《复合材料 Composites》课程教学资源（学习资料）第五章 陶瓷基复合材料_SiC-SiC-31

E噩≈S Journal of the European Ceramic Society 22(2002)2349-2356 www.elsevier.com/locate/jeurc Tensile fracture behavior of continuous sic fiber-reinforced SiC matrix composites at elevated temperatures and correlation to in Situ constituent properties Shuqi guo*,, Yutaka Kagawa Institute of Industrial Science, The University of Tokyo, 7-22-1, Roppongi, Minato-ku, Tokyo 106-8558, Japan Received 21 March 2001; received in revised form 20 December 2001: accepted 6 January 2002 Abstract The tensile fracture behavior and tensile mechanical properties of polymer infiltration pyrolysis(PIP)-processed two-dimensional plain-woven fabric carbon-coated NicalonM SiC fiber and BN-coated Hi-Nicalon"M SiC fiber-reinforced SiC matrix composites have been investigated. Tensile testing of the composites was carried out in air between 298 and 1400 K. In situ fiber strength and interface shear stress were determined by fracture mirror size and pulled-out fiber length measurements. For the Nicalon/C/Sic. tensile strength remained nearly constant up to 800 K, and while the strength dropped from 140 MPa at 800 K to 41 MPa at 1200 K, with weakest link failure mode. For the Hi-Nicalon/BN/SiC, the tensile strength increased slightly with increase in test tem- perature up to 1200 K; however, a large decrease in the strength was observed at 1400 K. In the case of the Hi-Nicalon/BN/ SiC, the fracture was governed by fiber bundle strength. The temperature dependence of tensile strength and fracture behavior of both omposites was attributed to change of the in situ constituent properties with temperature. C 2002 Elsevier Science Ltd. All rights Keywords: Composites; Fiber strength; Interface shear stress; Mechanical properties; Polymer infiltration pyrolysis; SiC/ SiC 1. Introduction The damage evolution of the composites usually includes two fundamental regimes: (i)at lower stresses Continuous fiber-reinforced ceramic matrix compo- matrix cracking originating from defects in the matrix sites(CFCCs) have become an important class of and followed by interface debonding between the fiber materials for structural applications at elevated tem- and matrix, and (ii) at higher stresses, occurrence of peratures because of their improved flaw tolerance, fiber damage and ultimate failure.6. 7 If matrix cracking large fracture resistance, and noncatastrophic mode of can reach the fully-saturated state prior to composite failure comparing with monolithic ceramic materials. failure, the subsequent deformation and failure are Among CFCCs, SiC fiber-reinforced SiC matrix com- dominated entirely by the fiber flaw population, show- posites have been studied extensively in recent years ing a noncatastrophic failure. 6, 7 However, if the fibers because the Sic fiber shows a potential for applications are sufficiently weak and interface bonding compara- at elevated temperatures. These studies demonstrated tively strong due to interface reaction in an oxidizing that the tensile mechanical behavior and properties of environment at high temperatures, composite failure is the composites strongly depend on the in situ fiber also possible in the regime of matrix cracking, showing strength characteristics, interface properties and matrix a catastrophic fracture similar to the fracture of a cracking stress, as well as the fabrication processes. I-5 monolithic ceramic 8,9 The transition from non- catastrophic to catastrophic fracture depends on in situ constituent properties. Thus, it is important to under 81-0298-51-3613. stand this transition to clarify tensile fracture behavior go. jp(S. Guo). of the composites and to allow correlation with the in Materials Science. I-I Namiki. Tsukuba situ constituent properties. In situ constituent properties of the composites such as fiber strength, interface shear 0955-2219/02/S. see front matter C 2002 Elsevier Science Ltd. All rights reserved. PII:S0955-2219(02)00028-6

Tensile fracture behavior of continuous SiC fiber-reinforced SiC matrix composites at elevated temperatures and correlation to in situ constituent properties Shuqi Guo*,1, Yutaka Kagawa Institute of Industrial Science, The University of Tokyo, 7-22-1, Roppongi, Minato-ku, Tokyo 106-8558, Japan Received 21 March 2001; received in revised form 20 December 2001; accepted 6 January 2002 Abstract The tensile fracture behavior and tensile mechanical properties of polymer infiltration pyrolysis (PIP)-processed two-dimensional plain-woven fabric carbon-coated NicalonTM SiC fiber and BN-coated Hi-NicalonTM SiC fiber-reinforced SiC matrix composites have been investigated. Tensile testing of the composites was carried out in air between 298 and 1400 K. In situ fiber strength and interface shear stress were determined by fracture mirror size and pulled-out fiber length measurements. For the Nicalon/C/SiC, tensile strength remained nearly constant up to 800 K, and while the strength dropped from 140 MPa at 800 K to 41 MPa at 1200 K, with weakest link failure mode. For the Hi-Nicalon/BN/SiC, the tensile strength increased slightly with increase in test tem￾perature up to 1200 K; however, a large decrease in the strength was observed at 1400 K. In the case of the Hi-Nicalon/BN/SiC, the fracture was governed by fiber bundle strength. The temperature dependence of tensile strength and fracture behavior of both composites was attributed to change of the in situ constituent properties with temperature. # 2002 Elsevier Science Ltd. All rights reserved. Keywords: Composites; Fiber strength; Interface shear stress; Mechanical properties; Polymer infiltration pyrolysis; SiC/SiC 1. Introduction Continuous fiber-reinforced ceramic matrix compo￾sites (CFCCs) have become an important class of materials for structural applications at elevated tem￾peratures because of their improved flaw tolerance, large fracture resistance, and noncatastrophic mode of failure comparing with monolithic ceramic materials. Among CFCCs, SiC fiber-reinforced SiC matrix com￾posites have been studied extensively in recent years because the SiC fiber shows a potential for applications at elevated temperatures. These studies demonstrated that the tensile mechanical behavior and properties of the composites strongly depend on the in situ fiber strength characteristics, interface properties and matrix cracking stress, as well as the fabrication processes.1
5 The damage evolution of the composites usually includes two fundamental regimes: (i) at lower stresses, matrix cracking originating from defects in the matrix and followed by interface debonding between the fiber and matrix, and (ii) at higher stresses, occurrence of fiber damage and ultimate failure.6,7 If matrix cracking can reach the fully-saturated state prior to composite failure, the subsequent deformation and failure are dominated entirely by the fiber flaw population, show￾ing a noncatastrophic failure.6,7 However, if the fibers are sufficiently weak and interface bonding compara￾tively strong due to interface reaction in an oxidizing environment at high temperatures, composite failure is also possible in the regime of matrix cracking, showing a catastrophic fracture similar to the fracture of a monolithic ceramic.8,9 The transition from non￾catastrophic to catastrophic fracture depends on in situ constituent properties. Thus, it is important to under￾stand this transition to clarify tensile fracture behavior of the composites and to allow correlation with the in situ constituent properties. In situ constituent properties of the composites such as fiber strength, interface shear 0955-2219/02/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved. PII: S0955-2219(02)00028-6 Journal of the European Ceramic Society 22 (2002) 2349–2356 www.elsevier.com/locate/jeurceramsoc * Corresponding author. Fax: +81-0298-51-3613. E-mail address: guo.shuqi@nims.go.jp (S. Guo). 1 National Institute for Materials Science, 1-1 Namiki, Tsukuba, Ibaraki 305-0044, Japan

S Guo, Y. Kagawa/Journal of the European Ceramic Society 22(2002)2349-2356 stress and matrix cracking stress, can be determined by The polycarbosilane had a melting point of 514 K,an a tensile test and fractographic analysis and measure- average molecular weight of 2410, and its chemical ments. 0,I Thus, it is possible to evaluate correlation of composition was: 60 mass% Si, 40 mass% C,<l e mechanical behavior to in situ constituent proper- mass%O. The five infiltrated pre-preg sheets were ties. Although the correlation of tensile fracture beha- stacked and pressed, and the stacked sheets were then vior to in situ constituent properties has been rep for precursor. The composite precursor was pyrolyzed at in CVI-processed composites, 3. 2 this correlation PIP-processed composites is not well known. In the 1273K in a high purity N2 atmosphere. Approximately present study, tensile testing of the two PIP-processed 40 vol. of pores existed in the pyrolyzed composite SiC fiber-reinforced SiC matrix composites was carried due to the low yield weight of the polycarbosilane. To out at room and elevated temperatures. The matrix further reduce the porosity of the composite, a multiple racking stress, in situ fiber strength and interface shear PIP process was applied. After 10 PIP cycles, the total stress were obtained. The correlation of the tensile frac- fiber volume fraction, f, and the total porosity of the ture behavior of the composites to in situ constituent composite were 0.28 and s0.09, respectively. Here- properties was discusse after. the NicalonTM Sic fiber and Hi-Nicalon M Sic fiber-reinforced SiC matrix composites fabricated are denoted as Nicalon/ C/SiC and Hi-Nicalon/BN/ SiC 2. Experimental procedure respectively 2. 1. Composite materials 2. 2. Tensile test The composite materials used in the present study The composite panels were cut into a dog-bone type were 2D plain-woven fabric SiC fiber-SiC matrix com- tensile test specimen with the long axis parallel to one of posites fabricated by the Pip process. To compare the the fiber axis directions. The shape and dimensions of effect of fiber strength and interface properties on the the specimen for monotonic tension testing are shown in tensile fracture behavior of the composites, NicalonM Fig. 1. Quasi-static monotonic tensile testing was car- and Hi-NicalonM SiC fibers(Nippon Carbon Co Ltd ried out using a servo-hydraulic MTs 808 testing system Tokyo, Japan)in which the surfaces were respectively (MTS System Co., MI, USA)with a constant crosshead coated with 0.04 um amorphous carbon and 0.4 um displacement rate of 0.5 mm/ min in ambient air at room bn by chemical vapor deposition(CVD) were used as temperature (298 K), 800, 1200 and 1400 K. Three reinforcements. The typical properties and chemical composite specimens were used for each measurement composition of the two fibers are listed in Table 1. The An electric furnace attached to the mts testing system coated SiC fibers were formed a plain-woven fabric provided the heating Axial strain was measured directly sheet, with 16x 16 numbers of fiber per inch. The fabric from the gauge length of the specimen by using a con- sheets averaging 150x 150 mm in size were cut from the tact extensometer (MTS Model 632.59, MTS System formed plane-woven sheet. The cut fabric sheets were Co., MI). Before the loading, the specimen was heated infiltrated with a polycarbosilane solution containing a fine B-SiC powder. The fine B-Sic powder had an aver- age diameter of N4 um and its addition effectively (A)Top View 90° Bundle reduced pore content in the SiC matrix after pyrolysis 4.5 0° Bundle Copper T able i Typical properties and chemical composition of the two fibers Fibres Nicalon Hi. Nicalon TM 25 25 Fibre properties Tensile strength (GPa) Youngs modulus (GPa) Elongation (B) Side View Average fibre radius Number of fibres Copper Tab Chemical compositions Si (wt%)56.6 (wt.%)31.7 (wt%) Fig. 1. Shape and dimensions of the specimen for monotonic tension testing

stress and matrix cracking stress, can be determined by a tensile test and fractographic analysis and measure￾ments.10,11 Thus, it is possible to evaluate correlation of the mechanical behavior to in situ constituent proper￾ties. Although the correlation of tensile fracture beha￾vior to in situ constituent properties has been reported in CVI-processed composites,3,12 this correlation for PIP-processed composites is not well known. In the present study, tensile testing of the two PIP-processed SiC fiber-reinforced SiC matrix composites was carried out at room and elevated temperatures. The matrix cracking stress, in situ fiber strength and interface shear stress were obtained. The correlation of the tensile frac￾ture behavior of the composites to in situ constituent properties was discussed. 2. Experimental procedure 2.1. Composite materials The composite materials used in the present study were 2D plain-woven fabric SiC fiber–SiC matrix com￾posites fabricated by the PIP process. To compare the effect of fiber strength and interface properties on the tensile fracture behavior of the composites, NicalonTM and Hi-NicalonTM SiC fibers (Nippon Carbon Co. Ltd., Tokyo, Japan) in which the surfaces were respectively coated with 0.04 mm amorphous carbon and 0.4 mm BN by chemical vapor deposition (CVD) were used as reinforcements. The typical properties and chemical composition of the two fibers are listed in Table 1.13 The coated SiC fibers were formed a plain-woven fabric sheet, with 1616 numbers of fiber per inch. The fabric sheets averaging 150150 mm in size were cut from the formed plane-woven sheet. The cut fabric sheets were infiltrated with a polycarbosilane solution containing a fine b-SiC powder. The fine b-SiC powder had an aver￾age diameter of 4 mm and its addition effectively reduced pore content in the SiC matrix after pyrolysis.4,5 The polycarbosilane had a melting point of 514 K, an average molecular weight of 2410, and its chemical composition was: 60 mass% Si, 40 mass% C, <1 mass% O. The five infiltrated pre-preg sheets were stacked and pressed, and the stacked sheets were then cured at 523 Kin ambient air to obtain the composite precursor. The composite precursor was pyrolyzed at 1273 Kin a high purity N2 atmosphere. Approximately 40 vol.% of pores existed in the pyrolyzed composite due to the low yield weight of the polycarbosilane. To further reduce the porosity of the composite, a multiple PIP process was applied. After 10 PIP cycles, the total fiber volume fraction, f, and the total porosity of the composite were 0.28 and 0.09, respectively. Here￾after, the NicalonTM SiC fiber and Hi-NicalonTM SiC fiber-reinforced SiC matrix composites fabricated are denoted as Nicalon/C/SiC and Hi-Nicalon/BN/SiC, respectively. 2.2. Tensile test The composite panels were cut into a dog-bone type tensile test specimen with the long axis parallel to one of the fiber axis directions. The shape and dimensions of the specimen for monotonic tension testing are shown in Fig. 1. Quasi-static monotonic tensile testing was car￾ried out using a servo-hydraulic MTS 808 testing system (MTS System Co., MI, USA) with a constant crosshead displacement rate of 0.5 mm/min in ambient air at room temperature (298 K), 800, 1200 and 1400 K. Three composite specimens were used for each measurement. An electric furnace attached to the MTS testing system provided the heating. Axial strain was measured directly from the gauge length of the specimen by using a con￾tact extensometer (MTS Model 632.59, MTS System Co., MI). Before the loading, the specimen was heated Table 1 Typical properties and chemical composition of the two fibers Fibres NicalonTM SiC Hi-NicalonTM SiC Fibre properties Tensile strength (GPa) 3.0 2.8 Young’s modulus (GPa) 220 270 Elongation (%) 1.4 1.0 Average fibre radius (mm) 7 7 Number of fibres per bundles 500 500 Chemical compositions Si (wt.%) 56.6 62.4 C (wt.%) 31.7 37.1 O (wt%) 11.7 0.5 Fig. 1. Shape and dimensions of the specimen for monotonic tension testing. 2350 S. Guo, Y. Kagawa / Journal of the European Ceramic Society 22 (2002) 2349–2356

S Guo, Y. Kagawa/Journal of the European Ceramic Society 22(2002)2349-2356 to the test temperature at a rate of 50C/min and then response. The large nonlinear regime is observed and held for a10 min to give a uniform temperature dis- remains nearly constant up to 800 K, however, this tribution in the specimen. After the tension testing, the regime is sharply reduced at 1200 K and nearly dis fracture surfaces of the specimens were characterized appears, showing a brittle fracture similar to that of using an optical microscope and scanning electron monolithic ceramics. Fig. 3 shows the typical tensile microscope(SEM) stress-strain curves of the Hi-Nicalon/ BN/SiC at room and elevated temperatures. The curves show a tensile 2.3. In situ fiber strength and effective interface shear stress-strain behavior similar to that of the Nicalon/C/ stress SiC: the linear deformation regime and nonlinear deformation regime are present, but the nonlinear (i) In situ fiber strength: in situ fiber strength in the regime is larger for all the test temperatures compared composite, ofu, was estimated from the fracture mirror to the Nicalon/C/SiC, showing a larger fracture resis radius, Im, of a pulled-out fiber. 0 The tensile strength, tance, especially at and above 1200 K Ofu, of the fiber is given by: The Youngs modulus of the composite, Ec, was obtained from the initial slope of these curves. The apparent matrix cracking stress, omc, was determined at m the transition point from a linear response to a non linear response, because the deviation from linear beha- where Kr is the(mode I)fracture toughness of the fiber. vior is often attributed to initiation of the matrix The fracture toughness of the SiC fiber is reported to be cracking. 18, 19 However, studies have shown that matrix A1.0 MPa m/ 14 This value is assumed to be indepen dent of the test tempe rature dependence up to 1100 K has been observed for N 800K lonM SiC fiber. 14 Assuming that the fiber failure follows the weakest-link principle, the statistical distribution of 298K fiber strength is described according to Weibull's two- 100 parameter statistical distribution theory. 5 Failure probability of the fiber, which is associated with fiber strength, is obtained using the mean rank method (ii)Interface shear stress: interface shear stress, Ti, of Sic fiber-reinforced ceramic matrix composites is expressed using an average fiber pullout length, L 1200K 0.5 Fig. 2. Typical urves of the Nicalon/ C/SiC at room and where the Rr is the radius of fiber(7 um), i(m) is a nondimensional function determined from the statistics 300 of fiber failure properties and takes a value close to Hi-Nicalon/BN/SiC 1200K unity for m>3. The interface shear stress determined represents the load transverse capacity from the fiber to he matrix, reflecting the interface sliding resistance in G200 the debonding interface. 1400K 3. Results 3.. Tensile mechanical behavior (i Tensile stress-strain curves: Fig. 2 shows the typi- cal tensile stress-strain curves of the Nicalon/ C/SiC at room aI nd elevated temperatures. All the curves exhibit Strain, E(%) a linear response near the origin and a following gra- Fig 3. Typical monotonic tensile stress-strain curves of the Hi-Nica- dual decrease of the slope up to fracture, i.e. nonlinear lon/BN/SiC at room and elevated temperatures

to the test temperature at a rate of 50 C/min and then held for 10 min to give a uniform temperature dis￾tribution in the specimen. After the tension testing, the fracture surfaces of the specimens were characterized using an optical microscope and scanning electron microscope (SEM). 2.3. In situ fiber strength and effective interface shear stress (i) In situ fiber strength: in situ fiber strength in the composite, fu, was estimated from the fracture mirror radius, rm, of a pulled-out fiber.10 The tensile strength, fu, of the fiber is given by:10 fu ¼ 3:5Kf ffiffiffiffiffi rm p ð1Þ where Kf is the (mode I) fracture toughness of the fiber. The fracture toughness of the SiC fiber is reported to be 1.0 MPa m1/2. 14 This value is assumed to be indepen￾dent of the test temperature because only a slight dependence up to 1100 Khas been observed for Nica￾lonTM SiC fiber.14 Assuming that the fiber failure follows the weakest-link principle, the statistical distribution of fiber strength is described according to Weibull’s two￾parameter statistical distribution theory.15 Failure probability of the fiber, which is associated with fiber strength, is obtained using the mean rank method. (ii) Interface shear stress: interface shear stress, i, of SiC fiber-reinforced ceramic matrix composites is expressed using an average fiber pullout length, L, as14,16 i ¼ lðmÞRffu 4L ð2Þ where the Rf is the radius of fiber ( 7 mm), l(m) is a nondimensional function determined from the statistics of fiber failure properties and takes a value close to unity for m>3.17 The interface shear stress determined represents the load transverse capacity from the fiber to the matrix, reflecting the interface sliding resistance in the debonding interface. 3. Results 3.1. Tensile mechanical behavior (i) Tensile stress–strain curves: Fig. 2 shows the typi￾cal tensile stress–strain curves of the Nicalon/C/SiC at room and elevated temperatures. All the curves exhibit a linear response near the origin and a following gra￾dual decrease of the slope up to fracture, i.e. nonlinear response. The large nonlinear regime is observed and remains nearly constant up to 800 K; however, this regime is sharply reduced at 1200 Kand nearly dis￾appears, showing a brittle fracture similar to that of monolithic ceramics. Fig. 3 shows the typical tensile stress–strain curves of the Hi-Nicalon/BN/SiC at room and elevated temperatures. The curves show a tensile stress–strain behavior similar to that of the Nicalon/C/ SiC; the linear deformation regime and nonlinear deformation regime are present, but the nonlinear regime is larger for all the test temperatures compared to the Nicalon/C/SiC, showing a larger fracture resis￾tance, especially at and above 1200 K. The Young’s modulus of the composite, Ec, was obtained from the initial slope of these curves. The apparent matrix cracking stress, mc, was determined at the transition point from a linear response to a non￾linear response, because the deviation from linear beha￾vior is often attributed to initiation of the matrix cracking.18,19 However, studies have shown that matrix Fig. 2. Typical monotonic tensile stress–strain curves of the Nicalon/ C/SiC at room and elevated temperatures. Fig. 3. Typical monotonic tensile stress–strain curves of the Hi-Nica￾lon/BN/SiC at room and elevated temperatures. S. Guo, Y. Kagawa / Journal of the European Ceramic Society 22 (2002) 2349–2356 2351

S Guo, Y. Kagawa/Journal of the European Ceramic Society 22(2002)2349-2356 cracking usually initiates well below this stress 20.21 tested at room and elevated temperatures. The pulled- Thus, in the present study, the apparent matrix cracking out fibers are observed in the composite specimens stress obtained should be higher than the true matrix fractured at and below 800 K, however, this fiber pull- cracking stress. The matrix cracking behavior clearly out behavior is not seen in the composite tested at 1200 was observed in the polished longitudinal cross-section K, showing a brittle fracture fashion. Fig. 6 shows the of the composite specimens tested at room and elevated macroscopic fracture appearances of the Hi-Nicalon temperatures(Fig 4). The transverse matrix cracks are BN SiC tested at room and elevated temperatures. Dif- formed through the transverse fiber bundle, and some of fering from the Nicalon /C/SiC, the room temperature them are arrested at the interface between the long- fracture behavior of the Hi-Nicalon/BN/SiC shows a itudinal and transverse fiber bundles fracture path which is jagged and stepped across the The average values of Ec and ome that were calculated thickness, and there is extensive fiber and fiber tow pull from the duplicate tests at each temperature are shown out. Although the fracture surface became smoother in Table 2, together with the tensile strength, OTs, and with increasing test temperature, the pulled-out fibers le strain to failure, Ec. For the Nicalon/C/SiC, the ten- are observed over all test temperatures sile mechanical properties remain nearly constant from Fig. 7 shows SEM micrographs of the fracture sur- 298 to 800 K and that they degrade sharply by 1200 K, faces of both the composites. Although all fracture sur in particular, the tensile strength and strain to failure faces of the composites tested at room and hig dropped from 140 MPa and 0. 38% at 800 K to 41 MPa temperatures showed a fibrous fracture surface, only and 0.075% at 1200 K. For the Hi-Nicalon/ BN/SiC, on few and short pulled-out fibers are observed in the the other hand, the mechanical properties such as Ec, fracture surface of the Nicalon/C/SiC tested at 1200 K ome, Ec and ars remain nearly constant up to 1200 K This is probably attributed to the silica(Sio2) formation and begin to degrade at 1400 K; they are much higher at the interface by oxidation of both the matrix and than those of the Nicalon/C/SiC at all test temperatures fiber, because of air penetration into the gaps at the interface resulting form the elimination of the C-coating (ii) Fracture surface observations: Fig. 5 shows the layer by oxidation above 700 K.,3 The pulled-out macroscopic fracture appearance of the Nicalon/ C/Sic length of fiber was measured using the method reported elsewhere. 3 For the Nicalon/C/SiC, the average pulled ut fiber length is≈50mat298K,≈70pmat800K and a20 um at 1200 K. For the Hi-Nicalon/BN/SIC. the pulled-out fiber length is a300 um at 298 K, N250 mat1200Kand≈190μmatl400K. It is clear that he pulled-out fiber length of the Hi-Nicalon/ BN/SiC is much larger than that of the Nicalon/ C/Sic at room and elevated temperatures. Moreover, for the Nicalon/ C/Sic tested at and below 800 K there are regions in all the bundles in which the fiber -fracture locations are essentially coplanar with one another compared to the Hi-Nicalon/BN/SIC. 10m 3.2. In situ fiber strength Fig 4. An example of optical photographs of polished longitudinal The in situ fiber strength characteristics of the Nica- cross-section of the two composites after monotonic tension testing. lon TM Sic fiber and the hi-Nicalon TM Sic fiber were showing matrix cracking(T=298 K, Hi-Nicalon/BN/SiC obtained using Eq (I)and the results are summarized in Table 2 Tensile experimental results of both the Nicalon/C/SiC and Hi-Nicalon/ BN/SIC Composite temperature T(K) modulus E(GPa) cracking stress me(MPa)strength oTs(MPa) failure e(%)strength oTs Measured tensile Strain to Predicted tensi Nicalon/C/SiC 58±5 65±8 136±19 0.42士 55±4 5±5 0.38士 17: 1200 33±3 4l±5 0.075±0.01512 Hi-Nicalon/BN/SIC 298 80±5 75±9 226±l1 0.84±0.12223 76士4 0±7 237士6 0.9士0.06228 0±5 197±15 0.68±0.09209

cracking usually initiates well below this stress.20,21 Thus, in the present study, the apparent matrix cracking stress obtained should be higher than the true matrix cracking stress. The matrix cracking behavior clearly was observed in the polished longitudinal cross-section of the composite specimens tested at room and elevated temperatures (Fig. 4). The transverse matrix cracks are formed through the transverse fiber bundle, and some of them are arrested at the interface between the long￾itudinal and transverse fiber bundles. The average values of Ec and mc that were calculated from the duplicate tests at each temperature are shown in Table 2, together with the tensile strength, TS, and the strain to failure, c. For the Nicalon/C/SiC, the ten￾sile mechanical properties remain nearly constant from 298 to 800 Kand that they degrade sharply by 1200 K, in particular, the tensile strength and strain to failure dropped from 140 MPa and 0.38% at 800 Kto 41 MPa and 0.075% at 1200 K. For the Hi-Nicalon/BN/SiC, on the other hand, the mechanical properties such as Ec, mc, c and TS remain nearly constant up to 1200 K and begin to degrade at 1400 K; they are much higher than those of the Nicalon/C/SiC at all test temperatures. (ii) Fracture surface observations: Fig. 5 shows the macroscopic fracture appearance of the Nicalon/C/SiC tested at room and elevated temperatures. The pulled￾out fibers are observed in the composite specimens fractured at and below 800 K, however, this fiber pull￾out behavior is not seen in the composite tested at 1200 K, showing a brittle fracture fashion. Fig. 6 shows the macroscopic fracture appearances of the Hi-Nicalon/ BN/SiC tested at room and elevated temperatures. Dif￾fering from the Nicalon/C/SiC, the room temperature fracture behavior of the Hi-Nicalon/BN/SiC shows a fracture path which is jagged and stepped across the thickness, and there is extensive fiber and fiber tow pull￾out. Although the fracture surface became smoother with increasing test temperature, the pulled-out fibers are observed over all test temperatures. Fig. 7 shows SEM micrographs of the fracture sur￾faces of both the composites. Although all fracture sur￾faces of the composites tested at room and high temperatures showed a fibrous fracture surface, only few and short pulled-out fibers are observed in the fracture surface of the Nicalon/C/SiC tested at 1200 K. This is probably attributed to the silica (SiO2) formation at the interface by oxidation of both the matrix and fiber, because of air penetration into the gaps at the interface resulting form the elimination of the C-coating layer by oxidation above 700 K.22,23 The pulled-out length of fiber was measured using the method reported elsewhere.3 For the Nicalon/C/SiC, the average pulled￾out fiber length is 50 mm at 298 K, 70 mm at 800 K and 20 mm at 1200 K. For the Hi-Nicalon/BN/SiC, the pulled-out fiber length is 300 mm at 298 K , 250 mm at 1200 Kand 190 mm at 1400 K. It is clear that the pulled-out fiber length of the Hi-Nicalon/BN/SiC is much larger than that of the Nicalon/C/SiC at room and elevated temperatures. Moreover, for the Nicalon/ C/SiC tested at and below 800 Kthere are regions in all the bundles in which the fiber-fracture locations are essentially coplanar with one another compared to the Hi-Nicalon/BN/SiC. 3.2. In situ fiber strength The in situ fiber strength characteristics of the Nica￾lonTM SiC fiber and the Hi-NicalonTM SiC fiber were obtained using Eq. (1) and the results are summarized in Fig. 4. An example of optical photographs of polished longitudinal cross-section of the two composites after monotonic tension testing, showing matrix cracking (T=298 K, Hi-Nicalon/BN/SiC). Table 2 Tensile experimental results of both the Nicalon/C/SiC and Hi-Nicalon/BN/SiC Composite materials Test temperature T (K) Young’s modulus Ec (GPa) Apparent matrix cracking stress mc (MPa) Measured tensile strength TS (MPa) Strain to failure c (%) Predicted tensile strength TS Nicalon/C/SiC 298 58 5 65 8 136 19 0.42 0.1 177 800 55 4 55 5 140 12 0.38 0.08 175 1200 49 4 33 3 41 5 0.075 0.015 127 Hi-Nicalon/BN/SiC 298 80 5 75 9 226 11 0.84 0.12 223 1200 76 4 70 7 237 6 0.9 0.06 228 1400 60 3 50 5 197 15 0.68 0.09 209 2352 S. Guo, Y. Kagawa / Journal of the European Ceramic Society 22 (2002) 2349–2356

S Guo, Y. Kagawa/Journal of the European Ceramic Society 22(2002)2349-2356 2353 罐翻 0.1mm 0.1mm (b) 0.1mm 0.1mm Fig. 5. Macroscopic fracture appearance of the Nicalon/C/SiC tested Fig. 6. Macroscopic fracture appearance of the Hi-Nicalon/BN/SiC at(a)298K,(b)800Kand(c)1200K tested at(a) 298 K,(b)1200 K and(c)1400 K. Table 3. The in situ strength of Nicalon TM SiC fiber is elevated temperature are expected for the carbon 1765 MPa at 298 K and 1705 MPa at 800 K; however, it coated NicalonTM fibers investigated in the present decreases to 1235 MPa at 1200 K. The degradation of study, especially microstructural and stoichimetric fiber strength due to exposure to high temperatures is changes of the fiber. On the other hand, the in situ already documented in the literature. 24,25 The fiber strength of Hi-Nicalon M SiC fiber remains nearly the strength decreased m 30 and 70%, respectively, after same value at the temperatures of 298 and 1200 K, and exposure at 1273 and 1573 K for 12 h in wet-air atmo- it begins to decrease above 1200 K. The fiber strength at sphere, 4 and the tensile strength drops from 2000 MPa 1400 K is lower by N11%, compared to that at room at room temperature to 1000 MPa at 1573 K in air. 5 temperature. Strength decrease of the Hi-NicalonTM This reduction is explained by the microstructural and SiC fiber at high temperature has been documented in stoichiometric changes and void formation in the fiber the literature. 26 Degradation in the strength of the fiber at higher temperatures. Similar changes, although not above the temperature of 1573 K in ambient air as severe because of a short-term heat exposure at observed and the major reason given for this was

Table 3. The in situ strength of NicalonTM SiC fiber is 1765 MPa at 298 Kand 1705 MPa at 800 K; however, it decreases to 1235 MPa at 1200 K. The degradation of fiber strength due to exposure to high temperatures is already documented in the literature.24,25 The fiber strength decreased 30 and 70%, respectively, after exposure at 1273 and 1573 Kfor 12 h in wet-air atmo￾sphere,24 and the tensile strength drops from 2000 MPa at room temperature to 1000 MPa at 1573 Kin air.25 This reduction is explained by the microstructural and stoichiometric changes and void formation in the fiber at higher temperatures. Similar changes, although not as severe because of a short-term heat exposure at elevated temperature, are expected for the carbon￾coated NicalonTM fibers investigated in the present study, especially microstructural and stoichimetric changes of the fiber. On the other hand, the in situ strength of Hi-NicalonTM SiC fiber remains nearly the same value at the temperatures of 298 and 1200 K, and it begins to decrease above 1200 K. The fiber strength at 1400 Kis lower by 11%, compared to that at room temperature. Strength decrease of the Hi-NicalonTM SiC fiber at high temperature has been documented in the literature.26 Degradation in the strength of the fiber above the temperature of 1573 Kin ambient air was observed and the major reason given for this was grain Fig. 5. Macroscopic fracture appearance of the Nicalon/C/SiC tested at (a) 298 K, (b) 800 K and (c) 1200 K. Fig. 6. Macroscopic fracture appearance of the Hi-Nicalon/BN/SiC tested at (a) 298 K, (b) 1200 K and (c) 1400 K. S. Guo, Y. Kagawa / Journal of the European Ceramic Society 22 (2002) 2349–2356 2353

S Guo, Y. Kagawa/Journal of the European Ceramic Society 22(2002)2349-2356 (b 38s 100um 100um 产100m 100m Fig. 7. Scanning electron micrographs of fracture surfaces of the NicalonM SiC/C/SiC tested at(a)298 K, and(b)1200 K and the Hi-Nicalon/BN/ Sic tested at(c)298 K, and (d)1400 K Table 3 In situ fiber tensile strength characteristics and interface shear stress of both the Nicalon/C/SiC and Hi-Nicalon/BN/SiC at room and elevated temperatures In situ fibre Weibull Fibre pullout Interfacial shear Number of materials odulus m length L (um) stress Ti(MPa) measurements 0±35 59±7 ±10 126±15 Hi-Nicalon/BN/SiC 2199 8.7 300±140 13±5 2187 250±110 15±5 9.6 90±90 17±7 growth. 26 Nearly the same temperature dependence was values at room and elevated temperatures are much observed as in the present study lower than those of the Nicalon/ C/SiC. This should be attributed to a better oxidation resistance of BN-coating 3.3. Interface shear stress on the Hi-Nicalon M fiber surface than C-coating on Nicalon fiber surface The Interface shear stresses of the Nicalon/C/SiC and the Hi-Nicalon/ BN/SiC composites are obtained using Eq(2), and the results are also summarized in Table 3. 4. Discussion For the Nicalon/ C/SiC, the interfacial shear stress is 59 MPa at 298K. 44 MPa at 800 K and 126 MPa at 1200 The experimental results indicated that the K. The interfacial shear stress decreases at 800 K from fracture behavior and properties of both the nica that of 298 K and then increases again at 1200 K. For SiC and the Hi-Nicalon/ BN/SiC depended on test tem- the Hi-Nicalon/BN/SiC, the interface shear stress tends perature. The temperature dependence essentially origi- to increase with increasing test temperature, and the nates from the change of in situ constituent properties

growth.26 Nearly the same temperature dependence was observed as in the present study. 3.3. Interface shear stress The Interface shear stresses of the Nicalon/C/SiC and the Hi-Nicalon/BN/SiC composites are obtained using Eq. (2), and the results are also summarized in Table 3. For the Nicalon/C/SiC, the interfacial shear stress is 59 MPa at 298 K, 44 MPa at 800 K and 126 MPa at 1200 K. The interfacial shear stress decreases at 800 K from that of 298 Kand then increases again at 1200 K. For the Hi-Nicalon/BN/SiC, the interface shear stress tends to increase with increasing test temperature, and the values at room and elevated temperatures are much lower than those of the Nicalon/C/SiC. This should be attributed to a better oxidation resistance of BN-coating on the Hi-NicalonTM fiber surface than C-coating on NicalonTM fiber surface.22,23 4. Discussion The experimental results indicated that the tensile fracture behavior and properties of both the Nicalon/C/ SiC and the Hi-Nicalon/BN/SiC depended on test tem￾perature. The temperature dependence essentially origi￾nates from the change of in situ constituent properties Table 3 In situ fiber tensile strength characteristics and interface shear stress of both the Nicalon/C/SiC and Hi-Nicalon/BN/SiC at room and elevated temperatures Composite materials Test temperature T (K) In situ fibre strength fu (MPa) Weibull modulus m Fibre pullout length L (mm) Interfacial shear stress i (MPa) Number of measurements Nicalon/C/SiC 298 1765 6.2 50 35 59 7 70 800 1705 6.7 70 40 44 10 70 1200 1235 6.9 20 10 126 15 70 Hi-Nicalon/BN/SiC 298 2199 8.7 300 140 13 5 70 1200 2187 9.1 250 110 15 5 70 1400 1961 9.6 190 90 17 7 70 Fig. 7. Scanning electron micrographs of fracture surfaces of the NicalonTM SiC/C/SiC tested at (a) 298 K, and (b) 1200 K and the Hi-Nicalon/BN/ SiC tested at (c) 298 K, and (d) 1400 K. 2354 S. Guo, Y. Kagawa / Journal of the European Ceramic Society 22 (2002) 2349–2356

S Guo, Y. Kagawa/Journal of the European Ceramic Society 22(2002)2349-2356 with test temperature. Assuming that the fiber stress is uniform and the fiber failure is non-interactive. the ten- sile strength of the composites, aTS, is approximatel given by f iou Noncatastrophic Fracture WLF OH (3b) 02}B where obu is the tensile strength of fiber bundle, fi is the 口1200K◆1400K volume fraction of longitudinal fiber bundle (L f/2) ofu is in situ fiber strength, m is the Weibull modulus of fiber, e is the base of the natural logarithm and r is the Fig. 8. Correlation of tensile strength of the composites normalized by The predicted values are listed in Table 2 to compare the predicted values to in situ fiber strength and interface shear stress with the measured ones. For the Hi-Nicalon/BN/SiC, (BF: brittle fracture, solid symbols: Hi-Nicalon/BN/SiC, dotted sym. the predicted values nearly coincide with the experi- mental results for all the test temperatures. The con- sistency indicates that the tensile strength is governed by the fiber bundle strength and ultimate fracture strength fiber properties and that the full potential of the fibers is is lower than that predicted by fiber bundle model utilized in the composite. This means that the matrix Therefore, although the Nicalon/ C/SiC exhibits a non cracking reached a full-saturated state prior to failure, linear/noncatastrophic behavior in tension, this compo- showing a noncatastrophic fracture mode(Figs. 3, 6 site failed prematurely as compared to the Hi-Nicalon/ and 7) that belongs to a global load sharing(GLS) BN/ SiC(Figs. 2 and 3) condition resulting from the lower interface shear stress Fig. 8 shows correlation of tensile strength and frac- (Table 3). On the contrary, for the Nicalon/C/SiC, the ture behavior of the both composites to fiber strength predicted values are higher than the measured values; in and interface shear stress. It is found that the curve particular, at 1200 K this discrepancy is considerably could be divided into two characteristic regimes: (1) noticeable. Although the fibrous fracture surfaces are of/Ti> 100, tensile strength of the composite coincides observed [Fig. 7(a) and(b)], there is only a modest with the values predicted by fiber strength with a non- amount of fiber pulled-out in the composite, especially catastrophic fracture under GLs condition, and (ii) at 1200 K where the brittle fracture surface is found In f/Ti 30), the tensile strength of composite another, indicating a strong correlation between breaks. 80% of predicted value(Table 2) with a weakest link This strong correlation probably results from local failure (WLF) mode; in particular as ofu/Ti is lower strong fiber/matrix bonding. Previous studies revealed (typical of/Ti< 10), the tensile strength is only 30% of that the fiber/matrix debonding in CVI-processed Nica- the predicted value(Table 2) with a brittle fracture lonBN/C/SiC is controlled by the thin silica layer fashion. This suggests that the full potential of the fibers resulting from the changes that have occurred near the is utilized only when the ratio of the interface shear surface of the metastable Nicalon SiC fiber during pro- stress to the fiber strength is below a critical value. In the cessing 28.29 Although in the present study the Nicalon/ present work, the critical ratio is approximately 1/100 C/SiC was processed by using non-CVI process but PIP process, this thin silica layer is excepted to be present at the interface, in particular, for the composite sp ecimens 5. Conclusions tested above 700 K oxidation of the C-coating pro- moted formation of silica layer resulting from oxidation of both the matrix and fiber. The interface gener- 1. For the Nicalon/C/SiC, Youngs modulus, ten ated by the silica layer might be the weakest link in the sile strength and strain to failure remained nearly interfacial phase sequence. This silica layer formed dur- constant up to 800 K and significantly then ing processing or/and oxidation generally is dis- decreased at 1200 K. A noncatastrophic fracture continuities and may be regarded as surface flaws which was observed at and below 800K. while a brittle locally weaken the fibers. 28,29 In the case of the weakest fracture fashion similar to monolithic ceramics link, fracture of composite is no longer dominated by occurred at 1200 K

with test temperature. Assuming that the fiber stress is uniform and the fiber failure is non-interactive, the ten￾sile strength of the composites, TS, is approximately given by27 TS ¼ fLbu ð3aÞ bu ¼ 1 me 1=m fu 1 þ 1 m ð3bÞ where bu is the tensile strength of fiber bundle, fL is the volume fraction of longitudinal fiber bundle (fLf/2), fu is in situ fiber strength, m is the Weibull modulus of fiber, e is the base of the natural logarithm and is the gamma function. The predicted values are listed in Table 2 to compare with the measured ones. For the Hi-Nicalon/BN/SiC, the predicted values nearly coincide with the experi￾mental results for all the test temperatures. The con￾sistency indicates that the tensile strength is governed by fiber properties and that the full potential of the fibers is utilized in the composite. This means that the matrix cracking reached a full-saturated state prior to failure, showing a noncatastrophic fracture mode (Figs. 3, 6 and 7) that belongs to a global load sharing (GLS) condition resulting from the lower interface shear stress (Table 3). On the contrary, for the Nicalon/C/SiC, the predicted values are higher than the measured values; in particular, at 1200 Kthis discrepancy is considerably noticeable. Although the fibrous fracture surfaces are observed [Fig. 7(a) and (b)], there is only a modest amount of fiber pulled-out in the composite, especially at 1200 Kwhere the brittle fracture surface is found. In addition, there are regions in all the bundles in which the fiber-fracture locations are essentially coplanar with another, indicating a strong correlation between breaks. This strong correlation probably results from local strong fiber/matrix bonding. Previous studies revealed that the fiber/matrix debonding in CVI-processed Nica￾lon/BN/C/SiC is controlled by the thin silica layer resulting from the changes that have occurred near the surface of the metastable Nicalon SiC fiber during pro￾cessing.28,29 Although in the present study the Nicalon/ C/SiC was processed by using non-CVI process but PIP process, this thin silica layer is excepted to be present at the interface, in particular, for the composite specimens tested above 700 Koxidation of the C-coating pro￾moted formation of silica layer resulting from oxidation of both the matrix and fiber.22,23 The interface gener￾ated by the silica layer might be the weakest link in the interfacial phase sequence. This silica layer formed dur￾ing processing or/and oxidation generally is dis￾continuities and may be regarded as surface flaws which locally weaken the fibers.28,29 In the case of the weakest link, fracture of composite is no longer dominated by the fiber bundle strength and ultimate fracture strength is lower than that predicted by fiber bundle model. Therefore, although the Nicalon/C/SiC exhibits a non￾linear/noncatastrophic behavior in tension, this compo￾site failed prematurely as compared to the Hi-Nicalon/ BN/SiC (Figs. 2 and 3). Fig. 8 shows correlation of tensile strength and frac￾ture behavior of the both composites to fiber strength and interface shear stress. It is found that the curve could be divided into two characteristic regimes: (i) fu/Ti5100, tensile strength of the composite coincides with the values predicted by fiber strength with a non￾catastrophic fracture under GLS condition, and (ii) fu/Ti<100, tensile strengths of the composite are lower than the predicted values; typically, as fu/Ti is higher (typical fu/Ti530), the tensile strength of composite is 80% of predicted value (Table 2) with a weakest link failure (WLF) mode; in particular as fu/Ti is lower (typical fu/Ti410), the tensile strength is only 30% of the predicted value (Table 2) with a brittle fracture fashion. This suggests that the full potential of the fibers is utilized only when the ratio of the interface shear stress to the fiber strength is below a critical value. In the present work, the critical ratio is approximately 1/100. 5. Conclusions 1. For the Nicalon/C/SiC, Young’s modulus, ten￾sile strength and strain to failure remained nearly constant up to 800 Kand significantly then decreased at 1200 K. A noncatastrophic fracture was observed at and below 800 K, while a brittle fracture fashion similar to monolithic ceramics occurred at 1200 K. Fig. 8. Correlation of tensile strength of the composites normalized by the predicted values to in situ fiber strength and interface shear stress (BF: brittle fracture, solid symbols: Hi-Nicalon/BN/SiC, dotted sym￾bols: Nicalon/C/SiC). S. Guo, Y. Kagawa / Journal of the European Ceramic Society 22 (2002) 2349–2356 2355

S Guo, Y. Kagawa/Journal of the European Ceramic Society 22(2002)2349-2356 2. For the Hi-Nicalon/ BN/SiC, tensile strength and at elevated temperatures. J. Am. Ceram Soc., 1995. 78(2), 388- strain to failure increased slightly with increasing 394 test temperature from 298 to 1200 K and then 10. Rice. R. W Treatise on Materials Science and Technology, vol decreased at 1400 K, while Youngs modulus Il, p. 199. Academic Press, New York, 1978 I1. Thouless, M. D. and Evans. A. G. Effects of pullout on the decreased with increasing test temperature. A echanical properties of cera matrIx noncatastrophic fracture was observed for all the Metall,1988,36.517-522 test temperature 12. Heredia, F. E, Spearing, S. M. Evans, A. G, Mosher, P and 3. In situ constituent properties of both composites Curtin, W.A. Mechanical properties of continuous.fiber-rein- were determined by tensile stress-strain curves forced carbon matrix composites and relationship to constituent properties. J. A. Ceram Soc., 1992. 75(11), 3017-302 and fracture surface measurements. The con- 3. Takeda. M. Sakamoto. J. Saeki. A. and Ichikawa. H. Mechan. stituent properties of Hi-Nicalon/BN/SiC were ical and structural analysis of silicon carbide fiber Hi-nicalon type superior to those of Nicalon/C/SiC, especially at S. Ceram. Eng. Sci. Proc., 1996. 17(4-5), 35-42. ≥1200K Thouless. M. D. Sbaizero. O. Sigl. L. S. and Evans. A. G 4. The tensile strengths of the Hi-Nicalon/BN/SiC Effect of interface mechanical properties on pullout in a Sic- fiber-reinforced lithium aluminum silicate glass-ceramic. J.Amm. were governed by the fibers and could be pre- Ceran.Soc.,1989,72(4),525-532. dicted using fiber bundle model on base of in situ 15. Weibull, W, A statistical distribution function of wide applic- fiber strength. On the contrary, the fracture of ability. J. Appl. Mech,19518.293-298 the Nicalon/ C/SiC was not be dominated by the 16. Thouless, M. D. and Evans, A. G. Effect of pull-out on the fiber bundle, with weakest link failure mode echanical properties of ceramic matrix composites. Acta Metall,1988,36,517-522 17. Beyerle, D. S, Spearing, S M, Zok, F. w. and Evans. A. G, Damage and failure in unidirectional ceramic-matrix composites. Acknowledgements J.Am. Ceran.Soc.,l992,75,27192725 18. Singh.R.N. Influence of interface shear stress on first-matrix The authors thank mr Fujikura and cracking stress in ceramic-matrix composites. J. Am. Ceram. Soc. Tanaka, Ultra-High Temperature Materials Research Center Co, Ltd, Japan, for their help in the tensile test. Marshall. d. B. and brennan. J. J. Effect of interfaces on the The first author(SQ G. would also like to thank the properties of fiber-reinforced ceramics. J. Am. Ceram. Soc., 1990 Japan Society for the Promotion of Science 73(6).1691-169 financial support of his research in Japan car. P. G. an noduli reductions in nicalon-CAs composites under static fat ue and cyclic fatigue. J. Am. Ceram. Soc., 1993, 73, 1720-1728 21. Kim.R.Y and Pagano, N.J. Crack initiation in unidirectional References rittle-matrix composites J. Am. Ceram. Soc., 1991, 74(5), 1082- 1. Morscher, G. N, Tensile stress rupture of SiCe/SiCm mini- 22. Llorca, J, Elices, M. and Celemin, J. A, Toughness and micro- composites with carbon and boron nitride interphases at elevated ructural degradation at high temperature in SiC fiber-reinforced temperatures in air.J.Am. Ceram Soc., 1997, 80(8). 2029-2042. ceramics. Acta Mater. 1998, 46(7). 2441-2453. 2. Lipetzky, P, Dvorak, G.J. and Stoloff, N.S., Tensile properties 23. Ogbuji, L U.J. T, A pervasive mode of oxidative degradation in a of a SiCSiC composite. Mater. Sci. Eng, 1996, A216, 11-19. Sic-SiC composite, 1998, 81(11). 2777-2784 3. Singh, D, Singh, J. P and Wheeler. M. J, Mechanical behavior 24. Clark, T.J., Arons, R. M. and Stamatoff, J. B, Thermal degra- of Sic(f/SiC composites and correlation to in situ fiber strength dation of nicalon SiC fibers. Cera. Eng. Sci. Proc., 1985. 6(7-8). at room and elevated temperatures. J. Am. Ceram. Soc., 1996. 576-588 793),591-596 K. C, Hodder. R. S and Tressler, R. E, 4. Shin, D. w. and Tanaka, T, Low-temperature processing of Strengths of fibers at elevated temperatures. J. Am 2),284-288 Soc.,1994,77(1),97-104 26. Hollon, G, Pailler, R, Naslain, R. and Olry, P. composite mposition and mechanical behavior at high temperat ygen-free Hi-Nicalon fiber. In High-Temperature re of uhe latrix Composites I1, Vol 58. ed. A. G. Evans and R. Naslain. J. Am. Ceram. Soc., 1999, 82(6). 1579- Ceram. Trans., The American Ceramic Society, 1995, pp. 299- 6. Lissart, N. and Lamon, J, Damage and failure in ceramic matrix minicomposites: experimental study and model. Acta Mater, 27. Coleman, B. D, On the strength of classical fibers, fiberbundles 1997,45(3),1025-10 Mech. Phys. Solids. 1958.7.60-70 properties of ceramic. 28. Naslain. R, Dugne, O, Guette, A, Sevely, J, Robin-Brosse, C, matrix composites. J.Am. Ceram Soc., 1991, 74(11), 2837-28 Rocher, J. P and Cotteret, J, Boron nitride interphase in ceramic 8. Steyer, T. E, Zok, F. w. and Wall, D. P, Stress rupture of an matrix composites. J. Am. Ceram. Soc., 1991, 740(10), 2482-2488 chanced nicalon/ silicon carbide composite at intermediate tem- : 9. Prouhet. S. Camus, G, Labrugere. C, Guette, A. and Martin peratures.J.Am. Ceram. (8),2140-2146 E, Mechanical characterization of Si-C(O) fiber/SiC (CVn 9. Xu. H. H.K. Braun, L. M., Ostertag. C. P, Krause. R. F. and matrix composites with a BN-interphase J. Am. Ceram. Soc Lloyd, l.K. Failure modes of SiC-fiber/Si, Na matrix composites 1994,77(3),649-656

2. For the Hi-Nicalon/BN/SiC, tensile strength and strain to failure increased slightly with increasing test temperature from 298 to 1200 Kand then decreased at 1400 K, while Young’s modulus decreased with increasing test temperature. A noncatastrophic fracture was observed for all the test temperatures. 3. In situ constituent properties of both composites were determined by tensile stress-strain curves and fracture surface measurements. The con￾stituent properties of Hi-Nicalon/BN/SiC were superior to those of Nicalon/C/SiC, especially at 51200 K. 4. The tensile strengths of the Hi-Nicalon/BN/SiC were governed by the fibers and could be pre￾dicted using fiber bundle model on base of in situ fiber strength. On the contrary, the fracture of the Nicalon/C/SiC was not be dominated by the fiber bundle, with weakest link failure mode. Acknowledgements The authors thank Mr. M. Fujikura and Dr. R. Tanaka, Ultra-High Temperature Materials Research Center Co., Ltd., Japan, for their help in the tensile test. The first author (S.Q.G.) would also like to thank the Japan Society for the Promotion of Science for its financial support of his research in Japan. References 1. Morscher, G. N., Tensile stress rupture of SiCf/SiCm mini￾composites with carbon and boron nitride interphases at elevated temperatures in air. J. Am. Ceram. Soc., 1997, 80(8), 2029–2042. 2. Lipetzky, P., Dvorak, G. J. and Stoloff, N. S., Tensile properties of a SiCf/SiC composite. Mater. Sci. Eng., 1996, A216, 11–19. 3. Singh, D., Singh, J. P. and Wheeler, M. J., Mechanical behavior of SiC(f)/SiC composites and correlation to in situ fiber strength at room and elevated temperatures. J. Am. Ceram. Soc., 1996, 79(3), 591–596. 4. Shin, D. W. and Tanaka, T., Low-temperature processing of ceramic woven fabric/ceramic matrix composites. J. Am. Ceram. Soc., 1994, 77(1), 97–104. 5. Takeda, M., Kagawa, Y., Mitsuno, S., Imai, Y. and Ichikawa, H., Strength of Hi-NicalonTM/silicon-carbide-matrix composite fabricated by the multiple polymer infiltration-pyrolysis process. J. Am. Ceram. Soc., 1999, 82(6), 1579–1581. 6. Lissart, N. and Lamon, J., Damage and failure in ceramic matrix minicomposites: experimental study and model. Acta Mater., 1997, 45(3), 1025–1044. 7. Curtin, W. A., Theory of mechanical properties of ceramic￾matrix composites. J. Am. Ceram. Soc., 1991, 74(11), 2837–2845. 8. Steyer, T. E., Zok, F. W. and Wall, D. P., Stress rupture of an enhanced nicalon/silicon carbide composite at intermediate tem￾peratures. J. Am. Ceram. Soc., 1998, 81(8), 2140–2146. 9. Xu, H. H. K., Braun, L. M., Ostertag, C. P., Krause, R. F. and Lloyd, I. K., Failure modes of SiC-fiber/Si3N4-matrix composites at elevated temperatures. J. Am. Ceram. Soc., 1995, 78(2), 388– 394. 10. Rice, R. W., Treatise on Materials Science and Technology, Vol. II, p. 199. Academic Press, New York, 1978. 11. Thouless, M. D. and Evans, A. G., Effects of pullout on the mechanical properties of ceramic matrix composites. Acta Metall., 1988, 36, 517–522. 12. Heredia, F. E., Spearing, S. M., Evans, A. G., Mosher, P. and Curtin, W. A., Mechanical properties of continuous-fiber-rein￾forced carbon matrix composites and relationship to constituent properties. J. Am. Ceram. Soc., 1992, 75(11), 3017–3025. 13. Takeda, M., Sakamoto, J., Saeki, A. and Ichikawa, H., Mechan￾ical and structural analysis of silicon carbide fiber Hi-nicalon type S. Ceram. Eng. Sci. Proc., 1996, 17(4–5), 35–42. 14. Thouless, M. D., Sbaizero, O., Sigl, L. S. and Evans, A. G., Effect of interface mechanical properties on pullout in a SiC- fiber-reinforced lithium aluminum silicate glass-ceramic. J. Am. Ceram. Soc., 1989, 72(4), 525–532. 15. Weibull, W., A statistical distribution function of wide applic￾ability. J. Appl. Mech., 1951, 18, 293–298. 16. Thouless, M. D. and Evans, A. G., Effect of pull-out on the mechanical properties of ceramic matrix composites. Acta Metall., 1988, 36, 517–522. 17. Beyerle, D. S., Spearing, S. M., Zok, F. W. and Evans, A. G., Damage and failure in unidirectional ceramic-matrix composites. J. Am. Ceram. Soc., 1992, 75, 2719–2725. 18. Singh, R. N., Influence of interface shear stress on first-matrix cracking stress in ceramic-matrix composites. J. Am. Ceram. Soc., 1990, 73(10), 2930–2937. 19. Cao, H. C., Bischoff, E., Sbaizero, O., Ruhle, M., Evans, A. G., Marshall, D. B. and Brennan, J. J., Effect of interfaces on the properties of fiber-reinforced ceramics. J. Am. Ceram. Soc., 1990, 73(6), 1691–1699. 20. Karandikar, P. G. and Chou, T. W., Damage development and moduli reductions in nicalon-CAS composites under static fati￾gue and cyclic fatigue. J. Am. Ceram. Soc., 1993, 73, 1720–1728. 21. Kim, R. Y. and Pagano, N. J., Crack initiation in unidirectional brittle-matrix composites. J. Am. Ceram. Soc., 1991, 74(5), 1082– 1090. 22. Llorca, J., Elices, M. and Celemin, J. A., Toughness and micro￾structural degradation at high temperature in SiC fiber-reinforced ceramics. Acta Mater., 1998, 46(7), 2441–2453. 23. Ogbuji, L. U. J. T., A pervasive mode of oxidative degradation in a SiC–SiC composite, 1998, 81(11), 2777–2784. 24. Clark, T. J., Arons, R. M. and Stamatoff, J. B., Thermal degra￾dation of nicalon SiC fibers. Ceram. Eng. Sci. Proc., 1985, 6(7–8), 576–588. 25. Pysher, D. J., Goretta, K. C., Hodder, R. S. and Tressler, R. E., Strengths of ceramic fibers at elevated temperatures. J. Am. Ceram. Soc., 1989, 72(2), 284–288. 26. Chollon, G., Pailler, R., Naslain, R. and Olry, P., Structure, composition and mechanical behavior at high temperature of the oxygen-free Hi-Nicalon fiber. In High-Temperature Ceramic￾Matrix Composites II, Vol. 58. ed. A. G. Evans and R. Naslain. Ceram. Trans., The American Ceramic Society, 1995, pp. 299– 304. 27. Coleman, B. D., On the strength of classical fibers, fiberbundles. J. Mech. Phys. Solids, 1958, 7, 60–70. 28. Naslain, R., Dugne, O., Guette, A., Sevely, J., Robin-Brosse, C., Rocher, J. P. and Cotteret, J., Boron nitride interphase in ceramic matrix composites. J. Am. Ceram. Soc., 1991, 74(10), 2482–2488. 29. Prouhet, S., Camus, G., Labrugere, C., Guette, A. and Martin, E., Mechanical characterization of Si-C(O) fiber/SiC (CVI) matrix composites with a BN-interphase. J. Am. Ceram. Soc., 1994, 77(3), 649–656. 2356 S. Guo, Y. Kagawa / Journal of the European Ceramic Society 22 (2002) 2349–2356