REACTIVITY OF DAWSONITE IN CARBON SEQUESTRATION

Although prominent in a host of modeling studies, dawsonite (NaAlCO₃(OH)₂) rarely occurs in nature and is not a precipitant in experimental simulations of a carbon repository (e.g., Kaszuba et al., 2003, 2005). Given the potential importance of dawsonite in the performance of a geologic carbon sequestration site, the stability of this mineral under physical-chemical conditions relevant to geologic storage and sequestration of carbon requires critical examination. Short-term experiments demonstrate that dawsonite is readily synthesized from gibbsite and kaolinite in concentrated NaHCO₃ solutions. A series of hydrothermal fluid-mineral experiments using flexible cell hydrothermal apparatus (Dickson cells) were used to assess the long-term geochemistry and reactivity of dawsonite in more geologically reasonable solutions: moderately saline (0.05 molal) NaHCO₃ solution with synthetic dawsonite at 50 and 75^oC, 200 bars; 1 molal NaHCO₃ brine with Georgia kaolinite at 75^oC, 200 bars; and 1 molal NaCl brine with Georgia kaolinite at 75^oC, 200 bars. In this last experiment, the brine-mineral system was reacted at pressure and temperature for ~ 45 days to approach steady state, then injected with supercritical CO₂ and allowed to react an additional 45 days. Reacted fluid was periodically sampled and analyzed for pH, SiO₂, Al, Na, Cl, and CO₂. Select samples were analyzed for colloidal particle-size-distribution and concentration.

At temperature-pressure conditions expected for a carbon repository, dawsonite dissolves to yield 10-100 times more Al in NaHCO₃ solution than predicted using existing thermodynamic data. Al-hydroxide colloids, a significant fraction of which are smaller than 0.45 microns, developed in solution. The colloidal fraction that is less than 0.45 microns, a commonly used filter size, accounts for at least some of this Al surplus. The Al surplus and the colloid abundance also appear to be related to the NaHCO₃ concentration. A number of computer calculations of carbon repository integrity predict the formation of dawsonite in response to CO₂ injection. Dissolution of dawsonite and formation of mobile Al hydroxide colloids can undermine the reliability of these predictions. The geochemical behavior of Al will be a key to determining the potential of dawsonite to sequester carbon.