Green light powered molecular state motor enabling eight-shaped unidirectional rotation

Molecular motors convert external energy into directional motions at the nano-scales. To date unidirectional circular rotations and linear motions have been realized but more complex directional trajectories remain unexplored on the molecular level. In this work we present a molecular motor powered by green light allowing to produce an eight-shaped geometry change during its unidirectional rotation around the central molecular axis. Motor motion proceeds in four different steps, which alternate between light powered double bond isomerizations and thermal hula-twist isomerizations. The result is a fixed sequence of populating four different isomers in a fully unidirectional trajectory possessing one crossing point. This motor system opens up unexplored avenues for the construction and mechanisms of molecular machines and will therefore not only significantly expand the toolbox of responsive molecular devices but also enable very different applications in the field of miniaturized technology than currently possible.


HTI Synthesis
Precursors 2 and 3 were prepared following a previously reported protocol according to the following scheme. 1
Meta-chloroperoxybenzoic acid (235 mg, 1.36 mmol, 1 equiv.) was added at 0 °C in portions and the solution was further stirred for 10 min at 0 °C. Subsequently, a saturated sodium carbonate solution was added until the pH was adjusted to 7, the aqueous phase was extracted with CH2Cl2 (3 x 50 mL), and the combined organic phases were dried over Na2SO4. After removing the solvent in vacuo the crude product was purified by column chromatography (SiO2, ihex/EtOAc = 7/3  1/1). The isomers 1-A and 1-D were further purified by HPLC (Chiralpak IC Semiprep column, nheptane/EtOAc = 7/3). All isomers were separately crystallized from nheptane/EtOAc. The four isomers of the title compound were isolated as orange to red crystals (405 mg, 0.97 mmol, 71 %, sum of all isomers).

Physical and photophysical properties Thermal isomerizations
Amberized NMR tubes were charged with 1 mg to 2.5 mg of the respective isomer A, B, C, or D and 0.7 mL of 1,2-dichlorobenzene-d4, DMSO-d6, or MeCN-d4/D2O: 8/2. Subsequently the NMR tubes were heated to the appropriate temperatures for isomerization reactions to occur and the kinetics were followed by 1 H NMR measurements in defined time intervals. The unchanging equilibrium ratio of isomers after prolonged heating were obtained from integration of the corresponding signals in the 1 H NMR spectrum.
The kinetics of thermal isomer interconversions differ strongly in the different solvents, however the most prominent isomerization is the thermal Hula-Twist isomerization of E configured C and D to Z configured B and A, respectively in all solvents. These isomerizations are unimolecular first order reactions and proceed towards an equilibrium composition with only the two interconverting isomers present. The kinetics can therefore generally expressed for the reaction XY according to Supplementary Equation 1. Likewise the reverse rate constant is defined and can be determined from separate measurements starting from the opposite isomer Y.
By using the Eyring equation (Supplementary Equation 4) the free activation enthalpies ΔG * can be calculated from the rate constants of the corresponding reaction.
The obtained free activation enthalpies ΔG * for the thermal isomerizations between A to B, C to D, and vice versa in the different solvents as well as the corresponding extrapolated half-lives at 25 °C are given in Table S1 together with the equilibrium compositions obtained at high temperatures.
However in nonpolar solvents like 1,2-dichlorobenzene-d4 not only the thermal hula twist reaction occurs but also single bond rotations and double bond isomerizations. To model the kinetics of this more complicated scenario the Markov matrix approach from our previous publication 2 was used by replacing the probability elements of the matrix with the converted Eyring equation (Supplementary Equation5).

(Supplementary Equation 5)
The kinetics were then modelled by multiplication of M1 with the concentrations of the isomers A, B, C, D to give the corresponding concentration vector of the next time point (Supplementary Equation6). The sum of the concentrations of the isomers were previously normalized to 100 to enable an easier presentation of the data. [C] [D] • The different ΔG * values for each isomerization were optimized against a refinement factor R, which was calculated according to Supplementary Equation 7.
With [Xm]t being the measured concentration of X at time t, [Xc]t being the calculated concentration of X at time t, and x being the total amount of acquired measurement points. The closer R becomes to 1 the better is the obtained fit of the data. Measurement points below a concentration of 0.01 were not taken into account for the calculation of the R factor as the error of integration might be very high at low concentrations.
From these measurements we could quantify the full ground state energy profile of HTI 1 in 12DCB. The equilibrium isomer composition at high temperatures delivers the energy differences between the corresponding states according to the relation of the change of Gibbs free energy and the equilibrium constant.

Quantum yield measurements in different solvents
The photochemical quantum yields of the different photoconversion reactions φ were calculated as the ratio between the numbers of isomerized molecules n(molecules isomerized) and the number of absorbed photons to be minimum structures since no imaginary frequencies have been found. All other structures were shown to be first order saddle points on the potential energy surface since only one imaginary vibrational mode has been found confirming them to be transition state structures.

Crystal structure analysis
All X-ray intensity data were measured on a Bruker D8 Venture TXS system equipped with a multilayer mirror optics monochromator and a Mo Kα rotating-anode X-ray tube (λ = 0.71073 Å). The data collections were performed at 103 K. The frames were integrated with the Bruker SAINT Software package 5 . Data were corrected for absorption effects using the Multi-Scan method (SADABS) 6