[60]fullerene-pyrrolidine-N-oxides

Article abstract of DOI:10.1021/jo052388s

Eight members of a new family of fullerene derivatives, [60]fulleropyrrolidine-N-oxides, have been synthesized and characterized. Facile oxidation, by a peracid, of the parent [60]fulleropyrrolidine gave clean conversions into the product molecules, in wh

Full text of DOI:10.1021/jo052388s

[60]Fullerene-Pyrrolidine-N-oxides
Peter Brough,† Cedric Klumpp,†,‡ Alberto Bianco,‡ Stephane Campidelli,† and
Maurizio Prato*,†
Dipartimento di Scienze Farmaceutiche, UniVersita` degli Studi di Trieste, 34127 Trieste, Italy, and Institut
de Biologie Mole´culaire et Cellulaire, UPR 9021 CNRS, 67084 Strasbourg, France
prato@units.it
ReceiVed NoVember 18, 2005
Eight members of a new family of fullerene derivatives, [60]fulleropyrrolidine-N-oxides, have been
synthesized and characterized. Facile oxidation, by a peracid, of the parent [60]fulleropyrrolidine gave
clean conversions into the product molecules, in which the tertiary amine is transformed into a quaternary
amine bearing an oxygen atom. The reaction is very selective, favoring the nitrogen atom of the pyrrolidine
1
ring in preference to epoxidation of the fullerene cage. The H NMR shows an AB quartet splitting
pattern, characteristic of nonequivalent hydrogens in the pyrrolidine ring and at a chemical shift
displacement of 0.8 ppm downfield. Other methods of characterization are described, including MS,
differential scanning calorimetry, thermogravimetric analysis, HPLC, UV/vis, and IR. Conclusive evidence
for the formation of an N-oxide moiety is provided by the synthesis, oxidation, and NMR characterization
of a novel [60]fulleropyrrolidine containing a 15N isotope, showing an 85 ppm downfield heteroatom
chemical shift. Preliminary details of the effects of substitution on the reactivity of the pyrrolidine ring
are also reported.
Introduction
functionalization.4 Alkylation using iodoalkanes gives the cor-
responding quaternary ammonium salt, which has been used in
the study of fullerene aggregation and also possesses interesting
electrochemical properties.5,6 The pyrrolidine nitrogen atom may
also be a source of some unwanted problems when applied to
biological systems, in that tertiary amines are protonated under
physiological conditions. DNA-interaction studies have shown
The fascinating chemistry surrounding the C60 cage has
emerged since the development of its first synthesis in workable
quantities.1,2 Of the many ways to functionalize the sphere, the
1,3-dipolar cycloaddition of azomethine ylides provides a great
deal of flexibility owing to the readily available starting
materials.3 As part of this reaction, a basic nitrogen atom is
introduced onto the fullerene cage in the form of a pyrrolidine
ring. Although the reactivity is much reduced in comparison to
a normal amine, it still can be used as a point of further
(4) Bagno, A.; Claeson, S.; Maggini, M.; Martini, M. L.; Prato, M.;
Scorrano, G. Chem.sEur. J. 2002, 8, 1015-1023.
(5) (a) Guldi, D. M.; Goulomis, A.; Vazquez, P.; Torres, T.; Georgakilas,
V.; Prato, M. J. Am. Chem. Soc. 2005, 127, 5811-5813. (b) Nakashima,
N.; Ishii, T.; Shirakusa, M.; Nakanishi, T.; Murakami, H.; Sagara, T.
Chem.sEur. J. 2001, 7, 1766-1772. (c) Cassell, A. M.; Asplund, C. K.;
Tour, J. M. Angew. Chem., Int. Ed. 1999, 38, 2403-2405. (d) Angelini,
G.; Cusan, C.; De Maria, P.; Fontana, A.; Maggini, M.; Pierini, M.; Prato,
M.; Schergna, S.; Villani, C. Eur. J. Org. Chem. 2005, 9, 1884-1891. (e)
Georgakilas, V.; Pellarini, F.; Prato, M.; Guldi, D. M.; Melle-Franco, M.;
Zerbetto, F. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 5075-5080. (f) Guldi,
D. M.; Zerbetto, F.; Georgakilas, V.; Prato, M. Acc. Chem. Res. 2005, 38,
38-43. (g) Brough, P.; Bonifazi, D.; Prato, M. Tetrahedron, in press (doi:
10.1016/j.tet.2005.08.125).
† Universita` degli Studi di Trieste.
‡ Institut de Biologie Mole´culaire et Cellulaire.
(1) (a) Wudl, F. Acc. Chem. Res. 1992, 25, 157-161. (b) Diederich, F.;
Isaacs, L.; Philp, D. Chem. Soc. ReV. 1994, 23, 243-255. (c) Balch, A. L.;
Olmstead, M. M. Chem. ReV. 1998, 98, 2123-2165. (d) Diederich, F.;
Kessinger, R. Acc. Chem. Res. 1999, 32, 537-545. (e) Hirsch, A.; Nuber,
B. Acc. Chem. Res. 1999, 32, 795-804.
(2) Kra¨tschmer, W.; Lamb, L. D.; Fostiropoulos, K.; Huffman, D. R.
Nature 1990, 347, 354.
(3) Maggini, M.; Scorrano, G.; Prato, M. J. Am. Chem. Soc. 1993, 115,
9798-9799.
(6) Da Ros, T.; Prato, M.; Carano, M.; Ceroni, P.; Paolucci, F.; Roffia,
S. J. Am. Chem. Soc. 1998, 120, 11645-11648.
10.1021/jo052388s CCC: $33.50 © 2006 American Chemical Society
Published on Web 02/04/2006
2014
J. Org. Chem. 2006, 71, 2014-2020

[60]Fullerene-Pyrrolidine-N-oxides
SCHEME 1. Oxidation of [60]Fulleropyrrolidines Using mCPBA
that protonated tertiary amines form an attractive electrostatic
interaction with phosphates.7 A method of masking this behavior
is to convert the amine into an amine-N-oxide, which forms a
repulsive electrostatic interaction with the DNA phosphates. In
a more general sense, the amine-N-oxide is also known to be a
useful protecting group in synthetic work.8
Results and Discussion
Fulleropyrrolidine derivatives 1a-c, 2a-c, and 3a,b were
synthesized by the well-known 1,3-dipolar cycloaddition reac-
tion (Scheme 1), using the appropriate amino acids and
aldehydes.3
Many efficient oxidants are known for the conversion of
amines into their oxides, such as dimethyldioxirane,11 HOF‚CH3-
CN,12 bis(trimethylsilyl)peroxide,13 and Mg-Al/H2O2.14 In this
work, we have used 3-chloroperoxybenzoic acid (mCPBA), a
reagent introduced in the 1970s for oxidizing substituted amines
and is commonly used in nitroxide chemistry, owing to its good
solubility in organic solvents.15,16 The former is an important
consideration, because fullerene molecules are, generally, soluble
in solvents such as dichloromethane or THF but much less
soluble in alcoholic/aqueous media. The mCPBA used in this
work was used as received to work with commercially available
starting materials.
In this work, we present the exploration of the oxidative
transformation of [60]fullerene-pyrrolidines into their corre-
sponding N-oxides, in which the tertiary amine is oxidized to a
quaternary amine bearing an oxygen atom, (R)3N+O- (Scheme
1). We envisaged some useful properties stemming from the
ability to unmask the N-oxide back into its parent amine by
thermal treatment.9 We have judiciously chosen the oxidant and
conditions to prevent or minimize epoxidation reactions that
are known, from the literature, to occur on C60.10
(7) Patterson, L.H.; Craven, M. R.; Fisher, G. R.; Teesdale-Spittle, P.
Oncol. Res. 1994, 6, 533.
(8) Greene, T. W.; Wuts, P. G. M. ProtectiVe Groups in Organic
Synthesis, 2nd ed.; John Wiley & Sons: New York, 1991; p 375.
(9) Ma, S.; Chowdury, S. K.; Alton, K. B. Anal. Chem. 2005, 77, 3676-
3682.
We have split the eight new N-oxides into three groups
according to their functionalization, revealing valuable observa-
tions regarding their synthesis, stability, and physical properties.
(10) (a) Fusco, C.; Seraglia, R.; Lucchini, V. J. Org. Chem. 1999, 64,
8363-8368. (b) Ko, W. B.; Nam, J. H.; Hwang, S. H. Ultrasonics 2004,
42, 611-615. (c) Antipin, A. V.; Shishlov, N. M.; Grabovskii, S. A.;
Kabal’nova, N. N. Russ. Chem. Bull. 2004, 53, 800-802. (d) Elemes, Y.;
Silverman, S. K.; Sheu, C.; Kao, M.; Foote, C. S.; Alvarez, M. M.; Whetten,
R. L. Angew. Chem., Int. Ed. Engl. 1992, 31, 351-353.
(11) Ferrer, M.; Sanchez-Baeza, F.; Messeguer, A. Tetrahedron 1997,
53, 15877-15888.
(13) Murray, R. W.; Singh, M. Synth. Commun. 1989, 19, 3509.
(14) Choudary, B. M.; Barathi, B.; Venkat Reddy, C.; Lakshmi Kentam,
M.; Raghavan, K. V. Chem. Commun. 2001, 1736-1737.
(15) Chapelet-Lectourneux, G.; Lemaire, H.; Rassat, A. Bull. Soc. Chim.
Fr. 1966, 3283-3290.
(16) (a) Brough, P.; Chiarelli, R.; Pecaut, J.; Rassat, A.; Rey, P. Chem.
Commun. 2003, 2722-2723. (b) Hirel, C.; Vostrikova, K. E.; Pecaut, J.;
Ovcharenko, V. I.; Rey, P. Chem.sEur. J. 2001, 7, 2007-2014.
(12) Rozen, S. Pure Appl. Chem. 1999, 71, 481-487.
J. Org. Chem, Vol. 71, No. 5, 2006 2015

Brough et al.
FIGURE 2. HPLC of N-oxide molecule 4a.
these products were all of a similar value, hindering their
separation, and we did not pursue characterization any further.
Evidently, the known oxidation reaction of C60 was a significant
side reaction.10 Thus, control of reaction conditions and stoi-
chiometry is essential for obtaining N-oxide products in good
purity.
It is useful to note that the addition of a large excess of
oxidant (10 equiv) results in the addition of multiple oxygen
atoms along with partial decomposition of the fullerene mol-
ecule.
FIGURE 1. Oxidation of 4c under dilute conditions and (inset) using
concentrated conditions.
The first series, molecules 1a-c, consists of fairly bulky groups
occupying the tertiary position on the pyrrolidino nitrogen.
Particular attention was made toward the incorporation of a
substituent capable of hydrogen bonding to the N-oxide, such
as the Boc-protected amine function. This is a useful tool that
aids stabilization.17 The second series of molecules (2a-c),
bearing a small methyl substituent, were chosen to minimize
steric crowding around the nitrogen atom while introducing
some functionalization at the ortho position. Last, to provide
insight into the effects of steric hindrance, we describe molecules
3a,b, in which both the nitrogen of the pyrrolidine and its ortho
position are substituted.
We began our research by oxidizing the molecules 1a-c, in
which the nitrogen of the pyrrolidine ring bears only one
substituent, denoted R. The known epoxidation reactions of
fullerene itself were initially of some concern, and to minimize
this undesired side reaction, we employed dilute conditions,
typically 0.05 mmol of fullerene molecule 1c in 100 mL of
CHCl3. The addition of the mCPBA oxidant over a long period
of time (1 h) and also using a dilute solution (0.05 mmol in 25
mL of CHCl3) gave a product of mass corresponding to the
addition of an oxygen atom (M + 16), 4c (Figure 1). Its identity
was confirmed by NMR and its purity was confirmed by HPLC
(Figure 2). The yield of product was sufficient (30%), but higher
conversions would be much more useful. With this in mind,
we conducted the same reaction using more concentrated
solutions. Oxidation using rapid addition times (5 min) furnished
a mixture of products that result from the addition of one, two,
and three oxygen atoms. An example is shown in Figure 1
(inset), in which compound 1c was dissolved in CHCl3 (8 mL)
and was oxidized using a concentrated solution of mCPBA (1
equiv in 1 mL of CHCl3). The mass spectrum shows that the
product 4c contains two major fractions, M + 16 and M + 32,
and also a small amount of M + 48. We attribute these further
addition products, M + 32 and M + 48, to the epoxidation of
the fullerene cage. The chromatographic retention factors of
The solubility of fullerene-N-oxide molecules in organic
solvents is different to their unoxidized counterparts, for
example, compound 1b dissolves well in toluene, whereas the
N-oxide derivative 4b is only sparingly soluble. This is a
common theme spanning all of the molecules described in this
work. The N-oxide molecules are much more soluble in solvent
systems of greater polarity, for example, toluene/EtOH (8:2)
and CHCl3/EtOH (8:2). This proved to be very useful in this
work, in which we were able to perform 13C NMR analysis on
compound 5a owing to its good solubility in an 8:2 mixture of
CDCl3/CD3OD.
The starting [60]fullerene-pyrrolidine molecules, in the cases
of 1a-c and 2a-c, can be recuperated in their original states
from the column chromatography (toluene/ethyl acetate (7:3)
as eluent) in 45-55% yield. Mass spectra data corresponded
to those of the starting materials. This is useful information but
at the same time very curious. In the recovered compounds there
is an absence of any oxidized materials, which ought to be
present if the oxidation mechanism was indiscriminate with
respect to the C60 cage and the pyrrolidino group. Indeed, careful
monitoring of the reaction progress by TLC revealed that the
N-oxide is the first product to arrive, and then during the latter
stages a second product ensues, containing two or more oxygen
atoms. These observations suggest that the reaction takes place
preferentially on the nitrogen of the pyrrolidine, and the fullerene
cage is only oxidized during the latter stages. This is consistent
with an initial nucleophilic attack of the pyrrolidino nitrogen
atom, which is known to resemble a tertiary amine, on a
hydroxyl group of the peracid.18
From the unoxidized to the oxidized form there is a distinct
change in the 1H NMR chemical shift of the hydrogens
belonging to the pyrrolidine group (Figure 3). The former shows
as a singlet in the 4.4-4.6 ppm region, owing to rapid pyramidal
(17) O’Neil, I. A.; Potter, A. J.; Southern, J. M.; Steiner, A.; Barkley, J.
V. Chem. Commun. 1998, 2511-2512.
(18) Ibne-Rasa, K. M.; Edwards, J. O. J. Am. Chem. Soc. 1962, 84, 763.
2016 J. Org. Chem., Vol. 71, No. 5, 2006

[60]Fullerene-Pyrrolidine-N-oxides
performed using paraformaldehyde and the standard conditions
of the 1,3-dipolar cycloaddition, giving the 15N-containing [60]-
fulleropyrrolidine, 13. Oxidation to the N-oxide 14 using
mCPBA was carried out using the same method as that described
for molecule 1b.
Compound 14 contains a 15N atom in the pyrrolidine group,
and this was used as a direct NMR probe to observe the
difference between the nitrogen atoms of 13 and those of its
1
oxidized counterpart 14. The H-15N correlation spectrum is
shown in Figure 4, in which a large difference is observed in
chemical shift. The nitrogen atom of 13 shows a signal at -340
ppm, whereas in the N-oxide 14 is it observed at -255 ppm.
This exceptionally large downfield shift signifies a distinct
change in the nitrogen atom and is firm evidence for the
formation of an N-oxide moiety.
Both compounds 13 and 14 are present on the same NMR
spectrum, owing to a partial deoxygenation of the latter into
the former. It seems to occur in the solid state, but at present,
the process is not fully understood.
FIGURE 3. 1H NMR of 4b (top) vs 1b (bottom) in CDCl3.
inversion, whereas in the case of [60]pyrrolidino-N-oxides, they
show as an AB quartet in the 5.2-5.5 ppm region. This splitting
pattern was affected by different solvent systems in which well-
separated AB quartets were observed for compound 4a in
CDCl3, but the same molecule in CDCl3/CD3OD (8:2) gave a
singlet.
The UV/vis profiles of molecules [60]fullerene-pyrrolidine-
N-oxide, 4b, and the nonoxidized precursor 1b, at room
temperature, are shown together in Figure 5. Both compounds
share common absorption bands at 700, 325, and 430 nm, a
characteristic of [6,6]-bridged monoadducts. These are consistent
with previously reported mono-addition products of fullerenes.21
Surprisingly, the presence of an N-oxide moiety on the pyrro-
lidine has little or no influence on the absorption maxima, in
comparison to the precursor molecule. Other [60]fulleropyrro-
lidine molecules bearing formal cations exhibit a UV/vis blue
shift from the normal 325-317 nm.22 However, this is not the
case in the amine oxides. We speculate that the special push-
pull characteristics of this group, that is, it can act as an electron
donor and/or electron acceptor, negate any observable changes
in the π-electron system.
Fullerenes, in which the carbon cage has been epoxidized,
show distinct signals in the 13C NMR, typically the sp3 C-O
bond appears at around 90 ppm.10d In most cases, this peak was
absent, indicating that in these molecules the oxygen atoms are
not present on the fullerene cage. However, a peak correspond-
ing to the epoxidized fullerene was clearly observed in the ortho-
phenyl molecule, 5c. Even tuning the reaction conditions to favor
one oxygen addition (using 0.5 equiv of mCPBA) failed to give
a pure product. The reason for this is not clear at this time, but
it could be attributed to an electronic effect, that is, the phenyl
group deactivating the amine and making it a poorer nucleophile.
In summary of the analytical data, the mass spectra are
1
1
consistent with the addition of one oxygen atom. In the H
The H NMR is a useful indication for the formation of
NMR, the pyrrolidine hydrogens are split and moved further
downfield toward the 5.0-5.5 ppm region, indicating that the
nitrogen atom is much more deshielded and is not susceptible
to pyramidal inversion. Signals corresponding to the epoxidation
of the fullerene cage were absent in the 13C NMR data. Upon
oxidation, the nitrogen chemical shift of a 15N-labeled molecule
shifts 85 ppm downfield, owing to the different electronic
environment. Thus, we can conclude that these molecules
possess the N-oxide group.
Synthetically valuable molecules23 can be prepared by rear-
rangement reactions of tertiary amine oxides, such as those of
Meisenheimer24 and Cope.25 Deoxygenation is also a known
occurrence under thermal activation.9 The [60]fulleropyrrolidine-
N-oxides belong to this class of molecules, and in this capacity,
we have carried out some initial studies into their thermal
behavior.
N-oxide products; however, it is an indirect method of charac-
terization, that is, the presence of an amine oxide is inferred by
observing its effect on the hydrogens of the pyrrolidine. A more
elegant method is to directly observe the spectroscopic change
of the nitrogen atom from before to after oxidation. To
conclusively show the presence of the nitrogen bearing an oxide
moiety, we have synthesized the fullerene molecule 14 using a
nitrogen-15 isotope in the pyrrolidine ring system.
The synthesis of the 15N-labeled amino acid commenced from
the commercially available 15N glycine (Scheme 2). The first
step was to protect the amine function using a tert-butyloxy-
carbonyl (Boc) group, giving molecule 7 (95% yield).19 Protec-
tion of the carboxylic acid was then carried out by a reaction
with benzyl bromide, furnishing 8 in 76% yield. To further
functionalize the amine group, it was first necessary to remove
the Boc (compound 9) and replace it with another protecting
group, 2-nitrobenzenesulfonyl chloride (nosyl chloride), obtain-
ing compound 10. This step was essential because the secondary
amine retained sufficient reactivity for the subsequent Mit-
sunobu-type reaction with triethylene glycol monomethyl ether,
giving molecule 11 in 45% yield.20 Deprotection of the nosyl
group in 11, using 2-mercaptoethanol and a strong base, gave
the 15N amino acid derivative, 12. Coupling to the C60 was then
(21) Kordatos, K.; Da Ros, T.; Prato, M.; Bensasson, R. V.; Leach, S.
Chem. Phys. 2003, 293, 263-280.
(22) Kordatos, K.; Da Ros, T.; Bosi, S.; Vazquez, E.; Bergamin, M.;
Cusan, C.; Pellarini, F.; Tomberli, V.; Baiti, B.; Pantarotto, D.; Georgakilas,
V.; Spalluto, G.; Prato, M. J. Org. Chem. 2001, 66, 4915-4920.
(23) (a) O’Neil, I. A.; Cleator, E.; Elena Ramos, V.; Chorlton, A. P.;
Tapolczay, D. J. Tetrahedron Lett. 2004, 45, 3655-3658. (b) Szabo, A.;
Galambos-Farago, A.; Mucsi, Z.; Timari, G.; Vasvari-Debreczy, L.;
Hermecz, I. Eur. J. Org. Chem. 2004, 687-694. (c) Buston, J. E. H.;
Coldham, I.; Mulholland, K. R. Synlett 1997, 322-324.
(19) Goodman, M.; Felix, A.; Moroder, L.; Toniolo, C. Methods of
Organic Chemistry (Houben-Weyl); Thieme: Stuttgart, 2002; Vol. E22a.
(20) Mitsunobu, O. Synthesis 1981, 1-28.
(24) Meisenheimer, J. Ber. Dtsch. Chem. Ges. 1919, 52, 1667.
(25) Cope, A. C.; Trumbull, E. R. Org. React. 1960, 11, 317-349.
J. Org. Chem, Vol. 71, No. 5, 2006 2017

Brough et al.
SCHEME 2. Synthesis of 15N-Labeled [60]Fullerene-Pyrrolidine-N-oxide 14; (i) Boc2O, NaOH; (ii) Benzyl Bromide, DBU; (iii)
HCl; (iv) Nosyl Chloride, Et3N; (v) Triethylene Glycol Monomethyl Ether, DEAD; (vi) 2-Mercaptoethanol, LiOH; (vii) C60,
Paraformaldehyde; (viii) mCPBA
When a three heat-cool cycle (50-200 °C) is used, dif-
ferential scanning calorimetry (DSC) analysis of 5a shows a
broad exothermic peak at 134.8 °C during the first cycle (Figure
6). Notably, subsequent thermal profiles of the same molecule
did not show this energy transition. All other N-oxide com-
pounds studied gave similar traces, see Supporting Information
of their precursor molecules. Thus, under thermal activation the
N-oxide function had been lost and the compound had returned
to its former [60]fullerene-pyrrolidine 2a. Of course, this does
not exclude the possibility that an oxygen atom was transferred
to the fullerene cage. To elucidate the whereabouts of the
1
for further information. H NMR analysis of the heat-cycled
solid showed peaks and splitting patterns corresponding to those
FIGURE 4. 1H-15N correlation spectrum of molecules 13 and 14.
Inset: zoom of the peak at ∼5.35 ppm. The intensity of this correlation
peak was weak because the JNH coupling constant of the heteronuclear
multiple-bond correlation pulse sequence was not optimized.
FIGURE 5. UV of N-oxide 4b (s) vs normal 1b (- ‚ -); the inset is
zoomed between 400 and 750 nm.
2018 J. Org. Chem., Vol. 71, No. 5, 2006

[60]Fullerene-Pyrrolidine-N-oxides
SCHEME 3. Oxidation and Thermal Deoxygenation of
[60]Fulleropyrrolidine-N-oxides
direct match for the theoretical (2.02%), within experimental
error. It is worthy to note that at higher temperatures other
reactions may take place, for example, the fascinating retro-
cycloaddition reaction of pyrrolidinofullerenes in which the
pyrrolidine ring is deconstructed, reforming the parent C60.26
In the solid state, the deoxygenation reaction did not occur
cleanly, and the product contained a significant amount of
degraded molecules. To improve on the quality of the final
materials, we explored solution and suspension deoxygenation.
The first solvent we examined was CHCl3, a solvent able to
dissolve 4b, yet not able to participate in hydrogen bonding.
Under reflux conditions, there was no discernible deoxygenation
after 12 h. Protic solvents are known to assist in the oxygen
removal of N-oxide molecules, so we examined refluxing
compound 4c in toluene/methanol (4:1), Scheme 3.27 After 12
h, mass spectrometry and NMR showed that complete conver-
sion to 1c had occurred, with a small amount of impurities. It
is also possible to gently convert 4c into 1c by suspending the
solid in water and warming at 40 °C. Using this method, we
could see a clean but slow conversion (approximately 50% after
24 h).
FIGURE 6. DSC profile of 5a.
To explore further the generality of the oxidation, we
introduced a bulky group at the ortho position to the pyrrolidino
nitrogen (Scheme 1). Fullerene molecules 3a,b were easily
prepared by employing benzaldehyde and 3-pyridine carbox-
aldehyde, in the place of formaldehyde, in the cycloaddition
reaction. Oxidation did not proceed smoothly, as was the case
for 1a-c. After the addition of 1 equiv of mCPBA, a quantity
that previously generated reasonable product yields, only traces
of oxidized material could be observed by TLC. Further
oxidation, using another 1 equiv of oxidant, gave sufficient
quantities of product, which were isolated using conventional
chromatography. HPLC analysis of the products revealed that
they contained a mixture of several different products, all having
similar retention times. This was verified by mass spectra, which
showed multiple oxygen additions, and also by NMR that were
characterized by many broad peaks. With conventional chro-
matography, separation of these molecules proved to be futile,
and we did not continue the study of this complicated mixture.
We speculate that steric crowding around the pyrrolidine
group tends to inhibit oxidation of the nitrogen atom. In this
case, the fullerene cage may be preferentially epoxidized at
many different positions. Electronic effects may also be of
significance, in which electron-withdrawing substituents may
partially deactivate the pyrrolidine nitrogen toward nucleophilic
attack. An example of this is the oxidation of 2b and 2c, in
which the former, bearing an ortho-alkyl group, proceeds
FIGURE 7. Close-up view of the TGA of 5a decomposition in the
range of 50-200 °C.
oxygen, we performed mass spectrometry on some thermally
treated molecules, revealing that the oxygen atom had been lost
and the mass of the major product corresponded to that of the
parent molecule. Some degradation was also present, as seen
in signals below 720 mass units.
To eliminate degradation effects from addends we examined
the thermal profiles of precursor molecules, 1b-c, 2a-c, and
3a,b, and no exothermic or endothermic peaks were observed.
Temperatures significantly higher than those corresponding to
deoxygenation resulted in the breakdown of the tert-butylcar-
bonyl group in molecule 1a, under thermal gravimetric analysis
(TGA).
Using a conventional ramp-heating profile, TGA would
normally show the loss of mass corresponding to an oxygen
atom. However, the DSC data already showed that deoxygen-
ation in the N-oxide molecules 4a-c and 5a-c is a slow
progressive process, occurring through a wide temperature range.
Additionally, the percentage weight loss of one oxygen atom
would be very small (1-2%) and difficult to observe. In fact,
we found there was a gradual weight loss in the TGA analyses,
but accurate measurement was difficult owing to the shallowness
of the gradient. A much better method was to quickly ramp the
temperature to a point in the middle of the deoxygenation
temperature range and to measure the weight loss as a function
of time. Figure 7 shows an example in the case of N-oxide
molecule 5a in which the observed weight loss (2.20%) is a
(26) Mart´ın, N.; Altable, M.; Filippone, S.; Mart´ın-Domenech, A.;
Echegoyen, L.; Cardona, C. M. Angew. Chem., Int. Ed. 2006, 45, 110-
114.
(27) Hans-Rene, B.; Gambarotti, C.; Jensen, V. R.; Gonzalez, R. R. J.
Org. Chem. 2005, 70, 3218-3224.
J. Org. Chem, Vol. 71, No. 5, 2006 2019

Brough et al.
1.51 (br s, 9H, CH3).13C NMR: δ 154.73, 146.19, 146.01, 145.94,
145.37, 145.23, 144.49, 143.05, 142.58, 142.14, 142.01, 141.83,
140.11, 136.15, 70.66, 67.75, 54.20, 28.54. IR-DRIFT (KBr): 2980,
2790, 1705, 1510, 1120, 770, 525.
smoothly, while the latter, bearing an ortho-phenyl group, gives
reduced yields of N-oxide and significant quantities of epoxi-
dized products.
General Procedure BsOxidation of Compounds 1a-c, 2a-
c, 3a,b, and 13. To a solution of compounds 1a-c, 2a-c, 3a,b,
and 13 (0.0497 mmol in 100 mL CHCl3) was added, dropwise,
mCPBA (1 equiv, 0.0497 mmol, in 25 mL CHCl3) over a period
of 1 h. Progress was monitored by TLC. The solution was stirred
for an additional 30 min and then filtered through a short column
of SiO2, first washing with toluene to remove organic compounds
and then with eluent solution to take the product mixture. A
concentration on a rotavap to a small amount of liquid (ap-
proximately 3 mL) and flash chromatography on SiO2 (toluene/
EtOH or CHCl3/EtOH mixtures as eluent) gave the product (4a-
c, 5a-c, 6a,b, and 14) as a brown solid.
Conclusion
[60]Fullerene-pyrrolidine-N-oxides have been prepared in
moderate yields (20-40%) using the mCPBA oxidation of [60]-
fullerene-pyrrolidines. Selective oxidation of the nitrogen atom
was favored under dilute conditions, whereas concentrated
solutions furnished mixtures of products. Molecules in which
the pyrrolidine is sterically crowded tend to shift oxidation away
from the pyrrolidino-nitrogen atom and onto the fullerene
moiety, giving a multitude of products.
A novel 15N-labeled [60]fullerene-pyrrolidine molecule was
synthesized and oxidized to the corresponding N-oxide. The
presence of a 15N-oxide moiety was confirmed using 15N NMR,
in which the chemical shift of the nitrogen atom was displaced
85 ppm further downfield in comparison to the unoxidized
precursor.
Thermal profiling showed that when heated (80-180 °C) in
both solid state and solution, deoxygenation occurs such that
the oxygen atom from the N-oxide is irreversibly lost.
We have prepared a number of molecules that should be
attractive compounds for the study of coordination complexes
and biological applications.
Compound 4a. C69H18N2O3 (MW, 922.89), 32.8% yield (15.5
mg, 0.016 mmol). The eluent used was CHCl3/EtOH (9:1). UV
(CH2Cl2) λmax, nm: 255, 322, 430, 703. MS (CHCl3/CH3OH, 1:1)
m/z 924 M+. Purity was checked on HPLC and estimated at 97%.
1H NMR (CDCl3): δ 5.42-5.22 (AB quartet, 4H, J ) 10.27 Hz,
pyrrolidino Hs), 4.15 (br s, 4H, -CH2CH2), 1.50 (s, 9H, tert-
butylcarbonyl). 13C NMR (CDCl3/CD3OD, 8:2): δ 153.92, 151.51,
147.11, 146.05, 145.59, 145.72, 145.59, 145.44, 145.08, 145.00,
144.68, 144.47, 144.06, 142.39, 141.81, 141.61, 141.46, 141.24,
139.75, 137.78, 135.34, 79.89, 68.86, 35.63, 28.12. IR-DRIFT
(KBr): 2973, 1710, 1518, 1373, 1288, 1266, 1174, 767, 545 cm-1
.
Acknowledgment. We would like to thank Dr. Davide
Bonifazi for the NMR spectra and many useful discussions, Dr
Tatiana Da Ros for the HPLC data, and Mr Fabio Hollan for
the mass spectra. This work was supported by European Union,
Human Potential Networks WONDERFULL (Contract HPRN-
CT-2002-00177), the University of Trieste, MIUR (PRIN 2004,
Prot. 2004035502), and the CNRS. We are grateful to Dr.
Roland Graff and Dr. Lionel Allouche (Louis Pasteur University,
Strasbourg) for recording the heteronuclear NMR data. Ce´dric
Klumpp is grateful to Universita` italo-francese/Universite´
franco-italienne for financially supporting his international Ph.D.
Experimental Section
General Procedure for 1,3-Dipolar Cycloaddition. A toluene
solution (200 mL) of C60 (200.0 mg, 0.28 mmol) and the appropriate
aldehyde (0.28 mmol) and amino acid (0.28 mmol) was heated to
reflux for 3 h. After cooling to room temperature, the product was
purified by column chromatography (toluene or toluene/ethyl
acetate, 7:3). The brown solids 1a-c, 2a-c, and 3a,b were
dissolved in CH2Cl2 and precipitated by the addition of CH3OH or
Et2O. This precipitation procedure was repeated another two times.
Compound 1a.28 C69H18N2O2 (MW, 906.89), 30.7% yield (78.8
mg, 0.086 mmol). MS (CHCl3/MeOH, 1:1) m/z 907 M+. UV (CH2-
Cl2) λmax, nm: 255, 322, 430, 703. 1H NMR: δ 4.47 (s, 4H,
pyrrolidine), 3.71-3.65 (m, 2H, CH2), 3.29-3.23 (m, 2H, CH2),
Supporting Information Available: Experimental procedure,
1H and 13C or ES-MS spectra for compounds 1c, 2b,c, 4a-c, 5a-
c, and 7-14. TGA analysis for 5a, DSC analyses for 4c and 5b,c,
and HPLC analysis of 6a. This material is available free of charge
via the Internet at http://pubs.acs.org.
(28) Kordatos, K.; Da Ros, T.; Bosi, S.; Vazquez, E.; Bergamin, M.;
Cusan, C.; Pellarini, F.; Tomberli, V.; Baiti, B.; Pantarotto, D.; Georgakilas,
V.; Spalluto, G.; Prato, M. J. Org. Chem. 2001, 66, 4915-4920.
JO052388S
2020 J. Org. Chem., Vol. 71, No. 5, 2006