Structural model for amorphous aluminosilicates

• Published in: The Journal of chemical physics Vol. 156; no. 6; pp. 064503 - 64519
• Main Authors: Gammond, Lawrence V. D., Youngman, Randall E., Zeidler, Anita, Aitken, Bruce G., Salmon, Philip S.
• Format: Journal Article
• Language: English
• Published: United States 14.02.2022
• ISSN: 0021-9606; 1089-7690
• DOI: 10.1063/5.0079607

An analytical model is developed for the composition-dependent structure of the amorphous aluminosilicate materials (M2O)x(Al2O3)y(SiO2)1−x−y and (MO)x(Al2O3)y(SiO2)1−x−y, where 0 ≤ x ≤ 1 and 0 ≤ y ≤ 1. The model is based on a simple set of reactions and contains a single adjustable parameter p (0 ≤ p ≤ 1). The latter is found from 27Al solid-state nuclear magnetic resonance (NMR) experiments in the regime where R = x/y ≥ 1, aided by new experiments on the magnesium and zinc aluminosilicate systems. The parameter p decreases linearly as the cation field strength of M+ or M2+ increases, as per the observation previously made for the degree of aluminum avoidance [Lee et al., J. Phys. Chem. C 120, 737 (2016)]. The results indicate that as the cation field strength increases, there are less fourfold coordinated aluminum atoms to contribute toward the glass network, and Al–O–Al bonds become more prevalent in a progressive breakdown of Loewenstein’s aluminum avoidance rule. The model gives a good account of the composition-dependent fraction of non-bridging oxygen (NBO) atoms for R ≥ 1, as assessed from the results obtained from solid-state NMR experiments. An extension of the model to (M2O3)x(Al2O3)y(SiO2)1−x−y glasses leads, however, to an excess of NBO atoms, the proportion of which can be reduced by invoking network-forming fivefold coordinated Al atoms and/or oxygen triclusters. The model provides a benchmark for predicting the structure-related properties of aluminosilicate materials and a starting point for predicting the evolution in the structure of these materials under the extreme conditions encountered in the Earth’s interior or in processes such as sharp-contact loading.