S. Zhu et al. /Composites Science and Technology 59(1999)833-851 Creep at 1300C in Argon Standard SiC/SiC (a)180 MPa (b)45 MPa(left) (c)45 MPa(right) Fig. 19. Crack propagation paths on cross sections of crept specimens in argon. (a)1300.C, 180 MPa; (b)and(c)1300C, 45 MPa fiber architecture [28]. Since the stress is lower than the stresses and 4.5 at high stresses. This result is similar to matrix cracking stress, few matrix cracks exist in the the present result in the similar temperature and stress crept specimens(Fig. 19). Evans and Zok [80] stated conditions. However, creep strain of AlO3/SIC that the rupture ductility of the ceramic fibers is typi- order of magnitude higher than that of SiC/SiC cally quite low, and consequently, matrix cracking by Although Al2O3 fibers are less creep resistant than stress transfer often leads to creep rupture with brittle NicalonTM fibers, their creep onset temperatures and characteristics ep ductility are similar [41]. Since the difference in The present tensile creep strain rates at a given stress thermal expansion coefficient between the fibers and nd temperature were in the same order of magnitude as matrix in Al2O3/SiC is larger than that in SiC/SiC, the the flexural creep rates by Vicens et al. [92](Fig. 13). matrix cracking in Al2O3 /Sic is much more extensive The apparent activation energy for tensile creep was than that in SiC/SiC. Thus, the constraint of the matrix also similar to that for flexural creep [92]. However, the on creep of fibers in AlO3 /SiC is less. The stress expo- stress exponent for tensile creep was found to be larger nent for creep of Al2O3/SiC at high stresses is similar to than that for flexural creep. These phenomena were that of Al2O3 fibers. Therefore, it was suggested that often seen in creep tests in monolithic ceramics [93]. In creep rate of Al2O3/SiC was controlled by creep of the monolithic ceramics, the higher stress exponent in ten- bridging fibers at high stresses. Evidently, creep beha- slle creep in nexus ral creep is generally attributed to vior of Sic/SiC is different from that of Al2O3/SiC cavitation or microcracking in tension but not in com 2D SiCr/CAS composite has the same fibers and dif- ferent matrix as SiC/SiC. Creep strain of SiC/CAS [26] To examine the effects of fibers, we compare the pre- is also one order of magnitude higher than that of SiC/ ent results with the tensile creep behavior of an SiC. Here it should be noted that the matrix contributes AlO3(D/CVI-SiC composite (Al2O3/SiC) with a 2D little resistance to creep strain of the composite at woven fiber architecture [28-30]. The steady-state creep 1200 C. Moreover, the stress exponent for creep of rate in Al,O3/SiC exhibited two distinct regimes SiC/CAS is consistent with that of NicalonTM fibers depending on creep stress at the temperatures of 950 to indicating that the composite creep rate appears to be 1100 C. The stress exponent for creep was 9.5 at low primarily controlled by creep of NicalonTMfibers®ber architecture [28]. Since the stress is lower than the matrix cracking stress, few matrix cracks exist in the crept specimens (Fig. 19). Evans and Zok [80] stated that the rupture ductility of the ceramic ®bers is typi￾cally quite low, and consequently, matrix cracking by stress transfer often leads to creep rupture with brittle characteristics. The present tensile creep strain rates at a given stress and temperature were in the same order of magnitude as the ¯exural creep rates by Vicens et al. [92] (Fig. 13). The apparent activation energy for tensile creep was also similar to that for ¯exural creep [92]. However, the stress exponent for tensile creep was found to be larger than that for ¯exural creep. These phenomena were often seen in creep tests in monolithic ceramics [93]. In monolithic ceramics, the higher stress exponent in ten￾sile creep than in ¯exural creep is generally attributed to cavitation or microcracking in tension but not in com￾pression. To examine the e€ects of ®bers, we compare the pre￾sent results with the tensile creep behavior of an Al2O3(f)/CVI-SiC composite (Al2O3/SiC) with a 2D woven ®ber architecture [28±30]. The steady-state creep rate in Al2O3/SiC exhibited two distinct regimes depending on creep stress at the temperatures of 950 to 1100C. The stress exponent for creep was 9.5 at low stresses and 4.5 at high stresses. This result is similar to the present result in the similar temperature and stress conditions. However, creep strain of Al2O3/SiC is one order of magnitude higher than that of SiC/SiC. Although Al2O3 ®bers are less creep resistant than NicalonTM ®bers, their creep onset temperatures and creep ductility are similar [41]. Since the di€erence in thermal expansion coecient between the ®bers and matrix in Al2O3/SiC is larger than that in SiC/SiC, the matrix cracking in Al2O3/SiC is much more extensive than that in SiC/SiC. Thus, the constraint of the matrix on creep of ®bers in Al2O3/SiC is less. The stress expo￾nent for creep of Al2O3/SiC at high stresses is similar to that of Al2O3 ®bers. Therefore, it was suggested that creep rate of Al2O3/SiC was controlled by creep of the bridging ®bers at high stresses. Evidently, creep beha￾vior of SiC/SiC is di€erent from that of Al2O3/SiC. 2D SiCf/CAS composite has the same ®bers and dif￾ferent matrix as SiC/SiC. Creep strain of SiC/CAS [26] is also one order of magnitude higher than that of SiC/ SiC. Here it should be noted that the matrix contributes little resistance to creep strain of the composite at 1200C. Moreover, the stress exponent for creep of SiCf/CAS is consistent with that of NicalonTM ®bers, indicating that the composite creep rate appears to be primarily controlled by creep of NicalonTM ®bers. Fig. 19. Crack propagation paths on cross sections of crept specimens in argon. (a) 1300C, 180 MPa; (b) and (c) 1300C, 45 MPa. 846 S. Zhu et al. / Composites Science and Technology 59 (1999) 833±851