Availableonlineatwww.sciencedirect.com Internationa ° Science Direct Journalof Fatique ELSEVIER International Journal of Fatigue 30(2008)502-516 www.elsevier.com/locate/ijfatigue Effects of frequency and environment on fatigue behavior of an oxide-oxide ceramic composite at1200°C查 M. B. Ruggles-Wrenn a, *. G. Hetrick a. SS.Baek Department of Aeronautics and Astronautics, Air Force Institute of Technology Patterson Air Force Base. OH 45433-7765. USA Agency for Defense Decelopment, Daejeo Received 7 August 2006: received in revised form 3 April accepted 8 April 2007 Available online 27 April 2007 Abstract The effect of frequency on fatigue behavior of an oxide-oxide continuous fiber ceramic composite( CFCC) was investigated at 1200C In aboratory air and in steam environment. The composite consists of a porous alumina matrix reinforced with laminated, woven mull- ite-alumina(NextelM720)fibers, has no interface between the fiber and matrix, and relies on the porous matrix for flaw tolerance Tension-tension fatigue tests were performed at frequencies of 0. I and 10 Hz for fatigue stresses ranging from 75 to 170 MPa. Fatigue run-out was defined as 10 cycles at the frequency of 0. 1 Hz and as 10 cycles at the frequency of 10 Hz. The CFCC exhibited excellent fatigue resistance 170 MPa(88% UTS at 1200C). The material retained 100% of its tensile strength. Presence of steam significantly degraded the fatigue performance, with the degradation being most pronounced at 0. 1 Hz. Com posite microstructure, as well as damage and failure mechanisms were investigated. Examination of fracture surfaces revealed higher degrees of fiber pull-out in specimens tested at 10 Hz, indicating weakening of the fiber/matrix interface. a qualitative spectral anal showed evidence of silicon species migration from the fiber to the matrix. Published by elsevier Ltd Keywords: Ceramic-matrix composites(CMCs): Oxides; Fatigue: High-temperature properties; Mechanical testing: Fractography 1. Introduction temperatures, together with a reduced need for cooling air, allow for improved high-temperature performance Advances in aerospace technologies have raised the when compared to conventional nickel-based superalloy mand for structural materials that exhibit superior [2]. Advanced reusable space launch vehicles will likely long-term mechanical properties and retained properties incorporate CMCs in critical propulsion components [3]. under high temperature, high pressure, and varying envi- However, these applications require exposure to oxidizing ronmental factors [1]. Ceramic-matrix composites(CMCs), environments. Therefore the thermodynamic stability and capable of maintaining excellent strength and fracture oxidation resistance of CMCs are vital issues. toughness at high temperatures, continue to attract atten Non-oxide fiber/ non-oxide matrix composites generally tion as candidate materials for such applications. Addition- exhibit poor oxidation resistance [4, 5], particularly at inter ally, the lower densities of CMCs and their higher use mediate temperatures(800C). The degradation involves oxidation of fibers, fiber coatings, and matrices and is typ- ically accelerated by the presence of moisture [6-8]. Using a w The views expressed are those of the authors and do not reflect the non-c oxide fiber/oxide matrix or oxide fiber/non-oxide official policy or position of the United States Air Force, Department of matrix composites generally does not substantially improve Defense or the US government the high temperature oxidation resistance [9]. The need for Corresponding author. Tel: +l 937 255 3636x4641: fax: +1 937 656 environmentally stable composites motivated the develop- E-mail address: marina. ruggles-wrenn(@afit. edu(M.B. Ruggles. ment of CMCs based on environmentally stable oxide con- stituents [10-18] 0142-1123S-see front matter Published by Elsevier Ltd. doi:10.1016/ .fatigue.200704004
Effects of frequency and environment on fatigue behavior of an oxide–oxide ceramic composite at 1200 C q M.B. Ruggles-Wrenn a,*, G. Hetrick a , S.S. Baek b a Department of Aeronautics and Astronautics, Air Force Institute of Technology, Wright-Patterson Air Force Base, OH 45433-7765, USA b Agency for Defense Development, Daejeon, South Korea Received 7 August 2006; received in revised form 3 April 2007; accepted 8 April 2007 Available online 27 April 2007 Abstract The effect of frequency on fatigue behavior of an oxide–oxide continuous fiber ceramic composite (CFCC) was investigated at 1200 C in laboratory air and in steam environment. The composite consists of a porous alumina matrix reinforced with laminated, woven mullite–alumina (Nextel720) fibers, has no interface between the fiber and matrix, and relies on the porous matrix for flaw tolerance. Tension–tension fatigue tests were performed at frequencies of 0.1 and 10 Hz for fatigue stresses ranging from 75 to 170 MPa. Fatigue run-out was defined as 105 cycles at the frequency of 0.1 Hz and as 106 cycles at the frequency of 10 Hz. The CFCC exhibited excellent fatigue resistance in laboratory air. The fatigue limit was 170 MPa (88% UTS at 1200 C). The material retained 100% of its tensile strength. Presence of steam significantly degraded the fatigue performance, with the degradation being most pronounced at 0.1 Hz. Composite microstructure, as well as damage and failure mechanisms were investigated. Examination of fracture surfaces revealed higher degrees of fiber pull-out in specimens tested at 10 Hz, indicating weakening of the fiber/matrix interface. A qualitative spectral analysis showed evidence of silicon species migration from the fiber to the matrix. Published by Elsevier Ltd. Keywords: Ceramic–matrix composites (CMCs); Oxides; Fatigue; High-temperature properties; Mechanical testing; Fractography 1. Introduction Advances in aerospace technologies have raised the demand for structural materials that exhibit superior long-term mechanical properties and retained properties under high temperature, high pressure, and varying environmental factors [1]. Ceramic–matrix composites (CMCs), capable of maintaining excellent strength and fracture toughness at high temperatures, continue to attract attention as candidate materials for such applications. Additionally, the lower densities of CMCs and their higher use temperatures, together with a reduced need for cooling air, allow for improved high-temperature performance when compared to conventional nickel-based superalloys [2]. Advanced reusable space launch vehicles will likely incorporate CMCs in critical propulsion components [3]. However, these applications require exposure to oxidizing environments. Therefore the thermodynamic stability and oxidation resistance of CMCs are vital issues. Non-oxide fiber/non-oxide matrix composites generally exhibit poor oxidation resistance [4,5], particularly at intermediate temperatures (800 C). The degradation involves oxidation of fibers, fiber coatings, and matrices and is typically accelerated by the presence of moisture [6–8]. Using a non-oxide fiber/oxide matrix or oxide fiber/non-oxide matrix composites generally does not substantially improve the high temperature oxidation resistance [9]. The need for environmentally stable composites motivated the development of CMCs based on environmentally stable oxide constituents [10–18]. 0142-1123/$ - see front matter Published by Elsevier Ltd. doi:10.1016/j.ijfatigue.2007.04.004 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 US Government. * Corresponding author. Tel.: +1 937 255 3636x4641; fax: +1 937 656 4032. E-mail address: marina.ruggles-wrenn@afit.edu (M.B. RugglesWrenn). www.elsevier.com/locate/ijfatigue Available online at www.sciencedirect.com International Journal of Fatigue 30 (2008) 502–516 International Journalof Fatigue
M.B. Ruggles-Wrenn et al. International Journal of fatigue 30(2008)502-516 503 It is widely accepted that in order to avoid brittle frac- more dramatic deterioration of fatigue performance, reduc- ture behavior in CMCs and improve the damage tolerance, ing the time to failure in fatigue to the much shorter failure a weak fiber/matrix interface is needed, which serves to times observed in creep These results suggest that time- deflect matrix cracks and to allow subsequent fiber pull- dependent damage processes associated with creep may t[19-22]. It has been demonstrated that similar crack- influence cyclic lifetime. Therefore loading frequency is deflecting behavior can be achieved by using a matrix with likely to have an effect on fatigue durability of the finely distributed porosity instead of a separate interface N720/A composite between matrix and fibers [23]. This concept has been suc- The objective of this study is to investigate effects of cessfully demonstrated for oxide-oxide composites loading frequency and steam environment on fatigue [10, 14, 18, 24-28]. Resulting oxide/oxide CMCs exhibit behavior of N720/A, an oxide-oxide CMC, at 1200C in damage tolerance combined with inherent oxidation resis- air and in steam environments at the loading frequencies tance. An extensive review of the mechanisms and mechan- of 0.1-10 Hz. Results show that the loading frequenc ous matrix CMCs is given in [29] has a marked effect on fatigue life, especially in steam envi- ical properties o ial applications, CFCCs will be subject ronment. The composite microstructure, as well as damage In many potential ap to fatigue loading under a wide range of frequencies. Sev- and failure mechanisms are discussed. eral studies examined high-temperature fatigue perfor mance of CMCs at loading frequencies <10 Hz [30-36]. 2. Experimental procedure At higher frequencies(ranging from 10 to 375 Hz), a strong fect of loading frequency on fatigue life has been demon- 2. 1. Material strated for CMCs with weak fiber-matrix interfaces tested at room temperature [37-39). It was reported that fatigue The material studied was NextelTM720/Alumina life decreased sharply as the loading frequency increased. (N720/A), a commercially available oxide-oxide ceramic This decrease in fatigue life was attributed to frictional composite(COI Ceramics, San Diego, CA), consisting of heating and interface and fiber damage. More recently, it a porous alumina(Al2O3) matrix reinforced with Nex has been shown that the room-temperature fatigue life of telTM720 mullite-alumina fibers composed of 85% Al2O3 certain ceramic-matrix composites with a strong fiber- and 15% Sio2 by weight. The composite was supplied in matrix interface shows little dependence on the loading fre- a form of 2.8 mm thick plates, comprised of twelve 0/900 quency [40]. Vanswijgenhoven et al. [41] found that at woven layers, with a density of 2.77 g/cm and a fiber vol 1200C the fatigue limit of a Nicalon-fabric-reinforced ume of approximately 45%. Matrix porosity was N24% CMC was unaffected by the loading frequency, while the The fiber fabric was infiltrated with the matrix in a sol number of cycles to failure increased and the time to failure gel process. The laminate was dried with a"vacuum bag decreased with increase in frequency technique under low pressure and low temperature, ther Porous matrix oxide/oxide CMCs exhibit several behav- pressureless sintered [46]. No coating was applied to the ior trends that are distinctly different from those exhibited fibers. The damage tolerance of the N720/A composite is by traditional CMCs with a fiber-matrix interface. Most enabled by a porous matrix. Representative micrographs Sic-fiber-containing CMCs exhibit longer life under static of the untested material are presented in Fig. I. Fig. la loading and shorter life under cyclic loading [42]. For these shows 00 and 90 fiber tows as well as numerous matrix materials, fatigue is significantly more damaging than cracks. In the case of the as-processed material, most are creep. Zawada et al. [43] examined the high-temperature shrinkage cracks formed during processing rather than mechanical behavior of a porous matrix Nextel 610/Alumi- matrix cracks generated during loading. Porous nature of nosilicate composite Results revealed excellent fatigue per- the matrix is seen in Fig. Ib formance at 1000C. Conversely, creep lives were short indicating low creep resistance and limiting the use of that 2. 2. Mechanical testing CMC to temperatures below 1000C. Ruggles-Wrenn et al [44] demonstrated that NextelTM720/Alumina(N720/A) A servocontrolled MTS mechanical testing machine composite exhibits excellent fatigue resistance in laboratory equipped with hydraulic water-cooled collet grips, a com air at 1200C. The fatigue limit(based on a run-out condi- pact two-zone resistance-heated furnace, and two tempera tion of 105 cycles)was 170 MPa(88% UTS at 1200C). ture controllers was used in all tests. An MTS TestStar Furthermore, the composite retained 100% of its tensile digital controller was employed for input signal generation strength. However, creep loading was found to be consid- and data acquisition. Strain measurement was accom- erably more damaging. Creep run-out (defined as 100 h plished with an MTS high-temperature air-cooled uniaxial at creep stress) was achieved only at stress levels belov extensometer. For elevated temperature testing, thermo- 50% UTS Mehrman et al. [45] reported that introduction couples were bonded to the specimens using alumina of a short hold period at the maximum stress into the fati- cement(Zircar)to calibrate the furnace on a periodic basis gue cycle significantly degraded the fatigue performance of The furnace controller(using a non-contacting thermocou- N720/A composite at 1200C in air. In steam, superpose- ple exposed to the ambient environment near the test spec tion of a hold time onto a fatigue cycle resulted in an even imen) was adjusted to determine the power setting needed
It is widely accepted that in order to avoid brittle fracture behavior in CMCs and improve the damage tolerance, a weak fiber/matrix interface is needed, which serves to deflect matrix cracks and to allow subsequent fiber pullout [19–22]. It has been demonstrated that similar crackdeflecting behavior can be achieved by using a matrix with finely distributed porosity instead of a separate interface between matrix and fibers [23]. This concept has been successfully demonstrated for oxide–oxide composites [10,14,18,24–28]. Resulting oxide/oxide CMCs exhibit damage tolerance combined with inherent oxidation resistance. An extensive review of the mechanisms and mechanical properties of porous matrix CMCs is given in [29]. In many potential applications, CFCCs will be subject to fatigue loading under a wide range of frequencies. Several studies examined high-temperature fatigue performance of CMCs at loading frequencies 610 Hz [30–36]. At higher frequencies (ranging from 10 to 375 Hz), a strong effect of loading frequency on fatigue life has been demonstrated for CMCs with weak fiber–matrix interfaces tested at room temperature [37–39]. It was reported that fatigue life decreased sharply as the loading frequency increased. This decrease in fatigue life was attributed to frictional heating and interface and fiber damage. More recently, it has been shown that the room-temperature fatigue life of certain ceramic–matrix composites with a strong fiber– matrix interface shows little dependence on the loading frequency [40]. Vanswijgenhoven et al. [41] found that at 1200 C the fatigue limit of a Nicalon-fabric-reinforced CMC was unaffected by the loading frequency, while the number of cycles to failure increased and the time to failure decreased with increase in frequency. Porous matrix oxide/oxide CMCs exhibit several behavior trends that are distinctly different from those exhibited by traditional CMCs with a fiber–matrix interface. Most SiC-fiber-containing CMCs exhibit longer life under static loading and shorter life under cyclic loading [42]. For these materials, fatigue is significantly more damaging than creep. Zawada et al. [43] examined the high-temperature mechanical behavior of a porous matrix Nextel 610/Aluminosilicate composite. Results revealed excellent fatigue performance at 1000 C. Conversely, creep lives were short, indicating low creep resistance and limiting the use of that CMC to temperatures below 1000 C. Ruggles-Wrenn et al. [44] demonstrated that Nextel720/Alumina (N720/A) composite exhibits excellent fatigue resistance in laboratory air at 1200 C. The fatigue limit (based on a run-out condition of 105 cycles) was 170 MPa (88% UTS at 1200 C). Furthermore, the composite retained 100% of its tensile strength. However, creep loading was found to be considerably more damaging. Creep run-out (defined as 100 h at creep stress) was achieved only at stress levels below 50% UTS. Mehrman et al. [45] reported that introduction of a short hold period at the maximum stress into the fatigue cycle significantly degraded the fatigue performance of N720/A composite at 1200 C in air. In steam, superposition of a hold time onto a fatigue cycle resulted in an even more dramatic deterioration of fatigue performance, reducing the time to failure in fatigue to the much shorter failure times observed in creep. These results suggest that timedependent damage processes associated with creep may influence cyclic lifetime. Therefore loading frequency is likely to have an effect on fatigue durability of the N720/A composite. The objective of this study is to investigate effects of loading frequency and steam environment on fatigue behavior of N720/A, an oxide–oxide CMC, at 1200 C in air and in steam environments at the loading frequencies of 0.1–10 Hz. Results show that the loading frequency has a marked effect on fatigue life, especially in steam environment. The composite microstructure, as well as damage and failure mechanisms are discussed. 2. Experimental procedure 2.1. Material The material studied was Nextel720/Alumina (N720/A), a commercially available oxide–oxide ceramic composite (COI Ceramics, San Diego, CA), consisting of a porous alumina (Al2O3) matrix reinforced with Nextel720 mullite–alumina fibers composed of 85% Al2O3 and 15% SiO2 by weight. The composite was supplied in a form of 2.8 mm thick plates, comprised of twelve 0/90 woven layers, with a density of 2.77 g/cm3 and a fiber volume of approximately 45%. Matrix porosity was 24%. The fiber fabric was infiltrated with the matrix in a sol– gel process. The laminate was dried with a ‘‘vacuum bag’’ technique under low pressure and low temperature, then pressureless sintered [46]. No coating was applied to the fibers. The damage tolerance of the N720/A composite is enabled by a porous matrix. Representative micrographs of the untested material are presented in Fig. 1. Fig. 1a shows 0 and 90 fiber tows as well as numerous matrix cracks. In the case of the as-processed material, most are shrinkage cracks formed during processing rather than matrix cracks generated during loading. Porous nature of the matrix is seen in Fig. 1b. 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 using alumina cement (Zircar) 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 M.B. Ruggles-Wrenn et al. / International Journal of Fatigue 30 (2008) 502–516 503
M. B. Ruggles-Wrenn et al. International Journal of Fatigue 30(2008)502-516 Fig. 1. As-received material: (a)overview, optical microscope and (b) porous nature of the matrix is evident (SEM to achieve the desired temperature of the test specimen. Fatigue run-out was defined as 10 cycles at 0.1 and The determined power setting was then used in actual tests. 1.0 Hz, and as 10 cycles at 10 Hz. The 10 cycle count The power setting for testing in steam was determined by value represents the number of loading cycles expected in placing the specimen instrumented with thermocouples in aerospace applications at that temperature. Fatigue run- steam environment and repeating the furnace calibration out limits were defined as the highest stress level, for which procedure. Thermocouples were not bonded to the test run-out was achieved. Note that in the case of all run-out specimens after the furnace was calibrated. Tests in steam tests, the failure of specimen did not occur when the test environment employed an alumina susceptor (tube with was terminated. Cyclic stress-strain data were recorded end caps), which fits inside the furnace. The specimen gage throughout each test. Thus stiffness degradation as well section is located inside the susceptor, with the ends of the as strain accumulation with fatigue cycles and/or time specimen passing through slots in the susceptor. Steam is could be examined. All specimens that achieved run-out introduced into the susceptor(through a feeding tube) in were subjected to tensile test to failure at 1200C in labo- a continuous stream with a slightly positive pressure, expel- ratory air to determine the retained strength and stifness ling the dry air and creating 100% steam environment inside the susceptor(see Fig. 2). 23. Microstructural characterization All tests were performed at 1200C. In all tests, a spec imen was heated to test temperature in 25 min, and held at Fracture surfaces of failed specimens were examined temperature for additional 15 min prior to testing. Dog using SEM(FEI Quanta 200 HV) as well as an optical bone shaped specimens of 152 mm total length with a 10- microscope(Zeiss Discovery V12). The SEM specimen mm-wide gage section shown in Fig. 2 were used in all tests. were carbon coated. In addition, energy-dispersive X-ray Tensile tests were performed in stroke control with a con- spectroscopy (EDS) analysis was performe using an stant displacement rate of 0.05 mm/s at 1200C in labora- EDAX Genesis 4000 EDS system tory air. The effects of frequency on the fatigue behavior were evaluated in tension-tension fatigue tests conducted 3 Results and discussion at the frequencies of 0. 1 and 10 Hz at 1200oC, in labora tory air and in steam environments. Fatigue data at 3.1. Monotonic tension 1.0 Hz from prior work [44] is included for comparison All fatigue experiments were carried out in load control Tensile results obtained at 1200C were consistent with with the ratio R(minimum to maximum stress) of 0.05. those reported earlier [44, 47]. The ultimate tensile strength (UTS)was 190 MPa, elastic modulus, 76 GPa, and failure strain, 0.38%. It is worthy of note that in all tests reported herein, the failure occurred within the gage section of the extensometer Creep was shown to be considerably more damaging than cyclic loading to oxide-oxide CMCs with porous Fig. 2. Test specimen, dimensions matrix [43, 44]. Recently Ehrman et al. [45] demonstrated
to achieve the desired temperature of the test specimen. The determined power setting was then used in actual tests. The power setting for testing in steam was determined by placing the specimen instrumented with thermocouples in 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 (see Fig. 2). All tests were performed at 1200 C. 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 shown in Fig. 2 were used in all tests. Tensile tests were performed in stroke control with a constant displacement rate of 0.05 mm/s at 1200 C in laboratory air. The effects of frequency on the fatigue behavior were evaluated in tension–tension fatigue tests conducted at the frequencies of 0.1 and 10 Hz at 1200 C, in laboratory air and in steam environments. Fatigue data at 1.0 Hz from prior work [44] is included for comparison. All fatigue experiments were carried out in load control with the ratio R (minimum to maximum stress) of 0.05. Fatigue run-out was defined as 105 cycles at 0.1 and 1.0 Hz, and as 106 cycles at 10 Hz. The 105 cycle count value represents the number of loading cycles expected in aerospace applications at that temperature. Fatigue runout limits were defined as the highest stress level, for which run-out was achieved. Note that in the case of all run-out tests, the failure of specimen did not occur when the test was terminated. Cyclic stress–strain data were recorded throughout each test. Thus stiffness degradation as well as strain accumulation with fatigue cycles and/or time could be examined. All specimens that achieved run-out were subjected to tensile test to failure at 1200 C in laboratory air to determine the retained strength and stiffness. 2.3. Microstructural characterization Fracture surfaces of failed specimens were examined using SEM (FEI Quanta 200 HV) as well as an optical microscope (Zeiss Discovery V12). The SEM specimens were carbon coated. In addition, energy-dispersive X-ray spectroscopy (EDS) analysis was performed using an EDAX Genesis 4000 EDS system. 3. Results and discussion 3.1. Monotonic tension Tensile results obtained at 1200 C were consistent with those reported earlier [44,47]. The ultimate tensile strength (UTS) was 190 MPa, elastic modulus, 76 GPa, and failure strain, 0.38%. It is worthy of note that in all tests reported herein, the failure occurred within the gage section of the extensometer. 3.2. Tension–tension fatigue Creep was shown to be considerably more damaging than cyclic loading to oxide–oxide CMCs with porous matrix [43,44]. Recently Mehrman et al. [45] demonstrated Fig. 1. As-received material: (a) overview, optical microscope and (b) porous nature of the matrix is evident (SEM). R=50 50.0 76.0 8.0 9.0 5.0 Fig. 2. Test specimen, dimensions in mm. 504 M.B. Ruggles-Wrenn et al. / International Journal of Fatigue 30 (2008) 502–516
M.B. Ruggles-Wrenn et al. International Journal of fatigue 30(2008)502-516 Summary of fatigue results for the N720/A composite at 1200C, in laboratory air and steam environments Test environment Max stress(MPa) Cycles to failure Time to failure(h) Fatigue at 0.1 HE Laboratory air 193 Laboratory air 136,1212 2.132 Steam 56,093 35 Steam 48.6 115 150 75 21 12 0.03 Fatigue at 1.0 Hz 33. Laboratory air 40.7 1.144 Laboratory air ,47 52 Laboratory 09.436 25 Steam 100,780 0 0.714 Steam 08 Steam 150 l1,782 Steam 02 0.06 0.81 fatigue at 10 Hz 1,0000102 0.77 Run-out, failure of specimen did not occur when the test was terminated Data from Ruggles-Wrenn et al. [44] that introduction of a short hold period(at the maximum Furthermore, in steam the influence of the loading fre- stress)into the fatigue cycle significantly degraded the fati- quency on fatigue life becomes dramatic. In laboratory gue performance of N720/A composite at 1200C in air air, the high 170 MPa fatigue limit was obtained at both and in steam. These results suggest that the loading rate 0.1 and 1.0 Hz In steam, the fatigue performance was best plays a significant role in damage development. Therefore at the frequency of 10 Hz. Yet, even at 10 Hz the in-steam investigation of the effects of frequency on fatigue behav- fatigue limit is only 150 MPa(78% UTS at 1200 C), ior, especially when conducted in steam environment is noticeably lower than what could be expected in air. At critical to assessing the durability of a given porous matrix the loading frequency of 1.0 Hz, the fatigue limit drops oxide-oxide CMC to 125 MPa(69% UTS at 1200C). As the frequency Tension-tension fatigue tests were conducted at the fre- decreases by another order of magnitude, the fatigue per quencies of 0. I and 10 Hz at 1200C in air and in steam. formance deteriorates drastically. At 0.1 Hz, run-out was Results are summarized in Table l. Results are also pre- not achieved even at the low stress level of 75 MPa(39% sented in Fig. 3 as the stress vs. time to failure curves. UTS at 1200C) Results of fatigue tests at 1.0 Hz from the prior study [44]are included in Table I and in Fig 3 for comparison Data in Table I show that the loading frequency has little T=1200'C, Steam ffect on fatigue performance in air. For the frequencies of 0 I and 1.0 Hz the fatigue limit in air was 170 MPa(88% UTS at 1200 C UTS at 1200C). This fatigue limit is based on the run- out condition of 10 cycles, approximate number of load ing cycles expected in aerospace applications at 1200C. could have resulted in a lower fatigue limit. Because the o 0 It is recognized that a more rigorous run-out condition fatigue performance was expected to improve with increas- ing loading frequency, no tests were conducted in air at the 01.0 Hz, Ruggles-Wrenn et al, 2006 口10Hz frequency of 10 Hz. In view of the excellent fatigue resis- tance and high fatigue limit obtained in air at the frequen cies 0. 1 and 1.0 Hz, an equally high in-air fatigue limit ould be anticipated at 10 Hz. Presence of steam causes noticeable degradation in fatigue performance. At all load- Fig 3. Fatigue S-N curves for NextelT720/alumina ceramic composite at ing frequencies investigated, the fatigue limits obtained 1200C in steam environment. Fatigue data at 1.0 Hz from ruggles Wrenn et al. [44]. Arrow indicates that failure of specimen did not occur steam are significantly lower than those obtained in air. when the test was terminated
that introduction of a short hold period (at the maximum stress) into the fatigue cycle significantly degraded the fatigue performance of N720/A composite at 1200 C in air and in steam. These results suggest that the loading rate plays a significant role in damage development. Therefore investigation of the effects of frequency on fatigue behavior, especially when conducted in steam environment is critical to assessing the durability of a given porous matrix oxide–oxide CMC. Tension–tension fatigue tests were conducted at the frequencies of 0.1 and 10 Hz at 1200 C in air and in steam. Results are summarized in Table 1. Results are also presented in Fig. 3 as the stress vs. time to failure curves. Results of fatigue tests at 1.0 Hz from the prior study [44] are included in Table 1 and in Fig. 3 for comparison. Data in Table 1 show that the loading frequency has little effect on fatigue performance in air. For the frequencies of 0.1 and 1.0 Hz the fatigue limit in air was 170 MPa (88% UTS at 1200 C). This fatigue limit is based on the runout condition of 105 cycles, approximate number of loading cycles expected in aerospace applications at 1200 C. It is recognized that a more rigorous run-out condition could have resulted in a lower fatigue limit. Because the fatigue performance was expected to improve with increasing loading frequency, no tests were conducted in air at the frequency of 10 Hz. In view of the excellent fatigue resistance and high fatigue limit obtained in air at the frequencies 0.1 and 1.0 Hz, an equally high in-air fatigue limit could be anticipated at 10 Hz. Presence of steam causes noticeable degradation in fatigue performance. At all loading frequencies investigated, the fatigue limits obtained in steam are significantly lower than those obtained in air. Furthermore, in steam the influence of the loading frequency on fatigue life becomes dramatic. In laboratory air, the high 170 MPa fatigue limit was obtained at both 0.1 and 1.0 Hz. In steam, the fatigue performance was best at the frequency of 10 Hz. Yet, even at 10 Hz the in-steam fatigue limit is only 150 MPa (78% UTS at 1200 C), noticeably lower than what could be expected in air. At the loading frequency of 1.0 Hz, the fatigue limit drops to 125 MPa (69% UTS at 1200 C). As the frequency decreases by another order of magnitude, the fatigue performance deteriorates drastically. At 0.1 Hz, run-out was not achieved even at the low stress level of 75 MPa (39% UTS at 1200 C). Table 1 Summary of fatigue results for the N720/A composite at 1200 C, in laboratory air and steam environments Test environment Max. stress (MPa) Cycles to failure Time to failure (h) Failure strain (%) Fatigue at 0.1 Hz Laboratory air 170 100,017a 278a 1.93a Laboratory air 170 136,121a 378a 2.13a Steam 75 56,093 156 3.35 Steam 100 17,498 48.6 1.80 Steam 125 1850 5.14 1.15 Steam 150 75 0.21 0.67 Steam 170 12 0.03 0.53 Fatigue at 1.0 Hzb Laboratory air 100 120,199a 33.4a 0.63a Laboratory air 125 146,392a 40.7a 1.14a Laboratory air 150 167,473a 46.5a 1.66a Laboratory air 170 109,436a 30.4a 2.25a Steam 100 100,780a 28.0a 0.71a Steam 125 166,326a 46.2a 1.08a Steam 150 11,782 3.27 1.12 Steam 170 202 0.06 0.81 Fatigue at 10 Hz Steam 150 1,000,010a 27.8a 0.77a Steam 170 11,387 0.32 1.03 a Run-out, failure of specimen did not occur when the test was terminated. b Data from Ruggles-Wrenn et al. [44]. 0 50 100 150 200 250 0.01 0.1 1 10 100 1000 Time to failure (h) ) aP M( ssert S 0.1 Hz 1.0 Hz, Ruggles-Wrenn et al, 2006 10 Hz UTS at 1200 ˚C T = 1200 ˚C, Steam Fig. 3. Fatigue S–N curves for Nextel720/alumina ceramic composite at 1200 C in steam environment. Fatigue data at 1.0 Hz from RugglesWrenn et al. [44]. Arrow indicates that failure of specimen did not occur when the test was terminated. M.B. Ruggles-Wrenn et al. / International Journal of Fatigue 30 (2008) 502–516 505
506 M.B. Ruggles-Wrenn et al. International Journal of Fatigue 30 (2008)502-516 Evolution of the hysteresis response of N720/A with Fig 4 reveal that ratcheting, defined as progressive increase fatigue cycles is typified in Fig. 4, which shows hysteresis in accumulated strain with increasing number of cycles stress-strain loops for tests conducted in steam at various continues throughout the test. Effects of loading frequency loading frequencies In all tests, regardless of the frequency, and test environment on hysteresis response are illustrated the most extensive damage occurs on the first cycle, where in Figs. 5a and b, respectively. It is seen that the permanent considerable permanent strain is seen upon unloading. strain produced during the first cycle decreases with Afterwards hysteresis loops stabilize quickly. Results in increasing frequency. It is also seen that larger permanent strain is produced in steam than in air Of importance in cyclic fatigue is the reduction in stifl- ness(hysteresis modulus determined from the maximum T=1200C, Steam and minimum stress-strain data points during a loa Omax= 125 MPa f=0.1 Hz cycle), reflecting the damage development during fatigue cycling. Change in modulus is shown in Fig. 6, where nor malized modulus (i.e. modulus normalized by the modulus obtained in the second cycle) is plotted vs. fatigue cycles. The first cycle was not used for normalization because of Cycle 1 the large permanent strain offset upon unloading. It is note- 00 ycle 50 worthy that although all in-air tests achieved run-out, a decrease in normalized modulus with cycling was still observed(Fig 6a) Modulus loss increased with increasing fatigue stress level. In air at the frequency of 1.0 Hz, the 1251.50 Strain(%) a200 1.0 Hz. Ruggles-Wrenn, 2006 T=1200C, Steam Cycle Omax= 125 MPa Cycle 2 Cycle 1025 T=1200 C, Steam 0250.500.751.00125 0.50 00 Strain (%) Strain(%) 1200℃c, Steam Cycle 1 T=1200"c Cycle 1000 gma=170 MPa 1.001.25150 0.25 0.75 1.00 Strain (%) Strain(%) Fig 4. Typical evolution of stress-strain hysteresis response of N720/A Fig. 5. The stress-strain response of N720/A ceramic composite at with fatigue cycles at 1200C in steam:(a)at 0. I Hz and 125 MPa, (b)at 1200C: (a)in steam environment at three different loading frequencies 1.0 Hz and 125 MPa, data from Ruggles-Wrenn et al. [44](c) at 10 Hz and (b) at 0. I Hz in air and in steam environments. Curves shifted by O1% and 170 MPa for clarity. Data at 1.0 Hz from Ruggles- Wrenn et al. [44]
Evolution of the hysteresis response of N720/A with fatigue cycles is typified in Fig. 4, which shows hysteresis stress–strain loops for tests conducted in steam at various loading frequencies. In all tests, regardless of the frequency, the most extensive damage occurs on the first cycle, where considerable permanent strain is seen upon unloading. Afterwards hysteresis loops stabilize quickly. Results in Fig. 4 reveal that ratcheting, defined as progressive increase in accumulated strain with increasing number of cycles, continues throughout the test. Effects of loading frequency and test environment on hysteresis response are illustrated in Figs. 5a and b, respectively. It is seen that the permanent strain produced during the first cycle decreases with increasing frequency. It is also seen that larger permanent strain is produced in steam than in air. 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 fatigue cycling. Change in modulus is shown in Fig. 6, where normalized modulus (i.e. modulus normalized by the modulus obtained in the second cycle) is plotted vs. fatigue cycles. The first cycle was not used for normalization because of the large permanent strain offset upon unloading. It is noteworthy that although all in-air tests achieved run-out, a decrease in normalized modulus with cycling was still observed (Fig. 6a). Modulus loss increased with increasing fatigue stress level. In air at the frequency of 1.0 Hz, the 0.00 0.25 0.50 0.75 1.00 1.25 1.50 Strain (%) T = 1200 ˚C, Steam max = 125 MPa f = 0.1 Hz Cycle 1000 Cycle 1 Cycle 50 Cycle 1820 0 50 100 150 0.00 0.25 0.50 0.75 1.00 1.25 1.50 ) aP M( ssert S ) aP M( ssert S 0 50 100 150 ) aP M( ssert S T = 1200 ˚C, Steam max = 125 MPa f = 1.0 Hz Cycle 100000 Cycle 10000 Cycle 1025 Cycle 25 Cycle 2 Cycle 1 a b c 0 50 100 150 200 0.00 0.25 0.50 0.75 1.00 1.25 1.50 Strain (%) Strain (%) T = 1200 ˚C, Steam σmax = 170 MPa f = 10 Hz Cycle 1 Cycle 10000 Cycle 1000 Cycle 50 Cycle 10 Fig. 4. Typical evolution of stress–strain hysteresis response of N720/A with fatigue cycles at 1200 C in steam: (a) at 0.1 Hz and 125 MPa, (b) at 1.0 Hz and 125 MPa, data from Ruggles-Wrenn et al. [44], (c) at 10 Hz and 170 MPa. 0 50 100 150 200 0.00 0.25 0.50 0.75 1.00 Strain (%) ) aP M( ssert S T = 1200 ˚C, Steam σmax = 170 MPa 0.1 Hz 10 Hz 1.0 Hz, Ruggles-Wrenn, 2006 0.00 0.25 0.50 0.75 1.00 Strain (%) T = 1200 ˚C σmax = 170 MPa f = 0.1 Hz Air Steam a 0 50 100 150 200 ) aP M( ssert S b Fig. 5. The stress–strain response of N720/A ceramic composite at 1200 C: (a) in steam environment at three different loading frequencies and (b) at 0.1 Hz in air and in steam environments. Curves shifted by 0.1% for clarity. Data at 1.0 Hz from Ruggles-Wrenn et al. [44]. 506 M.B. Ruggles-Wrenn et al. / International Journal of Fatigue 30 (2008) 502–516
M.B. Ruggles-Wrenn et al. International Journal of fatigue 30(2008)502-516 modulus loss is likely due to the early failure of fiber bun 100MPa,1.0H125MPa,1.0Hz dles. while the modulus loss observed in other tests con 150MPa,1.0÷170MPa,1.0Hz ducted at 0.1 and 1.0 Hz is likely due to the progressive cracking of the matrix. Increase in frequency by another order of magnitude to 10 Hz has an even more appreciable 日1.04 normalized modulus was only 5% in the 150 MPa test, which achieved a run-out. and 8% in the 170 MPa test Continuous decrease in modulus observed both in air and in steam suggests progressive damage with continued T=1200°c,Air cycling. Because the fatigue damage is still evolving at 10 cycles (10 cycles at 10 Hz), the 10(10 at 10 Hz) fati 1.E+001E+011.E+021.E+031.E+041.E+051E+061.E+07 gue limit does not meet the criteria of a true endurance fati Cycles(N) gue limit proposed by Sorensen et al. [48 ]and may not be a true endurance fatigue limit 91目901h+12Ma10h 75 MPa 0.1 Hz Maximum cyclic strains as functions of cycle number for fatigue tests conducted at 1200C in air and in steam are 1.4EI-0150 MPa, 0.1 Hz 150 MPa, 1.0 Hz 0150 MPa, 10 Hz presented in Figs. 7a and b, respectively. It is seen that rat- 170MPa01h·170MPa1.0H170MPa,10H cheting takes place in all tests conducted at 1200C. In lab oratory air the rate of strain accumulation increases with increasing fatigue stress level. On the other hand, the loading frequency does not appear to have a strong effect on strain accumulation rate. For the fatigue stress of T= 1200C. Steam E+001.E+011.E+021.E+031.E+04 1E+061.E+07 a T=1200c.A ●100MPa,1.0H Fig. 6. Normalized modulus vs fatigue cycles at 1200C(a) in laborator 馨170MPa,1.0Hz ir and(b) in steam Data at 1.0 Hz from Ruggles-Wrenn et al. [44] A170 MPa, 0.1 Hz normalized modulus dropped 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. In air, the loading frequency appears to have little effect on the modulus change with cycles. Modulus loss of 18% observed in the 170 MPa test conducted at 0. 1 Hz is not significantly different from that at 1.0 Hz. Changes in normalized modulus as well as the 1E+001.E+011.E+021.E+031.E+041.E+051.E+06 influence of loading rate on modulus evolution become Cycle(N) more pronounced in steam(Fig. 6b). While in air the b reduction in normalized modulus was limited to 18% T= 1200 C Steam (170 MPa test at 0.1 Hz), in steam the normalized modulus 125MPa,0.1 loss reached 30%(170 MPa tests at 0. 1 and at 1.0 Hz. As 30日|→150MPa,01h 18%in the 75 MPa test, 20% in the 100 MPa test, s ge 2.5 F -170 MPa,0.1 Hz in air, in steam modulus loss increases with increasing fati- gue stress level. At 0.1 Hz, normalized modulus loss wa 2.0 F1-o-125 MPa,1.0H the 125 and 150 MPa tests, and 30% in the 170 MPa test. i 15F1-0-170 MPa 1.o t/ -o-150 Mpa, 10 Hz Increase in frequency by an order of magnitude had a 10E-170Mpa,10z noticeable effect on the modulus evolution with cycles. At 1.0 HZ. normalized modulus loss in run-out fatigue tests was 10% in the 100 MPa test and 16%o in the 125 MPa test In the 150 MPa test normalized modulus dropped by 17%o, 1E+001.E+011E+021E+031E+041.E+051.E+06 and in the 170 MPa test, by 30%0. Note that the normalized modulus loss of 30% was observed in the 170 MPa tests Fig. 7. Maximum strain vs fatigue cycles at 1200C: (a)in laboratory air conducted at both 0.1 and 1.0 Hz. in the case of the and(b)in steam environment Data at 1.0 Hz from Ruggles-Wrenn et al 170 MPa tests, which produced very short fatigue lifetimes, [44]
normalized modulus dropped 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. In air, the loading frequency appears to have little effect on the modulus change with cycles. Modulus loss of 18% observed in the 170 MPa test conducted at 0.1 Hz is not significantly different from that at 1.0 Hz. Changes in normalized modulus as well as the influence of loading rate on modulus evolution become more pronounced in steam (Fig. 6b). While in air the reduction in normalized modulus was limited to 18% (170 MPa test at 0.1 Hz), in steam the normalized modulus loss reached 30% (170 MPa tests at 0.1 and at 1.0 Hz). As in air, in steam modulus loss increases with increasing fatigue stress level. At 0.1 Hz, normalized modulus loss was 18% in the 75 MPa test, 20% in the 100 MPa test, 21% in the 125 and 150 MPa tests, and 30% in the 170 MPa test. Increase in frequency by an order of magnitude had a noticeable effect on the modulus evolution with cycles. At 1.0 Hz, normalized modulus loss in run-out fatigue tests was 10% in the 100 MPa test and 16% in the 125 MPa test. In the 150 MPa test normalized modulus dropped by 17%, and in the 170 MPa test, by 30%. Note that the normalized modulus loss of 30% was observed in the 170 MPa tests conducted at both 0.1 and 1.0 Hz. In the case of the 170 MPa tests, which produced very short fatigue lifetimes, modulus loss is likely due to the early failure of fiber bundles, while the modulus loss observed in other tests conducted at 0.1 and 1.0 Hz is likely due to the progressive cracking of the matrix. Increase in frequency by another order of magnitude to 10 Hz has an even more appreciable effect on the modulus change. At 10 Hz, the reduction in normalized modulus was only 5% in the 150 MPa test, which achieved a run-out, and 8% in the 170 MPa test. Continuous decrease in modulus observed both in air and in steam suggests progressive damage with continued cycling. Because the fatigue damage is still evolving at 105 cycles (106 cycles at 10 Hz), the 105 (106 at 10 Hz) fatigue limit does not meet the criteria of a true endurance fatigue limit proposed by Sorensen et al. [48] and may not be a true endurance fatigue limit. Maximum cyclic strains as functions of cycle number for fatigue tests conducted at 1200 C in air and in steam are presented in Figs. 7a and b, respectively. It is seen that ratcheting takes place in all tests conducted at 1200 C. In laboratory air the rate of strain accumulation increases with increasing fatigue stress level. On the other hand, the loading frequency does not appear to have a strong effect on strain accumulation rate. For the fatigue stress of 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 1.E+07 Cycles (N) sul udo Mdezil a mr oN 100 MPa, 1.0 Hz 125 MPa, 1.0 Hz 150 MPa, 1.0 Hz 170 MPa, 1.0 Hz 170 MPa, 0.1 Hz T = 1200 ˚C, Air 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 1.E+07 Cycles (N) sul udo Mdezil a mr oN 75 MPa, 0.1 Hz 100 MPa, 0.1 Hz 100 MPa, 1.0 Hz 125 MPa, 0.1 Hz 125 MPa, 1.0 Hz 150 MPa, 0.1 Hz 150 MPa, 1.0 Hz 150 MPa, 10 Hz 170 MPa, 0.1 Hz 170 MPa, 1.0 Hz 170 MPa, 10 Hz T = 1200 ˚C, Steam a b Fig. 6. Normalized modulus vs. fatigue cycles at 1200 C (a) in laboratory air and (b) in steam. Data at 1.0 Hz from Ruggles-Wrenn et al. [44]. 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 Cycle (N) 100 MPa, 1.0 Hz 150 MPa, 1.0 Hz 170 MPa, 1.0 Hz 170 MPa, 0.1 Hz T = 1200 ˚C, Air 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 Cycle (N) Strain % Strain % 75 MPa, 0.1 Hz 100 MPa, 0.1 Hz 125 MPa, 0.1 Hz 150 MPa, 0.1 Hz 170 MPa, 0.1 Hz 100 MPa, 1.0 Hz 125 MPa, 1.0 Hz 150 MPa, 1.0 Hz 170 MPa, 1.0 Hz 150 Mpa, 10 Hz 170 Mpa, 10 Hz T = 1200 ˚C, Steam a b Fig. 7. Maximum strain vs. fatigue cycles at 1200 C: (a) in laboratory air and (b) in steam environment. Data at 1.0 Hz from Ruggles-Wrenn et al. [44]. M.B. Ruggles-Wrenn et al. / International Journal of Fatigue 30 (2008) 502–516 507
508 10-mm 10 mm 10m Fig 8. Fracture surfaces(optical micrographs) of the N720/A specimens tested in cyclic fatigue at 1200C in steam at 0.1 Hz:(a)omax= 150 MPa, fr=0.21 h, Ef=0.67%;(b)omax= 125 MPa, Ir=5. 14 h, Er=1. 15%; and (c)omax= 100 MPa, tr= 486h, Ef= 1.80%. Size of the damage zone, the amount of fiber pullout and cyclic fatigue lifetime increase with decreasing fatigue stress level. 170 MPa, the evolution of strain with cycles observed at fiber pullout(see Fig 8), which accounts for larger accu- 0.1 Hz is similar to that at 1.0 Hz. Note that all tests con- mulated strains. In steam the evolution of maximum strain ducted in air achieved fatigue run-out. As seen in Figs. with cycles is strongly influenced by the loading frequency. 7a and b, strains accumulated in steam are considerably For a given fatigue stress level, the rate of strain accumula lower than those accumulated in air at the same fatigue tion increases with decreasing frequency. In the case of the stress and loading frequency. Generally, lower strain accu- 150 MPa tests, specimen cycled at 0. 1 Hz accumulated mulation with cycling indicates that less damage has 0.59% strain during the first 50 cycles, while those tested occurred, and that it is mostly limited to some additional at 1.0 and 10 Hz accumulated 0. 32% and 0. 25% strain matrix cracking. However, lower accumulated strains respectively. However, higher rate of strain accumulation observed in steam invariably correspond to shorter fatigue does not necessarily translate into higher failure strain. lives. In this case lower accumulated strains are more likely The decrease in loading frequency causes a dramatic Results in Table I reveal that in steam at a given loading strain accumulation. Once again considering the time for due to early bundle failures leading to specimen failure. decrease in fatigue life, hence allowing much less time for frequency, both the fatigue life and the accumulated strain tests, it is seen that the fatigue life of 11, 782 cycles increase with decreasing fatigue stress. This trend is partic-(3.27 h) produced at 1.0 Hz allowed for accumulated fail- ularly pronounced at the loading frequency of 0. 1 Hz. The ure strain of 1. 12%0, while at 0. 1 Hz a much shorter life 170 MPa test failed after 12 cycles accumulating only of 202 cycles (0.56 h)allowed for accumulated strain of 0.53% strain, while the 75 MPa test survived 56, 093 cycles only 0.67% and accumulated a much larger strain of 3.35%. Specimens Retained strength and stiffness of the specimens, which with longer cyclic lives also exhibited larger amounts of achieved a run-out, are summarized in Table 2, where Table 2 Retained properties of the N720/A specimens subjected to prior fatigue in laboratory air and in steam environment at 1200oC Fatigue Retained strength Strength retention Retained modulus Modulus retention Strain at failur MPa) environment (MPa) (%0) (GPa) %) Prior fatigue at 0. 1 Hz ≥100 0.3 ≥100 Prior fatigue at 1.0Hz ≥100 ≥100 ≥100 0.51 100 l74 Prior fatigue at 10 Hz 0.48 Retained properties measured at 1200C in laboratory air Data from Ruggles-Wrenn et al. [44]
170 MPa, the evolution of strain with cycles observed at 0.1 Hz is similar to that at 1.0 Hz. Note that all tests conducted in air achieved fatigue run-out. As seen in Figs. 7a and b, strains accumulated in steam are considerably lower than those accumulated in air at the same fatigue stress and loading frequency. Generally, lower strain accumulation with cycling indicates that less damage has occurred, and that it is mostly limited to some additional matrix cracking. However, lower accumulated strains observed in steam invariably correspond to shorter fatigue lives. In this case lower accumulated strains are more likely due to early bundle failures leading to specimen failure. Results in Table 1 reveal that in steam at a given loading frequency, both the fatigue life and the accumulated strain increase with decreasing fatigue stress. This trend is particularly pronounced at the loading frequency of 0.1 Hz. The 170 MPa test failed after 12 cycles accumulating only 0.53% strain, while the 75 MPa test survived 56,093 cycles and accumulated a much larger strain of 3.35%. Specimens with longer cyclic lives also exhibited larger amounts of fiber pullout (see Fig. 8), which accounts for larger accumulated strains. In steam the evolution of maximum strain with cycles is strongly influenced by the loading frequency. For a given fatigue stress level, the rate of strain accumulation increases with decreasing frequency. In the case of the 150 MPa tests, specimen cycled at 0.1 Hz accumulated 0.59% strain during the first 50 cycles, while those tested at 1.0 and 10 Hz accumulated 0.32% and 0.25% strain, respectively. However, higher rate of strain accumulation does not necessarily translate into higher failure strain. The decrease in loading frequency causes a dramatic decrease in fatigue life, hence allowing much less time for strain accumulation. Once again considering the 150 MPa tests, it is seen that the fatigue life of 11,782 cycles (3.27 h) produced at 1.0 Hz allowed for accumulated failure strain of 1.12%, while at 0.1 Hz a much shorter life of 202 cycles (0.56 h) allowed for accumulated strain of only 0.67%. Retained strength and stiffness of the specimens, which achieved a run-out, are summarized in Table 2, where Fig. 8. Fracture surfaces (optical micrographs) of the N720/A specimens tested in cyclic fatigue at 1200 C in steam at 0.1 Hz: (a) rmax = 150 MPa, tf = 0.21 h, ef = 0.67%; (b) rmax = 125 MPa, tf = 5.14 h, ef = 1.15%; and (c) rmax = 100 MPa, tf = 48.6 h, ef = 1.80%. Size of the damage zone, the amount of fiber pullout and cyclic fatigue lifetime increase with decreasing fatigue stress level. 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) Fatigue environment Retained strength (MPa) Strength retention (%) Retained modulus (GPa) Modulus retention (%) Strain at failure (%) Prior fatigue at 0.1 Hz 170 Air 194 P100 51.7 82 0.38 170 Air 196 P100 55.6 80 0.39 Prior fatigue at 1.0 Hza 100 Air 194 P100 56.6 74 0.44 125 Air 199 P100 54.9 73 0.44 150 Air 199 P100 43.4 72 0.53 170 Air 192 P100 40.7 67 0.51 100 Steam 174 90 47.6 84 0.40 125 Steam 168 88 52.0 80 0.43 Prior fatigue at 10 Hz 150 Steam 184 96 55.4 93 0.48 Retained properties measured at 1200 C in laboratory air. a Data from Ruggles-Wrenn et al. [44]. 508 M.B. Ruggles-Wrenn et al. / International Journal of Fatigue 30 (2008) 502–516
M B. Ruggles-Prenn et al International Journal of fatigue 30(2008)502-516 results from prior work [44]are included for comparison. It is seen that all specimens tested in air exhibited increase in d=AK?=A(Yava) tensile strength, irrespective of the fatigue stress level or the where a is the crack size, t is time, A and n are material con- loa ding frequency. However, considerable stiffness loss (18-33%)was observed. The N720/A composite relies on the applied stress intensity factor, Y is a geometric factor and a is the applied stress. If the cyclic and static fatigue g,as evidenced by a decrease in stiffness, and induces by subcritical(slow)crack growth, then the cyclic fatigue the weakening of the fiber-matrix interfaces. By causing lifetime can be predicted from the static fatigue data by additional matrix cracking, prior fatigue in air serves using a linear elastic crack growth model [59]. Evans and maintain the matrix porosity at the minimum level needed Fuller[59] proposed that the ratio of the time to failure un for crack deflection. As a result, prior fatigue in air effec- der cyclic fatigue, taf, to the time to failure under static fa- ively improves the damage tolerance and long-term dura- tigue, Ist, under the same maximum applied stress omax can bility of the composite as demonstrated by an improved be obtained as tensile strength of the pre-fatigued specimens. Conversely, prior fatigue in steam caused reduction in both strength (1) and stifness. Strength loss in steam was limited to 12% and stifness loss, to 20%. Strength and stiffness degrada where the applied stress d(n)=omaxf(n) is a periodic func tion increases with decreasing frequency of prior fatigue. tion of time, t is the period, and 0<fo< I for tensile As the frequency of the 170 MPa tests decreased from o loading. For static loading, the applied stress o(0)=omax to 1.0 Hz, strength loss increased from 4% to 12% and stiff- ness loss, from 7% to 20%. The discrepancy between the For cyclic loading with the ratio R(minimum to maximum retained modulus of a run-out specimen and the decrease stress)and a triangular waveform the periodic function f() hysteresis modulus observed during fatigue testing most is expressed as likely stems from different methods used to determine the R+ R f(t)= or0≤t≤ retained and hysteresis moduli Recent work [44]revealed the degrading effect of steam 1-(1-R)(￥-1)for≤t≤t on fatigue as well as on creep performance of the N720/A In the case of a trapezoidal waveform, where ih is the hold composite at 1200C. Mehrman et al. [45] showed that the time at maximum stress, fL is the loading/unloading time, superposition of a hold time at maximum stress onto a fati- and t=(h IL) is the period, the function f(n) is given by gue cycle drastically reduced cyclic lifetime of N720/A at 1200C in steam. The current study revealed a marked R+(1-R) or0≤t≤ influence of the loading frequency on fatigue life in steam. f()=I for≤t≤h+号(4) These findings suggest that a unique time-dependent failure 1-2(t-4-)forh+是≤t≤ mechanism may be operating under both cyclic and static loading at 1200C in steam The crack growth exponent n is readily obtained from the It is recognized that stress corrosion of the N720 fibers static fatigue data at 1200C in steam [44]. Noting that sta may be the mechanism behind reduced creep resistance of tic fatigue lifetime in steam is related to the applied stress N720/A composite at 1200 C in steam. Earlier studies as: [49-54] suggested that static fatigue (i.e. delayed fracture under a sustained constant load)of silica-based glasses sf=4.5×1026-113 (5) was a chemical process, where subcritical (slow) crack ( the applied stress is in megapascals and time is in seconds) growth resulted from and was controlled by a stress- the crack growth exponent n=11. 3 is determined. Using enhanced chemical reaction between glass and water in R=0.05 and Eqs.(2)-(4), the ratio of cyclic to static life- the environment. Michalske and Bunker [55-57]examined times, r, can be calculated for the triangular waveform as the role of mechanical strain in accelerating chemical reac- well as for the trapezoidal waveforms with the loading/ tions between the Si-o bonds at the crack tip and environ- unloading time tL= l s and hold times fh of 10 and 100 s mental molecules and found that the highly strained Si-o The predicted cyclic fatigue lifetimes are plotted in Fig. 9 onds reacted with water at least 8 orders of magnitude together with the experimental results obtained in steam faster than the unstrained bonds. Michalske and Bunker The results of fatigue tests from prior work [44, 45] are also [57] proposed a quantitative chemical-kinetics-based model shown in Fig. 9 for comparison. It is seen that the predicted to predict the rate of crack growth in silica glass in humid cyclic lifetimes are in excellent agreement with the experi- condition as a function of the applied stress. This model mental results obtained in fatigue tests with the trapezoidal describes a fracture rate law in which the crack growth rate waveform and with the triangular waveform at 0. I Hz, increases exponentially with the applied stress intensity. indicating that there is no apparent cyclic effect on the For many glass and ceramic materials subcritical crack crack growth and that the slow crack growth is growth can be described by a power law [5 the dominant failure mechanism. Conversely, predictions
results from prior work [44] are included for comparison. It is seen that all specimens tested in air exhibited increase in tensile strength, irrespective of the fatigue stress level or the loading frequency. However, considerable stiffness loss (18–33%) was observed. The N720/A composite relies on the porous matrix for crack deflection and damage tolerance. Fatigue cycling promotes progressive matrix cracking, as evidenced by a decrease in stiffness, and induces the weakening of the fiber–matrix interfaces. By causing additional matrix cracking, prior fatigue in air serves to maintain the matrix porosity at the minimum level needed for crack deflection. As a result, prior fatigue in air effectively improves the damage tolerance and long-term durability of the composite as demonstrated by an improved tensile strength of the pre-fatigued specimens. Conversely, prior fatigue in steam caused reduction in both strength and stiffness. Strength loss in steam was limited to 12% and stiffness loss, to 20%. Strength and stiffness degradation increases with decreasing frequency of prior fatigue. As the frequency of the 170 MPa tests decreased from 10 to 1.0 Hz, strength loss increased from 4% to 12% and stiff- ness loss, from 7% to 20%. 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. Recent work [44] revealed the degrading effect of steam on fatigue as well as on creep performance of the N720/A composite at 1200 C. Mehrman et al. [45] showed that the superposition of a hold time at maximum stress onto a fatigue cycle drastically reduced cyclic lifetime of N720/A at 1200 C in steam. The current study revealed a marked influence of the loading frequency on fatigue life in steam. These findings suggest that a unique time-dependent failure mechanism may be operating under both cyclic and static loading at 1200 C in steam. It is recognized that stress corrosion of the N720 fibers may be the mechanism behind reduced creep resistance of N720/A composite at 1200 C in steam. Earlier studies [49–54] suggested that static fatigue (i.e. delayed fracture under a sustained constant load) of silica-based glasses was a chemical process, where subcritical (slow) crack growth resulted from and was controlled by a stressenhanced chemical reaction between glass and water in the environment. Michalske and Bunker [55–57] examined the role of mechanical strain in accelerating chemical reactions between the Si–O bonds at the crack tip and environmental molecules and found that the highly strained Si–O bonds reacted with water at least 8 orders of magnitude faster than the unstrained bonds. Michalske and Bunker [57] proposed a quantitative chemical-kinetics-based model to predict the rate of crack growth in silica glass in humid condition as a function of the applied stress. This model describes a fracture rate law in which the crack growth rate increases exponentially with the applied stress intensity. For many glass and ceramic materials subcritical crack growth can be described by a power law [58]: da dt ¼ AKn I ¼ A Y r ffiffiffi a p n ð1Þ where a is the crack size, t is time, A and n are material constants dependent on temperature and environment, KI is the applied stress intensity factor, Y is a geometric factor and r is the applied stress. If the cyclic and static fatigue failure mechanisms are indeed identical and dominated by subcritical (slow) crack growth, then the cyclic fatigue lifetime can be predicted from the static fatigue data by using a linear elastic crack growth model [59]. Evans and Fuller [59] proposed that the ratio of the time to failure under cyclic fatigue, tcf, to the time to failure under static fatigue, tsf, under the same maximum applied stress rmax can be obtained as: r ¼ tcf tsf ¼ s Z s 0 rðtÞ rmax n dt 1 ð2Þ where the applied stress r(t) = rmaxf(t) is a periodic function of time, s is the period, and 0 6 f(t) 6 1 for tensile loading. For static loading, the applied stress r(t) = rmax. For cyclic loading with the ratio R (minimum to maximum stress) and a triangular waveform the periodic function f(t) is expressed as: f ðtÞ ¼ R þ ð1 RÞ 2t s for 0 6 t 6 s 2 1 ð1 RÞ 2t s 1 for s 2 6 t 6 s ( ð3Þ In the case of a trapezoidal waveform, where th is the hold time at maximum stress, tL is the loading/unloading time, and s = (th + tL) is the period, the function f(t) is given by: f ðtÞ ¼ R þ ð1 RÞ 2t tL for 0 6 t 6 tL 2 1 for tL 2 6 t 6 th þ tL 2 1 2ð1RÞ tL t th tL 2 for th þ s 2 6 t 6 s 8 >>>: ð4Þ The crack growth exponent n is readily obtained from the static fatigue data at 1200 C in steam [44]. Noting that static fatigue lifetime in steam is related to the applied stress as: tsf ¼ 4:5 1026r11:3 max ð5Þ (the applied stress is in megapascals and time is in seconds), the crack growth exponent n = 11.3 is determined. Using R = 0.05 and Eqs. (2)–(4), the ratio of cyclic to static lifetimes, r, can be calculated for the triangular waveform as well as for the trapezoidal waveforms with the loading/ unloading time tL = 1 s and hold times th of 10 and 100 s. The predicted cyclic fatigue lifetimes are plotted in Fig. 9 together with the experimental results obtained in steam. The results of fatigue tests from prior work [44,45] are also shown in Fig. 9 for comparison. It is seen that the predicted cyclic lifetimes are in excellent agreement with the experimental results obtained in fatigue tests with the trapezoidal waveform and with the triangular waveform at 0.1 Hz, indicating that there is no apparent cyclic effect on the crack growth rate and that the slow crack growth is the dominant failure mechanism. Conversely, predictions M.B. Ruggles-Wrenn et al. / International Journal of Fatigue 30 (2008) 502–516 509
es-Wrenn et al. International Journal of fatigue 8)502-510 a1000 undergo additional matrix sintering and subsequent loss of T=1200°c, Steam matrix porosity. Under static loading, the densification of the matrix and the stress corrosion of the fibers work together to accelerate failure and to reduce lifetime of the composite. In contrast, cyclic loading counteracts matrix densification by causing progressive matrix cracking and weakening of the fiber-matrix interface [44, 45]. Apparently the beneficial effects of the mechanical cycling on damage ue 1.0 Hz, Ruggles-Wrenn 2006 tolerance of the N720/A composite compensate for the negative effects of the stress corrosion on the fibers, result ing in improved durability and longer lifetimes under cyclic loadin 1E+011.E+021E÷031.E+041.E+051.E+061.E+07 Time to Failure(s) 3.3. Composite microstructure b T=1200°c, Steam Fracture surfaces of several specimens tested at 1200C in steam were examined to gain a better understanding of the effects of the cyclic loading frequency on the damage and failure mechanisms in this porous matrix CMC. The influence of the loading frequency on the fracture surface phy is illustrated in Figs. a Cyclic Fatigue with 10 s Hold, Ehrman 2006 surfaces obtained in cyclic fatigue tests conducted with the maximum stress of 170 MPa at 0.1, 1.0, and 10 Hz, respec A Static Fatigue, Ruggles-wrenn 2006 tively. It is seen that the specimen tested at 0. 1 Hz, which icted from Static Data survived only 0.03 h(12 cycles) produced a short damage zone(2 mm in length). The fracture surface is dominated 1.5+01 1.5+02 1.E+03 1. E+04 1.5+05 1.+06 by coordinated failure of the fiber bundles with small iso- Time to Failure(s) lated areas of fiber pullout. A longer damage zone and a Fig 9. Comparison between cyclic fatigue lifetimes predicted from static somewhat brusher failure surface are obtained at 1.0 Hz tigue data and experimental results obtained for N720/A ceramic A dramatically different fracture surface topography is pro- omposite at 1200C in steam: (a) in cyclic tests with triangular wave form duced by the specimen tested at 10 Hz, which also exhibited fatigue tests with hold time from Ehrman et al f t al. [44), resuts ta the longest lifetime surviving 0.32 h(11,387 cycles).At seen 14 mm. The fibers in the 0o tows in each cloth layer exhi bit uncorrelated random failure producing a brushy fr significantly underestimate the cyclic lifetimes measured in ture surface. The fiber pullout is extensive and the tests conducted with the triangular waveform at higher fre- variation in pull-out length is considerable. It is recognized quencies. At l and 10 Hz only the cyclic lifetimes measured that the increase in the spatial correlation in the fiber fai for omax= 170 MPa are within the range of the predicted ure locations is among the main manifestations of the life, while the cyclic lifetimes produced for the applied matrix densification [60,61]. The near-planar fracture sur- stress levels <170 MPa are at least an order of magnitude face obtained in the 0. 1 Hz test indicates the loss of matrix larger than those predicted from static data. Apparently, porosity and subsequent matrix densification. As a result for applied stresses <170 MPa, cyclic loading with a short the composite exhibits decreased damage tolerance, brittle duration of maximum stress has a beneficial effect on life- fracture behavior and a short lifetime. Conversely, a time compared to static loading fibrous failure surface obtained at 10 Hz. demonstrates that To gain insight into the beneficial effect of mechanical the mechanical cycling has restored the matrix porosity to cling, recall that the N720/A composite derives its dam- the minimum level required to deflect the matrix cracks and age tolerance from a porous matrix. Therefore the stability to allow subsequent fiber pullout. The CMc exhibits of the matrix porosity against densification is critical to the improved damage tolerance and delayed failure, as evi- composites long-term durability. The loss of matrix poros- dences by a considerably longer cyclic lifetime obtained y would inhibit crack deflection, reduce damage tolerance at 10 Hz. These conclusions are confirmed by examining and accelerate failure. Recent studies [60, 61] demonstrated the SEM micrographs of the aforementioned fracture sur that for a composite consisting of NextelTM720 fibers in a faces shown in Fig. ll. The fracture surface produced at porous alumina matrix, a porosity reduction of 6% was 0. 1 Hz(Fig. lla)is dominated by planar regions of coordi- bserved after a 10-min exposure at 1200C, which was nated fiber failure, indicative of matrix densification due to caused by additional sintering of the matrix. It is likely that additional sintering. The fracture surface obtained at N720/A specimens subjected to loading at 1200 C in steam 1.0 Hz(Fig. 11b)still exhibits large areas of planar
significantly underestimate the cyclic lifetimes measured in tests conducted with the triangular waveform at higher frequencies. At 1 and 10 Hz only the cyclic lifetimes measured for rmax = 170 MPa are within the range of the predicted life, while the cyclic lifetimes produced for the applied stress levels <170 MPa are at least an order of magnitude larger than those predicted from static data. Apparently, for applied stresses <170 MPa, cyclic loading with a short duration of maximum stress has a beneficial effect on lifetime compared to static loading. To gain insight into the beneficial effect of mechanical cycling, recall that the N720/A composite derives its damage tolerance from a porous matrix. Therefore the stability of the matrix porosity against densification is critical to the composite’s long-term durability. The loss of matrix porosity would inhibit crack deflection, reduce damage tolerance and accelerate failure. Recent studies [60,61] demonstrated that for a composite consisting of Nextel720 fibers in a porous alumina matrix, a porosity reduction of 6% was observed after a 10-min exposure at 1200 C, which was caused by additional sintering of the matrix. It is likely that N720/A specimens subjected to loading at 1200 C in steam undergo additional matrix sintering and subsequent loss of matrix porosity. Under static loading, the densification of the matrix and the stress corrosion of the fibers work together to accelerate failure and to reduce lifetime of the composite. In contrast, cyclic loading counteracts matrix densification by causing progressive matrix cracking and weakening of the fiber–matrix interface [44,45]. Apparently the beneficial effects of the mechanical cycling on damage tolerance of the N720/A composite compensate for the negative effects of the stress corrosion on the fibers, resulting in improved durability and longer lifetimes under cyclic loading. 3.3. Composite microstructure Fracture surfaces of several specimens tested at 1200 C in steam were examined to gain a better understanding of the effects of the cyclic loading frequency on the damage and failure mechanisms in this porous matrix CMC. The influence of the loading frequency on the fracture surface topography is illustrated in Figs. 10a–c displaying fracture surfaces obtained in cyclic fatigue tests conducted with the maximum stress of 170 MPa at 0.1, 1.0, and 10 Hz, respectively. It is seen that the specimen tested at 0.1 Hz, which survived only 0.03 h (12 cycles) produced a short damage zone (2 mm in length). The fracture surface is dominated by coordinated failure of the fiber bundles with small isolated areas of fiber pullout. A longer damage zone and a somewhat brushier failure surface are obtained at 1.0 Hz. A dramatically different fracture surface topography is produced by the specimen tested at 10 Hz, which also exhibited the longest lifetime surviving 0.32 h (11,387 cycles). At seen in Fig. 10c, the length of the damage zone has increased to 14 mm. The fibers in the 0 tows in each cloth layer exhibit uncorrelated random failure producing a brushy fracture surface. The fiber pullout is extensive and the variation in pull-out length is considerable. It is recognized that the increase in the spatial correlation in the fiber failure locations is among the main manifestations of the matrix densification [60,61]. The near-planar fracture surface obtained in the 0.1 Hz test indicates the loss of matrix porosity and subsequent matrix densification. As a result, the composite exhibits decreased damage tolerance, brittle fracture behavior and a short lifetime. Conversely, a fibrous failure surface obtained at 10 Hz, demonstrates that the mechanical cycling has restored the matrix porosity to the minimum level required to deflect the matrix cracks and to allow subsequent fiber pullout. The CMC exhibits improved damage tolerance and delayed failure, as evidences by a considerably longer cyclic lifetime obtained at 10 Hz. These conclusions are confirmed by examining the SEM micrographs of the aforementioned fracture surfaces shown in Fig. 11. The fracture surface produced at 0.1 Hz (Fig. 11a) is dominated by planar regions of coordinated fiber failure, indicative of matrix densification due to additional sintering. The fracture surface obtained at 1.0 Hz (Fig. 11b) still exhibits large areas of planar 10 100 1000 1.E+01 1.E+02 1.E+03 1.E+04 1.E+05 1.E+06 1.E+07 Time to Failure (s) ) aP M( ssert S mu mi xa M Cyclic Fatigue 0.1 Hz Cyclic Fatigue 1.0 Hz, Ruggles-Wrenn 2006 Cyclic Fatigue 10 Hz Static Fatigue, Ruggles-Wrenn 2006 Cyclic Fatigue - Predicted from Static Data T = 1200 °C, Steam 10 100 1000 1.E+01 1.E+02 1.E+03 1.E+04 1.E+05 1.E+06 Time to Failure (s) ) aP M( ssert S mu mi xa M Cyclic Fatigue with 10 s Hold, Mehrman 2006 Cyclic Fatigue with 100 s Hold, Mehrman 2006 Static Fatigue, Ruggles-Wrenn 2006 Cyclic Fatigue with Hold Time - Predicted from Static Data T = 1200 °C, Steam a b Fig. 9. Comparison between cyclic fatigue lifetimes predicted from static fatigue data and experimental results obtained for N720/A ceramic composite at 1200 C in steam: (a) in cyclic tests with triangular wave form and (b) in cyclic tests with hold time at maximum stress. Static fatigue data and cyclic fatigue data at 1.0 from Ruggles-Wrenn et al. [44], results of fatigue tests with hold time from Mehrman et al. [45]. 510 M.B. Ruggles-Wrenn et al. / International Journal of Fatigue 30 (2008) 502–516
M B. Ruggles-Wrenn et al. International Journal of fatigue 30(2008)502-516 10 mm 10mm -10 mm Fig. 10. Fracture surfaces(optical micrographs)of the N720/A specimens tested in cyclic fatigue with maximum stress of 170 MPa at 1200C in steam:(a) 0 I Hz, Ir=0.03h, (b)1.0 Hz, tr=0.06 h, and (c)10 Hz, tr=0.32 h. Size of the damage zone and cyclic fatigue lifetime increase with increasing frequenc 物子 Fig. 11. Fracture surfaces(SEM micrographs)of the N720/A specimens tested in cyclic fatigue with the maximum stress of 170 MPa at 1200C in steam (a)0.1 Hz; (b)1.0 Hz;(c)10 Hz. The spatial correlation in the fiber failure locations within an individual tow increases with decreasing loading frequency fracture; however, areas of uncoordinated brushy failure maintaining matrix porosity at the level sufficient to enable are also visible. Conversely the fracture surface produced crack-deflecting behavior and fiber pullout. fiber fracture where individual fibers are clearly discern- s. Micrographs in Figs. 10 and 1l reveal that the fracture at 10 Hz(Fig. llc)is dominated by areas of uncorrelated rface appearance may be correlated with time to failure able a brushy fracture surface with extensive fiber pullout Predominantly planar fracture surface(for example see demonstrates that mechanical cycling at 10 Hz has success- Fig. 1la), indicative of a fast failure process where crack fully counteracted the effects of matrix densification by fronts propagate rapidly through both matrix and fibers
fracture; however, areas of uncoordinated brushy failure are also visible. Conversely the fracture surface produced at 10 Hz (Fig. 11c) is dominated by areas of uncorrelated fiber fracture where individual fibers are clearly discernable. A brushy fracture surface with extensive fiber pullout demonstrates that mechanical cycling at 10 Hz has successfully counteracted the effects of matrix densification by maintaining matrix porosity at the level sufficient to enable crack-deflecting behavior and fiber pullout. Micrographs in Figs. 10 and 11 reveal that the fracture surface appearance may be correlated with time to failure. Predominantly planar fracture surface (for example see Fig. 11a), indicative of a fast failure process where crack fronts propagate rapidly through both matrix and fibers, Fig. 10. Fracture surfaces (optical micrographs) of the N720/A specimens tested in cyclic fatigue with maximum stress of 170 MPa at 1200 C in steam: (a) 0.1 Hz, tf = 0.03 h, (b) 1.0 Hz, tf = 0.06 h, and (c) 10 Hz, tf = 0.32 h. Size of the damage zone and cyclic fatigue lifetime increase with increasing frequency. Fig. 11. Fracture surfaces (SEM micrographs) of the N720/A specimens tested in cyclic fatigue with the maximum stress of 170 MPa at 1200 C in steam: (a) 0.1 Hz; (b) 1.0 Hz; (c) 10 Hz. The spatial correlation in the fiber failure locations within an individual tow increases with decreasing loading frequency. M.B. Ruggles-Wrenn et al. / International Journal of Fatigue 30 (2008) 502–516 511