Topological pruning enables ultra-low Rayleigh scattering in pressure-quenched silica glass

Silica glass is the most indispensable material in optical communication applications due to its superior optical properties.The transmission loss of silica glass has been reduced over the past 30 years by continuous efforts toward decreasing density fluctuations by lowering of fictive temperature,e.g.,through improvements in processing or doping.A recent study has shown that shrinkage of structural voids by hot compression is a promising way to further decrease the loss.However,an atomic understanding of the pressure effect is still lacking.Here,using molecular simulations,we connect the void shrinkage to topological pruning of silica network.Two physical models predict that the Rayleigh scattering loss of pressure-quenched silica glass can be reduced by>50%when the glass is quenched at an appropriate pressure(4 GPa in our simulation).Our studies are consistent with available experimental results and demonstrate topologically optimized structure can give desirable properties for optical applications of silica as well as other glasses with similar network structure.

Implicit glass model for simulation of crystal nucleation for glass-ceramics

Predicting crystal nucleation behavior in glass-ceramic materials is important to create new materials for high-tech applications.Modeling the evolution of crystal microstructures is a challenging problem due to the complex nature of nucleation and growth processes.We introduce an implicit glass model(IGM)which,through the application of a Generalized Born solvation model,effectively replaces the glass with a continuous medium.This permits the computational efforts to focus on nucleating atomic clusters or undissolved impurities that serve as sites for heterogeneous nucleation.We apply IGM to four different systems:binary barium silicate(with two different compositions),binary lithium silicate,and ternary soda lime silicate and validate our precipitated compositions with established phase diagrams.Furthermore,we nucleate lithium metasilicate clusters and probe their structures with SEM.We find that the experimental microstructure matches the modeled growing cluster with IGM for lithium metasilicate.

Predicting glass structure by physics-informed machine learning

Machine learning(ML)is emerging as a powerful tool to predict the properties of materials,including glasses.Informing ML models with knowledge of how glass composition affects short-range atomic structure has the potential to enhance the ability of composition-property models to extrapolate accurately outside of their training sets.Here,we introduce an approach wherein statistical mechanics informs a ML model that can predict the non-linear composition-structure relations in oxide glasses.This combined model offers an improved prediction compared to models relying solely on statistical physics or machine learning individually.Specifically,we show that the combined model accurately both interpolates and extrapolates the structure of Na_(2)O–SiO_(2)glasses.Importantly,the model is able to extrapolate predictions outside its training set,which is evidenced by the fact that it is able to predict the structure of a glass series that was kept fully hidden from the model during its training.