Computational and Theoretical Chemistry 1053 (2015) 123–129

Contents lists available at ScienceDirect

Computational and Theoretical Chemistry

journal homepage: www.elsevier.com/locate/comptc

NCI analysis of the interaction cation


p in complexes with molecular

bowls derived from fullerene
Iván González-Veloso a, Jorge A. Carrazana-García b, Daniela Josa a, Jesús Rodríguez-Otero a,⇑,
Enrique M. Cabaleiro-Lago b
a
Centro Singular de Investigación en Química Biolóxica e Materiais Moleculares (CIQUS), Universidade de Santiago de Compostela, Rúa Jenaro de la Fuente, s/n 15782 Santiago
de Compostela, Galicia, Spain
b
Departamento de Química Física, Facultad de Ciencias, Universidade de Santiago de Compostela, Campus de Lugo. Av. Alfonso X El Sabio, s/n 27002 Lugo, Galicia, Spain

a r t i c l e i n f o a b s t r a c t

Article history: A systematic study based on the non-covalent interactions (NCI) index was carried out in order to analyze
Received 14 July 2014 the characteristics of the cation


p interaction in complexes formed by alkaline cations (Li+, Na+, K+, Rb+,
Received in revised form 2 October 2014 and Cs+) and molecular bowls of increasing size, going from corannulene, C20H10, to the fully closed buck-
Accepted 9 October 2014
minsterfullerene, C60. For this purpose all the possible inner complexes with the cation on the bottom of
Available online 17 October 2014
the bowl were studied. SCF densities were employed at the B3LYP/6-31 + G⁄ level (LANL2DZ effective core
potential basis set was included for Rb+ and Cs+). NCI plots allowed to visualize the characteristics of the
Keywords:
cation


p interaction; showing the spatial arrangement of the interaction and locating the regions of
Non-covalent interactions
NCI
different interaction strength. Moreover, when the comparison of NCI plot is carried out in an adequate
DFT way, an excellent correlation with the calculated complexation energies was found. Finally, in order to
QTAIM obtain complementary data, QTAIM calculations were carried out for several complexes, locating critical
Cation


p interactions points and bond paths.
Ó 2014 Elsevier B.V. All rights reserved.

1. Introduction aromatic systems [9]. In previous work, it has been shown that
the polarization contribution can be the main responsible of the
Non-covalent interactions with the participation of aromatic stability of a cation


p complex, even when the electrostatic
rings play a crucial role in a great variety of phenomena in physics, contribution is repulsive [10].
chemistry and biochemistry [1–4]. The ubiquity of these non- The so-called buckybowls are aromatic curved systems formed by
covalent interactions points out the need for a deep knowledge joining six and five carbon rings in a similar way as in fullerenes, the
and understanding of the characteristics of the interaction at a five carbon rings introducing the curvature on the delocalized system
microscopic level. [13,14]. The simplest of these bowls, corannulene C20H10, has already
When aromatic molecules participate in a non-covalent interac- been considered in different studies in order to determine its
tion, it usually corresponds to one of the following three types: structure, barrier for inversion and also its capability for coordinating
p


p interaction, XH


p hydrogen bond or ion


p interaction cations [10,13,15–18]. The curvature of corannulene produces
[4,5]. More specifically, cation


p complexes of aromatic different molecular electrostatic potential (MEP) by the convex and
compounds have received great attention in recent years, in part the concave face. The MEP of corannulene is more negative by the
as a result of the role they are believed to play in protein folding, convex face so coordination of cations is mostly favored by this face
neurological signaling, the functioning of certain ion channels [10,16–18]. However, as the size of the bowl grows it becomes more
and other diverse phenomena [3,4,6–8]. A general picture has polarizable, thus increasing induction contribution, and its structure
emerged indicating that the interaction is mainly controlled by gets more closed and similar to that of fullerene. As a consequence of
the electrostatic contribution coming from the interaction of the the reduction on the concave cavity available for locating the cation,
ion and the quadrupole moment of the aromatic system [3,9–12]. the steric repulsion in concave complexes could be important,
However, it has been shown that in cation


p systems the role especially in the case of bulky cations [10,17].
played by polarization can be important, especially in extended In a previous work, complexes formed by alkaline cations and a
series of increasingly large buckybowls derived from fullerene
have been studied focusing in their energetic and structural
⇑ Corresponding author. characteristics [10]. In the present work, the focus is the analysis

http://dx.doi.org/10.1016/j.comptc.2014.10.012
2210-271X/Ó 2014 Elsevier B.V. All rights reserved.
124 I. González-Veloso et al. / Computational and Theoretical Chemistry 1053 (2015) 123–129

of the interaction by means of a topological analysis of the electron its derivatives [21]. In order to obtain information about the inter-
density following the spirit of the Quantum Theory of Atoms in action in the complexes, the plotting of non-covalent interaction
Molecules [19]. This theory, formulated by Bader and co-workers regions was performed using the recently developed NCIPLOT pro-
has proven to be a valuable tool for the qualitative and quantitative gram [20]. This method enables identification of non-covalent
analysis of molecular structure and properties, including the study interactions and it is based on the peaks that appear in the reduced
of cation


p interactions. Using an accurate mathematical density gradient (RDG) at low densities. Plotting an RDG isosurface
implementation the topology of the charge density or other for a low value of RDG, the non-covalent regions clearly appear in
three-dimensional density function, the system under study is ana- the supramolecular complex. These interactions correspond to both
lyzed and various kinds of descriptors can be calculated. The usual favorable and unfavorable interactions. In order to differentiate
analysis of non-covalent interactions by applying QTAIM theory is between them, the sign of the second density Hessian eigenvalue
focused in the location of intermolecular critical bonds and bonds times the density is color-mapped onto the isosurface: blue, green
paths, their characteristics being employed as a tool for gaining an red indicate strongly attractive interactions, weak vdW
insight into the characteristics of the interaction. However, in large interactions, and strongly repulsive interactions, respectively. SCF
systems, QTAIM provides an overwhelming amount of informa- densities were employed; B3LYP/6-31 + G⁄ level was used for all
tion: a large number of critical points, their characteristics, a large the complexes and the effective core potential LANL2DZ basis set
set of bond paths, etc. Therefore, it could be desirable a method was included for Rb+ and Cs+. In order to a get a better visualization
allowing visualizing in a quick and simple way the most relevant of the NCI plots, only the RDG surfaces inside the bowl are shown in
features of the interaction. The non covalent interaction index Figs. 2–9.
(NCI) has been devised with this purpose [20,21]. In the present On the other hand, the topology of the electron density was
work, the NCI approach is applied to a series of complexes formed analyzed according to the AIM theory developed by Bader [19].
by the alkaline cations with a series of buckybowls of increasing For this purpose the Multiwfn program was employed and the dif-
size derived from fullerene [10]. Starting from corannulene, new ferent critical points and paths between them were obtained [22].
layers of atom carbons are added to finally reach fullerene. After The same electron density than that for NCIPLOT was used.
each layer is added, the bowl is properly saturated with the neces- In all cases the wavefunction were obtained with the Gauss-
sary hydrogen atoms. This way, a series of bowls is constructed ian09 program package [23] and the geometry of the complexes
showing significantly different molecular electrostatic potentials, was taken from Ref. [10].
which will reflect in the intensity and characteristics of the inter-
action. Also, as the bowl becomes larger and more closed, steric
effects are expected to appear in complexes formed by the concave 3. Results and discussion
face. It is expected that their relevance in the interaction will be
revealed by the NCI analysis. In order to carry out a systematic comparison, all the complexes
where the alkaline cation is placed inside the bowl (on the bottom
of it) were chosen. Fig. 1 shows the structure of the bowls and that
2. Computational details
of C60, buckminsterfullerene. Figs. 2–7 show the computed NCI
plots for the interaction of the five alkaline cations (Li+, Na+, K+,
The non-covalent interactions (NCI) index identifies interactions
Rb+, Cs+) with P1, P2, P3, P4, P5, and C60, respectively. Fig. 2 shows
in a chemical system solely on the basis of the electron density and
a clear trend running from Li+ to Cs+: as the cation size increases

Fig. 2. NCIPLOT gradient isosurfaces (0.55 a.u.) of P1


X+ (X = Li, Na, K, Rb, Cs)
Fig. 1. Structure of the bowls P1–P5 and C60. complexes. Complexation energies (kcal mol 1) in parentheses taken from Ref. [10].
 I. González-Veloso et al. / Computational and Theoretical Chemistry 1053 (2015) 123–129 125

Fig. 3. NCIPLOT gradient isosurfaces (0.55 a.u.) of P2


X+ (X = Li, Na, K, Rb, Cs)
complexes. Complexation energies (kcal mol 1) in parentheses taken from Ref. [10].

Fig. 5. NCIPLOT gradient isosurfaces (0.55 a.u.) of P4


X+ (X = Li, Na, K, Rb, Cs)
complexes. Complexation energies (kcal mol 1) in parentheses taken from Ref. [10].

attractive interaction) and is transformed gradually to reach an

almost fully green (which indicates a weak interaction) for the
P1


Cs+ complex. If the NCI plots are compared with the values
of the complexation energy (in parentheses) an excellent correla-
tion emerges. Fig. 3 (which corresponds to bowl P2) is very similar
to Fig. 2. Only a couple of minor details can be observed. First, the
RDG surfaces are slightly bigger than those of Fig. 2 (as expected,
since the bowl is bigger, then the polarizable p cloud is bigger).
And second, the surface becomes green sooner: for Rb+ is already
fully green. This happens since interaction with P2 is slightly
weaker than that with P1. The trends follow for complexes with
P3 (Fig. 4): surfaces continue to get bigger and now already for
K+ the fully green is achieved. This fact is emphasized for com-
plexes with P4 (Fig. 5), and even faintest ochre (indicative of a
slight repulsive interaction) is found for the P4


Cs+ complex.
Another significant fact takes place in the P4 complexes: the bulk-
ier cations (K+, Rb+, Cs+) leave the bottom of the bowl, ascending in
order to get a better interaction with the increasing p cloud. This
leads to an annular form of the RDG surfaces. By contrast, the
smaller and polarizing cations, Li+ and Na+, are still anchored to
the bottom of the bowl. Trends continue for complexes with P5
(Fig. 6); surfaces continue to get bigger and the colored ochre
regions continue to grow in P5


Cs+ complex (even incipient ochre
regions already appears in the P5


Rb+ complex). The RDG sur-
faces for the bulkiest cations are close to becoming an inner sphere,
Fig. 4. NCIPLOT gradient isosurfaces (0.55 a.u.) of P3


X+ (X = Li, Na, K, Rb, Cs) while the smallest cations do not lift off the bottom of the bowl.
complexes. Complexation energies (kcal mol 1) in parentheses taken from Ref. [10]. Finally, the situation does not change much for complexes with
C60 (Fig. 7). Practically the only difference is that the RDG surfaces
for the bulkiest cations finally close completely becoming
(and the polarizing power decreases) the RDG surface is getting spherical.
bigger but the interaction is getting weaker. So, for the P1


Li+ It is worth noting that Figs. 2–7 are carried out with the same
complex the surface is clearly blue (which indicates a strongly scale running from 0.025 to +0.025 a.u. If for each figure a specific
126 I. González-Veloso et al. / Computational and Theoretical Chemistry 1053 (2015) 123–129

Fig. 7. NCIPLOT gradient isosurfaces (0.55 a.u.) of C60


X+ (X = Li, Na, K, Rb, Cs)
+
Fig. 6. NCIPLOT gradient isosurfaces (0.55 a.u.) of P5


X (X = Li, Na, K, Rb, Cs) complexes. Complexation energies (kcal mol 1) in parentheses taken from Ref. [10].
complexes. Complexation energies (kcal mol 1) in parentheses taken from Ref. [10].

scale is used, color differences could perhaps be more noticeable.

However, a single scale was chosen so comparison can be made
uniformly in all figures, contrasting the difference between the
behaviors of the different bowls in a correct way. In summary,
Figs. 2–7 show for each bowl how the NCI plots can be useful to
visualize the characteristics of the interaction in the complexes.
For each figure the RDG surfaces show the spatial arrangement of
the interaction, and the color scale allows comparing its strength
in the different regions; being the correlation with the values of
complexation energy very good. These figures show the behavior
for each bowl with the five alkaline cations. Conversely, Fig. 8
shows the NCI plots for all the complexes between Li+ and the
six increasing bowls. Despite trying different color scales, it was
not possible to find one that allows a clear differentiation of the
behavior of the six complexes. Finally an isosurface of 0.65 a.u.
with a scale running from 0.020 to +0.015 was chosen. However,
despite the important difference of complexation energy (almost
20 kcal/mol between the most favorable complex, P1


Li+, and
the least favorable one, C60


Li+) the differences of the NCI plots
are almost negligible: the size of the surface remains constant
and only a slight variation of the edge of the surface takes place
(first becoming greener and then redder); also a very small red area
in the center of the surface appears in the largest bowls. Thus,
unlike in Figs. 2–7, NCI plots are practically insensitive to the var-
iation of complexation energy. However, if all the complexes with Fig. 8. NCIPLOT gradient isosurfaces (0.65 a.u.) of X


Li+ (X = P1, P2, P3, P4, P5, C60)
Cs+ are analyzed together, the NCI plots (see Fig. 9) give rise to good complexes. Complexation energies (kcal mol 1) in parentheses taken from Ref. [10].
results: in this case the change of the bowl progressively leads to a
change of the RDG surfaces, which grow and change their colors. color appears. So, in the biggest bowls there are regions of strong
For P1


Cs+ the surface is mainly blue (strong attractive interac- blue, but countered and overcome by wide red regions with strong
tion) and as the bowl increases the ochre color and then the red repulsive interactions. Therefore, for Li+ complexes NCI plots are
 I. González-Veloso et al. / Computational and Theoretical Chemistry 1053 (2015) 123–129 127

Fig. 9. NCIPLOT gradient isosurfaces (0.55 a.u.) of X


Cs+ (X = P1, P2, P3, P4, P5, C60)
complexes. Complexation energies (kcal mol 1) in parentheses taken from Ref. [10].

almost insensitive to the change of complexation energy, while for

Cs+ complexes NCI plots reflects the change of complexation
energy in a rather suitable way. The behavior of the complexes
with Na+, K+, and Rb+ remains in an intermediate position, which
progressively changes from Li+ to Cs+. Then, why do NCI plots work
well when the same bowl is studied but do not work well when
different bowls with the same cation (except for the heaviest ones) Fig. 10. QTAIM molecular graphs of X


Li+ (X = P1, P2, P3, P4, P5, C60) showing the
are studied? In order to obtain more data and to explain the appar- bond paths (lines) and the different critical points: violet = nuclear critical point,
ent occasional malfunction of the NCI plots a QTAIM analysis was NCP (3, 3), orange = bond critical point, BCP (3,-1), yellow = ring critical point, RCP
carried out for all the complexes between Li+ and the six increasing (3, +1), green = cage critical point, CCP (3, +3). (For interpretation of the references to
colour in this figure legend, the reader is referred to the web version of this article.)
bowls of Fig. 8. The results are shown in Fig. 10. As expected from
Figs. 2–7 and 8, in all cases the Lithium cation remains anchored to
the bottom of the bowl. This is clearly illustrated by the existence
of five bond paths, BPs, between the cation and the five carbon 50%) must be kept in mind. This fact will continue happening when
atoms of the bottom. This five BPs are accompanied by their five P3, P4, . . . are included in the comparison (although progressively
Bond Critical Points, BCPs (orange in Fig. 10). Moreover, there are less dramatically). Obviously this will affect much to the total elec-
five Ring Critical Points, RCPs (yellow) corresponding to each of tron density and is very likely that this is the cause of the apparent
the triangular faces of the pentagonal pyramid that emerges. Con- malfunction of Fig. 8. Thus, the progressive increase of the total
sequently, a Cage Critical Point, CGP (green) appears inside this electron density blurs the progressive decrease of the interaction
pyramid. It is worth noting that all these eleven critical points strength. However, when comparison is established between the
are virtually coplanar. This pyramidal pattern remains unchanged same bowl and different cations (as it is done in Figs. 2–7), this
for all the Li+ complexes (Fig. 10). Obviously, due to symmetry, problem almost vanishes because the difference in total electron
electron densities of the five BCPs of the edges of the pentagonal density is rather modest. Therefore, it seems clear that if an ade-
pyramid are identical for each complex and their values are quate comparison between NCI plots wants to be established then
0.019, 0.018, 0.018, 0.019, 0.020, 0.020 a.u., for the complexes with the difference of total electron density between complexes must be
P1–P5, and C60, respectively. Therefore, these values do not reflect not very large. Unlike complexes with Li+ (Fig. 8), complexes with
the decreasing of complexation energy as the bowl of the size Cs+ (Fig. 9) give rise to reasonable NCI plots, despite the increasing
increases either; even the largest value corresponds to the complex size of the bowls. The difference between both cations is clear: Li+
with the smallest complexation energy. Thus, at least in this case, has 2 electrons while Cs+ is much bulkier and has 54 electrons. This
the electron density at the BCP that binds the cation with the bowl huge size of Cs+ largely mitigates the increase of the total electron
does not correlate well with the strength of the interaction. The density (the relative increase is substantially lower) as the bowl
source of this bad correlation could be the large difference in the grows. This fact, together with the large increase of the complexa-
total electron density of the complexes being compared. When tion energy, results in the reasonable behavior of NCI plots for
complexes P1


Li+ and P2


Li+ are compared, the huge increase complexes with Cs+ (Fig. 9). As expected, for complexes with Na+,
of the number of electrons (from 132 to 192, an increase of nearly K+, and Rb+ the behavior lies between that of Li+ and that of Cs+.
128 I. González-Veloso et al. / Computational and Theoretical Chemistry 1053 (2015) 123–129

For each bowl the reduced density gradient (RDG) surfaces show
the spatial arrangement of the interaction, and the color scale
(corresponding to the second density Hessian eigenvalue) allows
comparing its strength in the different regions. Moreover, the NCI
results correlates very well with the values of complexation
energy.
Only Li+, and to a lesser extent Na+, remain anchored to the
bottom of all bowls. However, the bulkiest cations, as the size of
the bowls increases, quickly leave the bottom and form a large
interaction surface, which finally becomes a full sphere with C60.
As the size of the surface increases, the interaction is increasingly
weak (which is easily visualized by the disappearance of the blue
tone and its replacement by ocher or reddish).
On the other hand, when the comparison of NCI plots is
established between complexes formed by the same cation and
the different increasing bowls, the results are not so good, except
in the case of the bulkiest cations (especially Cs+). So, in these cases
the NCI plots are almost insensitive to the change of complexation
energy. A possible explanation of this problem could be the very
large difference of total electron density between the complexes
compared. As the size of the bowl increases, the total number of
electrons increases very much, and therefore, the comparison of
RDG surfaces becomes worthless. Thus, it seems clear that if an
adequate comparison between NCI plots wants to be established
then the molecular systems being compared cannot have very
different sizes.
The NCI plots, like QTAIM, are based on the electron density and
its derivatives. The NCI index facilitates the visualization and
Fig. 11. QTAIM molecular graphs of X


Cs+ (X = P1, P2, P3, P4, P5, C60) showing the enables identification of non-covalent interactions in a quick and
bond paths (lines) and the different critical points: violet = nuclear critical point,
simple way, which is especially useful in large systems where
NCP (3, 3), orange = bond critical point, BCP (3, 1), yellow = ring critical point,
RCP (3, +1), green = cage critical point, CCP (3, +3). (For interpretation of the QTAIM provides an excessive amount of information, very
references to colour in this figure legend, the reader is referred to the web version of awkward to be analyzed.
this article.)

Acknowledgments
In order to assess the validity of the results obtained at the
B3LYP level, several calculations were performed using a long- The authors want to express their gratitude to the CESGA (Cen-
range corrected version of B3LYP and a substantially larger basis tro de Supercomputación de Galicia) for the use of their computers.
set. For this purpose we have chosen all the complexes with the D.J. thanks the Spanish Ministry of Education for FPU scholarship.
heaviest cation (Cs+) since it is likely to be the case in which the
greatest deviations emerge. These results (see Supportting
Appendix A. Supplementary material
Information) provide no noteworthy novelty with respect to those
already mentioned.
Supplementary data associated with this article can be found, in
Finally, a QTAIM analysis was carried out for all the complexes
the online version, at http://dx.doi.org/10.1016/j.comptc.2014.
between Cs+ and the six increasing bowls of Fig. 9; the results are
10.012.
shown in Fig. 11. Only in the case of the smallest bowl P1, the
behavior is similar to that of Li+ and a pattern with a pentagonal
pyramid is obtained. However, already for the second bowl P2, References
the cation leaves the bottom of the bowl and five BPs are estab-
[1] H.-J. Schneider, A. Yatsimirski, Principles and Methods in Supramolecular
lished with more external regions, specifically with BCPs between Chemistry, Wiley, New York, 2000.
carbon atoms of the third layer (counting from the center). With P3 [2] A.J. Stone, The Theory of Intermolecular Forces, Oxford University Press,
and P4 the cation joins to the most external carbon atoms: in the Oxford, 2013.
[3] A.S. Mahadevi, G.N. Sastry, Cation-p interaction: its role and relevance in
first case directly to the ten external carbon atoms, and in the sec-
chemistry, biology, and material science, Chem. Rev. 113 (2013) 2100–2138.
ond case the five BPs go to the BCPs between the external carbon [4] L.M. Salonen, M. Ellermann, F. Diederich, Aromatic rings in chemical and
atoms. Finally, for the quasi-closed, P5, and closed (C60) bowls, biological recognition: energetics and structures, Angew. Chem. Int. Ed. Engl.
50 (2011) 4808–4842.
the Cs+ cation is placed in the center of the bowl, establishing mul-
[5] S. Tsuzuki, T. Uchimaru, Magnitude and physical origin of intermolecular
tiple connections with the surrounding carbon atoms. Again, and interactions of aromatic molecules: recent progress of computational studies,
this time referring to QTAIM analysis, the behavior of the interme- Curr. Org. Chem. 10 (2006) 745–762.
diate cations, Na+, K+, and Rb+, places in an intermediate position [6] E.A. Meyer, R.K. Castellano, F. Diederich, Interactions with aromatic rings in
chemical and biological recognition, Angew. Chem. Int. Ed. Engl. 42 (2003)
between Li+ and Cs+. 1210–1250.
[7] D.A. Dougherty, Cation-p interactions involving aromatic, J. Nutr. 137 (2007)
1504S–1508S.
4. Conclusions [8] D.A. Dougherty, The cation-p interaction, Acc. Chem. Res. 46 (2013) 885–893.
[9] I. Soteras, M. Orozco, F.J. Luque, Induction effects in metal cation-benzene
According to our results, NCI plots can be a useful tool to complexes, Phys. Chem. Chem. Phys. 10 (2008) 2616–2624.
[10] J.A. Carrazana-Garcia, J. Rodriguez-Otero, E.M. Cabaleiro-Lago, DFT study of the
visualize the characteristics of the interaction cation


p in the interaction between alkaline cations and molecular bowls derived from
complexes formed by alkaline cations and increasing buckybowls. fullerene, J. Phys. Chem. B 115 (2011) 2774–2782.
 I. González-Veloso et al. / Computational and Theoretical Chemistry 1053 (2015) 123–129 129

[11] E.M. Cabaleiro-Lago, J. Rodríguez-Otero, Á. Peña-Gallego, Effect of [20] J. Contreras-García, E.R. Johnson, S. Keinan, R. Chaudret, J.-P. Piquemal, D.N.
microhydration on the guanidinium


benzene interaction, J. Chem. Phys. Beratan, W. Yang, NCIPLOT: a program for plotting noncovalent interaction
135 (2011) 214301–214309. regions, J. Chem. Theory Comput. 7 (2011) 625–632.
[12] J.A. Carrazana-Garcia, J. Rodriguez-Otero, E.M. Cabaleiro-Lago, A computational [21] E.R. Johnson, S. Keinan, P. Mori-Sánchez, J. Contreras-García, A.J. Cohen, W.
study of anion-modulated cation-p interactions, J. Phys. Chem. B 116 (2012) Yang, Revealing Noncovalent interactions, J. Am. Chem. Soc. 132 (2010) 6498–
5860–5871. 6506.
[13] M.A. Petrukhina, L.A. Scott, Fragments of Fullerenes and Carbon Nanotubes, [22] T. Lu, F. Chen, Multiwfn: a multifunctional wavefunction analyzer, J. Comput.
John Wiley & Sons Inc., Hoboken, New Jersey, 2012. Chem. 33 (2012) 580–592.
[14] M.A. Petrukhina, L.T. Scott, Coordination chemistry of buckybowls: from [23] M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb, J.R. Cheeseman,
corannulene to a hemifullerene, Dalton Trans. (2005) 2969–2975. G. Scalmani, V. Barone, B. Mennucci, G.A. Petersson, H. Nakatsuji, M. Caricato,
[15] U.D. Priyakumar, G.N. Sastry, Cation-p interactions of curved polycyclic X. Li, H.P. Hratchian, A.F. Izmaylov, J. Bloino, G. Zheng, J.L. Sonnenberg, M.
systems: M+ (M = Li and Na) ion complexation with buckybowls, Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y.
Tetrahedron Lett. 44 (2003) 6043–6046. Honda, O. Kitao, H. Nakai, T. Vreven, J.M.J.A., J.E. Peralta, F. Ogliaro, M.
[16] D. Vijay, H. Sakurai, V. Subramanian, G.N. Sastry, Where to bind in buckybowls? Bearpark, J.J. Heyd, E. Brothers, K.N. Kudin, V.N. Staroverov, R. Kobayashi, J.
The dilemma of a metal ion, Phys. Chem. Chem. Phys. 14 (2012) 3057–3065. Normand, K. Raghavachari, A. Rendell, J.C. Burant, S.S. Iyengar, J. Tomasi, M.
[17] J.R. Green, R.C. Dunbar, When do molecular bowls encapsulate metal cations?, Cossi, N. Rega, N.J. Millam, M. Klene, J.E. Knox, J.B. Cross, V. Bakken, C. Adamo, J.
J Phys. Chem. A 115 (2011) 4968–4975. Jaramillo, R. Gomperts, R.E. Stratmann, O. Yazyev, A.J. Austin, R. Cammi, C.
[18] P. Garcia-Novo, A. Campo-Cacharron, E.M. Cabaleiro-Lago, J. Rodriguez-Otero, Pomelli, J.W. Ochterski, R.L. Martin, K. Morokuma, V.G. Zakrzewski, G.A. Voth,
Interaction between anions and substituted molecular bowls, Phys. Chem. P. Salvador, J.J. Dannenberg, S. Dapprich, A.D. Daniels, Ö. Farkas, J.B. Foresman,
Chem. Phys. 14 (2012) 104–112. J.V. Ortiz, J. Cioslowski, D.J. Fox. Gaussian 09, revision A.02. 2009.
[19] R.F.W. Bader, Atoms in Molecules: A Quantum Theory, Clarendon Press,
Oxford, 1990.