Currently there is a great deal of interest in carbon dioxide for the recovery of both heavy and light oils. This paper deals with the efficiency of gaseous carbon dioxide as a recovery agent for moderately viscous oils. The paper gives model results, and compares and contrasts the findings with laboratory and field test observations, pointing out the range of conditions over which carbon dioxide is likely to be effective. Also, the current status of carbon dioxide for light oils is considered.

It was found that over the viscosity range of 1 to 1000 centipoises, carbon dioxide was superior to nitrogen injection, natural depletion, or waterflooding for oil viscosities above 70 cp. The gain over waterflooding was as much as 9 percentiles in oil recovery, being greater for the more viscous crudes. Oil saturation was an important variable as oil recovery decreased rapidly with a decrease in saturation. Another significant factor affecting ultimate oil recovery was the critical gas saturation. Viscous oils showed a 27% increase in recovery as the critical gas saturation varied from 0 to 10%. The blowdown recovery upon curtailment of carbon dioxide injection was about 1%; field values are as high as 4%. Reasons for this discrepancy are outlined. The amount of carbon dioxide left in the reservoir was used as a measure of the efficiency of the process; it was high for low oil saturations, especially for the more viscous oils.

An economic analysis of the carbon dioxide injection process showed that the economics are tenuous; a variety of factors in addition to the oil price would determine the economic viability of the process.


While there is little debate that a significant amount of oil remains held in the ground by current technical and economic constraints, opinion is widespread as to the proper recovery technique or techniques to unlock these reserves. Infill drilling and a handful of alternative recovery methods such as thermal, miscible and improved mobility floods compete for the over 330 billion barrels of United States and Canadian oil that remains in place.

Carbon dioxide injection, as one of these processes, has long been thought of as a miscible processes, has long been thought of as a miscible process best applied in light oils with gravities process best applied in light oils with gravities greater than 20 degrees API. However, immiscible carbon dioxide flooding, as part of the suite of enhanced oil recovery methods being tested, may be promising in the case of heavy, moderately viscous oils where carbon dioxide injection improves recovery by lowering oil viscosity and promoting swelling.

Deposits of heavy oil total several trillion barrels in the U.S., Venezuela and Canada. In the U.S. alone, there are over 2,000 heavy oil reservoirs occurring in 1500 fields in 26 states. The total resource is estimated at 106.8 billion barrels, of which some 60.9 billion barrels occur in fields which do not offer favorable conditions for thermal methods.


Oil displacement by carbon dioxide injection relies on a number of mechanisms related to the phase behavior of CO2-crude oil mixtures. Strongly dependent on reservoir temperature, pressure and crude oil composition, the dominant displacement characteristics exhibited usually fall into one of five categories as shown in Figure 1.

At reservoir pressures below 1000 psia (Region I), the major effects of carbon dioxide injection on oil recovery appear due to the solubility of carbon dioxide in the crude oil.

At reservoir pressures above 1000 psia several phase behavior characteristics may dominate the phase behavior characteristics may dominate the recovery process. If reservoir pressure is increased and falls into Region II, Holm reports the start of substantial hydrocarbon extraction by gaseous carbon dioxide. Holm and Yellig agree that this gas may, if conditions are favorable (Region III), vaporize enough of the C2-C30 fractions of the reservoir crude to establish miscibility.

P. 137

This content is only available via PDF.
You can access this article if you purchase or spend a download.