Chemistry of 1-fluoro-2,3,4-triphenylcyclobutadiene dimers

Article abstract of DOI:10.1021/jo701900k

(Chemical Equation Presented) The reaction of 2,4-dichloro-1,1-difluoro-3- phenyl-2-cyclobutene 1 with excess phenyllithium and subsequent transformations of the products have been reinvestigated. The phenyllithium reaction appears to proceed through the

Full text of DOI:10.1021/jo701900k

Chemistry of 1-Fluoro-2,3,4-triphenylcyclobutadiene Dimers
Angshuman R. Choudhury,† Deepak Chopra,† Tayur N. Guru Row,*,†
Kuppuswamy Nagarajan,‡ and John D. Roberts*,§
Solid State & Structural Chemistry Unit, Indian Institute of Science, Bangalore 560012, India, Hikal R&D
Center, Bannerghatta Road, Bangalore 560076, India, and Crellin Laboratory, DiVision of Chemistry &
Chemical Engineering, California Institute of Technology, Pasadena, California 91125
ssctng@sscu.iisc.ernet.in; robertsj@caltech.edu
ReceiVed August 29, 2007
The reaction of 2,4-dichloro-1,1-difluoro-3-phenyl-2-cyclobutene 1 with excess phenyllithium and
subsequent transformations of the products have been reinvestigated. The phenyllithium reaction appears
to proceed through the intermediacy of a fluorotriphenylcyclobutadiene 2 to produce a well-characterized
dimeric trans-hexaphenyldifluorotricyclooctadiene 3a. Subsequent transformations of 3a gave a pen-
taphenyldihydrodifluoropentalene 4, which on acid hydrolysis formed a pentaphenyldihydropentalenone
5. When 3a was photolyzed in benzene, after purification, it afforded 6, an isomer of 5, probably by way
of 7, an isomer of 4. Thermolysis of 3a also provided, in low yield, a substance believed to be a
pentaphenylfluorophenanthrene 8. Along with isolation of 3a, and probably arising from a different isomer
of the 3 family, was a pentaphenylfluorophenanthrene 9, which was suspected of being an isomer of 8.
Single-crystal X-ray studies were used to derive structures for 4, 5, 6, and 9. Formation of the unusual
and intriguing transformation products has at least been rationalized.
Introduction
to form dimers with double bonds substituted exclusively with
phenyl groups, there are four possible different dimers, provided
we consider both the “sofa” 10a and “tub” 10b forms.
In the course of investigations 45 years ago of the reaction
of phenyllithium with 2,4-dichloro-1,1-difluoro-3-phenyl-2-
cyclobutene 1, it was discovered that excess phenyllithium gave
a mixture of products and, through at least partial intermediacy
of fluorotriphenylcyclobutadiene 2, formed a dimer 3a, C44H30F2,
mp 200 °C (d) in about 17% yield.1-3 Dimerizations of
cyclobutadienes to 2,3,7,6-tricyclo[01,4.05,8]octadienes are well-
known reactions, but in the case of 2, there are a rather large
number of possible structures. However, omitting enantiomers,
assuming cis addition and also that cycloaddition occurs only
Originally,3 without today’s more powerful methods of
structural characterization, it was necessary to proceed in a
stepwise manner. Thus, UV and Raman spectral data indicated
only tetrasubstituted cis-stilbene-like double bonds. Proton NMR
showed only phenyl proton chemical shifts, while fluorine NMR
gave a single shift for the two fluorine atoms. These data allowed
us to conclude that the dimer had the tricyclooctadiene
framework with six phenyl groups and two fluorine atoms.
Dipole moment determinations further narrowed the four
structural possibilities (two sofa and two tub) uniquely to 3a,3
† Indian Institute of Science.
‡ Hikal R&D Center.
§ California Institute of Technology.
(1) Kitahara, Y.; Caserio, M. C.; Scardiglia, F.; Roberts, J. D. J. Am.
Chem. Soc. 1960, 82, 3106-3111.
(2) Nagarajan, K.; Caserio, M. C.; Roberts, J. D. ReV. Chimie 1962, 7,
1109-1117.
(3) Nagarajan, K.; Caserio, M. C.; Roberts, J. D. J. Am. Chem. Soc. 1964,
86, 449-453.
10.1021/jo701900k CCC: $37.00 © 2007 American Chemical Society
Published on Web 11/03/2007
9732
J. Org. Chem. 2007, 72, 9732-9735

1-Fluoro-2,3,4-triphenylcyclobutadiene Dimers
SCHEME 1. Reaction Synopsis
SCHEME 2. Possible Isomerization Processes of 3a
which was confirmed by a difficult (for that period) single-
crystal X-ray study.4 Efforts to prove the structure of 3a by
chemical degradation reactions were unsuccessful, but in the
process of trying, several crystalline compounds were obtained,
of which only two could even be assigned tentative structures,
in our original work. Scheme 1 is a roadmap of the various
reactions and the general character of their products. The
continued availability of the samples and current interest in F,
F interactions in crystals of fluoroaromatics,5 prompted further
investigation of these substances.
Results and Discussion
Our current studies of 3a have confirmed the molecular
formula, C44H30F2, by its mass spectrum. The proton NMR
spectrum in CDCl3 at 300 MHz (vs 40 MHz in 1960) gave only
signals in the aromatic region. Strikingly, 3a showed two
approximate doublets at 6.75 (4H) and 7.64 ppm (4H), which
were not coupled to each other. The other 22 protons were seen
as indistinct multiplets between 6.9 and 7.4 ppm. The doublets
at 6.75 and 7.64 ppm were unexpected and could well involve
the ortho protons of the phenyl groups adjacent to the fluorine
6 (see Scheme 2). Despite the plausibility of the mechanism,
the product was not 6 but was shown quite surprisingly by
single-crystal X-ray diffraction to be 5 (see Figure 1). To make
matters more interesting, 5 cannot be formed by a route similar
to Scheme 2, either from 3a or 3a’s very reasonably formed
isomer 3b, by as simple a mechanism as that proposed for 3a
to 6. However, a somewhat more complicated mechanism via
3c and 4 leading to 5 is shown in Scheme 3, which requires
that 3c be formed by heating 3a.
atoms in 3a. The 13C NMR spectrum of 3a at 75 MHz showed
13
quaternary carbons at 99.2 ppm with
J
values of 274 Hz
1CF
and 13J2CF of 28 Hz. The saturated quaternary carbons carrying
phenyl groups were at 69.38 ppm as triplets arising from
coupling with the nearby F atoms.
When heated above its melting point, 210-215 °C, for 75 s,
3a gave a red melt, which when triturated with hexane afforded
an 85% yield of yellow crystalline 4. Traces of an isomer of 4,
thought to be 7, were detected in this reaction. Compound 4,
C44H30F2, was isomeric with 3a, but with two magnetically
nonequivalent geminal fluorine atoms (1JFF ) 250 Hz) in its
NMR spectrum. On hydrolysis with sulfuric acid, 4 afforded
yellow 5, C44H30O, with a ketonic CdO band in the IR at 1710
cm-1. The ketone was apparently hindered at the carbonyl group
because it did not form a 2,4-dinitrophenylhydrazone derivative.
A possible mechanism for conversion of 3a to a gem-difluoro
compound would involve ionization of one of the equivalent
fluorines, a conventional cationic cyclobutyl-to-cyclopropylm-
ethyl rearrangement, followed by fluoride ion return to afford
a gem-difluoride, which after hydrolysis would give the ketone
(4) Fritchie, C. J.; Hughes, E. W. J. Am. Chem. Soc. 1962, 84, 2257-
2258.
(5) Choudhury, A. R.; Nagarajan, K; Guru Row, T. N. Cryst. Eng. 2004,
6, 43.
FIGURE 1. ORTEP representation of 5.
J. Org. Chem, Vol. 72, No. 25, 2007 9733

Choudhury et al.
SCHEME 3. Possible Formation of 5 from 3c
SCHEME 4. Possible Route for Formation of 3c from 3a
FIGURE 2. ORTEP representation of 4.
SCHEME 5. Possible Isomerization Route from 3b to
Phenanthrene Derivative 9
FIGURE 3. ORTEP representation of ketone 6.
nylcyclooctatetraene or octaphenylcyclo[01,4]octadiene.6 That 4
is formed preferentially by heating 3a is likely to be the result
of strong steric interactions between the phenyl groups expected
for 7 compared to 4, as is evident in Figure 2 and seen quite
dramatically by comparing the crystal structures of 5 (Figure
1) and 6 (Figure 3).
Oddly, photolysis of 3a in benzene using a medium-pressure
mercury lamp, followed by chromatography on alumina, af-
forded 15% of a yellow ketone of mp 199-199 °C, which by
single-crystal X-ray diffraction turned out to be 6 (Figure 3),
which may have been formed from the corresponding difluoride
7 by hydrolysis catalyzed by alumina in the chromatographic
purification.
FIGURE 4. ORTEP representation of 9.
One reasonable way of having that happen would be to have
3a open to the cyclooctatetraene 11 and then isomerize to 3c
by way of the corresponding cubane 12, or else 3a going directly
to 12 and thence to 3c, both as shown in Scheme 4. Octaphe-
nylcubane has been reported via the intermediacy of octaphe-
In the reaction of phenyllithium with 1, a small yield (3.5%)
of a second compound 9 was consistently obtained as white
(6) Freedman, H. H.; Petersen, D. R. J. Am. Chem. Soc. 1962, 84, 2837-
2838.
9734 J. Org. Chem., Vol. 72, No. 25, 2007

1-Fluoro-2,3,4-triphenylcyclobutadiene Dimers
SCHEME 6. Possible Route from 3a to 8
bond reorganizations to form a phenanthrene derivative. This
cyclization has a parallel in isomerization of a cyclohexenyl-
cyclobutadiene to a tetrahydronaphthalene.7
When 3a was heated in boiling decalin for 2 min, hydrogen
fluoride was evolved and a 20% yield of a white substance 8
of mp 337-338 °C was obtained, along with larger amounts of
4 and 5. The solid 8 was C44H29F by mass spectrometry and
thus an isomer of 9. The resemblance was heightened by its
UV spectrum in cyclohexane with absorption maxima at 210
and 276 nm with shoulders at 228 and 324 nm. Unfortunately,
useful proton and carbon NMR spectra or suitable crystals for
single crystals could not be obtained for single-crystal diffrac-
tions studies, although XRD indicated crystallinity. It is possible
that the product is a mixture of isomers, but a principle
component could be structure 8, which would arise in a
straightforward way from 3a by Scheme 6, analogous to Scheme
5.
Acknowledgment. Dedicated to Professor S. Swaminathan
on the occasion of his 81st birthday. T.N.G. acknowledges
funding (DST 0361) from the Department of Science and
Technology, Government of India. At the California Institute
of Technology, the research was supported by the National
Science Foundation through grants funded in 1956-60 and
currently under Grant No. CHE-0104273.
fluffy needles, mp 273 °C. This compound had the molecular
formula C44H29F, as shown by elemental analysis and molecular
weight determined by isothermal distillation. Its 19F NMR
spectrum had a single broad peak, while its 13C spectrum showed
a quaternary carbon coupled to fluorine with a 13JCF of 244 Hz.
The UV spectrum in cyclohexane had maxima at 210, 234, an
280 nm with a shoulder at 326 nm. These data were suggestive
of an aromatic hydrocarbon, and this was confirmed by single-
crystal X-ray diffraction to be 9 (see Scheme 5 and Figure 4)
assumed to be formed through the intermediacy of a different
dimer of 1, designated as 3b, which could lead to 9, first by an
ortho-electrophilic cyclization on a phenyl group, followed by
Supporting Information Available: Details of the X-ray
diffraction studies, packing diagrams, and CIFs. This material is
available free of charge via the Internet at http://pubs.acs.org.
JO701900K
(7) Sharts, C. M.; Roberts, J. D. J. Am. Chem. Soc. 1961, 83, 871-878.
J. Org. Chem, Vol. 72, No. 25, 2007 9735