Development of structural descriptors to predict dissolution rate of volcanic glasses: Molecular dynamic simulations

Kai Gong, Elsa OlivettiJournal of the American Ceramic Society 2021 · DOI · arXiv

Abstract

Establishing the composition–structure–property relationships for amorphous materials is critical for many important natural and engineering processes, including the dissolution of highly complex volcanic glasses. In this investigation, we performed force field molecular dynamics (MD) simulations to generate detailed structural representations for 10 natural $CaO–MgO–Al_2O_3–SiO_2–TiO_2–FeO–Fe_2O_3–Na_2O–K_2O$ glasses with compositions ranging from rhyolitic to basaltic. Based on the resulting atomic structural representations at 300 K, we have calculated the partial radial distribution functions, nearest interatomic distances and coordination number, which are consistent with the literature data on silicate‐based glasses. Based on these structural attributes and classical bond valence models, we have introduced a novel structural descriptor, that is, average metal–oxygen (M–O) bond strength parameter, which has captured the log dissolution rates of the 10 glasses at both acidic and basic conditions (based on literature data) with $R^2$ values of ~0.80–0.92 based on linear regression. This structural descriptor is seen to outperform several other structural descriptors also derived from MD simulation results, including the average metal oxide dissociation energy, the average self‐diffusion coefficient of all the atoms at 2000 K, and the energy barrier of self‐diffusion. Furthermore, we showed that the MD‐derived descriptors generally exhibit better predictive performance than the degree of depolymerization parameter commonly used to describe glass and mineral reactivity. The results suggest that the structural descriptors derived from MD simulations, especially the average M–O bond strength parameter, are promising structural descriptors for connecting composition with dissolution rates of highly complex natural glasses.

Introduction

Volcanic glasses constitute approximately 12% of the Earth's exposed continental crust. Their dissolution is a fundamental driver of global geochemical cycles, influencing everything from soil fertility to the long-term regulation of atmospheric $CO_2$. In engineering, these glasses are increasingly used as sustainable alternatives to Portland cement, where their reactivity (dissolution rate) determines the final strength of the concrete.

The challenge lies in their complexity. Volcanic glasses are amorphous and contain a wide variety of oxides ($CaO, MgO, Al_2O_3, SiO_2, TiO_2, FeO, Fe_2O_3, Na_2O, K_2O$). Traditional methods to predict reactivity, such as the NBO/T parameter (the ratio of non-bridging oxygens to network formers), are often too simplistic for such multi-component systems. There is a critical need for a "structural descriptor" that can link a glass's chemical composition to its physical durability.

Datasets and Sample Characterization

The study utilized a dataset of ten natural volcanic glasses with well-documented chemical compositions and experimental dissolution rates measured at pH 4 and pH 10.6. These compositions span a wide range, from ~47 to ~73 wt. % $SiO_2$, covering various volcanic types including rhyolite, dacite, and basalt.

Chemical composition (wt. %) and experimental log dissolution rates for the ten volcanic glass samples.

Methods

The authors employed force field Molecular Dynamics (MD) simulations using the Pedone force field to generate detailed 3D atomic models of the ten natural volcanic glasses. The core innovation is the introduction of the average metal-oxygen (M-O) bond strength parameter ($S_{M-O}$).

Unlike previous descriptors that use fixed dissociation energies for each element, this new parameter is derived from classical bond valence models. It treats bond strength as a dynamic variable that depends on the actual interatomic distances and coordination numbers (how many neighbors an atom has) extracted from the MD simulations. The researchers compared this new descriptor against several others, including the average metal oxide dissociation energy (AMODE), the average self-diffusion coefficient (ASDC), and the traditional NBO/T parameter.

Key Results and Insights

  • Superior Accuracy: The average M-O bond strength parameter achieved an impressive correlation with experimental data, with $R^2$ values of 0.80 to 0.92 via linear regression, significantly outperforming the traditional NBO/T parameter ($R^2 \approx 0.54–0.63$).
  • Structure Over Composition: The MD simulations revealed that while bond distances remain relatively stable across different glass types, the coordination numbers (especially for modifiers like Na and K) vary significantly, which dictates reactivity.
  • The Iron Factor: The study highlights the distinct roles of iron valency. $Fe^{3+}$ acts primarily as a "network former" (strengthening the glass), while $Fe^{2+}$ acts as a "network modifier" (weakening it). Accurately modeling this distinction was vital for predicting the behavior of basaltic glasses.
  • Universal Trend: The model successfully captured the inverse relationship between bond strength and dissolution: as the calculated average bond strength increases, the log dissolution rate decreases linearly.

Discussion and Conclusion

A major insight is that current descriptors assume "congruent dissolution" (the glass dissolves uniformly). The authors suggest that the lower predictive power in basaltic glasses at high pH is likely due to incongruent dissolution, where elements like Fe and Mg leach much slower than Si.

Future work will involve refining force fields to better capture the coordination of minor elements and developing more complex models that account for incongruent leaching behaviors. These structural descriptors derived from MD simulations are promising for connecting composition with dissolution rates in fields ranging from geochemistry to the long-term stability of nuclear waste glasses.

Summary updated

Abstract

Abstract Establishing the composition–structure–property relationships for amorphous materials is critical for many important natural and engineering processes, including the dissolution of highly complex volcanic glasses. In this investigation, we performed force field molecular dynamics (MD) simulations to generate detailed structural representations for 10 natural CaO–MgO–Al 2 O 3 –SiO 2 –TiO 2 –FeO–Fe 2 O 3 –Na 2 O–K 2 O glasses with compositions ranging from rhyolitic to basaltic. Based on the resulting atomic structural representations at 300 K, we have calculated the partial radial distribution functions, nearest interatomic distances and coordination number, which are consistent with the literature data on silicate‐based glasses. Based on these structural attributes and classical bond valence models, we have introduced a novel structural descriptor, that is, average metal–oxygen (M–O) bond strength parameter, which has captured the log dissolution rates of the 10 glasses at both acidic and basic conditions (based on literature data) with R 2 values of ∼0.80–0.92 based on linear regression. This structural descriptor is seen to outperform several other structural descriptors also derived from MD simulation results, including the average metal oxide dissociation energy, the average self‐diffusion coefficient of all the atoms at 2000 K, and the energy barrier of self‐diffusion. Furthermore, we showed that the MD‐derived descriptors generally exhibit better predictive performance than the degree of depolymerization parameter commonly used to describe glass and mineral reactivity. The results suggest that the structural descriptors derived from MD simulations, especially the average M–O bond strength parameter, are promising structural descriptors for connecting composition with dissolution rates of highly complex natural glasses.
BibTeX
@article{gong2021development,
  author = {Kai Gong and Elsa Olivetti},
  title = {{Development of structural descriptors to predict dissolution rate of volcanic glasses: Molecular dynamic simulations}},
  journal = {Journal of the American Ceramic Society},
  year = {2021},
  doi = {10.1111/jace.18265},
}
Figure