E≈S Journal of the European Ceramic Society 20(2000)531-535 Mechanical characterisation of mullite-based ceramic matrix composites at test temperatures up to 1200c P.W.M.Peters.*,B. Daniels a, F. Clemens b, W.D. Vogel b DLR Institute of Materials Research, 51170 Koln Daimler Chrysler/ Dornier Fors lung, 88039 Friedrichshafen, Germany Accepted 10 August 1999 Abstract Nextel 610 fibre-reinforced mullite-based matrix fabricated by Dornier Forschung was characterised at DLR Institute of Mate- rials Research. The material was produced by the polymer route after coating the fibres with a 0 I um thick carbon layer. The composite was manufactured by infiltrating the fibres with a slurry containing a diluted polymer and mullite powder, curing in an autoclave and subsequently heat treating and pyrolysis of the polymer. A final heat treatment in air is performed to remove the carbon coating and to reduce the residual stresses A(0/90/0/ 90/0/90)s-laminate was produced with an average fibre volume fraction of 45.6% and a porosity of 15.9%. Dog-bone-type tensile specimens with a width of 10 mm were cut from the plate by water jet and tested at temperatures up to 1200C in air. The tensile strength at room temperature measured 177. 4 MPa and linearly decreased to 145.2 MPa at a temperature of 800C. A stronger decrease occurred at 1000 and 1200 C. In contradiction to ceramic matrix com- posites manufactured by the Cvl-route the stress-strain behaviour is nearly linear up to failure. The modulus of the composite(at room temperature 108.8 GPa)is analysed on the basis of the expected moduli of the fibres and the mullite matrix. It can be con- cluded that the contribution of the matrix to the modulus of the composite is low, caused by porosity and components other than mullite. The intralaminar shear strength at room temperature measured 36 MPa. This value reflecting shear transfer capability of fibre to matrix limits the amount of fibre pull-out. C 2000 Elsevier Science Ltd. All rights reserved. Keywords: Aluminosilicate fibres; Composites; Laminates: Mechanical properties; Mullite matrix 1. Introduction giving rise to a loss of the crack deflective capability, if interfacial reaction products increase fibre/matrix bond Ceramic matrix composites(CMCs) are under devel- ing. Application of protective coatings is a possibility to opment for high temperature applications as e. g in aero reach short and medium term oxidative resistance. 2.3 engines, rocket nozzles and re-entry heat shields. In Long term oxidation resistance for thousands of comparison with monolythic ceramics these fibre rein- hours, however, can only be reached if the material forced ceramics are attractive due to the increased system is composed of components each of which is toughness and decreased flaw sensitivity. First genera- oxidation resistant. For this reason in recent years tion CMCs were usually produced on the basis of che- CMCs with oxide fibres and oxide matrices have drawn mical vapour deposition (impregnation) making use of much attention. In the absence of weak layers sur- fibres surrounded by a carbon coating. These CMCs rounding the fibres a different crack deflection mechan- how on loading multiple matrix cracking which leads to ism must be activated as is for example possible on the a gradual loss of the contribution of the matrix to the basis of a different coefficient of thermal expansion of composite stiffness. This so-called quasi-ductility is fibre and matrix and a special geometrical arrangement reached due to the fact that matrix cracks do not pen of closely packed fibres in a finescale porous matrix. 4 trate the fibres but deflect at or in the C-layer. The main Crack deflection in the usual sense can be realized with draw back of these CMCs is, that in high temperature oxidative environments the carbon layer is consumed crack deflective materials(e. g. monazites 5.6) porous interphases 5,6 fugitive interphases 0955-2219/00/S. see front matter C 2000 Elsevier Science Ltd. All rights reserved PII:S0955-2219(99)00250
Mechanical characterisation of mullite-based ceramic matrix composites at test temperatures up to 1200C P.W.M. Peters a,*, B. Daniels a , F. Clemens b, W.D. Vogel b a DLR Institute of Materials Research, 51170 KoÈln, Germany bDaimlerChrysler / Dornier Forschung, 88039 Friedrichshafen, Germany Accepted 10 August 1999 Abstract Nextel 610 ®bre-reinforced mullite-based matrix fabricated by Dornier Forschung was characterised at DLR Institute of Materials Research. The material was produced by the polymer route after coating the ®bres with a 0.1 mm thick carbon layer. The composite was manufactured by in®ltrating the ®bres with a slurry containing a diluted polymer and mullite powder, curing in an autoclave and subsequently heat treating and pyrolysis of the polymer. A ®nal heat treatment in air is performed to remove the carbon coating and to reduce the residual stresses. A (0/90/0/90/0/90)s-laminate was produced with an average ®bre volume fraction of 45.6% and a porosity of 15.9%. Dog-bone-type tensile specimens with a width of 10 mm were cut from the plate by water jet and tested at temperatures up to 1200C in air. The tensile strength at room temperature measured 177.4 MPa and linearly decreased to 145.2 MPa at a temperature of 800C. A stronger decrease occurred at 1000 and 1200C. In contradiction to ceramic matrix composites manufactured by the CVI-route the stress±strain behaviour is nearly linear up to failure. The modulus of the composite (at room temperature 108.8 GPa) is analysed on the basis of the expected moduli of the ®bres and the mullite matrix. It can be concluded that the contribution of the matrix to the modulus of the composite is low, caused by porosity and components other than mullite. The intralaminar shear strength at room temperature measured 36 MPa. This value re¯ecting shear transfer capability of ®bre to matrix limits the amount of ®bre pull-out. # 2000 Elsevier Science Ltd. All rights reserved. Keywords: Aluminosilicate ®bres; Composites; Laminates; Mechanical properties; Mullite matrix 1. Introduction Ceramic matrix composites (CMCs) are under development for high temperature applications as e.g. in aero engines, rocket nozzles and re-entry heat shields. In comparison with monolythic ceramics these ®bre reinforced ceramics are attractive due to the increased toughness and decreased ¯aw sensitivity. First generation CMCs were usually produced on the basis of chemical vapour deposition (impregnation) making use of ®bres surrounded by a carbon coating. These CMCs show on loading multiple matrix cracking which leads to a gradual loss of the contribution of the matrix to the composite stiness.1 This so-called quasi-ductility is reached due to the fact that matrix cracks do not penetrate the ®bres but de¯ect at or in the C-layer. The main drawback of these CMCs is, that in high temperature oxidative environments the carbon layer is consumed giving rise to a loss of the crack de¯ective capability, if interfacial reaction products increase ®bre/matrix bonding. Application of protective coatings is a possibility to reach short and medium term oxidative resistance.2,3 Long term oxidation resistance for thousands of hours, however, can only be reached if the material system is composed of components each of which is oxidation resistant. For this reason in recent years CMCs with oxide ®bres and oxide matrices have drawn much attention. In the absence of weak layers surrounding the ®bres a dierent crack de¯ection mechanism must be activated as is for example possible on the basis of a dierent coecient of thermal expansion of ®bre and matrix and a special geometrical arrangement of closely packed ®bres in a ®nescale porous matrix.4 Crack de¯ection in the usual sense can be realized with . crack de¯ective materials (e.g. monazites 5,6) . porous interphases 5,6 . fugitive interphases. 0955-2219/00/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved. PII: S0955-2219(99)00250-2 Journal of the European Ceramic Society 20 (2000) 531±535 * Corresponding author
P.W.M. Peters et al. / Journal of the European Ceramic Society 20(2000)531-535 In the present work the mechanical behaviour of a the polymer, which is also cross-linked and pyrolyzed mullite-based CMC with oxide fibres(Nextel 610)and a One to three reinfiltration cycles show the best results fugitive interface is characterized. The material is pro- As a final process, the composite is annealed in air to duced by Dornier Forschung by the polymer route and reduce internal stresses and to remove the carbon coat tested at DLR ing on the fibres. The microstructure of the composite is shown in Fig. 2. The matrix is characterized by a homogeneous distribution of the mullite and the Si-O- 2. Materials and experiments C phase and a small grain size of the particles compared to the fibre dimensions The material investigated(designated Nextel 610/ A crossply laminate(0/90/0/90/0/ 90)s with, in total, 12 Umox)is produced by Dornier Forschung. The manu- layers was produced. The average thickness of the plate facturing technique is based on the polymer infiltration was 2.00 mm(ply thickness 0.167 mm), the fibre volume nd pyrolysis process. 7. It comprises three steps. First, fraction and the porosity measured Ve=0.456 and the fibre bundles are coated with a carbon layer of about 0. 1 um thickness. The fibres are pulled through a vessel containing a solution of a carbon precursor polymer Thus, a polymer film on each single fibre is produced and subsequently converted into a carbon layer by pyrolysis in a furnace The infiltration technique is similar to the production of fibre reinforced plastics. The fibres can be infiltrated by filament winding(as shown in Fig. 1), resin transfer molding or wet lamination with the slurry out of diluted silicon polymer and mullite powder. After infil tration, the polymer is cross-linked and densified in an autoclave to form a solid green body Subsequently the polymer is pyrolyzed in an oven with an inert atmosphere. Thus, a ceramic matrix is roduced that consists of mullite and a Si-O-C phase Fig. 2 structural details on polished surface of a Nextel 610/ To minimize porosity, the composite is reinfiltrated with Mox crossply laminate CMC Production by Infiltration and Pyrolysis of Polymers precursor ● solvent of prepregs filament winding ceramic powder filament winding by Resin Transfer Moulding (RTM under autoclave conditions out forming tools 200-300C.10-20bar 1100·1600C protecton integral structure from CMC Fig. I. Production of mullite/mullite CMCs by poylmer impregnation and pyrolysis
In the present work the mechanical behaviour of a mullite-based CMC with oxide ®bres (Nextel 610) and a fugitive interface is characterized. The material is produced by Dornier Forschung by the polymer route and tested at DLR. 2. Materials and experiments The material investigated (designated Nextel 610/ Umox) is produced by Dornier Forschung. The manufacturing technique is based on the polymer in®ltration and pyrolysis process.7,8 It comprises three steps. First, the ®bre bundles are coated with a carbon layer of about 0.1 mm thickness. The ®bres are pulled through a vessel containing a solution of a carbon precursor polymer. Thus, a polymer ®lm on each single ®bre is produced and subsequently converted into a carbon layer by pyrolysis in a furnace. The in®ltration technique is similar to the production of ®bre reinforced plastics. The ®bres can be in®ltrated by ®lament winding (as shown in Fig. 1), resin transfer molding or wet lamination with the slurry out of a diluted silicon polymer and mullite powder. After in®ltration, the polymer is cross-linked and densi®ed in an autoclave to form a solid green body. Subsequently the polymer is pyrolyzed in an oven with an inert atmosphere. Thus, a ceramic matrix is produced that consists of mullite and a Si±O±C phase. To minimize porosity, the composite is rein®ltrated with the polymer, which is also cross-linked and pyrolyzed. One to three rein®ltration cycles show the best results. As a ®nal process, the composite is annealed in air to reduce internal stresses and to remove the carbon coating on the ®bres. The microstructure of the composite is shown in Fig. 2. The matrix is characterized by a homogeneous distribution of the mullite and the Si±O± C phase and a small grain size of the particles compared to the ®bre dimensions. A crossply laminate (0/90/0/90/0/90)s with, in total, 12 layers was produced. The average thickness of the plate was 2.00 mm (ply thickness 0.167 mm), the ®bre volume fraction and the porosity measured Vf=0.456 and Fig. 1. Production of mullite/mullite CMCs by poylmer impregnation and pyrolysis.7 Fig. 2. Macrostructural details on polished surface of a Nextel 610/ Umox crossply laminate. 532 P.W.M. Peters et al. / Journal of the European Ceramic Society 20 (2000) 531±535
P.W.M. Peters et al. / Journal of the European Ceramic Society 20(2000)531-535 Vp=0.159, respectively. The density of the plate reached Parallel sided specimens with a width of 10 mm as 3.0015 g/cm3. Dog-bone type specimens were cut from indicated in Table 3 were also cut from the plates but he plate in 00-direction by waterjet. For a first test ser- now with the aid of a diamond wire machine. Specimens ies waisting of the specimens from a width of 15 mm in cut under an angle of 45 with respect of the 0-direc the grip range to a 10 mm wide gauge section was per- tion are used to measure the shear modulus g and the formed according to the dimensions in Fig 3 and Table intralaminar shear strength at room temperature. Fur- 1. As most of the specimens tested in the first test series ther 10 mm wide specimens were cut transverse to the broke near the radius the shape of the specimen was 0%-direction. After this these specimens were ground adapted to give a smoother change of width. The x-y with a diamond wheel on both surfaces in order to coordinates for the changed specimen shape is given in remove the surface layers so that finally specimens with Table 2 the orientation 900.2/02/900.2 remain. The stiffness of the The dog-bone specimens were loaded up to failure in crossply specimens with two different relative amounts a tensile test performed at different temperatures of transverse plies (Voe=50% and V9oe=16.6% between room temperature and 1200.C as a respectively) allows the determination of the moduli of In order to Itee a homogeneous temperature dis- the parallel and transverse ply (El, En tribution, a holding time of 10 min was applied after reaching the required temperature. Heating of the mid- dle section of the specimen is realized by inductive 3. Results and discussion heating of a susceptor. The temperature is measured with the aid of a thermocouple in the middle of the A calculation of the longitudinal and transverse ply specimen. The heating coil with a height of 100 mm and stifness En and Ei with the aid of the moduli of the consisting of 6 windings was adjusted in such a way, two types of crossply specimens delivers the values that the temperature in the 40 mm gauge length was as indicated in Table 4. The modulus parallel to the fibres homogeneous as possible. The maximum temperature can be approximated very well by the rule of mixtures deviation in the gauge length of 40 mm measured + 3% -o% at a temperature of 1200 C in the middle Stress- strain behaviour at all temperatures is measured with the aid of a clip gauge with a gauge length of 20 mm as schematically presented in Fig 4 Fig. 3. Dog-bone specimen dimensions for(0/90/0/90/0/90), tensile Table I Radius of dog-bone specimen(grip length 55 mm) Fig 4. Schematics of experimental set-up for high temperature tensile X0.0001.6122955456758307.78290009931 esting of Nextel 610/Mox Y0.0000.0270.0600.ll80.1790.3140.4320.545 X11.50712.55513.24513.75514.13114.38215000 Y0.8141.0681.2911.4921.6771.8272.50 Test specimens to characterize the mechanical behaviour of Nextel 610/Mox Table 2 Specimen type Test Radius of adapted dog-bone temperature propertie X0.0001612 5.8307.7829.00010 (0/90/0/90/0/90) 10 mm dog-bone RT-1200C E, UTS 0.1790.3140.4320.54 902/02/90 10 mm paral RT X15 (+45/-45/+45/10 mm parallel RT E(G), UTS(S) 1.522012.5 45/+45/-45)
Vp=0.159, respectively. The density of the plate reached 3.0015 g/cm3 . Dog-bone type specimens were cut from the plate in 0-direction by waterjet. For a ®rst test series waisting of the specimens from a width of 15 mm in the grip range to a 10 mm wide gauge section was performed according to the dimensions in Fig. 3 and Table 1. As most of the specimens tested in the ®rst test series broke near the radius the shape of the specimen was adapted to give a smoother change of width. The x±y coordinates for the changed specimen shape is given in Table 2. The dog-bone specimens were loaded up to failure in a tensile test performed at dierent temperatures between room temperature and 1200C as a maximum. In order to guarantee a homogeneous temperature distribution, a holding time of 10 min was applied after reaching the required temperature. Heating of the middle section of the specimen is realized by inductive heating of a susceptor. The temperature is measured with the aid of a thermocouple in the middle of the specimen. The heating coil with a height of 100 mm and consisting of 6 windings was adjusted in such a way, that the temperature in the 40 mm gauge length was as homogeneous as possible. The maximum temperature deviation in the gauge length of 40 mm measured +3% ÿ0% at a temperature of 1200C in the middle. Stress± strain behaviour at all temperatures is measured with the aid of a clip gauge with a gauge length of 20 mm as schematically presented in Fig. 4. Parallel sided specimens with a width of 10 mm as indicated in Table 3 were also cut from the plates but now with the aid of a diamond wire machine. Specimens cut under an angle of 45 with respect of the 0-direction are used to measure the shear modulus G and the intralaminar shear strength at room temperature. Further 10 mm wide specimens were cut transverse to the 0-direction. After this these specimens were ground with a diamond wheel on both surfaces in order to remove the surface layers so that ®nally specimens with the orientation 900.2/02/900.2 remain. The stiness of the crossply specimens with two dierent relative amounts of transverse plies (V90=50% and V90=16.6% respectively) allows the determination of the moduli of the parallel and transverse ply (E==; E?). 3. Results and discussion A calculation of the longitudinal and transverse ply stiness E== and E? with the aid of the moduli of the two types of crossply specimens delivers the values indicated in Table 4. The modulus parallel to the ®bres can be approximated very well by the rule of mixtures: Fig. 3. Dog-bone specimen dimensions for (0/90/0/90/0/90)s tensile test. Table 1 Radius of dog-bone specimen (grip length 55 mm) X 0.000 1.612 2.955 4.567 5.830 7.782 9.000 9.931 Y 0.000 0.027 0.060 0.118 0.179 0.314 0.432 0.545 X 11.507 12.555 13.245 13.755 14.131 14.382 15.000 Y 0.814 1.068 1.291 1.492 1.677 1.827 2.500 Table 2 Radius of adapted dog-bone specimen (grip length 40 mm) X 0.000 1.612 2.955 4.567 5.830 7.782 9.000 10 Y 0.000 0.027 0.060 0.118 0.179 0.314 0.432 0.54 X 15 20 25 30 Y 1.03 1.52 2.01 2.5 Table 3 Test specimens to characterize the mechanical behaviour of Nextel 610/Umox Specimen Specimen type Test temperature Mechanical properties (0/90/0/90/0/90)s 10 mm dog-bone RT-1200C E, UTS 900.2/02/900.2 10 mm parallel RT E (+45/ÿ45/+45/ ÿ45/+45/ÿ45)s 10 mm parallel RT E(G), UTS(S) Fig. 4. Schematics of experimental set-up for high temperature tensile testing of Nextel 610/Umox. P.W.M. Peters et al. / Journal of the European Ceramic Society 20 (2000) 531±535 533
P.W.M. Peters et al. /Journal of the European Ceramic Society 20(2000)531-535 Table 4 Elastic constants and intralaminar shear strengths of different CMCs with a porosity content of 10-15% Process Material Er GPa Em GPa Ey GPa Matrix contr. to En% E, GPa Gn GPa S MPa 21.8 CVI SiC/SIC E∥=ErVr+EmV linearly to 145.2 MPa at 800C. Above 800C a stronger in the absence of porosity. Taking Er(Nextel 610)=331 drop occurs. The modulus is maintained around 100 GPa? and Em=220 GPao for mullite the resulting GPa up to a maximum temperature of 1000C.At modulus of E/= 271 GPa is substantially higher than 1200 C a strong drop in modulus takes place. The the value given in Table 4 stress-strain curves at the different temperatures are Due to porosity and the presence of the probably not practically linear up to failure with the tendency for continuous Si-O-C network in which the mullite pa larger non-linearity at larger temperatures. Non-linear cles are embedded, the matrix contributes only 6% of type of failure giving rise to the before-mentioned quasi the theoretical modulus of mullite ductility can be a result of A comparison made in Table 4 with other CMCs(C/c and Sic/SiC) shows that this behaviour is characteristic (a) matrix cracking for CMCs produced by the pyrolysis of a polymer matrix b) fibre failure resulting in loose bundle type of failure (PP). If the matrix is produced by chemical vapour It has been discussed that the matrix contributes little impregnation(CVI) a continuous stiff matrix with a max- to the modulus of the composite Matrix cracking, thus, imum contribution of the matrix [ Em(1-Vp-Vr)] can hardly give rise to a quasi-ductile behaviour. The can be possible. I only phenomenon which can produce a substantial The shear modulus Gl, can be determined with the amount of non-linearity is fibre failure due to a loose aid of a tensile test on a +45-specimen, if the strain in bundle type of failure. Condition for this type of failure and transverse to the loading direction, &n and et is a low capability of transferring stresses between fibre respectively is measured. Fig. 5 shows the result of such and the matrix an experiment. The shear modulus follows from This low stress transfer capability usually leads to a considerable pull-out. Microscopical observations of G=a/(E1-E1) (2) fracture surfaces, given in Fig. 7, indicate in agreement with the occurrence of little non-linearity only limited whereas the intra-laminar shear strength is given by pull-out of broken fibres, mainly near porous areas. No substantial difference between the fracture surfaces pro- S=Omax/2 ( duced at room temperature and e.g. at 800C test tem- perature occurred. Fracture surfaces illustrate that a The results of the tensile tests at temperatures ranging homogeneous fugitive interface with a theoretical thick from room temperature to 1200oC are indicated in Fig. ness of 0. l um is not present around all fibres but local 6. The strength of the(0/90/0/90/0/90)s specimens at room temperature measures 177. 4 MPa and decreases -+-modulus 目 「 transverse40 axial 80 40140050 01015 200400600 010001200 strain Fig. 6. Tenile strength and modulus of Nextel 610/Mox crossply Fig. 5. Stress-strain curve of +45 specimen at room temperature. laminate as a function of the test temperature
E== EfVf EmVm 1 in the absence of porosity. Taking Ef (Nextel 610)=331 GPa9 and Em 220 GPa10 for mullite the resulting modulus of E== 271 GPa is substantially higher than the value given in Table 4. Due to porosity and the presence of the probably not continuous Si±O±C network in which the mullite particles are embedded, the matrix contributes only 6% of the theoretical modulus of mullite. A comparison made in Table 4 with other CMCs (C/C and SiC/SiC) shows that this behaviour is characteristic for CMCs produced by the pyrolysis of a polymer matrix (PP). If the matrix is produced by chemical vapour impregnation (CVI) a continuous sti matrix with a maximum contribution of the matrix [ Em 1 ÿ Vp ÿ Vf] can be possible.11 The shear modulus G==;? can be determined with the aid of a tensile test on a 45-specimen, if the strain in and transverse to the loading direction, "l and "t respectively, is measured. Fig. 5 shows the result of such an experiment. The shear modulus follows from: G = " l ÿ "t 2 whereas the intra-laminar shear strength is given by S max=2 3 The results of the tensile tests at temperatures ranging from room temperature to 1200C are indicated in Fig. 6. The strength of the (0/90/0/90/0/90)s specimens at room temperature measures 177.4 MPa and decreases linearly to 145.2 MPa at 800C. Above 800C a stronger drop occurs. The modulus is maintained around 100 GPa up to a maximum temperature of 1000C. At 1200C a strong drop in modulus takes place. The stress-strain curves at the dierent temperatures are practically linear up to failure with the tendency for larger non-linearity at larger temperatures. Non-linear type of failure giving rise to the before-mentioned quasiductility can be a result of (a) matrix cracking (b) ®bre failure resulting in loose bundle type of failure. It has been discussed, that the matrix contributes little to the modulus of the composite. Matrix cracking, thus, can hardly give rise to a quasi-ductile behaviour. The only phenomenon which can produce a substantial amount of non-linearity is ®bre failure due to a loose bundle type of failure. Condition for this type of failure is a low capability of transferring stresses between ®bre and the matrix. This low stress transfer capability usually leads to a considerable pull-out. Microscopical observations of fracture surfaces, given in Fig. 7, indicate in agreement with the occurrence of little non-linearity only limited pull-out of broken ®bres, mainly near porous areas. No substantial dierence between the fracture surfaces produced at room temperature and e.g. at 800C test temperature occurred. Fracture surfaces illustrate that a homogeneous fugitive interface with a theoretical thickness of 0.1 mm is not present around all ®bres but local Table 4 Elastic constants and intralaminar shear strengths of dierent CMCs with a porosity content of 10±15% Process Material Ef GPa Em GPa E== GPa Matrix contr. to E==% E? GPa G//? GPa S MPa PP C/C 331 35 206 0 0 6 21.8 PP N610/Umox 325 220 163 6 54.4 23.0 36 CVI SiC/SiC 200 469 265 100 130 65 ± Fig. 5. Stress-strain curve of 45 specimen at room temperature. Fig. 6. Tenile strength and modulus of Nextel 610/Umox crossply laminate as a function of the test temperature. 534 P.W.M. Peters et al. / Journal of the European Ceramic Society 20 (2000) 531±535
P.W.M. Peters et al. / Journal of the European Ceramic Society 20(2000)531-535 o the modulus of the laminate(E0/o=108.4 GPa at room temperature) is small, so that the matrix cannot contribute to a quasi-ductile stress-strain behaviour The intra-laminar shear strength(at room temperature 36 MPa)is rather high, which leads to limited fibre pull out visible at fracture surfaces of crossply specimens References I. Beyerle, D.s., Spearing Damage and failure in un ceramic matrIx composites. 2. Strife. J. R. Ceramic coat for carbon-carbon composites. eramic Bulletin. 1988.. 369-374 Fig.7.SEM-picture of fracture surface of 0o-ply in crossply specimen 3. Ohlhors, C.W.Vaughn, W.L. and Barratt, D.M., Current tested at room temperature research in oxidation-resistant carbon-carbon composites at NASA Langley, NASAReport N93-12456, 149-158, 1993 Tu. W. Lange, F. F. and Evans, A. G, Concept for damage. points of fibre/matrix contact are visible. These contact tolerant ceramic composite with "strong"interfaces. J. Am points realise fibre/matrix stress transfer quantified by Ceran.Soc.,1996,792).417-424 the intra-laminar shear strength indicated in Table 4. A 5. Marshall. D. B. Davis. J.B.. Morgan. P. E D. and Porter, J.R. possibility to increase quasi-ductile effects is to increase Interface materials for damage-tolerant oxide composites. Key the thickness of the fugitive interface, which however is Engineering materials. 1997. 127-131 27-36 6. Cain, M. G. Cain, R. L, Tye, A, Rian, P, Lewis, M. H and accompanied by a reduction of the intralaminar shear Gent, J. Structure and stability of synthetic interfaces in CMCs strength Key Engineering Materials, 1997. 127-131, 37-51 7. Knabe. H. Haug, T, Schafer. W. and Waldenmaier, T, Oxide Ceramic Matrix Composites for Aerospace Applications. Dornier Gmbh und dornier luftfahrt gmbh. friedrichshafen conclusions 8. Duran, A, Aparicio, M., Rebstock, K. and Vogel, w. D, Rein- filtration processes for polymer derived fibre reinforced ceramics. A Nextel 610 fibre reinforced mullite-based ceramic Key Engineering Materials. 1997. 127-131. 287-294 matrix on the basis of a fugitive interface has been 9. Clemens, F, private communication. characterised. The crossply laminate(0/90/0/90/0/90)s 10. Schneider. H. Okada. K. and Pask J. Mullite and Mullite with a room temperature strength of 177.4 GPa showed Ceramics, John Wiley Sons. New York 11. Peters, P. W. M., Martin, E. and Pluvinage, P, Influence of por. a moderate drop in strength up to 800oC and a stronger osity and fibre coating on engineering elastic moduli of fibre. decrease up to 1200oC. The contribution of the matrix reinforced ceramics(SiC/SiC) Composites, 1995, 26(2), 108-114
points of ®bre/matrix contact are visible. These contact points realise ®bre/matrix stress transfer quanti®ed by the intra-laminar shear strength indicated in Table 4. A possibility to increase quasi-ductile eects is to increase the thickness of the fugitive interface, which however is accompanied by a reduction of the intralaminar shear strength. 4. Conclusions A Nextel 610 ®bre reinforced mullite-based ceramic matrix on the basis of a fugitive interface has been characterised. The crossply laminate (0/90/0/90/0/90)s with a room temperature strength of 177.4 GPa showed a moderate drop in strength up to 800C and a stronger decrease up to 1200C. The contribution of the matrix to the modulus of the laminate (E0/90=108.4 GPa at room temperature) is small, so that the matrix cannot contribute to a quasi-ductile stress±strain behaviour. The intra-laminar shear strength (at room temperature 36 MPa) is rather high, which leads to limited ®bre pull out visible at fracture surfaces of crossply specimens. References 1. Beyerle, D. S., Spearing, S. M., Zok, F. W. and Evans, A. G., Damage and failure in unidirectional ceramic matrix composites. J. Am. Ceram. Soc, 1992, 75(10), 2719±2725. 2. Strife, J. R., Ceramic coatings for carbon±carbon composites. Ceramic Bulletin, 1988, 67, 369±374. 3. Ohlhors, C.W., Vaughn, W.L. and Barratt, D.M., Current research in oxidation-resistant carbon±carbon composites at NASA Langley, NASA-Report N93-12456, 149±158, 1993. 4. Tu, W., Lange, F. F. and Evans, A. G., Concept for damagetolerant ceramic composite with ``strong'' interfaces. J. Am. Ceram. Soc., 1996, 79(2), 417±424. 5. Marshall, D. B., Davis, J. B., Morgan, P. E. D. and Porter, J. R., Interface materials for damage-tolerant oxide composites. Key Engineering Materials, 1997, 127±131, 27±36. 6. Cain, M. G., Cain, R. L., Tye, A., Rian, P., Lewis, M. H. and Gent, J., Structure and stability of synthetic interfaces in CMCs. Key Engineering Materials, 1997, 127±131, 37±51. 7. Knabe, H., Haug, T , SchaÈfer, W. and Waldenmaier, T., Oxide Ceramic Matrix Composites for Aerospace Applications. Dornier GmbH und Dornier Luftfahrt GmbH, Friedrichshafen. 8. Duran, A., Aparicio, M., Rebstock, K. and Vogel, W. D., Rein- ®ltration processes for polymer derived ®bre reinforced ceramics. Key Engineering Materials, 1997, 127±131, 287±294. 9. Clemens, F., private communication. 10. Schneider, H., Okada, K., and Pask, J., Mullite and Mullite Ceramics, John Wiley & Sons, New York. 11. Peters, P. W. M., Martin, E. and Pluvinage, P., In¯uence of porosity and ®bre coating on engineering elastic moduli of ®brereinforced ceramics (SiC/SiC). Composites, 1995, 26(2), 108±114. Fig. 7. SEM-picture of fracture surface of 0-ply in crossply specimen tested at room temperature. P.W.M. Peters et al. / Journal of the European Ceramic Society 20 (2000) 531±535 535