Examining supercritical CO₂ dissolution kinetics during carbon sequestration through column experiments
MetadataShow full item record
Carbon sequestration is a method of capturing and storing excess anthropogenic CO₂ in the subsurface. When CO₂ is injected, the temperature and pressure at depth turn it into a supercritical (SC) fluid, where density is that of a liquid, but viscosity and compressibility resemble a gas. Ultimately the SC CO₂ is trapped at depth either by low permeability sealing layers, by reactions with minerals, or by dissolving into fluids. The injected CO₂ is buoyant and initially exists as a non-aqueous hydrophobic layer floating on top of the subsurface brine, up against the upper sealing formation, but over time it will dissolve into the brine and potentially react with minerals. The details of that initial dissolution reaction, however, are only poorly understood, and I address three basic questions for this research: What is the fundamental kinetics of SC CO₂ dissolution into water? How fast does dissolved CO₂ diffuse away from the source point? And what geochemical conditions influence the dissolution rate? To answer these questions I employed a high pressure flow-through approach using a column packed with coarse quartz sand. The system was both pressure and temperature controlled to have either liquid or SC CO₂ present, and was typically run at 100 Bar, 0.5 to 2.5 mls/min, and 28-60°C. After establishing the hydraulic parameters for the column using two conservative tracers (Br, As), injections (5 and 20 [mu]l) were made either as aqueous solutions equilibrated to high pressure CO₂, or as pure liquid or SC CO₂ into 0.1 mmol NaOH. For all experiments the pH of the system was monitored, and [CO₂] over time was calculated from those data. For injections of brine with dissolved CO₂, transport was conservative and was nearly identical to the conservative tracers. The CO₂ quickly mixes in the column and does not react with the quartz. The liquid and SC CO₂ injections, however, do not act conservatively, and have a very long tailing breakthrough curve that extends to tens of pore volumes. I hypothesize that the SC CO₂ is becoming trapped as a droplet or many droplets in the pore spaces, and the long breakthrough tail is related either to the rate of dissolution into the aqueous phase, the diffusion of dissolved CO₂ away from the phase boundary, or the reaction with the NaOH, limited to the narrow contact zones in the pore throats. Because of the speed at which acid-base reactions occur (nanosecond kinetics), I infer that the rate limiting step is either surface dissolution or diffusion. From plots of ln[CO₂] v. time I obtained values for k, the specific rate of the dissolution reaction R=-k[CO₂]. No trend for k was seen with respect to changes in temperature, but k did show a trend with respect to changing flow rate. k increased from an average value of 3.05x10⁻³ at 0.5 ml/min to an average value of 3.38x10⁻³ at 1.6 ml/min, and then held constant at the higher flow rates, up to 2.5 ml/min. I interpret these data to show that at low flow rates, the reaction is diffusion limited; the fluid nearest the contact zone becomes saturated with dissolved CO₂. At higher flow rates, the fluid is moving fast enough that saturation cannot occur, and the kinetics of the dissolution reaction dominate. Simple geometric models indicate that the CO₂/water interface is shaped like a spherical cap, indicating that the snapped-off CO₂ is forming a meniscus in the pore throat, limiting the surface area across which dissolution can occur.