Availableonlineatwww.sciencedirect.com °scⅰ ence Direct c000055 Part A: applied science and manufacturing ELSEVIER Composites: Part A 37(2006)2029-2040 www.elsevier.com/locate/compositesa Effects of steam environment on high-temperature mechanica behavior of Nextel M720/alumina(N720/A)continuous fiber ceramic composite M B. Ruggles-Wrenn", S Mall, C.A. Eber, L B. Harlan Department of Aeronautics and Astronautics, Air Force Institute of Technology, WPAFB, OH 45433-7765, US.A Received 15 February 2005: received in revised form 30 November 2005; accepted 13 December 2005 Abstract Mechanical behavior of an oxide-oxide continuous fiber ceramic composite( CFCC)consisting of a porous alumina matrix reinforced with laminated, woven mullite/alumina fibers(Nextel M720)was investigated at 1200 and 1330oC in laboratory air and in 100%steam environments. CFCC has no interface between the fiber and matrix, and relies on the porous matrix for flaw tolerance Tension-tension fatigue behavior was studied for fatigue stresses ranging from 100 to 170 MPa at 1200C, and for fatigue stresses of 50 and 100 MPa at 1330C. Tensile creep behavior was examined for creep stresses ranging from 80 to 154 MPa at 1200C, and for creep stresses of 50 and 100 MPa at 1330C. At 1200C, the CFCC exhibited excellent fatigue resistance in laboratory air. The fatigue limit(based on a run-out condition of 10 cycles)was 170 MPa(88% UTS at 1200C). The material retained 100% of its tensile strength. Presence of steam caused noticeable degradation in fatigue performance at 1200C Fatigue resistance at 1330C was poor In creep tests, primary and secondary creep regimes were observed. Minimum creep rate was reached in all tests. At 1200oC, creep rates were 10-8-10-s-I and maximum time to rupture was 255 h. At 1330oC, creep rates were 10-7-10-5s-I and maximum time to rupture was 87 h Presence of steam accel erated creep rates and dramatically reduced creep life Published by elsevier Ltd Keywords: A Ceramic-matrix composites(CMCs); A Fibres; A Creep 1. Introduction aerospace applications. Compared to the conventional nickel-based superalloys, CMCs offer improved high- Aerospace components require structural materials that temperature performance at reduced weight. Advanced have superior long-term mechanical properties and can be reusable space launch vehicles will likely incorporate exposed to severe environmental conditions, such as high fiber-reinforced CMCs in critical propulsion components temperature, high pressure, or water vapor. Ceramic- In these applications, CMCs will be subjected to mechani- matrix composites(CMCs), capable of maintaining excel- cal loading in complex environments. For example, a typ- lent strength and fracture toughness at high temperatures ical service environment for a reusable rocket engine continue to attract attention as candidate materials for turbopump rotor includes hydrogen, oxygen and steam, at pressures >200 atm [l]. Ceramic-matrix composites are also being considered for aerospace turbine engine applica- The views expressed are those of the authors and do not reflect the tions. Higher material operating temperatures and official policy or position of the United States Air Force, Department of decreased cooling air requirement are the significant Defense or the u.s. government Corresponding author. Tel. +937 255 3636x4641; fax: +937656 4032 advantages that CMCs offer to the aerospace engine design E-mailaddress:marina.ruggles-wrenn(@afit.edu(M.B.Ruggles.community.However,theseapplication sure to oxidizing environments. Recently CMCs have been 1359-835X/S.see front matter Published by Elsevier Ltd doi:10.1016/j.compositesa.2005.12.008
Effects of steam environment on high-temperature mechanical behavior of NextelTM720/alumina (N720/A) continuous fiber ceramic composite q M.B. Ruggles-Wrenn *, S. Mall, C.A. Eber, L.B. Harlan Department of Aeronautics and Astronautics, Air Force Institute of Technology, WPAFB, OH 45433-7765, USA Received 15 February 2005; received in revised form 30 November 2005; accepted 13 December 2005 Abstract Mechanical behavior of an oxide–oxide continuous fiber ceramic composite (CFCC) consisting of a porous alumina matrix reinforced with laminated, woven mullite/alumina fibers (NextelTM720) was investigated at 1200 and 1330 C in laboratory air and in 100% steam environments. CFCC has no interface between the fiber and matrix, and relies on the porous matrix for flaw tolerance. Tension–tension fatigue behavior was studied for fatigue stresses ranging from 100 to 170 MPa at 1200 C, and for fatigue stresses of 50 and 100 MPa at 1330 C. Tensile creep behavior was examined for creep stresses ranging from 80 to 154 MPa at 1200 C, and for creep stresses of 50 and 100 MPa at 1330 C. At 1200 C, the CFCC exhibited excellent fatigue resistance in laboratory air. The fatigue limit (based on a run-out condition of 105 cycles) was 170 MPa (88% UTS at 1200 C). The material retained 100% of its tensile strength. Presence of steam caused noticeable degradation in fatigue performance at 1200 C. Fatigue resistance at 1330 C was poor. In creep tests, primary and secondary creep regimes were observed. Minimum creep rate was reached in all tests. At 1200 C, creep rates were 108 –105 s 1 and maximum time to rupture was 255 h. At 1330 C, creep rates were 107 –105 s 1 and maximum time to rupture was 87 h. Presence of steam accelerated creep rates and dramatically reduced creep life. Published by Elsevier Ltd. Keywords: A. Ceramic-matrix composites (CMCs); A. Fibres; A. Creep 1. Introduction Aerospace components require structural materials that have superior long-term mechanical properties and can be exposed to severe environmental conditions, such as high temperature, high pressure, or water vapor. Ceramic– matrix composites (CMCs), capable of maintaining excellent strength and fracture toughness at high temperatures continue to attract attention as candidate materials for aerospace applications. Compared to the conventional nickel-based superalloys, CMCs offer improved hightemperature performance at reduced weight. Advanced reusable space launch vehicles will likely incorporate fiber-reinforced CMCs in critical propulsion components. In these applications, CMCs will be subjected to mechanical loading in complex environments. For example, a typical service environment for a reusable rocket engine turbopump rotor includes hydrogen, oxygen and steam, at pressures >200 atm [1]. Ceramic–matrix composites are also being considered for aerospace turbine engine applications. Higher material operating temperatures and decreased cooling air requirement are the significant advantages that CMCs offer to the aerospace engine design community. However, these applications also require exposure to oxidizing environments. Recently CMCs have been 1359-835X/$ - see front matter Published by Elsevier Ltd. doi:10.1016/j.compositesa.2005.12.008 q The views expressed 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 U.S. Government. * Corresponding author. Tel.: +937 255 3636x4641; fax: +937 656 4032. E-mail address: marina.ruggles-wrenn@afit.edu (M.B. RugglesWrenn). www.elsevier.com/locate/compositesa Composites: Part A 37 (2006) 2029–2040
B. Ruggles-Wrenn et al. Composites: Part A 37(2006)2029-2040 demonstrated in various turbine components [2]. Many of of intermittent moisture exposure on the high-temperature these demonstration components have exhibited acceler- fatigue durability of five different CMCs, including N720/ ated degradation after only a few hours in service environ- A Fatigue testing of N720 /A was performed at 1200C, ment. It is now widely recognized that the thermodynamic fatigue stress levels were <120 MPa. Cyclic loading and stability and oxidation resistance of CMCs have become moisture exposure were applied alternately Zawada et al. Important issues. observed no degradation in fatigue performance or Sintered structural ceramics are known to exhibit degra- retained strength with intermittent moisture exposure dation in high-temperature environments. Non-oxide fiber/ In the present study, fatigue and creep-rupture testing of non-oxide matrix composites generally show poor oxida- N720/A specimens was conducted both in laboratory air tion resistance [3, 4]. The degradation involves oxidation and in 100% steam environment at high temperatures of fibers and fiber coatings, and is accelerated by the pres-(1200 and 1330C). Applied stress levels used in both ence of moisture [5-7]. Numerous studies addressed oxida- creep-rupture and fatigue tests(<170 MPa) were consider- tion of Sic in moist environments [8-13]. Opila and Hann ably higher than those employed in previous studies. As is [9] and Pila [10, 11] reported that the presence of water seen in detail, effects of steam environment on fatigue and vapor increased the rate of Sio growth on SiC at high especially on creep performance cannot be neglected temperature, which led to accelerated rates of Sic reces- sion. Degradation of BN fiber coatings in moist environ- 2. Experimental procedure ments has also been a subject of extensive research [14-20]. Non-oxide fiber/oxide matrix composites or oxide 2.1. Material fiber/non-oxide matrix composites do not exhibit high oxidation resistance either. For these materials. the high The composite studied was a commercially available permeability constant for the diffusion of oxygen results material(N720/A, COI Ceramics, San Diego, CA)consist- in rapid oxygen permeation through the oxide matrix ing of NextelM720 fibers in a porous alumina matrix, sup- [21]. These considerations motivated the development of plied in a form of 2.8 mm thick plates. The plates consisted environmentally stable or: Composites(CFCCs) based on of 12 0/900 woven layers, with a density of 2.78g/cm The main advantage of CMCs over monolithic ceramics fiber coating. The fiber fabric was infiltrated with the is their superior toughness, tolerance to the presence of matrix in a sol-gel process. After drying with a"vacuum cracks and defects, and non-catastrophic mode of failure. bag"technique under low pressure and low temperature, It is now well recognized that CFCCs can be designed to the composite was pressureless sintered [49]. Matrix poros- exhibit non-brittle fracture behavior and improved damage ity was w24%. Such porosity level renders the matrix suffi- tolerance by introducing a weak fiber/matrix interface, ciently weak and gives the composite excellent damage which serves to deflect matrix cracks and to allow subse- tolerance during loading. Representative micrographs of quent fiber pull-out [31-33]. It has recently been demon- the untested as-received material are shown in Fig. I strated that similar crack-deflecting behavior can also be Fig. l(a) shows 0 and 90 fiber tows as well as numerous achieved by means of a finely distributed porosity in the matrix cracks. In the case of untested material, most are matrix instead of a separate interface between matrix and shrinkage cracks formed during processing rather than fibers [34]. This microstructural design philosophy implic- matrix cracks generated during loading. Porous nature of itly accepts the formation of strong interfaces. It builds the matrix is seen in Fig. on the experience with porous interlayers as crack deflec tion paths [35, 36] and extends the concept to utilize a por- 2. 2. Mechanical testing ous matrix as a surrogate. The concept has been successfully demonstrated for oxide-oxide composites A servocontrolled MTS mechanical testing machine [22, 26, 30, 37-41]. Resulting oxide/oxide CFCCs exhibit equipped with hydraulic water-cooled collet grips, a com damage tolerance combined with inherent oxidation resis- pact two-zone resistance-heated furnace, and two tempera tance. An extensive review of the mechanisms and mechat ture controllers was used in all tests. An mrs Teststar ical properties of porous-matrix CMCs is given in [42]. digital controller was employed for input signal generation The objective of this study is to investigate effects of and data acquisition. Strain measurement was accom- steam environment on high-temperature mechanical plished with an MTS high-temperature air-cooled uniaxial behavior and durability of an oxide-oxide CFCC, consist- extensometer. For elevated temperature testing, thermo- ing of a porous alumina matrix reinforced with the couples were bonded to the specimens to calibrate the fur- Nextel720 fibers. Several previous studies examined nace on a periodic basis. The furnace controller(using a high-temperature mechanical behavior of this material non-contacting thermocouple exposed to the ambient envi [2, 43], and [44]. Unlike its counterpart reinforced with ronment near the test specimen) was adjusted to determine Sic fibers [45-48], the CFCC exhibited steady-state creep. the power setting needed to achieve the desired tempera- Creep rates were 10-8-10-7s-l, similar to those expected ture of the test specimen. Thus determined power setting from fibers alone. Zawada et al. [2] investigated the effect was then used in actual tests. The power setting for testing
demonstrated in various turbine components [2]. Many of these demonstration components have exhibited accelerated degradation after only a few hours in service environment. It is now widely recognized that the thermodynamic stability and oxidation resistance of CMCs have become important issues. Sintered structural ceramics are known to exhibit degradation in high-temperature environments. Non-oxide fiber/ non-oxide matrix composites generally show poor oxidation resistance [3,4]. The degradation involves oxidation of fibers and fiber coatings, and is accelerated by the presence of moisture [5–7]. Numerous studies addressed oxidation of SiC in moist environments [8–13]. Opila and Hann [9] and Opila [10,11] reported that the presence of water vapor increased the rate of SiO2 growth on SiC at high temperature, which led to accelerated rates of SiC recession. Degradation of BN fiber coatings in moist environments has also been a subject of extensive research [14–20]. Non-oxide fiber/oxide matrix composites or oxide fiber/non-oxide matrix composites do not exhibit high oxidation resistance either. For these materials, the high permeability constant for the diffusion of oxygen results in rapid oxygen permeation through the oxide matrix [21]. These considerations motivated the development of continuous fiber ceramic composites (CFCCs) based on environmentally stable oxide constituents [22–30]. The main advantage of CMCs over monolithic ceramics is their superior toughness, tolerance to the presence of cracks and defects, and non-catastrophic mode of failure. It is now well recognized that CFCCs can be designed to exhibit non-brittle fracture behavior and improved damage tolerance by introducing a weak fiber/matrix interface, which serves to deflect matrix cracks and to allow subsequent fiber pull-out [31–33]. It has recently been demonstrated that similar crack-deflecting behavior can also be achieved by means of a finely distributed porosity in the matrix instead of a separate interface between matrix and fibers [34]. This microstructural design philosophy implicitly accepts the formation of strong interfaces. It builds on the experience with porous interlayers as crack deflection paths [35,36] and extends the concept to utilize a porous matrix as a surrogate. The concept has been successfully demonstrated for oxide–oxide composites [22,26,30,37–41]. Resulting oxide/oxide CFCCs exhibit damage tolerance combined with inherent oxidation resistance. An extensive review of the mechanisms and mechanical properties of porous-matrix CMCs is given in [42]. The objective of this study is to investigate effects of steam environment on high-temperature mechanical behavior and durability of an oxide–oxide CFCC, consisting of a porous alumina matrix reinforced with the NextelTM720 fibers. Several previous studies examined high-temperature mechanical behavior of this material [2,43], and [44]. Unlike its counterpart reinforced with SiC fibers [45–48], the CFCC exhibited steady-state creep. Creep rates were 108 –107 s 1 , similar to those expected from fibers alone. Zawada et al. [2] investigated the effect of intermittent moisture exposure on the high-temperature fatigue durability of five different CMCs, including N720/ A. Fatigue testing of N720/A was performed at 1200 C, fatigue stress levels were 6120 MPa. Cyclic loading and moisture exposure were applied alternately. Zawada et al. observed no degradation in fatigue performance or retained strength with intermittent moisture exposure. In the present study, fatigue and creep-rupture testing of N720/A specimens was conducted both in laboratory air and in 100% steam environment at high temperatures (1200 and 1330 C). Applied stress levels used in both creep-rupture and fatigue tests (6170 MPa) were considerably higher than those employed in previous studies. As is seen in detail, effects of steam environment on fatigue and especially on creep performance cannot be neglected. 2. Experimental procedure 2.1. Material The composite studied was a commercially available material (N720/A, COI Ceramics, San Diego, CA) consisting of NextelTM720 fibers in a porous alumina matrix, supplied in a form of 2.8 mm thick plates. The plates consisted of 12 0/90 woven layers, with a density of 2.78 g/cm3 and a fiber volume of approximately 44%. There was no fiber coating. The fiber fabric was infiltrated with the matrix in a sol–gel process. After drying with a ‘‘vacuum bag’’ technique under low pressure and low temperature, the composite was pressureless sintered [49]. Matrix porosity was 24%. Such porosity level renders the matrix suffi- ciently weak and gives the composite excellent damage tolerance during loading. Representative micrographs of the untested as-received material are shown in Fig. 1. Fig. 1(a) shows 0 and 90 fiber tows as well as numerous matrix cracks. In the case of untested material, most are shrinkage cracks formed during processing rather than matrix cracks generated during loading. Porous nature of the matrix is seen in Fig. 1(b). 2.2. Mechanical testing A servocontrolled MTS mechanical testing machine equipped with hydraulic water-cooled collet grips, a compact two-zone resistance-heated furnace, and two temperature controllers was used in all tests. An MTS TestStar digital controller was employed for input signal generation and data acquisition. Strain measurement was accomplished with an MTS high-temperature air-cooled uniaxial extensometer. For elevated temperature testing, thermocouples were bonded to the specimens to calibrate the furnace on a periodic basis. The furnace controller (using a non-contacting thermocouple exposed to the ambient environment near the test specimen) was adjusted to determine the power setting needed to achieve the desired temperature of the test specimen. Thus determined power setting was then used in actual tests. The power setting for testing 2030 M.B. Ruggles-Wrenn et al. / Composites: Part A 37 (2006) 2029–2040
M B. Ruggles-Wrenn et al / Composites: Part 437(2006)2029-2040 Creep-rupture tests were conducted in load control in accordance with the procedure in ASTM standard C 1337. In each test, stress-strain data were recorded during the loading to the creep stress level and the actual creep period. Thus both total strain and creep strain could be cal culated and examined. Creep-rupture tests were carried out at 1200 and 1330C, in laboratory air and in steam environments Tension-tension fatigue tests were carried out in load control with an R ratio(minimum stress divided by the maximum stress)of 0.05 at a frequency of 1 Hz. Fatigue run-out was set to 10 cycles. The 10 cycle count value rep- resents the number of loading cycles expected in aerospace applications at that temperature. Fatigue run-out limits were defined as the highest stress level. for which run-out throughout each test. In order to assess the damage devel opment in the composite, stiffness hysteresis energy density, as well as strain accumulation with fatigue cycles were examined. To determine retained strength and stiffness, specimens that achieved run-out were subjected to tensile test to failure at the temperature of the fatigue test. Fatigue tests were performed at 1200 and 1330C, in laboratory air and in steam environments 23. characterization Fracture surfaces of failed specimens were examined using SEM (Model 360FE, Leica)as well as optical micros- copy. The SEM specimens were gold coated 10.0p 3. Results and discussion 3.. Monotonic tension porous nature of the matrix is evident(SEM) Tensile stress-strain behavior at 23. 1200 and 1330 C is in steam environment was determined by placing the spec- shown in Fig. 2. The stress-strain curves obtained at 23 imen instrumented with thermocouples in 100% steam and 1200C are nearly linear to failure. Such linear behav- environment and repeating the furnace calibration proce- ior indicates that there is little additional matrix cracking dure. Thermocouples were not bonded to the test speci mens after the furnace was calibrated. Tests in steam environment employed an alumina susceptor(tube with end caps), which fits inside the furnace. The specimen gage section is located inside the susceptor, with the ends of the 200[1200c specimen passing through slots in the susceptor. Steam is a continuous stream with a slightly positive pressure, expe- 150| introduced into the susceptor(through a feeding tube)in 23c 1330c ling the dry air and creating 100% steam environment inside the susceptor In all tests, a specimen was heated to test temperature in 25 min, and held at temperature for additional 15 min prior to testing. Dog bone shaped specimens of 152 mm total length with a 10-mm-wide gage section were used in all .000250500.751.001.251.501.75200 Tensile tests were performed in stroke control with a Strain(%) constant displacement rate of 0.05 mm/s. Tensile tests were Fig. 2. Tensile stress-strain curves for Nextel TM720/alumina ceramic conducted at 23, 1200 and 1330C in laboratory air. composite at 23, 1200 and 1330C
in steam environment was determined by placing the specimen instrumented with thermocouples in 100% steam environment and repeating the furnace calibration procedure. Thermocouples were not bonded to the test specimens after the furnace was calibrated. Tests in steam environment employed an alumina susceptor (tube with end caps), which fits inside the furnace. The specimen gage section is located inside the susceptor, with the ends of the specimen passing through slots in the susceptor. Steam is introduced into the susceptor (through a feeding tube) in a continuous stream with a slightly positive pressure, expelling the dry air and creating 100% steam environment inside the susceptor. In all tests, a specimen was heated to test temperature in 25 min, and held at temperature for additional 15 min prior to testing. Dog bone shaped specimens of 152 mm total length with a 10-mm-wide gage section were used in all tests. Tensile tests were performed in stroke control with a constant displacement rate of 0.05 mm/s. Tensile tests were conducted at 23, 1200 and 1330 C in laboratory air. Creep-rupture tests were conducted in load control in accordance with the procedure in ASTM standard C 1337. In each test, stress–strain data were recorded during the loading to the creep stress level and the actual creep period. Thus both total strain and creep strain could be calculated and examined. Creep-rupture tests were carried out at 1200 and 1330 C, in laboratory air and in steam environments. Tension–tension fatigue tests were carried out in load control with an R ratio (minimum stress divided by the maximum stress) of 0.05 at a frequency of 1 Hz. Fatigue run-out was set to 105 cycles. The 105 cycle count value represents the number of loading cycles expected in aerospace applications at that temperature. Fatigue run-out limits were defined as the highest stress level, for which run-out was achieved. Cyclic stress–strain data were recorded throughout each test. In order to assess the damage development in the composite, stiffness degradation, changes in hysteresis energy density, as well as strain accumulation with fatigue cycles were examined. To determine retained strength and stiffness, specimens that achieved run-out were subjected to tensile test to failure at the temperature of the fatigue test. Fatigue tests were performed at 1200 and 1330 C, in laboratory air and in steam environments. 2.3. Characterization Fracture surfaces of failed specimens were examined using SEM (Model 360FE, Leica) as well as optical microscopy. The SEM specimens were gold coated. 3. Results and discussion 3.1. Monotonic tension Tensile stress–strain behavior at 23, 1200 and 1330 C is shown in Fig. 2. The stress–strain curves obtained at 23 and 1200 C are nearly linear to failure. Such linear behavior indicates that there is little additional matrix cracking Fig. 1. As-received material: (a) overview, optical microscope and (b) porous nature of the matrix is evident (SEM). 0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 0 50 100 150 200 250 Strain (%) Stress (MPa) 23°C 1330°C 1200°C Fig. 2. Tensile stress–strain curves for NextelTM720/alumina ceramic composite at 23, 1200 and 1330 C. M.B. Ruggles-Wrenn et al. / Composites: Part A 37 (2006) 2029–2040 2031
M.B. Ruggles-Wrenn et aL. Composites: Part A 37(2006)2029-2040 and that fiber-matrix debonding is insignificant. Material exhibits typical fiber-dominated composite behavior. Fiber 口1200°c,Ar■1330°c,Air fracture appears to be the dominant damage mode. At △1200c,Seam▲1330c, Steam 23C, the ultimate tensile strength (UTS) was 169 MPa UTS at 1200C lastic modulus, 60 GPa, and failure strain, 0.35%. At 2 150 1200C, the UTS, elastic modulus and failure strain were 5 192 MPa, 75 GPa and 0.38%, respectively. These results UTS at1330°c agree well with those reported by COI Ceramics [50] The stress-strain behavior changes dramatically at 1330C. The stress-strain curve at 1330C is linear up to the proportional limit( 74 MPa), where non-linear behav- r sets in. At 1330C, the UTS, elastic modulus and fail- ure strain were 120 MPa, 42 GPa and 1.7/, respectively 1.E+001.E4011.E+021.E+081.E+041.E+05 While the UTS and the elastic modulus are significantly lower than those at 1200C, failure strain increases almost Fig 3. Fatigue S-N curves for NextelTM720/alumina ceramic composite tenfold at 1200 and 1330C, in laboratory air and in steam environment It is important to note that in all tension tests, as well as in all other tests reported herein, the failure occurred within the gage section of the extensometer. run-out condition of 10 cycles, approximate number o loading cycles expected in aerospace applications 3.2. Tension-tension fatigue 1200oC. It is believed that a more rigorous run-out condi tion would have resulted in a lower fatigue limit. Presence Degradation of fatigue performance in high-tempera of steam(a highly oxidizing environment) causes noticeable ture oxidizing environments remains among the key con- degradation in fatigue performance. At 1200C, the in- erns that must ddressed before using CMCs in steam fatigue limit is only 125 MPa (65% UTS 1200C). As seen in Fig. 3, increase in temperature from advanced aerospace applications. Therefore high-tempera- 1200 to 1330C results in significant degradation of the ture fatigue tests, especially when conducted in steam envi- in-air fatigue performance. Even at the low fatigue stress ronment are critical to assessing the durability of a given level of 50 MPa(42% UTS at 1330C) the run-out was CMC Tension-tension fatigue tests with a ratio, R of 0.05 not achieved As expected, steam environment even further vere performed at 1200 and 1330%C in air and in steam degraded an already poor fatigue resistance nvironments. Results are summarized in Table 1. where Of importance in cyclic fatigue is the reduction in stiff the maximum stress level and number of cycles to failure. and minima.sIs modulus determined from the maximum um stress-strain data points during a load Results are also presented in Fig. 3 as stress vs cycles to cycle), reflecting the damage development during cycling failure(S-N) curves for both temperatures and environ- normalized modulus(i. e. modulus normalized by the mod- ments.At 1200C the in-air fatigue limit was 170 MPa ulus obtained in the first cycle) is plotted vs fatigue cycles (88% UTS at 1200C). This fatigue limit is based on a It is noteworthy that although all in-air tests achieved Table 1 gue results for the N720/A composite at 1200 and °C.in ory air and steam environments Cycles to failure - A125 MPa, Air +125 MPa, Steam1-1H2oc Eh-100MPa, Air -100 MPa, Steam T= 120 Max stress(MPa) -e-150 MPa, Air --150 MPa, Steam Laboratory air 120,199 Laboratory air - 170 MPa, Air -+170 MPa, Stea Laboratory air 67,4732 Laboratory air l09.436 100.7804 Steam 1663262 Steam at1330°C 1E001E011E21E:01E:041E:051E:06 ry 25,852 Fig 4. Normalized modulus vs fatigue cycles at 1200C in laboratory
and that fiber-matrix debonding is insignificant. Material exhibits typical fiber-dominated composite behavior. Fiber fracture appears to be the dominant damage mode. At 23 C, the ultimate tensile strength (UTS) was 169 MPa, elastic modulus, 60 GPa, and failure strain, 0.35%. At 1200 C, the UTS, elastic modulus and failure strain were 192 MPa, 75 GPa and 0.38%, respectively. These results agree well with those reported by COI Ceramics [50]. The stress–strain behavior changes dramatically at 1330 C. The stress–strain curve at 1330 C is linear up to the proportional limit (74 MPa), where non-linear behavior sets in. At 1330 C, the UTS, elastic modulus and failure strain were 120 MPa, 42 GPa and 1.7%, respectively. While the UTS and the elastic modulus are significantly lower than those at 1200 C, failure strain increases almost tenfold. It is important to note that in all tension tests, as well as in all other tests reported herein, the failure occurred within the gage section of the extensometer. 3.2. Tension–tension fatigue Degradation of fatigue performance in high-temperature oxidizing environments remains among the key concerns that must be addressed before using CMCs in advanced aerospace applications. Therefore high-temperature fatigue tests, especially when conducted in steam environment are critical to assessing the durability of a given CMC. Tension–tension fatigue tests with a ratio, R of 0.05, were performed at 1200 and 1330 C in air and in steam environments. Results are summarized in Table 1, where test temperature and environment are shown together with the maximum stress level and number of cycles to failure. Results are also presented in Fig. 3 as stress vs cycles to failure (S–N) curves for both temperatures and environments. At 1200 C the in-air fatigue limit was 170 MPa (88% UTS at 1200 C). This fatigue limit is based on a run-out condition of 105 cycles, approximate number of loading cycles expected in aerospace applications at 1200 C. It is believed that a more rigorous run-out condition would have resulted in a lower fatigue limit. Presence of steam (a highly oxidizing environment) causes noticeable degradation in fatigue performance. At 1200 C, the insteam fatigue limit is only 125 MPa (65% UTS at 1200 C). As seen in Fig. 3, increase in temperature from 1200 to 1330 C results in significant degradation of the in-air fatigue performance. Even at the low fatigue stress level of 50 MPa (42% UTS at 1330 C) the run-out was not achieved. As expected, steam environment even further degraded an already poor fatigue resistance. Of importance in cyclic fatigue is the reduction in stiff- ness (hysteresis modulus determined from the maximum and minimum stress–strain data points during a load cycle), reflecting the damage development during cycling. Change in modulus at 1200 C is shown in Fig. 4, where normalized modulus (i.e. modulus normalized by the modulus obtained in the first cycle) is plotted vs fatigue cycles. It is noteworthy that although all in-air tests achieved Table 1 Summary of fatigue results for the N720/A composite at 1200 and 1330 C, in laboratory air and steam environments Test environment Max stress (MPa) Cycles to failure Fatigue at 1200 C Laboratory air 100 120,199a Laboratory air 125 146,392a Laboratory air 150 167,473a Laboratory air 170 109,436a Steam 100 100,780a Steam 125 166,326a Steam 150 11,782 Steam 170 202 Fatigue at 1330 C Laboratory air 50 97,282 Laboratory air 100 1,519 Steam 50 25,852 Steam 100 347 a Run-out. 0 50 100 150 200 250 1.E+00 1.E+01 1.E+02 1.E+03 1.E+04 1.E+05 1.E+06 Cycles (N) Stress (MPa) 1200°C, Air 1330°C, Air 1200°C, Steam 1330°C, Steam UTS at 1200°C UTS at 1330°C Fig. 3. Fatigue S–N curves for NextelTM720/alumina ceramic composite at 1200 and 1330 C, in laboratory air and in steam environment. 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 1.E+00 1.E+01 1.E+02 1.E+03 1.E+04 1.E+05 1.E+06 Cycles (N) Normalized Modulus (E/Eo) 100 MPa, Air 100 MPa, Steam 125 MPa, Air 125 MPa, Steam 150 MPa, Air 150 MPa, Steam 170 MPa, Air 170 MPa, Steam T = 1200°C f = 1 Hz R = 0.05 Fig. 4. Normalized modulus vs fatigue cycles at 1200 C in laboratory air and in steam environment. 2032 M.B. Ruggles-Wrenn et al. / Composites: Part A 37 (2006) 2029–2040
M B. Ruggles.Wrenn et al Composites: Part A 37(2006)2029-2040 run-out, a decrease in normalized modulus with cycling was still observed. Modulus loss increased with increasing Cycle 125 T=1200°c the 150 MPa test, and 17% in the 170 MPa test. Decrease in F Cycle 1 fatigue stress level In air, normalized modulus was reduced by 5%o in the 100 MPa test, 7% in the 125 MPa test, 8% in Cycle 103225 normalized modulus becomes more pronounced in steam 3 Cycle 2 environment. while modulus was limited to 17%, normalized modulus loss reached 30% in steam. In steam environment. normalized modulus loss in run-out fatigue tests was 10% in the 100 MPa test and 16% in the 125 MPa test. normalized frequency 1 Hz modulus in the 150 MPa test dropped by 17%, and in the 170 MPa test, by a significant 30% prior to failure at 202 00.1020.3040.50.60.70.80.9 cycles. Continuous decrease in modulus observed both in Strain(%) air and steam environments suggests progressive damage Fig. 6. Typical evolution of a stress-strain hysteresis loop with fatigue with continued cycling Because the fatigue damage is still cycles. the criteria of a true endurance fatigue limit proposed by Sorensen et al. [51] and may not be a true endurance presented in Fig. 7(a) and(b)for laboratory air and steam fatigue limit environments, respectively. It is seen that ratcheting takes Normalized modulus evolution with cycling at 1330.C is place in all fatigue tests conducted in air at 1200C. Onset shown in Fig. 5. Modulus loss in the 100 MPa in-air test was of ratchets g depends on the ma aximum tigue stress. Ear- % noticeably greater than the 5% modulus loss in the lier onset of ratcheting is observed in tests with higher max corresponding 1200.C test. This suggests accelerated dam- imum stress levels. In the 100 MPa test, there is little age growth, but may also be indicative of accelerated fiber change in accumulated strain for up to 10,000 cycles, only degradation at hig gher temperature. Contrary to the expec. tations, presence of steam caused little additional modulus degradation Modulus loss in steam was limited to 17% T=1200°c Hysteresis loops for a 100 MPa test conducted in air at Fatigue in Air 1200C are presented in Fig. 6. Results in Fig. 6 are repre- sentative of the hysteresis loop evolution with cycling observed in all fatigue tests reported herein. It is seen that most extensive damage occurs on the first cycle. Afterwards hysteresis loops stabilize quickly. Results in Fig. 6 reveal that ratcheting, defined as progressive increase in accumu lated strain with increasing number of cycles, continues throughout the test. Maximum and minimum cyclic strains as functions of cycle number for fatigue tests conducted at 1200C are E+001.E+011.E+021.E+031.E+041.E+051E+06 (a) Cycles(N T=1330°c 非翻 T=1200°c Fatigue in Steam R=0.05 1.0 1.E+021.E+0 E+001E+011E+021.E+031.E+041.E+051E+06 Cycles(N) Fig. 5. Normalized modulus vs fatigue cycles at 1330C in laboratory Fig. 7. Maximum and minimum strains as functions of cycle number at and in steam environment 1200C:(a)in laboratory air and(b)in steam environment
run-out, a decrease in normalized modulus with cycling was still observed. Modulus loss increased with increasing fatigue stress level. In air, normalized modulus was reduced by 5% in the 100 MPa test, 7% in the 125 MPa test, 8% in the 150 MPa test, and 17% in the 170 MPa test. Decrease in normalized modulus becomes more pronounced in steam environment. While in air the reduction in normalized modulus was limited to 17%, normalized modulus loss reached 30% in steam. In steam environment, normalized modulus loss in run-out fatigue tests was 10% in the 100 MPa test and 16% in the 125 MPa test. Normalized modulus in the 150 MPa test dropped by 17%, and in the 170 MPa test, by a significant 30% prior to failure at 202 cycles. Continuous decrease in modulus observed both in air and steam environments suggests progressive damage with continued cycling. Because the fatigue damage is still evolving at 105 cycles, the 105 fatigue limit does not meet the criteria of a true endurance fatigue limit proposed by Sorensen et al. [51] and may not be a true endurance fatigue limit. Normalized modulus evolution with cycling at 1330 C is shown in Fig. 5. Modulus loss in the 100 MPa in-air test was 15%, noticeably greater than the 5% modulus loss in the corresponding 1200 C test. This suggests accelerated damage growth, but may also be indicative of accelerated fiber degradation at higher temperature. Contrary to the expectations, presence of steam caused little additional modulus degradation. Modulus loss in steam was limited to 17%. Hysteresis loops for a 100 MPa test conducted in air at 1200 C are presented in Fig. 6. Results in Fig. 6 are representative of the hysteresis loop evolution with cycling observed in all fatigue tests reported herein. It is seen that most extensive damage occurs on the first cycle. Afterwards hysteresis loops stabilize quickly. Results in Fig. 6 reveal that ratcheting, defined as progressive increase in accumulated strain with increasing number of cycles, continues throughout the test. Maximum and minimum cyclic strains as functions of cycle number for fatigue tests conducted at 1200 C are presented in Fig. 7(a) and (b) for laboratory air and steam environments, respectively. It is seen that ratcheting takes place in all fatigue tests conducted in air at 1200 C. Onset of ratcheting depends on the maximum fatigue stress. Earlier onset of ratcheting is observed in tests with higher maximum stress levels. In the 100 MPa test, there is little change in accumulated strain for up to 10,000 cycles, only 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 1.E+00 1.E+01 1.E+02 1.E+03 1. E+04 1.E+05 1.E+06 Cycles (N) Normalized Modulus (E/Eo) 100 MPa, Air 100 MPa, Steam 50 MPa, Steam T = 1330°C f = 1 Hz R = 0.05 Fig. 5. Normalized modulus vs fatigue cycles at 1330 C in laboratory air and in steam environment. 0 20 40 60 80 100 120 140 160 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 Strain (%) Stress (MPa) T = 1200°C Fatigue in Air Max Stress = 100 Mpa frequency = 1 Hz R = 0.05 Cycle 10000 Cycle 103225 Cycle 1025 Cycle 125 Cycle 1 Cycle 2 Fig. 6. Typical evolution of a stress–strain hysteresis loop with fatigue cycles. 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 1.E+00 1.E+01 1.E+02 1.E+03 1.E+04 1.E+05 1.E+06 Cycles (N) Strain (%) T = 1200°C Fatigue in Steam T = 1200°C Fatigue in Air 100 Mpa 125 MPa 150 MPa 170 MPa Strain (%) Cycles (N) 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 1.E+00 1.E+01 1.E+02 1.E+03 1.E+04 1.E+05 1.E+06 100 MPa 150 MPa 170 MPa (a) (b) Fig. 7. Maximum and minimum strains as functions of cycle number at 1200 C: (a) in laboratory air and (b) in steam environment. M.B. Ruggles-Wrenn et al. / Composites: Part A 37 (2006) 2029–2040 2033
B. Ruggles-Wrenn et al. Composites: Part A 37(2006)2029-2040 then does ratcheting begin. Conversely, in the 150 MPa test itcheting begins after 1000 cycles, and in the 170 MP -e-100 MPa Air T=1200°c test, after only 250 cycles. Earlier ratcheting is accompa- 40 150 MPa. Air f=1H 2-170 MPa. Air R=0.05 nied with higher strain accumulation. Maximum strains ≥ accumulated in the 100. 150 and 170 MPa tests were +100 MPa. Steam t-125 MPa Steam 0.6%, 1.7%/, and 2.4%, respectively -150 MPa. steam As seen in Fig. 7(b), presence of steam accelerates rat cheting at 1200oC. For a given maximum stress, specimens 2o tested in steam exhibited a much earlier onset of ratcheting than those tested in air. In the 100 MPa test conducted in steam, ratcheting begins after 100 cycles, in the 150 MPa test, after 50 cycles, and in the 170 MPa test, after mere 10 cycles. Strain accumulations in the 100, 125, 150 and 1.E+001.E+011.E+021.E+031.E+041.E+051.E+06 170 MPa tests conducted in steam were 0.7%. 1.0% 0.7%, and 0.8%, respectively. Strains accumulated in the Fig 9. H 60 and 170 MPa tests in steam are considerably lower laboratory air and sis energy density(HED)vS fatigue cycles at 1200.Cin than those produced for the same fatigue stress levels in air. Generally, lower strain accumulation with cycling indi- cates that less damage has occurred, and that it is mostly limited to some additional matrix cracking. However, in 母100MPa,Air T the case of 150 and 170 MPa tests conducted in steam ◆100MPa. Stean low accumulated strains are more likely due to early bundle 70 ilures leading to specimen failure in accelerated ratcheting as well as in larger strain accumu- 9 lations with cycling(see Fig. 8). In the 100 MPa test at 1330C, ratcheting begins immediately, strain is accumu- lated rapidly, reaching a significant 3.9% at failure. At 1330C, presence of steam results in earlier failure. Fur- 10 m thermore, in steam environment, shorter cyclic lives are 2+001.E+011.E+021.E+031.E+041.E+051E+06 accompanied with lower strain accumulations Cycles(N) The hysteresis energy density(HED) behavior is shown in Figs. 9 and 10 for 1200 and 1330oC, respectively. The Fig 10. Hysteresis energy density (HED) vs fatigue cycles at 1330"C in HED values at 1200C are fairly small, with the average of xl0 kJ/m. Most traditional composites with interfaces and classical fiber debonding typically produce HED val- a slight decrease in HED with continued cycling. However, ues>80kJ/m'when fatigued above the proportional limit. upon closer examination the 1200C data reveals that at It is seen that for each stress level tested at 1200C, the 10 cycles the hed becomes stable for all tests except HED exhibits a significant decrease within the first 10 the 150 MPa test conducted in steam. Among tests repre- cycles. From this cycle number on there appears to be only sented in Fig 9, only the 150 MPa test in steam did not nieve run-out. In conventional composites, a decrease in HED with fatigue cycling is generally attributed to deg- T=1330°c radation of interfacial shear resistance at the fiber matrix 35·100MPa, Steam interface. For brittle matrix composites, it was also -A50 MPa. Steam observed [32] that continuous damage development, such as matrix cracking and fiber/matrix debonding, in a cycli cally loaded specimen may have a significant effect on the HED behavior. The hed behavior at 1330C is qualita- tively similar to that observed at 1200C. However,aver age HED values obtained in 100 MPa tests were higher at 1330C than at 1200C. The presence of steam appears to have little effect on the hed behavior at both tempe 李导是 tures investigated E+001.E+011E+021.E+031.E+0 E+051.E+06 Retained strength and stiffness of the fatigue specimens, Cycles(N) which achieved fatigue run-out, are summarized in Table 2. Fig 8. Maximum and minimum strains as functions of cycle number at Evaluation of retained properties is useful in assessing the 330C in laboratory air and in steam environment damage state of the composite subjected to prior loading
then does ratcheting begin. Conversely, in the 150 MPa test ratcheting begins after 1000 cycles, and in the 170 MPa test, after only 250 cycles. Earlier ratcheting is accompanied with higher strain accumulation. Maximum strains accumulated in the 100, 150 and 170 MPa tests were 0.6%, 1.7%, and 2.4%, respectively. As seen in Fig. 7(b), presence of steam accelerates ratcheting at 1200 C. For a given maximum stress, specimens tested in steam exhibited a much earlier onset of ratcheting than those tested in air. In the 100 MPa test conducted in steam, ratcheting begins after 100 cycles, in the 150 MPa test, after 50 cycles, and in the 170 MPa test, after mere 10 cycles. Strain accumulations in the 100, 125, 150 and 170 MPa tests conducted in steam were 0.7%, 1.0%, 0.7%, and 0.8%, respectively. Strains accumulated in the 150 and 170 MPa tests in steam are considerably lower than those produced for the same fatigue stress levels in air. Generally, lower strain accumulation with cycling indicates that less damage has occurred, and that it is mostly limited to some additional matrix cracking. However, in the case of 150 and 170 MPa tests conducted in steam, low accumulated strains are more likely due to early bundle failures leading to specimen failure. In air environment, increase in test temperature results in accelerated ratcheting as well as in larger strain accumulations with cycling (see Fig. 8). In the 100 MPa test at 1330 C, ratcheting begins immediately, strain is accumulated rapidly, reaching a significant 3.9% at failure. At 1330 C, presence of steam results in earlier failure. Furthermore, in steam environment, shorter cyclic lives are accompanied with lower strain accumulations. The hysteresis energy density (HED) behavior is shown in Figs. 9 and 10 for 1200 and 1330 C, respectively. The HED values at 1200 C are fairly small, with the average of 10 kJ/m3 . Most traditional composites with interfaces and classical fiber debonding typically produce HED values P80 kJ/m3 when fatigued above the proportional limit. It is seen that for each stress level tested at 1200 C, the HED exhibits a significant decrease within the first 10 cycles. From this cycle number on there appears to be only a slight decrease in HED with continued cycling. However, upon closer examination the 1200 C data reveals that at 104 cycles the HED becomes stable for all tests except the 150 MPa test conducted in steam. Among tests represented in Fig. 9, only the 150 MPa test in steam did not achieve run-out. In conventional composites, a decrease in HED with fatigue cycling is generally attributed to degradation of interfacial shear resistance at the fiber matrix interface. For brittle matrix composites, it was also observed [32] that continuous damage development, such as matrix cracking and fiber/matrix debonding, in a cyclically loaded specimen may have a significant effect on the HED behavior. The HED behavior at 1330 C is qualitatively similar to that observed at 1200 C. However, average HED values obtained in 100 MPa tests were higher at 1330 C than at 1200 C. The presence of steam appears to have little effect on the HED behavior at both temperatures investigated. Retained strength and stiffness of the fatigue specimens, which achieved fatigue run-out, are summarized in Table 2. Evaluation of retained properties is useful in assessing the damage state of the composite subjected to prior loading. 100 MPa, Air T = 1330°C 100 MPa, Steam 50 MPa, Steam 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 1.E+00 1.E+01 1.E+02 1.E+03 1.E+04 1.E+05 1.E+06 Cycles (N) Strain (%) Fig. 8. Maximum and minimum strains as functions of cycle number at 1330 C in laboratory air and in steam environment. 100 MPa, Air 150 MPa, Air 170 MPa, Air 100 MPa, Steam 125 MPa, Steam 150 MPa, Steam T = 1200°C f = 1 Hz R = 0.05 0 5 10 15 20 25 30 35 40 45 50 1.E+00 1.E+01 1.E+02 1.E+03 1.E+04 1. E+05 1.E+06 Cycles (N) Hysteretic Energy Density (kJ/m3) Fig. 9. Hysteresis energy density (HED) vs fatigue cycles at 1200 C in laboratory air and in steam environment. T = 1330°C f = 1 Hz R = 0.05 0 10 20 30 40 50 60 70 80 90 100 1.E+00 1. E+01 1.E+02 1. E+03 1.E+04 1. E+05 1.E+06 Cycles (N) Hysteretic Energy Density (kJ/m3) 100 MPa, Air 50 MPa, Steam 100 MPa, Steam Fig. 10. Hysteresis energy density (HED) vs fatigue cycles at 1330 C in laboratory air and in steam environment. 2034 M.B. Ruggles-Wrenn et al. / Composites: Part A 37 (2006) 2029–2040
M B. Ruggles-Wrenn et al Composites: Part A 37(2006)2029-2040 35 Table 2 etained properties of the N720/A specimens and in steam envi nt at I200° Fatigue stress(MPa) Retained strength(MPa ion (%) Retained modulus (GPa) Modulus retention (% Strain at failure (% Prior fatigue in laboratory air 000 0.44 0.44 43.4 0.53 40.7 Prior fatigue in steam environment 98 6 432730 0.40 125 It is seen that specimens tested in air exhibited no loss of N720/A CMC(the same material as used in this study) tensile strength, irrespective of the fatigue stress level. Campbell et al. [56] exposed N720 /a to a water-vapor However, considerable stifness loss (28-33%)was environment for 1000 h at 1200C. Strength loss of observed. Stifness degradation increases with increasing 15% was observed after exposure prior fatigue stress level. Full retention of tensile strength In the present study, a high fatigue limit(88% UTS)and suggests that no fatigue damage occurred to the fibers. 100% strength retention are observed for specimens tested The reduction in stiffness is most likely due to additional in air. Presence of steam noticeably degrades fatigue perfor- matrix cracking. Conversely, prior fatigue in steam caused mance of the material. In steam environment, fatigue limit reduction of both strength and stiffness. Strength loss in is significantly lower(65% UTS)and strength retention is steam was limited to 12% and stifness loss, to 20%. In this limited to 90%. It is noteworthy that strength losses similar case, the loss of strength may be associated with the envi- to those observed by Campbell et al. [56] after 1000 h of no- ronmental degradation of the fibers, while both fiber degra- load exposure are seen after only 28 h(10 cycles at a fre- dation and progressive matrix cracking may account for quency of 1 Hz) of fatigue cycling in steam at 1200C. The the loss of stifness. The discrepancy between the retained strength loss is strongly influenced by the loading condi modulus of a run-out specimen and the decrease in hyster- tions. In a given high-temperature environment, strength esis modulus observed during fatigue testing most likely degradation increases with increasing fatigue load. stems from different methods used to determine the retained and hysteresis moduli Results in Table 2 demon- 3.3. Creep rupture strate that fatigue in air did not cause reduction in strength However, prior fatigue in steam environment resulted in Results of the creep-rupture tests are summarized noticeable strength loss, which cannot be neglected. Fiber Table 3, where test temperature and environment are degradation represents a possible source of composite deg- shown together with the creep stress level and time to rup- radation. Nextel720 fibers consist of alumina grains with ture. It is noteworthy that all specimens failed within the an approximate diameter of 0. 1 um distributed among lar- extensometer gage section ger(0.5 um)mullite grains, consisting of many smaller sub- Creep curves obtained at 1200 and 1330C are pre- grains [52]. Reported observations of the response of the sented in Figs. ll and 12, respectively. Time scale in Figs fibers to thermal exposure are somewhat conflicting Deleg- ll and 12 is reduced in order to clearly show creep curves lise et al. [52] observed significant degradation only above 1400C for 5 h exposure times, while Milz et al. [53] Table 3 reported severe degradation after 2 h at 1300C. Petry nary of creep-rupture results for the N720/A composite at 1200 and and Mah [54] report a small reduction in strength after C, in laboratory air and in steam environments 2 h at 1100C. The causes of degradation are not well Environment Creep stress(MPa) Time to rupture (s) understood; surface grooving, structural coarsening [54 and local impurity enrichment have been suggested [53]. air Creep at I200°c 917,573 Evidence in literature also suggests that, under specific con-Air 147. ditions. reactions between the sio in the fiber and mois-Air ture may occur. Wannaparhun et al. [55] showed that, at Air 1100C in water-vapor environment, Sio, could be lea- Steam 05480 165,77 ched out of the NextelM720 fiber. Exposure to water st vapor resulted in the formation of volatile Si(oH)4 and Steam 154 was responsible for the loss of the mullite phase in the fiber. Creep at 1330%C Furthermore, formation of volatile Si(oH)4 also resulted in Air 313,198 surface recondensation of these silicon species with theAir Al,O3 matrix at the specimen surface, in turn causing an Steam 0000 l1,088 increase in aluminosilicate content at the surface of the Steam
It is seen that specimens tested in air exhibited no loss of tensile strength, irrespective of the fatigue stress level. However, considerable stiffness loss (28–33%) was observed. Stiffness degradation increases with increasing prior fatigue stress level. Full retention of tensile strength suggests that no fatigue damage occurred to the fibers. The reduction in stiffness is most likely due to additional matrix cracking. Conversely, prior fatigue in steam caused reduction of both strength and stiffness. Strength loss in steam was limited to 12% and stiffness loss, to 20%. In this case, the loss of strength may be associated with the environmental degradation of the fibers, while both fiber degradation and progressive matrix cracking may account for the loss of stiffness. The discrepancy between the retained modulus of a run-out specimen and the decrease in hysteresis modulus observed during fatigue testing most likely stems from different methods used to determine the retained and hysteresis moduli. Results in Table 2 demonstrate that fatigue in air did not cause reduction in strength. However, prior fatigue in steam environment resulted in noticeable strength loss, which cannot be neglected. Fiber degradation represents a possible source of composite degradation. NextelTM720 fibers consist of alumina grains with an approximate diameter of 0.1 lm distributed among larger (0.5 lm) mullite grains, consisting of many smaller subgrains [52]. Reported observations of the response of the fibers to thermal exposure are somewhat conflicting. Deleglise et al. [52] observed significant degradation only above 1400 C for 5 h exposure times, while Milz et al. [53] reported severe degradation after 2 h at 1300 C. Petry and Mah [54] report a small reduction in strength after 2 h at 1100 C. The causes of degradation are not well understood; surface grooving, structural coarsening [54] and local impurity enrichment have been suggested [53]. Evidence in literature also suggests that, under specific conditions, reactions between the SiO2 in the fiber and moisture may occur. Wannaparhun et al. [55] showed that, at 1100 C in water–vapor environment, SiO2 could be leached out of the NextelTM720 fiber. Exposure to water vapor resulted in the formation of volatile Si(OH)4 and was responsible for the loss of the mullite phase in the fiber. Furthermore, formation of volatile Si(OH)4 also resulted in surface recondensation of these silicon species with the Al2O3 matrix at the specimen surface, in turn causing an increase in aluminosilicate content at the surface of the N720/A CMC (the same material as used in this study). Campbell et al. [56] exposed N720/A to a water–vapor environment for 1000 h at 1200 C. Strength loss of 15% was observed after exposure. In the present study, a high fatigue limit (88% UTS) and 100% strength retention are observed for specimens tested in air. Presence of steam noticeably degrades fatigue performance of the material. In steam environment, fatigue limit is significantly lower (65% UTS) and strength retention is limited to 90%. It is noteworthy that strength losses similar to those observed by Campbell et al. [56] after 1000 h of noload exposure are seen after only 28 h (105 cycles at a frequency of 1 Hz) of fatigue cycling in steam at 1200 C. The strength loss is strongly influenced by the loading conditions. In a given high-temperature environment, strength degradation increases with increasing fatigue load. 3.3. Creep rupture Results of the creep-rupture tests are summarized in Table 3, where test temperature and environment are shown together with the creep stress level and time to rupture. It is noteworthy that all specimens failed within the extensometer gage section. Creep curves obtained at 1200 and 1330 C are presented in Figs. 11 and 12, respectively. Time scale in Figs. 11 and 12 is reduced in order to clearly show creep curves Table 2 Retained properties of the N720/A specimens subjected to prior fatigue in laboratory air and in steam environment at 1200 C Fatigue stress (MPa) Retained strength (MPa) Strength retention (%) Retained modulus (GPa) Modulus retention (%) Strain at failure (%) Prior fatigue in laboratory air 100 194 P100 56.6 74 0.44 125 199 P100 54.9 73 0.44 150 199 P100 43.4 72 0.53 170 192 P100 40.7 67 0.51 Prior fatigue in steam environment 100 174 90 47.6 84 0.40 125 168 88 52.0 80 0.43 Table 3 Summary of creep-rupture results for the N720/A composite at 1200 and 1330 C, in laboratory air and in steam environments Environment Creep stress (MPa) Time to rupture (s) Creep at 1200 C Air 80 917,573 Air 100 147,597 Air 125 15,295 Air 154 968 Steam 80 165,777 Steam 100 8,966 Steam 125 869 Steam 154 98 Creep at 1330 C Air 50 313,198 Air 100 4,244 Steam 50 11,088 Steam 100 40 M.B. Ruggles-Wrenn et al. / Composites: Part A 37 (2006) 2029–2040 2035
M B. Ruggles-Wrenn et al. Composites: Part A 37(2006)2029-2040 125 MPa, creep strain accumulation increases from 1%to T=1200°c 3.4%. However, for the creep stress of 154 MPa, creep strain accumulation reaches only about 0.6%/. It is note- worthy that in all creep tests, the accumulated creep strain significantly exceeded failure strain obtained in tension test It is important to recognize that the total strain incurred in 100 MPa. Steam a creep-rupture test represents a sum of two contributions: (1)that due to the initial loading up to the specific creep stress level, and (2) that accumulated during the actual 80 MPa, A creep od. for to 90% of total strai during the creep period. However, for specimen tested at 50000 00000 150000 200000 creep stress of 154 MPa, creep strain accounted only for Time(s) 33% of the total strain The creep curves produced in air environment at T=1200°c 1330C are qualitatively similar to those obtained at 1200C. Creep strain accumulation decreases from 5%to approximately 4% as the creep stress increases from 50 to 100 MPa. In both creep tests, close to 100% of the total 125 MPa, Steam strain was incurred during the creep period. Creep strain accounts for 98% of the total strain in the 50 MPa test 154 MPa, Steam and for 94% of the total strain in the 100 MPa test Results in Figs. Il and 12 demonstrate that specimens tested in steam environment produced creep curves that are qualitatively similar to those produced in air. As 154 MPa. Air expected, strains incurred during the initial loading to a given creep stress were not affected by the steam. Conversely, creep strains accumulated in steam Time(s) environment were significantly different from those accu Fig.11.Creep strain vs time for Nextel TM720 /alumina ceramic composite mulated in air. For both test temperatures and creep stress at 1200C in laboratory air and in steam environment: (a)at 80 and levels 100 MPa, presence of steam resulted in lower creep strains and much lower creep lifetimes at both test ter atures. At 1200C, creep strains produced in steam ronment for creep stress levels of 100, 125 and 154 100 MPa. Steam 100 MPa. Air were, respectively, 53%, 73% and 92% lower than those produced at the same stress levels in air. At 1330C and creep stress of 100 MPa, creep strain accumulated in steam 50 MPa,Air was 60% lower than that accumulated in air Minimum creep rate was reached in all tests. Creep rate 1000 2000 Time(s) as a function of applied stress is presented in Fig. 13, where results of the present investigation are plotted together with Fig. 12. Creep strain vs time for NextelM720/alumina ceramic composite the data from Wilson and Visser [57] for Nextel 720 fibers at 1330C in laboratory air and in steam environment To further facilitate comparison between the creep proper ties of the fibers and the composite, the Nextel 720 fiber produced at higher stress levels. It is seen that all creep data adjusted for Vr=0. 22(volume fraction of the on-axis curves generated at 1200C in air environment exhibit pri- fibers in the N720/A composite)is also shown For both mary and secondary creep regimes. Transition from pri- temperatures, the minimum creep rates increase with ary to secondary creep occurs during the first 10% of increasing applied stress. At 1200C, as creep stress creep life. Secondary creep appears to be nearly linear to increases from 80 to 154 MPa creep rate increases by two failure. As the creep stress level increases from 80 MPa to orders of magnitude. Creep rates of the composite obtained
produced at higher stress levels. It is seen that all creep curves generated at 1200 C in air environment exhibit primary and secondary creep regimes. Transition from primary to secondary creep occurs during the first 10% of creep life. Secondary creep appears to be nearly linear to failure. As the creep stress level increases from 80 MPa to 125 MPa, creep strain accumulation increases from 1% to 3.4%. However, for the creep stress of 154 MPa, creep strain accumulation reaches only about 0.6%. It is noteworthy that in all creep tests, the accumulated creep strain significantly exceeded failure strain obtained in tension test. It is important to recognize that the total strain incurred in a creep-rupture test represents a sum of two contributions: (1) that due to the initial loading up to the specific creep stress level, and (2) that accumulated during the actual creep period. For specimens tested at creep stresses 6125 MPa, close to 90% of total strain was accumulated during the creep period. However, for specimen tested at creep stress of 154 MPa, creep strain accounted only for 33% of the total strain. The creep curves produced in air environment at 1330 C are qualitatively similar to those obtained at 1200 C. Creep strain accumulation decreases from 5% to approximately 4% as the creep stress increases from 50 to 100 MPa. In both creep tests, close to 100% of the total strain was incurred during the creep period. Creep strain accounts for 98% of the total strain in the 50 MPa test, and for 94% of the total strain in the 100 MPa test. Results in Figs. 11 and 12 demonstrate that specimens tested in steam environment produced creep curves that are qualitatively similar to those produced in air. As expected, strains incurred during the initial loading to a given creep stress were not affected by the presence of steam. Conversely, creep strains accumulated in steam environment were significantly different from those accumulated in air. For both test temperatures and creep stress levels <100 MPa, specimens tested in steam accumulated more creep strain than specimens tested in air. At 1200 C, creep strain produced in the 80 MPa test conducted in steam was 67% higher than that produced at the same creep stress in air. At 1330 C and creep stress of 50 MPa, creep strain accumulated in steam was 20% higher than that accumulated in air. For creep stress levels P100 MPa, presence of steam resulted in lower creep strains and much lower creep lifetimes at both test temperatures. At 1200 C, creep strains produced in steam environment for creep stress levels of 100, 125 and 154 MPa were, respectively, 53%, 73% and 92% lower than those produced at the same stress levels in air. At 1330 C and creep stress of 100 MPa, creep strain accumulated in steam was 60% lower than that accumulated in air. Minimum creep rate was reached in all tests. Creep rate as a function of applied stress is presented in Fig. 13, where results of the present investigation are plotted together with the data from Wilson and Visser [57] for Nextel 720 fibers. To further facilitate comparison between the creep properties of the fibers and the composite, the Nextel 720 fiber data adjusted for Vf = 0.22 (volume fraction of the on-axis fibers in the N720/A composite) is also shown. For both temperatures, the minimum creep rates increase with increasing applied stress. At 1200 C, as creep stress increases from 80 to 154 MPa creep rate increases by two orders of magnitude. Creep rates of the composite obtained Fig. 11. Creep strain vs time for NextelTM720/alumina ceramic composite at 1200 C in laboratory air and in steam environment: (a) at 80 and 100 MPa, (b) at 125 and 154 MPa. 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 0 500 1000 1500 2000 Time (s) Strain (%) T = 1330°C 100 MPa, Air 50 MPa, Air 100 MPa, Steam 50 MPa, Steam Fig. 12. Creep strain vs time for NextelTM720/alumina ceramic composite at 1330 C in laboratory air and in steam environment. 2036 M.B. Ruggles-Wrenn et al. / Composites: Part A 37 (2006) 2029–2040
M B. Ruggles.Wrenn et al Composites: Part 4 37(2006)2029-2040 1.E02 given creep stress, creep-rupture life at 1330C is consider ■1200°c,Ar ably reduced compared to that at 1200C. With the creep 1E03 1330°c,Ar run-out condition defined as 100 h. 80 MPa was the run .E o a c steam N720 Fiber at v, =0. 2 out stress at 1200C. At 1330C the run-out was not 1200°c achieved. Even at a low creep stress of 50 MPa the rupture g1E05 time was 87h100 MPa, and 82% for the applied stress of 80 MPa. At 1000 1330C, presence of steam reduced creep lives by 96-98% Creep Stress(MPa) It is recognized that Nextel 720 fiber has the best creep Fig.13. Minimum creep rate as a function of applied stress at 1200 and performance of any commercially available polycrystalline 1330C in laboratory air and in steam environment Data for Nextel 720 oxide fiber. The superior high-temperature creep perfor fibers (Wilson [57])are also shown. mance of the Nextel720 fibers results from the high con tent of mullite, which has a much better creep resistance in air environment were close to the N720 fiber data than alumina [57]. Wannaparhum et al. [55] reported that adjusted for Vr=0. 22. For a given creep stress, creep rates exposure of the N720/A composite to water vapor at in steam were approximately an order of magnitude higher 1100 C could result in an increase in Al,O3 content of than those in air. Fitting the experimental results(obtained the fiber because of the loss of Sio, from its mullite phase in either air or steam) with a temperature-independent The loss of the mullite phase in the fiber may be the mech Norton-Bailey equation of the form anism behind the higher creep rates and reduced creep resistance observed in steam. However, further experiments ields the stress exponents that nsiderably higher quantifying the mullite loss from the fiber exposed to water than that reported for the N720 fibers alone [57]. It is be- vapor at high temperature would be required to ascertain lieved that the higher stress exponents are due to a contri- bution from the matrix. As expected, creep resistance 3. 4. Composite microstructure decreases dramatically with increasing temperature. Fc the creep stress of 100 MPa, creep rate at 1330C was al- most two orders of magnitude higher that at 1200C Fracture surface of a specimen tested in creep is shown The presence of steam further accelerates creep and de. in Fig. 15. It should be noted that the appearance of the fracture surface was not significantly affected by any of grades creep resistance. At 1330C, the presence of steam the following factors: test temperature(1200 vs 1330oC), increases the minimum creep rate by a factor of 100 Stress-rupture behavior is summarized in Fig test environment (air vs steam), test type(tension vs creep creep stress is plotted vs time to rupture at 1200 and 1330C vs tension-tension fatigue). Fracture surfaces of similar ere in air and in steam environments. At both temperatures appearance were produced in all tests. Micrographs in creep life decreases with increasing applied stress. For a ig. 15 are typical and representative of fracture surfaces obtained in all tests in this study. It is seen in Fig. 15(a) that the fracture plane is not well defined. The fibers in the 0o tows in each cloth layer exhibit UTS at1200°c 180 random failure producing fiber pull-out △1330°c,Ar shows that the 0 fiber tows break over a wide range of ▲1330°C. Steam axial locations, in general spanning the entire width of 1204 UTS at1330°c the specimen. The locations of the fiber breaks within an individual tow also exhibit a broad distribution, typically Ml mm in length, as seen in Fig. 15(c). It is important to note that no matrix holes were observed on the fracture surface. In conventional CFCCs with "weak"' interfaces the fiber pull-out results in formation of matrix holes, where broken fibers slide out of the matrix. However, in 1.E+00 1.E+01 1.E+02 1.E+03 1. E+04 1.E+05 1.E+06 1. E+07 leave matrix sockets but causes fragmentation of interven- Time(s) ing matrix in the region of strain localization. Some of the Fig. 14. Creep stress vs time to rupture at 1200 and 1330C in laborator matrix debris and matrix still bonded to the fibers are seen air and in steam environment in Fig. 15(d)
in air environment were close to the N720 fiber data adjusted for Vf = 0.22. For a given creep stress, creep rates in steam were approximately an order of magnitude higher than those in air. Fitting the experimental results (obtained in either air or steam) with a temperature-independent Norton–Bailey equation of the form e_ ¼ Arn yields the stress exponents that are considerably higher than that reported for the N720 fibers alone [57]. It is believed that the higher stress exponents are due to a contribution from the matrix. As expected, creep resistance decreases dramatically with increasing temperature. For the creep stress of 100 MPa, creep rate at 1330 C was almost two orders of magnitude higher that at 1200 C. The presence of steam further accelerates creep and degrades creep resistance. At 1330 C, the presence of steam increases the minimum creep rate by a factor of 100. Stress-rupture behavior is summarized in Fig. 14, where creep stress is plotted vs time to rupture at 1200 and 1330 C in air and in steam environments. At both temperatures, creep life decreases with increasing applied stress. For a given creep stress, creep-rupture life at 1330 C is considerably reduced compared to that at 1200 C. With the creep run-out condition defined as 100 h, 80 MPa was the runout stress at 1200 C. At 1330 C the run-out was not achieved. Even at a low creep stress of 50 MPa the rupture time was 87 h < 100 h. The short creep lives produced at 1330 C indicate that this oxide/oxide CMC should not be used under sustained loading at temperatures above 1200 C. Presence of steam dramatically reduced creep lives at both test temperatures. At 1200 C, reduction in creep life due to steam was at least 90% for applied stress levels P100 MPa, and 82% for the applied stress of 80 MPa. At 1330 C, presence of steam reduced creep lives by 96–98%. It is recognized that NextelTM720 fiber has the best creep performance of any commercially available polycrystalline oxide fiber. The superior high-temperature creep performance of the NextelTM720 fibers results from the high content of mullite, which has a much better creep resistance than alumina [57]. Wannaparhum et al. [55] reported that exposure of the N720/A composite to water vapor at 1100 C could result in an increase in Al2O3 content of the fiber because of the loss of SiO2 from its mullite phase. The loss of the mullite phase in the fiber may be the mechanism behind the higher creep rates and reduced creep resistance observed in steam. However, further experiments quantifying the mullite loss from the fiber exposed to water vapor at high temperature would be required to ascertain this. 3.4. Composite microstructure Fracture surface of a specimen tested in creep is shown in Fig. 15. It should be noted that the appearance of the fracture surface was not significantly affected by any of the following factors: test temperature (1200 vs 1330 C), test environment (air vs steam), test type (tension vs creep vs tension–tension fatigue). Fracture surfaces of similar appearance were produced in all tests. Micrographs in Fig. 15 are typical and representative of fracture surfaces obtained in all tests in this study. It is seen in Fig. 15(a) that the fracture plane is not well defined. The fibers in the 0 tows in each cloth layer exhibit random failure producing fiber ‘‘pull-out’’. Fig. 15(b) shows that the 0 fiber tows break over a wide range of axial locations, in general spanning the entire width of the specimen. The locations of the fiber breaks within an individual tow also exhibit a broad distribution, typically 1 mm in length, as seen in Fig. 15(c). It is important to note that no matrix holes were observed on the fracture surface. In conventional CFCCs with ‘‘weak’’ interfaces, the fiber pull-out results in formation of matrix holes, where broken fibers slide out of the matrix. However, in the present composite, the pull-out of the fibers does not leave matrix sockets but causes fragmentation of intervening matrix in the region of strain localization. Some of the matrix debris and matrix still bonded to the fibers are seen in Fig. 15(d). 1.E-09 1.E-08 1.E-07 1.E-06 1.E-05 1.E-04 1.E-03 1.E-02 10 100 1000 Creep Stress (MPa) Creep Rate (s-1) 1200°C, Air 1200°C, Steam 1330°C, Air 1330°C, Steam N720 Fiber, 1200°C Wilson, 2001 N720 Fiber at Vf = 0.22 1200°C Fig. 13. Minimum creep rate as a function of applied stress at 1200 and 1330 C in laboratory air and in steam environment. Data for Nextel 720 fibers (Wilson [57]) are also shown. 0 20 40 60 80 100 120 140 160 180 200 1.E+00 1.E+01 1.E+02 1.E+03 1.E+04 1.E+05 1.E+06 1.E+07 Time (s) Stress (MPa) 1200°C, Air 1200°C, Steam 1330°C, Air 1330°C, Steam UTS at 1200°C UTS at 1330°C Fig. 14. Creep stress vs time to rupture at 1200 and 1330 C in laboratory air and in steam environment. M.B. Ruggles-Wrenn et al. / Composites: Part A 37 (2006) 2029–2040 2037
M B. Ruggles-Wrenn et al. Composites: Part A 37(2006)2029-2040 (b) 1.0 mm 1.0mm 100pm Fig. 15. Fracture surface of a typical NextelM720/alumina CMC specimen tested in creep at 1200.C:(a)overall view showing general extent of fiber pullout, (b) fiber pullout across width of specimen, (c) fiber pullout within a 0 bundle, (d) matrix particles bonded to the fiber, (e)region of fairly coordinated fiber fracture in the 0 tow and(f) nearly planar fracture of the 90 fiber tows. tows exhibit flatter, more coordinated fracture topography. air and in 100% steam environment at 1200 and 133082 Fig. 15(e) shows that some fracture regions in the 0 fiber ceramic composite were characterized in laborator Close examination reveals that most of the fibers fracture on different planes, suggesting that a single crack front 4. 1. Fatigue behavior did not cause this fracture topography. However, one can see a few pairs of fibers(arrows)which exhibit planar frac Tension-tension fatigue behavior of the N720/A CFCC ture. These fiber pairs appear to have a common fracture was studied for fatigue stress levels of 100-170 MPa at such common fracture origin is produced during fiber pro- at 1330C. Results suggest the following conclusions. 9 origin where they touch. Haslam et al. [58] suggest that 1200C, and for fatigue stress levels of 50 and 100 M cessing, when adjacent fibers in the bundle stick to each other and sinter together along their cylindrical axis. a typ -(1)The N720 /A composite exhibits excellent fatigue ical fracture of the 90 fiber tows in a cloth layer is shown resistance in laboratory air at 1200C. The fatigue in Fig. 15(f). Here, the fracture surface topography can be on a run -out condition of 105 characterized as nearly planar. cycles)is 170 MPa(88% UTS at 1200C). The mate- rial retains 100% of its tensile strength. However 4. Concluding remarks considerable stiffness loss (30-50%)is observed (2)Presence of steam causes noticeable degradation in The tension-tension fatigue behavior and the creep-rup- fatigue performance at 1200C. The fatigue limit in ture behavior of the NextelM720/Alumina continuous steam environment is 125 MPa (65% UTS at
Fig. 15(e) shows that some fracture regions in the 0 tows exhibit flatter, more coordinated fracture topography. Close examination reveals that most of the fibers fracture on different planes, suggesting that a single crack front did not cause this fracture topography. However, one can see a few pairs of fibers (arrows) which exhibit planar fracture. These fiber pairs appear to have a common fracture origin where they touch. Haslam et al. [58] suggest that such common fracture origin is produced during fiber processing, when adjacent fibers in the bundle stick to each other and sinter together along their cylindrical axis. A typical fracture of the 90 fiber tows in a cloth layer is shown in Fig. 15(f). Here, the fracture surface topography can be characterized as nearly planar. 4. Concluding remarks The tension–tension fatigue behavior and the creep-rupture behavior of the NextelTM720/Alumina continuous fiber ceramic composite were characterized in laboratory air and in 100% steam environment at 1200 and 1330 C. 4.1. Fatigue behavior Tension–tension fatigue behavior of the N720/A CFCC was studied for fatigue stress levels of 100–170 MPa at 1200 C, and for fatigue stress levels of 50 and 100 MPa at 1330 C. Results suggest the following conclusions: (1) The N720/A composite exhibits excellent fatigue resistance in laboratory air at 1200 C. The fatigue limit in air (based on a run-out condition of 105 cycles) is 170 MPa (88% UTS at 1200 C). The material retains 100% of its tensile strength. However, considerable stiffness loss (30–50%) is observed. (2) Presence of steam causes noticeable degradation in fatigue performance at 1200 C. The fatigue limit in steam environment is 125 MPa (65% UTS at Fig. 15. Fracture surface of a typical NextelTM720/alumina CMC specimen tested in creep at 1200 C: (a) overall view showing general extent of fiber pullout, (b) fiber pullout across width of specimen, (c) fiber pullout within a 0 bundle, (d) matrix particles bonded to the fiber, (e) region of fairly coordinated fiber fracture in the 0 tow and (f) nearly planar fracture of the 90 fiber tows. 2038 M.B. Ruggles-Wrenn et al. / Composites: Part A 37 (2006) 2029–2040