Diffusion models for the early-stage SON68 glass dissolution in a hyper-alkaline solution

• Published in: Applied geochemistry 2019-12, Vol.111, p.104439, Article 104439
• Main Authors: Liu, Sanheng, Ferrand, Karine, Aertsens, Marc, Jacques, Diederik, Lemmens, Karel
• Format: Article
• Language: English
• Subjects: Diffusion-controlled dissolution; Hyper-alkaline solution; Nuclear glass dissolution; Surface area to solution volume effect
• ISSN: 0883-2927; 1872-9134
• DOI: 10.1016/j.apgeochem.2019.104439

Understanding the dissolution of nuclear waste glass in hyper-alkaline conditions is critical to predicting its safe disposal in cement-dominated disposal conditions. In this work, solution data from the early-stage SON68 glass dissolution in a KOH solution (pH 13.5 and 30 °C) were interpreted with models based on silicon (Si) diffusion controlled stoichiometric dissolution of the glass. The solution data show a linear increase of boron (B) concentration with the square root of time and a decrease of the glass dissolution rate with an increase in the ratio of glass surface area to solution volume (SA/V). Three models from the literature, based on either aqueous pore diffusion or solid diffusion with/without resistance of the surface boundary layer, can fit the evolution of solution concentrations for the major elements of the glass equally well. However, the fitted Si diffusion coefficients differ by several orders of magnitude. Since all these fitted diffusion coefficients are lower than diffusion coefficients generally reported for aqueous pore diffusion, they are more indicative of solid diffusion processes. These results show that even under such hyper-alkaline conditions where dissolution and precipitation reactions are usually considered to prevail over solid diffusion, solid diffusion can still be an important dissolution mechanism. Those two dissolution mechanisms do not necessarily contradict each other. It is postulated that incorporation of potassium (K) from the solution onto the glass surface provides a diffusion barrier against the release of glass components into the solution. Therefore, the early-stage glass dissolution rate under hyper-alkaline conditions should be lower than the predicted glass network dissolution rates extrapolated based simply on pH effects. •Both aqueous pore diffusion and solid diffusion models fit a given set of solution data equally well.•Aqueous pore diffusion models result in SA/V dependent diffusion coefficients.•Solid diffusion models result in SA/V independent diffusion coefficients.•Solid diffusion may control glass dissolution even under hyper-alkaline conditions.•Potassium incorporation into the glass surface layer provides some protective effects on glass dissolution.