Crystalline-like ordering in melt-quenched network glasses?

• Evidence for crystalline-like ordering in glasses summarised
• Glass structure theories tested against modern experimental data
• New data for natural opal
• Limitations of the original Zachariasen–Warren random network theory outlined
• Cybotactic theory yields the closest description of the structure of real glasses.

The existence of crystalline-like ordering in melt-quenched glasses is still a subject of debate in the literature, although it is not such an emotive issue as it was during the almost 50 years of the great crystallite vs. random network controversy. However, whereas the latter is now relegated to history, arguments concerning the extent (if any) of crystalline-like ordering in glasses still arise, especially in connection with newly discovered glass-forming systems. The objective of the present paper, therefore, is to present a rational discussion of the present perception of the relationship between the structures of vitreous and crystalline phases and that of the melt from which they are formed. The limitations of the various experimental techniques for studying the intermediate-range order in glasses are assessed, together with the role of chemistry in determining both the short- and intermediate-range order. It is concluded that, whilst there are no unambiguous data that uniquely establish the presence of crystalline-like ordering in glasses, there is strong circumstantial evidence for its existence, e.g. in respect of their devitrification behaviour, and especially of that for glasses having more than one component.