Effect of the Callovian-Oxfordian clayey fraction on borosilicate glass alteration

Document available in extended abstract form only. In France, high-level nuclear waste (HLW) is confined in a glass matrix packaged into stainless steel canister and carbon steel overpack. The HLW should be buried in a geological clay formation like, potentially, the Callovian-Oxfordian (COx) clay-stone located in the north-eastern Parisian basin. The COx clay-stone contains minerals that can feed the near-field with soluble Mg. Such minerals are carbonates (ankerite, dolomite) as well as clay minerals (chlorite, illite, interstratified illite/smectite). Previous laboratory experiments have proved that aqueous solutions of Mg salts could significantly increase the alteration rate of nuclear glass (Jollivet et al., 2012). This motivated to go a step further by studying the alteration of nuclear glass put in contact with Mg minerals. A first set of experiments have revealed that the rate of glass dissolution was increased with hydro-magnesite (4MgCO3.Mg(OH)2.4H2O, a chemically simple model mineral) and dolomite. In both cases, Mg coming from carbonate dissolution reacts with Si, provided by the glass, in order to form Mg silicates (Debure et al., 2012). In that case, Si consumption sustains glass alteration. Mg silicate precipitation also consumes protons; therefore the interdiffusion of alkali within the glass alteration layer eventually becomes a driving force that sustains Mg silicate precipitation. The second set of experiments, presented here, aimed at better characterizing the role of the COx clayey fraction. The separation of the clayey phases of the COx clay-stone has been made in collaboration with the LEM lab (Nancy, France) by a sequence of sieving, acidic dissolution of carbonates, NaCl washing and sedimentation (Rivard, 2011). According to XRD and infrared analyses, the clayey fraction was mainly composed of kaolinite, illite, interstratified illite/smectite and chlorite (plus a little residual amount of quartz). This first step aimed to remove easily exchangeable Mg and to check whether remaining Mg, especially structural Mg could still feed the solution. The cation exchange capacity (CEC) and the exchangeable cation population after treatment were measured by the cobaltihexamine test, indicating an enrichment of Na (from 14% to 79% in equivalents) of the exchangeable population at the expense of Ca (from 50% to 4%), Mg (from 15% to 6%) and K (from 18% to 7%) compared to the pristine clay-stone. Synthetic borosilicate Mg-free glasses were used as simplified references of the French nuclear glass R7T7. Batch experiments were performed in closed system at 90 deg. C during 150 days. The clay/glass (C/G) weight ratio ranged from 0.01 to 100 for a liquid/solid (L/S) weight ratio of 20. The specific surface was estimated to be around 0.06 m2/g and 100 m2/g for the glass powder and the clayey fraction, respectively. Chemical analysis of the batch solutions were carried out with time, as well as XRD, SEM and TOF-SIMS analysis of the solids after completion of the batch experiments. Modeling was performed with the GRAAL kinetic model of glass dissolution that considers elements and water diffusion in the alteration layer (Frugier et al., 2008), implemented within the reactive transport code HYTEC (van der Lee et al., 2003). The experiments showed that the higher the clay/glass ratio, the lower the pH90deg.C and the higher the glass alteration. As shown in Fig. 1, the pH90deg.C varied from 8.9 for the lowest clay/glass proportion (C/G 1) to 5.4 for the highest one (C/G = 100). The corresponding mean glass dissolution rates are 15 nm/d and 300 nm/d, respectively. A batch test performed with the sole clayey fraction in pure water yielded a pH90deg.C of 5.2, whereas the test made with the glass alone led to a pH90deg.C around 9. The main effect of the clayey fraction in the batch tests seems to drop the pH down to more acidic values that concentrations of B (a tracer of the glass alteration) on the one hand, and of the concentrations of Ca and Na on the other hand, was indicative of a dditional processes such as cation exchange. Furthermore, Si concentration in glass/water and in glass/clay experiments is higher than in the sole clay experiments. In case of C/G = 1, where pH was close to 8.9, alteration was three times higher than in pure water at the same pH. This effect could be due to potential Si-bearing-phases precipitation, such as Mg silicates. At higher C/G ratio, when pH decreased below 8, the analysis of the solid phases clearly showed the formation of an alteration layer around the glass grains (Fig 1: case of C/G = 10, pH = 6.7), but there was no evidence of Mg silicate neo-formation. The modeling results supported the predominant role of pH in the control of glass dissolution, as well as cation exchange and no precipitation of Mg silicate phases below pH 7. The model failed to reproduce the pH buffering of the clay to values as low as 5.5. Finally, whatever the mineral: hydro-magnesite, dolomite or poor exchangeable Mg clay fraction, the glass alteration rate is higher in presence of Mg minerals than in pure water. Similar batch test performed with the COx clay-stone are under progress along these lines. (authors)