Abstract
Silicon carbide-based hot-section jet engine components enable increased fuel efficiency compared to state-of-the-art metal-based components but require environmental barrier coatings (EBCs) to inhibit degradation from the harsh engine environment. High temperature steam (a product of fuel combustion) and molten glass (the result of dirt and ash ingested into the engine hot section) can react with, damage, and reduce EBC lifetimes. The molten glass is primarily composed of Calcium-Magnesium-Aluminosilicates (CMAS). Studying CMAS-EBC interactions is crucial for enhancing CMAS mitigation strategies and EBC lifetime models. The work presented in this dissertation explores CMAS-steam synergy on Yb-silicate degradation, investigates EBC compositional effects on molten CMAS wetting behavior, and elucidates the effect of a fifth oxide (denoted by ‘X’) on CMXAS glass properties.
The effects of steam on CMAS degradation mechanisms are not well-known. CMAS-EBC degradation was performed at 1300°C in an environment-controlled tube furnace for 4-, 24-, and 96-hour durations in steam (90% H2O/10% O2) compared to dry-O2 (100% O2) and lab air environments. Three dense, model EBC materials were investigated: a phase-pure Yb2Si2O7, a nominal 20vol% Yb2SiO5 in Yb2Si2O7, and a nominal 25vol% Yb2Si2O7 in Yb2SiO5. The three substrates were loaded with ~40mg of Ca33-Mg9-Al13-Si45 (single cation oxide mol%) pre-reacted CMAS. Experiments were repeated in triplicate to evaluate the effect of steam on CMAS-EBC interactions. Steam showed an increase in molten CMAS spreading compared to the dry-O2 environment on both substrates. The 20Yb2SiO5/Yb2Si2O7 substrate reduced CMAS infiltration by an order of magnitude compared to the phase-pure Yb2Si2O7 substrates. In general, CMAS transport across the polished model EBC surface preferentially occurred along grain and/or phase boundaries. Steam also increased CMAS reactivity with phase-pure Yb2Si2O7 but did not affect CMAS reactivity with 20Yb2SiO5/Yb2Si2O7.
This work also assessed the effect of rare-earth disilicate (REDS) composition and processing method on CMAS wetting using a heating microscope to quantify contact angle and spreading dimensions. Substrates included freestanding atmospheric plasma spray (APS) REDS coatings (for RE = Y, La, Nd, Gd, Yb, Lu), dense phase-pure spark plasma sintered (SPS) Yb2Si2O7, and a dense SPS two-phase mixture of 20 vol% Yb2SiO5 in Yb2Si2O7. CMAS (Nominally Ca33-Mg9-Al13-Si45 in single cation mol%) was loaded as a 10 mg cylindrical rod atop the specimen surface, and heated in stagnant lab air to temperatures of 1250°C. The heating microscope was used to monitor the evolution of molten CMAS contact angle, width, and height as a function of time. Post-exposure CMAS contact angle, width, and height did not show systematic trends with composition. CMAS spreading and reactivity were also examined using plan view SEM/EDS and XRD. CMAS spreading decreased with RE cation size, correlating inversely with RE-apatite phase stability. APS YbMS increased CMAS spreading compared to dense APS YbDS. Processing effects showed polished SPS Yb-silicates and unpolished APS YbDS both increased CMAS spreading relative to dense polished APS YbDS. CMAS transport was observed to spread along grain boundaries and channels provided by surface roughness or porosity. Highly connected porosity promoted CMAS infiltration over surface spreading. The reactive wetting mechanism on REDS coating materials is hypothesized to be reaction-limited with substrate wettability increasing as rare-earth apatite formation driving force decreases.
Finally, this work investigated common natural-forming (X = Na2O, FeO, FeO2, TiO2) and coating-derived (X = Y2O3, ZrO2, HfO2, La2O3, Nd2O3, Gd2O3, Yb2O3, Lu2O3) oxide additions to CMXAS glasses - where X denotes a fifth oxide constituent and their effect on CMAS viscosity, coefficient of thermal expansion, softening temperature, and dilatometric glass transition temperature (CTE, Td, Tg). CMAS viscosity, CTE, Td, and Tg can indicate relative threat of CMAS attack or premature coating failure. Glass property relationships were elucidated by cation size effects and allow inferences to glass structure to be made. Iron oxide valence, group IV metal, and rare-earth metal cations – including one dual cation addition (Y3+ and Yb3+) – were explored. The baseline CMAS, nominally Ca33–Mg9–Al13–Si45 (single cation mol%), was synthesized from constituent oxide powders. Natural-forming oxide additions consistently operated as network modifiers. Coating-derived oxide additions behaved as network modifiers in the molten liquid state but acted as network formers in the condensed amorphous state. Fe3+ additions were shown to have the greatest effect of all additions on glass properties, exhibiting the greatest propensity for CMAS attack. Trends observed between dilatometric CMXAS glass properties allow for CMXAS properties to be inferred should one property (CTE, Td, Tg) be known. FactSage viscosity calculations were most consistent with experimental data compared to equivalent Thermo-Calc calculations. Molten CMAS and CMXAS viscosity measurements were shown to correlate with net cation field strength, offering a promising alternative in property prediction for EBC lifetime methodology where thermodynamic data are not available. Coating performance should consider the effect of coating constituents on CMAS viscosity and CTE, dissolution, and precipitation behaviors.
