October 1998 Communications of the American Ceramic Sociery 2739 Fig. 1. Cross section of dense RBAO body (98.6% TD) with 3. Platinum-coated sapphire fiber pushed out of dense RBAO 1550C. It was decided not to polish the sample in order to prevent ix showing debon nd ductile deformation of the coating remain visible on the surface of the sappie 5 damage to the ductile interphase. Therefore, smal defects fro the ductility of the metal phase. 2 Both matrix and coating show no differences between sintered specimens and samples lI. Results and discussion annealed for up to 50 h in air Vickers indentations gave first indications about a weak vith continuous fiber reinforcement can be fabricated. because metal interphase. The cracks introduced by, the indentor run of the good flowing characteristics of the powder and the plas- tic deformability of the metallic aluminum during pressing. almost defect-free green compacts can be obtained by simple matrix are weak enough to allow the desired crack deflection tion is limited because of the size and rigidity of the sapphire and debonding. In Fig. 2, a sample cracked in the cleavage fibers as well as the simple hand laying technique used for apparatus is shown. In this case the debonding has mainly green body fabrication, fiber volume contents up to =20% were taken place between fiber and coating and coating deforms achieved. High green densities of >60%TD and the volume plastically. The smooth fiber surfaces prove the absence of expansion related to the oxidation of aluminum to alumina detrimental reactions between fiber. matrix and coating These results already show the suitability of platinum coat- and. therefore no defects are formed. 19 In accordance with the and both matrix and fiber. As shown in Fig 3, debond can results of Lange 20 these defect-free compacts can be sintered in occur between all components of the composite simulta- air without cracking. Final densities of 98.6% TD were ob- interphase itself. The corresponding pushout diagram(Fig. 4) verage grain size of the sintered specimens was 1.5 and 1.7 shows no sharp debonding peak stress but a soft transition um for a sintering time of 2 h and an annealing time of 50 h between debonding and sliding out of the fiber, correlated with respectively. It can be concluded that grain coarsening is al the ductile deformation behavior of the platinum coating. After most completely inhibited by the zirconia particles.2I debonding, pushout forces decrease because of the lower co- during processing. No indications of detrimental reactions be- the decreasing embedding length of the fiber. However, the ditionally, no gaps etween fibers, matrix, and coating were lated from the maximum of the pushout curves are between K,aBAO=7.5×1056-9×10°k4=7×106-85×10 50 F·482N atd=19.67 um Fig. 2. Cracked sample showing debonding between platinum slurry coating and fiber. The interphase deforms plasticallyIII. Results and Discussion Using RBAO as matrix material, homogeneous composites with continuous fiber reinforcement can be fabricated. Because of the good flowing characteristics of the powder and the plas￾tic deformability of the metallic aluminum during pressing, almost defect-free green compacts can be obtained by simple powder metallurgical processing. Although fiber volume frac￾tion is limited because of the size and rigidity of the sapphire fibers as well as the simple hand laying technique used for green body fabrication, fiber volume contents up to ≈20% were achieved. High green densities of >60% TD and the volume expansion related to the oxidation of aluminum to alumina (DV 4 28%) lead to reduced shrinkage on sintering compared with conventional alumina ceramics.15 Therefore, the stress regime in these composites is expected to be at a very low level and, therefore, no defects are formed.19 In accordance with the results of Lange20 these defect-free compacts can be sintered in air without cracking. Final densities of 98.6% TD were ob￾tained by pressureless sintering in air for 2 h at 1550°C. The average grain size of the sintered specimens was 1.5 and 1.7 mm for a sintering time of 2 h and an annealing time of 50 h, respectively. It can be concluded that grain coarsening is al￾most completely inhibited by the zirconia particles.21 As shown in Fig. 1, the platinum coating remains fully intact during processing. No indications of detrimental reactions be￾tween the components of the composite have been found. Ad￾ditionally, no gaps between fibers, matrix, and coating were formed due to low CTE mismatch (aAl203 ≈ 7 × 10−6 –8.5 × 10−6 K−1, aRBAO ≈ 7.5 × 10−6 –9 × 10−6 K−1, aPt ≈ 9 × 10−6 K−1) and the ductility of the metal phase.12 Both matrix and coating show no differences between sintered specimens and samples annealed for up to 50 h in air. Vickers indentations gave first indications about a weak metal interphase. The cracks introduced by the indentor run inside the coating around the fibers. Using the cleavage appa￾ratus, it can be shown that, even at unstable crack growth, both the interface fiber/coating as well as the interface coating/ matrix are weak enough to allow the desired crack deflection and debonding. In Fig. 2, a sample cracked in the cleavage apparatus is shown. In this case the debonding has mainly taken place between fiber and coating and coating deforms plastically. The smooth fiber surfaces prove the absence of detrimental reactions between fiber, matrix, and coating. These results already show the suitability of platinum coat￾ings as a weak interphase for oxide composites. First single fiber pushout tests confirm weak bonding between the coating and both matrix and fiber. As shown in Fig. 3, debonding can occur between all components of the composite simulta￾neously, combined with a plastic deformation of the platinum interphase itself. The corresponding pushout diagram (Fig. 4) shows no sharp debonding peak stress but a soft transition between debonding and sliding out of the fiber, correlated with the ductile deformation behavior of the platinum coating. After debonding, pushout forces decrease because of the lower co￾efficient of friction for sliding compared to static friction and the decreasing embedding length of the fiber. However, the detailed investigation of frictional sliding is beyond the scope of this paper. The values of the interfacial shear stress calcu￾lated from the maximum of the pushout curves are between Fig. 1. Cross section of dense RBAO body (98.6% TD) with sap￾phire fiber, PVD-coated with platinum, after annealing for 50 h at 1550°C. It was decided not to polish the sample in order to prevent damage to the ductile interphase. Therefore, small defects from cutting remain visible on the surface of the sapphire fiber. Fig. 2. Cracked sample showing debonding between platinum slurry coating and fiber. The interphase deforms plastically. Fig. 3. Platinum-coated sapphire fiber pushed out of dense RBAO matrix showing debonding and ductile deformation of the coating. Fig. 4. Typical pushout curve of platinum-coated sapphire fiber (re￾calculated taking elastic deformation of the testing apparatus into account). October 1998 Communications of the American Ceramic Society 2739