Translational research on clinically failed zirconia implants.

OBJECTIVES: To provide fractographic analysis of clinically fractured zirconia implants recovered with their cemented crown. To calculate bending moments, corresponding stress and crack onset location on the implant's fracture surface using a mathematical model integrating spatial coordinates of the crown-implant part and occlusal loading obtained from 2D and 3D images. METHODS: 15 fractured zirconia implants parts (11 posterior and 4 anterior) with their all- ceramic crowns still cemented on it were recovered. The implants were first generations from four manufacturers (AXIS Biodental, Z-Systems, Straumann, Swiss Dental Solutions). The time-to-failure varied between 2weeks and 9years. Fractography was performed identifying the failure origin and characteristic surface crack features. From 2D and 3D digital images of the crown-implant part, spatial coordinates anchoring the crown's occlusal contacts with the implant's central axis and reference plane were integrated in a mathematical model spreadsheet. Loads of 500 N in total were selectively distributed over identified occlusal contacts from wear patterns. The resultant bending and torsion moments, corresponding shear, tensile, maximum principal stress and von Mises stress were calculated. The fracture crack onset location on the implant's fracture surface was given by an angular position with respect to an occlusal reference and compared with the location of the fracture origin identified from fractographic analysis. RESULTS: Implants fractured from the periphery of the smaller inner diameter between two threads at the bone-entrance level except for one implant which failed half-way within the bone. The porous coating (AXIS Biodental) and the large grit alumina sandblasting (Z-System) created surface defects directly related to the fracture origin. The model spreadsheet showed how occlusal loading with respect to the implant's central axis affects bending moments and crack onset. Dominant loads distributed on contacts with important wear pattern provided a calculated crack onset location in good agreement with the fractographic findings of the fracture origin. SIGNIFICANCE: Recovered broken zirconia implant parts with their restorative crowns can provide not only information regarding the failure origin using fractography but also knowledge regarding occlusal crown loading with respect to the implant's axis. The mathematical model was helpful in showing how occlusal loading affects the location of the fracture initiation site on clinical zirconia implant fracture cases.

Fatigue behavior in water of Y-TZP zirconia ceramics after abrasion with 30 µm silica-coated alumina particles.

OBJECTIVE: The use of a 30 µm alumina-silica coated particle sand (CoJet™ Sand, 3M Espe), has shown to enhance the adhesion of resin cements to Y-TZP. The question is whether or not sandblasting 30 µm particles does negatively affect the fatigue limit (S-N curves) and the cumulative survival of Y-TZP ceramics. METHOD: Four zirconia materials tested were: Zeno (ZW) (Wieland), Everest ZS (KV) (KaVo), Lava white (LV) and Lava colored (LVB) (3M Espe). Fatigue testing (S-N) was performed on 66bar of 3 mm × 5 mm × 40 mm with beveled edges for each zirconia material provided by the manufacturers. One half of the specimens were CoJet sandblasted in the middle of the tensile side on a surface of 5 mm × 6 mm. Cyclic fatigue (N=30/group) (sinusoidal loading/unloading at 10 Hz between 10% and 100% load) was performed in 3-point-bending in a water tank. Stress levels were lowered from the initial static value (average of N=3) until surviving 1 million cycles. Fatigue limits were determined from trend lines. Kaplan-Meier survival analysis was performed to determine the failure stress at the median percentile survival level for 1 million of cycles before and after sandblasting. The statistical analyses used the log-rank test. Characterization of the critical flaw was performed by SEM for the majority of the failed specimens. RESULTS: The fatigue limits "as received" (ctr) were: LV=720 MPa, LVB=600 MPa, KV=560 MPa, ZW=470 MPa. The fatigue limits "after CoJet sandblasting" were: LV=840 MPa, LVB=788 MPa, KV=645 MPa, ZW=540 MPa. The increase in fatigue limit after sandblasting was 15% for Zeno (ZW) and Everest (KV), 17% for Lava (LV) and 31% for Lava colored (LVB). The KM median survival stresses in MPa were: ZW(ctr)=549 (543-555), ZW(s)=587 (545-629), KV(ctr)=593 (579-607), KV(s)=676 (655-697), LVB(ctr)=635 (578-692), LVB(s)=809 (787-831), LV(ctr)=743 (729-757), LV(s)=908 (840-976). Log-rank tests were significantly different (p<0.001) for all sandblasted groups vs. the "as received" except for Zeno (Wieland) (p=0.295). Failures started from both intrinsic and machined flaws. SIGNIFICANCE: 30 µm particle sandblasting did significantly improve the fatigue behavior of three out of four Y-TZP ceramic materials and can therefore be recommended for adhesive cementation procedures. This study was supported in part by grants from the Swiss Society for Reconstructive Dentistry (SSRD) and 3M Espe.