Molecular Dynamics Modelling of Barium Silicate and Barium Fluorozirconate Glasses

Advancement in science and technology has profoundly depended on new types of glass innovation. The glasses that were studied in this project are binary barium silicate glasses, binary barium fluorozirconate glasses, ZBLAN glasses and ?Eu?^(3+) doped ZBLAN glass (the ZBLAN glasses are based on binary barium fluorozirconate glass). The high atomic number of barium in the barium silicate glasses provides high mass and high electron density providing its applications for heat and X-ray shielding. The phenomena such as phase separation in the barium silicate glass will affect its properties of durability and electrical conductivity. On the other hand, ZBLAN glasses have a broad infrared optical transmission window due to the weaker bonding/interaction of F^- ions. Due to the presence of lanthanum in the composition ZBLAN glass can be easily doped with rare-earth ions such as ?Eu?^(3+) giving it many optical applications such as optical amplifier and fibre lasers.

Hence, it's essential to study the structure of these glasses to understand their properties for applications. This thesis used the classical molecular dynamics modelling technique to study the static atomic structure of glass. Generally, fluoride glasses can be formed by totally replacing oxygen atoms in oxide glasses by fluorine atoms. The oxide silicate glasses are common glasses that follow the Zachriasen rules of glass formation but the fluorozirconate glasses do not and lack fixed structural units.

The structure analysis was performed at short-range order (e.g. coordination number, bond length and bond angle), medium-range order (e.g. network connectivity) and long- range order (e.g. phase separation). The related crystals were also simulated in similar conditions to the glasses to compare their atomic structure. Normally at short-range order glass

structure is similar to its related crystal but the differences between them starts from the position and number of next nearest neighbours and increases thereafter. Additionally, the

new methods such as rotational invariants and grid analysis were used to scrutinise structural units and phase separation respectively.

The model of barium silicate glass shows good agreement with experimental diffraction data. The typical bond length and coordination number for Ba were 2.97 Å and approximately 7 respectively. The model did not show any phase separation at low Ba content and hence for further investigation very large models of alkaline earth silicate glasses were studied to see how Ba, Ca and Mg are distributed in the glass. The grid analysis was used to see the distributions which show homogeneity for Ba and Ca and inhomogeneity for Mg cation.

The structural units of fluorozirconate glasses were carefully studied as they do not follow the Zachriasen glass model. The coordination number for Zr was mixture of 7 and 8. The rotational invariant analysis shows that the structural units of ZrF_n polyhedra for coordination number 7 and 8 were similar to Augmented Triangular Prism and Biaugmented Triangular Prism respectively. However, rotational invariant values for BaF_n polyhedra tend more towards random.

The large complex model of ?Eu?^(3+) doped ZBLAN glass was made as it is studied for optical applications. The initial analysis was to observe whether Zr and Ba has similar structural roles as in binary fluorozirconate glass system which they do. Considering the extra elements in ZBLAN glass, Al behaves like a network former and has octahedra structural units whereas La and Na behave like modifiers. In the glass Eu was uniformly distributed with predominantly coordination number of 8 and does not have well defined structural units.