Abstract
Silicon Carbide (SiC) based Ceramic Matrix Composites (CMCs) entered service in aircraft turbine engines as replacements for some Ni-base superalloy components in 2016. The CMCs consist of SiC fibers, a BN interphase coating, and a SiC-based matrix and have several benefits over traditional superalloys, including higher operating temperature and lower density, which both contribute to increased engine efficiency. The CMC constituents (fibers, interphase, and matrix), however, have the potential to be exposed to hot combustion gases (including H2O and O2), if the environmental barrier coating delaminates or cracks, or there is a crack which penetrates into the CMC itself. To be able to predict the lifetime of CMCs, their oxidation behavior in the presence of the hot combustion gases must be understood.
The oxidation behavior of CMCs is studied in this work through oxidation of individual CMC constituents (fibers and stand-alone matrix) as well as CMCs in both dry and wet O2 environments using ThermoGravimetric Analysis (TGA). The temperatures studied were 800-1300°C, a relevant temperature range for turbine engine applications. Microscopy of oxidized materials helped to inform the key oxidation mechanisms at play.
Stand-alone SiC fibers were studied first. The oxidation behavior of Hi-Nicalon SiC fibers in dry O2 was similar to that of bulk SiC oxidized in the same conditions. In wet O2, however, oxidation at the higher temperatures (1200-1300°C) resulted in crystalline oxides that cracked in situ from hoop stresses.
The stand-alone matrix material consisted of SiC particulates in a continuous Si phase, with regions that contained boron. This material was oxidized in both dry and wet O2 using TGA. The TGA results showed slightly increased parabolic oxidation rates at 1200 and 1300°C for the matrix material, as compared with bulk SiC, due to the presence of boron. The oxidation kinetics at 800°C were also affected by boron, as neither linear nor parabolic oxidation behavior was observed through TGA. The oxidation behavior of the matrix material was considered as a baseline for CMC oxidation behavior.
The CMCs that were oxidized had a chemical vapor deposited SiC seal coat on all but one surface of the coupons. All CMC constituents were exposed to the oxidizing environments (O2 or H2O) on the exposed CMC face and reacted simultaneously to form SiO2 and B2O3 reaction products. Observations with TGA, scanning electron microscopy, and transmission electron microscopy indicated that the BN interphase in the CMCs sealed rapidly with the formation of borosilicate glass droplets in all oxidizing conditions. In the interior of the composite, where the partial pressure of the oxidant is reduced, SiC preferentially oxidized at the SiC/BN interfaces, as expected from thermodynamic considerations. Overall, oxidation of CMCs was minimal, reaching only ~20 µm below the surface in the most extreme conditions tested (wet O2 at 1300°C for 100h).
Thermally grown borosilicate glass droplets sealed the BN interphase on the exposed CMC surface. The exact composition of the borosilicate phase, however, could not be determined with typical spectroscopic techniques. A study of stand-alone borosilicate glasses of known compositions was conducted to probe the composition and properties of the thermally grown borosilicate glass. Borosilicate glasses ranging from 74-100 wt% B2O3 (balance SiO2) were synthesized in a box furnace and heated using a hot stage microscope to understand the temperature and viscosity of each composition on forming a spherical droplet shape. The resulting relationships between composition, temperature, and viscosity were used to determine the range in composition of the thermally grown borosilicate glasses on CMCs was likely between 43-63 wt% B2O3.
All of the results—from the stand-alone glasses, oxidation of individual CMC constituents, and oxidation of CMCs—were integrated together to develop a description of CMC oxidation mechanisms. The identified mechanisms are valid for all temperatures and environments studied here, and describe the borosilicate glass sealing of the exposed BN interphase, along with the preferential oxidation of SiC phases (fibers and matrix) below the exposed CMC surface. In addition to the mechanistic description, a comparison to current CMC oxidation models was conducted. Mechanisms and parameters in the existing models were assessed and values of these parameters consistent with observations of the present study were recommended for improved SiC/BN/SiC life prediction models.
