Technical Report NTB 91-04

Grimsel Test SiteThe radionuclide migration experimentOverview of investigations 1985 – 1990

This paper provides an overview of the investigations conducted from 1985 to 1990 as a part of the radionuclide migration experiment which is currently in progress in the Nagra underground research laboratory at the Grimsel pass in the Central Swiss Alps. The major aims of the project are (1) to test the extrapolation of laboratory sorption data to field conditions, (2) to analyse retardation processes in a fractured rock, (3) to improve and develop the necessary methodologies for site characterization and (4) to test existing geochemical, hydrodynamic, and solute transport models or their associated data bases. Field and modeling work are complemented by an extensive laboratory support programme.

The field experiments are carried out in a water conducting fracture within a tectonically active, granitic rock mass. On the scale of several meters this fracture can be described as a two-dimensional feature. The major water flow paths are the product of relatively recent tectonic deformation (alpine uplift). Brittle processes reactivated older, ductile shear zones and led to the formation of a few centimeter wide, asymmetrical array of conductive fractures in an otherwise relatively impermeable mylonitic fabric. These fractures are partially filled with highly porous, unconsolidated micaceous fault gouge. The migration fracture and the surrounding environs were extensively characterized in terms of mineralogy, hydrochemistry and hydrology prior to performing the migration test.

The water in the migration fracture is anoxic, has a high pH (~ 9.6) and a low ionic strength (0.0012 M) with the cation content dominated by Na+ and Ca2+ while F-, HCO3-, Cl-, SO42-­ and SiO(OH)3- are the major anions. Further studies included the definition of ambient microbiological and colloidal populations as well as alteration and natural trace element transport processes in the fracture vicinity.

Hydrogeological exploration involved the drilling of 8 boreholes (6 to 24 m long). Subsequent hydraulic testing revealed local heterogeneities of the transmissivity from 5·10-6 to about 10-8 m2s-1 (single hole tests). On a scale of several meters, however, the fracture appears rather homogeneous with an average transmissivity of about 2·10-6 m2s-1 based on the results of cross-hole tests. Various hydrodynamic models satisfactorily predict stationary conditions (pressure distribution, flow) but have some difficulty in reproducing the observed transient response from long-term pumping tests.

At distances between 6 and 12 m from the tunnel wall, hydraulic pressures average around 1.5 bar. This hydraulic setting is well suited to maintain artificially stressed, closed flow fields (unequal strength dipoles with input/output flow ratios always less than 1/2) for tracer experiments. A series of pilot experiments utilizing conservative tracers (uranine, 4He and 82Br-) was performed in order to test the equipment and to identify appropriate borehole constellations and flow conditions (travel times) for the planned migration experiments. A variety of novel techniques, including quartz fiber fluorometry (optrodes) and the He tracer method were developed and successfully tested.

For the first series of reactive tracer migration experiments, isotopes of Na+ and Sr2+ were chosen on the basis of laboratory studies, consisting of rock-water interaction, batch radionuclide sorption and dynamic sorption experiments. These studies indicated that at low concentrations the sorption of the Na+ and Sr2+ tracers at low concentrations is dominated by reversible cation (isotope) exchange under equilibrium conditions and is proportional to the available cation exchange capacity CEC. Sorption coefficients (Kd) for Na and Sr under the chemical conditions prevailing in the migration fracture were expected to range from 0.3 - 0.4 ml g-1 and from 16 - 25 ml g-1, respectively.

The first field tests performed, using a pulse injection of short-lived 24Na+, showed a small retardation of the breakthrough peak. To prolong the interaction time of the tracer with fracture materials, a complex setup for continuous injection was constructed with tracer dosage and input/output flow being provided by highly precise (modified) HPLC pumps. In an attempt to minimize uncertainties from instrumental dispersion, equipment volumes were reduced as much as possible and down-hole tracer analyses implemented. A 7-day continuous injection of 22Na+, together with uranine (and pulses of 82Br- and 4He) was monitored over 2 months. Utilizing the same equipment, several pulse injection runs with 24Na+, uranine, 82Br- and 123I-, and input/output ratios of 1/3 and 1/15, were carried out. Pulse tests with moderately sorbing 85Sr2+ were still underway at the end of the period considered in this report.

Laboratory rock-water interaction studies, a field-scale equilibration experiment and a preliminary interpretation of Na tracer breakthrough experiments by a dual-porosity modeling approach yield roughly consistent sorption values. With the transport model calibrated by the Na experiments, detailed predictions for the breakthrough of Sr are available. Ongoing field experiments with 85Sr2+ will provide a critical test for the applicability of the radionuclide retardation model. After the tests with chemically simple and weakly sorbing tracers, it is planned to use chemically more complex nuclides of relevance to repository safety assessment studies (including isotopes of Cs, Se, Ni, Tc, and potentially Pd, Sn and Np). The final stage of the migration experiment is intended to be an excavation of part or all of the tracer flow path.

The Grimsel migration experiment demonstrates conclusively how the combined efforts of modeling, laboratory and field investigations can substantially widen the understanding of radionuclide transport in a geological environment.