Anion-templated rotaxane formation

Article abstract of DOI:10.1021/ja027519a

The development of an acyclic chloride anion template in which the chloride anion is coordinatively unsaturated and available for subsequent complexation to various hydrogen bond donating components is described. This template orients a neutral hydrogen b

Full text of DOI:10.1021/ja027519a

Published on Web 09/24/2002
Anion-Templated Rotaxane Formation
James A. Wisner,† Paul D. Beer,*,† Michael G. B. Drew,‡ and Mark R. Sambrook†
Contribution from the Department of Chemistry, Inorganic Chemistry Laboratory,
UniVersity of Oxford, Oxford, OX1 3QR, U.K., and Department of Chemistry,
UniVersity of Reading, Reading, RG6 6AD, U.K.
Received July 1, 2002
Abstract: The development of an acyclic chloride anion template in which the chloride anion is coordinatively
unsaturated and available for subsequent complexation to various hydrogen bond donating components is
described. This template orients a neutral hydrogen bond donating ligand and a pyridinium cation
orthogonally to one another. Incorporation of second-sphere interactions between the ligand and the
pyridinium cation improved the efficacy of the chloride template. These results were exploited in the
construction of a chloride anion-templated [2]rotaxane which, after anion template removal, was studied
with regards to its anion recognition properties. Encirclement of the neutral macrocycle around the dumbbell-
shaped pyridinium cation in the [2]rotaxane produced a dramatic increase in its selectivity for chloride
anions as compared to the noninterlocked cation. This is interpreted as a function of the anion template
used to create the [2]rotaxane superstructure.
Introduction
templation accumulating in the literature, demonstrating the
efficacy of such an approach. Examples include Hawthorne’s
The synthesis of macrocyclic compounds and supramolecular
arrays in recent years has been facilitated largely by the use of
template methodology to their construction.1 To date, most of
these schemes depend on the use of neutral or cationic species
to achieve the desired template effect, utilizing a variety of
noncovalent forces such as metal-ligand coordination, hydrogen
bonding, π-π stacking, and solvophobic interactions.1,2 The
absence of anionic species as more general participants in this
arena is often attributed to the relatively small ratio of charge
to radius and sizable solvation energy as compared to cations.3
However, there are a growing number of reports of anion
mercuracarborands, Lehn’s circular double helicates, Fujita’s
coordination nanotubes, Mingos and Vilar’s amidinothiourea
cages, Ward and McCleverty’s tetrahedral Co2+ boxes, and
Dunbar’s tetranickel box, all of which depend on particular
anionic species for their formation.4
It is within this context that there are very few instances of
template synthesis employing anions in the construction of
interpenetrated or mechanically bonded compounds (termed
pseudorotaxanes, rotaxanes, catenanes, and knots). Stoddart et
al. have published solid-state evidence for hexafluorophosphate
anion templation in the formation of a [2]pseudorotaxane from
dibenzylammonium threads and a polyether macrocycle.5 In a
similar system, Montalti and Prodi determined that threading
of a dialkylammonium axle through a dibenzo-24-crown-8 wheel
to form a [2]pseudorotaxane could be controlled by the
interaction of the cation with chloride anions in solution.6 The
only examples of anion-templated formation of interlocked
species are those developed by Vo¨gtle et al., and studied by
others, in which phenolates, thiophenolates, and sulfonamide
anions are noncovalently bound to a cyclic tetralactam.7,8
Subsequent covalent trapping of these intermediates by a
* To whom correspondence should be addressed. Fax: (+44) 1865
27290. E-mail: paul.beer@chem.ox.ac.uk.
† University of Oxford.
‡ University of Reading.
(1) (a) Raymo, F. M.; Stoddart, J. F. Templated Organic Synthesis; Wiley:
Weinheim, Germany, 2000. (b) Hubin, T. J.; Busch, D. H. Coord. Chem.
ReV. 2000, 200-202, 5. (c) ComprehensiVe Supramolecular Chemistry;
Sauvage, J.-P., Hosseini, M., Eds.; Elsevier: New York, 1996; Vol. 9.
(2) (a) Molecular Catenanes, Rotaxanes, and Knots; Sauvage, J.-P., Dietrich-
Buchecker, C., Eds.; Wiley-VCH: Weinheim, Germany, 1999. (b) Stoddart,
J. F.; Nepogodiev, S. A. Chem. ReV. 1998, 98, 1959. (c) Vo¨gtle, F.;
Duennwald, T.; Schmidt, T. Acc. Chem. Res. 1996, 29, 451. (d) Seel, C.;
Parham, A. H.; Safarowsky, O.; Huebner, G. M.; Vo¨gtle, F. J. Org. Chem.
1999, 64, 7236. (e) Hunter, C. A. J. Am. Chem. Soc. 1992, 114, 5303. (f)
Dietrich-Buchecker, C. O.; Sauvage, J. P.; Kern, J. M. J. Am. Chem. Soc.
1984, 106, 3043. (g) Leigh, D. A.; Lusby, P. J.; Teat, S. J.; Wilson, A. J.;
Wong, J. K. Y. Angew. Chem., Int. Ed. 2001, 40, 1538. (h) Johnston, A.
G.; Leigh, D. A.; Pritchard, R. J.; Deegan, M. D. Angew. Chem., Int. Ed.
Engl. 1995, 34, 1209. (i) Jeon, Y.-M.; Whang, D.; Kim, J.; Kim, K. Chem.
Lett. 1996, 503. (j) Fujita, M.; Ibukuro, F.; Hagihara, H.; Ogura, K. Nature
1994, 367, 720. (k) Glink, P. T.; Schiavo, C.; Stoddart, J. F.; Williams, D.
J. Chem. Commun. 1996, 1483. (l) Anelli, P. L.; Ashton, P. R.; Ballardini,
R.; Balzani, V.; Delgado, M.; Gandolfi, M. T.; Goodnow, T. T.; Kaifer,
A. E.; Philp, D.; Pietraszkiewicz, M.; Prodi, L.; Reddington, M. V.; Slawin,
A. M. Z.; Spencer, N.; Stoddart, J. F.; Vicent, C.; Williams, D. J. J. Am.
Chem. Soc. 1992, 114, 193.
(4) (a) Yang, X.; Knobler, C. C.; Hawthorne, F. Angew. Chem., Int. Ed. Engl.
1991, 11, 1507. (b) Hasenknopf, B.; Lehn, J.-M.; Kneisel, B. O.; Baum,
G.; Fenske, D. Angew. Chem., Int. Ed. Engl. 1996, 35, 1838. (c)
Hasenknopf, B.; Lehn, J.-M.; Boumediene, N.; Dupont-Gervais, A.; Van
Dorsselaer, A.; Kneisel, B.; Fenske, D. J. Am. Chem. Soc. 1997, 119, 10956.
(d) Fujita, M.; Nagao, S.; Ogura, K. J. Am. Chem. Soc. 1995, 117, 1649.
(e) Aoyagi, M.; Biradha, K.; Fujita, M. J. Am. Chem. Soc. 1999, 121, 7457.
(f) Vilar, R.; Mingos, D. M. P.; White, A. J. P.; Williams, D. J. Angew.
Chem., Int. Ed. 1998, 37, 1258. (g) Flemming, 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, 1279. (h) Campos-Fernandez, C. S.; Clerac, R.;
Dunbar, K. R. Angew. Chem., Int. Ed. 1999, 38, 3477.
(3) (a) Supramolecular Chemistry of Anions; Bianchi, A., Bowman-James, K.,
Garcia-Espana, E., Eds.; Wiley-VCH: New York, 1997. (b) Beer, P. D.;
Gale, P. A. Angew. Chem., Int. Ed. 2001, 40, 486. (c) Berger, M.;
Schmidtchen, F. P. Chem. ReV. 1997, 97, 1609.
(5) 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, 2068.
(6) Montalti, M.; Prodi, L. Chem. Commun. 1998, 1461.
9
10.1021/ja027519a CCC: $22.00 © 2002 American Chemical Society
J. AM. CHEM. SOC. 2002, 124, 12469-12476
12469

A R T I C L E S
Wisner et al.
Figure 1. (a) An anion-templated [2]pseudorotaxane. (b) The “clipping”
method for rotaxane formation based on an anion template.
sterically demanding electrophile generates rotaxane products
with some of the highest yields reported for mechanically
bonded systems. Vo¨gtle’s method depends on reaction of the
anionic component during formation of the products, eliminating
a possible contribution to their function. Hence, the anionic
template does not “live on” within the resulting construction
and, as such, is unavailable for manipulation of its co-
conformational disposition. The conservation of the anion
template within such a structure has the added potential of
creating an anion recognition domain specific to the template
used to construct it.
Figure 2. Structures of ion pairs 1 and hydrogen bond donating ligands 2,
3, and 4.
proved to be problematic, prompting the exploration of other
methods to generate the desired mechanical bond. This has
resulted in the selection of a “clipping” methodology (Figure
1b), which has met with success in other systems.11 Herein, we
describe the development of an acyclic chloride anion template
and its application in the synthesis of a [2]rotaxane which is
evaluated in terms of its use as a chloride anion receptor.
We have recently shown that chloride anions can be used to
template the formation of [2]pseudorotaxanes selectively.9 The
basis of this approach involves a cationic thread designed to
partially fill the coordination sphere of the strongly ion-paired
chloride anion in a noncompetitive solvent. Thus, the anion
remains coordinatively unsaturated, leaving an empty meridian
for complexation by a neutral hydrogen bond donating ligand
disposed orthogonally to the cation (Figure 1a). Selectivity
observed in this system for chloride anions over other anionic
species is a function of anion recognition by both the cation
and the neutral ligand. This preference is further established
through effective second-sphere coordination of the cationic
thread by the remaining components of the macrocyclic
framework, which act in concert with the anion recognition
event. Participation of a less complementary anion in this design
results in a disparity with the coordination environment created
in the superstructure and an ensuing reduction in the ability of
the cationic thread to reside within the cavity of the macrocycle.
Considering these results, our efforts have now turned to the
question of whether this anion template approach can be applied
to the more challenging task of constructing interlocked
molecules and what possible properties these products might
display. The obvious synthetic route to generate a rotaxane by
“stoppering” the ends of a preformed pseudorotaxane10 has
Development of an Acyclic Anion Template
Three ligands were initially synthesized and assessed for their
ability to coordinate the chloride anion in the ion pair 1a (Figure
2). Crabtree has shown that simple isophthalamide ligands are
good receptors for chloride anions in noncompetitive solvents.12
The syntheses of 1, 2, and 5 have been described elsewhere.9
Compounds 3 and 4 were synthesized from 5 by SN2 addition
to 2 equiv of either 1-bromohexane or the tosylate of ethyl-
eneglycol monoallyl ether, respectively, in a mixture of DMF
and K2CO3 (Scheme 1).
Ligand 2 incorporates a chloride-specific anion-binding cleft
formed by the amide protons and the isophthalamide ring proton
at the 2-position. Likewise, the cation 1+ contains the same
functionality in addition to the tight ion-pairing provided by
the positively charged pyridinium ring, serving to locate the
anion within its hydrogen bonding cleft in nonpolar solvents.
The strength and nature of the ligation of 1a by anion receptor
2 were established by titration of the ligand with the ion pair in
(10) (a) Chichak, K.; Walsh, M. C.; Branda, N. R. Chem. Commun. 2000, 10,
847. (b) Blanco, M.-J.; Chambron, J.-C.; Heitz, V.; Sauvage, J.-P. Org.
Lett. 2000, 2, 3051. (c) Rowan, S. J.; Cantrill, S. J.; Stoddart, J. F. Org.
Lett. 1999, 1, 129. (d) Loeb, S. J.; Wisner, J. A. Chem. Commun. 1998,
24, 2757. (e) Anderson, S.; Claridge, T. D. W.; Anderson, H. L. Angew.
Chem., Int. Ed. Engl. 1997, 36, 1310.
(11) (a) Glink, P. T.; Oliva, A. I.; Stoddart, J. F.; White, A. J. P.; Williams, D.
J. Angew. Chem., Int. Ed. 2001, 40(10), 1870. (b) Gatti, F. G.; Leigh, D.
A.; Nepogodiev, S. A.; Slawin, A. M. Z.; Teat, S. J.; Wong, J. K. Y. J.
Am. Chem. Soc. 2001, 123, 5983. (c) Hunter, C. A.; Low, C. M. R.; Packer,
M. J.; Spey, S. E.; Vinter, J. G.; Vysotsky, M. O.; Zonta, C. Angew. Chem.,
Int. Ed. 2001, 40, 2678. (d) Jeong, K.-S.; Choi, J. S.; Chang, S.-Y.; Chang,
S.-Y. Angew. Chem., Int. Ed. 2000, 39, 1692.
(12) (a) Kavallieratos, K.; Bertao, C. M.; Crabtree, R. H. J. Org. Chem. 1999,
64, 1675. (b) Kavallieratos, K.; de Gala, S. R.; Austin, D. J.; Crabtree, R.
H. J. Am. Chem. Soc. 1997, 119, 2325.
(7) (a) Hu¨bner, G. M.; Gla¨ser, J.; Seel, C.; Vo¨gtle, F. Angew. Chem., Int. Ed.
1999, 38, 383. (b) Reuter, C.; Wienard, W.; Hu¨bner, G. M.; Seel, C.; Vo¨gtle,
F. Chem.-Eur. J. 1999, 5, 2692.
(8) (a) Schalley, C. A.; Silva, G.; Nising, C. F.; Linnartz, P. HelV. Chim. Acta
2002, 85, 1578. (b) Mahony, J. M.; Shukla, R.; Marshall, R. A.; Beatty, A.
M.; Zajicek, J.; Smith, B. D. J. Org. Chem. 2002, 67, 1436. (c) Deetz, M.
J.; Shukla, R.; Smith, B. D. Tetrahedron 2002, 58, 799.
(9) Wisner, J. A.; Beer, P. D.; Drew, M. G. B. Angew. Chem., Int. Ed. 2001,
40, 3606.
9
12470 J. AM. CHEM. SOC. VOL. 124, NO. 42, 2002

Anion-Templated Rotaxane Formation
A R T I C L E S
Scheme 1. Synthesis of 3 and 4
Table 1. Association Constants, Free Energies of Complexation,
and Selected 1H NMR Shifts of 2, 3, and 4 with 1a in CD2Cl2-d2 at
298 K
chemical shift (∆δ in ppm)a
complex
K (M-1
)
−∆G (kJ mol-1
)
c
d
g and h
a
Figure 3. UV-vis spectra of 1:1 mixtures of 1a with 2, 3, and 4 in
methylene chloride at 5 × 10-3 M.
1a‚2
1a‚3
1a‚4
25
50
260
7.98
9.69
13.8
0.02
0.08
0.31
0.14
0.23
0.70
-0.03
-0.19
the observation of the amide and isophthalamide ring protons
of 4, which now experience marked downfield shifts during the
titration with 1a, denoting an even greater engagement of the
anion within its hydrogen bond donating cleft. Larger shifts of
the hydroquinone ring protons upfield in the 1H NMR spectra,
induced by π-stacking with the pyridinium ring of 4, are
observed, leading to the conclusion that the influence of the
terminal ether oxygen atoms serves to orient the rings in this
manner to a greater extent. This inference is supported by the
UV-vis spectrum of the 1:1 complex of 1a and 4, which
displays a more intense charge-transfer absorption than that of
1a with 3 (Figure 3).
As in the case of ligand 2, titration of 3 or 4 with hexa-
fluorophosphate derivative 1b gave no detectable change in the
system. Thus, the template hinges around the chloride anion
recognition event, the efficacy of which can then be enhanced
by the inclusion of ancillary groups to provide favorable second-
sphere coordination of the cation.
a Shifts determined from a 1:1 mixture of the two components at 5 ×
10-3 M.
methylene chloride and by analyzing the resulting shifts in the
1H NMR spectra. As can be seen in Table 1, the shifts of both
the amide and the aryl protons in the ligand are small, and the
calculated association constant (25 M-1, -∆G ) 7.98 kJ mol-1
)
is low. This is interpreted as a consequence of strong polarization
of the chloride anion within the cleft of the cation, making it a
poor hydrogen bond acceptor and resulting in a weak interaction
with hydrogen bond donating ligand 2. A similar titration of 2
with 1b displayed no observable change in the system, confirm-
ing the assumption that the association is a result of chloride
anion recognition.
Ligands 3 and 4 were designed to investigate whether second-
sphere coordination of the chloride anion could improve the
anion template interaction. Ligand 3 contains electron-rich
hydroquinone rings in its pendant arms, which were expected
to reach around the chloride anion and engage the electron-
poor pyridinium ring in a classic donor-acceptor π-stacking
interaction.13 This modification yields a modest increase in the
free energy of association over that of 2 (Ka ) 50 M-1, -∆G
) 9.69 kJ mol-1) when titrated with 1a. The magnitudes of the
shifts of the amide and isophthalamide ring protons also increase
by a significant margin in both cases, indicating a greater ability
of the ligand to participate in hydrogen bonding with the anion.
The presence of π-stacking between the ligand and the cation
is revealed by small upfield shifts of the hydroquinone protons
during the titration. This is further demonstrated by the
observation of a charge-transfer absorption at approximately 380
nm upon the admixture of the two colorless components,
conferring a pale yellow color to the solution (Figure 3).
Ligand 4 includes both hydroquinone rings in its structure
as well as the addition of terminal ether linkages designed to
augment the second-sphere coordination of the chloride anion
by providing C-H‚‚‚O hydrogen bonds to the N+-Me group
and R-protons of the pyridinium ring in 1a. The success of this
strategy is borne out by the resulting increase in the free energy
of association (Ka ) 260 M-1, -∆G ) 13.8 kJ mol-1) over
the other two ligands. This enhancement is again exhibited in
Anion-Templated Rotaxane Formation
As mentioned previously, a logical extension of the successful
anion coordination of the ion pair 1a by ligand 4 is the formation
of a rotaxane utilizing this anion template via a clipping pro-
cedure. Toward this end, the stoppered thread-shaped com-
pound 8 was synthesized as depicted in Scheme 2. Coupling of
the triphenylmethylaniline15 6 to 3,5-pyridinedichlorocarbonyl
produced dumbbell-shaped pyridine compound 7, which was
then quaternized with methyl iodide to give the corresponding
pyridinium thread as the iodide salt 8b. Anion exchange of
8b with NH4Cl or AgPF6 formed the desired 8a and 8c,
respectively.
The generation of a [2]rotaxane from complex 4‚8a was
effected by the use of Grubbs’ ring-closing metathesis (RCM)
on the allylic end groups of the ligand (Scheme 3).14 Orientation
of the arms of 4 around dumbbell-shaped 8a, in a manner similar
to that described with 1a, now serves not only to support the
chloride anion template but to facilitate the ring closure reaction
about the ion pair, resulting in a mechanical bond to produce
the [2]rotaxane 9a in 47% yield. Although the macrocyclization
reaction has the possibility of producing both cis and trans
(14) (a) Weck, M.; Mohr, B.; Sauvage, J. P.; Grubbs, R. H. J. Org. Chem. 1999,
64, 5463. (b) Grubbs, R. H.; Chang, S. Tetrahedron 1998, 54, 4413.
(15) Gibson, H. W.; Lee, S.-H.; Engen, P. T.; Lecavalier, P.; Sze, J.; Shen, Y.
X.; Bheda, M. J. Org. Chem. 1993, 58, 3748.
(13) (a) Gillard, R. E.; Raymo, F. M.; Stoddart, J. F. Chem.-Eur. J. 1997, 3,
1933. (b) Claessens, C. G.; Stoddart, J. F. J. Org. Phys. Chem. 1997, 10,
254. (c) Langford, S. J.; Stoddart, J. F. Pure Appl. Chem. 1996, 68, 1255.
9
J. AM. CHEM. SOC. VOL. 124, NO. 42, 2002 12471

A R T I C L E S
Wisner et al.
Scheme 2. Synthesis of Ion Pairs 8a-c
Scheme 3. Synthesis of [2]Rotaxanes 9a and 9b
ion pair 1b and further demonstrates the chloride anion template
upon which the synthetic scheme depends.16
Analysis of the 1H NMR spectrum of 9a in comparison with
4 and 8a reveals the interlocked nature of the components
(Figure 4). Large downfield shifts of the amide and their
mutually ortho-isophthalamide ring protons in the macrocycle
(∆δ ) 2.24 and 0.70 ppm, respectively) indicate the complex-
ation of the anion within this hydrogen bond donating cleft.
This is mirrored by a concomitant shift upfield of the amide
and para-pyridinium ring protons in the dumbbell-shaped axle
(∆δ ) -0.61 and -0.99 ppm, respectively), a consequence of
polarization of the chloride anion away from the cationic cleft
by the neutral wheel. Inclusion of the pyridinium ring within
the macrocyclic cavity generates π-stacking-induced upfield
shifts of the protons at the ortho position (∆δ ) -0.22 ppm)
and those attached to the hydroquinone rings which split into
separate signals (∆δ ) -0.42 and -0.64 ppm) due to the
asymmetric environment presented by the interlocked cation.
The presence of the pyridinium ring within the cavity is also
supported by C-H‚‚‚O hydrogen bonding between the polyether
chain of the macrocycle and the N+-Me group of the cation,
these protons shifting downfield (∆δ ) 0.58 ppm) as a result
of the interaction.
Single crystals of 9a suitable for X-ray diffraction analysis
were grown by the slow diffusion of diisopropyl ether into a
chloroform solution of the [2]rotaxane. The structural analysis
reveals the expected interlocked product in which the macro-
cycle encircles the ion pair (Figure 5a). The chloride anion fits
well into the cleft of the pyridinium ring, forming hydro-
gen bonds with five hydrogen atoms directed toward it with
X-H‚‚‚Cl distances of C(55A) 3.633, N(53A) 3.337, C(1A)
3.335, N(13A) 3.403, and C(19A) 3.821 Å and X-H‚‚‚Cl angles
of 147, 168, 155, 167, and 145°, respectively (Figure 5b). These
five atoms form a pentagonal girdle for the chloride anion, and
least squares planes calculations show that C(55A), N(53A),
C(1A), N(13A), and C(19A) are coplanar to within 0.03 Å, the
anion being 0.48 Å from the plane described by them. The
angles of intersection between the central pyridinium and the
isomers at the double bond of the product, only the trans isomer
was observed. This is presumably a function of the metathesis
reaction, which equilibrates to a thermodynamic mixture, the
trans isomer being the more stable derivative.
The analogous reaction of 8c with ligand 4 failed to produce
any detectable mechanically bonded products. This is predictable
considering the complete lack of affinity of 4 for the similar
(16) Analogous reactions of 8b or the bromide pyridinium stoppered thread with
4 also failed to produce any evidence for rotaxane formation.
9
12472 J. AM. CHEM. SOC. VOL. 124, NO. 42, 2002

Anion-Templated Rotaxane Formation
A R T I C L E S
Figure 4. 1H NMR spectra of 4, 8a, and 9a in CDCl3 at 298 K.
case, the C-H‚‚‚Cl and N-H‚‚‚Cl angles are 167, 112, and
170°, respectively, indicating that only the two nitrogen atoms
form strong hydrogen bonds. The orientation of the phenyl
ring around C(52) is particularly interesting as it is tilted out
of the plane such that N(59)-C(57)-C(53)-C(52) and
N(41)-C(42)-C(51)-C(52) torsion angles are 30.3 and -38.0°,
respectively. Thus, the chloride anion can be considered as
having seven hydrogen bonds, with those of the pyridinium
cation forming a pentagonal plane, with the two contributed by
the isophthalamide macrocycle being located in approximately
axis positions.
The macrocycle surrounds the cation such that the two
hydroquinone rings are approximately parallel and sandwich
the central pyridinium ring. Least squares plane calculations
show that the two phenyl rings in the macrocycle are 6.95 Å
apart and are parallel to within 9.4°. These two rings therefore
allow enough room for π-stacking with the enclosed cation.
However, these two rings do not overlap the pyridinium ring
completely but rather two atoms in one amide linkage and two
carbon atoms in the pyridinium ring. Thus, atoms N(53A),
C(51A), C(6A), and C(1A) form the sandwich, and their plane
intersects those of the two hydroquinone rings at angles of 11.5
and 3.5°, respectively. Distances between these four atoms and
the overlapping atoms in the hydroquinone rings range from
3.37 to 3.80 Å. The positioning of the pyridinium ring is not
symmetric with respect to the macrocycle and veers to one side
such that C(5A) is 3.07 Å from O(21) and the N+-Me group
C(4A) is 3.50 Å from O(24). Presumably the reason for this
asymmetry is the formation of a hydrogen bond between the
methyl group C(4A) and the ether oxygen O(24). The H‚‚‚O
distance is 2.61 Å, and the C-H‚‚‚O angle is 154°. This
hydrogen atom is also directed at the alkenyl double bond of
the macrocycle, but the distances from these carbon atoms of
3.57 and 3.66 Å are too long to indicate any C-H‚‚‚π
interaction.
Figure 5. (a) Stick representation of the solid-state structure of [2]rotaxane
9a illustrating the interlocked nature of the components (red, macrocycle;
blue, pyridinium cation; green, chloride anion as CPK sphere). Solvent
molecules have been removed, and only one occupancy of the tert-butyl
groups is shown for clarity. (b) Stick representation of the solid-state
structure of [2]rotaxane 9a with atom numbering scheme (chloride anion
as green sphere). Solvent molecules have been removed, and only one tert-
butyl phenyl group of the (tert-butylphenyl)2(phenyl)methyl stoppers is
shown for clarity.
two adjacent phenyl rings of the cation are 18.8 and 11.7°,
respectively.
The positioning of the isophthalamide ring is asymmetric with
respect to the cation, being tilted toward an aromatic tert-butyl
phenyl group in one of the stoppers. Indeed, it would appear
that the reasons for this asymmetry are C-H‚‚‚π interactions
By contrast, complexation of the anion by the isophthalamide
macrocycle is far more distorted. The chloride anion is close to
three hydrogen atoms directed toward the cavity. Distances are
N(59) 3.394, C(52) 3.369, and N(41) 3.458 Å. However, in this
9
J. AM. CHEM. SOC. VOL. 124, NO. 42, 2002 12473

A R T I C L E S
Wisner et al.
Table 2. Association Constants (M-1) Determined by Titration of
8c and 9b with TBA Salts of Three Anions in 1:1 CDCl3:CD3OD at
298 K
the [2]rotaxane superstructure, which is of complementary
topology for the chloride anion. Larger anions such as dihy-
drogen phosphate and acetate must either penetrate this pocket,
which should be sterically demanding and displace the pyri-
dinium cation from the macrocyclic cavity, or bind to the
periphery of the rotaxane, a configuration which must involve
less than the full complement of possible hydrogen bond donors
that the [2]rotaxane can present to a chloride anion.17
receptor
anion
Cl-
8c
9b
K11
K11
K11
K12
125
260
22 000
140
K11
K11
K11
K12
1130
300
100
40
-
H2PO4
AcO-
Conclusions
between two of the hydrogen atoms on the phenyl ring of the
isophthalamide ring (viz. H(55) and H(56)) and two of the
carbon atoms of the tert-butyl phenyl stopper (C24A and C25A)
with C-H‚‚‚C distances of 3.36, 3.45, 3.04, and 3.82 Å.
The results presented here have demonstrated the rational
development of a chloride anion template involved in the
orthogonal coordination of a neutral hydrogen bond donating
ligand and a pyridinium cation. Efficacy of the template was
improved by the incremental incorporation of π-stacking and
hydrogen bond second-sphere interactions between the cationic
and neutral components, which in concert aid the complexation
of the chloride anion template. This anionic template was then
applied to the formation of a [2]rotaxane by a clipping procedure
using ring-closing metathesis. Comparison of the anion recogni-
tion properties of the dumbbell-shaped pyridinium cation and
the [2]rotaxane derived from it has illustrated that incorporation
of a macrocyclic hydrogen bond donating ligand about the
central axle can dramatically effect its selectivity for a particular
anion. The rotaxane product exhibits a dramatic increase in its
selectivity for chloride as compared to the noninterlocked cation.
Anion Recognition
One of the central motives for the development of the anion-
templated [2]rotaxane described was the possibility of employing
it as a chloride-specific anion receptor once the templating anion
had been removed. Toward this end, the chloride anion in 9a
was replaced with the noncompetitive PF6 anion by treatment
with AgPF6 in methylene chloride to yield [2]rotaxane 9b.
Comparison of the CDCl3 1H NMR spectra of the two rotaxanes
indicates that replacement of the chloride anion with hexafluo-
rophosphate does little to displace the cationic dumbbell 8+ from
the cavity of the macrocycle in this solvent. This conclusion is
supported by the small differences between the π-stacking-
induced shifts of the hydroquinone protons (∆δ ) 0.06-0.08
ppm) in the macrocyclic components of both species with the
pyridinium cation.
We anticipate that the principles exploited here can be
developed further to include different anionic, cationic, and
neutral components, resulting in new interlocked receptors
selective for the anion template employed. Research is currently
proceeding in our laboratories toward this end in addition to
elaboration and refinement of the system presented here to
further improve selectivity of the [2]rotaxane for the templating
chloride anion.
To determine the effect that rotaxane formation has upon
anion recognition, both free dumbbell-shaped 8c and [2]rotaxane
9b were titrated with three different anions in a 1:1 mixture of
CDCl3:CD3OD (Table 2), monitoring the resulting shift of the
1
proton at the para-position of the pyridinium ring using H
NMR. This solvent mixture was chosen due to problems with
the solubility of either the starting compounds or the complexed
products during titration in other competitive solvents such as
DMSO-d6 or pure CD3OD. As can be seen from the table, the
free axle 8c shows a much greater preference for binding acetate
anions over dihydrogen phosphate and chloride, as well as 1:1
and 1:2 (receptor:anion) binding stoichiometry. Titrations of 8c
with chloride and dihydrogen phosphate anions both display
1:1 stoichiometry, but the association constants are moderate
with the least preference for chloride, even though the hydrogen
bond donating cleft of the cation is of the correct size and shape
for this particular anion. These results may be interpreted as a
function of the basicity of the oxoanions being the contributing
factor in their greater affinity for this receptor.
Titrations of [2]rotaxane 9b with the same anions display
the opposite selectivity pattern in comparison with 8c. As a result
of encirclement of the pyridinium cation by the macrocyclic
ligand, the rotaxane now binds chloride anions more strongly
than acetate and dihydrogen phosphate. Acetate is still bound
with 1:1 and 1:2 stoichiometries, but the magnitudes of the
interactions are reduced greatly in relation to 8c. The extent of
anion recognition with dihydrogen phosphate remains essen-
tially unchanged (Ka ) 300 M-1) by the transition to interlocked
9b. These differences in anion selectivity can be ascribed to
the creation of a unique hydrogen bond donating pocket
formed by the diamide clefts of the cation and macrocycle in
Experimental Section
General Methods. Chemicals were purchased from Aldrich and used
as received. All reactions were carried out under an atmosphere of dry
N2 unless otherwise stated. All solvents were dried and distilled before
use according to standard laboratory procedures. Bis(4-tert-butylphe-
nyl)phenyl-methanol15 and 2-allyloxyethanol-p-toluenesulfonate18 were
prepared according to literature methods. Chromatography was per-
formed on 240-400 mesh silica gel-60. 1H and 13C NMR spectra were
recorded on Varian 300 MHz Mercury VX Works and 500 MHz Unity
Plus spectrometers. UV-vis spectra were recorded at a concentration
of 5 × 10-3 M on a Perkin-Elmer Lambda 6 UV-vis spectrometer.
(3). Compound 5 (0.300 g, 0.687 mmol) and 1-bromohexane (0.308
g, 1.72 mmol) were dissolved in a mixture of dimethylformamide (30
mL) and K2CO3 (0.200 g, 1.44 mmol) and were heated to 60 °C for 16
h. The mixture was cooled to room temperature and filtered to remove
insolubles. The solvent was then removed by rotoevaporation, and the
remaining light brown oil was chromatographed (silica gel, 95:5 CHCl3:
MeOH) to give the product as a white solid (0.154 g, 37%). 1H NMR
(CDCl3, 300 MHz, TMS, 298 K): δ (ppm) 8.20 (s, 1H), 7.93 (d, J )
7.6 Hz, 2H), 7.51 (t, J ) 7.6 Hz, 1H), 6.83 (m, 8H), 6.75 (b, 2H), 4.11
(m, 4H), 3.90 (m, 8H), 1.77 (m, 4H), 1.46 (m, 4H), 1.36 (m, 8H), 0.92
(m, 6H). 13C NMR: δ (ppm) 166.7, 153.8, 152.4, 134.8, 130.2, 129.2,
125.6, 115.7, 115.6, 69.0, 67.7, 40.3, 32.2, 29.9, 26.3, 23.2, 14.7. Elem.
1
(17) Analogous H NMR titration experiments of 9b with bromide and iodide
showed the rotaxane does not bind either halide.
(18) Maynard, H. D.; Grubbs, R. H. Macromolecules 1999, 32, 6917.
9
12474 J. AM. CHEM. SOC. VOL. 124, NO. 42, 2002

Anion-Templated Rotaxane Formation
A R T I C L E S
Anal. Calcd for C36H48N2O6: C, 71.50; H, 8.00; N, 4.63. Found: C,
71.30; H, 7.72; N, 4.52.
147.1, 144.8, 143.8, 135.3, 134.6, 131.8, 131.2, 130.7, 127.6, 125.9,
124.5, 119.8, 77.6, 64.3, 34.8, 31.9. Elem. Anal. Calcd for C74H78-
ClN3O2: C, 82.53; H, 7.30; N, 3.90. Found: C, 82.22; H, 7.70; N,
4.00.
(4). Compound 5 (0.300 g, 0.687 mmol) and 2-allyloxyethanol-p-
toluenesulfonate (0.440 g, 1.72 mmol) were dissolved in a mixture of
dimethylformamide (30 mL) and K2CO3 (0.200 g, 1.44 mmol) and were
heated to 60 °C for 16 h. The mixture was cooled to room temperature
and filtered to remove insolubles. The solvent was then removed by
rotoevaporation, and the remaining light brown oil was chromato-
graphed (silica gel, 95:5 CHCl3:MeOH) to give the product as a white
(8c). Compound 8b (0.150 g, 0.128 mmol) and AgPF6 (0.048 g,
0.19 mmol) were suspended in MeCl2 (30 mL) and stirred in the absence
of light for 16 h. The resulting mixture was filtered through a pad of
Celite, and the solvent was evaporated to give the pure product as a
1
yellow solid (0.146 g, 96%). H NMR (CDCl3, 300 MHz, TMS, 298
1
solid (0.174 g, 42%). H NMR (CDCl3, 300 MHz, TMS, 298 K): δ
K): δ (ppm) 10.02 (b, 2H), 9.36 (s, 1H), 8.76 (s, 2H), 7.49 (d, J ) 9.0
Hz, 4H), 7.19 (m, 30H), 3.76 (s, 3H), 1.26 (s, 32H). 13C NMR: δ
(ppm) 147.8, 146.4, 145.2, 144.1, 143.1, 135.2, 134.5, 131.2, 130.5,
130.1, 127.0, 125.3, 124.0, 119.4, 77.2, 63.9, 34.6, 31.7. Elem. Anal.
Calcd for C74H78F6N3O2P: C, 74.92; H, 6.63; N, 3.54. Found: C, 74.48;
H, 6.68; N, 3.84.
(ppm) 8.20 (s, 1H), 7.91 (d, J ) 8.4 Hz, 2H), 7.47 (t, J ) 8.4 Hz, 1H),
6.90 (b, 2H), 6.82 (m, 8H), 5.92 (m, 1H), 5.29 (d, J ) 16.4 Hz, 1H),
5.20 (d, J ) 10.6 Hz, 1H), 4.07 (m, 12H), 3.83 (m, 4H), 3.77 (t, J )
5.4 Hz, 4H). 13C NMR: δ (ppm) 166.4, 153.0, 152.3, 134.4, 134.2,
129.8, 128.7, 125.2, 117.2, 115.5, 115.2, 72.3, 68.6, 68.0, 67.2, 39.9.
Elem. Anal. Calcd for C34H40N2O8: C, 67.53; H, 6.67; N, 4.63.
Found: C, 67.19; H, 6.88; N, 4.58.
(9a). Compounds 4 (0.125 g, 0.207 mmol) and 8a (0.125 g, 0.116
mmol) were dissolved in dichloromethane and stirred for 5 min. Grubbs’
catalyst was added (20 mol %), and the solution was stirred for a further
16 h. The solvent was evaporated, and the red-brown solid was
chromatographed (silica gel, 93:7 CHCl3:MeOH), collecting the first
yellow band eluted. The fractions were evaporated and crystallized by
slow diffusion of diisopropyl ether into chloroform to yield the final
4-[Bis(4-t-butylphenyl)phenylmethyl]aniline (6). Bis(4-tert-butyl-
phenyl)phenyl-methanol (6 g, 16.1 mmol) was dissolved in acetyl
chloride (30 mL) and refluxed for 12 h. The resulting solution was
cooled to room temperature, and the solvent was removed under vacuum
to leave a yellow solid. The solid was dissolved in aniline (30 mL)
and heated to 100 °C for 48 h, during which time it changed from
deep red to dark purple. The solution was again cooled to room
temperature and added slowly to a rapidly stirred solution of concen-
trated HCl (16 mL) in water (300 mL) to form a purple solid. The
solid was filtered off, dissolved in MeCl2 (200 mL), and passed through
a plug of silica (50 g). The silica was further eluted with MeCl2 (200
mL), and the solvent was evaporated to leave the crude orange/brown
product. Recrystallization from toluene/hexane gave the product as a
white crystalline solid (5.19 g, 72%). 1H NMR (CDCl3, 300 MHz, TMS,
298 K): δ (ppm) 7.22 (m, 9H), 7.10 (d, J ) 8.5 Hz, 4H), 6.97 (d, J )
1
product as a bright yellow crystalline solid (0.091 g, 47%). H NMR
(CDCl3, 300 MHz, TMS, 298 K): δ (ppm) 10.31 (b, 2H), 9.66 (b,
1H), 9.14 (b, 2H), 8.89 (b, 1H), 8.55 (b, 2H), 8.00 (d, J ) 7.5 Hz,
2H), 7.78 (d, J ) 9.0 Hz, 4H), 7.16 (m, 31H), 6.40 (d, J ) 8.7 Hz,
4H), 6.18 (d, J ) 8.7 Hz, 4H), 6.02 (b, 2H), 4.37 (b, 3H), 4.08 (b,
8H), 3.82 (b, 4H), 3.75 (b, 8H), 1.33 (s, 36H). 13C NMR: δ (ppm)
167.1, 158.2, 153.5, 151.9, 148.6, 147.1, 145.4, 144.7, 143.6, 135.0,
134.1, 133.6, 131.9, 131.7, 131.2, 130.8, 130.2, 128.9, 127.6, 126.0,
124.5, 120.3, 77.6, 71.4, 69.8, 68.4, 66.2, 64.3, 41.1, 34.9, 31.9. Elem.
Anal. Calcd for C106H114ClN5O10: C, 77.00; H, 6.95; N, 4.24. Found:
C, 76.48; H, 6.72; N, 4.01.
9.0 Hz, 2H), 6.57 (d, J ) 9.0 Hz, 2H), 3.60 (b, 2H), 1.27 (s, 18H). 13
C
NMR: δ (ppm) 148.3, 147.7, 144.3, 144.0, 137.5, 132.2, 131.3, 130.9,
127.3, 125.7, 124.3, 114.3, 63.8, 34.8, 31.9. Elem. Anal. Calcd for
C33H37N: C, 88.54; H, 8.33; N, 3.13. Found: C, 88.43; H, 8.38; N,
3.19.
(9b). Compound 9a (0.050 g, 0.030 mmol) was dissolved in
methylene chloride (10 mL), and AgPF6 (0.012 g, 0.045 mmol) was
added to the solution, which was then stirred in the absence of light
for 16 h. The resulting mixture was filtered through Celite, which was
washed with two further aliquots of methylene chloride (2 × 5 mL).
The solvent was evaporated to yield the pure product as a yellow green
(7). 4-[Bis(4-tert-butylphenyl)phenylmethyl]aniline (1.00 g, 2.23
mmol) and triethylamine (0.34 g, 3.4 mmol) were dissolved in MeCl2
(40 mL), and 3,5-pyridinedichlorocarbonyl (0.228 g, 1.12 mmol) was
added as a solid to the reaction mixture. The resulting solution was
stirred for 4 h, after which time the solvent was removed by rotary
evaporation to leave a yellow solid. The residue was refluxed in ethanol
(50 mL) for 1 h and, after cooling to room temperature, filtered off to
give the pure product as a white solid (0.977 g, 85%). 1H NMR (DMSO-
d6, 300 MHz, TMS, 298 K): δ (ppm) 10.59 (b, 2H), 9.19 (s, 2H), 8.70
(s, 1H), 7.66 (d, J ) 9.0 Hz, 4H), 7.14 (m, 30H), 1.27 (s, 32H). Elem.
Anal. Calcd for C73H75N3O2: C, 85.42; H, 7.37; N, 4.09. Found: C,
85.05; H, 7.39; N, 4.20.
1
solid (0.048 g, 90%). H NMR (CDCl3, 300 MHz, TMS, 298 K): δ
(ppm) 10.14 (b, 2H), 9.06 (b, 1H), 8.91 (b, 2H), 8.62 (b, 1H), 8.17 (d,
J ) 7.5 Hz, 2H), 7.96 (b, 2H), 7.67 (d, J ) 9.0 Hz, 4H), 7.53 (t, J )
9.0 Hz, 1H), 7.25 (m, 22H), 7.13 (d, J ) 7.5 Hz, 8H), 6.48 (d, J ) 8.7
Hz, 4H), 6.24 (d, J ) 8.7 Hz, 4H), 5.85 (b, 2H), 4.12 (b, 3H), 4.01 (b,
4H), 3.96 (b, 4H), 3.81 (b, 8H), 3.75 (b, 4H), 1.33 (s, 36H). 13C NMR:
δ (ppm) 166.2, 157.6, 152.3, 151.5, 148.0, 146.4, 144.4, 144.1, 142.9,
134.2, 133.8, 133.3, 131.4, 131.0, 130.6, 130.2, 129.6, 128.6, 127.0,
125.5, 123.9, 119.7, 115.4, 114.8, 114.3, 77.2, 72.4, 70.9, 69.2, 67.6,
63.9, 40.1, 34.6, 31.7. Elem. Anal. Calcd for C106H114F6N5O10P: C,
72.21; H, 6.52; N, 3.97. Found: C, 71.45; H, 6.34; N, 3.98.
1H NMR Titrations. Titrations were performed with a starting
concentration of 2, 3, 4, 8c, and 9b at 5 × 10-3 M and the addition of
the appropriate aliquots of titrant with a microsyringe. Data from all
titrations were analyzed using the program EQNMR19 to obtain
association constants. Data were obtained during titrations of 2, 3, and
4 by observing the shifts in the proton resonances of the amide groups
and the 2-position of the isophthalamide ring upon addition of 1a, which
gave the same values of association constants within error. Data were
obtained during titrations of 8c and 9b by observing shifts in the
resonance of the proton at the para-position of the pyridinium ring
upon addition of TBA salts of the three anions. All titrations fit 1:1 or
1:2 binding models as appropriate. The stoichiometry of complexation
was verified by Job plot analysis where necessary. Job plot analysis of
the titration of 9b with dihydrogen phosphate indicated that the
(8b). Compound 7 (1.00 g, 0.974 mmol) was suspended in a solution
of acetone (50 mL) and methyl iodide (10 mL) and refluxed for 72 h.
The resulting mixture was cooled to room temperature, and the yellow
1
solid was filtered off to give pure product (0.983 g, 86%). H NMR
(DMSO-d6, 300 MHz, TMS, 298 K): δ (ppm) 10.90 (b, 2H), 9.61 (s,
2H), 9.45 (s, 1H), 7.67 (d, J ) 9.0 Hz, 4H), 7.14 (m, 30H), 4.48 (s,
3H), 1.26 (s, 32H). 13C NMR: δ (ppm) 160.4, 148.5, 147.3, 144.0,
143.8, 136.3, 134.1, 131.6, 130.9, 130.6, 128.3, 125.2, 120.2, 80.0,
35.1, 32.1. Elem. Anal. Calcd for C74H78IN3O2: C, 76.07; H, 6.73; N,
3.60. Found: C, 75.07; H, 6.77; N, 3.82.
(8a). Compound 8b (0.150 g, 0.128 mmol) was suspended in
chloroform (100 mL) and washed with 1 M aqueous NH4Cl (8 × 100
mL) and water (2 × 100 mL). The organic layer was dried with MgSO4,
filtered, and rotoevaporated to give the pure product as a yellow solid
1
(0.130 g, 94%). H NMR (CDCl3, 300 MHz, TMS, 298 K): δ (ppm)
10.92 (b, 2H), 10.65 (s, 1H), 8.77 (s, 2H), 7.80 (d, J ) 9.0 Hz, 4H),
7.19 (m, 30H), 3.79 (s, 3H), 1.26 (s, 32H). 13C NMR: δ (ppm) 148.5,
(19) Hynes, M. J. J. Chem. Soc., Dalton Trans. 1993, 311.
9
J. AM. CHEM. SOC. VOL. 124, NO. 42, 2002 12475

A R T I C L E S
Wisner et al.
stoichiometry of complexation was 1:1. However, the data were only
fit satisfactorily with the inclusion of a weak 2:1 (anion:receptor)
contribution (K21 ) 50 M-1). Estimated errors are e10% for Ka values
was refined on F 2 using Shelxl.22 The final R values were R1 0.113
and wR2 0.286 for 4797 data with I > 2σ(I). Details of the structure
have been deposited at the CCDC Reference No. 187740. Copies of
the data can be obtained free of charge on application to CCDC, 12
Union Road, Cambridge CB2 1EZ, U.K. (fax: (+44) 1223-336-033;
e-mail: deposit@ccdc.cam.ac.uk).
greater than 100 M-1 and e20% for Ka values less than 100 M-1
.
X-ray Crystallography. Crystal data for 9a‚1.5C2H5OH‚
1.5CH3OH‚0.5H2O: C112H131.5ClN5O12.5, M ) 1783.17, triclinic, space-
group P-1, a ) 14.731(23), b ) 18.514(27), c ) 21.150(27), R )
100.25(1), â ) 106.68(1), γ ) 101.87(1)°, U ) 5257 Å3, Z ) 2, Fcalc
) 1.127 g cm-3. Intensity data were collected with Mo KR radiation
using the Marresearch Image Plate System. The crystal was positioned
at 70 mm from the Image Plate. One hundred frames were measured
at 2° intervals with a counting time of 4 min to give 19 156 independent
reflections. Data analysis was carried out with the XDS program.20 The
structure was solved using direct methods with the Shelx86 program.21
Several of the tert-butyl groups were disordered, and two sets of
positions for the three methyl groups were refined with occupancies x
and 1 - x. The asymmetric unit also contained several solvent
molecules, three ethanol and three methanol molecules, and one water
molecule, all refined with 50% occupancy. Apart from the disordered
atoms and the solvent molecules, all non-hydrogen atoms were refined
with anisotropic thermal parameters. The hydrogen atoms were included
in geometric positions and given thermal parameters equivalent to 1.2
times those of the atom to which they were attached. The structure
Acknowledgment. We thank EPSRC for a postdoctoral
fellowship (J.A.W.) and a studentship (M.R.S.). The EPSRC
and the University of Reading are thanked for funds for the
Image Plate System. Financial support for Dr. Wisner from the
Natural Sciences and Engineering Research Council of Canada
is also acknowledged.
Supporting Information Available: Tables giving the crystal
data and structure refinement information, bond lengths and
bond angles, atomic and hydrogen coordinates, and isotropic
and anisotropic displacement coordinates for 9a (PDF). This
material is available free of charge via the Internet at
http://pubs.acs.org.
JA027519A
(20) Kabsch, W. J. Appl. Crystallogr. 1988, 21, 916.
(21) Sheldrick, G. M. Acta Crystallogr. 1990, A46, 467.
(22) Sheldrick, G. M. Program for crystal structure refinement. University of
Gottingen, 1993.
9
12476 J. AM. CHEM. SOC. VOL. 124, NO. 42, 2002