S. Tariolle et al. / Journal of Solid State Chemistry 177(2004)487-49 Milling/ mixing Polyamide powders allowed to obtain different pore olvent dispersant+ ceramic powder sizes(from 4 to 20 um), corn starch led to spherical pores of approximately 10 um. Corn starch led to the Addition of Binder plasticizer highest level of porosity (47 vol% for 55 vol% of corn starch). In addition, the pyrolyzation was easier with starch [Addition of Pore forming agent In the multi-layered materials, the thickness was 7O um for dense layers and 80 um for porous layers De-airing The microstructure of dense layers and of ligaments in porous layers were similar (grain size 1.2 um, aspe ape casting ratio 1.6)191 Different types of multi-layered materials were pre- Thermo-compaction are Alternate dense-porous materials with porous layer Burning out made with pore forming agent or with porous layer made in the absence of sintering aid: Whatever the material, the Sintering layers had uniform thickness and were parallel each Fig. 2. Steps for preparation of ceramic-ceramic laminates. other. The grain size in the porous layers was similar to that of the dense ones. The mean grain size was 0.8 um Specimens were diamond polished using different The porosity in the layer made with corn starch was terconnected and had a mean size of 10 um. In the stages to obtain a mirror finish. To reveal the micro- porous layer made in the absence of sintering aid, the structure, materials were etched. For B.C, polished porosity was finer surfaces were electrolytically etched with KOH solution Weak interfaces made with graphite spray: These For SiC, polished surfaces were plasma etched (in specimens were produced by graphite coating via CF4/8 vol% O2 atmosphere) spraying and drying on each layer of B4C before thermocompaction. However, this method led to pro- 2.3. Sample characterization blems of uniformity and reproducibility of the weak interfaces Density of materials was calculated from measured Dense layers had a thickness of 100 um(Fig. 4)and eight and the geometrically determined volume. The a relative density of 0.94 porosity was determined by helium pycnometry for closed porosity; however, the size distribution and 3. 2. fracture behavior proportion of open porosity were determined by 3. 2. Sic Image analysis was used for characterization of grain Mechanical properties of monolithic materials have been measured: Young modulus, toughness, fracture and pore size using micrographs obtained by optical energy [15]. Clegg's criterion Eq (3)predicted crack microscopy(for porosity) and SEM(for grain size) Crack propagation in multi-layered materials was deflection for the SiC alternate multi-layered materials evaluated using 3 point-bending fracture tests on with porous layers containing a porosity larger than notched specimens (20mm span, INSTRON 8562, 37 vol%. but no significant crack deflection was displacement measured by LVDT sensor, cross-head observed in our case(Fig. 5) speed 0.025 mm/min) Two issues must be considered in a discussion of the results. First, if we consider the second phase, at grain boundaries, formed by the sintering additives, the chemical composition of this phase, determined from 3. Results and discussion local chemical analysis, continuously changes from layers and there brunt chan 3. Macrostructural and microstructural in fracture energy. As a consequence of this difference in characterization the chemistry between the dense layers and the ligament the 3.. Sic Secondly, the use of graphite platelets appears Different polyamide powders and corn starch were more favorable to deflect cracks than equi-axed corn tested as pore forming agents [91 starch. Then, pore shape appears to be an importantSpecimens were diamond polished using different stages to obtain a mirror finish. To reveal the micro￾structure, materials were etched. For B4C, polished surfaces were electrolytically etched with KOH solution. For SiC, polished surfaces were plasma etched (in CF4/8 vol% O2 atmosphere). 2.3. Sample characterization Density of materials was calculated from measured weight and the geometrically determined volume. The porosity was determined by helium pycnometry for closed porosity; however, the size distribution and proportion of open porosity were determined by mercury porosimetry. Image analysis was used for characterization of grain and pore size using micrographs obtained by optical microscopy (for porosity) and SEM (for grain size). Crack propagation in multi-layered materials was evaluated using 3 point-bending fracture tests on notched specimens (20 mm span, INSTRON 8562, displacement measured by LVDT sensor, cross-head speed 0.025 mm/min). 3. Results and discussion 3.1. Macrostructural and microstructural characterization 3.1.1. SiC Different polyamide powders and corn starch were tested as pore forming agents [9]. Polyamide powders allowed to obtain different pore sizes (from 4 to 20 mm), corn starch led to spherical pores of approximately 10 mm. Corn starch led to the highest level of porosity (47 vol% for 55 vol% of corn starch). In addition, the pyrolyzation was easier with starch. In the multi-layered materials, the thickness was 70 mm for dense layers and 80 mm for porous layers. The microstructure of dense layers and of ligaments in porous layers were similar (grain size 1.2 mm, aspect ratio 1.6) [9]. 3.1.2. B4C Different types of multi-layered materials were pre￾pared (Fig. 3, Table 2): Alternate dense–porous materials with porous layer made with pore forming agent or with porous layer made in the absence of sintering aid: Whatever the material, the layers had uniform thickness and were parallel each other. The grain size in the porous layers was similar to that of the dense ones. The mean grain size was 0.8 mm. The porosity in the layer made with corn starch was interconnected and had a mean size of 10 mm. In the porous layer made in the absence of sintering aid, the porosity was finer. Weak interfaces made with graphite spray: These specimens were produced by graphite coating via spraying and drying on each layer of B4C before thermocompaction. However, this method led to pro￾blems of uniformity and reproducibility of the weak interfaces. Dense layers had a thickness of 100 mm (Fig. 4) and a relative density of 0.94. 3.2. Fracture behavior 3.2.1. SiC Mechanical properties of monolithic materials have been measured: Young modulus, toughness, fracture energy [15]. Clegg’s criterion Eq. (3) predicted crack deflection for the SiC alternate multi-layered materials with porous layers containing a porosity larger than 37 vol%, but no significant crack deflection was observed in our case (Fig. 5). Two issues must be considered in a discussion of the results. First, if we consider the second phase, at grain boundaries, formed by the sintering additives, the chemical composition of this phase, determined from local chemical analysis, continuously changes from dense to porous layers and there are no abrupt changes in fracture energy. As a consequence of this difference in the chemistry between the dense layers and the ligament in the porous layers, Eq. (3) is no longer valid. Secondly, the use of graphite platelets appears more favorable to deflect cracks than equi-axed corn starch. Then, pore shape appears to be an important ARTICLE IN PRESS Addition of Binder + Plasticizer Milling / Mixing Solvent + dispersant+ ceramic powder Addition of Pore forming agent De-airing Tape casting Stacking Thermo-compaction Burning out Sintering Fig. 2. Steps for preparation of ceramic–ceramic laminates. S. Tariolle et al. / Journal of Solid State Chemistry 177 (2004) 487–492 489