Search for a command to run...
Bend tests, both uniaxial and biaxial, are widely used to determine the strength of ceramics. Besides the use of simple specimen shapes that are easy to machine, a main advantage of these tests is the elimination of alignment and gripping problems occurring in pure tensile tests. In the present work, main attention is focused on the required accuracy of these bend tests and the determination of an appropriate mixed-mode fracture criterion for the prediction of biaxial strength from uniaxial data. For this purpose, the performance of both uniaxial (three- and four-point) bend tests and biaxial (ball-on-ring and ring-on-ring) tests has been studied by means of strain gauge measurements. An experimental accuracy of about 1% has been reached by the application of free rollers for the uniaxial tests. In the case of three-point bending, a correction has to be applied that accounts for wedging stresses in the specimen. Satisfying results regarding the accuracy of the biaxial tests were achieved using ball bearings. These ball bearings do not affect the axisymmetrical stress configuration within the center of the specimens. Subsequently, different types of alumina have been tested at room and elevated temperatures. The results of these tests have been interpreted by a weakest-link analysis based on the assumption that fracture initiated from surface flaws. This analysis indicated that different fracture criteria should be applied in order to predict the biaxial strength of these materials from uniaxial data. The applied fracture criteria show differences in shear stress sensitivity, indicating that different fracture mechanisms take place in the tested materials that have a different microstructure. Whether the best applicable fracture criterion for a material can be linked to its microstructure is the subject of ongoing research.