December 2007 Communications of the American Ceramic Society 4053 10000 (strong 1000 300 Composit i Composite Oxidation temperature (C Fig 4. Static fatigue lifetimes for Nicalon"/CAS tested in air at 4500 Fig. 5. Environmental embrittlement 525, and 600C. Arrows indicate samples that survived after 1000 h dation of Nicalon"/CAS, demonstrat Note that all testing is conducted above the unction of oxidation temperature and aterial, which was determined to be 116+4 MPa in three-point bend. failure: .. brittle failure. Note that the ided into regions of high, medium, and low strength. oughness of Nicalon fibers are typically on the order degradation noted in prior studies is readily apparent, and dem- Powell et al. presented a shear-lag model that de- onstrates the issues relating to interphase stability quite clearly he residual thermal stress state in Nicalon/ CAS after While Nicalon"/ CAS can now be viewed primarily as a model It was noted in that study that surface effects are sig- material for studies such as these, the present observations of such that the residual stress after fabrication and cool- low temperature degradation have significant implications for om temperature will be sufficient to result in a surface more advanced generations of non-oxide ceramic matrix com- egion equivalent to at least one-or two-fiber diam- posites. For example, the recent approach of developing com- eters in depth. Clearly, in the present case with pa emoval posites with multiple Sic/c interphase layers, where the carbon of the carbon layer this effect is exacerbated, and there will b yers are designed to be very thin to allow rapid sealing, may complete debonding for the relatively small sample size of a ot prevent degradation when flexure bar tion at low temperatures in the absence of a protective coating. It is apparent that the interfacial instabilities have a signifi Even when such a coating is present, the demands on it will cant effect upon the c ite macro-mechanical behavior. Re- gnificant as it must protect the material at wide extremes of moval of the carbon interface results in clamping of the fiber by temperature, in environments where stress is likely to be applied the surrounding matrix. In this instance both the interfacial At high temperatures some degree of coating compliance is like- debond energy and the frictional sliding stress are increased ly to be a typical feature however this may not be the case at substantially. The result of this interface modification is that nterfacial debonding criteria, where the interface fracture latively low temperatures, where the simple glass-sealing tech- iques described earlier will be ineffective. energy is significantly lower than the fiber fracture energy (i.e. Gi<0.25Gr, where Gr is the fiber fracture energy), are no longer atisfied. Consequently, debonding and fiber slie ng nt occur. and a transition to brittle failure is observed Figure 4 demonstrates the effect of combined applied stress Nicalon"/CAS composites have been exposed to oxidizing heat d oxidation temperature upon the static fatig treatments between 375 and 600 C. for both unstressed and CAS/Nicalon"composites. It is clear that, for a static fatigue-loaded conditions. While strength is essentially re- tress,there is a strong correlation between the composite life- tained after unstressed oxidation at the lowest temperatures (i. e, time and the oxidation temperature, in accordance with the un- 375 and 450C), a transition to brittle failure occurs with in- loaded oxidation response. In each of these cases the static time (i.e. between 100 and 1000 h at 375( applied stress exceeds the microcracking stress, omc, of the com- and between 10 and 100 h at 450C). At 525 and 600oC posite in flexure(-116.0+4.0 MPa); consequently the compos- strength decreases with increasing oxidation time, and the fail- ite matrix is microcracked for the entire test duration un ure mode is brittle for all examples. These changes in mechanical failure. In this instance, the presence of microcracking in the behavior arise from removal of the carbon-based fiber /matrix matrix significantly increases the number of paths for oxygen terphase, via"pipe-line oxidation ngress to the fiber-matrix interface. Generally similar observa creased friction present at the fiber/matrix interface(arisin tions to these were made by Yasmin and Bowen, for cyclic due to the residual stress state in this particular system, with fatigues loading at 800C. At room temperature, they noted that the higher CTE matrix effectively"clamping-down""onto the II composite was 200 MPa(defined as successful completion under static fatigue loading in oxidizing environments, with life- of 10 cycles). However, the fatigue limit was reduced signifi- times significantly reduced under increasing temperature when to 100 MPa at 800C, which corresponds with an ples are loaded above the materials proportional limit, ome stress just below the microcracking stress of this mate when matrix microcracking will occur. Based upon the current r applied cyclic stresses above omc, the fatigue lifetimes study and previously published data, an environmental embrit were dramatically reduced at the elevated test temperature. tlement failure mechanism map has been developed for Nica- Based upon the data outlined in Fig. 1, in combination with lon"/CAS, which highlights the region of intermediate that presented in previous work for the oxidation behavior of temperature embrittlement as a function of ex /CAs at higher temperature ture and time. The present work has demonstrated that oxida to develop a failure mechanism map for Nicalon"/CAS after tion degradation can occur in CMCs with a carbon-based oxidation exposure (Fig. 5). The intermediate temperature terphase, even at temperatures as low as 375%C, which maysurface roughness of Nicalont fibers are typically on the order of 5 nm.33 Powell et al. 34 presented a shear-lag model that de￾scribes the residual thermal stress state in Nicalont/CAS after synthesis. It was noted in that study that surface effects are sig￾nificant, such that the residual stress after fabrication and cool￾ing to room temperature will be sufficient to result in a surface debonded region equivalent to at least one- or two-fiber diam￾eters in depth. Clearly, in the present case with partial removal of the carbon layer this effect is exacerbated, and there will be near complete debonding for the relatively small sample size of a flexure bar. It is apparent that the interfacial instabilities have a signifi- cant effect upon the composite macro-mechanical behavior. Re￾moval of the carbon interface results in clamping of the fiber by the surrounding matrix. In this instance, both the interfacial debond energy and the frictional sliding stress are increased substantially. The result of this interface modification is that interfacial debonding criteria, where the interface fracture energy is significantly lower than the fiber fracture energy (i.e., Gio0.25Gf, where Gf is the fiber fracture energy),35 are no longer satisfied. Consequently, debonding and fiber sliding no longer occur, and a transition to brittle failure is observed. Figure 4 demonstrates the effect of combined applied stress and oxidation temperature upon the static fatigue lifetime of CAS/Nicalont composites. It is clear that, for a given applied stress, there is a strong correlation between the composite life￾time and the oxidation temperature, in accordance with the un￾loaded oxidation response. In each of these cases the static applied stress exceeds the microcracking stress, smc, of the com￾posite in flexure (B116.074.0 MPa); consequently the compos￾ite matrix is microcracked for the entire test duration until failure. In this instance, the presence of microcracking in the matrix significantly increases the number of paths for oxygen ingress to the fiber–matrix interface. Generally similar observa￾tions to these were made by Yasmin and Bowen,22 for cyclic fatigues loading at 8001C. At room temperature, they noted that the fatigue limit for an identical cross-ply Nicalont/CAS Type￾II composite was B200 MPa (defined as successful completion of 106 cycles). However, the fatigue limit was reduced signifi- cantly to B100 MPa at 8001C, which corresponds with an applied stress just below the microcracking stress of this mate￾rial. For applied cyclic stresses above smc, the fatigue lifetimes were dramatically reduced at the elevated test temperature. Based upon the data outlined in Fig. 1, in combination with that presented in previous work for the oxidation behavior of Nicalont/CAS at higher temperatures,21,26,27 it is possible to develop a failure mechanism map for Nicalont/CAS after oxidation exposure (Fig. 5). The intermediate temperature degradation noted in prior studies is readily apparent, and dem￾onstrates the issues relating to interphase stability quite clearly. While Nicalont/CAS can now be viewed primarily as a model material for studies such as these, the present observations of low temperature degradation have significant implications for more advanced generations of non-oxide ceramic matrix com￾posites. For example, the recent approach of developing com￾posites with multiple SiC/C interphase layers, where the carbon layers are designed to be very thin to allow rapid sealing,36 may not prevent degradation when the material is exposed to oxida￾tion at low temperatures in the absence of a protective coating. Even when such a coating is present, the demands on it will be significant as it must protect the material at wide extremes of temperature, in environments where stress is likely to be applied. At high temperatures some degree of coating compliance is like￾ly to be a typical feature, however this may not be the case at relatively low temperatures, where the simple glass-sealing tech￾niques described earlier will be ineffective.21 IV. Conclusions Nicalont/CAS composites have been exposed to oxidizing heat￾treatments between 3751 and 6001C, for both unstressed and static fatigue-loaded conditions. While strength is essentially re￾tained after unstressed oxidation at the lowest temperatures (i.e., 3751 and 4501C), a transition to brittle failure occurs with in￾creasing exposure time (i.e., between 100 and 1000 h at 3751C, and between 10 and 100 h at 4501C). At 5251 and 6001C, strength decreases with increasing oxidation time, and the fail￾ure mode is brittle for all examples. These changes in mechanical behavior arise from removal of the carbon-based fiber/matrix interphase, via ‘‘pipe-line’’ oxidation, and the subsequently in￾creased friction present at the fiber/matrix interface (arising due to the residual stress state in this particular system, with the higher CTE matrix effectively ‘‘clamping-down’’ onto the Nicalont fiber). A similar embittlement mechanism is apparent under static fatigue loading in oxidizing environments, with life￾times significantly reduced under increasing temperature when samples are loaded above the materials proportional limit, smc, when matrix microcracking will occur. Based upon the current study and previously published data, an environmental embrit￾tlement failure mechanism map has been developed for Nica￾lont/CAS, which highlights the region of intermediate temperature embrittlement as a function of exposure tempera￾ture and time. The present work has demonstrated that oxida￾tion degradation can occur in CMCs with a carbon-based interphase, even at temperatures as low as 3751C, which may 1 10 100 1000 0 100 200 300 400 500 600 450 525 600 0.001 0.01 0.1 Lifetime (hrs) Applied stress (MPa) Fig. 4. Static fatigue lifetimes for Nicalont/CAS tested in air at 4501, 5251, and 6001C. Arrows indicate samples that survived after 1000 h. Note that all testing is conducted above the proportional limit for this material, which was determined to be 11674 MPa in three-point bend. 0.1 1 10 100 1000 10000 0 200 400 600 800 1000 1200 Composite Brittle (strong) Brittle (medium strength) Brittle (weak) Composite Oxidation temperature (°C) Oxidation time (hrs) Fig. 5. Environmental embrittlement failure mechanism map for oxi￾dation of Nicalont/CAS, demonstrating the failure mode observed as a function of oxidation temperature and exposure duration. m, composite failure; , brittle failure. Note that the brittle failure regions are subdi￾vided into regions of high, medium, and low strength. December 2007 Communications of the American Ceramic Society 4053