Peter Binev
Approximation Methods In Electron Microscopy
Electron microscopy and in particular Scanning Transmission Electron Microscopy (STEM) has made tremendous advances in the instrumentation in the last couple of decades. These improvements allowed the scientists to observe materials at the picometre scale. However, the use of standard data processing has not allowed the extraction of all the available information. The structure of the noise is very complex and is far from being fully understood. The dominant presence of spatial distortions requires new mathematical tools and methods to better understand and process the STEM data.
The talk will give a short introduction to the specifics of STEM data and will focus on its modeling that allows its better understanding from a mathematical point of view. A particular topic is the non-rigid registration and averaging of series of aberration-corrected Z-contrast STEM images to demonstrate sub-picometre precision measurements of atom positions and achieve five to seven times better precision than previous methods. Another one is improvements of the alignment procedure for electron tomography and the development of a tomographic reconstruction based on partial segmentation. Some problems related to energy-dispersive X-ray spectroscopy will be discussed as well.