COMPOSITES SCIENCE AND TECHNOLOGY ELSEⅤIER Composites Science and Technology 62(2002)2179-2188 www.elsevier.com/locate/compscitech Consolidation of polymer-derived Sic matrix composites processing and microstructure Masaki Kotania,*. I, Takahiro Inoue, Akira Kohyama, Kiyohito Okamura Yutai Katoh Graduate School of Energy Science, Kyoto Univ Yoshidahonmachi, Sakyo-ku, Kyoto 606-8501, Japan bNational Institute of Advanced Industrial Science and Technology, 1-2-1 Namiki, Tsukuba, Ibaraki, 305-8942 Jape Institute of Advanced Energy, Kyoto University, Gokasho, Uji, Kyoto, 611-0011 Japan dOsaka prefecture University, 1-1 Gakuen-cho, Sakai, Osaka, 599-8531 Japan Received 19 October 2001: received in revised form 26 July 2002: accepted 27 July 2002 Abstract SiC fiber reinforced Sic matrix composite(SiC/SiC composite) has been developed by polymer impregnation and pyrolysis (PIP) method, which consists of impregnation, curing, consolidation, and re-impregnation and pyrolysis. As a prospective approach to fabricate a high performance composite, consolidation conditions, such as curing temperature to make a green body, pressure and heating rate during consolidation, were systematically controlled for effective consolidation. Because of its advantage in controlling hysical characteristic, polyvinylsilane(pvs) that is liquid thermosetting organo-silicic compound was utilized as the matrix pre- cursor. Based on the pyrolytic behavior of PVs, effects of the process conditions on microstructure of the consolidated bodies were accurately characterized. To relate those microstructure with mechanical property, flexural tests were performed for the composites fter multiple PIP processing. Consequently, process conditions to make a high performance composite could be appeared. Struc- C 2002 Elsevier Science Ltd. All rights reserved. er tural conditions to be optimized for further improvement in mechanical and environmental properties were also discussed Keywords: A Ceramic-matrix composites(CMCs): A. Preceramic polymer; B Curing: B Porosity: B Mechanical properties 1. Introduct shape and geometry, microstructural control, and cost Since silicon carbide possesses such superior proper Main challenge of this process has been made to ties as strength at elevated temperature, oxidation resis- reduce pores and cracks which were formed due to gas tance and microstructural stability under irradiation, evolution and volumetric shrinkage of a preceramic there have been many efforts on r& d of Sic/Sic polymer during pyrolysis [8-11]. Gas evolution causes composite for use in aerospace vehicles and fusion the inhomogenization of matrix. Volumetric shrinkage power reactor [1-5]. Among potential fabrication pro- directly gives rise to the formation of pores. After a cesses of ceramics matrix composites (CMCs), PIP polymer is hardened, both events lead to crack initia- method is one of most promising methods because of its tion. At that time, fiber distribution is determined advantages in the viewpoints of impregnation efficiency Although repetition of PIP processing is much useful to among fibers, large-scale fabrication with complicated fill such defects with polymer-pyrolyzed product microstructure produced in first PIP processing would be influential on final microstructure. Therefore, effec- Corresponding author at present address. Tel: +81-298-68-2336: tive consolidation to yield minimum amount of crack initiation and appropriate fiber distribution are very and Development, National Space Development Agency e Present address: Thermal Engineering Group, Office important technical issue a high per mance SiC/Sic composite Tsukuba Space Center 2-1-1 Sengen, Tsukuba, Ibaraki, n order to control the distribution of fibers and pores, utilization of high volumetric yield polymer, fille 0266-3538/02/S. see front matter C 2002 Elsevier Science Ltd. All rights reserved. PII:S0266-3538(02)00151-3
Consolidation of polymer-derived SiC matrix composites: processing and microstructure Masaki Kotania,*,1, Takahiro Inoueb, Akira Kohyamac , Kiyohito Okamurad, Yutai Katohc a Graduate School of Energy Science, Kyoto University, Yoshidahonmachi, Sakyo-ku, Kyoto 606-8501, Japan bNational Institute of Advanced Industrial Science and Technology, 1-2-1 Namiki, Tsukuba, Ibaraki, 305-8942 Japan c Institute of Advanced Energy, Kyoto University, Gokasho, Uji, Kyoto, 611-0011 Japan dOsaka Prefecture University, 1-1 Gakuen-cho, Sakai, Osaka, 599-8531 Japan Received 19 October 2001; received in revised form 26 July 2002; accepted 27 July 2002 Abstract SiC fiber reinforced SiC matrix composite (SiC/SiC composite) has been developed by polymer impregnation and pyrolysis (PIP) method, which consists of impregnation, curing, consolidation, and re-impregnation and pyrolysis. As a prospective approach to fabricate a high performance composite, consolidation conditions, such as curing temperature to make a green body, pressure and heating rate during consolidation, were systematically controlled for effective consolidation. Because of its advantage in controlling physical characteristic, polyvinylsilane (PVS) that is liquid thermosetting organo-silicic compound was utilized as the matrix precursor. Based on the pyrolytic behavior of PVS, effects of the process conditions on microstructure of the consolidated bodies were accurately characterized. To relate those microstructure with mechanical property, flexural tests were performed for the composites after multiple PIP processing. Consequently, process conditions to make a high performance composite could be appeared. Structural conditions to be optimized for further improvement in mechanical and environmental properties were also discussed. # 2002 Elsevier Science Ltd. All rights reserved. Keywords: A. Ceramic-matrix composites (CMCs); A. Preceramic polymer; B. Curing; B. Porosity; B. Mechanical properties 1. Introduction Since silicon carbide possesses such superior properties as strength at elevated temperature, oxidation resistance and microstructural stability under irradiation, there have been many efforts on R & D of SiC/SiC composite for use in aerospace vehicles and fusion power reactor [1–5]. Among potential fabrication processes of ceramics matrix composites (CMCs), PIP method is one of most promising methods because of its advantages in the viewpoints of impregnation efficiency among fibers, large-scale fabrication with complicated shape and geometry, microstructural control, and cost [6,7]. Main challenge of this process has been made to reduce pores and cracks which were formed due to gas evolution and volumetric shrinkage of a preceramic polymer during pyrolysis [8–11]. Gas evolution causes the inhomogenization of matrix. Volumetric shrinkage directly gives rise to the formation of pores. After a polymer is hardened, both events lead to crack initiation. At that time, fiber distribution is determined. Although repetition of PIP processing is much useful to fill such defects with polymer-pyrolyzed product, microstructure produced in first PIP processing would be influential on final microstructure. Therefore, effective consolidation to yield minimum amount of crack initiation and appropriate fiber distribution are very important technical issue for making a high performance SiC/SiC composite. In order to control the distribution of fibers and pores, utilization of high volumetric yield polymer, filler 0266-3538/02/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved. PII: S0266-3538(02)00151-3 Composites Science and Technology 62 (2002) 2179–2188 www.elsevier.com/locate/compscitech * Corresponding author at present address. Tel.: +81-298-68-2336; fax: +81-298-68-2968. E-mail address: kotani.masaki@nasda.go.jp (M. Kotani). 1 Present address: Thermal Engineering Group, Office of Research and Development, National Space Development Agency of Japan, Tsukuba Space Center 2-1-1 Sengen, Tsukuba, Ibaraki, 305-8505 Japan
M. Kotani et al. /Composites Science and Technology 62(2002)2179-2188 addition to the polymer, and pressurization during -SiHy-CHy-CHy-and-C(SiH3)H-CHy-, with the ratio consolidation is expected to be much beneficial. How- of 1: 1 [21]. It was synthesized by radical polymerization ever, even with an advanced polymer of high ceramic of vinylsilane(CH2=CH-SiH3)in an autoclave Density yield of more than 80 wt % large volumetric shrinkage and molecular weight of the polymer were 0.91 Mg m-3, unavoidably occurs due to big change of density from and 957(number average )/2780(weight average)(Mw/ polymer to ceramic [12-14]. Filler loading into polymer Mn=2.9), respectively. It is transparent liquid with the can improve apparent volume shrinkage, and also dis viscosity of about 70 cP at room temperature. Pyrolysis perse pores finely [15, 16]. In addition, filler particles chemistry during polymer-to-ceramic conversion has could replenish pores if those are active to produce car- been studied so far [22] bide [17-19]. Contrary to this, it might decline impreg- To take advantage of PVSs superior characteristics, nation efficiency and weaken a pyrolyzed product. its pyrolytic behaviors such as thermosetting, shrinkage Pressurization is expected to fix a fibrous preform andand gas evolution were precisely inspected prior to eliminate pores outside, as far as the polymer retains composite fabrication. Thermogravimetric and differ fluidity. Even after a polymer is hardened, it might be ential thermal analysis (TG-DTA)was performed under beneficial for reducing crack initiation formed due to the heating rate of 300 K/h and the flowing rate of Ar of internal gas pressure. As another option to reduce pores 0. 2 1 /min As the reference material, high purity alumina and cracks in a consolidated body, curing prior to powder was utilized. DTA curve was determined by pressurization is prospective, because the volumetric deducting a blank from the original analytical data fraction of a final pyrolyzed product out of a polymeric Differently from this experiment, TGa were performed precursor is increased at various heating rates in flowing Ar(1 I/ min). Mor- In this work, process development for high perfor- phological analysis was performed for isolated inter mance polymer-derived SiC/Sic composite was per- mediates of PVS in room temperature. Densities of the formed. As the matrix precursor, PVS, which was intermediates were measured by picnometry using dis liquid polycarbosilane with a lot of functional Si-H tilled water. For these experiments, the samples were bonds, was applied because of its advantages of suffi- prepared by heating the polymer in same conditions as cient stability at ambient temperature, low viscosity, curing(400 K). According to those data, volumetric and continuous thermosetting behavior. Rheological residues at various temperatures were estimated with the properties such as viscosity and wettability are much following equation important characteristic in filler dispersion, impregna tion into small area among fibers. PVS and its slurry (a1473/or) UT= with Sic particles appeared to be impregnated very well into a continuous SiC fiber preform without dilution Recent work also demonstrated its superior rheological property in fiber production, namely finer SiC fiber was where v is volumetric residue(%), o mass residue (%) successfully synthesized by blending PVS with conven- and p density(Mg m-3). Subscripts of the characters tional polycarbosilane [20]. Owing to continuous ther- correspond to pyrolyzing temperature(K). Total quan mosetting behavior during pyrolysis, its physical tity of evolved gas was monitored as a function of tem- characteristics could be accurately controlled by heat perature at intervals of 100 K. The sample was heated in treatment. To make a composite of high density and a closed silica tube of already known volume at 300 K/h uniform fiber distribution, main efforts was paid for in vacuum. The quantities were estimated from the optimizing consolidation conditions; such as curing change of gas pressure temperature to prepare a green body, pressure and In the fabrication of composite, Hi-Nicalon"M,which heating rate to make a consolidated body. In con- is continuous SiC-C) fiber produced by Nippon Car sequence, the effects of the process parameters on den- bon Co, Ltd (Japan), was employed as the reinforce- sity and microstructure were clearly revealed. And those ment SiC particles of mean particle size of 0. 27 um were were discussed on the basis of pyrolytic behavior of the utilized as the filler material. It was commercially man polymer. The relationship between microstructure and factured as ultra fine grade of Betarundum by IB mechanical properties of the composites was character- DEN Co, Ltd ( Japan). All samples were unidirectional ized by flexural test composites. Those were fabricated in the following procedures composed of four steps 2. Experimental procedure 1. To prepare a unidirectional fibrous preform, the fiber tow was uniformly wound in size of 40x20 The polymer used as the matrix precursor was poly- mm. Then it was heated up to 873 K in vacuum vinylsilane, which is developed by Mitsui chemical, inc. for removing sizing agent that quite a little (Japan). It was composed of two kinds of unit structure applied by the manufacturer
addition to the polymer, and pressurization during consolidation is expected to be much beneficial. However, even with an advanced polymer of high ceramic yield of more than 80 wt.%, large volumetric shrinkage unavoidably occurs due to big change of density from polymer to ceramic [12–14]. Filler loading into polymer can improve apparent volume shrinkage, and also disperse pores finely [15,16]. In addition, filler particles could replenish pores if those are active to produce carbide [17–19]. Contrary to this, it might decline impregnation efficiency and weaken a pyrolyzed product. Pressurization is expected to fix a fibrous preform and eliminate pores outside, as far as the polymer retains fluidity. Even after a polymer is hardened, it might be beneficial for reducing crack initiation formed due to internal gas pressure. As another option to reduce pores and cracks in a consolidated body, curing prior to pressurization is prospective, because the volumetric fraction of a final pyrolyzed product out of a polymeric precursor is increased. In this work, process development for high performance polymer-derived SiC/SiC composite was performed. As the matrix precursor, PVS, which was a liquid polycarbosilane with a lot of functional Si–H bonds, was applied because of its advantages of suffi- cient stability at ambient temperature, low viscosity, and continuous thermosetting behavior. Rheological properties such as viscosity and wettability are much important characteristic in filler dispersion, impregnation into small area among fibers. PVS and its slurry with SiC particles appeared to be impregnated very well into a continuous SiC fiber preform without dilution. Recent work also demonstrated its superior rheological property in fiber production, namely finer SiC fiber was successfully synthesized by blending PVS with conventional polycarbosilane [20]. Owing to continuous thermosetting behavior during pyrolysis, its physical characteristics could be accurately controlled by heat treatment. To make a composite of high density and uniform fiber distribution, main efforts was paid for optimizing consolidation conditions; such as curing temperature to prepare a green body, pressure and heating rate to make a consolidated body. In consequence, the effects of the process parameters on density and microstructure were clearly revealed. And those were discussed on the basis of pyrolytic behavior of the polymer. The relationship between microstructure and mechanical properties of the composites was characterized by flexural test. 2. Experimental procedure The polymer used as the matrix precursor was polyvinylsilane, which is developed by Mitsui chemical, inc. (Japan). It was composed of two kinds of unit structures, –SiH2–CH2–CH2– and –C(SiH3)H–CH2–, with the ratio of 1:1 [21]. It was synthesized by radical polymerization of vinylsilane (CH2¼CH–SiH3) in an autoclave. Density and molecular weight of the polymer were 0.91 Mg m3 , and 957 (number average)/2780 (weight average) (Mw/ Mn=2.9), respectively. It is transparent liquid with the viscosity of about 70 cP at room temperature. Pyrolysis chemistry during polymer-to-ceramic conversion has been studied so far [22]. To take advantage of PVS’s superior characteristics, its pyrolytic behaviors such as thermosetting, shrinkage and gas evolution were precisely inspected prior to composite fabrication. Thermogravimetric and differential thermal analysis (TG-DTA) was performed under the heating rate of 300 K/h and the flowing rate of Ar of 0.2 l/min. As the reference material, high purity alumina powder was utilized. DTA curve was determined by deducting a blank from the original analytical data. Differently from this experiment, TGA were performed at various heating rates in flowing Ar (1 l/min). Morphological analysis was performed for isolated intermediates of PVS in room temperature. Densities of the intermediates were measured by picnometry using distilled water. For these experiments, the samples were prepared by heating the polymer in same conditions as curing (400 K). According to those data, volumetric residues at various temperatures were estimated with the following equation. T ¼ ð Þ !1473=!T 1473 T ð1Þ where is volumetric residue (%), ! mass residue (%), and density (Mg m3 ). Subscripts of the characters correspond to pyrolyzing temperature (K). Total quantity of evolved gas was monitored as a function of temperature at intervals of 100 K. The sample was heated in a closed silica tube of already known volume at 300 K/h in vacuum. The quantities were estimated from the change of gas pressure. In the fabrication of composite, Hi-NicalonTM, which is continuous SiC–(C) fiber produced by Nippon Carbon Co., Ltd. (Japan), was employed as the reinforcement. SiC particles of mean particle size of 0.27 um were utilized as the filler material. It was commercially manufactured as ultra fine grade of BetarundumTM by IBIDEN Co., Ltd. (Japan). All samples were unidirectional composites. Those were fabricated in the following procedures composed of four steps. 1. To prepare a unidirectional fibrous preform, the fiber tow was uniformly wound in size of 4020 mm. Then it was heated up to 873 K in vacuum for removing sizing agent that quite a little applied by the manufacturer, 2180 M. Kotani et al. / Composites Science and Technology 62 (2002) 2179–2188
M. Kotani et al. /Composites Science and Technology 62(2002)2179-2188 2181 make a cured sheet, the preform was dippe 3. Results and discussion into PVS's slurry at the filler content of 25 wt% in an ambient environment, and then heated up 3. 1. Pyrolytic behavior of the polymer to curing temperature at 300 K/h in Ar, To make a consolidated body, the cured sheets Fig. I shows TG-DTA curves of PVs at 300 K/h in were stacked, and then heated up to 1473 K at Ar. Mass degradation continuously occurred from 380 300 K/h in Ar under unidirectional pressure, to 800 K. It was accelerated along with temperature and 4. To make a composite, the consolidated body was became highest between 650 and 700 K. Most of the subjected to six times of re-impregnation and mass change had finished below 700 K. Mass yield of a pyrolysis(subsequent PIP processing) of PVs pyrolyzed product up to 1473 K was 32.6%. As for without pressurization DTA curve, a downward tendency was seen from 400 K. It should be related with the mass degradation due to In all heat treatments, target temperature was kept fo the emission of polymer components. Between 500 and 10 min. Although the composites were fabricated at 600 K, there could be identified a continuous depres almost same amount of fiber, those thickness differed in sion. Based on previous reports [23, 24], it might be the range between 1. 5 and 2.5 mm, depending on the related with cross-linking reaction. Big endothermic consolidation conditions. Thus, fiber volume fraction of peak at 700 K implied a drastic change of the molecular the samples was dependent on consolidation condition. structure to form Si-C backbone, where the fragmenta- Process optimizations were performed by evaluating the tion of polymer structure simultaneously occurred [25] consolidation bodies. Sample I Ds were set to reflect the Appearances of isolated samples of PVs heated up to process conditions (curing temperature/K, pressure various temperatures were exhibited in Table 1. On the applied during consolidation/MPa). whole, the polymer showed continuous thermosetting Densimetry for the samples was performed by Archi- from transparent liquid to porous brownish glassy solid nedean method after every PIP processing. Relative between 600 and 700 K. The polymer pyrolyzed up to density(dg) was defined as ceramic volume fraction in a 583 K was not so much different from original one other bulk, calculated with the following equation than a slight increase of viscosity. Heated up to 603 K gelation was recognized. Referred to the TGa curve, dR a-sor)100%) (2) mass degradation had already reached more than 20 here. Then, the polymer continuously thermoset with further mass degradation and gradual coloration. Below where as is weight of a specimen measured in water, oc 673 K, the pyrolyzed products were free of pore. But, and op weights of a specimen with without water in all pyrolyzed above 693 K, the polymer became brownish open pores measured in atmosphere, and pH, o)r density and quite insoluble in solvents, and frothed vigorously of water at the temperature of T. dR could be estimated These features suggested that the polymer evolved much only for the samples after consolidation, because closed gas and subsequently lost almost its plasticity at this pore might be formed in subsequent PIP processing. temperature Apparent density (dA) was defined as average density of Fig. 2 shows densities and volumetric residues of PVs all constituents, expressed in the following equation as a function of temperature. Although it is not easy to da= P(H2O)7(Mg m-) It depends on the ratio of fiber and matrix in a con- solidated body. According to densities of the fiber(2.74 Mg- m-3)and the matrix theoretically derived from the slurry(2.92 Mgm-3), the density was proportional to matrix content in all constituents, microstructural characterizations were performed after consolidation nd subsequent PIP processing, using optical micro- scope(OM)and scanning electron microscope(SEM) Three-point flexural test was performed at room tem- perature. Dimensions of test specimens were 30 mm length x 4 mm width x I mm height Span and crosshead 20 40060080010001200 speed were 25 mm and 0.5 mm/min, respectively. UIti mate flexural strength (u) and work-of-fracture Temperature /K W..F)were obtained from the peak load and the area Fig. 1. Thermogravimetric and differential thermal analysis(TG of load-crosshead displacement chart. DTA)curves for PVS at a heating rate of 300 K/h in Ar
2. To make a cured sheet, the preform was dipped into PVS’s slurry at the filler content of 25 wt.% in an ambient environment, and then heated up to curing temperature at 300 K/h in Ar, 3. To make a consolidated body, the cured sheets were stacked, and then heated up to 1473 K at 300 K/h in Ar under unidirectional pressure, 4. To make a composite, the consolidated body was subjected to six times of re-impregnation and pyrolysis (subsequent PIP processing) of PVS without pressurization. In all heat treatments, target temperature was kept for 10 min. Although the composites were fabricated at almost same amount of fiber, those thickness differed in the range between 1.5 and 2.5 mm, depending on the consolidation conditions. Thus, fiber volume fraction of the samples was dependent on consolidation condition. Process optimizations were performed by evaluating the consolidation bodies. Sample IDs were set to reflect the process conditions (curing temperature/K, pressure applied during consolidation/MPa). Densimetry for the samples was performed by Archimedean method after every PIP processing. Relative density (dR) was defined as ceramic volume fraction in a bulk, calculated with the following equation. dR ¼ 1 !D !C !S ðH2OÞT 100ð%Þ ð2Þ where !S is weight of a specimen measured in water, !C and !Dweights of a specimen with/without water in all open pores measured in atmosphere, and ðH2OÞT density of water at the temperature of T. dR could be estimated only for the samples after consolidation, because closed pore might be formed in subsequent PIP processing. Apparent density (dA) was defined as average density of all constituents, expressed in the following equation. dA ¼ !D !D !S ðH2OÞT Mg m3 ð3Þ It depends on the ratio of fiber and matrix in a consolidated body. According to densities of the fiber (2.74 Mg.m3 ) and the matrix theoretically derived from the slurry (2.92 Mg.m3 ), the density was proportional to matrix content in all constituents. Microstructural characterizations were performed after consolidation and subsequent PIP processing, using optical microscope (OM) and scanning electron microscope (SEM). Three-point flexural test was performed at room temperature. Dimensions of test specimens were 30 mm length4 mm width1 mm height. Span and crosshead speed were 25 mm and 0.5 mm/min, respectively. Ultimate flexural strength (u) and work-of-fracture (W.O.F) were obtained from the peak load and the area of load-crosshead displacement chart. 3. Results and discussion 3.1. Pyrolytic behavior of the polymer Fig. 1 shows TG–DTA curves of PVS at 300 K/h in Ar. Mass degradation continuously occurred from 380 to 800 K. It was accelerated along with temperature and became highest between 650 and 700 K. Most of the mass change had finished below 700 K. Mass yield of a pyrolyzed product up to 1473 K was 32.6%. As for DTA curve, a downward tendency was seen from 400 K. It should be related with the mass degradation due to the emission of polymer components. Between 500 and 600 K, there could be identified a continuous depression. Based on previous reports [23,24], it might be related with cross-linking reaction. Big endothermic peak at 700 K implied a drastic change of the molecular structure to form Si–C backbone, where the fragmentation of polymer structure simultaneously occurred [25]. Appearances of isolated samples of PVS heated up to various temperatures were exhibited in Table 1. On the whole, the polymer showed continuous thermosetting from transparent liquid to porous brownish glassy solid between 600 and 700 K. The polymer pyrolyzed up to 583 K was not so much different from original one other than a slight increase of viscosity. Heated up to 603 K, gelation was recognized. Referred to the TGA curve, mass degradation had already reached more than 20% here. Then, the polymer continuously thermoset with further mass degradation and gradual coloration. Below 673 K, the pyrolyzed products were free of pore. But, pyrolyzed above 693 K, the polymer became brownish and quite insoluble in solvents, and frothed vigorously. These features suggested that the polymer evolved much gas and subsequently lost almost its plasticity at this temperature. Fig. 2 shows densities and volumetric residues of PVS as a function of temperature. Although it is not easy to Fig. 1. Thermogravimetric and differential thermal analysis (TGDTA) curves for PVS at a heating rate of 300 K/h in Ar. M. Kotani et al. / Composites Science and Technology 62 (2002) 2179–2188 2181
M. Kotani et al. /Composites Science and Technology 62(2002 )2179-2188 Table I 8.0 Appearances of isolated pyrolyzed products of PVs up to various Temp.(K) 6.0 Transparent liquid(100 cP) Transparent viscous liquid 4.0 anslucent rubbery solid, non porous 3.0 Transparent glassy solid, non porous Yellowish glassy solid, non porous 0 Brownish glassy solid, non porous 1.0 0.0 measured accurate density for a light polymer-pyrolyze intermediate by picnometry, it could be apparently Temperature /K shown that the densities were increased almost linearly Fig 3. Gas evolution behavior of PVS as a function of temperature with temperature. Thus, the volumetric residue showed similar behavior to the mass change. where drastic TGA curves for PVS at various heating rates are volumetric shrinkage occurred between 600 and 700 K shown in Fig. 4. Clear effect of heating rate was seen in and most of volume change had finished below 700 K. ceramic yield, where it was improved from 32%(600 K/ Volumetric yield of the polymer after pyrolysis up to h)to 37%(10 K /h). It might be owing to the increase of 1473 K was estimated to be 11% time for cross linking at around 700 K [25]. Another As an unavoidable negative influence on matrix mor- noticeable feature appeared in the mass degradation phology, gas evolution behavior during pyrolysis was behavior of early stage of pyrolysis(400-600 K). It monitored as a function of temperature in Fig 3. Each increased as the heating rate was slowed down. There point represents total amount of gases evolved from 100 fore, it was considered that the ceramics yield of PVs K below to representative temperature As gas evolution vas not so influenced by the amount of loss of low was negligible below 600 K, mass degradation detected molecular fraction, but the degree of cross linking below 600 K was proved to be due to the evaporation of Thus, sufficient time for cross linking was quite impor low molecular oligomers [26]. PVS heated up to more tant for efficient consolidation. than 400 K, those fractions might volatilize from a cru According to these results, it was approved that main cible and deposited on the cool part of the heating decomposition of PVs occurred between 600 and 700 K apparatus. Gas evolution was drastically increased from with great amount of gas evolution and mass degrada 600 to 700 K. This event would be related with the tion. As inorganization was highly proceeded, the poly- fragmentation of polymer structure. It was also sug- mer would turn poor of fluidity above 700 K Since fiber gested by a big endothermic peak detected in Dta distribution and matrix homogeneity in a composite curve. Then, gas evolution gradually declined along cannot be improved after the polymer has lost fluidity with temperature fiber alignment, stacking of prepreg sheets and shaping 1000 1500 Temperature /K Temperature /K Fig. 2. Densities and volumetric residues of Pvs as a function of temperature. Fig 4. TGA curves for PvS at various heating rates in Ar
measured accurate density for a light polymer-pyrolyzed intermediate by picnometry, it could be apparently shown that the densities were increased almost linearly with temperature. Thus, the volumetric residue showed similar behavior to the mass change, where drastic volumetric shrinkage occurred between 600 and 700 K and most of volume change had finished below 700 K. Volumetric yield of the polymer after pyrolysis up to 1473 K was estimated to be 11%. As an unavoidable negative influence on matrix morphology, gas evolution behavior during pyrolysis was monitored as a function of temperature in Fig. 3. Each point represents total amount of gases evolved from 100 K below to representative temperature. As gas evolution was negligible below 600 K, mass degradation detected below 600 K was proved to be due to the evaporation of low molecular oligomers [26]. PVS heated up to more than 400 K, those fractions might volatilize from a crucible and deposited on the cool part of the heating apparatus. Gas evolution was drastically increased from 600 to 700 K. This event would be related with the fragmentation of polymer structure. It was also suggested by a big endothermic peak detected in DTA curve. Then, gas evolution gradually declined along with temperature. TGA curves for PVS at various heating rates are shown in Fig. 4. Clear effect of heating rate was seen in ceramic yield, where it was improved from 32% (600 K/ h) to 37% (10 K/h). It might be owing to the increase of time for cross linking at around 700 K [25]. Another noticeable feature appeared in the mass degradation behavior of early stage of pyrolysis (400–600 K). It increased as the heating rate was slowed down. Therefore, it was considered that the ceramics yield of PVS was not so influenced by the amount of loss of low molecular fraction, but the degree of cross linking. Thus, sufficient time for cross linking was quite important for efficient consolidation. According to these results, it was approved that main decomposition of PVS occurred between 600 and 700 K with great amount of gas evolution and mass degradation. As inorganization was highly proceeded, the polymer would turn poor of fluidity above 700 K. Since fiber distribution and matrix homogeneity in a composite cannot be improved after the polymer has lost fluidity, fiber alignment, stacking of prepreg sheets and shaping Table 1 Appearances of isolated pyrolyzed products of PVS up to various temperatures Temp. (K) Appearance r.t. Transparent liquid (100 cP) 583 Transparent viscous liquid 603 Transparent gel 623 Translucent rubbery solid, non porous 653 Transparent glassy solid, non porous 673 Yellowish glassy solid, non porous 693 Brownish glassy solid, non porous Fig. 2. Densities and volumetric residues of PVS as a function of temperature. Fig. 4. TGA curves for PVS at various heating rates in Ar. Fig. 3. Gas evolution behavior of PVS as a function of temperature. 2182 M. Kotani et al. / Composites Science and Technology 62 (2002) 2179–2188
M. Kotani et al. / Composites Science and Technology 62(2002)2179-2188 2183 should be completed below 700 K. Curing processing Fig. 6(b), the break of a product into fibers was caused prior to consolidation might be effective for densifica- by squeeze of slurry from fiber preform. Successive tion of a body because effective volumetric yield of a consolidation was accomplished under the pressure of 1 precursor was improved [see Formula(1)]. If the poly- 5 and 10 MPa. Similar profiles were shown as a function mer was cured up to 600 and 700 K, it could be of curing temperature under these pressures;i.e.con- improved by almost twice. Also, further densification tinuous change and particular high density could be might be possible by crushing pores with appropriate presented. It was considered that the green bodies were pressure, because the polymer continuously underwent efficiently consolidated owing to well-balanced relation viscous liquid and plastic solid in this range of tem- ship between the physical characteristics of a precursor perature. At that time, a precursor impregnated into and pressure in these conditions. As pressure was fibrous body had to be soft enough for plastic defor- increased, curing temperature at which the particular mation under external forces. In addition, heating rate high relative density appeared rose, and those values during the first pyrolysis should be slow so as not only was increased. It nsidered that this increase of to control crack initiation and dilatation but to improve relative density was owing to the improvement of effec ceramic yield. These requirements have to be fully taken tive yield of a precursor. In this optimization, relative into consideration for process development density could attain to 70% in the condition of (623, 10) Fig. 7 shows apparent densities of the consolidated 3. 2. Optimization of consolidation conditions bodies under various conditions. Apparent density defined in this work provides useful information about 3.2.1. Curing temperature and pressure microstructure, because it depends on the ratio of the As thermosetting of PVS from viscous liquid to solid occurred in very short range between 583 and 663 K curing temperature was precisely controlled with the interval of 20 K. Fig. 5 shows relative densities of the consolidated bodies under various conditions between curing temperature and pressure. Clear effects of the process conditions on relative density could be identi- fied. Fig. 6(a) and(b) exhibits representative appear ances of the samples consolidated in the conditions of g (583, 0) and(603, 20) respectively, showing extreme. 551 cases related with pressure. As a noticeable feature of Fig 5, the results of 0 MPa are notably inferior to those of other pressure. Many large pores among inhomo- geneously distributed fibers and bundles could be seen in Fig. 6(a). This figure indicates that pressure was very important for appropriate fiber distribution and densi- fication, though the fabrication process of a composite without pressurization is very attractive to develop near-net shape production of complicated-shaped com- K ponents. On the contrary, excessive pressure also gave a Fig. 5. Relative densities of as-consolidated bodies under various negative influence for microstructure. As shown in conditions between curing temperature and pressure (b) Fig. 6. Optical micrographs of as-consolidated bodies under the conditions of (a)(583, 0)and(b)(623, 15)
should be completed below 700 K. Curing processing prior to consolidation might be effective for densification of a body because effective volumetric yield of a precursor was improved [see Formula (1)]. If the polymer was cured up to 600 and 700 K, it could be improved by almost twice. Also, further densification might be possible by crushing pores with appropriate pressure, because the polymer continuously underwent viscous liquid and plastic solid in this range of temperature. At that time, a precursor impregnated into a fibrous body had to be soft enough for plastic deformation under external forces. In addition, heating rate during the first pyrolysis should be slow so as not only to control crack initiation and dilatation but to improve ceramic yield. These requirements have to be fully taken into consideration for process development. 3.2. Optimization of consolidation conditions 3.2.1. Curing temperature and pressure As thermosetting of PVS from viscous liquid to solid occurred in very short range between 583 and 663 K, curing temperature was precisely controlled with the interval of 20 K. Fig. 5 shows relative densities of the consolidated bodies under various conditions between curing temperature and pressure. Clear effects of the process conditions on relative density could be identi- fied. Fig. 6(a) and (b) exhibits representative appearances of the samples consolidated in the conditions of (583, 0) and (603, 20) respectively, showing extreme cases related with pressure. As a noticeable feature of Fig. 5, the results of 0 MPa are notably inferior to those of other pressure. Many large pores among inhomogeneously distributed fibers and bundles could be seen in Fig. 6 (a). This figure indicates that pressure was very important for appropriate fiber distribution and densi- fication, though the fabrication process of a composite without pressurization is very attractive to develop near-net shape production of complicated-shaped components. On the contrary, excessive pressure also gave a negative influence for microstructure. As shown in Fig. 6(b), the break of a product into fibers was caused by squeeze of slurry from fiber preform. Successive consolidation was accomplished under the pressure of 1, 5 and 10 MPa. Similar profiles were shown as a function of curing temperature under these pressures; i. e. continuous change and particular high density could be presented. It was considered that the green bodies were efficiently consolidated owing to well-balanced relationship between the physical characteristics of a precursor and pressure in these conditions. As pressure was increased, curing temperature at which the particular high relative density appeared rose, and those values was increased. It was considered that this increase of relative density was owing to the improvement of effective yield of a precursor. In this optimization, relative density could attain to 70% in the condition of (623,10). Fig. 7 shows apparent densities of the consolidated bodies under various conditions. Apparent density defined in this work provides useful information about microstructure, because it depends on the ratio of the Fig. 5. Relative densities of as-consolidated bodies under various conditions between curing temperature and pressure. Fig. 6. Optical micrographs of as-consolidated bodies under the conditions of (a) (583, 0) and (b) (623, 15). M. Kotani et al. / Composites Science and Technology 62 (2002) 2179–2188 2183
M. Kotani et al. /Composites Science and Technology 62(2002 )2179-2188 matrix. Continuous profiles were shown as a function of curing temperature in each pressure. The main feature to be noted is that the density of 1, 5, and 10 MPa represented a peak at the temperature at which the optimized relative density presented. It suggested that the porosity of a consolidated body was reduced by increasing matrix ratio. These features implied close relationship of apparent density with the physical char acteristics of a precursor. It is found that effective con- solidation for making a highly densified body with small nt of could be done for Fig. 8(aHd) exhibits SEM micrographs of the con solidated bodies produced in the conditions of (583, 1) (603,5), 0)and(643, 10), respectively. The micrographs (a),(b) and (c) correspond to micro- structures of highly consolidated bodies, as described in Fig. 5. Those are prospective candidates for high 7. apparent densities of as-consolidated bodies under various strength composites. In the conditions of (a) and(b), tions between curing temperature and pressure. fine and fair uniform fiber distribution was observed a (b) 之 100/Fm Fig. 8. SEM micrographs of as-consolidated bodies under the conditions of (a)(583, 1).(b)(603, 5),(c)(623, 10), and(d)(643, 10)
matrix. Continuous profiles were shown as a function of curing temperature in each pressure. The main feature to be noted is that the density of 1, 5, and 10 MPa represented a peak at the temperature at which the optimized relative density presented. It suggested that the porosity of a consolidated body was reduced by increasing matrix ratio. These features implied close relationship of apparent density with the physical characteristics of a precursor. It is found that effective consolidation for making a highly densified body with small amount of pores could be done for appropriately cured green body under well-balanced pressure. Fig. 8(a)–(d) exhibits SEM micrographs of the consolidated bodies produced in the conditions of (583, 1), (603, 5), (623, 10) and (643, 10), respectively. The micrographs (a), (b) and (c) correspond to microstructures of highly consolidated bodies, as described in Fig. 5. Those are prospective candidates for highstrength composites. In the conditions of (a) and (b), fine and fair uniform fiber distribution was observed. Fig. 8. SEM micrographs of as-consolidated bodies under the conditions of (a) (583, 1), (b) (603, 5), (c) (623, 10), and (d) (643, 10). Fig. 7. Apparent densities of as-consolidated bodies under various conditions between curing temperature and pressure. 2184 M. Kotani et al. / Composites Science and Technology 62 (2002) 2179–2188
M. Kotani et al. Composites Science and Technology 62(2002)2179-2188 Apparent difference was not identified between these 3. 2.2. Heating rate conditions. There seemed to be enough matrix as to Fig. 10 shows relative densities of the consolidated bind fibers tightly in both samples, though many pores bodies at various heating rates. There could be con- of various sizes still remained all over the specimen. sidered two main influences by heating rate, i.e. volu- These pores would be formed due to volume shrinkage metric yield of a polymer-pyrolyzed product as of matrix precursor. While, stratiform structure where demonstrated in Fig. 4 and gas evolution rate. Volu- both fiber rich and matrix rich areas were formed was metric yield of a precursor directly affected porosity recognized all over the specimen in micrograph(c). It while gas evolution may cause crack initiation if it is too due to poor fluidity of the precursor. This structure tinuously improved by slowing the heating rate down o would take over intermediate layer between cured sheet much. In case of 583 K. relative densities were co could provide a negative influence on mechanical beha 60 K/h. However, it declined at low heating rate from viors such as delamination, in spite of superior relative 30 to 10 K/h. While, relative densities of the composites density [27]. Micrograph(d) is a typical example of the at 643 K were continuously improved as heating rate microstructure cured at beyond 623 K. Many cracks as became slower all over the range, although those differ well as lamination layers were observed all over the ences were fairly smaller. There could not be seen the specimen. It would be initiated by pressure applied for declination of the density that was shown in case of 583 excessively hardened sheets. K. According to the previous report that showed that With above careful inspection for the consolidated the evolution of Sic structure didn, t depend on only bodies, it was demonstrated that the density and micro- temperature but time exposed at high temperature [2 structure of a polymer-derived composite depend on thethe crystallization of the polymer-pyrolyzed product consolidation conditions, which closely related with the could be employed as one of reason for the declination. characteristics of a precursor. Fig. 9 shows schematic But, considering that the declination was shown only ummary of the effect of consolidation conditions in the for the green bodies cured up to 583 K, it was possibly aspects of density and microstructure. Microstructures related with heating condition under pressure before the seemed to be classified into following four condition polymer was solidified. As fiber distribution in a green areas: (1)effective consolidation was achieved with well- body cured up to 643 K could not be changed during balanced pressure for cured sheets, where the homo- consolidation so much, the improvement of the density geneity of matrix was dependent on curing temperature, was considered to be owed only to the increase of cera- (2)pores and cracks were formed due to insufficient mic yield. Greater increase and subsequent decline of supression,(3) fibrous body with little matrix was pro- relative density which were shown in case of 583 K must duced due to excess pressure, (4)laminate structure and include any other factors than the yield. crack were formed due to poor fluidity of a precursor Fig. 1l exhibits cross sectional micrographs of the From the standpoint of density and microstructure, the consolidated bodies at the heating rates of (a)600,(b) condition(603, 5)seemed to be most promising process 30 and(c)10 K/h, respectively. Those composites were candidate for high performance composite production derived from the green bodies cured up to 583 K. There 583K 2 1000 10000 Curing Temperature Heating Rate/Kh Fig 9. Schematic summary of consolidation conditions in the aspects Fig. 10. Relative densities of as-consolidated bodies at various heating of density and microstructure
Apparent difference was not identified between these conditions. There seemed to be enough matrix as to bind fibers tightly in both samples, though many pores of various sizes still remained all over the specimen. These pores would be formed due to volume shrinkage of matrix precursor. While, stratiform structure where both fiber rich and matrix rich areas were formed was recognized all over the specimen in micrograph (c). It would take over intermediate layer between cured sheets due to poor fluidity of the precursor. This structure could provide a negative influence on mechanical behaviors such as delamination, in spite of superior relative density [27]. Micrograph (d) is a typical example of the microstructure cured at beyond 623 K. Many cracks as well as lamination layers were observed all over the specimen. It would be initiated by pressure applied for excessively hardened sheets. With above careful inspection for the consolidated bodies, it was demonstrated that the density and microstructure of a polymer-derived composite depend on the consolidation conditions, which closely related with the characteristics of a precursor. Fig. 9 shows schematic summary of the effect of consolidation conditions in the aspects of density and microstructure. Microstructures seemed to be classified into following four condition areas: (1) effective consolidation was achieved with wellbalanced pressure for cured sheets, where the homogeneity of matrix was dependent on curing temperature, (2) pores and cracks were formed due to insufficient supression, (3) fibrous body with little matrix was produced due to excess pressure, (4) laminate structure and crack were formed due to poor fluidity of a precursor. From the standpoint of density and microstructure, the condition (603, 5) seemed to be most promising process candidate for high performance composite production. 3.2.2. Heating rate Fig. 10 shows relative densities of the consolidated bodies at various heating rates. There could be considered two main influences by heating rate, i.e. volumetric yield of a polymer-pyrolyzed product as demonstrated in Fig. 4 and gas evolution rate. Volumetric yield of a precursor directly affected porosity, while gas evolution may cause crack initiation if it is too much. In case of 583 K, relative densities were continuously improved by slowing the heating rate down on 30 K/h. However, it declined at low heating rate from 30 to 10 K/h. While, relative densities of the composites at 643 K were continuously improved as heating rate became slower all over the range, although those differences were fairly smaller. There could not be seen the declination of the density that was shown in case of 583 K. According to the previous report that showed that the evolution of SiC structure didn’t depend on only temperature but time exposed at high temperature [28], the crystallization of the polymer-pyrolyzed product could be employed as one of reason for the declination. But, considering that the declination was shown only for the green bodies cured up to 583 K, it was possibly related with heating condition under pressure before the polymer was solidified. As fiber distribution in a green body cured up to 643 K could not be changed during consolidation so much, the improvement of the density was considered to be owed only to the increase of ceramic yield. Greater increase and subsequent decline of relative density which were shown in case of 583 K must include any other factors than the yield. Fig. 11 exhibits cross sectional micrographs of the consolidated bodies at the heating rates of (a) 600, (b) 30 and (c) 10 K/h, respectively. Those composites were derived from the green bodies cured up to 583 K. There Fig. 9. Schematic summary of consolidation conditions in the aspects of density and microstructure. Fig. 10. Relative densities of as-consolidated bodies at various heating rates. M. Kotani et al. / Composites Science and Technology 62 (2002) 2179–2188 2185
M. Kotani et al. /Composites Science and Technology 62(2002)2179-2188 Fig. 11. SEM micrographs of as-consolidated bodies at the heating rates of (a)600,(b)30, and(c)10 K/h. could be seen more cracks as the heating rate increased. This feature should be related with gas evolution rate during consolidation and provide main influence on 700 1 MPa relative density. Since these kinds of morphologies would remain after multiple densification processing, ■10MPa heating condition during consolidation should be also appropriately selected to control these negative influences Fig. 12 exhibits average flexural strengths of the composites, which were summarized as the functions of (a) curing temperature and (b) pressure. The strengths were distributed between 200 and 700 MPa and closely dependent on curing temperature, as shown in (a) Curing temperature /K Those became higher as curing temperature was low ered. The composites, which were obtained from highly 583K (623, 10), didnt show remarkable high strength. This implied that densification of a consolidated body did not lead to apparent mechanical improvement directly. for pressure, no representative depende shown in(b) Fig. 13 represents flexural strengths of the composites consolidated at various heating rates. Unified tendency for all composites was not seen. The effect of crack dis- tribution that was shown in Fig. Il on mechanical per formance was not be revealed. On the whole. the composites fabricated at lower curing temperature Pressure/MPa showed higher strength. And strengths became more Fig. 12. Flexural strengths of the composites fabricated under various constant as curing temperature increased. These fea- conditions of (a)curing temperature and (b)pressure. tures were also shown in Fig. 12. According to these results, only curing temperature hardening of the slurry as curing temperature increased was approved to be influential factor for flexural Therefore, it was found that the flexural strength had strength. For this factor, any change in microstructure close dependence on fiber volume fraction. In spite of would be brought about to affect flexural strength. concentrated effort on reducing porosity, remarkable Fig. 14 exhibited the relationship between fiber volume improvement of mechanical performance was not fraction and curing temperature of the consolidated achieved for r the comp osites obtained from highly com- bodies. Close dependence where fiber volume fraction pacted consolidated body. As for fracture behavior, was linearly decreased along with curing temperature almost all composites tested showed non-catastrophic was shown. It was considered that compacting efficiency feature, even though no interfacial layer was presently of a green body by pressurization would decline due to introduced. Since no big difference was presented in
could be seen more cracks as the heating rate increased. This feature should be related with gas evolution rate during consolidation and provide main influence on relative density. Since these kinds of morphologies would remain after multiple densification processing, heating condition during consolidation should be also appropriately selected to control these negative influences. 3.2.3. Flexural test Fig. 12 exhibits average flexural strengths of the composites, which were summarized as the functions of (a) curing temperature and (b) pressure. The strengths were distributed between 200 and 700 MPa and closely dependent on curing temperature, as shown in (a). Those became higher as curing temperature was lowered. The composites, which were obtained from highly consolidated bodies in the conditions of (603, 5) and (623, 10), didn’t show remarkable high strength. This implied that densification of a consolidated body did not lead to apparent mechanical improvement directly. As for pressure, no representative dependence was shown in (b). Fig. 13 represents flexural strengths of the composites consolidated at various heating rates. Unified tendency for all composites was not seen. The effect of crack distribution that was shown in Fig. 11 on mechanical performance was not be revealed. On the whole, the composites fabricated at lower curing temperature showed higher strength. And strengths became more constant as curing temperature increased. These features were also shown in Fig. 12. According to these results, only curing temperature was approved to be influential factor for flexural strength. For this factor, any change in microstructure would be brought about to affect flexural strength. Fig. 14 exhibited the relationship between fiber volume fraction and curing temperature of the consolidated bodies. Close dependence where fiber volume fraction was linearly decreased along with curing temperature was shown. It was considered that compacting efficiency of a green body by pressurization would decline due to hardening of the slurry as curing temperature increased. Therefore, it was found that the flexural strength had close dependence on fiber volume fraction. In spite of concentrated effort on reducing porosity, remarkable improvement of mechanical performance was not achieved for the composites obtained from highly compacted consolidated body. As for fracture behavior, almost all composites tested showed non-catastrophic feature, even though no interfacial layer was presently introduced. Since no big difference was presented in Fig. 11. SEM micrographs of as-consolidated bodies at the heating rates of (a) 600, (b) 30, and (c) 10 K/h. Fig. 12. Flexural strengths of the composites fabricated under various conditions of (a) curing temperature and (b) pressure. 2186 M. Kotani et al. / Composites Science and Technology 62 (2002) 2179–2188
M. Kotani et al. / Composites Science and Technology 62(2002)2179-2188 号400 800 um Heating rate/Kh Fig. 13. Flexural strengths of the composites fabricated under various heating rate. n 10 MPa cog=o日 Fig. 15. Fracture surface of the composite consolidated in the condi- tions of (a)(583, 5). and(b)(603, 5) 600 fibers are hindered This feature suggested that fracture behavior of a composite was related with matrix micro- Curing temperature /K structure as well as fiber/matrix interfacial properties Fig. 14. Fiber volume fraction of each composite summarized as a And those factors might be related each other for frac- unction of curing temperature. ture behavior. It was considered that embrittlement occurred more easily as porosity of matrix was those stress-strain curves, those work- of fractures were decreased. Thus, main factors in mechanical perfor- nearly proportional to flexural strength. Therefore, it mance, such as fiber volume fraction, porosity and could be concluded that the consolidation condition interfacial property, should be syncetically controlled should be optimized to raise fiber volume fraction To acquire universal relationship among these factors is rather than decrease porosity for improving flexural essential for the final goal for establishing a process to strength of a composite, as far as the composite broke in fabricate high performance SiC/Sic composite for non-catastrophic fracture mode severe environments Basically, technical development to form homo- geneous matrix with smaller amount of defects is quite important for improving the performance and reliability 4. Conclusion of a composite. But the process optimization to reduce porosity didn,t lead to apparent improvement of flex- To fabricate a high performance SiC/SiC composite ural strength. Possible reason for this result was by PIP method, systematic process characterizations for appeared in fracture surface exhibited in Fig. 15. effective consolidation were performed with precise Although many pullout fibers of significant length inspection about the pyrolytic behavior of the polymeric shown in (a) were generally observed, those were precursor, polyvinylsilane. The following conclusions remarkably decreased in the composites obtained from were found highly consolidated bodies as represented in(b). In this case, toughening mechanism between fiber and matrix 1. Pvs was found to occur continuous thermoset would decline because mechanical independences of ting from viscous liquid to glassy solid between
those stress-strain curves, those work-of fractures were nearly proportional to flexural strength. Therefore, it could be concluded that the consolidation condition should be optimized to raise fiber volume fraction rather than decrease porosity for improving flexural strength of a composite, as far as the composite broke in non-catastrophic fracture mode. Basically, technical development to form homogeneous matrix with smaller amount of defects is quite important for improving the performance and reliability of a composite. But the process optimization to reduce porosity didn’t lead to apparent improvement of flexural strength. Possible reason for this result was appeared in fracture surface exhibited in Fig. 15. Although many pullout fibers of significant length shown in (a) were generally observed, those were remarkably decreased in the composites obtained from highly consolidated bodies as represented in (b). In this case, toughening mechanism between fiber and matrix would decline because mechanical independences of fibers are hindered. This feature suggested that fracture behavior of a composite was related with matrix microstructure as well as fiber/matrix interfacial properties. And those factors might be related each other for fracture behavior. It was considered that embrittlement occurred more easily as porosity of matrix was decreased. Thus, main factors in mechanical performance, such as fiber volume fraction, porosity and interfacial property, should be syncetically controlled. To acquire universal relationship among these factors is essential for the final goal for establishing a process to fabricate high performance SiC/SiC composite for severe environments. 4. Conclusion To fabricate a high performance SiC/SiC composite by PIP method, systematic process characterizations for effective consolidation were performed with precise inspection about the pyrolytic behavior of the polymeric precursor, polyvinylsilane. The following conclusions were found. 1. PVS was found to occur continuous thermosetting from viscous liquid to glassy solid between Fig. 13. Flexural strengths of the composites fabricated under various heating rate. Fig. 14. Fiber volume fraction of each composite summarized as a function of curing temperature. Fig. 15. Fracture surface of the composite consolidated in the conditions of (a) (583, 5), and (b) (603, 5). M. Kotani et al. / Composites Science and Technology 62 (2002) 2179–2188 2187
M. Kotani et al. / Composites Science and Technology 62(2002)2179-2188 600 and 700 K, with large amount of mass Preparation and properties of monolithic and composite ceramics degradation and gas evolution. PVS and its produced by polymer pyrolysis. Am Ceram Soc Bull lurry with Sic particles could be compound 1983:62(8):916-23 very well into Hi-Nicalon preform without dilu- 9 Yoshida H, Miyata N, Sagawa M. Ishikawa S, Naito K,Eno- tion, owing to its low viscosity and excellent moto N, Yamagishi C. Preparation of unidirectionally reinforced carbon-SiC composite by repeated infiltration of polycarbosilane wettability against SiCPcS fiber J Ceram Soc Japan 1992: 100(4): 454-8. 2. Efficient densification could be achieved under [10] Shin Dw, Tanaka H. Low-temperature processing of ceramic the consolidation condition in which the physical oven fabric/ceramic matrix composites. J Am Ceram Soc 1994: characteristics of the precursor were well 77(1):97-104. [1 Tanaka T, Tamari N, Kondoh I, Iwasa M. Fabrication and balanced with external pressure. Lowering heat- echanical properties of 3-dimensional Tyranno fiber reinforced ing rate was also approved to be beneficial for Sic composites by repeated infiltration of polycarbosilane. J mproving relative density of a consolidated Ceram Soc Japan 1996: 104(5): 454-7 body. By the process optimization, acon [2] Hurwitz Fl, Calomino AM. Mechanical behavior of a Hi-Nica- solidated body of 70% of relative density with lonM/SiC composite having a polycarbosilane derived matrix. Ceram Eng Sci Proc 1999: 20(3): 251-8. favorable fiber distribution successfully pro- duced only one time PIP processing [13 Interrante LV, Whitmarsh CW, Sherwood w, Wu HJ, Lewis R, Maciel G. High yield polycarbosilane precursors to stoichio- 3. Requirements in consolidation processing to metric SiC Synthesis, Pyrolysis and Application. Mater Res Soc fabricate strong composite were fundamentally Symp Proc 1994:346:593-603 acquired. It was approved that increasing fib [14 Interrante LV, Rushkin 1, Shen Q. Linear and hyper polycarbosilanes with Si-CH]Si bridging groups:a volume fraction was most important factor for e producti materials. Appl Organometal Chem 1998: 12: 695-705 composite showed non-catastrophic fracture [5] Kotani M, Kohyama A, Okamura K, Inoue T. Fabrication behavi nce SiC/SiC composite by polymer impregnation [16 Kotani M, Kohyama A, Katoh Y, Okamura K. Effect of Sic article dispersion on microstructure and mechanical properties Acknowledgements [17 Jamet J, Spann JR, Rice RW, Lewis D, Coblenz ws Ceramic- er composite processing via polymer-filler matrices Ceram Eng This work is performed as a part of'R&d of Com- Sci proc1984;5(7):677-94 [18 Suttor D, Erny T, Grail P, Goedeke H, Hung T Fiber reinforced posite Materials for Advanced Energy Systems'research cmc with polymer/filler derived matrix. In: Hausner H, Hirano S, project, supported by Core Research for Evolutional Messing gL, et al, editors. Ceramic Transactions, vol. 51. Wes- Science and Technology(CREST). The authors are berville OH: Am Ceram Soc: 1995.p 211-5. grateful to Dr. M. Itoh(Mitsui chemicals, inc )for pro- (9) Greil P. Active-filler-controlled pyrolysis of preceramic polymers iding polyvinylsilane [20 Idesaki A, Narisawa M, Okamura K, Sugimoto M, Morita Y, Seguchi T, et al. Fine silicon carbide fibers synthesized from beam curing. J Mater Sci 2001: 36: 357-62. References 21 Itoh M, Iwata K, Kobayashi M, Takeuchi R, Kabeya T. Pre- [1 Brewer D. HSR/ EPM combustor materials development pro- 1998:31:560915 Mater Sci Eng 1999: A(261): 284-91 22 Boury B, Corriu RJP, Leclercq D, Mutin PH, Planeix JP,Vioux K, Tezuka A, Funayama O, Isoda T, Terada Y, Kato S, A. Poly(vinylsilane): a precursor to silicon carbide. I. Preparation 4. Fabrication and pressure testing of a gas-turbine com- and characterization. Organometallics 1991: 10: 1457-61 manufactured by a preceramic-polymer-impregnation 23]Boury B, Corriu RP, Douglas WE Poly(carbosilane precursors Comp Sci Technol 1999: 59: 853-9 of silicon carbide: rhe effect of cross-linking on ceramic residue. RE Recent development in fibers and interphases for Chem Mater 1991: 3 487-9 high temperature ceramic matrix composites Composites: Part A [24] Hurwitz Fl, Kacik TA, Bu XY, Masnovi J, Heimann PJ, Beyer K Pyrolytic conversion of methyl- and vinylsilane polymer to Si- 1 Kohyama A, Katoh Y, Hinoki T, Zhang W, Kotani M. Progress C ceramics. J Mater Sci 1995: 30: 3130-6 25 Corriu RP, Leclercq D, Mutin PH, Planeix JM, Vioux A ystems: CREST-ACE program Proc &th Eur Conf Comp Mater Mechanism of pyrolysis of polycarbosilanes: Poly(silylethylene) and poly( dimethylsilylethylene) Organometallics 1993: 12: 454-62 I Snead LL, Jones R, Kohyama A, Fenici P Status of silicon car- [26] Schmidt WR, Interrante LV, Doremus RH, Trout TK, Marchetti s, Maciel GE. Pyrolysis chemistry of an organometallic pre- [6 Cornie JA, Chiang YM, Uhlmann DR, Mortensen A, Collins cursor to silicon carbide. Chem Mater 1991: 3(2): 257 JM. Processing of metal and ceramic matrix composites Ceram 27 Jessen TL, Greenhut III V.A., Friel JJ. J Am Cerar Bul986;65(2:293-303 82(10)2753-61 [7 Jones R, Szweda A, Petrak D. Polymer derived ceramic matrix [28 Soraru GD, Babonneau F, Mackenzie JD. Structural evolutions mposites. Composites: Part A 1999: 30: 569-75 from polycarbosilane to SiC ceramic. J Mater Sci 1 [8 Walker Jr, BE, Rice RW, Becher PF, Bender BA, Coblenz wS
600 and 700 K, with large amount of mass degradation and gas evolution. PVS and its slurry with SiC particles could be compound very well into Hi-Nicalon preform without dilution, owing to its low viscosity and excellent wettability against SiCPCS fiber. 2. Efficient densification could be achieved under the consolidation condition in which the physical characteristics of the precursor were well balanced with external pressure. Lowering heating rate was also approved to be beneficial for improving relative density of a consolidated body. By the process optimization, a consolidated body of 70% of relative density with favorable fiber distribution was successfully produced only one time PIP processing. 3. Requirements in consolidation processing to fabricate strong composite were fundamentally acquired. It was approved that increasing fiber volume fraction was most important factor for strong composite production as far as the composite showed non-catastrophic fracture behavior. Acknowledgements This work is performed as a part of ‘R&D of Composite Materials for Advanced Energy Systems’ research project, supported by Core Research for Evolutional Science and Technology (CREST). The authors are grateful to Dr. M. Itoh (Mitsui chemicals, inc.) for providing polyvinylsilane. References [1] Brewer D. HSR/EPM combustor materials development program. Mater Sci Eng 1999;A(261):284–91. [2] Sato K, Tezuka A, Funayama O, Isoda T, Terada Y, Kato S, Iwasa M. Fabrication and pressure testing of a gas-turbine component manufactured by a preceramic-polymer-impregnation method. Comp Sci Technol 1999;59:853–9. [3] Tressler RE. Recent development in fibers and interphases for high temperature ceramic matrix composites. Composites: Part A 1999;30:429–37. [4] Kohyama A, Katoh Y, Hinoki T, Zhang W, Kotani M. Progress in the development of SiC/SiC composite for advanced energy systems: CREST-ACE program. Proc 8th Eur Conf Comp Mater 1998;4:15–22. [5] Snead LL, Jones R, Kohyama A, Fenici P. Status of silicon carbide composites for fusion. J Nucl Mater 1996:233–7. [6] Cornie JA, Chiang YM, Uhlmann DR, Mortensen A, Collins JM. Processing of metal and ceramic matrix composites. Ceram Bull 1986;65(2):293–303. [7] Jones R, Szweda A, Petrak D. Polymer derived ceramic matrix composites. Composites: Part A 1999;30:569–75. [8] Walker Jr, BE, Rice RW, Becher PF, Bender BA, Coblenz WS. Preparation and properties of monolithic and composite ceramics produced by polymer pyrolysis. Am Ceram Soc Bull 1983;62(8):916–23. [9] Yoshida H, Miyata N, Sagawa M, Ishikawa S, Naito K, Enomoto N, Yamagishi C. Preparation of unidirectionally reinforced carbon-SiC composite by repeated infiltration of polycarbosilane. J Ceram Soc Japan 1992;100(4):454–8. [10] Shin DW, Tanaka H. Low-temperature processing of ceramic woven fabric/ceramic matrix composites. J Am Ceram Soc 1994; 77(1):97–104. [11] Tanaka T, Tamari N, Kondoh I, Iwasa M. Fabrication and mechanical properties of 3-dimensional Tyranno fiber reinforced SiC composites by repeated infiltration of polycarbosilane. J Ceram Soc Japan 1996;104(5):454–7. [12] Hurwitz FI, Calomino AM. Mechanical behavior of a Hi-NicalonTM/SiC composite having a polycarbosilane derived matrix. Ceram Eng Sci Proc 1999;20(3):251–8. [13] Interrante LV, Whitmarsh CW, Sherwood W, Wu HJ, Lewis R, Maciel G. High yield polycarbosilane precursors to stoichiometric SiC. Synthesis, Pyrolysis and Application. Mater Res Soc Symp Proc 1994;346:593–603. [14] Interrante LV, Rushkin I, Shen Q. Linear and hyperbranched polycarbosilanes with Si-CH2-Si bridging groups: a synthetic platform for the construction of novel functional polymeric materials. Appl Organometal Chem 1998;12:695–705. [15] Kotani M, Kohyama A, Okamura K, Inoue T. Fabrication of high performance SiC/SiC composite by polymer impregnation and pyrolysis method. Ceram Eng Sci Proc 1999;20(4):309–16. [16] Kotani M, Kohyama A, Katoh Y, Okamura K. Effect of SiC particle dispersion on microstructure and mechanical properties of polymer-derived SiC/SiC composite. J Mater Sci (in press). [17] Jamet J, Spann JR, Rice RW, Lewis D, Coblenz WS. Ceramic- fiber composite processing via polymer-filler matrices. Ceram Eng Sci Proc 1984;5(7):677–94. [18] Suttor D, Erny T, Grail P, Goedeke H, Hung T. Fiber reinforced cmc with polymer/filler derived matrix. In: Hausner H, Hirano S, Messing GL, et al, editors. Ceramic Transactions, vol. 51. Westerville OH: Am Ceram Soc; 1995. p. 211–5. [19] Greil P. Active-filler-controlled pyrolysis of preceramic polymers. J Am Ceram Soc 1995;78(4):835–48. [20] Idesaki A, Narisawa M, Okamura K, Sugimoto M, Morita Y, Seguchi T, et al. Fine silicon carbide fibers synthesized from polycarbosilane-polyvinylsilane polymer blend using electron beam curing. J Mater Sci 2001;36:357–62. [21] Itoh M, Iwata K, Kobayashi M, Takeuchi R, Kabeya T. Preparations and properties of poly(vinylsilane)s. Macromolecules 1998;31:5609–15. [22] Boury B, Corriu RJP, Leclercq D, Mutin PH, Planeix JP, Vioux A. Poly(vinylsilane): a precursor to silicon carbide. 1. Preparation and characterization. Organometallics 1991;10:1457–61. [23] Boury B, Corriu RJP, Douglas WE. Poly(carbosilane precursors of silicon carbide: rhe effect of cross-linking on ceramic residue. Chem Mater 1991;3:487–9. [24] Hurwitz FI, Kacik TA, Bu XY, Masnovi J, Heimann PJ, Beyene K. Pyrolytic conversion of methyl- and vinylsilane polymer to Si– C ceramics. J Mater Sci 1995;30:3130–6. [25] Corriu RJP, Leclercq D, Mutin PH, Planeix JM, Vioux A. Mechanism of pyrolysis of polycarbosilanes: Poly(silylethylene) and poly(dimethylsilylethylene). Organometallics 1993;12:454–62. [26] Schmidt WR, Interrante LV, Doremus RH, Trout TK, Marchetti PS, Maciel GE. Pyrolysis chemistry of an organometallic precursor to silicon carbide. Chem Mater 1991;3(2):257–67. [27] Jessen TL, Greenhut III V.A., Friel JJ. J Am Ceram Soc 1999; 82(10):2753–61. [28] Soraru GD, Babonneau F, Mackenzie JD. Structural evolutions from polycarbosilane to SiC ceramic. J Mater Sci 1990;25:3886– 93. 2188 M. Kotani et al. / Composites Science and Technology 62 (2002) 2179–2188