Defect imaging and analysis for predicting the reliability of ceramics

Abstract

Some outstanding properties of ceramics, (e.g. high hardness, wear and temperature resistance), make them attractive for engineering applications. However, their brittleness and a broad scatter in mechanical strength are serious drawbacks that often hinder their application. The main reason for high variability in mechanical strength is the combination of low fracture toughness and fracture initiating defects with variable size and geometry.

Initially, the variation in mechanical strength, described by the Weibull modulus was determined. Subsequently, fracture models plus the fractography were used to predict strength and compared with the results of mechanical testing. Applying high-throughput microscopy and image analysis on the same materials enables prediction of defect populations, their size and nature. Based on these results, the prediction of mechanical properties is addressed.

This study used tested four-point bending bars and focussed on the detection of intrinsic defects. Investigated materials were Al2O3 in two grades, a Y-TZP ZrO2 with added Al2O3-dispersions, and commercial SiC in two grades, one with high-porosity. Broken bars were examined by fractography, followed by a quantitative microstructure analysis of microsections.

Analogous defect information (e.g. size, geometry) as found by fractography could be obtained by ceramography. A correlation with the mechanical properties could be demonstrated. The Weibull moduli deduced from the defect distribution acquired from image analysis are in good accordance with the findings from the strength tests, with one of the SiC materials the only exception. The impact of failure initiating defects on fracture can be evaluated and explained using defect models.

The thesis demonstrates the significant advantages of the ceramographic approach developed in this work in obtaining microstructural data to estimate the mechanical reliability of ceramics. The approach is fast and gives additional insights. The accuracy of predicting the Weibull Modulus is high plus the investigations can be carried out non-destructively on real components as it is shown in this work. Furthermore, the capability of determining microstructural defects allows a deeper understanding of processing issues and supports quality improvement of ceramic components.