Anion-templated assembly of pseudorotaxanes: Importance of anion template, strength of ion-pair thread association, and macrocycle ring size

Article abstract of DOI:10.1021/ja046278z

A wide range of pseudorotaxane assemblies containing positively charged pyridinium, pyridinium nicotinamide, imidazolium, benzimidazolium and guanidinium threading components, and macrocyclic isophthalamide polyether ligands have been prepared using a general anion templation procedure. In noncompetitive solvent media, coupling halide anion recognition by a macrocyclic ligand with ion-pairing between the halide anion and a strongly associated cation provides the driving force for interpenetration. Extensive solution ¹H NMR binding studies, thermodynamic investigations, and single-crystal X-ray structure determinations reveal that the nature of the halide anion template, strength of the ion-pairing between the anion template and the cationic threading component, and to a lesser extent favorable second sphere π-π aromatic stacking interactions between the positively charged threading component and macrocyclic ligand, together with macrocyclic ring size, affect the efficacy of pseudorotaxane formation.

Full text of DOI:10.1021/ja046278z

Published on Web 01/26/2005
Anion-Templated Assembly of Pseudorotaxanes: Importance
of Anion Template, Strength of Ion-Pair Thread Association,
and Macrocycle Ring Size
Mark R. Sambrook,† Paul D. Beer,*,† James A. Wisner,†,‡ Rowena L. Paul,†
Andrew R. Cowley,† Fridrich Szemes,† and Michael G. B. Drew§
Contribution from the Department of Chemistry, Inorganic Chemistry Laboratory, UniVersity of
Oxford, South Parks Road, Oxford, OX1 3QR, U.K., Department of Chemistry, The UniVersity of
Western Ontario, Chemistry Building, London, Ontario, N6A 5B7, Canada, and Department of
Chemistry, UniVersity of Reading, Reading, RG6 6AD, U.K.
Received June 23, 2004; E-mail: paul.beer@chem.ox.ac.uk
Abstract: A wide range of pseudorotaxane assemblies containing positively charged pyridinium, pyridinium
nicotinamide, imidazolium, benzimidazolium and guanidinium threading components, and macrocyclic
isophthalamide polyether ligands have been prepared using a general anion templation procedure. In
noncompetitive solvent media, coupling halide anion recognition by a macrocyclic ligand with ion-pairing
between the halide anion and a strongly associated cation provides the driving force for interpenetration.
Extensive solution 1H NMR binding studies, thermodynamic investigations, and single-crystal X-ray structure
determinations reveal that the nature of the halide anion template, strength of the ion-pairing between the
anion template and the cationic threading component, and to a lesser extent favorable second sphere
π-π aromatic stacking interactions between the positively charged threading component and macrocyclic
ligand, together with macrocyclic ring size, affect the efficacy of pseudorotaxane formation.
Introduction
formation of, in particular, inorganic-based polymetallic cage
complexes8 and circular double helicates9 have now appeared,
Imaginative template methodologies are increasingly being
used in the construction of complex supramolecular architec-
tures,1 ranging from molecular cagelike assemblies and helicates
to mechanically interlocked supramolecules, such as rotaxanes,
catenanes, and knots.2 Cationic and neutral species have
dominated the templated synthetic strategies reported to date
by employing metal-ligand coordination,3 π-π stacking in-
teractions,4 hydrogen bonding,5 and solvophobic effects6 to
effect assembly between two or more components. The chal-
lenge of exploiting anions to direct supramolecular assembly
remains largely under-developed, which may be attributed to
their diffuse nature, pH dependence, and relative high solvation
energy as compared to that of cations.7 Although various
serendipitous discoveries where anions have templated the
strategic anion-templated syntheses and assemblies are rare.10
(3) (a) Dietrich-Buchecker, C. O.; Sauvage. J.-P. Tetrahedron Lett. 1983, 24,
5095-5098. (b) Dietrich-Buchecker, C. O.; Sauvage, J.-P.; Kern, J.-M. J.
Am. Chem. Soc. 1984, 106 (10), 3043-3045. (c) Weck, M.; Mohr, B.;
Sauvage, J.-P.; Grubbs, R. H. J. Org. Chem. 1999, 64 (15), 5463-5471.
(d) Dietrich-Buchecker, C. O.; Sauvage, J.-P. Angew. Chem., Int. Ed. Engl.
1989, 28, 189-192. (e) Dietrich-Buchecker, C. O.; Sauvage, J.-P. Angew.
Chem., Int. Ed. Engl. 1989, 28, 189-192. (f) Fujita, M.; Ibukuro, F.;
Hagihara, H.; Ogura, K. Nature 1994, 367, 720-722. (g) Fujita, M.;
Ibukuro, F.; Yamaguchi, K.; Ogura, K. J. Am. Chem. Soc. 1995, 117 (14),
4175-4176. (h) Padilla-Tosta, M. E.; Fox, O. D.; Drew, M. G. B.; Beer,
P. D. Angew. Chem., Int. Ed. 2001, 40 (22), 4235-4239. (i) Wu, C.;
Lecavalier, P. R.; Shen, Y. X.; Gibson, H. W. Chem. Mater. 1991, 3 (4),
569-572.
(4) (a) Ashton, P. R.; Ballardini, R.; Balzani, V.; Beˇlohradsky, M.; Gandolfi,
M. T.; Philp, D.; Prodi, L.; Raymo, F. M.; Reddington, M. V.; Spencer,
N.; Stoddart, J. F.; Venturi, M.; Williams, D. J. J. Am. Chem. Soc. 1996,
118 (21), 4931-4951. (b) Stoddart, J. F.; Tseng, H.-R. Proc. Natl. Acad.
Sci. U.S.A. 2002, 99 (8), 4797-4800. (c) Raymo, F. M.; Houk, K. N.;
Stoddart, J. F. J. Am. Chem. Soc. 1998, 120 (36), 9318-9322. (d) Ashton,
P. R.; Baxter, I.; Fyfe, M. C. T.; Raymo, F. M.; Spencer, N.; Stoddart, J.
F.; White, A. J. P.; Williams, D. J. J. Am. Chem. Soc. 1998, 120 (10),
2297-2307.
(5) (a) Hunter, C. A. J. Am. Chem. Soc. 1992, 114 (13), 5303-5311. (b) Adams,
H.; Carver, F. J.; Hunter, C. A. Chem. Commun. 1995, 809-810. (c) Vo¨gtle,
F.; Meier, S.; Hoss, R. Angew. Chem., Int. Ed. Engl. 1992, 31, 1619-
1622. (d) Ottens-Hildebrandt, S.; Nieger, M.; Rissanen, K.; Rouvinen, J.;
Meier, S.; Harder, G.; Vo¨gtle, F. J. Chem. Commun. 1995, 777. (e) Johnston,
A. G.; Leigh, D. A.; Pritchard, R. J.; Deegan, M. D. Angew. Chem., Int.
Ed. Engl. 1995, 34, 1209-1212.
(6) Anderson, S.; Anderson, H. L. Angew. Chem., Int. Ed. Engl. 1996, 35,
1956-1958.
(7) Beer, P. D.; Gale P. A. Angew. Chem., Int. Ed. 2001, 40 (3), 486-516.
(8) Fleming, J. S.; Mann, K. L. V.; Carraz, C.-A.; Psillakis, E.; Jeffery, J. C.;
McCleverty, J. A.; Ward, M. D. Angew. Chem., Int. Ed. 1998, 37 (9), 1279-
1281.
† University of Oxford.
‡ The University of Western Ontario.
§ University of Reading.
(1) (a) Raymo, F. M.; Stoddart, J. F. Templated Organic Synthesis; Wiley:
Weinheim, Germany, 2000. (b) ComprehensiVe Supramolecular Chemistry;
Sauvage, J.-P., Hosseini, M., Eds.; Elsevier: New York, 1996; Vol. 9. (c)
Davis, A. V.; Yeh, R. M.; Raymond, K. N. Proc. Natl. Acad. Sci. U.S.A.
2002, 99 (8), 4793-4796. (d) Furlan, R. L. E.; Otto, S.; Sanders, J. K. M.
Proc. Natl. Acad. Sci. U.S.A. 2002, 99 (8), 4801-4804. (e) Krische, M. J.;
Lehn, J.-M. Struct. Bonding 2000, 96, 3-29. (f) Greig, L. M.; Philp, D.
Chem. Soc. ReV. 2001, 30, 287-302. (f) Gale, P. A. Annu. Rep. Prog.
Chem., Sect. B 2002, 98, 581-605. (g) Ward, M. D. Annu. Rep. Prog.
Chem., Sect A 2002, 98, 285-320. (h) Lindsey, J. S. New. J. Chem. 1991,
15 (2-3), 153-180.
(2) (a) Hubin, T. J.; Busch, D. H. Coord. Chem. ReV. 2000, 200-202, 5-52.
(b) Breault, G. A.; Hunter, C. A.; Mayers, P. C. Tetrahedron 1999, 55,
5265-5293. (c) Panova, I. G.; Topchieva, I. N. Russ. Chem. ReV. 2001,
70 (1), 23-44. (d) Sauvage, J.-P. Acc. Chem. Res. 1990, 23 (10), 319-
327.
(9) (a) Hasenknopf, B.; Lehn, J.-M.; Kniesel, B. O.; Baum, G.; Fenske, D.
Angew. Chem., Int. Ed. Engl. 1996, 35 (16), 1838-1840. (b) Hasenknopf,
B.; Leh, J.-M.; Boumediene, N.; Dupont-Gervais, A.; Van Dorsselaer, A.;
Kneisel, B.; Fenske, D. J. Am. Chem. Soc. 1997, 119 (45), 10956-10962.
9
2292
J. AM. CHEM. SOC. 2005, 127, 2292-2302
10.1021/ja046278z CCC: $30.25 © 2005 American Chemical Society

Anion-Templated Assembly of Pseudorotaxanes
A R T I C L E S
Results and Discussion
Thread and Macrocycle Ligand Design for Anion-Tem-
plated Pseudorotaxane Formation. When the strategic design
features for integrating anionic guest species into a pseudoro-
taxane superstructure are taken into account (see Figure 1), the
anion template is required to simultaneously coordinate two
components in an orthogonal fashion. Crabtree and co-workers
have shown that simple isophthalamide molecules, such as 1a,
are receptors for anions in chloroform.17 In particular, they bind
chloride, bromide, and iodide anions exclusively in a 1:1
stoichiometry, which is unsuitable for the formation of the
desired template stoichiometry of 2 receptor:1 anion. This
problem was overcome by the utilization of a cationic receptor
containing an amide cleft anion recognition site as one of the
ligands. As a consequence of favorable electrostatics and
increased acidity of the amide and aryl protons, the pyridinium
bisamide receptor 2+ is designed to bind anions strongly in
noncompetitive solvent media. Indeed, in acetone solution, ion-
pairing between the pyridinium cation and the chloride anion
in 2a is very strong, as confirmed by 1H NMR titration
experiments with tetrabutylammonium (TBA) Cl-, where an
association constant of >105 M-1 was obtained for the equi-
librium:18
Figure 1. Anion-templated self-assembly of pseudorotaxanes. Recognition
of the anion by the macrocycle results in the formation of an interpenetrated
structure.
This is especially the case when employing anions in the
construction of interpenetrated compounds. Stoddart et al.
reported the organizational role a hexafluorophosphate anion
can play in the assembly of large [5]- and [6]-pseudorotaxanes
from dibenzylammonium threads and polyether macrocycles.11
Vo¨gtle et. al. have used phenolates, thiophenolates, and sul-
fonamide anions which when complexed to a cyclic tetralactam
and subsequently reacted with a sterically demanding electro-
phile, affords [2]-rotaxane products in high yields.12 This elegant
anion-templated “snapping” methodology has recently been
exploited by Smith in the preparation of a series of ion-pair
binding rotaxanes13 and in a related “stoppering” method of
rotaxane formation by Schalley and co-workers.14
Kassoc
Herein, we report the use of a rational design procedure to
develop a general method of using anions to template the
formation of a wide range of pseudorotaxanes, based on the
coupling of anion recognition with ion-pairing (Figure 1).15
TBA+Cl- + 2d y z 2a + TBA+PF6
-
Complexation of the chloride anion occurs in the hydrogen
bond donor cleft formed by the amide groups and their mutually
ortho-aryl proton. The resulting tight ion-pair 2a leaves the
chloride anion with an unsaturated coordination sphere and
presents an empty meridian available for binding another
hydrogen bond donor ligand in the absence of a competitive
solvent. As a preliminary test of this hypothesis, 1H NMR
titration of 1b with 2a in acetone-d6 produced a 1:1 association
constant of 100 M-1, while an analogous titration experiment
of 1b with 2d gave no evidence of association.
This templation strategy is designed to operate in noncom-
petitive solvent media where the anion of the ion-pair is strongly
associated with the potential cationic threading component and
importantly remains coordinatively unsaturated. Subsequent
anion recognition by the macrocyclic ligand results in pseu-
dorotaxane formation as the cationic thread strongly associates
with the complexed anion within the macrocyclic cavity. The
success of this general anion template procedure is illustrated
with the halide-directed assembly of a series of novel pseu-
dorotaxanes containing pyridinium, pyridinium nicotinamide,
imidazolium, benzimidazolium and guanidinium threading
components, and macrocyclic diamide polyether ligands. Solu-
tion 1H NMR binding studies and isothermal calorimetry (ITC),
together with X-ray crystal structure analyses, demonstrate that
the nature of the anion template and the strength of ion-pairing
in the threading guest species are both crucial to the thermo-
dynamic stability of pseudorotaxane assembly.16
The macrocycle component was designed to act as the wheel
through which an ion-pair, such as 2a, would thread to form
the [2]-pseudorotaxane. It was anticipated that an isophthalamide
(17) (a) Kavallieratos, K.; Bertao, C. M.; Crabtree, R. H. J. Org. Chem. 1999,
64 (5), 1675-1683. (b) Kavallieratos, K.; de Gala, S. R.; Austin, D. J.;
Crabtree, R. H. J. Am. Chem. Soc. 1997, 119 (9), 2325-2326.
(18) It is also feasible to view this association constant as representative of the
equilibrium:
Other possible equilibria in solution are:
(10) (a) Vilar R. Angew. Chem., Int. Ed. 2003, 42 (13), 1460-1477. (b) Beer,
P. D.; Sambrook, M. R. In Encylopedia of Nanoscience and Nanotechnol-
ogy; Marcel Dekker Inc.: New York, in press.
(11) Fyfe, M. C. T.; Glink, P. T.; Menzer, S.; Stoddart, J. F.; White, A. J. P.;
Williams, D. J. Angew. Chem., Int. Ed. Engl. 1997, 36 (19), 2068-2070.
(12) (a) Seel, C.; Vo¨gtle, F. Chem.sEur. J. 2000, 6 (1), 21-24. (b) Reuter, C.;
Wienand, W.; Hubner, G. M.; Seel, C.; Vo¨gtle, F. Chem.sEur. J. 1999, 5
(9), 2692-2697. (c) Hubner, G. M.; Gla¨ser, J.; Seel, C.; Vo¨gtle, F. Angew.
Chem., Int. Ed. 1999, 38 (3), 383-386.
(13) (a) Mahoney, J. M.; Shukla, R.; Marshall, R. A.; Beatty, A. M.; Zajicek,
J.; Smith, B. D. J. Org. Chem. 2002, 67 (5), 1757-1767. (b) Deetz, M. J.;
Shukla, R.; Smith, B. D. Tetrahedron 2002, 58, 799-805. (c) Shukla, R.;
Deetz, M. J.; Smith, B. D. Chem. Commun. 2000, 2397-2398.
(14) Ghosh, P.; Mermagen, O.; Schalley, C. A. Chem. Commun. 2002, 2628-
2629.
TBA salts are used frequently in anion-binding studies due to their
noncompetitive nature; the dissociation constant, K1, has been previously
determined to be 16.6 × 104 M-1 in acetone: (a) Savedoff, L. G. J. Am.
-
Chem. Soc. 1966, 88 (4), 664-667. The PF6 anion is also commonly
used in anion-binding studies due to its noncoordinating nature, that is,
(15) Jones, J. W.; Gibson, H. W. J. Am. Chem. Soc. 2003, 125 (23), 7001-
the dissociation constant, K2, is very large, and the pyridinium countercation
can be assumed to be “free” in solution (dissociation constant for TBA+PF6
-
7004.
(16) Part of this work has been published as preliminary communications: (a)
Wisner, J. A.; Beer, P. D.; Drew, M. G. B. Angew. Chem., Int. Ed. 2001,
40 (19), 3606-3609. (b) Wisner, J. A.; Beer, P. D.; Berry, N. G.;
Tomapatanaget, B. Proc. Natl. Acad. Sci. U.S.A. 2002, 99 (8), 4983-4986.
≈ 6300 M-1 in CH2Cl2): (b) Nelsen, S. F.; Ismagilov, R. F. J. Phys. Chem.
A 1999, 103, 5373-5378. For ion-pairing with platinum complex cations,
see: (c) Romeo, R.; Arena, G.; Scolaro, L. M.; Plutino, M. R. Inorg. Chim.
Acta 1995, 240, 81-92.
9
J. AM. CHEM. SOC. VOL. 127, NO. 7, 2005 2293

A R T I C L E S
Sambrook et al.
Figure 2. Structures of 1a,b and 2a-d.
anion recognition fragment incorporated within the macrocyclic
cavity would satisfy the coordinatively unsaturated chloride
anion. In addition, the integration of hydroquinone groups and
polyether functionalities into the cyclic framework would
stabilize the cationic pyridinium component of the ion-pair in
the interpenetrated structure.
1
Figure 4. Observed perturbations in the aromatic region of the H NMR
of macrocycle 4 in acetone-d6 at 293 K upon addition of 1 equiv of TBA
chloride. Bottom: 4. Top: 4 + TBACl.
Table 1. Association Constants, Ka,exp (M-1), for TBA Salts as
Determined by 1H NMR Titrations in Acetone-d6 at 293 K (errors
less than 10%)
The new macrocycles, potential pyridinium, and pyridinium
nicotinamide ion-pair threads (Figure 3) were prepared using
standard organic synthetic procedures as described in the
Supporting Information.
anionic
guest
3
4
5
Cl
Br
I
2420
480
45
2560
470
50
2800
685
60
1H NMR Binding Studies. The anion binding ability of
macrocycles 3-5 was investigated initially by the addition of
1 molar equiv of anion, as its tetrabutylammonium (TBA) salt,
to a solution of the macrocycle in acetone-d6. Upon addition of
Cl-, Br-, or I-, there is an observed downfield shift in the
macrocyclic amide and aryl protons (e and d), indicative of anion
binding within the amide cleft. In addition, a downfield shift in
the aromatic protons h is observed, whereas aromatic protons i
affinities and selectivities between the three macrocycles. In
all cases, chloride anions are the most strongly bound, followed
by bromide and then iodide. This binding trend is a result of
the near-identical nature of the isophthalamide anion binding
cleft in the three systems. Chloride anions are bound most tightly
as a result of increased hydrogen bond acceptor ability and
through size match between the anion binding site and the ionic
radius of the guest. Both bromide and iodide are larger anions
as well as poorer hydrogen bond acceptors. No binding is
observed for the large, noncoordinating hexafluorophosphate
anion.
1
remain unperturbed (see Figure 4). No perturbations in the H
NMR spectra of macrocycles 3-5 are observed upon addition
of TBA hexafluorophosphate, indicative of an absence of anion
binding.
1
Quantitative anion H NMR titration experiments, in which
the macrocycle amide (e), aryl (d), and hydroquinone (h) protons
were monitored, gave a series of titration curves which were
subsequently analyzed using the WinEQNMR19 program, and
determined association constants are presented in Table 1. This
association constant is defined by the following equilibrium:
Initial evidence of pseudorotaxane formation came from the
observed shifts in the 1H NMR spectra of 2a and macrocycle 4
in a 1:1 mixture in acetone-d6 (Figure 5). A downfield shift in
the macrocycle amide protons and aryl proton is indicative of
anion complexation. In addition, upfield shifts of the amide
protons of 2a are postulated to be a result of polarization of the
chloride anion toward the hydrogen bond donors of the
macrocycle, which reduces the strength of hydrogen bonding
between the pyridinium cation thread and the anion. Further
evidence of the thread residing within the cavity is provided
Ka,exp
macrocycle + TBA‚X- y z macrocycle‚X- + TBA+
Binding stoichiometries were confirmed by Job plot analyses.
As expected, there is little difference in the anion binding
Figure 3. Structures of 3-5 and 6a-d.
9
2294 J. AM. CHEM. SOC. VOL. 127, NO. 7, 2005

Anion-Templated Assembly of Pseudorotaxanes
A R T I C L E S
expected considering the similarities between the two systems.21
This would also limit the possible existence of a 4‚Cl- complex
in solution, and hence K2 will be virtually nonexistent. The lack
of perturbations in the macrocyclic hydroquinone proton
environment upon addition of pyridinium hexafluorophosphate
thread, 2d, indicates that K3 is zero, which, again, is as expected
due to the lack of a suitable templating anion. Conclusive proof
for the ion-pair being the lone guest species in the pseudoro-
taxane assembly process comes from two important experimen-
tal observations. First, agreement between the association
constants obtained by analysis of the macrocyclic amide protons,
which are perturbed by the anion, and the macrocyclic hydro-
quinone protons, which undergo upfield perturbations as a result
of the thread residing within the cavity, indicate that no free
anion is complexed by the macrocycle.22 Second, the association
constants were found to be constant over a range of macrocycle
concentrations (1.0, 1.5, 2.0, and 2.5 × 10-3 M). If the observed
1H NMR shift was considered to be the weighted average of
the two complexes, 4‚Cl- and 4‚2a, then the experimentally
determined association constant can be considered to be
representative of the two species and is expressed by one of
the following two equations (Kipc, defined by eq 1, represents
the association constant for the ion-pair complex, e.g., 4‚2a):23
1
Figure 5. Observed perturbations in the aromatic region of the H NMR
of macrocycle 4 in acetone-d6 at 293 K upon addition of 1 equiv of thread
2a. Bottom: 4. Middle: 4‚2a. Top: 2a.
by significant upfield shifts in the macrocyclic hydroquinone
protons and a shift in the pyridinium aryl proton, highly
indicative of π-π donor-acceptor stacking interactions.20 It
was observed that these 1H NMR shift magnitudes were largest
for 2a and diminished in size, 2b > 2c. No shifts were observed
in the 1:1 mixtures of the macrocycles with thread 2d.
Analogous results were also obtained in the 1H NMR spectra
of a 1:1 mixture of the pyridinium nicotinamide thread 6a and
the macrocycles in deuterated acetone. Downfield shifts in the
macrocycle amide and aryl proton indicate the binding of an
anion within the cleft. Upfield shifts in the nicotinamide amide
proton provide further evidence of threading, as polarization of
the chloride anion toward the macrocycle perturbs the electronic
environment. Upfield shifts in the macrocycle hydroquinone
protons once more indicate the presence of π-π donor-
acceptor interactions in the complex and thus the thread residing
within the cavity. As with the pyridinium threads, the magnitude
of the observed shifts followed the trend Cl- > Br- > I-, with
no shifts being observed for 6d, the hexafluorophosphate thread.
A quantitative 1H NMR titration experiment for macrocycle
4 with pyridinium thread 2a in acetone-d6 at 293 K was carried
out in which the macrocyclic amide, aryl, and hydroquinone
protons were monitored and the data analyzed using the
WinEQNMR19 program. Initial assumptions led us to believe
that this association constant, Ka,exp, was for the following
equilibrium:
K11/2K
Ka,exp
)
2 + Kipc for K2[4] << 1
[2a]
and
(K1K2)1/2
([2a][8])1/2
Ka,exp
)
+ Kipc for K2[4] >> 1
The first term in both equations represents the fraction of
complex present in solution that is the macrocycle-anion
complex, 4‚Cl-, and the second is the fraction that is the ion-
pair complex, 4‚2a. The invariance of the association constant
with changing host and guest concentrations, therefore, indicates
that no 4‚Cl- complex is formed, and thus the experimentally
determined association constant is a true measure of the ion-
pair association.
Quantitative 1H NMR titration experiments were then carried
out for all of the pyridinium-based threads, monitoring both the
macrocyclic anion binding protons and the hydroquinone
protons, and the obtained binding curves (Figure 6) were then
analyzed using the WinEQNMR19 program (Table 2). It is
noteworthy that Table 2 shows pseudorotaxane formation is
Ka,exp
thread+‚X- + macrocycle y z thread+X-‚macrocycle (1)
It is important, however, to consider other possible equilibria
in solution, which could potentially be:
(19) Hynes, M. J. J. Chem. Soc., Dalton. Trans. 1993, 311-312.
(20) (a) Loeb, S. J.; Wisner, J. A. Chem. Commun. 2000, 845-846. (b) Allwood,
B. L.; Shahriari-Zavareh, H.; Stoddart, J. F.; Williams, D. J. Chem.
Commun. 1987, 1058-1061. (c) Kickham, J. E.; Loeb, S. J.; Murphy, S.
L. Chem.sEur. J. 1997, 3 (8), 1203-1213.
K1
\
thread+ + X- y
z thread+‚X-
(21) It is difficult to accurately determine such large binding constants; however,
a linear plot was obtained for shift perturbations on additions of guest up
to 1 equiv, at which point a plateau was observed and no perturbations
seen upon further additions of TBA chloride.
K2
macrocycle + X- y z macrocycle‚X-
\
(22) For titrations with 8 at initial concentrations of 1.0, 1.5, 2.0, and 2.5 ×
10-3 M at 1 equiv of guest 2a (from a 0.075 M solution) added, ∆δamide
0.42, 0.47, 0.51, 0.53; ∆δaryl ) 0.54, 0.61, 0.67, 0.69; and ∆δhydroquinone
)
)
K3
thread+ + macrocycle y z thread+‚macrocycle
\
-0.32, -0.37, -0.40, -0.417, respectively. Binding constants determined
for 8 at initial concentrations of 1.5, 2.0, and 2.5 × 10-3 M are Kamide
)
2450, 2460, and 2450 M-1, and Karyl and Khydroquinone determined values
are in agreement within 10% experimental error.
The large association constant, K1, for the pyridinium thread
with a chloride counteranion indicates that the concentration
of free anion in solution tends to zero. A binding constant of
>105 M-1 was also obtained upon titration of TBA chloride
with the pyridinium nicotinamide thread, 6d, as would be
(23) If macrocycle 8 also complexed Cl- in significant concentrations, then the
determined association constants would be concentration-dependent, and
a plot of 1/{[2a][8]}1/2 would give a straight line with an intercept equal to
the true ion-pair complexation constant. For full mathematical treatment,
see ref 15.
9
J. AM. CHEM. SOC. VOL. 127, NO. 7, 2005 2295

A R T I C L E S
Sambrook et al.
loss on pseudorotaxane formation will be greatest for the larger
macrocycle (vide infra). The smallest macrocycle could be said
to possess the greatest size complementarity match with the
pyridinium threads, which is also confirmed by the single-crystal
X-ray structural evidence (vide infra).
While anion templation is still strong enough for pseudoro-
taxane formation, the pyridinium nicotinamide threads associate
to a lesser extent with the macrocycles than do the pyridinium
threads. A possible reason for this is the loss of stability provided
by second-sphere pyridinium N+-CH3‚‚‚polyether hydrogen
bonds (see Figure 7 and crystal structures in Figures 10 and
13). The trend across the range of macrocycles matches that
observed for the pyridinium threads, which suggests that the
size complementarity requirements of the nicotinamide threads
do not differ significantly from those of the pyridinium
analogues, especially in respect to the aromatic stacking
interactions.
Figure 6. Example titration curves for threads 2a ([), 2b (9), and 2c (2)
with macrocycle 4 monitoring the ortho-aryl proton in acetone-d6 at 293 K
(initial host concentration, 1.5 × 10-3 M).
Table 2. Association Constants, Ka,exp (M-1), for Threads 2a-c
and 6a-c as Determined by 1H NMR Titrations in Acetone-d6 at
293 K (errors less than 10%)
thread
3
4
5
2a
6a
2b
6b
2c
6c
9500
1900
610
2400
950
700
240
65
980
320
120
180
<50a
50
200
<50a
65
<50a
a Data could not be fitted using WinEQNMR due to very weak binding.
Figure 8. Structures of 7a-d and 8a-d.
Imidazolium, Benzimidazolium, and Guanidinium Pseu-
dorotaxane Assemblies. Given the success of producing
interpenetrated structures with pyridinium-based threading
components, it was of interest to apply this anion templation
principle further to include other potential cationic threads, such
as imidazolium, benzimidazolium, and guanidinium. Imidazo-
lium compounds have recently gained a great deal of attention
due to many of their derivatives being room temperature ionic
liquids (RTILs),24 as well as readily accessible precursors to
N-heterocyclic carbenes.25 Furthermore, a variety of acyclic
imidazolium derivatives have been shown recently to recognize
anions,26 with chloride and bromide anions exhibiting a tem-
plation role in the synthesis of bisimidazolium macrocycles.27
Standard synthetic procedures were used to prepare the imida-
zolium and benzimidazolium salts (Figure 8).28
The guanidinium moiety is a well-established entity for anion
binding, possessing both a positive charge and suitable hydrogen
bond donors.29 Potential guanidinium threads, 11a-d, were
prepared as shown in Scheme 1. Reaction of n-hexylisothio-
cyanate with n-hexylamine gave the thiourea, 9, which was
methylated using iodomethane to give the thiouronium iodide
salt, 10. Substitution with methylamine gave the guanidinium
iodide salt, 11c, which was converted to the chloride, bromide,
templated preferentially by the chloride anion pyridinium salts
2a and 6a, with no association being observed for the PF6
-
threads, 2d and 6d. This trend is a result of the complementary
amide cleft size and geometry for chloride in preference to that
of the larger halides, highlighting the essential role that the anion
plays in the assembly process. In sharp contrast to the TBA
anion salt-binding experiments, there is a large variation in
binding strength across the macrocycles, with the smallest
macrocycle, 3, binding the pyridinium chloride threads, 2a and
6a, with a much greater affinity than for macrocycle 4 or 5.
The lowest association constants for chloride threading are
observed for both systems with the largest macrocycle, 5.
Possible reasons for the variation in association constants
between macrocycles can be arrived at by considering the
influence of the secondary interactions on the stability of the
interpenetrated complex. It can be postulated that the smallest
macrocycle maximizes both π-π stacking and hydrogen
bonding interactions with both pyridinium systems, and thus
association is strong. The largest macrocycle, therefore, reduces
the efficacy of both aromatic stacking and hydrogen bonding,
primarily as a result of the polyether chain being too long and
thus distorting the macrocycle geometry. In addition, the entropic
Figure 7. (a) Pyridinium-based pseudorotaxane assembly. (b) Pyridinium nicotinamide pseudorotaxane assembly (chloride anion represented by gray sphere).
9
2296 J. AM. CHEM. SOC. VOL. 127, NO. 7, 2005

Anion-Templated Assembly of Pseudorotaxanes
A R T I C L E S
Scheme 1
and hexafluorophosphate salts via the free guanidine base, 12.
differ significantly upon complexation with the free anion or
the ion-pair, and thus they provide little information on the
complexes present in solution.
Initial evidence for threading of imidazolium and benzimi-
1
dazolium pseudorotaxanes came from H NMR spectra of 1:1
mixtures of threads 7a and 8a with macrocycle 4 in acetone-
d6. Downfield shifts in the macrocycle amide and o-aryl protons
were observed along with upfield shifts in the CH proton in
the 2-position of the ion-pair, indicative of the anion being bound
within the amide cleft of the macrocycle. No perturbations of
The initial assumption was, again, that these Ka,exp determina-
tions were a direct measurement of the pseudorotaxane forma-
tion defined in eq 1. Interestingly though, the ion-pair association
measured by titration of TBA chloride into a solution of known
concentration of imidazolium hexafluorophosphate in acetone-
d6 at 293 K was determined to be 1800 M-1. This value is
significantly less than the ion-pairing strength of the pyridinium
threads with chloride; it is also lower than the binding affinities
displayed by the macrocycles for chloride anions. This led us
to question whether the imidazolium cation will “follow” the
anion into the macrocycle or simply compete for complexation
of the anion with the macrocycle. As with the pyridinium
systems, further equilibria and their association constants, K1
to K3, were also considered.
In this case, K1, determined to be 1800 M-1, indicates that
some “free” anion may be present in solution, and hence
association constant K2 may be significant and contribute to
the experimentally determined constant, Ka,exp. This relatively
smaller magnitude of ion-pair association is likely, despite the
strong electrostatic attraction between the two species, to stem
from the presence of only one C-H‚‚‚X- hydrogen bond, in
comparison to two amide NH‚‚‚X- and one CH‚‚‚X- hydrogen
bond in the pyridinium ligands.
1
the H NMR spectra were observed for 7d and 8d, with the
magnitude of observed shifts for threads a-c following the trend
Cl- > Br- > I-. Larger upfield shifts in the macrocycle
hydroquinone protons were observed for 7a over 8a (0.17 and
0.018 ppm, respectively), indicative of more significant π-π
stacking interactions. Despite the imidazolium threads having
a positive charge density greater than that of the benzimidazo-
lium analogues, the latter possess an increased aromatic surface
area, and this is thought to be the reason for the increase in the
role of π-π stacking interactions in the final complex.
1
As with the previous pseudorotaxane systems, a H NMR
titration was initially carried out with thread 7a and macrocycle
4 and an association constant, Ka,exp, determined as 1720 M-1
(acetone-d6, 293 K). This association constant was found to be
identical (within experimental error) for the binding curves
obtained by monitoring the amide and anion binding aryl proton.
It is unlikely, however, that these proton environments would
(24) (a) Davis, J. J.; Fox, P. A. Chem. Commun. 2003, 1209-1212. (b) Seddon,
K. R. J. Chem. Technol. Biotechnol. 1997, 68, 351-356. (c) Branco, L.
C.; Rosa, J. N.; Ramos, J. J. M.; Afonso, C. A. M. Chem.sEur J. 2002, 8
(16), 3671-3677. (d) Dzyuba, S. V.; Bartsch, R. A. Chem. Commun. 2001,
16, 1466-1467.
(25) (a) Arduengo, A. J.; Harlow, R. L.; Kline, M. J. Am. Chem. Soc. 1991,
113 (1), 361-363. (b) Arduengo, A. J.; Rasika Dias, H. V.; Harlow, R. J.;
Kline, M. J. Am. Chem. Soc. 1992, 114 (14), 5530-5534. (c) Herrmann,
W. A. Angew. Chem., Int. Ed. 2002, 41, 1290-1309.
Investigation of the association constant determined by
monitoring the shift perturbations of the macrocycle’s hydro-
quinone protons, h, gave an association constant of 500 M-1
for macrocycle 4 titrated with the potential imidazolium chloride
thread, 7a. These shift perturbations must be a result of the ion-
pair residing within the cavity, as the anion does not cause
significant perturbations to this proton environment (as dem-
onstrated with the TBA salt studies discussed previously) and
the cationic thread itself cannot associate without a suitable
anion, as shown by the lack of any proton shift changes with
the imidazolium hexafluorophosphate ion-pair.30
(26) (a) Alcalde, E.; Alvarez-Ru´a, C.; Garc´ıa-Granda, S.; Garc´ıa-Rodriguez, E.;
Mequida, N.; Pe´rez-Garc´ıa, L. Chem. Commun. 1999, 295-296. (b) Ihm,
H.; Yun, S.; Kim, H. G.; Kim, J. K.; Kim, K. S. Org. Lett. 2002, 4 (17),
2897-2900. (c) Sato, K.; Arai, S.; Yamagishi, T. Tetrahedron Lett. 1999,
40 (28), 5219-5222.
(27) Alcalde, E.; Ramos, S.; Pe´rez-Garc´ıa, L. Org. Lett. 1999, 1 (7), 1035-
1038.
(28) Dzyuba, S. V.; Bartsch, R. A. Chem. Commun. 2001, 1466-1467.
(29) (a) Best, M. D.; Tobey, S. L.; Anslyn, E. V. Coord. Chem. ReV. 2003,
240, 3-15. (b) Schmidtchen, F. P.; Berger, M. Chem. ReV. 1997, 97, 1609-
1646.
9
J. AM. CHEM. SOC. VOL. 127, NO. 7, 2005 2297

A R T I C L E S
Sambrook et al.
Table 3. Experimentally Determined Association Constants, Ka,exp
and Ion-Pair Association Constants, Kipc (eq 1), for Threads 7a-c
and 8a-c as Determined by 1H NMR Titrations in Acetone-d6 at
293 K, with Initial Host Concentration at 1.5 × 10-3 M and Guest
Feed Solution at 0.0375 M for Imidazolium and 0.01875 M for
Benzimidazolium
,
3
4
5
a
b
a
b
a
b
thread
Ka,exp
Kipc
Ka,exp
Kipc
Ka,exp
Kipc
7a
8a
7b
8b
7c
8c
1340
1630
380
390
50
400
320
190
240
c
1720
2600
480
500
30
500
250
300
230
c
1690
1200
520
370
60
230
200
c
110
c
40
80
30
90
50
c
a Estimated error ) 10%. b Estimated error ) 20%. c Shifts too small
to determine binding constant; estimate less than 20 M-1
Figure 9. Proposed orientation of imidazolium thread 7a within pseudoro-
taxane complex (chloride anion represented by gray sphere).
.
Further investigation of the 4‚7a complex formation in
solution was achieved by titration of the ion-pair guest species
into varying concentrations of the host molecule and the
association constant determined. It was found that the association
constants do vary with concentration, indicating the presence
of both 4‚7a and 4‚Cl- complexes in solution. Significantly,
however, the association constant determined by monitoring the
hydroquinone protons involved in π-π stacking interactions is
invariant with concentration, within experimental error limits.
Typical values for Ka,exp at macrocycle 4 concentrations of
1.5 × 10-3, 1.75 × 10-3, and 2.25 × 10-3 M are 1720, 1630,
and 1500 M-1, respectively, from monitoring the macrocycle
aryl proton, d. A plot of Ka,exp versus 1/{[7a][4]}1/2, or 1/
{[4]}1/2, gave a straight line, and the value of Kipc was
identical ion-pairing for the imidazolium and benzimidazolium
is unsurprising considering the similarity in methine proton pKa
values reported for related N,N′-methyl derivatives: imidazolium
) 23.0 and benzimidazolium ) 21.6 in water at 25 °C.33 Table
3 shows that relatively weaker pseudorotaxane complexes are
formed with the benzimidazolium threads. This finding, coupled
with similar Ka,exp measurements, indicates that the macrocycle-
anion complex is more predominant when the cationic thread
is benzimidazolium. This must arise from either less effective
secondary stabilizing interactions or from a greater entropic
deficit on binding. It is notable, however, that the association
constants for the benzimidazolium pseudorotaxanes can be
determined with greater accuracy due to the larger magnitude
of hydroquinone proton perturbations.
determined from the intercept to be 520 M-1 31
Importantly,
.
In contrast to the pyridinium chloride threads (Table 1),
altering the macrocycle polyether ring size has a less dramatic
affect on the association constant magnitudes of the imidazolium
chloride guest species. Neither the imidazolium nor benzimi-
dazolium threads possess the ability to hydrogen bond to the
polyether chain in the same manner as the pyridinium salts,
2a-d, due to a lack of suitable hydrogen bond donor function-
ality.34
A range of Ka,exp values were determined for the guanidinium
threads with macrocycles 3-5 in acetone-d6 at 293 K. Macro-
cyclic amide and aryl protons were monitored, and in all cases,
the determined values were considerably smaller in magnitude
than those of all of the previous pseudorotaxane assemblies
(Table 4). The strength of ion-pairing was again measured by
titration of TBA chloride into a solution of the hexafluorophos-
phate salt, 11d, in acetone-d6 at 293 K. Monitoring the NH
protons and analysis of the obtained titration curve gave a
binding constant of 1500 M-1. This is significantly less effective
than the ion-pairing in the pyridinium systems and, like the
imidazolium assemblies, implies that there will be a significant
concentration of macrocycle-anion complex present in solution.
In all of the previous pseudorotaxane-assembled systems,
evidence for the thread residing within the cavity was seen by
significant upfield perturbations in the macrocyclic hydro-
the value of Kipc determined via this method matches, within
experimental error, the value determined via monitoring the
macrocycle’s hydroquinone protons, h. Table 3 gives the
experimentally determined association constants from monitor-
ing the amide and aryl protons to give Ka,exp, that is, the
association constant representing the formation of both
macrocycle‚X- and macrocycle‚ion-pair complexes, and those
determined by monitoring the hydroquinone protons to give Kipc,
that is, the association constant for pseudorotaxane formation.
Table 3 shows that the most strongly associated pseudoro-
taxanes are found with the chloride imidazolium and benzimi-
dazolium salts, which mirrors the macrocycles’ preference for
this halide guest.32 In contrast to the pyridinium systems,
however, the magnitude of the chemical shift perturbations
caused by π-π stacking is significantly smaller, hence larger
1
errors will be present due to the sensitivity limit of H NMR.
Large perturbations of the imidazolium methine proton are
observed upon titration, indicating that the orientation of the
imidazolium threads within the macrocycle is as shown in Figure
9.
Ion-pairing in the benzimidazolium chloride thread 8a was
measured by titration of TBA chloride into a solution of thread
8d in acetone-d6 at 293 K and determined to be 1800 M-1. The
(30) Due to the small shift perturbations in the hydroquinone protons caused
by the imidazolium thread residing in the cavity, the errors for these
association constants are estimated to be (20% as a result of the sensitivity
limit of 1H NMR.
(33) Amyes, T. L.; Diver, S. T.; Richard, J. P.; Rivas, F. M.; Toth, K. J. Am.
Chem. Soc. 2004, 126 (13), 4366-4374.
(34) Attempts to further probe the imidazolium assembling motif by variable-
1
(31) Estimated errors on this value are (30%.
temperature H NMR proved fruitless, as even at -70 °C, the complexes
(32) It is interesting to note the presence of an as-yet undetermined 1:2 host:
guest binding contribution to the binding constants for the bromide and
iodide threads, which is not present when the amide protons or the aryl
proton is monitored. This contribution to the association is, although
noticeable by Job plot analyses, very small in magnitude.
and free species were found to still be in fast exchange. One noticeable
feature, however, was the larger downfield shift observed in a mixture of
8:12a (1.5 × 10-3:3 × 10-3 M-1) upon cooling, indicating that the nature
of the complexation processes, as expected, is exothermic (∆δamide ) 1.17,
1.21, 1.25, and 1.27 ppm at 20, 10, 0, and -10 °C, respectively).
9
2298 J. AM. CHEM. SOC. VOL. 127, NO. 7, 2005

Anion-Templated Assembly of Pseudorotaxanes
A R T I C L E S
K-1 for 2a‚3. Unfavorable entropic contributions in both systems
arise from the loss of rotational and translational degrees of
freedom upon pseudorotaxane association. Taking into account
the conclusions drawn from the single-crystal X-ray diffraction
evidence (vide infra), it is likely that increasing the length of
the polyether chain results in a greater entropic deficit upon
binding in order to maintain optimum anion recognition and
π-π stacking interactions.
Table 4. Experimentally Determined Association Constants, Ka,exp
(M-1), for Threads 11a-c as determined by 1H NMR Titrations in
Acetone-d6 at 293 K (errors less than 10%)a
thread
3
4
5
11a
11b
11c
390
310
100
340
320
125
400
370
90
a Ka,exp represents formation of both macrocycle-anion and macrocycle-
ion-pair complexes.
Table 5. Isothermal Calorimetric Data for Complexes 2a‚3 and
2a‚4 Determined in 1,2-Dichloroethane at 298 Ka
1
1
1
thread
macrocycle
n
K
a (M-
)
∆
H (cal mol-
)
∆
S (cal K-
)
2a
2a
3
4
1
1
3.58 × 104
-1.186 × 104
-18.9
-295
3.06 × 105
-9.45 × 104
a Errors < 10%.
quinone protons, and these perturbations could also be used to
determine association constants. In the guanidinium systems,
there is no electron-deficient aromatic surface within the thread,
and hence no π-π donor-acceptor interactions are present in
the complex. Proof for the threading process was therefore
obtained by undertaking intermolecular 1H NMR NOE experi-
ments. Saturation of the macrocyclic hydroquinone proton
environment, i, resulted in a significant increase in the intensity
of both guanidinium N+-methyl and NHCH2 protons. These
guanidinium proton peak intensity perturbations arise from the
proximity of the guanidinium thread residing within the cavity
of the macrocycle.
Figure 10. Solid-state structure of 2a‚3. Left: view in the plane of the
macrocyclic ring. Right: view through the annulus of the macrocyclic ring.
Chloride anion represented as CPK sphere for clarity.
The lack of π-π donor-acceptor interactions in the pseu-
dorotaxane complex results in an absence of secondary binding
site with which to accurately measure the macrocycle-ion-pair
association constant, Kipc, and it becomes extremely difficult
to accurately determine these values. Of significance though,
is the small Ka,exp values themselves, indicating that, although
pseudorotaxane formation is occurring, it is a less effective
assembling motif than the previously discussed pyridinium- and
imidazolium-based systems.
Calorimetric Data. Isothermal calorimetry (ITC) allows the
determination of ∆G0, ∆H0, and ∆S0 for a ligand-receptor
interaction by the measurement of stepwise changes in the
enthalpy of interaction during the course of a single titration
experiment.35
Figure 11. Solid-state structure of 2a‚5. Left: view in the plane of the
macrocyclic ring. Right: view through the annulus of the macrocyclic ring.
Chloride anion represented as CPK sphere for clarity.
Binding constants for the pseudorotaxane systems in acetone
1
were thought, from studying the H NMR titration data, to be
too low for ITC determination. A less competitive solvent
system, 1,2-dichloroethane, was thus chosen and titrations were
carried out (Table 5).
The association constants determined by this method, in
1
agreement with the H NMR titration data (Table 2), show a
1:1 stoichiometry for pseudorotaxane formation. As expected,
the assembly process is enthalpically driven, with pseudoro-
taxane complex 2a‚4 being more enthalpically favored. How-
ever, it is noteworthy that there is a greater entropic deficit upon
binding for 2a‚4, -295 cal K-1, in comparison to -18.9 cal
Figure 12. Solid-state structure of 2b‚3. Left: view in the plane of the
macrocyclic ring. Right: view through the annulus of the macrocyclic ring.
Bromide anion represented as CPK sphere for clarity.
(35) (a) Turnbull, W. B.; Daranas, A. H. J. Am. Chem. Soc. 2003, 125 (48),
14859-14866. (b) Stulz, E.; Scott, S. M.; Bond, A. D.; Otto, S.; Sanders,
J. K. M. Inorg. Chem. 2003, 42, 3086-3096. (c) Weber, P. C.; Salemme,
F. R. Curr. Opin. Struct. Biol. 2003, 13, 115-121. (d) Leavitt, S.; Freire,
E. Curr. Opin. Struct. Biol. 2001, 11, 560-566. (e) Rakharsky, M. V.;
Inoue, Y. Chem. ReV. 1998, 98, 1875-1917. (f) Izatt, R. M.; Terry, R. E.;
Haymore, B. L.; Hansen, L. D.; Dalley, N. K.; Avondet, A. G.; Christensen,
J. J. J. Am. Chem. Soc. 1976, 98, 7620-7630.
Solid-State Evidence of Pseudorotaxane Formation. Fur-
ther evidence of pseudorotaxane assembly comes from the solid
state. Pale yellow crystals of macrocycle 7 with threads 2a, 2b,
9
J. AM. CHEM. SOC. VOL. 127, NO. 7, 2005 2299

A R T I C L E S
Sambrook et al.
anion geometry
Table 6. Selected Crystallographic Data
complex
N
−H‚‚‚
X
distance (Å)
angle (deg)
2a‚3
N(1)‚‚‚Cl(1)
N(2)‚‚‚Cl(1)
N(4)‚‚‚Cl(1)
N(5)‚‚‚Cl(1)
N(1)‚‚‚Cl(1)
3.383(3)
3.396(3)
3.386(4)
3.359(3)
3.349(2)
165.0(4)
172.0(4)
170.0(4)
157.0(6)
168.0(4)
octahedron
2a‚5
distorted trigonal
bipyramid
N(2)‚‚‚Cl(1)
N(4)‚‚‚Cl(1)
N(5)‚‚‚Cl(1)
N(1)‚‚‚Br(1)
N(2)‚‚‚Br(1)
N(4)‚‚‚Br(1)
N(5)‚‚‚Br(1)
N(1)‚‚‚Cl(1)
N(2)‚‚‚Cl(1)
N(3)‚‚‚Cl(1)
3.376(2)
3.400(2)
3.303(2)
3.490(5)
3.365(5)
3.424(6)
3.510(8)
3.345(2)
3.385(2)
3.302(2)
176.0(5)
165.0(6)
140.0(7)
171.0(2)
160.0(2)
171.0(2)
173.0(2)
169.0(2)
164.0(2)
169.0(3)
2b‚3
distorted tetrahedron
octahedron
12a‚4
Figure 13. Solid-state structure of 12a‚4. Left: view in the plane of the
macrocyclic ring. Right: view through the annulus of the macrocyclic ring.
Chloride anion represented as CPK sphere and hydrogens omitted for clarity.
and 6a and of macrocycle 5 with thread 2a were grown by slow
diffusion of diisopropyl ether into a 1:1 solution of the two
components in acetone, chloroform, or 1,2-dichloromethane. 1H
NMR spectra of these crystals dissolved in acetone-d6 are
identical to that of a 1:1 mixture of the separate compounds. In
all cases, single-crystal X-ray analysis reveals the ion-pair is
threaded through the cavity of the macrocycle in the manner
intended. In structure 2a‚3 (Figure 10), the chloride anion is
observed to have an octahedral geometry and engages in six
hydrogen bonds with the hydrogen bond donor clefts of the
macrocycle and the thread, as expected from solution evidence.
Four short N-H‚‚‚Cl- hydrogen bond contacts are observed,
and NH‚‚‚Cl distances are reported in Table 6. The pyridinium
methyl group sits out of the plane of the polyether ring but still
engages in N+CH3‚‚‚O hydrogen bonding, with a typical
distance of 3.370 Å. The hydroquinone rings are approximately
parallel with respect to each other (within 6°) at a distance of
∼7 Å and sandwich the central pyridinium ring, indicative of
aromatic stacking interactions.
The crystal structure of thread 2b with the larger macrocycle,
5 (Figure 11), reveals that the chloride anion appears to possess
distorted tetrahedral geometry and hydrogen bonds to the two
molecular components through four short N-H‚‚‚Cl contacts
(Table 6). Hydrogen bonding between the pyridinium methyl
group and the polyether chain is significantly shorter (2.987-
3.175 Å) than in the previous structure as the methyl group
appears to sit within the polyether chain. The hydroquinone
groups interact with the pyridinium ring in the same manner as
in 2a‚3, indicating that optimal aromatic stacking interactions
are permitted despite the larger macrocycle size. This is likely
to be due to the large flexibility inherent in the pentaethylene
chain and results in a greater entropic loss upon complexation.
The structure of the bromide-templated pseudorotaxane
complex, 2b‚3 (Figure 12), shows the bromide anion possesses
distorted trigonal bipyramidal geometry with the four N-H‚‚
‚Br- bonding distances being noticeably longer than those in
the analogous chloride complex, 2a‚3. The methyl group again
resides outside of the macrocycle, and observed hydrogen
bonding is in the range of 3.383-3.446 Å.
nicotinamide ring is sandwiched between the two parallel
hydroquinone units, and the lack of a suitable hydrogen donor
in the correct position on the thread means that the polyether
chain plays no role in further stabilization of the pseudorotaxane
complex.
Table 6 contains N-H‚‚‚X- (X ) Cl or Br) hydrogen bond
lengths and the geometry of the templating anion within the
pseudorotaxane complex for the four structures discussed.
Conclusions
A general anion templation procedure for assembling a wide
range of pseudorotaxanes which contain positively charged
pyridinium, pyridinium nicotinamide, imidazolium, benzimi-
dazolium and guanidinium threading components, and macro-
cyclic isophthalamide polyether ligands has been developed. The
templation strategy is based on coupling halide anion recognition
with ion-pairing. In noncompetitive solvent media, a halide
anion of a potential cationic threading component continues to
be coordinatively unsaturated. Anion complexation by a mac-
rocyclic ligand results in the interpenetration of the strongly
associated cationic component of the ion-pair leading to a
1
pseudorotaxane assembly. Extensive H NMR titration inves-
tigations reveal that the thermodynamic stability of the pseu-
dorotaxane assembly depends critically on the nature of the
halide anion template, with chloride proving to be the optimum
template anion, and also, importantly, on the strength of the
ion-pairing between the anion template and the cationic thread.
This ion-pair association strength is dictated by the character
of the positively charged threading component. In the case of
pyridinium-based systems, ion-pairing is extremely strong,
whereas for imidazolium and guanidinium threads, weaker ion-
pair association results in free anion-macrocycle complexation
competing with pseudorotaxane formation. In addition, the
efficacy of favorable second-sphere π-π aromatic stacking
interactions between the positively charged pyridinium threading
component and the hydroquinone units of the respective
macrocyclic ligand, together with favorable second-sphere
pyridinium N+-CH3‚‚‚polyether macrocycle hydrogen bonding
interactions, is dictated by macrocyclic polyether ring size. In
contrast, imidazolium and guanidinium pseudorotaxane stability
is not affected significantly by macrocycle ring size. Preliminary
isothermal calorimetric studies indicate, as expected, that
pseudorotaxane assembly is an enthalpically driven process,
opposed by entropy. Single-crystal X-ray structural determina-
The interpenetrated nature of the pyridinium nicotinamide
threads with the macrocycle is demonstrated in the solid state
by a structure of the chloride thread 6a residing within the cavity
of macrocycle 4 (Figure 13). In this system, only three N-H‚
‚‚Cl hydrogen bonds are possible; however, the chloride anion
still retains an octahedral geometrical preference (Table 6). The
9
2300 J. AM. CHEM. SOC. VOL. 127, NO. 7, 2005

Anion-Templated Assembly of Pseudorotaxanes
A R T I C L E S
14.0. ESI MS [C14H32N3 - Cl]+: m/z observed 242.2598, calcd
242.2596. Elemental analysis calcd for C14H32ClN3: C, 60.51; H, 11.61;
N, 15.12. Found: C, 59.34; H, 11.86; N, 15.31.
tions reveal solid-state evidence of pseudorotaxane formation,
highlighting the role of anion recognition, aromatic π-π
stacking, and hydrogen bonding interactions in the stabilization
of the interpenetrated supramolecular assembly. Efforts are
currently being made to exploit this anion templation strategy
to produce related rotaxane36 and catenane37 derivatives for
anion sensing and molecular machinelike behavior.38
N,N′-Di(1-hexyl)-N′′-methylguanidine Hydrobromide (17b). To
a solution of guanidine (18) (0.6 g, 2.5 mmol) in methanol (20 mL)
was added NH4Br (0.24 g, 2.5 mmol), and the solution was refluxed
for 2 h. Evaporation of solvent resulted in an oily residue which was
dried in vacuo to give the title compound in quantitative yield as pale-
yellow waxy material (0.81 g). 1H NMR (300 MHz, CDCl3, TMS): δ
7.28 (q, J ) 4.8 Hz, 1H, dN+HCH3), 6.54 (t, J ) 5.1 Hz, 2H, CH2-
CH2NHC), 3.36 (q, J ) 6.0 Hz, 4H, CH2CH2NHC), 2.98 (d, J ) 4.8
Hz, 3H, dN+HCH3), 1.59 (qi, J ) 7.2 Hz, CH2CH2NHC), 1.36-1.24
(m, 12H, CH3(CH2)3CH2CH2NHC), 0.88 (t, J ) 6.6 Hz, 6H, CH2CH3).
13C NMR (300 MHz, CDCl3, TMS): δ 155.5, 42.3, 31.3, 29.1, 29.0,
26.4, 22.5, 14.0. ESI MS [C14H32N3 - Br]+: m/z observed 242.2596
(100), calcd 242.2596. Elemental analysis calcd for C14H32BrN3: C,
52.17; H, 10.01; N, 13.04. Found: C, 52.12; H, 10.17; N, 12.80.
N,N′-Di(1-hexyl)-N′′-methylguanidine Hydroiodide (17c). A solu-
tion of (S)-methyl derivative (16) (6.3 g, 16.3 mmol) in ethanol (40
mL) containing methylamine (6.5 g, 210 mmol) was stirred at 42-45
°C for 3 h and at 70 °C for 2 h. After evaporation of the volatile parts,
the oily residue was dried in vacuo to give the title compound in
Experimental Section
General Methods. Chemicals were purchased from Aldrich and used
as received with the exception of thionyl chloride which was distilled
prior to use according to standard laboratory procedures. Deionized
water was used in all cases. All reactions were carried out under an
atmosphere of dry nitrogen, unless otherwise stated. 1H and 13C NMR
were recorded on Varian 300 MHz Mercury VX Works and 500 MHz
Unity Plus spectrometers. Mass spectrometry was carried out using a
Micromass LCT electrospray mass spectrometer. Elemental analyses
were performed by the Inorganic Chemistry Laboratory Microanalysis
service. All melting point determinations are uncorrected.
N,N′-Di(1-hexyl)thiourea (15). To a solution of n-hexyl isothiocy-
anate (2.86 g, 20 mmol) in THF (50 mL) was added n-hexylamine
(2.23 g, 22 mmol), and the reaction mixture was stirred under reflux
for 3 h. The volume of the mixture was reduced (10 mL), and hexane
(50 mL) was added. Finally, the solution was concentrated to 35 mL,
and the product crystallized at -20 °C overnight. The white crystalline
product was separated, washed with cold hexane, and dried in vacuo
to give the title compound in 94% yield (4.6 g) as white powder. Mp
34-38 °C. 1H NMR (300 MHz, CDCl3, TMS): δ 5.80 (br s, 2H, CH2-
CH2NHCdS), 3.38 (br s, 4H, CH2CH2NHCdS), 1.58 (qi, J ) 6.1 Hz,
4H, CH2CH2NHCdS), 1.38-1.27 (m, 12H, CH3(CH2)3CH2CH2NH),
0.87 (t, J ) 6.7 Hz, 6H, CH3CH2). 13C NMR (CDCl3): δ 181.4, 44.4,
31.4, 29.0, 26.6, 22.5, 14.0. FI MS [C13H28N2S]+: m/z observed
244.1975 (100), calcd 244.1973. Elemental analysis calcd for
C13H28N2S: C, 63.88; H, 11.55; N, 11.46. Found: C, 64.04; H, 11.64;
N, 11.66.
1
quantitative yield as colorless viscous oil (6 g). H NMR (300 MHz,
CDCl3, TMS): δ 6.91 (q, J ) 4.9 Hz, 1H, dN+HCH3), 6.23 (t, J )
5.4 Hz, 2H, CH2CH2NHC), 3.43 (q, J ) 6.6 Hz, 4H, CH2CH2NHC),
3.05 (d, J ) 4.6 Hz, 3H, dN+HCH3), 1.66 (qi, J ) 7.6 Hz, CH2CH2-
NHC), 1.36-1.26 (m, 12H, CH3(CH2)3CH2CH2NHC), 0.86 (t, J ) 6.8
Hz, 6H, CH2CH3). 13C NMR (300 MHz, CDCl3, TMS): δ 154.7, 42.7,
31.4, 29.3, 29.0, 26.3, 22.5, 14.0. ESI MS [C14H32N3 - I]+: m/z
observed 242.2587, calcd 242.2596. Elemental analysis calcd for C14H32-
IN3: C, 45.53; H, 8.73; N, 11.38. Found: C, 46.06; H, 8.64; N, 11.51.
N,N′-Di(1-hexyl)-N′′-methylguanidine Hexafluorophosphate (17d).
Obtained from guanidine (18) (0.48 g, 2 mmol) and NH4PF6 (0.33 g,
2 mmol), employing the above procedure, 17d resulted in quantitative
1
yield as pale yellow oil (0.77 g). H NMR (300 MHz, CDCl3, TMS):
δ 5.66 (q, J ) 5.0 Hz, 1H, dN+HCH3), 5.29 (t, J ) 5.3 Hz, 2H, CH2-
CH2NHC), 3.22 (q, J ) 5.6 Hz, 4H, CH2CH2NHC), 3.93 (d, J ) 4.8
Hz, 3H, dN+HCH3), 1.65 (qi, J ) 6.5 Hz, CH2CH2NHC), 1.40-1.28
(m, 12H, CH3(CH2)3CH2CH2NHC), 0.89 (t, J ) 6.6 Hz, 6H, CH2CH3).
13C NMR (300 MHz, CDCl3, TMS): δ 155.0, 41.8, 31.2, 28.4, 27.8,
26.1, 22.4, 13.9. ESI MS [C14H32N3 - PF6]+: m/z observed 242.2586
N,N′-Di(1-hexyl)-(S)-methylisothiouronium Iodide (16). To a
solution of thiourea (15) (4.15 g, 17 mmol) in ethanol (40 mL) was
added iodomethane (4.82 g, 34 mmol), and the mixture was stirred at
40-45 °C for 4 h and at 70 °C for 1 h. Evaporation of the solvent and
drying of the oily residue in vacuo gave the title compound in 98%
1
yield as colorless oil (6.4 g). H NMR (300 MHz, CDCl3, TMS): δ
(100), calcd for C14H32N3 (M - PF6)+ 242.2596. Elemental analysis
+
8.55 and 7.94 (2 × br s, 2H, CH2CH2N+HdC(SCH3)NHCH2CH2), 3.76
and 3.46 (2 × br s, 4H, CH2CH2N+HdC(SCH3)NHCH2CH2), 2.86 (s,
3H, SCH3), 1.71 (br s, 4H, CH2CH2NH), 1.36-1.26 (m, 12H, CH3-
(CH2)3CH2CH2NH), 0.84 (t, J ) 6.0 Hz, 6H, CH3CH2). 13C NMR (300
MHz, CDCl3, TMS): δ 45.3, 31.3, 29.4, 26.4, 22.5, 16.3, 14.1. ESI
MS [C14H31N2S - I]+: m/z observed 259.2199, calcd 259.2208.
Elemental analysis calcd for C14H31IN2S: C, 43.52; H, 8.09; N, 7.25.
Found: C, 44.17; H, 8.09; N, 7.43.
calcd for C14H32F6N3P: C, 43.41; H, 8.33; N, 10.85. Found: C, 44.06;
H, 9.04; N, 11.02.
N,N′-Di(1-hexyl)-N′′-methylguanidine (18). A solution of guanidine
hydroiodide (17c) (0.96 g, 2.6 mmol) in CH2Cl2 (35 mL) was vigorously
stirred with 6 M sodium hydroxide solution (6 mL) for 30 min. The
organic layer was separated, dried over K2CO3, evaporated, and the
residue dried in vacuo to give the free base in quantitative yield as
1
colorless viscous oil (0.63 g). H NMR (300 MHz, DMSO-d6, TMS):
N,N′-Di(1-hexyl)-N′′-methylguanidine Hydrochloride (17a). To
a solution of guanidine (18) (0.6 g, 2.5 mmol) in methanol (20 mL)
was added NH4Cl (0.14 g, 2.5 mmol), and the solution was refluxed
for 2 h. Evaporation of solvent resulted in an oily residue which was
dried in vacuo to give the title compound in quantitative yield as
δ 4.82 (br s, 2H, CH2CH2NHC), 2.87 (t, J ) 6.8 Hz, 4H, CH2CH2-
NHC), 2.53 (s, 3H, dNCH3), 1.42-1.26 (m, 16H, CH3(CH2)4CH2-
NHC), 0.84 (t, J ) 6.3 Hz, 6H, CH2CH3). 13C NMR (300 MHz, DMSO-
d6, TMS): δ 152.9, 43.0, 31.7, 30.8, 30.3, 26.9, 22.6, 14.4. ESI MS
[C14H31N3 + H]+: m/z observed 242.258, calcd 242.2596.
Typical 1H NMR Titration Procedure. All titrations were per-
formed with the starting concentration of hosts 7, 8, and 9 at 1.5 ×
10-3 M and the addition of appropriate aliquots of titrant (0.2, 0.4-
1.8, 2.0, 2.5, 3.0, 4.0, 5.0, 7.0, and 10.0 equiv) to the NMR sample
with a 10, 25, or 50 µl microsyringe as appropriate. The macrocycle
aryl proton d was followed during the course of the titration, and the
data were fitted and analyzed to give association constants using the
WinEQNMR program.21 Following the macrocycle amide protons e
gave association constants in agreement within experimental error.
Unless stated otherwise, all titration curves fit 1:1 binding models and
estimated errors were less than 10%.
1
colorless solid (0.69 g). H NMR (300 MHz, CDCl3, TMS): δ 7.80
(q, J ) 4.7 Hz, 1H, dN+HCH3), 6.95 (t, J ) 5.3 Hz, 2H, CH2CH2NHC),
3.41 (q, J ) 5.8 Hz, 4H, CH2CH2NHC), 3.02 (d, J ) 4.8 Hz, 3H,
dN+HCH3), 1.65 (qi, J ) 7.2 Hz, CH2CH2NHC), 1.39-1.26 (m, 12H,
CH3(CH2)3CH2CH2NHC), 0.88 (t, J ) 6.6 Hz, 6H, CH2CH3). 13C NMR
(300 MHz, CDCl3, TMS): δ 155.7, 42.5, 31.4, 29.1, 29.0, 26.4, 22.5,
(36) Wisner, J. A.; Beer, P. D.; Drew, M. G. B.; Sambrook, M. R. J. Am. Chem.
Soc. 2002, 124 (42), 12469-12476.
(37) Sambrook, M. R.; Beer, P. D.; Wisner, J. A.; Paul, R. L.; Cowley, A. J.
Am. Chem. Soc. 2004, 126, 15364-15365.
(38) Keaveney, C. M.; Leigh, D. A. Angew Chem., Int. Ed. 2004, 43 (10), 1222-
1224.
9
J. AM. CHEM. SOC. VOL. 127, NO. 7, 2005 2301

A R T I C L E S
Sambrook et al.
Calorimetry. All binding experiments were performed in anhydrous
1,2-dichloroethane purchased from Aldrich and stored over activated
3 Å molecular sieves. The solvent was ultrasonically degassed under
vacuum before use. All binding experiments were performed on a VP-
ITC Isothermal Calorimeter from Microcal Inc. (Northampton, MA),
with a stirring rate of 300 rpm, a constant temperature of 25 °C, and
50-70 injections of 3-5 µL per injection. Experiments were run with
cell concentrations of the macrocycle in the range of 0.05-5 mmol
and ion-pair concentrations in the titration syringe at 10 times the
concentration used in the cell. The range used was necessary to generate
data over a wide range of Ka values observed that would give acceptable
curve fitting during data analysis. Dilution effects for all titrations were
determined by a second experiment adding the same ion-pair solution
into a pure 1,2-dichloroethane solution and subtracting the result from
the raw titration to produce the final binding curve. Association
constants and thermodynamic parameters were obtained in all cases
by fitting the final binding curve using a one-site model in the Origin
7 software supplied with the instrument. Curves were fitted by variation
of all three experimentally determined parameters (n, Ka, ∆H) and gave
excellent agreement to the data for all titrations. Note that in all curves,
the first three data points were removed due to systematic errors
generated by loading of the syringe into the cell.
X-ray Crystallography. Crystal data for 2a‚3‚0.8H2O: M )
1005.06, monoclinic, space group P121/c1, a ) 10.871(2), b )
15.8369(2), c ) 33.7208(3), R ) 90, â ) 90.7775(3), γ ) 90°, U )
5808.73(14) Å3, Z ) 4, R ) 0.1344, wR ) 0.1022. Crystal data for
2a‚5: M ) 1078.79, monoclinic, space group P121/n1, a ) 16.0174,
b ) 10.3218(1), c ) 35.6876(3), R ) 90, â ) 100.8331(4), γ ) 90°,
U ) 795.03(8) Å3, Z ) 4, R ) 0.0425, wR ) 0.0471. Crystal data for
2b‚3: M ) 1035.13, monoclinic, space group P121/c1, a ) 10.88110(10),
b ) 16.0952(2), c ) 33.5237(4), R ) 90, â ) 90.6073(4), γ ) 90°, U
) 5870.79(11) Å3, Z ) 4, R ) 0.1071, wR ) 0.1404. Crystal data for
12a‚4: M ) 977.68, monoclinic, space group P121/n1, a ) 11.03740-
(10), b ) 21.5071(3), c ) 22.7595(3), R ) 90, â ) 100.9476(5), γ )
90°, U ) 5304.38(11) Å3, Z ) 4, R ) 0.0426, wR ) 0.0551. Single
crystals of 2a‚3, 2b‚3, 2a‚5, and 12a‚4 were coated with perfluoropoly-
ether oil, mounted on a glass fiber, and cooled rapidly to 150 K in a
stream of cold N2 using an Oxford Cryosystems CRYOSTREAM unit.
Diffraction data were measured using an Enraf-Nonius KappaCCD
diffractometer (graphite monochromated Mo KR radiation, λ ) 0.71073
Å). Intensity data were processed using the DENZO-SMN package.39
The structure was solved by direct methods using the SIR92 program.40
Full-matrix least-squares refinement was carried out using the CRYS-
TALS program suite.41 The N-H atoms were located in the difference
Fourier maps and the coordinates and isotropic thermal parameters
subsequently refined. All other hydrogen atoms were positioned
geometrically after each cycle of refinement. A Chebychev polynomial
weighting scheme was applied. In compounds 2b‚3 and 12a‚4, disorder
was encountered in the hexyl chains of the thread. The disorder was
modeled appropriately. Crystallographic data for the compounds are
listed in the Supporting Information.
Acknowledgment. We thank EPSRC for postdoctoral fel-
lowships (J.A.W. and F.S.) and a studentship (M.R.S.). Financial
support for Dr. Wisner from the Natural Sciences and Engineer-
ing Research Council of Canada is also acknowledged.
Supporting Information Available: Synthetic procedures and
characterization data for threads 2a-d, macrocycles 3-5, and
threads 6a-d. Tables giving the crystal data and structure
refinement information, bond lengths and bond angles, atomic
and hydrogen coordinates, and isotropic and anisotropic dis-
placement coordinates for structures 2a‚7, 2a‚9, 2a‚7, and 12a‚
8. Figure demonstrating the 1H NMR intermolecular NOE
experiment carried out for complex 17a‚8. This material is
available free of charge via the Internet at http://pubs.acs.org.
JA046278Z
(39) Otwinowski, Z.; Minor, W. Processing of X-ray Diffraction Data Collected
in Oscillation Mode. Methods Enzymol. 1997, 276.
(40) Alotmare, A.; Cascarano, G.; Giacovazzo, G.; Guagliardi, A.; Burla, M.
C.; Polidori, G.; Camalli, M. J. Appl. Crystallogr. 1994, 27 (3), 435-435.
(41) Watkin, D. J.; Prout, C. K.; Carruthers, J. R.; Betteridge, P. W.; Cooper,
R. I. CRYSTALS, issue 11; Chemical Crystallography Laboratory: Oxford,
U.K., 2001.
9
2302 J. AM. CHEM. SOC. VOL. 127, NO. 7, 2005