Vol. 80. No Table I. Processing Conditions for Laminated Matrix Composites Temperature of 910-950 5 33232 10-959 900-967 gradient CVI process. In this process, a pressure gradient forces g/cm, respectively, and that the volume of carbon deposited the reagent stream to flow through a preform which is subjected vas twice that of the SiC deposited. This latter assumption is to a temperature gradient. The details of the equipment and an approximation based on observed microstructures rimental procedure have been explained els An entire cross section of each composite disk was mounted where. Briefly, the preforms consisted of 40 layers of T-300 epoxy and polished. The polished sections were observed plain weave carbon cloth, 4. 8 cm in diameter, oriented at Oo and via scanning electron microscopy to permit observation of the 0. These layers were stacked in a graphite preform holder and composite microstructure. Several samples were fractured lightly compacted, giving a height of -0.8 cm. Two types of sing flexure, to observe the propagation of cracks. Transmis preform holders, namely, type 2 and 3, which are described in a sion electron microscopy was used to more clearly observe the prior publication, were used. The type 2 and 3 preform holders thinner layers and to determine the phases deposited extended 5.1 and 7.6 cm above the gas injector, respectively The height of the preform holder influe the temperatu rm.The temperature lI. Results reform olders wenene 350 and o 1 5oc respectively The objective of this work was realized; laminated matrix he operating conditions for the infiltration experim composites containing numerous very thin layers were success- given in Table I. A thin carbon interface was deposited before fully prepared. The infiltration time, density, and porosity of the the deposition of the laminated matrix. The interface was laminated matrix composites(L-1,-3, and-5), and carbon deposited by flowing 40 cm/min of methane and 160 matrix composites(L-2 and-6) used as controls, are given in of hydrogen through the preform for 20 min. The temp Table Il. The infiltration time for the laminated composites of the bottom of the preform during the interface depo was 4.5-8 h versus -4 h for the carbon matrix composites was-975C. This step was followed by deposition of C and apparently, this is the result of Sic deposition being slower SiC, alternately. Carbon was deposited from a reagent mixture than carbon deposition for the conditions used here. The open containing 50% propylene-50%hydrogen, and the total flow porosity of the laminated matrix and that of the carbon matrix rate was 400 cm/min. The Sic layers were deposited using samples are similar, but the closed porosity values are higher of hydrogen. The deposition time for each laminate layer was conditions used for SiC infiltration require Indicates that the 50 cm/min of methyltrichlorosilane(MTS)and 500 cm/min for the laminated matrix composites. This 5 min except for L-5, where each Sic layer was deposited for to achieve similar levels of closed poros: caiustment in order nated matrix composites were achieved in each case(Figs. 1-3) using only a carbon matrix for the purpose of comparison with Both the C and Sic layers were generally continuous with the the lmcs ourse of deerasitin then aminated matn precoated between electron microscopy and electron diffraction verified that the 900 and 961C. This temperature variation was caused by ging the reagent stream, thereby altering the thermal con expected. The number of layers at a given location depended on tivity of the gas between the water-cooled gas injector and the space between the fibers. In a cloth layup, as used in the reform. The thermal conductivity of the propylene/hydro- present work, the distance between the fibers within a tow was gen mixture was lower than that of the MTS/hydrogen mixture 2-3 um(micropores), and the distance between the tows for the concentrations used in the present work. Consequentl 50-100 um(macropores). The tows became densified early in the temperature increased when the propylene/hydrogen mi the infiltration process, and most of the infiltration time was ture was used as the reagent, and the temperature decreased spent on filling the macropores found between the cloth layers when the reagent was changed to MTS/hydrogen. about 60 elapsed between ending the deposition of one layer and starting within a tow(Fig 3). However, all the layers were observed in the deposition of the next layer. During this interval, hydrogen the matrix deposited within the mas,Ofthe den s0.5 um in flowed through the composite As shown in fi 3, layers of posite was determined us were achieved. The thickness of the deposited layers Archimedes'principle with methanol(p =0.79 g/ ncreased with increasing distance from the fiber sur- open-pore volume was calculated by weighing the composite ng the deposition process. The thickness of the initial saturated with methanol. These two values were added to obtain ayers was as small as 0.01 um and increased to.5 ur ne bulk volume. To calculate total porosity it was assumed that the end of the deposition process. The increase in the depo ne densities of the deposited carbon and Sic were 1.9 and 3.2 rate, i.e., layer thickness, with infiltration time was caused by Table Il. Properties of the Infiltrated Composites tal No of Infiltration time Weight gain Bulk de Run No 6.67 12 1.67 164 490 11 1.65 13.7 32 12.88 17.7 50.8 7.57114 Journal of the American Ceramic Society— Lackey et al. Vol. 80, No. 1 Table I. Processing Conditions for Laminated Matrix Composites Temperature of Carbon deposition SiC deposition Preform preform bottom time per cycle time per cycle Run No. type (8C) (min) (min) L-1 3 910–950 5 5 L-2 3 915–954 L-3 2 900–961 5 5 L-5 3 910–959 5 10 L-6 2 900–967 gradient CVI process. In this process, a pressure gradient forces g/cm3 , respectively, and that the volume of carbon deposited the reagent stream to flow through a preform which is subjected was twice that of the SiC deposited. This latter assumption is to a temperature gradient. The details of the equipment and an approximation based on observed microstructures. general experimental procedure have been explained else- An entire cross section of each composite disk was mounted where.24 Briefly, the preforms consisted of 40 layers of T-300 in epoxy and polished. The polished sections were observed plain weave carbon cloth, 4.8 cm in diameter, oriented at 08 and via scanning electron microscopy to permit observation of the 908. These layers were stacked in a graphite preform holder and composite microstructure. Several samples were fractured, lightly compacted, giving a height of ;0.8 cm. Two types of using flexure, to observe the propagation of cracks. Transmis- preform holders, namely, type 2 and 3, which are described in a sion electron microscopy was used to more clearly observe the prior publication,24 were used. The type 2 and 3 preform holders thinner layers and to determine the phases deposited. extended 5.1 and 7.6 cm above the gas injector, respectively. The height of the preform holder influences the temperature and III. Results the temperature gradient through the preform. The temperature differences between the hot and cold sides for the type 2 and 3 The objective of this work was realized; laminated matrix preform holders were ;3508 and ;1508C, respectively. composites containing numerous very thin layers were success- The operating conditions for the infiltration experiments are fully prepared. The infiltration time, density, and porosity of the given in Table I. A thin carbon interface was deposited before laminated matrix composites (L-1, -3, and -5), and carbon the deposition of the laminated matrix. The interface was matrix composites (L-2 and -6) used as controls, are given in deposited by flowing 40 cm3 /min of methane and 160 cm3 /min Table II. The infiltration time for the laminated composites of hydrogen through the preform for 20 min. The temperature was 4.5–8 h versus ;4 h for the carbon matrix composites; of the bottom of the preform during the interface deposition apparently, this is the result of SiC deposition being slower was ;9758C. This step was followed by deposition of C and than carbon deposition for the conditions used here. The open SiC, alternately. Carbon was deposited from a reagent mixture porosity of the laminated matrix and that of the carbon matrix containing 50% propylene–50% hydrogen, and the total flow samples are similar, but the closed porosity values are higher rate was 400 cm3 /min. The SiC layers were deposited using for the laminated matrix composites. This indicates that the 50 cm3 /min of methyltrichlorosilane (MTS) and 500 cm3 /min conditions used for SiC infiltration require adjustment in order of hydrogen. The deposition time for each laminate layer was to achieve similar levels of closed porosity. 5 min except for L-5, where each SiC layer was deposited for Scanning electron microscopy showed that the desired lami- 10 min. Two infiltration runs (L-2 and -6) were conducted nated matrix composites were achieved in each case (Figs. 1–3). using only a carbon matrix for the purpose of comparison with Both the C and SiC layers were generally continuous with the the LMCs. exception of the first few layers in sample L-3. Transmission The temperature of the bottom of the preform during the electron microscopy and electron diffraction verified that the course of depositing the laminated matrix fluctuated between deposits were turbostratic carbon and crystalline SiC, as 9008 and 9618C. This temperature variation was caused by expected. The number of layers at a given location depended on changing the reagent stream, thereby altering the thermal con- the space between the fibers. In a cloth layup, as used in the ductivity of the gas between the water-cooled gas injector and present work, the distance between the fibers within a tow was the preform. The thermal conductivity of the propylene/hydro- 2–3 mm (micropores), and the distance between the tows was gen mixture was lower than that of the MTS/hydrogen mixture 50–100 mm (macropores). The tows became densified early in for the concentrations used in the present work. Consequently, the infiltration process, and most of the infiltration time was the temperature increased when the propylene/hydrogen mix- spent on filling the macropores found between the cloth layers ture was used as the reagent, and the temperature decreased and tows within a cloth.21 when the reagent was changed to MTS/hydrogen. About 60 s Hence, not all layers were observed elapsed between ending the deposition of one layer and starting within a tow (Fig. 3). However, all the layers were observed in the matrix deposited within the macropores (Fig. 2). the deposition of the next layer. During this interval, hydrogen was flowed through the composite. As shown in Figs. 1–3, layers of C and SiC ,0.5 mm in The apparent volume of the composite was determined using thickness were achieved. The thickness of the deposited layers Archimedes’ principle with methanol (r 5 0.79 g/cm generally increased with increasing distance from the fiber sur- 3 ). The open-pore volume was calculated by weighing the composite face during the deposition process. The thickness of the initial saturated with methanol. These two values were added to obtain layers was as small as 0.01 mm and increased to ;0.5 mm near the bulk volume. To calculate total porosity it was assumed that the end of the deposition process. The increase in the deposition the densities of the deposited carbon and SiC were 1.9 and 3.2 rate, i.e., layer thickness, with infiltration time was caused by Table II. Properties of the Infiltrated Composites Fiber content Total No. of Infiltration time Weight gain Bulk density Total porosity Open porosity Run No. (vol%) cycles (h) (g) (g/cm3 ) (%) (%) L-1 50.6 40 6.67 12.90 1.672 16.4 5.94 L-2 49.0 4.25 11.74 1.658 9.2 4.97 L-3 56.7 27 4.50 14.74 1.700 13.7 5.56 L-5 51.8 32 8.00 12.88 1.647 17.7 8.90 L-6 50.8 3.60 13.11 1.692 7.6 7.57