Image Segmentation for FIB-SEM Serial Sectioning of a Si/C-Graphite Composite Anode Microstructure Based on Preprocessing and Global Thresholding.

The choice of materials that constitute electrodes and the way they are interconnected, i.e., the microstructure, influences the performance of lithium-ion batteries. For batteries with high energy and power densities, the microstructure of the electrodes must be controlled during their manufacturing process. Moreover, understanding the microstructure helps in designing a high-performance, yet low-cost battery. In this study, we propose a systematic algorithm workflow for the images of the microstructure of anodes obtained from a focused ion beam scanning electron microscope (FIB-SEM). Here, we discuss the typical issues that arise in the raw FIB-SEM images and the corresponding preprocessing methods that resolve them. Next, we propose a Fourier transform-based filter that effectively reduces curtain artifacts. Also, we propose a simple, yet an effective, global-thresholding method to identify active materials and pores in the microstructure. Finally, we reconstruct the three-dimensional structures by concatenating the segmented images. The whole algorithm workflow used in this study is not fully automated and requires user interactions such as choosing the values of parameters and removing shine-through artifacts manually. However, it should be emphasized that the proposed global-thresholding method is deterministic and stable, which results in high segmentation performance for all sectioning images.