Contaminant identification during laser cleaning of thermal barrier coatings

Abstract

Laser ablation has previously been demonstrated as a technique for removal of contaminants from engine-run thermal barrier coatings. In this paper, laser-induced breakdown spectroscopy (LIBS) was applied during ablation cleaning to quantify the contaminant concentrations. Calibration LIBS spectra were acquired from pressed powdered samples of calcium oxide, magnesium oxide, alumina, and silica, which are major components of surface contaminants (referred to as CMAS) in gas turbine engines. Proper orthogonal decomposition (POD) was applied to the calibration spectra from the CMAS components to generate an orthogonal mode set for reducing measured data. Since these POD modes can represent mixed compositions, an approach was derived to separate the modes into chemically unique spectra while retaining their orthogonal and normal properties. Results using these renormalized modes show that the calcium and magnesium oxide composition in CMAS can be generally quantified. Quantification for alumina and silica is limited due to their inherently lower signal. Measurements were applied to turbine blades run in a gas turbine engine to examine the distribution of CMAS contaminants across the surface. Results showed the CMAS contaminants on the blades were abundant in calcium and magnesium oxides, which represented over 85% of the LIBS spectra. Approximately 14% of the measured CMAS contaminant spectra were residual, representing noise and unidentified species.