Glass Reflections
Cambridge 7th to 9th September

Presenting Author:
Ondrej Gedeon

article posted 20 Mar 2015

Ondrej Gedeon

Having graduated in Mathematics and Physics and obtained his PhD in Chemistry, Ondrej Gedeon now heads a research group at University of Chemical Technology, Department of Glass and Ceramics, in Prague, Czech Republic.

His scientific interest is parted into atomistic simulations, mostly Molecular Dynamics of silicate systems, interaction of glass with electron beam, and characterization of materials by means of spectroscopic and microscopic methods, among them mainly SEM, AFM, EPMA, and XPS.

Current projects include:
i) MD simulation of bulk and surface of silicate systems,
ii) modification of glass surface by electron beam, and
iii) characterization of corroded glass fibres.

Reflections about Molecular dynamics and glass transition

Ondrej Gedeon
University of Chemical Technology, Technicka 5, 166 28 Prague 6

Glass transition, despite of many decades effort, resists of any simple, unique and widespread acceptable explanation. The transition is often presented as a set of a few particular effects, not always clear how much they are essential for the transition.

Molecular dynamics (MD) offers a powerful tool that provides, in principle, a complete picture of melt/glass ensemble evolution. Therefore, observation of glass melt, adequately cooled down, should offer the detailed description of glass transition. Having "full information" about the system MD well documents the lack of the microscopic and structural description of glass; putting two pictures next to each other, one snapshot taken above and the other one below glass transition, there is no simple method how to decide which is which (see Figure below). MD nevertheless suffers a few chronically known drawbacks, among them short modelling time, resulting in the extremely fast cooling rate, is the most serious one. A wide-spread view on "MD glass transition" is that it corresponds to high Tool's temperature and that there is is no straightforward way how to relate it with real glass.

Not speaking about the limitation imposed by MD, natural questions arise. Is the observed "MD glass transition" the real one? Or in other words, if we should have been able to prepare the real glass according the recipe given by MD had we observed glass transition? Accepting the view that MD produces real glass transition and that glass transition is a macroscopic/thermodynamic (TD) phenomenon it seems the main obstacle to bridge the MD simulations and TD formalism is the evaluation of configurational entropy (CE) that seems to play a decisive role in glass transition.

In the presented contribution CE entropy based on rings topology, and therefore implicitly including medium-range order, is introduced and discussed. Two set of potentials and various TD conditions were used to study the robustness of the suggested method for vitreous silica. CE relaxation is calculated on base of TD equations and compared with CE obtained directly from MD simulation. Both approaches yield results that can be matched with high accuracy. It is also shown viscosity given by Adam-Gibbs equation for configurational entropy of MD melt is able to fit experimental data.

Reflection symmetry for the presented two structures cannot be of course declared. However, replacing geometry by topology the situation is being less self-evident. The snapshots only document lacking morphisms in non-crystalline substances in comparison with crystal samples where symmetries strongly simplified their description.

Snapshots of vitreous silica (blue - silicon, red - oxygen) as obtained from MD simulations are presented on figure. Atoms from box 44.87 x 44.87 x 4.49 3 are visualised, the last value is for the direction perpendicular to the picture. The left snapshot is structure at temperature below glass transition while the right snapshot corresponds to structure above glass transition. Note only subtle differences between these two pictures.