could not be foamed when rapidly depressurized at room temperature. Consequently, cell nucleation could occur heterogeneously at high energy regions such as the boundaries between the solid LDPE and melt polysiloxane phases and
the boundaries between the solid microbead and melt polysiloxane phases. Thus, it was expected that the addition of
LDPE would increase the cell density of the foamed body. However, the opposite results were obtained. Thus, we surmised from the results that a 10 wt.% addition of LDPE does not make more nucleation sites per unit volume than a
4 wt.% addition of 20 μm-microbead. Note that the addition of 10 wt.% LDPE decreased the content of microbeads from
40 % (SiOC1) to 36 % (SiOC2) (see Table 1). This means that the size of the dispersed LDPE in the extruded blends
should be larger than the size of the microbead, i.e., ~20 μm. A previous paper [12] confirmed that the microbead in
polysiloxane and microbead blends acted as a nucleating site for the cells during foaming. Thus, the use of a smaller
microbead led to a higher population density of nucleation sites per unit volume, resulting in a higher overall cell density.
The present results suggest that smaller microbead is a more efficient nucleating agent than LDPE and larger microbead
in polysiloxane-LDPE-microbead blends. The porosities of foamed blends were ~53.5 % for SiOC1,
~71.0 % for SiOC2, and ~74.9 % for SiOC3 (Fig. 3). Since the number of nucleation sites per unit volume of SiOC1
was larger than the LDPE-containing specimens (SiOC2 and SiOC3), LDPE-containing specimens (Fig. 1(b) and (c))
nucleated a smaller number of cells and that smaller number of cells led to faster cell growth rate, resulting in larger cells
(see Fig. 1(a)). Thus, the addition of LDPE increased the porosity of foamed blends. When the foamed specimen was cross-linked and pyrolyzed, the LDPE and polymer microbeads were burned out and only SiOC ceramic remained. The cross-linking process ensured that the foamed specimens would maintain their foamed structures during pyrolysis; in consequence, porous SiOC ceramics were finally fabricated. By combining the foaming and the pyrolysis processes, a homogeneous, opencell SiOC ceramic was produced at a temperature as low as1200 oC.Typical microstructures of the porous SiOC ceramics after pyrolysis at 1200o C are shown in Fig. 4. As shown in Fig. 4, open-cells were formed for all specimens. The morphology of the cells were duplex: (1) spherical cells that were replicated from the polymer microbeads, indicating that the microbead shape was retained in the blend until the polymer microbeads reached their decomposition temperature and (2) relatively large elongated or equiaxed cells that were nucleated and grown by foaming using CO2 or formed by decomposing LDPE during pyrolysis. The cell densities of the SiOC foams were 1.1×109 cells/cm3
for SiOC1, 7.1×108 cells/cm3 for SiOC2, and 2.1×108 cells/cm3for SiOC3 (Fig. 2). It was observed for the foamed specimens that the addition of LDPE resulted in a lower cell density and the use of a larger microbead resulted in a lower cell density when the same content of LDPE was added. Tendencies and characteristics observed in the foaming process
were also observed in this process, with respect to the cell density of the foamed and pyrolyzed specimens. The cell
densities of the foamed and pyrolyzed specimens (2.1×108 ~1.1×109 cells/cm3, see Fig. 2) were higher than those of the
foamed specimens (2.6×107~1.8×108 cells/cm3 ) because ofboth the burn-out of the polymer microbead and the shrinkagethat accompanied the pyrolysis.The porosities of the SiOC foams were ~77.4 % for SiOC1,~84.6 % for SiOC2, and ~89.9 % for SiOC3 (Fig. 3). Thesame tendency exhibited in the foamed blends was obtainedafter pyrolysis. The porosity increased as LDPE was added