634 S.R. Choi et al. /Journal of the European Ceramic Society 25(2005)1629-1636 O Nicalon/BSAS W200 △H- Nicalon/BSAS Theoretical 0 20 Prelaod factor, a x100 [9 Prelaod factor, a x100 [% C/sic (2-D) ension1200°c Tension1200°c Theoretical Theoretical 100 10 020406080100 Prelaod factor, a x100 [ Fig. 5. Results of preload tests, plotted with ultimate tensile strength as a function of preload factor for(a) Nicalon/BSAS and Hi-Nicalon/BSAS, (b)SiCr/MAs, (c)SiCr/SiC (enhanced), and(d)Cr/SiC ("standard"and"enhanced")ceramic matrix composites at elevated temperatures in air. A theoretical prediction based on Eq (4)is included for comparison for each composite material where ofp is ultimate tensile strength with a preload stress or test time was greater than 80% of fracture stress or a is in the range of0 s a < 1. It is noted from Eq.(4) total test time that ultimate tensile strength under preload is more sensi- The insignificant difference in ultimate tensile strength of tive to higher preload factor a and lower n value, because the Cr/sic composite between two preloads of 0 and 80% of much augmented delayed failure occurring under these is particularly noteworthy. The total test time at no preload conditions was about 5-7 h for the Cr/SiC composite(standard and en- Each solid line in Fig. 5 indicates the theoretical predic- hanced), while the respective total test time with an 80% tion based on Eq(4), together with the estimated value of preload was about 1-2h. Despite this appreciable difference n and the ultimate tensile strength with no preload. The pre- in test time(or exposure time for oxidation) between the two diction, despite a limited number of test specimens used preloads, the resulting strength difference was minimal. This in good agreement with the experimental data except for the suggests that oxidation would not be a sole source of the SiCH/MAS composite. Note that Eq(4)was derived based composite failure as well as of the rate dependency. A further on the power-law, slow crack growth formulation, Eq.(2) tudy using increased number of test specimens would reveal Therefore, the reasonable applicability of the preload anal- more detail aspects of failure mechanism(s) involved in the sis to the current composites suggests that delayed failure Cr/SiC composite. However, it is important to state at this process of these composites would be the one governed by point that apart from detailed understanding of a complex the power-law type of slow crack growth, as expressed in oxidation kinetics associated with the C/Sic composite, the from the figure that the overall difference in ultimate tensile results of both ultimate tensile strength and preload tes E s Eq.(2). This is consistent with the observations of the pre- composite failure can be described phenomenologically vious preload studies using other CMCs. It is also noted the simple power-law formulation of Eq(2), based on the trength between two preloads (a=0 and 80%)was insignif- icant, resulting in a considerable saving (80%)of test time 3.3. Constant stress(stress rupture)tests by applying the 80% preload. This indicates that any signifi cant crack growth that would control ultimate tensile strength A summary of the results of constant stress(stress rupture of a composite did not occur even when the applied stress to testing for the Nicalon/BSAS composite(batches A and B)at test specimen during test reached up to 80%of its fracture 1100C is presented in Fig. 6, where time to failure was plot tress. Conversely, the crack growth or damage to control fi- ted against applied stress in log-log scales. a decrease intime nal catastrophic failure would have occurred when applied to failure with increasing applied stress, which represents a1634 S.R. Choi et al. / Journal of the European Ceramic Society 25 (2005) 1629–1636 Fig. 5. Results of preload tests, plotted with ultimate tensile strength as a function of preload factor for (a) Nicalon/BSAS and Hi-Nicalon/BSAS, (b) SiCf/MAS, (c) SiCf/SiC (“enhanced”), and (d) Cf/SiC (“standard” and “enhanced”) ceramic matrix composites at elevated temperatures in air. A theoretical prediction based on Eq. (4) is included for comparison for each composite material. where σfp is ultimate tensile strength with a preload and α is in the range of 0 ≤ α < 1. It is noted from Eq. (4) that ultimate tensile strength under preload is more sensi￾tive to higher preload factor α and lower n value, because of much augmented delayed failure occurring under these conditions. Each solid line in Fig. 5 indicates the theoretical predic￾tion based on Eq. (4), together with the estimated value of n and the ultimate tensile strength with no preload. The pre￾diction, despite a limited number of test specimens used, is in good agreement with the experimental data except for the SiCf/MAS composite. Note that Eq. (4) was derived based on the power-law, slow crack growth formulation, Eq. (2). Therefore, the reasonable applicability of the preload anal￾ysis to the current composites suggests that delayed failure process of these composites would be the one governed by the power-law type of slow crack growth, as expressed in Eq. (2). This is consistent with the observations of the pre￾vious preload studies using other CMCs.1 It is also noted from the figure that the overall difference in ultimate tensile strength between two preloads (α = 0 and 80%) was insignif￾icant, resulting in a considerable saving (∼80%) of test time by applying the 80% preload. This indicates that any signifi- cant crack growth that would control ultimate tensile strength of a composite did not occur even when the applied stress to test specimen during test reached up to 80% of its fracture stress. Conversely, the crack growth or damage to control fi- nal catastrophic failure would have occurred when applied stress or test time was greater than 80% of fracture stress or total test time. The insignificant difference in ultimate tensile strength of the Cf/SiC composite between two preloads of 0 and 80% is particularly noteworthy. The total test time at no preload was about 5–7 h for the Cf/SiC composite (standard and en￾hanced), while the respective total test time with an 80% preload was about 1–2 h. Despite this appreciable difference in test time (or exposure time for oxidation) between the two preloads, the resulting strength difference was minimal. This suggests that oxidation would not be a sole source of the composite failure as well as of the rate dependency. A further study using increased number of test specimens would reveal more detail aspects of failure mechanism(s) involved in the Cf/SiC composite. However, it is important to state at this point that apart from detailed understanding of a complex oxidation kinetics associated with the Cf/SiC composite, the composite failure can be described phenomenologically by the simple power-law formulation of Eq. (2), based on the results of both ultimate tensile strength and preload tests. 3.3. Constant stress (stress rupture) tests A summary of the results of constant stress (stress rupture) testing for the Nicalon/BSAS composite (batches A and B) at 1100 ◦C is presented in Fig. 6, where time to failure was plot￾ted against applied stress in log–log scales. A decrease in time to failure with increasing applied stress, which represents a