N P Bansal /Journal of the European Ceramic Sociery 29(2009)525-535 X-ray diffraction(XRD) patterns were recorded at room tem- 2. Hi-Nicalon/BN/SIC/BSAS CMC rature using a step scan procedure(0.02%/20 step, time/step v=043)#1-29-96 0.5 or I s)on a Phillips ADP-3600 automated diffractometer BN/SIC-coated Hi-Nicalon fiber equipped with a crystal monochromator employing Cu Ko radi- measured from dimensions as by the archimedes method Microstructures of the polished cross-sections and fracture surfaces were observed in a JEOL JSM-840A scanning electron microscope. Prior to analysis, a thin layer of carbon was evaporated onto the SEM specimens for electrical conductivity. The elemental compositions of the fibersurface coatings were TGA curves for BSAS monolith, BN/SiC-coated Hi-Nicalon fiber and analyzed with a scanning Auger microprobe(Fisons Instru- alon/BN/SC/BSAS composite recorded at a heating rate of 5.C/min in ments Microlab Model 310-F) The fibers for this analysis were mounted on a stainless steel sample mount by tacking the ends with colloidal graphite. Depth profiling was per- exposures in air and mechanical testing. The volume fraction of formed by sequential ion-beam sputtering and Auger analysis fibers in the composite was found to be -0.32. The ion etching was done with 3 keV argon ions rastered For high-temperature annealing, the CMC bars were rested over an approximately I mm2 area. The etch rate in Ta2O5 on the edges of an alumina boat placed inside a programmable under these conditions was 0.05 nm/s. Auger electron spec- box furnace. The furnace temperature was raised at a heating troscopy(AES)analysis of the coated Hi-Nicalon fibers was rate of 20C/min. CMC bars were annealed at 550, 800, 900, performed using an electron beam current of approximately 1000, 1100, and 1200C for 100 h in stagnant ambient air and 1.5 nA. The beam was rastered over a 2 um x 20 um area of furnace cooled. Dimensions and weight of each test bar were the fiber with the long axis of the area aligned with the long recorded before and after annealing fiber axis. Spectra were acquired in integral mode at beam Mechanical properties were determined from apparent energy of 2 keV and depth profiles were generated by plot tress-strain curves recorded from a three-point flexure test ting elemental peak areas against ion etch time. The atomic speed of 1.27 mm/min(0.05 in /min) and support span(L)of the spectrometer transmsionrr dividing the peak areas by flexure test bars. Stress, o, was calculated from beam theory, sensitivity factors were derived from spectra of ion etched Si, assuming a linear elastic beam, using the equation B, SiC, BN, and T102 standards. The depth scale is from the Ta2O5 calibration and no attempt has been made to adjust (1) for the actual etch rate for each material. Only the fibers with a smooth surface coating. rather than those having thick where b and h are the width and thickness of the test sample and and rough coating morphologies, were used for Auger analy- P is the load. The yield stress, y, was taken from the onset fsis deviation from linearity in the stress-strain curve. Elastic mod ulus of the composite was determined from the linear portion of 3. Results and discussion the stress-strain curve Cyclic fiber push-in tests were performed using a desktop 3.1. Scanning Auger analysis apparatus previously described, but with the addition of a sym- metrically placed pair of capacitance gauges for displacement Elemental composition depth profiles obtained from scan- measurements. Thin sections of the composites, cut normal to ning Auger microprobe analysis for the BN/SiC coatings on the fiber axis with a diamond saw, and polished down to a 0. 1-um Hi-Nicalon fibers are shown in Fig. 1. The coating consists of finish on both top and bottom faces were tested. Final specimen 0. 15 um thick Si-rich SiC followed by 0.6 um of carbon rich thickness was typically about 3 mm. Fibers were pushed in using"BN. In addition, unintentionally deposited carbon layer is also a 700-included-angle conical diamond indenter with a 10-um present between the SiC and"BN"coatings. Another predom diameter flat base. To prevent the sides of the conical inden- nantly carbon layer is also seen between the"BN"coating and ter from impacting the matrix, push-in distances were restricted the fiber surface. Presence of free Si has also been detected to just a couple of microns. Unless otherwise noted, each test in the SiC coating layer by Raman microspectroscopy. This is consisted of five cycles of loading and unloading between a consistent with the results of another study22 which found the selected maximum load and a minimum load of 0.01 N at room Sic layer to be rich in Si from scanning Auger analysi mperature in ambient atmosphere Thermogravimetric analysis(TGA)was carried out at a heat- 3.2. Microstructural analy ing rate of 5C/min under flowing air from room temperature to faced with a computerized data acquisition and analysis system. SEM micrographs taken from the polished cross-section of 1500C using a PerkinElmer TGA-7 system, which was inter- he unidirectional hot pressed composite are shown in Fig. 2.N.P. Bansal / Journal of the European Ceramic Society 29 (2009) 525–535 527 Fig. 3. TGA curves for BSAS monolith, BN/SiC-coated Hi-Nicalon fiber and Hi-Nicalon/BN/SiC/BSAS composite recorded at a heating rate of 5 ◦C/min in air. exposures in air and mechanical testing. The volume fraction of fibers in the composite was found to be ∼0.32. For high-temperature annealing, the CMC bars were rested on the edges of an alumina boat placed inside a programmable box furnace. The furnace temperature was raised at a heating rate of 20 ◦C/min. CMC bars were annealed at 550, 800, 900, 1000, 1100, and 1200 ◦C for 100 h in stagnant ambient air and furnace cooled. Dimensions and weight of each test bar were recorded before and after annealing. Mechanical properties were determined from apparent stress–strain curves recorded from a three-point flexure test using an Instron 4505 universal testing machine at a cross-head speed of 1.27 mm/min (0.05 in./min) and support span (L) of 40 mm. Strain gauges were glued to the tensile surfaces of the flexure test bars. Stress, σ, was calculated from beam theory, assuming a linear elastic beam, using the equation: σ = 3PL 2bh2 (1) where b and h are the width and thickness of the test sample and P is the load. The yield stress, σy, was taken from the onset of deviation from linearity in the stress–strain curve. Elastic mod￾ulus of the composite was determined from the linear portion of the stress–strain curve. Cyclic fiber push-in tests were performed using a desktop apparatus previously described,21 but with the addition of a sym￾metrically placed pair of capacitance gauges for displacement measurements. Thin sections of the composites, cut normal to the fiber axis with a diamond saw, and polished down to a 0.1-m finish on both top and bottom faces were tested. Final specimen thickness was typically about 3 mm. Fibers were pushed in using a 70◦-included-angle conical diamond indenter with a 10-m diameter flat base. To prevent the sides of the conical inden￾ter from impacting the matrix, push-in distances were restricted to just a couple of microns. Unless otherwise noted, each test consisted of five cycles of loading and unloading between a selected maximum load and a minimum load of 0.01 N at room temperature in ambient atmosphere. Thermogravimetric analysis (TGA) was carried out at a heat￾ing rate of 5 ◦C/min under flowing air from room temperature to 1500 ◦C using a PerkinElmer TGA-7 system, which was inter￾faced with a computerized data acquisition and analysis system. X-ray diffraction (XRD) patterns were recorded at room tem￾perature using a step scan procedure (0.02◦/2θ step, time/step 0.5 or 1 s) on a Phillips ADP-3600 automated diffractometer equipped with a crystal monochromator employing Cu K radi￾ation. Density was measured from dimensions and mass as well as by the Archimedes method. Microstructures of the polished cross-sections and fracture surfaces were observed in a JEOL JSM-840A scanning electron microscope. Prior to analysis, a thin layer of carbon was evaporated onto the SEM specimens for electrical conductivity. The elemental compositions of the fiber surface coatings were analyzed with a scanning Auger microprobe (Fisons Instru￾ments Microlab Model 310-F). The fibers for this analysis were mounted on a stainless steel sample mount by tacking the ends with colloidal graphite. Depth profiling was per￾formed by sequential ion-beam sputtering and Auger analysis. The ion etching was done with 3 keV argon ions rastered over an approximately 1 mm2 area. The etch rate in Ta2O5 under these conditions was 0.05 nm/s. Auger electron spec￾troscopy (AES) analysis of the coated Hi-Nicalon fibers was performed using an electron beam current of approximately 1.5 nA. The beam was rastered over a 2 m × 20m area of the fiber with the long axis of the area aligned with the long fiber axis. Spectra were acquired in integral mode at beam energy of 2 keV and depth profiles were generated by plot￾ting elemental peak areas against ion etch time. The atomic concentrations were calculated by dividing the peak areas by the spectrometer transmission function and the sensitivity fac￾tors for each peak, then scaling the results to total 100%. The sensitivity factors were derived from spectra of ion etched Si, B, SiC, BN, and TiO2 standards. The depth scale is from the Ta2O5 calibration and no attempt has been made to adjust for the actual etch rate for each material. Only the fibers with a smooth surface coating, rather than those having thick and rough coating morphologies, were used for Auger analy￾sis. 3. Results and discussion 3.1. Scanning Auger analysis Elemental composition depth profiles obtained from scan￾ning Auger microprobe analysis for the BN/SiC coatings on Hi-Nicalon fibers are shown in Fig. 1. The coating consists of ∼0.15m thick Si-rich SiC followed by ∼0.6m of carbon rich “BN”. In addition, unintentionally deposited carbon layer is also present between the SiC and “BN” coatings. Another predomi￾nantly carbon layer is also seen between the “BN” coating and the fiber surface. Presence of free Si has also been detected13 in the SiC coating layer by Raman microspectroscopy. This is consistent with the results of another study22 which found the SiC layer to be rich in Si from scanning Auger analysis. 3.2. Microstructural analysis SEM micrographs taken from the polished cross-section of the unidirectional hot pressed composite are shown in Fig. 2.