Availableonlineatwww.sciencedirect.com SCIENCE DIRECTo composites Part A: applied science ELSEVIER Composites: Part A 35(2004)33-40 www.elsevier.com/locate/composites Degradation of Sic/Sic composite due to exposure at high temperatures in vacuum in comparison with that in air S. Ochiai,, S. Kimura, H. Tanaka, M. Tanaka, M. Hojo, K. Morishita, H Okuda H Nakayama, M. Tamura, K Shibata, M. Sato "International Innovation Center, Kyoto University, Sakyo-ku, Kyoto 606-8501, Japan Graduate School of Engineering, Kyoto University, Sakyo kae 60-8501, Japan Japan Ultra-high Temperature Materials Research institute, 3-1-8 Higashimachi, Tajimi City, Gifit 507-0801 dapan Ultra-high Temperature Materials Research Institute, 573-3 Okiube, Ube ciry, Yamaguchi 755-000 Received 25 April 2003: revised 1 August 2003: accepted 5 September 2003 Abstract Room temperature residual strength of the SiC/Sic composite exposed in vacuum at high temperatures(823-1673 K) was studied and compared with that exposed in air. The vacuum-exposed composite showed only the fiber-pullout type fracture, and the pullout length increased with increasing exposure temperature and time, while the fractured mode of the air-exposed one changed with progressing oxidation; from the fiber-pullout type to the nonfiber-pullout one characterized by the overall fracture perpendicular to the tensile axi without fiber-pullout. The reduction in residual strength in the case of vacuum exposure was attributed mainly to the extension of the decomposition-induced defects on the fiber surface into fiber, while that in the case of air exposure mainly to the extension of the crack made by premature fracture of the Sio2 layer into the fiber. A simple model based on the kinetics of the growth of the defects and fracture exposure, which could describe the experimental resule posite strength as a function of exposure temperature and time for the vacuum mechanics was presented to describe the variation of co C 2004 Elsevier ltd. all rights reserved Keywords: A Ceramic-matrix composites(CMCs); B Environmental degradation; B Strength; C Damage mechanics 1. Introduction strength of the composite could be attributed to the reduction in fiber strength due to the extension In the preceding paper [1], the degradation due to the crack made by the premature fracture of the SiOz layer into exposure in air at high temperatures of the Sic/Sic fiber. (3)A simple model based on the kinetics of the growth composed of Si-Zr-C-O fiber(ZMI fiber with a describe the variation of composite strength as a function of composition of Sizr<ooI C14400.32, produced by Ube exposure temperature and time, which could describe the Companies) and ZrSiO4 particles (30 mass%)-dispersec experimental results The main results are summarized In the present work, the degradation behavior due to the follows.(1)With increasing exposure temperature and time, high temperature exposure in vacuum was studied by using (a)the residual strength of the exposed composite decreased the same specimens in order to compare the result for air and(b)the fracture mode changed from the fiber-pullout exposure and to try to describe the reduction as a function of type to the nonfiber-pullout one characterized by the overall exposure temperature and time by modeling also for the fracture perpendicular to the tensile axis without fiber- vacuum environment. Concerning the thermal stability of the pullout. (2)The main reason for the reduction in residual SiC fibers made from the precursors such as polycarbosilane and polytititanocarbosilane, extensive studies have been Corresponding author. Tel : +81-75-753-4834; fax:+81-75-753-4841 carried out [2-9]. Among them, Simao et al. [6 demon E-imail address: ochiai @iic. kyoto-u ac jp(S Ochiai) strated that the decomposition leads to the crystallization of 1359-835X/S- see front matter 2004 Elsevier Ltd. All rights reserved. doi: 10. 1016/j-compositesa. 2003.09.006
Degradation of SiC/SiC composite due to exposure at high temperatures in vacuum in comparison with that in air S. Ochiaia,*, S. Kimurab , H. Tanakab , M. Tanakab , M. Hojob , K. Morishitab , H. Okudaa , H. Nakayamac , M. Tamurad , K. Shibatad , M. Satoe a International Innovation Center, Kyoto University, Sakyo-ku, Kyoto 606-8501, Japan b Graduate School of Engineering, Kyoto University, Sakyo-ku, Kyoto 606-8501, Japan c Japan Ultra-high Temperature Materials Research Institute, 3-1-8 Higashimachi, Tajimi City, Gifu 507-0801, Japan d Japan Ultra-high Temperature Materials Research Institute, 573-3 Okiube, Ube city, Yamaguchi 755-0001, Japan e Ube Research Laboratory, Ube Industries Ltd, 1978-5 Kogushi, Ube city, Yamaguchi 755-8633, Japan Received 25 April 2003; revised 1 August 2003; accepted 5 September 2003 Abstract Room temperature residual strength of the SiC/SiC composite exposed in vacuum at high temperatures (823–1673 K) was studied and compared with that exposed in air. The vacuum-exposed composite showed only the fiber-pullout type fracture, and the pullout length increased with increasing exposure temperature and time, while the fractured mode of the air-exposed one changed with progressing oxidation; from the fiber-pullout type to the nonfiber-pullout one characterized by the overall fracture perpendicular to the tensile axis without fiber-pullout. The reduction in residual strength in the case of vacuum exposure was attributed mainly to the extension of the decomposition-induced defects on the fiber surface into fiber, while that in the case of air exposure mainly to the extension of the crack made by premature fracture of the SiO2 layer into the fiber. A simple model based on the kinetics of the growth of the defects and fracture mechanics was presented to describe the variation of composite strength as a function of exposure temperature and time for the vacuum exposure, which could describe the experimental results. q 2004 Elsevier Ltd. All rights reserved. Keywords: A. Ceramic–matrix composites (CMCs); B. Environmental degradation; B. Strength; C. Damage mechanics 1. Introduction In the preceding paper [1], the degradation due to the exposure in air at high temperatures of the SiC/SiC composite was studied using the composite specimens composed of Si–Zr–C–O fiber (ZMI fiber with a composition of SiZr,0.01C1.44O0.32, produced by Ube Companies) and ZrSiO4 particles (30 mass%)–dispersed Si–Zr–C matrix. The main results are summarized as follows. (1) With increasing exposure temperature and time, (a) the residual strength of the exposed composite decreased and (b) the fracture mode changed from the fiber-pullout type to the nonfiber-pullout one characterized by the overall fracture perpendicular to the tensile axis without fiberpullout. (2) The main reason for the reduction in residual strength of the composite could be attributed to the reduction in fiber strength due to the extension of the crack made by the premature fracture of the SiO2 layer into fiber. (3) A simple model based on the kinetics of the growth of the SiO2 layer and fracture mechanics was presented to describe the variation of composite strength as a function of exposure temperature and time, which could describe the experimental results. In the present work, the degradation behavior due to the high temperature exposure in vacuum was studied by using the same specimens in order to compare the result for air exposure and to try to describe the reduction as a function of exposure temperature and time by modeling also for the vacuum environment. Concerning the thermal stability of the SiC fibers made from the precursors such as polycarbosilane and polytititanocarbosilane, extensive studies have been carried out [2–9]. Among them, Simoo et al. [6] demonstrated that the decomposition leads to the crystallization of 1359-835X/$ - see front matter q 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.compositesa.2003.09.006 Composites: Part A 35 (2004) 33–40 www.elsevier.com/locate/compositesa * Corresponding author. Tel.: þ81-75-753-4834; fax: þ81-75-753-4841. E-mail address: ochiai@iic.kyoto-u.ac.jp (S. Ochiai)
S Ochiai et al./ Composites: Part A 35(2004)33-40 amorphous fiber into B-SiC, involving the generation of both the vacuum and air exposures Tensile test was carried out at Sio and Co gases even in Ar atmosphere [6]. This suggests room temperature at a crosshead speed of 8.3 X 10 m/s t the decomposition in vacuum will be enhanced due to for a gage length 50 mm. The fracture surface of the asier generation of the gases in comparison with that in A composite was observed with scanning electron microscope According to Chollon et al. [7]. the decomposition takes (SEM (Jeol, JSM-541OLS)) place at the fiber surface, resulting in a silicon depleted layer and Sic crystal growth, and these features are thought to include the formation of surface flaws of large size and thus a 3. Results and discussion decrease in fiber strel If the surface defects are the strength-determining factor 3. 1. Residual strength of vacuum-exposed composite for the fiber exposed in vacuum at high temperatures fracture mechanical approach can be applied for description Fig. I shows the weight change due to the exposure in of the fiber strength. It is expected that the fracture vacuum and air For 973 K-exposure, the weight change was mechanical model, proposed in the preceding paper [1] to minor in both environments. With increasing exposure describe the residual strength of the air-exposed composite, temperature, the weight tends to increase in the case of can be employed with a slight modification for description exposure. On contrary, the weight tends to decrease in of the residual strength of the vacuum-exposed composite In this work. first the residual strength and fracture mode of he composite exposed in vacuum at high temperatures will be studied experimentally to reveal the differences and 8 similarities of the degradation behavior between the air and vacuum exposures. Then a model will be presented to describe the variation of the strength of the vacuum-exposed 5 function of by modifying the former model [1] and will be applied to the experimental results 2. Experimental procedure Exposure time, t(s) The same SiC/SiC mini-composite specimens used for the air exposure test [1] were used for the present vacuum exposure test. As the fiber, Si-Zr-C-O fiber(ZMI fiber with a composition of Sizr<o0r C14400.32) modified by a pecial treatment was used. The modifie ed fiber had the 1473K following microstructural features [10]: (a) the top surface is a carbon layer with a thickness of a few nanometers and (b) the carbon concentration is graded in the circumferential o Vacuum region within 20-30 nm from the top surface. For the matrix, the same precursor as the ZMI fiber with dispersed ZrSiO4 particles was used. The specimens were fabricated Exposure time, t(s) by the polymer impregnation and pyrolysis(PIP)method at Jbe Companies [1, 11]. The fiber volume fraction and cross sectional area of thus fabricated composite specimens were 0.45 and 0.17 mm- on an average, respectively 1673K 1573and1673Kfor3.6×102-36×105 s in a vacuum O Vacuum chamber, which was evacuated continuously by a vacuum 5 pump to the residual pressure of around 2.7x 10 Pa. All specimens were heated at a constant rate of 0.67 K/s up to the prescribed temperature, and after the exposure for the prescribed time, they were cooled down to room tempera- ture at the same rate of 0.67 K/s as well as those for air Exposure time, t(s) The weight change due to the exposure was measured to Fig. 1. Weight change of the composite due to the exposure in vacuum and monitor the difference in degradation process between in air at 973, 1473 and 1673 K
amorphous fiber into b-SiC, involving the generation of both SiO and CO gases even in Ar atmosphere [6]. This suggests that the decomposition in vacuum will be enhanced due to easier generation of the gases in comparison with that in Ar. According to Chollon et al. [7], the decomposition takes place at the fiber surface, resulting in a silicon depleted layer and SiC crystal growth, and these features are thought to include the formation of surface flaws of large size and thus a decrease in fiber strength. If the surface defects are the strength-determining factor for the fiber exposed in vacuum at high temperatures, fracture mechanical approach can be applied for description of the fiber strength. It is expected that the fracture mechanical model, proposed in the preceding paper [1] to describe the residual strength of the air-exposed composite, can be employed with a slight modification for description of the residual strength of the vacuum-exposed composite. In this work, first the residual strength and fracture mode of the composite exposed in vacuum at high temperatures will be studied experimentally to reveal the differences and similarities of the degradation behavior between the air and vacuum exposures. Then a model will be presented to describe the variation of the strength of the vacuum-exposed composite as a function of exposure temperature and time by modifying the former model [1] and will be applied to the experimental results. 2. Experimental procedure The same SiC/SiC mini-composite specimens used for the air exposure test [1] were used for the present vacuum exposure test. As the fiber, Si–Zr–C–O fiber (ZMI fiber with a composition of SiZr,0.01C1.44O0.32) modified by a special treatment was used. The modified fiber had the following microstructural features [10]; (a) the top surface is a carbon layer with a thickness of a few nanometers and (b) the carbon concentration is graded in the circumferential region within 20–30 nm from the top surface. For the matrix, the same precursor as the ZMI fiber with dispersed ZrSiO4 particles was used. The specimens were fabricated by the polymer impregnation and pyrolysis (PIP) method at Ube Companies [1,11]. The fiber volume fraction and crosssectional area of thus fabricated composite specimens were 0.45 and 0.17 mm2 on an average, respectively. The specimens were exposed at 973, 1123, 1273, 1473, 1573 and 1673 K for 3.6 £ 102 –3.6 £ 105 s in a vacuum chamber, which was evacuated continuously by a vacuum pump to the residual pressure of around 2.7 £ 1023 Pa. All specimens were heated at a constant rate of 0.67 K/s up to the prescribed temperature, and after the exposure for the prescribed time, they were cooled down to room temperature at the same rate of 0.67 K/s as well as those for air exposure. The weight change due to the exposure was measured to monitor the difference in degradation process between the vacuum and air exposures. Tensile test was carried out at room temperature at a crosshead speed of 8.3 £ 1026 m/s for a gage length 50 mm. The fracture surface of the composite was observed with scanning electron microscope (SEM (Jeol, JSM-5410LS)). 3. Results and discussion 3.1. Residual strength of vacuum-exposed composite Fig. 1 shows the weight change due to the exposure in vacuum and air. For 973 K-exposure, the weight change was minor in both environments. With increasing exposure temperature, the weight tends to increase in the case of air exposure. On contrary, the weight tends to decrease in Fig. 1. Weight change of the composite due to the exposure in vacuum and in air at 973, 1473 and 1673 K. 34 S. Ochiai et al. / Composites: Part A 35 (2004) 33–40
S Ochiai et aL./Composites: Part A 35 (2004)33-40 the vacuum exposure at higher temperatures. The weight 3. 2. Degradation mechanism loss increased with increasing time and it became larger than 30% when the specimens were exposed at 1473 and The as-supplied composite showed fiber-pullout type 1673 K for more than 3.6X 10s. Such a large weight loss fracture mode. In the case of air-exposure, with progress of could be attributed to the decomposition of fiber and matrix oxidation, the fracture mode changed from the fiber-pullout involving generation of Sio and CO gases [6,7] type to the nonfiber-pullout, but the reduction in composite The average room temperature strength of the as- strength did not correspond to the change in fracture supplied specimens was 630 MPa. Fig. 2 shows the mode. Namely, most specimens for the strength range of measured room temperature-residual strength of the speci- 130-630 MPa showed fiber-pullout type, and only the mens exposed in vacuum at(a)973, ()1123, (c)1273, (d) severely degraded specimens with the strength range of 1473,(e)1573 and(f)1673 K, plotted against exposure time 60-130 MPa showed the nonfiber-pullout one On the other t. As shown later in Section 3.3, from the analysis and hand, in the vacuum exposure, all specimens showed the modeling, the parameters, which can describe the exper- fiber-pullout type. The fiber-pullout length increased with imental data, were estimated. The solid lines show the increasing temperature and time. The variation of the results calculated by such estimated values. Within the fracture morphology with increasing time in the case of range of the exposure time investigated (up to 1473 K-exposure is shown in Fig 3. In this example, the t= 3.6x 10 s), following features are read strength of the specimen exposed for 3. 6X 10-s was nearly the same as that of the as-supplied specimens(630 MPa)but (1)For 973, 1123 and 1273 K-exposure, no or only a slight it decreased with increasing exposure time, falling to around reduction in strength arose as in(a)-(c), respectively. 100 MPa for the exposure time of 3. 6x 10 s, as has been (i) For 1473 K-exposure, the strength for the 3.6X 10 shown in Fig. 2(d (shortest exposure time in this work) was comparable Fig. 4 shows the appearance of the matrix and fiber in the to that of the as-supplied (nonexposed) specimens, but composite specimens exposed at the same temperature of it decreased with increasing time as in(d) 1473 K in vacuum. (a)-(c) show the defects in the matrix, (iii) For 1573 and 1673 K-exposure, the strength decreased on the side surface of the pulled-out fiber and in the fracture with increasing time surface of the fiber, respectively. In this way, many defects 1000 1000 973K 1123K 1273K g100-10810105105要10-10-10410510 02103104105106 Exposure time, t(s) time, t(s) 1000 1000 n1000 1473K 1573K 画1098101-1:/重10 02103104105106 Exposure time, t(s) Exposure time, t (s) Exposure time, t(s) Fig. 2. Relation between the residual tensile strength and exposure time in vacuum. The solid lines refer to the analytical results based on proposed model whose details will be shown in Section 3.3
the vacuum exposure at higher temperatures. The weight loss increased with increasing time and it became larger than 30% when the specimens were exposed at 1473 and 1673 K for more than 3.6 £ 104 s. Such a large weight loss could be attributed to the decomposition of fiber and matrix involving generation of SiO and CO gases [6,7]. The average room temperature strength of the assupplied specimens was 630 MPa. Fig. 2 shows the measured room temperature-residual strength of the specimens exposed in vacuum at (a) 973, (b) 1123, (c) 1273, (d) 1473, (e) 1573 and (f) 1673 K, plotted against exposure time t. As shown later in Section 3.3, from the analysis and modeling, the parameters, which can describe the experimental data, were estimated. The solid lines show the results calculated by such estimated values. Within the range of the exposure time investigated (up to t ¼ 3:6 £ 105 s), following features are read (i) For 973, 1123 and 1273 K-exposure, no or only a slight reduction in strength arose as in (a)–(c), respectively. (ii) For 1473 K-exposure, the strength for the 3.6 £ 102 s (shortest exposure time in this work) was comparable to that of the as-supplied (nonexposed) specimens, but it decreased with increasing time as in (d). (iii) For 1573 and 1673 K-exposure, the strength decreased with increasing time. 3.2. Degradation mechanism The as-supplied composite showed fiber-pullout type fracture mode. In the case of air-exposure, with progress of oxidation, the fracture mode changed from the fiber-pullout type to the nonfiber-pullout, but the reduction in composite strength did not correspond to the change in fracture mode. Namely, most specimens for the strength range of 130–630 MPa showed fiber-pullout type, and only the severely degraded specimens with the strength range of 60–130 MPa showed the nonfiber-pullout one. On the other hand, in the vacuum exposure, all specimens showed the fiber-pullout type. The fiber-pullout length increased with increasing temperature and time. The variation of the fracture morphology with increasing time in the case of 1473 K-exposure is shown in Fig. 3. In this example, the strength of the specimen exposed for 3.6 £ 102 s was nearly the same as that of the as-supplied specimens (630 MPa) but it decreased with increasing exposure time, falling to around 100 MPa for the exposure time of 3.6 £ 105 s, as has been shown in Fig. 2(d). Fig. 4 shows the appearance of the matrix and fiber in the composite specimens exposed at the same temperature of 1473 K in vacuum. (a)–(c) show the defects in the matrix, on the side surface of the pulled-out fiber and in the fracture surface of the fiber, respectively. In this way, many defects Fig. 2. Relation between the residual tensile strength and exposure time in vacuum. The solid lines refer to the analytical results based on proposed model, whose details will be shown in Section 3.3. S. Ochiai et al. / Composites: Part A 35 (2004) 33–40 35
S. Ochiai et al./ Composites: Part A 35(2004)33-40 5μm Imm (a)3,6×103s (d) Fig. 3. Appearance of the fractured SiC/SiC composite exposed in vacuuM at1473Kfor(a)3.6×102,(b)3.6×103,(c)3.6×104and(d)3.6×105s. are formed at the fiber surface due to the decomposition of the fiber and matrix through the generation of Sio and co gases [6,7]. The result mentioned above indicates the (b)3.6×104 (1) The defects shown in Fig. 4 are thought to act to reduce e strength as indicated by Hollon et al. [7]. The rength not only of the fiber but also of the matrix is reduced (2) All specimens showed the fiber-pullout fracture mod accompanied by the interfacial debonding. The increase in fiber-pullout length with increasing exposure time ( Fig 3)means that the interfacial strength between fiber and matrix decreases. The decrease in interfacial strength could be attributed to the consumption of the carbon at the interface and to the thinning of fiber and 5 matrix due to the decomposition. The amount of the carbon at the interface is however small. It is deduced (c)3.6×10 that the carbon is consumed in the initial stage. The large amount of the continuing weight loss in the late stage Fig 4. Appearance of the matrix and fiber in the composite specimens (Fig. 1)could be attributed to the decomposition of fiber posed at the same temperature of 1473 K in vacuum.(a)-(c) show the and matrix defects in the matrix, on the side surface of the pulled-out fiber and in the racture surface of the fiber, respectively. (3)As shown in our recent analysis and simulation of debonding in the composite [12]. the fracture mode determined by the relative ratio of interfacial strength to bonding strength decreased far more than the fiber component strength, and the fracture mode changes strength with progress of the decomposition. gradually from fiber-pullout to nonfiber-pullout typ (4) Some portions of the matrix had flown away upon with increasing relative ratio. If we assume that the fracture of the specimens, as shown in Fig 3(c)and(d) interfacial bonding strength is kept during the vacuum This phenomenon could be attributed to the reduced exposure, the pullout length of fiber decreases with strength of the matrix and interface. Such a fracture decreasing fiber strength. However, it is not the case behavior suggests that the matrix could not support the since the fiber-pullout length increases with decreasing applied load and most applied load was supported by the fiber strength. Thus it is suggested that the interfacial fiber
are formed at the fiber surface due to the decomposition of the fiber and matrix through the generation of SiO and CO gases [6,7]. The result mentioned above indicates the followings. (1) The defects shown in Fig. 4 are thought to act to reduce the strength as indicated by Chollon et al. [7]. The strength not only of the fiber but also of the matrix is reduced. (2) All specimens showed the fiber-pullout fracture mode, accompanied by the interfacial debonding. The increase in fiber-pullout length with increasing exposure time (Fig. 3) means that the interfacial strength between fiber and matrix decreases. The decrease in interfacial strength could be attributed to the consumption of the carbon at the interface and to the thinning of fiber and matrix due to the decomposition. The amount of the carbon at the interface is, however, small. It is deduced that the carbon is consumed in the initial stage. The large amount of the continuing weight loss in the late stage (Fig. 1) could be attributed to the decomposition of fiber and matrix. (3) As shown in our recent analysis and simulation of debonding in the composite [12], the fracture mode is determined by the relative ratio of interfacial strength to component strength, and the fracture mode changes gradually from fiber-pullout to nonfiber-pullout type with increasing relative ratio. If we assume that the interfacial bonding strength is kept during the vacuum exposure, the pullout length of fiber decreases with decreasing fiber strength. However, it is not the case since the fiber-pullout length increases with decreasing fiber strength. Thus it is suggested that the interfacial bonding strength decreased far more than the fiber strength with progress of the decomposition. (4) Some portions of the matrix had flown away upon fracture of the specimens, as shown in Fig. 3(c) and (d). This phenomenon could be attributed to the reduced strength of the matrix and interface. Such a fracture behavior suggests that the matrix could not support the applied load and most applied load was supported by the fiber. Fig. 3. Appearance of the fractured SiC/SiC composite exposed in vacuum at 1473 K for (a) 3.6 £ 102 , (b) 3.6 £ 103 , (c) 3.6 £ 104 and (d) 3.6 £ 105 s. Fig. 4. Appearance of the matrix and fiber in the composite specimens exposed at the same temperature of 1473 K in vacuum. (a)–(c) show the defects in the matrix, on the side surface of the pulled-out fiber and in the fracture surface of the fiber, respectively. 36 S. Ochiai et al. / Composites: Part A 35 (2004) 33–40
S Ochiai et aL./Composites: Part A 35 (2004)33-40 (5) The reduction in composite strength is very large(from of the exposure temperature (T) and time(t)is derived as 630 MPa to less than 100 MPa). Such a large strength follows reduction cannot be accounted for by the decrease in The strength of the surface-damaged fiber is determined interfacial strength since (i) the interfacial strength is by the effective defect size depending on the shape and the originally made low to allow interfacial debonding for fracture toughness of the fiber [17]. In the present work, we suppression of mode I crack extension, and (ii)even if assume that the fracture toughness of fiber is not changed the interfacial bonding strength decreases to zero, the and the effective size a of the strength-determining defects omposite can retain the strength fiber bundle increases with increasing T and t is expressed by strength x fiber volume fraction which is not so low if the fiber strength is retained. a= Br'exp(-O/RT) (6)Thus, in the case of vacuum exposure, the reduction in where B is the constant, n is the reaction index depending on strength of composite could be attributed mainly to the the rate-controlling process, e is the activation energy and R reduction in fiber strength due to the formation of is the gas constant. In the case of air exposure, n was 1/2 for defects on fiber surface(Fig 4(b)and (c)), if the fracture thin oxide layer since the rate controlling for toughness-values of the fiber before and after the formation of SiO2 layer is the diffusion of oxygen through exposure were not so much different, while, in the Sio2 layer [6]. In the present case of vacuum exposure, as case of air exposure, the reduction could be attributed to the rate controlling process is unknown, the value of n the extension of the crack made by the premature cannot be given in advance. It will be shown later to be 1/2 fracture of SiO2 layer into fiber [1]. from the fitting of the model to experimental results (7) As mentioned in(6), the strength-determining factor is Under the existence of the defect with an effective size a different between the vacuum exposure(formation of the fiber stress at fracture op is given by defect accompanied by the gas generation in the decomposition process)and air exposure(formation df=K/y(a) of crack due to a premature fracture of the Sio2 layer However it is common in vacuum and where Kle is the fracture toughness of the fiber and y is the that the reduction in strength of composite could be finite correction factor, being 1. 12 for small surface flaw attributed to the reduction in fiber strength, which could The fracture toughness Klc of the Nicalon [18] and Tyranno be described by the fracture mechanical approach as [19] fibers has been reported to be around 1 MPa m".For Klc =l MPa m, the strength is calculated to be 1.5 GPa for a=300 nm and 0.9 GPa for 1 um. In this way, the strength of the fiber is seriously reduced when the defects 3.3. Modeling to describe the composite strength as a are formed on fiber surface function of vacuum exposure temperature and time Noting the strength of composite determined by propagation of the defect into the fiber as d and substitutin own above the main reason for the degradation car Eqs.(1)and(2)into dc= Adf, we have be attributed to the reduction in fiber strength due to the formation of defects in the case of vacuum exposure if the d C=Snt exp(@n2RT) fracture toughness of the fiber after the exposure is not so where Su is a constant given by a(B)"K/(Y()) much different from that before it, and due to the premature When the defect formed by the vacuum exposure is fracture of the brittle SiO, layer in the case of air exposure. smaller than the intrinsic one, the value of oc calculated by In our preceding work [1], the approach to describe the Eq (3)is higher than the original strength ocu In this case, strength of fiber with a cracked reaction layer [13-16] was he fiber is broken not by the extrinsic defect formed by the applied to describe the strength of the air-exposed vacuum exposure but by the intrinsic flaws. Such a case composite as a function of temperature and time, by when the effective size of the extrinsic defect is combining the kinetics of growth of the Sioz layer on the than the strength-determining defect of the original fiber surface and fracture mechanics. In this work, the similarly as it arises when the thickness of the procedure presented in our preceding work is applied as to reaction layer is smaller than the defect of the original fiber describe the strength of vacuum-exposed composite with in the case of cracking of the reaction layer [1, 13-16) surface-damaged fibers as a function of exposure tempera- As the size of the extrinsic defect increases with time. the ture and time variation of the composite strength with time can be divided As indicated in Section 3. 2, most applied stress is into two stages; Stage I where the strength is the same as supported by the fiber, while the fiber is degraded by the that of as-supplied specimen(eu=ocu)and Stage Il where formation of the defects Due to the reason mentioned in(4) the strength decreases with increasing time, obeying Eq. ( 3), in Section 3. 2, the strength of composite Ocu is taken to be (u= op) proportional to the fiber strength o(ocu Aofu where A is a The variation of the cu as a function of t is schematically constant). The defect-determined fiber strength as a function own in Fig. 5(a). At a given exposure temperature
(5) The reduction in composite strength is very large (from 630 MPa to less than 100 MPa). Such a large strength reduction cannot be accounted for by the decrease in interfacial strength since (i) the interfacial strength is originally made low to allow interfacial debonding for suppression of mode I crack extension, and (ii) even if the interfacial bonding strength decreases to zero, the composite can retain the strength ‘fiber bundle strength £ fiber volume fraction’, which is not so low if the fiber strength is retained. (6) Thus, in the case of vacuum exposure, the reduction in strength of composite could be attributed mainly to the reduction in fiber strength due to the formation of defects on fiber surface (Fig. 4(b) and (c)), if the fracture toughness-values of the fiber before and after the exposure were not so much different, while, in the case of air exposure, the reduction could be attributed to the extension of the crack made by the premature fracture of SiO2 layer into fiber [1]. (7) As mentioned in (6), the strength-determining factor is different between the vacuum exposure (formation of defect accompanied by the gas generation in the decomposition process) and air exposure (formation of crack due to a premature fracture of the SiO2 layer). However, it is common in vacuum and air exposures that the reduction in strength of composite could be attributed to the reduction in fiber strength, which could be described by the fracture mechanical approach as shown below. 3.3. Modeling to describe the composite strength as a function of vacuum exposure temperature and time As shown above, the main reason for the degradation can be attributed to the reduction in fiber strength due to the formation of defects in the case of vacuum exposure if the fracture toughness of the fiber after the exposure is not so much different from that before it, and due to the premature fracture of the brittle SiO2 layer in the case of air exposure. In our preceding work [1], the approach to describe the strength of fiber with a cracked reaction layer [13–16] was applied to describe the strength of the air-exposed composite as a function of temperature and time, by combining the kinetics of growth of the SiO2 layer on the fiber surface and fracture mechanics. In this work, the procedure presented in our preceding work is applied as to describe the strength of vacuum-exposed composite with surface-damaged fibers as a function of exposure temperature and time. As indicated in Section 3.2, most applied stress is supported by the fiber, while the fiber is degraded by the formation of the defects. Due to the reason mentioned in (4) in Section 3.2, the strength of composite scu is taken to be proportional to the fiber strength sfuðscu ¼ Asfu where A is a constant). The defect-determined fiber strength as a function of the exposure temperature ðTÞ and time ðtÞ is derived as follows. The strength of the surface-damaged fiber is determined by the effective defect size depending on the shape and the fracture toughness of the fiber [17]. In the present work, we assume that the fracture toughness of fiber is not changed and the effective size a of the strength-determining defects increases with increasing T and t is expressed by a ¼ Btn expð2Q=RTÞ ð1Þ where B is the constant, n is the reaction index depending on the rate-controlling process, Q is the activation energy and R is the gas constant. In the case of air exposure, n was 1/2 for thin oxide layer since the rate controlling process for formation of SiO2 layer is the diffusion of oxygen through SiO2 layer [6]. In the present case of vacuum exposure, as the rate controlling process is unknown, the value of n cannot be given in advance. It will be shown later to be 1/2 from the fitting of the model to experimental results. Under the existence of the defect with an effective size a; the fiber stress at fracture sp f is given by sp f ¼ KIc=YðpaÞ 1=2 ð2Þ where KIc is the fracture toughness of the fiber and Y is the finite correction factor, being 1.12 for small surface flaw. The fracture toughness KIc of the Nicalon [18] and Tyranno [19] fibers has been reported to be around 1 MPa m1/2. For KIc ¼ 1 MPa m1=2; the strength is calculated to be 1.5 GPa for a ¼ 300 nm and 0.9 GPa for 1 mm. In this way, the strength of the fiber is seriously reduced when the defects are formed on fiber surface. Noting the strength of composite determined by the propagation of the defect into the fiber as sp c and substituting Eqs. (1) and (2) into sp c ¼ Asp f ; we have sp c ¼ SIIt 2n=2 expðQ=2RTÞ ð3Þ where SII is a constant given by AðBÞ 2nKIc={YðpÞ 1=2} When the defect formed by the vacuum exposure is smaller than the intrinsic one, the value of sp c calculated by Eq. (3) is higher than the original strength s0 cu In this case, the fiber is broken not by the extrinsic defect formed by the vacuum exposure but by the intrinsic flaws. Such a case arises when the effective size of the extrinsic defect is smaller than the strength-determining defect of the original fiber, as similarly as it arises when the thickness of the reaction layer is smaller than the defect of the original fiber in the case of cracking of the reaction layer [1,13–16]. As the size of the extrinsic defect increases with time, the variation of the composite strength with time can be divided into two stages; Stage I where the strength is the same as that of as-supplied specimen ðscu ¼ s0 cuÞ and Stage II where the strength decreases with increasing time, obeying Eq. (3), ðscu ¼ sp c Þ. The variation of the scu as a function of t is schematically shown in Fig. 5(a). At a given exposure temperature, S. Ochiai et al. / Composites: Part A 35 (2004) 33–40 37
S Ochiai et al./ Composites: Part A 35(2004)33-40 as shown by the slope of the solid lines in the streng reduced region in Fig. 2. From the fitting of the model to No reduction experimental data in Stage II in Fig. 2, the constant SI and (ocu=Oco) the activation energy g for n= 1/2 were estimated to be In(ocu) 0.35 MPa/s and 220 kJ/mol, respectively. The solid lines in Fig. I show the calculated In(ocu)-In(t)diagram by using thus estimated values, describing fairly well the experimental results At each temperature, the calculated lines in Fig. 2 do not Shift to right necessarily go through the center of the measured values, since the values of n, Sn and Q were estimated as to fit the all exposure data covering all temperatures investigated. The validity of the estimated values of n= 1/2, Sn=0.35 MPa/sand Critical time Q=220 kJ/mol can be demonstrated as follows. Eq. (3) In(t) with n= 1/2 can be re-written as In(ocu)=(-1/4)(In(t)-20/kT)+In(Sm) (4) Low Temp. Using Eq. (4), we can superimpose all data of Tcu in graph by plotting In(ocu) against In(t)-20/kT. Fig. 6 shows the results of such plotting for vacuum exposure, gether with those for air exposure [1] for reference. The Intermediate Temp. i solid lines show the calculated ones by using the foremen tioned oc=630 MPa, n= 1/2, S=0.35 MPa/s 4and 2=220 kJ/mol for the vacuum exposure and by using 1000 D In(t) Fig. 5. Schematic representation of variation of te strength with exposure time. (a) The principle variation of the In(oc) as a function of n(n) which would be realized if the tre time covers zero to infinite seconds. (b) The variation of the In(o)as a function of In(n) in the limited exposure time range as in the present experiment. The strengths of the composite exposed at low, intermediate and high temperature vary alon AB, ACD and Ef, respectively. 15 the time satisfying dc= c is noted as the critical time, n t-2O/RT below and beyond which Stages I and ll arise, respectively (a)Air variation shown in Fig. 5(a) would be realized if the exposure time covers zero to infinite seconds. As the exposure time is limited to 3.6 x 102-36x 10s in the 乏 present work, both Stages of I and Il do not appear necessarily at respective temperature, since the critical time, corresponding to the transition from Stage I to Il, becomes shorter and longer when the exposure temperature is high and low, respectively. Which of Stages I and Il arise in the In(cu-In() curve in the limited time range at different temperatures is schematically shown in Fig. 5(b). The trengths of the composite exposed at low, intermediate and s high temperature vary along AB (only Stage I arises), ACD (Stage I arises up to the time C(critical time), beyond which Stage II arises)and EF (only Stage II arises), respectively. n t-2/RT The slope of In(ocu)-In(t) curve in Stage Il is -n/2 While the data were scattered, n= 1/2 showed a good description of the slope at all temperatures investigated Fig. 6. Measured values of In(ocu) plotted against In(r)-2o/kT
the time satisfying s0 cu ¼ sp c is noted as the critical time, below and beyond which Stages I and II arise, respectively. The variation shown in Fig. 5(a) would be realized if the exposure time covers zero to infinite seconds. As the exposure time is limited to 3.6 £ 102 –3.6 £ 105 s in the present work, both Stages of I and II do not appear necessarily at respective temperature, since the critical time, corresponding to the transition from Stage I to II, becomes shorter and longer when the exposure temperature is high and low, respectively. Which of Stages I and II arise in the lnðscuÞ–lnðtÞ curve in the limited time range at different temperatures is schematically shown in Fig. 5(b). The strengths of the composite exposed at low, intermediate and high temperature vary along AB (only Stage I arises), ACD (Stage I arises up to the time C (critical time), beyond which Stage II arises) and EF (only Stage II arises), respectively. The slope of lnðscuÞ 2 lnðtÞ curve in Stage II is 2n=2: While the data were scattered, n ¼ 1=2 showed a good description of the slope at all temperatures investigated as shown by the slope of the solid lines in the strengthreduced region in Fig. 2. From the fitting of the model to experimental data in Stage II in Fig. 2, the constant SII and the activation energy Q for n ¼ 1=2 were estimated to be 0.35 MPa/s1/4 and 220 kJ/mol, respectively. The solid lines in Fig. 1 show the calculated lnðscuÞ 2 lnðtÞ diagram by using thus estimated values, describing fairly well the experimental results. At each temperature, the calculated lines in Fig. 2 do not necessarily go through the center of the measured values, since the values of n; SII and Q were estimated as to fit the all data covering all temperatures investigated. The validity of the estimated values of n ¼ 1=2; SII ¼ 0:35 MPa=s 1=4 and Q ¼ 220 kJ/mol can be demonstrated as follows. Eq. (3) with n ¼ 1=2 can be re-written as lnðscuÞ¼ð21=4Þ{lnðtÞ 2 2Q=kT} þ lnðSIIÞ ð4Þ Using Eq. (4), we can superimpose all data of scu in one graph by plotting lnðscuÞ against lnðtÞ 2 2Q=kT: Fig. 6 shows the results of such plotting for vacuum exposure, together with those for air exposure [1] for reference. The solid lines show the calculated ones by using the aforementioned s0 cu ¼ 630 MPa; n ¼ 1=2; SII ¼ 0:35 MPa/s1/4 and Q ¼ 220 kJ/mol for the vacuum exposure and by using Fig. 5. Schematic representation of variation of composite strength with exposure time. (a) The principle variation of the lnðscuÞ as a function of lnðtÞ; which would be realized if the exposure time covers zero to infinite seconds. (b) The variation of the lnðscuÞ as a function of lnðtÞ in the limited exposure time range as in the present experiment. The strengths of the composite exposed at low, intermediate and high temperature vary along AB, ACD and EF, respectively. Fig. 6. Measured values of lnðscuÞ plotted against lnðtÞ 2 2Q=kT: 38 S. Ochiai et al. / Composites: Part A 35 (2004) 33–40
S Ochiai et aL./Composites: Part A 35 (2004)33-40 the estimated values in the preceding work [1] for the ai Based on the results mentioned above in combination exposure For both air-and vacuum exposures, the models with the kinetics of growth of the defects and fracture describe well the experimental results. In this way, when the mechanics, a simple model was presented to describe the all data are superimposed in one graph, the validity of the variation of composite strength as a function of exposure estimated values for both exposures is clearly shown. temperature and time for the vacuum exposure, which Only Stages I and II were found in the case of vacuum could describe the experimental results exposure On the other hand, in the case of air exposure, Stage Ill, where the strength goes down rather slowly with increasing time in comparison with that in Stage I wasAcknowledgements found in addition to Stages l and l. Such a difference arises from the difference in fracture mechanism. The fiber is The present work was supported by the New Energy and fractured directly from the defect in the case of vacuum xposure. Thus, the strength is reduced monotonically with Japan ndustrial Technology Development Organization(NEDO), increasing time. On the other hand, in the case of air exposure, the fiber is fractured by the extension of the crack made by fracture of the brittle Sio2 layer. In this process, the crack does not extend upon formation but it extends after References increment of applied stress when the Sio2 is thin( Stage l), [1 Ochiai S, Kimura S, Tanaka M, Tanaka H, Hojo M, Morishita K. while it extends upon formation when the Sio2 layer is thick Okuda H, Nakayama H, Shibata K, Sato M. Residual strength of SiC/ (Stage I[1]. Thus the reduction in strength is described by SiC composite exposed at high temperatures in air as a function of three stages in the air exposure exposure temperature and time. Compos Part A, in press. The behavior mentioned above might be, however, [2] Bunsell AR, Berger MH. Fine ceramic fibres. New York: Marcel limited to short time exposure. If the saturation of reaction Dekker: 1999 [3] Bunsell AR, Berger MH. Fine diameter ceramic fibres. J Eur Ceramic occurs or the decomposition speed becomes very low for very Soc2000:20:2249-60 long exposure time, the effective defect size would not obey [4] Shimo T, Chen H, Okamura K. High-temperature stability of Eq (1)but would tend to grow very slowly. In such a case, an icalon under Ar or O2 atmosphere. J Mater Sci 1994: 29: 456-63 additional stage where the reduction in strength is not 5] Kakimoto K, Shimo T, Okamura K, Seguchi T, Sato M, Kumagawa strongly dependent on time could be expected to arise. On K, Yamamura T. High-temperature crystallization behavior of si this point, further study under long-term exposure is needed Ti-C-O fiber cured by electron beam irradiation. J Jpn Inst Metals 199458(2):229-34 [6] Shimo T, Okamura K, Hayatsu T Effect of atmosphere on pyrolysis of Nicalon J Mater Sci 1996: 31: 4407-13 4. Conclusions [71 Hollon G, Pailler R, Naslain R, Laanani F, Monthioux M, Olry P. Thermal stability of PCS-derived SiC fibre with a low oxygen content The room temperature residual strength of SiC(ZMI fiber (Hi-Nicalon). J Mater Sci 1997: 32: 327-47 [8] Hollon C, Pailler R, Naslain R, Olry P. bricated by Ube Companies )/SiC composite exposed at microstructure and mechanical behavior at high high temperatures(823-1673 K) in vacuum was studied fibre with a low oxygen content (Hi-Nicalon) and compared with that exposed in air. Main results are 1133-47 summarized as follows [9] Jia N, Bodet R, Tressler RE. Effects of microsturute instability on the reep behavior of Si-C-O(Nicalon)fibers in argon. J Am Ceram Soc 376(12):3051-60 1. The residual strength decreased with increasing exposure [10] Ultra high temperature materials research institute. Study on emperature and time both for vacuum and air exposures durability and life prediction of continuous fiber reinforced The variation of the fracture mode was, however, quite ceramic matrix composite. Research report NEDO-lTK-9911 different. In case of vacuum exposure, only the fiber 2000p.6-9 pullout type occurred and also the pullout length [11] Tanaka Y, Inoue Y, Miyamoto N, Sato M, Yamamura T Properties of Si-Zr-C-O Fiber/SizrC composites dispersed increased with increasing temperature and time, while ZrsiO4 particles in the matrix. Int J Mater Prod Technol 2001 in the case of air exposure, the fracture mode changed 16(1-3):197-205 from fiber-pullout to nonfiber-pullout one with increas- [12] Ochiai S, Hojo M, Schulte K, Fiedler B Nondimensional simulation ing temperature and time of influence of toughness of interface on tensile stress-strain 2. The degradation of the composite due to the exposure at behavior of unidirectional microcomposite. Compos Part A 2001 32(6:749-61 high temperatures could be attributed to the reduction in [13] Metcalfe AG, Klein KJ Interface in metal matrix composites.Effects fiber strength for both environments, but the mechanism of the interface on longitudinal tensile properties. New York: was different: the extension of the decomposition Academic Press: 1974. P. 125-68 induced surface defects into fiber in the case of vacuum [14] Shorshorov MK, Ustinov LM, Zirlin AM, Olefilebko VL, Vonogra- exposure and the extension of the crack made by dov Lv. Brittle interface layers and the tensile strength of metal matrix-fibre composites. J Mater Sci 1979: 14: 1850-61 premature fracture of the SiOz layer into fiber in [15] Ochiai A, Murakami Y. Tensile strength of composites with brittle case of air reaction zones at interface. J Mater Sci 1979: 14: 831-40
the estimated values in the preceding work [1] for the air exposure. For both air- and vacuum exposures, the models describe well the experimental results. In this way, when the all data are superimposed in one graph, the validity of the estimated values for both exposures is clearly shown. Only Stages I and II were found in the case of vacuum exposure. On the other hand, in the case of air exposure, Stage III, where the strength goes down rather slowly with increasing time in comparison with that in Stage II was found in addition to Stages II and I. Such a difference arises from the difference in fracture mechanism. The fiber is fractured directly from the defect in the case of vacuum exposure. Thus, the strength is reduced monotonically with increasing time. On the other hand, in the case of air exposure, the fiber is fractured by the extension of the crack made by fracture of the brittle SiO2 layer. In this process, the crack does not extend upon formation but it extends after increment of applied stress when the SiO2 is thin (Stage II), while it extends upon formation when the SiO2 layer is thick (Stage III) [1]. Thus the reduction in strength is described by three stages in the air exposure. The behavior mentioned above might be, however, limited to short time exposure. If the saturation of reaction occurs or the decomposition speed becomes very low for very long exposure time, the effective defect size would not obey Eq. (1) but would tend to grow very slowly. In such a case, an additional stage where the reduction in strength is not strongly dependent on time could be expected to arise. On this point, further study under long-term exposure is needed. 4. Conclusions The room temperature residual strength of SiC(ZMI fiber fabricated by Ube Companies)/SiC composite exposed at high temperatures (823–1673 K) in vacuum was studied and compared with that exposed in air. Main results are summarized as follows 1. The residual strength decreased with increasing exposure temperature and time both for vacuum and air exposures. The variation of the fracture mode was, however, quite different. In case of vacuum exposure, only the fiberpullout type occurred and also the pullout length increased with increasing temperature and time, while in the case of air exposure, the fracture mode changed from fiber-pullout to nonfiber-pullout one with increasing temperature and time. 2. The degradation of the composite due to the exposure at high temperatures could be attributed to the reduction in fiber strength for both environments, but the mechanism was different: the extension of the decompositioninduced surface defects into fiber in the case of vacuum exposure and the extension of the crack made by premature fracture of the SiO2 layer into fiber in the case of air exposure. 3. Based on the results mentioned above in combination with the kinetics of growth of the defects and fracture mechanics, a simple model was presented to describe the variation of composite strength as a function of exposure temperature and time for the vacuum exposure, which could describe the experimental results. Acknowledgements The present work was supported by the New Energy and Industrial Technology Development Organization (NEDO), Japan. References [1] Ochiai S, Kimura S, Tanaka M, Tanaka H, Hojo M, Morishita K, Okuda H, Nakayama H, Shibata K, Sato M. Residual strength of SiC/ SiC composite exposed at high temperatures in air as a function of exposure temperature and time. Compos Part A, in press. [2] Bunsell AR, Berger MH. Fine ceramic fibres. New York: Marcel Dekker; 1999. [3] Bunsell AR, Berger MH. Fine diameter ceramic fibres. J Eur Ceramic Soc 2000;20:2249–60. [4] Shimoo T, Chen H, Okamura K. High-temperature stability of Nicalon under Ar or O2 atmosphere. J Mater Sci 1994;29:456–63. [5] Kakimoto K, Shimoo T, Okamura K, Seguchi T, Sato M, Kumagawa K, Yamamura T. High-temperature crystallization behavior of Si– Ti–C–O fiber cured by electron beam irradiation. J Jpn Inst Metals 1994;58(2):229–34. [6] Shimoo T, Okamura K, Hayatsu T. Effect of atmosphere on pyrolysis of Nicalon. J Mater Sci 1996;31:4407–13. [7] Chollon G, Pailler R, Naslain R, Laanani F, Monthioux M, Olry P. Thermal stability of PCS-derived SiC fibre with a low oxygen content (Hi-Nicalon). J Mater Sci 1997;32:327–47. [8] Chollon C, Pailler R, Naslain R, Olry P. Correlation between microstructure and mechanical behavior at high temperatures of a SiC fibre with a low oxygen content (Hi-Nicalon). J Mater Sci 1997;32: 1133–47. [9] Jia N, Bodet R, Tressler RE. Effects of microsturute instability on the creep behavior of Si–C–O (Nicalon) fibers in argon. J Am Ceram Soc 1993;76(12):3051–60. [10] Ultra high temperature materials research institute. Study on durability and life prediction of continuous fiber reinforced ceramic matrix composite. Research report NEDO-ITK-9911. 2000; p. 6–9. [11] Tanaka Y, Inoue Y, Miyamoto N, Sato M, Yamamura T. Properties of Si–Zr–C–O Fiber/SiZrC composites dispersed ZrSiO4 particles in the matrix. Int J Mater Prod Technol 2001; 16(1–3):197–205. [12] Ochiai S, Hojo M, Schulte K, Fiedler B. Nondimensional simulation of influence of toughness of interface on tensile stress–strain behavior of unidirectional microcomposite. Compos Part A 2001; 32(6):749–61. [13] Metcalfe AG, Klein KJ. Interface in metal matrix composites. Effects of the interface on longitudinal tensile properties. New York: Academic Press; 1974. p. 125–68. [14] Shorshorov MK, Ustinov LM, Zirlin AM, Olefilebko VL, Vonogradov LV. Brittle interface layers and the tensile strength of metal matrix–fibre composites. J Mater Sci 1979;14:1850–61. [15] Ochiai A, Murakami Y. Tensile strength of composites with brittle reaction zones at interface. J Mater Sci 1979;14:831–40. S. Ochiai et al. / Composites: Part A 35 (2004) 33–40 39
S. Ochiai et al./ Composites: Part A 35(2004)33-40 [16] Ochiai S, Hojo M, Tanaka M. Mechanical interactions between fiber [18] Kondo M, Imai Y, Tezuka H, Koh and cracked coating layer and their influences on fiber strength. fibers extracte Compos Part A 1999: 30: 451-6 [17] Ochiai S, Inoue T, Fujita T, Hojo M, H. Leucht R 1463-9 Degradation mechanism of SiC/Super due to [19 Matsunaga K, Ochiai S, Osamura K, Waku Y, YamamuraT. Influence of interfacial reaction. Metall Mat Tans 1999:30A(10) heat-treatment on mechanical property of Si-Ti-C-O fiber-reinforced 2713-20. aluminum matrix composites. J Jpn Inst Metals 1993: 57: 1035-40
[16] Ochiai S, Hojo M, Tanaka M. Mechanical interactions between fiber and cracked coating layer and their influences on fiber strength. Compos Part A 1999;30:451–61. [17] Ochiai S, Inoue T, Fujita T, Hojo M, Dudek HJ, Leucht R. Degradation mechanism of SiC/Super a2 composite due to interfacial reaction. Metall Mater Trans 1999;30A(10): 2713–20. [18] Kondo M, Imai Y, Tezuka H, Kohyama A. Fracture strength of PCS– SiC fibers extracted from metal matrix composites: effects of surface defects and a new evaluation method. Tetsu to Hagane 1989;75(9): 1463–9. [19] Matsunaga K, Ochiai S, Osamura K, Waku Y, Yamamura T. Influence of heat-treatment on mechanical property of Si–Ti–C–O fiber-reinforced aluminum matrix composites. J Jpn Inst Metals 1993;57:1035–40. 40 S. Ochiai et al. / Composites: Part A 35 (2004) 33–40