Inductive and Mesomeric Effects of the [60]Fulleropyrrolidine Fragment and [60]Fullerene Sphere: A Quantitative Evaluation Based on Theory and Experiments

Article abstract of DOI:10.1002/ejoc.201100997

Inductive and mesomeric effects of the [60]fulleropyrrolidine (Pyr=C ₆₀) and the [60]fullerene (C₆₀) molecular frameworks have been investigated by linear free energy relationship analyses. The electronic effects of these moieties have been studied by expressly designing [60]fulleropyrrolidine derivative 1, which has a C-H bond α to both a carbonyl group (C-H_α) and a C₆₀ cage. The extent of polarization of such a bond was then used to probe electron-withdrawing perturbation induced by the fullerene sphere. Thermodynamic measurements based on both theoretical and experimental approaches allowed the acidity of ketone 1 to be measured; this resulted in about a 1 × 10⁶-fold increase in acidity with respect to that of the structurally correlated acetophenone. Experimental and theoretical kinetic determinations were used to determine the stereolability of ketone 1, which was, in fact, chiral due to the stereogenic carbon α to the carbonyl group. Also, these results confirmed the very strong tendency for the C₆₀ sphere to promote breaking of the C-H_α bond. Thismeans that the Pyr=C₆₀ and C ₆₀ fragments are able to express electron-withdrawing effects stronger than those of halogen atoms, but less effective than those of carbonyl derivatives (e.g., ketones, carboxylic acids or acyl derivatives). Finally, the σ_p, σ_I and σ_R ^- descriptors of inductive and/or mesomeric effects were successfully estimated for both Pyr=C₆₀ and C₆₀ fragments thanks to the integrated use of an original procedure based on semi-empirical calculations and Hammett/Taft correlations involving mono- or dual-parameter equations.

Full text of DOI:10.1002/ejoc.201100997

FULL PAPER
DOI: 10.1002/ejoc.201100997
Inductive and Mesomeric Effects of the [60]Fulleropyrrolidine Fragment and
[60]Fullerene Sphere: A Quantitative Evaluation Based on Theory and
Experiments
Alessia Ciogli,[a] Antonella Fontana,[b] Francesco Gasparrini,[a] Michele Maggini,[c]
Mario Giovannoli,[a] Marco Pierini,*[a] Filippo Busolo,[c] Gabriella Siani,[b] and
Claudio Villani[a]
Keywords: Fullerenes / Kinetics / Thermodynamics / Linear free energy relationships
Inductive and mesomeric effects of the [60]fulleropyrrolidine
(Pyr=C60) and the [60]fullerene (C60) molecular frameworks
have been investigated by linear free energy relationship
analyses. The electronic effects of these moieties have been
studied by expressly designing [60]fulleropyrrolidine deriva-
tive 1, which has a C–H bond α to both a carbonyl group (C–
Hα) and a C60 cage. The extent of polarization of such a bond
was then used to probe electron-withdrawing perturbation
induced by the fullerene sphere. Thermodynamic measure-
ments based on both theoretical and experimental ap-
proaches allowed the acidity of ketone 1 to be measured; this
resulted in about a 1ϫ106-fold increase in acidity with re-
spect to that of the structurally correlated acetophenone. Ex-
perimental and theoretical kinetic determinations were used
to determine the stereolability of ketone 1, which was, in fact,
chiral due to the stereogenic carbon α to the carbonyl group.
Also, these results confirmed the very strong tendency for the
C60 sphere to promote breaking of the C–Hα bond. This
means that the Pyr=C60 and C60 fragments are able to ex-
press electron-withdrawing effects stronger than those of
halogen atoms, but less effective than those of carbonyl de-
rivatives (e.g., ketones, carboxylic acids or acyl derivatives).
–
Finally, the σp, σI and σR descriptors of inductive and/or
mesomeric effects were successfully estimated for both
Pyr=C60 and C60 fragments thanks to the integrated use of
an original procedure based on semi-empirical calculations
and Hammett/Taft correlations involving mono- or dual-
parameter equations.
been undertaken to characterize C60 as a substituent. Gen-
erally, it is well known that the C60 sphere is able to express
electron-withdrawing properties as both a fragment coval-
ently linked to a suitable molecular derivative[3] and as a
guest involved in supramolecular systems.[4] In particular,
these effects were studied for the basic properties and nu-
cleophilicity of a fulleropyrrolidine derivative,[3] as well as
for the acidity of a benzoic acid para-substituted with a
cyclopropylfullerene, which allowed a σp value for the C60
sphere to be assessed.[3] However, although of significant
importance, the quantitative evaluations reported were
based on a C60 cage separated from the reaction site by
more than a single σ bond; therefore, affording information
that was probably not solely as a result of the pure C60
fragment. Herein, we focused our attention on the possibil-
ity of overcoming this limitation by resorting to a carefully
designed model molecule that incorporated a C60 sphere.
We obtained the required information through an inte-
grated experimental and theoretical approach.
Introduction
Fullerene derivatives have a number of attractive features
that make them well suited for a variety of applications,
ranging from new materials[1] to nanomedicine.[2] Mono-
functionalized [60]fullerenes are easily obtained by repro-
ducible synthetic procedures and are amenable to thorough
physico-chemical characterization, particularly in view of
their higher solubility in common solvents than that of pris-
tine [60]fullerene (C60). Functional fragments introduced
onto the C60 spheroid alter its properties and the induced
perturbations have been studied and understood for a large
number of substituents. However, the effects of the C60 frag-
ment on the physico-chemical properties of a molecular
framework are less well known and no systematic study has
[a] Dipartimento di Chimica e Tecnologie del Farmaco,
Università “La Sapienza”,
00185 Roma, Italy
Fax: +39-064-9912780
E-mail: marco.pierini@uniroma1.it
[b] Dipartimento di Scienze del Farmaco,
Università “G. d’Annunzio”
66013 Chieti, Italy
Results and Discussion
[c] Dipartimento di Chimica Organica, Università di Padova,
35131 Padova, Italy
In the first step of our study, we focused on the design of
a [60]fullerene derivative endowed with structural properties
Supporting information for this article is available on the
WWW under http://dx.doi.org/10.1002/ejoc.201100997.
Eur. J. Org. Chem. 2012, 193–202
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
193

M. Pierini et al.
FULL PAPER
suitable to allow effective and quantitative analyses of the yielding the same kind of information, might be based on
electron-withdrawing ability of the C60 sphere. As a funda- the stereogenic feature of the carbon atom α to the C=O
mental requirement, it was ensured that a structure was group (Cα), which renders 1 a chiral species, the stereo-
present in the molecular framework able to act as a sensible chemical stability of which is directly linked to C–Hα acid-
probe of the fullerene effect. Electron-donating or -with- ity. In fact, proton abstraction from Cα would necessarily
drawing effects of molecular moieties substituted for suit- lead to a loss of the configuration of this centre, which
able species can be properly monitored through either ki- could be inverted or restored by subsequent reprotonation.
netic or thermodynamic measurements (e.g., values of Within an established solvent and base catalyst, the rate
either rate constants of substitution/isomerization reactions constant governing this process (i.e., the enantiomerization
or equilibrium constants of isomerization/acid–base reac- rate constant of 1, ke) will be strictly related to the acidity
tions, respectively), providing that their relative position of Hα,[5] and therefore, also to the electron-withdrawing
within the structure may directly affect the sensitivity of the properties of C60. Thus, measuring ke might also provide a
reaction site by inductive and/or mesomeric perturbation. decisive contribution toward achieving our goal. However,
For this purpose, we considered it particularly attractive to before starting the non-trivial synthesis of compound 1, we
correlate the electron-withdrawing properties of a fullerene searched, by assessing acidity, for prior theoretical confirm-
sphere with the acid strength of a C–H bond α to a keto ation of the suitability of such a ketone to be an effective
group and also in close proximity to the C60 fragment. Tak- model in this study. Initially, we faced the problem of the
ing into account the well-established, effective procedure unsuitability of common commercial or free computer soft-
available for the synthesis of fulleropyrrolidines through ware, such as the well-known packages ACD/pKa,[6a]
1,3-dipolar cycloaddition,[3,4] this molecular framework was Marvin,[6b] Mo-Ka,[6c] or the online computer program
the structural unit conceived to host the required keto SPARC.[6d] Indeed, these software estimate pKa values on
group. Accordingly, compound 1 was the fullerene deriva- the basis of extensive experimental parameterization of the
tive designed for our purposes (Scheme 1).
chemical/physical effects related to a large number of mo-
lecular moieties typically encountered within chemical or-
ganic species[6e] (in ACD/pKa, however, the pKa estimation
of carboxylic acids is not available). Unfortunately, the C60
sphere was not among the parameterized groups. For this
reason, we were forced to perform the pKa estimation by
elaborating an empirical approach specifically for monoke-
tones and based on multi-parameter Equation (1) contain-
ing density charge values (δi) as molecular descriptors of
acidity, which are very effective for this purpose.[5a,5b,7]
pKa = ΣCiδi + C0
(1)
The δi values were calculated by a semi-empirical procedure
(Hamiltonian AM1) on both the Hα and oxygen atoms of
the neutral ketone form and the oxygen atom of the enolate
ion of the monoketones used as empirical references (see
Scheme 2), the acidity values of which in water (pKa) were
available from the literature (see Table S1 in the Supporting
Information).
Scheme 1. Compounds involved in the present study. TEG: triethy-
lene glycol.
In this species the hydrogen α to the C=O group (Hα) is
also α to the fullerene sphere. It is therefore intended that
the extent of polarization of the C–Hα bond may act as a
probe of the electron-withdrawing properties of C60, which,
in turn, should be reflected in the acidity of Hα. Thus, mea-
Scheme 2. Multi-parameter equation and molecular descriptors
used within the theoretical estimation of pKa values of monocar-
bonyl ketones.
On the basis of statistical analysis, the 3 descriptors in
surement of the acidity of compound 1 (i.e., the pKa values Equation (1) were selected from a set of 14 candidates (see
in water or other solvents) compared with that of other suit- the Supporting Information), whereas parameters C0 and
ably substituted ketones should afford the required infor- Ci were optimized by linear regression analysis of a set of
mation. In addition, an independent approach, capable of equations related to 42 (n = 42) pKa data points (Table S1
194
www.eurjoc.org
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2012, 193–202

Inductive and Mesomeric Effects of [60]Fulleropyrrolidine
and Scheme S1 in the Supporting Information). The final was isolated in 38% yield through the cycloaddition of
statistical significance of each δi descriptor was checked by [60]fullerene with the azomethine ylide formed when N-
a T test. The exclusion of randomness of the obtained re- methylglycine and phenylglyoxal were heated to reflux in
sults was checked by an F test. After such an analysis (for chlorobenzene for 2 h. HPLC analysis was employed to as-
more details, see the Exp. Section), we obtained Equa- sess the purity of 1 (Figure S1 in the Supporting Infor-
tion (2).
mation), whereas NMR spectroscopy (Figures S2 and S3 in
the Supporting Information) and mass spectrometry (Fig-
ure S4 in the Supporting Information) were used to validate
the proposed molecular structure. Further details are re-
ported in the Exp. Section.
pKa = –148.133δHα – 28.7157δO1 – 21.8184δO2 + 13.3510
(2)
with n = 42, R2 = 0.950, and standard deviation (SD) =
0.9. All three employed descriptors were highly significant
and any reasonable possibility of casualness was excluded
To achieve preliminary experimental evidence of the abil-
ity of C60 to promote polarization of the C–Hα bond in 1,
we studied the tendency of this ketone to enantiomerize. In
fact, for the monoketones, there was a clear linear relation-
ship between enolization rate constant (kt) and the acidity
of the C–Hα moiety.[5a] Thus, due to the stereogenic nature
of the Cα atom in 1, the propensity of this ketone to enolize
will also give a direct measure of its stereolability, according
to the relationship that exists between enantiomerization
and enolization rate constants: kt = 2ke.[5a,9,10] For this pur-
pose, the kinetic determination of the enantiomerization
rate constant of 1 was performed by monitoring the racemi-
zation process of the pure enantiomers [enantiomeric excess
(ee) values greater than 99.8%], previously separated by
enantioselective HPLC at the semi-preparative level using a
bonded polysaccharide-based stationary phase in a thermo-
statted reactor (Figure 1). Stereoisomerization was per-
formed at 50 °C with 1,8-diazabicyclo[5.4.0]undec-7-ene
(DBU) as the basic catalyst employed at different concen-
trations in three different media: pure chloroform, mix1 =
hexane/2-propanol 90:10 (v/v) and mix2 = hexane/2-propa-
nol/chloroform 85:10:5 (v/v/v).
(F = 7ϫ10–26). By inserting the density charge values δHα
,
δO1 and δO2 calculated on the sensitive Hα, O1, and O2
atoms within the optimized structure of 1 (see Scheme 2)
into Equation (2), we obtained a pKa value of 11.9Ϯ0.9 for
this fullero ketone. A first rough comparison between this
result and the typical acidity expressed by simple alkyl
ketones, in particular, by the more strictly correlated aceto-
phenone (Scheme 1 and Table 1), suggests an impressive
ability of the fulleropyrrolidine molecular framework to
promote C–Hα bond polarization, which, for example,
would lead to an ionization equilibrium constant about
1ϫ106-fold higher than that of acetophenone.
Table 1. Co-logarithms/logarithms of the ionization constants (Ka)
and second-order enolization rate constants promoted by OH– (kt)
of the carbonyl derivatives in water analyzed in this study.
pKa
logk[e]
pKa
logk[e]
pKa
logk
10.1[a]
10.1[b]
11.9[c]
17.1[c]
17.7[d]
11.3[e]
18.3[f]
19.3[f]
20.5[g]
15.9[f]
3.07[i]
4.36[j]
8
9
13.3[f]
1.91[f] 15
–
2.11[g]
1
2
–0.41[k]
14.8[g]
0.55[g] 16
21.0[g] –2.65[g]
The progress of the racemization process was monitored
by resolving the enantiomers of 1 through enantioselective
HPLC and then integrating the areas of the relevant peaks.
Measurements of pseudo-first-order enantiomerization rate
3
4
5
6
7
3.20[e]
–1.08[f]
–1.14[f]
–1.72[g]
0.20[f]
10
11
8.9[g]
5.1[h]
4.30[g] 17
8.09[l] 18
8.9[g]
5.2[h]
3.24[f]
7.51[h]
2.03[g]
–
12 19.0[g] –1.72[g] 19
12.7[g]
7.0[g]
13 10.7[g]
14 25.0
3.53[g] 20
21
–
11.0[g]
–
1
constants ke were then carried out by fitting a plot of ee
[a] Obtained by titration in the presence of cetyltrimethylammo-
nium bromide (CTAB). [b] Obtained by titration in the presence of
Triton X-100 (TX-l00). [c] Assessed by Equation (3). [d] Assessed
by Marvin software.[6b] [e] From ref.[5a] [f] From ref.[5b] [g] From
ref.[8] [h] From ref.[7a] [i] Assessed by Equation (3) with pKa = 10.
[j] Assessed by Equation (3) with pKa = 12. [k] Assessed by Equa-
tion (3) with pKa = 17.4. [l] Assessed by the difference between the
logkt values from compounds 6 and 18 and the sum with logkt of
5.
values as a function of time through the equation ln(ee)
= (–21ke)t. Eventually, second-order enantiomerization rate
constants, ke, were obtained from the slope of linear plots
1
of ke versus [DBU]. The obtained results, given in Table 2,
clearly indicated the sensitivity of 1 to changes in configura-
tion under quite mild conditions; this would be expected
for strongly activated polarization of the C–Hα bond.
For comparison, the enolization rate constant of 1 in
To gain a more direct indication of the effect of only the mix1 was only twofold smaller than that calculated under
C60 sphere, we estimated the pKa value of ketone 2 the same solvent and temperature conditions for ketone 3,
[Scheme 1, pKa = 17.1 and 17.7 from Equation (2) and shown in Scheme 1, the acidity of which in water was re-
Marvin software, respectively; average pKa = 17.4], the ported to be 11.3 pKa units.[5a] Therefore, this finding shows
structure of which corresponds to that of 1 without the C60 the reliability, within an estimated error of 1 pKa unit, of
moiety. Inspection of Table 1, shows that ΔpKa between the obtained pKa value of 11.9. Nevertheless, we considered
ketones 1 and 2 amounts to about 6 units, that is, an acidity it essential to confirm further this value through experimen-
gap larger than that produced by monochlorine or phenyl tal pKa determination. To perform this measurement, we
substitution on acetone, but also very similar to that of faced the non-trivial problem of the scant solubility of pris-
monobromination of the same reference ketone. Therefore, tine and mono-substituted fullerenes in water. It is well
these theoretical findings confirmed the suitability of com- known that strong hydrophobic interactions drive the ful-
pound 1 to act as an effective model for our purposes and lerene cages to form colloidal clusters in different hydrophi-
induced us to carry out its synthesis. Fulleropyrrolidine 1 lic media.[12] However, the water solubility of pristine
Eur. J. Org. Chem. 2012, 193–202
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
195

M. Pierini et al.
FULL PAPER
It has recently been demonstrated that the proton trans-
fer reaction takes place within an enolate–water cluster in
the head polar region of the micelle, where the reacting
partners actually experience environmental conditions, for
example, micropolarity effects, very similar to those of bulk
water.[16,17] These characteristics enable us to consider the
micellar environments described as suitable for our intents.
In fact, the solubility of 1 in an aqueous micellar solution
of CTAB or TX-100 allowed us to study its acid properties
in water. Consequently, the acid dissociation constant, Ka,
was measured potentiometrically by titrating 1 with stan-
dardized NaOH. The obtained pKa values were
10.06Ϯ0.03 and 10.1Ϯ0.5 in the presence of CTAB and
TX-100, respectively (both values are averages of two deter-
minations). These figures are impressive because they sug-
gest an acidity for 1 even stronger than that assessed by
calculation. However, the effect produced by the micelles
on the ionization reaction has to be taken into account and
some considerations deserve to be highlighted.
Previous studies on the keto–enol interconversion of
some heterocyclic ketones have demonstrated that the ap-
parent acidities of enols and ketones increase on passing
from water to surfactant solutions.[18–20] As a relevant ex-
water
ample, the ΔpKa (pKa
– pKamicellar solution) of 2-phenyl-
acetylfuran is 2.3 in the presence of the cationic CTAB.[19]
The increase of acidity upon micellization has been ex-
plained in terms of electrostatic interaction between the
enolate and the CTAB micelle in which the enolate is stabi-
lized. However, an acidity increase has also been observed
in the presence of a neutral zwitterionic surfactant (ΔpKa =
1.1).[19] These results suggest that other effects must also
play a role in the observed increase of acidity upon micelliz-
Figure 1. (a) Enantioselective HPLC traces of enantiomers a and b
of compound 1 detected by UV (top left) and circular dichroism
(CD; top right); analytical control of the purified enantiomers
(traces revealed by UV detector on middle and bottom left; online
CD spectra of the first and second eluted enantiomers on middle
right). (b) Racemization trend in solvent mix1. (c) Racemization ation and that thermodynamic data of ionization achieved
trend in solvent mix2. For the analytical conditions used, see the
Exp. Sect.
through such an approach should be considered as the
lower limit of a pKa range that might extend its extremes
up to about 1 or 2.3 pKa units, depending on the nature of
the surfactant and the structure of the analyzed substrate.
Then, it seemed reasonable to assume that the experimental
Table 2. Second-order enantiomerization rate constants of ketones
acidity of ketone 1 lay between the pKa values 10.1 and 12.4
from the determination with CTAB and between 10.1 and
11.2 from the determination with Triton; these values were
in good agreement, by associated determination errors, with
experimental results and theoretical estimations. From the
largest and smallest possible values of acidity of 1 (i.e., the
extreme pKa values of 10 and 12), the relevant enolization
rate constants, kt, in water for the process catalyzed by the
OH– ions at 25 °C may be estimated through Equation (3),
which was reported within ref.[5a] in which p is the number
of acidic hydrogen atoms considered for statistical correc-
1 and 3 in organic media at 50 °C.
Solvent
Basic catalyst
logke
1
1
1
3
mix1
mix2
CHCl3
mix1
DBU
DBU
DBU
DMA
–2.83
–2.98
–3.34
–2.49[a]
[a] Extrapolated to 50 °C by a straight line regression (R2 = 0.983)
from four relevant kinetic data at temperatures of 30, 35, 40, 45 °C
(ref.[11]).
[60]fullerene,[13] higher fullerenes[14] and some fullerene de- tion.
rivatives[15] has been successfully promoted by the addition
log(ke/p) = 0.6434 log(Ka/p) + 10.79
(3)
of surfactants, such as TX-100, BrIJ 35, sodium dodecyl
sulfate (SDS) and cetyltrimethylammonium chloride. In the
The data obtained, reported in Table 3, can then be com-
micellar system, the hydrophobic fullerene cage is expected pared with the enolization kinetic data related to several
to be incorporated in the micellar core, while the polar other ketones mono- or disubstituted at the Cα atom with
groups are located close to the aqueous phase. In this sys- different functional groups. To dissect the contribution of
tem, the formation of clusters is prevented and ensures the the elected substituents from all of the possible contri-
dissolution of fullerene monomers.[15]
butions affecting kЈ, we considered the differences given by
196
www.eurjoc.org
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2012, 193–202

Inductive and Mesomeric Effects of [60]Fulleropyrrolidine
Equation (4) in which i refers to the generic ith group weaker electron-withdrawing effect than moieties involving
ranked in Table 3, whereas ref. refers to the relevant ketone carbonyl as the framework directly linked to the Cα atom
with hydrogen instead of the i group.
(i.e., –COCH3 or –COOEt). The same conclusions may also
be stated when considering the fulleropyrrolidine frame-
work as a whole substituting group, with the only difference
being a more marked effect. This evidence leads to possible
reconsideration of the small electron-withdrawing ability at-
tributed to the C60 sphere in the research work reported in
ref.[3a] and quantified as the σp Hammett parameter (σp =
0.06). In such a study, in fact, the fullerene cage incorpo-
rated into the chemical species used for the evaluation (i.e.,
compound 22 of Scheme 1) is linked to the para position of
a benzoic acid through a carbon bridge, so that its effect in
the ionization process of the benzoic acid should be more
properly referred to as the cyclopropylfullerene molecular
framework (hereafter symbolized as CϽC60). Thus, the
small σp value measured for this composite moiety might be
explained by the well-known strong dependence that relates
inductive effects to the number of sigma bonds separating
the group from the reaction site or from a site directly con-
jugated to the centre of reaction (in this case, two σ bonds
between the para position of the benzoic acid and the C60
sphere). In addition, it has also to be stressed that the elec-
tron-withdrawing effect expressed by σp is inclusive of both
inductive and resonance factors and that the latter could
play a non-negligible part in the process, even in opposition
to inductive perturbation. Therefore, we determined if it
would be possible to assess a σp value at the C60 cage start-
ing from the kinetic datum of stereolability achieved for the
fulleroketone 1, which features a direct link from the C60
group to the reaction site (only a σ bond between Cα and
C60). For this purpose, we plotted Δlog(kt) values related to
10 mono- and 4 disubstituted ketones against the relevant
σp values available in the literature (see Table 3).[21] Linear
correlation analysis afforded quite good agreement among
the data, with an R2 coefficient of 0.911, see Equation (5).
Δlog(kt) = log(kt)i – log(kt)ref.
(4)
Table 3. Kinetic data used in the Hammett and Taft correlations
[Equations (5) and (7)] of linear free energy relationship (LFER)
values and an estimation of σp values.
Group C [a] Pe
σI
σR
σp
σm
σp
Pc
[b]
–
[c]
[d]
(ref.)
H
5
0.00
0
0
0
0
–0.09 0.21
(5)
6
(5)
7
(5)
8
(5)
9
CH3
Ph
Br
–0.58 –0.04 –0.11
–0.17 –0.07 –0.15 –0.82
–0.01 0.06 0.04 1.25
0.23 0.39 0.21 2.81
0.23 0.37 0.07 2.32
1.34 0.1
3.05 0.5
0.04
–0.19
Cl
1.69 0.47 –0.23
5.44 0.28 0.47
9.23 0.76 0.47
–0.58 0.27 –0.45
(5)
COCH3 10
(5)
0.5
0.38 0.44 5.48
NO2
11
(6)
0.78 0.71 0.81 9.18
–0.27 0.12 –0.15 –0.68
0.45 0.37 0.36 4.77
0.09 0.05 0.23
OCH3 12
(5)
COOEt 13
4.67 0.3
3.25
0.34
(5)
–
SO3
15
(5)
16
(5)
17
(5)
18
(5)
19
CH3
CH3
Cl
–1.51 –0.08 –0.22
4.38 0.94 –0.46
8.65 0.72 0.36
3.17 0.26 0.23
3.48 0.40 0.03[e]
–0.34 –0.14 –0.24 –1.86
0.46 0.74 0.34 4.42
0.61 0.64 0.75 8.14
0.28 0.02 0.22 3.74
Cl
CH3
NO2
CH3
COOEt (5)
C60
1
(2)
1
(4)
2
0.25 3.16[f]
0.37 4.46[f]
0.31 3.74[f]
0.44 5.04[f]
–0.03 0.53[f]
4.77
4.15 0.47 0.05[e]
5.44
0.24[e]
0.67 0.03 0.04[e]
0.22[e]
Pyr=C60
Pyr
σp = 0.0978Δlog(kt) – 0.0918
(5)
(4)
When analysis was attempted with σm instead of the σp
parameters, the regression afforded a correlation of unac-
ceptable quality (R2 = 0.689), indicating that resonance
contributions were involved in the enolization process of
the ketones analyzed during enolate formation. Equation
(5) was then used to estimate the σp values of all functional
groups employed within the relevant correlation, as well as
those of [60]fulleropyrrolidine (Pyr=C60, σp = 0.31–0.44),
[a] C: carbonyl derivative; ref.: carbonyl derivative used as the refer-
ence. [b] Pe: experimental log(keC/keref). [c] σp parameter calculated
from Equation (5). [d] Pc: log(keC/keref) calculated from Equa-
tion (7). [e] Values that minimize the difference between experimen-
tal and calculated Δlog(kt). [f] Values calculated by inserting the
σR data averaged from those obtained by kinetic (Table 3) and
thermodynamic Taft correlations into Equation (7) (see below in
Table 5).
–
pyrrolidine (Pyr, σp = –0.03) and [60]fullerene (C60, σp
=
In order to remove the contribution of the C60 sphere from 0.25–0.37) molecular frameworks (Table 3). The standard
that expressed by the whole fullero-pyrrolidine framework, error was 0.1 σp units. Thus, by comparison of the [60]ful-
[3]
the enolization rate constant of ketone 2 was assessed by lerene effect with that reported for the framework CϽC60
Equation (3), using average pKa = 17.4 estimated by both it was possible to evaluate the reduction of electron-with-
Equation (1) and Marvin software (see above) for this pur- drawing ability triggered by insertion of a sigma bond be-
pose.
tween C60 and the reaction site (about 0.2–0.3 σp units).
Inspection of Table 3 shows that the C60 cage demon- Intrigued by the good findings achieved through the use of
strates an electron-withdrawing effect stronger than that re- the descriptor σp, but not σm, we were encouraged to verify
lated to halogen atoms (chlorine and bromine) and groups if it was possible to separate pure inductive effects from
–
such as phenyl or –SO3 . Additionally, the C60 cage has a those of resonance. Therefore, we repeated the correlation
Eur. J. Org. Chem. 2012, 193–202
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
197

M. Pierini et al.
FULL PAPER
analysis that led to straight line 5 by resorting to following exercise electron-withdrawing or -donating effects through
dual-parameter Equation (6) of the type proposed by the quantity EEX (EEX = electron-withdrawing effect; X
Taft.[22]
distinguishes the functional group).[5a] In such an approach,
the charge density on a hydrogen atom of a methane mole-
cule (δH ) is used as the reference to perform a compari-
–
Δlog(kt) = p1σI + p2σR + p3
(6)
CH4
son with the equivalent quantity calculated when a gem-
hydrogen is substituted with a selected X functional group
(δHX). In the absence of direct conjugation between the X
group and the reaction site, EEX should essentially monitor
purely inductive effects. We evaluated the EEref.X quantities
for nine functional groups, the σI values of which were ho-
mogeneously distributed over a rather wide range from the
weakly electron-donating methyl moiety to the strongly
electron-withdrawing nitro group. The substitution was re-
lated to the structure of three reference molecules: (i) meth-
ane (R1H); (ii) N-methylpyrrolidine (R2H) and (iii) N-
methyl-[60]fullero-pyrrolidine (R3H; see Scheme 3).
This equation can take into account both inductive and
mesomeric effects through the use of the two distinct de-
scriptors σI and σR–, respectively, and modulate their effect
by means of the calculable regression coefficients p1 and p2,
respectively; p3 is a constant term also determined by the
–
regression. By excluding the SO3 group from the analysis
–
(due to the unavailability of the relevant σR value), the
regression afforded a very good correlation [Equation (7)]
with R2 = 0.991 and a standard error of 0.34 Δlog(kt) units
(Figure 2). Thus, Equation (7) might be a useful tool for
testing inductive and resonance effects of C60 or other
groups, providing at least one σ value (either σI or σR–) is
available independently.
Δlog(kt) = 7.713σI + 6.606σR + 0.2111
(7)
–
Scheme 3. Structures used to assess σI values through EEX versus
σI correlations.
Therefore, according to the procedure described in detail
in the Exp. Section, 27 estimations of EEX were performed
for 3 classes of compounds, each of which derived from the
basic structures of R1H (ensemble R1X), R2H (ensemble
R
2X) and R3H (ensemble R3X), respectively. It was found
that σI and EEX values from the ensembles of compounds
1X (i.e., EE1X), R2X (i.e., EE2X) and R3X (i.e., EE3X) corre-
lated in a satisfactory way, according to Equations (8), (9)
and (10).
R
Figure 2. Taft correlations (diamond data points) performed by
using kinetic (a) and thermodynamic data (b) on the carbonyl com-
pounds reported in Tables 3 and 5, respectively. Square and triangle
points: estimation of Δlog(kt) and ΔpKa values for the C60 and
–
Pyr=C60 fragments, respectively, starting from the average σR
σI = 11.632EE1X – 0.0328(R2 = 0.926; SD = 0.08, n = 9)
σI = 11.872EE2X – 0.0625(R2 = 0.943; SD = 0.07, n = 9)
(8)
(9)
(these data refer to the findings obtained by considering the acidity
of 1 as an average of the extremes of the range 10–12 considered
in the text, i.e., pKa = 11).
Recently some of us reported a very simple theoretical
procedure, based on semi-empirical calculations, suitable
for characterizing the ability of a molecular framework to σI = 14.506EE3X – 0.0658(R2 = 0.946; SD = 0.07, n = 9)
(10)
198
www.eurjoc.org
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2012, 193–202

Inductive and Mesomeric Effects of [60]Fulleropyrrolidine
Consequently, the σI values associated with the pyrrolid- analysis indicating a great significance of the descriptors σI
–
ine, fulleropyrrolidine and C60 molecular frameworks have and σR and a reasonable absence of casualness in the ob-
been assessed and are collected in Table 4 by inserting the tained results.
following EEX charge differences into Equation (8):
–
ΔpKa = 13.7925σI + 10.2149σR + 0.59009
(13)
R3H
R1H
R2H
R1H
EEPyr=C60 = δH
– δH ; EEPyr = δH
– δH ; EEC60
R3H
R2H
= δH
– δH
.
Table 5. Thermodynamic data used in the LFER-type Taft corre-
lation [Equation (13)].
Table 4. EEX data employed within the correlations expressed by
Equations (8), (9) and (10) and estimation of σI parameters by
Equation (8).
[a]
[b]
–
[c]
Group
H
D
(ref.) Pe
σI
σR
Pc
5
(5)
6
(5)
7
(5)
8
0.00
0
0
0.59
–1.09
2.38
X
EER1X
EER21X
EER3X
σI[a]
σI[b]
CH3
Ph
–1.3
–0.04
0.1
–0.11
0.04
CH3
H
N(CH3)2
Ph
COCH3
I
Cl
Br
CN
NO2
Pyr
Pyr=C60
C60
0.005
0.000
0.004
0.017
0.036
0.042
0.032
0.042
0.044
0.069
0.006
0.043
0.037
–0.006
0.000
–0.010
0.006
0.023
0.034
0.033
0.036
0.032
0.057
–0.002
0.000
–0.010
0.004
0.018
0.025
0.026
0.029
0.029
0.047
–0.04
0.00
0.06
0.10
0.29
0.39
0.47
0.50
0.53
0.76
–0.02
0.00
–0.08
0.11
0.34
0.45
0.41
0.48
0.47
0.77
0.04
0.47
0.40
3.36
6.00
4.44
Br
0.5
–0.19
–0.23
0.47
5.55
(5)
9
Cl
0.47
0.28
0.76
0.27
0.3
4.72
(5)
10
(5)
11
(6)
12
(5)
13
COCH3
NO2
OCH3
10.4
9.25
14.17
0.23
8.6
0.47
15.87
–0.28
8.20
–0.45
0.34
CO-
OEt
[a] σI values given in ref.[23]. [b] σI values calculated with Equa-
tion (8).
(5)
14
(5)
16
(5)
17
(5)
18
(5)
19
(5)
NH2
–5.8
–1.7
10.4
14.1
6.6
0.12
–0.08
0.92
–0.48
–0.22
–0.46
0.36
–2.66
–2.76
8.58
CH3
CH3
Cl
By inspecting the values given in Table 4, some impor-
tant conclusions may be derived. As far as the inductive
effect is concerned, the calculated σI values qualify both C60
Cl
and Pyr=C60 as among the strongest electron-withdrawing CH3
0.72
14.20
6.53
NO2
moieties considered in this study, with effects very close to
CH3
those associated with halogen atoms and the cyano group
CO-
0.26
0.23
and not too far from that of the alkylsulfonate group (σI =
OEt
0.59). As expected, the Pyr molecular framework is among
the less active ones, very close to the properties of the struc-
tural analogue –N(CH3)2, with which it has a strong struc-
tural analogy. With the σI values in hand, we were now able
to assess the σR parameters relevant to the Pyr=C60, Pyr
and C60 moieties by inserting their calculated σI values into Pyr=C60
Equation (7) and solving it with respect to the unknown
term. The achieved values, reported in Table 3, indicated
marginal resonance effects or C60; therefore, it seems rea-
sonable to infer that the estimated σP parameter should also
include minimal resonance contributions. All of the ob-
tained findings were confirmed by repeating the Taft analy-
sis on Ka data, according to the Equations (11) and (12).
COCH3
Br
COCH3
CH3
20
(5)
21
(5)
1
(2)
1
(4)
2
(4)
12.3
8.27
0.72
0.24
0.40
0.47
0.03
0.28
0.36
13.38
7.58
C60
5.40
7.40
6.30
8.30
0.90
–0.07[d]
0.13[d]
5.89[e]
7.91[e]
6.93[e]
8.93[e]
1.13[e]
–
–0.08[d]
0.12[d]
Pyr
–0.01[d]
[a] D: carbonyl derivative; ref.: carbonyl derivative used as the refer-
ence. [b] Pe: experimental Ϫlog(KaD/Karef.). [c] Pc: –log(KaD/Ka
)
ref.
calculated by using Equation (13). [d] Values that minimize the
difference between experimental and calculated ΔpKa. [e] Values
calculated by inserting σR data averaged from those obtained by
kinetic (Table 3) and thermodynamic (Table 5) Taft correlations
into Equation (13) .
–
ΔpKa = (pKa)i – (pKa)ref.
(11)
Note, the quality of the correlation increases to R2
0.977 if the datum related to the –NH2 group (which is
=
–
ΔpKa = v1σI + v2σR + v3
(12)
The substrates selected for analysis (see Table 5) were characterized by a relatively high uncertainty) is excluded
many of the mono- and disubstituted monocarbonyl com- from the regression.
pounds already considered within the kinetic study that led
to Equation (7) and for which experimental data for acidity
When the values of σI and σR– assessed for Pyr=C60 and
60 by kinetic analysis were inserted into Equation (13), the
C
in water were available in the literature. Also, in this case, relevant calculated values of ΔpKa resulted in good agree-
regression analysis afforded a good correlation coefficient ment with experimental values, within an error less, in the
[R2 = 0.960, Figure 2 and Equation (13)], with the statistical worst case, than 20% for Pyr=C60 and 19% for C60. This
Eur. J. Org. Chem. 2012, 193–202
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
199

M. Pierini et al.
FULL PAPER
–
chased from Bucky USA (99%). The reaction to prepare fullero-
pyrrolidine 1 was monitored by TLC using Merck pre-coated silica
gel 60-F254. NMR spectra were recorded in CDCl3 on a Bruker AC
250 spectrometer at room temperature. Chemical shifts are given in
parts per million (δ) relative to tetramethylsilane. UV/Vis spectra
were recorded on a Perkin–Elmer Lambda 45 spectrophotometer.
FTIR characterization was carried out on a Nicolet 5700 FTIR
spectrophotometer. ESI-MS spectra were recorded with a MSD SL
Trap mass spectrometer (Agilent Technologies): derivative 1 was
dissolved in methanol in the presence of 0.1% (v/v) formic acid.
clearly indicates good reliability of the assessed σI and σR
parameters, especially those related to the purely inductive
effects. As a refinement, by inserting only the σI data into
Equation (13), for both Pyr=C60 and C60, we recalculated
the relevant σR values as those that minimized ΔpKa.
Eventually, the averaged σR values from the kinetic and
thermodynamic Taft analyses were also included in Equa-
tions (7) and (13) and the new estimations of Δlog(kt) and
ΔpKa were compared with those of related experimental
–
–
data. The final errors were equal to or less than 10% for HPLC analyses were performed on a Shimadzu LC10 ADVP bi-
nary pump system equipped with a Jasco 875 detector at 340 nm.
Derivative 1 was eluted on a Nacalai Tesque Cosmosyl Buckyprep
analytical column (250ϫ4.6 mm, eluent: toluene, flow:
1 mLminϪ1, retention time for 1: 5.01 min).
Pyr=C60 and 9% for C60 from both the kinetic and thermo-
dynamic data (Table 6).
Table 6. Estimation of σR– parameters from Equations (7) and (13).
%E[b] σR
%E[d] σR
%E[f] %E[g]
–[a]
–[c]
–[e]
σI
0.40 0.028 –18 –0.070
0.223 –13 0.130
Pyr=C60 0.47 0.047 –20 –0.075
0.243 –15 0.120
0.03 0.035 –51 –0.010
σR
Synthesis of Fulleropyrrolidine 1: [60]Fullerene (150 mg,
0.21 mmol), N-methylglycine (64 mg, 0.72 mmol) and phenylgly-
oxal (150 mg, 1.12 mmol) were dissolved in chlorobenzene (50 mL)
and brought to reflux temperature. The reaction was monitored by
TLC [SiO2, toluene, Rf (1) = 0.4] and stopped just after the forma-
tion of traces of multiple adducts to [60]fullerene detectable by TLC
with a lower retention factor than that of the pyrrolidine monoad-
duct. The crude mixture, loaded on top of a SiO2 flash column
chromatography, was eluted with toluene to remove unreacted [60]
fullerene and then the product was isolated. The fractions contain-
ing fulleropyrrolidine 1 were concentrated at reduced pressure to a
small volume and transferred to a centrifuge tube. The product was
precipitated from acetonitrile, washed twice with the same solvent
and dried under reduced pressure to give 1 as a brownish powder
(71 mg, 38%). 1H NMR (250 MHz, CDCl3, 25 °C): δ = 2.9 (s, 3
H), 4.3 (d, 1 H), 5.0 (d, 1 H), 5.6 (s, 1 H), 7.4–7.6 (m, 5 H) ppm.
13C NMR (62.9 MHz, CDCl3/CS2, 25 °C): δ = 40.0, 69.2, 70.1,
73.8, 81.6, 128.3, 128.8, 129.0, 129.7, 133.9, 135.5 136.0, 136.7,
137.1, 139.4, 139.5, 140.3, 140.4, 141.6, 141.7, 141.8, 142.0, 142.1,
142.2, 142.5, 142.6, 142.7, 143.1, 144.3, 144.5, 144.7, 145.2, 145.4,
145.5, 145.6, 145.7, 145.9, 146.0, 146.1, 146.2, 146.3, 147.3, 151.5,
C60
19
13
19
15
44
–0.021
0.177
–0.014
0.182
0.013
9
6
–9
–7
10
7
–10
–8
Pyr
22
–26
[a] Assessed by re-solving Equation (7) with σI values reported in
Table 4. [b] Errors associated with the log(keD/keref.) quantities cal-
culated by using the σR values in Equation (7). [c] Assessed by re-
–
solving Equation (13) with σI values reported in Table 4. [d] Errors
associated with the log(KaD/Karef.) quantities calculated by using
–
–[a]
–[c]
the σR values in Equation (13). [e] Average of the σR and σR
values reported in Table 6. [f] Errors associated with the log(ke
/
D
keref.) quantities calculated by using the σR values in Equation (7).
–
[g] Errors associated with the log(KaD/Karef.) quantities calculated
–
by using the σR values in Equation (13).
Conclusions
Derivative 1 was employed as a suitable model to per-
form a quantitative estimation of the inductive and meso-
meric effects of the C60 cage and the composite fulleropyr-
rolidine molecular framework Pyr=C60. An original theo-
retical approach was specifically used to address the estima-
tion of the pKa value of 1 for which commercial software
cannot be employed, due to the absence of a suitable pa-
rameter for the fullerene fragment. A significant increase
of acidity, with respect to the unsubstituted analogue, was
theoretically assessed and, next, confirmed by experimental
determination of the aqueous pKa value in the presence of
suitable surfactants. The effective electron-withdrawing ef-
fects of the C60 sphere were also evidenced by kinetic mea-
surements of second-order enantiomerization rate constants
of 1. Through LFER analyses and dedicated semi-empirical
calculations such properties were successfully rationalized
by a reliable assessment of Hammett (σp) and Taft (σI and
σR–) parameters for the molecular frameworks Pyr=C60 and
C60. We believe that the obtained results may actively con-
tribute towards more in-depth knowledge of the properties
of [60]fullerene derivatives, especially with respect to their
perspective applications in the field of electronic devices.
152.8, 154.2, 154.6, 195.7 ppm. IR (KBr): ν = 2946, 2848, 2780,
˜
1687, 1594, 1446, 1228, 1215, 767, 688, 527 cm–1. UV/Vis (CH2Cl2)
λmax (ε) 255 (45237), 318 (14650), 430 (2080) nm. ESI-MS
(C70H11NO): m/z 882 [M + H]+. Figures of the relevant IR, NMR,
ESI-MS and UV/Vis spectra are reported in the Supporting Infor-
mation (Figures S1–S6).
Enantiomer Separation: Apparatus: Analytical liquid chromatog-
raphy was performed on a JASCO chromatograph equipped with
a Rheodyne model 7725i 20 μL loop injector, a PU-980 HPLC
pump, a Mod Jasco-975 single-wavelength absorbance detector and
a Jasco Mod 995-CD circular dichroism detector. Chromatographic
data were collected and processed by using Borwin software (Jasco
Europe, Italy). Semi-preparative liquid chromatography was per-
formed on a Waters chromatograph (Waters Associates) equipped
with a Rheodyne model 7012 500 μL loop injector, a UV Spectro-
Monitor 4100 spectrophotometer and a refractive index (Waters
R401 detectors).
Chromatographic Procedures: Enantiomers of 1 were collected by
semi-preparative HPLC using Chiralpak-IA column (250 ϫ 10 mm
I.D.), n-hexane/dichloromethane/2-propanol 60:10:30 as the eluent
(flow rate 4.0 mLminϪ1 and T = 25 °C), UV (λ = 300 nm) and IR
detections. Sample
1 was dissolved in mobile phase (c ≈
15 mgmLϪ1) and each injection was loaded with 200 μL (process
yield ≈ 90%). The enantiomeric excesses were determined by using
the analytical Chiralpak-IA column (250ϫ4.6 mm I.D.), the same
mobile phase employed for semi-preparative mode at flow rate
Experimental Section
General: Phenylglyoxal, N-methylglycine and all solvents were pur-
chased from Aldrich and used as received. [60]Fullerene was pur-
200
www.eurjoc.org
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2012, 193–202

Inductive and Mesomeric Effects of [60]Fulleropyrrolidine
1.00 mLminϪ1 and the UV/CD detection at 300 nm. The chroma-
tographic steps afforded 1-first eluted (25 mg, k = 1.67, ee 99.9%)
and 1-second eluted (29 mg, k = 3.05, α = 1.80, ee 99.9%) isolated
samples.
each conformer generated by rotating the X group around the C–
X bond of an angle-step of 36°, from 0 to 360°, according to the
dedicated procedure implemented in SPARTAN. Generalizing, for
compounds belonging to the classes R1X, R2X and R3X and ranked
in Scheme 3, the hydrogen charge differences EEX were calculated
according to EEref.X = δH
refers to one of the molecules R1H, R2H or R3X employed as the
reference instead of CH4.
Enantiomerization of 1: Solutions of the individual enantiomers of
1 (concentration about 0.3 mgmL–1) were held at constant tem-
perature (50 °C) in mix1, mix2 or CHCl3 and variable amounts
of DBU (concentrations ranging from 1ϫ10–2 to 4ϫ10–2 m) were
added. Samples were withdrawn at fixed time intervals and ana-
lyzed by HPLC on an analytical Chiralpak-IA column
(250ϫ4.6 mm I.D.) with n-hexane/dichloromethane/2-propanol
60:10:30 as the eluent at flow rate 1.00 mLmin–1 and UV/CD de-
tection at 300 nm. The ee values were estimated from the integrated
areas in the chromatograms according to Equations (14) and (15).
ref.X
– δHref., in which the superscript ref.
Regression Analyses: Linear regression procedures based on multi-
parameter equations were performed within the estimation of acid-
ity of monocarbonyl ketones [Equation (2)] and Taft correlation
analyses [Equations (7) and (13)]. The obtained results were ana-
lyzed according to F and T tests to exclude casualness (index F)
and to quantify the statistical weight of each of the two descriptors
(index T and factors ti). Linear regression analyses and F and T
tests were performed by the dedicated functions implemented in
the Microsoft Office Excel® 2003 program. Probability related to
the t-Student distribution in the T test was set to 0.05. In the three
quoted cases, T values of 2.02, 2.22 and 2.16 were calculated. Re-
gressions corresponding to Equations (2), (7) and (13) led to the
following F indexes and ti factors: (i) 1ϫ10–24, 9.49, 3.20, 5.36; (ii)
6ϫ10–11, 23.81, 20.94; and (iii) 8ϫ10–10, 12.41, 10.18.
ee (%) ϫ (R*) = (a – b#)/(a + b#) ϫ 100
(14)
(15)
ee (%) ϫ (S*) = (b – a#)/(a + b#) ϫ 100
in which a and b are the peak areas of the (R*) and (S*) enantio-
mers remaining after enantiomerization, whereas a# and b# are the
de novo enantiomeric peak areas. At 25 °C and in the absence of
basic components in the eluent, the extent of enantiomerization
during the chromatographic analysis was undetectable. According
to the equation ln(ee) = (–21ke)t, pseudo-first-order enantiomeriz-
ation rate constants were calculated from plots of ln(ee) versus
time. Eventually, second-order enantiomerization rate constants ke
were then obtained from the slope of linear plots of 1ke versus
[DBU].
Supporting Information (see footnote on the first page of this arti-
cle): Evaluation of acidity of ketones by semi-empirical calculations
(EASC method); HPLC analyses; NMR, ESI-MS, IR and UV/Vis
spectroscopic data of 1.
Acknowledgments
Experimental pKa Determination of Compound 1: Potentiometric ti-
trations were performed by using a standardized 2ϫ 10–2 moldm–3
NaOH solution as the titrant. TX-100 was used without further
purification. CTAB was purified by crystallization from acetone.
The concentration of the surfactants were kept above the critical
micellar concentration (CMC) values.[24,25] The appropriate
amount of 1 was dissolved in CS2 to obtain a 2ϫ10–3 moldm–3
solution. A 20 mm aqueous surfactant solution was added whilst
stirring and nitrogen gas was purged in the two-phase system until
a limpid solution was obtained.
We wish to thank the Ministero dell’Università e della Ricerca
(MIUR) for financial support (PRIN 2008, project 20085M27SS_
004).
[1] a) F. Wudl, Acc. Chem. Res. 1992, 25, 157–161; b) M. Prato, J.
Mater. Chem. 1997, 7, 1097–1109; c) J.-F. Nierengarten, New
J. Chem. 2004, 28, 1177–1191; d) J. L. Segura, N. Martín,
D. M. Guldi, Chem. Soc. Rev. 2005, 34, 31–47; e) D. M. Guldi,
F. Zerbetto, V. Georgakilas, M. Prato, Acc. Chem. Res. 2005,
38, 38–43; f) J. Roncali, Chem. Soc. Rev. 2005, 34, 483–495; g)
N. Martín, L. Sánchez, M. Á. Herranz, B. Illescas, D. M.
Guldi, Acc. Chem. Res. 2007, 40, 1015–1024; h) T. Nakanishi,
Chem. Commun. 2010, 46, 3425–3436; i) A. M. López, A. Ma-
teo-Alonso, M. Prato, J. Mater. Chem. 2011, 21, 1305–1318; j)
F. Giacalone, N. Martin, Adv. Mater. 2010, 22, 4220–4248; k)
A. M. Lopez, A. Mateo-Alonso, M. Prato, J. Mater. Chem.
2011, 21, 1305–1318; l) R. Po, M. Maggini, N. Camaioni, J.
Phys. Chem. 2010, 26, 1889–1893; m) N. Martin, F. Giacalone,
in: Fullerene Polymers, Wiley-VCH, Weinheim, Germany, 2009.
[2] a) A. W. Jensen, S. R. Wilson, D. I. Schuster, Bioorg. Med.
Chem. 1996, 4, 767–79; b) N. Tagmatarchis, H. Shinohara,
Mini-Rev. Med. Chem. 2001, 1, 339–348; c) S. Bosi, T. Da Ros,
G. Spalluto, M. Prato, Eur. J. Med. Chem. 2003, 38, 913–923;
d) M. Satoh, I. Takayanagi, J. Pharmacol. Sci. 2006, 100, 513–
518; e) R. Partha, L. Jodie, J. L. Conyers, Int. J. Nanomed.
2009, 4, 261–275; f) C. Constantin, M. Neagu, R. M. Ion, M.
Gherghiceanu, C. Stavaru, Nanomedicine 2010, 5, 307–17; g) E.
Nakamura, H. Isobe, Acc. Chem. Res. 2003, 36, 807–815; h)
A. Bianco, T. Da Ros, in: Biological Applications of Fullerenes,
in: Fullerenes – Principles and Applications (Eds.: F. Langa, J.-
F. Nierengarten), Royal Society of Chemistry, Cambridge, UK,
2007; i) P. Witte, F. Beuerle, U. Hartnagel, R. Lebovitz, A. Sav-
ouchkina, S. Sali, D. Guldi, N. Chronakis, A. Hirsch, Org. Bi-
omol. Chem. 2007, 5, 3599–3613; j) G.-F. Liu, M. Filipovic´, I.
Ivanovic´-Burmazovic´, F. Beuerle, P. Witte, A. Hirsch, Angew.
Computational Details: All molecular modelling calculations were
performed with the computer program SPARTAN ’04 (Wave-
function Inc., 18401 Von Karman Avenue, Suite 370 Irvine, CA
92612). Conformational searches, based on the molecular mechan-
ics force field MMFF were performed on the ketones and their
enolate forms used for the regression analysis, leading to Equa-
tion (2) (see the Supporting Information), as well as on the ketone
and enolate forms of ketones 1 and 2. All of the structures inside
an energy window of 3 kcalmol–1 were further optimized by semi-
empirical calculations (Hamiltonian AM1). Coulson’s charge
densities on sensible atoms reported in Scheme 2 were averaged in
terms of the relevant Boltzmann probability (at 25 °C) of each con-
formation. All structures belonging to the classes of compounds
R1X, R2X and R3X were optimized at the AM1 semi-empirical level
and the atomic charges were calculated according to the Coulson
definition. The quantity EEX was expressed through EEX = δHx
δH , in which δHx refers to the charge of a hydrogen atom of a
–
CH4
methane molecule substituted at the carbon by an X group (i.e., of
CH4
a hydrogen gem to the X group), whereas δH
refers to the hydro-
gen charge in the pristine methane molecule. For X groups of non-
spherical symmetry with respect to the C–X bond, the charge den-
sity on the hydrogen atom gem to the X group, δHx, was obtained
by averaging the charge values evaluated on the same atom inside
Eur. J. Org. Chem. 2012, 193–202
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
201

M. Pierini et al.
FULL PAPER
Chem. 2008, 120, 4055; Angew. Chem. Int. Ed. 2008, 47, 3991–
3994.
Kleinoder, J. Gasteiger, J. Chem. Inf. Model. 2006, 46, 2256–
2266.
[3] a) A. Bagno, S. Claeson, M. Maggini, M. L. Martini, M. Prato,
G. Scorrano, Chem. Eur. J. 2002, 8, 1015–1023; b) L. Ech-
egoyen, L. E. Echegoyen, Acc. Chem. Res. 1998, 31, 593–601; c)
A. Hirsch, M. Brettreich, in: Fullerenes Wiley-VCH, Weinheim,
Germany, 2005; d) T. Akasaka, F. Wudl, S. Nagase, in: Chemis-
try of Nanocarbons, Wiley-VCH, Weinheim, Germany, 2010; e)
N. Martin, Chem. Commun. 2006, 2093–2104; f) Y. Matsuo,
A. Iwashita, Y. Abe, C.-Z. Li, K. Matsuo, M. Hashiguchi, E.
Nakamura, J. Am. Chem. Soc. 2008, 130, 15429–15436.
[4] a) T. Haino, M. Yanase, Y. Fukazawa, Angew. Chem. 1997,
109, 288; Angew. Chem. Int. Ed. Engl. 1997, 36, 259–260; b)
R. M. Williams, J. W. Verhoeven, Recueil Trav. Chim. Pays-Bas
1992, 111, 531–532; c) F. Diederich, M. Gomez-Lopez, Chem.
Soc. Rev. 1999, 28, 263–277; d) D. Bonifazi, O. Enger, F. Died-
erich, Chem. Soc. Rev. 2007, 36, 390–414; e) L. Sanchez, N.
Martin, D. M. Guldi, Angew. Chem. 2005, 117, 5508; Angew.
Chem. Int. Ed. 2005, 44, 5374–5382; f) D. Canevet, M. Gallego,
H. Isla, A. de Juan, E. M. Perez, N. Marin, J. Am. Chem. Soc.
2011, 133, 3184–3190; g) F. G. Perez, E. M. Sanchez, N. Mar-
tin, Angew. Chem. 2008, 120, 1110; Angew. Chem. Int. Ed. 2008,
47, 1094–1097.
[5] a) R. Cirilli, R. Costi, R. Di Santo, F. Gasparrini, F. La Torre,
M. Pierini, G. Siani, Chirality 2009, 21, 24–34; b) G. Angelini,
P. De Maria, A. Fontana, M. Pierini, G. Siani, J. Org. Chem.
2007, 72, 4039–4047; c) T. L. Amyes, J. P. Richard, J. Am.
Chem. Soc. 1996, 118, 3129–3141.
[6] a) ACD/Labs Package v. 6.00, (pKa DB Module), Toronto
(Canada), 2002; b) Marvin v. 4.1.1.3, 2007, ChemAxon (http://
www.chemaxon.com); c) Mo-Ka, Molecular Discovery, Middle-
sex, UK, 2009; d) S. Hilal, S. W. Karickhoff, L. A. Carreira,
Quantum Struct. Acta Relat. 1995, 14, 348–355; e) J. Manches-
ter, G. Walkup, O. Rivin, Z. You, J. Chem. Inf. Model. 2010,
50, 565–571.
[8]
[9]
[10]
J. P. Guthrie, Can. J. Chem. 1979, 57, 1177–1185.
V. Schurig, Chirality 2005, 17, S205–S226.
E. L. Eliel, S. H. Wilen, in: Stereochemistry of Carbon Com-
pounds, McGraw-Hill, New York, 1994, p. 426.
R. Cirilli, R. Costi, R. Di Santo, F. La Torre, M. Pierini, G.
Siani, Anal. Chem. 2009, 81, 3560–3570.
a) D. M. Guldi, H. Hungerbühker, K.-D. Asmus, J. Phys.
Chem. 1995, 99, 13487–13493; b) P. De Maria, A. Fontana, C.
Gasbarri, D. Velluto, Soft Matter 2006, 2, 594–602.
a) H. Hungerbühker, D. M. Guldi, K.-D. Asmus, J. Am. Chem.
Soc. 1993, 115, 3386–3387; b) A. Beeby, J. Eastoe, R. K.
Heenan, J. Chem. Soc., Chem. Commun. 1994, 517–518.
D. M. Guldi, J. Phys. Chem. B 1997, 101, 9600–9605.
D. M. Guldi, J. Phys. Chem. A 1997, 101, 3895–3900.
P. De Maria, A. Fontana, G. Cerichelli, J. Chem. Soc. Perkin
Trans. 2 1997, 2329–2334.
P. De Maria, A. Fontana, G. Siani, A. Arcadi, M. Chiarini,
Arkivoc 2002, 11, 17–31.
P. De Maria, A. Fontana, C. Gasbarri, G. Siani, Tetrahedron
2005, 61, 7176–7183.
M. Aschi, R. Zappacosta, P. De Maria, G. Siani, A. Fontana,
A. Amadei, Int. J. Quantum Chem. 2011, 111, 1293–1305.
P. Marracino, A. Amadei, F. Apollonio, G. d’Inzeo, M. Liberti,
A. Di Crescenzo, A. Fontana, R. Zappacosta, M. Aschi, J.
Phys. Chem. B 2011, DOI: 10.1021/jp2018567.
C. Hansch, A. Leo, S. H. Unger, K. H. Kim, D. Xikaitani, E. J.
Lien, J. Med. Chem. 1973, 16, 1207–1216.
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
[24]
[25]
S. Ehrenson, R. T. C. Brownlee, R. W. Taft, Prog. Phys. Org.
Chem. 1973, 10, 1–80.
C. Hansch, A. Leo, in: Substituent Constants for Correlation
Analysis in Chemistry and Biology, Wiley, New York, 1979.
J. H. Fendler, in: Membrane Mimetic Chemistry, 1982, Wiley,
New York, p. 9.
[7] a) R. Ballini, S. Cerritelli, A. Fontana, P. De Maria, M. Pierini,
G. Siani, Eur. J. Org. Chem. 2000, 1641–1646; b) W. Cabri, I.
D’Acquarica, P. Simone, M. Di Iorio, M. Di Mattia, F. Gas-
parrini, F. Giorgi, A. Mazzanti, M. Pierini, M. Quaglia, C.
Villani, J. Org. Chem. 2011, 76, 1751–1758; c) J. Zhang, T.
A. De la Maza, J. L. Parra, Biochem. J. 1994, 303, 907–914.
Received: July 8, 2011
Published Online: November 11, 2011
202
www.eurjoc.org
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2012, 193–202