Fluid-Rock Interactions Resulting From the Underground Disposal of Carbon Dioxide

C. A. Rochelle Fluid Processes Group, British Geological Survey, Keyworth, Nottingham, NG12 SGG, UK


K. Bateman Fluid Processes Group, British Geological Survey, Keyworth, Nottingham, NG12 SGG, UK

J. Pearce Mineralogy and Petrology Group, British Geological Survey, Keyworth, Nottingham, NG12 SGG, UK

It is now generally accepted that anthropogenic CO2 emissions are contributing to the global rise in atmospheric CO2 levels causing a rise in the radiative heating of the Earth's surface and lower atmosphere - an enhanced greenhouse effect. For large point sources (e.g. power stations), CO2 emissions could be reduced by separating it from flue gases, compressing it, then transporting it by pipeline for deep underground disposal. At depths below approximately 800 m, subsurface conditions would allow the CO2 to exist as a supercritical fluid, having a density of approximately 0.7 kg.l-1. This means that it would take up relatively little volume compared to the gaseous form. Previous studies (Gunter et al., 1993; Czernichowski-Lauriol et al., 1996) have shown that disposal of CO2 into silicoclastic rock types (e.g. sandstones) is preferable to that in carbonate rock types (e.g. chalk) because of the greater potential for pH buffering and net carbonate mineral precipitation, resulting in the long term immobilisation of CO2. For Europe, the large scale disposal of CO2 is most practicable within deep Triassic sandstones beneath the North Sea, which have potential storage capacity equivalent to several hundred years of present-day CO2 production (Holloway et al., 1995). However, for safe, long term disposal, it is important to understand how CO2 will interact with host rocks, associated porewaters, and caprocks. This communication briefly summarises the findings of recent experimental investigations and natural analogue studies.

Laboratory experiments were conducted at 80 °C and 200 bars pressure, conditions representative of deep sandstones beneath the North Sea. Under such conditions, the supercritical CO2 has a solubility in pure water of approximately 6 g.1-1. However, in order to represent more closely saline porewaters found within deep sandstones, a O.SSM NaCl solution was used as the aqueous phase in the experiments, in which supercritical CO2 has a solubility of approximately 5 g.1-1. This salinity (approximately that of seawater) allowed for computer modelling of the solutions by conventional programmes and databases (Czernichowski-Lauriol et al., 1996). Two main experimental approaches have been used to study CO2-porewater-host rock interactions. Firstly, a range of sandstones and caprocks were reacted in simple, long term batch experiments containing both supercritical CO2 and NaCl solution. Experiments lasted up to 8 months to investigated the types, and significance, of reaction mechanisms occurring. Secondly, more complex flow-through experiments were conducted, in which cores of sandstone were flushed with both dry supercritical CO2, and NaCl solution previously equilibrated with supercritical CO2. These lasted up to two months and investigated the reactions in a flowing system more akin to that expected during disposal operations.

In general, the reaction of CO2 with sandstone minerals led to K-feldspar dissolution, and precipitation of clay phases and
a little calcite. This is similar to the observations of previous workers (Gunter et al., 1993) who suggested dissolution of
Ca-feldspar followed by the formation of calcite and kaolinite. However, North Sea sandstones do not contain appreciable
Ca-feldspar because this is easily altered during diagenesis, but they often contain anhydrite cement. Our experiments showed anhydrite dissolution followed by precipitation of elongate calcite crystals. Anhydrite dissolution will probably be the largest source of Ca within North Sea sandstones, and hence have a major control on calcite precipitation. Being less dense than water, supercritical CO2 will rise within the host formation until it reaches a low permeability lithology. Therefore, samples of different caprocks were reacted in the batch experiments to investigate caprock integrity. In caprocks, little reaction was observed with the clay phases, however K-feldspar and dolomite within mudstones dissolved. Studies of natural accumulations of CO2 within the USA indicate that the sealing capacity of both mudstone and anhydrite caprocks are not adversely affected by the presence of CO2 over geologic timescales, possibly due to the precipitation of secondary phases and sealing off of the reacting minerals. However, the natural CO2 fields show a decrease in permeability within the sandstones due to the formation of authigenic clays following initial mineral dissolution.


This abstract is published with the permission of the Director of the British Geological Survey.


Czernichowski-Lauriol, I., Rochelle, C., Bateman, K., Pearce, J., Sanjuan, B. & Blackwell, P., Proceedings of the international symposium on the scientific and engineering aspects of deep injection disposal of hazardous and industrial wastes, Lawrence Berkeley Laboratory, May 10-13, 1994, Academic Press (1996, in press).

Gunter, W.D., Perkins, E.H. & McCann, T.J., Proceedings of the IEA Carbon Dioxide Disposal Symposium (Oxford, England) 29-31, March 1993, Energy Conversion Management, 34, 941-948, Pergamon Press, Oxford (1993).

Holloway, S., Heederik, J.P., van der Meer, L.G.H., Czernichowski-Lauriol, I., Harrison, R., Lindeberg, E., Summerfield, I.R., Rochelle, C., Schwarzkopf, T., Kaarstad, O. & Berger, B., The underground disposal of carbon dioxide, final report of JOULE II project No. CT92-003 1, summary report (1995).