KF and CsF recognition and extraction by a calix[4]crown-5 strapped calix[4]pyrrole multitopic receptor

Article abstract of DOI:10.1021/ja310673p

On the basis of ¹H NMR spectroscopic analyses and single crystal X-ray crystal structural data, the ion-pair receptor 1, bearing a calix[4]pyrrole for anion binding and calix[4]arene crown-5 for cation recognition, was found to act as a receptor for both CsF and KF ion-pairs. Both substrates are bound strongly but via different binding modes and with different complexation dynamics. Specifically, exposure to KF in 10% CD₃OD in CDCl₃ leads first to complexation of the K⁺ cation by the calix[4]arene crown-5 moiety. As the relative concentration of KF increases, then the calix[4]pyrrole subunit binds the F^- anion. Once bound, the K⁺ cation and the F^- anion give rise to a stable 1:1 ion-pair complex that generally precipitates from solution. In contrast to what is seen with KF, the CsF ion-pair interacts with receptor 1 in two different modes in 10% CD₃OD in CDCl₃. In the first of these, the Cs⁺ cation interacts with the calix[4]arene crown-5 ring weakly. In the second interaction mode, which is thermodynamically more stable, the Cs ⁺ cation and the counteranion, F^-, are simultaneously bound to the receptor framework. Further proof that system 1 acts as a viable ion-pair receptor came from the finding that receptor 1 could extract KF from an aqueous phase into nitrobenzene, overcoming the high hydration energies of the K⁺ and F^- ions. It was more effective in this regard than a 1:1 mixture of the constituent cation and anion receptors (4 and 5).

Full text of DOI:10.1021/ja310673p

Article
pubs.acs.org/JACS
KF and CsF Recognition and Extraction by a Calix[4]crown‑5
Strapped Calix[4]pyrrole Multitopic Receptor
Sung Kuk Kim,† Vincent M. Lynch,† Neil J. Young,∥ Benjamin P. Hay,*,∥ Chang-Hee Lee,*,‡
Jong Seung Kim,*,§ Bruce A. Moyer,*,∥ and Jonathan L. Sessler*,†,⊥
†Department of Chemistry and Biochemistry, The University of Texas at Austin, 105 East 24th Street, Stop A5300, Austin, Texas
78712-1224, United States
‡Department of Chemistry, Kangwon National University, Chun-Chon 200-701, Korea
§Department of Chemistry, Korea University, Seoul 136-701, Korea
∥Chemical Sciences Division, Oak Ridge National Laboratory, Oak Ridge, Tennessee 37830-6119, United States
⊥Department of Chemistry, Yonsei University, Seoul 120-749, Korea
S
* Supporting Information
ABSTRACT: On the basis of 1H NMR spectroscopic analyses
and single crystal X-ray crystal structural data, the ion-pair
receptor 1, bearing a calix[4]pyrrole for anion binding and
calix[4]arene crown-5 for cation recognition, was found to act
as a receptor for both CsF and KF ion-pairs. Both substrates
are bound strongly but via different binding modes and with
different complexation dynamics. Specifically, exposure to KF
in 10% CD3OD in CDCl3 leads first to complexation of the K+
cation by the calix[4]arene crown-5 moiety. As the relative
concentration of KF increases, then the calix[4]pyrrole subunit
binds the F− anion. Once bound, the K+ cation and the F−
anion give rise to a stable 1:1 ion-pair complex that generally
precipitates from solution. In contrast to what is seen with KF, the CsF ion-pair interacts with receptor 1 in two different modes
in 10% CD3OD in CDCl3. In the first of these, the Cs+ cation interacts with the calix[4]arene crown-5 ring weakly. In the second
interaction mode, which is thermodynamically more stable, the Cs+ cation and the counteranion, F−, are simultaneously bound to
the receptor framework. Further proof that system 1 acts as a viable ion-pair receptor came from the finding that receptor 1 could
extract KF from an aqueous phase into nitrobenzene, overcoming the high hydration energies of the K+ and F− ions. It was more
effective in this regard than a 1:1 mixture of the constituent cation and anion receptors (4 and 5).
complexities associated with tracking multiple ionic species, a
problem compounded by the poor solubility of many of the
ion-pairs (salts) that are commonly tested.1−8
INTRODUCTION
■
In recent years, increasing attention has focused on so-called
ditopic receptors, systems with two disparate binding sites.1,2
Much of this interest reflects the fact that in many cases ditopic
receptors containing both anion and cation binding sites display
enhanced affinities for targeted ion-pairs as compared to
analogous monotopic or single ion receptors.1,2 While not
established unequivocally, these enhanced recognition features
are thought to result from allosteric effects and enhanced
electrostatic interactions between the cobound ions present in
the ion-pair complexes. More broadly, the putative improve-
ments in affinity and selectivity that can be achieved relative to
single site receptors have made ion-pair receptors attractive for
use in salt solubilization, ion extraction, trans-membrane ion
transport, ion sensing applications, and as logic gates.3−8
However, despite their significant advantages, only a limited
number of well-characterized ion-pair receptors are currently
known. We attribute this to difficulties of molecular design and
synthesis associated with incorporating two disparate binding
motifs into a single framework as well as to experimental
One of the key unresolved questions in ion-pair receptor
design concerns the importance of internuclear charge
separation. Examples of the cation- and anion-binding sites
being proximal and remote are known.1−8 It is perhaps most
straightforward to build ion-pair receptors by linking cation
receptors with anion receptors, making it a challenge to realize
the internuclear distance desired for maximum stability of the
ion-pair receptor complex. If charge separation is too large, then
there should be no advantage of an ion-pair receptor over a
simple mixture of separate molecular cation and anion
receptors. The question remains as to what internuclear
distances are needed for effective intramolecular cooperativity
and what are the molecular-design and matrix factors that come
into play for controlling cooperativity. The linked calix[4]-
Received: October 30, 2012
Published: December 3, 2012
© 2012 American Chemical Society
20837
dx.doi.org/10.1021/ja310673p | J. Am. Chem. Soc. 2012, 134, 20837−20843

Journal of the American Chemical Society
Article
pyrrole−calix[4]crown ion-pair receptors that we have been
studying represent an ideal platform for investigating this
question.8−10 Not only do they present an interesting
laboratory of binding sites in the same molecule, but the
strongest binding sites are at opposite ends of the molecule.
They thus represent useful systems wherein the effect of various
external factors, including the choice of anion, cation, and
solvent, may be probed in a well-controlled manner.
In an effort to understand in greater detail the correlation
between receptor structure and ion-pair binding function, we
report herein the KF and CsF binding properties of ion-pair
receptor 1 and provide evidence not only for site-selective
binding but also for dynamic cation recognition behavior in the
case of CsF. We also show that receptor 1 acts as an extractant
for KF under conditions of liquid−liquid two-phase extraction
and that it is more effective for this purpose than the anion
receptor 4, the cation receptor 5, or an equimolar mixture of
these two monotopic binding agents. This finding provides
support for the appealing notion that effectively designed ion-
pair receptors can display features that are superior to the sum
of their constituent parts.
NMR time scale (Figure S1). This slow exchange kinetics
presumably reflects strong complexation between the fluoride
anion and receptor 1 as the result of direct anion−receptor
interactions involving the calix[4]pyrrole moiety. This con-
clusion was further supported by the observation that over the
course of the addition, the singlet ascribed to the pyrrolic NH
proton resonance is shifted downfield to δ ≈ 12.7 ppm (Δδ ≈
6.0 ppm) and becomes split into a doublet (J = 40.0 Hz). This
latter splitting is a typical feature of calix[4]pyrrole−F− anion
complexes and is ascribed to a coupling of the bound fluoride
anion with the NH protons (Figure S1).11 The triplet peaks
corresponding to the β-pyrrolic protons were likewise shifted
upfield during the titration and ultimately appeared as two
singlets resonating at 5.83 and 5.68 ppm, respectively (Figure
S1). These latter chemical shift changes are attributed to an
increase in the electron density of the pyrrole subunits resulting
from the interaction between the anion and the pyrrolic NH
protons.
The interactions between receptor 1 and the F− anion (as the
tetraethylammonium (TEA) salt) in acetonitrile were further
probed by isothermal titration calorimetry (ITC). The resulting
data revealed that the interaction between TEAF [2.97 mM]
and receptor 1 [0.20 mM] is favorable in terms of both
enthalpy and entropy (Ka = 6.0 × 105 M−1; ΔG = −31 kJ/mol;
ΔH = −16 kJ/mol; and TΔS = 15 kJ/mol; Table S1 and Figure
S3).
Given the favorable binding observed for TEAF, an effort was
made to understand the cation dependence (if any) on the
recognition features of 1. In principle, ion-pair receptor 1 can
form complexes with cations via three distinct recognition
motifs, the calix[4]crown-5 ring, the ethylene glycol spacers
between the calix[4]arene and the calixpyrrole unit, and the π-
electron-rich concave calix[4]pyrrole “cup” formed by anion
complexation. As a consequence, salts could be bound to 1 via a
number of binding modes.8−10,12 Four possible limiting binding
modes for the complexation of CsF are shown in Figure 1 and
RESULTS AND DISCUSSION
In a first set of analyses, the ability of ion-pair receptor 1 to
bind halide anions in CDCl3 solution was investigated via H
■
1
NMR spectroscopy. In analogy to what was found to be true in
the case of receptors 2 and 3 (studied previously), significant
changes were observed in the 1H NMR spectrum of receptor 1
when it was subjected to titration with soluble fluoride anion
salts, such as tetrabutylammonium fluoride (TBAF) (Figure
S1).8,9 However, other anions tested (TBACl, TBABr, and
TBAI) induced no appreciable chemical shift changes in the 1H
NMR spectra of receptor 1 in CDCl3 (Figure S2). We thus
conclude that, as is true for 2 and 3, receptor 1 is highly
selective for the fluoride anion under these solvent con-
ditions.8,9
As shown in Figure S1, receptor 1 in its ion-free form
displays a broad singlet at δ = 6.71 ppm for the NH protons
and two triplets about δ = 6.04 ppm and δ = 5.95 ppm for the
β-pyrrolic protons in its 1H NMR spectrum. These peak
patterns are in accord with what has been seen for compounds
2 and 3.8,9 The observed triplets are thought to reflect long-
range couplings between the CH and NH protons. Support for
this latter conclusion came from deuterium exchange experi-
ments carried out using 10% CD3OD in CDCl3 (loss of NH
signal; simplification of CH resonances into doublets).8,9 When
receptor 1 was titrated with TBAF in CDCl3, two sets of
distinguishable resonances for all proton signals were seen in
Figure 1. Views of the four limiting binding modes proposed for the
interaction of receptor 1 with CsF.
are designated crown/crown, crown/pyrrole, glycol/pyrrole,
and pyrrole/pyrrole, respectively. In this nomenclature, the first
part of the name indicates the location of the cation, and the
latter part of the name indicates the location of the anion.
Molecular mechanics calculations were performed to evaluate
the relative stabilities of the binding modes shown in Figure 1
in the case of KF and CsF in the absence of solvent. The results
of these calculations are summarized in Table 1.
1
the H NMR spectra recorded before saturation was achieved.
These signals were ascribed to the free form of 1 and its
fluoride anion complex, respectively. This observation of two
sets of peaks is consistent with the binding/release equilibrium
between compound 1 and the fluoride anion being slow on the
The data in Table 1 provide some general insights into the
KF and CsF recognition features of receptor 1. First,
irrespective of the binding mode, there is a strong preference
for the binding of K+ rather than Cs+. Second, the preferred
20838
dx.doi.org/10.1021/ja310673p | J. Am. Chem. Soc. 2012, 134, 20837−20843

Journal of the American Chemical Society
Article
1
the H NMR spectrum, corresponding the β-pyrrolic protons,
are seen to appear. In contrast, other peaks corresponding to
the aromatic protons of the calix[4]arene moiety were merely
broadened (Figure S7). This is consistent with the presence of
two different kinds of binding interactions involving receptor 1
and the CsF ion-pair. In one of the modes, only the Cs+ cation,
but not the F− anion, is weakly bound to the crown-5 ring to
form a cesium complex ([1·Cs+]F−) wherein the F− counter-
anion is not cobound (crown/crown binding mode). This form
is characterized by a general broadening of the aromatic and
crown ether signals in the 1H NMR spectrum, but not
significant changes in the chemical shifts. This complex
([1·Cs+]F−) exists in fast equilibrium with the free receptor,
as evidenced by the peak broadening seen for the aromatic
protons of the calix[4]arene moiety. In contrast, in the other
binding mode, the Cs+ cation and the F− anion are bound
concurrently and strongly to receptor 1. In this case, the cation
and anion are stabilized by the ethylene glycol spacers and the
calix[4]pyrrole moiety, respectively (glycol/pyrrole mode; cf.,
Figures 1 and 2). This latter complexation mode, which
Table 1. Calculated Gas-Phase Binding Energies for Ion-
Pairs Bound to 1 via Different Limiting Binding Modes
a
binding energy (kcal/mol)
ion-pair crown/crown crown/pyrrole glycol/pyrrole pyrrole/pyrrole
KF
−204.4
−183.9
−186.5
−165.6
−200.6
−193.1
−196.6
−182.5
CsF
a
ΔE = E(complex) − E(ligand) − E(cation) − E(anion).
gas-phase binding modes, crown/crown and glycol/pyrrole, are
those that minimize the distance between the two cobound
ions (i.e., K+ or Cs+ and F−). The KF ion-pair exhibits a
preference for the crown/crown mode, whereas CsF exhibits a
preference for the glycol/pyrrole mode. In the absence of
solvation, these preferences primarily result from maximizing
the electrostatic attraction between the cation and anion.13
The solution-phase anion and cation binding behavior of 1
1
was investigated via H NMR spectroscopy using a mixed
solvent system consisting of CDCl3 and CD3OD (9/1, v/v).
This particular choice of solvents was dictated by the solubility
of the salts under study. A further consideration is that this
solvent system was used in the case of 2 and 3,8,9 thus
permitting direct comparisons with these previously studied
systems. As shown in Figures S4 and S5, the addition of KClO4
and CsClO4 led to a significant change in the proton signals for
the calix[4]arene crown-5 moiety but only a modest and no
shift in the pyrrolic CH resonances for KClO4 and CsClO4,
respectively. Such observations are interpreted in terms of the
K+ and Cs+ cations being bound to the calix[4]arene crown-5
without the perchlorate anion being cobound to the calix[4]-
pyrrole subunit.10 On the basis of a quantitative analysis of
chemical shifts involved, it is concluded that receptor 1 binds
the K+ cation more strongly than it does the Cs+ cation (Ka =
6.5 × 106 M−1 for K+ vs Ka = 3.3 × 104 M−1 for Cs+).10
In contrast to the above, treatment of 1 with 5 equiv of
TBAF induced no appreciable change in the 1H NMR
spectrum in CDCl3/CD3OD (9/1, v/v), indicating that
receptor 1 is unable to bind the F− anion (or the large TBA+
cation) effectively in this more polar solvent system (relative to
pure CDCl3; vide supra; Figure S6). Presumably, the F− anion
is strongly solvated in this mixed solvent system. This solvation
may be enhanced by the fact that the TBAF used for these
studies is not anhydrous. However, in the presence of
coordinating cations, such as cesium and potassium, the F−
anion is bound to the receptor 1, forming strong ion-pair
complexes wherein the cation is bound to the calix[4]arene
crown-5 moiety and the anion to the calix[4]pyrrole cavity
(crown/pyrrole mode; cf., Figures 1, S4, and S5). Support for
this conclusion comes from NMR spectral studies carried out in
CD3OD/CDCl3 (1:9, v/v). For instance, as shown in Figure
S5c, the addition of CsF caused all proton signals of receptor 1,
including the β-pyrrolic resonances of the calix[4]pyrrole
moiety, to shift; this is as expected for a binding mode wherein
both the calix[4]arene crown-5 and the calix[4]pyrrole moiety
take part in ion-pair complexation. The presumed strong
interaction between the F− anion and receptor 1 was further
evidenced by the observation that the singlet pyrrolic NH signal
is split into a doublet. This splitting, which occurs upon
exposure to CsF, is accompanied by a remarkable downfield
shift in the NH proton signals (from 6.87 to 12.0 ppm).
Changes in other spectral regions are also seen.
Figure 2. Proposed binding interactions involving 1 and KF and CsF
in CD3OD/CDCl3 (1/9, v/v). On the basis of an analysis of the NMR
spectra, a so-called pyrrole−pyrrole binding mode, wherein the Cs+
cation is bound in the “cup” of the calix[4]pyrrole and the F− is
complexed by the pyrrole NH protons, is ruled out. See text.
involves a slow equilibrium, is similar to what was seen in the
case of the CsF complex of 3 and gives rise to the new
9
1
distinguishable peaks in the H NMR spectrum. After the Cs+
cation is bound to the ethylene glycol spacer, it moves to the
crown-5 ring to form a thermodynamically more stable ion-pair
complex ([1·CsF]) (crown/pyrrole mode; cf., Figures 1 and 2).
Under conditions where 1 is titrated with 0.72−1.38 equiv of
CsF in CD3OD/CDCl3 (1:9, v/v), a new set of sharp peaks in
20839
dx.doi.org/10.1021/ja310673p | J. Am. Chem. Soc. 2012, 134, 20837−20843

Journal of the American Chemical Society
Article
Figure 3. Proposed binding interactions involving 1 and various K+, Cs+, and F− salts in CD3OD/CDCl3 (1/9, v/v). On the basis of an analysis of
1
the H NMR spectra, we rule out other possible binding modes. See text.
The affinity of receptor 1 for CsF in 10% MeOH in CHCl3
was quantified by isothermal titration calorimetry (ITC). This
analysis, which is subject to ≤15% error, revealed that the
interaction of CsF [2.95 mM] with 1 [0.19 mM] is highly
exothermic and enthalpy-driven (ΔH = −64 kJ/mol; ΔG =
−26 kJ/mol; Ka = 4.1 × 104 M−1; Table S1 and Figure S8). The
binding constant of receptor 1 for CsF in 10% CD3OD in
CDCl3 is lower than that of 2 having the calix[4]arene crown-6
(Ka = 3.8 × 105 M−1) and higher than that of the crown-free
receptor 3 (Ka = 1.3 × 104 M−1) (Table S1).8,9 These
differences are ascribed to the reduced interaction between the
crown-5 ring of 1 and the Cs+ cation relative to the
corresponding crown-6 system (2).
Receptor 1 also forms a thermodynamically stable 1:1 ion-
pair complex with KF wherein the K+ cation is bound to the
calix[4]arene crown-5 ring and the F− anion is bound to the
calix[4]pyrrole moiety (crown/pyrrole mode; cf., Figure 1).
However, in terms of binding dynamics on the NMR time
scale, the complexation of KF is quite different from that of
CsF. Upon addition of KF to a solution of receptor 1 in 10%
CD3OD in CDCl3, the proton signals of the calix[4]arene
crown-5 undergo a significant shift, while those of the
calix[4]pyrrole moiety remain largely unchanged (Figure S4).
Such observations are consistent with receptor 1 coordinating
the K+ cation first through the calix[4]arene crown-5 ring
without the F− anion being bound to the calix[4]pyrrole moiety
(crown/crown mode). This complex ([1·K+]F) exists in slow
equilibrium with the free receptor, as evidenced by the
observation of two sets of distinguishable peaks in the proton
NMR spectrum (Figure S4). Specifically, the signals are
ascribable to both the K+ complex ([1·K+]F−) and the free
receptor (but no other species seen in the NMR spectrum).
Once the K+ is bound to receptor 1, the resulting potassium
complex ([1·K+]) facilitates the binding of the F− counter-
anion. In what is then a sequential process on the NMR time
scale, the counteranion then binds to the calix[4]pyrrole moiety
to give the KF ion-pair complex ([1·KF]; crown/pyrrole
mode), which then precipitates from solution (Figures 1−3).
As a result of the observed precipitation, the complexation of
receptor 1 with KF is irreversible under these solution phase
conditions. Presumably, this reflects the fact that KF is bound
by 1 and that the resulting ion-pair−receptor complex is less
well solvated than the ions K+ and F− alone.
To provide support for the above conclusion, the precipitate
1
was collected by filtration, and its H NMR spectrum was
recorded in nitrobenzene-d5 solution (in which the precipitate
and putative complex are soluble). The resulting spectrum
revealed proton signals ascribable to both the calix[4]arene
crown-5 and the calix[4]pyrrole moieties with both sets of
resonances having undergone significant chemical shift changes
relative to the ion-free receptor 1. This is consistent with the K+
and the F− ions being cobound in nitrobenzene to the
20840
dx.doi.org/10.1021/ja310673p | J. Am. Chem. Soc. 2012, 134, 20837−20843

Journal of the American Chemical Society
Article
calix[4]arene crown-5 and the calix[4]pyrrole subunits,
respectively (Figure S9).
occurred. This phase change is ascribed to the formation of
an insoluble [1·KF] ion-pair complex as a result of cation
exchange, replacement of Cs+ by K+ (Figures 3 and 4). In
contrast, the addition of KClO4 to solutions of free receptor 1
in this same solvent mixture or to solutions of CsF alone (i.e.,
in the absence of receptor 1) did not lead to precipitation. We
thus conclude that adding KClO4 to a solution of complex
[1·CsF] gives rise to the corresponding KF complex, [1·KF]
(Figures 3 and 4).
Precipitation was also observed when TBAF was added to a
solution of the potassium complex of receptor 1 formed from
potassium perchlorate ([1·K+]ClO4 ) in 10% CD3OD in
−
CDCl3. This observation leads us to infer that under these
conditions the KF ion-pair complex of 1 is formed as the result
of counteranion exchange (Figure 3).
The stepwise binding seen for receptor 1 and KF on the
NMR time scale mirrors what was seen in the case of
compound 2 and CsF.9 Therefore, it is concluded that
receptors 1 and 2, both of which have two different binding
sites for specific cations (favoring K+ and Cs+ binding in the
case of receptors 1 and 2, respectively), display similar
complexation dynamics. Specifically, the cations interact first
with the receptors via the stronger cation recognition site
exclusively. Only then does complexation of the counteranion
occur within the calix[4]pyrrole anion binding site (Figure 2).
This behavior reflects the fact that receptor 1 bears several
potential cation binding sites. In this specific case, both the
calix[4]arene crown-5 and the ethylene glycol spacers interact
with the Cs+ cation (cf., Figure 2).
The fact that receptor 1 binds KF and CsF ion-pairs strongly
led us to consider that this receptor could have use as an
extractant. These fluoride salts are particularly challenging to
extract from water in that the fluoride anion (ΔGhyd = −465 kJ/
mol for F−) is strongly hydrated as compared to other
monoanionic ions such as Cl− (ΔGhyd = −340 kJ/mol), Br−
(ΔGhyd = −315 kJ/mol), I− (ΔGhyd = −275 kJ/mol), and NO3
−
(ΔGhyd = −300 kJ/mol).14 To test this possibility, H NMR
spectroscopy was used in conjunction with a two-phase system
consisting of D2O and nitrobenzene-d5. Upon exposure of
receptor 1 in C6D5NO2 to aqueous (D2O) solutions of KF,
1
1
significant changes in the H NMR spectrum were observed
(Figure S10). The proton signals of the calix[4]pyrrole moiety,
as well as those of the calix[4]arene crown-5 ring, were seen to
undergo significant shifts. These changes were particularly large
in the case of the NH signals (Δδ ≈ 5.1 ppm). This is
consistent with the interactions between the fluoride anions
and the receptor being very strong. On the basis of the absence
of free receptor signals and the saturation-like behavior seen as
the amount of KF in the aqueous layer increases, we conclude
1
To provide support for the inferences drawn from the H
NMR spectral measurements, that receptor 1 binds both KF
and CsF but displays high selectivity for KF over CsF, we
investigated whether cation metathesis would occur when a
precomplexed CsF ion-pair complex was exposed to K+. This
study was carried out by adding KClO4 to solutions of the
[1·CsF] complex in 10% CD3OD/CDCl3 (Figures 3 and 4).
When a solution of KClO4 in a mixture of methanol-d4/
chloroform-d (1/9, v/v) was added to a solution of complex
[1·CsF] in the same solvent, formation of a precipitate
1
that to the limits of the H NMR spectral analysis, 100% of
receptor 1 originally present in the nitrobenzene phase exists as
the KF ion-pair complex when the initial KF (in the aqueous
phase) and receptor (in the organic phase) are 20 and 4.0 M,
respectively. We thus conclude that it is an effective KF
extractant.15
In contrast to what is seen with KF, exposure of the
nitrobenzene solution of receptor 1 (4.0 M) to an aqueous CsF
solution (20 M) leads to no discernible change in the 1H NMR
spectrum (Figure S10). This observation is taken as evidence
that 1 is unable to extract CsF under these liquid−liquid
extraction conditions. The selectivity for KF over CsF under
conditions of extraction is noteworthy given the fact that the
Cs+ cation (ΔGhyd = −250 kJ/mol) is less well hydrated than
the K+ cation (ΔGhyd = −295 kJ/mol) (vide supra).14,16
A central question is whether receptor 1 would function
more effectively than an equimolar mixture of the isolated
anion and cation receptors from which it is formally derived. To
address this issue, we first tested the ability of the simple cation
and anion receptors (4 and 5) to extract KF from a water phase
into a nitrobenzene phase. Upon contacting nitrobenzene-d5
solutions of (separately) 4 and 5 (4.0 M, respectively) with
D2O solutions containing KF (20 M), no appreciable chemical
shift changes were observed (Figures S12 and S13). This
finding leads us to conclude that neither the anion receptor 4
nor the cation receptor 5 on their own is able to extract
appreciable quantities of KF under our standard liquid−liquid
extraction conditions.
We next tested an equimolar mixture of 4 and 5. In this case,
and in contrast to what was seen for the individual receptors,
evidence for extraction was seen, although the level of efficacy
was reduced as compared to what was observed in the case of
the ion-pair receptor 1. In particular, when an equimolecular
mixture of 4 and 5 in nitrobenzene-d5 was contacted with an
Figure 4. Precipitation induced via cation metathesis observed upon
adding the K+ cation (as KClO4) to the preformed CsF complex of 1
in CD3OD/CDCl3 (1/9, v/v).
20841
dx.doi.org/10.1021/ja310673p | J. Am. Chem. Soc. 2012, 134, 20837−20843

Journal of the American Chemical Society
Article
aqueous (D2O) solution of KF, significant chemical shift
changes were observed in the 1H NMR spectrum (Figure S14).
For instance, two sets of distinguishable resonances were seen
for all of the proton signals in the case of the cation receptor 5.
These signals were readily attributable to the substrate-free
form of 5 and its K+ complex, respectively. This observation led
us to the inference that the binding/release equilibrium
between compound 5 and the potassium cation is slow on
the NMR time scale (Figure S14). In the case of calix[4]pyrrole
4, the NH signal could not be seen in the mixture after contact
with an aqueous KF solution, whereas the β-pyrrolic CH
resonance was shifted 0.1 ppm to slightly higher field but
without peak splitting. Considered in concert, these two
findings provide support for the notion that under the
conditions of the experiment calix[4]pyrrole 4 interacts with
the fluoride anion but with binding and release both being fast
on the NMR time scale (Figure S14).
Å for the Cs+···O separations. The distances between the Cs+
ion and the aromatic carbon atoms in the meta- and para-
positions with respect to the phenoxy groups are on the order
of 3.40−3.46 Å. It is thus inferred that π-metal interactions are
playing a role in stabilizing the complex (Figure 5). As
compared to the distances observed in the CsF complex of
receptor 2 (3.08−3.36 Å for the Cs+···O distance and 3.43−
3.63 Å for the π-metal interaction), those observed between the
bound Cs+ cation and the calix[4]arene crown-5 of receptor 1
are much shorter. This leads us to suggest that receptor 1 holds
the Cs+ cation more tightly than does receptor 2 despite the
fact that the calix[4]arene crown-6 binds the Cs+ cation much
more strongly than does the calix[4]arene crown-5.
In complex 1·CsF, the F− ion is bound to the NH protons of
the calix[4]pyrrole with the relevant N···F− distances being
2.77−2.81 Å. One methanol molecule is also hydrogen bonded
to the F− anion, the O···F− distance being 2.58 Å. In analogy to
what was seen for the complex of compound 2 with CsF, there
is no direct interaction between the Cs+ and F− ions bound to
receptor 1. While lattice effects may be playing a role, we take
this as meaning that the stabilization energy arising from the
formation of the complex, 1·CsF, is large enough to offset the
presumed Columbic energy penalty caused by what appears to
be an unfavorable ion separation. The distance between the Cs+
ion and the F− ion in the complex was found to be 10.29 Å
(Figure S11). Again, this leads us to suggest that the formation
of a strong complex, 1·CsF containing a methanol molecule, is
energetically more stable than the corresponding contact ion-
pair complex. This is true despite the large separation between
the Cs+ and F− ions.
1
Comparing the integral ratios in the H NMR spectrum of
the K+ complex of compound 5 with those seen in the
corresponding ion-free form revealed that ∼33% of the
receptor mixture takes part in KF complexation under
conditions of extraction. Because this number is ∼0% in the
absence of 4, we conclude that synergy between the anion
receptor 4 and the cation receptor 5 is necessary for the
effective extraction of KF. On the other hand, an equimolar
mixture of 4 and 5 appears less effective for extraction than the
ion-pair receptor 1, as evidenced by the fact that, to the limits of
NMR spectral analysis, ∼100% of the latter species participates
in KF extraction.
Further support for the KF and CsF binding modes inferred
for receptor 1 from the solution phase NMR spectroscopic
studies discussed above came from single crystal X-ray
diffraction analyses. Suitable single crystals of the CsF complex
were obtained by subjecting a chloroform/methanol solution of
receptor 1 to slow evaporation in the presence of excess cesium
fluoride. The resulting crystal structure revealed that receptor 1
forms a 1:1 complex with cesium fluoride, 1·CsF, in the solid
state wherein the Cs+ cation is bound to the calix[4]arene
crown-5 moiety (Figure 5). This binding mode is observed
despite the fact that the calix[4]arene crown-5 is too small to
bind the Cs+ cation strongly.10 The Cs+ ion is encapsulated by
the calix[4]arene crown ether ring with distances of 2.84−3.03
The structure of the KF complex of receptor 1 in the solid
state was also determined by X-ray crystallography. Suitable
single crystals were obtained by allowing a chloroform/
methanol solution of the CsF complex to undergo slow
evaporation in the presence of one molar equivalent of KClO4.
As shown in Figure 6, in the resulting structure, the K+ cation is
Figure 6. Two different views of the single crystal X-ray diffraction
structure of 1·KF·CH3OH. Solvent molecules not involved in
stabilization of the ion-pair complex have been removed for clarity.
bound to the crown ether at K+···O distances that vary between
2.77 and 2.87 Å. Evidence for a π-metal interaction between the
aromatic rings of the calix[4]arene and the K+ cation is seen in
the K+···C distances of 3.06−3.23 Å (Figure 6). The F− anion is
hydrogen bonded to the four pyrrolic NH protons with N···F−
distances of 2.76−2.79 Å. A methanol molecule is also bound
within the receptor and interacts with the fluoride anion. The
Figure 5. Two different views of the single crystal X-ray diffraction
structure of 1·CsF·CH3OH. Solvent molecules not involved in
stabilization of the ion-pair complex have been removed for clarity.
20842
dx.doi.org/10.1021/ja310673p | J. Am. Chem. Soc. 2012, 134, 20837−20843

Journal of the American Chemical Society
Article
relevant O···F− distance is 2.69 Å. The separation between the
Geosciences, and Biosciences, Office of Basic Energy Sciences,
U.S. DOE.
K+ and the cobound F− was found to be 10.74 Å (Figure S11).
REFERENCES
CONCLUSIONS
On the basis of both H NMR spectroscopic analyses and
■
■
1
(1) Kim, S. K.; Sessler, J. L. Chem. Soc. Rev. 2010, 39, 3784−3809
and references cited therein.
single crystal X-ray crystal structural data, we conclude that the
ion-pair receptor 1, which contains a calix[4]pyrrole for anion
binding and calix[4]arene crown-5 for cation recognition, is
able to complex both the CsF and the KF ion-pairs strongly.
However, the underlying binding occurs via very different
modes for these two salts. Specifically, in the case of KF, the
calix[4]arene crown-5 moiety binds the K+ cation first and then
the calix[4]pyrrole subunit binds the F− anion. This gives a
stable 1:1 ion-pair complex that generally precipitates from
solution. In contrast, the CsF ion-pair interacts with receptor 1
in two different ways. In one of these modes, the Cs+ cation
interacts with the calix[4]arene crown-5 ring weakly, albeit in
fast equilibrium. In this case, the counteranion is not cobound.
In the second interaction mode, which in 10% CD3OD in
CDCl3 is thermodynamically more stable, the Cs+ and F− ions
are simultaneously bound to the receptor framework through
the ethylene glycol spacers and the calix[4]pyrrole moiety,
respectively. In this case, dynamic binding behavior is seen in
10% CD3OD in CDCl3 with the Cs+ cation being bound first to
the ethylene glycol spacers and only subsequently by the
calix[4]arene crown-5 subunit. While forming complexes with
both CsF and KF, the present system displays ion-pair
selectivity. This is manifest by the fact that addition of
KClO4 to the CsF complex of 1 leads to cation metathesis and
formation of the more stable KF complex. The stability of the
latter ion-pair complex is reflected in the ability of receptor 1
function as more effective extractant for KF in a two-phase
aqueous-nitrobenzene test system than either 4 or 5, alone or in
combination. We thus conclude that suitably designed ditopic
receptors, such as 1, could offer advantages over simple
mixtures of the corresponding anion and cation binding
subunits.
(2) (a) McConnell, A. J.; Beer, P. D. Angew. Chem., Int. Ed. 2012, 51,
5052−5061. (b) Smith, B. D. In Ion-pair Recognition by Ditopic
Receptors, Macrocyclic Chemistry: Current Trends and Future Prospec-
tives; Gloe, K., Antonioli, B., Eds.; Kluwer: London, U.K., 2005; pp
137−152. (c) Kirkovits, G. J.; Shriver, J. A.; Gale, P. A.; Sessler, J. L. J.
Inclusion Phenom. Macrocyclic Chem. 2001, 41, 69−75.
(3) (a) Pfeifer, J. R.; Reiβ, P.; Koert, U. Angew. Chem., Int. Ed. 2006,
45, 501−504. (b) Sisson, A. L.; Shah, M. R.; Bhosale, S.; Matile, S.
Chem. Soc. Rev. 2006, 35, 1269−1286. (c) Nakamura, T.; Akutagawa,
T.; Honda, K.; Underhill, A. E.; Coomber, A. T.; Friend, R. H. Nature
1998, 394, 159−162. (d) Gokel, G. W.; Leevy, W. M.; Weber, M. E.
Chem. Rev. 2004, 104, 2723−2750. (e) Davis, A. P.; Sheppard, D. N.;
Smith, B. D. Chem. Soc. Rev. 2007, 36, 348−357.
(4) (a) Chrisstoffels, L. A. J.; De Jong, F.; Reinhoudt, D. N.; Sivelli,
S.; Gazzola, L.; Casnati, A.; Ungaro, R. J. Am. Chem. Soc. 1999, 121,
10142−10151. (b) Rudkevich, D. M.; Mercer-Chalmers, J. D.;
Verboom, W.; Ungaro, R.; Reinhoudt, D. N. J. Am. Chem. Soc.
1999, 117, 6124−6125. (c) Tong, C. C.; Quesada, R.; Sessler, J. L.;
Gale, P. A. Chem. Commun. 2008, 6321−6323.
(5) (a) Mahoney, J. M.; Stucker, K. A.; Jiang, H.; Carmichael, I.;
Brinkmann, N. R.; Beatty, A. M.; Noll, B. C.; Smith, B. D. J. Am. Chem.
Soc. 2005, 127, 2922−2928. (b) Deetz, M. J.; Shang, M.; Smith, B. D.
J. Am. Chem. Soc. 2000, 122, 6201−6207. (c) Mahoney, J. M.; Beatty,
A. M.; Smith, B. D. Inorg. Chem. 2004, 43, 7617−7621. (d) Mahoney,
J. M.; Davis, J. P.; Smith, B. D. J. Org. Chem. 2003, 68, 9819−6820.
(e) Mahoney, J. M.; Beatty, A. M.; Smith, B. D. J. Am. Chem. Soc. 2001,
123, 5847−5858. (f) Mahoney, J. M.; Nawaratna, G. U.; Beatty, A. M.;
Duggan, P. J.; Smith, B. D. Inorg. Chem. 2004, 43, 5902−5907.
(g) Mahoney, J. M.; Marshall, R. A.; Beatty, A. M.; Smith, B. D.;
Camiolo, S.; Gale, P. A. J. Supramol. Chem. 2003, 1, 289−292.
(6) Reeske, G.; Bradtmoller, G.; Schurmann, M.; Jurkschat, K. Chem.-
̈
̈
Eur. J. 2007, 13, 10239−10245.
(7) Kim, S. K.; Gross, D. E.; Cho, D. G.; Lynch, V. M.; Sessler, J. L. J.
Org. Chem. 2011, 76, 1005−1012.
(8) Kim, S. K.; Sessler, J. L.; Gross, D. E.; Lee, C.-H.; Kim, J. S.;
Lynch, V. M.; Delmau, L. H.; Hay, B. P. J. Am. Chem. Soc. 2010, 132,
5827−5836.
ASSOCIATED CONTENT
■
S
* Supporting Information
(9) Sessler, J. L.; Kim, S. K.; Gross, D. E.; Lee, C.-H; Kim, J. S.;
Lynch, V. M. J. Am. Chem. Soc. 2008, 130, 13162−13166.
1H NMR spectroscopic data, ITC analyses, and X-ray structural
data for 1·KF·(CH3OH)3 (CCDC 826579) and
1·CsF·(CH3OH)2·(CHCl3)2 (CCDC 826578). This material
is available free of charge via the Internet at http://pubs.acs.org.
(10) Kim, S. K.; Vargas-Zuniga, G. I.; Hay, B. P.; Young, N. J.;
́
̃
Delmau, L. H.; Masselin, C.; Lee, C.-H.; Kim, J. S.; Moyer, B. A.;
Lynch, V. M.; Sessler, J. L. J. Am. Chem. Soc. 2012, 134, 1782−1792.
(11) Sato, W.; Miyaji, H.; Sessler, J. L. Tetrahedron Lett. 2000, 41,
6731−6736.
AUTHOR INFORMATION
Corresponding Author
sessler@cm.utexas.edu; haybp@ornl.gov; chhlee@kangwon.ac.
kr; moyerba@ornl.gov; jongskim@korea.ac.kr
■
(12) Custelcean, R.; Delmau, L. H.; Moyer, B. A.; Sessler, J. L.; Cho,
W.-S.; Gross, D.; Bates, G. W.; Brooks, S. J.; Light, M. E.; Gale, P. A.
Angew. Chem., Int. Ed. 2005, 44, 2537−2542.
(13) The nature of the strap calix[4]pyrroles is thought to favor the
cone conformation. See: Lee, C.-H.; Miyaji, H.; Yoon, D.-W.; Sessler,
J. L. Chem. Commun. 2008, 24−34.
(14) Marcus, Y. J. Chem. Soc., Faraday Trans. 1991, 87, 2995−2999.
(15) This observation does not address the background effects (if
any) of KF transfer in the organic phase. It informs us only about the
fate of the receptor. Nevertheless, it does allow conclusions about the
efficiency of KF binding and extraction to be inferred.
(16) Previously, we had shown that receptor 1 may be used to effect
the extraction of KCl. See ref 10.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work was supported by the Office of Basic Energy
Sciences, U.S. Department of Energy (DOE) (grant DE-FG02-
01ER15186 to J.L.S.), the Korea NRF grant (MEST 2009-
0087013 to C.-H.L.), CRI project of KRF (No. 2011-0000420
to J.S.K.), and the Korean WCU program (R32-2010-000-
10217-0) administered through the National Research
Foundation of Korea funded by the Ministry of Education,
Science and Technology (MEST). B.P.H., B.A.M., and N.J.Y.
acknowledge support from the Division of Chemical Sciences,
20843
dx.doi.org/10.1021/ja310673p | J. Am. Chem. Soc. 2012, 134, 20837−20843