Investigating the imidazolium-anion interaction through the anion-templated construction of interpenetrated and interlocked assemblies

Article abstract of DOI:10.1002/chem.201102005

The interaction between imidazolium cations and coordinating anions is investigated through the anion-templated assembly of interpenetrated and interlocked structures. The orientation of the imidazolium motif with respect to anion binding, and hence the h

Full text of DOI:10.1002/chem.201102005

FULL PAPER
DOI: 10.1002/chem.201102005
Investigating the Imidazolium–Anion Interaction through the Anion-
Templated Construction of Interpenetrated and Interlocked Assemblies
Graeme T. Spence,[a] Christopher J. Serpell,[a] Jo¼o Sardinha,[b] Paulo J. Costa,[b]
Vꢀtor Fꢁlix,[b] and Paul D. Beer*[a]
1
Abstract: The interaction between imi-
dazolium cations and coordinating
anions is investigated through the
anion-templated assembly of interpene-
trated and interlocked structures. The
orientation of the imidazolium motif
with respect to anion binding, and
hence the hydrogen bond donor ar-
rangement, was varied in acyclic recep-
tors, interpenetrated assemblies, and
the first mono-imidazolium interlocked
systems. Their anion recognition prop-
erties and co-conformations were stud-
ied by solution-phase H NMR investi-
gations, solid-state structures, molecu-
lar dynamics simulations, and density
functional theory calculations. Our
findings suggest that the imidazolium-
anion binding interaction is dominated
by electrostatics with hydrogen-bond-
ing contributions having weak orienta-
tional dependence.
Keywords: anions · molecular rec-
ognition · rotaxanes · supramolec-
ular chemistry · template synthesis
Introduction
With the objective of constructing selective and potent
anion receptors, we have established an anion templation
strategy for the synthesis of interlocked host structures,
which possess topologically unique, well defined three-di-
mensional binding cavities.[13,14] By employing different
anion-binding motifs within the components of rotaxanes, a
variety of selective interlocked host systems for chloride,
bromide, and iodide anions have been prepared containing
pyridinium,[15] triazolium,[16] and iodotriazolium[17] axle
motifs, respectively. However, while the imidazolium group
has been exploited in the assembly of a number of pseudor-
otaxanes[18–20] and recent bis- and tetra-imidazolium rotax-
ane systems,[21] to the best of our knowledge no mono-imida-
zolium interlocked structures have been synthesized, either
for anion recognition or in others area of supramolecular
chemistry.
We report herein solution-phase and solid-state investiga-
tions, including computational MD simulations and density
functional theory (DFT) calculations, into the nature of the
interaction between imidazolium cations and coordinating
anions in the anion-templated assembly of interpenetrated
and interlocked structures. Importantly, the impact this has
on anion recognition was investigated by manipulating the
orientation of the imidazolium motif with respect to poten-
tial hydrogen bonds for anion-binding in acyclic, interpene-
trated and interlocked systems. This was achieved by methyl
group substitution of the imidazolium group dictating that
anion binding would occur through favorable hydrogen
bonding to C2-H in 4,5-dimethylimidazolium derivatives
(Scheme 1a), and to C4/5-H in 2-methylimidazolium deriva-
tives (Scheme 1b).
The imidazolium motif is widely employed in many diverse
areas of chemistry, such as ionic liquids (ILs),[1] liquid crys-
tals,[2] highly effective stabilizers for transition-metal nano-
particles,[3] and in organometallic chemistry as precursors for
N-heterocyclic carbene (NHC) ligands.[4] Also, since the first
imidazolium anion receptors were reported by Sato et al.[5]
and Alcalde et al.[6] in 1999, this group has been used exten-
sively in the areas of anion recognition and sensing.[7]
In anion coordination chemistry, the idea of charge-assist-
ed hydrogen bonding directs the design of anion receptors,[8]
which often incorporate several imidazolium groups for con-
vergent hydrogen-bond donation.[9] However, to the best of
our knowledge, no investigation into the nature of the imi-
dazolium-anion binding interaction has been undertaken in
the field of anion recognition. While this interaction in ILs
has been extensively studied and is still a matter of lively
debate,[10–12] the corresponding interaction in anion coordi-
nation chemistry is fundamentally different due to the fact
that anion-binding studies are carried out in dilute solution,
and due to the geometrically constrained arrangement of hy-
drogen-bond donors that exist in anion host systems.
[a] G. T. Spence, Dr. C. J. Serpell, Prof. P. D. Beer
Chemistry Research Laboratory, Department of Chemistry
University of Oxford, Mansfield Road, Oxford, OX1 3TA (UK)
Fax :ACHTUNGTRENNUNG(+44)1865-272690
E-mail: paul.beer@chem.ox.ac.uk
[b] Dr. J. Sardinha, Dr. P. J. Costa, Prof. V. Fꢀlix
Departamento de Quꢁmica, CICECO and
Secżo Autꢂnoma de CiÞncias da Safflde
Universidade de Aveiro, 3810-193 Aveiro (Portugal)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201102005.
Chem. Eur. J. 2011, 17, 12955 – 12966
ꢃ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
12955

midpoint between the two nitrogens, as noted by Tsuzuki
and co-workers[11]), than in 2·PF6.
Crystals of 1·Cl and 2·Cl suitable for single-crystal X-ray
diffraction structural analysis were obtained by leaving the
viscous oils to stand for 2–3 months (Figure 1).[24] As well as
Scheme 1. Controlling the orientation of the imidazolium group with re-
spect to anion binding by methyl group substitution.
Results and Discussion
Anion binding of simple acyclic imidazolium derivatives:
Simple acyclic imidazolium derivatives 1·PF6 and 2·PF6, con-
taining the 4,5-dimethyl- and 2-methylimidazolium motifs,
respectively, were synthesized as non-coordinating hexa-
fluorophosphate salts (Scheme 2). The anion-binding prop-
Scheme 2. 4,5-Dimethyl- and 2-methylimidazolium derivatives 1·X and
2·X, respectively.
1
erties of these systems were studied initially by H NMR ti-
tration experiments in CDCl3. Upon addition of tetrabutyl-
ACHTUNGTRENNUNGammonium (TBA) halide and dihydrogen phosphate salts,
significant downfield shifts were observed for C2-H in 1·PF6
and C4/5-H in 2·PF6 (Dd= +2.11 ppm and +0.33 ppm, re-
spectively, after 10.0 equivalents of TBACl were added), in-
dicating that binding was occurring in the expected modes.
WinEQNMR2[22] analysis of the titration data gave the 1:1
stoichiometric association constants shown in Table 1, with
Figure 1. Crystal structures of a) 1·Cl and b) 2·Cl. The center of positive
charge is marked by a dark blue sphere, with corresponding distances to
anions shown in dotted blue lines. H···Cl close contacts are displayed
using dashed black lines.
Table 1. Association constants, K, of 1·PF6 and 2·PF6 with anions in
CDCl3 at 293 K. Obtained by monitoring the imidazolium proton(s).
Errors are given in parentheses.
observing hydrogen bonds between the imidazolium ring
protons and the anion (and co-crystallized atmospheric
water), there are a number of close contacts between the
alkyl protons a- to the imidazolium ring and the halide
anions. In 1·Cl the closest chloride to the center of positive
charge is positioned at 4.188(2) Š through a C2-H hydrogen
Anion
1·PF6
2·PF6
K [mꢀ1
]
K [mꢀ1
]
Clꢀ
3044 (108)
2736 (186)
1609 (115)
4301 (572)
782 (42)
523 (60)
402 (44)
1000 (83)
Brꢀ
Iꢀ
ꢀ
H2PO4
ꢀ
bond with a C H···Cl angle of 149.08 and C···Cl distance of
the strength of anion binding for both systems following the
trend in anion basicity. Acyclic receptor 1·PF6 exhibits much
stronger anion-binding affinity than 2·PF6. This can be ra-
tionalized by the differing acidities of the respective imida-
zolium protons involved in hydrogen bonding.[10,23] Also, the
electrostatic contribution to binding is greater in 1·PF6, due
to the smaller distance between the bound anion and the
center of imidazolium positive charge (considered to be
3.461(1) Š (Figure 1a). The two next-nearest chloride
anions are at a distance of 5.119(2) and 5.637(1) Š from the
center of positive charge. As expected, in the 2·Cl structure
the chloride closest to the center of positive charge is fur-
ther away, at a distance of 4.847(6) Š, and is coordinated by
an imidazolium ring C4/5-H hydrogen bond with a C···Cl dis-
ꢀ
tance of 3.528(4) Š and a C H···Cl angle of 150.28 (Fig-
ure 1b). In this case, the next two closest anions are at
12956
www.chemeurj.org
ꢃ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 12955 – 12966

Imidazolium Interpenetrated and Interlocked Assemblies
FULL PAPER
5.090(2) and 5.195(11) Š. Although the observable parame-
ters pertaining to imidazolium ring hydrogen bonding
strength (distances and angles) imply that there is little dif-
ference in the anion–proton interactions between the two
structures, in 2·Cl the halide anion is at a relatively longer
distance from the center of positive charge. This suggests
that the greater anion binding strength of 1·PF6 observed in
CDCl3 solution is a consequence of the actual geometry of
the complex permitting a closer approach of opposite charg-
es rather than a function of classical hydrogen bonding.
Pseudorotaxane assembly investigations: The ability of a va-
riety of positively charged pyridinium, imidazolium, and
guanidinium threading components to form anion-templated
pseudorotaxane assemblies with isophthalamide macrocycles
has been investigated previously.[18] For an unsubstituted
imidazolium thread, the strongest associations were ob-
served when chloride was the templating anion. Also, as
control systems for an investigation into halogen bonding in-
terpenetrative assembly, pseudorotaxanes between 1·Cl and
2·Cl and macrocycle 3, stabilized by halide-anion binding
and p–p stacking interactions, have been recently report-
ed.[19] In contrast to the anion binding trend observed in
Table 1, 2·Cl was found to form a significantly more stable
interpenetrative assembly with 3 in CDCl3 than 1·Cl (245mꢀ1
and 97mꢀ1, respectively).[19] It was postulated that this result
was due to a beneficial cooperative chelating hydrogen-
bonding mode with the halide anion, along with potential
ꢀ
C H···O hydrogen bonding from the imidazolium 2-methyl
group of 2·Cl to the polyether chain of 3.
To further study these interpenetrated assemblies in the
context of this investigation, additional evidence for the so-
lution-phase co-conformations of these pseudorotaxanes was
provided by 2D 1H NMR ROESY spectroscopy. Diagnostic
regions of the spectra are shown in Figure 2 and Figure 3,
along with the assigned through-space intercomponent cor-
relations. These indicate that both 1·Cl and 2·Cl threads are
in the expected orientation within the macrocycle, with the
imidazolium protons pointing towards the anion binding
cleft.
Crystals suitable for single-crystal X-ray diffraction struc-
tural determination were prepared for 1·3·Cl and 2·3·Cl by
slow evaporation of dilute CHCl3 solutions of the pseudoro-
taxane assemblies, and analyzed by using synchrotron radia-
tion at Diamond Light Source beamline I19. In contrast to
the solution-phase evidence, the solid-state structure of the
4,5-dimethylimidazolium system 1·3·Cl revealed the opposite
Figure 2. Selected regions of 1H ROESY NMR spectrum of 1·3·Cl in
CDCl3 at 293 K. Orientational intercomponent correlations are marked
on the structure.
of the expected co-conformation, with the imidazolium C2
proton pointing towards the polyether chain of the macro-
Chem. Eur. J. 2011, 17, 12955 – 12966
ꢃ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
12957

P. D. Beer et al.
Figure 4. Crystal structure of interpenetrated assembly 1·3·Cl. The center
of positive charge is marked by a dark blue sphere, with corresponding
distances to anions shown in dotted blue lines. H···Cl and H···O close
contacts are displayed using dashed black lines. Protons omitted for clari-
ty where appropriate. The chloride anion from the adjacent assembly is
depicted as a transparent sphere.
AHCTUNGTREGcNNNU ycle, and the methyl groups directed towards the chloride
anion (Figure 4). This results in a large distance between the
centre of the positive charge and the anion, which is appa-
rently templating the pseudorotaxane, at 5.885(2) Š. How-
ever, examination of the entire structure reveals that the
chloride anion occupying the cavity of the adjacent pseudo-
AHCTUNGTREGrNNUN otaxane is in fact slightly closer at 5.272(3) Š. This provides
increased electrostatic stabilisation and, along with close
contacts between imidazolium methyl protons and the two
nearest anions (Figure 4), aids the construction of a crystal-
line network.
Conversely, disorder was observed in the 2·3·Cl structure,
owing to the thread existing in two contrasting co-conforma-
tions of approximately equal occupancy—with either the
4,5-imidazolium protons (A) or the 2-methyl group (B)
pointing towards the anion (occupancies 0.514(8) and
0.486(8) respectively) (Figure 5). In co-conformation A, the
chloride anion in the pseudorotaxane cavity is the closest
anion to the positive charge at 4.994(1) Š, the next nearest
being at 5.853(3) Š. However, when the methyl group
points towards the anion in the cavity (B), the closest anion
is located in an adjacent pseudorotaxane at 5.182(3) Š, com-
pared to 5.367(1) Š for the chloride in the same assembly.
Additional inter-assembly hydrogen bonds are also observed
in both co-conformations, although slightly longer than close
contacts, further strengthening the network of intermolecu-
lar forces required for crystal formation.
Interestingly, the equivalent bromide structure 2·3·Br is
almost exactly isostructural with 2·3·Cl, albeit with a greater
degree of disorder, consistent with the chaotropic nature of
the more diffuse halide anion (see Figure S30 in the Sup-
porting Information).
These findings suggest that, while pseudorotaxane forma-
tion in solution appears to operate through a charge-assisted
hydrogen-bonding mechanism within each assembly, their
solid-state orientations are much less dependent on this in-
teraction due to the proximity of adjacent units. This can be
rationalized by considering both the electrostatic and hydro-
Figure 3. Selected regions of 1H ROESY NMR spectrum of 2·3·Cl in
CDCl3 at 293 K. Orientational intercomponent correlations are marked
on the structure.
12958
www.chemeurj.org
ꢃ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 12955 – 12966

Imidazolium Interpenetrated and Interlocked Assemblies
FULL PAPER
mations and to provide further insights into these assem-
blies, MD simulations and DFT calculations were carried
out on 1·3·Cl and 2·3·Cl pseudorotaxanes.
Pseudorotaxanes 1·3·Cl and 2·3·Cl were investigated by
means of MD simulations using the AMBER 11 package[25]
with the GAFF[26] parameters. For each chloride interpene-
trative assembly, two inverse co-conformations were select-
ed: for 1·3·Cl, the conformation shown in Figure 4 (denoted
A) and the opposite with the C2-H proton pointing towards
the chloride anion (denoted B); for, 2·3·Cl, co-conforma-
tions A and B as depicted in Figure 5. Since only 2·3·Cl crys-
tallized in both co-conformations, the starting structures for
the modeling studies were generated in the gas phase
through MD conformational search methods. All computa-
tional details, including the procedure to generate the start-
ing structures, are given in Supporting Information.
The selected structures (1·3·Cl-A, 1·3·Cl-B, 2·3·Cl-A, and
2·3·Cl-B) were immersed in periodic cubic boxes of CHCl3
molecules and their dynamical behavior was investigated by
MD simulations for 20 ns using the previously equilibrated
structures. Furthermore, for 1·3·Cl-A and 2·3·Cl-B assembly
arrangements, three simulation runs were performed to en-
hance the sampling of the conformational space in an at-
tempt to observe the rotation of the imidazolium threads to
the structures observed by NMR spectroscopy (1·3·Cl-B and
2·3·Cl-A). In all simulations the chloride anion was kept
firmly bonded to the isophthalamide macrocyclic cleft
ꢀ
through two simultaneous N H···Cl hydrogen bonds, as in
our previous study on halogen-bonding interpenetrative as-
semblies.[19] The stabilities of the imidazolium thread orien-
tations were assessed by monitoring the interaction distan-
ces as follows: C2-H···Cl, C4-CH3···Cl, and C5-CH3···Cl for
1·3·Cl structures; and C2-CH3···Cl, C4-H···Cl, and C5-H···Cl
for 2·3·Cl structures. For each starting co-conformation of
both assemblies, the relevant interaction distances (H···Cl,
or C···Cl when methyl groups are involved) were followed
throughout the simulation time (Figure 6).[27]
For 1·3·Cl-A, rotation of the imidazolium thread was ob-
served in one simulation (Figure 6, top, left), resulting in
conversion from this starting co-conformation to 1·3·Cl-B
with representative snapshots shown in Figure 7. Initially,
both imidazolium methyl groups (C4/5-CH3) are pointing to-
wards the chloride anion, while the C2-H establishes hydro-
gen bond interactions with the oxygens of the polyether
loop (not shown). At approximately 8.5 ns of the simulation
run, the imidazolium motif starts to rotate affording an in-
termediate conformation, labeled 1·3·Cl-C, where the imida-
zolium thread is almost perpendicular to both phenyl rings
of the macrocycle with one methyl group and the C2-H
pointing at them. From about 8.7 ns onwards, similar dy-
namical behavior to 1·3·Cl-B is observed (Figure 6, top,
right), with the C2-H interacting with the chloride, indicating
a definitive conversion of the original 1·3·Cl-A into the
1·3·Cl-B interpenetrative assembly co-conformation.
Figure 5. Crystal structures of interpenetrated assembly 2·3·Cl, showing
both geometries of the disordered thread (A and B). The centre of posi-
tive charge is marked by a dark blue sphere, with corresponding distan-
ces to anions shown in dotted blue lines. H···Cl and H···O close contacts
are displayed using dashed black lines. Protons omitted for clarity where
appropriate. The anion from the adjacent assembly is depicted as a trans-
parent sphere.
gen-bonding interactions observed. However, due to the
level of complexity and multitude of interactions present
within the crystalline networks, little can be inferred about
the direct balance between electrostatics and imidazolium
hydrogen bonding in the anion-templated assembly of these
interpenetrative systems.
According to the MD simulation, the co-conformation
1·3·Cl-C is extremely short-lived (1.2 ps); no plateau is ob-
served in the transition of the C2-H···Cl or C4/5-CH3···Cl dis-
Modeling studies: To attempt to account for the disparity in
the solution-phase and solid-state pseudorotaxane co-confor-
Chem. Eur. J. 2011, 17, 12955 – 12966
ꢃ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
12959

P. D. Beer et al.
Figure 6. Evolution of the distances for the interactions: C2-H···Cl, C4-CH3···Cl, and C5-CH3···Cl in 1·3·Cl-A and 1·3·Cl-B and C2-CH3···Cl, C4-H···Cl, and
C5-H···Cl in 2·3·Cl-A and 2·3·Cl-B during MD simulations in CHCl3. For 1·3·Cl-A and 2·3·Cl-B interpenetrated assemblies, only one of the three simula-
tion runs is shown.
Figure 8. Surface representation (blue) of the histogram built with the po-
Figure 7. Representative snapshots of 1·3·Cl-A MD run taken at different
simulation times and illustrating the rotational conversion 1·3·Cl-A
sitions occupied by the C2 hydrogen atom during the 1·3·Cl-A ! 1·3·Cl-
C ! 1·3·Cl-B transition observed between 8 and 9 ns of the MD run.
ꢀ
(left)!1·3·Cl-C (center)!1·3·Cl-B (right). The N H···Cl hydrogen bonds
The rotation axis is perpendicular to the macrocyclic plane.
ꢀ
are drawn as yellow dashed lines, while the C H···Cl hydrogen bonds are
drawn as blue (C2-H···Cl), red (C4-CH3···Cl), and green (C5-CH3···Cl), fol-
lowing the color scheme used in Figure 6.
each run, which is in agreement with the intrinsic nature of
the unconstrained MD simulations and the complexity of
the rotation of the pseudorotaxane assembly. In other
words, due to the extreme co-conformational change ob-
served in the first simulation it is understandable that this
conversion was not observed in every simulation. Further-
more, when the intermediate 1·3·Cl-C was subjected to MD
simulation experiments, it always converted to 1·3·Cl-B. De-
spite the starting co-conformation 1·3·Cl-A not being ob-
served in solution by NMR spectroscopy, it remains stable
throughout the 20 ns of the remaining two MD simulations
(see Figure S32 in the Supporting Information) indicating
that it can be considered as a stable minimum.
tances. Indeed the grid of points built with the positions oc-
cupied by the C2 hydrogen between 8 and 9 ns of this MD
simulation, represented in Figure 8, shows that this atom is
primarily pointing towards the oxygen atoms of the poly-
ether linkage or the chloride (large individual “clouds”),
whereas a very small group of points is seen facing the
phenyl group of the macrocycle (this region of points is
highlighted with a red circle in Figure 8).
However, this thread rotation event was not observed in
the other two MD simulation runs for 1·3·Cl-A. This fact in-
dicates a different sampling of the conformational space in
12960
www.chemeurj.org
ꢃ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 12955 – 12966

Imidazolium Interpenetrated and Interlocked Assemblies
FULL PAPER
In contrast, when the simulation starts from the 1·3·Cl-B
assembly, no rotational conversion was observed (Figure 6,
top right). These findings suggest that co-conformation
1·3·Cl-B is preferred in solution, in agreement with the
NMR data. Similarly, starting structures 2·3·Cl-A and 2·3·Cl-
B were stable over the entire time of the simulation (see
Figure 6, bottom left and right), including in the additional
runs for 2·3·Cl-B (see Figure S32). The co-conformations for
the 2·3·Cl-A and 2·3·Cl-B simulations, obtained from cluster-
ing the MD trajectories, are comparable to those observed
in the X-ray single crystal structures (Figure 5) and are rep-
resented in Figure S33.
Synthesis of rotaxanes: Taking into account the anion-bind-
ing properties and chloride-anion-templated pseudorotaxane
assembly results of imidazolium derivatives 1·PF6 and 2·PF6,
it was of interest to prepare novel rotaxane host systems
containing 4,5-dimethyl- and 2-methylimidazolium axle com-
ponents. Investigating the anion-recognition properties of
such systems, with contrasting interlocked cavity hydrogen-
bond donor arrangements, would provide further insights
into the imidazolium–anion interaction.
4,5-Dimethyl- and 2-methylimidazolium axle components
4·Cl and 5·Cl (Scheme 3) were synthesized as detailed in the
Supporting Information (see Scheme S2 and S3). For rotax-
Hence, thread rotation from
2·3·Cl-B to the co-conformation
observed in solution (2·3·Cl-A)
was not observed within the
MD simulation time. This may
be due to insufficient sampling
of the conformational space or
because the starting structure is
trapped in a potential energy
surface well. Therefore we
cannot infer which co-confor-
mation is more stable in solu-
tion using this approach.
Subsequently, in order to
complement this study, the
structural and energetic prefer-
ences of 1·3·Cl and 2·3·Cl co-
conformations were investigat-
ed by DFT calculations, using
as starting geometries the rep-
resentative MD and X-ray crys-
tal structures. These studies
were accompanied with the cal-
culation of the electron densi-
Scheme 3. Axle components 4·Cl and 5·Cl and macrocycle precursor 6 used in rotaxane formation.
ties of the individual imidazolium threads, and are detailed
in the Supporting Information. For 1·3·Cl, co-conformation
B was calculated to be 7.86 kcalmolꢀ1 lower in energy than
A using the most representative structures obtained from
the MD simulations. Whereas, the differences were much
smaller for 2·3·Cl co-conformations, where A was found to
be lower in energy than B by only 0.23 kcalmolꢀ1 using the
MD structures, and 1.55 kcalmolꢀ1 using the X-ray crystal
structures. These results are consistent with the solution-
phase NMR structural findings and MD simulations behav-
ior.
In summary, by using a combination of MD and DFT sim-
ulations, the apparent disparity between the experimentally
determined solution-phase and solid-state co-conformations
of the pseudorotaxanes can be rationalized. The theoretical
studies confirm and quantify the higher stabilities of the ob-
served solution structures which are dictated by individual
and specific anion templation interactions. Conversely, the
existence of inter-assembly interactions (which aid the for-
mation of a crystalline network) results in the formation of
the opposite co-conformers in the solid state.
ane formation, a ꢄclippingꢅ method was employed using
Grubbsꢅ-catalyzed ring-closing metathesis (RCM) of bis-
vinyl-appended macrocycle precursor 6.[14] Whilst the iso-
phthalamide unit in macrocycle 3 was functionalized with a
tert-butyl group to aid solubility for the pseudorotaxane as-
sembly 2D H NMR ROESY experiments, macrocycle pre-
cursor 6 contains an electron-withdrawing nitro isophthal-
AHCTUNGTREGaNNNU mide substituent to increase the acidity of the amide pro-
tons for enhanced rotaxane–anion binding strength.
Rotaxane 7·Cl was synthesized by reaction of an equimo-
lar solution of 4·Cl and 6 in dry CH2Cl2 with Grubbsꢅ
second-generation RCM catalyst (Scheme 4). The product
was isolated in 30% yield following purification by prepara-
tive silica thin layer chromatography. The 1H NMR spec-
trum of 7·Cl, along with those of axle 4·Cl and macrocycle
precursor 6, is shown in Figure 9.
Upon rotaxane formation, an upfield shift of imidazolium
proton 1 and downfield shifts of isophthalamide protons b
and c were observed. These correspond to polarization of
the templating chloride anion away from the imidazolium
group of the axle and towards the hydrogen-bond donors of
1
Chem. Eur. J. 2011, 17, 12955 – 12966
ꢃ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
12961

P. D. Beer et al.
Scheme 4. Synthesis of 4,5-dimethylimidazolium rotaxanes 7·Cl and 7·PF6.
The analogous 2-methylimidazolium rotaxanes 8·Cl and
8·PF6 were prepared by using the same method (Scheme 5).
In this case, a reduced yield of 19% was obtained for the ro-
taxanation step. Interestingly, this trend in rotaxane yields
between the two substituted imidazolium motifs is contrary
to the previously reported stability trend observed for pseu-
dorotaxane formation, where the 2-methylimidazolium
thread 2·Cl formed a significantly more stable interpenetra-
tive assembly with macrocycle 3 than the 4,5-dimethyl ana-
1
logue 1·Cl.[19] Similar shifts in the H NMR spectra of 8·Cl
and 8·PF6 were observed as discussed above for the forma-
tion of 7·Cl and 7·PF6 (Figure 10). However, a greater hy-
droquinone splitting was observed for 8·Cl compared with
7·Cl, indicating different extents of p–p stacking in the two
systems.
Further characterization of the rotaxane structures was
obtained by 2D H NMR ROESY spectroscopy. A number
Figure 9. 1H NMR spectra in CDCl3 at 293 K of a) axle 4·Cl, b) rotaxane
7·Cl, and c) macrocycle precursor 6.
1
the macrocycle. The signals associated with the hydroqui-
none environments f and g have split slightly and moved sig-
nificantly upfield in the rotaxane, owing to favorable donor–
acceptor p–p stacking interactions between the imidazolium
axle and the macrocycle. It is notable that an upfield shift of
the axle methyl protons 2 was also observed, presumably
due to the hydroquinone ring currents.
The chloride template was exchanged for the non-coordi-
nating hexafluorophosphate anion to afford 7·PF6 in quanti-
tative yield. Upfield shifts in the 1H NMR signals corre-
sponding to the interlocked cavity protons involved in anion
binding (b, c, and 1) were observed upon removal of the co-
ordinating chloride anion. A greater splitting in the hydro-
quinone proton signals (f and g) was also observed, suggest-
ing increased p–p stacking interactions in 7·PF6 compared
with 7·Cl.
Figure 10. 1H NMR spectra in CDCl3 at 293 K of a) axle 5·Cl, b) rotaxane
8·Cl, and c) macrocycle precursor 6.
12962
www.chemeurj.org
ꢃ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 12955 – 12966

Imidazolium Interpenetrated and Interlocked Assemblies
FULL PAPER
Scheme 5. Synthesis of 2-methyl imidazolium rotaxanes 8·Cl and 8·PF6.
of through-space axle–macrocycle intercomponent correla-
tions were observed (Figure 11) confirming the interlocked
nature of the systems. The 2D ROESY spectra are shown in
the Supporting Information (Figure S4, S8, S12, and S16).
Although difficulties in identifying some correlations were
experienced due to the broadening of several proton peaks
and crowded aromatic regions, the spectra strongly suggest
that the orientations of the axles within the macrocycles is
as expected—with the imidazolium ring protons pointing to-
wards the macrocycleꢅs isophthalamide anion binding cleft.
Importantly this is the case for the 4,5-dimethyl- and 2-
methylimidazolium rotaxanes, both as the chloride salts and
after exchange to the non-coordinating hexafluorophosphate
anions.
1
Figure 11. Orientational H-1H ROESY intercomponent correlations observed for 7·Cl, 7·PF6, 8·Cl, and 8·PF6.
Chem. Eur. J. 2011, 17, 12955 – 12966
ꢃ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
12963

P. D. Beer et al.
Rotaxane anion-binding properties: The anion binding prop-
erties of rotaxanes 7·PF6 and 8·PF6 were investigated by
1H NMR titration experiments using TBA salts in 1:1
CDCl3/CD3OD.[28] Downfield shifts for the C-H peaks in-
volved in anion binding (b, 1 and 1’) were observed through-
out these titrations, however it was not possible to follow
the amide protons (c) due to deuterium exchange. The spec-
tra for the titrations of 7·PF6 and 8·PF6 with chloride are
shown in Figure 12 and Figure 13.
Table 2 shows that for both rotaxane hosts 7·PF6 and
8·PF6, the halides are bound considerably more strongly
than the more basic oxoanions dihydrogen phosphate and
acetate. Importantly, this is a reversal in selectivity com-
pared with 1·PF6 and 2·PF6, which has been achieved by the
creation of a three-dimensional interlocked cavity of com-
plementary size and shape for the spherical halide guest spe-
cies. The oxoanions are too large to penetrate the respective
rotaxaneꢅs interlocked binding pocket. The suggestion of a
different binding mode, external to the rotaxane cavity, for
dihydrogen phosphate and acetate is supported by the dif-
ferential behavior of the methyl signal 2 in 7·PF6 during the
titrations. Slight downfield shifts of 0.05–0.07 ppm are ob-
served upon addition of the halide anions, whereas these
protons shift upfield by 0.02 ppm with dihydrogen phos-
phate and do not shift with acetate (<0.01 ppm) (shown in
Figure S20 in the Supporting Information). Hence, the con-
trasting binding modes displayed by 7·PF6 are thought to
alter the position of the imidazolium methyl protons with
respect to the hydroquinone ring currents in differing ways.
It is noteworthy that no selectivity is exhibited between
the halides by either rotaxane. This indicates that there is a
certain amount of flexibility within the interlocked struc-
tures, allowing the cavity to expand and encapsulate the
spherical halide anions of increasing size. This is in contrast
to similar anion-templated interlocked host systems, where
halide discrimination is observed.[16,17]
Figure 12. 1H NMR spectra for solutions in 1:1 CDCl3/CD3OD at 293 K
of a) rotaxane 7·PF6, b) rotaxane 7·PF6 plus 1.0 equivalent of TBACl.
Despite the contrasting hydrogen-bonding arrangements
of the imidazolium groups in 7·PF6 and 8·PF6, there is no
significant difference in the anion-binding selectivity of the
rotaxanes. Hence, it can be concluded that the distinct imi-
dazolium hydrogen-bond donor arrangement within the in-
terlocked binding cavity is of little importance in determin-
ing the anion-recognition properties of the rotaxane. This is
contrary to what might be expected by incorporating geo-
metrically contrasting imidazolium hydrogen-bonding motifs
into the topologically well-defined cavities of interlocked
host structures, and suggests that imidazolium hydrogen-
bond interactions exhibit weak orientational dependence.
Importantly the halide-binding affinity exhibited by 7·PF6
is stronger than that shown by 8·PF6 (674mꢀ1 and 446mꢀ1 re-
spectively for bromide), however, interestingly the differ-
ence is much less pronounced than that observed for 1·PF6
and 2·PF6. Incorporation of substituted imidazolium anion-
binding motifs into the rotaxane host systems appears to
have reduced the impact of the hydrogen-bond donor ar-
rangement on the imidazolium–anion interaction. Hence,
the observed halide-binding affinities of both rotaxanes
7·PF6 and 8·PF6 may be considered to be dominated by elec-
trostatics, with the corresponding axle imidazolium cation to
halide anion distance dependence possibly explaining the
difference in binding strengths between the two rotaxane
host systems.
Figure 13. 1H NMR spectra for solutions in 1:1 CDCl3/CD3OD at 293 K
of a) rotaxane 8·PF6, b) rotaxane 8·PF6 plus 1.0 equivalent of TBACl.
Association constants were determined from the titration
data by monitoring cavity proton
b
and using
WinEQNMR2[22] (Table 2). These protons are directly in-
volved in anion binding, and their signals could be accurate-
ly monitored throughout all titration experiments.
Table 2. Association constants, K, of rotaxanes 7·PF6 and 8·PF6 with
anions in 1:1 CDCl3/CD3OD at 293 K. Obtained by monitoring proton b.
Errors are given in parentheses.
Anion
7·PF6
8·PF6
K [mꢀ1
]
K [mꢀ1
]
Clꢀ
652 (52)
674 (21)
578 (38)
173 (7)
99 (9)
360 (24)
446 (30)
360 (27)
153 (15)
108 (6)
Brꢀ
Iꢀ
ꢀ
H2PO4
OAcꢀ
12964
www.chemeurj.org
ꢃ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 12955 – 12966

Imidazolium Interpenetrated and Interlocked Assemblies
FULL PAPER
an MBraun MPSP-800 column. Water was de-ionized and microfiltered
using a Milli-Q Millipore machine. All tetrabutylammonium salts and
Grubbsꢅ second-generation catalyst were stored in a vacuum desiccator
prior to use.
Conclusion
An investigation into the imidazolium–anion interaction was
undertaken through the anion-templated assembly and syn-
thesis of interpenetrated and interlocked structures. Solu-
tion-phase studies, solid-state structures, computational MD
simulations, and DFT calculations suggest that the strength
of anion binding, commonly thought to occur through
charge-assisted hydrogen bonding, is influenced more by
electrostatics than conventional hydrogen bonding.
In particular, the solid-state structure orientations of imi-
dazolium threading components within the pseudorotaxane
assemblies appear to show weak dependence on any direc-
tional interactions with the templating chloride anion and
contrasting co-conformations with respect to solution-phase
NMR studies. This was investigated by using MD simula-
tions, which illustrate 4,5-dimethylimidazolium thread rota-
tion within the chloride interpenetrated assembly from the
solid-state structure to the co-conformation observed in so-
lution. In contrast, the equivalent event was not observed
for the 2-methylimidazolium derivative.
Further evidence was obtained from the preparation of
the first mono-imidazolium interlocked structures and their
anion binding properties. The topologically well-defined,
three-dimensional interlocked binding cavities within the
host rotaxanes displayed significant selectivity for the spher-
ical halide anions over the larger, more basic oxoanions,
however with limited preference between the halides. The
only difference in the anion-binding properties between the
rotaxanes, which incorporate imidazolium groups of con-
trasting hydrogen-bond donor geometries, was stronger
binding for the 4,5-dimethylimidazolium system. This sug-
gests that the imidazolium–halide anion binding interaction
within the interlocked cavity is electrostatically dominated
and less dependent on the exact hydrogen-bonding arrange-
ment.
It is concluded that viewing the interaction between the
imidazolium cation and counter anion as charge-assisted hy-
drogen bonding is inadequate when considering the con-
struction, co-conformations and properties of interpenetrat-
ed and interlocked assemblies in the field of anion-coordina-
tion chemistry. We suggest that such interactions might be
better described as charge-driven hydrogen bonding. Our
findings for the acyclic, pseudorotaxane and rotaxane sys-
tems described herein indicate that, in anion recognition,
electrostatic interactions dominate with respect to imidazoli-
um–anion binding, with the hydrogen-bonding contributions
having weak orientational dependence, and are therefore of
fundamental significance for the future design of imidazol-
NMR spectra were recorded on Varian Mercury 300, Varian Unity Plus
500 and Bruker AVII 500 (with 13C Cryoprobe) spectrometers. Mass
spectrometry was carried out on a Bruker micrOTOF spectrometer. Mi-
crowave reactions were carried out using a Biotage Initiator 2.0 micro-
wave. Melting points were recorded on a Gallenkamp capillary melting
point apparatus and are uncorrected.
Literature procedures were used in the preparation of 1·Cl, 2·Cl,[19] 3,[18]
and 6.[14] The syntheses of 1·PF6, 2·PF6, 4·Cl, and 5·Cl are given in the
Supporting Information.
Rotaxane 7·Cl: Axle 4·Cl (200 mg, 0.164 mmol) and macrocycle precursor
6 (106 mg, 0.164 mmol) were dissolved in dry CH2Cl2 (50 mL) and stirred
for 1 h. Grubbsꢅ second-generation catalyst (10% by weight, 10.6 mg)
was then added and the reaction mixture was left to stir at room temper-
ature for two days under N2. The solvent was removed in vacuo and the
crude product was purified by preparative silica thin-layer chromatogra-
phy (94:6 CH2Cl2/CH3OH followed by 4:6 CH2Cl2/CH3CN) to yield the
product as an off white solid (90 mg, 0.049 mmol, 30%). M.p. 167–1698C
(decomp); 1H NMR (500 MHz, CDCl3): d=10.03 (s, 1H; Hb), 9.79 (s,
1H; H1), 9.36 (br s, 2H; Hc), 9.09 (s, 2H; Ha), 7.23–7.26 (m, 12H; H9),
7.08–7.13 (m, 16H; H7 & H8), 6.64–6.66 (m, 4H; H6), 6.36–6.42 (m, 8H;
Hf & Hg), 5.82 (br s, 2H; Hk), 4.15–4.17 (m, 4H; He), 3.97 (br s, 4H; Hj),
3.92–3.94 (m, 4H; Hd), 3.81–3.82 (m, 4H; Hh), 3.69–3.70 (m, 4H; Hi),
3.62–3.64 (m, 4H; H5), 3.57–3.60 (m, 4H; H3), 1.94–1.97 (m, 4H; H4),
1.88 (s, 6H; H2), 1.27 ppm (s, 54H; H10); 13C NMR (75.5 MHz, CDCl3): d
= 164.2, 155.7, 152.7, 152.3, 149.0, 148.3, 144.0, 140.5, 135.6, 132.4, 130.6,
128.8, 126.4, 126.0, 124.1, 114.5, 113.8, 112.7, 70.9, 69.2, 67.7, 66.4, 63.7,
63.0, 43.8, 40.4, 34.3, 31.4, 28.7, 8.0 ppm; HRMS (ESI): m/z: [MꢀCl]+
1808.0551 C117H140N5O12 (calcd 1808.0526).
Rotaxane 7·PF6: Rotaxane 7·Cl (50 mg, 0.027 mmol) was dissolved in
CH2Cl2 (10 mL) and washed with 0.1m NH4PF6 (aq) (10”10 mL) and
water (4”10 mL). The organic fraction was dried over MgSO4 and in
vacuo to yield the product as an off white solid (49 mg, 0.025 mmol,
93%). M.p. 158–1598C (decomp); 1H NMR (300 MHz, CDCl3): d =9.02
(s, 2H; Ha), 8.40 (br s, 1H; Hb), 7.97 (s, 1H; H1), 7.23–7.26 (m, 12H; H9),
3
7.08–7.15 (m, 16H; H7 & H8), 6.68–6.71 (m, 4H; H6), 6.57 (d, J =8.8 Hz,
4H; Hf), 6.33 (d, 3J =8.8 Hz, 4H; Hg), 5.83 (br s, 2H; Hk), 4.08–4.11 (m,
4H; He), 3.89–3.95 (br s, 8H; Hd & Hj), 3.76–3.79 (m, 8H; Hh & H5),
3.61–3.68 (m, 8H; Hi & H3), 1.91–1.98 (m, 4H; H4), 1.89 (s, 6H; H2),
1.31 ppm (s, 54H; H10); 19F NMR (282.5 MHz, CDCl3): d=ꢀ70.1 ppm (d,
1J= 760 Hz; PF6); 13C NMR (75.5 MHz, CDCl3): d=164.4, 155.7, 153.0,
152.2, 149.0, 148.3, 144.0, 140.5, 136.4, 132.3, 130.6, 129.6, 126.9, 126.6,
124.1 114.9, 113.8, 112.8, 70.8, 69.3, 67.5, 66.6, 63.5, 63.0, 44.0, 40.0, 34.3,
31.4, 28.7, 7.8 ppm; HRMS (ESI): m/z: [MꢀPF6]+ 1808.0535
C
117H140N5O12 (calcd 1808.0526).
Rotaxane 8·Cl: Axle 5·Cl (200 mg, 0.166 mmol) and macrocycle precursor
6 (108 mg, 0.166 mmol) were dissolved in dry CH2Cl2 (50 mL) and stirred
for 1 h. Grubbsꢅ second-generation catalyst (10% by weight, 10.8 mg)
was then added and the reaction mixture was left to stir at room temper-
ature for two days under N2. The solvent was removed in vacuo and the
crude product was purified by preparative silica thin-layer chromatogra-
phy (95:5 CH2Cl2/CH3OH followed by ethyl acetate) to yield the product
as an off white solid (58 mg, 0.031 mmol, 19%). M.p. 162–1648C
(decomp); 1H NMR (300 MHz, CDCl3): d =9.96 (s, 1H; Hb), 9.40 (br s,
2H; Hc), 9.08 (s, 2H; Ha), 7.47 (s, 2H; H1’), 7.24–7.27 (m, 12H; H9’),
7.09–7.16 (m, 16H; H7ꢅ & H8’), 6.67–6.70 (m, 4H; H6’), 6.55 (d, 3J=
8.8 Hz, 4H; Hf), 6.33 (d, 3J=8.8 Hz, 4H; Hg), 5.82 (br s, 2H; Hk), 4.17–
4.20 (m, 4H; He), 3.97–4.02 (m, 4H; Hd), 3.92 (br s, 4H; Hj), 3.78–3.79
(m, 4H; Hh), 3.65–3.73 (m, 12H; Hi & H3ꢅ & H5’), 2.07 (s, 3H, H2’), 1.86–
1.94 (m, 4H; H4’), 1.31 ppm (s, 54H; H10’); 13C NMR (75.5 MHz, CDCl3):
d=164.3, 155.7, 153.2, 152.0, 148.9, 148.4, 143.9, 142.5, 140.6, 135.7, 132.4,
130.6, 129.2, 126.3, 124.1, 121.2, 115.1, 113.7, 112.8, 70.8, 69.4, 67.4, 67.0,
63.5, 63.0, 45.1, 39.9, 34.3, 31.4, 28.8, 8.7 ppm; HRMS (ESI): m/z:
[MꢀCl]+ 1794.0417 C116H138N5O12 (calcd 1794.0388).
ACHTUNGTRENNUNGium-based anion–host systems.
Experimental Section
General considerations: Commercially available solvents and chemicals
were used without further purification unless otherwise stated. Dry sol-
vents were obtained by purging with nitrogen and then passing through
Chem. Eur. J. 2011, 17, 12955 – 12966
ꢃ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
12965

P. D. Beer et al.
Rotaxane 8·PF6: Rotaxane 8·Cl (29 mg, 0.016 mmol) was dissolved in
CH2Cl2 (10 mL) and washed with 0.1m NH4PF6 (aq) (10”10 mL) and
water (4”10 mL). The organic fraction was dried over MgSO4 and in
vacuo to yield the product as an off white solid (25 mg, 0.13 mmol,
81%). M.p. 156–1608C (decomp); 1H NMR (300 MHz, CDCl3): d=9.01
(s, 2H, Ha), 8.48 (s, 1H; Hb), 7.22–7.26 (m, 12H; H9’), 7.07–7.17 (m, 16H;
H7ꢅ & H8’), 6.96 (s, 2H; H1’), 6.71–6.74 (m, 4H; H6’), 6.59 (d, 3J=8.8 Hz,
4H; Hf), 6.33 (d, 3J =8.8 Hz, 4H; Hg), 5.80 (br s, 2H; Hk), 4.17–4.20 (m,
4H; He), 3.90–3.94 (m, 4H; Hd), 3.89 (br s, 4H; Hj), 3.81–3.84 (m, 4H;
H5’), 3.62–3.75 (m, 12H; Hh & Hi & H3’), 2.07 (s, 3H; H2’), 1.89–1.96 (m,
4H; H4’), 1.31 ppm (s, 54H; H10’); 19F NMR (282.5 MHz, CDCl3): d=
ꢀ70.4 (d, 1J=759 Hz; PF6); 13C NMR (125.7 MHz, CDCl3): d=164.5,
155.8, 153.2, 152.2, 148.9, 148.4, 144.0, 142.8, 140.6, 136.4, 132.4, 130.6,
129.6, 126.4, 124.1, 120.7, 115.3, 113.9, 112.8, 70.8, 69.4, 67.4, 66.8, 63.6,
63.1, 45.2, 39.9, 34.3, 31.4, 28.5, 8.7 ppm; HRMS (ESI): m/z: [MꢀPF6]+
1793.9404 C116H138N5O12 (calcd 1794.0388).
[12] K. Fumino, A. Wulf, R. Ludwig, Angew. Chem. 2008, 120, 8859–
8862; Angew. Chem. Int. Ed. 2008, 47, 8731–8734; W. Zhao, F.
Leroy, B. Heggen, S. Zahn, B. Kirchner, S. Balasubramanian, F.
Müller-Plathe, J. Am. Chem. Soc. 2009, 131, 15825–15833; S. B. C.
Lehmann, M. Roatsch, M. Schçppke, B. Kirchner, Phys. Chem.
Chem. Phys. 2010, 12, 7473–7486; A. Wulf, K. Fumino, R. Ludwig,
Angew. Chem. 2010, 122, 459–463; Angew. Chem. Int. Ed. 2010, 49,
449–453; C. Roth, T. Peppel, K. Fumino, M. Kçckerling, R. Ludwig,
Angew. Chem. Int. Ed. 2010, 49, 10221–10224; T. Cremer, C. Kol-
beck, K. R. J. Lovelock, N. Paape, R. Wçlfel, P. S. Schulz, P. Was-
serscheid, H. Weber, J. Thar, B. Kirchner, F. Maier, H.-P. Steinrück,
Chem. Eur. J. 2010, 16, 9018–9033.
[13] J. A. Wisner, P. D. Beer, M. G. B. Drew, M. R. Sambrook, J. Am.
Chem. Soc. 2002, 124, 12469–12476; M. D. Lankshear, P. D. Beer,
Coord. Chem. Rev. 2006, 250, 3142–3160; M. D. Lankshear, P. D.
Beer, Acc. Chem. Res. 2007, 40, 657–668.
[14] M. R. Sambrook, P. D. Beer, M. D. Lankshear, R. F. Ludlow, J. A.
Wisner, Org. Biomol. Chem. 2006, 4, 1529–1538.
[15] L. M. Hancock, L. C. Gilday, S. Carvalho, P. J. Costa, V. Fꢀlix, C. J.
Serpell, N. L. Kilah, P. D. Beer, Chem. Eur. J. 2010, 16, 13082–
13094.
Acknowledgements
[16] K. M. Mullen, J. Mercurio, C. J. Serpell, P. D. Beer, Angew. Chem.
2009, 121, 4875–4878; Angew. Chem. Int. Ed. 2009, 48, 4781–4784.
[17] N. L. Kilah, M. D. Wise, C. J. Serpell, A. L. Thompson, N. G. White,
K. E. Christensen, P. D. Beer, J. Am. Chem. Soc. 2010, 132, 11893–
11895.
[18] M. R. Sambrook, P. D. Beer, J. A. Wisner, R. L. Paul, A. R. Cowley,
F. Szemes, M. G. B. Drew, J. Am. Chem. Soc. 2005, 127, 2292–2302.
[19] C. J. Serpell, N. L. Kilah, P. J. Costa, V. Fꢀlix, P. D. Beer, Angew.
Chem. 2010, 122, 5450–5454; Angew. Chem. Int. Ed. 2010, 49, 5322–
5326.
[20] A. K. Mandal, M. Suresh, A. Das, Org. Biomol. Chem. 2011, 9,
4811–4817; H.-Y. Gong, B. M. Rambo, E. Karnas, V. M. Lynch,
K. M. Keller, J. L. Sessler, J. Am. Chem. Soc. 2011, 133, 1526–1533.
[21] H.-Y. Gong, B. M. Rambo, W. Cho, V. M. Lynch, M. Oh, J. L. Sess-
ler, Chem. Commun. 2011, 47, 5973–5975; C. J. Serpell, R. Chall,
A. L. Thompson, P. D. Beer, Dalton Trans., DOI:10.1039/
C1DT10186B.
[22] M. J. Hynes, J. Chem. Soc. Dalton Trans. 1993, 311–312.
[23] A. M. Magill, B. F. Yates, Aust. J. Chem. 2004, 57, 1205–1210.
[24] We have previously published the structure of 1·Cl,[19] however it
was not subjected to the analysis detailed in this paper.
G.T.S. thanks the EPSRC for a studentship (EP/F011504. C.J.S. thanks
the EPSRC and Johnson Matthey for a CASE studentship and the
EPSRC for a PhD+ award. P.J.C. thanks FCT for the postdoctoral grant
SFRH/BPD/27082/2006. V.F. acknowledges the financial support from
FCT under the project PTDC/QUI/68582/2006 with co-participation of
the European Community funds FEDER, QREN, and COMPETE. We
are grateful to Diamond Light Source for the award of beamtime
(MT1880/MT1858) and to the beamline scientists for their assistance, and
we acknowledge Oxford Chemical Crystallography for use of their instru-
ments.
[1] T. Welton, Chem. Rev. 1999, 99, 2071–2084; H. Weingärtner,
Angew. Chem. 2008, 120, 664–682; Angew. Chem. Int. Ed. 2008, 47,
654–670.
[2] K. Binnemans, Chem. Rev. 2005, 105, 4148–4204.
[3] J. Dupont, J. D. Scholten, Chem. Soc. Rev. 2010, 39, 1780–1804; M.-
A. Neouze, J. Mater. Chem. 2010, 20, 9593–9607.
[4] H. Jacobsen, A. Correa, A. Poater, C. Costabile, L. Cavallo, Coord.
Chem. Rev. 2009, 253, 687–703; S. Wei, X.-G. Wei, X. Su, J. You, Y.
Ren, Chem. Eur. J. 2011, 17, 5965–5971.
[5] K. Sato, S. Arai, T. Yamagishi, Tetrahedron Lett. 1999, 40, 5219–
5222.
[6] E. Alcalde, N. Mesquida, L. Perez-Garcia, C. Alvarez-Rua, S.
Garcia-Granda, E. Garcia-Rodriguez, Chem. Commun. 1999, 295–
296.
[7] J. Yoon, S. K. Kim, N. J. Singh, K. S. Kim, Chem. Soc. Rev. 2006, 35,
355–360; Z. Xu, S. K. Kim, J. Yoon, Chem. Soc. Rev. 2010, 39,
1457–1466.
[25] D. A. Case, T. A. Darden, T. E. Cheatham III, C. L. Simmerling, J.
Wang, R. E. Duke, R. Luo, R. C. Walker, W. Zhang, K. M. Merz, B.
Roberts, B. Wang, S. Hayik, A. Roitberg, G. Seabra, I. Kolossvai,
K. F. Wong, F. Paesani, J. Vanicek, J. Liu, X. Wu, S. R. Brozell, T.
Steinbrecher, H. Gohlke, Q. Cai, X. Ye, J. Wang, M.-J. Hsieh, G.
Cui, D. R. Roe, D. H. Mathews, M. G. Seetin, C. Sagui, V. Babin, T.
Luchko, S. Gusarov, A. Kovalenko, P. A. Kollman, AMBER 11,
University of California, San Francisco, 2010.
[26] J. Wang, R. M. Wolf, J. W. Caldwell, P. A. Kollman, D. A. Case, J.
Comput. Chem. 2004, 25, 1157–1174.
[8] H. Ihm, S. Yun, H. G. Kim, J. K. Kim, K. S. Kim, Org. Lett. 2002, 4,
2897–2900.
[27] Interactions involving the C2-X, C4-X and C5-X (X=H or CH3) sites
to chloride are drawn in blue, red, and green respectively as can be
seen in Figure 9 and 10. This colour scheme was maintained for the
subsequent modeling figures presented in the Supporting Informa-
tion.
[9] K. Chellappan, N. J. Singh, I.-C. Hwang, J. W. Lee, K. S. Kim,
Angew. Chem. 2005, 117, 2959–2963; Angew. Chem. Int. Ed. 2005,
44, 2899–2903; W. W. H. Wong, M. S. Vickers, A. R. Cowley, R. L.
Paul, P. D. Beer, Org. Biomol. Chem. 2005, 3, 4201–4208; H.-T. Niu,
Z. Yin, D. Su, D. Niu, J. He, J.-P. Cheng, Dalton Trans. 2008, 3694–
3700; P. P. Neelakandan, D. Ramaiah, Angew. Chem. 2008, 120,
8535–8539; Angew. Chem. Int. Ed. 2008, 47, 8407–8411; C. J. Ser-
pell, J. Cookson, A. L. Thompson, P. D. Beer, Chem. Sci. 2011, 2,
494–500; Z. Xu, N. J. Singh, S. K. Kim, D. R. Spring, K. S. Kim, J.
Yoon, Chem. Eur. J. 2011, 17, 1163–1170.
[28] Due to the insolubility of TBAOAc in pure CDCl3, titrations with
TBAOAc were not able to be carried out for the simple acyclic imi-
dazolium derivatives. However, they were possible for the rotaxane
host systems which were carried out in 1:1 CDCl3/CD3OD.
[10] P. A. Hunt, B. Kirchner, T. Welton, Chem. Eur. J. 2006, 12, 6762–
6775.
[11] S. Tsuzuki, H. Tokuda, M. Mikami, Phys. Chem. Chem. Phys. 2007,
9, 4780–4784.
Received: June 29, 2011
Published online: October 5, 2011
12966
www.chemeurj.org
ꢃ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 12955 – 12966