SEI TERIALS ENGE& ENGIEERN ELSEVIER Materials Science and Engineering A300(2001)196-202 www.elsevier.com/locate/msea Mechanical properties of 3D fiber reinforced C/SiC composites Yongdong Xu, Aifei Cheng, Litong Zhang, Hongfeng Yin, Xiaowei Yin State Key Laboratory of solidifi n Polytechnical Unicersity, Xian, Shaanxi 710072, People's Republic of China Received 9 May 2000: received in revised form 17 August 2000 abstract High toughness and reliable three dimensional textile carbon fiber reinforced silicon carbide composites were fabricated by chemical vapor infiltration. Mechanical properties of the composite materials were investigated under bending, shear, and impact loading. The density of the composites was 2.0-2. 1 g cm after the three dimensional carbon preform was infiltrated for 30 h The values of flexural strength were 441 MPa at room temperature, 450 MPa at 1300C, and 447 MPa at 1600oC. At elevated temperatures(1300 and 1600.C), the failure behavior of the composites became some brittle because of the strong interfacial bonding caused by the mis-match of thermal expansion coefficients between fiber and matrix. The shear strength was 30.5 MPa. The fracture toughness and work of fracture were as high as 20.3 MPa m"2 and 12.0 kJ m-2, respectively. The composites exhibited excellent uniformity of strength and the Weibull modulus, m, was 23. 3. The value of dynamic fracture toughness was 62- measured by Charpy impact tests. C 2001 Elsevier Science B.v. All rights reserved Keywords: 3D textile C/SiC composites; Toughness; Reliability; Chemical vapor infiltration 1. Introduction rently being applied to the extent that they could be Even when they have been employed, relatively low Continuous fiber reinforced ceramic matrix com stress applications and large safety factors were usually posites(CFCCs)are very interesting structural materi considered. The main reason is the difficulty and uncer- als because of their higher performance compared with tainty that exist in determining their failure strength, super-alloy at elevated temperatures, and higher frac fracture toughness, operating lifetime in severe condi ture toughness compared with monolithic ceramics [I tions because the nature of the deformation and failure potential to be used in advanced aero-engines. Among 13,14 of the composites were very complicated 4]. For this reason, CFCCs are considered as the most these CFCCs. both carbon fiber and silicon carbide This paper examined the mechanical properties over fiber reinforced silicon carbide composites(C/SiC and a large temperature range of 3D textile C/SiC com- posite materials produced by chemical vapor infiltra- SiC/SiC)are most promising and have been received tion. The aims of current contribution are (1)to considerable interest [1, 5-8. Many investigations hav develop the understanding of the effects of architecture been conducted on two dimensional woven C/SiC and on the mechanical properties and the damage behavior SiC/SiC composite materials. Recently, attention has of the composites, (2) to expand the experimental been focused on three dimensional woven or braided knowledge for the design of the three dimensional ceramic matrix composite materials in order to meet textile composite materials mechanical and thermal properties requirements along he thickness of the composites [9-12] espite the attractiveness of fiber-reinforced com- 2. Materials and experimental procedures tes as engineering components, they are not cu 2. 1. Fabrication of the composites orresponding author. Tel. +86-29-8491427: fax: +86-29. 8491000 PAN-based carbon fiber was employed and each yarn contained 3000 filaments. The three-dimensional 0921-5093/01/s- see front matter o 2001 Elsevier Science B.V. All rights reserved PI:S0921-509300)01533-1
Materials Science and Engineering A300 (2001) 196–202 Mechanical properties of 3D fiber reinforced C/SiC composites Yongdong Xu *, Laifei Cheng, Litong Zhang, Hongfeng Yin, Xiaowei Yin State Key Laboratory of Solidification Processing, Northwestern Polytechnical Uni6ersity, Xian, Shaanxi 710072, People’s Republic of China Received 9 May 2000; received in revised form 17 August 2000 Abstract High toughness and reliable three dimensional textile carbon fiber reinforced silicon carbide composites were fabricated by chemical vapor infiltration. Mechanical properties of the composite materials were investigated under bending, shear, and impact loading. The density of the composites was 2.0–2.1 g cm−3 after the three dimensional carbon preform was infiltrated for 30 h. The values of flexural strength were 441 MPa at room temperature, 450 MPa at 1300°C, and 447 MPa at 1600°C. At elevated temperatures (1300 and 1600°C), the failure behavior of the composites became some brittle because of the strong interfacial bonding caused by the mis-match of thermal expansion coefficients between fiber and matrix. The shear strength was 30.5 MPa. The fracture toughness and work of fracture were as high as 20.3 MPa m1/2 and 12.0 kJ·m−2 , respectively. The composites exhibited excellent uniformity of strength and the Weibull modulus, m, was 23.3. The value of dynamic fracture toughness was 62 kJ·m−2 measured by Charpy impact tests. © 2001 Elsevier Science B.V. All rights reserved. Keywords: 3D textile C/SiC composites; Toughness; Reliability; Chemical vapor infiltration www.elsevier.com/locate/msea 1. Introduction Continuous fiber reinforced ceramic matrix composites (CFCCs) are very interesting structural materials because of their higher performance compared with super-alloy at elevated temperatures, and higher fracture toughness compared with monolithic ceramics [1– 4]. For this reason, CFCCs are considered as the most potential to be used in advanced aero-engines. Among these CFCCs, both carbon fiber and silicon carbide fiber reinforced silicon carbide composites(C/SiC and SiC/SiC) are most promising and have been received considerable interest [1,5–8]. Many investigations have been conducted on two dimensional woven C/SiC and SiC/SiC composite materials. Recently, attention has been focused on three dimensional woven or braided ceramic matrix composite materials in order to meet mechanical and thermal properties requirements along the thickness of the composites [9–12]. Despite the attractiveness of fiber-reinforced composites as engineering components, they are not currently being applied to the extent that they could be. Even when they have been employed, relatively low stress applications and large safety factors were usually considered. The main reason is the difficulty and uncertainty that exist in determining their failure strength, fracture toughness, operating lifetime in severe conditions because the nature of the deformation and failure behavior of the composites were very complicated [13,14]. This paper examined the mechanical properties over a large temperature range of 3D textile C/SiC composite materials produced by chemical vapor infiltration. The aims of current contribution are (1) to develop the understanding of the effects of architecture on the mechanical properties and the damage behavior of the composites, (2) to expand the experimental knowledge for the design of the three dimensional textile composite materials. 2. Materials and experimental procedures 2.1. Fabrication of the composites PAN-based carbon fiber was employed and each yarn contained 3000 filaments. The three-dimensional * Corresponding author. Tel.: +86-29-8491427; fax: +86-29- 8491000. E-mail address: ydxu@nwpu.edu.cn (Y. Xu). 0921-5093/01/$ - see front matter © 2001 Elsevier Science B.V. All rights reserved. PII: S0921-5093(00)01533-1
Y. Xu et al. Materials Science and Engineering 4300 (2001)196-202 500 400 200 1.0 3.0 Displacement, mm Fig. I. Stress-displacement curve for 3D C/SiC textile composites at 1600C (3-D) fabric preform was braided by four-step process- thinned to 75 um, dimpled to a center using Ar+ions ing. The fiber volume fraction was 40%. In the present at an incident angle of 15. The specimens were exam- experiment, isothermal /forced flow chemical vapor ined with an operated voltage of 200 kV. infiltration was employed to deposit pyrolytic carbon layer and silicon carbide which has been described previously in detail [Il, 12]. A carbon layer of 200 nm 3. Results and discussion thickness was deposited on the surface of carbon fiber as interfacial layer with butane prior to densification. 3. 1. Flexural loading Methytrichosilane(MTS, CH3,)was used for depo- sition of SiC and carried by bubbling hydrogen. Typical The mechanical properties of fiber reinforced com- onditions used for the densification of silicon carbide posite materials are largely governed by the stress matrix are 1100C, a hydrogen to MTS mol ratio of 10, transfer capability of fiber /matrix interface. The and a pressure of a few kPa. Argon(Ar) was employed interfacial compatibility is related to the interfacial as diluent gas to slow the rate of deposition shear stress. which characterizes the combination of stress necessary to de-bond the interface and the fric- 2. 2. Mechanical properties measurement tional forces developed at the interface. The density of Mechanical properties of the composite materials vere characterized under bending, shear, and impact loading. Flexural strength was measured using the three-point-bending method at temperatures from room temperature up to the elevated temperatures(1300 and 1600C)in vacuum. Shear strength was measured using the short beam bending method with a span of 15 mm edged-notched beam method. The impact tests gle fracture toughness was determined with the sir performed with instrumented Charpy equipment. The sample size was3.0×20×70mm, and the impact velocity of 3 ms-1 2.3. Microstructure observation and surface analysi The density of the samples was determined by the water displacement method. The microstructure of frac- ture surface was observed by a scanning electron micro- scope. Transmission electron microscope samples were prepared by cutting 300 um thickness using a low speed diamond saw. The composites were mechanically Fig. 2. Fiber fracture in the SiC matrix at 1600C
Y. Xu et al. / Materials Science and Engineering A300 (2001) 196–202 197 Fig. 1. Stress–displacement curve for 3D C/SiC textile composites at 1600°C. (3-D) fabric preform was braided by four-step processing. The fiber volume fraction was 40%. In the present experiment, isothermal/forced flow chemical vapor infiltration was employed to deposit pyrolytic carbon layer and silicon carbide, which has been described previously in detail [11,12]. A carbon layer of 200 nm thickness was deposited on the surface of carbon fiber as interfacial layer with butane prior to densification. Methytrichosilane (MTS, CH3SiCl3) was used for deposition of SiC and carried by bubbling hydrogen. Typical conditions used for the densification of silicon carbide matrix are 1100°C, a hydrogen to MTS mol ratio of 10, and a pressure of a few kPa. Argon (Ar) was employed as diluent gas to slow the rate of deposition. 2.2. Mechanical properties measurement Mechanical properties of the composite materials were characterized under bending, shear, and impact loading. Flexural strength was measured using the three-point-bending method at temperatures from room temperature up to the elevated temperatures (1300 and 1600°C) in vacuum. Shear strength was measured using the short beam bending method with a span of 15 mm. Fracture toughness was determined with the single edged-notched beam method. The impact tests were performed with instrumented Charpy equipment. The sample size was 3.0×20×70 mm, and the impact velocity of 3 ms−1 . 2.3. Microstructure obser6ation and surface analysis The density of the samples was determined by the water displacement method. The microstructure of fracture surface was observed by a scanning electron microscope. Transmission electron microscope samples were prepared by cutting 300 mm thickness using a low speed diamond saw. The composites were mechanically thinned to 75 mm, dimpled to a center using Ar+ ions at an incident angle of 15°. The specimens were examined with an operated voltage of 200 kV. 3. Results and discussion 3.1. Flexural loading The mechanical properties of fiber reinforced composite materials are largely governed by the stresstransfer capability of fiber/matrix interface. The interfacial compatibility is related to the interfacial shear stress, which characterizes the combination of stress necessary to de-bond the interface and the frictional forces developed at the interface. The density of Fig. 2. Fiber fracture in the SiC matrix at 1600°C
Y. Xu et al. Materials Science and Engineering 4300 (2001)196-202 262228KV 23∠1mmD3 Fig 3. Fracture surface of 3D C/Sic composites under fiexural loading. 261728KV X251mD38 Fig 4. Fracture surface of 3D C/SiC composites under shear loading 200 8120 Displacement, mm Fig. 5. Failure behavior of notched 3D C/Sic composites with a notch
198 Y. Xu et al. / Materials Science and Engineering A300 (2001) 196–202 Fig. 3. Fracture surface of 3D C/SiC composites under flexural loading. Fig. 4. Fracture surface of 3D C/SiC composites under shear loading. Fig. 5. Failure behavior of notched 3D C/SiC composites with a notch
Y. Xu et al. Materials Science and Engineering 4300 (2001)196-202 Table I glass phase at the grain boundary. For C/Sic com Flexural strength data for 3D C/SiC composites posites, there is no glass phase in materials. Hence, the Number dr MPa flexural strength of the composites was nearly constant when the temperature ranged from room temperature up to 1600C. The average values flexural strength were 441 MPa at room temperature, 450 MPa at 1300oC and 447 MPa at 1600C. From Fig. l. it was also observed that the failure behavior of the composites was varied with the increase of the temperature. At room temperature, the stress drop was very gradual 435 after the maximum stress point. This observation sug gested that the fracture energy of the materials was very high. However, the failure behavior became brittle and the composites exhibited steep stress drops after the maximum stress point at high temperatures(Fig. Ib and c) The variation of failure behavior of composites was caused by alteration of the interfacial bonding between fiber and matrix. The t300 carbon fiber is anan- 475 sotropic material and usually characterized by two thermal expansion coefficients (TECs), a radial TEC (7.0 x 10-6oC-)and a longitudinal TEC(-01 to A prefor cm-3 after the three infiltrated SiC matrIx Is 4.8 10-6oC-[15, 16]. Hence, the composites was 2.0-2.1 1.1 x 10-6C-). The TEc of chemical vapor dimensional carbon m was chemical vapor the tensile stress within the interfacial phase along the infiltrated for 30 h. Fig. la shows the typical failure fiber radial direction was generated after the composites behavior of 3D C/SiC textile composites at room tem- were cooled down from the infiltration temperature perature,which is different from that of monolithic(1100 C)to room temperature. It was easy for the ceramIcs The present composite materials exhibited the carbon fiber to debond and be pulled out from the substantial non-linear failure behavior. From the silicon carbide matrix. Above the infiltration tempera stress-displacement curve, it could be observed that the ture, however, the interfacial stress became compressive failure of 3D C/SiC composites occurred in a controlled which led to the tight bond between the fiber and manner. In addition, the materials exhibited significant matrix. In addition, the tensile stress was generated in ailure deflection (1.2 mm) the carbon fiber. In such case. the carbon fiber was In general, the strength of monolithic ceramics(such easily damaged in the silicon carbide matrix as illus- as silicon nitride, mullite) decreases significantly at ele- trated in Fig. 2. As a result, the composites exhibited vated temperature due to the softening and sliding of brittle failure behavior because it was difficulty for the 234 95 Fig. 6. Weibull plot for flexural strength of 3 C/Sic composites
Y. Xu et al. / Materials Science and Engineering A300 (2001) 196–202 199 Table 1 Flexural strength data for 3D C/SiC composites sf , MPaNumber 3941 4062 4183 4 425 4265 4276 7 430 4358 4359 10 438 44011 44312 44413 44914 45915 45916 47017 18 473 47519 47820 glass phase at the grain boundary. For C/SiC composites, there is no glass phase in materials. Hence, the flexural strength of the composites was nearly constant when the temperature ranged from room temperature up to 1600°C. The average values flexural strength were 441 MPa at room temperature, 450 MPa at 1300°C, and 447 MPa at 1600°C. From Fig. 1, it was also observed that the failure behavior of the composites was varied with the increase of the temperature. At room temperature, the stress drop was very gradual after the maximum stress point. This observation suggested that the fracture energy of the materials was very high. However, the failure behavior became brittle and the composites exhibited steep stress drops after the maximum stress point at high temperatures (Fig. 1b and c). The variation of failure behavior of composites was caused by alteration of the interfacial bonding between fiber and matrix. The T300 carbon fiber is an anisotropic material and usually characterized by two thermal expansion coefficients (TECs), a radial TEC (7.0×10−6 °C−1 ) and a longitudinal TEC (−0.1 to 1.1×10−6 °C−1 ). The TEC of chemical vapor infiltrated SiC matrix is 4.8×10−6 °C−1 [15,16]. Hence, the tensile stress within the interfacial phase along the fiber radial direction was generated after the composites were cooled down from the infiltration temperature (1100°C) to room temperature. It was easy for the carbon fiber to debond and be pulled out from the silicon carbide matrix. Above the infiltration temperature, however, the interfacial stress became compressive which led to the tight bond between the fiber and matrix. In addition, the tensile stress was generated in the carbon fiber. In such case, the carbon fiber was easily damaged in the silicon carbide matrix as illustrated in Fig. 2. As a result, the composites exhibited brittle failure behavior because it was difficulty for the the composites was 2.0–2.1 g cm−3 after the three dimensional carbon preform was chemical vapor infiltrated for 30 h. Fig. 1a shows the typical failure behavior of 3D C/SiC textile composites at room temperature, which is different from that of monolithic ceramics. The present composite materials exhibited the substantial non-linear failure behavior. From the stress–displacement curve, it could be observed that the failure of 3D C/SiC composites occurred in a controlled manner. In addition, the materials exhibited significant failure deflection (1.2 mm). In general, the strength of monolithic ceramics (such as silicon nitride, mullite) decreases significantly at elevated temperature due to the softening and sliding of a Fig. 6. Weibull plot for flexural strength of 3D C/SiC composites
Y. Xu et al. Materials Science and Engineering 4300 (2001)196-202 SiC matrix 100nm 0999415KV84”忘 Fig. 7. Insufficient uniformity of interfacial layer. fiber to be pulled out from the matrix. Moreover, the bonding is usually considered as a kind of weak interfa composites showed significant non-linear failure behav- cial bonding because of the residual pores in the com- ior at 1600 C, which was attributed to creep of the posites caused by the bottom neck effectduring the manometer grain size of silicon carbide matrix. The chemical vapor infiltration process. Accordingly, both mis-match along the fiber axis of TECs between the fiber pull-out and bundle pull-out were observed at silicon carbide matrix and the fiber resulted in many room temperature(Fig. 3). At elevated temperatures micro-cracks in the matrix. It is believed that these 1300 and 1600 C), the fiber /matrix interfacial bonding micro-cracks have some contribution to the non-linear became strong but the bundle/bundle interfacial bond ailure behavior of the materials by deflection of th ing was still weak enough. In this case, the bundle pull-out was dominated, which resulted in the brittle Microstructural observations revealed that the failure failure behavior at high temperatures behavior of 3D textile C/SiC composites was domi- nated by the damage mode. As discussed above, the interfacial bonding between the carbon fiber and silicon 3. 2. Failure behavior of notched specimen under shear carbide matrix was dependent on the properties of loadin interfacial phase and temperature. However, the inter- facial bonding between fiber bundle and bundle was Shear strength of 3D C/SiC composites was mea- only influenced by the density of the composites but sured by the short shear beam method of three-point- independent of the temperature. For the three dimen- bending. The shear strength was calculated by the sional textile CFCCs, the bundle/bundle interfacial following equation [17]
200 Y. Xu et al. / Materials Science and Engineering A300 (2001) 196–202 Fig. 7. Insufficient uniformity of interfacial layer. Fig. 8. Impact fracture surface of C/SiC composites. fiber to be pulled out from the matrix. Moreover, the composites showed significant non-linear failure behavior at 1600°C, which was attributed to creep of the manometer grain size of silicon carbide matrix. The mis-match along the fiber axis of TECs between the silicon carbide matrix and the fiber resulted in many micro-cracks in the matrix. It is believed that these micro-cracks have some contribution to the non-linear failure behavior of the materials by deflection of the main crack. Microstructural observations revealed that the failure behavior of 3D textile C/SiC composites was dominated by the damage mode. As discussed above, the interfacial bonding between the carbon fiber and silicon carbide matrix was dependent on the properties of interfacial phase and temperature. However, the interfacial bonding between fiber bundle and bundle was only influenced by the density of the composites but independent of the temperature. For the three dimensional textile CFCCs, the bundle/bundle interfacial bonding is usually considered as a kind of weak interfacial bonding because of the residual pores in the composites caused by the ‘bottom neck effect’ during the chemical vapor infiltration process. Accordingly, both fiber pull-out and bundle pull-out were observed at room temperature (Fig. 3). At elevated temperatures (1300 and 1600°C), the fiber/matrix interfacial bonding became strong but the bundle/bundle interfacial bonding was still weak enough. In this case, the bundle pull-out was dominated, which resulted in the brittle failure behavior at high temperatures. 3.2. Failure beha6ior of notched specimen under shear loading Shear strength of 3D C/SiC composites was measured by the short shear beam method of three-pointbending. The shear strength was calculated by the following equation [17]
Y. Xu et al. Materials Science and Engineering 4300 (2001)196-202 4bh (1) metal materials, coefficient of variance(Cv) is used to describe the scatter of the strength and is usually less where P is the fracture load (N), b and h are width and than 1.2%. where cv is defined as the ratio of mean- height, respectively squared error divided by average strength. Obviously, The shear behavior of stress-displacement curve was CV of 3D C/SiC composites(5. 1%)was as four times as similar to that of bending behavior and shown in F hat of metal materials la. The value obtained for shear strength was 30.5 It is well established that the strength of brittle MPa. Different from two dimensional cfcs and the naterials obeys Weibull distribution. For CFCCs, the laminated composites, no layer debonding was ob- three basic constituents(fiber, interphase, and matrix served in the present composites. TEM observation are essentially brittle and cracking involves defect-in- revealed shear fracture of carbon fiber in the com- duced random failures. Therefore, the strength statisti- sites(Fig. 4a). For 3D C/SiC composites, the typical cal distribution of 3D C/Sic composites can be shear fracture surface was illustrated in Fig. 4b described by using Weibull equation 3.3. Failure behavior of notched specimen under F(G)=1-e (≥0;a,u>0)(3) fiexural loading where f is the cumulative distribution function g is the In order to determine the fracture tol samples applied stress. m, Weibull modulus, is the most impor were notched and tested in a three-p using a tant measurement of materials which chara span of 40 mm and a cross-head of 0.05 scatter of the materials. Large values of m indicate mm'min-. The value of fracture toughness(K,s) was uniformity, while small m represents large scatter. ou is calculated by using of the Griffith expression the scale parameter or characteristic strength because the maximum density of the Weibull distribution is ac (2) located at the position of o=Ou Fig. 6 shows the result of Weibull plot of InIn(l-F) versus In o. It could where os is the fracture load, Y the geometrical factor be found from Fig. 6 that the flexural strength of 3D for an edge crack in a three-point bend beam, and c is C/Sic composites also obeys Weibull distribution. The he notch depth result is consistent with those of SicTiAl and Sic/Sic The fracture toughness calculated from the Eq(2) composites [19, 20]. The slope corresponding to Weibull was 20.3 MPa 2, which is much higher than that of modulus m is 23.3 monolithic ceramic materials(3-5 MPam ). It should For C/SiC composites, the scatter of flexural strength be pointed out that this value of Kle could not represent is influenced by three factors, i.e. porosity, interfacial he real toughness of the materials. The reason is that situation between fiber and matrix, and the extend of he Eq (2)is based on the linear elastic mechanics and damage of the fibers during fabrication. It is well not valid for the CFCCs with substantial non-linear known that residual pores in the texture are considered behavior as a kind of defects and act as failure source. In Usually, the work of fracture was also used to de- addition, the strength is decreased because the effective cribe toughness of the materials. It can be obtained bearing volume of the materials is decreased by the from the characteristic area under the load-displace- pores. In 3D C/Sic composites, the pores are usually ment curve divided by the cross section of the specimen. present in two kinds(i)small pores(10 um) between In order to determine the work of fracture effectively, fiber and fiber within the fiber bundle are controlled by we defined the characteristic area(A which started the spacing of fibers and volume fraction of fiber, (ii) from initial point to the 10% drop of the curve(Fig. 5). the large pores between fiber bundle and bundle (10- This gives an average work of fracture as high as 12.0 10 um)[21]. In CVI process, the infiltration of the kJ-m. This value is twice that of 2D laminated pores is mainly dependent on the diffusion of the ceramic matrix composites. This value is much higher gaseous reactant species. Due to the limitation of the han that of both the laminated silicon carbide matrix CVI process and structure complexity of fabrics, it is nitmlposites(4625 J'm-2)and over two orders of mag- very difficulty to fill in the pores completely The interface is very important for CFCCs during con nitride, 80J'm-[18 fabrication or from environmental attack during use. In general, the interface should be strong enough to allow 3.4. Reliability statistics of flexural strength load transfer from the matrix to the fibers under stress but weak enough so that an advancing matrix crack The flexural strength data of 3D C/SiC composites can be deflected by the fibers. However, the carbon re listed in Table 1. Average strength and mean- fiber is usually attacked by the chemical reaction with quared error were 441 and 22.3 MPa, respectively. For residual O, moisture, and HCI [22, 23]. In addition
Y. Xu et al. / Materials Science and Engineering A300 (2001) 196–202 201 t= 3P 4bh (1) where P is the fracture load (N), b and h are width and height, respectively. The shear behavior of stress–displacement curve was similar to that of bending behavior and shown in Fig. 1a. The value obtained for shear strength was 30.5 MPa. Different from two dimensional CFCCs and the laminated composites, no layer debonding was observed in the present composites. TEM observation revealed shear fracture of carbon fiber in the composites (Fig. 4a). For 3D C/SiC composites, the typical shear fracture surface was illustrated in Fig. 4b. 3.3. Failure beha6ior of notched specimen under flexural loading In order to determine the fracture toughness, samples were notched and tested in a three-point bend using a span of 40 mm and a cross-head speed of 0.05 mm·min−1 . The value of fracture toughness (K1c) was calculated by using of the Griffith expression, sc= K1c Y c (2) where sc is the fracture load, Y the geometrical factor for an edge crack in a three-point bend beam, and c is the notch depth. The fracture toughness calculated from the Eq. (2) was 20.3 MPa·m1/2 , which is much higher than that of monolithic ceramic materials (3–5 MPa·m1/2 ). It should be pointed out that this value of K1c could not represent the real toughness of the materials. The reason is that the Eq. (2) is based on the linear elastic mechanics and is not valid for the CFCCs with substantial non-linear behavior. Usually, the work of fracture was also used to describe toughness of the materials. It can be obtained from the characteristic area under the load–displacement curve divided by the cross section of the specimen. In order to determine the work of fracture effectively, we defined the characteristic area (Ac) which started from initial point to the 10% drop of the curve (Fig. 5). This gives an average work of fracture as high as 12.0 kJ·m−2 . This value is twice that of 2D laminated ceramic matrix composites. This value is much higher than that of both the laminated silicon carbide matrix composites (4625 J·m−2 ) and over two orders of magnitude greater than monolithic ceramic materials (silicon nitride, 80 J·m−2 )[18]. 3.4. Reliability statistics of flexural strength The flexural strength data of 3D C/SiC composites are listed in Table 1. Average strength and meansquared error were 441 and 22.3 MPa, respectively. For metal materials, coefficient of variance (CV) is used to describe the scatter of the strength and is usually less than 1.2%. Where CV is defined as the ratio of meansquared error divided by average strength. Obviously, CV of 3D C/SiC composites (5.1%) was as four times as that of metal materials. It is well established that the strength of brittle materials obeys Weibull distribution. For CFCCs, the three basic constituents (fiber, interphase, and matrix) are essentially brittle and cracking involves defect-induced random failures. Therefore, the strength statistical distribution of 3D C/SiC composites can be described by using Weibull equation, F(s)=1−exp − s su mn (s]0; s,su\0) (3) where F is the cumulative distribution function. s is the applied stress. m, Weibull modulus, is the most important measurement of materials, which characterizes the scatter of the materials. Large values of m indicate uniformity, while small m represents large scatter. su is the scale parameter or characteristic strength because the maximum density of the Weibull distribution is located at the position of s=su. Fig. 6 shows the result of Weibull plot of ln ln(1−F)−1 versus ln s. It could be found from Fig. 6 that the flexural strength of 3D C/SiC composites also obeys Weibull distribution. The result is consistent with those of SiC/TiAl and SiC/SiC composites [19,20]. The slope corresponding to Weibull modulus m is 23.3. For C/SiC composites, the scatter of flexural strength is influenced by three factors, i.e. porosity, interfacial situation between fiber and matrix, and the extend of damage of the fibers during fabrication. It is well known that residual pores in the texture are considered as a kind of defects and act as failure source. In addition, the strength is decreased because the effective bearing volume of the materials is decreased by the pores. In 3D C/SiC composites, the pores are usually present in two kinds (i) small pores (100 mm) between fiber and fiber within the fiber bundle are controlled by the spacing of fibers and volume fraction of fiber, (ii) the large pores between fiber bundle and bundle (102 – 103 mm) [21]. In CVI process, the infiltration of the pores is mainly dependent on the diffusion of the gaseous reactant species. Due to the limitation of the CVI process and structure complexity of fabrics, it is very difficulty to fill in the pores completely. The interface is very important for CFCCs during fabrication or from environmental attack during use. In general, the interface should be strong enough to allow load transfer from the matrix to the fibers under stress, but weak enough so that an advancing matrix crack can be deflected by the fibers. However, the carbon fiber is usually attacked by the chemical reaction with residual O, moisture, and HCl [22,23]. In addition,
Y. Xu et al. Materials Science and Engineering 4300 (2001)196-202 T-300 carbon fiber is very irregular and there are many 20.3 MPam/2 and 12.0 kJm-3, respectively. The com micro-depressions on its surface(Fig. 7). These micro- posites exhibited high uniformity of strength and the depressions have two kinds of effects on the com- Weibull modulus m was 23.3. The value of dynamic posites. On one hand, the micro-depressions lead to the fracture toughness was 62 kJ m-2 as measured by strong interfacial bond between fiber and matrix by the Charpy impact tests mechanical interlock of rough surface between fiber and matrix On the other hand. the insufficient unifor mity of interfacial layer thickness would result in the Acknowledgements scatter of flexural strengt The authors wish to thank the National natural 3.5.In Scientific Foundation of China. Chinese Aeronautics Foundation. and National Defense Foundation of Instrumented Charpy impact tests on un-notched China for the financial supports samples were conducted to determine the energy ab- sorbing capability and dynamic fracture behavior of the composite materials. The dynamic fracture toughness (ax) was calculated using the following equation, References △W [R. Naslain, CVI composites, in: R. Warren (Ed ) Ceramic Matrix Composites, Chapman Hall, London, 1992, p. 199 2 T.M. Besmann, B w. Sheldon, R.A. Lowden, Science 253(1991) Where Aw is the absorbing energy of materials during impact processing, b and h are the thickness and width 5] T Ishikawa, SKajii, K. Matsanaga, T.Hogan,Y. Kohtoku, T of specimen, respectively Nagasawa, Science 282(1998)1295-1297 (K.M. Prewo, Am. Ceram Soc. Bull. 68( 1989)395 The value of a is 62 kj m- for three dimensional 5 M. Wang, C. Laird, J. Mater. Sci. 31(1996)206 C/SiC composite materials, which could be compared [6].J. Brenann, Mater. Sci. Eng. Al26(1990)20 with that of super-alloy(ax=80-160 kJ'm-2). The [7 D Singh, J.P. Singh, M.J. Wheeler, J. Am. Ceram Soc impact fracture surface was shown in Fig. &a. It should (1996)591. be noted that the materials were still connected together [8S. Prouhet, G. Camus, C. Labrugee, A. Guette, J. Am. Ceram but not broken into pieces like iron metal. In addition, Soc.77(1994)649. 9 F K. Ko, Am. Ceram. Soc. Bull. 68(1989)401 the composites could also be penetrated by a steel nail P. Pluvinage, A P. Majidi, T.w. Chou, J. Mater. Sci. 31(1996) without severe fracture(Fig. 8b). It is obvious that three dimensional C/SiC composite materials exhibit an Y.D. xu, LT J Am Ceram Soc. 80(1997)1897. excellent impact damage tolerance L F. Cheng, D t. Yan, Carbon 36(1998) [13]M. Saki, J. Ceram Soc. Jpn. 99(10)(1991)983-992 [14 A.G. Evans, J. Am. Ceram Soc. 73(1990)187 4. Conclusions [ J. Lackey, J.A. Hanigofsky, G B. Freeman, R.D. Hardin, A Am. Ceram High toughness and reliable three dimensional textile [6] F Lamouroux, X. Bourrat, R. Naslain, Carbon 31(1993)1273 carbon fiber reinforced silicon carbide composites were [7R D. Wu, H.R. Sheng, Behavior of Textile Structural Comppp- osites, Science Press, Beijing, 1998, p. 205 obtained by chemical vapor infiltration. Mechanical [18]WJ. Clegg, K.M.N.Kendall,N. Alford, T.W. Button, JD properties of the composite materials were investigated Brichall, Nature 347(1990)455-457 under bending, shear, and impact loading. The density (9 R. Naslain, J. Lamon, R, Paller, X. Bourrat, A. Guette, F of the composites was 2.0-2. 1 g cm-3after the 3-D carbon preform was infiltrated for 30 h. 3D textile 20S. Ochiai, M. Hojo, M. Tanaka, Composites A30(1999)451 /SiC composites exhibited excellent mechanical prop- [21]Y Xu, L Cheng, L Zhang, Carbon 37(1999)1179 erties at temperatures from room temperature up to (22Y. Xu, L. Cheng, L Zhang, H Ying, C. You, J Mater. Sci., in 1600C. The values of flexural strength were 441 MPa t room temperature, 450 MPa at 1300 C, and 447 [23] T.M. Besmann, R.A. Lowden, R.A. Lowden, Overview of chem- ical vapor infiltration, in: R. Naslain(Ed. ) High Temperature MPa at 1600oC. The shear strength was 30.5 MPa. The Ceramic Matrix Composites, Woodhead Publications, Bordeaux fracture toughness and work of fracture were as high as 1993,p.215
202 Y. Xu et al. / Materials Science and Engineering A300 (2001) 196–202 T-300 carbon fiber is very irregular and there are many micro-depressions on its surface (Fig. 7). These microdepressions have two kinds of effects on the composites. On one hand, the micro-depressions lead to the strong interfacial bond between fiber and matrix by the mechanical interlock of rough surface between fiber and matrix. On the other hand, the insufficient uniformity of interfacial layer thickness would result in the scatter of flexural strength. 3.5. Impact loading Instrumented Charpy impact tests on un-notched samples were conducted to determine the energy absorbing capability and dynamic fracture behavior of the composite materials. The dynamic fracture toughness (ak) was calculated using the following equation, ak=DW bh (4) Where DW is the absorbing energy of materials during impact processing, b and h are the thickness and width of specimen, respectively. The value of ak is 62 kJ·m−2 for three dimensional C/SiC composite materials, which could be compared with that of super-alloy (ak=80–160 kJ·m−2 ). The impact fracture surface was shown in Fig. 8a. It should be noted that the materials were still connected together but not broken into pieces like iron metal. In addition, the composites could also be penetrated by a steel nail without severe fracture (Fig. 8b). It is obvious that three dimensional C/SiC composite materials exhibit an excellent impact damage tolerance. 4. Conclusions High toughness and reliable three dimensional textile carbon fiber reinforced silicon carbide composites were obtained by chemical vapor infiltration. Mechanical properties of the composite materials were investigated under bending, shear, and impact loading. The density of the composites was 2.0–2.1 g·cm−3 after the 3-D carbon preform was infiltrated for 30 h. 3D textile C/SiC composites exhibited excellent mechanical properties at temperatures from room temperature up to 1600°C. The values of flexural strength were 441 MPa at room temperature, 450 MPa at 1300°C, and 447 MPa at 1600°C. The shear strength was 30.5 MPa. The fracture toughness and work of fracture were as high as 20.3 MPa·m1/2 and 12.0 kJ·m−2 , respectively. The composites exhibited high uniformity of strength and the Weibull modulus m was 23.3. The value of dynamic fracture toughness was 62 kJ·m−2 as measured by Charpy impact tests. Acknowledgements The authors wish to thank the National Natural Scientific Foundation of China, Chinese Aeronautics Foundation, and National Defense Foundation of China for the financial supports. References [1] R. Naslain, CVI composites, in: R. Warren (Ed.), Ceramic Matrix Composites, Chapman & Hall, London, 1992, p. 199. [2] T.M. Besmann, B.W. Sheldon, R.A. Lowden, Science 253 (1991) 1104. [3] T. Ishikawa, S. Kajii, K. Matsanaga, T. Hogani, Y. Kohtoku, T. Nagasawa, Science 282 (1998) 1295–1297. [4] K.M. Prewo, Am. Ceram. Soc. Bull. 68 (1989) 395. [5] M. Wang, C. Laird, J. Mater. Sci. 31 (1996) 2065. [6] J.J. Brenann, Mater. Sci. Eng. A126 (1990) 203. [7] D. Singh, J.P. Singh, M.J. Wheeler, J. Am. Ceram. Soc. 79 (1996) 591. [8] S. Prouhet, G. Camus, C. Labrugee, A. Guette, J. Am. Ceram. Soc. 77 (1994) 649. [9] F.K. Ko, Am. Ceram. Soc. Bull. 68 (1989) 401. [10] P. Pluvinage, A.P. Majidi, T.W. Chou, J. Mater. Sci. 31 (1996) 232–241. [11] Y.D. Xu, L.T. Zhang, J. Am. Ceram. Soc. 80 (1997) 1897. [12] Y.D. Xu, L.T. Zhang, L.F. Cheng, D.T. Yan, Carbon 36 (1998) 1051. [13] M. Saki, J. Ceram. Soc. Jpn. 99 (10) (1991) 983–992. [14] A.G. Evans, J. Am. Ceram. Soc. 73 (1990) 187. [15] J. Lackey, J.A. Hanigofsky, G.B. Freeman, R.D. Hardin, A. Prasad, J. Am. Ceram. Soc. 78 (1995) 1564–1570. [16] F. Lamouroux, X. Bourrat, R. Naslain, Carbon 31 (1993) 1273. [17] R.D. Wu, H.R. Sheng, Behavior of Textile Structural Compppposites, Science Press, Beijing, 1998, p. 205. [18] W.J. Clegg, K.M.N. Kendall, N. Alford, T.W. Button, J.D. Brichall, Nature 347 (1990) 455–457. [19] R. Naslain, J. Lamon, R. Pailler, X. Bourrat, A. Guette, F. Langlais, Composites A30 (1999) 537–547. [20] S. Ochiai, M. Hojo, M. Tanaka, Composites A30 (1999) 451– 461. [21] Y. Xu, L. Cheng, L. Zhang, Carbon 37 (1999) 1179. [22] Y. Xu, L. Cheng, L. Zhang, H. Ying, C. You, J. Mater. Sci., in press. [23] T.M. Besmann, R.A. Lowden, R.A. Lowden, Overview of chemical vapor infiltration, in: R. Naslain (Ed.), High Temperature Ceramic Matrix Composites, Woodhead Publications, Bordeaux, 1993, p. 215.