Citation
Studies in the arylcyclopropene series.

Material Information

Title:
Studies in the arylcyclopropene series.
Series Title:
Studies in the arylcyclopropene series.
Creator:
Grubbs, Robert Howard
Publisher:
Robert Howard Grubbs
Publication Date:
Language:
English

Subjects

Subjects / Keywords:
Aluminum ( jstor )
Borohydrides ( jstor )
Bromides ( jstor )
Ethers ( jstor )
Hydrides ( jstor )
Lithium ( jstor )
Pyridines ( jstor )
Sodium ( jstor )
Solvents ( jstor )
Ultraviolet spectrum ( jstor )

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright [name of dissertation author]. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Resource Identifier:
021877670 ( alephbibnum )
13369029 ( oclc )

Downloads

This item has the following downloads:


Full Text









STUDIES IN THE ARYLCYCLOPROPENE

SERIES




















By
ROBERT HOWARD GRUBBS


A THESIS PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF MASTER OF SCIENCE











UNIVERSITY OF FLORIDA
August, 1965 (-.














ACKNOWLEDGMENTS


The author extends his sincere appreciation to his supervisory committee,

to the faculty of the Chemistry Department of the University of Florida, and to

his colleagues for their interest and friendly cooperation during the course of

this investigation. He takes particular pleasure in expressing gratitude to his

research director, Dr. M. A. Battiste, for his constant sincerity and willing-

ness in providing the counsel, guidance, and encouragement necessary for the

realization of this project; and to Dr. J. F. Helling for his aid in the preparation

of this thesis.

The author gratefully acknowledges the financial support of a portion of this

research by the National Science Foundation.









TABLE OF CONTENTS


Page


ACKNOWLEDGEMENTS........................................ ii

LIST OF TABLES .............................................. iv

LIST OF FIGURES ........................................ .... v

CHAPTER

I. REDUCTION OF TRIPHENYLCYCLOPROPENYL BROMIDE WITH
COMPLEX METAL HYDRIDES

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

Discussion and Results............................ 6

II. THE PREPARATION AND REARRANGEMENT OF TETRAARYLCYCLO-
PROPENES

Introduction ....................................... 19

Discussion and Results ............................. 24

III. EXPERIMENTAL

Reduction of Triphenylcyclopropenyl Bromide with Sodium
Borohydride ................................ ........ 47

Reduction of Triphenylcyclopropenyl Bromide in Pyridine
with Lithium Aluminum Hydride........................ 51

Preparation of Tetraarylcyclopropenes .................. 57

Thermal Rearrangements .............................. 65

Acid Catalyzed Rearrangements......................... 72

BIBLIOGRAPHY.......................... .......................... 74

BIOGRAPHICAL SKETCH ........................................... 76














LIST OF TABLES


Table Page

1. Reduction with Sodium Borohydride. . . . . . . . 7

2. Lithium Aluminum Hydride in Pyridine. . . . . . ... 11

3. Reaction of Triphenylcyclopropenyl Bromide with Activated
Aromatic Nuclei ......................... .. . 21

4. Grignard Preparation ..................... . 26-27

5. Reaction of Triphenylcyclopropenyl Bromide with Anisole. . 29












LIST OF FIGURES


Figure Page

1. Ultraviolet Spectra of Tetraphenylcyclopropene, (XV), (XIX), and
(XXI) . .. . . .. .. .. ........ .. .. . .. . .331

2. Ultraviolet spectrum of 3-(p-tolyl)-l, 2, 3-triphenylcyclopropene. 32

3. Ultraviolet spectrum of 3-(o-anisyl)-l, 2, 3-triphenylcyclopropene. 33

4. Ultraviolet spectrum of 3-(p-chlorophenyl)-1, 2, 3-triphenylcyclo-
propene . . . . . . . . . . . . . . . . 34

5. Mechanistic Scheme for Thermal Rearrangement of Tetraphenylcyclo-
propene ............. .................. 39












CHAPTER I

REDUCTION OF TRIPHENYLCYCLOPROPENYL
BROMIDE WITH COMPLEX METAL HYDRIDES

INTRODUCTION

The recent interest in non-benzenoid aromatic compounds has led to the prepa-
1-6
ration of various stable carbocyclic cations in the cyclopropene series. The

need for a convenient route to the parent hydrocarbons from these cations is

obvious. Not only are the parent hydrocarbons needed for comparison purposes

but they are inherently interesting themselves. The recent report of Breslow on

the preparation of triphenylcyclopropene from triphenylcyclopropenyl bromide by

lithium aluminum hydride reductionI and the work by Brown and Bell on the

sodium borohydride reduction of carbonium ions formed under solvolytic condi-

tions indicate that the reduction of a stable carbonium ion with complex metal

hydrides could provide useful synthetic information as well as further insight into

the chemistry of these reducing agents. The reduction of stable cations employing

complex metal hydrides allows observation of the products arising directly from

hydride attack without the complications inherent in reductions involving carbonyl

groups.

The triphenylcyclopropenyl cation (I) was chosen for a study of this nature

since many of its reaction products have been identified, and the majority of these
1,6
are solids. Triphenylcyclopropene (II) has also been found to be an excellent

dienophile, and has found uses in this role as an intermediate in the preparation

of heptaphenyltropylidene and various bicyclic systems. 8 This added utility

demands a convenient high yield route to this species.
1






2

Sodium borohydride would, at the outset, appear to be a poor reducing agent

for triphenylcyclopropenyl cation since it has been found that the cation (I) under-

goes irreversible ring opening reactions, 6 and/or ether formation in protic sol-

vents and that sodium borohydride behaves as a poor reducing agent in aprotic

solvents.

The triphenylcyclopropenyl cation (I) is converted entirely to its covalent

ether (III) at spectroscopic concentrations (ca. 10-4) in 95% ethanol.3



S + ROH H (1)

III

H2


1 OH- 0 H20 (2)
H IV
III

The percentage of non-solvated cation (I) in solution will be determined by the

original concentration of cation, the pK of the cation in that solvent system, and

the pH of the solution. For example, calculations using pK (2.8)3 of the cation (I)

in 23% aqueous ethanol show that the ratio of cation (I) to covalent ether (III) in a

0.10 molar 23% aqueous ethanolic solution of cation (I) is nearly 9:1. The trend in

the pK's (the pK of the cation in water is 3.1)3 indicates that the percentage of the

covalent ether (III) would be somewhat greater under non-aqueous conditions. With

these considerations in mind, the at first surprisingly successful reduction (75-85%

yields) of triphenylcyclopropenyl cation (I) by sodium borohydride in ethanoll7 is

understandable. This results indicates that sodium borohydride is potentially






3

a good trapping agent for stable cations under solvolytic conditions and promoted

the further investigation of reductions of the cation (I) with sodium borohydride in

both aqueous and anhydrous, protic solvents.

Earlier work by H. C. Brown0 indicate that ketones were reduced under

anhydrous conditions by sodium borohydride in diglyme, but more recent work has

indicated the actual rate of reaction is very slow and that the major portion of the

isolated products arises in the aquoeus work up. The possibility that the solvated

cation (I) might be more easily reduced by sodium borohydride that ketones and the

fact that heterogeneous reductions of the cation (I) with lithium aluminum hydride

in ether give good yields of triphenylcyclopropene (II)1 suggested that the investiga-

tion of sodium borohydride reductions of (I) in ethers and other aprotic solvents

might be fruitful, both from a mecahnistic as well as synthetic point of view.

Hydroboration of the cyclopropene double bond was anticipated to be a possible side

reaction in aprotic solvents.

Lithium aluminum hydride in ether is a more powerful reducing agent toward
14
most functional groups than sodium borohydride in protic solvents,4 but the

insolubility of the cation (I) in ether solvents impairs the ability of lithium aluminum

hydride to bring about the desired reductions. It was hoped that an aprotic solvent

could be found that would give a homogeneous lithium aluminum hydride-cation (I)

solution, increase the yield of cyclopropene (II) and cut down the required reaction

time. The successful reduction of ketones with lithium aluminum hydride in
11-15
pyridine by Lansbury suggested that pyridine would be an excellent solvent for

the desired reductions. Lansbury's work with the reduction of various ketones with

lithium aluminum hydride in pyridine resulted from the need for solvent systems








for reduction of compounds which were insoluble in ethers and could not be re-

duced by sodium borohydride in protic solvents. It was found that ketones were

readily reduced by lithium aluminum hydride in pyridine on addition of the reducing

agent to a pyridine solution of the ketone. Studies of ketones which were readily

cleaved, those where the living anion would be fairly stable, indicated that the

lithium aluminum hydride solutions were basic enough to bring about the cleavage.14

The suggested scheme for the reduction and cleavage of a ketone in fresh lithium

aluminum hydride-pyridine solution is:14

SO A1H C5H5N O
II 5 3 5 5 1
R-C C 0 + A1H --- R C C3 RY CC + C5H5N:A1H
H H



O (3)
R- ---3 R-C-H + 03-C-
I 3 3
H

To further demonstrate the basicity of lithium aluminum hydride in pyridine,
16
it was used to generate the trityl anion from triphenylmethane.6 The basicity of

the lithium aluminum hydride-pyridine solution decreases with age. This was

demonstrated by the addition of an aged solution of lithium aluminum hydride in

pyridine to a colored solution of trityl anions. The color due to the anion vanished.

This result suggests that the nature of the reducing species in the solution is

changing with age.

It was found that an aged (24 hours) solution of lithium aluminum hydride in

pyridine was a more selective reducing agent than a fresh solution in that it reduced

aryl ketones in preference to alkyl ketones. 11,12 The mechanistic details of the








reaction have not been determined, although nuclear magnetic resonance spectra

studies indicate that the dihydropyridyl aluminum complex (XI) is formed. This

species (XI) can be crystallized from concentrated, aged (24 hours) solutions.


The ratio of 1, 2-dihydropyridyl to 1, 4-dihydropyridyl is approximately 1:1.

It was felt that the study of the reduction of a stable cation by this species

would be of interest in that it could possibly provide a convenient system for small

scale reductions of the cation (I) and give further insight into the nature of the

reagent.


Q\7








DISCUSSION AND RESULTS


As can be seen from Table I, the sodium borohydride reductions all gave the

same products but in different proportions. A formalized reaction scheme which

shows the mode of formation of the observed product is:





(I)+ ROH ,H
Br (III) (4)
R

BH4 H

BH2 (IV)0H




(II) (V)



As seen from the scheme, the three reactions competing with reduction in protic

solvents are: reversible ether or carbinol formation, irreversible ring opening,

and decomposition of the hydride by solvent. To minimize these side reactions the

reductions were attempted in aprotic solvents. The fact that the two reduction

reactions in anhydrous ether solutions, one in which sodium borohydride is soluble

(diglyme 0) and one in which it is not, gave the same yield (25%) suggested that

reduction could be occurring during work up. This possibility was investigated by

quenching an ether-sodium borohydride reaction after only five minutes of stirring

with 5% hydrochloric acid solution (pH 1). The results indicate that the major

portion of the reduction must have occurred during work up. In fact the yield (55%)









TABLE 1

REDUCTIONS WITH SODIUM BOROHYDRIDE


Reaction Ratio of Equiv. Yield of Yield of
Time cation/ BH Triphenylcyclopropene Ketone


Ether 24 hr. 1:1 24 21

Anhydrous
diglyme 3 hr. 1:1.2 26 trace

60% Aqueous
diglyme 3 hr. 1:1.2 77 trace

Water pH 1 5,min. 1:1 55 ---

Water pH 9 15 min 1:1 13 --

Pyridine 3 hr. 1:1 70 trace






8

from this reaction was higher than that from the anhydrous reactions which had

been stirred for periods up to 24 hours, although the yield is very close to that

obtained by Battiste and Breslow76 (50%) from a similar reaction. The difference

in yields is probably due to variations in the quenching rates and the amount of

hydrogen bromide in the cation (I) which reacts with the borohydride before water

is added. Somewhat surprising is the fact that the cation (I) can compete so well

for the hydride (55% yield of cyclopropene II) in a strongly acidic solution. When

the pH of the solution was raised to 9, only a 13% yield of cyclopropene (II) was

obtained. Although the yield is low and the reaction was not run under equilibrium

conditions, this yield is significant in that the cation (I) is hydrolized very rapidly

under basic conditions. 6 It indicates that sodium borohydride can compete with

hydroxide for the cation and is a fairly good trapping agent for a stable cation even in

basic solution.

The good yield of triphenylcyclopropene from sodium borohydride reduction in

aqueous diglyme gives considerable information about borohydride reductions of

cations. First, as in the ketone reductions, 6the protic solvent appears to aid in

hydride transfer,7,10 but most interesting is the fact that the reduction proceeds

almost to the exclusion of hydrolysis of the cation despite possible destruction of

the borohydride ion in 60% aqueous diglyme solution (pH 5).

Studies of the reduction of ketones by sodium borohydride in pyridine by Richie

and others19 26 indicate that these reductions are very sluggish under anhydrous

conditions, but that the reaction is fairly rapid if lithium borohydride is used or if

lithium ion is added to the sodium borohydride solutions. The effect of the lithium

ion was rationalized to be due to its ability to increase the polarity of the carbonyl

group or to increase the dissociation of the sodium borohydride in solution.19 The







9

reduction of triphenylcyclopropenyl cation (I) in pyridine with sodium borohydride

gave good yields of triphenylcyclopropene (II) (70%). The reaction was followed by

ultraviolet spectroscopy and found to be complete within 30 minutes. This shows

that sodium borohydride is a reactive species in pyridine solution and suggests that

the role of the lithium ion in the reduction of ketones, in the same system, is to

enchance the polarization of the carbonyl group. The isolation of triphenylcyclopro-
18
pene in high yields indicates that the by-product of the reaction, C H N:BH 18

is not an effective hydroborating agent, at least towards cyclopropene double bonds.

The reductions of triphenylcyclopropenyl cation (I; Br-) with lithium aluminium

hydride in pyridine gave surprisingly low yields of triphenylcyclopropene. The

yields from fresh solutions were 8-9% and from aged solutions, 2-4%. The major

portion of the reaction mixtures was inseparable oils, which were mainly ketonic

and were similar to those oils obtained from hydrolysis of the cation (I) itself. The

major cyclopropenyl component, obtained in 8-15% yields, was identified as 3, 5-

bis(triphenylcyclopropenyl)pyridine (VII). Structure assignment was made as follows.

The compound (VII) showed absorption in the infrared at 5. 5 (characteristic of

1, 2-disubstituted cyclopropene) and an ultraviolet spectrum which was character-

istic of 1, 2-diphenylcyclopropene chromophore. The analysis and molecular weight

(620, osomometer) satisfy the formula C47H38N. Using the calculated molecular

weight of 612, the molar extinction coefficient (53, 000) of the 316 millimicron band

is approximately twice that for 1, 2-diphenylcyclopropenes indicating that there are

two triphenylcyclopropenyl rings per molecule. The analytical and spectral data

are consistent with a molecule made up of two triphenylcyclopropene moieties

attached somehow to a pyridine nucleus. Compound (VII) showed three major peaks







10

in the nuclear magnetic resonance spectra. The tau values for these peaks were

1.43 doublett), 2.31 (complex multiplet, and 2.75 tau (complex multiplet, with

the area ratio of the first peak to the second being ca. 9:2. Tetraphenylcyclopro-

pene shows a similar band at 2.27 tau which corresponds to the ortho protons on

the phenyls conjugated with the double bond. Then with two triphenylcyclopropene

rings in the molecule as shown by the analytical data, the multiple at 2.31 tau in

VII should correspond to no less than 8 protons. The tau values for pyridine are

21
1.50 (ortho protons), 2.64 (para proton), and 3.10 (meta protons). The pertinent
22
coupling constants being J = 4 cps, J = 1.9 cps. The position, area and
2-6 2-4
splitting (since a constant of 0.4 would not be seen) indicate that the 1.43 tau peak

of VII corresponds to two ortho protons on the pyridine ring. Possibly the added

area in the 2.31 tau peak is due to the one remaining proton on the pyridine nucleus.

This reduces the number of possible structures to two, the 3, 5- or the 3,4-bis-

(triphenylcyclopropenyl)pyridine. The similarity of the physical properties,

especially the melting point, and the spectra of 1, 3-bis(triphenylcyclopropenyl)-

benzene to those of VII suggested that the correct structure was the 3, 5- and not

the 3,4- structure. The following mechanistic considerations also support this

structure. The addition of the cation (I) to the pyridine nucleus is similar to the

addition of cation to other activated nuclei. 38 The calculated electron densities for

pyridine, as well as experimental observations, indicate that electrophiles should

attack at the 3 and 5 positions. The electron density at these positions should be

even greater in a dihydropyridine. All of these considerations are consistent with

3, 5-bis(triphenylcyclopropenyl)pyridine (VII).













TABLE 2


LITHIUM ALUMINUM HYDRIDE IN PYRIDINE


Yield
Age of Reaction Equiv. Yield Yield of of
Solution (hr.) Time (hr.) cation/LAH Triphenyl Cyclopropene VII Ketone

0
Solid added 3 1:1.66 7% 8 % 20

0.5 3 1:1 8% 9.1% 17

24 3 1:1 2% 10 %+ 21

72 3 1:1 4% 8.5% 20
(15%)







12
The nature of the lithium aluminium hydride solutions will be examined before

a mechanistic scheme is suggested for the formation of the observed products. The

reaction of lithium aluminium hydride with cyclopropene (II) in pyridine provides

insight into the nature of the nucleophilic species present.

It was found that triphenylcyclopropene (II) was dimerized by excess lithium

aluminium hydride if the solid hydride was added to a pyridine solution of the

cyclopropene (II) but no dimerization was observed if the lithium aluminium hydride

pyridine solution had been aged before the cyclopropene (II) was added. Breslow

and Dowd1 found that triphenylcyclopropene was readily dimerized in a strongly

basic solution such as sodamide in liquid ammonia and these authors suggested the

following mechanism which does not involve formation of the triphenylcyclopropenyl

anion.




SNH H (5)

--o --


II VII






The same scheme should also apply in the lithium aluminium hydride-pyridine reac-

tion since a fresh lithium aluminium hydride-pyridine solution is less basic than

liquid ammonia-sodamide solution, as evidenced by the low yield of trityl anion

formed in this solution, 1and proton abstraction from triphenylcyclopropene would

be even less likely. These reactions then allow examination of the nucleophylic

species present in lithium aluminium hydride solution in a case where there are






13

no acidic or reducible species involved.

The mechanism (6) suggested by Lansbury for the metalation of acidic hydro-

carbons in pyridine lithium aluminum hydride solutions involves A1H4 as the

basic species which donates a hydride to the acidic proton with formation of a
14
molecule of hydrogen, the anion of the hydrocarbon, and AH314


(6)

03C-H + H-AIH3 + : 03C- + A1H3N + H2




Lansbury argues that pyridine is better able to solvate the alane (A1H3) than

ether. Thus enhancing the basicity of lithium aluminum hydride over that usually

observed in ether solvents. The basicity of the solution decreases with age to the

point that it appears able to donate a proton to a triaryl- or diarylmethide ion.16

The suggested rational for this behavior is that the hydrogen remaining on the

aluminum after the complexation reaction has proceeded to the point that three

dihydropyridyls have added to the aluminum, becomes acidic, due to the electron
16
withdrawing effect of the dihydropyridyls, and protonates the anion.6 His results

show that there is an initial build up and subsequent reduction of the amount of

anion formed. Using fluorene as the organic acid, he found that after 1 hour 52%,
16
5 hours 71%, and 18 hours 48% metalation had occurred.

Lansbury's mechanistic scheme (6) cannot be applied to the dimerization of

triphenylcyclopropene. If the basic species were A1H4 as suggested, the attack

of this species on the double bond of triphenylcyclopropene (II) would result in

reduction. This being the case there must be another strongly basic species present.





14

As can be seen from the reductions of ketones with lithium aluminum deuteride11

in pyridine, there is a considerable amount of the species XII formed within 3

hours







A1H3 1 Li + -- Li + 1H,

XII XIII







This species, due to the solvating power of pyridine toward alane (A1H ), should

dissociate to a considerable extent producing dihydropyridyl lithium (XI) which

should catalyze the dimerization of triphenylcyclopropene. As the solution ages

the concentration of XII decreases due to the formation of XI therefore decreasing

the concentration of Xm assuming that XI is dissociated to a small extent. This

assumption seems reasonable since the bulky groups that would remain attached to

the aluminum on dissociation would decrease the solvation of the resulting aluminum

species. The electron withdrawing effect on the dihydropyridyls would also favor

association. These considerations then suggest that the basic species in fresh

lithium aluminum hydride pyridine solution is XII or XIII.

The reported results from the reaction of lithium aluminum hydride in pyri-

dine with acidic hydrocarbons can be readily explained with this species serving as

the base. The following scheme, suggested by the above considerations, is consis-

tent with the apparent reversibility of the reaction and the effect of excess lithium

aluminum hydride on the reaction. (Excess lithium aluminum hydride gives a









faster rate of anion formation but has little effect on the maximum concentration of

anion produced). 16



N Li + 03CH > N-H -9-. 03C-Li+


(8)




-AlH 3 Li+ -A H2- N Li XI

XII



An aged solution of XI would equilibrate with XII to reduce the basicity of the solu-

tion due to the reasons given above. This is in agreement with the observed quench-

ing of a solution of trityl anions by aged lithium aluminum hydride-pyridine solu-
16
tions.

The above considerations indicate that there is a dihydropyridyl alumino com-

plex, either XI or XII, present in both aged and fresh lithium aluminum hydride

pyridine solutions. There are three points of attack open to the cation (I) on the

dihydropyridyl nucleus: (1) the 2 or 4 position to remove hydride and give cyclo-

propene, (2) the nitrogen to effectively trap out the cation and yield ketone on

hydrolysis, (3) or on the 3 and 5 carbons to yield VII. This scheme


Ba


Br b+
c






16




a 4 AI-Br C5H5N



(9)


Attack on N H20 other
ketones








VII



is consistent with the results, but of course, it cannot be considered a proven

mechanism. The similarity of product distribution of those species arising from

routes other than reduction from the different solutions suggests that the equili-

brium (7) plays a small role in these reactions and the reactivity of the cation

toward the pyridyl species is essentially the same whether the pyridyl is com-

plexed with the aluminum or is a free anion. This is reasonable on the basis of
26
the small steric requirement of XI found in reduction of ketones, 2and the

apparently small steric requirements for reduction of the cation as found in the

reduction of I with lithium tritertiarybutoxyaluminohydride.

The difference in yields of triphenylcyclopropene (II) from fresh (8%) and

aged (2-4%) solutions is probably due to reduction of the cation (I) by the lithium

aluminum hydride in the fresh solution before the reducing agent reacts with the

pyridine.






17

The high yields of ketone would be expected from a scheme of this nature

since the nitrogen would be the preferential point of attack on the dihydropyridyl

by an electrophile since it has a non-bonding pair of electrons. This part of the

scheme is also supported by the fact that the cation (I) is converted in pyridine

solution with an added secondary amine base to a species which appears to give a

ketone on hydrolysis.

It would be expected that the major product from reaction (9c) would be the

mono-alkylated product. The most obvious reason for not observing this product

is that it was obscured in the complex oils, although the ultraviolet spectra of the

oils remaining after crystallization showed only trace amounts of cyclopropenyl

material and this showed a spectrum identical to that of VII, but presumably the

mono-substituted pyridine would absorb in the same region. At any rate, it is

evident that the bis(triphenylcyclopropenyl)pyridine was the major cyclopropene

product formed in the reduction other than triphenylcyclopropene. More specu-

lative explanations are that the bulky triphenylcyclopropenyl group once bonded to

the dihydropyridyl nucleus distorts the alumino complex to such an extent that the

mono-alkylated dihydropyridyl is easily alkylated again, or that the dihydropyridyl-

1, 2, 3-triphenylcycloporpene is unstable under aqueous conditions and does not

aromatise as the bis-(triphenylcyclopropenyl)dihydropyridinyl does and yields a

complex mixture of products.

Another mechanism (10) was considered originally but was ruled out due to the

fact that sodium borohydride gave good yields of triphenylcyclopropene in pyridine.

Sodium borohydride would be expected to reduce the pyridine-cation complex

almost as readily as lithium aluminium hydride or lithium tetrakis(N-dihydro-






18

pyridyl)aluminate. The ultraviolet spectrum of triphenylcyclopropenyl bromide in


CHN HO
C5 5 LAH Ketones


(10)
Br-
LAH -0
H H N

II VII

pyridine, however, does suggest that there is some complexation of the pyridine

with the cation (I) since the spectrum shows an absorption pattern very similar to

a covalent triphenylcyclopropene.

From a synthetic point of view, it is evident from the results of this investi-

gation that the reduction of triphenylcyclopropenyl cation (I) in protic solvents are

not as complex as first anticipated and that the reduction of the cation (I) in aqueous

solvents with sodium borohydride gives high yields of triphenylcyclopropene. The

reduction under these conditions require short reaction times and give a high

purity product after a short work up procedure.

Mechanistically the study has provided insight into the nature of sodium boro-

hydride and lithium aluminum hydride in pyridine and has provided more informa-

tion about lithium tetrakis(N-dihydropyridyl)aluminate (XI) and related species.








CHAPTER II


THE PREPARATION AND THERMAL REARRANGEMENT OF
TETRAARYLCYCLOPROPENES

INTRODUCTION



The literature provides many examples of rearrangements of small
ring compounds to yield systems involving less bond angle strain, but
relatively few rearrangements have been reported in the cyclopropene
series. The following are some pertinent examples of thermal
rearrangements recently observed in the latter series.


(1)

4- Q4 CH=CHO

XXVIII
XXVII


(24)


XIV


CHICHCH
2


0
S00


(33)


These involve, except in the last instance, fairly complex pathways and









only formalized mechanisms have been proposed for each. As an example, the
mechanism suggested for the dimerization of triphenylcyclopropene and subse-
quent rearrangement of the dimer is given below.







_H-H
-/ -. --

XXVII & XXVIII
The complexity of the thermal rearrangements of these cyclopropene
derivatives prevents any conclusion to be drawn as to the mode of cyclopropene
ring cleavage and subsequent reactions of the ring fragment (s). Any attempt to
study the mechanism of such thermal rearrangements should first concern itself
with the initial ring cleavage step and the nature of the species produced. The
ideal cyclopropene derivative for a study of this nature would be one which yields
only primary products from ring cleavage.
Tetraarylcyclopropenes have no labile hydrogen, exhibit moderate thermal
stability, and have the essential features of the previous examples of cyclopropene
thermal rearrangements.
Tetraphenylcyclopropene, prepared by the reaction of phenylmagnesium
bromide with triphenylcyclopropenyl bromide, was initially investigated as a
8
possible deinophile in high temperature Diels-Alder reactions, but it was found
that at high temperatures the cyclopropene rearranged to another hydrocarbon.
Further investigation with the above mechanistic considerations in mind showed
that tetraphenylcyclopropene had rearranged after heating at 2500 for
two hours to 1,2,3-triphenylindene. The reaction was apparently




2500


XIX










TABLE 3

REACTION OF TRIPHENYLCYCLOPROPENYL BROMIDE
WITH ACTIVATED AROMATIC NUCLEI


+ ArH---


Ar= p -dimethylaminophenyl

p-methoxyphenyl

2,4-dimethoxyphenyl

2-methoxynaphthyl

anthron-10-yl


m.p. 189-190

m.p. 124-1260

m.p. 1930

m.p. 123-1250

m.p. 198-199.5


B. Fbhlish and P. Biirgle, Angew. Chem., internal. edit., Vol. 3, 699 (1964)


L'









quantitative.34 Simple mechanistic considerations indicate that 1,2,3-

triphenylindene is the primary product from the species produced from thermal

cleavage of the tetraphenylcyclopropene ring. Thus tetraphenylcyclopropene is
suitable for the desired studies.

With the large body of information available on the effect of substituents
on various reactive intermediates, it was felt that the nature of the fragment
produced from rupture of the tetraarylcyclopropene ring could best be studied

by determining the effect of substituents on the rate of the reaction. The substituents

could also serve as labels to show the geometry of the intermediate species. The

most apparent method of introducing substituents into the tetraphenylcyclopropene
system is by the reaction of the appropriate arylmagnesium bromide with
triphenylcyclopropenyl bromide (I).
Reduction of the cation (I) by the less reactive Grignard reagents40 to give
bistriphenylcyclopropene (XIV) or hexaphenylbenzene (X)24 was an anticipated
problem.




2 +- 2 ArMgX --
Br

I XIV X

In attempts to prepare octaphenyltropylidene from heptaphenyltropylium cation

and phenylmagnesium bromide, complete reduction was observed.17 Reduction
has also been observed in attempts to prepare tetraphenylmethane from triphenyl-
methyl bromide and phenylmagnesium bromide.40

Shortly after the work on this problem had begun, the preparation of several

tetraarylcyclopropenes was reported by Fdhlish and Birgle.28 These authors
report that the triphenylcyclopropenyl cation substitutes onto activated aromatic

nuclei in 10:1 acetic acid-pyridine solution. The melting points of some of their

derivatives however, were not in accord with those of the tetraarylcyclopropenes
prepared by the Grignard method. The assignedstructures and melting points of
the products of the reported reaction are given in Table III. Since little information
about the procedure or yields were given in the report, this reaction was








reinvestigated to determine if this reaction could be utilized to prepare tetraarylcyclo-
propenes not obtainable by the use of Grignard reagents and to identify the products
which appeared not to have the assigned cyclopropene structure.

The ultraviolet absorption of the diphenylcyclopropene chromaphore appears
1-5
to be affected by changes in the groups attached to the saturated carbon. This

effect is very apparent in the derivatives of triphenyl and diphenylcyclopropene.

The 3-aryl-1,2,3-triphenylcyclopropenes provide a system in which subtle

electronic and steric effects may be observed.

This study in the tetraarylcyclopropene series was considered as a three

prong problem: (1) the preparation and properties of the cyclopropenyl compounds,

(2) study the products of rearrangement, and (3) determination of the effect of the

substituents on the rate of the rearrangement. The author's original goal was the

completion of the first phase of the problem but investigation of various problems

arising in the preparations gave results which led to preliminary investigations

of the other two phases.









DISCUSSION AND RESULTS


The reaction of aryl Grignard reagents with triphenylcyclopropenyl cation (I)

gave good yields of tetraarylcyclopropenes. The yields in most cases were only
slightly lower than 50%, based on the starting cation (I). Various reaction
conditions and work-up procedures were used depending on the nature of the
particular reaction.
The cation (I) is only slightly soluble in ether solvents so the reactions are
heterogeneous. To increase the homogeneity of the reaction, tetrahydrofuran
was used initially as the solvent because the cation (I) appeared to be more soluble
in this ether than in ethyl ether and its higher boiling point should permit a faster
reaction rate. In later reactions it was found that the reactions proceeded smoothly in
refusing ethyl ether or in some cases in ethyl ether ar room temperature making
the use of tetrahydrofuran impractical since it is hard to dry. The reaction
reported for the formation of 1, 3-(bistriphenylcyclopropenyl)benzene (XXII) was
run in tetrahydrofuran since the pra di-Grignard reagent could reportedly be
42
obtained only at the higher temperatures provided by this solvent,4 but later

preparations of the bistriphenylcyclopropenyl benzene (XXII) in ethyl ether gave
yields as high as those in tetrahydrofuran. These reactions gave nearly 45%
yields of products arising from the di-Grignard, (XIX) and (XXV), suggesting
that the meta di-Grignard is not as difficult to form as para di-Grignard.
The reactions appeared to be essentially complete after 24 hours but shorter
reaction periods have given lower yields. As an example, only 19% of the desired
cyclopropene was obtained in the reaction of p-chlorophenyl magnesium bromide
with cation (I) after 6 hr.
To check the dependence of the yield on concentration of Grignard reagent in
the more sterically hindered reactions, the preparation of 3-mesityl-1, 2,3-
triphenylcyclopropene was run at two concentrations. The first reaction in which
the concentration of Grignard reagent was 0.27 M, gave 18% of the desired
cyclopropene. The second run at a higher concentration, 1. 0 M, gave a 52% yield.
The yield of the desired cyclopropene from the less hindered Grignard reagents appeared









to be less concentration dependant. Since there is a considerable amount of
hydrogen bromide associated with the bromide salt (I) from its preparation,
there was at least a 100% excess of Grignard reagent used in all reactions.
The tetraarylcyclopropene is formed before hydrolysis, therefore it can be
separated from the insoluble inorganic solids and unreacted cation (I) by filtration
of the reaction mixture before water is added.


S+ ArMgX ----- + MgXBr


This procedure gave good results but the magnesium salts usually clogged
the filter and made filtration difficult. It was found that the tetraarylcyclopropenes
could be easily separated from the hydrolysis products of (I) by column
chromatography and this made the difficult filtration unnecessary. In cases where
the hydrocarbon resulting from hydrolysis of the Grignard reagent was high
boiling, the excess Grignard reagent was quenched by pouring the reaction
mixture onto powdered Dry Ice. The Dry Ice mixture was then acidified and
extracted with ether. The acid resulting from carbonation of the excess Grignard
reagent was then extracted from the ether solution with aqueous sodium bicarbonate.
This procedure eliminated the problem of removing the high boiling liquids which
came off the chromatography column together with the tetraarylcyclopropenes.








TABLE 4


GRIGNARD PREPARATI NS


Br" + ArMgX


Ar


Ar mp. Yield (cm.-l) U.V. Spectra (cyclohexane) max (log )


R-Anisyl (XXI)


0-Anisyl (XVII)


Mesityl (XV)

-Naphthyl (XXXI)


Phenyla (XIX)




yl b (XXV)


160-1610


197-197.50


150-1510

189-1900


178-1790




235-2360


40%


44%


45%

52%


47%




17%


1810


1810


1820


230 (4.56), 300(sh), 312(4.36),
334(4.20)34

286(4.26), 303(4.31), 317(4.34),
333(4.29)c


280(4.14), 300(4.20)


.292.5(4.35), 299(sh), 305(sh), 318(4.27),
333(sh) (ca. 4.2)34


1825




1800


280(4.53), 305(4.33), 318(4.35),
3.35(4.31)



228, 302, 316(4.73), 333c


"T"


-.








TABLE 4--Continued


Ar mp. Yield (cm.-1) U.V. Spectra (cyclohexane) max (log )

p-Tolyl (XVI) 134.5-1360 51% 1830 230(4.53), 307(4.32), 318(4.30),
336 (4.26)
E-Chlorophenyl 154-1550 19% 1825 230(4.56), 304(4.33), 317(4.35),
(XVIII) 334(4.31)


aphenyl lithium used.

bPrepared from the bis-grignard of meta-dibromo benzene.

CAcetonitrile solvent.









The yield of reduction products was much lower than expected from the
considerations given in the introduction. Only trace amounts of hexaphenylbenzene
(X) were isolated from the reaction with the meta-di-Grignard and only small
amounts of impure material which appeared to contain traces of the bistriphenyl-
cyclopropene (XIV) were obtained from the preparation of XV and XVIII. Consider-
ing the steric requirements of the mesitylmagnesium bromide, it is surprising
that such a high yield of the cyclopropene derivative (XV) is obtained from the
reaction and such a small amount of reduction product is produced.
Shortly after the work on this problem had begun, Fohlish and Burgle28
reported the preparation of tetraarylcyclopropenes by the reaction of triphenyl-
cyclopropenyl bromide (I) with activated aromatic nuclei. Their results are given
in Table III. The melting point for their p-anisyl derivative (XXI), 124-126 was
approximately 400 lower than that of 3-p-anisyl-1,2,3-triphenylcyclopropene, m.p.
160-1260, prepared in our laboratory by the Grignard method.
The only reaction conditions given were that the cation (I) was heated with the

appropriate aromatic compound in 10:1 acetic acid-pyridine. In hopes of finding
this method useful for preparation of tetraarylcyclopropene derivatives not
obtainable by the Grignard method an investigation to determine the optimum
reaction conditions was undertaken. The results of this investigation are
summarized on Table V. Reasonable yields (based on starting cation (I) could be
realized only when a large excess of the aromatic reactant was used. This would
make the method practical only for micro scale preparations, especially in cases
where the aromatic compound is expensive.








TABLE 5

REACTION OF TRIPHENYLCYCLOPROPENYL BROMIDE
WITH ANISOLE


y Br-


Anisole HAc -
Pyridine


XXI


Equiv. of Equiv. of Equiv. of
Products
Reaction pyridine Anisole Cation

Cation
+ anisole 0.81 8.4 0.87 A light green fluorescent oil


Cation
+ anisole 5.3 21.0 2.7 10% XXI


Cation
+ anisole 3.7 33.6 2.7 21% XXI


Cation
+ anisole 2.8 84.0 2.7 47% XXI









The reaction was also investigated to determine the structure of the product,
m.p. 124-1260, that had been isolated and reported from the reaction of anisole
with the cation (I). Since the compound was reported to show absorption at 5.5
microns in the infrared (characteristic of a 1, 2-diphenylcyclopropene), it was
felt that the reported compound was possibly another tetraarylcyclopropene. The
high melting point of the o-anisyl derivative, m.p. 197-197.5, rules out this

most obvious possibility. The problem at this point appeared to be more complex

than had been anticipated and lead to further work which gave results that prompted

the investigation of the rearrangements of tetraarylcyclopropenes in greater detail
than had been originally planned but did not lead to a positive identification of
the reported compound. These results will be discussed later.

The ultraviolet and infrared spectra of the tetraarylcyclopropenes prepared

inthis work showed larger substituent effects than originally anticipated. The
spectral data is tabulated in Table IV. and reproductions of the ultraviolet
spectra shown in Figures 1-4. The characteristic 1,2-diphenylcyclopropene
-1 1-5
peak (1820 cm ) is present in all the infrared spectra on the tetraarylcyclo-
propenes prepared. The position of the peak appears to be affected somewhat by
the substituents, but the spectra were recorded on a Perkin Elmer Infracord
infrared spectrometer and are not sufficiently accurate for substituent effect
correlations.
The ultraviolet spectra demonstrated the effect of substituents on the
wavelength of maximum absorption as well as the overall shape of the spectrum
for the 1,2-diphenylcyclopropenyl chromophore in each of the tetraarylcyclopro-
penes. The effect can be attributed to either electronic or steric factors. The latter
are more apparent and are the most easily explained and will be discussed first.
The spectrum of tetraphenylcyclopropene (XIX) will be used as the reference in all
*the discussion. The 3-mesityl derivative (XV) (Figure 1) exhibited a blue shift
in the long wavelength absorption band and a marked decrease in the absorption
intensity for this band. Models of the compound show that there is a definite
interaction between the ortho methyls on the mesityl ring and the phenyls conjugated
with the double bond of the cyclopropene ring. Less interaction results if the









Tetraphenylcyclopropene (XIX) ---------
3-Mesityl-1,2 3-triphenylcyclopro-
pene XV) - -
3-(p-anisyl)-l,2 5-triphenylcyclo-
propene XXI) -- --


N


N -


spectra of (XV), (XIX), and (XXI).


/ \


0
O

1
(IJ


240


260 280 300 320 340
X in mu


Fig. 1.-Ultraviolet























1


0
r--



10-








5-


240


Fig. 2.-Ultraviolet spectrum of 3-(p-tolyl)-l,2,3-triphenylcyclopropene(XVI).


I 2 I i I I034
260 280 300 320 340
X in mu





























S1

0
O
r-l


X mu
Fig. 3.-Ultraviolet spectrum of 3-(o-anisyl)-1,2,3-triphenylcyclpopopen(XVII).


C-.























0






10-







5-
260 280 300 320 340
i in mu
Fig. 4-.-Ultraviolet spectrum of 3-(p-chlorophenyl)-1 ,2,3-triphenylcyclopro-
pene(XVIII).









phenyls are out of the plane of the cyclopropene ring. This would decrease the

conjugation which would, by analogy with cis and trans stilbenes,41 account for
a major part of the observed effect. Apparently, neither the steric effect nor the
electronic effect of one ortho substituent is very great. The o-anisyl derivative
(XVI) shows a spectrum which is only slightly different from that of tetraphenycy-
clopropene. This point is also supported by the fact that the 2,4-dimethoxy
derivative (XXII) shows a spectrum very similar to that of the p-anisyl derivative

(XXI).
As can be seen from Figures 1-4, the substituents in the para position effect
the wavelengths of maximum absorption very little, but have a pronounced effect
on the relative intensities of the three characteristic disubstituted cyclopropene
peaks. Although there have not been enough compounds prepared for substituent
effect-spectra shape correlation, there are enough differences in the spectra of the
p-chloro, p-methoxy, and p-methyl derivatives to allow a speculative examination
of the nature of the substituent effects. The chloro group deactivatess the ring
slightly) affects the spectra very little. The methoxy group (activates the ring
strongly) gives a pronounced change in the shape of the spectra. The methyl group
(activates the ring slightly) gives a different but very definite change in the relative
absorption intensities. It is apparent then that electron donating groups have the
largest effect on the shape of the absorption spectra. At first glance, the saturated
3-carbon would be expected to be able to transmit inductive effects very inefficiently.
The quite large effect of the substituents on the absorption spectra of the cyclopropene
chromophore is consistent with the methylene carbon having considerable sp2 character
since trigonal carbons are more efficient at transmitting inductive effects than a
tetrahedral carbon.30 The data collected to date indicates that with more examples
at hand the ultraviolet spectra of tetraarylcyclopropenes can possibly provide
further insight into the intriguing problem of the electronic structure of the cyclopro-
pene ring.
In the investigation of the reaction of the cation (I) with anisole in acetic acid-
pyridine mixtures which was discussed earlier in this chapter, it was found that a
green fluorescent material was obtained when the reaction was run with insufficient






36
pyridine present. This material did not show a typical cyclopropene ultraviolet
spectrum, but exhibited a spectrum very similar to that of 1,2, 3-triphenylindene
(XX). Crystals melting over a wide temperature range were obtained from the oils
after various solvent manipulations, suggesting a mixture of products. Since it was
known that tetraphenylcyclopropene rearranged to triphenylindene (XX) on heating
above 2000, 3-p-anisyl-1,2,3-triphenylcyclopropene (XXI) was rearranged to a
mixture of indenes by the same procedure. Attempts to completely separate the
mixture were futile. Chromatography yielded two crops of crystals that gave,
m.p. 138-1450 and 145-1480, respectively. The mixtures showed differentabsorp-
tions in the infrared and ultraviolet and each appeared to be a mixture of at least
two different isomers. The infrared and ultraviolet spectra of the higher melting
material were similar to those of the product form the electrophilic addition
reaction mentioned above. This further supported the supposition that the addition
reaction was yielding indenes and suggested two possible modes of formation of the
products.
The first possibility is similar to that suggested by Breslow and Battiste29 for
the rearrangement of diphenyl-(2,3-diphenylcyclopropenyl)- carbinol. This scheme
is as follows:

Q r --c
Br OMe
OMe OMe Y"
pyr. HOMe OMe
OMe



XXI -
The other possibility is that the tetraarylcyclopropene is formed and then rearranged
under the reaction conditions. 3-p-Anisyl-1,2,3-triphenylcyclopropene was found
to be stable in acetic acid-pyridine solution in the absence of any proton catalysis.
The fact that the desired cyclopropene could be obtained when excess pyridine was
present suggested the possibility that the hydrogen bromide associated with the
bromide salt (I) was lowering the pH of the solution to such a point that the tetra-
arylcyclopropene was undergoing an acid catalized rearrangement. When the









tetraarylcyclopropene (XXI) was heated at 900 in acetic acid with added hydrobromic
acid a green fluorescent oil was obtained which showed infrared and ultraviolet
spectra very similar to the spectra of the higher melting material obtained from the
thermal rearrangement. In contrast to this, tetrasubstituted cyclopropenes have
been shown to be very stable to base. 1This suggests that the indenes from the
electrophilic addition reaction were probably formed from the tetraarylcyclopropene
31 37
(III) by a mechanism similar to that proposed by Breslow31 and Battiste37 for the
acid catalyzed rearrangement of 1,2-diphenylcyclopropenyl carboxylic acid and by
Breslow and Wolf for the acid catalized rearrangement of triphenylcyclopropene.32
The mechanism for the acid catalyzed rearrangement of tetraarylcyclopropenes is:









XIX ; XX



This scheme is apparently general in the tri and tetraaryl cyclopropene series.
The mechanism indicates that the group attached originally to the methylene carbon
of the cyclopropene should wind up in the 3- position of the resulting indene. This
is supported by the fact that triphenylcyclopropene undergoes acid catalyzed
rearrangement to 1,2-diphenylindene (XXIX)32 and by evidence in the literature that
arylindenes are not readily isomerized by acid.27






XXI XX


XXX


XXIX









By this mechanism then only two isomers should be formed from the acid catalyzed
rearrangement of a 3-aryl-1,2,3-triphenylcyclopropene. These two isomers
would be the 3-aryl-1,2-diphenylindene (C) and the triphenylindene substituted in
the 6 position (D).




+ ++ H H
(D) X





(C)

The acid catalyzed rearrangement reactions provide a convenient source of
triarylindenes of known structure which promised to simplify the study of the
products from the thermal rearrangement of tetraarylcyclopropenes. With this
information at hand it was felt that a further study of the thermal products of
3-aryl-1,2,3-triphenylcyclopropenes should prove to be fruitful.
Tetraphenylcyclopropene had been found to rearrange to 1,2,3-triphenylindene
on heating at 2500 for 2 hours. The reaction was shown to be actually a thermal
and not an acid catalyzed reaction at the surface of the glass by running the reaction
in a caustic coated vessel.34 The same product was obtained in the same length of
time under these conditions. The preliminary rate studies demonstrate the same
point although the rate does appear to be depressed slightly by washing the reaction
tubes with base instead of distilled water before the reaction is run. These same
rate studies indicate that the rate is first order.
Mechanistic considerations suggest three possible modes of thermal cleavage of
the cyclopropene ring. There could be a heterolytic cleavage giving rise to ionic
intermediates. This process appears unfavorable since either a vinyl carboniumion
or carbanion would have to be formed. The former is energetically unfavorable and
the latter is not consistent with results to be presented later. A concerted mechanism,
another possibility, would be energetically unfavorable since the bond in the cyclopropene
ring would have to cleave at the same time as the aromaticity of the phenyl ring was













A





Shift
) o


Mechanistic Scheme for Thermal 1, 3 H
Rearrangement of Shift
Tetraarylcyclopropenes
Figure 5










being destroyed. The third possibility is a homolytic cleavage giving rise to
diradical intermediates. The homolytic cleavage would appear to be the
44
favored mode of cleavage since in analogous vinyl cyclopropane rearrangements
where there are groups present in the cyclopropane which will stabilize a radical
38
species the intermediate has been shown to be diradical in nature. The
mechanism involving homolytic cleavage with a diradical intermediate is presented
in Figure 5.
As can be seen from the mechanistic scheme (Figure 5), the thermal rearrange-

ment could yield four different isomers of the triarylindene. Even if one of the

paths to a product were unfavorable, thermal equilibration of the indenes could still
give rise to the complex mixture which would be very difficult to separate. This
was found to be the case in attempts to isolate the thermal products from the anisyl
derivative (XXI). Preliminary attempts at separation by the use of thin layer
chromatography indicate that this technique may provide a means of separation
of the various mixtures of isomers.
From the mechanistic schemes for acid catalyzed rearrangement and thermal
rearrangements of tetraarylcyclopropenes the products from the acid catalyze
reaction should be the same as those resulting from 1,2 hydrogen shift in the
thermal rearrangement. The acid catalyzed rearrangement them provide a means
of determining those products which arise from 1,2 hydrogen shift and those
arising from 1,3 hydrogen shifts.
The goals of the study of thermal products are three in number:
(1) the determination of the preferred path of hydrogen migration, (2) the effect
of product stability on the preferred path of hydrogen migration and (3) the deter-
mination of the physical properties of the products in preparation for a study of
the kinetics.
Although many difficulties were encountered in the attempt to study the thermal
and acid catalyzed products from the anisyl derivative (XXI), the results implied
that information could be obtained if a tetraarylcyclopropene were studied which

would yield fewer products. The mesityl derivative (XV) was the compound

selected since the ortho positions on one of the phenyls are blocked by methyls. This should








decrease the number of thermal products possible to two.
On heating at 2000 for two hours, 3-mesityl-1,2,3-triphenylcyclopropene (XV)
yielded two products which showed typical indene absorption in the ultraviolet and
were easily separated by solvent manipulations. Until structure assignment is made,
the higher melting isomer, m.p. 181-1820, will be designated as (A) and the lower
melting isomer, m.p. 131-1320, as (B). The ultraviolet spectra of the two compounds
first suggested that (A) was the 1-mesityl-2,3-diphenylindene (XXIII) and (B) was the
3-mesityl-l-2-diphenylindene (XXIV) since








O 0




XXIII XXIV


isomer (A) showed absorption at essentially the same wavelength as triphenylindene
and (B) absorbed at shorter wavelengths. It was felt that the bulky mesityl group
in the 3-position would decrease conjugation sterically and therefore (XXIV) would
show absorption at shorter wavelengths than 1,2,3-triphenylindene. Attempts to
isomerize B to A and A to B under basic conditions failed.43 The stability of B in
base under aerobic conditions suggested that it was not a 1-hydroindene as origi-
nally thought. When 3-mesityl-1,2,3-triphenyl cyclopropene was heated at 85-900
under acidic conditions it was converted quantitatively to (A). Mechanistic consid-
erations and the fact that 1,2-diphenylindene32 was obtained exclusively from
triphenylcyclopropene under the same conditions show that the isomer (A) is
3-mesityl-1,2-diphenylindene (XIV). This contention is supported by the results of
27
Koelch which indicate that arylindenes are not rearranged under acidic conditions.
The possibility that the product from acid catalyzed rearrangement is the thermo-
dynamically favored product is ruled out by the fact that 1,2-diphenylindene is









isomerized almost quantitatively to the 2,3 isomer on heating at 2500 for 2 hrs.
A mesityldiphenylindene prepared by a route which does not allow definite
assignment of structure showed the same ultraviolet absorption and retention time
on a thin layer chromatography plate as A. The method of preparation was:



0^Mgx^ _HI O
V_ M0gX or XXIII


XXIV
Preparation of the two isomers (XXIII) and (XXIV) is now under way.
The nuclear magnetic resonance spectra of the two isomers, A and B, were
very revealing. The isomer A showed resonance signals at 2.84 tau (aromatic)
4.73 tau (indenyl proton), a doublet at 7.65 tau (methyl protons) and a singlet at
8.12 tau (methyl protons). The area ratio of the 7.65 tau peak to the 8.12 tau peak
was 2.1. The assignment of the 4.75 tau peak was made by comparison to an m.m.r.
spectrum of 1,2,3-triphenylindene. (XX) The non-equivalence of the methyl protons
is consistent with the assigned structure. Models of the two possible isomers
(XXII and (XXIV) show that the methyls on the mesityl group of XXIII are all in
essentially the same environment but in XXIV interaction of the mesityl or the
methyls with the 4-hydrogen on the fused benzene ring as well as with orthohydro-
gens on the 2-phenyl, restrict free rotation, producing a different magnetic environ-
ment for the two ortho methyls, thus explaining their nonequivalence.
Sompound B showed three resonance signals due to methyl protons. These were
at 7.74 tau, 8.10 tau and 8.3 tau. The fact that the methyls are highly non-equiva-
lent indicates that the mesityl nucleus has been destroyed. There was no signal
in the 4-5 tau region showing that :it was not a 1-hydroindene as suspected from the
chemical data. The methyl showing a resonance signal at 7.74 tau is in an
environment similar to that of the ortho methyl in A. The 8. 10 tau signal indicates
that there is one methyl on a phenyl similar to the para proton on A which is not
being affected by steric interactions. The third signal at 8.3 tau was in the same
region as the signal from the methyl protons (8.29 tau) of 1-methyl-1,2,3-triphenyl-
indene (XXV). The structure, 1,4,6-trimethyl-1,2,3-triphenylindene (XXVI), is








most consistent with this data.

Me



MeMe 2

XXVI

The apparent mechanism for formation of (XXVI) involves a 1,2-methyl shift
as shown.






-0k -- XXVI



XV

The attack of the intermediate species on the mesityl ring at a methyl position
suggests that the species is electron deficient and is most consistent with the
diradical representation.
The isolation of the 3-mesityindene (XXIV) isomer only would indicate on first
glance that 1-2 hydrogen migration is the preferred path to product, but the lack
of information about the thermal behavior of the mesitylindenes (XXIII) and (XXIV)
precludes any definite conclusions on this point.
With the above information at hand, work was begun on the third phase of the
problem, the determination of the kinetics of the thermal rearrangement of
tetraarylcyclopropenes. The nonvolatility of both the products and reactants and
the fact that both absorbed in the same region in the ultraviolet prevented the use
of vapor phase chromatography and ultraviolet spectroscopy for following the rate
of the reaction. The observation that 1,2,3-triphenylindene (XX) fluoresced strongly
in the visible region while tetraphenylcyclopropene did not, suggested that the
reaction could be followed by fluorimetry. Initial investigations show that the
rearrangement of tetraphenylcyclopropene can be followed by this technique. Since









all the triaryl indenes prepared to date fluoresce, the method would appear to be
general. This technique, which has rarely been used in organic chemistry, could
prove to be very useful for following reactions involving a fluorescing species.
.Spectrophotometric techniques involving emission measurements are more versatile
thank absorption techniques. In absorption measurements there is only one variable,
the wavelength of light passed through the sample. In emission measurements there
are two variables, the wavelength of excitation and the wavelength of emission.
In conclusion, the work to date has provided convenient procedures for the
preparation of tetraarylcyclopropenes, given further insight into cyclopropene
rearrangements, suggested some interesting studies of triarylindenes, and has
provided a method for the investigation of the kinetics of the rearrangements of
tetraarylcyclopropenes.














CHAPTER III

EXPERIMENTAL

Physical Measurements



General

Melting points were taken on either a Fisher-Johns melting point block or a

Thomas-Hoover unimelt capillary melting point apparatus. All melting points

are uncorrected.

Carbon-hydrogen analyses were determined by Geller Microanalytical

Laboratories, Charleston, West Virginia.

Spectra

The infrared spectra were recorded on a Perkin-Elmer Infracord spectro-

photometer. Solids were examined in potassium bromide pellets and oils were

examined on sodium chloride plates.

The ultraviolet spectra were recorded on a Cary 14 recording spectrometer.

The solvents employed were suitable for use in spectrophotometry.

The nuclear magnetic resonance spectra were recorded in deuterochloroform,

with tetramethylsilane as the internal standard. A Varian DP-60 instrument was

used and the chemical shifts were calculated from measurements of sideband

spectra.






46

Procedure

Triphenylcyclopropenyl bromide (I) was prepared by the procedure of

Breslow and H. W. Chang, 3 and had been dried in a drying pistol at 500 for at

least 24 hours prior to use. All anhydrous reactions were run under an argon

atmosphere.














REDUCTION OF TRIPHENYLCYCLOPROPENYL BROMIDE WITH SODIUM

BOROHYDRIDE

The sodium borohydride used was commercial grade. Diglyme was dis-

tilled from calcium hydride and stored over Dri-Na. The ethyl ether was com-

mercial anhydrous ether. All reaction vessels were dried at 1200 for at least

2 hours before use. All liquids and solutions were transferred by means of a

syringe.

Reduction with sodium borohydride in ethyl ether

Triphenylcyclopropenyl bromide (1.74 g., 5.00 m moles) was suspended in

50 ml. of anhydrous ether. Sodium borohydride (0.19 g., 5.0 m moles) was

added in small portions to the suspension. The reaction mixture, after having

been stirred 24 hours at room temperature (35 ), was poured slowly into 100 ml.

of 5% hydrochloric acid. The solid that separated was taken up in 50 ml. of ether,

and the layers were separated. The aqueous layer was extracted twice with 25 ml.

portions of ether. The extracts were combined with the original organic layer

and dried over magnesium sulfate. After filtration and evaporation of the sol-

vent, a mixture of solid and oil was obtained which was dissolved in 10% benzene-

hexane, and chromatographed over alumina (40 cc in hexane) using the solvent

sequence: hexane, 10% benzene-hexane, 15% benzene-hexane,20% benzene-

hexane, 30% benzene-hexane, 50% benzene-hexane, 50% benzene-ether, ether,

methylene chloride, and methanol.







48
The yield of triphenylcyclopropene (II) was 23.6% (0.316 g.), m.p. 108-1090,

(lit. m.p. 112-1130)1, of benzylidene desoxybenzoin (IV) 0. 035 g., m.p. 96-1000,

(lit. m.p. 100-101o)6, of isobenzylidene desoxybenzoin (V) 0. 087 g., m.p. 85-88,

(lit. m.p. 89-90) 23, and 0.239 g. of a mixture of IV and V, m.p. 81-86o. The

total yield of crystalline ketone was 21% with various ketonic oils (infrared) re-

maining. An alcohol (0. 016 g. ), m.p. 227-2300, was isolated but not identified.

The compound exhibited infrared abosrptions at 2.87 microns.

Reduction with sodium borohydride in anhydrous diglyme

Triphenylcyclopropenyl bromide (I) (1.74 g., 5. 00 m moles) was suspended in

50 ml. of anhydrous diglyme. Sodium borohydride (0.227 g., 6. 00 m moles) was

added in small portions. The yellow slurry immediately became white. The re-

action was stirred at room temperature for 3.25 hours before it was poured into

75 ml. of water. The aqueous solution was extracted 3 times with 50 ml. portions

of ether. The extracts were combined and washed with 50 ml. portions of 5%

aqueous sodium bicarbonate and water before drying over magnesium sulfate. The

organic solution was concentrated to a thick oil by heating on a water bath under

reduced pressure. The oil was dissolved in 10% benzenehexane, and chromato-

graphed as described previously. Triphenylcyclopropene (II) was obtained in 24%

yield (0.315 g.), m.p. 109-111o. The ketonic oils yielded no crystals, but a trace

of solid alcohol (2.87 1) was isolated from the 1:1 benzene-ether fraction.

Reduction with sodium borohydride in 60% aqueous diglyme

Triphenylcyclopropenyl bromide (I) (2.74 g., 5. 00 m moles) was suspended

in 30 ml. of diglyme. Sufficient water (20 ml. ) was added to give the desired con-

centration before 0. 227 g. (6. 00 m moles) of sodium borohydride was introduced








into the solution in small portions. The pH of a 60% aqueous diglyme solution was

5. The reaction was stirred for 3 hours, then worked up and chromatographed as

described for reaction I-A. The yield of triphenylcyclopropene (II) (m.p. 109-111)

was 77% (1. 024 g.). Only trace amounts of benzilidene dexoxybenzoin (IV) were

isolated.

Reduction with sodium borohydride in strongly acidic solution

A suspension of triphenylcyclopropenyl bromide (I) (1. 00 g., 2.72 m moles)

and sodium borohydride (0.11 g., 2.75 m moles) in 25 ml. of ethyl ether was

stirred for 5 minutes. The mixture was then poured fairly rapidly into a stirred

5% hydrochloric acid solution. The aqueous layer (pH 1) was separated and the

organic layer washed with 5% aqueous sodium bicarbonate and water. After drying

(MgSO4), the solution was concentrated to a viscous oil. Chromatography over

20 cc. of Merck alumina, with 5% benzene-hexane, as the elutant, gave 0.400 g.

(55%) of triphenylcyclopropene (II), m.p. 108-1090.

Reduction with sodium borohydride in aqueous base

Triphenylcyclopropenyl bromide (1.00 g., 2.72 m moles) was added in small

portions to a well-stirred 5% aqueous sodium bicarbonate solution (35 ml.) of

sodium borohydride (0.11 g., 2.76 m moles). After approximately half the cation (I)

had been added, ca. 5 ml. of ether was added to take up the oil which had formed

on the top of the water solution. The pH remained between 9 and 10 (measured with

pH paper) throughout the reaction. After all the cation (I) had been added, the

mixture was stirred for 15 minutes before the layers were separated. After wash-

ing the organic layer with 5% hydrochloric acid and water and evaporating the solvent,








the oil was chromatographed over 20 cc. of alumina. The yield of triphenylcyclo-

propene (II) (m.p. 109-1110) was 13% (0.096 g.).

Reduction with sodium borohydride in pyridine

Sodium borohydride (0.19 g., 5.0 m moles) was added in small portions over

a 10 minute period to 1.74 g. (5. 00 m moles) of triphenylcyclopropenyl bromide (I)

dissolved in 50 ml. of anhydrous pyridine. The yellow mixture was stirred 3 hours

at room temperature and then poured into 100 ml. of 5% hydrochloric acid. The

solid that formed was taken up in ether, the ether layer separated, and the aqueous

layer extracted with four 50 ml. portions of ether. The extracts were combined and

washed with 50 ml. portions of 5% hydrochloric acid, 5% aqueous sodium bicarbonate,

and water. The extracts were dried over magnesium sulfate, filtered, and concen-

trated under reduced pressure to a viscous oil (1.45 g.). The oil was chromatograph-

ed as before yielding 70% (.914 g.) of triphenylcyclopropene, m.p. 107-1100, after

one recrystallization. Only trace amounts of benzilidene desoxybenzoin could be

isolated from the ketonic oils, but .015 g. of a red,.crystalline compound (VI) was

isolated, m.p. 147-149. It was identified as 2, 3-diphenylindenone from its melting

point (lit. m.p. 151-152)25 and its infrared spectrum.

In a separate experiment, triphenylcyclopropenyl bromide (0.348 g., 1. 00 m-

moles) was dissolved in 10 ml. of anhydrous pyridine, sodium borohydride added in

small portions, and mixture allowed to stir for 3 hours. An aliquot of this solution

was taken before the borohydride was added and 30 minutes, 1 hour and 3 hours,

after addition. The aliquots were appropriately diluted with acetonitrile to give

2. 0 x 10-5M solutions. The ultraviolet spectra of these solutions were recorded

using an acetonitrile reference solution that contained 2 x 10-2 pyridine. A shift
using an acetonitrile reference solution that contained 2 x 10 % pyridine. A shift







51

of X from 300mm, shown by the sample taken before addition of the reducing

agent, to 313 mnL, shown by the sample taken after the addition of the hydride,

indicated that the reduction was essentially complete in the first 30 minutes.

The ultraviolet spectrum of the reaction products after work up was the same as

that of the reaction mixture. The base line for the spectra of the samples con-

taining sodium borohydride was lowered off scale in the 355-370 mrL region and

the pyridine showed strong absorption in the 220-270 mnL region. The ultraviolet
acetonitrile
spectra of authentic samples of triphenylcyclopropenyl bromide (I) X -287. 5,
max
acetonitrile
300, 316 millimicrons and triphenylcyclopropene (II) X -303, 314, 332 milli-
max
microns were identical to those of the aliquots taken from the reaction.


REDUCTION OF TRIPHENYLCYCLOPROPENYL BROMIDE IN PYRIDINE WITH

LITHIUM ALUMINUM HYDRIDE

The lithium aluminum hydride-pyridine solutions were prepared by the method

of Lansbury and Peterson.1 Solutions were centrifuged, transferred to a second

argon-flushed tube fitted with a rubber septum, and aged under positive argon pres-

sure. All transferring of liquids was accomplished with a syringe. The lithium

aluminum hydride used was not purified. Pyridine, dried and distilled from barium

oxide, was stored over potassium hydroxide pellets until used. All reaction vessels

had been dried at least 2 hours at 1200.

Reduction with fresh lithium aluminum hydride solution

Lithium aluminum hydride (0. 195 g., 5.00 m moles) was added in small

portions to 50 ml. of 0.1 M solution of triphenylcyclopropenyl bromide (I) (5.0 m-

moles) in pyridine. After an initial vigorous reaction, the reaction was stirred






52


for 3 hours at room temperature. The reaction was quenched by adding 50 ml. of

methanol. The methanolic solution was then poured into 100 ml. of 5% hydrochloric

acid, and the aqueous layer extracted with four 50 ml. portions of ether. Extracts

were combined, washed with 50 ml. portions of 5% hydrochloric acid, 5% aqueous

sodium bicarbonate, and water, and dried over magnesium sulfate. After evapora-

tion of solvent, 2. 0 g. of yellow oil were obtained. The oil was chromatographed

over Merck alumina as described previously. The first component off the column

was triphenylcyclopropene (II), identified from its infrared spectrum and melting

point; m.p. 106-1070 after one recrystallization from ether-hexane. The crude

yield of cyclopropene was 7% (. 090 g. ). From the third through fifth fractions,

two crystalline compounds were separated by fractional crystallization. They were

found to be benzilidene desoxybenzoin (IV) and isobenzilidene desoxybenzoin (V) from

their melting points and infrared spectra. Benzilidene desoxybenzoin was isolated

in 16% yield (. 322 g.) m.p. 96-1010 (lit. m.p. 100-101), 6 isobenzilidenedesoxy-

benzoin in 2% yield (. 038 g.), m.p. 80-850 (lit. m.p. 87-88o), 23 and various mix-

tures of the two, m.p. 93-950, in 2% yield (. 034 g.). The remaining ketonic oils

from these fractions showed an infrared spectrum very similar to the oils obtained

from hydrolysis of the cation and appeared to be predominantly a mixture of the

benzilidene desoxybenzoins. The benzene and 1:1 benzene-ether fraction yielded a

fourth compound (VII) (0. 124 g.), m.p. 235-2370 (green melt) after two recrystalli-

zations from methylenechloride-hexane; acetnitrile (log e) 333(45, 000), 316(52, 600),
max
KBr
302(50, 600), 288(84, 000) mp ; X = 5.49, 12.8, 13.21, 14.28, 14.54 microns.
max

Anal. Calcd. for C47H33N; mw. .612, C, 92.15; H, 5.53; N, 2.14. Found: mw

620 (vapor phase osmometer); C, 92.2; H, 5. 54; N, 2. 29.








Reduction with fresh lithium aluminum hydride solution

Triphenylcyclopropenyl bromide (I) (1.74 g., 5. 00 m moles) was dissolved

in 30 ml. of anhydrous pyridine and 20 ml. of a 30 minute aged lithium aluminum

hydride-pyridine solution (0. 25 M, 5. 0 m moles) was added dropwise with a syringe

over a 15 minute period. The reaction was stirred at room temperature for 3

hours and then worked up as in A above. Slow elution chromatography on alumina,

using the already given solvents, yielded 8% (0. 110 g.) of triphenylcyclopropene

in several fractions melting 104-1110. Benzilidene desoxybenzoin, 17% (0. 245 g.),

and 0.138 g. of VII, m.p. 210-2300, were also obtained.

Reduction with aged lithium aluminum hydride solution

A solution of lithium aluminum hydride (20 ml., 0.25 M in LAH) that had been

aged 24 hours was added over a 10 minute period to 1.74 g. (5. 00 m moles) of

triphenylcyclopropenyl bromide dissolved in 30 ml. of pyridine. The reaction was

stirred for 3 hours at room temperature and worked up as in run A. Slow elution

chromatography yielded 2.4% (. 0321 g.) of II, m.p. 104-1080, 0.205 g. of IV,

m.p. 100-1010, .079 g. of a mixture of IV and V (total yield of ketone, 21%), and

10% (0.157 g.) of VII, m.p. 198-2000 (235-2360 after two recrystallizations from

methylene chloride-hexane).

Reduction with concentrated, aged lithium aluminum hydride solution

To 1. 04 g. (3. 00 m moles) of solid cation (I) was added dropwise 29 ml. of

0.25 M (.0075 moles) lithium aluminum hydride-pyridine solution that had been

aged for 3 days. The reaction mixture became red and viscous on addition of the

reducing agent. After stirring for three hours at room temperature, the reaction

was worked up as before. Chromatography over alumina gave; 034 g. (4%) of






54


triphenylcyclopropene, (m.p. 107-1100), 0.168 g. of IV, 091 g. of V (total yield

of ketone 20%) and .076 g. (8%) of VII. A brown oil (.154 g.) remained from which

no more crystalline VII could be obtained but the ultraviolet spectrum indicated

that it was ca. 50% in VII. Trace amounts of 2, 3-diphenylindenone (infrared

spectra 5.89 I), m.p. 145-1490, (lit. m.p. 151-1520)25 were also isolated.

Stability of triphenylcyclopropene in lithium aluminum hydride-pyridine solution

Triphenylcyclopropene (0. 134 g., 0.5 m moles) was dissolved in 5 ml. of

pyridine. Lithium aluminum hydride (0.65 m moles, .025 g.) was added in small

portions to the triphenylcyclopropene solution. The reaction mixture was stirred

3 hours at room temperature and poured into 20 ml. of 5% hydrochloric acid. A

white solid formed which was taken up in 20 ml. of ethermethylene chloride. The

layers were separated, the aqueous layer extracted with 15 ml. of ether and the

extract was combined with the original layer, and the organic layer was washed

with 15 ml. of water. After drying over magnesium sulfate, the solvent was

separated, leaving a slightly yellow solid, which on recrystallization from ether-

hexane gave 0.108 g. (80%) of solid, m.p. 107-1090, along with a yellow residue.

The reaction was repeated using 0. 268 g. (1. 00 m moles) of triphenylcyclo-

propene in 10 ml. of pyridine and .095 g. (2.5 m moles) of lithium aluminum

hydride. The reaction was vigorous, turning from a yellow to a blue-green color.

After stirring 3 hours at room temperature the solution was worked up as above,

yielding 0. 246 g. (92%) of a slightly yellow material, m.p. 178-1850. After one

recrystallization from benzene-hexane, this material melted at 182-1830, turned

green at 2250, and blue at temperatures above 2400, ( lit. m.p. 179. 5-180. 5o).

This material was identified as the triphenylcyclopropene dimer (VIII) by its

melting point and infrared and ultraviolet spectra.








To 0. 124 g. (0.46 m moles) of triphenylcyclopropene was added 7 ml. of an

0. 25M (1.75 m moles) of a lithium aluminium hydride solution which had been

aged for 48 hours. The reaction was stirred for three hours and worked up as

above. A greyish solid was obtained (0.11 g.) m.p. 108-1100 (90% recovery).

Reaction of triphenylcyclopropenyl bromide in pyridine with added amine

To a 0. 093 g/ml. solution of triphenylcyclopropenyl cation in pyridine con-

tained in an ultraviolet spectra cell was added 0.25 ml. of diethyl amine. The

solution showed absorptions, before addition of the amine, at: 324, 306 and 292 mnL.

The optical density of the highest peak, the 324 mL maxima, was ca. 0.73.

Immediately after the addition of the amine, a new peak began to appear at 330 mn.

which reached its maximum optical density, ca. 1.45, after 3 minutes, and the

original cyclopropene peaks vanished. The spectra remained constant for 1 hour,

then ca. 10 drops of water were added. The spectra changed very slowly, and

after 48 hours a new peak had appeared at 315 millimicrons and the original peak

had vanished. This last spectra was very similar, although not identical to the

ultraviolet spectra of benzilidene desoxybenzoin run under similar conditions.

Benzilidene desoxybenzoin shows absorption at 310 and 320 m[.

Reduction of triphenylcyclopropenyl bromide with lithium tritertiarybutoxyalumino-

hydride

Triphenylcyclopropenyl bromide (I) (1.74 g., 5. 00 m moles) was partially

dissolved in 25 ml. of tetrahydrofuran. A stirred suspension (1.99 g., 7.5 m moles)

of lithium tritertiarybutoxyaluminohydride in 25 ml. of tetrahydrofuran was added

dropwise to the cation (I) mixture. All the solid had disappeared by the end of the

addition and a yellow solution resulted. The solution was-stirred for 20 hours at






56

room temperature and worked up as described in the previous reaction. Chroma-

tography yielded 0.766 g. of triphenylcyclopropene (57. 2%, m. p. 108-1110), 0. 013 g.

of 2,3-diphenylindenone, m.p. 148-1500, and 0.103 g. (7%) of benzilidene desoxy-

benzoin, m.p. 94-980.

Preparation of 1, 3-bistriphenylcyclopropenyl benzene (IX)

The 1, 3-bis Grignard was prepared by adding 3. 8 ml. (. 020 moles) of 1, 3-di-

bromobenzene in 40 ml. of anhydrous tetrahydrofuran to 1.08 g. of magnesium

turnings. An iodine crystal was used to initiate the reaction. A greyish suspen-

sion had formed after 5 hours of refluxing under an argon atmosphere. The

resulting suspension was added dropwise to 13. 88 g. (. 040 moles) of triphenyl-

cyclopropenyl bromide (I) partially dissolved in 75 ml. of tetrahydrofuran. After

refluxing for 3 hours, the reaction was hydrolyzed with 6% aqueous ammonium

chloride solution. The layers were separated, the aqueous layer extracted with

50 ml. portions of ether and methylene chloride, the extracts combined, and the

organic solution dried over magnesium sulfate. Chromatography of the residue

after evaporation of the solvents yielded three major compounds. The first, from

the 25% benzene-hexane fraction, was tetraphenylcyclopropene (1.74 g., 25%),

m.p. 175-1770 after three recrystallizations, identified by comparison to an au-

thentic sample. The 50% benzene-hexane fraction yielded the desired product

(5. 5 p in infrared). This compound (4. 53 g., 17%) melted 231-2330, after three

recrystallizations from benzene-hexane. The relatively pure material was percu-

lated through a short alumina column in 25% benzene-hexane and recrystallized

twice from benzene-hexane to give an analytical sample, m.p. 235-2360 (green

melt above 2700). This material gave a typical aryl cyclopropene absorption in the









ultraviolet, X 228, 302, 316 (4173), 332 millimicrons. Anal. Calcd. for
max

C48H34: C, 94.5; H, 5.6. Found: C, 94.41; H, 5.98. The fifth fraction yielded

benzilidene desoxybenzoin, and the remaining fractions gave a solid (. 055 g.)

which melted above 3000. The melting behavior and the infrared spectrum,

identical to an authentic sample, showed that this was hexaphenylbenzene (X). 1,17


PREPARATION OF TETRAARYLCYCLOPROPENES

The ether solvents were analytical grade and were distilled from lithium

aluminum hydride immediately before use. All reaction vessels were dried at

100-1200 for at least two hours before use. All solvents were transferred with

a syringe and all reactions were run under an argon atmosphere.

The concentration of active Grignard was determined by adding 1 ml. of the

Grignard reagent to 10.0 ml. of 0.100 M hydrochloric acid with a syringe and back

titrating the excess acid with 0.100 M aqueous sodium hydroxide to a phenolphthalein

end point.

The magnesium used was finely cut turnings that had been washed with methyl-

ene chloride and heated with an iodine crystal.

The chromatograms were run on a column of 20 g. of Merck alumina packed

in hexane. Eluting solvents were hexane, 5, 10, 20, 40, 60% benzene-hexane ,

benzene, 50% benzene-ether, ethyl ether, and methanol. A 50 ml portion of each

of the solvents was used except in cases where it was observed that a band was

coming off the column and then a portion large enough to completely elute the band

was used.








3-Mesityl-1, 2, 3-triphenylcyclopropene (XV)

To 0.360 g. (15.0 m moles) of magnesium covered with ether was added a

small portion (approximately 15 drops) of 1.96 ml. (13.0 m moles) of mesityl

bromide. The reaction was initiated by heating and the addition of 10 F1. of methyl

iodide. The remaining mesityl bromide in ether (20 ml.) was added over a 20

minute period. After having been refluxed for 4 hours, the yellow solution was

0.552 M in active Grignard. The Grignard solution was checked by adding 4 ml. of

the solution to Dry Ice. After appropriate work up, mesitoic acid was obtained,

m.p. 149-1510, (Lit. m.p. 150-152o).35 A portion (14 ml., 7.75 m moles) of the

mesityl Grignard prepared above was added dropwise over a 10 minute period to

1. 00 g. (2. 90 m moles) of triphenylcyclopropenyl bromide suspended in 15 ml. of

tetrahydrofuran. The reaction was exothermic and a large portion of the cation (I)

went into solution during the course of the addition. The reaction mixture was

stirred under reflux for 46 hours. The resulting suspension was filtered and the

solid washed twice with 35 ml. portions of methylene chloride. The filtrates were

combined and washed with 25 ml. portions of water, 5% hydrochloric acid, 5%

aqueous sodium bicarbonate, and water. Evaporation of the solvent after drying

over magnesium sulfate gave a viscous oil. The oil was placed on a chromato-

graphy column in 10% benzene-hexane and the previously described elution scheme

followed. The excess mesitylene came off in the first fraction. The next three

fractions were combined and recrystallized from chloroform-methanol. This

material (0. 203 g., 18%) showed absorption in the infrared at 1810 cm-1 and

melted 150-151o after three recrystallizations from methanol-chloroform. Although

this material did not show a typical arylcyclopropene spectrum (broad absorption at:








cyclohexane
Scylh 280 (4.14), 300 mF (4. 20)), the other data indicated that it was the
max
desired compound. Anal. calcd. for C30H26: C, 93.22; H, 6.78. Found:

C, 93.02; H, 6.87.

The fifth fraction yielded 0.150 g. of a yellow oil, which after crystallization

from methylenechloride-cyclohexane melted at 159-1650, showed absorption in the

ultraviolet region at 315 and 238 mp. and 1820 cm-1 in the infrared. Fractions

six through nine yielded 0. 324 g. of material, m.p. 95-970, which was identified

as benzylidene desoxybenzoin (IV) by its melting point (lit. m.p. 100-1010) and

infrared spectrum.

The above reaction was repeated on a larger scale using higher concentrations

of reactants and a different work up procedure. A 1. 805M solution of mesityl

Grignard was prepared by refluxing 5.88 ml. (38.4 m moles) of mesityl bromide

with 1. 08 g. (45. 0 m moles) of magnesium in approximately 40 ml. of anhydrous

ether for 4.5 hours. A portion of this solution (12 ml.; 22 m moles) was added

dropwise to a suspension of 3. 00 g. (8.75 m moles) of triphenylcyclopropene

bromide in 10 ml. of ether. The reaction was exothermic and all of the cation (I)

went into solution. The resulting solution was stirred at room temperature for

36 hours. The excess mesityl Grignard reagent was removed by pouring the

reaction mixture onto 50 cc. of powdered Dry Ice. After the viscous mass had

warmed to room temperature, 75 ml. of ether and 50 ml. of water were added

and the layers separated. The organic layer was washed with two 25 ml. portions

of aqueous sodium bicarbonate, 15 ml. of 0.1 Nhydrochloric acid and 25 ml. of

water. After having been dried over magnesium sulfate, the solvent was evapo-

rated to give 4. 0 g. of oily solid. Chromatography of the solid over 25 g. of







alumina using 5% benzene-petroleum ether as the eluent yielded three 50 ml.

fractions containing the desired compound. The first fraction yielded only trace

amounts, the second 1.25 g. of slightly yellow crystals, m.p. 139-1450 (raised

to 151-1520 after one recrystallization from methanol-chloroform), and the third,

two crops of total weight 0.114 g. yellow prisms, m.p. 150-1510, m.p. 142-1470,

respectively. The total yield of the desired cyclopropene was 52%.

3-p-Tolyl-1, 2, 3 -triphenylcyclopropene (XVI)

To 0.36 g. (15 m moles) of magnesium turnings was added 2 ml. of a satu-

rated solution of p-bromotoluene in ether. The reaction began immediately. The

addition of the remaining p-bromotoluene (2. 06 g.; 12. 0 m moles) in 25 ml. of

ether was completed in 20 minutes. Most of the magnesium had disappeared after

the mixture had refluxed for 2. 5 hours. The solution was 0.41 M in active Grignard.

The p-tolyl Grignard (18 ml., 7.4 m moles) was added dropwise to 1.00 g.

(2.90 m moles) of triphenylcyclopropenyl bromide suspended in 15 ml. of tetra-

hydrofuran. After a 50 hour period of reflux with vigorous stirring, the reaction

was worked up by the filtration procedure described above. The first six fractions

off the chromatography column yielded 0.529 g. (51%) of material, m.p. 127-1310

after one recrystallization from methanol-chloroform. A sample prepared for

analysis, m.p. 134.5-1360, showed absorption in the ultraviolet region (milli-
-1
micron, log E) at 230 (4.53), 307 (4.32), 318 (4.30), 336 (4.26) and at 1830 cm-1

in the infrared.

Anal. calcd. for C28 H 22: C, 93.81; H, 6.19. Found: C, 93.58; H, 6.27.

The seventh and eighth fractions yielded 0.132 g. of benzylidene desoxybenzoin,

m.p. 96-1000.









3-o-Anisyl-1, 2, 3 -triphenylcyclopropene (XVII)

o-Bromoanisole (approximately 0.1 ml.) was added to magnesium turnings

(0.34 g.; 18. 0 m moles) that were covered with 3 ml. of ether. The reaction was

initiated by heating with an infrared lamp. The remaining 22 ml. of ether and

o-bromoanisole (total of 1. 87 ml.; 15. 0 m moles) was added over a 15 minute

period. A 0.725 M solution of active Grignard reagent was obtained after 1 hour

of refluxing and 2 hours of stirring at room temperature. The solution was red

black. Triphenylcyclopropenyl bromide (1. 00 g.; 2.9 m moles) was suspended

in 15 ml. of tetrahydrofuran and 16. 5 ml. of 0.725 M Grignard solution (12. 0 m

moles) was added dropwise over a 5 minute period. During the exothermic reac-

tion the cation went into solution and a white solid appeared shortly after the con-

clusion of the addition. The reaction was stirred 20. 5 hours at room temperature.

Dry carbon dioxide was bubbled into the reaction for 1 hour before the entire reac-

tion mixture was poured onto 20 cc. of crushed Dry Ice. After the resulting

viscous mass had warmed to room temperature, 40 ml. of 5% hydrochloric acid

and 25 ml. of methylene chloride were added. The organic layer was separated

and extracted with 50 ml. and 25 ml. portions of 5% aqueous sodium bicarbonate.

After a final wash with water, the solution was dried over magnesium sulfate and

filtered. Chromatography of the oily solid remaining after evaporation of the solvents
-1
give 0.424 g. (39%) of material which showed absorption at 1810 cm in the infrared

m.p. 197-197.50. This compound showed an ultraviolet spectrum that was typical

acetonitrile
of the diphenylcyclopropene chromophore, max (log E)n 286 (4.26), 303 (4.31),

317 (4.34), 333 mni (4.29).

Anal. calcd. for C28H 220: C, 89.80; H, 5.92. Found: C, 89.58; H, 5.81.
--- 28 22









3-(p-Chlorophenyl)-l, 2, 3-triphenylcyclopropene (XVIII)

A few crystals of 4-chlorobromobenzene were added to 0.132 g. (13. 0 m

moles) of magnesium turnings that were covered with ether. After gently warm-

ing the flask to initiate the reaction, the remainder of the 4-chlorobromobenzene

(total of 2. 29 g.; 12. 0 m moles) dissolved in 20 ml. of ether was added over a

20 minute period. A solution that was 0. 59 M in active Grignard was obtained

after 2 hours of reflux. A portion (6 ml.; 3. 54 m moles) of this solution was

added slowly to 1. 00 g. (2.90 m moles) of triphenylcyclopropenyl bromide

suspended in 15 ml. of ether. The refraction was refluxed six hours and worked

up by the filtration method. The first fraction from the column gave a colorless

oil (0.203 g., 19%) which solidified on standing, m.p. 135-1400. An analytical

sample of this material, m.p. 154. 5-155. 50, showed absorption at kcyclohexane
max
230 (4. 56), 304 (4.33), 317 (4.35), 334 mrn (4.31) in the ultraviolet and at 1825 cm1

in the infrared.

Anal. calcd. for C27H Cl: C, 85.06; H, 5.05. Found: C, 85.07, H, 5.09.

The next three fractions yielded 0.176 g. of benzylidene desoxybenzoin, m.p.

97-1000. The sixth and seventh fractions yielded 0.215 g. of yellow oil from which

traces of crystals, m.p. 165-1950 were isolated.

Tetraphenylcyclopropene (XIX)

A 0.76 M solution of phenyllithium was prepared by stirring 3.14 g. of bromo-

benzene in 15 ml. of ether with 0. 30 g. of lithium wire. A portion of this solution

(13 ml.; 10. 0 m moles) was added dropwise with a syringe to 1.74 g. (5. 00 m moles)

of triphenylcyclopropenyl bromide suspended in 10 ml. of ether. The resulting

black solution was stirred for 7. 5 hours at room temperature before 15 ml. of water









was added with cooling. After separation of the layers, the aqueous layer was

extracted with 20 ml. of ether. The organic layers were combined and washed

with 50 ml. of 0. 1 N hydrochloric acid, two 25 ml. portions of aqueous sodium

bicarbonate, and 25 ml. of water. The yellow solid, obtained after evaporation

of the dried solution to dryness, was recrystallized from ether-hexane. Three

crops of material were obtained (0.798 g.; 47%), m.p., after two recrystalliza-
13
tions, 177-179, (lit. m.p. 177-1780)8 cyclohexane 280 (4.,6), 305 (4.33), 318
max
(4.35), 335 m L (4.31). This compound, identified as tetraphenylcyclopropene,
-1
shows the typical disubstituted cyclopropene peak at 1825 cm in the infrared.

3-(p-Anisyl)-l, 2, 3-triphenylcyclopropene (XXI)34

To 0.52 g. (22.0 m moles) of magnesium and an iodine crystal was added

dropwise a solution of 2.7 ml. (21.0 m moles) of p-bromoanisole in 50 ml. of ether

and the mixture refluxed for 2 hours under argon. The resulting Grignard reagent

was added to a suspension of 4. 0 g. (9.35 m moles) of triphenylcyclopropenyl cation

(I) in 75 ml. of ether. After a 2 hour reflux period the reaction mixture was worked

up as in the previous reactions. Chromatography over ca. 40 g. of alumina yielded

two fractions which contained the desired cyclopropene (5. 5 microns in infrared).

The total yield of cyclopropene was 1.22 g. (40%), m.p. 160-1610. The product

(XXI) showed major absorption in the infrared (KBr) at 3.3, 3.4, 3.5, 5.5, 6.28,

6.6, 6.7, 6.8, 6.9, 7.8, 8.1, 8.55, 9.8, 12.05, 12.8, 13.2, 14.2, 14.7 microns

and in the ultraviolet (cyclohexane) at 230 (4. 56), 300 (sh), 312 (4.36), 334 (420)

millimicrons.

Anal. Calcd. for C2H220: C, 89.8; H, 5.92. Found: C, 89.44; H, 6.02.
28 22








3-p-Anisyl-1, 2, 3-triphenylcyclopropene (XXI)

A. Triphenylcyclopropenyl bromide (0.300 g.; 0. 90 m moles) was dissolved

in a solution of 2 ml. of glacial acetic acid (2% acetic anhydride added), 1 ml. of

anisole and 3 drops of pyridine. The mixture was refluxed for 22 hours. After

cooling the mixture 10 ml. of water and 5 ml. of methylene chloride were added.

The layers were separated and the aqueous layer was extracted with two 10 ml.

portions of methylene chloride. After washing the organic layer three times with

10 ml. portions of 5% hydrochloric and 5% aqueous sodium bicarbonate, once with

water and drying over magnesium sulfate, the solvents were evaporated. The oily

residue was chromatographed over 15 g. of Merck alumina. The first 4 fractions

yielded a green, fluorescent oil which showed infrared absorption very similar to

that of the higher melting material from the thermal rearrangement. Crystals were

obtained which a broad range between 65 and 1400 depending on the number of

recrystallizations.

B. Triphenylcyclopropenyl bromide (1. 00 g., 2.90 m moles) was refluxed

with 2. 5 ml. of anisole, 6 ml. of acetic acid and 0. 6 ml. of pyridine for 20 hours.

The reaction was worked up as in A. The first two fractions from the chromato-

graphy column gave 0.103 g. (10%) of crystals, m.p. 157-1580, after trituration

with ether. The melting point and infrared spectrum were identical to those of

an authentic sample of 3-p-anisyl-1, 2, 3-triphenylcyclopropene. 3The later

fractions yielded approximately 0. 035 g. of 2, 3-diphenylindenone, m. p. 151-1520.

(lit. m.p. 151-152)25

C. The above reaction was repeated using 1.00 g. (2.90 m moles) of cation (I),

3.3 ml. of 10:1 acetic acid-pyridine, and 4.0 ml. of anisole. The crude yield of the









desired cyclopropene (5. 5 micron peak in infrared) was 0. 231 g. (21%), obtained

in two fractions, m.p. 143-149, 157-158. Diphenylindenone, m.p. 150-1520,

(0. 021 g.) was also obtained.

D. The best yield of the desired cyclopropene was obtained when the volume

of anisole was increased to 10 ml. and the volume of 10:1 acetic acid-pyridene was

decreased to 3 ml. Using these reaction conditions 0.520 g. (47%) of crude 3-p-

anisyl-1, 2, 3-triphenylcyclopropene was obtained from 1. 00 g. of cation (I). The

excess anisole was removed by heating the reaction mixture on a steam bath under

a stream of air.

3-(2, 4-Dimethoxyphenyl)-1, 2, 3-triphenylcyclopropene (XXII)

Triphenylcyclopropenyl bromide (1. 00 g.; 2. 90 m moles) was heated at 85-900

for 24 hours with 10 ml. of m-dimethoxybenzene and 3 ml. of 10:1 acetic acid

pyridine. After aqueous work up and evaporation of the low boiling solvents on a

rotary evaporator, the excess dimethoxybenzene was removed by heating the

reaction mixture on a steam bath under a stream of air. Chromatography over

25 g. of Merck alumina yielded 0.243 g. (23%) of white crystals, m.p. 160-1650,

which showed absorption at 5. 5 microns in the infrared. The melting point was

raised to 180-1820 after 4 recrystallizations. The ultraviolet spectrum was

consistent with the cyclopropene structure showing absorption at: Xchloroform
max
331 (shoulder), 305, and 290. 5 m..


THERMAL REARRANGEMENTS

Tetraphenylcyclopropene

Tetraphenylcyclopropene (0. 200 g.) was sealed in a tube which had been






66

evacuated to 0. 5 mm. pressure. After heating at 235-2400 for 2 hours in a Wood's

metal bath and cooling to room temperature, a clear glass was obtained. The glass

was crystallized from methanol-chloroform to give 0.180 g. (90%) of 1, 2, 3-tri-

phenylindene (XX), m.p. 129-1310 (lit. m.p. 131-132). 39 An analytical sample,

m.p. 131-1320, showed ultraviolet absorption (cyclohexane) at 306 (4.18), 241 mi

(4.38) and infrared absorption (KBr) at 1595, 1495, 1455, 1440, 911, 790, 773,
-1
760 (w), 751, 739, 719 (w), and 695 (s) cm- A fluroescence spectra of the com-

pound showed X (excitation) 355 and X (emission) 425 millimicrons.
max max
3-p-Anisyl-1, 2, 3-triphenylcyclopropene

3-p-Anisyl-1, 2, 3-triphenylcyclopropene (0. 350 g.) was heated at 2000 for 4

hours in a sublimation apparatus under an argon atmosphere. A greenish, yellow

solid collected on the condenser. The contents of the reaction vessel were dissolved

in petroleumether and chromatographed on 30 g. of Woelm neutral alumina, activity

grade 1. The column was developed with 10% benzene-petroleum ether. Two bands

that were fluorescent under ultraviolet light were formed. The solvent sequence

used was 100 ml. mixtures of petroleum ether, benzene, ethyl ether, chloroform,

and methanol. The polarity of the solvents was gradually increased at such a rate

as to keep the band moving at a reasonable rate. Fractions of 40 ml. were taken.

The fractions were combined according to the fluctuation of the weight of material

per fraction. There was a total of 46 fractions taken. The first two yielded 0. 042 g.

of white solid, m.p. 138-1450 after one recrystallization from ether-cyclohexane.

This material showed infrared absorption (KBr) at 3.25, 3.4, 3.5, 6.21, 6.58,

6.71, 6.8, 6.9, 7.8, 8.08, 8.55, 9.68, 11.88, 12.35, 12.85, 13.08, 13.28, 13.58,

14.3, and 14. 5 microns and ultraviolet absorption (max ) in chloroform at 315 and

244 millimicrons.








The next nine fractions yielded crystals which were various mixtures (infra-

red) of the first components and the later fractions. Fractions 12-14 yielded

0.085 g. of crystals, m.p. 145-1480 after two recrystallizations from ether-

cyclohexane. These crystals showed infrared absorption at (KBr) 3.3, 3.4, 3.5,

6.25, 6.9, 7.41, 7.8, 7.9, 8.1 (w), 8.2, 9.7, 11.8, 12.06, 12.3, 12.75, 13.1 (w),

13.3, 13.4, 13.6, 14.3, and 14.5 microns and ultraviolet absorption in chloroform

at 315 and 240 millimicrons.

The remaining fractions gave only traces of yellow or brown oils

3-Mesityl-l, 2, 3-triphenylcyclopropene

3-Mesityl-l, 2, 3-triphenylcyclopropene (0. 050 g.) was heated at 190-2000 for

two hours in a small sublimation apparatus which had been flushed with argon. On

cooling a fluorescent oil formed on the condenser and sides of the flask. The oil

from the condenser (0. 100 g., 10%) was triturated with methanol-chloroform to

yield white crystals (A), m.p. 179-1810. The material from the bottom part of the

apparatus was crystallized from methanol-chloroform to give three crops of

crystals. The first crop (0. 010 g., 20%) was the same compound as that on the

condenser, m.p. 175-1780. The next two crops yielded 0. 014 g. (28%) of crystals, (B)

131-132.

The higher melting material (A) showed absorption (max ) in the ultraviolet
max

in chloroform at 330 shoulder) 315 shoulder) 307, and 240 millimicrons and in the

infrared (KBr) at 3.3, 3.4, 3.5, 6.28, 6.7, 7.3, 9.35, 9.75, 11.0, 11.62, 11.9 (w),

12.02 (w), 13.1, 13.55, 14.1 (w), 14.4, and 14. 55 microns. The lower melting

material (B) showed absorption (X ) in the ultraviolet in chloroform at 297. 5
and 240 millimicrons and in the infrared (KBr) at 3.3, 3.4, 3.5, 6.29, 6.7, 6.91,
and 240 millimicrons and in the infrared (KBr) at 3.3, 3.4, 3.5, 6.29, 6.7, 6.91,







68

9.3, 9.75, 11.6, 11.7, 11.9, 12.5, 13.0, 13.2, 13.61, 14.1, and 14.35 microns.

The nuclear magnetic spectra of the two compounds were considerably dif-

ferent. The higher melting compound (A) showed the following resonance signals

in the n. m. r. (CDC13): complex multiple at 2. 84 tau, sharp signlet at 4.73 tau,

a doublet at 7.65 tau, and a sharp singlet at 8.12 tau. The area ratio of the last

three peaks was 1:6:3. The lower melting material (B) showed the following

resonance signals (CDC13): a complex multiple at 2. 80 tau, and sharp singlets
3
at 7.74, 8.10, and 8.30 tau. The last three peaks were of equal area.

1, 2-Diphenylindene (XXIX)

1, 2-Diphenylindene (0. 025 g.) was sealed in an 8 mm. tube. The tube had

been evacuated to approximately 0. 5 mm. before sealing. After having been

heated for 2 hours at 190-1950 in a Wood's metal bath with the top of the tube

protruding above the liquid level, the major portion of the material had sublimed

into the upper portion of the tube. The tube was opened and the white needles were

scraped out, m.p. 175-1780.

The above experiment was repeated with the tube completed submerged. The

starting material was recovered essentially unchanged, m.p. 175-1770

When the temperature was raised to 250 100, in a similar experiment, a

new product was obtained, m.p. 103-1060: Xchloroform 315 (shoulder), 307, 240.
max

This new solid was identified as 2, 3-diphenylindene (XXX) by its melting point

(lit. m.p. 107-108)39 and ultraviolet spectrum.

Preparation of a mesityldiphenylindene

2, 3-Diphenylindenone (VI) (0. 025 g.) was dissolved in 3 ml. of ether and 3 ml.

of 0. 552 M mesitylmagnesium bromide. The resulting red mixture was stirred









until the color had vanished (over night). After aqueous work up, a green oil was

obtained which did not show absorption in the carbonyl region (VI showed absorp-

tion at 5. 87 microns) but showed a new peak near 2. 8 microns (alcohol). The

resulting oil was heated with 5% hydriodic acid in acetic acid. After work up and

chromatography on approximately 5 g. of alumina, using petroleum ether as the

eluent, a fluorescent oil was obtained which showed ultraviolet absorption at 330

(shoulder), 315 (shoulder), 307, and complex absorption after 290 millimicrons.

This oil showed the same retention time on a thin layer chromatogram (5:1 hexane-

benzene as eluent, silica gel HF254 adsorbent) as A found from the thermal re-
254
arrangement of 3-mesityl-l, 2, 3-triphenylcyclopropene.

Preparation of 1-methyl-i, 2, 3-triphenylindene (XXV)

n-Butyllithium (0.6 ml. of 1. 5 M solution in hexane) was added dropwise to

0.180 g. (5.0 m moles) of 1, 2, 3-triphenylindene dissolved in 8 ml. of ethyl ether.

The resulting solution was refluxed for 3 hours before 0. 5 ml. of methyl iodide

(10 m moles) was added in small portions. After addition, the cloudy solution was

refluxed for 12 hours. Aqueous work up gave a yellow oil which was chromato-

graphed on 25 g. of Merck alumina. Three 30 ml. fractions which yielded color-

less oils were collected after elution with petroleum ether. The first two fractions

showed absorption in the ultraviolet at 303. 5 and 240 millimicrons. The third

fraction contained unreacted 1, 2, 3-triphenylindene as identified by its ultraviolet

absorption at 307 and 240 millimicrons. Crystallization of the first two fractions

was attempted from methanol-chloroform and from petroleum ether. Finally an

oily solid was obtained from acetic acid which gave a white powder (0. 045 g.)

after grinding in a morter and pestle, m.p. 100-1050. Although the melting point






70

is high (lit. m.p. 96-980), 36 resonance signals at 2.8 tau (complex multiple) and

8. 29 tau showed that the compound was 1-methyl-i, 2, 3-triphenylindene.

Stability of mesityl indene (A) to basic conditions

To a saturated solution of ethanolic potassium hydroxide was added 0. 050 g.

of (A). The solution had been purged of oxygen by bubbling a stream of argon

through the ethanol before the reactants were added. The reaction mixture was

refluxed for 3. 5 hours under an argon atmosphere before it was neutralized with

20% hydrochloric acid. The organic layer, after normal work up, yielded a

yellow oil which crystallized on trituration with methanol-chloroform, m.p. 178-

1800. The ultraviolet spectra and melting point showed it to be starting material.

The same results were obtained when n-butyllithium was used as the base.

When the above reaction was repeated with ethanolic potassium hydroxide

solution which had not been deoxygenated, a new solid was obtained, m.p. 183-

1860. This material showed ultraviolet absorption at 345 (shoulder), 329, 316,

309 (shoulder), and 240 millimicrons, and infrared absorption at 2. 8 microns.

Stability of mesityl indene (B) to basic conditions

A solution of 0. 010 g. in 10 ml. saturated ethanolic potassium hydroxide was

heated at 85-900 for 23 hours. The solution had not been deoxygenated and was

protected from the atmosphere only by a drying tube of calcium chloride. The

ultraviolet spectrum of the product after aqueous work up was identical to that of

the starting material.

Preliminary rate determination

Standard solutions of 1, 2, 3-triphenylindene were prepared. It was found that

a plot of In %T versus concentration was linear from concentration of 1. 00 to






71

-2
1. 00 x 10 mg. /ml. The %T values were determined on an American Instrument

Company, Fluoro-Microphotometer. Interference filters with maximum trans-

mittance at 440 and 365 millimicrons were used as the emission and excitation

filters. A 1 mg. /ml. solution of tetraphenylcyclopropene showed only slight

absorbance and no fluorescence with this set of filters. Solutions containing

1 mg. /ml. of total material were prepared that were 0, 5, 10, 20, 30, 40, 50,

70, and 100% 1, 2, 3-triphenylindene with enough tetraphenylcyclopropene added to

give the constant 1 mg. /ml. concentration. A plot of In %T versus percent tri-

phenylindene gave a smooth curve. Sample tubes (8 mm. x 15-20 cm.) were

prepared by two methods. All the tubes were first boiled in detergent and then

rinsed 10 times with distilled water. The first set was dried after rinsing. The

second set was soaked in concentrated ammonium hydroxide for at least three

hours and then rinsed with 5% aqueous ammonium hydroxide. Samples of tetra-

phenylcyclopropene (0. 0100 g.) were weighed into the first set of tubes and half of

the tubes prepared by the second method. The remaining tubes in the second set

were filled with 1. 000 ml. of a 10. 00 mg. /ml. solution of tetraphenylcyclopropene

in decahydronaphthalene. The tubes were sealed after they had been evacuated

and/or degassed to 0.25 mm. pressure. The tubes were introduced into a constant

temperature bath (192.4 + 0. 20) and allowed to remain for a prescribed period of

time. The temperature bath was a chamber filled with equilibrated decahydro-

naphthalene vapors at atmospheric pressure. After the prescribed time period

(time was kept to within 2 sec.), the tubes were opened and the contents transferred

quantitatively to 10 ml. volumetric flasks with 10 ml. of dichloroethane

percent conversion of tetraphenylcyclopropene to triphenylindene was then deter-






72

mined by comparing the %T of these solutions to those of known percent concentra-

tion. Plots of In (percent conversion) versus time were linear but the reaction

rate at this temperature was too rapid to give accurate rate constants although the

relative rates give a qualitative estimate of the effect of the various methods of

sample preparation of the rate of the reaction. Within 5 minutes, 97 percent

conversion was observed in the tubes washed with distilled water only, 60-70

percent conversion in the tubes that had been washed with base, and 20-30 percent

conversion in those tubes which had been washed with base and in which the sample

was in solution.

ACID CATALYZED REARRANGEMENTS

Tetraphenylcyclopropene (XIX)

Tetraphenylcyclopropene (0. 010 g.) was refluxed in a mixture of 2 ml. of

acetic acid, 0.4 ml. of benzene and 3 drops of concentrated hydrochloric acid for

24 hours. The reaction mixture was extracted with ether and the ether later was

washed with 5% aqueous sodium bicarbonate until there was no apparent acid-base

reaction in the separatory funnel. The ether solution was dried over magnesium

sulfate and filtered. After evaporation of the solvent, an oily residue (0. 012 g.)

was obtained which crystallized on trituration with cyclohexane-ether, m.p. 128-

1300. The infrared and ultraviolet spectra of this compound was identical with

that of 1, 2, 3-triphenylindene obtained from the thermal rearrangement of tetra-

phenylcyclopropene.

3 -(p-Anisyl)-l, 2, 3-triphenylcyclopropene (XXI)

3-(p-Anisyl)-l, 2, 3-triphenylcyclopropene (0. 050 g.) was refluxed for 24 hours

in a mixture of 3 ml. of benzene, 2 ml. of acetic acid, and 2 drops of 48% aqueous









hydrobromic acid. After working up as described in preceding experiments,

0. 055 g. of brown oil was obtained. The infrared spectrum of the product,

especially in the ether region (7.5-8.5 microns), and the melting point, m.p. 145-

1490 after two recrystallizations from ether-cyclohexane, were essentially identi-

cal to those of the higher melting material obtained from the thermal rearrange-

ment.

3-Mesityl-1, 2, 3 -triphenylcyclopropene (XV)

A solution of 0. 025 g. of 3-mesityl-l, 2, 3-triphenylcyclopropene in 4 ml. of

acetic acid, 1 ml. of benzene, and 0.25 ml. of concentrated hydrochloric acid was

refluxed for 17 hours. After aqueous work up, 0. 016 g. of crystalline material,

m.p. 179-1800, was obtained after trituration with ether-cyclohexane. The infra-

red and ultraviolet spectra of this material was identical to that of the higher

melting material (A) obtained from the thermal rearrangement.

Triphenylcyclopropene (II)

To a mixture of 30 ml. of acetic acid, 6 ml. of benzene, and 1. 8 ml. of

hydrochloric acid was added 0. 385 g. of triphenylcyclopropene. The mixture was

stirred at 90-950 for 20 hours. A white solid (0. 170 g.) crystallized from the

solution when it was cooled to room temperature, m.p. 178-1800. A second crop

of crystals, m.p. 176-1780, (0.190 g.) was obtained by concentrating the filtrate

under reduced pressure. This material showed absorption in the ultraviolet at

335 (shoulder), 320, 308, and 240 millimicrons. This compound was identified as

39
1, 2-diphenylindene by its melting point (lit. m.p. 177-1780), infrared and

ultraviolet spectra, and mechanistic considerations.









26. P. T. Lansbury and R. E. Macleay, J. Am. Chem. Soc., 87, 831 (1965).

27. C. F. Koelsch, J. Am. Chem. Soc., 56, 1337 (1934).

28. B. Fohlish and P. Burgle, Angew. Chem. Internat. Edit., 3, 699 (1964).

29. R. Breslow and M. A. Battiste, J. Am. Chem. Soc., 82, 3626 (1960).

30. E. S. Gould, Mechanism and Structure in Organic Chemistry, Holt, Rinehart
and Winston, Inc., New York, 1959, p. 207.
31. R. Breslow, B. Winter, and M. Battiste, J. Org. Chem., 24, 415 (1959).

32. R. Breslow and P. Wolf, unpublished work, cited in ref. 44.

33. A. Small and R. Breslow, unpublished work, cited in ref. 44.

34. M. A. Battiste and J. Todd, unpublished work.

35. Org. Syn., Coll. Vol. III, p. 555.
36. C. F. Koelsch and R. V. White, J. Org. Chem., 6. 602 (1941).

37. R. Breslow and M. A. Battiste, unpublished work, cited in ref. 44.

38. N. P. Neureiter, J. Org. Chem., 24, 2044 (1959).

39. Ref. 25, p. 391.

40. M. S. Kharasch and Otto Reinmuth, Grignard Reactions of Non-metallic
Substances, Prentice Hall, Inc., New York, 1954, p. 1051.
41. J. R. Dyer, Applications of Absorption Spectroscopy of Organic Compounds,
Prentice Hall, Inc., Englewood Cliffs, N. J., 1965, p. 20.
42. H. Gilman, N. J. Beaber and J. L. Jones, Rec. tray. chim., 48, 597 (1928).

43. G. Bergson, Acta. Chem. Scand., 17, 2691 (1963).
44. R. Breslow, Molecular Rearrangements, Part I, P. De Mayo, Ed., Interscience
Publishers, Inc., New York, N. Y., 1963, pp. 236-261.









27. C. F. Koelch, J. Am. Chem. Soc., 56, 1337, (1934).

28. B. Fohlish and P. Burgle, Angew, Chem. Internat. Edit., 3, 699, (1964).

29. R. Breslow and M. A. Battiste. J. Am. Chem. Soc., 82, 3626, (1960).

30. E. S. Gould, Mechanism and Structure in Organic Chemistry, Holt, Rinehart
and Winston, Inc., New York, 1959, p. 207

31. R. Breslow, B. Winter, and M. Battiste, J. Org. Chem., 24, 415, (1959).

32. R. Breslow and P. Wolf, unpublished work, cited in ref. 44.

33. A. Small and R. Breslow, unpublished work, cited in ref. 44.

34. M. A. Battiste and J. Todd, unpublished work.

35. Org. Syn., Coll. Vol. III, p. 555.

36. C. F. Koelch and R. V. White, J. Org. Chem., 6, 602, (1941).

37. R. Breslow and M. A. Battiste, unpublished work, cited in ref. 44.

38. N. P. Neureiter, J. Org. Chem., 24, 2044, (1959).

39. Elsevier Encyclopedia of Organic Chemistry, Vol. 12 A., series III, Elsevier
Publishing Company, Inc., New York-Amsterdam, 1948, (a.) p. 237, (b.) p. 391.

40. M. S. Kharasch and Otto Reinmuth, Grignard Reagents, Prentice Hall, Inc.,
New York, 1954, p. 1051.

41. J. R. Dyer, Applications of Absorption Spectroscopy of Organic Compounds,
Printice Hall, Inc., Englewood Cliffs, N. J., 1965, p. 20.

42. H. Gilman, N. J. Beaber and J. L. Jones, Rec. tray. chim., 48, 597, (1929).

43. G. Bergson, Acta. Chem. Scand. 17, 2691, (1963).

44. R. Breslow, Molecular Rearrangements, Part I, Paul De Mayo, Editor
Interscience Publishers, Inc., New York, N.Y., 1963, pp 236 261.















BIOGRAPHICAL SKETCH


Robert Howard Grubbs was born on February 27, 1942, in Calvert City,
Kentucky. He attended the public schools in Paducah, Kentucky, graduating
from Paducah Tilghman High School in June of 1960. He entered the University
of Florida in September of 1960, and obtained a Bachelor of Science degree in
December of 1963. Since entering graduate school at the University of Florida
in January of 1963, he has held teaching and research assitantships while working
toward a Master of Science degree. He plans to continue his education at
Columbia University.













This thesis was prepared under the direction 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 Master of Science.

August 14, 1965






Dean, College of Arts and Sciences




Dean, Graduate School


Supervisory Committee:




Chairman




// v~n 7/<=2e







Internet Distribution Consent Agreement


In reference to the following dissertation or thesis:


AUTHOR:
Grubbs, Robert
TITLE:
Studies in the arylcyclopropene series (record number: 554225)
PUBLICATION DATE:
1965


I, ,b4 /, /<'l_, ( Ld as copyright holder for the aforementioned
dissertation or thesis, hereby grant specific and limited archive and distribution rights to the Board of
Trustees of the University of Florida and its agents. I authorize the University of Florida to digitize and
distribute the dissertation or thesis described above for nonprofit, educational purposes via the Internet
or successive technologies.

This is a non-exclusive grant of permissions for specific off-line and on-line uses for an indefinite
term. Off-line uses shall be limited to those specifically allowed by "Fair Use" as prescribed by the
terms of United States copyright legislation (cf, Title 17, U.S. Code) as well as to the maintenance and
preservation of a digital archive copy. Digitization allows the University of Florida to generate image-
and text-based versions as appropriate and to provide and enhance access using search software.

This grant op e io ibits use of the digitized versions for commercial use or profit.


Signature of Copyright Holder

(&1rl- // Gl Q J6
Printed or Typed Name of Copyright Holder/Licensee

Personal information blurred


7-


Date of Signature

Please print, sign and return to:
Cathleen Martyniak
UF Dissertation Project
Preservation Department
University of Florida Libraries
P.O. Box 117007
Gainesville, FL 32611-7007




Full Text
TABLE 5
REACTION OF TRIPHENYLCYCLOPROPENYL BROMIDE
WITH ANISOLE
Reaction
Equiv. of
pyridine
Equiv. of
Anisle
Equiv. of
Cation
Products
Cation
+ anisle
0.81
8.4
0.87
A light green fluorescent oil
Cation
+ anisle
5.3
21.0
2.7
10% XXI
Cation
+ anisle
3.7
33.6
2. 7
21% XXI
Cation
+ anisle
2.8
84.0
2. 7
47% XXI


TABLE 4--Continued
Ar
mp.
Yield
(cm."^-)
U.V. Spectra (cyclohexane) max (log )
£-Tolyl (XVI)
134.5-1360
51%
1830
230(4.53), 307(4.32), 318(4.30),
336(4.26)
£-Chlorophenyl
(XVIII)
154-155
19%
1825
230(4.56), 304(4.33), 317(4.35),
334(4.31)
aPhenyl lithium used.
^Prepared from the bis-grignard of meta-dibromo benzene.
cAcetonitrile solvent.
to
<1


STUDIES IN THE ARYLCYCLOPROPENE
SERIES
By
ROBERT HOWARD GRUBBS
A THESIS PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF MASTER OF SCIENCE
UNIVERSITY OF FLORIDA
August, 1965

ACKNOWLEDGMENTS
The author extends his sincere appreciation to his supervisory committee,
to the faculty of the Chemistry Department of the University of Florida, and to
his colleagues for their interest and friendly cooperation during the course of
this investigation. He takes particular pleasure in expressing gratitude to his
research director, Dr. M. A. Battiste, for his constant sincerity and willing
ness in providing the counsel, guidance, and encouragement necessary for the
realization of this project; and to Dr. J. F. Helling for his aid in the preparation
of this thesis.
The author gratefully acknowledges the financial support of a portion of this
research by the National Science Foundation.

TABLE OF CONTENTS
Page
ACKNOWLEDGEMENTS ii
LIST OF TABLES iv
LIST OF FIGURES v
CHAPTER
I. REDUCTION OF TRIPHENYLCYCLOPROPENYL BROMIDE WITH
COMPLEX METAL HYDRIDES
Introduction 1
Discussion and Results 6
II. THE PREPARATION AND REARRANGEMENT OF TETRAARYLCYCLO-
PROPENES
Introduction 19
Discussion and Results 24
m. EXPERIMENTAL
Reduction of Triphenylcyclopropenyl Bromide with Sodium
Borohydride 47
Reduction of Triphenylcyclopropenyl Bromide in Pyridine
with Lithium Aluminum Hydride 51
Preparation of Tetraarylcyclopropenes 57
Thermal Rearrangements 65
Acid Catalyzed Rearrangements 72
BIBLIOGRAPHY 74
BIOGRAPHICAL SKETCH 76
in

LIST OF TABLES
Table Page
1. Reduction with Sodium Borohydride 7
2. Lithium Aluminum Hydride in Pyridine 11
3. Reaction of Triphenylcyclopropenyl Bromide with Activated
Aromatic Nuclei 21
4. Grignard Preparation 26-27
5. Reaction of Triphenylcyclopropenyl Bromide with Anisle. . 29
xv

LIST OF FIGURES
Figure Page
1. Ultraviolet Spectra of Tetraphenylcyclopropene, (XV), (XIX), and
(XXI) 31
2. Ultraviolet spectrum of 3-(p-tolyl)-l, 2,3-triphenylcyclopropene. 32
3. Ultraviolet spectrum of 3-(o-anisyl)-l, 2, 3-triphenylcyclopropene. 33
4. Ultraviolet spectrum of 3-(p-chlorophenyl)-l, 2,3-triphenylcyclo
propene 34
5. Mechanistic Scheme for Thermal Rearrangement of Tetraphenylcyclo
propene 39
v

CHAPTER I
REDUCTION OF TRIPHENYLCYCLOPROPENYL
BROMIDE WITH COMPLEX METAL HYDRIDES
INTRODUCTION
The recent interest in non-benzenoid aromatic compounds has led to the prepa-
3 6
ration of various stable carbocyclic cations in the cyclopropene series. The
need for a convenient route to the parent hydrocarbons from these cations is
obvious. Not only are the parent hydrocarbons needed for comparison purposes
but they are inherently interesting themselves. The recent report of Breslow on
the preparation of triphenylcyclopropene from triphenylcyclopropenyl bromide by
1 7
lithium aluminum hydride reduction and the work by Brown and Bell on the
sodium borohydride reduction of carbonium ions formed under solvolytic condi
tions indicate that the reduction of a stable carbonium ion with complex metal
hydrides could provide useful synthetic information as well as further insight into
the chemistry of these reducing agents. The reduction of stable cations employing
complex metal hydrides allows observation of the products arising directly from
hydride attack without the complications inherent in reductions involving carbonyl
groups.
The triphenylcyclopropenyl cation (I) was chosen for a study of this nature
since many of its reaction products have been identified, and the majority of these
1 6
are solids. Triphenylcyclopropene (II) has also been found to be an excellent
dienophile, and has found uses in this role as an intermediate in the preparation
9 8
of heptaphenyltropylidene and various bicyclic systems. This added utility
demands a convenient high yield route to this species.
1

2
Sodium borohydride would, at the outset, appear to be a poor reducing agent
for triphenylcyclopropenyl cation since it has been found that the cation (I) under-
0
goes irreversible ring opening reactions, and/or ether formation in protic sol
vents and that sodium borohydride behaves as a poor reducing agent in aprotic
solvents.
The triphenylcyclopropenyl cation (I) is converted entirely to its covalent
-4 3
ether (HI) at spectroscopic concentrations (ca. 10 M) in 95% ethanol.
The percentage of non-solvated cation (I) in solution will be determined by the
original concentration of cation, the pK of the cation in that solvent system, and
3
the pH of the solution. For example, calculations using pK (2.8) of the cation (I)
in 23% aqueous ethanol show that the ratio of cation (I) to covalent ether (HI) in a
0.10 molar 23% aqueous ethanolic solution of cation (I) is nearly 9:1. The trend in
3
the pK's (the pK of the cation in water is 3.1) indicates that the percentage of the
covalent ether (IH) would be somewhat greater under non-aqueous conditions. With
these considerations in mind, the at first surprisingly successful reduction (75-85%
17
yields) of triphenylcyclopropenyl cation (I) by sodium borohydride in ethanol is
understandable. This results indicates that sodium borohydride is potentially

3
a good trapping agent for stable cations under solvolytic conditions and promoted
the further investigation of reductions of the cation (I) with sodium borohydride in
both aqueous and anhydrous, protic solvents.
Earlier work by H. C. Brown10 indicate that ketones were reduced under
anhydrous conditions by sodium borohydride in diglyme, but more recent work has
indicated the actual rate of reaction is very slow and that the major portion of the
7
isolated products arises in the aquoeus work up. The possibility that the solvated
cation (I) might be more easily reduced by sodium borohydride that ketones and the
fact that heterogeneous reductions of the cation (I) with lithium aluminum hydride
in ether give good yields of triphenylcyclopropene (II)1 suggested that the investiga
tion of sodium borohydride reductions of (I) in ethers and other aprotic solvents
might be fruitful, both from a mecahnistic as well as synthetic point of view.
Hydroboration of the cyclopropene double bond was anticipated to be a possible side
reaction in aprotic solvents.
Lithium aluminum hydride in ether is a more powerful reducing agent toward
14
most functional groups than sodium borohydride in protic solvents, but the
insolubility of the cation (I) in ether solvents impairs the ability of lithium aluminum
hydride to bring about the desired reductions. It was hoped that an aprotic solvent
could be found that would give a homogeneous lithium aluminum hydride-cation (I)
solution, increase the yield of cyclopropene (II) and cut down the required reaction
time. The successful reduction of ketones with lithium aluminum hydride in
11-15
pyridine by Lansbury suggested that pyridine would be an excellent solvent for
the desired reductions. Lansbury's work with the reduction of various ketones with
lithium aluminum hydride in pyridine resulted from the need for solvent systems

4
for reduction of compounds which were insoluble in ethers and could not be re
duced by sodium borohydride in protic solvents. It was found that ketones were
readily reduced by lithium aluminum hydride in pyridine on addition of the reducing
agent to a pyridine solution of the ketone. Studies of ketones which were readily
cleaved, those where the living anion would be fairly stable, indicated that the
lithium aluminum hydride solutions were basic enough to bring about the cleavage.1
The suggested scheme for the reduction and cleavage of a ketone in fresh lithium
14
aluminum hydride-pyridine solution is:
O O A1H C H N 9~
R-C C0, + A1H" R C C0X R C C0f + C H N:A1H
3 4 X 3 i 5 5 3
i
H
o
R-C-H + 0 -C-
o
(3)
To further demonstrate the basicity of lithium aluminum hydride in pyridine,
16
it was used to generate the trityl anion from triphenylmethane. The basicity of
the lithium aluminum hydride-pyridine solution decreases with age. This was
demonstrated by the addition of an aged solution of lithium aluminum hydride in
pyridine to a colored solution of trityl anions. The color due to the anion vanished.
This result suggests that the nature of the reducing species in the solution is
changing with age.
It was found that an aged (24 hours) solution of lithium aluminum hydride in
pyridine was a more selective reducing agent than a fresh solution in that it reduced
11 12
aryl ketones in preference to alkyl ketones. The mechanistic details of the

5
reaction have not been determined, although nuclear magnetic resonance spectra
studies indicate that the dihydropyridyl aluminum complex (XI) is formed. This
species (XI) can be crystallized from concentrated, aged (24 hours) solutions.
The ratio of 1,2-dihydropyridyl to 1, 4-dihydropyridyl is approximately 1:1.
It was felt that the study of the reduction of a stable cation by this species
would be of interest in that it could possibly provide a convenient system for small
scale reductions of the cation (I) and give further insight into the nature of the
reagent.

DISCUSSION AND RESULTS
As can be seen from Table I, the sodium borohydride reductions all gave the
same products but in different proportions. A formalized reaction scheme which
shows the mode of formation of the observed product is:
As seen from the scheme, the three reactions competing with reduction in protic
solvents are: reversible ether or carbinol formation, irreversible ring opening,
and decomposition of the hydride by solvent. To minimize these side reactions the
reductions were attempted in aprotic solvents. The fact that the two reduction
reactions in anhydrous ether solutions, one in which sodium borohydride is soluble
(diglyme10) and one in which it is not, gave the same yield (25%) suggested that
reduction could be occurring during work up. This possibility was investigated by
quenching an ether-sodium borohydride reaction after only five minutes of stirring
with 5% hydrochloric acid solution (pH 1). The results indicate that the major
portion of the reduction must have occurred during work up. In fact the yield (55%)
6

TABLE 1
REDUCTIONS WITH SODIUM BOROHYDRIDE
Solvent
Reaction
Time
Ratio of Equiv.
cation/BH,,
Yield of
Triphenyl cy cl opropene
Yield of
Ketone
Ether
24 hr.
4
1:1
24
21
Anhydrous
diglyme
3 hr.
1:1.2
26
trace
60% Aqueous
diglyme
3 hr.
1:1.2
77
trace
Water pH 1
5 min.
1:1
55

Water pH 9
15 min
1:1
13

Pyridine
3 hr.
1:1
70
trace

8
from this reaction was higher than that from the anhydrous reactions which had
been stirred for periods up to 24 hours, although the yield is very close to that
76
obtained by Battiste and Breslow (50%) from a similar reaction. The difference
in yields is probably due to variations in the quenching rates and the amount of
hydrogen bromide in the cation (I) which reacts with the borohydride before water
is added. Somewhat surprising is the fact that the cation (I) can compete so well
for the hydride (55% yield of cyclopropene II) in a strongly acidic solution. When
the pH of the solution was raised to 9, only a 13% yield of cyclopropene (II) was
obtained. Although the yield is low and the reaction was not run under equilibrium
conditions, this yield is significant in that the cation (I) is hydrolized very rapidly
0
under basic conditions. It indicates that sodium borohydride can compete with
hydroxide for the cation and is a fairly good trapping agent for a stable cation even in
basic solution.
The good yield of triphenylcyclopropene from sodium borohydride reduction in
aqueous diglyme gives considerable information about borohydride reductions of
0
cations. First, as in the ketone reductions, the protic solvent appears to aid in
7 10
hydride transfer, but most interesting is the fact that the reduction proceeds
almost to the exclusion of hydrolysis of the cation despite possible destruction of
the borohydride ion in 60% aqueous diglyme solution (pH 5).
Studies of the reduction of ketones by sodium borohydride in pyridine by Richie
19 26
and others indicate that these reductions are very sluggish under anhydrous
conditions, but that the reaction is fairly rapid if lithium borohydride is used or if
lithium ion is added to the sodium borohydride solutions. The effect of the lithium
ion was rationalized to be due to its ability to increase the polarity of the carbonyl
group or to increase the dissociation of the sodium borohydride in solution.19 The

9
reduction of triphenylcyclopropenyl cation (I) in pyridine with sodium borohydride
gave good yields of triphenylcyclopropene (II) (70%). The reaction was followed by
ultraviolet spectroscopy and found to be complete within 30 minutes. This shows
that sodium borohydride is a reactive species in pyridine solution and suggests that
the role of the lithium ion in the reduction of ketones, in the same system, is to
enchance the polarization of the carbonyl group. The isolation of triphenylcyclopro-
18
pene in high yields indicates that the by-product of the reaction, C H N:BH ,
5 5 3
is not an effective hydroborating agent, at least towards cyclopropene double bonds.
The reductions of triphenylcyclopropenyl cation (I; Br) with lithium aluminium
hydride in pyridine gave surprisingly low yields of triphenylcyclopropene. The
yields from fresh solutions were 8-9% and from aged solutions, 2-4%. The major
portion of the reaction mixtures was inseparable oils, which were mainly ketonic
and were similar to those oils obtained from hydrolysis of the cation (I) itself. The
major cyclopropenyl component, obtained in 8-15% yields, was identified as 3, 5-
bis(triphenylcyclopropenyl)pyridine (VII). Structure assignment was made as follows.
The compound (VII) showed absorption in the infrared at 5. 5 p (characteristic of
1, 2-disubstituted cyclopropene) and an ultraviolet spectrum which was character
istic of 1, 2-diphenylcyclopropene chromophore. The analysis and molecular weight
(620, osomometer) satisfy the formula C^H^N. Using the calculated molecular
weight of 612, the molar extinction coefficient (53, 000) of the 316 millimicron band
is approximately twice that for 1, 2-diphenylcyclopropenes indicating that there are
two triphenylcyclopropenyl rings per molecule. The analytical and spectral data
are consistent with a molecule made up of two triphenylcyclopropene moieties
attached somehow to a pyridine nucleus. Compound (VII) showed three major peaks

10
in the nuclear magnetic resonance spectra. The tau values for these peaks were
1.43 (doublet), 2.31 (complex multiplet), and 2.75 tau (complex multiplet), with
the area ratio of the first peak to the second being ca. 9:2. Tetraphenylcyclopro-
pene shows a similar band at 2. 27 tau which corresponds to the ortho protons on
the phenyls conjugated with the double bond. Then with two triphenylcyclopropene
rings in the molecule as shown by the analytical data, the multiplet at 2. 31 tau in
VII should correspond to no less than 8 protons. The tau values for pyridine are
21
1.50 (ortho protons), 2.64 (para proton), and 3.10 (meta protons). The pertinent
22
coupling constants being J = .4 cps, J =1.9 cps. The position, area and
2-0 2-4
splitting (since a constant of 0.4 would not be seen) indicate that the 1.43 tau peak
of VII corresponds to two ortho protons on the pyridine ring. Possibly the added
area in the 2.31 tau peak is due to the one remaining proton on the pyridine nucleus.
This reduces the number of possible structures to two, the 3, 5- or the 3,4-bis-
(triphenylcyclopropenyl)pyridine. The similarity of the physical properties,
especially the melting point, and the spectra of 1,3-bis(triphenylcyclopropenyl)-
benzene to those of VII suggested that the correct structure was the 3,5- and not
the 3,4- structure. The following mechanistic considerations also support this
structure. The addition of the cation (I) to the pyridine nucleus is similar to the
38
addition of cation to other activated nuclei. The calculated electron densities for
pyridine, as well as experimental observations, indicate that electrophiles should
attack at the 3 and 5 positions. The electron density at these positions should be
even greater in a dihydropyridine. All of these considerations are consistent with
3, 5-bis(triphenylcyclopropenyl)pyridine (VII).

TABLE 2
LITHIUM ALUMINUM HYDRIDE IN PYRIDINE
Age of
Solution (hr.)
Reaction
Time (hr.)
Equiv.
cation/LAH
Yield
Triphenyl Cyclopropene
Yield of
VII
Yield
of
Ketone
0
Solid added
3
1:1.66
71
8 1
20
0.5
3
1:1
8%
9.11
17
24
3
1:1
21
i1
o
21
72
3
1:1
4%
8.51
(151)
20

12
The nature of the lithium aluminium hydride solutions will be examined before
a mechanistic scheme is suggested for the formation of the observed products. The
reaction of lithium aluminium hydride with cyclopropene (II) in pyridine provides
insight into the nature of the nucleophilic species present.
It was found that triphenylcyclopropene (II) was dimerized by excess lithium
aluminium hydride if the solid hydride was added to a pyridine solution of the
cyclopropene (II) but no dimerization was observed if the lithium aluminium hydride
pyridine solution had been aged before the cyclopropene (II) was added. Breslow
and Dowd1 found that triphenylcyclopropene was readily dimerized in a strongly
basic solution such as sodamide in liquid ammonia and these authors suggested the
following mechanism which does not involve formation of the triphenylcyclopropenyl
anion.
H H
VII
II
The same scheme should also apply in the lithium aluminium hydride-pyridine reac
tion since a fresh lithium aluminium hydride-pyridine solution is less basic than
liquid ammonia-sodamide solution, as evidenced by the low yield of trityl anion
16
formed in this solution, and proton abstraction from triphenylcyclopropene would
be even less likely. These reactions then allow examination of the nucleophylic
species present in lithium aluminium hydride solution in a case where there are

13
no acidic or reducible species involved.
The mechanism (6) suggested by Lansbury for the metalation of acidic hydro
carbons in pyridine lithium aluminum hydride solutions involves AlH^ as the
basic species which donates a hydride to the acidic proton with formation of a
14
molecule of hydrogen, the anion of the hydrocarbon, and A1H .
3
(6)
03C-H + H-AlHj + :KQ> 0f- + A Lansbury argues that pyridine is better able to solvate the alane (A1H ) than
O
ether. Thus enhancing the basicity of lithium aluminum hydride over that usually
observed in ether solvents. The basicity of the solution decreases with age to the
' 16
point that it appears able to donate a proton to a triaryl- or diarylmethide ion.
The suggested rational for this behavior is that the hydrogen remaining on the
aluminum after the complexation reaction has proceeded to the point that three
dihydropyridyls have added to the aluminum, becomes acidic, due to the electron
16
withdrawing effect of the dihydropyridyls, and protonates the anion. His results
show that there is an initial build up and subsequent reduction of the amount of
anion formed. Using fluorene as the organic acid, he found that after 1 hour 52%,
16
5 hours 71%, and 18 hours 48% metalation had occurred.
Lansbury's mechanistic scheme (6) cannot be applied to the dimerization of
triphenylcyclopropene. If the basic species were AlH^ as suggested, the attack
of this species on the double bond of triphenylcyclopropene (II) would result in
reduction. This being the case there must be another strongly basic species present.

14
As can be seen from the reductions of ketones with lithium aluminum deuteride^
in pyridine, there is a considerable amount of the species XII formed within 3
hours
XII xni
(7)
This species, due to the solvating power of pyridine toward alane (A1H ), should
O
dissociate to a considerable extent producing dihydropyridyl lithium (XIII) which
should catalyze the dimerization of triphenylcyclopropene. As the solution ages
the concentration of XII decreases due to the formation of XI therefore decreasing
the concentration of XIII assuming that XI is dissociated to a small extent. This
assumption seems reasonable since the bulky groups that would remain attached to
the aluminum on dissociation would decrease the solvation of the resulting aluminum
species. The electron withdrawing effect on the dihydropyridyls would also favor
association. These considerations then suggest that the basic species in fresh
lithium aluminum hydride pyridine solution is XII or XIII.
The reported results from the reaction of lithium aluminum hydride in pyri
dine with acidic hydrocarbons can be readily explained with this species serving as
the base. The following scheme, suggested by the above considerations, is consis
tent with the apparent reversibility of the reaction and the effect of excess lithium
aluminum hydride on the reaction. (Excess lithium aluminum hydride gives a

15
faster rate of anion formation but has little effect on the maximum concentration of
anion produced).16
An aged solution of XI would equilibrate with XII to reduce the basicity of the solu
tion due to the reasons given above. This is in agreement with the observed quench
ing of a solution of trityl anions by aged lithium aluminum hydride-pyridine solu-
16
tions.
The above considerations indicate that there is a dihydropyridyl alumino com
plex, either XI or XII, present in both aged and fresh lithium aluminum hydride
pyridine solutions. There are three points of attack open to the cation (I) on the
dihydropyridyl nucleus: (1) the 2 or 4 position to remove hydride and give cyclo-
propene, (2) the nitrogen to effectively trap out the cation and yield ketone on
hydrolysis, (3) or on the 3 and 5 carbons to yield VII. This scheme

16
is consistent with the results, but of course, it cannot be considered a proven
mechanism. The similarity of product distribution of those species arising from
routes other than reduction from the different solutions suggests that the equili
brium (7) plays a small role in these reactions and the reactivity of the cation
toward the pyridyl species is essentially the same whether the pyridyl is com-
plexed with the aluminum or is a free anion. This is reasonable on the basis of
26
the small steric requirement of XI found in reduction of ketones, and the
apparently small steric requirements for reduction of the cation as found in the
reduction of I with lithium tritertiarybutoxyaluminohydride.
The difference in yields of triphenylcyclopropene (II) from fresh (8%) and
aged (2-4%) solutions is probably due to reduction of the cation (I) by the lithium
aluminum hydride in the fresh solution before the reducing agent reacts with the
pyridine.

T
17
The high yields of ketone would be expected from a scheme of this nature
since the nitrogen would be the preferential point of attack on the dihydropyridyl
by an electrophile since it has a non-bonding pair of electrons. This part of the
scheme is also supported by the fact that the cation (I) is converted in pyridine
solution with an added secondary amine base to a species which appears to give a
ketone on hydrolysis.
It would be expected that the major product from reaction (9c) would be the
mono-alkylated product. The most obvious reason for not observing this product
is that it was obscured in the complex oils, although the ultraviolet spectra of the
oils remaining after crystallization showed only trace amounts of cyclopropenyl
material and this showed a spectrum identical to that of VII, but presumably the
mono-substituted pyridine would absorb in the same region. At any rate, it is
evident that the bis(triphenylcyclopropenyl)pyridine was the major cyclopropene
product formed in the reduction other than triphenylcyclopropene. More specu
lative explanations are that the bulky triphenylcyclopropenyl group once bonded to
the dihydropyridyl nucleus distorts the alumino complex to such an extent that the
mono-alkylated dihydropyridyl is easily alkylated again, or that the dihydropyridyl-
1, 2,3-triphenylcycloporpene is unstable under aqueous conditions and does not
aromatise as the bis-(triphenylcyclopropenyl)dihydropyridinyl does and yields a
complex mixture of products.
Another mechanism (10) was considered originally but was ruled out due to the
fact that sodium borohydride gave good yields of triphenylcyclopropene in pyridine.
Sodium borohydride would be expected to reduce the pyridine-cation complex
almost as readily as lithium aluminium hydride or lithium tetrakis(N-dihydro-

18
pyridyl)aluminate. The ultraviolet spectrum of triphenylcyclopropenyl bromide in
pyridine, however, does suggest that there is some complexation of the pyridine
with the cation (I) since the spectrum shows an absorption pattern very similar to
a covalent triphenylcyclopropene.
From a synthetic point of view, it is evident from the results of this investi
gation that the reduction of triphenylcyclopropenyl cation (I) in protic solvents are
not as complex as first anticipated and that the reduction of the cation (I) in aqueous
solvents with sodium borohydride gives high yields of triphenylcyclopropene. The
reduction under these conditions require short reaction times and give a high
purity product after a short work up procedure.
Mechanistically the study has provided insight into the nature of sodium boro
hydride and lithium aluminum hydride in pyridine and has provided more informa
tion about lithium tetrakis(N-dihydropyridyl)aluminate (XI) and related species.

CHAPTER II
THE PREPARATION AND THERMAL REARRANGEMENT OF
TETRAARYLCYCLOPROPENES
INTRODUCTION
The literature provides many examples of rearrangements of small
ring compounds to yield systems involving less bond angle strain, but
relatively few rearrangements have been reported in the cyclopropene
series. The following are some pertinent examples of thermal
rearrangements recently observed in the latter series.
These involve, except in the last instance, fairly complex pathways and
19

20
only formalized mechanisms have been proposed for each. As an example, the
mechanism suggested for the dimerization of triphenylcyclopropene and subse
quent rearrangement of the dimer is given below.
The complexity of the thermal rearrangements of these cyclopropene
derivatives prevents any conclusion to be drawn as to the mode of cyclopropene
ring cleavage and subsequent reactions of the ring fragment (s). Any attempt to
study the mechanism of such thermal rearrangements should first concern itself
with the initial ring cleavage step and the nature of the species produced. The
ideal cyclopropene derivative for a study of this nature would be one which yields
only primary products from ring cleavage.
Tetraarylcyclopropenes have no labile hydrogen, exhibit moderate thermal
stability, and have the essential features of the previous examples of cyclopropene
thermal rearrangements.
Tetraphenylcyclopropene, prepared by the reaction of phenylmagnesium
bromide with triphenyleyclopropenyl bromide, was initially investigated as a
8
possible deinophile in high temperature Diels-Alder reactions, but it was found
that at high temperatures the cyclopropene rearranged to another hydrocarbon.
Further investigation with the above mechanistic considerations in mind showed
that tetraphenylcyclopropene had rearranged after heating at 250 for
two hours to 1,2,3-triphenylindene. The reaction was apparently
XIX
XX

TABLE 3
Ar=
REACTION OF TRIPHENYLCYCLOPROPENYL BROMIDE
WITH ACTIVATED AROMATIC NUCLEI
Br + Aril
p -dimethylaminophenyl
p-methoxyphenyl
2,4-dimethoxyphenyl
2 -methoxy naphthyl
anthron-10-yl
m.p. 124-126
m.p. 193
m.p. 123-125
m.p. 198-199.5
B. Fohlish and P. Biirgle, Angew. Chem. internat. edit. Vol. 3, 699 (1964)

22
34
quantitative. Simple mechanistic considerations indicate that 1,2,3-
triphenylindene is the primary product from the species produced from thermal
cleavage of the tetraphenylcyclopropene ring. Thus tetraphenylcyclopropene is
suitable for the desired studies.
With the large body of information available on the effect of substituents
on various reactive intermediates, it was felt that the nature of the fragment
produced from rupture of the tetraarylcyclopropene ring could best be studied
by determining the effect of substituents on the rate of the reaction. The substituents
could also serve as labels to show the geometry of the intermediate species. The
most apparent method of introducing substituents into the tetraphenylcyclopropene
system is by the reaction of the appropriate arylmagnesium bromide with
triphenylcyclopropenyl bromide (I).
Reduction of the cation (I) by the less reactive Grignard reagents^0 to give
24
bistriphenylcyclopropene (XIV) or hexaphenylbenzene (X) was an anticipated
problem.
In attempts to prepare octaphenyltropylidene from heptaphenyltropylium cation
and phenylmagnesium bromide, complete reduction was observed.-^7 Reduction
has also been observed in attempts to prepare tetraphenylmethane from triphenyl-
methyl bromide and phenylmagnesium bromide. ^0
Shortly after the work on this problem had begun, the preparation of several
tetraarylcyclopropenes was reported by Fohlish and Biirgle.^ These authors
report that the triphenylcyclopropenyl cation substitutes onto activated aromatic
nuclei in 10:1 acetic acid-pyridine solution. The melting points of some of their
derivatives however, were not in accord with those of the tetraarylcyclopropenes
prepared by the Grignard method. The assignedstructures and melting points of
the products of the reported reaction are given in Table III. Since little information
about the procedure or yields were given in the report, this reaction was

23
reinvestigated to determine if this reaction could be utilized to prepare tetraarylcyclo-
propenes not obtainable by the use of Grignard reagents and to identify the products
which appeared not to have the assigned cyclopropene structure.
The ultraviolet absorption of the diphenylcyclopropene chromaphore appears
1-5
to be affected by changes in the groups attached to the saturated carbon. This
effect is very apparent in the derivatives of triphenyl and diphenylcyclopropene.
The 3-aryl-l,2,3-triphenylcyclopropenes provide a system in which subtle
electronic and steric effects may be observed.
This study in the tetraarylcyclopropene series was considered as a three
prong problem: (1) the preparation and properties of the cyclopropenyl compounds,
(2) study the products of rearrangement, and (3) determination of the effect of the
substituents on the rate of the rearrangement. The author's original goal was the
completion of the first phase of the problem but investigation of various problems
arising in the preparations gave results which led to preliminary investigations
of the other two phases.

DISCUSSION AND RESULTS
The reaction of aryl Grignard reagents with triphenylcyclopropenyl cation (I)
gave good yields of tetraarylcyclopropenes. The yields in most cases were only
slightly lower than 50%, based on the starting cation (I). Various reaction
conditions and work-up procedures were used depending on the nature of the
particular reaction.
The cation (I) is only slightly soluble in ether solvents so the reactions are
heterogeneous. To increase the homogeneity of the reaction, tetrahydrofuran
was used initially as the solvent because the cation (I) appeared to be more soluble
in this ether than in ethyl ether and its higher boiling point should permit a faster
reaction rate. In later reactions it was found that the reactions proceeded smoothly in
reflusing ethyl ether or in some cases in ethyl ether ar room temperature making
the use of tetrahydrofuran impractical since it is hard to dry. The reaction
reported for the formation of 1,3-(bistriphenylcyclopropenyl)benzene (XXII) was
run in tetrahydrofuran since the pra di-Grignard reagent could reportedly be
42
obtained only at the higher temperatures provided by this solvent, but later
preparations of the bistriphenylcyelopropenyl benzene (XXII) in ethyl ether gave
yields as high as those in tetrahydrofuran. These reactions gave nearly 45%
yields of products arising from the di-Grignard, (XIX) and (XXV), suggesting
that the meta di-Grignard is not as difficult to form as para di-Grignard.
The reactions appeared to be essentially complete after 24 hours but shorter
reaction periods have given lower yields. As an example, only 19% of the desired
cyclopropene was obtained in the reaction of p-chlorophenyl magnesium bromide
with cation (I) after 6 hr.
To check the dependence of the yield on concentration of Grignard reagent in
the more sterically hindered reactions, the preparation of 3-mesityl-1,2,3-
triphenylcyclopropene was run at two concentrations. The first reaction in which
the concentration of Grignard reagent was 0.27 M_, gave 18% of the desired
cyclopropene. The second run at a higher concentration, 1.0 M, gave a 52% yield.
The yield of the desired cyclopropene from the less hindered Grignard reagents appeared
24

25
to be less concentration dependant. Since there is a considerable amount of
hydrogen bromide associated with the bromide salt (I) from its preparation,
there was at least a 100% excess of Grignard reagent used in all reactions.
The tetraarylcyclopropene is formed before hydrolysis, therefore it can be
separated from the insoluble inorganic solids and unreacted cation (I) by filtration
of the reaction mixture before water is added.
-j- MgXBr
+ ArMgX
This procedure gave good results but the magnesium salts usually clogged
the filter and made filtration difficult. It was found that the tetraarylcyclopropenes
could be easily separated from the hydrolysis products of (I) by column
chromatography and this made the difficult filtration unnecessary. In cases where
the hydrocarbon resulting from hydrolysis of the Grignard reagent was high
boiling, the excess Grignard reagent was quenched by pouring the reaction
mixture onto powdered Dry Ice. The Dry Ice mixture was then acidified and
extracted with ether. The acid resulting from carbonation of the excess Grignard
reagent was then extracted from the ether solution with aqueous sodium bicarbonate.
This procedure eliminated the problem of removing the high boiling liquids which
came off the chromatography column together with the tetraarylcyclopropenes.

TABLE 4
GRIGNARD PREPARATIONS
Ar
mp.
Yield
(cm.*"*)
U.V. Spectra (cyclohexane) max (log )
ja-Anisyl
(XXI)
160-161
40%
1810
230(4.56), 300 (sh), 312(4.36),
334(4.20)34
jo-Anisyl
(XVII)
197-197.5
44%
1810
286(4.26), 303(4.31), 317(4.34),
333(4.29)
Mesityl
(XV)
150-151
45%
1820
280(4.14), 300(4.20)
-Naphthyl (XXXI)
189-190
52%
292.5(4.35), 299 (sh), 305 (sh) 318(4.27),
333(sh) (ca. 4.2)34
Phenyl3
(XIX)
178-179
47%
1825
280(4.53), 305(4.33), 318(4.35),
3.35(4.31)
V
w*
b (XXV)
235-236
17%
1800
228, 302, 316(4.73), 333
to
O

TABLE 4--Continued
Ar
mp.
Yield
(cm."^-)
U.V. Spectra (cyclohexane) max (log )
£-Tolyl (XVI)
134.5-1360
51%
1830
230(4.53), 307(4.32), 318(4.30),
336(4.26)
£-Chlorophenyl
(XVIII)
154-155
19%
1825
230(4.56), 304(4.33), 317(4.35),
334(4.31)
aPhenyl lithium used.
^Prepared from the bis-grignard of meta-dibromo benzene.
cAcetonitrile solvent.
to
<1

28
The yield of reduction products was much lower than expected from the
considerations given in the introduction. Only trace amounts of hexaphenylbenzene
(X) were isolated from the reaction with the meta-di-Grignard and only small
amounts of impure material which appeared to contain traces of the bistriphenyl-
cyclopropene (XIV) were obtained from the preparation of XV and XVIII. Consider
ing the steric requirements of the mesitylmagnesium bromide, it is surprising
that such a high yield of the cyclopropene derivative (XV) is obtained from the
reaction and such a small amount of reduction product is produced.
28
Shortly after the work on this problem had begun, Fohlish and Burgle
reported the preparation of tetraarylcyclopropenes by the reaction of triphenyl-
cyclopropenyl bromide (I) with activated aromatic nuclei. Their results are given
in Table III. The melting point for their p-anisyl derivative (XXI), 124-126, was
approximately 40 lower than that of 3-p-anisyl-l,2,3-triphenylcyclopropene, m.p.
160-126, prepared in our laboratory by the Grignard method.
The only reaction conditions given were that the cation (I) was heated with the
appropriate aromatic compound in 10:1 acetic acid-pyridine. In hopes of finding
this method useful for preparation of tetraarylcyclopropene derivatives not
obtainable by the Grignard method an investigation to determine the optimum
reaction conditions was undertaken. The results of this investigation are
summarized on Table V. Reasonable yields (based on starting cation (I) could be
realized only when a large excess of the aromatic reactant was used. This would
make the method practical only for micro scale preparations, expecially in cases
where the aromatic compound is expensive.

TABLE 5
REACTION OF TRIPHENYLCYCLOPROPENYL BROMIDE
WITH ANISOLE
Reaction
Equiv. of
pyridine
Equiv. of
Anisle
Equiv. of
Cation
Products
Cation
+ anisle
0.81
8.4
0.87
A light green fluorescent oil
Cation
+ anisle
5.3
21.0
2.7
10% XXI
Cation
+ anisle
3.7
33.6
2. 7
21% XXI
Cation
+ anisle
2.8
84.0
2. 7
47% XXI

30
The reaction was also investigated to determine the structure of the product,
m.p. 124-126, that had been isolated and reported from the reaction of anisle
with the cation (I). Since the compound was reported to show absorption at 5.5
microns in the infrared (characteristic of a 1,2-diphenylcyclopropene), it was
felt that the reported compound was possibly another tetraarylcyclopropene. The
high melting point of the o-anisyl derivative, m.p. 197-197.5, rules out this
most obvious possibility. The problem at this point appeared to be more complex
than had been anticipated and lead to further work which gave results that prompted
the investigation of the rearrangements of tetraarylcyclopropenes in greater detail
than had been originally planned but did not lead to a positive indentification of
the reported compound. These results will be discussed later.
The ultraviolet and infrared spectra of the tetraarylcyclopropenes prepared
inthis work showed larger substituent effects than origionally anticipated. The
spectral data is tabulated in Table IV. and reproductions of the ultraviolet
spectra shown in Figures 1-4. The characteristic 1,2-diphenylcyclopropene
-j 5
peak (1820 cm ) is present in all the infrared spectra on the tetraarylcyclo
propenes prepared. The position of the peak appears to be affected somewhat by
the substituents, but the spectra were recorded on a Perkin Elmer Infracord
infrared spectrometer and are not sufficiently accurate for substituent effect
correlations.
The ultraviolet spectra demonstrated the effect of substituents on the
wavelength of maximum absorption as well as the overall shape of the spectrum
for the 1,2-diphenylcyclopropenyl chromophore in each of the tetraarylcyclopro
penes. The effect can be attributed to either electronic or steric factors. The latter
are more apparent and are the most easily explained and will be discussed first.
The spectrum of tetraphenylcyclopropene (XIX) will be used as the reference in all
the discussion. The 3-mesityl derivative (XV) (Figure 1) exhibited a blue shift
in the long wavelength absorption band and a marked decrease in the absorption
intensity for this band. Models of the compound show that there is a definite
interaction between the ortho methyls on the mesityl ring and the phenyls conjugated
with the double bond of the cyclopropene ring. Less interaction results if the

Fig. 1.-Ultraviolet spectra of (XV), (XIX), and (XXI).

Fig. 2.-Ultraviolet spectrum of 3-(p-tolyl)-l,2,3-triphenylcyclopropene(XVI).
CO
tO

Fig. 3.-Ultraviolet spectrum of 3-(o-anisyl)-l,2,3-triphenylcyclopropene(XVII).
CO
CO

Fig. 4.-Ultraviolet spectrum of 5-(p-chlorophenyl)-l,2,5-triphenylcyclopro-
pene(XVIII).

35
phenyls are out of the plane of the cyclopropene ring. This would decrease the
conjugation which would, by analogy with cis and trans stilbenes,41 account for
a major part of the observed effect. Apparently, neither the steric effect nor the
electronic effect of one ortho substituent is very great. The o-anisyl derivative
(XVI) shows a spectrum which is only slightly different from that of tetraphenycy-
clopropene. This point is also supported by the fact that the 2,4-dimethoxy
derivative (XXII) shows a spectrum very similar to that of the p-anisyl derivative
(XXI).
As can be seen from Figures 1-4, the substituents in the para position effect
the wavelengths of maximum absorption very little, but have a pronounced effect
on the relative intensities of the three characteristic disubstituted cyclopropene
peaks. Although there have not been enough compounds prepared for substituent
effect-spectra shape correlation, there are enough differences in the spectra of the
p-chloro, p-methoxy, and p-methyl derivatives to allow a speculative examination
of the nature of the substituent effects. The chloro group (deactivates the ring
slightly) affects the spectra very little. The methoxy group (activates the ring
strongly) gives a pronounced change in the shape of the spectra. The methyl group
(activates the ring slightly) gives a different but very definite change in the relative
absorption intensities. It is apparent then that electron donating groups have the
largest effect on the shape of the absorption spectra. At first glance, the saturated
3-carbon would be expected to be able to transmit inductive effects very inefficiently.
The quite large effect of the substituents on the absorption spectra of the cyclopropene
chromophore is consistent with the methylene carbon having considerable sp^ character
since trigonal carbons are more efficient at transmitting inductive effects than a
30
tetrahedral carbon. The data collected to date indicates that with more examples
at hand the ultraviolet spectra of tetraarylcyclopropenes can possibly provide
further insight into the intriguing problem of the electronic structure of the cyclopro
pene ring.
In the investigation of the reaction of the cation (I) with anisle in acetic acid-
pyridine mixtures which was discussed earlier in this chapter, it was found that a
green fluorescent material was obtained when the reaction was run with insufficient

36
pyridine present. This material did not show a typical cyclopropene ultraviolet
spectrum, but exhibited a spectrum very similar to that of 1,2,3-triphenylindene
(XX). Crystals melting over a wide temperature range were obtained from the oils
after various solvent manipulations, suggesting a mixture of products. Since it was
known that tetraphenylcyclopropene rearranged to triphenylindene (XX) on heating
above 200, 3-p-anisyl-1,2,3-triphenyl cyclopropene (XXI) was rearranged to a
mixture of indenes by the same procedure. Attempts to completely separate the
mixture were futile. Chromatography yielded two crops of crystals that gave,
m.p. 138-145 and 145-148, respectively. The mixtures showed differentabsorp
tions in the infrared and ultraviolet and each appeared to be a mixture of at least
two different isomers. The infrared and ultraviolet spectra of the higher melting-
material were similar to those of the product form the electrophilic addition
reaction mentioned above. This further supported the supposition that the addition
reaction was yielding indenes and suggested two possible modes of formation of the
products.
29
The first possibility is similar to that suggested by Breslow and Battiste for
the rearrangement of diphenyl-(2,3-diphenylcyclopropenyl)- carbinol. This scheme
is
The other possibility is that the tetraarylcyclopropene is formed and then rearranged
under the reaction conditions. 3-p-Anisyl-l,2,3-triphenylcyclopropene was found
to be stable in acetic acid-pyridine solution in the absence of any proton catalysis.
The fact that the desired cyclopropene could be obtained when excess pyridine was
present suggested the possibility that the hydrogen bromide associated with the
bromide salt (I) was lowering the pH of the solution to such a point that the tetra
arylcyclopropene was undergoing an acid catalized rearrangement. When the

37
tetraarylcyclopropene (XXI) was heated at 90 in acetic acid with added hydrobromic
acid a green fluorescent oil was obtained which showed infrared and ultraviolet
spectra very similar to the spectra of the higher melting material obtained from the
thermal rearrangement. In contrast to this, tetrasubstituted cyclopropenes have
been shown to be very stable to base. ^ This suggests that the indenes from the
electrophilic addition reaction were probably formed from the tetraarylcyclopropene
31 37
(III) by a mechanism similar to that proposed by Breslow and Battiste for the
acid catalyzed rearrangement of 1,2-diphenylcyclopropenyl carboxylic acid and by
32
Breslow and Wolf for the acid catalized rearrangement of triphenylcyclopropene.
The mechanism for the acid catalyzed rearrangement of tetraarylcyclopropenes is:
This scheme is apparently general in the tri and tetraaryl cyclopropene series.
The mechanism indi cates that the group attached originally to the methylene carbon
of the cyclopropene should wind up in the 3- position of the resulting indene. This
is supported by the fact that triphenylcyclopropene undergoes acid catalyzed
32
rearrangement to 1,2-diphenylindene (XXIX) and by evidence in the literature that
27
arylindenes are not readily isomerized by acid.

38
By this mechanism then only two isomers should be formed from the acid catalyzed
rearrangement of a 3-aryl-l,2,3-triphenylcyclopropene. These two isomers
would be the 3-aryl-l,2-diphenylindene (C) and the triphenylindene substituted in
the 6 position (D).
The acid catalyzed rearrangement reactions provide a convenient source of
triarylindenes of known structure which promised to simplify the study of the
products from the thermal rearrangement of tetraarylcyclopropenes. With this
information at hand it was felt that a further study of the thermal products of
3-aryl-l, 2,3-triphenyl cyclopropenes should prove to be fruitful.
Tetraphenylcyclopropene had been found to rearrange to 1,2,3-triphenylindene
on heating at 250 for 2 hours. The reaction was shown to be actually a thermal
and not an acid catalyzed reaction at the surface of the glass by running the reaction
34
in a caustic coated vessel. The same product was obtained in the same length of
time under these conditions. The preliminary rate studies demonstrate the same
point although the rate does appear to be depressed slightly by washing the reaction
tubes with base instead of distilled water before the reaction is run. These same
rate studies indicate that the rate is first order.
Mechanistic considerations suggest three possible modes of thermal cleavage of
the cyclopropene ring. There could be a heterolytic cleavage giving rise to ionic
intermediates. This process appears unfavorable since either a vinyl carboniumion
or carbanion would have to be formed. The former is energetically unfavorable and
the latter is not consistent with results to be presented later. A concerted mechanism,
another possibility, would be energetically unfavorable since the bond in the cyclopropene
ring would have to cleave at the same time as the aromaticity of the phenyl ring was

shift
A
Mechanistic Scheme for Thermal
Rearrangement of
Tetraarylcyclopropenes
Figure 5
Shift
CO
CO

40
being destroyed. The third possibility is a homolytic cleavage giving rise to
diradical intermediates. The homolytic cleavage would appear to be the
44
favored mode of cleavage since in analogous vinyl cyclopropane rearrangements
where there are groups present in the cyclopropane which will stabilize a radical
38
species the intermediate has been shown to be diradical in nature. The
mechanism involving homolytic cleavage with a diradical intermediate is presented
in Figure 5.
As can be seen from the mechanistic scheme (Figure 5), the thermal rearrange
ment could yield four different isomers of the triarylindene. Even if one of the
paths to a product were unfavorable, thermal equilibration of the indenes could still
give rise to the complex mixture which would be very difficult to separate. This
was found to be the case in attempts to isolate the thermal products from the anisyl
derivative (XXI). Preliminary attempts at separation by the use of thin layer
chromatography indicate that this technique may provide a means of separation
of the various mixtures of isomers.
From the mechanistic schemes for acid catalyzed rearrangement and thermal
rearrangements of tetraarylcyclopropenes the products from the acid catalyze
reaction should be the same as those resulting from 1,2 hydrogen shift in the
thermal rearrangement. The acid catalyzed rearrangement them provide a means
of determining those products which arise from 1,2 hydrogen shift and those
arising from 1,3 hydrogen shifts.
The goals of the study of thermal products are three in number:
(1) the determination of the preferred path of hydrogen migration, (2) the effect
of product stability on the preferred path of hydrogen migration and (3) the deter
mination of the physical properties of the products in preparation for a study of
the kinetics.
Although many difficulties were encountered in the attempt to study the thermal
and acid catalyzed products from the anisyl derivative (XXI), the results implied
that information could be obtained if a tetraarylcyclopropene were studied which
would yield fewer products. The mesityl derivative (XV) was the compound
selected since the ortho positions on one of the phenyls are blocked by methyls. This should

41
decrease the number of thermal products possible to two.
On heating at 200 for two hours, 3-mesityl-1,2,3-triphenylcyclopropene (XV)
yielded two products which showed typical indene absorption in the ultraviolet and
were easily separated by solvent manipulations. Until structure assignment is made,
the higher melting isomer, m.p. 181-182, will be designated as (A) and the lower
melting isomer, m.p. 131-132, as (B). The ultraviolet spectra of the two compounds
first suggested that (A) was the l-mesityl-2,3-diphenylindene (XXIII) and (B) was the
3-mesityl-l-2-diphenylindene (XXIV) since
isomer (A) showed absorption at essentially the same wavelength as triphenylindene
and (B) absorbed at shorter wavelengths. It was felt that the bulky mesityl group
in the 3-position would decrease conjugation sterically and therefore (XXIV) would
show absorption at shorter wavelengths than 1,2,3-triphenylindene. Attempts to
43
isomerize B to A and A to B under basic conditions failed. The stability of B in
base under aerobic conditions suggested that it was not a 1-hydroindene as origi
nally thought. When 3-mesityl-l,2,3-triphenyl cyclopropene was heated at 85-90
under acidic conditions it was converted quantitatively to (A). Mechanistic consid-
32
erations and the fact that 1,2-diphenylindene was obtained exclusively from
triphenyl cyclopropene under the same conditions show that the isomer (A) is
3-mesityl-1,2-diphenylindene (XIV). This contention is supported by the results of
27
Koelch which indicate that arylindenes are not rearranged under acidic conditions.
The possibility that the product from acid catalyzed rearrangement is the thermo
dynamically favored product is ruled out by the fact that 1,2-diphenylindene is

42
isomerized almost quantitatively to the 2,3 isomer on heating at 250 for 2 hrs.
A mesityldiphenylindene prepared by a route which does not allow definite
assignment of structure showed the same ultraviolet absorption and retention time
on a thin layer chromatography plate as A. The method of preparation was:
XXIV
Preparation of the two isomers (XXIII) and (XXIV) is now under way.
The nuclear magnetic resonance spectra of the two isomers, A and B, were
very revealing. The isomer A showed resonance signals at 2.84 tau (aromatic)
4. 73 tau (indenyl proton), a doublet at 7.65 tau (methyl protons) and a singlet at
8.12 tau (methyl protons). The area ratio of the 7.65 tau peak to the 8.12 tau peak
was 2.1. The assignment of the 4. 75 tau peak was made by comparison to an m.m.r.
spectrum of 1,2,3-triphenylindene. (XX) The non-equivalence of the methyl protons
is consistent with the assigned structure. Models of the two possible isomers
(XXIII and (XXIV) show that the methyls on the mesityl group of XXIII are all in
essentially the same environment but in XXIV interaction of the mesityl or the
methyls with the 4-hydrogen on the fused benzene ring as well as with ortho hydro
gens on the 2-phenyl, restrict free rotation, producing a different magnetic environ
ment for the two ortho methyls, thus explaining their nonequivalence.
Sompound B showed three resonance signals due to methyl protons. These were
at 7.74 tau, 8.10 tau and 8.3 tau. The fact that the methyls are highly non-equiva
lent indicates that the mesityl nucleus has been destroyed. There was no signal
in the 4-5 tau region showing that :it was not a 1-hydroindene as suspected from the
chemical data. The methyl showing a resonance signal at 7. 74 tau is in an
environment similar to that of the ortho methyl in A. The 8.10 tau signal indicates
that there is one methyl on a phenyl similar to the para proton on A which is not
being affected by steric interactions. The third signal at 8.3 tau was in the same
region as the signal from the methyl protons (8.29 tau) of 1-methyl-1,2,3-triphenyl
indene (XXV). The structure, 1,4,6-trimethyl-l,2,3-triphenylindene (XXVI), is

43
most consistent with this data.
Me
Me
XXVI
The apparent mechanism for formation of (XXVI) involves a 1,2-methyl shift
as shown.
XXVI
XV
The attack of the intermediate species on the mesityl ring at a methyl position
suggests that the species is electron deficient and is most consistent with the
diradical representation.
The isolation of the 3-mesityindene (XXIV) isomer only would indicate on first
glance that 1-2 hydrogen migration is the preferred path to product, but the lack
of information about the thermal behavior of the mesitylindenes (XXIII) and (XXIV)
precludes any definite conclusions on this point.
With the above information at hand, work was begun on the third phase of the
problem, the determination of the kinetics of the thermal rearrangement of
tetraarylcyclopropenes. The nonvolatility of both the products and reactants and
the fact that both absorbed in the same region in the ultraviolet prevented the use
of vapor phase chromatography and ultraviolet spectroscopy for following the rate
of the reaction. The observation that 1,2,3-triphenylindene (XX) fluoresced strongly
in the visible region while tetraphenylcyclopropene did not, suggested that the
reaction could be followed by fluorimetry. Initial investigations show that the
rearrangement of tetraphenylcyclopropene can be followed by this techanique. Since

44
all the triaryl indenes prepared to date fluoresce, the method would appear to be
general. This technique, which has rarely been used in organic chemistry, could
prove to be very useful for following reactions involving a fluorescing species.
Spectrophotometric techniques involving emission measurements are more versatile
thank absorption techniques. In absorption measurements there is only one variable,
the wavelength of light passed through the sample. In emission measurements there
are two variables, the wavelength of excitation and the wavelength of emission.
In conclusion, the work to date has provided convenient procedures for the
preparation of tetraarylcyclopropenes, given further insight into cyclopropene
rearrangements, suggested some interesting studies of triarylindenes, and has
provided a method for the investigation of the kinetics of the rearrangements of
tetraaryl cyclopropenes.

CHAPTER III
EXPERIMENTAL
Physical Measurements
General
Melting points were taken on either a Fisher-Johns melting point block or a
Thomas-Hoover unimelt capillary melting point apparatus. All melting points
are uncorrected.
Carbon-hydrogen analyses were determined by Geller Microanalytical
Laboratories, Charleston, West Virginia.
Spectra
The infrared spectra were recorded on a Perkin-Elmer Infracord spectro
photometer. Solids were examined in potassium bromide pellets and oils were
examined on sodium chloride plates.
The ultraviolet spectra were recorded on a Cary 14 recording spectrometer.
The solvents smployed were suitable for use in spectrophotometry.
The nuclear magnetic resonance spectra were recorded in deuterochloroform,
with tetramethylsilane as the internal standard. A Varian DP-60 instrument was
used and the chemical shifts were calculated from measurements of sideband
spectra.
45

46
Procedure
Triphenylcyclopropenyl bromide (I) was prepared by the procedure of
3
Breslow and H. W. Chang, and had been dried in a drying pistol at 50 for at
least 24 hours prior to use. All anhydrous reactions were run under an argon
atmosphere.

47
REDUCTION OF TRIPHENYLCYCLOPROPENYL BROMIDE WITH SODIUM
BOROHYDRIDE
The sodium borohydride used was commercial grade. Diglyme was dis
tilled from calcium hydride and stored over Dri-Na. The ethyl ether was com
mercial anhydrous ether. All reaction vessels were dried at 120 for at least
2 hours before use. All liquids and solutions were transferred by means of a
syringe.
Reduction with sodium borohydride in ethyl ether
Triphenylcyclopropenyl bromide (1.74 g., 5.00 m moles) was suspended in
50 ml. of anhydrous ether. Sodium borohydride (0.19 g., 5.0 m moles) was
added in small portions to the suspension. The reaction mixture, after having
been stirred 24 hours at room temperature (35), was poured slowly into 100 ml.
of 5% hydrochloric acid. The solid that separated was taken up in 50 ml. of ether,
and the layers were separated. The aqueous layer was extracted twice with 25 ml.
portions of ether. The extracts were combined with the original organic layer
and dried over magnesium sulfate. After filtration and evaporation of the sol
vent, a mixture of solid and oil was obtained which was dissolved in 10% benzene-
hexane, and chromatographed over alumina (40 cc in hexane) using the solvent
sequence: hexane, 10% benzene-hexane, 15% benzene-hexane,20% benzene-
hexane, 30% benzene-hexane, 50% benzene-hexane, 50% benzene-ether, ether,
methylene chloride, and methanol. ,

48
The yield of triphenylcyclopropene (II) was 23.6% (0.316 g.), m.p. 108-105,
(lit. m.p. 112-1130)1, of benzylidene desoxybenzoin (IV) 0. 035 g., m.p. 96-100,
0
(lit. m.p. 100-101) of isobenzylidene desoxybenzoin (V) 0. 087 g., m.p. 85-88,
23
(lit. m.p. 89-90) and 0. 239 g. of a mixture of IV and V, m.p. 81-86. The
total yield of crystalline ketone was 21% with various ketonic oils (infrared) re
maining. An alcohol (0. 016 g.), m.p. 227-230, was isolated but not identified.
The compound exhibited infrared abosrptions at 2.87 microns.
Reduction with sodium borohydride in anhydrous diglyme
Triphenylcyclopropenyl bromide (I) (1.74 g., 5. 00 m moles) was suspended in
50 ml. of anhydrous diglyme. Sodium borohydride (0.227 g., 6. 00 m moles) was
added in small portions. The yellow slurry immediately became white. The re
action was stirred at room temperature for 3. 25 hours before it was poured into
75 ml. of water. The aqueous solution was extracted 3 times with 50 ml. portions
of ether. The extracts were combined and washed with 50 ml. portions of 5%
aqueous sodium bicarbonate and water before drying over magnesium sulfate. The
organic solution was concentrated to a thick oil by heating on a water bath under
reduced pressure. The oil was dissolved in 10% benzenehexane, and chromato
graphed as described previously. Triphenylcyclopropene (II) was obtained in 24%
yield (0.315 g.), m.p. 109-111. The ketonic oils yielded no crystals, but a trace
of solid alcohol (2. 87 p) was isolated from the 1:1 benzene-ether fraction.
Reduction with sodium borohydride in 60% aqueous diglyme
Triphenylcyclopropenyl bromide (I) (2.74 g., 5. 00 m moles) was suspended
in 30 ml. of diglyme. Sufficient water (20 ml.) was added to give the desired con
centration before 0. 227 g. (6. 00 m moles) of sodium borohydride was introduced

49
into the solution in small portions. The pH of a 60% aqueous diglyme solution was
5. The reaction was stirred for 3 hours, then worked up and chromatographed as
described for reaction I-A. The yield of triphenylcyclopropene (II) (m.p. 109-111)
was 77% (1.024 g.). Only trace amounts of benzilidene dexoxybenzoin (IV) were
isolated.
Reduction with sodium borohydride in strongly acidic solution
A suspension of triphenylcyclopropenyl bromide (I) (1.00 g., 2.72 m moles)
and sodium borohydride (0.11 g., 2.75 m moles) in 25 ml. of ethyl ether was
stirred for 5 minutes. The mixture was then poured fairly rapidly into a stirred
5% hydrochloric acid solution. The aqueous layer (pH 1) was separated and the
organic layer washed with 5% aqueous sodium bicarbonate and water. After drying
(MgSO^), the solution was concentrated to a viscous oil. Chromatography over
20 cc. of Merck alumina, with 5% benzene-hexane, as the elutant, gave 0.400 g.
(55%) of triphenylcyclopropene (II), m.p. 108-109.
Reduction with sodium borohydride in aqueous base
Triphenylcyclopropenyl bromide (1.00 g., 2. 72 m moles) was added in small
portions to a well-stirred 5% aqueous sodium bicarbonate solution (35 ml.) of
sodium borohydride (0.11 g., 2.76 m moles). After approximately half the cation (I)
had been added, ca. 5 ml. of ether was added to take up the oil which had formed
on the top of the water solution. The pH remained between 9 and 10 (measured with
pH paper) throughout the reaction. After all the cation (I) had been added, the
mixture was stirred for 15 minutes before the layers were separated. After wash
ing the organic layer with 5% hydrochloric acid and water and evaporating the solvent,

50
the oil was chromatographed over 20 cc. of alumina. The yield of triphenylcyclo-
propene (II) (m.p. 109-111) was 13% (0.096 g.).
Reduction with sodium borohydride in pyridine
Sodium borohydride (0.19 g., 5.0 m moles) was added in small portions over
a 10 minute period to 1.74 g. (5. 00 m moles) of triphenylcyclopropenyl bromide (I)
dissolved in 50 ml. of anhydrous pyridine. The yellow mixture was stirred 3 hours
at room temperature and then poured into 100 ml. of 5% hydrochloric acid. The
solid that formed was taken up in ether, the ether layer separated, and the aqueous
layer extracted with four 50 ml. portions of ether. The extracts were combined and
washed with 50 ml. portions of 5% hydrochloric acid, 5% aqueous sodium bicarbonate,
and water. The extracts were dried over magnesium sulfate, filtered, and concen
trated under reduced pressure to a viscous oil (1.45 g.). The oil was chromatograph
ed as before yielding 70% (.914 g.) of triphenylcyclopropene, m.p. 107-110, after
one recrystallization. Only trace amounts of benzilidene desoxybenzoin could be
isolated from the ketonic oils, but 015 g. of a red, crystalline compound (VI) was
isolated, m.p. 147-149. It was identified as 2, 3-diphenylindenone from its melting
25
point (lit. m.p. 151-152) and its infrared spectrum.
In a separate experiment, triphenylcyclopropenyl bromide (0.348 g., 1.00 in
moles) was dissolved in 10 ml. of anhydrous pyridine, sodium borohydride added in
small portions, and mixture allowed to stir for 3 hours. An aliquot of this solution
was taken before the borohydride was added and 30 minutes, 1 hour and 3 hours,
after addition. The aliquots were appropriately diluted with acetonitrile to give
_5
2. 0 x 10 M solutions. The ultraviolet spectra of these solutions were recorded
-2
using an acetonitrile reference solution that contained 2x10 % pyridine. A shift

51
f ^max from 300mp, shown by the sample taken before addition of the reducing
agent, to 313 mp, shown by the sample taken after the addition of the hydride,
indicated that the reduction was essentially complete in the first 30 minutes.
The ultraviolet spectrum of the reaction products after work up was the same as
that of the reaction mixture. The base line for the spectra of the samples con
taining sodium borohydride was lowered Off scale in the 355-370 mp region and
the pyridine showed strong absorption in the 220-270 mp region. The ultraviolet
acetonitrile
spectra of authentic samples of triphenylcyclopropenyl bromide (I) X -287.5,
max
acetonitrile
300, 316 millimicrons and triphenylcyclopropene (II) \ -303, 314, 332 milli-
max
microns were identical to those of the aliquots taken from the reaction.
REDUCTION OF TRIPHENYLCYCLOPROPENYL BROMIDE IN PYRIDINE WITH
LITHIUM ALUMINUM HYDRIDE
The lithium aluminum hydride-pyridine solutions were prepared by the method
11
of Lansbury and Peterson. Solutions were centrifuged, transferred to a second
argon-flushed tube fitted with a rubber septum, and aged under positive argon pres
sure. All transferring of liquids was accomplished with a syringe. The lithium
aluminum hydride used was not purified. Pyridine, dried and distilled from barium
oxide, was stored over potassium hydroxide pellets until used. All reaction vessels
had been dried at least 2 hours at 120.
Reduction with fresh lithium aluminum hydride solution
Lithium aluminum hydride (0.195 g., 5. 00 m moles) was added in small
portions to 50 ml. of 0.1 M solution of triphenylcyclopropenyl bromide (I) (5.0 in
moles) in pyridine. After an initial vigorous reaction, the reaction was stirred

52
for 3 hours at room temperature. The reaction was quenched by adding 50 ml. of
methanol. The methanolic solution was then poured into 100 ml. of 5% hydrochloric
acid, and the aqueous layer extracted with four 50 ml. portions of ether. Extracts
were combined, washed with 50 ml. portions of 5% hydrochloric acid, 5% aqueous
sodium bicarbonate, and water, and dried over magnesium sulfate. After evapora
tion of solvent, 2.0 g. of yellow oil were obtained. The oil was chromatographed
over Merck alumina as described previously. The first component off the column
was triphenyleyclopropene (II), identified from its infrared spectrum and melting
point; m.p. 106-107 after one recrystallization from ether-hexane. The crude
yield of cyclopropene was 7% (. 090 g.). From the third through fifth fractions,
two crystalline compounds were separated by fractional crystallization. They were
found to be benzilidene desoxybenzoin (IV) and isobenzilidene desoxybenzoin (V) from
their melting points and infrared spectra. Benzilidene desoxybenzoin was isolated
in 16% yield (.322 g.) m.p. 96-101 (lit. m.p. 100-101), isobenzilidenedesoxy-
23
benzoin in 2% yield (.038 g.), m.p. 80-85 (lit. m.p. 87-88), and various mix
tures of the two, m.p. 93-95, in 2% yield (. 034 g.). The remaining ketonic oils
from these fractions showed an infrared spectrum very similar to the oils obtained
from hydrolysis of the cation and appeared to be predominantly a mixture of the
benzilidene desoxybenzoins. The benzene and 1:1 benzene-ether fraction yielded a
fourth compound (VII) (0.124 g.), m.p. 235-237 (green melt) after two recrystalli
zations from methylene chloride-hexane; V
acetonitrile
(log e) 333(45, 000), 316(52,600),
max
302(50,600), 288(84, 000) mlJt; X
KBr
= 5.49, 12.8, 13.21, 14.28, 14.54 microns.
max
Anal. Caled, for C H N; mw. 612, C, 92.15; H, 5.53; N, 2.14. Found: mw
47 33
620 (vapor phase osmometer); C, 92.2; H, 5.54; N, 2.29.

53
Reduction with fresh lithium aluminum hydride solution
Triphenylcyclopropenyl bromide (I) (1.74 g., 5.00 m moles) was dissolved
in 30 ml. of anhydrous pyridine and 20 ml. of a 30 minute aged lithium aluminum
hydride-pyridine solution (0. 25 M, 5. 0 m moles) was added dropwise with a syringe
over a 15 minute period. The reaction was stirred at room temperature for 3
hours and then worked up as in A above. Slow elution chromatography on alumina,
using the already given solvents, yielded 8% (0.110 g.) of triphenylcyclopropene
in several fractions melting 104-111. Benzilidene desoxybenzoin, 17% (0. 245 g.),
and 0.138 g. of VII, m.p. 210-230, were also obtained.
Reduction with aged lithium aluminum hydride solution
A solution of lithium aluminum hydride (20 ml., 0. 25 M in LAH) that had been
aged 24 hours was added over a 10 minute period to 1.74 g. (5. 00 m moles) of
triphenylcyclopropenyl bromide dissolved in 30 ml. of pyridine. The reaction was
stirred for 3 hours at room temperature and worked up as in run A. Slow elution
chromatography yielded 2.4% (. 0321 g.) of n, m.p. 104-108, 0. 205 g. of IV,
m.p. 100-101, 079 g. of a mixture of IV and V (total yield of ketone, 21%), and
10% (0.157 g.) of VII, m.p. 198-200 (235-236 after two recrystallizations from
methylene chloride-hexane).
Reduction with concentrated, aged lithium aluminum hydride solution
To 1. 04 g. (3. 00 m moles) of solid cation (I) was added dropwise 29 ml. of
0. 25 M (. 0075 moles) lithium aluminum hydride-pyridine solution that had been
aged for 3 days. The reaction mixture became red and viscous on addition of the
reducing agent. After stirring for three hours at room temperature, the reaction
was worked up as before. Chromatography over alumina gave; 034 g. (4%) of

54
triphenylcyclopropene, (m.p. 107-110), 0.168g. of IV, 091 g. of V (total yield
of ketone 20%) and .076 g. (8%) of VII. A brown oil (.154 g.) remained from which
no more crystalline VII could be obtained but the ultraviolet spectrum indicated
that it was ca. 50% in VII. Trace amounts of 2,3-diphenylindenone (infrared
spectra 5. 89 p), m.p. 145-149, (lit. m.p. 151-152) were also isolated.
Stability of triphenylcyclopropene in lithium aluminum hydride-pyridine solution
Triphenylcyclopropene (0.134 g. ,0.5m moles) was dissolved in 5 ml. of
pyridine. Lithium aluminum hydride (0.65 m moles, 025 g.) was added in small
portions to the triphenylcyclopropene solution. The reaction mixture was stirred
3 hours at room temperature and poured into 20 ml. of 5% hydrochloric acid. A
white solid formed which was taken up in 20 ml. of ethermethylene chloride. The
layers were separated, the aqueous layer extracted with 15 ml. of ether and the
extract was combined with the origional layer, and the organic layer was washed
with 15 ml. of water. After drying over magnesium sulfate, the solvent was
separated, leaving a slightly yellow solid, which on recrystallization from ether-
hexane gave 0.108 g. (80%) of solid, m.p. 107-109, along with a yellow residue.
The reaction was repeated using 0. 268 g. (1.00 m moles) of triphenylcyclo
propene in 10 ml. of pyridine and 095 g. (2. 5 m moles) of lithium aluminum
hydride. The reaction was vigorous, turning from a yellow to a blue-green color.
After stirring 3 hours at room temperature the solution was worked up as above,
yielding 0. 246 g. (92%) of a slightly yellow material, m.p. 178-185. After one
recrystallization from benzene-hexane, this material melted at 182-183, turned
green at 225, and blue at temperatures above 240, ( lit. m.p. 179. 5-180. 50).1
This material was identified as the triphenylcyclopropene dimer (VIH) by its
melting point and infrared and ultraviolet spectra.

55
To 0.124 g. (0.46 m moles) of triphenylcyclopropene was added 7 ml. of an
0. 25 M (1.75 m moles) of a lithium aluminium hydride solution which had been
aged for 48 hours. The reaction was stirred for three hours and worked up as
above. A greyish solid was obtained (0.11 g.) m.p. 108-110 (90% recovery).
Reaction of triphenylcyclopropenyl bromide in pyridine with added amine
To a 0. 093 g/ml. solution of triphenylcyclopropenyl cation in pyridine con
tained in an ultraviolet spectra cell was added 0. 25 ml. of diethyl amine. The
solution showed absorptions, before addition of the amine, at: 324, 306 and 292 mp.
The optical density of the highest peak, the 324 mP maxima, was ca. 0.73.
Immediately after the addition of the amine, a new peak began to appear at 330 mp
which reached its maximum optical density, ca. 1.45, after 3 minutes, and the
origional cyclopropene peaks vanished. The spectra remained constant for 1 hour,
then ca. 10 drops of water were added. The spectra changed very slowly, and
after 48 hours a new peak had appeared at 315 millimicrons and the origional peale
had vanished. This last spectra was very similar, although not identical to the
ultraviolet spectra of benzilidene desoxybenzoin run under similar conditions.
Benzilidene desoxybenzoin shows absorption at 310 and 320 mp.
Reduction of triphenylcyclopropenyl bromide with lithium tritertiarybutoxyalumino-
hydride
Triphenylcyclopropenyl bromide (I) (1.74 g., 5. 00 m moles) was partially
dissolved in 25 ml. of tetrahydrofuran. A stirred suspension (1. 99 g., 7.5 m moles)
of lithium tritertiarybutoxyaluminohydride in 25 ml. of tetrahydrofuran was added
dropwise to the cation (I) mixture. All the solid had disappeared by the end of the
addition and a yellow solution resulted. The solution was'stirred for 20 hours at

56
room temperature and worked up as described in the previous reaction. Chroma
tography yielded 0. 766 g. of triphenylcyclopropene (57.2%, m.p. 108-111), 0. 013 g.
of 2, 3-diphenylindenone, m.p. 148-150, and 0.103 g. (7%) of benzilidene desoxy-
benzoin, m.p. 94-98.
Preparation of 1,3-bistriphenylcyclopropenyl benzene (IX)
The 1, 3-bis Grignard was prepared by adding 3. 8 ml. (. 020 moles) of 1, 3-di-
bromobenzene in 40 ml. of anhydrous tetrahydrofuran to 1. 08 g. of magnesium
turnings. An iodine crystal was used to initiate the reaction. A greyish suspen
sion had formed after 5 hours of refluxing under an argon atmosphere. The
resulting suspension was added dropwise to 13. 88 g. (. 040 moles) of triphenyl -
cyclopropenyl bromide (I) partially dissolved in 75 ml. of tetrahydrofuran. After
refluxing for 3 hours, the reaction was hydrolyzed with 6% aqueous ammonium
chloride solution. The layers were separated, the aqueous layer extracted with
50 ml. portions of ether and methylene chloride, the extracts combined, and the
organic solution dried over magnesium sulfate. Chromatography of the residue
after evaporation of the solvents yielded three major compounds. The first, from
the 25% benzene-hexane fraction, was tetraphenylcyclopropene (1.74 g., 25%),
m.p. 175-177 after three recrystallizations, identified by comparison to an au
thentic sample. The 50% benzene-hexane fraction yielded the desired product
(5. 5 p in infrared). This compound (4. 53 g. 17%) melted 231-233, after three
recrystallizations from benzene-hexane. The relatively pure material was percu-
lated through a short alumina column in 25% benzene-hexane and recrystallized
twice from benzene-hexane to give an analytical sample, m.p. 235-236 (green
melt above 270). This material gave a typical aryl cyclopropene absorption in the

57
ultraviolet, X. 228, 302, 316 (4173), 332 millimicrons. Anal. Caled, for
max
^48^34: C, 94. 5; H, 5.6. Found: C, 94.41; H, 5.98. The fifth fraction yielded
benzilidene desoxybenzoin, and the remaining fractions gave a solid (. 055 g.)
which melted above 300. The melting behavior and the infrared spectrum,
1 17
identical to an authentic sample, showed that this was hexaphenylbenzene (X).
PREPARATION OF TETRAARYLCYCLOPROPENES
The ether solvents were analytical grade and were distilled from lithium
aluminum hydride immediately before use. All reaction vessels were dried at
100-120 for at least two hours before use. All solvents were transferred with
a syringe and all reactions were run under an argon atmosphere.
The concentration of active Grignard was determined by adding 1 ml. of the
Grignard reagent to 10. 0 ml. of 0.100 M hydrochloric acid with a syringe and back
titrating the excess acid with 0.100M aqueous sodium hydroxide to a phenolphthalein
end point.
The magnesium used was finely cut turnings that had been washed with methyl
ene chloride and heated with an iodine crystal.
The chromatograms were run on a column of 20 g. of Merck alumina packed
in hexane. Eluting solvents were hexane, 5, 10, 20, 40, 60% benzene-hexane ,
benzene, 50% benzene-ether, ethyl ether, and methanol. A 50 ml portion of each
of the solvents was used except in cases where it was observed that a band was
coming off the column and then a portion large enough to completely elute the band
was used.

58
3-Mesityl-1, 2, 3-triphenylcyclopropene (XV)
To 0.360 g. (15.0 m moles) of magnesium covered with ether was added a
small portion (approximately 15 drops) of 1.96 ml. (13. 0 m moles) of mesityl
bromide. The reaction was initiated by heating and the addition of 10 pi. of methyl
iodide. The remaining mesityl bromide in ether (20 ml.) was added over a 20
minute period. After having been refluxed for 4 hours, the yellow solution was
0. 552 M in active Grignard. The Grignard solution was checked by adding 4 ml. of
the solution to Dry Ice. After appropriate work up, mesitoic acid was obtained,
3 5
m.p. 149-151, (Lit. m.p. 150-152). A portion (14 ml., 7.75 m moles) of the
mesityl Grignard prepared above was added dropwise over a 10 minute period to
1. 00 g. (2. 90 m moles) of triphenylcyclopropenyl bromide suspended in 15 ml. of
tetrahydrofuran. The reaction was exothermic and a large portion of the cation (I)
went into solution during the course of the addition. The reaction mixture was
stirred under reflux for 46 hours. The resulting suspension was filtered and the
solid washed twice with 35 ml. portions of methylene chloride. The filtrates were
combined and washed with 25 ml. portions of water, 5% hydrochloric acid, 5%
aqueous sodium bicarbonate, and water. Evaporation of the solvent after drying
over magnesium sulfate gave a viscous oil. The oil was placed on a chromato
graphy column in 10% benzene-hexane and the previously described elution scheme
followed. The excess mesitylene came off in the first fraction. The next three
fractions were combined and recrystallized from chloroform-methanol. This
material (0.203 g., 18%) showed absorption in the infrared at 1810 cm 1 and
melted 150-151 after three recrystallizations from methanol-chloroform. Although
this material did not show a typical arylcyclopropene spectrum (broad absorption at:

59
. cyclohexane
k,,,. 280 (4.14), 300 mp (4. 20)), the other data indicated that it was the
Ill cLX
desired compound. Anal, caled, for C H : C, 93.22; H, 6.78. Found:
30 20
C, 93.02; H, 6.87.
The fifth fraction yielded 0.150 g. of a yellow oil, which after crystallization
from methylenechloride-cyclohexane melted at 159-165, showed absorption in the
ultraviolet region at 315 and 238 mp and 1820 cm ^ in. the infrared. Fractions
six through nine yielded 0.324 g. of material, m.p. 95-97, which was identified
as benzylidene desoxybenzoin (IV) by its melting point (lit. m.p. 100-101) and
infrared spectrum.
The above reaction was repeated on a larger scale using higher concentrations
of reactants and a different work up procedure. A 1.805M solution of mesityl
Grignard was prepared by refluxing 5. 88 ml. (38.4 m moles) of mesityl bromide
with 1. 08 g. (45. 0 m moles) of magnesium in approximately 40 ml. of anhydrous
ether for 4. 5 hours. A portion of this solution (12 ml.; 22 m moles) was added
dropwise to a suspension of 3. 00 g. (8.75 m moles) of triphenylcyclopropene
bromide in 10 ml. of ether. The reaction was exothermic and all of the cation (I)
went into solution. The resulting solution was stirred at room temperature for
36 hours. The excess mesityl Grignard reagent was removed by pouring the
reaction mixture onto 50 cc. of powdered Dry Ice. After the viscous mass had
warmed to room temperature, 75 ml. of ether and 50 ml. of water were added
and the layers separated. The organic layer was washed with two 25 ml. portions
of aqueous sodium bicarbonate, 15 ml. of 0.1 N hydrochloric acid and 25 ml. of
water. After having been dried over magnesium sulfate, the solvent was evapo
rated to give 4. 0 g. of oily solid. Chromatography of the solid over 25 g. of

60
alumina using 5% benzene-petroleum ether as the eluent yielded three 50 ml.
fractions containing the desired compound. The first fraction yielded only trace
amounts, the second 1.25 g. of slightly yellow crystals, m.p. 139-145 (raised
to 151-152 after one recrystallization from methanol-chloroform), and the third,
two crops of total weight 0.114 g. yellow prisms, m.p. 150-151, m.p. 142-147,
respectively. The total yield of the desired cyclopropene was 52%.
3-p-Tolyl-l, 2, 3-triphenylcyclopropene (XVI)
To 0.36 g. (15 m moles) of magnesium turnings was added 2 ml. of a satu
rated solution of p-bromotoluene in ether. The reaction began immediately. The
addition of the remaining p-bromotoluene (2. 06 g.; 12.0 m moles) in 25 ml. of
ether was completed in 20 minutes. Most of the magnesium had disappeared after
the mixture had refluxed for 2. 5 hours. The solution was 0.41M in active Grignard.
The p-tolyl Grignard (18 ml. ,7.4m moles) was added dropwise to 1. 00 g.
(2.90 m moles) of triphenylcyclopropenyl bromide suspended in 15 ml. of tetra-
hydrofuran. After a 50 hour period of reflux with vigorous stirring, the reaction
was worked up by the filtration procedure described above. The first six fractions
off the chromatography column yielded 0. 529 g. (51%) of material, m.p. 127-131
after one recrystallization from methanol-chloroform. A sample prepared for
analysis, m.p. 134. 5-136, showed absorption in the ultraviolet region (milli
micron, log E) at 230 (4.53), 307 (4.32), 318 (4.30), 336 (4.26) and at 1830 cm'1
in the infrared.
Anal, caled, for C, 93.81; H, 6.19. Found: C, 93.58; H, 6.27.
28 22
The seventh and eighth fractions yielded 0.132 g. of benzylidene desoxybenzoin,
m.p. 96-100.

61
3-o-Anisyl-l, 2, 3-triphenylcyclopropene (XVII)
o-Bromoanisole (approximately 0.1 ml.) was added to magnesium turnings
(0.34 g.; 18. 0 m moles) that were covered with 3 ml. of ether. The reaction was
initiated by heating with an infrared lamp. The remaining 22 ml. of ether and
o-bromoanisole (total of 1. 87 ml.; 15. 0 m moles) was added over a 15 minute
period. A 0. 725 M solution of active Grignard reagent was obtained after 1 hour
of refluxing and 2 hours of stirring at room temperature. The solution was red
black. Triphenylcyclopropenyl bromide (1. 00 g.; 2.9 m moles) was suspended
in 15 ml. of tetrahydrofuran and 16. 5 ml. of 0.725 M Grignard solution (12. 0 m
moles) was added dropwise over a 5 minute period. During the exothermic reac
tion the cation went into solution and a white solid appeared shortly after the con
clusion of the addition. The reaction was stirred 20. 5 hours at room temperature.
Dry carbon dioxide was bubbled into the reaction for 1 hour before the entire reac
tion mixture was poured onto 20 cc. of crushed Dry Ice. After the resulting
viscous mass had warmed to room temperature, 40 ml. of 5% hydrochloric acid
and 25 ml. of methylene chloride were added. The organic layer was separated
and extracted with 50 ml. and 25 ml. portions of 5% aqueous sodium bicarbonate.
After a final wash with water, the solution was dried over magnesium sulfate and
filtered. Chromatography of the oily solid remaining after evaporation of the solvents
give 0.424 g. (39%) of material which showed absorption at 1810 cm 1 in the infrared
m.p. 197-197.5. This compound showed an ultraviolet spectrum that was typical
of the diphenylcyclopropene chromophore, ^^^(log^E) 286 (4*26) 303 (4.31),
317 (4.34), 333 m^ (4.29).
Anal, caled, for C^H0: C, 89.80;H, 5.92. Found: C, 89.58; H, 5.81.
28 22

62
3-(p-ChlorophenyI)-l, 2, 3-triphenylcyclopropene (XVIH)
A few crystals of 4-chlorobromobenzene were added to 0.132 g. (13. 0 m
moles) of magnesium turnings that were covered with ether. After gently warm
ing the flask to initiate the reaction, the remainder of the 4-chlorobromobenzene
(total of 2. 29 g.; 12. 0 m moles) dissolved in 20 ml. of ether was added over a
20 minute period. A solution that was 0. 59 M in active Grignard was obtained
after 2 hours of reflux. A portion (6 ml.; 3.54 m moles) of this solution was
added slowly to 1. 00 g. (2. 90 m moles) of triphenylcyclopropenyl bromide
suspended in 15 ml. of ether. The refraction was refluxed six hours and worked
up by the filtration method. The first fraction from the column gave a colorless
oil (0.203 g., 19%) which solidified on standing, m.p. 135-140. An analytical
sample of this material, m.p. 154. 5-155. 5, showed absorption at ^cyc^^exan
max
230 (4. 56), 304 (4.33), 317 (4.35), 334 mp (4.31) in the ultraviolet and at 1825 cm ^
in the infrared.
Anal, caled, for C H Cl: C, 85.06; H, 5.05. Found: C, 85.07, H, 5.09.
i i j. y
The next three fractions yielded 0.176 g. of benzylidene desoxybenzoin, m.p.
97-100. The sixth and seventh fractions yielded 0. 215 g. of yellow oil from which
traces of crystals, m.p. 165-195 were isolated.
Tetraphenylcyclopropene (XIX)
A 0.76M solution of phenyllithium was prepared by stirring 3.14 g. of bromo-
benzene in 15 ml. of ether with 0. 30 g. of lithium wire. A portion of this solution
(13 ml.; 10. 0 m moles) was added dropwise with a syringe to 1.74 g. (5. 00 m moles)
of triphenylcyclopropenyl bromide suspended in 10 ml. of ether. The resulting-
black solution was stirred for 7.5 hours at room temperature before 15 ml. of water

63
was added with cooling. After separation of the layers, the aqueous layer was
extracted with 20 ml. of ether. The organic layers were combined and washed
with 50 ml. of 0.1 N hydrochloric acid, two 25 ml. portions of aqueous sodium
bicarbonate, and 25 ml. of water. The yellow solid, obtained after evaporation
of the dried solution to dryness, was recrystallized from ether-hexane. Three
crops of material were obtained (0.798 g.; 47%), m.p., after two recrystalliza
tions, 177-179, (lit. m.p. 177-1780)8 \Cycl0lieXane 280 (4.5ff), 305 (4.33), 318
m.cLx
(4.35), 335 mp (4.31). This compound, identified as tetraphenylcyclopropene,
shows the typical disubstituted cyclopropene peak at 1825 cm 1 in the infrared.
34
3-(p-Anisyl)-l, 2, 3-triphenylcyclopropene (XXI)
To 0. 52 g. (22. 0 m moles) of magnesium and an iodine crystal was added
dropwise a solution of 2.7 ml. (21.0 m moles) of p-bromoanisole in 50 ml. of ether
and the mixture refluxed for 2 hours under argon. The resulting Grignard reagent
was added to a suspension of 4. 0 g. (9.35 m moles) of triphenylcyclopropenyl cation
(I) in 75 ml. of ether. After a 2 hour reflux period the reaction mixture was worked
up as in the previous reactions. Chromatography over_ca 40 g. of alumina yielded
two fractions which contained the desired cyclopropene (5. 5 microns in infrared).
The total yield of cyclopropene was 1.22 g. (40%), m.p. 160-161. The product
(XXI) showed major absorption in the infrared (KBr) at 3. 3, 3.4, 3.5, 5. 5, 6. 28,
6.6, 6.7, 6.8, 6.9, 7.8, 8.1, 8.55, 9.8, 12.05, 12.8, 13.2, 14.2, 14.7 microns
and in the ultraviolet (cyclohexane) at 230 (4.56), 300 (sh), 312 (4.36), 334 (420)
millimicrons.
Anal. Caled, for C H O: C, 89.8; H, 5.92. Found: C, 89.44; H, 6.02.
28 22

64
3-p-Anisyl-l, 2, 3-triphenylcyclopropene (XXI)
A. Triphenylcyclopropenyl bromide (0.300 g.; 0. 90 m moles) was dissolved
in a solution of 2 ml. of glacial acetic acid (2% acetic anhydride added), 1 ml. of
anisle and 3 drops of pyridine. The mixture was refluxed for 22 hours. After
cooling the mixture 10 ml. of water and 5 ml. of methylene chloride were added.
The layers were separated and the aqueous layer was extracted with two 10 ml.
portions of methylene chloride. After washing the organic layer three times with
10 ml. portions of 5% hydrochloric and 5% aqueous sodium bicarbonate, once with
water and drying over magnesium sulfate, the solvents were evaporated. The oily
residue was chromatographed over 15 g. of Merck alumina. The first 4 fractions
yielded a green, fluorescent oil which showed infrared absorption very similar to
that of the higher melting material from the thermal rearrangement. Crystals were
obtained which a broad range between 65 and 140 depending on the number of
recrystallizations.
B. Triphenylcyclopropenyl bromide (1. 00 g., 2. 90 m moles) was refluxed
with 2. 5 ml. of anisle, 6 ml. of acetic acid and 0. 6 ml. of pyridine for 20 hours.
The reaction was worked up as in A. The first two fractions from the chromato
graphy column gave 0.103 g. (10%) of crystals, m.p. 157-158, after trituration
with ether. The melting point and infrared spectrum were identical to those of
34
an authentic sample of 3-p-anisyl-l, 2,3-triphenylcyclopropene. The later
fractions yielded approximately 0.035 g. of 2, 3-diphenylindenone, m.p. 151-152.
(lit. m.p. 151-152) .
C. The above reaction was repeated using 1.00 g. (2. 90 m moles) of cation (I),
3.3 ml. of 10:1 acetic acid-pyridine, and 4. 0 ml. of anisle. The crude yield of the

65
desired cyclopropene (5. 5 micron peak in infrared) was 0. 231 g. (21%), obtained
in two fractions, m.p. 143-149, 157-158. Diphenylindenone, m.p. 150-152,
(0. 021 g.) was also obtained.
D. The best yield of the desired cyclopropene was obtained when the volume
of anisle was increased to 10 ml. and the volume of 10:1 acetic acid-pyridene was
decreased to 3 ml. Using these reaction conditions 0.520 g. (47%) of crude 3-p-
anisyl-1, 2, 3-triphenylcyclopropene was obtained from 1. 00 g. of cation (I). The
excess anisle was removed by heating the reaction mixture on a steam bath under
a stream of air.
3-(2, 4-Dimethoxyphenyl)-l, 2, 3-triphenylcyclopropene (XXII)
Triphenylcyclopropenyl bromide (1. 00 g.; 2. 90 m moles) was heated at 85-90
for 24 hours with 10 ml. of m-dimethoxybenzene and 3 ml. of 10:1 acetic acid
pyridine. After aqueous work up and evaporation of the low boiling solvents on a
rotary evaporator, the excess dimethoxybenzene was removed by heating the
reaction mixture on a steam bath under a stream of air. Chromatography over
25 g. of Merck alumina yielded 0. 243 g. (23%) of white crystals, m.p. 160-165,
which showed absorption at 5. 5 microns in the infrared. The melting point was
raised to 180-182 after 4 recrystallizations. The ultraviolet spectrum was
consistent with the cyclopropene structure showing absorption at: ^c^rolrm
331 (shoulder), 305, and 290.5 mp.
THERMAL REARRANGEMENTS
Tetraphenylcyclopropene
Tetraphenylcyclopropene (0.200 g.) was sealed in a tube which had been

66
evacuated to 0. 5 mm. pressure. After heating at 235-240 for 2 hours in a Wood's
metal bath and cooling to room temperature, a clear glass was obtained. The glass
was crystallized from methanol-chloroform to give 0.180 g. (90%) of 1,2,3-tri-
phenylindene (XX), m.p. 129-131 (lit. m.p. 131-132). An analytical sample,
m.p. 131-132, showed ultraviolet absorption (cyclohexane) at 306 (4.18), 241 mp
(4.38) and infrared absorption (KBr) at 1595, 1495, 1455, 1440, 911, 790, 773,
760 (w), 751, 739, 719 (w), and 695 (s) cm 1. A fluroescence spectra of the com
pound showed X (excitation) 355 and X (emission) 425 millimicrons,
max max
3-p-Aiisyl-l, 2,3-triphenylcyclopropene
3-p-Anisyl-l, 2,3-triphenylcyclopropene (0.350 g.) was heated at 200 for 4
hours in a sublimation apparatus under an argon atmosphere. A greenish, yellow
solid collected on the condenser. The contents of the reaction vessel were dissolved
in petroleumether and chromatographed on 30 g. of Woelm neutral alumina, activity
grade 1. The column was developed with 10% benzene-petroleum ether. Two bands
that were fluorescent under ultraviolet light were formed. The solvent sequence
used was 100 ml. mixtures of petroleum ether, benzene, ethyl ether, chloroform,
and methanol. The polarity of the solvents was gradually increased at such a rate
as to keep the band moving at a reasonable rate. Fractions of 40 ml. were taken.
The fractions were combined according to the fluctuation of the weight of material
per fraction. There was a total of 46 fractions taken. The first two yielded 0. 042 g.
of white solid, m.p. 138-145 after one recrystallization from ether-cyclohexane.
This material showed infrared absorption (KBr) at 3. 25, 3.4, 3.5, 6.21, 6.58,
6.71, 6.8, 6.9, 7.8, 8.08, 8.55, 9.68, 11.88, 12.35, 12.85, 13.08, 13.28, 13.58,
14.3, and 14. 5 microns and ultraviolet absorption (X ) in chloroform at 315 and
max
244 millimicrons.

67
The next nine fractions yielded crystals which were various mixtures (infra
red) of the first components and the later fractions. Fractions 12-14 yielded
0.085 g. of crystals, m.p. 145-148 after two recrystallizations from ether-
cyclohexane. These crystals showed infrared absorption at (KBr) 3.3, 3.4, 3.5,
6.25, 6.9, 7.41, 7.8, 7.9, 8.1 (w), 8.2, 9.7, 11.8, 12.06, 12.3, 12.75, 13.1 (w),
13.3, 13.4, 13.6, 14.3, and 14. 5 microns and ultraviolet absorption in chloroform
at 315 and 240 millimicrons.
The remaining fractions gave only traces of yellow or brown oils
3-Mesityl-1, 2, 3-triphenylcyclopropene
3-Mesityl-l, 2, 3-triphenylcyclopropene (0. 050 g.) was heated at 190-200 for
two hours in a small sublimation apparatus which had been flushed with argon. On
cooling a fluorescent oil formed on the condenser and sides of the flask. The oil
from the condenser (0.100 g., 10%) was triturated with methanol-chloroform to
yield white crystals (A), m.p. 179-181. The material from the bottom part of the
apparatus was crystallized from methanol-chloroform to give three crops of
crystals. The first crop (0. 010 g., 20%) was the same compound as that on the
condenser, m.p. 175-178. The next two crops yielded 0. 014 g. (28%) of crystals, (B)
131-132.
The higher melting material (A) showed absorption (X ) in the ultraviolet
ITLclX
in chloroform at 330 (sholder), 315 (sholder), 307, and 240 millimicrons and in the
infrared (KBr) at 3.3, 3.4, 3.5, 6.28, 6.7, 7.3, 9.35, 9.75, 11.0, 11.62, 11.9 (w),
12. 02 (w), 13.1, 13.55, 14.1 (w), 14.4, and 14. 55 microns. The lower melting
material (B) showed absorption (X ) in the ultraviolet in chloroform at 297.5
HL3X
and 240 millimicrons and in the infrared (KBr) at 3.3, 3.4, 3.5, 6.29, 6.7, 6.91,

68
9.3, 9.75, 11.6, 11.7, 11.9, 12.5, 13.0, 13.2, 13.61, 14.1, and 14.35 microns.
The nuclear magnetic spectra of the two compounds were considerably dif
ferent. The higher melting compound (A) showed the following resonance signals
in the n. m. r. (CDC1 ): complex multiple! at 2. 84 tau, sharp signlet at 4.73 tau,
O
a doublet at 7.65 tau, and a sharp singlet at 8.12 tau. The area ratio of the last
three peaks was 1:6:3. The lower melting material (B) showed the following-
resonance signals (CDC1 ): a complex multiplet at 2.80 tau, and sharp singlets

at 7. 74, 8.10, and 8.30 tau. The last three peaks were of equal area.
1, 2-Diphenvlindene (XXIX)
1, 2-Diphenylindene (0. 025 g.) was sealed in an 8 mm. tube. The tube had
been evacuated to approximately 0. 5 mm. before sealing. After having been
heated for 2 hours at 190-195 in a Wood's metal bath with the top of the tube
protruding above the liquid level, the major portion of the material had sublimed
into the upper portion of the tube. The tube was opened and the white needles were
scraped out, m.p. 175-178.
The above experiment was repeated with the tube completed submerged. The
starting material was recovered essentially unchanged, m.p. 175-177.
When the temperature was raised to 250*10, in a similar experiment, a
new product was obtained, m.p. 103-106: ^Cilj-0ri0rm 315 (shoulder), 307, 240.
max
This new solid was identified as 2, 3-diphenylindene (XXX) by its melting point
o 39
(lit. m.p. 107-108u) and ultraviolet spectrum.
Preparation of a mesityldiphenylindene
2, 3-Diphenylindenone (VI) (0. 025 g.) was dissolved in 3 ml. of ether and 3 ml.
of 0. 552M mesitylmagnesium bromide. The resulting red mixture was stirred

69
until the color had vanished (over night). After aqueous work up, a green oil was
obtained which did not show absorption in the carbonyl region (VI showed absorp
tion at 5. 87 microns) but showed a new peak near 2. 8 microns (alcohol). The
resulting oil was heated with 5% hydriodic acid in acetic acid. After work up and
chromatography on approximately 5 g. of alumina, using petroleum ether as the
eluent, a fluorescent oil was obtained which showed ultraviolet absorption at 330
(shoulder), 315 (shoulder), 307, and complex absorption after 290 millimicrons.
This oil showed the same retention time on a thin layer chromatogram (5:1 hexane-
benzene as eluent, silica gel HF adsorbent) as A found from the thermal re-
Orr
arrangement of 3-mesityl-i, 2, 3-triphenylcyclopropene.
Preparation of 1-methyl-1, 2, 3-triphenylindene (XXV)
n-Butyllithium (0. 6 ml. of 1. 5M solution in hexane) was added dropwise to
0.180 g. (5. 0 m moles) of 1, 2, 3-triphenylindene dissolved in 8 ml. of ethyl ether.
The resulting solution was refluxed for 3 hours before 0. 5 ml. of methyl iodide
(10 m moles) was added in small portions. After addition, the cloudy solution was
refluxed for 12 hours. Aqueous work up gave a yellow oil which was chromato
graphed on 25 g. of Merck alumina. Three 30 ml. fractions which yielded color
less oils were collected after elution with petroleum ether. The first two fractions
showed absorption in the ultraviolet at 303. 5 and 240 millimicrons. The third
fraction contained unreacted 1, 2,3-triphenylindene as identified by its ultraviolet
absorption at 307 and 240 millimicrons. Crystallization of the first two fractions
was attempted from methanol-chloroform and from petroleum ether. Finally an
oily solid was obtained from acetic acid which gave a white powder (0. 045 g.)
after grinding in a morter and pestle, m.p. 100-105. Although the melting point

70
3 0
is high (lit. m.p. 96-98), resonance signals at 2. 8 tau (complex multiplet) and
8.29 tau showed that the compound was 1-methyl-1, 2,3-triphenylindene.
Stability of mesityl indene (A) to basic conditions
To a saturated solution of ethanolic potassium hydroxide was added 0. 050 g.
of (A). The solution had been purged of oxygen by bubbling a stream of argon
through the ethanol before the reactants were added. The reaction mixture was
refluxed for 3. 5 hours under an argon atmosphere before it was neutralized with
20% hydrochloric acid. The organic layer, after normal work up, yielded a
yellow oil which crystallized on trituration with methanol-chloroform, m.p. 178-
180. The ultraviolet spectra and melting point showed it to be starting material.
The same results were obtained when n-butyllithium was used as the base.
When the above reaction was repeated with ethanolic potassium hydroxide
solution which had not been deoxygenated, a new solid was obtained, m.p. 183-
186. This material showed ultraviolet absorption at 345 (shoulder), 329, 316,
309 (shoulder), and 240 millimicrons, and infrared absorption at 2.8 microns.
Stability of mesityl indene (B) to basic conditions
A solution of 0. 010 g. in 10 ml. saturated ethanolic potassium hydroxide was
heated at 85-90 for 23 hours. The solution had not been deoxygenated and was
protected from the atmosphere only by a drying tube of calcium chloride. The
ultraviolet spectrum of the product after aqueous work up was identical to that of
the starting material.
Preliminary rate determination
Standard solutions of 1, 2, 3-triphenylindene were prepared. It was found that
a plot of In %T versus concentration was linear from concentration of 1. 00 to

71
1. 00 x 10 ^ mg. /ml. The %T values were determined on an American Instrument
Company, Fluoro-Microphotometer. Interference filters with maximum trans
mittance at 440 and 365 millimicrons were used as the emission and excitation
filters. A 1 mg./ml. solution of tetraphenylcyclopropene showed only slight
absorbance and no fluorescence with this set of filters. Solutions containing
lmg./ml. of total material were prepared that were 0, 5, 10, 20, 30, 40, 50,
70, and 100% 1,2,3-triphenylindene with enough tetraphenylcyclopropene added to
give the constant 1 mg./ml. concentration. A plot of In %T versus percent tri
phenylindene gave a smooth curve. Sample tubes (8 mm. x 15-20 cm.) were
prepared by two methods. All the tubes were first boiled in detergent and then
rinsed 10 times with distilled water. The first set was dried after rinsing. The
second set was soaked in concentrated ammonium hydroxide for at least three
hours and then rinsed with 5% aqueous ammonium hydroxide. Samples of tetra
phenylcyclopropene (0. 0100 g.) were weighed into the first set of tubes and half of
the tubes prepared by the second method. The remaining tubes in the second set
were filled with 1. 000 ml. of a 10. 00 mg./ml. solution of tetraphenylcyclopropene
in decahydronaphthalene. The tubes were sealed after they had been evacuated
and/or degassed to 0.25 mm. pressure. The tubes were introduced into a constant
temperature bath (192.4 I 0. 2) and allowed to remain for a prescribed period of
time. The temperature bath was a chamber filled with equilibrated decahydro
naphthalene vapors at atmospheric pressure. After the prescribed time period
(time was kept to within 2 sec.), the tubes were opened and the contents transferred
quantitatively to 10 ml. volumetric flasks with 10 ml. of dichloroethane
percent conversion of tetraphenylcyclopropene to triphenylindene was then deter-

72
mined by comparing the %T of these solutions to those of known percent concentra
tion. Plots of In (percent conversion) versus time were linear but the reaction
rate at this temperature was too rapid to give accurate rate constants although the
relative rates give a qualitative estimate of the effect of the various methods of
sample preparation of the rate of the reaction. Within 5 minutes, 97 percent
conversion was observed in the tubes washed with distilled water only, 60-70
percent conversion in the tubes that had been washed with base, and 20-30 percent
conversion in those tubes which had been washed with base and in which the sample
was in solution.
ACID CATALYZED REARRANGEMENTS
Tetraphenylcyclopropene (XIX)
Tetraphenylcyclopropene (0. 010 g.) was refluxed in a mixture of 2 ml. of
acetic acid, 0.4 ml. of benzene and 3 drops of concentrated hydrochloric acid for
24 hours. The reaction mixture was extracted with ether and the ether later was
washed with 5% aqueous sodium bicarbonate until there was no apparent acid-base
reaction in the separatory funnel. The ether solution was dried over magnesium
sulfate and filtered. After evaporation of the solvent, an oily residue (0. 012 g.)
was obtained which crystallized on trituration with cyclohexane-ether, m.p. 128-
130. The infrared and ultraviolet spectra of this compound was identical with
that of 1, 2, 3-triphenylindene obtained from the thermal rearrangement of tetra
phenylcyclopropene.
3-(p-Anisyl)-l, 2, 3-triphenylcyclopropene (XXI)
3-(p-Anisyl)-l, 2, 3-triphenylcyclopropene (0. 050 g.) was refluxed for 24 hours
in a mixture of 3 ml. of benzene, 2 ml. of acetic acid, and 2 drops of 48% aqueous

73
hydrobromic acid. After working up as described in preceding experiments,
0. 055 g. of brown oil was obtained. The infrared spectrum of the product,
especially in the ether region (7.5-8. 5 microns), and the melting point, m.p. 145-
149 after two recrystallizations from ether-cyclohexane, were essentially identi
cal to those of the higher melting material obtained from the thermal rearrange
ment.
3-Mesityl-l, 2, 3-triphenylcyclopropene (XV)
A solution of 0. 025 g. of 3-mesityl-l, 2,3-triphenylcyclopropene in 4 ml. of
acetic acid, 1 ml. of benzene, and 0.25 ml. of concentrated hydrochloric acid was
refluxed for 17 hours. After aqueous work up, 0. 016 g. of crystalline material,
m.p. 179-180, was obtained after trituration with ether-cyclohexane. The infra
red and ultraviolet spectra of this material was identical to that of the higher
melting material (A) obtained from the thermal rearrangement.
Triphenylcyclopropene (II)
To a mixture of 30 ml. of acetic acid, 6 ml. of benzene, and 1.8 ml. of
hydrochloric acid was added 0.385 g. of triphenylcyclopropene. The mixture was
stirred at 90-95 for 20 hours. A white solid (0.170 g.) crystallized from the
solution when it was cooled to room temperature, m.p. 178-180. A second crop
of crystals, m.p. 176-178, (0.190 g.) was obtained by concentrating the filtrate
under reduced pressure. This material showed absorption in the ultraviolet at
335 (shoulder), 320, 308, and 240 millimicrons. This compound was identified as
39
1, 2-diphenylindene by its melting point (lit. m.p. 177-178), infrared and
ultraviolet spectra, and mechanistic considerations.

75
26. P. T. Lansbury and R. E. Macleay, J. Am. Chem. Soc., 87, 831 (1965).
27. C. F. Koelsch, J. Am. Chem. Soc. 56, 1337 (1934).
28. B. Fohlish and P. Burgle, Angew. Chem. Internat. Edit., 3, 699 (1964).
29. R. Breslow and M. A. Battiste, J. Am. Chem. Soc., 82, 3626 (1960).
30. E. S. Gould, Mechanism and Structure in Organic Chemistry, Holt, Rinehart
and Winston, Inc., New York, 1959, p. 207.
31. R. Breslow, B. Winter, and M. Battiste, J, Org. Chem., 24, 415 (1959).
32. R. Breslow and P. Wolf, unpublished work, cited in ref. 44.
33. A. Small and R. Breslow, unpublished work, cited in ref. 44.
34. M. A. Battiste and J. Todd, unpublished work.
35. Org. Syn., Coll. Vol. HI, p. 555.
36. C. F. Koelsch and R. V. White, J. Org. Chem., 6_, 602 (1941).
37. R. Breslow and M. A. Battiste, unpublished work, cited in ref. 44.
38. N. P. Neureiter, J. Org. Chem., 24, 2044 (1959).
39. Ref. 25, p. 391.
40. M. S. Kharasch and Otto Reinmuth, Grignard Reactions of Non-metallic
Substances, Prentice Hall, Inc., New York, 1954, p. 1051.
41. J. R. Dyer, Applications of Absorption Spectroscopy of Organic Compounds,
Prentice Hall, Inc., Englewood Cliffs, N. J., 1965, p. 20.
42. H. Gilman, N. J. Beaber and J. L. Jones, Rec. trav. chim., 48, 597 (1928).
43. G. Bergson, Acta. Chem. Scand., 17, 2691 (1963).
44. R. Breslow, Molecular Rearrangements, Part I, P. De Mayo, Ed., Interscience
Publishers, Inc., New York, N. Y., 1963, pp. 236-261.

75
27. C. F. Koelch, J. Am. Chem. Soc. J56, 1337, (1934).
28. B. Fohlish and P. Burgle, Angew, Chem. Internat. Edit., 3, 699, (1964).
29. R. Breslow and M. A. Battiste. J. Am. Chem. Soc., 82, 3626, (1960).
30. E. S. Gould, Mechanism and Structure in Organic Chemistry, Holt, Rinehart
and Winston, Inc., New York, 1959, p. 207
31. R. Breslow, B. Winter, and M. Battiste, J. Org. Chem., 24, 415, (1959).
32. R. Breslow and P. Wolf, unpublished work, cited in ref. 44.
33. A. Small and R. Breslow, unpublished work, cited in ref. 44.
34. M. A. Battiste and J. Todd, unpublished work.
35. Org. Syn., Coll. Vol. Ill, p. 555.
36. C. F. Koelch and R. V. White, J. Org. Chem., 6, 602, (1941).
37. R. Breslow and M. A. Battiste, unpublished work, cited in ref. 44.
38. N. P. Neureiter, J. Org. Chem. 24, 2044, (1959).
39. Elsevier Encyclopedia of Organic Chemistry, Vol. 12 A., series III, Elsevier
Publishing Company, Inc., New York-Amsterdam, 1948, (a.)p. 237, (b.) p. 391.
40. M. S. Kharasch and Otto Reinmuth, Grignard Reagents, Prentice Hall, Inc.,
New York, 1954, p. 1051.
41. J. R. Dyer, Applications of Absorption Spectroscopy of Organic Compounds,
Printice Hall, Inc., Englewood Cliffs, N. J., 1965, p. 20.
42. H. Gilman, N. J. Beaber and J. L. Jones, Rec. trav. chim., 48, 597, (1929).
43. G. Bergson, Acta. Chem. Scand. 17, 2691, (1963).
44. R. Breslow, Molecular Rearrangements, Part I, Paul De Mayo, Editor
Interscience Publishers, Inc., New York, N. Y., 1963, pp 236 261.

BIOGRAPHICAL SKETCH
Robert Howard Grubbs was born on February 27, 1942, in Calvert City ,
Kentucky. He attended the public schools in Paducah, Kentucky, graduating
from Paducah Tilghman High School in June of 1960. He entered the University
of Florida in September of 1960, and obtained a Bachelor of Science degree in
December of 1963. Since entering graduate school at the University of Florida
in January of 1963, he has held teaching and research assitantships while working
toward a Master of Science degree. He plans to continue his education at
Columbia University.
76-

This thesis was prepared under the direction 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 Master of Science.
August 14, 1965
Dean, College of Arts and Sciences
Dean, Graduate School
Supervisory Committee:
Chairman

Internet Distribution Consent Agreement
In reference to the following dissertation or thesis:
AUTHOR:
Grubbs, Robert
TITLE:
Studies in the arylcyclopropene series (record number: 554225)
PUBLICATION DATE:
1965
dfri //, (oi-ifUi
as copyright holder for the aforementioned
I,
dissertation or thesis, hereby grant specific and limited archive and distribution rights to the Board of
Trustees of the University of Florida and its agents. I authorize the University of Florida to digitize and
distribute the dissertation or thesis described above for nonprofit, educational purposes via the Internet
or successive technologies.
This is a non-exclusive grant of permissions for specific off-line and on-line uses for an indefinite
term. Off-line uses shall be limited to those specifically allowed by "Fair Use" as prescribed by the
terms of United States copyright legislation (cf, Title 17, U.S. Code) as well as to the maintenance and
preservation of a digital archive copy. Digitization allows the University of Florida to generate image-
and text-based versions as appropriate and to provide and enhance access using search software.
This grant ofrpepnii
ibits use of the digitized versions for commercial use or profit.
Signature of Copyright Holder
Grufh
Printed or Typed Name of Copyright Holder/Licensee
Personal information blurred
Date of Signature
Please print, sign and return to:
Cathleen Martvniak
UF Dissertation Project
Preservation Department
University of Florida Libraries
P.O. Box 117007
Gainesville, FL 32611-7007



57
ultraviolet, X. 228, 302, 316 (4173), 332 millimicrons. Anal. Caled, for
max
^48^34: C, 94. 5; H, 5.6. Found: C, 94.41; H, 5.98. The fifth fraction yielded
benzilidene desoxybenzoin, and the remaining fractions gave a solid (. 055 g.)
which melted above 300. The melting behavior and the infrared spectrum,
1 17
identical to an authentic sample, showed that this was hexaphenylbenzene (X).
PREPARATION OF TETRAARYLCYCLOPROPENES
The ether solvents were analytical grade and were distilled from lithium
aluminum hydride immediately before use. All reaction vessels were dried at
100-120 for at least two hours before use. All solvents were transferred with
a syringe and all reactions were run under an argon atmosphere.
The concentration of active Grignard was determined by adding 1 ml. of the
Grignard reagent to 10. 0 ml. of 0.100 M hydrochloric acid with a syringe and back
titrating the excess acid with 0.100M aqueous sodium hydroxide to a phenolphthalein
end point.
The magnesium used was finely cut turnings that had been washed with methyl
ene chloride and heated with an iodine crystal.
The chromatograms were run on a column of 20 g. of Merck alumina packed
in hexane. Eluting solvents were hexane, 5, 10, 20, 40, 60% benzene-hexane ,
benzene, 50% benzene-ether, ethyl ether, and methanol. A 50 ml portion of each
of the solvents was used except in cases where it was observed that a band was
coming off the column and then a portion large enough to completely elute the band
was used.


9
reduction of triphenylcyclopropenyl cation (I) in pyridine with sodium borohydride
gave good yields of triphenylcyclopropene (II) (70%). The reaction was followed by
ultraviolet spectroscopy and found to be complete within 30 minutes. This shows
that sodium borohydride is a reactive species in pyridine solution and suggests that
the role of the lithium ion in the reduction of ketones, in the same system, is to
enchance the polarization of the carbonyl group. The isolation of triphenylcyclopro-
18
pene in high yields indicates that the by-product of the reaction, C H N:BH ,
5 5 3
is not an effective hydroborating agent, at least towards cyclopropene double bonds.
The reductions of triphenylcyclopropenyl cation (I; Br) with lithium aluminium
hydride in pyridine gave surprisingly low yields of triphenylcyclopropene. The
yields from fresh solutions were 8-9% and from aged solutions, 2-4%. The major
portion of the reaction mixtures was inseparable oils, which were mainly ketonic
and were similar to those oils obtained from hydrolysis of the cation (I) itself. The
major cyclopropenyl component, obtained in 8-15% yields, was identified as 3, 5-
bis(triphenylcyclopropenyl)pyridine (VII). Structure assignment was made as follows.
The compound (VII) showed absorption in the infrared at 5. 5 p (characteristic of
1, 2-disubstituted cyclopropene) and an ultraviolet spectrum which was character
istic of 1, 2-diphenylcyclopropene chromophore. The analysis and molecular weight
(620, osomometer) satisfy the formula C^H^N. Using the calculated molecular
weight of 612, the molar extinction coefficient (53, 000) of the 316 millimicron band
is approximately twice that for 1, 2-diphenylcyclopropenes indicating that there are
two triphenylcyclopropenyl rings per molecule. The analytical and spectral data
are consistent with a molecule made up of two triphenylcyclopropene moieties
attached somehow to a pyridine nucleus. Compound (VII) showed three major peaks


69
until the color had vanished (over night). After aqueous work up, a green oil was
obtained which did not show absorption in the carbonyl region (VI showed absorp
tion at 5. 87 microns) but showed a new peak near 2. 8 microns (alcohol). The
resulting oil was heated with 5% hydriodic acid in acetic acid. After work up and
chromatography on approximately 5 g. of alumina, using petroleum ether as the
eluent, a fluorescent oil was obtained which showed ultraviolet absorption at 330
(shoulder), 315 (shoulder), 307, and complex absorption after 290 millimicrons.
This oil showed the same retention time on a thin layer chromatogram (5:1 hexane-
benzene as eluent, silica gel HF adsorbent) as A found from the thermal re-
Orr
arrangement of 3-mesityl-i, 2, 3-triphenylcyclopropene.
Preparation of 1-methyl-1, 2, 3-triphenylindene (XXV)
n-Butyllithium (0. 6 ml. of 1. 5M solution in hexane) was added dropwise to
0.180 g. (5. 0 m moles) of 1, 2, 3-triphenylindene dissolved in 8 ml. of ethyl ether.
The resulting solution was refluxed for 3 hours before 0. 5 ml. of methyl iodide
(10 m moles) was added in small portions. After addition, the cloudy solution was
refluxed for 12 hours. Aqueous work up gave a yellow oil which was chromato
graphed on 25 g. of Merck alumina. Three 30 ml. fractions which yielded color
less oils were collected after elution with petroleum ether. The first two fractions
showed absorption in the ultraviolet at 303. 5 and 240 millimicrons. The third
fraction contained unreacted 1, 2,3-triphenylindene as identified by its ultraviolet
absorption at 307 and 240 millimicrons. Crystallization of the first two fractions
was attempted from methanol-chloroform and from petroleum ether. Finally an
oily solid was obtained from acetic acid which gave a white powder (0. 045 g.)
after grinding in a morter and pestle, m.p. 100-105. Although the melting point


25
to be less concentration dependant. Since there is a considerable amount of
hydrogen bromide associated with the bromide salt (I) from its preparation,
there was at least a 100% excess of Grignard reagent used in all reactions.
The tetraarylcyclopropene is formed before hydrolysis, therefore it can be
separated from the insoluble inorganic solids and unreacted cation (I) by filtration
of the reaction mixture before water is added.
-j- MgXBr
+ ArMgX
This procedure gave good results but the magnesium salts usually clogged
the filter and made filtration difficult. It was found that the tetraarylcyclopropenes
could be easily separated from the hydrolysis products of (I) by column
chromatography and this made the difficult filtration unnecessary. In cases where
the hydrocarbon resulting from hydrolysis of the Grignard reagent was high
boiling, the excess Grignard reagent was quenched by pouring the reaction
mixture onto powdered Dry Ice. The Dry Ice mixture was then acidified and
extracted with ether. The acid resulting from carbonation of the excess Grignard
reagent was then extracted from the ether solution with aqueous sodium bicarbonate.
This procedure eliminated the problem of removing the high boiling liquids which
came off the chromatography column together with the tetraarylcyclopropenes.


73
hydrobromic acid. After working up as described in preceding experiments,
0. 055 g. of brown oil was obtained. The infrared spectrum of the product,
especially in the ether region (7.5-8. 5 microns), and the melting point, m.p. 145-
149 after two recrystallizations from ether-cyclohexane, were essentially identi
cal to those of the higher melting material obtained from the thermal rearrange
ment.
3-Mesityl-l, 2, 3-triphenylcyclopropene (XV)
A solution of 0. 025 g. of 3-mesityl-l, 2,3-triphenylcyclopropene in 4 ml. of
acetic acid, 1 ml. of benzene, and 0.25 ml. of concentrated hydrochloric acid was
refluxed for 17 hours. After aqueous work up, 0. 016 g. of crystalline material,
m.p. 179-180, was obtained after trituration with ether-cyclohexane. The infra
red and ultraviolet spectra of this material was identical to that of the higher
melting material (A) obtained from the thermal rearrangement.
Triphenylcyclopropene (II)
To a mixture of 30 ml. of acetic acid, 6 ml. of benzene, and 1.8 ml. of
hydrochloric acid was added 0.385 g. of triphenylcyclopropene. The mixture was
stirred at 90-95 for 20 hours. A white solid (0.170 g.) crystallized from the
solution when it was cooled to room temperature, m.p. 178-180. A second crop
of crystals, m.p. 176-178, (0.190 g.) was obtained by concentrating the filtrate
under reduced pressure. This material showed absorption in the ultraviolet at
335 (shoulder), 320, 308, and 240 millimicrons. This compound was identified as
39
1, 2-diphenylindene by its melting point (lit. m.p. 177-178), infrared and
ultraviolet spectra, and mechanistic considerations.


Fig. 2.-Ultraviolet spectrum of 3-(p-tolyl)-l,2,3-triphenylcyclopropene(XVI).
CO
tO


This thesis was prepared under the direction 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 Master of Science.
August 14, 1965
Dean, College of Arts and Sciences
Dean, Graduate School
Supervisory Committee:
Chairman


2
Sodium borohydride would, at the outset, appear to be a poor reducing agent
for triphenylcyclopropenyl cation since it has been found that the cation (I) under-
0
goes irreversible ring opening reactions, and/or ether formation in protic sol
vents and that sodium borohydride behaves as a poor reducing agent in aprotic
solvents.
The triphenylcyclopropenyl cation (I) is converted entirely to its covalent
-4 3
ether (HI) at spectroscopic concentrations (ca. 10 M) in 95% ethanol.
The percentage of non-solvated cation (I) in solution will be determined by the
original concentration of cation, the pK of the cation in that solvent system, and
3
the pH of the solution. For example, calculations using pK (2.8) of the cation (I)
in 23% aqueous ethanol show that the ratio of cation (I) to covalent ether (HI) in a
0.10 molar 23% aqueous ethanolic solution of cation (I) is nearly 9:1. The trend in
3
the pK's (the pK of the cation in water is 3.1) indicates that the percentage of the
covalent ether (IH) would be somewhat greater under non-aqueous conditions. With
these considerations in mind, the at first surprisingly successful reduction (75-85%
17
yields) of triphenylcyclopropenyl cation (I) by sodium borohydride in ethanol is
understandable. This results indicates that sodium borohydride is potentially


BIOGRAPHICAL SKETCH
Robert Howard Grubbs was born on February 27, 1942, in Calvert City ,
Kentucky. He attended the public schools in Paducah, Kentucky, graduating
from Paducah Tilghman High School in June of 1960. He entered the University
of Florida in September of 1960, and obtained a Bachelor of Science degree in
December of 1963. Since entering graduate school at the University of Florida
in January of 1963, he has held teaching and research assitantships while working
toward a Master of Science degree. He plans to continue his education at
Columbia University.
76-


51
f ^max from 300mp, shown by the sample taken before addition of the reducing
agent, to 313 mp, shown by the sample taken after the addition of the hydride,
indicated that the reduction was essentially complete in the first 30 minutes.
The ultraviolet spectrum of the reaction products after work up was the same as
that of the reaction mixture. The base line for the spectra of the samples con
taining sodium borohydride was lowered Off scale in the 355-370 mp region and
the pyridine showed strong absorption in the 220-270 mp region. The ultraviolet
acetonitrile
spectra of authentic samples of triphenylcyclopropenyl bromide (I) X -287.5,
max
acetonitrile
300, 316 millimicrons and triphenylcyclopropene (II) \ -303, 314, 332 milli-
max
microns were identical to those of the aliquots taken from the reaction.
REDUCTION OF TRIPHENYLCYCLOPROPENYL BROMIDE IN PYRIDINE WITH
LITHIUM ALUMINUM HYDRIDE
The lithium aluminum hydride-pyridine solutions were prepared by the method
11
of Lansbury and Peterson. Solutions were centrifuged, transferred to a second
argon-flushed tube fitted with a rubber septum, and aged under positive argon pres
sure. All transferring of liquids was accomplished with a syringe. The lithium
aluminum hydride used was not purified. Pyridine, dried and distilled from barium
oxide, was stored over potassium hydroxide pellets until used. All reaction vessels
had been dried at least 2 hours at 120.
Reduction with fresh lithium aluminum hydride solution
Lithium aluminum hydride (0.195 g., 5. 00 m moles) was added in small
portions to 50 ml. of 0.1 M solution of triphenylcyclopropenyl bromide (I) (5.0 in
moles) in pyridine. After an initial vigorous reaction, the reaction was stirred


37
tetraarylcyclopropene (XXI) was heated at 90 in acetic acid with added hydrobromic
acid a green fluorescent oil was obtained which showed infrared and ultraviolet
spectra very similar to the spectra of the higher melting material obtained from the
thermal rearrangement. In contrast to this, tetrasubstituted cyclopropenes have
been shown to be very stable to base. ^ This suggests that the indenes from the
electrophilic addition reaction were probably formed from the tetraarylcyclopropene
31 37
(III) by a mechanism similar to that proposed by Breslow and Battiste for the
acid catalyzed rearrangement of 1,2-diphenylcyclopropenyl carboxylic acid and by
32
Breslow and Wolf for the acid catalized rearrangement of triphenylcyclopropene.
The mechanism for the acid catalyzed rearrangement of tetraarylcyclopropenes is:
This scheme is apparently general in the tri and tetraaryl cyclopropene series.
The mechanism indi cates that the group attached originally to the methylene carbon
of the cyclopropene should wind up in the 3- position of the resulting indene. This
is supported by the fact that triphenylcyclopropene undergoes acid catalyzed
32
rearrangement to 1,2-diphenylindene (XXIX) and by evidence in the literature that
27
arylindenes are not readily isomerized by acid.


55
To 0.124 g. (0.46 m moles) of triphenylcyclopropene was added 7 ml. of an
0. 25 M (1.75 m moles) of a lithium aluminium hydride solution which had been
aged for 48 hours. The reaction was stirred for three hours and worked up as
above. A greyish solid was obtained (0.11 g.) m.p. 108-110 (90% recovery).
Reaction of triphenylcyclopropenyl bromide in pyridine with added amine
To a 0. 093 g/ml. solution of triphenylcyclopropenyl cation in pyridine con
tained in an ultraviolet spectra cell was added 0. 25 ml. of diethyl amine. The
solution showed absorptions, before addition of the amine, at: 324, 306 and 292 mp.
The optical density of the highest peak, the 324 mP maxima, was ca. 0.73.
Immediately after the addition of the amine, a new peak began to appear at 330 mp
which reached its maximum optical density, ca. 1.45, after 3 minutes, and the
origional cyclopropene peaks vanished. The spectra remained constant for 1 hour,
then ca. 10 drops of water were added. The spectra changed very slowly, and
after 48 hours a new peak had appeared at 315 millimicrons and the origional peale
had vanished. This last spectra was very similar, although not identical to the
ultraviolet spectra of benzilidene desoxybenzoin run under similar conditions.
Benzilidene desoxybenzoin shows absorption at 310 and 320 mp.
Reduction of triphenylcyclopropenyl bromide with lithium tritertiarybutoxyalumino-
hydride
Triphenylcyclopropenyl bromide (I) (1.74 g., 5. 00 m moles) was partially
dissolved in 25 ml. of tetrahydrofuran. A stirred suspension (1. 99 g., 7.5 m moles)
of lithium tritertiarybutoxyaluminohydride in 25 ml. of tetrahydrofuran was added
dropwise to the cation (I) mixture. All the solid had disappeared by the end of the
addition and a yellow solution resulted. The solution was'stirred for 20 hours at


75
27. C. F. Koelch, J. Am. Chem. Soc. J56, 1337, (1934).
28. B. Fohlish and P. Burgle, Angew, Chem. Internat. Edit., 3, 699, (1964).
29. R. Breslow and M. A. Battiste. J. Am. Chem. Soc., 82, 3626, (1960).
30. E. S. Gould, Mechanism and Structure in Organic Chemistry, Holt, Rinehart
and Winston, Inc., New York, 1959, p. 207
31. R. Breslow, B. Winter, and M. Battiste, J. Org. Chem., 24, 415, (1959).
32. R. Breslow and P. Wolf, unpublished work, cited in ref. 44.
33. A. Small and R. Breslow, unpublished work, cited in ref. 44.
34. M. A. Battiste and J. Todd, unpublished work.
35. Org. Syn., Coll. Vol. Ill, p. 555.
36. C. F. Koelch and R. V. White, J. Org. Chem., 6, 602, (1941).
37. R. Breslow and M. A. Battiste, unpublished work, cited in ref. 44.
38. N. P. Neureiter, J. Org. Chem. 24, 2044, (1959).
39. Elsevier Encyclopedia of Organic Chemistry, Vol. 12 A., series III, Elsevier
Publishing Company, Inc., New York-Amsterdam, 1948, (a.)p. 237, (b.) p. 391.
40. M. S. Kharasch and Otto Reinmuth, Grignard Reagents, Prentice Hall, Inc.,
New York, 1954, p. 1051.
41. J. R. Dyer, Applications of Absorption Spectroscopy of Organic Compounds,
Printice Hall, Inc., Englewood Cliffs, N. J., 1965, p. 20.
42. H. Gilman, N. J. Beaber and J. L. Jones, Rec. trav. chim., 48, 597, (1929).
43. G. Bergson, Acta. Chem. Scand. 17, 2691, (1963).
44. R. Breslow, Molecular Rearrangements, Part I, Paul De Mayo, Editor
Interscience Publishers, Inc., New York, N. Y., 1963, pp 236 261.


22
34
quantitative. Simple mechanistic considerations indicate that 1,2,3-
triphenylindene is the primary product from the species produced from thermal
cleavage of the tetraphenylcyclopropene ring. Thus tetraphenylcyclopropene is
suitable for the desired studies.
With the large body of information available on the effect of substituents
on various reactive intermediates, it was felt that the nature of the fragment
produced from rupture of the tetraarylcyclopropene ring could best be studied
by determining the effect of substituents on the rate of the reaction. The substituents
could also serve as labels to show the geometry of the intermediate species. The
most apparent method of introducing substituents into the tetraphenylcyclopropene
system is by the reaction of the appropriate arylmagnesium bromide with
triphenylcyclopropenyl bromide (I).
Reduction of the cation (I) by the less reactive Grignard reagents^0 to give
24
bistriphenylcyclopropene (XIV) or hexaphenylbenzene (X) was an anticipated
problem.
In attempts to prepare octaphenyltropylidene from heptaphenyltropylium cation
and phenylmagnesium bromide, complete reduction was observed.-^7 Reduction
has also been observed in attempts to prepare tetraphenylmethane from triphenyl-
methyl bromide and phenylmagnesium bromide. ^0
Shortly after the work on this problem had begun, the preparation of several
tetraarylcyclopropenes was reported by Fohlish and Biirgle.^ These authors
report that the triphenylcyclopropenyl cation substitutes onto activated aromatic
nuclei in 10:1 acetic acid-pyridine solution. The melting points of some of their
derivatives however, were not in accord with those of the tetraarylcyclopropenes
prepared by the Grignard method. The assignedstructures and melting points of
the products of the reported reaction are given in Table III. Since little information
about the procedure or yields were given in the report, this reaction was


23
reinvestigated to determine if this reaction could be utilized to prepare tetraarylcyclo-
propenes not obtainable by the use of Grignard reagents and to identify the products
which appeared not to have the assigned cyclopropene structure.
The ultraviolet absorption of the diphenylcyclopropene chromaphore appears
1-5
to be affected by changes in the groups attached to the saturated carbon. This
effect is very apparent in the derivatives of triphenyl and diphenylcyclopropene.
The 3-aryl-l,2,3-triphenylcyclopropenes provide a system in which subtle
electronic and steric effects may be observed.
This study in the tetraarylcyclopropene series was considered as a three
prong problem: (1) the preparation and properties of the cyclopropenyl compounds,
(2) study the products of rearrangement, and (3) determination of the effect of the
substituents on the rate of the rearrangement. The author's original goal was the
completion of the first phase of the problem but investigation of various problems
arising in the preparations gave results which led to preliminary investigations
of the other two phases.


LIST OF FIGURES
Figure Page
1. Ultraviolet Spectra of Tetraphenylcyclopropene, (XV), (XIX), and
(XXI) 31
2. Ultraviolet spectrum of 3-(p-tolyl)-l, 2,3-triphenylcyclopropene. 32
3. Ultraviolet spectrum of 3-(o-anisyl)-l, 2, 3-triphenylcyclopropene. 33
4. Ultraviolet spectrum of 3-(p-chlorophenyl)-l, 2,3-triphenylcyclo
propene 34
5. Mechanistic Scheme for Thermal Rearrangement of Tetraphenylcyclo
propene 39
v


shift
A
Mechanistic Scheme for Thermal
Rearrangement of
Tetraarylcyclopropenes
Figure 5
Shift
CO
CO


DISCUSSION AND RESULTS
As can be seen from Table I, the sodium borohydride reductions all gave the
same products but in different proportions. A formalized reaction scheme which
shows the mode of formation of the observed product is:
As seen from the scheme, the three reactions competing with reduction in protic
solvents are: reversible ether or carbinol formation, irreversible ring opening,
and decomposition of the hydride by solvent. To minimize these side reactions the
reductions were attempted in aprotic solvents. The fact that the two reduction
reactions in anhydrous ether solutions, one in which sodium borohydride is soluble
(diglyme10) and one in which it is not, gave the same yield (25%) suggested that
reduction could be occurring during work up. This possibility was investigated by
quenching an ether-sodium borohydride reaction after only five minutes of stirring
with 5% hydrochloric acid solution (pH 1). The results indicate that the major
portion of the reduction must have occurred during work up. In fact the yield (55%)
6


50
the oil was chromatographed over 20 cc. of alumina. The yield of triphenylcyclo-
propene (II) (m.p. 109-111) was 13% (0.096 g.).
Reduction with sodium borohydride in pyridine
Sodium borohydride (0.19 g., 5.0 m moles) was added in small portions over
a 10 minute period to 1.74 g. (5. 00 m moles) of triphenylcyclopropenyl bromide (I)
dissolved in 50 ml. of anhydrous pyridine. The yellow mixture was stirred 3 hours
at room temperature and then poured into 100 ml. of 5% hydrochloric acid. The
solid that formed was taken up in ether, the ether layer separated, and the aqueous
layer extracted with four 50 ml. portions of ether. The extracts were combined and
washed with 50 ml. portions of 5% hydrochloric acid, 5% aqueous sodium bicarbonate,
and water. The extracts were dried over magnesium sulfate, filtered, and concen
trated under reduced pressure to a viscous oil (1.45 g.). The oil was chromatograph
ed as before yielding 70% (.914 g.) of triphenylcyclopropene, m.p. 107-110, after
one recrystallization. Only trace amounts of benzilidene desoxybenzoin could be
isolated from the ketonic oils, but 015 g. of a red, crystalline compound (VI) was
isolated, m.p. 147-149. It was identified as 2, 3-diphenylindenone from its melting
25
point (lit. m.p. 151-152) and its infrared spectrum.
In a separate experiment, triphenylcyclopropenyl bromide (0.348 g., 1.00 in
moles) was dissolved in 10 ml. of anhydrous pyridine, sodium borohydride added in
small portions, and mixture allowed to stir for 3 hours. An aliquot of this solution
was taken before the borohydride was added and 30 minutes, 1 hour and 3 hours,
after addition. The aliquots were appropriately diluted with acetonitrile to give
_5
2. 0 x 10 M solutions. The ultraviolet spectra of these solutions were recorded
-2
using an acetonitrile reference solution that contained 2x10 % pyridine. A shift


14
As can be seen from the reductions of ketones with lithium aluminum deuteride^
in pyridine, there is a considerable amount of the species XII formed within 3
hours
XII xni
(7)
This species, due to the solvating power of pyridine toward alane (A1H ), should
O
dissociate to a considerable extent producing dihydropyridyl lithium (XIII) which
should catalyze the dimerization of triphenylcyclopropene. As the solution ages
the concentration of XII decreases due to the formation of XI therefore decreasing
the concentration of XIII assuming that XI is dissociated to a small extent. This
assumption seems reasonable since the bulky groups that would remain attached to
the aluminum on dissociation would decrease the solvation of the resulting aluminum
species. The electron withdrawing effect on the dihydropyridyls would also favor
association. These considerations then suggest that the basic species in fresh
lithium aluminum hydride pyridine solution is XII or XIII.
The reported results from the reaction of lithium aluminum hydride in pyri
dine with acidic hydrocarbons can be readily explained with this species serving as
the base. The following scheme, suggested by the above considerations, is consis
tent with the apparent reversibility of the reaction and the effect of excess lithium
aluminum hydride on the reaction. (Excess lithium aluminum hydride gives a


68
9.3, 9.75, 11.6, 11.7, 11.9, 12.5, 13.0, 13.2, 13.61, 14.1, and 14.35 microns.
The nuclear magnetic spectra of the two compounds were considerably dif
ferent. The higher melting compound (A) showed the following resonance signals
in the n. m. r. (CDC1 ): complex multiple! at 2. 84 tau, sharp signlet at 4.73 tau,
O
a doublet at 7.65 tau, and a sharp singlet at 8.12 tau. The area ratio of the last
three peaks was 1:6:3. The lower melting material (B) showed the following-
resonance signals (CDC1 ): a complex multiplet at 2.80 tau, and sharp singlets

at 7. 74, 8.10, and 8.30 tau. The last three peaks were of equal area.
1, 2-Diphenvlindene (XXIX)
1, 2-Diphenylindene (0. 025 g.) was sealed in an 8 mm. tube. The tube had
been evacuated to approximately 0. 5 mm. before sealing. After having been
heated for 2 hours at 190-195 in a Wood's metal bath with the top of the tube
protruding above the liquid level, the major portion of the material had sublimed
into the upper portion of the tube. The tube was opened and the white needles were
scraped out, m.p. 175-178.
The above experiment was repeated with the tube completed submerged. The
starting material was recovered essentially unchanged, m.p. 175-177.
When the temperature was raised to 250*10, in a similar experiment, a
new product was obtained, m.p. 103-106: ^Cilj-0ri0rm 315 (shoulder), 307, 240.
max
This new solid was identified as 2, 3-diphenylindene (XXX) by its melting point
o 39
(lit. m.p. 107-108u) and ultraviolet spectrum.
Preparation of a mesityldiphenylindene
2, 3-Diphenylindenone (VI) (0. 025 g.) was dissolved in 3 ml. of ether and 3 ml.
of 0. 552M mesitylmagnesium bromide. The resulting red mixture was stirred


54
triphenylcyclopropene, (m.p. 107-110), 0.168g. of IV, 091 g. of V (total yield
of ketone 20%) and .076 g. (8%) of VII. A brown oil (.154 g.) remained from which
no more crystalline VII could be obtained but the ultraviolet spectrum indicated
that it was ca. 50% in VII. Trace amounts of 2,3-diphenylindenone (infrared
spectra 5. 89 p), m.p. 145-149, (lit. m.p. 151-152) were also isolated.
Stability of triphenylcyclopropene in lithium aluminum hydride-pyridine solution
Triphenylcyclopropene (0.134 g. ,0.5m moles) was dissolved in 5 ml. of
pyridine. Lithium aluminum hydride (0.65 m moles, 025 g.) was added in small
portions to the triphenylcyclopropene solution. The reaction mixture was stirred
3 hours at room temperature and poured into 20 ml. of 5% hydrochloric acid. A
white solid formed which was taken up in 20 ml. of ethermethylene chloride. The
layers were separated, the aqueous layer extracted with 15 ml. of ether and the
extract was combined with the origional layer, and the organic layer was washed
with 15 ml. of water. After drying over magnesium sulfate, the solvent was
separated, leaving a slightly yellow solid, which on recrystallization from ether-
hexane gave 0.108 g. (80%) of solid, m.p. 107-109, along with a yellow residue.
The reaction was repeated using 0. 268 g. (1.00 m moles) of triphenylcyclo
propene in 10 ml. of pyridine and 095 g. (2. 5 m moles) of lithium aluminum
hydride. The reaction was vigorous, turning from a yellow to a blue-green color.
After stirring 3 hours at room temperature the solution was worked up as above,
yielding 0. 246 g. (92%) of a slightly yellow material, m.p. 178-185. After one
recrystallization from benzene-hexane, this material melted at 182-183, turned
green at 225, and blue at temperatures above 240, ( lit. m.p. 179. 5-180. 50).1
This material was identified as the triphenylcyclopropene dimer (VIH) by its
melting point and infrared and ultraviolet spectra.


49
into the solution in small portions. The pH of a 60% aqueous diglyme solution was
5. The reaction was stirred for 3 hours, then worked up and chromatographed as
described for reaction I-A. The yield of triphenylcyclopropene (II) (m.p. 109-111)
was 77% (1.024 g.). Only trace amounts of benzilidene dexoxybenzoin (IV) were
isolated.
Reduction with sodium borohydride in strongly acidic solution
A suspension of triphenylcyclopropenyl bromide (I) (1.00 g., 2.72 m moles)
and sodium borohydride (0.11 g., 2.75 m moles) in 25 ml. of ethyl ether was
stirred for 5 minutes. The mixture was then poured fairly rapidly into a stirred
5% hydrochloric acid solution. The aqueous layer (pH 1) was separated and the
organic layer washed with 5% aqueous sodium bicarbonate and water. After drying
(MgSO^), the solution was concentrated to a viscous oil. Chromatography over
20 cc. of Merck alumina, with 5% benzene-hexane, as the elutant, gave 0.400 g.
(55%) of triphenylcyclopropene (II), m.p. 108-109.
Reduction with sodium borohydride in aqueous base
Triphenylcyclopropenyl bromide (1.00 g., 2. 72 m moles) was added in small
portions to a well-stirred 5% aqueous sodium bicarbonate solution (35 ml.) of
sodium borohydride (0.11 g., 2.76 m moles). After approximately half the cation (I)
had been added, ca. 5 ml. of ether was added to take up the oil which had formed
on the top of the water solution. The pH remained between 9 and 10 (measured with
pH paper) throughout the reaction. After all the cation (I) had been added, the
mixture was stirred for 15 minutes before the layers were separated. After wash
ing the organic layer with 5% hydrochloric acid and water and evaporating the solvent,


CHAPTER II
THE PREPARATION AND THERMAL REARRANGEMENT OF
TETRAARYLCYCLOPROPENES
INTRODUCTION
The literature provides many examples of rearrangements of small
ring compounds to yield systems involving less bond angle strain, but
relatively few rearrangements have been reported in the cyclopropene
series. The following are some pertinent examples of thermal
rearrangements recently observed in the latter series.
These involve, except in the last instance, fairly complex pathways and
19


38
By this mechanism then only two isomers should be formed from the acid catalyzed
rearrangement of a 3-aryl-l,2,3-triphenylcyclopropene. These two isomers
would be the 3-aryl-l,2-diphenylindene (C) and the triphenylindene substituted in
the 6 position (D).
The acid catalyzed rearrangement reactions provide a convenient source of
triarylindenes of known structure which promised to simplify the study of the
products from the thermal rearrangement of tetraarylcyclopropenes. With this
information at hand it was felt that a further study of the thermal products of
3-aryl-l, 2,3-triphenyl cyclopropenes should prove to be fruitful.
Tetraphenylcyclopropene had been found to rearrange to 1,2,3-triphenylindene
on heating at 250 for 2 hours. The reaction was shown to be actually a thermal
and not an acid catalyzed reaction at the surface of the glass by running the reaction
34
in a caustic coated vessel. The same product was obtained in the same length of
time under these conditions. The preliminary rate studies demonstrate the same
point although the rate does appear to be depressed slightly by washing the reaction
tubes with base instead of distilled water before the reaction is run. These same
rate studies indicate that the rate is first order.
Mechanistic considerations suggest three possible modes of thermal cleavage of
the cyclopropene ring. There could be a heterolytic cleavage giving rise to ionic
intermediates. This process appears unfavorable since either a vinyl carboniumion
or carbanion would have to be formed. The former is energetically unfavorable and
the latter is not consistent with results to be presented later. A concerted mechanism,
another possibility, would be energetically unfavorable since the bond in the cyclopropene
ring would have to cleave at the same time as the aromaticity of the phenyl ring was


66
evacuated to 0. 5 mm. pressure. After heating at 235-240 for 2 hours in a Wood's
metal bath and cooling to room temperature, a clear glass was obtained. The glass
was crystallized from methanol-chloroform to give 0.180 g. (90%) of 1,2,3-tri-
phenylindene (XX), m.p. 129-131 (lit. m.p. 131-132). An analytical sample,
m.p. 131-132, showed ultraviolet absorption (cyclohexane) at 306 (4.18), 241 mp
(4.38) and infrared absorption (KBr) at 1595, 1495, 1455, 1440, 911, 790, 773,
760 (w), 751, 739, 719 (w), and 695 (s) cm 1. A fluroescence spectra of the com
pound showed X (excitation) 355 and X (emission) 425 millimicrons,
max max
3-p-Aiisyl-l, 2,3-triphenylcyclopropene
3-p-Anisyl-l, 2,3-triphenylcyclopropene (0.350 g.) was heated at 200 for 4
hours in a sublimation apparatus under an argon atmosphere. A greenish, yellow
solid collected on the condenser. The contents of the reaction vessel were dissolved
in petroleumether and chromatographed on 30 g. of Woelm neutral alumina, activity
grade 1. The column was developed with 10% benzene-petroleum ether. Two bands
that were fluorescent under ultraviolet light were formed. The solvent sequence
used was 100 ml. mixtures of petroleum ether, benzene, ethyl ether, chloroform,
and methanol. The polarity of the solvents was gradually increased at such a rate
as to keep the band moving at a reasonable rate. Fractions of 40 ml. were taken.
The fractions were combined according to the fluctuation of the weight of material
per fraction. There was a total of 46 fractions taken. The first two yielded 0. 042 g.
of white solid, m.p. 138-145 after one recrystallization from ether-cyclohexane.
This material showed infrared absorption (KBr) at 3. 25, 3.4, 3.5, 6.21, 6.58,
6.71, 6.8, 6.9, 7.8, 8.08, 8.55, 9.68, 11.88, 12.35, 12.85, 13.08, 13.28, 13.58,
14.3, and 14. 5 microns and ultraviolet absorption (X ) in chloroform at 315 and
max
244 millimicrons.


ACKNOWLEDGMENTS
The author extends his sincere appreciation to his supervisory committee,
to the faculty of the Chemistry Department of the University of Florida, and to
his colleagues for their interest and friendly cooperation during the course of
this investigation. He takes particular pleasure in expressing gratitude to his
research director, Dr. M. A. Battiste, for his constant sincerity and willing
ness in providing the counsel, guidance, and encouragement necessary for the
realization of this project; and to Dr. J. F. Helling for his aid in the preparation
of this thesis.
The author gratefully acknowledges the financial support of a portion of this
research by the National Science Foundation.


42
isomerized almost quantitatively to the 2,3 isomer on heating at 250 for 2 hrs.
A mesityldiphenylindene prepared by a route which does not allow definite
assignment of structure showed the same ultraviolet absorption and retention time
on a thin layer chromatography plate as A. The method of preparation was:
XXIV
Preparation of the two isomers (XXIII) and (XXIV) is now under way.
The nuclear magnetic resonance spectra of the two isomers, A and B, were
very revealing. The isomer A showed resonance signals at 2.84 tau (aromatic)
4. 73 tau (indenyl proton), a doublet at 7.65 tau (methyl protons) and a singlet at
8.12 tau (methyl protons). The area ratio of the 7.65 tau peak to the 8.12 tau peak
was 2.1. The assignment of the 4. 75 tau peak was made by comparison to an m.m.r.
spectrum of 1,2,3-triphenylindene. (XX) The non-equivalence of the methyl protons
is consistent with the assigned structure. Models of the two possible isomers
(XXIII and (XXIV) show that the methyls on the mesityl group of XXIII are all in
essentially the same environment but in XXIV interaction of the mesityl or the
methyls with the 4-hydrogen on the fused benzene ring as well as with ortho hydro
gens on the 2-phenyl, restrict free rotation, producing a different magnetic environ
ment for the two ortho methyls, thus explaining their nonequivalence.
Sompound B showed three resonance signals due to methyl protons. These were
at 7.74 tau, 8.10 tau and 8.3 tau. The fact that the methyls are highly non-equiva
lent indicates that the mesityl nucleus has been destroyed. There was no signal
in the 4-5 tau region showing that :it was not a 1-hydroindene as suspected from the
chemical data. The methyl showing a resonance signal at 7. 74 tau is in an
environment similar to that of the ortho methyl in A. The 8.10 tau signal indicates
that there is one methyl on a phenyl similar to the para proton on A which is not
being affected by steric interactions. The third signal at 8.3 tau was in the same
region as the signal from the methyl protons (8.29 tau) of 1-methyl-1,2,3-triphenyl
indene (XXV). The structure, 1,4,6-trimethyl-l,2,3-triphenylindene (XXVI), is


18
pyridyl)aluminate. The ultraviolet spectrum of triphenylcyclopropenyl bromide in
pyridine, however, does suggest that there is some complexation of the pyridine
with the cation (I) since the spectrum shows an absorption pattern very similar to
a covalent triphenylcyclopropene.
From a synthetic point of view, it is evident from the results of this investi
gation that the reduction of triphenylcyclopropenyl cation (I) in protic solvents are
not as complex as first anticipated and that the reduction of the cation (I) in aqueous
solvents with sodium borohydride gives high yields of triphenylcyclopropene. The
reduction under these conditions require short reaction times and give a high
purity product after a short work up procedure.
Mechanistically the study has provided insight into the nature of sodium boro
hydride and lithium aluminum hydride in pyridine and has provided more informa
tion about lithium tetrakis(N-dihydropyridyl)aluminate (XI) and related species.


64
3-p-Anisyl-l, 2, 3-triphenylcyclopropene (XXI)
A. Triphenylcyclopropenyl bromide (0.300 g.; 0. 90 m moles) was dissolved
in a solution of 2 ml. of glacial acetic acid (2% acetic anhydride added), 1 ml. of
anisle and 3 drops of pyridine. The mixture was refluxed for 22 hours. After
cooling the mixture 10 ml. of water and 5 ml. of methylene chloride were added.
The layers were separated and the aqueous layer was extracted with two 10 ml.
portions of methylene chloride. After washing the organic layer three times with
10 ml. portions of 5% hydrochloric and 5% aqueous sodium bicarbonate, once with
water and drying over magnesium sulfate, the solvents were evaporated. The oily
residue was chromatographed over 15 g. of Merck alumina. The first 4 fractions
yielded a green, fluorescent oil which showed infrared absorption very similar to
that of the higher melting material from the thermal rearrangement. Crystals were
obtained which a broad range between 65 and 140 depending on the number of
recrystallizations.
B. Triphenylcyclopropenyl bromide (1. 00 g., 2. 90 m moles) was refluxed
with 2. 5 ml. of anisle, 6 ml. of acetic acid and 0. 6 ml. of pyridine for 20 hours.
The reaction was worked up as in A. The first two fractions from the chromato
graphy column gave 0.103 g. (10%) of crystals, m.p. 157-158, after trituration
with ether. The melting point and infrared spectrum were identical to those of
34
an authentic sample of 3-p-anisyl-l, 2,3-triphenylcyclopropene. The later
fractions yielded approximately 0.035 g. of 2, 3-diphenylindenone, m.p. 151-152.
(lit. m.p. 151-152) .
C. The above reaction was repeated using 1.00 g. (2. 90 m moles) of cation (I),
3.3 ml. of 10:1 acetic acid-pyridine, and 4. 0 ml. of anisle. The crude yield of the


TABLE 2
LITHIUM ALUMINUM HYDRIDE IN PYRIDINE
Age of
Solution (hr.)
Reaction
Time (hr.)
Equiv.
cation/LAH
Yield
Triphenyl Cyclopropene
Yield of
VII
Yield
of
Ketone
0
Solid added
3
1:1.66
71
8 1
20
0.5
3
1:1
8%
9.11
17
24
3
1:1
21
i1
o
21
72
3
1:1
4%
8.51
(151)
20


52
for 3 hours at room temperature. The reaction was quenched by adding 50 ml. of
methanol. The methanolic solution was then poured into 100 ml. of 5% hydrochloric
acid, and the aqueous layer extracted with four 50 ml. portions of ether. Extracts
were combined, washed with 50 ml. portions of 5% hydrochloric acid, 5% aqueous
sodium bicarbonate, and water, and dried over magnesium sulfate. After evapora
tion of solvent, 2.0 g. of yellow oil were obtained. The oil was chromatographed
over Merck alumina as described previously. The first component off the column
was triphenyleyclopropene (II), identified from its infrared spectrum and melting
point; m.p. 106-107 after one recrystallization from ether-hexane. The crude
yield of cyclopropene was 7% (. 090 g.). From the third through fifth fractions,
two crystalline compounds were separated by fractional crystallization. They were
found to be benzilidene desoxybenzoin (IV) and isobenzilidene desoxybenzoin (V) from
their melting points and infrared spectra. Benzilidene desoxybenzoin was isolated
in 16% yield (.322 g.) m.p. 96-101 (lit. m.p. 100-101), isobenzilidenedesoxy-
23
benzoin in 2% yield (.038 g.), m.p. 80-85 (lit. m.p. 87-88), and various mix
tures of the two, m.p. 93-95, in 2% yield (. 034 g.). The remaining ketonic oils
from these fractions showed an infrared spectrum very similar to the oils obtained
from hydrolysis of the cation and appeared to be predominantly a mixture of the
benzilidene desoxybenzoins. The benzene and 1:1 benzene-ether fraction yielded a
fourth compound (VII) (0.124 g.), m.p. 235-237 (green melt) after two recrystalli
zations from methylene chloride-hexane; V
acetonitrile
(log e) 333(45, 000), 316(52,600),
max
302(50,600), 288(84, 000) mlJt; X
KBr
= 5.49, 12.8, 13.21, 14.28, 14.54 microns.
max
Anal. Caled, for C H N; mw. 612, C, 92.15; H, 5.53; N, 2.14. Found: mw
47 33
620 (vapor phase osmometer); C, 92.2; H, 5.54; N, 2.29.


8
from this reaction was higher than that from the anhydrous reactions which had
been stirred for periods up to 24 hours, although the yield is very close to that
76
obtained by Battiste and Breslow (50%) from a similar reaction. The difference
in yields is probably due to variations in the quenching rates and the amount of
hydrogen bromide in the cation (I) which reacts with the borohydride before water
is added. Somewhat surprising is the fact that the cation (I) can compete so well
for the hydride (55% yield of cyclopropene II) in a strongly acidic solution. When
the pH of the solution was raised to 9, only a 13% yield of cyclopropene (II) was
obtained. Although the yield is low and the reaction was not run under equilibrium
conditions, this yield is significant in that the cation (I) is hydrolized very rapidly
0
under basic conditions. It indicates that sodium borohydride can compete with
hydroxide for the cation and is a fairly good trapping agent for a stable cation even in
basic solution.
The good yield of triphenylcyclopropene from sodium borohydride reduction in
aqueous diglyme gives considerable information about borohydride reductions of
0
cations. First, as in the ketone reductions, the protic solvent appears to aid in
7 10
hydride transfer, but most interesting is the fact that the reduction proceeds
almost to the exclusion of hydrolysis of the cation despite possible destruction of
the borohydride ion in 60% aqueous diglyme solution (pH 5).
Studies of the reduction of ketones by sodium borohydride in pyridine by Richie
19 26
and others indicate that these reductions are very sluggish under anhydrous
conditions, but that the reaction is fairly rapid if lithium borohydride is used or if
lithium ion is added to the sodium borohydride solutions. The effect of the lithium
ion was rationalized to be due to its ability to increase the polarity of the carbonyl
group or to increase the dissociation of the sodium borohydride in solution.19 The


43
most consistent with this data.
Me
Me
XXVI
The apparent mechanism for formation of (XXVI) involves a 1,2-methyl shift
as shown.
XXVI
XV
The attack of the intermediate species on the mesityl ring at a methyl position
suggests that the species is electron deficient and is most consistent with the
diradical representation.
The isolation of the 3-mesityindene (XXIV) isomer only would indicate on first
glance that 1-2 hydrogen migration is the preferred path to product, but the lack
of information about the thermal behavior of the mesitylindenes (XXIII) and (XXIV)
precludes any definite conclusions on this point.
With the above information at hand, work was begun on the third phase of the
problem, the determination of the kinetics of the thermal rearrangement of
tetraarylcyclopropenes. The nonvolatility of both the products and reactants and
the fact that both absorbed in the same region in the ultraviolet prevented the use
of vapor phase chromatography and ultraviolet spectroscopy for following the rate
of the reaction. The observation that 1,2,3-triphenylindene (XX) fluoresced strongly
in the visible region while tetraphenylcyclopropene did not, suggested that the
reaction could be followed by fluorimetry. Initial investigations show that the
rearrangement of tetraphenylcyclopropene can be followed by this techanique. Since


4
for reduction of compounds which were insoluble in ethers and could not be re
duced by sodium borohydride in protic solvents. It was found that ketones were
readily reduced by lithium aluminum hydride in pyridine on addition of the reducing
agent to a pyridine solution of the ketone. Studies of ketones which were readily
cleaved, those where the living anion would be fairly stable, indicated that the
lithium aluminum hydride solutions were basic enough to bring about the cleavage.1
The suggested scheme for the reduction and cleavage of a ketone in fresh lithium
14
aluminum hydride-pyridine solution is:
O O A1H C H N 9~
R-C C0, + A1H" R C C0X R C C0f + C H N:A1H
3 4 X 3 i 5 5 3
i
H
o
R-C-H + 0 -C-
o
(3)
To further demonstrate the basicity of lithium aluminum hydride in pyridine,
16
it was used to generate the trityl anion from triphenylmethane. The basicity of
the lithium aluminum hydride-pyridine solution decreases with age. This was
demonstrated by the addition of an aged solution of lithium aluminum hydride in
pyridine to a colored solution of trityl anions. The color due to the anion vanished.
This result suggests that the nature of the reducing species in the solution is
changing with age.
It was found that an aged (24 hours) solution of lithium aluminum hydride in
pyridine was a more selective reducing agent than a fresh solution in that it reduced
11 12
aryl ketones in preference to alkyl ketones. The mechanistic details of the


3
a good trapping agent for stable cations under solvolytic conditions and promoted
the further investigation of reductions of the cation (I) with sodium borohydride in
both aqueous and anhydrous, protic solvents.
Earlier work by H. C. Brown10 indicate that ketones were reduced under
anhydrous conditions by sodium borohydride in diglyme, but more recent work has
indicated the actual rate of reaction is very slow and that the major portion of the
7
isolated products arises in the aquoeus work up. The possibility that the solvated
cation (I) might be more easily reduced by sodium borohydride that ketones and the
fact that heterogeneous reductions of the cation (I) with lithium aluminum hydride
in ether give good yields of triphenylcyclopropene (II)1 suggested that the investiga
tion of sodium borohydride reductions of (I) in ethers and other aprotic solvents
might be fruitful, both from a mecahnistic as well as synthetic point of view.
Hydroboration of the cyclopropene double bond was anticipated to be a possible side
reaction in aprotic solvents.
Lithium aluminum hydride in ether is a more powerful reducing agent toward
14
most functional groups than sodium borohydride in protic solvents, but the
insolubility of the cation (I) in ether solvents impairs the ability of lithium aluminum
hydride to bring about the desired reductions. It was hoped that an aprotic solvent
could be found that would give a homogeneous lithium aluminum hydride-cation (I)
solution, increase the yield of cyclopropene (II) and cut down the required reaction
time. The successful reduction of ketones with lithium aluminum hydride in
11-15
pyridine by Lansbury suggested that pyridine would be an excellent solvent for
the desired reductions. Lansbury's work with the reduction of various ketones with
lithium aluminum hydride in pyridine resulted from the need for solvent systems


75
26. P. T. Lansbury and R. E. Macleay, J. Am. Chem. Soc., 87, 831 (1965).
27. C. F. Koelsch, J. Am. Chem. Soc. 56, 1337 (1934).
28. B. Fohlish and P. Burgle, Angew. Chem. Internat. Edit., 3, 699 (1964).
29. R. Breslow and M. A. Battiste, J. Am. Chem. Soc., 82, 3626 (1960).
30. E. S. Gould, Mechanism and Structure in Organic Chemistry, Holt, Rinehart
and Winston, Inc., New York, 1959, p. 207.
31. R. Breslow, B. Winter, and M. Battiste, J, Org. Chem., 24, 415 (1959).
32. R. Breslow and P. Wolf, unpublished work, cited in ref. 44.
33. A. Small and R. Breslow, unpublished work, cited in ref. 44.
34. M. A. Battiste and J. Todd, unpublished work.
35. Org. Syn., Coll. Vol. HI, p. 555.
36. C. F. Koelsch and R. V. White, J. Org. Chem., 6_, 602 (1941).
37. R. Breslow and M. A. Battiste, unpublished work, cited in ref. 44.
38. N. P. Neureiter, J. Org. Chem., 24, 2044 (1959).
39. Ref. 25, p. 391.
40. M. S. Kharasch and Otto Reinmuth, Grignard Reactions of Non-metallic
Substances, Prentice Hall, Inc., New York, 1954, p. 1051.
41. J. R. Dyer, Applications of Absorption Spectroscopy of Organic Compounds,
Prentice Hall, Inc., Englewood Cliffs, N. J., 1965, p. 20.
42. H. Gilman, N. J. Beaber and J. L. Jones, Rec. trav. chim., 48, 597 (1928).
43. G. Bergson, Acta. Chem. Scand., 17, 2691 (1963).
44. R. Breslow, Molecular Rearrangements, Part I, P. De Mayo, Ed., Interscience
Publishers, Inc., New York, N. Y., 1963, pp. 236-261.


5
reaction have not been determined, although nuclear magnetic resonance spectra
studies indicate that the dihydropyridyl aluminum complex (XI) is formed. This
species (XI) can be crystallized from concentrated, aged (24 hours) solutions.
The ratio of 1,2-dihydropyridyl to 1, 4-dihydropyridyl is approximately 1:1.
It was felt that the study of the reduction of a stable cation by this species
would be of interest in that it could possibly provide a convenient system for small
scale reductions of the cation (I) and give further insight into the nature of the
reagent.


65
desired cyclopropene (5. 5 micron peak in infrared) was 0. 231 g. (21%), obtained
in two fractions, m.p. 143-149, 157-158. Diphenylindenone, m.p. 150-152,
(0. 021 g.) was also obtained.
D. The best yield of the desired cyclopropene was obtained when the volume
of anisle was increased to 10 ml. and the volume of 10:1 acetic acid-pyridene was
decreased to 3 ml. Using these reaction conditions 0.520 g. (47%) of crude 3-p-
anisyl-1, 2, 3-triphenylcyclopropene was obtained from 1. 00 g. of cation (I). The
excess anisle was removed by heating the reaction mixture on a steam bath under
a stream of air.
3-(2, 4-Dimethoxyphenyl)-l, 2, 3-triphenylcyclopropene (XXII)
Triphenylcyclopropenyl bromide (1. 00 g.; 2. 90 m moles) was heated at 85-90
for 24 hours with 10 ml. of m-dimethoxybenzene and 3 ml. of 10:1 acetic acid
pyridine. After aqueous work up and evaporation of the low boiling solvents on a
rotary evaporator, the excess dimethoxybenzene was removed by heating the
reaction mixture on a steam bath under a stream of air. Chromatography over
25 g. of Merck alumina yielded 0. 243 g. (23%) of white crystals, m.p. 160-165,
which showed absorption at 5. 5 microns in the infrared. The melting point was
raised to 180-182 after 4 recrystallizations. The ultraviolet spectrum was
consistent with the cyclopropene structure showing absorption at: ^c^rolrm
331 (shoulder), 305, and 290.5 mp.
THERMAL REARRANGEMENTS
Tetraphenylcyclopropene
Tetraphenylcyclopropene (0.200 g.) was sealed in a tube which had been


40
being destroyed. The third possibility is a homolytic cleavage giving rise to
diradical intermediates. The homolytic cleavage would appear to be the
44
favored mode of cleavage since in analogous vinyl cyclopropane rearrangements
where there are groups present in the cyclopropane which will stabilize a radical
38
species the intermediate has been shown to be diradical in nature. The
mechanism involving homolytic cleavage with a diradical intermediate is presented
in Figure 5.
As can be seen from the mechanistic scheme (Figure 5), the thermal rearrange
ment could yield four different isomers of the triarylindene. Even if one of the
paths to a product were unfavorable, thermal equilibration of the indenes could still
give rise to the complex mixture which would be very difficult to separate. This
was found to be the case in attempts to isolate the thermal products from the anisyl
derivative (XXI). Preliminary attempts at separation by the use of thin layer
chromatography indicate that this technique may provide a means of separation
of the various mixtures of isomers.
From the mechanistic schemes for acid catalyzed rearrangement and thermal
rearrangements of tetraarylcyclopropenes the products from the acid catalyze
reaction should be the same as those resulting from 1,2 hydrogen shift in the
thermal rearrangement. The acid catalyzed rearrangement them provide a means
of determining those products which arise from 1,2 hydrogen shift and those
arising from 1,3 hydrogen shifts.
The goals of the study of thermal products are three in number:
(1) the determination of the preferred path of hydrogen migration, (2) the effect
of product stability on the preferred path of hydrogen migration and (3) the deter
mination of the physical properties of the products in preparation for a study of
the kinetics.
Although many difficulties were encountered in the attempt to study the thermal
and acid catalyzed products from the anisyl derivative (XXI), the results implied
that information could be obtained if a tetraarylcyclopropene were studied which
would yield fewer products. The mesityl derivative (XV) was the compound
selected since the ortho positions on one of the phenyls are blocked by methyls. This should


STUDIES IN THE ARYLCYCLOPROPENE
SERIES
By
ROBERT HOWARD GRUBBS
A THESIS PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF MASTER OF SCIENCE
UNIVERSITY OF FLORIDA
August, 1965


36
pyridine present. This material did not show a typical cyclopropene ultraviolet
spectrum, but exhibited a spectrum very similar to that of 1,2,3-triphenylindene
(XX). Crystals melting over a wide temperature range were obtained from the oils
after various solvent manipulations, suggesting a mixture of products. Since it was
known that tetraphenylcyclopropene rearranged to triphenylindene (XX) on heating
above 200, 3-p-anisyl-1,2,3-triphenyl cyclopropene (XXI) was rearranged to a
mixture of indenes by the same procedure. Attempts to completely separate the
mixture were futile. Chromatography yielded two crops of crystals that gave,
m.p. 138-145 and 145-148, respectively. The mixtures showed differentabsorp
tions in the infrared and ultraviolet and each appeared to be a mixture of at least
two different isomers. The infrared and ultraviolet spectra of the higher melting-
material were similar to those of the product form the electrophilic addition
reaction mentioned above. This further supported the supposition that the addition
reaction was yielding indenes and suggested two possible modes of formation of the
products.
29
The first possibility is similar to that suggested by Breslow and Battiste for
the rearrangement of diphenyl-(2,3-diphenylcyclopropenyl)- carbinol. This scheme
is
The other possibility is that the tetraarylcyclopropene is formed and then rearranged
under the reaction conditions. 3-p-Anisyl-l,2,3-triphenylcyclopropene was found
to be stable in acetic acid-pyridine solution in the absence of any proton catalysis.
The fact that the desired cyclopropene could be obtained when excess pyridine was
present suggested the possibility that the hydrogen bromide associated with the
bromide salt (I) was lowering the pH of the solution to such a point that the tetra
arylcyclopropene was undergoing an acid catalized rearrangement. When the


LIST OF TABLES
Table Page
1. Reduction with Sodium Borohydride 7
2. Lithium Aluminum Hydride in Pyridine 11
3. Reaction of Triphenylcyclopropenyl Bromide with Activated
Aromatic Nuclei 21
4. Grignard Preparation 26-27
5. Reaction of Triphenylcyclopropenyl Bromide with Anisle. . 29
xv


53
Reduction with fresh lithium aluminum hydride solution
Triphenylcyclopropenyl bromide (I) (1.74 g., 5.00 m moles) was dissolved
in 30 ml. of anhydrous pyridine and 20 ml. of a 30 minute aged lithium aluminum
hydride-pyridine solution (0. 25 M, 5. 0 m moles) was added dropwise with a syringe
over a 15 minute period. The reaction was stirred at room temperature for 3
hours and then worked up as in A above. Slow elution chromatography on alumina,
using the already given solvents, yielded 8% (0.110 g.) of triphenylcyclopropene
in several fractions melting 104-111. Benzilidene desoxybenzoin, 17% (0. 245 g.),
and 0.138 g. of VII, m.p. 210-230, were also obtained.
Reduction with aged lithium aluminum hydride solution
A solution of lithium aluminum hydride (20 ml., 0. 25 M in LAH) that had been
aged 24 hours was added over a 10 minute period to 1.74 g. (5. 00 m moles) of
triphenylcyclopropenyl bromide dissolved in 30 ml. of pyridine. The reaction was
stirred for 3 hours at room temperature and worked up as in run A. Slow elution
chromatography yielded 2.4% (. 0321 g.) of n, m.p. 104-108, 0. 205 g. of IV,
m.p. 100-101, 079 g. of a mixture of IV and V (total yield of ketone, 21%), and
10% (0.157 g.) of VII, m.p. 198-200 (235-236 after two recrystallizations from
methylene chloride-hexane).
Reduction with concentrated, aged lithium aluminum hydride solution
To 1. 04 g. (3. 00 m moles) of solid cation (I) was added dropwise 29 ml. of
0. 25 M (. 0075 moles) lithium aluminum hydride-pyridine solution that had been
aged for 3 days. The reaction mixture became red and viscous on addition of the
reducing agent. After stirring for three hours at room temperature, the reaction
was worked up as before. Chromatography over alumina gave; 034 g. (4%) of


CHAPTER III
EXPERIMENTAL
Physical Measurements
General
Melting points were taken on either a Fisher-Johns melting point block or a
Thomas-Hoover unimelt capillary melting point apparatus. All melting points
are uncorrected.
Carbon-hydrogen analyses were determined by Geller Microanalytical
Laboratories, Charleston, West Virginia.
Spectra
The infrared spectra were recorded on a Perkin-Elmer Infracord spectro
photometer. Solids were examined in potassium bromide pellets and oils were
examined on sodium chloride plates.
The ultraviolet spectra were recorded on a Cary 14 recording spectrometer.
The solvents smployed were suitable for use in spectrophotometry.
The nuclear magnetic resonance spectra were recorded in deuterochloroform,
with tetramethylsilane as the internal standard. A Varian DP-60 instrument was
used and the chemical shifts were calculated from measurements of sideband
spectra.
45


xml record header identifier oai:www.uflib.ufl.edu.ufdc:UF0008983200001datestamp 2009-02-09setSpec [UFDC_OAI_SET]metadata oai_dc:dc xmlns:oai_dc http:www.openarchives.orgOAI2.0oai_dc xmlns:dc http:purl.orgdcelements1.1 xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.openarchives.orgOAI2.0oai_dc.xsd dc:title Studies in the arylcyclopropene series.dc:creator Grubbs, Robert Howarddc:publisher Robert Howard Grubbsdc:date 1965dc:type Bookdc:identifier http://www.uflib.ufl.edu/ufdc/?b=UF00089832&v=00001000554225 (alephbibnum)13369029 (oclc)dc:source University of Floridadc:language English


DISCUSSION AND RESULTS
The reaction of aryl Grignard reagents with triphenylcyclopropenyl cation (I)
gave good yields of tetraarylcyclopropenes. The yields in most cases were only
slightly lower than 50%, based on the starting cation (I). Various reaction
conditions and work-up procedures were used depending on the nature of the
particular reaction.
The cation (I) is only slightly soluble in ether solvents so the reactions are
heterogeneous. To increase the homogeneity of the reaction, tetrahydrofuran
was used initially as the solvent because the cation (I) appeared to be more soluble
in this ether than in ethyl ether and its higher boiling point should permit a faster
reaction rate. In later reactions it was found that the reactions proceeded smoothly in
reflusing ethyl ether or in some cases in ethyl ether ar room temperature making
the use of tetrahydrofuran impractical since it is hard to dry. The reaction
reported for the formation of 1,3-(bistriphenylcyclopropenyl)benzene (XXII) was
run in tetrahydrofuran since the pra di-Grignard reagent could reportedly be
42
obtained only at the higher temperatures provided by this solvent, but later
preparations of the bistriphenylcyelopropenyl benzene (XXII) in ethyl ether gave
yields as high as those in tetrahydrofuran. These reactions gave nearly 45%
yields of products arising from the di-Grignard, (XIX) and (XXV), suggesting
that the meta di-Grignard is not as difficult to form as para di-Grignard.
The reactions appeared to be essentially complete after 24 hours but shorter
reaction periods have given lower yields. As an example, only 19% of the desired
cyclopropene was obtained in the reaction of p-chlorophenyl magnesium bromide
with cation (I) after 6 hr.
To check the dependence of the yield on concentration of Grignard reagent in
the more sterically hindered reactions, the preparation of 3-mesityl-1,2,3-
triphenylcyclopropene was run at two concentrations. The first reaction in which
the concentration of Grignard reagent was 0.27 M_, gave 18% of the desired
cyclopropene. The second run at a higher concentration, 1.0 M, gave a 52% yield.
The yield of the desired cyclopropene from the less hindered Grignard reagents appeared
24


72
mined by comparing the %T of these solutions to those of known percent concentra
tion. Plots of In (percent conversion) versus time were linear but the reaction
rate at this temperature was too rapid to give accurate rate constants although the
relative rates give a qualitative estimate of the effect of the various methods of
sample preparation of the rate of the reaction. Within 5 minutes, 97 percent
conversion was observed in the tubes washed with distilled water only, 60-70
percent conversion in the tubes that had been washed with base, and 20-30 percent
conversion in those tubes which had been washed with base and in which the sample
was in solution.
ACID CATALYZED REARRANGEMENTS
Tetraphenylcyclopropene (XIX)
Tetraphenylcyclopropene (0. 010 g.) was refluxed in a mixture of 2 ml. of
acetic acid, 0.4 ml. of benzene and 3 drops of concentrated hydrochloric acid for
24 hours. The reaction mixture was extracted with ether and the ether later was
washed with 5% aqueous sodium bicarbonate until there was no apparent acid-base
reaction in the separatory funnel. The ether solution was dried over magnesium
sulfate and filtered. After evaporation of the solvent, an oily residue (0. 012 g.)
was obtained which crystallized on trituration with cyclohexane-ether, m.p. 128-
130. The infrared and ultraviolet spectra of this compound was identical with
that of 1, 2, 3-triphenylindene obtained from the thermal rearrangement of tetra
phenylcyclopropene.
3-(p-Anisyl)-l, 2, 3-triphenylcyclopropene (XXI)
3-(p-Anisyl)-l, 2, 3-triphenylcyclopropene (0. 050 g.) was refluxed for 24 hours
in a mixture of 3 ml. of benzene, 2 ml. of acetic acid, and 2 drops of 48% aqueous


60
alumina using 5% benzene-petroleum ether as the eluent yielded three 50 ml.
fractions containing the desired compound. The first fraction yielded only trace
amounts, the second 1.25 g. of slightly yellow crystals, m.p. 139-145 (raised
to 151-152 after one recrystallization from methanol-chloroform), and the third,
two crops of total weight 0.114 g. yellow prisms, m.p. 150-151, m.p. 142-147,
respectively. The total yield of the desired cyclopropene was 52%.
3-p-Tolyl-l, 2, 3-triphenylcyclopropene (XVI)
To 0.36 g. (15 m moles) of magnesium turnings was added 2 ml. of a satu
rated solution of p-bromotoluene in ether. The reaction began immediately. The
addition of the remaining p-bromotoluene (2. 06 g.; 12.0 m moles) in 25 ml. of
ether was completed in 20 minutes. Most of the magnesium had disappeared after
the mixture had refluxed for 2. 5 hours. The solution was 0.41M in active Grignard.
The p-tolyl Grignard (18 ml. ,7.4m moles) was added dropwise to 1. 00 g.
(2.90 m moles) of triphenylcyclopropenyl bromide suspended in 15 ml. of tetra-
hydrofuran. After a 50 hour period of reflux with vigorous stirring, the reaction
was worked up by the filtration procedure described above. The first six fractions
off the chromatography column yielded 0. 529 g. (51%) of material, m.p. 127-131
after one recrystallization from methanol-chloroform. A sample prepared for
analysis, m.p. 134. 5-136, showed absorption in the ultraviolet region (milli
micron, log E) at 230 (4.53), 307 (4.32), 318 (4.30), 336 (4.26) and at 1830 cm'1
in the infrared.
Anal, caled, for C, 93.81; H, 6.19. Found: C, 93.58; H, 6.27.
28 22
The seventh and eighth fractions yielded 0.132 g. of benzylidene desoxybenzoin,
m.p. 96-100.


12
The nature of the lithium aluminium hydride solutions will be examined before
a mechanistic scheme is suggested for the formation of the observed products. The
reaction of lithium aluminium hydride with cyclopropene (II) in pyridine provides
insight into the nature of the nucleophilic species present.
It was found that triphenylcyclopropene (II) was dimerized by excess lithium
aluminium hydride if the solid hydride was added to a pyridine solution of the
cyclopropene (II) but no dimerization was observed if the lithium aluminium hydride
pyridine solution had been aged before the cyclopropene (II) was added. Breslow
and Dowd1 found that triphenylcyclopropene was readily dimerized in a strongly
basic solution such as sodamide in liquid ammonia and these authors suggested the
following mechanism which does not involve formation of the triphenylcyclopropenyl
anion.
H H
VII
II
The same scheme should also apply in the lithium aluminium hydride-pyridine reac
tion since a fresh lithium aluminium hydride-pyridine solution is less basic than
liquid ammonia-sodamide solution, as evidenced by the low yield of trityl anion
16
formed in this solution, and proton abstraction from triphenylcyclopropene would
be even less likely. These reactions then allow examination of the nucleophylic
species present in lithium aluminium hydride solution in a case where there are


62
3-(p-ChlorophenyI)-l, 2, 3-triphenylcyclopropene (XVIH)
A few crystals of 4-chlorobromobenzene were added to 0.132 g. (13. 0 m
moles) of magnesium turnings that were covered with ether. After gently warm
ing the flask to initiate the reaction, the remainder of the 4-chlorobromobenzene
(total of 2. 29 g.; 12. 0 m moles) dissolved in 20 ml. of ether was added over a
20 minute period. A solution that was 0. 59 M in active Grignard was obtained
after 2 hours of reflux. A portion (6 ml.; 3.54 m moles) of this solution was
added slowly to 1. 00 g. (2. 90 m moles) of triphenylcyclopropenyl bromide
suspended in 15 ml. of ether. The refraction was refluxed six hours and worked
up by the filtration method. The first fraction from the column gave a colorless
oil (0.203 g., 19%) which solidified on standing, m.p. 135-140. An analytical
sample of this material, m.p. 154. 5-155. 5, showed absorption at ^cyc^^exan
max
230 (4. 56), 304 (4.33), 317 (4.35), 334 mp (4.31) in the ultraviolet and at 1825 cm ^
in the infrared.
Anal, caled, for C H Cl: C, 85.06; H, 5.05. Found: C, 85.07, H, 5.09.
i i j. y
The next three fractions yielded 0.176 g. of benzylidene desoxybenzoin, m.p.
97-100. The sixth and seventh fractions yielded 0. 215 g. of yellow oil from which
traces of crystals, m.p. 165-195 were isolated.
Tetraphenylcyclopropene (XIX)
A 0.76M solution of phenyllithium was prepared by stirring 3.14 g. of bromo-
benzene in 15 ml. of ether with 0. 30 g. of lithium wire. A portion of this solution
(13 ml.; 10. 0 m moles) was added dropwise with a syringe to 1.74 g. (5. 00 m moles)
of triphenylcyclopropenyl bromide suspended in 10 ml. of ether. The resulting-
black solution was stirred for 7.5 hours at room temperature before 15 ml. of water


63
was added with cooling. After separation of the layers, the aqueous layer was
extracted with 20 ml. of ether. The organic layers were combined and washed
with 50 ml. of 0.1 N hydrochloric acid, two 25 ml. portions of aqueous sodium
bicarbonate, and 25 ml. of water. The yellow solid, obtained after evaporation
of the dried solution to dryness, was recrystallized from ether-hexane. Three
crops of material were obtained (0.798 g.; 47%), m.p., after two recrystalliza
tions, 177-179, (lit. m.p. 177-1780)8 \Cycl0lieXane 280 (4.5ff), 305 (4.33), 318
m.cLx
(4.35), 335 mp (4.31). This compound, identified as tetraphenylcyclopropene,
shows the typical disubstituted cyclopropene peak at 1825 cm 1 in the infrared.
34
3-(p-Anisyl)-l, 2, 3-triphenylcyclopropene (XXI)
To 0. 52 g. (22. 0 m moles) of magnesium and an iodine crystal was added
dropwise a solution of 2.7 ml. (21.0 m moles) of p-bromoanisole in 50 ml. of ether
and the mixture refluxed for 2 hours under argon. The resulting Grignard reagent
was added to a suspension of 4. 0 g. (9.35 m moles) of triphenylcyclopropenyl cation
(I) in 75 ml. of ether. After a 2 hour reflux period the reaction mixture was worked
up as in the previous reactions. Chromatography over_ca 40 g. of alumina yielded
two fractions which contained the desired cyclopropene (5. 5 microns in infrared).
The total yield of cyclopropene was 1.22 g. (40%), m.p. 160-161. The product
(XXI) showed major absorption in the infrared (KBr) at 3. 3, 3.4, 3.5, 5. 5, 6. 28,
6.6, 6.7, 6.8, 6.9, 7.8, 8.1, 8.55, 9.8, 12.05, 12.8, 13.2, 14.2, 14.7 microns
and in the ultraviolet (cyclohexane) at 230 (4.56), 300 (sh), 312 (4.36), 334 (420)
millimicrons.
Anal. Caled, for C H O: C, 89.8; H, 5.92. Found: C, 89.44; H, 6.02.
28 22


T
17
The high yields of ketone would be expected from a scheme of this nature
since the nitrogen would be the preferential point of attack on the dihydropyridyl
by an electrophile since it has a non-bonding pair of electrons. This part of the
scheme is also supported by the fact that the cation (I) is converted in pyridine
solution with an added secondary amine base to a species which appears to give a
ketone on hydrolysis.
It would be expected that the major product from reaction (9c) would be the
mono-alkylated product. The most obvious reason for not observing this product
is that it was obscured in the complex oils, although the ultraviolet spectra of the
oils remaining after crystallization showed only trace amounts of cyclopropenyl
material and this showed a spectrum identical to that of VII, but presumably the
mono-substituted pyridine would absorb in the same region. At any rate, it is
evident that the bis(triphenylcyclopropenyl)pyridine was the major cyclopropene
product formed in the reduction other than triphenylcyclopropene. More specu
lative explanations are that the bulky triphenylcyclopropenyl group once bonded to
the dihydropyridyl nucleus distorts the alumino complex to such an extent that the
mono-alkylated dihydropyridyl is easily alkylated again, or that the dihydropyridyl-
1, 2,3-triphenylcycloporpene is unstable under aqueous conditions and does not
aromatise as the bis-(triphenylcyclopropenyl)dihydropyridinyl does and yields a
complex mixture of products.
Another mechanism (10) was considered originally but was ruled out due to the
fact that sodium borohydride gave good yields of triphenylcyclopropene in pyridine.
Sodium borohydride would be expected to reduce the pyridine-cation complex
almost as readily as lithium aluminium hydride or lithium tetrakis(N-dihydro-


47
REDUCTION OF TRIPHENYLCYCLOPROPENYL BROMIDE WITH SODIUM
BOROHYDRIDE
The sodium borohydride used was commercial grade. Diglyme was dis
tilled from calcium hydride and stored over Dri-Na. The ethyl ether was com
mercial anhydrous ether. All reaction vessels were dried at 120 for at least
2 hours before use. All liquids and solutions were transferred by means of a
syringe.
Reduction with sodium borohydride in ethyl ether
Triphenylcyclopropenyl bromide (1.74 g., 5.00 m moles) was suspended in
50 ml. of anhydrous ether. Sodium borohydride (0.19 g., 5.0 m moles) was
added in small portions to the suspension. The reaction mixture, after having
been stirred 24 hours at room temperature (35), was poured slowly into 100 ml.
of 5% hydrochloric acid. The solid that separated was taken up in 50 ml. of ether,
and the layers were separated. The aqueous layer was extracted twice with 25 ml.
portions of ether. The extracts were combined with the original organic layer
and dried over magnesium sulfate. After filtration and evaporation of the sol
vent, a mixture of solid and oil was obtained which was dissolved in 10% benzene-
hexane, and chromatographed over alumina (40 cc in hexane) using the solvent
sequence: hexane, 10% benzene-hexane, 15% benzene-hexane,20% benzene-
hexane, 30% benzene-hexane, 50% benzene-hexane, 50% benzene-ether, ether,
methylene chloride, and methanol. ,


13
no acidic or reducible species involved.
The mechanism (6) suggested by Lansbury for the metalation of acidic hydro
carbons in pyridine lithium aluminum hydride solutions involves AlH^ as the
basic species which donates a hydride to the acidic proton with formation of a
14
molecule of hydrogen, the anion of the hydrocarbon, and A1H .
3
(6)
03C-H + H-AlHj + :KQ> 0f- + A Lansbury argues that pyridine is better able to solvate the alane (A1H ) than
O
ether. Thus enhancing the basicity of lithium aluminum hydride over that usually
observed in ether solvents. The basicity of the solution decreases with age to the
' 16
point that it appears able to donate a proton to a triaryl- or diarylmethide ion.
The suggested rational for this behavior is that the hydrogen remaining on the
aluminum after the complexation reaction has proceeded to the point that three
dihydropyridyls have added to the aluminum, becomes acidic, due to the electron
16
withdrawing effect of the dihydropyridyls, and protonates the anion. His results
show that there is an initial build up and subsequent reduction of the amount of
anion formed. Using fluorene as the organic acid, he found that after 1 hour 52%,
16
5 hours 71%, and 18 hours 48% metalation had occurred.
Lansbury's mechanistic scheme (6) cannot be applied to the dimerization of
triphenylcyclopropene. If the basic species were AlH^ as suggested, the attack
of this species on the double bond of triphenylcyclopropene (II) would result in
reduction. This being the case there must be another strongly basic species present.


30
The reaction was also investigated to determine the structure of the product,
m.p. 124-126, that had been isolated and reported from the reaction of anisle
with the cation (I). Since the compound was reported to show absorption at 5.5
microns in the infrared (characteristic of a 1,2-diphenylcyclopropene), it was
felt that the reported compound was possibly another tetraarylcyclopropene. The
high melting point of the o-anisyl derivative, m.p. 197-197.5, rules out this
most obvious possibility. The problem at this point appeared to be more complex
than had been anticipated and lead to further work which gave results that prompted
the investigation of the rearrangements of tetraarylcyclopropenes in greater detail
than had been originally planned but did not lead to a positive indentification of
the reported compound. These results will be discussed later.
The ultraviolet and infrared spectra of the tetraarylcyclopropenes prepared
inthis work showed larger substituent effects than origionally anticipated. The
spectral data is tabulated in Table IV. and reproductions of the ultraviolet
spectra shown in Figures 1-4. The characteristic 1,2-diphenylcyclopropene
-j 5
peak (1820 cm ) is present in all the infrared spectra on the tetraarylcyclo
propenes prepared. The position of the peak appears to be affected somewhat by
the substituents, but the spectra were recorded on a Perkin Elmer Infracord
infrared spectrometer and are not sufficiently accurate for substituent effect
correlations.
The ultraviolet spectra demonstrated the effect of substituents on the
wavelength of maximum absorption as well as the overall shape of the spectrum
for the 1,2-diphenylcyclopropenyl chromophore in each of the tetraarylcyclopro
penes. The effect can be attributed to either electronic or steric factors. The latter
are more apparent and are the most easily explained and will be discussed first.
The spectrum of tetraphenylcyclopropene (XIX) will be used as the reference in all
the discussion. The 3-mesityl derivative (XV) (Figure 1) exhibited a blue shift
in the long wavelength absorption band and a marked decrease in the absorption
intensity for this band. Models of the compound show that there is a definite
interaction between the ortho methyls on the mesityl ring and the phenyls conjugated
with the double bond of the cyclopropene ring. Less interaction results if the


67
The next nine fractions yielded crystals which were various mixtures (infra
red) of the first components and the later fractions. Fractions 12-14 yielded
0.085 g. of crystals, m.p. 145-148 after two recrystallizations from ether-
cyclohexane. These crystals showed infrared absorption at (KBr) 3.3, 3.4, 3.5,
6.25, 6.9, 7.41, 7.8, 7.9, 8.1 (w), 8.2, 9.7, 11.8, 12.06, 12.3, 12.75, 13.1 (w),
13.3, 13.4, 13.6, 14.3, and 14. 5 microns and ultraviolet absorption in chloroform
at 315 and 240 millimicrons.
The remaining fractions gave only traces of yellow or brown oils
3-Mesityl-1, 2, 3-triphenylcyclopropene
3-Mesityl-l, 2, 3-triphenylcyclopropene (0. 050 g.) was heated at 190-200 for
two hours in a small sublimation apparatus which had been flushed with argon. On
cooling a fluorescent oil formed on the condenser and sides of the flask. The oil
from the condenser (0.100 g., 10%) was triturated with methanol-chloroform to
yield white crystals (A), m.p. 179-181. The material from the bottom part of the
apparatus was crystallized from methanol-chloroform to give three crops of
crystals. The first crop (0. 010 g., 20%) was the same compound as that on the
condenser, m.p. 175-178. The next two crops yielded 0. 014 g. (28%) of crystals, (B)
131-132.
The higher melting material (A) showed absorption (X ) in the ultraviolet
ITLclX
in chloroform at 330 (sholder), 315 (sholder), 307, and 240 millimicrons and in the
infrared (KBr) at 3.3, 3.4, 3.5, 6.28, 6.7, 7.3, 9.35, 9.75, 11.0, 11.62, 11.9 (w),
12. 02 (w), 13.1, 13.55, 14.1 (w), 14.4, and 14. 55 microns. The lower melting
material (B) showed absorption (X ) in the ultraviolet in chloroform at 297.5
HL3X
and 240 millimicrons and in the infrared (KBr) at 3.3, 3.4, 3.5, 6.29, 6.7, 6.91,


35
phenyls are out of the plane of the cyclopropene ring. This would decrease the
conjugation which would, by analogy with cis and trans stilbenes,41 account for
a major part of the observed effect. Apparently, neither the steric effect nor the
electronic effect of one ortho substituent is very great. The o-anisyl derivative
(XVI) shows a spectrum which is only slightly different from that of tetraphenycy-
clopropene. This point is also supported by the fact that the 2,4-dimethoxy
derivative (XXII) shows a spectrum very similar to that of the p-anisyl derivative
(XXI).
As can be seen from Figures 1-4, the substituents in the para position effect
the wavelengths of maximum absorption very little, but have a pronounced effect
on the relative intensities of the three characteristic disubstituted cyclopropene
peaks. Although there have not been enough compounds prepared for substituent
effect-spectra shape correlation, there are enough differences in the spectra of the
p-chloro, p-methoxy, and p-methyl derivatives to allow a speculative examination
of the nature of the substituent effects. The chloro group (deactivates the ring
slightly) affects the spectra very little. The methoxy group (activates the ring
strongly) gives a pronounced change in the shape of the spectra. The methyl group
(activates the ring slightly) gives a different but very definite change in the relative
absorption intensities. It is apparent then that electron donating groups have the
largest effect on the shape of the absorption spectra. At first glance, the saturated
3-carbon would be expected to be able to transmit inductive effects very inefficiently.
The quite large effect of the substituents on the absorption spectra of the cyclopropene
chromophore is consistent with the methylene carbon having considerable sp^ character
since trigonal carbons are more efficient at transmitting inductive effects than a
30
tetrahedral carbon. The data collected to date indicates that with more examples
at hand the ultraviolet spectra of tetraarylcyclopropenes can possibly provide
further insight into the intriguing problem of the electronic structure of the cyclopro
pene ring.
In the investigation of the reaction of the cation (I) with anisle in acetic acid-
pyridine mixtures which was discussed earlier in this chapter, it was found that a
green fluorescent material was obtained when the reaction was run with insufficient


TABLE OF CONTENTS
Page
ACKNOWLEDGEMENTS ii
LIST OF TABLES iv
LIST OF FIGURES v
CHAPTER
I. REDUCTION OF TRIPHENYLCYCLOPROPENYL BROMIDE WITH
COMPLEX METAL HYDRIDES
Introduction 1
Discussion and Results 6
II. THE PREPARATION AND REARRANGEMENT OF TETRAARYLCYCLO-
PROPENES
Introduction 19
Discussion and Results 24
m. EXPERIMENTAL
Reduction of Triphenylcyclopropenyl Bromide with Sodium
Borohydride 47
Reduction of Triphenylcyclopropenyl Bromide in Pyridine
with Lithium Aluminum Hydride 51
Preparation of Tetraarylcyclopropenes 57
Thermal Rearrangements 65
Acid Catalyzed Rearrangements 72
BIBLIOGRAPHY 74
BIOGRAPHICAL SKETCH 76
in


56
room temperature and worked up as described in the previous reaction. Chroma
tography yielded 0. 766 g. of triphenylcyclopropene (57.2%, m.p. 108-111), 0. 013 g.
of 2, 3-diphenylindenone, m.p. 148-150, and 0.103 g. (7%) of benzilidene desoxy-
benzoin, m.p. 94-98.
Preparation of 1,3-bistriphenylcyclopropenyl benzene (IX)
The 1, 3-bis Grignard was prepared by adding 3. 8 ml. (. 020 moles) of 1, 3-di-
bromobenzene in 40 ml. of anhydrous tetrahydrofuran to 1. 08 g. of magnesium
turnings. An iodine crystal was used to initiate the reaction. A greyish suspen
sion had formed after 5 hours of refluxing under an argon atmosphere. The
resulting suspension was added dropwise to 13. 88 g. (. 040 moles) of triphenyl -
cyclopropenyl bromide (I) partially dissolved in 75 ml. of tetrahydrofuran. After
refluxing for 3 hours, the reaction was hydrolyzed with 6% aqueous ammonium
chloride solution. The layers were separated, the aqueous layer extracted with
50 ml. portions of ether and methylene chloride, the extracts combined, and the
organic solution dried over magnesium sulfate. Chromatography of the residue
after evaporation of the solvents yielded three major compounds. The first, from
the 25% benzene-hexane fraction, was tetraphenylcyclopropene (1.74 g., 25%),
m.p. 175-177 after three recrystallizations, identified by comparison to an au
thentic sample. The 50% benzene-hexane fraction yielded the desired product
(5. 5 p in infrared). This compound (4. 53 g. 17%) melted 231-233, after three
recrystallizations from benzene-hexane. The relatively pure material was percu-
lated through a short alumina column in 25% benzene-hexane and recrystallized
twice from benzene-hexane to give an analytical sample, m.p. 235-236 (green
melt above 270). This material gave a typical aryl cyclopropene absorption in the


10
in the nuclear magnetic resonance spectra. The tau values for these peaks were
1.43 (doublet), 2.31 (complex multiplet), and 2.75 tau (complex multiplet), with
the area ratio of the first peak to the second being ca. 9:2. Tetraphenylcyclopro-
pene shows a similar band at 2. 27 tau which corresponds to the ortho protons on
the phenyls conjugated with the double bond. Then with two triphenylcyclopropene
rings in the molecule as shown by the analytical data, the multiplet at 2. 31 tau in
VII should correspond to no less than 8 protons. The tau values for pyridine are
21
1.50 (ortho protons), 2.64 (para proton), and 3.10 (meta protons). The pertinent
22
coupling constants being J = .4 cps, J =1.9 cps. The position, area and
2-0 2-4
splitting (since a constant of 0.4 would not be seen) indicate that the 1.43 tau peak
of VII corresponds to two ortho protons on the pyridine ring. Possibly the added
area in the 2.31 tau peak is due to the one remaining proton on the pyridine nucleus.
This reduces the number of possible structures to two, the 3, 5- or the 3,4-bis-
(triphenylcyclopropenyl)pyridine. The similarity of the physical properties,
especially the melting point, and the spectra of 1,3-bis(triphenylcyclopropenyl)-
benzene to those of VII suggested that the correct structure was the 3,5- and not
the 3,4- structure. The following mechanistic considerations also support this
structure. The addition of the cation (I) to the pyridine nucleus is similar to the
38
addition of cation to other activated nuclei. The calculated electron densities for
pyridine, as well as experimental observations, indicate that electrophiles should
attack at the 3 and 5 positions. The electron density at these positions should be
even greater in a dihydropyridine. All of these considerations are consistent with
3, 5-bis(triphenylcyclopropenyl)pyridine (VII).


46
Procedure
Triphenylcyclopropenyl bromide (I) was prepared by the procedure of
3
Breslow and H. W. Chang, and had been dried in a drying pistol at 50 for at
least 24 hours prior to use. All anhydrous reactions were run under an argon
atmosphere.


TABLE 3
Ar=
REACTION OF TRIPHENYLCYCLOPROPENYL BROMIDE
WITH ACTIVATED AROMATIC NUCLEI
Br + Aril
p -dimethylaminophenyl
p-methoxyphenyl
2,4-dimethoxyphenyl
2 -methoxy naphthyl
anthron-10-yl
m.p. 124-126
m.p. 193
m.p. 123-125
m.p. 198-199.5
B. Fohlish and P. Biirgle, Angew. Chem. internat. edit. Vol. 3, 699 (1964)


41
decrease the number of thermal products possible to two.
On heating at 200 for two hours, 3-mesityl-1,2,3-triphenylcyclopropene (XV)
yielded two products which showed typical indene absorption in the ultraviolet and
were easily separated by solvent manipulations. Until structure assignment is made,
the higher melting isomer, m.p. 181-182, will be designated as (A) and the lower
melting isomer, m.p. 131-132, as (B). The ultraviolet spectra of the two compounds
first suggested that (A) was the l-mesityl-2,3-diphenylindene (XXIII) and (B) was the
3-mesityl-l-2-diphenylindene (XXIV) since
isomer (A) showed absorption at essentially the same wavelength as triphenylindene
and (B) absorbed at shorter wavelengths. It was felt that the bulky mesityl group
in the 3-position would decrease conjugation sterically and therefore (XXIV) would
show absorption at shorter wavelengths than 1,2,3-triphenylindene. Attempts to
43
isomerize B to A and A to B under basic conditions failed. The stability of B in
base under aerobic conditions suggested that it was not a 1-hydroindene as origi
nally thought. When 3-mesityl-l,2,3-triphenyl cyclopropene was heated at 85-90
under acidic conditions it was converted quantitatively to (A). Mechanistic consid-
32
erations and the fact that 1,2-diphenylindene was obtained exclusively from
triphenyl cyclopropene under the same conditions show that the isomer (A) is
3-mesityl-1,2-diphenylindene (XIV). This contention is supported by the results of
27
Koelch which indicate that arylindenes are not rearranged under acidic conditions.
The possibility that the product from acid catalyzed rearrangement is the thermo
dynamically favored product is ruled out by the fact that 1,2-diphenylindene is


TABLE 4
GRIGNARD PREPARATIONS
Ar
mp.
Yield
(cm.*"*)
U.V. Spectra (cyclohexane) max (log )
ja-Anisyl
(XXI)
160-161
40%
1810
230(4.56), 300 (sh), 312(4.36),
334(4.20)34
jo-Anisyl
(XVII)
197-197.5
44%
1810
286(4.26), 303(4.31), 317(4.34),
333(4.29)
Mesityl
(XV)
150-151
45%
1820
280(4.14), 300(4.20)
-Naphthyl (XXXI)
189-190
52%
292.5(4.35), 299 (sh), 305 (sh) 318(4.27),
333(sh) (ca. 4.2)34
Phenyl3
(XIX)
178-179
47%
1825
280(4.53), 305(4.33), 318(4.35),
3.35(4.31)
V
w*
b (XXV)
235-236
17%
1800
228, 302, 316(4.73), 333
to
O


16
is consistent with the results, but of course, it cannot be considered a proven
mechanism. The similarity of product distribution of those species arising from
routes other than reduction from the different solutions suggests that the equili
brium (7) plays a small role in these reactions and the reactivity of the cation
toward the pyridyl species is essentially the same whether the pyridyl is com-
plexed with the aluminum or is a free anion. This is reasonable on the basis of
26
the small steric requirement of XI found in reduction of ketones, and the
apparently small steric requirements for reduction of the cation as found in the
reduction of I with lithium tritertiarybutoxyaluminohydride.
The difference in yields of triphenylcyclopropene (II) from fresh (8%) and
aged (2-4%) solutions is probably due to reduction of the cation (I) by the lithium
aluminum hydride in the fresh solution before the reducing agent reacts with the
pyridine.


44
all the triaryl indenes prepared to date fluoresce, the method would appear to be
general. This technique, which has rarely been used in organic chemistry, could
prove to be very useful for following reactions involving a fluorescing species.
Spectrophotometric techniques involving emission measurements are more versatile
thank absorption techniques. In absorption measurements there is only one variable,
the wavelength of light passed through the sample. In emission measurements there
are two variables, the wavelength of excitation and the wavelength of emission.
In conclusion, the work to date has provided convenient procedures for the
preparation of tetraarylcyclopropenes, given further insight into cyclopropene
rearrangements, suggested some interesting studies of triarylindenes, and has
provided a method for the investigation of the kinetics of the rearrangements of
tetraaryl cyclopropenes.


61
3-o-Anisyl-l, 2, 3-triphenylcyclopropene (XVII)
o-Bromoanisole (approximately 0.1 ml.) was added to magnesium turnings
(0.34 g.; 18. 0 m moles) that were covered with 3 ml. of ether. The reaction was
initiated by heating with an infrared lamp. The remaining 22 ml. of ether and
o-bromoanisole (total of 1. 87 ml.; 15. 0 m moles) was added over a 15 minute
period. A 0. 725 M solution of active Grignard reagent was obtained after 1 hour
of refluxing and 2 hours of stirring at room temperature. The solution was red
black. Triphenylcyclopropenyl bromide (1. 00 g.; 2.9 m moles) was suspended
in 15 ml. of tetrahydrofuran and 16. 5 ml. of 0.725 M Grignard solution (12. 0 m
moles) was added dropwise over a 5 minute period. During the exothermic reac
tion the cation went into solution and a white solid appeared shortly after the con
clusion of the addition. The reaction was stirred 20. 5 hours at room temperature.
Dry carbon dioxide was bubbled into the reaction for 1 hour before the entire reac
tion mixture was poured onto 20 cc. of crushed Dry Ice. After the resulting
viscous mass had warmed to room temperature, 40 ml. of 5% hydrochloric acid
and 25 ml. of methylene chloride were added. The organic layer was separated
and extracted with 50 ml. and 25 ml. portions of 5% aqueous sodium bicarbonate.
After a final wash with water, the solution was dried over magnesium sulfate and
filtered. Chromatography of the oily solid remaining after evaporation of the solvents
give 0.424 g. (39%) of material which showed absorption at 1810 cm 1 in the infrared
m.p. 197-197.5. This compound showed an ultraviolet spectrum that was typical
of the diphenylcyclopropene chromophore, ^^^(log^E) 286 (4*26) 303 (4.31),
317 (4.34), 333 m^ (4.29).
Anal, caled, for C^H0: C, 89.80;H, 5.92. Found: C, 89.58; H, 5.81.
28 22


59
. cyclohexane
k,,,. 280 (4.14), 300 mp (4. 20)), the other data indicated that it was the
Ill cLX
desired compound. Anal, caled, for C H : C, 93.22; H, 6.78. Found:
30 20
C, 93.02; H, 6.87.
The fifth fraction yielded 0.150 g. of a yellow oil, which after crystallization
from methylenechloride-cyclohexane melted at 159-165, showed absorption in the
ultraviolet region at 315 and 238 mp and 1820 cm ^ in. the infrared. Fractions
six through nine yielded 0.324 g. of material, m.p. 95-97, which was identified
as benzylidene desoxybenzoin (IV) by its melting point (lit. m.p. 100-101) and
infrared spectrum.
The above reaction was repeated on a larger scale using higher concentrations
of reactants and a different work up procedure. A 1.805M solution of mesityl
Grignard was prepared by refluxing 5. 88 ml. (38.4 m moles) of mesityl bromide
with 1. 08 g. (45. 0 m moles) of magnesium in approximately 40 ml. of anhydrous
ether for 4. 5 hours. A portion of this solution (12 ml.; 22 m moles) was added
dropwise to a suspension of 3. 00 g. (8.75 m moles) of triphenylcyclopropene
bromide in 10 ml. of ether. The reaction was exothermic and all of the cation (I)
went into solution. The resulting solution was stirred at room temperature for
36 hours. The excess mesityl Grignard reagent was removed by pouring the
reaction mixture onto 50 cc. of powdered Dry Ice. After the viscous mass had
warmed to room temperature, 75 ml. of ether and 50 ml. of water were added
and the layers separated. The organic layer was washed with two 25 ml. portions
of aqueous sodium bicarbonate, 15 ml. of 0.1 N hydrochloric acid and 25 ml. of
water. After having been dried over magnesium sulfate, the solvent was evapo
rated to give 4. 0 g. of oily solid. Chromatography of the solid over 25 g. of


28
The yield of reduction products was much lower than expected from the
considerations given in the introduction. Only trace amounts of hexaphenylbenzene
(X) were isolated from the reaction with the meta-di-Grignard and only small
amounts of impure material which appeared to contain traces of the bistriphenyl-
cyclopropene (XIV) were obtained from the preparation of XV and XVIII. Consider
ing the steric requirements of the mesitylmagnesium bromide, it is surprising
that such a high yield of the cyclopropene derivative (XV) is obtained from the
reaction and such a small amount of reduction product is produced.
28
Shortly after the work on this problem had begun, Fohlish and Burgle
reported the preparation of tetraarylcyclopropenes by the reaction of triphenyl-
cyclopropenyl bromide (I) with activated aromatic nuclei. Their results are given
in Table III. The melting point for their p-anisyl derivative (XXI), 124-126, was
approximately 40 lower than that of 3-p-anisyl-l,2,3-triphenylcyclopropene, m.p.
160-126, prepared in our laboratory by the Grignard method.
The only reaction conditions given were that the cation (I) was heated with the
appropriate aromatic compound in 10:1 acetic acid-pyridine. In hopes of finding
this method useful for preparation of tetraarylcyclopropene derivatives not
obtainable by the Grignard method an investigation to determine the optimum
reaction conditions was undertaken. The results of this investigation are
summarized on Table V. Reasonable yields (based on starting cation (I) could be
realized only when a large excess of the aromatic reactant was used. This would
make the method practical only for micro scale preparations, expecially in cases
where the aromatic compound is expensive.


58
3-Mesityl-1, 2, 3-triphenylcyclopropene (XV)
To 0.360 g. (15.0 m moles) of magnesium covered with ether was added a
small portion (approximately 15 drops) of 1.96 ml. (13. 0 m moles) of mesityl
bromide. The reaction was initiated by heating and the addition of 10 pi. of methyl
iodide. The remaining mesityl bromide in ether (20 ml.) was added over a 20
minute period. After having been refluxed for 4 hours, the yellow solution was
0. 552 M in active Grignard. The Grignard solution was checked by adding 4 ml. of
the solution to Dry Ice. After appropriate work up, mesitoic acid was obtained,
3 5
m.p. 149-151, (Lit. m.p. 150-152). A portion (14 ml., 7.75 m moles) of the
mesityl Grignard prepared above was added dropwise over a 10 minute period to
1. 00 g. (2. 90 m moles) of triphenylcyclopropenyl bromide suspended in 15 ml. of
tetrahydrofuran. The reaction was exothermic and a large portion of the cation (I)
went into solution during the course of the addition. The reaction mixture was
stirred under reflux for 46 hours. The resulting suspension was filtered and the
solid washed twice with 35 ml. portions of methylene chloride. The filtrates were
combined and washed with 25 ml. portions of water, 5% hydrochloric acid, 5%
aqueous sodium bicarbonate, and water. Evaporation of the solvent after drying
over magnesium sulfate gave a viscous oil. The oil was placed on a chromato
graphy column in 10% benzene-hexane and the previously described elution scheme
followed. The excess mesitylene came off in the first fraction. The next three
fractions were combined and recrystallized from chloroform-methanol. This
material (0.203 g., 18%) showed absorption in the infrared at 1810 cm 1 and
melted 150-151 after three recrystallizations from methanol-chloroform. Although
this material did not show a typical arylcyclopropene spectrum (broad absorption at:


71
1. 00 x 10 ^ mg. /ml. The %T values were determined on an American Instrument
Company, Fluoro-Microphotometer. Interference filters with maximum trans
mittance at 440 and 365 millimicrons were used as the emission and excitation
filters. A 1 mg./ml. solution of tetraphenylcyclopropene showed only slight
absorbance and no fluorescence with this set of filters. Solutions containing
lmg./ml. of total material were prepared that were 0, 5, 10, 20, 30, 40, 50,
70, and 100% 1,2,3-triphenylindene with enough tetraphenylcyclopropene added to
give the constant 1 mg./ml. concentration. A plot of In %T versus percent tri
phenylindene gave a smooth curve. Sample tubes (8 mm. x 15-20 cm.) were
prepared by two methods. All the tubes were first boiled in detergent and then
rinsed 10 times with distilled water. The first set was dried after rinsing. The
second set was soaked in concentrated ammonium hydroxide for at least three
hours and then rinsed with 5% aqueous ammonium hydroxide. Samples of tetra
phenylcyclopropene (0. 0100 g.) were weighed into the first set of tubes and half of
the tubes prepared by the second method. The remaining tubes in the second set
were filled with 1. 000 ml. of a 10. 00 mg./ml. solution of tetraphenylcyclopropene
in decahydronaphthalene. The tubes were sealed after they had been evacuated
and/or degassed to 0.25 mm. pressure. The tubes were introduced into a constant
temperature bath (192.4 I 0. 2) and allowed to remain for a prescribed period of
time. The temperature bath was a chamber filled with equilibrated decahydro
naphthalene vapors at atmospheric pressure. After the prescribed time period
(time was kept to within 2 sec.), the tubes were opened and the contents transferred
quantitatively to 10 ml. volumetric flasks with 10 ml. of dichloroethane
percent conversion of tetraphenylcyclopropene to triphenylindene was then deter-


70
3 0
is high (lit. m.p. 96-98), resonance signals at 2. 8 tau (complex multiplet) and
8.29 tau showed that the compound was 1-methyl-1, 2,3-triphenylindene.
Stability of mesityl indene (A) to basic conditions
To a saturated solution of ethanolic potassium hydroxide was added 0. 050 g.
of (A). The solution had been purged of oxygen by bubbling a stream of argon
through the ethanol before the reactants were added. The reaction mixture was
refluxed for 3. 5 hours under an argon atmosphere before it was neutralized with
20% hydrochloric acid. The organic layer, after normal work up, yielded a
yellow oil which crystallized on trituration with methanol-chloroform, m.p. 178-
180. The ultraviolet spectra and melting point showed it to be starting material.
The same results were obtained when n-butyllithium was used as the base.
When the above reaction was repeated with ethanolic potassium hydroxide
solution which had not been deoxygenated, a new solid was obtained, m.p. 183-
186. This material showed ultraviolet absorption at 345 (shoulder), 329, 316,
309 (shoulder), and 240 millimicrons, and infrared absorption at 2.8 microns.
Stability of mesityl indene (B) to basic conditions
A solution of 0. 010 g. in 10 ml. saturated ethanolic potassium hydroxide was
heated at 85-90 for 23 hours. The solution had not been deoxygenated and was
protected from the atmosphere only by a drying tube of calcium chloride. The
ultraviolet spectrum of the product after aqueous work up was identical to that of
the starting material.
Preliminary rate determination
Standard solutions of 1, 2, 3-triphenylindene were prepared. It was found that
a plot of In %T versus concentration was linear from concentration of 1. 00 to


Fig. 4.-Ultraviolet spectrum of 5-(p-chlorophenyl)-l,2,5-triphenylcyclopro-
pene(XVIII).


Internet Distribution Consent Agreement
In reference to the following dissertation or thesis:
AUTHOR:
Grubbs, Robert
TITLE:
Studies in the arylcyclopropene series (record number: 554225)
PUBLICATION DATE:
1965
dfri //, (oi-ifUi
as copyright holder for the aforementioned
I,
dissertation or thesis, hereby grant specific and limited archive and distribution rights to the Board of
Trustees of the University of Florida and its agents. I authorize the University of Florida to digitize and
distribute the dissertation or thesis described above for nonprofit, educational purposes via the Internet
or successive technologies.
This is a non-exclusive grant of permissions for specific off-line and on-line uses for an indefinite
term. Off-line uses shall be limited to those specifically allowed by "Fair Use" as prescribed by the
terms of United States copyright legislation (cf, Title 17, U.S. Code) as well as to the maintenance and
preservation of a digital archive copy. Digitization allows the University of Florida to generate image-
and text-based versions as appropriate and to provide and enhance access using search software.
This grant ofrpepnii
ibits use of the digitized versions for commercial use or profit.
Signature of Copyright Holder
Grufh
Printed or Typed Name of Copyright Holder/Licensee
Personal information blurred
Date of Signature
Please print, sign and return to:
Cathleen Martvniak
UF Dissertation Project
Preservation Department
University of Florida Libraries
P.O. Box 117007
Gainesville, FL 32611-7007


48
The yield of triphenylcyclopropene (II) was 23.6% (0.316 g.), m.p. 108-105,
(lit. m.p. 112-1130)1, of benzylidene desoxybenzoin (IV) 0. 035 g., m.p. 96-100,
0
(lit. m.p. 100-101) of isobenzylidene desoxybenzoin (V) 0. 087 g., m.p. 85-88,
23
(lit. m.p. 89-90) and 0. 239 g. of a mixture of IV and V, m.p. 81-86. The
total yield of crystalline ketone was 21% with various ketonic oils (infrared) re
maining. An alcohol (0. 016 g.), m.p. 227-230, was isolated but not identified.
The compound exhibited infrared abosrptions at 2.87 microns.
Reduction with sodium borohydride in anhydrous diglyme
Triphenylcyclopropenyl bromide (I) (1.74 g., 5. 00 m moles) was suspended in
50 ml. of anhydrous diglyme. Sodium borohydride (0.227 g., 6. 00 m moles) was
added in small portions. The yellow slurry immediately became white. The re
action was stirred at room temperature for 3. 25 hours before it was poured into
75 ml. of water. The aqueous solution was extracted 3 times with 50 ml. portions
of ether. The extracts were combined and washed with 50 ml. portions of 5%
aqueous sodium bicarbonate and water before drying over magnesium sulfate. The
organic solution was concentrated to a thick oil by heating on a water bath under
reduced pressure. The oil was dissolved in 10% benzenehexane, and chromato
graphed as described previously. Triphenylcyclopropene (II) was obtained in 24%
yield (0.315 g.), m.p. 109-111. The ketonic oils yielded no crystals, but a trace
of solid alcohol (2. 87 p) was isolated from the 1:1 benzene-ether fraction.
Reduction with sodium borohydride in 60% aqueous diglyme
Triphenylcyclopropenyl bromide (I) (2.74 g., 5. 00 m moles) was suspended
in 30 ml. of diglyme. Sufficient water (20 ml.) was added to give the desired con
centration before 0. 227 g. (6. 00 m moles) of sodium borohydride was introduced


CHAPTER I
REDUCTION OF TRIPHENYLCYCLOPROPENYL
BROMIDE WITH COMPLEX METAL HYDRIDES
INTRODUCTION
The recent interest in non-benzenoid aromatic compounds has led to the prepa-
3 6
ration of various stable carbocyclic cations in the cyclopropene series. The
need for a convenient route to the parent hydrocarbons from these cations is
obvious. Not only are the parent hydrocarbons needed for comparison purposes
but they are inherently interesting themselves. The recent report of Breslow on
the preparation of triphenylcyclopropene from triphenylcyclopropenyl bromide by
1 7
lithium aluminum hydride reduction and the work by Brown and Bell on the
sodium borohydride reduction of carbonium ions formed under solvolytic condi
tions indicate that the reduction of a stable carbonium ion with complex metal
hydrides could provide useful synthetic information as well as further insight into
the chemistry of these reducing agents. The reduction of stable cations employing
complex metal hydrides allows observation of the products arising directly from
hydride attack without the complications inherent in reductions involving carbonyl
groups.
The triphenylcyclopropenyl cation (I) was chosen for a study of this nature
since many of its reaction products have been identified, and the majority of these
1 6
are solids. Triphenylcyclopropene (II) has also been found to be an excellent
dienophile, and has found uses in this role as an intermediate in the preparation
9 8
of heptaphenyltropylidene and various bicyclic systems. This added utility
demands a convenient high yield route to this species.
1


Fig. 1.-Ultraviolet spectra of (XV), (XIX), and (XXI).


TABLE 1
REDUCTIONS WITH SODIUM BOROHYDRIDE
Solvent
Reaction
Time
Ratio of Equiv.
cation/BH,,
Yield of
Triphenyl cy cl opropene
Yield of
Ketone
Ether
24 hr.
4
1:1
24
21
Anhydrous
diglyme
3 hr.
1:1.2
26
trace
60% Aqueous
diglyme
3 hr.
1:1.2
77
trace
Water pH 1
5 min.
1:1
55

Water pH 9
15 min
1:1
13

Pyridine
3 hr.
1:1
70
trace


20
only formalized mechanisms have been proposed for each. As an example, the
mechanism suggested for the dimerization of triphenylcyclopropene and subse
quent rearrangement of the dimer is given below.
The complexity of the thermal rearrangements of these cyclopropene
derivatives prevents any conclusion to be drawn as to the mode of cyclopropene
ring cleavage and subsequent reactions of the ring fragment (s). Any attempt to
study the mechanism of such thermal rearrangements should first concern itself
with the initial ring cleavage step and the nature of the species produced. The
ideal cyclopropene derivative for a study of this nature would be one which yields
only primary products from ring cleavage.
Tetraarylcyclopropenes have no labile hydrogen, exhibit moderate thermal
stability, and have the essential features of the previous examples of cyclopropene
thermal rearrangements.
Tetraphenylcyclopropene, prepared by the reaction of phenylmagnesium
bromide with triphenyleyclopropenyl bromide, was initially investigated as a
8
possible deinophile in high temperature Diels-Alder reactions, but it was found
that at high temperatures the cyclopropene rearranged to another hydrocarbon.
Further investigation with the above mechanistic considerations in mind showed
that tetraphenylcyclopropene had rearranged after heating at 250 for
two hours to 1,2,3-triphenylindene. The reaction was apparently
XIX
XX


Fig. 3.-Ultraviolet spectrum of 3-(o-anisyl)-l,2,3-triphenylcyclopropene(XVII).
CO
CO


15
faster rate of anion formation but has little effect on the maximum concentration of
anion produced).16
An aged solution of XI would equilibrate with XII to reduce the basicity of the solu
tion due to the reasons given above. This is in agreement with the observed quench
ing of a solution of trityl anions by aged lithium aluminum hydride-pyridine solu-
16
tions.
The above considerations indicate that there is a dihydropyridyl alumino com
plex, either XI or XII, present in both aged and fresh lithium aluminum hydride
pyridine solutions. There are three points of attack open to the cation (I) on the
dihydropyridyl nucleus: (1) the 2 or 4 position to remove hydride and give cyclo-
propene, (2) the nitrogen to effectively trap out the cation and yield ketone on
hydrolysis, (3) or on the 3 and 5 carbons to yield VII. This scheme