Experimental assessment of toughness in ceramic matrix composites using the J-integral with digital image correlation part I: methodology and validation

Abstract

The complex nature of damage in ceramic matrix composites (CMCs) renders conventional methods of measuring quantitative fracture properties impractical. This paper is the first of two-part series that assesses the feasibility of using the J-integral with full-field deformation data from digital image correlation (DIC) to characterize toughness in continuous fiber SiC/SiC CMC laminates. It provides a resource for best practices when incorporating experimentally measured, full-field deformation data into J-integral evaluations of toughness. The techniques discussed are important for researchers attempting to measure fracture properties in advanced materials with damage mechanisms that have yet to be well characterized. Two methods for evaluating potential energy release rates J are presented: (1) numerical integration over a line contour and (2) Gaussian integration over an area contour. Accuracy and path independency for both methods were verified using analytically derived deformation fields for a center-cracked, infinite plate of isotropic material under equi-biaxial tension. Inherent noise in the deformation data and necessary contour truncation at the crack surfaces reduced accuracy and introduced path dependency. However, this was mitigated by careful noise filtering of the deformation data prior to the evaluation of J. Applying the line and area integrals to DIC data from tapered, double-cantilever beam, acrylic compact tension specimens resulted in the findings that (1) both integrals measured fracture toughness within a range of published values; (2) both integrals captured the constant stress intensity factor behavior that is characteristic of the tapered beam geometry; and (3) the area integral measurements were consistently larger than line integral measurements. The area integral is more accurate than the line integral as it samples more data points, reducing its sensitivity to experimental noise. Although variability in experimental data can be minimized, it cannot be eliminated. Measurement error is inevitable; thus, the line and area integrals must be used with caution when characterizing quantitative properties including fracture toughness.