In the reac tion without olefin, the yield of heptafulvalene was 33%; and in the reaction with dimethyl fumarate, the yield of XLI was 38%, while the dimer yield dropped to 9%. Reaction of heptafulvalene with dimethyl fumarate in a reaction analogous to that proposed by Lemal (6) to explain the products of reaction of electrophiles with tetraamino ethylenes (vide supra) is also a possibility as the source of XLI. This path was excluded by photolysis of a solution of heptafulvalene and the olefin in tetrahydrofuran. Although the dimer decomposed under the photolysis conditions and a 0=0 ---------'>Dimethyl Fumarate hv Decomposition colorlesB solution remained, no trace of the adduct, XLI, could be found. The only isolable product was dimethyl fuma rate. Another possible route involves diazo addition to the olefin, followed by rapid decomposition of the resulting pyrazoline. ri'his reaction sequence is quite common in car bene chemistry, and, in a number of cases, both the diazo compound and the pyrazoline are isolable intermediates (1,2).

PAGE 39

33 XXXVI O< ~ N ~T XT I c-c-coocH.,. ,J COOCH 3 3 Unsuccessful attempts to isolate diazocycloheptatriene have already been described, and an attempt to intercept a 1pyrazoline intermediate was likewise unsuccessful. It has been shown ( 56) that unstable 1--pyrazolines can be inter cepted by conversion to the more stable 2-pyrazolines in the presence of added base. Isolation of the 2-pyrazoline at the expense of the cyclopropane is evidence that the 1-pyra zoline is an intermediate in formation of the cyclopropane. -N 2 ----;;:,R~X R~y Accordingly, the salt, XXXVI, was photol.yzed in the presence of dimethylfumarate and a tenfold excess of triethyl amine. Nitrogen evolution was in excess of 80% and was evi dence, in itself, that little or no pyrazoline was being trapped. The reaction mixture was worked up in the usual mann e r to give a 16% yield of heptafulvalen e and a 49 % yield

PAGE 40

34 of the spiro compound. No evidence of a 2 ~ pyrazoline was found. This does not elimnate the possibility of the dia zo pyrazoline route to XLI; but if the pyrazolirie is actually an intermediate, it is apparently rapidly decomposed under the reaction conditions. Another likely route involves decomposition of the sodium salt to an unstable diazocycloheptatriene which rapidly loses nitrogen to form the carbene, or alternatively, concerted decomposition of the salt to give cycloheptatri enylidene directly, whic~ in either case, adds to the elec tron-deficient double bond to form XLI. _J ,----)',XLI There has been a variety of debate on the assignment of multiplicities to carbenes on the basis of sterospecifi city or nonsterospecificity of their addition to olefins (1). Skell has pointed out that a singlet carbene can undergo concerted (and therefore, stereospecific) addition without violating spin conservation principles (57). A triplet, how ever, cannot undergo addition without a spin-inversion cess and may proceed at first to a diradical species such as

PAGE 41

35 XLIX. If the spin-inversion process is slow compared to rotation about the carbon-carbon single bond, then the end R'\ 1 C >-~ R / J.) s. I. R><] R/C1 ~--c C ?/ 1 R 1 2) ring R XLIX closure result would be a nonstereospecific addition. Consequently, nonstereospecificity is generally considered to indicate triplet carbene, and stereospecificity is assigned to sing let carbene. This argument is not without weaknesses, how ever, for Gaspar and Hammond (58) have warned that since the relative rates of spin inversion and carbon-carbon bond rota tion are not known, it is quite conceivable that the latter might be the slower process. Conversely, there is no assur ance that a singlet carbene will add in a concerted fashion just because there is spin conservation. It is possible, therefore, to envision a stereospecific triplet addition and a nonstereospecific singlet addition. In the case of cycloheptatrienylidene, there is a good possibility that singlet addition to an olefin could proceed through a zwitterion such as L to give nonstereospecific addi tion products. The addition process has been shown to be stereospecific, however; so if Lis indeed an intermediate, then ring closure must be faster than rotation about the single bond.

PAGE 42

36 R R + \ __ / --~; ... R --> Product's L It is generally conceded that carbenes are generated in a singlet spin state, although this is not necessarily true (59). The singlet carbene then either reacts as a singlet or undergoes intersystem crossing if there is a more stable triplet state available. In cases of this type, added inert solvent should decrease the overall degree of stereo specificity by providing more chances for collisional deacti vation to the more stable triplet. M. Jones and K. R. Rettig have studied a case where this apparently occurs (60). They irradiated 9-diazofluorene in cis-2-butene and observed the cis/trans ratio of the spire products as the reaction was diluted with hexafluorobenzene. They observed a rapid decrease in this ratio with dilutio~which they attributed to an increase in intersystem crossing to produce the triplet which then added nonstereospecifically to the cis-2-butene. Addition of 50 mole% of hexafluorobenzene was sufficient to decrease the cis/trans ratio to 0.6. In an analogous case, Jones and coworkers irradiated dimethyl diazomalonate in the presence of an inert solvent and observed the degree of ste reospecificity of addition of the carbene to ~i~-4-methyl-2pentcne (61). Above 95 mole% of hexafluorobenzene, they

PAGE 43

37 observed a sharp decrease in the amount of cis product,which they again attributed to triplet production. Therefore, from the two examples above, it might be expected that carben e s with triplet ground states can undergo intersystem crossing from the initially produced singlet under conditions of high dilution. Considering tetrahydrofuran as an inert solvent in the photolytic decomposition of the salt with the cis olefin, maleinitrile, the mole% of the inert solvent can be calculated as 99.3 %. Since the reaction is still highly stereoselective at this point, it can be assumed that the carbene is reacting in its ground state or that the transition to the ground state is strongly forbidden. Since the addition process is stereospecific, any transition state for the addition must not allow rotation before ring closure. At the same time, a species such as the zwitterion L, should be stable and could easily arise in the addition process. Perhaps a combination of these can produce a logical intermediate. A transition state involving much greater bonding on one side of the form ing three-membered ring, such as LI, would still incorporate stereospecific addition. LI

PAGE 44

38 rhis "nonsynchronous concerted addition II v1ould cor,respond to addition of a 1,1-dipole to an olefin. Low-temperature nrnr studies of cycloheptatriene have shown that the ring inversion process can be stopped at -150 indicating an activation energy of 6.3 Kcal/mole (62). The spiro compound, XLI, could have a lower E because of the a spire three-membered ring. The c 7 c 1 c 2 angle in the sevenmembered ring of XLI would be greater than the corresponding angle in cycloheptatriene. In the limiting case, where E = O, the seven-membered ring of XLI would be planar. The a low-temperature spectrum was examined from -30 to -95, and as expected, there was no change. Unfortunately, lower temperatures could not be obtained. V a ---.....!I,,,,. ----=-The spirononatriene, XLI, did not undergo Diels-Alder reactions with common dienophiles as do some derivatives of cycloheptatriene (63). A potential route to a spiropentene involving a Diels-Alder addition to XLI was therefore pre cluded (64). E. Other Olefins as Cycloheptatrienylidene Acceptors. The list of electrophilic olefins tried as trapping agents is quite short. The presence of strongly electron withdrawing goups on the olefin makes it quite unstable to

PAGE 45

39 basic conditions which are necessary for the generation of the carbene. In a number of cases where a tenfold excess of the electron-deficient olefin was used, not a trace could be found at the end of the photolysis. Perhaps a number of the olefins listed would easily form adducts with the carbene if it could be generated under less basic conditions. Maleic anhydride, fumaramide, methyl cinnamate, dimethyl acetylenedicarboxylate, methyl B-acetylacrylate, and the mixed methyl ester-dimethylarnide of fumaric acid were all tried, but none gave evidence of a spiro compound. It should be mentioned, however, that in each case except fumaramide, methylcinnarnate, and dimethyl acetylenedicarboxylate, there was no olefin at the end of the photolysis. This could mean that the olefin completely decomposed before the photolysis was well under way or alternatively, the spire adduct, like the olefin, was unstable to the reaction conditions. The addition of cycloheptatrienylidene to olefins sug gests a potential source of [!n,6] spirarenes (65,66). The spirarenes consist of two orthogonal pi systems joined by a common spire carbon in which m and n represent the p orbitals in each ring. There is no easy synthetic route to these com pounds, and only a few have been made. Cycloheptatrienyli dene would form one of the pi systems, and it would become

PAGE 46

40 attached to an olefin by a spiro carbon. If suitable accep tors could be found, then ~. 6] [4. 6] and [6. 6] spirarenes could be synthesized. The spirarene system involves perhaps the easiest synthesis to visualize. Addition of the carbene to an acety lene to give the spiro ~-~nonatetraene, LII, would be a direct route. Dimethyl acetylenedicarboxylate was tried as a trapping agent, but no spiro compound was isolated. There + 0 : +j(--1 COOCH 3 900cH 3 C HI C I COOCH 3 --~LIII Cl LV COOCH 3 LII NaI >r1 ~\,,1 1 acetone base '> LVI LIV COOCH 3

PAGE 47

41 was no heptafulvalene produced in this reaction which may indicate formation of the spirarene did occur, but the com pound subsequently decomposed. An indirect synthesis could involve addition to a halogen-containing olefin followed by dehalogenation or dehydrohalogenation. 1,2-Dibromoethylene was tried with the hope that the adduct, Lill, could be debromi nated to LIV, but again no spire compound was obtained. l!ep tafulvalene was the only product in a yield of 44%. A third attempt at formation of a D.Q spirarene involved methyl trans-B-chloroacrylate. Dehydrochlorination of the adduct, LV, would give the desired system, LVI, but here also, the adduct was not detected, and the dimer (39%) was the only product. Synthesis of the remaining two spirarenes is not as easy to imagine. A possible route to the system lies in a 1,4 cycloaddition reaction between the carbene and a diene followed by conversion of the resulting cyclopentene ring to a cyclopentadiene ring to give LVII. The addition process would be the reverse of several symmetry-allowed fragmentation reactions studied by Lemal (67,68). Work is still in progress on this system. R R R R R LVII

PAGE 48

42 An interesting entry into the spirarene system could be provided by benzoquinone. Conversion of the ene dione, LVill, to the di.ene would be followed by conversion to the triene, LIX, by analogy with the norcaradiene-cyclo heptatriene system. Unfortunately, the instability of ben zoquinone to base precluded this reaction. 0 II + o-r-1 fl 0 LVIII LIX F. Decomposition of 2,4,6-Cycloheptatrienone p-Toluenesul fonylhydrazon~ Sodium Salt under tigh Vacuum. A convenient method of preparing some diazo compounds involves pyrolysis of the dry sodium salt of the p-toluene sulfonylhydrazone of the corresponding ketone under high vacuum (69). As the diazo compound is formed, it distills out of the reaction mixture into a cold trap in yields up to 95%. If the analogous pyrolysis of XXXVI would give diazo cycloheptatriene or possibly even cycloheptatrienylidene, perhaps it could be trapped at low temperature. The reaction vessel employed was a 500 ml boiling flask fitted with a cold finger. The salt, XXXVI, was generated from the p-toluenesul fonylhydrazone, XXXV, with sodium hydride in tetrahydrofuran. Removal of the solvent in vacuo gave the salt as a brown, muddy solid which still contained traces of tetrahydrofuran.

PAGE 49

43 After heatina to 50 overniaht at ca. J J 10 nun of Hg, there still remained appreciable quantities of solvent. Since higher temperatures led to slow decomposition of the salt, it was used without further purification. Best results were obtained when the solid was distributed as a thin, uniform film over the inner surface of the reaction flask. The decomposition could be carried out photochemically or thermally to give the same results. Phtolysis of the solid was generally much slower, however, and only the outer layer of salt decomposed. In contrast, the thermal reaction was much smoother. When the temperature was grad o ually increased to 150 the decomposition proceeded slowly as evidenced by a gradual increase in pressure as nitrogen was evolved. If the reaction vessel was quickly lowered into a bath at 150, a rapid increase in pressure was noted, but here also, the reaction proceeded quite smoothly with little or no "bumping''. Decomposition was complete within 2-3 sec in this case. When a dry-ice-acetone cold finger was used, either rapid or slow decomposition of the salt gave at first a white coating of tetrahydrofuran. As the decomposition pro ceeded, the cold finger became discolored so that at the end of the reaction, a smooth, deep red coating was observed. After admitting an atmosphere of dry nitrogen, the finger was allowed to warm to room temperature. There was no color change during the entire warming period. Had the diazo com pound been trapped, considerable distortion of the smooth

PAGE 50

44 surface would have been expected as nitrogen was evolved. Examination of the material on the finger showed only hepta fulvalene and tetrahydrofuran. Sodium p-toluenesulfinate accounted for most of the solid residue in the reaction flask. The formation of hepatfulvalene probably proceeds through the intermediacy of diazocycloheptatriene or cyclo heptatrienylidene rather than by dimerization of the salt, XXXVI (vide supra). The heptafulvalene must be formed in one of three places; the dimerization process can occur 1) in the dry salt as the reactive intermediate is genera ted, 2) in the space between the residue and the cold finger, and 3) on the cold finger either before or after the ''cool ing process" occurs. The first possibility involves a heterc geneous reaction and the second, a gas phase reaction. In the third case, if the carbene or the diazo compound conden ses on the finger, there is the possibility that it may react while still "hot" or vibrationally excited, or that it may 11 cool 11 somewhat before reaction. If the latter case actually occurs, then trapping the reactive intermediate at lower temperatures may prevent subsequent reaction. With this in mind, the reaction was repeated using liquid nitro gen in the cold finger. Slow decomposition produced a smooth, light brown coat on the finger. After nitrogen was bled into the system, it was allowed to slowly warm. Soon after the liquid nitrogen in the finger evaporated (ca. -150), noticeable darkening

PAGE 51

45 of the smooth coat was observed, and again only heptafulva lene and solvent were found. No gas evolution could be detected on the cold finger which might indicate decomposition of a diazo intermediate, but this does not preclude this possibility. A reason a ble explanation, however, is that the carbene actually condensed on the finger and did not dimerize until warmed somewhat. The dimerization process would cause the darkening observed. Several attempts were made to trap the reactive inter mediate with dimethyl fumarate. The olefin was sublimed onto the finger through small orifices during the entire decornposi tion to provide intimate contact with the intermediate. Although darkening of the finger was again observed as the system was warmed, no spire compound could be found. This was not entirely unexpected, for at the low temperature, the dimerization process may well be much faster than addi tion to the olefin. A possible alternative involves the use of an olefin which is a liquid at dry-ice temperature, so that as the intermediate comes in contact with the finger, it can dissolve and undergo addition in the solution phase rather than the solid phase as in the case of dimethyl furna rate. 1,2-Difluoroethylene and l,1,1,4,4,4-hexafluoro-2butene are electron-deficient olefins which are being consi dered. Further work i n this direction is anticipated.

PAGE 52

CHAPTER III ~l'HERMAL REARRANGEMENT OF 8, 9-DICARBOMETHOXYSPIRO [2. G] NONA-2, t1, 6-'ERIENE rro 1,2-DICARBOMETHOXYINDANE During the search for a second isomer of the ~piro nonatriene, a sample was injected into a glc. On a 1/8" X 5' 20% SE30 on 60/80 das ChromZ column, a sample of XLI gave two peaks with approximate retention times of 8 and 10 min. The ratio of the peak areas in order of their elution was 0.052:t.oo, and repeated recrystallization of the spire compound produced no change in this ratio. If these two peaks were due to cis and tr?ns XLI, some change in this ratio would then be expected. When a solution of XLI was treated with acid, rearrange ment to the heptafulvene occurred. A sample of this solution showed only a single peak due to the rearrangement product. Both of the above two peaks had disappear~d indicating that both were unstable to acid. A possible explanation for the two peaks lay in the isomerization of the pure isomer of XLI to a mixture of cis and trans isomers of XLI in the injection port of the glc. This would explain why recrystallization failed to change the area ratio. Isolation of the two com ponents by preparative scale glc was the next obvious step. 46

PAGE 53

47 Reasonable separation of the two peaks was obtained on a 1/4" X 5' column of the material described previously at 150 and 50 psi. From the larger of the two peaks, a white crystalline solid was obtained with mp 70-72. Charac teristics of the compound were completely different from XLI, and on the basis of the physical properties given below, it was assigned the structure of trans-1,2-dicarbomethoxyindane, LXa. The nmr spectrum showed an aromatic multiplet of four protons at T2.70, a doublet comprising one proton at 15.48, and a multiplet of three protons between 15.85 and 6.85 partially obscured by nonequivalent methyl ester peaks at 16.17 and 6.25. The mass spectrum was identical to that of XLI, and it is very likely that XLI underwent the same ther mal rearrangement in the inlet of the mass spectrometer. The structure of LXa was confirmed by comparison with an authentic sample synthesized according to Cook and Stephenson (70) COOCH 3 II COOCH 3 COOCH 3 H LXa LXb A small quantity of the minor peak was obtained as a 50:50 mixture with LXa and appeared to be similar since ir and uv of the mixture were identical to those of the trans

PAGE 54

48 indane. Hass spectra. of the two peaks obtained as they were eluted directly from the column into the mass spectrometer were also superimposable, so the minor peak was assigned the structure of the cis indane, LXb. This provides a logical explanation for the behavior of XLI on glc analysis. Apparently, pure trar:i,~ XLI rearranges thermally in the injector to a 95:5 mixture of the trans and cis indane. This isomerization can also be followed conve niently by high-temperature nmr. At 95, XLI is apparently stable indefinitely in heptane, but at 131 in n-decane, the reaction is 90% complete in 3.5 hr. The mechanism of this reaction is a source for interesting conjecture. A logical sequence of events would involve rupture of the three-membered ring and reclosure to form a new ring at a different site either concerted or in a step wise fashion with concomitant or subsequent arornatization. XLI COOCH 3 COOCH 3 L COOCH 3 COOCH 3 LX ~-COOCH, .., H LXI

PAGE 55

49 Such a sequence could involve an intermediate such as LXI which could aromatize by a concerted hydride shift-bond migration step. The formation of LXI would involve bond migration from c 1 to c 7 and at first appears to be an interesting example of a 1,7-sigmatropic rearrangement. However, a suprafacial 1,7 shift is symmetry forbidden by a concerted thermal pro cess involving a symmetrical orbital on the migrating atom (71). J. A. Berson (72) has shown that an antisymmetric orbital can be involved in a thermal, suprafacial 1,3 migra tio~ and the same should be true for a 1,7 shift. This results in inversion of configuration at the migrating atom. --->If the thermal conversion of XLI to LXI could occur via a sigmatropic shift, it must be either antarafacial or an inversion process at the migrating atom. Examination of a molecular model of XLI makes it quite obvious that an antara facial process would be essentially impossible. The second alternative would correspond physically to a 1,2 shift with inversion of configuratio~which is highly unlikely. An alternative route involves the zwitterion, L, which would form by a heterolytic cleavage of a cyclopropyl bond.

PAGE 56

50 This would be a relatively stable species since the positive charge is. stabilized by tropylium-type resonance and the negative charge by resonance with the adjacent carbonyl. Attack by the anion on c 2 of the tropylium system would lead directly to LXI. This intermediate would also be expected to give nonstereospecific products whicl1 could explain the mixture of cis and trans indane from the apparently pure isomer of XLI. It should be recalled, however, th~t this zwitterion was also mentioned as a possible intermediate in the addition of cycloheptatrienylidene to dimethyl fumarate. Although it cannot be ruled out in either case, L cannot be an intermediate in both schemes. Although a 1,7 sigmatropic shift in the transition from XLI to LXI is forbidden in a concerted thermal process, it is symmetry allowed in a concerted photochemical process (71). Thus, it appeared possible that photolysis of the spire com pound at a wave length lower than that used to generate it might result in the formation of LXI which at room tempera ture might not undergo further thermal reaction to LX. Accord ingly, XLI was photolyzed in a quartz vessel in benzene solu tion. After 3.5 hr of photolysis with an i~mersion lamp, the solvent was removed to give an oily residue. The nmr spectrum of the oil showed decomposition of most of the spiro compound, but no significant new absorptions attributable to a species such as LXI were found nor was the spectrum of LX observed. If Lis indeed an intermediate in the conversion, then its ionic character might cause a rate increase in more polar

PAGE 57

51 solvents. Samples of XLI in triglyme and ndecane were placed in an oil bath at 150 and removed at five-minute intervals to be examined by nmr. The approximate half-lives for formation of indane in n-decane and triglyrne respec tively were 20 min and 8 min. This corresponds to a rate increase of 2.5 in the more polar solvent and is not incon sistent with the above mechanism. Careful examination of the spectra from these two reactions revealed no transient species which might indicate a preequilibrium of XLI with a species such as LXI. The dicyanospirononatrienes, XLVIa and XLVIh, are pre sently being examined for an analogous thermal rearrangement.

PAGE 58

CHAPTER IV ADDI'l'IONAL SCHEMES EXAMINED 1~.s ROUTES TO CYCLOHEPTATRIENYLIDENE A. Base-promoted Decomposition of Methyl NNi troso-N-cyclo he_ptaFrTeny} s::arbama tea r -:2l N,-r-T =q}_methyi-~N7-=-ni troso-N I --cy_~l?h~P.:. tatrienylurea. Another route to diazocycloheptatriene or cycloheptatri enylidene which was explored was base-induced decomposition of nitrosoureas and nitrosocarbamates. Previous work in this laboratory (31,33) had shown that treatment of N,N dimethyl-N' -ni troso-N' (2, 3-diphenyl-2-cyclopropenyl) urea, LXII, with potassium t-butoxide in the presence of dimethyl fumarate gave the spiropentene, LXIV. Methyl N-nitroso-N (2,3-diphenyl-2-cyclo p rofenyl)carbamate, LXIII, also proved tc be a good source of LXIV. By analogy, the corresponding N-cycloheptatrienyl compounds should provide a route to XLI. Ph ~NOO y N-c':-B Ph base ----~__/ LXIII, B = OCH 3 52 Ph COOCH 3 Ph LXIV

PAGE 59

53 The ureas and carbamates are readily synthesized by treatment of the corresponding isocyanate with an amine or an alcohol. Accordingly, cyclohepta-2,4,6-trien-lylisocya nate was synthesized by addition of potassium cyanate to an aqueous solution of tropylium bromide (73) and by rearrange ment of the acid azide under anhydrous conditions (74). Addi tion of diethylamine or methanol to a solution of the iso cyanate gave the urea, LXV, and the carbamate, LXVI, respectively. B-H N-~-B O H 0 LXV, B = N(CH 3 ) 2 LXVI, B = OCH 3 0 Nitrosation of the carbamate was carried out at -20 by addinq a cold ether solution of dinitrogen tetroxide until the green color persisted fo~ 5 min. Addition of a large excess of the dinitrogsn tetroxide solution led to markedly decreased yields. After washing the reaction mj_xture with base and drying, a dark yellow oil remained. This oil was passed through a column of deactivated silica gel to give a pale yellow solution. Concentration under a stream of dry nitrogen afforded relatively pure nitrosocarbamate, LXVII, as a yellow oil. At room temperature, the oil decomposed 1 1 tt bl b t +o 0 _,_ t s ow y w1. l1 DUD _1ng, u a~ unc..er n1 .... rogen 1. t,vas apparently stable indefinitely.

PAGE 60

54 LXVI LXVII Stirring the nitrosourea in hexane with sodium meth oxide and several drops of methanol led to decomposition with quantitative gas evolution. Chromatography of the reaction mixture on deactivated basic alumina gave heptafulvalene (7%), but no other major product. When dimethyl fumarate was added to the reaction mixture, the spire adduct, XLI, was obtained in 4% yield. spirononatriene LXVII heptafulvalene The major product of this reaction was shown to be methoxysuccinate, LXVIII, which results from addition of rnethoxide across the double bond of the olefin. This pro vides confirmation of the mode of isomerization of the cis olefin to the trans in the presence of a base. A stirred solution of di~ethylfu rn arate and sodium methoxide in meth anol gave rise to the sa ~ e product in good yield.

PAGE 61

55 /COOCIL~ /COOCE3 CH30, /COOCH3 /c = C -<..~z: /c C --H_,+-~ CH300C CH300C 'ocH3 ,,,c, OCH 3 H COOCH 3 LXVIII The use of potassium t-butoxide as the base in the nitro socarbamate decomposition gave slightly higher yields. Hep tafulvalene was obtained in 13% yield without dimethyl fuma rate, and with the olefin, the yield of spiro adduct was 4.5%. The product of base addition across the olefinic dou ble bond, t-butoxysuccinate, was not observed presumably because of the bulky size of the base. The N,N-dimethylurea, LXV, was nitrosated as in the case of the carbareate, but the product was quite unstable and rapidly darkened during the workup. When the conditions of the workup were modified so that the nitrosourea, LXIX, LXV LXIX 0 was never allowed to warm to O reasonable yields were obtained. Low temperature nmr of the unstable compound con firmed its structure by analogy with the nitrosocarbamate, but as the sample was warmed to 10, the spectrum disappeared. An!ong the Feaks which appeared in the spectrum above 10,

PAGE 62

56 were those of N,N-dimethylcycloheptatrienylamine. This was confirmed by synthesizing a sample with dimethyl amine and tropylium bromide. In the case of the diphenylcyclopropenylnitrosourea, LXX, decomposition without base led to the carbamate, LXXI, which also gave rise to the spiropentene via an a-elimina tion when treated with potassium t-buto x ide. An attempt was made to convert LXIX into the correspond ing carbamate, but dimethylcycloheptatrienylamine was the only identifiable product. Apparently carbon dioxide as well as nitrogen is easily lost from the cycloheptatrienyl nitresourea. LXX Ph~ fi ll I o--c-N (ctr 3 ) 2 -------' -N2 Ph LXXI LXIX -------> -N 2 -co 2 O N(CH) 3 2 Treatment of a cold solution of the nitrosourea with sodium methoxide and a few drops of methanol gave besides the amine, a 2 % yield of heptafuJvalene. In a similar reac tion with dimethyl fumarate present, a 0.5% yield of the

PAGE 63

57 spirononatriene was found. These yields are based on starting urea and not on the nitrosourea. Since neither of the above routes appeared promising, they were not pursued further. B. a-Elimination Schemes. A number of other methods were examined as possible routes to cycloheptatrienylidene. Since it was highly desir able to find a route which did not involve the possible inter mediacy of diazocycloheptatriene, several a-elimin~tions were considered. a-Dehalogenations can frequently be effected by strong bases or metals (1,2). Accordingly, chlorotropy lium chloride was treated with electrolytic copper dust in tetrahydrofuran. After stirring overnight, the mixture had turned cloudy, but examination by uv and tlc showed the dimer, heptafulvalene, had not formed. Similar treatment with sodium metal followed by aqueous workup gave no tropone indicating a reaction with the chloride had taken place. Once again however, no evidence of heptafulval ene was found. The chloride was also treated with n-butyllithium and tbutyllithium with similar results. Previous work in this laboratory had shown that treat ment of diphenylcyclcpropenyl acetates, benzoates, and p-nitro benzoates with strong base in the presence of olefins gave products which indicated generation of the carbene (33). Cycloheptatrienyl acetate was synthesized (75) and treated with potassium t-butoxide in the presence of dimethyl fumarate. After refluxing for two days in heptane, the mixture

PAGE 64

58 ,-,as examined by glc and tlc. No cvic1ence was found to indicate that either heptafulvalene or the spiro compound, XLI, was formed. Similar results were observed when a solution of the acetate and dimethyl fumarate weJ:e treated with tbutyllithium at -50. Since this a-elimination route did not appear promising, it was not pursued further. Another method concerned d e ca~boxylation of a carboxy tropylium salt to give the carbene directly. Carboxytropy lium bromide has been reported and was found to be quite unstable (76). Since fluoroborate salts are in general much more stable than the corresponding bromides, a synthesis of carboxytropyliurn fluoroborate was undertaken. A solution of 2,4,6-cycloheptatriene carboxylic acid in dichloromethane was added dropwise to a solution of trityl fluoroborate in dichloromethane at room temperature. A light brown solid formed, and after thirty minutes of stirring, it was fil tered and dried. The fluoroborate charred at 150 and melted with gas evolution at 165. Recrystallization from nitroo methane-benzene gave white plates which charred at 153 and melted with gas evolution at 165. Their spectrum showed -1 a carbonyl peak at 1722 cm and an absorption at 272 m was found in the uv spectrum. The nrnr spectrum in ni tromethane showed a three-proton multiplet at T2.0-2.6 and a broad singlet at T2.80. The analysis indicated the empiri cal formula as c 8 E 7 o 2 BF 4 This salt was quite stable at room temperature under nitrogen as opposed to the behavior of the corres p onding bromide.

PAGE 65

59 Ph 3 CBF 4 0 ;> ~co 2 II + BF 4 LXXII Further confirmation of the structure of LXXII was obtained when the salt was reduced with lithium aluminumtri(t-butoxy)hydride. A mixture of three isomeric cyclo heptatriene carboxylic acids was obtained in 50% yield. 2,4,6-Cycloheptatriene carboxylic acid was not found and was not an expected product since it would involve hydride attack at the most hindered position of the tropylium ring. 1,3,5-Cycloheptatriene 1-carboxylic acid, LXXIII, composed 35% while the other two isomeric acids made up the remain ing 65%. Q : 35% OOH LXXIII OCOOH LiAl(OtBu) 3 H -------COOH 65% 11ooc-O:

PAGE 66

60 If the carboxylic acid proton could be removed from LXXII, decarboxyla tion should lead directly to cycloheptatrienylidene. The acid-salt was treated with a number of bases including lithium carbonate, sodium hydride, triethyl amine, and lithium metal. In each case, reaction occurred 7 -co 2 but the product was normally a gum my oil which resisted all efforts at characterization. Further work is anticipated in this system.

PAGE 67

CHAPTER V SUM:t,,11\RY The base-promoted decomposition of 2,4,6-cyclohepta trienone p-toluenesulfonylhydrazone has been investigated as a source of cycloheptatrienylidene. Either photolytically or thermally, the decomp6sition gives up to 70% of the un stable dimer, heptafulvalene, when no acceptors are present. In the presence of dimethyl fumarate, an electron-deficient olefin, 8,9-dicarbomethoxyspiro(2.~ nona-2,4,6-triene can be isolated in 50% yield. This spirononatriene is the adduct expected if the carbene adds to the olefin. With fumaroni trile, an analogous spirononatriene is formed. The cis isomer of fumaronitrile, maleinitrile, gives the cis adduct indicating that the addition process is highly stereoselec tive if not stereospecific This stereospecificity at high dilution is given as evidence that the carbene is adding in a singlet state which may be the ground state of the species. Evidence is presented against the diazo compound as the reactive intermediate in the formation of the spirononatri enes. The spirononatrienes were found to undergo a facile, acid-catalyzed rearrangement to the corresponding heptaful61

PAGE 68

62 venes. Above 130, the adducts rearrange to 1,2-disubsti tuted indanes, and an interesting mechanism is postulated to explain this thermal isomerization. When electron-rich double bonds are used as acceptors, only ti1e dimer is formed. Thus, it appears that the reactive intermediate is nucleophilic as well as stereospecific in its reaction with olefins. Base-catalyzed decomposj_tion of methyl N-cyclohepta trienyl-N-nitrosocarbamate and N,N-dimethyl-N'-nitroso-N' cycloheptatrienylurea wa~ also investigated as a carbene source. The spirononatriene is formed when dimethyl fuma rate is present, and heptafulvalene is a product in the absence of the olefin; but in each case, the yields are quite low, and side reactions predominate.

PAGE 69

CHAP'l'ER VI EXPERIMENTl\L A. General. Melting points were taken in a Thomas Hoover Unimclt apparatus and are uncorrected. Elemental analyses were performed by Galbraith Laboratories, Inc., Knoxville, Tennessee. Ultraviolet spectra were recorded on a Cary 14. or Cary 15 double-beam spectrophotometer using 1 cm. silica cells. Infrared spectra were recorded with a Beckmann Model IRlO spectrophotometer. In all cases where the KBr technique was not used, sodium chloride plates were substituted. Nuclear magnetic resonance spectra were determined on a Varian A-60A High resolution spectrometer with a variable temperature probe. Chemical shifts are reported in tau values from internal tetramethylsilane standard. Mass spectra were determined on a Hitachi model ruru-6E Mass Spectrometer. Analytical thin layer chromatography {tlc) was accom plished on 2" X 8" plates coated in these laboratories with 0.25 rrun layers of E. Merck HF 254 silica gel; components were visualized by their quenching of fluorescence under ultraviolet light. Gas-liquid chromatography {glc) was conducted on a Wilkens Aerograph Model 600-B, or 600-D Hi-Fi instruments, with flame ionization detectors using 63

PAGE 70

nitrogen carrier gas; preparative scale work was accomplished on a Wilkens Aerograph Dual Column Temperature Programmer Gas Chromatograph, usin~ helium carrier gas. All photolyses were carried out with a Hanovia, 450 watt, high-pressure mercury immersion lamp, and unless otherwise noted, a pyrex vessel was used. Tetrahydrofuran for the photolysis reactions was dried by distillation 64 frorn lithium aluminum hydride. rhe sodium hydride was obtained from Alfa Inorganics, Inc. It was weighed as a dispersion, washed three times with pentane, and used as a powder for each reaction. All technical hydrocarbon solvents were distilled before use. The cycloheptatriene used in the following reactions was generously donated by Shell Chemical Co. and contained 9% toluene. B. Chlorocycloheptatrien_xlium Chloride. This compound has been reported (47) and was prepared as described with the following exception: thionyl chloride was used as sol vent and halogenating agent. Tropone was added dropwise to an excess of thionyl chloride at o 0 (reverse addition caused vigorous decomposition of the sample with no produc tion of the desired product). At the end of the addition, the solution was refluxed gently for 5 min on a steam bath. Removal of the excess thionyl chloride on a rotary evapora tor gave the chloride as a yellow crystalline solid which was used without further purification.

PAGE 71

65 c. 2, 4_, 6-Cyclohep tatrienone p -'l'oluenesulfonylhydrazone Hydrochlorid~. A solution of 450 mg (2.42 mmoles) of toluenesulfonylhydrazine in 5 ml of absolute ethanol was added rapidly with stirring to a solution of 304 mg (2.45 mmoles) of chlorotropylium chloride in 5 ml of absolute ethanol. The dark red solution was rapidly stirred at room temperature for 30 min. The yellow solid which had formed was filtered, washed with ether, and dried in vacuo to give 600 mg (90%) of the salt. Three recrystallizations from methanol-ether gave an analytical sample, mp 170d (st), liquefaction and gas evolution at 210; KBr vmax 3180, 2930, 2700 (broad), 1638, 1595, 1522, 1498, 1450, 1348, 1254, 1233, 1190, 1165, 1090, 1030, 905, 833, 812, 766, 710, 648, -1 583, 555, 546, 442, and 375 cm ; ~EtOH 229 lE 13200) and max 315 m (e: 15700); D 0 t'I 22.56 (complex pattern, 10 H, aroma Fi~ tic and cycloheptatriene ring protons) and 7.73 (singlet, 3 H, methyl protons); m/e ( 70 ev) 274, 156, 155, 139, 119, (base peak), 107, 91, 90, 89, 78, 77, 65, 53, 51, 50, 39, 38, and 36; metastable peaks, 46.5 and 60.5. Anal. Calcd for c 14 e 15 N 2 o 2 scl: C, 54.10; H, 4.83; N, 9.03. Found: C, 54.37; H, 5.03, N, 9.26. D. 2,4,6-Cycloheptatrienone p-Toluenesulfonylhydrazone. To a vigorously stirred mixture of 50 ml of 10% aqueous NaHco 3 solution and 25 ml of dichloromethane was added 1.00 g (3.22 mmoles) of the hydrochloride salt. After 15_min, gas evolution had ceased and the two layers were separated. The aqueous layer was washed twice with 10 ml portions of

PAGE 72

66 dichloromethane. The organic fractions were combined, dried over magnesium sulfate, and concentrated on a rotary evapo rator to give the p-toluenesulfonylhydrazone as a dark red solid. Recrystallization from benzene-pentane afforded 850 mg (96%) of XXXV. The p-toluenesulfonylhydrazone has two distinct crystalline forms: 0 dark red pl.ates, mp 142.5-143.5 and light red needles, mp 144-J 45. Both forms gave identical infrared spectra in nujol mulls. Several recrystallizations from benzene-pentane gave an analytical sample; KBr vmax 3240, 2038, 1603, 1567, 1471, 1393, 1304, 1310, 1228, 1187, 1170, 1092, 1042, 1003, 920, 864, 813, 743, 700, 660, 565, 547 --1 ,EtOH cm A 2 4 5 ( 7 9 0 0) and 315 m ( 1310 0) ; max 1.84-2.70 (A 2 B 2 4 H, aromatic protons), 3.33-3.83 (complex pattern, 6 H, cycloheptatriene ring protons) and 7.55 (sing let, 3 H, methyl protons); m/e (70 ev) 274 (molecular ion), 180, 179, 156, 155, 139, 119 (base peak), 107, 91, 90, 89, 78, 77, 65, 63, 61, 50, and 39. Anal. Calcd for c 14 H 14 N 2 o 2 s; C, 61.30; 10.22. Found: C, 61.50, H, 5.13; N, 10.06. H, 5.11, 1\T .L. '4 E. Photolysis of 2, 4, 6-Cycloheptatrienone p-Toluenesul fony_;!..:_ hydrazone Sodium Salt. Formation of Heptafulvalene. In a typical reaction 0.500 g (1.82 mmoles) of the p-toluenesul fonylhydrazone was dissolved in 120 ml of dry tetrahydrofuran. The sodium salt was formed by addition of 50 mg (2.08 mmoles) of sodium hydride powder to the dark red solution. When gas evolution had ceased (15 min), the photolysis vessel was swept with a stream of dry nitrogen for 10 minutes to

PAGE 73

67 remove hydrogen gas. Photolysis of the dark slurry was fol lowed by nitrogen evolution which stopped at 35 ml (85%) after 3 hr. The cold finger was coated with a light brown solid which ir showed to be sodium p-toluenesulfinate~ The red solution was poured into 300 ml of water and extracted with three 50 ml portions of pentan e The pentane extracts were combined, dried over magnesium sulfate, and concentrated to a crude black solid. This solid was taken up in 10 ml of 5% ether in hexane and chromatographed on 20 g of basic alumina (Woelm, activity grade III) using 5% ether in hexane as eluent. The heptafulvalene was eluted rapidly as a sharp, dark red band. Uv determination of the yield as described below gave 107 mg (65%) of the dimer. Concentration of the solution gave a crude black crystalline solid which after several recrystallizations from methanol-water (pH 7-9) or sublimation gave permanganate-colored plates with mp 119121 [lit. (52) mp 12 2 J ; VKBr 2900, 1540, 1440, 1265, max 975, 935, 898, 888, 835, 805, 790, 781, and 712 -1 >..EtOH cm max 234 (E 24000) and 362 m (E 25000) [lit. ( 5 2) >.. 234 max (E 22000) and 362 m (E 21000)]; CDC1 3 TTMS 4.09 (structured singlet); rn/e (70 ev) 180 (molecular ion and base peak), 165, 152, 139, 115, 102, 90, 89, 76, and 63; metastable peaks, 130, 68.8, and 56.8. A satisfactory analysis was not obtained for this com pound. F. Determination of ~he Yi~~j_.....?.,f_~~p!_~;ulvale~ ~Due to the losses incurred in handling heptafulvalene, the yields

PAGE 74

68 were determined by a spectroscopic method. The reaction mixtures were first chromatographed on basic alumi.na (Woelm, activity grade III) using 5% ether in hexane as eluent. This allowed rapid elution of the dimer which decomposes slowly on the column, while still obtaining excellent sep aration from all other components of the reaction mixtures. The fractions which contained heptafulvalene were then com bined and diluted accurately to give a solution estimated to be 10-S M. The exact concentration was then calculated from the uv absorption of the 362 m peak. The extinction coefficient for this absorption is 21000 as reported by Doering (52). The extinction coefficients calculated in these laboratories on three samples of heptafulvalene which were recrystallized three times from methanol-water (pH 7-9) and triply sublimed, were 26000, 25000, and 21400. The last sample was left overnight under nitrogen and may have partially decomposed, so the value selected was 25000. G. Photolysis of 2,4,6-Cycloher2_!:atrienone p-Toluene sulfonylhydrazone Sodium Salt in the Presence of Dimethyl Fumarate. Formation of 8, 9-Dicarbomethox y ':3piro (2. 6] nona-~ 2,4,6-triene. In a typical reaction l.OOg (3.65 mrnoles) of the p-toluenesulfonylhydrazone was dissolved in 120 ml of dry tetrahydrofuran. The sodium salt was formed by addition of 100 mg (4.15 ni.moles) of sodium hydride powder to the dark red solution with vigorous stirring. When gas evolu tion had ceased (ca. 15 min), the photolysis vessel was swept with a stream of dry nitrogen to remove hydrogen gas.

PAGE 75

69 To this slurry was added 2.50 g (17.4 mmol e s) of dimethyl fumarate. After 4 hr of photolysis, the nitrogen evolution had stopped at 70 ml (85%). The light red solution was poured into 500 ml of water and extracted with tl1ree 200 ml portions of pentane. The pentane extracts were combined, dried over magnesium sulfate, and concentrated to a dark solid. This residue was taken up in 20 ml of 5% ether in hexane and chrornatographed on 50 g of basic alumina (Woelrn, activity grade III). Heptafulvalene came off rapidly with 5% ether i n hexane and the yield, as determined by uv, was 38.7 mg (10.5%). Gradually increasing the ether concentra tion to 10% gave a mixture of dimethyl fumarate and the spirononatriene compound. Fractional crystallization gave 814 mg of recovered dimethyl fumarate and 421 mg (49.6%) of XLI with mp 72-74. Two recrystallizations from methanolo water (pH 7-9) gave an analytical sample with_mp 76.5-77.5 ; VKBr 3080, 3020, 3002, 2970, 1720, 1450, 1440, 1340, 1270, max 1200, 1158, 1078, 1018, 895, 885, 872, 745, 705, 645, 590, 502, and 330 cm-l,. ~EtOH 262 m (E 2800), ~CDCl3 3.20max TMS 3.83 (complex pattern, 4 H, 3,4,5, and 6 protons on cycloheptatriene ring), 4.55 (doublet, 2 H, 2 and 7 protons on cyclo heptatriene ring), 6.31 (singlet, 6 H, ester methyls), and 7.69 (singlet, 2 H, cyclopropyl protons); m/e (70 ev) 234 (molecular ion), 203, 175, 174, 116, 115 (base peak), 91, 90, and 89; metastable peaks, 131 and 76. Anal. Calcd for c 13 H 14 o 4 : C, 66.66; H, 5.98. Found: C, 66.50; H, 5.93.

PAGE 76

70 H. Isomerization __ of __ 8 ,_9-Dicarbomethoxyspixo [2. 6] nona-_2, 4, 6triene _to 8-Carbornethoxy-8-carbomethoxymethylmethylenecyclo ~e:eta-::-_2, 4, 6-triene on l-~cid Alumi::]a _:_ A solution of 42 mg (0.18 nm~les) of the spirononatriene in 5 ml of anhydrous ether was placed on a dry column of 5 g of acid alumina (Woelm, activity grade I). Upon contact, the column imme diately turned bright yellow. The yellow band was eluted with acetone and concentrated to a dark yellow oil on a rotary evaporator. This oil was was taken up in pentane and cooled in dry ice to give a fluffy yellow solid. Recrys tallization from pentane gave 38 mg (91%) of the heptafulvene as yellow needles with mp 48-50. Two further recrystallizations gave an analytical vKBr 3020 2960 2910 1750 0 sample with mp 49.5-51 ; max ' 1705, 1642, 1570, 1442, 1355, 1278, 1203, 1184, 1118, 1030, 822, 783, 744, 660, and 545 -1 cm AEtOH 245 (E 10000) and 340 max mp (E: 13000): CDCl3 'tTMS 2.57 (poorly resolved doublet, 2 H, 2 and 7 protons on cycloheptatriene ring), 3.40-3.85 (multiplet, 4 H, 3,4,5, and 6 protons on cycloheptatriene ring), 6.29 and 6.34 (singlets, 6 H, ester methyls), and 6.65 (singlet, 2 H, methylene protons); m/e (70 ev) 234 (molecular ion), 203, 175 (base peak), 115, 91, 89, and 59; metastable peaks, 131 and 76. Anal. Calcd for c 13 H 14 o 4 : C, 66.66; H, 6.02. Found: C, 66.48; H, 5.88. I. Photolysis __ of 2, 4, 6_:_~ycl~_hepta trteno::-ie p-Tol1.1enesu 1 fonyl hydra zone ~9liiun~Salt in the Presence of Fumaronitrile. Format.ion of trans-8 9-~iC:_y a,!_l ~~pj..r9_[2 _._ 6] nona-2 ,_4, 6--trieI_':._E:_~

PAGE 77

71 To a solution of 548 mg (2.0 mmoles) of the p-toluenesul fonylhydrazone in 120 ml of dry tetrahydrofuran was added 62 mg (2.58 mmoles) of sodium hydride powder. Vigorous gas evolution occurred, and the sodium salt precipitated to form a dark brown slurry. When the reaction was complete (ca. 15 min), the reaction flask was flushed well with dry nitrogen to remove all hydrogen ga~ and 780 mg (10.0 mmoles) of fumaronitrile was added. After 2.5 hr of photolysis, the nitrogen evolution had stopped at 42.5 ml (89%). The dark red solution was poured into 300 ml of water and extracted with three 100 ml portions of pentane. The pentane extracts were combined, dried over magnesium sulfate, and concentrated to a dark oily solid on a rotary evaporator. Tlc of the crude residue showed heptafulvalene as a dark spot and a second colorless spot in 33% ether in pentane. The second spot turned bright yellow after standing on the dry plate for 15 min, consistent with the behavior of the adduct of cycloheptatrienylidene and dimethyl fumarate. The nmr spec trum of the crude residue showed dimer, fumaronitrile, and a set of absorptions consistent with a spiro compound. The residue was taken up in 10% ether in hexane and chromato graphed on 25 g of basic alumina (Woelm, activity grade III) using 10% ether in hexane as eluent. The dimer was eluted rapidly as a characteristic red band, and the yield was found to be 33 mg (18.5%). When the solvent polarity was increased to 20% ether in hexane, a white solid was obtained which was shown to be the adduct, X~VIa, 97 mg (29%) with

PAGE 78

72 0 mp 110-114 Three recrystallizations from benzene -hexane gave an analytical sample with mp 121 123: KBr v 3030, max 2250, 1528, 1440, 1400, 1238, 1212, 1265, 1030, 970, 875, -1 ,EtOH 8 3 0 7 6 0 7 0 5 6 6 0 6 2 5 6 2 2 4 8 5 and 4 5 0 cm ; I\ max 262 m (c:: 2760); CDCl3 T 1 ~ 8 3.10-3.60 (complex pattern, 4 H, 3,4,5, and 6 protons of cycloheptatriene ring), 4.53 (doublet, 2 H, 2 and 7 protons on cycloheptatriene ring), and 8.01 (sing-let, 2 H, cyclopropyl protons); m/e (70 ev) 168 (molecular ion), 167, 141, 140, 128, 115, 90 (b~se peak), 89, 77, and 63; metastable peaks, 117.7 and 48.3. Anal. Calcd for c 11 H 8 N 2 : C, 78.55; H, 4.79; N, 16.65. Found: C, 7 8. 8 2; H, 4. 81; N, 16. 7 0. J. Isomeriza tion of trans-8, 9-Dicyanospiro [2. 6] nona-2 1 4, 6_:_ triene to 8-Cyano-8-cyanomethylmethvlenecyclohepta-2,4,6-tri ene on Acid Alumina. A solution of 20 mg (0.12 mmoles) of the spire compound in 2 ml of anhydrous ether was added to a dry column of 5 g of acid alumina (i ioelm, activity grade I). An additional 2 ml of anhydrous ether was added to wash the solution onto the column. The bright orange band which rapidly developed was then eluted with acetone to give a red solution. This solution was concentrated to 1 ml under a stream of dry nitrogen, and pentane was added to the cloud point. After cooling in the refrigerator for 1 hr, the red crystals of the heptafulvene were filtered and dried to give 15 mg (75%) with mp 97-99. Recrystallization from 0 benzene-hexane gave red plates with mp 99.5100.5 ; vKBr 2255, 2195, 1640, 1548, 1520, 1474, 1408, 1272, 1160, max

PAGE 79

73 912, 862, 821, 760, 715, 580, 523, and 500 cm-l 2 4 7 ( s 9 5 0 0) and 3 tl 5 m ( c: 170 0 0) ; CDCl3 1 TMS 2.83-3.30 (unresolved multiplet, 2 H, 2 and 7 protons on cyclohepta triene ring), 3.50 (broad singlet, 4 H, 3,4,5, and 6 protons on cycloheptatriene ring), and 6.72 (singlet, 2 H, methylene protons); rn/e (70 ev) 168 (molecular ion), 167, 141, 140, 128, 115, 90 (base peak), 89, 77, and 63; metastable peaks, 117.7, 92.5, 80.0, and 48.3. Anal. Calcd for c 11 H 8 N 2 : C, 78.55; H, 4.79; N, 16.65. Found: C, 78.43; H, 4.72; N, 16.51. K. Photolysis of 2, 4, 6-S~ycloheptatrienone pToluenesulfon y lhydrazone Sodium Salt in the Presence of Maleinitrile. Forma .. tion of cis-8, 9-Dicyanospiro [2. 6] nona--2, 4, 6-triene. This reaction was carried out as described in section I, with one exception; maleinitrile was substituted for fumaronitrile. The yield of heptafulvalene was 17.6%. The spirononatriene was obtained in 30% yield and was assigned the ?is structure on the basis of the physical characteristics given below. An analytical sample was obtained after several recrystalli zations from ether-pentane with mp 116-118.5; vKBr 3020 max 2245, 1532, 1444, 1399, 1378, 1353, 1315, 1274, 1208, 1169, 1058, 1009, 988, 940, 888, 811, 755, 702, 668, 619, 558, 516, -1 EtOH CDCl3 444, and 372 cm ; A 262 m (E: 2720); -rTI'S 3 .10-3. 35 max (complex pattern, 2 H, 4 and 5 protons on cycloheptatriene ring), 3.40-3.80 (complex pattern, 2 H, 3 and 6 protons on cycloheptatriene ring), 4.25-4.55 (broad doublet, 1 H, 2 proton on cycloheptatriene ring), 4.55-4.85 (broad doublet,

PAGE 80

7/1, 1 H, 7 proton on cycloheptatriene ring), and 7.97 (singlet, 2 H, cyclopropyl protons); m/e (70 ev) 168 (molecular ion), 167, 141, 140, 128, 115, 90 (base peak), 89, 77, and 63; metastable peaks, 117.7 and 48.3. Anal. Calcd for c 11 H 8 N 2 : C, 78.55; H, 4.79; N, 16.65. Found: C, 78.35; H, 4.77; N, 16.60. L. ~r.iotolysis of 2, 4, 6-::_cyclohe12 ta tEJ_~nc.?_!.1e p-Toluenesulf onyl gydrazon~Sodium Salt in the Presence of Other Acceptors. In a typical reaction, 0.500 g (1.82 mmoles) of the p-toluene sulfonylhydrazone, XXXV, in 120 ml of dry tetrahydrofuran was converted to the sodium salt with 53 mg (2.20 mmoles) of sodium hydride powder. After gas evolution ceased, the re action flask was flushed with nitrogen and a tenfold excess of the acceptor was added. The reaction was followed by nitrogen evolution and was normally complete within 3 hr. The crude reaction mixture was examined with tlc and nmr for evidence of an adduct. The yield of heptafulvalene was determined as in the previous reactions and varied from 3070%. No evidence of an adduct was found in the reaction with cis--2-butene, cyclohexene, Npipe::.idylcyclohexene, trans-p dinitrostilbene, 1,2-dibromoethylene, methyl trans-5-chloro acrylate, maleic anhydride, fumaramide, methyl cinnamate, dimethyl acetylenedicarboxylate, methyl S-acetylacrylate, and the mixed methyl ester-dimethylamide of fumaric acid. enesulfonylhydrazone Sodium Sal ~-~~i t~~nd without D~_me_!:hyl Fumarate. Two 10 ml portions of dry diglyme (distilled from

PAGE 81

75 lithium aluminum hydride) were thermostated at 101.5 under nitrog e n in a constant temperature bath containing ref lu x in, ; dioxane. 'l'o one sample was added 2 8 8 mg ( 2. 0 0 mmoles) of dimethyl fumarat e Employing a solid addition tube, 368 mg (1.00 mmoles) of the salt was added in each case. The rates of both reactions were followed by nitrogen evolution and were virtually identical, giving half-lives of ca. 11 min. Both solutions were evaporated to dryness in vacuo and worked up as usual. The yield of heptafulvalene in the reaction without dimethyl fumarate was 29.5 rng (33%). In the reaction with dimethyl fumarate, the yield of the spire compound was found to be 90 mg (38%), while the heptafulvalene had dropped to 8.0 mg (9%). N. Phot_?..:"b.._Ysis of Heptafu1 valene in the presence of Dimethyl_ Fumarate. A solution of 17.3 mg (0.096 mmoles) of heptafulva lene and 69 mg (0.48 rnmoles) of dimethyl fumarate in 50 ml of hexane was photolyzed with two Kenmore Sunlamps through a pyrex filter for 15 hr. At the end of this time, the colorless solution was concentrated and examined for products. Tlc, uv, and glc of the reaction mixture showed that no spirononatriene had been formed. Dimethyl fumarate was the only identifiable compound. O. Photolysis of 2,4,6-Cycloheptatrienone p-Toluenesulfonyl hydrazone Sodium Salt in the Presence of Dimethyl Fumarate and Triethylamine. This reaction was identical to the decom position described in section G except for the presence of 3.7 g (36.5 mmoles) of triethylamine. The nmr spectrum was

PAGE 82

76 examined carefully for the presence of a 2-pyrazoline in the reaction mixtu r e, but no evidence for one wa s found. The yields of heptafulvalene and spirononatr~ene wer e 53 mg (16 % ) and 420 mg (49%), respectively. P. Thermal Decompo s ition of 2,4,6-Cyclohep;atrienone p-Tolu enesulfonylhydrazone Sodium Salt under Vacuum. In a typical reaction, 250 mg (0.91 mm o les) of the p-toluenesulfonylhydra zone, XXXV, in 5 ml of tetra.hydrofuran was converted to its sodium salt with 26 mg (1.1 mmoles) of sodium hydride powder. As the solvent was removed oh a rotary evaporator, the salt was distributed uniformly on the inside of the 500 ml flask. The flask was then fitted with a cold finger. The salt was dried at so 0 and ca. l03 mm of Hg overnight. Liquid nitro gen was added to the cold finger, and the reaction flask was heated slowly with an oil bath to 150. The finger was coated with a light red material at the end of this time. An atmosphere of dry nitrogen was bled into the system, and the liquid nitrogen was removed. As the finger began to warm (ca. -150), a noticeable darkening of the coating occurred. After warming to room temperature, the coating was examined and found to consist of tetrahydrofuran and heptafulvalene, 16 4 mg ( 2 0 % ) Q. Thermal Decomposition of 2,4,6-Cycloheptatrienone o-Tolu enesulfonylhydrazone Sodium Salt under Vacuum with Dimethyl Furna.rate. A uniform coating of the sodium salt was prepared from 250 mg (0.91 mmoles) of the p-toluenesulfony1.hydrazone and 26 mg (1.1 mmoles) of sodium hydride as described above.

PAGE 83

77 The reaction fla s k had been modified by adding two small tubes pointed at the cold finger through which dimethyl furnarate could be sublimed. The decomposition was also carried out as described above with the fumarate subliming onto the finger during the entire reaction. Again a light red coat was observ ed and again a notice a ble darkening occurred. When the ma terial on the finger was examined after the reaction, with nmr, glc, and tlc, only heptafulvalene, tetrahydrofuran, and dimethyl fumarate were found. No trace of a spirononatriene was found. R. Thermal Rearrangement of 8, 9-Dicarb<_?methoxvspiro [2, 6] nona-2, 4, 6-triene to 1,2-Dicarbornethoxyindane. A solution of 100 mg (0.43 ITu'1loles) of the spirononatriene in 1 ml of ether was injected into the preparative glc on a 3/8" X 5' 20% SE30 on 60/80 Gas ChromZ column at 150 and 50 psi in 0.1 ml portions. The two peaks which were eluted were collected at room temperature. The larger peak (retention time ~a. 8 min) gave white crystals, mp 68-70. Recrystallization from benzene-hexane gave 30 mg (30%) of trans-1,2-dicarbomethoxy indane with mp 70-72. This sample was identical with a sam ple synthesized by the method of Cook and Stephenson (70). Anal. Calcd for c 13 H 14 o 4 : C, 66.66; H, 6.02. Found: C, 66.51; H, 5.99. The minor peak (retention time ca. 6 min) also gave a white solid, 2 mg, with mp 65-69. Glc showed this to be a 50:50 mixture of the two peaks. Ir and uv of the solid were identical with those of the trans isomer above, so the minor

PAGE 84

78 peak \Jas assigned the structure of ci~-1, 2-dicarbomethoxy indane. S. 'I'hermal Rearrangement. of 8, 9-Dic~~!?.?met0oxyspiro [2, 6]:: !1ona-2, 4, 6-triene to 1, 2-Dicarbomethoxy indane in n-J:?ecane and Triglyme. Two solutions were made consisting of 32.7 mg (0.14 mmoles) in 0.5 ml of n-decu.ne and 33.3 mg (0.14 mrfloles) in 0.5 ml of triglyme. The solutions were placed in nmr tubes and suspended in an oil bath at 150. The reaction was followed by removing the tubes at 5 min intervals and examining their nmr spectra. The approximate half-lives as determined by the appearance of the aromatic protons of the indane were 20 min and 8 min for the n-decane solution and the triglyme solution, respectively. T. Methyl }i=:.fY~loheDta trienylcarbarna te. To a solution of 13.3 g (0.10 moles) of 2,4,6-cycloheptatrienyl-l-isocyanate 0 in 150 ml of dry benzene at 10 was added 5.4 g (0.10 moles) of sodium methoxide in 30 ml of methanol. Reaction was vigo rous and exothermal. An ir spectrum of the crude mixture -1 showed no isocyanate peak and a new carbonyl at 1680 cm After filtration, the filtrate was concentrated on a rotary evaporator to a dark oil. After dissolving the oil in a small amount of dichloromethane, pentane was added slowly. The solid which formed was filtered and dried to give 7.3 g (44%) of the crude carbamate with mp 69-72. No attempt was made to improve this yield. Three recrystallizations from hexane gave a pure sample with 0 mp 78.5-80.5 ; VKBr max 3270, 3080, 2945, 1680, 1555, 1310, 1255, 1050, 700, and

PAGE 85

440 --1 cm \EtOH 253 m ( c 4400) max 1--i ~CDCl3 3 34 {complex TMS pattern, 2 H, 3 and 6 protons on cycloheptatriene ring), 79 4.53 (complex pattern, 2 H, 2 and 7 protons on cyclohepta triene ring), 4.83 (broad peak, 1 H, nitrogen proton), 5.97 (quartet, 1 H, methyne proton), and 6.34 (singlet, 3 H, methyl protons); m/e (70 ev) 165 (molecular ion), 150, 133, 106 (base peak) 91, 79, 77, 65, and 59. Anal. Calcd for c 9 H 11 No 2 : C, 65.44; H, 6.71; O, 19.37. Found: C, 65.63; H, 6.73; 0, 19.38. U. .Methyl N-Ni troso-N-cyclohepta trie~y_l~-~.E~_ama te. A stirred slurry of 250 mg (1. 51 :rrffnoles) of methyl Ncycloheptat.rien ylcarbamate, 1.23 g (15.0 mmoles) of anhydrous sodium acetate, and 0.5 g of anhydrous sodium sulfate in. 15 ml of dry dichloromethane was cooled to -25. A 0.48 M ether solution of dinitrogen tetroxide (prepared by bubbling the gas into a tared volumetric flask of ether at -so 0 and noting the increase in weight) was added in 3 ml portions until the green color persisted for at least 10 min. The excess dinitrogen tetroxide and approximately half of the dichloromethane were removed in vacuo and 10 ml of anhydrous ether was added. The r e action ~ixture was then washed three times with ice-cold aqueous sodium bicarbonate solution saturated with salt and twice with ice-cold salt solution. The aqueous layers were combined and washed twice with dichloromethane. All organic fractions were combined a~d dried over sodium sulfate, then filtered through magnesium sulfate to remove the last traces of water. After removing the solvent in vacuo, the dark oil was taken

PAGE 86

80 up in 1:1 ether-pentane solution and passed through a short silica gel column which had been deactivated by packing in water-saturated ether. The yellow eluent was collected until it was colorless and concentrated under a stream of nitrogen. The yellow oil (237 mg, 80%) was reasonably pure by nmr; vplates 3020, 2960, 1755, 1510, 1440, 1360, 1310, max l 1200, 1150, 1070, 770, and 705 cm ; 3.38 (broad triplet, 2 H, 4 and 5 protons on cycloheptatriene ring), 3.75-4.10 (complex pattern, 2 H, 3 and 6 protons on cyclo heptatriene ring), 4.70-5.10 (complex pattern, 2 H, 2 and 7 protons on cycloheptatriene ring), 5.20-5.50 (broad triplet, 1 H, methyne proton), and 5.89 (singlet~ 3 H, methyl protons). V. Decom2-.9sition of Methyl N-Nitroso-N-cycloheptatrienyl carbamate with Sodium Methoxide. To a stirred solution of 240 mg (1.23 ITmoles) of the nitrosocarbamate in 30 ml of anhydro~s ether were added 850 mg (15. 7 mrnoles) of sodium methoxide and 2-3 drops of methanol. Quantitative nitrogen evolution occurred within 30 min. The solid was removed by vacuum filtration and the filtrate was concentrated to a dark oil on a rotary evaporator. This oil was taken up in 10 ml of 5% ether in hexane and chromatographed on 30 g of basic alumina (Woelm, Activity grade III) using 5% ether in hexane as eluent. The heptafulvalene came off rapidly as a dark red band, and the yield, as determined by uv, was 8 mg ( 7 % ) W. Decomposition of Methvl N-Nitroso-N-cvcloheptatrienvl---=--------~~---------~--carbamate with Sodium .Methoxide in the Presence of Di m eth.-_yJ:.

PAGE 87

81 Fumarate. To a stirred solution of 237 mg (1.22 mmoles) of the nitrosocarbamate in 30 ml of anhydrous ether were added 1.75 g (12.2 mrnoles) of dimethyl fumarate and 2-3 drops of methanol. After the olefin had completely dissolved, 600 mg (11.1 mmoles) of sodium methoxide was added. Quanti tative nitrogen evolution occurred within l hr and 15 min. The solid was filtered and discarde~ and the filtrate was concentrated under a stream of dry nitrogen to 10 ml. The remaining solution was evaporated to dryness on a rotary evap orator and taken up in 10 ml of 5% ether in hexane. Careful chromatography of this solution on 30 g of basic alumina (Woelm, activity grade III) using 5% ether in hexane and gradually proceeding to 8% ether in hexane gave 7 mg (4%) of the spirononatriene which was pure by glc. X. Decomposition of Methyl N-NitrosoN-cycloheptatrienyl__:_ carbamate with Potassium t-Butoxide. A solution of 210 mg (1.08 mmoles) of the nitrosocarbamate in 30 ml of anhydrous ether was cooled to o 0 with stirring. To this was added 550 mg (5.8 rnraoles) of potassium t-butoxide, and the dark mix ture was allowed to warm to room temperature. Gas evolution was measured during the next 2.5 hr to be 20 ml (83%). The mixture was filtered, and the filtrate evaporated to dryness on a rotary evaporator. The oily residue was taken up in 10 ml of 5% ether in hexane and chromatographed on 20 g of basic alumina (Woelm, activity grade III) using 5% ether in hexane as eluent to give the dimer as a dark red band. Determination of the concentration of a solution of the dimer showed

PAGE 88

82 13 mg ( 13 ~) carbamate with Potassium t-Butoxide in the Presence of Dimethyl Fumarate~ A solution of 200 mg (1.03 mmoles) of the nitrosocarbamate and 1.50 g (10.4 mmoles) of dimethyl fumarate in 30 ml of anhydrous ether was cooled to o 0 with stirring. To this was added 700 mg (6.2 mmoles) of potassium t-butoxide, and the mixture was allowed to warm to room tem perature. Gas evolution over the next 4 hr was 20 ml (87%). The orange slurry was poured into 50 ml of water and extracted with three 25 ml portions of dichloromethane. The organic extracts were combined, dried over sodium sulfate, and con centrated to an oily solid on a rotary evaporator. The residue was taken up in 10 ml of 5% ether in hexane and chro rnatographed on 30 g of basic alumina (Woelm, activity grade III) using 5% ether in hexane as eluent to give 108 mg of a white crystalline solid which was shown by nmr to consist of 11 mg (4.5%) of the spirononatriene and 99 mg of dimethyl fuma rate. Glc and tlc showed these to be the only components. Z. N,N-Dimethyl-N'-cycloheptatrienylurea (77). To a solution of 2. 5 g ( 19. 0 rnmoles) of 2, 4, 6-cyclohepta trien-1 -ylisocyao nate in 20 ml of anhydrous ether at O was added 3 g (66 rnrnoles) of dimethylarnine. After stirring the solution for 10 min, the ether and excess dimethyl.amine were removed on a rotary evaporator to give 1.3 g (38%) of the urea as.a crude solid with mp 110-115. Three recrystallizations from benzenehexane gave an analytical sample with mp 124-127; \JKBr max

PAGE 89

83 3320, 3010, 2930, 1625, 1525, 1380, 1302, 1235, 1198, 1127, 1069, 1058, 1010, 923, 864, 857, 772, 748, 730, 703, 625, 595, 461, 418, 393, and 262 cm1 ; ~:=~H 258 m (E 4200); CDCl TTMS 3 3.35 (complex pattern, 2 H, 4 and 5 protons on cycloheptatriene ring), 3.80 (complex pattern, 2 H, 3 and 6 pro tons on cycloheptatriene ring), 4.48 (complex pattern, 2 H, 2 and 7 protons on cycloheptatriene ring), 5.0 (broad doub let, 1 H, nitrogen proton), 5.83 (quartet, 1 H, methyne pro ton), and 7.18 (singlet, 6 H, methyl protons); m/e (70 ev) 178 (molecular ion), 163, 133, 132, 106, 104, 91, 72 (base peak), 65, 46, 45, 44, and 15; metastable peaks, 149 and 99.5. Anal. Calcd for c 10 HJ_ 4 N 2 O: C 67.39; H, 7.93; O, 8.98. Found: C, 6 7 2 3; H, 7 9 4 ; 0, 9. 11 AA. N, N-Dimethyl-N 1 -ni troso-N '-cycloheptatrienylurea. A mixture of 500 mg (2.81 mmoles) of the urea, 2.30 g (28.1 mmoles) of sodium acetate, and 0.5 g of sodium sulfate in 20 ml of dry dichloromethane was cooled to -25 with stirring. A solution of 0.29 M dinitrogen tetroxide in anhydrous ether was added in 3 ml portions until the green color persisted for at least 5 min. The excess dinitrogen tetroxide and ca. 10 ml of dichloromethane were removed in vacuo while maina taining the temperature between -25 and -30 The nitrosoo urea decomposed rapidly above O so the following steps were carried out as quickly as possible. The reaction mix ture was added to 10 ml of cold anhydrous ether then washed twice with ice-cold 10% aqueous sodium bicarbonate solution

PAGE 90

84 and twice with ice-cold aqueous salt solution. The organi c layer was quickly cool ed to -20, and sodium sulfate was added. After 15 min, the mixture was filtered to give an orange solution which was k ep t below 40 and used without further purification; -rS~ 1 Csl 3 (-2 o 0 ) 2. 9 4-3. 30 (poorly resolved trip11 let, 2 H, 4 and 5 protons on cycloheptatriene ring), 3.44 3.84 (compl.ex pattern, 2 H, 3 and 6 protons on cyclohepta triene ring), 4.36-4.76 (complex pattern, 2 II, 2 and 7 pro tons on cycloheptatriene ring), 5.29 (triplet, 1 methyne proton), and 6.71 and 6.80 (singlets, 6 H, methyl groups). Upon warming to 10, the cycloheptatriene ring pattern broad ened and lost all resolution. The methyl peaks disappeared, and two new singlets appeared at T 6.96 and 7.61. BB. Dec~mrosition of N,N-Dimethyl-N'-nitroso-N'-cycloheptatrienylurea with Sodium Methoxide. To a solution of the nitrosourea prepared from 200 mg (1.15 mmoles) of the urea in 30 ml of anhydrous ether at -20 were added 300 mg (5.55 mmoles) of sodium nethoxide and 2-3 drops of methanol. The dark mixture was allowed to warm slowly to room temperature with stirring. After filtration, the filtrate was taken to dryness on a rotary evaporator. The oily residue was taken up in 5 ml of 5% ether in hexane and chromatographed on 12 g of basic alumina (Woelm, activity grade III) using 5% ether in hexane as eluent. Heptafulvalene was eluted as a dark red band, and uv determination of the yield showed 2.1 mg (2.1%, based on starting urea). CC. ~_ecomp_osi t ion of N, N-Dimethyl-N' -ni troso-)1' -cyclohepta

PAGE 91

85 trienvlurea with Sodium Hethoxide in the Presence of Dime thy]. Fumarate. To a solution of the nitrosourea prepared from 125 mg (0: 70 nunoles) of the u r eu. in 25 ml of anhydrous ether at -20 was added a cold (0) mixture of 40 ml of anhy drous ether, 1.0 g (7.65 nunoles) of dimethyl fumarate, 300 mg (5.55 nunoles) of sodium metho x ide, and 2-3 drops of methanol. The dark reaction mixture was allowed to warm slmvly to room temperature with stirring. After filtration, the filtrate was taken to dryness on a rotary evaporator. The crude solid was taken up in 5% ether in hexane and chromatographed on 20 g of basic alumina (Woelm, activity grade III) using 5% ether in hexane. Elution of the spirononatriene was fol lowed by glc on a 5', 20% SE30 column at 160 and an inlet pressure of 20 psi. Using S-methoxynaphthalene as an inter nal standard, the yield was determined by glc to be 0.80 mg (0.5%, based on starting urea). DD. CarboxyE__ycloheptatrienylium Fluoroborate. A solution of 0.500 g (3.68 mmoles) of 2,4,6-cycloheptatrienecarboxy lic acid in 10 ml of dichlorornethane was added dropwise over a 30 min period to a stirred solution of 1.50 g (4.55 mmoles) of trityl fluoroborate (78). The dark mixture was then refluxed for 1.5 hr under nitrogen. The mixture was filtered to give a light brown solid which, after drying in vacuo, weighed 665 mg (82%). The crude carboxytropylium fluorobo rate charred at 150, and melted with gas evolution at 165. After several recrystallizations from nitromethane-benzene, an analytical sample was obtained which charred at 153 and

PAGE 92

melted with gas evolution at 165; VKBr 3010, 2760, 2563, max 86 2420, 1722, 1472, 1439, 1384, 1214, 1070, 850, 792, 734, 686, -1 534, 522, and 488 cm \EtOH 272 m (s 5700) CH3N02 max TTMS -0.1 (broad singlet, 1 H, carboxylic acid proton), 0.0-0.4 (complex pattern, 2 H, 2 and 7 protons on cycloheptatriene ring), and 0.5 (broad singlet, 4 H, 3,4,5, and 6 protons on cycloheptatriene ring). Anal. Calcd for c 8 H 7 o 2 BF 4 : C, 43.25; H, 3.16. Found: C, 43.52; H, 2.99.

PAGE 93

LIST OF REFERENCES 1. J. Hine, "Divalent Carbon," Ronald Press, New York, 1964. 2. W. I
PAGE 94

88 16. R. Breslow, LT __ !,m~ Chem. S~~.:..:_, _?_0_, 3719 (1958). 17. H. W. Wanzlick and H.J. Kleiner, Anaew. Chem., Internc1 t. Ed --~E-_g 1 ~, 6 5 ( 19 G 4} -~-18. H. Quast and S. Hunig, Angew. Chem. Internat. Ed. Engl., I, soo (1964). 19. J. Metzger, et al., Bnll. Soc. Chim. Fr., 1964 2857. 20. H. W. Wanzlick, et al., lmcrew. _C:he~_Internat. Ed. Engl., ~, 126 (1966). 21. H. Wahl and J. J. Vorsanger, Bull. Soc. Chim. Fr., 1965, 3359. 22. H. Balli, Angew, Chem., Internat. Ed. Engl., l, 809 (1964). 23. H. Quast and E. Frankenfeld, Engl., i, 691 (1965). Ancrew. Chem., Internat. Ed. -"~-----'---------24. U. Schollkopf and F. Gerhart, Angew. Chem., Intcrnat. Ed. Eng]., 805 (1967). 25. D. M. Lemal, E. P. Gosselink, and A. Ault, Tetrahedron Lett., 19, 579. 26. R. W. Hoffmann and H. Hauser, Tetrahedron Lett_:_, 19_i_~, 197. 27. R. \. Hoffmann and H. Hauser, Tetrahedron, _21, 891 (1965). 28. D. M. Lemal, E. P. Gosselink, and S. D. McGregor, ~._ __ l=nL_ Chem. Soc.,~' 582 (1966). 29. R. W. Hoffmann and H. Hauser, Tetrahedron Lett., 19~_!, 1365. 30. R. J. Crawford and R. Raap, Proc. Chem. Soc., 1961_, 370. 31. W. M. Jones and J.M. Denham, J. Am. Chem. Soc.,~' 944 (1964). 32. W. M. Jones and il. E. Stowe, 'I'etrahedron Lett., 1964, 3459. 33. M. E. Stowe, Ph.D. Dissertation, University of Florida, August, 1967.

PAGE 95

89 34. R. Breslow and L. J. Altman, J. l\In. Chem. Soc.,~, 504 (1966). 35. I. 1.r,oritani, ~-~~!.:., :!'.~~cr_"?.:!1edron Lett., 1966, 373. 36. I. Moritani, et a~_., J. __ .-Z\rn. __ Chem. __ Soc., ~, 1259 (1967). 37. s.-I. Murahashi, I. Moritani, and M. Nishino, Chem. Soc_., ~-' 1257 (1967). 38. Private communication with R. Hoffmann. J. Am. 39. A preliminary account of this work has already appeared. W. M. Jones and C. L. Ennis, J. Am. Chem. Soc_., 89, 3069 (1967). 40. W.R. Bamford and T. S. Stevens, J. Chem. Soc., 195~, 4735. 41. P Radlick, ~_Qrg. Chem., ~, 960 (1964). 4 2. Cf. D. M. G. Lloyd, "Carbocyclic Non--benzenoid }\rorna tic Compounds", Elsevier Publishing Co., Nm York, N. Y. 1966, p. 135. 43. N. L. Bauld and Y. S. Rim, J. Am. Chem. Soc.,~, 6763, (1967). 44. H.J. Dauben and D. F. Rhoades, J. Am. Chem. Soc.,__, 6764 (1967). 45. T. Mukaki, Bull. Chem. Soc., Japan, ]l_, 238 (1960). 46a. G. Sunnagawa and N. Soma, Japanese Patent 12674 (1962); Chem. Anstr., _Q_, P4064h (1964). 46b. Y. Kitihara, T. Asao, and M. Funamizu, Japanese Patent 11629 (1964); Chem. Abstr., 61, Pl602le (1964). 47. Y. Kitihara, T. Asao, M. Funamizu, Japanese Patent 11628 (1964); Chem. Abstr., __!, Pl602le (1964). 48. W. Kirmse, B.-G. Von Bulow, and H. Schepp, Anal. Chem., 691, 41 (1966). 49. C. G. Overberger, J.-P. Anselme, and J. G. Lombardino, "Organic Compounds with Nitrogen..;Nitrogen Bonds," Ronald Press, New York, 1966. 50. R. Baltzly, et al., J. Org. Che~.,.?.._, 3669 (1961).

PAGE 96

90 51. D. G. Farnum, ~! :__9_1:9. Che~ ? :.. ., 8 7 0 ( J. 9 6 3) 52. W. von E. Doering in "Theoretical Organic Chemistry. The Kekule Symposium," Academic Press, Inc., New York; N. Y., 1959. 53. A. Streitwieser, Jr., "Molecular Orbital Theory for Organic Chemists," John Wiley and Sons, Inc., New York, N. Y., 1961. 54. D. ~1. Patel, M. E. H. Howden, and J. D. Roberts, J. Am. Chem. S?C., ~2_, 3218 (1963). 55. R. Huisgen, Angew. Chem., Internat. Ed. Engl., I, 633 (1963). 56. w. M. Jones, T. H. Glenn, and D. G. Baarda, J. Org. Che~., ~' 2887 (1963). 57. P. S. Skell and R. C. Woodworth, J. run. Chem. ~o~., 78, 4496 (1956). 58.. P. P. Gaspar and G. S. Harmond in "Carbene Chemistry," w. Kirmse, Ed., Academic Press, New York, N. Y., 1964, Chapter 12. 59. M. Jones, Jr. and K. R. Rettig, J. Am. Chem. Soc., 87_, 4015 (1965). 60. M. Jones, Jr. and K. R. Rettig, J. Am. Chem. Soc., 87, 4013 (1965). 61. M. Jones, Jr., A. Kulczycki, Jr., and K. F. Hun~el, 'l'etrahecl.ron Lett., 1967, 183. 62. F. A. L. Anet, J. Am. Chem. Soc., ~' 458 (1964). 63. Cf. K. Alder and G. Jacobsj Chem. Ber., 86, 1528 (1953), K. Alder, R. Muders, W. Krane, and P. Wirtz, J'.,_nn. 6 2 7, 59 (1960) and earlier references cited therei~ -64. This reaction was suggested by Prof. J. A. Berson at the Univ. of Wisconsin. 65. H. E. Simmons and T. Fukunaga, J. Am. Chem. Soc., ~2_, 5208 (1967). 66. R. Hoffmann, A. Imamura, and G.D. Zeiss, J. Am. Chem. ~oc., ~' 5215 (1967). 67. D. M. Lemal and S. D. McGregor, J. Am. Chem. Soc., 88 1335 (1966).

PAGE 97

91 68. S. D. McGregor and D. M. Lemal, J. Am. Chem. Soc., ~ ~' 2858 (1966). 69. G. M. Kaufman, J. A. Smith, G. G. Vander Stouw, and H. Shechter, J ._Am._Chem._Soc., 8_7, 935 (1965). 70. J. W. Cook and E. F. M. Stephenson, J. Chem. Soc., 1949, 842. 71. R. B. Wood w ard and R. Hoffmann, J. Am. Chem. Soc.,_]_, 2511 (1965}. 72. J. A. Berson and G. L. Nelson, J Arn. Chem. Soc., 5503 (1967). 73. W. von E. Doering and L. E. Helgen, J. Chem. S_?_~, 1961, 482. 74. D. N. Kursanov, M. E. Vol'pin, I. S. Akhrem, and I. Ya. Kachkurova, Izvest. Akad, Nauk S.S. S.R., Otdel. Khim. Na:-1k, 19 5 7, 13 71; Chem. Abs tr. 5 2, 7 26n; -cr-5"T~--75. C. M. OrJ ando, Jr. and K. Weiss, ~: __ '?r'l.. Chem., ]:].__, 4715 (1962). 76. A. W. Johnson, A. Langemann, ar.d M. 'l'isler, J. Chem. ~oc., 1955, 1622. 77. This compound was first prepared by W. M. Jones. 78. H.J. Dauben, Jr., L. R. Honnen, and K. M. Harmon, J. Org. Chem_., I~, 1442 (1960).

PAGE 98

BIOGRAPHICAL SKETCH Cecil Lawrence Ennis was born August 14, 1943, in Auburn, Alabama. He was graduated from Auburn High School in June, 1960 and entered Auburn University the following September. While there he was a member cf Phi Lambda Upsi lon chemical honorary society. He obtained his Bachelor of Science degree in chemistry in June, 1964 and enrolled in the Graduate School of the University of Florida the following September. He was an Arts and Sciences Feliow and Gulf Oil Fellow during his graduate study. He is presently a member of The American Chemical Society and The Chemical Society of London. Mr. Ennis is married to the former Linda Rae Hayes. He is a First Lieutenant in the U.S. Air Force and will enter active duty in April, 1968. 92

PAGE 99

This dissertation was prepared under the direction of the chairman of the candidate's supervisor y commit t ee. and has been approved by all member s of that committe e It was submitted to the Dean of the College of Arts and Sciences and to the Graduate Council, and was approved as partial fulfillment of the requirements for the degree of Doctor of Philosophy. March 1968 Supervi~y Committee: \; .... '. '-. ~~\, ... -~ >, )' ,. __ ~':::,.~ ~-.!z< ::. ~ '\ Chairman 0 / J ~ ,. ~ --. -w ~ L f J1, ,__:__"' .1;c l -o e_a_n_,_c_o 1-'-1-ea ~! :_ ~ L s -;~-d_S_c_i_e_n_c_e_s_ -------------------Dean, Graduate School