Journal of the American Ceramic Society-Nannetti et al Vol. 87. No. 7 Table I. Composites Densification Behavior at Various Stages of Preparation After CvI After Slurry Filtration arent densit Slurry int, PIP cycles Porosity Apparent density 153 20 153 0 197 14 1.4 1.41 2.0 10 1.48 1.93 2.63 2-D: composite made of stacked fabric, 3-D composite with fibers placed along three dimensions: CP: coarse SiC powder, FP: fine SiC powder. II. Experimental Procedure IlL. Results and Discussion (1) Composite Manufacturing (1) Densification Behavior Two-dimensional preforms were obtained stacking and pressing The typical densification behavior of the 2-D and 3-D compos- yranno SA(Ube Industries, Ube, Japan) plain woven cloths in a ites, measured at various stages of preparation, is summarized in T perforated graphite holder until the desired volume fraction Table L. The specimens related to both 2-D and 3-D composites are (-40%)was achieved. The 3-D preform architecture had a relative identified by the number of PIP cycles and by the Sic powder tiber distribution of 40: 40: 20 in the r y z directions, respectively, slurry infiltration step before PIP, using coarse powder(CP and a 36% fiber volume fraction pecimens)and fine powder(FP specimens). The slurry infiltration A thin pyrolytic carbon layer (-0.2 um) was deposited onto permitted homogeneous filling with SiC powder-15%-19% 2-D and 3-D-preforms fibers by isothermal chemical vapor infil (depending on initial preform porosity and SiC powder particle tration (I-CVI) using methane(CH,) precursor. A SiC thin layer ze)of the preform geometrical volume, roughly corresponding to (-0.2 um) was also deposited by I-CVI using methyltrichlorosi one-third of the composite total porosity after CVI lane(CH, SiCI,). The purpose was to increase preform stiffness and to prevent damages during subsequent processing A The porosity decrease versus PIP cycles number is shown in g I for two typical 3-D preforms. It is interesting to note that the To fill the large interbundle porosity inside the preforms and to SiC powder introduced through the slurry infiltration did not seem reduce the number of steps required by the typical PIP process, to affect significantly the subsequent PIP steps efficiency. This some samples were subjected to low-pressure infiltration, through effect can be clearly observed in Fig. I where, especially for the the preform thickness, of a dilute SiC powder aqueous slurry (6 early cycles, similar porosity trends versus PIP cycles are shown for 3-D preforms infiltrated with and without the powder infiltra- average grain size of -5 um (F 1000, Logitech, Glasgow tion step(3-D-CP-7 PIP and 3-D-14 PIP specimens, respectively. Scotland) and a finer one, with average grain size of-2 um(Dow Table D). The Sic powder infiltration step proved to sensibly Ceram, Dortmund, Germany ) Some preforms were also processed increase the preforms density, allowing final lower porosity levels without performing the SiC powder infiltration step with a reduced number of PIP cycles) than the samples produced The samples underwent several polymer impregnation/pyrolysis ly by CVI and PIP (Table 1). Malta, NY)as SiC-precursor, polymer pyrolysis was con- with powder infiltration and those densified only by PIP evidences alternatively at 1 100 and 1300 C in an argon atmospl the Sic powder filling efficiency in preforms large interbundle After densification, some composites were heated up to 1700. to voids. Figure 2 shows a polished cross section of a typical 3-D crystallize the PIP-derived SiC matrix, as performed by Berbon et specimen where a large void, delimited by three fiber bundles. al. for C/SiC composites. appears homogeneously filled with SiC particles bound by the polymer derived SiC. In addition, the presence of a continuous network of SiC powder greatly reduced the length and the size of (2) Characterization the cracks induced by the polymer shrinkage on pyrolisis.These are clearly evident in Fig. 3, which shows the cross section of a Composite apparent density values were computed by mass and geometric volume. The reported porosity values refer to total 2-D composite processed without the slurry infiltration step and porosity and were calculated assuming volume additivity for each phase present in the composite. Fiber density was assumed to be 3.02 g/cm, SiC powder density 3.20 g/cm, while the polymer derived SiC density, measured by helium picnometry, was 2.51 g/em(after pyrolysis at 1300C) -3D-14PIP The modulus of rupture(MOR) was measured by three-point 8-3D-CP-7PIP nding test at room temperature (span 40 mm, crosshead speed 0.5 mm/min) using 3 mm x 6 mm X 50 mm specimens. Elastic constants were measured by the longitudinal and transverse reso- nant frequencies using an e-meter resonant frequencies tester James Instruments, Chicago, IL). Thermal diffusivity was mea sured by the laser flash technique under vacuum. The specific heat 王王10% value of the finished composites was determined by differential 章8% scanning calorimetry (DSC 7, Perkin-Elmer, Norwalk, CT)at nea room temperature. An estimation of the thermal conductivity (TC) 01234567891011121314 was thus obtained by multiplying the thermal diffusivity data by parent density and specific heat. The composite microstructure was investigated by optical microscopy and scanning electron microscopy (SEM), performed on polished cross sections. Fig. I. Porosity versus number of PIP cycles for 3-D composite cessed with and without SiC slurry infiltration