- Email: [email protected]
SiC composite at elevated temperature
Ceramics International xxx (xxxx) xxx–xxx
Contents lists available at ScienceDirect
Ceramics International journal homepage: www.elsevier.com/locate/ceramint
Tensile creep behavior of three-dimensional four-step braided SiC/SiC composite at elevated temperature ⁎
Xin Jinga,b, Xiaoguang Yangb,c, Duoqi Shib,c,d, , Hongwei Niue a
School of Power and Energy, Northwestern Polytechnical University, Xi'an 710072, China School of Energy and Power Engineering, Beihang University, Beijing 100191, China c Collaborative Innovation Center of Advanced Aero-engine, Beijing 100191, China d Beijing Key Laboratory of Aero-Engine Structure and Strength, Beijing 100191, China e China Flight Test Establishment, Xi’an China b
A R T I C L E I N F O
A BS T RAC T
Keywords: SiC/SiC composites Creep Fracture Ceramic matrix composites
This article presents experimental results for tensile creep deformation and rupture behavior of threedimensional four-step braided SiC/SiC composites at 1100 °C and 1300 °C in air. The creep behavior at 1300 °C exhibited a long transient creep regime and the creep rate decreased continuously with time. The creep behavior at 1100 °C exhibited an apparent steady-rate regime and the creep deformation was smaller than that at 1300 °C. However, the creep rupture time at both temperatures showed little difference. The mechanisms controlling creep deformation and rupture behavior were analyzed.
1. Introduction Ceramic matrix composites (CMCs), particularly those containing braided or woven fiber preforms, are under development for high temperature applications as e.g. in aero engines, rocket nozzles, and reentry heat shields. The SiC/SiC composites exhibited good tensile strength (~350 MPa) at room temperature [1,2]. However, a significant decrease to approximate 200 MPa in tensile strength was observed at 1100 °C and 1300 °C in air, which was due to oxidation embrittlement initiated from the surfaces. Moreover, composites with environmental barrier coating (EBC) maintained good mechanical properties for short-term at 1300°C [2]. The creep behavior of SiC/SiC (C/SiC, SiC/C) composites has been widely investigated [3–19] because long-time mechanical properties at high temperatures are important for the application of CMCs. The creep behaviors are different in a combination of different fiber and matrix. Generally, stress redistribution occurs during creep and the rule of mixture is violated. Therefore the creep mismatch ratio (CMR) is proposed which is defined as a ratio of the creep rate of the fiber to that of the matrix [20]. When CMR < 1, stress transferred from matrix to fiber. The embedded fiber control the deformation and rupture behavior of CMCs. When CMR > 1, stress transferred from fiber to matrix at first, so the creep behavior is dominated by matrix properties. Once greater matrix stress results in matrix cracking, the stress transferred to bridging fiber once again. Zhu investigates creep and
⁎
fatigue behavior of SiC fibers (Nicalon and Hi-Nicalon) reinforced CVI derived SiC matrix composites in air and argon at temperatures ranging from 1000 °C to 1300 °C [3–6]. Both creep and fatigue resistance of Hi-Nicalon/SiC is similar to that of enhanced SiC/SiC, but much better than standard SiC/SiC. For these SiC/SiC of which the creep is controlled by matrix, better performance of the matrix can effectively improve the CMCs creep resistance. Recently research of Morscher demonstrates that the porous matrix, for example the SiC matrix derived by PIP process, does not carry significant load, which leads to higher creep strain rate of composites compared to that contains CVI or MI matrix [7] . Moreover, the creep behavior of the CMCs composites is not only depended on the constituent creep properties, but also influenced by their textile structures. The creep damage and failure of CMC is relevant to matrix cracking behavior, which is influenced by fiber preform. The creep behavior of the CMCs composites which has different type of fiber preform was not widely investigated. Previous research was mainly focused on woven CMCs. For woven CMCs, transverse matrix cracking initiated in transverse tows or inter-tow matrix, penetrating the longitudinal tows, leading to ultimate failure [9,10,12,18]. In general, the first generation Si-C-O fibers and carbon-rich interface in SiC/SiC-based composites possessed relative low oxidation resistance at elevated temperature in air, limiting the long-term performance of SiC-based CMCs [21–23]. The oxidation embrittlement propagates through matrix cracking, and the oxygen reacts with fiber,
Corresponding author at: School of Energy and Power Engineering, Beihang University, Beijing 100191, China. E-mail address: [email protected] (D. Shi).
http://dx.doi.org/10.1016/j.ceramint.2017.02.076 Received 5 September 2016; Received in revised form 22 January 2017; Accepted 17 February 2017 0272-8842/ © 2017 Elsevier Ltd and Techna Group S.r.l. All rights reserved.
Please cite this article as: Jing, X., Ceramics International (2017), http://dx.doi.org/10.1016/j.ceramint.2017.02.076
Ceramics International xxx (xxxx) xxx–xxx
X. Jing et al.
interface and matrix. In order to overcome this drawback, besides the application of second or third generation more creep resistant SiC fiber [24], glass phase in the matrix [4,5], SiBC [18], glass sealant [25] and coatings [26] were applied on the composites to act as an oxygen diffusion barrier and make the composites exhibited good mechanical properties in air. The present work investigated the tensile creep behavior of threedimensional four-step braided KD-1 fiber/PyC interface/SiC matrix composites at 1100 °C and 1300 °C in air. Composites with and without coatings were crept at 1300°C, and composites with coatings were crept at 1100 °C. Rupture behaviors and failure mechanisms of composites after creep testing was also evaluated by scanning electron microscopy (SEM) and computed tomography (CT). Based on the experimental results, the creep controlling mechanism and the coating effect will be discussed with the experimental results.
Stress/MPa
300
200
RT 1100U 1300U 1300C
100
0 0.0
0.2
0.4
0.6
0.8
Strain/%
2. Material and experimental procedures
Fig. 2. Typical monotonic tensile stress-strain curve for coated and uncoated composites (U means uncoated and C means coated).
The CMC [1], are based on the SiC/SiC system with SiC fibers as reinforcement. The yarns consisting 1.2K fibers were braided into a 3D four directional preform. The fiber volume fraction was approximately 46.5%. The preform was surrendered to a chemical vapor deposition (CVD) process to prepare a ~200 nm thick pyrolysed carbon (PyC) layer. Polymer impregnation and pyrolysis (PIP) process with a polymer precursor of liquid polyvinycarbosilane (LPVCS) was adopted to dense the matrix. 14 cycles of impregnation and pyrolysis were repeated to reduce the composites porosity. The original densified preform was a flat plate of which the size is 130 mm×60 mm×4 mm. The tensile test specimens shown in Fig. 1 were cut by water-jet cutting machine. The overall specimen dimensions were roughly 127 mm ×14 mm×4 mm and the central reduced section were 30 mm× 6 mm×4 mm. For some specimens to be tested in elevated temperature, after machining to the geometry suitable for mechanical testing, a two-layer EBC was applied on the specimen surfaces to protect the composite against oxidation. The inner layer was mullite and the outer layer was erbium silicate. Both of the layer thickness was 0.1 mm. Monotonic tension and creep tests were conducted under stress control in a servo-hydraulic SHIMADZU test machine, following the general guidelines of ASTM standard C1337 [27]. The specimens were griped by edge load, passive grip interface and the alignment of the testing system was verified at the beginning of the test series. The high temperature furnace and contact-type Epsilon-3548 high temperature
extensometer used for these experiments have also been shown in Fig. 1. Details about the experiment system were in ref [2]. The loading rate at the beginning of the creep test was 3 MPa/s and creep stress range was 40–120 MPa. Creep tests for coated specimens were conducted at 1100 °C and 1300 °C, while for uncoated specimens the tests were only conducted at 1300 °C. After the creep test, fracture surface, cross-section and side surface of the specimens were observed using by scanning electron microscopy (SEM). The inner micro structure evolution was evaluated by computed tomography (CT). X-ray diffraction (XRD) analysis was conducted in order to investigate crystallization of SiC/SiC during creep tests.
3. Results and discussion 3.1. Monotonic tensile behavior Monotonic tensile stress–strain curves of the SiC/SiC composites at room temperature(RT),1100 °C and1300 °C are shown in Fig. 2. All the curves were similar, starting with a linear response, and being followed by a non-linear pseudo-ductile fracture behavior. The ultimate tensile stress(UTS) of composites with coating was considerably higher than
Fig. 1. Test configuration and specimens used for monotonic loading and creep testing at elevated temperatures.
2
Ceramics International xxx (xxxx) xxx–xxx
X. Jing et al.
Fig. 3. Fracture surface of (a) coated and (b) uncoated SiC/SiC composites monotonic tensile tested at 1300 ℃, and magnification of (c) fiber pull-out region and (d) flat region.
similar to the coated specimens. In high stress creep tests, the uncoated specimens showed more rapid deformation and the tertiary creep occurred quite early (Fig. 4c). In low stress creep tests, primary creep was still dominated, but the creep strains were higher than the coated specimens under the same stress (Fig. 4d). Fig. 5 plots the creep strain rate versus time relation of coated specimens at 1100 °C and 1300 °C. The creep rate in 1100 °C tests (Fig. 5a) exhibited decreasing-steady (- increasing) phenomenon. The steady state stages was obvious, starting earlier when the stress was higher. In 1300 °C creep tests (Fig. 5b), the decreasing creep rate of unfailed specimens showed almost linear relation with time with in the test period. In many cases, steady state creep rate, ε̇cr , under applied stress, σ, is represented by the power law equation:
without coating and was close to the UTS at room temperature. The coated specimens maintain the elastic modulus (about 70% of the RT modulus) of composites at 1300 °C, but the modulus for uncoated specimens drop almost 55%. The proportional limits at 1100 °C and 1300 °C almost have the same value of 80 MPa (40%UTS). Fracture surfaces of the coated and uncoated specimens tested at 1300 ℃ were shown in Fig. 3. A flat region without fiber pullout was observed for uncoated specimen, indicating the oxidation occurred rapidly for uncoated composites at elevated temperature [28,29]. Additionally, the absence of the flat oxidation damage region on fracture surface of coated specimen demonstrated the EBC had good oxidation resistance during short-term (not more than 1 h). 3.2. Creep deformation behavior
εcṙ = Aσ n exp (−Q /RT )
The relationship between creep deformation and time at 1100 ℃ and 1300 ℃ is plotted in Fig. 4. At same stress level, the material showed more creep resistance at 1100 °C than 1300 °C. In creep tests at 1100°C (Fig. 4a), quasi-steady state creep dominated the creep deformation. The creep deformation curves of failed specimens exhibited typical three-stage characteristics. After a short primary creep stage, the creep strain rates decreased to a minimum rate ranging from 10−9~10−6/s. Although the creep time during the steady state was long, the creep deformation was small. At high stresses, rapid acceleration of creep rate became more evident just prior to fracture. In creep tests for the coated specimens at 1300 °C (Fig. 4b), transient creep behavior dominated the creep deformation, which was quite different from 1100 °C. At high stresses, the primary creep transferred to tertiary creep directly, without an obvious steady-rate stage. In 40 MPa creep test which accidentally terminated in 30 h, primary creep lasted for the whole tests and no clear evidence of steady creep state was found. The 60 MPa creep test survived after 100 h, steady creep was found after 30 h transient creep. Creep deformation of the uncoated specimens at 1300 °C was
(1)
where A is a dimensionless constant, n the stress exponent, Q the apparent activation energy, R the universal gas constant, and T the absolute temperature. The relationship between applied stress and quasi–steady state creep rate is shown in Fig. 6a. In the low stress region, the creep rate was insensitive to stress and n was almost unity, while the value was large in high stress range. The transition stress where the stress exponent changes was correspond to the proportional limit (60–80 MPa) in the monotonic testing at elevated temperature. Similar results have been reported for Si-Ti-C-O fiber/SiC-based matrix composites [11] and 0°/90° Nicalon SiC/SiC [20] . The stress exponent in the low stress region was similar to Si-C-O fibers, indicating the creep deformation was controlled by fiber bundles. The higher value of n in high stress range was attributed to matrix cracking [11]. Slow crack growth and bridging fiber creep increased the creep rate. It should be noticed that the time point for determining the steady creep strain rate of each specimen was different, especially for the case of 1300 °C tests which did not show secondary creep state. The creep rate decreases gradually from an initial high value until a minimum rate is attained essentially at the point of failure. The timing for calculate 3
Ceramics International xxx (xxxx) xxx–xxx
X. Jing et al.
Fig. 4. Typical tensile creep strain versus time of the composite: (a) tested at 1100 ℃ for coated specimens, (b) tested at 1300 ℃ for coated specimens, (c) tested at 1300 ℃ for uncoated specimens under high stress, (d) tested at 1300 ℃ for uncoated specimens under low stress.
creep rate at high stress was usually less than 2000 s, while in low stress range the time point was usually more than 104 or 105 s. At high stress level, the creep curves terminated early, giving high minimum
creep rates and low rupture lives. In contrast, the creep rate under low stress decayed continuously until fracture occurred after long time, resulting in lower minimum rate. Therefore, a similar relationship was 4
Ceramics International xxx (xxxx) xxx–xxx
X. Jing et al.
Fig. 6. (a)Quasi-steady state creep rate versus applied stress for composites. (b)Creep strain rate at different time versus applied stress for the composites at 1300 °C. Fig. 5. Typical tensile creep strain rate versus time of the composite: (a) tested at 1100 ℃ for coated specimens, (b) tested at 1300 ℃ for coated specimens.
applied for determining ε̇cr at fixed creep time t, as the work done by [12]. The Eq. (1) was utilized to determine the stress dependence at 5×102, 1×103, 5×103, 1×104, 5×104, and 1×105 s for specimens tested at 1300 °C. The stress exponents, n, were determined from the log ε̇cr vs log σ plots (Fig. 6b). Values of n decreased slightly with increasing time and all of them were close to unity, indicating the creep deformations were controlled by Newtonian-type viscous flow. These stress exponents were similar to Nicalon fibers, indicating that the creep rate of the composites was controlled by fiber creep. The microstructure of Si-C-O fibers is unstable above 1200 °C. Fig. 7 shows XRD patterns of the coated composites before and after creep tests at 1300 °C. The results demonstrated the increase in volume and size of the SiC grains. Bodet [30,31] reported that the decomposition of Si-C-O fiber results in a decrease of the SiOxCy volume fraction and an increase in SiC volume fraction, leading to an increase in fiber viscosity and, therefore, creep resistance. The microstructure evolution during creep explained long transient-stage creep deformation at 1300 °C. Due to weak creep resistance of porous matrix, creep deformation was controlled by the fibers, accompanied by cracking of the weak brittle matrix. The matrix, especially derived through PIP process, lost load-bearing capacity in creep test [7]. Due to the stress redistribution during creep, the external load was transferred to fibers. The creep data for the fiber is limited, so the Nicalon fiber data [32] were used here to investigate the creep deformation as both of them are the first generation SiC fibers and similar in chemical and microstructural aspects. Comparing the creep rates of composites tested at 1300 °C at 1×105 s with the creep data of Nicalon fibers, the applied stresses on
Fig. 7. XRD patterns from the composite before and after creep tests at 1300 °C in air.
composites were about one fourth of the stresses on Nicalon fibers at same creep rate (Fig. 8). Given the fiber volume fraction of 0.465 for the SiC/SiC composites, the stress on fiber should be 2.15(1/0.465) times of the applied stress. The discrepancy was attributed to two reasons. On one hand, the off-axis loading of the fiber bundles impaired their load-bearing capacity [33]. Previously study informed that the 20° off-axis loading led to about 35% loss of the strength for these 3D fourdirectional CMCs [2,33]. It was reasonable to consider that creep 5
Ceramics International xxx (xxxx) xxx–xxx
X. Jing et al.
Fig. 8. Comparison of the stress-creep rate relationship of the SiC/SiC composites and Nicalon fibers at 1300 °C.
resistance was influenced by the fiber bundle directions in the preform. On the other hand, the content of O in the fibers is higher than Nicalon fibers, which indicated the KD-1 fibers contain less amount of β-SiC grains and more amorphous Si-O-C. The increase of β-SiC content improves the viscosity and, therefore, creep resistance [31]. 3.3. Creep rupture behavior Fig. 10. Scanning electron micrographs showing typical fracture surface observed in composite creep tested in air: (a) Creep tested at 1100 ℃ under a stress of 120 MPa, the arrows illustrate the defect for oxidation ingression paths, (b) Creep tested at 1300 ℃ under a stress of 40 MPa, the arrows illustrate the undamaged bridging fiber bundles.
Creep rupture behavior of coated specimens tested at 1100 °C and 1300 °C and uncoated specimens tested at 1300 °C are summarized in Fig. 9. Although the data was seen to have considerable scatter, it would appear that a change in creep strength degradation behavior of uncoated specimens occurred in short-term region within 1000 s. The difference in strength degradation behavior was not notable between uncoated and coated specimens tested at 1300 °C for long-term. The temperature effect on strength degradation seemed to be also insignificant between 1100 °C and 1300 °C for coated composites. The threshold stress of creep rupture within 100 h was 40–60 MPa for all three cases. In short term tensile experiments, the coating effectively prevented oxidation damage invasion and the composites hold a relative high ultimate strength. However, when the composites were conducted in long-term loading including fatigue and creep tests, ingression of oxygen into the specimen still took place and flat regions were observed
on the fracture surfaces of these specimens [11,12]. When the stress was relative high, the morphology of the fracture surface was similar to uncoated specimens under monotonic test at elevated temperatures. Fig. 10(a) shows the fracture surface of coated specimen tested at 1100 °C with a creep stress of 120 MPa. The undamaged region occupied almost 47% area of whole cross section and it was not located in the central zone, indicating that the oxygen ingression did not propagate evenly along the periphery region of the specimen but initiated at surfaces or corners where coating had defects and penetrated the specimen through the unbridged cracks. The large crack opening displacements associated with a growing surface crack would then allow direct oxygen penetration, promoting fiber failure in the crack plane. In this way, the dominant crack would grow rapidly until the remaining unbroken fiber bundles were unable to carry the load, resulting in a pull-out zone in one region which may or may not be located centrally on the fracture surface. When the stress was lower and the loading time was longer, the flat fracture surface caused by oxidation damage occupied the most area of the specimen fracture surface. The fiber pullout regions sometimes could not be connected to a continuous piece. The fracture surface of specimens tested at 1300 ℃ under a creep stress of 40 MPa is shown in Fig. 10(b). Several distributed fiber pullout regions were embedded in the flat surface. The area occupied by pullout fiber bundles regions added up to 15% of the whole cross section area. The area of each undamaged regions was small and each region usually consisted only one or two fiber bundles, which indicated that to an extent, failure occurred by growth and link up of several separately-nucleated cracks. Additionally, the bundle/bundle interface still took effect on deflecting the damage propagation, leading to fiber bundles bridging the crack. As the load became lower, the microstructure had more influence on the oxidation damage propagation. For this SiC/SiC composite, creep of the fibers was accompanied by cracking of the brittle matrix. To understand the damage during the
Fig. 9. Creep rupture data for uncoated composites at 1300 ℃ and coated composites at 1100 ℃ and 1300 ℃.
6
Ceramics International xxx (xxxx) xxx–xxx
X. Jing et al.
Fig. 11. Cracks distribution in the specimens at 1300 °C under a creep stress of (a)(c)40 MPa and (b)(d) 120 MPa.
failure and sudden fracture occurred by fiber pull-out. Fig. 11 shows the cracks in the composites under 40 MPa and 120 MPa at 1300 °C. The crack density was greater for 40 MPa than 120 MPa. The cracks for 40 MPa were all bridged by fibers, but some cracks for 120 MPa were unbridged. After examining the crack in other specimens, it could be concluded that at lower stresses, the long rupture time lead to denser crack distribution. However, as stress increased, more unbridged cracks were observed. The unbridged cracks were difficult to be observed for the specimens tested under medium stress level, except the crack leading to ultimate fracture. The matrix cracking evolution, therefore, should be not only depend on stress, but also influenced by time. The EBC played an important role on maintaining the mechanical properties of the SiC/SiC composites for short-term loading at elevated temperatures. As the creep time increase, the difference for coated and uncoated composites became smaller, including the creep strain rate and rupture time. The effect of coating degraded as the load and time
creep, the damage processes leading to failure were clarified by microstructural examination and fractographic studies. The fibers in undamaged region were thought as ‘effective fibers’ that contribute to the composite ultimate strength [11,34,35]. It could be possible to conclude that oxidation damage propagated in the composites gradually through the flat unbridged matrix crack plane and the specimens failed when the fiber bundles in the rest undamaged region could not sustain the external load. According to the examination, most of the cracks were located in the fiber-free matrix region, arrested by the individual bundles of braided preforms. The tunneling cracks between fiber bundles made the stress redistribution, leading to fiber bear more load. Thus indicate that the fiber dominate the composites creep deformation behavior. Few cracks penetrated into fiber bundles normal to the loading direction, especially near the surface, because of oxidation embrittlement. Then oxidation-assisted fiber failure occurred in crack tip, leading to matrix cracks propagate inward. That explained the fracture surface contains the flat crack growth zones with fiber 7
Ceramics International xxx (xxxx) xxx–xxx
X. Jing et al.
Fig. 12. Cracks on coating for the specimens under (a)10 MPa, and (b) 40 MPa for 30 h at 1300 °C. (c) cracks between coating and SiC/SiC for 10 MPa specimens.
Acknowledgement
increased. Fig. 12(a) shows cracks were observed on the coating of the specimen which was tested for 30 h at 1300 °C under a very small load (not more than 10 MPa). As the stress increased to 40 MPa (Fig. 12(b)), more cracks were observed. Some pieces of the coatings began to be peeled off (Fig. 12(c)), leading to the area underwent more severe oxidation damage. The coating, therefore, could protect the composites effectively from oxidation ingress as long as the integrity was maintained. Once the oxygen invaded the composites, the mechanical performance of the SiC/SiC composites would decrease dramatically.
The research was supported by National Natural Science Foundation of China (No. NSFC51275023). References [1] Z. Luo, X. Zhou, J. Yu, Mechanical properties of SiC/SiC composites by PIP process with a new precursor at elevated temperature, Mater. Sci. Eng.: A 607 (5) (2014) 155–161. [2] J. Xin, et al., Fiber strength measurement for KD-I(f)/SiC composites and correlation to tensile mechanical behavior at room and elevated temperatures, Ceram. Int. 41 (1) (2015) 299–307. [3] S. Zhu, et al., Monotonic tension, fatigue and creep behavior of SiC-fiber-reinforced SiC-matrix composites: a review, Compos. Sci. Technol. 59 (6) (1999) 833–851. [4] S. Zhu, et al., Creep and fatigue behavior in Hi-Nicalon-Fiber-reinforced silicon carbide composites at high temperatures, J. Am. Ceram. Soc. 82 (1) (1999) 117–128. [5] S. Zhu, et al., Creep and fatigue behavior in an enhanced SiC/SiC composite at high temperature, J. Am. Ceram. Soc. 81 (9) (1998) 2269–2277. [6] S. Zhu, et al., Tensile creep behavior of a SiC-fiber/SiC composite at elevated temperatures, Compos. Sci. Technol. 57 (12) (1998) 1629–1637. [7] G.N. Morscher, Tensile creep and rupture of 2D-woven SiC/SiC composites for high temperature applications, J. Eur. Ceram. Soc. 30 (11) (2010) 2209–2221. [8] G.N. Morscher, et al., Creep in vacuum of woven Sylramic-iBN melt-infiltrated composites, Compos. Sci. Technol. 71 (1) (2011) 52–59. [9] G.N. Morscher, V.V. Pujar, Creep and stress–strain behavior after creep for SiC fiber reinforced, melt-infiltrated SiC matrix composites, J. Am. Ceram. Soc. 89 (5) (2006) 1652–1658. [10] G.N. Morscher, et al., Tensile creep and fatigue of Sylramic-iBN melt-infiltrated SiC matrix composites: retained properties, damage development, and failure mechanisms, Compos. Sci. Technol. 68 (15–16) (2008) 3305–3313. [11] T. Ogasawara, et al., Tensile creep behavior of 3-D woven Si-Ti-C-O fiber/SiC-based matrix composite with glass sealant. J. Mater. Sci. vol. 35(4). pp. 785–793. [12] T. Ogasawara, et al., Tensile creep behavior and thermal stability of orthogonal three-dimensional woven TyrannoTM ZMI fiber/silicon-titanium-carbon-oxygen matrix composites, J. Am. Ceram. Soc. 85 (2) (2002) 393–400. [13] B. Wilshire, Creep property comparisons for ceramic-fibre-reinforced ceramic– matrix composites, J. Eur. Ceram. Soc. 22 (8) (2002) 1329–1337. [14] B. Wilshire, F. Carreño, Deformation and damage processes during tensile creep of ceramic-fibre-reinforced ceramic–matrix composites, J. Eur. Ceram. Soc. 20 (4) (2000) 463–472. [15] G. Boitier, J.L. Chermant, J. Vicens, Understanding the creep behavior of a 2.5D Cf–SiC composite: II. Experimental specifications and macroscopic mechanical creep responses, Mater. Sci. Eng.: A 289 (1–2) (2000) 265–275. [16] G. Boitier, J. Vicens, J.L. Chermant, Understanding the creep behavior of a 2.5D Cf–SiC composite-I. Morphology and microstructure of the as-received material, Mater. Sci. Eng.: A 279 (1–2) (2000) 73–80.
4. Conclusion The tensile creep behavior of a three-dimensional four-step braided SiC fiber/SiC ceramic matrix composite with and without coating at 1100 °C and 1300 °C in air has been investigated. The following conclusions have been made. (1) The creep deformation for the composites at 1100 °C contained primary and secondary stage while the creep deformation for composites at 1300 °C mainly exhibited primary stage. (2) The fibers control the rates of creep deformation. The stress exponent was similar to that of Si-C-O fibers found in literature. (3) The matrix cracking was different under high or low stress. The dominant crack grows rapidly by direct ingress of oxygen along the crack plane under high stress. In contrast, cracks form extensively throughout the specimens under low stress. The reduced crack growth rates lead to longer time and high strains before failure. The undamaged area ratio increases with decreasing stress. (4) The creep rupture behavior of the composite is suggested to depend on the factor of unbridged cracking propagation and residual strength. As the developing cracks penetrate into the fiber bundles, crack growth rates are limited by crack-bridging fibers, but fiber failure is promoted by oxygen ingress. Creep fracture then takes place when the area of the developing cracks attain the critical fraction of the test-piece cross-section required for sudden failure by fiber pull-out.
8
Ceramics International xxx (xxxx) xxx–xxx
X. Jing et al.
15) (2002) 2741–2747. [27] ASTM, Standard Test Method for Creep and Creep Rupture of Continuous FiberReinforced Advanced Ceramics Under Tensile Loading at Elevated Temperatures, in C1337-10, 2010. [28] I.J. Davies, et al., Fibre strength parameters measured in situ for ceramic-matrix composites tested at elevated temperature in vacuum and in air, Compos. Sci. Technol. 59 (6) (1999) 801–811. [29] A.G. Evans, et al., Models of high-temperature, environmentally assisted embrittlement in ceramic-matrix composites, J. Am. Ceram. Soc. 79 (9) (1996) 2345–2352. [30] R. Bodet, et al., Microstructural stability and creep behavior of Si-C-O (Nicalon) fibers in carbon monoxide and argon environments, J. Am. Ceram. Soc. 79 (10) (1996) 2673–2686. [31] N. Jia, R. Bodet, R.E. Tressler, Effects of microstructural instability on the creep behavior of Si-C-O (Nicalon) fibers in argon, J. Am. Ceram. Soc. 76 (12) (1993) 3051–3060. [32] G. Simon, A.R. Bunsell, Creep Behaviour and Structural Characterization at High Temperatures of Nicalon SiC Fibres, vol. 19(11), 1984, pp. 3658–3670. [33] G.N. Morscher, H.M. Yun, J.A. DiCarlo, In-plane cracking behavior and ultimate strength for 2D woven and braided melt-infiltrated SiC/SiC composites tensile loaded in off-axis fiber directions, J. Am. Ceram. Soc. 90 (10) (2007) 3185–3193. [34] I.J. Davies, et al., Fibre strength parameters measured in situ for ceramic-matrix composites tested at elevated temperature in vacuum and in air, Compos. Sci. Technol. 59 (6) (1999) 801–811. [35] I.J. Davies, et al., Case study of failure in a glass-sealed SiC/SiC-based composite creep tested at 1100 °C in air, Mater. Lett. 48 (3–4) (2001) 205–209.
[17] G. Boitier, J. Vicens, J.L. Chermant, Understanding the creep behavior of a 2.5D Cf–SiC composite. III. From mesoscale to nanoscale microstructural and morphological investigation towards creep mechanism, Mater. Sci. Eng.: A 313 (1–2) (2001) 53–63. [18] J.L. Chermant, et al., The creep mechanism of ceramic matrix composites at low temperature and stress, by a material science approach, J. Eur. Ceram. Soc. 22 (14– 15) (2002) 2443–2460. [19] C.H. Weber, et al., High temperature deformation and rupture in SiC-C composites, Mater. Sci. Eng.: A 196 (1–2) (1995) 25–31. [20] J.W. Holmes, X. Wu, Elevated temperature creep behavior of continuous fiber reinforced ceramics, in: S. Nair, K. Jakus, S. Nair, K. Jakus (Eds.), High Temperature Mechanical Behavior of Ceramic Compsites, ButterworthHeinemann, 1995, pp. 193–259. [21] J. DiCarlo, H. Yun, Non-oxide (silicon carbide) fibers, in: N. Bansal, N. Bansal (Eds.), Handbook of Ceramic Composites, Kluwer Academic, 2005. [22] F.E. Heredia, et al., Oxidation embrittlement probe for ceramic-matrix composites, J. Am. Ceram. Soc. 78 (8) (1995) 2097–2100. [23] H. Lin, P.F. Becher, Stress-temperature-lifetime working envelope of Nicalon fiberreinforced SiC matrix composites in air, Compos. Part A: Appl. Sci. Manuf. 28 (11) (1997) 935–942. [24] A.R. Bunsell, A. Piant, A review of the development of three generations of small diameter silicon carbide fibres, J. Mater. Sci. 41 (3) (2006) 823–839. [25] I.J. Davies, et al., Case study of failure in a glass-sealed SiC/SiC-based composite creep tested at 1100 °C in air, Mater. Lett. 48 (3–4) (2001) 205–209. [26] H.E. Eaton, G.D. Linsey, Accelerated oxidation of SiC CMC's by water vapor and protection via environmental barrier coating approach, J. Eur. Ceram. Soc. 22 (14–
9
Contents lists available at ScienceDirect
Ceramics International journal homepage: www.elsevier.com/locate/ceramint
Tensile creep behavior of three-dimensional four-step braided SiC/SiC composite at elevated temperature ⁎
Xin Jinga,b, Xiaoguang Yangb,c, Duoqi Shib,c,d, , Hongwei Niue a
School of Power and Energy, Northwestern Polytechnical University, Xi'an 710072, China School of Energy and Power Engineering, Beihang University, Beijing 100191, China c Collaborative Innovation Center of Advanced Aero-engine, Beijing 100191, China d Beijing Key Laboratory of Aero-Engine Structure and Strength, Beijing 100191, China e China Flight Test Establishment, Xi’an China b
A R T I C L E I N F O
A BS T RAC T
Keywords: SiC/SiC composites Creep Fracture Ceramic matrix composites
This article presents experimental results for tensile creep deformation and rupture behavior of threedimensional four-step braided SiC/SiC composites at 1100 °C and 1300 °C in air. The creep behavior at 1300 °C exhibited a long transient creep regime and the creep rate decreased continuously with time. The creep behavior at 1100 °C exhibited an apparent steady-rate regime and the creep deformation was smaller than that at 1300 °C. However, the creep rupture time at both temperatures showed little difference. The mechanisms controlling creep deformation and rupture behavior were analyzed.
1. Introduction Ceramic matrix composites (CMCs), particularly those containing braided or woven fiber preforms, are under development for high temperature applications as e.g. in aero engines, rocket nozzles, and reentry heat shields. The SiC/SiC composites exhibited good tensile strength (~350 MPa) at room temperature [1,2]. However, a significant decrease to approximate 200 MPa in tensile strength was observed at 1100 °C and 1300 °C in air, which was due to oxidation embrittlement initiated from the surfaces. Moreover, composites with environmental barrier coating (EBC) maintained good mechanical properties for short-term at 1300°C [2]. The creep behavior of SiC/SiC (C/SiC, SiC/C) composites has been widely investigated [3–19] because long-time mechanical properties at high temperatures are important for the application of CMCs. The creep behaviors are different in a combination of different fiber and matrix. Generally, stress redistribution occurs during creep and the rule of mixture is violated. Therefore the creep mismatch ratio (CMR) is proposed which is defined as a ratio of the creep rate of the fiber to that of the matrix [20]. When CMR < 1, stress transferred from matrix to fiber. The embedded fiber control the deformation and rupture behavior of CMCs. When CMR > 1, stress transferred from fiber to matrix at first, so the creep behavior is dominated by matrix properties. Once greater matrix stress results in matrix cracking, the stress transferred to bridging fiber once again. Zhu investigates creep and
⁎
fatigue behavior of SiC fibers (Nicalon and Hi-Nicalon) reinforced CVI derived SiC matrix composites in air and argon at temperatures ranging from 1000 °C to 1300 °C [3–6]. Both creep and fatigue resistance of Hi-Nicalon/SiC is similar to that of enhanced SiC/SiC, but much better than standard SiC/SiC. For these SiC/SiC of which the creep is controlled by matrix, better performance of the matrix can effectively improve the CMCs creep resistance. Recently research of Morscher demonstrates that the porous matrix, for example the SiC matrix derived by PIP process, does not carry significant load, which leads to higher creep strain rate of composites compared to that contains CVI or MI matrix [7] . Moreover, the creep behavior of the CMCs composites is not only depended on the constituent creep properties, but also influenced by their textile structures. The creep damage and failure of CMC is relevant to matrix cracking behavior, which is influenced by fiber preform. The creep behavior of the CMCs composites which has different type of fiber preform was not widely investigated. Previous research was mainly focused on woven CMCs. For woven CMCs, transverse matrix cracking initiated in transverse tows or inter-tow matrix, penetrating the longitudinal tows, leading to ultimate failure [9,10,12,18]. In general, the first generation Si-C-O fibers and carbon-rich interface in SiC/SiC-based composites possessed relative low oxidation resistance at elevated temperature in air, limiting the long-term performance of SiC-based CMCs [21–23]. The oxidation embrittlement propagates through matrix cracking, and the oxygen reacts with fiber,
Corresponding author at: School of Energy and Power Engineering, Beihang University, Beijing 100191, China. E-mail address: [email protected] (D. Shi).
http://dx.doi.org/10.1016/j.ceramint.2017.02.076 Received 5 September 2016; Received in revised form 22 January 2017; Accepted 17 February 2017 0272-8842/ © 2017 Elsevier Ltd and Techna Group S.r.l. All rights reserved.
Please cite this article as: Jing, X., Ceramics International (2017), http://dx.doi.org/10.1016/j.ceramint.2017.02.076
Ceramics International xxx (xxxx) xxx–xxx
X. Jing et al.
interface and matrix. In order to overcome this drawback, besides the application of second or third generation more creep resistant SiC fiber [24], glass phase in the matrix [4,5], SiBC [18], glass sealant [25] and coatings [26] were applied on the composites to act as an oxygen diffusion barrier and make the composites exhibited good mechanical properties in air. The present work investigated the tensile creep behavior of threedimensional four-step braided KD-1 fiber/PyC interface/SiC matrix composites at 1100 °C and 1300 °C in air. Composites with and without coatings were crept at 1300°C, and composites with coatings were crept at 1100 °C. Rupture behaviors and failure mechanisms of composites after creep testing was also evaluated by scanning electron microscopy (SEM) and computed tomography (CT). Based on the experimental results, the creep controlling mechanism and the coating effect will be discussed with the experimental results.
Stress/MPa
300
200
RT 1100U 1300U 1300C
100
0 0.0
0.2
0.4
0.6
0.8
Strain/%
2. Material and experimental procedures
Fig. 2. Typical monotonic tensile stress-strain curve for coated and uncoated composites (U means uncoated and C means coated).
The CMC [1], are based on the SiC/SiC system with SiC fibers as reinforcement. The yarns consisting 1.2K fibers were braided into a 3D four directional preform. The fiber volume fraction was approximately 46.5%. The preform was surrendered to a chemical vapor deposition (CVD) process to prepare a ~200 nm thick pyrolysed carbon (PyC) layer. Polymer impregnation and pyrolysis (PIP) process with a polymer precursor of liquid polyvinycarbosilane (LPVCS) was adopted to dense the matrix. 14 cycles of impregnation and pyrolysis were repeated to reduce the composites porosity. The original densified preform was a flat plate of which the size is 130 mm×60 mm×4 mm. The tensile test specimens shown in Fig. 1 were cut by water-jet cutting machine. The overall specimen dimensions were roughly 127 mm ×14 mm×4 mm and the central reduced section were 30 mm× 6 mm×4 mm. For some specimens to be tested in elevated temperature, after machining to the geometry suitable for mechanical testing, a two-layer EBC was applied on the specimen surfaces to protect the composite against oxidation. The inner layer was mullite and the outer layer was erbium silicate. Both of the layer thickness was 0.1 mm. Monotonic tension and creep tests were conducted under stress control in a servo-hydraulic SHIMADZU test machine, following the general guidelines of ASTM standard C1337 [27]. The specimens were griped by edge load, passive grip interface and the alignment of the testing system was verified at the beginning of the test series. The high temperature furnace and contact-type Epsilon-3548 high temperature
extensometer used for these experiments have also been shown in Fig. 1. Details about the experiment system were in ref [2]. The loading rate at the beginning of the creep test was 3 MPa/s and creep stress range was 40–120 MPa. Creep tests for coated specimens were conducted at 1100 °C and 1300 °C, while for uncoated specimens the tests were only conducted at 1300 °C. After the creep test, fracture surface, cross-section and side surface of the specimens were observed using by scanning electron microscopy (SEM). The inner micro structure evolution was evaluated by computed tomography (CT). X-ray diffraction (XRD) analysis was conducted in order to investigate crystallization of SiC/SiC during creep tests.
3. Results and discussion 3.1. Monotonic tensile behavior Monotonic tensile stress–strain curves of the SiC/SiC composites at room temperature(RT),1100 °C and1300 °C are shown in Fig. 2. All the curves were similar, starting with a linear response, and being followed by a non-linear pseudo-ductile fracture behavior. The ultimate tensile stress(UTS) of composites with coating was considerably higher than
Fig. 1. Test configuration and specimens used for monotonic loading and creep testing at elevated temperatures.
2
Ceramics International xxx (xxxx) xxx–xxx
X. Jing et al.
Fig. 3. Fracture surface of (a) coated and (b) uncoated SiC/SiC composites monotonic tensile tested at 1300 ℃, and magnification of (c) fiber pull-out region and (d) flat region.
similar to the coated specimens. In high stress creep tests, the uncoated specimens showed more rapid deformation and the tertiary creep occurred quite early (Fig. 4c). In low stress creep tests, primary creep was still dominated, but the creep strains were higher than the coated specimens under the same stress (Fig. 4d). Fig. 5 plots the creep strain rate versus time relation of coated specimens at 1100 °C and 1300 °C. The creep rate in 1100 °C tests (Fig. 5a) exhibited decreasing-steady (- increasing) phenomenon. The steady state stages was obvious, starting earlier when the stress was higher. In 1300 °C creep tests (Fig. 5b), the decreasing creep rate of unfailed specimens showed almost linear relation with time with in the test period. In many cases, steady state creep rate, ε̇cr , under applied stress, σ, is represented by the power law equation:
without coating and was close to the UTS at room temperature. The coated specimens maintain the elastic modulus (about 70% of the RT modulus) of composites at 1300 °C, but the modulus for uncoated specimens drop almost 55%. The proportional limits at 1100 °C and 1300 °C almost have the same value of 80 MPa (40%UTS). Fracture surfaces of the coated and uncoated specimens tested at 1300 ℃ were shown in Fig. 3. A flat region without fiber pullout was observed for uncoated specimen, indicating the oxidation occurred rapidly for uncoated composites at elevated temperature [28,29]. Additionally, the absence of the flat oxidation damage region on fracture surface of coated specimen demonstrated the EBC had good oxidation resistance during short-term (not more than 1 h). 3.2. Creep deformation behavior
εcṙ = Aσ n exp (−Q /RT )
The relationship between creep deformation and time at 1100 ℃ and 1300 ℃ is plotted in Fig. 4. At same stress level, the material showed more creep resistance at 1100 °C than 1300 °C. In creep tests at 1100°C (Fig. 4a), quasi-steady state creep dominated the creep deformation. The creep deformation curves of failed specimens exhibited typical three-stage characteristics. After a short primary creep stage, the creep strain rates decreased to a minimum rate ranging from 10−9~10−6/s. Although the creep time during the steady state was long, the creep deformation was small. At high stresses, rapid acceleration of creep rate became more evident just prior to fracture. In creep tests for the coated specimens at 1300 °C (Fig. 4b), transient creep behavior dominated the creep deformation, which was quite different from 1100 °C. At high stresses, the primary creep transferred to tertiary creep directly, without an obvious steady-rate stage. In 40 MPa creep test which accidentally terminated in 30 h, primary creep lasted for the whole tests and no clear evidence of steady creep state was found. The 60 MPa creep test survived after 100 h, steady creep was found after 30 h transient creep. Creep deformation of the uncoated specimens at 1300 °C was
(1)
where A is a dimensionless constant, n the stress exponent, Q the apparent activation energy, R the universal gas constant, and T the absolute temperature. The relationship between applied stress and quasi–steady state creep rate is shown in Fig. 6a. In the low stress region, the creep rate was insensitive to stress and n was almost unity, while the value was large in high stress range. The transition stress where the stress exponent changes was correspond to the proportional limit (60–80 MPa) in the monotonic testing at elevated temperature. Similar results have been reported for Si-Ti-C-O fiber/SiC-based matrix composites [11] and 0°/90° Nicalon SiC/SiC [20] . The stress exponent in the low stress region was similar to Si-C-O fibers, indicating the creep deformation was controlled by fiber bundles. The higher value of n in high stress range was attributed to matrix cracking [11]. Slow crack growth and bridging fiber creep increased the creep rate. It should be noticed that the time point for determining the steady creep strain rate of each specimen was different, especially for the case of 1300 °C tests which did not show secondary creep state. The creep rate decreases gradually from an initial high value until a minimum rate is attained essentially at the point of failure. The timing for calculate 3
Ceramics International xxx (xxxx) xxx–xxx
X. Jing et al.
Fig. 4. Typical tensile creep strain versus time of the composite: (a) tested at 1100 ℃ for coated specimens, (b) tested at 1300 ℃ for coated specimens, (c) tested at 1300 ℃ for uncoated specimens under high stress, (d) tested at 1300 ℃ for uncoated specimens under low stress.
creep rate at high stress was usually less than 2000 s, while in low stress range the time point was usually more than 104 or 105 s. At high stress level, the creep curves terminated early, giving high minimum
creep rates and low rupture lives. In contrast, the creep rate under low stress decayed continuously until fracture occurred after long time, resulting in lower minimum rate. Therefore, a similar relationship was 4
Ceramics International xxx (xxxx) xxx–xxx
X. Jing et al.
Fig. 6. (a)Quasi-steady state creep rate versus applied stress for composites. (b)Creep strain rate at different time versus applied stress for the composites at 1300 °C. Fig. 5. Typical tensile creep strain rate versus time of the composite: (a) tested at 1100 ℃ for coated specimens, (b) tested at 1300 ℃ for coated specimens.
applied for determining ε̇cr at fixed creep time t, as the work done by [12]. The Eq. (1) was utilized to determine the stress dependence at 5×102, 1×103, 5×103, 1×104, 5×104, and 1×105 s for specimens tested at 1300 °C. The stress exponents, n, were determined from the log ε̇cr vs log σ plots (Fig. 6b). Values of n decreased slightly with increasing time and all of them were close to unity, indicating the creep deformations were controlled by Newtonian-type viscous flow. These stress exponents were similar to Nicalon fibers, indicating that the creep rate of the composites was controlled by fiber creep. The microstructure of Si-C-O fibers is unstable above 1200 °C. Fig. 7 shows XRD patterns of the coated composites before and after creep tests at 1300 °C. The results demonstrated the increase in volume and size of the SiC grains. Bodet [30,31] reported that the decomposition of Si-C-O fiber results in a decrease of the SiOxCy volume fraction and an increase in SiC volume fraction, leading to an increase in fiber viscosity and, therefore, creep resistance. The microstructure evolution during creep explained long transient-stage creep deformation at 1300 °C. Due to weak creep resistance of porous matrix, creep deformation was controlled by the fibers, accompanied by cracking of the weak brittle matrix. The matrix, especially derived through PIP process, lost load-bearing capacity in creep test [7]. Due to the stress redistribution during creep, the external load was transferred to fibers. The creep data for the fiber is limited, so the Nicalon fiber data [32] were used here to investigate the creep deformation as both of them are the first generation SiC fibers and similar in chemical and microstructural aspects. Comparing the creep rates of composites tested at 1300 °C at 1×105 s with the creep data of Nicalon fibers, the applied stresses on
Fig. 7. XRD patterns from the composite before and after creep tests at 1300 °C in air.
composites were about one fourth of the stresses on Nicalon fibers at same creep rate (Fig. 8). Given the fiber volume fraction of 0.465 for the SiC/SiC composites, the stress on fiber should be 2.15(1/0.465) times of the applied stress. The discrepancy was attributed to two reasons. On one hand, the off-axis loading of the fiber bundles impaired their load-bearing capacity [33]. Previously study informed that the 20° off-axis loading led to about 35% loss of the strength for these 3D fourdirectional CMCs [2,33]. It was reasonable to consider that creep 5
Ceramics International xxx (xxxx) xxx–xxx
X. Jing et al.
Fig. 8. Comparison of the stress-creep rate relationship of the SiC/SiC composites and Nicalon fibers at 1300 °C.
resistance was influenced by the fiber bundle directions in the preform. On the other hand, the content of O in the fibers is higher than Nicalon fibers, which indicated the KD-1 fibers contain less amount of β-SiC grains and more amorphous Si-O-C. The increase of β-SiC content improves the viscosity and, therefore, creep resistance [31]. 3.3. Creep rupture behavior Fig. 10. Scanning electron micrographs showing typical fracture surface observed in composite creep tested in air: (a) Creep tested at 1100 ℃ under a stress of 120 MPa, the arrows illustrate the defect for oxidation ingression paths, (b) Creep tested at 1300 ℃ under a stress of 40 MPa, the arrows illustrate the undamaged bridging fiber bundles.
Creep rupture behavior of coated specimens tested at 1100 °C and 1300 °C and uncoated specimens tested at 1300 °C are summarized in Fig. 9. Although the data was seen to have considerable scatter, it would appear that a change in creep strength degradation behavior of uncoated specimens occurred in short-term region within 1000 s. The difference in strength degradation behavior was not notable between uncoated and coated specimens tested at 1300 °C for long-term. The temperature effect on strength degradation seemed to be also insignificant between 1100 °C and 1300 °C for coated composites. The threshold stress of creep rupture within 100 h was 40–60 MPa for all three cases. In short term tensile experiments, the coating effectively prevented oxidation damage invasion and the composites hold a relative high ultimate strength. However, when the composites were conducted in long-term loading including fatigue and creep tests, ingression of oxygen into the specimen still took place and flat regions were observed
on the fracture surfaces of these specimens [11,12]. When the stress was relative high, the morphology of the fracture surface was similar to uncoated specimens under monotonic test at elevated temperatures. Fig. 10(a) shows the fracture surface of coated specimen tested at 1100 °C with a creep stress of 120 MPa. The undamaged region occupied almost 47% area of whole cross section and it was not located in the central zone, indicating that the oxygen ingression did not propagate evenly along the periphery region of the specimen but initiated at surfaces or corners where coating had defects and penetrated the specimen through the unbridged cracks. The large crack opening displacements associated with a growing surface crack would then allow direct oxygen penetration, promoting fiber failure in the crack plane. In this way, the dominant crack would grow rapidly until the remaining unbroken fiber bundles were unable to carry the load, resulting in a pull-out zone in one region which may or may not be located centrally on the fracture surface. When the stress was lower and the loading time was longer, the flat fracture surface caused by oxidation damage occupied the most area of the specimen fracture surface. The fiber pullout regions sometimes could not be connected to a continuous piece. The fracture surface of specimens tested at 1300 ℃ under a creep stress of 40 MPa is shown in Fig. 10(b). Several distributed fiber pullout regions were embedded in the flat surface. The area occupied by pullout fiber bundles regions added up to 15% of the whole cross section area. The area of each undamaged regions was small and each region usually consisted only one or two fiber bundles, which indicated that to an extent, failure occurred by growth and link up of several separately-nucleated cracks. Additionally, the bundle/bundle interface still took effect on deflecting the damage propagation, leading to fiber bundles bridging the crack. As the load became lower, the microstructure had more influence on the oxidation damage propagation. For this SiC/SiC composite, creep of the fibers was accompanied by cracking of the brittle matrix. To understand the damage during the
Fig. 9. Creep rupture data for uncoated composites at 1300 ℃ and coated composites at 1100 ℃ and 1300 ℃.
6
Ceramics International xxx (xxxx) xxx–xxx
X. Jing et al.
Fig. 11. Cracks distribution in the specimens at 1300 °C under a creep stress of (a)(c)40 MPa and (b)(d) 120 MPa.
failure and sudden fracture occurred by fiber pull-out. Fig. 11 shows the cracks in the composites under 40 MPa and 120 MPa at 1300 °C. The crack density was greater for 40 MPa than 120 MPa. The cracks for 40 MPa were all bridged by fibers, but some cracks for 120 MPa were unbridged. After examining the crack in other specimens, it could be concluded that at lower stresses, the long rupture time lead to denser crack distribution. However, as stress increased, more unbridged cracks were observed. The unbridged cracks were difficult to be observed for the specimens tested under medium stress level, except the crack leading to ultimate fracture. The matrix cracking evolution, therefore, should be not only depend on stress, but also influenced by time. The EBC played an important role on maintaining the mechanical properties of the SiC/SiC composites for short-term loading at elevated temperatures. As the creep time increase, the difference for coated and uncoated composites became smaller, including the creep strain rate and rupture time. The effect of coating degraded as the load and time
creep, the damage processes leading to failure were clarified by microstructural examination and fractographic studies. The fibers in undamaged region were thought as ‘effective fibers’ that contribute to the composite ultimate strength [11,34,35]. It could be possible to conclude that oxidation damage propagated in the composites gradually through the flat unbridged matrix crack plane and the specimens failed when the fiber bundles in the rest undamaged region could not sustain the external load. According to the examination, most of the cracks were located in the fiber-free matrix region, arrested by the individual bundles of braided preforms. The tunneling cracks between fiber bundles made the stress redistribution, leading to fiber bear more load. Thus indicate that the fiber dominate the composites creep deformation behavior. Few cracks penetrated into fiber bundles normal to the loading direction, especially near the surface, because of oxidation embrittlement. Then oxidation-assisted fiber failure occurred in crack tip, leading to matrix cracks propagate inward. That explained the fracture surface contains the flat crack growth zones with fiber 7
Ceramics International xxx (xxxx) xxx–xxx
X. Jing et al.
Fig. 12. Cracks on coating for the specimens under (a)10 MPa, and (b) 40 MPa for 30 h at 1300 °C. (c) cracks between coating and SiC/SiC for 10 MPa specimens.
Acknowledgement
increased. Fig. 12(a) shows cracks were observed on the coating of the specimen which was tested for 30 h at 1300 °C under a very small load (not more than 10 MPa). As the stress increased to 40 MPa (Fig. 12(b)), more cracks were observed. Some pieces of the coatings began to be peeled off (Fig. 12(c)), leading to the area underwent more severe oxidation damage. The coating, therefore, could protect the composites effectively from oxidation ingress as long as the integrity was maintained. Once the oxygen invaded the composites, the mechanical performance of the SiC/SiC composites would decrease dramatically.
The research was supported by National Natural Science Foundation of China (No. NSFC51275023). References [1] Z. Luo, X. Zhou, J. Yu, Mechanical properties of SiC/SiC composites by PIP process with a new precursor at elevated temperature, Mater. Sci. Eng.: A 607 (5) (2014) 155–161. [2] J. Xin, et al., Fiber strength measurement for KD-I(f)/SiC composites and correlation to tensile mechanical behavior at room and elevated temperatures, Ceram. Int. 41 (1) (2015) 299–307. [3] S. Zhu, et al., Monotonic tension, fatigue and creep behavior of SiC-fiber-reinforced SiC-matrix composites: a review, Compos. Sci. Technol. 59 (6) (1999) 833–851. [4] S. Zhu, et al., Creep and fatigue behavior in Hi-Nicalon-Fiber-reinforced silicon carbide composites at high temperatures, J. Am. Ceram. Soc. 82 (1) (1999) 117–128. [5] S. Zhu, et al., Creep and fatigue behavior in an enhanced SiC/SiC composite at high temperature, J. Am. Ceram. Soc. 81 (9) (1998) 2269–2277. [6] S. Zhu, et al., Tensile creep behavior of a SiC-fiber/SiC composite at elevated temperatures, Compos. Sci. Technol. 57 (12) (1998) 1629–1637. [7] G.N. Morscher, Tensile creep and rupture of 2D-woven SiC/SiC composites for high temperature applications, J. Eur. Ceram. Soc. 30 (11) (2010) 2209–2221. [8] G.N. Morscher, et al., Creep in vacuum of woven Sylramic-iBN melt-infiltrated composites, Compos. Sci. Technol. 71 (1) (2011) 52–59. [9] G.N. Morscher, V.V. Pujar, Creep and stress–strain behavior after creep for SiC fiber reinforced, melt-infiltrated SiC matrix composites, J. Am. Ceram. Soc. 89 (5) (2006) 1652–1658. [10] G.N. Morscher, et al., Tensile creep and fatigue of Sylramic-iBN melt-infiltrated SiC matrix composites: retained properties, damage development, and failure mechanisms, Compos. Sci. Technol. 68 (15–16) (2008) 3305–3313. [11] T. Ogasawara, et al., Tensile creep behavior of 3-D woven Si-Ti-C-O fiber/SiC-based matrix composite with glass sealant. J. Mater. Sci. vol. 35(4). pp. 785–793. [12] T. Ogasawara, et al., Tensile creep behavior and thermal stability of orthogonal three-dimensional woven TyrannoTM ZMI fiber/silicon-titanium-carbon-oxygen matrix composites, J. Am. Ceram. Soc. 85 (2) (2002) 393–400. [13] B. Wilshire, Creep property comparisons for ceramic-fibre-reinforced ceramic– matrix composites, J. Eur. Ceram. Soc. 22 (8) (2002) 1329–1337. [14] B. Wilshire, F. Carreño, Deformation and damage processes during tensile creep of ceramic-fibre-reinforced ceramic–matrix composites, J. Eur. Ceram. Soc. 20 (4) (2000) 463–472. [15] G. Boitier, J.L. Chermant, J. Vicens, Understanding the creep behavior of a 2.5D Cf–SiC composite: II. Experimental specifications and macroscopic mechanical creep responses, Mater. Sci. Eng.: A 289 (1–2) (2000) 265–275. [16] G. Boitier, J. Vicens, J.L. Chermant, Understanding the creep behavior of a 2.5D Cf–SiC composite-I. Morphology and microstructure of the as-received material, Mater. Sci. Eng.: A 279 (1–2) (2000) 73–80.
4. Conclusion The tensile creep behavior of a three-dimensional four-step braided SiC fiber/SiC ceramic matrix composite with and without coating at 1100 °C and 1300 °C in air has been investigated. The following conclusions have been made. (1) The creep deformation for the composites at 1100 °C contained primary and secondary stage while the creep deformation for composites at 1300 °C mainly exhibited primary stage. (2) The fibers control the rates of creep deformation. The stress exponent was similar to that of Si-C-O fibers found in literature. (3) The matrix cracking was different under high or low stress. The dominant crack grows rapidly by direct ingress of oxygen along the crack plane under high stress. In contrast, cracks form extensively throughout the specimens under low stress. The reduced crack growth rates lead to longer time and high strains before failure. The undamaged area ratio increases with decreasing stress. (4) The creep rupture behavior of the composite is suggested to depend on the factor of unbridged cracking propagation and residual strength. As the developing cracks penetrate into the fiber bundles, crack growth rates are limited by crack-bridging fibers, but fiber failure is promoted by oxygen ingress. Creep fracture then takes place when the area of the developing cracks attain the critical fraction of the test-piece cross-section required for sudden failure by fiber pull-out.
8
Ceramics International xxx (xxxx) xxx–xxx
X. Jing et al.
15) (2002) 2741–2747. [27] ASTM, Standard Test Method for Creep and Creep Rupture of Continuous FiberReinforced Advanced Ceramics Under Tensile Loading at Elevated Temperatures, in C1337-10, 2010. [28] I.J. Davies, et al., Fibre strength parameters measured in situ for ceramic-matrix composites tested at elevated temperature in vacuum and in air, Compos. Sci. Technol. 59 (6) (1999) 801–811. [29] A.G. Evans, et al., Models of high-temperature, environmentally assisted embrittlement in ceramic-matrix composites, J. Am. Ceram. Soc. 79 (9) (1996) 2345–2352. [30] R. Bodet, et al., Microstructural stability and creep behavior of Si-C-O (Nicalon) fibers in carbon monoxide and argon environments, J. Am. Ceram. Soc. 79 (10) (1996) 2673–2686. [31] N. Jia, R. Bodet, R.E. Tressler, Effects of microstructural instability on the creep behavior of Si-C-O (Nicalon) fibers in argon, J. Am. Ceram. Soc. 76 (12) (1993) 3051–3060. [32] G. Simon, A.R. Bunsell, Creep Behaviour and Structural Characterization at High Temperatures of Nicalon SiC Fibres, vol. 19(11), 1984, pp. 3658–3670. [33] G.N. Morscher, H.M. Yun, J.A. DiCarlo, In-plane cracking behavior and ultimate strength for 2D woven and braided melt-infiltrated SiC/SiC composites tensile loaded in off-axis fiber directions, J. Am. Ceram. Soc. 90 (10) (2007) 3185–3193. [34] I.J. Davies, et al., Fibre strength parameters measured in situ for ceramic-matrix composites tested at elevated temperature in vacuum and in air, Compos. Sci. Technol. 59 (6) (1999) 801–811. [35] I.J. Davies, et al., Case study of failure in a glass-sealed SiC/SiC-based composite creep tested at 1100 °C in air, Mater. Lett. 48 (3–4) (2001) 205–209.
[17] G. Boitier, J. Vicens, J.L. Chermant, Understanding the creep behavior of a 2.5D Cf–SiC composite. III. From mesoscale to nanoscale microstructural and morphological investigation towards creep mechanism, Mater. Sci. Eng.: A 313 (1–2) (2001) 53–63. [18] J.L. Chermant, et al., The creep mechanism of ceramic matrix composites at low temperature and stress, by a material science approach, J. Eur. Ceram. Soc. 22 (14– 15) (2002) 2443–2460. [19] C.H. Weber, et al., High temperature deformation and rupture in SiC-C composites, Mater. Sci. Eng.: A 196 (1–2) (1995) 25–31. [20] J.W. Holmes, X. Wu, Elevated temperature creep behavior of continuous fiber reinforced ceramics, in: S. Nair, K. Jakus, S. Nair, K. Jakus (Eds.), High Temperature Mechanical Behavior of Ceramic Compsites, ButterworthHeinemann, 1995, pp. 193–259. [21] J. DiCarlo, H. Yun, Non-oxide (silicon carbide) fibers, in: N. Bansal, N. Bansal (Eds.), Handbook of Ceramic Composites, Kluwer Academic, 2005. [22] F.E. Heredia, et al., Oxidation embrittlement probe for ceramic-matrix composites, J. Am. Ceram. Soc. 78 (8) (1995) 2097–2100. [23] H. Lin, P.F. Becher, Stress-temperature-lifetime working envelope of Nicalon fiberreinforced SiC matrix composites in air, Compos. Part A: Appl. Sci. Manuf. 28 (11) (1997) 935–942. [24] A.R. Bunsell, A. Piant, A review of the development of three generations of small diameter silicon carbide fibres, J. Mater. Sci. 41 (3) (2006) 823–839. [25] I.J. Davies, et al., Case study of failure in a glass-sealed SiC/SiC-based composite creep tested at 1100 °C in air, Mater. Lett. 48 (3–4) (2001) 205–209. [26] H.E. Eaton, G.D. Linsey, Accelerated oxidation of SiC CMC's by water vapor and protection via environmental barrier coating approach, J. Eur. Ceram. Soc. 22 (14–
9