MORE carbon dioxide is making its way into our atmosphere as we burn fossil fuels and deforest tropical lands. Most experts agree that increased emissions of greenhouse gases-especially carbon dioxide-are responsible for an overall warming of our planet over the last 150 years. In 1991, Norway became the first country to impose a federal tax on atmospheric CO2 emissions from combustion-based point sources such as coal-fired power plants. Shortly thereafter, this tax-$55 per ton of CO2-was extended to include emissions associated with offshore oil and gas production. The day is not far off when other countries, possibly including the U.S., will follow Norway's lead, thus creating a strong financial incentive to develop strategies for safe disposal of CO2 waste streams. One such strategy is to capture excess CO2 and inject it underground, where it will remain sequestered from the atmosphere for thousands of years. Geochemist James W. Johnson is heading a Livermore team that is developing criteria for identifying subsurface geologic formations that could be used for CO2 sequestration. "Our work is part of a long-term Department of Energy effort to identify optimal sites for sequestering CO2," says Johnson. Although CO2 injection is a technique commonly used for enhancing the recovery of oil, large-scale injection for the sole purpose of isolating CO2 from the atmosphere is occurring at just one place today: the offshore Sleipner facility, owned and operated by Statoil, Norway's state oil company. Located beneath the Norwegian sector of the North Sea, the extensive Sleipner West natural gas field is characterized by a high (9 percent) concentration of CO2, well above the 2.5-percent limit imposed by European export specifications. Statoil strips excess CO2 from the recovered gas in a tower on its offshore production platform before exporting the gas to the European community. Injecting the captured CO2 into a confined aquifer-800 meters below the seabed and 2,500 meters above the Sleipner West hydrocarbon reservoir-results in no tax on Statoil for its atmospheric emissions. Since 1996, Statoil has injected about a million tons of CO2 per year and saved $55 million per year in taxes. The injection facility cost just $80 million to construct, and its operation accounts for less than 1 percent of overall production costs. At Sleipner, geologic sequestration has proved to be an environmentally sound and financially prudent disposal option for excess CO2. Starting with simulations of CO2 injection at the Sleipner site, Johnson and his collaborators, Carl Steefel and John Nitao, are developing a general modeling capability for analyzing CO2 sequestration in geologic formations. This Livermore team is uniquely qualified to forge this capability, given their experience in developing an internationally recognized suite of reactive transport simulators (GIMRT, NUFT), supporting geochemical software (SUPCRT92), and thermodynamic-kinetic databases (GEMBOCHS). Using this integrated toolbox, they have begun to identify the geochemical, hydrologic, and structural constraints on successful geologic CO2 sequestration. Eventually, they will correlate these constraints with the characteristics of potential geologic formations, rank their overall sequestration performance based on this correlation, and thus identify optimal injection sites.

Modeling a Dynamic System Reactive transport modeling integrates the geochemical, hydrological, and mechanical processes that characterize dynamic geologic systems. These processes, which include chemical reactions, fluid flow, heat transfer, and mechanical stress and strain, are interdependent and must be modeled simultaneously to simulate the true behavior of geologic systems. Simultaneous modeling was not possible for complex geologic systems until the advent of massively parallel supercomputers. Now, Johnson's team is producing the first-ever reactive transport simulations of CO2 injection and sequestration within geologic formations. Their initial Sleipner simulations examine what happens to CO2 after it is pumped into its watery grave. At Sleipner, the storage formation is a highly porous, fluid-saturated sandstone aquifer, sealed at both the top and bottom by thick, relatively impermeable shale. The CO2 moves through the formation via several migration processes and at the same time is trapped by various sequestration processes. The CO2 migrates by displacing ambient water, with which it is largely immiscible, and by rising relative to this water, owing to its lower density. It also moves faster than the ambient fluid because of its lower viscosity. As the CO2 plume migrates, some of it may react with formation minerals to precipitate carbonates (mineral trapping), some dissolves into the formation waters (solubility trapping), and some may eventually be isolated within anticlinal structures bound by the shale cap (hydrodynamic trapping). Understanding the relative effectiveness of these competing migration and sequestration processes is the key to identifying sites that will provide optimal sequestration performance. Reactive transport modeling represents a unique capability for quantifying this balance of processes.

First Results The results of preliminary, two-dimensional NUFT and GIMRT simulations of CO2 injection at Sleipner are shown in the three figures here, which illustrate the relative effectiveness of various sequestration processes after one year of injection. The figure above illustrates the profound dependence of the CO2 plume's location on the absence or presence of thin shale barriers within the aquifer. Without these layers, the CO2 plume rises quickly to the aquifer caprock, where it then migrates laterally beneath this impermeable seal. When low-permeability shale units are present, as they are at Sleipner, they effectively retard the plume's vertical migration while promoting its lateral extension. The shale layers not only delay the arrival of the CO2 plume at the caprock but also increase tremendously the volumetric extent of plume interaction with the aquifer and thus the potential for solubility and mineral trapping. The figure at top shows the spatial distribution of aqueous CO2 concentrations when shale layers are present. It indicates that about 3 percent of the total injected CO2 has dissolved into the ambient formation waters. Thus, solubility trapping represents a small but measurable contribution to aggregate sequestration. The contribution of mineral trapping is also small but measurable. Precipitation of carbonates requires the presence of appropriate elements within formation minerals. In this Sleipner simulation, only a small concentration of one such element (calcium) is present in a single formation mineral (plagioclase), also of small concentration. Hence, mineral trapping is limited to minor calcite precipitation at the expense of plagioclase dissolution-a very slow process relative to solubility trapping. After 1 year, calcite precipitation has sequestered less than 1 percent of the injected CO2. In this preliminary 1-year simulation, solubility and mineral sequestration account for less than 4 percent of the injected CO2. However, the relative effectiveness of solubility and especially of mineral trapping may be significantly increased over longer time frames within formations whose ambient fluid composition and mineralogy are different from Sleipner's. Johnson is quick to note, "Our research is first-cut reactive transport modeling of the complex CO2 injection-sequestration problem." Other potentially significant effects will be evaluated in future work.

A Collaboration Begins The preliminary reactive-transport simulations of CO2 injection that Johnson's team carried out at Sleipner used site-specific technical data available in the public domain, but these data are insufficient for further detailed modeling efforts. Livermore recently initiated a collaboration with the International Energy Association (IEA), which coordinates research and development and monitoring of the Saline Aquifer CO2 Storage (SACS) project at Sleipner. As part of the collaboration, IEA-SACS will supply Livermore with additional Sleipner data, which will permit more highly resolved simulations. These improved models will yield new insights into the current injection process and perhaps ways to improve sequestration performance at Sleipner. For Livermore and the Department of Energy, obtaining more data for the unique Sleipner CO2 sequestration project-and developing a general modeling capability based on Sleipner simulations-is invaluable. The problem of excess CO2 must be solved, geologic sequestration represents a potentially promising solution, and reactive-transport modeling provides a unique way to identify optimal geologic formations for sequestration in the U.S. —Katie Walter

Key Words: carbon dioxide sequestration, reactive transport modeling.

For more information contact James W. Johnson (925) 423-7352 (jwjohnson@llnl.gov).