Mineral dissolution and precipitation during CO2 injection at the Frio-I Brine Pilot: Geochemical modeling and uncertainty analysis

• Geochemical models for Frio-I field test are constructed.
• Uncertainties due to activity correction and mineral surface areas are estimated.
• Estimated 0.02 wt.% of reservoir rock mass is dissolved during the field test.
• Frio “C” mineral assemblage has minimal reactivity over 1000 years.

During the Frio-I Brine Pilot CO2 injection experiment in 2004, distinct geochemical changes in response to the injection of 1600 tons of CO2 were recorded in brine samples collected from the monitoring well. Previous geochemical modeling studies have considered dissolution of calcite and iron oxyhydroxides, or release of adsorbed iron, as the most likely sources of the increased ion concentrations. In this modeling study we explore possible alternative sources of the increasing calcium and iron, based on the data from the detailed petrographic characterization of the Upper Frio Formation “C”. Particularly, we evaluate whether dissolution of pyrite and oligoclase (anorthite component) can account for the observed geochemical changes. Due to kinetic limitations, dissolution of pyrite and anorthite cannot account for the increased iron and calcium concentrations on the time scale of the field test (10 days). However, dissolution of these minerals is contributing to carbonate and clay mineral precipitation on the longer time scales (1000 years). We estimated that during the field test dissolution of calcite and iron oxide resulted in ∼0.02 wt.% loss of the reservoir rock mass. The reactive transport models were constructed for 25 and 59 °C temperature and using Pitzer and B-dot activity correction methods. These models predict carbonate minerals, dolomite and ankerite, as well as clay minerals kaolinite, nontronite and montmorillonite, will precipitate in the Frio Formation “C” sandstone as the system progresses toward chemical equilibrium during a 1000-year period. Cumulative uncertainties associated with using different thermodynamic databases, activity correction models (Pitzer vs. B-dot), and extrapolating to reservoir temperature, are manifested in the difference in the predicted mineral phases. However, these models are consistent with regards to the total volume of mineral precipitation and porosity values which are predicted to within 0.002%.

Figure optionsDownload as PowerPoint slide