The alteration layer that forms on the surface of the nuclear waste glass plays a crucial role in controlling the long-term dissolution behavior. To accurately predict this behavior, it is essential to understand the chemical structure and mass transfer characteristics within the alteration layer. Unfortunately, obtaining comprehensive information about the alteration layer from experimental approaches is limited due to the difficulties in analyzing nanometer-scale surfaces. In this study, molecular dynamics (MD) calculations, which can model realistic atomic structures, were conducted to evaluate the alteration layer’s structural properties and the water molecules’ kinetic properties.

The structure within an alteration layer may change significantly depending on environmental conditions such as water content. Therefore, we created several alteration layer models with water content as a parameter and systematically investigated whether the characteristics of the alteration layer changed. The water content was set in 10 wt% increments from 10 to 100 wt%, and the ratio of SiO2 and H2O was adjusted at each water content so that the total number of atoms was 9,000 atoms. MD calculations were performed for each model by annealing at 6000 K under constant volume until the structure reached full equilibrium, then cooling at a rate of 20 K/ps to obtain the equilibrium structure at 300 K. The atomic structure was sampled and used in the analysis to evaluate the silicate backbone structure and water mobility. The reactive force field (ReaxFF) optimized to simulate the glass-water interface was used to build the alteration layer model.

Structural analysis of the silicate framework revealed that with increasing water content, the Si bridging structure was changed depending on water content. The Q4 and Q3 structures, which have four or three cross-linked oxygen atoms, were dominant in the model with low water content. In contrast, the high water content structure showed that the bridging structure of the silicate framework was almost broken, and Si species about 80% is in the Q0 structure without bridging oxygen. Pore size distribution and the volume fraction filled with water were examined to characterize the porous structure. The pore was defined as the excluded volume with all water molecules in the cell.

All pore size distributions with different water contents were found to be Gaussian line shapes, and the mean pore size was increased with increasing water content. The mean pore radius for the model with the lowest water content (10wt%) was 1.8%, which is expected to be constrained water dynamics in such a small pore. Water cluster analysis on the basis of hydrogen bonds between water molecules was performed to investigate the aggregated state of water.

The results revealed that most of the water was isolated within the pores with lower water content, and the translational movement of water was significantly limited. To evaluate the translational motion of water, the self-diffusion coefficient was calculated from the mean square displacement (MSD) of water. The diffusion coefficients dramatically increase with increasing water content. To organize the relationship between diffusion coefficients and porous structure, the diffusion data were fitted with linear function against mean porosity. The fitting successfully explained the diffusion data, meaning that the geometric parameters of the porous structure for alteration layers can model the mass transfer.