Insights into Silicate and Oxide Melt Structure from Amorphous, Non-Glass-Forming Materials

Abstract:

Many silicate and oxide liquids of interest in the Earth sciences and in technology cannot readily be quenched to glasses, either because of low silica contents (and hence low viscosity at the melting point and accompanying liquid ‘fragility’) or because of liquid-liquid unmixing at high temperature. Although in-situ, high temperature structural tools have been in use for decades and are rapidly developing, many methods are still most informative for glass samples quenched to ambient pressure and temperature, e.g. high-resolution solid-state NMR.

Amorphous oxides, including alumina and silicate compositions, have widespread technological applications. These are generally deposited by a variety of high-energy sputtering methods, as films of thicknesses of 10’s to 100’s of nm. Using Al-27, Si-29, and O-17 NMR, we have recently shown that for such films, very similar short-range structure is seen in materials made by very different kinetic pathways, such as sol-gel synthesis vs. ion-beam sputtering. This path-independent structure suggests that these materials pass through transient equilibrium states during their formation, probably that of deeply supercooled liquids just above glass transition temperatures. In the HfO2-SiO2 and ZrO2-SiO2 systems, for example, samples have well-resolved O-17 NMR spectra, allowing quantitation of O sites with only Hf(Zr) neighbors (so-called “free” oxide ions), with mixed Hf(Zr) and Si neighbors, and Si only. The observed oxygen speciation agrees well with a simple thermodynamic model of one of the most fundamental equilibria in silicate systems, namely the reaction of bridging (Si-O-Si) and “free” (e.g. OHf3 and OHf4) oxide ions to produce “non-bridging” oxygens (e.g. Si-OHf2). This new approach to sampling such structural equilibria in compositions far outside the range of normal glass-forming liquids may provide new insights into more geological compositions as well, as well as in more general models of silicate melt chemistry.