Engineering topics
Publications and source records attributed to Ogbuji, Linus U. J. T..
Two types of pest behavior in SiC/BN/SiC composites are distinguished and illustrated: one intrinsic and progressive, the other extrinsic and catastrophic. Their similarities and differences are presented. Some recent remedies for SiC/BN/SiC pest are discussed.
SiC-SiC composites are prone to pest degradation at intermediate temperatures where the formation of a protective scale of silica by direct oxidation of SiC is too slow to offer effective protection from attack by ambient gases. The attack results in the removal of the BN interphase and its replacement by a borosilicate liquid (which dissolves and weakens the fiber) and finally by SiO2, (which embrittles the composite by bonding fiber to fiber or to matrix). This paper reports a study aimed at mapping out the temperature range of such pest degradation and the severity of its effect.
The oxidative durability of a SiC-SiC composite with Hi-Nicalon fiber and BN interphase was investigated at 800 C (where pesting is known to occur in SiC-SiC composites) for exposure durations of up to 500 hours and in a variety of oxidant mixes and flow rates, ranging from quasi-stagnant room air, through slow flowing O2 containing 30-90% H2O, to the high-velocity flame of a burner rig. Degradation of the composite was determined from residual strength and fracture strain in post-exposure mechanical tests and correlated with microstructural evidence of damage to fiber and interphase. The severest degradation of composite behavior was found to occur in the bumer rig, and is shown to be connected with the high oxidant velocity and substantial moisture content, as well as a thin sublayer of carbon indicated to form between fiber and interphase during composite processing.
The oxidation behavior of reaction-formed silicon carbide (RFSC) ceramics was investigated in the temperature range of 1100 to 1400 C. The oxidation weight change was recorded by TGA; the oxidized materials were examined by light and electron microscopy, and the oxidation product by x-ray diffraction analysis (XRD). The materials exhibited initial weight loss, followed by passive weight gain (with enhanced parabolic rates, k(sub p)), and ending with a negative (logarithmic) deviation from the parabolic law. The weight loss arose from the oxidation of residual carbon, and the enhanced k(sub p) values from internal oxidation and the oxidation of residual silicon, while the logarithmic kinetics is thought to have resulted from crystallization of the oxide. The presence of a small amount of MoSi, in the RFSC material caused a further increase in the oxidation rate. The only solid oxidation product for all temperatures studied was silica.
A suitable interphase material for non-oxide ceramic-matrix composites must be resistant to oxidation. This means it must exhibit a slow rate of oxidation, and its oxidation product must be such as to ensure that the system survives oxidation when it does occur. Because the current benchmark interphase materials, carbon and boron nitride, lack these qualities, a porous fiber coating was developed to satisfy both the mechanical and oxidative requirements of an interphase for the SiC/SiC and SiC/Si2N4 composites that are of interest to NASA. This report presents the interphase microstructure achieved and the resulting characteristics of fiber push-out from a matrix of reaction-bonded silicon nitride (RBSN), both as-fabricated and after substantial annealing and oxidation treatments.
The results of two-step oxidation experiments on chemically-vapor-deposited Si3N4 and SiC at 1350 C show that a correlation exists between the presence of a Si2N2O interphase and the strong oxidation resistance of Si3N4. During normal oxidation, k sub p for SiC was 15 times higher than that for Si3N4, and the oxide scale on Si3N4 was found by SEM and TEM to contain a prominent Si2N2O inner layer. However, when oxidized samples are annealed in Ar for 1.5 h at 1500 C and reoxidized at 1350 C as before, three things happen: the oxidation k sub p increases over 55-fold for Si3N4, and 3.5-fold for SiC; the Si3N4 and SiC oxidize with nearly equal k sub p's; and, most significantly, the oxide scale on Si3N4 is found to be lacking an inner Si2N2O layer. The implications of this correlation for the competing models of Si3N4 oxidation are discussed.
Microstructural and energy dispersive spectrometry evidence is produced, from transmission electron microscopy, to show that a silicon oxynitride inner layer is produced by the oxidation of silicon nitride in dry oxygen at 1350 C as proposed by Tressler et al. (1989). However, details of the microstructures at the oxide/nitride interface do not agree entirely with the rest of the Tressler model for the oxidation of Si3N4.
The effect of thermal excursion and thermal cycling on the oxidation stability of chemical vapor-deposited (CVD) SiC and Si3N4 was studied at 1350 C. Thermal cycling alone produced no noticeable change in oxidation kinetics. However, TEM showed that oxide scales grown in cycles consist of alternating layers of SiO2 and Si2N2O. When the oxidation of CVD SiC or Si3N4 at 1350 C was interrupted with a 1.5-h annealing in Ar at 1500 C, the kinetics of reoxidation at 1350 C were found to be drastically increased. The SiC and Si3N4 then oxidized essentially at the same rate, which is over 50 times the preannealing rate, and comparable to the expected oxidation rate of these materials at 1500 C.
Severe difficulties are encountered in the preparation of oxidized Si3N4 specimens for XTEM transmission electromicroscopic inspection, in virtue of the extreme difference between Si3N4 and SiO2 mechanical properties. Attention is presently given to a preparation method in which an overlayer of the nitride is always occluded; this protects the oxide through most of the thinning that specimen preparation entails. An XTEM image of the oxide/nitride interface is presented.