doi: /nature10587 Contents

Transcription

1 SUPPLEMENTARY INFormatION doi: /nature10587 Contents Section 1 Section 2 Section 3 Section 4 Section 5 STM imaging details Analysis of the linearity of the movement Additional examples of control of movement of the molecule Stereoisomers and adsorption geometries of the four-wheeled molecule Synthesis and characterisation of both meso and racemic isomers of four-wheeled molecule 1

2 Section 1. STM imaging details and model figures Supplementary Figure 1 Sketch of the manipulation experiment. The STM tip is positioned above the molecule and a voltage pulse is applied with voltages higher than 600 mv. The tipsurface distance is adjusted to ensure that the current remains between 30 and 50 pa during the manipulation pulse. Subsequently, the surface is scanned under mild conditions, i.e., with voltages and currents below 60 mv and 50 pa, respectively (the bias voltage (U) is applied to the sample). The resulting STM image shows where the molecule moved, after the voltage pulse was applied. The whole process is repeated multiple times. The sum of the individual steps shows the trajectory travelled by the molecule. Helix-reversal manipulations are performed correspondingly, but without induction of movement at voltages between 200 and 450 mv. Supplementary Figure 2 Stability of the STM contrasts upon continuous scanning at mild conditions. a, 3D view of a typical four-lobe STM contrast ((R,R-R,R)- or (S,S-S,S)-enantiomer). b-d, Three subsequent STM images (6.9 nm x 7.6 nm, I=43 pa, U=47 mv) recorded from the same area of the surface. No modification of STM contrasts is observed on repeated scanning. 2

3 Discussion on STM contrasts of the molecule Relevant factors contributing to STM contrast of the molecule: 1) Stereochemistry of the compound used; meso-(r,s-r,s)-isomer or (R,R-R,R), (S,S-S,S) and (R,R-S,S) isomer. 2) Landing geometry; each molecule can land on a surface from two different sides with an additional possibility of flipping the front and the back halves with respect to each other. 3) Conformation of the motor wheels; each motor wheel can adopt two different conformations, a stable and a metastable conformation (both can be observed at 7 K). The number of possible conformers for each isomer with the specific landing geometry is 16 (2 4 = 16 conformers). 4) Conformation of the alkyl side chains (omitted from the model for clarity). Supplementary Figure 3 Variations of STM contrast for different conformers of (R,R- R,R)- or (S,S-S,S)-enantiomers. a, STM image of four molecules adsorbed on the surface upon sublimation. Different STM contrasts are related to different conformations of the four motor wheels of the molecule. b-j, Randomly chosen frames (scanning parameters, 7.0 nm x 7.8 nm, I = 47 pa, U = 47 mv) of the Supplementary Movie 1 showing variations of STM contrasts of the same molecule after single movement steps when voltage pulses were applied (manipulation conditions: I = pa, U = mv). The STM contrasts are reversibly altered between four and three lobe structures. 3

4 Supplementary Figure 4 STM contrasts of individual isomers of the molecule. This figure shows molecular models and expected STM contrast shapes of three situations; correctly landed meso-isomer, wrongly landed meso-isomer and (R,R-R,R)-isomer. In each case all four motor wheels adopt the so-called stable conformation. Side view and top view give an impression of the 3D geometry of the isomers. The areas highlighted by the pink ovals show anticipated STM signals due to protruding aromatic parts of the molecules (note that also other factors play a role in the experimentally observed images, see discussion above). The bottom panel of the figure shows STM images found for the related isomers. The assignment of the STM contrasts to particular isomer must be taken with caution as various conformers of each isomer can give similar contrasts. 4

5 Section 2. Analysis of the linearity of the movement In order to evaluate linearity vs. randomness of the translational movement of the molecule upon excitation via inelastic tunnelling a statistical evaluation of the trajectories travelled by the molecules is presented below. Since the orientation of the molecule solely based on STM appearance is far from trivial as the molecule exhibits complex behaviour (Supp. Fig. 3), we used the following strategy: Always three neighbouring positions of the molecule (positions 1-2-3, 2-3-4, 3-4-5, etc.) from the trajectories (Supp. Fig. 5a) are connected by two lines. We arbitrarily set a rule that if an angle between the connecting lines is between 180 and 120 degrees (Supp. Fig. 5b) than the incremental movement falls into a category of preferentially linear forward movement. If the angle is between 120 and 0 degrees (Supp. Fig. 5b) the movement falls into a category of random movement. Since the random translational movement (due to the imperfection of the spinning) of the (R,R-R,R)- and (S,S-S,S)-enantiomers has the same probability to happen in all directions, an anticipated result is that in 33% of the cases the movement will be preferentially linear and in 66% of the case the movement will be random. In an ideal case the movement of the meso-isomer is expected to be 100% linear. However, a deviation from linearity can be anticipated, because simultaneous activation of all four motors is not expected. Supplementary Figure 5c shows trajectories travelled by the meso-isomer (left) and the (R,R-R,R)- or (S,S-S,S) enantiomer (right). Following the above mentioned analysis the trajectory travelled by the enantiomer shows in 30% of the cases linear steps which is close to 33% as expected for random movement (Supp. Fig. 5d). In contrast 90% of the steps travelled by the meso-form (Supp. Fig. 5d) fall into the category of preferentially linear movement. From all the collected data on the movement of the meso-isomer with the active adsorption geometry we found that 77% of the incremental movements were preferentially linear. This deviates significantly from the expected 33% of random movement. It should be noted that any, either randomly or directionally, diffusing molecules (as oppose to our molecular system with processive motors) should never show a positive outcome in this analysis as the forward and reverse movement has the same probability to happen. 5

6 Supplementary Figure 5 Linearity treatment of the movement of the meso-isomer and the enantiomer. a-b, Graphical principle of the linearity treatment. A longer trajectory (a) is divided into incremental three point trajectories. If the three point trajectory falls into the o interval (b) the incremental movement is treated as preferentially linear. If the three point trajectory falls into the o interval (b) the incremental movement is treated as random. c, Trajectory of the meso-isomer and the enantiomer from Figures 2b and 4e of the main text. d, Graph showing the theoretical (blue) and experimental (green) percentage of the incremental movements of the enantiomer (random) and meso-isomer (preferentially linear) corresponding to their trajectories travelled (c). S 1 (1 N 1 ns ) (x 100 gives %) N Supplementary Equation 1 Calculation of the standard deviation from the probability for successful events. S: standard deviation, N: number of attempts, n S : number of successful attempts). 6

7 Section 3. Additional examples of control of movement of the molecule Supplementary Figure 6 Additional examples of movement of the molecule. a, Preferential directional movement of the meso-form of the molecule; sequence of eight subsequent STM images showing directional translational movement of the meso-molecule upon seven voltage pulses with +750 mv (I = pa). The molecule moves from the centre of the image to the left. The frame size corresponds to 10.2 nm x 9.3 nm and the scanning conditions were I = 74 pa, U = 47 mv. b, Behaviour of the meso-isomer adsorbed with a geometry that does not allow for directional translational movement. The graphical sketch shows two geometries with all four methyl groups pointing upwards or downwards. Sequence of subsequent eleven STM images of the meso-isomer upon ten voltage pulses (U = mv, I = pa). The frame size is 14 nm x 9 nm and the scanning conditions are I = 74 pa, U = 47 mv. The sequence shows that the contrast is changed while the molecule exhibits only restricted movement. c, Selection of individual position from the trajectory shown in Figure 4e showing a random motion of a single enantiomer of the racemic version of (R,R-R,R) and (S,S-S,S) isomers. STM images (scanning parameters: 7.0 nm x 7.8 nm, I = 47 pa, U = 47 mv) showing 17 individual steps of the 7

8 trajectory travelled by the (R,R-R,R) or (S,S-S,S)-isomer of the molecule upon 17 excitations via inelastic tunnelling (manipulation parameters: I = pa, U = mv). The corresponding movie of the whole trajectory is provided as Supplementary Movie 3. The observed STM contrasts of the molecule include structures where one lobe is darker and images with three-lobe or even two-lobe structures. Irreversible changes in the structure of the molecule, however, are excluded, because the original four-lobe contrast always reappears after further manipulation. Electron excitation with different polarisations Supplementary Figure 7 Extended series of polarity dependent voltage pulsing from Figure 4d of the main text. a-h, Negative voltage excitation induces contrast changes in the molecule ((R,R-R,R)- or (S,S-S,S)-enantiomer), but no movement (scan area = 7.0 nm x 7.8 nm, I = 43 pa, U = 47 mv). i-l, The positive voltage excitation leads to the translational movement and contrast changes (helix inversion). Polarity dependent excitation manipulations were tested on several molecules and always resulted in identical observations. 8

9 Section 4. Stereoisomers and adsorption geometries of the four-wheeled molecule Supplementary Figure 8 Absorption geometries and expected resulting propulsion direction of the molecule. This figure shows cartoons of possible absorption geometries of the isomers of the molecule. Black wedges in the molecular cartoons represent the methyl groups pointing upwards and the dashed wedges represent methyl groups pointing downwards. a, Three possible landing geometries of the meso-isomer on a surface. Concerted action of all four motors for the meso-isomer that is correctly landed is expected to render linear directional propulsion while for the two remaining geometries the effect of the propulsion of the individual motors should be cancelled out and no movement should be observed. b, The sum of propulsion directions of the individual (R,R-R,R)- and (S,S-S,S)-enantiomers should lead to spinning movement while in the of (R,R-S,S)-isomer no movement should be observed. 9

10 Section 5. Synthesis and characterisations of both meso and racemic isomers of four-wheeled molecule General Remarks Chemicals were purchased from Aldrich or Acros; solvents were reagent grade, distilled and dried before use according to standard procedures. Column chromatography was performed on silica gel (Aldrich 60, mesh). 13 C NMR spectra were recorded on a Varian VXR-400 (101 MHz), or a Varian AMX500 (126 MHz). 1 H NMR spectra were recorded on a Varian VXR- 400 (400 MHz) or a Varian AMX500 (500 MHz). Chemical shifts are denoted in -unit (ppm) relative to CDCl 3 ( 1 H = 7.26, 13 C = 77.0). For 1 H NMR spectroscopy, the splitting parameters are designated as follows: s (singlet), d (doublet), t (triplet), m (multiplet), dd (double doublet). HRMS (EI) spectra were obtained with a Thermo Scientific LTQ Orbitrap XL spectrometer. MALDI-TOF spectra were obtained with Voyager DE-Pro instrument. HPLC analyses were performed on a Waters HPLC system equipped with a 600E solvent delivery system and a 996 Photodiode Array Detector with a Chiralcel AD (Diacel) column. 10

11 Synthetic scheme

12 Synthetic scheme

13 Experimental section for the synthesis of the four-wheeled molecule 1 2,7-di(hex-1-ynyl)-9H-carbazole (3): 2,7-Dibromocarbazole (2) (0.5 g, 1.54 mmol) was dissolved in a mixture of a dry Et 3 N/THF (5:5 ml) and the solution was degassed via three freeze-pump-thaw cycles. PdCl 2 (PPh 3 ) 2 (10 mol%) and CuI (10 mol%) were added and the reaction mixture was stirred at room temperature for 5 min. 1-Hexyne (8 equiv, 1.38 ml, mmol) was slowly added (with syringe pump, addition in 0.5 h) to the reaction mixture and the mixture was stirred at 60 C for 16 h. The mixture was filtered over celite and the celite cake was washed with DCM and EtOAc. The solvents were removed in vacuum and the solid residue was purified by a column chromatography to yield 3 as a brown solid (0.48 g, 1.47 mmol, 95%); TLC (Pentane:EtOAc, 95:5 v/v): RF = 0.46; 1 H-NMR (400 MHz, CDCl 3 ): δ 8.00 (s, 1H), 7.91 (d, J = 8.1 Hz, 2H), 7.38 (s, 2H), 7.28 (dd, J = 8.1, 1.3 Hz, 2H), 2.47 (t, J = 7.0 Hz, 4H), (m, 4H), (m, 4H), 0.99 (t, J = 7.3 Hz, 6H); 13 C-NMR (101 MHz, CDCl 3 ): δ , , , , , , 90.18, 81.44, 30.90, 22.05, 19.20, HRMS (m/z): [M+1] + calcd for C 24 H 26 N, ; found, ,7-dihexyl-9H-carbazole (4): Catalytic hydrogenation of compound 3 (0.48 g, 1.47 mmol) with H 2 (gas) in the presence of 10% Pd-C (0.10 g) in EtOAc (25 ml) was finished in 5 h at rt. The crude reaction mixture was filtered over celite and the celite cake was washed with DCM and EtOAc. The solvents were removed in vacuum and the solid residue was purified by column chromatography to yield 4 as a yellow viscous oil (0.48 g, 1.44 mmol, 98%); TLC (Pentane:EtOAc, 95:5, v/v): RF = 0.54: 1 H-NMR (400 MHz, CDCl 3 ): δ 7.90 (d, J = 8.1 Hz, 2H), 7.83 (s, 1H), 7.20 (s, 2H), 7.04 (d, J = 8.1 Hz, 2H), 2.76 (t, J = 8.1 Hz, 4H), (m, 4H), (m, 12H), 0.89 (t, J = 5.5 Hz, 6H); 13 C-NMR (101 MHz, CDCl 3 ): δ , , 13

14 121.34, , , , 36.52, 31.95, 31.79, 29.06, 22.64, HRMS (m/z): [M+1] + calcd for C 24 H 34 N, ; found, ,8-dihexyl-2,10-dimethyl-10,11-dihydro-1H-dicyclopenta[c,g]carbazole-3,9(2H,6H)-dione (5) Compound 4 (0.70 g, 2.09 mmol) was added slowly to a suspension of AlCl 3 (1.11 g, 8.36 mmol) and methacryloyl chloride (0.30 ml, 4.18 mmol) in DCM (50 ml) at 78 C. The mixture was subsequently allowed to stir at rt overnight. The mixture was carefully hydrolyzed on ice, and the organic layer separated. The aqueous phase was extracted with DCM (3 30 ml), and the combined organic fractions were washed with (aq.) NaHCO 3 and dried over Na 2 SO 4. The solvents were removed in vacuum and the solid residue purified by column chromatography to provide 5 (0.23 g, 0.48 mmol, 23%). 1 H-NMR (400 MHz, CDCl 3 ): δ 8.66 (s, 1H), 7.20 (s, 2H), 4.03 (dd, J = 16.9, 8.1 Hz, 2H), (m, 4H), (m, 2H), (m, 2H), (m, 4H), (m, 4H), 1.41 (d, J = 7.3 Hz, 6H), (m, 8H), 0.88 (t, J = 6.8 Hz, 6H); 13 C-NMR (101 MHz, CDCl 3 ): δ , , , , , , , 42.02, 36.73, 32.35, 31.78, 31.08, 29.38, 22.70, 17.01, HRMS (m/z): [M+1] + calcd for C 32 H 42 NO 2, ; found, tert-butyl-4,8-dihexyl-2,10-dimethyl-3,9-dioxo-2,3,10,11-tetrahydro-1hdicyclopenta[c,g]carbazole-6(9h)-carboxylate (6): The mixture of compound 5 (40 mg, 0.09 mmol), Boc 2 O (27.8 mg, 0.13 mmol) and 4-DMAP (11 mg, 0.09 mmol) was dried using a 14

15 vacuum pump; the atmosphere was changed to N 2 and the mixture dissolved in dry 1,4-dioxane (5 ml). The mixture was stirred for 5 h at rt, quenched with water, followed by extraction with EtOAc. The organic solution was dried with Na 2 SO 4 the solvent was removed in vacuum and the residue was then purified by column chromatography; TLC (Pentane:EtOAc, 92:8 v/v): RF = Compound 6 was obtained as yellow oil (36 mg, 0.06 mmol, 69%). 1 H-NMR (400 MHz, CDCl 3 ): δ (s, 2H), 4.07 (dd, J = 16.8, 8.0 Hz, 2H), 3.31 (dd, J = 16.9, 3.8 Hz, 2H), (m, 4H), (m, 2H), 1.80 (s, 9H), (m, 4H), (m, 4H), 1.41 (d, J = 7.4 Hz, 6H), (m, 8H), 0.89 (t, J = 7.0 Hz, 6H); 13 C-NMR (101 MHz, CDCl 3 ): δ , , , , , , , , 85.59, 42.05, 37.36, 32.54, 31.75, 31.11, 29.37, 28.31, 22.68, 16.67, HRMS (m/z): [M+1] + calcd for C 37 H 50 NO 4, ; found, ,9-di(9H-fluoren-9-ylidene)-4,8-dihexyl-2,10-dimethyl-2,3,6,9,10,11-hexahydro-1Hdicyclopenta[c,g]carbazole (11): Lawesson s reagent (0.70 g, 1.73 mmol) was added to a stirred solution of 6 (0.25 g, 0.44 mmol) in toluene (10 ml). After heating at 100ºC for 3 h, the red reaction mixture was cooled to room temperature followed by purification through short column chromatography (Pentane:EtOAc, 9:1 v/v) to give as a first fraction a red colored solution of thioketone 7 which was used, without further purification, after concentrated in vacuo. To a solution of thioketone 7 in toluene (8 ml), 9-diazo-9H-fluorene 8 (0.34 g, 1.77 mmol) was added and the mixture was heated at 80 C for 16 h. To a crude solution of the resulting episulfide 9, PPh 3 (0.35 g, 1.32 mmol) was added and the mixture was heated for an additional 5 h at 80 ºC. The reaction mixture was concentrated in vacuo. The crude mixture was purified by column chromatography; TLC (Pentane:EtOAc, 99:1 v/v): RF = 0.68, (racemic mixture of enantiomers [(R,R)- and (S,S)-]) and RF = 0.66 (meso-(r,s)-isomer) yielding a mixture of diastereomers 10 together as an orange oil. Deprotection of the Boc-group was achieved by dissolving compound 15

16 10 (0.21 g, 0.24 mmol, 55% from compound 6) in dry THF (10 ml) and TBAF (1 M in THF, 5 equiv, 1.2 ml) was added at rt. The reaction mixture was stirred at 60 C for 6 h, the solvent was evaporated and a mixture of diastereomers 11 was obtained as an orange oil (180 mg, 0.23 mmol, 96%). The residue was purified by column chromatography and meso and racemic compounds 11 were separated at this point. Meso-(R,S)-11: TLC (Pentane:EtOAc, 95:5 v/v): RF = 0.37 (45 mg); 1 H-NMR (400 MHz, CDCl 3 ): δ 8.33 (s, 1H), 7.95 (d, J = 7.5 Hz, 2H), 7.82 (dd, J = 6.1, 2.3 Hz, 2H), 7.74 (d, J = 7.2 Hz, 2H), 7.49 (d, J = 7.7 Hz, 2H), (m, 4H), 7.25 (t, J = 8.7 Hz, 2H), 7.14 (s, 2H), 7.04 (t, J = 7.7 Hz, 2H), (m, 2H), 3.79 (dd, J = 15.2, 5.6 Hz, 2H), 3.44 (d, J = 15.0 Hz, 2H), (m, 2H), (m, 2H), 1.61 (s, 4H), 1.51 (d, J = 6.7 Hz, 6H), (m, 12H), 0.69 (t, J = 7.0 Hz, 6H); 13 C-NMR (126 MHz, CDCl 3 ): δ , , , , , , , , , , , , , , , , , , , , 44.71, 42.95, 34.78, 32.36, 31.52, 28.52, 22.34, 19.39, Racemic mixture of (R,R)- and (S,S)-11: TLC (Pentane:EtOAc, 95:5 v/v): RF = 0.42 (135 mg); 1 H-NMR (400 MHz, CDCl 3 ): δ 8.30 (s, 1H), 7.93 (d, J = 6.4 Hz, 2H), 7.85 (dd, J = 5.1, 2.3 Hz, 2H), 7.78 (d, J = 7.6 Hz, 2H), 7.60 (d, J = 7.9 Hz, 2H), 7.38 (dd, J = 6.2, 2.4 Hz, 4H), (m, 4H), 7.08 (t, J = 7.6 Hz, 2H), (m, 2H), 3.84 (dd, J = 14.7, 5.6 Hz, 2H), 3.19 (d, J = 14.9 Hz, 2H), (m, 2H), (m, 2H), (m, 4H), 1.34 (d, J = 6.7 Hz, 6H), (m, 12H), 0.67 (t, J = 7.0 Hz, 6H); 13 C-NMR (101 MHz, CDCl 3 ): δ , , , , , , , , , , , , , , , , , , , , 44.67, 42.84, 34.59, 32.24, 31.62, 28.49, 22.41, 19.40, HRMS (m/z): [M+1] + calcd for C 58 H 58 N, ; found, Starting from this step all the further steps were performed separately for racemic and meso compounds using identical reaction procedures. 16

17 3,9-di(9H-fluoren-9-ylidene)-4,8-dihexyl-6-(4-iodophenyl)-2,10-dimethyl-2,3,6,9,10,11- hexahydro-1h-dicyclopenta[c,g]carbazole (14): To a mixture of compound 11 (24.7 mg, mmol), CuI (0.61 mg, mmol), K 2 CO 3 (6.6 mg, mmol) and 1-bromo-4-iodobenzene (12) (10.9 mg, mmol) in DMF (8 ml) under N 2 atmosphere. A solution of 1,3-di(pyridin-2-yl)propane-1,3-dione (0.7 mg, mmol) in DMF (2 ml) was added via syringe and the reaction mixture was stirred at 120 C for 16 h. After cooling, the mixture was extracted with EtOAc, washed with water (3 10 ml) and dried over Na 2 SO 4. The organic solvent was evaporated to afford the crude product. For a subsequent Finkelstein reaction, crude compound 13 (23 mg, mmol), CuI (10 mol%) and dried NaI (6 equiv) were dissolved in dry 1,4-dioxane (5 ml) and then N,N-dimethylethane-1,2- diamine (2 equiv.) were added via a syringe. The reaction mixture was stirred at 120 C for 1 h and further heated to 130 C for 16 h. The mixture was filtrated over celite and the celite cake was washed with DCM. The solvents were removed in vacuum and the solid residue was purified by column chromatography to give 14 as an orange oil. Meso-(R,S)-14: TLC (Pentane:EtOAc, 99:1 v/v): RF = 0.74 (7 mg, 58%); 1 H-NMR (400 MHz, CDCl 3 ): δ 8.07 (d, J = 8.6 Hz, 2H), 7.94 (d, J = 6.8 Hz, 2H), 7.82 (d, J = 7.0 Hz, 2H), 7.74 (d, J = 7.4 Hz, 2H), (m, 4H), (m, 4H), 7.25 (t, J = 8.7 Hz, 2H), 7.12 (s, 2H), 7.03 (t, J = 7.1 Hz, 2H), (m, 2H), 3.79 (dd, J = 14.5, 5.3 Hz, 2H), 3.44 (d, J = 15.2 Hz, 2H), (m, 2H), (m, 2H), (m, 4H), 1.51 (d, J = 6.7 Hz, 6H), (m, 12H), 0.68 (t, J = 7.1 Hz, 6H); 13 C-NMR (101 MHz, CDCl 3 ): δ , , , , , , , , , , , , , , , , , , , , , , , 91.11, 44.11, 43.12, 34.88, 32.63, 31.59, 29.70, 22.44, 19.84, Meso-(R,S)-14 was analysed by chiral HPLC using a 17

18 Chiralcel AD column (Heptane:i-propanol, 95:5 v/v): A single fraction was obtained with retention time = 4.36 min. Racemic mixture of (R,R)- and (S,S)-14: TLC (Pentane:EtOAc, 99:1 v/v): RF = 0.79 (16.2 mg, 67%); 1 H-NMR (400 MHz, CDCl 3 ): δ 8.07 (d, J = 8.6 Hz, 2H), 7.96 (d, J = 7.9 Hz, 2H), (m, 2H), 7.78 (d, J = 7.6 Hz, 2H), 7.62 (d, J = 7.9 Hz, 2H), 7.51 (d, J = 8.6 Hz, 2H), (m, 4H), 7.30 (t, J = 7.4 Hz, 2H), 7.15 (s, 2H), 7.10 (t, J = 6.1 Hz, 2H), (m, 2H), 3.89 (dd, J = 15.1, 5.7 Hz, 2H), 3.23 (d, J = 15.0 Hz, 2H), (m, 2H), (m, 2H), 1.36 (d, J = 6.4 Hz, 6H), (m, 4H), (m, 12H), 0.65 (t, J = 7.1 Hz, 6H); 13 C- NMR (101 MHz, CDCl 3 ): δ , , , , , , , , , , , , , , , , , , , , , , , 92.91, 44.71, 42.97, 34.79, 32.36, 31.52, 28.53, 22.39, 19.40, The racemic mixture of (R,R)-14 and (S,S)-14 was analysed by chiral HPLC using a Chiralcel AD column (Heptane:i-propanol, 95:5 v/v): the first eluted fraction at t = 3.07 min and the second fraction at t = 3.87 min, respectively. HRMS (m/z): [M+1] + calcd for C 64 H 61 IN, ; found, (4-ethynylphenyl)-3,9-di(9H-fluoren-9-ylidene)-4,8-dihexyl-2,10-dimethyl-2,3,6,9,10,11- hexahydro-1h-dicyclopenta[c,g]carbazole (16): Compound 14 (35.1 mg, mmol) was dissolved in dry Et 3 N/THF (10mL,1:1 ratio) and the solution was degassed via three freeze-pump-thaw cycles. PdCl 2 (PPh 3 ) 2 (5 mol%, 1.3 mg, mmol), CuI (5 mol%, 0.4 mg, mmol) and trimethylsilylacetylene (0.05 ml, mmol) 18

19 were added and the reaction mixture was stirred at 50 C for 16 h. The reaction crude was filtered over celite and the celite cake washed with DCM and EtOAc. The solvents were removed in vacuum and the crude compound 15 was dissolved in THF (3 ml). TBAF (1 M in THF, ml, mmol) was added at rt and the reaction mixture was stirred for 30 min. The solvent was evaporated and the residue was purified by flash chromatography over a short path of silica gel using DCM as solvent to yield 16 as orange viscous oil. Meso-(R,S)-16: TLC (Pentane:EtOAc, 99:1 v/v): RF = 0.74 (5 mg, 71%); 1 H-NMR (400 MHz, CDCl 3 ): δ 7.95 (d, J = 7.2 Hz, 2H), 7.87 (d, J = 8.5 Hz, 2H), 7.82 (d, J = 8.6 Hz, 2H), 7.74 (d, J = 7.3 Hz, 2H), 7.69 (d, J = 8.3 Hz, 2H), 7.49 (d, J = 7.9 Hz, 2H), (m, 4H), 7.25 (t, J = 8.2 Hz, 2H), 7.15 (s, 2H), 7.03 (t, J = 7.0 Hz, 2H), (m, 2H), 3.79 (dd, J = 15.2, 5.7 Hz, 2H), 3.45 (d, J = 15.0 Hz, 2H), (m, 2H), (m, 2H), (m, 4H), 1.52 (d, J = 6.7 Hz, 6H), (m, 12H), 0.68 (t, J = 7.0 Hz, 6H). Racemic mixture of (R,R)- and (S,S)-16: TLC (Pentane:EtOAc, 99:1 v/v): RF = 0.75 (22.7 mg, 73%); 1 H-NMR (400 MHz, CDCl 3 ): δ 7.93 p.p.m. (dd, J = 5.7, 3.1 Hz, 2H), (m, 4H), 7.78 (d, J = 7.5 Hz, 2H), 7.72 (d, J = 8.5 Hz, 2H), 7.61 (d, J = 7.9 Hz, 2H), (m, 4H), 7.29 (t, J = 7.5 Hz, 2H), 7.17 (s, 2H), 7.10 (t, J = 7.6 Hz, 2H), (m, 2H), 3.89 (dd, J = 15.0, 5.9 Hz, 2H), 3.25 (s, 1H), 3.22 (d, J = 15.0 Hz, 2H), (m, 2H), (m, 2H), 1.35 (d, J = 6.7 Hz, 6H), (m, 4H), (m, 12H), 0.64 (t, J = 7.1 Hz, 6H). HRMS (m/z): [M+1] + calcd for C 66 H 62 N, ; found, ,4-bis(4-(3,9-di(9H-fluoren-9-ylidene)-4,8-dihexyl-2,10-dimethyl-2,3,10,11-tetrahydro-1Hdicyclopenta[c,g]carbazol-6(9H)-yl)phenyl)buta-1,3-diyne (1): To a solution of PdCl 2 (PPh 3 )

20 (10 mol%, 2.0 mg, mmol), CuI (10 mol%, 0.5 mg, mmol) in dry Et 3 N/THF (5:2 ml) was degassed via three freeze-pump-thaw cycles. Compound 16 (24.5 mg, mmol) in THF (3 ml) was slowly added via syringe mixture was stirred at 60 C for 16 h. After cooling down to room temperature, the solution was filtrated over celite and the celite cake was washed with DCM and EtOAc. The solvents were removed in vacuum and the solid residue purified by column chromatography to yield a 1 as an orange solid. Meso-(R,S-R,S)-1: TLC (Pentane:DCM, 9:1 to 6:4 v/v): RF = 0.41 (2.8 mg, mmol, 80 %); 1 H NMR (400 MHz, CDCl 3 ) δ 7.96 (d, J = 8.4 Hz, 8H, ArH), 7.83 (d, J = 6.6 Hz, 4H, ArH), 7.75 (dd, J = 7.7, 3.6 Hz, 8H, ArH), 7.50 (d, J = 7.8 Hz, 4H, ArH), (m, 8H, ArH), 7.27 (d, J = 4.8 Hz, 4H, ArH), 7.20 (s, 4H, ArH), 7.05 (t, J = 7.1 Hz, 4H, ArH), (m, 4H, CH), 3.81 (dd, J = 15.3, 6.1 Hz, 4H, CH 2 ), 3.46 (d, J = 15.3 Hz, 4H, CH 2 ), (m, 4H, CH 2 ), (m, 4H, CH 2 ), 1.53 (d, J = 7.8 Hz, 12H, CH 3 ), (m, 8H, CH 2 ), (m, 10H, CH 2 ), (m, 14H, CH 2 ), 0.70 (t, J = 7.1 Hz, 12H, CH 3 ); 13 C NMR (126 MHz, CDCl 3 ) δ , , , , , , , , , , , , , , , , , , , , , , , , , , , , , 81.44, 75.04, 45.43, 44.72, 42.98, 41.66, 38.37, 34.79, 32.36, 32.12, 31.53, 29.71, 29.48, 29.37, 28.54, 22.40, 22.37, 21.39, 19.40, 13.94, (lower number of aromatic carbons is due to overlap). The meso-isomer was analysed by chiral HPLC using Chiralcel AD column with a mixture of n-heptane: i-propanol (95: 5 v/v) as the eluent. Only a single fraction was found with retention time t = 8.49 min (Supp. Fig. 9). MALDI-TOF MS (trans-2-[3-(4-tert-butylphenyl)-2-methyl-2-propenylidene]-malononitrile as the matrix): m/z calcd for C 132 H 120 N 2 : , found Racemic mixture of (R,R-R,R), (R,R-S,S) and (S,S-S,S) 1: TLC (Pentane: DCM, 9:1 to 7:3 v/v): RF = 0.23 (20.8 mg, mmol, 85 %); mp C; 1 H NMR (400 MHz, CDCl 3 ) δ 7.94 (t, J = 7.3 Hz, 8H, ArH), (m, 4H, ArH), 7.79 (d, J = 8.2 Hz, 8H, ArH), 7.63 (d, J = 7.9 Hz, 4H, ArH), (m, 8H, ArH), 7.30 (t, J = 7.5 Hz, 4H, ArH), 7.22 (s, 4H, ArH), 7.11 (t, J = 7.6 Hz, 4H, ArH), (m, 4H, CH), 3.90 (dd, J = 15.3, 5.7 Hz, 4H, CH 2 ), 3.23 (d, J = 14.8 Hz, 4H, CH 2 ), (m, 4H, CH 2 ), (m, 4H, CH 2 ), 1.35 (d, J = 6.7 Hz, 12H, CH 3 ), (m, 16H, CH 2 ), (m, 16H, CH 2 ), 0.66 (t, J = 7.0 Hz, 12H, CH 3 ); 13 C NMR (101 MHz, CDCl 3 ) δ , , , , , , , , , 20

21 138.11, , , , , , , , , , , , , , , , , , , , 81.47, 75.09, 44.74, 43.01, 34.81, 33.82, 32.33, 31.93, 31.61, 31.53, 29.70, 29.67, 29.16, 28.96, 28.55, 22.69, 22.39, 19.40, 14.12, (lower number of aromatic carbons is due to overlap). The racemic mixture was analysed by chiral HPLC using a Chiralcel AD column with a mixture of n-heptane: i-propanol (95: 5) as the eluent. As expected the mixture was composed of 3 fractions; the first eluted fraction at t = 3.18 min, the second fraction at t = 3.54 min and the third fraction at t = 4.37 min, respectively (Supp. Fig. 10). MALDI-TOF MS (trans-2-[3-(4-tert-butylphenyl)-2-methyl-2-propenylidene]malononitrile as the matrix): m/z calcd for C 132 H 120 N 2 : , found

22 Chiral HPLC data for four-wheeled molecule 1 UV/Vis spectra Supplementary Figure 9 Chiral HPLC of the meso--(r,s-r,s)-isomer of the four-wheeled molecule 1 shows the retention time of the meso--(r,s-r,s)-isomer. 22

23 UV/Vis spectra Supplementary Figure 10 Chiral HPLC of the mixture of (R,R-R,R), (R,R-S,S) and (S,S-S,S) isomers of the four-wheeled molecule 1 shows the presence and separation of all three isomers. 23

24 Legends for Supplementary Movies Supplementary Movie 1 Directional movement of the meso-isomer of the molecule. The movie shows directional translational movement of the meso-isomer of the molecule on Cu(111) after excitation by tunnelling electrons emanating from the STM tip. The movie is composed of 12 frames with size (10.2 nm x 9.3 nm) based on STM images (I = 74 pa, U = 47 mv). To track the movement of the molecule, the scanned area was moved upwards between frame 8 and 9. The frames of the movie are thus composed of the scanned area of the surface and the black background representing a surface area that was scanned at the later stage. Supplementary Movie 2 Helix inversion of the wheels of the molecule at low bias voltages. The movie shows modification of STM contrasts of the molecule induced by helix inversion in the motor wheels. The helix inversions of the motor wheels in the movie are induced by electron excitations at low voltages (voltage range 200 to 350 mv) The frames of the movie constitute of the STM images with dimensions 7.0 nm x 7.8 nm recorded at tunnelling current I = 43 pa and voltage U = 47 mv). Supplementary Movie 3 Random movement of the molecule after electron excitation. The movie shows random movement of the molecule (from a mixture of (R,R-R,R) and (S,S-S,S)- isomers) on Cu(111) after excitation by tunnelling electrons emanating from the STM tip. The movie is composed of 60 subsequent STM images (7.0 nm x 7.8 nm, I = 47 pa, U = 47 mv). After each image the STM tip is brought on top of the molecules and an excitation voltage pulse is applied resulting in movement of the molecule. Supplementary Movie 4 Anticipated linear propelled movement of a molecular model of the meso-form. The movie shows propelled movement of the meso-isomer of the four-wheeled molecule on a surface via consecutive double bond isomerisation and helix inversion processes of the motor units in concerted unidirectional fashion. For clarity, the hexyl groups are substituted by methyl groups. 24