Imaging with STEM detector, experiments vs. simulation

0431334 - ÚPT 2015 CZ eng A - Abstract
Mika, Filip - Konvalina, Ivo - Walker, Christopher
Imaging with STEM detector, experiments vs. simulation.
9th International Conference on Charged Particle Optics. Book of Abstracts. Brno: Institute of Scientific Instruments AS CR, v. v. i, 2014. s. 64. ISBN 978-80-87441-11-4.
[International Conference on Charged Parrticle Optics /9./. 31.08.2014-05.09.2014, Brno]
R&D Projects: GA TA ČR TE01020118; GA MŠMT(CZ) LO1212; GA MŠMT ED0017/01/01
Institutional support: RVO:68081731
Keywords : STEM * Monte-Carlo simulations * transmitted electrons
Subject RIV: JA - Electronics ; Optoelectronics, Electrical Engineering

Knowledge of angular and energy distribution of Transmitted Electrons (TE) is very important for the understanding of image formation in Scanning Transmission Electron Microscopy (STEM). Monte Carlo (MC) simulations are mostly used for theoretical study of these distributions. In this work the SEM-DRAFT (based on Geant4) software were used for carrying out the simulation of electron propagation through thin foils. The study was done by both an experiment and a simulation. We simulate real arrangement of specimen chamber including magnetic field of objective lens and geometrical arrangement of the STEM detector. TEs, the angular and energy distributions of which are given by MC simulations, are traced in EOD software which allows to simulate trajectories of electrons in electromagnetic fields. The experiments was performed on the SEM, Magellan 400, equipped by a multi-annular semiconductor STEM detector. The detector is divided into six concentric annuli and imaged in bright field (BF), four dark field (DF1-4) and high angle annular dark field (HAADF) modes. The last HAADF annulus is subdivided into six angular segments, but for our case we used it as one segment. During the experimental study we imaged light and heavy element foils of 100 nm thickness (silicon, gold). The primary energy range was between 15 keV–30 keV and we collected the image from each of six annuli and measured the average brightness of a defined image area to determine the signal intensity. The images were taken under the same settings of the brightness, the contrast and the primary current. Fig. 2 shows the results of MC simulations vs. experiment of signal intensity for Silicon. The differences in intensity can be explained as that simulation does not yet include complete of experimental arrangement, e.g. does not reflect the effect of multiple reflection and generation of a signal electrons from the sample holder, the chamber wall or the surface of the STEM detector.
Permanent Link: http://hdl.handle.net/11104/0236395