[60]Fullerene as a substituent

Article abstract of DOI:10.1002/1521-3765(20020301)8:5<1015::AID-CHEM1015>3.0.CO;2-Q

The substituent effect of the dihydro[60]fullerenyl group and its hydrophobic parameters have been evaluated quantitatively. The substituent constant has been determined from the pK value of a fullerene-based, para-substituted benzoic acid 1 in 80% dioxan

Full text of DOI:10.1002/1521-3765(20020301)8:5<1015::AID-CHEM1015>3.0.CO;2-Q

G. Scorrano et al.
The NMR titration curve of a fulleropyr-
rolidine, with a pK(BD+) of 5.6 in 85%
dioxane/water, is shown. The backdrop
depicts the HOMO and HOMO(À4)
surfaces (calculated at the PM3 level) of
N-methylfulleropyrrolidine; this suggests
a low basicity of the amino nitrogen.
For more details see the following pages.
Chem. Eur. J. 2002, 8, No. 5
¹ WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002
0947-6539/02/0805-1015 $ 17.50+.50/0
1015

FULL PAPER
[60]Fullerene as a Substituent
Alessandro Bagno,[a] Sofia Claeson,[b] Michele Maggini,[a] Maria Luisa Martini,[a]
Maurizio Prato,[c] and Gianfranco Scorrano*[a]
Abstract: The substituent effect of the
dihydro[60]fullerenyl group and its hy-
drophobic parameters have been eval-
uated quantitatively. The substituent
constant has been determined from the
pK value of a fullerene-based, para-
substituted benzoic acid 1 in 80% diox-
ane/water (v/v) by NMR spectroscopy.
The resulting Hammett s value of 0.06,
consistent with a small electron-with-
drawing effect of C60, is a consequence
of the fact that only inductive effects can
be transmitted through the two tetra-
coordinate carbon atoms between the
fullerene p system and the para-position
of the benzoic acid moiety in 1. The
parameter p, which describes the hydro-
phobic character of the substituent C60,
has been evaluated as the difference
between that of 1 and model compound
2. The p value, which is larger than 3,
indicates that the fullerene cage imparts
high hydrophobicity to the molecule to
which it is attached. Finally, we have
evaluated how the fullerene spheroid
influences the acid base properties and
nucleophilicity of the pyrrolidine nitro-
gen in a suitably functionalized fullero-
pyrrolidine. The fulleropyrrolidine
4
‡
(pKBD ˆ 5.6) is six orders of magnitude
less basic and 1000 times less reactive
‡
than its model 3 (pKBD ˆ 11.6). This
may be related to through-space inter-
actions of the nitrogen lone pair and the
fullerene p system.
Keywords: basicity
¥ fullerenes ¥
fulleropyrrolidines ¥ partition coef-
ficient ¥ substituent effects
negative potentials is observed in cyclic voltammetry of
derivatives when compared to those for pristine C60.[7]
Introduction
Following the discovery[1] and first macroscopic isolation[2] of
fullerenes, major advances in the understanding of the basic
principles of their chemical reactivity[3] have given rise to a
large number of derivatives with attractive properties for
biological applications[4] and as new materials.[5] However, the
pace of development has left comparatively little time to
collect data on the fundamental characteristics of the full-
erene moiety as a substituent.
Monofunctionalization of C60, the most widely studied
fullerene, does not substantially alter its electronic and
electrochemical properties. The UV/Vis absorption features
of C60 are, in fact, mostly retained in its derivatives.[6] Only a
slight shift of the reduction processes (ꢀ50 150 mV) to more
On the contrary, little is known on how the fullerene
substituent modifies the chemical properties of groups linked
to it (is it electron withdrawing? And to what extent?) also
with regard to its hydrophobic properties (to what extent?).
Recently, various calculations have been made on a number
of physicochemical properties of C60. Among these, logP
values for C60 have been calculated;[8] however, no conclu-
sions could be drawn since experimental data were not
available. The basic properties of C60 anions have also been
studied:[9] this revealed that the basicity increases from the
3À
mildly basic monoanion to the highly basic C60
.
Herein, we report on the quantitative evaluation of the
substituent effect of C60, on the determination of its hydro-
phobic parameters, and of its effect on the basic properties
and on the nucleophilicity of a fulleropyrrolidine derivative.
[a] Prof. G. Scorrano, Dr. A. Bagno, Prof. M. Maggini, Dr. M. L. Martini
Centro Meccanismi CNR, Dipartimento di Chimica Organica Uni-
¡
versita di Padova
Via Marzolo 1, 35131 Padova (Italy)
Fax : (‡39)049-8275239
Results
E-mail: gianfranco.scorrano@unipd.it
Substituent effect: The most widely used parameter to
measure the effect of a substituent is the Hammett constant
s,[10] which is evaluated by the measurement of its effect on the
ionization constant of a substituted benzoic acid in water. The
known extremely low solubility of C60 derivatives in water
dictates the use of an organic solvent that contains water, and
[b] Dr. S. Claeson
ERASMUS visiting student from University of Gˆteborg (Sweden).
Present address: AstraZeneca, Mˆlndal (Sweden)
[c] Prof. M. Prato
¡
Dipartimento di Scienze Farmaceutiche, Universita di Trieste
Piazzale Europa 1, 34127 Trieste (Italy)
1016
¹ WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002
0947-6539/02/0805-1016 $ 17.50+.50/0
Chem. Eur. J. 2002, 8, No. 5

[60]Fullerene as a Substituent
1015 1023
one for which the substituent effects on the ionization of
benzoic acids are already available. To ensure adequate
solubility, we prepared the fullerene-based benzoic acid
derivative 1 that contains a triethylene glycol chain, and the
reference compound 2, which lacks the fullerene moiety.
Methanofullerene 1 has been synthesized through the se-
quence described in Scheme 1. It is reasonably soluble in 80%
dioxane/water (v/v), which was then chosen as the solvent for
the study of the substituent effects (1 for benzoic acids ˆ
1.68).[11]
process, and 3) the use of a reference signal whose chemical
shift is insensitive to solvent and pH changes.
The problem of sensitivity was easily solved with the use of
a 400-MHz instrument. However, since the NMR technique
has rarely been applied to ionization equilibria in mixed
organic solvents, the other two points were carefully exam-
ined.
The most obvious choice for the signal to be monitored is
that of the aromatic protons which, in the present case of p-
substituted benzoic acids, give rise to the familiar four-line
pattern, whose splitting and appearance depend on the
relative electron-withdrawing powers of the two substituents.
p-Nitrobenzoic acid (PNBA) was chosen as a test acid for
calibration purposes.
The problem concerning the reference for the chemical
shifts has been thoroughly examined for concentrated acid
solutions.[13] It was concluded that the best reference should
have the same charge type and solvation characteristics as the
ion being examined. In the present case, in which the solvent
composition remains constant and dilute acid base solutions
are employed, the problem is expected to be of minor
importance. Two substances were chosen as candidates,
Me4NCl and MeSO3Na; both are ionic species not protonated
under the conditions of interest.
A test determination of the ionization constant of p-
nitrobenzoic acid was carried out in 60% dioxane. The pKa
values thus obtained were 5.97 and 5.96 with Me4NCl and
MeSO3Na as the reference compounds, respectively; all four
aromatic signals yielded identical values. These values com-
pare very well with the literature value of 5.99,[11] considering
the different techniques, and point out that all four aromatic
protons are suitable probes, and both standards work equally
well. However, the Me4NCl signal is close to that of dioxane
The studies of ionization equilibria are usually carried out
by monitoring the change in UV spectra of the acid as a
‡
function of pH (or other scales of H activity).[10] However, in
the present case the use of UV spectroscopy is severely
hampered because: 1) the UV spectrum is dominated by the
strong absorption of the C60 group, rather distant from the site
of ionization; 2) the strong absorption of the solvent dioxane
which totally masks the region of interest. An alternative to
UV is NMR spectroscopy,[12] where the change in the chemical
shift of a suitable signal is monitored as a function of pH. This
approach requires that the neutral and ionized forms undergo
fast exchange on the NMR timescale, so that the measured
chemical shift n is the weighted average of the values for the
neutral (nHA) and the ionized (nAÀ) species: n([HA] ‡ [AÀ]) ˆ
nHA[HA] ‡ nAÀ[AÀ], in full analogy with the expression
commonly adopted in UV measurements. This method has
been widely adopted for studies in concentrated acids and
bases, where similar problems apply.[12]
The most important problems connected to NMR measure-
ments are: 1) the sensitivity is much lower than for UV
measurements; 2) the need for a suitable signal from the
species under investigation whose chemical shift can be easily
detected and changes only as a consequence of the ionization
CO2CH3
1. SOCl2
2. CH2N2
CO2H
CO2tBu
CO2CH3
CO2H
CO2CH3
TEG-OH
H+
COCHN2
NH
Cl3CCOC(CH3)3
BF3
HCl
H2O
C-CH2-TEG
C-CH2O(CH2CH2O)3-CH3
TEG
C-CH2-TEG
O
O
O
5
8
9
1. H2NNH2
2. MnO2
6
Pd/C
H2
CO2H
CO2CH3
CO2tBu
HCl
H2O
CH2CH2-TEG
CH2CH2-TEG
C-CH2-TEG
7
2
N2
10
C60
ButO2C
toluene
HO2C
ButO2C
ButO2C
CH2-TEG
CH2-TEG
CH2-TEG
CH2-TEG
CF3SO3H
CH2Cl2
chlorobenzene
+
∆
1
12
12
11
Scheme 1. Synthesis of 1 and 2. See Experimental Section for details.
Chem. Eur. J. 2002, 8, No. 5
¹ WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002
0947-6539/02/0805-1017 $ 17.50+.50/0
1017

G. Scorrano et al.
FULL PAPER
and its intensity is adversely affected because of the partial
saturation occurring during measurements requiring satura-
tion of the solvent signal. As a consequence, MeSO3Na was
used as the standard in all subsequent measurements.
tion of the hydrophobicity of the whole molecule. The
parameter p is given by p ˆ logPX À logPH, where PX and
PH are the partition coefficients for the substituted and
unsubstituted substrate, respectively. A positive value for p
means that the substituent prefers the octanol phase (hydro-
phobic), while a negative p value indicates hydrophilic
character of the substituent.
Measurements in 80% dioxane gave somewhat different
results: PNBA yielded a pKa value of 8.25, compared to a
literature value of 8.15.[11] The origin of this discrepancy is
probably related to the uncertainty in the correction for
activity coefficients (see Experimental Section) as well as to
the problem (i.e. reproducibility) arising in connection with
the use of a glass electrode in highly nonaqueous media.
Because all our measurements were carried out under
identical conditions, pK differences will not be affected,
although some uncertainty will result in the s values. The
resulting pKa values are collected in Table 1.
The measurements for the relatively water-soluble com-
pounds (benzoic acid, PNBA, and 2), were carried out without
difficulty. On the contrary, the extremely low aqueous
solubility of 1 prompted us to determine logP values for the
model compound 2 at two different concentrations to test for
its consistency (Table 2). In any case, the solubility of 1 was
close to the detection threshold of the HPLC technique we
used; thus, only a very small peak with a signal-to-noise ratio
(S/N) of about 2could be obtained, at best. After checking
with samples of known concentration, we set an upper limit
for cwater of 6 Â 10À9 molLÀ1. To confirm the presence of 1,
5 mL of the aqueous phase after partitioning were concen-
trated to dryness under reduced pressure and the residue,
dissolved in octanol, was checked by HPLC. A peak was
clearly detected at the retention time of derivative 1 whose
concentration, calculated for the 10 mL parent solution,
qualitatively matched the upper limiting value estimated
above. The results are reported in Table 2.
Table 1. pKa Values of benzoic acids in 60% and 80% dioxane/water at
308C.
Acid
60% Dioxane
ref. [11] this work
80% Dioxane
ref. [11]
this work
benzoic acid
2
p-nitrobenzoic acid
7.12
5.99
2
7.19
5.96
9.2
9.52 Æ 0.09
8.25 Æ 0.01
9.42 Æ 0.04
8.15
1
From the data for 2, one can evaluate the hydrophobicity of
the -(CH2CH2O)4CH3 group as p ˆ À0.60. Therefore, we can
only estimate a lower limit for the logP value for 1 as ꢀ4,
which corresponds to a p value of >3. Assuming additivity of
the substituent effects, one can correct for the hydrophilic
character of the -(CH2CH2O)4CH3 group and hence estimate
the p value for the fullerene group itself as being larger than 4.
The -(CH2CH2O)4CH3 group, linked in the para position to
the aromatic ring of 2, renders the benzoic acid less acidic by
0.3 pK units. This is consistent with a small electron-donating
effect of the methylene groups; the s value of the group can
be estimated as s ˆ (9.22 À 9.52)/1.68 ˆ À0.18. The group
linked to the para position in 1 includes the fullerene; overall,
it is an acid-weakening group, but less than the
-(CH2CH2O)4CH3 group (s ˆ (9.22 À 9.42)/1.68 ˆ À0.12).
If we allow the additivity of the two parts of the substituent,
we can then estimate that the fullerene moiety has an acid-
strengthening effect of 0.1 unit. We can also estimate a s value
for the fullerene of 0.1/1.68 ˆ 0.06, indicating a small electron-
withdrawing effect for the group.
Basicity and reactivity of fulleropyrrolidines: One of the best
methods to functionalize fullerenes is the cycloaddition of
azomethine ylides.[15] This produces fulleropyrrolidines which
are characterized by the presence of a pyrrolidine ring fused
to a 6,6 ring-junction of C60. The pyrrolidine nitrogen atom is
close to the fullerene sphere: how are its basic and nucleo-
philic properties affected?
Hydrophobicity parameters: The two parameters of interest
are the partition coefficient P and the parameter p, which
describes the hydrophobic character of the substituent.[14] The
We are again faced with the problem of solubility. We
synthesized, therefore, model compounds N-methyl-2-[2(2-
methoxyethoxy)ethoxy]methylpyrrolidine (3) and N-methyl-
2-[2(2-methoxyethoxy)ethoxy]methylfulleropyrrolidine (4),
according to the sequences shown in Scheme 2.
Fulleropyrrolidine 4 is soluble (0.3 mgmlÀ1) in an 85%
dioxane/water mixture. The NMR spectrum of 4 in this
solvent shows that the signals considered to be suitable for
partition coefficient is evaluated by the expression P ˆ coctanol
/
[cwater(1 À a)], where coctanol and cwater are the concentration of
the substrate in equilibrated solutions in octanol and water,
respectively, and a is the degree of dissociation of the acid in
the aqueous phase. The partition coefficient gives an indica-
Table 2. Hydrophobicity parameters.
[b]
Compound
pH[a]
pKa
a
coctanol [m]
cwater [m]
logP
1.78
1.93
1.18
1.16
p[c]
0.15[d]
À 0.60
À 0.62
> 3
benzoic acid
p-nitrobenzoic acid
2
2
1
4.46
4.21
4.21
4.15
4.00
4.19
3.41
4.37
4.37
4.31
0.651
0.863
0.409
0.376
0.329
5.23 Â 10À3
4.63 Â 10À3
3.13 Â 10À3
1.26 Â 10À4
1.01 Â 10À4
0.25Â 10À3
0.40Â 10À3
0.35Â 10À3
0.14Â 10À4
< 6 Â10À9
> 4
[a] pH value of the aqueous phase after partitioning. [b] The pKa values for benzoic and p-nitrobenzoic acid were taken from ref. [21]. Those for 1 and 2 were
computed with the Hammett equation and the s values obtained in the previous section. [c] p ˆ logPX À logPH. [d] 0.02(ref. [14]).
1018
¹ WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002
0947-6539/02/0805-1018 $ 17.50+.50/0
Chem. Eur. J. 2002, 8, No. 5

[60]Fullerene as a Substituent
1015 1023
CH3
CH3
N
Cl(CH2CH2O)2CH3
KOH, DMSO
N
CH2OH
CH2O(CH2CH2O)2CH3
CH2O(CH2CH2O)2CH3
3
CH3
N
CH3
O
CH
HN
toluene, ∆
R = -CH2O(CH2CH2O)2CH3
CO2H
C60
+
R
4
Scheme 2. Synthesis of 3 and 4. See Experimental Section for details.
generating the chemical shift versus pH curve are close to the
solvent signals. Therefore, solvent suppression by presatura-
tion led to an excessive decrease in their intensity. To
overcome this problem, we used perdeuterated solvents
([D8]dioxane and D2O), DCl as acid, and Me4NCl as the
‡
internal reference. Obviously, the acidity constant is a pKBD
value. The pK data are collected in Table 3, while Figure 1
shows the NMR titration curves of pyrrolidines 3 and 4.
‡
Table 3. pKBH Values for substituted pyrrolidines in dioxane/water at
308C.
Figure 1. Chemical shift variations, relative to Me4NCl (TEMA), of a
methylene proton at position 5 of the pyrrolidine ring in a) 1 and b) 2 versus
pD (85% [D8]dioxane/D2O at 308C).
Compound
80% Dioxane/water
pKBH
85% [D8]Dioxane/D2O
pKBD
‡
‡
N-methylpyrrolidine
3
4
11.7 Æ 0.1
11.1 Æ 0.1
11.6 Æ 0.1
5.6 Æ 0.1
in 1 the fullerene p system is separated from the para position
of the benzoic acid ring by the two tetracoordinate carbon
atoms that bear the three-membered ring, so that only
inductive effects can be transmitted to the reacting center.
Therefore, it is not surprising that the polar effect is minimal.
With regards to the hydrophobic character, the estimated
value (p > 3) is largely positive which implies that the
fullerene substituent favors, relative to H, the octanol phase
and hence has a strong hydrophobic character. The magnitude
of p can be compared with that of cyclohexyl (2.51) and
ferrocenyl (2.46) groups,[14] while logP is similar to that of
pyrene (4.88).[14]
The basic strength of N-methylpyrrolidine compares well
with literature data for other tertiary amines in 80% dioxane/
‡
water, for example cyclohexyldimethylamine (pKBH ˆ 11.08)
‡
and triethylamine (pKBH ˆ 11.39).[16] The decrease in pK
upon going from N-methylpyrrolidine to derivative 3 shows a
small electron-withdrawing effect of the -CH2-OR substituent
(cf. sp for CH2OCH3 ˆ 0.03). On the contrary, the difference
of six orders of magnitude between the basicity of pyrrolidine
3 and that of fulleropyrrolidine 4 is impressive.
By itself, the value of 4 for logP appears to be inconsistent
À
The reactivity of pyrrolidine 3 and that of fulleropyrroly-
dine 4 has been compared in the Menschutkin reaction with
methyl iodide in CDCl3. The solvent was chosen to allow
monitoring of the reaction by NMR spectroscopy. Compound
3 reacts at 258C with a rate constant of 2.1 Â 10À3 LmolÀ1 sÀ1,
whereas the rate constant for fulleropyrrolidine 4 is, under the
same conditions, 4.2 Â 10À6 LmolÀ1 sÀ1, that is, the reaction of
the fullerene derivative is retarded by 1000-fold.
with the large size and functionalities (mainly C C bonds) of
the fullerene derivative; thus, for example, Abraham et al.[8]
estimated a much larger logP value (12.6) for the partition
between water and water/wet 1-octanol for fullerene. How-
ever, one must bear in mind that the value obtained in this
work is only a lower limit for logP or p because of the
extremely unfavorable water/octanol partition of 1 despite the
added hydrophilic group.
In addition, it is worth mentioning that in polar solvents
hydrophilic fullerene monoadducts can form supramolecular
aggregates.[17] Similar aggregation phenomena have been
demonstrated for C60 itself in solvents having a dielectric
permittivity (e) larger than ꢀ13 (e values for water and
1-octanol are 78.3 and 10.3, respectively).[18] This occurrence
introduces the additional difficulties that a) the nature of the
Discussion
The low s value (0.06) evaluated for the fullerene as a
substituent is, at first sight, surprising given the fact that the
fullerene core is a well-known electron acceptor.[7] However,
Chem. Eur. J. 2002, 8, No. 5
¹ WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002
0947-6539/02/0805-1019 $ 17.50+.50/0
1019

G. Scorrano et al.
FULL PAPER
‡
‡
dissolved species is solvent-dependent, and b) the solubility of
such aggregates might differ substantially from that of the
monomeric species. Therefore, it is worthwhile to make an
independent estimation of the relative magnitude of logP
values of 1 and 2 as follows: assuming logP ˆ 12.6 for 1, as
proposed by Abraham, and a value of 1.2for 2, the Gibbs
energies of transfer (octanol ! water) can be estimated as
DGto!w ˆ 2.303RTlogP ˆ 17 and 2kcalmol À1 for 1 and 2,
respectively. This difference can be accounted for as follows:
the energetics of transfer of both solutes is dictated by
a) cavitation energies and b) solute solvent interactions. The
latter term can be further dissected into polar contribution
(i.e. dipole dipole and hydrogen-bonding interactions) and
dispersive contributions (mainly the result of the solute and
solvent polarizabilities). Polar contributions can be assumed
equal for 1 and 2, since both species share the same number
and type of suitable functionalities. Conversely, their polar-
izabilities (a) are substantially different: an HF/3-21G
calculation yields average volume polarizabilities of a ˆ 88
and 24 ä3 for 1 and 2, respectively (the value for 1 compares
favorably with the value for C60 itself, a ˆ 86 ä3).[19] The
polarizability of water (1.45 ä3) is also substantially smaller
than that of 1-octanol (17 ä3, estimated from that of di-n-
butyl ether[20]). From the classical expression for the disper-
sive interaction of two spherical particles with given a and
ionization potential I, [Eq. (1)], it is qualitatively apparent
difference between pKBD and pKBH is of the order of 0.2
0.6 units; obviously the same 6 pK units difference will
probably hold in the pH scale for 3 and 4). Similar results
have been recently reported by D×Souza et al., who ascribed
the differences to inductive effects.[22a] Siegel et al. determined
the relative basicity of N-methylfulleropyrrolidine and N,N-
dimethylaniline by means of a competitive titration NMR
method in a fully nonaqueous medium (CS2/CDCl3).[22b] These
measurements indicated the two bases to have essentially
equal pK values; however, even assuming that solvent effects
on the proton-transfer equilibrium are the same, it is not
possible to derive a quantitative assessment of the basicity of
the fullerene derivative, owing to the large uncertainty
associated with measuring acid strengths in nonaqueous,
nonpolar media.
Polar effects cannot be responsible for this large decrease in
basicity (see above and compare with the s value of 0.06). On
the other hand, steric effects can also be discarded on the
following grounds: although the basicity of aliphatic and
aromatic nitrogen bases is indeed affected by steric hindrance
(the hindered species being less basic), the known basicity
differences amount to only 1 2p K units.[23] Furthermore,
such relatively large effects are only encountered when the
nitrogen atom lies within a cavity formed by the bulky
substituents, for example in the cases of 2,4- versus 2,6-di-tert-
butylpyridine, or piperidine versus cis-2,6-di-tert-butylpiper-
idine. Even then, the observed steric effects have been
demonstrated to be connected with the solvation of the
protonated base rather than with protonation itself.[23b] In
contrast, the optimized structure of N-methylfulleropyrroli-
dine using the PM3 semiempirical method (Figure 2) shows
2
E ˆ À a1a
3
I1I2
1
(1)
2 I1 ‡ I2 r612
that the interaction between the most polarizable solute
solvent pair (i.e., 1 and 1-octanol) would be the most
stabilizing. However, a quantitative assessment is hampered
by the need of estimating the approach distance (difficult to
define for nonspherical particles), and by the difficulty in
accounting for the complex collective phenomena taking
place in a bulk medium.
In fact, however, the strongest factor operating in the
transfer appears to be the cavitation energy. According to the
scaled particle theory (SPT),[21] this term depends primarily on
the relative diameter of solute and solvent molecules, so that
the most endergonic process is the creation of a cavity in a
solvent whose molecules are very different in diameter from
those of the solute. The Gibbs energies of transfer (octanol !
water) can be simply derived from the solvation (gas ! sol-
vent) energies. Thus, even allowing for a fairly wide uncer-
tainty in the molecular diameters, according to the SPT the
DGto!w is much less favorable for
1 than for 2 by
ꢀ20 kcalmolÀ1, which is in the same range as the difference
in the experimental DlogP. This arises mainly because the
cavitation energy of 1 in water is very unfavorable (owing to
the very large difference in size between water and 1).
Hence, our calculated estimate agrees with that calculated
by Abraham: both data point to an extremely high hydro-
phobic character of the fullerene sphere. The experimental
result only yields a lower limit for logP. The large exper-
imental difficulties, including aggregation of fullerene in
water, do not allow for a better estimate.
Figure 2. a) HOMO and b) HOMO(À4) surfaces of N-methylfullero-
pyrrolidine calculated with the PM3 semiempirical method.
that the pyrrolidine ring is in the envelope conformation, with
the lone pair in the axial position pointing toward the
fullerene sphere. The nitrogen atom is therefore far away
from the fullerene core, so that the solvation of the
ammonium ion can be expected to be free of any steric
influence.
Conversely, it is well known that the aqueous basicity of
‡
‡
The pKBD values of compounds 3 and 4 differ by 6 units!
anilines (pKBH ꢀ 4 5) is much lower than that of aliphatic
‡
‡
(For other weak acids, such as RCO2H and ArNH3 , the
amines (pKBH ꢀ 10 11). In this case, the effect stems from an
1020
¹ WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002
0947-6539/02/0805-1020 $ 17.50+.50/0
Chem. Eur. J. 2002, 8, No. 5

[60]Fullerene as a Substituent
1015 1023
(see partition coefficient measurements) were performed on a Shimadzu
HPLC unit (Shimadzu LC-8A pump, 20 mL injections, SCL-8A system
controller, SPD-6A spectrophotometric detector at l ˆ 254 or 340 nm
(derivative 1)). Reactions were monitored by thin-layer chromatography
on Macherey-Nagel plates (Polygram SILG/UV254, 0.2mm thickness).
Flash column chromatography was performed on 230 400 mesh silica gel
(Macherey-Nagel). Reaction yields were not optimized and refer to pure,
isolated products. PM3 and ab initio calculations were performed with
Spartan4[28] and Gaussian98,[29] respectively.
intrinsically lower basicity of anilines (in the gas phase,
cyclohexylamine is more basic than aniline by 11 kcalmolÀ1),
the solvation energy of the respective ammonium ions being
similar.[24] This is commonly ascribed to a participation of the
nitrogen lone pair to the benzene resonance; we suggest that
the lower basicity of the fulleropyrrolidine derives from a
similar effect. In fact, PM3 calculations show that the HOMO
frontier orbital of N-methylfulleropyrrolidine (Figure 2a) is
localized almost exclusively on the fullerene. Only the
HOMO(À4) molecular orbital, shown in Figure 2b, displays
electron density on the pyrrolidine nitrogen. Therefore, it
appears that the nitrogen lone pair lies in a lower energy
orbital and, as such, is not readily available for protonation.
This is also consistent with the reduced nucleophilic reac-
tivity: the 1000-fold decrease in reaction rate on going from
pyrrolidine to its fullerene derivative is of the expected order
of magnitude.
Materials: C60 was purchased from Bucky-USA (99.5%). All other
reagents were purchased from Aldrich. Fulleropyrrolidine 4[30] was
prepared as reported in the literature.
2-[2-(2-Methoxyethoxy)ethoxymethyl]-1-methylpyrrolidine (3): 1-Methyl-
2-pyrrolidine-methanol (2.0 g, 17 mmol) and bis(ethyleneoxy)methyl
chloride (4.7 g, 34 mmol) were added to a suspension of KOH (3.8 g) in
DMSO (33 mL). The mixture was stirred at room temperature for 2.5 h.
Brine (100 mL) was added followed by extraction of the aqueous phase
with CH2Cl2. The crude oily residue, obtained after concentration of the
organic phase, was distilled (b.p. 1708C, 2mmHg) to yield derivative
3
(1.2g; 34%). An analytical sample was obtained by further bulb-to-bulb
distillation. 1H NMR (250 MHz, CDCl3, 2 58C, TMS): d ˆ 1.48 1.94 (m,
4H), 2.10 2.20 (dd, 1H), 2.30 2.42 (dd, 1H), 2.36 (s, 3H), 2.98 3.08 (td,
1H), 3.34 (s, 3H), 3.32 3.65 (m, 8H); 13C NMR (62.5 MHz, CDCl3, 2 58C,
TMS): d ˆ 57.29, 58.49, 64.25, 70.04, 70.10, 70.19, 71.48; HR-MS calcd for
C11H23NO3 217.1672; found 217.1639 (Æ0.005).
There are other examples of through-space interactions
between the fullerene p system and orbitals of the added
functional group. In spiromethanofullerenes, for instance, a
similar effect has been called periconjugation.[25] Another
example has been reported in the EPR study of radical anions
Methyl 4-(2-Diazoacetyl)benzoate (5): Terephthalic acid monomethyl ester
(1.0 g, 5.6 mmol) was dissolved in freshly distilled thionyl chloride (10 mL)
and 3 drops of DMF. The mixture was heated to reflux temperature for 3 h,
then excess thionyl chloride was removed in vacuo. Crude methyl
4-chlorocarbonylbenzoate (m.p. 54 568C; 1H NMR (250 MHz, CDCl3,
258C, TMS) d ˆ 3.98 (s, 3H), 8.18 (s, 4H); 13C NMR (62.5 MHz, CDCl3,
258C, TMS) d ˆ 52.7, 130.0, 131.2, 136.0, 136.5, 165.2, 168.0) was dissolved
in diethyl ether (20 mL) and added to a solution of diazomethane in diethyl
ether (ꢀ16 mmol) cooled on an ice-bath. When the addition was complete
(15 min), the cooling bath was removed and the solution was stirred at
room temperature for 3 h. The solvent and excess diazomethane were
removed with a stream of nitrogen. The residue was purified by flash
chromatography (SiO2, eluent: toluene/ethyl acetate 1:1, Rf ˆ 0.7) to yield 5
as yellow crystals (860 mg; 76%). M.p. 107 1098C; 1H NMR (200 MHz,
CDCl3, 2 58C, TMS): d ˆ 3.95 (s, 3H), 5.95 (s, 1H), 7.82(m, 2H), 8.12(m,
2H); 13C NMR (62.5 MHz, CDCl3, 2 58C, TMS): d 52.1, 54.7, 126.4, 129.5,
133.3, 139.7, 165.8, 185.1; IR (neat): nÄ ˆ 2366, 1717, 1680 cmÀ1; GC-MS
of fulleropyrrolidine 4 by Brustolon and co-workers.[26]
A
further case of nitrogen lone pair delocalization into the
fullerene core is that of fulleroproline 13 for which we
measured the kinetic parameters of the trans ! cis isomer-
ization.[27]
H3C
O
O
CH3
C
N
C
N
CO2tBu
CO2tBu
.
‡
(70 eV): m/z (%): 163 (100) [M À CHN2]
.
trans
cis
Methyl 4-(2-{2-[2-(2-methoxyethoxy)ethoxy]ethoxy}acetyl)benzoate (6):
Triethylene glycol monomethyl ether (TEG-OH) (1.4 mL, 8.1 mmol)
followed by 5 drops of BF3 ¥ Et2O were added to a solution of 5 (850 mg,
4.17 mmol) in diethyl ether (120 mL). The solution was left standing for
two days at room temperature. Aqueous workup afforded an oily residue
which was purified by flash column chromatography (SiO2, toluene/ethyl
13
The activation enthalpy for the isomerization process is
about 7 kcalmolÀ1 lower than that of the proline analogue.
This means that in 13 the nitrogen lone pair is less available
for conjugation with the carbonyl with a consequent loss in the
1
acetate 1:1) to yield 6 (820 mg; 58%). H NMR (200 MHz, CDCl3, 2 58C,
TMS): d ˆ 3.36 (s, 3H), 3.50 3.80 (m, 12H, triethylene glycol), 3.96 (s, 3H),
8.0 (m, 2H), 8.12 (m, 2H); 13C NMR (62.5 MHz, CDCl3, 2 58C, TMS): d ˆ
51.8, 58.3, 69.7, 69.8, 70.1, 70.2, 71.1, 73.7, 127.2, 129.1, 133.4, 137.3, 165.4,
195.4; IR (neat): nÄ ˆ 1724, 1704, 1281, 1108 cmÀ1; GC-MS (70 eV): m/z (%):
À
C N double-bond character.
.
‡
340 (10) [M]
.
Methyl 4-(2-{2-[2-(2-methoxyethoxy)ethoxy]ethoxy}ethyl)benzoate (7):
Pd/C (ꢀ15 mg) was added to a solution of 6 (580 mg, 1.71 mmol) in
ethanol (20 mL) and the mixture was stirred under hydrogen atmosphere
(1 atm) for four days. After filtration of the catalyst, the solvent was
evaporated in vacuo and the product purified by flash column chromatog-
raphy (SiO2, ethyl acetate/methanol 95:5) to yield 7 (320 mg; 58%) as a
clear oil. 1H NMR (250 MHz, CDCl3, 2 58C, TMS): d ˆ 2.92 (t, 2H), 3.37 (s,
3H), 3.55 (m, 2H), 3.5 3.8 (m, 14H), 3.89 (s, 3H), 7.28 (m, 2H), 7.94 (m,
2H); 13C NMR (62.5 MHz, CDCl3, 2 58C, TMS): d 36.23, 51.96, 58.99, 70.31,
70.47, 70.53, 70.58, 71.59, 71.90, 128.90, 129.59, 144.59, 167.02; IR (neat): nÄ ˆ
Experimental Section
Instrumentation: 1H and 13C NMR spectra were recorded on Bruker
AC200, AC250 and AM400 spectrometers. Chemical shifts are given
relative to tetramethylsilane. UV/Vis absorption spectra were collected on
a Perkin-Elmer Lambda5 spectrophotometer. FT-IR spectra were collect-
ed on a Perkin-Elmer 1720X spectrophotometer. MALDI (matrix-assisted
laser desorption ionization) mass spectra were obtained in positive linear
mode at 15 kV acceleration voltage on a Reflex time-of-flight mass
spectrometer (Bruker), with 2,5-dihydroxybenzoic acid as the matrix. GC-
MS analyses were performed on a Hewlett-Packard electron-impact mass
spectrometer5970 coupled with a gas chromatograph8890 equipped with a
30 m  0.25 mm Alltech EC-1 column. The exact mass determination was
performed on a VGZAB2F instrument (70 eV, 200 mA). Isocratic elutions
‡
1700, 1224, 1111 cmÀ1; GC-MS (70 eV): m/z (%): 326 (10)[M] .
4-(2-{2-[2-(2-Methoxyethoxy)ethoxy]ethoxy}ethyl)benzoic acid (2): A sus-
pension of ester 7 (420 mg, 1.29 mmol) in aqueous HCl (10%, 80 mL) was
heated to reflux temperature for 2h then the solvent was removed in vacuo.
The product, dissolved in CH2Cl2 (15 mL) was filtered through a pad of
Chem. Eur. J. 2002, 8, No. 5
¹ WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002
0947-6539/02/0805-1021 $ 17.50+.50/0
1021

G. Scorrano et al.
FULL PAPER
celite. Evaporation of the solvent afforded the benzoic acid derivative 2 in
nearly quantitative yield as a deliquescent solid that was used for acidity
and partition coefficient measurements without further purification.
1H NMR (250 MHz, CDCl3, 2 58C, TMS): d ˆ 2.96 (t, 2H), 3.38 (s, 3H),
and the mixture was stirred at room temperature overnight. After the
solvent had been evaporated at reduced pressure, the residue was
transferred in a centrifuge tube with the aid of diethyl ether, washed
thoroughly with the same solvent and then with acetonitrile to give 1
(25 mg, 86%) as a brownish solid; 1H NMR (250 MHz, CDCl3/CS2 2:1,
258C, TMS): d ˆ 3.38 (s, 3H), 3.55 (m, 2H), 3.63 (m, 6H), 3.70 (m, 2H), 3.82
(m, 2H), 4.86 (s, 2H), 8.13 (m, 2H), 8.28 (m, 2H); 13H NMR (62.5 MHz,
CDCl3/CS2 2:1, 25 8C, TMS): d ˆ 50.71, 59.06, 70.46, 70.59, 70.65, 70.71,
71.31, 71.56, 71.88, 129.18, 129.89, 132.48, 137.62, 137.84, 140.74, 140.94,
141.64, 141.87, 141.96, 142.83, 142.87, 142.92, 143.0, 143.61, 144.09, 144.36,
144.41, 144.52, 144.65, 144.83, 144.99, 145.06, 145.49, 147.00, 147.61, 192.35;
3.5 3.9 (m, 14H), 7.32(m, 2H), 8.01 (m, 2H);
13C NMR (62.5 MHz,
CDCl3, 2 58C, TMS): d ˆ 36.26, 58.99, 70.28, 70.43, 70.50, 70.54, 71.54, 71.87,
127.40, 129.03, 130.18, 145.43, 171.40; IR (KBr): nÄ ˆ 1712, 1249, 1107 cmÀ1
;
elemental analysis calcd (%) for C16H24O6 ¥ 0.5H2O (321.4): C 59.79, H 7.84;
found: C 59.38, H 7.74.
4-(2-{2-[2-(2-Methoxyethoxy)ethoxy]ethoxy}acetyl)benzoic acid (8): The
hydrolysis of the ester function of benzoate 6 (470 mg, 1.38 mmol) was
performed as previously described for derivative 7. Benzoic acid 8 (440 mg,
98%) was isolated and used for the next reaction without further
purification. 1H NMR (200 MHz, CDCl3, 2 58C, TMS): d ˆ 3.38 (s, 3H),
3.5 3.9 (m, 12H, triethylene glycol), 4.88 (s, 2H), 8.0 (m, 2H), 8.15 (m,
2H); 13C NMR (62.5 MHz, CDCl3, 2 58C, TMS): d ˆ 58.83, 61.38, 70.12,
70.24, 70.28, 70.38, 70.63, 70.78, 71.71, 72.34, 74.28, 127.75, 130.14, 133.96,
138.12, 168.42, 196.05; IR (KBr): nÄ ˆ 1684, 1700 cmÀ1; EI-MS (70 eV): m/z:
IR (KBr): nÄ ˆ 2864, 1720, 1681, 1105, 526 cmÀ1; UV/Vis (octanol): lmax
ˆ
‡
219, 256, 322, 428 nm; MALDI MS C76H22O6 (1031): m/z: 1054 [M‡Na] .
Acidity measurements: 1,4-Dioxane was freed from peroxides by perco-
lation through activated neutral alumina. Stock solutions were prepared by
dissolving the acid in the dioxane/water solution (0.1m in NaClO4)
containing CH3SO3Na; the final concentrations were 10À3 m for p-nitro-
benzoic acid, 5 Â 10À3 m for 2 and CH3SO3Na. The stock solution of 1 was
prepared by dissolving the acid in dioxane and adding the required amount
of water containing NaClO4 and CH3SO3Na (final concentration 5 Â
10À4 m). 16 24 aliquots were then brought to the desired pH by addition
of dilute NaOH or HCl in the same solvent. The pH-meter was calibrated
with a standard buffer solution; no difference was observed if the glass
electrode was conditioned in 3m KCl or in 80% dioxane. pH-Meter
readings (B values) were corrected to pH values as recommended by
van Uitert and Fernelius[32] as pH ˆ B ‡ D with 0.1 mol kgÀ1 as the molality
of the electrolyte, and x ˆ 0.45 as the mole fraction of the organic solvent.
.
‡
326 [M]
.
tert-Butyl
4-(2-{2-[2-(2-Methoxyethoxy)ethoxy]ethoxy}acetyl)benzoate
(9): A solution of tert-butyl-2,2,2-trichloroacetimidate (750 mg, 3.4 mmol)
in n-hexane (7 mL) at room temperature was added to a solution of acid 8
(440 mg, 1.35 mmol) in CH2Cl2 (7 mL). One drop of BF3 ¥ Et2O was added
and the mixture stirred at room temperature for 20 h. After addition of
solid Na2CO3, the solution was filtered through a pad of celite and then
evaporated under reduced pressure. The product was purified by flash
column chromatography (SiO2, eluent: ethyl acetate) to yield 9 (190 mg;
o
The value of D (ˆ logUH ‡ log1/g) was estimated by interpolation of the
1
37%) as a clear oil. H NMR (200 MHz, CDCl3, 2 58C, TMS): d ˆ 1.62(s,
o
published values of logUH and log1/g so that for the 80% dioxane/water
9H), 3.38 (s, 3H), 3.48 3.82(m, 12H, triethylene glycol), 4.85 (s, 2H), 7.95
(m, 2H), 8.08 (m, 2H); 13C NMR (62.5 MHz, CDCl3, 2 58C, TMS): d ˆ 28.1,
59.0, 70.47, 70.53, 70.79, 70.91, 71.9, 74.4, 81.8, 127.7, 129.6, 136.1 137.7, 164.7,
196.2; IR (neat) 1705 1715 (br) cmÀ1; GC-MS (70 eV): m/z (%): 263 (20)
solution a value of D ˆ 2.52 was calculated. However, the uncertainty
related to the sparse values for interpolation, and the steepness of the curve
of D versus x limits the accuracy of D to Æ0.1. 1H NMR measurements were
carried out on a Bruker AM400 instrument operating at 400 MHz without
a lock. The probe temperature (308C) was checked with a sample of neat
ethylene glycol. Typically, for each measurement 32 128 transients were
accumulated in 32K data points. Chemical shifts in Hz are referenced to
internal CH3SO3Na. The strong resonance of the solvent was suppressed by
multiple presaturation (40 cycles, 55 ms each with total irradiation time of
2.2 s). The total acquisition time was 15 20 min per sample. The resulting
values of chemical shift as a function of pH were fitted to Equation (2)
where I ˆ [HA]/[AÀ] and logI ˆ ÀpH ‡ pKa.
.
‡
{M À [(OCH2CH2)2OCH3]}
.
Homofullerene 11: The hydrazone of 9 was prepared by dissolving 9
(80 mg, 0.21 mmol) in absolute ethanol (7 mL). Excess anhydrous hydra-
zine (0.1 mL) was added and the mixture was stirred at reflux temperature
for 1.5 h. The solvent was removed on the rotary evaporator and excess
hydrazine under high vacuum (0.1 torr) at 408C. The oily residue was
dissolved in chloroform (10 mL), and dry MgSO4 (200 mg) was added
followed by portionwise addition of MnO2 (200 mg). After stirring at room
temperature for 10 min, the resulting yellow solution containing diazo
derivative 10 (IR (neat): nÄ ˆ 2052 (nN N), 1708 (nC O) cmÀ1) was filtered
ˆ
ˆ
InHA ‡ nAÀ
1 ‡ I
through a pad of celite in order to remove all solids. Because of the
instability towards SiO2 and concentration, the diazo compound was added
directly to a solution of C60 (100 mg, 0.14 mmol) in toluene (120 mL). The
solution was stirred at room temperature for 1.5 h. The solvents were
removed in vacuo and the residue was purified by flash column
chromatography (SiO2, eluent: toluene ! toluene/ethyl acetate 75:25).
Yield: 45 mg (30%); 1H NMR (250 MHz, CDCl3, 2 58C, TMS): d ˆ 1.63 (s,
9H), 3.37 (s, 3H), 3.4 3.7 (m, 12H, triethylene glycol), 4.83 (s, 2H), 8.09
(m, 2H), 8.17 (m, 2H); 13C NMR (62.5 MHz, CDCl3, 2 58C, TMS): d ˆ 28.2,
50.8, 59.0, 61.1, 70.4, 70.5, 70.6, 70.7, 71.3, 71.8, 81.2, 125.3, 128.2, 129.0,
129.26, 129.3, 131.0, 131.4, 132.4, 134.9, 137.15, 137.2, 138.0, 138.3, 138.7,
139.4, 140.41, 140.45, 141.31, 141.38, 142.11, 142.16, 142.29, 149.32, 142.82,
142.94, 142.97, 143.09, 143.15, 143.50, 143.54, 143.7, 143.8, 143.9, 144.2,
144.8, 145.1, 145.2, 147.3, 150.2, 165.0; UV/Vis (cyclohexane): lmax ˆ 214,
n ˆ
(2)
The values of pKa, nHA, and nAÀ were optimized by nonlinear least-squares
fitting. The calculated limiting chemical shifts (nHA and nAÀ) were then used
to calculate logI values; those between À1 and 1 were used to calculate pK
values which were eventually averaged to yield the recommended values,
along with their standard deviations.
Basicity measurements: The general procedure is the same as described for
the acidity measurements, except that the solvent was 85% [D8]dioxane/
D2O and hence D ˆ 3.29; pH readings were corrected to pD by adding
0.40.[16] The chemical shifts were referenced to internal Me4NCl. All
measurements were carried out at 308C. Data were processed as described
for the acidity measurements.
‡
260, 332, 423 nm; MALDI MS C80H30O6 (1086): m/z: 1087 [M‡H] .
Kinetic measurements: The reactions were carried out in CDCl3 as the
solvent at 258C. Stock solutions of CH3I, compounds 3 and 4 in CDCl3 were
prepared and thermostatted at 258C. Aliquots were transferred into a
NMR tube and the initial concentration determined by integrating the CH3
signal against that of CHCl3 present as impurity in the deuterated solvent.
The CHCl3 concentration, determined by integration of a known solution
Methanofullerene 12: Homofullerene 11 was converted quantitatively to
methanofullerene 12 according to the literature[31] by heating overnight a
0.5 mm solution of 11 in 1,2-dichlorobenzene. 1H NMR (250 MHz, CDCl3,
258C, TMS): d ˆ 1.64 (s, 9H), 3.37 (s, 3H), 3.55 (m, 2H), 3.59 3.68 (m,
6H), 3.70 (m, 2H), 3.80 (m, 2H), 4.83 (s, 2H), 8.05 (m, 2H), 8.16 (m, 2H);
13C NMR (62.5 MHz, CDCl3, 2 58C, TMS): d ˆ 28.22π 50.79, 59.05, 70.55,
70.68, 70.72, 71.41, 71.84, 71.92, 81.27, 129.26, 131.94, 132.96, 137.74, 137.98,
140.46, 140.81, 141.04, 142.08, 142.14, 142.95, 143.0, 143.04, 143.14, 143.76,
144.16, 144.46, 144.56, 144.67, 144.71, 144.80, 145.03, 145.10, 145.20, 145.71,
147.32, 148.01, 165.44; UV/Vis (cyclohexane): lmax ˆ 216, 264, 327, 430 nm;
of
a
standard (cubane dimethyl ester) is 2.51 Â 10À2 m. The starting
concentration of the reactants was evaluated as cX ˆ cS(AX/AS), where cX
and cS are the concentrations of the reactant and of CHCl3, respectively,
and AX and AS the integrated area of the CH3 signal for the reagent and of
the methine proton for chloroform. The concentrations were: a) 1.4 Â
10À2 m for 3 and 8.0 Â 10À2 m for CH3I and b) 4 Â 10À3 m for 4 and 4.8 Â
10À1m for CH3I. Rate constants were evaluated by the usual plots of
concentration versus time.
‡
MALDI MS C80H30O6 (1086): m/z: 1109 [M‡Na] .
Methanofullerene 1: Trifluoromethanesulfonic acid (0.2mL) was added to
a solution of methanofullerene 12 (30 mg, 0.28 mmol) in CH2Cl2 (5 mL),
1022
¹ WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002
0947-6539/02/0805-1022 $ 17.50+.50/0
Chem. Eur. J. 2002, 8, No. 5

[60]Fullerene as a Substituent
1015 1023
Partition coefficients measurements: Deionized water was shaken with
distilled octanol for five minutes. Octanol-saturated water was boiled for
30 min to remove CO2 and was kept under N2 atmosphere during the
cooling process. The substrate (Table 2) was dissolved in octanol. The
octanol solution (10 mL) and water (10 mL) were stirred vigorously for
1.25 h under a nitrogen atmosphere. The phases were left standing for
10 min so that they could separate. The pH was measured (Metrohm632
pH-meter) for the aqueous phase and the concentration of the substrate
determined in both phases by means of HPLC with the aid of external
standards (CH3CN solutions of naphthalene for benzoic and p-nitrobenzoic
acid; b-naphthol for compound 2 and benzo[a]pyrene for compound 1. 1 ml
of CH3CN standard solution was added to 1 mL of the octanol or aqueous
phase after partitioning). Isocratic elution was performed on a Shimadzu
HPLC station (see instrumentation). Conditions for benzoic acid, p-
nitrobenzoic acid and 2: column Techsphere, ODS (C18), 250 Â 4.6 mm, 5m,
flow 1 mLminÀ1. Eluent: CH3CN/H2O 65:35, 0.05% TFA (benzoic and p-
nitrobenzoic acid); CH3CN/H2O 35:65, 0.05% TFA (derivative 2). Con-
ditions for derivative 1: column Vydac201TP510 (C18), 150 Â 4.6 mm, 5 m,
flow 1 mLminÀ1. Eluent: CH3CN/CHCl3 60:40, 0.05% TFA.
2655; e) P. Timmerman, H. L. Anderson, F. Rudiger, J.-F. Nierengart-
en, T. Habicher, P. Seiler, F. Diederich, Tetrahedron 1996, 52, 4925;
f) G. Van Lier, B. Safi, P. Geerlings, J. Chem. Soc. Perkin Trans. 2 1998,
349; g) C. A. Reed, R. D. Bolskar, Chem. Rev. 2000, 100, 1075.
[10] L. P. Hammett in Physical Organic Chemistry, McGraw-Hill, New
York, 1970.
[11] D. V. Jahargirdar, D. N. Shelke, R. G. Deshpande, J. Chem. Soc. Perkin
Trans. 2 1977, 1513.
[12] a) A. Bagno, G. Scorrano, Acc. Chem. Res. 2000, 33, 609; b) A. Bagno,
G. Scorrano, R. A. More O×Ferrall, Rev. Chem. Intermed. 1987, 7, 313.
[13] A. Bagno, V. Lucchini, G. Scorrano, Bull. Soc. Chim. Fr. 1987, 563.
[14] C. Hansch, A. J. Leo in Substituent Constants for Correlation Analysis
in Chemistry and Biology, Wiley, New York, 1979.
[15] M. Prato, M. Maggini, Acc. Chem. Res. 1998, 31, 519.
[16] P. K. Glasoe, F. A. Long, J. Phys. Chem. 1960, 64, 188.
[17] a) D. M. Guldi, M. Prato, Acc. Chem. Res. 2000, 33, 695; b) G.
Angelini, P. De Maria, A. Fontana, M. Pierini, M. Maggini, F.
Gasparrini, G. Zappia, Langmuir 2001, 17, 6404.
[18] S. Nath, H. Pal, A. V. Sapre, Chem. Phys. Lett. 2000, 327, 143.
[19] P. Norman, Y. Luo, D. Jonsson, H. ägren, J. Chem. Phys. 1997, 106,
8788.
[20] CRC Handbook of Chemistry and Physics, 77th ed. (Ed.: D. R. Lide),
CRC Press, Boca Raton, 1997.
Acknowledgement
[21] R. A. Pierotti, Chem. Rev. 1976, 76, 717.
We thank G. Walder and C. Sotriffer (ERASMUS visiting students from
the University of Innsbruck), Dr. F. Della Negra, and Dr. S. Bucella for
their help in working out detailed experimental procedures. We are grateful
to Dr. R. Seraglia (CNR, Padova) for mass spectra and to Dr. M. Zancato
(Department of Pharmaceutical Science, University of Padova) for
elemental analyses. This work was supported by MURST (contract
no. 9803194198).
[22] a) F. D×Souza, M. E. Zandler, R. D. Gollapalli, K. Klodzimierz, J.
Phys. Chem. A 2000, 104, 6887; b) R. J. Steffens, K. K. Baldridge, J. S.
Siegel, Helv. Chim. Acta 2000, 83, 2644.
[23] a) E. M. Arnett, G. Scorrano, Adv. Phys. Org. Chem. 1976, 13, 83;
b) R. W. Alder, Chem. Rev. 1989, 89, 1215.
[24] R. W. Taft, Prog. Phys. Org. Chem. 1983, 14, 247.
[25] M. Eiermann, R. C. Haddon, B. Knight, Q. C. Li, M. Maggini, N.
Martin, T. Ohno, M. Prato, T. Suzuki, F. Wudl, Angew. Chem. 1995,
107, 1755; Angew. Chem. Int. Ed. 1995, 34, 1591.
[26] M. Brustolon, A. Zoleo, G. Agostini, M. Maggini, J. Phys. Chem. A
1998, 102, 6331.
[1] H. W. Kroto, J. R. Heath, S. C. O×Brien, R. F. Curl, R. E. Smalley,
Nature 1985, 318, 162.
[2] W. Kr‰tschmer, L. D. Lamb, K. Fostiropoulos, D. R. Huffman, Nature
1990, 347, 354.
[27] A. Bianco, V. Lucchini, M. Maggini, M. Prato, G. Scorrano, C. Toniolo,
J. Peptide Sci. 1998, 4, 364.
[3] a) F. Diederich, C. Thilgen, Science 1996, 271, 317; b) A. Hirsch in
Fullerenes and Related Structures, Topics in Current Chemistry,
Vol. 199, Springer, Berlin 1999; c) Fullerenes: Chemistry, Physics and
Technology (Eds.: K. Kadish, R. Ruoff), Wiley Interscience, 2000.
[4] a) T. Da Ros, M. Prato, Chem. Commun. 1999, 663; b) A. W. Jensen,
S. R. Wilson, D. I. Schuster, Bioorg. Med. Chem. 1996, 4, 767; c) E.
Nakamura, H. Tokuyama, S. Yamago, T. Shiraki, Y. Sugiura, Bull.
Chem. Soc. Jpn. 1996, 69, 2143.
[5] a) C. A. Mirkin, W. B. Caldwell, Tetrahedron 1996, 52, 5113; b) M.
Prato, J. Mater. Chem. 1997, 7, 1097; c) N. Martin, L. Sanchez, B.
Illescas, I. Perez, Chem. Rev. 1998, 98, 2527; d) K. E. Geckeler, S.
Samal, Polym. Int. 1999, 48, 743; e) H. Imahori, Y. Sakata, Eur. J. Org.
Chem. 1999, 2445; f) D. M. Guldi, M. Prato, Acc. Chem. Res. 2000, 33,
695; g) J. L. Segura, N. Martin, J. Mater. Chem. 2000, 10, 2403.
[6] C. S. Foote, Top. Curr. Chem. 1994, 169, 347.
[28] Spartan 4.0, Wavefunction, Inc. 18401 von Karman Ave. 370, Irvine,
CA 92715
[29] Gaussian98, Revision A.7, M. J. Frisch, G. W. Trucks, H. B. Schlegel,
G. E. Scuseria, M. A. Robb, J. R. Cheeseman, V. G. Zakrzewski, J. A.
Montgomery, Jr., R. E. Stratmann, J. C. Burant, S. Dapprich, J. M.
Millam, A. D. Daniels, K. N. Kudin, M. C. Strain, O. Farkas, J. Tomasi,
V. Barone, M. Cossi, R. Cammi, B. Mennucci, C. Pomelli, C. Adamo,
S. Clifford, J. Ochterski, G. A. Petersson, P. Y. Ayala, Q. Cui, K.
Morokuma, D. K. Malick, A. D. Rabuck, K. Raghavachari, J. B.
Foresman, J. Cioslowski, J. V. Ortiz, B. B. Stefanov, G. Liu, A.
Liashenko, P. Piskorz, I. Komaromi, R. Gomperts, R. L. Martin,
D. J. Fox, T. Keith, M. A. Al-Laham, C. Y. Peng, A. Nanayakkara, C.
Gonzalez, M. Challacombe, P. M. W. Gill, B. Johnson, W. Chen, M. W.
Wong, J. L. Andres, C. Gonzalez, M. Head-Gordon, E. S. Replogle,
J. A. Pople, Gaussian, Inc.: Pittsburgh PA, 1998.
[7] L. Echegoyen, L. E. Echegoyen, Acc. Chem. Res. 1998, 31, 593.
[8] M. H. Abraham, C. E. Green, W. E. Acree, J. Chem. Soc. Perkin Trans.
2 2000, 281.
[30] M. Maggini, G. Scorrano, M. Prato, G. Brusatin, P. Innocenzi, M.
Guglielmi, A. Renier, R. Signorini, M. Meneghetti, R. Bozio, Adv.
Mater. 1995, 7, 404.
[9] a) P. J. Fagan, P. J. Krusic, D. H. Evans, S. A. Lerke, E. Johnston, J.
Am. Chem. Soc. 1992, 114, 9697; b) D. E. Cliffel, A. J. Bard, J. Phys.
Chem. 1994, 98, 8140; c) M. E. Niyazymbetov, D. H. Evans, S. A.
Lerke, P. A. Cahill, C. C. Henderson, J. Phys. Chem. 1994, 98, 13093;
d) M. E. Niyazymbetov, D. H. Evans, J. Electrochem. Soc. 1995, 142,
[31] F. Diederich, L. Isaacs, D. Philp, Chem. Soc. Rev. 1994, 23, 243.
[32] G. Van Uitert, W. C. Fernelius, J. Am. Chem. Soc. 1954, 76, 5887.
Received: June 28, 2001 [F3381]
Chem. Eur. J. 2002, 8, No. 5
¹ WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002
0947-6539/02/0805-1023 $ 17.50+.50/0
1023