Infrared study of intercomponent interactions in a switchable hydrogen-bonded rotaxane

Article abstract of DOI:10.1002/chem.200701531

The macrocycle in rotaxane 1 is preferentially hydrogen bonded to the succinamide station in the neutral form, but can be moved to the naphthalimide station by one-electron reduction of the latter. The hydrogen bonding between the amide NH groups of the macrocycle and the C=O groups in the binding stations in the thread was studied with IR spectroscopy in different solvents in both states. In addition, the solvent effect on the vibrational frequencies was analyzed; a correlation with the solvent acceptor number (AN) was observed. The conformational switching upon reduction could be detected by monitoring the hydrogen-bond-induced shifts of the v(CO) frequencies of the C=O groups of the succinamide and the reduced naphthalimide stations. The macrocycle was found to shield the encapsulated station from the solvent: wavenumbers of v(CO) bands of the C=O groups residing inside the macrocycle cavity remain unaffected by the solvent polarity.

Full text of DOI:10.1002/chem.200701531

FULL PAPER
DOI: 10.1002/chem.200701531
Infrared Study of Intercomponent Interactions in a Switchable Hydrogen-
Bonded Rotaxane
Dhiredj C. Jagesar,[a] Frantisˇek Hartl,[b] Wybren Jan Buma,[a] and Albert M. Brouwer*[a]
Abstract: The macrocycle in rotaxane 1
is preferentially hydrogen bonded to
the succinamide station in the neutral
form, but can be moved to the naph-
thalimide station by one-electron re-
duction of the latter. The hydrogen
bonding between the amide NH groups
of the macrocycle and the C=O groups
in the binding stations in the thread
was studied with IR spectroscopy in
different solvents in both states. In ad-
dition, the solvent effect on the vibra-
tional frequencies was analyzed; a cor-
relation with the solvent acceptor
number (AN) was observed. The con-
formational switching upon reduction
could be detected by monitoring the
hydrogen-bond-induced shifts of the
n(CO) frequencies of the C=O groups
of the succinamide and the reduced
naphthalimide stations. The macrocycle
was found to shield the encapsulated
station from the solvent: wavenumbers
of n(CO) bands of the C=O groups re-
siding inside the macrocycle cavity
remain unaffected by the solvent polar-
ity.
Keywords:
electrochemistry
·
IR
spectroscopy molecular
·
shuttles · rotaxanes
Introduction
principle, capable of converting chemical energy into me-
chanical work. Molecular switches have also been proposed
to be promising candidates for key elements in molecular
scale information processing devices. Interlocked systems
have been designed in which hydrogen bonding,[6–9] p–p-in-
teractions,[10,11] metal coordination,[12,13] Coulombic forces,
and hydrophobic interactions[14,15] are utilized, not only to
allowtemplate-directed synthesis, but also to control the in-
tercomponent interactions. To fully exploit the potential of
these molecules and for the design of newmolecules with
improved performance, a detailed understanding of the in-
tercomponent interactions is mandatory.
Exploitation of intercomponent mobility in rotaxanes and
other interlocked molecular assemblies has emerged as an
appealing subject of research during the past decade.[1–5] The
key challenge is to control the intercomponent degrees of
freedom, which can be accomplished by the modulation of
the weak interactions between the components. In the case
of rotaxanes, changing the interactions between the macro-
cyclic ring and binding stations in the thread can result in re-
versible transitions between structurally well-defined states
in which the ring binds to different parts of the thread (co-
conformers). If this shuttling is signaled by an observable
change in a physical or chemical property of the system, the
molecule behaves as a switch. These molecular switches are
also referred to as molecular motors, because they are, in
In this report, we describe the infrared study of hydrogen-
bond interactions in rotaxane 1 containing two potential
binding stations for the macrocycle: a succinamide and a
naphthalimide station. The succinamide group is known to
be a better hydrogen-bond acceptor than the naphthalimide
station. Hence, in the thermodynamically favored conformer
the macrocycle resides at the succinamide station.[16,17]
The switching function in rotaxane 1 is based on the shut-
tling of the macrocycle between the two stations, which is
triggered by one-electron reduction of the naphthalimide
station, either electrochemically[16] or photochemically[17]
(Scheme 1). The thus-formed radical anion has a higher hy-
drogen-bonding affinity towards the macrocycle, which re-
sults in a net translational motion of the macrocycle towards
the naphthalimide radical anion to form the energetically
[a] D. C. Jagesar, Prof. Dr. W. J. Buma, Prof. Dr. A. M. Brouwer
Van ’t Hoff Institute for Molecular Sciences
University of Amsterdam
Nieuwe Achtergracht 129
1018 WS Amsterdam (The Netherlands)
Fax (+31)205-255-670
E-mail: a.m.brouwer@uva.nl
[b] Dr. F. Hartl
Van ’t Hoff Institute for Molecular Sciences
University of Amsterdam
Nieuwe Achtergracht 166
Cꢀ
1018 WV Amsterdam (The Netherlands)
more favorable co-conformer ni-1 .
Chem. Eur. J. 2008, 14, 1935 – 1946
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1935

gen-bonding interactions.[19–21]
These studies, therefore, usual-
ly focus on the detection of hy-
drogen bonds between amide
groups. The amide C=O
stretching (n(CO), often called
ꢀ
the amide I band) and N H
stretching (ns(NH)) frequencies
are strongly affected by hydro-
gen bonds,[22–24] making the
amide group an exceptionally
good probe for the study of hy-
drogen-bond-stabilized
con-
formers. Hydrogen bonding be-
tween amide groups decreases
the NH and C=O bond orders
The shuttling process was studied in a time-resolved way
by using the electronic absorption spectrum of the naphtha-
limide radical anion as a probe.[17] With this method, the
shuttling was revealed by monitoring the precise position of
the strong electronic absorption band near 420 nm, which
undergoes a blue shift when the macrocycle binds to the
radical anion.
An alternative approach to study the structural changes
during the shuttling process is by using the macrocycle as a
probe, as in contrast to the succinamide and naphthalimide
stations, this component is directly involved at every stage
of the process. In the course of the shuttling cycle, hydrogen
bonds of the macrocycle to the succinamide station are
broken, and after the shuttling newhydrogen bonds are
formed with the naphthalimide radical anion. Thus, direct
monitoring of hydrogen bonds during the shuttling would be
an ideal way to study the process.
in both acceptor amide and donor amide.[24,25] Therefore, for
amides involved in hydrogen bonding, generally, large red
shifts of the n(CO) and n(NH) frequencies are observed.
So far, only a fewreports on the application of IR spec-
troscopy in studies of intercomponent interactions in inter-
locked molecules have been published in the literature.[26–30]
The first one was reported for a hydrogen-bonded[2] cate-
nane. In particular, the characterization of amide C=O and
NH vibrations in a benzylic amide [2]catenane, and the con-
sequences of the interlocked structure for these vibrations
have been presented.[26] The studied [2]catenane was an in-
terlocked dimer of the macrocycle contained in 1. In a later
work, IR spectroscopy was used to study external effects on
the intercomponent hydrogen bonding in this [2]catenane
incorporated in different inorganic salt matrices.[27] The salt
matrix was shown to influence the number and geometry of
hydrogen bonds between the macrocycle amide groups. The
only reported IR study of interlocked molecules in solution
was done on a succinamide rotaxane. 2D IR spectroscopy
was used to determine the conformations of this rotaxane
by means of the dipolar interactions of the C=O groups.[29]
Hydrogen bonds in rotaxanes and the effect on the vibra-
tional frequencies of the involved donor and acceptor
IR spectroscopy is a promising technique for this purpose.
It provides structural information, and can be applied with
the submicrosecond time resolution needed to monitor the
shuttling process.[18] IR spectroscopy has been widely used
in the analysis of secondary structural motifs in proteins and
peptides, which are determined to a large extent by hydro-
Scheme 1. Shuttling cycle of rotaxane 1 upon one-electron reduction and reoxidation of the naphthalimide station.
1936
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 1935 – 1946

Switchable Hydrogen-Bonded Rotaxanes
FULL PAPER
groups have also been analyzed with quantum chemical cal-
culations.[25] Calculation of the vibrational frequencies in ro-
taxane mimics (hydrogen-bonded complexes of isophtalic
amide donor and different acceptors) showed pronounced
red shifts of n(CO) bands of both the acceptor (35–65 cmꢀ1)
and donor.
The goal of the present work is to study the nature and
strength of the intercomponent hydrogen bonds in the neu-
tral and reduced forms of rotaxane 1 and thread 2 and to in-
vestigate howthese interactions are modified by medium
polarity. Furthermore, we want to correlate these interac-
tions with properties, such as shuttling rate and co-confor-
mer distribution. In addition, the solvent effect on the vibra-
tional frequencies will be analyzed. For a detailed under-
standing of the relationship between spectral features and
molecular conformations, the C=O stretching modes of 1
and 2 will be characterized. Additional information is ob-
Figure 1. The different n(CO) bands of C=O groups in the succinamide
and naphthalimide units and the macrocycle. The intramolecularly hydro-
gen-bonded conformations of the succinamide station (4a and 4b,
Scheme 2) are not included in this overview.
ꢀ
tained from an analysis of the N H stretching vibrations.
The structures and energies of the co-conformers of rotax-
ane 1 are to a large extent determined by the formation of
hydrogen bonds. Hydrogen bonding is strongly affected by
the solvent polarity, being more favored in solvents of lower
polarity. To study the effect of the solvent polarity on the
hydrogen bonding in 1, IR spectra were recorded in struc-
turally similar non-hydrogen bonding chlorinated solvents
with different dielectric constants (e): chloroform (CHCl3,
e=4.70), dichloromethane (CH2Cl2, e=8.93) and 1,2-di-
chloroethane (ClCH2CH2Cl, e=10.36). Also weakly hydro-
gen-bond-accepting solvents, THF (e=7.58) and butyroni-
trile (PrCN, e=27.2) were used.
A
ꢁ1700 cmꢀ1. This n(CO) band was chosen as a standard for
two reasons: this C=O group was found to be free of inter-
actions with other parts of the molecule and the correspond-
ing n(CO) band is well separated from the other bands. The
absorption coefficient of the ns(CO)ni band at ꢁ1700 cmꢀ1 is
(6ꢂ2)”102 Lmolꢀ1 cmꢀ1. Due to the procedure for sample
preparation, the concentration of each solution, and, there-
fore, the molar absorption coefficients, were known only ap-
proximately.
Examination of the molecular structures of 1 and 2 leads
us to predict the appearance of a complex combination of
IR bands originating from different n(CO) modes of C=O
groups of the succinamide (succ-CO) and naphthalimide (ni-
CO) stations and the macrocycle (macro-CO). In the IR
spectra of the radical anions, drastic changes of the n(CO)
Modelcompounds : The IR spectra of compounds 1–4 in
CH2Cl2 are depicted in Figure 2A. The IR spectrum of 3
shows the characteristic features of naphthalimides:[32] four
bands are observed in the C=O region. The bands at 1698
and 1660 cmꢀ1 are assigned to the symmetric (ns(CO)ni) and
antisymmetric (nas(CO)ni) C=O stretching modes, respective-
ly, while the other two bands at 1630 and 1603 cmꢀ1 origi-
nate from aromatic ring vibrations (n(Ar)ni). The B3LYP cal-
culation predicts a similar band pattern: four bands in the
spectral range of interest, at 1707 (ns(CO)ni), 1673
(nas(CO)ni), 1618 (n(Ar)ni), and 1565 cmꢀ1 (n(Ar)ni).
In the C=O stretching range of the IR spectrum of 4 in
CH2Cl2, a broad band at 1672 cmꢀ1 with a tail on the low
frequency side is observed. This band was analyzed by fit-
ting it to a sum of Lorentzian profiles, and it was found to
be composed of three contributions (Figure 3D). The two
components at 1676 and 1669 cmꢀ1 arise from unperturbed
C=O stretching. The red-shifted broad band at 1657 cmꢀ1 is
attributed to C=O stretching in the intramolecular hydro-
gen-bonded seven-membered ring conformations 4a and 4b
(Scheme 2).[33] Assuming that hydrogen bonding does not
change the absorption coefficient of the C=O stretching
mode, we can estimate from the relative contributions of the
three Lorentzians to the absorption band that approximately
Cꢀ
bands of the naphthalimide C=O groups (ni -CO) are to be
expected. To facilitate the assignment of the bands, model
compounds for the succinamide (4) and naphthalimide (3)
stations were synthesized and studied. Unfortunately, the
macrocycle itself could not be used in this study because of
its very poor solubility in non-hydrogen bonding solvents.[31]
Model compounds 3 (neutral molecule and radical anion)
and 4 will be shown to allow separate assignment of the
n(CO) bands in the IR spectra to different C=O groups
present in the succinamide and naphthalimide moieties of 1,
Cꢀ
Cꢀ
1 , 2 and 2 . Also, B3LYP/6–31G(d) calculations have been
performed to predict the nature and frequencies of the vi-
brational modes and further support the spectral analysis.
Results and Discussion
An overviewof the different n(CO) bands expected in the
IR spectra of 1–4 is given in Figure 1.
To allowcomparison of the IR bands in the different com-
pounds, all spectra were scaled to the intensity of naphthal-
Chem. Eur. J. 2008, 14, 1935 – 1946
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1937

A. M. Brouwer et al.
Figure 3. IR spectra of 4 in THF (NH region (A) and amide I region (B))
and CH2Cl2 (NH region (C) and amide I region (D)). The bands in B and
D were fitted to a sum of three Lorentzian profiles.
not perfect, but the calculation supports the notion that
there are two modes with distinctly different frequencies. A
similar conclusion was drawn from 2D IR experiments on a
related succinamide rotaxane.[29] The presence of two dis-
tinct bands implies that the C=O transition moments are not
antiparallel to each other but almost perpendicular. This is
in nice agreement with the B3LYP calculation, which pre-
dicts a dihedral angle of 1228. In the case of 1, the bifurcated
sets of hydrogen bonds with the macrocycle NH groups re-
quire an antiparallel configuration of the C=O groups,[34]
and indeed, for rotaxane 1, only one n(CO)succ band is ob-
served (Figure 5B). In the internally hydrogen-bonded
gauche conformation (corresponding to 4a), one band is cal-
culated at essentially the same wavenumber as that of the
trans form (1716 cmꢀ1), the other one at a lower frequency
(1680 cmꢀ1). In the experimental spectrum, the latter band
is found at 1657 cmꢀ1. A comparison of the calculated ener-
gies indicates that the hydrogen-bonded form has a lower
energy for the isolated molecules (ꢀ2.3 kcalmolꢀ1), but in
CH2Cl2, the extended form is predicted to be more stable by
1.2 kcalmolꢀ1. This is in good agreement with our interpre-
tation of the IR spectra of 4, which concludes that the ex-
tended form is the predominant species.
Figure 2. Amide I region of the IR absorbance spectra of 1–4 in
A) CH2Cl2 and B) THF. The spectra were scaled to the intensity of the
ns(CO)ni band at ꢁ1700 cmꢀ1
.
one third of 4 exists in these hydrogen-bonded conforma-
tions.
For the trans conformation of 4, the B3LYP calculations
predict the vibrational bands of the n(CO) modes at 1713
and 1699 cmꢀ1, with similar intensities. Experimentally, a
band is found at 1672 cmꢀ1, which can be modeled as a sum
of two Lorentzian peaks at 1676 and 1669 cmꢀ1. The agree-
ment between experimental and calculated frequencies is
In the amide I range of the IR spectrum of 4 in THF, the
bands are slightly blue-shifted with respect to the corre-
sponding peaks in CH2Cl2, and are narrower. The NH
stretching in 4, on the other hand, reveals the opposite
changes, the bands being red-shifted and broadened in THF
(Figure 3A and B). Hydrogen
bonding between THF and the
NH groups of 4 is most likely
to be the reason for these ef-
fects. THF is a hydrogen-bond
acceptor, hence, only a strong
direct effect on the NH
stretching mode is present, and
Scheme 2. Intramolecularly hydrogen-bonded conformations 4a and 4b.
a weaker indirect effect on the
1938
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 1935 – 1946

Switchable Hydrogen-Bonded Rotaxanes
FULL PAPER
Figure 4. Amide I region of IR absorbance spectra of rotaxane 1 and
thread 2 in CHCl3, CH2Cl2, ClCH2CH2Cl, PrCN, and THF. All spectra
have been scaled to the intensity of the ns(CO)ni band at ꢁ1700 cmꢀ1
.
C=O stretching. The smaller contribution of the low-fre-
quency Lorentzian to the band indicates that the intramolec-
ular hydrogen bonding leading to 4a and 4b is disrupted to
a large extent.
Apart from the free NH-stretching band, a broad red-
shifted band belonging to NH stretching in 4a and 4b is ob-
served in CH2Cl2. This band pattern is characteristic for
molecules in which intramolecular hydrogen bonding be-
tween amide groups is possible.[34]
Rotaxane and thread: The IR spectra of 2 in CHCl3,
CH2Cl2, ClCH2CH2Cl, PrCN, and THF are similar
(Figure 4). In the spectral range of interest, the IR spectrum
of 2 is equal to the sum of the spectra of 3 and 4, indicating
the absence of significant interactions between the naphtha-
limide and succinamide stations. In the chlorinated solvents,
all n(CO) modes are affected by the changes in solvent po-
larity, showing a blue shift of the corresponding bands with
increasing dielectric constant. Only the aromatic ring vibra-
tion at 1603 cmꢀ1 remains unaffected.
Figure 5. A) Difference absorbance IR spectra (1 minus 2) in CHCl3,
CH2Cl2, ClCH2CH2Cl, PrCN, and THF. The spectra were obtained by
subtracting the thread spectrum from the rotaxane spectrum after scaling
at the ns(CO)ni band. The arrows indicate the changes in this range of sol-
vents. B) Fitting result of the difference absorbance spectra. c is the
absorbance difference spectrum in THF. The spectra were fitted to a sum
of three or four Lorentzians (g). The peak positions in all solvents are
given as numbers.
The IR spectrum of 1 displays a more complicated combi-
nation of bands (Figure 4). An important observation is that
the IR spectrum does not showany indication of the exis-
tence of co-conformer ni-1, the ns(CO)ni band at 1698 cmꢀ1
in THF remains completely unaffected compared to 2. This
is in agreement with NMR spectroscopic studies on co-con-
former distribution in 1 in different solvents; the predomi-
nant co-conformer (>95%) is succ-1.[16] Crystal structures
of succinamide-based rotaxanes showthat two sets of bifur-
cated hydrogen bonds can exist between the macrocycle NH
groups and the C=O groups of the succinamide station.[16]
To obtain a clearer picture of the additional bands in 1,
difference absorbance spectra were constructed by subtract-
ing the scaled IR spectra of 2 from those of 1 (Figure 5A).
As stated above, the n(CO) modes localized on the naphtha-
limide station remain unaffected in 1, so corresponding
bands are expected to vanish in the difference spectrum.
The positive peaks in the difference spectra are contribu-
tions from the C=O stretching vibrations of the macrocycle
amide groups (n(CO)macro). Shifts of bands of the succina-
mide station due to interactions with the macrocycle may
give rise to negative or positive bands in the difference spec-
trum.
The difference spectra were fitted to a sum of Lorentzian
peaks; the results are displayed in Figure 5B. In THF, the
n(CO)macro band is found at 1667 cmꢀ1. This band gradually
shifts to 1655 cmꢀ1 in CHCl3. In THF and PrCN, a weak
band is observed at 1645 and 1648 cmꢀ1, respectively, which
may be attributed to hydrogen-bonded C=O groups of the
macrocycle. Hydrogen bonding between the succ-NH
groups and the macro-CO groups is possible due to the flex-
Chem. Eur. J. 2008, 14, 1935 – 1946
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1939

A. M. Brouwer et al.
ibility of the macrocycle: the macro-CO groups can point
inward to from hydrogen bonds with the succ-NH groups.
Molecular modeling and crystal structures of related rotax-
anes[35–37] showthat this binding motif can exist, but it proba-
bly plays a minor role. In the chlorinated solvents, this band
was not found, probably because it is obscured by the strong
n(CO)macro and n(CO)succ bands. Due to the red shift of the
n(CO)macro band towards the n(CO)succ band, the overlap is
larger in the chlorinated solvents.
solvent refractive index by using: p=(n2ꢀꢀ1)/
C
The solvent-induced shifts of the IR frequencies of 1 and
2 were evaluated with this LSER method. The wavenumbers
of the ns(CO)ni and ns(Ar)ni bands of 2 were taken directly
from the IR spectra. Because of band overlap, the n(CO)succ
band of 2 and n(CO)macro band of 1 were obtained from
peak fitting (Figure 5B). The strongest correlation was
found with the solvent acceptor number. The results of this
analysis are presented in Figure 6 and Table 1.
The negative band at 1680 cmꢀ1 (in CHCl3 1675 cmꢀ1) rep-
resents the decrease of free succ-CO groups in 1 compared
to 2. The positive band near 1640 cmꢀ1 can be assigned to
stretching of the hydrogen-bonded C=O groups in the succi-
namide station. The red shift of the hydrogen-bonded rela-
tive to free n(CO)succ in, for example, CH2Cl2 is 34 cmꢀ1.
This is 14 cmꢀ1 larger than in the case of conformers 4a and
4b relative to 4 and is caused by strong hydrogen bonding
with NH groups in the macrocycle. The hydrogen bonding is
stronger in 1 because two bifurcated hydrogen bonds with
each succ-CO are possible, while intramolecular hydrogen
bonds in 4a and 4b do not have optimal geometries. The
calculated red shifts of the n(CO) band in analogous hydro-
gen-bonded complexes of isophtalic amide and different ac-
ceptors obtained by Reckien and co-workers (35–40 cmꢀ1),
agree well with our experimentally observed red shift.[25]
Remarkably, the hydrogen-bonded n(CO)succ band lies
within 2 cmꢀ1 in the chlorinated solvents (1639–1641 cmꢀ1),
while all other bands related to the amide or imide groups
in 1 and 2 are solvatochromic (see below). The absence of a
solvent effect implies shielding of the C=O groups from the
solvent by the macrocycle. In PrCN, this band appears at
1634 cmꢀ1, which is almost the same as in THF. The differ-
ence in behavior compared with the chlorinated solvents is
probably caused by the hydrogen-bond accepting properties
of PrCN and THF. Hydrogen bonding of the solvents with
the NH groups of the amides in the succinamide moiety re-
duces the corresponding C=O stretching frequency.
*
Figure 6. Plot of ns(CO)ni, n(Ar)ni, and n(CO)succ in 2 ( ) and n(CO)macro
*
in 1 ( ) wavenumbers versus solvent acceptor number (AN). The ob-
tained values of n0, and a are listed in Table 1.
Table 1. Coefficients n0 and a, and standard deviations were obtained by
using the LSER [Eq. (1)] analysis with the solvent acceptor number.
Mode
Compound
n0
ns(CO)ni
n(CO)succ
n(CO)macro
n(Ar)ni
2
2
1
2
1705.8ꢂ0.5
1682ꢂ2
ꢀ0.40ꢂ0.03
ꢀ0.33ꢂ0.10
ꢀ0.72ꢂ0.12
ꢀ0.15ꢂ0.02
Solvent effects: The IR frequencies in 1 and 2 undergo sub-
stantial red shifts in the solvent range THF, PrCN,
ClCH2CH2Cl, CH2Cl2, and CHCl3. The shifts are not system-
atic with the dielectric constant of the solvent; hence evalua-
tion of specific solvent effects, such as hydrogen bonding
and donor–acceptor interactions is required. The evaluation
of specific and nonspecific solvent effects is often conducted
by using linear solvation energy relationships (LSER).[38]
The effect of a solvent on a given vibrational mode can be
described by Equation (1).[39]
1673ꢂ2
1632.2ꢂ0.3
The solvent acceptor number is a measure of the solvent
electrophilicity.[40] The negative value for a demonstrates a
red shift of the corresponding vibrational mode, which in
turn is a result of bond weakening. So, in the more electro-
philic solvents, electron density is withdrawn from the bonds
associated with the vibrational mode. The absolute values of
a reveal that the C=O stretching frequencies, in particular,
those of the exposed C=O groups in the macrocycle, are
more sensitive to the solvent acceptor number than the C=C
stretching in the aromatic ring of the naphthalimide
(n(Ar)ni).
X
n ¼ n0 þ
aiPi
ð1Þ
i
In this equation, n is the wavenumber of the absorption
band in the particular solvent, Pi is one of the solvent pa-
rameters: acceptor number[40] (AN), donor number (DN),
polarity (Y; derived from the static dielectric constant as:
NH stretching of 1 and 2: The NH region in the IR spectra
of 1 and 2 is depicted in Figure 7. Despite the poor signal-
Y=(esꢀ1)/
A
1940
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 1935 – 1946

Switchable Hydrogen-Bonded Rotaxanes
FULL PAPER
which has a more positive oxidation potential than the bare
Cꢀ
radical anion 2 .
Amide I: Upon one-electron reduction of model compound
Cꢀ
3, three newbands of
3
appear at 1616, 1565, and
1534 cmꢀ1 at the expense of the n(CO)ni bands at 1703 and
1667 cmꢀ1 and the n(Ar)ni band at 1631 cmꢀ1 in 3 (Figure 8).
Figure 7. NH-stretching region of IR absorbance spectra of rotaxane 1
and thread 2 in CHCl3, CH2Cl2, ClCH2CH2Cl, PrCN, and THF. All spec-
tra were scaled to the intensity of the ns(CO)ni band at ꢁ1700 cmꢀ1
.
to-noise ratio, clear bands can be distinguished. The NH-
stretching band of free NH groups in thread 2 in CHCl3 ap-
pears at 3443 cmꢀ1 and shifts to 3436 and 3432 cmꢀ1 in
CH2Cl2 and ClCH2CH2Cl, respectively. The contribution of
the hydrogen-bonded NH stretching due to conformers with
an intrasuccinamide hydrogen bond is present as a broad
band centered around 3300 cmꢀ1. Although it is impossible
to make a good estimate of the integrated intensity of this
band, it appears that only a small fraction of the NH groups
are hydrogen bonded. In PrCN and THF, the NH-stretching
bands are significantly broadened and red-shifted to 3385
and 3339 cmꢀ1, respectively, due to hydrogen bonds between
the solvent molecules and the succinamide NH groups.
In the NH-stretching region of the IR spectrum of rotax-
ane 1, bands that can be assigned to the succinamide station
and the macrocycle can be distinguished. In the chlorinated
solvents, the NH stretching of the succinamide station is
centered near 3425 cmꢀ1, which is at a slightly lower fre-
quency (shift ꢁ10 cmꢀ1) than in thread 2. The macrocycle
NH stretching appears as a broad band with a maximum at
3365 cmꢀ1. While the amide I band shows a substantial sol-
vent effect in the chlorinated solvents, the macrocycle NH
stretching is unaffected by the solvent polarity. These NH
groups are situated in the interior of the macrocycle and are
shielded from the solvent. In PrCN and THF, this band is
red-shifted, indicating that the solvent also forms hydrogen
bonds with the macrocycle when its NH groups are not in-
volved in hydrogen bonds with the thread.
Figure 8. Amide I region of the IR spectra of the naphthalimide model
compound 3 (c), 3 (g), and difference spectrum (3 minus 3)
(a) in THF.
Cꢀ
Cꢀ
Cꢀ
Calculation of the vibrational frequencies of 3 predicts the
three bands to be stronger than for the neutral molecule,
and shifted to 1633 (ns(CO)ni), 1589 (nas(CO)ni) and
1532 cmꢀ1 (n(Ar)ni), which agrees well with the experimental
result. An additional aromatic ring vibration at 1558 cmꢀ1 is
predicted, but with a very low intensity. The calculation
shows that the normal modes associated with the three
bands are similar in the radical anion and in the neutral
molecule, and also that the order of the frequencies is the
same. The C=O stretching modes have a larger shift than
the aromatic ring vibration, but the latter is substantial as
Cꢀ
well. Because the singly occupied molecular orbital in 3 is
strongly delocalized over the whole p-system, not only the
C=O, but also the C=C bonds are weakened compared to
those in 3.
Cꢀ
A similar picture emerges for 2 (Figure 9). In this case,
the characteristic band pattern of the succinamide station
between 1700 and 1640 cmꢀ1, in 2 obscured by the nas(CO)ni
band, is clearly visible in the radical anion. The tail on the
Cꢀ
red side of the ns(CO)ni band of 2 at 1616 cmꢀ1 (not pres-
Cꢀ
ent in the spectrum of 3 ), supports the existence of a small
fraction of 2 in which hydrogen bonds exist between the
Cꢀ
IR spectroelectrochemistry: The electrochemistry of 1 and 2
was already studied in detail.[17] The CV response of 2 in
THF exhibits a reversible, one-electron wave corresponding
to reduction and reoxidation of the naphthalimide group
(ꢀ1.41 V versus Fc/Fc+). Rotaxane 1, on the other hand,
displays an irreversible behavior. This irreversibility is due
succinamide station and naphthalimide radical anion (folded
conformations, see Scheme 3).
In the course of the reduction of rotaxane 1, the three
naphthalimide bands at 1702, 1667 and 1631 cmꢀ1 gradually
ꢀ1
disappeared and newbands at 1591, 1546 and 1518 cm
Cꢀ
to shuttling: the species formed after reduction is ni-1 ,
grewin. The initially obscured n(CO)succ and n(CO)macro
Chem. Eur. J. 2008, 14, 1935 – 1946
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1941

A. M. Brouwer et al.
Figure 10. NH-stretching region of the IR spectra of neutral (c) and
reduced (g) thread 1 (1 reduced) and rotaxane 2 (2 reduced) in
THF.
Figure 9. IR spectra of neutral (c) and reduced (g) top: rotaxane 1
(1 reduced) and bottom: thread 2 (2 reduced) in THF.
Cꢀ
Cꢀ
Cꢀ
Cꢀ
Cꢀ
bands in 1 are visible in 1 . The n(CO)macro band has shifted
to a lower wavenumber and appears at 1658 cmꢀ1 (Figure 9).
ence absorbance spectra between radical anion and neutral
molecule for rotaxane 1 and thread 2 (Figure 11). In both
the rotaxane and thread, three negative bands are present.
NH stretching: The NH stretching region in the IR spectrum
of the neutral molecule and radical anion of the rotaxane (1
Cꢀ
For 2 , three prominent newbands appear in the IR spec-
Cꢀ
Cꢀ
and 1 ) and the thread (2 and 2 ) in THF are depicted in
trum at 1616, 1565 and 1534 cmꢀ1. These bands are shifted
Cꢀ
Cꢀ
Figure 10. In the spectrum of 2 , an intensity decrease of
to lower wavenumbers in 1 and appear at 1591, 1551 and
the NH-stretching band at 3340 cmꢀ1 is observed, while the
red-shifted band of more strongly hydrogen-bonded NH
stretching increases. This supports the presence of hydrogen
bonds between the NH groups of the succinamide station
and the C=O groups of the naphthalimide radical anion in
Cꢀ
the folded conformation (folded-2 , Scheme 3). The relative
intensities of the free and hydrogen-bonded NH-stretching
Cꢀ
bands indicate that the extended-2 conformation is the
predominant species.
The NH-stretching region of reduced rotaxane 1 reveals
Cꢀ
some interesting features. The NH-stretching band of the
less-strongly bound NH groups in the macrocycle
(3365 cmꢀ1) has almost disappeared. The band of the NH
stretching at ꢁ3300 cmꢀ1 is more intense than in neutral 1.
Cꢀ
The stronger hydrogen-bonding interactions in 1 compared
to 1 support the notion that the macrocycle is indeed hydro-
gen bonded to the naphthalimide radical anion.
Different spectra: A convenient way to analyze the process-
Figure 11. Difference absorbance spectra (radical anion minus neutral) of
es accompanying the reduction is by comparing the differ-
rotaxane 1 (g) and thread 2 (c) in THF.
Cꢀ
Scheme 3. Structures of extended (predominant) and the folded conformations of 2 .
1942
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 1935 – 1946

Switchable Hydrogen-Bonded Rotaxanes
FULL PAPER
1518 cmꢀ1. The larger shift to lower wavenumbers of the
Table 2. Wavenumbers of IR modes of neutral and reduced 1, 2, and 3 in
THF.
Cꢀ
naphthalimide radical anion bands in 1 with respect to
Mode
Neutral [cmꢀ1
]
1
Radical anion [cmꢀ1
Cꢀ Cꢀ
]
Cꢀ
those of 2 provides clear evidence that translational move-
Cꢀ
3
2
3
2
1
ment of the macrocycle from the succinamide station to the
Cꢀ
ns(CO)ni
n(CO)macro
n(CO)macro
n(CO)succ
n(CO)succ
nas(CO)ni
n(Ar)ni
1703
–
–
–
–
1702
–
–
1702
1616
–
–
–
–
1616
–
–
–
–
1591
1658
-
1681
–
naphthalimide radical anion has taken place in 1 . This
1667[a]
1648[b]
Cꢀ
spectral red shift in 1 is induced by hydrogen bonding be-
tween the NH groups of the macrocycle and the carbonyls
1680[a]
1658[b]
1667
1631
Cꢀ
Cꢀ
1633[b]
1667
1631
of the radical anion. A red shift in 2 (compared to 3 ) ap-
pears only as a shoulder on the main band (1616 cmꢀ1), and
hydrogen bonding of the imide anion with the succinamide
1667
1631
1565
1534
1565
1534
1551
1518
Cꢀ
[a] Free C=O. [b] Hydrogen bonded C=O.
NH groups (folded-2 ) is of minor importance. In analogy
with the situation in 2 , the extended conformation of 1 is
expected to be the predominant species.
Cꢀ
Cꢀ
In the difference absorbance spectrum of the rotaxane,
two additional positive bands at 1658 and 1681 cmꢀ1 are
present. The first band (1681 cmꢀ1) has the same position as
the n(CO)succ band in 4, and belongs to n(CO)succ, which has
been liberated after shuttling. The second band at 1658 cmꢀ1
is from the macrocycle C=O stretching. This is red-shifted
by 9 cmꢀ1 due to an indirect effect of hydrogen bonding to
the NH groups of the amides. The computational study of
Reckien et al. reports shifts of ꢀ5 to ꢀ16 cmꢀ1 for similar
Cꢀ
cases.[25] In the predominant co-conformer ni-1 (Scheme 1),
the hydrogen bonds from the macrocycle NH groups to the
radical anion are stronger than those in co-conformer succ-
Cꢀ
1 , which causes further weakening of the macrocycle C=O
bond and, as a consequence, a more red-shifted C=O
stretching is observed. Another consequence is that also the
amide II band (ꢁ1520 cmꢀ1 in 1) of the macrocycle amide
groups is expected to change. A shift to higher frequency is
predicted by computation.[25] The precise changes in the
amide II band could not be resolved due to the overlap with
the nas(CO)ni and n(Ar)ni bands of the naphthalimide radical
anion. Conversely, the apparent changes in the latter bands
can also be influenced by changes in the amide II bands.
The (unexpected) shift of the n(Ar)ni peak from 1534 cmꢀ1
in thread 2 to 1518 cmꢀ1 in the rotaxane 1 is, therefore, un-
certain. The most important bands present in the IR spectra
of neutral and reduced 1 and 2 are listed in Figure 12 and
Table 2.
Figure 13. Difference absorbance spectra (radical anion minus neutral) of
rotaxane 1 (g) and thread 2 (c) in butyronitrile.
ns(CO)ni band remains at almost the same position
(1592 cmꢀ1) as in THF. This, again, is clear evidence that the
macrocycle shields the carbonyl groups of the station to
which it is hydrogen bonded from the solvent. This was also
previously concluded from the electronic absorption spectra
Cꢀ
of the radical anion in ni-1 : the absorption maximum of
Cꢀ
Cꢀ
ni-2 shifts to the blue in more polar solvents, but in ni-1 ,
it does not depend on the medium polarity because the
effect of the hydrogen bonds made with the macrocycle is
stronger than that of solvation.[17]
To analyze the solvent effect on the switching, IR spec-
troelectrochemical experiments of 1 and 2 were also carried
Cꢀ
out in PrCN (Figure 13). For 2 in PrCN, the ns(CO)ni band
shows a solvatochromic red shift (relative to THF) and ap-
pears at 1612 cm . For ni-1 in PrCN, however, the
The difference absorbance spectrum (radical anion versus
neutral molecule) of the rotaxane in THF and PrCN suggest
that a small fraction of the rad-
ꢀ1
Cꢀ
ical anion is not hydrogen
bonded: absorbance from free
radical anion is observed at
1616, 1563 and 1534 cmꢀ1
(in
A
THF), suggesting that a small
Cꢀ
fraction of 1 exists as the
succ-1
Cꢀ
co-conformer. This
seems to conflict with previous
studies on the co-conformer
Cꢀ [16,17]
Cꢀ
Cꢀ
Figure 12. Structures of the naphthalimide radical anion in rotaxane (1 ) and thread (2 ), with wavenumbers
of the most important IR modes characteristic for the binding of the macrocycle to the naphthalimide radical
anion.
distribution of 1 .
Howev-
er, the presence of bands at-
Chem. Eur. J. 2008, 14, 1935 – 1946
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1943

A. M. Brouwer et al.
tributable to the imide radical anion in a non-hydrogen-
bonded form does not necessarily mean that the molecules
exist in the succ-1 co-conformer. Cyclic voltammetry, UV/
Conclusion
Cꢀ
The amide I region in the IR spectra of 1 and 2 in solvents
of different polarity (CHCl3, CH2Cl2, ClCH2CH2Cl, PrCN,
and THF) was analyzed. In the major co-conformer of 1,
the macrocycle is situated at the succinamide station. Both 1
and 2 showbands due to the naphthalimide station, originat-
ing from symmetric and antisymmetric C=O stretching and
aromatic ring vibrations. All these modes are solvatochro-
mic and exhibit a red shift with increasing solvent acceptor
number. In the IR spectrum of 2, n(CO)succ bands of both
free and hydrogen-bonded C=O groups were observed. In
rotaxane 1, a major part of the succinamide C=O groups
was found to be hydrogen bonded to the NH groups of the
macrocycle. The shift with respect to the free n(CO)succ in 2
is large (47 cmꢀ1) because two bifurcated hydrogen bonds
can be formed with each carbonyl.[16]
Vis spectroelectrochemistry and transient absorption spec-
troscopy experiments with 1, reveal almost quantitative con-
Cꢀ
Cꢀ
Cꢀ
Cꢀ
version of succ-1 to ni-1 (ratio ni-1 /succ-1 >1500).[17]
Cꢀ
Assuming full conversion to ni-1 in the present IR experi-
ments, we can conclude that in this co-conformer not all the
C=O groups of the naphthalimide anion are involved in hy-
drogen bonding, and also that more than one type of hydro-
gen bond with different geometry and strength is present.
An additional indication for this is the fact that the ns(CO)ni
Cꢀ
band of ni-1 (1591 cmꢀ1) is broad, probably composed of
bands from C=O stretching of differently hydrogen bonded
C=O groups.
The switching is driven by the fact that hydrogen bonding
between a macrocycle and naphthalimide radical anion is
energetically favored over hydrogen bonding between the
macrocycle and succinamide station, that is, the hydrogen
The IR spectroelectrochemical experiments unambiguous-
ly showthat the macrocycle undergoes translational move-
ment from the succinamide to the one-electron-reduced
naphthalimide station. Upon reduction of thread 2, three
newbands of the radical anion appear in the spectrum,
while the bands from the neutral naphthalimide disappear.
In rotaxane 1, the same bands showup, but shifted to lower
frequencies. This red shift of the C=O stretching modes
(25 cmꢀ1) is a result of bond weakening caused by hydrogen
bonding with the NH groups of the macrocycle. The
n(CO)macro band also shows a shift to lower frequency
Cꢀ
bonds between macro-NH and ni -CO are stronger than be-
tween macro-NH and succ-CO. This is confirmed by the fact
that the n(CO)macro band exhibits a red shift of 9 cmꢀ1 in ni-
Cꢀ
1
compared to succ-1. The red shift is obviously caused by
Cꢀ
stronger hydrogen bonds in ni-1 compared to succ-1.
From the fact that hydrogen bonds in ni-1 are stronger
Cꢀ
than in succ-1, one would expect a larger hydrogen-bond-in-
duced red shift of the ns(CO)ni band relative to the n(CO)succ
band. However, the red shift of the n(CO)succ band of 1 in
THF is 47 cmꢀ1 (1680–1633 cmꢀ1), while the largest shift ob-
(9 cmꢀ1 in THF) because the amide N H groups are in-
ꢀ
Cꢀ
served in the spectrum of ni-1 is for the ns(CO)ni band
volved in hydrogen bonds (indirect effect). Also, after shut-
tling a band from the liberated succinamide station appears
in the spectrum.
(25 cmꢀ1). The observation that the ns(CO)ni band in the rad-
ical anion exhibits a smaller red shift than the n(CO)succ
band in the neutral molecule, means that the former is less
sensitive to hydrogen bonding, but the reason for this is not
obvious.
A remarkable observation is the shielding effect of the
Cꢀ
macrocycle. The succ-CO and ni -CO groups encapsulated
by the macrocycle are isolated from the solvent. The corre-
sponding n(CO) bands do not showany dependence on the
solvent polarity.
To understand in more detail the influence of the strength
of the H-bonding interactions on the equilibrium between
Cꢀ
Cꢀ
succ-1 and ni-1 , experiments with structurally modified
rotaxanes are in progress. In these newmolecular shuttles,
the naphthalimide station was replaced by other aromatic
imides. Due to the different size of the aromatic core, the
electron density on the carbonyl groups in the radical anion
will be different. An important result of the study with a
naphthalene bisimide system is that despite the smaller driv-
ing force than in the naphthalene monoimide system, shut-
tling still occurs to a large extent (>80%).[41] This supports
ExperimentalSection
Syntheses: Compounds 1 and 2 were prepared as described in previous
work.[17] The syntheses of compounds 3 and 4 are depicted in Scheme 4.
Compound 3 was synthesized from precursor 5 in 89% yield. Compound
5 was prepared by using a literature procedure.[17] Compound 4 was syn-
thesized in two steps from commercially available 2,2-diphenylethylamine
and succinic anhydride in 71% yield.
Cꢀ
Cꢀ
the idea that in 1 , the conversion to the ni-1 co-confor-
N-Butyl-2,5-di-tert-butylnaphthalimide (3): A solution of 5 (205 mg,
0.66 mmol) and N-butyl amine (481 mg, 6.6 mmol) in ethanol (25 mL)
was refluxed overnight and stirred at room temperature for 3 d. After
this time, the white precipitate was filtered off and washed with ethanol
and dried in air. Product 3 was obtained as a white powder. Yield:
215 mg (89%); 1H NMR (400 MHz, CDCl3): d=8.64 (d, 2H; arom. H),
mer is close to quantitative. Therefore, the minor fraction of
Cꢀ
apparently free C=O groups in ni-1 , represented by the
bands at 1614 and 1563 cmꢀ1 (in THF), must represent a
Cꢀ
fraction of species in which not all ni -COꢁs are hydrogen
ꢀ
ꢀ
ꢀ
8.12 (d, 2H; arom. H), 4.20 (t, 2H; N CH2 C), 1.72 (quin., 2H; N
bonded to the macrocycle NH groups.
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
CH2 CH2 CH2), 1.47 (m, 20H; tBu + CH2 CH2 CH3), 0.98 ppm (t,
ꢀ
3H; CH2 CH3); elemental analysis calcd (%) for C24H31NO2: C 78.86, H
8.55, N 3.83, O, 8.75; found: C 78.20, H 8.50, N 3.98, O 9.24.
N-(2,2-Diphenylethyl)succinamic acid (6): A mixture of 2,2-diphenyl-
AHCTREUNG
1944
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 1935 – 1946

Switchable Hydrogen-Bonded Rotaxanes
FULL PAPER
the calculation of IR frequencies in a sol-
vent medium, the polarisable continuum
model (PCM)[44, 45] was used. The computed
wavenumbers were scaled by
a standard
scaling factor of 0.9614.[46]
Acknowledgements
We thank C. Mahabiersing for technical
support and Prof. D.A. Leigh (Edinburgh,
UK) for providing us with a synthetic build-
ing block for thread 2. This research was
supported by The Netherlands Organization
for the Advancement of Research (NWO).
Scheme 4. Syntheses of naphthalimide and succinamide model compounds 3 and 4.
under reduced pressure and the white solid obtained was triturated with
water to remove excess succinic anhydride and polar byproducts. After
drying, 6 was obtained as a light-brown powder. Yield: 6.78 g (78%);
1H NMR (400 MHz, CD3OD): d=7.38–7.16 (m, 10H; arom. H), 4.24 (t,
[1] V. Balzani, A. Credi, S. Silvi, M. Venturi, Chem. Soc. Rev. 2006, 35,
1135–1149.
[2] H. Tian, Q. C. Wang, Chem. Soc. Rev. 2006, 35, 361–374.
[3] I. Willner, B. Basnar, B. Willner, Adv. Funct. Mater. 2007, 17, 702–
717.
[4] S. Saha, J. F. Stoddart, Chem. Soc. Rev. 2007, 36, 77–92.
[5] E. R. Kay, D. A. Leigh, F. Zerbetto, Angew. Chem. 2007, 119, 72–
196; Angew. Chem. Int. Ed. 2007, 46, 72–191.
[6] C. M. Keaveney, D. A. Leigh, Angew. Chem. 2004, 116, 1242–1244;
Angew. Chem. Int. Ed. 2004, 43, 1222–1224.
ꢀ
ꢀ
ꢀ
J=8.1 Hz, 1H; (Ph)2CH C), 3.81 (d, J=8.1 Hz, 2H; C CH2 N), 2.47 (t,
ꢀ
ꢀ
J=7.0 Hz, 2H; C CH2 COOH), 2.34 ppm (t, 2H, J=7.0 Hz, 2H;
ꢀ
ꢀ
HNOC CH2 C).
N-Butyl-N’-(2,2-diphenylethyl)succinamide (4): A solution of 6 (168 mg,
0.566 mmol) in CH2Cl2 (10 mL) was added to a mixture of N-(3-dimethy-
laminopropyl)-N’-ethylcarbodiimide hydrochloride (EDC) (297 mg,
1.55 mmol) and 1-hydroxybenzotriazole (HOBt) (123 mg, 0.913 mmol) in
CH2Cl2 (10 mL). The mixture was stirred for 10 min. A solution of butyl
amine (41.4 mg, 0.566 mmol) in CH2Cl2 (5 mL) was added, and the reac-
tion mixture was stirred overnight. The solvent was evaporated and the
residue was dissolved in EtOAc. The solution was washed subsequently
with 1m HCl (3”), saturated NaHCO3 (3”), and water (3”). The organ-
ic layer was dried over MgSO4 and the solvent was evaporated, affording
4 as a white powder. Yield 182 mg (91%); 1H NMR (400 MHz, CDCl3):
d=7.39–7.18 (m, 10H; arom. H), 5.85 (brs, 2H; NHs), 4.16 (t, 1H;
[7] V. Serreli, C. F. Lee, E. R. Kay, D. A. Leigh, Nature 2007, 445, 523–
527.
[8] A. M. Brouwer, S. M. Fazio, N. Haraszkiewicz, D. A. Leigh, C. M.
Lennon, Photochem. Photobiol. Sci. 2007, 6, 480–486.
[9] J. Berna, D. A. Leigh, M. Lubomska, S. M. Mendoza, E. M. Perez, P.
Rudolf, G. Teobaldi, F. Zerbetto, Nat. Mater. 2005, 4, 704–710.
[10] J. D. Badjic, C. M. Ronconi, J. F. Stoddart, V. Balzani, S. Silvi, A.
Credi, J. Am. Chem. Soc. 2006, 128, 1489–1499.
[11] V. Balzani, M. Clemente-Leon, A. Credi, B. Ferrer, M. Venturi,
A. H. Flood, J. F. Stoddart, Proc. Natl. Acad. Sci. USA 2006, 103,
1178–1183.
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
(Ph)2CH C), 3.88 (dd, 2H; (Ph)2CH CH2 N), 3.19 (m, 2H; N CH2
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
CH2), 2.40 (s, 4H; OC CH2 CH2 CO), 1.46 (quin., 2H; N CH2 CH2
ꢀ
ꢀ
ꢀ
ꢀ
CH2), 1.34 (quin., 2H; CH2 CH2 CH3), 0.92 (t, 3H; CH2 CH2 CH3); el-
emental analysis calcd (%) for C22H28N2O2: C 74.97, H 8.01, N 7.95, O
9.08; found: C 74.83; H 8.56; N 7.88; O 8.67.
[12] J. P. Collin, D. Jouvenot, M. Koizumi, J. P. Sauvage, Eur. J. Inorg.
Chem. 2005, 1850–1855.
[13] M. C. Jimenez-Molero, C. Dietrich-Buchecker, J. P. Sauvage, Chem.
Eur. J. 2002, 8, 1456–1466.
[14] E. J. F. Klotz, T. D. W. Claridge, H. L. Anderson, J. Am. Chem. Soc.
2006, 128, 15374–15375.
[15] H. Isobe, S. Sato, E. Nakamura, Org. Lett. 2002, 4, 1287–1289.
[16] A. Alteri, F. G. Gatti, E. R. Kay, D. A. Leigh, D. Martel, F. Paolucci,
A. M. Z. Slawin, J. K. Y. Wong, J. Am. Chem. Soc. 2003, 125, 8644–
8654.
[17] A. M. Brouwer, C. Frochot, F. G. Gatti, D. A. Leigh, L. Mottier, F.
Paolucci, S. Roffia, G. W. H. Wurpel, Science 2001, 291, 2124–2128.
[18] L. Hviid, J. W. Verhoeven, A. M. Brouwer, M. N. Paddon-Row, J. X.
Yang, Photochem. Photobiol. Sci. 2004, 3, 246–251.
[19] B. Ishimoto, K. Tonan, S. Ikawa, Spectrochim. Acta A 1999, 55,
2321–2327.
[20] R. A. Shaw, A. Perczel, H. H. Mantsch, G. D. Fasman, J. Mol. Struct.
1994, 324, 143–150.
[21] G. P. Dado, S. H. Gellman, J. Am. Chem. Soc. 1993, 115, 4228–4245.
[22] P. Bruni, E. Giorgini, E. Maurelli, G. Tosi, Vib. Spectrosc. 1996, 12,
249–255.
Infrared spectra
Sample preparation: All IR spectra were measured with 2–4 mm solu-
tions. All solutions were prepared in Schlenk flasks by dissolving weighed
amounts of compound under dry N2 in freshly distilled dry solvent to
avoid the presence of water in the samples. The solvent was added with a
syringe (the volume of the obtained solution was known only approxi-
mately). The sample solutions were then transferred to the IR cell with a
syringe.
Infrared spectroelectrochemistry: IR spectroelectrochemical measure-
ments were performed in an optically transparent thin-layer electrochem-
ical (OTTLE) cell[42] equipped with CaF2 optical windows and a minigrid
platinum working electrode. The optical beam can pass directly through
the working electrode, allowing the redox processes taking place in the
thin solution layer surrounding the working electrode to be studied spec-
troscopically. The controlled-potential electrochemical conversions were
carried out with a PA4 potentiostat (EKOM, Polnµ, Czech Republic). A
ꢀ1
slowscan rate of 2 mVs was used to allow quantitative electrochemical
conversion. The spectra were recorded at different points on the thin-
layer voltammograms; the scanning of the potential was paused during
the
recording.
Dry
tetrabutylammonium
hexafluorophosphate
(Bu4NPF6), crystallized twice from methanol, was used as supporting
electrolyte. IR spectra were recorded on Nicolet 6700 FTIR or Bruker
Vertex 70 FTIR spectrometers.
[23] G. P. Dado, S. H. Gellman, J. Am. Chem. Soc. 1992, 114, 3138–3139.
[24] W. Herrebout, K. Clou, H. O. Desseyn, N. Blaton, Spectrochim.
Acta A 2003, 59, 47–59.
Calculations: Calculations of the IR frequencies in the amide I region of
the model compounds were carried out at the B3LYP/6–31G(d) level.[43]
For the calculations of the C=O vibrations in 4, the butyl group was re-
placed by a methyl group, while for 3 and 3 , the butyl group was re-
placed by an ethyl group, and the tert-butyl groups by methyl groups. For
[25] W. Reckien, B. Kirchner, S. D. Peyerimhoff, J. Phys. Chem. A 2006,
110, 12963–12970.
[26] M. Fanti, C. A. Fustin, D. A. Leigh, A. Murphy, P. Rudolf, R. Cau-
dano, R. Zamboni, F. Zerbetto, J. Phys. Chem. A 1998, 102, 5782–
5788.
Cꢀ
Chem. Eur. J. 2008, 14, 1935 – 1946
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1945

A. M. Brouwer et al.
[27] C. A. Fustin, D. A. Leigh, P. Rudolf, D. Timpel, F. Zerbetto, Chem-
PhysChem 2000, 1, 97–100.
[40] U. Mayer, V. Gutman, W. Greger, Monatsh. Chem. 1975, 106, 1235.
[41] J. Baggerman, PhD Thesis, University of Amsterdam, Amsterdam,
2006.
[28] T. J. Huang, A. Flood, C. W. Chu, S. Kang, T. F. Guo, T. Yamamoto,
H. R. Tseng, B. D. Yu, Y. Yang, J. F. Stoddart, Third IEEE Confer-
ence on Nanotechnology (San Francisco), 2003, 2, 698–701.
[29] O. F. A. Larsen, P. Bodis, W. J. Buma, J. S. Hannam, D. A. Leigh, S.
Woutersen, Proc. Natl. Acad. Sci. USA 2005, 102, 13378–13382.
[30] E. DeIonno, H. R. Tseng, D. D. Harvey, J. F. Stoddart, J. R. Heath,
J. Phys. Chem. B 2006, 110, 7609–7612.
[31] A. G. Johnston, D. A. Leigh, A. Murphy, J. P. Smart, M. D. Deegan,
J. Am. Chem. Soc. 1996, 118, 10662–10663.
[32] I. Grabchev, C. Petkov, V. Bojinov, Dyes Pigm. 2004, 62, 229–234.
[33] S. H. Gellman, G. P. Dado, G. B. Liang, B. R. Adams, J. Am. Chem.
Soc. 1991, 113, 1164–1173.
[34] D. T. McQuade, S. L. McKay, D. R. Powell, S. H. Gellman, J. Am.
Chem. Soc. 1997, 119, 8528–8532.
[35] D. S. Marlin, D. G. Cabrera, D. A. Leigh, A. M. Z. Slawin, Angew.
Chem. 2006, 118, 1413–1418; Angew. Chem. Int. Ed. 2006, 45, 1385–
1390.
[36] D. A. Leigh, M. A. F. Morales, E. M. Perez, J. K. Y. Wong, C. G.
Saiz, A. M. Z. Slawin, A. J. Carmichael, D. M. Haddleton, A. M.
Brouwer, W. J. Buma, G. W. H. Wurpel, S. Leon, F. Zerbetto,
Angew. Chem. 2005, 117, 3122–3127; Angew. Chem. Int. Ed. 2005,
44, 3062–3067.
[37] F. Biscarini, M. Cavallini, D. A. Leigh, S. Leon, S. J. Teat, J. K. Y.
Wong, F. Zerbetto, J. Am. Chem. Soc. 2002, 124, 225–233.
[38] W. R. Fawcett, P. Politzer, J. S. Murray, Theoretical and Computa-
tional Chemistry: Quantitative Treatments of Solvent/Solute Interac-
tions, Elsevier, Amsterdam, 1994, pp. 183–212.
ˇ
ˇ
[42] M. Krejcik, M. Danek, F. Hartl, J. Electroanal. Chem. 1991, 317,
179–187.
[43] Gaussian 03 (Revision B.3), M. J. Frisch, G. W. Trucks, H. B. Schle-
gel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, J. A. Montgom-
ery, T. Vreven, K. N. Kudin, J. C. Burant, J. M. Millam, S. S. Iyengar,
J. Tomasi, V. Barone, B. Mennucci, M. Cossi, G. Scalmani, N. Rega,
G. A. Petersson, H. Nakatsuji, M. Hada, M. Ehara, K. Toyota, R.
Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao,
H. Nakai, M. Klene, X. Li, J. E. Knox, H. P. Hratchian, J. B. Cross,
C. Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev,
A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, P. Y. Ayala, K.
Morokuma, G. A. Voth, P. Salvador, J. J. Dannenberg, V. G. Zakr-
zewski, S. Dapprich, A. D. Daniels, M. C. Strain, O. Farkas, D. K.
Malick, A. D. Rabuck, K. Raghavachari, J. B. Foresman, J. V. Ortiz,
Q. Cui, A. G. Baboul, S. Clifford, J. Cioslowski, B. B. Stefanov, G.
Liu, A. Liashenko, P. Piskorz, I. Komaromi, R. L. Martin, D. J. Fox,
T. Keith, M. A. Al-Laham, C. Y. Peng, A. Nanayakkara, M. Challa-
combe, P. M. W. Gill, B. Johnson, W. Chen, M. W. Wong, C. Gonza-
lez and J. A. Pople, Gaussian, Inc., Pittsburgh PA, USA, 2003.
[44] S. Miertus, E. Scrocco, J. Tomasi, Chem. Phys. 1981, 55, 117–129.
[45] M. Cossi, V. Barone, R. Cammi, J. Tomasi, Chem. Phys. Lett. 1996,
255, 327–335.
[46] A. P. Scott, L. Radom, J. Phys. Chem. 1996, 100, 16502–16513.
Received: September 27, 2007
[39] W. R. Fawcett, A. A. Kloss, J. Phys. Chem. 1996, 100, 2019–2024.
Published online: December 6, 2007
1946
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 1935 – 1946