Microstructural Evolution of Silica on Single Crystal Silicon Carbide. Part II: Influence of Impurities and Defects

Abstract

Passive oxidation of single crystal silicon carbide (6H–SiC) resulted in the formation of a vitreous silica layer which crystallized gradually at temperatures near 1300°–1400°C. During this process, statistically distributed devitrification centers appeared and disk‐like aligned crystal plates (radialites) formed. The crystallization process did not necessarily start at structural defects although these often act as areas of preferred nucleation. A second structural transition from disk‐like radialites to small crystalline spheres (globulites) was not connected to structural defects but a consequence of the presence of impurities. Alkaline and earth‐alkaline elements are common contaminations within the atmospheres of typical furnaces fitted with alumina tubes. Globulite formation was a process of recrystallization catalyzed by an impurity‐related melt formed on top of the devitrified areas. Crystallization caused exsolution of reaction gas and local impurity enrichments as the solubility for these phases is much higher in the devitrifying vitreous silica matrix. Both clean and impurity‐loaded oxidation produced specific morphologies of scales and interfaces.

Oxidation kinetics strength of Hi‐NicalonTM‐S SiC fiber after oxidation in dry and wet air

“…These glass scales were 1.8 μm thick, which is consistent with the Deal‐Grove kinetic parameters for glass scale formation in (Table ), so although crystallization rates were drastically changed, oxidation rates for glass scale formation were not significantly affected. Impurities may be removed during glass crystallization, so FIB‐SIMS analysis of glass scales formed in the “dirty” furnace that was not baked out, with sizing burned off and with sizing dissolved off, were compared with glass scales formed in a baked‐out furnace after 100 hours at 800°C and 0.33 hours at 1200°C. Absolute element concentrations are very difficult to determine from SIMS without extensive calibration for the particular element and crystalline phase, so results are shown as count ratios between fibers oxidized in “dirty” (no bake‐out) and “clean” (baked‐out) furnaces (Figure ).…”

Silicon Carbide Oxidation in Steam up to 2 MPa

“…Note that in case of environments other than steam, silica formed on the surface of SiC is initially amorphous and can undergo slow crystallization . According to Wagstaff and Richards, water vapor has little effect on cristobalite nucleation but increases the parabolic crystallization rate by about a factor of four.…”

Hi‐NicalonTM‐SSiC Fiber Oxidation and Scale Crystallization Kinetics

“…Crystallization of SiO 2 scales to cristobalite or tridymite reduces oxygen permeability, and is a source of confusion in interpretation of SiC oxidation data that does not distinguish crystalline and amorphous scales . Crystallization of SiO 2 scales on nominally pure SiC is reported to start at temperatures of 1400°C, at temperatures as low as 700°C, and at temperatures in between . Disparity in these results illustrates the extreme sensitivity of silica crystallization to SiC impurities and trace H 2 O and other gases in the environment …”

“…Factors that affect SiC oxidation rates include impurities in the fiber, particularly alkali and alkali earths. These change oxidation rates, reduce scale viscosity, and lower temperatures for scale crystallization . Moisture has similar effects .…”