Trapping and keeping carbon dioxide emissions
Down to Earth:
Plant life has evolved to take CO2 from the atmosphere naturally. Rocks? Naturally not! And yet, rocks are now being used to sequester CO2 emissions from power plants and industrial facilities. In test sites around the world, the CO2 captured from those emitters is transported as a compressed fluid and injected underground into formations of porous rock.
Injecting the supercritical CO2 into a porous rock formation that is overlain by an impermeable, nonporous rock layer traps the CO2 and prevents it from migrating upward, thus keeping it, at least for a while, from returning to the atmosphere. However, the behavior of the fluid in the pore space plays an important role in its movement and containment, and will determine how long the CO2 will remain underground. And when we understand these factors, we will be closer to knowing whether carbon sequestration is a viable solution to combat global warming.
Getting TOUGH at Berkeley
At the Lawrence Berkeley National Laboratory's Earth sciences division, Christine Doughty uses the Transport of Unsaturated Groundwater and Heat (TOUGH) model to simulate the movement of water vapor and heat in porous media. She has customized this model so she can compare field measurements with simulations, and thus determine just how much injected CO2 will stay trapped in an aquifer as bubbles within the porous media.
Carbon dioxide underground is a tricky business. At specific temperatures and pressures, it behaves as a supercritical fluid, filling its container like a gas but with a density like that of a liquid. Supercritical CO2 is buoyant underground due to its low density relative to the native brine, and this makes leakage (as it tries to move up toward the surface) a concern.
In Doughty's version of the TOUGH model, the amount of supercritical CO2 that stays trapped in the formation is based on a process called capillary trapping or residual gas saturation, by which water in the aquifer starts to move back into the rock pores as a plume of buoyant CO2 rises. The water stays close to the mineral grains, flowing around bubbles of CO2.
"Individual CO2 bubbles break off from the plume and are trapped," explains Doughty. The trapping happens after the injection of CO2 ceases and the supercritical plume is no longer expanding in all directions. On a macroscopic scale, residual trapping occurs when the fraction of CO2 in the pore space decreases to the "residual gas saturation," an amount small enough to be split into bubbles, which then cease to move. Such a strong trapping mechanism is good for storage because it reduces the potential for leakage.
In the model, the residual gas saturation can be adjusted for different values and compared to field images showing how much the CO2 has spread. Doughty's "tough" work at Berkeley, then, is providing solid ground (so to speak) upon which to build a case for sequestration.
It takes two
Investigation in the field is equally important to determining sequestration's future. To this end, scientists are carefully monitoring a large-scale CO2 injection at Cranfield Field, Mississippi, conducted as part of the Southeast Regional Sequestration Partnership (SECARB). The observation aims to provide the information needed to increase confidence in CO2 geological storage capacity and retention.[1,7] An injection well drilled in 2009 has conducted more than 3 million tons of CO2 into a 25-m-thick layer of the Lower Tuscaloosa Formation. Two observation wells are equipped with state-of-the-art monitoring systems that track and monitor the CO2 plume as it moves through the reservoir.
Temperature and pressure gauges inside the observation wells monitor the subsurface conditions. Additionally, the resident brine and the brine-plus-CO2 mix is constantly sampled, so that models can be compared with the length of time it takes for injected CO2 to arrive at the monitoring wells.
Because there are two observation wells, a two-dimensional profile of the progressing plume of CO2 can be derived. A string of seismic sources in one well sends signals to a string of sensors in the other. Seismic waves travel more slowly in CO2, and so the signals' arrivals are timed in order to show where the CO2 is located. This cross-well imaging is performed before and after CO2 injection; The technique provides information about the fluid properties as well as images of the CO2 distribution in the plane between the wells.
A similar set of images developed at the South Liberty field near Dayton, Texas—in which images were compared with modeled CO2 distributions—indicate that a "high" value for trapping is indeed the case. Those CO2 distributions were based on different values of a measure of how much residual gas is trapped.
At Cranfield, however, cross-well imaging will also make use of electric resistivity tomography. In this technology, the source string comprises electrodes, the signals of which will be used to make inferences about subsurface resistivity. "Brine is electrically conductive, while CO2 is not," explains Doughty. This means that images will use spatial variation of electrical resistivity to show how the CO2 plume changes over time.
Another trapping method relies on CO2 dissolution in the aquifer's resident brine. The process makes the brine denser, whereupon it sinks. "This is exciting because dense brine can be used to set up a convective cell," says Doughty. Scientists also want to know how CO2 reacts with minerals in the rock formation to form carbonate particles, still another factor that affects trapping.
CO2 in water or water in CO2
In addition to modeling and field studies, Pacific Northwest National Laboratory's (PNNL) Carbon Sequestration Initiative supports experiments to investigate the long-term viability of CO2 storage in deep geological reservoirs. Carbon dioxide dissolved in deep saline aquifers becomes highly reactive with minerals. Silicate minerals such as olivine, serpentine, pyroxene, and plagioclases, for instance, will potentially react with CO2 to produce metal carbonates, the only trapping mechanism that permanently converts the greenhouse gas into a stable form.
Most work in this area has focused on aqueous-dominated systems in which dissolved CO2 reacts to form crystalline carbonate minerals. Less attention has been paid to reactions that occur between minerals in the host rock and supercritical CO2 that has been saturated with dissolved water. But experiments are now underway to expose carbonated minerals to variably hydrated supercritical CO2 at different temperatures and pressures. In this way, scientists hope to determine what affects carbonation rates.[4,5]
In particular, the research aims to explain how the reactions occur and what phases form. "It's going to be important near the well borehole, where they do the injection," says John Loring, a geochemist at PNNL. Important because the borehole holds a very high concentration of supercricital CO2, which will not yet have dissolved into the aqueous phase—that's a process that happens deeper in the well. "You could have reactions that clog the well," explains Loring.
The other vital reason to study the hydrated CO2 is that the supercritical phase is more buoyant than the aqueous phase, so it could rise and interact with the minerals in the caprock. If that happens, the dominant phase in contact with the caprock could be the supercritical phase. Since the caprock is ultimately what holds the CO2 down, it is important to understand the chemical reactions that take place there.
The PNNL team works mostly with minerals found in basaltic formations, while their colleagues in the field inject CO2 into a test well within the extensive Columbia River basalts in eastern Washington State. The main reactive mineral of interest is olivine, a magnesium iron silicate common in Earth's subsurface. But the PNNL team also looks at serpentine minerals including antigorite, and pyroxene-like wollastonite. Those minerals differ in their chemical compositions and crystal structures.
When a mineral such as forsterite, the magnesium-rich species of olivine, is exposed to supercritical CO2 hydrate, the water adsorbs to the forsterite surface. "What we're finding is that the phases that form and their rates of formation depend on the thickness of the adsorbed water layer," says Loring. Thus, it appears that the thickness of the water film determines the rate and extent of carbonate mineral formation.
Whereas the water film thickness is on the order of angstroms to nanometers, the carbonate particles that form are polycrystalline on the order of hundreds of nanometers. Huge particles form by transport of ions through very thin water films. It's mechanisms such as this that Loring and his colleagues want to understand better.
The tools of the trade
To carry out its Carbon Sequestration Initiative, PNNL has developed a range of tools to observe the carbonation reaction as it happens. Reaction vessels contain hydrated supercritical CO2 that reacts with minerals at temperatures (around 50°C) and pressures (90 to 180 atm) representative of an injection well. Various spectroscopic techniques are then brought to bear on the physical and chemical composition of the solids as a function of time.
Additionally, quartz crystal microbalances are used to measure the change in mass of vessel contents as a precipitate forms. If a clay is exposed to hydrated supercritical CO2, the clay will take up water. With quartz crystal microbalance, changes in mass as small as 10 ng can be measured in samples that weigh as little as a few μg.
"We try to attack a problem with as many different techniques as we can," says Loring. Because each technique has different advantages and disadvantages, a combination provides the most detailed information about the carbonate products and how they form.
Now that the EPA has finalized its requirements for geologic sequestration as long-term storage, pilot sites in the US are injecting large volumes of CO2 into rock formations. A detailed knowledge of what makes the CO2 stay contained, and an ability to predict how it might move through the pore space, is vital for the success of these projects.
Trapping CO2 could turn out to be among the most important waste management projects our species has deployed to date. With our technological intervention, rocks could join plants in sequestering atmospheric CO2 and slowing global warming.
- V. Nunez-Lopez, S. D. Hovorka, "Subsurface monitoring of large-scale CO2 injection at SEACARBC's Phase III Cranfield Site," Carbon Management Technology Conference, 2012.
- T. M. Daly et al., "Time-lapse crosswell seismic and VSP monitoring of injected CO2 in a brine aquifer," Environ. Geol. 54, 1657 (2008).
- C. Doughty et al., "Site characterization for CO2 geologic storage and vice versa: the Frio brine pilot, Texas, USA as a case study," Environ. Geol. 54, 1635 (2008).
- Q. R. S. Miller et al., "Insights into silicate carbonation processes in water-bearing supercritical CO2 Fluids," Int. J Greenhouse Gas Control 15, 104 (2013).
- J. S. Loring et al., "In situ infrared spectroscopic study of forsterite carbonation in wet supercritical CO2," Environ. Sci. Technol. 45, 6204 (2011).
- J. S. Loring et al., "In situ infrared spectroscopic study of brucite carbonation in dry to water-saturated supercritical carbon dioxide," J Phys. Chem. 116, 4768 (2012).
- J. Doetsch et al., "Constraining CO2 simulations by coupled modeling and inversion of electrical resistance and gas composition data," Int. J Greenhouse Gas Control (in press).