Corrosion Science 50(2008)3132-3138 Contents lists available at Science Direct Corrosion science ELSEVIER journalhomepagewww.elsevier.com/locate/corsci Thermal and mechanical stabilities of Hi-Nicalon SiC fiber under annealing and creep in various oxygen partial pressures - Sha,, T Hinoki, A Kohyama Graduate School of energy Science, Kyoto University, Gokasho, Uji Kyoto 611-0011, Japan b Institute of Advanced Energy, Kyoto University. Gokasho, Uji, Kyoto 611-0011, Japan ARTICLE INF O A BSTRACT Article history Thermal and mechanical stabilities were investigated on Hi-Nicalon SiC fibers under annealing and creep Accepable 18May2008 13 August 2008 in various oxygen partial pressures by mass change, mechanical properties as well as microstructural online 19 August 2008 features. In the case of mechanical stability, the tensile strength of fiber is strongly dependent on the oxy- en partial pressure of testing environment, but a weak dependence of BSr creep resistance on oxygen artial pressure is appeared. The analyses of surface morphology and mass change indicated the thermal Keywords: A Ceramic tability of fiber under annealing in different environments was different. At different exposure temper B. X-ray diffraction C Oxidatio by the stress applied through BSR test. This means the thermal stability of fibers is related to not only t <posed environments, but also the mechanical state of fibers. G 2008 Elsevier Ltd. All rights reserved. 1 Introduction pressures. In such case, the Sic materials would be oxidized in pas Ceramic-matrix composites(CMCs)are been considering as the degradation of Sic materials in oxidizing environments strongly potential structural materials for advanced energy generation sys- depends on the oxidation mechanism. Jacobson[6] has generalized tems and propulsion systems. As we know, the monolithic ceramic the oxidation degradation mechanism of Sic materials in varied naterials are very brittle in the fracture behaviour. with the rein environments, but it is still insufficient because of the complexity forcement of fibers, the fracture toughness and impact capability of of service environments. The key question concerns the oxidation ceramic materials could be improved significantly. Thus, the load- kinetics: passive and active oxidation. This topic has given rise to bearing SiC fibers as reinforcements in CMCs are backbone, and much controversy for Sic materials, because the temperature acceptable performance of high temperature CMCs strongly de- boundaries for the oxidation kinetics are quite dependent not only pends upon judicious selection and incorporation of ceramic fiber on the materials themselves(purity and crystallinity) but also on reinforcement with the proper chemical, physical and mechanical the specific service environment(exposure temperature, oxygen properties. Low fiber strength and thermal stability could result partial pressure and mechanical state ) Furthermore, rarely is one in low fracture toughness and accelerate sub-critical crack mechanism operative in performance degradation of SiC materials. gation in CMCs In the past few decade years, extensive efforts have In practice, several mechanisms operate simultaneously een made to develop the SiC fibers in order to satisfy the require- SiC fibers as the reinforcements of structural materials for higl ments from the high temperature technologies. Especially, in re- temperature technologies, the most desired critical properties are cent years, the Sic fibers with high thermal stability, which were excellent high temperature mechanical and thermal stabilities. cured by electron-beam irradiation [1], are considering as the Thus, in an specific service environment, the environmental dura- promising reinforcement in CMCs. However, the mechanical and bility and the response of reinforced fibers to service environment thermal stabilities of SiC fibers as reinforcements in CMCs are very are major concern and they should be revealed for the reliability sensitive to their purity, crystallinity and service environments[2- evaluation of CMCs. In our previous works [7, 8] the fracture fea- 4 including thermal and loading history. ures for this fiber have been characterized, but no attempts were For high temperature applications, the CMCs are often sub- made to correlate the environment with microstructure and high jected to oxidative environments with different oxygen partial temperature mechanical properties. For practie dication of MCs, it is requested to accumulate Corresponding author. Tel: +81 774 38 3463: fax: +81 774 38 3467 veal the degradation mechanism of Sic fibers with a consideration E-mailaddress:shajianjun720@yahoo.com(.].Sha). of the mechanical and thermal stabilities 0010-938X/s front matter 2008 Elsevier Ltd. all rights reserved. do:101016/ J- corsi2008.08003
Thermal and mechanical stabilities of Hi-Nicalon SiC fiber under annealing and creep in various oxygen partial pressures J.J. Sha a,*, T. Hinoki b , A. Kohyama b aGraduate School of Energy Science, Kyoto University, Gokasho, Uji, Kyoto 611-0011, Japan b Institute of Advanced Energy, Kyoto University, Gokasho, Uji, Kyoto 611-0011, Japan article info Article history: Received 18 May 2008 Accepted 13 August 2008 Available online 19 August 2008 Keywords: A. Ceramic B. SEM B. X-ray diffraction C. Oxidation abstract Thermal and mechanical stabilities were investigated on Hi-Nicalon SiC fibers under annealing and creep in various oxygen partial pressures by mass change, mechanical properties as well as microstructural features. In the case of mechanical stability, the tensile strength of fiber is strongly dependent on the oxygen partial pressure of testing environment, but a weak dependence of BSR creep resistance on oxygen partial pressure is appeared. The analyses of surface morphology and mass change indicated the thermal stability of fiber under annealing in different environments was different. At different exposure temperatures and oxygen partial pressure levels, the different oxidation regimes are responsible for the strength degradation and microstructure change of fibers. Furthermore, the microstructure change is also affected by the stress applied through BSR test. This means the thermal stability of fibers is related to not only the exposed environments, but also the mechanical state of fibers. 2008 Elsevier Ltd. All rights reserved. 1. Introduction Ceramic–matrix composites (CMCs) are been considering as the potential structural materials for advanced energy generation systems and propulsion systems. As we know, the monolithic ceramic materials are very brittle in the fracture behaviour. With the reinforcement of fibers, the fracture toughness and impact capability of ceramic materials could be improved significantly. Thus, the loadbearing SiC fibers as reinforcements in CMCs are backbone, and acceptable performance of high temperature CMCs strongly depends upon judicious selection and incorporation of ceramic fiber reinforcement with the proper chemical, physical and mechanical properties. Low fiber strength and thermal stability could result in low fracture toughness and accelerate sub-critical crack propagation in CMCs. In the past few decade years, extensive efforts have been made to develop the SiC fibers in order to satisfy the requirements from the high temperature technologies. Especially, in recent years, the SiC fibers with high thermal stability, which were cured by electron-beam irradiation [1], are considering as the promising reinforcement in CMCs. However, the mechanical and thermal stabilities of SiC fibers as reinforcements in CMCs are very sensitive to their purity, crystallinity and service environments [2– 4] including thermal and loading history. For high temperature applications, the CMCs are often subjected to oxidative environments with different oxygen partial pressures. In such case, the SiC materials would be oxidized in passive/active oxidation regime [5]. As we know, the performance degradation of SiC materials in oxidizing environments strongly depends on the oxidation mechanism. Jacobson [6] has generalized the oxidation degradation mechanism of SiC materials in varied environments, but it is still insufficient because of the complexity of service environments. The key question concerns the oxidation kinetics: passive and active oxidation. This topic has given rise to much controversy for SiC materials, because the temperature boundaries for the oxidation kinetics are quite dependent not only on the materials themselves (purity and crystallinity), but also on the specific service environment (exposure temperature, oxygen partial pressure and mechanical state). Furthermore, rarely is one mechanism operative in performance degradation of SiC materials. In practice, several mechanisms operate simultaneously. SiC fibers as the reinforcements of structural materials for high temperature technologies, the most desired critical properties are excellent high temperature mechanical and thermal stabilities. Thus, in an specific service environment, the environmental durability and the response of reinforced fibers to service environment are major concern and they should be revealed for the reliability evaluation of CMCs. In our previous works [7,8], the fracture features for this fiber have been characterized, but no attempts were made to correlate the environment with microstructure and high temperature mechanical properties. For practical application of CMCs, it is requested to accumulate experimental data and to reveal the degradation mechanism of SiC fibers with a consideration of the mechanical and thermal stabilities. 0010-938X/$ - see front matter 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.corsci.2008.08.003 * Corresponding author. Tel.: +81 774 38 3463; fax: +81 774 38 3467. E-mail address: shajianjun720@yahoo.com (J.J. Sha). Corrosion Science 50 (2008) 3132–3138 Contents lists available at ScienceDirect Corrosion Science journal homepage: www.elsevier.com/locate/corsci
J Sha et al. /Corrosion Science 50(2008)3132-3138 3133 herefore, for understanding the mechanical and thermal sta bilities and failure mechanism of Sic fibers over a wide range of 一A temperatures and varied environments, this work proceeded a complementary investigation on the microstructure features and high temperature properties of Hi-Nicalon fibers under annealing and creep in various oxygen partial pressures at elevated temper- atures, and attempted to clarify the correlation between the envi- ronment with mechanical and thermal stabilities with this result a further discussion was made on the environment-pertinent 502 Tk 2. Experimental The SiC fiber examined in this study is Hi-Nicalon(C/Si atomic ratio: 1.38, oxygen: 0.5 wt%, diameter: 14 um). This fiber was an- nealed and crept in air(O2: 20%, dew point: 3C), high-purity Ar 1000110012001300140015001600 (HP-Ar, 02: 2 ppm, dew point:-55C)and ultra high-purity Ar HP-Ar, 02: 0. 1 ppb: dew point:-5.5C)under flowing atmo- here with a pressure of 10 Pa and held for 1 h at desired tem- Fig 1. Mass change of Hi-Nicalon hiber under annealing at elevated temperatures in peratures ranging from 1000 to 1500C. Fibers' annealing was fferent atmospheres for 1h. performed on the 5 cm fragments which were positioned in the hot zone of furnace chamber. The mass change was measured by electronic balance(mass resolution: #0. 1 mg). After annealing at 3.3. Tensile properties and creep resistance 1500C, individual fiber was carefully separated and pulled out from the fiber bundle for single fiber tensile test by a technique Fig 3 shows the tensile stress-strain curve for fibers un as described in our previous studies[8-10 annealing in different environments at 1500C. The tensile stress The creep resistance was assessed by bend stress relaxation (BSR)method which was developed by Morscher [11]. The detailed ven that conversion of load to tensile stress would depend on the configuration of test jig can be found elsewhere 121 By means of specific diameter of each fiber. In this work, the individual fiber this configuration, the influence of environment on BSR creep diameter was measured by SEM image but not using the average resistance could be evaluated. The meter used to index the diameter. The true value of ultimate tensile strength can be read reep resistance is the bend stress relaxation parameter m, which from the stress-strain curve is defined as: m=1-Ro/Ra, where ro and Ra are, respectively the Fig 4 shows the dependence of mean strength on the testing curvature for the initially imposed bend strain and the residual environments. The fiber's strength decreased with decreasing the curvature after thermal exposure and strain removal, if m-1, no oxygen partial pressure It should be noted during the specimen relaxation has occurred: if m=0, complete relaxation has occurred preparation that fibers with low strength became very difficult to (Ro=Ra). Furthermore, the surface morphologies of fibers were set without breaking them. The mean strength we gave will conse- examined by the observation of field-emission scanning electron quently not take the weakest fibers into account(no strength could microscopy(FE-SEM): the phase in fibers was analyzed by X-ray be obtained ) Due to this shortcoming, overestimation of tensile diffraction(XRD). strength is likely. As observed for fibers after annealing in UH Ar at 1500C for 1 h, the fibers are too fragile to measure the 3. Results Furthermore, Fig. 4 shows the dependence of 1-h BSR 3.1. Mass chang resistance m on testing environments at 1300C. The BSr Fig. 1 shows the mass change for fibers under annealing in different oxygen partial pressure atmospheres for 1 h at elevated is observed for fibers under UHP-Ar in air, but the mass gain at 1400C is lower than that at 1300C. For the fibers under annealing in inert atmospher (HP-Ar and UHP-Ar), a mass loss is observed at temperature be- yond1300°C 3. 2. X-ray diffraction patten Fig 2 shows X-ray diffraction patterns for fibers under anneal ing in different oxygen partial pressures for 1 h at temperatures ranging from 1300 to 1500C. Obviously, the fibers under anneal ing in air result in the formation of silica peaks with different height at 20=22 which are identified as the cristobalite but the silica peak does not appear for fibers under annealing in inert atmospheres. Furthermore, it can be seen in Fig. 2 that B-Sic peaks 1520253035404550556065707580 become sharper for fibers under annealing at temperature over 2 Theta 1300C in comparison to as-received fibers, but it is more obvious Fig. 2. X-ray diffraction patterns for Hi-Nicalon fibers under annealing in different for fibers under annealing at 1500C. oxygen partial pressure atmospheres,◆:阝Sic:▲: cristobalite
Therefore, for understanding the mechanical and thermal stabilities and failure mechanism of SiC fibers over a wide range of temperatures and varied environments, this work proceeded a complementary investigation on the microstructure features and high temperature properties of Hi-Nicalon fibers under annealing and creep in various oxygen partial pressures at elevated temperatures, and attempted to clarify the correlation between the environment with mechanical and thermal stabilities. With this result, a further discussion was made on the environment-pertinent performance. 2. Experimental The SiC fiber examined in this study is Hi-Nicalon (C/Si atomic ratio: 1.38, oxygen: 0.5 wt%, diameter: 14 lm). This fiber was annealed and crept in air (O2: 20%, dew point: 3 C), high-purity Ar (HP-Ar, O2: 2 ppm, dew point: 5.5 C) and ultra high-purity Ar (UHP-Ar, O2: 0.1 ppb: dew point: 5.5 C) under flowing atmosphere with a pressure of 105 Pa and held for 1 h at desired temperatures ranging from 1000 to 1500 C. Fibers’ annealing was performed on the 5 cm fragments which were positioned in the hot zone of furnace chamber. The mass change was measured by electronic balance (mass resolution: ±0.1 mg). After annealing at 1500 C, individual fiber was carefully separated and pulled out from the fiber bundle for single fiber tensile test by a technique as described in our previous studies [8–10]. The creep resistance was assessed by bend stress relaxation (BSR) method which was developed by Morscher [11]. The detailed configuration of test jig can be found elsewhere [12]. By means of this configuration, the influence of environment on BSR creep resistance could be evaluated. The parameter used to index the creep resistance is the bend stress relaxation parameter m, which is defined as: m = 1 R0/Ra, where R0 and Ra are, respectively, the curvature for the initially imposed bend strain and the residual curvature after thermal exposure and strain removal, if m = 1, no relaxation has occurred; if m = 0, complete relaxation has occurred (R0 = Ra). Furthermore, the surface morphologies of fibers were examined by the observation of field-emission scanning electron microscopy (FE-SEM); the phase in fibers was analyzed by X-ray diffraction (XRD). 3. Results 3.1. Mass change Fig. 1 shows the mass change for fibers under annealing in different oxygen partial pressure atmospheres for 1 h at elevated temperatures. A mass gain is observed for fibers under annealing in air, but the mass gain at 1400 C is lower than that at 1300 C. For the fibers under annealing in inert atmosphere (HP-Ar and UHP-Ar), a mass loss is observed at temperature beyond 1300 C. 3.2. X-ray diffraction pattern Fig. 2 shows X-ray diffraction patterns for fibers under annealing in different oxygen partial pressures for 1 h at temperatures ranging from 1300 to 1500 C. Obviously, the fibers under annealing in air result in the formation of silica peaks with different height at 2h = 22 which are identified as the cristobalite, but the silica peak does not appear for fibers under annealing in inert atmospheres. Furthermore, it can be seen in Fig. 2 that b-SiC peaks become sharper for fibers under annealing at temperature over 1300 C in comparison to as-received fibers, but it is more obvious for fibers under annealing at 1500 C. 3.3. Tensile properties and creep resistance Fig. 3 shows the tensile stress–strain curve for fibers under annealing in different environments at 1500 C. The tensile stress was calculated from the load acquired during the tensile test. Given that conversion of load to tensile stress would depend on the specific diameter of each fiber. In this work, the individual fiber diameter was measured by SEM image but not using the average diameter. The true value of ultimate tensile strength can be read from the stress–strain curve. Fig. 4 shows the dependence of mean strength on the testing environments. The fiber’s strength decreased with decreasing the oxygen partial pressure. It should be noted during the specimen preparation that fibers with low strength became very difficult to set without breaking them. The mean strength we gave will consequently not take the weakest fibers into account (no strength could be obtained). Due to this shortcoming, overestimation of tensile strength is likely. As observed for fibers after annealing in UHPAr at 1500 C for 1 h, the fibers are too fragile to measure the strength. Furthermore, Fig. 4 shows the dependence of 1-h BSR creep resistance m on testing environments at 1300 C. The BSR creep -2 -1 0 1 2 3 4 1000 1100 1200 1300 1400 1500 1600 Air HP-Ar UHP-Ar Mass change (%) Temperature(°C) Fig. 1. Mass change of Hi-Nicalon fiber under annealing at elevated temperatures in different atmospheres for 1 h. Intensity 15 20 25 30 35 40 45 50 55 60 65 70 75 80 2 Theta Air HP-Ar UHP-Ar As-received 1500 C 1400 C 1300 C 1500 C 1400 C 1300 C 1500 C 1400 C 1300 C Fig. 2. X-ray diffraction patterns for Hi-Nicalon fibers under annealing in different oxygen partial pressure atmospheres, : b SiC; N: cristobalite. J.J. Sha et al. / Corrosion Science 50 (2008) 3132–3138 3133
J Sha et al/ Corrosion Science 50(2008)3132-3138 3.4.2. Fibers under annealing and creep in HP-Ar n As-received ig. 6 shows the morphologies of Hi-Nicalon SiC fibers under annealing and creep in HP-Ar at elevated temperatures for 1 h. For the fibers under annealing(Fig. 6a)and creep(Fig. 6d)at 1300C, fibers surface did not show obvious oxygen attack. For the Hi-Nicalon fibers under creep at 1400C(Fig. 6e), the bubbles were formed on the surface of fiber, but it did not appear under annealing condition except for the slight grain growth(Fig. 6b). The fibers annealing at 1500C have coarse surface accompanying with the formation of pits( Fig. 6c). In particular, for the fiber under creep at 1500C(Fig. 6f). it gave a porous microstructure with 0.5 3.4.3. Fibers under annealing and creep in UHP-Ar Fig. 7 shows the morphologies of Hi-Nicalon fibers under annealing and creep in UHP-Ar at elevated temperatures for 1 h. No obvious oxidation was observed for hi-Nicalon fibers under Fig. 3. Tensile test for Hi-Nicalon fibers under annealing in different annealing at 1300C, but for fibers under creep condition, bubbles under annealing condition, many large pits was produced on the surface of fibers( Fig. 7b), meanwhile, the tensile side of crept fiber showed the needle-like grains with a length of about 5 um(Fig. 7e). O Stress relaxation parameter In particular, for the Hi-Nicalon fibers under annealing and 1500C, the fibers were oxidized more severely and much coarse- grained surface was produced( Fig. 7c and f). It is clear these huge crystals 25 4 Discussion The oxidation mechanism of Sic is very complex, which may be ategorized the passive and active oxidation regimes. The regime of oxidative reaction is quite dependent on the actual chemica omposition of fibers, oxygen partial pressure and exposure tem perature as well as other factors 5, 13-15. The thermo-chemical correlation and oxidation dynamics for active and passive oxida- As-received Air HP-Ar UHP-A tion of silicon carbide have been investigated experimentally and theoretically in the literatures [5, 13-15]. From which we could know passive oxidation would lead to the formation of Sio2 film Fig. 4. The tensile strength re and stress relaxation parameter(m)of Hi- protecting the Sic materials from further oxidation attack, in the licalon fibers under varied oxygen partial pressures, tensile strengths wer neasured on fibers under annealing at 1500C; stress relaxation parameters wer case of active oxidation the sio formation rate is too low to seal xamined at1300°C he surface of materials, resulting in the mass loss and pits formation The oxygen partial pressure for the transition from passive oxi- resistance of Hi-Nicalon fibers under high oxygen partial pressure dation to active oxidation at 1300C was po=10-2 Pa[5].The is somewhat lower than that of fibers in low oxygen partial pres- nominal Po, of testing environments in this study was about ure, namely, a weak dependence of BSr creep resistance on oxy- 2x 10" Pa(air ). 0.2 Pa(HP-Ar)and 10 Pa(UHP-Ar), indicating gen partial pressure is observed hat fiber would be oxidized in passive and active oxidation regime is quite possible. And also, with increasing the exposure tempera- 3.4. Morphologies of fibers ture, the critical value of po, for the transition from passive to ac- tive oxidation would be increased. The selection of the test 3.4.1. Fibers under annealing and creep in air condition in this work would be helpful in understanding the oxi- Fig. 5 shows the morphologies of Hi-Nicalon SiC fibers under dation behavior of SiC fibers. nnealing and creep in air at elevated temperatures for 1 h. The In Fig. 1, the mass gain for fibers under annealing in air is due to Sioz film formed uniformly on the surface of fibers during anneal- the passive oxidation of fibers with atmospheric oxygen. However, ing at elevated temperatures( Fig. 5a-c). the mass gain is not smoothly increased with increasing the tem or the Hi-Nicalon fibers under annealing at 1500C Fig. 5b). perature. For fiber under annealing in air at 1400C. it was oxi- and creep at temperatures above 1300C. the cracks were found dized mildly. The fiber under annealing in air at 1500C was vithin silica film(Fig. 5d and e). Some patterns were also observed again subjected to a high degree of oxidation. This result is consis within the silica film formed on the surface of Hi-Nicalon fibers tent with the Xrd pattern showed in Fig. 2: ( i)the cristobalite peak (Fig. 5d and e) under creep condition. at 20=22 for fiber under annealing in air indicated that fiber was From the cross-section of fibers under annealing at 1500c oxidized mainly in passive oxidation regime; (ii)a wide cristobalite (Figs 2e and 3e), we could see that the silica layer consisted of a peak(20=22)with relatively low intensity for fibers under concentric sheath and the fibers surface was blanketed well by annealing at 1400C was observed in comparison to that of fibers silica layer. No cracks were found between the silica layer and under annealing at 1300 and 1500C, showing the passive oxida- non-oxidized fiber core tion at 1400C was refrained. This is due to the quick formation
resistance of Hi-Nicalon fibers under high oxygen partial pressure is somewhat lower than that of fibers in low oxygen partial pressure, namely, a weak dependence of BSR creep resistance on oxygen partial pressure is observed. 3.4. Morphologies of fibers 3.4.1. Fibers under annealing and creep in air Fig. 5 shows the morphologies of Hi-Nicalon SiC fibers under annealing and creep in air at elevated temperatures for 1 h. The SiO2 film formed uniformly on the surface of fibers during annealing at elevated temperatures (Fig. 5a–c). For the Hi-Nicalon fibers under annealing at 1500 C Fig. 5b), and creep at temperatures above 1300 C, the cracks were found within silica film (Fig. 5d and e). Some patterns were also observed within the silica film formed on the surface of Hi-Nicalon fibers (Fig. 5d and e) under creep condition. From the cross-section of fibers under annealing at 1500 C (Figs. 2e and 3e), we could see that the silica layer consisted of a concentric sheath and the fiber’s surface was blanketed well by silica layer. No cracks were found between the silica layer and non-oxidized fiber core. 3.4.2. Fibers under annealing and creep in HP-Ar Fig. 6 shows the morphologies of Hi-Nicalon SiC fibers under annealing and creep in HP-Ar at elevated temperatures for 1 h. For the fibers under annealing (Fig. 6a) and creep (Fig. 6d) at 1300 C, fibers surface did not show obvious oxygen attack. For the Hi-Nicalon fibers under creep at 1400 C (Fig. 6e), the bubbles were formed on the surface of fiber, but it did not appear under annealing condition except for the slight grain growth (Fig. 6b). The fibers annealing at 1500 C have coarse surface accompanying with the formation of pits (Fig. 6c). In particular, for the fiber under creep at 1500 C (Fig. 6f), it gave a porous microstructure with many pits formation. 3.4.3. Fibers under annealing and creep in UHP-Ar Fig. 7 shows the morphologies of Hi-Nicalon fibers under annealing and creep in UHP-Ar at elevated temperatures for 1 h. No obvious oxidation was observed for Hi-Nicalon fibers under annealing at 1300 C, but for fibers under creep condition, bubbles and large grains were formed on the surface (Fig. 7d). At 1400 C under annealing condition, many large pits was produced on the surface of fibers (Fig. 7b), meanwhile, the tensile side of crept fiber showed the needle-like grains with a length of about 5 um (Fig. 7e). In particular, for the Hi-Nicalon fibers under annealing and creep at 1500 C, the fibers were oxidized more severely and much coarsegrained surface was produced (Fig. 7c and f). It is clear these huge crystals grew outward from the fiber surface. 4. Discussion The oxidation mechanism of SiC is very complex, which may be categorized the passive and active oxidation regimes. The regime of oxidative reaction is quite dependent on the actual chemical composition of fibers, oxygen partial pressure and exposure temperature as well as other factors [5,13–15]. The thermo-chemical correlation and oxidation dynamics for active and passive oxidation of silicon carbide have been investigated experimentally and theoretically in the literatures [5,13–15]. From which we could know passive oxidation would lead to the formation of SiO2 film protecting the SiC materials from further oxidation attack, in the case of active oxidation, the SiO2 formation rate is too low to seal the surface of materials, resulting in the mass loss and pits formation. The oxygen partial pressure for the transition from passive oxidation to active oxidation at 1300 C was pO2 ¼ 101 2 Pa [5]. The nominal pO2 of testing environments in this study was about 2 104 Pa (air), 0.2 Pa (HP-Ar) and 105 Pa (UHP-Ar), indicating that fiber would be oxidized in passive and active oxidation regime is quite possible. And also, with increasing the exposure temperature, the critical value of pO2 for the transition from passive to active oxidation would be increased. The selection of the test condition in this work would be helpful in understanding the oxidation behavior of SiC fibers. In Fig. 1, the mass gain for fibers under annealing in air is due to the passive oxidation of fibers with atmospheric oxygen. However, the mass gain is not smoothly increased with increasing the temperature. For fiber under annealing in air at 1400 C, it was oxidized mildly. The fiber under annealing in air at 1500 C was again subjected to a high degree of oxidation. This result is consistent with the XRD pattern showed in Fig. 2: (i) the cristobalite peak at 2h = 22 for fiber under annealing in air indicated that fiber was oxidized mainly in passive oxidation regime; (ii) a wide cristobalite peak (2h = 22) with relatively low intensity for fibers under annealing at 1400 C was observed in comparison to that of fibers under annealing at 1300 and 1500 C, showing the passive oxidation at 1400 C was refrained. This is due to the quick formation 0 0.5 1 1.5 2 2.5 3 0 0.2 0.4 0.6 0.8 As-received Air HP-Ar Tensile strength (GPa) Strain (%) 1 Fig. 3. Tensile test curves for Hi-Nicalon fibers under annealing in different atmospheres at 1500 C for 1 h. 0 0.5 1 1.5 2 2.5 3 3.5 4 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Tensile strength Stress relaxation parameter Tensile strength (GPa) Stress relaxation parameter,m As-received Air HP-Ar UHP-Ar Annealing atmosphere Fig. 4. The tensile strength retention and stress relaxation parameter (m) of HiNicalon fibers under varied oxygen partial pressures, tensile strengths were measured on fibers under annealing at 1500 C; stress relaxation parameters were examined at 1300 C. 3134 J.J. Sha et al. / Corrosion Science 50 (2008) 3132–3138
J Sha er al. / Corrosion Science 50(2008)3132-3138 3135 1400°C d 1500°C 150C 1500°C Fig. 5. Morphologies of Hi-Nicalon fibers under annealing(a-c)and creep(d-e)in air at elevated temperatures for 1 h. of silica film in the initial stage at 1400C. protecting the fiber from obtained from the different conditions(Fig. 2), it is clear that no further oxidative reaction cristobalite peaks were observed for fibers under annealing in inert The observation of surface and cross-section morphologies in atmosphere. However, the B-Sic peak intensity increased with Fig 5 also confirmed that fibers were mainly oxidized in passive decreasing the oxygen partial pressure, indicating the further crys- oxidation regime characterized by the formation of silica film tallization and grain coarsening occurred during annealing. From when fibers exposed in air at high temperatures, but the passive this primary evidence, we thought that fibers were mainly oxidized oxidation was enhanced under Bsr creep condition due to in- in active oxidation regime. creased oxygen permeation Cracks in the silica layer observed in In inert atmosphere, the passive to active transition is a key Fig 5 resulted from the different coefficient of thermal expansion point on the change of microstructure and properties of Sic mate- in the fiber core and the silica layer, which put the silica layer in rials. The oxygen partial pressure for the transition from passive tension state on cooling. On the other hand, beta SiOz transforms oxidation to active oxidation was Po,=10 N2 Pa at 1300C into alpha Sioz at 300-370C with an accompanying volume [5]. The po, value in inert atmosphere in this study was about change [16), can also generate the stress in silica layer resulting 0.2 Pa(HP-Ar)and 10-5Pa(UHP-Ar) Hence, these oxygen partial in the formation of cracks. The formed SiO2 layer not only can re- pressures fall into the range of 10-N2 Pa, indicating that the ac- frain the further oxidation of SiC, but also can suppress the decom- tive oxidation could occur in the present condit osition of amorphous phase by restricting the outward diffusion From the observations of surface morphologies in Figs. 6 and 7 of Sio and CO gases produced through the following reaction [17: we can see the active oxidation initialized at different tempera SiCxO- SiC(s)+C(s)+Sio(g)+ Co(g tures in different conditions. As observed in Fig. 6a-C, active oxida tion characterized by the formation of bubbles and pits for fibers Additionally, for Hi-Nicalon fibers under creep at temperature der annealing in HP-Ar seems starting at 1400C, while the ac above 1300C, the local decomposition of SicxOy phase(oxygen: tive oxidation was enhanced for fibers under annealing in UHP-Al 0.5 wt%) and oxidation of free carbon(C: 0.38) were enhanced as shown in Fig. 7a-C. Meanwhile, the temperatures for active oxi- and introduced the defects into the silica layer, which might be dation of fibers under creep condition, was shift to low value as the reason for resulting in the formation of patterns as shown in shown in Fig. 6d-e and Fig. 7d-e, indicating the transition from passive oxidation to active oxidation was accelerated under creep nder annealing in inert atmosphere(HP-Ar and UHP- condition due to the stress applied by bsr test, the oxygen attack Ar) tures above 1300C, mass loss was observed( fig. 1). to the surface of fiber would become easier leading to more bub- if we make a comparison among the Xrd patterns bles and pits was formed. Especially, in the tensile side, the oxygen
of silica film in the initial stage at 1400 C, protecting the fiber from further oxidative reaction. The observation of surface and cross-section morphologies in Fig. 5 also confirmed that fibers were mainly oxidized in passive oxidation regime characterized by the formation of silica film when fibers exposed in air at high temperatures, but the passive oxidation was enhanced under BSR creep condition due to increased oxygen permeation. Cracks in the silica layer observed in Fig. 5 resulted from the different coefficient of thermal expansion in the fiber core and the silica layer, which put the silica layer in tension state on cooling. On the other hand, beta SiO2 transforms into alpha SiO2 at 300–370 C with an accompanying volume change [16], can also generate the stress in silica layer resulting in the formation of cracks. The formed SiO2 layer not only can refrain the further oxidation of SiC, but also can suppress the decomposition of amorphous phase by restricting the outward diffusion of SiO and CO gases produced through the following reaction [17]: SiCxOy ! SiCðsÞ þ CðsÞ þ SiOðgÞ þ CoðgÞ ð1Þ Additionally, for Hi-Nicalon fibers under creep at temperatures above 1300 C, the local decomposition of SiCxOy phase (oxygen: 0.5 wt%) and oxidation of free carbon (C: 0.38) were enhanced and introduced the defects into the silica layer, which might be the reason for resulting in the formation of patterns as shown in Fig. 5. For fibers under annealing in inert atmosphere (HP-Ar and UHPAr) at temperatures above 1300 C, mass loss was observed (Fig. 1). Furthermore, if we make a comparison among the XRD patterns obtained from the different conditions (Fig. 2), it is clear that no cristobalite peaks were observed for fibers under annealing in inert atmosphere. However, the b-SiC peak intensity increased with decreasing the oxygen partial pressure, indicating the further crystallization and grain coarsening occurred during annealing. From this primary evidence, we thought that fibers were mainly oxidized in active oxidation regime. In inert atmosphere, the passive to active transition is a key point on the change of microstructure and properties of SiC materials. The oxygen partial pressure for the transition from passive oxidation to active oxidation was pO2 ¼ 101 2 Pa at 1300 C [5]. The pO2 value in inert atmosphere in this study was about 0.2 Pa (HP-Ar) and 105 Pa (UHP-Ar). Hence, these oxygen partial pressures fall into the range of 101 2 Pa, indicating that the active oxidation could occur in the present condition. From the observations of surface morphologies in Figs. 6 and 7, we can see the active oxidation initialized at different temperatures in different conditions. As observed in Fig. 6a–c, active oxidation characterized by the formation of bubbles and pits for fibers under annealing in HP-Ar seems starting at 1400 C, while the active oxidation was enhanced for fibers under annealing in UHP-Ar as shown in Fig. 7a–c. Meanwhile, the temperatures for active oxidation of fibers under creep condition, was shift to low value as shown in Fig. 6d–e and Fig. 7d–e, indicating the transition from passive oxidation to active oxidation was accelerated under creep condition. Due to the stress applied by BSR test, the oxygen attack to the surface of fiber would become easier leading to more bubbles and pits was formed. Especially, in the tensile side, the oxygen Fig. 5. Morphologies of Hi-Nicalon fibers under annealing (a–c) and creep (d–e) in air at elevated temperatures for 1 h. J.J. Sha et al. / Corrosion Science 50 (2008) 3132–3138 3135
J Sha et al/ Corrosion Science 50(2008)3132-3138 l300°C a1300°C 1400°C b1400"C Bubble 1500CF c1500 f Fig. 6. Morphologies of Hi-Nicalon fibers under annealing (a-c)and creep(d-f) in HP-Ar at elevated temperatures for 1 h attack rate should be higher than other places of fiber, namely ap- Based on the above analysis, the strength degradation of fibers plied stress enhanced the oxidation. different oxygen partial pressures could be attributed to differ L Concerning the huge grain deposited on the skin of the fibers, ent microstructures caused by different oxidative mechanism. lich might be produced by the quick decomposition of amor- The strength degradation of fiber under annealing in air is not so phous phase at high temperatures through reaction (1)and the much comparing to that of fibers under annealing in inert atmo- eactions between Sio and carbon/ carbon oxide by chemical gas sphere as shown in Fig. 4. This is due to the sio2 layer stopping The grain coarsening can also be enhanced by ap- the further oxygen attack and suppressing the decomposition of plied stress due to the increased diffusivity of C/Si at grain bound- amorphous phase(Fig. 5). The main reason for strength degrada aries as shown in Figs. 6 and 7. tion of Hi-Nicalon fibers in inert atmosphere is due to the damaged Noteworthy is that the active oxidation is infinite slow if t structure from active oxidation and decomposition of amorphous oxygen partial pressure is very low [18. But in present work, the as well as grain growth as observed in Figs. 6 and 7 oxygen content in inert atmosphere seems sufficient to cause obvi- The creep resistance of Sic fibers slightly increased with us active oxidation. This means the oxygen content remained in decreasing the oxygen partial pressure as shown in Fig 4, which furnace chamber should be higher than nominal value. This is quite is likely controlled by the grain coarsening and oxygen content of possible in case. Because the furnace wall might absorb water va- fiber. Firstly, by reaction (1), the decomposition of amorphous por at room temperature, thus, the release of water vapor at the phase in low oxygen partial pressure environments is easier than eginning of test would increase the oxygen content. It is very that in high oxygen partial pressure environments. In this case, ficult to monitor the humidity of furnace chamber at high temper- due to the coalescence of B Sic grain [20 the grain coarsening gen will be reduced by flow inert gas. During this process, the pas- decreased creep rate [21]. On the other hand, under the low a B atures. At that time, if the oxygen partial pressure is higher than would occur. The grain coarsening increased with decreasing ox the critical transition value, a thin silica film could be formed on gen partial pressure has been observed in other studies [3. 13].Ar the fiber surface With times increase, the content of residual oxy- also, the concurrent grain growth in tensile creep has resulted in sive and active oxidation can proceed concurrently. The formation partial pressure atmosphere, the easier decomposition of amor- f bubbles in the silica scale can provide some indication for this phous phase distributed at grain boundaries should result in a sta- process [ 19]. When the oxygen partial pressure in furnace chamber ble grain boundary structure. Therefore, the somewhat high creep decreased to critical transition value, then the silica film disappear resistance in low oxygen partial pressures could be attributed to gradually and oxidation of Sic fiber was mainly proceeded in active the concurrent grain growth and the reduction of oxygen content oxidation regime. of fiber during BSR test as well as other factors
attack rate should be higher than other places of fiber, namely applied stress enhanced the oxidation. Concerning the huge grain deposited on the skin of the fibers, which might be produced by the quick decomposition of amorphous phase at high temperatures through reaction (1) and the reactions between SiO and carbon/carbon oxide by chemical gas phase reaction. The grain coarsening can also be enhanced by applied stress due to the increased diffusivity of C/Si at grain boundaries as shown in Figs. 6 and 7. Noteworthy is that the active oxidation is infinite slow if the oxygen partial pressure is very low [18]. But in present work, the oxygen content in inert atmosphere seems sufficient to cause obvious active oxidation. This means the oxygen content remained in furnace chamber should be higher than nominal value. This is quite possible in case. Because the furnace wall might absorb water vapor at room temperature, thus, the release of water vapor at the beginning of test would increase the oxygen content. It is very dif- ficult to monitor the humidity of furnace chamber at high temperatures. At that time, if the oxygen partial pressure is higher than the critical transition value, a thin silica film could be formed on the fiber surface. With times increase, the content of residual oxygen will be reduced by flow inert gas. During this process, the passive and active oxidation can proceed concurrently. The formation of bubbles in the silica scale can provide some indication for this process [19]. When the oxygen partial pressure in furnace chamber decreased to critical transition value, then the silica film disappear gradually and oxidation of SiC fiber was mainly proceeded in active oxidation regime. Based on the above analysis, the strength degradation of fibers in different oxygen partial pressures could be attributed to different microstructures caused by different oxidative mechanism. The strength degradation of fiber under annealing in air is not so much comparing to that of fibers under annealing in inert atmosphere as shown in Fig. 4. This is due to the SiO2 layer stopping the further oxygen attack and suppressing the decomposition of amorphous phase (Fig. 5). The main reason for strength degradation of Hi-Nicalon fibers in inert atmosphere is due to the damaged structure from active oxidation and decomposition of amorphous as well as grain growth as observed in Figs. 6 and 7. The creep resistance of SiC fibers slightly increased with decreasing the oxygen partial pressure as shown in Fig. 4, which is likely controlled by the grain coarsening and oxygen content of fiber. Firstly, by reaction (1), the decomposition of amorphous phase in low oxygen partial pressure environments is easier than that in high oxygen partial pressure environments. In this case, due to the coalescence of b SiC grain [20], the grain coarsening would occur. The grain coarsening increased with decreasing oxygen partial pressure has been observed in other studies [3,13]. And also, the concurrent grain growth in tensile creep has resulted in a decreased creep rate [21]. On the other hand, under the low oxygen partial pressure atmosphere, the easier decomposition of amorphous phase distributed at grain boundaries should result in a stable grain boundary structure. Therefore, the somewhat high creep resistance in low oxygen partial pressures could be attributed to the concurrent grain growth and the reduction of oxygen content of fiber during BSR test as well as other factors. Fig. 6. Morphologies of Hi-Nicalon fibers under annealing (a–c) and creep (d–f) in HP-Ar at elevated temperatures for 1 h. 3136 J.J. Sha et al. / Corrosion Science 50 (2008) 3132–3138
Sha et al. /Corrosion Science 50(2008)3132-3138 1300°C a13°C d 1400°C b1400C 1500°C c1500°C Fig. 7. Morphologies of Hi-Nicalon fibers under annealing(a-c)and creep(d-f) in UHP-Ar at elevated temperatures for 1 h. 5. Conclusion [2] T. Shimon, H. Takeuchi, K Okamura, Thermal stability of polycarbosilane- silicon carbide fibers under reduced pressures. ]. Am. Ceram Soc. 8 Thermal and mechanical stabilities were investigated on H- 3/G.4)566-570 T. Shibayama, H. Takahashi, Microstructural evolution of Hi-Nicalon" Nicalon fibers under annealing and creep in various oxygen partial pressures by mass cnange, mecnanical properties as well as mIcro- [) T. Shimo, K okamura, M. Ito, M. Takeda, High-temperature stability of low- de fber heat-treated under different atmos (1)In the case of mechanical stability, the tensile strength of 51 B. Schneider, A Guette, R. Naslain, M Catal, A Costecalde A theoretical and fibers was strongly dependent on the oxygen partial pres sure of testing environment, but a weak dependence of BSR creep resistance on oxygen partial pressure was [6NS. Jacobson, Corrosion of stlicon-based ceramics in combustion observed [71J-] Sha, JS. Park, T. Hinoki, A Kohyama, J. (2)Mass change, SEM observations and XRD examinations indi of increasing the exposure temperature and decreasing the nanges were accelerated by stress applied by BSR test, indi [10] J Sha, T Nozawa, I.S. Park, Y Katoh, A Kohyama, Effect of heat tr cating the thermal stability of fber was related to not only Nuc Mate the two factors mentioned above but also the mechanical [11] G.N. Morscher, J-A Dicarlo, A simple test for thermomechanical eva state of fibers. The different oxidation regimes were respon- [12] J]. Sha, Performance of Sic-based fibers under severe environments and its sible for the strength degradation and microstructure change of fibers [13] T Shimao, K Okamura, Y polycarbosilane-derived silicon carbide fibers heated in Ar-o gas mixtures, J. ater.sci.37(2002)1793-1800 [14] T Shimo, Y. Morisada, K Okamura, Oxidation behavior of Si-M-C-O fibers References nder wide range of oxygen partial pressures. ] Mater Sci 37(2002)4361 [1] M. Takeda, A. Urano, J. Sakamoto, Y. Imai, Microstructure and oxidative [15]KG. Nickel, Corrosion of non-oxide ceramics, Ceram. Int. 23(1997)127-133. segradation behavior of silicon carbide fiber Hi-Nicalon type S. J Nucl. Mater. [16] T. Shimoo, F. Toyoda, K Okamura, Thermal stability of 3-363(1998)1594-1599 bjected to selected oxidation treatment, J.Am. Ceram. Soc..83(2000)1450-1456
5. Conclusion Thermal and mechanical stabilities were investigated on HiNicalon fibers under annealing and creep in various oxygen partial pressures by mass change, mechanical properties as well as microstructural features. The main results are summarized as follows: (1) In the case of mechanical stability, the tensile strength of fibers was strongly dependent on the oxygen partial pressure of testing environment, but a weak dependence of BSR creep resistance on oxygen partial pressure was observed. (2) Mass change, SEM observations and XRD examinations indicated the thermal stability of fiber was dependent on the exposure temperature and oxygen content of environments, namely, the degree of oxidation of fiber increased with increasing the exposure temperature and decreasing the oxygen partial pressure. And also, the microstructure changes were accelerated by stress applied by BSR test, indicating the thermal stability of fiber was related to not only the two factors mentioned above, but also the mechanical state of fibers. The different oxidation regimes were responsible for the strength degradation and microstructure change of fibers. References [1] M. Takeda, A. Urano, J. Sakamoto, Y. Imai, Microstructure and oxidative degradation behavior of silicon carbide fiber Hi-Nicalon type S, J. Nucl. Mater. 258–363 (1998) 1594–1599. [2] T. Shimoo, H. Takeuchi, K. Okamura, Thermal stability of polycarbosilanederived silicon carbide fibers under reduced pressures, J. Am. Ceram. Soc. 84 (2001) 566–570. [3] G. He, T. Shibayama, H. Takahashi, Microstructural evolution of Hi-NicalonTM SiC fibers annealed and crept in various oxygen partial pressure atmospheres, J. Mater. Sci. 35 (2000) 1153–1164. [4] T. Shimoo, K. Okamura, M. Ito, M. Takeda, High-temperature stability of lowoxygen silicon carbide fiber heat-treated under different atmosphere, J. Mater. Sci. 35 (2000) 3733–3739. [5] B. Schneider, A. Guette, R. Naslain, M. Cataldi, A. Costecalde, A theoretical and experimental approach to the active-to-passive transition in the oxidation of silicon carbide: experiments at high temperatures and low total pressures, J. Mater. Sci. 33 (1998) 535–547. [6] N.S. Jacobson, Corrosion of Silicon-based ceramics in combustion environments, J. Am. Ceram. Soc. 76 (1993) 3–28. [7] J.J. Sha, J.S. Park, T. Hinoki, A. Kohyama, J. Yu, Tensile properties and creep behavior of SiC-based fibers under various oxygen partial pressures, Mater. Sci. Forum 475–479 (2005) 1333–1336. [8] J.J. Sha, J.S. Park, T. Hinoki, A. Kohyama, Strength and fracture properties of advanced SiC-based fibers, Mech. Compos. Mater. 42 (2006) 527–534. [9] G.E. Yougnblood, C. Lewinsohn, R.H. Jones, A. Kohyama, Tensile strength and fracture surface characterization of Hi-NicalonTM SiC fibers, J. Nucl. Mater. 289 (2001) 1–9. [10] J.J. Sha, T. Nozawa, J.S. Park, Y. Katoh, A. Kohyama, Effect of heat treatment on the tensile strength and creep resistance of advanced SiC fibers, J. Nucl. Mater. 329–333 (2004) 592–596. [11] G.N. Morscher, J.A. Dicarlo, A simple test for thermomechanical evaluation of ceramic fibers, J. Am. Ceram. Soc. 75 (1992) 136–140. [12] J.J. Sha, Performance of SiC-based fibers under severe environments and its mechanistic analysis, Doctoral Thesis, Kyoto Universiy, Kyoto, Japan, 2005. [13] T. Shimoo, K. Okamura, Y. Morisada, Active-to-passive oxidation transition for polycarbosilane-derived silicon carbide fibers heated in Ar–O2 gas mixtures, J. Mater. Sci. 37 (2002) 1793–1800. [14] T. Shimoo, Y. Morisada, K. Okamura, Oxidation behavior of Si–M–C–O fibers under wide range of oxygen partial pressures, J. Mater. Sci. 37 (2002) 4361– 4368. [15] K.G. Nickel, Corrosion of non-oxide ceramics, Ceram. Int. 23 (1997) 127–133. [16] T. Shimoo, F. Toyoda, K. Okamura, Thermal stability of low-oxygen silicon carbide fiber (Hi-Nicalon) subjected to selected oxidation treatment, J. Am. Ceram. Soc. 83 (2000) 1450–1456. Fig. 7. Morphologies of Hi-Nicalon fibers under annealing (a–c) and creep (d–f) in UHP-Ar at elevated temperatures for 1 h. J.J. Sha et al. / Corrosion Science 50 (2008) 3132–3138 3137
J Sha et al/ Corrosion Science 50(2008)3132-3138 [17] T Shimon, K Okamura, L Tst Seguchi, Thermal stability of low-oxygen [20] M.H. Jaskowiak, A Dicarlo, Pressure effects on the thermal-stability of silicon- Sic fibers under different conditions, J Mater. Sci. 34 (1999)5623-5631 carbide fibers. J Am Ceram Soc. 72(1989)192-15 [18] T. Shimon, H. Takeuchi, K. Okamura, Thermal stability of polycarbosilane- [21JA Dicarlo, H M Yun, J.B. Hurst, Fracture mechanisms for SiC fibers and sic/sic ved silicon carbide fibers under reduced pressures, Am Ceram Soc. 84 under stress-rupture conditions at high temperatures, AppL. Math. Comput.152(2004)473-481 [19] KL Luthra, Some new perspectives on oxidation of silicon carbide& silicon nitride. J Am Ceram Soc. 74 (1991)1095-1103
[17] T. Shimoo, K. Okamura, I. Tsukada, T. Seguchi, Thermal stability of low-oxygen Sic fibers under different conditions, J. Mater. Sci. 34 (1999) 5623–5631. [18] T. Shimoo, H. Takeuchi, K. Okamura, Thermal stability of polycarbosilanederived silicon carbide fibers under reduced pressures, J. Am. Ceram. Soc. 84 (2001) 566–570. [19] K.L. Luthra, Some new perspectives on oxidation of silicon carbide & silicon nitride, J. Am. Ceram. Soc. 74 (1991) 1095–1103. [20] M.H. Jaskowiak, J.A. Dicarlo, Pressure effects on the thermal-stability of siliconcarbide fibers, J. Am. Ceram. Soc. 72 (1989) 192–197. [21] J.A. Dicarlo, H.M. Yun, J.B. Hurst, Fracture mechanisms for SiC fibers and SiC/SiC composites under stress-rupture conditions at high temperatures, Appl. Math. Comput. 152 (2004) 473–481. 3138 J.J. Sha et al. / Corrosion Science 50 (2008) 3132–3138