203 Journal of the American Ceramic Sociery-Morscher Vol. 80. No 8 cracked, the fibers were carrying the load at the matrix crack For the Nicalon minicomposites(Fig. 4(a)), a definite dif- nd Bn interphase minicomposites. The C-Nic minicomposite life was far more sensitive to stress than that of 3MBN-N minicomposites, where life was only mildly dependent on stress. For the Hi-Nicalon minicomposites(Fig. 4(b), no real difference was apparent between the two BN-coated systems ss dependence ior stress rupture at 750 and 950C, with a slightly more signifi- cant stress dependence at 1200C A few samples that did not fail were tested at room tem- rature to determine the retained strength after stressed oxi- dation. The 3MBN-Nic microcomposites were very difficult to handle after the experiment and could not be tested. The BN Hn composites could be handled and tested. The retained strength at room temperature of two minicomposites that were originally tested at 1200C for 882 h(328 MPa)and 965 h(385 Fig. 2. Optical micrograph of a polished longitudinal section of a MPa)was 525 and 792 MPa, respectively (-25% and 38% respectively, of the ultimate strength at room temperature) tion after the minicomposite was cycled 50 times to a load of 55 kg at (2)Microstructural Observations room temperature (A) C-Nic Minicomposites: The fracture surfaces of C- Nic minicomposites that have been tested at 700@C are shown The data for all the stress-rupture results are shown in in Fig. 5. Figure 5(a)shows a typical fracture surface for short graphical form on log scales in Figs. 4(a) and(b). The mini- time(<l h) failure at 700C. There is moderate fiber pullout, composites always failed in the hottest region of the furnace. and the fracture surface appears to be essentially the same as The data is plotted as the stress on the fibers, assuming that the room-temperature fast-fracture surfaces. For longer rupture fibers carried the full load. The number of fibers were counted times(2-10 h), the fracture surfaces are characterized by long for each minicomposite (Table I) from several (at least five) fiber pullout lengths(Fig. 5(b). The pullout length is almost 2 shed cross sections. The number of fibers for a given mini mm for this broomlike fracture surface. The fracture surfaces type was found to differ from sample to sample by for the longest rupture times(106 h) are planar(Fig. 5(c) less than ten fibers. The highest number of fibers that were Some fibers appeared to be in contact with the matrix; this counted from the observed cross sections was used as the fiber was true even of samples that have been tested at higher loads tow count. The average fiber diameter was determined from the for shorter times( Fig. 5(d)). a glass or reaction layer could not average of fifty fibers(Table I). The cross-sectional area was be discerned between the fiber and matrix layer. However, the then determined predominant feature was fibers in contact with the matrix. For The data were plotted as stress applied to the fibers because longer rupture times, the planar fracture surface was charac he minicomposite area varies along the length, and the con- terized by most fibers contacting the matrix(Fig. 5(e)).How- sistent attribute of the minicomposites was the number of fiber ever, fracture of these fibers usually did not occur at the point per sample. Also, because the minicomposites were pre- of fiber/matrix contact(Figs. 5(d)and(e)). For all he C-Nic fracture surfaces, the interphase area that was originally carbon as vacant The volume fraction of fibers was determined from or C-Nic and pB minicomposites. The other minicomposites were not ana- cal fracture surfaces for HN-PBN ruptured ne 6 shows typi- the same for the other two bl 1000r15 I< Temperature 4500 E00 Load 3500 G00 2.3to14 09 failed in 2500 40 4 5 AE Energy per event1000 0100020003000400050006000700080009000 Time Fig. 3. Typical tensile data for a room-temperature precrack experiment and a high-temperature stress-rupture experiment, including load, temperature, and AE activity versus timThe data for all the stress-rupture results are shown in graphical form on log scales in Figs. 4(a) and (b). The mini￾composites always failed in the hottest region of the furnace. The data is plotted as the stress on the fibers, assuming that the fibers carried the full load. The number of fibers were counted for each minicomposite (Table I) from several (at least five) polished cross sections. The number of fibers for a given mini￾composite type was found to differ from sample to sample by less than ten fibers. The highest number of fibers that were counted from the observed cross sections was used as the fiber tow count. The average fiber diameter was determined from the average of fifty fibers (Table I). The cross-sectional area was then determined. The data were plotted as stress applied to the fibers because the minicomposite area varies along the length,¶ and the con￾sistent attribute of the minicomposites was the number of fibers per sample. Also, because the minicomposites were pre￾cracked, the fibers were carrying the load at the matrix crack and minicomposite failure always occurred at a matrix crack. For the Nicalon minicomposites (Fig. 4(a)), a definite dif￾ference in stress-rupture behavior exists between the carbon and BN interphase minicomposites. The C-Nic minicomposite life was far more sensitive to stress than that of 3MBN-Nic minicomposites, where life was only mildly dependent on stress. For the Hi-Nicalon minicomposites (Fig. 4(b)), no real difference was apparent between the two BN-coated systems. Both minicomposite types showed mild stress dependence for stress rupture at 750° and 950°C, with a slightly more signifi￾cant stress dependence at 1200°C. A few samples that did not fail were tested at room tem￾perature to determine the retained strength after stressed oxi￾dation. The 3MBN-Nic microcomposites were very difficult to handle after the experiment and could not be tested. The BN– HN composites could be handled and tested. The retained strength at room temperature of two minicomposites that were originally tested at 1200°C for 882 h (328 MPa) and 965 h (385 MPa) was 525 and 792 MPa, respectively (∼25% and 38%, respectively, of the ultimate strength at room temperature). (2) Microstructural Observations (A) C-Nic Minicomposites: The fracture surfaces of C￾Nic minicomposites that have been tested at 700°C are shown in Fig. 5. Figure 5(a) shows a typical fracture surface for short time (<1 h) failure at 700°C. There is moderate fiber pullout, and the fracture surface appears to be essentially the same as room-temperature fast-fracture surfaces. For longer rupture times (2–10 h), the fracture surfaces are characterized by long fiber pullout lengths (Fig. 5(b)). The pullout length is almost 2 mm for this broomlike fracture surface. The fracture surfaces for the longest rupture times (106 h) are planar (Fig. 5(c)). Some fibers appeared to be in contact with the matrix; this was true even of samples that have been tested at higher loads for shorter times (Fig. 5(d)). A glass or reaction layer could not be discerned between the fiber and matrix layer. However, the predominant feature was fibers in contact with the matrix. For longer rupture times, the planar fracture surface was charac￾terized by most fibers contacting the matrix (Fig. 5(e)). How￾ever, fracture of these fibers usually did not occur at the point of fiber/matrix contact (Figs. 5(d) and (e)). For all the C-Nic fracture surfaces, the interphase area that was originally carbon was vacant. (B) BN Interphase Minicomposites: Figure 6 shows typi￾cal fracture surfaces for HN-PBN ruptured minicomposites; however, the observations are the same for the other two BN- ¶ The volume fraction of fibers was determined from several digitized images of minicomposite polished cross sections and found to vary over the range of 0.2–0.25 for C-Nic and PBN-HN minicomposites. The other minicomposites were not ana￾lyzed. Fig. 3. Typical tensile data for a room-temperature precrack experiment and a high-temperature stress-rupture experiment, including load, temperature, and AE activity versus time. Fig. 2. Optical micrograph of a polished longitudinal section of a PBN-HN minicomposite showing a relatively uniform crack distribu￾tion after the minicomposite was cycled 50 times to a load of 55 kg at room temperature. 2032 Journal of the American Ceramic Society—Morscher Vol. 80, No. 8