Experimental and theoretical investigation of the reactivity of partially fluorinated radicals


Material Information

Experimental and theoretical investigation of the reactivity of partially fluorinated radicals
Physical Description:
vii, 186 leaves : ill. ; 29 cm.
Bartberger, Michael David, 1970-
Publication Date:


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


Thesis (Ph.D.)--University of Florida, 1998.
Includes bibliographical references (leaves 175-184).
Statement of Responsibility:
by Michael David Bartberger.
General Note:
General Note:

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 029218667
oclc - 40097324
System ID:

Full Text








Among the great number of individuals with whom I have interacted throughout

the course of my education at the University of Florida and elsewhere, I wish to express

my sincere appreciation to the special few that have motivated, challenged, and inspired


I extend my deepest gratitude to Prof. William R. Dolbier, Jr., an outstanding

scientist and truly exceptional educator, for his excellent guidance, support, and

friendship throughout the course of my graduate career. My appreciation for the

knowledge he has shared with me, as well as his patience and level of understanding,

particularly during periods of difficulty and stress, can not be overstated.

I wish to thank the two finest classroom instructors I have ever had-my first

college level chemistry teacher, Dr. Jeanette Madea, for her profound influence in my

decision to pursue a career in the chemical sciences, and Prof. Seth Elsheimer, for my

initial exposure to the fascinating area of organofluorine chemistry in 1990 and his

friendship thereafter.

I am indebted to Dr. Max Muir for introducing me to computational chemistry.

The experience I have gained in the use of molecular orbital methods as a tool for the

understanding of chemical reactivity is due entirely to him. Special thanks go to Prof.

Benjamin Horenstein for his helpful discussions and generosity with regard to

computational resources.

I thank my colleagues, past and present, in both the Dolbier group and the

Department of Chemistry as a whole. A few bear special mention-Dr. Keith Palmer, for

his friendship and advice during my first year in the group; Dr. Xiao Xin Rong and He-Qi

Pan, for their camaraderie and early assistance with radical kinetics; Dr. Conrad

Burkholder, for numerous stimulating discussions; and Dr. Henryk Koroniak, Michelle

Fletcher, Lian Luo, Feng Tian, and Kevin Ley for their friendship (and tolerance!)

throughout the course of my stay in the department.

I wish to thank my graduate committee, particularly the "organic" portion thereof--

Profs. Merle Battiste and Kirk Schanze, for their advice and encouragement. Also,

special thanks go to Prof. R. J. Bartlett for taking seriously my interest in theoretical

methodology and the invitations to participate in his workshops on Applied Molecular

Orbital Theory.

I am especially grateful to my very best friend, Cynthia Dawn Zook, for her

unrelenting moral support and encouragement over the last several years. Finally, I

wish to acknowledge my parents, George Charles and Beverly Jean, for instilling in me

the work ethic which has likely had as much to do with the successful completion of this

work as any of the chemistry I ever learned.


ACKNOW LEDGEMENTS ............................................................................................... ii

ABSTRACT.................................................................................................................... vi



Introduction .............................................................................................................. 1
Radical Chain Processes .......................................................................................... 2
Hydrogen Atom Abstraction Reactions...................................................................... 4
Intermolecular Radical Addition Reactions ................................................................ 9
Intramolecular Addition Reactions: Radical Cyclizations.......................................... 13
Methods for Determination of Organic Radical Kinetics........................................... 22
Conclusion .............................................................................................................. 26

2 THE FLUORINE SUBSTITUENT IN ORGANIC SYSTEMS................................. 28

Introduction ............................................................................................................. 28
Structure, Bonding, and Reactivity in Saturated Systems........................................ 29
Structure, Bonding, and Reactivity in Unsaturated Systems.................................... 31
Fluorine Non-Bonded Interactions in Reactive Intermediates.................................. 33
Fluorine Steric Effects .............................................................................................35
The Fluorine Substituent in Free Radicals............................................................... 37
Organofluorine Radical Reactivity ........................................................................... 40
Conclusion .............................................................................................................. 48

REACTIONS .......................................................................... .......... .. ............. 49

Introduction ............................................................................................................. 49
Precursor Syntheses and Competitive Kinetic Studies ............................................ 50
Discussion ............................................................................................................... 61
Conclusion .............................................................................................................. 76

INTRAMOLECULAR CYCLIZATION REACTIONS ..................... ............................ 77

Introduction ............................................................................................................. 77
Precursor Syntheses and Competitive Kinetic Studies............................................ 78
Discussion ............................................................................................................... 89
Conclusion ............................................................................................................ 100

5 EXPERIM ENTAL................................................................................................... 102

General Methods- Experimental............................................................................ 102
General Methods- Theoretical ............................................................................... 103
Synthetic Procedures ............................................................................................ 103
Com petitive Kinetic Procedures............................................................................. 129

APPENDIX A: SELECTED 19F NM R SPECTRA.......................................................... 136

DATA IN TABLES 3-3 AND 3-4............................................................................. 172

REFERENCES............................................................................................................ 175

BIOGRAPHICAL SKETCH .......................................................................................... 185

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



Michael David Bartberger

May 1998

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

The reactivities of a series of partially-fluorinated radicals towards intermolecular

addition, hydrogen abstraction, and intramolecular cyclization have been investigated.

Based on competitive kinetic techniques and absolute rate contents for addition of these

radicals to styrene obtained by laser flash photolysis, absolute rate constants for

abstraction of hydrogen from tributylstannane have been determined for 1,1-

difluoroalkyl, 2,2-difluoroalkyl, 1,1,2,2-tetrafluoroalkyl, 3-perfluoroalkyl, and

pentafluoroethyl radicals. Fluorination at the 3-position of an alkyl radical was found to

exert a negligible effect on the kinetics of hydrogen abstraction. All other systems

exhibit rate enhancements relative to non-fluorinated analogues, the magnitudes of

which are dependent upon the degree and location of fluorine substitution. A parallel

computational study was performed utilizing density functional calculations, providing

estimates of carbon-carbon and carbon-hydrogen bond dissociation energies (BDEs) for

hydrofluorocarbons. The observed kinetic enhancements were attributed to a

combination of structural, charge transfer, and enthalpic effects, due to the pyramidal

nature of 1-fluoralkyl radicals, increased electrophilic character induced by successive

fluorination, and thermodynamics of carbon-carbon and carbon-hydrogen bond

formation. From the computation of partial atomic charges in fluoroalkanes, the

contrasting effect of 1-fluorination on carbon-carbon and carbon-hydrogen BDEs and

the consistent strengthening effect of such substitution at the 2-position have been

explained on the basis of Coulombic interactions.

Based on the rate constants obtained for hydrogen abstraction, absolute rate

constants for 5-exo and 6-endo intramolecular cyclization for a series of partially

fluorinated 5-hexenyl radicals have been obtained. These observed rates of cyclization

may be rationalized by the same combination of effects influencing their bimolecular

addition reactions. In some cases, the rate of 6-endo closure is dramatically

accelerated relative to the parent hydrocarbon without the introduction of reversibility of

ring closure.




The discovery of the first free radical, triphenylmethyl, by Moses Gomberg1 in

1900 initiated considerable effort directed toward the understanding of radical reactivity.

However, only after a series of pioneering investigations undertaken more than thirty

years later were the primary mechanistic pathways available to organic free radicals well

elucidated.2- These studies, most notably those of Kharasch et al.,47 demonstrated that

most radical processes can be expressed in terms of a small number of elementary

steps, or variations thereof, as shown below in Figure 1-1.8

A" + B A-B coupling / homolysis (1-1)

A* + B-D A-B + D substitution (SH2) (1-2)

A* + B=D A-B-D addition / p-fission (1-3)

A' + e A- ; electron transfer (1-4)

A' e -- A+

Figure 1-1. The Elementary Mechanistic Pathways of Free Radical Reactions.

Despite these breakthroughs, organic free radical reactions continued for years

to be regarded as unpredictable, unselective, and in general inadaptable to synthetic

application. Fortunately, subsequent kinetic and thermochemical studies have served to

uncover the factors governing the reactivity of organic radicals, and consequently in

recent years sentiment toward the utility of free radicals in synthesis has drastically

changed. Indeed, the number of elegant works in the literature based on radical

mediated transformations is a testament to their applicability in the construction of

natural products and other complex synthetic targets.915 It is the purpose of this

introductory chapter to acquaint the reader with the fundamental types of free radical

processes which occur in organic systems, as well as to provide an overview of the

wealth of physical studies which have given rise to the current level of understanding of

organic radical reactivity.

Radical Chain Processes

Most free radical reactions occur via a sequence of chain events, propagated by

intermediate steps during the course of the reaction. An example illustrating a

competition between two potential pathways is provided in Figure 1-2.

Initiation: In-In 2 In (1-5)

In* + M-H In-H + M" (1-6)

Propagation: M" + R-X -- M-X + R" (1-7)
R" + M-H H-- R-H + M' (1-8)
R" k -- R'" (1-9)

Propagation: R" + M-H -- R'-H + M- (1-10)

Figure 1-2. Radical Chain Process Involving Competition Between Rearrangement
versus Hydrogen Atom Transfer from a Donor Molecule M-H.

Homolysis of an initiator, typically accomplished by thermal or photochemical

means, provides a source of (often metal centered) radicals M" (equation 1-6) from

which intermediate radicals R are generated by reaction with a suitable precursor R-X

(equation 1-7). This species encounters one of two fates: trapping, in this case by

hydrogen atom donor, to yield R-H (equation 1-8) or transformation via a unimolecular or

bimolecular process (equation 1-9) to form radical R", itself then trapped producing

R'-H. In either case, additional metal radicals are formed and the chain process

continued via the propagation steps given in equations 1-7, 1-8 and 1-10.

The distribution of products R-H and R'-H is governed by the relative propensity

of R* toward rearrangement versus trapping (that is, kr and kH,) the latter dependent on

both the nature of R" and the type of trapping agent employed. In systems where

trapping is fast relative to rearrangement (kH >> kr), the partitioning radical R is

converted to R-H with little or no rearranged product. However, if kH and kr are of

comparable magnitude, product mixtures result. An understanding of the reactivity of a

radical intermediate toward potential competing processes is therefore essential for the

design of useful kinetic experiments, as well as for the development of effective synthetic


For an efficient chain process, it is necessary that the propagation steps are

rapid relative to chain termination steps, thereby maintaining a low but constant

concentration of radical intermediates. Besides the obvious practical benefit (higher

product yields) resulting from such a condition, the occurrence of undesired chain

termination side reactions such as disproportionation and radical-radical coupling,

possibly complicating kinetic analyses, is minimized. In many cases, this may be

achieved by judicious selection of the type and concentrations of precursor R-X and

trapping agent.

This procedure enjoys wide application in both kinetic and synthetic studies

requiring the controlled generation of radical intermediates. One of its variants, likely the

most commonly used procedure for the indirect (competitive) determination of the rates

of organic radical reactions, is based on the trialkylstannane reduction of an alkyl halide

(the 'Tin Hydride Method").16-18 Other modifications of this general procedure exist,

accommodating a variety of radical precursors and trapping agents; a discussion of time-

resolved and competitive techniques utilized in radical kinetic measurements is provided

later in the chapter.

Finally, it is important to note the implication of kinetic control in the above

discussion. That is, that the distribution of products R-H and R'-H may be ascribed to

the relative values of kH and kr hinges on the absence of any thermodynamic

equilibration of products under the reaction conditions. This is of vital importance in the

design and interpretation of competitive kinetic studies and is discussed in detail in

Chapter 3.

Radical reactivity is dependent on the "complex interplay" of thermodynamic,

steric, and polar considerations.19 The relationship between enthalpies of activation and

heats of reaction, the basis of the thermochemical kinetic approach of Benson,20 was

recognized early on and holds for a number of radical addition and substitution

reactions, where the order of reactivity often parallels exothermicity.21-23 This relation

has led to such overgeneralizations as "radical reactions follow the most exothermic

available pathway" or ". . afford the most stable possible product."8 However, reaction

thermochemistry is not the sole, nor even predominant decisive factor in the outcome of

radical reactions. Nonbonding interactions and the electronic influence of substituents in

ground and transition states (which may be rationalized in terms of Frontier Molecular

Orbital (FMO) theory)24'25 will also play a role. A discussion of the combination of these

effects as manifested in hydrogen atom abstractions and inter- and intramolecular

additions, the most commonly occurring and well-characterized reactions of organic free

radicals, will now be presented.

Hvdrogen Atom Abstraction Reactions

The vast majority of free radical applications involve the use of an organometallic

hydride of the type R3M-H (most commonly, where M = Sn, Si, or Ge) as a hydrogen

atom donor and chain propagation agent, the properties of which have been the focus of

extensive investigation by kineticists. Metal-hydrogen bond dissociation energies

(BDEs) along with activation parameters and associated absolute rate constants for


hydrogen atom transfer to n-alkyl hydrocarbon radicals by a series of donors R3M-H

have been determined and are provided below in Table 1-1 .16, 26-32

Table 1-1. Bond Dissociation Energies with Activation Parameters and Rate Constants
for Hydrogen Atom Transfer to Hydrocarbon Radicals by R3M-H.

R3M-H BDE. kcal mol1 log A E, kcal mol'1 k. _,10 M s' (298 K)

nBu3SnH 73.7 9.06 3.65 2.3

(TMS)3SiH 79.0 8.86 4.47 0.38

nBu3GeH 82.6 8.44 4.70 0.093

(TMS)2Si(CH3)H 82.9 8.89 5.98 0.032

Et3SiH 90.1 8.66 7.98 0.00064

Analysis of the data demonstrates that for hydrogen atom abstraction by

structurally similar radicals from this series of donors, a direct relation holds between the

rate of transfer and the strength of the metal-hydrogen bond being broken. This is

depicted graphically in Figure 1-3. In addition, it is noted that in each case the pre-

exponential term in the Arrhenius relation remains relatively constant. Thus, the rate

variations within the series are due almost entirely to differences in activation energies.

16 -
16m = -0.50403
S14- b =52.313
4r2 = 0.96858


8 -
6 '-

72 74 76 78 80 82 84 86 88 90 92

BDE, kcal mol"1

Figure 1-3. Plot of In kH for Alkyl Radicals versus M-H Bond Dissociation Energies for
Hydrogen Atom Donors R3M-H in Table 1-1.

However, as previously mentioned, relative thermodynamics is not the only factor

which influences the kinetics of hydrogen atom transfer. The fast donor thiophenol

(PhSH) reacts with primary alkyl radicals with a rate constant of 1.36 x 108 M"' s1 at

298 K,33 and has been employed as a trapping agent in competitive kinetic studies

involving strained or otherwise highly reactive radicals with rearrangement rates upward

of 101" s1 and thus with lifetimes on the picosecond timescale.34 This enhanced rate of

transfer, not commensurate with its S-H BDE of 82.0 kcal mol1,35 gives rise to a severe

deviation from the plot in Figure 1-3 and indicates the presence of other influences.

Table 1-2. Absolute Rate Constants for Hydrogen Atom Transfer to tert-Butoxyl
Radicals by R3M-H.

R3M-H kH. 106 M1 s-' (300 K)

nBu3SnH 220

(TMS)3SiH 110

nBu3GeH 80

Et3SiH 5.7

Further evidence may be found in the rates of hydrogen atom transfer to tert-

butoxyl radicals by the same series of hydrogen atom donors, provided in Table 1-2 and

illustrated graphically in Figure 1-4.3637 It is observed that for tert-butoxyl radicals, rates

of hydrogen abstraction are at least two orders of magnitude greater than those of their

n-alkyl counterparts. Although the relative strengths of the newly formed C-H or 0-H

bonds will certainly play a role, the difference in BDE between tBuO-H and n-alkyl C-H

bonds (105 and 100 kcal mol1, respectively)38 is not sufficient to explain the increase in

reactivity, especially in light of the fact that such rapid hydrogen atom abstractions

should proceed with early transition states.36

At this time, the absolute rates of reduction of tert-butoxyl radicals by thiophenol

have yet to be determined. However, a series of competition studies by Hartung and

Gallou39 involving 4-pentenyl-l-oxy radicals and utilizing naphthalene 2-thiol (NpSH) as

a trapping agent have determined a ratio [ kH (NpSH) / kH (nBu3SnH) ] of 1.4. By

comparison, n-alkyl radicals afford the ratio [ kH (PhSH) / kH (nBu3SnH) ] = 59.1.

Although a leveling effect may be partly responsible for the compressed ratio of rates for

tert-butoxyl radicals (which are indeed within an order of magnitude of the diffusion-

controlled limit)40 it is logical to assume based on the aforementioned examples that

hydrogen abstraction reactivity will be governed to some extent by factors other than

simple relative BDE values of the donor species.

2Om = -0.22212
19- b =35.942
2'~~r ="S^r 0.91532
0 18



15 1 T -------------
72 74 76 78 80 82 84 86 88 90 92
BDE, kcal" mol[1

Figure 1-4. Plot of In kH for tert-Butoxyl Radicals vs. M-H Bond Dissociation Energies for
Hydrogen Atom Donors R3M-H in Table 1-2.

Chatgilialoglu et al.3637 have attributed such differences in reactivity to a

polarized, or charge separated, transition state of the type depicted in Figure 1-5. Here

it can be seen that in the case of an electropositive metal hydride donor, hydrogen atom

transfer to alkoxyl radicals (b) is facilitated by greater stabilization of partial negative

charge on oxygen relative to carbon, with a resultant decrease in activation barrier.

*+ 6- 8* 8-
(a) R3M ----- H ----- R (b) R3M ----- H --- OR

Figure 1-5. Charge Polarized Transition State for Hydrogen Abstraction from R3M-H by
(a) Alkyl and (b) Alkoxyl radicals.

In the case of thiol donors, the opposite situation ensues. The greater

electronegativity of sulfur relative to tin (or other metal atom) gives rise to a reversal in

the transition state charge distribution. This arrangement, involving a partially negatively

charged sulfur atom, is better suited to the more nucleophilic alkyl radical, where in the

alkoxyl case a less- or non-polarized transition state results. Such a mismatch in the

latter is partially responsible for the decrease in rate enhancement for hydrogen

abstraction from thiols by alkoxyl radicals, relative to their alkyl analogues.

Frontier Molecular Orbital Theory of Atom Abstraction Reactions

FMO theory provides a satisfying rationale for the kinetic characteristics of

hydrogen abstraction reactions of free radicals. In general terms, radicals are species

containing an unpaired electron in a singly occupied molecular orbital (SOMO), which in

the ground state of the radical is its highest occupied orbital. According to the FMO

concept, during the course of the reaction this SOMO will interact with both the highest

occupied (HOMO) and lowest unoccupied (LUMO) molecular orbitals of the donor

molecule. Such interactions between these "frontier molecular orbitals"24 are not

necessarily equal. The extent of SOMO-HOMO and SOMO-LUMO interaction is

governed by their energy values, the strongest interaction occurring between orbitals

closest in energy.

It is these values, influenced by atom type as well as neighboring substituents,

from which the relative descriptors such as "nucleophilic" and "electrophilic" are derived

and provide the basis for the previously described concept of "polar factors." Electron

donating substituents generally serve to raise both HOMO and LUMO energies, with

electron withdrawing groups resulting in lowering. Radicals possessing a low energy

SOMO will display electrophilic character, whereas a higher energy SOMO gives rise to

a more nucleophilic species. Figure 1-6 depicts the FMO interactions between radical

and donor in each of these cases.

During the abstraction process, the primary interaction involves the radical

SOMO and the a and C* orbitals of the donor M-H bond. The antibonding a* orbitals of

the donor are typically quite high in energy, and thus in atom abstraction reactions the

SOMO-HOMO interaction dominates.


-- HOMO ', ^-- HOMC

(a) (b) *

Figure 1-6. FMO Diagram Illustrating the SOMO-HOMO Interaction Between (a) a
Nucleophilic Alkyl and (b) an Electrophilic Alkoxyl radical.

Here it is seen that the lower-energy SOMO of the alkoxyl radical (ca. -12 eV, as

determined from ionization potential measurements)25 leads to a reinforced interaction

with the donor HOMO (case b). This greater stabilizing interaction results in a lowered

activation barrier and hence a more facile transfer reaction, compared to the more

nucleophilic alkyl radical, (case a) whose SOMO energies range from -6.9 to -9.8 eV.25

Some of the most striking examples of such "polar" factors involve systems

where fluorine substitution has taken place at, or adjacent to, the radical center. This is

elaborated upon in Chapters 2 and 3, where the effects of fluorination on the structure

and reactivity of free radicals are discussed and compelling evidence provided based on

kinetic studies of hydrogen transfer to such partially and fully fluorinated alkyl radicals.

Intermolecular Radical Addition Reactions

Over the past twenty years, the intermolecular addition reactions of free radicals

(as well as their intramolecular cyclization counterparts) have become an important

addition to the arsenal of C-C bond formation methods available to the synthetic organic

chemist. Their mild means of generation from a variety of precursors and tolerance for a

wide variety of functional groups provide distinct advantages over ionic processes.

Alkyl radical additions to carbon-carbon double bonds are highly exothermic, as a

new a bond is formed at the expense of a n bond (in the case of methyl radical addition

to ethylene (Figure 1-7), AHrxn= ca. -22.6 kcal mol-1).41'42 In accordance with the

Hammond postulate,43 such additions should proceed via early transition states, with low

barriers of activation. This is indeed the case, as supported by a wealth of both

experimental44 and theoretical4,1'4549 data.

CH; + CH2=CH2 CH3CH2CH2"

Ea = 7.9 kcal molr1
AHrxn = -22.6 kcal mol-r1

Figure 1-7. Addition of Methyl Radical to Ethylene, Yielding n-Propyl Radical.
Experimental Activation Energies and Heats of Reaction are Shown.

FMO Theory of Radical Additions

Quantum mechanical molecular orbital calculations at levels of ab initio theory

ranging from UHF to UQCISD(T) and varying basis sizes from 3-21G to 6-311G(2df,p)

are consistent in their characterization of the transition structure for the above reaction

G I::pyr(CH3)= 101.9

2.246 A 109.1o 173.40

S1.382 A

Figure 1-8. UHF/6-31G(d) Transition Structure and Relevant Geometrical Parameters
for Addition of Methyl Radical to Ethylene.







' -

,'4| HOMC

(a) (b) -

Figure 1-9. FMO Diagram for Addition of (a) Nucleophilic and (b) Electrophilic Radicals
to Alkenes.

Table 1-3. Some Relative Rates of Addition of Methyl and tert-Butyl Radicals to Alkenes








KreL I (3



KeL( ^^03 3C






a Data for ethyl radical.

(Figure 1-8), which possesses an incipient C-C bond distance of ca. 2.23 2.27 A. The

C-C-C attack angle of 109.1 is rationalized in FMO terms based on a primary interaction

between the radical SOMO and the LUMO of the alkene. It is in such reactions with high

exothermicities and early transition states that FMO interactions are most

substantial.2450 This postulate enjoys experimental support; for the t-butyl radical, a

correlation exists between rates of addition to alkenes and the experimentally

determined electron affinities of the latter.21'51 Such an FMO interaction for nucleophilic

radicals is shown in Figure 1-9 (case a) and is influenced by substituents on both radical

and olefin, a raising of radical SOMO and/or lowering of alkene MO energies

strengthening the SOMO-LUMO interaction and enhancing the rate of addition as seen

from the data in Table 1-3.21,52 Here, the greater nucleophilic character of t-butyl relative

to methyl is evident from its enhanced rate of addition to olefins bearing electron

withdrawing groups.

As previously discussed, electronegative substituents at the radical center which

substantially lower its SOMO energy will impart electrophilic character and reinforce the

transition state SOMO-HOMO interaction (Figure 1-9, case b). Indeed, it has been

shown that rates of addition of dicyanomethyl53 and perfluorinated54 radicals correlate

with the ionization potentials of the substrate alkenes. The intermediate behavior of
"ambiphilic" radicals, such as malonyl and (tert-butoxycarbonyl)methyl, has also been

documented, these species yielding "U"-shaped correlations between rates of addition

and alkene IP and EA values.5557

A more thorough presentation of kinetic results obtained to date for the addition

of partially and fully fluorinated radicals to alkenes is given in Chapters 2 and 3.

Steric Effects; Regiochemical Preferences in Addition

Competition studies on both nucleophilic and electrophilic radicals have provided

for some generalizations in terms of the regiochemical preference for addition to

unsymmetrically substituted olefins.52 The preferred orientation of radical addition

occurs to the unsubstituted end of the double bond, attributed to steric repulsion but also

influenced by the effect of substituents on the coefficients of the HOMO and LUMO of

the alkene. Such FMO effects can be the decisive factor in polysubstituted olefins if

steric effects are in opposition. Strongly spin delocalizing substituents on the alkene

reinforce such sterically induced regiochemical preferences and exert slight rate

enhancing effects; however, as such additions occur through early transition states the
effect of exothermicity on the kinetics of addition should be minimal.

The concepts introduced in the aforementioned discussion on intermolecular
radical additions extend to their intramolecular cyclization analogues, an overview of
which will now be presented.

Intramolecular Addition Reactions: Radical Cyclizations

The intramolecular addition reactions of alkenyl radicals enjoy a strong foothold

among the available strategies for the construction of cyclic organic molecules. In
addition to their synthetic utility, the kinetic, regioelectronic and stereoelectronic
characteristics of radical cyclization reactions as a function of substituent continue to fuel
an abundance of fundamental physical organic structure-reactivity investigations, more

than thirty years after the first report of the archetypal radical ring closure, cyclization of
hex-5-en-1-yl radical 1 (Figure 1-10).58

250 C
S6 + 0

(98%) (2%)
1 2 3
Figure 1-10. Cyclization of Hex-5-en-1-yl Radical 1 to Cyclopentylcarbinyl (2) and
Cyclohexyl (3) Radicals. At 250 C, 5-exo Closure Dominates 49: 1.

The most striking aspect of this reaction lies in the preferred regiochemistry of
addition. In the case of the parent hydrocarbon, 5-exo59 cyclization dominates (Eact [5-
exo] = 6.8 kcal mol1, Eac [6-endo] = 8.5 kcal mol-1)27'60 yielding the less

thermodynamically stable primary cyclopentylcarbinyl radical 2. This finding has
provided the driving force for a number of experimental61 and theoretical62"68

investigations geared toward the understanding of radical cyclization regiochemistry in

the hydrocarbon and related substituted systems.

Early explanations,69 later advanced by semiempirical techniques,65 attributed

this result to a less negative entropy of activation for cyclization to 2. Although

experiment demonstrates this to be true, the difference (AS1,5 AS*,6 = 2.8 eu)60 is

insufficient to completely account for the observed regiochemistry; the preferred mode of

cyclization resulting primarily from enthalpic (AH*1,6 AH*1,5 = 1.7 kcal mol1) rather than

entropic factors. Ab initio computations66 lend support to this conclusion.

Transition structures for 5-exo and 6-endo cyclization of 1 have been located

using a variety of theoretical treatments. The UHF/6-31G(d) structures leading to 2 and

3 are shown in Figures 1-11 and 1-12.

^ Ifc-JflL e0 = 109.70 ^ s
2.186 A ',

Figure 1-11. Two Views of the UHF/6-31G(d) "5-exo-chair" Transition Structure for
Cyclization Hex-5-en-l-yl Radical 1 to Cyclopentylcarbinyl Radical 2.

098.40 '

Figure 1-12. Two Views of the UHF/6-31G(d) "6-endo-chair" Transition Structure for
Cyclization of Hex-5-en-1-yl Radical I to Cyclohexyl radical 3.

Spellmeyer and Houk66 have also postulated additional "boat-like" transition
structures on the basis of molecular mechanics calculations parameterized by ab initio
results of model systems. Inclusion of these "boat-like" structures as viable competing
pathways was found to be necessary for the accurate prediction of regio- and
stereoselectivities in cyclizations of alkyl substituted and heteroalkenyl radicals. The
existence of such structures is corroborated by higher level ab initio treatments
performed as part of the present study and are shown in Figures 1-13 and 1-14.

O~~rM^^ Sj~s0 9108.20 d ~e

rK:is) 2.192 A

Figure 1-13. Two Views of the UHF/6-31G(d) "5-exo-boaf Transition Structure for
Cyclization of Hex-5-en-l-yl Radical 1 to Cyclopentylcarbinyl Radical 2.

0 =100.00 "
J^T \2.245 A %Am. r

Figure 1-14. Two Views of the UHF/6-31G(d) "6-endo-twist-boaf' Transition Structure
for Cyclization of Hex-5-en-l-yl Radical 1 to Cyclohexyl Radical 3.

Upon inspection of the forming bond lengths and angles of the 5-exo structure in
Figure 1-11, its similarity to the transition structure for addition of methyl radical to
ethylene is readily apparent. The C-C-C angle of attack, 109.7, is practically identical to

that in Figure 1-8 and fits the requirement for overlap of the radical SOMO with the n*

orbital of the alkene moiety. From Figure 1-12 it is observed that this angle is

significantly reduced (98.4) in the 6-endo approach. Thus the required disposition of

centers for optimal FMO overlap is more readily achieved in the 5-exo transition

structure, leading to the kinetically preferred cyclization product.

Substituent Effects on the Kinetics and Regiochemistry of 5-Hexenyl Cyclizations

Radical 1 cyclizes to cyclopentylmethyl radical 2 with a rate constant (kcs) of

2.3 x 105 s' at 25 C.27 In the parent hydrocarbon, 6-endo closure competes to a very

minor extent (kc6 = ca. 4.7 x 103 s1). However, it will be shown that the rates and

regiochemical preferences can be substantially affected by substitution at both the

radical center and terminal alkene.

Alkyl Substitution: Steric Effects

Beckwith et al. have reported rate constants for a number of alkyl-substituted

hexenyl radicals.70,71 The rates of 5-exo closure as a function of gem-dialkyl substitution

on the aliphatic portion of the hexenyl chain is shown in Table 1-4.

It is seen that substitution at the radical center has a nearly negligible effect on

the rate of cyclization, due to offsetting polar and steric considerations. Conversely, a

significant (> 10-fold) rate enhancement is observed with internal substitution (systems

6, 8, and 9), accelerated by relief of steric compression between alkyl groups during ring

formation (the "Thorpe-lngold", or "gem-dimethyl" effect).72

Intermediate kinetic behavior would be expected of monosubstituted 5-hexenyl

systems. The data in Table 1-5 show this to be the case. In addition, a stereochemical

preference for cis- or trans-dimethylcyclopentanes, depending on the location of the

substituent on the chain, is observed. This has been rationalized by Beckwith et al.71

Table 1-4. 5-exo Cyclization Rate Constants for gem-Disubstituted 5-Hexenyl Radical

Cyclization Reaction kc5, 105 s1 (298 K)

1 2

4 5

^ 36c

6 7

= __52c

8 7

= 32c

9 5
a Reference 27. b Reference 28. c Reference 71.

based on the cyclization transition structure depicted in Figure 1-11. Although "early" in
terms of the forming C-C bond, the overall orientation of atoms in this structure is quite
product-like. According to this rationale, substituents on the aliphatic fragment are
likened to those in chair cyclohexane, which then occupy an equatorial position in the
chair transition structure. Minor products are assumed to derive from the occupation of

axial positions. The latter has been disputed by Spellmeyer and Houk,6 whose model

indicates that such secondary products originate from an equatorial disposition of

substituents in the boat-like transition structure of Figurel-12, rather than from an axial

orientation in the chair.

Table 1-5. Cis- and Trans- 5-exo Cyclization Rate Constants for Monosubstituted 5-
Hexenyl Systems. Rate Constants are for 298 K.

Cyclization Reaction kc5 (cis), 105 s"1 kc5 (trans), 105 s-1

+ 1.18 0.42a
10 cis-11 trans-11

+ 2.4" b 4 b

12 cis-13 trans-13

'N +
U Q + 7 7Ob 2.4

14 cis-13 trans-13

+ 0.75b 3.60

15 cis-11 trans-11
a Reference 73. b Reference 71.

In the above examples, 5-exo products are formed either predominantly (> 97%)

or exclusively. Substitution at the vinyl group leads to marked changes in regiochemical

ratios as indicated by the data in Table 1-6. Replacement of hydrogen by methyl (16) or

isopropyl (19) at C5 results in preferential formation 6-membered rings 18 and 21.

Inspection of the data indicates that this shift in regiochemistry is not due to a significant

Table 1-6. Rate Constants at 338 K for 5-exo and 6-endo Cyclization for Vinyl-
substituted 5-Hexenyl Radicals.

Cyclization Reaction kc5, 105 s1 kc6, 105 s1

+ )9.4 0.19 ab
1 2 3
wI d

0.21b 0. O37b

16 17 18

+ 0 0.21b 0.66b

19 20 21

+ 1. 0.19b

22 23 24

S+ 22b < 0.1b

25 26 27
a Calculated from the Arrhenius parameters given in reference 27. b Reference 70.

extent to rate enhancement for 6-endo closure, but rather a substantial (44-fold)
retardation of 5-membered ring formation due to a combination of 1,5 steric hindrance
and back strain engendered at Cs upon adaptation of sp3 character. Substitution at both
C5 and C6 again favors 5-exo closure, the rate of which decreased relative to the parent
system. Disubstitution at C6 (25) gives rise to a slight (2.3-fold) rate enhancement for 5-
exo closure, sufficiently explained on thermodynamic grounds, which dominates 6-endo

cyclization by a factor of at least 220. The above data indicates that the kinetic and

regiochemical characteristics of alkyl-substituted 5-hexenyl cyclizations may be

sufficiently rationalized by steric considerations. The effects of substitution by

conjugating, heteroatom-containing groups is outlined below.

FMO Considerations

Studies of 5-hexenyl systems bearing "polar" subsitutents have been

investigated by Newcomb.16'74-77 In line with those of intermolecular radical additions,

the kinetics of radical ring closure will be influenced by the impact of substitutents on the

SOMO-HOMO and SOMO-LUMO interaction in the cyclization transition state.

Table 1-7. 5-exo Cyclization Rate Constants For oa-Donor- and a-Acceptor-substituted
6,6-Diphenyl-5-hexenyl radicals. Rate Constants are for 298 K.

Ph &" Ph


System X Y kc5. 107 s1 (298 K)a

28-> 29 H H 4

30- 31 H CH3 2

32-> 33 CH3 CH3 1

34- 35 H OCH3 4

36-> 37 H CO2CH2CH3 3.7

38 -39 CH3 CO2CH2CH3 0.04

40 41 CH3 CN 0.03
a Calculated from the Arrhenius parameters provided in reference 76.

As seen from the data in Table 1-7, only a very minor effect is exerted by either

a-donor or a-acceptor substituents, relative to parent system 28. The marked decrease

in rate for 38 and 40 is attributed to an increase in activation energy due to enforced

planarity at the radical site induced by the Tc-delocalizing substituents CO2CH2CH3 and


Table 1-8. Absolute Rate Constants at 298 K for 5-exo Cyclization of 5-Hexenyl
Systems Bearing Vinylic Donor and Acceptor Substituents.

Cyclization Reaction

kc5, 105 s-1 (298 K)



NC 44





* CN



H3CON ___ A
L^ .--^ 1000b

46 47
a Reference 27. b Calculated from the Arrhenius parameters provided in reference 74.

Substitution at the vinyl terminus, especially by strong resonance-withdrawing

groups can significantly accelerate the rate of ring closure. Although possessing a

radical stabilizing group, methoxy analog 42 enjoys only a very minor increase in rate.

The donor substituent raises the energies of the frontier orbitals, increasing the SOMO-

HOMO interaction but widening the SOMO-LUMO energy gap, the latter more important

for relatively nucleophilic alkyl radicals (Figure 1-9). Substituents which serve to lower

the FMO energies should reinforce the SOMO-LUMO interaction, leading to rate

enhancement. The nearly 300-fold increase resulting from cyano substitution (44)
reflects such an effect. The slight rate increase 46 relative to 45 has been explained on

the basis of the suggested slight extra "push-pull," or "captodative" stabilization
manifested in donor-acceptor disubstituted systems.79'80

The importance of kinetic control was previously mentioned. Care must be taken
in assessing the potential for reversibility in such intramolecular additions, which may
obscure the effect of steric and/or polar influences on reaction kinetics. This is

demonstrated in the 5-exo:6-endo product ratios of highly stabilized systems 48 and
5 .81,82

Table 1-9. Product Ratios for 5-Hexenyl Cyclization Reactions Under Full or Partial
Thermodyamic Control.
Cyclization Reaction % 5-exo % 6-endo

.*^ 'rPh Ph^V

h <22 > 78
48 49 50

-U CO + L C02 16 84

51 52 53

Given the data and discussion provided in the above sections, a review of the
direct and competitive techniques utilized in the determination of rates of organic radical

reactions is now in order.

Methods for Determination of Organic Radical Kinetics

The development of indirect competitive methods, in conjunction with laser flash
photolytic generation and time-resolved detection of transient intermediates, has greatly

expanded the dynamic range available for the measurement of radical reactions,

especially those at the upper end of the kinetic scale. Such advances have provided for
the use of a variety of precursors and the accurate determination of rate constants for

reactions approaching the diffusion-controlled limit in solution.16

Laser Flash Photolysis: Direct Measurement of Addition Rates

In the time-resolved laser flash photolysis method, described in detail in the

literature,83 radicals are generated from precursors possessing a suitable chromophore

by a laser pulse of the appropriate wavelength. Alkyl iodides, diacyl peroxides, and the

0-acylthiohydroxamic esters of Barton et al.84 are most commonly utilized in this regard

(Figure 1-15).

(a) R-l ,-1 R

| R' R hv ||0 -C02
(b) R- 0 2 R1 O fast 2 R
O O-

( N hv 0 -CO2
(C) R 0 R 0 fast
Figure 1-15. Laser Flash Photolytic Generation of Radicals R" from (a) Iodide, (b) Diacyl
Peroxide, and (c) 0-Acylthiohydroxamic Ester Precursors.

These radicals so generated undergo further reaction, usually bimolecular
addition or unimolecular cyclization to a (typically phenyl-substituted) double bond (a

styrene in the case of bimolecular additions). The increase in the characteristic

absorption of this intermediate benzyl radical (Xmax 320 nm) is then followed in a time-

dependent manner by UV-visible spectroscopy (Figure 1-16). For bimolecular additions,

this experimental growth curve is fit to the expression in Equation 1-11, yielding absolute

rate of addition kwd. With such data in hand, this addition can now serve as a competing

basis reaction for the determination of rates of other transformations involving the same

or structurally similar radical.

kobs =ko + kadd [alkene] (1-11)

-\ R

hv -G /
R-X hvo R --- / ax ca. 320 nm
kadd 1 G

Figure 1-16. LFP Generation of R* and Detection of Transient Benzyl Radical Adduct
for Determination of Absolute Rate of Addition kad.

Indirect Methods (Competitive Techniques)

Indirect kinetic methods involve the partitioning of an intermediate between two

competing pathways, one with a known rate constant and the other whose rate constant

is to be determined. Post facto product analyses, typically by chromatographic or

spectroscopic means, provide a ratio of rate constants from which the new kinetic value

is obtained. Such radical kinetic measurements usually involve competition between two

bimolecular reactions or a bimolecular reaction competing with unimolecular

rearrangement; examples of both instances are provided in Figures 1-17 and 1-18.

Determination of rates of hydrogen abstraction by this method involves the

generation of R" in the presence of two trapping agents; in Figure 1-17, styrene and the

hydrogen atom donor. Both traps are typically present in excess to ensure pseudo-first

order behavior. Radical R* may undergo addition to styrene (with known rate constant

kadd) forming the intermediate benzyl radical, itself trapped with excess hydrogen donor,

yielding closed shell product with a rate which is kinetically unimportant provided the

addition reaction is irreversible. Alternatively, R* is trapped directly by hydrogen atom

donor with a rate constant kH, which may be obtained from the pseudo-first order relation

[reduced] [kH ] [ R1 [M-H ]
[ adduct ] [ kadd ] [ R I [ CH2=CHR']

where [reduced] and [adduct] are the final product concentrations, [M-H] is the
concentration of hydrogen atom donor, and [CH2=CHR'] the concentration of alkene trap.
For accurate kinetic determinations, a series of runs is performed where trap
concentrations are varied, (again, maintaining at least a five-fold excess) a plot of
product ratios versus that of trapping agents yielding the ratio kH I kay.

(vary) VR R

k /add --- ^{S--- ~

\ M-H (vary)

R "----- ^ --H


Figure 1-17. Competition Between Bimolecular Addition to an Alkene with Rate
Constant kad and Bimolecular Trapping by Hydrogen Atom Donor with Rate Constant

Determination of rates of cyclization are performed in a similar manner, using
hydrogen atom abstraction (with the known value of kH) as the competitive basis
reaction. Unimolecular rearrangement competes with bimolecular trapping with an
excess of hydrogen atom donor (Figure 1-18) to yield intermediate cyclic radicals, further
trapped to form characterizable products.
The ratio of products of cyclization versus hydrogen abstraction are obtained
from the pseudo-first order relation in Equation 1-13, a plot of the ratio of products of
hydrogen abstraction to those of cyclization as a function of trapping agent concentration
affording ratios kcs I kH and kc6 I kH.
Finally, the importance of an efficient chain process should be reemphasized.
High conversions of precursors, although important for any radical reaction, are crucial in

competitive kinetic experiments. The reliability of data resulting from indirect methods is

directly dependent on high "mass balance" values, those of 90% or greater typically

being desired.

[reduced] [kH] [R] [M-H] (1-13)
[cyclized] [kcn ][R*]

M-H (vary)

\ kc5 jL

\ l^ \ M -H1"
-k -6 0 0

Figure 1-18. Competition Between Bimolecular Hydrogen Atom Abstraction with Rate
Constant kad and Unimolecular 5-exo and 6-endo Cyclization with Rate Constants kc5
and kc6.

Such competitive processes have been employed extensively by the Dolbier
research group in the investigation of the rates of addition, hydrogen atom abstraction,

and cyclization of a variety of fluorinated open shell systems, providing the first

quantitative kinetic data for this class of reactive intermediates.54,85-91


The proceeding discussions have attempted to provide the reader with an
introduction to the chemistry of organic free radicals. Kinetic data for hydrogen

abstraction, addition, and cyclization reactions of hydrocarbon radicals, important
benchmarks for comparison of reactivity with other substituted systems, was provided.


Substituent effects, rationalized on the basis of a combination of thermodynamic, steric

and FMO considerations, were discussed.

The following chapter provides a review of the effects of fluorine substitution in

organic molecules, including fluorinated radicals. Previous research efforts in this area

by the Dolbier group are summarized, setting the stage for the presentation of results of

the current study.




Incorporation of fluorine into organic molecules often imparts dramatic alterations

in structure and reactivity. These effects are induced by three major characteristics

inherent to the fluorine atom: extreme electronegativity, non-bonded electron pairs, and

relatively small size.

Fluorine possesses the highest electronegativity of all the elements, with a value

of 4.10 on the Pauling scale, compared to oxygen (3.50), chlorine (2.83), bromine (2.74),

carbon (2.50), and hydrogen (2.20).92 As a substituent in organic systems, this results in

strong inductive withdrawal of electron density through the a molecular framework and

highly polarized bonds with substantial ionic character.

Three non-bonding pairs of electrons in 2p orbitals similar in size to those of

other second-row elements provide for optimal overlap, and therefore an offsetting back

donation of electron density into the molecule to which it is bonded.

The accepted van der Waals radius of fluorine, 1.47 A, suggests minimal steric

impact in comparison with other halogens (chlorine, 1.73 A; bromine, 1.84 A; iodine,

2.01 A; carbon, 1.70 A; oxygen, 1.52 A; hydrogen, 1.20 A).93 This has allowed for the

complete replacement of hydrogen by fluorine in organic systems, a feat not possible to

such an extent with any other element.

The following sections, based on a number of excellent reviews,9497 provide an

introduction to the fascinating behavior exhibited by fluorinated stable molecules and

reactive intermediates due to a combination of the above effects.

Structure. Bonding, and Reactivity in Saturated Systems

The data provided in Tables 2-1 and 2-29'97 reveal a trend unique to fluorine

within the halogenated methanes. An incremental shortening of C-F interatomic

distances, with a resultant increase in bond dissociation energies, is observed as the

series is traversed. No such trend exists for any other member of the halomethane

family; on the contrary, it is seen from the data in Table 2-2 that such C-X BDE values

instead decrease with increasing halogen content. Strengthening of C-H bonds is also

observed within the fluoromethanes (CH3F, 101.3 kcal mol1; CH2F2, 103.2 kcal mol1;

CF3H, 106.7 kcal mol-1).95

Table 2-1. Carbon-Halogen Interatomic Distances (Angstroms) of Halomethanes


F 1.385 1.357 1.332 1.319

Cl 1.781 1.772 1.758 1.767

Br 1.939 1.934 1.930 1.942

Table 2-2. Carbon-Halogen Bond Dissociation Energies (D, kcal mol') of


F 108.3 119.5 127.5 130.5

Cl 82.9 81.0 77.7 72.9

Br 69.6 64 62 56.2

I 57.2 51.3 45.7 -

Data for geminally fluorinated ethanes parallel that of the methane series,

demonstrating a progressive strengthening and shortening of both C-C and C-F bonds

with increasing fluorination (Table 2-3). Conversely, vicinal fluorination gives rise to the

opposite effect on C-C bonds, a steady lengthening and weakening being observed.

A variety of hypotheses have been put forth to explain the observed trends. One

rationalization, invoked by Pauling98 and based on valence bond theory, involves "double

bond, no bond" resonance of the type depicted in Figure 2-1.

Table 2-3. Interatomic Distances and Dissociation Energies of Fluoroethanes

Ethane r (C-C). A Do (C-C). kcal mol-1 r (C-F). A Do (C-F). kcal mol-r1

CH3-CH3 1.532 90.4

CH3-CH2F 1.502 91.2 1.398 107.9

CH3-CHF2 1.498 95.6 1.343 Unknown

CH3-CF3 1.494 101.2 1.335 124.8

CH2F-CF3 1.501 94.6 Unknown 109.4 (CH2F)

CF3-CF3 1.545 98.7 Unknown 126.8

As the degree of geminal fluorination is increased, the number of such resonance

forms involving doubly bonded fluorine increases (0, 2, 6, and 12 in the case of CH3F,

CH2F2, CH3F, and CF4, respectively). This is supported by ab initio calculations at the

RHF/4-31G and 4-31G(d) levels,99-101 which illustrate back donation of electron density

from fluorine into the C-F o* orbitals. It is further observed that the overlap population

between the 2p orbitals of carbon and those of fluorine increases continually with

successive fluorination; in contrast, such carbon-chlorine overlap populations decrease

steadily from CH3CI to CCl4.

F F-
;L F ^ -----"-=

Figure 2-1. "Double Bond, No Bond" Resonance in Geminally Fluorinated Alkanes.

Alternative explanations based on hybridization schemes have also been
advanced. It is postulated that for electronegative elements bound to carbon,

rehybridization occurs causing an increase in the amount of p character directed toward

the substituent. Thus, in CH3F, the C-F bond possesses greater p character, with

greater s character in the C-H bonds. This rationale accounts not only for incremental

C-F bond strengthening, but also for the observed changes in geometry within the

fluoromethane series. Accumulation of p character in C-F bonds should lead to a

decrease in FCF bond angle, accompanied by HCH widening. This is consistent with

experimental observation (ZFCF in CF4, 109.5; CHF3, 108.7; CH2F2, 108.3;102'103 for

CH2F2, ZHCH = 113.70).103

Finally, a more recent argument has been advanced by Wiberg,104 on the basis

of Coulombic interactions between carbon and fluorine substituents. From charge-fitting

treatments based on calculated electrostatic potentials, a linear increase in positive

charge on carbon is observed, while the degree of negative charge on each of the

fluorine substituents remains quite constant. Thus, incremental fluorine substitution

strengthens not only new, but also previous, C-F bonds. This finding will be further

discussed in Chapter 3, as such ESP-derived charges calculated for larger fluorinated

systems as part of the present study were found to provide a cogent explanation for the

remarkable and contrasting effects of fluorination on the strengths of both C-C and a and

13 C-H bonds.

Structure. Bonding, and Reactivity in Unsaturated Systems

The most reliable structural and n-BDE data for the fluoroethylenes is provided in

Table 2-4.9'7 Vinylic fluorine substitution results in shorter C=C bonds than in the

parent hydrocarbon, and shorter C-F bonds than fluoroalkanes bearing the same

number of geminal or vicinal fluorine substituents. Ab initio investigations by Radom99

and others105107 attribute the C-F bond contraction to delocalization of fluorine 2p

electrons into the C-C n bond (depicted in resonance terms in Figure 2-2). Computed

atomic charges are consistent with this conclusion.

Table 2-4. Interatomic Distances, Angles, and n Bond Dissociation Energies of


r(C=C),A 1.339 1.333 1.316 1.309 1.311

r(C-F),A 1.348 1.324 1.336 1.319

ZHCH, deg. 117.8 114.7 119.3 -

ZHCF, deg. 111.3 114.0 -

ZFCF, deg. 109.7 109.1 112.6

7 D0, kcal mol' 63-64 Unknown 62.8 Unknown 52.3

The marked decrease in FCF bond angles has been rationalized by Bemrett108

and Kollman109 on the basis of hybridization arguments; Epiotis has advanced an

alternative explanation involving nonbonded attraction between fluorine atoms.110111'
>\ F F

Figure 2-2. Fluorine 2p Electron Delocalization in Unsaturated Systems.

Photoelectron12 and electron attachment1"3 spectral data for the fluoroethylene

series are provided in Table 2-5. A significant lowering (over a range of 3.1 eV) of the a

MO energies is observed, with only a slight (ca. 0.3 eV) variation in the n MOs.

Stabilization of the a MOs is ascribed to extensive delocalization over the fluorine

substituents; such an interaction within the n system is diminished and counteracted by

strong C-F antibonding overlap.112 A steady increase in electron attachment energies

with successive fluorination can also be seen, attributed to destabilization of n* resulting

from an antibonding interaction with the fluorine 2p AOs.113

Heats of hydrogenation provided in Table 2-695 illustrate the reactivities of

fluorinated alkenes. In general, transformation of a polyfluorinated olefin into a saturated

Table 2-5. Vertical Ionization Potentials (I. P.) and Electron Attachment Energies (EA) for
the Fluoroethenes.

Ethene I.P., eV a I.P., eV E.A., eV

CH2=CH2 10.6 12.85 1.78

CH2=CHF 10.58 13.79 1.91

CH2=CF2 10.72 14.79 1.84

cis-CHF=CHF 10.43 13.97 2.18

trans-CHF=CHF 10.38 13.90 2.39

CF2=CHF 10.53 14.64 2.45

CF2=CF2 10.52 15.95 3.00

derivative is more exothermic than for the parent hydrocarbon. This arises from a

combination of the destabilizing effect of polyfluorination on double bonds and the

thermodynamic preference for gem-difluoro substitution at saturated carbon. The

deviation of CH2=CHF in Table 2-6 is explained by the preference of a single fluorine

substituent to reside at the vinylic position (Figure 2-3).114

Table 2-6. Heats of Hydrogenation of the Fluoroethenes.

Ethene AH (H,), kcal mol1

CH2=CH2 -32.6

CH2=CHF -29.7

CH2=CF2 -38.8

CF2=CF2 -45.7

Fluorine Non-Bonded Interactions in Reactive Intermediates

The r-donor ability of the fluorine substituent is reflected in its activating and

ortho, para-directing character in electrophilic aromatic substitution reactions,115

consistent with "3C NMR measurements of fluorobenzene, where shielding of these

positions is observed.1"6

H CH2F 12 H3C F 12 H3C H

H H AH = -3.34 kcal mol-1 H H AH = +0.65 kcal mol1- H F

54 Z-55 E-55

H3C F 12 H CF2H

H F AH =-2.5 kcal mol-1 H H

56 57

Figure 2-3. Thermodynamic Equilibria in Mono- and Difluoropropenes.

Delocalization of fluorine's nonbonded electrons into the vacant 2p orbital on

carbon more than compensates for its inductive withdrawal in a-fluoro carbocations,

resulting in net stabilization. In the gas phase, carbocation stability increases along the

series *CH3 < CF3 < *CH2F < CHF2 and *CH2CH3 << +CF2CH3 = +CHFCH3.117'118 The

+CF3 cation has been observed in the gas phase, with many others having been

successfully generated in solution.119121 In contrast, fluorination at the 1 position and

beyond destabilizes carbocations due to inductive effects; simple alkyl 13-fluoro

carbocations not benefiting from additional stabilizing factors have yet to be detected.

Electron pair repulsion in a-fluoro carbanions results in a strong preference for

pyramidal geometries, ab initio calculations122-124 predicting an FCF angle of ca. 99.5

and an inversion barrier of 119 kcal mol1 for CF3. Although fluorination does increase

C-H bond acidities in such pyramidal systems,125 a destabilizing effect is observed in

cases such as the 9-fluorofluorenyl anion (Figure 2-4, X = F) where coplanarity is forced

between the 2p orbitals on fluorine and the remainder of the R system.126 Fluorine

substitution in the p position stabilizes carbanions through a combination of inductive

and hyperconjugative effects, (Figure 2-5) the latter supported by X-ray crystallographic

data of perfluoroalkyl anion salts as well as through calculation.'22"127128



Figure 2-4. Fluorine Destabilization in

F o



X kexc (reI

D 1

F 0.125

Cl 400

Br 700

Planar 9-Halofluorenyl Anions.


Figure 2-5. Negative Hyperconjugation in p-Fluorocarbanions.

Fluorine Steric Effects

The minimal spatial requirements of fluorine, the smallest non-hydrogenic

substituent, would imply a very minor steric impact on reaction kinetics and

thermochemistry. In most cases this is true; however, examples of steric inhibition in

reactions of fluoro-substituted systems do exist, typically in conformational and other

dynamic processes occurring via highly congested transition states. This is illustrated in

Figure 2-6;129'130 the meta ring flip in 61 (X = F) exhibiting the largest known rate

retardation induced by a single fluorine substituent.

The disparate behavior observed in the Cope rearrangements of d,I- and meso-

62 (Figure 2-7)131 provides a particularly striking example of a fluorine steric effect.

Transformation of d,l-62 to 63 proceeds via a typical chair-like transition structure, where

in meso-62 a "boat-like" structure is required for C1 C6 bond formation. The higher AH*

x x

q: )/

X = H, AG < 6 kcal morl'1 (340 K)
X = F, AG" = 11.1 kcal mol"1 (340 K)

kH / kF= 1011 (298 K)

Figure 2-6. Inhibition of Conformational Dynamics by Fluorine Substitution.

and positive ASt implies a dissociative, rather than concerted, transition state for

rearrangement of meso-62. This is induced by severe electrostatic repulsion between

the high charge densities of the terminal fluorine substituents, separated by less than the

sum of their van der Waals radii in the Cope transition state.132

H1 CF2




AH* = 22.4 kcal molr1
AS* = -17.5 eu




meso-62 AHW = 49.5 kcal mol1 63
AS* = +8.1 eu
Figure 2-7. Chair- versus Boat-Constrained Cope Rearrangement Reactions of
Terminally Fluorinated Dienes d,I- and meso-62.


Steric effects are enhanced by perfluoroalkylation and branching. Cyclohexane
A values133 and modified Taft steric parameters1'34 demonstrate that CF3 is at least as
large as isopropyl; evidence exists135 to suggest that perfluoroisopropyl and tert-butyl are
comparable in size. The most remarkable example of the above affects is the existence
of perfluorinated radical 64, (Figure 2-8) found to be persistent by ESR even in the
presence of molecular oxygen.136 The astounding kinetic stability of 64 derives from
steric sheltering of the unpaired electron by the neighboring perfluoroalkyl groups.

2: 'F1 9

Figure 2-8. Scherer's Persistent Perfluoroalkyl Radical.

The Fluorine Substituent in Free Radicals

Early application of organofluorine radical chemistry was comprised of the chain-
mediated addition of polyhalomethanes and ethanes to olefins, first by Haszeldine137 and
soon thereafter by Tarrant.138'139 Subsequent relative rate studies by Stefani et al.140
followed by those of Tedder141'142 clearly demonstrate the contrasting behavior of
fluorinated and non-fluorinated radicals in their bimolecular additions to alkenes.
The fluorine substituent has a substantial effect on the structure of organic
radicals as well as dramatic, but comprehensible, alterations in hydrogen abstraction
and addition reactivity in comparison to hydrocarbon systems, resulting primarily from

fluorine's potent a-withdrawing character.

Structural Aspects

In contrast to the planar, n-type methyl radical, a-fluorination results in

increasingly pyramidal, a-type radicals, as indicated by electron paramagnetic

resonance measurements143 and ab initio theoretical studies.144147 Calculations by

Pasto at the UHF/4-31G level indicate barriers to inversion of 0.5, 6.8, and 25.1

kcal mol1 for CH2F, CHF2, and "CF3, respectively.147 Geometries of the fluoromethyl

radicals computed at the UHF/6-31G(d) level are provided in Figure 2-9.

I 90.0o I 101.1o I 105.80 I 107.60

Figure 2-9. UHF/6-31G(d) Pyramidalization Angles of *CH3, "CH2F, "CHF2, and *CF3.

Inversion barriers in the series appear somewhat sensitive to the level of theory

employed and substantially increase with the inclusion of polarization functions. SCF

calculations by Dykstra145 employing a polarized double- basis result in an inversion

barrier of 33 kcal mol1 for "CF3; inclusion of electron correlation (QCISD(T)/

6-31G(d)//UHFI6-31G(d), present work) affords a value of 29.4 kcal mol-1. Such

successive deviation from planarity is due to a combination of effects; relief of It

repulsion between the singly-occupied orbital on carbon and the fluorine 2p electrons is

further reinforced by overlap between the carbon 2p and C-F antibonding orbitals. This

strong tendency for pyramidalization has been shown to be responsible for the low n

bond energy in tetrafluoroethylene (Tables 2-4 and 2-6).148

The structures of fluorinated C2 radicals have been theoretically probed by

Paddon-Row and Chen et al.149-152 Alkyl substitution induces slight pyramidalization as

observed in the ethyl radical, (Figure 2-10) due to hyperconjugation between the SOMO

and the staggered P3 C-H bond.153 Fluorination at the radical center exerts an effect

similar to that observed in the methyl series, while the structures of alkyl radicals are

found to be relatively insensitive to P-fluorination.

1 94.60 I 94.4

105.7 / 105.9

Figure 2-10. UHF/6-31G(d) Pyramidalization Angles for Ethyl Radicals CH3CH2",
CF3CH2", CH3CF2", and CF3CF2.

Radical Stability

FMO theory dictates that for radicals bearing electronegative substituents with

lone pairs (F, OH, NH2, SH) an inductive, destabilizing influence exists, countered by

stabilization resulting from delocalization of the unpaired electron.154 Thus, in the a

sense, fluoroalkyl radicals are destabilized. Furthermore, the opposing i-stabilizing

effect of the fluorine lone pairs decreases with pyramidalization of the radical site, due to

diminished overlap with the 2p AO on carbon.

The progressive decrease in stability of alkyl radicals with a- or 3-fluorination has

been illustrated by Pasto,80147 in the form of calculated radical stabilization energies

(RSE) based on isodesmic reactions (Table 2-7). The aforementioned increase in C-H

bond dissociation energies along the fluoromethane series lends experimental support.

Although some degree of 3 C-F hyperconjugative interaction is observed in 2-

fluoro substituted radicals,152 such a contribution to overall radical stability is minor and

inductive destabilizing influences dominate, as indicated by experimental and theoretical

C-H BDEs for the 2-fluoroethanes (Table 2-8).

Table 2-7. Isodesmic Equation and Radical Stabilization Energies (RSE, kcal mol-1,
4-31G) for a- and -Substituted Systems. Positive Values Denote Radical Stabilization.

Xn'CH3-n + CH4 -* XnCH4-n + 'CH3 (2-1)


F +1.64 CH3 +3.27

F2 +0.56 OH +5.73

F3 -4.21 OCH3 +5.30

CH2F +1.46 CN +5.34

CHF2 +0.16 NH2 +10.26

CF3 -1.34 *NH3 -4.07

SH +5.66

*SH2 -3.17

Table 2-8. Experimental and Theoretical C-H Bond Dissociation Energies (kcal mol'1)
of 2-Fluoroethanes.

101.1ab 103.6' 106.7a.f

97.7c 99.6c 101.3c 102.0c

100.0d 104.3d

102.0e 104.le 105.9e 107.le
a Experimental Value; Reference 155. b Experimental Value, Reference 156.
c MP2/6-311G(d,p)//MP2/6-31G(d,p); Reference 157. d B3LYP/6-31G(d); Reference 91
and Present Study. e MP2/6-311+G(3df,2p)//MP2/6-31G(d); Reference 158.
f Experimental Value; Reference 159.

Organofluorine Radical Reactivity

As a result of the a-withdrawing character of the fluorine substituent and

interaction of the SOMO with C-F o* orbitals, fluoroalkyl radicals should possess lowered

SOMO energies and therefore exhibit enhanced SOMO-HOMO interactions in

comparison with reactions of their hydrocarbon counterparts. Experimental ionization

potential and electron affinity data, although sparse, has been compiled in a recent

review by Dolbier90 and demonstrates the greater absolute electronegativity of the

fluoroalkyl radicals. Calculated quantities, inferred from Koopmans' theorem160 or based

on radical-ion energy differences, follow the expected trend although quantitative

agreement is often lacking.161162 The combination of such FMO, geometric, and

enthalpy factors in hydrogen atom abstraction, intermolecular addition, and cyclization

reactions of fluoroalkyl radicals is now discussed.

Hvdroaen Atom Abstraction Reactions

A review by Tedder163 has underlined the importance of polar and enthalpic

factors in radical abstraction reactions. Activation parameters for abstraction by methyl

and trifluoromethyl radicals from a series of hydrogen donors are provided in Table 2-9.

Table 2-9. Arrhenius Parameters for Hydrogen Atom Abstraction by Methyl and
Trifluoromethyl Radicals.

CH3" CF3"
H-Donor Ea log A EAa oqg A

CH3-H 14.2 8.8 11.3 8.9

CH3CH2-H 11.8 8.8 6.9 8.4

(CH3)2CH-H 10.1 8.8 6.5 8.1

(CH3)3C-H 8.0 8.3 4.9 7.7

H-Cl 2.5b 5.Oc
a In kcal mol-1. b AHOXnn = -1 kcal mol-1. c AHOrxn = -3 kcal mol1.

In the first four examples, the decrease in activation barrier for both CH3* and

CF3 abstractions from alkanes are in line with the greater stability of the product radical;

in each case, the barrier to abstraction by CF3 is substantially lowered. In contrast,

abstraction of hydrogen atom from HCI by CF3* is much less facile, occurring with a

barrier twice that of CH3* in spite of its slightly greater exothermicity.

In collaboration with Lusztyk and Ingold at NRCC, LFP-determined rates of

perfluoroalkyl radical addition to a number of alkenes have been determined by the

Dolbier group.54'85'88'91 This has afforded, via competitive kinetic techniques, absolute

rate constants for hydrogen abstraction by the perfluoro-n-heptyl radical from a series of

donors, summarized in Table 2-10.54'86'87'89 For comparison, abstraction rate constants

for hydrocarbon n-alkyl radicals were provided in Table 1-1.

Table 2-10. Absolute Rate Constants for Hydrogen Atom Abstraction for Perfluoro-n-
heptyl Radicals. Rate Constants are at 303 K.

Et3SiH (TMS),Si(CH3)H nBu3GeH (TMS)SiH nBu3SnH PhSH

kH(n-C7F15"), 0.75 16 15 51 203 0.28
106 M1 s-1

For the first five donors in the series, a substantial rate enhancement is observed

over hydrocarbon radicals, ranging from 75-fold in the case of nBu3SnH to nearly 900-

fold in the case of Et3SiH, after slight temperature correction to 303 K. Although

hydrogen transfer to perfluoroalkyl radicals is more exothermic (see the previous

discussion on C-H BDEs) this is insufficient to account for a nearly three order of

magnitude difference in rate constants. Furthermore, thiophenol, an excellent donor to

hydrocarbon radicals, is found to suffer a greater than 400-fold decrease in transfer rate

to perfluoroalkyls.

These characteristics are explained by the ability of the radical-donor pair to

accommodate charge transfer interactions in the hydrogen transfer transition state, as

previously discussed in Chapter 1. The relatively electropositive donor agents

(stannanes, germanes, and silanes) lead to a more favorable polarity matchup with the

electronegative perfluoroalkyl radical than with the hydrocarbon. Conversely, transfer

from the more electronegative thiophenol results in a non-polarized or polarity-

mismatched transition state. The existence of such polar effects were confirmed by a

correlation between rates of hydrogen transfer to perfluoro-n-heptyl radicals by a series

of substituted thiophenols, versus their Hammett a+ constants (Figure 2-11).87 The

resulting p value of -0.56, when compared to that obtained in the case of tert-butoxyl

(-0.30)164 again reflects the high electrophilicity of perfluoroalkyl radicals.

0.6 -

0.4 -



-0.2 -

-0.4 7--
-1.00 -0.75 -0.50 -0.25

0.00 0.25 0.50 0.75 1.00

Figure 2-11. Hammett Plot for Hydrogen Abstraction from para-Substituted Thiophenols
by Perfluoro-n-heptyl Radical.

Intermolecular Addition Reactions

Relative rates of addition of small fluorocarbon radicals to fluorinated and non-

fluorinated olefins have been extensively investigated by Tedder and Walton,

culminating in a critical review in 1980.52 Table 2-11 illustrates the relative reactivity of

Table 2-11. Relative Rates of Addition of the Fluoromethyl Radicals to Ethene and






kdd (CF4)/ka (CdH4) (437 K)




the fluoromethyl radicals towards ethylene and tetrafluoroethylene. Additional studies,

utilizing a number of unsymmetrical methyl- and trifluoromethyl-substituted olefins,

solidified the ascription of relative rates and regiochemical preferences to a combination

of polar and steric influences.

Absolute rates of addition of perfluoroalkyl radicals to alkenes have been

determined by laser flash photolysis, a subset of the data acquired to date presented in

Table 2-12.90 The dramatic rate acceleration enjoyed by the perfluoroalkyl radicals

versus hydrocarbon n-alkyls is readily apparent, ranging from factors of 300 to 30,000 in

the case of the heptafluoropropyl radical addition to electron-rich alkenes. The rate of

addition to pentafluorostyrene, in contrast, is increased by only a factor of 42.

Table 2-12. Absolute Rate Constants for Addition of Perfluoroalkyl Radicals to Alkenes.

kdd, 106 M-1 s- (298 K)
Alkene C3F7 C7F15 "C8F17 CF CF3 RCH*

Styrene 43 46 46 79 53 0.12

a-Methylstyrene 78 89 94 87 0.059

13P-Methylstyrene 3.8 3.7 7.0 17

Pentafluorostyrene 13 23 26 0.31

4-Methylstyrene 61

4-Methoxystyrene 65

4-Chlorostyrene 36

4-(CF3)Styrene 35

1-Hexene 6.2 7.9 16 0.0002

Such enhancements may potentially be attributed to a combination of factors.

Relative reaction enthalpies should play a role, as a stronger C-C bond (from CH3-CH3

versus CF3-CH3 BDE data in Table 2-3, ca. 11 kcal mol-1) is formed upon perfluoroalkyl

addition. However, only slight increases (factors of 5-7) in addition rates of the

perfluoroalkyls to styrene versus 1-hexene are observed, despite the greater

exothermicity (ca. 16 kcal mol1) of the former. This demonstrates the relatively minor

importance of reaction enthalpy, in accord with the early transition states expected for

radical addition.

The pyramidal, a-character of the fluoroalkyl radicals should afford a kinetic

advantage (further discussed in Chapter 3) over the planar hydrocarbon, the LFP-

measured rate of addition of 1,1-difluoropentyl radical to styrene, 2.7 x 106 M'1 s1,88

giving rise to a 22-fold enhancement relative to n-alkyls.

The primary factor responsible for such striking increases in reactivity is believed

to be charge transfer influences (Figure 2-12) similar to those postulated for hydrogen

atom transfer reactions. The lowered SOMO energies of the perfluoroalkyl radicals and

resulting enhanced SOMO-HOMO interactions with alkenes in the addition transition

state leads to substantial rate enhancement. Supporting evidence, in the form of a

Hammett relation involving para-substituted styrenes and correlations between addition

rates and alkene ionization potentials, has been offered.54 Ab initio computations concur

with the experimental findings and are discussed in Chapter 3.


Figure 2-12. Polarized Transition State for Addition of Perfluoroalkyl Radicals to

Intramolecular Cyclization Reactions

Cyclopolymerization of fluorinated monomers has long been known as a means

for the generation of macromolecular materials with unique physical properties.165"168

However, despite the widespread popularity of the unimolecular 5-hexenyl radical

cyclization for the generation of five-membered rings, synthesis of fluorinated cyclic

products utilizing radical methodology has received limited attention.90 This is somewhat

surprising, in light of the current interest and demonstrated importance of fluorinated

analogues and mimics of pharmaceutical and agricultural agents.169

Until recently, no quantitative kinetic data existed for cyclization reactions of

fluorinated radicals. With the competitively-determined rate constants of hydrogen

abstraction by perfluoroalkyl radicals serving as basis reactions, rates of cyclization of a

number of fluorinated 5-hexenyl systems have now been determined,86,89,'170 examples of

which are provided in Figure 2-13.

Most obvious of the data is the remarkable rate acceleration in the 5-exo

cyclizations of 65 and 68, occurring with 163- and 41-fold increases relative to the parent

hydrocarbon I and ascribed to charge-transfer effects analogous to those occurring in

bimolecular additions. Consequently, with no such polarity matchup in 71 (which

involves cyclization of an electrophilic radical onto an electron-deficient double bond)

only a minor increase in rate is observed. The 5-exo cyclization rates of 71 and 74, in

line with those other hexenyl systems bearing fluorinated double bonds,170 demonstrate

the lack of kinetic impact of vinylic substitution.

Especially surprising is the degree to which 65 and 68 undergo 6-endo closure,

the former with a 1040-fold, and the latter a 700-fold acceleration relative to 1. This is

further discussed in Chapter 4, in which the reactivities of a series of lightly-fluorinated 5-

hexenyl systems are investigated.

In the proceeding discussions of hydrogen abstraction, addition, and cyclization

reactivity, the observed kinetic behavior was found to be due to a combination of

geometric and polar effects induced by polyfluorination. Related studies of partially

fluorinated alkyl radicals, addressed in the next two chapters, will aid in separating the

effects of fluorination at the a and p positions and beyond, providing insight into the

extent to which the effects of perfluorination on the reactivity of organic radicals are the

sum of their parts.

Table 2-13. Some Absolute Rate Constants for 5-exo and 6-endo
Fluorinated 5-Hexenyl Radicals. Rate Constants are for 303 K.

Cyclization Reaction

kc5, 105 s-1

Cyclization of

kc6, 105s1

F F2


^ CF2
F2CC.. CF2


F ,F.





- F2C CF2

F2C 2

75 76



+ 0



+ r-CF2


+ "- I I
F2C,.. CF2


+ C(F2





An introduction to the general structural and reactivity characteristics imparted by

fluorine substitution in organic systems has been provided. Such substitution can either

be stabilizing or destabilizing, depending on the nature of the ground state molecule,

intermediate, or reaction in question.

A summary of the first absolute kinetic data obtained for reactions of fluorinated

radicals has been presented. Results of these studies, involving per- or otherwise highly

fluorinated systems, demonstrate the combined influence of polarity, structural, and

enthalpic factors. The following studies of the addition, hydrogen abstraction, and

cyclization kinetics of partially fluorinated alkyl radicals will serve to dissect the relative

magnitudes of these influences on organic radical reactivity.




Initial studies of the addition rates of some partially-fluorinated radicals to

alkenes88 demonstrated observable rate enhancements relative to n-alkyls, though not

nearly as great as those of perfluoroalkyl systems. Such reactivity derives from the

combination of structural and polar characteristics induced by fluorine substitution.

The current study extends the amount of absolute rate data assembled for the

addition of partially fluorinated radicals to olefins. In addition, through competitive kinetic

techniques, absolute rate constants for hydrogen abstraction from tri-n-butyltin hydride

(nBu3SnH) have been determined. Such kinetic data is necessary for the determination

of absolute rates of cyclization of partially fluorinated 5-hexenyl radicals, discussed in

Chapter 4, and allows for the partitioning of the gross reactivity characteristics of

perfluoralkyl radicals into the separate influences of a, P3, and y fluorination. This is

accomplished by use of the following systems: RCH2CH2CF2' (a,a-difluoro),

RCH2CF2CH2 (P3,P13-difluoro), RCH2CF2CF2* (a,cc,3,,3-tetrafluoro), CF3CF2 (a,a,p13,13,13-

pentafluoro), and RfCH2CH2 (y-perfluoro).

The existence of charge transfer stabilization in the transition states for

fluorinated radical addition to alkenes is corroborated by ab initio calculations, and the

thermodynamics of C-H and C-C bonding and radical stabilization in hydrofluorocarbons

rationalized on the basis of Coulombic interactions.

Precursor Syntheses and Competitive Kinetic Studies

In each of the competitive kinetic runs, hydrofluorocarbon radicals were

generated from bromide or iodide precursors by photoassisted C-X bond homolysis.

These radicals subsequently underwent competitive trapping with known, varying

concentrations of styrene or nBu3SnH, adjusted to ensure pseudo-first-order kinetic

behavior and to allow for accurately measurable amounts of trapping products, as

depicted in Figure 1-17 and in greater detail below.

1.,1-Difluorohex-l-vl Radical (77)

The synthesis of bromide precursor 80 was achieved in two steps (Figure 3-1) in

a straightforward manner. Copper(l)-mediated addition of dibromodifluoromethane to 1-

pentene (78), based on a modification by Gonzalez et al.171 of a procedure by Burton

and Kehoe172 afforded 1,3-dibromo-1,1-difluorohexane 79 in typical yield.

Regioselective displacement of the internal bromine was accomplished via sodium

borohydride reduction in DMSO, providing precursor 80 contaminated with a small

amount of overreduction product 81. Pure samples of each were obtained by

preparative GC separation, the former utilized in the competition run and the latter for

spectral comparison with kinetic NMR data.

CF2Br2 Br
s --CF2Br
CuCI (cat.) (57.3%)
78 79

----- ^^^CF2Br + CF^*2Hl
80 81

Figure 3-1. Preparation of 1-Bromo-1,1-difluorohexane 80 and 1,1-Difluorohexane 81.

CF2Br hv
80 CAD6





Figure 3-2. kH I kaw Competitive

L ~77

+ nBu3Sn"

81 F2H

S ./ /^C. nBu3SnH
F 1 ______-- _


+ nBu3Sn"

Kinetic Scheme for 1,1-Difluorohex-l-yl Radical 77.

Photolysis of 80 as a C6D6 solution in the presence of an excess of nBu3SnH and

styrene (Figure 3-2) afforded intermediate radical 77. Subsequent entrapment by these

agents (both irreversible processes) yielded 81 and 82, respectively; the latter further

trapped by nBu3SnH to yield 3,3-difluoro-l-phenyloctane 83. Throughout the course of

the reaction, nBu3Sn* radicals are generated to propagate the chain process via

abstraction of halogen from precursor 80.

Product ratios for varied concentrations of trapping agents were determined by

19F NMR analysis according to the pseudo-first-order relation in Equation 3-1,


[kH] [ 77] [nBu3SnH]
[kadd] [ 77] [ C6H5CH=CH2 ]


a plot of product ratios obtained for each data point versus that of trapping agents

affording the ratio kH I kadd. The stability of trapped products under the reaction

conditions and lack of appreciable side reactions were demonstrated by the high

conversion of precursor 80 to 81 and 83 versus an internal standard of a,a,a,-

trifluorotoluene, (< -63.24) indicating in turn the high efficiency of the chain process and

reliability of the kinetic results. A partial 19F NMR spectrum of the first of six data points

is provided in Figure 3-3, a doublet of triplets (-CF2H, 4 -116.0) observed for 81 versus

an overlapping triplet of triplets at -99.1 (-CF2-) for 83. Full kinetic data and yields are

given in Table 3-1 below, the plot of which located in Figure 3-4. The slope of the line

(3.39 0.02) in conjunction with the known absolute rate constant for addition of

.. .... .. . ... .. ......) ---.--.-.-.. .

-98 -99 -100 -101 -102 -103 -104 -105 -106 -107 -108 -109 -110 -111 -112 -113 -114 -115 -116 -117 -118

Figure 3-3. Partial 19F NMR Spectrum of Data Point 1 for kH I kaW Competition of 1,1-
Difluorohex-1-yl Radical 77.

Table 3-1. Competitive Kinetic Data for kH I kadd Competition of 1,1-Difluorohex-l-yl
Radical 77.

[801 CH5CH=CH2 1 [F nBuSnH 1 / [C6H5CH=CH2 1 f 811/ F 83 1 % Yield

0.094 2.01 0.719 2.30 95

0.094 1.81 0.847 2.73 95

0.094 1.61 1.01 3.26 96

0.094 1.41 1.21 3.93 97

0.094 1.21 1.49 4.88 96

0.094 1.01 1.87 6.21 95


1,1-difluoropent-l-yl radical to styrene, 2.7 ( 0.5) x 106 M1 s-1, resulted in an absolute

rate constant kH of 9.1 ( 1.7) x 106 M1 s1. It should be noted that the accuracy of such

derived kH values can be no better than those reported in the LFP determinations of kadd,

the error estimates in the former reflecting both the least-squares fit of the line and

propagated error of the latter. Synthesis of styrene adduct 83 was performed by classic

means for characterization and spectral comparison (Figure 3-5).

0.50 0.75 1.00 1.25 1.50
[ nBu3SnH ] / [ C6H5CH=CH2 ]

Figure 3-4. Plot of the Data in Columns 3 and 4 of Table 3-1.


Na2Cr2O7 / H2S04





1. CH3(CH2)4CHO

2. H30*



Figure 3-5. Preparation of Styrene Adduct 83.






2.2-Difluorohex-l-vl Radical (88)

In accordance with literature precedent,173 a-bromination of 2-hexanone in the

presence of urea in acetic acid selectively afforded 1-bromo isomer 90 in 72.2% yield.

Subsequent treatment with diethylaminosulfurtrifluoride (DAST) provided bromo

precursor 91, originally purified by preparative GC for use in the kinetic study. However,

a sluggish chain reaction (further hindered by the strong UV absorption of the excess

styrene present in the kinetic samples) led to the preparation of iodo precursor 92

(Figure 3-6) via Finkelstein transformation at elevated temperature.

0 Br2 0 DAST
89 90
.'/^^ C-cCH2Br Nal C c.CH21
F2 (CH3)2CO F2
(59.7%) (87.5%)
91 92

Figure 3-6. Preparation of 1-lodo-2,2-Difluorohexane, (92) Precursor to 2,2-Difluorohex-
1-yl Radical 88.

10 -
9 Coefficients:
-" m = 27.3
M 8 b =-0.065
r 2 = 0.998
V) 7-


5 I --- I I -
0.20 0.22 0.24 0.26 0.28 0.30 0.32 0.34

[ nBu3SnH ] / [ C6H5CH=CH2]

Figure 3-7. Plot for kH I kad Competition of 2,2-Difluorohex-l-yl Radical 88.

The competition plot for kH I kadd determination is found in Figure 3-7; raw data for

this and all remaining experiments in this chapter may be found in Chapter 5. With the

absolute rate constant for addition of 2,2-difluoropent-l-yl radical to styrene known from

LFP experiments, a kH value of 1.4 ( 0.5) x 107 M1 s1 was determined. Authentic

samples of 93 and 98 were prepared for spectral comparison and characterization as

illustrated below in Figure 3-8.

c-^_/I .^CH2Br


" - c .-C H 3


N. Br


. YMgBr
( ,;5"-

1. CH3(CH2)3CHO

2. H3O+



Na2Cr2O7 / H2S04



Figure 3-8. Preparation of Hydrogen Abstraction Product 93 and Styrene Adduct 98.

1.1.2,2-Tetrafluorobut-l-vl (99) and 1.1,2.2-Tetrafluorohex-l-yl (100) Radicals

At the time of this study, no absolute rate constants for addition of a 1,1,2,2-

tetrafluorinated radical to alkenes had been determined. Thus, in order to obtain a kH

value for such a system, a precursor suitable for absolute kad measurements was first

required. Bromide precursors, although in most cases sufficient for competition

experiments, are ineffective under the LFP operating conditions used in our kad

determinations (308 nm excimer laser pulses) due to their relatively short wavelength

chromophore and small extinction coefficient (for 102, Ew = 37.3 M1 cm1, ;,ax = 218

nm, cyclohexane solvent). The instability of 0-acylthiohydroxamic esters of the

perfluoroalkanoic acids has been noted by Barton.174 With neither these nor diacyl

peroxide precursors lending themselves to isolation and / or shipment to the NRCC in

Canada, the synthesis of a suitable iodide precursor was undertaken.

1-Bromo-1,1,2,2-tetrafluorohexane (102, Figure 3-9) was prepared in one step

from 6-bromo-5,5,6,6-tetrafluorohex-1-ene (supplied by Halocarbons, Inc.) and its

transformation to the corresponding iodide or carboxylic acid (the latter of which could be

converted to the iodide via Hunsdieker methodology) attempted under a variety of

conditions (Figure 3-10).

1. BH3 Me2S
,- c,.-CF2Br 2. C5HlCO2H Nc--CF2Br
F2 Tetraglyme F2
101 102

Figure 3-9. Preparation of 1-Bromo-1,1,2,2-tetrafluorohexane (102).

Although the conversion of perfluoroalkyl iodides to their lighter analogues is

known in the literature,175176 downward transhalogenation of perfluoroaliphatic halides is

exceedingly difficult. Indeed, all attempts at conversion of 102 to the corresponding

iodide were met with failure. Perfluoroalkyl Grignard reagents, generated either directly

(and in low yield) or via transmetallation by an alkylmagnesium halide, utilize iodide

starting materials.177178 Investigations of perfluoroalkylzinc halides by Miller179 resulted

in a similar conclusion; perfluoro-n-propyl iodide may be converted to the organozinc

reagent in ca. 60-80% yield (as determined by aqueous hydrolysis or capture with

halogen electrophiles) after a brief induction period. On the contrary, reaction of

heptafluoro-n-propyl bromide with zinc in 1,2-dimethoxyethane afforded no product after

1) Mg
2) 12


1) t-BuLi, -80C
2) 12

1) t-BuLi, -100C
2) 12

1) MeLi, -100C
2) 12


1) MeLi, -100C
2) CO2


CuCI (cat.)

Et3N 3HF / NIS
CH2CI2 0 C










Figure 3-10. Attempted Preparation of lodo Analogue of 102.

73 hours at 900 C; a 60% yield of the zinc derivative was obtained (inferred via

hydrolysis) after a period of 1.5 months. Attempted lithiation of 102 at low temperature

resulted in the formation of p-fluoride elimination product 1,1,2-trifluoro-l-hexene, (103)

as identified by its 19F NMR spectrum. Additionally, neither Cu(l)-induced addition of

1,2-diiodotetrafluoroethane to 1-pentene nor iodofluorination of 103 utilizing the

triethylamine trihydrofluoride / N-halosuccinimide methodology of Alvernhe et al.180 were



,,-- c.CF2Br





Synthesis of a C4 iodide was achieved through modification of a DuPont literature

procedure,181 whereby 1,4-diiodo-1,1,2,2-tetrafluorobutane was produced in 40.1% yield

via direct thermal addition of 1,2-diiodotetrafluoroethane to ethylene (Figure 3-11). DBU-

induced elimination of hydrogen iodide in ether followed by diimide hydrogenation with

hydrazine-hydrogen peroxide in methanol afforded tetrafluoroiodo precursor 107 in 2.8%

overall yield after preparative GC purification, which was sent to the NRCC for absolute

kadd measurements.





I c .CF21




Figure 3-11. Preparation of 1,1,2,2-Tetrafluoro-l -iodobutane (107).

0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5
[ nBu3SnH ] / [ C6H5CH=CH2 ]

Figure 3-12. Plot for kH I kaw Competition of 1,1,2,2-Tetrafluorohex-l-yl Radical 100.

Bromide 102 was utilized in the kH I kadd competition, the kinetic plot for which

given in Figure 3-12. An authentic sample of styrene addition product 110 was prepared


by slow syringe pump addition of nBu3SnH to a heated, irradiated solution of 102 and

styrene in benzene (Figure 3-13).











Figure 3-13. Preparation of Hydrogen Abstraction Product 108 and Styrene Adduct 110.

2-f[Perfluorohexylleth-l-vi Radical (111)

2-[Perfluorohexyl]-1l-iodoethane 112 was provided as a gift from Prof. Neil Brace.

The kH I kadd competition plot is found in Figure 3-14; hydrogen abstraction product

1-[perfluorohexyl]ethane (113) and styrene adduct 1-[perfluorohexyl]-4-phenylbutane

(117) were prepared as shown in Figure 3-15.

6 Coefficients:
% m = 16.0
5 b =-0.371
r 2 = 0.999


2 1I I
0.20 0.25 0.30 0.35 0.40 0.45
[nBu3SnH ] / [ C6H5CH=CH2 ]

Figure 3-14. Plot for kH I kadd Competition of 2-[Perfluorohexyl]eth-1-yl Radical 111.




+ N


C6F13 -




Et3B (cat.)_



Figure 3-15. Preparation of Hydrogen Atom Abstraction Product 113 and Styrene

Pentafluoroethyl Radical (118)

lodopentafluoroethane (119) was obtained from PCR, Inc. Due to the high

volatility of both this precursor (bp 12-13 C) and hydrogen atom abstraction product

pentafluoroethane (120, bp -48.5 C) 119 was handled as a solution in degassed C6D6.

The kH I kadd competition experiment (Figure 3-16) was performed in tubes which were

quickly flame-sealed upon injection of an aliquot of the chilled precursor stock solution.

0.40 0.45 0.50 0.55 0.60 0.65 0.70 0.75 0.80
[ nBu3SnH ] / [ C6H5CH=CH2]
Figure 3-16. Plot for kH / kadd Competition of Pentafluoroethyl Radical 118.




Samples of 120 and 122 were prepared under free radical conditions for characterization

purposes as shown in Figure 3-17.


119 C6eH5CH=CH2 F2 ]
L 121

nBu3SnH F3C2'
C61-16 121

nBu3SnH F3cC- -


Figure 3-17. Preparation of Hydrogen Atom Abstraction Product 120 and Styrene
Adduct 122.


Absolute rate constants for addition and hydrogen atom abstraction for systems

studied in this chapter are provided in Table 3-2. For comparison, data for hydrocarbon
(n-pentyl) and perfluoroalkyl (perfluoro-n-heptyl) radicals are included.

It is seen from the data that the reactivity trends for partially fluorinated radical

additions to styrene are generally adhered to in hydrogen abstractions. However, the

actual rate constants for the latter are observed to differ (on the average, by a factor of
11) from those for addition, and span a narrower range. The decrease in abstraction
rate ratios as a function of substitution is due to the proximity (within an order of
magnitude) to diffusion control for the more reactive radicals 100, 118, and 127.

C-H and C-C bond dissociation energies of the fluoroalkanes should reflect the
relative thermochemistry of hydrogen atom abstraction and addition to alkenes by their
respective radicals. However, very few experimentally determined BDE values for such

Table 3-2. Absolute Rate Constants for Hydrogen Abstraction from Tributyltin Hydride
and Addition to Styrene by Partially Fluorinated Radicals. Rate Constants are for 298 K.
Radical k (M-1 s-1) kk (M-1 s1) kH

RCH2CH2 1.2 xlO 105 b 1 2.4 X106 c 1
(123) a

RCH2CF2" 2.7 (0.5) x 106 e 22.5 9.1 ( 1.7) x 106 f 3.8
(124, 77) d

RCF2CH2" 5.2(1.8)x 105 e 4.3 1.4 (0.5) x 107 f 5.8
(125,88) g

RCF2CF2" 2.0 (0.1) x 107' 167 9.2 (0.8) x 107 38
(99,100) h

RfCH2CH2" 1.3(0.2) x 105 e 1.1 2.1 (0.3) x 106' 0.9

CF3CF2= 7.9 (1.0) x 107 658 3.2 ( 0.3) x 108 133

C7F15* 4.6 (0.6) x 107J 383 2.0 ( 0.3) x 108j 83
a R = C3H7. b Reference 182, After Modification for Temperature and Other Factors in
Table III of Reference 183. c Reference 27. d For kadd Experiment, R = C3H7 (124); for
kH Experiment, R = C4H9 (77). e Reference 88. Reference 91 and Present Study. g For
kadd Experiment, R = C3H7 (125); for kH Experiment, R = C4H9 (88). h For kadd
Experiment, R = C2H5 (99); for kH Experiment, R = C4H9 (100). For kadd Experiment,
Rf = C4F9 (126); for kH Experiment, Rf = C6F13 (111). 1 Reference 54.

systems have been reported. Theoretical studies by Boyd157'184 at the MP2 level have

provided reasonably accurate C-H and C-C BDEs for C2 hydrofluorocarbons. A more

recent investigation based on isodesmic reactions by Marshall and Schwartz,158

published after the completion of the present study, has yielded C-H BDEs of

appreciably high quality for some linear (C2) and branched (up to C4) polyfluoroalkanes.

Out of interest in determining additional C-H and C-C BDEs for larger (through C4)

fluorinated n-alkyls, the geometries of a series of partially fluorinated ethanes, propanes,

and butanes, along with their respective radicals generated from terminal C-H or C-CH3

bond cleavage, were optimized at the hybrid density functional level. This DFT method

was chosen due to its implicit consideration of electron correlation at only slightly greater

computational expense than that of Hartree-Fock theory. Utilizing the three-parameter

exchange functional of Becke'85 and the correlation functional of Lee, Yang, and Parr186

with the 6-31G(d) basis, bond dissociation energies obtained in this "direct" fashion were

found to be, in the cases where such values are known, within 1-3 kcal mol1 of those

determined experimentally and in comparable or better agreement with experiment than

the MP2/6-31 1 G(d,p) values obtained by Boyd for C2 systems. Tabulated experimental

(where known) and calculated C-H and C-C BDE values are provided below in Tables

3-3 and 3-4, respectively.

Table 3-3. Theoretical and Experimental C-H Bond Dissociation Energies.

C-H Bond Calculated BDE. kcal mol1 a Experiment

CH3CH2-H 100.0 101.1 + 1 e
97.7 c
102.0 d

CH3CF2-H 97.4 b 99.5 2.5f
97.0 c

CF3CH2-H 104.3 b 106.7 1'
102.0 c
107.1 d

CF3CF2-H 99.5 b 102.7 0.5
99.7 c
104.6 d

CH3CH2CH2-H 100.3 b 100.4 0.6 e

CH3CH2CF2-H 97.7b

CH3CF2CH2-H 103.1 b

CH3CF2CF2-H 100.1 b

CF3CH2CH2-H 101.4 b

CF3CF2CH2-H 103.8 b
a Reported as Do (298.15 K). b B3LYP/6-31G(d); Reference 91 and Present Study.
c MP2/6-311 G(d,p)//MP2/6-31 G(d,p); Reference 157. d MP2/6-311 +G(3df,2p)//MP2/
6-31G(d); Reference 158. e Reference 187. f Reference 95.

Table 3-4. Experimental and Theoretical C-C Bond Dissociation Energies.

C-C Bond Calculated BDE. kcal mol1 a Experiment

CH3-CH3 89.4 b 90.4 0.2 d
90.6 c

CF3-CH3 99.6 b 101.2 1.1 d
103.3 c

CH3CH2-CH3 86.3 b

CH3CF2-CH3 91.4 b

CF3CH2-CH3 91.4 b

CF3CF2-CH3 95.5b

CH3CH2CH2-CH3 86.7 b

CH3CH2CF2-CH3 91.6 b

CH3CF2CH2-CH3 89.9 b

CH3CF2CF2-CH3 95.4 b

CF3CH2CH2-CH3 87.8 b
a Reported as Do (298.15 K). b B3LYP/6-31G(d); Reference 91 and Present Study.
c MP2/6-311G(d,p)//MP2/6-31G(d,p); Reference 157. d Reference 95.

Inspection of the data reveals interesting trends within both the C-H and C-C

BDE series. From Table 3-3, it is observed that a-fluorination results in a weakening of

C-H bonds, on the order of 1-3 kcal mol1, as predicted by the various levels of theory

and supported by experiment in the case of ethane versus 1,1-difluoroethane.

Conversely, P-fluoro substitution results in a 3-5 kcal molr1 increase in terminal C-H

BDEs. Furthermore, strengthening of C-C bonds (Table 3-4) is observed for all systems

examined, whether substituted at the a, P3, or even (albeit diminished) Y position. An

explanation for this behavior, consistent with the given BDE and other thermochemical

data, is provided later in the chapter.

With the relative thermodynamics of C-H and C-C bond formation investigated,

attention is turned to polarity effects. As mentioned previously, rate constants for

addition of the perfluoroalkyl radicals to subtituted styrenes (Table 2-12) are observed to

increase with decreasing ionization potential (styrene, IP 8.43 eV; a-methylstyrene, IP

8.19 eV; pentafluorostyrene, IP 9.20 eV).188 In contrast, 1,1-difluoropentyl radical (124)

was found to add with rates equal, within experimental error, for all three olefins.88 From

this observation, along with derived absolute electronegativities for the radicals CH3*

(4.96), CH2F" (4.73), CHF2" (4.91), and CF3" (5.74) it was concluded that a-fluoro

substitution alone does not impart electrophilicity to an alkyl radical, and may instead

give rise to nucleophilic behavior.88

A series of theoretical studies by Wong et al.48'49,189,190 have assessed the degree

of charge transfer interaction in the transition states for methyl, hydroxymethyl,

cyanomethyl and tert-butyl radical additions to a series of monosubstituted olefins.

Based in part on the computation of partial charges, it was concluded that the addition of

methyl radical was governed primarily by enthalpic effects, with no evidence for

nucleophilic character arising from either Mulliken or Bader-based charge-fitting

schemes. Hydroxymethyl and tert-butyl were found to be nucleophilic, with cyanomethyl

exhibiting substantial electrophilicity.

In order to determine such tendencies for the fluorinated radicals under

investigation, transition structures for the addition of hydrocarbon and fluoro-substituted

ethyl radicals to ethylene and C, of propene have been located at the UHF/6-31G(d)

level. Partial atomic charges were then computed, based on fits to the electrostatic

potential at points selected according the Merz-Kollman-Singh scheme.91'192

Previous studies by Houk et al. utilizing the 3-21G basis found that addition of

ethyl radical to ethylene occurs preferentially via a gauche conformation, with an

incipient C-C-C-C dihedral angle of ca. 600.46 Such gauche and anti structures

computed at the UHF/6-31(d) level are depicted in Figure 3-18, the former ca. 0.1

kcal mol'1 lower in energy than that of (b). Inclusion of electron correlation at the spin-

projected PMP2/6-311G(d,p)//UHF/6-31G(d) level increases this energy difference to ca.

0.3 kcal mol1.

103.80 \%
2.227 A \ 1100

2.231 A 109.90
\ 109n9

Figure 3-18. UHF/6-31G(d) (a) Gauche and (b) Anti Transition Structures for Addition of
Ethyl Radical to Ethylene. Relevant Geometrical Parameters are Shown.

Consistent with UHF/3-21G results, the preferred mode of addition of the ethyl

and radical to C1 of propene involves a gauche arrangement of the radical with the

alkene C=C i bond and a transoid orientation of the methyl group of the radical with

respect to that of the olefinic C2 carbon (Figure 3-19). This 'gauche-transoid' structure

lies ca. 0.1 and 0.4 kcal mol1 below the 'anti' and 'gauche-cisoid' conformers,

respectively, at the UHF/6-31(d) level after zero-point energy correction. Inclusion of

correlation effects (PMP2/6-311G(d,p)//UHF/6-31G(d)) yields energy differences of 0.3

and 0.4 kcal mol1. These orientation preferences extend to the fluoroethyl series, as

seen in Tables 3-5 and 3-6.

, 2.224 A

( a.

2.228 A

(b) CJ* ^

\ 2.219 A

(c) J

NJ 0 0
Figure 3-19. UHF/6-31G(d) (a) 'Gauche-transoid', (b) 'Anti', and (c) 'Gauche-cisoid'
Transition Structures for Addition of Ethyl Radical to C1 of Propene.

Table 3-5. Geometric Parameters and Total and Relative Energies of Transition
Structures for Addition of Fluorinated Methyl and Ethyl Radicals to Ethylene.
Radical r (C-C) r (C=C) z C-C-C E_ ZPE _
(Al LAl (Deg.) (Deg.) (au) a

CH3" 2.246 1.382 109.1 101.9 -117.575692 0.089165 -

CF3" 2.300 1.372 106.6 108.6 -414.156036 0.067955 -

CH3CH2 a 2.227 1.384 110.0 103.8 -156.612999 0.120578 0.0
(-157.243689) (0.0)

CH3CH2 b 2.231 1.383 109.9 103.3 -156.612807 0.120455 0.05
(-157.243039) (0.34)

CH3CF2 a 2.235 1.378 110.0 108.6 -354.333278 0.105597 0.0
(-355.404529) (0.0)

CH3CF2 b 2.245 1.378 106.3 108.2 -354.332326 0.105470 0.53
(-355.403222) (0.75)

CF3CH2 a 2.258 1.380 109.0 104.1 -453.202939 0.097486 0.0
(-454.490483) (0.0)

CF3CH2 b 2.258 1.380 107.4 103.5 -453.203135 0.097452 -0.14
(-454.490294) (0.10)

CF3CF2 a 2.275 1.375 108.2 108.6 -650.904439 0.081977 0.0
(-652.632413) (0.0)
CF3CF2 b 2.280 1.375 104.7 108.1 -650.904343 0.081948 0.04
(-652.632131) (0.16)
a Gauche. b Anti. c Degree of Radical Pyramidalization. d UHF/6-31G(d) Values;
PMP2/6-311G(d,p)//UHF/6-31G(d) Values in Parentheses. e Relative Conformer Energy
Differences, kcal mol-1. UHF/6-31G(d); PMP2/6-311G(d,p)IIUHFI6-31G(d) Values in

Inspection of the data reveals the similarity in both angle of attack and forming

C-C bond length, regardless of either the nature of the radical or conformation of the

transition structure. Incipient bond lengths range from ca. 2.22 to 2.30 A, generally

slightly longer for additions of the fluorinated members of the series. These somewhat

earlier transition states are also reflected in the shorter olefinic C=C bonds, 1.382 A in

the case of methyl radical addition to ethylene versus 1.372 A for trifluoromethyl, in turn


Table 3-6. Geometric Parameters and Total and Relative Energies of Transition
Structures for Addition of Fluorinated Methyl and Ethyl Radicals to C1 of Propene.




CH3CH2 a

CH3CH2" b


CH3CF2" a

CH3CF2" b

CH3CF2" c

CF3CH2 a

CF3CH2" b

CF3CH2" c

CF3CF2" a

CF3CF2" b

CF3CF2 c

r (C-C)






























z C-C-C















a Gauche-transoid. b Gauche-cisoid. c Anti.d Degree of Radical Pyramidalization.
e UHF/6-31G(d) Values; PMP2/6-311 IG(d,p)//UHF/6-31G(d) Values in Parentheses.
'Relative Conformer Energy Differences, kcal mol'. UHF/6-31G(d); PMP2/
6-311G(d,p)//UHF/6-31G(d) Values in Parentheses.



























































in accord with the greater exothermicity for the latter (AErxn = -22.35 kcal molr1 versus

-34.52 kcal mol-1, respectively, at the [QCISD(T)/6-311G(d,p)]'/UHF/6-31(d) level, and

C-C BDE data in Table 3-4).

Gauche-transoid addition of ethyl radical to C, of propene occurs via a transition

structure with a forming C-C interatomic distance of 2.224 A and and a C=C bond length

of 1.385 A, in comparison with 2.274 A and 1.375 A for addition of pentafluoroethyl along

the same trajectory. Attack angles appear slightly smaller for additions of the fluorinated

radicals, most notably in the case of CH3* versus CF3 and consistent with a reinforced

SOMO-HOMO interaction for the latter. However, such differences are barely

significant, and in the ethyl series appear to be influenced more by the conformation of

the transition structure than the nature of the attacking radical, in line with the previously

observed insensitivity of transition state geometry to additions of both nucleophilic and

electrophilic radicals.47'48

Of note is the degree of pyramidalization at the radical site in the addition

transition structure, ranging from 102-105 for radicals of the RCH2* type and 108-109

for a-fluorinated species. Considering the pyramidal nature of the ground states of the

latter as well (Figures 2-9 and 2-10) it follows that x-fluoroalkyl radicals enjoy a kinetic

advantage over their hydrocarbon analogues in that little or no additional bending is

necessary to accommodate the addition transition structure. The energetic cost of

pyramidalization of the methyl and tert-butyl radicals to the same extent as required for

their addition to ethylene has been computed at 1.5 and 1.6 kcal molr', respectively, at

the RMP2/6-31G(d)ilUHF/6-31G(d) level.49

Calculated degrees of charge transfer (CT) between radical and olefin moieties

of the addition transition structures are provided in Table 3-7. Where applicable, lowest

energy conformations (gauche in the case of ethyl radical additions to ethylene, gauche-

transoid for additions to C1 of propene) were used in the determination of the

electrostatic potential-derived charges.

Table 3-7. Calculated Charge Transfer Data (Electrons) for Transition Structures of
Hydrocarbon and Fluorinated Methyl and Ethyl Radical Addition to Ethylene and C1 of

Radical Ethylene a Propene b

CH3" -0.019 -0.004

CF3" -0.013 -0.006

CH3CH2" +0.036 +0.052

CH3CF2" +0.030 +0.052

CF3CH2" -0.047 -0.037

CF3CF2" -0.045 -0.034

Note: Derived from UHF/6-31G(d) Electrostatic Potentials. Negative Values Denote
Electron Transfer from Alkene to Radical. a Gauche Transition Structure. b Gauche-
transoid Transition Structure.

Consistent with previous investigation, addition of CH3* to ethylene involves only

a slight degree of charge transfer (-0.019 e) from olefin to radical (Mulliken and Bader

analyses yield values of -0.017 and -0.011, respectively)48 and even less so for addition

to propene. Somewhat surprisingly, CF3" addition is also predicted to occur without

appreciable polarization.

Along the ethyl series, such interactions appear more clearly defined. Ethyl

radical addition to both ethylene and propene involves a shift of electron density from the

radical to the alkene (0.036 and 0.052 e, respectively) well in accord with the expected

nucleophilicity of the alkyl radicals. CH3CF2* is predicted to be nucleophilic as well,

exhibiting transition state polar characteristics very similar to those of its hydrocarbon

counterpart and consistent with the experimentally deduced non-electrophilicity of

the a-fluoro radicals.

A striking reversal in these trends occurs upon fluorination at the p3 carbon atom,

regardless of the nature of the radical site itself. Here it is seen that both CF3CH2* and

CF3CF2 exhibit substantial electrophilicty, with ca. 0.04 0.05 units of electron density

transferred from the alkene to the radical center. To place such values into some

degree of perspective, addition of the strongly electrophilic cyanomethyl radical to C2 of

electron-rich vinylamine is predicted to occur with a transfer of ca. 0.11 electrons from

CH2=CHNH2 to "CH2CN.48

It is especially noteworthy that the degree of CT in the case of CF3CH2" and

CF3CF2" addition is practically unaffected by fluorination at the a carbon (-0.047 versus

-0.045 and -0.037 versus -0.034, respectively). This, along with the demonstrated lack

of kinetic impact of fluorine substitution at the y position (Table 3-2) leads to the

conclusion that the electrophilic character of the perfluoroalkyl radicals derives

exclusively from substitution at the 2-position.

With geometric, enthalpic, and polar considerations for hydrofluorocarbon

radicals having been addressed, the influences of each of these effects on determined

kadd and kH values are now discussed.

a .a-Difluoroalkvl Radicals (77, 124)

The 1,1-difluoroalkyl radicals, as mentioned previously, benefit from

pyramidalization at their radical site, leading to a more facile adoption of the transition

structure for addition or hydrogen atom abstraction. However, such an advantage is

counteracted by the experimentally and theoretically demonstrated lack of electrophilicity

for such species. Moreover, the terminal C-H bond weakening effect of gem-difluoro

substitution (2 3 kcal mol1, Table 3-3) leads to a slight thermodynamic disadvantage

for hydrogen abstraction by the corresponding radical relative to the hydrocarbon. Thus,

the 3.8-fold rate enhancement enjoyed by 77 may be completely ascribed to the a-type,

pyramidal nature of its ground state, attenuated by enthalpic and polarity factors working

in opposition. In contrast, the strengthening effect of gem-difluorination on C-C bonds

compliments that of pyramidal geometry, leading to a more substantial (22.5-fold) rate

enhancement for the addition of 124 to styrene versus hydrocarbon 123.

B.3-Difluoroalkvl Radicals (88. 125)

The rate enhancements for addition (4.3) and hydrogen abstraction (5.8)

observed for 2,2-difluoroalk-l-yl radicals are due to a complimentary combination of

polar and enthalpic effects. The terminal C-H bond in 2,2-difluoropropane is predicted to

be 2.8 kcal mol' stronger than that of propane itself; similarly, gem-difluorination at C2

of butane leads to a 3.2 kcal mol"1 strengthening of its C3-C4 bond. In spite of these

favorable considerations, the near-planar ground state geometry of 88 results in a

modest net rate acceleration for hydrogen abstraction from nBu3SnH. This also

functions to oppose the CT and enthalpic advantages present in the addition reaction of

125, giving rise to only a slight rate increase relative to n-alkyls and certainly diminished

in comparison with that enjoyed by 124.

-y-Fluorinated Radicals (111, 126)

Due to the near-planar geometric character of 3-fluoroalk-l-yls and the lack of

effect (ca. 1 kcal mol1) on terminal C-H and C-C BDE values, fluorination beyond two

carbon atoms removed from the radical site exhibits a negligible effect on the rates of

both addition and hydrogen transfer. The reactivities of 111 and 126 are found to be,

within experimental error, identical to those of the corresponding hydrocarbon.

.ac.13.-Tetrafluoroalkvl Radicals (99.100) and Pentafluoroethyl Radical (118)

Radicals substituted at the a and P3 positions benefit from both pyramidal

geometries and electrophilic character. Since the degree of pyramidalization of 1,1-

difluoroalkyl radicals remains constant regardless of substitution at the 3- and further

positions, geometrically induced influences on the reactivities of such polyfluorinated

radicals are expected to be uniform. Consequently, rate enhancements for radicals of

the type RCH2CF2CF2, RfCF2CF2CF2", and CF3CF2" versus RCH2CH2CF2" (for kH: 38,

83, and 133 versus 3.8; for kadd: 167, 383 and 658 versus 22.5; all relative to n-alkyls)

derive from either an increasing degree of transition state charge transfer stabilization,

increasingly greater exothermicity of reaction, or a combination of both. The relevant

C-H BDE data in Table 3-3 yields no direct correlation between reaction rate and

enthalpy for the polyfluorinated radical series, with values of 97.7, 100.1, and 99.5

kcal mol' corresponding to hydrogen abstraction by radicals 77, 100, and 118.

Similarly, terminal C-C BDEs of 91.6, 95.4, and 95.5 kcal mol1, equated with the

additions of 124, 99, and 118, illustrate that although P-fluoro substitution should lead to

rate enhancement on thermochemical grounds, such an effect does not account for the

incremental acceleration across the series.

With the degree of radical electrophilicity related to its substitution at the 2-

position and the potential for additional 13 C-F a* delocalization made possible by the

"extra" fluorine substituent in CF3CF2* relative to RCH2CF2CF2" and RfCF2CF2CF2", it

follows that the increasing resonance and inductive withdrawal ability of these groups

relative to RCH2CH2CF2* sufficiently explain both the enhanced reactivity of these

radicals as a whole, as well as the observed trend.

Fluorine Substituent Effects on Bond Dissociation Energies; Coulombic Interactions

As previously discussed, substitution by fluorine in hydrocarbons gives rise to

nearly additive and sometimes opposite effects on C-H and C-C homolytic bond

dissociation energies. For example, the aforementioned 1-3 kcal mol' weakening

effect of a,a-difluoro substitution and the 3-5 kcal mol-1 strengthening of terminal C-H

bonds by p-fluorination lead to near cancellation in the case of pentafluoroethane and

1,1,2,2-tetrafluoropropane, yielding BDE values very near those of the parent

hydrocarbon (Table 3-3.) Furthermore, the 4.9 kcal mol1 strengthening brought about

by substitution at the breaking C-C bond is reinforced by an additional 3-5 kcal mol"1

upon further fluorine incorporation at the p position, leading to net increases of nearly 10

kcal mol-1 over the parent in the C2-C3 homolysis of 1,1,1,2,2-pentafluoropropane and

cleavage of the terminal C-C bond of 2,2,3,3-tetrafluorobutane.

The opposite effects of aa-difluoro substitution on C-H and C-C bond

dissociation energies bear special mention. Experimental BDE values in the

fluoromethanes are in accord with the general RSE expectations of Pasto (Table 2-7) in

that although substitution at a radical center by a single fluorine is stabilizing, its further

incorporation leads to a successive decrease in RSE, resulting in net destabilization for

CF3'. This is consistent with the incremental strengthening of C-F bonds along the

fluoromethane series, leading to a C-H BDE in CF3H which is 1.9 kcal mol1 stronger

than that of methane itself (methane BDE, 104.8 kcal morl1; see discussion below Table

2-2) and the comparatively weaker C-H bonds in CHsF and CH2F2.

The considerable stability of the 2,2-difluoroalkanes relative to their 1,1-difluoro

isomers, demonstrated by the isodesmic reaction in Equation 3-2 (calculated from

B3LYP/6-31G(d) total energies and zero-point corrections) provides the underlying

reason for why the stability trends observed above do not extend to C-C bonds.

CH3CF2CH3 + CH3CH3 CH3CF2H + CH3CH2CH3 (3-2)

AErxn = + 7.8 kcal mol1

In Chapter 2, the Wiberg rationale of electrostatic attraction for the incremental

strengthening and shortening of C-F bonds in the fluoromethanes was introduced.

Similarly, it is found that such a Coulombic-based argument sufficiently explains the

observed effects of fluorine substitution on C-H and C-C bond dissociation energies.

Atomic charges for select hydrofluorocarbons based on the B3LYP/6-31G(d)

electrostatic potential (Merz-Kollman radii) are provided in Table 3-8.

The dipolar nature of the C-C bond in 1,1,1-trifluoroethane and its resultant

increase in BDE relative to ethane and hexafluoroethane (Table 2-3) was first postulated

by Rodgers.16 Such stabilization due to increased C-C bond ionicity is seen in

Equations 3-3 (derived from experimental heats of formation188) and 3-4 (from B3LYP/6-

31G(d) total and zero-point energies).

Table 3-8. Atomic Charges in Hydrofluorocarbons, Based on B3LYP/6-31G(d) Density.

Hvdrofluorocarbon LF _Cg LC "HXHoi

CaH3Ca,,H3 -0.055 +0.018

C,,H2F2 -0.200 +0.320 +0.041

CpH3CaF2H -0.228 +0.467 -0.386 +0.020 +0.110(2H)
1,1 -Difluoroethane +0.135 (1 H)

CH3C,,F2CpH3 -0.245 +0.631 -0.477 +0.133 (4H)
2,2-Difluoropropane +0.140 (2H)

CF3CF3 + CH3CH3 2 CH3CF3 (3-3)

AErxn = -16.9 kcal mol"1

CH3CH2CH2CH3 + CH3CF2CF2CH3 2 CH3CF2CH2CH3 (3-4)

AErxn = -5.0 kcal mol1-

The significant electrostatic attraction between adjacent carbon atoms in both

CH3CF2H and CH3CF2CH3 is readily apparent from the data in Table 3-8, providing an

explanation for the strengthening of these bonds relative to their hydrocarbon or

perfluorocarbon analogues. In addition, C-H repulsion in difluoromethane and 1,1-

difluoroethane is predicted, in accord with the observed weakening of these bonds

compared to those of methane and ethane. Conversely, the strong attraction between


the P carbon and hydrogen atoms of CH3CF2H (XC, -0.386; xHavg, +0.118) and

CH3CF2CH3 (.C, -0.477; XHavg, +0.135) is consistent with their greater theoretical and

experimental BDEs.


Based on time-resolved kwd measurements, absolute rate constants for

hydrogen abstraction from tri-n-butyltin hydride have been determined for a series

partially fluorinated radicals. The reactivities of such radicals towards nBu3SnH follow

those of addition to alkenes. The enhanced reactivity of a,a-difluoroalkyl radicals in

hydrogen abstraction reactions derives exclusively from their pyramidal geometry.

p-Fluorination leads to a favorable combination of polar and thermodynamic factors in

both addition and hydrogen transfer reactions, giving rise to the exceptional reactivity of

CF3CF2* and the perfluoroalkyl radicals as a whole. In Chapter 4, the kH values so

obtained are utilized in the determination of absolute rates of cyclization for partially

fluorinated 5-hexenyl radicals.

A self-consistent rationale for the impact of fluorine substitution on C-H and C-C

bond dissociation energies based on electrostatic considerations was offered, providing

new understanding of the thermochemistry of bonding and radical stabilization in





The intramolecular addition reactions of 5-hexen-l-yl radicals continue to attract

the attention of synthetic and physical organic chemists alike. Such cyclizations to

(predominantly) 5-exo products have been utilized as probes for the detection of radical

intermediates and as basis reactions for the competitive determination of absolute

kinetic data for a number of free radical transformations.16 Rationalization of the rates,

and especially the regio- and stereochemistry, of intramolecular radical additions on the

basis of force field63'64'66'87 and molecular orbital62,65'68'68 techniques has proven to be one

of the greatest successes of theory in the prediction of organic reactivity. Due in no

small part to such structure-reactivity studies, application of free radical methodology to

the singular and tandem construction of 5-membered rings has been equally exploited,

providing for the assembly of functionalized organic systems under mild conditions, often

accomplished with a high degree of stereocontrol.11'15

Determination of absolute rates of cyclization of per- and other highly fluorinated

5-hexenyl systems86,89 170 have aided in solidifying the understanding of the effect of

fluorine substitution on the reactivity of organic radicals, though at the same time

generating a number of new questions, particularly with regard to cyclization


In order to examine the potentially more subtle influences of partial fluorination

on 5-hexenyl radical reactivity, and to obtain a set of data through which the effect of

incremental gem-difluoro substitution along the aliphatic portion of the 5-hexenyl chain

may be assessed, absolute rates of 5-exo and 6-endo cyclization for some partially

fluorinated 5-hexenyl radicals have been determined based on competitive kinetic

technique and the absolute rates of hydrogen abstraction obtained in Chapter 3.

Precursor Syntheses and Competitive Kinetic Studies

As in the bimolecular addition versus hydrogen abstraction competition studies,

bromide precursors were utilized in the generation of partially fluorinated 5-hexenyl

radicals. Photolysis by UV irradiation (Rayonet photoreactor) in the presence of known,

varying concentrations of hydrogen atom donor, carefully adjusted to ensure pseudo-first

order kinetic behavior and to allow for accurately measurable amounts of cyclization and

hydrogen abstraction products, provided the kinetic ratio kH I kcn. Absolute rate

constants for 5-exo and (where applicable) 6-endo cyclization were then determined

from the known value of hydrogen abstraction rate constant kH, illustrated in Figure 1-18

and in greater detail below.

1.1-Difluorohex-5-en-1-vyl Radical (128)

Synthesis of bromide 135 was achieved in six steps in ca. 14.5% overall yield,

starting from commercially available 3-buten-l-ol (129, Figure 4-1). Curiously, direct

addition of dibromodifluoromethane to 129 could not be induced, even through extended

reaction time at elevated temperatures. Although the presence of the alcohol

functionality in the alkene starting material would not have been expected to exhibit a

detrimental effect (in light of the hydroxylic nature of the ethanolamine / terft-butanol

cosolvent medium) protection of the hydroxyl moiety as its tert-butyldimethylsilyl ether

130 (TBDMSCI, imidazole in dimethylformamide) followed by dibromodifluoromethane

addition indeed afforded 1,3-dibromo-1,1-difluoro adduct 131 in good yield. Highly

selective displacement of the internal bromine yielded bromodifluoromethyl derivative

129 130

CuCI (cat.)


BrF2C\7 O






Figure 4-1. Preparation of 6-Bromo-6,6-difluorohex-1-ene, Precursor to 1,1-Difluorohex-
5-en-1-yl Radical 128.

132 with virtually no overreduction product, as monitored via 1"F NMR through high

conversion of starting material. Lewis acid deprotection via the method of Cort193 and

subsequent pyridinium chlorochromate oxidation provided aldehyde 134, further

subjected to Wittig olefination to yield precursor 135.

Generation of 128 was achieved via irradiation of a solution of 135 in CeD6 in the

presence of excess nBu3SnH (Figure 4-2) and an internal standard of a,ca,a-

trifluorotoluene. Direct capture of 128 by hydrogen atom donor afforded reduction

product 6,6-difluorohex-l-ene 136, whereas intermediate 137, subsequently trapped by

nBu3SnH to yield spectroscopically observable cyclization product 138, was generated

via irreversible, unimolecular rearrangement with rate constant kc5 (no 6-endo cyclization

was observed, within NMR detection limits, (ca. 4%) for 128). During the course of the


r2C \/\ OTBDMS


2 \AH


CF2Br hv ^[ ]
135 C6D6 128

nBu3SnH /,. /,\/CF2H + nBu3Sn.

S k136

F_ /nBu3SnH F
kc5 F] F

137 138
+ nBu3Sn

Figure 4-2. kH / kc Competitive Kinetic Scheme for 1,1-Difluorohex-5-en-1-yl Radical

reaction, tributylstannyl radicals generated by transfer of hydrogen atom from nBu3SnH
to 128 and 137 served to propagate the chain process via bromine abstraction from 135.
Product ratios for varied concentrations of nBu3SnH were determined by 19F
NMR analysis according to the pseudo-first-order relation in Equation 4-1,

[136] [kH] [1281 [nBu3SnH (4-1)
[138] [kC5] [128]

a plot of which obtained for each data point versus nBu3SnH concentration providing the
ratio kH / kcs. Exceptionally clean spectra and high mass balances were obtained for
each kinetic point, indicating the efficiency of the radical chain process and reliability of
the obtained rate constant ratios. A partial 19F spectrum of the first of six data points is
provided in Figure 4-3, a doublet of triplets (-CF2H, 0 -116.2) observed for 136 versus

overlapping doublets of doublets of triplets at 4 -100.3 and -107.8 for the diastereotopic
-CF2- resonances of 138. Kinetic data and product yields are given in Table 4-1, a plot
of which found in Figure 4-4. The slope of the line (2.57 + 0.05) in conjunction with the

-96 -98 -100 -102

-104 -106 -108 -110

-112 -114

-116 -118

Figure 4-3. Partial 19F NMR Spectrum of
Difluorohex-5-en-l-yl Radical 128.

Table 4-1. Competitive Kinetic Data for kH
Radical 77.








[nBu3SnH 1







Data Point 1 for kH I kc Competition of 1,1-

/ kc Competition of 1,1-Difluorohex-5-en-1-yl








% Yield







m = 2.57
b =-0.175
r I = 0.999

0.6 0.7 0.8 0.9 1.0 1.1

Figure 4-4. Plot of the Data in Columns 2 and 3 of Table 4-1.

4.0 -
2.5 -
2.0 -

known absolute rate constant for hydrogen atom abstraction from nBu3SnH by 1,1-

difluorohex-1-yl radical 77, 9.1 (+ 1.7) x 106 M"1 s'1, resulted in a kcs value of

3.5 ( 0.59) x 106 s1 for 5-exo closure of 128, with errors in kc reflecting both the least-

squares fit of the line and propagated error in kH. Syntheses of hydrogen atom transfer

and cyclization products 136 and 138 were performed as shown in Figure 4-5.

4 ,/CF2Br nBu3SnH ___/ /,/CF2H
135 Mesitylene 136

CH2C12 F

139 138

Figure 4-5. Preparation of Hydrogen Abstraction and 5-Exo Cyclization Products 136
and 138.

2.2-Difluorohex-5-en-1-vyl Radical (140)

Bromide 144 was obtained in a three-step synthesis starting from 1,2-epoxy-5-

hexene (Figure 4-6). Regiospecific ring opening by a Corey94 procedure afforded

bromohydrin 142, converted to the corresponding a-haloketone via Jones oxidation.

Treatment of 143 with DAST in dichloromethane afforded precursor 144 in 40.9% overall

yield, purified by preparative GC for competitive kinetic study.

In contrast to the virtually regiospecific 5-exo closure of 128, a broad singlet

resonance at -95.8, comprising approximately 9% of cyclized products, was observed

in the 1"F NMR spectra for the kH / kc competition of 140. This is attributed to competing

6-endo cyclization to 148 (Figure 4-7), the presence of which was confirmed by spectral

comparison with that of an authentic sample of 149.

THF / H20


Na2Cr2O7 / H2S04






Figure 4-6. Preparation of 6-Bromo-5,5-difluorohex-1-ene, Precursor to 2,2-Difluorohex-
5-en-1-yl Radical 140.

,/xv, CF2CH2Br






F ]





Figure 4-7. kH I kc Competitive Kinetic

Scheme for 2,2-Difluorohex-5-en-1 -yl Radical











Cyclizations of 1p,3-difluoroalkyl radicals have appeared in the synthetic

literature,195 utilized in the generation of alkoxy-substituted gem-difluorocyclopentane,

cyclohexane, and tetrahydropyran derivatives, though reported to undergo addition in an

exo-specific manner. However, the regiochemical behavior exhibited in the cyclization of

140 provided experimental verification of that previously predicted on the basis of ab

initio calculations, performed as part of the present study and elaborated upon in the

Discussion section of the chapter. Competition plots for kH / kc5 and kH I kce are found in

Figures 4-8 and 4-9, respectively. Preparation of hydrogen abstraction and 5-exo and 6-

endo cyclization products was performed as shown in Figure 4-10.

6 Coefficients:
I- m=12.5
S5 b =-0.924
r2 = 0.999

2 -1 I I I I11
0.25 0.30 0.35 0.40 0.45 0.50 0.55 0.60


Figure 4-8. Plot of kH I kc5 Competition of 2,2-Difluorohex-5-en-1-yl Radical 140.

60 coefficients:
0 m=132
50 b = -14.0 &
40 r 2 = 0.995
-" 40

S30 -

20 ------
0.25 0.30 0.35 0.40 0.45 0.50 0.55 0.60


Figure 4-9. Plot of kH / kc6 Competition of 2,2-Difluorohex-5-en-1-yl Radical 140.

F2CH2Br nBu3SnH M-W /VCF2CH3
144 Mesitylene 145


150 147

O F\ F

.^ DAST .

151 149

Figure 4-10. Preparation of Hydrogen Abstraction and 5-Exo and 6-Endo Cyclization
Products 145,147, and 149.

1.1,2,2-Tetrafluorohex-5-en-1-vyl Radical (152)

Bromide 101 (Halocarbons, Inc.) served as the precursor to a,a,pj3-

tetrafluorinated radical 152. However, attempts at determination of accurate kH I kc

ratios using nBu3SnH as a trapping agent met with failure, leading primarily to reduction

product 153, with only minor amounts of 155 and 177 evident in the 19F NMR baseline

which could not be integrated accurately over a span of hydrogen atom donor

concentrations (Figure 4-11). In principle, lowering the concentration of both radical

precursor (typically in the 0.05 0.1 M range) and trapping agent (while still maintaining

pseudo-first order conditions) should effectively decrease the amount of reduction

product and allow for a greater degree of cyclization to be observed. However, too great

of a decrease in precursor concentration leads to decreased NMR signal to noise ratios,

the necessity of longer acquisition times per sample, and increased potential for the

introduction of systematic error.

hv F 1
^?c^-\ CFzBr ----- ",v.' c,.C F2
F2 nBu3SnH CF2
F2 F6D6
101 c6D 152

nBu3SnH ^..c-CF2H
kH F2


virtually no cyclization products
kc5 observed


Figure 4-11. Attempted kH / kc Competition of 1,1,2,2-Tetrafluorohex-5-en-1-yl Radical
152 with nBu3SnH as Trapping Agent.

As it was evident that any cyclization reaction of 152 occurred with a rate

constant too low to be competitive with transfer of hydrogen from nBu3SnH, attention

was turned to alternative trapping agents. With the rate of hydrogen atom transfer to

perfluoroalkyl radicals by a number of reducing agents having been accurately

determined (Table 2-10), it was decided to investigate the suitability of

tris(trimethylsilyl)silane ((TMS)3SiH) as a competitive trapping agent for the calibration of

cyclization rate constants for 152, due to its approximately four-fold decrease in

hydrogen transfer rate to perfluoroalkyls relative to nBu3SnH. For such a competition to

be of kinetic value, however, it was necessary to determine rate constant kH for

tetrafluoroalkyl radical 100 with (TMS)3SiH, using its known rate of addition to styrene as

a competing basis reaction. The plot for the kH ((TMS)3SiH) / kadd (styrene) competition

of 100 is provided in Figure 4-12.


With the kH ((TMS)3SiH) value of 1.8 ( 0.1) x 107 M1 s1 for 100 in hand (which,

along with its kH (nBu3SnH) of 9.2 ( 0.8) x 107 M1 s-', (Table 3-2) may be compared

with 5.1 x 107 M1 s1 and 2.0 x 108 M1 s1, respectively, for perfluoro-n-alkyl radicals;

Table 2-10) rate constants kc5 and kc6 for 152 were then determined (Figures 4-13 and

4-14.) Use of this slower hydrogen transfer agent allowed for sufficient competitive

(including significant 6-endo) cyclization such that accurate kH / kcn ratios could be

obtained. Isolation of products 153, 155, and 157 was achieved by slow syringe pump

addition of nBu3SnH to a heated, irradiated solution of 101 in mesitylene (Figure 4-15).

2.2 -
S2.0 Coefficients:
0 m=0.913
1.8 -
,e- * b =0.189
1. r=0.999
1.2 -
1 .0 1 I I I 1 I
0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0
[ (TMS)3SiH ] / [ C6H5CH=CH2]

Figure 4-12. Plot for kH ((TMS)3SiH) / kadd (Styrene) Competition of 1,1,2,2-
Tetrafluorohex-1-yl Radical 100.

2.5 -
S.. Coefficients:
u) 2.0 m=2.11
tOe b =-0.242 J
S 1.5 r =0.998


0 .5 1 1 1 1 1-1
0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2
[ (TMS)3SiH ]

Figure 4-13. Plot of kH / kc5 Competition of 1,1,2,2-Tetrafluorohex-5-en-1-yl Radical 152.

0.6 0.7 0.8 0.9 1.0

[ (TMS)3SiH ]

Figure 4-14. Plot of kH I kc6 Competition of 1,1,2,2-Tetrafluorohex-5-en-1-yl Radical 152.

^, -^^ .^-CF2Br


.. CF2


,,,,s,/ CF2H









Figure 4-15. Preparation of Hydrogen
Products 153,155, and 157.

Abstraction and 5-Exo and 6-Endo Cyclization

1.1 1.2


Absolute rate constants of cyclization for radicals 128, 140, and 152 are given in

Table 4-2. For comparison, such kcs and (where applicable) kc6 values for parent

hydrocarbon 1 and fluorinated radicals 65 and 68 are also provided, the latter two

systems along with those of the current study found to give rise to the greatest impact on

cyclization kinetics and regiochemistry. Recent studies of 5-hexenyl systems bearing

vinylic fluorine substituents have demonstrated that the effect of such substitution is

relatively minor, with no 6-endo products observed within the detection limits imposed by

NMR analysis and kc5 (re) values with respect to 1 ranging from ca. 0.09 to 2.3.170

Cyclization Kinetics

Rates of intramolecular addition of partially-fluorinated radicals to alkenes should

be governed by the same combination of steric, polar, and thermodynamic factors which

influence the reactivity of their intermolecular counterparts. As seen by comparison of

the data in Tables 3-2 and 4-2, the reactivity characteristics of the above radicals in

unimolecular cyclization reactions, particularly 5-exo closure, generally reflect those

observed in bimolecular additions. This is logical in light of the similarity of their

transition structures, elaborated upon in Chapters 1 and 3.Y

The pyramidal nature of a,a-difluoroalkyl radicals, combined with the more

favorable thermodynamics of C-C bond formation involving fluorinated carbon (see

related discussions in Chapters 2 and 3, along with cyclization transition structures and

energies of reaction below) provide sufficient explanation for the 13-fold increase in rate

of 5-exo ring closure of 128 relative to 1. The factor of 22.5 observed for addition of 77

versus 124 to styrene is consistent with the observed cyclization rate ratios.

The increase in kcs of 4.1 enjoyed by 140 parallels that of bimolecular addition of

125 to styrene, (4.3) due to its increased electrophilicty over both hydrocarbon 1 and


Table 4-2. Absolute Rate Constants for 5-Exo and 6-Endo Cyclization of Partially
Fluorinated 5-Hexenyl Radicals. Rate Constants are for 303 K; Relative kcn Values in

Cyclization Reaction

kc5, 105 s-1

k6, 105 s-1

0.05 a

+ 0


N/A b,c






11 (3.8) b

87 (t 4.1) b

1.1 (0.34)" b

19 ( 1.1) b

F2C-.. -CF2

+ r CF2
F2C,. CF2

440 ( 46) d

110 (1.7) d

8 Reference 16. b Current Study. c 6-Endo Cyclization Not Observed Within 19F NMR
Detection Limits (Approximately 4%). d Reference 170.

jF2 C





CF cF2





1 CFF2
` CCF2




52 (6.4) d

35 ( 4.4) d

a,a-difluorocarbon 128 and greater exothermicity of addition relative to n-alkyls, though

tempered by the effectively planar, n-nature of its radical center.

a,a,p3,13-Tetrafluorinated radical 152, as in the case of bimolecular additions,

benefits from favorable thermodynamics of addition as well as its electrophilicty and

G-character, leading to the 32-fold increase in kcs compared to parent 1. It should be

noted that such unimolecular cyclizations possess an inherent entropic advantage over

their bimolecular analogues, generally proceeding with log A values ca. 2 units larger

than those for the latter18 and resulting in a leveling of rate ratios relative to

intermolecular additions.

Upon additional fluorination of the aliphatic moiety of the 5-hexenyl chain (65, 68)

such radicals undertake perfluoroalkyl character, leading to further increase in reactivity

akin to that observed for C7Fs* (127, kadd (.e) = 383) versus CH3CH2CF2CF2" (99,

kadd (re1) = 167, relative to n-alkyl) in bimolecular additions to styrene. Geminal

difluorination at the allylic position (68) serves to diminish the transition state SOMO-

HOMO interaction, and hence kcs and kcm, relative to 65.

Cyclization Regiochemistrv

The significant degree to which 152, 65, 68, and even 140 undergo 6-endo

cyclization is particularly striking, with six-membered ring formation in 140 occurring with

a rate nearly half, and 65 and 68 more than an order of magnitude greater than, that of

5-exo closure for hydrocarbon 1. In comparison, 5-hexenyl systems bearing alkyl

substituents along the aliphatic fragment exhibit regiochemical profiles similar to that of

the unsubstituted parent.70'71

The question of potential reversibility in the above cyclizations has been

addressed, in light of the greater relative thermodynamic stability of secondary

cyclohexyl radicals. Upon independent generation of 5-exo adduct radical 69 from

precursor 1-(iodomethyl)-2,2,3,3,4,4,5,5-octafluorocyclopentane in the presence of

hydrogen atom donor triethylsilane in C6D6, the only product observed after complete

consumption of starting material was that resulting from direct capture of 69 by Et3SiH.170

The lack of 6-endo or ring-opened products originating from 69, coupled with the ab initio

predictions based on relative energies of cyclization transition structures described

below, demonstrates that the regiochemical characteristics of fluorinated 5-hexenyl

radical cyclizations are indeed kinetic in nature.

Of further note is that system 68, which undergoes the greatest percentage

(24.1%) of 6-endo closure (that is, exhibiting the least selectivity) is not the most

reactive. Hexafluoro system 65, though forming 67 with a rate constant 1.5 times that of

analogous closure of 68 to 70, does so only to an extent of 10.6% of total cyclized


A combined ab initio I molecular mechanics approach has allowed for accurate

regiochemical predictions for a number of alkyl and heteroalkyl intramolecular radical

additions.66 In order to examine the effect of the degree and location of fluorine

substitution on transition structure geometry and energetic, as well as on activation

barriers and reaction enthalpy, the "chair-like" and "boat-like" 5-exo and 6-endo

cyclization transition structures for the parent hydrocarbon and various fluorinated

5-hexenyl systems, along with their respective open-chain radicals and products of

5- and 6-membered ring closure, have been investigated with ab initio techniques.

In accordance with a UMINDO/3 investigation of Bischof,62 the lowest energy

conformation of 5-hexenyl radical 1 was found to be an all-trans methylene chain in a

gauche orientation with the internal vinyl hydrogen (Figure 4-16.) Alignment of the singly

occupied orbital of 1 with the adjacent C-H bond was found to be slightly preferred (ca.

0.1 kcal morl1) over similar C-C alignment at the UHF/6-31G(d) + ZPE level.

From the calculated structures and energies of cyclization products

(cyclopentylmethyl and cyclohexyl radicals) it was possible to compute energies of

reaction for hydrocarbon 1 and its fluorinated analogues. Total energies of reactant and

product radicals and exothermicities of 5-exo and 6-endo cyclization for 1,128, 152, and

65 are provided below in Table 4-3.

r(C=C) = 1.318A

Figure 4-16. Lowest Energy Conformation of 5-Hexen-l-yl Radical 1; SOMO (On Right)
Aligned with Adjacent C-H Bond. UHF/6-31G(d) Optimized Geometry.

As expected from C-C BDE data, (Table 3-4) intermolecular additions of the

fluorinated species are, as a whole, more exothermic than for parent 1. However, no

direct correlation exists between either absolute rates of 5-exo and 6-endo addition or

relative percentage of 6-membered ring formation and its corresponding reaction

exothermicity. Although a steady increase in both kc5 and kce is observed along the

series (1 -> 128 --> 152 -> 65), both cyclizations of 65 are predicted to be less

exothermic than those of 152. Furthermore, relative enthalpies (AEn(15.s) AErxn(.e)) are

found to rise with the degree of fluorination, favoring 6-endo closure in consistent

manner for both levels of theory employed. This is at variance with the lesser extent of

6-endo closure in 65 compared to 152.

Total, zero-point, and relative energies along with pertinent geometrical

parameters for the UHF/6-31G(d) cyclization transition structures of 1, (depicted in

Figures 1-11 1-14) 128, 140, 152, and 65 are reported in Table 4-4. Although the

calculated energy differences between "chair" and "boat" forms of either 5-exo or 6-endo

transition structures are quite consistent among the theoretical methods, energies of the

"6-endo-chaif' and "6-endo-twist-boaf' structures relative to the "5-exo-chaie" and

"5-exo-boaf' appear to be overestimated at the PMP2/6-311G(d,p)//UHF/6-31G(d) level

compared to both UHF and QCISD(T) results. Bearing this in mind, relative transition

Full Text
xml version 1.0 standalone yes
PageID P431
ErrorID 1012
ErrorText orientation error: 0143-0177