K K. Chawla et al. Journal of the European Ceramic Society 20(2000)551-559 50 um Monazites Saphikon 10 um 10μm Fig 4. Interface ng at the(five-dip) monazite/fiber interface with cracks produced by two 98N(15 s) indentations in the matrix:(a) low magnification SEl, and (c) backscattered electron image of the same area shown in(b). The bright phase surrounding the Saphikon fi Note the interfacial debonding occurred along the Saphikon fiber/ monazite interface, but not along the polycrystalline just as likely that the interphase cracked, rather than ated clamping pressures, the polycrystalline alumina, separation occurring at one the two interfaces. This monazite interface was much rougher than the single could be caused by the microcracking of the coating. crystal Saphikon fiber/monazite interface When a vickers indentor with a 49n indentation was Fiber pushout tests showed that both five- and 10-dip used to produce cracks in the fiber, it resulted in crack coated fibers slid smoothly from the alumina matrix and arrest, crack deflection, and interfacial debonding in five shout load did not lead to matrix cracking. The and 10-dip coated composites. Again, debonding was shear stress-displacement curve and SEM images for a observed at the smooth Saphikon fiber/monazite inter- five-dip coated fiber composite with 122 um fiber diam face but not at the rough polycrystalline alumina / mon- eter and 134 um specimen thickness are shown in Fig. 6 azite interface. This is contrary to the analysis of In Fig. 6(b), the debonded monazite coating is stuck to interfacial debonding due to Morgan and Marshall, the matrix, i.e. the debonding occurred mostly along the According to Morgan and Marshall, only when a crack Saphikon fiber/monazite interface. Fig. 7 shows the grew from monazite to polycrystalline alumina was shear stress-displacement curve and SEM images for a interfacial debonding expected and observed, not vice five-dip coated fiber composite with 163 um fiber diam versa. However, the present work showed that when eter and 145 um specimen thickness. A sinusoidal var cracks approached any one of the two interfaces, poly- iation in the shear stress-displacement curve can be crystalline alumina monazite or single crystal(Saphi- seen. This was probably caused by sinusoidal asperities kon)alumina/monazite interfaces, interfacial debonding existing on the fiber surface. occurred at the smooth single crystal alumina/monazite A comparison of pushout shear stress/displacement interface rather than at the rough interface between curves of five- and 10-dip coated composites showed polycrystalline alumina/monazite. It would appear that that the five-dip coated composites debonded at higher in our case, the interfacial roughness played a very shear stress values than the 10-dip coated composites, important role in interfacial debonding. In spite of a and also the frictional shear stress in these composites certain surface roughness on the Saphikon fiber, which was higher. The reasons are as follows. Debonding was grown into the fiber during manufacture and gener- usually involves a Mode II (shear) fracture phenomenonjust as likely that the interphase cracked, rather than separation occurring at one the two interfaces. This could be caused by the microcracking of the coating. When a Vickers indentor with a 49 N indentation was used to produce cracks in the ®ber, it resulted in crack arrest, crack de¯ection, and interfacial debonding in ®ve and 10-dip coated composites. Again, debonding was observed at the smooth Saphikon ®ber/monazite inter￾face but not at the rough polycrystalline alumina/mon￾azite interface. This is contrary to the analysis of interfacial debonding due to Morgan and Marshall,7 According to Morgan and Marshall,7 only when a crack grew from monazite to polycrystalline alumina was interfacial debonding expected and observed, not vice versa. However, the present work showed that when cracks approached any one of the two interfaces, poly￾crystalline alumina/monazite or single crystal (Saphi￾kon) alumina/monazite interfaces, interfacial debonding occurred at the smooth single crystal alumina/monazite interface rather than at the rough interface between polycrystalline alumina/monazite. It would appear that in our case, the interfacial roughness played a very important role in interfacial debonding. In spite of a certain surface roughness on the Saphikon ®ber, which was grown into the ®ber during manufacture and gener￾ated clamping pressures, the polycrystalline alumina/ monazite interface was much rougher than the single crystal Saphikon ®ber/monazite interface. Fiber pushout tests showed that both ®ve- and 10-dip coated ®bers slid smoothly from the alumina matrix and the pushout load did not lead to matrix cracking. The shear stress±displacement curve and SEM images for a ®ve-dip coated ®ber composite with 122 mm ®ber diam￾eter and 134 mm specimen thickness are shown in Fig. 6. In Fig. 6(b), the debonded monazite coating is stuck to the matrix, i.e. the debonding occurred mostly along the Saphikon ®ber/monazite interface. Fig. 7 shows the shear stress±displacement curve and SEM images for a ®ve-dip coated ®ber composite with 163 mm ®ber diam￾eter and 145 mm specimen thickness. A sinusoidal var￾iation in the shear stress±displacement curve can be seen. This was probably caused by sinusoidal asperities existing on the ®ber surface. A comparison of pushout shear stress/displacement curves of ®ve- and 10-dip coated composites showed that the ®ve-dip coated composites debonded at higher shear stress values than the 10-dip coated composites, and also the frictional shear stress in these composites was higher. The reasons are as follows. Debonding usually involves a Mode II (shear) fracture phenomenon. Fig. 4. Interfacial debonding at the (®ve-dip) monazite/®ber interface with cracks produced by two 98 N (15 s) indentations in the matrix: (a) low magni®cation SEI, (b) high magni®cation SEI, and (c) backscattered electron image of the same area shown in (b). The bright phase surrounding the Saphikon ®ber is monazite. Note the interfacial debonding occurred along the Saphikon ®ber/monazite interface, but not along the polycrystalline alumina/monazite interface. 554 K.K. Chawla et al. / Journal of the European Ceramic Society 20 (2000) 551±559