An investigation of the effect of sand grain size on the refining of a pure soda-lime-silica glass in a laboratory furnace
JSGT 1958 V42 T020-T031
Sixty gram melts of glass of composition (weight per cent) 73·5 SiO2, 10·0 CaO, 16·5 Na2O, were founded for varying times at 1400°C using four different sand grain sizes from 0·34 to 0·095 mm. Discs were cast from the melts and measurements of the total number of seed and their size distribution made on photographic enlargements of the annealed discs. The results showed that the total number of seed decreased exponentially with time. As the sand grain size decreased the seediness at the batch-free time increased, but the rate of refining also increased, and the most homogeneous and least seedy glass was produced most quickly by using the finest sand. The data enabled the rates of disappearance of particular sizes of seed to be derived and again the relation between number and time was exponential. The refining was therefore not controlled by the rise of bubbles through the melt. Relatively rapid convection currents were shown to exist in the melt but later work (not fully reported here) showed that the elimination of convection did not affect the refining. It was concluded that the exponential relation was not due to convection but to some other process that caused small seed to disappear much more rapidly than predicted from Stokes' law. The size of sand grains used affected the rate of disappearance of a particular size of bubble as well as the number and size distribution at the batch-free time. The results for the series using the finest sand showed an anomaly that could not be accounted for solely by the change in homogeneity of the melts.
Michael Cable
