Synthesis, binding properties and theoretical studies of p-tert-butylhexahomotrioxacalix[3]arene tri(adamantyl)ketone with alkali, alkaline earth, transition, heavy metal and lanthanide cations

Article abstract of DOI:10.1016/j.tet.2008.11.005

p-tert-Butylhexahomotrioxacalix[3]arene tri(adamantyl)ketone (1b) was synthesized for the first time. Compound 1b was obtained in a cone conformation in solution at room temperature, as established by NMR spectroscopy (¹H and ¹³C). The binding properties of ligand 1b for alkali, alkaline earth, transition, heavy metal and lanthanide cations have been assessed by phase transfer and proton NMR titration experiments. Molecular mechanics and ab initio techniques were also employed to complement the NMR data. The results are compared to those obtained with other closely related homooxacalixarene derivatives. Although triketone 1b is a weak extractant, it shows a strong peak selectivity for Na⁺ and also some preference for Ag⁺. Proton NMR titrations indicate the formation of 1:1 complexes between 1b and the cations studied, and also that they should be located inside the cavity defined by the phenoxy and carbonyl oxygen atoms. Although the molecular mechanics results show little correlation with the NMR data, a good agreement was obtained with the ab initio models.

Full text of DOI:10.1016/j.tet.2008.11.005

Tetrahedron 65 (2009) 496–503
Contents lists available at ScienceDirect
Tetrahedron
journal homepage: www.elsevier.com/locate/tet
Synthesis, binding properties and theoretical studies of p-tert-butylhexa-
homotrioxacalix[3]arene tri(adamantyl)ketone with alkali, alkaline earth,
transition, heavy metal and lanthanide cations
,
b
,
c
a,
Paula M. Marcos a , Jose R. Ascenso , Manuel A.P. Segurado b, Raul J. Bernardino d, Peter J. Cragg e
*
´
a Centro de Cieˆncias Moleculares e Materiais, Faculdade de Cieˆncias da Universidade de Lisboa, Edif´ıcio C8, 1749-016 Lisboa, Portugal
b
´
Faculdade de Farmacia da Universidade de Lisboa, Av. Prof. Gama Pinto, 1649-003 Lisboa, Portugal
c
´
Instituto Superior Tecnico, Complexo I, Av. Rovisco Pais, 1049-001 Lisboa, Portugal
d
a
´
´
´
Escola Superior de Tecnologia do Mar, Instituto Politecnico de Leiria, Santuario N. Sra. dos Remedios, Apartado 126, 2524-909 Peniche, Portugal
e School of Pharmacy and Biomolecular Sciences, University of Brighton, Brighton BN2 4GJ, UK
a r t i c l e i n f o
a b s t r a c t
Article history:
p-tert-Butylhexahomotrioxacalix[3]arene tri(adamantyl)ketone (1b) was synthesized for the first time.
Compound 1b was obtained in a cone conformation in solution at room temperature, as established by
NMR spectroscopy (1H and 13C). The binding properties of ligand 1b for alkali, alkaline earth, transition,
heavy metal and lanthanide cations have been assessed by phase transfer and proton NMR titration
experiments. Molecular mechanics and ab initio techniques were also employed to complement the NMR
data. The results are compared to those obtained with other closely related homooxacalixarene de-
rivatives. Although triketone 1b is a weak extractant, it shows a strong peak selectivity for Naþ and also
some preference for Agþ. Proton NMR titrations indicate the formation of 1:1 complexes between 1b and
the cations studied, and also that they should be located inside the cavity defined by the phenoxy and
carbonyl oxygen atoms. Although the molecular mechanics results show little correlation with the NMR
data, a good agreement was obtained with the ab initio models.
Received 29 September 2008
Received in revised form 26 October 2008
Accepted 4 November 2008
Available online 8 November 2008
Keywords:
Calixarenes
Hexahomotrioxacalix[3]arene ketone
Metal cation extraction
Proton NMR titration
Molecular mechanics
Ab initio
Ó 2008 Elsevier Ltd. All rights reserved.
1. Introduction
symmetry. However, and despite these features, the studies of the
ionophoric abilities of hexahomotrioxacalix[3]arenes bearing car-
The importance of calixarenes1–3 in host–guest chemistry, even
after more than three decades of research, is due to their relatively
easy functionalization on both the upper and lower rims, giving
a great variety of derivatives, which can become versatile molecular
platforms for the design of supramolecular structures.
In particular, the ability of carbonyl containing substituents on
the lower rim of calixarenes to bind metal ions, predominantly
alkali and alkaline earth,4,5 but also transition, heavy metal6,7 and
lanthanide8,9 cations, has been largely studied. At the same time,
homooxacalixarenes,10,11 which derive from calixarenes by re-
placement of one or more CH2 bridges by CH2OCH2 groups, have
also been investigated. Among them, hexahomotrioxacalix[3]ar-
enes have received significant attention as receptors, mainly due to
their structural features:12 a cavity formed by a 18-membered ring,
only two basic conformations (cone and partial cone) and a C3-
bonyl groups at their lower rims towards metal cations are scarce
compared to those of the calix[4]arene analogues. A few examples
of derivatives containing ester,13,14 carboxylic acid15 and amide16,17
groups are reported in the literature.
Following our previous studies on the synthesis and binding
properties of homooxacalixarenes bearing carbonyl groups at the
lower rim, namely amide18,19 and ketone20,21 functions, towards
a large variety of cations, we obtained a new hexahomotrioxacal-
ix[3]arene derivative. This paper reports the synthesis, the NMR
conformational analysis and the binding properties of p-tert-
butylhexahomotrioxacalix[3]arene tri(adamantyl)ketone (1b) to-
wards alkali, alkaline earth, transition (Mn2þ, Fe2þ, Co2þ, Ni2þ, Cu2þ
and Zn2þ), heavy metal (Agþ, Cd2þ and Pb2þ) and lanthanide (La3þ
,
Ce3þ, Pr3þ, Nd3þ, Sm3þ, Eu3þ, Gd3þ, Dy3þ, Er3þ and Yb3þ) cations.
These properties have been assessed by extraction studies with the
corresponding metal picrates from an aqueous solution into
dichloromethane and by proton NMR titrations. In the absence of
X-ray structures, theoretical studies using molecular mechanics
and ab initio techniques were employed to complement the in-
formation given by the NMR data. The results are compared to
those obtained with other closely related homooxacalixarene
* Corresponding author. Centro de Cieˆncias Moleculares e Materiais, Faculdade de
ˆ
´
Ciencias da Universidade de Lisboa, Edifıcio C8, 1749-016 Lisboa, Portugal. Tel.:
þ351 21 7500111; fax: þ351 21 7500979.
E-mail address: pmmarcos@fc.ul.pt (P.M. Marcos).
0040-4020/$ – see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tet.2008.11.005

P.M. Marcos et al. / Tetrahedron 65 (2009) 496–503
497
derivatives, and discussed in terms of substituents, size and con-
formational effects of the macrocycles.
Tables 1 and 2. The corresponding values for compounds 1c19 and
2b20,21 are included for comparison.
To the best of our knowledge, this is the first report on the
synthesis and ion binding properties of a hexahomotrioxacalix[3]-
arene ketone derivative.
The results obtained with ketone 1b range from 4 to 20% for the
alkali cations and from 2 to 4% for the alkaline earth cations (Table 1).
These data reveal that 1b is a weak extractant for both groups of
cations, except for Naþ (20%), for which it shows a significant
preference. In fact, ketone 1b, although displaying much lower
extraction percentages than triamide 1c, exhibits a sharper peak
t-Bu
t-Bu
t-Bu
selectivity for Naþ than 1c (SNa
¼3:9 and 1:6 for 1b and 1c, re-
þ=Kþ
spectively). However, within these series of cations, both hexa-
homotrioxacalix[3]arene derivatives show similar extraction
profiles. Compared to its dihomooxacalix[4]arene tetraketone an-
alogue 2b, compound 1b is a much weaker extractant for all the
cations. This is due presumably to its higher conformational flexi-
bility. The replacement of the methylene bridges by dimethyl-
OR
OR
OR
OR
OR
O
O
O
OR
RO
O
t-Bu
t-Bu
t-Bu
eneoxa bridges in triketone 1b increases its flexibility. The
D
Gs
t-Bu
1a R = H
2a R = H
barriers for conformational inversion in CDCl3 are <9 and
12.9 kcal molꢀ1 for the parent calixarenes 1a and 2a, respectively.23
Furthermore, 1b possesses only three ketone groups and conse-
quently only six donating sites can surround the cations, compared
to the eight sites of the tetraketone.
In the case of the transition and heavy metal cations (Table 1)
a similar situation can be observed. Compound 1b is also a poor
extractant showing, however, some preference for Agþ (13%). An
extraction profile favouring Agþ over Pb2þ is now observed for 1b
1b R = CH2COAd
2b R = CH2COAd
1c R = CH2CON(Et)2
2. Results and discussion
2.1. Synthesis and NMR conformational analysis
Treatment of compound 1a with 1-adamantyl bromomethyl
ketone and sodium hydride in dry THF under reflux furnished the
hexahomotrioxacalix[3]arene triketone 1b. Proton and carbon-13
NMR experiments were carried out in chloroform at room tem-
perature to establish its conformation.
(SAgþ
(SPb2þ
¼2:6), an inverse trend to that shown by triamide 1c
=Pb2þ
=Agþ
¼2). The ketone behaviour is similar to those already
found for other dihomooxa ketone derivatives.21 Although these
ligands contain hard oxygen donor atoms, they display a clear
preference for Agþ, a soft Lewis acid.
The C3-symmetry in triketone 1b is reflected by its proton and
carbon NMR spectra. The 1H NMR spectrum displays one AB quartet
for the CH2 bridge protons and one singlet for the tert-butyl and
aromatic protons. In addition, the OCH2COAd groups exhibit one
singlet for the CH2 protons and a series of multiplets for the Ad
groups. The 13C NMR spectrum shows a pattern containing five
downfield resonances arising from the carbonyl and aromatic car-
bon atoms, two midfield resonances arising from the methylene
carbon atoms of the OCH2COAd and CH2OCH2 groups, and six up-
field resonances arising from the quaternary carbon atoms of the
Ad and tert-butyl groups (2 lines), the secondary and tertiary car-
bon atoms of the Ad group (3 lines) and the methyl carbon atoms of
the tert-butyl group (1 line). All resonances were assigned by DEPT
experiments.
The extraction of lanthanide cations by both adamantyl ketones
was studied in this work (Table 2). Triketone 1b is practically unable
to discriminate among the ten lanthanides investigated and is
a weak phase transfer agent (%E ranges from 5 to 7). The lack of
donating groups may be even more relevant here, as lanthanides
require high coordination numbers (8 or 9).24 Tetraketone 2b
shows higher percentages (%E ranges from 6 to 23) and preference
for the heavy lanthanides Gd3þ and Yb3þ (23 and 21%, respectively).
The extraction trend observed for both ketones is similar, but this
behaviour is the opposite of that found by us for the corresponding
amide derivatives.25
As was mentioned before, this kind of calixarene can exist in
only two basic conformations (cone and partial cone). Thus, those
spectra clearly indicate the cone conformation for compound 1b.
2.3. Proton NMR studies
To obtain further information on the cation binding behaviour of
triketone 1b, specifically concerning the binding sites, proton NMR
titrations in CDCl3 were performed. The cations studied were Naþ,
Kþ, Ca2þ, Sr2þ, Ba2þ, Zn2þ, Agþ, Pb2þ, La3þ, Eu3þ and Yb3þ. Different
situations were found after the addition of the salts to the ligand.
No complexation was observed in the cases of Zn2þ, Eu3þ and Yb3þ
and in all other cases the experimental data were consistent with
the formation of 1:1 complexes only.
2.2. Extraction studies
The ionophoric properties of triketone 1b towards alkali, alka-
line earth, transition, heavy metal and lanthanide cations were
evaluated by the standard picrate extraction method.22 The results,
expressed as a percentage of cation extracted (%E), are reported in
Table 1
Percentage extraction of alkali, alkaline earth, transition and heavy metal picrates into CH2Cl2 at 25 ꢁCa
Liþ
Naþ
Kþ
Rbþ
Csþ
Mg2þ
Ca2þ
Sr2þ
Ba2þ
Mn2þ
Fe2þ
Co2þ
Ni2þ
Cu2þ
Zn2þ
Agþ
Cd2þ
Pb2þ
Ionic radiusb/Å
0.78
4.0
23
0.98
20
82
1.33
5.2
85
1.49
5.6
66
1.65
5.5
30
0.78
2.0
13
1.06
4.1
15
1.27
3.9
23
1.43
3.9
63
0.83
2.4
7.0
0.78
5.2
11
0.75
1.5
5.2
39
0.69
3.6
7.0
0.73
4.3
31
0.75
3.3
21
1.15
13
80
0.95
3.0
15
1.18
5.0
43
1b
2bc
1cd
25
50
32
27
20
17
34
41
55
19
19
45
24
15
40
37
80
a
Values with uncertainties less than 5%.
b
Goldschmidt, V. M. Skrifter Norske Videnskaps-Akad. Oslo, I, Mat.-Naturv. Kl, 1926 and Shannon, R. D.; Prewitt, C. T. Acta Crystallogr. 1969, B25, 925; 1970, B26, 1046; data
quoted in Marcus, I. Ion Properties; Marcel Dekker: New York, NY, 1997; pp 46–47.
c
Data taken from Refs. 20 and 21.
Data taken from Ref. 19.
d

498
P.M. Marcos et al. / Tetrahedron 65 (2009) 496–503
and the lowest for La3þ
(
Dd¼0.05 ppm). This behaviour of ketone 1b
Table 2
Percentage extraction of lanthanide picrates into CH2Cl2 at 25 ꢁCa
is closer to that of the hexahomotrioxacalix[3]arene amide de-
rivative (1c), studied previously.19
La3þ Ce3þ Pr3þ Nd3þ Sm3þ Eu3þ Gd3þ Dy3þ Er3þ Yb3þ
Ionic radiusb/Å 1.03 1.01 0.99 0.98 0.96 0.95 0.94 0.91 0.89 0.87
1b
2b
For other calixarenes, the variation in chemical shifts experi-
enced by the equatorial methylene protons of the ArCH2Ar bridges
is downfield and much smaller than that of the axial protons, but
the ArCH2OCH2Ar bridges behave differently, as reported be-
fore.13,18–21 The equatorial methylene protons of the oxygen
bridges move upfield and experience larger shift variations than
the axial ones. These results suggest that the oxygen bridge con-
formation changes significantly upon complexation, with the equa-
torial protons undergoing a higher shielding effect. The maximum
6.0
9.3
5.1
7.7
4.5
6.0
5.6
8.4
5.9
14
5.7
7.4
5.9
23
5.3
7.9
6.0
12
6.8
21
a
Values with uncertainties less than 5%.
Shannon, R. D.; Prewitt, C. T. Acta Crystallogr. 1969, B25, 925; 1970, B26, 1046;
b
data quoted in Marcus, I. Ion Properties; Marcel Dekker: New York, NY, 1997; pp
46–47.
Variable amounts of the salts were added to ketone 1b and the
proton spectra were recorded after each addition. These titrations
indicate similar behaviours for Naþ and the divalent cations. With
[salt]/[ligand] ratios lower than one, sharp signals of the complexed
and uncomplexed ligand are present in the spectra (for Ca2þ and
Sr2þ they are slightly broad), indicating that on the NMR time scale
the exchange rate between the two species is slow, at room tem-
perature. This behaviour reflects a high affinity of ligand 1b for
these cations. Upon reaching 1:1 ratios, all the signals for the free
ligand disappear and those of the complexed ligand remain un-
altered after subsequent addition of the salts, indicating a 1:1
metal-to-ligand stoichiometry. In contrast, titrations of 1b with Kþ
and Agþ initially induce broadening of the signals until the [salt]/
[ligand] ratio reaches the unity, when the signals become sharp.
This indicates a fast exchange rate between the two species on the
NMR time scale, at room temperature. Again, the NMR titrations
suggest a 1:1 stoichiometry, since no further spectral changes are
observed after subsequent additions of the salts.
value was recorded for Naþ
La3þ
(Dd¼0.30 ppm) and the minimum for
(Dd¼0.11 ppm). All the monovalent cations show higher shift
changes than the divalent ones. Gutsche26 established that the
difference in the chemical shifts between the axial and equatorial
protons of the ArCH2Ar bridges indicates the degree of flattening
of the cone conformation. A value around 0.9 ppm means a regu-
lar cone conformation and zero a regular 1,3-alternate confor-
mation. Upon complexation with all eight cations the DdHax–Heq
increases (Table 3). If Gutsche’s criterion is also applicable to the
CH2OCH2 bridges, it indicates that the phenyl groups in triketone
1b are more flattened than those in calix[4]arenes and diho-
mooxacalix[4]arenes, and stand up when the cations enter into
the ionophoric cavity, resulting in a more regular cone confor-
mation upon complexation. Interestingly, the highest difference
is observed for Kþ
Dd¼0.31 ppm).
(
Dd¼0.61 ppm) and the lowest for Naþ
(
The deshielding effect observed for the aromatic protons in-
Proton NMR data of the free and complexed ligand 1b are col-
lected in Table 3. Complexation of the cations affects all the proton
chemical shifts in the ligand, except in the cases of Zn2þ, Eu3þ and
Yb3þ where there are no variations, as already mentioned. The
largest variations are recorded for the aromatic protons and the
methylene protons of the OCH2CO groups, which move downfield,
and by the oxygen bridge equatorial methylene protons (CH2OCH2),
which move upfield. The smallest variations are observed for the
protons of the adamantyl and tert-butyl groups. Identical situation
was already found for triamide 1c.19
dicates the involvement of the phenolic oxygens in complexation,
as reported previously.27 The largest shift changes shown by the
protons adjacent to the oxygen donor atoms can be explained in
terms of variations of both shielding and deshielding effects of the
aromatic rings and carbonyl groups upon cation binding to those
donor atoms. Therefore, this suggests that the cations must be in-
side the cavity defined by the phenoxy and carbonyl oxygen atoms
and bound through metal–oxygen interactions.
The magnitude of the chemical shift variations for triketone 1b
follows the order: Naþ>AgþzCa2þzPb2þ>Sr2þzKþ>Ba2þ>La3þ
.
From this trend, Naþ is the best bound cation, followed by com-
parable spectral changes among Agþ, Ca2þ and Pb2þ. These results
reflect the extraction data, particularly the preference for the
monovalent cations Naþ and Agþ. Although lanthanides are hard
Lewis acids and have strong affinity towards hard oxygen donor
atoms, they are the poorest bound cations. This may be due to the
fact that lanthanide cations require high coordination numbers (8
or 9) and therefore the six donating sites available in triketone 1b
are not enough to promote a strong complexation. That magnitude
order for 1b suggests that, more important than the charge (mono,
di or trivalent) and nature (hard, soft or intermediate) of the cat-
ions, the size is the main factor affecting complexation. An average
cation size of 1.09 Å seems to best match the cavity dimensions of
1b. It is worthwhile to point out that among the three lanthanides
studied, only La3þ (ionic radius of 1.03 Å) was complexed.
It has been observed that the chemical shift changes of the
methylene protons of the OCH2CO groups in dihomooxacalix[4]-
arene amide18 and ketone20,21 derivatives are upfield for the
monovalent cations and downfield for the divalent ones. A close
examination of the spectral changes upon complexation indicates
that triketone 1b behaves in a slightly different way. Except for
Ba2þ, which shows a small upfield shift, all the cations experience
downfield shift changes. However, the divalent cations still show
much higher variations than the monovalent and also the trivalent
cations. The highest value was recorded for Pb2þ
(Dd¼0.68 ppm)
Table 3
Proton chemical shifts (d, ppm) of ligand 1b and its 1:1 metal complexes in CDCl3
t-Bu
Ad
CH2OCH2
eq
OCH2CO
ArH
ax
Dd
2.4. Theoretical studies
1b
1.09
1.23
1.20
1.22
1.19
1.13
1.10
1.24
1.21
1.13
1.10
1.10
1.67–2.02
1.70–2.12
1.70–2.08
1.73–2.12
1.72–2.10
1.68–2.05
1.67–2.03
1.70–2.11
1.75–2.14
1.68–2.04
1.67–2.03
1.67–2.03
4.49
4.19
4.23
4.33
4.29
4.31
4.48
4.24
4.31
4.38
4.48
4.49
4.75
4.50
4.84
4.82
4.86
4.70
4.76
4.69
4.83
4.86
4.75
4.76
0.26
0.31
0.61
0.49
0.57
0.39
0.28
0.45
0.52
0.48
0.27
0.27
4.85
5.03
4.95
5.27
5.09
4.78
4.86
5.02
5.53
4.90
4.87
4.87
6.91
7.25
7.13
7.23
7.19
6.99
6.91
7.28
7.21
6.95
6.91
6.91
1bþNaþ
1bþKþ
Energy minimized structures were calculated using molecular
mechanics for 1b and its metal cation complexes to mirror those
used in the NMR experiments. In previous work this method has
met with some success,19 however, here analysis of the molecular
mechanics results revealed little correlation with the NMR data.
Table 4 gives data related to the coordination environment around
the cations studied. Naþ, Kþ, Ca2þ and Zn2þ were predicted to be
bound within the cavity defined by the lower rim and bridging
oxygen atoms. For these cations the geometry of 1b was quite
1bþCa2þ
1bþSr2þ
1bþBa2þ
1bþZn2þ
1bþAgþ
1bþPb2þ
1bþLa3þ
1bþEu3þ
1bþYb3þ

P.M. Marcos et al. / Tetrahedron 65 (2009) 496–503
499
and Agþ complexes (Fig. 2b) shows the cations slightly below the
plane defined by the three phenoxy oxygen atoms, interacting
mainly with the three carbonyl oxygen atoms and also with two of
the phenoxy oxygen atoms (‘3þ2’ interaction), as shown by the
distances given in Table 5. In the case of Zn2þ, Sr2þ and Ba2þ
complexes (Fig. 2c) a ‘3þ2’ interaction mode is also observed, but
with the adamantyl groups more distant from each other. For Ca2þ
and La3þ complexes (Fig. 2d), the cations coordinate with the six
oxygen atoms in a higher position within 1b, especially for Ca2þ. For
all these complexes, the presence of the cation imposes slightly
shorter distances among the tert-butyl groups and slightly larger
distances among the adamantyl groups, comparing to free ligand
1b. The results of the ab initio optimized structures towards Ca2þ
and La3þ are consistent with the NMR data. Both cations induce
approximately the same chemical shift variations for the aromatic
and methylenic protons of the OCH2CO groups as, for example, for
Table 4
Summary of molecular modelling experiments
Complex
Position of cationa
Mnþ–O
Aromatic centroid
separation
interactionsb
1b$Naþ
ꢀ1.92
Carbonyl
2.35
2.37c
2.49c
2.69c
2.74c
2.75
2.36
2.55c
2.65
5.5
5.6
5.4
Alkyl ether
Aryl ether
Alkyl ether
Aryl ether
Carbonyl
Carbonyl
Alkyl ether
Aryl ether
1b$Kþ
ꢀ2.07
ꢀ1.61
1b$Ca2þ
1b$Sr2þ
1b$Ba2þ
1b$Zn2þ
5.44d
3.45d
ꢀ1.61
5.4
5.4
5.4
Carbonyl
Aryl ether
Alkyl ether
2.11
2.66
4.20c
1b$Agþ
1b$Pb2þ
1b$La3þ
ꢀ3.78
8.13d
9.70d
5.4
5.2
5.2
Ca2þ
(Dd¼0.32 and 0.42 ppm, respectively). For the remaining
cations a slight discrepancy between the NMR and the ab initio data
was found, mainly in the case of Naþ, Kþ and Agþ (see Table 3). For
these cations the aromatic protons undergo a larger chemical shift
variation than the OCH2CO methylenic protons, suggesting
a stronger interaction with the phenoxy oxygen atoms than with
the carbonyl oxygen atoms. Finally for Pb2þ complex (Fig. 2e) the
cation is in the lowest position within 1b, in the same plane as the
adamantyl groups, interacting with the three carbonyl oxygen
atoms only, and presenting the highest dipole moment. This is in
agreement with the NMR data, as Pb2þ induces the largest chemical
shift variation for the methylene protons of the OCH2CO groups
All distances in Å.
a
Relative to the plane defined by the three alkyl ether oxygen atoms in 1b.
Less than 5 Å; mean distance of three closest contacts unless otherwise stated.
Single interaction.
b
c
d
Outside macrocyclic cavity.
regular with the three aromatic rings adopting pseudo-C3v symmetry.
Agþ interacts weakly with the lower rim of 1b and the remaining
cations, Sr2þ, Ba2þ, Pb2þ and La3þ, do not appear to interact at all. In
these five examples the macrocycle has a highly distorted upper
rim geometry. Examples of these binding motifs are illustrated in
Figure 1. It has been noted that gas phase models of solution be-
haviour often break down at low levels of theory28 so a more so-
phisticated ab initio approach was adopted.
Although the comparison between gas phase and condensed
phase structures is not direct, as evidenced by the molecular me-
chanics results, all the structures were optimized solvent free in the
ab initio simulations. They are therefore expected to correlate pri-
marily with gas state complex geometries and not necessarily with
the extractability from aqueous solution data, in which the solvent
will influence the metal–ligand interactions and relative com-
plexation energies.
(
Dd¼0.68 ppm) and a much smaller one (Dd¼0.30 ppm) for the
aromatic protons. This difference suggests that Pb2þ cation strongly
interacts with the carbonyl oxygens than with the phenoxy oxy-
gens. Moreover, it also induces the largest chemical shifts for the
adamantyl protons, among all the cations.
The ab initio calculations also predict that the oxygen atoms of
the ArCH2OCH2Ar bridges do not participate in metal binding, since
average distances from the cations to the mean plane of the three
bridge oxygen atoms are too great for such an interaction (Table 5).
A similar situation was already mentioned for a closely related
hexahomotrioxacalix[3]arene derivative.13
In order to determine the electrostatic equilibriums in the dif-
ferent complexes, maps of electrostatic potential were calculated
Four different binding modes were obtained for the metal
complexes studied. They are illustrated in Figure 2, together with
the predicted structure of ligand 1b (Fig. 2a). The structural data for
1b complexes are reported in Table 5. The binding mode for Naþ, Kþ
over an electronic isodensity surface of
these surfaces, the red regions are those of relatively lower or more
negative electrostatic potential and the blue regions correspond to
r
¼0.04 Åꢀ3 (Fig. 3). On
Figure 1. Optimized molecular mechanics structures for 1b–metal complexes (a) Naþ, Kþ, Ca2þ and Zn2þ, (b) Agþ and (c) Sr2þ, Ba2þ, Pb2þ and La3þ. All the hydrogens were removed
for clarity.

500
P.M. Marcos et al. / Tetrahedron 65 (2009) 496–503
Figure 2. Optimized ab initio structures for 1b–metal complexes (a) 1b, (b) Naþ, Kþ and Agþ, (c) Zn2þ, Sr2þ and Ba2þ, (d) Ca2þ and La3þ and (e) Pb2þ. All the hydrogens were
removed for clarity.
higher or more positive electrostatic potential, as seen in the as-
sociated scale. Each cation produces a visible reorganization of all
the charges in the ligand as shown, for example, in the case of Zn2þ
(Fig. 3c). Although this structure is very similar to that of 1b
(Fig. 3a), the presence of high positive potential zones can make
this complex less likely to be stabilized by the solvent. Contrarily,
the Naþ complex (Fig. 3b) shows a charge distribution closer to that
of ligand 1b, which may indicate a better solvent stabilization for
this complex. These interpretations are in agreement with the NMR
data, where Naþ is the best bound cation.
From the complexation energies reported in Table 6, the fol-
lowing order was observed: La3þ>Zn2þ>Ca2þ>Sr2þ>Pb2þ>Ba2þ
>Naþ>Agþ>Kþ. As with the molecular mechanics results, this
order of stability does not exactly match the NMR sequence as the
two most stable complexes observed in the ab initio model (La3þ
and Zn2þ) are among the least stable complexes in the NMR. This
Table 5
Structural data of 1b complexes (HF/6-31G)
carb–metalb (Å)
Ophen.–metalb (Å)
(C–C)tert (Å)
Dipole moment
(Debye)
Metal
Ocarb charged
Ophen charged
c
Position
O
cationa (Å)
charged
1b
d
d
d
6.37, 7.13, 7.43
6.23
6.14
8.35, 6.54, 6.92
6.14
5.80
6.77, 5.95, 6.18
6.17
6.24
6.44
7.30
4.26
4.28
4.95
7.63
7.90
7.26
3.98
17.32
4.51
d
ꢀ0.561
ꢀ0.784
1b$Naþ
1b$Kþ
ꢀ3.37
ꢀ3.77
ꢀ1.64
ꢀ3.54
ꢀ3.75
ꢀ3.24
ꢀ3.61
ꢀ5.58
ꢀ2.61
2.28
2.65
2.45
2.44
2.62
2.02
2.43
2.31
2.44
2.42, 2.45, 4.80
2.85, 2.96, 5.14
2.49
2.59, 2.62, 5.15
2.80, 2.84, 5.30
2.12, 2.22, 4.80
2.70, 2.71, 4.91
4.43, 4.57, 5.56
2.54
þ0.722
þ0.789
þ1.694
þ1.500
þ1.528
þ1.460
þ0.586
þ1.556
þ2.239
ꢀ0.628
ꢀ0.850, ꢀ0.839, ꢀ0.795
ꢀ0.813, ꢀ0.818, ꢀ0.793
ꢀ0.912
ꢀ0.626
1b$Ca2þ
1b$Sr2þ
1b$Ba2þ
1b$Zn2þ
1b$Agþ
1b$Pb2þ
1b$La3þ
ꢀ0.696
ꢀ0.710, ꢀ0.707, ꢀ0.769
ꢀ0.755, ꢀ0.705, ꢀ0.702
ꢀ0.699, ꢀ0.805, ꢀ0.726
ꢀ0.611
ꢀ0.791, ꢀ0.904, ꢀ0.905
ꢀ0.883, ꢀ0.882, ꢀ0.789
ꢀ0.788, ꢀ0.922, ꢀ0.956
ꢀ0.817
ꢀ0.781, ꢀ0.785, ꢀ0.846
ꢀ0.747
ꢀ0.796
ꢀ0.993
a
Distance between the cation position and the plane defined by the three bridge oxygen atoms.
b
Distance between the cation and each one of the oxygen atoms of the three carbonyl groups (Ocarb) or the three phenoxy groups (Ophen). When values are similar the mean
is presented.
c
Distance between two adjacent tertiary carbons from the tert-butyl groups.
Mulliken charges.
d

P.M. Marcos et al. / Tetrahedron 65 (2009) 496–503
501
Figure 3. Electrostatic potential for 1b–metal complexes. The electrostatic potential (in a.u.) is represented over a constant electronic isodensity
r
(in Åꢀ3) surfaces of volume Vs (in
Å
ꢀ3). All figures correspond to
r
¼0.05 Åꢀ3. (a) 1b, (b) Naþ, (c) Zn2þ and (d) Ca2þ
.
discrepancy is almost certainly related to the energy of metal
desolvation and number of solvent molecules found in the metal’s
primary coordination sphere in the complex. In order to de-
termine the importance of this process, the desolvation energies
(Table 7) were calculated for each cation with six methanol
molecules, except for Agþ and La3þ, where the coordination
numbers are 2 and 9, respectively.29 The larger desolvation ener-
gies associated with the more positive cations can make the
complexation process considerably harder, and therefore be re-
sponsible for the behaviour observed, especially in the case of
Zn2þ. For this cation, the desolvation barrier is almost two times
the barriers associated with the other divalent cations, which may
explain the lack of interaction between Zn2þ and 1b observed in
the NMR data.
Table 6
Total energies E (a.u.) and complexation energies
D
Ec (kcal molꢀ1 a for the systems
)
studied
HF/3-21Gb
HF/6-31Gb
E
DEc
E
D
Ec
3. Conclusions
1b
ꢀ3437.10452
ꢀ3597.97122
ꢀ4033.26479
ꢀ4110.42396
ꢀ6554.00591
ꢀ3462.02825c
ꢀ5206.03696
ꢀ8610.88236
ꢀ3440.05393c
ꢀ3870.94318c
d
ꢀ3454.74367
ꢀ3616.55910
ꢀ4053.82257
ꢀ4131.33155
ꢀ3484.90005c
ꢀ3479.63556c
ꢀ5231.93634
ꢀ3600.73660c
ꢀ3457.67496c
ꢀ3888.51075c
d
1b$Naþ
1b$Kþ
ꢀ121.08
ꢀ96.44
ꢀ323.63
ꢀ222.25
ꢀ221.66
ꢀ412.53
ꢀ111.25
ꢀ226.76
ꢀ557.16
ꢀ97.99
Hexahomotrioxacalix[3]arene triketone 1b was synthesized and
obtained in a cone conformation in solution at room temperature.
Extraction studies with metal picrates from an aqueous solution
into CH2Cl2 have shown that ketone 1b is a weak extracting agent.
However, 1b exhibits a sharp peak selectivity for Naþ and also some
preference for Agþ. The low extraction percentages obtained with
1b are, presumably, due to its higher conformational flexibility
compared to its dihomooxacalix[4]arene tetraketone analogue 2b.
Moreover, 1b possesses only three ketone groups and consequently
only six donating sites can surround the cations. The NMR data
support the extraction results, particularly the preference for the
monovalent cations Naþ and Agþ. Ligand 1b forms 1:1 complexes
ꢀ67.28
ꢀ303.71
ꢀ235.82
ꢀ201.68
ꢀ364.33
ꢀ90.66
1b$Ca2þ
1b$Sr2þ
1b$Ba2þ
1b$Zn2þ
1b$Agþ
1b$Pb2þ
1b$La3þ
ꢀ215.38
ꢀ512.24
a
D
Ec corresponds to the process 1bþMnþ/[1bꢀMnþ], where Mnþ represents
each one of the cations.
b
Optimized geometries at this level of theory.
Energies obtained with the referred level of theory for all atoms except the
c
cation. For the cations the Stuttgart/Dresden ECP basis set was used.

502
P.M. Marcos et al. / Tetrahedron 65 (2009) 496–503
Table 7
Desolvation energies
a
D
Ed (kcal molꢀ1
) for the systems studied
Naþ
Kþ
Ca2þ
Sr2þ
Ba2þ
Zn2þ
Agþ
Pb2þ
La3þ
HF/6-31Gb
D
Ed
þ87.49
þ54.16
þ241.95
þ222.14c
þ188.42c
þ352.55
þ61.42c
þ205.83c
þ640.00c
a
D
Ed corresponds to the process [M(MeOH)6]nþ/Mnþþ6MeOH, where M represents each one of the cations; for Agþ and La3þ the coordination numbers are 2 and 9,
a
respectively.
b
Optimized geometries at this level of theory.
Energies obtained with the referred level of theory for all atoms except the cation. For the cations the Stuttgart/Dresden ECP basis set was used.
c
with Naþ, Kþ, Ca2þ, Sr2þ, Ba2þ, Agþ, Pb2þ and La3þ, and no com-
plexation with Zn2þ, Eu3þ and Yb3þ. The structures of these com-
plexes deduced from NMR experiments show that the cations
should be encapsulated in the cavity composed of the phenoxy and
carbonyl oxygen atoms. A certain discrepancy was observed be-
tween the molecular mechanics results and the NMR data. Never-
theless, at a higher level of theory, the ab initio model predicted
positions for the bound cations that agree with the NMR data, al-
though it does not satisfactorily explain the strength of the metal
binding.
transition metal picrate preparation have already been described.30
All the other metal picrates (alkali, alkaline earth, heavy and lan-
thanides) were also prepared using the same method. Thus, they
were prepared by adding an excess of the respective metal car-
bonate in a hot (80–90 ꢁC) and saturated aqueous picric acid so-
lution. After filtration of the unreacted carbonate, the solution was
concentrated and cooled until precipitation was complete. The
crystals of metal picrates were then recrystallized from hot aqueous
solutions. For the Ag(I) picrate preparation, silver oxide was used
instead of silver carbonate. Due to the light sensitivity of the silver
salts, the procedure was done avoiding direct light.
4. Experimental
4.1. Synthesis
4.3. Proton NMR titration experiments
Several aliquots (up to 2–3 equiv) of the salt solutions (0.25 M)
in CD3OD were added to CDCl3 solutions (0.5ꢂ10ꢀ2 M) of ligand 1b
directly in the NMR tube. The salts used were Na and K thiocya-
nates, Ca, Sr and Pb perchlorates, and Ba, Ag, Zn, La, Eu and Yb
triflates. Due to the low solubility of Eu triflate in MeOH, it was
necessary to decrease the concentration of the ligand
(0.25ꢂ10ꢀ3 M) and of the salt (4.17ꢂ10ꢀ3 M). The spectra were
recorded on a Bruker Avance III 500 Spectrometer after each ad-
dition of the salts. The temperature of the NMR probe was kept
constant at 22 ꢁC.
All chemicals were reagent grade and were used without further
purification. Melting points were measured on a Stuart Scientific
apparatus and are uncorrected. 1H and 13C NMR spectra were
recorded on a Bruker Avance III 500 MHz spectrometer, with TMS
as internal reference. Elemental analysis was determined on
a Fisons EA 1108 microanalyser.
4.1.1. 7,15,23-Tri-tert-butyl-25,26,27-triadamantylketone-
2,3,10,11,18,19-hexahomo-3,11,19-trioxacalix[3]arene (1b)
To a stirred solution of 2.0 g (3.47 mmol) of p-tert-butylhexa-
homotrioxacalix[3]arene (1a) in 100 mL of dry THF was added
0.63 g (21 mmol) of sodium hydride (80% oil dispersion) followed
by 4.12 g (16 mmol) of 1-adamantyl bromomethyl ketone. The re-
action mixture was refluxed under N2 for 24 h. The solvent was
then evaporated, and a yellow oily residue was obtained and
poured into a volume of 160 mL of H2O/CH2Cl2. This mixture was
stirred for about 1 h and, after separating the phases, the organic
layer was washed several times with 1 M HCl solution and water,
and dried with anhydrous Na2SO4. Evaporation of the solvent gave
a yellow oily residue, to which cold ethanol was added. A white
solid was obtained by filtration and then recrystallization from
dichloromethane/ethanol afforded the pure compound 1b (0.84 g,
22 %) as white crystals: mp 205–207 ꢁC. 1H NMR (CDCl3, 500 MHz)
4.4. Experimental theoretical methods
4.4.1. Molecular modelling studies
The coordinates of the ligand and a cation held within the cavity
defined by the amide groups were taken from an X-ray structure in
which adamantyl groups were substituted for amides.31 The initial
structure for each complex was generated by systematically
changing the cation. An unconstrained geometry optimisation was
then carried out on each structure using the Merck Molecular Force
Field (MMFF94) within the Spartan ’06 software suite.32
4.4.2. Ab initio studies
The structures of ketone 1b and its cation complexes were fully
optimized without symmetry constraints using the Hartree–Fock
(HF) method, which is well known to produce accurate geometries.
An initial guess for all the structures was made using the small basis
set 3-21G before reoptimized using the larger 6-31G basis set. For
the larger cations, the binding energies were calculated by using
effective core potentials (ECP) on the metal, derived from energy
adjusted ab initio pseudopotential methods in the Gaussian 03
software package.33 The vibrational frequency calculations were
performed at the same level of theory to check that all structures
were global minima of the potential energy surface (denoted by an
absence of negative vibrational frequencies) and to correct the
computed energies for zero-point energies as well as translational,
rotational, and vibrational contributions to the enthalpy. All cal-
culations were performed with the Gaussian 03 and all the figures
and electrostatic potential surfaces were constructed using the
Molekel software.34
d
1.09 (s, 27H, C(CH3)3),1.67–2.02 (m, 45H, Ad), 4.49, 4.75 (ABq,12H,
J¼13.3 Hz, CH2OCH2), 4.85 (s, 6H, OCH2CO), 6.91 (s, 6H, ArH); 13C
NMR (CDCl3, 125.8 MHz)
d 27.9 (CH, Ad), 31.5 (C(CH3)3), 34.2
(C(CH3)3), 36.6, 38.0 (CH2, Ad), 45.1 (C, Ad), 69.4 (CH2OCH2), 75.1
(OCH2CO), 126.0 (ArH), 131.1, 145.9, 154.0 (Ar), 210.0 (CO). Anal.
Calcd for C72H96O9: C, 78.22; H, 8.75. Found: C, 77.85; H, 8.98.
4.2. Extraction studies
Equal volumes (5 mL) of aqueous solutions of metal picrates
(2.5ꢂ10ꢀ4 M) and solutions of the calixarenes (2.5ꢂ10ꢀ4 M) in CH2Cl2
were vigorously shaken for 2 min, and then thermostated in a wa-
ter bath with mechanical stirring at 25 ꢁC overnight. After complete
phase separation, the concentration of picrate ion in the aqueous
phase was determined spectrophotometrically (lmax¼354 nm). For
each cation–calixarene system the absorbance measurements were
repeated, at least, four times. Blank experiments showed negligible
picrate extraction in the absence of a calixarene. The details of the
This study started with a search for minimum energy confor-
mations using several initial geometries, taking in account that

P.M. Marcos et al. / Tetrahedron 65 (2009) 496–503
12. Araki, K.; Inada, K.; Otsuka, H.; Shinkai, S. Tetrahedron 1993, 49, 9465.
503
each of the cations could electrostatically interact with: (a) one or
more of the three aromatic rings; (b) one or more oxygens from the
oxygen bridges; (c) with one or more oxygens of the phenoxy and
carbonyl groups. After optimizing these initial possibilities, the
global minimum configurations were used for the final optimiza-
tion study with the larger basis set.
The electrostatic potential for the 1b–metal complexes were
drawn over a surface of constant electronic isodensity using Molekel
software. All the data needed to generate these surfaces and elec-
trostatic potentials were calculated with Gaussian 03 software.
13. Araki, K.; Hashimoto, N.; Otsuka, H.; Shinkai, S. J. Org. Chem. 1993, 58, 5958.
14. Yamato, T.; Zhang, F.; Tsuzuki, H.; Miura, Y. Eur. J. Org. Chem. 2001, 1069.
15. Arnaud-Neu, F.; Cremin, S.; Harris, S.; McKervey, M. A.; Schwing-Weill, M. J.;
Schwinte´, P.; Walker, A. J. Chem. Soc., Dalton Trans. 1997, 329.
16. Matsumoto, T.; Nishio, S.; Takeshita, M.; Shinkai, S. Tetrahedron 1995, 51, 4647.
17. Yamato, T.; Zhang, F. J. Inclusion Phenom. Macrocyclic Chem. 2001, 39, 55.
18. Marcos, P. M.; Fe´lix, S.; Ascenso, J. R.; Segurado, M. A. P.; Pereira, J. L. C.; Kha-
zaeli-Parsa, P.; Hubscher-Bruder, V.; Arnaud-Neu, F. New J. Chem. 2004, 28, 748.
19. Marcos, P. M.; Ascenso, J. R.; Cragg, P. J. Supramol. Chem. 2007, 19, 199.
20. Marcos, P. M.; Fe´lix, S.; Ascenso, J. R.; Segurado, M. A. P.; Mellah, B.; Abidi, R.;
Hubscher-Bruder, V.; Arnaud-Neu, F. Supramol. Chem. 2006, 18, 285.
21. Marcos, P. M.; Fe´lix, S.; Ascenso, J. R.; Segurado, M. A. P.; Thue´ry, P.; Mellah, B.;
Michel, S.; Hubscher-Bruder, V.; Arnaud-Neu, F. New J. Chem. 2007, 31, 2111.
22. Pedersen, C. J. Am. Chem. Soc. 1970, 92, 391.
23. Gutsche, C. D.; Bauer, L. J. J. Am. Chem. Soc. 1985, 107, 6052.
24. Cotton, S. A. C. R. Chim. 2005, 8, 129.
25. Marcos, P. M.; Ascenso, J. R.; Segurado, M. A. P.; Cragg, P. J.; Hubscher-Bruder, V.;
Arnaud-Neu, F. manuscript in preparation.
26. Gutsche, C. D. In Calixarenes; Stoddart, J. F., Ed.; The Royal Society of Chemistry:
Cambridge, 1989.
The relative complexation energies (DEc corresponding to the
process 1bþMnþ/[1bꢀMnþ], where Mnþ represents each one
of the cations) were calculated as the energy difference at 298 K
between the final energy of the complexes and the sum of the
energies of 1b and the cation in study, at the same theoretical
level.35
27. Arduini, A.; Pochini, A.; Reverberi, S.; Ungaro, R.; Andreetti, G.; Ugozzoli, F.
Tetrahedron 1986, 42, 2089.
28. Sheehan, R.; Cragg, P. J. Supramol. Chem. 2008, 20, 443.
References and notes
29. Lee, J. D. Concise Inorganic Chemistry; Wiley-Blackwell: Oxford, 1999.
30. Marcos, P. M.; Ascenso, J. R.; Segurado, M. A. P.; Pereira, J. L. C. J. Inclusion
Phenom. Macrocyclic Chem. 2002, 42, 281.
1. Gutsche, C. D. In Calixarenes Revisited; Stoddart, J. F., Ed.; The Royal Society of
Chemistry: Cambridge, 1998.
31. Cragg, P. J.; Allen, M. C.; Steed, J. W. Chem. Commun. 1999, 553.
32. Spartan ’06, Wavefunction Inc., Irvin, CA 92612, USA.
2. Calixarenes in Action; Mandolini, L., Ungaro, R., Eds.; Imperial College Press:
London, 2000.
33. Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman,
J. R.; Montgomery, J. A., Jr.; Vreven, T.; Kudin, K. N.; Burant, J. C.; Millam, J. M.;
Iyengar, S. S.; Tomasi, J.; Barone, V.; Mennucci, B.; Cossi, M.; Scalmani, G.; Rega,
N.; Petersson, G. A.; Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.;
Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Klene, M.; Li,
X.; Knox, J. E.; Hratchian, H. P.; Cross, J. B.; Adamo, C.; Jaramillo, J.; Gomperts, R.;
Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.;
Ayala, P. Y.; Morokuma, K.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Zakrzewski,
V. G.; Dapprich, S.; Daniels, A. D.; Strain, M. C.; Farkas, O.; Malick, D. K.; Rabuck, A.
D.; Raghavachari, K.; Foresman, J. B.; Ortiz, J. V.; Cui, Q.; Baboul, A. G.; Clifford, S.;
Cioslowski, J.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.;
Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.;
Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen, W.; Wong, M. W.; Gonzalez, C.;
Pople, J. A. Gaussian 03, Revision B.03; Gaussian: Pittsburgh, PA, 2003.
34. Flu¨ kiger, P.F. Molekel, Molecular Visualisation Software; University of Geneva
(http://www.bioinformatics.org/molekel/wiki/).
3. Calixarenes 2001; Asfari, Z., Bo¨hmer, V., Harrofield, J., Vicens, J., Eds.; Kluwer
Academic: Dordrecht, 2001.
4. McKervey, M. A.; Schwing-Weill, M. J.; Arnaud-Neu, F. In Comprehensive Su-
pramolecular Chemistry; Lehn, J. M., Gokel, G. W., Eds.; Elsevier: Oxford, 1996;
Vol. 1, pp 537–603.
5. Arnaud-Neu, F.; McKervey, M. A.; Schwing-Weill, M. J. In Calixarenes 2001;
Asfari, Z., Bo¨hmer, V., Harrowfield, J., Vicens, J., Eds.; Kluwer Academic: Dor-
drecht, 2001; pp 385–406.
6. Ludwig, R. Fresenius’ J. Anal. Chem. 2000, 367, 103.
7. Roundhill, D. M.; Shen, J. Y. In Calixarenes 2001; Asfari, Z., Bo¨hmer, V., Har-
rowfield, J., Vicens, J., Eds.; Kluwer Academic: Dordrecht, 2001; pp 407–420.
8. Arnaud-Neu, F. Chem. Soc. Rev. 1994, 235.
9. Arnaud-Neu, F. Radiochim. Acta 2003, 91, 659.
10. Masci, B. In Calixarenes 2001; Asfari, Z., Bo¨hmer, V., Harrowfield, J., Vicens, J.,
Eds.; Kluwer Academic: Dordrecht, 2001; pp 235–249.
11. (a) Shokova, E. A.; Kovalev, V. V. Russ. J. Org. Chem. 2004, 40, 607; (b) Shokova,
E. A.; Kovalev, V. V. Russ. J. Org. Chem. 2004, 40, 1547.
35. Bernardino, R. J.; Cabral, B. J. C. Supramol. Chem. 2002, 14, 57.