Force-controlled lifting of molecular wires

Lifting a single molecular wire off the surface with a combined frequency-modulated atomic force and tunneling microscope it is possible to monitor the evolution of both the wire configuration and the contacts simultaneously with the transport conductance experiment. In particular, critical points where individual bonds to the surface are broken and instabilities where the wire is prone to change its contact configuration can be identified in the force gradient and dissipation responses of the junction. This additional mechanical information can be used to unambiguously determine the conductance of a true molecular wire, that is, of a molecule that is contacted via a pointlike “crocodile clip” to each of the electrodes but is otherwise free.

Figure 1

Simulated junction geometries during tip retraction. (a) PTCDA adsorbed on Ag(111). The molecule is shown in the adsorption site that is used in the simulation. The tip trajectory employed in experiment and simulation is shown in red (its projection into the $xy$ plane is shown as a shadow). An animation of this tip trajectory is provided in the supplemental material (Ref. 14). Tip positions corresponding to the four junction geometries in panels (b) to (e) are marked in blue. The two angles $\phi$ and $\theta$ describe the orientation of the molecule in the junction. (b) The second carboxylic oxygen is detached from the surface. (c) The almost planar molecule is lifted into the upright configuration. (d) Upright molecule ($\phi ={90}^{\circ },\theta =0^{\circ }$) bound to Ag(111) via two carboxylic oxygen atoms. (e) Molecular wire configuration, that is, $\phi ={90}^{\circ },\theta ={15}^{\circ }$. The coordinate $z_{c.m.}$ describes the motion of the molecule after it is detached from the surface.

Figure 2

Stiffness of the tip/PTCDA/Ag(111) junction. (a) Two-dimensional histogram of 121 $\Delta f\left(z\right)$ curves recorded while lifting up single isolated PTCDA molecules. After each manipulation cycle the molecule remained at the tip. In 65$\%$ of the cases it was possible to redeposit the molecule back onto the surface by having the tip approach the surface and applying a voltage of +0.6 V to the sample. The inset shows $\Delta f\left(z\right)$ for a single approach/retraction cycle: Approach is shown in black, retraction in red. For all curves in the histogram, the $\Delta f$ signal during tip approach was subtracted from the $\Delta f$ data taken during retraction. (b) Same dataset as in panel (a), but each curve is shifted on the horizontal axis by individual values $z_0$ such that the peak belonging to feature B is aligned with the corresponding peak in the simulated $\Delta f$ curve (dashed line) (Ref. 24). The inset shows a histogram of shift distances $z_0$. (c) Averaged dissipation signal for all 121 retractions of panel (b). Before averaging, each curve was shifted by the same $z_0$ as the corresponding $\Delta f$ curve in panel (b). The dissipation is a measure of the energy needed to sustain a constant oscillation amplitude of the tuning fork sensor.

Figure 3

Force, potential energy, and conductance of the molecular wire. (a) Histogram of the 35 integrable $\frac{dF}{dz}$ curves (i.e., no abrupt jumps for $z>15$ Å) from the dataset in Fig. 2b. The solid red line displays the averaged experimental data. The black dashed line shows the simulated $\Delta f$ curve (Ref. 24). The tip heights that correspond to the junction geometries in Fig. 1 are indicated and labeled as b, c, d, and e. The upper horizontal axis ($\phi$, $\theta$, $z_{c.m.}$) indicates the coordinates of molecular motion, enforced by the tip trajectory shown in Fig. 1a. (b) Black dashed line, force on the tip as calculated in the force-field simulation; red solid lines, force on the tip as calculated by integrating the averaged experimental $\frac{dF}{dz}$ in panel (a) [red solid line in panel (a)] for $z>16$ Å and for $z<15$ Å. The branch for $z<15$ Å was shifted along the vertical axis such that it starts at the value where the left section of the curve ended. (c) Black dashed line, potential energy of PTCDA in the junction as calculated in the force-field simulation; red solid lines, potential energy of PTCDA in the junction as calculated by integrating the experimental $F\left(z\right)$ curves in panel (b) [red solid line in panel (b)]. The integration constant for $z<15$ Å was again chosen to match the value where the left section of the curve ended. (d) Histogram of the junction conductance $\frac{dI}{dV}\left(z\right)$ at a bias voltage of $V=-0.5$ mV for the same dataset as in panels (a)–(c). At $z\approx 17$ Å the last molecule-substrate bond breaks and the junction enters the tunneling regime. The inset is a line profile through the histogram parallel to the vertical axis in the corridor between $z=16.25$ Å and $z=17.0$ Å.