COMPOSITES SCIENCE AND TECHNOLOGY ELSEVIER Composites Science and Technology 61(2001)1331-1338 www.elsevier.com/locate/compsci Notch sensitivity of fatigue life in a Sylramic M/Sic composite at elevated temperature J.C. McNulty, M.Y. He, F W. Zok Materials Department, University of California, Santa Barbara, CA 93106, US.A Received 16 March 2000: received in revised form 20 February 2001; accepted 7 March 2001 The effects of holes and notches on the fatigue life of an advanced SylramicTM/SiC composite at 815.C have been examined.At this temperature, fracture occurs by an oxidative embrittlement mechanism, common to most Sic-based composites In unnotched specimens, embrittlement is manifested at stresses above the matrix cracking limit, omc, leading to fracture following relatively short exposure times(100 h). As a consequence, a fatigue threshold is obtained at a stress, oth a Ome. This threshold is due to the absence of an easy path for oxygen ingress when matrix cracks are not present. In center-hole specimens, an analogous threshold is btained at a stress, oth S Omc/ke, where ke is the elastic stress concentration factor(=2.5). That is, once the cracking limit is exceeded at the hole edge, embrittlement and fracture ensue. The threshold stress for center-notch specimens with stress con- centration ke 7 is numerically similar to that of the center-hole specimens with ke 2.5, indicating some tolerance to local stress levels above the global matrix cracking limit in sharply notched geometries. Non-linear finite-element calculations of the stresses in the center-hole and center-notch specimens are used to infer the local conditions associated with the threshold a key result is that the damage tolerance and notch insensitivity normally associated with inelastic straining cannot be exploited at temperatures at which the embrittlement mechanism operates. The implication is that composite structures with holes and notches must be designed extremely conservatively to ensure long lifetimes(>100 h).@ 2001 Elsevier Science Ltd. All rights reserved Keywords: Notch sensitivity; Fatigue; SiC composite; Embrittlement 1. Introduction based CFCCs exhibit inelastic deformation associated with matrix cracking and interface debonding, provid Continuous-fiber ceramic composites(CFCCs)based ing damage tolerance and strength retention in the pre- on SiC constituents have been under development for sence of holes and notches [3-5] the past two decades, mainly for use in advanced gas The most significant problem hindering SiC--CFCCs is turbine engines. The motivation for this activity is the oxidation embrittlement. The embrittlement most com- desire to increase operating temperatures and hence monly occurs by oxygen ingress through matrix cracks improve engine efficiency and performance. Additional followed by reaction of the oxygen with the fibers and the benefits associated with the elimination of film cooling fiber coatings [6-22]. The problem is particularly severe of combustor liners and the ensuing reductions in No at intermediate temperatures(500-900oC)[6, 15, 23-34] emissions have also been identified [1, 2 It is expected to be exacerbated by cyclic loading, since The selection of Sic as the main constituent in high- under such conditions the reaction gases contained performance CFCCs is based on a number of attractive within the crack are expelled during unloading and the physical and mechanical characteristics. Notably, their oxidizing atmosphere drawn into the composite through ow diffusivity results in good creep resistance at high matrix cracks during reloading. Cyclic loading may also temperatures. Furthermore, their high thermal con- accelerate fiber fracture in regions where the fibers have ductivity and low thermal expansion coefficient lead to bonded to the surrounding matrix as a consequence of high thermal shock resistance. In addition, the Sic- the oxidation process, thereby limiting the extent of further sliding along the fiber/matrix interface Corresponding author. Tel: +1-805-893-8699: fax: 1-805-893 The present paper examines the embrittlement phe- nomenon in an advanced Sic-CFCC which is currently E-imail address: zok(@engineering. ucsb. edu(F w. Zok) of interest to both the aerospace and the land-base 0266-3538/01/ S.see front matter C 2001 Elsevier Science Ltd. All rights reserved. PII:S0266-3538(01)00032-X
Notch sensitivity of fatigue life in a Sylramic TM/SiC composite at elevated temperature J.C. McNulty, M.Y. He, F.W. Zok* Materials Department, University of California, Santa Barbara, CA 93106, USA Received 16 March 2000; received in revised form 20 February 2001; accepted 7 March 2001 Abstract The effects of holes and notches on the fatigue life of an advanced SylramicTM/SiC composite at 815C have been examined. At this temperature, fracture occurs by an oxidative embrittlement mechanism, common to most SiC-based composites. In unnotched specimens, embrittlement is manifested at stresses above the matrix cracking limit, mc, leading to fracture following relatively short exposure times (100 h). As a consequence, a fatigue threshold is obtained at a stress, th mc. This threshold is due to the absence of an easy path for oxygen ingress when matrix cracks are not present. In center-hole specimens, an analogous threshold is obtained, at a stress, th mc=ke, where ke is the elastic stress concentration factor (= 2.5). That is, once the cracking limit is exceeded at the hole edge, embrittlement and fracture ensue. The threshold stress for center-notch specimens with stress concentration ke ¼ 7 is numerically similar to that of the center-hole specimens with ke ¼ 2:5, indicating some tolerance to local stress levels above the global matrix cracking limit in sharply notched geometries. Non-linear finite-element calculations of the stresses in the center-hole and center-notch specimens are used to infer the local conditions associated with the threshold. A key result is that the damage tolerance and notch insensitivity normally associated with inelastic straining cannot be exploited at temperatures at which the embrittlement mechanism operates. The implication is that composite structures with holes and notches must be designed extremely conservatively to ensure long lifetimes (> 100 h). # 2001 Elsevier Science Ltd. All rights reserved. Keywords: Notch sensitivity; Fatigue; SiC composite; Embrittlement 1. Introduction Continuous-fiber ceramic composites (CFCCs) based on SiC constituents have been under development for the past two decades, mainly for use in advanced gas turbine engines. The motivation for this activity is the desire to increase operating temperatures and hence improve engine efficiency and performance. Additional benefits associated with the elimination of film cooling of combustor liners and the ensuing reductions in NOx emissions have also been identified [1,2]. The selection of SiC as the main constituent in highperformance CFCCs is based on a number of attractive physical and mechanical characteristics. Notably, their low diffusivity results in good creep resistance at high temperatures. Furthermore, their high thermal conductivity and low thermal expansion coefficient lead to high thermal shock resistance. In addition, the SiCbased CFCCs exhibit inelastic deformation associated with matrix cracking and interface debonding, providing damage tolerance and strength retention in the presence of holes and notches [3–5]. The most significant problem hindering SiC–CFCCs is oxidation embrittlement. The embrittlement most commonly occurs by oxygen ingress through matrix cracks, followed by reaction of the oxygen with the fibers and the fiber coatings [6–22]. The problem is particularly severe at intermediate temperatures (500–900C) [6,15,23–34]. It is expected to be exacerbated by cyclic loading, since under such conditions the reaction gases contained within the crack are expelled during unloading and the oxidizing atmosphere drawn into the composite through matrix cracks during reloading. Cyclic loading may also accelerate fiber fracture in regions where the fibers have bonded to the surrounding matrix as a consequence of the oxidation process, thereby limiting the extent of further sliding along the fiber/matrix interface. The present paper examines the embrittlement phenomenon in an advanced SiC–CFCC which is currently of interest to both the aerospace and the land-base 0266-3538/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved. PII: S0266-3538(01)00032-X Composites Science and Technology 61 (2001) 1331–1338 www.elsevier.com/locate/compscitech * Corresponding author. Tel.: +1-805-893-8699; fax: +1-805-893- 8486. E-mail address: zok@engineering.ucsb.edu (F.W. Zok).
1332 J C. McNulty et al. /Composites Science and Technolog y 61(2001)1331-1338 power-generation industries. Experiments are performed to assess the fatigue performance at a moderately high temperature(815oC). This temperature falls in the regime in which embrittlement has been reported for other SiC based CFCCs, yet is also relevant to the service environ- ments in which CFCC components are expected to operate in turbine engine applications. The sensitivity of fatigue life to stress concentrations is probed through tests on specimens with circular holes and notches as well as straight(unnotched)specimens. The study of the effects of stress concentrations is motivated by the need, in some instances, to use through-holes for either attachment or cooling in engine components. In a broader context, it might also be used to infer the extent of degradation due to foreign impact damage, assuming at the simplest level ( that the damaged region is unable to sustain load and hence acts similarly to a hole or notch. Some additional insights into the local conditions required for embrittle- ment and fracture in the notched geometries are obtained through calculations of the stresses in the notched geo metries, by the use of a non-linear constitutive law appropriate to CFCS 2. Materials and experimental measurements All experiments were performed on a composite com prising SylramicTM SiC! fibers in a plain-weave archi tecture and a Sic matrix produced by a proprietary hybrid process involving chemical vapor infiltration(CVI) and reactive melt infiltration(MD). The microstructure of the (b) composite is shown in Fig. 1. The notable features include the presence of a uniform Bn coating on the fibers, an overcoat'of SiC around the fiber tows(produced by CVn) and a Sic-based two-phase matrix (produced by reactive MI). The matrix is essentially fully dense, with out the large pores that are invariably present in CFCCs made by CVI alone. The absence of pores in the matrix has two consequences. First, it increases the stress at the onset of matrix cracking relative to that of the cvi CFCCs containing pores. Secondly, it is expected to improve the thermal conductivity, especially in the dire tion transverse to the fibers The fiber volume fraction To assess the effects of stress concentrators on low cycle fatigue(LCF), three types of specimens were used: (i) standard dog-bone(unnotched) specimens; (ii) straight tensile specimens of width 2W=31.8 mm, with a center hole of diameter 2a=6.35 mm (a/w=0.2); and (ii) straight tensile specimens of width 2W=31. 8 mm, with a Fig. 1. Scanning electron micrographs of a cross-section through the omposite:(a) at low magnification, showing the longitudinal and center notch of length 2a=6.35 mm (a/W=0.2)and transverse fiber tows; (b)at higher magnifications, showing the melt notch root radius p A0.2 mm. The specimens were tow. The micrographs were taken in backscatter electron imaging infiltrated matrix surrounding a fiber tow, and (c) the interior of the prepared by electrodischarge machining. Four spe mens of each geometry were tested at a temperature of (BED mode, thereby revealing atomic number contrast. The coatings on the fibers are bN. The Sic within the tows was produced by chemical vapor infiltration process, prior to melt-infiltration of th Produced by Dow Corning. remainder of the matrix
power-generation industries. Experiments are performed to assess the fatigue performance at a moderately high temperature (815C). This temperature falls in the regime in which embrittlement has been reported for other SiCbased CFCCs, yet is also relevant to the service environments in which CFCC components are expected to operate in turbine engine applications. The sensitivity of fatigue life to stress concentrations is probed through tests on specimens with circular holes and notches as well as straight (unnotched) specimens. The study of the effects of stress concentrations is motivated by the need, in some instances, to use through-holes for either attachment or cooling in engine components. In a broader context, it might also be used to infer the extent of degradation due to foreign impact damage, assuming at the simplest level that the damaged region is unable to sustain load and hence acts similarly to a hole or notch. Some additional insights into the local conditions required for embrittlement and fracture in the notched geometries are obtained through calculations of the stresses in the notched geometries, by the use of a non-linear constitutive law appropriate to CFCCs. 2. Materials and experimental measurements All experiments were performed on a composite comprising SylramicTM SiC1 fibers in a plain-weave architecture and a SiC matrix produced by a proprietary hybrid process involving chemical vapor infiltration (CVI) and reactive melt infiltration (MI). The microstructure of the composite is shown in Fig. 1. The notable features include the presence of a uniform BN coating on the fibers, an ‘overcoat’ of SiC around the fiber tows (produced by CVI) and a SiC-based two-phase matrix (produced by reactive MI). The matrix is essentially fully dense, without the large pores that are invariably present in CFCCs made by CVI alone. The absence of pores in the matrix has two consequences. First, it increases the stress at the onset of matrix cracking relative to that of the CVI CFCCs containing pores. Secondly, it is expected to improve the thermal conductivity, especially in the direction transverse to the fibers. The fiber volume fraction is 35%. To assess the effects of stress concentrators on low cycle fatigue (LCF), three types of specimens were used: (i) standard dog-bone (unnotched) specimens; (ii) straight tensile specimens of width 2W=31.8 mm, with a center hole of diameter 2a=6.35 mm ð Þ a=W ¼ 0:2 ; and (iii) straight tensile specimens of width 2W=31.8 mm, with a center notch of length 2a=6.35 mm ð Þ a=W ¼ 0:2 and notch root radius 0:2 mm. The specimens were prepared by electrodischarge machining. Four specimens of each geometry were tested at a temperature of 1 Produced by Dow Corning. Fig. 1. Scanning electron micrographs of a cross-section through the composite: (a) at low magnification, showing the longitudinal and transverse fiber tows; (b) at higher magnifications, showing the meltinfiltrated matrix surrounding a fiber tow, and (c) the interior of the tow. The micrographs were taken in backscatter electron imaging (BEI) mode, thereby revealing atomic number contrast. The coatings on the fibers are BN. The SiC within the tows was produced by a chemical vapor infiltration process, prior to melt-infiltration of the remainder of the matrix. 1332 J.C. McNulty et al. / Composites Science andTechnology 61 (2001) 1331–1338
J C. McNulty et al. /Composites Science and Technolog y 61(2001)1331-1338 815C in ambient air. The specimens were heated using an at a stress of 150-165 MPa. There was no apparent effect induction heating system and a steel susceptor. Tempera- of temperature on this response, except that the fracture tures were measured using thermocouples at three loca- strain at 815C was slightly lower than that at 20C tions on the specimen surface: at the center and at points Optical microscopy of polished cross-sections through the located +20 mm from the center. The temperatures at fractured test specimens revealed periodic matrix cracks. these three locations were maintained within +5C of the as commonly observed in this class of CFCC. targeted temperature( 815oC)for the duration of each test The effects of holes on the net-section strength at Steel tabs were affixed to the specimen ends using a high room temperature are shown in Fig. 3. The strength temperature ceramic adhesive and the tabbed sections decreases with increasing hole size, to 70% of the then inserted into hydraulic wedge grips. Additionally, unnotched strength at 2a=6.3 mm. Similar reductions several tests were performed at room temperature on have been reported previously for SiC/SiC, SiC/MAS [5] straight (unnotched) specimens as well as center hole and all-oxide ceramic composites [35]. These effects can ecimens with hole diameters of either 2a=3. 18 or 6.35 be rationalized on the basis of the stress redistribution mm and with a/w=0.2. The latter tests were used to due to inelastic straining around the hole as well as the assess the extent of notch sensitivity under ambient size-scale dependence of the local conditions needed to conditions. These results formed the basis on which the precipitate fracture, as described in Section 4 notch sensitivity of lCF life was subseq uently assessed The LCF behavior of 815C is shown in Fig. 4. Several The elevated temperature tests included one mono- notable features emerge. (i) The monotonic tension tests tonic tension test on each specimen type at a stressing on the center-hole and center-notch geometries yield rate of do/dt= 100 MPa/min, typically leading to frac- strengths similar to one another, oc 160-165 MPa, Ire in <3 min. The remaining three specimens were subject to LCF tests with a stress ratio of R=0.o1. The ading spectrum comprised relatively rapid loading and unloading(each taking a I min), and a hold time of 2 h at the peak stress. This spectrum was chosen to simulate he loading conditions that a gas turbine engine compo- nent might encounter in service. The peak stress levels were selected on the basis of the monotonic tensile test results. In cases in which fracture did not occur. the test were terminated after I week( 160-170 h)of loading The room temperature tests were also performed at a stressing rate of 100 MPa/min redictions(d=0.75 mm) The monotonic tensile test results on the unnotched pecimens are plotted in Fig. 2. At both 20 and 815C, the tensile response is essentially bilinear, with the tran Hole Diameter, 2a(mm) sition in slope occurring at the onset of matrix cracking Fig 3. Effects of center holes on room-temperature tensile strength The solid line represents the prediction of the point stress fracture model for a characteristic distance of d=0.75 ● Center hole 口 Center Notch 250 =254 GPa Center hole °° 2 0.0 Strain (%) g. 2. Tensile notched specimens at both 20 815C. Note the strong similarity in the stress-strain response at the wo test temperatures. t difference being the strains at which Fig 4. Summary of lCf life at 815C. Arrows pointing to the right fracture occurs. The solid lines represent a bilinear fit to the data, used indicate runout; arrows pointing to the left indicate test results for calibrating the nonlinear constitutive law obtained from monotonic tension tests
815C in ambient air. The specimens were heated using an induction heating system and a steel susceptor. Temperatures were measured using thermocouples at three locations on the specimen surface: at the center and at points located 20 mm from the center. The temperatures at these three locations were maintained within 5C of the targeted temperature (815C) for the duration of each test. Steel tabs were affixed to the specimen ends using a high temperature ceramic adhesive and the tabbed sections then inserted into hydraulic wedge grips. Additionally, several tests were performed at room temperature on straight (unnotched) specimens as well as center hole specimens with hole diameters of either 2a=3.18 or 6.35 mm and with a=W ¼ 0:2. The latter tests were used to assess the extent of notch sensitivity under ambient conditions. These results formed the basis on which the notch sensitivity of LCF life was subsequently assessed. The elevated temperature tests included one monotonic tension test on each specimen type at a stressing rate of d=dt ¼ 100 MPa/min, typically leading to fracture in 43 min. The remaining three specimens were subject to LCF tests with a stress ratio of R=0.01. The loading spectrum comprised relatively rapid loading and unloading (each taking 1 min), and a hold time of 2 h at the peak stress. This spectrum was chosen to simulate the loading conditions that a gas turbine engine component might encounter in service. The peak stress levels were selected on the basis of the monotonic tensile test results. In cases in which fracture did not occur, the tests were terminated after 1 week ( 160–170 h) of loading. The room temperature tests were also performed at a stressing rate of 100 MPa/min. The monotonic tensile test results on the unnotched specimens are plotted in Fig. 2. At both 20 and 815C, the tensile response is essentially bilinear, with the transition in slope occurring at the onset of matrix cracking at a stress of 150–165 MPa. There was no apparent effect of temperature on this response, except that the fracture strain at 815C was slightly lower than that at 20C. Optical microscopy of polished cross-sections through the fractured test specimens revealed periodic matrix cracks, as commonly observed in this class of CFCC. The effects of holes on the net-section strength at room temperature are shown in Fig. 3. The strength decreases with increasing hole size, to 70% of the unnotched strength at 2a=6.3 mm. Similar reductions have been reported previously for SiC/SiC, SiC/MAS [5] and all-oxide ceramic composites [35]. These effects can be rationalized on the basis of the stress redistribution due to inelastic straining around the hole as well as the size-scale dependence of the local conditions needed to precipitate fracture, as described in Section 4. The LCF behavior of 815C is shown in Fig. 4. Several notable features emerge. (i) The monotonic tension tests on the center-hole and center-notch geometries yield strengths similar to one another, c160–165 MPa, Fig. 2. Tensile response of unnotched specimens at both 20 and 815C. Note the strong similarity in the stress-strain response at the two test temperatures, the key difference being the strains at which fracture occurs. The solid lines represent a bilinear fit to the data, used for calibrating the nonlinear constitutive law. Fig. 3. Effects of center holes on room-temperature tensile strength. The solid line represents the prediction of the point stress fracture model for a characteristic distance of d=0.75 mm. Fig. 4. Summary of LCF life at 815C. Arrows pointing to the right indicate runout; arrows pointing to the left indicate test results obtained from monotonic tension tests. J.C. McNulty et al. / Composites Science andTechnology 61 (2001) 1331–1338 1333
J.C. McNulty et al. /Composites Science and Technolog y 61(2001)1331-1338 despite large differences in the elastic stress concentration factors: ke=2.5 vs. 7. 1(see Appendix). These strengths are c60% of the unnotched monotonic strength, u N 270 MPa, measured at 815C. Evidently the degree of notch sensitivity at this temperature is only slightly greater than that at 25C.(ii) In the unnotched specimens, a fatigue threshold(defined by tr> 160h)was obtained at a stress level, oth N 165 MPa, corresponding closely to the matrix cracking limit: ome A 150-160 MPa(Fig. 2) relations between the threshold stress and the matrix cracking stress had been established previously for static loading of a CVI SiC/SiC at a temperature of 900C [31]. However, the cracking stress of the latter material was considerably lower(60-70 MPa), a consequence of the matrix porosity. (iii)In the center-hole and cen- 等0m ter-notch specimens, the threshold was considerably ower, oth N 60-85 MPa, representing 240-50% of the unnotched threshold. Essentially the same behavior was obtained with notches and holes. (iv) For all three spe- cimen geometries, the stress rupture curves were extre- mely shallow, with the exception of the regime at very short fracture times (r<<I h). Evidently fracture occurs very rapidly at stresses above the threshold level. The implication is that life prediction at stresses above the threshold is likely to be exceedingly difficult Representative fractured test specimens were exam ined in a scanning electron microscope (SEM). Typical Fig. 5. Fracture surface of a center hole specimen tested under features are shown in Figs. 5 and 6 Specimens subjected monotonic tension at 815C to monotonic tensile loading at elevated temperature exhibited little fiber pullout(Fig. 5), suggesting a rela- 4. Modeling of notch sensitivity tively high interfacial strength. However, there was no indication of oxidation on the fracture surfaces follow- In light of the extremely shallow stress rupture curves, ng the short exposure times( 3 min) of these test the modelling activity focused predominantly on the pecimens. Similarly low levels of pullout were observed effects of notches and holes on the threshold stress levels on the fracture surfaces of specimens tested at room A secondary priority was the notch sensitivity of the emperature monotonic tensile strength at ambient temperature By contrast, the LCF specimens exposed to elevated temperature for many hours exhibited significant levels 4.I. Notch sensitivity of monotonic tensile strengt of oxidation on the fracture plane. Fig. 6 shows micro- graphs of a center-hole specimen, tested at a peak stress, The notch sensitivity of strength under ambient(non ap=85 MPa. The regions of the fracture surface near oxidizing) conditions is dictated by two factors: ()the the hole edge exhibited particularly severe oxidation, as extent to which inelastic straining mitigates the stress manifest by the presence of a glassy layer on the axial concentration at the notch tip, and (i) the dependence fiber tows. Similar features were also observed on the of the fracture condition on the stress gradients that center-notch specimens The inference from the presence exist ahead of the notch. The former effects can be cal- of the glassy layer on the fiber fracture surfaces is that culated by finite element methods using a non-linear the fibers are failing progressively, starting near the tip constitutive law appropriate to CFCCs In the present of the notch or hole(where the stress concentration is at study, the law developed by Genin and Hutchinson [36] a maximum) and proceeding through the remaining has been used for such calculations. This constitutive (unnotched) section of the material. This sequence is law is based on a phenomenological description of the consistent with that observed in other SiC-based development of inelastic strain in cross-ply CFCCs CFCCS, as detailed in [15] under biaxial stressing. The pertinent functions are
despite large differences in the elastic stress concentration factors: ke=2.5 vs. 7.1 (see Appendix). These strengths are 60% of the unnotched monotonic strength, u 270 MPa, measured at 815C. Evidently the degree of notch sensitivity at this temperature is only slightly greater than that at 25C. (ii) In the unnotched specimens, a fatigue threshold (defined by tf5 160h) was obtained at a stress level, O th 165 MPa, corresponding closely to the matrix cracking limit: mc 150–160 MPa (Fig. 2). Similar correlations between the threshold stress and the matrix cracking stress had been established previously for static loading of a CVI SiC/SiC at a temperature of 900C [31]. However, the cracking stress of the latter material was considerably lower (60–70 MPa), a consequence of the matrix porosity. (iii) In the center-hole and center-notch specimens, the threshold was considerably lower, th 60–85 MPa, representing 40–50% of the unnotched threshold. Essentially the same behavior was obtained with notches and holes. (iv) For all three specimen geometries, the stress rupture curves were extremely shallow, with the exception of the regime at very short fracture times (tf<<1 h). Evidently fracture occurs very rapidly at stresses above the threshold level. The implication is that life prediction at stresses above the threshold is likely to be exceedingly difficult. 3. Fractography Representative fractured test specimens were examined in a scanning electron microscope (SEM). Typical features are shown in Figs. 5 and 6. Specimens subjected to monotonic tensile loading at elevated temperature exhibited little fiber pullout (Fig. 5), suggesting a relatively high interfacial strength. However, there was no indication of oxidation on the fracture surfaces following the short exposure times ( 3 min) of these test specimens. Similarly low levels of pullout were observed on the fracture surfaces of specimens tested at room temperature. By contrast, the LCF specimens exposed to elevated temperature for many hours exhibited significant levels of oxidation on the fracture plane. Fig. 6 shows micrographs of a center-hole specimen, tested at a peak stress, p ¼85 MPa. The regions of the fracture surface near the hole edge exhibited particularly severe oxidation, as manifest by the presence of a glassy layer on the axial fiber tows. Similar features were also observed on the center-notch specimens. The inference from the presence of the glassy layer on the fiber fracture surfaces is that the fibers are failing progressively, starting near the tip of the notch or hole (where the stress concentration is at a maximum) and proceeding through the remaining (unnotched) section of the material. This sequence is consistent with that observed in other SiC-based CFCCs, as detailed in [15]. 4. Modeling of notch sensitivity In light of the extremely shallow stress rupture curves, the modelling activity focused predominantly on the effects of notches and holes on the threshold stress levels. A secondary priority was the notch sensitivity of the monotonic tensile strength at ambient temperature. 4.1. Notch sensitivity of monotonic tensile strength The notch sensitivity of strength under ambient (nonoxidizing) conditions is dictated by two factors: (i) the extent to which inelastic straining mitigates the stress concentration at the notch tip, and (ii) the dependence of the fracture condition on the stress gradients that exist ahead of the notch. The former effects can be calculated by finite element methods using a non-linear constitutive law appropriate to CFCCs. In the present study, the law developed by Genin and Hutchinson [36] has been used for such calculations. This constitutive law is based on a phenomenological description of the development of inelastic strain in cross-ply CFCCs under biaxial stressing. The pertinent functions are Fig. 5. Fracture surface of a center hole specimen tested under monotonic tension at 815C. 1334 J.C. McNulty et al. / Composites Science andTechnology 61 (2001) 1331–1338
J.C. McNulty et al. /Composites Science and Technolog y 61(2001)1331-1338 calibrated using the results of tensile tests performed both into account [5, 35]. Furthermore, previous studies have along one of the fiber axes(in the 00/90 orientation) and revealed that the characteristic distance inferred from at 45 to the fiber axes. The constitutive law is then the model and the experiment is rather insensitive to the implemented in a finite-element code for the purpose of details of the composite microstructure, falling in the calculating stress and strain fields in non-uniform spe narrow range of 0.5-0.75 mm for SiC/SiC [5], SiC/ men or component geometries. The capability of the MAs [5] and an all-oxide CFCC [35] model to predict the latter response has been validated This approach has been used to model the strength of through comparisons with strain measurements on spe- the center-hole specimens of the SylramicM/SiC com- cimens containing notches and circular holes [5, 35]. The posite tested at room temperature, using a characteristic model tacitly assumes that the material response is distance, d=0. 75 mm. The effects of inelastic straining on scale-independent. This presents some limitations on the the stress distribution ahead of a hole are illustrated by the size-scale of features that can be accurately simulated, as results plotted in Fig. 7(a). The strength predictions are discussed in Section 4.2. Additional details of the imple- plotted in Fig 3, for comparison with the experimental mentation and calibration procedures are described in measurements. Excellent agreement between experiment Refs. [36] and [5] and theory is obtained. Furthermore, the inferred char- he effects of stress gradients on fracture are less well acteristic distance is remarkably similar to the values nderstood. Nevertheless, a simple fracture criterion obtained in other CFCCs, indicating a commonalty in based on the attainment of a critical stress( taken as the the factors controlling fracture of the various composite unnotched strength)over a characteristic distance d along systems the incipient fracture plane has been found to adequately The same approach has been used to rationalize the describe notched strength of CFCCs, provided the effects effects of both notches and holes on the strength at 815oC. of inelastic straining on the stress distribution are taken In this case, the characteristic distance was inferred 2.0 g =75 MPa 150M g15 0.10 Distance from hole edge, X/(W-a) m tch length, 2a=6. 35 mm EM model predictions 00002 0060.080.10 Fig. 7. Typical stress distributions for(a) center-hole and(b)center- 10m notched specimens of the SylramicTM/SiC composite, calculated using the nonlinear constitutive law of Genin and Hutchinson [36]. Also shown for comparison are the corresponding elastic predictions. The Fig. 6. Fracture surface of a center hole specimen tested in fatigue at results demonstrate the role of inelastic straining in mitigating stress ap=85 MPa at8l5°C
calibrated using the results of tensile tests performed both along one of the fiber axes (in the 0/90 orientation) and at 45 to the fiber axes. The constitutive law is then implemented in a finite-element code for the purpose of calculating stress and strain fields in non-uniform specimen or component geometries. The capability of the model to predict the latter response has been validated through comparisons with strain measurements on specimens containing notches and circular holes [5,35]. The model tacitly assumes that the material response is scale-independent. This presents some limitations on the size-scale of features that can be accurately simulated, as discussed in Section 4.2. Additional details of the implementation and calibration procedures are described in Refs. [36] and [5]. The effects of stress gradients on fracture are less well understood. Nevertheless, a simple fracture criterion based on the attainment of a critical stress (taken as the unnotched strength) over a characteristic distance d along the incipient fracture plane has been found to adequately describe notched strength of CFCCs, provided the effects of inelastic straining on the stress distribution are taken into account [5,35]. Furthermore, previous studies have revealed that the characteristic distance inferred from the model and the experiment is rather insensitive to the details of the composite microstructure, falling in the narrow range of 0.5–0.75 mm for SiC/SiC [5], SiC/ MAS [5] and an all-oxide CFCC [35]. This approach has been used to model the strength of the center-hole specimens of the SylramicTM/SiC composite tested at room temperature, using a characteristic distance, d=0.75 mm. The effects of inelastic straining on the stress distribution ahead of a hole are illustrated by the results plotted in Fig. 7(a). The strength predictions are plotted in Fig. 3, for comparison with the experimental measurements. Excellent agreement between experiment and theory is obtained. Furthermore, the inferred characteristic distance is remarkably similar to the values obtained in other CFCCs, indicating a commonalty in the factors controlling fracture of the various composite systems. The same approach has been used to rationalize the effects of both notches and holes on the strength at 815C. In this case, the characteristic distance was inferred Fig. 6. Fracture surface of a center hole specimen tested in fatigue at p ¼ 85 MPa at 815C. Fig. 7. Typical stress distributions for (a) center-hole and (b) centernotched specimens of the SylramicTM/SiC composite, calculated using the nonlinear constitutive law of Genin and Hutchinson [36]. Also shown for comparison are the corresponding elastic predictions. The results demonstrate the role of inelastic straining in mitigating stress concentrations. J.C. McNulty et al. / Composites Science andTechnology 61 (2001) 1331–1338 1335
J.C. McNulty et al. /Composites Science and Technolog y 61(2001)1331-1338 Unnotched 方1 Center 100 = 50 一 Fig 8. Comparisons of the measured and predicted monotonic tensile strengths of specimens with holes and notches at 815C. The strength underestimated using the elastic stress concentration factor in com- Time to Fracture, t, (hours) bination with a critical stress fracture criterion, i.e. on=Ouke. In contrast, the model based on the inelastic stress distribution coupled with the point-stress criterion provides an adequate representation, for d0.3-0.5mm. a.250 Unnotched from comparisons of the measured and predicted not- ched strengths, assuming various values for d in the model. The comparisons are presented in Fig. 8. The 150 inferred characteristic distance is d 0.3-0.5 mm. con Center notch d=0. 3 mm erably smaller than the value obtained from the room 100 temperature tests (d N 0.75 mm). This reduction in d reflects a degradation in notch-sensitivity. It is unknown whether this degradation is intrinsic to the fracture mechanism at elevated temperature, it reflects the onset a of oxidative embrittlement or it is associated with scat Time to Fracture, t(hours) ter in the experimental data Fig. 9. Comparisons of measured and predicted thresholds at 815C 4.2. Notch sensitivity of the threshold stress for (a) center-hole and(b)center- notch specimens, based on the non- linear finite element model and the point-stress fracture criterion. An analogous stress-based fracture model has been used to rationalize the threshold stress levels in the cen in combination with the matrix cracking stress, where- ter-notch and center-hole specimens. The threshold upon the predicted threshold becomes stress was taken to be that for which the longitudinal stress,o, along the incipient fracture plane exceeds the oth=oth/ke =Omc/ke (1) unnotched threshold, ogh N 165 MPa, over a character- stic distance, duh, ahead of the hole or notch. As before This represents a lower-bound estimate of the thresh- the characteristic distance was obtained by comparing old stress For the center-hole specimen with a/W=0. 2, the calculated results with the experimental measure- ke=2.5 [37](Appendix) and thus oth N 66 MPa [also ments. Such comparisons are plotted in Fig 9. plotted on Fig. 9(a)]. This prediction is in reasonable or the center-hole specimens, the inferred distance agreement with the range obtained from the experi- spans the range dth 0-0.2 mm: the variation arising ments(60-85 MPa), reaffirming the poor notch sensi from the difference between the lowest stress at which tivity associated with LCF fracture at the test fracture had occurred, A 85 MPa, and the highest stress temperature. Furthermore, the extreme sensitivity of the at which run-out was obtained, A 60 MPa. Such low fracture time to the presence of matrix cracks is con- values of dth are indicative of extremely strong notch sen sistent with the oxidation embrittlement characteristics sitivity. Indeed, in the limit of dth=0, the threshold corre- of other Sic-based CFCCs at intermediate temperatures sponds to the case where the maximum stress(at the hole [12, 15, 17, 22, 23, 27, 38] edge)is equal to the unnotched threshold stress. Since the For the center-nlotch specimens, the inferred char- unnotched threshold is dictated by the matrix cracking acteristic distance is somewhat larger, duh A03-0.4 stress and the material response up cracking limit mm, indicating improved damage tolerance and lower is elastic, a conservative estimate of the threshold can be notch sensitivity. Furthermore, the lower bound esti obtained using the elastic stress concentration factor ke mate of the threshold stress, Oth N 23 MPa, based on
from comparisons of the measured and predicted notched strengths, assuming various values for d in the model. The comparisons are presented in Fig. 8. The inferred characteristic distance is d 0.3–0.5 mm, considerably smaller than the value obtained from the room temperature tests (d 0.75 mm). This reduction in d reflects a degradation in notch-sensitivity. It is unknown whether this degradation is intrinsic to the fracture mechanism at elevated temperature, it reflects the onset of oxidative embrittlement, or it is associated with scatter in the experimental data. 4.2. Notch sensitivity of the thresholdstress An analogous stress-based fracture model has been used to rationalize the threshold stress levels in the center-notch and center-hole specimens. The threshold stress was taken to be that for which the longitudinal stress, , along the incipient fracture plane exceeds the unnotched threshold, o th 165 MPa, over a characteristic distance, dth, ahead of the hole or notch. As before, the characteristic distance was obtained by comparing the calculated results with the experimental measurements. Such comparisons are plotted in Fig. 9. For the center-hole specimens, the inferred distance spans the range dth 00:2 mm: the variation arising from the difference between the lowest stress at which fracture had occurred, 85 MPa, and the highest stress at which run-out was obtained, 60 MPa. Such low values of dth are indicative of extremely strong notch sensitivity. Indeed, in the limit of dth=0, the threshold corresponds to the case where the maximum stress (at the hole edge) is equal to the unnotched threshold stress. Since the unnotched threshold is dictated by the matrix cracking stress and the material response up to the cracking limit is elastic, a conservative estimate of the threshold can be obtained using the elastic stress concentration factor ke in combination with the matrix cracking stress, whereupon the predicted threshold becomes th ¼ o th=ke ¼ mc=ke ð1Þ This represents a lower-bound estimate of the threshold stress. For the center-hole specimen with a=W ¼ 0:2, ke=2.5 [37] (Appendix) and thus th 66 MPa [also plotted on Fig. 9(a)]. This prediction is in reasonable agreement with the range obtained from the experiments (60–85 MPa), reaffirming the poor notch sensitivity associated with LCF fracture at the test temperature. Furthermore, the extreme sensitivity of the fracture time to the presence of matrix cracks is consistent with the oxidation embrittlement characteristics of other SiC-based CFCCs at intermediate temperatures [12,15,17,22,23,27,38]. For the center-notch specimens, the inferred characteristic distance is somewhat larger, dth 0:30:4 mm, indicating improved damage tolerance and lower notch sensitivity. Furthermore, the lower bound estimate of the threshold stress, th 23 MPa, based on Fig. 8. Comparisons of the measured and predicted monotonic tensile strengths of specimens with holes and notches at 815C. The strength is underestimated using the elastic stress concentration factor in combination with a critical stress fracture criterion, i.e. n ¼ u=ke. In contrast, the model based on the inelastic stress distribution coupled with the point-stress criterion provides an adequate representation, for d 0.3–0.5 mm. Fig. 9. Comparisons of measured and predicted thresholds at 815C for (a) center-hole and (b) center-notch specimens, based on the nonlinear finite element model and the point-stress fracture criterion. 1336 J.C. McNulty et al. / Composites Science andTechnology 61 (2001) 1331–1338
J C. McNulty et al. /Composites Science and Technolog y 61(2001)1331-1338 1337 1)and the elastic stress concentration factor, times in the post-cracking regime are not likely to be ke=7. 1 [39], is significantly below the measured value, useful, because of the precipitous drop in fracture time C70-85 MPa. This difference is believed to be asso- to unacceptably low levels ciated with the steep stress gradients that exist in the The LCF threshold for center-hole specimens is simi notch tip region coupled with the intrinsic volume larly dictated by the matrix cracking limit. Once the ependence of the matrix strength. That is, the high cracking stress is exceeded locally (even over a very small stress levels exist over only small volumes ahead of the distance ahead of the hole), fracture ensues. A reasonably notch (over distances p), reducing significantly the accurate estimate of the threshold is obtained through the probability of encountering a matrix flaw to initiate elastic stress concentration factor in combination with the cracking. Finite element calculations for the center-not- matrix cracking stress [Eqn.(1)). This highly notch-sen ched geometry used in the present study (with 2a=6.35 sitive behavior suggests the need for an extremely con- mm and p=0.2 mm)reveal that the elastic stress servative design approach for CFCC structures diminishes rapidly with distance from the notch tip, by containing holes. Furthermore, it precludes the exploi- N50%at a distance of 0. 1 mm [Fig. 7(b)]. By contrast, tation of the inelastic deformation mechanisms that the reduction in elastic stress in the center-hole specimen operate in these materials and impart the requisite with a hole diameter of 6.35 mm is only A6%at the damage tolerance in non-oxidizing environments same distance from the hole edge. )It is surmised that The LCF threshold for center-notch specimens is com- even a modest volume dependence of strength might parable to that of the center hole specimens. Recognizing lead to a significant elevation in the maximum local that the stress concentration facto er in the not- stress at the onset of cracking. This elevation in the local ched geometry, the composite appears to exhibit superior cracking stress may account for the rather high threshold damage tolerance in the notched geometry. However, this of the notched specimens. If so, the beneficial effects of apparent damage tolerance may be a result of the volume inelastic straining on the stress concentration may be dependence of the matrix cracking stress, not a con- overestimated by the non-linear constitutive law used in sequence of local inelastic straining. Further experiments this study, because, in essence, the volume dependence of are needed to determine the statistical parameters asso- the cracking stress is neglected. A further implication is ciated with the matrix strength and to assess their role in that the inferred characteristic distance (dth a 0.3-0.4 the conditions for fracture ahead of a notch. Additional mm)may be anomalously high. It remains to be estab- experiments are also needed to assess the extent of lished which of these effects is dominant catter in fatigue life, for the purpose of validating the failure models presented here Appendix. Elastic stress concentrations in notched tensile The notch sensitivity of the tensile strength of the specimens SylramicTM/SiC composite at ambient temperature is consistent with the trends observed in other fccs The elastic stress concentration factors k. were cal- Typically a reduction of a25% in net-section strength is culated assuming that the composite is elastically iso- obtained for open holes of diameter 6 mm. This modest tropic. This assumption can be justified on the basis that amount of notch sensitivity can be rationalized on the the Youngs moduli measured in the 0 /90 and +450 basis of the stress redistribution that occurs as a result orientations were within 2% of one another (250 GPa) of local inelastic straining and the size-scale dependence This behavior is attributable to the fact that both the of the conditions at the onset of fracture through the matrix and the fibers are essentially pure Sic point stress fracture criterion. The notch sensitivity at For a specimen of width 2W and with a center hole of elevated temperature appears to be reduced slightly, as diameter 2a, ke is given by [37] manifest in a reduction in the inferred characteristic distance in the point stress criterion The unnotched LCF threshold for the SylramicM/ Sic ke=0=2+(1-a) (A) composite at 815C is dictated by the matrix cracking stress. Above the cracking stress, fracture occurs rapidly typically in >1)is given by [391
Eq. (1) and the elastic stress concentration factor, ke=7.1 [39], is significantly below the measured value, 70–85 MPa. This difference is believed to be associated with the steep stress gradients that exist in the notch tip region coupled with the intrinsic volume dependence of the matrix strength. That is, the high stress levels exist over only small volumes ahead of the notch (over distances ), reducing significantly the probability of encountering a matrix flaw to initiate cracking. Finite element calculations for the center-notched geometry used in the present study (with 2a=6.35 mm and =0.2 mm) reveal that the elastic stress diminishes rapidly with distance from the notch tip, by 50% at a distance of 0.1 mm [Fig. 7(b)]. (By contrast, the reduction in elastic stress in the center-hole specimen with a hole diameter of 6.35 mm is only 6% at the same distance from the hole edge.) It is surmised that even a modest volume dependence of strength might lead to a significant elevation in the maximum local stress at the onset of cracking. This elevation in the local cracking stress may account for the rather high threshold of the notched specimens. If so, the beneficial effects of inelastic straining on the stress concentration may be overestimated by the non-linear constitutive law used in this study, because, in essence, the volume dependence of the cracking stress is neglected. A further implication is that the inferred characteristic distance (dth 0.3–0.4 mm) may be anomalously high. It remains to be established which of these effects is dominant. 5. Summary The notch sensitivity of the tensile strength of the SylramicTM/SiC composite at ambient temperature is consistent with the trends observed in other CFCCs. Typically a reduction of 25% in net-section strength is obtained for open holes of diameter 6 mm. This modest amount of notch sensitivity can be rationalized on the basis of the stress redistribution that occurs as a result of local inelastic straining and the size-scale dependence of the conditions at the onset of fracture through the point stress fracture criterion. The notch sensitivity at elevated temperature appears to be reduced slightly, as manifest in a reduction in the inferred characteristic distance in the point stress criterion. The unnotched LCF threshold for the SylramicTM/SiC composite at 815C is dictated by the matrix cracking stress. Above the cracking stress, fracture occurs rapidly, typically in 4100 h. Moreover, the LCF life curve for stresses slightly above the threshold is extremely shallow, indicating a very strong sensitivity of fracture time to applied stress. For the lifetime requirements of most gas turbine engine components (103 –104 h), the allowable stress levels would have to remain below this threshold. A further implication is that models to predict fracture times in the post-cracking regime are not likely to be useful, because of the precipitous drop in fracture time to unacceptably low levels. The LCF threshold for center-hole specimens is similarly dictated by the matrix cracking limit. Once the cracking stress is exceeded locally (even over a very small distance ahead of the hole), fracture ensues. A reasonably accurate estimate of the threshold is obtained through the elastic stress concentration factor in combination with the matrix cracking stress [Eqn. (1)]. This highly notch-sensitive behavior suggests the need for an extremely conservative design approach for CFCC structures containing holes. Furthermore, it precludes the exploitation of the inelastic deformation mechanisms that operate in these materials and impart the requisite damage tolerance in non-oxidizing environments. The LCF threshold for center-notch specimens is comparable to that of the center hole specimens. Recognizing that the stress concentration factor is higher in the notched geometry, the composite appears to exhibit superior damage tolerance in the notched geometry. However, this apparent damage tolerance may be a result of the volumedependence of the matrix cracking stress, not a consequence of local inelastic straining. Further experiments are needed to determine the statistical parameters associated with the matrix strength and to assess their role in the conditions for fracture ahead of a notch. Additional experiments are also needed to assess the extent of scatter in fatigue life, for the purpose of validating the failure models presented here. Appendix. Elastic stress concentrations in notched tensile specimens The elastic stress concentration factors, ke; were calculated assuming that the composite is elastically isotropic. This assumption can be justified on the basis that the Young’s moduli measured in the 0/90 and 45 orientations were within 2% of one another (250 GPa). This behavior is attributable to the fact that both the matrix and the fibers are essentially pure SiC. For a specimen of width 2W and with a center hole of diameter 2a, ke is given by [37]: ke o net ¼ 2 þ 1 a W 3 ðA1Þ where o is the maximum longitudinal stress (at the hole edge) and net is the applied net-section stress. For the center hole geometry used in the present experiments, a=W ¼ 0:2 and thus ke=2.51. The elastic stress concentration factor for a specimen of width 2W and with a center notch of length 2a and root radius (with a= 1) is given by [39]: J.C. McNulty et al. / Composites Science andTechnology 61 (2001) 1331–1338 1337
1338 J.C. McNulty et al. Composites Science and Technology 61(2001)1331-1338 [18 Lara-Curzio E. Ferber MK. Stress rupture of continuous fiber 2 ceramic composites at intermediate temperatures. J Mater Sci For the center notch geometry used in the present [19] Lara-Curzio E. Oxidation induced stress-rupture of fiber bundles. experiments, a/w=0.2 and a/p= 15.9, and thus ke J Eng Mater Technol 1998: 120: 105-9 20]Llorca J, Elices M, Celemin JA. Toughness and microstructural 7.1 Acta Mater1998:46(7):244l-53 21Lara-Curzio E. Analysis of oxidatin-assisted stress-rupture of References ntinuous fiber-reinforced ceramic composites at intermediate [Bannister RL, Ceruvu NS, Little DA. McQuiggan G. Develop. [22] Naslain R, Lamon J, Pailler R, Bourrat X, Guette A, Langlais F. ment requirements for an advanced gas turbine system. Trans Micro/minicomposites: a useful approach to the design and SME1995;117:724-33 development of non-oxide CMCs Composites 1999: 30A: 547-7 2 Smith KO, Fahme A. Experimental assessment of the er [23 Luh EY, Evans AG. High-temperature mechanical properties of benefits of a ceramic gas turbine combustor. Proc. ASME IGT amic matrix composite. J Am Ceram Soc 1987: 70: 466-9 Conference. 1996 [24 Wetherhold RC, Zawada LP. Heat treatments as a method of 3 Droillard C, Lamon J Fracture toughness of 2D woven SiC/Sic protection for a ceramic fiber-glass matrix composite. J Am ilayered interphases. J Am Ceram Soc Ceram Soc1991;74(8):2097-100 99679(4):84958. 25 Henager Jr. CH, Jones RH. Sub-critical crack growth in CV 4 Kagawa Y, Goto K. Notch sensitivity of two-dimensional woven silicon carbide reinforced with Nicalon fibers- experiment and Sic fiber-reinforced Sic matrix composite fabricated by the model. J Am Ceram Soc 1994 77(9): 2381-94 olymer conversion process. J Mater Sci Letters 1997: 16: 8 26 Raghuraman S, Stubbins JF. Ferber MK, Wereszczak AA Crack McNulty JC, Zok Fw, Genin G. Evans AG. Notch-sensitivity of agation in SiCr/Sic matrix composite under static effects of 212-215: straining and volume-dependent strength. J Am Ceram Soc [27] Plucknett KP, Sutherland S, Daniel AM, Cain RL, West G 6 Prewo KM. Fatigue and stress rupture of silicon carbide fibre- Taplin DMR, Lewis MH. Environmental ageing effects in a si inforced glass-ceramics. J Mater Sci 1987: 22: 2695-701 con carbide fiber-reinforced glass-ceramic matrix composite [7 Cooper RF, Chyung K. Structure and chemistry of fb Microscopy 1995: 177(3): 251-63 nterfaces in silicon carbide fiber-reinforced glass-ceramic com 28 Plucknett KP, Lewis MH. Inhibition of intermediate posites: an electron roscopy stud degradation of calcium aluminosilicate/Nicalon by erature pretreatment. J Mater Sci Lett 1995; 14: 1223-6 [8 Bischoff E, Ruhle M, Sbaizero O. Evans AG. Microstructural [29 Reynaud P Cyclic fatigue of ceramic-matrix composites at ambi- ies of the interfacial zone of a sic-fiber- rein ech1996:56:809-14 luminosilicate glass-ceramic. J Am Ceram Soc 1989: 72(5): 741-5. [30] Mizuno M, Zhu S, Nagano Y, Sakaida Y, Ki Bonney LA, Cooper RF. Reaction-layer interfaces in SiC-fiber 1. Cyclic fatigue behavior of Sic/siC composites at room and high temperatures. J Am Ceram Soc 1996: 79(12): 3065-77. electron microscopy analysis. J Am Ceram Soc 1990, 73(10): 2916-2 31 Steyer TE, Zok FW, Walls DP Stress rupture of an enhanced [10 Qi G, Spear KE, Pantano CG. Carbon-layer formation at silicon Nicalon/SiC composite at intermediate temperatures. JAm carbide-glass interfaces Mater Sci and Eng 1993 A162: 4 Ceram Soc 1998: 81: 2140--6. [1 Pharaoh MW, Daniel AM, Lewis MH. Stability of interfaces in [32] Zhu S, Mizuno M, Nagano Y, Cao J, calcium aluminosilicate matrix/Nicalon SiC fibre composites nd fatigue behavior in an enhanced Mater sci lett1993:12(13):998-1001. [12 Filipuzzi L, Camus G, Naslain R, Thebault J. Oxidation 33] Lee SS, Zawada LP, Stachler JM CA. Mechanical ms and kinetics of ID-SiC/C/SiC composite materials: I behavior and high temperature performance of a woven nicalon/ an experimental approach. J Am Ceram Soc 1994: 77(2): 459-66. Si-N-C matrix composite. J Am Ceram Soc 1998: 81(7): 1797-811 [13 Filipuzzi L, Naslain R. Oxidation mechanisms and kinetics of 34 Morscher GN. Tensile stress rupture of SiCrSicm mini- ID-SiC/C/SiC composite materials: Il, modeling. J Am Ceram composites with carbon and boron nitride interphases at elevated Soc1994:77(2):467-80 temperatures in air. J Am Ceram Soc 1997: 80(8): 2029-42 [14 Huger M, Fargeot D, Gault C. Ultrasonic characterization of 35 Heathcote JA. Gong X-Y, Yang J, Ramamurty U, Zok Fw. In- oxidation mechanisms in Nicalon/C/SiC composites. JAm ne mechanical properties of an all-oxide ceramic composite Ceram Soc199477(10):255460. Am Ceram Soc I999;82(10):2721-30. [15 Heredia FE, McNulty JC, Zok FW, Evans AG. An oxidation [36] Genin G, Hutchinson Jw. Composite laminates in plane stress soc roget probe for ceramic matrix composites. JAm Ceram nstitute deling and stress redistribution due to matrix [16 Labrugere C, Guillaumat L, Guette A, Naslain R. Effect of age. [37] Peterson RE Stress Concentration Factors. New York: Wiley, ng treatments at high temperatures on the microstructure and echanical behavior of 2D Nicalon/C/SiC composites. 2: ageing [38] Singh D, Singh JP, Wheeler MJ. Mechanical behavior of Sicd Sic composites and correlation to in situ fiber strength at re and elevated temperatures. J Am Ceram Soc 1996: 79(3): 591 [7 Lara-Curzio E. Stress rupture of Nicalon/SiC continuous fiber [39] Suresh S. Fatigue of Materials. Cambridge: Cambridge Uni- ceramic composites in air. J Am Ceram Soc 1997: 80(12): 3268-72 ersity Press, 1991( Chapter 8)
ke o net ¼ 2 1 a W ffiffiffiffiffi a r sec a W 1=2 ðA2Þ For the center notch geometry used in the present experiments, a=W ¼ 0:2 and a= ¼ 15:9, and thus ke ¼ 7:1: References [1] Bannister RL, Ceruvu NS, Little DA, McQuiggan G. Development requirements for an advanced gas turbine system. Trans ASME 1995;117:724–33. [2] Smith KO, Fahme A. Experimental assessment of the emissions benefits of a ceramic gas turbine combustor. Proc. ASME IGT Conference, 1996. [3] Droillard C, Lamon J. Fracture toughness of 2D woven SiC/SiC CVI composites with multilayered interphases. J Am Ceram Soc 1996;79(4):849–58. [4] Kagawa Y, Goto K. Notch sensitivity of two-dimensional woven SiC fiber-reinforced SiC matrix composite fabricated by the polymer conversion process. J Mater Sci Letters 1997;16:850–4. [5] McNulty JC, Zok FW, Genin G, Evans AG. Notch-sensitivity of fiber-reinforced ceramic-matrix composites: effects of inelastic straining and volume-dependent strength. J Am Ceram Soc 1999;82(5):1217–28. [6] Prewo KM. Fatigue and stress rupture of silicon carbide fibrereinforced glass-ceramics. J Mater Sci 1987;22:2695–701. [7] Cooper RF, Chyung K. Structure and chemistry of fiber-matrix interfaces in silicon carbide fiber-reinforced glass-ceramic composites: an electron microscopy study. J Mater Sci 1987;22:3148– 60. [8] Bischoff E, Ruhle M, Sbaizero O, Evans AG. Microstructural studies of the interfacial zone of a SiC-fiber-reinforced lithium aluminosilicate glass-ceramic. J Am Ceram Soc 1989;72(5):741–5. [9] Bonney LA, Cooper RF. Reaction-layer interfaces in SiC-fiberreinforced glass ceramics: a high-resolution scanning transmission electron microscopy analysis. J Am Ceram Soc 1990;73(10):2916–21. [10] Qi G, Spear KE, Pantano CG. Carbon-layer formation at siliconcarbide-glass interfaces. Mater Sci and Eng 1993;A162:45–52. [11] Pharaoh MW, Daniel AM, Lewis MH. Stability of interfaces in calcium aluminosilicate matrix/Nicalon SiC fibre composites. J Mater Sci Lett 1993;12(13):998–1001. [12] Filipuzzi L, Camus G, Naslain R, Thebault J. Oxidation mechanisms and kinetics of 1D-SiC/C/SiC composite materials: I, an experimental approach. J Am Ceram Soc 1994;77(2):459–66. [13] Filipuzzi L, Naslain R. Oxidation mechanisms and kinetics of 1D-SiC/C/SiC composite materials: II, modeling. J Am Ceram Soc 1994;77(2):467–80. [14] Huger M, Fargeot D, Gault C. Ultrasonic characterization of oxidation mechanisms in Nicalon/C/SiC composites. J Am Ceram Soc 1994;77(10):2554–60. [15] Heredia FE, McNulty JC, Zok FW, Evans AG. An oxidation embrittlement probe for ceramic matrix composites. J Am Ceram Soc 1995;78:2097–100. [16] Labrugere C, Guillaumat L, Guette A, Naslain R. Effect of ageing treatments at high temperatures on the microstructure and mechanical behavior of 2D Nicalon/C/SiC composites. 2: ageing under CO and influence of a SiC seal-coating. J Eur Ceram Soc 1997;17:541–57. [17] Lara-Curzio E. Stress rupture of Nicalon/SiC continuous fiber ceramic composites in air. J Am Ceram Soc 1997;80(12):3268–72. [18] Lara-Curzio E, Ferber MK. Stress rupture of continuous fiber ceramic composites at intermediate temperatures. J Mater Sci Lett 1997;16:23–6. [19] Lara-Curzio E. Oxidation induced stress-rupture of fiber bundles. J Eng Mater Technol 1998;120:105–9. [20] Llorca J, Elices M, Celemin JA. Toughness and microstructural degradation at high temperature in SiC fiber-reinforced ceramics. Acta Mater 1998;46(7):2441–53. [21] Lara-Curzio E. Analysis of oxidatin-assisted stress-rupture of continuous fiber-reinforced ceramic composites at intermediate temperatures. Composites 1999;30A:549–54. [22] Naslain R, Lamon J, Pailler R, Bourrat X, Guette A, Langlais F. Micro/minicomposites: a useful approach to the design and development of non-oxide CMCs. Composites 1999;30A:547–7. [23] Luh EY, Evans AG. High-temperature mechanical properties of a ceramic matrix composite. J Am Ceram Soc 1987;70:466–9. [24] Wetherhold RC, Zawada LP. Heat treatments as a method of protection for a ceramic fiber-glass matrix composite. J Am Ceram Soc 1991;74(8):2097–100. [25] Henager Jr. CH, Jones RH. Sub-critical crack growth in CVI silicon carbide reinforced with Nicalon fibers — experiment and model. J Am Ceram Soc 1994;77(9):2381–94. [26] Raghuraman S, Stubbins JF, Ferber MK, Wereszczak AA. Crack propagation in SiCf/SiC ceramic matrix composite under static and cyclic loading conditions. J Nuclear Mater 1994;212–215: 840–844. [27] Plucknett KP, Sutherland S, Daniel AM, Cain RL, West G, Taplin DMR, Lewis MH. Environmental ageing effects in a silicon carbide fiber-reinforced glass-ceramic matrix composite. J Microscopy 1995;177(3):251–63. [28] Plucknett KP, Lewis MH. Inhibition of intermediate temperature degradation of calcium aluminosilicate/Nicalon by high temperature pretreatment. J Mater Sci Lett 1995;14:1223–6. [29] Reynaud P. Cyclic fatigue of ceramic-matrix composites at ambient and elevated temperatures. Comp Sci Tech 1996;56:809–14. [30] Mizuno M, Zhu S, Nagano Y, Sakaida Y, Kagawa Y, Watanabe M. Cyclic fatigue behavior of SiC/SiC composites at room and high temperatures. J Am Ceram Soc 1996;79(12):3065–77. [31] Steyer TE, Zok FW, Walls DP. Stress rupture of an enhanced Nicalon/SiC composite at intermediate temperatures. J Am Ceram Soc 1998;81:2140–6. [32] Zhu S, Mizuno M, Nagano Y, Cao J, Kagawa Y, Kaya H. Creep and fatigue behavior in an enhanced SiC/SiC composite at elevated temperature. J Am Ceram Soc 1998;81(9):2269–77. [33] Lee SS, Zawada LP, Staehler JM, Folsom CA. Mechanical behavior and high temperature performance of a woven nicalon/ Si–N–C matrix composite. J Am Ceram Soc 1998;81(7):1797–811. [34] Morscher GN. Tensile stress rupture of SiCf/SiCm minicomposites with carbon and boron nitride interphases at elevated temperatures in air. J Am Ceram Soc 1997;80(8):2029–42. [35] Heathcote JA, Gong X-Y, Yang J, Ramamurty U, Zok FW. Inplane mechanical properties of an all-oxide ceramic composite. J Am Ceram Soc 1999;82(10):2721–30. [36] Genin G, Hutchinson JW. Composite laminates in plane stress: constitutive modeling and stress redistribution due to matrix cracking. J Am Ceram Soc 1997;80:1245–55. [37] Peterson RE. Stress Concentration Factors. New York: Wiley, 1974 p. 111. [38] Singh D, Singh JP, Wheeler MJ. Mechanical behavior of SiCf/ SiC composites and correlation to in situ fiber strength at room and elevated temperatures. J Am Ceram Soc 1996;79(3):591–6. [39] Suresh S. Fatigue of Materials. Cambridge: Cambridge University Press, 1991 (Chapter 8). 1338 J.C. McNulty et al. / Composites Science andTechnology 61 (2001) 1331–1338