Glass corrosion

Abstract

Borosilicate glasses are currently used for the immobilization of highly radioactive waste and are materials of choice for many biomedical and research industries. They are metastable materials that corrode in aqueous solutions, reflected by the formation of silica-rich corrosion rims. Until now, there is no consensus in the scientific community about the reaction and transport mechanism(s) and the rate-limiting steps involved in the corrosion of silicate glasses. Most models have the basic assumption in common that ion release from the glass network is occurring via interdiffusion and that the glass network itself is not being disrupted, only modified. On the contrary stands the interface-coupled dissolution-precipitation (ICDP) model, which first was developed for mineral replacement reactions and was recently adapted to glass corrosion. It is based on the congruent dissolution of the glass network that is spatially and temporally coupled to the precipitation and polymerization of silica, forming the amorphous corrosion rim. The dissolution of a radionuclide-binding glass matrix is naturally a sensitive issue for the safe disposal of vitrified high-level nuclear waste. A sound description of the reaction mechanisms and the identification of the rate-limiting steps is essential to predict the long-term corrosion of silicate glasses, particularly when time scales must reach several thousand to millions of years as required by safety regulations for a nuclear repository.<br /> In three project studies, the broad spectrum of borosilicate glass corrosion was investigated from the first surface precipitates at an inward-moving solution-glass interface, over the dynamic development of the corrosion rim itself, and the tracing of individual species within the corrosion rim and across the rim-glass interface. Results of atomic force microscopy and single-pass flow-through experiments deliver strong evidence for a significant compositional difference between the surfacial and bulk solution. Hence, local supersaturation of the interfacial solution with respect to amorphous silica at the glass surface can explain how precipitation of silica can occur when the bulk solution is still undersaturated. To study the dynamic development of the corrosion rim in space and time, a novel fluid cell-based <i>in situ</i> Raman spectroscopy method devised. This method allows monitoring the congruent dissolution of the glass and simultaneously the precipitation and polymerization of the silica-based corrosion rim at elevated temperatures in space and time without the need to terminate the running experiment. For the first time, the formation of a silica- and water-rich zone at the interface between glass and corrosion rim could be observed in operando. Commonly such zones were identified post mortem as gaps or cracks between pristine glass and corrosion rim, and, hence, referred to as result of sample drying. However, these results show that these discontinuities are a primary feature of the corrosion process itself and that the dissolution process must proceed within the therein present interfacial fluid. Lastly, multi-isotope tracer (²H, ¹⁸O, ¹⁰B, ³⁰Si, ⁴⁴Ca) experiments were performed on pristine and already corroded glass monoliths of different glass compositions. Results of transmission electron microscopy and analyses by nanoscale secondary ion mass spectrometry reveal a nanometre-sharp interface between the silica-based corrosion rim and the glass, where decoupling of isotope tracer occurs, while proton diffusion and ion exchange can be observed within the glass. Moreover, a dense layer was observed between the corrosion rim and glass, which appears to the quenched silica-rich interfacial (pore) solution, which was observed in operando in the <i>in situ</i> Raman experiment.<br /> As these new findings cannot be explained by solid-state diffusion processes, nor the classical ICDP process accounts for ion exchange in the glass, a unifying mechanistic model is proposed, which accounts for all critical observations so far made on naturally and experimentally corroded glasses. The main corrosion rim forming process is based on the interface-coupled glass dissolution-silica precipitation reaction. However, a diffusion front over several tens of nanometres may evolve in the glass ahead of the dissolution interface once transport limitations cause the dissolution rate to become slower than the diffusion rate of individual species (D_H = 1.3 × 10^-23 m² s^-1).