Borate melt structure: a short review

The historical development of the study of borate melt structure is briefly reviewed. Focus is on the short range order, and in particular the temperature dependence of the boron–oxygen coordination number, n_BO(T), and its relationship to melt composition. Early in the 20th century, measurements of borosilicate melt viscosities revealed rich composition– temperature behaviour, and by mid-century it was suggested that this was related to a decline in n_BO(T) with increasing T. Measurements of n_BO(T) began as early as 1972, when Stepanov & Novikov unwittingly demonstrated the fictive temperature, T_f, dependence of n_BO in aluminoborosilicate glass fibres using ¹¹B NMR. Since that time the number of studies of n_BO(T) grew rapidly, including further ex situ ¹¹B NMR, but also spurred by the development of in situ high temperature Raman and NMR spectroscopies, x-ray and neutron diffraction, and statistical, thermodynamic and molecular dynamics modelling methodologies. In recent years ab initio and polarisable ion molecular dynamics, based on density functional theory, and non-resonant inelastic x-ray scattering techniques have been applied to borate melts. This wealth of effort has led to the understanding that B–O coordination typically declines with increasing temperature. The magnitude of such changes are strongly composition dependent, ranging from Δn_BO/ΔT≈0 in e.g. aluminoborates and low modifier sodium borates, up to |Δn_BO/ΔT|≤1×10^–3 K^–1 in e.g. alkali diborates and borosilicates. The possibility of Δn_BO/ΔT>0 remains, and although not observed to date, has been predicted for compositions in the vicinity of sodium metaborate. Directions for future research include: in situ study of the competing effects of temperature and pressure on n_BO(P,T), the interaction of temperature induced B–O coordination change with other variable coordination network cations (Al³⁺, Ga³⁺, Ge⁴⁺, Ti⁴⁺), and further accurate and extensive measurements of n_BO(T) to provide stringent tests of models.