Thermodynamic characterization of halide-π interactions in solution using "two-wall" aryl extended calix[4]pyrroles as model system

Article abstract of DOI:10.1021/ja412098v

Herein, we report our latest experimental investigations of halide-π interactions in solution. We base this research on the thermodynamic characterization of a series of 1:1 complexes formed between halides (Cl ^-, Br^-, and I^-

Full text of DOI:10.1021/ja412098v

Article
pubs.acs.org/JACS
Thermodynamic Characterization of Halide−π Interactions in
Solution Using “Two-Wall” Aryl Extended Calix[4]pyrroles as Model
System
Louis Adriaenssens,† Guzman Gil-Ramírez,† Antonio Frontera,‡ David Quinonero,‡
́
̃
Eduardo C. Escudero-Adan,† and Pablo Ballester*,†,§
́
†Institute of Chemical Research of Catalonia (ICIQ), Avgda Països Catalans 16, 43007 Tarragona, Spain
‡Departament de Química, Universitat de les Illes Balears, Crta Valldemossa Km 7.5, 07122 Palma, Spain
§Catalan Institution for Research and Advanced Studies (ICREA), Passeig Lluís Companys 23, 08010 Barcelona, Spain
S
* Supporting Information
ABSTRACT: Herein, we report our latest experimental investigations of
halide−π interactions in solution. We base this research on the
thermodynamic characterization of a series of 1:1 complexes formed
between halides (Cl−, Br−, and I−) and several α,α-isomers of “two-wall”
calix[4]pyrrole receptors bearing two six-membered aromatic rings in
opposed meso positions. The installed aromatic systems feature a broad
range of electron density as indicated by the calculated values for their
electrostatic surface potentials at the center of the rings. We show that a
correlation exists between the electronic nature of the aromatic walls and
the thermodynamic stability of the X−⊂receptor complexes. We give
evidence for the existence of both repulsive and attractive interactions
between π systems and halide anions in solution (between 1 and −1 kcal/
mol). We dissect the measured free energies of binding for chloride and
bromide with the receptor series into their enthalpic and entropic thermodynamic quantities. In acetonitrile solution, the binding
enthalpy values remain almost constant throughout the receptor series, and the differences in free energies are provoked
exclusively by changes in the entropic term of the binding processes. Most likely, this unexpected behavior is owed to strong
solvation effects that make up important components of the measured magnitudes for the enthalpies and entropies of binding.
The use of chloroform, a much less polar solvent, limits the impact of solvation effects revealing the expected existence of a
parallel trend between free energies and enthalpies of binding. This result indicates that halide−π interactions in organic solvents
are mainly driven by enthalpy. However, the typical paradigm of enthalpy−entropy compensation is still not observed in this less
polar solvent.
with the more established noncovalent interaction of cations
with π-systems, cation−π interactions.7
INTRODUCTION
■
In 1986, Chowdhury and Kebarle reported gas-phase studies
indicating that hexafluorobenzene (C6F6), in spite of not having
hydrogen atoms binds chloride quite strongly.1 Thermody-
namically stable 1:1 gas-phase adducts with the general formula
[C6F6−X]− for X− = Cl−, Br−, and I− could be detected and
electronic structure calculations supported the location of the
anion above the π-cloud of the aromatic system.2 In addition,
thermochemical data for the gas-phase clustering reactions of
the halide ions with C6F6 indicated that the bonding strength in
the [C6F6−X]− complexes derived from electrostatic (Cou-
lombic) interaction. These interesting findings remained
dormant in the literature until 2002, when four computational
studies reported the existence of general attractive interactions
between anions centered over the face of π-systems in six
membered cyclic3−5 and heterocyclic6 arenes featuring positive
quadrupole values. The term “anion−π interaction” was coined
to describe this binding motif due to its close structural analogy
While attractive cation−π interactions are in tune with
chemical intuition, the attractive interaction between an anion
and a π system is decidedly contrary to the normal chemist’s
instinct. Indeed, since their introduction to the greater chemical
community in 2002, anion−π interactions have inspired a large
number of publications. Numerous and extensive investigations
on the physical nature of anion−π interactions and other
influencing effects (hydrogen bonds, cations, ring size, etc.)
have been performed using electronic structure methods.
Anions nestled in the π electron cloud of aromatic rings in
single-crystal X-ray structures of supramolecular complexes are
frequently highlighted as evidence of attractive anion−π
interactions.8 Several thorough searches of the Cambridge
Structural Database (CSD)3,4,9−11 and Protein Data Bank
Received: November 27, 2013
Published: February 4, 2014
© 2014 American Chemical Society
3208
dx.doi.org/10.1021/ja412098v | J. Am. Chem. Soc. 2014, 136, 3208−3218

Journal of the American Chemical Society
Article
(PDB)12 have been carried out to demonstrate the existence of
anion−π close contacts in archived solid-state structures.
However, experimental evidence supporting the existence of
attractive anion−π interactions involving charge-neutral
aromatic systems in solution is still scarce.9,13−17 Gas-phase
and theoretical studies assign bond energies for supramolecular
anion−π interactions in the range of 4−17 kcal mol−1, quite
close to those characterizing some cation−π interactions.18,19
Meanwhile, the values of the stability constants experimentally
determined for complexes involving anion−π interactions in
solution often do not provide an indisputable assessment of the
magnitude of the interaction in this phase.20−27 So far, this
limitation has prevented a confident comparison of the energies
that can be assigned to anion−π interactions in solution with
respect to their cation−π counterparts. In any case, anion−π
interactions have been claimed as a new supramolecular force
opening new opportunities for the design of sensors,
supramolecular hosts, catalysts, and materials.28 Functional
systems based on anion−π interactions have also been
reported.26,29−33
We became interested in the experimental study and
quantification of anion−π interactions in solution with the
aim of increasing the understanding of their physical nature,
their magnitude, and their potential applications. The extensive
use of benzene panels in shaping the cavity of many molecular
and supramolecular containers, i.e., cavitands derived from
calix[4]arenes, resocin[4]arenes, and cycloveratrilenes, centered
our attention on the quantification of anion−π interactions
involving substituted six-membered (phenyl) and charge-
neutral rings. In 2008, we designed and synthesized a series
of anion receptors based on the α,α,α,α-stereoisomer of aryl-
extended calix[4]pyrroles 1 bearing four meso-phenyl sub-
stituents “four-wall” calixpyrroles (Figure 1, left).34 Different
This series of “four-wall” calix[4]pyrrole receptors 1 was also
used to quantify chloride−π(phenyl) interactions in acetonitrile
solution. Quantification was achieved by taking the chloride
complex of octamethyl calix[4]pyrrole (Cl−⊂2) as a reference
for the binding energy attributed to the primary hydrogen
bonding interaction between the chloride anion and the
calix[4]pyrrole core.35 The difference in binding energy
between “four-wall” complexes Cl−⊂1 and “no-wall” reference
Cl−⊂2 (ΔΔG values) represents the contribution to binding
from the interaction between the “four walls” of receptors 1 and
the chloride anion. These values correlated well with the
corresponding Hammett constants for the para substituents.
This demonstrated that the observed chloride−π(phenyl)
interactions were mainly dominated by electrostatics. In
acetonitrile solution, this simple model system allowed us to
determine that the interaction between Cl− and a p-NO2-
substituted phenyl group was very slightly attractive; we
assigned a magnitude of ca. −0.1 kcal mol−1 to this interaction.
In this paper, we report the further study of anion−π
interactions in solution using a series of “two-wall” calix[4]-
pyrrole receptors 3. The “two-wall” receptors 3 have several
advantages over the “four-wall” analogues. These advantages
enable a much more detailed analysis of anion−π interactions.
For example, they allow multiple electron-withdrawing
substituents to be placed on each of the two meso-phenyl
walls. In this way, contrary to the original “four-wall” receptors
1, one can incorporate two meta nitro groups or five fluorine
atoms around the meso-phenyl substituents in 3. Furthermore,
additional meso-phenyl substituents were incorporated in
receptor series 3, i.e., pyridyl and p-azidophenyl. The availability
of these new receptors widens the range of aromatic
electrostatic surface potentials (ESP) or quadrupole values
that can be interrogated in relation to the energetic magnitude
of their interaction with chloride. The deletion of two phenyl
meso substituents in the “two-wall” receptor series 3 provided
an important increase in the thermodynamic stability of the
corresponding chloride complexes compared to the their “four-
wall” counterparts 1 that feature negative values of ESP at the
center of the phenyl rings. This finding was significant as it
rendered the series of “two-wall” receptors 3 a suitable model
system to investigate the interaction of bromide and iodide with
various aromatic rings. These two halides formed complexes
with the four walls receptors 1, deprived of electron-
withdrawing groups in their phenyl substituents, which were
thermodynamically too weak for the accurate quantification of
their binding constants.36 Finally, in the study presented here
we disclose data regarding the enthalpic and entropic
components of the anion−π interaction. By dissecting the
free binding energy in this way we were able to come to
important conclusions regarding dominant factors controlling
anion−π interactions and the ability of our model system to
analyze and quantify anion−π interactions in solution.
Figure 1. Molecular structures of the inclusion complexes of chloride
with “four-wall” aryl extended calixpyrroles 1 (left), “no-wall”
octamethylcalix[4]pyrrole 2 (middle), and “two-wall” aryl extended
calix[4]pyrroles 3 (right).
para substituents were used to modify the charge density and
charge distribution on the surface of the π-system. The extent
to which this modification affected the electronics of the
aromatic ring was quantified using the corresponding Hammett
constants for the para substituents. The formation of four
hydrogen bonds between the chloride ion and the NH groups
of the calix[4]pyrrole unit in 1 induces the receptor to adopt
the cone conformation which situates the anion in the deep
aromatic cavity paneled by four fixed phenyl walls. The total
binding energy, dominated by the strong hydrogen bonding
interactions present in the inclusion complexes Cl−⊂1, was
modulated by the weaker anion−π interactions. Thus, the
association constant values determined for the series Cl−⊂1
increased with the electronic-withdrawing character of the para
substituent.
RESULTS AND DISCUSSION
■
Synthesis and Characterization of the Receptor
Series. The synthesis of “two-wall” calix[4]pyrrole receptors
α,α-3 was accomplished using two subsequent acid-mediated
condensations (Scheme 1). The first, using pyrrole as the
solvent and a methyl phenyl ketone 4, constructed the aryl-
substituted dipyromethane 5, and the second, using acetone as
solvent, effected cyclocondensation to give “two-wall” calix[4]-
pyrrole receptors 3.37 The methoxy-functionalized receptor
α,α-3a was obtained through the reaction of the parent
3209
dx.doi.org/10.1021/ja412098v | J. Am. Chem. Soc. 2014, 136, 3208−3218

Journal of the American Chemical Society
Article
methane, acetonitrile, or acetone solutions revealing that the
calix[4]pyrrole core adopted a 1,2 alternate conformation in the
solid state (Figure 3).
Scheme 1. Synthetic Scheme for the Preparation of the
Series of Methylaryldipyromethanes 5b−i and “Two-Wall”
Calix[4]pyrroles 3a−i
Figure 3. X-ray structures of (a) α,α-3b, (b) α,α-3c, (c) α,α-3h, and d)
α,α-3g. Calix[4]pyrroles all in 1,2 alternate conformation. The α,α-3
calixpyrroles are depicted in stick representation.
1H NMR Titration Experiments with the Halides.
Geometries of the Halide Complexes in Solution and
in the Solid State. 1H NMR spectroscopy was used to probe
the binding of calixpyrroles α,α-3a−h40 with chloride in
acetonitrile solution.
1
In particular, H NMR was useful in obtaining structural
information on the geometry of the complexes and their
binding dynamics. We used tetrabutylammonium chloride
(TBACl) as the source of the guest chloride anion. In
acetonitrile solution and in the range of concentrations used,
the tetrabutylammonium salt is highly dissociated and is
expected to produce simple anionic 1:1 complexes Cl−⊂α,α-
3.41
In a typical example, the initial addition of 0.5 equiv of
TBACl to an acetonitrile solution of α,α-3c (1.1 mM) induced
shifting and broadening of most of the proton signals of the
receptor (Figure 4).
hydroxyl compound α,α-3i with methyl iodide in DMF solution
using cesium carbonate as base. The syntheses of 3b,26 3e,26
3f,38 3g,26 and 3h39 have been previously described by our
group and others via similar routes.
1
The H NMR spectra of calixpyrroles α,α-3 in CD3CN
solution display two signals for the β-pyrrolic protons
resonating close to 5.8 ppm (H3&4, Figure 2). The pyrrole
NH protons appear downfield shifted at δ ≈ 8 ppm, probably
due to hydrogen bonding with solvent molecules.
Exceptions to this general behavior were observed in the
titrations of receptors α,α-3h (3,5-(NO2)2-Ph) and α,α-3g
(C6F5) with chloride. These two receptors contain aromatic
walls with multiple, strong electron-withdrawing substituents.
Figure 2. Selected regions of the 1H NMR spectra of two-wall
calixpyrrole α,α-3c in CD3CN solution at 298 K. The molecular
structure indicates proton assignments.
All proton signals are sharp and well resolved. The number of
resonances is indicative of either a fast interconversion between
conformers (alternate and cone) or a single locked con-
formation in solution. Thus, consistent with their symmetry,
the α,α-3a−h isomers displayed three signals with the same
integral values for the meso methyl protons. The structures of
several α,α-3 and α,β-3 (see the Supporting Information)
calixpyrroles were also confirmed by X-ray crystallography.
Single crystals of the α,α-3 isomers grew from dichloro-
1
Figure 4. Selected regions of the H NMR spectra acquired for the
titration of α,α-3c with TBACl in CD3CN solution at 298 K. [α,α-3c]
= 1.1 mM. See Figure 2 for proton assignment; primed numbers
indicate proton signals for the bound receptor.
3210
dx.doi.org/10.1021/ja412098v | J. Am. Chem. Soc. 2014, 136, 3208−3218

Journal of the American Chemical Society
Article
In these two cases, the addition of 0.4 equiv of TBACl induced
the appearance of a new set of proton signals alongside the
signals belonging to the free receptor. We assigned the new set
of signals to protons of the bound receptor in the inclusion
complexes Cl−⊂α,α-3g&h. Separate signals for the fluorine
atoms in the free and bound receptor α,α-3g were also detected
in the 19F NMR spectra (Figure 5).
CH···Cl− hydrogen bonds. The upfield shifts observed for the
signals corresponding to the fluorine atoms in bound α,α-3g
(C6F5) are also consistent with the existence of chloride−π
interactions. No significant changes were detected for the
chemical shift values of the proton signals corresponding to the
TBA cation. Taking into consideration all the chemical shift
changes, we propose the formation of anionic complexes
Cl−⊂α,α-3 in acetonitrile solution featuring a binding geometry
in which the calix[4]pyrrole core of the receptor adopts the
cone conformation and forms four hydrogen bonds between
the pyrrole NHs and the bound chloride. The bound chloride is
sandwiched in the cleft defined by the two axially oriented
meso-phenyl groups (Figure 6, a). This binding geometry is
reminiscent of that displayed by the inclusion complexes of
chloride with the four-wall analogues Cl−⊂α,α,α,α-1 (Figure 6,
b).34
Figure 6. (a) CAChe-minimized structure showing the proposed
geometry for the anionic complex Cl−⊂α,α-3f; (b) X-ray structure of
the analogous four-wall complex Cl−⊂α,α,α,α-1f.34
1
1
Figure 5. Selected regions of the H NMR spectra acquired for the
It is worth noting that, while the H NMR signal for the β-
titration of α,α-3g (C6F5) with TBACl in CD3CN solution at 298 K.
[α,α-3g] = 1.3 mM. The corresponding 19F NMR spectra are depicted
as insets. See Scheme 1 for proton assignments; primed symbols
indicate proton and fluorine signals for the bound receptor. Arrows are
used to allocate this set of proton signals in the slow chemical
exchange.
pyrrole protons of the four wall receptors α,α,α,α-1 moved
systematically upfield on chloride binding, the signals for the β-
pyrrole protons H3 and H4 of the α,α-3 receptors, with the
exception of α,α-3g, shifted in opposite directions upon the
introduction of chloride. The signal for H3, located in closer
proximity to the phenyl ring than H4, experienced downfield
shifts while the signal for H4 was upfield shifted. This
observation suggests the existence in solution of a conforma-
tional change of the calix[4]pyrrole core of the “two wall”
receptors α,α-3, from alternate to cone conformation upon
chloride binding. The solid-state structures are also in support
of this hypothesis since, in all cases, they show the free α,α-3
receptors in the alternate conformation (Figure 3). However, in
the inclusion complexes Cl−⊂α,α-3 the receptors are in cone
conformation (Figure 7) vide infra.42 This is in contrast with
the solid state and solution observations for “four-wall”
receptors, α,α,α,α-1, which indicate that the cone conformer
is preferred both in the free and bound state.
1
In the H NMR titrations of all receptors α,α-3a−h, the
addition of 1 equiv of TBACl gave rise to the observation of a
single set of sharp proton signals corresponding to the receptor
in the bound state. The addition of more than 1 equiv of
TBACl did not induce additional changes to the proton signals
of the receptors. Taken together, the above-described results
indicate that the association constant values for the complexes
of the receptors α,α-3a−h binding Cl− are too high to be
1
measured accurately with H NMR titrations (>104 M−1) and
that receptors α,α-3h (di-NO2) and α,α-3g (F5) form
complexes with Cl− that are kinetically stable on the 1H
NMR time scale.
The formation of anionic complexes Cl−⊂α,α-3 was also
detected by mass spectrometry using electrospray ionization
(ESI) and negative detection mode. Receptors α,α-3 produced
anionic complexes with chloride in the gas phase. Using very
mild ionization conditions (almost zero voltage), it was possible
to detect the peak corresponding to the mass of the Cl−⊂α,α-3
complexes as the exclusive ion in the 200−1000 m/z range.
As briefly mentioned above, the binding geometry we
propose for the inclusion complexes Cl−⊂α,α-3 in solution is in
complete agreement with the structures observed for the
corresponding chloride complexes in the solid state. Single
crystals of the 1:1 complexes of several α,α-3 receptors with
tetramethylammonium chloride (TMACl) were obtained by
slow evaporation of TMACl-saturated acetonitrile or acetone
While NMR spectroscopy was unsuitable for the accurate
measurement of binding constants, the technique offered
important information about the geometry of the binding
complexes Cl−⊂α,α-3 in acetonitrile solution. Upon chloride
binding, the proton signal of the pyrrole NHs (H8) was shifted
significantly downfield (Δδ = 3.00 to 3.23 ppm) while the
signals corresponding to the protons of the aromatic walls
(H1&2) moved upfield (Δδ ≈ −0.1 ppm). The downfield shift
experienced by the signal belonging to the NHs is consistent
with the formation of four hydrogen bonds with the chloride
anion. The upfield shift registered for the aromatic protons
constitutes clear evidence for the interaction of the chloride ion
with the face of the π-system and not with the CH groups via
3211
dx.doi.org/10.1021/ja412098v | J. Am. Chem. Soc. 2014, 136, 3208−3218

Journal of the American Chemical Society
Article
support their use as a model system for the quantification of
anion-π interactions in solution. In all solid state structures of
the chloride complexes Cl−⊂α,α-3, the packing of the crystals
revealed that the tetramethyl ammonium cation is in close
contact with the bound chloride anion and included into the
electronically rich aromatic cavity defined by the pyrrole rings
of an adjacent and identical calix[4]pyrrole unit (Figure 7).
The high thermodynamic stability exhibited by the chloride
complexes of the “two-wall” receptors α,α-3 (estimated Ka
values higher than 104 M−1 independently of the electron
character of the meso-phenyl rings) prompted us to investigate
the binding properties of the “two-wall” system with other
halides. The complexation of bromide and iodide with
1
receptors α,α-3b-h was first studied using H NMR titrations.
The chemical shift changes experienced by the receptor series
upon binding bromide or iodide parallel those observed for
chloride (Figure S9 (Supporting Information) for I− and Figure
S10 (Supporting Information) for Br−). The most notable
difference in the binding behavior with the larger halides was
the magnitude of the complexation-induced shift experienced
Figure 7. (a) X-ray structure of the Cl−⊂α,α-3h complex. (b)
Expansion (side and top views) of the region relevant to the chloride-π
interaction; important geometrical parameters between the Cl− and
the aromatic ring are indicated. (c) Partial packing of the X-ray
structure of TMA·Cl−⊂α,α-3h. Distances in angstroms.
1
by the H NMR signal corresponding to the pyrrole NHs. As
the halide radii increased, the difference between the free and
bound chemical shift values of the NH protons decreased.
Thus, the change in chemical shift for the NH protons binding
bromide is Δδ ≈ 2.5 ppm and only Δδ ≈ 1.9 ppm for iodide,
compared to Δδ = 3.0 to 3.2 ppm for chloride.45 This behavior
is anticipated given the expected trend in hydrogen bond
strength between different halides and the calix[4]pyrroles 3
(I− weaker than Br− weaker than Cl−).
solutions of the host. The solid-state structures of the inclusion
complexes revealed that the Cl− atom is not located directly
perpendicular to the centroid of the phenyl rings but somewhat
offset. The primary hydrogen bonding interactions formed
between the Cl− and the pyrrole NHs (dCl ···N ≈ 3.3 Å)
−
Single crystals suitable for X-ray analysis grew from
acetonitrile solutions of receptor α,α-3g (F5) saturated with
tetramethylammonium chloride, bromide, and iodide salts. The
X-ray structures revealed the formation of inclusion complexes
TMA·X−⊂α,α-3g with structural features similar to those
described above (Figure 8).The comparison of the crystal
structures of the inclusion complexes of the three halides with
the same receptor TMA·X−⊂α,α-3g exposed several similarities
and differences. The general binding geometry of the 1:1
complexes and the conformation adopted by the receptor is
almost identical independently of the included anion. The
distance between the centroids of the C6F5 aromatic rings is
maintained more or less constant throughout the series of
complexes (dcentroids ≈ 7.9 Å).This result demonstrated that the
inclusion of the bigger iodide anion is not sterically hindered.
The increase in radius of the anion is also responsible for the
observed change in distance between the centroid defined by
the four N atoms of the pyrrole rings and the center of the
halide. In turn, this difference in the placement of the halide
translates to a modification of the tilt angle θ that characterizes
the position of the anion with respect to the center of the
aromatic C6F5 ring. For the TMA·I−⊂α,α-3g complex the
geometry of the iodide−π interaction is very close to that
predicted by theory, having the I− in contact with five of the
ring carbons to within ∑vdW + 0.4 Å and a tilt angle of θ ≈
80°. Taken together, these results show that in the solid state,
bromide, iodide and chloride behave similarly in the way that
they bind two-wall calix[4]pyrrole receptors α,α-3 and that the
receptor series constitutes a valuable and valid model system for
the quantification of halide−π interactions in solution.
dominate and pull the anion deep inside the aromatic cleft. The
geometries of the Cl−/arene contacts in the complexes are
characterized by tilt angles (θ) in the range of 66° to 70° and
distances between the mean plane defined by the six carbon
atoms of the arene and the Cl− (dplane) larger than 3.6 Å (3.60−
3.99 Å) (see Figure 7).
The observed geometrical parameters of the chloride−π
interactions in solid-state structures of the “two wall” complexes
are not in ideal agreement with those calculated for chloride−π
complexes involving C6 arenes.3,4 In the calculated structures,
the Cl− is located above the ring centroid and in close contact
with all six arene carbon atoms; that is, all Cl−···C contacts have
a distance < ∑vdw radii (1.70 Å + 1.75 Å = 3.45 Å). This
translates to a dplane value <3.16 Å. However, taking into
consideration the recent theoretical results described by
Estarellas et al.,43 which indicate that anion−π interactions
have less restricted directionality preferences than cation−π
interactions and their proposal of the existence of anion−π
contacts at distances < ∑wdw radii +0.4 Å = 3.85 Å, the crystal
structures of the inclusion complexes Cl−⊂α,α-3 can be
considered as convincing examples of Cl− interacting with C6
π systems.
An experimental finding that strongly supports this latter
hypothesis is based on the trend observed for dplane in these
structures. Generally, this distance decreases as the number and
strength of the electron-withdrawing substituents of the ring
increase. For example, for the complex Cl−⊂α,α-3a (OMe, see
Figure S3, Supporting Information) dplane = 3.99 Å. For the
complex Cl−⊂α,α-3h (3,5-(NO2)2-Ph, Figure 7) this distance
drops to dplane = 3.60 Å.44 We relate this distance modulation to
the strength and character, repulsive or attractive, of the Cl−−π
interactions present in the Cl−⊂α,α-3 complexes. In short, the
geometrical parameters of the Cl−-arene contacts encountered
in the solid-state structures of the complexes Cl−⊂α,α-3
1
Using H NMR spectroscopy (receptor concentration ≈1
mM), “two-wall” receptors lacking electron-withdrawing
substituents on their meso-phenyl walls α,α-3a (OMe) and
α,α-3b (H) did not show detectable interactions with iodide.46
3212
dx.doi.org/10.1021/ja412098v | J. Am. Chem. Soc. 2014, 136, 3208−3218

Journal of the American Chemical Society
Article
titration experiments are summarized in Table 1. However,
similarly to their interaction with chloride, receptors α,α-3g and
α,α-3h containing the most electron-poor aromatic walls
showed binding with bromide that was too strong to be
accurately measured by NMR spectroscopy.
Isothermal Titration Calorimetry Experiments. Ther-
modynamic Characterization of the Inclusion Com-
plexes X−⊂α,α-3 and the Halide−π Interaction. For the
accurate determination of the binding constant values that were
1
estimated to be higher than 104 M−1 from the results of H
NMR titration experiments, we relied on isothermal titration
calorimetry (ITC) experiments.47,48 An ITC experiment
affords, in a single run, information on the enthalpy and
entropy of the interaction, complex stoichiometry, and
association constant values. All the ITC experiments performed
with the α,α-3 receptor series and chloride showed excellent fits
to the theoretical binding isotherm for a 1:1 complex
formation. We include in Figure 9 two representative examples
Figure 8. Front, side, and top views of the X-ray structures of 1:1
inclusion complexes of receptor α,α-3g with tetramethylammonium
halide salts: (a) Cl−, (b) Br−, and (c) I−. Selected geometrical
parameters for one of the two anion−π contacts are indicated. The
black dots represent calculated centroids. The structures are shown
with all atoms in stick representation. Using the criteria introduced in
ref 43 for the existence of anion−π contacts, the crystal structures of
the three inclusion complexes X−⊂α,α-3g can be considered as
convincing examples of X− interacting with C6 π systems. Distances in
angstroms.
1
However, H NMR spectroscopic titrations of the examples
comprised of more electron-poor walls proved to be suitable for
the determination of binding constants. Thus, using this
technique, we obtained binding data for the interaction of
iodide with receptors α,α-3d,f,g,h. 1H NMR spectroscopic
titration was also a viable technique to determine the binding
constants for some of the “two wall” calixpyrrole receptors with
Figure 9. (Top) Raw data (heat vs time) for the ITC titrations of (a)
α,α-3b (R = H) and (b) α,α-3h (di-NO2) receptors with TBACl.
(Bottom) Normalized integration of heat vs Cl−/α,α-3 molar ratio; the
fit of the 1:1 theoretical binding isotherm (red line) to the
experimental data is also shown. The values returned in the fitting
procedure for all ITC experiments are summarized in Tables 1 and 2.
[α,α-3b] = 1.3 mM and [α,α-3h] = 0.1 mM.
1
bromide. The thermodynamic data obtained in the H NMR
Table 1. Association Constant Values (Ka, M−1) and Free Energies of Complexation (ΔG, kcal/mol) Measured in MeCN at 298
K for the Inclusion Complexes of the Three Halide Anions with the Receptor Series α,α-3 and the Reference Octamethyl
a
Calix[4]pyrrole 2, and Statistically Corrected Free Energy Values Calculated for the Anion−π Interactions (ΔΔG, kcal/mol)
halide
chloride
bromide
ΔG
iodide
ΔG
b
e
e
e
receptor
2(Me)
ESP
n/a
Ka × 10−3
ΔG
ΔΔG
Ka × 10−3
ΔΔG
Ka
ΔΔG
d
d
c
108
−6.9
−5.5
−6.0
−6.6
−6.9
−6.5
−7.4
−7.8
−8.5
n/a
3.6
−4.9
n/a
12.6
−1.5
n/a
d
3a(OMe)
3b(H)
−16.1
−15.9
−9.7
−9.2
−8.9
−0.9
7.7
10.6
0.7
0.4
d
c
26.5
0.8
−3.9
−4.7
−4.6
−4.3
−5.0
−6.2
−6.3
0.5
0.1
d
c
3c(Br)
67.8
0.1
2.8
d
d
c
c
3d(Py)
124
0.0
2.3
0.1
18.2
33.5
−1.7
−2.1
−3.5
−0.1
−0.3
−1.0
d
c
3e(N3)
60.1
0.2
1.4
0.3
d
d
3f(NO2)
3g(C6F5)
3h(di-NO2)
277
−0.3
−0.5
−0.8
4.7
−0.1
−0.6
−0.7
d
d
553
32.3
f
d
d
c
11.8
1790
39.5
380
a
ITC titration experiments were repeated at least twice, and the reported Ka is the mean of the values obtained from the fit of the integrated heat
data to a 1:1 binding model. NMR titration experiments were also repeated at least twice, and the reported Ka is the mean of the values obtained
from the fit of the chemical shift changes observed for the proton signals to a 1:1 binding model. Errors for ΔG are assumed to be in the region of
b
c
d
20%. Value of the electrostatic surface potential at the center of the aromatic ring. Determined by NMR spectroscopic titration. Determined by
e
f
−
−
ITC. ΔΔG = (ΔGX ⊂3 − ΔGX ⊂2)/2. Data from the titration of α,α-3g with TBAI failed to fit well to a 1:1 model (see the Supporting Information
for details).
3213
dx.doi.org/10.1021/ja412098v | J. Am. Chem. Soc. 2014, 136, 3208−3218

Journal of the American Chemical Society
Article
of ITC experiments. The association constant values for all
Cl−⊂α,α-3 inclusion complexes and the bromide inclusion
complexes with octamethyl calix[4]pyrrole 2 (Me), α,α-3e
(Py), α,α-3f (NO2), α,α-3g (F5), and α,α-3h (di-NO2) were
determined using ITC experiments.
As alluded to above, in acetonitrile solution, and at the
concentration range used for the ITC titrations, both the salt
used as the source for the chloride/bromide/iodide anion
(TBA+·X−) and the resulting complex (X−⊂α,α-3·TBA+) are
considered to be fully dissociated.41,49,50 This is evidenced by
the fact that the halide-binding constants measured for TMA
and TBA salts are the same. Thus, the equation we used for the
calculation of the stability constants of the 1:1 complexes
formed by receptors α,α-3 and halide anions (X− = Cl−, Br−,
I−) in acetonitrile solution is as follows and has units of M−1.
supported by the solid state structures of the Cl−⊂α,α-3
complexes where it can be clearly seen that the distances
between the pyrrole nitrogen and the chloride remain constant
throughout the receptor series.
Thus, we considered the values of the free energies of
binding for the complexes X−⊂α,α-3 as the sum of two
different and independent types of intermolecular interaction:
(a) hydrogen bonding and (b) anion−π interaction. On the
basis of this hypothesis and for any one of the investigated
halides, the differences in the free energies of binding calculated
between any two of the complexes X−⊂α,α-3 provide a direct
measurement of the relative interaction energy of the anion
with the different meso-aromatic π-systems.
The free energies of binding determined for the complexes of
“no-wall” octamethyl calix[4]pyrrole with the different halides
X−⊂2 constitute suitable references of the contribution solely
from hydrogen bonding to the binding of halides with the
receptor series α,α-3. By subtracting this value from the free
energy of binding for the same halide with a “two-wall”
calix[4]pyrrole α,α-3 (eq 2) we calculated the contribution to
total binding from the anion−π interactions. This value was
then statistically corrected (divided by two) to calculate the
value of the anion−π interaction between the halide and one of
the various meso-aromatic systems in the complexes X−⊂α,α-3
(Table 1).
[X−⊂α,α‐3]
Ka =
[X−][α,α‐3]
(1)
Association constant values Ka and the derived Gibbs free
energies (ΔG) for the 1:1 inclusion complexes of the α,α-3
receptor series with chloride, bromide, and iodide anions in
acetonitrile at 298 K are listed in Table 1. For a given α,α-3
receptor, the stability constant value determined for the iodide
complex is predictably smaller than that with bromide which, in
turn, is predictably smaller than that with chloride. This
observation supports the importance of electrostatics (mainly
hydrogen bonds) in the overall binding affinity of the receptor
series for the halides. The influence of the type of halide on the
energetics of the anion−π interaction will be discussed below.
Traditionally, Hammett plots have been used to rationalize
and quantify electrostatic trends in experimental data. We and
others51 have noticed that the Hammett constants of meta-
substituents attached to aromatic rings, σm, show a linear
correlation with the electrostatic surface potential (ESP) values
calculated at the center of those aromatic rings. Similar plots
using B3LYP/6-31G* level calculations show that the same
trend exists for Hammett constants of para-substituents, σp. In
the work presented here, there is a mixture of para
monosubstitution and meta-disubstitution of phenyl rings, in
addition to the pentafluorophenyl and pyridyl groups. To better
describe this diverse range of aromatic systems, we calculated
the ESP values at the center of the ring in lieu of using the
Hammett constants. In this way, we simply generated a scale
that expresses the electronic properties of the surfaces above
the meso-aromatic rings employed in this study.52 The values of
the calculated ESPs at the center of each meso-aromatic group
are listed in Table 1 and become more positive going through
the series of receptors from α,α-3a to α,α-3h.
ΔG(X− ⊂ 3) − ΔG(X− ⊂ 2)
ΔΔG =
(2)
2
In this calculation, we assume that the anion−π interactions
are additive. That is, there is little to no effect caused by one
anion−π interaction on the other. Such an assumption seems
reasonable given theoretical studies on the interactions of
halides with triazine rings, which suggest that anion−π
interactions are indeed additive.54
For the three halides, the determined ΔΔG values correlate
well with the ESP values calculated at the centroid of the
respective aromatic walls (Figure 10).55
A likely explanation to the observed trend is consistent with
anion−π interactions being governed by electrostatics. The
We have mentioned above that the main interaction in the
stabilization of the X−⊂α,α-3 complexes is the formation of
hydrogen bonds between the halide and the pyrrole NHs of the
receptor. Throughout the receptor series α,α-3, we observed a
close similarity in the values of the chemical shift of the free
pyrrole NH protons. Likewise, the complexation induced shift
experienced by the NHs upon halide binding is very similar and
is only a function of the halide type. Furthermore, calculations
of EPN values at the pyrrole nitrogen nuclei of representative
dipyromethanes point to similar pKa values for the pyrrole NH
groups across the series of receptors α,α-3a to α,α-3h and
reference system 2.53 These observations suggest that, for a
given anion, it is sensible to consider that the strength of the
principal hydrogen bonding interaction remains constant
throughout the series X−⊂α,α-3 and X−⊂2. This claim is also
Figure 10. Plot of the experimental values for the anion-π interaction
■
▲
derived from the binding of chloride (green ), bromide (red ), or
●
iodide ( purple ) with two-wall calix[4]pyrroles α,α-3 correlate with
the ESP values calculated at the centroid of the aromatic walls. For
clarity, data for calixpyrrole α,α-3e is not shown. See the Supporting
Information for a complete plot.
3214
dx.doi.org/10.1021/ja412098v | J. Am. Chem. Soc. 2014, 136, 3208−3218

Journal of the American Chemical Society
Article
slopes of the correlation lines for each halide are very similar.56
While the existence of electrostatic interactions also serves to
explain why bromide owing to the increase in its atomic radius
binds slightly weaker than chloride, they do not account for the
noticeably more favorable interactions measured for iodide,
about 0.2 kcal mol−1 more favorable. It looks as though
describing the interaction as purely electrostatic and dependent
on the ESP value at the ring’s center is reasonable but perhaps
slightly too simplistic. In fact, a plain electrostatic anion−π
interaction should favor the chloride anion due to its higher
charge density and reduced size. The increased size and
polarizability of the iodide anion with respect to the other
halides is likely the reason for its stronger anion−π interactions.
We speculate that the induction of dipoles is maximized in the
case of iodide due to its larger volume and reduced
electronegativity. Most likely, the electrostatic attraction in
the iodide−π complexes is enhanced by the high polarizability
Table 2. Experimental Values for the Enthalpic (ΔH, kcal/
mol) and Entropic (TΔS, kcal/mol, 298 K) Components of
the Free Energy of Binding for the Complexes of Chloride
and Bromide with “Two-Wall” Calix[4]pyrroles α,α-3 in
a
Acetonitrile Solution
chloride
bromide
receptor
ESP
ΔH
TΔS
ΔH
TΔS
−0.9
2
n/a
−8.7
−6.9
−6.1
−6.2
−6.2
−6.2
−6.2
−6.4
−5.8
−1.6
−0.8
−0.6
0.3
−5.7
3a(OMe)
3b(H)
−16.1
−15.9
−9.7
−9.2
−8.9
−0.9
7.7
3c(Br)
3d(py)
0.8
−5.1
−0.5
3e(N3)
0.3
3g(NO2)
3h(F5)
−1.2
1.4
−5.0
−4.8
−4.2
0
1.3
2.1
3i(di-NO2)
11.8
2.7
a
The values presented are the mean of those obtained from at least
of the iodide, resulting in additional Debye−Huckel attractions
̈
two ITC titration experiments. Errors for ΔH and TΔS are assumed to
be in the region of 20%.
between the partial positive charge on the aromatic surface and
the induced dipole of the anion or vice versa through partial
charge transfer. van der Waals interactions are also maximized
for I− since the surface contact in the idoide−π complexes is
larger. Finally, the geometry of the I−−π contact observed in
the solid state for the complex I−⊂α,α-3g (Figure 8, c) closely
resembles that theoretically predicted as the energy minimum
for anion−π interactions involving a C6 aromatic ring.
The anion−π interactions between the electron-rich surfaces
of the phenoxy and phenyl groups (α,α-3a and α,α-3b,
respectively) and chloride, bromide and iodide (in the latter
case extrapolated from the corresponding regression line) are
all faintly repulsive (∼0.3−0.7 kcal mol−1). Then, as the ESP
values of the aromatic systems become increasingly positive, the
halide−π interactions become attractive. Eventually, the
interactions of all halides with the very π-acidic C6F5 and
(3,5-(NO2)2-Ph) systems (α,α-3g and α,α-3h respectively) are
noticeably attractive (∼ −0.7 to −1.0 kcal/mol).
As discussed above, the binding of chloride with all receptors
α,α-3a−h and the binding of bromide with the more electron
poor receptors α,α-3d and α,α-3f−h was investigated by ITC.
Thus, it was possible to dissect the free energies of binding into
their enthalpic (ΔH) and entropic (TΔS) components (Table
2 and Figure 15).
Figure 11. Plot of the experimental values for the enthalpic (lower,
circles, ΔH, kcal/mol) and entropic (upper, diamonds, TΔS at 298 K,
kcal/mol) components of the free energy of binding for the complexes
of chloride (green) or bromide (burgandy) with two-wall calix[4]-
pyrroles α,α-3 plotted against the ESP values at the center of the
respective meso-aromatic rings. For clarity, data for calixpyrrole α,α-3c
is not shown. See the Supporting Information for a complete plot.
As expected for binding processes taking place in a nonprotic
organic solvent and dominated by the formation of hydrogen
bonds, all events were mainly enthalpically driven and highly
exothermic. Also, as expected for complexation processes
relying on electrostatic forces (hydrogen bonds and anion−π
interactions), the binding of receptors α,α-3 with chloride
(−5.8 to −7.6 kcal/mol) was more exothermic than with
bromide (−4.2 to −5.1 kcal/mol). However, quite surprisingly
for us, the magnitude of the enthalpies of binding for a
common halide remained relatively constant across the series of
receptors α,α-3 or even displayed a slightly inverted trend with
respect to the free energies of binding (Figure 11, lower).
Indeed, the binding of chloride to the most electron-rich
calixpyrroles α,α-3a and α,α-3b is actually slightly more
enthalpically favorable than the binding of chloride with the
most electron poor examples α,α-3g and α,α-3h. A similar
trend was seen for the binding of bromide.57 Clearly, the almost
constant value of the enthalpic term (ΔH) for halide binding
across the series of calixpyrroles α,α-3 does not parallel the
trend seen for the free energies of binding (ΔG) and
consequently neither for the free energy values ascribed to
the halide-π interactions (ΔΔG).
At first sight, these results may be considered as a rare
example of a binding process showing a linear free energy
relationship that is not a consequence of an enthalpy−entropy
compensation effect.58,59 We are accustomed to relate an
increase in stability (ΔG) for a series of related 1:1 complexes
formed in nonprotic organic solvents to a gain in enthalpy
(stronger binding) resulting in a more organized complex (loss
of entropy, weaker binding). Such a scenario will correspond to
the most common enthalpy−entropy compensation effect. We
rationalized that the unexpected trends observed here for the
thermodynamic quantities ΔH and TΔS are blurred by
solvation effects that occur during the binding process. On
the one hand, both the magnitudes and the trend of the free
energy values determined for the series of complexes correlate
well with our simplistic picture of a 1:1 binding process in the
gas phase, where energetically more favorable anion-π
interactions should lead to thermodynamically more stable
3215
dx.doi.org/10.1021/ja412098v | J. Am. Chem. Soc. 2014, 136, 3208−3218

Journal of the American Chemical Society
Article
complexes. On the other hand, the dissection of Gibbs energies
into their enthalpic and entropic terms highlights the
shortcomings associated with the interpretation of these
magnitudes using the same simplistic physical model of a 1:1
complex formation that fails to consider solvation effects.
During the binding process, many of the constrained
acetonitrile molecules involved in solvating the receptor and
the anion will be released to the bulk solution. This will result
in two important effects. First, if the molecules released to the
bulk solvent form less strong interactions between themselves
than when solvating the halide and the receptor, this will be
accompanied by a loss of enthalpy. This means that the
measured enthalpy for the binding process is actually the
combination of the enthalpies of host−guest binding and
[Cl ⊂ α,α ‐ 3·MTOA]
[MTOACl][α,α ‐ 3]
Ka =
(4)
Thus, the differences in stability constants for the complexes
formed between the calix[4]pyrrole receptor series α,α-3b-−h
and MTOACl in chloroform can be directly ascribed to the
different binding affinities for chloride. This can be done
because only the anion-binding site is significantly modified. In
this way, the binding data obtained for the receptor series and
MTOACl in chloroform (see eq 4) can be related to that
obtained with TBACl in acetonitrile (see eq 1). In accordance
with this hypothesis, across the receptor series, the values
obtained for the free energies of the Cl−-π interaction (ΔΔG)
in chloroform with MTOACl matched very well with those
observed in acetonitrile with TBACl (Table S3 and Figure S6,
Supporting Information). Conversely, ΔH and TΔS for the
complexes between MTOACl and the receptor series α,α-3b-h
in chloroform painted a different picture to that seen in
acetonitrile with TBACl63 (Table S2 (Supporting Information),
and Figure 12, chloroform results in blue and acetonitrile
results in green).
−
solvent reorganization (ΔHobs = ΔHX ⊂3 + ΔHsolvation), both
terms being complex dependent. Second, the desolvation of
host and guest is typically characterized by a gain in entropy
owing to the release of several solvent molecules to the bulk
solution. Conversely, the formation of the 1:1 complex occurs
under loss of translational and rotational entropy of one of the
two components and may also restrict the conformational
motion of the host. The sum of these terms provides the overall
entropy value for the inclusion process in solution and explains
why a gain in entropy can be observed experimentally. We must
conclude that, in general, solvation effects render the physical
interpretation of the enthalpic and entropic components of free
energies of binding very complicated.
Due to the unexpected thermodynamic behavior seen in our
system, we performed binding experiments in other solvents.
Not surprisingly, analysis of chloride binding in acetone
revealed a very similar situation to that seen in acetonitrile
(see Table S2 and Figures S4 and S21 in the Supporting
Information). In particular, we looked to move the system into
a nonpolar solvent that would minimize solvation effects and
better approximate the gas phase host−guest complexation. In
order to do so, we turned from the tetrabutylammonium
(TBA) chloride salt, which due to ion pairing and lack of
cation-receptor size complementarity shows only weak
interactions in nonpolar solvents like chloroform,60 to the
methyl trioctylammonium chloride salt (MTOACl). Methyl-
trialkyl ammonium chloride salts are known to form receptor-
separated ion paired complexes with calix[4]pyrroles in
chlorinated solvents.61 Contrary to the above-described binding
experiments carried out in acetonitrile solution where
bimolecular anionic complexes X−⊂α,α-3 were formed, the
binding of MTOACl in chloroform yields trimolecular neutral
ion-paired complexes Cl−⊂3·MTOA. Despite this significant
difference, in chloroform solution, inclusion complexes between
MTOACl and “two-wall” calixpyrroles display the same binding
geometry regarding the anion and the aryl-extended
calixpyrrole core.62 In other words, the complex Cl−⊂3·
MTOA will still place the anion in the aromatic cleft,
sandwiched between the two phenyl meso-substituents and
interacting with their π systems. Furthermore, if one considers
that in chloroform, MTOACl is highly ion-paired, as well as the
resulting complex Cl−⊂3·MTOA, then, as in acetonitrile, the
binding event can be considered as the formation of a 1:1
complex, exhibiting an association constant reported in units of
M−1 (eqs 3 and 4).
Figure 12. Plot of the experimental values for the enthalpic (lower,
circles, ΔH, kcal/mol) and entropic (upper, diamonds, TΔS at 298 K,
kcal/mol) components of the free energy of binding for the complexes
of TBACl in acetonitrile (green) or MTOACl in chloroform (blue)
with two-wall calix[4]pyrroles α,α-3 plotted against the ESP values at
the center of the respective meso-aromatic rings.
In chloroform, the slope of the line defined by the entropic
quantities (TΔS) across the receptor series is still positive but is
much less pronounced than in acetonitrile. The entropy values
are slightly negative or very close to zero. Significantly, in the
less polar solvent the enthalpic component of binding (ΔH) is
still highly exothermic but becomes more negative (attractive)
going from Cl⊂α,α-3b·MTOA to Cl⊂α,α-3h·MTOA. These
results strongly suggest that, as expected, anion−π interactions
are mainly enthalpic in nature and that the unexpected
enthalpic and entropic quantities observed in acetonitrile are
due to solvation effects. Our findings also serve as a caveat for
the direct interpretation of the enthalpic and entropic
magnitudes of binding processes where major changes in
solvation of host and/or guest are experienced upon complex
formation.
CONCLUSION
In conclusion, we have characterized thermodynamically the
binding of chloride, bromide, and iodide with a variety of two-
■
α,α ‐ 3 + MTOACl ⇄ Cl ⊂ α,α ‐ 3·MTOA
(3)
3216
dx.doi.org/10.1021/ja412098v | J. Am. Chem. Soc. 2014, 136, 3208−3218

Journal of the American Chemical Society
Article
Notes
wall calix[4]pyrrole receptors α,α-3. The receptor series α,α-3
exhibits a variety of electronically different aromatic walls.
While bound to the calix[4]pyrroles, the anions are wedged in
the cleft formed by the two aromatic walls and sandwiched
between them. The anions are, therefore, forced to interact,
either repulsively or attractively, with the corresponding π
systems. Halide binding becomes progressively more attractive
as the number and electron withdrawing character of the
aromatic substituents increases (more positive ESP value at the
center of the six member ring). This may be considered as
evidence for the presence of anion-π interactions and the
electrostatic nature of those interactions. Comparison of
binding energies for the “two-wall “system α,α-3 with binding
energies for the “no-wall” reference calix[4]pyrrole 2 allows a
rough quantification of the strengths of the various halide-π
interactions. Chloride and bromide experience interactions with
π systems that are very similar in strength and reach attractive
magnitudes of up to −0.7 kcal mol−1. Iodide experiences
slightly more attractive interactions with π systems, up to −1.0
kcal mol−1, a phenomenon that we ascribe to the increased size
and polarizablility of iodide in relation to the other halides.
The binding processes were analyzed more thoroughly via
their enthalpic and entropic terms. Unexpectedly, in acetonitrile
solution, the increase in free energy of binding (ΔG) of halides
with receptors having aromatic rings with the more positive
ESP values, correlates with a more positive entropic
contribution (TΔS) to binding. In addition, the free energies
values and their enthalpic contributions (ΔH) show a subtle
inverted trend. We ascribed this unexpected behavior to strong
solvation effects taking place during the binding process.
Indeed, upon changing to a less polar solvent, chloroform, the
binding data displayed the expected relationship of free energy
and enthalpy of binding in nonpolar organic solvents, by which
an increase in free energy is reflected by an increase in binding
enthalpy. However, even in chloroform solution the model
system under study does not obey the common enthalpy−
entropy compensation paradigm. We suggest that the loss of
entropy leading to complex formation is compensated by a
significant gain in entropy accompanying the almost complete
desolvation of the ion-pair. This desolvation is required for the
formation of receptor-separated ion-paired Cl−⊂α,α-3·TOA
complexes. Finally, our results indicate that anion−π
interactions in polar and nonpolar organic solvents can be
weakly attractive −0.7 to −1 kcal mol−1 and that they are
mainly driven by enthalpy.64
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We thank Gobierno de Espana MINECO (projects CTQ2011-
̃
23014, CTQ2011-27512/BQU and CONSOLIDER INGEN-
IO CSD2010-00065, FEDER funds), Generalitat de Catalunya
DURSI (2009SGR00686), and ICIQ Foundation for funding.
D.Q. thanks the MICINN of Spain for a “Ramon
contract.
́
y Cajal”
REFERENCES
(1) Chowdhury, S.; Kebarle, P. J. Chem. Phys. 1986, 85, 4989.
(2) Hiraoka, K.; Mizuse, S.; Yamabe, S. J. Phys. Chem. 1987, 91, 5294.
(3) Quinonero, D.; Garau, C.; Rotger, C.; Frontera, A.; Ballester, P.;
Costa, A.; Deya, P. M. Angew. Chem., Int. Ed. 2002, 41, 3389.
(4) Quinonero, D.; Garau, C.; Frontera, A.; Ballester, P.; Costa, A.;
Deya, P. M. Chem. Phys. Lett. 2002, 359, 486.
■
(5) Alkorta, I.; Rozas, I.; Elguero, J. J. Am. Chem. Soc. 2002, 124,
8593.
(6) Mascal, M.; Armstrong, A.; Bartberger, M. D. J. Am. Chem. Soc.
2002, 124, 6274.
(7) Ma, J. C.; Dougherty, D. A. Chem. Rev. (Washington, DC, U.S.)
1997, 97, 1303.
(8) Giese, M.; Albrecht, M.; Bannwarth, C.; Raabe, G.; Valkonen, A.;
Rissanen, K. Chem. Commun. 2011, 47, 8542.
(9) Berryman, O. B.; Bryantsev, V. S.; Stay, D. P.; Johnson, D. W.;
Hay, B. P. J. Am. Chem. Soc. 2007, 129, 48.
(10) Ahuja, R.; Samuelson, A. G. CrystEngComm 2003, 5, 395.
(11) Mooibroek, T. J.; Black, C. A.; Gamez, P.; Reedijk, J. Cryst.
Growth Des. 2008, 8, 1082.
(12) Robertazzi, A.; Krull, F.; Knapp, E. W.; Gamez, P.
CrystEngComm 2011, 13, 3293.
(13) Chen, Y.; Wang, D. X.; Huang, Z. T.; Wang, M. X. Chem.
Commun. 2011, 47, 8112.
(14) Berryman, O. B.; Johnson, D. W. Chem. Commun. 2009, 3143.
(15) Dawson, R. E.; Hennig, A.; Weimann, D. P.; Emery, D.;
Ravikumar, V.; Montenegro, J.; Takeuchi, T.; Gabutti, S.; Mayor, M.;
Mareda, J.; Schalley, C. A.; Matile, S. Nature Chem. 2010, 2, 533.
(16) Chudzinski, M. G.; McClary, C. A.; Taylor, M. S. J. Am. Chem.
Soc. 2011, 133, 10559.
(17) Wang, D.-X.; Wang, M.-X. J. Am. Chem. Soc. 2013, 135, 892.
(18) Garau, C.; Frontera, A.; Quinonero, D.; Ballester, P.; Costa, A.;
Deya, P. M. Chem. Phys. Lett. 2004, 392, 85.
(19) Garau, C.; Frontera, A.; Quinonero, D.; Ballester, P.; Costa, A.;
Deya, P. M. J. Phys. Chem. A 2004, 108, 9423.
(20) Rosokha, Y. S.; Lindeman, S. V.; Rosokha, S. V.; Kochi, J. K.
Angew. Chem., Int. Ed. 2004, 43, 4650.
(21) Berryman, O. B.; Sather, A. C.; Hay, B. P.; Meisner, J. S.;
Johnson, D. W. J. Am. Chem. Soc. 2008, 130, 10895.
(22) Schneider, H.; Vogelhuber, K. M.; Schinle, F.; Weber, J. M. J.
Am. Chem. Soc. 2007, 129, 13022.
(23) Wang, D. X.; Zheng, Q. Y.; Wang, Q. Q.; Wang, M. X. Angew.
Chem., Int. Ed. 2008, 47, 7485.
(24) Chifotides, H. T.; Schottel, B. L.; Dunbar, K. R. Angew. Chem.,
Int. Ed. 2010, 49, 7202.
(25) Sokkalingam, P.; Kim, D. S.; Hwang, H.; Sessler, J. L.; Lee, C. H.
Chem. Sci. 2012, 3, 1819.
(26) Sokkalingam, P.; Kee, S. Y.; Kim, Y.; Kim, S. J.; Lee, P. H.; Lee,
C. H. Org. Lett. 2012, 14, 6234.
(27) Yoo, J.; Park, I. W.; Kim, T. Y.; Lee, C. H. Bull. Korean Chem.
ASSOCIATED CONTENT
■
S
* Supporting Information
General experimental methods for the synthesis and binding
studies, fits of the titration data, additional experimental results
discussed in the text and their associated figures, spectral data
for new compounds, and X-ray crystallographic files of 5c, 5d,
5h, α,α-3a, α,α-3b, α,α-3c, α,α-3g, α,α-3h, α,β-3c, α,β-3g, and
α,β-3h and the inclusion complexes TMACl⊂α,α-3a,
TMACl⊂α,α-3b, TMACl⊂α,α-3c, TMACl⊂α,α-3g,
TMACl⊂α,α-3h, TBABr⊂α,α-3f, TMABr⊂α,α-3g,
TMABr⊂α,α-3h, and TMAI⊂α,α-3g. This material is available
free of charge via the Internet at http://pubs.acs.org.
Soc. 2010, 31, 630.
(28) Frontera, A.; Gamez, P.; Mascal, M.; Mooibroek, T. J.; Reedijk,
J. Angew. Chem., Int. Ed. 2011, 50, 9564.
(29) Gorteau, V.; Bollot, G.; Mareda, J.; Perez-Velasco, A.; Matile, S.
J. Am. Chem. Soc. 2006, 128, 14788.
AUTHOR INFORMATION
■
Corresponding Author
pballester@iciq.es
(30) Mareda, J.; Matile, S. Chem.−Eur. J. 2009, 15, 28.
3217
dx.doi.org/10.1021/ja412098v | J. Am. Chem. Soc. 2014, 136, 3208−3218

Journal of the American Chemical Society
Article
(31) Vargas-Jentzsch, A.; Emery, D.; Mareda, J.; Metrangolo, P.;
Resnati, G.; Matile, S. Angew. Chem., Int. Ed. 2011, 50, 11675.
(32) Zhao, Y. J.; Domoto, Y.; Orentas, E.; Beuchat, C.; Emery, D.;
Mareda, J.; Sakai, N.; Matile, S. Angew. Chem., Int. Ed. 2013, 52, 9940−
99.
(33) Zhao, Y.; Beuchat, C.; Domoto, Y.; Gajewy, J.; Wilson, A.;
Mareda, J.; Sakai, N.; Matile, S. J. Am. Chem. Soc. 2014, DOI: 10.1021/
ja412290r.
(53) Calculations of the electrostatic potential values at the nitrogen
nuclei in the pyrrole units of phenyl substituted dipyromethanes hint
at a decrease of two pKa units for the pyrrolic NHs of α,α-3h with
respect to α,α-3b and octamethyl reference 2. Liu, S. B.; Pedersen, L.
G. J. Phys. Chem. A 2009, 113, 3648−3655. According to the pKa slide
rule, such a minimal change in the difference of donor-acceptor
acidities, ΔpKa, should have a reduced impact in the strength of
hydrogen bonding between specific halide and the calix[4]pyrrole core
for the receptor series α,α-3 and reference 2 studied in this paper.
Gilli; Pretto, L.; Bertolasi, V.; Gilli, G. Acc. Chem. Res. 2009, 42, 33−44
These findings in combination with the 1H NMR data discussed in the
text provide strong support to the claim that the primary hydrogen-
bonding interaction of the calix[4]pyrrole core is not significantly
perturbed by the introduction of aromatic walls or their modification
by different substituent groups..
́
(34) Gíl-Ramírez, G.; Escudero-Adan, E. C.; Benet-Buchholz, J.;
Ballester, P. Angew. Chem., Int. Ed. 2008, 47, 4114.
(35) If the four CaxialH···Cl− interactions present in the Cl−⊂2
complex were highly favorable in the binding energetics, the use of the
free enthalpy of binding of the complex as a reference to quantify Cl−-
phenyl interactions with the tetraaryl calix[4]prrole system would
induce an energetic reduction to our estimates. We assume that this is
not the case and that the CaxialH···Cl− interactions in the complex can
be considered to be isoenergetic with NCCH3···Cl− interactions
present in the solvated anion.
(54) Garau, C.; Quinonero, D.; Frontera, A.; Ballester, P.; Costa, A.;
Deya, P. M. J. Phys. Chem. A 2005, 109, 9341.
(55) The determined values of ΔΔG’s for the Cl−−π interaction
using the “two wall” system are in good agreement with those reported
previously in ref 34 using “four wall” analogues. However, it is worth
noting that the calculated magnitudes for repulsive Cl−−π interactions
(R = H, OMe, and Br) are slightly larger when derived from “four
wall” receptors. In trying to attenuate repulsive Cl−−π interactions,
“two wall” receptors can adopt a cone conformation of the bound state
in which the opposite phenyl groups bend away from the anion while
the axial meso-methyl groups get into a closer contact, possibly
establishing weak CH-Cl−−interactions. Interestingly, in the “four
wall” analogues the same conformational adaptation induces that the
other two meso-phenyl groups, instead of the methyl groups, are
brought in close proximity to the anion, thus avoiding the reduction of
repulsive Cl−-π interactions.
(36) The increased thermodynamic stability of complexes between
anions and the two-wall receptors α,α-3 recently allowed us to analyze
and quantify nitrate−π interactions in solution.26 These results showed
that, similarly to our aforementioned investigation of Cl−-π
interactions, nitrate−π interactions are mainly electrostatic in nature
and can be slightly attractive when very electron-withdrawing groups
are substituents of the aryl group.
(37) See the Supporting Information for detailed descriptions of the
syntheses.
(38) Bruno, G.; Cafeo, G.; Kohnke, F. H.; Nicolo, F. Tetrahedron
2007, 63, 10003.
(39) Park, J. S.; Yoon, K. Y.; Kim, D. S.; Lynch, V. M.; Bielawski, C.
W.; Johnston, K. P.; Sessler, J. L. Proc. Natl. Acad. Sci. U.S.A. 2011, 108,
20913.
(40) Calix[4]pyrrole α,α-3i was not used in binding experiments
because of competitive hydrogen bonding between the chloride anion
and the para hydroxy groups on the meso phenyl rings.
(41) de Namor, A. F. D.; Shehab, M. J. Phys. Chem. B 2003, 107,
6462.
(56) The slope of the correlation line for the interaction of the nitrate
anion with the same receptor series is also very similar. Adriaenssens,
L.; Estarellas, C.; Jentzsch, A. V.; Belmonte, M. M.; Matile, S.;
Ballester, P. J. Am. Chem. Soc. 2013, 135, 8324.
(57) Variable-temperature 1H NMR titrations were run with TBAI in
acetonitrile. ΔH and TΔS were determined by the construction of
Van’t Hoff plots. Indeed, plotting these values in front of the ESP
values of the corresponding aromatic rings revealed the same
unexpected behavior as seen for the binding of chloride and bromide.
See the Supporting Information for details.
(42) X-ray structures suggest that when in the 1,3 alternate
conformation H3 is shielded by the meso-aromatic rings. Hence, in
1
the H NMR spectrum, the signal belonging to H3 appears upfield
with respect to the signal corresponding to H4. Upon binding, the
conformational change to cone conformation moves H3 away from the
(58) Anslyn, E. V.; Dougherty, D. A. Modern Physical Organic
Chemistry; University Science Books: Sausalito, CA, 2006; Chapters 4
and 8.
1
anisotropic current of the meso-aromatic ring and so the H NMR
signal corresponding to H3 is shifted downfield past the signal for H4.
(43) Estarellas, C.; Bauza, A.; Frontera, A.; Quinonero, D.; Deya, P.
M. Phys. Chem. Chem. Phys. 2011, 13, 5696.
(59) Dunitz, J. D. Chem. Biol. 1995, 2, 709.
(60) Sessler, J. L.; Gross, D. E.; Cho, W. S.; Lynch, V. M.;
Schmidtchen, F. P.; Bates, G. W.; Light, M. E.; Gale, P. A. J. Am. Chem.
Soc. 2006, 128, 12281.
(61) Gross, D. E.; Schmidtchen, F. P.; Antonius, W.; Gale, P. A.;
Lynch, V. M.; Sessler, J. L. Chem.−Eur. J. 2008, 14, 7822.
(62) See the Supporting Information
(63) The binding between MTOACl and receptors α,α-3 was also
investigated in acetonitrile. As expected, the results were almost
identical to those obtained with TBACl. This confirms that in
acetonitrile the ion pair is completely dissociated and that the unusual
thermodynamics of binding seen in acetonitrile are solely due to subtle
differences in the energy of solvation. See the Supporting Information
for details.
(64) Arranz-Mascaros, P.; Bazzicalupi, C.; Bianchi, A.; Giorgi, C.;
Godino-Salido, M.-L.; Gutierrez-Valero, M.-D.; Lopez-Garzon, R.;
Savastano, M. J. Am. Chem. Soc. 2013, 135, 102.
(44) Both the quoted distances represent the shorter of the two dplane
distances measured in the inclusion complexes constituting the
asymmetric unit of the crystal.
(45) In the cases where 100% complexation of the receptor was not
achieved the chemical shift values of the bound pyrrole NHs could be
obtained from the nonlinear fit of the titration data using the
HypNMR 2006 software.
(46) The reduced solubility of these receptors in acetonitrile
hampered our attempts to detect binding at higher concentrations.
(47) Gonzalez-Alvarez, A.; Frontera, A.; Ballester, P. J. Phys. Chem. B
2009, 113, 11479 and references cited therein.
(48) Wiseman, T.; Williston, S.; Brandts, J. F.; Lin, L. N. Anal.
Biochem. 1989, 179, 131.
(49) No changes were detected in the chemical shift values of the
TBA cation during the NMR titrations. In addition, we have not
observed any concentration dependence or common ion (cation)
effect during the determination of the Ka values using ITC.
(50) de Namor, A. F. D.; Khalife, R. Phys. Chem. Chem. Phys. 2010,
12, 753.
(51) Cockroft, S. L.; Perkins, J.; Zonta, C.; Adams, H.; Spey, S. E.;
Low, C. M. R.; Vinter, J. G.; Lawson, K. R.; Urch, C. J.; Hunter, C. A.
Org. Biomol. Chem. 2007, 5, 1062.
(52) Wheeler, S. E.; Houk, K. N. J. Phys. Chem. A 2010, 114, 8658.
3218
dx.doi.org/10.1021/ja412098v | J. Am. Chem. Soc. 2014, 136, 3208−3218