3192 Journal of the American Ceramic Society-Morscher et al. VoL. 90. No. 10 Table IV. Ultimate Strength Properties of Composites in the (5) Implications for Architecture Design of MI SiC/SiC Literature Tested in Multiple Directions From the observations of this study, one can summarize several Composite architecture Orientation ffas(GPa)(MPa) onstituent and architectural guidelines that can be applied to Matrix-dominated composites uture designs of components fabricated with ceramic compos- Nic-MAS 0/90 laminate 00.185130385 ites in general and nonoxide Syl-iBN MI SiC/SiC composites in 300.185120147 articular where high off-axis strengths are required. It is as- 450.37110157 sumed that the design goals will be to achieve as high a matrix Nic-CVI23 2 D woven0/9000.2220190 racking stress as possible as well as a high UTS along the prin- 200.2210170 cipal stress directions within the components 450.4210170 First, the matrix constituent should display a high stiffness Fiber-dominated composites and high strain capability by utilizing a high-modulus compo- C- epoxy 00.25NA461 sition, such as SiC, and a fabrication approach that results in as 150.25NA274 low a porosity as possible, such as MI SiC. For the Sic/sic composites of this study, the high-modulus low-porosity MI 300.25NA143 450.5NA12 matrix allows the composite elastic modulus to be fairly inde pendent of architecture and in-plane testing direction. Further oxy braid0/±45 00.1540.8417 more, as is the case with the MI process, the 0/±60 axial 11 31.6318 fabrication process should(if possible)result in a residual com- 0/±60hoop300.4649.9400 pressive stress on the matrix critical flaws or weak portion of the 0/±45hoop 5 0.42 19.8 165 fiber architecture after final composite fabrication(see Table D) CVI, chemical vapor infiltration: UTS. ultimate tensile strength. In contrast to residual stresses caused by thermal expansion mismatch between the fiber and the matrix the residual stress of the mi process appears to be independent of temperature up to at least 81 45). This is an important design consideration for composite Second, the fiber constituent should be as strong as possible and architecture selection when appreciable off-axis loading ap n its as-produced condition and should retain a high fraction of plications are pursued this strength after composite fabrication. The on-axis UTS val Note that the Nic/CVI SiC ,composites had the highest ues for the Sylramic-iBN MI composites of this study(see Table elative off-axis strength. As discussed above, the eq. (3)anal D)and other studieswhen normalized by the fiber volume ysis does not account for such material properties as interfacial fraction are the highest( 2400 MPa)displayed to date by ding stress and fiber bending stiffness, which will affect the any woven Sic-based fiber in an MI composite. For off-axis mechanics of fiber fracture when aligned at an angle within a UTS, high fiber strength is also important for obtaining high transverse matrix crack. The absolute off-axis strengths of the fiber load-carrying ability during bending within matrix cracks. iC/CVI SiC composites were similar to the HN/MI and In addition, the fiber should have as high a modulus as possible Sylramic/MI and less than the Syl-iBN/MI composites. Also in order to shift composite loads away from the matrix flaws or Culto of the Nic/CVI SiC composites were poor, indicating a low rom weak portions of the fiber architecture and onto the load- n sI ter processing) fiber strength in comparison with the bearing fiber. Also, in combination with the interphase coatin othe es, possibly resulting in higher than ex- the fiber surface conditions should be such as to provide hig pected normalized off-axis fiber strength results. Another pos- interfacial shear to inhibit large crack-bridging fiber lengths that bility is that the beneficial effects of a lower fiber modulus and will both statistically reduce fiber strength within the cracks interfacial sliding stress are becoming evident for the nicalon auge length effect)and allow more fiber bending for off-axis composites. Nevertheless, it is suggested that the two lines in ading. Due to the relatively high surface roughness of the Fig. Il represent, albeit crudely, off-axis UTS for matrix-dor Iramic-ibn fiber and due to the stiffness of the combined inated and fiber-dominated CMC. These relationships can be iBN/BN interphase coating, interfacial shear strengths in Sylra used by designers as a"rule-of-thumb"when considering the mic- iBN/BN/SiC systems are approximately 70 Mpa, much type of composite and fiber architecture to be used in a com- than observed with other smoother fiber types and or ponent with off-axis stresses. carbon interfacial coatings when processed in the same manner and with similar thickness 13 Third, the type of fiber architecture and orientation of fiber must be judiciously selected in relation to the directions and magnitudes of the principal stresses within the CMC compo- Matrix Dominated SiC/SiC CMC nent. a primary guideline shown in this study is to effective fiber volume fractions as high as possible in these pi a 0. pal stress directions, both to reduce stress on matrix flaws and/ r weak 90 minicomposites and to increase the stress and strain for ultimate failure. However as shown here. the conventional 0.61 Dominated Nic/MAS [22] approach of putting the primary fiber axes directly along the 0.5 principal stress directions may not be required. For example for Nic/CVI the 0/90 panel aligned at 45 to the primary fiber axes, the ae Epoxy PW[181 onset cracking stress increased from 190 to 220 MPa due in part to the removal of minicomposites being perpendicular to the boxy braid [191 oading direction. But UTS values decreased significantly from Al203/mullite[20] 410 to 242 MPa because of fiber strength loss within open ma C/C& C/CVI SiC [21] rix cracks. However, for the braided panel aligned at a smaller 3 to the primary fiber axes, onset stresses remained high at 220 MPa and UTS only degraded to 350 MPa. Generally, it ngle of loaded-fibers off axis, degrees is thought that high us values are desirable for co Fig. 11. Effect of testing direction and composite type on the relative materials, but for nonoxide ceramic composites, such as MI SiC/ trength retention of the fibers for the composite data in Tables l, l SiC, structural life degrades in a complex manner above the cracking stress due to environment ingress through the open451). This is an important design consideration for composite and architecture selection when appreciable off-axis loading ap￾plications are pursued. Note that the Nic/CVI SiC2,21,23 composites had the highest relative off-axis strength. As discussed above, the Eq. (3) anal￾ysis does not account for such material properties as interfacial sliding stress and fiber bending stiffness, which will affect the mechanics of fiber fracture when aligned at an angle within a transverse matrix crack. The absolute off-axis strengths of the NiC/CVI SiC composites were similar to the HN/MI and Sylramic/MI and less than the Syl-iBN/MI composites. Also, sult0 of the Nic/CVI SiC composites were poor, indicating a low in situ (after processing) fiber strength in comparison with the other SiC/SiC composites, possibly resulting in higher than ex￾pected normalized off-axis fiber strength results. Another pos￾sibility is that the beneficial effects of a lower fiber modulus and interfacial sliding stress are becoming evident for the Nicalon composites. Nevertheless, it is suggested that the two lines in Fig. 11 represent, albeit crudely, off-axis UTS for matrix-dom￾inated and fiber-dominated CMC. These relationships can be used by designers as a ‘‘rule-of-thumb’’ when considering the type of composite and fiber architecture to be used in a com￾ponent with off-axis stresses. (5) Implications for Architecture Design of MI SiC/SiC Components From the observations of this study, one can summarize several constituent and architectural guidelines that can be applied to future designs of components fabricated with ceramic compos￾ites in general and nonoxide Syl-iBN MI SiC/SiC composites in particular where high off-axis strengths are required. It is as￾sumed that the design goals will be to achieve as high a matrix cracking stress as possible as well as a high UTS along the prin￾cipal stress directions within the components. First, the matrix constituent should display a high stiffness and high strain capability by utilizing a high-modulus compo￾sition, such as SiC, and a fabrication approach that results in as low a porosity as possible, such as MI SiC. For the SiC/SiC composites of this study, the high-modulus low-porosity MI matrix allows the composite elastic modulus to be fairly inde￾pendent of architecture and in-plane testing direction. Further￾more, as is the case with the MI process, the composite fabrication process should (if possible) result in a residual com￾pressive stress on the matrix critical flaws or weak portion of the fiber architecture after final composite fabrication (see Table I). In contrast to residual stresses caused by thermal expansion mismatch between the fiber and the matrix, the residual stress of the MI process appears to be independent of temperature up to at least 8151C.25 Second, the fiber constituent should be as strong as possible in its as-produced condition and should retain a high fraction of this strength after composite fabrication. The on-axis UTS val￾ues for the Sylramic-iBN MI composites of this study (see Table I) and other studies7 when normalized by the fiber volume fraction are the highest (B2400 MPa) displayed to date by any woven SiC-based fiber in an MI composite. For off-axis UTS, high fiber strength is also important for obtaining high fiber load-carrying ability during bending within matrix cracks. In addition, the fiber should have as high a modulus as possible in order to shift composite loads away from the matrix flaws or from weak portions of the fiber architecture and onto the load￾bearing fiber. Also, in combination with the interphase coating, the fiber surface conditions should be such as to provide high interfacial shear to inhibit large crack-bridging fiber lengths that will both statistically reduce fiber strength within the cracks (gauge length effect) and allow more fiber bending for off-axis loading. Due to the relatively high surface roughness of the Sylramic-iBN fiber and due to the stiffness of the combined iBN/BN interphase coating, interfacial shear strengths in Sylra￾mic-iBN/BN/SiC systems are approximately 70 Mpa,7 much higher than observed with other smoother fiber types and/or carbon interfacial coatings when processed in the same manner and with similar thickness.13,26 Third, the type of fiber architecture and orientation of fibers must be judiciously selected in relation to the directions and magnitudes of the principal stresses within the CMC compo￾nent. A primary guideline shown in this study is to achieve effective fiber volume fractions as high as possible in these prin￾cipal stress directions, both to reduce stress on matrix flaws and/ or weak 901 minicomposites and to increase the stress and strain for ultimate failure. However, as shown here, the conventional approach of putting the primary fiber axes directly along the principal stress directions may not be required. For example for the 0/90 panel aligned at 451 to the primary fiber axes, the AE onset cracking stress increased from 190 to 220 MPa due in part to the removal of minicomposites being perpendicular to the loading direction. But UTS values decreased significantly from 410 to 242 MPa because of fiber strength loss within open ma￾trix cracks. However, for the braided panel aligned at a smaller 231 to the primary fiber axes, onset stresses remained high at B220 MPa and UTS only degraded to B350 MPa. Generally, it is thought that high UTS values are desirable for composite materials, but for nonoxide ceramic composites, such as MI SiC/ SiC, structural life degrades in a complex manner above the cracking stress due to environment ingress through the open Table IV. Ultimate Strength Properties of Composites in the Literature Tested in Multiple Directions Composite Fiber architecture Orientation f0 or f
off-axis E (GPa) UTS (MPa) Matrix-dominated composites Nic-MAS22 0/90 laminate 0 0.185 130 385 30 0.185 120 147 45 0.37 110 157 Nic-CVI23 2D woven 0/90 0 0.2 220 190 20 0.2 210 170 45 0.4 210 170 Fiber-dominated composites C - epoxy PW18 Plain Weave 0 0.25 NA 461 15 0.25 NA 274 30 0.25 NA 143 45 0.5 NA 127 C - epoxy braid19 0/745 axial 0 0.15 40.8 417 0/760 axial 0 0.11 31.6 318 0/760 hoop 30 0.46 49.9 400 0/745 hoop 45 0.42 19.8 165 CVI, chemical vapor infiltration; UTS, ultimate tensile strength. 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0 10 20 30 40 50 Angle of loaded-fibers off axis, degrees Norm. Effective Fiber Strength Al2O3/mullite [20] Nic/C [2] C/epoxy PW [18] Syl-iBN/MI (this study) Syl-iBN/MI Braid (this study) Syl/MI [22] Nic/MAS [22] Nic/CAS [2] HN/MI [22] Nic/CVI [2] C/epoxy braid [19] Nic/CVI [21] C/C & C/CVI SiC [21] Fiber Axis Loading Axis angle Fiber Dominated CMC Matrix Dominated SiC/SiC CMC Nic/CVI [23] Fig. 11. Effect of testing direction and composite type on the relative strength retention of the fibers for the composite data in Tables I, III, and IV. 3192 Journal of the American Ceramic Society—Morscher et al. Vol. 90, No. 10