• TABLE OF CONTENTS
HIDE
 Title Page
 Dedication
 Acknowledgement
 Table of Contents
 List of Figures
 Abstract
 Introduction
 Results
 Discussion
 Thermal generation of diradica...
 Experimental
 Appendix: NMR and IR spectra of...
 References
 Biographical sketch






Group Title: Cycloaddition reactions of 1,1-difluoroallene /
Title: Cycloaddition reactions of 1,1-difluoroallene
CITATION PDF VIEWER THUMBNAILS PAGE IMAGE ZOOMABLE
Full Citation
STANDARD VIEW MARC VIEW
Permanent Link: http://ufdc.ufl.edu/UF00097477/00001
 Material Information
Title: Cycloaddition reactions of 1,1-difluoroallene
Physical Description: ix, 152 leaves : ill. ; 28 cm.
Language: English
Creator: Piedrahita, Carlos A., 1950-
Publication Date: 1978
Copyright Date: 1978
 Subjects
Subject: Difluoroallene   ( lcsh )
Chemistry thesis Ph. D
Dissertations, Academic -- Chemistry -- UF
Genre: bibliography   ( marcgt )
non-fiction   ( marcgt )
 Notes
Thesis: Thesis--University of Florida.
Bibliography: Bibliography: leaves 146-151.
Additional Physical Form: Also available on World Wide Web
General Note: Typescript.
General Note: Vita.
Statement of Responsibility: by Carlos A. Piedrahita.
 Record Information
Bibliographic ID: UF00097477
Volume ID: VID00001
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: alephbibnum - 000065653
oclc - 04361977
notis - AAH0867

Downloads

This item has the following downloads:

PDF ( 5 MBs ) ( PDF )


Table of Contents
    Title Page
        Page i
        Page i-a
    Dedication
        Page ii
    Acknowledgement
        Page iii
    Table of Contents
        Page iv
    List of Figures
        Page v
        Page vi
    Abstract
        Page vii
        Page viii
        Page ix
    Introduction
        Page 1
        Page 2
        Page 3
        Page 4
        Page 5
        Page 6
        Page 7
        Page 8
        Page 9
        Page 10
        Page 11
        Page 12
        Page 13
        Page 14
        Page 15
        Page 16
        Page 17
        Page 18
        Page 19
        Page 20
        Page 21
        Page 22
        Page 23
        Page 24
        Page 25
        Page 26
        Page 27
        Page 28
        Page 29
        Page 30
        Page 31
    Results
        Page 32
        Page 33
        Page 34
        Page 35
        Page 36
        Page 37
        Page 38
        Page 39
        Page 40
        Page 41
        Page 42
        Page 43
    Discussion
        Page 44
        Page 45
        Page 46
        Page 47
        Page 48
        Page 49
        Page 50
        Page 51
        Page 52
        Page 53
        Page 54
        Page 55
        Page 56
        Page 57
        Page 58
        Page 59
        Page 60
        Page 61
        Page 62
        Page 63
        Page 64
        Page 65
        Page 66
        Page 67
        Page 68
        Page 69
        Page 70
        Page 71
        Page 72
        Page 73
        Page 74
    Thermal generation of diradicals
        Page 75
        Page 76
        Page 77
        Page 78
        Page 79
    Experimental
        Page 80
        Page 81
        Page 82
        Page 83
        Page 84
        Page 85
        Page 86
        Page 87
        Page 88
        Page 89
        Page 90
        Page 91
        Page 92
        Page 93
        Page 94
        Page 95
        Page 96
        Page 97
        Page 98
        Page 99
        Page 100
        Page 101
        Page 102
        Page 103
        Page 104
        Page 105
        Page 106
        Page 107
        Page 108
        Page 109
        Page 110
        Page 111
        Page 112
        Page 113
        Page 114
        Page 115
        Page 116
        Page 117
    Appendix: NMR and IR spectra of relevant compounds
        Page 118
        Page 119
        Page 120
        Page 121
        Page 122
        Page 123
        Page 124
        Page 125
        Page 126
        Page 127
        Page 128
        Page 129
        Page 130
        Page 131
        Page 132
        Page 133
        Page 134
        Page 135
        Page 136
        Page 137
        Page 138
        Page 139
        Page 140
        Page 141
        Page 142
        Page 143
        Page 144
        Page 145
    References
        Page 146
        Page 147
        Page 148
        Page 149
        Page 150
        Page 151
    Biographical sketch
        Page 152
        Page 153
        Page 154
Full Text










CYCLOADDITION REACTIONS OF 1,1-DIFLUOROALLENE


By

CARLOS A. PIEDRAHITA

















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








UNIVERSITY OF FLORIDA


1978































To Mariann












ACKNOWLEDGEMENTS

I would like to thank my research director, Dr. W.R.

Dolbier, Jr., for the support and encouragement he has

provided during the past five years. His enthusiasm for

chemistry has helped me through the often wearisome process

of research. I am also grateful to Dr. K.N. Houk of

Louisiana State University and to the members of his

research group for their helpful and ea -r collaboration.

Thanks are also due to my fellow graduate students, who

are too numerous to mention, for their friendship and

moral support. My thanks also go to my parents and

family, whose love and sacrifices have helped me to

attain this goal. A debt of gratitude is also owed to my

"other" parents in Pennsylvania for their concern and

support. Last, but not least, to my wife Mariann, whose

love and understanding have made the past four years the

happiest of my life, thank you for your help. Your bright

smile and understanding words brightened many days.


iii













TABLE OF CONTENTS


Page


ACKNOWLEDGEMENTS

LIST OF FIGURES

ABSTRACT

CHAPTERS:

I. INTRODUCTION


RESULTS


III. DISCUSSION

IV. THERMAL GENERATION OF DIRADICALS

V. EXPERIMENTAL

APPENDIX-NMR AND IRSPECTRA OF RELEVANT COMPOUNDS

REFERENCES

BIOGRAPHICAL SKETCH


iii

v

vii



1


44

75

80

119

146

152












LIST OF FIGURES


Figure Page

1 Correlation diagram for the formation
of cyclobutane from two molecules of
ethylene by an (s,s) mode of addition. 6

2 Frontier molecular orbitals of isoprene
acrylonitrile, and 1,1-dicyanoethylene.
From ref. 49. 23

3 Totally diradical mechanism for difluoro-
allene cycloadditions. 47

4 Concerted [2+4] and diradical [2+2]
mechanism for difluoroallene cycloadditions 57

5 Frontier orbitals of difluoroallene, furan,
cyclopentadiene and dienes 30a, 30b, and
30c. 70

6 Mechanism for the pyrolysis of 38. 78

7 Calibration curve for the reaction of 1,1-
difluoroallene and 1,3-butadiene. 89

8 Calibration curve for the reaction of 1,1-
difluoroallene and 2-trimethylsilyloxy-
1,3-butadiene. 102

9 1H nmr spectrum of 19. 119
19
10 1F nmr spectrum of 19. 120

11 Photoelectron spectrum of 19. 121

12 1H nmr spectrum of 24. 122

13 H nmr spectrum of 26. 123

14 1H nmr spectrum of 27. 124

15 H nmr spectrum of 28. 125

16 1H nmr spectrum of 29. 126

17 H nmr spectrum of 31a. 127






Figure

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35


nmr spectrum of 33a.

nmr spectrum of 31b and 32b.

nmr spectrum of 33b.

nmr spectrum of 34b.

nmr spectrum of 31c and 32c.

nmr spectrum of 33c.

nmr spectrum of 38.

nmr spectrum of 39.

spectrum of 19.

spectrum of 24.

spectrum of 28.

spectrum of 29.

spectrum of 31a.

spectrum of 31b and 32b.

spectrum of 33b.

spectrum of 33c.

spectrum of 38.

spectrum of 39.


Page

128

129

130

131

132

133

134

135

136

137

138

139

140

141

142

143

144

145











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


CYCLOADDITION REACTIONS OF 1,1-DIFLUOROALLENE

By

Carlos A. Piedrahita

June 1978

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

Cycloaddition reactions were carried out between 1,1-

difluoroallene and six dienes. The reactions with cyclo-

pentadiene, furan, and hexachlorocyclopentadiene formed only

[2+4] adducts. The reactions with 1,3-butadiene, isoprene,

and 2-trimethylsilyloxy-l,3-butadiene resulted in formation

of both [2+2] and [2+4] adducts. In all six cases, [2+4]

cycloaddition occurred exclusively at the non-fluorinated

allene n bond. In contrast, the [2+2] cycloaddition resulted

in predominant addition to the fluorinated allene n bond. The

[2+4] adducts were characterized by a lack of orientational

preference with respect to the diene, as shown in the reac-

tions with isoprene and 2-trimethylsilyloxy-l,3-butadiene.

However, the [2+2] adducts showed a strong orientational

preference with respect to the diene.

These results could be viewed in terms of two distinct

mechanistic alternatives. The first would envision all of


vii






the cycloaddition products arising from diradicals by non-

concerted mechanisms, whereas the second would entail a

dichotomy of mechanism involving competition between a

concerted [2+4] reaction and a non-concerted, diradical

[2+2] reaction.

Several factors combine to make the first alternative

unlikely. The formation of significant amounts of [2+2]

and [2+4] adducts, despite the unfavorable cisoid-transoid

diene equilibrium, as well as the decrease in the [2+4]/[2+2]

product ratio when the cycloaddition temperature is in-

creased, suggest reaction paths with different activation

parameters. This would be inconsistent with the totally

diradical mechanism. Finally, the observed contrasting

regiospecificities of the [2+2] and [2+4] cycloadditions,

with respect to both the location of the CF2 group and the

orientational preference of the diene moiety, along with

the small steric effect of fluorine make the totally

diradical mechanism an improbable and unattractive alter-

native.

The second alternative provided the simplest and most

reasonable explanation of the results. The [2+2] cyclo-

addition results were totally consistent with a reaction

involving diradical intermediates. The Frontier Molecular

Orbital Theory of Cycloadditions effectively rationalized

the [2+4] results in terms of a concerted mechanism. Ab

initio (STO-3G) calculations carried out on difluoroallene

indicated that the C2-C3 7 and 7* orbitals were lowered in


viii







and the C1-C2 r and T* orbitals raised in energy, relative

to those of allene. The inductive effect of fluorine lower-

ed all the orbital energies. Thus the HOMO is the C -C2 T

orbital and the LUMO the C2-C3 T* orbital. A photoelectron

spectrum confirmed the calculations and provided an experi-

mental value for the HOMO energy.

The LUMO of difluoroallene interacted with the diene

HOMOs to give a stronger interaction than the diene-LUMO-

difluoroallene-HOMO interaction. The calculated LUMO

coefficients for C2 and C3 of difluoroallene were nearly

identical and accounted for the diene moiety's lack of

regioselectivity.

The cycloaddition reactions of 1,1-difluoroallene have

provided a unique, easily perceivable mechanistic probe,

through which concerted and non-concerted mechanisms may

be distinguished by simple product identification. The

particularly reactive nature of difluoroallene should make

it useful in the study of other cycloaddition processes.












CHAPTER I
INTRODUCTION

Cycloadditions General

Cycloaddition reactions have intrigued organic chemists

for nearly fifty years. Perhaps the best known and most

general thermal cycloaddition is the [2+4] or Diels-Alder

reaction.1, 2 It involves reaction between conjugated

dienes, such as butadiene, cyclopentadiene, and their

derivatives, and other unsaturated molecules (known as

dienophiles), such as maleic anhydride and acrylonitrile,

to form six-membered rings. Although the six atoms in-

volved in forming the ring are usually all carbon atoms,

that need not be so. Examples are known in which atoms




/CN CN


I + I ---- ^ [I I





other than carbon, mainly oxygen and nitrogen, are part

of the reacting system of either the diene, the dienophile,

or both.3

The reaction also exhibits extraordinary degrees of

stereospecificity,4 stereoselectivity,5 and regioselectivity


-1-







as illustrated below. That, along with the reaction's

generality, makes it an important synthetic tool in organic

chemistry.

The first modern report of this reaction was that of

Diels and Alder in 1928. Since then it has been extensive-
ly reviewed in the literature.8-12
ly reviewed in the literature.


Stereospecificity







Stereoselectivity







Regioselectivity


Es


Es


+ H0
+H 0

0 0O



Ph Ph CN


+ CN CN
CN


80 : 20



Somewhat less common than [2+4] cycloadditions are

[2+2] thermal cycloadditions of alkenes to form cyclo-

butanes.13 This reaction is not as general as the [2+4]

reaction. Ethylene and other simple alkenes reluctantly

undergo dimerization to form four-membered rings.14


0'







However, fluoroalkenes have been found to undergo this

reaction with great ease.5 Tetrafluoroethylene, for

example, can smoothly dimerize to perfluorocyclobutane at

200c*with the reverse reaction occurring above 5000.16, 17



C2 C2 2000
II + fl _.. ,
11 1+ -5000
CF2 CF2



Cycloaddition also takes place when the fluoroalkene

is only one of the reactive components and an olefinic

hydrocarbon is the other; as in the reaction of propene
18
and tetrafluoroethylene.18 In addition to standard olefins,

H
CF2 CHI 2000 F4 H
2 2 4 -
Il + I CH
CF C 3
2 C\ H 3
H CH3

fluoroalkenes also undergo reaction with certain conjugated

dienes, such as isoprene and butadiene, to form cyclo-
19
butanes. Although in these cases the possibility of

cyclohexene formation via [2+4] cycloaddition exists, the

[2+2] adducts are found to be the dominant products. These

curious results will be explored in detail later.

Ketenes and allenes also undergo [2+2] cycloaddition
20, 21
either by dimerization,2 21 or as one of the reactive
22, 23
components in the cycloaddition.2 23 Other kinds of

cycloadditions include 1,3-dipolar cycloadditions24 which



*The symbol "o" is used throughout to indicate Celcius
temperature (OC).







form five-membered ring heterocycles and cycloadditions of

allyl cations to dienes to give seven-membered rings.25

Obviously, a broad spectrum of addends can be used and an

assortment of ring sizes may be formed by thermal cyclo-
26
additions. Photochemical cycloadditions provide even
27
more possibilities.2

Cycloaddition Mechahisms

Cycloaddition reactions of the [2+2] and [2+4] sort

may be viewed in terms of two possible mechanisms. The

first is the concerted mechanism in which both new a bonds

are formed simultaneously. It is experimentally character-

ized by stereospecificity and a lack of solvent effect on

rate. The second is the stepwise diradical mechanism which

involves a diradical intermediate and is characterized by

the absence of stereospecificity. A third mechanism con-

sidered to be stepwise dipolar and involving a zwitterionic

intermediate has been observed in certain specific cyclo-

additions. Because of its limited applicability, the

dipolar mechanism will not be discussed.

Examination of the [2+2] cycloaddition in terms of the
28
Woodward-Hoffmann orbital symmetry theory28 affords valuable

insight into its mechanism. The dimerization of two ethyl-

ene molecules to cyclobutane will serve as an example due to

the system's simplicity. However, the conclusions drawn

will apply to all [2+2] cycloadditions. The simplest

approach geometry for ethylene dimerization is the supra7

facial, suprafacial (s,s) approach as shown below, since




-5-









Q Q
I







(s,s)


it yields maximum orbital overlap between the bonding centers.

The following correlation diagram illustrates that this

approach is thermally a symmetry-forbidden process and

may not take place concertedly.

However, there are other geometries of approach

possible, such as the suprafacial, antarafacial (s,a)

approach and the antarafacial, antarafacial (a,a) approach.






-


/ I
I /


"sfi ^ ^D -Q^/
^T} r ^firp


(a,a)


-^ -'


(s,a)



















*A --AAo*




H*AA -


H*AS


SS -


--- ASo
SSo


Figure 1: Correlation diagram for the formation of cyclo-
butane from two molecules of ethylene by an
(s,s) mode of addition.








By orbital symmetry techniques similar to the ones used for

the (s,s) case, one can determine that the (a,a) approach

is symmetry-forbidden while the (s,a) approach is symmetry-

allowed. Excited state cycloadditions would reverse the

above conclusions, allowing (s,s) and (a,a) processes but

forbidding (s,a) reactions. The simple illustration above

does not show the actual geometry 6f approach for the al-

lowed (s,a) process. The optimum arrangement for achiev-

ing maximum overlap of the bonding orbitals requires the

ethylene molecules to approach each other orthogonally as

shown below. However, even this approach involves relatively






















inefficient orbital overlap. In addition, non-bonded inter-

actions between the substituents on the reactants act to

hinder the orbitals' approach to each other. These effects

combine to make the symmetry-allowed [ 2S + 7T2 a]I mode of

addition very unfavorable. Not unexpectedly, it has been
/ \
/ \
/ \



\c







inefficient orbital overlap. In addition, non-bonded inter-

actions between the substituents on the reactants act to

hinder the orbitals' approach to each other. These effects

combine to make the symmetry-allowed [ 2s + r2 mode of

addition very unfavorable. Not unexpectedly, it has been





-8-


found that, in general, [2+2] cycloadditions proceed by

stepwise mechanisms. Furthermore, no example of concerted,

thermal [ 2s + 2 ] cycloadditions has yet been unequivocably

established.29' 30

The stepwise diradical mechanism for cyclobutane form-

ation was first proposed on the basis of the observed
13'
orientation in the cycloadducts. That is, the prevalence

of "head-to-head" over "head-to-tail" adducts, with the

"head" of a reactant molecule being the end which would

form the most stable radical. Thus the dimerization of

acrylonitrile leads to 1,2-dicyanocyclobutane, since the

diradical intermediate in which both radical centers are

stabilized by a-cyano-groups (head-to-head arrangement) is

more stable than the alternative diradical (head-to-tail

arrangement), which would lead to 1,3-dicyanocyclobutane.

Another example involves the addition of l,l-dichloro-

2,2-difluoroethylene to butadiene which proceeds as shown

below to give cyclobutane 2 in 92% yield.31 This suggests

the intermediacy of diradical 1 and not 3 or 4. Further-

more, the diradical 1 is estimated to be about 8 kcal/mole

lower in energy than 3, and more than 21 kcal/mole lower
32
than 4. Such energy differences are great enough to

insure that the orientation is almost all as shown.







F
CF
I 2 pF F
+ 1 P. / r F
CC 2 CCl2 1
Cl
1 2
Cl
Cl CH2 *CC2

2 *CF2 F

3 4 F


As stated previously, absence of stereospecificity in

the products is an experimental characteristic of the step-

wise diradical mechanism. The elegant and extensive work of
32, 33
Bartlett and his co-workers3 33 on the stereochemistry of

the [2+2] cycloaddition of fluoroalkenes has placed the

stepwise diradical mechanism for [2+2] cycloadditions on

firm ground.

The addition of tetrafluoroethylene to either cis- or

trans-1,2-dideuterioethylene gives identical product mix-
34
tures. This result, showing total lack of stereo-

specificity, is inconsistent with a concerted mechanism




CF D D
II + -
CF2 or F4 F4 + F4 D
D D 4 D


D




-10-


but totally in agreement with a stepwise diradical mech-

anism. Furthermore, the addition of l,l-dichloro-2,2-

difluoroethylene to the cis and trans double bonds of

the isomeric 2,4-hexadienes shows loss of configuration

at the double bond which becomes a part of the cyclobutane

ring but not at the other double bond.35 When the isomers

of 1,4-dichlorobutadiene are used instead of the 2,4-

hexadienes, the stereochemical results yield exactly the
36
same pattern. Use of the less reactive tetrafluoro-

ethylene instead of the l,l-dichloro-2,2-difluoroethylene

in the addition with 2,4-hexadiene also yields similar

stereochemical results.37

The above evidence is completely consistent with the

stepwise diradical mechanism. Loss of configuration at

the reacting double bond can be attributed to the internal

rotation of the diradical intermediate prior to cyclization.

In fact, it was estimated that internal rotation of the

intermediate occurs ten times faster than ring closure.35

Since the initial diradical conformation is unlikely to

be right for immediate ring closure, it is reasonable to

assume that rotations about the a bonds can occur in the

time required for the proper, ring-forming conformation

to be attained. The almost complete retention of configur-

ation in the other double bond is due to its being part

of an allylic radical which has a barrier to rotation.




-11-


Examination of recovered diene in the above cyclo-

additions shows small amounts of isomerization which do

not occur when only the dienes are subjected to the reaction

conditions. These results are attributed to formation of

the usual diradical, followed by rotation and reversion
32
to alkene and isomerized diene. Measurement of rates of

cleavage relative to ring closure was carried out for the

1,4-dichlorobutadiene system and found to equal 0.23 to 0.34.

Bartlett and his group have also studied the case of

competitive concerted [2+4] versus stepwise diradical [2+2]

cycloadditions. Generally, the concerted [2+4] cycloaddi-

tion is the most favored mode of addition between a 1,3-

diene and a monoene. However in some cases, for example the

addition of l,l-dichloro-2,2-difluoroethylene to butadiene,31

[2+2] cycloaddition dominates the reaction. Obviously,

diradical formation is sufficiently favorable compared to

the concerted process to dominate the reaction.

An important point to consider in these cases is that

acyclic dienes, such as 1,3-butadiene, exist in an equili-

brium between cisoid and transoid conformations, with the

transoid form being usually favored over the cisoid. The

diradical formed from the transoid diene is locked by its


transoid


cisoid





-12-


geometry into forming only cyclobutanes. However, a di-

radical formed from the cisoid diene could close to either

cyclobutane or cyclohexene.38 39 The cisoid conformation

also favors the concerted [2+4] cycloaddition as it cannot

occur from the transoid conformation.

A re-examination of the cycloaddition of butadiene

to l,l-dichloro-2,2-difluoroethylene'reveals that at 800,

99% of the product is cyclobutane, but a trace (1%) of

cyclohexene is also formed.38 Investigation of the cyclo-

adducts over the temperature range 600 to 1760 reveals that

the amount of cyclohexene formed varies from 0.9% at 600

to 2.3% at 1760. It is found that the fraction of cyclo-

hexene product has exactly the same temperature dependence

as the fraction of cisoid butadiene. Indeed, the ratio

of percentage of cyclohexene formed to percentage of cisoid

butadiene is approximately constant over the entire temper-

ature range. The authors interpret these results as evi-

dence that the small amounts of cyclohexene formed arise

from closure of the cis-diradical that results when the

alkene attacks the cisoid form of butadiene. However,

increasing amounts of cisoid butadiene also favor the

occurence of a competing concerted [2+4] addition.

A simple way to determine whether [2+2] or [2+4]

adducts arise by competing stepwise and concerted mech-

anisms would be to study the stereoselectivity of the

reaction. This has been done for the reaction of cis- and

trans-1,2-dichloro-l,2-difluoroethylene with cyclopentadiene.40




-13-


The trans alkene (containing 1% cis alkene) was allowed

to react with cyclopentadiene, resulting in formation of

both [2+2] and [2+4] adducts. The [2+4] adducts composed

97.3% of the total product; 96.7% of it was the trans and

0.6%, the cis [2+4] adduct. The four [2+2] isomers

composed 2.7% of the total product. The two trans isomers,

differing only in the position of the double bond, accounted

for 1.2% and 1.0%, while the two cis isomers comprised

0.2% and 0.3% of the total product. When 95% stereo-

chemically pure cis alkene was subjected to the same reac-

tion, the [2+4] adducts composed 94.0% of the total pro-

duct with 87.6% cis and 6.4% trans adducts. The four [2+2]

adducts accounted for 6.0% of the total product. The two

cis isomers comprised 2.4% and 1.7% of it, while the two

trans isomers consisted of 0.9% and 1.0% of the total pro-

duct. The 18-32% loss of configuration in the [2+2] adducts

is too large to be accounted for by 1-5% cis-trans impurity

in the reacting alkene, and consistent with that observed
35-37
in diradical cycloadditions. These results then are

indicative of simultaneous, competitive [2+4] concerted,

stereospecific cycloaddition and [2+2] stepwise, non-

stereospecific diradical cycloaddition.

The mechanism of the [2+4] cycloaddition has been

studied even more extensively than that of the [2+2]

reaction.1 12 The prototypical [2+4] cycloaddition of

butadiene and ethylene yields cyclohexene. Its examination
in terms of the Woodward-Hoffmann rules28 will yield results
in terms of the Woodward-Hoffmann rules will yield results





-14-


which will be applicable to all [2+4] cycloadditions.

Approach of the two components in an (s,s) fashion, as

shown below, provides for the greatest amount of orbital

overlap.







)0


>.5 /








Of course, the diene can only react in a cisoid con-

figuration. Dienes which are fused transoid do not react

in a [2+4] cycloaddition, and acyclic dienes in which there

exists a transoid-cisoid equilibrium can only form [2+4]

products from the cisoid conformation. The construction

of a correlation diagram for this reaction illustrates

that the (s,s) approach is a thermal, symmetry-allowed

process and may take place in a concerted fashion.28





-15-


_A *
44


4 *A--All*

7T*A-

3 *S--

- - ~- -

%2A-


ns__SH

Sis-
ISA

_Ao
^-Ac1





By using Woodward and Hoffmann's selection rules for

thermal polyene cycloadditions, it can be determined that

the (s,a) approach is symmetry-forbidden but the (a,a)

approach is symmetry-allowed. However, because of poor

orbital overlap in the (a,a) approach, the [2+4] cyclo-

addition normally takes place in an (s,s) fashion.

As mentioned previously, [2+4] cycloadditions exhibit

a high degree of stereospecificity with respect to both

diene and dienophile. This is a point of strong evidence

for the concerted mechanism since a stepwise mechanism

should result in non-stereospecific addition due to compe-

tition between ring closure and bond rotation.





-16-


The small effect of solvent polarity on the rate of

[2+4] cycloadditions is inconsistent with it occurring via

a stepwise, zwitterionic mechanism. The stepwise, diradical

mechanism for the [2+4] cycloaddition was first proposed

in 1936.41 More recently its main proponent has been

Firestone, whose arguments have been recently reviewed.42

However, the evidence is clearly against the diradical mech-

anism and the concerted mechanism has been almost universally

accepted. Formation of a diradical should be equally

possible from both cisoid and transoid dienes. However,

the diradicals formed from transoid dienes would lead to

cyclobutanes ([2+2] addition). Since most dienes and

dienophiles react only through the cisoid diene to give

exclusively six-membered rings, this is good evidence for

a concerted mechanism. Also the work of Bartlett and

his co-workers on competitive [2+2] and [2+4] cycloadditions

as discussed earlier, provides evidence for the concerted-

ness of the [2+4] reaction.

The extraordinary degree of regioselectivity observed

in many [2+4] cycloadditions has been explained in terms

of the concerted mechanism by the use of frontier orbital

theory.43 However, the diradical mechanism has also been

used to account for the observed regioselectivity.44 A

reaction recently reported45 has served as a test between

the concerted and diradical mechanism. Diene 5 reacted

with dienophile 6 to give a quantitative yield of the two




-17-


products, 7 and 8, in the proportions shown. These

results are inconsistent with the diradical mechanism,




MeO I MeQ Me

U+ Il +

5 6 '7 8

65% 85%

MeO



9



since formation of the most stable diradical 9 should

lead to 8 as the major product. However, the formation

of 7 as the major product is completely consistent with

the concerted mechanism and frontier orbital theory.

Frontier Molecular Orbital Theory

At the same time that Woodward and Hoffmann were

developing their fundamental theory on the conservation

of orbital symmetry, other workers were developing what

has come to be known as frontier molecular orbital (FMO)

theory.

The FMO theory focuses on the interaction between

the highest occupied molecular orbital (HOMO) and the

lowest unoccupied molecular orbital (LUMO) of two reacting

species. The HOMO and the LUMO are the frontier orbitals.

The theoretical basis for FMO theory lies in quantum





-18-


46
mechanics and the use of perturbation theory.46 This is

an approximate quantum mechanical method in which the

interaction of the molecular orbitals of the two reacting

species are treated as perturbations on each other.

The interaction of any two orbitals gives rise to two

new orbitals as shown below. By neglecting overlap,




T--j-j- C-i












the energy difference between the original and perturbed

orbitals is given by the second-order perturbation expres-

sion (1), where c. and Ec. are the energies of the orbitals



AE = (H 2/(c )



4t and di respectively. The closer in energy the two

orbitals are, the greater the interaction will be, since

the (c.-c.) term is in the denominator of (1). This

expression will hold for the interaction of any two orbitals

of the reacting species.





-19-


If the interaction were to take place between an

occupied orbital of one reactant and an occupied orbital

of the other, destabilization would result because of the

formation of a closed shell. Of course, the interaction

of the two unoccupied orbitals is trivial and has no

influence on the reactants. It is the interactions of

occupied with unoccupied orbitals that produce stabilization

or lowering of the system's energy and lead to bonding.

Of these stabilizing interactions, that involving the HOMO

of one species and the LUMO of the other, and vice-versa,

will be the most important since these orbitals are closest

in energy. That is why frontier orbitals are so important.

The other occupied-unoccupied interactions contribute to

the lowering of the system's energy but they are less

important than the HOMO-LUMO contribution.

Fukui has used FMO theory to explain the Woodward-
47
Hoffmann rules for pericyclic reactions. Simply stated,

the method is this. The frontier orbitals for the two

reacting species are determined. If the symmetry of the

HOMO of one component is such that it can overlap with the

LUMO of the other component, then the reaction is symmetry-

allowed and may occur concertedly. If the orbitals are of

the wrong symmetry for overlap, the reaction is symmetry-

forbidden.

For example, the thermal cycloaddition of two ethylene

molecules to form cyclobutane has been found to be symmetry-




-20-


allowed in an (s,a) fashion but forbidden in an (s,s)
fashion by use of Woodward and Hoffmann's symmetry correla-
28
tion techniques. The same results can be obtained from
FMO theory, by looking at the frontier orbitals and their
symmetry interaction as shown below. Similarly, if we look


LUMO 10

i ""-" antibonding


HOO
G0 0


(s,s)


0 \ LUMO

I /


0 0 OMO
0 0


(s,a)


at the cycloaddition of butadiene and ethylene to form
cyclohexene, the Woodward-Hoffmann approach determines
that the (s,s) reaction is symmetry-allowed. The same
conclusion can be reached by looking at the frontier or-
bitals of the two components as shown below. To determine


LUMO


HOMO


HOMO


LUMO





-21-


whether the reaction is allowed or not, it does not matter

which pair of frontier orbitals are used, as long as the

HOMO is from one component and the LUMO from the other.

The FMO theory has been very successful in rationalizing

reactivity and regioselectivity for ionic, radical, photo-
48
chemical, and thermal pericyclic reactions.48 Its most

dramatic success has been in the area of thermal pericyclic

reactions, with cycloadditions having the largest share
49
of success.

In using FMO theory to determine the allowedness of

a cycloaddition, it was unimportant which HOMO-LUMO pair

was chosen since either one would give the same answer.

However, to explain the effect of substituents on the rate

and regioselectivity of the cycloaddition, it must be known

which is the more important pair of frontier orbitals.

This choice can be made by following a very simple principle.

The dominant HOMO-LUMO pair will be the one with the small-

est difference in energy, or to put it another way, the HOMO

and LUMO which are closest in energy will be the more

important pair. Furthermore, the smaller the energy gap,

the faster the reaction should be.

The regioselectivity of the reaction can be explained

by looking at the orbital coefficients of the HOMO and LUMO.

These coefficients may be calculated by sophisticated

quantum mechanical methods or they may be estimated by
43
qualitative techniques.43 Once the correct HOMO-LUMO pair

has been chosen and the coefficients determined, the





-22-


preferred regioisomer can be predicted in the following

way. The larger terminal coefficients on each component

will become preferentially bonded and yield the predicted

product.

An example of the practicality of the FMO theory

will serve to clarify the above statements. This example

is taken from Houk's review on the'FMO theory of cyclo-
49
additions.49 The [2+4] cycloadditions of isoprene to

acrylonitrile and isoprene to 1,1-dicyanoethylene will

be examined. The frontier orbitals for the three mole-

cules are as shown in Figure 2. The orbital energies

have been obtained from experimental data and the orbital

coefficients have been calculated by quantum mechanical

methods. In the isoprene-acrylonitrile case, it can be

seen that the isoprene HOMO-acrylonitrile LUMO interaction

is the most important, because these orbitals are the

closest in energy. The "para" product will be the preferred

regioisomer because the largest coefficients are on C1

of isoprene and C2 of acrylonitrile. By uniting these

two sites, the "para" product is formed. Experimentally,

the "para" isomer predominates (70% at 20) 50

1,1-Dicyanoethylene has a lower LUMO energy than

acrylonitrile. This brings it closer to the isoprene HOMO,

resulting in greater interaction and faster reaction

rate. The data for isoprene has not been measured, but

cyclopentadiene reacts 100 times faster with 1,1-dicyano-

ethylene than with acrylonitrile.51 The regioselectivity





-23-


NC CN


0.8
SLUMO

.55
--.41
-.42
.56


HOMO -8.89
HOMO ~

.63
.42
-.33
-.49


-0.02 LU
LUMO


.75
-.59
CN


\-10 92
L HOMO


-1.54
S LUMO


.66
-.49
NC CN





-11.38
HOMO


.60
.49
CN


.61
.45

NC CN


Figure 2: Frontier molecular orbitals of isoprene, acrylo-
nitrile, and 1,1-dicyanoethylene. From ref. 49.


-11





-24-


of the 1,1-dicyanoethylene-isoprene reaction is greater

(91% "para" at 200) than that for the acrylonitrile case,

because the difference between the 1,1-dicyanoethylene

LUMO coefficients is greater and the interaction stronger.

In order to apply the FMO theory of cycloadditions to

all cases, the properties of the reactant frontier orbitals

must be known. Time consuming, computer-aided calculations

can yield these properties, but simple generalizations can

often supply enough information to make a prediction or

rationalize an experimental result. The methods for making

these generalizations will be covered in detail in

chapter three.

Allene Cycloadditions

Molecules such as allene have long been objects of

curiosity to chemists, due to their cumulated double bond

structure. In fact, the tendency of allenes to dimerize
52
was recognized 65 years ago by Lebedew. Allenes are

similar to simple alkenes in that they undergo both [2+2]

and [2+4] cycloadditions and on heating dimerize to give

dimethylenecyclobutanes. 26, 53

In recent years, Dai and Dolbier have investigated
54, 55
a broad spectrum of allene cycloaddition reactions54

using secondary deuterium isotope effects as the mech-

anistic probe. The reactions examined included two [2+4]

cycloadditions, and four [2+2] cycloadditions. Both intra-

and intermolecular isotope effects were determined





-25-


and they provided information about the product-determining

and rate-determining steps, respectively.

1,1-Dideuterioallene was used in intramolecular

competition experiments to determine the intramolecular

isotope effects. Competition experiments using tetra-

deuterioalene and undeuterated allene were used to deter-

mine the intermolecular isotope effects. A normal

(i.e., kH/kD >1) isotope effect was observed for all of the

[2+2] processes but the [2+4] processes yielded inverse

(i.e., kH/kD
A concerted reaction in which both new a bonds are

formed should give rise to an inverse intramolecular

isotope effect, and since in this case the rate-determining

and product-forming steps are one, the intermolecular

isotope effect should be of the same magnitude. The results

confirm these predictions for the [2+4] reactions which

according to orbital symmetry rules should be concerted.




Cl Cl6 Cl

CH
2 CD



I I D
CD2

10
Intramolecular kH/kD = 0.90 0.02


Intermolecular kH/kD = 0.90 0.04





-26-


The normal isotope effect observed for the [2+2] processes

is inconsistent with anticipations for a concerted reaction

but can be rationalized for a two-step reaction. Further-

more, the significantly different intra- and intermolecular

isotope effects observed, as shown below for the reaction

with acrylonitrile, indicate that the rate-determining and


C2 D H H

I + +D CN CH


II D
CN *-

CD, 212
10 Intramolecular k /k = 1.13 1.21

Intermolecular k /kD = 1.04 + 0.05

Equilibrium k /kD = 0.92 + 0.01



product forming steps cannot be identical, so that these

reactions must have multi-step mechanisms. Interestingly,

this [2+2] cycloaddition was complicated by a concurrent

degenerate methylenecyclobutane thermal isomerization. When

carried to complete equilibration, the measured equilibrium

isotope effect was inverse. This result was expected, due

to deuterium's thermodynamic preference for sp3 bonding.

Thus it could not be responsible for the observed kinetic

effect. A detailed examination of the allene-acrylonitrile

reaction can give further insight into the reaction mech-

anism and explain the observed kinetic isotope effects.





-27-


H CN


D H D
C=C=C "H
\ H
+ H
13
H H

H CN

HH CN

D CN
D fast

H H H

14


It is significant that in non-concerted cycloadditions,

allene consistently forms the initial C-C bond at the center,

sp-hybridized carbon.56 This should give rise to an allylic

radical intermediate. The rate-determining step of the

two-step mechanism shown above involves no hybridization

change at C-l or C-3 and should give rise to a small inter-

molecular isotope effect. This expectation is borne out

by the observed isotope effect of 1.04 0.05.

The product-forming step involves the destruction of

the stabilized diradical 13. Although the combination of

two simple, non-stabilized radicals has negligible activa-

tion energy, the destruction of stabilized diradicals such

as 13 seems to require an activation energy of approximately

13 kcal/mole.55 Probably, the source of this activation
13 kcal/mole. Probably, the source of this activation




-28-


energy is the rotation of a methylene group on the

allyl radical towards the orthogonal geometry, as

shown in 14, necessary for a bond formation. There-

fore the transition state 14 will have little or no

a bonding and the isotope effect could be considered a

"steric" isotope effect. It would be derived from a

"loosening" of the rigidity and steric requirements

of the planar allyl system, thus allowing the -CH2

to rotate faster than the -CD2 out of a sterically

congested situation, and giving rise to the observed

normal isotope effect.

From these studies, it was concluded that allene

behaves as a normal alkene undergoing concerted

[2+4] cycloadditions and multi-step [2+2] cyclo-

additions. Just as important, the use of kinetic

secondary deuterium isotope effects as mechanistic

tools apparently allows the distinction between

concerted and non-concerted mechanisms.

The use of fluorine as an allene substituent,

instead of deuterium, may also provide a mechanistic

tool which, like the isotope effects above, would

allow a distinction between concerted and non-

concerted cycloaddition mechanism. This is based





-29-


on the dramatic thermodynamic preference for fluorine

to be bound to sp3-hybridized carbon rather than
2 57
sp -hybridized carbon.57 That preference is believed

to account for the enhanced reactivity of fluoroolefins
15, 58
in cycloaddition reactions. 58

Effect of Fluorine as a Substituent

Fluorine's uniqueness as a substituent derives mostly

from three factors. The first two factors are fluorine's

high electronegativity and the presence of fluorine's

non-bonded electron pairs. Third is the fact that as a

second period element, fluorine has orbital dimensions

which make possible excellent overlap both in forming a

bonds with carbon and in T-conjugative interaction with

contiguous carbon i systems. The last factor acts to

magnify the effect of the first two.

The effective orbital overlap and high electroneg-

ativity give rise to a polar, very short C-F a bond (1.317
59
vs 1.766 for a C-C1 bond). Furthermore, the bond strength

of the C-F bond seems strongly dependent upon carbon's

hybridization. There is a significant thermodynamic

advantage for bonding to an sp3-hybridized carbon instead

of an sp2-hybridized carbon.





-30-


This can best be seen by examining the equilibrium

in the butadiene-cyclobutene system. The cyclization of

the hydrocarbon is significantly endothermic but the

equilibrium for the perfluorosystem favors the cyclo-

butene by 11.7 kcal/mole.57 These results can be ration-

alized as arising from an electronegativity effect (it has



F6 F



AHl = -11.7 kcal/mole


been pointed out that the 2p character of the atomic orbital

used by carbon to form a C-X o bond should increase with

increasing electronegativity of X),60 and from H-conjuga-

tive destabilization due to antibonding interactions of

fluorine's lone pairs with an sp -carbon's I system.

Because of the short C-F a bond, and the good H overlap

with carbon's 2p orbitals, this "repulsion" seems to be

especially significant for fluorine.

Evidence for this effect comes from the observation

that a-fluorine substitution apparently destabilizes

carbanions relative to a-chlorine.61 This occurs in spite

of the inductive stabilizing power of fluorine towards

anions, which is manifested in systems having fluorine

substituted 6 to the anionic site.61 The stereochemistry

about the anionic site is critical in the case of





-31-


a-substitution, with the destabilization increasing dra-

matically when the carbanion is planar instead of tetra-

hedral.61

Florine's effectiveness as a f-electron donor becomes

apparent in its ability to stabilize a cationic center.

In spite of its electron withdrawing inductive effect,

fluorine actually activates the papa position in the

chlorination of fluorobenzene, and has a slightly negative
61
a + value. Other examples of fluorine's H-electron

donor capacity to stabilize a-electron deficient sites

have been discussed.

A study of the cycloaddition reactions of 1,1-difluoro-

allene, particularly those in which a partition of products

between [2+2] and [2+4] reactions could occur, might allow

distinction between concerted and stepwise mechanisms.

This would provide a unique probe of cycloaddition mechanisms,

with more dramatic results than those of the isotope effect

studies using dideuterioallene.













CHAPTER II
RESULTS

The first reported synthesis of 1,1-difluoroallene

was in a 1956 patent.62 The pyrolysis of 1-methylene-

2,2,3,3-tetrafluorocyclobutane, obtained from tetrafluoro-

ethylene and allene,63 in a quartz tube at 8000 and 6mm

pressure was reported to give 1,1-difluoroallene and 1,1-

difluoroethylene as volatile products. In 1957 Blomquist

and Longone published an alternative, five-step synthesis

beginning with ethylene and dibromodifluoromethane.64

1,3-Dibromo-l,l-difluoropropane (15) was obtained in 50%



CH C2 + CF2Br -- CF2BrCH2CH2Br

15

CF2BrCBr = CH2 4-- CF2BrCHBrCH2Br 4-- CF2BrCH = CH2

18 17 16



S CF2 = C = CH2

19



yield from the free radical addition of dibromodifluoro-

methane to ethylene. Dehydrobromination of 15 with aqueous

potassium hydroxide gave a 60% yield of propene 16, which

was photobrominated to give the propane 17 in 86% yield.


-32-





-33-


Treatment of 17 with cold 10% alcoholic potassium hydroxide

resulted in a 62% yield of propene 18. Debromination of 18

with zinc in refluxing ethanol occurred readily to give

a 56% yield of 1,1-difluoroallene (19). The overall yield

of allene from 15 was 18%.

In 1960, Knoth and Coffman published a third synthe-

sis of 1,1-difluoroallene.65 They'reported that the pyrol-

ysis of l-acetoxymethyl-2,2,3,3-tetrafluorocyclobutane,

prepared in 70% yield from allyl acetate and tetrafluoro-

ethylene, in a quartz tube at 8500 and 1 mm pressure yields

1,1-difluoroallene in 25-40% yield. The best synthesis

of 1,1-difluoroallene thus far reported and the one used to

prepare 19 for this study is that of Drakesmith, Stewart,
66
and Tarrant. Photobromination of 3,3,3-trifluoropropene

(20) gave an 85% yield of the propane (21). Dehydrobromina-

tion of 21 by potassium hydroxide in the absence of solvent



CF3CH = CH CF3CHBrCH Br -- F CF CBr = CH2

20 21 22



CF = C = CH2 -- CF3CLi = CH2

19 23


gave a 95% yield of trifluoroisopropenylbromide (22). Treat-

ment of 22 with n-butyllithium solution at -900 resulted in

formation of 23 which gave 1,1-difluoroallene when allowed





-34-


to warm to room temperature. Distillation of the crude

difluoroallene thus obtained gave a distillate of 75%

difluoroallene content. The remaining 25% was composed of

10% 3,3,3-trifluoropropene, 9% 3,3-difluoropropyne, 2%

ether and 4% unidentified products none greater than 1%

concentration. Further distillation did not increase the

difluoroallene content, and the 75% pure difluoroallene

was used for the cycloaddition reactions. The allene was

formed in 18% yield from 22 and 14% yield from 20. Difluoro-

allene of 97% purity could be obtained by preparative glpc,

and it was used to obtain spectral data. The allene struc-

ture was confirmed by comparison of its ir spectrum to that

reported in the literature.6 The 1C nmr of 19 has been
67 68 '
reported, along with its microwave spectrum, mass
64
spectrum, and complete analysis of its ir and Raman
69 1
spectrum. The H nmr (100 MHz, CC14) showed a triplet

at 6 6.04 (J = 3.5 Hz) and the 1F nmr a triplet at TFA

- 26.7 (JHF = 3.5 Hz). The photoelectron spectrum of

difluoroallene, determined by Houk and Domelsmith,70

showed a first ionization band at 9.79 eV, and a second

ionization band at 11.42 eV.

In their paper, Knoth and Coffman briefly examined

some of the cycloaddition chemistry of difluoroallene

including its dimerization.65 Initially, two of their

reactions were re-examined. Excess cyclopentadiene reacted

with difluoroallene at 00 to give, in near quantitative

yield, 5-difluoromethylenebicyclo[2.2.1]hept-2-ene (24) as




-35-


had been found previously.65 Analysis by glpc indicated

that only 24 appeared to have formed. The structure of



CF2 CF2

II +
C

124
CH2
19

24 was confirmed by nmr, ir, mass spectrum and exact mass
-i
analysis. Its strong C = C stretching band at 1775 cm-

was consistent with the terminal fluoroolefin (C = CF2)
71, 72
structure.71, 72 (For complete ir data, see experimental

section). Its 1H nmr bore similarities to other known

bicyclo[2.2.1] systems.73' 74 (For complete 1H and 19

nmr data, see experimental section.)

Preparation of 24 was also carried out by an indepen-

dent scheme as shown below. Dehydronorcamphor (25),
75
prepared by the method of Bartlett and Tate, was treated

CH2 CN

OAc

NC OAc




24 2CF 25d

24 25




-36-


with the Burton-Naae difluoro-Wittig reagent76' 77 to yield

24 in 27% yield. The material obtained in this manner was

identical in all respects to that obtained from cyclopenta-

diene and difluoroallene.

The second reaction of Knoth and Coffman to be re-examined

was that between acrylonitrile and difluoroallene. It

was carried out in a sealed tube, with benzene as solvent,
19
for nine hours at 1250. Analysis by glpc and F nmr indi-

cated that only two products were present in a 4:1 ratio.

The major product was 2,2-difluoro-3-cyanomethylenecyclo-

butane (26) and the minor product was 3-cyano-l-difluoro-

methylenecyclobutane (27). The structures of the products


CF2 CH

II I! ic 2 CCF2
C + C -~ +

S CN HF H H
CH2 NC F NC H

19 26 27


were verified by nmr, ir, mass spectrum, and exact mass

analysis. These results are consistent with those of Knoth

and Coffman,6 who found the reaction to give 26 as the

major product as well as an unidentified isomer which

contained a difluoromethylene (C = CF2) group. It was

determined that rearrangement of 27 to 26 will not occur

at 1250. Only when 27 is heated to 2400 does conversion

to 26 begin to occur.




-37-


In order to examine the [2+4] cycloadditions of

difluoroallene with other cyclic dienes, its reactions

with furan and hexachlorocyclopentadiene were carried out.

Except for the reactions with cyclopentadiene and acrylo-

nitrile, all cycloadditions of difluoroallene were carried

out in sealed glass tubes with an excess of diene, small

amounts of hydroquinone, and in the absence of any solvent.

The reaction with furan was carried out for 20 min at 500


CF2

; o -, cb"
C +

C2 28
19

An H nmr of the reaction mixture reveals the presence of

only one product. This was purified by preparative glpc

and its structure confirmed as 5 -difluoromethylene-

7-oxabicyclo[2.2.l]hex-2-ene (28) by its nmr. Hexachloro-

cyclopentadiene reacted with difluoroallene at 1000 for

2.5 hours. Molecular distillation of the residue, after

removal of unreacted diene, yielded a small amount of yellow

CF2 C
C2 6 Cl

C + CF2

CH
CH2 29
19




-38-


liquid verified to be 1,2,3,4,7,7-hexachloro-5-difluoro-

methylenebicyclo[2.2.l]hept-2-ene (29) by nmr, ir, mass

spectrum, and exact mass analysis.

Three cycloadditions of difluoroallene which produce

both [2+2] and [2+4] adducts were carried out and examined

in terms of their total regioselectivity. The dienes used

were 1,3-butadiene (30a), isoprene' (30b), and 2-trimethyl-

silyloxy-1,3-butadiene (30c). The latter was synthesized
78, 79
using the method of Jung.7 79 As indicated below, these

reactions yield difluoromethylene cyclohexenes and vinyl

methylenecyclobutanes as products. Analyses of the reaction

mixtures were carried out by glpc and determination of the

absolute total yield was carried out using carefully con-

structed calibration curves, and hydrocarbons as internal

standards. For more details, the experimental section may

be consulted.


X CF2 -F2 X CF2



CH
II
CH2 31 32


30a, X = CH2 F2

30b, X = CH3 + X + X H
/X F H F H H
30c, X = OSiMeH F H H

33 34 35





-39-


The reaction of difluoroallene and 1,3-butadiene

resulted in formation of two products in an overall yield

of 87%. The major product was 4-difluoromethylenecyclo-

hexene (31a), and the other product was 2,2-difluoro-3-

vinylmethylenecyclobutane (33a). The structures of these

compounds were confirmed by ir, nmr, mass spectrum, and

exact mass analysis. Carrying out'the reaction at a

temperature 300 higher resulted in a similar product dis-

tribution.

Isoprene and difluoroallene gave four different cyclo-

addition products in an overall yield of 85%; two [2+4]

products in approximately equal amounts and two [2+2] pro-

ducts in a 2.6:1 ratio. The two [2+4] adducts were 1-methyl-

4-difluoromethylenecyclohexene (31b) and 2-methyl-4-difluoro-

methylenecyclohexene (32b) which were in an apparently in-

separable mixture. The major [2+2] adduct was 2,2-difluoro-

3-methyl-3-vinylmethylenecyclobutane (33b) and the minor

product 2,2-difluoro-3-isopropenylmethylenecyclobutane (34b).

The structures of all these compounds were verified by

spectroscopic analysis. When the reaction was carried out

300 higher, it gave similar product patterns.

In the reaction between 2-trimethylsilyloxy-l,3-

butadiene and difluoroallene four cycloaddition adducts

were obtained in an overall yield of 60%. The two [2+4]

adducts were l-trimethylsilyloxy-4-difluoromethylenecyclo-

hexene (31c) and 2-trimethylsilyloxy-4-difluoromethylene-

cyclohexene (32c), formed in equal amounts. The [2+2]





-40-


adducts were in a 2.2:1 ratio. The major product was

2,2-difluoro-3-trimethylsilyloxy-3-vinylmethylenecyclo-

butane (33c) and the minor one was 2,2-difluoro-3-(l-tri-

methylsilyloxy)vinylmethylenecyclobutane (34c). All of

the structures were confirmed by nmr, ir, mass spectrum,

and exact mass analysis. When the reaction was carried

out at 200 higher temperature, it gave identical product

distribution. The results of the cycloaddition of difluoro-

allene to 30a, 30b, and 30c are summarized in Table 1.

In order to determine whether the 10% 3,3,3-trifluoro-

propene impurity in the difluoroallene was reacting with

the dienes, reactions between 16a, 16b, and 16c and

3,3,3-trifluoropropene were carried out. In all three

cases, the 3,3,3-trifluoropropene yielded no glpc detectable

products under cycloaddition reaction conditions.

To test the behavior of a diradical containing the

CF2 moiety when it cyclizes to a six-membered ring, the

diradicals were generated by an alternate route involving

pyrolyses of some [2+2] adducts of difluoroallene. The

static pyrolyses of 33a, 33b, and 34b were carried out at

2100 for 14 hours, using a 5% solution of the adduct in

n-decane. The flow pyrolysis of 33a was carried out at

6940. All four pyrolyses resulted in extensive decomposi-

tion yielding no product or recovered starting material.

This was attributed to surface effects causing elimination

of HF and decomposition.





-41-


Table 1

Yields of Cycloaddition Adducts


Diene Temp C
Used


Time
(hrs)


Relative Yield
31 32'


a,c


Absolute


STotal
33 34 Yield


1100

1400

1100

1400

1400

1600


7 63

2 52 --

7.5 (29)b (35)b

3 (25)b (30)b

5.5 23 23

5.5 23 23


aRelative yields of products
of time of reaction.


37

48


-- 87%

-- 82%


26 10 85%

32 12 66%

37 17 60%

37 17 64%


were found to be independent


Distinction between 31 and 32 was not made; all spectra
of 31 and 32 were of mixtures and relative amounts were
determined by capillary glpc; product identifications were
by spectroscopic analyses of glpc isolated materials.
c
In these reactions, traces of the exocyclic CF2 adduct 35
were detected, the ratio of 33+34:35 never being smaller
than about 20:1.


30a

30a

30b

30b

30c

30c




-42-


The [2+2] adducts of cyclopentadiene and difluoro-

allene are not available from the cycloaddition of the

two reactants. One of the cycloadducts was synthesized as

set forth in the scheme below. Reaction of cyclopenta-

diene with dichloroketene, generated in situ from dichloro-

acetylchloride and triethylamine, gives 7,7-dichlorobicyclo-

[3.2.0]hept-2-ene-6-one (36) in 78o yield.80 Reduction

0

C + 0 1
CC12 +1
C1
36




$yc2 j--


38 37



of 36 with two equivalents of tri-n-butyltin hydride results

in a 70% yield of bicyclo[3.2.0]hept-2-ene-6-one (37).80

This was treated with the Burton-Naae difluoro-Wittig

reagent76, 77 to produce a 27% yield of 6-difluoromethylene-

bicyclo[3.2.0]hept-2-ene (38). The structure of this

compound was confirmed by ir, nmr, mass spectrum, and exact

mass analysis.

In contrast to the pyrolyses of 33a, 33b, and 34b,

the pyrolysis of 38 was very clean. It was carried out in




-43-


the gas phase for 10-11 hours in a well conditioned Pyrex

vessel heated to 2220 in a fused salt high temperature

thermostat. Pyrolysis resulted in the formation of two

products in a 4.1:1 ratio and complete disappearance of

starting material. Both products appeared to be stable

to the pyrolysis conditions. The major product was found

to be 5-difluoromethylenebicyclo[2'2.l]hept-2-ene (24) which

was identical in all respects to that obtained by the cyclo-

addition of difluoroallene and cyclopentadiene. The minor

product was 7,7-difluoro-6-methylenebicyclo[3.2.0]hept-2-

ene (39) whose structure was confirmed by ir, nmr, mass

spectrum, and exact mass analysis. The pyrolysis of 24

was carried out at 2860 and resulted in slow decomposition

of 24 without formation of 38 or 39. The behavior of the








FF2 F 2 H 2
F


38 24 39



diradical generated from pyrolysis of 38 is not consistent

with the cycloaddition results, but seems to be governed

by other factors to be discussed in chapter IV.














CHAPTER III
DISCUSSION

The reactions of difluoroallene with 1,3-butadiene

(30a), isoprene (30b), and 2-trimethylsilyloxy-l,3-

butadiene (30c) yielded significant amounts of both [2+2]

and [2+4] adducts, as summarized in Table 1. The [2+4]

adducts were exclusively those with an exocyclic difluoro-

methylene group, while the [2+2] adducts were, in stark

contrast, almost entirely those with endocyclic CF2 (see

note c in Table 1). It is also evident that orientation

effects in the [2+4] and [2+2] reactions are markedly

different. Although no meaningful orientational preference

was observed in the [2+4] adducts of 30b and 30c, the ratios

of 33:34 in these reactions were 2.6:1 and 2.2:1, respec-

tively.

All of these results may be viewed in terms of two

distinct mechanistic alternatives. The first would envi-

sion all of the cycloaddition products arising from diradi-

cals by non-concerted, stepwise mechanisms, whereas the

second would entail a dichotomy of mechanism involving

competition between a concerted [2+4] reaction and a

non-concerted [2+2] reaction, with the latter occurring

via a stepwise, diradical pathway.


-44-





-45-


The mechanism involved in the first alternative is

illustrated in Figure 3, using 1,3-butadiene as the diene.

The butadiene is initially attacked at a terminal carbon,

and the difluoroallene forms its initial bond at the center

carbon,56 both resulting in formation of allyl radicals.

Of course, butadiene exists in an equilibrium between the

cisoid and transoid conformations.

Transoid butadiene is attacked by difluoroallene to

form diradical 40. This diradical has three options. It

can dissociate and revert to reactants, rotate to the

cisoid diradical or cyclize to a cyclobutane ring. The

formation of 40 has been found to be irreversible. In

the thermal rearrangement experiments described in chapter

2, there were no cases which resulted in evidence of retro-

cycloaddition. Since these rearrangements were carried out

at temperatures much higher than those of the cycloadditions,

it seems unreasonable that a reversion to reactants would

occur during cycloaddition, which didn't occur from the

diradical at higher temperatures.

A six-membered-ring product cannot be formed from 40

because the butadiene moiety is locked into its transoid

configuration. Rotation to the cisoid diradical would have

an energy barrier equal to the resonance energy of the

allyl radical. Experimentally this has been found to be81

14.3 2 kcal/mole. The third option left to diradical 40

is 1,4-cyclization to form a cyclobutane ring. O'Neal





















0






0
-4
V
>,
u

a,
a,





C:
0
Q)






0
:1






r-4

U4



0
ra












-C-)
u












a,
ri











u


-4

Cd




-4



0






a,




r14





-47-


/ x
"a a


1














\a ^


So1





-48-


82
and Benson have estimated the energy barrier for such cycli-

zation to be equal to 7.4 kcal/mole. A comparison of the rate

of cyclization to the rate of allyl radical rotation, using

the two energy barriers above, shows that at 1100 cyclization

to the cyclobutane is about 11,000 times faster than allyl

radical rotation. The principal course for diradical 40 to

follow, then, would be cyclizationito the cyclobutane. The

preferential cyclization of the difluoroallyl radical moiety

at the difluoromethylene terminus follows from the previously

discussed preference for the CF2 group to be sp3 hybrid-

ized.57, 61 In fact, both ab initio (STO-3G) and MINDO/3

calculations performed by Houk and Strozier83 indicate that

the difluoroallyl radical, although planar, simultaneously
84
pyramidalizes and rotates at the CF2 terminus more easily

than at the CH2 terminus.

From the cisoid butadiene two diradicals of interest,

namely 41 and 42, could form. Cyclization to the cyclo-

hexene would occur by simple rotation of the indicated bond,

which would bring the p orbitals into ideal position for

bonding. In order to account for the observation that

the cyclohexenes obtained in these reactions are exclu-

sively those with an exocyclic difluoromethylene group,

certain steric effects must be invoked. These involve

the assumption that fluorine is a much more sterically

demanding substituent than hydrogen. Thus cisoid diradical

42 should be sterically more hindered than 41 because the

fluorine in 42 would crowd the hydrogens more. Diradical




-49-


41 would then be formed preferentially, and continue on

to the observed product. Because the cyclization requires

little twisting, formation of the cyclohexene would be

relatively easy.

In summary, the totally diradical mechanism would be

as set forth in Figure 3. The [2+2] cycloaddition would

take place by irreversible formation of transoid diradical

40. The cyclobutane product would then be formed from it

by cyclization. The [2+4] cycloaddition would occur by

preferential formation of cisoid diradical 41 over 42.

Cyclization of 41 would form the observed product, 31a.

Although this mechanism seems to explain the experimental

observations, there are various factors which work against

it.

At 1100, the ratio of transoid to cisoid butadiene

is about 11.38 Because of this, the transoid diradical 40

should be the predominant diradical formed, and should lead

to a predominance of [2+2] adduct in the product. However,

in all these reactions the [2+2] and [2+4] adducts were

formed in comparable amounts. Since the possibility of

reversible diradical formation has already been ruled out,

all diradicals formed must go on to product. The same

steric effects which favor transoid butadiene in the

butadiene equilibrium should also favor formation

of the transoid diradical. This makes the




-50-


transoid-cisoid butadiene equilibrium unfavorable in

terms of the observed products.

Thermochemically, there should be little difference

in the heats of formation of diradicals 40 and 41, other

than that due to the cisoid vs transoid geometry of one

of the allyl radicals formed (2-3 kcal/mole).12 The

activation energies (Ea) of cyclifation for both 40 and

41 should be small (%7.4 kcal/mole),82 and as such, an

increase in reaction temperature should do little to the

ratio of [2+4] to [2+2] products. When the cycloaddition

of difluoroallene and butadiene was carried out at 1400

(300 higher), the [2+4]/[2+2] ratio went from 1.7 to 1.1,

a 35% decrease. Similarly, the reaction with isoprene

at 1400 produced a 28% decrease. The changes observed

with butadiene and isoprene are inconsistent with the wholly

diradical mechanism. Furthermore, if the diradical mech-
38
anism were operating, the slight increase (3%)38 of cisoid

butadiene would predict a slight increase in the [2+4] pro-

duct. Thus, it is difficult to see how both [2+2] and [2+4]

products could be arising from diradicals.

The steric effects invoked to affect the conformational

equilibrium of cisoid diradicals 41 and 42 so as to favor

formation of 41 are inconsistent with the fact that experi-

mentally, fluorine is a very small substituent. For example,

in fluorocyclohexane the equatorial position is favored over

the axial by only 0.12 kcal/mole, in contrast to the

1.7 kcal/mole value for a methyl substituent.8 86





-51-


Examination of some estimates for van der Waals radii8

indicate that a fluorine substituent will take up less
0 0
space (2.23 A) than a methyl (3.05 A) but more than a
O
hydrogen (1.51 A). The construction of models of diradicals

41 and 42 plainly shows that there would be little, if any,

conformational preference for 41 over 42, and that the

steric effect of fluorine upon cyclization of 42 should

be minimal. Most importantly, the slight differences

which might exist between 41 and 42 or their ability to

cyclize, should not totally preclude formation of 43.

In fact, not even traces of 43 can be detected.

The cycloadditions of substituted allenes can provide

a measure of the importance of steric effects on conforma-

tional equilibria of diradicals such as 41 and 42. The

[2+2] reactions of 1,1-dimethylallene with acrylonitrile

and chlorotrifluoroethylene formed products with both




o
CF 2

H I\
CFCl \CN






F Cl F Cl H + H
F F N N

39% 22% 21% 11%




-52-


encb- and exocyclic methyl groups in reasonable amounts.

Apparently the steric effect of the methyl groups was

not large enough to prevent their incorporation into the
89
cyclobutane rings. A study of the [2+41 reaction of

cyclopentadiene and methylallene found that the methyl






+ +


17% 9%




group was again incorporated in the ring as well as on the
90
exocyclic methylene group. A third example involves the

photochemical (triplet) cycloadditions of quinones with

substituted allenes. Because this is a photochemical

reaction, diradicals are undoubtedly involved in the mech-

anism. Dimethylallene reacts with phenanthroquinone 44

and shows no hesitation about incorporating the methyl

groups into the ring. Similarly, methoxyallene reacts

with tetrachlorobenzoquinone 45 to yield exclusively the

product which contains the methoxy group in the ring.

From all of the above examples, it appears that steric

effects due to the allene substituents are not very impor-

tant in determining conformational equilibria of diradicals

formed in allene cycloadditions. Significant amounts of.




-53-


+ hv


44 51% 17%


Cl OC Cl Cl OCH


Shv 3 1
Cl C1 Cl

45 54% 0%


both the exocyclic and more sterically crowded endocyclic

substituted products were formed, with methoxyallene form-

ing only the endocyclic product. The unimportance of the

steric effects of methyl and methoxy, both larger substi-

tuents than fluorine, cannot support the idea that the

steric effect of fluorine would lead to the specificity

observed in the [2+4] cycloadditions of difluoroallene.

Thus, the argument that steric effects influence the equi-

librium of diradicals 41 and 42 (see Figure 3) to preferen-

tially form 41 and product 31a does not satisfactorily

explain the results obtained and makes diradical [2+4]

cycloaddition unlikely.

The observed contrast in orientation effects for the

[2+2] and [2+4] cycloadditions is difficult to explain in

terms of the diradical mechanism. While the preference for

33 over 34 in the [2+2] reactions is consistent with preferen-

tial formation of the more stable transoid diradical 46 over





-54-


the less stable diradical 47, the lack of orientational

preference in the [2+4] reaction is inconsistent with it


CF2


CH2


CH2


F

x
H C
2


4
F
F

H2C


proceeding by a diradical pathway. Reaction of 2-substi-

tuted cisoid butadiene with difluoroallene should result

in preferential formation of the more stable cisoid


CH2 49
.--


CF


x


CH2


CF


x





-55-


diradical 48 over the less stable species 49, and lead to

preferential formation of cyclohexene 32. Indeed, Bartlett

has found that cycloadditions of 2-substituted butadienes

in which there is competition between [2+2] and [2+4]

reactions generally exhibit similar orientational preferences



CC1 F F
+ 2 F Fj

CF2 C1- Cl
Cl H Cl,

15% 83%

C1
C 1 C l
F Cl

.2% 1.4%

39
between the two processes. There are exceptions to this,

but they are cases, such as with 2-t-butylbutadiene, where

a significant steric effect exists at the 2-position, and
39
hinders attack at the 1-position.39 Bartlett's results

can be interpreted as supportive of the diradical mechanism.

However, the unique results obtained with difluoroallene,

using dienes with sterically small 2-substituents, can be

correlated with the diradical mechanism only by claiming

a different regiospecificity of difluoroallene on cisoid

and transoid diene. This claim lacks any analogy in the

literature, and the results must be taken as evidence that

the diradical mechanism is not operating.




-56-


To summarize, the unfavorable conformational equilib-

rium of the diene, and the observed contrasting regiospeci-

ficities of these [2+2] and [2+4] cycloadditions, with

respect to both the location of the CF2 group and the

orientational preference of the diene moiety, make the

totally diradical mechanism of Figure 3 an unlikely and

unattractive alternative.

The second mechanistic alternative proposed explains

the results in the simplest and most satisfying manner.

They ensue from a competition between a concerted [2+4]

cycloaddition and a non-concerted [2+2] cycloaddition. The

results of the [2+2] reaction are completely consistent

with the intervention of diradical intermediates as shown

on Figure 4. The exclusive formation of cyclobutanes

containing endocyclic CF2 is completely consistent with

preferential cyclization of the CF2 end in the diradicals.83

This is due to fluorine's preference for bonding to sp3

hybridized carbons.5 61 The orientational preference

for formation of cyclobutane 33 over 34 follows from prefer-

ential formation of the more stable diradical 46. Since

both of these aspects of the [2+2] diradical cycloaddition

have been previously discussed in detail, no further comment

seems necessary.

The formation of significant amounts of both [2+2]

and [2+4] products in these cycloadditions, despite the

unfavorable transoid/cisoid butadiene ratio (11.2 at 110?),

is an indication that the [2+2] and [2+4] reactions are






-57-


c-l











x


U U


N 0 N

U U


r-4

r14


Q)





0
-l


NO
0


c-)




0
H-l

4-
*rl










0
















U)
-1g
0

+ -
(M-H
ro




u
>,





-0-










4J


u




rrl


*r l
fc




-58-


fairly competitive. Rate constants, as defined by the
91
Arrhenius equation (2), are made up of entropy and

enthalpy contributions. The enthalpy contribution is due



k = A e-Ea/RT (2)



to the activation energy, Ea, whicl is related to the

enthalpy of activation AH*, by equation (3). At 1100



E = AH* + RT (3)
a


Ea and AH* differ by only 0.76 kcal/mole. The entropy

contribution is due to the entropy of activation, AS*,

which enters equation (2) as part of the preexponential
91
factor A, as defined in equation (4). Of course,


SKT n AS*/R (4)
A = ee (4)



experimentally, the AS* is not obtained from equation (4)

but from the free energy of activation, AG*. The basic

thermodynamic equation which relates AH* and AS* to the

AG* is equation (5).



AG* = AH* TAS* (5)



If two reactions are competitive, i.e., their rates

are comparable, the AG* of the reactions must also be comparable.





-59-


Then, if the AS* of one reaction is much different than

that of the other, the AH* of the reactions will also have

to be different, to account for the comparable rates and

AG*. It can be seen from (5) that, for a constant AG*,

the relationship between AH* and AS* is such that as the

AS* becomes more negative (larger) the AH* becomes less

positive (smaller) and vice-versa.' Since AH* and E
a
differ by a very small factor, the above also holds when

AH* is replaced by Ea
a
A search of the literature can provide average values

of AS* and E for both [2+2] and [2+4] cycloadditions. The
a
AS* values were calculated (at 3000K) from the reported

log A and equation (4). The E were used as reported.
a
91-95
Nine [2+2] cycloadditions91-95 and eleven [2+4] cycloaddi-
12 91 96-99
tions 9 9 were examined. The [2+2] reactions had

an average AS* = -23 9 e.u. and an average E = 27 4
a
kcal/mole. The [2+4] reactions had an average AS* =

-36 5 e.u. and an average E = 20 5 kcal/mole. Of
a
course, there are [2+2] cycloadditions with lower E and
a
more negative AS* than certain [2+4] cycloadditions, but

on the average [2+2] cycloadditions have larger E and
a
less negative AS* than [2+4] cycloadditions, whose concerted

transition state favors a more negative AS* and lower E
a
The E and AS* of the competitive [2+2] and [2+4]
a
cycloadditions of difluoroallene are not known. However,

from the average values above and the fact that the rates

of cycloaddition are comparable, certain generalizations





-60-


may be made. The [2+4] reaction should have a more nega-

tive AS* than the [2+2], implying a lower E for the [2+4]
a
reaction, to keep the rates comparable. As an example,

it can be calculated that two reactions whose AS* differ

by 13 e.u. and whose rates, k1 and k2 are in a 2:1 ratio

will have a difference of 5 kcal/mole in their activation

energies at 1100. An increase in he reaction temperature

will help the higher Ea process, in this case the [2+2],

more than the lower Ea process, the [2+4] reaction. Further-

more, the [2+4] reaction with its more negative AS* will

be hurt more than the [2+2] reaction by the higher temper-

ature. Both of these factors should lead to an increase of

[2+2] product over [2+4] product as the cycloaddition

temperature is raised. This is indeed what is observed.

At first glance, the observed orientation of the CF2

group (totally exocyclic, sp2 hybridized) in the majority

of the [2+4] adducts seemed a bit strange. If the cyclo-

additions were concerted, and all the evidence indicates

that they were, then thermodynamic and kinetic control

should have been synonymous. This line of reasoning would

predict [2+4] cycloadducts with endocyclic, sp hybridized

CF2 groups, totally opposite to the observed results.

There is no doubt that the observed and isolated [2+4]

adducts were, in every case the kinetically

controlled products. The cycloadditions were carried out

at temperatures from 00 to 1600, which are rather low to.

cause isomerization of the products by retro [2+4] processes.




-61-


The cyclopentadiene-acrolein adduct, for example, under-

goes retro-cycloaddition between 1920 and 2420 with an

activation energy of 34 kcal/mole, and the cyclopentadiene-
12
maleic anhydride adduct dissociates above 1650.

The Frontier Molecular Orbital (FMO) Theory of
43, 49
Cycloadditions43, 49 can be used to effectively ration-

alize the results of the [2+4] cycloadditions in terms

of a concerted mechanism. Calculations [ab initio (STO-

3G)] carried out by Houk and Gandour on 1,1-difluoroallene

indicate that its LUMO is the C2-C3 i* orbital, not the
100
C -C2 i* orbital. The C2-C3 i and f* orbitals are

lowered in energy, relative to those of allene, by mix-

ing with the CF2 acceptor orbital, while the CI-C2

and T* orbitals are raised by antibonding interactions

with the fluorine lone pairs. The inductive effect

of fluorine acts to lower all orbital energies. The

actual calculated orbital energies are as follows.

The second lowest unoccupied molecular orbital (SLUMO)

is the C -C2 T* orbital with a calculated energy of

-1.4 eV, and the LUMO is the C2-C3 T* orbital with a

calculated energy of -0.8 eV. The HOMO is the C1-C2

r orbital with a calculated energy of -7.61 eV and the

second highest occupied molecular orbital (SHOMO) is

the C2-C3 7 orbital which has a calculated energy of -9.26 eV.





-62-


Experimental confirmation of fluorine's effect on the

molecular orbitals of difluoroallene was obtained from

photoelectron spectroscopy (PES),70 which directly measures

the ionization potential (IP), and thereby the energy, of

the occupied orbitals.01 The PES of difluoroallene

(97% glpc pure) showed the first IP to be due to the C1-C2

T orbital (HOMO). The IP was decreased (9.79 eV), relative

to allene's (10.3 eV). Therefore, the orbital had a higher

energy (orbital energy is the negative of the IP) than the

allene orbital. The second ionization band, due to the

C2-C3 n orbital (SHOMO), was raised to 11.42 eV, relative to

allene's 10.3 eV, indicating a lower SHOMO energy for

difluoroallene. These experimental results are in complete
100
agreement with the ab initio calculations which predicted

a raised C1-C2 T orbital (HOMO) and a lowered C2-C3

orbital (SHOMO), with respect to allene. Besides confirming

the effect of fluorine on difluoroallene's orbitals, the

PES has provided absolute, experimental values for the orbital

energies of the HOMO and SHOMO. The HOMO has an energy of

-9.79 eV and the SHOMO an energy of -11.42 eV.

Theoretical calculations, such as those carried out

on difluoroallene, could be performed to determine the

effect of substituents on FMO energies of dienes and dien-

ophilies. However, the time and expense involved in such

computer-aided calculations precludes them being used for

a large number of cases. Furthermore, for many experimental




-63-


and physical organic chemists the maze of numbers involved

in such calculations is unfamiliar and confusing territory.

For those reasons, Houk developed a set of generalizations,

based mostly on experimental results, about the frontier
43, 49
orbital energies of substituted dienes and dienophiles,43

to be used in a qualitative manner. The effect of fluorine

on the orbitals of difluoroallene could have been predicted

using Houk's generalizations. An examination of these

generalizations will be instructive in further clarifying

the difluoroallene case.

There are three general classes of substituents in

common dienophiles. They are conjugating (C) substituents

such as -CH=CH2, Ph, etc; electron-withdrawing (Z) substi-

tuents like CN, CHO, etc; and electron-donating (X) substi-

tuents such as alkyl, OR, NR2, etc. The effect of the

conjugating substituents on the FMO energies is easily
48
rationalized by looking at a simple case.48 The addition

of a vinyl substituent to ethylene gives the conjugated

1,3-butadiene or vinylethylene system. The relationship

between the molecular orbitals of ethylene and butadiene
48
is well known. The HOMO of butadiene is raised in energy,

relative to the ethylene HOMO, because of the antibonding

between C2 and C3; this antibonding was missing from the

isolated r bonds. The LUMO of butadiene is lowered in

energy, relative to the ethylene LUMO, because of the

bonding between C2 and C3; this bonding was absent from

the isolated 7 bonds. So, C substituents raise the energy

of the HOMO and lower the energy of the LUMO.





-64-


LUMO 7Tn ----- -----* LUMO

LUMO
w0 0




Woo
_,, ___ OMHOMO

HOMO ---- HOMO


AL 0--<


ethylene butadiene ethylene



Electron-donating (X) substituents also affect the
dienophile FMO energies in a simple way.48 For example, in
methyl vinyl ether, a typical X substituted olefin, a lone
pair of electrons is brought into conjugation with the double
bond. A good model for this case is the allyl anion. The
orbital effect of going from ethylene to the allyl anion
can be effectively rationalized as shown below. The HOMO
of the allyl anion is raised in energy, relative to ethylene,
because of the change from bonding (in ethylene) to non-
bonding orbital. Similarly the LUMO of the allyl anion is
raised in energy, relative to ethylene, because of the





-65-


3, \LUMO




*t \
LUMO LUM .




0 ",-
--- HOMO -


/

HOHO -- /







ethylene allyl anion p-orbital


increased antibonding in the 3 orbital, relative to the

c* orbital of ethylene. Thus X substituents raise both

HOMO and LUMO energies.

The electron withdrawing (Z) substituents act to lower

both HOMO and LUMO energies, relative to ethylene. But,

because many of these substituents are both electron with-

drawing and conjugating (e.g., CN, COH, COR) the overall
49
lowering of the HOMO is less than that of the LUMO. This

class of substituent does not easily lend itself to ration-

alization via concrete examples as have the previous two.

However, qualitatively it would be expected that removal





-66-


of electron density from the olefinic bond would decrease

electron-electron repulsion and "stabilize" the bond, i.e.

lower the frontier orbital energies, relative to ethylene.

Of course, the electron withdrawal would make the bond more

reactive to cycloaddition. This is verified by the obser-

vation that the majority of [2+4] cycloadditions take place
. 1, 26
with Z substituted olefins as dienophiles. For 1-

and 2-substituted dienes the substituent effects are similar,

with C substitution raising the HOMO and lowering the LUMO,

X substitution raising both HOMO and LUMO, and Z substitu-

tion lowering both HOMO and LUMO. However, because of

differences in orbital coefficients between the C-l and

C-2 of butadiene systems, the magnitude of energy changes

is smaller in the 2-substituted dienes.

It is now easy to see that "normal" [2+4] cycloadditions,

e.g., between butadiene and acrylonitrile, are between Z

substituted dienophiles and unsubstituted or X substituted

dienes. The reaction is dominated by the highly stabilizing

diene-HOMO-Dienophile-LUMO interaction, because it has the

smallest energy gap. Inverse electron demand [2+4] cyclo-

additions, 26 occur when the diene-LUMO-dienophile-HOMO

interaction is greatest, leading to reaction between X

substituted dienophiles and Z substituted dienes.

It is now possible to qualitatively determine the

effect of fluorine on the orbitals of difluoroallene. The

fluorine substituents act as both electron donors (X) and




-67-


electron acceptors (Z). It is well known57' 61 that

repulsion between non-bonded electron pairs on fluorine,

and i electrons of the carbon-carbon double bond destabilize

fluoroolefins and raise the energy of the system, relative

to the parent hydrocarbon. Difluoroallene is no different

than any other fluoroolefin. The C1-C2 i and 7* orbitals are

raised in energy, relative to those of allene, by anti-

bonding interactions (repulsion) between the fluorine lone

pairs and the C -C2 7 electrons. The C2-C3 orbitals are

not affected by this repulsion, because they are orthogonal

and unconjugated to the C -C2 orbitals. Direct repulsion

between the fluorine lone pairs and the C2-C3 orbitals can-

not occur because the CF2 group is too far away. So,

fluorine acts as an X substitutent and raises the energy

of C1-C2 w and T* orbitals. The electron withdrawing

inductive and field effects of fluorine make it act as a

Z substituent also. The energies of both the C -C2 T and

T*, and C2-C3 I and T* orbitals are lowered. However,

because of the strong repulsion effect on the C -C2 orbitals,

those orbitals are lowered only slightly by the Z effect,

making the net result a raising of the C1-C2 7 and T*

orbitals, relative to allene. On the other hand, the

C2-C3 i and T* orbitals do not feel the repulsion effect,

and the Z effect can effectively lower the orbital energies.

The final arrangement of orbitals on difluoroallene then,

has the C -C2 7 orbital as the HOMO and the C2-C3 T* orbital





-68-


as the LUMO. These are exactly the same results previously

obtained from the ab initio calculations.

Although knowledge of the frontier orbital energies

of the cycloaddends can determine which HOMO-LUMO pair

will dominate the reaction as well as the relative rate of

cycloaddition, to determine the favored regioselectivity

of the products, the frontier orbital coefficients must

be known. Houk has provided generalizations about the

relative magnitudes of the orbital coefficients for C, Z,
43
and X substituted dienes and olefins.4 However, these

coefficients are best determined by calculations.
100
The ab initio (STO-3G) calculations performed on

difluoroallene also yielded the HOMO and LUIMO orbital

coefficients. The LUMO coefficients, 0.73 at C2 and -0.78

at C3 are nearly identical, while the HOMO coefficients are

0.51 at C1 and 0.61 at C2. These results complete the pic-

ture of the difluoroallene frontier orbitals as shown below.



\ .73 .78 H
C = C= C
7I-
F H
LUMO
E = -0.8 eV



S.51 .61 H
C = C C
F H
HOMO

E = -9.79 eV




-69-


It is now quite clear why concerted [2+4] cycloaddi-

tion occurs at the C2-C3 bond of difluoroallene. The di-

fluoroallene is acting as a normal dienophile, and using

its LUMO to interact with the HOMOs of the dienes used.
49 102
The frontier orbital energies for isoprene, butadiene,0

cyclopentadiene,103 furan,104 105 and 2-trimethylsilyloxy-

butadiene43 are shown in Figure 5. The energy of the last
43
diene was estimated from Houk's generalizations. It can

be seen that for all the above dienes, the diene-HOMO-di-

fluoroallene-LUMO interactions are greater than the di-

fluoroallene-HOMO-diene-LUMO interactions.

The frontier orbital energies for hexachlorocyclopenta-

diene have not been reported in the literature. However,

the attempted inverse electron demand cycloaddition with

difluoroallene, which should have resulted in reaction with

the allene's Cl-C2 bond, failed to occur. Instead, a normal

[2+4] cycloaddition at the C2-C3 bond of difluoroallene

formed 29 as the only product. Since inverse electron

demand cycloadditions have dominant dienophile HOMO-diene

LUMO interactions, the formation of 29 implies that, for

hexachlorocyclopentadiene and difluoroallene, the diene-

HOMO-dienophile-LUMO interaction is the dominant one.

The observed near absence of regioselectivity in the

[2+4] adducts of 30b and 30c can now be easily accounted

for. The "para" and "meta" cyclohexenes formed inseparable

mixtures whose composition was determined by capillary

glpc. In the case of isoprene, the products were in a 1.2:1





-70-


6



5/
IL I
I






2
1I
/



/
















-7
S- I/











-9 / -
I /
I /


0 / /


















-10 -
-8 I

\ I

\ I /
















F CF = C=CH
2i2

Figure 5: Frontier orbitals of difluoroallene, furan,
cyclopentadiene and dienes 30a, 30b, and 30c.
cyclopentadiene and dienes 30a, 30b, and 30c.




-71-


ratio and with 2-trimethylsilyloxy-1,3-butadiene they were

formed in equal amounts. Both of the dienes have the largest
43
HOMO coefficient on C-1. The coefficient for isoprene is
49
0.63.49 The coefficient for 2-trimethylsilyloxy-l,3-buta-

diene is not known, but for 2-methoxy-1,3-butadiene, a good

model, the coefficient06 is 0.65. The LUMO coefficients of

difluoroallene are nearly identical, implying a 1:1 ratio of

"para" and "meta" cyclohexenes, with perhaps a very slight

preference for formation of the "para" isomer. For comparison,

CF2

X ( X ,1 C F 2 XC F 2
)I X


CH2 "meta" "para"



30 19 32 31


acrylonitrile, which gives a 2.9:1 ratio of "para" to "meta"

adducts with isoprene at 200, has LUMO coefficients of

0.75 and -0.59 at the unsubstituted and substituted termini,

respectively.50

Evidence for distinguishing the mechanistic alternatives

for the cycloadditions of difluoroallene has been presented.

A totally diradical mechanism cannot explain the results in

a satisfying manner. Rate determining (hence irreversible)

formation of the product-forming diradicals 40, 41, and 42

from cisoid and transoid dienes would result in an unfavor-

able cisoid-transoid equilibrium, which should lead to a





-72-


predominance of [2+2] product, contrary to the observed

results. Isomerization of the transoid diradical to

the cisoid diradical would be unfavorable, due to the

relatively high barrier to rotation in a allylic radical.

Thus, the observed temperature effect on the [2+4]/

[2+2] product ratios implies different activation

parameters for the two processes.

The invocation of steric effects which might favor

exclusive formation of cisoid diradical 41 over 42

is inconsistent with the experimental fact that fluorine

is a small substituent. Examination of molecular models

of 41 and 42 suggested little, if any, conformational

preference for 41 over 42. If a steric effect existed,

it should not preclude formation of at least traces of

cyclohexene 43, in analogy with cycloadditions of methyl-

substituted allenes, which yield both endo- and exocyclic

adducts. In fact, no traces of 43 were detected. Finally,

the observed contrasting regiospecificities of the [2+2]

and [2+4] cycloadditions, with respect to both the loca-

tion of the CF2 group and the orientational preferences

of the diene moiety, make the totally diradical mechanism

an unlikely and unattractive alternative.

A mechanism involving competition between a concerted

[2+4] cycloaddition, and a non-concerted [2+2] cycloaddition





-73-


explained the results in the simplest and most reasonable

manner. The [2+2] cycloaddition results were totally con-

sistent with a reaction involving diradical intermediates.

The known differences in E and AS* of [2+2] and [2+4]
a
cycloadditions could account for the formation of comparable

amounts of both types of products.

The Frontier Molecular Orbital (FMO) Theory of Cyclo-

additions effectively rationalized the [2+4] results in

terms of a concerted mechanism. Ab initio (STO-3G) calcula-

tions carried out on 1,1-difluoroallene indicated that its

HOMO was the C -C2 7 orbital and the LUMO was the C2-C3
100
7* orbital. The fluorine substituents lowered the

energy of the C2-C3 r and n* orbitals, relative to allene.

A photoelectron spectrum of difluoroallene confirmed the

lowering and raising of the T orbitals and provided an

experimental value for the HOMO energy.70 The effect of

fluorine on difluoroallene's orbitals could also be quali-
43, 49
tatively rationalized from Houk's generalizations.43

Difluoroallene acted as a normal dienophile in the [2+4]

cycloadditions. Its LUMO interacted with the diene HOMOs

to give a stronger interaction than the diene-LUMO-difluoro-

allene-HOMO interaction. All of the cycloadditions occurred

at the C2-C3 bond, and yielded products with exocyclic CF2.

The calculated LUMO coefficients for C2 and C3 of difluoro-

allene were nearly identical and accounted for the lack of

regioselectivity in the [2+4] adducts of isoprene and tri-

methylsilyloxybutadiene.





-74-


The cycloaddition reactions of 1,1-difluoroallene have

provided a unique, easily perceivable mechanistic probe,

through which concerted and non-concerted mechanisms may

be distinguished by simple product identification. Valuable

insight into the effect of fluorine substituents on

cumulated system orbitals and the behavior of fluorine

substituted dienophiles has been gained from the use of

FMO theory. The particularly reactive nature of difluoro-

allene in cycloadditions should make it useful in the study

of many other cycloaddition processes.













CHAPTER IV
THERMAL GENERATION OF DIRADICALS

As a means of testing the behavior of diradicals

containing a CF2 group, when they cyclize to six-membered

rings, diradicals were generated by pyrolyses of some [2+2]

adducts of difluoroallene. The static pyrolyses of 33a,

33b, and 34b, and the flow pyrolysis of 33a, resulted in

extensive decomposition attributable to HF elimination via

surface effects. However, a gc-ms study of the pyrolyzates


CH CH 2 CH 2
CH3
F F F
H F CH F H F

33a 33b 34b


did show the presence of one product in small amounts. Its

mass spectrum identified it as a cyclohexene isomer of the

respective cyclobutane, but no decision could be made as to

the actual structure from the available information.

In contrast to the above pyrolyses, the gas-phase,

static pyrolysis of compound 38 was very clean. Two pro-

ducts, 24 and 39, were formed in a 4.1:1 ratio, with complete

disappearance of starting material. Of course, 24 was also

obtained as the exclusive product from the cycloaddition,of

difluoroallene and cyclopentadiene. These pyrolysis results


-75-




-76-


H H F F
CF2 A CF2 H

H
H H H
4.1 : 1

38 24 39






may appear inconsistent with the cycloaddition results, but

they need not be.

Formation of 24 and 39 from pyrolysis of 38 could be

viewed as supportive of the totally diradical mechanism for

difluoroallene cycloaddition. However, a careful look at
107
some results recently published by Hasselmann indicates

that the pyrolysis of 38 is best interpreted by an alterna-

tive explanation.

Hasselmann has studied the thermal rearrangement of

the specifically methyl substituted hydrocarbons 50 to 53.

In each case, 50 to 53 rearrange into one another and into

the four methylated 5-methylenebicyclo[2.2.1]hept-2-enes

54 to 57, with the product of a formal [1.3] sigmatropic

rearrangement being formed in slightly greater yields.

Three mechanistic alternatives were proposed to explain

the results.

The first alternative, competing concerted [3.3] and

[1.3] sigmatropic rearrangments, was eliminated by using




-77-


50 51 52 53









54 55 56 57


activation energy and steric arguments. The second and third

alternatives both involved diradicals. The former was a two-

step mechanism with equilibrated diradicals as intermediates

and the latter was a course involving non-equilibrated di-

radicals or diradical transition states of comparable energy.

A choice between alternatives two and three was made by exam-

ining the product ratio 54/55. The ratio of 54/55 formed

from 52 was 1.33 and 54/55 from 53 was 1.43. If the diradi-

cals formed were equilibrated intermediates, the values of

54/55 from 50 and 51 should have been between 1.33 and 1.43;

however, 54/55 from 50 and 51 was found to be 0.134 and

0.657 respectively. Hasselmann then concluded that formation

of non-equilibrated diradicals offered the best explanation

for his experimental results.

The exclusive formal [1.3] sigmatropic rearrangement

of 38 is not surprising when seen in the light of Hassel-

mann's results.107 As shown on Figure 6, a [3.3] rearrangement




-78-


:CF2


H





H


H-


720 rotation


CF2


-H


1800


252 rotation


rotation


80%


F
H -CH

H H
H

39

20%


Figure 6: Mechanism for the pyrolysis of 38




-79-


would require either passing the CF2 group past the CH2

group of the ring (apparently a relatively unlikely confor-

mational motion) or have the CF2 group go the other way,

where it must somehow be prevented from forming 39 if it is

to get to the [3.3] product 58. Apparently, under these

pyrolysis conditions, 24 and 39 are just the easiest products

to get to.

Invocation of a barrier to passage of the CF2 group

past the CH2 group of the ring might seem paradoxical, in

view of the previous chapter's arguments for the small size

of fluorine substituents.8' 8 However, Hasselmann has

found, by using deuterium, that even in the case of hydrogen

a small barrier exists for passage over the CH2 group of
108
the ring. Since fluorine is larger than hydrogen, a

similar barrier would be expected in the diradical formed

from pyrolysis of 38.

Finally, it must realized that in the pyrolysis

of 38 the diradical formed is a specific one, which forms

24 and 39 because they are the easiest products to get to.

However, in the cycloadditions of difluoroallene, a choice

would exist as to which diradical would form, and the ini-

tially formed diradical would lead to the products most

easily formed from it. This implies cycloaddition products

with little regioselectivity with regards to the CF2 group,

contrary to the observed results.














CHAPTER V
EXPERIMENTAL

All boiling points are uncorrected. Infrared spectra

were determined on both a Perkin-Elmer model 137 and a

Beckman model IR-10 spectrophotometer, and all absorption
-1
bands are reported in cm The spectra of most liquids

were determined as films between NaCl plates while those

of gases were done in a 6 cm gas cell with NaCI windows

at 20-30 mm pressure. Proton magnetic resonance spectra

were recorded on a Varian XL-100 spectrometer unless

otherwise specified; chemical shifts are reported in parts
19
per million downfield from internal TMS. All F NMR

spectra were recorded on the XL-100 instrument at 94.1 IMHz

with chemical shifts reported in parts per million from

external trifluoroacetic acid. Mass spectra were determined

on an AEI-MS 30 spectrometer at 70 eV. Exact mass analyses

were also determined on the AEI-MS 30.

The glpc preliminary analyses and preparative scale

separations were carried out on a Varian Aerograph model

90-P gas chromatograph with thermal conductivity detector.

Glpc yield studies were performed on a Hewlett-

Packard model 5710A gas chromatograph with flame ioniza-

tion detector. The sample composition was determined with

a Vidar Autolab 6300 Digital Integrator.


-80-




-81-


Four columns were used in glpc work and are refer-

enced as follows:

(1) column A 20% SE-30 on Chrom. P regular 15 ft

x 0.125 in copper

(2) column B 20% SE-30 on Chrom. P regular 15 ft

x 0.75 in copper

(3) column C 15% ODPN on Chtom. P regular 20 ft

x 0.75 in aluminum

(4) column D 5% ODPN on Chrom. P regular 20 ft x

0.125 in copper

All reagents which are not referenced were commercially

available.

1,2-Dibromo-3,3,3-triflurorpropane (21)

A 1 liter three-necked flask was fitted with a gas

inlet tube, mechanical stirrer and reflux condenser topped

off by a small dry ice condenser and drying tube. Bromine

(453 g, 2.8 moles) was added to the flask and stirring

begun. A standard Sears sun lamp was set up to shine on

the flask. Then 3,3,3-trifluoropropene (20) was bubbled

in at a moderate rate. The reaction flask was cooled with

an ice bath for the first hour of gas addition. The tri-

fluoropropene was bubbled in for five hours until the

solution was almost colorless. The sun lamp was then turned

off, gas addition stopped and the reflux condenser replaced

with a set-up for simple distillation. The gas inlet tube

was replaced by a stopper. An oil bath was placed around





-82-


the flask and distillation begun. This yielded 614 g

(85%, based on bromine) of the propane 21 as a colorless,

dense liquid: bp 113-1140; ir 3040, 2990, 1430, 1365,

1340, 1275, 1250, 1210, 1180, 1140, 1110, 1050, 1030 (weak),

965, 915, 900, 860, 755, 720, 670, 660, 630; mass spectrum

m/e (rel intensity) 256 (1.6), 254 (0.8), 177 (94.6), 175

(100), 173 (4.0), 171 (2.0), 131 (6.0), 129 (6.3), 113

(17.5), 111 (18.2), 82 (1.5), 80 (1.2), 69 (29.6), 51 (11.2).

28 (23.6); exact mass calcd for C3H3Br2F3: 253.85520

found: 253.85506; nmr (CC14) 1H 6 4.58-4.20 (m, 1H), 4.04-

3.52 (m, 2H); F 6 7.81 (d, 3F).

Trifluoroisopropenyl Bromide (22)

This compound was prepared using the procedure of

Drakesmith, Stewart, and Tarrant.6

1,1-Difluoroallene (19)

The procedure used was a modification of that reported

by Drakesmith, Stewart, and Tarrant.66 An oven-dried 500 ml

three-necked flask was equipped with a magnetic stir bar,

low temperature thermometer, gas dispersion tube and

parallel adapter. On the parallel adapter were placed a

200 ml vacuum jacketed, pressure equalizing dropping funnel,

and a standard Friedrichs condenser. The condenser was

connected to a 100 ml stopcock-equipped vacuum trap by means

of glass tubing. The trap outlet was connected to a U-tube

mineral oil bubbler. Line nitrogen was passed through

concentrated sulfuric acid and a "Drierite" drying tube

before entering the reaction flask. The gas dispersion




-83-


tube was used to introduce the nitrogen into the system.

The entire system was flushed with nitrogen for one hour.

The nitrogen flow was then reduced and the flask charged

with trifluoroisopropenyl bromide (35 g, 0.2 mole) in 150 ml

anhydrous ether (freshly opened can). n-Butyl lithium

solution (15% in hexanes, 85.8 g of solution, 12.8 g n-BuLi,

0.2 mole) was placed in the dropping funnel then cooled

to -78. The trap was also cooled to -780, and ice water

was circulated through the reflux condenser by a small

mechanical pump. An insulated crystallizing dish half

filled with heptane (tech. grade) was placed around the

reaction flask. Then the flask was slowly cooled to -900

by slow addition of liquid nitrogen to the heptane.

Dropwise addition of the cooled n-butyl lithium solution

was begun. The reaction temperature was kept at -90o50

during the entire addition. As the n-butyl lithium was

added, the reaction mixture darkened in color. After the

n-butyl lithium addition was complete, the reaction mix-

ture was kept at -900 for 15 min then allowed to warm to

room temperature. A slow stream of nitrogen was bubbled

through the solution to remove dissolved difluoroallene.

Warming of the reaction mixture caused it to become even

darker in color and lithium fluoride precipitated out

as a white fluffy solid. After room temperature was

reached, nitrogen was passed through the solution for an

additional five hours. The crude overgases collected in.

the trap were analyzed by gas ir and shown to contain




-84-


ether, 1,1-difluoroallene and small quantities of 3,3,3-

trifluoropropene. A distillation of this material was

performed in the following manner. The crude overgases

were transferred on the vacuum line to a storage vessel.

It was connected in series with three traps using glass

adapters for the connections between traps. The entire

system, up to the stopcock of the storage vessel, was

then evacuated. The storage vessel was connected to the

first trap, which was cooled to -780. The second and third

traps were cooled to -1960 and the vacuum pump was connect-

ed to the third trap. After the system pressure had gone

down to 0.025 mm, the stopcock on the storage vessel

was opened and its cooling bath taken away. The crude

overgases were allowed to evaporate from the storage

vessel at a moderate rate. Occasional cooling of the ves-

sel was necessary. After the storage vessel was empty,

the trap contents were checked by gas ir. Trap 1 was found

to contain mostly ether with some 1,1-difluoroallene. In

trap 2 there was a mixture of 1,1-difluoroallene, ether,

and 3,3,3-trifluoropropene. Trap 3 was empty. The con-

tents of trap 1 were redistilled to give two fractions.

The first one contained only ether and was discarded.

The second one was combined with the contents of trap 2

from the first distillation. This material was then put

through the distillation process several times until the

gas ir indicated no ether present. A gas ir of this





-85-


material is identical to that reported for 1,1-difluoro-

allene64 with peaks also appearing for 3,3,3-trifluoro-

propene. A gas sample of the final distillate was taken

on the vacuum line, diluted with argon and analyzed by

glpc (column A at 800). The average of three analyses

gave the following distillate composition: 10% CF3CH=CH2,

75% CF2=C=CH2, 2% ether, 9% CF2H-CEC-H, 4% other products,

none greater than 1% concentration. Further distillation

of this material was found not to significantly increase

the difluoroallene (DFA) content. The 75% DFA mixture

was used in most reactions of 1,1-difluoroallene.

GLPC Purification of 1,1-Difluoroallene

The difluoroallene volatiles which contain 75% DFA

were transferred, on the vacuum line, to a vessel of the

following design: A small (%50 ml) round bottom flask with

septum port for 7 x 11 mm septums and 2 mm high-vacuum

stopcock with T 10/30 male joint. An atmosphere of argon

was bled into the vacuum line and the vessel. The vessel

stopcock was closed and the vessel stored at -780 during

the entire purification procedure. The glpc purification

was carried out using column B at 300. The volume injected

varied but was always between 50 and 100 microliters. The

material to be injected was removed from the vessel with a

cooled syringe. To cool the syringe, a glass sleeve was

placed around the barrel and filled with powdered dry ice.

Once the sample was in the syringe, the syringe was

quickly withdrawn from the vessel, removed from the sleeve





-86-


and then the sample was injected into the gc. The difluoro-

allene peak was collected using a spiral trap cooled to

-78. The material thus collected was glpc analyzed and

found to have 97% difluoroallene content. It was used for
13 67 68 64, 69
spectroscopic studies. The C nmr, microwave, ir, '
64 69
mass, and Raman spectra of difluoroallene have been

reported in the literature. The h nmr (CC14) gave a
19
triplet at 6 6.04 while the F nmr showed a triplet at

6 -26.7. The coupling constant was J = 3.5 Hz. The

photoelectron spectrum, as determined by Houk and Domel-
70
smith, showed a first ionization band at 9.79 eV and

a second ionization band at 11.42 eV. The vibrational

spacings observed were, in order of decreasing intensity:


9.79 eV band 11.42 eV band
-i -i
1361 cm1 1215 cm1

1832 410

897 712

%983


Reaction of 1,1-Difluoroallene and 1,3-Butadiene

A medium-wall glass ampule of approximately 20 ml

capacity was oven-dried, then cooled under nitrogen. A

few crystals of hydroquinone were then added. The ampule

was connected to the vacuum line and evacuated. 1,3-Buta-

diene (2.66 g, 0.0493 mole) was measured out on the

vacuum line and condensed into the ampule. Similarly,





-87-


1,1-difluoroallene volatiles (>50% DFA content) were

measured out to give 0.0197 mole of gas which was condensed

into the ampule. The ampule was sealed under vacuum and

placed in a stirred, preheated (1100) oil bath. After

7 hours, the ampule was removed from the oil bath, cooled

to -78, and opened. The product mixture was analyzed

and the components isolated by glpc (column C at 70).

There were two main components in a ratio of 1.7:1. The

earlier eluting component was the cyclohexene 31a: ir 3040,

2980, 2930, 1765, 1440, 1350, 1335, 1315, 1275, 1255, 1220,

1210, 1075, 990, 940, 865, 770, 660; mass spectrum m/e

(rel intensity) 130 (100), 129 (13.2), 127 (16.2), 115

(82.9), 110 (3.9), 109 (19.2), 95 (10.6), 79 (68.3),

77 (34.6), 76 (3.7), 54 (16.0), 51 (23.4), 39 (33.8);

exact mass calcd for C7H F2: 130.05930 found: 130.05919;

nmr (CDC13) H 6 5.72 (m, 2H), 2.70 (m, 2H), 2.20 (m, 4H);
19
F 6 -19.0 (AB quartet, 2F, J = 61 Hz). The latter

component was the cyclobutane 33a: ir (gas) 3100, 2995,

1835 (weak), 1755 (weak), 1640 (weak), 1435, 1380 (weak),

1300, 1205, 1165, 1115, 1085, 1045, 990, 920, 770, 720; mass

spectrum m/e (rel intensity) 130 (2.0), 129 (12.6), 115 (100),

110 (8.3), 109 (22.3), 95 (8.5), 90 (36.0), 79 (58.2),

77 (18.1), 76 (4.4), 54 (52.4), 53 (20.9), 39 (74.4); exact

mass calcd for C7H F2: 130.05930 found: 130.05921;

nmr (CDC13) H 6 6.14 5.73 (m, 1H), 5.64 5.44 (m, 1H),

5.38 5.02 (m, 3H), 3.68 3.12 (m, 1H), 3.04 2.22 (m, 2H);
19 quartet, 2F, = 212 Hz)
F 6 -21.7 (AB quartet, 2F, J = 212 Hz).
AB





-88-


Preparation of the Calibration Curve for the Reaction of
1,1-Difluoroallene and 1,3-Butadiene

A small vial was tared and a measured amount of

n-octane (Chem. Serv. Pract. Grade) added to it. A mea-

sured amount of cyclohexene 31a was added to the octane.

The weights of octane and 31a were recorded and the mixture

analyzed by glpc (column A at 800). In this manner, three

samples containing different 31a/octane weight ratios were

prepared and analyzed. After the analysis, the 31a/octane

area ratios were calculated and a plot made of area ratios

vs weight ratios. The data points were plotted and the

best straight line drawn through them. This plot could then

be used to determine the weight ratio for a known area

ratio. In turn, the weight of 31a for a given 31a/octane

weight ratio could be determined when the weight of octane

was known. From this, the reaction yields were determined.



Sample Weight of Weight of Wt. ratio Area ratio
n-octane 31a 31a/octane 31a/n-octane

1 17.6 mg 18.4 mg 1.04 .74

2 10.2 mg 19.7 mg 1.93 1.2

3 4.4 mg 22.5 mg 5.11 3.2



The line drawn from the above data was checked for fit

on a linear regression-least mean squares program using a

Texas Instruments TI-57 calculator. The measured correla-

tion coefficient was 0.9995.




-89-


1 2 3 4 5


Weight ratio


31a/octane


Figure 7: Calibration curve for the reaction of 1,1-
difluoroallene and 1,3-butadiene.




-90-


Yield Determination in the Reaction of Difluoroallene and
1,3-Butadiene

A medium-wall glass ampule of approximately 7 ml

capacity was oven-dried and cooled under argon. A few crys-

tals of hydroquinone were added and 37 mg of n-octane

(Chem. Serv. Pract. Grade) was carefully syringed in. The

ampule was attached to the vacuum line, cooled, evacuated,

and degassed. 1,3-Butadiene (141 mg, 2.6 mmoles) was mea-

sured out on the vacuum line and condensed into the ampule.

Similarly, difluoroallene volatiles (75% DFA content) were

measured out to give 71.3 mg (.938 mmoles) of difluoroallene

which was condensed into the ampule. The ampule was sealed

under vacuum and placed in a stirred, preheated (1100) oil

bath. After 423 minutes, the ampule was removed from the

oil bath, cooled to -78, and opened. The contents were

analyzed by glpc (column A at 80). With the help of the

previously prepared calibration curve, it was found that

cyclohexene 31a composed 63% of the total product and was

formed in 55% yield (59 mg). The cyclobutane 33a composed

37% of the total product and was formed in 32% yield (35 mg).

The total yield of these two products was found to be 87%

(94 mg). The reaction had gone to 89% conversion (from un-

reacted difluoroallene). The products 31a and 33a accounted

for 89% of all observed glpc "product" peaks.

Determination of relative yields of products at various

reaction times were also carried out. Six ampules were

prepared, each containing approximately the same amounts




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

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