An electrochemically and thermally switchable donor-acceptor [c2]daisy chain rotaxane

Article abstract of DOI:10.1002/anie.201308498

Although motor proteins are essential cellular components that carry out biological processes by converting chemical energy into mechanical motion, their functions have been difficult to mimic in artificial synthetic systems. Daisy chains are a class of rotaxanes which have been targeted to serve as artificial molecular machines because their mechanically interlocked architectures enable them to contract and expand linearly, in a manner that is reminiscent of the sarcomeres of muscle tissue. The scope of external stimuli that can be used to control the musclelike motions of daisy chains remains limited, however, because of the narrow range of supramolecular motifs that have been utilized in their templated synthesis. Reported herein is a cyclic daisy chain dimer based on π-associated donor-acceptor interactions, which can be actuated with either thermal or electrochemical stimuli. Molecular dynamics simulations have shown the daisy chain's mechanism of extension/contraction in the ground state in atomistic detail. Flexing molecular muscles: A bistable [c2]daisy chain rotaxane, based on π-associated donor-acceptor interactions between naphthalene diimide and dioxynaphthalene recognition units, is obtained using click chemistry. A combination of experimental data and molecular dynamics simulations demonstrates that the daisy chain's extensile and contractile motions can be controlled, either by redox chemistry (a thermodynamically driven process), or by temperature (an entropy-driven process). Copyright

Full text of DOI:10.1002/anie.201308498

Angewandte
Chemie
DOI: 10.1002/anie.201308498
Molecular Switches
An Electrochemically and Thermally Switchable Donor–Acceptor
[c2]Daisy Chain Rotaxane**
Carson J. Bruns, Jianing Li, Marco Frasconi, Severin T. Schneebeli, Julien Iehl, Henri-
Pierre Jacquot de Rouville, Samuel I. Stupp, Gregory A. Voth,* and J. Fraser Stoddart*
Abstract: Although motor proteins are essential cellular
components that carry out biological processes by converting
chemical energy into mechanical motion, their functions have
been difficult to mimic in artificial synthetic systems. Daisy
chains are a class of rotaxanes which have been targeted to
serve as artificial molecular machines because their mechan-
ically interlocked architectures enable them to contract and
expand linearly, in a manner that is reminiscent of the
sarcomeres of muscle tissue. The scope of external stimuli
that can be used to control the musclelike motions of daisy
chains remains limited, however, because of the narrow range
of supramolecular motifs that have been utilized in their
templated synthesis. Reported herein is a cyclic daisy chain
dimer based on p-associated donor–acceptor interactions,
which can be actuated with either thermal or electrochemical
stimuli. Molecular dynamics simulations have shown the daisy
chainꢀs mechanism of extension/contraction in the ground state
in atomistic detail.
muscles. Cyclic dimers, known as [c2]daisy chains, are of
considerable interest because their large-amplitude mechan-
ical motions can be harnessed to design artificial molecular
machines[3] (AMMs) inspired by the magnificent biomolecu-
lar machines such as myosin and kinesin, which transduce
chemical energy into mechanical energy to carry out the
functions that sustain life processes. The actuation of switch-
able [c2]daisy chains has been controlled by ionic,[2] pH,[4] and
photochemical[5] stimuli. A few pH-switchable poly[c2]daisy
chains have also been prepared,[6] the most recent of which
demonstrated[6d] a remarkable approximate 6 mm change in
contour length when actuated in solution, thus bringing the
dream of creating daisy chain AMMs closer to reality.[6e]
Although a variety of recognition motifs have been applied
to the construction of daisy chains, including transition
metals,[2] crown ether cation recognition,[7] hydrophobic
forces,[8] and anion binding,[9] to our knowledge, neither
electrochemically switchable nor thermally switchable daisy
chains have so far graced the literature. Since we and others
have utilized[10] p-associated donor–acceptor (D–A) interac-
tions to construct electrochemically switchable bistable
catenanes and rotaxanes, analogous D–A daisy chains
appear to us to offer new ways of controlling molecular
muscles. Although there have been several reports[11] of D–A
supramolecular daisy chains, that is, pseudorotaxanes, molec-
ular daisy chains of this ilk have remained elusive. Further-
more, to our knowledge no computational analyses of
[c2]daisy chains have been undertaken, despite their potential
to deepen our understanding at the atomistic level of the
structures and dynamics of mechanically interlocked archi-
tectures that exceed simple [2]rotaxanes and [2]catenanes in
sophistication and complexity. Herein, we report a neutral
D–A bistable [c2]daisy chain, based on naphthalene diimide
(NDI) and a 1,5-dioxynaphthalene (DNP)-containing crown
ether, along with its molecular dynamics and thermal and
electrochemical switching properties in solution.
M
olecular daisy chains[1] are a class of rotaxanes composed
of self-complementary monomers with cross-threaded archi-
tectures which enable them to express extensile and contrac-
tile motions, thus earning them the title[2] of molecular
[*] C. J. Bruns, Dr. M. Frasconi, Dr. S. T. Schneebeli, Dr. J. Iehl,
Dr. H.-P. Jacquot de Rouville, Prof. S. I. Stupp, Prof. J. F. Stoddart
Department of Chemistry, Northwestern University
2145 Sheridan Road, Evanston, IL 60208 (USA)
E-mail: stoddart@northwestern.edu
Dr. J. Li, Prof. G. A. Voth
Department of Chemistry, Institute for Biophysical Dynamics
James Frank Institute and Computation Institute
The University of Chicago
5735 South Ellis Avenue, Chicago, IL 60637 (USA)
E-mail: gavoth@uchicago.edu
[**] This research was supported by the Non-Equilibrium Energy
Research Center (NERC), an Energy Frontiers Research Center
(EFRC) funded by the US Department of Energy, Office of Science,
Office of Basic Energy Sciences under Award Number DE-
The design of the daisy chain was inspired by the NDI-
based catenanes and rotaxanes introduced by the group of
Sanders[12] and adapted by us[13] to construct switchable
[2]rotaxanes. We decided to couple a naphtho[35]crown-10
(N35C10) macrocycle to an alkyne-terminated NDI unit so as
to obtain the monomer N35C10-NDI (see the Supporting
Information), which was stoppered in CHCl3 with the azide-
bearing stopper precursor StN3 using a copper(I)-catalyzed
azide–alkyne cycloaddition[14] (Scheme 1). Purification of the
product mixture by flash column chromatography and
recycling gel permeation chromatography (GPC) in CHCl3
yielded the monomer 1 as the major product (68%) and the
[c2]daisy chain 2 as the most abundant minor product (7%).
SC0000989. This research is part of the Joint Center of Excellence in
Integrated Nano-Systems (JCIN) at King Abdulaziz City for Science
and Technology (KACST) and Northwestern University (NU)
(Project 34-494). The authors would like to thank both KACST and
NU for their continued support of this research. J.L. and G.A.V. were
supported by a grant from the Office of Naval Research (ONR Grant
N00014-13-1-0058). S.T.S. thanks the International Institute for
Nanotechnology (IIN) at Northwestern University (NU) for an IIN
Postdoctoral Fellowship and the QUEST high-performance com-
puting center at NU for a research allocation of computer time.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201308498.
Angew. Chem. Int. Ed. 2014, 53, 1953 –1958
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1953

.
Angewandte
Communications
Scheme 1. Synthesis of the neutral crown ether NDI monomer 1 and
the [c2]daisy chain 2 using the copper(I)-catalyzed azide–alkyne cyclo-
addition.
Figure 1. Partial 1H NMR spectra (600 MHz, CDCl3, 335 K) comparing
the chemical shifts for the aromatic protons in 1 (b) and 2 (c) with
those for the uncomplexed control compounds NDI(C9H18CCH)2 (a)
and N35C10-NHBoc (d).
The compound 2 has a shorter retention time than 1 on GPC
(see Figure S1a in the Supporting Information), an observa-
tion which is consistent with its larger size, while also verifying
that the stopper is bulky enough to prevent the ring from
deslipping over long timescales. Solutions of 1 and 2 are
reddish in color, owing to a charge-transfer (CT) interaction
between the NDI and DNP units in both compounds. The
absorption spectra (see Figure S1b) of 1 and 2 show CT
absorption bands at lmax = 516 and 470 nm, respectively, with
a substantially higher extinction coefficient for the CT band in
the case of 2. The existence of a stabilizing CT interaction in
1 explains the modest production of 2, since N35C10-NDI
self-complexation most likely competes with dimerization
during the stoppering reaction. The high-resolution mass
spectra of compounds 1 and 2 match well (see Figure S2) with
the predicted spectra. Since they are composed of identical
components, both compounds give rise to signals at m/z
ꢀ 1589 corresponding to the [M+H]+ ion of 1 and the
[M+2H]2+ ion of 2, with isotope patterns indicative of singly
and doubly charged species, respectively. We also observed
the [M+H]+ signal of 2 at m/z ꢀ 3179.
locked structure of the daisy chain. The significant line
broadening of the NDI and crown ether signals are only
observed in the case of 2, thus indicating that the correspond-
ing protons participate in exchange processes occurring on
a timescale similar to that of the NMR timescale.
We appealed to variable-temperature (VT) 1H NMR
spectroscopy (Figure 2) to evaluate the dynamics of 2. At
335 K, the exchange processes that take place in 2 are rapid
enough to give a single set of time-averaged signals. These
The 1H NMR spectra of 1 and 2 were assigned (see
1
Section S4.1) with the aid of two-dimensional H–1H corre-
lation spectroscopy. Figure 1 presents a comparison of the
chemical shifts of the aromatic protons of compounds 1 and 2
with the signals for the same protons in the uncomplexed NDI
and N35C10 controls, NDI(C9H18CCH)2 and N35C10-
NHBoc, in CDCl3 at 335 K. The 2/6, 3/7, and 4/8 DNP
protons as well as the NDI protons of both 1 and 2 uniformly
resonate at lower frequencies compared with those of the
control compounds. These shifts corroborate the evidence
from the absorption spectra, thus indicating that a DNP–NDI
CT interaction is being expressed in both compounds. The
resonance for the crown etherꢀs resorcinol proton HR1 is
dramatically shifted to lower frequencies in the [c2]daisy
chain 2 but not in the monomer 1, an observation which is
consistent with the stacking of resorcinol units on top of the
NDI unit, and is expected only in the mechanically inter-
1
Figure 2. The VT H NMR spectrum (600 MHz, CDCl3) of 2. The
triazole proton HTri (boxed in magenta) and adjacent methylene proton
Hk (boxed in black) are diagnostic signals for quantifying parameters
associated with translational isomerism.
1954
www.angewandte.org
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 1953 –1958

Angewandte
Chemie
signals broaden and then separate as the temperature is
elucidating the dynamics of molecular machines at the
1
lowered and exchange rates become slower on the H NMR
atomistic level. Although molecular dynamics (MD) simu-
lations have played an important role in helping us to
understand the one-dimensional sliding motions of rings
along dumbbells in simple [2]rotaxanes,[15] little effort has
been devoted to simulating and understanding the dynamics
of more complex rotaxane architectures such as daisy chains.
We have therefore used atomistic MD simulations in explicit
CHCl3 solvent to investigate the mechanism by which
isomerism occurs in 2. Since the two rings of 2 can potentially
slide along their dumbbells independently of each other, two
independent collective variables should be suitable to de-
scribe its correlated switching mechanism. We chose the end-
to-end distance (defined as the distance between the meth-
ylene carbon atoms of the stoppers) and the ring-to-ring
distance (defined as the distance between the centroids of
each crown ether) as the reaction coordinates of a two-
dimensional free-energy surface. At 228 K (Figure 3a), two
major free-energy basins are observed, which correspond to
the GSCC (Structure I) and the MSCC (Structures IVand V).
The MD simulations predict the GSCC to be favored by
(1.5 ꢂ 1.2) kcalmolꢁ1, with an activation barrier to transla-
tional isomerism of approximately 12 kcalmolꢁ1, which is in
reasonable agreement with the result from dynamic 1H NMR
spectroscopy. Remarkably, the free-energy basin correspond-
ing to the MSCC disappears (Figure 3b) when the simulations
are run at 335 K,[16] thus indicating that 2 is bistable at low
temperatures and monostable at high temperatures. This
unusual result most likely arises from the entropic contribu-
tion to the free energy which becomes more important at
higher temperatures, since the extended state of the GSCC
affords more conformational flexibility than the MSCC. The
ability to control the translational isomerism of 2 with
temperature provides a basis for entropy-driven[17] thermal
actuation. Beyond assisting in the interpretation of exper-
imental data, the MD calculations performed on 2 describe, in
unprecedented detail, the mechanism of the GSCC!MSCC
timescale. By contrast, the VT 1H NMR spectrum (see
Figure S8) of 1 shows no line broadening or coalescence
1
over the same temperature range. The large collection of H
signals of 2 at low temperatures points to a distribution of co-
conformational isomers which equilibrate slowly on the
1H NMR timescale. We attribute the minor set of signals to
a metastable state co-conformation (MSCC) in which the
crown ether macrocycles of 2 encircle the triazole units,
instead of encircling the NDI units, as expected in the ground-
state co-conformation (GSCC). This conclusion is supported
by the separation of the triazole proton HTri resonance and
that of the adjacent methylene proton Hk, each into two sets
of signals below 278 K. The more intense, higher frequency
HTri signal corresponds to the GSCC, while the lower
frequency signal represents the MSCC. The MSCC HTri
signal undergoes a second act of separation below 248 K,
most likely because the pirouetting motions of the triazole
ꢁ
and DNP units (rotations around their central C C bonds;
see Figure S10) also enter a slow exchange regime on the
1H NMR timescale. Integration of the HTri signals at 228 K
reveals a GSCC/MSCC ratio of approximately 4:3. Assuming
a coalescence temperature Tc of approximately 273 K for the
triazole signals, the rate kc governing the exchange between
GSCC and MSCC is estimated to be 69 sꢁ1. By relying on
lineshape analysis to estimate (Section S4.3) rate constants
above Tc, we have calculated a free energy of activation
(DGꢀ) of (13.65 ꢂ 0.04) kcalmolꢁ1 for the mechanical con-
traction/extension motion of the bistable daisy chain.
Although these parameters represent a weighted average of
the activation barrier for both MSCC!GSCC and GSCC!
MSCC conversions, the 4:3 GSCC/MSCC distribution corre-
sponding to DG8 < 0.2 kcalmolꢁ1 dictates that DGꢀ is nearly
equal for both processes.
Computer simulation is a powerful tool, not only for
validating the interpretation of experimental data, but also for
Figure 3. Two-dimensional free-energy profiles and graphical illustrations of representative structures for the translational isomerization of the
[c2]daisy chain 2, obtained with well-tempered metadynamics simulations in explicit CHCl3 at a) 228 K and b) 335 K. These results demonstrate
that temperature can be used to control the extension and contraction of 2, since bistability is maintained only at low temperatures.
Angew. Chem. Int. Ed. 2014, 53, 1953 –1958
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
1955

.
Angewandte
Communications
with respect to the two co-conformations adopted by the daisy
chain, associated with the crown ethersꢀ encirclement of
either the NDI or triazole units (denoted by the prefixes “ndi”
and “tri”) in each of its three NDI redox states (NDI, NDICꢁ,
and NDI2(Cꢁ)), which are populated according to the equilib-
rium constants K1, K2, and K3, respectively (Figure 5).
(I) In 2, process I (ndi2!ndi22(Cꢁ)) is shifted to higher
potentials (Figure 4a) with respect to 1 because the
crown ether hinders ET to NDI. The peak broadening
and increase in potential of process I at higher scan rates
(Figure 4b) can be rationalized by the fact that the
molecular dynamics become slow on the timescale of
the CV scan, and thus a wider range of ndi2 (co)-
conformations in which the crown ether more effec-
tively hinders ET to NDI are sampled at the electrode.
A very weak signal[18] occurring at the same potential as
the control might be assignable to the MSCC (tri2!
tri22(Cꢁ)). This assignment is supported by a growth in
intensity of the minor signal at lower temperatures (see
1
Figure S13), since the results from H NMR spectros-
copy and MD simulations suggest that the equilibrium
(K1) shifts toward the MSCC at lower temperatures.
(II) At slow scan rates, the daisy chain has time enough to
relax from ndi22(Cꢁ) to tri22(Cꢁ), which exposes the NDICꢁ
unit and yields identical potentials for process II in
compounds 1 and 2. This process is shifted to slightly
higher potentials at fast scan rates, most likely because
the scan rate surpasses the ndi22(Cꢁ)!tri22(Cꢁ) conversion
rate, thus leaving the NDICꢁ units encircled during the
generation of 24(Cꢁ).
Figure 4. Cyclic voltammograms (CV) of 1 and 2 (1 mm in CH2Cl2,
298 K, 0.1m NBu4PF6). a) A comparison between the CV of 1 and 2 at
50 mVsꢁ1 shows that the first reduction and final re-oxidation waves
(I/IV) of 2 are dramatically shifted with respect to 1, while the second
redox couple (II/III) is identical in both compounds. b) The variable
scan rate CV of 2 shows that the peak potential of processes I, II, and
IV are scan-rate-dependent.
conversion. The free energy surface in Figure 3a suggests that
2 does not contract by a concerted pathway. Instead, 2 is more
likely to adopt a compact, folded geometry stabilized by side-
on DNP–NDI interactions (Structure II) in which only one
ring encircles a triazole unit, before passing through an
unstable transition state (Structure III) to access the MSCC.
The MSCC, defined as any co-conformation with both rings
encircling the triazole units, has a broad basin and includes
a co-conformation (Structure IV) in which a side-by-side
NDI–NDI interaction, reminiscent of J-aggregation, tran-
spires. Although there is a distribution of tail-to-tail distances
in both GSCC and MSCC, it is clear that the GSCC is 1–2 nm
longer on average than the MSCC.
(III) The ET of process III oxidizes 24(Cꢁ) back to 22(Cꢁ) and is
essentially identical for 1 and 2 across all scan rates.
We have shown previously[13b–c] that electrochemical
reduction of the NDI unit eliminates its affinity for aromatic
crown-10 ethers, constituting a basis for operating a redox
switch. We have used cyclic voltammetry (Figure 4) at 298 K
in Ar-purged CH2Cl2 with 0.1m NBu4PF6 as a supporting
electrolyte to elucidate the redox-controlled switching mech-
anism of 2 at a concentration of 1 mm. All potentials are
referenced against an Ag/AgCl electrode (SSCE). Since 1 is
identical to the components of 2 in its constitution without
being interlocked, its cyclic voltammogram (CV) is an ideal
control for comparison (Figure 4a) with 2. The CV of 1 at
50 mVsꢁ1 shows two reversible redox processes at E1/2
=
ꢁ0.66 and ꢁ1.06 V, corresponding to the first and second
reduction, respectively, of NDI. The potential of these
processes are invariant (see Figure S12) over a wide range
of scan rates (10–1000 mVsꢁ1). Each of the electron-transfer
(ET) processes (I–IV in Figure 4) in the CV of 2 are discussed
Figure 5. The electrochemical switching mechanism which governs the
contractile actuation of 2.
1956
www.angewandte.org
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 1953 –1958

Angewandte
Chemie
3.97–3.90 (m, 28H), 3.87–3.84 (m, 8H), 3.74–3.70 (m, 4H), 3.63–3.69
Since the NDI unit is only exposed in the contracted
state of the daisy chain, we conclude that tri24(Cꢁ) vastly
outcompetes ndi24(Cꢁ) (K3 ! 1), as a shift in potential
would be expected at least in the fast scan regime if
ndi24(Cꢁ) was populated to a significant extent.
(m, 8H), 3.57–3.53 (m, 4H), 2.72 (t, J = 7.5 Hz, 4H), 1.83 (p, J =
7.5 Hz, 4H), 1.71 (p, J = 7.5 Hz, 4H), 1.53 (p, J = 7.5 Hz, 4H), 1.49–
1.38 (m, 16H), 1.32 ppm (s, 54H). 13C NMR (150 MHz, CDCl3,
335 K): d = 163.3, 162.6, 159.1, 156.0, 153.3, 152.2, 150.1, 148.6, 144.2,
141.0, 138.0, 132.5, 131.8, 131.5, 130.9, 125.6, 125.5, 125.2, 125.0, 124.1,
123.7, 121.5, 114.4, 113.4, 109.1, 103.9, 100.1, 71.5, 71.4, 71.2, 71.2, 70.2,
69.6, 67.6, 67.2, 66.8, 63.3, 53.7, 49.7, 42.6, 34.4, 31.4, 29.6, 29.5, 29.5,
29.4, 29.4, 28.4, 27.4, 25.8 ppm. HRMS (ESI-TOF-MS): m/z calcd for
(IV) Since the electron-rich crown ether can facilitate re-
oxidation of NDICꢁ by donating its electron density, the
significant shift in potential of process IV in 2 with
respect to 1 suggests that ndi22(Cꢁ) is also well-repre-
sented in the equilibrium population. By combining this
result with process II we conclude that K2 is of
intermediate strength (K3 < K2 < K1). 22(Cꢁ) most likely
equilibrates between extended and contracted states
more rapidly than 2, since having only a weak
[NDICꢁ···crown ether] interaction will lower the activa-
tion barrier for the shuttling process.
C
C
194H227N10O30 [M+H]+ 3178.6607, found 3178.6619; m/z calcd for
194H228N10O30 [M+2H]2+ 1589.8340, found 1589.8376.
Received: September 30, 2013
Revised: November 3, 2013
Published online: January 21, 2014
Keywords: cyclic voltammetry · donor-acceptor systems ·
.
molecular dynamics · molecular recognition · rotaxanes
The radical anion-driven electrochemical switching mech-
anism[19] portrayed in Figure 5 is further supported by
spectroelectrochemical experiments, which reveal (see Fig-
ure S14) absorption bands for 22(Cꢁ) that are characteristic of
NDI radical anions, with absorption maxima near l = 680 and
750 nm.
In conclusion, we have described a neutral donor–
acceptor molecular muscle having dual-mode switching
capabilities controlled by stimuli, either thermal or electro-
chemical, which are unprecedented for molecular daisy
chains. We have also demonstrated the power of molecular
dynamics simulations as a means of understanding atomisti-
cally the mechanisms of isomerism occurring in daisy chains.
This bistable [c2]daisy chain is an excellent candidate for
designing artificial molecular machines that function by
transducing thermal or electrochemical energy into mechan-
ical energy on the nanoscale.
[1] J. Rotzler, M. Mayor, Chem. Soc. Rev. 2013, 42, 44 – 62.
[2] a) M. C. Jimꢃnez, C. Dietrich-Buchecker, J.-P. Sauvage, Angew.
Chem. 2000, 112, 3422 – 3425; Angew. Chem. Int. Ed. 2000, 39,
3284 – 3287; b) M. Jimenez-Molero, C. Dietrich-Buchecker, J.-P.
Sauvage, Chem. Eur. J. 2002, 8, 1456 – 1466; c) M. C. Jimenez-
Molero, C. Dietrich-Buchecker, J.-P. Sauvage, Chem. Commun.
2003, 1613 – 1616; d) J.-P. Collin, C. Dietrich-Buchecker, P.
GaviÇa, M. C. Jimenez-Molero, J.-P. Sauvage, Acc. Chem. Res.
2001, 34, 477 – 487.
[3] A. Coskun, M. Banaszak, R. D. Astumian, J. F. Stoddart, B. A.
Grzybowski, Chem. Soc. Rev. 2012, 41, 19 – 30.
[4] a) J. Wu, K. C.-F. Leung, K. D. Benꢄtez, J.-Y. Han, S. J. Cantrill,
L. Fang, J. F. Stoddart, Angew. Chem. 2008, 120, 7580 – 7584;
Angew. Chem. Int. Ed. 2008, 47, 7470 – 7474; b) C. Romuald, E.
Busseron, F. Coutrot, J. Org. Chem. 2010, 75, 6516 – 6531.
[5] R. E. Dawson, S. F. Lincoln, C. J. Easton, Chem. Commun. 2008,
3980 – 3982.
[6] a) P. G. Clark, M. W. Day, R. H. Grubbs, J. Am. Chem. Soc. 2009,
131, 13631 – 13633; b) L. Fang, M. Hmadeh, J. Wu, M. A. Olson,
J. M. Spruell, A. Trabolsi, Y.-W. Yang, M. Elhabiri, A.-M.
Albrecht-Gary, J. F. Stoddart, J. Am. Chem. Soc. 2009, 131,
7126 – 7134; c) M. Hmadeh, L. Fang, A. Trabolsi, M. Elhabiri,
A.-M. Albrecht-Gary, J. F. Stoddart, J. Mater. Chem. 2010, 20,
3422 – 3430; d) G. Du, E. Moulin, N. Jouault, E. Buhler, N.
Giuseppone, Angew. Chem. 2012, 124, 12672 – 12676; Angew.
Chem. Int. Ed. 2012, 51, 12504 – 12508; e) C. J. Bruns, J. F.
Stoddart, Nat. Nanotechnol. 2012, 8, 9 – 10.
[7] a) S. J. Rowan, S. J. Cantrill, J. F. Stoddart, A. J. P. White, D. J.
Williams, Org. Lett. 2000, 2, 759 – 762; b) S.-H. Chiu, S. J. Rowan,
S. J. Cantrill, J. F. Stoddart, A. J. P. White, D. J. Williams, Chem.
Commun. 2002, 2948 – 2949; c) S. H. Ueng, S.-Y. Hsueh, C.-C.
Lai, Y.-H. Liu, S.-M. Peng, S.-H. Chiu, Chem. Commun. 2008,
817 – 819; d) E. N. Guidry, J. Li, J. F. Stoddart, R. H. Grubbs, J.
Am. Chem. Soc. 2007, 129, 8944 – 8945; e) M. Zhang, S. Li, S.
Dong, J. Chen, B. Zheng, F. Huang, Macromolecules 2011, 44,
9629 – 9634; f) O. A. Bozdemir, G. Barin, M. E. Belowich, A. N.
Basuray, F. Beuerle, J. F. Stoddart, Chem. Commun. 2012, 48,
10401 – 10403.
Experimental Section
Atomistic MD simulations of 2 were set up with the Desmond-
Maestro package[20] using the generic OPLS-AA force field[21] in an
explicit CHCl3 solvent box (ca. 90 ꢁ 90 ꢁ 90 ꢂ) with periodic boun-
dary conditions, and the constructs (ca. 30000 atoms) were equili-
brated with the six-step standard protocol for 60 ps. After equilibra-
tion, well-tempered metadynamics simulations,[22] using the tail-to-tail
and ring-to-ring distances as collective variables, were performed in
the NPT constant ensemble (228 and 335 K, 1 bar, Martyna–Tobias–
Klein coupling scheme[23]) with a 3 fs timestep. The metadynamics
simulations were terminated upon convergence ( ꢀ 600 to 800 ns).
2: DN35C10-NDI (65 mg, 64 mmol), StN3 (64 mg, 110 mmol), 2,6-
lutidine (8 mL), and [Cu(MeCN)4PF6] (10 mg, 27 mmol) were stirred
at ambient temperature in CHCl3 (700 mL) for 24 h under an
atmosphere of N2. A solution of EDTA (50 mg) in H2O (20 mL)
was added and the mixture was extracted with CH2Cl2 (3 ꢁ 10 mL).
The organic extracts were combined, concentrated, and subjected to
chromatography (SiO2: eluting first with 1:3 CH2Cl2/EtOAc and then
with 1:50 MeOH/EtOAc). The red-colored fractions were combined
and further purified by recycling GPC in CHCl3 to afford 1 (69 mg,
68%) and 2 (7 mg, 7%), where 2 was the faster-eluting product.
1H NMR (600 MHz, CDCl3, 335 K): d = 8.60 (s, 8H), 7.39 (s, 2H), 7.24
(d, J = 8.6 Hz, 12H), 7.10 (d, J = 8.9, 4H), 7.08 (d, J = 8.6 Hz, 12H),
6.90 (d, J = 8.4 Hz, 4H), 6.74 (d, J = 8.9 Hz, 4H), 6.41 (br, 4H), 6.38
(br, 4H), 5.96 (br, 4H), 5.11 (br, 2H), 4.64 (t, J = 5.2 Hz, 4H), 4.45 (br,
4H), 4.32 (t, J = 5.2 Hz, 4H), 4.14–4.11 (m, 4H), 4.06–4.00 (m, 12H),
[8] a) T. Hoshino, M. Miyauchi, Y. Kawaguchi, H. Yamaguchi, A.
Harada, J. Am. Chem. Soc. 2000, 122, 9876 – 9877; b) L. Cao, L.
Isaacs, Org. Lett. 2012, 14, 3072 – 3075; c) J. Rotzler, S. Drayss, O.
Hampe, D. Hꢅussinger, M. Mayor, Chem. Eur. J. 2013, 19, 2089 –
2101.
[9] N. H. Evans, P. D. Beer, Chem. Eur. J. 2011, 17, 10542 – 10546.
[10] G. Barin, A. Coskun, M. M. Foudah, J. F. Stoddart, Chem-
PlusChem 2012, 77, 159 – 185.
Angew. Chem. Int. Ed. 2014, 53, 1953 –1958
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
1957

.
Angewandte
Communications
[11] a) P. R. Ashton, I. W. Parsons, F. M. Raymo, J. F. Stoddart,
A. J. P. White, D. J. Williams, R. Wolf, Angew. Chem. 1998, 110,
2016 – 2019; Angew. Chem. Int. Ed. 1998, 37, 1913 – 1916; b) D.
Cao, C. Wang, M. A. Giesener, Z. Liu, J. F. Stoddart, Chem.
Commun. 2012, 48, 6791 – 6793; c) N. Yamaguchi, D. S. Nagve-
kar, H. W. Gibson, Angew. Chem. 1998, 110, 2518 – 2520; Angew.
Chem. Int. Ed. 1998, 37, 2361 – 2364; d) N. L. Strutt, H. Zhang,
M. A. Giesener, J. Lei, J. F. Stoddart, Chem. Commun. 2012, 48,
1647 – 1649.
[12] a) D. G. Hamilton, J. K. M. Sanders, J. E. Davies, W. Clegg, S. J.
Teat, Chem. Commun. 1997, 897 – 898; b) D. G. Hamilton, J. E.
Davies, L. Prodi, J. K. M. Sanders, Chem. Eur. J. 1998, 4, 608 –
620; c) D. G. Hamilton, M. Montalti, L. Prodi, M. Fontani, P.
Zanello, J. K. M. Sanders, Chem. Eur. J. 2000, 6, 608 – 617;
d) M. J. Gunter, N. Bampos, K. D. Johnstone, J. K. M. Sanders,
New J. Chem. 2001, 25, 166 – 173.
[13] a) S. A. Vignon, T. Jarrosson, T. Iijima, H.-R. Tseng, J. K. M.
Sanders, J. F. Stoddart, J. Am. Chem. Soc. 2004, 126, 9884 – 9885;
b) T. Iijima, S. A. Vignon, H.-R. Tseng, T. Jarrosson, J. K. M.
Sanders, F. Marchioni, M. Venturi, E. Apostoli, V. Balzani, J. F.
Stoddart, Chem. Eur. J. 2004, 10, 6375 – 6392; c) H.-P. Jacquot
de Rouville, J. Iehl, C. J. Bruns, P. L. McGrier, M. Frasconi,
A. A. Sarjeant, J. F. Stoddart, Org. Lett. 2012, 14, 5188 – 5191.
[14] a) V. V. Rostovtsev, L. G. Green, V. V. Fokin, K. B. Sharpless,
Angew. Chem. 2002, 114, 2708 – 2711; Angew. Chem. Int. Ed.
2002, 41, 2596 – 2599; b) C. W. Tornøe, C. Christensen, M.
Meldal, J. Org. Chem. 2002, 67, 3057 – 3064.
[15] a) S. S. Jang, Y. H. Jang, Y.-H. Kim, W. A. Goddard III, A. H.
Flood, B. W. Laursen, H.-R. Tseng, J.-F. Stoddart, J. O. Jeppesen,
J. W. Choi, D. W. Steuerman, E. DeIonno, J. R. Heath, J. Am.
Chem. Soc. 2005, 127, 1563 – 1575; b) Y. H. Jang, S. S. Jang, W. A.
Goddard III, J. Am. Chem. Soc. 2005, 127, 4959 – 4964; c) S. S.
Jang, Y. H. Jang, Y.-H. Kim, W. A. Goddard III, J. W. Choi, J. R.
Heath, B. W. Laursen, A. H. Flood, J. F. Stoddart, K. Nørgaard,
T. Bjørnholm, J. Am. Chem. Soc. 2005, 127, 14804 – 14816;
d) Y. H. Jang, W. A. Goddard III, J. Phys. Chem. B 2006, 110,
7660 – 7665; e) M. E. Foster, K. Sohlberg, J. Chem. Theory
Comput. 2007, 3, 2221 – 2233; f) H. Kim, W. A. Goddard III, S. S.
Jang, W. R. Dichtel, J. R. Heath, J. F. Stoddart, J. Phys. Chem. A
2009, 113, 2136 – 2143; g) K. Phoa, J. B. Neaton, V. Subramanian,
Nano Lett. 2009, 9, 3225 – 3229; h) Q. Zhang, Y. Tu, H. Tian, Y.-
L. Zhao, J. F. Stoddart, H. ꢂgren, J. Phys. Chem. B 2010, 114,
6561 – 6566; i) A. Credi, M. Semeraro, S. Silvi, M. Venturi,
Antioxid. Redox Signaling 2011, 14, 1119 – 1165; j) C. Casati, P.
Franchi, R. Pievo, E. Mezzina, M. Lucarini, J. Am. Chem. Soc.
2012, 134, 19108 – 19117; k) A. Coskun, J. M. Spruell, G. Barin,
W. R. Dichtel, A. H. Flood, Y. Y. Botros, J. F. Stoddart, Chem.
Soc. Rev. 2012, 41, 4827 – 4859.
[16] To rule out any bias which may arise from incomplete sampling,
MD simulations were also started from the MSCC, but these
experiments always resulted in fast equilibration to the GSCC at
335 K.
[17] For a rare example of entropy-driven switching in rotaxanes, see:
G. Bottari, F. Dehez, D. A. Leigh, P. J. Nash, E. M. Pꢃrez, J. K. Y.
Wong, F. Zerbetto, Angew. Chem. 2003, 115, 6066 – 6069; Angew.
Chem. Int. Ed. 2003, 42, 5886 – 5889.
1
[18] The same sample of 2 was used for all H NMR spectroscopic
and CVexperiments. Thus, the minor signal cannot be attributed
to an NDI impurity because there is no evidence for such an
impurity above the detection limit (< 5%) of 1H NMR spec-
troscopy. A similar minor signal for process I was also observed
in an analogous NDI-crown ether [2]rotaxane. See Ref. [13c].
[19] For other examples of electrochemically switchable [2]rotaxanes
driven by the reduction of imide-functionalized naphthalene
units, see: a) 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; b) G. Fioravanti, N. Haraszkie-
wicz, E. R. Kay, S. M. Mendoza, C. Bruno, M. Marcaccio, P. G.
Wiering, F. Paolucci, P. Rudolf, A. M. Brouwer, D. A. Leigh, J.
Am. Chem. Soc. 2008, 130, 2593 – 2601.
[20] a) Desmond Molecular Dynamics System, version 3.0, D. E.
Shaw Research, New York, NY, 2011. Maestro-Desmond
Interoperability Tools, version 3.0, Schrçdinger, New York,
NY, 2011; b) K. J. Bowers et al., “Scalable Algorithms for
Molecular Dynamics Simulations on Commodity Clusters,”
Proceedings of the ACM/IEEE Conference on Supercomputing
(SC06), Tampa, Florida, 2006, November 11 – 17; c) Z. Guo, U.
Mohanty, J. Noehre, T. K. Sawyer, W. Sherman, G. Krilov, Chem.
Biol. Drug Des. 2010, 75, 348 – 359; d) D. Shivakumar, J.
Williams, Y. Wu, W. Damm, J. Shelley, W. Sherman, J. Chem.
Theory Comput. 2010, 6, 1509 – 1519.
[21] G. A. Kaminski, R. A. Friesner, J. Tirado-Rives, W. L. Jorgen-
sen, J. Phys. Chem. B 2001, 105, 6474 – 6487.
[22] A. Barducci, G. Bussi, M. Parrinello, Phys. Rev. Lett. 2008, 100,
020603.
[23] G. J. Martyna, D. J. Tobias, M. L. Klein, J. Chem. Phys. 1994, 101,
4177 – 4191.
1958
www.angewandte.org
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 1953 –1958