Development and application of an innovative method for studying the diffusion of dissolved gases in porous saturated media

Many countries consider clay-based materials for the safe disposal of high- and intermediate level radioactive waste, either because of the choice of argillaceous formations to host the repository, or as a component in the engineered barrier system. The clays under consideration have a high capacity to retain radionuclides (strong fixation to clay), limited water flow by means of low permeability and high self-sealing properties by capillary sealing efficiency. In Belgium, Boom Clay is considered as a potential host rock. Within a geological repository, the production of gas is unavoidable whereby the dominant process is anaerobic corrosion of metals producing hydrogen. In a first stage, the generated gas will dissolve in the porewater and dissipate by diffusion. If the rate of gas generation is larger than the diffusive flux, a free gas phase will form which might have negative effects on the performance of the barriers. In order to obtain a reliable estimate about the balance between gas generation and gas dissipation, sound diffusion coefficients for dissolved gases are essential. Our first goal was to develop a suitable technique to measure diffusion coefficients of dissolved gases. Besides diffusion coefficients for a variety of dissolved gases (He, Ne, Ar, CH4, C2H6 and Xe), also diffusion coefficients for hydrogen are needed. It is well known that experiments with hydrogen often suffer from experimental problems such as microbial conversion of H2 into CH4. Therefore, it was necessary to find a way to avoid/minimize this microbial activity during diffusion experiments. It is known that diffusion coefficients in free water (D0) depend on the size of the molecule. More specifically, the size dependency of D0 can be described by an exponential function. In this study, it was observed that an exponential relationship was also found for the effective diffusion coefficient (Deff) as a function of molecule size, which confirmed that molecule size also influences diffusion of molecules in porous materials. A third objective was to investigate whether similar exponential relationships could be found for other clayey materials, which finally could be used to estimate diffusion coefficients based on the size of the diffusing molecule. As dissolved gases are considered to be conservative tracers, they are useful for the assessment of the transport properties and pore structure by means of their diffusion coefficients. Thus, these diffusion coefficients would also depend on the petrophysical properties of the material, characterising its pore structure. Hence, the last goal of this research study was to investigate the influence of different petrophysical properties on the diffusive behaviour of dissolved gases, thus, allowing coupling between measured petrophysical parameters, Deff and molecule size. In order to answer these questions, an innovative method was developed to measure the diffusion coefficient of dissolved gases using the double through-diffusion methodology. This allowed to measure the diffusion coefficients of two dissolved gases in a single experiment with a high precision. When using hydrogen, a complex sterilisation procedure combining heat sterilisation, gamma irradiation, gas filtration and the use of a microbial inhibitor was developed, which eliminated microbiological disturbances. By using this procedure, for the first time, reliable and accurate diffusion coefficients for dissolved hydrogen were obtained for three different samples of the Boom Clay. The obtained diffusion coefficients enable a more precise assessment of the problems related to H2 production/dissipation in a repository environment. To investigate the relationship between the molecule size and their diffusion coefficients in more detail, diffusion experiments with gases of different sizes and HTO were performed on different clay-rich / argillaceous samples (Boom Clay, Eigenbilzen Sands, Callovo-Oxfordian Clay, Opalinus Clay and bentonite). Similar to the relationship between D0 and molecular size, for all samples under investigation, a reliable relationship between the molecular size and effective diffusion coefficient was obtained which can be described by an exponential function. The difference in distance between the Deff and D0 curves relates to the geometrical factor (G ~ D0/Deff). This geometric factor provides information on how the porous network influences diffusing molecules and account for the tortuosity and constrictivity of the sample. For the samples of the Boom Clay and the Eigenbilzen Sands, the exponential coefficient is very similar to the D0 relationship. Similar exponential coefficients indicate that the geometric factor will be quasi constant when the size of the diffusing molecule increases. This matches with the experimental results, where the difference in G between the smallest and the largest molecule is less than 3. However, for the other clayey samples (COX, OPA and bentonite), the exponential factors differ from the one of the D0 relationship, hence G varies strongly with the size of the diffusing molecule, which is also experimentally observed. In literature, diffusion coefficients are often estimated by using a constant value of G for a certain sample or formation (often derived from diffusion experiments with HTO). Based on the data presented in this work, one can conclude that this approach is not always correct and it can lead to a substantial overestimation of the diffusion coefficient. Therefore, we propose an alternative method to estimate diffusion coefficients of dissolved gases, based on the exponential relationship that has been observed on a large set of diversified samples. By measuring experimentally the effective diffusion coefficient of two unreactive, dissolved gases possessing a different size, one can determine the exponential function and as a consequence, one can derive the diffusion coefficients of other dissolved gases (with a size in between the two measured gases) based on their size. When using this approach for one of our samples, the predicted and measured diffusion coefficients differ by less than 30%, which is deemed satisfactory for predictive gas dissipation calculations. In order to investigate how the transport properties of a dissolved gas molecule can be linked to the petrophyiscal and petrographical properties of a clay-rich sample, the main focus was on clay-dominated Boom Clay samples (Putte and Terhagen Member) and more sandy Eigenbilzen Sands. For these samples, a detailed petrophysical analysis has been performed. Diffusivity and hydraulic conductivity of the Boom Clay and Eigenbilzen Sands are very different. Petrophysical analysis showed large differences in mineralogy and grain size distribution: samples of the Boom Clay are rich in clay minerals and contain a large weight percent’s (> 67%) clay fraction ( 43%) sand fraction (> 62 µm). These differences in composition are also reflected by their microstructure. The Boom Clay samples are characterised by a clay supported matrix with some homogeneously distributed quartz grains; pores are not visible by the techniques used ( 250 nm) which allows enhanced transport of dissolved gases and water. Hence, a clear link was found between the transport properties and the petrophysical/petrographical properties of the samples