A robust method to characterize rumpling in high-strength bond coats

Abstract

A new method has been developed to nondestructively characterize bond coats that rumple during thermal cycling. This method employs Fourier transforms of 2D optical profilometer data and is applied to coatings that have differing tendencies to rumple such as (Pt,Ni)Al coatings and high-strength γ′ or γ+γ′ bond coats. The method isolates the 2D periodic undulations characteristic of rumpling from other inhomogeneous time-dependent phenomena that occur at the surface, including HfO2 and NiAl2O4 spinel oxide growth, bond coat swelling, and noisy data. The method is benchmarked against the traditional rumpling indicators seen in the literature, surface roughness and tortuosity, using synthetic datasets to explore the strengths and limitations of the algorithm. The versatility of this method to quantify rumpling is demonstrated on experimental datasets from a β-phase bond coat with high levels of noisy or missing data. Application of the method provides a quantification of the rumpling amplitude and wavelength that are confirmed with metallographic cross sections of the samples and predicted by the Balint and Hutchinson rumpling model. The algorithm correctly measures (in contrast to the RMS surface roughness) an increasing surface undulation amplitude during initial furnace cycling, followed by a decrease in undulation amplitude near the end of the coating life as TGO spallation becomes prevalent.