Zhongbo Lee

Incorporate elastic and inelastic scattering into image calculation for low-voltage transmission electron microscope

Day of Ph.D. defence: 28.03.2014



In this work, we have studied the dependence of image contrast on different parameters for low voltage TEM by means of image calculation. For pure elastic scattering, we have utilized a semi-experimental model which calculates the image by averaging over the energy distribution of the elastically scattered imaging electrons, derived from the experimental EELS data. The calculations were performed for graphene under the accelerating voltage of 80 kV and 20 kV. We have investigated the influence of geometrical aberrations, chromatic aberrations, energy distribution of the imaging electrons as well as other damping effects on the image contrast. Based on the calculations, we have shown the ways to obtain the optimum imaging conditions for low voltage TEMs and initiated the discussion of inelastic scattering under the accelerating voltage of 20 kV. For the calculation involving inelastic scattering, we have utilized the existing mutual coherence model. In order to improve the computational efficiency, we have derived a new approximation to factorize the mixed dynamic form factor, one of the key quantities for the image calculation involving inelastic scattering. The new approximation can be applied for different imaging conditions with enough accuracy. Experimental images are recorded with finite electron dose. We have explored the dependence of the signal-to-noise ratio, the atom contrast and the specimen resolution on the electron dose and the sampling of the detector. A modified definition of the atom contrast considering finite electron dose is introduced, and this definition is more reasonable for evaluating the object visibility in the experimental images than other existing contrast definitions. We have analyzed the 80 kV experimental HRTEM images of graphene based on the new definition of dose-dependent contrast and found good qualitative agreement between the experimental and calculated images.

Ph.D. thesis