The need to increase the efficiency of oil recovery and environmental concerns are bringing to prominence the use of carbon dioxide (CO₂) as a tertiary recovery agent. Assessment of the impact of flooding with CO₂ all eligible reservoirs in the United States not yet undergoing enhanced oil recovery (EOR) requires making the best possible use of the experience gained in 40 years of applications. Review of the publicly available literature has located relevant CO₂-EOR information for 53 units (fields, reservoirs, pilot areas) in the United States and 17 abroad.

As the world simultaneously faces an increasing concentration of CO₂ in the atmosphere and a higher demand for fossil fuels, the CO₂-EOR process continues to gain popularity for its efficiency as a tertiary recovery agent and for the potential for having some CO₂ trapped in the subsurface as an unintended consequence of the enhanced production (Advanced Resources International and Melzer Consulting, 2009). More extensive application of CO₂-EOR worldwide, however, is not making it significantly easier to predict the exact outcome of the CO₂ flooding in new reservoirs. The standard approach to examine and manage risks is to analyze the intended target by conducting laboratory work, running simulation models, and, finally, gaining field experience with a pilot test. This approach, though, is not always possible. For example, assessment of the potential of CO₂-EOR at the national level in a vast country such as the United States requires making forecasts based on information already available.

Although many studies are proprietary, the published literature has provided reviews of CO₂-EOR projects. Yet, there is always interest in updating reports and analyzing the information under new perspectives. Brock and Bryan (1989) described results obtained during the earlier days of CO₂-EOR from 1972 to 1987. Most of the recovery predictions, however, were based on intended injections of 30 percent the size of the reservoir’s hydrocarbon pore volume (HCPV), and the predictions in most cases badly missed the actual recoveries because of the embryonic state of tertiary recovery in general and CO₂ flooding in particular at the time. Brock and Bryan (1989), for example, reported for the Weber Sandstone in the Rangely oil field in Colorado, an expected recovery of 7.5 percent of the original oil in place (OOIP) after injecting a volume of CO₂ equivalent to 30 percent of the HCPV, but Clark (2012) reported that after injecting a volume of CO₂ equivalent to 46 percent of the HCPV, the actual recovery was 4.8 percent of the OOIP. Decades later, the numbers by Brock and Bryan (1989) continue to be cited as part of expanded reviews, such as the one by Kuuskraa and Koperna (2006). Other comprehensive reviews including recovery factors are those of Christensen and others (2001) and Lake and Walsh (2008). The Oil and Gas Journal (O&GJ) periodically reports on active CO₂-EOR operations worldwide, but those releases do not include recovery factors. The monograph by Jarrell and others (2002) remains the most technically comprehensive publication on CO₂ flooding, but it does not cover recovery factors either.

This chapter is a review of the literature found in a search for information about CO₂-EOR. It has been prepared as part of a project by the U.S. Geological Survey (USGS) to assess the incremental oil production that would be technically feasible by CO₂ flooding of all suitable oil reservoirs in the country not yet undergoing tertiary recovery.