For the Autumn 2025 semester the LSME is offering the following student projects. For any other project interest, please contact [email protected] or [email protected].
Â
Atomic column detection, location and ferroelectric distortion mapping
Scanning transmission electron microscopy (STEM) can image materials with a resolution at the sub-Ă scale. This allows resolving individual atomic columns in crystals. The magnetic, ferroelectric, and electronic properties of modern functional materials are often governed by subtle changes in the positions of the atoms in the crystalline lattice. However, automatic detection and precise location of the atomic columns from a STEM image is not trivial. At the LSME group we are developing an open-source software (âUCsplitâ) to perform this task.
Your role in this project will be to become proficient with the usage of âUCsplitâ to map ferroelectric distortions in perovskite thin films, evaluate the performance of different algorithms included in the software, and, eventually, contribute to its development.
Contact: [email protected]
 Â
Exploring the behaviour of signal-to-noise ratio in image pre-processing and template pattern matching for SPED datasets
Scanning precession electron diffraction (SPED) is a powerful technique in transmission electron microscopy, that is used to reduce the effect of dynamical diffraction and enhance the quality of the acquired diffraction patterns, thereby allowing for a more accurate structural analysis. This project focuses on analysing the influence of signal-to-noise ratio (SNR) in the pre-processing and template pattern matching stages of SPED datasets analysis workflow. The student will leverage an in-house developed Python-based algorithm designed to address key challenges in conventional SPED data analysis workflows, including noise resilience, low spatial mapping resolution, and slow processing speeds. By examining how variations in SNR influence algorithm performance, the project aims to optimise data pre-processing strategies and improve template pattern matching reliability. Although the primary work will involve data analysis, the student will assist in microscopy experiments in order to correlate their theoretical knowledge with the experimental technique. This experience will provide insights into experimental parameters that impact SNR and the subsequent data processing pipeline.
Contact: [email protected]
 Â
EELS simulations of dielectric photonic nanocavities
In a recent study of optical nanocavities, made from dielectric silicon, using low-loss electron energy-loss spectroscopy (EELS), we identified that the EELS data give valuable insights into different optical eigenmodes with a spatial resolution of a few 10s of nanometers â so far below the diffraction limit. While previously we focused on the spectral peaks, it turns out that spectral dips may also be interesting. The aim in this project will be to investigate the dips via simulation of the EEL spectra. Towards this goal, a previously-developed COMSOL approach will be adapted and applied to our experimental case. It is noted that extensive experimental data is already available for comparison with, and validation of, the EELS simulations. Note that this project is suitable for the EPFL minor in photonics.
Contact: [email protected]