July 1997 High-Temperature Transverse Fracture Toughness 1819 breaking the fibers bridging the crack surface without pulling from the micrograph of the crack in Fig. 14. The crack is tightly them out of the matrix. This process is test time dependent, i.e., closed near the top side. Crack opening near the top side is no e slower the fracture the more effective the oxidation process more than a few micrometers even after handling between The increased interfacial bond strength results in a planar frac- test and the mounting of the sample(which was cut from the ture surface without noticeable fiber pullout, possibly lowering cracked DT specimen). The crack opening, which was prob- e transverse fracture toughness of the composite. However, bly much smaller than that during and after the failure of the the decrease in the transverse composite fracture toughness at DT specimen, might not have allowed air to penetrate through 800 and 1000@C is believed to be caused mainly by lower to the top side matrix fracture toughness at high temperatures relative to that The possible occurrence of high-temperature stress corrosion at room temperature, As much as 50% reduction has been crack growth was also explored by performing two dt tests observed in CAS matrix toughness at 1000 C 45 holding the load(P)constant to give Gr of about 20 and 25 J/m Microdebonding tests were also performed along the crack for about 3 h at 800C(G, is determined from Eq (I)replacing n cross sections of dt specimens fractured at temperatures G, and p。 by gr and p, ely). Examination of the from 20 to 1000C(1-2 h long tests ) One such cross section tested specimens under the light microscope did not reveal any fibers bridging the crack surface crack growth Cracks grow only when Gy is close to the critical aused the two arms of the dt specimens to remain intact even value, GIc, indicating that no significant stress corrosion crack after the failure. The crack is more opened on the bottom side growth parallel to the fibers is observed in the Nicalon/CAS-II of the specimen(with a maximum crack opening of about 30 composite at lower GI, at least for the experiments carried out um)and less opened on the top side due to the nature of the in this study. However, Spearing et al. 46 reports significant loading as discussed previously. Only the yery close stress corrosion matrix cracking on the same material system (within less than a fiber diameter distance)to the crack were under longitudinal loading conditions, at str required to develop matrix cracks in short-duration, monotonic Fiber-matrix bond strength distributions along the crack be. loading tests, even at room temperature but for much longer tween the two(top and bottom) surfaces of the DT specimens stress application times(as long as 278 h) Some environmental stress cracking might be expected by in Fig. 15. No significant change in fiber-matrix bond strengt oxidative removal of the interfacial carbon layer at high tem- was observed from 20 to 600C. The most significant increase peratures even for the short exposure times of this study. How- in the interfacial strength was observed at 1000C. However, a ever, this seems to be prevented by subsequent replacement of critical observation is that oxidation was effective only halfway the oxidized carbon layer with a new interfacial phase closing across the fracture surface from the bottom side while the bone the gap between fiber and matrix, and sealing the oxidation strengths on the other half were similar to those of control process off. 14 4174 opposites(except for the oxidized fibers very close to the top surface exposed to the environment). The results for the frac ture at 800%C were similar in nature but with lower increases in V. Summary and Conclusions Dt test technique was successfully utilized to determine The explanation for this selective oxidation may be drawn fracture toughness(Grc) for cracks parallel to the fibers in a c10 top bottom Dislance from Top Surfoce/Thickness fig. 15. Interfacial bond strength distribution along the crack from top to bottom face for DT specimens fractured(1-2 h long tests)at differentJuly 1997 High-Temperature Transverse Fracture Toughness 1819 breaking the fibers bridging the crack surface without pulling them out of the matrix. This process is test time dependent, i.e., the slower the fracture the more effective the oxidation process. The increased interfacial bond strength results in a planar frac￾ture surface without noticeable fiber pullout, possibly lowering the transverse fracture toughness of the composite. However, the decrease in the transverse composite fracture toughness at 800" and 1OOO"C is believed to be caused mainly by lower matrix fracture toughness at high temperatures relative to that at room temperature. As much as 50% reduction has been observed in CAS matrix toughness at 1000°C."5 Microdebonding tests were also performed along the crack on cross sections of DT specimens fractured at temperatures from 20" to 1OOO"C (1-2 h long tests). One such cross section is shown in Fig. 14. The fibers bridging the crack surface caused the two arms of the DT specimens to remain intact even after the failure. The crack is more opened on the bottom side of the specimen (with a maximum crack opening of about 30 km) and less opened on the top side due to the nature of the loading as discussed previously. Only the fibers very close (within less than a fiber diameter distance) to the crack were tested. Fiber-matrix bond strength distributions along the crack be￾tween the two (top and bottom) surfaces of the DT specimens tested at temperatures of 20", 600", 800", and 1OOO"C are given in Fig. 15. No significant change in fiber-matrix bond strength was observed from 20" to 600°C. The most significant increase in the interfacial strength was observed at 1OOO"C. However, a critical observation is that oxidation was effective only halfway across the fracture surface from the bottom side, while the bond strengths on the other half were similar to those of control composites (except for the oxidized fibers very close to the top surface exposed to the environment). The results for the frac￾ture at 800°C were similar in nature but with lower increases in bond strength. The explanation for this selective oxidation may be drawn from the micrograph of the crack in Fig. 14. The crack is tightly closed near the top side. Crack opening near the top side is not more than a few micrometers even after handling between the test and the mounting of the sample (which was cut from the cracked DT specimen). The crack opening, which was prob￾ably much smaller than that during and after the failure of the DT specimen, might not have allowed air to penetrate through to the top side. The possible occurrence of high-temperature stress corrosion crack growth was also explored by performing two DT tests holding the load (P) constant to give GI of about 20 and 25 J/mz for about 3 h at 800°C (GI is determined from Eq. (1) replacing GI, and P, by GI and P, respectively). Examination of the tested specimens under the light microscope did not reveal any crack growth. Cracks grow only when GI is close to the critical value, GI,, indicating that no significant stress corrosion crack growth parallel to the fibers is observed in the NicalodCAS-Il composite at lower GI, at least for the experiments carried out in this study. However, Spearing et aL46 reports significant stress corrosion matrix cracking on the same material system under longitudinal loading conditions, at stresses below that required to develop matrix cracks in short-duration, monotonic loading tests, even at room temperature but for much longer stress application times (as long as 278 h). Some environmental stress cracking might be expected by oxidative removal of the interfacial carbon layer at high tem￾peratures even for the short exposure times of this study. How￾ever, this seems to be prevented by subsequent replacement of the oxidized carbon layer with a new interfacial phase closing the gap between fiber and matrix, and sealing the oxidation process 0ff.14-17*42 V. Summary and Conclusions DT test technique was successfully utilized to determine fracture toughness (GI,) for cracks parallel to the fibers in a 0 a 2 x * 20oc 0 600OC 8OOOC A lO0O~C 0 0.0 0.2 A 4A 4 1. rn 0. 0 A x I T 1 I . 0. bottom I I I I 0.4 0.6 0.8 I Dislance from Tap Surface/Thickness 0 Fig. 15. temperatures (bond strengths with mows exceeded the measurement capacity of the apparatus). Interfacial bond strength distribution along the crack from top to bottom face for DT specimens fractured (1-2 h long tests) at different