The Second Matrix Diffusion Experiment in the Water Phase of the Repro Project: WPDE 2

REPRO experiments (Retention properties of rock matrix) study rock matrix retention properties under realistic conditions deep in the bedrock in ONKALO underground characterisation facility at Olkiluoto, Finland. The objective is to investigate tracer transport in rock matrix that is representative for the near-field of the nuclear waste repository, and to demonstrate that the assumptions made in the safety case of the final nuclear waste repository are in line with the site evidence. REPRO is composed of several parallel laboratory and in-situ experiments that study retention properties using different experimental configurations. Present report describes the second in-situ tracer test, the WPDE 2-test (Water Phase Diffusion Experiment 2). This experiment was started in January 2013 and it continued until May 2014. It was performed as a matrix diffusion experiment in a single drill hole. Tracer transport took place along an artificial fracture on the circumference of a 2 m long packed-off section of the drill hole. The artificial fracture was formed by placing a cylindrical flow guide in center of the drill hole section. This set-up enables matrix diffusion through the wall of the drill hole into the rock matrix that surrounds the experimental section and sorption of radionuclides onto the mineral surface. The first in-situ tracer test WPDE1 (Water Phase Diffusion Experiment 1) was performed in 2012 (Poteri et al. 2017). There were only two differences in the experimental set-up of the WPDE 1 and WPDE 2 experiments. The flow rate in WPDE 2 was half of the flow rate in WPDE 1 and composition of the tracer cocktail was slightly different. The main retention process studied by WPDE experiments was matrix diffusion and sorption in the pore space of the rock matrix. Retention by matrix diffusion is very sensitive to changes in flow rate. This means that the change in retention due to the different flow rates in WPDE 1 and WPDE 2 is an indication of the dynamics of the main retention process. Repeating the test using two different flow rates significantly increases confidence on the identification of the main retention process. WPDE 2 test was executed by injecting a short pulse of a cocktail of five different tracers (HTO, 22Na, 36Cl, 85Sr and 133Ba) to the water flowing (10 µL/min) through the test section of the drill hole. The experimental performance of the test was successful. Measured breakthrough curves were analyzed using two independent models. The main peaks of the breakthrough curves had the same characters as in the WPDE 1 experiment such that first breakthrough took place much earlier than was estimated based on the volume of the flow channel and the raise of the breakthrough curve was very steep. However, the main peak in WPDE 2 was wider than in WPDE 1 and it had indications of double peaks, contrary to WPDE 1 that had only a single peak. Interpretation of the breakthrough curves indicated clearly that retention observed in the breakthrough curves was caused by matrix diffusion to the rock surrounding the drill hole and sorption on the minerals in the pore space of the rock matrix. Both models used to analyze the breakthrough curves gave similar retention properties, taking into account the error range of the results. Results for the non-sorbing HTO also agree well with the corresponding results from laboratory. For sorbing tracers the distribution coefficients determined from in-situ experiment appears to be smaller than the distribution coefficients determined in laboratory. Tracer retention properties determined from WPDE 2 breakthrough curves followed the order that was expected based on their sorption properties (22Na<85Sr<133Ba). Pore diffusivity in the rock matrix for HTO was estimated from modelling results by assuming HTO as non-sorbing and using porosity of the rock matrix determined in laboratory (ε ~ 0.6%). This gave diffusivity of Dp ~ (1...2)×10‑11 m2/s, that agrees well with the laboratory result Dp=(1.0 ± 0.3)×10-11 m2/s and also with the corresponding result from WPDE 1 –experiment (Dp ~ 2×10‑11 m2/s). Accessible diffusion porosity for 36Cl was determined to be ε 0.3% (in WPDE 1 -experiment ε 0.2%) and diffusivity Dp < 8×10-12 m2/s (in WPDE 1 ‑experiment Dp < 1.2×10‑11 m2/s). Diffusivities determined for sorbing tracers were Da ~ (2–3)×10‑13 m2/s for 22Na (in WPDE 1 Da ~ 1.6×10‑13 m2/s) and Da ~ (4–10)×10‑14 m2/s for 85Sr (not included in WPDE 1). These are higher than corresponding laboratory values by a factor of about 2–5. Distribution coefficient determined for 22Na was Kd ~ 1×10-4 m3/kg and for 85Sr it was Kd ~ 3–4×10‑4 m3/kg. Corresponding distribution coefficients determined in laboratory are about an order of magnitude higher for 22Na and about a factor three higher for 85Sr. Results for 133Ba are very uncertain, because the recovery was only about 30% of the injected activity. These results indicated that distribution coefficient for 133Ba would be about an order of magnitude higher, and diffusivity (Da) about the same amount smaller in-situ than the corresponding values from laboratory. But, these values are very uncertain as noted above. For example, also possible sorption on the surface of the drill hole wall could have considerable influence on the results of this tracers.