the signals of the xylyl and guanidinium protons of BPX26C6
and thread 2-H·PF6, respectively, are consistent with the
proposed geometry of the [2]pseudorotaxane [(BPX26C6⊃2-
H)·PF6], in which the cationic guanidinium ion is located
between the two xylyl groups of BPX26C6. Using 1H NMR
spectroscopic dilution experiments, we determined the as-
sociation constant (Ka) for the interaction between the two
species in CDCl3/CD3CN (10:1) to be 200 ( 20 M-1.10 We
Scheme 3. Synthesis of the Molecular Switch 17-H·2PF6
1
observed negligible shifts for the signals in the H NMR
spectra of the same mixture in CD3CN, CD3NO2, and CDCl3/
CD3CN (1:1), relative to those of the free species, indicating
that the interactions between the host and guest are very weak
in moderately polar solvents.
To prove the existence of the [2]pseudorotaxane [(BPX-
26C6⊃2-H)·PF6] in solution, we prepared a corresponding
interlocked [2]rotaxane from a related complex formed
between the threadlike guanidinium salt 9-H·PF6 and
BPX26C6. We synthesized 9-H·PF6 from the amine 10
(Scheme 2) following a reaction pathway similar to that used
to synthesize 2-H·PF6. The reaction of 3,5-di-tert-butylbenzyl
bromide (16), BPX26C6, and the threadlike guanidinium salt
9-H·PF6 (each 150 mM) in CHCl3/CH3CN (10:1) at room
temperature gave the [2]rotaxane 17-H·2PF6 (Scheme 3) and
the dumbbell-shaped salt 18-H·2PF6 in 18 and 52% yield,
respectively, after ion exchange and column chromatogra-
phy.11 This result confirmed that BPX26C6 encircles thread-
like guanidinium ions in the form of [2]pseudorotaxanes in
CDCl3/CD3CN (10:1) solution.
We suspected that [N-H···O] hydrogen bonding between
the BPX26C6 unit and the guanidinium station of the
[2]rotaxane 17-H·2PF6 would be more energetically favorable
in CDCl3 than would be [N+C-H···O] hydrogen bonding and
any π-π interactions between the BPX26C6 unit and the
pyridinium station; i.e., BPX26C6 should encircle the
guanidinium station in nonpolar solvents. The observation
that only small shifts in the positions of the signals of the
pyridinium unit (HR and Hꢀ) and its adjacent CH2 group of
the [2]rotaxane 17-H·2PF6 in CDCl3 relative to those in the
1H NMR spectrum of the free dumbbell-shaped salt 18-
H·2PF6 under the same conditions and the lack of cross
signals between these protons and the aromatic protons of
the BPX26C6 unit in 2D NOSEY spectra supported that the
macrocyclic component is not residue on the pyridinium
station of the [2]rotaxane 17-H·2PF6 in CDCl3 (see the
Supporting Information). The addition of CD3CN to the
CDCl3 solution of the [2]rotaxane 17-H·2PF6 led to gradual,
yet significant, upfield shifts in the positions of the signals
of the pyridinium (and adjacent CH2) and xylyl protons in
1
the H NMR spectra (Figure 2). From a comparison with
1
(11) Data for [2]rotaxane 17-H·2PF6: mp 105-106 °C; H NMR (400
the spectra of the dumbbell-shaped salt 18-H·2PF6 in these
solvents and from the cross signals observed between the
aromatic protons of the BPX26C6 unit and the pyridinium/
CH2 protons in the 2D NOSEY spectrum of 17-H·2PF6 in
CD3CN, we infer that the BPX26C6 unit prefers to encircle
the pyridinium station of this [2]rotaxane in relatively polar
solvents. Thus, the [2]rotaxane 17-H·2PF6 behaves as a
reversible molecular switch in which the BPX26C6 unit
resides selectively on the guanidinium station in nonpolar
solvents and on the pyridinium station in polar solvents.12
In addition to modifying the polarity of the solvent system,
we expected that we could also operate the molecular switch
by introducing, into a solution of the [2]rotaxane 17-H·2PF6
in a nonpolar solvent, an anion that binds tightly to
guanidinium ions, thereby forcing the BPX26C6 unit to
migrate to the pyridinium station (Scheme 4). Indeed, after
adding 3 equiv of tetra-n-butylammonium chloride (Bu4NCl)
to a CD3NO2 solution of the [2]rotaxane 17-H·2PF6, we
observed significant upfield shifts of the signals of the
MHz, CDCl3): δ ) 1.34 (s, 18H), 1.37 (s, 18H), 3.65-3.69 (m, 16H),
3.80-3.98 (br, 4H), 4.22-4.32 (m, 4H), 5.47 (br, 2H), 5.86 (br, 2H),
6.22-6.38 (br, 2H), 6.92 (s, 8H), 7.05 (s, 2H), 7.26(s, 2H), 7.35-7.39 (br,
2H), 7.40 (s, 1H), 7.53 (s, 1H), 7.80 (d, J ) 8 Hz, 2H), 8.11-8.19 (br,
2H), 8.57-8.61 (br, 2H). 13C NMR (100 MHz, CDCl3): δ ) 31.3, 31.5,
35.0, 35.0, 44.1, 45.8, 65.0, 69.8, 70.5, 73.4, 121.9, 122.3, 123.6, 124.2,
125.0, 127.9, 128.7, 128.8, 131.0, 132.5, 134.5, 136.7, 140.1, 143.6, 151.2,
152.5, 153.8, 155.4. HR-MS(ESI): m/z calcd for [17-H·PF6]+ C67H92N4O6PF6
1193.6659, found 1193.6659; [17-H]2+ C67H92N4O6 524.3509, found
524.3508. Data for Dumbbell 18-H·2PF6: mp 112-115 °C. 1H NMR (400
MHz, CD3CN): δ ) 1.28 (s, 18H), 1.32 (s, 18H), 4.36 (d, 6 Hz, 2H), 4.51
(d, 6 Hz, 2H), 5.65 (s, 2H), 6.25 (br, 2H), 6.75-6.78 (m, 2H), 7.13 (s,
2H), 7.37 (s, 2H), 7.40-7.44 (m, 3H), 7.56 (s, 1H), 7.86 (d, J ) 8 Hz,
2H), 8.22 (d, J ) 7 Hz, 2H), 8.76 (d, J ) 7 Hz, 2H). 13C NMR (100 MHz,
CD3CN): 31.4, 31.5, 35.4, 35.6, 44.9, 46.0, 64.9, 121.7, 122.2, 123.7, 124.2,
125.4, 128.3, 128.6, 132.6, 133.3, 135.1, 140.8, 144.2, 151.4, 152.4, 155.8
(one signal is missing, possibly because of signal overlapping). HR-
MS(ESI): m/z calcd for [18-H·PF6]+ C43H60N4PF6 777.4460, found 777.4459;
[18-H]2+ C43H60N4 316.2409, found 316.2408.
(12) For other molecular switches that can be operated through changes
in solvent polarity, see: (a) Lane, A. S.; Leigh, D. A.; Murphy, A. J. Am.
Chem. Soc. 1997, 119, 11092–11093. (b) Ambrosi, G.; Dapporto, P.;
Formica, M.; Fusi, V.; Giorgi, L.; Guerri, A.; Micheloni, M.; Paoli, P.;
Pontellini, R.; Rossi, P. Chem.-Eur. J. 2003, 9, 800–810. (c) Chiang, P.-
T.; Cheng, P.-N.; Lin, C.-F.; Liu, Y.-H.; Lai, C.-C.; Peng, S.-M.; Chiu,
S.-H. Chem.-Eur. J. 2006, 12, 865–876.
Org. Lett., Vol. 11, No. 3, 2009
615