Transport behaviour of low molecular weight organic compounds in multi-mineral clay systems. A comparison between measured and predicted values

Determining the retardation properties of argillaceous media for dissolved low molecular weight organic compounds (LMW-OC's) is still an open issue in the assessment of the migration behaviour of C-14 in the near and far field of a deep geological repository of radioactive wastes. Here we report retardation factors of potential carriers of organic C-14 (carboxylates and alcohols) in binary clay mixtures (kaolinite and illite) as well as in Opalinus Clay (OPA), the favoured host rock for the deep geological disposal of radioactive waste in Switzerland. The percolation technique with a pulse injection of the LMW-OC's was used for all measurements. The transferability of sorption values of the LMW-OC's gained on the single-mineral systems (kaolinite and illite, respectively) to the binary clay mixtures and to clay rock (OPA) was tested. For binary clay mixtures, percolation experiments were carried out using tritiated water and Cl-36(-) as the reference tracers. The retardation factors predicted by the component additivity approach (CA) were consistent with the measured ones. Deuterated water and bromide were used as reference tracers in experiments with OPA. Complex breakthrough curves of bromide and the test compounds were observed here. Heterogeneities of the pore space are proposed as the main reason to explain the complex structure of breakthrough curve. Additionally, (a)biotic degradation or transformation of the organic compounds over the long experimental duration possibly further added to the complexity of the results in some cases. The derivation of sorption distribution ratios (R-d) , determined from the breakthrough curves was therefore inherently associated with large uncertainties. The transferability of retardation data gained on less complex systems to the conditions of OPA was tested using the CA and the global composite (GC) approach. Significant discrepancies were found between the measured values and those predicted by the CA approach. The discrepancies were found to be smaller using the GC approach. In that case, the results obtained for single- and binary clay systems were also used for calibration purposes. Retardation factors of the LMW-OC's in OPA were then predicted based on two empiric relationships. The observed discrepancies are most probably related to the different chemical composition of the aqueous solution (e.g. ionic strength, pH), the various mineralogical features (e.g. compaction), and the different hydraulic conditions within the clay systems (e.g. Darcy flux). The absence of anion-sorption in OPA may be particularly explained by differences in the pH of the OPA pore water and the pore water in compacted kaolinite samples.