Modeling Interfacial Glass‐Water Reactions: Recent Advances and Current Limitations

Abstract

Describing the reactions that occur at the glass-water interface and control the development of the altered layer constitutes one of the main scientific challenges impeding existing models from providing accurate radionuclide release estimates. Radionuclide release estimates are a critical component of the safety basis for geologic repositories. The altered layer (i.e., amorphous hydrated surface layer and crystalline reaction products) represents a complex region, both physically and chemically, sandwiched between two distinct boundaries-pristine glass surface at the inner most interface and aqueous solution at the outer most interface. Computational models, spanning different length and timescales, are currently being developed to improve our understanding of this complex and dynamic process with the goal of accurately describing the mesoscale changes that occur as the system evolves. These modeling approaches include geochemical simulations (i.e., classical reaction path simulations and glass reactivity in allowance for alteration layer simulations), Monte Carlo simulations, and molecular dynamics methods. Discussed in this manuscript are the advances and limitations of each modeling approach placed in the context of the glass-water reaction and how collectively these approaches provide insights into the mechanisms that control the formation and evolution of altered layers. New results are presented as examples of each approach.

Atomistic computer simulations of water interactions and dissolution of inorganic glasses

“…Application of MC methods to glass dissolution involve solving coupled hydrolysis/dissolution and condensation reactions. 3,75 MC simulations which include a time component are term kinetic MC (KMC). KMC evolves the system from state to state based on reaction rates and can enable simulations of infrequent events and process longer time scales compared to MD simulations.…”

“…76,84 Kerisit et al 89,90 have further developed the MC simulation methodology for glass dissolutions to take into consideration glass structure features such as NBO and corrosion conditions such as dynamic flow-through experiments. 3 In simulations of flowthrough corrosion of borosilicate glasses, it was found that at high flow rate conditions, thick alteration layer was formed and glass dissolves congruently. At low flow rates, highly incongruent dissolution was observed with formation of a permanent protective layer similar to static dissolution conditions.…”

“…The initial stage is interdiffusion of proton or hydronium ions from the solution with sodium or other alkali ions in the glass. 3 This is followed by the hydrophilic attack of water on the Si-O-Si or Si-O-Al linkages that lead to hydroxylation of the silicate glass network. The remaining hydrolyzed glass skeleton then undergoes condensation and repolymerization to form the hydrated nanoporous silica rich gel layer which can be protective, decreasing dissolution to a residual rate.…”

“…10,11 Further development resulted in the formation of the GRAAL model (Glass Reactivity with Allowance for the Alteration Layer) of nuclear waste glass dissolution, which include the formation of a passivating reactive interphase (PRI) formed through hydrolysis and condensation in the hydrated gel layer. 3,4,12 The PRI has been identified as a barrier to water transport into the glass and solvated ions into the solution, resulting in transport inhibition effect on glass alteration. While evidence of PRI is limited, Gin et al 13 used nanoSIMS to identify a dense region 0.5 micron thick between the glass and gel regions, reported to be the first direct evidence of PRI from experiment.…”

Transport- and surface reaction-controlled SON68 glass dissolution at 30°C and 70°C and pH=13.7

“…Since the reaction affinity model based on the Transition State Theory (TST) cannot accurately describe all of the processes that control glass corrosion in the residual rate regime, transport-limited models such as GM2001 (Grambow and Mueller, 2001), r(t) (Ribet et al, 2001), BRAG (Aertsens, 2007), GRAAL (Frugier et al, 2008) and STORM (Pierce and Bacon, 2011) models were developed. In transport-limited models surface layers form a barrier to the transport of water to the reacting glass surface and/or the transport of reaction products away from the reacting glass surface (Vienna et al, 2013;Pierce et al, 2014).…”