The[4 + 2] cycloaddition of anthracene with C60F18; anthracene goes ring walking

Article abstract of DOI:10.1039/b009441m

From the[4 + 2] cycloaddition of anthracene at the curved face of C₆₀F₁₈ we have isolated two main isomers of C_s and C₁ symmetry (isomers1 and 2, respectively) in ca. 3 : 2 relative yields; isomer 2 contained tr

Full text of DOI:10.1039/b009441m

View Article Online / Journal Homepage / Table of Contents for this issue
The [4 ϩ 2] cycloaddition of anthracene with C60F18; anthracene
goes ring walking
Anthony G. Avent,a Olga V. Boltalina,b Joan M. Street,c Roger Taylor*a and Xian-Wen Weia
2
a The Chemistry Laboratory, CPES School, Sussex University, Brighton, UK BN1 9QJ
b Chemistry Department, Moscow State University, Moscow 119899, Russia
c Chemistry Department, The University, Southampton, UK SO17 1BJ
Received (in Cambridge, UK) 24th November 2000, Accepted 6th April 2001
First published as an Advance Article on the web 2nd May 2001
From the [4 ϩ 2] cycloaddition of anthracene at the curved face of C60F18 we have isolated two main isomers of Cs
and C1 symmetry (isomers 1 and 2, respectively) in ca. 3 : 2 relative yields; isomer 2 contained traces of a third isomer,
also of Cs symmetry. Of the four mono adducts that can, in principle, be formed by this cycloaddition, the yields
of the three obtained reflect differences in steric hindrance between the fluorine and aromatic addends. The major
Cs isomer (1) is reasonably stable, but the more crowded C1 isomer (2) reverts readily to C60F18 and anthracene on
standing, and also undergoes rearrangement to isomer (1). This is the first example of spontaneous migration of a
cycloaddend from one 6 : 6-bond to another on the same fullerene cage surface. The UV/VIS spectrum of 1 is
significantly different from that for C60F18, suggesting that some electronic interaction may exist between the
addend and the cage. Anthraquinone is produced as a by-product of the reaction (and also from that between
either C60 or C60Cl6 and anthracene) showing the fullerene cage acts here as an oxidant.
The reaction of [60]fullerene with anthracene to give a 1 : 1-
Introduction
addition product has been reported a number of times,5–8 and
gives also a brown insoluble precipitate,7 which we confirm
(but it appears only upon removal of the reaction solvent). The
1 : 1-addition product is unstable, especially on heating, and
readily undergoes the retro Diels–Alder reaction to give the
original components. A bis adduct of D2h symmetry was also
isolated8 and this is produced quantitatively from the
mono-adduct on heating in the solid state;9 by contrast, five
bis-adducts are obtained by heating the mono-adduct in
solution.9 The 1 : 1-adduct between [60]fullerene and 9,10-
dimethylanthracene is also unstable at room temperature
(apparently more so than the anthracene adduct, due probably
to steric interactions between the methyls and the cage),
consequently this addend has been used as a removable
blocking and directing group.10 Lastly, the 1 : 1 adduct between
anthracene and C70Ph8 has been described.11
The fluorofullerene C60F18 is unique in possessing a flat aromatic
face, strong electron-withdrawal due to the presence of the
fluorines, and a substantial ‘curved fullerene’ region where
cycloadditions can occur.1,2 If suitable donor addends can be
attached, the molecule promises to be exceptionally useful
in applications involving electron-transfer properties, and we
plan to investigate the chemistry of C60F18 in this context. The
compound is available currently only in ca. tens of milligrams,
and so our initial experiments will be focused on determining
the kinds of reactions that are possible. Subsequently, as we are
able to produce larger quantities, promising derivatives will be
produced in amounts suitable for electrochemical and related
experiments, in order to determine the donor–acceptor
electronic interactions.
The use of C60Cl6 and C60Br6,8 in the formation of charge-
transfer complexes with organic donors (i.e. with no covalent
bonding between them) has been reported recently, it being
anticipated that the increased electron withdrawal by the halogen-
ated cage would increase the electron transfer.3 However, these
halogenofullerenes proved to be unsuitable for this purpose
because strong donors caused dehalogenation (the stability
following the order C60Cl6 > C60Br8 > C60Br6) whilst weak
donors showed no charge transfer. The greater strength of the
C–F bond causes fluorofullerenes to be much more resistant
to halogen loss than these other halogenofullerenes; it is
possible to obtain their EI mass spectra, something which is
impossible with chloro- or bromofullerenes. Fluorofullerenes
can be expected therefore to be more stable towards the form-
ation of donor–acceptor derivatives.
Reaction of anthracene with C60F18 can, in principle, give
four products from addition at the curved surface. The
bonds across which addition can occur are labelled a–d in
Fig. 1. Addition across either of bonds a and c will give a
product of Cs symmetry, whilst addition across bond b and d
will give a product of C1 symmetry. Steric hindrance between
the anthracene addend and fluorines will increase from a→d,
with the latter being very improbable because of the extent
of this hindrance. Force field calculations give the heats of
formation as 698.8, 710.1, 716.3 and 766.6 kcal molϪ1, respec-
tively, and whereas the differences a→b and b→c are 11.3 and
6.2 kcal molϪ1, respectively, the difference c→d is 50.3 kcal
molϪ1. The number of 1H NMR signals obtained for each
isomer will be: a, 4 × 2 (sp2 region), 1 × 2(sp3); b, 8 × 1(sp2),
2 × 1(sp3); c 4 × 2(sp2), 2 × 1(sp3); d, 8 × 1(sp2), 2 × 1(sp3). The
number of 19F NMR signals will be 10, 18, 10, 18, respectively,
with the readily distinguished upfield multiplets (due to the
fluorines attached to sp3 carbons themselves attached to sp3
carbons) comprising resonance combinations: 1 × 2 F ϩ 1 ×
1 F; 3 × 1 F; 1 × 2 F ϩ 1 × 1 F; 3 × 1 F.
The only derivative of C60F18 made so far, C60F15Ph3
(‘triumphene’), was produced by electrophilic substitution
of the C60F18 electrophile into benzene, catalysed by FeCl3 (this
can be described also as nucleophilic substitution of fluorine
by phenyl).4 Since no cycloadditions are known, we decided in
the first instance to carry out the reaction with anthracene.
994
J. Chem. Soc., Perkin Trans. 2, 2001, 994–997
This journal is © The Royal Society of Chemistry 2001
DOI: 10.1039/b009441m

View Article Online
Fig. 1 Schlegel diagram for C60F18 (᭹ = F), showing the four different
bonds (a–d) across which cycloaddition can occur.
Fig. 3 19F NMR spectrum of isomer 1.
Fig. 4 Location of the fluorines identified from the 2 D 19F NMR
spectrum of 1.
Fig. 2 The structure of isomer 1 (addition across bond a); ‘visible’
fluorines only are shown.
and 3.2 Hz, HCЈ), 7.45 (2 H, dd, J 5.4 and 3.2 Hz, HC), 5.74 (2 H,
s, HA). The 5.74 ppm peak showed 3.8% NOE couplings to
each of the 7.74 and 7.60 ppm peaks, confirming these latter to
be the BЈ and B hydrogens (Fig. 2) The pattern is consistent
only with the product from addition across the a bond (Fig. 1)
as shown in Fig. 2. Moreover, formation of this isomer in the
highest yield is compatible with it being the least hindered one.
During the acquisition of the spectrum, traces of the peaks for
anthracene appeared, showing that the adduct slowly reverts to
products (see also below).
Experimental
An excess of anthracene (20 mg, 0.112 mmol) was added to
C60F18 (6 mg, 0.006 mmol) dissolved in toluene (20 cm3, HPLC
grade), and the bright yellow mixture was heated under reflux
for 24 h under argon, during which time the colour changed
to orange. The crude product was filtered with exclusion of
atmospheric moisture (to avoid any nucleophilic replacement
of fluorine) from a small amount of fine red crystals which
were produced, and separated by HPLC (high pressure liquid
The 19F NMR spectrum (Fig. 3) showed ten lines at
δF Ϫ130.48 (1 F, d, J 20 Hz, 1), Ϫ134.29 (2 F, s, 2), Ϫ134.49 (2 F,
dd, J 4 and 16 Hz, 3), Ϫ135.72 (2 F, s, 4), Ϫ136.03 (2 F, s, 5),
Ϫ145.95 (2 F, d, J 30 Hz, 6), Ϫ147.20 (2 F, s, 7), Ϫ149.58 (2 F,
dt, J 5 and 30 Hz, 8), Ϫ157.31 (1 F, dt, J 10 and 20 Hz, 9),
Ϫ157.42 (2 F, dt, J 10 and 18 Hz, 10); some of the singlets
were broad and unresolved. The spectrum showed similarities
chromatography) using
a 10 mm × 250 mm Cosmosil
Buckyprep column, with toluene elution at a flow rate of 4.7 ml
minϪ1. Fractions were obtained with retention times of 41 min
(recovered C60F18 ca. 1 mg), 13 min (isomer 1), 9 min (isomer 2),
combined yields of these two isomers ca. 3 mg, 2.4 min
(recovered anthracene) and 3 min (anthraquinone); the 1 : 2
isomer ratio was approximately 3 : 2. EI mass spectra were
12
to that of Cs C60F17CF3 in having two almost superimposed
1
obtained at 70 eV, H NMR were run at 500 MHz, and 19F
upfield peaks (in a 2 : 1 intensity ratio) due to the unique
fluorines attached to carbons themselves surrounded by three
sp3 carbons.
NMR spectra were run at 376.5 MHz.
We were unable to identify the red crystals (which were very
insoluble in toluene, chloroform etc.), but they appeared to be a
derivative of C60F18 because this was regenerated after a toluene
suspension was allowed to stand for 2–3 weeks.
The bold numbers above identify the fluorines in Fig. 4, and
were deduced from a 2 D spectrum. As in the case of C60F18
the fluorines nearest to the central aromatic ring (nos. 3, 4 and
5) are more downfield than those further away (nos. 6, 7 and 8).
Relative to the corresponding fluorines in C60F18, peaks 2 and
6–8 are more upfield, the rest are more downfield. Analyses
such as these will be valuable in analysing the structures of
other derivatives of C60F18, which we hope to produce in due
course.
Traces of C60F18 arising from the retro cycloaddition are
evident, the two main peaks for this occurring at Ϫ143.43 and
Ϫ136.03 ppm [coincident with peak no. 5, but evident on the
2 D spectrum (not shown)]. The integrated intensities show the
amount to be 5.5%, which represents the extent of the retro
reaction after ca. 2 weeks at room temperature. This derivative
therefore appears to have similar stability to that of the C60:
anthracene complex.
Results and discussion
Isomer 1
The EI mass spectrum (70 eV) showed peaks only for C60F18
(1062 amu with some normal fragmentation of C60 at 720 amu)
and anthracene (178 amu). Since both of these components
were absent from the HPLC-purified sample, it is clearly an
adduct (which can only be a 1 : 1 adduct because of steric
considerations).
1
The structure was confirmed by the H NMR (500 MHz
CDCl3), which gave peaks at δH 7.74 (2 H, dd, J 5.4 and 3.2 Hz,
HBЈ), 7.60 (2 H, dd, J 5.4 and 3.2 Hz, HB), 7.47 (2 H, dd, J 5.4
J. Chem. Soc., Perkin Trans. 2, 2001, 994–997
995

View Article Online
Fig. 6 19F NMR spectrum of isomer 2. Peaks marked ᭹ are due to
isomer 1 produced by isomerisation, and peaks marked × are due to the
second Cs isomer arising from addition across bond c.
Fig. 5 The structure of isomer 2 (addition across bond b); ‘visible’
fluorines only are shown.
Isomer 2
As for isomer 1 the EI mass spectrum consisted only of the
anthracene and C60F18 fragmentation ions. The 1H NMR
spectrum (500 MHz, CDCl3) showed 2 to be unsymmetrical,
giving ten lines overall at δH 7.65 (1 H, d, J 7.4 Hz, HC), 7.60
(1 H, d, J 7.5 Hz, HD), 7.57 (1 H, d, J 7.4 Hz, HE), 7.47 (1 H, d,
J 7.1 Hz, HF), 7.442 (1 H, dd, J 7.4 and 1.3 Hz, HG), 7.405 (1 H,
dd, J 7.6 and 1.3 Hz, HH), 7.392 (1 H, dd, J 7.5 and 1.3 Hz, HI),
7.365 (1 H, dd, J 7.4 and 1.3 Hz, HJ), 5.623 (1 H, s, HA), 5.495
(1 H, s, HB). The relationships between HC-F and HG-J, Fig. 5)
were determined by a 2D-COSY experiment. Couplings (NOE)
between HA–HD, HA–HE, HB–HC, and HB–HF were determined
as 8.0, 7.3, 9.3 and 7.5%, respectively. Some other double
doublets, of about half the above intensities were evident in the
7.34–7.45 ppm region, and these are attributed to a trace of a
third isomer of Cs symmetry (produced by addition across
bond c) as shown also by the 19 F NMR spectrum, Fig. 6.
Isomer 2 is less stable than isomer 1, significant amounts of
anthracene being evident from the 1H NMR spectrum, showing
that decomposition was occurring fairly rapidly. In principle,
isomer 2 could involve addition across bond d. However, given
the extreme steric hindrance that models indicate for the latter,
this possibility can be ruled out.
Fig. 7 Location of the fluorines identified from the 2 D 19F NMR
spectrum of 2.
solved at this time; separation of anthracene from a 1 : 1 adduct
with [60]fullerene and recombination with a second cage takes
place in the solid state.9 Interconversions from a 6,6-closed-
to a 6,5-open structure (methanofullerene to homofullerene)
have been reported previously,14 but these are mechanistically
different since they involve simultaneous 1,3-shifts of two C–C
bonds.
The eighteen lines for isomer 2 appeared at δF -127.25 (1 F, d,
J 20 Hz, 1), Ϫ130.46 (1 F, d, J 20 Hz, 2), Ϫ134.48 (1 F, sh 3),
Ϫ135.20 (1 F, s, 4), Ϫ136.25 (1 F, s, 5), Ϫ136.65 (1 F, sh 6),
Ϫ137.43 (1 F, d, J 20 Hz, 7), Ϫ137.57 (1 F, s, 8), Ϫ139.24 (1 F,
ddd, J 6, 6 and 33 Hz, 9), Ϫ141.14 (1 F, dd, J 9 and 26 Hz, 10),
Ϫ142.68 (1 F, ddd, J 6, 6 and 26 Hz, 11), Ϫ143.53 (1 F, d, J 28
Hz, 12), Ϫ143.87 (1 F, d, J 28 Hz, 13), Ϫ144.40 (1 F, ddd, J 5, 5
and 26 Hz, 14), Ϫ149.85 (1 F, ddd, J 7, 7 and 28 Hz, 15),
Ϫ157.19 (1F, dt, J 11 and 21 Hz, 16), Ϫ157.49 (1 F, dt, J 10 and
19 Hz, 17), Ϫ158.25 (1 F, dt, J 10 and 20 Hz, 18). The bold
numbers identify the fluorines in Fig. 7 as deduced from a 2 D
spectrum. The small amount of compound prevented identi-
fication of some of the couplings, hence some assignments
may be interchanged; the couplings 17→2, 18→7, 13→11→9,
4→12→15, 16→1→6→10→14, and 16→8 were clearly identi-
fied. As for isomer 1 (and C60F18 derivatives in general) the
fluorines nearest to the central benzenoid ring (nos. 3, 4, 5, 6,
8, 9) are more downfield than those further out (nos. 10, 11,
12, 13, 14, 15). The peak 7 (and hence peak 18) was assigned on
the basis of its marked upfield shift assumed to arise from the
proximity of the anthracene addend.
The 19F NMR spectrum, run ca. 2 weeks after formation of
the adduct, is shown in Fig. 6, all of the associations between
peaks from the same component being determined from a 2 D
spectrum. There are four main peaks arising from C60F18
produced by the retro Diels–Alder reaction, eighteen peaks due
to isomer 2, and (marked ×) nine of the ten peaks due to the
presence of the alternative Cs isomer produced by addition
across bond c (Fig. 1); their intensities relative to those for iso-
mer 2 appear greater because of the higher symmetry.
Most significant, however, are the peaks marked ᭹ due to
isomer 1 produced from isomerisation of isomer 2 (and possibly
also some of the second symmetrical isomer). We are confident
that isomer 1 was not present originally in isomer 2 because the
former eluted after the latter in the HPLC separation and so
would not be present due to ‘tailing’ and, moreover, there was
no trace of it in the 1H NMR spectrum that was run ca. 10 days
before the 19F NMR spectrum. This appears to be the first
example of spontaneous migration of a cycloaddend across the
same cage surface. Such a migration can only be detected in the
presence of a second addend (in our case the fluorine atoms),
and has been observed previously only during electrochemistry
of bis malonates (shown to be intramolecular through the use
of a standard cross-over procedure).13 In view of the lability of
the anthracene it is possible that separation–recombination
occurs in our case, but this mechanistic aspect cannot be readily
The presence of the components in the 19F NMR spectrum
of 2 is readily seen in the region where the resonances appear
for the fluorines attached to sp3-hybridised carbons, themselves
attached to sp3-hybridised carbons (Fig. 8). This shows the
intense off-scale peak due to regenerated C60F18, the three equal
intensity double triplets for peaks 16–18 of isomer 2 (marked
A), the two double triplets for peaks 9 and 10 of isomer 1
(marked B – one of these is partially superimposed on one of
996
J. Chem. Soc., Perkin Trans. 2, 2001, 994–997

View Article Online
conjugation needs to be considered. Although there is a major
difference between the spectra for the two derivatives there is a
precedent for this.15
Oxidation to anthraquinone
As noted in the experimental, anthraquinone was isolated
from the reaction mixture. This was identified by two equal
intensity multiplets at δH 8.30 and 7.83 (lit.16 8.21–8.38 and
7.67–7.83). This same product was also obtained from the
reaction between both [60]fullerene and anthracene, and C60Cl6
and anthracene. The fullerene cage is evidently acting as an
oxidant. This seems not to have been reported before in the
reaction with anthracene, but previously we reported that [60]-
fullerene will readily oxidise H2S to sulfur.17
Further reactions of C60F18 are planned.
Fig. 8 Expanded Ϫ(157–159) ppm region of the 19F NMR spectrum
of 2.
Acknowledgements
We thank INTAS (grant no. 97-30027), NATO, and the Royal
Society (Joint Project Grant) for financial support (to O. V. B.
and R. T.), and the Royal Society for a Sino-British Trust
Award Fellowship (to X.-W. W.).
References
1 O. V. Boltalina, V. Yu. Markov, R. Taylor and M. P. Waugh, Chem.
Commun., 1996, 2549.
2 I. S. Neretin, K. A. Lyssenko, M. Yu. Antipin, Yu. L. Slovokhotov,
O. V. Boltalina, P. A. Troshin, A. Yu. Lukonin, L. N. Sidorov and
R. Taylor, Angew. Chem., Int. Ed., 2000, 39, 3273.
3 Y. Yoshida, A. Otasuka, O. V. Drozdova and G. Saito, J. Am. Chem.
Soc., 2000, 122, 7244.
4 O. V. Boltalina, J. M. Street and R. Taylor, Chem. Commun., 1998,
1827.
5 K. Komatsu, Y. Murata, A. Miyabo, K. Takeuchi and T. S. M. Wan,
Fullerene Sci. Technol., 1993, 1, 231; K. Komatsu, Y. Murata,
N. Sugita, K. Takeuchi and T. S. M. Wan, Tetrahedron Lett., 1993,
34, 8473.
6 J. A. Schleuter, J. H. Seaman, S. Taha, H. Cohen, K. R. Lykke,
H. H. Wang and J. M. Williams, J. Chem. Soc., Chem. Commun.,
1993, 972.
7 F. Wudl, Acc. Chem. Res., 1992, 25, 157.
8 M. Tsuda, T. Akayuki, T. Nogami, S. Kurono and M. Ohashi,
J. Chem. Soc., Chem. Commun., 1993, 1296.
9 B. Kräutler, T. Müller, J. Maynollo, K. Gruber, C. Kratky,
P. Ochsenbein, D. Schwarzenbach and H.-B. Bürgi, Angew. Chem.,
Int. Ed. Engl., 1996, 35, 1204.
10 I. Lamparth, C. Maichle-Mössmer and A. Hirsch, Angew. Chem.,
Int. Ed. Engl., 1995, 34, 1607.
Fig. 9 UV/VIS spectrum (dichloromethane) for 1.
the bands for isomer 2) and two other double triplets for the
other symmetrical isomer (marked C).
11 A. G. Avent, P. R. Birkett, A. D. Darwish, H. W. Kroto, R. Taylor
and D. R. M. Walton, Fullerene Sci. Technol., 1997, 5, 643.
12 O. V. Boltalina, P. B. Hitchcock, P. A. Troshin, J. M. Street and
R. Taylor, J. Chem. Soc., Perkin Trans. 2, 2000, 2410.
13 R. Kessinger, M. Gómez-López, C. Boudon, J.-P. Gisselbrecht,
M. Gross, L. Echegoyen and F. Diederich, J. Am. Chem. Soc., 1998,
120, 8545.
14 T. Suzuki, Q. Li, K. C. Khemani and F. Wudl., J. Am. Chem. Soc.,
1992, 114, 7301; F. Diederich, L. Isaacs and D. Phelp, J. Chem. Soc.,
Perkin Trans. 2, 1994, 391; A. B. Smith, R. M. Strongin. L. Brard,
G. T. Furst, W. J. Romanow, K. G. Owens, R. J. Goldshmidt and
R. C. King, J. Am. Chem. Soc., 1995, 117, 5492.
UV spectra (dichloromethane)
The spectrum of C60F18 shows two bands at 253.3 nm (main)
and 342.4 nm (this latter has a small shoulder at ca. 362 nm).
The spectrum for isomer 2 is fairly similar, the main band
appearing at 255.2 nm, the minor band at 341.7 nm with
a second shoulder at ca. 378 nm. [This similarity could in
principle be attributed to the rapid reversion to C60F18 and
anthracene, though none of the sharp anthracene bands (326.2,
341.9, 359.3, 378.7 nm) could be seen.] By contrast, the
spectrum for 1 (Fig. 9) shows substantial differences, with two
major bands at 232.9 and 265.1 (main) nm, and many minor
bands at 342.7, 399.6, 428.7, 484.0 and 518.4 nm. This suggests
significant electronic interaction between the addend and the
cage, though the effect of interruption of the cage surface
15 Y. Nakamura, M. Taki, S. Tobita, H. Shizuka, H. Yokoi,
K. Ishiguro, Y. Sawaki and J. Nishimura, J. Chem. Soc., Perkin
Trans. 2, 1999, 127.
16 Aldrich Library of 1H NMR spectra, Aldrich Chemical Co., 1985.
17 A. D. Darwish, H. W. Kroto, R. Taylor and D. R. M. Walton,
Fullerene Sci. Technol., 1993, 1, 571.
J. Chem. Soc., Perkin Trans. 2, 2001, 994–997
997