Change in bond parameters & induced stress in the alteration layer of corroded boro-aluminosilicate glass revealed by structural analysis
Open Access
- Author:
- Kaya, Huseyin
- Graduate Program:
- Materials Science and Engineering
- Degree:
- Doctor of Philosophy
- Document Type:
- Dissertation
- Date of Defense:
- May 12, 2022
- Committee Members:
- Jon-Paul Maria, Major Field Member
Seong Kim, Chair & Dissertation Advisor
John Mauro, Major Field Member
Jorge Sofo, Outside Unit & Field Member
John Mauro, Program Head/Chair - Keywords:
- glass
glass corrosion
nuclear waste glass
optical spectroscopy
ellipsometry
thin films
porous materials - Abstract:
- Although nuclear energy is the second largest source of low-carbon and emission-free electricity, the task associated with storing spent nuclear fuel remains to be resolved. The best available technology for storage is immobilizing nuclear waste in a borosilicate glass matrix which is collected in geological repositories. Due to possible changes in geological conditions over long storage time frames, it is possible that the stored glass encounters with underground water and corrode. Thus, understanding corrosion behavior of borosilicate glass is critical to maintain the integrity of nuclear waste long enough for the radioactivity to decay to acceptable levels. Glass corrosion rates depend on multiple factors including glass composition, pH, temperature, thermal history of glass, and solution composition. Among these factors, glass composition is especially important as it modifies the corrosion behavior non-linearly. For example, Al increases corrosion resistance when added in low amounts but decreases it when added in high amounts. Hence, data for a certain glass composition is not necessarily helpful to understand corrosion behavior of other compositions. Until recently, nuclear waste glass programs of different countries had focused on their own glass compositions, and this slowed down the efforts to understand basics of borosilicate glass corrosion. To consolidate research efforts across the globe, international scientific community decided to focus on the same composition which is a model surrogate glass called International Simple Glass (ISG). The work carried out in this dissertation focuses on the corrosion behavior of ISG for the most part. Despite the extensive research efforts, borosilicate glass corrosion is still poorly understood. This is mostly because glass corrosion proceeds through multiple mechanisms such as ion-exchange, network hydrolysis and precipitation of corrosion products that depend on composition, pH and solution composition in different fashions. These mechanisms might also cause feedback effect for each other. Furthermore, slow corrosion rates and buried interfaces complicate experimental characterization approaches. In order to obtain corrosion layers that are thick enough to characterize long experiment durations are needed. In geological repositories, underground water expected to encounter glass with around neutral pH levels. Around neutral pH levels, aqueous corrosion reactions cause formation of nanoporous surface layers that are referred to as the alteration layer. Although it is not clear how, the alteration layer may act as a passivating layer and decrease corrosion rates. It is hypothesized that this passivation occurs as a result of the re-organization of the alteration layer via hydrolysis and re-condensation reactions. The re-organization can modify the pore structure of the alteration layer and lead to pore clogging. As a consequence, diffusion of water across the alteration layer is hampered. Thus, it is important to understand the structure of the alteration layer and how intrinsic (glass composition) and extrinsic factors (pH and ions present in the corrosion solution) affect it. Conventional techniques to study structure of glassy materials, including nuclear magnetic resonance (NMR) and x-ray scattering, are not adequate to investigate the network structure of the alteration layer. NMR is an inherently insensitive technique, and thus, it requires a high surface area. To obtain high surface area, the glass sample is crushed which can alter its network structure. On the other hand, x-ray scattering probes deeper than the typical alteration layer thicknesses of a few microns. In this dissertation, vibrational spectroscopy and spectroscopic ellipsometry are proposed to study the structure of the alteration layer. Vibrational spectroscopy of glass provides information on the glass network vibrations. In specular reflection infrared (SR-IR) spectroscopy, fingerprint regions for the chemical bonds that make up the glass network of silica and silicate glasses include Si-O stretching (centered around ~ 1100 cm-1) and Si-O-Si bending (centered around ~ 450 cm-1) vibration bands. Additionally, the spectra of boro-aluminosilicate glasses include B-O stretching (centered around ~900 cm-1 for 4-coordinated and ~ 1350 cm-1 for 3-coordinated) vibration bands. Vibrational spectra of glasses with different thermal histories, in other words with different network structures, differ. For instance, the maximum reflectance wavenumber of the Si-O stretching band shifts with the fictive temperature of vitreous silica. Moreover, molecular dynamics calculations revealed that the maximum reflectance wavenumber correlates with the Si-O bond length distribution of silica and silicate glasses. Empirical observations show that the alteration layer of ISG eventually achieves to good passivation decreasing corrosion rates by 3-5 orders of magnitude. When investigated with SR-IR spectroscopy, the network vibrations of the alteration layer of ISG was found to be different than those of the bulk ISG. Upon corrosion, the maximum of the Si-O stretching band in the spectrum is blue-shifted from that of bulk ISG towards those of vitreous silica and porous Vycor glass. According to the results of molecular dynamics calculations for silicate glasses, this blue-shift in the Si-O stretching band corresponds to a decrease in the Si-O bond length distribution. Since deviations in the bond length from the optimum could make the chemical bonds less stable, this could potentially have implications on the apparent corrosion rates. However, it should be noted that glass is not a thermodynamic equilibrium state of matter, and a decrease in the bond length distribution does not necessarily mean a deviation from the optimum bond length. It is likely that during the re-organization the network structure of the alteration layer accommodates a thermodynamically more favorable configuration that is less prone to dissociation. Passivation behavior of the alteration layer depends on the corrosion conditions. For instance, when ISG is corroded at pH 7, the thickness of the alteration layer is ~8 times higher than if it is corroded at pH 9. One factor that explains this difference in passivation behaviors is possible structural differences in these alteration layers that formed at different pH levels. The silanol condensation reactions, which is the driving force for the reorganization in the alteration layer and the subsequent reduction of corrosion rates, is known to be pH-dependent. Then it can be hypothesized that the structures of the alteration layers formed at different pH levels are different. Possible differences in the network structure might cause different chemical reactivities as well as alter the interaction with the water molecules modifying their diffusion. Another implication of the structural differences between the bulk ISG and the alteration layer is the possibility of internal stress in the alteration layer. Chemical strengthening, also known as the ion-exchange process, is a widely used technology to generate stressed layers on glass surface to improve mechanical performance. In ion-exchange, induced stress originates from the size difference between hosting and exchanging ions. During aqueous corrosion, soluble glass constituents leave their sites. To maintain charge neutrality, these sites become occupied by hydrous species (H+ or H3O+). This exchange could also generate stress in the alteration layer. Internal stress might have two consequences. Firstly, under stress, chemical bonds that make up the glass network may become more reactive increasing apparent corrosion rates. Secondly, if this internal stress builds up and reaches to a critical threshold, it might cause mechanical deformation in the alteration layer which could cause cracking or flaking of the layer exposing fresh bulk glass to the corrosion solution. Findings of the carried work revealed that the alteration layer has a network structure that is different than that of the bulk glass, and the structure depends on the corrosion conditions. Hydrous species in the alteration layer, as well as the hydrogen bonding interactions between them, were also found contingent on the corrosion conditions. Another important discovery is that the alteration layer has internal stress. These findings might have implications for the corrosion behavior of boroalumino-silicate glasses and need to be accounted for in the glass corrosion modelling efforts.