Enhanced Oil Recovery — a complete guide to thermal, gas, and chemical methods

Primary and secondary recovery typically leave between half and two-thirds of the original oil in place still trapped in the reservoir. Enhanced oil recovery is the family of techniques that targets that residual oil — by adding heat, by injecting gases that mix with the crude, or by injecting chemicals that change how oil and water move through rock. This guide walks through every major EOR method, the physics behind each, the reservoir conditions where each works, and the screening criteria practitioners use to decide which one to deploy.

Why oil is left behind

To understand EOR, you first have to understand why so much oil stays underground after conventional production. Recovery efficiency is the product of two distinct factors: displacement efficiency at the pore scale, and sweep efficiency at the reservoir scale.

Displacement efficiency is governed by capillary forces. Oil is trapped in pore throats as disconnected droplets, held in place by interfacial tension between oil and water. The dimensionless number that controls whether a droplet moves is the capillary number, which balances viscous forces against capillary forces.

In a typical waterflood, the capillary number is around 10^-7, far too low to mobilize residual oil. To meaningfully reduce residual oil saturation, the capillary number must increase by three to four orders of magnitude. There are only two practical levers: increase the viscous force (hard — it would require impractical injection rates), or reduce interfacial tension by orders of magnitude. Most chemical and miscible EOR methods attack this term directly.

Sweep efficiency, the other factor, is governed by the mobility ratio between displacing and displaced fluids. When water — which is far less viscous than heavy or even medium crude — pushes oil, it fingers through and bypasses large volumes. Improving the mobility ratio is the second great lever of EOR.

The three families of EOR

EOR methods — sometimes called tertiary recovery, though the term is imprecise since EOR can be applied at any stage — divide naturally into three families based on what is injected and the dominant recovery mechanism.

Thermal methods

Thermal EOR works on a single principle: heat reduces oil viscosity, often dramatically. A heavy crude at 10,000 cp at reservoir temperature can drop below 10 cp when heated to 200 °C. Viscosity reduction improves the mobility ratio, mobilizes oil that was effectively immobile, and is the only practical route to recovering the world's vast heavy-oil and bitumen resources. Thermal methods account for the largest share of EOR production worldwide.

Steam flooding (steam drive)

Continuous steam injection through dedicated injector wells creates a steam zone that advances through the reservoir, heating and displacing oil toward producers. As steam condenses at the advancing front, it gives up its latent heat to the formation and the oil. The mechanism is a combination of viscosity reduction, steam distillation of lighter components, thermal expansion, and a hot-waterflood behind the condensation front.

Steam flooding is the most widely applied thermal method and one of the most mature EOR technologies overall. It is best suited to relatively shallow reservoirs (heat losses make deep steam injection uneconomic), with high porosity and permeability to accept steam, and thick pay zones to limit vertical heat loss to overburden and underburden. Recovery factors of 50–65% of OOIP are achievable in favorable settings.

Steam quality and the heat the steam carries

The reason steam is so much more effective than hot water is latent heat. When a kilogram of water is boiled to steam, it absorbs roughly 2,000–2,250 kJ of latent heat of vaporization — several times the sensible heat needed to raise that same water from reservoir temperature to boiling. The steam carries this large energy payload into the formation and releases it all at once when it condenses at the steam front. Steam quality — the mass fraction that is actually vapor rather than entrained hot water — therefore governs how much heat each unit of injected steam delivers. Wellbore and surface heat losses degrade quality before the steam reaches the sand, which is one reason steam injection is depth-limited.

The classical model for how the heated zone grows is the Marx-Langenheim solution (1959), which treats steam-zone expansion as an energy balance: injected heat goes partly into expanding the steam zone and partly into conduction losses to the over- and underburden. The heated area grows quickly at first and then asymptotes as conduction losses catch up with injection — giving the engineer a way to estimate heated volume, oil-displacement rate, and the optimal time to convert from steam drive to blowdown.

Cyclic steam stimulation (CSS, "huff and puff")

CSS uses a single well for both injection and production in alternating cycles. Steam is injected for a period of weeks (the "huff"), the well is shut in to allow heat to soak into the formation, and then the same well is put on production (the "puff") until rates decline, at which point the cycle repeats. CSS gives faster payback than steam flooding because each well produces almost immediately, but ultimate recovery is typically lower since only the near-wellbore region is heated. It is frequently used as a precursor to steam flooding.

Steam-assisted gravity drainage (SAGD)

SAGD uses a pair of horizontal wells drilled one above the other, typically 5 m apart. Steam is injected continuously into the upper well, forming a steam chamber that grows upward and outward. Heated oil and condensate drain by gravity to the lower production well. SAGD is the enabling technology for in-situ bitumen recovery in oil sands too deep to mine, and it can achieve recovery factors above 50% in suitable reservoirs. It requires thick, high-quality sand with vertical permeability and minimal shale barriers that would impede chamber growth.

In-situ combustion (fire flooding)

In-situ combustion ignites a portion of the oil in the reservoir and sustains a combustion front by injecting air or oxygen. The burning front — typically 350–600 °C — moves through the reservoir, and ahead of it a cascade of zones forms: a combustion zone, a cracking/coking zone, a steam and vaporization zone, a condensation bank, and finally an oil bank being driven toward the producers. The fuel consumed is the heaviest fraction of the crude (coke), so the process effectively upgrades the produced oil in place.

Combustion can be dry (air only) or wet (water co-injected to recover heat from behind the front and improve efficiency). It reaches deeper reservoirs than steam since heat is generated in the formation rather than transported from surface. But it is operationally complex and difficult to control — front stability, oxygen breakthrough, and corrosion are persistent challenges — which has limited its field adoption despite its thermodynamic appeal.

Gas & miscible methods

Gas injection methods displace oil with an injected gas. Their power comes from miscibility — when the injected gas and reservoir oil mix into a single phase, interfacial tension vanishes and residual oil saturation can approach zero. The key parameter is the minimum miscibility pressure (MMP): the reservoir pressure above which the gas and oil become miscible. Below MMP, the displacement is immiscible and far less effective.

How miscibility develops — the three drive mechanisms

Miscibility is not a single phenomenon. There are two fundamentally different ways a gas can become miscible with oil, and a third hybrid that describes most real CO₂ floods.

• First-contact miscibility (FCM): the injected fluid mixes with the oil in all proportions on the very first contact, with no phase boundary at any mixing ratio. LPG and rich solvent slugs can be first-contact miscible. FCM is the simplest case but requires expensive injectants, so it is rare in continuous floods.
• Vaporizing (lean) gas drive: a lean gas (methane, nitrogen, flue gas) progressively extracts intermediate components (C₂–C₆) from the oil into the advancing gas front. After repeated contacts the leading edge of the gas becomes enriched enough to be miscible with fresh oil ahead. Miscibility is generated dynamically at the displacement front.
• Condensing (enriched) gas drive: the opposite mechanism — an enriched gas carries intermediates that condense into the oil at the trailing edge, progressively enriching the oil until it becomes miscible with the injected gas behind the front.

In practice, CO₂ displacement is best described as a combined condensing/vaporizing mechanism: CO₂ both condenses into the oil and vaporizes intermediates out of it, and the dynamic miscibility that develops is a blend of the two idealized drives. This is why CO₂ achieves miscibility at far lower pressure than nitrogen — it has access to both pathways and is an unusually effective solvent for intermediate hydrocarbons.

MMP is measured in the laboratory with a slim-tube test (a long, thin sand-packed coil where recovery is measured against pressure; MMP is the pressure above which recovery flattens near 90–95%) or a rising-bubble apparatus, and estimated with correlations (Holm & Josendal, Yellig & Metcalfe, Cronquist) as a screening first pass. MMP rises with reservoir temperature and with the heaviness of the oil, and falls as the injected gas is enriched with intermediates.

CO₂ miscible flooding

Carbon dioxide is the most important miscible gas in practice. It does not become miscible with oil on first contact at typical reservoir conditions, but achieves miscibility through a multiple-contact (dynamic) process: as CO₂ contacts the oil, it extracts intermediate hydrocarbons (C₂–C₆) into the gas phase and condenses into the oil, and after repeated contacts the compositions converge until the interface disappears. CO₂ also swells the oil and reduces its viscosity, both of which aid recovery.

CO₂ flooding requires sufficient pressure to reach MMP — typically 1,100–2,500 psi depending on oil composition and temperature — which sets a practical minimum depth. It works best on light to medium oils (API generally above 25–27°) where MMP is achievable below the fracture pressure. A persistent challenge is the unfavorable mobility ratio: CO₂ is far less viscous than oil, so it tends to finger and override, which is why it is so often deployed as WAG (see below). CO₂ flooding has the additional appeal of sequestering carbon, making it central to carbon-capture-and-storage economics.

Nitrogen and flue gas

Nitrogen achieves miscibility only at very high pressures, so it is restricted to deep, light-oil reservoirs — but it is cheap to source (from air) and chemically inert. It is sometimes used to maintain reservoir pressure or to drive a previously injected solvent slug. Flue gas (a mix of N₂ and CO₂ from combustion) offers a lower-cost alternative where pure CO₂ is unavailable, trading some miscibility performance for supply economics.

Hydrocarbon gas injection

Injecting hydrocarbon gases — lean natural gas, enriched gas, or LPG (propane/butane) — can achieve miscibility with light oils. Enriched-gas and LPG slug processes were among the earliest miscible methods. They are highly effective but economically sensitive: the injected hydrocarbons have their own market value, so the method is most attractive where gas is stranded or where it can be recovered and recycled.

Water-alternating-gas (WAG)

WAG addresses the central weakness of all gas injection — poor sweep due to low gas viscosity and gravity override — by alternating slugs of gas and water. The water improves the mobility ratio and the vertical sweep, while the gas provides the miscible displacement at the pore scale. CO₂-WAG is now the dominant form of CO₂ EOR in mature basins. Variants include simultaneous water-and-gas (SWAG) and foam-assisted processes that use surfactant to stabilize the gas as foam and further improve conformance.

Chemical methods

Chemical EOR injects water-soluble chemicals to attack the capillary and mobility problems directly. The three core agents — polymer, surfactant, and alkali — are used alone or in combination. Chemical methods occupy the medium-oil middle ground and are most often applied to reservoirs that have already been waterflooded, where significant mobile oil remains but conventional waterflooding can no longer recover it.

Polymer flooding

Polymer flooding adds high-molecular-weight, water-soluble polymers — typically partially hydrolyzed polyacrylamide (HPAM) or xanthan biopolymer — to injection water. The polymer thickens the water, dramatically improving the mobility ratio and forcing the injected fluid to sweep zones it would otherwise bypass. Polymer flooding does not reduce residual oil saturation at the pore scale; it improves volumetric sweep efficiency. It is the simplest, cheapest, and most widely deployed chemical method.

The practical limits are set by polymer stability. HPAM degrades at high temperature and high salinity (especially with divalent ions like Ca²⁺ and Mg²⁺), losing viscosity. Xanthan tolerates salinity better but is vulnerable to bacterial degradation. Polymer flooding favors moderate temperatures, moderate salinity, and good reservoir continuity so the thickened water can do its work.

Resistance factor, rheology, and inaccessible pore volume

Polymer performance is quantified by two field-measured parameters. The resistance factor is the ratio of water mobility to polymer-solution mobility — effectively how much harder the polymer makes it to push fluid through the rock. The residual resistance factor is the permeability reduction that persists after the polymer slug has passed, because adsorbed and retained polymer continues to restrict the high-permeability paths even once injection reverts to water.

HPAM solutions are non-Newtonian. At the low shear rates deep in the reservoir they are shear-thinning (viscosity drops as shear rises), which helps injectivity near the wellbore where shear is high. But at very high shear — in pumps, chokes, and perforations — HPAM can undergo irreversible mechanical degradation, snapping the polymer chains and permanently losing viscosity. Facility and completion design has to keep shear below that threshold.

A subtler effect is inaccessible pore volume (IAPV): the large polymer molecules cannot enter the smallest pores that water can, so the polymer front actually travels faster than a passive tracer in the same flood — sometimes arriving at producers ahead of where simple volumetrics would predict. IAPV partially offsets retention losses and must be accounted for in slug sizing.

Surfactant flooding

Surfactants are amphiphilic molecules that reduce the oil-water interfacial tension by orders of magnitude — from tens of dynes/cm to ultralow values around 10^-3 dynes/cm. This is the direct attack on the capillary number described earlier: lower IFT raises N_c by the three to four orders of magnitude needed to mobilize trapped residual oil. Mobilized oil coalesces into an oil bank that is then displaced to producers. Surfactant flooding can recover oil that no other mechanism reaches, but the chemicals are expensive, sensitive to reservoir salinity and temperature, and prone to adsorption losses on the rock surface.

Winsor phase behavior — where ultralow IFT comes from

The ultralow interfacial tension that makes surfactant flooding work is not a property of the surfactant alone — it emerges from the phase behavior of the surfactant-oil-brine system. P. A. Winsor classified these systems into three types based on how the surfactant partitions between the oil and water phases, and which microemulsion phase forms. Understanding Winsor phase behavior is the single most important concept in surfactant formulation design.

• Winsor Type I (lower-phase microemulsion): the surfactant is more soluble in water. An oil-in-water microemulsion forms in equilibrium with excess oil on top. The surfactant sits mostly in the lower (aqueous) phase. Typical at low salinity.
• Winsor Type II (upper-phase microemulsion): the surfactant is more soluble in oil. A water-in-oil microemulsion forms in equilibrium with excess water below. The surfactant sits mostly in the upper (oleic) phase. Typical at high salinity.
• Winsor Type III (middle-phase microemulsion): a third, bicontinuous microemulsion phase forms between excess oil above and excess water below. The surfactant concentrates in this middle phase, which co-solubilizes large and roughly equal volumes of both oil and water. This is the regime that produces the lowest IFT.

As a formulation variable — most commonly salinity — is increased, the system transitions Type I → Type III → Type II. The transition is driven by the surfactant's affinity shifting from water toward oil. The middle-phase (Type III) window is the target: it is here that interfacial tension between the microemulsion and both excess phases drops to the ultralow values (10^-3 dynes/cm or lower) needed to mobilize residual oil.

Alkaline flooding

Alkaline (caustic) flooding injects an alkali such as sodium carbonate or sodium hydroxide. With crudes that contain natural organic acids, the alkali reacts in situ to generate surfactant — soap — at the oil-water interface, lowering IFT without the cost of synthetic surfactant. Alkali also alters rock wettability and reduces surfactant adsorption when used in combination. Its effectiveness depends strongly on the acid number of the crude.

ASP — alkaline-surfactant-polymer

ASP combines all three agents to capture their synergies in a single process. The alkali generates in-situ soap and protects the surfactant from adsorption, the surfactant drives IFT to ultralow values to mobilize residual oil, and the polymer provides the mobility control to sweep the mobilized oil bank efficiently. ASP is the most powerful chemical EOR process and has produced incremental recoveries above 20% of OOIP in field projects. The cost is operational complexity: careful formulation design, sensitivity to reservoir conditions, and scaling and emulsion problems in surface facilities.

Emerging & hybrid methods

Microbial EOR (MEOR)

MEOR uses microorganisms — either injected or stimulated in situ with nutrients — to produce biosurfactants, biopolymers, gases, and acids that mobilize oil. The appeal is low cost and the use of renewable, biodegradable agents. In practice MEOR has remained largely at the pilot scale: controlling microbial activity deep in a reservoir, and predicting and sustaining the desired metabolic products, are difficult. It is best regarded as a promising but still-maturing technology.

Low-salinity and smart waterflooding

Injecting water of carefully engineered ionic composition — typically lower salinity than formation brine, or with tuned ratios of specific ions — can shift rock wettability toward more water-wet conditions and release additional oil. The mechanisms are still debated (multi-ion exchange, double-layer expansion, fines migration), but the low cost and simplicity have made low-salinity waterflooding one of the most actively researched EOR areas of the past two decades.

Foam and conformance control

Foam — gas dispersed in a surfactant-stabilized liquid — is increasingly used to improve the conformance of gas-injection and steam processes. By raising the apparent viscosity of the gas phase and selectively blocking high-permeability thief zones, foam diverts injected fluids into the unswept matrix. It is a control technology layered onto other methods rather than a standalone process.

Screening criteria

The seminal work on EOR screening is Taber, Martin, and Seright's "EOR Screening Criteria Revisited" (SPE 12069, 1983 and updated in 1997), which compiled technical screening guidelines from field experience across all major methods. These criteria — refined by the National Petroleum Council's 1984 study and the Green & Willhite SPE textbook — remain the industry's first-pass filter. The table below summarizes typical ranges; they are starting points, not hard rules, and every prospect requires reservoir-specific evaluation.

The patterns in the table reflect the underlying physics. Thermal methods take the heavy, viscous, shallow oils that gas and chemical methods cannot touch. Miscible gas takes the light, low-viscosity oils where MMP is achievable. Chemical methods occupy the medium range. Depth limits are mechanistic: steam fails deep because of heat loss; gas fails shallow because MMP cannot be reached below fracture pressure.

Typical incremental recovery

Screening tells you what can work; the next question is how much oil each method typically recovers. The figures below are broad industry ranges for incremental recovery above waterflood, expressed as a percentage of original oil in place (OOIP). Actual results vary enormously with reservoir quality, heterogeneity, and execution — treat these as orientation, not promises.

Two patterns stand out. Thermal methods on heavy oil deliver the largest absolute recoveries simply because there is so much immobile oil to mobilize. And the chemical methods that attack residual oil directly — surfactant and ASP — can match thermal numbers on the right reservoir, which is why they continue to attract investment despite their complexity.

Choosing a method

In practice the choice of EOR method is driven first by oil type and reservoir conditions, then narrowed by economics and operational practicality. A workable mental model:

• Heavy oil and bitumen (< 20° API): thermal is almost always the answer. Shallow and mineable — surface processing. Deep — SAGD or CSS. Steam flooding for accessible heavy oil with good injectivity.
• Light to medium oil (> 25° API), deep, low viscosity: miscible gas. CO₂ if a source is available and MMP is reachable; hydrocarbon gas if stranded gas exists; nitrogen for very deep light-oil reservoirs.
• Medium oil, post-waterflood, moderate conditions: chemical. Polymer for a mobility problem; surfactant or ASP where significant residual oil remains and the economics justify the chemical cost.
• Always cross-check against: heterogeneity (thief zones wreck any flood), existing well spacing and infrastructure, chemical/water compatibility, and — increasingly — the carbon balance, where CO₂ EOR can turn a cost into a credit.

The systematic way to run this first pass is a screening tool that scores every method against your reservoir's parameters at once, so you see the full ranking rather than evaluating methods one at a time. That is exactly what the RFour Energy EOR screening tool does.

Three takeaways

1. Every method pulls one of two levers. EOR either reduces interfacial tension to mobilize trapped oil, or improves the mobility ratio to sweep more of the reservoir — usually by reducing oil viscosity, eliminating IFT through miscibility, or thickening the displacing fluid. Understanding which lever a method pulls tells you when it will and won't work.
2. Oil type sets the family. Heavy oil goes thermal, light oil goes miscible gas, medium oil goes chemical. The screening criteria are not arbitrary — they are the physics of viscosity reduction, minimum miscibility pressure, and chemical stability expressed as ranges.
3. Screening narrows; it doesn't decide. Published criteria after Taber et al. quickly eliminate unsuitable methods, but the final selection always requires lab work, simulation, a pilot, and economics. Use screening to focus expensive study on the two or three candidates that survive the first pass.

References & further reading:
Taber, J. J., Martin, F. D., Seright, R. S. (1983). EOR Screening Criteria Revisited. SPE 12069.
Taber, J. J., Martin, F. D., Seright, R. S. (1997). EOR Screening Criteria Revisited — Part 1 & 2. SPE Reservoir Engineering, 12(3).
Green, D. W., Willhite, G. P. (1998). Enhanced Oil Recovery. SPE Textbook Series Vol. 6.
Lake, L. W. (1989). Enhanced Oil Recovery. Prentice Hall.
National Petroleum Council (1984). Enhanced Oil Recovery. NPC Report, Washington D.C.
Stosur, G. J., Hite, J. R., Carnahan, N. F., Miller, K. (2003). The Alphabet of Oil Recovery Processes. SPE 84864.
Sheng, J. J. (2011). Modern Chemical Enhanced Oil Recovery: Theory and Practice. Elsevier.
Butler, R. M. (1991). Thermal Recovery of Oil and Bitumen. Prentice Hall.
Holm, L. W., Josendal, V. A. (1974). Mechanisms of Oil Displacement by Carbon Dioxide. JPT, 26(12).
Marx, J. W., Langenheim, R. H. (1959). Reservoir Heating by Hot Fluid Injection. Trans. AIME, 216.
Yellig, W. F., Metcalfe, R. S. (1980). Determination and Prediction of CO₂ Minimum Miscibility Pressures. JPT, 32(1).
Stalkup, F. I. (1983). Miscible Displacement. SPE Monograph Vol. 8.
Winsor, P. A. (1954). Solvent Properties of Amphiphilic Compounds. Butterworths, London.
Huh, C. (1979). Interfacial Tensions and Solubilizing Ability of a Microemulsion Phase. J. Colloid Interface Sci., 71(2).
Healy, R. N., Reed, R. L., Stenmark, D. G. (1976). Multiphase Microemulsion Systems. SPE 5565.
Geffen, T. M. (1973). Oil Recovery Potential from Laboratory Surfactant Floods. JPT, 25(7).