Abstract
Diffusion parameters for HTO, 36Cl−, and 125I− were determined on Upper Toarcian argillite samples from the Tournemire Underground Research Laboratory (Aveyron, France) using the through diffusion technique. The direction of diffusion was parallel to the bedding plane. The purpose of the present study was 3-fold; it was intended (i) to confirm the I− interaction with Upper Toarcian argillite and to verify the effects of initial I− concentration on this affinity, as previously observed by means of radial diffusion experiments, (ii) to highlight any discrepancy between Cl− and I− diffusivity, and (iii) to investigate the effect of an increase of the ionic strength of the solution on the anionic tracers' diffusive behaviour. The results show that the effective diffusion coefficient (De) and diffusion accessible porosity (εa) values obtained with an ionic strength (I.S.) synthetic pore water of 0.01 eq L−1 are: De = 2.35-2.50 × 10−11 m2 s−1 and εa = 12.0-15.0% for HTO, and De = 14.5-15.5 × 10−13 m2 s−1 and εa = 2.5-2.9% for 36Cl−. Because of anionic exclusion effects, anions diffuse slower and exhibit smaller diffusion accessible porosities than HTO, taken as a water tracer. The associated effective diffusion coefficient (De) and rock capacity factor (α) obtained for 125I are: De = 7.00-8.60 × 10−13 m2 s−1 and α = 4.3-7.2%. Such values make it possible to calculate low 125I distribution ratios (0.0057 < RD < 0.0192 mL g−1) which confirm the trend indicating that the 125I rock capacity factor increases with the decrease of the initial I− concentration. Additional through-diffusion experiments were carried out with a higher ionic strength synthetic pore water (I.S. = 0.11 eq L−1). No evolution of HTO diffusion parameters was observed. The anionic tracers' effective diffusion coefficient increased by a factor of two but no clear evolution of their accessible porosity was observed. Such a paradox could be related to the particularly small mean pore size of the Upper Toarcian argillite of Tournemire. The most significant finding of this study is the large discrepancy (factor of two) between the values of the effective diffusion coefficient for 125I and 36Cl. Whatever the ionic strength of the synthetic solution used, 125I exhibited De values two times lower than those of 36Cl. A detailed explanation for this difference cannot be given at present even if a hypothesis based on ion-pairing or on steric-exclusion cannot be excluded. This makes questionable the assumption usually made for quantifying 125I sorption and postulating that 36Cl and 125I would diffuse in the same porosity. In other terms, at Tournemire, 125I sorption could be more pronounced than previously indicated.