J Mater Sci(2007)42:5046-5056 5047 of SiC fibers with several techniques, and revealed 14 um), Hi-Nicalon Type-S(HNLS: C/Si= 1.05, some other original features: The carbon-rich layer on oxygen=0.2 wt %, diameter: 12 um) and Tyranno the surface of the HNLS fiber(80 nm) is much thicker SA(TySA(Grade 3): C/Si= 1.05, oxygen 0.5 wt%, than that of HNL fiber (20 um); Tyranno-SA fiber alumina l wt%, diameter: 7 um). It is clear that HNL (about 10 um in diameter) has a carbon-rich core fiber contained excess carbon and low oxygen content indicating that near-stoichiometric composition is only The latter two have near-stoichiometric composition effective near edge region. Similar phenomenon on and high crystallinity. These fibers were put in a Tyranno-SA fiber was also observed by Colomban graphite crucible and then annealed in Ar under a et al. using the Raman Spectroscopy [9, 10]. Bunsell pressure of 10 Pa and held for 1 h at desired tempe et al. 11 also reported that both Hi-Nicalon Type s ature from 1, 300 to 1, 900C. The annealing condition and Tyranno-SA fiber contain excess carbon at triple has been described in more detail elsewhere [12] points of grain boundaries. The microstructure of Sic materials in high tempe rature an nd oxidative environ- Characterization ment is very sensitive to composition. Consequently, it is necessary to characterize the microstructure of Sic After annealing, several techniques were used to fibers at elevated temperatures in order to know the characterize the fiber microstructure features and degradation mechanism and to predict high tempera- fracture properties. The XRD with CuKx irradiation ture performance of CMCs. was applied to examine the present phase, and the In our former work [12], the tensile properties of the apparent crystallite size of B-Sic was estimated by annealed Sic fibers were investigated by tensile test, employing the Scherrer formula[13] and fundamental analysi on the microstructure of these fibers was performed by means of field-emission L=K. i/(D.cos 0) scanning electron microscopy(FE-SEM) and X-ray diffraction(XRD), but no attempts were made to where K is a constant(taken as 0.9),i the Cuko correlate the microstructure and the fractur length (i.e, 1=0.154056 D the half-value ties.As observed in this work [12], the fracture of the width of p-Sic (111) peak and e the Bragg angle Hi-NicalonM and the Hi-Nicalon M Type S fibers (0=17.50 for B-Sic (111)) mainly originated from critical flaw and showed a clear FE-SEM image analyses were carried out within a fracture mirror zone. Therefore, for practical applica- fiber tow and along the fiber length to check the tion of CMCs, it is requested to accumulate experi- diameter variation. Result showed that these fibers had mental data and to reveal the degradation mechanism a wide diameter variation within a tow. The diameter of Sic fibers with a consideration of the thermal- variations within a tow are 10.78-16.60 um for the chemical stability HNL fiber. 10.85-14 04 um for the hnls fiber. 520- In order to identify the factors, which affect the high 9.67 um for the TySA fiber, respectively. Based on this temperature performance of SiC fibers, this work result, the individual fiber diameter was measured for proceeded a complementary investigation on the precise strength calculation. FE-SEM was also used to microstructure features and fracture properties of Sic characterize the fibers surface morphologies and fibers annealed at elevated temperatures, and fracture surface. attempted to clarify the correlation between the Furthermore, the critical flaw size and fracture mechanical properties and the microstructure. The mirror size on the fracture surfaces of fiber fragments fracture toughness and the critical fracture energy at were measured by means of FE-SEM examination elevated temperatures were estimated using the frac- The fracture fragments were obtained by single fila- ture mechanics by measurement of the critical flaw ment tensile tests, which were performed at room Irror sIZ temperature in ambient atmosphere using a mechani cal testing apparatus (Instron Corp. Model 5581 according to ASTM-recommended procedures [14] The load was applied at a constant strain rate of Experimental procedure 2 x 10/s and measured by a load-cell of 2.5N. The individual filaments had a gauge length of 25.4 mm and Materials and annealing condition were aligned and glued on cardboard fixture with epoxy. For each fiber type, the total number of tests is The fibers examined in this study were Hi-NicalonM 220 The detailed testing procedure has been presented (HNL: C/Si= 1.39, oxygen=0.5 wt%, diameter: in the literatures [12, 15of SiC fibers with several techniques, and revealed some other original features: The carbon-rich layer on the surface of the HNLS fiber (80 nm) is much thicker than that of HNL fiber (20 um); Tyranno-SA fiber (about 10 um in diameter) has a carbon-rich core indicating that near-stoichiometric composition is only effective near edge region. Similar phenomenon on Tyranno-SA fiber was also observed by Colomban et al. using the Raman Spectroscopy [9, 10]. Bunsell et al. [11] also reported that both Hi-Nicalon Type S and Tyranno-SA fiber contain excess carbon at triple points of grain boundaries. The microstructure of SiC materials in high temperature and oxidative environ￾ment is very sensitive to composition. Consequently, it is necessary to characterize the microstructure of SiC fibers at elevated temperatures in order to know the degradation mechanism and to predict high tempera￾ture performance of CMCs. In our former work [12], the tensile properties of the annealed SiC fibers were investigated by tensile test, and fundamental analysis on the microstructure of these fibers was performed by means of field-emission scanning electron microscopy (FE-SEM) and X-ray diffraction (XRD), but no attempts were made to correlate the microstructure and the fracture proper￾ties. As observed in this work [12], the fracture of the Hi-NicalonTM and the Hi-NicalonTM Type S fibers mainly originated from critical flaw and showed a clear fracture mirror zone. Therefore, for practical applica￾tion of CMCs, it is requested to accumulate experi￾mental data and to reveal the degradation mechanism of SiC fibers with a consideration of the thermal￾chemical stability. In order to identify the factors, which affect the high temperature performance of SiC fibers, this work proceeded a complementary investigation on the microstructure features and fracture properties of SiC fibers annealed at elevated temperatures, and attempted to clarify the correlation between the mechanical properties and the microstructure. The fracture toughness and the critical fracture energy at elevated temperatures were estimated using the frac￾ture mechanics by measurement of the critical flaw size/mirror size. Experimental procedure Materials and annealing condition The fibers examined in this study were Hi-NicalonTM (HNL: C/Si = 1.39, oxygen = 0.5 wt%, diameter: 14 um), Hi-NicalonTM Type-S (HNLS: C/Si = 1.05, oxygen = 0.2 wt%, diameter: 12 um) and TyrannoTM SA (TySA (Grade 3): C/Si = 1.05, oxygen < 0.5 wt%, alumina < 1 wt%, diameter: 7 um). It is clear that HNL fiber contained excess carbon and low oxygen content. The latter two have near-stoichiometric composition and high crystallinity. These fibers were put in a graphite crucible and then annealed in Ar under a pressure of 105 Pa and held for 1 h at desired temper￾ature from 1,300 to 1,900 C. The annealing condition has been described in more detail elsewhere [12]. Characterization After annealing, several techniques were used to characterize the fiber microstructure features and fracture properties. The XRD with CuKa irradiation was applied to examine the present phase, and the apparent crystallite size of b-SiC was estimated by employing the Scherrer formula [13]: L = K k=ðD cos hÞ ð1Þ where K is a constant (taken as 0.9), k the CuKa wavelength (i.e., k = 0.154056 nm), D the half-value width of b-SiC (111) peak and h the Bragg angle (h = 17.5 for b-SiC (111)). FE-SEM image analyses were carried out within a fiber tow and along the fiber length to check the diameter variation. Result showed that these fibers had a wide diameter variation within a tow. The diameter variations within a tow are 10.78–16.60 um for the HNL fiber, 10.85–14.04 um for the HNLS fiber, 5.20– 9.67 um for the TySA fiber, respectively. Based on this result, the individual fiber diameter was measured for precise strength calculation. FE-SEM was also used to characterize the fiber’s surface morphologies and fracture surface. Furthermore, the critical flaw size and fracture mirror size on the fracture surfaces of fiber fragments were measured by means of FE-SEM examination. The fracture fragments were obtained by single fila￾ment tensile tests, which were performed at room temperature in ambient atmosphere using a mechani￾cal testing apparatus (Instron Corp. Model 5581) according to ASTM-recommended procedures [14]. The load was applied at a constant strain rate of 2 · 10–4/s and measured by a load-cell of 2.5 N. The individual filaments had a gauge length of 25.4 mm and were aligned and glued on cardboard fixture with epoxy. For each fiber type, the total number of tests is ‡20. The detailed testing procedure has been presented in the literatures [12, 15]. J Mater Sci (2007) 42:5046–5056 5047 123