Computational Modeling of Silicate Glasses A Quantitative Structure-Property Relationship Perspective

Computational Modeling of Silicate Glasses A Quantitative Structure-Property Relationship Perspective 

Notwithstanding the huge improvement in experimental methodologies, like X-ray Absorption Fine Structure, Neutron Diffraction, Nuclear Magnetic Resonance, Infrared and Raman spectroscopy, the elucidation of the glass structure still remain a difficult task [4]. In fact, quite often, difficulties in data interpretation of multi-component glasses and apparent contradictory structural evidences from different techniques have to be faced. 

Dipaitmiento di Sdenze Chimiche e Geologiche, Universita degU Studi di Modena e Reggio Emilia, Ma Campi 183,41125 Modena, Italy e-mail menziani unimore.it 

The advent of computational simulation techniques as an accepted component of material development is one of the most important advances in material research. Molecular Dynamics (MD) is nowadays well established as a powerful tool to provide an atomic scale picture of the structure and insight into the behavior of complex glasses in different environments and under different conditions. Recent advances in the construction of interatomic potential allow the correct quantitative evaluation of the numerical values of stractural, mechanical, thermal, electrical and transport properties for simple glasses [9-15]. However, accurate and reliable evaluation of the same properties for multicomponent glasses has proved far more difficult. 

In these cases, i.e. when a direct comparison with experimental observables is not possible, the results of Molecular Dynamics simulations can be used to provide the numerical representation of structure (codified by stmctural descriptors) to be related with the experimental properties of interest through mathematical models. This implies a shift from empirical composition-property relationships to computational structure-property relationships, thus acquiring an immense practical importance in the development of predictive and interpretative models [16]. This approach, called Quantitative Structure-Property Relationships (QSPR), is well known and extensively applied in the area of drug discovery, and chemical toxicology modeling. However, its application in the field of material design is only recently being explored [17-19]. 

---