G N Morscher et al /Composites Science and Technology 68(2008 )3305-3313 331 Oxidized Not oxidized Fig. 11. Typical unbridged edge-crack growth in specimen tested at 30 Hz fatigue, 179 MPa for 8. 1 x 10 cycles (75 h). Tensile specimen inset indicates direction of unbridged crack propagation through cross-section Not Oxidized oXidized pical unbridged face-crack growth in specimen tested at 30 Hz fatigue, 220 MPa for 178,493 cycles (1. h). Tensile specimen inset indicates direction of by oxidation. For example, the fact that the room temperature been well documented, the cause of which can be linked to atom ultimate residual strength of the 110-MPa 2036-h creep speci- ic diffusional mechanisms [17]. However, another degradation mens was only slightly greater than one half the strength of the possibility would be due to diffusion of free Si from the compos as-produced strength with no evidence for an oxidation-induced ite MI matrix through the CvI Sic which has been observed to at- cause for failure indicates fiber degradation. Two potential mech- tack fibers and degrade composites with no stress above 1350C be considered for this behavior. Fiber degradation [18. This may appear more unlikely because no such mechanism could be due to an intrinsic creep-controlled flaw growth mecha- has been observed at temperatures lower than or equal to nism. Slow crack growth and creep degradation of Sic fibers has 1300C for the times considered here, although stress may en-by oxidation. For example, the fact that the room temperature ultimate residual strength of the 110-MPa 2036-h creep speci￾mens was only slightly greater than one half the strength of the as-produced strength with no evidence for an oxidation-induced cause for failure indicates fiber degradation. Two potential mech￾anisms can be considered for this behavior. Fiber degradation could be due to an intrinsic creep-controlled flaw growth mecha￾nism. Slow crack growth and creep degradation of SiC fibers has been well documented, the cause of which can be linked to atom￾ic diffusional mechanisms [17]. However, another degradation possibility would be due to diffusion of free Si from the compos￾ite MI matrix through the CVI SiC which has been observed to at￾tack fibers and degrade composites with no stress above 1350 C [18]. This may appear more unlikely because no such mechanism has been observed at temperatures lower than or equal to 1300 C for the times considered here, although stress may en￾Fig. 11. Typical unbridged edge-crack growth in specimen tested at 30 Hz fatigue, 179 MPa for 8.1 106 cycles (75 h). Tensile specimen inset indicates direction of unbridged crack propagation through cross-section. Fig. 12. Less typical unbridged face-crack growth in specimen tested at 30 Hz fatigue, 220 MPa for 178,493 cycles (1.6 h). Tensile specimen inset indicates direction of unbridged crack propagation through cross-section. G.N. Morscher et al. / Composites Science and Technology 68 (2008) 3305–3313 3311