Samenvatting

This report outlines the origin of aberrations in a Low Energy Electron Microscope (LEEM) and how to measure and automatically correct them. The aberrations in a LEEM are mainly caused by the cathode objective lens and are a limiting factor for the achievable resolution. To optimise the resolution of the microscope it is therefore necessary to cancel out the aberrations by using an electron mirror, which can introduce aberrations of the same magnitude but with opposite sign of the cathode objective lens. Aberrations can be measured by using Low Energy Electron Diffraction (LEED) patterns. In a LEED pattern spots are displaced from the centre spot as a function of the angle through the system. The magnitude of this displacement depends on defocus and spherical aberrations. By fitting a function of the form C1α + C3α3, where C1 stands for defocus, and C3 stand for third order spherical aberrations, and α for the angle, defocus en third order spherical aberrations can be extracted.

To automatically correct aberrations present in the microscope, LEED patterns are analysed with a program written with the programming language Python. Due to limiting time the present form of this program only contains a noise filter, which uses Wiener deconvolution in the frequency domain, and spot detection where a 3x3 pattern is shifted over the filtered pattern. To ensure that the spot detection works properly ten theoretical LEED patterns are generated with pre-set values for defocus en third order spherical aberration. These patterns are analysed by the written code to find the spot locations which are then exported to OriginProtm to make a fit. From these fits the pre-set values for defocus en third order spherical aberrations are recovered. The results of the fits gave an average uncertainty of 0,30% and 1,31% for respectively defocus en third order spherical aberrations. This shows that the spot detection works properly. In the final part of the report there will be a description about how to implement order recognition using Delaunay-triangulation and Hasse-diagrams in the future.