Availableonlineatwww.sciencedirect.com SCIENCEDIRECT° COMPOSITES CIENCE AND TECHNOLOGY ELSEVIER Composites Science and Technology 66(2006)993-1000 www.elsevier.com/locate/compscitech Tensile creep behavior of notched two-dimensional-C/SiC composite Wu Xiaojun, Qiao Shengru, Hou Juntao, Zhao Qing, Han Dong, Li M Box 547, Department of Materials Science and Engineering of Northwestern Polytechnical Unirersity, 710072 Xian Shaanxi, China Received 15 November 2004: received in revised form 14 August 2005: accepted 16 August 2005 ailable online 4 October 2005 Abstract Tensile creep tests of two-dimensional-C/SiC specimens with double-edge arc notches have been carried out at 1100, 1300 and 1500C vacuum. The matrix cracks on the surface and resonance frequency were examined at different creeping times. At 1100 C, the creep strains of both smooth and notched specimens were concentrated at the transient stage and the steady creep rates were nearly zer whereas steady creep rates of notched specimens and smooth specimens were similar at 1500C. It has been observed that the creep dam- age mainly concentrated at the area near the notches Micro-cracks appeared in the area near the notches and on the cross-points of the woven fiber bundles, and the longitudinal fibers near the notches fractured easily. Both types of curves, namely quantity of micro-cracks vS. time and micro-crack width vs time, were extremely similar as for the creep curves. In general, micro-cracks developed fast during the first 10 h. It has been noticed that within the first 2 h, the micro-cracks near the notches grow faster than those far from the notches, hereas the growth rate of micro-cracks far from notches was faster than those near the notches after 2 h. This phenomenon indicates the stress redistribution during creep. Damage curves at 1300 and 1500 C have similar trend, though the damage and the quantity of micro-cracks at 1500oC are higher than those at 1300oC o 2005 Elsevier Ltd. All rights reserved Keywords: 2D-C/SiC; Notch; Tensile creep: Micro-cracks; Damage 1. Introduction ites. Current work intends to study creep beh notched 2D-C/SiC composites in details The high-temperature creep performance of ID, 2D 2.5D and 3D C/SiC composites has been investigated. 2. Experiments Within the experimental temperature range(<1500C) the creep behavior of C fibers and Sic matrix is not obvi- 2.1. Materials and specimens S. The creep behavior of these composites is mainly caused by damage, such as matrix cracking, fiber fracture, Cross-woven C/SiC composites have been manufac fiber/matrix interfacial debonding, interfacial sliding, etc. tured here. PAN T-300 carbon fibers were woven into The elastic modulus of these composites decreases signifi- cross-woven carbon cloth, and then stacked into pre- cantly because of creep damage. Matrix cracking and inter- form. Each fiber bundle is composed of about 3000 car facial sliding are the main contributors to creep strain bon fibers and each fiber has a diameter of 7 um.The 14. In many applications, the composite component carbon fiber preform was first coated with a layer of unavoidably has rounded notches, but there is little knowl- pyrocarbon with the thickness of 0.2 um by chemical edge about creep behavior of notched 2D-C/SiC compos- vapor deposition (CVD), and then the Sic matrix was deposited by a chemical vapor infiltration(CVI) proces The composite contains 40 vol% fiber and 17 vol% poros- ity, the density is about 2.01 g/cm. The creep specimens ing author. Tel /fax: +02988492084. have the length of 80 mm and thickness of 3 mm, E-lmailaddresscml1072002(@163.com(w.xiaojun). shown in Fig. 1. All the specimens have double-edge 02663538/S. see front matter 2005 Elsevier Ltd. All rights reserved doi:10.1016j.compscitech.2005.08.008
Tensile creep behavior of notched two-dimensional-C/SiC composite Wu Xiaojun *, Qiao Shengru, Hou Juntao, Zhao Qing, Han Dong, Li Mei Box 547, Department of Materials Science and Engineering of Northwestern Polytechnical University, 710072 Xian Shaanxi, China Received 15 November 2004; received in revised form 14 August 2005; accepted 16 August 2005 Available online 4 October 2005 Abstract Tensile creep tests of two-dimensional-C/SiC specimens with double-edge arc notches have been carried out at 1100, 1300 and 1500 C in vacuum. The matrix cracks on the surface and resonance frequency were examined at different creeping times. At 1100 C, the creep strains of both smooth and notched specimens were concentrated at the transient stage and the steady creep rates were nearly zero, whereas steady creep rates of notched specimens and smooth specimens were similar at 1500 C. It has been observed that the creep damage mainly concentrated at the area near the notches. Micro-cracks appeared in the area near the notches and on the cross-points of the woven fiber bundles, and the longitudinal fibers near the notches fractured easily. Both types of curves, namely quantity of micro-cracks vs. time and micro-crack width vs. time, were extremely similar as for the creep curves. In general, micro-cracks developed fast during the first 10 h. It has been noticed that within the first 2 h, the micro-cracks near the notches grow faster than those far from the notches, whereas the growth rate of micro-cracks far from notches was faster than those near the notches after 2 h. This phenomenon indicates the stress redistribution during creep. Damage curves at 1300 and 1500 C have similar trend, though the damage and the quantity of micro-cracks at 1500 C are higher than those at 1300 C. 2005 Elsevier Ltd. All rights reserved. Keywords: 2D-C/SiC; Notch; Tensile creep; Micro-cracks; Damage 1. Introduction The high-temperature creep performance of 1D, 2D, 2.5D and 3D C/SiC composites has been investigated. Within the experimental temperature range (<1500 C), the creep behavior of C fibers and SiC matrix is not obvious. The creep behavior of these composites is mainly caused by damage, such as matrix cracking, fiber fracture, fiber/matrix interfacial debonding, interfacial sliding, etc. The elastic modulus of these composites decreases signifi- cantly because of creep damage. Matrix cracking and interfacial sliding are the main contributors to creep strain [1–4]. In many applications, the composite component unavoidably has rounded notches, but there is little knowledge about creep behavior of notched 2D-C/SiC composites. Current work intends to study creep behavior of notched 2D-C/SiC composites in details. 2. Experiments 2.1. Materials and specimens Cross-woven C/SiC composites have been manufactured here. PAN T-300 carbon fibers were woven into cross-woven carbon cloth, and then stacked into preform. Each fiber bundle is composed of about 3000 carbon fibers and each fiber has a diameter of 7 lm. The carbon fiber preform was first coated with a layer of pyrocarbon with the thickness of 0.2 lm by chemical vapor deposition (CVD), and then the SiC matrix was deposited by a chemical vapor infiltration (CVI) process. The composite contains 40 vol% fiber and 17 vol% porosity, the density is about 2.01 g/cm3 . The creep specimens have the length of 80 mm and thickness of 3 mm, as shown in Fig. 1. All the specimens have double-edge 0266-3538/$ - see front matter 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.compscitech.2005.08.008 * Corresponding author. Tel./fax: +02988492084. E-mail address: cm11072002@163.com (W. Xiaojun). www.elsevier.com/locate/compscitech Composites Science and Technology 66 (2006) 993–1000 COMPOSITES SCIENCE AND TECHNOLOGY
Xiaojun et al. / Composites Science and Technology 66(2006)993-1000 thickness 3 500℃ 170MPa 排 0.02 100℃,205MPa Fig. 1. Schematic of tensile creep specimen(in mm). arc notches, and each notch has 2mm radius and and 1500 C with different level of applied stress( Arrow indicates failure) 0.5 mm depth and the creep rate reduces constantly. In the stationary 2.2. Test procedure stage, at 1100oC, the creep rate tends to zero in the cases of both the notched specimen with 95 MPa applied nomi- The tensile creep tests were performed on an ultra-high- nal stress and the smooth specimen with 205 MPa applied temperature creep machine in 10-Pa vacuum, at 1100, nominal stress. However, at 1500oC, both the creep curves 1300 and 1500C Creep clamps are made of 3 D-C/C com- of the notched specimen with 95 MPa applied nominal posite,creep deformations were monitored by LVDT,and stress and the smooth specimen with 170 MPa applied temperatures were measured by W-Re thermo-couples. nominal stress are composed of a series of steps whose po- The surfaces of several specimens were polished for the sition rises constantly with creeping time, and their creep morphology observation on the HITACHI S-570 scanning rate is nearly in the same magnitude(10-/h). At current electron microscop M). At defined times of creep test, experimental period of creeping time(<100 h), no tertiary such as 0,0.5, 2, 10, 25 and 50 h, the matrix cracks on the stage has been observed, no matter the samples were rup- surface and resonance frequency of the polished specimens tured or not were examined. The amount of cracks was counted and the widths of cracks were measured statistically under SEM, 3.2. sem observation thus their variations with creeping time were obtained. In the mean time, resonance frequencies of these specimens Micrographs of the near notch area and the area far from have also been measured by the VS-300-2 type electronic the notch at 95 MPa applied stress and 1500 C are shown in vibration platform and WFC-3 type electric vortex mea- Figs. 4 and 5, respectively. During the process of creeping, there were a large number of transverse(perpendicular to the loading axis)micro-cracks and a few longitudinal(par- 3. Experimental results llel to the loading direction) micro-cracks appeared on the surface of the specimen. With the increase of creeping 3.1. Creep curves time. the number of transverse micro-cracks increased inten- sively, and the width of micro-cracks also increased, though Creep curves of notched and non-notched specimens are the number of the longitudinal micro-cracks only increased presented in Figs. 2 and 3, respectively. Note that the creep slightly By focusing on fixed areas(their magnitude of areas curves are made up of a series of steps, and can roughly be are the same as Figs. 4 and 5)and counting the observed divided into a transient and a stationary stage. This is in amount of micro-cracks. the statistical result of micro agreement with the literature on non-notched 2D-C/Sic cracks vs creep time curves have been plotted in Fig.6.Note [5]. In transient creep stage, creep strain increases rapidly that in general the number of the matrix micro-cracks con- stantly increased and their width expanded with creeping 0.04 1500°C time. During the first several hours of creeping, the amount of micro-cracks increased rapidly and the width of the crack opened quickly. As creeping time increased the amount of micro-cracks gradually saturated and the crack width 80.02 1100°C reached a steady value. A typical matrix micro-crack shown in Fig. 7 has been selected for the measurement of the crack width vs creeping time, and the result is shown 010 in Fig 8. Comparing the curves in Figs. 6 and 8 with the creeping curve of notched samples which were obtained un- der the same conditions in Fig. 2, the statistical curves are 2. Creep curves of notched 2- C/SiC specimens in vacuum and with very similar with the creeping curve itself and they all MPa applied stress at 1100 and 1500C showed similar development tendency with time
arc notches, and each notch has 2 mm radius and 0.5 mm depth. 2.2. Test procedure The tensile creep tests were performed on an ultra-hightemperature creep machine in 104 Pa vacuum, at 1100, 1300 and 1500 C. Creep clamps are made of 3D-C/C composite, creep deformations were monitored by LVDT, and temperatures were measured by W–Re thermo-couples. The surfaces of several specimens were polished for the morphology observation on the HITACHI S-570 scanning electron microscopy (SEM). At defined times of creep test, such as 0, 0.5, 2, 10, 25 and 50 h, the matrix cracks on the surface and resonance frequency of the polished specimens were examined. The amount of cracks was counted and the widths of cracks were measured statistically under SEM, thus their variations with creeping time were obtained. In the mean time, resonance frequencies of these specimens have also been measured by the VS-300-2 type electronic vibration platform and WFC-3 type electric vortex measurement instrument with the resolution of 1 Hz. 3. Experimental results 3.1. Creep curves Creep curves of notched and non-notched specimens are presented in Figs. 2 and 3, respectively. Note that the creep curves are made up of a series of steps, and can roughly be divided into a transient and a stationary stage. This is in agreement with the literature on non-notched 2D-C/SiC [5]. In transient creep stage, creep strain increases rapidly and the creep rate reduces constantly. In the stationary stage, at 1100 C, the creep rate tends to zero in the cases of both the notched specimen with 95 MPa applied nominal stress and the smooth specimen with 205 MPa applied nominal stress. However, at 1500 C, both the creep curves of the notched specimen with 95 MPa applied nominal stress and the smooth specimen with 170 MPa applied nominal stress are composed of a series of steps whose position rises constantly with creeping time, and their creep rate is nearly in the same magnitude (105 /h). At current experimental period of creeping time (<100 h), no tertiary stage has been observed, no matter the samples were ruptured or not. 3.2. SEM observation Micrographs of the near notch area and the area far from the notch at 95 MPa applied stress and 1500 C are shown in Figs. 4 and 5, respectively. During the process of creeping, there were a large number of transverse (perpendicular to the loading axis) micro-cracks and a few longitudinal (parallel to the loading direction) micro-cracks appeared on the surface of the specimen. With the increase of creeping time, the number of transverse micro-cracks increased intensively, and the width of micro-cracks also increased, though the number of the longitudinal micro-cracks only increased slightly. By focusing on fixed areas (their magnitude of areas are the same as Figs. 4 and 5) and counting the observed amount of micro-cracks, the statistical result of microcracks vs. creep time curves have been plotted in Fig. 6. Note that in general the number of the matrix micro-cracks constantly increased and their width expanded with creeping time. During the first several hours of creeping, the amount of micro-cracks increased rapidly and the width of the crack opened quickly. As creeping time increased, the amount of micro-cracks gradually saturated and the crack width reached a steady value. A typical matrix micro-crack as shown in Fig. 7 has been selected for the measurement of the crack width vs creeping time, and the result is shown in Fig. 8. Comparing the curves in Figs. 6 and 8 with the creeping curve of notched samples which were obtained under the same conditions in Fig. 2, the statistical curves are very similar with the creeping curve itself and they all showed similar development tendency with time. Fig. 1. Schematic of tensile creep specimen(in mm). Fig. 2. Creep curves of notched 2D-C/SiC specimens in vacuum and with 95 MPa applied stress at 1100 and 1500 C. Fig. 3. Creep curves of smooth 2D-C/SiC specimens in vacuum at 1100 and 1500 C with different level of applied stress (Arrow indicates failure). 994 W. Xiaojun et al. / Composites Science and Technology 66 (2006) 993–1000
w. Xiaojun et al./ Composites Science and Technology 66(2006)993-1000 H的B0hR0.5h 200um 2h国10h 25h顺AB50h 200um 200um ig. 4. Microcracks evolution at the notch front of 2D-C/Sic tensile creep along time at 1500C and 95 MPa in vacuum. Because of the stress concentration around the notch, gitudinal fiber bundles can cause serious damage and redis racks tend to appear more in this area, especially at notch tribution of stress. root, rupture has also been observed on the longitudinal fi- By using the same counting method, the amount of mi- rs near the notch root, as shown in Fig 9(a). Similarly, cro-cracks near the notch area has been counted during there is stress concentration around the weave porosity, creep course at 95 MPa and 1300C. The statistical re- and micro-cracks were easily observed near the porosity, sults of micro-cracks vs time on the top and on the lat- as seen in Fig. 9(b). Micro-cracks generally propagate eral side of the specimen are plotted in Fig. 10, the within bundles, as shown in Fig 9(c), though some of them results of 1500C at same stress level (95 MPa)have also may expand through the fiber interface and cause fracture been presented in the same graph. Note that all the statis- of longitudinal fiber bundles, as shown in Fig 9(d). As tical curves are very similar and can all be roughly divided the longitudinal fiber bundles of 2D-C/SiC normally bear into two stages: the fast developing stage of micro-cracks the main load applied on the sample [6, the fracture of lon- and the following slow increasing stage. The difference is
Because of the stress concentration around the notch, cracks tend to appear more in this area, especially at notch root, rupture has also been observed on the longitudinal fi- bers near the notch root, as shown in Fig. 9(a). Similarly, there is stress concentration around the weave porosity, and micro-cracks were easily observed near the porosity, as seen in Fig. 9(b). Micro-cracks generally propagate within bundles, as shown in Fig. 9(c), though some of them may expand through the fiber interface and cause fracture of longitudinal fiber bundles, as shown in Fig. 9(d). As the longitudinal fiber bundles of 2D-C/SiC normally bear the main load applied on the sample [6], the fracture of longitudinal fiber bundles can cause serious damage and redistribution of stress. By using the same counting method, the amount of micro-cracks near the notch area has been counted during creep course at 95 MPa and 1300 C. The statistical results of micro-cracks vs. time on the top and on the lateral side of the specimen are plotted in Fig. 10, the results of 1500 C at same stress level (95 MPa) have also been presented in the same graph. Note that all the statistical curves are very similar and can all be roughly divided into two stages: the fast developing stage of micro-cracks and the following slow increasing stage. The difference is Fig. 4. Microcracks evolution at the notch front of 2D-C/SiC tensile creep along time at 1500 C and 95 MPa in vacuum. W. Xiaojun et al. / Composites Science and Technology 66 (2006) 993–1000 995
w. Xiaojun et al. Composites Science and Technology 66(2006)993-1000 雨Hahw05h 200pm 200m R的10h 200um 200um 时啊25h50h Fig. 5. Microcracks evolution far from the notch of 2D-C/SiC tensile creep along time at 1500C and 95 MPa in vacuum. that, in both situations-on top surface of the specimen locates at the root of the notches. The stress value at the and on the lateral side of the specimen, there are obvi- area far from notch is only 71.1 MPa. According to the sta ously more micro-cracks at 1500C than at 1300C tistical results of micro-cracks by SEM, the growth rate of micro-cracks during creep process is 3.3. The stress redistribution during creep process ANSYS Software was used to simulate initial longitudi- where u is the micro-cracks growth rate, s stands for quan nal stress distribution of notched specimen. Fig. ll shows tity of micro-cracks, t for creeping time. At 1500C and stress distribution immediately after loading, before creep 95 MPa, micro-crack growth rates at the area far from strain accumulated. Because of the existence of notches, notches and near the notches(observed areas are equal the initial stress level near the notches is higher than that have been calculated by Eq(1), and the results have been far from the notches and the maximum stress(132. 8 MPa) gathered in Table 1, where Di is micro-crack growth rate for
that, in both situations-on top surface of the specimen and on the lateral side of the specimen, there are obviously more micro-cracks at 1500 C than at 1300 C. 3.3. The stress redistribution during creep process ANSYS software was used to simulate initial longitudinal stress distribution of notched specimen. Fig. 11 shows stress distribution immediately after loading, before creep strain accumulated. Because of the existence of notches, the initial stress level near the notches is higher than that far from the notches and the maximum stress (132.8 MPa) locates at the root of the notches. The stress value at the area far from notch is only 71.1 MPa. According to the statistical results of micro-cracks by SEM, the growth rate of micro-cracks during creep process is v ¼ Ds=Dt; ð1Þ where v is the micro-cracks growth rate, s stands for quantity of micro-cracks, t for creeping time. At 1500 C and 95 MPa, micro-crack growth rates at the area far from notches and near the notches (observed areas are equal) have been calculated by Eq. (1), and the results have been gathered in Table 1, where v1 is micro-crack growth rate for Fig. 5. Microcracks evolution far from the notch of 2D-C/SiC tensile creep along time at 1500 C and 95 MPa in vacuum. 996 W. Xiaojun et al. / Composites Science and Technology 66 (2006) 993–1000
w. Xiaojun et al./ Composites Science and Technology 66(2006)993-1000 z8E5品3 of microcracks at front z巴E quantity of microcracks at the side near notch - quantity of microcracks at front H quantity of microcracks at the far from notch side far from the notch Fig. 6. Amount of micro-cracks as a function of time for D-C/SiC tensile creep specimen at 1500C and 95 MPa in vacuum. (a)on the top surface of the specimen and (b) on the lateral side of the specimen 0.5h h Fig. 7. The crack width variation with creeping time in the near notch area and on the lateral side of the 2D-C/Sic tensile creep specimen at 1500C and the near notch areas on the top surface of the specimen, U2 10 is for near notch areas on the lateral side of the specimen D3 is for the areas far from the notch on the top surface, and D4 is for the areas far from the notch on the lateral side sur- face. Notice that the growth rate of micro-cracks decreases 2 constantly. The fastest growth stage of micro-cracks is 10203040 within 10 h: the growth rate of the micro-cracks of the near 50 60 notch area has been faster than that for the far notch area within 2 h, in contrast the growth rate of micro-cracks at Fig 8. Evolution of crack width as a function of time for a notched 2D.c/ the area far from notch has been faster than that of the Sic tensile creep specimen at 1500C and 95 MPa in vacuum. near notch area in the period of 2-10 h
the near notch areas on the top surface of the specimen, v2 is for near notch areas on the lateral side of the specimen, v3 is for the areas far from the notch on the top surface, and v4 is for the areas far from the notch on the lateral side surface. Notice that the growth rate of micro-cracks decreases constantly. The fastest growth stage of micro-cracks is within 10 h: the growth rate of the micro-cracks of the near notch area has been faster than that for the far notch area within 2 h, in contrast the growth rate of micro-cracks at the area far from notch has been faster than that of the near notch area in the period of 2–10 h. 0 0 10 20 30 40 50 60 quantity of microcracks at front near notch quantity of microcracks at front far from notch 0 10 20 30 40 50 0 20 40 Quantity of microcracks/ N quantity of microcracks at the side near notch quantity of microcracks at the side far from the notch 60 t (h) t (h) 10 20 30 40 50 60 Quantity of microcracks/ N a b Fig. 6. Amount of micro-cracks as a function of time for notched 2D-C/SiC tensile creep specimen at 1500 C and 95 MPa in vacuum, (a) on the top surface of the specimen and (b) on the lateral side of the specimen. Fig. 7. The crack width variation with creeping time in the near notch area and on the lateral side of the 2D-C/SiC tensile creep specimen at 1500 C and 95 MPa in vacuum, 50·. 0 2 4 6 8 10 12 0 10 20 30 40 50 60 t (h) Width of microcrack/ μm Fig. 8. Evolution of crack width as a function of time for a notched 2D-C/ SiC tensile creep specimen at 1500 C and 95 MPa in vacuum. W. Xiaojun et al. / Composites Science and Technology 66 (2006) 993–1000 997
w. Xiaojun et al Composites Science and Technology 66(2006)993-1000 -RB-1560-5 83628Kv氵5" b下362Kv氵iu酯 IG52D-C/SIC-RB-1591-5 [m52T-CZSIC-RB-15IR-5 00715Ky xi50 200 d 83628Kv Fig 9. SEM micrographs of 2D-C/SiC tensile crept specimen at the near notch area at 1500C and 95 MPa in vacuum:(a)near the notch of the specimen after 0.5 h creeping, 150x;(b) near the notch of the specimen after 0.5 h creeping, 150x;(c) near the notch on lateral side of the specimen after 0.5 h creeping. 150x;(c) near the notch on lateral side of the specimen after 25 h creeping, 150x;(d) near the notch of the specimen after 25 h creeping, 1000x 3.4. The damage evaluated by elastic modulus resistance and residual strength, etc. In the present work, the creep damage has been evaluated by elastic modulus There are many methods to evaluate the damage of cera- The resonance frequency of the sample is measured by mic matrix composites, such as elastic modulus, electrical stopping the creeping test at different creep times. As elastic modulus is proportional to the square of the resonance fre- quency, variation of elastic modulus could be obtained indirectly by varying the resonance frequency. The damage parameter, D, Is given 一一b D=1-(E/E0)=(-f2)后2, where f and E are the resonance frequency and elastic modulus of the damaged material at time I, respectively ,, fo and Eo are the resonance frequency and elastic modu- lus of un-damaged material, respectively. Fig. 12 show the variation of damage with creeping time at different temperatures. It appears that the trend of damage curves Fig 10. The amount of micro-cracks as a function of time for 2 D-C/Sic under 1300 and 1500oC are similar. At the transient tensile creep specimen at the near notch area at 1300 and 1500.C and creep stage, damage curves increase rapidly with time 95MPa in vacuum: (a)was measured on the top surface of the specimen at after experience a relatively slow and slightly decreasing )was mea (c)was measured on the top surface of the specimen at 1300C, (d)was development stage, the damage curves start to increase measured on the lateral side of the specimen at 1300C agaln
3.4. The damage evaluated by elastic modulus There are many methods to evaluate the damage of ceramic matrix composites, such as elastic modulus, electrical resistance and residual strength, etc. In the present work, the creep damage has been evaluated by elastic modulus. The resonance frequency of the sample is measured by stopping the creeping test at different creep times. As elastic modulus is proportional to the square of the resonance frequency, variation of elastic modulus could be obtained indirectly by varying the resonance frequency. The damage parameter, D, is given by D ¼ 1 ðE=E0Þ¼ðf 2 0 f 2 Þ=f 2 0 ; ð2Þ where f and E are the resonance frequency and elastic modulus of the damaged material at time t, respectively; f0 and E0 are the resonance frequency and elastic modulus of un-damaged material, respectively. Fig. 12 shows the variation of damage with creeping time at different temperatures. It appears that the trend of damage curves under 1300 and 1500 C are similar. At the transient creep stage, damage curves increase rapidly with time; after experience a relatively slow and slightly decreasing development stage, the damage curves start to increase again. Fig. 9. SEM micrographs of 2D-C/SiC tensile crept specimen at the near notch area at 1500 C and 95 MPa in vacuum: (a) near the notch of the specimen after 0.5 h creeping, 150·; (b) near the notch of the specimen after 0.5 h creeping, 150·; (c) near the notch on lateral side of the specimen after 0.5 h creeping, 150·; (c) near the notch on lateral side of the specimen after 25 h creeping, 150·; (d) near the notch of the specimen after 25 h creeping, 1000·. 0 10 20 30 40 50 60 0 10 20 30 40 50 60 t (h) Quantity of microcracks a b c d Fig. 10. The amount of micro-cracks as a function of time for 2D-C/SiC tensile creep specimen at the near notch area at 1300 and 1500 C and 95 MPa in vacuum: (a) was measured on the top surface of the specimen at 1500 C, (b) was measured on the lateral side of the specimen at 1500 C, (c) was measured on the top surface of the specimen at 1300 C, (d) was measured on the lateral side of the specimen at 1300 C. 998 W. Xiaojun et al. / Composites Science and Technology 66 (2006) 993–1000
w. Xiaojun et al./ Composites Science and Technology 66(2006)993-1000 999 indicates that during the creeping course the stress of near he maximum notch field reduced to some extent whereas the stress in the stress at the far field from notch increased. The reason of this phenom notch root enon is that the cracking in the matrix relaxed the stress concentration around the notches and then the relaxation changed the creeping stress in many positions on the notched sample as the creeping time increased. Similar sults have been reported for other materials [7,8] Even though 2D-C/SiC is cross-woven composite, the fi- bers can still be misaligned with respect to the loading direc tion, therefore fibers and fiber bundles were not fully F=86.6 MPa stretched in the matrix. During the process of creeping, G=102 MP: the longitudinal fibers have been stretched gradually and H=117. 4 MPa 1=132.8MPa tend to be realigned along the stress axis, and this caused the increase of the elastic modulus. as the creased, the creep damage accumulated and caused the drop of the elastic modulus. These two types of process devel- Stress at far from oped separately and competed with each other. After cer- the notch tain creeping time, fibers have been gradually straightened and oriented along the applied stress direction. This proces I indicates the maximum value and e indicates of stress value far from contributed the most to increase the elastic modulus. The development of damage is relatively slow at this stage, it only slows down the increase of elastic modulus to some Growth rate of micro-cracks at different creep time for 2D. c/Sic tensile extent and induced slight drop of the damage parameter D creep test at 1500C and 95 MPa in vacuum As the creeping test went on, fibers were further straight- 0-2h) 0-25(h) ened and oriented along the stress direction till the ela 4 On the other hand continuous development of the creeping damage caused the further increase of the damage parameter D, and this is associated with the constant dropping of the elastic mod This is th In both notched and smooth samples, creeping rates -1300°c were almost zero at 1100C, though creep curves at 1500G 1500C for both notched and smooth specimen were com posed of a series of obvious steps which continuously rose 20.2 with creeping time. When the micro-damage accumulation reached to certain extent, the lager damage took place which leads a step in creep curves. At the same temperature the creeping rate of notched sample at 95 MPa and smooth sample at 170 MPa are at the same level (10/h).There- fore, it is obvious that temperature, other than the applied stress level, can influence the stable creeping rate the most At 95 MPa, the comparison of damage curves in Fig 12 evolution as a function of time for notched 2D-C/sic and the crack quantity curves in Fig. 10 shows that these specimens at 95 MPa in vacuum. curves are very similar and in all the curves there is a fast hanging stage and a slow changing stage. In the slow 4. Discussion changing stage, damage value at 1500C was always larg than that at 1300 oC. the amount of cracks at 1500oC was As the creeping time increased, micro-crack growth also higher than that at 1300C. This means that damage rates near the notch and far from it both reduced con- and crack tend to happen more in the creeping specimens stantly and became closer. When micro-cracks were reach- at 1500C than at 1300C. Matrix cracking can cause ing saturation, the amount of micro-cracks near the notch the fiber fracture, and the fiber fracture leads the stress was obviously higher than that far from notch. Within 2 h, redistribution and more matrix cracking. Fiber fracture the growth rate of micro-cracks at the near notch fields was and matrix cracking are the main damage forms of C/Sic faster than that far from notches, in contrast the growth composite [24]. rate of micro-cracks far from notches was faster than that The mismatch of thermal expansion between C fiber and of near notch fields within the time period of 2-10 h. It SiC matrix cause thermal residual stresses in the composite
4. Discussion As the creeping time increased, micro-crack growth rates near the notch and far from it both reduced constantly and became closer. When micro-cracks were reaching saturation, the amount of micro-cracks near the notch was obviously higher than that far from notch. Within 2 h, the growth rate of micro-cracks at the near notch fields was faster than that far from notches, in contrast the growth rate of micro-cracks far from notches was faster than that of near notch fields within the time period of 2–10 h. It indicates that during the creeping course the stress of near notch field reduced to some extent, whereas the stress in the far field from notch increased. The reason of this phenomenon is that the cracking in the matrix relaxed the stress concentration around the notches, and then the relaxation changed the creeping stress in many positions on the notched sample as the creeping time increased. Similar results have been reported for other materials [7,8]. Even though 2D-C/SiC is cross-woven composite, the fi- bers can still be misaligned with respect to the loading direction, therefore fibers and fiber bundles were not fully stretched in the matrix. During the process of creeping, the longitudinal fibers have been stretched gradually and tend to be realigned along the stress axis, and this caused the increase of the elastic modulus. As the creeping time increased, the creep damage accumulated and caused the drop of the elastic modulus. These two types of process developed separately and competed with each other. After certain creeping time, fibers have been gradually straightened and oriented along the applied stress direction. This process contributed the most to increase the elastic modulus. The development of damage is relatively slow at this stage, so it only slows down the increase of elastic modulus to some extent and induced slight drop of the damage parameter D. As the creeping test went on, fibers were further straightened and oriented along the stress direction till the elastic modulus stopped further increasing. On the other hand, continuous development of the creeping damage caused the further increase of the damage parameter D, and this is associated with the constant dropping of the elastic modulus. This is the possible explanation for Fig. 12. In both notched and smooth samples, creeping rates were almost zero at 1100 C, though creep curves at 1500 C for both notched and smooth specimen were composed of a series of obvious steps which continuously rose with creeping time. When the micro-damage accumulation reached to certain extent, the lager damage took place which leads a step in creep curves. At the same temperature the creeping rate of notched sample at 95 MPa and smooth sample at 170 MPa are at the same level (105 /h). Therefore, it is obvious that temperature, other than the applied stress level, can influence the stable creeping rate the most. At 95 MPa, the comparison of damage curves in Fig. 12 and the crack quantity curves in Fig. 10 shows that these curves are very similar and in all the curves there is a fast changing stage and a slow changing stage. In the slow changing stage, damage value at 1500 C was always larger than that at 1300 C; the amount of cracks at 1500 C was also higher than that at 1300 C. This means that damage and crack tend to happen more in the creeping specimens at 1500 C than at 1300 C. Matrix cracking can cause the fiber fracture, and the fiber fracture leads the stress redistribution and more matrix cracking. Fiber fracture and matrix cracking are the main damage forms of C/SiC composite [2–4]. The mismatch of thermal expansion between C fiber and SiC matrix cause thermal residual stresses in the composite, Fig. 11. Stress distribution on a notched specimen of 2D-C/SiC composite (I indicates the maximum value and E indicates of stress value far from notch area). Table 1 Growth rate of micro-cracks at different creep time for 2D-C/SiC tensile creep test at 1500 C and 95 MPa in vacuum v (pieces/h) 0–2 (h) 2–10 (h) 10–25 (h) 25–50 (h) v1 20 0.375 0.4 0.01 v2 13 1 0.13 0.2 v3 6.5 2.75 0.2 0.125 v4 6 1.875 0.27 0.01 0 0.1 0.2 0.3 0.4 0 10 20 30 40 50 60 t(h) D 1300˚C 1500˚C Fig. 12. Damage evolution as a function of time for notched 2D-C/SiC specimens at 95 MPa in vacuum. W. Xiaojun et al. / Composites Science and Technology 66 (2006) 993–1000 999
w. Xiaojun et al Composites Science and Technology 66(2006)993-1000 3μm Fig 13. Micrographs of lateral side of the specimen near the notch area after 0.5 h creep with 95 MPa applied stress in vacuum, (a)1300C(b)1500C. and the minimum stress level is at about 1000C, which is 3. Amount of micro-cracks vs. time curve and micro-crack near the fabrication temperature of the 2D-C/SiC. The width vs time curve were very similar to the creeping residual thermal stress increases as the temperature de curves, and all these curves included a rapid increase creases or increases from 1000C. On one hand, this ther stage at the beginning followed by a slower increase mal residual stress can cause interfacial debonding. as In general, micro-cracks developed fast within 10 h. shown in Fig. 13. Under the same creeping stress, it ap The growth rate of the micro-cracks near the notch peared that fiber/matrix interface has larger scale of deb was faster than that far from the notch within 2 h onding at 1500C than at 1300oC. The interfacial whereas growth rate of micro-cracks far from notch debonding is one of the factors to deduce creep strain, in was faster than near the notch at the period of 2-10 h. addition, interfacial debonding can easily introduce fiber These revealed the stress redistribution during the creep sliding and this phenomenon on the macroscopic scale process. shows as the increase of creeping strain On the other hand, 4. The trends of damage curves at 1300 and 1500C were he thermal stress may transfer into tensile stress on fibers similar. The damage value at 1500C was always larger when the temperature is above the 2D-C/SiC fabrication than that at 1300C, and the amount of cracks at temperature [9] and this tensile stress can cause fiber frac l500° was also higher than that at1300°. ture. Then the fiber fracture introduces the stress redistri- bution within the neighbor-fibers of the broken fiber and e surroun ing SiC matrix, and eventually cause matrix ref crack. The existence of interfacial debonding make the ma- ix cracks easily propagate along the fiber/matrix interface u shengru Qiao, Zhongxue Yang. Dong Han, Mei Li. Tensile creep other than perpendicular to the fibers, and this can reduce damage and mechanism of 3D-C/SiC composite. J Mater Eng the stress concentration effect at crack tips. Therefore 2004: 1(4: 34-6 [in Chinese] interface debonding can improve the notch strength of [2] Boitier G, Vicens 3, Chermant JL Carbon fiber nano-creep in creep- the notched specimens. tested Cf-SiC composites. Scr Mater 1998: 38(6): 93 3] Boitier G, Chermant JL, Vicens J Multiscale investigation of the creep behavior of a 2.5D Cf-SiC composite J Mater Sci 1999: 34: 2759-67 4] Boitier G, Chermant JL, Vicens J. Understanding the creep behavior 5 Conclusions of a 2.5D Cf-SiC composite. Il. Experimental specifications and macroscopic mechanical creep responses. Mater Sci Eng A 2000:289:265-75 1. At 1100C, no matter in which stress level, the creep [5] Dong Han, Shengru Qiao, Mei Li, Juntao Hou, Xiaojun Wu. strains of both smooth and notched specimen were con centrated on transient creep stage while the steady creep composites. Acta Metall Sin 2004: 17(4): 569-74 rates were nearly zero. At 1500C, steady creep rate of [6] Gamus G, Guillaumat L, Baste S. Development of damage in a 2D notched specimen with 95 MPa creep stress level and the woven C/Sic posite under mechanical loading: I. mechanical smooth specimen at 170 MPa creep stress level were at 7 McNulty John C et al. Notch-sensitively of fiber-reinforced ceramic- matrix composites effects of inelastic straining and volume-dependent has been observed within 100 h creep test strength. J Am Ceram Soc 1999: 82(5): 1217-28 2. Creep damage mainly concentrated in the near notch [8] Charles et al. Notched tensile creep testing of ceramics. Mater Sci Eng area. Micro-cracks tended to appear on the near notch area and the cross-points of the woven fibers. Fractures []Shengru Qiao, Mei Li, Dong Han, Litong Zhang. Flexural perfor- easily occurred on the longitudinal fibers near the notch post-heat-treatment on flexural performance. J Mech Strengt 2003:25(5):495-8[ in Chinese
and the minimum stress level is at about 1000 C, which is near the fabrication temperature of the 2D-C/SiC. The residual thermal stress increases as the temperature decreases or increases from 1000 C. On one hand, this thermal residual stress can cause interfacial debonding, as shown in Fig. 13. Under the same creeping stress, it appeared that fiber/matrix interface has larger scale of debonding at 1500 C than at 1300 C. The interfacial debonding is one of the factors to deduce creep strain, in addition, interfacial debonding can easily introduce fiber sliding and this phenomenon on the macroscopic scale shows as the increase of creeping strain. On the other hand, the thermal stress may transfer into tensile stress on fibers when the temperature is above the 2D-C/SiC fabrication temperature [9] and this tensile stress can cause fiber fracture. Then the fiber fracture introduces the stress redistribution within the neighbor-fibers of the broken fiber and the surrounding SiC matrix, and eventually cause matrix crack. The existence of interfacial debonding make the matrix cracks easily propagate along the fiber/matrix interface other than perpendicular to the fibers, and this can reduce the stress concentration effect at crack tips. Therefore, interface debonding can improve the notch strength of the notched specimens. 5. Conclusions 1. At 1100 C, no matter in which stress level, the creep strains of both smooth and notched specimen were concentrated on transient creep stage while the steady creep rates were nearly zero. At 1500 C, steady creep rate of notched specimen with 95 MPa creep stress level and the smooth specimen at 170 MPa creep stress level were at the same magnitude (105 /h). No tertiary creep stage has been observed within 100 h creep test. 2. Creep damage mainly concentrated in the near notch area. Micro-cracks tended to appear on the near notch area and the cross-points of the woven fibers. Fractures easily occurred on the longitudinal fibers near the notch area. 3. Amount of micro-cracks vs. time curve and micro-crack width vs. time curve were very similar to the creeping curves, and all these curves included a rapid increase stage at the beginning followed by a slower increase. In general, micro-cracks developed fast within 10 h. The growth rate of the micro-cracks near the notch was faster than that far from the notch within 2 h, whereas growth rate of micro-cracks far from notch was faster than near the notch at the period of 2–10 h. These revealed the stress redistribution during the creep process. 4. The trends of damage curves at 1300 and 1500 C were similar. The damage value at 1500 C was always larger than that at 1300 C, and the amount of cracks at 1500 C was also higher than that at 1300 C. References [1] shengru Qiao, Zhongxue Yang, Dong Han, Mei Li. Tensile creep damage and mechanism of 3D-C/SiC composite. J Mater Eng 2004;1(4):34–6 [in Chinese]. [2] Boitier G, Vicens J, Chermant JL. Carbon fiber nano-creep in creeptested Cf-SiC composites. Scr Mater 1998;38(6):937–43. [3] Boitier G, Chermant JL, Vicens J. Multiscale investigation of the creep behavior of a 2.5D Cf-SiC composite. J Mater Sci 1999;34:2759–67. [4] Boitier G, Chermant JL, Vicens J. Understanding the creep behavior of a 2.5D Cf-SiC composite. II. Experimental specifications and macroscopic mechanical creep responses. Mater Sci Eng A 2000;289:265–75. [5] Dong Han, Shengru Qiao, Mei Li, Juntao Hou, Xiaojun Wu. Comparison of fatigue and creep behavior of 2D and 3D-C/SiC composites. Acta Metall Sin 2004;17(4):569–74. [6] Gamus G, Guillaumat L, Baste S. Development of damage in a 2D woven C/SiC composite under mechanical loading:I. mechanical characterization. Compos Sci Technol 1996;56:1363–72. [7] McNulty John C et al. Notch-sensitively of fiber-reinforced ceramicmatrix composites effects of inelastic straining and volume-dependent strength. J Am Ceram Soc 1999;82(5):1217–28. [8] Charles et al. Notched tensile creep testing of ceramics. Mater Sci Eng 1995;A203:217–21. [9] Shengru Qiao, Mei Li, Dong Han, Litong Zhang. Flexural performance of 3D-C/SiC composites at high temperature and influence of post-heat-treatment on flexural performance. J Mech Strength 2003;25(5):495–8 [in Chinese]. Fig. 13. Micrographs of lateral side of the specimen near the notch area after 0.5 h creep with 95 MPa applied stress in vacuum, (a) 1300 C (b) 1500 C. 1000 W. Xiaojun et al. / Composites Science and Technology 66 (2006) 993–1000