Bond valence in silicate glasses

Abstract

In contrast with crystals, glass structure cannot be unambiguously determined by experimental techniques, and molecular dynamics (MD) simulations are often used to model such systems. Ion–ion interactions are modeled by empirical potentials, which are validated by comparison between the structural information extracted from experimental results and the simulated structure. Bond valence (BV) theory is often used to check the reliability of experimentally determined crystal structures. However, it has not been applied in a similar fashion to the results of MD simulations of disordered systems. To illustrate the use of BV theory in silicate glasses, BV calculations were performed in a 1500 atom-3D-periodic MD simulation box of a glass of composition CaO–FeO–2SiO 2 (mol%) with two sets of Born–Mayer–Huggins potentials. The structural models are analyzed from the point of view of BV theory. The BV sums around the 1500 atoms show that the BV model is globally verified. In addition, we show that the BV sums are sensitive to changes in MD potential parameters and can thus be used to test the plausibility of the MD model of a silicate glass. This new strategy opens possibilities in the derivation of more robust and constrained structural models of silicate glasses and melts.