Abstract
The peak of the variation in the critical strain energy release rate G 1Ic with porosity, as previously measured on sintered glass beads, is modelled. The approach is an adaptation of an existing model, based on the description of the actual microstructure as resulting from the densification of an initial stacking of spheres. A previous model made use of a unique reference toughness K Ic0 of a hypothetical bulk material and as such was valid only if the micromechanisms of crack propagation were the same over the whole range of porosity. In the present case, the observed change in fracture micromechanisms with porosity in sintered glass beads is considered. The analysis of the microstructure and crack path morphologies enables specific toughnesses to be attributed to the mechanisms observed at the extreme values of porosity: crack pinning by pores at low porosities and rupture of the sintering necks at high porosities. The transition from one to the other is expressed by an evolution with porosity of the virtual reference toughness K Ic0 . The two-parameter model fits the experimental data and also predicts the effect of bead size on the G 1Ic peak extent. It is finally proposed that a similar approach could be used for other materials, other properties and other microstructural parameters