Strumming a Single Chemical Bond

Atomic force microscopy and scanning tunneling microscopy can image the internal structure of molecules adsorbed on surfaces. One reliable method is to terminate the tip with a nonreactive adsorbate, often a single CO molecule, and to collect data at a close distance where Pauli repulsion plays a strong role. Lateral force microscopy, in which the tip oscillates laterally, probes similar interactions but has the unique ability to pull the CO over a chemical bond, load it as a torsional spring, and release it as it snaps over with each oscillation cycle. This produces measurable energy dissipation. The dissipation has a characteristic decay length in the vertical direction of 4 pm, which is 13 times smaller than the decay length in typical STM or AFM experiments.

Figure 1

PTCDA on Cu(111). (a) Sketch of the LFM setup with a CO adsorbed to the end of a metal tip. The tip oscillates laterally in the $x$ direction with an amplitude $A$. (b) Current image of a PTCDA island and (c) the corresponding LFM image. Molecular schematics follow the color scheme: O red, C gray, H white, Cu copper.

Figure 2

Excitation in LFM data. (a) LFM image of a single PTCDA molecule and (b) corresponding excitation image. The scale bar for both (a) and (b) is shown in (b). The direction of oscillation is horizontal. (c) (Inset, right) Sketch of two atoms considered in this model. (Main) Moving the CO tip over a bond at $x=z=0\mathrm{m}$ causes the CO to tilt, but it cannot reach the universal low energy position. The background grayscale shows the potential energy (in meV) of the O atom as acted upon by the two C atoms, modeled with Morse potentials. As the tip passes from left to right ($i$ to ii), the CO cocks over the bond (iii) until it snaps down (iv). (d) The angle that the CO makes with the vertical as a function of the $x$ position of the metal tip apex. (e) The lateral force on the apex makes a hysteresis loop around the snap. The area of this loop is the dissipated energy, $E_{\mathrm{diss}}$.

Figure 3

Dissipation as a function of oscillation amplitude (a) The six atoms that were included in this model are shown darker than the others. (b) Experimentally measured dissipation linescans across a bond at various amplitudes can be reproduced by a snapping model shown in (c) and described in the text. (d) The frequency shift across the bonds can also be well represented by the model, shown in (e).

Figure 4

The maximum energy dissipation above a bond as a function of the height of the tip. Black dots are data, red circles are the model output. The blue line is an exponential fit to the data from $z=5$ to 15 pm, to characterize the decay length.