First Advisor

Peter Moeck

Term of Graduation

Spring 2026

Date of Publication

7-8-2026

Document Type

Thesis

Degree Name

Master of Science (M.S.) in Physics

Department

Physics

Language

English

Physical Description

1 online resource (xii, 240 pages)

Abstract

Though it is an impressive technology, Scanning Tunneling Microscopy (STM) produces images commonly plagued by small systematic errors, often called "drift". These effects have a variety of causes, most notably nonlinear behavior of piezoelectric actuator tubes responsible for the positioning of the scanning probe or sample. One result of these distortions is a potential incorrect classification of the symmetry groups and projected Laue classes of recorded images from crystals—made possible by using recently developed objective methods. An implementation of a known method for detecting and correcting these errors is presented. This method operates strictly in Fourier space so as to specifically correct global distortions and to limit the impact of significant but strictly localized symmetry breakings such as those caused by sudden probe slips or defects in a sample. The efficacy of the implementation through the correction of images whose distortion effects lead to incorrect classifications of well modeled and understood physical structures, and the more accurate objective classification of the images after correction is demonstrated. These methods are applied both to synthetic simulated images created in Photoshop that model highly oriented pyrolytic graphite (HOPG) as well as actual STM images of crystals of varying quality. In addition, the methods are applied to images captured via a variety of other imaging methods including aberration-corrected Transmission Electron Microscopy and Atomic Force Microscopy. While the causes of the systematic errors vary based on the imaging technique employed, the methods make improvements to either the objective symmetry classification or to composite metrics based on lattice parameters, regardless of imaging technique. Finally, a small comparison is executed between this method and another more modern and involved method based on feature indexing and localized corrections to demonstrate its potentially outsized capability relative to its complexity.

Persistent Identifier

https://archives.pdx.edu/ds/psu/45057

Included in

Physics Commons

Share

COinS