Ion-exchange mechanisms and interfacial reaction kinetics during aqueous corrosion of sodium silicate glasses

Glass, a ubiquitous material, finds applications ranging from everyday objects to high-tech devices. Its interaction with water, however, leads to property changes that are critical to understand for various applications, from surface modification during processing to long-term corrosion in disposal. While water incorporation can improve glass strength and reduce brittleness, the complex nanoscale interfacial reactions occurring over varying timescales pose a significant challenge to understanding glass-water interactions. Techniques like NMR, SIMS, XPS, FTIR, and Raman spectroscopy, coupled with atomistic simulations, have provided insights. Glass dissolution is a complex, multi-stage process depending on material properties and environmental factors. Commonly observed stages include initial fast ion-exchange, hydrolysis of bridging oxygen bonds, and formation of a porous silica-rich layer. Hydration, hydrolysis, and ion-exchange are fundamental processes, with ion-exchange leading to selective leaching and initial fast dissolution. However, detailed atomistic mechanisms, particularly for ion-exchange, remain elusive. This study aims to elucidate the fundamental mechanisms of ion-exchange and hydration in sodium silicate glass – an archetype for more complex silicate glasses – using ReaxFF-based MD simulations, leveraging recently refined potentials that enable the simulation of chemical reactions and bond formation/breakage. Previous work using this approach identified three stages of water-sodium silicate glass interaction, but a deeper understanding of the ion-exchange process is needed to fully comprehend glass corrosion.

Various computational methods, including ab initio molecular dynamics (AIMD), classical molecular dynamics (MD), and Monte Carlo (MC) simulations, have been used to study glass structures and glass-water interactions. MD simulations, offering a balance between detail and computational cost, are widely employed, while AIMD, though accurate, is expensive for the large systems needed to study glasses. Previous MD studies have focused on sodium silicate glass structures, but accurate modeling of glass-water interfacial reactions requires reactive potentials capable of simulating water dissociation. Several potentials have been proposed, including extensions of the CLAYFF force field and the ReaxFF reactive force field. The ReaxFF approach, fitted to first-principles data, allows the simulation of chemical reactions and bond formation/breakage. Recent refinements of the ReaxFF parameters for Na-Si-O-H systems have improved accuracy and enabled detailed investigations of glass-water interfacial reactions. While most practical glasses are multi-component, sodium silicate glass serves as a valuable model system for understanding the fundamentals of glass-water interactions.

The study utilized reactive force field (ReaxFF) based molecular dynamics (MD) simulations to investigate the interfacial reactions between sodium silicate glass (22.8 Na₂O-77.2 SiO₂) and water. The glass model, consisting of 8001 atoms, was generated using a simulated melt-quench process with a pairwise potential, followed by relaxation with the ReaxFF potential. Two surfaces were created by cleaving the bulk glass and annealed. A water box containing 3000 H₂O molecules was created and placed on top of the glass surface. The system was relaxed at reaction temperatures (300 K, 350 K, 400 K, and 450 K) under NPT and then NVT ensembles. The simulations were performed using the LAMMPS simulation package. The ReaxFF potential included terms for Coulombic, van der Waals, bond, angle, torsion, over- and under-coordination penalties, and lone pair interactions. Key parameters in the ReaxFF potential were obtained through fitting to first principles data from various sources including bulk sodium silicate crystal phases, glass-water interactions, and diffusion energetics. Ion-exchange processes were monitored by tracking interatomic distances between relevant atoms (H, O, Na). Statistical analyses, including calculation of concentrations of involved species and reaction frequencies, were performed to determine reaction kinetics.

The simulations revealed several key ion-exchange mechanisms: 1. **Na⁺/H⁺ exchange:** This occurs between oxygen ions on a single silicon-oxygen tetrahedron or adjacent tetrahedra. The process involves water molecules as intermediates, facilitating proton transfer. 2. **Clustered reaction:** Two non-bridging oxygens (NBOs) react with an adjacent water molecule, forming silanol groups. This mechanism is proposed as a significant pathway for water transport into the glass after the initial surface reactions deplete surface NBOs. 3. **Silanol formation vs. re-formation:** Statistical analyses indicated that silanol formation is more favored than re-formation within the first 3 ns of the simulation. 4. **Water's dual role:** Water molecules act as both intermediates in proton transfer and terminators of clustered reactions. 5. **Two types of silanol re-formation:** Two mechanisms for silanol re-formation were observed: (a) self-exchange (proton transfer between two NBOs on the same silicon tetrahedron) and (b) adjacent-exchange (proton transfer between NBOs on different silicon tetrahedra). Both involve Na⁺/H⁺ ion exchange. 6. **Continuous reactions:** A complex reaction involving multiple steps, including both silanol formation and adjacent-exchange re-formation, resulting in the formation of two silanol groups on adjacent tetrahedra, was also observed. 7. **Reaction kinetics:** The overall reaction rate was analyzed based on the concentration changes of H₂O and NBOs. Two distinct stages were observed: an initial fast stage dominated by water diffusion and silanol formation, followed by a slower stage where silanol re-formation becomes more prevalent. The transition between the two stages is gradual. 8. **Reaction frequencies:** Analyses of the frequencies of the three ion-exchange processes (silanol formation and the two types of re-formation) showed that silanol formation has the highest frequency, followed by adjacent-exchange re-formation, and lastly, self-exchange re-formation. Higher temperatures increased the frequencies of all three reactions.

The findings directly address the research question by providing detailed atomic-level mechanisms of ion-exchange during the aqueous corrosion of sodium silicate glass. The identification of the clustered reaction as a key water transport mechanism is a significant contribution. The observation of two distinct reaction stages, with different dominant processes and rates, provides a more refined picture of glass-water interactions. The simulations demonstrate the importance of both water molecules and the local glass structure in influencing the reaction pathways and rates. The slower diffusion of Na⁺ compared to H⁺ explains the reversibility observed in some proton transfer processes. The study's limitations are acknowledged; the relatively short simulation times (3 ns) restrict the observation to initial stages of corrosion, and the model glass, though an archetype, may not fully represent all aspects of more complex glass compositions. Despite these limitations, the results provide valuable insights into the fundamental processes of glass corrosion, complementing experimental observations and paving the way for more advanced predictive models.

This study used ReaxFF MD simulations to uncover the detailed mechanisms of ion-exchange during the aqueous corrosion of sodium silicate glass. The findings reveal the importance of water-mediated proton transfer, clustered reactions, and the interplay between silanol formation and re-formation. Two distinct reaction stages were identified, differing in their dominant processes and rates. The study highlights the significance of local glass structure and the dual roles of water molecules. Future research could explore the effects of different glass compositions, extended simulation timescales to access later stages of corrosion, and the incorporation of more complex solution chemistries.

The simulations were limited to a relatively short timescale (3 ns), which restricts observation of the long-term behavior of the glass. The relatively high temperatures used to accelerate the reactions might also introduce a small degree of deviation from experimental conditions. The sodium silicate glass serves as a model system; the findings might not be directly transferable to all glass compositions. Finally, the complexity of the system makes isolating individual reaction pathways for quantitative kinetic analysis challenging.