A series of glassmaking experiments have been done to explore the influence, if any, of the presence of alkali chlorides in a typical soda-lime-silica glass batch on the final composition of the glass. The experiments have shown that concentrations of chlorides up to the limit of solubility of chlorine in the melt are actively contributing alkali ions to the glass-forming process, and that at higher chloride concentrations in the batch a separate salt melt forms, known as galle. At equilibrium conditions, the alkali ratio in the galle is different from the initial alkali ratio in the batch and differs from the ratio in the co-existing silicate melt. Significantly, there is full exchange of alkali ions between the two melt systems and the addition of pure potassium chloride to a batch containing pure sodium carbonate (=soda ash) results in the formation of a mixed alkali glass and a mixed alkali galle, with complementary alkali ratios. The alkali earth elements (i.e. calcium and magnesium), in contrast, are not taking part in these ion exchange reactions, and seem not to be affected by chlorides present in the batch. These findings are particularly relevant when comparing analyses of different plant ashes with archaeological glasses; the alkali ion ratios between ash and glass are only likely to be similar when no galle is forming together with the glass melt; in the presence of more than a few weight percent of chlorides in the batch, it is likely that the alkali ratio in the glass will be increasingly shifted away from the total alkali ratio in the batch as the chlorine content increases. While the argument is developed specifically for Late Bronze Age halophytic plant ash glasses, the results are likely to be valid in principle also for any other glasses based on halogen-rich batches and containing more than one alkali metal.
