Suitability of Existing Numerical Model Codes and Thermodynamic Databases for the Prognosis of Calcite Dissolution Processes in Near-Surface Sediments Due to a CO2 Leakage Investigated by Column Experiments

Abstract

Among the risks of CO2 storage is the potential of CO2 leakage into overlaying formations and near-surface potable aquifers. Through a leakage, the CO2 can intrude into protected groundwater resources, which can lead to groundwater acidification followed by potential mobilisation of heavy metals and other trace metals through mineral dissolution or ion exchange processes. The prediction of pH buffer reactions in the formations overlaying a CO2 storage site is essential for assessing the impact of CO2 leakages in terms of trace metal mobilisation. For buffering the pH-value, calcite dissolution is one of the most important mechanisms. Although calcite dissolution has been studied for decades, experiments conducted under elevated CO2 partial pressures are rare. Here, the first study for column experiments is presented applying CO2 partial pressures from 6 to 43 bars and realising a near-natural flow regime. Geochemical calculations of calcite dissolution kinetics were conducted using PHREEQC together with different thermodynamic databases. Applying calcite surface areas, which were previously acquired by N2-BET or calculated based on grain diameters, respectively, to the rate laws according to Plummer et al. (Am J Sci 278:179–216, doi: 10.2475/ajs.278.2.179 , 1978) or Palandri and Kharaka (US Geol Surv Open file Rep 2004–1068:71, 2004) in the numerical simulations led to an overestimation of the calcite dissolution rate by up to three orders of magnitude compared to the results of the column experiments. Only reduction of the calcite surface area in the simulations as a fitting procedure allowed reproducing the experimental results. A reason may be that the diffusion boundary layer (DBL), which depends on the groundwater flow velocity and develops at the calcite grain surface separating it from the bulk of the solution, has to be regarded: The DBL leads to a decrease in the calcite dissolution rate under natural laminar flow conditions compared to turbulent mixing in traditional batch experiments. However, varying the rate constants by three orders of magnitudes in a field scale PHREEQC model simulating a CO2 leakage produced minor variations in the pH buffering through calcite dissolution. This justifies the use of equilibrium models when calculating the calcite dissolution in CO2 leakage scenarios for porous aquifers and slow or moderate groundwater flow velocities. However, the selection of the thermodynamic database has an impact on the dissolved calcium concentration, leading to an uncertainty in the simulation results. The resulting uncertainty, which applies also to the calculated propagation of an aquifer zone depleted in calcite through dissolution, seems negligible for shallow aquifers of approximately 60 m depth, but amounts to 35 % of the calcium concentration for aquifers at a depth of approximately 400 m.