FAILURE OF ALL-CERAMIC FIXED PARTIAL DENTURES IN-VITRO AND IN-VIVO - ANALYSIS AND MODELING

Hertzian cone cracks visible at the loading site of 20 all-ceramic fixed partial dentures (FPDs), tested in vitro, led to the hypotheses that failure was due to the propagation of localized contact damage crack systems (Hertzian stress state) and that such damage was an unlikely clinical failure mode. Fractographic analysis of the 20 laboratory-failed and nine clinically-failed all-ceramic FPDs allowed for definitive testing of these hypotheses and a comparison between in vitro and in vivo failure behavior. In all cases, failure occurred in the FPD connectors (none from contact damage), with approximately 70 to 78% originating from the interface between the core and veneer ceramics. The coincidence between failure origins provides strong evidence that the in vitro test modeled aspects of structural behavior having clinical importance. The fractographic observations, coupled with the in vitro failure load data, furnished very specific boundary conditions which were applied to constrain mathematical models of FPD connector failure. Finite element analysis (FEA) of the laboratory FPDs found that maximum principal tensile stresses would occur at locations consistent with the fractographic observations only if: (1) there were appropriate elastic moduli differences between the ceramics; and (2) a small amount of abutment rotation was allowed. Weibull failure probability (P-f) calculations, incorporating FEA stress profiles, very closely replicated the laboratory failure distribution only when: (1) the veneer ceramic was much weaker than the core ceramic; and (2) the Weibull modulus of the core-veneer interface was much lower than that for the free veneer surface (i.e., the interface is of lower quality with regard to defects). This combined fractographic and mathematical analysis of FPD connectors suggests that the core-veneer interface is an important failure source and that the veneering ceramic overwhelmingly controls load-bearing capability. Observations from failed clinical restorations provided critical guidance in validating a laboratory test and focusing a mathematical failure model.