Medium-Range Order in Iron Phosphate Glass Models Obtained Using Various Randomization Techniques: A Molecular...

Glasses are known to have medium-range order (MRO), but their link to any experimentally measurable quantity is still ambiguous. The first sharp diffraction peak (FSDP) in structure factor S(q) obtained from diffraction experiments on glasses has been associated with this MRO (∼7-15 Å), but understanding the fundamental origin of this universal peak is still an open problem. We have addressed this issue for a complex glass, i.e., iron phosphate glass (IPG), through atomistic models generated from a hybrid approach (our in-house-developed MC code + molecular dynamics simulation). IPG is a technologically important glass with applications in waste vitrification, bioactive glass, laser glass material, anode material for batteries, etc., and is seen as a strengthened substitute for borosilicate glasses. We performed a comparative study by generating glass models from different initial configurations and randomization techniques. The developed IPG models were first validated with existing data on short-range order (SRO) and MRO through the study of pair correlation functions, bond angle distributions, and coordination number for SRO and rings distribution, FSDP in structure factor, and void size distribution for MRO. The study of coordination environment of oxygen is specifically shown to aid in understanding glass formation through topological constraint theory. Thereafter, to understand the fundamental origin of FSDP in S(q), structure factors were calculated corresponding to the individual ring sizes present in the model. The relative contribution of these individual S(q)'s in the total experimental S(q) is estimated using an inverse fitting approach. The contributions thus obtained directly correlated with ring size percentages in the models for the considered q-range. In particular, the melt-quenched model obtained from the MC model as an initial structure is found to reproduce most experimental features seen in IPG. Through this exercise, we can connect the rings distribution of an atomistic glass model with an experimentally measurable quantity like FSDP in S(q) for a complex glass-like IPG. This gives physical meaning to the rings distribution while also proving that this structural descriptor is a useful tool for validation of MRO in simulation-produced models of glass.