Journal of the American Ceramic SociefyPeres-Rigueiro et al. Vol. 82. No. 12 (A) Fibe Fiber 100nn 10 0.8 T0.6 0.4 0.2 00 05 Energy(ke E Fig. 5. TEM photographs of SH-C-N-matrix composite, showing BN coating(A)in the as-received condition and(B)after I h at 1200C in air. Energy spectra of the BN/fiber interlayer in the as-received condition and after I h at 1200oC are shown below the corresponding micrographs to this assumption, the differences in oxidation between this surface. The formation of SiO on the fiber surface immedi- composite and its Al, O3-matrix counterpart should be in the ately led to the nucleation of defects, severely reducing the structure and/or composition of this interface. In particular, the fiber strength and, consequently the composite strength and high carbon content of the interfacial layer, which is beneficial toughness. To assess the importance of this phenomenon, the in to fiber/matrix debonding in the presence of a crack, may be a situ fiber strength in the as-received composite, at 25C, and negative factor from the point of view of oxidation resistance, after I h at 8000, 1000, and 1200C was determined from the because carbon is rapidly removed at high temperature, open- radius of the mirror region on the fiber fracture surface. An ing the path for the entry of oxygen. This hypothesis is sup empirical relationship between the fracture mirror radius, ams ported by the results reported by Sun et al.2 on a BMAs and the fiber strength, S, was shown by numerous authors to be glass-ceramic matrix reinforced with Nicalon fibers with a dual of the form, Sic/BN coating. The composite in that study was subjected to a heat treat- Sa A (1) ment("ceraming')at 1200@C in argon before high-temperature where Am usually is denoted the mirror constant, estimated at testing to crystallize the matrix. This treatment also affected the 2.51 MPam for Nicalon SiC fibers. 16, 7, The fracture su interface: Two thin sublayers(one carbon rich and another faces of -130 fibers were analyzed for each temperature. Most SiO2 rich) that were present at the BN/fiber interface in the of the fibers exhibited a distinct mirror-mist-hackle structure as-pressed composite disappeared in the ceramed material. 9(type II), in which the mirror radius could be determined easily As a result, the oxidation resistance of the bn and the bn/fiber However, this radius was too short to be measured accurately interface was excellent, as shown by the analyses of samples in a small fraction of fibers(type 1). In addition, no distinct subjected to creep tests in air at 1100.C over hundreds of fracture mirror boundary was seen in other fibers, and the whole fiber fracture surface was specular(type Ill). The pro- portions of fibers with type L, Il, and Ill fracture surfaces V. Fiber degradation shown in Table I for each temperature These results can be used to compute the fiber-failure prob- As indicated already, the BN/fiber interlayer in the Al2O3. ability, F, as a function of the fiber strength, for each tempera- natrix composite seemed to change by oxidation of the fiber ture. The strength data computed from the mirror radiusto this assumption, the differences in oxidation between this composite and its Al2O3-matrix counterpart should be in the structure and/or composition of this interface. In particular, the high carbon content of the interfacial layer, which is beneficial to fiber/matrix debonding in the presence of a crack, may be a negative factor from the point of view of oxidation resistance, because carbon is rapidly removed at high temperature, open￾ing the path for the entry of oxygen. This hypothesis is sup￾ported by the results reported by Sun et al.12 on a BMAS glass-ceramic matrix reinforced with Nicalon fibers with a dual SiC/BN coating. The composite in that study was subjected to a heat treat￾ment (“ceraming”) at 1200°C in argon before high-temperature testing to crystallize the matrix. This treatment also affected the interface: Two thin sublayers (one carbon rich and another SiO2 rich) that were present at the BN/fiber interface in the as-pressed composite disappeared in the ceramed material.19 As a result, the oxidation resistance of the BN and the BN/fiber interface was excellent, as shown by the analyses of samples subjected to creep tests in air at 1100°C over hundreds of hours. V. Fiber Degradation As indicated already, the BN/fiber interlayer in the Al2O3- matrix composite seemed to change by oxidation of the fiber surface. The formation of SiO2 on the fiber surface immedi￾ately led to the nucleation of defects, severely reducing the fiber strength and, consequently, the composite strength and toughness. To assess the importance of this phenomenon, the in situ fiber strength in the as-received composite, at 25°C, and after 1 h at 800°, 1000°, and 1200°C was determined from the radius of the mirror region on the fiber fracture surface. An empirical relationship between the fracture mirror radius, am, and the fiber strength, S, was shown by numerous authors to be of the form, Sam 4 Am (1) where Am usually is denoted the mirror constant, estimated at 2.51 MPazm1/2 for Nicalon SiC fibers.16,17,27 The fracture sur￾faces of ∼130 fibers were analyzed for each temperature. Most of the fibers exhibited a distinct mirror-mist-hackle structure (type II), in which the mirror radius could be determined easily. However, this radius was too short to be measured accurately in a small fraction of fibers (type I). In addition, no distinct fracture mirror boundary was seen in other fibers, and the whole fiber fracture surface was specular (type III). The pro￾portions of fibers with type I, II, and III fracture surfaces are shown in Table I for each temperature. These results can be used to compute the fiber-failure prob￾ability, F, as a function of the fiber strength, for each tempera￾ture. The strength data computed from the mirror radius are Fig. 5. TEM photographs of Si–C–N-matrix composite, showing BN coating (A) in the as-received condition and (B) after 1 h at 1200°C in air. Energy spectra of the BN/fiber interlayer in the as-received condition and after 1 h at 1200°C are shown below the corresponding micrographs. 3498 Journal of the American Ceramic Society—Pe´rez-Rigueiro et al. Vol. 82, No. 12