Characterisation of Functional Materials by Electron Microscopy and Machine Learning Techniques

Student thesis: Phd


This thesis demonstrates the implementation and development of advanced electron microscopy techniques for characterisation of nanostructures found in functional materials. A number of previously unreported phases were determined through this project that have an effect on the overall property of the materials. The first part of the project involved implementing precession electron diffraction tomography (PEDT) and kinematical refinement for structure determination of the sub-micron and nano-scaled phases found within BiMnO3 and La-doped SrTiO3, respectively. The sub-micron phase in BiMnO3 was determined to be bismuth oxide consisted of Bi2O2 layers separated by 4.8 Å. The crystal structure of a novel nano-phase in La-SrTiO3 was determined through PEDT. Oxygen occupancies were refined with a novel approach that only included the reflections that were believed to be directly associated with the oxide ions in the structure. The second part of the project introduced a technique for automated analysis of the microstructure through merging of simultaneously acquired signals of converged beam electron diffraction (CBED) and energy dispersive X-ray spectroscopy (EDS) data. Combining the simultaneously acquired datasets provided a deeper understanding of the microstructure than analysing either data in isolation. The covariance between the different signals improved the efficiency of machine learning approaches which can highlight the regions that require further analysis more readily. Using this approach, previously unknown structures were discovered in Co2FeSi and Fe2VAl systems. A crystal structure was determined for the unknown phase in Co2FeSi that explains the measurements in terms of a new stacking sequence on {111} lattice planes. This phase was also found within Fe2VAl, indicating the possibility of being present in other types of X2YZ Full-Heusler alloys.
Date of Award1 Aug 2023
Original languageEnglish
Awarding Institution
  • The University of Manchester
SupervisorAlexander Eggeman (Supervisor) & Sarah Haigh (Supervisor)


  • transmission electron microscopy
  • crystallography
  • functional materials
  • machine learning

Cite this