Electronegativity-Based QSPR Analysis for Understanding Structure-Property Relationships of Glass Materials

Glass materials have been widely used from daily life applications to high technology areas. However, the growing demands from applications, especially those with critical requirements, make it challenging to develop new functional glass materials through the traditional trial and error method. Recently, a methodology by combining molecular dynamics (MD) simulation and quantitative structure-property relationship (QSPR) analysis has shown its ability to correlate the structure and properties of glass, thus promising in predicting unknown glass properties. However, the established MD-QSPR method usually relies on experimental inputs, which limits its capability for applicable systems. In this work, a general-purpose analytical method combined with electronegativity was proposed without experimental inputs in the descriptor construction process. The descriptor in the new method consists of three parts, counting for contributions from composition, structure, and energy, respectively. Experimentally measured glass properties including density, hardness, glass transition temperature (T_g), and elemental leaching rates (η) have been correlated with descriptors based on different structure inputs. In addition, various types of glass have been employed to validate the generality of this method. The results show that the new method can help construct general-purpose models of the structure-property relationship and provide new ideas to design new functional glass.