A review of sample thickness effects on high-resolution transmission electron microscopy imaging.

ABSTRACT

High-resolution transmission electron microscopy (HRTEM) is an important approach to analyzing material structures. However, in reality, preparing a sufficiently thin sample for use in HRTEM, based on which images could be interpreted by weak phase object approximation theory, is difficult. During the imaging process, the thickness of the sample has two primary effects-a dynamical effect and a non-linear effect. Both are reviewed in this paper. Considering only the dynamical effect, the Bloch wave method and multislice theory have been proposed to understand the relationship between sample thickness and imaging. These methods exhibit high accuracy but high complexity as well. Sacrificing accuracy, pseudo-weak phase object approximation (PWPOA) theory can provide clues to the relationship in reciprocal space with greater simplicity. Meanwhile, in real space, channeling theory describes the dynamical effect with sufficient accuracy, and with the 1s state approximation, i.e., for a certain range of thicknesses, it provides a physical image and simplified expression with which to describe the relationship between the exit wave and sample thickness. As for the non-linear effect, a method of separating linear and non-linear information using a combination of transmission cross-coefficient theory and PWPOA theory was recently proposed. The variation of non-linear and linear imaging with sample thickness has also been discussed. A deep understanding has been acquired regarding the effects of the sample thickness, but a complete understanding of the HRTEM imaging process for thick samples has remained elusive. This understanding is crucial to the retrieval of structure from HRTEM images.