J.R. Verkouteren, R.B. Marinenko, and D.S. Bright

Objective: To determine compositional and crystalline phase changes upon heating of plasma-sprayed (Zr,Y)Ox ceramic coatings.

Problem: Ceramic coatings based on yttria-stabilized zirconia are used as thermal barriers on aircraft and land-based turbines, and on diesel engines (Pratt and Whitney, General Electric, Caterpillar, METCO). The coatings fail after a number of thermal cyclings; the exact causes are not known. Redistribution of yttria in the coating from the tetragonal (t) phase to a low-yttria monoclinic (m) phase and a high-yttria cubic (c) phase is thought to contribute to the failure. Bulk diffraction methods (neutron and x-ray) are used primarily for phase identification in these materials, but the analysis is challenging due to the similarity in structure between the t and c phases. The lattice parameters of each phase are used to track the redistribution of yttria, but the challenging nature of the analysis raises doubts as to the accuracy of the results. Testing the current understanding of the chemical and phase dynamics in these materials falls under the CSTL program for Chemical Characterization of Materials.

Approach: Representative materials obtained from a collaboration with MSEL include coatings prepared by plasma spray using two different feedstock powders. Portions of the coatings were heated to temperatures above 1000 ºC to simulate in-service conditions. Elemental wavelength dispersive spectrometry (WDS) x-ray compositional mapping and microbeam x-ray diffraction (mXRD) were used to relate yttria concentrations to the observed phases. Phase compositions and element distributions were determined from the maps with NIST software. Compositionally distinct areas were identified from the maps and analyzed by mXRD to determine the crystalline phase. These results were compared to those from bulk neutron and x-ray diffraction.

WDS compositional x-ray map (a) with low-yttria areas in red. microXRD pattern (b) of t-ZrO2 from 10 micrometer wide area.

Results and Future Plans: The distribution of yttria and the resulting phase compositions are dependent upon the choice of feedstock powders used in the preparation of the coatings. One coating was quite inhomogeneous, with a much larger range in yttria concentration for the c phase than recognized by bulk diffraction. After heating, this material became more homogeneous, although a c phase with a relatively high yttria content was retained. The second coating was homogeneous as produced but formed a low-yttria phase upon heating. This low-yttria phase is not monoclinic, as would be expected from phase equilibria considerations, but is instead tetragonal (see Figure). The bulk composition shifts toward slightly higher yttria concentrations to accommodate the formation of the low-yttria t phase. This phase was not detected by bulk diffraction, and, in fact, there is no reference in the literature to a t phase with such low yttria concentrations (0.01 mole fraction). These results show that individual coatings display different responses to heating, in some cases following routes not predicted by phase equilibria studies. The next step will be to study coatings on substrates before and after failure to determine if there are any correlations between compositional/phase changes and failure mechanisms.

Web Contact micro@nist.gov