Effect of Substituants on the Structure of the Vinyl Radical: Calculations and Experiments

Article abstract of DOI:10.1021/jo962373h

Literature reports indicate either a linear or a bent configuration for a-substituted vinyl radicals. In order to get a better insight into this structural question, calculations of the structure of some a-substituted vinyl radicals were conducted by the DFT approach [BLYP/6-31G(d,p) and B3LYP/ 6-311G(2d,2p)]. For vinyl radicals bearing α-type substituents (Me, SH, Cl, OH, F), the bent form is found to be the minimum energy structure; the inversion barrier of the E and Z forms is found to markedly increase as the electronegativity of the group increases. Vinyl radicals bearing π-type substituents (CH=CH₂, CHO, CN, C₆H₅) are found to be linear. The effect of β-substituents is much lower, and the β-fluoro- and β-ethenylvinyl radicals, representative of α- and π-type substituents, are calculated to be bent. Comparison with other literature calculations, obtained by different methods for some similar vinylic radicals, supports our computational results. Experimental support was also sought, with particular regards to the inversion barrier of the bent radicals. Since a very high barrier is calculated for the α-fluorovinyl radical, its inversion rate could be slower than the "sampling" time of a proper chemical reaction. Indeed, the α-fluorovinyl radicals, generated from PhCH=C(F)Br by reaction with Bu₃SnH/AIBN, retained the configuration of their precursors in the overall hydrodebromination process.

Full text of DOI:10.1021/jo962373h

4072
J . Org. Chem. 1997, 62, 4072-4077
Effect of Su bstitu en ts on th e Str u ctu r e of th e Vin yl Ra d ica l:
Ca lcu la tion s a n d Exp er im en ts
Carlo Galli,* Alessandra Guarnieri, Heinz Koch,† Paolo Mencarelli,* and Zvi Rappoport‡
Dipartimento di Chimica and Centro CNR Meccanismi di Reazione, Universita` “La Sapienza”,
00185 Roma, Italy, Department of Chemistry, Ithaca College, Ithaca, New York 14850, and Department of
Organic Chemistry, The Hebrew University, J erusalem 91904, Israel
Received December 19, 1996X
Literature reports indicate either a linear or a bent configuration for R-substituted vinyl radicals.
In order to get a better insight into this structural question, calculations of the structure of some
R-substituted vinyl radicals were conducted by the DFT approach [BLYP/6-31G(d,p) and B3LYP/
6-311G(2d,2p)]. For vinyl radicals bearing σ-type substituents (Me, SH, Cl, OH, F), the bent form
is found to be the minimum energy structure; the inversion barrier of the E and Z forms is found
to markedly increase as the electronegativity of the group increases. Vinyl radicals bearing π-type
substituents (CHdCH2, CHO, CN, C6H5) are found to be linear. The effect of â-substituents is
much lower, and the â-fluoro- and â-ethenylvinyl radicals, representative of σ- and π-type
substituents, are calculated to be bent. Comparison with other literature calculations, obtained
by different methods for some similar vinylic radicals, supports our computational results.
Experimental support was also sought, with particular regards to the inversion barrier of the bent
radicals. Since a very high barrier is calculated for the R-fluorovinyl radical, its inversion rate
could be slower than the “sampling” time of a proper chemical reaction. Indeed, the R-fluorovinyl
radicals, generated from PhCHdC(F)Br by reaction with Bu3SnH/AIBN, retained the configuration
of their precursors in the overall hydrodebromination process.
In a recent investigation on the SRN1(V) reaction,1 as
we have dubbed the vinylic counterpart of the SRN1
process of aryl halides2 (Scheme 1, where RX is the vinyl
or aryl substrate, and Y- is a suitable nucleophile),
definite evidence has been gathered for the intermediacy
of vinyl radicals in the propagation cycle.
Sch em e 2
Sch em e 1
RX
-8 RX•- (initiation)
(1)
(2)
(3)
(4)
(5)
hν/Y
from the solvent (SH, in eq 5), to give the hydrodehalo-
genated product, and combination with the nucleophile,
to eventually afford the substitution product (eqs 3 and
4). The stereoconvergence observed indicates that in-
termediate radical 5 has either a linear sp structure or
an “average linear” structure due to a rapidly intercon-
verting E,Z mixture of sp2 bent radicals. From our
results it was difficult to decide between the two struc-
tural hypotheses.1
RX•- f R• + X-
R• + Y- f RY•-
RY•- + RX f RY + RX•-
R• + SH f RH + S•
As for the stereochemical course of the SRN1(V) process,
investigation of the photostimulated reaction of the E and
Z isomers of a suitable vinyl halide (i.e., 1) with pina-
colone enolate ion (Me3CCOCH2-; 2) as the nucleophile
revealed complete loss of the original stereochemistry of
(E)-1 and (Z)-1 in the substituted (3) and hydrodehalo-
genated (4) products (Scheme 2, where An is anisyl).1
This outcome has been attributed to the intermediacy
of a substituted R-phenylvinyl radical [AnC(Ph)dC(Ph)•,
5], which partitions between hydrogen atom abstraction
If we remember that the structures of the “close
neighbors” of the vinyl radical, viz. the vinyl anion and
the vinyl cation, are bent and configurationally stable for
the former,3 while linear or bridged for the latter,4 an
intermediate structure for the vinyl radical seems rea-
sonable.5,6
(3) (a) Miller, S. I.; Lee, W. G. J . Am. Chem. Soc. 1959, 81, 6313. (b)
Walborsky, H. M.; Turner, L. M. Ibid. 1972, 94, 2273. (c) Corey, E. J .;
Ulrich, P.; Fitzpatrick, J . M. Ibid. 1976, 98, 222. (d) Froelicher, S. W.;
Freiser, B. S.; Squires, R. R. Ibid. 1986, 108, 2853. Here, an E f Z
isomerization barrier of 26.6 kcal/mol for 1-propenyl ions is calculated.
(e) Chou, P. K.; Kass, S. R. Ibid. 1991, 113, 4357.
(4) (a) Mayr, H.; Schneider, R.; Wilhelm, D.; Schleyer, P. v. R. J .
Org. Chem. 1981, 46, 5336. (b) Rappoport, Z. In Reactive Intermediates;
Abramovitch R. A., Ed.; Plenum Press: New York, 1983; Vol. 3, pp
427-615. (c) Berkowitz, J .; Mayhew, C. A.; Ruscic, B. J . Chem. Phys.
1988, 88, 7397. (d) Lindh, R.; Rice, J . E.; Lee, T. J . Ibid. 1991, 94,
8008.
† Ithaca College.
‡ The Hebrew University.
X Abstract published in Advance ACS Abstracts, May 15, 1997.
(1) Galli, C.; Gentili, P.; Guarnieri, A.; Rappoport, Z. J . Org. Chem.
1996, 61, 8878.
(2) (a) Bunnett, J . F. Acc. Chem. Res. 1978, 11, 413. (b) Rossi, R. A.;
De Rossi, R. H. Aromatic Substitution by the SRN1 Mechanism; ACS
Monograph 178; American Chemical Society: Washington, DC, 1983.
(c) Rossi, R. A.; Pierini, A. B.; Pen˜e´n˜ory, A. B. In The Chemistry of
Functional Groups. Supplement D2: The Chemistry of Halides, Pseudo-
halides and Azides; Patai, S., Rappoport, Z., Eds.; Wiley: Chichester,
1995; Chapter 24.
(5) J enkins, P. R.; Symons, M. C. R.; Booth, S. E.; Swain, C. J .
Tetrahedron Lett. 1992, 33, 3543.
(6) Gillespie, R. J . Chem. Soc. Rev. 1992, 21, 59.
S0022-3263(96)02373-0 CCC: $14.00 © 1997 American Chemical Society

Effect of Substituents on the Structure of the Vinyl Radical
J . Org. Chem., Vol. 62, No. 12, 1997 4073
ized by a very small inversion barrier. The latter would
make the two intrinsically nonequivalent vinylic â-hy-
drogens appear equivalent, similar to a linear structure.
Indeed it was this “apparent” equivalence of the â-hy-
drogens on the esr timescale that suggested the linear
structure in the earlier esr studies.8a,b Very recently, a
new esr investigation confirmed a linear structure for an
R-phenylvinyl radical intermediate.8f Consequently, the
structural description of the R-substituted vinyl radicals
appears elusive and certainly difficult to determine by
experimental methods. A computational approach may
be easier for settling the question. The desirable goal of
such a computational endeavor would be to explore the
potential energy surface for a number of substituted vinyl
radicals, in order to ascertain if a linear or bent structure
represents the energy minimum and, whenever appropri-
ate, to calculate the energy barrier to inversion between
the bent isomeric forms.
There exists circumstantial esr evidence7 that the
parent vinyl radical has two sp2 bent structures that are
rapidly interconverting (4 × 1010 s-1 at 300 K), due to an
inversion barrier as low as 3 kcal/mol.7a The sp linear
structure could be the transition structure for this
inversion, instead of a real intermediate.
It is reported that R-substituents can modify this
situation to a major extent. For example, the R2CdC-
(R′)• radical, where R′ ) Ph, appears linear by esr
investigation8 and by muon spin rotation spectroscopy,9
whereas when R′ ) Cl or Br, the evidence favors a bent
stable structure.10 Analysis of the stereochemistry of the
olefinic products formed in a radical addition to 1-alkynes,
where an R-substituted alkenyl radical is the reactive
intermediate, supports11 a bent, stable structure for the
R-MeO-alkenyl radical and points8f to a linear structure
for the R-CO2Me and R-SiMe3 vinyl radicals. However,
stereochemical results suggest11 that the R-CN and the
R-Ph vinyl radicals could be bent and rapidly equilibrat-
ing at elevated temperatures (240-260 °C), but are more
likely to be linear at lower temperatures. In the case of
R-alkylvinyl radicals, a bent and fast equilibrating struc-
ture was corroborated,8f,11 thereby confirming the struc-
ture deduced from esr investigation.7
In spite of the importance of vinyl radicals among the
fundamental intermediates of organic chemistry, it was
surprising to find, when we started this investigation,
that there was only a thorough study of the potential
energy surface of the parent ethenyl radical by post
Hartree-Fock ab initio methods.13 However, there was
no computational information about the relative energy
of the linear and bent forms of R-substituted vinyl
radicals. Likewise, no evaluation was available for the
energy barrier to interconversion of isomeric E,Z bent
vinyl radicals as a function of the substituent. A knowl-
edge of the barrier could allow the investigation of the
inversion rate, which is important for interpreting the
stereochemical outcome of reactions involving vinyl
radicals as intermediates.
For these reasons, and as an implementation of our
previous stereochemical investigation,1 we have under-
taken a theoretical scrutiny of the following R-substituted
vinyl radicals:
Conflicting literature reports appear to afflict mostly
π-type substituents, such as the phenyl or carbonyl
groups. These substituents, which could allow stabiliza-
tion of the adjacent vinyl radical by delocalization of the
unpaired electron, are likely to impose a linear structure
at the radical center in order to achieve maximum orbital
overlap, as it is shown for the R-CN group.
This expectation appeared to be confirmed by an old
esr study on the pertinent radicals generated in solid
matrices.8a,b However, a more recent esr investigation
of these radicals, generated in solution by UV photolysis
of suitable peroxide precursors,12 questions this structure,
for at least the R-carbomethoxyvinyl and R-cyanovinyl
radicals, which seem to have a bent structure character-
These radicals were chosen in order to investigate the
effect of σ-type substituents (structures 6-11) and of
π-type substituents (structures 12-15) and bear some
resemblance with those experimentally examined by
us.1,14 Two â-substituted vinyl radicals (structures 16
and 17) were also studied. The results of the calcula-
tions, carried out by using the DFT approach on these
structures, are presented below. In order to compare
calculations and experiment, the stereochemistry and
products of a reaction leading to the formation of one of
these R-substituted vinyl radical intermediates was
investigated, and we present the results along with other
experimental literature data.15
(7) (a) Fessenden, R. W.; Schuler, R. H. J . Chem. Phys. 1963, 39,
2147. (b) Cochran, E. L.; Adrian, F. J .; Bowers, V. A. Ibid. 1964, 40,
213. (c) Fessenden, R. W. Ibid. 1967, 71, 74.
(8) (a) Bonazzola, L.; Feinstein, S.; Marx, R. Mol. Phys. 1971, 22,
689. (b) Neilson, G. W.; Symons, M. C. R. J . Chem. Soc., Perkin Trans.
2 1973, 1405. (c) Kasai, P. H.; Whipple, E. B. J . Am. Chem. Soc. 1967,
89, 1033. (d) Bennett, J . E.; Howard, J . A. Chem. Phys. Lett. 1971, 9,
460. (e) Panek, E. J .; Kaiser, L. R.; Whiteside, G. M. J . Am. Chem.
Soc. 1977, 99, 3708. (f) Rubin, H.; Fischer, H. Helv. Chim. Acta 1996,
79, 1670.
(9) Rhodes, C. J .; Roduner, E. J . Chem. Soc., Perkin Trans. 2 1990,
1729.
(10) (a) Liu, M. S.; Soloway, S.; Wedegaertner, D. K.; Kampmeier,
J . A. J . Am. Chem. Soc. 1971, 93, 3809. (b) Singer, L. A.; Kong, N. P.
Tetrahedron Lett. 1967, 643.
(11) Metzger, J . O.; Blumenstein, M. Chem. Ber. 1993, 126, 2493.
(12) Korth, H.-G.; Lusztyk, J .; Ingold, K. U. J . Chem. Soc., Perkin
Trans. 2 1990, 1997.
(13) Paddon-Row, M. N.; Pople, J . A. J . Phys. Chem. 1985, 89, 2768.
(14) (a) Galli, C.; Gentili, P.; Rappoport, Z. J . Org. Chem. 1994, 59,
6786. (b) Amatore, C.; Galli, C.; Gentili, P.; Guarnieri, A.; Schottland,
E.; Rappoport, Z. J . Chem. Soc., Perkin Trans. 2 1995, 2341.
(15) Simamura, O. Top. Stereochem. 1969, 4, 1.

4074 J . Org. Chem., Vol. 62, No. 12, 1997
Galli et al.
Ta ble 1. In ver sion En er gy Ba r r ier s a n d Bon d An gles
F ou n d for Ben t Ra d ica ls 6-11, 16, a n d 17 As Obta in ed
fr om DF T Ca lcu la tion s (BLYP /6-31G(d ,p )//BLYP /
6-31G(d ,p ). F or Ra d ica ls 6, 9, a n d 11 th e Cor r esp on d in g
Va lu es Obta in ed a t a High er Level of Th eor y (B3LYP /
6-311G(2d ,2p )//B3LYP /6-311G(2d ,2p )) Ar e Rep or ted in
P a r en th eses
Resu lts
Com p u ta tion a l Deta ils. Since it is known that the
electron correlation should be taken into account when
dealing with inversion barriers in vinyl radicals,13 and
with the aim of keeping the computational effort within
reasonable limits, we have chosen the DFT approach as
a theoretical method. It is known that this method may
achieve a greater accuracy than the Hartree-Fock (HF)
theory, at a modest increase in cost. The DFT approach
overcomes the major limitation of the HF theory, i.e. the
neglect of electron correlation, by including some of the
electron correlation effects. Its results may be compared
to the post HF MP2 method.16 The DFT approach has
been already used as a theoretical tool for the investiga-
tion of fluoro-substituted vinyl radicals.17 While our
study was in progress, a paper appeared where ab initio
calculations for a certain number of R-substituted vinyl
radicals, but neither including the phenyl and ethenyl
groups nor the formyl group, were reported.18 Some
results of this paper are in agreement with ours, but
others, in the case of the π-type substituents, are in
contrast with ours and with previous8 findings. Since
our approach differs from that employed in the paper
quoted above,18 it is interesting to compare the results
of the two approaches.
a
Corrected for the zero point energies.
All the calculations were carried out by using either
the Gaussian 92/DFT19 or the Gaussian 94 packages.20
The Density Functional approach (DFT) was used with
the Becke’s 1988 exchange functional, which includes the
Slater exchange along with corrections involving the
gradient of the density,21 and with the correlation func-
tional of Lee, Yang and Parr, which includes both local
and nonlocal terms22,23 (BLYP). A 6-31G(d,p) basis set,
a split valence basis set supplemented with polarization
p-functions on hydrogens and d-functions on heavy
atoms, was used. For each radical, starting both from a
linear structure and a bent structure, a full optimization
was carried out in order to find the nearest stationary
point. For all the structures obtained, the spin contami-
nation due to states of multiplicity higher than the
doublet state was negligible, in that the S2 parameter
was, in all cases, well within the 10% of the expectation
value for a doublet (0.75). The stationary points found
were characterized by a frequency calculation. For the
linear structure featuring one imaginary frequency, and
therefore found to be a saddle point, the normal mode,
corresponding to the imaginary frequency, was animated
by using the visualization program MolWin.24 In this
way it was verified that the displacements that compose
the mode lead to the two isomeric bent forms, thus
confirming that the linear form is indeed the transition
state for the inversion. For the radicals found to be bent,
the barrier of inversion was obtained as the difference
between the total energies of the bent and the linear
form. This energy difference was then corrected for the
differences in the zero point energies. In order to check
the reliability of the BLYP/6-31G(d,p) model used here,
the calculations for radicals 6, 9, and 11 have been
repeated at a higher level of theory by using the Becke’s
three-parameter functional, with the correlation func-
tional of Lee, Yang, and Parr (B3LYP). A triple ú (zeta)
basis set with polarization 2p-function on hydrogen and
2d-function on heavy atoms was used [6-311G(2d,2p)].
Com p u ta tion a l Resu lts. Radicals 6-11, bearing
σ-type substituents, were found to be bent. The corre-
sponding linear forms were characterized as the transi-
tion state for the inversion between the two isomeric bent
forms. The values of the energy barrier to the inversion,
together with the bond angles at the radical center
(namely, the angle between the double bond and the
substituent bond) found for the minimum-energy struc-
tures, are reported in Table 1. The nature of the
substituent greatly affects the inversion barrier and, to
a lesser extent, the bond angle at the radical center. In
Figures 1 and 2 the height of the barrier and the bond
angle values, respectively, are plotted against the elec-
tronegativity25 of the R-substituent. It is evident from
both Table 1 and Figures 1 and 2 that the results from
the B3LYP/6-311G(2d,2p) calculations strictly follow
those obtained from the BLYP/6-31G(d,p) model.
(16) Bartolotti, L. J .; Flurchick, K. In Reviews in Computational
Chemistry; Lipkowitz, K. B., Boyd, D. B., Eds.; VCH Publishers: New
York, 1996; Vol 7, Chapter 4.
(17) (a) Barone, V.; Adamo, C.; Russo, N. Chem. Phys. Lett. 1993,
212, 5. (b) Barone, V.; Adamo, C.; Russo, N. Int. J . Quantum Chem.
1994, 52, 963.
(18) Guerra, M. Res. Chem. Intermed. 1996, 22, 369.
(19) Frisch, M. J .; Trucks, G. W.; Schlegel, H. B.; Gill, P. M. W.;
J ohnson, B. G.; Wong, M. W.; Foresman, J . B.; Robb, M. A.; Head-
Gordon, M.; Replogle, E. S.; Gomperts, R.; Andres, J . L.; Raghavachari,
K.; Binkley, J . S.; Gonzalez, C.; Martin, R. L.; Fox, D. J .; Defrees, D.
J .; Baker, J .; Stewart, J . J . P.; Pople, J . A. Gaussian 92/ DFT, Revision
F.2; Gaussian, Inc.: Pittsburgh, PA, 1993.
(20) Frisch, M. J .; Trucks, G. W.; Schlegel, H. B.; Gill, P. M. W.;
J ohnson, B. G.; Robb, M. A.; Cheeseman, J . R.; Keith, T.; Petersson,
G. A.; Montgomery, J . A.; Raghavachari, K.; Al-Laham, M. A.;
Zakrzewski, V. G.; Ortiz, J . V.; Foresman, J . B.; Cioslowski, J .;
Stefanov, B. B.; Nanayakkara, A.; Challacombe, M.; Peng, C. Y.; Ayala,
P. Y.; Chen, W.; Wong, M. W.; Andres, J . L.; Replogle, E. S.; Gomperts,
R.; Martin, R. L.; Fox, D. J .; Binkley, J . S.; Defrees, D. J .; Baker, J .;
Stewart, J . P.; Head-Gordon, M.; Gonzalez, C.; Pople, J . A. Gaussian
94, Revision C.2; Gaussian, Inc.: Pittsburgh PA, 1995.
(21) Becke, A. D. Phys. Rev. A 1988, 38, 3098.
Radicals 12-15, bearing π-type substituents, were
found to be linear. For all of them the π system of the
substituent is orthogonal to the π system of the double
bond and coplanar with the p orbital at the radical center.
(22) Lee, C.; Yang, W.; Parr, G. Phys. Rev. B 1988, 37, 785.
(23) Miehlich, B.; Savin, A.; Stoll, H.; Preuss, H. Chem. Phys. Lett.
1989, 157, 200.
(24) Copyright 1993-1994 Pavel V. Ganelin. All rights reserved.
(25) Bergman, D.; Hinze, J . Angew. Chem., Int. Ed. Engl. 1996, 35,
150.

Effect of Substituents on the Structure of the Vinyl Radical
J . Org. Chem., Vol. 62, No. 12, 1997 4075
F igu r e 3. Minimum energy structures calculated (BLYP/6-
31G(d,p)) for radicals 11 (left) and 15 (right).
(120°) and a linear (180°) one for these radical centers.
The double bond-to-substituent angle gradually ap-
proaches the “normal” sp2 angle on increasing the elec-
tronegativity of the R-substituent (Figure 2). The height
of the barrier to inversion also depends strongly on the
electronic effect of the R-substituent, and a quasilinear
relationship is observed between the former and the
electronegativity of the R-substituent (Figure 1). In
comparison with the effect of the R-substituent, it can
be verified that the â-substituent effect (structures 16
and 17) is greatly diminished, as the case of the fluorine
substituent demonstrates, and the related radicals are
bent, regardless of the σ- or π-character of the substitu-
ent.
F igu r e 1. Inversion barrier vs electronegativity of the R-sub-
stituent as obtained from DFT calculations: (b) BLYP/6-31G-
(d,p)//BLYP/6-31G(d,p); (4) B3LYP/6-311G (2d,2p)//B3LYP/6-
311G(2d,2p).
The computational results for the structure of vinylic
radicals substituted with R-σ-substituents are in accord
with available spectroscopic evidence,7,10 and also with
results of recent ab initio calculations.18 The barriers
calculated by us are ca. 2-4 kcal/mol lower than those
from reference 18. However, whereas our values are
corrected for the differences in the zero point energies,
such correction was not mentioned in reference 18.
The results obtained by our calculations for the inver-
sion barrier of the bent radicals lend themselves to
experimental corroboration. For example, given that the
rate constant of H-atom abstraction by a vinyl radical
from Bu3SnH is ca. 108 M-1 s-1 at 30 °C,26 bent vinyl
radicals endowed with an inversion barrier greater than
6-7 kcal/mol should behave as configurationally stable
toward this reaction. The case of R-fluorovinyl radical
was investigated since, in view of its very high calculated
inversion barrier (19.5 kcal/mol), it was expected to give
retention of configuration in reaction with Bu3SnH. We
note that inversion barriers comparable in height, and
of the order of 25-27 kcal/mol, are indeed reported for
configurationally stable vinyl anions.3a,d
Exp er im en ta l Su p p or t. Retention of configuration
in the case of the R-MeO and R-Cl groups (our calculated
barriers 13.3 and 11.1 kcal/mol, Table 1) has indeed been
reported previously10 and attributed to a slower inversion
rate of the substituted vinyl radical, as controlled by the
electronegativity of the substituent, than the rate of the
“sampling” chemical reaction (hydrogen atom abstraction,
in general). A check of the computational prediction for
the inversion barrier of the R-fluorovinyl radical was
carried out with precursor PhCHdC(Br)F (19). It was
synthesized as an E,Z-mixture (ca. 80:20),27 and the two
diastereoisomers were separated by preparative GC.
Proton NMR showed that the more abundant geometric
isomer of 19 was the E, as judged from the J HF coupling
constant of 33 Hz, which has a typical value for a trans
F igu r e 2. Bond angle at the radical centre vs electronegativity
of the R-substituent as obtained from DFT calculations: (b)
BLYP/6-31G(d,p); (4) B3LYP/6-311G(2d,2p).
The â-substituted vinyl radicals 16 and 17, bearing a
σ-type and a π-type substituent, respectively, were found
to be bent. The minimum energy structures for radicals
11 and 15 are drawn in Figure 3.
Discu ssion
σ-Su bstitu ted Vin yl Ra d ica ls. The calculations
concerning the vinyl radicals 6-11 indicate that they
exist as two isomeric bent forms, which interconvert
through the linear structure. The bond angle values at
the radical center range from 130° to 141°, which are
values larger than expected from an sp2 hybridization
and suggest a structure “intermediate” between a bent
(26) Curran, D. P. Synthesis 1988, 417.
(27) Haar, V. R., J r. Thesis, University of Iowa (1973); University
Microfilms 74-16,693.

4076 J . Org. Chem., Vol. 62, No. 12, 1997
Galli et al.
Sch em e 3
Sch em e 5
Sch em e 4
precursor. Indeed, as shown in Scheme 5, the hydrode-
halogenation of either E- or Z-R-propylvinyl halides with
Bu3SnH/AIBN is reported as leading to stereoconver-
gence.30
Comparable results are reported for other chemical
trapping studies with R-alkylvinyl radicals.11,15 This
experimental behavior is in keeping with the bent fast-
equilibrating structure predicted by the calculations for
the R-alkylvinyl radical intermediate.
H-F relationship in an unsaturated system.28 Consis-
tently, (Z)-19 had a smaller J HF coupling constant (15
Hz). Reaction of pure (E)-19 with Bu3SnH/AIBN in
refluxing benzene gave a single hydrodebrominated
product (i.e., â-fluorostyrene), presenting GC and GC-
MS retention times significantly different (ca. 0.2 min,
on capillary columns) from that of the single hydrode-
brominated isomer resulting from the analogous reaction
with (Z)-19. Proton NMR of the crude reaction product
from (E)-19 showed the expected large (60 Hz) J HF
coupling constant due to the geminal H-F relationship28
and the small (9 Hz) J HH coupling constant due to the
cis H-H relationship. This spectral evidence is in
keeping with the formation of cis â-fluorostyrene from
(E)-19 (Scheme 3). Retention of configuration in the bent
R-fluorovinyl radical (Z)-20 and (E)-20 is therefore sub-
stantiated; it supports the result of the calculations, and
confirms previous experimental findings with R-halovinyl
radicals.10
One might wonder what would have been the possible
stereochemical outcome of a reaction similar to that
reported in Scheme 3 if the R-fluorovinyl radical inter-
mediate were instead linear. Our previous investigation
of the stereochemical behavior of the R-phenylvinyl
radical intermediate, which is calculated to be linear, in
reaction with Bu3SnH has demonstrated loss of the
configuration of the vinyl halide precursors (Scheme 4).1
Hence, either stereoconvergence or at least obtaining
a cis,trans mixture of â-fluorostyrenes would appear as
a likely outcome of the Bu3SnH/AIBN reaction of (E)-19
and (Z)-19 through a linear radical intermediate. Ob-
taining an unbalanced diastereomeric mixture is indeed
reported29 even for an R-phenylvinyl radical (i.e., a linear
radical) featuring a steric hindrance of the â-substituents
more similar to that in 20 than to that in 5. Thus, the
hypothesis of an R-fluorovinyl radical being linear may
be safely disregarded, on the basis of the different
stereochemical outcome that linear radicals do induce.
Support from the literature can also be found for other
cases of substituents investigated with our calculations.
For example, the R-methylvinyl radical intermediate,
with the very small calculated inversion barrier of 3.1
kcal/mol, is predicted to give rise to products that have
lost the original configuration of a suitable E- or Z-parent
π-Su bstitu ted Vin yl Ra d ica ls. The calculations of
vinyl radicals 12-15 indicate that they exist in the linear
forms, with substituent-to-double bond angles of 180°.
The equilibrium geometry found shows that the π-system
of the substituent is orthogonal to the π-system of the
double bond and is coplanar with the p orbital at the
radical center, thus suggesting a delocalization of the
unpaired electron by the adiacent π-system of the sub-
stituent. These findings deserve some comments with
respect to the experimental information already available
for π-substituted radicals, and also to the ab initio results
recently published.18
First, most experimental evidence points toward a
linear structure for these radicals,8,19 even though con-
trasting reports can also be found. For example, kinetic
results11 point toward a bent structure for radical 14 and
15, while spectroscopic results8 suggest a linear one. One
should concede that, when the inversion barrier of the
bent structure is only few kcal/mol, it may be difficult to
experimentally distinguish between the two possibilities.
This was our conclusion in the previous paper, where the
stereoconvergence in the SRN1(V) reaction (Scheme 2),
which is caused by the intermediacy of the R-phenylvinyl
radical 5, could be ascribed either to the linear radical
or to a fast equilibrating pair of bent radicals. Second,
among the three π-substituents examined earlier,18
namely, CtCH, CN, and COOH, only the cyano group is
in common with ours (structure 14). We calculate this
radical to be linear, whereas the UMP4(SDQ) results
indicate a bent structure with a barrier to inversion of
1.4 kcal/mol.18 Such a value, which is probably not
corrected for the zero point energy, is very small, thus
indicating a very flat surface. Moreover, as noted,18 such
calculation may be affected by spin contamination,
whereas our results should not be affected by such a
problem. Incidentally, preliminary results concerning
the structure of a vinyl radical R-substituted with a 2,4,6-
(28) (a) Becker, E. D. In High Resolution NMR; Academic Press:
New York, 1980; 2nd Ed.; Chapter 5. (b) Masure, D.; Sauvetre, R.;
Normant, J . F.; Villieras, J . Synthesis 1976, 761.
(29) Ogawa, A.; Yokoyama, K.; Yokoyama, H.; Obayashi, R.; Kambe,
N.; Sonoda, N. J . Chem. Soc., Chem. Commun. 1991, 1748.
(30) Curran, D. P.; J asperse, C. P.; Totleben, M. J . J . Org. Chem.
1991, 56, 7169.

Effect of Substituents on the Structure of the Vinyl Radical
J . Org. Chem., Vol. 62, No. 12, 1997 4077
trimethylphenyl group31 indicate that at a lower level of
theory (ROHF/3-21G) this radical is more likely to be
bent (bond angle: 144.1°), but it becomes linear, as our
R-phenylvinyl radical 15, when calculated by the DFT
approach.
We conclude that the DFT approach employed here
provides an overall picture which is consistent with the
experimental body of evidence concerning vinyl radicals
and as such can be viewed as a promising method for
dealing with organic structures.
(smaller bond angle, in Figure 2); as a consequence, the
higher will become the s character of the orbital occupied
by the odd electron. Conversely, radicals endowed with
higher inversion barriers (Einv > 30 kcal/mol), such as
the acyl42 YsC(•)dO and phosphonyl43 Y3P(•)dO, are
firmly bent, and therefore not significantly influenced by
the nature of the R-substituent(s) Y.
Exp er im en ta l Section
Ca lcu la tion s a n d In str u m en ta tion . Calculations were
conducted on a DEC Alpha AXP 3000/500. GC and GC-MS
analyses were performed on methyl silicone gum capillary
columns.1 Preparative GC separation of the E,Z-mixture of
Vin yl vs Cyclop r op yl a n d Oth er Ra d ica ls. The
bonds in a cyclopropane ring bear some resemblance to
π-bonds.32 It is therefore interesting to compare informa-
tion33 on the structures of the cyclopropyl anion, radical,
and cation, with that on their vinylic counterparts. The
cyclopropyl anion is a bent configurationally stable
species, while the cyclopropyl radical is a bent, rapidly
inverting species,33 with kinv of 108 s-1 at -175 °C and
1011 s-1 at 71 °C,34,35 due to a 3.9 kcal/mol inversion
barrier.36 Angular strain causes the cyclopropyl cation
to be best viewed as an open allyl structure,33 even though
a closed bent cyclopropyl cation has been substantiated
in the presence of some specific structural features.37
19 was executed on
a Carlo Erba Fractovap ATC/f gas
chromatograph. 1H NMR spectra were taken in CDCl3 at 300
MHz on a Bruker AC 300 instrument.
Syn th esis of â-Br om o-â-flu or ostyr en e (19). A dry 500
mL three-neck roundbottom flask with reflux condenser,
thermometer, septum, and a Tru-bore stirrer was flushed with
N2 prior to adding 26.0 g (99 mmol) of Ph3P and 150 mL of
freshly distilled anhydrous diglyme. The contents of the flask
were heated to 70 °C prior to the injection of 12.7 g (47 mmol)
of CFBr3. The solution turned from clear to yellow with a
precipitate forming. At 90 °C, 5.0 g (47 mmol) of PhCHO was
injected and the mixture heated for 3 h at 120 °C. The dark
solution was cooled and the product codistilled with water. The
lower layer of the distillate was washed with water to remove
diglyme, dried (MgSO4), and vacuum distilled, bp 80-82 °C
(at 11 Torr), to give 7.1 g (35 mmol; 75% yield) of a E,Z mixture
of 19 (>97% pure by gas chromatography).27 1H NMR δ 5.9
(J HF ) 33 Hz: E isomer; 88%), 6.6 (J HF ) 15 Hz: Z isomer;
12%). 19F NMR δ -68.3 (E), -65.7 (Z).28 The two diastereo-
mers of 19 were separated (> 99% pure) by means of prepara-
tive gas chromatography on a 4 m × 4 mm methyl silicone
fluid (550/SF) packed column at 170 °C.
Calculations (MP3/6-31G*//6-31G*) for an R-chloro-
cyclopropyl radical36 gives a barrier to cis,trans isomer-
ization (i.e., 7.9 kcal/mol) that is 3.2 kcal/mol smaller than
that for the corresponding R-chlorovinyl radical (i.e., 9).
More generally, the effect of R-substituents on the
inversion rate (kinv) of a cyclopropyl σ-radical had sug-
gested a decrease of kinv on increasing the electronega-
tivity of the substituent.38 This is in agreement with the
trend for substituted vinyl σ-radicals obtained from our
and Guerra’s calculations.18 An analogous correlation
between inversion barriers and electronegativity was
found for symmetrically R-substituted methyl (Y3C•)39 and
silyl radicals (Y3Si•),40 and it therefore appears to be a
common feature of radicals endowed with small-to-
medium inversion barriers. The reason for the correla-
tion between electronegativity of the substituent and
effects that it causes on the geometry of a radical center,
has been debated.6,33,41 Our preferred explanation is that
of Pauling,41b which suggests changes of the s character
of the orbital occupied by the odd electron in response to
a partially ionic character of the C-substituent bond(s).
The more electronegative the substituent X, the more
polar the C-X bond and the higher its p character
Hyd r od ebr om in a tion of 19 w ith Bu 3Sn H. A solution
of 50 mg of (E)-19 (0.25 mmol), 8 mg AIBN (0.05 mmol), and
330 µL Bu3SnH (1.24 mmol) in 0.6 mL benzene was irradiated1
at 350 nm for 3 h.14b Direct GC-MS analysis (no workup) of
the reaction mixture revealed formation of fluorostyrene
(retention time: 5.7 min). The reaction mixture was eluted
on a short hand-made chromatographic column (a Pasteur
pipette filled with silica gel) with the aid of a little C6D6 as
the eluent, in order to remove tin-containing compounds. A
1
fraction containing the reaction product was analyzed by H
NMR (no suppression of C6H6 was needed), and the vinylic
pattern (J HH 9 Hz), along with the J HF geminal coupling
constant (60 Hz, see text), allowed the assignment of the
structure of cis-â-fluorostyrene to this product. Similarly, (Z)-
19 gave trans-â-fluorostyrene (GC-MS retention time: 5.9
1
min), as judged by its H NMR (J HH 15 Hz). The contamina-
tion of either one of the â-fluorostyrene isomers in the other
isomer was lower than the detection limit of <0.1%.
Ack n ow led gm en t . We thank Dr. Nico Sanna
(CASPUR, Roma) for his kind assistance during the
calculations.
(31) Yamataka, H.; Rappoport, Z. Work in progress.
(32) (a) de Meijere, A. Angew. Chem., Int. Ed. Engl. 1979, 18, 809.
(b) Wiberg, K. B. Acc. Chem. Res. 1996, 29, 229.
Su p p or tin g In for m a tion Ava ila ble: Coordinates, total
energies, zero-point vibrational energies, and unscaled vibra-
tional frequencies of the stationary points obtained from BLYP/
6-31G(d,p)//BLYP/6-31G(d,p) calculations for radicals 6-17
and from B3LYP/6-311G(2d,2p)//B3LYP/6-311G(2d,2p) calcu-
lations for radicals 6, 9, and 11 (10 pages). This material is
contained in libraries on microfiche, immediately follows this
article in the microfilm version of the journal, and can be
ordered from the ACS; see any current masthead page for
ordering information.
(33) Boche, G.; Walborsky, H. M. In Cyclopropane derived radical
intermediates; updates from The Chemistry of Functional Groups;
Patai, S., Rappoport, Z., Eds.; Wiley: Chichester, 1990; Chapters 1
and 3.
(34) Fessenden, R. W.; Schuler, R. H. J . Chem. Phys. 1965, 43, 2704.
(35) J ohnston, L. J .; Lusztyk, J .; Wayner, P. P. M.; Abeywickrema,
A. N.; Beckwith, A. L. J .; Scaiano, J . C.; Ingold, K. U. J . Am. Chem.
Soc. 1985, 107, 4594.
(36) Apeloig, Y.; Nakash, M. J . Am. Chem. Soc. 1994, 116, 10781.
(37) Hart, H.; Martin, R. A. J . Am. Chem. Soc. 1960, 82, 6362.
(38) Altman, L. J .; Baldwin, R. C. Tetrahedron Lett. 1971, 2531.
(39) Guerra, M. J . Phys. Chem. 1995, 99, 81.
(40) Guerra, M. J . Am. Chem. Soc. 1993, 115, 11926.
(41) (a) Hoffmann, R.; Lipscomb, W. N. J . Chem. Phys. 1962, 36,
2179, 3498. (b) Pauling, L. Ibid. 1969, 51, 2767. (c) Wells, P. R. Prog.
Phys. Org. Chem. 1968, 6, 111. (d) Bingham, R. C.; Dewar, M. J . S. J .
Am. Chem. Soc. 1973, 95, 7180.
J O962373H
(42) Guerra, M. J . Chem. Soc., Perkin Trans. 2 1996, 779.
(43) Guerra, M. Chem. Phys. Lett. 1995, 237, 61.