One hundred years ago, world-famous scientist A. A. Lebedev performed a set of classical measurements on annealed optic crown glasses. He found that these glasses exhibited characteristic endothermic effects in a particular temperature range. To explain these phenomena, Lebedev proposed a hypothesis that the glasses contain tiny quartz crystals. This initial hypothesis was quickly disapproved, and the origin of the endothermic effect of glasses remains an unsolved puzzle. This work uses recently proposed nanoflake model of silica glass structure to explain the endothermic effect of various glasses. The new model differs from the popular continuous random network theory in that it emphasizes the medium-range ordering structure of glasses. According to the nanoflake based theory, the endothermic effect of glasses is caused by the transition from ordered one-dimensional structures into disordered structure in glasses. The new theory also predicts that the temperature range of the endothermic effect is dependent on both glass composition and cooling rates during glass formation.

The anomalous density-temperature relationship of vitreous silica with low hydroxyl content is explained by the formation of medium-range ordering structure in the glass transition process. The ordered medium-range structure has the shape of a "nanoflake" and consists of two layers of SiO₄ tetrahedra, bonded by O atoms located in the middle of the structure. The nanoflakes interact with their surrounding structures through both covalent chemical bonds and van der Waals bonds. In the formation of the van der Waals bonds, the orientation of SiO₄ tetrahedra can change, which results in an increase in distance between the nanoflakes and their surrounding structures. Thus, there is a slight volume enlargement associated with the formation of nanoflakes. Since the nanoflakes' formation starts at a temperature near 1480 °C, and the population of the nanoflakes grows continuously as temperature decreases until about 950 °C, the bulk volume of silica glass increases in the temperature range from about 1480 °C to 950 °C. Therefore, the density anomaly of silica glass can be explained as a byproduct of forming of medium-range ordering structure in the silica glass transition.

To investigate the possible structural differences between silica glass fibers and bulk silica glasses, electron energy loss spectroscopy (EELS) has been used to study the short-range and medium-range structures of both forms of silica glasses. The short-range structure of silica glass, such as the coordination and symmetry, was investigated by the energy loss near edge structure (ELNES) of Si L2,3-edges. The ordering structure in the medium-range was analyzed by the exponential optical absorption edge also known as the Urbach edge of the glasses. The optical absorption data were obtained from the low energy loss spectrum of EELS through Kramers-Kronig analysis. The results show that silica fiber has the same short-range structure as the bulk specimen, but is significantly more disordered than the bulk glasses.

Correlative microscopy is used to compare the performance and nanophase structures of soda-lime-silica, fused silica and borosilicate glasses using Vickers indentation crack analysis and the transmission electron microscopy (TEM) Fresnel contrast method. It is found that the observed indentation cracking behaviour is correlated to the nanophase separation structure of these glasses. The so-called "normal" cracking behaviour of soda-lime-silica glass is influenced by its spinodal nanophase separation; while the "anomalous" cracking behaviour of fused silica is due to the uniform single phase structure. Borosilicate glass has a droplet nanophase separation and shows "intermediate" cracking behaviour. These results indicate that in order to produce low britleness glasses it is important to control nanophase separation structure of a glass.