The homogenising process in glass was studied by mixing two different glasses, placed in a known initial distribution. The
initial orientation was to have a horizontal interface between the two glasses because usually the inhomogeneities lie in parallel
horizontal layers in a crucible containing glass unstirred by convection or a mechanical stirrer.
The homogeneity was measured quantitatively using an etching and interferometery technique developed in this work. The degree of homogeneity was defined in terms of the intensity of segregation, which varies from one in a fully segregated system to zero in a fully mixed system. Mixing was studied using three different simple stirrers.
The intensity of segregation was plotted against number of revolutions of the stirrer. From these graphs it can be seen that
the value of intensity of segregation calculated from the etching data has a quantitative significance for following the process of
homogenisinatian.
The flow patterns observed in thin sections of the glass melts, were completely different from those observed in room temperature model experiments. This was because the glasses mixed were of different viscosities and densities, while the model had used liquids, of equal viscosity and density.
The stirrers studied were not very efficient mixing devices but they were designed to compare results in glasses with the previous, room temperature model experiments rather than to be most efficient. The disc stirrer was a convenient standard for comparison of the performance of other stirrers. The addition of a blade to the disc helped in "breaking up the interface "between the two layers. The complex disc stirrer was only efficient when the lower disc was above the liquid-liquid interface. Although no stirrer was extremely efficient, all produced more homogeneous glass than were obtained from unstirred melts.