A pentiptycene-derived light-driven molecular brake

Article abstract of DOI:10.1021/ol800689a

(Chemical Equation Presented) A room-temperature light-driven molecular brake (1), consisting of a pentiptycene rotator, a 3,5-dinitrophenyl brake, and a photoisomerizable ethenyl spacer, is reported. The rotation rates of the rotator differ by about 9 orders of magnitude between the brake-on (cis-1) and brake-off (trans-1) states.

Full text of DOI:10.1021/ol800689a

ORGANIC
LETTERS
2008
Vol. 10, No. 11
2279-2282
A Pentiptycene-Derived Light-Driven
Molecular Brake
Jye-Shane Yang,*,† Yao-Ting Huang,† Jinn-Hsuan Ho,† Wei-Ting Sun,†
Hsin-Hau Huang,† Ying-Chih Lin,*,† Shing-Jong Huang,† Shou-Ling Huang,†
Hsiu-Feng Lu,‡ and Ito Chao*,‡
Department of Chemistry and Instrumentation Center, National Taiwan UniVersity,
Taipei, Taiwan 10617, and Institute of Chemistry, Academia Sinica, Taipei,
Taiwan 11529
jsyang@ntu.edu.tw; yclin@ntu.edu.tw; ichao@chem.sinica.edu.tw
Received March 26, 2008
ABSTRACT
A room-temperature light-driven molecular brake (1), consisting of a pentiptycene rotator, a 3,5-dinitrophenyl brake, and a photoisomerizable
ethenyl spacer, is reported. The rotation rates of the rotator differ by about 9 orders of magnitude between the brake-on (cis-1) and brake-off
(trans-1) states.
The ability to control specific motions of subunits in
molecules or supramolecular ensembles is essential for the
development of molecular machines.1 As photons are cleaner,
faster, and longer-range control elements (stimuli) compared
to chemicals and electrons,2 many efforts have been devoted
to constructing prototypes of light-driven or photocontrollable
linear and rotary motions.1–3 However, the progress in some
aspects falls behind the corresponding chemical-controlled
systems. One particular example is “molecular brakes”, in
which the Brownian rotary motion of a rigid wheel compo-
nent about a single bond can be slowed down by a brake
unit. Since the first chemical-driven molecular brake reported
by Kelly and co-workers in 1994,4 several others have also
been developed.5 The first approach toward a light-driven
counterpart was reported by Feringa and co-workers in 1997
using a cis-trans interconvertable sterically overcrowded
alkene.6 However, the desired rotary motion is restricted for
both the trans and cis isomers at room temperature because
of unexpectedly high and similar rotational barriers (19.7
and 19.0 kcal mol-1, respectively, at 303 K). In a later report
on a similar system, the rate of rotation (krot) was determined
to be less than 0.1 s-1 for all the stable and unstable states
7
at 298 K. To date, an effective room-temperature photo-
controllable molecular brake has yet to be demonstrated.
† National Taiwan University.
(4) Kelly, T. R.; Bowyer, M. C.; Bhaskar, K. V.; Bebbington, D.; Garcia,
A.; Lang, F.; Kim, M. H.; Jette, M. P. J. Am. Chem. Soc. 1994, 116, 3657–
3658.
‡ Academia Sinica.
(1) Balzani, V.; Venturi, M.; Credi, A. Molecular DeVices and
Machines-A Journey into the Nano World; Wiley-VCH: Weinheim, 2003.
(2) (a) Raymo, F. M. Angew. Chem., Int. Ed. 2006, 45, 5249–5251. (b)
Credi, A. Aust. J. Chem. 2006, 59, 157–169. (c) Saha, S.; Stoddart, J. F.
Chem. Soc. ReV. 2007, 36, 77–92.
(5) (a) Jog, P. V.; Brown, R. E.; Bates, D. K. J. Org. Chem. 2003, 68,
8240–8243. (b) Annunziata, R.; Benaglia, M.; Cinquini, M.; Raimondi, L.;
Cozzi, F. J. Phys. Org. Chem. 2004, 17, 749–751. (c) Alfonso, I.; Burguete,
M. I.; Luis, S. V. J. Org. Chem. 2006, 71, 2242–2250.
(3) (a) Kottas, G. S.; Clarke, L. I.; Horinek, D.; Michl, J. Chem. ReV.
2005, 105, 1281–1376. (b) Kay, E. R.; Leigh, D. A.; Zerbetto, F. Angew.
Chem., Int. Ed. 2007, 46, 72–191. (c) Feringa, B. L. J. Org. Chem. 2007,
(6) Schoevaars, A. M.; Kruizinga, W.; Zijlstra, R. W. J.; Veldman, N.;
Spek, A. L.; Feringa, B. L. J. Org. Chem. 1997, 62, 4943–4948.
(7) ter Wiel, M. K. J.; van Delden, R. A.; Meetsma, A.; Feringa, B. L.
Org. Biomol. Chem. 2005, 3, 4071–4076.
72, 6635–6652
.
10.1021/ol800689a CCC: $40.75
Published on Web 05/03/2008
2008 American Chemical Society

We report herein the first example of a room-temperature
light-driven molecular brake (1) that displays distinct rates
of rotation in the brake-on versus brake-off states. The rigid
pentiptycene group8 in 1 serves as a four-bladed wheel, and
the trans-cis photoisomerizable dinitrostyryl group behaves
as a photoresponsive brake component (Figure 1). In the trans
under 60 °C, because the salt decomposes at higher tem-
peratures. The stilbene backbone was finally constructed by
the Wittig reaction of 4 and formylpentiptycene 5, the
synthesis of which has recently been reported.9 The resulting
trans and cis isomers of 1 (2: 1) can be readily separated by
column chromatography.
1
Figure 2 shows the aromatic region of H and 13C NMR
spectra of trans-1 and cis-1 in DMSO-d6 at 298 K. The
Figure 1. Molecular structures of a light-driven molecular brake
(1) corresponding to the schematic illustrations shown in the
Abstract graphic. The trans (trans-1) and cis (cis-1) isomers
correspond to the brake-off and brake-on states. The numerical
labels for protons and carbons are for discussion of the VT NMR
spectra.
form (trans-1, the brake-off state), the pentiptycene rotator
undergoes free rotation (krot ≈ 109 s-1) due to a small
rotational barrier (∆G‡ ≈ 4.5 kcal/mol at 298 K). However,
such a rotation is nearly blocked (krot ) 3 s-1, ∆G‡ ≈ 16.4
kcal mol-1 at 298 K) in the cis form (cis-1, the brake-on
state) as the result of an intercalation of the dinitrophenyl
group into the U-shaped cavities of the rotator. As a result,
the rate of rotation in the brake-on and brake-off states differs
by nearly 9 orders of magnitude at 298 K.
1
Figure 2
.
Aromatic region of (a) H and (b) 13C NMR spectra of
trans-1 and cis-1 in DMSO-d6 at 298 K. The solid circles denote
peaks due to the dinitrostyryl group and the Arabic numerals denote
the corresponding protons and carbons labeled in Figure 1.
presence of only one set of signals for trans-1 suggests that
both the pentiptycene rotator and the dinitrophenyl brake can
undergo free rotation about the Cvinyl-Caryl single bonds. In
contrast, the observation of two sets of signals for the
pentiptycene methine nuclei but not for the dinitrophenyl
nuclei indicates that rotation of the rotator but not the brake
component is slower than the NMR time scale. Except for
some quaternary carbons, the signals can be unambiguously
assigned on the basis of a series of 2D NMR experiments,
including COSY, HSQC, NOESY, and ROESY (Supporting
Information, Figures S5-S11), and the numerical labels in
Figure 2 correspond to the labeled protons and carbons in
Figure 1.
The synthesis of molecular brake 1 is outlined in Scheme
1. The commercially available 3,5-dinitrobenzyl alcohol 2
Scheme 1
The information about rotational barriers and rates for the
pentiptycene rotator in cis-1 can be obtained from variable-
temperature (VT) NMR studies and spectral simulations. As
shown in Figure 3a, some of the proton peaks at 298 K
coalesce at higher temperatures. For example, a coalescence
temperature (Tc) near 348 K is found for protons H3 and
was first converted with thionyl chloride to 3,5-dinitrobenzyl
chloride 3, which then reacted with triphenylphosphine to
form the phosphonium salt 4. The latter reaction was kept
H3′, corresponding to an energy barrier of ∆G‡
) 16.9
(348K)
( 0.2 kcal mol-1. However, the multiplicity of proton signals
and the presence of more than one type of spin systems in
(8) For a recent review on pentiptycene chemistry, see: Yang, J.-S.; Yan,
J.-L. Chem. Commun. 2008, 1501–1512.
(9) Yang, J.-S.; Ko, C.-W. J. Org. Chem. 2006, 71, 844–847.
Org. Lett., Vol. 10, No. 11, 2008
2280

Figure 3. Pentiptycene peripheral phenylene (blade) region of the (a) experimental proton and (b) carbon and (c) simulated carbon VT
NMR spectra of cis-1 (9 and 60 mM for proton and carbon, respectively, DMSO-d6, 500 MHz). Values of temperature (T, K) and
interconversion rate (k, s-1) between the two isoenergetic states are also given for every trace. Simulated spectra were obtained with the
Topspin 2.0 program.10
the phenylene blades of the pentiptycene rotator prevent a
direct spectral simulation. Thus, a VT 13C NMR experiment
was carried out, and the resulting spectra (Figure 3b) were
simulated (Figure 3c).10 The results suggest that the rotation
is nearly blocked at 298 K, and the rate constant (k) for
interconversion between the two isoenergetic conformers of
cis-1 is only 6 s-1. In other words, it takes ca. 0.33 s for a
360° rotation (i.e., krot ) k/2). The activation energy (Ea )
14.8 ( 0.5 kcal mol-1) and enthalpic (∆H‡ ) 14.1 ( 0.5
kcal mol-1) and entropic (∆S‡ ) - 7.6 ( 1.4 cal K-1 mol-1)
contributions to the free energy of activation (∆G‡ ) 16.4
and 16.8 kcal mol-1 at 298 and 348 K, respectively) were
obtained by Arrhenius and Eyring plots (Supporting Infor-
mation, Figures S12 and S13). Since the rotational barrier
is mainly due to an enthalpic factor, the transition state is
destabilized mainly due to steric effect. It should also be
noted that the values of ∆G‡ at 348 K deduced from 1H and
13C VT NMR are essentially the same.
the dinitrophenyl group as a result of severe steric interactions
with H1, H1′, and the bridgehead hydrogen atom, as revealed
by the transition structure (Figure 4b). Strains in bond angles
Figure 4. DFT-derived structures for cis-1: (a) the optimized
conformation and (b) the transition structure along the pentiptycene
rotation coordinate.
To obtain structural information about the transition
structure and the optimized conformation of cis-1 along the
pentiptycene rotation coordinates, we have carried out DFT
(BMK/6-311+G**//B3LYP/6-31G*) calculations.11 The cal-
culation results are justified by the good agreement of the
calculated (16.75 kcal mol-1) and the NMR-determined ∆G‡
value (16.4 kcal mol-1) at 298 K. The brake moiety in the
structurally optimized cis-1 (Figure 4a) is, as expected,
located at the U-shaped cavities, although the protruding
pentiptycene blades create two types of cavities (U- vs
V-shaped). The V-shaped cavities are in fact inaccessible to
and bond lengths also account in part for the high energy of
the transition state. The structural coordinates are supplied
as Supporting Information.
An attempt to evaluate the rotational barriers for the
pentiptycene rotator in trans-1 and the dinitrophenyl brake
in cis-1 with VT 1H NMR was hampered by their low energy
barriers, as we could not observe decoalescence of the signals
even at a temperature as low as 183 K (in CD2Cl2). In view
of the good agreement of DFT calculations and NMR
experiments in the pentiptycene rotation barrier of cis-1, DFT
calculations have been applied to retrieve the corresponding
information for that in trans-1 (4.45 kcal mol-1)12 and the
brake rotation in cis-1 (6.85 kcal mol-1). With a calculated
∆G‡ value differing by 12.3 kcal mol-1 for the pentiptycene
(10) All line shape simulations were performed using the Bruker Topspin
2.0 program.
(11) To expedite the calculations, the octyl group was replaced by a
methyl group. The BMK functional has been shown to be a good functional
in terms of thermochemistry and kinetics for main-group elements. Zhao,
Y.; Truhlar, D. G. Acc. Chem. Res. 2008, 41, 157–167.
Org. Lett., Vol. 10, No. 11, 2008
2281

rotation in trans-1 versus cis-1 at 298 K, the difference in
rotation rate is in the order of 109.13
Photoswitching between trans-1 and cis-1 has been
investigated in dichloromethane solutions.14,15 On the basis
of their relative molar absorptivity (ε), the wavelengths of
306 (εtrans-1/εcis-1 ) 3.35) and 254 nm (εtrans-1/εcis-1 ) 1.20)
were adopted for the trans-1 f cis-1 and the cis-1 f trans-1
conversion, respectively. The corresponding ratio of [trans-
1]/[cis-1] determined by 1H NMR is 25/75 and 45/55 in the
photostationary states. As monitored by absorption spectra
(Figure 5), photoswitching between the two photostationary
states is quite robust (Figure 5, inset).
In summary, the pentiptycene-derived stilbene 1 has been
prepared and investigated as a photocontrollable molecular
brake that functions at room temperature. Both experimental
and computational results reveal that at 298 K rotation of
the four-bladed pentiptycene (the rotator) is “free” in trans-1
but is nearly blocked in cis-1. The brake-on (cis-1) and brake-
off (trans-1) states differ by a rotation rate of ∼109-fold and
can be interconverted through the ethylene trans-cis pho-
toisomerization reactions. Studies on the related systems are
in progress in our laboratory in order to gain insights into
substituent effects on the brake performance.
Figure 5. Absorption spectra of trans-1 (curve a) and cis-1 (curve
d) and their photostationary states irradiated with alternating 306-
(curves c) and 254-nm (curves b) UV light irradiation in dichlo-
romethane. Inset shows the changes in absorbance at 322 nm
starting from trans-1 (10 µM) for 7 switching cycles.
Acknowledgment. Financial support for this work was
provided by the National Science Council, Academia Sinica,
and National Taiwan University. The computing time granted
by the National Center for High-Performance Computing and
the Computing Center of Academia Sinica is acknowledged.
We also thank Mr. G.-J. Huang for measuring the photoi-
somerization quantum yields.
(12) A barrier of ∼1000 cm-1 (2.86 kcal mol-1) was found for the
rotation of a single phenyl ring in trans-stilbene and its 4,4′-disubstituted
derivatives. Arp, Z.; Chiang, W.-Y.; Laane, J.; Sakamoto, A.; Tasumi, M.
J. Phys. Chem. A 2002, 106, 3479–3484.
(13) Even with a consideration of 1-2 kcal/mol uncertainty in calculated
∆G‡ value for pentiptycene rotation in trans-1, the ratio of the on-off rotation
rate is still around 107-108.
(14) The quantum yields for the trans-1 f cis-1 and cis-1 f trans-1
photoisomerization in dichloromethane are 0.17 and 0.21, respectively
.
Supporting Information Available: Experimental meth-
ods, synthesis of intermediates 3 and 4, 1D and 2D NMR
spectra and DFT-derived structural coordinates for trans-1
and cis-1, and the Arrhenius and Eyring plots for cis-1. This
material is available free of charge via the Internet at
http://pubs.acs.org.
(15) Thermal isomerization of the double bond in 1 has been evaluated
by DFT (UBMK/6-311+G**//UB3LYP/6-31G*) modeling. The calculated
free energy barrier was 45.22 kcal mol-1 relative to cis-1, and the free
energy difference between cis-1 and trans-1 was 0.76 kcal mol-1. These
values are similar to the computational and experimental data reported for
the parent stilbene (41-46 kcal mol-1 for the thermal isomerization barrier
and 2.3-5.7 kcal mol-1 for the energy difference between the cis and trans
isomers), see: (a) Meier, H. Angew. Chem., Int. Ed. 1992, 31, 1399–1420.
(b) Han, W.-G.; Lovell, T.; Liu, T.; Noodleman, L. ChemPhysChem 2002,
3, 167–178.
OL800689A
2282
Org. Lett., Vol. 10, No. 11, 2008