Cation Complexation by Chemically Modified Calixarenes. 11. Complexation and Extraction of Alkali Cations by Calix[5]- And -[6]arene Ketones. Crystal and Molecular Structures of Calix[5]arene Ketones and Na+ and Rb+ Complexes

Article abstract of DOI:10.1021/jo9713114

A series of four calix[5]arenes and three calix[6]arenes (R-calixarene-OCH₂COR¹) (R = H or Bu^t) with alkyl ketone residues (R¹ = Me or Bu^t) on the lower rim have been synthesized, and their affinity for complexation of alkali cations has been assessed through phase-transfer experiments and stability constant measurements. The conformations of these ketones have been probed by ¹H NMR and X-ray diffraction analysis, and by molecular mechanics calculations. Pentamer 3 (R = R¹ = Bu^t) possesses a symmetrical cone conformation in solution and a very distorted cone conformation in the solid state. Pentamer 5 (R = H, R¹ = Bu^t) exists in a distorted 1,2-alternate conformation in the solid state, but in solution two slowly interconverting conformations, one a cone and the other presumed to be 1,2-alternate, can be detected. X-ray structure analysis of the sodium and rubidium perchlorate complexes of 3 reveal the cations deeply encapsulated by the ethereal and carbonyl oxygen atoms in distorted cone conformations which can be accurately reproduced by molecular mechanics calculations. The phase-transfer and stability constant data reveal that the extent of complexation depends on calixarene size and the nature of the alkyl residues adjacent to the ketonic carbonyls with tert-butyl much more efficacious than methyl.

Full text of DOI:10.1021/jo9713114

J . Org. Chem. 1998, 63, 489-501
489
Ca tion Com p lexa tion by Ch em ica lly Mod ified Ca lixa r en es. 11.
Com p lexa tion a n d Extr a ction of Alk a li Ca tion s by Ca lix[5]- a n d
-[6]a r en e Keton es. Cr ysta l a n d Molecu la r Str u ctu r es of
Ca lix[5]a r en e Keton es a n d Na + a n d Rb+ Com p lexes
Steven E. J . Bell, J ulie K. Browne, Vickie McKee, M. Anthony McKervey,* J ohn F. Malone,
Maeve O’Leary, and Andrew Walker
School of Chemistry, The Queen’s University, Belfast BT9 5AG, N. Ireland
Francoise Arnaud-Neu, Olivier Boulangeot, Olivier Mauprivez, and Marie-J ose Schwing-Weill
Laboratoire de Chimie-Physique, URA au CNRS 405, ECPM, 1, rue Blaise Pascal,
67000 Strasbourg, France
Received J uly 17, 1997X
A series of four calix[5]arenes and three calix[6]arenes (R-calixarene-OCH2COR1) (R ) H or But)
with alkyl ketone residues (R1 ) Me or But) on the lower rim have been synthesized, and their
affinity for complexation of alkali cations has been assessed through phase-transfer experiments
and stability constant measurements. The conformations of these ketones have been probed by
1H NMR and X-ray diffraction analysis, and by molecular mechanics calculations. Pentamer 3 (R
) R1 ) But) possesses a symmetrical cone conformation in solution and a very distorted cone
conformation in the solid state. Pentamer 5 (R ) H, R1 ) But) exists in a distorted 1,2-alternate
conformation in the solid state, but in solution two slowly interconverting conformations, one a
cone and the other presumed to be 1,2-alternate, can be detected. X-ray structure analysis of the
sodium and rubidium perchlorate complexes of 3 reveal the cations deeply encapsulated by the
ethereal and carbonyl oxygen atoms in distorted cone conformations which can be accurately
reproduced by molecular mechanics calculations. The phase-transfer and stability constant data
reveal that the extent of complexation depends on calixarene size and the nature of the alkyl residues
adjacent to the ketonic carbonyls with tert-butyl much more efficacious than methyl.
The ability of carbonyl-containing substituents on the
lower rim of calixarenes1 to act as ligating podands for
encapsulation of metal and ammonium cations is now
well established.2-4 The majority of studies describe the
use of ester,5-8 amide,9,10 and carboxylic acid deriva-
tives.11 Apart from one published study5,12 the use of
ketonic carbonyl groups as potential binding sites on the
calixarene substructure is unexplored, despite the greater
donicity of ketones as compared with esters. We have
conducted a systematic study of the behavior of ketones
in the pentamer and hexamer series in complexation and
extraction of alkali cations.
The pentamers are particularly attractive members of
the calixarene homology, though they are the least
studied. They possess larger peripheries than their calix-
[4]arene counterparts, and are presumably more mobile,
yet retain the capability to adopt true cone conformations
of C5v symmetry. And although pentamers have not yet
received the same degree of scrutiny regarding their
conformational characteristics, some important features
are beginning to emerge. Gutsche and co-workers13 have
recently completed a comprehensive study of a whole
series of p-tert-butylcalix[5]arene ethers and esters, the
results of which reveal that even though calix[5]arenes
resemble tetramers in having four up/down conforma-
tions (cone, partial cone, 1,2-alternate, 1,3-alternate),
there are many more opportunities for conformational
subtleties associated with the degree of tilt of the
individual aryl subunits within each of these four con-
formational categories. Furthermore, the NMR method
alone may no longer provide a definitive distinction
between the noncone conformations.
X Abstract published in Advance ACS Abstracts, December 1, 1997.
(1) Gutsche, C. D. Calixarenes. In Monographs in Supramolecular
Chemistry; Stoddart, J . F., Ed.; Royal Society of Chemistry: London,
1989; Vol. 1.
(2) Schwing-Weill, M.-J .; McKervey, M. A. In Calixarenes, a Versatile
Class of Macrocyclic Compounds; Vicens, J ., Bohmer, V., Eds.; Kluwer
Academic Publishers: Dordrecht, 1990; pp 149-172.
(3) Schwing-Weill, M.-J .; Arnaud-Neu, F.; McKervey, M. A. J . Phys.
Org. Chem. 1992, 5, 496.
(4) Ungaro, R.; Pochini, A. Reference 2, pp 127-147.
(5) Arnaud-Neu, F.; Collins, E. M.; Deasy, M.; Ferguson, G.; Harris,
S. J .; Kaitner, B.; Lough, A. J .; McKervey, M. A.; Marques, E.; Ruhl,
B. L.; Schwing-Weill, M.-J .; Seward, E. M. J . Am. Chem. Soc. 1989,
111, 8681.
(6) Chang, S. K.; Cho, I. J . Chem. Soc., Perkin Trans. 1 1986, 211.
(7) Arduini, A.; Pochini, A.; Reverberi, S.; Ungaro, R.; Andreetti, G.
D.; Ugozzoli, F.; Tetrahedron 1986, 42, 2089.
(8) Danil de Namor, A. F.; Gil, E.; Losa Tanco, M. A.; Pacheco
Tanaka, D. A.; Pulcha Salazar, L. E.; Schulz, R. A.; Wang, J . J . Phys.
Chem. 1995, 99, 16776.
(9) Arduini, A.; Pochini, A.; Reverberi, S.; Ungaro, R.; Andreetti, G.
D.; Ugozzoli, F. J . Incl. Phenom. 1988, 6, 119.
(10) Arnaud-Neu, F.; Schwing-Weill, M.-J .; Ziat, K.; Cremin, S.;
Harris, S. J .; McKervey, M. A. New. J . Chem. 1991, 15, 33.
(11) Arnaud-Neu, F; Barrett, G.; Harris, S. J .; McKervey, M. A.;
Owens, M.; Schwing-Weill, M. J .; P. Schwinte. Inorg. Chem. 1993, 32,
2644.
Syn th esis of Ca lixa r en e Keton es
The compounds chosen for study were p-tert-butylcalix-
[5]arene pentamethyl ketone 1 and its calix[6]arene
(12) Ferguson, G.; Kaitner, B.; McKervey, M. A.; Seward, E. M. J .
Chem. Soc., Chem. Commun. 1987, 584.
(13) Stewart, D. R.; Krawiec, M.; Kashyap, R. P.; Watson, W. H.;
Gutsche, C. D. J . Am. Chem. Soc. 1995, 117, 586.
S0022-3263(97)01311-X CCC: $15.00 © 1998 American Chemical Society
Published on Web 01/16/1998

490 J . Org. Chem., Vol. 63, No. 3, 1998
Bell et al.
Sch em e 1
F igu r e 1. A view of molecule 3 showing the molecular
conformation.
tert-butyl ketone 3 shares these features to a considerable
degree. The aromatic, methyleneoxy and both tert-butyl
group protons in the 1H NMR spectrum appear as
singlets with an AB pattern for the bridging methylene
groups, suggesting a symmetrical cone conformation.
Solid-state IR analysis, however, revealed two carbonyl
absorptions of unequal intensity at 1699 and 1729 cm-1
,
indicative of a less symmetrical arrangement with the
implication that the symmetrical solution state structure
is the result of a rapid conformation averaging process.
X-ray diffraction (Figure 1) revealed that the molecule
does possess a cone conformation in the solid state and
with the same type of distortion as that found in its
pentaester counterpart 12.14 The spatial void in the
upper rim cavity is partially occupied by inclination of
one tert-butyl group. The conformation of 3 is best
described by the angles that the aromatic rings (A-E)
make with the mean plane of the five macrocyclic ring
methylene groups, viz.: (137°: A), (56°: B), (140°: C),
(101°: D) and (107°: E). An angle greater than 90°
implies the aromatic ring is tilted in a manner that
pitches its attached tert-butyl group away from the calix
cavity while for angles less than 90° the reverse is true.
Inspection of these tilting angles reveals that while ring
B tilts in markedly, the remaining aromatic rings are
inclined outward but by very different amounts. The
inward disposition of ring B necessarily forces adjoining
rings A and C to adopt a more open arrangement in
relation to the remaining macrocyclic aromatic units.
This effect is reflected in the tilting angles of rings A
(137°) and C (140°) where the greater outward tilt
compared to rings D (101°) and E (107°) reduces steric
interaction between the tert-butyl groups on rings A and
C with that on ring B. This structural distortion may
also relieve the conformational strain imposed on the
adjacent bridging methylene carbon atoms by the incli-
nation of ring B. The inter-ring methylene carbons
deviate considerably from coplanarity as a result of
distortion with individual atoms as much as 0.43 Å from
the mean plane.
counterpart 2; p-tert-butylcalix[5]arene penta tert-butyl
ketone 3 and its hexamer counterpart 4; calix[5]arene
penta tert-butyl ketone 5, calix[5]arene pentamethyl
ketone 6, and calix[6]arene hexamethyl ketone 7.
All six compounds were prepared from the appropriate
calixarene (8-11) by exhaustive alkylation with either
chloroacetone or bromopinacolone (Scheme 1) in dry
acetone containing potassium carbonate and potassium
or sodium iodide, the latter to activate the electrophile
through halogen exchange. In the case of pentaketones
3 and 5 with pinacolone moieties on the lower rim, the
products were isolated as K+ complexes. Decomplexation
of 5 was accomplished by treatment with water at 60 °C;
prolonged treatment with aqueous ethanol (3 days under
reflux) was required to release ketone 3 from its complex.
Analytically pure samples of all six ketones were obtained
by recrystallization from ethanol-dichloromethane. 1H
NMR analysis provided some insight into the conforma-
tions of these ketones and their complexes in solution.
The crystal and molecular structures of 1, 3, and 5 and
of the sodium and rubidium perchlorate complexes of 3
have been determined by X-ray diffraction and compared
with structures simulated by molecular mechanics.
Con for m a tion of Ca lixa r en e Keton es
In our earlier work on calix[5]arene acetates of type
12 and 13 we found evidence in solution for the presence
of stable cone conformations of presumed averaged C5v
symmetry.14 X-ray diffraction confirmed the cone con-
formation in the solid state for 12, but further revealed
a distortion of the cone which does not approximate to
any regular geometric arrangement. One of the five aryl
rings tilts inward so that its tert-butyl group overhangs
the cavity on the upper rim, while the remaining four
tilt outward, as is normal, but to different extents. Penta
The pendent ketone side chains adopt very different
spatial arrangements. Ketone chains A, C, D and E take
up positions below the calix cage with ketone arm E
filling the lower rim cavity in much the same way as has
been observed with the pentaethyl ester. The inward
inclination of aryl ring B forces its attached ketone arm
to be directed away from the calix cage. The uncom-
(14) Barrett, G.; McKervey, M. A.; Malone, J . F.; Walker, A.; Arnaud-
Neu, F.; Guerra, L.; Schwing-Weill, M.-J .; Gutsche, C. D.; Stewart, D.
R. J . Chem. Soc., Perkin Trans. 2 1993, 1475.

Cation Complexation by Chemically Modified Calixarenes
J . Org. Chem., Vol. 63, No. 3, 1998 491
F igu r e 2. Iconographic representation and NMR symmetry patterns for noncone conformations of p-tert-butylcalix[5]arene ketones.
plexed pentaketone cavity is thus enclosed by a distorted
pentagonal pyramid composed of the five phenolic oxygen
atoms (O1A-O1E) and the carbonyl oxygen atom (O2E).
The remaining carbonyl oxygen atoms take up random
positions around the lower rim calix cage. Adjacent
phenolic O--O separations around the lower rim vary
between 3.30 Å (O1C-O1D) and 4.51 Å (O1B-O1C)
constants of the assigned partners. On this basis the
ArCH2Ar protons appear as three pairs of doublets at
3.29 (2H, HB), 3.34 (1H, HB′), 3.81 (2H, HB′′), 4.01 (2H,
HA′′), 4.13 (1H, HA′), and 4.21 (2H, HA) ppm. Such a
spectral pattern is commensurate with a partial cone, a
1,2-alternate or a 1,3-alternate conformational orienta-
tion. The OCH2 protons show an unusual arrangement
in this ketone producing two singlet resonances at 4.17
and 4.22 ppm and one pair of doublet resonances at 4.16
and 4.53 ppm in a 1:2:2 ratio, respectively. This implies
that the rigidifying process accompanying conformational
freezing is accentuating the nonequivalent nature of some
of these protons which are illustrated iconographically
in Figures 2a-c for the partial cone, 1,2-alternate, and
1,3-alternate conformers, respectively. The environment
of the OCH2 protons in 1 is such that four of the
nonequivalent hydrogen atoms appear as a singlet at 4.22
ppm rather than a pair of doublets which might have
been expected on the basis of the symmetry arguments
shown in Figure 2. This suggests that these four protons
are in chemically similar environments which is tenta-
tively possible for ketone arms D and C in Figure 2a and
ketone arms B and E in Figure 2c but not for ketone
chains C, D or B, E in Figure 2b. This implies, therefore,
that 1 may assume a partial cone arrangement (Figure
2a), or a 1,3-alternate orientation (Figure 2c), in solution.
The tert-butyl protons appear as three singlets at 1.15,
1.33, and 1.44 ppm in a 2:2:1 ratio, a pattern reminiscent
of a partial cone, 1,2-, or 1,3-alternate species. The
terminal methyl groups on the ketone chains also appear
as three singlets at 0.75, 1.00-1.09 (broadened), and 2.11
ppm in a 1:2:2 ratio. One methyl group occupies the most
shielded position at 0.75 ppm, a feature indicative of a
partial cone conformer. However, a partial cone con-
former would be expected to contain the remaining four
methyl groups in fairly similar environments which
should be reflected in relatively closely separated chemi-
cal shift positions. This is not the case for 1, in which
the other four CH3 protons are widely separated, two sets
being deshielded at 2.11 ppm and the remaining two
much more shielded at 1.00-1.09 ppm. This spectral
arrangement is much more consistent with a 1,2- or 1,3-
alternate conformation. The 13C NMR spectrum (-10 °C)
of 1 displays two downfield ArCH2Ar resonances between
the at 34.08 and 34.22 ppm for the methylene carbons
between the anti-aryl rings and only one upfield ArCH2-
Ar signal at 31.39 ppm for the syn-aryl rings, a pattern
consistent only with a 1,3-alternate conformation. On
highlighting the outward disposition of chain B.
A
molecule of methylene chloride is present within the
crystal lattice, but there are no short intermolecular
contacts to the calixarene. Some of the tert-butyl groups
of 3 were disordered with high thermal parameters and
where appropriate were restrained using the computa-
tional features (SAME, SADI) available in SHELXL 93.15
Changing the alkyl residue adjacent to the carbonyl
group in ketone 3 from tert-butyl to methyl as in 1 has a
profound effect on the conformations adopted in the
pentamer series. The 1H NMR spectrum of 1 at room
temperature consists of a series of broad, partly resolved
signals indicative of rapid fluxional behavior. It is
probable that conformational interconversion in 1 pro-
ceeds via the oxygen through the annulus pathway where
the ketonic methyl chains are sufficiently small to pass
through the cavity of the pentamer, unlike the sterically
more consuming ketonic tert-butyl chains in nonfluxional
3. At low temperatures in solution (-10 °C) this fast
fluxional motion in 1 is suppressed and a series of well-
1
resolved H NMR resonances result, from which a con-
1
formational assignment was tentatively made. The H
NMR spectrum of 1, at -10 °C, shows five sets of aryl H
resonances in a 1:1:1:1:1 ratio of two doublets (7.10, J )
1.8 Hz, 7.13 ppm, J ) 1.1 Hz) and three singlets (7.19,
7.37, and 7.58 ppm), illustrating immediately that 1 does
not adopt the same symmetrical cone arrangement that
was observed for 3. The ArCH2Ar bridging methylene
region is very complex, a problem compounded by the fact
that the OCH2 ketonic group protons also appear in this
area. To make an assignment it was assumed that the
OCH2 protons would overlap only with the more down-
field HA signals of the ArCH2Ar methylene protons and
would not interfere with the much more shielded, upfield,
HB of the ArCH2Ar methylene hydrogens. By employing
1
a series of decoupling experiments and a 2D H COSY
NMR spectrum of 1, it was possible to assign HB protons
coupled to their methylene HA counterparts, a process
ultimately verified by the close similarity of the coupling
(15) Sheldrick, G. M. University of Go¨ttingen, 1993. SHELXL 93.

492 J . Org. Chem., Vol. 63, No. 3, 1998
Bell et al.
(1H, HB′), 3.79 (2H, HB′′), 3.98 (2H, HA′′), and 4.12 ppm
(1H, HA′ and 2H, HA overlapping). This spectral arrange-
ment is consistent with a partial cone, 1,2-, or 1,3-
alternate conformation. The OCH2 protons appear as a
singlet (4.23 ppm), a pair of doublets (4.20 and 4.53 ppm),
and a multiplet (4.04 ppm exhibiting AB coupling) in a
1:2:2 ratio, respectively. Such a pattern is more consis-
tent with a partial cone or 1,3-alternate arrangement (see
Figure 2) where two sets of OCH2 protons lie in chemi-
cally similar environments. The terminal methyl groups
of the ketone chains appear as three singlets at 0.56, 1.43
(broadened), and 2.14 ppm in a 1:2:2 ratio, respectively.
The extreme upfield position (0.56 ppm) of just one
methyl group is consistent with a partial cone conforma-
tion. This observation is supported by the -10 °C 13C
NMR spectrum of 6 which reveals just one downfield
resonance at 36.71 ppm indicative of aligned anti aro-
matic rings and two upfield resonances at 26.95 and
28.82 ppm indicative of syn aromatic rings, suggesting a
partial cone or a 1,2-alternate conformer. Taking all the
spectral data into account, a partial cone isomer is chosen
as the most likely conformation. Solid-state IR spectro-
scopic analysis of 6 revealed two carbonyl stretches at
1723 and 1744 cm-1 in accord with the solution state
structure.
Ketone 5, the para-hydrogen analogue of 3, was first
isolated as a K+ complex (vide supra). The 1H NMR
spectrum of the complex clearly pointed to the presence
of a symmetrical cone conformation in solution. In
contrast, the spectrum of 5 in the uncomplexed state at
room temperature consisted of a complicated series of
sharp signals attributable to the presence of two fixed
(on the NMR time scale) conformations, a cone (17%) and
a noncone (83%). However, the isolation of one K+
complex indicated that conformational interconversion
was possible, though at a rate sufficiently slow on the
NMR time scale to distinguish two conformational iso-
mers. 1H NMR measurements at higher temperatures
did reveal the onset of coalescence of the signals repre-
senting the two conformations. This conformational
isomerism in 5 raises the question as to the nature of
the molecular motion responsible: a lower rim through
the annulus motion, as is the case with calix[4]arene
conformers, or an upper rim through the annulus motion?
Although the latter is not observed with calix[4]arenes,
even with para substituents as small as hydrogen, it has
been observed with calix[6]arenes, not just with para-
hydrogen substituents but also with para-tert-butyl sub-
stituents. Gutsche’s view from inspection of molecular
models that calix[5]arenes may fall between these two
situations is supported by the behavior of pentaketones
3 and 5. Thus, at normal temperatures calix[5]arenes
with p-tert-butyl groups can only interconvert, if at all
(cf 3), via the lower rim motion through the annulus.
With 5, in contrast, the bulky tert-butyl substituent on
the lower rim will surely inhibit this motion, but since
slow interconversion does occur we can conclude that it
does so by rotation of the upper rim para-hydrogen
substituent through the annulus.
F igu r e 3. A view of molecule 1 showing the molecular
conformation.
the basis of these spectral observations, the 1,3-alternate
isomer is deemed most probable for compound 1. Solid-
state IR spectroscopic analysis of 1 shows two carbonyl
absorptions at 1715 and 1738 cm-1 indicative of a
structural form which places the carbonyl oxygens in
nonidentical environments in accord with the solution
state structure. In contrast to our analysis of the solution
conformation of ketone 1 we found by X-ray diffraction
that in the solid state the arrangement is that of a very
distorted 1,2-alternate conformation (Figure 3). The
asymmetric unit comprises the calixarene and one mol-
ecule of dichloromethane solvate; there is no evidence of
any interaction between calixarene and solvate. The
calixarene conformation is irregular and can be described
in terms of the tilt angles of the phenyl rings to the plane
of the five macrocyclic methylene groups: (92°: A, 141°:
B, 87°: C, 7°: D and 86°: E). Rings A, C, and E are
almost perpendicular to the cavity, B is tilted with the
tert-butyl group outward and D almost coplanar with the
plane of the five methylene groups. Ring E is inverted
so that its tert-butyl group lies on the opposite rim to
those of rings A, B, and C.
Ketone 6 is also a methyl derivative of a calix[5]arene
but lacks a p-tert-butyl substituent. The 1H NMR
spectrum of 6 shows that this molecule is also fluxional
at room temperature. Aryl rotation in 6 can be ac-
complished by an oxygen through the annulus pathway
or a para hydrogen through the annulus pathway. The
former is supported by the behavior of 1, where it is
known that rotation of the para substituent through the
calix is precluded. Nevertheless, the molecule is flux-
ional. At -10 °C the conformational mobility of ketone
1
6 is suppressed and the H NMR spectrum now appears
as a series of sharp signals whose multiplicity is incon-
sistent with a cone conformation. The calix aromatic
protons appear as a series of coupled signals at 6.77-
7.57 ppm providing little information concerning the
precise conformation of 6. Again, an assumption, similar
to that made for ketone 1, was employed, regarding the
assignment of the ArCH2Ar methylene region. The HB
ArCH2Ar protons in 6 were thus judged to be the most
shielded hydrogens in this region. With the combination
The major conformational isomer of 5 displayed a very
complex arrangement of sharp signals with three tert-
butyl singlets (ratio 2:2:1) and a series of doublets and
multiplets which could be accommodated by either a 1,2-
alternate, a 1,3-alternate, or a partial cone conformation
without offering any compelling choice among the three.
The minor isomer of 5 was readily identified as cone from
1
of a series of decoupling experiments and a 2D H COSY
NMR analysis, the corresponding HA coupling partners
could be identified. On this basis the ArCH2Ar protons
again appear as three pairs of doublets, verifiable by
corresponding coupling constants, at 3.33 (2H, HB), 3.40

Cation Complexation by Chemically Modified Calixarenes
J . Org. Chem., Vol. 63, No. 3, 1998 493
terminations in methanol. Unfortunately, the latter,
which provide a more precise expression of the thermo-
dynamics of the complexation process, could not be
extended to all ketones because of solubility limitations.
Preliminary extraction experiments revealed a very large
kinetic effect in the behavior of the tert-butyl ketone 3.
Whereas all the other ketones reached extraction equi-
librium within minutes, it was necessary to allow several
hours with 3 before reaching reproducible extraction
levels. The data are summarized in Tables 1 and 2 which
also contain for comparison the published data for
pentaesters 12 and 13, tetramethyl ketone 14, and
hexaester 15. These data, illustrated graphically in
Figures 6 and 7, reveal dramatic effects of calix size and
substitution. The most striking aspect of the extraction
data for alkali cations is the strong dependency on the
nature of the alkyl residues adjacent to the ketonic
carbonyls on the lower rim in the pentamer derivatives
and, to a lesser extent, on the nature of the para
substituent on the upper rim. The combination of tert-
butyl substituents on both rims of the calixarene as in 3
produces the highest extraction levels for all alkali
cations with, however, little discrimination between them
after Li+. The levels of extraction approach those ob-
served earlier with tetraamides10 and pentaamides.22 Of
the two, tert-butyl substitution on the lower rim appears
to be the dominant influence. Thus, while replacing the
para tert-butyl groups by hydrogen substituents as in 5
has the effect of reducing the overall extraction levels,
e.g. Li+ down from 58.4 to 8.3%, with leveling off for the
larger cations now commencing at K+, replacing the lower
rim tert-butyl groups with methyl groups as in 1 has a
much greater effect, reducing the extraction levels to a
maximum of 11% (for Rb+). The combination of para-
hydrogen substituents on the upper rim and methyl
substituents on the lower rim in 6 leads to the lowest
levels of extraction in the entire pentaketone series. This
contrasts with the effect of methyl ketones in the tet-
ramer series, cf. compound 14, where there is a strong
sodium peak selectivity (42.3%). The reasons for the poor
extraction performance of compounds 1 and 6 may lie in
their conformational mobility and their inherent lack of
any elements of preorganization. Tetraketone 14, in
contrast, is already in a fixed cone conformation prior to
complexation. In the hexamer series ketonic binding
groups are uniformly very inefficient with very low levels
of alkali extraction for compounds 2, 4, and 7.
F igu r e 4. A view of molecule 5 showing the molecular
conformation.
1
the H NMR pattern. However, when crystalline 5 was
analyzed by X-ray diffraction, only one conformation, the
1,2-alternate, was found in the solid state. But interest-
ingly, when crystals of 5 were dissolved in CDCl3, the
1H NMR spectrum again revealed the presence of both
cone and (presumed) 1,2-alternate conformations in a 83:
17 ratio, confirming that interconversion does occur,
leading eventually, in the presence of K+ ion, to a cone
complex. The structure adopted in the solid state by 5
is shown in Figures 4 and 5a. The compound assumes a
very distorted 1,2-alternate orientation, implying that
crystallization induces formation of this isomeric form
preferentially. Two of the aromatic rings (groups A and
D) in Figure 4 lies in a flattened orientation in compari-
son to the relatively less tilted arrangements of the
remainder. Such extreme distortion from the more
conventional up/down arrangement of aryl rings is
reflected in the angular tilt of each aromatic unit from
the macrocyclic methylene mean plane, viz.: (76°: A),
(122°: B), (73°: C), (155°: D) and (56°: E). An angle of
<90° implies that the aromatic para proton on each ring
is directed into the calix cavity defined by the macrocyclic
rim in question. An angle of >90° reflects the reverse
situation. It must be noted, though, that the five bridging
methylene groups in a noncone structure such as this are
not coplanar, and significant deviations from the mean
plane (as much as 0.66 Å) are prevalent. Nevertheless,
the tilting angles provide a useful assessment of the
distorted nature of this conformation. The structure as
a whole provides a visual representation of the cavity size
in relation to the podands.
For stability constants in methanol, the log â values
(Table 2) for 5 reflect the trends in extraction with the
higher values (5.67-5.84) associated with the larger
cations. Unfortunately, comparative data for the double
tert-butyl combination in 3 could not be obtained due to
solubility limitations in methanol. Nevertheless, the
beneficial effects of lower rim tert-butyl groups are again
Since the conformation of the major conformer of
ketone 5 could not be unequivocally distinguished be-
tween a partial cone, a 1,2- or a 1,3-alternate isomer, it
is probable that the solution state mirrors the solid state,
and on this basis a 1,2-alternate arrangement is also
likely to be adopted preferentially in solution. The
remaining ketones, hexamers 2 and 7, were conforma-
tionally mobile at room temperature.
(16) Guilbaud, P.; Varnek, A.; Wipff, G. J . Am. Chem. Soc. 1993,
115, 8298. Varnek, A.; Wipff, G. J . Phys. Chem. 1993, 97, 10840.
(17) Grootenhuis, P. D. J .; Kallman, P. A.; Groenen, L. C.; Rein-
houdt, D. N.; van Hummel, G. J .; Ugozzoli, F.; Andreetti, G. D. J . Am.
Chem. Soc. 1990, 112, 4165-4176.
(18) Stewart, D. R.; Gutsche, C. D. Org. Prep. Proc. Intl. 1993, 25,
137.
(19) Gutsche, C. D.; Iqbal, M. Org. Synth. 1990, 68, 234.
(20) Pedersen, C. J . Fed. Proc. Fed. Am. Soc. Expl. Biol. 1968, 27,
1305.
(21) Vetrogon, V. I.; Lukyanenko, N. G.; Schwing-Weill, M.-J .;
Arnaud-Neu, F. Talanta 1994, 12, 2105-2112.
(22) Arnaud-Neu, F.; Barrett, G.; Fanni, S.; Marrs, D.; McGregor,
W.; McKervey, M. A.; Schwing-Weill, M.-J .; Vetrogon, V. I.; Wechsler,
S. J . Chem. Soc., Perkin Trans 2 1995, 453.
Com p lexa tion Stu d ies. Discu ssion
Physicochemical measurements of the efficacy of all
seven ketones in complexation of alkali metal cations
were conducted using biphase extraction of picrates from
water into dichloromethane and stability constant de-

494 J . Org. Chem., Vol. 63, No. 3, 1998
Bell et al.
F igu r e 5. Stereoviews of energy minimized structures of molecule 5 (b and c) and, for comparison, stereoviews (a) of the X-ray
structure of 5.
apparent from comparison of the log â values for 5 with
those of 6, where the latter is consistently the weaker
binding ligand. In the pentaester derivatives 12 and 13,
the higher levels of extraction associated with the tert-
butyl ester 13 are also reflected in stability constants.14
Obviously, factors other than the lipophilicity of the
ligands must be invoked to explain this behavior. As
shown by the thermodynamic parameters of complex-
ation of Cs+ (Table 3), the important increase of stability
(∆log â ) 2.74) accompanying the replacement of methyl
(compound 6) by tert-butyl groups (compound 5) is the
result of a tremendous increase in the enthalpy change
factors such as (i) the higher basicity of the carbonyl
oxygens due to the presence of tert-butyl groups, more
donating than methyl groups, (ii) the lower solvation of
the ligand due to steric hindrance, and (iii) a better
preorganization of the ligand. The small entropy term,
even slightly negative as compared to the value for Cs6+,
would suggest the importance of solvation effects.
On the other hand, the log â values for ketone 1 are
also very consistent with the performance of this ketone
in extraction, both measurements showing a slight
preference for Rb+. However, comparison between 1 and
6 reveals a slight preference for 1 with respect to the
larger cations, except Cs+, for which thermodynamic
parameters are very similar.
from -∆Hc ) 14 to -∆Hc ) 35 kJ mol-1
enthalpy contribution for Cs5+ may be related to different
. The large

Cation Complexation by Chemically Modified Calixarenes
J . Org. Chem., Vol. 63, No. 3, 1998 495
there data for any metal complex of a calixarene ketone
or ester. The structure adopted by the 3‚NaClO4 com-
plex, recrystallized from methanol, is shown in Figure
8.. In the solid state the complex adopts a distorted cone
conformation, although to a lesser extent than the free
ligand 3, where one of the aromatic tert-butyl groups is
oriented toward the hydrophobic upper rim cavity. The
tilt angles of the aromatic rings to the bridging methylene
mean plane again provides the best structural assess-
ment with angles of (144°: A), (102°: B), (99°: C), (134°:
D) and (74°: E). Aside from ring E the remaining
aromatic rings all have their tert-butyl groups pitched
away from the upper rim cavity. The inclination of E
again forces adjacent rings A and D to accentuate their
outward tilt, although this effect is much more pro-
nounced for ring A than for ring D. Adjacent phenolic
O--O separations defining the boundaries of the lower rim
calix cage vary between 3.09 Å (O1A-O1E) and 4.63 Å
(O1D-O1E) highlighting the concomitant inward/out-
ward distortion. The Na+ cation is held in a spatial
position between the phenolic oxygen plane and the
ketonic carbonyl oxygen atoms. This relatively small
guest lies off-center from the lower rim calix cage because
the oxygen atoms on ring E do not participate in binding
and the enclathration cage is instead composed of the
remaining ketonic chains. Of these chains, only the
phenolic oxygen atom attached to ring A is precluded
from electrostatic interaction to Na+. Its position in the
Na+ coordination cage is instead taken by the oxygen
atom of the included methanol solvent molecule with a
Na+--oxygen separation of only 2.33 Å. The Na+ cation
is thus surrounded by an 8 coordinate oxygen cage. The
calixarene oxygen--Na+ separations in this cage vary
between 2.28 Å (Na1-O2D) and 2.81 Å (Na1-O1D). The
perchlorate counteranion is situated just outside the
hydrophilic lower rim cavity and was found to be disor-
dered, unequally, over two sites. Again, as in the free
ligand case, some of the tert-butyl groups were disordered
and where possible were either refined as PART
(SHELXL-93)15 atoms or restrained in a tetrahedral
environment.
F igu r e 6. Percentage cation extracted (% E) from water into
dichloromethane of alkali picrates vs the cation ionic radius:
influence of calixarene size in the ketone series.
X-r a y Cr ysta llogr a p h ic An a lysis of P en ta -ter t-
F igu r e 7. Percentage cation extracted (% E) from water into
dichloromethane of alkali picrates vs cation ionic radius:
substituent effects.
bu tyl Keton e 3‚Rb+ClO4 Com p lex. X-ray crystal-
-
-
lographic analysis of the Rb+ClO4 complex of 3 was
accomplished after recrystallization of the complex from
ethanol. The structure of the complex is shown in Figure
9.
Ta ble 1. P er cen ta ge Extr a ction (%E) of Alk a li P icr a tes
in to CH2Cl2 a t 20 °C. Ar ith m etic Mea n of Sever a l
Exp er im en ts, Sta n d a r d Devia tion on th e Mea n : σn -1 e 1
In the solid state the structure again adopts a now
classical distorted cone conformation with one tert-butyl
aryl ring inclined to the other upper rim inclination. It
is thus apparent from both the Na+ and Rb+ complexes
of 3 that only small universal conformational changes in
the free form of the penta-tert-butyl ketone are needed,
to effect cationic inclusion. This was presumed on the
basis of the high broad extraction capabilities (Table 1)
shown by 3, but is now confirmed through solid-state
structural analysis. The tilt angles of the aromatic rings
to the macrocyclic ring methylene mean plane again
provide the best means of structural definition viz.: (110°:
A), (112°: B), (129°: C), (51°: D) and (127°: E). Apart from
the sharp inward inclination of D, the remaining aro-
matic rings all assume outward tilts with their attached
tert-butyl groups pitched away from the calix cavity by
varying degrees. Unlike the previous Na+ complex case,
the inward inclination of ring D in this case is more
pronounced, mirroring the tilting extent evidenced in the
compounds
Li+
Na+
K+
Rb+
Cs+
5-Me ketone 1
6-Me ketone 2
0.5
0
58.4
0.8
8.3
0
0
8
25
6.5
0
1.7
0.4
80
2.0
31.7
0.3
0
33
69
42.3
2.2
10.2
0.6
81.9
2.1
58
2.8
0
47
11.2
0.5
81
2.1
61.1
3.4
0
51
72
7.6
3.3
7.8
1.7
79
3.1
61.8
2.6
4.6
51
5-tert-butyl ketone 3
6-tert-butyl ketone 4
p-H-5-tert-butyl ketone 5
p-H-5Me ketone 6
p-H-6Me ketone 7
5-Et estera 12
5-tert-butyl estera 13
4-Me ketone 14
6-tert-butyl ester 15
74
8.7
2.1
68
9.6
8.4
a
Reference 14.
X-r a y Cr ysta llogr a p h ic An a lysis of P en ta -ter t-
bu tyl Keton e 3‚Na +ClO4-. The isolation of good quality
crystals of the sodium and rubidium perchlorate com-
plexes of ketone 3 provide an opportunity to probe their
solid-state structures by X-ray diffraction. No solid-state
structural data exists for calix[5]arene complexes, nor are

496 J . Org. Chem., Vol. 63, No. 3, 1998
Bell et al.
Ta ble 2. Log â Va lu esa Ca tion s Com p lexa tion in Meth a n ol, T ) 25 °C, I ) 0.1 m ol d m -3 Et4NClO4 or Et4NCl
compounds
Li+
Na+
<1.5d
4.8 ( 0.2d
<1.5d
4.4
5.1
5.1
K+
Rb+
Cs+
1
5
6
<1.5d
<1.5d
<1.5d
1.0
1.5
2.7
3.34 ( 0.06d
5.67 ( 0.01e
3.1 ( 0.2d
5.3
3.6 ( 0.2d
5.81 ( 0.08e
3.2 ( 0.3d
5.6
3.22 ( 0.08d
5.84 ( 0.03e
3.1 ( 0.2d
5.5
12b
13b
14c
6.1
3.1
5.8
3.6
5.3
3.1
a
b
d
Arithmetic mean of several experiments, standard deviation on the mean: σn-1
.
Reference 14. c Reference 5. Spectrophotometric
determination. e Potentiometric determination.
Ta ble 3. Th er m od yn a m ic P a r a m eter s of Com p lexa tion
interaction between the tert-butyl groups on rings C and
E to that on ring D. The adjacent phenolic O--O separa-
tions on the lower rim boundary vary between 3.25 Å
(O1B-O1C) and 4.41 Å (O1D-O1E) and serve to high-
light the uneven nature of the aromatic tilting angles.
The extremeties of phenolic O--O separation are similar
in magnitude to those in the free ligand pentaketone
(3.30 and 4.51 Å) implying that very little structural
change accompanies Rb+ inclusion. This feature is
further supported by the similarity between the respec-
tive aromatic ring tilting angles. The Rb+ cation is held
in a spatial position closer to the phenolic oxygen plane
than to the mean plane defined by the ketonic carbonyl
oxygen atoms. Unlike the Na+ complex, the Rb+ cation
occupies a more central position in the lower rim calix
cage despite the fact that the oxygen atoms on ring D do
not participate in binding. Figure 10 shows an overlay
of the two X-ray structures showing the relative position
of the Rb+ and Na+ cation within the cavity. As was
apparent for the Na+ case, the enclathration cage is
composed of, and defined by, the remaining ketonic
podands. Of the remaining arms, the carbonyl oxygen
atoms (O2) attached to rings A and B are not included
in the complexation process. The remaining phenolic and
carbonylic oxygen atoms define a 6 coordinate complex-
ation cage for Rb+. The calixarene--Rb+ separations in
this cage vary between 2.85 and 2.97 Å. The perchlorate
counteranion is not disordered in this case and occupies
a position beside the upper rim of the macrocycle. Some
of the tert-butyl groups possessed relatively large thermal
parameters and were either refined as disordered atoms
partitioned between two sites or tetrahedrally restrained.
X-ray diffraction has thus provided an insight into the
subtle conformational modifications imposed on the
penta-tert-butyl ketone by alkali cation encapsulation.
in Meth a n ol of Cs+ Com p lexes, T ) 25 °C
1
5
6
-∆Gc (kJ mol-1
-∆Hc (kJ mol-1
)
)
)
18.4 ( 0.5
15 ( 2
3 ( 2
33.3 ( 0.2
35 ( 0.1
-2 ( 1
18 ( 1
14 ( 2
4 ( 3
T∆Sc (kJ mol-1
∆Sc (J K-1 mol-1
)
10 ( 7
-7 ( 3
13 ( 10
F igu r e 8. A view of the Na+ complex of 3 showing the
molecular conformation.
Molecu la r Mech a n ics Ca lcu la tion s. Having ac-
quired some structural data from X-ray diffraction and
1H NMR spectra, it became of interest to see if molecular
mechanics calculations (MM+) could provide further
insight into these calix[5]arene systems, especially the
complexed state. Wipff and co-workers16 have already
demonstrated, in the most sophisticated studies to date,
how molecular dynamics can be used to probe the
relationship between ligand conformation, cation, and
solvent in calix[4]arene amide complexes of alkali cations
and Eu3+. For us the first step was to establish that MM
calculations could reliably predict structures of the larger
and conformationally more heterogeneous calix[5]arenes
of lower symmetry.
The energy-minimized MM+ calculated and X-ray
structures of ketone 3 are shown in Figure 11a,b. The
MM+ structure is adequately close to that of the crystal
structure; in particular the model reproduces well the
inward tilt of one of the aryl rings and, due to the
mechanical coupling present in the system, the increased
F igu r e 9. A view of the Rb+ complex of 3 showing the
molecular conformation.
free ligand pentaketone. Presumably the lack of solvent
inclusion in the upper rim hydrophobic cavity of the Rb+
complex permits a similar degree of inward tilt to the
free ligand. Again, the inward distortion of ring D on
the upper rim forces the adjoining aromatic rings C and
E to assume more open orientations reflected in greater
outward tilts of 129° for C and 127° for E in relation to
rings A and B. Such an effect alleviates the steric

Cation Complexation by Chemically Modified Calixarenes
J . Org. Chem., Vol. 63, No. 3, 1998 497
F igu r e 10. Superimposed stereoviews of the Na+ and Rb+ complexes of 3.
outward tilt of the two adjacent rings. It should be
emphasized that the starting geometry for this conforma-
tion was built with all aryl rings approximately parallel
and upright and that minimization to the tilted structure
did not depend on a biased starting geometry. It is clear
from closer inspection of the structure that the inward
tilt of the unique phenoxy ring must be accompanied by
a very substantial movement of its tert-butyl group
toward the center of the calixarene. In fact, space-filling
models show that the remaining four tert-butyl groups
form a cavity on the lower rim into which the fifth tert-
butyl group fits exactly.
modeling the Na+ complex of 3 with no methanol in the
cavity and then to explore the effect of adding the
methanol separately. The energy-minimized structures
of the Na+ complex with and without methanol are shown
in Figure 11e,f. In the modeled structure, the complex
solvent molecule is in approximately the correct position
as is the Na+ cation. Closer comparison of the modeled
and X-ray structures shown in Figure 11f,g shows the
model does not exactly reproduce the X-ray structure,
because the reorientation of the inward tilting ring is
slightly underestimated, as is the additional tilt of the
adjacent ring, but the overall fit to the X-ray structure
is satisfactory. This gives us some confidence that the
modeling procedure is reliable and that it is reasonable
to draw conclusions from the model structure for the Na+
complex in the absence of coordinated solvent. This
structure (shown in Figure 11e) retains the overall
geometry of the Rb+ complex, the main difference being
that the Na+ ion lies farther below the phenolic oxygens
than does the Rb+. This implies first that the large
difference in the positions of the metal cations observed
in the crystal structures (illustrated in Figure 10) is not
simply due to the presence of the encapsulated solvent
in the Na+ complex, but a significant part of the differ-
ence must arise from a preference of Na+ for the more
electron-rich carbonyl groups over the ether linkages,
irrespective of other factors. The position of the Na
cation is slightly more central in the model with MeOH
bound, so that the solvent inclusion does have a second-
ary effect on the metal ion position.
Second, changing the metal cation without also adding
a solvent molecule does not cause a large change in the
orientation of the inward tilting aryl ring in the model
structures. However, incorporation of a methanol solvent
molecule into the Na+ complex does force this aryl ring
outward so that it adopts the more upright orientation
observed in the X-ray structure (Figure 8). The two main
differences in the structure of the Rb+ and Na+ complexes
can thus be attributed to two separate influences: the
metal cations will bind at different positions to the
undistorted calixarene, while the encapsulated solvent
causes the ligand structure to distort.
The fact that the uncomplexed ketone 5 of the para-
dealkylated calixarene adopts a noncone conformation in
the solid state and a mixture of conformations in solution,
prompted us to extend the modeling studies to this
system. An extensive search of the conformers which this
molecule can adopt revealed many local minima on the
potential energy hypersurface, similar to the situation
The Rb+ complex of 3 has a very similar geometry to
that of the metal-free ligand, indicating a high degree of
preorganization for complexation. The MM+ structure
of the Rb+ complex is shown in Figure 11c, along with
the actual crystal structure (Figure 11d). It is clear that
there is again good agreement between the two struc-
tures, despite the very crude nature of the molecular
mechanics calculation. The similarity in ligand confor-
mation between the free calixarene 3 and its Rb+ complex
is reproduced and, more impressively, the actual position
of the cation in the cavity is also very close in the model
structure. It could be argued that this is simply because
the cation has found the lowest energy position in a
calixarene whose structure the modeling process has been
unable to disturb. However, the fact that similar calcu-
lations on tetraamide 16 and its K+ complex (whose X-ray
structures are both known)9 do reproduce the extensive
reorganization of the calixarene conformation on com-
plexation argues strongly against this view.
The structure of the Na+ (MeOH) complex of 3 shows
two interesting effects as compared to the Rb+ complex
(Figure 10). The first is that the Na+ ion lies much
farther below the plane of the phenolic oxygens than does
the Rb+ ion, despite the coordination by methanol at the
upper rim which would, if anything, be expected to pull
the cation even higher. The second is that the unique
inward tilting aryl ring is moved to a much more upright
position in the Na+ complex. In addition, the symmetric
arrangement of the phenolic rings is lost in the complex
with one of those adjacent to the unique ring tilted very
far outward and the other tilted significantly less.
However, on the basis of the X-ray data alone, it is not
possible to tell whether these effects are associated solely
with the Na+ ion or are due to the coordinated solvent
molecule. Since the modeling appeared to predict rea-
sonable structures for 3 and its Rb+ complex, it seemed
appropriate to attempt to address these questions by

498 J . Org. Chem., Vol. 63, No. 3, 1998
Bell et al.
F igu r e 11. X-ray and energy-minimized structures of molecule 3. (a) Energy-minimized structure of free ligand. (b) X-ray structure
of free ligand. (c) Energy-minimized structure of Rb+ complex. (d) X-ray structure of Rb+ complex. (e) Na+ complex without methanol
(energy minimized). (f) Na+ complex with methanol (energy minimized). (g) Na+ complex with methanol (X-ray).
encountered by Gutsche in his studies of calix[5]arene
ethers and esters.13 There is evidence from studies of
calix[4]arenes that molecular mechanics calculations are
not reliable in predicting the lowest energy conformers17
and there is no a priori reason to suppose that in the
conformationally more heterogeneous calix[5]arenes they
will prove to be more reliable, except that the increased
flexibility of the larger molecules may reduce the interac-
tions between each of the substituted arene rings and
allow comparison between relatively unstrained mol-
ecules. However, in view of the conformational hetero-
geneity, it is useful to have some guide to the geometries
of the possible conformers and an indication of those
which are either very high or reasonably low in energy,
to aid interpretation of other data. The structure of the
lowest energy conformer resulting from an extensive
search of the large number of related structures which
can be generated from 5 is shown in Figure 5b, along

Cation Complexation by Chemically Modified Calixarenes
J . Org. Chem., Vol. 63, No. 3, 1998 499
added and the solution filtered through Celite. The organic
layer was separated and washed in turn with 1 N HCl, brine,
and distilled water, each washing being back-extracted with
dichloromethane. The organic layer was dried with MgSO4
and concentrated to a pale orange solid. The solid was
dissolved in chloroform and subsequently precipitated by the
addition of methanol. Recrystallization from toluene afforded
calix[5]arene (5.22 g, 80%), mp >275 °C as a cream-colored
solid. [Found: C, 79.4; H, 5.4. C35H30O5 requires C, 79.2; H,
5.7%]; νmax (KBr/cm-1) 3300 (OH); δH (CDCl3), (300 MHz) 3.85
(10H, bs, ArCH2Ar), 6.84 (5H, t, J ) 7.51 Hz, ArH), 7.21 (10H,
d, J ) 7.51 Hz, ArH), 8.92 (5H, s, OH).
p-ter t-Bu tylca lix[5]a r en e P en ta -ter t-bu tyl Keton e (3).
A mixture of p-tert-butylcalix[5]arene18 (4.0 g, 4.94 × 10-3 mol),
bromopinacolone (4.42 g, 2.47 × 10-2 mol), potassium iodide
(4.10 g, 2.47 × 10-2 mol), anhydrous potassium carbonate (6.35
g, 4.60 × 10-2 mol), and dry acetone (60 mL) was heated under
reflux for 24 h. The cooled mixture was filtered through Celite,
and the filtrate and dichloromethane washings of the solid
residue were combined and washed successively with 3%
sulfuric acid (50 mL) and water (2 × 10 mL). The solution
was then dried and concentrated at reduced pressure to afford
a solid which was taken up in aqueous ethanol and heated
under reflux for 3 days. Dichloromethane extraction yielded
with the X-ray structure. In this particular case there
is good agreement between both structures. The next
lowest lying conformer is the cone, but there are several
low energy conformers within a few kcal mol-1 of each
other; for example, also included in the figure is a second
1,2-partial cone whose energy is approximately 2 kcal
mol-1 higher than that of the lowest energy form. In
solution, conformational averaging similar to that ob-
served for the cone conformer is clearly possible between
both 1,2-partial cone structures shown in Figure 5, since
their interconversion does not require a full inversion of
an aryl ring. This observation emphasizes the main
advantage provided by modeling of such systems which
is not that the modeling gives a structure very similar
to that of the crystal as the lowest energy conformer, but
that it can generate a series of low energy structures
which, although they cannot be observed crystallographi-
cally, may well be important in solution.
Exp er im en ta l Section
Unless stated otherwise, all reagents were obtained from
commercial sources and were used without further purifica-
tion.
Infrared spectra were obtained using a Perkin-Elmer 983G
grating spectrophotometer. Solid samples were dispersed in
KBr and recorded as clear pressed disks.
a
solid which on recrystallization from ethanol-dichlo-
romethane furnished 3 (2.83 g, 44%) as white crystals, mp
301-303 °C; 1H NMR (CDCl3, 300 MHz) δ 1.01 (45H, s), 1.22
(45H, s) 3.35 (5H, d, J ) 14.5 Hz), 4.73 (5H, d, J ) 14.4 Hz),
5.04 (10H, s), 6.87 (10H, s); IR (KBr) λmax 1729/, 1699. Anal.
Calcd for C85H120O10‚H2O; C, 77.30; H, 9.30. Found: C, 77.40;
H, 9.30; MS (ES+) m/e: 1319.7 (M + H + H2O)+, 1323.5 (M +
Na)+, 1324.3 (M + H + Na). C85H120O10 requires M+ 1300.9.
p-ter t-Bu tylca lix[5]a r en e P en ta m eth yl Keton e (1). A
mixture of p-tert-butylcalix[5]arene (5.0 g, 6.17 × 10-3 mol),
sodium iodide (8.67 g, 5.79 × 10-2 mol), chloroacetone (5.35 g,
5.79 × 10-2 mol), and potassium carbonate (8.00 g, 5.79 × 10-2
mol) in dry acetone (300 mL) was heated under reflux for 12
h. The cooled mixture was filtered through Celite, and the
filtrate and dichloromethane washings of the solid residue
were combined and concentrated at reduced pressure to afford
an oil. The crude product was suspended in water and stirred
at 60 °C for 2 h. Dichloromethane extraction yielded a solid
which was recrystallized from ethanol-dichloromethane to
1H NMR spectra were recorded at 300 MHz on a General
Electric QC 300 spectrometer and at 500 MHz on a General
Electric QE 500 instrument. 13C spectra were recorded at 125
MHz, using a General Electric QE 500 spectrometer. In all
cases tetramethylsilane (TMS) was used as an internal
standard and chloroform-d (deuteriochloroform) as solvent
unless otherwise stated. Chemical shifts are expressed in
parts per million (ppm or δ) downfield from the standard.
Elemental analyses were determined on a Perkin-Elmer
2400 CHN microanalyzer. Compounds quoted gave elemental
analyses to within (0.5% of theoretical values. Carbon values
in microanalysis are frequently low for calixarenes with
cavities capable of retaining solvent molecules. In some
instances acceptable elemental analysis was only possible if
one assumes the inclusion of one or more molecules of solvent.
EI mass spectra were recorded at 70 eV on a VG Autospec
instrument using a heated inlet system. Accurate molecular
weights were determined by the peak matching method using
perfluorokerosene as standard reference. ES mass spectra
were recorded on a Fisons VG-Quatro instrument with elec-
trospray inlet. Depending on the molecular weight of the
sample, bovine trypsinogen (600-2200 mass range), horse
heart myoglobin (700-1800 mass range), or a mix of poly-
(ethylene glycol)s (300-1400 mass range) were used for
calibration purposes.
1
afford 1 (4.31 g, 64%), as a white solid, mp 244-246 °C; H
NMR (CDCl3, 500 MHz, -10 °C) δ 0.75 (3H, s), 1.00-1.09 (6H,
bs), 1.15 (18H, s), 1.33 (18H, s), 1.44 (9H, s), 2.11 (6H, s), 3.29
(2H, d, J ) 13.6 Hz), 3.34 (1H, d, J ) 14.0 Hz), 3.81 (2H, d, J
) 14.7 Hz), 4.01 (2H, d, J ) 14.7 Hz,), 4.13 (1H, d, J ) 12.9
Hz), 4.16 (2H, d, J ) 16.5 Hz), 4.17 (2H, s), 4.21 (2H, d, J )
13.6 Hz), 4.22 (4H, s), 4.53 (2H, d, J ) 16.5 Hz), 7.10 (2H, d,
J ) 1.8 Hz), 7.13 (2H, d, J ) 1.1 Hz), 7.19 (2H, s), 7.37 (2H,
s), 7.58 (2H, s); δc (CDCL3) (-10 °C) 25.83, 26.04, 27.74, 30.96,
31.06, 31.17, 31.22, 31.33, 31.39, 31.49, 34.08, 34.22, 76.83,
76.89, 77.10, 125.76, 126.17, 126.54, 126.89, 127.23, 127.66,
127.70, 127.73, 131.65, 132.79, 133.16, 146.09, 146.49, 146.78,
150.74, 151.65, 204.41, 206.33, 210.12; IR (KBr) λmax 1738,
1715 (CdO). Anal. Calcd for C70H90O10: C, 77.00; H, 8.30.
Found: C, 76.70; H, 8.30: MS (ES+) m/e: 1091.5 (M + H)+,
1109.6 (M + H + H2O)+, 1113.5 (M + Na)+. C70H90O10 requires
M+ 1090.6.
Crystallographic data for 1 and 3‚NaClO4 were collected at
a low temperature on a Siemens P4 four-circle diffractometer.
The remaining data sets were collected at rt using a Siemens
P3 machine. All the structures were solved by direct meth-
ods23 and refined on F2 using SHELXL-93.15
Analytical TLC was performed on Merck Kieselgel 60254
plates. Preparative TLC was carried out with glass plates (20
× 20 cm) coated with Merck Kieselgel PF254+366 (21 g in 58
mL of H2O per plate). Flash chromatography was effected
using Merck Kieselgel 60 (230-400 mesh).
Commercial grade solvents were dried and purified by the
standard procedures as specified in Purification of Laboratory
Chemicals (Perrin, D. D.; Armarego, W. L. F.; Perrin, D. R.;
Pergamon Press: New York, 1966).
Ca lix[5]a r en e (9). A mixture of p-tert-butylcalix[5]arene
(10.00 g, 12.35 mmol) and finely ground aluminum trichloride
(16.47 g, 0.12mol) was stirred vigorously in dry toluene (100
mL) at room temperature for 12 h. Hydrochloric acid (1 N)
was then added dropwise to destroy excess aluminum trichlo-
ride. Dichloromethane (ca. 3 times the volume of toluene) was
Ca lix[5]a r en e P en ta ter t-Bu tyl Keton e (5). Treatment
of calix[5]arene with bromopinacolone in the presence of
potassium iodide and potassium carbonate exactly as described
above for tert-butyl ketone 3 afforded ketone 5 (53%), mp 210-
212 °C (from ethanol-dichloromethane); 1H NMR (CDCl3, 300
MHz) δ (cone isomer 5a ) 1.17 (45H, s) 3.31-3.40 (5H, d, J )
14.5 Hz), 4.67 (5H, d, J ) 14.5 Hz), 4.87 (10H, s), 6.73 (15H,
m); δ (1,2-alternate isomer 5b) 1.09 (9H, s), 1.12 (18H, s), 1.22
(18H, s), 3.34 (1H, d, J ) 14.4 Hz), 3.36 (2H, d, J ) 16.6 Hz),
3.85 (2H, d, J ) 16.9 Hz), 3.89 (4H, m), 4.42 (1H, d, J ) 14.4
Hz), 4.50 (2H, s), 4.51 (2H, d, J ) 16.7 Hz), 4.74 (4H, m), 4.77
(2H, d, J ) 16.7 Hz), 6.10 (2H, d, J ) 7.3 Hz), 6.64 (2H, m),
6.82 (2H, d, J ) 7.6 Hz), 6.94 (2H, m), 7.08 (2H, d, J ) 7.4
Hz), 7.24 (2H, m) 7.44 (2H, m), 7.54 (1H, t, J ) 6.2 Hz); λmax
(KBr) 1723 (CdO). Anal. Calcd for C65H80O10: C, 76.40; H,

500 J . Org. Chem., Vol. 63, No. 3, 1998
Bell et al.
7.90. Found: C. 76.20; H, 8.00: MS (ES+) m/e: 1021.4 (M +
H)+, 1039.3 (M + H + H2O)+, 1059.4 (M + H)+. C65H80O10
requires M+ 1020.56.
Ca lix[5]a r en e P en ta m eth yl Keton e (6). Treatment of
calix[5]arene with chloroacetone in the presence of potassium
iodide and potassium carbonate exactly as described above for
pentamethyl ketone (1) afforded 6 (73%) as white crystals, mp
223-225 °C (from ethanol-dichloromethane); 1H NMR (CDCl3,
500 MHz, -10 °C) δ 0.56 (3H, s), 1.43 (6H, bs), 2.14 (6H, s),
3.33 (2H, d, J ) 14.3 Hz), 3.40 (1H, d, J ) 14.3 Hz), 3.79 (2H,
d, J ) 14.0 Hz), 3.98 (2H, d, J ) 14.3 Hz), 4.04 (4H, m), 4.12
(3H, d, J ) 14.Hz, and d, J ) 14.3 Hz), 4.20 (2H, d, J ) 16.7
Hz), 4.23 (2H, s), 4.53 (2H, d, J ) 16.7 Hz), 6.77-6.83 (4H,
m), 7.07 (2H, t, J ) 7.5 Hz), 7.15-7.29 (5H, m), 7.33 (2H, d, J
) 7.5 Hz), 7.57 (2H, d, J ) 7.0 Hz); δc (CDCl3) (-10 °C) 25.27,
25.64, 25.95, 26.95, 28.82, 36.71, 76.41, 77.14, 77.20, 124.19,
124.56, 124.93, 128.20, 130.01, 130.51, 130.74, 131.49, 132.96,
133.11, 133.34, 133.59, 153.53, 153.63, 153.74, 203.65, 205.37,
209.53 (Cd0); IR λmax (KBr) 1744, 1723 (CdO). Anal. Calcd
for C50H50O10: C, 74.00; H, 6.20. Found: C, 73.5; H, 6.00; MS
(ES+) m/e: 811.4 (M + H)+, 833.4 (M + Na)+, 849.3 (M + K)+.
F igu r e 12. Partial charges used in the MM+ calculation of
the structures of Na3+ and Rb3+.
trile were also unsuccessful. The potentiometric measure-
ments were performed in the case of K+, Rb+, and Cs+ which
form very strong complexes with ligand 5. The method, based
on the competition with Ag+ as auxiliary cation, has been
described in detail elsewhere.10 log â value found for the Ag5+
complex and used in the calculations was 5.10 ( 0.05. The
stability constants derived from both types of measurements
(Table 2) were refined by the program Sirko.21
The ionic strength of the solutions was maintained at 0.01
mol dm-3 by use of Et4NCl and Et4NClO4, for the spectropho-
tometric and potentiometric measurements, respectively. Et4-
NCl was recrystallized twice from acetone, Et4NClO4 was
washed with acetone before being recrystallized twice from
water. Both salts were dried under vacuum for 24 h.
The metallic salts used were the following chlorides: LiCl
(Fluka, puriss.), NaCl (Prolabo, Normapur), KCl (Merck, p.a.,
Fluka, puriss.), and CsCl (Merck, p.a.), the following
perchlorates: AgClO4 (Fluka, puriss.), KClO4 (Prolabo, Nor-
mapur), RbCl (Sigma), and cesium nitrate (CsNO3) (Fluka,
purum).
C
50H50O10 requires M+ 810.33.
p-ter t-Bu tylcalix[6]ar en e Hexam eth yl Keton e (2). Treat-
ment of p-tert-butylcalix[6]arene19 with chloroacetone, potas-
sium iodide, and potassium carbonate exactly as described
above for pentamethyl ketone 1 afforded hexamethyl ketone
- (58%), as a colorless solid, mp 237-240 °C (ethanol-
dichloromethane); 1H NMR (CDCl3, 300 MHz) δ 1.17 (54H, s),
1.70-1.91 (18H, bs), 3.40-4.20 (24H, br), 7.05 (12H, bs); 1R
(KBr) λmax 1735 (CdO). Anal. Calcd for C84H108O12: C, 76.44;
H, 8.34. Found: C, 77.08; H, 8.25.
The enthalpies of complexation ∆Hc of cesium complexes
have been determined in the same solvent using an isoperibol
titration calorimeter (Tronac 450, Orem, UT). The experi-
mental details have been previously described.21 The ligand
concentrations were in the range 5 × 10-4 10-3 mol dm-3. ∆Hc
values were refined by program Sirko,21 the values of log â
determined by spectrophotometry or potentiometry being set
as constant. The entropies of complexation ∆Sc were derived
from the following equation: ∆Gc ) ∆Hc - T∆Sc, knowing ∆Gc
) -RT ln â (Table 3).
p-ter t-Bu tylca lix[6]a r en e Hexa -ter t-bu tyl Keton e (4).
Treatment of p-tert-butylcalix[6]arene with bromopinacolone,
potassium iodide, and potassium carbonate exactly as de-
scribed above for penta tert-butyl ketone 3 afforded hexaketone
1
(70%) as a colorless solid, mp 301-304 °C; H NMR (CDCl3,
500 MHz) δ 0.90-1.40 (108H, m), 3.30-3.70 (12H, m), 4.20-
4.50 (6H, m), 4.60-5.00 (6H, m), 6.60-7.40 (12H, m); IR (KBr)
λ
max 1728 (CdO). Anal. Calcd for C102H144O12‚2H2O: C, 76.60;
H, 9.01. Found: C, 76.90; H, 8.84.
Ca lix[6]a r en e Hexa m eth yl Keton e (7). Treatment of
calix[6]arene with chloroacetone, potassium iodide, and potas-
sium carbonate as described above for hexamethyl ketone
afforded hexamethyl ketone as a colorless solid (68%), mp ∼285
°C dec; 1H NMR (CDCl3, 300 MHz) δ 1.80 (18H, bs), 3.90 (12H,
bs), 4.18 (12H, bs), 6.90 (36H, t); IR (KBr) λmax 1733 (CdO).
Anal. Calcd for C60H60O12: C, 74.09; H, 6.17. Found: C, 74.36;
H, 6.15.
P icr a te Extr a ction Mea su r em en ts. Picrate extraction
experiments were performed following Pedersen’s pro-
cedure:20 5 mL of a 2.5 × 10-4 M aqueous picrate solution and
5 mL of a 2.5 × 10-4 M solution of calixarene in dichlo-
romethane were agitated in a stoppered glass tube with a
mechanical shaker for 2 min and then magnetically stirred in
a thermostated water-bath for a minimum of 30 min. In the
case of the p-tert-butyl calix[5]arene penta tert-butyl ketone
(3), 3 days of stirring were necessary to reach the extraction
equilibrium. The solutions were then left standing for 30 min
for phase separation, and the concentration of picrate remain-
ing in the aqueous phase was determined spectrophotometri-
cally. Experimental details as well as the preparation of alkali
picrates have been previously described.10 The results, ex-
pressed as the percentage (% E) of cation extracted are given
in Table 1.
Th er m od yn a m ic P a r a m eter Mea su r em en ts. The sta-
bility constants â, â being the concentration ratio [ML+]/[M+]-
[L] (with M+ ) cation and L ) ligand), have been determined
in methanol (Carlo Erba p.a.) by UV absorption spectropho-
tometry or by potentiometry. The spectrophotometric method
consisted of adding increasing amounts of the metal salt
solution into 2 mL of ligand solution (CL ca. 2.5 × 104 mol
dm-3) directly in the spectrophotometric cell (1 cm path
length), after the absence of any slow complexation kinetics
had been checked. Compound 3 was not soluble enough for a
spectrophotometric study in methanol. Attempts in acetoni-
Molecu la r Mech a n ics P r oced u r es. Calculations were
performed using the MM+ method implemented within the
“Hyperchem” molecular modeling package.25 This all-atom
force field is essentially an extended version of MM2.26 The
metal-free structures were minimized using dipole-dipole
interactions. The complexes were “built” by inserting the
appropriate metal cation approximately in the vicinity of the
pendent arms of the calixarene. The metal was fixed in this
position during the minimization while the calixarene struc-
ture was allowed to relax around it. Electrostatic interactions
were explicitly included in the standard way. Partial charges
for the calixarene were estimated from a semiempirical (AM1)
calculation on a single fragment, i.e., one aryl subunit, of the
pentamer. These charges were not scaled in any way and were
directly transferred, after some rounding off, and grouping of
adjacent charges, to the complex. The partial charges are
shown in Figure 12. In the case of structures which included
solvent, the solvent molecule was introduced deliberately
rotated 90° to its orientation in the X-ray structure, so as to
prevent as far as possible an unconscious bias in positioning
the molecule near to its known position in the X-ray picture.
No extensive conformational searches of the pendent ketone
arms were carried out since the large degree of approximation
in the model did not justify investigation of such fine detail.
The starting structures for all the cation complexes were
therefore based on the same model metal-free calixarene.
(23) Sheldrick, G. M. SHELXS 86. Acta Crystallogr., Sect A 1990,
46, 467.
(24) McKervey, M. A.. Schwing-Weill, M.-J .; Arnaud-Neu, F. In
Comprehensive Supramolecular Chemistry; Vogel, G. W., Vol. Ed.;
Pergamon: New York, 1996; Vol. 1, p 537.
(25) Hyperchem is available from Hypercube Inc., 419 Philips Street,
Waterloo, Ontario, N2L 3X2, Canada.
(26) Allinger, N. L. J . Am. Chem. Soc. 1977, 99, 8127.

Cation Complexation by Chemically Modified Calixarenes
J . Org. Chem., Vol. 63, No. 3, 1998 501
More extensive calculations were carried out on the metal-
free dealkylated complex 5. In this case we attempted to build
a complete set of the various conformers possible for this
sterically less demanding calix[5]arene by using as the initial
geometry the cone conformer and then systematically inverting
the aryl rings around the cavity. We found that because of
the large number of possible conformers more than one
starting geometry was needed to generate a closed set of
conformers. For example, three starting geometries for the
partial cone conformer were used, and these corresponded
either to inversion of the inward tilting aryl ring or one of the
two adjacent rings or one of the pair opposite. Energy
minimization of these structures was followed by a cross-
checking procedure where a new set of starting structures was
generated by reinversion of the aryl rings, i.e., the partial cone
was also built from an energy minimized 1,2-alternate con-
former. Energy minimzation of this new set of conformers
generated structures which could be compared with those from
the first build/minimization cycle. In many cases several
structures relaxed to a clearly distinguishable single conformer
with a characteristic aryl ring conformation, although the
orientation of the alkyl chains were different in each of the
relaxed structures. In such cases the lowest energy conformer
was chosen. To fully sample the potential energy hypersurface
it would be necessary to carry out a full conformational search
of each of the alkyl chain conformations within the same aryl
ring geometry. However, the number of starting geometries
rapidly becomes prohibitive; for example, a set of structures
where each of the five unique ketonic carbonyls is allowed only
to point inward (toward the core) or outward has 25 ) 32
members, if four approximately perpendicular orientations are
required the number rises to 45 ) 1024.
Su p p or tin g In for m a tion Ava ila ble: Tables 1-24 (sum-
mary of crystal data and structure refinement, atomic coor-
dinates, bond lengths and angles, anisotropic displacement
parameters, and hydrogen coordinates for compounds 1, 3,
3‚NaClO4, 3‚RbCl04 and 5) (51 pages). This material is
contained in libraries on microfiche, immediately follows this
article in the microfilm version of the journal, and can be
ordered from the ACS; see any current masthead for ordering
information.
J O9713114