Computer simulations have developed to a stage to play an important role in glass research. Simulations have gradually moved from university and academic labs to industrial environments and, in some situations, have demonstrated successes for new glass product development. However, there remain many challenges such as effectiveness of potential models, experimental validation, size and time scale limitations and so on in modeling glasses and amorphous materials. These challenges are the focus of this international workshop following the tradition of the first workshop held in Strasbourg, France July 2-4, 2012.

In the field of atomic-scale simulation of glasses, a strong interplay exists between classical molecular dynamics (CMD) and first-principles molecular dynamics (FPMD) computational strategies, both used (under different working hypothesis) to interpret and complement experimental data. Such a synergy between different approaches has unique features in the area of condensed matter modeling. In order to understand its origins, let us focus on a specific example, i.e. the comparison between the measured partial structure factors and the partial structure factors calculated by using CMD. In addition to the wealth of information resulting from such combined analysis of the atomic structure, this same comparison is quite insightful on the quality of the empirical potentials that are employed. The FPMD approach has highlighted the enormous challenge inherent in a characterization of the glassy state in terms of atomic-scale properties, i.e. interatomic distances, coordination numbers, bond angles, structural units and ring statistics. In specific cases, the FPMD schemes have proved to be (at least partially) successful, provided that a suitable exchange-correlation functional is selected in DFT. This choice of functional depends on the nature of the chemical bonding (e.g. the relative localization of the electron density) in the material under consideration.

Computational materials science can be used to model glasses effectively since it provides the atomic-scale tools that are necessary for accounting for their chemical differences. However, this is only true if one is willing to cope with the possible limitations arising from the empirical character of CMD by resorting to the more elaborate FPMD. Along these lines, the first purpose of our workshop will be to identify current achievements and properly assess the state-of-the-art in the atomic-scale modeling of glasses. We intend to focus on representative examples of systems where the detailed information on the structure provided by molecular dynamics has been instrumental in making progress in the area of glass science.

However, the above encouraging statements on the success of molecular dynamics methods hide several shortcomings and stumbling blocks that are related to a realistic simulation of glasses. In this context, the second (and by far the most challenging) purpose of this workshop will be to discuss, understand and propose solutions to a set of open issues (see the section devoted to the objectives) that all have, in common, the need for a quantitative knowledge of the structural properties of a glass in order to account for their other properties.