The durability of glass is of crucial importance for nuclear waste storage. Indeed, one solution of waste confinement studied in France consists in vitrifying the radioactive elements in a glass matrix. As the contact of the radioactive package with the geological water may be a critical issue, it is essential to understand the mechanisms of glass alteration or corrosion by water. The experiments done on BM32 GMT were based on the study of a glass reference derived from the glasses of the french atomic nuclear agency (CEA) dedicated to the packaging by eliminating a few minor elements in order to isolate the specific effect of the main cations (Ca, Zr, Al) which are incorporated in basic ternary borosilicate glass (SiO2/B2O3/Na2O with respectively 70, 15 and 15 % molar composition).

The previous experiments published so far were mostly carried out on glass powders by SAXS or SANS [1,2]. It was shown that a porous layer forms at the glass surface due to the leaching of the soluble elements (Na, B, …). Consequently, the dissolution of these soluble elements proceeds at the interface between the altered layer and the intact core. However, for high degrees of alteration and small grains, the alteration process can be slowed down by the decrease of the quantity of interface. Thus, alteration of glass monoliths was studied in situ at 90°C by x-ray in addition to glass powder in order to look at the early stages of the alteration kinetic. A specific cell was designed for this purpose. Some samples were elaborated few weeks before the X-ray experiment and characterized in cold water.

Imgspalla

Fig.1: Schematic structure of the monolithic glass at an intermediate step of alteration. With high photon energy (18 keV) in order to probe deeply up to the core intact glass thickness of the layers, their electronic densities and interfacial roughness between them can be extracted from X-ray reflectivity curves.

Morphological characterization consists in irradiating the glass under grazing incidence in order to enhance the sensitivity to the altered layer on the top of the intact core layer and collect scattered intensity at small angles near the direct and reflected beam (GISAXS). Inspired by the small angle x-ray scattering analysis, the developed methodology relies on the measurement of the absolute intensity in this reflection geometry. GISAXS patterns are then measured in one shot with a 2D CCD camera. By assuming an homogeneous altered layer in composition and porosity, the radial averaged scattered intensity is simply given by a geometric factor (corresponding to an integral product of the transmission factors in the Distorted Wave Born Approximation description ) that can be evaluated through measurements for a couple of incidence angles [3]. This last factor depends only on the product of the altered layer thickness and the difference of atomic cross section of the layer and the solvent. The altered and dissolved layers thickness, the porous volume and the specific surface area can be determined. The composition dependence of the various results can be then discussed (articles redaction in progress).

Imgspalla2 Fig. 2: Radial averaged intensity normalised by the number of incident photons as a function of momentum transfer during the in-situ alteration of glass (2% Zr). The signal amplitude increases in a autosimilar manner, corresponding to an increase of the volume without variation of porosity texture. The thickness of the altered layer increases in the first hours of the reaction. Its saturation appears after few days.

Experimental conditions can be summarized as follows: energy was varied between 18 and 25 keV, allowing an acceptable attenuation of the beam through 3cm of water, grazing angle were comprised between 0.04° and 0.2°, horizontal width of the beam onto a vertical surface was at most 80 µm with an horizontal divergence less than 0.2 mrad, detector was mounted 2 meters away from the sample with a vacuum tube in between comprising a Ta beamstop before the mylar exit foil. As main results, the alteration kinetic in-situ over "very short" time (from the geological point of view) reveals a quasi instantaneous nucleation of the pores (see fig ). From the Porod's law (q-4), the presence of a sharp interface between the residual skeleton and the water inside the pores can be evidenced. In the case of a glass with four elements, the nanoporous structure developed on a monolith and at the surface of micrometric grains when they are altered together is the same. GISAXS has proved to be an efficient method to measure in situ the morphology of the superficial layer of a glass monolith. Compared to classical powder studies [2], this technique opens the opportunity of seeking gradients and inhomogeneities in depth in complex glass compositions, closer to the nuclear reference used.

Références:

[1] O. Deruelle, O. Spalla, P. Barboux, J. Lambard J.of non Crystalline solids 261 237-251 (2000)
[2] Spalla O., Barboux, P., Sicard, L., Lyonnard, S., Bley, F. J. Non Cryst. Solid 347 56-68 (2004)
[3] Sicard L., Spalla, O., Barboux, P., Né, F., Taché, O. J. Non Cryst. Solid 345 230-233 (2004)