Synthetic Molecular Motors: Thermal N Inversion and Directional Photoinduced C=N Bond Rotation of Camphorquinone Imines

Article abstract of DOI:10.1002/anie.201506691

The thermal and photochemical E/Z isomerization of camphorquinone-derived imines was studied by a combination of kinetic, structural, and computational methods. The thermal isomerization proceeds by linear N inversion, whereas the photoinduced process occurs through C=N bond rotation with preferred directionality as a result of diastereoisomerism. Thereby, these imines are arguably the simplest example of synthetic molecular motors. The generality of the orthogonal trajectories of the thermal and photochemical pathways allows for the postulation that every suitable chiral imine qualifies, in principle, as a molecular motor driven by light or heat.

Full text of DOI:10.1002/anie.201506691

Angewandte
Chemie
International Edition: DOI: 10.1002/anie.201506691
German Edition: DOI: 10.1002/ange.201506691
Molecular Motors
Synthetic Molecular Motors: Thermal N Inversion and Directional
=
Photoinduced C N Bond Rotation of Camphorquinone Imines
Lutz Greb, Andreas Eichhçfer, and Jean-Marie Lehn*
[
6]
Abstract: The thermal and photochemical E/Z isomerization
of camphorquinone-derived imines was studied by a combina-
tion of kinetic, structural, and computational methods. The
thermal isomerization proceeds by linear N inversion, whereas
the photoinduced process occurs through C=N bond rotation
with preferred directionality as a result of diastereoisomerism.
Thereby, these imines are arguably the simplest example of
synthetic molecular motors. The generality of the orthogonal
trajectories of the thermal and photochemical pathways allows
for the postulation that every suitable chiral imine qualifies, in
principle, as a molecular motor driven by light or heat.
commercial compounds. The directionality was verified by
following the fate and interconversion of the diastereomeric
states A/A’/B/B’ (Figure 1a). More importantly, pursuing
a conjecture based on the orthogonality of the thermal and
photochemical configurational switching of imines, those
motors were tuned to have a two-step mode of action,
depending on the nature of the stator part and the N sub-
N
on-equilibrium processes and their coupling to directed
nanoscopic, microscopic, and, ultimately, macroscopic dis-
placements are intrinsic to life. Such tasks are usually
performed by biological macromolecules and have been
[
1]
perfected over millions of years. To unravel the basic
molecular mechanisms of directional displacement, it is
constructive to start with the simplest/smallest possible
molecular devices. Small-molecule systems acting as artificial
molecular machines and motors, with a 1000 times smaller
molecular weight than biological motor proteins, can be
classified by their mode of action (for example rotation,
Figure 1. Schematic comparison of a) a four-step and b) a two-step
motor.
[
2–4]
[7]
walking, contraction).
Among the most intensively studied
stituent (Figure 1b; A/A’). This mode of action is intrinsi-
synthetic molecular motors are the light-driven unidirectional
cally accessible only by imines and consists of a photochemical
[
3]
rotating overcrowded olefins. By fine adjustment of their
steric or electronic parameters, it has been shown in a striking
number of derivatives that the factors influencing the rota-
tional frequency of such motors are well understood and can
out-of-plane C=N bond rotation and a thermal in-plane
[8]
N inversion. The thermal process has experimentally and
[
9]
theoretically been proven to be a linear N inversion,
whereas the photochemical C=N bond isomerization has
[
4]
be tuned. Such devices have been shown to indeed effect
macroscopic changes by their coupling to surfaces or to
been only theoretically confirmed to be a rotational process,
[
10]
both in the S (p!p*) and in the S (n!p*) state. The
2
1
[
5]
hydrogels. Nevertheless, the field of light-driven rotating
molecular motors with functionalities other than olefins
remains little explored.
directionality is based on the assumption that the photo-
chemical C=N bond rotation should exhibit preferred direc-
tionality due to displacement through diastereotopic subspa-
ces spanned by chirality in the molecule, and thus ensuing
unsymmetrical excited-state surfaces. Thus, the two-step
mode-of-action should impart a motor-like motion, based
on analogy with the behavior of olefinic double bonds and on
theoretical considerations and plausibility. If imines with
chirality at a suitable distance indeed undergo a photochem-
ical C=N bond rotation with preferred directionality, conse-
We demonstrated recently that a four-step mode of
displacement of the rotor part around the stator could be
accomplished by using the special features of chiral imines,
easily synthesized in a one-step procedure starting from
[
*] Dr. L. Greb, Prof. Dr. J.-M. Lehn
Institut de Science et d’IngØnierie SupramolØculaires (ISIS)
UniversitØ de Strasbourg, 8 allØe Gaspard Monge
Strasbourg 67000 (France)
quently every suitable chiral imine would qualify as a molec-
ular motor. This essential question remains to be unambig-
uously answered.
Direct experimental proof of the photoinduced trajectory
is precluded by the fast dynamics in the excited state. It is thus
necessary to gain insight into the photochemical C=N bond
E-mail: lehn@unistra.fr
Dr. L. Greb, Dr. A. Eichhçfer
Institut für Nanotechnologie (INT)
Karlsruhe Institut für Technologie, Hermann-von-Helmholtz-Platz 1
7
6344 Eggenstein-Leopoldshafen (Germany)
rotation by indirect methods. Trapping of a photochemically
generated species on its rotational pathway in the two
different diastereoisomers would indicate a preferred direc-
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201506691.
Angew. Chem. Int. Ed. 2015, 54, 14345 –14348
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
14345

.
Angewandte
Communications
Table 1: Thermodynamic, kinetic, photochemical, and calculated
transition-state energies of 1–3.
tionality of rotation (Figure 1b; trapped states). Indeed, the
observation of diastereoselectivity in the N-cyclopropylimin-
Imine
1
2
3
1
-pyrroline photo-rearrangement of chiral N-cyclopropyl-
(R=cyclopropyl) (R=tBu) (R=CH CF )
2
3
benzylidene imines was originally used to elucidate the
underlying mechanism and it was concluded by comparison
E/Z (thermal)
92:8
36:64
99:1
24.7Æ0.6
À1.1Æ0.05
24.7
99:1
21:79
99:1
22.8Æ0.5
5.7Æ0.2
21.2
82:18
29:71
99:1
[
[
a]
b]
of experimental and theoretical results that a biased rotation
E/Z (PSS at 365 nm)
E/Z (PSS at 455 nm)
around the C=N bond plays the decisive role.[11]
°
À1
DH [kcalmol
]
19.0Æ0.8
Herein, we present a new class of synthetic imine
molecular motors, incorporating the natural product camphor
and its corresponding chirality, as essentially the simplest
example of a synthetic molecular motor. Identification of the
thermal N inversion pathway is supported by kinetic analysis
of the thermal isomerization in comparison with X-ray
molecular structures and computed transition-state energies.
The directionality in the photochemical C=N bond isomer-
°
DS [e.u.]
À19.7Æ1.9
24.7
°
À1
DG [ [kcalmol
]
293]
°
À1 [c]
calc. DE [kcalmol
]
24.2
21.7
25.6
[
[
[
a] Measured using a UV handlamp (6 W, l=365 nm) irradiating for 5 h.
b] Measured using a LED (5 W, l=455 nm) irradiating for 1 h.
c] Calculated electronic energies for linear transition states at the
B2PLYP-D3/def2-TZVP level + COSMO solvent correction. For further
details, see the Supporting Information.
ization is investigated by trapping of biradical intermediates
in a N-cyclopropylimin-1-pyrroline photo-rearrangement.
The orthogonality of the thermal and photochemical path-
ways thus highlights chiral imines as a unique class of
molecular motors.
half-life t1/2 = 10 min for 2, 7.5 h for 1). This difference would
not be expected if solely electronic factors were considered,
and instead indicates a steric ground-state destabilization for
2. Accordingly, the molecular structure of 2 shows a noticeable
enlargement of the C=NÀC bond angle (126 88 ) compared to
The camphorquinone-derived imines (1–3) were synthe-
sized by TiCl -mediated condensation of commercially avail-
4
able (R)-(À)-camphorquinone with suitable amines (specifi-
the less bulky and thermally more stable 1 (1198; Figure 2).
Interestingly, no significant torsion (pre-twist) around the
C=N bond was detected in either case, in agreement with
cally tert-butyl-, cyclopropyl-, or 2,2,2-trifluoroethylamines)
[
12,13]
to obtain the desired products in good yields (Scheme 1).
1
H-NOESY NMR and X-ray crystallographic structure analy-
sis confirmed the E configuration of the imines as the
thermodynamically favored compounds (Figure 2).
a linear N inversion pathway on the ground-state surface. The
activation entropies for 1 and 2 were small and concentration
independent, indicative of unimolecular processes. The lower
activation enthalpy and notably negative activation entropy
of 3 may be rationalized in terms of n-to-s* hyperconjugation
Irradiation of solutions of imines with UV light at l =
3
65 nm in CD CN resulted in the formation of photosta-
3
[
14]
tionary states (PSSs) containing imines in the Z configuration
in 64–79% yield (Table 1). The yellow-colored Z isomers
were entirely converted back into the E isomers upon
irradiation with blue light (l = 455 nm). The switching could
be cycled up to ten times without significant side reactions,
thus characterizing these imines as new two-way photo-
switches. The thermal Z-to-E back-isomerization was inves-
of the nitrogen lone pair to the s* (C-CF ) orbital. This
3
hyperconjugation would lead to stabilization of the transition
state but only in one over three possible conformations, thus
reducing the rotational degrees of freedom. Structures and
energies of both configurational isomers and transition states
were calculated by means of DFT (B2PLYP-D3/def2-TZVP/
[
15]
COSMO). The calculated bond angles and bond lengths are
in good agreement with the available X-ray solid-state
molecular structures and the relative energies of the isomers
reflect the preference for the E isomers. The transition-state
structures along the ground-state surface correspond in all
cases to a linear C=NÀC geometry. The calculated energies fit
1
tigated by time- and temperature-dependent H NMR spec-
°
troscopy, yielding free energies of activation (DG ) ranging
À1
from circa 21–25 kcalmol . The relative thermal stability of
the Z isomers of 1 and 2 was markedly different (at 293K,
with the experimentally determined energies for the thermal
Z-to-E back-isomerization of compounds 1–3 (Table 1), thus
confirming the N inversion pathway in the ground state.
Irradiation of a solution of the N-cyclopropylimine 1 in
CH CN or D O with light of shorter wavelength (l = 254 nm)
Scheme 1. Photochemical and thermal isomerization of camphorqui-
none imines. cPr=cyclopropyl.
3
2
than necessary for E/Z isomerization (l = 365 nm) effected
the photochemical N-cyclopropylimin-1-pyrroline rearrange-
ment (Scheme 2). The product distribution of the crude
mixture was determined by integration of the signals in the
1
H NMR spectrum for the imines and more accurately by gas
chromatography. The rearrangement reaction was performed
on a preparative scale (200 mg). The products were isolated
1
and their structures were confirmed by two-dimensional H-
1
COSY, H-NOESY, HSQC, and HMBC NMR spectra. In
addition to the expected pyrroline products 4, ethylene gas
( H NMR signal at d = 5.41 ppm) was formed as well as
isonitrile 5 (42%; confirmed by X-ray crystallographic
Figure 2. Molecular structures of imines 1 (left) and 2 (right). Values
1
given are the C
=NÀC bond angles. Atom colors: C=blue; H=white;
O=red; N=pink.
1
4346 www.angewandte.org
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2015, 54, 14345 –14348

Angewandte
Chemie
light from the visible part of the electromagnetic spectrum led
to formation of a PSS with an E/Z ratio of 78:22. Conse-
quently, the probability that a molecule now starting from the
E isomer reaches the S state, with subsequent reaction to
2
form endo-4, might be increased, as reflected in a higher
diastereomeric ratio.
Applied in the context of molecular motors, the N-
cyclopropyl imine rearrangement serves as a tool to inves-
tigate directionality in the photoinduced C=N bond rotation
Scheme 2. Photochemical N-cyclopropylimin-1-pyrroline rearrangement
of 1.
and should generally allow quantification of the efficiency of
imine molecular motors (and more precisely of the stator
parts). Of course, it must be noted that the rotational pathway
of the intermediate open-shell biradical species (from the
S2 state) may have a different selectivity than that of the pure
structure determination), most likely by a photoinduced
[16,17]
[
2+2] cycloreversion reaction.
Most importantly, the two
diastereomeric spirocyclic pyrroline products endo-4/exo-4
were obtained with an unbalanced d.r. value of 58:42, also
indicating a preferred directionality in the C=N bond rotation
C=N bond rotation (from the S state), but it should never-
1
(
see below).
theless serve as a guiding value.
Our assignment of the singlet nature of the excited states
The present work gives insight into the unique orthogonal
trajectories of thermal and photochemical isomerization of
chiral imines, finally providing experimental support for the
for E/Z isomerization (l = 365 nm) and photo-rearrangement
l = 254 nm) was supported by the photoreaction of 1 in the
presence of a 10-fold excess of piperylene, without a signifi-
(
[
7]
original conjecture. The camphorquinone-derived imines
[
18]
[11]
[19]
cant loss in efficiency. In analogy to the literature, we
consider that upon irradiation with light at l = 254 nm, the
S2 state leads to the opening of the cyclopropyl ring to give
(or other small-molecule chiral imines) are arguably the
simplest prototypical example of synthetic molecular motors.
The thermal N inversion is confirmed by kinetic analysis in
combination with molecular structures and calculated ener-
gies of the transition states. The diastereospecific trapping of
biradical intermediates provides support for the directional
photochemical C=N bond rotation. Consequently, every
a biradical before reaching the S /S2 conical intersection
1
(
CIS2/S1; Scheme 3). Thereafter, the pathway chosen on the
chiral imine could serve potentially as a molecular motor in
response to light and heat, driven by light as the cleanest fuel
and relaxing towards thermal equilibrium.
Scheme 3. Proposed ring opening and diastereoselective ring closure
after biased C=N bond rotation.
Acknowledgements
L.G. thanks the Alexander von Humboldt foundation for
a postdoctoral fellowship. We also thank the ERC (Advanced
Research Grant SUPRADAPT 290585) for financial support.
Exploratory experiments have been performed by Dr. Nema
Hafezi on another type of system. We thank Prof. Stefan
Bräse for sharing the Rayonet Photoreactor.
S1 surface, causing the C=N bond rotation, is guided by the
relative energies along the rotation vectors through the
diastereomeric spaces, reflected in the ratio obtained in the
final diastereospecific ring closing of the biradical. Impor-
tantly, besides the photo-rearrangement to the pyrroline,
E/Z isomerization took place at l = 254 nm with much
higher efficiency. The photostationary state at l = 254 nm
Keywords: density functional calculations · imines ·
isomerization · molecular motors · photochemistry
(
E/Z ratio = 48:53) was reached within 30 minutes of irradi-
How to cite: Angew. Chem. Int. Ed. 2015, 54, 14345–14348
Angew. Chem. 2015, 127, 14553–14556
ation whereas the complete conversion to 4/5 for the same
sample took 3 hours. The Frank–Condon regions and corre-
sponding relaxation pathways in the excited state should be
different in shape, depending on whether they are reached
from the E or the Z ground-state isomer. It can be assumed
that the preferred direction of rotation is different from both
sides, and that the observed d.r. value (58:42) is a lower-limit
level reflecting the PSS (48:53), hence concealing a higher
inherent selectivity. In other words, it is likely that if every
excitation event commenced solely from the E isomer,
a higher selectivity might be expected. Indeed, when the
reaction was performed with a medium-pressure mercury
[
[
1] a) M. Schliwa, Molecular motors, Wiley-VCH, Weinheim, 2003;
b) M. Schliwa, G. Woehlke, Nature 2003, 422, 759 – 765.
2] a) J.-P. Sauvage, V. Amendola, Molecular machines and motors,
Springer, Berlin, 2001; b) T. R. B. V. Kelly, Molecular machines,
Springer, Berlin, 2005; c) E. R. Kay, D. A. Leigh, F. Zerbetto,
Angew. Chem. Int. Ed. 2007, 46, 72 – 191; Angew. Chem. 2007,
119, 72 – 196; d) A. Credi, S. Silvi, M. Venturi, W. R. Browne,
Molecular machines and motors: recent advances and perspec-
tives, Springer, Berlin, 2014; e) D.-H. Qu, Q.-C. Wang, Q.-W.
Zhang, X. Ma, H. Tian, Chem. Rev. 2015, 115, 7543 – 7588.
3] N. Koumura, R. W. J. Zijlstra, R. A. van Delden, N. Harada,
B. L. Feringa, Nature 1999, 401, 152 – 155.
[
[
lamp in D O (irradiating with light of wavelengths over the
complete UV/Vis spectrum), the observed diastereomeric
ratio for endo-4/exo-4 increased to 68:32. Irradiation with
2
4] a) M. M. Pollard, M. Klok, D. Pijper, B. L. Feringa, Adv. Funct.
Mater. 2007, 17, 718 – 729; b) M. Klok, N. Boyle, M. T. Pryce, A.
Angew. Chem. Int. Ed. 2015, 54, 14345 –14348
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 14347

.
Angewandte
Communications
Meetsma, W. R. Browne, B. L. Feringa, J. Am. Chem. Soc. 2008,
30, 10484 – 10485; c) M. M. Pollard, A. Meetsma, B. L. Feringa,
1985, 68, 45 – 55; f) I. Tavernelli, U. F. Rçhrig, U. Rothlisberger,
Mol. Phys. 2005, 103, 963 – 981.
1
Org. Biomol. Chem. 2008, 6, 507 – 512; d) A. S. Lubbe, N.
Ruangsupapichat, G. Caroli, B. L. Feringa, J. Org. Chem. 2011,
[11] a) P. J. Campos, A. Soldevilla, D. Sampedro, M. A. Rodríguez,
Org. Lett. 2001, 3, 4087 – 4089; b) A. Soldevilla, D. Sampedro,
P. J. Campos, M. A. Rodríguez, J. Org. Chem. 2005, 70, 6976 –
6979; c) D. Sampedro, A. Soldevilla, M. A. Rodríguez, P. J.
Campos, M. Olivucci, J. Am. Chem. Soc. 2005, 127, 441 – 448.
[12] For experimental details, see the Supporting Information.
[13] S. E. Denmark, I. Rivera, J. Org. Chem. 1994, 59, 6887 – 6889.
[14] T. Asano, H. Furuta, H. J. Hofmann, R. Cimiraglia, Y. Tsuno, M.
Fujio, J. Org. Chem. 1993, 58, 4418 – 4423.
7
6, 8599 – 8610; e) J. Bauer, L. Hou, J. C. M. Kistemaker, B. L.
Feringa, J. Org. Chem. 2014, 79, 4446 – 4456.
[
5] a) T. Kudernac, N. Ruangsupapichat, M. Parschau, B. Macia, N.
Katsonis, S. R. Harutyunyan, K.-H. Ernst, B. L. Feringa, Nature
2
011, 479, 208 – 211; b) Q. Li, G. Fuks, E. Moulin, M. Maaloum,
M. Rawiso, I. Kulic, J. T. Foy, N. Giuseppone, Nat. Nanotechnol.
015, 10, 161 – 165.
2
[
[
[
6] L. Greb, J.-M. Lehn, J. Am. Chem. Soc. 2014, 136, 13114 – 13117.
7] J.-M. Lehn, Chem. Eur. J. 2006, 12, 5910 – 5915.
8] a) H. Kessler, D. Leibfritz, Tetrahedron 1970, 26, 1805 – 1820;
b) H. Kessler, D. Leibfritz, Tetrahedron Lett. 1970, 11, 1423 –
[15] For computational details, see the Supporting Information.
[16] a) H. Quast, G. Meichsner, Chem. Ber. 1987, 120, 1049 – 1058;
b) H. Quast, A. Fuß, A. Heublein, H. Jakobi, B. Seiferling,
Chem. Ber. 1991, 124, 2545 – 2554.
1
426; c) H. Kessler, Tetrahedron 1974, 30, 1861 – 1870; d) R.
[17] For X-ray solid-state molecular structure and a proposed
Knorr, J. Ruhdorfer, J. Mehlstäubl, P. Bçhrer, D. S. Stephenson,
Chem. Ber. 1993, 126, 747 – 754.
9] a) J. M. Lehn, B. Munsch, Theor. Chim. Acta 1968, 12, 91 – 94;
b) R. Macaulay, L. A. Burnelle, C. Sandorfy, Theor. Chim. Acta
mechanism,
see
the
Supporting
Information.
CCDC 1411350 (1), 1411351 (2), and 1411352 (5) contain the
supplementary crystallographic data for this paper. These data
are provided free of charge by The Cambridge Crystallographic
Data Centre.
[
1
973, 29, 1 – 7; c) H. Yamataka, S. C. Ammal, T. Asano, Y. Ohga,
Bull. Chem. Soc. Jpn. 2005, 78, 1851 – 1855; d) A. Gaenko, A.
Devarajan, L. Gagliardi, R. Lindh, G. Orlandi, Theor. Chem.
Acc. 2007, 118, 271 – 279; e) J. Gµlvez, A. Guirado, J. Comput.
Chem. 2010, 31, 520 – 531; f) S. He, Y. Tan, X. Xiao, L. Zhu, Y.
Guo, M. Li, A. Tian, X. Pu, N.-B. Wong, J. Mol. Struct.
Theochem. 2010, 951, 7 – 13.
[18] As a result of the concurrent thermal E/Z isomerisation and
experimental requirements, the exact determination of quantum
yields was not pursued.
[19] The imines reported in Ref. [11c] are not molecular motors since
the stereoinformation in those derivatives is lost by very fast
racemization of the chiral cyclopropyl rings.
[
10] a) P. Russegger, Chem. Phys. 1978, 34, 329 – 339; b) Y. Osamura,
S. Yamabe, K. Nishimoto, Int. J. Quantum Chem. 1980, 18, 457 –
4
62; c) P. Russegger, Chem. Phys. Lett. 1980, 69, 362 – 366; d) V.
Bonacic-Koutecky, M. Persico, J. Am. Chem. Soc. 1983, 105,
388 – 3395; e) V. Bona cˇ i c´ -Kouteck y´ , J. Michl, Theor. Chim. Acta
Received: July 20, 2015
Published online: October 9, 2015
3
1
4348 www.angewandte.org
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2015, 54, 14345 –14348