Results and discussion
Cocoons as volume-occupying fillers in foams
The apparent densities of the manufactured unreinforced and cocoon reinforced foams are presented in Fig. 6.3. The measured density of the neat foam (45 kg/m3) matched the datasheet value. It was evident that cocoon-reinforced foams had significantly higher densities than the unreinforced foams. Indeed, the various foams reinforced with B. mori cocoons had 1.4 times higher densities (at 59—63 kg/m3), while the G. postica cocoon reinforced foam was 2.6 times denser at 118 kg/m3. The higher density of the wild G. postica cocoon-reinforced foam compared to the domesticated B. mori cocoon-reinforced foams was due to the higher wall thickness of G. postica cocoon shells (Table 6.1) and consequently a higher filler weight fraction of 88%; the filler weight fraction of B. mori cocoon-reinforced foams was in the range of 57—68%. Often the density of both (unreinforced) cellular foams and conventional microsphere fillers is directly proportional to their compressive strength [15—17]. Therefore, for cocoon-reinforced foams to be substitutes to unreinforced foams, the specific compressive properties of the former would have to exceed that of the latter.
Figure 6.3 The volumetric composition and apparent densities of the various cocoon- reinforced syntactic foams.
The volumetric composition of the various foams is also described in Fig. 6.3. A high volume fraction of macrosphere fillers is attractive as it drastically reduces resin consumption. In the cocoon-reinforced foams, the cocoons had replaced between 40—70% by volume (60—90% by weight) of polyurethane resin. Foams reinforced with B. mori cocoons had a consistent fibrous cocoon volume fraction Vf of 2.7—3.1 %. This figure was much higher at Vf = 7.8% for G. postica cocoon- reinforced foams. The larger volume of G. postica cocoons also resulted in a larger void volume fraction of VV = 61.6 %, yielding a total nonmatrix volume fraction (Vf + vV) of just under 70%. In comparison, foams reinforced with B. mori cocoons with closed ends had a nonmatrix volume fraction of 50—55%. Pierced ends in B. mori cocoons were a potential entry point for the expanding foam and consequently their foams comprised of lower nonmatrix volume fraction of about 40%. The orientation of the nonspherical cocoons was also observed to marginally affect the volumetric composition with foams incorporating longitudinal-oriented cocoons having 5% higher total nonmatrix component than foams with transverse- and randomly-oriented cocoons.
Our results demonstrate that silkworm cocoon macrofillers, which are prolate spheroids, can yield high packing fractions ranging between 40 to 70% in their uncompressed state. This compares to geometrically-permissible, maximum achievable packing fractions of spherical microspheres n/6 ^ 52% (simple cubic packing), n^3/8 ^ 68% (centered cubic packing), and n/3^2 « 74% (hexagonal close cubic packing). Sherwood  has also shown that prolate spheroids represent a local maxima (at 41%) while spheres represent a local minimum (at 38%) for asymptotic packing densities close to the jamming limit. It is noteworthy however that if the particles were to be compressed, spheres allow more scope for an increase in the packing fraction than prolate spheroids . Therefore, silkworm cocoon macrofillers are more suitable for syntactic foams manufactured via casting rather than press molding.