Implementation Pitfalls for Carbonate Mineral Dissolution -- a Technical Note

TL;DR

This work identifies a systematic misreporting in the Palandri2004 carbonate-mineral dissolution rate parameters, where the reaction order was defined with respect to $P(\ce{CO2})$ rather than $a_{\ce{H2CO3}^\ast}$, causing substantial overestimations of dissolution timescales in reactive-transport models. By adopting a corrected formulation that references $a_{\ce{H2CO3}^\ast}$ and validating it against a calcite dissolution batch test, the authors demonstrate that Reaktoro and DuMuX simulations align with experimental data, reducing timescale errors by about an order of magnitude. The study emphasizes the necessity of validating embedded kinetics against experiments and highlights the impact of parameter reporting on model reliability. The findings have practical implications for geochemical simulations, CO2 sequestration studies, and the broader use of kinetic databases in reactive-transport modeling.

Abstract

In systems with slow reaction kinetics, such as mineral dissolution processes, chemical equilibrium cannot be assumed and an accurate understanding of reaction rates is essential; discrepancies in parameter reporting can greatly affect simulation results. This technical note identifies an issue with the reporting of rate parameters for carbonate mineral dissolution in a widely used database for reactive transport modeling based on Palandri and Kharaka 2004. This misrepresentation leads to a considerable overestimation of reaction timescales. Using the simulators Reaktoro and DuMuX, we simulated a simple calcite dissolution batch test and compared the results to experimental data. By adjusting the parameter to align with established literature, we demonstrate an improved fit between simulated and experimental data. Discrepancies in reaction timescales were reduced by an order of magnitude, emphasizing the importance of regular validation of simulations with experimental data.

Figures (1)

• Figure 1: Comparison of the originally reported values in Palandri2004, our proposed adjustment, simulation of the model from Plummer1978 in DuMuX and experimental data from Plummer1976. Evidently, reaction time scales are significantly overestimated by the original report.