AVS 45th International Symposium
    Surface Science Division Monday Sessions
       Session SS1+NS-MoA

Paper SS1+NS-MoA5
Multispectral Image Classifications of Si(001) Surface Electronic Structure

Monday, November 2, 1998, 3:20 pm, Room 308

Session: Novel Surface Probes
Presenter: K.M. Horn, Sandia National Laboratories
Authors: K.M. Horn, Sandia National Laboratories
B.S. Swartzentruber, Sandia National Laboratories
G.C. Osbourn, Sandia National Laboratories
Correspondent: Click to Email
We have imaged the electronic structure of Si(001) surfaces by applying multispectral image analysis techniques to multi-bias STM conductance data. Atomic surfaces are first characterized by recording conductance spectra, C(V), at each point in a 2D scan of the surface. The resulting 3D data set, (x, y, C(V) ), is then converted into a series of bias-dependent conductance images. These images are analyzed to produce a single, color-coded, classed image that reflects the surface’s electronic structure. The image analysis is performed by a computed grouping algorithm that identifies pixels sharing common conductance characteristics. The resulting classed images distinguish features not clearly resolved in a topographic image, and reveal stark electronic differences between topographically similar features. We first demonstrate the reliability of this classification technique on simple Si(001) features. Classed surfaces are then presented for various surface defects and Si and Ge structures that have been deposited on the Si(001) surface. These electronic structure images reveal features that are not readily visible or distinguished in a constant-current topograph. Direct comparison of the conductance spectra from these features confirms the classification result. This computer-based data reduction technique may prove useful in defect detection, validating surface models, and in understanding more complicated systems in which atomistic models are derived from a limited number of single-bias topographic images. Sandia is a multiprogram laboratory operated by Sandia Corporation, a Lockheed Martin Company, for the U.S.DOE under contract DE-AC04-94AL85000.