Title: Size Scale of Leveling Affects Fractal Failure Analysis of Y-TZP
Trent Wilkerson (Presenter)
University of Mississippi
Kartikeya Jodha, University of Mississippi Medical Center
Susana Salazar Marocho, University of Mississippi Medical Center
Gaurav Joshi, American Dental Association
Jason Griggs, University of Mississippi Medical Center
Objectives: Fractal geometry is useful in analysis of ceramic fracture surfaces. The fractal dimension (D) can be used to estimate fracture toughness of a material, which may help in identifying failure modes of prostheses. However, accurately measuring D values of fracture surfaces requires leveling the surfaces to be analyzed. Otherwise, smaller D values (negative bias) will result. Previous literature suggested that the surface can be leveled on a macroscopic scale – even when measuring D at the microscopic scale. We sought to discover if there was a difference between leveling on the scale of measurement and leveling on a larger scale when analyzing dental zirconia.
Methods: Impressions were taken of fractured four-point flexure bars (n=18) made of yttria-tetragonal polycrystalline zirconia (Y-TZP, ZirCAD, Ivoclar Vivadent) using light bodied impression material (Extrude, Kerr). Epoxy replicas (EpoxySet, AlliedHighTech) were made of two areas on each fracture surface, and 5µm x 5µm areas were scanned using an atomic force microscope (AFM, Bioscope Catalyst, Bruker) and were leveled using a first-order flattening operation. Each scanned area was divided into 2.5µm x 2.5µm quadrants. Image file format was converted using a custom script in MathCAD (PTC) and analyzed using FRACTALS software (Russ) with the Minkowski Cover algorithm.
Results: D values were not normally distributed (Shapiro-Wilk, p<0.05). Friedman repeated measures ANOVA on ranks showed that the images that were leveled and measured on the 5-µm scale had significantly higher D (associated with greater accuracy) than images leveled on the 5-µm scale but measured on the 2.5-µm scale (p<0.001). The median D values (inter-quartile range) for the 5-µm scale measurements was 2.152(2.139,2.160) while those for the quadrants measured at 2.5-µm scale were 2.140(2.129,2.150), 2.138(2.125,2.151), 2.138(2.127,2.151), and 2.136(2.120,2.154).
Conclusions: Y-TZP fracture surfaces should be leveled and measured at the same size scale to achieve greatest accuracy in D values. Leveling on larger scales is unacceptable.