The preparation and structures of non-hydrocarbon functionalised fullerene-diamine adducts

Article abstract of DOI:10.1039/b303952h

Addends based on C2-substituted piperazine are employed to generate C₆₀ monoadducts by photochemical addition, representing the first introduction of non-hydrocarbon addend functionality to fullerenes via the oxidative dehydrogenation reaction of diamines with fullerenes.

Full text of DOI:10.1039/b303952h

The preparation and structures of non-hydrocarbon functionalised
fullerene–diamine adducts†
Craig P. Butts* and Mikael Jazdzyk
School of Chemistry, University of Exeter, Exeter, UK EX4 4QD. E-mail: C.P.Butts@Exeter.ac.uk; Fax: +44
1392 263434; Tel: +44 1392 263455
Received (in Cambridge, UK) 10th April 2003, Accepted 13th May 2003
First published as an Advance Article on the web 29th May 2003
Addends based on C2-substituted piperazine are employed
to generate C60 monoadducts by photochemical addition,
representing the first introduction of non-hydrocarbon
addend functionality to fullerenes via the oxidative dehydro-
genation reaction of diamines with fullerenes.
piperazines were found to be slower than that of unsubstituted
piperazine 1h (entry 8, Table 1).
Reactions of piperazine derivatives with alkyl, hydroxyl,
ether and amide functionalities (1a–e, entries 1–5, Table 1) all
proceed readily and no bis-addition products were isolated. The
successful reaction of the amide derivatives 1d and 1e was
surprising given that attempted photochemical reactions of C60
with the ester 1g (entry 7, Table 1) gave no isolable products
despite observation of the expected colour change. Significantly
less than quantitative amounts of C60 were recovered from the
latter reaction mixture, suggesting that the observed colour
change is the result of some reaction of decomposition products
with the fullerene. Irradiation of the ester under reaction
conditions, but in the absence of C60, gave no noticeable
decomposition. Presumably the ester is unstable to the photo-
chemical conditions only in the presence of the strongly
electron accepting fullerene. As these reactions are proposed to
proceed via an electron transfer step, forming radical anion–
radical cation pairs, it seems likely that it is the radical cation of
the piperazine ester which is decomposing. Similarly, an
attempted reaction of a silyl ether derivative of 1b also gave no
isolable products, despite an observed colour change during
photolysis.
The numerous chemical reactions of fullerenes1 have been well
documented over the last decade. The Bingel–Hirsch addition–
elimination of halomalonates, Prato’s dipolar cycloadditions of
azomethine ylides and Diels–Alder reactions of quinodime-
thano-derivatives are some of the best examples of relatively
general, substrate tolerant methods for addition to C60, with
typical yields of up to 45%.2
The oxidative dehydrogenation reactions of C60 with primary
and secondary amines3 and diamines4 were established early in
the story of fullerene chemistry. The reactions of primary and
secondary amines with C60 are very difficult to control and give
different products depending on substrate and conditions. These
products include 1,2- and 1,4- monoadduct regioisomers,1c
fullerene dimers and fullerene oxides,5 in addition to the usual
array of multiple-addition products. Indeed, when considering
the monoadducts of amines or diamines, only the reaction of
piperazine with C60 gives a synthetically useful yield of a single
monoadduct (ca. 50% in the best case reported4). To the best of
our knowledge, every other amine or diamine reagent and
conditions give 520% yield of a fullerene monoadduct.4,6 More
significantly, no diamine reagent reported to date has contained
non-hydrocarbon functionality. Consequently the oxidative
dehydrogenation addition of secondary diamines (and amines)
to C60 has not been exploited to any significant extent as a
general method for the introduction of functionality to fullerene
derivatives.
The structures of adducts 2a–f were confirmed by 1H and 13
C
NMR spectroscopy, as well as MALDI-TOF mass spectrome-
try.
1
Recently, in a paper describing the assignment of the H
NMR spectrum of the piperazine–C60 adduct, we reported the
7
photochemical reaction of 2-methylpiperazine with C60. This
reaction gave the corresponding monoadduct in 41% yield,
which compares favourably with the yield achieved by Kampe
and others for the piperazine–C60 monoadduct.4 We also
reported an improved preparation for this latter adduct,
obtaining a 72% yield by photochemical reaction of stoichio-
metric quantities of fullerene and diamine. The following is a
report describing preparations of non-hydrocarbon function-
alised C60–piperazine adducts based on this method (Scheme 1),
illustrating the potential of these reactions as a more general
procedure for fullerene monoderivatisation.
Scheme 1 Reaction of C60 with substituted piperazines 1a–h.
Table 1 Yields of monoadducts 2a–h from the photochemical reaction† of
3 equiv. of piperazines 1a–h with C60 (80 mg) in toluene (80 ml)
A typical reaction‡ involved irradiation8 of a degassed
toluene solution of C60 (typically ca. 0.11 mmol) and 3
equivalents of a racemic piperazine derivative 1 under a
nitrogen atmosphere for 64 hours. A summary of the reaction
times and yields is given in Table 1, as well as the previously
reported piperazine and methylpiperazine reactions for compar-
ison. During the reaction, the solutions changed colour from
deep purple to brown, as is generally observed in fullerene
addition reactions. In all cases, reactions of substituted
Entry Addend Adduct
R
RA
t/h
Yield(%)
1
2
3
4
5
6
7
8
1a
1b
1c
1d
1e
1f
2a
2b
2c
2d
2e
2f
CH3
H
H
H
H
H
CH3
H
64
64
64
64
64
64
64
18
41
43
10
6
CH2OH
CH2OC6H13
CONH2
CONHC3H7
CH3
27
20a
b
1g
1h
2g
2h
CO2Et
H
—
H
72c
a Reaction employed 5 equivalents of 1f. b Colour change observed, but no
products isolated. c Reaction employed only 1 equivalent of piperazine
1h.
† Electronic supplementary information (ESI) available: full synthetic and
spectroscopic details. See http://www.rsc.org/suppdata/cc/b3/b303952h/
1530
CHEM. COMMUN., 2003, 1530–1531
This journal is © The Royal Society of Chemistry 2003

The 1H NMR spectra of 2a–f show all of the expected signals
derived from their corresponding starting materials 1a–f, albeit
shifted significantly downfield in 2a–f as is expected for
addends on a C60 cage. The structures of the adducts 2a–e are
presumed to be bicyclic DABCO analogues, as illustrated in
Scheme 1, similar to that of the known 2h.9 However for these
C2-substituted examples, there are two possible arrangements
for the C2-substituents, either exo or endo to the fullerene cage,
as shown schematically in Fig. 1. In all photochemical reactions
reported here, only one monoadduct was isolated. Our previous
report7 demonstrated that the methyl substituent in adduct 2a
10–62) and the sp3 hybridised fullerene carbons (d 77–81). The
amide carbonyl resonances at ca. d 171 are also observed for
both adducts 2d and 2e. The remaining observed resonances (d
136–154) correspond to sp2 hybridised fullerene carbons. While
not all of the expected 13C resonances for the fullerene cage
carbons in adducts 2a–f are fully resolved, numeric analysis
based on the smallest peak in the appropriate region suggests
that all of the expected carbon resonances are indeed present,
albeit overlapping in some cases.
In conclusion, we have developed methods for the first
efficient monoaddition of diamines containing non-hydro-
carbon functionality to C60. Good yields of the corresponding
monoadducts are achieved for cases where the piperazine
substituents are stable to the photochemical conditions. For all
monosubstituted piperazine adducts, the C2-substituent appears
to preferentially occupy an exo position on the piperazine ring,
however endo arrangements are energetically acceptable, albeit
less so, as demonstrated by the trans-2,5-dimethylpiperazine
adduct 2f. We are currently exploring methods of preparing
other polyfunctionalised examples and the synthetic potential of
adduct 2b, which can be used to access a wide variety of C60
derivatives by simple transformations of the hydroxyl func-
tional group.
1
occupies an exo-position and that exo protons give rise to H
NMR signals upfield of those corresponding to endo protons.
This matched conclusions drawn elsewhere10 for fullerene
Diels–Alder adducts, in which resonances corresponding to
endo groups in similar structures are observed downfield of the
corresponding exo protons.
We wish to thank Mr Simon Thorpe and Ms Sharon Spey at
University of Sheffield for assistance with mass spectrometry,
and the TMR networks USEFULL and BIOFULL (CT 960126
and 980192) for financial support for MJ. We acknowledge the
financial support of the School of Chemistry, University of
Exeter.
Fig. 1 Schematic diagram of addend conformations.
It is presumed that the C2-substituent in all of the mono-
substituted adducts 2b–2e also occupies an exo position. Indeed,
the H NMR spectra of 2b–e were all substantially similar to
1
that of the methyl derivative 2a with the 1H chemical shifts of
exo and endo protons fitting the trend of 2a and 2h if the C2-
substituent was assumed to be exo in all of these cases.11
Analysis of COSY and NOE experiments supported these
assignments although the crucial ‘1,3-diaxial’ NOE between the
C2-substituent and H6exo was not observed in the adducts 2b
and 2c and cannot exist in amides 2d and 2e as there are no C2-
methylene protons present.
Notes and references
‡ Photochemical experiments employed an unfiltered medium pressure
mercury lamp in a quartz water jacket, which was immersed in the reaction
solution. All reaction solutions and reagents were purged extensively with
nitrogen immediately before photolysis. The piperazine derivatives 1a, 1f
and 1g were all commercially available as the racemates, while 1b, 1d and
1e were prepared from the ester 1g. Piperazine derivative 1c was prepared
in turn from 1b. For full synthetic and spectroscopic details see ESI.†
In the case of the trans-2,5-dimethylpiperazine adduct 2f, one
methyl group must be exo to the cage, while the other methyl
group must be endo to the cage. This is shown by the loss of
symmetry in the 1H NMR spectrum of 2f compared to 1f, with
2 independent resonances being observed for the methyl groups
and 6 independent resonances for the ring protons in the former.
1 (a) A. Hirsch, Top. Curr. Chem., 1999, 199; (b) S. R. Wilson, D. I.
Schuster, B. Nuber, M. S. Meier, M. Maggini, M. Prato and R. Taylor,
Fullerenes: Chemistry, Physics, and Technology, ed. K. M. Kadish and
R. S. Ruoff, Wiley-Interscience, New York, 2000; (c) A. Hirsch,
Synthesis, 1995, 895.
2 All yields reported or discussed herein are absolute yields, rather than
yields relative to the quantity of C60 consumed.
3 A. Hirsch, Q. Li and F. Wudl, Angew. Chem., Int. Ed. Engl., 1991, 30,
1309.
4 (a) K. D. Kampe, N. Egger and M. Vogel, Angew. Chem., Int. Ed. Engl.,
1993, 32, 1174; (b) K. D. Kampe and N. Egger, Liebigs Ann., 1995,
115.
5 H. Isobe, N. Tomita and E. Nakamura, Org. Lett., 2000, 23, 3663.
6 For other examples see.(a) M. Diekers, A. Hirsch, S. Pyo, J. Rivera and
L. Echegoyen, Eur. J. Org. Chem., 1998, 1111; (b) M. Maggini, G.
Scorrano, A. Bianco, C. Toniolo and M. Prato, Tetrahedron Lett., 1995,
36, 2845.
7 C. P. Butts, R. W. A. Havenith, M. Jazdzyk, T. Drewello and S. Kotsiris,
Tetrahedron Lett., 2003, 44, 3565.
8 Sun et al. and Wang have previously reported the rapid photochemical
addition of piperazine to C60: (a) Y. P. Sun, B. Ma and C. E. Bunker, J.
Phys. Chem. A, 1998, 102, 7580; (b) N. X. Wang, Tetrahedron, 2002,
58, 2377.
1
While the loss of symmetry in the H NMR spectrum of 2f
might also be explained by formation of an adduct with only one
C60–N bond, the lack of any resonance corresponding to a C60–
1
H in the H NMR spectrum, and the identified mass of the
protonated parent ion (m/z = 833) by MALDI-TOF mass
1
spectrometry do not support this. The observed H chemical
shifts of the methyl group resonances also support the proposed
configuration, with one resonance arising at d 1.82 (exo-CH3)
and the other at d 1.97 (endo-CH3). The chemical shift of the
more upfield resonance is comparable to that observed for the
exo configured methyl group in adduct 2a (d 1.78), while the
resonance at d 1.97 fits the trend of signals arising from endo
substituents being downfield shifted. There is also an observed
NOE (ca. 2.5%) between the resonance at d 1.82 and a
resonance (d 4.15) which, according to the COSY spectrum, is
not in the same 1H coupling system of adduct 2f. This suggests
a ‘1,3-diaxial’ relationship between the two nuclei, thus
assigning them both to exo positions.
9 A. L. Balch, A. S. Ginwalla, M. M. Olmstead and R. Herbstirmer,
Tetrahedron, 1996, 52, 5021–5032.
10 M. F. Meidine, A. G. Avent, A. D. Darwish, H. W. Kroto, O. Ohashi, R.
Taylor and D. R. M. Walton, J. Chem. Soc., Perkin Trans. 2, 1994, 6,
1189 and reference 2 therein.
11 The only exception to the trend was the H3exo protons of adducts 2d and
2e, which were observed at > 4.20 ppm. This is readily explained by a
through-space deshielding effect of the adjacent C2-carbonyl group.
The incorporation of the endo substituent does appear to have
some energy cost, as the photochemical reaction of 1f with C60
was much slower than those of the monosubstituted substrates
and consequently 5 equivalents of substrate were employed for
this reaction.
In the 13C NMR spectra of all adducts, the most upfield
resonances correspond to the piperazine adduct carbons (d
CHEM. COMMUN., 2003, 1530–1531
1531