The effects of small crystal tilts on dynamical scattering: why simulated images are thinner than experimental ones

Abstract

It is now well-established that high-resolution transmission electron microscope image simulation can accurately reproduce experimental HRTEM images when imaging parameters are accurately known. These parameters include specimen properties such as unit cell structure, crystal orientation, and specimen thickness, as well as microscope parameters such as beam energy and coherence (both spatial and temporal), objective aperture size, and objective lens defocus and spherical aberration. Of these parameters, the most difficult to measure, and thus to include accurately, is the specimen thickness (and also the temporal coherence; on the other hand, although spatial coherence is easily measured, its effects are not always accurately included in all simulations). In many investigations that utilize matches between experimental and simulated HRTEM images, specimen thickness is treated as a disposable parameter; i.e. the experimental specimen thickness is not measured, but assigned to be the thickness that was used in the simulation program to produce the simulated image that best matched the experimental one.