December 2007 Communications of the American Ceramic Society Table I. Basic Physical Properties of the As-Received after oxidation at 375.C(Fig. 1). However, a transition from Nicalon"/CAS Composite, Including Three-Point Flexure typical"pseudo-ductile"composite failure(with a failure"tail Mechanical Data observed on the load-deflection curve indicative of fiber pull- out)to a brittle failure mode (with no failure"tail")occurs be- Measured value tween 100 and 1000 h(Fig. 2). Similar behavior is apparent at Fiber fraction(vol%) 38.1+4.5 450oC. however the transition to a brittle failure mode is accel 2.67+0.0 erated, occurring between 10 and 100 h of oxidation exposure cn At 525 and 600C, the maximum flexure strength decreases Flexure modulus(GPa) 92.0+5.0 Flexure strength, or(MPa) 0l0+14. continuously with increasing heat-treatment duration, although a slight strength increase is initially noted after oxidation at Flexure proportional limit, omc(MPa) 1160+4.0 525C for 10 h; an essentially brittle fracture mode is observed for all these examples. The fracture transition is accompanied by hange from a fibrous failure surface (i.e, considerable fiber composite properties is presented in Table I. Bars for flexure pull-out)to one where there is minimal fiber pull out testing were cut from a hot-pressed plate(thickness 2.35 mm) to dimensions of 50-mm length x 5-mm width. Weight change The influence of oxidation exposure time and temperature, after oxidation was determined to +0.0001 g. Preliminary ex- from 375. to 600C, upon the mass loss of the Nicalon"/CAS amination of samples oxidized at 300C for 1000 h showed neg- exure bars is demonstrated in Fig. 3. At 375C, the mass loss ligible change in weight (i.e, within measuring error).A exhibits a nominally linear dependence upon time. Increasing the consequence, oxidation temperatures from 375 to 600C were oxidation temperature results in significant deviation from the behavior of the material observed at 375C. at both 450. and (model 1200, Barnstead/Thermolyne, Dubuque, IA). for dura- however the samples have subsequently increased in mass after tions of up to 1000 h. Test bars were sited on a high-purity 1000 h exposure. This effect is even more apparent at 600C Al2O3 support, and inserted into and removed from the furnace where weight loss is observed within the first 10 h, with subse- It temperature. There was no evidence of sample/support read quent weight gain after this time. It was noted that for oxidation tion with this arrangement, for any of the oxidation condition periods in excess of 100 h, samples treated at 600C showed tudied. Flexure testing was performed at a displacement rate of essentially no further weight gain within experimental error a three-point bend fixture, with an ou loading span of 40 mm, giving a span to thickness ratio of 17: 1. At least two tests were performed for each combination 40-nm thick, which has a turbostratic-type layered graph tructure.In previous work on Nicalon"/CAs, scanning treatments, the samples were preloaded to the desired stress in a Auger microscopy analysis demonstrated that the carbon layer Sic four-point bend fixture, with inner and outer spans of 6.35 d40 ely, held within a Mosi, elemen t furnace in ccordance with the observation of weight loss noted in the air. Samples were then heated to the desired test temperature, present work( Fig. 3). The weight loss data demonstrates similar and held at this temperature until failure. Microstructures of the oxidative removal of carbon at all examined temperatures received and oxidized specimens were examined by field emis between 375 and 600C. However, it is also apparent that, sion scanning electron microscopy(FE-SEM: Hitachi S-4500. with the exception of exposure at 375.C, mass gain occurs after Tokyo, Japan) n initial period of loss. This behavior is indicative of oxidation of the exposed SiO-C fiber surface, forming amorphous Sio with an associated increase in mass. Several previous studies II. Results and discussion have noted the formation of sio, ligaments that bridge across The effects of low-temperature oxidation heat treatments the fiber/matrix interface within this intermediate temperature the maximum flexure strength of Nicalon"/ CAS are presented ange 12,13 scanni er microscopy has also demonstrate of the load/deflection curves obtained for he presence of Sio, regions on the fiber surface in the case of Nicalon"/CAS after this series of heat treatments is show icalon"/ CAS oxidized at 600C for 100 h. schematically in Fig. 2. It is apparent that strength is retained Ultimately, the exposed fiber ends will be sealed via Sio and even slightly increased, for exposure times up to 1000 h, formation and plugging, preventing further oxidation of the will essential ease, as noted for samples held at 600C for times in excess of 100 h(Fig. 3). It is clear that care must be taken in imparting any significance upon the apparent transition from linear oxi dation kinetics at 375C to the behavior at 450. and 525C. as carbon weight loss through oxidation will be masked by weight gain through SiO formation, thus complicating analysis. It is now well established that in this situation carbon removal occurs via a"pipe-line"oxidation process. Huger et al.developed a simple geometrical model to estimate the critical time, Ie, taken for sealing of the exposed fiber ends, such that carbon oxidation ceases. This model was based on oxidation of SiC-SiC compos- ites, where both the fiber and matrix can oxidize. Their a 国 proach was subsequently adapted for the case of high temperature sealing of composites where only the fiber can oxidize. For a nonoxidizing matrix the critical sealing time Ic, Is given by Fig. 1. Three point flexure strength of Nicalon"/CAS after isothermal B(-(1/01) xidation at temperatures between 375 and 600C, for up to 1000 h The as-received flexure strength, before oxidation exposure, was 501+14 where Br is the rate constant for oxidation of fibers in air. H is on interphase thickness, and er iscomposite properties is presented in Table I. Bars for flexure testing were cut from a hot-pressed plate (thickness B2.35 mm) to dimensions of 50-mm length
5-mm width. Weight change after oxidation was determined to 70.0001 g. Preliminary ex￾amination of samples oxidized at 3001C for 1000 h showed neg￾ligible change in weight (i.e., within measuring error). As a consequence, oxidation temperatures from 3751 to 6001C were subsequently examined. Unstressed oxidation heat treatments were performed in a conventional laboratory muffle furnace (model 1200, Barnstead/Thermolyne, Dubuque, IA), for dura￾tions of up to 1000 h. Test bars were sited on a high-purity Al2O3 support, and inserted into and removed from the furnace at temperature. There was no evidence of sample/support reac￾tion with this arrangement, for any of the oxidation conditions studied. Flexure testing was performed at a displacement rate of 0.5 mm/min using a three-point bend fixture, with an outer loading span of 40 mm, giving a span to thickness ratio of B17:1. At least two tests were performed for each combination of oxidation temperature and time. For the stressed fatigue heat treatments, the samples were preloaded to the desired stress in a SiC four-point bend fixture, with inner and outer spans of 6.35 and 40 mm, respectively, held within a MoSi2 element furnace in air. Samples were then heated to the desired test temperature, and held at this temperature until failure. Microstructures of the as-received and oxidized specimens were examined by field emis￾sion scanning electron microscopy (FE-SEM; Hitachi S-4500, Tokyo, Japan). III. Results and Discussion The effects of low-temperature oxidation heat treatments upon the maximum flexure strength of Nicalont/CAS are presented in Fig. 1. A summary of the load/deflection curves obtained for Nicalont/CAS after this series of heat treatments is shown schematically in Fig. 2. It is apparent that strength is retained, and even slightly increased, for exposure times up to 1000 h, after oxidation at 3751C (Fig. 1). However, a transition from typical ‘‘pseudo-ductile’’ composite failure (with a failure ‘‘tail’’ observed on the load–deflection curve indicative of fiber pull￾out) to a brittle failure mode (with no failure ‘‘tail’’) occurs be￾tween 100 and 1000 h (Fig. 2). Similar behavior is apparent at 4501C, however the transition to a brittle failure mode is accel￾erated, occurring between 10 and 100 h of oxidation exposure. At 5251 and 6001C, the maximum flexure strength decreases continuously with increasing heat-treatment duration, although a slight strength increase is initially noted after oxidation at 5251C for 10 h; an essentially brittle fracture mode is observed for all these examples. The fracture transition is accompanied by a change from a fibrous failure surface (i.e., considerable fiber pull-out) to one where there is minimal fiber pull out. The influence of oxidation exposure time and temperature, from 3751 to 6001C, upon the mass loss of the Nicalont/CAS flexure bars is demonstrated in Fig. 3. At 3751C, the mass loss exhibits a nominally linear dependence upon time. Increasing the oxidation temperature results in significant deviation from the behavior of the material observed at 3751C. At both 4501 and 5251C, mass loss increases to a maximum after 100 h exposure, however the samples have subsequently increased in mass after 1000 h exposure. This effect is even more apparent at 6001C, where weight loss is observed within the first 10 h, with subse￾quent weight gain after this time. It was noted that for oxidation periods in excess of 100 h, samples treated at 6001C showed essentially no further weight gain within experimental error. Nicalont/CAS composites possess an in situ formed carbon interlayer between the fiber and matrix, typically between 20 and 40-nm thick, which has a turbostratic-type layered graphite structure.21,26 In previous work on Nicalont/CAS, scanning Auger microscopy analysis demonstrated that the carbon layer is completely removed after oxidation at 4501C for 100 h,27 in accordance with the observation of weight loss noted in the present work (Fig. 3). The weight loss data demonstrates similar oxidative removal of carbon at all examined temperatures between 3751 and 6001C. However, it is also apparent that, with the exception of exposure at 3751C, mass gain occurs after an initial period of loss. This behavior is indicative of oxidation of the exposed Si–O–C fiber surface, forming amorphous SiO2 with an associated increase in mass. Several previous studies have noted the formation of SiO2 ligaments that bridge across the fiber/matrix interface within this intermediate temperature range.12,13 Scanning Auger microscopy has also demonstrated the presence of SiO2 regions on the fiber surface in the case of Nicalont/CAS oxidized at 6001C for 100 h.27 Ultimately, the exposed fiber ends will be sealed via SiO2 formation and plugging, preventing further oxidation of the interlayer.21 At this point, further mass gain will essentially cease, as noted for samples held at 6001C for times in excess of 100 h (Fig. 3). It is clear that care must be taken in imparting any significance upon the apparent transition from linear oxi￾dation kinetics at 3751C to the behavior at 4501 and 5251C, as carbon weight loss through oxidation will be masked by weight gain through SiO2 formation, thus complicating analysis. It is now well established that in this situation carbon removal occurs via a ‘‘pipe-line’’ oxidation process.18–20 Huger et al.28 developed a simple geometrical model to estimate the critical time, tc, taken for sealing of the exposed fiber ends, such that carbon oxidation ceases. This model was based on oxidation of SiC–SiC compos￾ites, where both the fiber and matrix can oxidize. Their ap￾proach was subsequently adapted for the case of high temperature sealing of composites where only the fiber can oxidize.27 For a nonoxidizing matrix the critical sealing time, tc, is given by: tc ¼ 1 Bf H ð Þ 1 ð1=yfÞ
2 where Bf is the parabolic rate constant for oxidation of the fibers in air, H is the carbon interphase thickness, and yf is the Table I. Basic Physical Properties of the As-Received Nicalont/CAS Composite, Including Three-Point Flexure Mechanical Data Property Measured value Fiber fraction (vol%) 38.174.5 Density (g/cm3 ) 2.6770.03 Flexure modulus (GPa) 92.075.0 Flexure strength, sf (MPa) 501.0714.0 Flexure proportional limit, smc (MPa) 116.074.0 0 100 200 300 400 500 600 10 100 1000 375 450 525 600 Oxidation time (hrs) Flexure strength (MPa) Fig. 1. Three point flexure strength of Nicalont/CAS after isothermal oxidation at temperatures between 3751 and 6001C, for up to 1000 h. The as-received flexure strength, before oxidation exposure, was 501714 MPa. December 2007 Communications of the American Ceramic Society 4051