Long term alteration of glass/iron systems in anoxic conditions: contribution of archaeological analogues to the study of mechanisms

The knowledge of glass alteration mechanisms arouses a great interest over the last decades, particularly in the nuclear field, since vitrification is used to stabilize high-level radioactive wastes in many countries. In the French concept, these nuclear glasses would be stored in geological repositories. This multi-barrier system (glass matrix, stainless steel container, low carbon steel over-container, geological barrier) must ensure the durable confinement of radionuclides. But laboratory experiments do not permit to predict directly the behaviour of these materials over typically a million-year timescale and the extrapolation of short-term laboratory data to long time periods remains problematic. Part of the validation of the predictive models relies on natural and archaeological analogues. Here, the analogues considered are vitreous slags produced as wastes by a blast furnace working during the 16. century in the iron making site of Glinet (Normandy, France). The choice of these specific artefacts is due to the presence of particular interface between corrosion products and glass matrix inside the blocks. Thus, they can help us to understand the influence of iron corrosion products from the steel containers on the glass alteration mechanisms and kinetics. A first part of this work concerns the characterization of the archaeological artefacts especially the interfacial area between glass and corrosion products inside cracks using micro and nano-beam techniques (μRaman spectroscopy, FEG-SEM, TEM, STXM...). This study has enabled to suggest an alteration process with different geochemical steps that leads to alteration profile observed. One of these steps is the precipitation of an iron silicate phase. In a second time, leaching experiments were set up on a synthetic glass of similar composition than the archaeological one to understand the first stages of alteration with and without iron. Two phenomena can be observed: silicon sorption and precipitation of iron silicate phase, the latter having a great impact on short term alteration kinetics. All these data underlines the importance of this iron silicate phase that seems to have a significant effect on long term alteration kinetics. In the context of nuclear glass alteration, the precipitation of this phase must be taken into account to predict long term behaviour, as it can maintain high rate of dissolution for a long time. (author)