Synthesis of C59(CHPh2)N from (C59N)2 and C59HN. The first derivatization of C59N

Full text of DOI:10.1021/ja964447c

2946
J. Am. Chem. Soc. 1997, 119, 2946-2947
the resulting azafullerenyl radical is expected to undergo free
radical reactions.
Synthesis of C59(CHPh2)N from (C59N)2 and C59HN.
The First Derivatization of C59N
Treatment of (C59N)2 with excess diphenylmethane in re-
fluxing o-dichlorobenzene (ODCB) for 48 h afforded, upon
chromatography (silica gel, toluene), a crude product. HPLC
analysis indicated one major constituent, consisting of 71% of
the isolable mixture. This was shown to be the substituted
azafullerene C59(CHPh2)N (2) by 1H and 13C NMR, FTIR, UV-
vis, mass spectrometry, and cyclic voltammetry. The hetero-
fullerene 2 was obtained in 42% isolated yield after purification
by HPLC on a Cosmosil Buckyprep semiprep column.8
Cheryl Bellavia-Lund, Rosario Gonza´lez,
Jan Cornelis Hummelen, Robin G. Hicks, Angela Sastre, and
Fred Wudl*
Institute for Polymers and Organic Solids
Departments of Chemistry and Materials
UniVersity of California
Santa Barbara, California 93106-5090
1
The H NMR spectrum of 2 showed signals corresponding
ReceiVed December 30, 1996
to two equivalent phenyl groups and a singlet corresponding to
the methine hydrogen at 6.09 ppm. The 34 lines observed
between 156 and 124 ppm (30 from the C59N moiety and 4
from the phenyl groups) in the 13C NMR spectrum are consistent
with the expected Cs symmetry of 2 depicted in eq 1. Another
peak at 65.2 ppm corresponds to the methine carbon atom and
a signal at 86.3 ppm must be due to the starred sp3 carbon atom
on the buckyball moiety. This carbon resonance is downfield
shifted by 14 ppm relative to the sp3 carbon of C59HN,6 as
expected on going from hydrogen to alkyl substitution. This
is in agreement with a “closed” structure in which the sp3 carbon
atom is bound to the nitrogen atom and the CHPh2 substituent,
as shown in 2 above.
While the NOE effect identifies most of the phenyl resonances
in the 13C NMR spectrum, proton-coupled 13C NMR distin-
guishes the resonance of the quaternary carbon of the phenyl
group (quartet at 139 ppm) from the surrounding fullerene
resonances. The coupled spectrum also supports the assignment
of the methine carbon, revealing a doublet at 65.2 (1JCH ) 127
Hz) and a â-coupled doublet at 86.7 (2JCH ) 5 Hz) for the sp3
carbon of the fullerene cage. Finally, positive ion electrospray
MS shows a molecular ion at m/z ) 889 and a strong peak at
m/z ) 722 corresponding to C59N+.
Since the incorporation of heteroatoms into the fullerene
skeleton is expected to modify its structural and electronic
properties,1 we prepared the first heterofullerene, “C59N”, which
was isolated as its dimer, (C59N)2.2 The isolation in the
condensed phase of this nitrogen-substituted fullerene allows
for the investigation of the chemical reactivity, physical proper-
ties,3,4 and possible applications of this new class of fullerenes.
As part of our continued interest in the azafullerenes, we have
initiated studies to prepare derivatives. Even though 15N-
labeling experiments5 and the isolation and full characterization
of C59HN have given us confidence that the starred carbons
(see eq 1, below) are sp3 hybridized, a principal reason for
derivatization of C59N was to glean information on the variability
of the chemical shift of the carbon atoms in question as a
function of substitution.
The UV-vis absorption of C59(Ph2CH)N is almost identical
to that of (C59N)2 and C59HN showing a broad band in the
visible at 448 nm; all three compounds are green in solution.
The similarities of the electronic spectra of the azafullerenes
supports the closed structure assigned to (C59N)2 depicted for
1, above.
Two possible processes for derivatization of C59N are (1)
deprotonation of the azafullerene C59HN6 followed by attack
of an electrophile and (2) homolytic dissociation of the
interdimer bond of (C59N)2 followed by free radical reactions.
Here, we report on the derivatization of azafullerene via
homolysis of its dimer.
6
The FTIR spectrum shows a typical pattern for a fullerene
derivative, in particular having absorptions similar to C59HN
in the fingerprint region (480-590 cm-1), where the strongest
peak is at 529.3 cm-1
.
According to theoretical calculations7 the interdimer bond of
(C59N)2 is relatively weak (18 kcal/mol) and should, under
photolysis or thermolysis conditions, undergo facile homolysis.
In the presence of a good hydrogen atom donor or radical source,
The diphenylmethane adduct of C59N exhibits two quasi-
reversible one-electron reduction waves; a third wave, which
is completely irreversible was also observed. This contrasts
the behavior of C59HN in which the reduction processes are
chemically irreversible. We interpreted the properties of the
latter as arising from facile hydrogen atom loss on reduction,
which is not possible in the diphenylmethane derivative. The
reduction potentials are presented in Table 1, along with those
of C60, (C59N)2, and C59HN for comparison. The potentials of
the (diphenylmethyl)azafullerene and hydroazafullerene are
comparable, and both are significantly more difficult to reduce
than the dimer, an observation consistent with the high elec-
tronegativity of a heterofullerene “substituent”. All of the
azafullerene compounds possess irreversible oxidation waves
at much lower potentials compared to those of C60.
(8) (Ph2CH)C59N: Diphenylmethane (2.4 mL) was added to a degassed
solution of (C59N)2 (30 mg) in HPLC grade 1,2-dichlorobenzene (25 mL).
The mixture was refluxed under argon for 48 h and chromatographed on
silica gel using toluene as eluent. The product was purified by HPLC using
a semipreparative Cosmosil Buckyprep column (30% hexane in toluene, 5
mL/min, UV/326 nm). Solvents were removed under vacuum, and the
product was washed with ether, centrifuged, and decanted three times to
remove ether-soluble components, including toluene. It was finally dried
under vacuum. Yield: 15 mg (42%). Please refer to Supporting Information
for spectroscopic data.
(1) Andreoni, W.; Gygi, F.; Parrinello, M. Chem. Phys. Lett. 1992, 190,
159.
(2) Hummelen, J. C.; Knight, B.; Pavlovich, J.; Gonza´lez, R.; Wudl, F.
Science 1995, 269, 1554. For C59N in the gas phase, see: Averdung, J.;
Luftmann, H.; Schlachter, I.; Mattay, J. Tetrahedron 1995, 51, 697.
Lamparth, I.; Nuber, B.; Schick, A.; Skiebe, T.; Grosser, T.; Hirsch, A.
Angew. Chem., Int. Ed. Engl. 1995, 35, 2257-2258. Another condensed
phase approach was recently disclosed: Nuber, B.; Hirsch, A. J. Chem.
Soc., Chem. Commun. 1996, 1421.
(3) Prassides, K.; Keshavarz-K, M.; Hummelen, J. C.; Andreoni, W.;
Giannozzi, P.; Beer, E.; Bellavia, C.; Cristofolini, L.; Gonzalez, R.; Lappas,
A.; Murata, Y.; Malecki, M.; Srdanov, V. I.; Wudl, F. Science 1996, 271,
1833.
(4) Brown, C. M.; Cristofolini, L.; Kordatos, K.; Prassides, K.; Bellavia,
C.; Gonza´lez, R.; Cheetham, A. K.; Zhang, J. P.; Andreoni, W.; Curioni,
A.; Fitch, A. N.; Pattison, P. Chem. Mater. 1996, 8, 2548-2550.
(5) Bellavia-Lund, C.; Keshavarz-K, M.; Gonza´lez, R.; Hummelen, J.-
C.; Hicks, R.; Wudl, F. Phosphorus, Sulfur Silicon Relat. Elem. 1997, in
press.
(6) Keshavarz-K., M.; Gonza´lez, M.; Hicks, R. G.; Srdanov, G.; Srdanov,
V. I.; Collins, T. A.; Hummelen, J. C.; Bellavia-Lund, C.; Pavlovich, J.;
Wudl, F. Nature 1996, 383, 147-150.
(7) Andreoni, W.; Curioni, A.; Holczer, K.; Prassides, K.; Keshavarz-
K, M.; Hummelen, J.-C.; Wudl, F. J. Am. Chem. Soc. 1996, 118, 11335-
11336.
S0002-7863(96)04447-2 CCC: $14.00 © 1997 American Chemical Society

Communications to the Editor
J. Am. Chem. Soc., Vol. 119, No. 12, 1997 2947
Table 1. Redox Potentials for C60 and some C59N Derivativesa
could be an adventitious initiator or C59N• from (C59N)2 (see
eq 5, below).
C60
(C59N)2
C59HN
C59N(CHPh2)
red
red
red
E1
E2
E3
-1123
-992
-1071
-1424
-1485
-1979
-2089
+886b
-1106b
-1082
-1450
-1913
-1500b
-1470
In a control experiment, treatment of C59HN6 with diphenyl-
methane in refluxing ODCB for 24 h, afforded, upon chroma-
tography (silica gel, toluene), one isolable band. HPLC analysis
of the crude product revealed the presence of 2 (73%), (C59N)2
(7%), and unreacted C59HN (5%). This experiment proves that
equation 4 is operative and that In• is due to a combination of
eqs 5 and 2.
Eox
+1300b,c
+1260d
+823b
+867b
a All potentials are in mV vs Ferrocene, in ODCB unless stated
otherwise. b Chemically irreversible. c Dubois, D.; Kadish, K. J. Am.
Chem. Soc. 1991, 113, 7773. d Chemically reversible, in 1,1,2,2-
tetrachloroethane solvent: Xie, Q.; Arias, F.; Echegoyen, L. J. Am.
Chem. Soc. 1993, 115, 9818.
We can attribute the mechanism of formation of 2 to be a
short free radical chain. As illustrated in eq 2, the initiation
step is proposed to be the homolysis of the azafullerene dimer
to C59N•
In support of the above dimerization, fresh, HPLC single
component samples of C59HN were found to dimerize on
standing at room temperature even in the solid state.
We have demonstrated above the relatively clean conversion
of (C59N)2 to a stable, soluble derivative in good yield. This
study shows that C59HN can also undergo homolytic cleavage
to yield the C59N radical, where all three species C59N•, C59-
HN, and (C59N)2 are involved in the process leading to 2. Full
characterization of the title molecule shows remarkable similari-
ties to (C59N)2 and C59HN, and in particular the UV-vis data
demonstrate that the electronic structure of all three molecules
is essentially identical. The 13C NMR spectra reveal a similar
pattern in the sp2 region of 2 to that of its precursors and an
unmistakable sp3 carbon at 86.1 ppm, illustrating the chemical
shift deviation as a result of substitution; an important stepping
stone for the absolute determination of the (C59N)2 structure.9
Research is currently in progress utilizing this methodology to
derive azafullerenes with unusual properties.
Subsequently, the steps in eqs 3a and 3b are plausible
propagation and termination steps. Hence, in the presence of
the hydrogen donor diphenylmethane, formation of C59HN was
expected to be one of the major products (eq 3a); however, it
was only 2% of the reaction mixture. The major product was
the diphenylmethane derivative of the azafullerene radical, 2,
(eq 3b).
Acknowledgment. We thank the National Science Foundation for
support through the Materials Research Laboratory and NSF DMR 95-
00888, as well as individual investigator grants. R.G. and A.S. thank
the Ministerio de Educacio´n y Cultura (Spain) for postdoctoral
fellowships. R.G.H. thanks the National Research Council of Canada
for a postdoctoral fellowship.
Two possible reasons for the low yield of C59HN are (1)
unlike 2, which appears to be quite stable under the “harsh”
reaction conditions, C59HN decomposes or polymerizes to yield
insoluble products or (2) C59HN undergoes a free radical
reaction to afford C59N• (eq 4). In eq 4, “In•” under the arrow
Supporting Information Available: Experimental details for the
preparation, purification, and characterization of 2 (6 pages). See any
current masthead page for ordering and Internet access instructions.
(9) Bellavia-Lund, C.; Collins, T.; Wudl, F. Submitted.
JA964447C