Spatial Zonation Limits Magnesite Dissolution in Porous Media

We investigate how mineral spatial distribution in porous media affects their dissolution rates. Specifically, we measure the dissolution rate of magnesite interspersed in different patterns in packed columns of quartz sand where the magnesite concentration (v/v) was held constant. The largest difference was observed between a Mixed column containing uniformly distributed magnesite and a One-zone column containing magnesite packed into one cylindrical center zone aligned parallel to the main flow of acidic inlet fluid (flow-parallel One-zone column). The columns were flushed with acid water at a pH of 4.0 at flow velocities of 3.6 or 0.36m/d. Breakthrough data show that the rate of magnesite dissolution is 1.6-2 times slower in the One-zone column compared to the Mixed column. This extent of rate limitation is much larger than what was observed in our previous work (14%) for a similar One-zone column where the magnesite was packed in a layer aligned perpendicular to flow (flow-transverse One-zone column). Two-dimensional reactive transport modeling with CrunchFlow revealed that ion activity product (IAP) and local dissolution rates at the grid block scale (0.1cm) vary by orders of magnitude. Much of the central magnesite zone in the One-zone flow-parallel column is characterized by close or equal to equilibrium conditions with IAP/Keq\u3e0.1. Two important surface areas are defined to understand the observed rates: the effective surface area (Ae) reflects the magnesite that effectively dissolves under far from equilibrium conditions (IAP/Keq\u3c0.1), while the interface surface area (AI) reflects the effective magnesite surface that lies along the quartz-magnesite interface. Modeling results reveal that the transverse dispersivity at the interface of the quartz and magnesite zones controls mass transport and therefore the values of Ae and AI. Under the conditions examined in this work, the value of Ae varies from 2% to 67% of the total magnesite BET surface area. Column-scale bulk rates RMgCO3,B (in units of mol/s) vary linearly with Ae and AI. Using Ae to normalize rates, we calculate a rate constant (10-9.56mol/m2/s) that is very close to the value of 10-10.0mol/m2/s under well-mixed conditions at the grid block scale. This implies that the laboratory-field rate discrepancy can potentially be caused by differences in the effective surface area. If we know the effective surface area of dissolution, we will be able to use the rate constant measured in laboratory systems to calculate field rates for some systems. In this work, approximately 60-70% of the Ae is at the magnesite-quartz interface. This implies that in some field systems where the detailed information that we have for our columns is not available, the effective mineral surface area may be approximated by the area of grains residing at the interface of reactive mineral zones. Although it has long been known that spatial heterogeneities play a significant role in determining physical processes such as flow and solute transport, our data are the first that systematically and experimentally quantifies the importance of mineral spatial distribution (chemical heterogeneity) on dissolution