The multicomponent oxide glass as a statistical ensemble of neighboring glassy compounds in the composition space

The rapid and low-cost development of multicomponent oxide glasses is a long-standing issue in condensed matters and materials science. However, traditional experimental approaches based on trial-and-error and empirical rules are time consuming and high-cost. Recently, with the proposal of the materials genome engineering, the "glass genome" has also attracted much attention, which can accelerate the development of multicomponent oxide glasses. Here, we present an alternative and more suitable description of a multicomponent oxide glass as a statistical ensemble of its neighboring glassy compounds (NGCs) in the composition space. Subsequently, the NGCs method was proposed to predict the structural features and physical properties of the multicomponent oxide glass by the statistical averages of structural features and physical properties of the NGCs. What is more, the NGCs method was systematically verified at the atomic-level view through molecular dynamics simulations. The predicted results through the NGCs method were in good agreement with the simulated results in multicomponent oxide glasses (Na2O-SiO2 and Na2O-CaO-SiO2 systems). This work supports the physical picture of multicomponent oxide glasses and further promotes the development of "glass genome," which will accelerate the research of glass materials.