5: Structures of Glasses

We know that the properties of three different samples of glasses of the same nominal composition, produced independently in three different laboratories, and annealed in the same manner, will be identical within reasonable limits. Our basic understanding of the solid state then indicates that these glasses have, if not identical, at least very similar structures. It follows that a lack of long range, periodic structure does not, in and of itself, preclude the existence of structure at a level that will control the properties of material. Most of the structural models found in the literature today actually only address the formation of a network. Very few directly address the issues of bond angle distributions, bond rotations, and bond length variations inherent to a random network model. Many of these models only attempt to explain trends in one property or type of spectrum, ignoring masses of other data for the same system which often cannot be explained using the proposed model. As a result, one should be very cautious in accepting any structural model for glasses without a thorough understanding of the limitations of the model and, unfortunately, possible bias on the part of those proposing the model.