Hyperspectral Raman line imaging of an aluminosilicate glass

An aluminosilicate glass, which is a model for glass formulations used as dental restorations, was examined by hyperspectral Raman line imaging. The data set consisted of more than 30 000 spectra, which were analyzed by using factor analysis. Nine score images were constructed from the nine significant factors identified. Three factors represent convolutions of noise, background, and offset. The other six factors represent Raman spectra of different bonding environments of the silicate tetrahedron. Three of those factors contain narrow Raman features. These are associated with a fully polymerized silica network, with a silicate tetrahedron with one nonbridging oxygen, and with an alumina-related inclusion or a silicate tetrahedron with two nonbridging oxygens. The last three significant factors contain broad Raman bands representing continua of slightly different bonding environments of silicate tetrahedra with 0-4 nonbridging oxygens. The score images reveal that the glass, although not homogeneous, has few regions with discrete heterogeneities. The different bonding networks commingle and could be interconnected.