R H. Jones, C.H. Henager: /Journal of the European Ceramic Sociery 25(2005)1717-1722 10 h VS/OE Argon +202 Pa O2 Deflection m Argon Fraction 0.001 Time(s) Domination Fig. 2. Displacement-time curves for single-edge 100011001200130014001500 SiC/SiC Hi-Nicalon composites tested at 1373K in ding for the indicated time In Ar the curve follows the expected p laws while in Ar+ Oz the curve becomes linear in time as given by Eq ( funct Map showing regions of dominant crack growth mechanism of oxygen concentration in the environment and temperature for 2.2.3. Fiber/matrix interphase removal(R) by oxidative SiC/SiC composites When the fiber/matrix interphase is a material that can 2.3. Crack growth mechanism map undergo a gaseous reaction with oxygen then interphase re moval in oxygen-containing environments becomes impor The proposed mechanism map is shown in Fig 3 Fiber tant, as shown by the data in Fig. 2. Previous research has creep relaxation(FR)dominates at high temperatures and the typical 0/90 plain weave CVI-SiC/SiCr materials with a tion energies are high( 600 kJ/mol)and oxidat. p activa- shown that the following interphase recession data applies tion energies are low (50 kJ/mol). Crack extension is dom- lox =3.3 x 10-4e-6014T p08891=R(T. Po,)t thus inated by interface removal (IR) for oxygen concentrations less than -o I when the temperatures are not extreme. We △lox=R(T,Po2)△t (5) observe both kinetic and activation energy changes in ou experimental data consistent with this hypothesis. We define where lox is the recession distance along the fiber/matrix in- the transition from Fr to iR by the locus of points where terphase from the crack face into the composite in m/s, T the crack growth from each mechanism is equal. At higher temperature in K, Po is fractional oxygen concentration, I oxygen concentrations, the fiber/matrix interphase oxidizes is exposure time and Ri is the interface recession rate. We assume that deb is maintained at its value determined by t and is replaced by a glass phase and either oxidation en brittlement(OE)or viscous sliding(VS)mechanisms may and the load on the fiber occur. We propose that interphase glass formation and chan- nel pinch-off (fiber-matrix bonding) at intermediate temper- 2.2. 4. Oxidation of the bridging fiber(OE and sOE) atures where the resultant glass phase is brittle results in O Exposed bridging fibers will undergo oxidation at elevated and the fibers fail since they can no longer slide relative to temperatures according to: 18 the matrix. We define the transition from ir to oe as the fb=1.84×10-7e-4016/Par0.5 (6) locus of points where the fiber/matrix channel pinches-off and oe occurs in a finite crack growth increment(2.5mm where fibox is the oxide thickness grown on the fiber in m, T in this case). At higher temperatures the glass phase is vis- is temperature in K, Po, is fractional oxygen concentration, cous, not brittle and Vs occurs. When the glass phase ha and t is exposure time in s a high viscosity, fibers can fail by rupture since the fiber/matrix bonding leads to stress concentrations- 2.2.5. Fiber stress rupture like replacing an interface with a low frictional sliding stress osed bridging fibe with a higher one. It follows that there will be a temperature rupture relationships obey time-dependent stress above which this stress concentration will not fail the fibers but this will depend on the details of the compositionally de- In(od=23E"R/[RT(n()+C) pendent glass viscosity. The transition from OE to VS/OE is schematic since we do not yet know the glass viscosity nor have we implemented a viscous sliding treatment for the of is fiber strength in MPa, and e, B, and C are fitted model. although such treatments exist. fiber oxidation also ters from single fiber rupture tests in results in loss of strength. which leads to embrittlement viaR.H. Jones, C.H. Henager Jr. / Journal of the European Ceramic Society 25 (2005) 1717–1722 1719 Fig. 2. Displacement-time curves for single-edge notched beam bars of SiC/SiC Hi-Nicalon composites tested at 1373 K in 4-point bending for the indicated time. In Ar the curve follows the expected fiber creep laws while in Ar + O2 the curve becomes linear in time as given by Eq. (5). 2.2.3. Fiber/matrix interphase removal (IR) by oxidative volatilization When the fiber/matrix interphase is a material that can undergo a gaseous reaction with oxygen then interphase re￾moval in oxygen-containing environments becomes impor￾tant, as shown by the data in Fig. 2. Previous research has shown that the following interphase recession data applies to the typical 0/90 plain weave CVI-SiC/SiCf materials with a CVI carbon interphase17. lox = 3.3 × 10−4e−6014/T P0.889 o2 t = Ri(T, Po2 )t thus lox = Ri(T, Po2 )t (5) where lox is the recession distance along the fiber/matrix in￾terphase from the crack face into the composite in m/s, T is temperature in K, Po2 is fractional oxygen concentration, t is exposure time and Ri is the interface recession rate. We assume that ldeb is maintained at its value determined by τ and the load on the fiber. 2.2.4. Oxidation of the bridging fiber (OE and VS/OE) Exposed bridging fibers will undergo oxidation at elevated temperatures according to:18 fibox = 1.84 × 10−7e−4016.8/T Po2 t 0.5 (6) where fibox is the oxide thickness grown on the fiber in m, T is temperature in K, Po2 is fractional oxygen concentration, and t is exposure time in s. 2.2.5. Fiber stress rupture (SR) Exposed bridging fibers will obey time-dependent stress rupture relationships as:19 ln(σf) = 2.3E −
β R [RT (ln(t) + C)] (7) where σf is fiber strength in MPa, and E, β, and C are fitted parameters from single fiber rupture tests in air. Fig. 3. Map showing regions of dominant crack growth mechanism as a function of oxygen concentration in the environment and temperature for SiC/SiC composites. 2.3. Crack growth mechanism map The proposed mechanism map is shown in Fig. 3. Fiber creep relaxation (FR) dominates at high temperatures and low oxygen concentrations, as expected, since creep activa￾tion energies are high (∼600 kJ/mol) and oxidation activa￾tion energies are low (∼50 kJ/mol). Crack extension is dom￾inated by interface removal (IR) for oxygen concentrations less than ∼0.1 when the temperatures are not extreme. We observe both kinetic and activation energy changes in our experimental data consistent with this hypothesis. We define the transition from FR to IR by the locus of points where the crack growth from each mechanism is equal. At higher oxygen concentrations, the fiber/matrix interphase oxidizes and is replaced by a glass phase and either oxidation em￾brittlement (OE) or viscous sliding (VS) mechanisms may occur. We propose that interphase glass formation and chan￾nel pinch-off (fiber–matrix bonding) at intermediate temper￾atures where the resultant glass phase is brittle results in OE and the fibers fail since they can no longer slide relative to the matrix. We define the transition from IR to OE as the locus of points where the fiber/matrix channel pinches-off and OE occurs in a finite crack growth increment (2.5 mm in this case). At higher temperatures the glass phase is vis￾cous, not brittle and VS occurs. When the glass phase has a high viscosity, fibers can fail by rupture since the local fiber/matrix bonding leads to stress concentrations20, rather like replacing an interface with a low frictional sliding stress with a higher one. It follows that there will be a temperature above which this stress concentration will not fail the fibers but this will depend on the details of the compositionally de￾pendent glass viscosity. The transition from OE to VS/OE is schematic since we do not yet know the glass viscosity nor have we implemented a viscous sliding treatment for the model, although such treatments exist21. Fiber oxidation also results in loss of strength, which leads to embrittlement via