FURI | Spring 2022

Sample Preparation and Microstructure Characterization of Novel Ceramic Matrix Composites

Accurate assessment of the performance of ceramic matrix composites (CMCs) requires capturing inherent manufacturing-induced flaws including defects, voids, and pre-existing cracks. Comprehensive microstructure characterization along with uncertainty quantification are necessary to generate representative volume elements (RVEs) suitable for high-fidelity physics-based multiscale predictive models. Therefore, the objective of this research is to construct a dictionary of high-resolution micrographs to train a deep learning (DL) framework capable of generating such experimentally informed RVEs. These models will be used for material properties homogenization, as well as damage initiation and localization. In this work, CMC samples are cut using a low-speed diamond wafering blade, cold mounted onto an epoxy resin, and subsequently polished using a series of Struers polishing disks, each with a successively smaller grit size (from 30 𝜇m down to 0.04 𝜇m). A multiscale material characterization study is then performed using confocal microscopy with high resolution and variable magnifications. Acquiring high-fidelity micrographs is found to be challenging and quite heavily dependent on the nuances of the equipment utilized.