A multi-mode-driven molecular shuttle: Photochemically and thermally reactive azobenzene rotaxanes

Article abstract of DOI:10.1021/ja053690l

The shuttling process of α-CyD in three rotaxanes (1-3) containing α-cyclodextrin (α-CyD) as a ring, azobenzene as a photoactive group, viologen as an energy barrier for slipping of the ring, and 2,4-dinitrobenzene as a stopper was investigated. The trans

Full text of DOI:10.1021/ja053690l

Published on Web 10/20/2005
A Multi-Mode-Driven Molecular Shuttle: Photochemically and
Thermally Reactive Azobenzene Rotaxanes
Hiroto Murakami,† Atsushi Kawabuchi,† Rika Matsumoto,† Takeshi Ido,† and
Naotoshi Nakashima*,‡
Contribution from the Department of Applied Chemistry, Faculty of Engineering, Nagasaki
UniVersity, Nagasaki 852-8521, Japan, and Department of Chemistry and Biochemistry,
Graduate School of Engineering, Kyushu UniVersity, Fukuoka 812-8581, Japan
Received June 21, 2005; E-mail: nakashima-tcm@mbox.nc.kyushu-u.ac.jp
Abstract: The shuttling process of R-CyD in three rotaxanes (1-3) containing R-cyclodextrin (R-CyD) as
a ring, azobenzene as a photoactive group, viologen as an energy barrier for slipping of the ring, and
2,4-dinitrobenzene as a stopper was investigated. The trans-cis photoisomerization of 1 by UV light
irradiation occurred in both DMSO and water due to the movement of R-CyD toward the ethylene group,
while the photoisomerization of 2 occurred in DMSO, but not in water. No photoisomerization was observed
for 3 in both water and DMSO. The activation parameters of 1 and 1-ref in DMSO are subject to a
compensation relation between ∆Sq and ∆Hq; however, in water, the ∆Sq terms are not compensated by
the ∆Hq terms. Alternating irradiation of the UV and visible lights resulted in a reversible change in the
induced circular dichroism (ICD) bands of trans-1 and cis-1. In contrast, after the UV light irradiation, the
ICD band of trans-2 decreased without the appearance of any bands of cis-2. The NMR spectra of 2 in
DMSO showed coalescence of the split signals for the methylene and for the viologen protons due to the
shuttling of R-CyD. Both the NOE differential spectra for cis-1 in water after UV light irradiation and 2 in
DMSO after heating to 120 °C showed the negative NOE peaks assigned to interior protons of R-CyD,
suggesting that R-CyD in cis-1 exists at the one ethylene moiety, and R-CyDs in cis-2 and 2 heated in
DMSO exist at the propylene moieties.
pH,11,19-22 redox,11,22-26 ion addition,17,27 solvent polar-
ity,12,14,16,28,29 and light13,14,21,23,28,30-38 have been used as external
Introduction
The construction of artificial devices and machines, both of
which function with external stimuli at the molecular level, has
attracted much attention in the nanoscience and nanotechnology
areas.1-5 Rotaxanes and catenanes, in which two or more
molecular components are mechanically interlocked, are suitable
candidates for the construction of molecular devices and
machines.6 Particularly, since the first proposal by Stoddart and
co-workers in 1991,7 molecular shuttles based on the rotaxane
structure, in which the shuttling process of a ring is controlled
by external stimuli, have been widely studied as one type of
molecular machines. Since the first molecular shuttle,7 many
different types of rotaxanes have been synthesized, and heat,7-18
stimuli, of which heat and light are the most convenient way to
(10) Loeb, S. J.; Wisner, J. A. Chem. Commun. 2000, 1939-1940.
(11) Bissell, R. A.; Co´rdova, E.; Kaifer, A. E.; Stoddart, J. F. Nature 1994,
369, 133-137.
(12) Anelli, P.-L. et al. Chem.sEur. J. 1997, 3, 1113-1135.
(13) Altieri, A.; Bottari, G.; Dehez, F.; Leigh, D. A.; Wong, J. K. Y.; Zerbetto,
F. Angew. Chem., Int. Ed. 2003, 42, 2296-2300.
(14) Bottari, G.; Dehez, F.; Leigh, D. A.; Nash, P. J.; Pe´rez, E. M.; Wong, J. K.
Y.; Zerbetto, F. Angew. Chem., Int. Ed. 2003, 42, 5886-5889.
(15) Ashton, P. R.; Philip, D.; Spencer, N.; Stoddart, J. F. J. Chem. Soc., Chem.
Commun. 1992, 1124-1128.
(16) Asakawa, M.; Brancato, G.; Fanti, M.; Leigh, D. A.; Shimizu, T.; Slawin,
A. M. Z.; Wong, J. K. Y.; Zerbetto, F.; Zhang, S. J. Am. Chem. Soc. 2002,
124, 2939-2950.
(17) Jiang, L.; Okuno, J.; Orita, A.; Otera, J. Angew. Chem., Int. Ed. 2004, 43,
2121-2124.
(18) Kawaguchi, Y.; Harada, A. Org. Lett. 2000, 2, 1353-1356.
(19) Cao, J.; Fyfe, M. C. T.; Stoddart, J. F. J. Org. Chem. 2000, 65, 1937-
1946.
† Nagasaki University.
‡ Kyushu University.
(1) Balzani, V.; Credi, A.; Raymo, F. M.; Stoddart, J. F. Angew. Chem., Int.
Ed. 2000, 39, 3248-3391.
(20) Badjic´, J. D.; Balzani, V.; Credi, A.; Silvi, S.; Stoddart, J. F. Science 2004,
303, 1845-1849.
(2) Ballardini, R.; Balzani, V.; Credi, A.; Gandolfi, M. T.; Venturi, M. Acc.
Chem. Res. 2001, 34, 445-455.
(21) Wang, Q.-C.; Qu, D. H.; Ren, J.; Chen, K.; Tian, H. Angew. Chem., Int.
Ed. 2004, 43, 1661-1665.
(3) Harada, A. Acc. Chem. Res. 2001, 34, 456-464.
(4) Schalley, C. A.; Beizai, K.; Vo¨gtle, F. Acc. Chem. Res. 2001, 34, 465-
476.
(22) Ashton, P. R.; Ballardini, R.; Balzani, V.; Baxter, I.; Credi, A.; Fyfe, M.
C. T.; Gandolfi, M. T.; Go´mez-Lo´pez, M.; Mart´ınez-D´ıaz, M.-V.; Piersanti,
A.; Spencer, N.; Stoddart, J. F.; Venturi, M.; White, A. J. P.; Williams, D.
J. J. Am. Chem. Soc. 1998, 120, 11932-11942.
(5) Collin, J.-P.; Dietrich-Buchecker, C.; Gavin˜a, P.; Jimenez-Molero, M. C.;
Sauvage, J.-P. Acc. Chem. Res. 2001, 34, 477-487.
(6) Sauvage, J.-P.; Dietrich-Buchecker, C. Molecular Catenanes, Rotaxanes
and Knots; Wiley-VCH: Weinheim, Germany, 1999.
(7) Anelli, P. L.; Spencer, N.; Stoddart, J. F. J. Am. Chem. Soc. 1991, 113,
5131-5133.
(8) Lee, J. W.; Kim, K.; Kim, K. Chem. Commun. 2001, 1042-1043.
(9) Chiu, S.-H.; Elizarov, A. M.; Glink, P. T.; Stoddart, J. F. Org. Lett. 2002,
4, 3561-3564.
(23) Armaroli, N.; Balzani, V.; Collin, J.-P.; Gavin˜a, P.; Sauvage, J.-P.; Ventura,
B. J. Am. Chem. Soc. 1999, 121, 4397-4480.
(24) Kihara, N.; Hashimoto, M.; Takata, T. Org. Lett. 2004, 6, 1693-1696.
(25) Altieri, A.; Gatti, F. G.; Kay, E. R.; Leigh, D. A.; Martel, D.; Paolucci, F.;
Slawin, A. M. Z.; Wong, J. K. Y. J. Am. Chem. Soc. 2003, 125, 8644-
8654.
(26) Tseng, H.-R.; Vignon, S. A.; Stoddart, J. F. Angew. Chem., Int. Ed. 2003,
42, 1491-1495.
9
10.1021/ja053690l CCC: $30.25 © 2005 American Chemical Society
J. AM. CHEM. SOC. 2005, 127, 15891-15899
15891

A R T I C L E S
Chart 1
Murakami et al.
iors using these compounds. In this paper, we discuss the
importance of the rotaxane structures for the design of a multi-
mode molecular shuttle. Several dual-mode-driven molecular
shuttles12,13,16,17,21-23 have already been reported; however, the
papers describing the triple-mode-driven molecular shuttles11,14
are very limited. Stoddart and co-workers reported the shuttling
processes of a rotaxane that consists of a tetracationic macro-
cycle and a thread, which possesses a benzidine and a biphenol
units as a station that are controlled by three external stimuli,
namely, heat, pH, and the redox reaction.11 Leigh and co-
workers described a heat-, light-, and solvent polarity-driven
molecular shuttle, in which the benzylic amid macrocycle
shuttles between the fumaramide and the maleamide stations.14
supply energy to the molecular shuttles. Light as the stimulus
includes three types of photochemical reactions to drive
molecular shuttles, namely, (i) photoinduced electron-transfer,
(ii) photoinduced chemical reaction, and (iii) photoisomerization
processes. Elegant examples of molecular shuttles driven by
photoinduced electron-transfer processes have been reported by
several groups.23,28,30-32,34 Abraham and co-workers described
a molecular shuttle driven by a photochemical reaction, in which
photoheterolysis was used to switch the position of a tetraca-
tionic macrocycle.35 Azobenzene is a suitable component to
construct a molecular shuttle driven by photoisomerization
because the trans-cis photoisomerization of azobenzene induces
a drastic change in its molecular occupied volume. We have
described the first photodriven molecular shuttling behavior
using 1, which possesses R-CyD as a ring molecule and an
azobenzene derivative as a threading molecule, in which R-CyD
moves back and forth from/to the azobenzene moiety to/from
the ethylene spacer by alternating UV and visible light irradia-
tion.36 Sohlberg and co-workers carried out AM1 semiempirical
electronic structure calculations for the light-driven molecular
shuttling behavior of 1.39 After our report, the photoisomeriza-
tion behaviors of an azobenzene,33,38 a stilbene,21,37 and a
maleamide14,28,31,32 have been used as the photoisomerization
groups.
Results and Discussion
Design and Synthesis. In this study, we use three rotaxanes
(1-3), containing a ring, a primary station as a photoactive unit,
and bulky terminal stoppers of R-CyD, azobenzene, viologen,
and 2,4-dinitrobenzene. As we previously described,36 the
rotaxane 1 is a short-distance molecular shuttling compound
that possesses ethylene spacers as a secondary station next to
the azobenzene unit. The rotaxane 2 is a long-distance molecular
shuttling compound that possesses propylene spacers as a
secondary station on opposite sides across the viologen units
from the azobenzene. In contrast, no secondary station exists
in the rotaxane 3; therefore, it is expected that compound 3 does
not work as a molecular shuttle.
The synthesis of 1 and its reference compound, 1-ref, have
already been briefly described elsewhere.36 We now describe
the details. The rotaxanes 1, and 2 and 3 were prepared
according to Schemes 1 and 2, respectively. The reaction of 6
with ca. 5 equiv of 4,4′-bipyridyl produced a “thread” compound
7 in 57% yield. The thread compounds 9 and 10 were also
obtained by the reaction of 4,4′-bis(chloromethyl)azobenzene40
with ca. 5 equiv each of 8 and 4,4′-bipyridyl in 86 and 57%
yields, respectively. The reactions of 7, 9, and 10 with 10-20
equiv of 2,4-dinitrofluorobenzene in the presence of R-CyD,
followed by anion exchange from Br- to ClO4- gave rotaxanes
1, 2, and 3 in 30, 68, and 4% yields, respectively. The poor
yield of 3 might be due to the weak binding between 10 and
R-CyD. Reference compounds 1-ref, 2-ref, and 3-ref were
synthesized by procedures similar to those of 1-3 in the absence
of R-CyD, and their yields were 48, 30, and 71%, respectively.
Formation of Rotaxane Structure. As we described in a
previous paper,36 the existence of the threads in the cavity of
In the present study, we describe (i) the design and synthesis
of two rotaxanes (2 and 3) in addition to the previously reported
compound 1 (Chart 1), (ii) the isomerization behavior of the
azobenzene moieties in these rotaxanes, and then (iii) the light,
heat, and solvent polarity-controlled molecular shuttling behav-
(27) Vignon, S. A.; Jarrosson, T.; Iijima, T.; Tseng, H.-R.; Sanders, J. K. M.;
Stoddart, J. F. J. Am. Chem. Soc. 2004, 126, 9884-9885.
(28) Bottari, G.; Leigh, D. A.; Pe´rez, E. M. J. Am. Chem. Soc. 2003, 125, 13360-
13361.
1
R-CyD was confirmed by H NMR spectroscopy. That is, the
(29) Lane, A. S.; Leigh, D. A.; Murphy, A. J. Am. Chem. Soc. 1997, 119,
11092-11093.
chemically equivalent protons on 1-ref were located in the same
electromagnetic environment, which is in sharp contrast to that
of 1, where the split signals with equal intensity41 appear. This
means that the respective protons on 1-ref are chemically
equivalent for C2 symmetry axis on the azo group, whereas 1
loses the symmetry due to the presence of R-CyD. This result
supports the formation of the rotaxane structure for 1.36 Similar
1H NMR spectral behavior was observed for 2 and 2-ref and
for 3 and 3-ref, indicating that both 2 and 3 also form the
rotaxane structure. In the spectra of 1-3, the signal of the
azobenzene protons next to the azo group splits 0.67, 0.62, and
(30) Benniston, A. C.; Harriman, A. Angew. Chem., Int. Ed. 1993, 32, 1459-
1461.
(31) Pe´rez, E. M.; Dryden, D. T. F.; Leigh, D. A.; Teobaldi, G.; Zerbetto, F. J.
Am. Chem. Soc. 2004, 126, 12210-12211.
(32) Brouwer, A. M.; Frochot, C.; Gatti, F. G.; Leigh, D. A.; Mottier, L.;
Paolucci, F.; Roffia, S.; Wurpel, G. W. H. Science 2001, 291, 2124-2128.
(33) Kauffmann, C.; Mu¨ller, W. M.; Vo¨gtle, F.; Weinman, S.; Abramson, S.;
Fuchs, B. Synthesis 1999, 849-853.
(34) Ashton, P. R.; Ballardini, R.; Balzani, V.; Credi, A.; Dress, K. R.; Ishow,
E.; Kleverlaan, C. J.; Kocian, O.; Preece, J. A.; Spencer, N.; Stoddart, J.
F.; Venturi, M.; Wenger, S. Chem.sEur. J. 2000, 6, 3558-3574.
(35) Abraham, W.; Grubert, L.; Grummt, U. W.; Buck, K. Chem.sEur. J. 2004,
10, 3562-3568.
(36) Murakami, H.; Kawabuchi, A.; Kotoo, K.; Nakashima, N. J. Am. Chem.
Soc. 1997, 119, 7605-7605.
(37) Stanier, C. A.; Alderman, S. J.; Claridge, T. D. W.; Anderson, H. L. Angew.
Chem., Int. Ed. 2002, 41, 1769-1772.
(40) Shinkai, S.; Shigematsu, K.; Kusano, Y.; Manabe, O. J. Chem. Soc., Perkin
(38) Qu, D. H.; Wang, Q.-C.; Ren, J.; Tian, H. Org. Lett. 2004, 6, 2085-2088.
(39) Sohlberg, K.; Sumpter, B. G.; Noid, D. W. J. Mol. Struct. 1999, 491, 281-
286.
Trans. 1 1981, 3279-3283.
(41) Sito, H.; Yonemura, H.; Nakamura, H.; Matsuo, T. Chem. Lett. 1990, 535-
538.
9
15892 J. AM. CHEM. SOC. VOL. 127, NO. 45, 2005

A Multi-Mode-Driven Molecular Shuttle
A R T I C L E S
Scheme 1. Synthesis of Rotaxane 1a
a Conditions: (i) 1,2-dibromoethane, Bu4NClO4, sat. Na2CO3 aq., ∆, 40%; (ii) HBr, EtOH, ∆; (iii) NaNO2, 1 M HCl aq., 0 °C; (iv) phenol, Na2CO3,
water, 0 °C, 90%; (v) 1,2-dibromoethane, Bu4NClO4, sat. Na2CO3 aq., ∆, 50%; (vi) 4,4′-bipyridyl, DMF, ∆, 57%; (vii) R-cyclodextrin, 2,4-dinitrofluorobenzene,
water, rt, NaClO4 (ion exchange), 30%.
Scheme 2. Synthesis of Rotaxanes 2 and 3a
a Conditions: (i) 3-bromopropylamine hydrobromide, MeCN, ∆, 58%; (ii) DMSO, ∆, 86%; (iii) R-cyclodextrin, 2,4-dinitrofluorobenzene, water, rt,
NaClO4 (ion exchange), 68%; (iv) 4,4′-bipyridyl, DMF, ∆, 57%; (v) R-cyclodextrin, 2,4-dinitrofluorobenzene, water, rt, NaClO4 (ion exchange), 4%.
0.62 ppm for 1, 2, and 3 in DMSO, which are larger than those
of the other protons. The split values of the signals for the other
protons on the rotaxanes gradually decreased with the increase
in the distance from the azo group, probably due to the lack of
anisotropy produced by R-CyD in 1. Thus, the obtained NMR
results indicate that R-CyD remains at the azobenzene moiety
on the rotaxanes. Furthermore, it was found that, in DMSO,
the protons of the methylene spacers in 3 produced double-
doublet and broad singlet signals, while in D2O, they gave two
sharp singlet signals (see Supporting Information, Figure S1).
Photoisomerization and Thermal Isomerization. The trans-
cis photoisomerization followed by cis-trans thermal isomer-
ization of the azobenzene moiety in the rotaxanes was examined
in water and in DMSO. The typical UV-visible spectra of 1 in
water and 2 in water and DMSO at the specified temperatures
are shown in Figure 2 (A, C, and E, respectively). The cis-
isomerization ratios are summarized in Table 1. UV light
irradiation in aqueous and DMSO solutions of 1 causes the
photoisomerization from the trans- to cis-form of the azobenzene
moiety, whereas, as we expected, no photoisomerization of 3
was observed in water and DMSO (see Supporting Information,
Figure S2). It is evident that the existence of the ethylene groups
in 1 as the secondary station between the azobenzene moiety
and the viologen moiety is important for the photoisomerization.
That is, R-CyD in 1 is readily movable to the ethylene group.
On the other hand, the rotaxane 2 shows the photoisomerization
behavior in DMSO; however, in water, the photoisomerization
of 2 was almost negligible. It is known that, in water, cationic
viologens act as an energy barrier for the slipping of R-CyD
due to the energetically unfavorable interaction between R-CyD
and the viologen moieties.41,42 For this reason, R-CyD is
expected to locate at the azobenzene moiety in 2 (and 3). When
1
The H NMR spectrum of 2 in DMSO also showed double-
doublet and broad singlet signals for the methylene spacers. The
appearance of the double-doublet and broad singlet signals
reveals that the geminal protons on one methylene next to the
azobenzene in 2 (and 3) are chemically nonequivalent and those
on the other methylene are chemically equivalent. This spectral
behavior might be due to the formation of hydrogen bonding10
between the hydroxy groups on R-CyD and the geminal protons
on the one methylene, of which the acidity is enhanced by the
electron attractive pyridinium group. The appearance of the
induced CD (ICD) for the rotaxanes is also evidence for the
formation of the rotaxane structure. More details are described
in the section of Molecular Shuttling Driven by Light and
Heating.
(42) Kawaguchi, Y.; Harada, A. J. Am. Chem. Soc. 2000, 122, 3797-3798.
9
J. AM. CHEM. SOC. VOL. 127, NO. 45, 2005 15893

A R T I C L E S
Murakami et al.
that cis-azobenzene is strongly stabilized in water by solvation.45
Therefore, the compound 1-ref in water is stabilized by the
solvation, which reflects the rate constant,46 whereas 1 prefers
the trans-form since the trans-form is stabilized via the
formation of the inclusion complex between R-CyD and the
trans-form, which leads to the acceleration of the rate constant.
Molecular Shuttling Driven by Light and Heating. The
shuttling processes of R-CyDs driven by light irradiation on 1
in water and 2 in DMSO were estimated using circular dichroism
(CD) spectroscopy. The measurement conditions are almost the
same as those stated in the section of Photoisomerization and
Thermal Isomerization. The measurement temperature of the
CD spectra of 1 was fixed at 5 °C, where the thermal
isomerization during the measurement is negligible. As we
previously described,36 the CD spectra of an aqueous solution
of 1 before photoirradiation show positive and negative induced
CD (ICD) bands (see Figure 2B), which are assigned to the
π-π* and n-π* transitions for trans-1 at 360 and 430 nm,
respectively. It is theoretically and experimentally known that
the band for the ICD of the long axis polarized transitions of
arene guests parallel to the axis of the CyD cavity is positive.47,48
After irradiation with UV light (360 nm), the positive ICD band
at 360 nm decreases together with the increase in the positive
ICD band at 312 nm and the negative ICD band at 430 nm,
which are assignable to the π-π* and n-π* transitions for cis-
1, respectively. The induced CD bands (λmax ) 360 and 435
nm) of compound 7, the thread compound of 1, in the presence
of R-CyD disappeared by the photoisomerization of the azoben-
zene moiety from the trans- to cis-forms (see Supporting
Information, Figure S3). On the basis of the data described
above, it is suggested that, for 1, R-CyD exists close to the
azobenzene unit when 1 is in the cis-form. Upon irradiation
with visible light (430 nm), the spectrum of cis-1 reverted to
that of the trans-1 with isosbestic points at 269 and 330 nm.
The photoisomerization of the rotaxane 2 occurred in DMSO,
but not in water as described in the section of Photoisomerization
and Thermal Isomerization. Figure 2F shows the CD spectra
of a DMSO solution of 2. Before photoirradiation, a positive
ICD peak, which is assignable to the π-π* transition for the
trans-2 at 330 nm was observed, while no n-π* transition band
was seen. In contrast, the ICD bands assigned to the π-π* and
n-π* transitions were observed in the CD spectra for the
aqueous solutions of 1 (as described above) and 2 (Figure 2D).
The presence of the n-π* transition band would be due to the
difference in the positioning of R-CyD on the azobenzene unit.
In water, R-CyD in trans-1 is believed to remain only on the
azo group due to the strong hydrophobic interaction, which
induces the ICD band for the n-π* transition. On the other
hand, in DMSO, the location of R-CyD in trans-2 is expected
to shift slightly from the azo group to the methylene moiety
probably due to the formation of hydrogen bonding10 as
described in the section of Formation of Rotaxane Structure.
This shift resulted in no ICD band of the n-π* transition. After
irradiation with UV light (350 nm) on a DMSO solution of 2,
the positive ICD band at 330 nm decreased without the
Figure 1. The 1H NMR spectra of 1-ref in DMSO-d6 (A) at 30 °C and 1
in DMSO-d6 (B) and D2O (C) at 30 °C.
R-CyD moved from the azobenzene moiety to the methylene
spacers in 2 (and 3), a part of R-CyD may still interact with the
azobenzene moiety, which hampers the photoisomerization due
to steric hindrance. In contrast, the energy barrier is expected
to be small in DMSO; thus, R-CyD allows moving from the
azobenzene moiety to one propylene moiety as the secondary
station beyond the viologen moiety. The observed behaviors
indicate that 2 functions as a molecular shuttle driven by dual-
stimuli, that is, light and solvent polarity. More details about
the motion of R-CyD are described in Molecular Shuttling
Driven by Light and Heating section.
The first-order rate constants (k1) and thermodynamic pa-
rameters for the cis-trans thermal isomerization are expected
to reflect the energy barrier arising from the rotaxane structure.
The reaction of 1 and 1-ref from the cis-form to the trans-form
in the dark was examined. The k1 values and activation
parameters are summarized in Table 2. The influence of the
rotaxane structure on the dark reaction can be estimated by
comparing the activation parameters of 1 to those of 1-ref. As
can be seen in Table 2, the ∆Sq values of 1-ref in water and
DMSO are more negative by 11.7 and 24.8 J mol-1‚K-1 than
those of 1, respectively. In addition, the ∆Sq values of 1-ref in
water and DMSO are different, while those of 1 are almost the
same. Solvation is known to affect the ∆Sq values.43,44 The
obtained results indicate that the transition state of the azoben-
zene moiety in 1-ref is more solvated than that in 1, for which
solvation is hindered due to the existence of R-CyD. Despite
the large difference in the ∆Sq values of 1 and 1-ref in DMSO,
their rate constants are almost the same. This is due to the
compensation relation between ∆Sq and ∆Hq.43,44 In contrast,
the rate constant of 1 in water is about 7-fold faster than that of
1-ref. Moreover, the ∆Hq values are almost the same, and the
∆Sq terms are not compensated by the ∆Hq terms. It is known
(45) Halpern, J.; Brady, G. W.; Winkler, C. A. Can. J. Res. 1950, 28B, 140-
155.
(46) Sanchez, A. M.; de Rossi, R. H. J. Org. Chem. 1996, 61, 3446-3451.
(43) Shinkai, S.; Kusano, Y.; Shigematsu, K.; Manabe, O. Chem. Lett. 1980,
1303-1306.
(44) Shinkai, S.; Nakaji, T.; Nishida, Y.; Ogawa, T.; Manabe, O. J. Am. Chem.
Soc. 1980, 102, 5860-5865.
(47) Shimizu, H.; Kaito, A.; Hatano, M. Bull. Chem. Soc. Jpn. 1979, 52, 2678-
2684.
(48) Ueno, A.; Fukushima, M.; Osa, T. J. Chem. Soc., Perkin Trans. 2 1990,
1067-1072.
9
15894 J. AM. CHEM. SOC. VOL. 127, NO. 45, 2005

A Multi-Mode-Driven Molecular Shuttle
A R T I C L E S
Figure 2. The UV-visible (A) and CD (B) spectra of 1 in water at 5 °C and the UV-visible (C and E) and CD (D and F) spectra of 2 in water (C and
D) and in DMSO (E and F) at 20 °C. [1] ) [2] ) 2.0 × 10-4 mol dm-3. The irradiation times of UV light (350 nm) are 0 (a, g, j), 2.0 (b), 4.0 (c), 8.0 (d),
13 (e), and 20 min (h, k). The spectra f, i, and l were obtained by visible light irradiation (420 nm) for 14 min (f) and 20 min (i, l) after the measurement
of e, h and k, respectively.
Table 1. The Photoisomerization Ratiosa of the Rotaxanes at
Specified Temperatures
Rotaxane 1b
Rotaxane 2b
H2O DMSO
Rotaxane 3b
H2O DMSO
temperature (°C)
H2O DMSO
5
20
30
67%
74%
62%
c
c
c
17%
68%
<2%
25%
∼0
∼0
a Photoisomerization ratios for the generation of the cis-forms of the
rotaxanes were calculated from the decrease in the absorbance at 350 nm,
where the absorbance of the cis-form is negligible. b [1] ) [2] ) [3] ) 2.0
× 10-5 mol dm-3
.
c The temperature is below the melting point of DMSO
(18.5 °C).
Table 2. The First-Order Rate Constants (k1) and Activation
Parameters for the Isomerization from cis- to trans-Forms in 1 and
1-ref in the Darka
Rotaxane 1
DMSO
1-ref
H2O
H2O
DMSO
b
k1 (×104) (s-1
)
2.9
94.8
84.2
0.91
97.7
87.7
0.43
99.6
85.5
1.1
97.1
79.6
∆Gq303 (kJ mol-1
)
b
∆Hq (kJ mol-1
∆Sq (kJ mol-1‚K-1
)
1
)
-34.8
-33.1
-46.5
-57.9
Figure 3. The H NMR spectra of a D2O solution of 1 at 5 (A), 30 (B),
55 (C), and 80 °C (D).
.
a [1] ) [1-ref] ) 2.0 × 10-5 mol dm-3 b Data at 303 K.
Figure 3 shows the data in the range of 5-80 °C for an aqueous
solution of 1. Accompanying the temperature increase from 5
to 55 °C, the signals of the viologen units shifted slightly due
to the enhancement of the molecular motion of the viologen
moieties, in which the chemical shifts for the protons of the
azobenzene moiety were almost constant, indicating that R-CyD
remains on the azobenzene unit in the temperature range of
5-80 °C. At 80 °C, the signals (denoted by the open circle in
Figure 3) of the viologens at 9.45 ppm disappeared, and the
two paired doublet signals (denoted by the closed circle in Figure
3) became two paired singlet signals. This suggests that a D-H
exchange reaction took place at the protons (proton f in Figure
1) on the viologen moieties in 1. To confirm the possibility of
D-H exchange, we prepared bis(2,4-dinitrophenyl)bipyridinium
dichloride as the model compound of 1. The 1H NMR spectrum
of this compound after heating in D2O at 80 °C for 1 h showed
appearance of a new band near 420 nm (Figure 2F), suggesting
the moving of R-CyD from the azobenzene moiety to the
secondary station far from the azobenzene in cis-2. With visible
light (430 nm) irradiation, the UV-visible spectrum of cis-2
reverts to that of trans-2; however, the CD band near 350 nm
recovered only 15% of the pristine CD band of trans-2. This
means that R-CyD still locates at the secondary station even
after the visible light irradiation. The obtained results reveal
that (i) the shuttling of R-CyD in 1 is driven by light, and (ii)
the shuttling of 2 can be controlled by two parameters, that is,
light and solvent polarity.
Dynamic NMR spectroscopy49 was carried out to examine
the movement of R-CyD by heating 1 in water and 2 in DMSO.
(49) Sandsto¨m, J. Dynamic NMR Spectroscopy; Academic Press: London, 1982;
pp 79-84.
9
J. AM. CHEM. SOC. VOL. 127, NO. 45, 2005 15895

A R T I C L E S
Murakami et al.
(∆Gq) and the rate constant at the coalescence temperature (kc)
were determined in order to investigate the movement of R-CyD
in 2. The rate constant of a site-exchange process at Tc is
determined by the chemical shift between the target signals (∆V),
which is used for the calculation of ∆Gq based on the Eyring
equation.49 From the double-doublet at 6.15 ppm for the one
methylene, ∆V and JAB are 25.2 and 13.7 Hz, respectively. The
values kc and ∆Gq at Tc (90 °C) for the first step were calculated
to be 94 s-1 and 80 kJ mol-1, respectively, and kc at 30 °C
extrapolated from the Eyring equation with the ∆Gq was
estimated to be 0.53 s-1. The slow rate constant at 30 °C for
the first step would be due to the hydrogen bonding10 between
the hydroxy groups of R-CyD and the protons of the methylene
spacer next to the pyridinium moiety (see Formation of Rotaxane
Structure section). From the signals between 8.6 and 8.8 ppm
for the one viologen, where ∆V is 51 Hz, the kc and ∆Gq at 150
°C for the second step were calculated to be 111 s-1 and 88 kJ
mol-1, respectively. The value kc at 30 °C was estimated to be
0.08 min-1. Harada and co-workers reported the shuttling
behavior of the rotaxane containing R-CyD as a ring and
dodecamethylene units linked with viologens as a thread in
DMSO.18 They determined kc at Tc (130 °C) and ∆Gq for
slipping of the R-CyD on the viologen unit to be 80 s-1 and 85
kJ mol-1, respectively, and kc at 30 °C was estimated to be 0.9
min-1. The rate constant for the slipping of R-CyD on the
viologen unit in 2 is 10 times lower than that of Harada’s
rotaxane. The difference in the slipping behavior between 2 and
Harada’s rotaxane would be due to the difference in the their
chemical structures; that is, Harada’s rotaxane has a linear
structure, and 2 is bent at the methylene spacers. The steric factor
is expected to play an important role in the slipping of R-CyD
through the viologen moieties.
Figure 4. The 1H NMR spectra of a DMSO-d6 solution of 2 at 30 (A), 60
(B) 90 (C), 120 (D), and 150 °C (E), and then, at 30 °C (F) after the
measurement of E.
that the peak intensities of a doublet signal assignable to the
four protons next to the nitrogen on the bipyridinium moiety
and the paired doublet signal decreased and a new singlet signal
appeared at 8.98 ppm. The molecular weights of bis(2,4-
dinitrophenyl)bipyridinium dichloride before and after the
heating obtained by FAB-MS measurements was 490.2 and
493.2, respectively, suggesting that the D-H exchange of three
protons in the four protons next to the nitrogen on the
bipyridinium moiety occurred. The acidity enhancement of the
protons by the introduction of 2,4-dinitrophenyl groups might
cause the D-H exchange.
In contrast, a coalescence of the split signals in the spectra
was observed for a DMSO solution of 2 at the temperature range
of 90-150 °C (Figure 4). We considered that the coalescence
comes from shuttling of R-CyD along the long axis of 2.
However, we have to be careful for the decomposition of the
compound at such high temperatures, so we have conducted a
HPLC (Hitachi, L-7100) experiment to detect R-CyD that should
be released by the decomposition of compound 2. The column
used was Cosmosil Sugar-D (Nakarai Tesque Inc.), and the
eluent was acetonitrile/water (65/35, v/v). It was found that only
a trace amount of R-CyD was detected for a DMSO-d6 solution
of 2 after heating to 150 °C. Therefore, we can conclude that
the decomposition of 2 is negligible in our experimental
condition. At 90 °C, the double-doublet signal (denoted by a
closed square) at 6.15 ppm that is assignable to the methylene
becomes a singlet signal. This means that the geminal protons
of the one methylene are equivalent on the NMR time scale.
At 150 °C, the eight doublet signals of the viologen moieties
then changed to four broad signals. This means that the protons
of the two viologen moieties exist in a similar environment.
These coalescences are derived from the shuttling of R-CyD
within the azobenzene moiety for the first step and then between
the azobenzene moiety and the propylene moieties. When the
temperature cycled back to 30 °C after heating to 150 °C, new
signals with low intensities appeared in the spectra (denoted
by the arrows in Figure 4). For example, the signals at 3.6 and
3.8 ppm in Figure 4F could be assignable to protons on the
propylene moiety. The appearance of new signals derived from
the movement of R-CyD from the azobenzene moiety to the
propylene moieties is also evident. The activation free energy
The CD spectra for the DMSO solution of 2 after heating to
the specified temperatures provided further evidence for the
movement of R-CyD in 2 (see Supporting Information, Figure
S4). Although the intensity of the ICD band of 2 was not much
different than heating to 100 °C, by further heating to 140 °C,
the CD intensity drastically decreased. This suggests that R-CyD
slips through the viologen units to the propylene moieties.
To obtain further evidence that R-CyDs stay at their secondary
stations in cis-1 and 2 after heating to 120 °C, we employed
NOE differential spectroscopy. As we described elsewhere,36
after UV light irradiation, the NOE differential spectrum for an
aqueous solution of 1, in which the peak for the protons of the
ethylene spacer as a secondary station at 5.30 ppm is perturbed,
shows a negative NOE peak50 at 3.65 ppm that is assignable to
the H-3 and/or H-5 protons of R-CyD, which are directed toward
the inside of the cavity (Figure 5A). This is evidence showing
that, after UV light irradiation, R-CyD exists at the ethylene
spacer moiety in 1. A DMSO solution of 2 at room temperature
after heating to 120 °C also gives a negative NOE peak at 3.37
ppm that is assignable to the H-3 and/or H-5 protons of R-CyD,
when the peak for the protons of the middle methylene in the
propylene spacer as the secondary station at 2.38 ppm was
1
perturbed (Figure 5B). Furthermore, the H NMR spectrum of
a DMSO solution of 2 at 30 °C after UV light irradiation shows
a new signal at 3.78 ppm that is assignable to the methylene
next to NH in the propylene units. The appearance of new
(50) Abraham, R. J.; Fisher, J.; Loftus, P. Introduction of NMR Spectroscopy;
John Wiley & Sons: New York, 1988.
9
15896 J. AM. CHEM. SOC. VOL. 127, NO. 45, 2005

A Multi-Mode-Driven Molecular Shuttle
A R T I C L E S
However, the shuttling process by the alternating photoirradia-
tion of UV light and visible light is irreversible.
Conclusions
We have described the design, synthesis, and the molecular
shuttling behavior of rotaxanes 1-3, which consist of R-CyD
as a ring, azobenzene as a photoactive moiety and a primary
station, viologen as a hydrophilic group and an energy barrier
for the slipping of R-CyD, and 2,4-dinitrobenzene as a stopper.
The 1H NMR and CD spectral measurements clarified that 1-3
form rotaxane structures, in which the R-CyD stays on the azo
group in 1-3 in water, and on one methylene moiety in 2 (and
3) in DMSO. The UV-visible spectral measurements have
revealed that 1 shows a reversible isomerization between the
trans- and cis-forms of the azobenzene moiety in both water
and DMSO by alternating irradiation of the UV and visible
lights, whereas the photoisomerization of 2 occurs in DMSO,
but not in water. For 3, the photoisomerization occurs neither
in water nor in DMSO. The influence of the rotaxane structure
on the first rate constants and thermodynamic parameters for
the cis-trans thermal isomerization of 1 is greater in water than
in DMSO since the presence of R-CyD hampers the solvation
by water. The CD and NOE differential spectral measurements
for the photochemically driven molecular shuttling processes
of 1 in water and 2 in DMSO demonstrated that R-CyD in 1
moves reversibly from the azobenzene moiety to the ethylene
moieties via the alternating photoirradiation of the UV and
visible lights. The reversible movement of R-CyD in 2 from/to
the azobenzene moiety to/form the propylene spacers also
occurs. The dynamic NMR, NOE differential, and CD spec-
troscopies clearly indicated that R-CyD in 2 in DMSO shuttles
(i) within the azobenzene moiety at temperatures below 100
°C, and (ii) between the azobenzene moiety and the propylene
moieties at temperatures above 100 °C. The rates of both
shuttling processes are slow possibly due to the formation of
some hydrogen bonding10 between the hydroxy groups of
R-CyD and the protons on the methylene next to the viologen
moieties and the bending structure of 2. We reached the
conclusion that the molecular shuttling processes for 1 and 2
are driven by three different modes, that is, by light for 1 and
heating and solvent polarity for 2.
Figure 5. The NOE differential spectra for (A) a D2O solution of 1 at 5
°C after UV light irradiation and (B) for a DMSO-d6 solution of 2 at 30 °C
after heating to 120 °C. The signal of the methylene spacer at 5.38 ppm for
1 and the signal of the middle methylene in the propylene spacer at 2.38
ppm for 2 were perturbed.
The stimuli used are orthodox; however, they provide a
powerful technique for controlling and monitoring the molecular
shuttling processes. The present study would afford the op-
portunity to design and develop nanoscale switching devices
based on multi-mode-driven molecular shuttles.
Figure 6. Schematic drawing for the shuttling processes of 1 and 2.
Experimental Section
signals was also observed in the NMR spectra at 30 °C after
heating to 120 °C (Figure 4F). This result suggests that R-CyD
exists at the propylene spacer moiety in cis-2 and 2 after heating
to 120 °C.
General Methods. 4,4′-Bipyridyl, 2,4-dinitrofluorobenzene, 1,2-
dibromoethane, p-hydroxyacetanilide, and 3-bromopropylamine hy-
drobromide were purchased from Tokyo Kasei, Inc. R-Cyclodextrin
was purchased from Wako Pure Chemical Inc. These reagents were
used as received. Solvents were purchased from Kishida Chemical Co.,
LTD, unless indicated otherwise. Dimethylformamide was dried over
CaH2 and distilled prior to use, and other solvents for the synthesis
were distilled prior to use. Dimethyl sulfoxide used in the measurements
was of spectral grade. Water was purified through an Ultrapure water
system, Milli-Q Plus (Millipore Co.). Its resistivity was over 18 MΩ
cm. Deuterated solvents were purchased from Merck LTD and Euriso-
Top SA. All other chemicals used were of reagent grade. 4,4′-Bis(2-
chloromethyl)azobenzene was synthesized according to the literature.40
Figure 6 is a schematic drawing for the shuttling processes
of 1 and 2. In rotaxane 1, R-CyD reversibly moves between
the azobenzene moiety and the ethylene spacer moiety by the
alternating photoirradiation of UV and visible light or temper-
ature (from cis-1 to trans-1). In rotaxane 2, R-CyD moves within
the azobenzene moiety at the temperatures below 100 °C and
moves though the axis of the rotaxane at temperatures above
100 °C. By UV light irradiation, the R-CyD in 2 also moves
from the azobenzene moiety to the propylene spacer moiety.
9
J. AM. CHEM. SOC. VOL. 127, NO. 45, 2005 15897

A R T I C L E S
Murakami et al.
1H and NOE differential NMR spectra were recorded on a JEOL
JNM-GX400 NMR in the presence of internal standards (TMS or DSS).
The UV-visible and circular dichroism (CD) spectral measurements
were carried out on a Hitachi U-3000 spectrophotometer and a JASCO
J-720 spectropolarimeter, respectively. Temperatures were maintained
to be constant within (0.1 °C (Neslab Instruments, Inc., Circulator
RTE-100). Photoirradiation for trans-cis photoisomerization of the
rotaxanes and the corresponding reference compounds was performed
on a spectrofluorometer (JASCO FP-770, light source: 150 W xenon
lamp). The slit width used in the experiment was 20 nm. The sample
concentration for the UV-visible spectral and circular dichroism (CD)
spectral measurements was 2.0 × 10-5 mol dm-3. FAB-MS and TOF-
MS spectral measurements were conducted on a JEOL JMS-700N and
an Applied Biosystems Voyager-DE Pro, respectively.
Photoisomerization ratios for the generation of cis-forms of the
rotaxanes were calculated from the decrease in the absorbance at around
350 nm, where the absorbance of the cis-form is negligible. The
reactions reached the photostationary states within 15 min. The rate
constants of the regeneration of the trans-forms from the cis-form in
the dark were determined by monitoring the increase in the absorption
band of the trans-form at around 350 nm. The OD plotted against
reaction time satisfied the first-order equation with r > 0.999. ∆Hq
and ∆Sq were determined from linear plots of ∆Gq against T, where
∆Gq at the given temperatures was calculated with Eyring equation
To a suspension of the brown solid in 1 M HCl aqueous solution
(126 mL) was added dropwise a solution of NaNO2 (0.75 g, 10.9 mmol)
in water (5 mL) at 0 °C to result in a transparent solution in a few
minutes. Excess amount of NaNO2 was quenched with sulfamic acid
by monitoring the iodostarch reaction. After the separation of a small
amount of insoluble parts by filtration, the mother liquid was added
dropwise to a solution of phenol (1.25 g, 13.4 mmol) and Na2CO3 (2.5
g, 23.6 mmol) in water (25 mL) at 0 °C. The solution immediately
became an ocherous suspension. After stirring for 1 h, the suspension
was adjusted to pH 4 with glacial acetic acid. The ocherous solid was
filtered off, washed with a large amount of water, and then recrystallized
twice from hexane then acetone to obtain 5 as a ocherous solid (2.8 g,
90%). Mp 147 °C. MS (EI) m/z 321 [M+]. 1H NMR (300 MHz, CDCl3,
TMS): δ 7.85 (m, 4H, ArH), 7.02 (d, 2H, ArH), 6.94 (d, 2H, ArH),
5.34 (s, 1H, OH), 4.36 (t, 2H, ethylene), 3.67 (t, 2H, ethylene) ppm.
IR (KBr) 3409 (νOH), 1594 (νNdN) cm-1
.
4,4′-Bis(2-bromoethoxy)azobenzene (6). To a suspension of 5 (2.5
g, 7.8 mmol), 1,2-dibromoethane (15.6 g, 81.2 mmol), and tetrabutyl-
ammonium perchlorate (0.07 g, 0.2 mmol) was added dropwise a
saturated Na2CO3 (10 mL) aqueous solution at 80 °C. After stirring
for 32 h at 80 °C, 1,2-dibromoethane was removed under reduced
pressure, and then the residue was extracted twice with chloroform
(each 20 mL portion). The chloroform solutions were combined, washed
with water, dried over anhydrous MgSO4, filtered, and then evaporated
to dryness. The crude solid was recrystallized twice from chloroform/
ethanol (1/2 v/v) to obtain 6 as a yellow solid (1.7 g, 50%). Mp 156
°C. Found: C, 45.18; H, 3.73; N, 6.53%. Calcd for C16H16Br2N2O2:
-∆Gq
RT
kBT
k1 ) κ
exp
(
)
h
1
C, 44.86; H, 3.77; N, 6.54%. H NMR (300 MHz, CDCl3, TMS): δ
where k1, kB, and h are the first-order rate, and Bolzmann and Planck
constants, respectively.
On dynamic NMR study,49 rate constants (kc) at Tc and for the first
step (an AB system) and the second step are written by the approximate
equations
7.86 (d, 4H, ArH), 7.03 (d, 4H, ArH), 4.37 (t, 4H, ethylene), 3.68 (t,
4H, ethylene) ppm. IR (KBr) 1594 (νNdN) cm-1
.
4,4′-Bis(4-(4′-pyridyl)pyridinium-2-ethoxy)azobenzene dibromide
(7). A solution of 6 (0.2 g, 0.47 mmol) and 4,4′-bipyridyl (0.4 g, 2.2
mmol) in dry DMF (3 mL) was stirred for 24 h at 75 °C. A precipitate
generated was filtered off and washed with DMF and diethyl ether to
obtain 7 as a yellow solid (0.2 g, 57%). Mp 262 °C. Found: C, 57.95;
H, 4.43; N, 10.82%. Calcd for C36H32Br2N6O2+0.5H2O: C, 57.69; H,
kc ) π[0.5{(∆ν)2 + 6JAB2}]1/2
and
1
4.44; N, 11.21%. H NMR (300 MHz, DMSO-d6, TMS): δ 9.32 (d,
4H, PyH), 8.87 (d, 4H, PyH), 8.70 (d, 4H, PyH), 8.06 (d, 4H, PyH),
π
kc )
∆ν2
7.84 (d, 4H, ArH), 7.14 (d, 4H, ArH), 5.14 (t, 4H, ethylene), 4.67 (t,
x
2
4H, ethylene) ppm. IR (KBr) 1600 (νNdN) cm-1
.
4,4′-Bis(4-(4′-(2,4-dinitrophenyl)pyridinium)pyridinium-2-ethoxy)-
azobenzene tetraperchlorate (1-ref). A solution of 7 (1.0 g, 1.3 mmol),
2,4-dinitrofluorobenzene (2.4 g, 13.0 mmol), and tetrabutylammonium
perchlorate (0.01 g) in water (60 mL) was stirred for 24 h at 25 °C.
The aqueous reaction mixture was washed twice with chloroform and
added sodium perchlorate (5.0 g) to precipitate 1-ref as a yellow solid
(0.8 g, 48%). Mp 229 °C. Found: C, 43.79; H, 2.94; N, 10.40%. Calcd
for C48H38Cl4N10O26: C, 43.90; H, 2.92; N, 10.67%. TOF-MS (R-
CHCA) m/z 1313.62 [(M + H)+]. 1H NMR (300 MHz, DMSO-d6,
TMS): δ 9.70 (d, 4H, PyH), 9.53 (d, 4H, PyH), 9.18 (d, 2H, ArH),
9.07 (d, 4H, PyH), 9.03 (dd, 2H, ArH), 8.95 (d, 4H, PyH), 8.42 (d,
2H, ArH), 7.86 (d, 4H, ArH), 7.16 (d, 4H, ArH), 5.23 (t, 4H, ethylene),
respectively, where ∆ν and JAB are the difference in the chemical shift
between the target split signals and coupling constant, respectively.
Eyring equation was used to calculate ∆Gq at the coalescence
temperature and to extrapolate values of k at 30 °C. ∆Gq was obtained
at 10% error range since many approximations are involved in this
semiquantitative treatment.
4-(2-Bromoethoxy)acetanilide (4). A mixture of p-hydroxyaceta-
nilide (2.0 g, 13.3 mmol), 1,2-dibromoethane (25 g, 133 mmol), and
tetrabutylammonium perchlorate (0.09 g, 0.26 mmol) in a saturated
aqueous solution of sodium carbonate (12 mL) was vigorously stirred
at 80 °C for 28 h. After evaporation of excess amount of 1,2-
dibromoethane and water, the residue was extracted with chloroform
(50 mL) and an insoluble solid (1,2-bis(4-acetamidophenoxy)ethane)
was filtered off. The extract was washed twice with a saturated aqueous
solution of sodium carbonate (50 mL), dried over anhydrous MgSO4,
filtered, and then evaporated to dryness. The resulting solid was
recrystallized twice from ethanol/water to give 4 as a white solid (1.4
g, 40%). Mp 127 °C. Found: C, 46.84; H, 4.70; N, 5.58%. Calcd for
C10H12BrNO2: C, 46.53; H, 4.69; N, 5.43%. 1H NMR (300 MHz, CD3-
OD, TMS): δ 7.43 (d, 2H, ArH), 6.88 (d, 2H, ArH), 4.27 (t, 2H,
ethylene), 3.66 (t, 2H, ethylene), 2.09 (s, 3H, CH3) ppm. IR (KBr)
4.73 (t, 4H, ethylene) ppm. IR (KBr) 1600 (νNdN), 1550, 1340 (νNO2
)
cm-1
.
Rotaxane 1. A solution of 7 (1.0 g, 1.3 mmol), 2,4-dinitrofluo-
robenzene (2.4 g, 13.0 mmol), and R-cyclodextrin (6.3 g, 6.5 mmol)
in water (60 mL) was stirred for 24 h at 25 °C. The aqueous reaction
mixture was washed twice with chloroform, added ammonium hexaflu-
orophosphate (5.0 g), and then evaporated. Acetonitrile was added to
the residue to remove an excess amount of the cyclodextrin. Tetra-
ethylammonium chloride (5.0 g) was added to the resulting acetonitrile
solution to produce a precipitate. After filtration, the solid product was
resolubilized in water and reprecipitated by sodium perchlorate (5.0 g)
to obtain 1 as a yellow solid (0.9 g, 30%). Mp 195 °C. Found: C,
43.57; H, 4.69; N, 6.11%. Calcd for C84H95Cl4N10O56: C, 43.51; H,
4.30; N, 6.04%. TOF-MS (R-CHCA) m/z 1885.69 [(M - 4ClO4)+].
1658 (νCdO) cm-1
.
4-(2-Bromoethoxy)-4′-hydroxyazobenzene (5). A solution of 4 (2.5
g, 9.7 mmol) in ethanol containing 20-30 wt % of HBr was refluxed
for 24 h. After evaporation of the solvent, the resulting light-brown
solid was used for next reaction without further purification.
9
15898 J. AM. CHEM. SOC. VOL. 127, NO. 45, 2005

A Multi-Mode-Driven Molecular Shuttle
A R T I C L E S
1H NMR (300 MHz, D2O, TMS): δ 9.43 (m, 6H, PyH, ArH), 9.35 (d,
2H, PyH), 9.32 (d, 2H, PyH), 8.95 (dd, 2H, ArH), 8.87 (d, 2H, PyH),
8.83 (d, 2H, PyH), 8.75 (d, 2H, PyH), 8.73 (d, 2H, PyH), 8.29 (d, 2H,
ArH), 8.27 (d, 2H, azoArH), 7.68 (d, 2H, azoArH), 7.32 (d, 2H,
azoArH), 7.21 (d, 2H, azoArH), 5.27 (m, 4H, ethylene), 4.96 (d, 6H,
CyD), 4.85 (t, 2H, ethylene), 4.78 (t, 4H, ethylene), 3.8-3.5 (m, 36H,
methylene), 6.08 (s, 2H, methylene), 5.32 (d, 6H, CyDOH), 5.19 (s,
6H, CyDOH), 4.77 (t, 4H, CCH2Py), 4.68 (d, 6H, CyDC1H), 4.27 (t,
6H, CyDOH), 3.67 (t, 4H, NCH2), 3.46 (m, 12H, CyDC6H), 3.4-3.1
(m, 24H, CyDCH), 2.37 (m, 4H, CH2CH2CH2) ppm. IR (KBr) 1610
(νNdN), 1523, 1334 (νNO2) cm-1
.
4,4′-Bis(4-(4′-pyridyl)pyridiniummethyl)azobenzene dichloride
(10). A solution of 4,4′-bis(chloromethyl)azobenzene (0.5 g, 1.8 mmol)
and 4,4′-bipyridyl (1.29 g, 8.3 mmol) in dry DMF (30 mL) was stirred
for 12 h at 85 °C. A precipitate generated was filtered off and washed
with DMF and diethyl ether to obtain 10 as an orange solid (0.2 g,
57%). Mp 262 °C. Found: C, 63.81; H, 5.03; N, 13.13%. Calcd for
CyD) ppm. IR (KBr) 3400 (νOH), 1600 (νNdN), 1550, 1340 (νNO2) cm-1
.
4-(4′-Pyridyl)-3-aminopropylpyridinium bromide hydrobromide
(8). A solution of 4,4′-bipyridyl (7.13 g, 46 mmol) and 3-bromopro-
pylamine hydrobromide (1.0 g, 4.6 mmol) in acetonitrile (100 mL) was
refluxed for 1.5 h. After cooling, a precipitate generated was filtered
and recrystallized from ethanol to give 8 as a white solid (1.0 g, 58%).
Mp 223 °C. Found: C, 38.73; H, 4.88; N, 10.06%. Calcd for C13H17N3-
Br2+1.6H2O: C, 38.66; H, 5.04 N, 10.40%. 1H NMR (300 MHz,
methanol-d4, TMS): δ 9.22 (dd, 2H, PyH), 8.85 (dd, 2H, PyH), 8.58
(dd, 2H, PyH), 8.03 (dd, 2H, PyH), 4.84 (t, 2H, CCH2Py), 3.15 (m,
1
C34H28N6Cl2+2.5H2O: C, 64.14; H, 5.22; N, 13.20%. H NMR (300
MHz, DMSO-d6, TMS): δ 9.44 (dd, 4H, PyH) 8.88 (dd, 4H, PyH),
8.70 (dd, 4H, PyH), 8.04 (dd, 4H, PyH), 7.95 (dd, 4H, ArH), 7.84 (dd,
4H, ArH), 6.04 (s, 4H, methylene) ppm. IR (KBr) 1600 (νNdN) cm-1
.
4,4′-Bis(4-(4′-(2,4-dinitrophenyl)pyridinium)pyridiniummethyl)-
azobenzene tetraperchlorate (3-ref). A solution of 10 (0.3 g, 0.5
mmol) and 2,4-dinitrofluorobenzene (2.2 g, 12.0 mmol) in water
(30 mL) was stirred for 44 h at 40 °C. The reaction mixture was washed
twice with chloroform and then was added sodium perchlorate
(2.0 g) to precipitate 3-ref as an orange solid (0.45 g, 71%). Mp
170 °C (dec). Found: C, 43.02; H, 3.07; N, 10.79%. Calcd for
C46H34Cl4N10O24+1.8H2O: C, 42.99; H, 2.95; N, 10.99%. TOF-MS
(R-CHCA) m/z 854.75 [(M - 4ClO4)+]. 1H NMR (400 MHz, DMSO-
d6, TMS): δ 9.69 (d, 4H, PyH), 9.60 (d, 4H, PyH), 9.18 (s, 2H, ArH),
9.04 (d, 4H, PyH), 9.03 (dd, 2H, ArH), 8.93 (d, 4H, PyH), 8.42 (d,
2H, ArH), 7.94 (d, 4H, azoArH), 7.80 (d, 4H, azoArH), 6.08 (s, 4H,
2H, NCH2), 2.45 (m, 2H, CH2CH2CH2) ppm. IR (KBr) 3442 (νNH) cm-1
.
4,4′-Bis(4-(4′-(3-aminopropyl)pyridinium)pyridiniummethyl)-
azobenzene dibromide(dichloride) bis(hydrobromide) (9). A solution
of 8 (0.16 g, 2.4 mmol) and 4,4′-bis(chloromethyl)azobenzene (0.1 g,
0.36 mmol) in dry DMSO (5 mL) was stirred at 70 °C for 3.5 h. A
precipitate generated was filtered and then washed with a large amount
of DMSO and small amount of acetone to give 9 as a yellow solid
(0.16 g, 86%). Found: C, 43.10; H, 4.65; N, 9.98%. Calcd for C40H46N8-
1
Cl2Br4+4.5H2O: C, 43.33; H, 4.98; N, 10.11%. H NMR (300 MHz,
D2O, TMS): δ 9.29 (d, 4H, PyH), 9.23 (d, 4H, PyH), 8.69 (d, 8H,
PyH), 8.08 (d, 4H, azoArH), 7.80 (d, 4H, azoArH), 4.95 (t, 4H, CCH2-
Py), 3.28 (t, 4H, NCH2), 2.58 (m, 4H, CH2CH2CH2) ppm. IR (KBr)
methylene) ppm. IR (KBr) 1600 (νNdN), 1550, 1340 (νNO2) cm-1
.
3430 (νNH) cm-1
.
Rotaxane 3. A solution of 9 (0.75 g, 1.3 mmol), 2,4-dinitrofluo-
robenzene (2.4 g, 13.0 mmol), and R-cyclodextrin (6.3 g, 6.5 mmol)
in water (50 mL) was stirred for 39 h at 25 °C. The reaction mixture
was washed with chloroform three times. After the addition of
ammonium hexafluorophosphate (5.0 g), the solvent was evaporated.
Acetonitrile was added to the residue to remove an excess amount of
the cyclodextrin. Tetraethylammonium chloride (4.0 g) was added to
the acetonitrile solution to produce a precipitate, which was collected
and dissolved in water, and then tetraethylammonium chloride was
added to give a precipitate, which was collected and washed with
methanol to give pure 3 as an orange solid (0.1 g, 4%). Mp 200 °C
(dec). Found: C, 42.36; H, 4.39; N, 6.01%. Calcd for C82H94-
Cl4N10O54+5.0H2O: C, 42.53; H, 4.53; N, 6.05%. TOF-MS (R-CHCA)
4,4′-Bis(4-(4′-(3-(2,4-dinitrophenylamino)propyl)pyridinium)-
pyridiniummethyl)azobenzene tetraperchlorate (2-ref). A solution
of 9 (0.3 g, 0.3 mmol) and 2,4-dinitrofluorobenzene (1.5 g, 8.2 mmol)
in dry DMF (40 mL) was stirred at 50 °C for 2 h. The solvent was
evaporated, and then the crude product was washed with a large amount
of chloroform and then methanol to produce 2-ref as an orange solid
(0.12 g, 30%). Mp 200 °C (dec). Found: C, 43.38; H, 3.63; N, 11.48%.
Calcd for C52H48Cl4N12O24+3.6H2O: C, 43.62; H, 3.86; N, 11.74%.
TOF-MS (R-CHCA) m/z 968.67 [(M - 4ClO4)+]. 1H NMR (400 MHz,
DMSO-d6, TMS): δ 9.54 (d, 4H, PyH), 9.36 (d, 4H, PyH), 8.85 (s,
2H, ArH), 8.78 (d, 4H, PyH), 8.73 (d, 4H, PyH), 8.27 (d, 2H, ArH),
7.96 (d, 4H, azoArH), 7.82 (d, 4H, azoArH), 7.27 (d, 2H, ArH), 6.06
(s, 4H, methylene), 4.77 (t, 4H, CCH2Py), 3.67 (t, 4H, NCH2), 2.37
1
m/z 1826.50 [(M - 4ClO4)+]. H NMR (300 MHz, D2O, TMS): δ
(m, 4H, CH2CH2CH2) ppm. IR (KBr) 1610 (νNdN), 1522, 1338 (νNO2
)
9.33-9.12 (m, 6H, PyH), 9.20 (d, 2H, PyH), 8.84 (dd, 1H, ArH), 8.83
(dd, 1H, ArH), 8.75-8.70 (m, 4H, PyH), 8.65 (d, 2H, PyH), 8.63 (d,
2H, PyH), 8.37 (d, 2H, azoArH), 8.18 (d, 1H, ArH), 8.16 (d, 1H, ArH),
7.84 (d, 2H, azoArH), 7.80 (d, 2H, azoArH), 7.71 (d, 2H, azoArH),
6.10 (s, 2H, methylene), 6.02 (s, 2H, methylene), 4.86 (s, 6H, CyD),
3.63-3.36 (m, 36H, CyD) ppm. IR (KBr) 3400 (νOH), 1600 (νNdN),
cm-1
.
Rotaxane 2. A mixture of R-cyclodextrin (2.0 g, 2.1 mmol) and 9
(0.25 g, 0.24 mmol) in water (40 mL) was stirred at room temperature
for 10 min, followed by the addition of 2,4-dinitrofluorobenzene (0.9
g, 4.9 mmol) and then 0.1 M NaOH aqueous solution to adjust to pH
9-10. After stirring at 40 °C for 184 h, the reaction mixture was washed
with 10 mL portions of chloroform four times to extract an excess
amount of 2,4-dinitrofluorobenzene. Sodium perchlorate (4.0 g, 33
mmol) was added to the aqueous layer, and then the resulting gummy
crude product was washed with a large amount of water then methanol
to obtain rotaxane 2 as an orange solid (62 mg, 68%). Mp 200 °C
1535, 1335 (νNO2) cm-1
.
Acknowledgment. We thank Dr. M. Kimura and Dr. T.
Fukuda of Nagasaki University for the helpful discussions on
the dynamic NMR spectroscopy. This work was supported by
Grants-in-Aid from the Ministry of Education, Science, Sports,
and Culture, Japan.
(dec). Found: C, 43.57; H, 4.74; N, 7.42%. Calcd for C88H108
-
Supporting Information Available: The 1H NMR, UV-vis
and CD spectra of 3, CD spectra of 7 and 2, and complete ref
12. This material is available free of charge via the Internet at
http://pubs.acs.org.
Cl4N12O54+3.5H2O: C, 43.99; H, 4.82; N, 6.99%. TOF-MS (R-CHCA)
m/z 1940.64 [(M - 4ClO4)+]. 1H NMR (400 MHz, DMSO-d6, TMS):
δ 9.49 (d, 2H, PyH), 9.41 (d, 2H, PyH), 9.36 (m, 4H, PyH), 8.87 (s,
2H, ArH), 8.86 (m, 2H, NH), 8.80 (d, 2H, PyH), 8.74 (d, 2H, PyH),
8.66 (d, 2H, PyH), 8.61 (d, 2H, PyH), 8.41 (d, 2H, azoArH), 8.38 (dd,
2H, ArH), 7.78 (m, 6H, azoArH), 7.77 (d, 2H, ArH), 6.15 (m, 2H,
JA053690L
9
J. AM. CHEM. SOC. VOL. 127, NO. 45, 2005 15899