Credoz, Anthony

This study presents the experimental and modeling results of CO 2 injection and transport in the vadose zone performed in PISCO2 facilities at the ES.CO2 center in Ponferrada (North Spain). During 46 days of experiments, 62.10 kg of CO 2 were injected through 16 micro-injectors in a 35 m 3 experimental unit fi lled with sandy material. Monitoring and mapping of surface CO 2 fl ux were performed periodically to assess the evolution of CO 2 migration through the soil and to the atmosphere. Numerical simulations were run using TOUGH2 code with EOS7CA research module considering two phases (gas and liquid) and three components (H 2 O, CO 2 , air). Two layers (sand, gravel) and atmosphere boundary were implemented taking into account heterogeneous soils, homogeneous soil, rainfall, temperature, and liquid saturation to allow a better understanding of CO 2 behavior in the vadose zone. This combined experimental and modeling approach shows that CO 2 leakage in the vadose zone quickly comes out through preferential migration pathways and spots with the ranges of fl uxes in the ground/surface interface from 2.5 to 600 g·m −2 ·day −1 . This gas channeling is mainly related to soil compaction and climatic perturbation. This has signifi cant implications for designadapted detection and monitoring strategies of early leakage in commercial CO 2 storage.

Diffusive gradients in thin fi lms (DGT) have been tested in CO2-rich, metal-bearing fl uids from springs in the Campo de Calatrava region in Central Spain, to assess their applicability as a monitoring tool in onshore CO2 storage projects. These fi lms are capable of adsorbing metals and recording changes in their concentration in water, sediments, and soils. Considering that CO2 dissolution promotes metal solubilization and transport, the use of these fi lms could be valuable as a monitoring tool of early leakage. A number of DGT have been deployed in selected springs with constant metal concentration. The studied waters show high concentrations of Fe, as high as 1 × 104 μg·L–1, Ni, Co, Zn, Cu, and Mn. Comparing re-calculated metal concentration in DGT with metal water concentration, two different metal behaviors are observed: (i) metals with sorption consistent with the metal concentration (i.e. plotting close to the 1:1 line in a [Me]DGT: [Me]water plot), and (ii) metals with non–linear sorption, with some data showing metal enrichment in DGT compared with the concentration in water. Metals in the fi rst group include Fe, Mn, Co, Ni, and U, and metals in the second group are Zn, Pb, Cr, Cu, and Al. From this research, it is concluded that the metals in the fi rst group can be used to monitor potential leakage by using DGT, providing effective leakage detection even considering low variations of concentrations, episodic metal release, and reducing costs compared with conventional, periodic water sampling.