unacceptable. Yet the fabrication of such TMCs has not yet been possible because of reactivity problems during consolidation. In CMCs the same issues dictate preferences for SiC/SiC or mullite/Al2O3 over SiC/Al2O3. Thermal conductivity is important for materials with limited ductility, because low k elevates both the transient thermal strains and the temperatures reached in operation. High-k materials are thus preferred, such as SiC.
DESIGN ISSUES
Inelastic Strain and Toughness
Inelastic deformation with attendant “ductility” enables stresses to be redistributed and thereby diminishes stress concentrations [8,9,15]. This peak stress reduction is particularly crucial to notch performance. In monolithic alloys, this effect is achieved through plasticity [16]. In CMCs, multiple matrix cracking with internal friction is the responsible mechanism [8,9]. In TMCs and PMCs, matrix plasticity is predominant [17]. Moreover, materials exhibiting inelastic deformation are amenable to well-established design strategies [18,19]. That is, an inelastic constitutive law is devised, preferably mechanism-based. Stresses are calculated by numerical methods, and a failure criterion implemented. The peak tensile stress is equated to a critical stress, such as the ultimate tensile strength (UTS) measured on test coupons.
The notch strength of a material is a good engineering indicator of its stress redistribution capacity; a notch sensitivity index establishes the relative importance attached to the design of holes and other strain concentrators [20]. There is a well-established methodology for assessing the notch performance of conventional materials. At the experimental level, it involves instrumented Charpy tests. At the mechanism level, it entails the use of finite element methods with plasticity to calculate peak stresses and to equate these to the critical stress [16]. Moreover, since the plastic zone around the notch is similar to that around a crack, the notch behavior scales with the fracture toughness. This connection allows the toughness to be used as materials selection input (see Figure 1).
For composites, the toughness is nonunique, because of large-scale inelastic deformation effects, such as bridging (LSB) and yielding (LSY) [19,21,22]. The notch sensitivity is then a more robust and useful measure of performance. Among composite materials the trend in notch sensitivity, from least to greatest, is surprisingly: CMC > PMC > TMC (Figure 3). The latter are quite notch-sensitive, because of plastic localization [23]. By contrast, CMCs combine a large inelastic zone with stochastic scaling effects to eliminate stress concentrations [9] (see Figure 3).
In ceramics and intermetallics, large stress concentrations persist because inelastic deformation mechanisms do not operate on a sufficiently macroscopic scale. They only occur in the 1- to 100-μm range around cracks, where they cause toughening [18, 24-26]. Consequently, notches cause a dramatic reduction in strength (see Figure 3). Moreover, the UTS is scale (size) dependent because of weakestlink statistics, and the distributions are often extreme value [27,28]. These problems compromise robust design practices [18,26].
Anisotropy
When reinforcements are used to provide a performance advantage, there are corresponding detriments associated with the anisotropy. The issues differ for PMCs/CMCs compared with TMCs. For the former, out-of-plane stresses can cause delamination and interlaminar shear cracking [8, 29-31]. These phenomena are dominated by the matrix properties, subject to manufacturing flaws and to weakest-link scaling. Incomplete understanding of the effects has resulted in the use of “knock-down factors” in