The three common crystallographic forms of YSZ are found present in the coating after exposure to CMAS, each in a different region and with a distinct morphology and Y content.Much of the coating away from the substrate and the exposed surface retains the initialt0structure and chemical composition (B7.6% YO1.5), as well as the columnar morphology with nanoscale intragranular porosity, indicating that the thermochemical attack was minimal in these regions. However, the YSZ near the bulk CMAS deposit is monoclinic, depleted in Y, and that closest to the substrate is cubic, enriched in Y relative to the original composition. Minor amounts of Ca are incorporated into solid solution for both of these forms, but not to any detectable degree in the retainedt0form. No other crystalline phases bearing major amounts of Zr or Y were found in the areas of the coating affected by CMAS.
The cumulative evidence suggests that both the Y-enriched and Y-depleted zirconia phases evolve through crystallization from the CMAS melt into which the originalt0is concurrently dissolved, with the characteristics of the precipitated phase determined by the local chemistry. The situations for the bulk and upper regions are discussed first, as the lower interaction zone also involves the precipitation of a second crystalline phase based on Al2O3rather than ZrO2. A complete analysis requires detailed knowledge of the phase equilibria in the system Al2O3–CaO–MgO–SiO2–Y2O3–ZrO2, which is not available. The concept, however, can be illustrated qualitatively with the aid of the CaO–SiO2–ZrO2liquidus projection in Fig. 9. CMAS is represented by a binary liquid with the same C:S ratio (B0.73) in contact with ZrO2, depicted by the tie line in this figure. At equilibrium, this composition should actually be solid below B14301C, consisting of a mixture of pseudowollastonite (a-CS) and tridymite (fig. 10359 Roth28). However, the crystallization of these phases is often suppressed kinetically and one may reasonably assume that the composition selected may exist as a supercooled liquid with chemical characteristics similar to the CMAS melt. In that case, the actual boundary of thet-ZrO2liquidus would not be given by the position of the L1CS1Z twofold saturation line in Fig. 1, but shifted toward the C–S binary at temperatures below 14001C, as indicated by the arrows. It is then readily apparent from Fig. 1 that the amount of Z needed to saturate the CS melt is quite modest (a few percent), in agreement with the ZrO2content detected in much of the CMAS within the coating. A corollary is that the presence of MgO and Al2O3in the original CMAS does not seem to affect substantially the solubility of ZrO2in the melt.
Fig. 1.Schematically modified liquidus projectionfor the ternaryCaO–SiO2–ZrO2. The original diagram (fig. Zr-287 Ondik and McMurdie36),as many published around that time, ignores the existence of the cubic 2 tetragonal transformation at high temperature as well as the existence of a two-phase miscibility gap in the liquid for the SiO2–ZrO2binary. The schematic presented here incorporates qualitatively more recent evidence from partial isopleths and revised binaries. In general,the location of the relevant L1CS1Z two-phase saturation line on the liquidus is reasonably close to that in the original diagram. The tie line shown has the same C:S ratio as the model calcium–magnesium alumino silicate.
本文由桑尧热喷涂网收集整理。本站文章未经允许不得转载；如欲转载请注明出处，北京桑尧科技开发有限公司网址：http://www.sunspraying.com/
|
