Leaching of borosilicate glasses. II. Model and Monte-Carlo simulations

A model for the corrosion of sodium borosilicate glasses is developed. Numerical simulations based on this model are worked out and compared with the results of the experimental companion paper. The glass structure is represented by a diamond lattice on the vertices of which the silicon and boron atoms are randomly distributed. The sodium cations are put in interstitial position and cause lattice depolymerization. Sodium and boron are assumed to be immediately dissolved when they stand at the solid–solution interface. The reaction of glass with water is described by three dissolution rate constants corresponding to silicon with one, two or three siloxane bonds. Dissolved silicon atoms (but not boron neither sodium) are allowed to condense again on the surface. The simulations show that the silicon concentration in solution saturates to a value which is controlled by the ratio of the dissolution to condensation rate constants. The silicon dissolution–condensation dynamics leads to a restructuring of the surface altered layer, which, in most cases, causes a corrosion blocking and limits the release of boron and sodium in solution. Since boron and sodium are thoroughly removed from the surface altered layer, the corrosion can be characterized by the final equivalent leached thickness in boron or sodium. The variation of this quantity is studied as a function of the glass composition and of the model parameters. This study put in evidence two regimes of corrosion, which are controlled by silica dissolution at low boron oxide content, and by boron leaching at high boron oxide content. With an appropriate choice of the silicon reactivity parameters, it is possible to reproduce quantitatively the variation of the equivalent leached thicknesses with the glass composition in two series of experiments at two different S/V ratios (S/V is the ratio of the initial glass surface area exposed to leaching to the volume of the solution). The corrosion blocking is explained in terms of percolation.