DI-B experiment:: planning, design and performance of an in situ diffusion experiment in the Opalinus Clay formation

The DI-B experiment is a long-term, natural-scale, in situ diffusion experiment, which is being performed in the Opalinus Clay formation at the Mont Terri Underground Rock Laboratory (URL), in Switzerland, employing nonradioactive tracers. One of the key aspects to be addressed for nuclear waste repository safety assessment purposes is the understanding of the transport mechanisms of the radionuclides contained in the radioactive waste. Consolidated clay formations display very low water hydraulic conductivities, so it is expected that the predominant transport process will be diffusion. The experimental set-up has been designed to withstand the site conditions and for monitoring and recording several physicochemical parameters (pH, conductivity, oxidation-reduction potential), as well as the pressures in the circuit and for the long-distance monitoring of the data acquisition system. The tracer selection has been made based on previous investigations carried out at CIEMAT, including a literature survey, laboratory sorption experiments and hydrogeochemical modeling for determining tracer stability under the physicochemical conditions to be expected in the site. The final selection includes (6)Li, (87)Rb, D (as D(2)O) and I (as I(-)). Hydrogeochemical modeling confirmed the stability of all the tracers selected. Batch sorption experiments showed that no sorption in the rock occurred in the case of (6)Li, D and I (conservative tracers), whereas (87)Rb was 100% sorbed. However, (87)Rb was chosen because of its analogy with Cs, a relevant radionuclide commonly present in the nuclear spent fuel. Diffusion experiments have been carried out at laboratory scale with Opalinus Clay samples to provide diffusion parameters for modeling purposes. Effective diffusion coefficients, perpendicular and parallel to the bedding planes of the rock, respectively, were (1.68 +/- 0.42) x 10(-11) and (4.02 +/- 0.30) x 10(-11) m(2)/s for tritium, and (2.70 +/- 0.27) x 10(-12) and (1.38 +/- 0.49) x 10(-11) m(2)/s for iodide. Additional through-diffusion experiments (parallel to the bedding) were performed with the nonsorbing tracer (36)Cl(-), in order to check the results obtained for iodide. The effective diffusion coefficient measured for chloride ions was (1.18 +/- 0.27) x 10(-11) m(2)/s, which is practically equal to the value obtained for iodide. Preliminary diffusion calculations have been carried out using two transport codes: GIMRT and CORE(2D), with conservative and nonconservative tracers, using effective diffusion coefficients (D(e)) obtained experimentally in the laboratory (through-diffusion experiments) or selected from the literature. The diffusion profiles obtained from the calculations showed slight variations, which were consistent with the different modeling approaches employed. The predictive modeling results have been used to determine the initial tracer concentration that should be added to the circuit to assure well-defined profiles at the end of the experiment. This long-term in situ diffusion experiment will also provide useful data for the interpretation of previous diffusion experiments performed at the Mont Terri URL. (C) 2004 Elsevier B.V. All rights reserved.