Determining the 3D atomic structure of nanomaterials is critical to understand their unique properties. Therefore, a thorough quantitative characterization by TEM is of great importance. The use of parameter estimation-based methods, and more recently by applying deep learning, allows us to extract reliable structural and chemical information from experimental STEM images. Recent progress enables us to extend this methodology to analyze time series of images using a hidden Markov model which is very promising for revealing dynamic structural changes resulting during in situ experiments. Moreover, progress in the quantitative analysis of heterogeneous materials has become possible by combining ADF STEM with EDX in a quantitative framework. This contribution aims to explain these recent developments using current state-of-the-art experimental examples.
Copyright. All Rights Reserved.