What can one infer about chemical bonding in glasses from their medium-range structural order?

Understanding the atomic structure of amorphous solids or glasses is a perennial, and ultimately insoluble, challenge, since the randomness involved means that structural information can only ever be obtained, from either experiment or structural modelling, in a statistical form, unlike the case for crystalline materials. The particular atomic structure, in a glass or crystal, is dictated by the underlying chemical bonding between the atoms. The question arises therefore, conversely, to what extent can one infer anything about the nature of the bonding in amorphous/glassy solids from some knowledge of the atomic structure? In this focused perspective article, I discuss the case of telluride glassy materials, and show that they represent a special case of bonding in chalcogenide materials. The near-linear axial structural configurations characteristic of certain telluride crystals, also prevalent in the corresponding glasses, and giving rise to a particular degree of linear medium-range structural order, are shown to arise from hyperbonding (e.g. three-centre/four-electron) interactions, rather than being associated with conventional (two-centre/two-electron) covalent bonds, as in other chalcogenide materials. These structural configurations are responsible for the unique ‘phase-change memory’ behaviour exhibited by some telluride materials. An experimental technique is proposed which should be capable of detecting such linear-character, medium-range structural order in telluride glasses.