Laboratory and Numerical Simulation Studies of Convectively Enhanced Carbon Dioxide Dissolution3 min read

Timothy J. Kneafsey and Karsten Pruess

As a possible means of reducing CO₂ emissions into the atmosphere, ESD’s Timothy J. Kneafsey and Karsten Pruess have recently studied the injection of carbon dioxide (CO₂) into deep saline aquifers and the benefits of enhanced CO₂ dissolution into the local brine. The method involves CO₂ being injected into a permeable, porous zone confined beneath a low-permeability cap rock, so that the CO₂ would remain in the aquifer for an extended period of time. Injected CO₂ would be in a supercritical state (scCO₂) and have lower density than the local existing brine; consequently, it would tend to rise to the top of the permeable interval and spread beneath the cap rock. After CO₂ is injected into the subsurface, it would either (1) exist in a mobile, separate, supercritical phase; (2) manifest itself as trapped scCO₂; (3) be dissolved in the host brine; or (4) be precipitated as solid minerals. The storage security and permanence increase as CO₂ is trapped, then dissolved, and then eventually chemically bound to solid phases, but the expected time scale of each of these modes increase sequentially as well. Any enhancement of the rate or magnitude of processes increasing security or permanence would be desirable.

Injected CO₂ will tend to spread under a confining cap rock, and at some distance from the injection well, there would likely be a nearly horizontal interface between a free CO₂ phase above and the aqueous phase below. At the CO₂/water interface, CO₂ will dissolve into the aqueous phase, and if the aqueous phase were immobile, the rate of CO₂ dissolution would be limited by the rate at which CO₂ could be removed from the interface by molecular diffusion. This process is slow, and the rate of CO₂ dissolution would decrease with time.

CO₂ dissolution into brine causes the brine density to increase on the order of 0.1% to 1%, depending on pressure, temperature, and salinity. This causes denser CO₂-rich brine to overlie the less dense local brine, inducing a gravitational instability. This instability can result in fluid convection, which could significantly increase the rate at which dissolved CO₂ is transferred from the interface with the overlying free CO₂, accelerating CO₂ dissolution. Kneafsey and Pruess have performed visualization and quantitative laboratory tests and numerical modelling to study this phenomenon.

In these laboratory visualization tests, CO₂ gas is introduced above the brine contained between two sheets of glass. The space between the sheets of glass is an analogue for the pore space within the reservoir in which the CO₂ would be emplaced. The brine contains water and a pH sensitive dye, sensing the CO₂ reaction with the water that forms an acidic solution. Initially, a diffusive front develops at the CO₂/water interface. This denser brine becomes unstable and forms fingers, which penetrate deeply into the cell, causing fresh brine to flow to the surface and enhancing CO₂ dissolution (Figure 1).

Kneafsey and Pruess have compared their experimental results to a numerical model using TOUGH2 and found good agreement (Figure 2). They are now looking at enhanced CO₂ dissolution in heterogeneous systems using both visualization and quantitative measurements, under conditions that are similar to natural reservoirs—to better understand this dissolution process. These measurements, when extended using numerical simulation, will be used to estimate how enhanced CO₂ dissolution resulting from density driven convection will affect geologic CO₂ sequestration at the field scale.

Figure 1. Experimental setup and the effects of CO₂ dissolution and density-driven convective at 1 minute, 35 minutes, and 68 minutes (clockwise from top left).

Figure 2. Comparison between density-driven convection in the modeled system (left) and experimental system (right).