The Science & Technology
of Glass
Cambridge - Monday 4th to
Wednesday 6th September 2017



Daniel T. Bowron
<daniel.bowron@stfc.ac.uk>

article posted 03 August 2017

Daniel Bowron is the leader of the Disordered Materials Group at the ISIS Neutron and Muon Source in Oxfordshire, UK, where he undertakes research into atomic and mesoscale structure-property relations in liquids and amorphous materials. His scientific and technical interests include the design, construction and exploitation of advanced Neutron and X-ray scattering/spectroscopy instrumentation, the development of novel data analysis software, and atomistic computer simulation and modelling of experimental data.

http://www.isis.stfc.ac.uk/People/daniel_bowron5212.html


Building experimentally consistent atomistic models of multicomponent glasses using diffraction data
Daniel T. Bowron
Science and Technology Facilities Council, ISIS Neutron and Muon Source, Rutherford-Appleton Laboratory, Harwell, Didcot, OX11 0QX, UK

The extraction of atomically resolved structural information from neutron and X-ray diffraction data collected from multicomponent glasses is a challenging problem. Although diffraction experiments are a powerful probe of atomic pair correlations in a material, a single measurement only provides structural insight via the total pair distribution function: a complex weighted sum of all the chemically specific site-site pair distributions that characterise a system. Formally, a material formed from N atomic components requires N(N+1)/2 partial pair distribution functions to properly describe its structure. In spite of the complexity of the total pair distribution function, it does provide a fundamental constraint on the structure of a material of interest, and any derived structural model must be consistent with this experimental data if it is to be considered acceptable. The challenge for the glass-structure scientist using these experimental methods, is thus to build such an acceptable model so that it can then be interrogated to provide insight into the likely atom-scale underpinnings of a materialís physical and chemical properties.

Over the past three decades, several methods have been developed for building atomistic models of structurally disordered materials that are consistent with experimental diffraction data. In this presentation, I will outline and illustrate how one such method, Empirical Potential Structure Refinement (EPSR), can be used to address this challenge. EPSR makes use of a classical Monte Carlo simulation engine to generate a three-dimensional atomistic model of a disordered material that is consistent with supplied neutron and/or X-ray diffraction data. The technique makes use of classical pairwise atomic interactions, based on Lennard-Jones potentials and Coulomb forces, to encode underlying physical and chemical constraints into a model. These interaction potentials are further combined with knowledge of a materialís compositional stoichiometry, density and any other available structural information, such as the structure of constituent molecular units, to provide a robust framework in which a model is produced. Once refined against the supplied experimental data, the resulting atomic configurations generated by EPSR can be interrogated to extract structural information on any specific partial pair distribution function, bond angle or coordination number distribution of interest.