Mapping the Electronic Surface Potential of Nanostructured Surfaces
(Result of the month 07/2009)

We present a method for the quantitative determination of the surface potential landscape of nanostructured surfaces based on the local analysis of the lowest field emission resonances (FERs) by scanning tunneling spectroscopy [1]. The method has a lateral resolution of ~1 nm and is applied to elucidate the site-specific adsorption properties of the strain relief pattern formed by two monolayers of Ag on Pt(111). The determined local surface potential is shown to be firmly related to variations of the local lattice parameter of the Ag layer. For the example of C60 fullerenes, we show that the surface potential difference of up to 0.35 eV is responsible for the site-selective immobilization on the strain relief pattern.

FIG. 1. STM images of the 2 ML Ag/Pt(111) strain relief pattern. (a) Topography image revealing the three domains of the pattern (T = 77 K, Vs = -2 V, I = 2 nA). (b) Height profile from A to B, as shown in (a). (c) Topography image of C60 adsorbed on the pattern (Vs = -2 V, I = 0.1 nA). Scale bars: 5 nm.
The Ag/Pt(111) strain relief pattern is formed by adsorption of two monolayers of silver on Pt(111) followed by annealing to 800 K [2]. This leads to the formation of a hexagonal pattern with 7 nm periodicity with a unit cell comprising three surface regions, called fcc, hcp1 and hcp2. Adsorption of C60 fullerenes reveals a highly inhomogeneous immobilization with respect to the different surface areas. As shown here, this is related to a lower surface potential in the hcp1 region, which influences the binding energy of C60 due to the partially ionic bonding character of C60 on silver [1].

FIG. 2. (a) Series of 100 dz/dV spectra (Δx = 0.18 nm) recorded along the line indicated in (b) and represented as a color-coded image. The white bars represent the spatial expectation values of the first four FERs. (b) Topography image of a single C60 adsorbed at a defect in the fcc region (Vs = 2 V, I = 0.1 nA).

In order to determine the lateral resolution of the surface potential determination based on the analysis of FERs we investigate the spatial evolution of the energy position of the FERs over a sharp defect, consisting of an isolated C60 molecule. As expected, the resolution is highest for the lowest FERs due to their residence very close to the surface. For comparison we display the expectation values for the lowest image potential states. The experimental resolution is systematically better than the calculated one. This can be understood from the increased constriction of the FERs to the surface due to the applied sample voltage. Restricting the analysis to the two lowest FERs reveals a lateral resolution of ~1 nm for the surface potential determination.

FIG. 3. (a) Experimental zv(Vs) curve (black line) and dzv/dV(Vs) curve (red line) recorded in the center of the hcp2 domain (feedback loop closed, Is=0.05 nA). (b) Relevant parameters of the model potential used for the simulation of the FERs. The unknown parameters are fitted to reproduce the energy position of the FERs in the hcp2 region, revealing Φt = 4.0 eV, zi = 0.17 nm, and z0 = 1.7 nm. The related energy levels of the modeled FERs are shown as blue vertical bars for n = 1, ... , 4 in (a). (c) Three-dimensional representation of the local surface potential as determined near the hcp1 region.

In order to quantitatively extract the surface potential from the FER analysis some unknown parameters need to be determined on either a surface region with known surface potential or by determination of the average surface potential by an alternative method, as is done here. In Fig. 3 the measured and simulated energy positions of the FERs in the hcp2 region for optimized parameters for tip work function, image plane position and z0 are displayed. These paramters are then kept constant and the local surface potential is determined by varying Δφ in the model potential in order to get the best agreement between measured and simulated energy positions of the FERs.