Local tunneling decay length and Kelvin probe force spectroscopy

In the past, current-distance spectroscopy has been widely applied to determine variations of the work function at surfaces. While for homogeneous sample areas this technique is commonly accepted to yield at least qualitative results, its applicability to atomic-scale variations has not been proven neither right nor wrong. Here we benchmark measurements of the current-distance decay constant against the well established Kelvin probe force spectroscopy for four distinctly different cases with atomic-scale variations of the local contact potential. The two techniques yield quite different results. Whereas the maps of the current-distance decay constant are consistent with being topographical artifacts, the Kelvin probe force spectroscopy maps show variations of the local contact potential difference in agreement with expected surface dipoles. This comparison clarifies that maps of the current-distance decay constant are not suited to directly characterize contact potential variations at surfaces on atomic length scales.

Figure 1

Illustration how the local topography can influence the apparent tunneling current decay length. If one assumes that profiles of constant current (dashed lines) follow the topography at different distances, beside protrusions and in trenches the line contours will become more sparse as compared to the flat terraces (see vertical dotted lines).

Figure 2

Exemplary curves of KPFS and $I\left(z\right)$ measurements. (a) KPFS parabola: raw data (blue) along with parabolic fit (red). The maximum position of the parabola gives the local VCPD value (green). (b) Logarithmic plot of $I\left(z\right)$ curves at different bias voltages: raw data (blue) along with exponential fits (red).

Figure 3

Individual neutral and negatively charged Au adatoms on bilayer NaCl/Cu(111) recorded with a metal tip. (a) Schematic side view of the adsorption geometry and charge distribution. (b) Current image of gold anion (left hand side) and neutral gold adatom (right hand side) $(z=12.5\AA ,V=0.1\mathrm{V})$. (c) VCPD map recorded at a height of $12.5\AA$. (d and e) $\kappa$ deduced from $I\left(z\right)$ curves for $12.5<z<15.5\AA$ at $V=-0.5$ and $V=0.5\mathrm{V}$ for panels (d) and (e), respectively. The color scale is the same for both maps. (Scale bars 5 Å.)

Figure 4

PTCDA on clean Cu(111) (top panels) and ClAnCN on bilayer NaCl/Cu(111) (bottom panels) investigated with CO functionalized tips. (a) Structure model of PTCDA. (b) Constant-height image of tunneling current extracted from grid data at a tip height of $9.6\AA$ and $0.1\mathrm{V}$ bias voltage. (c) VCPD map recorded at $12.5\AA$. (d and e) $\kappa$ maps extracted from $I\left(z\right)$ curves for $9.6 <z<12.6\AA$ at $V=-0.4$ and $V=0.75\mathrm{V}$ for panels (d) and (e), respectively. (f) Structure model of ClAnCN. (g) Constant-height image of tunneling current recorded at a tip height of $15.8\AA$ and $0.1\mathrm{V}$ bias voltage. (h) VCPD map recorded at $18.2\AA$. (i and k) $\kappa$ map extracted from $I\left(z\right)$ curves for $16.2 <z<18.2\AA$ at $V=-0.5$ and $V=0.5\mathrm{V}$ for panels (i) and (k), respectively. Panels (a), (c), (f), and (h) reprinted from [32]. (Scale bars 5 Å.)

Figure 5

${\mathrm{F}}_{12}{\mathrm{C}}_{18}{\mathrm{Hg}}_3$ and ${\mathrm{H}}_{12}{\mathrm{C}}_{18}{\mathrm{Hg}}_3$ on clean Cu(111) examined with a CO functionalized tip. (a) Structure models of ${\mathrm{F}}_{12}{\mathrm{C}}_{18}{\mathrm{Hg}}_3$ (left hand side) and ${\mathrm{H}}_{12}{\mathrm{C}}_{18}{\mathrm{Hg}}_3$ (right hand side) shown in the same orientation as the molecules under investigation. (b) Constant-height image of tunneling current of the two molecules $(z=9.6\AA ,V=0.1\mathrm{V})$. (c) VCPD map recorded at $12.0\AA$. (d) Side view onto constant-current line profiles across ${\mathrm{F}}_{12}{\mathrm{C}}_{18}{\mathrm{Hg}}_3$ and ${\mathrm{H}}_{12}{\mathrm{C}}_{18}{\mathrm{Hg}}_3$ molecules. The profiles were acquired along the line indicated in panel (e) at $V=-0.2\mathrm{V}$. When going from one constant-current profile to the adjacent one, the current changes by a factor of 2. The two dashed red lines highlight positions of trenches in the corresponding constant-height image. (e and f) $\kappa$ map extracted from $I\left(z\right)$ curves for $9.9<z<11.5\AA$ at $V=-0.2$ and $V=0.35\mathrm{V}$ for panels (e) and (f), respectively. Panels (a) and (c) reprinted from [32]. (Scale bars 5 Å.)