Cristallochimie des produits d'altération des vitraux médiévaux: application au vieillissement des déchets vitrifiés

Résumé.

Les vitraux médiévaux, altérés à même les verrières ou exhumés lors de fouilles archéologiques, fournissent un excellent analogue pour suivre le comportement à long terme de déchets vitrifiés altérés en conditions naturelles. Ils permettent en effet d'étudier l'influence des conditions d'altération et de la composition du verre sur sa dissolution, et de déterminer les cinétiques de dissolution des éléments piégés dans le verre sur des périodes de plus d'un millénaire. Ce travail montre de plus que la dissolution des verres silico-calciques et alcalins est incongruente et que la minéralogie des pellicules d'altération joue un rôle primordial quant au comportement à long terme des déchets vitrifiés, puisque même si le verre s'altère, les produits secondaires peuvent, dans certains cas, piéger la majorité des éléments polluants.

Abstract.

One of the major problems of modern society is to render inoffensive ultimate wastes produced by municipal, industrial or nuclear activity. These ultimate wastes contain high amounts of toxic elements and have to be stabilised before storage. Among the actual solidification–stabilisation processes, vitrification offers many advantages, such as a reduction of volume, no porosity and good chemical durability. However, before using silicate glasses to stabilise ultimate wastes, it is necessary to ensure the long-term behaviour of these materials and to study their solubility and the possible dispersion of polluting elements into the biosphere. Although many leaching experiments have been carried out, knowledge of the behaviour of vitrified wastes weathered over periods of 100 to 10,000 years remains sparse. Previous work (Sterpenich et Libourel 2001) has shown that medieval stained-glass windows provide a valuable opportunity to study the alteration and long-term behaviour of vitrified wastes under natural conditions of weathering. This study allowed the quantification of the influence of conditions of alteration and of glass composition on dissolution kinetics through the calculation of mean dissolution rates of toxic elements over periods of time of up to one millennium. However, the mineralogy and the chemistry of altered-layer stained glasses are complex. An essential second step is the study of the crystal chemistry of pollutants in order to determine the long-term stability of the phases containing them and thus to ensure the long-term containment of potentially toxic elements.

This work focuses on the identification of mineral phases formed through alteration of medieval stained glasses, either on windows or buried in the soil. Medieval stained-glass windows allow the study of the behaviour of a glass submitted to different alteration conditions over periods of more than one millennium. Due to their chemical complexity, medieval stained glasses permit the quantification of release rates of elements, in particular pollutants belonging to the vitreous silicate matrix. Medieval stained glasses altered on windows by meteoric waters develop an altered layer made of a hydrated silica gel, generally depleted in transition and heavy metals. Secondary products (sulphates and carbonates) precipitated on the surface or in the microcracks do not trap significant amounts of toxic elements. In contrast, stained glasses altered in the soil have a thick altered layer (up to more than 1.5 mm) comprised of a hydrated silica-gel and a hydroxyapatite-like phase [Ca₅OH(PO₄)₃]. A third phase precipitates in the altered layer close to the free surface of the glass or in the rare microcracks. It is a phase rich in manganese, probably in the form of manganite (γ-MnOOH) or hausmannite (Mn₃O₄).

The dissolution rates of elements are strongly influenced by the mineralogy of the altered layer. The partition coefficients were determined for several polluting elements for potassic buried stained glasses. Transition elements and heavy metals are strongly partitioned into the manganese-bearing phase. Lead, arsenic and zinc are also present in the phosphate phase. Hydrated silica-gel, consisting of around 80 wt% SiO₂ (from an anhydrous composition) principally traps aluminium, iron and sometimes a proportion of copper, zinc and lead.

In summary, the study of medieval stained glasses shows that the dissolution of calcium- and alkali-bearing silicate glasses is incongruent and the altered layer has a protective capacity which depends on the alteration conditions (secondary phases trapping pollutants, cracking phenomena, thickness of altered layer limiting the diffusion of leached species). The mineralogy of altered layers is also of prime interest for the long-term behaviour of vitrified wastes, as even if the glass alters, secondary phases can trap most of the potentially toxic elements.

This study also shows that particular attention must be paid to the choice of glass composition, in particular for vitrified waste formulation. For example, the presence of phosphorus in fresh glass favours the precipitation of secondary minerals which retain polluting elements in the matrix.