Effects of the variances of input parameters on water-mineral interactions during CO2 sequestration modeling.

Abstract: The mineral fixation of CO₂ has a significant but highly varying relevance in carbon capture and storage (CCS) models. One reason is the impact of the minerals’ concentrations, but the choice of the thermodynamic database on which the geochemical model calculations are based, has probably an equivalent or even more important impact on the model results. In this example, natural groundwater and a mineral assemblage consisting of calcite, dolomite, albite, and K-feldspar in two different concentrations were used to perform scenario simulations with the geochemical model PHREEQC. The four different thermodynamic datasets phreeqc.dat, wateq4f.dat, llnl.dat, and minteq.dat were used to calculate the amount of CO₂ which could be fixed through the reactions with the described minerals. In the simulations using low reactive mineral concentrations (1–2.5 wt%), the range of the amount of fixed CO₂ in the mineral and dissolved phases lies between 0.29 and 0.31 mol/kgw for the four chosen thermodynamic databases, which reflects the complete transformation of albite and K-feldspar to secondary minerals in all the simulations and only minor differences in the amounts of dissolved calcite and dolomite. On the other hand, the amount of fixed CO₂ shows a variance of between 1.51 and 2.17 mol/kgw when using high reactive mineral concentrations (10–25 wt%). Especially the transformation of the K-feldspar depended significantly on the chosen thermodynamic database which affected the calculated product minerals of the reaction. Using the phreeqc.dat, the wateq4f.dat, and the minteq.dat databases, the amount of CO₂ fixation due to its reaction with K-feldspar is up to six times larger than the amount calculated with the llnl.dat database. It is currently not clearly possible to decide which thermodynamic database should be used to receive the most realistic modeling results. Consequently, the uncertainty in the modeling results due to varying thermodynamic databases should be regarded in addition to the uncertainty resulting from factors like the heterogeneous distribution of mineral phases in the aquifers. It still has to be discussed whether a risk-based approach offers a secure way of modeling CCS related scenarios or whether worst-case scenarios have to be applied. [Copyright &y& Elsevier]