Machine learning-facilitated multiscale imaging for energy materials

Abstract

The relationship between the structure of a material and its properties, the so-called structure-property correlations, is at the center of materials science. The microstructure of a material is an essential feature for the optimization of physicochemical properties with improved performance. However, it remains a challenge to recognize and extract all relevant information from microscopic images. Machine learning (ML) has entered the field of multiscale characterization and visualization in energy materials. The aim of this review is to provide concise tutorials on multiscale imaging techniques and ML methods, showing how they can be incorporated for solving physical and chemical problems. With a particular focus on image segmentation, we discuss noteworthy applications of ML in X-ray and electron microscopy imaging for rechargeable batteries, solar cells, and fuel cells and discuss how ML can facilitate identifying microstructures, enhancing image quality, and tracking dynamic processes occurring both inside and at the materials interfaces.