March 2002 Efects of Thermal Aging on the Mechanical Properties of a Porous-Matrix Ceramic Composite 150 blunting cracks that emanate from fiber breaks. Evidently the extent of the changes in the matrix strength following aging are insufficient to noticeably alter the fiber bundle strength. This suggests the existence of a rather broad maximum or plateau in t fiber bundle strength, wherein the properties are insensitive to the matrix properties(top curve in Fig 9). The much larger changes in the extent of matrix sintering in aluminosilicate matrix CFCCs lead to extremely brittle fracture characteristics and low strength illustrated by the decreasing part of the top curve in Fig. 9 It is anticipated that there are two other behavioral regimes with 50 regard to the o°90°and±45° tensile response. In the former orientation, the strength is expected to increase with increasing matrix strength in the regime where the matrix strength is very low. This expectation is based in part on analogous behavior of dense-matrix CFCCs with weak interphases. Notably, when the interfacial sliding 00400600800100012001400 stress, To, is sufficiently low to ensure global load sharing character Aging Temperature(C) istics, the bundle strength is predicted to scale with T. (m+D)where m is the Weibull modulus of the fibers. For typical values of m (4-10), the exponent 1/(m+ 1)0. 1-0. 2, and thus the sensitivity to To is weak. Indeed, over the typical range of m values, twofold 凵20 may not be detected readily among the scatter in the experimental measurements. Furthermore, once To becomes sufficiently large, the load-sharing characteristics among broken fibers become more local- ●From pt ized and the fiber bundle strength then gradually diminishes. 0 nsion analogy, the fiber bundle strength in the porous-matrix composites is expected to follow a similar dependence on the matrix shear strength. initially increasing and then decreasing with increasing matrix strength. Furthermore, the seeming independence of the composite strength on the matrix strength in the present experiments may be a 5 consequence of a similarly weak dependence in composite propertie on matrix properties in the regime probed by these experiments coupled with some scatter in the experimental measurements. Con- sequently, it remains to be established more definitively whether the 0200400600800100012001400 present experimental results do indeed reside along the broad plateau Aging Temperature(C) in Fig.9 or whether they are in one of the adjacent regimes in which the strength is either gradually increasing or gradually decreasing with Fig. 8. Effects of thermal aging on(a) the matrix hardness and(b) the the degree of matrix sintering. matrix Young's modulus For similar reasons, a maximum in the +45 composite strength is also expected. That is, once the matrix strength achieves a sufficiently high value, the inelastic straining capabilities of the are diminished and the becomes increasingly sensitive to the presence of flaws, introduce Mullite/Alumina either during processing or as a consequence of mechanical 090° loading. In this regime, the strength is expected to be low and exhibit large variability. These hypotheses require further theoret 5品 From a technological viewpoint, the retention of the fiber- dominated properties after the 1200C aging treatment is partici Aluminosilicate larly encouraging. It reaffirms that, with the selection of a matrix Matrix with a stable pore structure coupled with a stable oxide fiber, these composites have the potential for long-term durability at temper atures to which they will be subjected in the targeted applications Otherwise, if the matrix is susceptible to appreciable sintering, the composites are prone to severe property degradation. Moreover, it ±45° is anticipated that further enhancement in the stability of the pore structure could be achieved through modifications to the matrix formulation, e.g., increasing the ratio of mullite to alumina, to ensure that the properties are retained for even longer time periods Degree of Matrix Sintering than those probed by the present experiments. However, these improvements may come at the expense of reduced off- Fig.9. Schematic showing the trends in the +45 and o%/90 tensile properties. This is the subject of current investigation strengths with the degree of matrix sintering Append llustrated by the lower(solid) curve in Fig. 9. In the 0 /90 orientation, the properties are largely fiber-dominated, with onl Youngs modulus of the matrix was inferred from the measured small contributions coming from the matrix. For instance, because Youngs moduli of the composites, measured both in the 0/90 of the extremely low value of matrix modulus in the as-processed and +45%orientations, using classical laminate theory. For this composite, even a twofold increase in this modulus following purpose, the composite is treated as a balanced, symmetric lay-up aging has only a small effect on the composite modulus. The main of unidirectional fiber composite plies. The calculation proceeds in role of the matrix in this orientation is to act as a medium for two steps. In the first, the properties of the laminate are expressedillustrated by the lower (solid) curve in Fig. 9. In the 0°/90° orientation, the properties are largely fiber-dominated, with only small contributions coming from the matrix. For instance, because of the extremely low value of matrix modulus in the as-processed composite, even a twofold increase in this modulus following aging has only a small effect on the composite modulus. The main role of the matrix in this orientation is to act as a medium for blunting cracks that emanate from fiber breaks. Evidently the extent of the changes in the matrix strength following aging are insufficient to noticeably alter the fiber bundle strength. This suggests the existence of a rather broad maximum or plateau in the fiber bundle strength, wherein the properties are insensitive to the matrix properties (top curve in Fig. 9). The much larger changes in the extent of matrix sintering in aluminosilicate matrix CFCCs lead to extremely brittle fracture characteristics and low strength, illustrated by the decreasing part of the top curve in Fig. 9. It is anticipated that there are two other behavioral regimes with regard to the 0°/90° and 45° tensile response. In the former orientation, the strength is expected to increase with increasing matrix strength in the regime where the matrix strength is very low. This expectation is based in part on analogous behavior of dense-matrix CFCCs with weak interphases. Notably, when the interfacial sliding stress, 0, is sufficiently low to ensure global load sharing character￾istics, the bundle strength is predicted to scale with 0 1/(m 1) where m is the Weibull modulus of the fibers.9 For typical values of m (4–10), the exponent 1/(m 1) 0.1–0.2, and thus the sensitivity to 0 is weak. Indeed, over the typical range of m values, twofold changes in 0 only alter the bundle strength by 10%, an effect which may not be detected readily among the scatter in the experimental measurements. Furthermore, once 0 becomes sufficiently large, the load-sharing characteristics among broken fibers become more local￾ized and the fiber bundle strength then gradually diminishes.10 By analogy, the fiber bundle strength in the porous-matrix composites is expected to follow a similar dependence on the matrix shear strength, initially increasing and then decreasing with increasing matrix strength. Furthermore, the seeming independence of the composite strength on the matrix strength in the present experiments may be a consequence of a similarly weak dependence in composite properties on matrix properties in the regime probed by these experiments, coupled with some scatter in the experimental measurements. Con￾sequently, it remains to be established more definitively whether the present experimental results do indeed reside along the broad plateau in Fig. 9 or whether they are in one of the adjacent regimes in which the strength is either gradually increasing or gradually decreasing with the degree of matrix sintering. For similar reasons, a maximum in the 45° composite strength is also expected. That is, once the matrix strength achieves a sufficiently high value, the inelastic straining capabilities of the composite are diminished and the strength of the composite becomes increasingly sensitive to the presence of flaws, introduced either during processing or as a consequence of mechanical loading. In this regime, the strength is expected to be low and exhibit large variability. These hypotheses require further theoret￾ical and experimental investigation. From a technological viewpoint, the retention of the fiber￾dominated properties after the 1200°C aging treatment is particu￾larly encouraging. It reaffirms that, with the selection of a matrix with a stable pore structure coupled with a stable oxide fiber, these composites have the potential for long-term durability at temper￾atures to which they will be subjected in the targeted applications. Otherwise, if the matrix is susceptible to appreciable sintering, the composites are prone to severe property degradation. Moreover, it is anticipated that further enhancement in the stability of the pore structure could be achieved through modifications to the matrix formulation, e.g., increasing the ratio of mullite to alumina, to ensure that the properties are retained for even longer time periods than those probed by the present experiments. However, these improvements may come at the expense of reduced off-axis properties. This is the subject of current investigation. Appendix Young’s modulus of the matrix was inferred from the measured Young’s moduli of the composites, measured both in the 0°/90° and 45° orientations, using classical laminate theory. For this purpose, the composite is treated as a balanced, symmetric lay-up of unidirectional fiber composite plies. The calculation proceeds in two steps. In the first, the properties of the laminate are expressed Fig. 8. Effects of thermal aging on (a) the matrix hardness and (b) the matrix Young’s modulus. Fig. 9. Schematic showing the trends in the 45° and 0°/90° tensile strengths with the degree of matrix sintering. March 2002 Effects of Thermal Aging on the Mechanical Properties of a Porous-Matrix Ceramic Composite 601