Biaxial flexure strength determination of endodontically accessed ceramic restorations

ObjectivesTo report analytic solutions capable of identifying failure stresses from the biaxial flexure testing of geometries representative of endodontic access cavities prepared through dental restorative materials.MethodsThe ring-on-ring biaxial flexure strength of annular discs with a central circular hole supported peripherally by a knife-edge support and loaded evenly at the upper edge of the central hole were solved using general expressions of deformations, moments and shears for flat plates of a constant thickness. To validate the solutions, finite element analyses were performed. A three-dimensional one-quarter model of the test was generated using a linear P-code FEA software and the boundary conditions represented the experimental test configuration whereby symmetry planes defined the full model. To enable comparison of the maximum principal stresses with experimental derived data, three groups of nominally identical feldspathic ceramic disks (n = 30) were fabricated. Specimens from Group A received a 4 mm diameter representative endodontic access cavity and were tested in ring-on-ring. Group B and C specimens remained intact and were tested in ring-on-ring and ball-on-ring, respectively, to give insight into strength scaling effects. Fractography was used to confirm failure origins, and statistical analysis of fracture strength data was performed using one-way ANOVAs (P < 0.05) and a Weibull approach.ResultsThe developed analytical solutions were demonstrated to deviate <1% from the finite element prediction in the configuration studied. Fractography confirmed the failure origin of tested samples to coincide with the predicted stress maxima and the area where fracture is observed to originate clinically. Specimens from the three experimental groups A–C exhibited different strengths which correlated with the volume scaling effects on measured strength.SignificanceThe solutions provided will enable geometric and materials variables to be systematically studied and remove the need for load-to-failure ‘crunch the crown’ testing.