Availableonlineatwww.sciencedirect.com COMPOSITES ScienceDirect SCIENCE AND TECHNOLOGY ELSEVIER Composites Science and Technology 68(2008)274-282 w.elsevier. com/locate/compscitech Effects of moisture on tensile stress rupture behavior of a SiC/SiC composite at elevated temperatures Shankar Mall, J.L. Ryba Department of Aeronautics and Astronautics, Air Force Institute of Technology, Bldg 640, 2950 Hobson Way, Wright-Patterson AFB, OH 45433-7765, USA Received 13 October 2006: received in revised form 20 February 2007: accepted 6 March 2007 vailable online 16 March 2007 Abstract This study characterized the moisture and temperature effects on the stress rupture behavior of ceramic matrix composite(CMC), Syl iBN/BN/SiC, which was composed of Sylramic(Syl) fibers with an in situ layer of boron nitride(iBN), boron nitride interphase(BN)and silicon carbide(SiC)matrix. For this purpose, stress rupture tests were conducted under 100% steam and laboratory air environments at three temperatures, 400C, 750C, and 950C to develop stress versus time to failure relationships At a given stress level, time to failur decreased with increase of the test temperature and moisture. However, the stress rupture strength at 750C under both laboratory and steam conditions was relatively low than the expected degradation based on the interpolation from those at 400-950C. Similar characteristics were observed in residual strength of the corresponding specimens which survived 100 h and also in the damage mecha nisms. The degradation of the bn interphase between the fiber and the matrix was the primary damage which was relatively more in the team environment than that in air laboratory. This damage was in the form of fracture, recession and loss of Bn interphase at 400C or 50C while there was pesting of fibers due to formation of borosilicate causing relatively more embrittlement of the tested CMC system at750°C Published by Elsevier Ltd Keywords: A Ceramics matrix composites: B Stress; C. Fiber; Silicon carbide 1. Introduction non-oxide components. CMCs have been manufactured with oxide fibers within an oxide matrix, such as the Nextel Ceramic matrix composites(CMCs )are the promising 720/alumina CMC system, non-oxide fibers within oxide materials for high temperature applications because of matrices, such as the carbon(C)/alumina-silicate CMC sys- their low density, high strength, high toughness and high tem, and with non-oxide fibers within non-oxide matrices, durability at elevated temperatures. A CMC is typically such as the silicon carbide fibers reinforced in silicon car- composed of, at the least, fibers and the matrix surround- bide matrix( SiC/SiC)CMC system. ing them. Some CMCs, such as the one tested in this study Nicalon(Nic), Hi-Nicalon(HN) and Sylramic(Syl)are incorporate also an interphase between fibers and matrix. few examples of silicon carbide fibers which are commer Ceramic fibers provide CMCs high strength and high elas- cially available [1-3]. As expected these fibers have different tic modulus in addition to their high temperature capabil- mechanical properties at elevated temperature and/or in ity. Ceramic matrix composites can consist of oxide and oxidizing environments. Besides fibers, another important constituent of CMCs is the fiber/matrix interphase. Carbon (C)and boron nitride(bn) have been two commonly used w The views expressed in this article are those of the authors and do ne reflect the official policy or position of the United States Air force. interphase materials in the SiC/SiC CMCs [2]. Unfortu- Department of Defense, or the US Government nately, carbon oxidizes and volatizes into CO and co readily in environments containing moisture. The disap- E-mail Shankar. Mall@afit. edu (S. Mall) pearance of the carbon generates a space between the fiber 0266-3538S.see front matter Published by Elsevier Ltd. doi: 10.1016j. compscitech. 2007.03.005
Effects of moisture on tensile stress rupture behavior of a SiC/SiC composite at elevated temperatures q Shankar Mall *, J.L. Ryba Department of Aeronautics and Astronautics, Air Force Institute of Technology, Bldg. 640, 2950 Hobson Way, Wright-Patterson AFB, OH 45433-7765, USA Received 13 October 2006; received in revised form 20 February 2007; accepted 6 March 2007 Available online 16 March 2007 Abstract This study characterized the moisture and temperature effects on the stress rupture behavior of ceramic matrix composite (CMC), SyliBN/BN/SiC, which was composed of Sylramic (Syl) fibers with an in situ layer of boron nitride (iBN), boron nitride interphase (BN) and silicon carbide (SiC) matrix. For this purpose, stress rupture tests were conducted under 100% steam and laboratory air environments at three temperatures, 400 C, 750 C, and 950 C to develop stress versus time to failure relationships. At a given stress level, time to failure decreased with increase of the test temperature and moisture. However, the stress rupture strength at 750 C under both laboratory air and steam conditions was relatively low than the expected degradation based on the interpolation from those at 400–950 C. Similar characteristics were observed in residual strength of the corresponding specimens which survived 100 h and also in the damage mechanisms. The degradation of the BN interphase between the fiber and the matrix was the primary damage which was relatively more in the steam environment than that in air laboratory. This damage was in the form of fracture, recession and loss of BN interphase at 400 C or 950 C while there was pesting of fibers due to formation of borosilicate causing relatively more embrittlement of the tested CMC system at 750 C. Published by Elsevier Ltd. Keywords: A. Ceramics matrix composites; B. Stress; C. Fiber; Silicon carbide 1. Introduction Ceramic matrix composites (CMCs) are the promising materials for high temperature applications because of their low density, high strength, high toughness and high durability at elevated temperatures. A CMC is typically composed of, at the least, fibers and the matrix surrounding them. Some CMCs, such as the one tested in this study, incorporate also an interphase between fibers and matrix. Ceramic fibers provide CMCs high strength and high elastic modulus in addition to their high temperature capability. Ceramic matrix composites can consist of oxide and non-oxide components. CMCs have been manufactured with oxide fibers within an oxide matrix, such as the Nextel 720/alumina CMC system, non-oxide fibers within oxide matrices, such as the carbon (C)/alumina-silicate CMC system, and with non-oxide fibers within non-oxide matrices, such as the silicon carbide fibers reinforced in silicon carbide matrix (SiC/SiC) CMC system. Nicalon (Nic), Hi-Nicalon (HN) and Sylramic (Syl) are few examples of silicon carbide fibers which are commercially available [1–3]. As expected these fibers have different mechanical properties at elevated temperature and/or in oxidizing environments. Besides fibers, another important constituent of CMCs is the fiber/matrix interphase. Carbon (C) and boron nitride (BN) have been two commonly used interphase materials in the SiC/SiC CMCs [2]. Unfortunately, carbon oxidizes and volatizes into CO and CO2 readily in environments containing moisture. The disappearance of the carbon generates a space between the fiber 0266-3538/$ - see front matter Published by Elsevier Ltd. doi:10.1016/j.compscitech.2007.03.005 q The views expressed in this article are those of the authors and do not reflect the official policy or position of the United States Air Force, Department of Defense, or the US Government. * Corresponding author. E-mail address: Shankar.Mall@afit.edu (S. Mall). www.elsevier.com/locate/compscitech Available online at www.sciencedirect.com Composites Science and Technology 68 (2008) 274–282 COMPOSITES SCIENCE AND TECHNOLOGY
S. Mall,J.L Ryba Composites Science and Technology 68(2008)274-282 275 and the matrix, thus increasing the access continually for reduction in the intermediate temperature range (i.e, from more exposure to the oxidizing environment. On the other 450 to 900C) under harsh environment. Therefore, the hand, BN also experiences oxidation, but not readily as the test temperatures of this study were selected in such man- ner that they were below this intermediate range, within In general CMCs experience embrittlement when this range, and above this range, respectively. Tests were exposed to oxidizing environment, and it has been referred conducted under 100% steam and laboratory air environ to as the oxidation embrittlement. This affects primarily ments to highlight oxidizing environments'effects In addi fiber/matrix interphase causing degradation in strength tion, monotonic tests were also conducted to establish the and toughness of the CMC. Oxidation embrittlement in baseline data. Finally, the detailed microscopic analyses several SiC/SiC CMC systems has been investigated in pre- were conducted to document the failure and damage vious studies [4-6] as well as in other CMC systems [7-12]. mechanisms It has been noted that the SiC/SiC CMC systems with BN on in the inte 2. Experi mediate temperature range(from 450 to 900C)under harsh environment since cracks in matrix material allow 2 material the outside environment to penetrate inside and attack the fiber/matrix interphase [7-12). This is due to oxidiza- As mentioned earlier, Syl- iBN/BN/SiC CMC was the tion of fiber/matrix interphase region forming the boria test material of this study. Honeywell Advanced Compos- (B2O3), which reacts with SiC. This generates a borosilicate ites, Inc. manufactured composite panels with Syl fibers melt, which is then solidified as a glass. This bonds fibers provided by NASa Glenn Research Center. The composite gether causing an embrittlement. In other words, oxida- material consisted of 8 plies of woven(5 Harness Satin) Syl tion in SiC/SiC CMC systems with BN interphase results tows containing 800 fibers. These preforms(i.e,woven in a solid brittle glass replacing the functional interphase plies) were treated in several steps. First, in situ BN pre- material, and thereby removes the inherent toughness forms had the interphase bn layer applied by chemical nechanisms. Thus, oxidation embrittlement is a concern vapor infiltration(CVI) process, resulting in a 10.64+ during the application of SiC/SiC CMC systems in humid 0.34% weight gain. Then, a thin layer of SiC was applied environments. One such example is the combustor liner by Cvi to the bn coated woven preform, resulting in a gas turbine engines [13] 53.40+.99% weight gain. SiC particle slurry was then One way to alleviate the oxidation embrittlement con- infiltrated into the porous network, resulting in a 30.10+ cern in the SiC/SiC CMC systems is to modify the fiber/ 1. 50%weight gain. Finally, molten Si was melt-infiltrated matrix interphase. One such modification involves the (Mi) to nearly fill the porous network, resulting in a pre-application of in situ grown BN layer on SiC fiber 13.92+0.80% weight gain. The resulting matrix was pre- before reinforcing in SiC matrix along with BN interphase dominantly SiC with some silicon (Si). The final volume [14]. This CMC system consisted of Syl fibers with in situ fractions were: for fibers, 37%, for BN, 6.6%, for CVI layer of boron nitride(iBN), Bn interphase and Sic SiC, 19.3%, for MI SiC and Si (including the porosity), matrix, and it is referred to as"Syl-iBN/BN/SiC". This 37.05%. The test specimens were cut in a dog-bone config- CMC system has been characterized for its stress rupture uration with the following nominal gage dimensions: and fatigue behavior under humid environment by the first 2. 1 mm thickness, 10.2 mm width and 100 mm length author and his colleagues [15-17], however these character-(Fig. 1) izations have been conducted in the intermediate tempera ture range (from 550 to 750C)only. In order to 2. 2. Test setup understand fully, it is necessary to investigate its behavior at other elevated temperatures also. This is focus of the All tests were conducted on a servo-hydraulics mechan present study. The specific objective of the present study ical testing machine equipped with hydraulic water-cooled haracterize the tensile stress rupture behavior of grips, a compact two-zone resistance-heated furnace, and Syl-iBN/BN/SiC at three elevated temperatures, 400 two temperature controllers. a digital controller was systems with bn intes hase have iet, sin ticant strength as po ire the rest da contests were conducted es ther in Fig. I. Dog-bone shaped specimen
and the matrix, thus increasing the access continually for more exposure to the oxidizing environment. On the other hand, BN also experiences oxidation, but not readily as the carbon. In general CMCs experience embrittlement when exposed to oxidizing environment, and it has been referred to as the oxidation embrittlement. This affects primarily fiber/matrix interphase causing degradation in strength and toughness of the CMC. Oxidation embrittlement in several SiC/SiC CMC systems has been investigated in previous studies [4–6] as well as in other CMC systems [7–12]. It has been noted that the SiC/SiC CMC systems with BN interphase have a significant strength reduction in the intermediate temperature range (from 450 to 900 C) under harsh environment since cracks in matrix material allow the outside environment to penetrate inside and attack the fiber/matrix interphase [7–12]. This is due to oxidization of fiber/matrix interphase region forming the boria (B2O3), which reacts with SiC. This generates a borosilicate melt, which is then solidified as a glass. This bonds fibers together causing an embrittlement. In other words, oxidation in SiC/SiC CMC systems with BN interphase results in a solid brittle glass replacing the functional interphase material, and thereby removes the inherent toughness mechanisms. Thus, oxidation embrittlement is a concern during the application of SiC/SiC CMC systems in humid environments. One such example is the combustor liner in gas turbine engines [13]. One way to alleviate the oxidation embrittlement concern in the SiC/SiC CMC systems is to modify the fiber/ matrix interphase. One such modification involves the pre-application of in situ grown BN layer on SiC fiber before reinforcing in SiC matrix along with BN interphase [14]. This CMC system consisted of Syl fibers with in situ layer of boron nitride (iBN), BN interphase and SiC matrix, and it is referred to as ‘‘Syl-iBN/BN/SiC’’. This CMC system has been characterized for its stress rupture and fatigue behavior under humid environment by the first author and his colleagues [15–17], however these characterizations have been conducted in the intermediate temperature range (from 550 to 750 C) only. In order to understand fully, it is necessary to investigate its behavior at other elevated temperatures also. This is focus of the present study. The specific objective of the present study was to characterize the tensile stress rupture behavior of Syl-iBN/BN/SiC at three elevated temperatures, 400 C, 750 C, and 950 C. As stated earlier, the SiC/SiC CMC systems with BN interphase have a significant strength reduction in the intermediate temperature range (i.e., from 450 to 900 C) under harsh environment. Therefore, the test temperatures of this study were selected in such manner that they were below this intermediate range, within this range, and above this range, respectively. Tests were conducted under 100% steam and laboratory air environments to highlight oxidizing environments’ effects. In addition, monotonic tests were also conducted to establish the baseline data. Finally, the detailed microscopic analyses were conducted to document the failure and damage mechanisms. 2. Experiments 2.1. Material As mentioned earlier, Syl-iBN/BN/SiC CMC was the test material of this study. Honeywell Advanced Composites, Inc. manufactured composite panels with Syl fibers provided by NASA Glenn Research Center. The composite material consisted of 8 plies of woven (5 Harness Satin) Syl tows containing 800 fibers. These preforms (i.e., woven plies) were treated in several steps. First, in situ BN preforms had the interphase BN layer applied by chemical vapor infiltration (CVI) process, resulting in a 10.64 ± 0.34% weight gain. Then, a thin layer of SiC was applied by CVI to the BN coated woven preform, resulting in a 53.40 ± 5.99% weight gain. SiC particle slurry was then infiltrated into the porous network, resulting in a 30.10 ± 1.50% weight gain. Finally, molten Si was melt-infiltrated (MI) to nearly fill the porous network, resulting in a 13.92 ± 0.80% weight gain. The resulting matrix was predominantly SiC with some silicon (Si). The final volume fractions were: for fibers, 37%, for BN, 6.6%, for CVI SiC, 19.3%, for MI SiC and Si (including the porosity), 37.05%. The test specimens were cut in a dog-bone configuration with the following nominal gage dimensions: 2.1 mm thickness, 10.2 mm width and 100 mm length (Fig. 1). 2.2. Test setup All tests were conducted on a servo-hydraulics mechanical testing machine equipped with hydraulic water-cooled grips, a compact two-zone resistance-heated furnace, and two temperature controllers. A digital controller was employed to generate and control all test commands as well as to acquire the test data. Tests were conducted either in Fig. 1. Dog-bone shaped specimen. S. Mall, J.L. Ryba / Composites Science and Technology 68 (2008) 274–282 275
J. L Ryba/ Composites Science and Technology 68 (2008)274-282 specimen fractured or the test run-out limit occurred, which was set at 100 h in this study. Henceforth, tests performed in laboratory air or in 100% steam environment will be referred to as tests in air or steam or the in-air or in-steam 3. Results and discussion Generator 3.. Monotonic test Ahmir Enclosure Fig 3 shows the typical monotonic tensile stress strain curves at room and three elevated temperatures(400C, 750C and 950C)of the Syl-iBN/BN/SiC CMC system This figure also shows the corresponding curves at 550C and 750C for the same material from a previous study [15]. A summary of material properties from the mono- Fig. 2. Schematic of the heating chamber with enclosure for steam tonic tests are presented in Table 1. The monotonic tensile environment stress strain curves at 750C from the present and previous studies are in good agreement with each other, and the the laboratory air or in the steam environment. For the lat ter an alumina susceptor (i.e, enclosure), as shown sche matically in Fig. 2, was used. The specimen gage section was located inside the susceptor. A commercially available steam generator was used which measured and controlled the desired humidity of the steam. Steam was introduced 550°c[15 into the susceptor through a feeding tube as a continuous stream with a slightly positive pressure, expelling the dry air and thereby creating a 100% steam environment inside the susceptor. R-type thermocouples were initially glued on 2 a test specimen in order to calibrate the furnace temper- 2 zoo ature for the elevated temperature tests, both for the steam and dry conditions. This allowed calibration of the furnace itially and on the periodic basis during course of this study within +2C. Thus power setting of the furnace for a desired temperature was determined which was then used 950°c in the actual tests. Other than furnace calibration thermo- couples were not attached to the test speci Strain 23. Test details ig. 3. Monotonic tensile stress strain of this study and from Two types of tests were conducted; monotonic tension previous study[I5y and stress rupture. Two test environments, having either air or 100% steam conditions, were used at Table I 400°C,750°Cand950°C. Monotonic tensile tests were Test data summary conducted at these temperatures and also at room tempera- Test Young's Ultimate Normalized Failure Sourd ture to get baseline information of the tested CMC system. conditions modulus strength ultimate strain These were conducted in the laboratory air environment (GPa) (MPa) only since a previous study showed that monotonic tensile RT Present behavior of the Syl-iBN /BN/SiC CMC system was depen dent on the temperature, but not on the humidity [17]. This 400C 231 334 0.81 0.36 Present was expected since the test duration in monotonic tests was not long enough to cause any effect from moisture in the 750°C 0.79 0.38 Present monotonic tension tests. The monotonic tensile test 0.74 Present involved loading of the specimen under the load control mode with a loading rate of 3333 N/min Stress rupture 550C 233 [5] tests involved application of a prescribed load in a short 750C 235 335 0.81 0.36 [5] duration(l min)and then keeping it constant until the Normalized with the ultimate tensile strength at room temperature
the laboratory air or in the steam environment. For the latter an alumina susceptor (i.e., enclosure), as shown schematically in Fig. 2, was used. The specimen gage section was located inside the susceptor. A commercially available steam generator was used which measured and controlled the desired humidity of the steam. Steam was introduced into the susceptor through a feeding tube as a continuous stream with a slightly positive pressure, expelling the dry air and thereby creating a 100% steam environment inside the susceptor. R-type thermocouples were initially glued on to a test specimen in order to calibrate the furnace temperature for the elevated temperature tests, both for the steam and dry conditions. This allowed calibration of the furnace initially and on the periodic basis during course of this study within ±2 C. Thus power setting of the furnace for a desired temperature was determined which was then used in the actual tests. Other than furnace calibration, thermocouples were not attached to the test specimens. 2.3. Test details Two types of tests were conducted; monotonic tension and stress rupture. Two test environments, having either laboratory air or 100% steam conditions, were used at 400 C, 750 C and 950 C. Monotonic tensile tests were conducted at these temperatures and also at room temperature to get baseline information of the tested CMC system. These were conducted in the laboratory air environment only since a previous study showed that monotonic tensile behavior of the Syl-iBN/BN/SiC CMC system was dependent on the temperature, but not on the humidity [17]. This was expected since the test duration in monotonic tests was not long enough to cause any effect from moisture in the monotonic tension tests. The monotonic tensile test involved loading of the specimen under the load control mode with a loading rate of 3333 N/min. Stress rupture tests involved application of a prescribed load in a short duration (1 min) and then keeping it constant until the specimen fractured or the test run-out limit occurred, which was set at 100 h in this study. Henceforth, tests performed in laboratory air or in 100% steam environment will be referred to as tests in air or steam, or the in-air or in-steam tests. 3. Results and discussion 3.1. Monotonic test Fig. 3 shows the typical monotonic tensile stress strain curves at room and three elevated temperatures (400 C, 750 C and 950 C) of the Syl-iBN/BN/SiC CMC system. This figure also shows the corresponding curves at 550 C and 750 C for the same material from a previous study [15]. A summary of material properties from the monotonic tests are presented in Table 1. The monotonic tensile stress strain curves at 750 C from the present and previous studies are in good agreement with each other, and the Fig. 2. Schematic of the heating chamber with enclosure for steam environment. 0 50 100 150 200 250 300 350 400 450 0 0.1 0.2 0.3 0.4 0.5 Strain % Stress (MPa) Room Temperature 4000 C 7500C 9500C 5500C [15] 7500C [15] Fig. 3. Monotonic tensile stress strain curves of this study and from previous study [15]. Table 1 Test data summary Test conditions Young’s modulus (GPa) Ultimate strength (MPa) Normalized ultimate strengtha Failure strain (mm/mm) Source RT 235 415 1 0.44 Present study 400 C 231 334 0.81 0.36 Present study 750 C 232 327 0.79 0.38 Present study 950 C 225 305 0.74 0.41 Present study 550 C 233 347 0.84 0.39 [15] 750 C 235 335 0.81 0.36 [15] a Normalized with the ultimate tensile strength at room temperature. 276 S. Mall, J.L. Ryba / Composites Science and Technology 68 (2008) 274–282
S. Mall,J.L Ryba Composites Science and Technology 68(2008)274-282 same at 550C from the previous study is also in agreement ature was lower than that at room temperature, the nor with its counterparts at 400C and 750C of this study. As malized stress at zero hour is not unity. The data, shown can be seen in Fig. 3 and Table 1, the ultimate tensile at 100 h, show those which did not fail up to 100 h which strength is reduced at the elevated temperatures relative was the run-out limit in this study. As can be seen in to that of room temperature. This reduction is almost same Fig. 4, both moisture and temperature have the detrimental in the temperature range used in this study(from 400 to effect on the stress rupture strength of the tested CMC sys 950C). The ultimate tensile strength at the elevated tem- tem. The composite survived up to at least 100 h at 400C peratures ranged from 74% to 84% of its counterpart at when loaded to a constant stress level of 72% and 60% of room temperature. However, there was one major differ- its ultimate tensile strength at room temperature under lab- nce in the initial linear portion of the stress strain curves oratory air and steam conditions, respectively. Their coun- between 950C and room temperature, 400C, 550C terparts at 750C and 950C were 52%(in air) or 49%(in and 750C. The stress strain curves at room temperature, steam) and 40%(in air) or 34%(in steam), respectively 400C, 550C and 750C show initially a linear portion Further, it is interesting to note that stress rupture strength up to about 200-225 MPa, which represents the portion generally reduced considerably due to exposure to moisture before any appreciable amount of matrix crack was devel- over a short duration oped in the composite. On the other hand, the stress strain 4 clearly shows that the moisture degraded the curve at 950C shows initial linear portion up to about CMC system at all three elevated temperatures used in this 100 MPa which is almost half of its counterpart at other study. This is elaborated in Fig. 5 where the interpolated temperatures. Thus, this CMCs system is degraded consid- normalized stress rupture strengths at 50 h for all six erably in this respect (i.e, appreciable amount of matrix ronments of this study from Fig 4 are plotted. This figure cracks had already developed when loaded up to compares the interpolated normalized stress rupture 100 MPa at 950C). The stress-strain curves after the strength at 50 h for all six test environments of this study knee-point show similar behavior at room and all elevated from Fig 4. This comparison suggests that stress rupture temperatures of this study, i.e., a continuously decreasing strength at 750C under both laboratory air and steam slope due to further matrix cracks and associated non-lin- conditions is relatively less than the expected degradation from 400C to 950C which is shown by the straight lines In other words, stress rupture strength at 750C in both 3.2. Stress rupture test test environments is less than the linearly interpolated val- ues between 400C and 950C. Thus, the tested Sic/sic Fig 4 shows stress rupture data in terms of the normal- CMC system with modified fiber/matrix interphase also ized stress versus time to failure relationships for both lab- performed relatively poor in the intermediate temperature oratory air and steam test environments. The y-axis of range(from 400 to 950C)compared to the outside of this figure is the applied stress in stress rupture tests was nor- range. This characteristic is similar to other SiC/SiC CMCs malized by the ultimate tensile strength at room tempera- systems as elaborated in the Introduction section ture. the data at zero hour in this figure are from the monotonic tensile tests conducted at each temperature, which have also been normalized. Since the ultimate tensile strength of the tested CMC system at each elevated temper S Air a Humid 素 0.6 400°c 708090100 950°c Fig 4. Stress rupture curves(normalized with room temperature ultimate Fig. 5. Estimated normalized rupture strength at 50 h at three tensile strength
same at 550 C from the previous study is also in agreement with its counterparts at 400 C and 750 C of this study. As can be seen in Fig. 3 and Table 1, the ultimate tensile strength is reduced at the elevated temperatures relative to that of room temperature. This reduction is almost same in the temperature range used in this study (from 400 to 950 C). The ultimate tensile strength at the elevated temperatures ranged from 74% to 84% of its counterpart at room temperature. However, there was one major difference in the initial linear portion of the stress strain curves between 950 C and room temperature, 400 C, 550 C and 750 C. The stress strain curves at room temperature, 400 C, 550 C and 750 C show initially a linear portion up to about 200–225 MPa, which represents the portion before any appreciable amount of matrix crack was developed in the composite. On the other hand, the stress strain curve at 950 C shows initial linear portion up to about 100 MPa which is almost half of its counterpart at other temperatures. Thus, this CMCs system is degraded considerably in this respect (i.e., appreciable amount of matrix cracks had already developed when loaded up to 100 MPa at 950 C). The stress–strain curves after the knee-point show similar behavior at room and all elevated temperatures of this study, i.e., a continuously decreasing slope due to further matrix cracks and associated non-linear displacements. 3.2. Stress rupture test Fig. 4 shows stress rupture data in terms of the normalized stress versus time to failure relationships for both laboratory air and steam test environments. The y-axis of figure is the applied stress in stress rupture tests was normalized by the ultimate tensile strength at room temperature. The data at zero hour in this figure are from the monotonic tensile tests conducted at each temperature, which have also been normalized. Since the ultimate tensile strength of the tested CMC system at each elevated temperature was lower than that at room temperature, the normalized stress at zero hour is not unity. The data, shown at 100 h, show those which did not fail up to 100 h which was the run-out limit in this study. As can be seen in Fig. 4, both moisture and temperature have the detrimental effect on the stress rupture strength of the tested CMC system. The composite survived up to at least 100 h at 400 C when loaded to a constant stress level of 72% and 60% of its ultimate tensile strength at room temperature under laboratory air and steam conditions, respectively. Their counterparts at 750 C and 950 C were 52% (in air) or 49% (in steam) and 40% (in air) or 34% (in steam), respectively. Further, it is interesting to note that stress rupture strength generally reduced considerably due to exposure to moisture over a short duration. Fig. 4 clearly shows that the moisture degraded the CMC system at all three elevated temperatures used in this study. This is elaborated in Fig. 5 where the interpolated normalized stress rupture strengths at 50 h for all six environments of this study from Fig. 4 are plotted. This figure compares the interpolated normalized stress rupture strength at 50 h for all six test environments of this study from Fig. 4. This comparison suggests that stress rupture strength at 750 C under both laboratory air and steam conditions is relatively less than the expected degradation from 400 C to 950 C which is shown by the straight lines. In other words, stress rupture strength at 750 C in both test environments is less than the linearly interpolated values between 400 C and 950 C. Thus, the tested SiC/SiC CMC system with modified fiber/matrix interphase also performed relatively poor in the intermediate temperature range (from 400 to 950 C) compared to the outside of this range. This characteristic is similar to other SiC/SiC CMCs systems as elaborated in the Introduction section. 0.3 0.4 0.5 0.6 0.7 0.8 0.9 0 10 20 30 40 50 60 70 80 90 100 Time (hr) Normalized Stress Steam Steam Steam Air Air Air 4000 C 7500 C 9500 C 4000 C 7500 C 9500 C 4000 C 7500C 9500C Fig. 4. Stress rupture curves (normalized with room temperature ultimate tensile strength). 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 Temperature (C) Normalized Stress Air Humid 4000C 7500C 9500C Fig. 5. Estimated normalized rupture strength at 50 h at three temperatures. S. Mall, J.L. Ryba / Composites Science and Technology 68 (2008) 274–282 277
278 J. L Ryba/ Composites Science and Technology 68 (2008)274-282 口 Stean 需0.6 80.5 950c Fig. 6. Normalized residual strength. Finally, the test specimens that survived 100 h of stress rupture were tested under the monotonic tension test in the laboratory air test environment to measure their resid ual strength. This provided the estimate of damage devel- oped during the tensile stress rupture test over 100 h. ig. 6 shows these residual strengths which are normalized by the ultimate tensile strength at room temperature. This comparison clearly shows that the residual strength is less in the steam test environment than that in the laborator air environment at each elevated temperature. In addition, it is lower at 750 oC in either air or steam environment than its counterparts at 400C and 950C. further, the residual strength at 750oC under the steam test environment is the minimum among all the six test conditions of this study This suggests that intermediate temperature range test con- dition has relatively more detrimental effect on the stress rupture performance of the tested CMC system under both Fig. 7. Fracture Surface from monotonic tensile test at 950 C,(a) at laboratory air and steam conditions lower magnification(100x)and (b) at higher magnification(8000x 3.3. Damage mechanisms (marked by arrows as few such examples) along with deb- All tested specimens of this study were examined with onding between fiber and interphase or between interphase the scanning electron microscope(SEM) to observe failure and Sic matrix and damage mechanisms. In the case of monotonic tension The comparison of the typical fractured surfaces from tests under all four test conditions (room temperature, stress rupture tests under both laboratory air and steam 400C, 750C and 950C), fracture surfaces were rough, test conditions are shown in Figs. 8-10. These show images but still relatively fat and perpendicular to the loading for tests at 400C, 750C, and 950C(from left to right) direction. There were no distinctive features of the fracture under steam and laboratory air conditions. Fig. 8 shows which can be attributed to the change in the test tempera- these at lower magnification(100x)of one repr tre. Fracture surface was a typical of a woven brittle sample from each test environment. It can be seen that CMC system where the longitudinal tows showed the fiber the laboratory air tests show a relatively rougher fracture pull-out that did not change with test temperature from plane compared to the steam tests. Also, the tests con- room temperature to 950C. Therefore for the sake of ducted at 750C, in the intermediate range, have relatively brevity, only one case of fracture surface is shown at both flatter fracture surface with very little fiber pullout(Fig. 8b lower and higher magnifications in Fig. 7, where fiber pull- and e) whereas the 400C and 950C tests show pullout of out and fracture in matrix at different planes are clearly evi- longitudinal tows and planar fracture within the tows dent as well as the presence of fractured BN interphase (Fig. 8a, d, c and f. This indicates that the crack
Finally, the test specimens that survived 100 h of stress rupture were tested under the monotonic tension test in the laboratory air test environment to measure their residual strength. This provided the estimate of damage developed during the tensile stress rupture test over 100 h. Fig. 6 shows these residual strengths which are normalized by the ultimate tensile strength at room temperature. This comparison clearly shows that the residual strength is less in the steam test environment than that in the laboratory air environment at each elevated temperature. In addition, it is lower at 750 C in either air or steam environment than its counterparts at 400 C and 950 C. Further, the residual strength at 750 C under the steam test environment is the minimum among all the six test conditions of this study. This suggests that intermediate temperature range test condition has relatively more detrimental effect on the stress rupture performance of the tested CMC system under both laboratory air and steam conditions. 3.3. Damage mechanisms All tested specimens of this study were examined with the scanning electron microscope (SEM) to observe failure and damage mechanisms. In the case of monotonic tension tests under all four test conditions (room temperature, 400 C, 750 C and 950 C), fracture surfaces were rough, but still relatively flat and perpendicular to the loading direction. There were no distinctive features of the fracture which can be attributed to the change in the test temperature. Fracture surface was a typical of a woven brittle CMC system where the longitudinal tows showed the fiber pull-out that did not change with test temperature from room temperature to 950 C. Therefore for the sake of brevity, only one case of fracture surface is shown at both lower and higher magnifications in Fig. 7, where fiber pullout and fracture in matrix at different planes are clearly evident as well as the presence of fractured BN interphase (marked by arrows as few such examples) along with debonding between fiber and interphase or between interphase and SiC matrix. The comparison of the typical fractured surfaces from stress rupture tests under both laboratory air and steam test conditions are shown in Figs. 8–10. These show images for tests at 400 C, 750 C, and 950 C (from left to right) under steam and laboratory air conditions. Fig. 8 shows these at lower magnification (100·) of one representative sample from each test environment. It can be seen that the laboratory air tests show a relatively rougher fracture plane compared to the steam tests. Also, the tests conducted at 750 C, in the intermediate range, have relatively flatter fracture surface with very little fiber pullout (Fig. 8b and e) whereas the 400 C and 950 C tests show pullout of longitudinal tows and planar fracture within the tows (Fig. 8a, d, c and f). This indicates that the cracks 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Temperature Normalized Residual Strength Air Steam 4000C 7500C 9500C Fig. 6. Normalized residual strength. Fig. 7. Fracture Surface from monotonic tensile test at 950 C, (a) at lower magnification (100·) and (b) at higher magnification (8000·). 278 S. Mall, J.L. Ryba / Composites Science and Technology 68 (2008) 274–282
S. Mall,J.L Ryba Composites Science and Technology 68(2008)274-28 279 400c.02h(a 8. Fracture surface from stress rupture tests at lower magnification (100x):(a)400C steam, (b)750C steam, (c)950C steam (d)400C air,(e) eair,(f)950°air 50c1510he Ambient Air Conditions Fig 9. Fracture surface from stress rupture tests at intermediate magnification(1000x): (a)400C steam.(b)750C steam, (c)950C steam (d)400C air,(e)750°cair,()950°air
Fig. 8. Fracture surface from stress rupture tests at lower magnification (100·): (a) 400 C steam, (b) 750 C steam, (c) 950 C steam, (d) 400 C air, (e) 750 C air, (f) 950 C air. Fig. 9. Fracture surface from stress rupture tests at intermediate magnification (1000·): (a) 400 C steam, (b) 750 C steam, (c) 950 C steam, (d) 400 C air, (e) 750 C air, (f) 950 C air. S. Mall, J.L. Ryba / Composites Science and Technology 68 (2008) 274–282 279
280 S. Mall,J.L Ryba / Composites Science and Technology 68(2008)274-282 T 400C 21.02(ay A30C19S(1 950c1:3-(c Steam Conditions E,0 0c1510h(e Fig. 10. Fracture surface from stress rupture tests at higher magnification(8000x) :(a)400C steam, (b)750C steam, (c)950C steam, (d)400C air, (e) 750°Cair,()950°C Steam Conditions (01 Ambient Air condition Fig.Il. Residual strength fracture surfaces.(a)40° C steam,(b)70° C stean,(c)950° C stean,(d)400°cair,(e)750° air and()950°C
Fig. 10. Fracture surface from stress rupture tests at higher magnification (8000·): (a) 400 C steam, (b) 750 C steam, (c) 950 C steam, (d) 400 C air, (e) 750 C air, (f) 950 C air. Fig. 11. Residual strength fracture surfaces. (a) 400 C steam, (b) 750 C steam, (c) 950 C steam, (d) 400 C air, (e) 750 C air and (f) 950 C air. 280 S. Mall, J.L. Ryba / Composites Science and Technology 68 (2008) 274–282
S. Mall, J. L Ryba/ Composites Science and Technology 68(2008)274-28 continued through the specimen virtually unimpeded in the condition at a given temperature, and likewise relatively intermediate temperature range tests while in the 400C more at 950C than at 400C in a given test environment. and 950C tests the fibers diverted the cracks for a short However, a different damage mechanism of the Bn inter- period before the entire tow failed phase is present at 750C. The bn interphase near the Fig.9 shows the fracture surface of the representative fiber/ matrix or fiber/fiber locations appears to have been specimens from all six stress rupture test environments at oxidized by the moisture under humid test environment. intermediate magnification(1000x). At this magnification, This occurred most probably due to the oxidation of Bn it can be seen that there is exterior and interior fiber deb- to form the boria(B2O3), which reacted with SiC to form onding, and fiber pullout and fracture along different a borosilicate melt. It then solidified where boron was lea planes at 400C(Fig. 9a and d) and 950C(Fig. 9c ched out from the condensed phase resulting in a glass with and f), i.e., below and above the intermediate range. There a very high SiO, content. This fused fibers together caused is very minimal fiber pullout with the fracture appearing the embrittlement(Fig. 10b and e), i.e., failure on the same lmost completely planar at 750C, i.e., within the inter- plane with no fiber pull-out, again a typical feature of fiber/ mediate range(Fig. 9b and e) matrix embrittlement. The planar fracture surface at At higher magnification(8000x) shown in Fig. 10 the 750C, as elaborated earlier, was due to the bonding details of degradation of the BN interphase material can together of the fibers and matrix by the silicate glass, so be observed. At 400C and 950C, the bn interphase that as the weakest fiber in a tow failed, the rest of the shows the evidence of fracture or recession(marked by tow also failed with it. These microscopic analyses thus arrows as few such examples). This is relatively more prom- clearly suggests that the tested CMC system experienced inent in the steam environment than the laboratory air test more embrittlement of fiber/ matrix interphase in the inter a)br. Monotonic Test-T 「(b)m118h 了(①曙12.31h 10 LmI Fig 12. Damage progression with time in 950C steam stress rupture tests-(a)5min, (b)1. 18 h,(c)4.2 h and (d)12.3 h
continued through the specimen virtually unimpeded in the intermediate temperature range tests while in the 400 C and 950 C tests the fibers diverted the cracks for a short period before the entire tow failed. Fig. 9 shows the fracture surface of the representative specimens from all six stress rupture test environments at intermediate magnification (1000·). At this magnification, it can be seen that there is exterior and interior fiber debonding, and fiber pullout and fracture along different planes at 400 C (Fig. 9a and d) and 950 C (Fig. 9c and f), i.e., below and above the intermediate range. There is very minimal fiber pullout with the fracture appearing almost completely planar at 750 C, i.e., within the intermediate range (Fig. 9b and e). At higher magnification (8000·) shown in Fig. 10 the details of degradation of the BN interphase material can be observed. At 400 C and 950 C, the BN interphase shows the evidence of fracture or recession (marked by arrows as few such examples). This is relatively more prominent in the steam environment than the laboratory air test condition at a given temperature, and likewise relatively more at 950 C than at 400 C in a given test environment. However, a different damage mechanism of the BN interphase is present at 750 C. The BN interphase near the fiber/matrix or fiber/fiber locations appears to have been oxidized by the moisture under humid test environment. This occurred most probably due to the oxidation of BN to form the boria (B2O3), which reacted with SiC to form a borosilicate melt. It then solidified where boron was leached out from the condensed phase resulting in a glass with a very high SiO2 content. This fused fibers together caused the embrittlement (Fig. 10b and e), i.e., failure on the same plane with no fiber pull-out, again a typical feature of fiber/ matrix embrittlement. The planar fracture surface at 750 C, as elaborated earlier, was due to the bonding together of the fibers and matrix by the silicate glass, so that as the weakest fiber in a tow failed, the rest of the tow also failed with it. These microscopic analyses thus clearly suggests that the tested CMC system experienced more embrittlement of fiber/matrix interphase in the interFig. 12. Damage progression with time in 950 C steam stress rupture tests – (a) 5 min, (b) 1.18 h, (c) 4.2 h and (d) 12.3 h. S. Mall, J.L. Ryba / Composites Science and Technology 68 (2008) 274–282 281
S. Mall,J.L Ryba / Composites Science and Technology 68(2008)274-282 mediate temperature range, which in turn caused a rela- Acknowledgements ively more reduction in stress rupture performance relative to that at below or above the intermediate temperature test The support of Ruth Sikorski, Propulsion Directorate, condition Air Force Research Laboratory, Wright-Patterson AFB Similar characteristics were observed in the fracture sur- as well as Drs. Jim DiCarlo and Greg Morscher, Ceramics faces of the residual strength tests(Fig. 11). The fracture Branch, NASA-Glenn Research Center is highly surfaces are relatively rougher in tests at 400C and appreciated 950C i.e outside the intermediate range(Fig. lla, c, d and f) than the more planar fracture in tests at 750"C References within the intermediate range(Fig. llb and e). Finally, 1g. 12 shows the effect of the exposure time on the damage []DiCarlo JA, Dutta S. Continuous ceramics fibers for progression. This is shown for one test condition, at 950C composites. In: Lehman R, El-Rahaiby S, Wachtman J in the steam environment, for the sake of brevity. This Handbook on continuous fiber reinforced ceramic matrix composites. clearly shows that damage of BN interphase increased with CIAC, Purdue University: 1995. p. 137-83 test exposure duration, and it is already prevalent at about 22] Ichikawa H, Ishikawa. Slicon carbide fibers (organometa 4h, or in-between l and 4 h. In summary, the micro- opic analysis suggests that the tested CMC composite [3] Yun HM, DiCarlo JA. Comparison of tensile, creep and rupture experienced relatively more embrittlement of the fiber/ trength properties of stoichiometric sic fibers. In: Proceedings of the matrix interphase in the presence of moisture and at the 3rd annual conference on composites, materials and structures, voL. intermediate temperature range. This resulted in relatively 20, Cocoa Beach;1999,p.259-72. more reduction in stress rupture performance in these test [4] Brennan JJ. Interfacial characterization of glass and matrix/nicalon SiC fiber composites. Mater Sci Res I 5] Cooper RF, Chyung K. Structure and chemistry of ceramics. J Mater Sci 1987- 22-3148-60. 4. Conclusions 6] Cao H, Bischoff E, Sbaizero O, Ruhle M, Evans AG, Marshall DB, et al. Effect of interfaces on the properties of fiber-reinforced The stress rupture behavior of silicon carbide fiber(Syl) ceramics. J Am Ceram Soc 1990: 73: 1691-9 [7 Morscher GN, Hurst J, Brewer D. intermediate-temperature stress reinforced in silicon carbide(SiC)matrix with a boron rupture of a woven Hi-nicalon, BN-interphase, SiC matrix in air. J nitride(BN) interphase( Syl-iBN/BN/SiC)was invest Am Ceram Soc 2000: 83: 1441-9. at three temperatures,400°C,750°Cand950°%C, [8] Morscher GN. Tensile stress rupture of SiC/SiCm minicomposites laboratory air and 100% steam test environments with carbon and boron nitride interphases at elevated temperatures in air. J Am Ceram Soc 1997: 80: 2029-42 rupture strength versus time to failure relationships were 9]Lara-Curzio E, Ferber MK, Tortorelli PE. Interface established under these six elevated temperature and envi- stress-rupture of Nicalon SiC CVCCs at Intermed ronment conditions. Time to failure in the steam test envi- tures. Key engineerin publica. ronment was shorter than that in the laboratory air test tions;1997.10691082. environment at a given stress level and at an elevated tem- [Io] Lara-Curzio E. stress-rupture of n/Sic continuous fiber °C. J Am Ceram soc perature. In addition time to failure decreased as the test an enhanced temperatures. J Am Ceram BN interphase between the fiber and the matrix. The pres- Soc1998;81:2140 ence of more moisture in the steam environment test condi- [12] Heredia FE, McNulty JC, Zok FW, Evans AG. Oxidation embrit- tion accelerated this process, and therefore it caused more tlement probe for ceramic-matrix composites. J Am Ceram Soc 199578:2097-10 reduction in stress rupture performance of the tested CMC [13] Brewer D. HSR/EPM combustor materials development program system in the steam test environment than in the laboratory Mater Sci Eng A 1999: 261: 284-94 air test environment at a given elevated temperature. Fur- [14] Yun HM, Gyenkenyesi JZ, Chen YL, Wheeler DR, DiCarlo JA ther, the tests conducted at 750C experienced relatively Tensile behavior of SiC/SiC composites reinforced by more damage than expected which resulted in more reduc sylramic fibers. In: Proceedings of the 25th annual confer composites, materials and structures, voL 20, Cocoa Beach; tion in stress rupture performance than the interpolated performance between 400C and 950C. Damage analysis [15] LaRochelle KJ, Mall S Temperature and moisture effects upon stress showed the degradation of BN interphase between fiber upture life of Syl-iBN/BN/SiC composites. Ceram Eng Sci Proc and matrix; however it was of different type at 400C or 2003:24:45964 950C versus 750C. It was in the form of fracture, reces- [16] Mall S Effects of moisture on fatigue behavior of SiC/SiC composite at elevated temperature. Mater Sci Eng: A 2005: 412: 165- sion and removal of Bn interphase at 400C or 950C. On [17] Mall S, LaRochelle KJ. Fatigue and stress-rupture behaviors of Sic/ the other hand. there was formation of borosilicate which Sic composite under humid environment at elevated temperature. caused pesting of fibers at 750C Compos Sci Technol 2006: 66: 2925-34
mediate temperature range, which in turn caused a relatively more reduction in stress rupture performance relative to that at below or above the intermediate temperature test condition. Similar characteristics were observed in the fracture surfaces of the residual strength tests (Fig. 11). The fracture surfaces are relatively rougher in tests at 400 C and 950 C i.e., outside the intermediate range (Fig. 11a, c, d and f) than the more planar fracture in tests at 750 C within the intermediate range (Fig. 11b and e). Finally, Fig. 12 shows the effect of the exposure time on the damage progression. This is shown for one test condition, at 950 C in the steam environment, for the sake of brevity. This clearly shows that damage of BN interphase increased with test exposure duration, and it is already prevalent at about 4 h, or in-between 1 and 4 h. In summary, the microscopic analysis suggests that the tested CMC composite experienced relatively more embrittlement of the fiber/ matrix interphase in the presence of moisture and at the intermediate temperature range. This resulted in relatively more reduction in stress rupture performance in these test conditions. 4. Conclusions The stress rupture behavior of silicon carbide fiber (Syl) reinforced in silicon carbide (SiC) matrix with a boron nitride (BN) interphase (Syl-iBN/BN/SiC) was investigated at three temperatures, 400 C, 750 C and 950 C, under laboratory air and 100% steam test environments. Stress rupture strength versus time to failure relationships were established under these six elevated temperature and environment conditions. Time to failure in the steam test environment was shorter than that in the laboratory air test environment at a given stress level and at an elevated temperature. In addition time to failure decreased as the test temperature increased in a given test environment. The primary damage mechanism involved the degradation of the BN interphase between the fiber and the matrix. The presence of more moisture in the steam environment test condition accelerated this process, and therefore it caused more reduction in stress rupture performance of the tested CMC system in the steam test environment than in the laboratory air test environment at a given elevated temperature. Further, the tests conducted at 750 C experienced relatively more damage than expected which resulted in more reduction in stress rupture performance than the interpolated performance between 400 C and 950 C. Damage analysis showed the degradation of BN interphase between fiber and matrix; however it was of different type at 400 C or 950 C versus 750 C. It was in the form of fracture, recession and removal of BN interphase at 400 C or 950 C. On the other hand, there was formation of borosilicate which caused pesting of fibers at 750 C. Acknowledgements The support of Ruth Sikorski, Propulsion Directorate, Air Force Research Laboratory, Wright-Patterson AFB as well as Drs. Jim DiCarlo and Greg Morscher, Ceramics Branch, NASA-Glenn Research Center is highly appreciated. References [1] DiCarlo JA, Dutta S. Continuous ceramics fibers for ceramic composites. In: Lehman R, El-Rahaiby S, Wachtman J, editors. Handbook on continuous fiber reinforced ceramic matrix composites. CIAC, Purdue University; 1995. p. 137–83. [2] Ichikawa H, Ishikawa T. Silicon carbide fibers (organometallic pyrolysis). In: Kelly A, Zweben C, Chou T, editors. Comprehensive composite materials, vol. I. Elsevier Science; 2000. p. 107–45. [3] Yun HM, DiCarlo JA. Comparison of tensile, creep and rupture strength properties of stoichiometric sic fibers. In: Proceedings of the 23rd annual conference on composites, materials and structures, vol. 20, Cocoa Beach; 1999, p. 259–72. [4] Brennan JJ. Interfacial characterization of glass and glass-ceramic matrix/nicalon SiC fiber composites. Mater Sci Res 1986;20:546–60. [5] Cooper RF, Chyung K. Structure and chemistry of fiber-reinforced ceramics. J Mater Sci 1987;22:3148–60. [6] Cao H, Bischoff E, Sbaizero O, Ruhle M, Evans AG, Marshall DB, et al. Effect of interfaces on the properties of fiber-reinforced ceramics. J Am Ceram Soc 1990;73:1691–9. [7] Morscher GN, Hurst J, Brewer D. intermediate-temperature stress rupture of a woven Hi-nicalon, BN-interphase, SiC matrix in air. J Am Ceram Soc 2000;83:1441–9. [8] Morscher GN. Tensile stress rupture of SiCf/SiCm minicomposites with carbon and boron nitride interphases at elevated temperatures in air. J Am Ceram Soc 1997;80:2029–42. [9] Lara-Curzio E, Ferber MK, Tortorelli PF. Interface oxidation and stress-rupture of NicalonTM/SiC CVCCs at Intermediate Temperatures. Key engineering materials. Switzerland: Trans Tech Publications; 1997. 1069–1082. [10] Lara-Curzio E. stress-rupture of Nicalon/SiC continuous fiber ceramic matrix composites in air at 950 C. J Am Ceram Soc 1997;80:3268–72. [11] Steyer TE, Zok FW, Walls DP. Stress rupture of an enhanced NicalonTM/SiC composite at intermediate temperatures. J Am Ceram Soc 1998;81:2140–6. [12] Heredia FE, McNulty JC, Zok FW, Evans AG. Oxidation embrittlement probe for ceramic-matrix composites. J Am Ceram Soc 1995;78:2097–100. [13] Brewer D. HSR/EPM combustor materials development program. Mater Sci Eng A 1999;261:284–94. [14] Yun HM, Gyenkenyesi JZ, Chen YL, Wheeler DR, DiCarlo JA. Tensile behavior of SiC/SiC composites reinforced by treated sylramic fibers. In: Proceedings of the 25th annual conference on composites, materials and structures, vol. 20, Cocoa Beach; 2001, p. 521–31. [15] LaRochelle KJ, Mall S. Temperature and moisture effects upon stress rupture life of Syl-iBN/BN/SiC composites. Ceram Eng Sci Proc 2003;24:459–64. [16] Mall S. Effects of moisture on fatigue behavior of SiC/SiC composite at elevated temperature. Mater Sci Eng: A 2005;412:165–70. [17] Mall S, LaRochelle KJ. Fatigue and stress-rupture behaviors of SiC/ SiC composite under humid environment at elevated temperature. Compos Sci Technol 2006;66:2925–34. 282 S. Mall, J.L. Ryba / Composites Science and Technology 68 (2008) 274–282