Steps Toward Automated Failure Analysis Using FIB-SEM 3D Tomography and AI-based Segmentation

In failure analysis, the process of searching for a defect in a single field-return sample involves a multitude of manual analysis steps, applying different methods and FA techniques. In high-volume failure analysis and construction analysis–as practiced in high-volume foundry fabs or in the context of a substantial number of RMA samples–automation can significantly reduce resource requirements and improve tool utilization . However, finding and characterizing a defect at a failure site within a sample using the slice-and-view method in a focused ion beam-scanning electron microscope (FIB-SEM) has not yet been automated and still requires constant monitoring by a human operator. Recent advances in FIB-SEM technology have made it possible to automatically acquire reliable, high-resolution slice-and-view datasets (FIB-SEM tomography). Furthermore, 3D visualization of such data segmented by artificial intelligence (AI)-based algorithms facilitates a comprehensive evaluation of defect sites by enabling an assessment from multiple directions. As a step toward full automation of FIB-SEM defect analysis, this paper uses a case study to explore approaches that could serve as a basis for automating defect detection in 3D FIB-SEM data. We evaluated approaches based on comparing experimental data with reference data, using conventional image processing, AI-based feature detection, or AI-based segmentation.