The Constitution of Glass

A controversial point in the study of glass is the question of its constitution. As glass cools there is no sudden change from fluidity to rigidity as in most cases of solidification, but a gradual decrease in fluidity. The same gradual change is also to be observed in some other properties, for example electrical conductivity, which Ambronn showed to be of an electrolytic nature. This gave rise to the view that glass is a solid solution. Devitrification of glass, which is aided by keeping the glass at softening temperature, thus conferring freedom of movement on the particles, is simply the crystallising out of the various constituents of the solid solution. The continuity of the properties of glass, however, has probably been taken for granted to a greater extent than is justified by experimental evidence. For example, examination of a number of glasses seems to show that at about 70° C. below softening temperature the rate of absorption of heat with rise of temperature undergoes a sudden increase, which is maintained up to a temperature slightly below the softening point. This range coincides with the annealing range within which the glass may be regarded as plastic. The coefficient of expansion remains constant until this plastic region is reached, when a sudden increase to a value four to seven times as great takes place.

According to Zulkowski, the durability of a glass is contingent upon the formation of double silicates, and not of a mixture or solid solution of simple silicates. Deviations from the recognised glass formulae produce poor glass, because they result in the presence of simple silicates which are more easily attacked. The formation of the double silicates probably takes place in the fining stage.

Another view has been put forward, founded on W. and D. Asch's conception of the structure of the silicic acid molecule. A good glass is to be regarded as a supercooled single chemical compound, with perhaps small quantities of impurities present. The molecule of the compound is a very large one, typical glass molecules being, for instance, 5Na₂0.7CaO.36SiO₂,5K₂O.7CaO.36SiO₂,6K₂O.2PbO.2ZnO.2BaO.36SiO₂, and 3Na₂O.3K₂O.3PbO.3CaO.36SiO₂. Substances forming large molecules tend to produce very viscous liquids. It is claimed in support of this theory that, when a glass devitrifies, the crystalline portion has the same composition as the vitreous portion, but this is not always the case. Both calcium silicate, and silica in the form of tridymite, may separate out. Peddle, judging from appearance and optical properties, concluded that the crystals forming in glass were wollastonite, but chemical analysis gave a composition approaching that of the original glass. This, however, he ascribed to occluded glass. In the case of barium glasses a barium silicate, BaSi₂O₅, separates out in large crystals, if present to the extent of 57 per cent, or more. The phenomenon of surface devitrification is probably due, at any rate partially, to the volatilisation of alkali from the surface, leaving a glass richer in silica, but recent observations seem to indicate that adsorbed water has an important influence.

Bradford compares the solidification of glass to the setting of a jelly, regarding glass therefore as colloidal. This view would seem to favour a structure similar to that required by the one-compound theory - that is, a silicon-oxygen net-work. It is well known that gelatine, on rapid stirring, becomes more mobile, owing, it is supposed, to the breaking-up of the linked chains of atoms forming the framework of the gel. It would be interesting to know if rapid stirring of glass which has reached a viscous state would make it more fluid again.

Quincke, from a study of the copper aventurine glasses which contain microscopic crystals of copper in a transparent glass developed in lines of similarly orientated octahedra, or distributed over plane or curved surfaces, concluded that glass has a jelly like structure built up of invisible foam walls separating foam cells, the wall and cell content consisting of a series of liquid phases, each containing several modifications of silicic acid.