S.R. Choi et al. /Journal of the European Ceramic Sociery 25(2005)1629-1636 635 ing configuration (constant stress-rate testing)to another Nicalon/BSAS(2-D) (constant stress)for a selected composite all support that delayed failure of the composites was controlled by the BATCH 'A power-law type of slow crack growth(or damage evolu- tion/accumulation), confirmed with not only the current 80} BATCH日 CMCs but the previous CMCs. It is noted that despite many 0000 differences in their processing, architecture, microstructure, and interface, compared with monolithic ceramics, CMCs still exhibit delayed failure, similar in principle to mono- lithic counterparts. This is simply due to the fact that macro- 310410510° topically, a CMC is a composite composed of two or more delayed-failure susceptible monolithic materials(co Time to failure, t, [s] stituents of fibers, matrices, interfaces, etc. ) The vulnera- bility to delayed failure would be greater in composite be- Fig.6.Results of constant stress("stress rupture")testing for Nicalon/BSAS cause of its more likelihood of chance to environmental composite(batches""(open triangle)and"B(closed triangle)at 1100C in air. The solid lines represent the predictions made based on Eq (5)from exposure by its inherently more open, porous microstruc the results of constant stress-rate testing( Fig. 2) tures, as compared to dense monolithic counterparts. The vulnerability of a composite, of course, will be increased ible to delaved susceptibility to delayed failure, was evident for the com- failure posite. The mode of fracture was similar to that in constant A subsequent importance drawn from the results of this stress-rate testing showing some fiber pullout with jagged work is that constant stress-rate testing, commonly utilized matrix cracking through the specimen-thickness direction. in monolithic ceramics, could be applicable to CMCs to de- At lower applied stresses, however, the fracture surfaces of termine their delayed-failure(or life prediction) parameters, both batches were somewhat flat with decreased fiber pull- at least for a short range of lifetimes, consistent with the pre- out, a change in fracture mode more likely to brittle fracture. vious observation. The merits of constant stress-rate testing The lines in Fig. 6 indicate life prediction from the constant are enormous in terms of test simplicity and test economy stress-rate data. The prediction, primarily applied to brittle (short test time and less data scatter)over other stress rup. monolithic materials, was made using the following relation ture or cyclic fatigue testing. A simplistic, phenomenolog based on the power-law formulation of Eq(2). ical law of delayed failure was only explored in this study to develop lifetime prediction testing and methodology for (5) CMCs, without accounting for detailed failure mechanisms associated with matrix/fiber interaction, matrix cracking and ts effect on slow crack growth, delayed failure of sustain- where tf and o are time to failure and applied constant stress ing fibers, and creep-associated deformation. 4-I8A micr respectively. Use of Eq.(5)together with n and d de- scopic level of study on this subject is thus needed. Finally termined in constant stress-rate testing allows one to pre- the results of this work suggest that care must be exer dict life under constant stress(stress rupture)loading. Even cised when characterizing elevated-temperature strength of though the small number of test specimens was used here composite materials. This is due to the fact that elevated due to limited material availability, the prediction was in rea- temperature strength has a relative meaning if a material ex- sonable agreement with experimental data at least for the hibits rate dependency: the strength simply depends on which Nicalon/BSAS composite. This indicates that the governing test rate one chooses(see Fig. 2). Therefore, use of at least failure law of the Nicalon/BSAS composite was very simi- two test rates(high and low) is recommended to better chan lar either in constant stress-rate or in constant stress loading. acterize high-temperature strength behavior of a composite Since the prediction(Eq (5))was made based on Eq (2), the material governing delayed-failure mechanism of the Nicalon/BSAS would be the one controlled by the power-law type of slow crack growth(Eq (2). Other CMCs also showed reasonable 4. Conclusions agreement in life between constant stress-rate and constant- stress loading configurations Elevated-temperature ultimate tensile strength of five different continuous fiber-reinforced ceramic composites 3.4. Implications including Nicalon/BSAS (2D), Hi-Nicalon/BSAS (2D), SiCr/MAS (ID), SiCr/SiC(2D woven: enhanced), and As seen in the preceding sections 3.1-3.3. the stren Cr/sic (2D woven: standard and enhanced), exhibited a dependency on test rate, the applicability of the preload tech- strong dependency on test rate, consistent with the behav- nique, and the reasonable life prediction from one load- ior observed in other CMCs as well as in many advancedS.R. Choi et al. / Journal of the European Ceramic Society 25 (2005) 1629–1636 1635 Fig. 6. Results of constant stress (“stress rupture”) testing for Nicalon/BSAS composite (batches “A” (open triangle) and “B” (closed triangle)) at 1100 ◦C in air. The solid lines represent the predictions made based on Eq. (5) from the results of constant stress-rate testing (Fig. 2). susceptibility to delayed failure, was evident for the com￾posite. The mode of fracture was similar to that in constant stress-rate testing showing some fiber pullout with jagged matrix cracking through the specimen-thickness direction. At lower applied stresses, however, the fracture surfaces of both batches were somewhat flat with decreased fiber pull￾out, a change in fracture mode more likely to brittle fracture. The lines in Fig. 6 indicate life prediction from the constant stress-rate data. The prediction, primarily applied to brittle monolithic materials, was made using the following relation based on the power-law formulation of Eq. (2). 1 tf = Dn +1 n + 1  σ−n (5) where tf and σ are time to failure and applied constant stress, respectively. Use of Eq. (5) together with n and D de￾termined in constant stress-rate testing allows one to pre￾dict life under constant stress (stress rupture) loading. Even though the small number of test specimens was used here due to limited material availability, the prediction was in rea￾sonable agreement with experimental data at least for the Nicalon/BSAS composite. This indicates that the governing failure law of the Nicalon/BSAS composite was very simi￾lar either in constant stress-rate or in constant stress loading. Since the prediction (Eq. (5)) was made based on Eq. (2), the governing delayed–failure mechanism of the Nicalon/BSAS would be the one controlled by the power-law type of slow crack growth (Eq. (2)). Other CMCs also showed reasonable agreement in life between constant stress-rate and constant￾stress loading configurations.1 3.4. Implications As seen in the preceding Sections 3.1–3.3, the strength dependency on test rate, the applicability of the preload tech￾nique, and the reasonable life prediction from one load￾ing configuration (constant stress-rate testing) to another (constant stress) for a selected composite all support that delayed failure of the composites was controlled by the power-law type of slow crack growth (or damage evolu￾tion/accumulation), confirmed with not only the current CMCs but the previous CMCs.1 It is noted that despite many differences in their processing, architecture, microstructure, and interface, compared with monolithic ceramics, CMCs still exhibit delayed failure, similar in principle to mono￾lithic counterparts. This is simply due to the fact that macro￾scopically, a CMC is a composite composed of two or more delayed–failure susceptible monolithic materials (con￾stituents of fibers, matrices, interfaces, etc.). The vulnera￾bility to delayed failure would be greater in composite be￾cause of its more likelihood of chance to environmental exposure by its inherently more open, porous microstruc￾tures, as compared to dense monolithic counterparts. The vulnerability of a composite, of course, will be increased if the constituents added are highly susceptible to delayed failure. A subsequent importance drawn from the results of this work is that constant stress-rate testing, commonly utilized in monolithic ceramics, could be applicable to CMCs to de￾termine their delayed–failure (or life prediction) parameters, at least for a short range of lifetimes, consistent with the pre￾vious observation.1 The merits of constant stress-rate testing are enormous in terms of test simplicity and test economy (short test time and less data scatter) over other stress rup￾ture or cyclic fatigue testing. A simplistic, phenomenolog￾ical law of delayed failure was only explored in this study to develop lifetime prediction testing and methodology for CMCs, without accounting for detailed failure mechanisms associated with matrix/fiber interaction, matrix cracking and its effect on slow crack growth, delayed failure of sustain￾ing fibers, and creep-associated deformation.14–18 A micro￾scopic level of study on this subject is thus needed. Finally, the results of this work suggest that care must be exer￾cised when characterizing elevated-temperature strength of composite materials. This is due to the fact that elevated￾temperature strength has a relative meaning if a material ex￾hibits rate dependency: the strength simply depends on which test rate one chooses (see Fig. 2). Therefore, use of at least two test rates (high and low) is recommended to better char￾acterize high-temperature strength behavior of a composite material. 4. Conclusions Elevated-temperature ultimate tensile strength of five different continuous fiber-reinforced ceramic composites, including Nicalon/BSAS (2D), Hi-Nicalon/BSAS (2D), SiCf/MAS (1D), SiCf/SiC (2D woven: enhanced), and Cf/SiC (2D woven: standard and enhanced), exhibited a strong dependency on test rate, consistent with the behav￾ior observed in other CMCs as well as in many advanced