[C-H···O] interactions as a control element in supramolecular complexes: Experimental and theoretical evaluation of receptor affinities for the binding of bipyridinium-based guests by catenated hosts

Article abstract of DOI:10.1021/ja982748b

Macrocyclic receptors incorporating two facing π-electron-rich aromatic surfaces, held at a distance of approximately 7 A? by polyether spacers, bind bipyridinium-based guests. This recognition motif, which is dictated by π- π stacking and [C-H···O] hydrogen-bonding interactions, has led to the development of efficient template-directed syntheses of mechanically interlocked molecules, such as catenanes and rotaxanes. By employing a supramolecularly assisted synthetic methodology based on these interactions, we have self-assembled two novel [3]catenanes, each incorporating two 1,5- dioxynaphtho-38-crown-10 components and one bipyridinium-based tetracationic cyclophane component. Single-crystal X-ray analyses of these [3]catenanes revealed that they possess internal cavities bounded on two opposite sites by π-electron-rich 1,5-dioxynaphthalene units separated by a distance of approximately 7-8 A?. Despite the presence of apparently ideal 'binding pockets', these mechanically interlocked compounds steadfastly refuse to bind bipyridinium-based guests in solution, as demonstrated by both ¹H NMR and UV-vis spectroscopy. AMBER* and HF/321G calculations on appropriate models show that the absence of [C-H···O] hydrogen-bonding interactions is responsible for the instability of these geometrically ideal complexes. The [C-H···O] bond appears to be quantitatively much more important than π- π stacking interactions in these particular systems.

Full text of DOI:10.1021/ja982748b

J. Am. Chem. Soc. 1999, 121, 1479-1487
1479
[C-H‚‚‚O] Interactions as a Control Element in Supramolecular
Complexes: Experimental and Theoretical Evaluation of Receptor
Affinities for the Binding of Bipyridinium-Based Guests by Catenated
Hosts1
K. N. Houk,*,† Stephan Menzer,‡ Simon P. Newton,§ Franc¸isco M. Raymo,†,§
J. Fraser Stoddart,*,†,§ and David J. Williams*,‡
Contribution from the Department of Chemistry and Biochemistry, UniVersity of California,
405 Hilgard AVenue, Los Angeles, California 90095-1569, School of Chemistry, UniVersity of Birmingham,
Edgbaston, Birmingham B15 2TT, U.K., and Department of Chemistry, Imperial College,
South Kensington, London SW7 2AY, U.K.
ReceiVed August 3, 1998
Abstract: Macrocyclic receptors incorporating two facing π-electron-rich aromatic surfaces, held at a distance
of approximately 7 Å by polyether spacers, bind bipyridinium-based guests. This recognition motif, which is
dictated by π-π stacking and [C-H‚‚‚O] hydrogen-bonding interactions, has led to the development of efficient
template-directed syntheses of mechanically interlocked molecules, such as catenanes and rotaxanes. By
employing a supramolecularly assisted synthetic methodology based on these interactions, we have
self-assembled two novel [3]catenanes, each incorporating two 1,5-dioxynaphtho-38-crown-10 components
and one bipyridinium-based tetracationic cyclophane component. Single-crystal X-ray analyses of these
[3]catenanes revealed that they possess internal cavities bounded on two opposite sites by π-electron-rich
1,5-dioxynaphthalene units separated by a distance of approximately 7-8 Å. Despite the presence of apparently
ideal “binding pockets”, these mechanically interlocked compounds steadfastly refuse to bind bipyridinium-
1
based guests in solution, as demonstrated by both H NMR and UV-vis spectroscopy. AMBER* and HF/3-
21G calculations on appropriate models show that the absence of [C-H‚‚‚O] hydrogen-bonding interactions
is responsible for the instability of these geometrically ideal complexes. The [C-H‚‚‚O] bond appears to be
quantitatively much more important than π-π stacking interactions in these particular systems.
Introduction
and two π-electron-rich dioxyarene rings, affording [2]- and
[3]catenanes, respectively, when the π-electron-rich recognition
sites are incorporated within macrocyclic polyethers. From the
inspection of CPK space-filling molecular models, the obvious
progression to [4]catenanes (Figure 1) incorporating two mac-
rocyclic polyethers and two tetracationic cyclophane compo-
nents appears simply to require the insertion of a third
p-phenylene ring into each of the neutral spacers separating the
(4) For accounts and reviews on [C-H‚‚‚O] hydrogen-bonding interac-
tions, see: (a) Desiraju, G. R. Acc. Chem. Res. 1991, 24, 290-296. (b)
Desiraju, G. R. Acc. Chem. Res. 1996, 29, 441-449. (c) Krishnamohan
Sharma, C. V.; Zaworotko, M. J. Chem. Commun. 1996, 2655-2656. (d)
Steiner, T. Chem. Commun. 1997, 727-734. (e) Berger, I.; Egli, M. Chem.s
Eur. J. 1997, 3, 1400-1404. (f) Bodige, S. G.; Rogers, R. D.; Blackstock,
S. C. Chem. Commun. 1997, 1669-1670.
(5) For accounts and reviews on [C-H‚‚‚π] interactions, see: (a) Nishio,
M.; Umezawa, Y.; Hirota, M.; Takeuchi, Y. Tetrahedron 1995, 51, 8665-
8701. (b) Umezawa, Y.; Tsuboyama, S.; Honda, K.; Uzawa, J.; Nishio, M.
Bull. Chem. Soc. Jpn. 1998, 71, 1207-1213. (c) Nishio, M.; Hirota, M.;
Umezawa, Y. The [C-H‚‚‚π] Interaction. EVidence, Nature, and Conse-
quences; Wiley: New York, 1998.
(6) (a) Anelli, P. L.; Ashton, P. R.; Ballardini, R.; Balzani, V.; Delgado,
M.; Gandolfi, M. T.; Goodnow, T. T.; Kaifer, A. E.; Philp, D.; Pietrasz-
kiewicz, 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,
193-218. (b) Asakawa, M.; Dehaen, W.; L’abbe´, G.; Menzer, S.; Nouwen,
J.; Raymo, F. M.; Stoddart, J. F.; Williams, D. J. J. Org. Chem. 1996, 61,
9591-9595.
The efficient template-directed syntheses2 of [2]- and [3]ca-
tenanes, incorporating π-electron-rich and π-electron-deficient
aromatic residues, rely principally upon the stabilizing and
cooperative effects of π-π stacking,3 [C-H‚‚‚O] hydrogen-
bonding,4 and [C-H‚‚‚π] interactions.5 The bipyridinium-based
tetracationic cyclophanes cyclobis(paraquat-p-phenylene)6 and
cyclobis(paraquat-4,4′-biphenylene)7 self-assemble8 around one
* Corresponding authors. Address correspondence to either K.N.H. or
J.F.S. at the University of California.
†University of California.
‡ Imperial College.
§ University of Birmingham.
(1) Molecular Meccano. 47. For part 46, see: Ashton, P. R.; Ballardini,
R.; Balzani, V.; Baxter, I.; Credi, A.; Fyfe, M. C. T.; Gandolfi, M. T.;
Go´mez-Lo´pez, M.; Martinez-Diaz, M.-V.; Piersanti, A.; Spencer, N.;
Stoddart, J. F.; Venturi, V.; White, A. J. P.; Williams, D. J. J. Am. Chem.
Soc. 1998, 120, 11932-11942.
(2) For accounts and reviews on the template-directed syntheses of
catenanes incorporating complementary π-electron-rich and π-electron-
deficient components, see: (a) Gillard, R. E.; Raymo, F. M.; Stoddart, J.
F. Chem.sEur. J. 1997, 3, 1933-1940. (b) Fyfe, M. C. T.; Stoddart, J. F.
Acc. Chem. Res. 1997, 30, 393-401. (c) Raymo, F. M.; Stoddart, J. F.
Chemtracts 1998, 11, 491-511. (d) Hamilton, D. G.; Davies, J. E.; Prodi,
L.; Sanders, J. K. M. Chem.sEur. J. 1998, 4, 608-620.
(3) For accounts and reviews on π-π stacking interactions, see: (a)
Hunter, C. A.; Sanders, J. K. M. J. Am. Chem. Soc. 1990, 112, 5525-
5534. (b) Hunter, C. A. Angew. Chem., Int. Ed. Engl. 1993, 32, 1584-
1586. (c) Cozzi, F.; Siegel, J. S. Pure Appl. Chem. 1995, 67, 683-689. (d)
Shetty, A. S.; Zhang, J.; Moore, J. S. J. Am. Chem. Soc. 1996, 118, 1019-
1027.
(7) (a) Asakawa, M.; Ashton, P. R.; Menzer, S.; Raymo, F. M.; Stoddart,
J. F.; White, A. J. P.; Williams, D. J. Chem.sEur. J. 1996, 2, 877-893.
(b) Ashton, P. R.; Menzer, S.; Raymo, F. M.; Shimizu, G. K. H.; Stoddart,
J. F.; Williams, D. J. Chem. Commun. 1996, 487-490.
10.1021/ja982748b CCC: $18.00 © 1999 American Chemical Society
Published on Web 02/09/1999

1480 J. Am. Chem. Soc., Vol. 121, No. 7, 1999
Houk et al.
Scheme 1
this paper, we report the experimental details for the formation
of the [3]catenanes 2‚4PF6 and 4‚4PF6 and attempt to explain
the failure to obtain the [4]catenanes 1‚8PF6 and 3‚8PF6. The
analyses of the crystal structures, combined with conformational
analyses using force field and ab initio quantum mechanical
calculations, have shown that the combined strength of numer-
ous [C-H‚‚‚O] hydrogen bonds can be remarkably large and
indeed much greater than π‚‚‚π stacking effects. Previous
suggestions6,9 of the importance of [C-H‚‚‚O] hydrogen bond-
ing in related systems have not provided quantitative measures
of the magnitude of these interations. Here, we report an
example where complex formation cannot occur in the absence
of [C-H‚‚‚O] hydrogen bonding. The role of [C-H‚‚‚O]
hydrogen bonding in biological systems and self-assembly
continues to receive increasing recognition and acceptance.4
Figure 1. Retrosynthetic analysis of [4]catenanes each incorporating
two macrocyclic polyethers and two tetracationic cyclophanes. Note
the similarities between the cavities of the macrocyclic polyether
1/5DN38C10 and the [3]catenanes 2‚4PF6 and 4‚4PF6.
two bipyridinium units to give cyclobis(paraquat-4,4′-ter-
phenylene). Thus, the self-assembly of the [4]catenane 1‚8PF6
should occur after the formation of the [3]catenane intermediate
2‚4PF6 which possesses a cavity large enough to permit the
catenation of a fourth ring. However, all our numerous attempts
to self-assemble the [4]catenane 1‚8PF6 by conventional template-
directed approaches have so far failed. The compound that did
result from the attempted template-directed synthesis was the
[3]catenane 2‚4PF6. Similarly, when the central phenylene ring
of the terphenylene units present in the reactants and products
is replaced by an acetylenic linkage, the outcome is exactly the
samesnamely, the [4]catenane 3‚8PF6 is not formed and it is
the [3]catenane 4‚4PF6 which is isolated. Furthermore, the
[3]catenanes 2‚4PF6 and 4‚4PF6 do not bind the π-electron-
deficient bipyridinium-based substrates paraquat bis(hexafluo-
rophosphate), [PQT][PF6]2, and diquat bis(hexafluorophosphate),
[DQT][PF6]2 (Figure 1), in solution. By contrast, the macrocyclic
polyether 1/5DN38C10, which possesses a sterically accessible
π-electron-rich cavity remarkably similar to those of the
[3]catenanes, strongly binds [PQT][PF6]2 and [DQT][PF6]2. In
Results and Discussion
[3]Catenane 2‚4PF6 Incorporating Cyclobis(paraquat-4,4′-
terphenylene). The salt 5‚2PF6 was synthesized (Scheme 1) in
two steps. Bromination of 4,4′-dimethyl-1,4-terphenyl, 6, af-
forded the dibromide 7, which was reacted subsequently with
4,4′-bipyridine to give 5‚2PF6 in a yield of 75%, after counterion
exchange. Reaction (Scheme 2) of 5‚2PF6 with 7 in the presence
of the macrocyclic polyether 1,5-dinaphtho-38-crown-10 (1/
5DN38C10) gave the [3]catenane 2‚4PF6sincorporating two
macrocyclic polyethers and one tetracationic cyclophane
componentsin a yield of 5% after counterion exchange.
Surprisingly, the expected [4]catenane 1‚8PF6sincorporating
two macrocyclic polyethers and two tetracationic cyclophane
componentsswas not formed. Even when the reaction was
repeated using ultrahigh-pressure conditions (12 kbar), only the
[3]catenane 2‚4PF6 was isolated, this time in the slightly better
yield of 10%.
(8) For accounts and reviews on self-assembly processes, see: (a)
Stoddart, J. F. in Chirality in Drug Design and Synthesis; Brown, C., Ed.;
Academic Press: London, 1990; pp 53-81. (b) Mallouk, T. E.; Lee, H. J.
Chem. Educ. 1990, 67, 829-834. (c) Constable, E. C. Angew. Chem., Int.
Ed. Engl. 1991, 30, 1450-1451. (d) Lindsey, J. S. New J. Chem. 1991, 15,
153-180. (e) Menger, F. M.; Lee, S. S.; Tao, X. AdV. Mater. 1995, 7,
669-671. (f) Ghadiri, M. R. AdV. Mater. 1995, 7, 675-677. (g) Hunter,
C. A. Angew. Chem., Int. Ed. Engl. 1995, 34, 1079-1081. (h) Lawrence,
D. S.; Jiang, T.; Levett, M. Chem. ReV. 1995, 95, 2229-2260. (i) Raymo,
F. M.; Stoddart, J. F. Curr. Opin. Colloid Interface Sci. 1996, 1, 116-126.
(j) Philp, D.; Stoddart, J. F. Angew. Chem., Int. Ed. Engl. 1996, 35, 1154-
1196.
(9) (a) Castro, R.; Nixon, K. R.; Evanseck, J. D.; Kaifer, A. E. J. Org.
Chem. 1996, 61, 7298-7303. (b) Asakawa, M.; Brown, C. L.; Menzer, S.;
Raymo, F. M.; Stoddart, J. F.; Williams D. J. J. Am. Chem. Soc. 1997,
119, 2614-2627.

[C-H‚‚‚O] Interactions in Supramolecular Complexes
J. Am. Chem. Soc., Vol. 121, No. 7, 1999 1481
Scheme 2
Figure 2. X-ray crystal structure of the [3]catenane 2‚4PF6.
symmetry, with the two 1,5-dinaphtho-38-crown-10 ringsseach
of which have local D2 symmetrysencircling the bipyridinium
units of the extended tetracationic cyclophane. This molecular
geometry is stabilized by a combination11 of π-π stacking,
[C-H‚‚‚O] hydrogen-bonding, and [C-H‚‚‚π] interactions,
similar to those observed2 in analogous [2]- and [3]catenanes.
A consequence of the local D2 symmetry of each macrocyclic
polyether component is a marked difference in the inclination
of the O‚‚‚O vectors of the 1,5-dioxynaphthalene rings with
respect to the N‚‚‚N axes of the encircled bipyridinium units.
The “inside” O‚‚‚O vector is inclined by 124°, while the
“alongside” O‚‚‚O vector is inclined by only 10° to the N‚‚‚N
axes. The tetracationic cyclophane displays a rigid open
geometry with a length (between the bipyridinium units) of ca.
15.10 Å and a breadth (between the centroids of the middle
phenylene rings of the terphenyl linkages) of 11.20 Å. A key
parameter of this molecular structure is the size of the void
formed between the two inside 1,5-dioxynaphthalene residues:
the mean interplanar separation is 7.93 Å, a distance that is ca.
0.9 Å greater than the separation observed between the 1,5-
dioxynaphthalene residues within each 1/5DN38C10 macro-
cycle.
Inspection of the packing of the [3]catenane molecules
revealed an absence of any intercatenane π-π stacking interac-
tions involving the bipyridinium units. However, symmetry-
related tetracationic cyclophanes in the crystallographic c
direction have their mean planes inclined by ca. 76° and are
arranged such that the end phenyl rings of each terphenylene
spacer are π-π stacked with their symmetry-related counterparts
to create (Figure 3) a folded mosaic-like array. The associated
interplanar separations and centroid-centroid distances are 3.50
and 3.80 Å, respectively.
The [3]catenane 2‚4PF6 was characterized unequivocally by
low- and high-resolution liquid secondary ion mass spectrometry
1
(LSIMS), H NMR spectroscopy, and X-ray crystallographic
analysis. The LSIMS spectrum revealed three peaks centered
on m/z 2534, 2389, and 2244 for [M - PF6]+, [M - 2PF6]+,
and [M - 3PF6]+, respectively, corresponding to the losses of
one, two, and three PF6- counterions, respectively. The 1H NMR
spectrum recorded in CD3CN at 350 K shows sharp and well-
resolved signals for the π-electron-rich and the π-electron-
deficient macrocyclic components. In particular, only one set
of signals is observed for each pair of naphthalene protons,
indicating that the circumrotation of the macrocyclic polyethers
through the cavity of the tetracationic cyclophane is fast on the
1H NMR time scale at 350 K. Similarly, the protons located in
the R- and â-positions, with respect to the nitrogen atoms, on
the bipyridinium units resonate as two doublets onlysi.e., the
local C2h symmetry which could be imposed by the 1,5-
dioxynaphthalene units is “lost” at this temperature as a result
of the fast dynamic processes on the 1H NMR time scale
involving their exchange or reorientation.10 On cooling of the
CD3CN solution of the [3]catenane down to 290 and then to
253 K, the circumrotation of the π-electron-rich through the
π-electron-deficient macrocyclic components becomes slow and
significant broadening of the resonances associated with the
naphthalene and bipyridinium protons is observed. Unfortu-
nately, as a result of the very low solubility of the [3]catenane
in solvents having freezing points lower than that of CD3CN,
1H NMR spectra of 2‚4PF6 at temperatures lower than 253 K
could not be recorded, thus preventing the determination of the
energy barrier associated with this circumrotation process.
The single-crystal X-ray analysis of the [3]catenane 2‚4PF6
shows (Figure 2) that the molecule has crystallographic Ci
The cavity within the [3]catenane 2‚4PF6 is lined by the
π-electron-rich 1,5-dioxynaphthalene units and is large enough,
at least in principle, to accommodate π-electron-deficient guests
such as [PQT][PF6]2 and [DQT][PF6]2. However, the 1H NMR
spectra of equimolar CD3CN solutions of the [3]catenane with
either [PQT][PF6]2 or [DQT][PF6]2 did not show any significant
(11) The mean interplanar separations between both the “inside” and
“alongside” 1,5-dioxynaphthalene residues and the bipyridinium unit are
in each case 3.5 Å. There are [C-H‚‚‚O] hydrogen bonds (i) between one
of the hydrogen atoms in the R-positions with respect to the nitrogen atoms
on the bipyridinium units and a central polyether oxygen atom ([C‚‚‚O],
3.38 Å; [H‚‚‚O], 2.46 Å; [C-H‚‚‚O], 161°) and (ii) between one of the
methylene groups of the tetracationic cyclophane and the adjacent oxygen
atom of the same polyether linkage ([C‚‚‚O], 3.30 Å: [H‚‚‚O], 2.45 Å;
[C-H‚‚‚O], 148°). Additionally, there are edge-to-face [C-H‚‚‚π] interac-
tions involving the hydrogen atoms in the 4- and 8-positions of the “inside”
1,5-dioxynaphthalene residues and the proximal p-phenylene rings of the
tetracationic cyclophane (respectively: [H‚‚‚π], 2.73 and 2.82 Å; [C-H‚‚‚π],
145 and 141°).
(10) (a) Asakawa, M.; Ashton, P. R.; Boyd, S. E.; Brown, C. L.; Gillard,
R. E.; Kocian, O.; Raymo, F. M.; Stoddart, J. F.; Tolley, M. S.; White, A.
J. P.; Williams, D. J. J. Org. Chem. 1997, 62, 26-37. (b) Bravo, J. A.;
Raymo, F. M.; Stoddart, J. F.; White, A. J. P.; Williams, D. J. Eur. J. Org.
Chem. 1998, 2565-2571.

1482 J. Am. Chem. Soc., Vol. 121, No. 7, 1999
Houk et al.
Figure 4. X-ray crystal structure of the [3]catenane 4‚4PF6.
tenane 4‚4PF6 in yields of 10 and 35%, respectively, after
counterion exchange. Again, the desired [4]catenane 3‚8PF6 was
not formed. Similarly, when the reaction was performed under
ultrahigh-pressure conditions (12 kbar), the [3]catenane 4‚4PF6
was isolated in a yield of 42%, together with trace amounts of
the [2]catenane. Again no [4]catenane was detected.
1
The [3]catenane 4‚4PF6 was characterized by LSIMS, H
NMR spectroscopy, and X-ray crystallography. In addition to
the molecular ion peak at m/z 2574 for [M]+, the LSIMS
spectrum of 4‚4PF6 showed three peaks centered on m/z 2429,
2284, and 2139 for [M - PF6]+, [M - 2PF6]+, and [M -
3PF6]+, respectively, corresponding to the losses of one, two,
and three PF6- counterions, respectively. The 1H NMR spectrum
of the [3]catenane 4‚4PF6, recorded in (CD3)2CO/CD3NO2 (9:
1, v/v) at 300 K, revealed only one set of signals for the three
pairs of naphthalene protons. In addition, the protons in the R-
and â-positions with respect to the nitrogen atoms on the
bipyridinium units resonate as two doublets only. As before,
the relative simplicity of this spectrum is a result of circum-
rotations of the macrocyclic polyethers through the cavity of
the tetracationic cyclophane which are fast on the 1H NMR time
scale at room temperature. Thus, the “inside” and “alongside”
1,5-dioxynaphthalene units give rise to only one set of signals
and cannot be distinguished. On cooling of the solution down
to 280 K, the rate of the circumrotation process becomes slower
Figure 3. Packing diagram of the [3]catenane 2‚4PF6. The macrocyclic
polyether components have been omitted for clarity.
difference from the 1H NMR spectra of the separate compounds,
even at temperatures as low as 253 Ksi.e., the lowest
temperature at which an 1H NMR spectrum of the [3]catenane
2‚4PF6 can be recorded (vide supra). Similarly, even when a
5-fold excess of the “potential” guest was employed, no changes
1
in the H NMR spectra were detected. Consistently, the UV-
vis spectra of the same solutions did not reveal any changes in
the absorbance and/or wavelength of the charge-transfer band
associated with the intracatenane donor-acceptor interactions.
In addition, we attempted also to “catenate” the [3]catenane 2‚
4PF6 with a second cyclobis(paraquat-4,4′-terphenylene) tetra-
cation by treating 2‚4PF6 with an excess of 5‚2PF6 and 7 in the
hope of generating the [4]catenane 1‚8PF6. However, only the
original [3]catenane 2‚4PF6 and some polymeric material
resulting from the reaction of the 5‚2PF6 with 7 were isolated
from the reaction mixture. These results show that, contrary to
reasonable expectations, the complexation of a π-electron-
deficient guest “inside” the cavity of the [3]catenane 2‚4PF6
does not occur. It was noted, however, that the separation
between the “inside” 1,5-dioxynaphthalene unitssi.e., the
distance between the recognition sites lining the cavity of the
potential receptorsmight be too large, judging by the solid-
state structure at least. It was, therefore, decided to try shortening
the distance between the two bipyridinium units in the tetra-
cationic cyclophane component by replacing the central phen-
ylene ring in both terphenylene units by an acetylenic linkage
in the hope of generating a [3]catenane possessing an internal
cavity of a more appropriate size for the binding of π-electron-
deficient guests.
1
and significant broadening of all signals is observed in the H
NMR spectrum. Further cooling to 210 K results in a more
complex 1H NMR spectrum which now shows two distinct sets
of signals for the “inside” and “alongside” 1,5-dioxynaphthalene
units. In particular, NOE experiments revealed that the protons
H4/8 of the “inside” 1,5-dioxynaphthalene units resonate at δ
4.15. In addition, the resonances associated with the protons in
the R-positions with respect to the nitrogen atoms on the
bipyridinium units separate into two sets of signals as a result
of the local C2h symmetry of the 1,5-dioxynaphthalene units.
By employment of the approximate coalescence treatment,12
a
free energy barrier of 12.5 kcal mol-1 was determined for the
circumrotation process at the coalescence temperature (280 K)
of the probe H4/8 protons of the 1,5-dioxynaphthalene units.
The single-crystal X-ray analysis of the [3]catenane 4‚4PF6
revealed (Figure 4) a Ci symmetric structure very similar to that
observed for 2‚4PF6, though with the desired shortening of the
dimensions of the tetracationic cyclophane and hence of the
separation of the “inside” 1,5-dioxynaphthalene units. The length
of the cyclophane is reduced to 13.47 Å, whereas the breadth
is unchanged at 11.15 Å. The accompanying interplanar
The [3]Catenane 4‚4PF6 Incorporating Cyclobis(paraquat-
4,4′-tolan). The salt 8‚2PF6 incorporating a tolan spacer was
prepared (Scheme 1) in two steps. The bromination of 9 was
followed by the reaction of the resulting dibromide 10 with 4,4′-
bipyridine, affording 8‚2PF6 in a yield of 50%, after counterion
exchange. The reaction (Scheme 2) of 8‚2PF6 with 10 in the
presence of the macrocyclic polyether 1/5DN38C10 afforded a
[2]catenanesincorporating one macrocyclic polyether and one
tetracationic cyclophane componentsin addition to the [3]ca-
(12) (a) Sutherland, I. O. Annu. Rep. NMR Spectrosc. 1971, 4, 71-235.
(b) Sandstro¨m, J. Dynamic NMR Spectroscopy; Academic Press: London,
1982.

[C-H‚‚‚O] Interactions in Supramolecular Complexes
J. Am. Chem. Soc., Vol. 121, No. 7, 1999 1483
Figure 5. Packing diagram of the [3]catenane 2‚4PF6.
separation between the “inside” dioxynaphthalene rings is
reduced to 6.77 Å, a value which should be near optimal for
the insertion of a π-electron-deficient guest. The replacement
of the terphenylene by tolan units within the tetracationic
cyclophane component is also accompanied by a marked change
in the conformation of one of the polyether linkages which
results in a minor alteration in the relative tilts of the O‚‚‚O
vectors of the “inside” and “alongside” 1,5-dioxynaphthalene
residues with respect to the N‚‚‚N axes of the encircled
bipyridinium units: the “inside” O...O vector is inclined by 128°
while the “alongside” O‚‚‚O vector is inclined by only 5°; cf.
124 and 10°, respectively in 2‚4PF6. Stabilization is again
achieved via a combination of π-π, [C-H‚‚‚O], and [C-H‚‚‚π]
interactions.13 The [3]catenane molecules form extended stacks
(Figure 5) with the planes of the acetylenic spacers of adjacent
lattice-translated molecules being ca. 3.28 Å apart; the shortest
intermolecular atom-atom contact is 3.43 Å between one of
the acetylenic carbon atoms of one molecule and one of the
quaternary aromatic carbon atoms in the diphenylacetylenic
spacer of the next and vice versa.
Figure 6. Fragmentation of the complex [1/5DN38C10][PQT]2+
performed in order to “isolate” the [C-H‚‚‚O] hydrogen bonds from
the π-π stacking interactions as well as the superimposed (black) solid-
state and (gray) calculated (simulated annealing, AMBER* force field)
geometries of the complex.
The single-crystal X-ray analysis of the corresponding complex
revealed (Figure 6) a pseudorotaxane geometry with the
bipyridinium unit of the guest sandwiched between the 1,5-
dioxynaphthalene units of the host. In addition, short distances
between the acidic hydrogen atoms in the R-positions with
respect to the nitrogen atoms on the bipyridinium units and some
of the polyether oxygen atoms were also noted. These observa-
tions suggested that the formation of this complex is a result of
π-π stacking interactions between the complementary aromatic
units as well as of [C-H‚‚‚O] hydrogen bonds. This solid-state
geometry was used as the input structure for a simulated
annealing molecular dynamics study using the AMBER* force
field15 in conjunction with the GB/SA solvation model16 for
H2O, as implemented in Macromodel.17 The calculated and
solid-state geometries differ (Figure 6) significantly in the
relative orientation of the aromatic units of host and guest. In
the calculated structure, the angles between the N‚‚‚N axis of
the bipiridinium unit of the guest and the O‚‚‚O axes of the
1,5-dioxynaphthalene units of the host are 126 and 138°
whereas, in the centrosymmetric solid-state structure, the angle
is 19°. To analyze the effect that the relative orientation of the
complementary aromatic units has upon the binding energy, the
model dimer complex shown in Figure 7a was employed. This
complex was constructed with Spartan18 by orienting the
bipyridinium and 1,5-dioxynaphthalene units parallel to each
other with a centroid-centroid separation19 of 3.51 Å. The angle
θ was varied from 0 to 180° in consecutive steps of 10° each,
Despite this “optimization” of the distance separating the
π-electron-rich 1,5-dioxynaphthalene recognition sites lining the
cavity of the [3]catenane, again, no complexation of either
1
[PQT][PF6]2 or [DQT][PF6]2 was detected by H NMR spec-
troscopy even at temperatures as low as 210 K. Binding
experiments were also carried out with neutral π-electron-
deficient guests, such as 1,4-benzoquinone and perfluorohydro-
quinone as alternatives to the positively charged [PQT][PF6]2
and [DQT][PF6]2 guests. However, no changes in the chemical
shifts or in the absorbance and/or wavelength of the charge-
transfer band associated with the [3]catenane were detected by
1H NMR and UV-vis spectroscopies, respectively.
Computational Evaluation of Binding Interactions. The
π-electron-rich macrocyclic polyether 1/5DN38C10 binds10,14
the π-electron-deficient bipyridinium salt [PQT]2+ in solution
(Ka ) 1190 M-1 in CD3CN at 25 °C) and in the solid state.
(13) The mean interplanar separations between both the “inside” and
“alongside” 1,5-dioxynaphthalene residues and the bipyridinium unit are
3.33 and 3.35 Å. There are [C-H‚‚‚O] hydrogen bonds (i) between
diametrically opposite pairs of hydrogen atoms located in the R-positions
with respect to the nitrogen atoms on the bipyridinium units and the oxygen
atoms in the proximal polyether linkages (respectivley: [C‚‚‚O], 3.26, 3.30,
and 3.23 Å; [H‚‚‚O], 2.43, 2.44, and 2.45 Å; [C-H‚‚‚O], 145, 148, and
138°) and (ii) between one of the methylene groups of the tetracationic
cylophane and one of the oxygen atoms in one of the polyether linkages
([C‚‚‚O], 3.12 Å; [H‚‚‚O], 2.39 Å; [C-H‚‚‚O], 133°). Additionally, there
are two edge-to-face [C-H‚‚‚π] interactions involving the “inside” 1,5-
dioxynaphthalene residues and the proximal p-phenylene rings of the
tetracationic cyclophane ([H‚‚‚π], 2.72 and 2.76 Å; [C-H‚‚‚π], 148 and
145°).
(15) Weiner, S. J.; Kollman, P. A.; Case, D. A.; Singh, V. C.; Ghio, C.;
Alagona, G.; Profeta, S., Jr.; Weiner, P. J. Am. Chem. Soc. 1984, 106, 765-
784.
(16) Still, W. C.; Tempczyk, A.; Hawley, R. C.; Hendrickson, T. J. Am.
Chem. Soc. 1990, 112, 6127-6129.
(17) Mohamadi, F.; Richards, N. G. J.; Guida, W. C.; Liskamp, R.;
Lipton, M.; Caufield, C.; Chang, G.; Hendrickson, T.; Still, W. C. J. Comput.
Chem. 1990, 11, 440-467.
(18) Spartan 4.1; Wavefunction Inc.: 18401 Von Karman Ave., 370
Irvine, CA 92715.
(19) A statistical survey of the Cambridge Crystallographic Database
(Allen, F. H.; Davies, J. E.; Galloy, J. J.; Johnson, O.; Kennard, P.; Macrae,
C. F.; Watson D. G. J. Chem. Inf. Comput. Sci. 1991, 31, 187-204) revealed
that the average centroid-centroid separation between stacked 1,5-
dioxynaphthalene and bipyridinium units is 3.51 Å.
(14) Ashton, P. R.; Chrystal, E. J. T.; Mathias, J. P.; Parry, K. P.; Slawin,
A. M. Z.; Spencer, N.; Stoddart, J. F.; Williams, D. J. Tetrahedron Lett.
1987, 28, 6367-6370.

1484 J. Am. Chem. Soc., Vol. 121, No. 7, 1999
Houk et al.
Figure 8. (a) Calculated (simulated annealing, AMBER* force field)
geometry of the complex [2]4+[PQT]2+. (b) Calculated (PRCG mini-
mization, AMBER* force field) geometry of the complex [4]4+[PQT]2+
.
[C-H‚‚‚O] hydrogen bonds. More precisely, the highly electron-
deficient bipyridinium protons interact strongly with the negative
electrostatic potential presented by the polyether oxygen atoms.
Deleting the polyether chains instead (I-III, Figure 6) results
in interaction energies18 of -21 and -23 kcal mol-1 for the
aromatic‚‚‚aromatic interactions in the solid-state and calculated
parent structures, respectively. These interaction energies result
from π-π stacking interactions. These observations suggest that
the contribution of the [C-H‚‚‚O] hydrogen bonds to the oVerall
binding energy is approximately 4 times higher than that
associated with the π-π stacking interactions!21 The interaction
energies are enormous for the gas-phase calculations with a
small basis set. Nevertheless, the very different magnitudes of
[C-H‚‚‚O] and [π‚‚‚π] interactions will be maintained in
solution.
Figure 7. (a) The model dimer complex employed to evaluate the
effect of the relative orientation of the aromatic units on the binding
energy. (b) Plot of the calculated (HF/3-21G*) binding energies of the
model dimer complex against the angle θ. (c) HF/3-21G* electrostatic
potential isosurface (isovalue ) -35 kcal mol-1) of 1,5-dimethoxy-
naphthalene.
and the geometries were optimized (HF/3-21G) at each step
constraining both the angle θ and the centroid-centroid
separation. The plot of the resulting binding energies ∆E against
the angle θ shows (Figure 7b) a difference of 6.0 kcal mol-1
between the most stable orientation (θ ) 10°) and the least
favorable one (θ ) 70°). The negative electrostatic potential
presented (Figure 7c) by the methoxy oxygen atoms is best
complemented by the highly electron-deficient pyridinium rings
when the angle θ is 10°. A nearly perpendicular orientation (θ
) 70°) of the O‚‚‚O and N‚‚‚N axes moves the negative and
positive regions apart destabilizing the complex.
To establish the relative contribution of the π-π stacking
and [C-H‚‚‚O] hydrogen-bonding interactions to the overall
binding energy of the complex [1/5DN38C10][PQT]2+, a series
of single-point HF/3-21G calculations were performed on model
compounds. The solid-state and calculated geometries of the
complex [1/5DN38C10][PQT]2+ were employed (Figure 6)
individually as the model complex structures. The 1,5-dioxy-
naphthalene rings were deleted without altering the relative
orientation of the remaining components. The resulting “com-
plex” (IV) and its separate “host” (V) and “guest” (VI)
components were subjected individually to single-point HF/3-
21G calculations. Subtracting the calculated energy of IV from
those of V and VI gave interaction energies20 of -82 and -78
kcal mol-1 for the solid-state and calculated “parent” structures,
respectively. These large interaction energies result from the
In a second set of model calculations, the solid-state structures
provided by the single-crystal X-ray analyses of the [3]catenanes
2
4+ and 44+ were used to construct the complexes between each
of the [3]catenanes and [PQT]2+. These complexes were
generated by introducing manually the guest into the cavities
of the hosts with Macromodel.17 The geometries of the resulting
complexes were optimized by employing the PRCG algorithm,22
and then they were subjected to simulated annealing using the
AMBER* force field and the GB/SA solvation model for H2O.
The global minimum found (Figure 8a) for the complex
[2]4+[PQT]2+ shows the guest to be inserted through the cavity
of the [3]catenane. By contrast, in the case of 44+, the guest is
(20) No significant differences were observed in the values of the
interaction energies when these were calculated with the inclusion of chloride
counterions.
(21) A similar effect was observed experimentally in a series of related
pseudorotaxanes incorporating bipyridinium and 1,4-dioxybenzene recogni-
tion sites: (a) 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, 193-218. (b) Asakawa, M.; Brown, C. L.; Menzer, S.; Raymo,
F. M.; Stoddart, J. F.; Williams, D. J. J. Am. Chem. Soc. 1997, 119, 2614-
2627.
(22) Polak, E.; Ribiere, G. ReV. Fr. Inf. Rech. Oper. 1969, 16-R1, 35-
43.

[C-H‚‚‚O] Interactions in Supramolecular Complexes
J. Am. Chem. Soc., Vol. 121, No. 7, 1999 1485
Figure 9. Fragmentation of the complexes [2]4+[PQT]2+ and
[4]4+[PQT]2+ performed in order to “isolate” the [C-H‚‚‚π] from the
π-π stacking interactions.
Figure 10. (a) The model dimer complex employed to evaluate the
effect of the relative orientation of the bipyridinium units on the binding
energy. (b) Plot of the calculated (single point, HF/3-21G*) binding
energies of a model dimer complex against the angle θ.
expelled from the cavity of the [3]catenane and the global
minimum obtained shows the guest well-separated from the host.
The geometry calculated for the complex [2]4+[PQT]2+
suggests that the hydrogen atoms in the R-positions with respect
to the nitrogen atoms on the bipyridinium units may be engaged
in [C-H‚‚‚π] interactions with the central aromatic ring of the
terpenylene spacers of the [3]catenane. To compare the relative
contribution of these [C-H‚‚‚π] interactions with that arising
from the π-π stacking interactions to the binding event, the
geometry of [2]4+[PQT]2+ calculated by simulated annealing
was employed as the “parent” structure from which a series of
model compounds (I-VI, Figure 9) were generated with
Spartan. Once again, the [PQT]2+ guest and the central aromatic
rings of the terphenylene spacers were extracted from the
“parent” structure, while their relative orientations were main-
tained. The energies of this “complex” (IV) and of its separate
“host” (V) and “guest” (VI) components were derived by
performing single-point HF/3-21G* calculations on each species.
The resulting energy difference20 (-16 kcal mol-1) arises mainly
from faVorable [C-H...π] interactions. Similarly (I-III, Figure
9), the contribution of the π-π stacking interactions20 (-22
kcal mol-1) was determined. The same procedure was applied
to [4]4+[PQT]2+ by employing the geometry (Figure 8b)
calculated after the PR/CG minimization as the “parent”
structure. Values of -10 and -16 kcal mol-1 were obtained20
for the [C-H‚‚‚π] and π-π stacking interactions, respectively.
Thus, on going from the complex [1/5DN38C10][PQT]2+ to
[2]4+[PQT]2+ and [4]4+[PQT]2+, the π-π stacking interactions
remain almost unaltered. However, the [C-H‚‚‚O] hydrogen
bonds are lost in faVor of [C-H‚‚‚π] interactions, which are
more than 4 times weaker.
oriented, within the build mode of Spartan, parallel to each other
with a centroid-centroid separation of 7.68 Åsi.e., the distance
separating the bipyridinium unit of the guest from one of the
bipyridinium units of the host in the complex [2]4+[PQT]2+
shown in Figure 8a. The angle θ was varied from 0 to 90° in
consecutive steps of 10° each, and single-point HF/3-21G
calculations were perfomed at each step. The plot of the resulting
binding energies ∆E against the angle θ shows (Figure 10b)
that energetically unfavorable orientations lie in the region
between 0 and 45°. By contrast, energetically favorable orienta-
tions are achieved when the value of θ is between 45 and 90°.
However, the associated interaction energies are significantly
lower than those calculated (vide supra) for the [C-H‚‚‚O],
[π‚‚‚π], and [C-H‚‚‚π] interactions, suggesting that their
contribution to the overall binding energy is relatively small.
Conclusions
To construct multiply interlocked [4]catenanes, we have
devised a two-stage self-assembly approach involving the
formation of [3]catenane intermediates possessing internal
cavities sufficiently large to allow the subsequent interlocking
of a fourth macrocyclic component. These [3]catenanes incor-
porate two 1,5-dioxynaphthalene macrocyclic components and
one tetracationic cyclophane component composed of two
bipyridinium recognition sites separated by two long spacerss
namely, terphenyl or tolan units. Surprisingly, however, the
expected [4]catenanes were not obtained and only the [3]-
catenane intermediates were isolated. Their single-crystal X-ray
analyses, indeed, revealed the presence of large cavities lined
by π-electron-rich 1,5-dioxynaphthalene units which seem to
be ideal for the inclusion of π-electron-deficient bipyridinium-
In addition to the decrease of stabilization energy arising from
the loss of [C-H‚‚‚O] hydrogen bonds in favor of weaker
[C-H‚‚‚π] interactions, the complex formation between the
tetracationic [3]catenanes 24+ and 44+ and the dicationic guest
[PQT]2+ might be disfavored also by repulsive electrostatic
interactions between the charged bipyridinium units of host and
guest. To analyze the effect of these interactions, a model dimer
complex composed of two identical [PQT][Cl]2 units (Figure
10a) was employed. The geometry of [PQT][Cl]2 was optimized
at the HF/3-21G level, and then two [PQT][Cl]2 units were
1
based guests. However, the solution H NMR and UV-vis
spectra of the [3]catenanes did not change significantly upon
the addition of either [PQT][PF6]2 or [DQT][PF6]2. A compu-
tational investigation indicates that the loss of favorable
[C-H‚‚‚O] hydrogen-bonding interactions prevents complex
formation. The magnitude of these interactions is very large.
The precise evaluation of the magnitude of these effects, now
known to be much greater than originally appreciated, are under

1486 J. Am. Chem. Soc., Vol. 121, No. 7, 1999
Houk et al.
[3]Catenane 2‚4PF6. Method A. The bis(hexafluorophosphate) salt
5‚2PF6 (50 mg, 0.06 mmol), the dibromide 6 (24 mg, 0.06 mmol), and
the macrocyclic polyether 1/5DN38C10 (140 mg, 0.22 mmol) sus-
pended in MeCN (10 mL) were stirred for 2 weeks at ambient
temperature. The mixture was filtered, and the solvent was removed
in vacuo. The resulting red solid was purified by column chromatog-
raphy [SiO2, 2 M NH4Cl/MeOH/MeNO2 (7:2:1, v/v/v)]. The fractions
containing the desired product were combined, and the solvent was
evaporated under reduced pressure. The residue was dissolved in H2O,
and a saturated aqueous solution of NH4PF6 was added to afford a red
precipitate. Crystallization of the resulting red solid by vapor diffusion
of i-Pr2O into a MeCN solution of the product gave the [3]catenane
2‚4PF6 as a red crystalline solid (6 mg, 4%): dec >210 °C; HRMS
m/z (LSIMS) calcd for [M - PF6]+ C132H136F18N4O20P3, 2531.8673,
found 2531.8732; LSIMS m/z 2534 [M - PF6]+, 2389 [M - 2PF6]+,
2244 [M - 3PF6]; 1H NMR (300 MHz, CD3CN, 300 K) δ 8.75-8.69
(b d, 8H), 7.75 (d, 8H), 7.85 (d, 8H), 7.53 (s, 8H), 6.68 (d, 8H), 6.55-
6.35 (b t, 8H), 6.23 (d, 8H), 5.84 (s, 8H), 3.98-3.74 (m, 64H). Crystals
suitable for X-ray crystallographic analysis were grown by vapor
diffusion of i-Pr2O into a MeCN solution of the [3]catenane. Crystal
evaluation in our laboratories. The importance of such [C-H‚‚‚O]
interactions continues to be illustrated in many biological
situations and manifested in supramolecular systems.
Experimental Section
General Procedures. Solvents were dried according to literature
methods.23 Dimethylformamide (DMF) was distilled from calcium
hydride under reduced pressure, and acetonitrile (MeCN) was distilled
from calcium hydride under an atmosphere of N2. Tetrahydrofuran
(THF) was refluxed over sodium acetophenone ketal and distilled under
N2. 4,4′-Dimethyl-p-terphenyl 6,24 4,4′-dimethyl-p-phenylacetylene 10,25
and 1,5-dioxynaphtho-38-crown-10 (1/5DN38C10)14 were prepared
according to literature procedures. Reactions requiring ultrahigh pres-
sures were carried out in Teflon vessels using a custom-built ultrahigh-
pressure reactor manufactured by PSIKA Pressure Systems Limited,
Glossop, UK. Thin-layer chromatography (TLC) was performed on
aluminum plates coated with Merck 5554 Kieselgel 60 F254. The plates
were air-dried and examined under a UV lamp and, then, if necessary,
developed in an iodine vapor tank. Column chromatography was
performed using Kieselgel 60 (0.040-0.063 mm mesh, Merck 9385).
Melting points were determined using an Electrothermal 9200 melting
point apparatus and are uncorrected. UV-visible spectra were recorded
at 25 °C on a Perkin-Elmer Lambda 2 spectrophotometer fitted with a
thermostated cell holder and operating under microcomputer control.
Electron impact mass spectroscopy (EIMS) was carried out on either
a Kratos Profile or a VG Prospec instrument. Liquid secondary ion
mass spectrometry (LSIMS) was carrried out on a VG Zabspec mass
spectrometer. 1H nuclear magnetic resonance (NMR) spectra were
recorded on either a Bruker AC300 (300 MHz) or a Bruker AMX400
(400 MHz) spectrometer using the deuterated solvent as lock and the
residual solvent as an internal reference. 13C NMR spectra were recorded
on a Bruker AC300 (75 MHz) spectrometer using the JMOD pulse
data:
C132H136N4O20‚4PF6‚4H2O, M ) 2822.46, monoclinic, a )
22.239(3) Å, b ) 12.223(2) Å, c ) 28.287(5) Å, â ) 95.69(1)°, V )
7651(2) Å3, space group P21/c, Z ) 2 (the structure has crystallographic
Ci symmetry), Dc ) 1.225 g cm-3, µ(Cu KR) ) 1.271 mm-1, λ )
1.541 78 Å, F(000) ) 2944. Data for a crystal having dimensions of
0.35 × 0.23 × 0.18 mm were measured on a Siemens P4/RA
diffractometer with Cu KR radiation (graphite monochromated) using
ω scans at 293 K. Of the 9595 independent reflections measured (θ e
55°), 5697 had Io > 2σ(Io) and were considered to be observed. The
data were corrected for Lorentz and polarization effects, but not for
-
absorption. The structure was solved by direct methods. One PF6
counterion was disordered and subsequently split into two half-occupied
positions. The ∆F maps showed the presence of unidentified solvent
molecules. These were refined as partially occupied H2O molecules.
Hydrogen atoms for the H2O molecules were not located; all other
hydrogen atoms were placed in idealized positions with Ueq(H) )
1.2Ueq(C) and were allowed to ride on their parent atoms. Anisotropic
refinement for all non-hydrogen atoms, with the exception of those of
the partially occupied H2O molecules, gave R1 ) 0.1253 and wR2 )
0.3384. All computations were carried out using the SHELXTL 5.03
package. Atomic coordinates and bond lengths and angles have been
deposited at the Cambridge Crystallographic Data Centre.
1
sequence (assuming JCH ) 143 Hz) or the PENDANT sequence.
Elemental analyses were performed by the University of Sheffied
Microanalytical Service.
4,4′-Bis(bromomethyl)-1,4-terphenyl, 7.26 4,4′-Dimethyl-1,4-ter-
phenyl, 6 (0.52 g, 2 mmol), and N-bromosuccimide (NBS) (0.79 g,
4.4 mmol) were suspended in CCl4 (50 mL) and heated under reflux.
AIBN was added in catalytic amounts over a period of 2.5 h. After
being cooled to room temperature, the mixture was filtered, and the
solid residues were washed with boiling CCl4 (5 × 30 mL). The solvent
was then evaporated under reduced pressure, and the product was
recrystallized from MeCO2Et twice to afford a white crystalline solid
(0.3 g, 40%): mp 102-104 °C (lit.27 265 °C, PhMe); LSIMS m/z 416
Method B. The bis(hexafluorophosphate) salt 5‚2PF6 (50 mg, 0.06
mmol), the dibromide 6 (24 mg, 0.06 mmol), and the macrocyclic
polyether 1/5DN38C10 (140 mg, 0.22 mmol) were dissolved in DMF
(7 mL), and the solution was subjected to a pressure of 12 kbar for 2
days at ambient temperature. The solvent was evaporated, and the
resulting residue was purified as described in method A to afford the
[3]catenane 2‚4PF6 (20 mg, 15%).
4,4′-Bis(bromomethyl)tolan, 10.27 4,4′-Bis(methyl)tolan (0.42 g, 2
mmol) and N-bromosuccimide (NBS) (0.79 g, 4.4 mmol) were
suspended in CCl4 (50 mL), the mixture was heated and heated under
reflux. AIBN was added in catalytic amounts over a period 2.5 h. After
being cooled to room temperature, the mixture was filtered, and the
solid residue was washed with boiling CCl4 (5 × 30 mL). The solvent
was evaporated under reduced pressure, and the product was recrystal-
lized from MeCO2Et twice to afford a white crystalline solid (0.4 g,
55%): mp 110-112 °C (lit.28 187-190 °C, benzene); LSIMS m/z 416
[M]+; 1H NMR (300 MHz, CDCl3, 300 K) δ 7.55-7.50 (m, 4H), 7.47-
7.83 (m, 4H), 4.58 (s, 4H).
1
[M]+; H NMR (300 MHz, CDCl3, 300 K) δ 7.68 (s, 4H), 7.62-7.60
(m, 4H), 7.50-7.48 (m, 4H), 4.58 (s, 4H).
4,4′-Bis(4,4′-pyridylpyridiniumylmethyl)-1,4′-terphenyl Bis(hexa-
fluorophosphate), 5‚2PF6. 4,4′-Bipyridine (0.63 g, 4 mmol) was
dissolved in MeCN (200 mL), and the solution was heated under reflux.
Then, the dibromide 7 (0.2 g, 0.5 mmol) was added gradually as a
solid in increasing amounts over a period of 8 h. After being cooled to
room temperature, the mixture was filtered. The resulting solid was
dissolved in 100 mL of Me2CO/H2O (1:1, v/v), and a concentrated
solution of NH4PF6 in H2O was added gradually until no further
precipitate was observed. The precipitate was filtered off and washed
with H2O, Et2O, and CHCl3 to afford 5‚2PF6 as a white powder (0.16
1
g, 60%): dec >280 °C, LSIMS m/z 713 [M - PF6]+; H NMR [300
MHz, (CD3)2CO, 300 K] δ 9.51-9.46 (m, 4H), 8.90-8.87 (b m, 4H),
8.74-8.71 (m, 4H), 8.05-8.00 (b m, 4H), 7.91-7.88 (m, 12H), 6.23
(s, 4H); 13C NMR (75 MHz, CDCl3, 300 K) δ 155.2, 152.4, 145.8,
142.4, 141.4, 140.1, 133.0, 130.6, 128.6, 128.4, 127.1, 122.7, 64.6.
Anal. Calcd for C40H32F12N2P2‚HPF6‚H2O: C, 45.36; H, 3.35; N, 5.29.
Found: C, 45.68; H, 3.14; N, 5.38.
4,4′-Bis(4,4′-pyridylpyridiniumylmethyl)tolan Bis(hexafluoro-
phosphate), 8‚2PF6. 4,4′-Bipyridine (1.70 g, 10.89 mmol) was dis-
solved in MeCN (200 mL), and the solution was heated under reflux.
The dibromide 10 (1.98 g, 5.45 mmol) was added gradually as a solid
in increasing amounts over a period 8 h. After being cooled to room
temperature, the mixture was filtered. The resulting solid was dissolved
in 100 mL of Me2CO/H2O (1:1, v/v), and a concentrated solution of
NH4PF6 in H2O was added until no further precipitate was observed.
(23) Furniss, B. S.; Hannaford, A. J.; Smith, P. W. G.; Tatchell, A. R.
Practical Organic Chemistry; Longman: New York, 1989.
(24) Hart, H.; Harada, K.; Du, C. J. F. J. Org. Chem. 1985, 50, 3104-
3110.
(25) Tao, W.; Nesbitt, S.; Heck, R. F. J. Org. Chem. 1990, 55, 63-69.
(26) Helm, A.; Heiler, D.; McLendon, G. J. Am. Chem. Soc. 1992, 114,
6227-6238.
(27) Raston, I.; Wennerstroem, O. Acta Chem. Scand. 1982, 36, 655-
660.

[C-H‚‚‚O] Interactions in Supramolecular Complexes
J. Am. Chem. Soc., Vol. 121, No. 7, 1999 1487
The precipitate was filtered-off and washed with H2O, Et2O and CHCl3
to afford the 8‚2PF6 as a white powder (2.19 g, 50%). mp dec > 280
°C; LSIMS m/z 1467 [2M-PF6]+, 661 [M - PF6]+; 1H NMR [300 MHz,
(CD3)2CO, 300 K] d 8.85-8.75 (m, 8H), 8.38-8.75 (m, 4H), 7.85-
8.78 (m, 4H), 7.77-7.74 (bm, 4H), 7.51-7.40 (m, 4H), 6.23 (s, 4H);
13C NMR (75 MHz, CDCl3, 300 K) d 157.2, 152.1, 146.0, 142.1, 134.4,
133.4, 130.6, 127.3, 125.2, 64.7, 52.3,; Anal.calcd for C36H28F12N2P2:
C 53.60, H 3.47, N 6.95; found C 53.43, H 3.38, N 6.85.
and were allowed to ride on their parent atoms. Anisotropic refinement
for all non-hydrogen atoms, with the exception of those of the partially
occupied MeOH, gave R1 ) 0.1178 and ωR2 ) 0.3166. All computa-
tions were carried out using the SHELXTL 5.03 package. Atomic
coordinates and bond lengths and angles have been deposited at the
Cambridge Crystallographic Data Centre.
Method B. The bis(hexafluorophosphate) salt 8‚2PF6 (78 mg, 0.1
mmol), the dibromide 10 (35 mg, 0.1 mmol), and the macrocyclic
polyether 1/5DN38C10 (154 mg, 0.24 mmol) were dissolved in DMF
(7 mL), and the solution was subjected to a pressure of 12 kbar for 2
days at ambient temperature. The solvent was evaporated, and the
resulting residue was purified as described in method A to afford the
[2]catenane (21 mg, 15%) and the [3]catenane 4‚4PF6 (87 mg, 35%).
Molecular Modeling. The solid-state structure of the complex
[1/5DN38C10][PQT]2+ was employed as the initial geometry. The
complexes [2]4+[PQT]2+ and [4]4+[PQT]2+ were constructed within the
input mode of Macromodel 5.017 by employing the solid-state structures
of the [3]catenanes as the initial geometries and then docking manually
the guest inside their cavities. All complexes with and without chloride
counterions were subjected to energy minimization employing the
Polak-Ribiere Conjugate Gradient (PRCG) method,22 the AMBER*
force field,15 and the Generalized-Born Surface Area (GB/SA) solvation
model16 for H2O, as implemented in Macromodel 5.0. The optimized
geometries were subjected individually to simulated annealing per-
formed by employing the AMBER* force field, the GB/SA solvation
model for H2O, and the PRCG minimization algorithm. The temperature
was gradually decreased from 500 to 50 K in 10 consecutive molecular
dynamics steps of 20.0 ps each. The bath constant and the time step
were mantained at 5.0 ps and 1.5 fs, respectively, in all steps. All HF/
3-21G* calculations were perfomed by using Spartan 4.118 and model
complexes and compounds generated after appropriate fragmentations
(I-VI, Figures 7 and 10) of the output structures of the simulated
annealing, in the case of [1/5DN38C10][PQT]2+ and [2]4+[PQT]2+, or
of the energy minimization, in the case of [4]4+[PQT]2+, [1/5DN38C10]-
[PQT][Cl]2, [2][PQT][Cl]6, and [4][PQT][Cl]6.
[3]Catenane 4‚4PF6. Method A. The bis(hexafluorophosphate) salt
8‚2PF6 (78 mg, 0.096 mmol), the dibromide 10 (35 mg, 0.096 mmol),
and the macrocyclic polyether 1/5DN38C10 (154 mg, 0.24 mmol)
suspended in MeCN (10 mL) were stirred for 2 weeks at ambient
temperature. The mixture was filtered and the solvent removed in vacuo.
The resulting red solid was purified by column chromatography [SiO2,
2 M NH4Cl/MeOH/MeNO2 (7:2:1, v/v/v)], which gave two different
products. Each product was dissolved in H2O, and a saturated aqueous
solution of NH4PF6 was added to afford a red precipitate. Crystallization
of the two products by vapor diffusion of benzene into MeCN solutions
of the compounds gave a [2]catenane (15 mg, 10%) and the [3]catenane
4‚4PF6 (25 mg, 35%). [2]Catenane: dec >210 °C; HRMS m/z (LSIMS)
calcd for [M - PF6]+, C88H84F18N4O10P3, 1791.5113, found 1791.5162;
LSIMS m/z 1937 [M]+, 1792 [M - PF6]+, 1647 [M - 2PF6]+, 1502
1
[M - 3PF6]+; H NMR (300 MHz, CD3CN, 300 K) δ 8.78-8.72 (m,
8H), 7.80-7.55 (m, 16H), 6.62-6.51 (m, 4H), 6.29-6.20 (m, 4H),
6.10-5.85 (m, 4H), 5.75 (s, 8H), 4.10-3.90 (m, 32H). [3]Catenane
4‚4PF6: dec >210 °C; HRMS m/z (LSIMS) calcd for [M - PF6]+,
C
124H128F18N4O20P3, 2427.8047, found 2427.8025; LSIMS m/z 2574
[M]+, 2429 [M - PF6]+, 2284 [M - 2PF6]+, 2139 [M - 3PF6]+; H
NMR (300 MHz, CD3CN, 300 K) δ 8.75-8.69 (b d, 8H), 7.75 (d,
8H), 7.85 (d, 8H), 7.65 (s, 8H), 6.68 (d, 8H), 6.55-6.47 (b t, 8H),
6.20 (d, 8H), 5.74 (s, 8H), 3.99-3.76 (m, 64H). Crystals suitable for
X-ray crystallographic analysis were grown by vapor diffusion of
MeOH into a MeCN solution of the [3]catenane. Crystal data:
C124H128N4O20‚4PF6‚4MeCN‚4MeOH, M ) 2802.48, monoclinic, a )
13.152(5) Å, b ) 14.344(3) Å, c ) 22.034(9) Å, R ) 98.98(2)°, â )
95.62(3)°, γ ) 103.21(2)°, V ) 3959(2) Å3, space group P1h, Z ) 1
1
(the structure has crystallographic Ci symmetry), Dc ) 1.175 g cm-3
,
Acknowledgment. This research was supported by the
Engineering and Physical Sciences Research Council, U.K., and
by the National Institute of General Medical Sciences, National
Institutes of Health.
µ(Cu KR) ) 1.205 mm-1, λ ) 1.54178 Å, F(000) ) 1460. Data for a
crystal having dimensions of 0.50 × 0.33 × 0.20 mm were measured
on a Siemens P4/RA diffractometer with Cu KR radiation (graphite
monochromated) using ω scans at 203 K. Of the 7976 independent
reflections measured (θ e 50°), 5740 had Io > 2σ(Io) and were
considered to be observed. The data were corrected for Lorentz and
polarization effects but not for absorption. The structure was solved
by direct methods. Solvent molecules found in the structure are
disordered: two MeCN molecules distributed over four positions and
two MeOH over four positions. Hydrogen atoms were placed in
idealized positions with Ueq(H) ) 1.2Ueq(C) and Ueq(H) ) 1.5Ueq(Me)
Supporting Information Available: Partial 1H NMR spec-
tra, additional structural diagrams, and tables listing atomic
coordinates, temperature factors, bond lengths and angles, and
torsion angles. This material is available free of charge on the
Internet at http://pubs.acs.org.
JA982748B