In a paper about CRT-panel foaming with Manganese oxide, they study the foaming of the glass with a high temperature microscope and made important conclusions about the transition from closed to open cells. By using the percolation limit, they define a temperature window for closed cells.
In the last paper, they foam CRT panel glass, soda lime glass and mixtures of these glasses with Fe2O3. CRT panel glass has the advantage to have a lower thermal conductivity (0.97 W/mK) than soda lime glass (1.1 W/mK) and on top of that, it should be rather cheap because there is no other use possible and the cellular glass manufacturer is paid to absorb the glass.
But in some cases, the glass in contact with Fe2O3 is prone to crystallization, which induces open cells. According to the paper, calcium phosphate is a crystallization inhibitor in this case and the cells remain closed. In our opinion, the use of a crystallization inhibitor to avoid open cells with a mineral foaming agent is the major contribution of this paper. It opens the door to 0.045 W/mK cellular glass at mineral wool prices with equivalent densities.