JAm. Ceram soc,8761996-100102004) ournal Preparation of Silicon-Titanium-Carbon-Oxygen Fabric/ Mullite Filler/Polytitanocarbosilane Laminates b Polymer Impregnation and Pyrolysis Method Yoshihiro Hirata, Kazunori Hayata, and Tomoyuki Maeda Department of Advanced Nanostructured Materials Science and Technology, Graduate School of Science and Engineering, Kagoshima University, Kagoshima 890-0065, Japan Tyranno Fiber Development Project, Ube Industries, Ltd, Ube 755-0067, Japan A polytitanocarbosilane(PTC, 20-50 mass%)xylene solution CVI or Rs process, the advantages of the pip process are was infiltrated into a porous, laminated woven fabric of 21-33 low-temperature processing and simple processing vol%Si-Ti-C-O fibers including 26-46 vol% mullite powder (low-cost processing). Low-temperature processing prey (filler)and decomposed at 1000C in an argon atmosphere radation of the fibers associated with pyrolysis, , cryst repeated eight times to produce a laminated composite of between the fibers and the matrix. -ve Sintering at the 68%0-85% of theoretical density. The effects of the poly In our previous papers, 38.49 a polytitanocarbosilane(PTC concentration and the fraction of mullite filler on the densifi- precursor of SiC fibers)xylene solution was impregnated into a cation rate and microstructure of the layered composite were rous Si-Ti-C-O fiber/mullite filler composite and decomposed at studied. The pseudoductility of the densified composite, as 000° C in an atmosphere. Mullite (3AL,O, 2SiO2) was a easured using four-point flexural testing, was caused by candidate oxide for high-temperature structural applications b buckling after the elastic deformation and was followed by cause of its high melting point, low thermal expansion coefficien delamination along the direction of the layered fabric. The and high creep resistance The densification of the composite strength and the energy of fracture were enhanced by control- proceeded by the formation of a polymer-derived solid in the ling the incorporation of mullite filler in the filament yarn (formation of a narrow pore-size distribution)and densifica pores. The PIP sequence was repeated eight times to produce a ered composite with 90% of theoretical density. This composite tion with a low-viscosity PTC solution. The composite with a nowed a significant pseudoductility with a flexural strength o higher strength provided a higher energy of fracture. The 90 MPa. ,4 Based on the above successful PIP process, we maximum energy of fracture reached 22 kJ/m in the compos studied the influence of the concentration of the ptc solution and ite with 330 MPa of strength in four-point flexure the role of mullite filler on the densification microstructure. and mechanical properties of the layered Si-Ti-C-O fabric/mullite L. Introduction powder/PTC composites. The decreased porosity of the composite with increased PIP sequences was analyzed theoretically, and the C ONTINUOUS FIBER-REINFORCED CERAMIC-MATRIX COMPOSITES esults were used to control the effective experimental parameters CFCCs) are known for their high damage tolerance with pseudoductility, and they can be applied in high-temperature structural material applications, such as aircraft engine compo- IL. Experimental Procedures nents or space plane engine parts. The development of me chanically reliable composites requires a weak chemical bond (I Green Laminated Composite between the fibers and matrix to enhance the bridging and pullout A micrometer-sized flakelike mullite powder with the following effects of fibers on crack propagation. The debonding length of properties was inserted as a filler into the open spaces between the fibers in the pullout depends on the roughness of the fiber surface continuous fibers: chemical composition(mass%)of 71.46 Al2O3, the strength of the fibers, and the residual stresses around the 28.13 SiO2,0.30 Zr02, 0. 10 TiO,, and 0.01 Na,O; and cumulative interfaces that result from the difference of thermal expansion particle-size distribution of 0.7 um(10%o), 1.7 um(50% ), and 3.8 coefficients and Youngs moduli between the fibers and the um(90%)(Chichibu Cement Co, Ltd, Tokyo, Japan).An matrix 2-35 aqueous mullite suspension with 40 vol% solids and a polyacrylic- The matrix phase of CFCCs is usually fabricated by ammonium dispersant(PAA; average molecular weight of 10 000) or near-net-shape processing methods, such as chemic at 0.65 mass% of the mullite was prepared to infiltrate into the PIP), 38-40 or reaction sintering(RS) nation and laminated fabric sheets. The pH of the suspension was adjusted to 8.5 using an NH OH solution to enhance the dispersion of the longer processing time for PIP sequences as compared with the negatively charged mullite particles by electrosteric stabilization with PAA. The suspensions were stirred for 24 h at room temperature, after which ultrasonic vibration at 20 kHz was applied for 5 min to disperse the particle agglomerates. Air bubbles T. Parthasarathy--contributing editor in the suspension were eliminated using a vacuum pump. Ten sheets of Si-Ti-C-0 plain-weave or satin-weave fabric(270 um thick, 25 mm wide, and 38 mm long) were laminated together in anuscript No. 186512. Received December 3, 2002: approved February 10, the aqueous mullite suspension under a pressure of 9.8 kPa (chemical composition(mass%)of 54.0 Si, 2.0 T1, 31.6 C, ember, American Ceramic Society 12. 4 0, diameter of ll um; tensile strength of 3.6 GPa, and
Preparation of Silicon-Titanium-Carbon-Oxygen Fabric/ Mullite Filler/Polytitanocarbosilane Laminates by Polymer Impregnation and Pyrolysis Method Yoshihiro Hirata,* Kazunori Hayata, and Tomoyuki Maeda Department of Advanced Nanostructured Materials Science and Technology, Graduate School of Science and Engineering, Kagoshima University, Kagoshima 890-0065, Japan Masaki Shibuya* Tyranno Fiber Development Project, Ube Industries, Ltd., Ube 755-0067, Japan A polytitanocarbosilane (PTC, 20–50 mass%)–xylene solution was infiltrated into a porous, laminated woven fabric of 21–33 vol% Si-Ti-C-O fibers including 26–46 vol% mullite powder (filler) and decomposed at 1000°C in an argon atmosphere. This polymer impregnation and pyrolysis method (PIP) was repeated eight times to produce a laminated composite of 68%–85% of theoretical density. The effects of the polymer concentration and the fraction of mullite filler on the densification rate and microstructure of the layered composite were studied. The pseudoductility of the densified composite, as measured using four-point flexural testing, was caused by buckling after the elastic deformation and was followed by delamination along the direction of the layered fabric. The strength and the energy of fracture were enhanced by controlling the incorporation of mullite filler in the filament yarn (formation of a narrow pore-size distribution) and densification with a low-viscosity PTC solution. The composite with a higher strength provided a higher energy of fracture. The maximum energy of fracture reached 22 kJ/m2 in the composite with 330 MPa of strength in four-point flexure. I. Introduction C ONTINUOUS FIBER-REINFORCED CERAMIC-MATRIX COMPOSITES (CFCCs) are known for their high damage tolerance with pseudoductility, and they can be applied in high-temperature structural material applications, such as aircraft engine components or space plane engine parts.1–11 The development of mechanically reliable composites requires a weak chemical bond between the fibers and matrix to enhance the bridging and pullout effects of fibers on crack propagation. The debonding length of fibers in the pullout depends on the roughness of the fiber surface, the strength of the fibers, and the residual stresses around the interfaces that result from the difference of thermal expansion coefficients and Young’s moduli between the fibers and the matrix.12–35 The matrix phase of CFCCs is usually fabricated by net-shape or near-net-shape processing methods, such as chemical vapor infiltration (CVI),17,36,37 polymer impregnation and pyrolysis (PIP),38–40 or reaction sintering (RS).41,42 Although it takes a longer processing time for PIP sequences as compared with the CVI or RS process, the advantages of the PIP process are low-temperature processing and simple processing equipment (low-cost processing). Low-temperature processing prevents degradation of the fibers associated with pyrolysis,43,44 crystallization, or grain growth,45 and it also inhibits sintering at the interfaces between the fibers and the matrix.46–48 In our previous papers,38,49 a polytitanocarbosilane (PTC, precursor of SiC fibers)–xylene solution was impregnated into a porous Si-Ti-C-O fiber/mullite filler composite and decomposed at 1000°C in an argon atmosphere. Mullite (3Al2O32SiO2) was a candidate oxide for high-temperature structural applications because of its high melting point, low thermal expansion coefficient, and high creep resistance.50–52 The densification of the composite proceeded by the formation of a polymer-derived solid in the pores. The PIP sequence was repeated eight times to produce a layered composite with 90% of theoretical density. This composite showed a significant pseudoductility with a flexural strength of 290 MPa.38,49 Based on the above successful PIP process, we studied the influence of the concentration of the PTC solution and the role of mullite filler on the densification, microstructure, and mechanical properties of the layered Si-Ti-C-O fabric/mullite powder/PTC composites. The decreased porosity of the composite with increased PIP sequences was analyzed theoretically, and the results were used to control the effective experimental parameters. II. Experimental Procedures (1) Green Laminated Composite A micrometer-sized flakelike mullite powder with the following properties was inserted as a filler into the open spaces between the continuous fibers: chemical composition (mass%) of 71.46 Al2O3, 28.13 SiO2, 0.30 ZrO2, 0.10 TiO2, and 0.01 Na2O; and cumulative particle-size distribution of 0.7 m (10%), 1.7 m (50%), and 3.8 m (90%) (Chichibu Cement Co., Ltd., Tokyo, Japan). An aqueous mullite suspension with 40 vol% solids and a polyacrylicammonium dispersant (PAA; average molecular weight of 10 000) at 0.65 mass% of the mullite was prepared to infiltrate into the laminated fabric sheets. The pH of the suspension was adjusted to 8.5 using an NH4OH solution to enhance the dispersion of the negatively charged mullite particles by electrosteric stabilization with PAA.53 The suspensions were stirred for 24 h at room temperature, after which ultrasonic vibration at 20 kHz was applied for 5 min to disperse the particle agglomerates. Air bubbles in the suspension were eliminated using a vacuum pump. Ten sheets of Si-Ti-C-O plain-weave or satin-weave fabric (270 m thick, 25 mm wide, and 38 mm long) were laminated together in the aqueous mullite suspension under a pressure of 9.8 kPa (chemical composition (mass%) of 54.0 Si, 2.0 Ti, 31.6 C, and 12.4 O; diameter of 11 m; tensile strength of 3.6 GPa; and T. Parthasarathy—contributing editor Manuscript No. 186512. Received December 3, 2002; approved February 10, 2004. *Member, American Ceramic Society. J. Am. Ceram. Soc., 87 [6] 996–1001 (2004) 996 journal
June 2004 Preparation of Si-Ti-C-0 Fabric/Mullite Filler/ Polytitanocarbosilane Laminates by PlP Method Youngs modulus of 186 GPa(Ube Industries, Ltd, Y amaguchi Japan). Excess suspension overflowed outside the mold The dried, green, laminated composites were calcined for I h at 1 100oC in an argon atmosphere. The characteristics of the lami nated composites are shown in Table I. In some samples,a pressure of 39 MPa was applied during the calcination to decreas the porosity of the composite before the PIP process. The fraction 10 of mullite powder filler in the calcined composite was 25-40 vol%, except for sample F. No filler was inserted into the fabric of sample F. The fiber fraction was in the range of 21-33 vol% 010203040506 (2) PIP Process with PTC Solution and Mechanical Polvtitanocarbosilane/ mass Properties of Densified Composites Figure I shows the viscosity of the xylene solution containing Fig. 1. Viscosity of the PTC-xylene solution used for the impregnation PTC of molecular weight 15 000(chemical composition(mass%) into the laminated porous composite of 44 S1, 2 T1, 42 C, 9 H, and 3O(Ube Industries )). The viscosity gradually increased to 30 mass% of the polymer and then rapidl creased at higher fractions of the polymer. In the present PIP process, precursor solutions containing 20-50 mass% polymer calcination of the laminated composites with a similar fiber fraction provided no increase of the packing density(samples A, nder vacuum to understand the concentration effect of the B, C, and D). The weave density affected the porosity of the polymer on the densification of the composites. The infiltrate laminated composite. The plain-weave fabric(17 yarn/2.54 cm x polymer was heated to 170C in air to cure the polymer, and it was 17 yarn/2. 54 cm, 800 filament/yarn) gave a lower porosity com- thermally decomposed at 1000C in an argon atmosphere to form posite than the satin-weave fabric (20 yarn/2.54 cm X 20 yarn/2.54 an inorganic residue. This PIP sequence was repeated eight times cm)at a similar fiber content because of the high packing to decrease the porosity. The density of the PIP-processed com- characteristic of mullite filler in the plain-weave fabric(samples E posite was measured using the Archimedes method with kerosene and G) The microstructures of the densified composites were observed The pore volume(Po) of the composite was important using optical microscopy and scanning electron microscopy in decreasing the porosity usil process In our previous (Model SM300, Topcon Technologies, Inc, Tokyo, Japan). The paper, we derived Eq.(1) porosity (P)after n PIP mposites were cut using a diamond wheel to test samples 4 mm sequences. ide, 6 mm high, and 38 mm long and were polished using Nos 600 and 2000 SiC papers. The flexural strength of each sample was P=Pol1-y-C easured at room temperature using the four-point flexural ethod with an upper span of 10 mm and a lower span of 30 mm and a crosshead speed of 0.5 mm/min( Model AGIOTA, Shimadzu where y is the ceramic yield of solids-to-polymer (y =0.88 for Seisakusho, Ltd, Kyoto, Japan). Two or three samples per PTC), D, the density of the polymer(1. 192 g/cm), Ds the density processing condition were tested. The appearance of the compos- of the solid formed (1.982 g/cm), and Cn the polymer concentra ites during bend testing was photographed using a digital camera. where I', is the volume of polymer and y, the volume of xyl Equation(1)indicates that(i) pore elimination efficiency (dP/dn) II. Results and discussio decreases as the number of PIP sequences increases, (ii)decreased Po and increased Cn effectively decrease the porosity for a given () Densification of the laminated Composites Using PIP The fractions of the fibers and the mullite filler laminated composites after the impregnation of PTC are shown in porosity Table I. The incorporation of the mullite filler in the laminated Figure 2(a) shows the decrease of abric greatly influenced the porosity before the PIp process. N fabric/mullite composite(sample B)with increasing Hp C-o mullite filler(sample F) resulted in 70% porosity. The high quences. In sample B, 20, 30, and 40 mass% polymer solutions fraction(46 vol%) of mullite filler in sample e decreased the were used in the first PIP sequence, the second through fourth PIP porosity to 33%. Samples A, B, C, and D with 31-33 vol% sequences, and fifth through eighth PIP sequences, respectively Si-Ti-C-O fabric contained 26-27 vol% mullite filler and 41-44 The solid line represents the porosity of open pores, calculated vol% porosity. The application of pressure of 39 MPa during the using Eq (1). The porosity from Eq (1)explains the trend of the Table I. Phase Compositions and Mechanical Properties of the Laminated Composites of the Si-Ti-C-O Fabrie-Mullite-PTC System after the Eighth PIP Sequence Fiber(vol%) 31.l( plain)30.8 32.9(plain) 32.9(plain) 21.2 (plain) 29.9(plain) 21.7( satin) Mullite filler(vol%) 25.5 ure during calcination 0 0 39 (MPa) PTC-derived solid(vol% 26.6 29.1 15.5 38.4 21.2 Closed pore(vol%) Young's modulus at elas 40±4 30±4 32±6 15±1 19±3 13±1 deformation(GPa) eda30m1223 3=13 111±2 69.1 12 7.2-7 03 139±2
Young’s modulus of 186 GPa (Ube Industries, Ltd., Yamaguchi, Japan)). Excess suspension overflowed outside the mold. The dried, green, laminated composites were calcined for 1 h at 1100°C in an argon atmosphere. The characteristics of the laminated composites are shown in Table I. In some samples, a pressure of 39 MPa was applied during the calcination to decrease the porosity of the composite before the PIP process. The fraction of mullite powder filler in the calcined composite was 25–46 vol%, except for sample F. No filler was inserted into the fabric of sample F. The fiber fraction was in the range of 21–33 vol%. (2) PIP Process with PTC Solution and Mechanical Properties of Densified Composites Figure 1 shows the viscosity of the xylene solution containing PTC of molecular weight 15 000 (chemical composition (mass%) of 44 Si, 2 Ti, 42 C, 9 H, and 3 O (Ube Industries)). The viscosity gradually increased to 30 mass% of the polymer and then rapidly increased at higher fractions of the polymer. In the present PIP process, precursor solutions containing 20–50 mass% polymer were impregnated into the porous green composites for 40 min under vacuum to understand the concentration effect of the polymer on the densification of the composites. The infiltrated polymer was heated to 170°C in air to cure the polymer, and it was thermally decomposed at 1000°C in an argon atmosphere to form an inorganic residue. This PIP sequence was repeated eight times to decrease the porosity. The density of the PIP-processed composite was measured using the Archimedes method with kerosene. The microstructures of the densified composites were observed using optical microscopy and scanning electron microscopy (Model SM300, Topcon Technologies, Inc., Tokyo, Japan). The composites were cut using a diamond wheel to test samples 4 mm wide, 6 mm high, and 38 mm long and were polished using Nos. 600 and 2000 SiC papers. The flexural strength of each sample was measured at room temperature using the four-point flexural method with an upper span of 10 mm and a lower span of 30 mm and a crosshead speed of 0.5 mm/min (Model AG10TA, Shimadzu Seisakusho, Ltd., Kyoto, Japan). Two or three samples per processing condition were tested. The appearance of the composites during bend testing was photographed using a digital camera. III. Results and Discussion (1) Densification of the Laminated Composites Using PIP The fractions of the fibers and the mullite filler in the green laminated composites after the impregnation of PTC are shown in Table I. The incorporation of the mullite filler in the laminated fabric greatly influenced the porosity before the PIP process. No mullite filler (sample F) resulted in 70% porosity. The high fraction (46 vol%) of mullite filler in sample E decreased the porosity to 33%. Samples A, B, C, and D with 31–33 vol% Si-Ti-C-O fabric contained 26–27 vol% mullite filler and 41–44 vol% porosity. The application of pressure of 39 MPa during the calcination of the laminated composites with a similar fiber fraction provided no increase of the packing density (samples A, B, C, and D). The weave density affected the porosity of the laminated composite. The plain-weave fabric (17 yarn/2.54 cm 17 yarn/2.54 cm, 800 filament/yarn) gave a lower porosity composite than the satin-weave fabric (20 yarn/2.54 cm 20 yarn/2.54 cm) at a similar fiber content because of the high packing characteristic of mullite filler in the plain-weave fabric (samples E and G). The pore volume (P0) of the starting composite was important in decreasing the porosity using the PIP process. In our previous paper,38 we derived Eq. (1) for the porosity (P) after n PIP sequences: P P0 1 Y Dp Ds Cp n (1) where Y is the ceramic yield of solids-to-polymer (Y 0.88 for PTC), Dp the density of the polymer (1.192 g/cm3 ), Ds the density of the solid formed (1.982 g/cm3 ), and Cp the polymer concentration (vol%) in the xylene solution (Cp Vp/(Vp Vx) Vp/P0, where Vp is the volume of polymer and Vx the volume of xylene). Equation (1) indicates that (i) pore elimination efficiency (dP/dn) decreases as the number of PIP sequences increases, (ii) decreased P0 and increased Cp effectively decrease the porosity for a given polymer, and (iii) a polymer with a high Y value (low loss of mass during pyrolysis of the polymer) is desirable for decreasing porosity. Figure 2(a) shows the decrease of porosity in the Si-Ti-C-O fabric/mullite composite (sample B) with increasing PIP sequences. In sample B, 20, 30, and 40 mass% polymer solutions were used in the first PIP sequence, the second through fourth PIP sequences, and fifth through eighth PIP sequences, respectively. The solid line represents the porosity of open pores, calculated using Eq. (1). The porosity from Eq. (1) explains the trend of the Fig. 1. Viscosity of the PTC–xylene solution used for the impregnation into the laminated porous composite. Table I. Phase Compositions and Mechanical Properties of the Laminated Composites of the Si-Ti-C-O Fabric–Mullite–PTC System after the Eighth PIP Sequence Property Value ABCDE FG Fiber (vol%) 31.1 (plain) 30.8 (plain) 32.9 (plain) 32.9 (plain) 21.2 (plain) 29.9 (plain) 21.7 (satin) Mullite filler (vol%) 25.9 25.5 25.9 26.5 45.9 0 28.0 Applied pressure during calcination at 1100°C (MPa) 0 0 0 39 39 0 0 PTC-derived solid (vol%) 26.6 29.1 24.4 22.2 15.5 38.4 21.2 Open pore (vol%) 11.0 9.6 13.7 14.9 9.2 31.7 24.8 Closed pore (vol%) 5.4 5.0 3.1 3.5 8.2 0 4.3 Young’s modulus at elastic deformation (GPa) 40 4 30 4 32 6 15 1 19 3 20 1 13 1 Strength (MPa) 312 17 264 14 227 9 111 2 172 2 69.1 0.3 139 2 Energy of fracture at 3.0 mm of displacement (kJm2 ) 19.4–22.3 9.9 18.5 10.2 12.8 7.2–7.4 4.4 June 2004 Preparation of Si-Ti-C-O Fabric/Mullite Filler/Polytitanocarbosilane Laminates by PIP Method 997
Journal of the American Ceramic Sociery-Hirata et al Vol 87. No 6 pores from the calculated line represents the decreasing elimi- 60 (a)Sample. Open Closed nation efficiency of open pores. With increasing PIP seque using the concentrated PTC solution, the pore channels be 40 narrow quickly. This causes difficulty in smooth flow of a high viscosity PTC solution into the inside of the composite, which xplains the suppression of closed-pore formation. That is, the increase in viscosity of the concentrated PTC solution decreases the densification rate of the composite 0 The effect of incorporation of mullite filler on the densification of the composites is understood by comparing the porosity among ● Closed pore samples B, E, and F(Table D). The f sample E, with Open pore, Eq(1) the highest content of mullite filler(46 vol%), disappeared more rapidly than the prediction using Eq(1). After the eighth PIP sequence with 30 mass% PTC solution, 9.2% of open pores and 8.2% of closed pores remained in sample E. Therefore, sample E exhibited a tendency to form more closed pores than samples A 0 and B when processed with PTC solutions of a similar viscosity 60 (e)Sample F This result was explained as follows. Incorporation of mullite filler in the Si-Ti-C-O fabric narrowed the pore size in the starting stage Open pore, Eq(I) of the PIP process and decreased the pore volume. As a resul oo O the open pores in the composite with a higher amount of mullite filler were eliminated more easily by the PlP process. Some pore 20 also were transformed to closed pores by the pyrolysis of PTC solution infiltrated in the narrow spaces of the composite As compared with the densification behavior of the composites 0 with mullite filler, the measured porosity of open pores in sample F with no mullite filler deviated greatly from the porosity predicted Number of impregnation sequence using Eq (1). Figure 2(c)shows that few closed pores were formed in sample F No filler in the layered fabric increased the porosity Fig.2. Porosity of the Si-Ti-C-O fabric/mullite filler/PTC system(sam. before the PIP process. The increased initial porosity decreased the oles B, C, and F) as a function of the number of PIP sequence. See Table densification rate of the composite. In addition, the PTC solution I for samples. in large pores had a possibility to elute outside the laminated composite durin pregnation process with the Ptc solutio After the eighth PIP process with 50 mass% PTC solution, 32% of ata alues higher than the measured open pores and 0% of closed pores remained in sample F porosity of open pores, which indicates more disappearance of When the satin-weave fabric of Si-Ti-C-0 fibers was laminated ores than the prediction. This discrepancy is associated with in the aqueous mullite suspension, open pores of 50 vol% were the formation of closed pores from the open pores. Figure 2(a) formed before the PIP process(sample G in Table I). The shows that the amount of closed pores in sample B increases densification behavior of sample G with 50 mass% PTC solution with increasing PIP sequences. Once closed pores we was similar to that of sample C shown in Fig. 2. That is, the formed, further densification of the composite was suppresse nterpretation for the densification in sample C also was applied because of limited impregnation by the polymer. That is, the for the decreased porosity in sample G. After the eighth PIP gradual deposition of the polymer-derived solid from the inside to process, 25% of open pores and 4% of closed pores remained in he outside of the composite was an important factor for the sample c oppression of the formation of closed pores. The densification ehavior of sample A with 20-30 mass% polymer solution was similar to the result of sample B with 20-40 mass% polymer (2) Mechanical Properties of the Densified Composites olution. In these samples, a plp sequence with a low concentra- able I summarizes the phase compositions of the laminated tion of PTC solution was followed by a plp sequence with a higher composites after the eighth PIP sequence. The fiber fractions were concentration of PTc solution to prevent the formation of closed in the range from 21 to 33 vol%. The PIp process controlled the pores that resulted from the pyrolysis of a concentrated polymer fractions of Ptc-derived solid and porosity of the composite, as (Fig. 1). However, 5%6% of closed pores in addition to discussed in section Il(1). The mechanical properties of the I 1%of open pores remained in samples A and b after th aminated composites in Table I were studied at room temperatur PIP sequence (Table I). In sample D, a 30 mass% PTC Figure 3(a) shows the typical stress-displacement curve for with a low viscosity was used in the first through fourth PIP sample A Samples A-D, with similar fractions of fiber and mullite equences, and, subsequently, a 50 mass% PTC solution was filler(Table D), showed elastic deformation in the initial stage of infiltrated into the composite in the fifth through eighth PIP the stress-displacement curves, followed by significant equences. The densification behavior of this sample also was pseudoductility. Samples A and c also showed an interesting similar to that of samples A and B. The amounts of open and increase of strength after 1.5 mm of displacement. The maximur closed pores of sample D were 14.9% and 3.5%, respectively, after strength at -0.5 mm of displacement is summarized in Table the eighth Plp sequence(Table n) and it became higher in the following order: sample D< sample Figure 2(b)shows the porosity in sample C with 50 mass% C< sample B sample A. A linear relationship was observed PTC solution as a function of the number of PIp sequence. The between the flexural strer and the y s moduli at the use of PTC solution of increased concentration may be an initial stage of the deformation for samples in Table I, except for effective method to decrease the porosity at a smaller number of sample F. The strength of the composite became higher with equences, as predicted by Eq(1) 2(b) shows that ncreased Youngs modulus. The Youngs moduli of samples A-D the measured porosity of open pores decreases along the with the similar fractions of fiber and mullite filler showed a calculated line to the second Plp sequence but deviates from the tendency to decrease when the concentration of PTc solution in ediction at h ther PIP sequences. On the other hand, the the PIP sequence increased. The microstructural inhomogeneity formation of closed pores is more suppressed in sample C tha that formed with the concentrated PtC solution may provide a sample A or B. These results lead to the following interpreta possible explanation for the strengths of samples A-D. The tion. The deviation in plus value of measured porosity of open detailed analysis will be studied
experimental data but gives values higher than the measured porosity of open pores, which indicates more disappearance of open pores than the prediction. This discrepancy is associated with the formation of closed pores from the open pores. Figure 2(a) shows that the amount of closed pores in sample B increases gradually with increasing PIP sequences. Once closed pores were formed, further densification of the composite was suppressed because of limited impregnation by the polymer. That is, the gradual deposition of the polymer-derived solid from the inside to the outside of the composite was an important factor for the suppression of the formation of closed pores. The densification behavior of sample A with 20–30 mass% polymer solution was similar to the result of sample B with 20–40 mass% polymer solution. In these samples, a PIP sequence with a low concentration of PTC solution was followed by a PIP sequence with a higher concentration of PTC solution to prevent the formation of closed pores that resulted from the pyrolysis of a concentrated polymer (Fig. 1). However, 5%–6% of closed pores in addition to 10%– 11% of open pores remained in samples A and B after the eighth PIP sequence (Table I). In sample D, a 30 mass% PTC solution with a low viscosity was used in the first through fourth PIP sequences, and, subsequently, a 50 mass% PTC solution was infiltrated into the composite in the fifth through eighth PIP sequences. The densification behavior of this sample also was similar to that of samples A and B. The amounts of open and closed pores of sample D were 14.9% and 3.5%, respectively, after the eighth PIP sequence (Table I). Figure 2(b) shows the porosity in sample C with 50 mass% PTC solution as a function of the number of PIP sequence. The use of PTC solution of increased concentration may be an effective method to decrease the porosity at a smaller number of PIP sequences, as predicted by Eq. (1). Figure 2(b) shows that the measured porosity of open pores decreases along the calculated line to the second PIP sequence but deviates from the prediction at higher PIP sequences. On the other hand, the formation of closed pores is more suppressed in sample C than sample A or B. These results lead to the following interpretation. The deviation in plus value of measured porosity of open pores from the calculated line represents the decreasing elimination efficiency of open pores. With increasing PIP sequences, using the concentrated PTC solution, the pore channels become narrow quickly. This causes difficulty in smooth flow of a high viscosity PTC solution into the inside of the composite, which explains the suppression of closed-pore formation. That is, the increase in viscosity of the concentrated PTC solution decreases the densification rate of the composite. The effect of incorporation of mullite filler on the densification of the composites is understood by comparing the porosity among samples B, E, and F (Table I). The open pores of sample E, with the highest content of mullite filler (46 vol%), disappeared more rapidly than the prediction using Eq. (1). After the eighth PIP sequence with 30 mass% PTC solution, 9.2% of open pores and 8.2% of closed pores remained in sample E. Therefore, sample E exhibited a tendency to form more closed pores than samples A and B when processed with PTC solutions of a similar viscosity. This result was explained as follows. Incorporation of mullite filler in the Si-Ti-C-O fabric narrowed the pore size in the starting stage of the PIP process and decreased the pore volume.38 As a result, the open pores in the composite with a higher amount of mullite filler were eliminated more easily by the PIP process. Some pores also were transformed to closed pores by the pyrolysis of PTC solution infiltrated in the narrow spaces of the composite. As compared with the densification behavior of the composites with mullite filler, the measured porosity of open pores in sample F with no mullite filler deviated greatly from the porosity predicted using Eq. (1). Figure 2(c) shows that few closed pores were formed in sample F. No filler in the layered fabric increased the porosity before the PIP process. The increased initial porosity decreased the densification rate of the composite. In addition, the PTC solution in large pores had a possibility to elute outside the laminated composite during the impregnation process with the PTC solution. After the eighth PIP process with 50 mass% PTC solution, 32% of open pores and 0% of closed pores remained in sample F. When the satin-weave fabric of Si-Ti-C-O fibers was laminated in the aqueous mullite suspension, open pores of 50 vol% were formed before the PIP process (sample G in Table I). The densification behavior of sample G with 50 mass% PTC solution was similar to that of sample C shown in Fig. 2. That is, the interpretation for the densification in sample C also was applied for the decreased porosity in sample G. After the eighth PIP process, 25% of open pores and 4% of closed pores remained in sample G. (2) Mechanical Properties of the Densified Composites Table I summarizes the phase compositions of the laminated composites after the eighth PIP sequence. The fiber fractions were in the range from 21 to 33 vol%. The PIP process controlled the fractions of PTC-derived solid and porosity of the composite, as discussed in section III(1). The mechanical properties of the laminated composites in Table I were studied at room temperature. Figure 3(a) shows the typical stress–displacement curve for sample A. Samples A–D, with similar fractions of fiber and mullite filler (Table I), showed elastic deformation in the initial stage of the stress–displacement curves, followed by significant pseudoductility. Samples A and C also showed an interesting increase of strength after 1.5 mm of displacement. The maximum strength at 0.5 mm of displacement is summarized in Table I, and it became higher in the following order: sample D sample C sample B sample A. A linear relationship was observed between the flexural strengths and the Young’s moduli at the initial stage of the deformation for samples in Table I, except for sample F. The strength of the composite became higher with increased Young’s modulus. The Young’s moduli of samples A–D with the similar fractions of fiber and mullite filler showed a tendency to decrease when the concentration of PTC solution in the PIP sequence increased. The microstructural inhomogeneity that formed with the concentrated PTC solution may provide a possible explanation for the strengths of samples A–D. The detailed analysis will be studied. Fig. 2. Porosity of the Si-Ti-C-O fabric/mullite filler/PTC system (samples B, C, and F) as a function of the number of PIP sequence. See Table I for samples. 998 Journal of the American Ceramic Society—Hirata et al. Vol. 87, No. 6
Preparation of Si-Ti-C-o Fabric/Mullite Filler Polytitanocarbosilane Laminates by PlP Method 00.511.2253 Displacement/mm omn Fig. 3. Stress-displacement curve(a) for sample A and the appearance(b)of sample A at the deformation stages of l-IV indicated in Fig. 3(a) The photograph shown in Fig 3(b) corresponds to the appear- previous paper: it increases as the relative densit ance of sample A at the deformation stages F-IV indicated in Fig. composite increases. The layered composite with 10 sheet 3(a). Buckling of the fabric occurred in the center of sample A at fabric and 8%-20% porosity exhibited 6-33 MPa of shea point Ill after elastic deformation(point ID). Further displacement along the layered fabric. The same energy absorption was accompanied by the gradual increase of the applied stress. The also is seen in samples B, C, D, and E appearance at point IV indicated no separation of the composite On the other hand, sample F, without mullite filler, showed a parallel to the direction of applied load. Figure 4 shows the different deformation behavior with relatively low strength (table microstructures before(Fig 4(a))and after(Fig. 4(b))the mea- I). After elastic deformation, sample F deformed to induce del surement of flexural strength of sample A. Sample A contained amination. No buckling was observed. The crack propagation large pores of 100-1000 um along the direction of the layered along the direction of the layered fabric reached the end of the fabric. The microstructure in Fig 4(b) indicates that many cracks, sample of 38 mm length. No macroscopic crack propagatio ormed at the center of sample A, propagated along the layered occurred parallel to the direction of applied load. As compared with the deformation behavior of samples A-F, sample G, with 2 based on the buckling of the SiC fabric and subsequent remarkable yol% SiC satin- weave fabric and 28 vol% mullite filler(Table D) delamination of the layered SiC fabric. The shear stress to induce fractured catastrophically with no pseudoductility. The total po the delamination of the layered SiC fabric has been examined in a rosity of open and closed pores of sample G was approximately (a) (b) 20kV 28x588ym 2mm 58pm54 mm Fig. 4. Microstructures(a)before and(b)after the measurement of flexural strength of sample A
The photograph shown in Fig. 3(b) corresponds to the appearance of sample A at the deformation stages I–IV indicated in Fig. 3(a). Buckling of the fabric occurred in the center of sample A at point III after elastic deformation (point II). Further displacement was accompanied by the gradual increase of the applied stress. The appearance at point IV indicated no separation of the composite parallel to the direction of applied load. Figure 4 shows the microstructures before (Fig. 4(a)) and after (Fig. 4(b)) the measurement of flexural strength of sample A. Sample A contained large pores of 100-1000 m along the direction of the layered fabric. The microstructure in Fig. 4(b) indicates that many cracks, formed at the center of sample A, propagated along the layered fabric. Therefore, the deformation behavior shown in Fig. 3 is based on the buckling of the SiC fabric and subsequent remarkable delamination of the layered SiC fabric. The shear stress to induce the delamination of the layered SiC fabric has been examined in a previous paper:49 it increases as the relative density of the composite increases. The layered composite with 10 sheets of SiC fabric and 8%–20% porosity exhibited 6–33 MPa of shear strength along the layered fabric. The same energy absorption mechanism also is seen in samples B, C, D, and E. On the other hand, sample F, without mullite filler, showed a different deformation behavior with relatively low strength (Table I). After elastic deformation, sample F deformed to induce delamination. No buckling was observed. The crack propagation along the direction of the layered fabric reached the end of the sample of 38 mm length. No macroscopic crack propagation occurred parallel to the direction of applied load. As compared with the deformation behavior of samples A–F, sample G, with 22 vol% SiC satin-weave fabric and 28 vol% mullite filler (Table I), fractured catastrophically with no pseudoductility. The total porosity of open and closed pores of sample G was approximately Fig. 3. Stress–displacement curve (a) for sample A and the appearance (b) of sample A at the deformation stages of I–IV indicated in Fig. 3(a). Fig. 4. Microstructures (a) before and (b) after the measurement of flexural strength of sample A. June 2004 Preparation of Si-Ti-C-O Fabric/Mullite Filler/Polytitanocarbosilane Laminates by PIP Method 999
Journal of the American Ceramic Sociery-Hirata et al. Vol 87. No. 6 3.0 mm displacement 4) No filler in the layered fabric increases the porosity before the PIP process. The increase in initial porosity decreases the densification rate of the composite ( The deformation mechanism of the densified laminated Si-Ti-C-O fabric with mullite filler is explained by the combina- tion of the following processes: elastic deformation- buckling of the Sic fabric - delamination along the direction of the layered fabric (6) The composite with a higher strength provides a higher energy of fracture. The maximum energy of fracture reaches 22 the composite with 330 MPa of strength. displacement 3.0 mm displacemen References J Kerans and T sarathy, "Crack Deflection in Ceramic Composites Strength/ MPa and Fiber Coating design Cr 24(1999) R.J. Kerans, R S. Hay, and N. J. Pagano, "The Role of the Fiber-Matrix Interface in Ceramic Composites, Fig. 5. Relationship between the energy of fracture at 0. 8 and 3.0 mm of 3A. G. Evans and F displacement for samples A-G and the flexural strength. Concepts for ttle-Matrix Composites, J. Am. Ceram. Soc., 76[5]1249-57(1993) that of sample F. However, the crack propagation behavior of Reinforced Brittle-Matrix Composites, Compos. Sci. Technol, 42, 3-24(1991) sample G was different from that of sample F. The crack formed N. J. Pagano, On the Micromechanical Failure Model in a Class of Ideal Brittle-Matrix Composites, Part 1. Coated-Fiber Composites, " Composites, Part B, in the tensile plane was deflected to the direction of applied load 29B,93-119(1998 trong bonding between the fibers and matrix caused such a "On the Micromechanical Failure Model in a Class of Ideal catastrophic failure behavior. A similar catastrophic fracture also Brittle-Matrix Composites, Part 2. Uncoated-Fiber Composites, "Composites, Part has been reported in the layered composite prepared from the 29B,121-30(1998) D. B. Marshall andA "Failure Mechanisms in Ceramic-Fiber/Ceramic- aluminosilicate fabric/mullite filler/mullite precursor polymer sys- Soc,68S225-31(1985) tem. 54 It was found from the observation of the fractured surface W. Hutchinson, "Interface Debonding and Fiber of the aluminosilicate fabric/mullite matrix composite that the low Cracking in Brittle- Matri sites,J.Am. Ceran. Soc., 72 [12] 2300-303 dispersibility of the mullite filler in the fiber yarn and the large amount of mullite precursor couple to form relatively strong Mechanical Properties on Pull Out in a SiC-Fiber-Reinforced Lithium Aluminum bonding between the fibers in the yarn, which resulted in britti fracture. The increased weave density of the present fabric P G. Charalambides and A, G, Evans, "Debonding Properties of Residually decreased the amount of mullite filler incorporated in the ya (samples E and g). The decreased amount of mullite filler may 2. Llora and R. N. Singh, "Influence of Fiber and Interfacial Properti Fracture Behavior of Fiber-Reinforced Ceramic Composites, J. Am. Ceram Soc. allow for direct contact between the fibers and result in the loss of 11]2882-90(1991) flexibility. Therefore, the nonuniform distribution of mullite filler in the fabric is responsible for the decrease in pseudoductility Table I and Fig. 5 summarize the energy of fracture at 0.8 and ites, "Composites, Part 4, 27 terfacial Failure in Ceramic Fiber/Glass Compos- 737-41(1996) 3.0 mm of displacement for samples A-G. The energy of fracture corresponds to the area surrounded by the nonlinear stress- Carbon Coatings for Oxide/Oxide Composites, " J. Am. Ceram Soc., 83 [2]329-36 displacement curve in Fig. 3. For a given displacement, the 210B. Bender, D Shad welL C Bulik. L. Inoorvati, and D. Lewis l "Efect of Fiber g. 5, the energy of fracture is plotted as a function of flexural energy of fracture pie with a higher strength provides a higher 17C. A. Lewinsohn, C. H. Henager Jr, R. H. Jones, and 3. 1. Eldridge,"Measurin ture. Especially, samples A and C, which show the Interphase Recession by Fiber Push-In Testing, J. Am. Ceram Soc., 84 14]866-68 adual increase of the deformation resistance at a larger displace- IE. Boakye, R. S. Hay, and M. D. Petry, "Continuous Coating of Oxide Fiber ment, exhibit significant energy of fracture, which reach 20 kJ/m Tows Using Liquid Precursors: Monazite Coatings on Nextel 720M,"J.Am. Ceram. at 3 mm of displacement. The surface energy of ceramic materials Soc,82192321-3l(1999 is in th from 1 to 12 nd T J Mackin, " Control of Interface fracture in the present composite is significantly high as compared Ceram Soc, 80[1232987-96(1997) with the surface energy P. E. D. Morgan and D. B. Marshall. "Functional Interfaces for Oxide/Oxide athy, E. Boakye, M. Cinibulk, and M. D. Petry, "Fabrication and IV. Conclusions 1)In the PIP process using a low-viscosity PTC solution, the monazite Monazite Multilayer Laminate,J.Am. Ceram Soc., 83[]802-808(2000) ensification of the laminated composite of the Si-Ti-C-O wover abric with mullite filler proceeds with the gradual formation of Alt P. E. D. Morgan and D. B. Marshall, "Ceramic Composites of Monazite and ncreasing number of PIP sequences 24M. G. Cain, R. L. Cain, M. H. Lewis, and J. Gent, " In Situ Reaction Rare-Earth The PIP sequence with a concentrated high-viscosity PtC 8071873-76(199 2M. K. Cinibulk and R. S. Hay, "Textured Magnetoplumbite Fiber-Matrix olution quick narrows the pore channel of the laminated Interphase Derived from SolGel Fiber Coatings,J. Am. Ceram. Soc., 79 [5] mposite, which causes difficulty in smooth flow of the PTC 1233-46(190 olution into the interior of the composite. As a result, the eN. lyi, S. Takekawa, and S. Kimura, "Crystal Chemistry of Hexaaluminates densification rate of the composite is decreased A. Parthasarathy, E, Boakye, K. A, Keller, and R s. Hi (3) Incorporation of mullite powder filler in the Si-Ti-C-0 Porous ZrO -SiO, and Monazite Coating Using Nextel M 720-Fiber-Reinforced fabric narrows the pore size in the starting stage of the PIP process Blackglass Minicomposites,J Am Ceram Soc., 84(7)1526-32(200 in addition to decreasing the pore volume. As a result, the open Langrange and E Mouch Matrix-SiC and Mullite 2D Woven Fabric Composites with or without Zirco pores in the composite with a higher amount of mullite filler are Coating Interphase: Elaboration and Properties, "J. Eur. Ceram Soc, 16 (2)301-14 more easily eliminated by the PIP process
that of sample F. However, the crack propagation behavior of sample G was different from that of sample F. The crack formed in the tensile plane was deflected to the direction of applied load. Strong bonding between the fibers and matrix caused such a catastrophic failure behavior. A similar catastrophic fracture also has been reported in the layered composite prepared from the aluminosilicate fabric/mullite filler/mullite precursor polymer system.54 It was found from the observation of the fractured surface of the aluminosilicate fabric/mullite matrix composite that the low dispersibility of the mullite filler in the fiber yarn and the large amount of mullite precursor couple to form relatively strong bonding between the fibers in the yarn, which resulted in brittle fracture. The increased weave density of the present fabric decreased the amount of mullite filler incorporated in the yarn (samples E and G). The decreased amount of mullite filler may allow for direct contact between the fibers and result in the loss of flexibility. Therefore, the nonuniform distribution of mullite filler in the fabric is responsible for the decrease in pseudoductility. Table I and Fig. 5 summarize the energy of fracture at 0.8 and 3.0 mm of displacement for samples A–G. The energy of fracture corresponds to the area surrounded by the nonlinear stress– displacement curve in Fig. 3. For a given displacement, the maintenance of high strength leads to a high energy of fracture. In Fig. 5, the energy of fracture is plotted as a function of flexural strength. The sample with a higher strength provides a higher energy of fracture. Especially, samples A and C, which show the gradual increase of the deformation resistance at a larger displacement, exhibit significant energy of fracture, which reach 20 kJ/m2 at 3 mm of displacement. The surface energy of ceramic materials is in the range from 1 to 12 J/m2 . 55 The measured energy of fracture in the present composite is significantly high as compared with the surface energy. IV. Conclusions (1) In the PIP process using a low-viscosity PTC solution, the densification of the laminated composite of the Si-Ti-C-O woven fabric with mullite filler proceeds with the gradual formation of closed pores with increasing number of PIP sequences. (2) The PIP sequence with a concentrated high-viscosity PTC solution quickly narrows the pore channel of the laminated composite, which causes difficulty in smooth flow of the PTC solution into the interior of the composite. As a result, the densification rate of the composite is decreased. (3) Incorporation of mullite powder filler in the Si-Ti-C-O fabric narrows the pore size in the starting stage of the PIP process in addition to decreasing the pore volume. As a result, the open pores in the composite with a higher amount of mullite filler are more easily eliminated by the PIP process. (4) No filler in the layered fabric increases the porosity before the PIP process. The increase in initial porosity decreases the densification rate of the composite. (5) The deformation mechanism of the densified laminated Si-Ti-C-O fabric with mullite filler is explained by the combination of the following processes: elastic deformation 3 buckling of the SiC fabric 3 delamination along the direction of the layered fabric. (6) The composite with a higher strength provides a higher energy of fracture. The maximum energy of fracture reaches 22 kJ/m2 in the composite with 330 MPa of strength. References 1 R. J. Kerans and T. A. Parthasarathy, “Crack Deflection in Ceramic Composites and Fiber Coating Design Criteria,” Composites, Part A, 30, 521–24 (1999). 2 R. J. Kerans, R. S. Hay, and N. J. Pagano, “The Role of the Fiber–Matrix Interface in Ceramic Composites,” Am. Ceram. Soc. Bull., 68 [2] 429–42 (1989). 3 A. G. Evans and F. W. Zok, “Review: The Physics and Mechanics of FiberReinforced Brittle-Matrix Composites,” J. Mater. Sci., 29, 3857–96 (1994). 4 J. B. Davis, J. P. A. Lofvander, and A. G. Evans, “Fiber Coating Concepts for Brittle-Matrix Composites,” J. Am. Ceram. Soc., 76 [5] 1249–57 (1993). 5 A. G. Evans, F. W. Zok, and J. B. Davis, “The Role of Interface in FiberReinforced Brittle-Matrix Composites,” Compos. Sci. Technol., 42, 3–24 (1991). 6 N. J. Pagano, “On the Micromechanical Failure Model in a Class of Ideal Brittle-Matrix Composites, Part 1. Coated-Fiber Composites,” Composites, Part B, 29B, 93–119 (1998). 7 N. J. Pagano, “On the Micromechanical Failure Model in a Class of Ideal Brittle-Matrix Composites, Part 2. Uncoated-Fiber Composites,” Composites, Part B, 29B, 121–30 (1998). 8 D. B. Marshall and A. G. Evans, “Failure Mechanisms in Ceramic-Fiber/CeramicMatrix Composites,” J. Am. Ceram. Soc., 68 [5] 225–31 (1985). 9 A. G. Evans, M. Y. He, and J. W. Hutchinson, “Interface Debonding and Fiber Cracking in Brittle-Matrix Composites,” J. Am. Ceram. Soc., 72 [12] 2300–303 (1989). 10M. D. Thouless, O. Sbaizero, L. S. Sigl, and A. G. Evans, “Effect of Interface Mechanical Properties on Pull Out in a SiC-Fiber-Reinforced Lithium Aluminum Silicate Glass-Ceramic,” J. Am. Ceram. Soc., 72 [4] 525–32 (1989). 11P. G. Charalambides and A. G. Evans, “Debonding Properties of Residually Stressed Brittle-Matrix Composite,” J. Am. Ceram. Soc., 72 [5] 746–53 (1989). 12J. Llora and R. N. Singh, “Influence of Fiber and Interfacial Properties on Fracture Behavior of Fiber-Reinforced Ceramic Composites,” J. Am. Ceram. Soc., 74 [11] 2882–90 (1991). 13K. K. Chawla, Z. R. Xu, and J.-S. Ha, “Processing, Structure, and Properties of Mullite Fiber/Mullite Matrix Composites,” J. Eur. Ceram. Soc., 16, 293–99 (1996). 14R. J. Yong and X. Yang, “Interfacial Failure in Ceramic Fiber/Glass Composites,” Composites, Part A, 27A, 737–41 (1996). 15K. A. Keller, T. Mah, T. A. Parthasarathy, and C. M. Cooke, “Fugitive Interfacial Carbon Coatings for Oxide/Oxide Composites,” J. Am. Ceram. Soc., 83 [2] 329–36 (2000). 16B. Bender, D. Shadwell, C. Bulik, L. Incorvati, and D. Lewis III, “Effect of Fiber Coatings and Composite Processings on Properties of Zirconia-Based Matrix–SiC Fiber Composites,” Am. Ceram. Soc. Bull., 65 [2] 363–69 (1986). 17C. A. Lewinsohn, C. H. Henager Jr., R. H. Jones, and J. I. Eldridge, “Measuring Interphase Recession by Fiber Push-In Testing,” J. Am. Ceram. Soc., 84 [4] 866–68 (2001). 18E. Boakye, R. S. Hay, and M. D. Petry, “Continuous Coating of Oxide Fiber Tows Using Liquid Precursors: Monazite Coatings on Nextel 720TM,” J. Am. Ceram. Soc., 82 [9] 2321–31 (1999). 19D.-H. Kuo, W. M. Kriven, and T. J. Mackin, “Control of Interfacial Properties through Fiber Coating Monazite Coatings in Oxide/Oxide Composites,” J. Am. Ceram. Soc., 80 [12] 2987–96 (1997). 20P. E. D. Morgan and D. B. Marshall, “Functional Interfaces for Oxide/Oxide Composites,” Mater. Sci. Eng., A, 162, 15–25 (1993). 21T. A. Parthasarathy, E. Boakye, M. Cinibulk, and M. D. Petry, “Fabrication and Testing of Oxide/Oxide Microcomposites with Monazite and Hibonite as Interlayers,” J. Am. Ceram. Soc., 82 [12] 3575–83 (1999). 22T. R. Mawdsley, D. Kovar, and J. W. Halloran, “Fracture Behavior of Alumina/ Monazite Multilayer Laminate,” J. Am. Ceram. Soc., 83 [4] 802–808 (2000). 23P. E. D. Morgan and D. B. Marshall, “Ceramic Composites of Monazite and Alumina,” J. Am. Ceram. Soc., 78 [6] 1553–63 (1995). 24M. G. Cain, R. L. Cain, M. H. Lewis, and J. Gent, “In Situ Reaction Rare-Earth Hexaaluminate Interphases,” J. Am. Ceram. Soc., 80 [7] 1873–76 (1997). 25M. K. Cinibulk and R. S. Hay, “Textured Magnetoplumbite Fiber–Matrix Interphase Derived from Sol–Gel Fiber Coatings,” J. Am. Ceram. Soc., 79 [5] 1233–46 (1996). 26N. Iyi, S. Takekawa, and S. Kimura, “Crystal Chemistry of Hexaaluminates: -Alumina and Magnetoplumbite Structure,” J. Solid State Chem., 83, 8–9 (1989). 27T. A. Parthasarathy, E. Boakye, K. A. Keller, and R. S. Hay, “Evaluation of Porous ZrO2–SiO2 and Monazite Coating Using NextelTM 720-Fiber-Reinforced BlackglassTM Minicomposites,” J. Am. Ceram. Soc., 84 [7] 1526–32 (2001). 28P. Colomban, E. Bruneton, J. L. Langrange, and E. Mouchon, “Sol–Gel Mullite Matrix–SiC and Mullite 2D Woven Fabric Composites with or without Zirconia Coating Interphase: Elaboration and Properties,” J. Eur. Ceram. Soc., 16 [2] 301–14 (1996). Fig. 5. Relationship between the energy of fracture at 0.8 and 3.0 mm of displacement for samples A–G and the flexural strength. 1000 Journal of the American Ceramic Society—Hirata et al. Vol. 87, No. 6
June 2004 Preparation of Si-Ti-C-o Fabric/Mullite Filler Polytitanocarbosilane Laminates by PlP Method 290Sudre, A G Razzell, L. Molliex, and M. Holmquist, "Alumina Single-Crystal 42D. Holz, S. Pagel, C. Bowen, S. Wu, and N. Claussen, "Fabrication of Fiber-Reinforced Alumina-Matrix for Com busto Low-to-Zero Shrinkage Reaction-Bonded Matrix Composites, "J. Eur. Ceram. Soc. 4]273-80(19989 1621255- R.J. Kerans and T. A. Parthasarathy, Theoretical Analysis of the Fiber Pullout shita, M. Matsuda, and M. Shibuya, "Observation of and Pushout Tests, Am. Ceram Soc., 74 [7 1585-96(1991 3P. DJero,R J Kerans, and T, A. Parthasarathy, "Effect of Interfacial Roughness Electron Microscopy,J. Ceram Soc. Jpn, 103 [11 1197-200 by Analytical Transmission on the Frictional Stress Measured Using Pushout Tests, "J. Am. Ceram. Soc., 74 [11] Y. Hirata, R. Yamashita, and M. Shibuya, "Microstructure and Crack Extension of Laminated Si-Ti-C-O Fiber/Mullite Matrix Composites, Mater. Chem Plrys, 69, R..Kerans,"The Role of Coating Compliance and Fiber/Matrix Interfacial 2525. Topography on Debonding in Ceramic Composites, Scr. Metall., 32 [4] 505-509 M. Schmucker, F. Flucht, and H. Shneider, " High-Temperature Behavior of Polycrystalline Aluminosilicate Fibers with Mullite Bulk Composition. L. Microstruc- 3R. J. Kerans, "Viability of Oxide Fiber Coatings in Ceramic Composites for Accommodation of Misfit Stresses"J. Anm Ce Soc,79同61664-68(199%6 R. Dong and Y. Hirata,"Processing of Ceria-Coated Alumina Fiber/Alumina 3T. A. Parthasarathy, D. R. Barlage, P. D. Jero, and R. J. Kerans, "Effect of 47Y. Kamino, Y. Hirata, and S Sameshima, "Preparation and Mechanical Proper- Composite,J.Am. Ceram. Soc., 77[12] 3232-36(1994) ties of Long Alumina Fiber/Alumina Matrix Composites, Mater. Letf, 28, 161-64 x. Li, Y, Matsuo and S. Kimura, "The Role of Fiber Coating on SiC-Fiber- R. Dong, H, Sueyoshi, M, Higo, Y. Uemura, and Y. Hirata,"Proc Reinforced Alumina Composite, " J Ceram Soc. Jpn, 100(4]499-503(19 Fiber/Alumina-Mullite Matrix Composite, J. Ceram. Process. Res. sC. Droillard and J. Lamon. "Fracture Toughness of 2-D-Woven SiC/Sic CVI-Composites with Multilayered Interphases,J.Am. Ceram Soc., 79[4]849-58 Y Nakamura, S Kasuga, J. Nakamura, and Y Hirata, "Mechanica 7T. K. Jacobsen and P. Brondsted, "Mechanical Properties of Two Plain-Woven in Higi-Temperature atrir Composites. Edited by w. Krenkel Chemical Vapor Infiltrated Silicon Carbide-Matrix Composites, J. Am. Ceram. Soc. and H. Shneider. Wiley-VCH, New York, 20 A. Aksay and J. A. Pask,"Stable and Metastable Equilibria in the System K. Hayata, "Infiltration an sis of Polytitanc- 41OAm, Ceram. Soc,581=12150712(195 carbosilane in a Si-Ti-C-O Fabric/Mullite Porous Composite, "J.Am. Ceram Soc., 83 15]1044-48(2000 and Optical Applications, J An. Ceram Soc., 74 [10]2343-58(1991 3"T. Ogasawara, T. Ishikawa, H. Ito, N. Watanabe, and L.J. Davies,"Multiple $2S. Somiya and Y, Hirata, "Mullite Powder Technology and Application in Cracking and Tensile Behavior for an Orthogonal 3-D Woven Si-Ti-C-O Fiber/Si- Takeshima, ""Effects of Particle Classification on Colloidal Ti-C-O Matrix Composite, " J. Am. Ceram Soc., 84[7] 1565-74(2001). .Hmmk08012( Processing of Mullite, Mater. Left, 16, 169-74(1993). mperature Dependence of Tensile Strength for a S4R. Dong, Y, Hirata, H. Sueyoshi, M. Higo, and Y. Uemura, "Polymer Impreg- ven Boron Nitride-Coated Hi-NicalonM SiC-Fiber-Reinforced Silicon Carbide nation and Pyrolysis(PIP) Method for the Preparation of Laminated Woven Matrix Composite, "J. Am. Ceram Soc, 74 [711585-96(1991). G. Morscher P. Pirouz, and A. H. Heuer,"Temperature Dependence of Fabric/Mullite Matrix Composites with Pseudoductility,J. Eur. Ceram. Soc., 24[11 Interfacial Shear Strength in SiC-Fiber-Reinforced Reaction-Bonded Silicon Nitride B. Lawn, Fracture of Brittle Solids, 2nd Ed, Cambridge Solid State Scie J. Am. Ceran. Soc., 73 [3]713-20(1990). Series, pp 51-56. Cambridge University Press, Cambridge, U.K., 1993
29O. Sudre, A. G. Razzell, L. Molliex, and M. Holmquist, “Alumina Single-Crystal Fiber-Reinforced Alumina-Matrix for Combustor Tiles,” Ceram. Eng. Sci. Proc., 19 [4] 273–80 (1998). 30R. J. Kerans and T. A. Parthasarathy, “Theoretical Analysis of the Fiber Pullout and Pushout Tests,” J. Am. Ceram. Soc., 74 [7] 1585–96 (1991). 31P. D. Jero, R. J. Kerans, and T. A. Parthasarathy, “Effect of Interfacial Roughness on the Frictional Stress Measured Using Pushout Tests,” J. Am. Ceram. Soc., 74 [11] 2793–801 (1991). 32R. J. Kerans, “The Role of Coating Compliance and Fiber/Matrix Interfacial Topography on Debonding in Ceramic Composites,” Scr. Metall., 32 [4] 505–509 (1995). 33R. J. Kerans, “Viability of Oxide Fiber Coatings in Ceramic Composites for Accommodation of Misfit Stresses,” J. Am. Ceram. Soc., 79 [6] 1664–68 (1996). 34T. A. Parthasarathy, D. R. Barlage, P. D. Jero, and R. J. Kerans, “Effect of Interfacial Roughness Parameters on the Fiber Push Out Behavior of Model Composite,” J. Am. Ceram. Soc., 77 [12] 3232–36 (1994). 35J. X. Li, Y. Matsuo and S. Kimura, “The Role of Fiber Coating on SiC-FiberReinforced Alumina Composite,” J. Ceram. Soc. Jpn., 100 [4] 499–503 (1992). 36C. Droillard and J. Lamon, “Fracture Toughness of 2-D-Woven SiC/SiC CVI-Composites with Multilayered Interphases,” J. Am. Ceram. Soc., 79 [4] 849–58 (1996). 37T. K. Jacobsen and P. Brondsted, “Mechanical Properties of Two Plain-Woven Chemical Vapor Infiltrated Silicon Carbide-Matrix Composites,” J. Am. Ceram. Soc., 84 [5] 1043–51 (2001). 38Y. Hirata, T. Matsura, and K. Hayata, “Infiltration and Pyrolysis of Polytitanocarbosilane in a Si-Ti-C-O Fabric/Mullite Porous Composite,” J. Am. Ceram. Soc., 83 [5] 1044–48 (2000). 39T. Ogasawara, T. Ishikawa, H. Ito, N. Watanabe, and I. J. Davies, “Multiple Cracking and Tensile Behavior for an Orthogonal 3-D Woven Si-Ti-C-O Fiber/SiTi-C-O Matrix Composite,” J. Am. Ceram. Soc., 84 [7] 1565–74 (2001). 40S. Guo and Y. Kagawa, “Temperature Dependence of Tensile Strength for a Woven Boron Nitride-Coated Hi-NicalonTM SiC-Fiber-Reinforced Silicon CarbideMatrix Composite,” J. Am. Ceram. Soc., 74 [7] 1585–96 (1991). 41G. Morscher, P. Pirouz, and A. H. Heuer, “Temperature Dependence of Interfacial Shear Strength in SiC-Fiber-Reinforced Reaction-Bonded Silicon Nitride,” J. Am. Ceram. Soc., 73 [3] 713–20 (1990). 42D. Holz, S. Pagel, C. Bowen, S. Wu, and N. Claussen, “Fabrication of Low-to-Zero Shrinkage Reaction-Bonded Matrix Composites,” J. Eur. Ceram. Soc., 16 [2] 255–60 (1996). 43Y. Hirata, R. Yamashita, M. Matsuda, and M. Shibuya, “Observation of Interfaces of Si-Ti-C-O Fiber/Mullite Matrix Composites by Analytical Transmission Electron Microscopy,” J. Ceram. Soc. Jpn., 103 [11] 1197–200 (1995). 44Y. Hirata, R. Yamashita, and M. Shibuya, “Microstructure and Crack Extension of Laminated Si-Ti-C-O Fiber/Mullite Matrix Composites,” Mater. Chem. Phys., 69, 252–60 (2001). 45M. Schmu¨cker, F. Flucht, and H. Shneider, “High-Temperature Behavior of Polycrystalline Aluminosilicate Fibers with Mullite Bulk Composition. I. Microstructure and Strength Properties,” J. Eur. Ceram. Soc., 16 [2] 281–85 (1996). 46R. Dong and Y. Hirata, “Processing of Ceria-Coated Alumina Fiber/Alumina Matrix Composite,” J. Ceram. Soc. Jpn., 108 [9] 823–29 (2000). 47Y. Kamino, Y. Hirata, and S. Sameshima, “Preparation and Mechanical Properties of Long Alumina Fiber/Alumina Matrix Composites,” Mater. Lett., 28, 161–64 (1996). 48R. Dong, H. Sueyoshi, M. Higo, Y. Uemura, and Y. Hirata, “Processing of SiC Fiber/Alumina–Mullite Matrix Composite,” J. Ceram. Process. Res., 2 [2] 72–78 (2001). 49Y. Nakamura, S. Kasuga, J. Nakamura, and Y. Hirata, “Mechanical Properties of Si-Ti-C-O Fabric/Polytitanocarbosilane/Mullite Laminated Composites”; pp. 721–27 in High-Temperature Ceramic-Matrix Composites. Edited by W. Krenkel, R. Naslain, and H. Shneider. Wiley-VCH, New York, 2001. 50I. A. Aksay and J. A. Pask, “Stable and Metastable Equilibria in the System SiO2–Al2O3,” J. Am. Ceram. Soc., 58 [11–12] 507–12 (1975). 51I. A. Aksay, D. M. Dabbs, and M. Sarikaya, “Mullite for Structural, Electronic, and Optical Applications,” J. Am. Ceram. Soc., 74 [10] 2343–58 (1991). 52S. Somiya and Y. Hirata, “Mullite Powder Technology and Application in Japan,” Am. Ceram. Soc. Bull., 70, 1624–32 (1991). 53Y. Hirata and K. Takeshima, “Effects of Particle Classification on Colloidal Processing of Mullite,” Mater. Lett., 16, 169–74 (1993). 54R. Dong, Y. Hirata, H. Sueyoshi, M. Higo, and Y. Uemura, “Polymer Impregnation and Pyrolysis (PIP) Method for the Preparation of Laminated Woven Fabric/Mullite Matrix Composites with Pseudoductility,” J. Eur. Ceram. Soc., 24 [1] 53–64 (2004). 55B. Lawn, Fracture of Brittle Solids, 2nd Ed., Cambridge Solid State Science Series; pp. 51–56. Cambridge University Press, Cambridge, U.K., 1993. June 2004 Preparation of Si-Ti-C-O Fabric/Mullite Filler/Polytitanocarbosilane Laminates by PIP Method 1001