CARBON DIOXIDE FLOODING FOR EKHANCED OIL RECOVERY:

PROMISE AND PROBLEMS

by

F.M. Orr, Jr., J.P. Heller and J.J. Taber

New Mexico Petroleum Recovery Research Center
r,ew' Mexico Institute of Hining and Technology

A'bstract

liW!!' Of the enhanced oil recovery methods currently being considered for
application to many of the nation's older oil fields, carbon dioxide (C0 2 )
flooding may offer the largest potential for additional oil recovery. The
physical mechanisms by which CO 2 contacts and mobilizes crude oil are
reviewed.
Influence on the displacement process of factors such as the
phase behcLvior of CO 2 -crude oil mixtures, swelling of oil by dissolved
CO , and reduction of oil viscosity are considered.
Adverse effects of
2
the viscous instability which occurs when very low viscosity CO displaces
2
more viscous oil and water are discussed. Advantages and disadvantages of
three potential methods for controlling the mobility of CO are reviewed:
2
thickening CO 2 with polym,eric additives, reduction of CO mobility by hig,h
2
-- water saturations, and use of surfactants to generate foam-like emulsions
of water and CO 2
We also review field experience to date and document
the recent surge in field activity.
Finally, a brief assessment of the
future of CO 2 flooding rt!search and practice is offered.

Presented at the Annual !-leeting of the American Oil Chemists Society

Toronto, May 2-6, 1982

"

"/

Introduction

Enhanced oil recovery processes attempt to recover oil left. behind by

conventional primary and secondary recovery methods.

Illien an oil field is

first di scovered, oil will usually flow into a producing well under the
natural pressure of the fluids present in the porous reservoir rocks.
fluid is removed from the reservoir, however, the pressure declines.

As
Oil

flows more slowly into the wells and may not flow to the surface without
pumping.

Even if pumps are installed to bring the oil to

pr.imary production usually

recovers

originally in place in the reservoir.

only about

15-20%

the surface,
of

To supply additional

the

oil

energy to

drive oil out of the porous rock and into the producing well, fluid must
be injected into the reservoir.
of gas or water
<

to

restore

Secondary recovery methods use injection

the driving

force

for

inexpensive fluid

and

Because water Is a very

because waterflooding usually produces as much or

more oil than Is obtained

standard

production methods.

during

technique

primary production, waterflooding

for

recovering

capillary forces,

between oil and water,

isolated droplets.

oil

left

has

by primary

However, water is not an ideal fluid for forcing oil

from the very small pores of a reservoir rock.

immiscible,

the

Huch of the oil now being produced in the United States

comes from fields which are being waterflooded.

the

through

.'.

reservoir rock.

become

flow

act

Because water and oil are

which arise from

the interfacial

tension

to trap a significant portion of the oil as

As lOater displaces oil from the reservoir rock,

the

oil saturation, or fraction of the volume of the pore space occupied by

oil, decreases

until

it

reaches

limiting

value,

the

"residual

oil

 0f.'~

.}~.
":

t,~

,.
saturation".

At this point, all o,f the remaining oil is trapped and no

longer flo~1ing.

Of the more than 400 billion barrels of oil discovered in

the United States,

nearly 300 billion barrels will remain after primary

and secondary recovery methods have been applied.

Thus,

the targe t

for

enhanced 01.1 recovery processes is a large one, indeed.

Tertiary recovery methods are designed

to overcome,

in one way or

another, the. capillary forces which trap the oil during waterflooding.
some methods,

surfactants

are

injected

which

reduce

the

In

interfacial

tension between oil and water and thus allow the trapped oil droplets to
be mobilized.
injected.

In other methods, fluids which are miscible with oil are

Because

injected fluid,

there are

no

interfaces

between

capillary forces are not present,

the oil wh;lch can be contacted can be recovered.

the oil

and

the

and in principle,

all

Unfortunately,

fluids

such as liquid propane. which are truly miscible with oil, are now much
too expensive for most field applic:ations.

The volume of fluid required

is large, so that the fluid chosen must be inexpensive and available in

large quantities.

One fluid
natural

which meets those requirements is carbon dioxide.

supplies of nearly pure CO 2 have

been found

Large

in New Mexico and

Colorado, and several large pi pelines to carry that CO 2 to Permian Basin

oil fields are planned.

CO 2 may also be available from power plant stack

gases, ammonia plants and coal gasification plants. l

CO 2 is not strictly miscible with crude oil, however.

oil, it fones two phases at typical reservoir conditions.

If mixed with
Under the right

conditions, however, it can displace oil nearly as efficiently as a truly

"

,"

miscible

solvent.

efficiently lies
mixtures.

The key
in the

to

understanding

phase behavior and

This paper reviews

contacts and
encountered

the

fluid

can

displace

properties of

physical mechanisms

mobilizes crude oil, describes

in field

how CO

applications of CO 2

the

oil

CO -oil
2

by which

CO

problems which will be

flooding,

and

documents

the

recent surge in field activity.

Displacement Mechanisms

When CO 2 is injected into a waterflooded oil reservoir, it displaces

some of the wa ter and mixes wi th the oil left behind by the wa terflood.
Fig.

1 describes

the behavior of binary mixtures of

CO

and

crude oil

(containing dissolved natural gas) from the Wasson field, a large field in
west Texas for which a CO 2 flood is planned.

The behavior of a particular

mixture depends on the CO 2 concentration and the pressure.

the original oil

(0% CO 2 )

is a

liquid

at

For instance,

pressures above 900 psia but

splits into liquid and vapor below that pressure.

A mixture containing 40

mole % CO 2 forms a single liquid phase above 1350 psia and a liquid and a
vapor at lower pressures.
is more complex.

At high CO 2 concentrations, the phase behavior

At low pressures liquid and vapor phases form.

pressure is increased, the vapor phase, which contains CO

hydrocarbon gases,

condenses

into a

second

liquid

phase.

As the

and the light

There

narrow pressure range over which two liquids and a vapor coexist.
those

pressures,

liquid.

two

liquids

form,

CO 2 -rich

liquid

and

is a
Above

an oil-rich

The phase behavior shown in Fig. 1 is typical of that observed

for crude oils

from

reservoirs at temperatures below about 120 F.

1-4

At

"

.
higher temperatures, the two-'phase reg:lon has the same general shape, but
there

Is no

three-phclse

region.

Instead

low-density vapor at low pressures and a

the

dense,

CO 2 -rich

phase

supercritical

is

phase at

high pressures.

It is apparent from Fig. 1 that CO 2 is quite soluble in crude 011 at

typical reservoir pressures, but it ,is not miscible in all proportions at
any reasonable

pressure.

For instance,

fraction IDust be nearly 0.7 before a

at

2500

second,

psia,

CO -rich

the CO

phase

mole

appears.

Thus, when CO 2 is injected into rock and contacts trapped oil, at first it
simply dissolves in the oil droplets.
volume,

however.

Fig.

as

the volume of

volume of the oil alone.

2 illustrates how dissolved CO

Permian Basin crude oils.

defined

The dissolved CO

The

S'.. elling

the saturated

factor

does occupy some

swells

reported

several

in Fig.

CO 2 -oil mixture divided

2 is

by the

Figs. 1 cmd 2 indicate that as the amount of CO

present increases, it takes higher pressure to force all of the CO 2 into

solution, and the oil expands as additional CO 2 dissolves in it.
CO 2 is injected into a 'o'aterflooded reservoir,

swell,

some

flow.

The

droplets will coalesce,

total

amount of oil

however, is limited.
residual

oil

displacement.
by

CO~

and

\;hich

hence,
can

Thus, as

trapped oil droplets lJill

some oil will begin to

be mobilized

by swelling,

Only the oil which has a saturation greater than the

saturation

can

flol.'

and

be

recovered

by an immiscible

Thus, if swelling alone accounted for the recovery of oil

flooding, the incremental recovery of oil would be relatively low

Fortunately, t!Jere is an additional mechanism which, under the right

Circumstances, can lead to a more efficient displacement.

This mechanism

""

..

'

is similar to the "Vaporizing Gas Drive" in which high-pressure methane is

used

displace oil. 5

to

As more CO 2

is

injected ~

becomes saturated and a CO -rich phase appears.

2

the

oil

eventually

This phase is not pure

CO 2 , however, but is a mixture of CO 2 and hydrocarbons extracted from the

oil.

Fig. 3

illus trates

the composi tional

behavior of mixtures

of CO

with recombined crude 6il from the Maljamar field,4 in which a pilot CO
. flood is being performed.
are

represented

hydrocarbons
th~

in

of

three

pseudocomponents:

and heavy hydrocarbons

(C

and

13

mixtures actually contain many more components.

in Fig. 3, mixtures of CO
fall

into

the

In Fig. 3, the compositions of CO 2 -oil mixtures

terms

(C -C I2 ),
I

CO

light

heavier),

though

In the system shown

% CO 2

and oil containing more than about 60 mol

two-phase

region,

equilibrium with an oil-rich phase.

in which

CO -rich
2

phase

is

in

Tie lines, which connect compositions

of phases in equilibrium, are shown in the two-phase region.

Any mixture

which has

split

an overall composition lying on a

tie

line will

As Fig. 3

phases whose compositions lie at the ends of the tie lines.

shows,

the

CO 2 -rich

hydrocarbons

if

phases

the

can contain substantial

overall mixture

contains

mole

enough

into

fractions
of

the

of

light

hydrocarbons which are preferentially extracted by the CO 2

The effect of this extraction behavior on the displacernentprocess is
illustrated in Fig. 4.

Mixtures of reservoir oil A wi th CO 2 lie on the

straight line connecting the CO

added

to

reservoir

reservoir,
CO

oil A,

as

it would

at

be

during

CO 2

the overall c01:lposition moves along that

dissolving in and swelling

reached

apex with the oil composition.

the

composition L I

the oil) until

5

At

this

As CO 2 is

injection into

straight: line

(with

the two-phase envelope is

point,

CO 2 -rich

phase

of

'.

. .11

"

.'

.
"

composition Ul

As discussed

appears ..

below,

this

phase

is much

less

viscous than the oil so it moves more rapidly under the imposed pressure
gradient, and thus the fluids of compo.sition Ul

011.

Mixtures of upper phase U1 with

reservoir oil A lie on a

straight line connecting their compositions.

phase U ,
1

the overall composition changes

composition L2 is reached.
U
2

mix with fresh reservoir

As the oil is diluted with

along

that

line until

is then repeated,

with

the upper

becoming increasingly enriched with light hydrocarbons

(L).

contacts,

enough

hydroca rbons
straight

upper

that

line

phase

eventually

contains

it is mi sc i ble wi th reservoir oil A.

connecting

the upper

phase

is said

to

have

"developed"

mis(:ibility with

the

phases

With repeated
extracted

That

composition with

composition no longer passes through the two-phase region.

CO 2

the

This time, the CO 2 -rich phase has composition

The contacting process

the

second

is,

the

the oil

In this case,
oil

through a

"multiple contact" process.

If the oil had had a different composition, say that of reservoir oil
B in Fig. 4, the development of miscibility would not have occurred.

The

contacting process would have proceeded as described above until the upper
phase composition had reached U
1

Then further contacts of the U phase

1

with oil B would produce no composition change, since those mixtures would
always fallon the same tie line.

Thus, oil compositions which lie in the

region of tie line extensions do not contain enough light hydrocarbons (L)
to develop miscibility.

The dividing line between oil compositions which

develop miscibility and

those which do not

shown in Fig. 4,
the

plait

point

is the

"critical tie

line"

the tie line which is tangent to the phase envelope at

(P)

where

upper

and

identical.

lower

phase

compositions

are

".

Because the phase behavioi of CO 2 -crude oil mixtures depends on the

pressure (see Fig.
changed by

the location of

1),

the

changing the system pressure.

critical

tie

line

can be

An increase in press~re causes

the two-phase region to. shrink, which displaces the critical tie line to
the left, .as shown in Fig. 5.

Tie line slopes may also change to reflect

more effective extraction of hydrocarbons at higher pressures.

Thus, an

oil for which the multiple contact process did not develop miscibility at
a low pressure could develop miscibility at a higher pressure.

Fig. 6 shows the effect of displacement pressure on the recovery of

oil from a long (40 ft.), slim (0.25 in. ID) tube packed with small glass
beads 4

(see reference

for a

review of laboratory techniques used

evaluate reservoir oils for CO 2 flooding).

to

At 800 psia, the displacement

..

-~.

was relatively inefficient.

Only about 50i. of the oil was recovered.

At

1000 psi, the displacement recovered 90% of the oil and at 1200 and 1400
psi,

nearly 95i. of

the oil was displaced.

Clearly,

CO 2

displacement

efficiency in this ideal porous flow system can be very high.

Detailed numerical
consistent with

the

the displacement process 7 ,8

simulations of

qualitative

description given above

displacements such as those performed in laboratory slim tubes.

for

are

ideal

In field

scale CO 2 floods, there are additional complexities, the effects of which

are not completely understood.
water which surrounds

trapped

For instance, in tertiary CO 2 floods, the

oil droplets may restrict

injected CO 2 to the oil droplets.

less efficient,

the

access

of

If mixing of CO 2 with trapped oil were

then the benefits of

the multiple

process described would also be reduced.

contact

extraction

...

."

-. .

.~

:,

...

'.'.

An addi tional
field-scale

CO

particularly if

and

perhaps more

displacements
the

important

3.rises

from

complexi ty

affecting

nonuniformities

reservoirs are heterogeneous

(as most

of

flow,

are).

This

problem is seriously magnified by the low viscosity of CO 2 as compared to

oil.

When CO

2 dissolves in oil, the viscosity of the resulting mixture is

Fig. 7 shows a correIa tion given by

lower than that of the original oil.

Simon and Graue 9 for'the effect on viscosity of saturating oil with CO 2 at

l20F.

If the saturation pressure is high enough, substantial viscosity

reductions

Unfortunately,

can occur.

however,

the

viscosity

CO -oil mixture is still much higher than that of pure CO 2 ,

2

of

the

and much of

the difficulty remains.

Viscous Instability and Mobility Control

,The discussion above

behavior of

has

CO 2 -Oil mixtures,

reservoir scale

is also

focussed
but

on

the

microscopic

the performance of

influenced

CO 2

or local

flood

by the macroscopllc behavior of

displacement regions which occur as CO 2 pushes oil.

on a
the

In any flood in which

a less viscous fluid displaces a more viscous one, the displacement front
is unstable.

As

result

of

the

instability,

an initially sharp

displacement front becomes convoluted and develops "fingers".

In Fig. 8, modified from Habermann,IO the difference between stable

and unstable

displacements

two-dimensional reservoir.
equal

is illustrated

for a

symmetry element

of

In Fig. 8a, the injected fluid has a viscosity

to or greater than that of

the fluid

in place.

The displacement

 ~:

".

front advances uniformly through the porous medium and at time t , when
3
injected fluid

breaks

through

reservoir has been swept.

into

the

producing

well,

In Fig. 8b, the injected fluid has

significantly lower than that of the fluid in place.

displacement is no longer uniform.
irregular.

most

The nonuniformi ties

of

the

viscosity

In this case, the

At time t l , the displacement front is

(viscous fingers)

grow. wi th addi tional

injection, and a much smaller portion of the reservoir has been swept when
injected fluid breaks through at time t 2
a

Thus, in displacements in which

fluid of low viscosity, such as CO 2 , drives one of higher viscosity,

oil, the displacement front is inherently unstable.

Such adverse mobility

displacements are characterized by early breakthrough and poor areal sweep

efficiency.

At reservoir pressures and temperatures, CO 2 has a viscosity of a few

hundredths of a centipoise, while the viscosities of oils for which CO 2
floods
Thus,

are

planned

usually lie in the range

in CO 2 floods,

viscosity)

is

in

of one

the

range

ten

to

one

hundred,

mobility paths from injection to production wells,

the

ten centipoise.

the mobility ratio (ratio of oil viscosity to CO 2

instabilities will cause early breakthrough of CO 2 ,

reduces

to

recovery rate of oil.

Thus

so

that

viscous

By establishing high
v~scous

fingering also

the overall effiCiency of the

displacement process is much lower in a reservoir than in the ideal slim

tube displacement described above.

An analogous instability occurs if a

denser fluid overlies a less dense one in a porous medium.

Indeed the two

situations can be treated simultaneously, and even played off against each
other if they are of opposite signs.
viscous

liquid

by

lower

A downward displacement of a heavy,

density,

less

viscous

fluid

can

be

gravitationally stabilized if the displacement front velocity is less than

a critical value.

In fclct, CO

floods at Weeks Island and Bay St. Elaine,

both in Louisiana, have been designed to take advantage of the difference

i'(1 density

between

injected

fluid

and

oil

to

stabilize

downward

displacements.

In many reserv?irs,

however.

the flow from injection to production

well is nearly horizontal, density differences cannot be used to stabilize

the flow, and frontal
adverse.

It

instabiliti,~s

is apparent,

will occur if the mobility ratio is

therefore,

controlling viscous fingering,

often

methods. is an important area of CO

that investigation of methods

referred

to

as mobility

eliminating

the basic cause of the problem.

mobility of

the injected CO ,
2

behind the frontal

By

all,aim at

somehow decreasing

the

thus increasing the pressure gradient

region, the forces which give rise to

fingers can be lessened or eliminated.

af ter all, maki ng

control

flooding research.

The mobility control methods which have been proposed

and.

for

the growth of

This approach has its drawbacks -

the eli splacing fluid

less rna bile inevi tably decreases

the rate at which it can be injected, and thus the rate at which the oil
can be produced.
breakthrough

and

Potential methods

Nevertheless.
poor reservoir
for

the

alternative

sweep are

controlling

the

described below.

10

effects

usually less

mobility

of

of

early

attractive.

injected

CO

are

'"
"

WAG (Water'Alternated with Gas)

A method that first appeared in the literature in the late.fifties is

based on the simple

principle

that

the mobility

to

fluid

can be

'd~creased by decreasing the proportion or "saturation" of that fluid ,11

This was first
suffer from

suggested

the

for solvent (such as

propane)

same difficulty as that described above

floods,

which

for CO 2 "

The

rock's relative permeability to the solvent can be decreased greatly by a

relatively small
Fig. 9a.

increase in water saturation.

This is illustrated in

When water is injected along with the solvent or miscible gas,

the overall mobility of the injected fluid is decreased because the water
fills some of the flow channels previously available to solvent.
the growth of

displacement

inhomogeneities

due

to

Hence,

instability
.-..

is

decreased.

A schematic of

the injection process is shown in Fig.

This method

is currently being used

in a number of CO 2

there are still some questions regarding the method.

CO

floods,

9b.

though

Injected water and

may not be distributed uniformly in the reservoir, and increased water

saturations may reduce microscopic displacement efficiency.

there is evidence that with a

For instance,

higher mobile water saturation,

larger

fraction of the residual oil stays "trapped" or shielded from contact with
the CO 2 12

"Foam~lsions" - CO? Dispersed in Water

The petroleum literature contains a number of

papers .which report

large pressure drops required to force foam through porous media.

Host of

these reports describe air foams composed of bubbles or cells of gas in a

11

....

.....

..
"

"

~t

water-continuous dispersion.

With the '"quality", or proportion of gas, at

E~ists

707. or greater, the aqueous component

films surrounding
p~incipal

the gas cells.

There

mostly as a network of thin

is general

agreement" that a

cause of the low mobi4ity (as evidenced by the higher pressure

gradients required) is the thermodynamie work required to distort and move

large numb,ers of water films through the necks of pores in the rock (Fig.
Those
"foamants".

films are stabilized

by the presence of

surfactants or

In the absence of the surfactant, the films disappear and the

mode of flow changes back into the usual situation when immiscible fluids
are flowing

simultaneously

in

well-defined channels accommodate

The effort

to

utilize

the

rock,

E~ach

and

separate,

fairly

flowing fraction.

clX~posed

"foams"

where

of CO

partly on the effort to find suitable surfactants.

and water is

based

Thus far, the approach

has followed the classic ideas of '"foam stability" ,14 and it appears that
a

number of

surfactants will

CO -oil-w31:er systems.
Z

produce foams which may function well

in

For ins tance. nonionic surfactants wi th sui tably

balanced T07a ter-hydrocarbon

solubilities

seem most

promising .15

In

addition, there is a need to study the differences between dispersions of

liquid or supercritical COZ in water, and the true foams.
ask how much of the reported
compressibility

of

air

experimental data resulted

foams,

and

how much

resulted

One can also

from

the large

from

the

same

phenomena as are seen in the flow of liquid-liquid dispersions.

Direct Thickening

Because at reservoir pressure CO 2 is a dense fluid -

12

its density is

'.
near that of oil - one can regard it as a potential solvent.

In fact, the

effectiveness of such dense CO 2 in displacing oil depends on this aspect

of the fluid.

It turns out that other substances besides hydrocarbons can

be dissolved in dense CO 2 , and that is the origin of the mobility control

method that we label as "Direct Thickening".

--The solutes of most

interest,

of course,

are

polymers.

Other

polymers have been used extensively to thicken wa ter. but these polymers
are not soluble in CO 2 ,

The ideal addi tive for CO 2 floods would be an

inexpensive polymer which, when dissolved at fairly low concentration in

liquid CO 2 , would raise the solution vise osi ty from .075 cp or so to, say,

1. 5 cp.

Such a moderate increase by a factor of about twenty would be

sufficient.

A great improvement in "areal sweep pattern"

would result,
.-.,.~

the amount of CO 2 required would decrease, and a very significant gain in

overall displacement efficiency would be observed.

Polymers which meet those requirements have not yet been found, but
high-pressure experiments have

shown that a

fairly large assortment of

hydrocarbon-soluble polymers are also soluble in CO 2 ,

raise the viscosity by only ten to twenty percent.
to

find

one

or more effective

currently active manufacturers,

polymers

among

but synthesi s

of

Those tested so far

It would be convenient
the

product

lists

new polymers

may

of
be

necessary.

Results of solubility measurements for some polymers in CO 2 are shown

in Table

1.

Those

Ii sted

are atactic,

relatively low molecular weight.

tested

have

been

insoluble.

straight-chain polymers

of

Higher molecular weight polymers we have

Table 2

13

reports

results

of

viscosity

....
","

measurements

for

some of

Because of

the systems.

the

10\01 molecular

weights of these polymers, and probably for other reasons,

increases .are much lower than those which can be obtained

the viscosity
for" aqueous

solutions containing similar concentrations of water-soluble polymers.

Chemicals for Mobility Control

The t.IO types of chemicals discussed as mobility control agents for

CO

The chemicals

floods must satisfy several additional requirements.

must survive in the environment in which they will be used, and must not
engage in destructive or inconvenient
present.

reactions

wi th other materials

They must not adsorb too readily on rock surfaces, and they must

tolerate pressures from I:me to five thousand psi (depending on the depth
of the reservoir) and temperatures from 20 to 120C (also depending on the
depth).

In addi tion, the reservoir contains crude oil with a variety of

organic

surfactants which commonly render rock oil-wet and destabilize

emulsions.

Also

concentrations

of

present

on the effectiveness
for

future

which

sometimes

calcium and magnesium as

carbonate and chloride ions.

subject

is water,

of

well

as

sodium,

high

and

of

The effect of these on the solubility, and

the mobility control methods

study.

contains

Finally,

mentioned,

is

any additive must be available

a
in

sufficient quantities at a price 10\01 enough to al10\ol economic use.

While

the

requirements

listed

ma:,

be

difficult

potential value of effective mobility control in CO

to

satisfy,

the

floods is very large.

It can be measured as a reduction in risk, an increase in profitability of

EaR projects, and an increase in petroleum supply in the years ahead.

14

.-.
"

.'

':

Field Activity in CO? Flooding

One advantage of CO 2 flooding is the broad range of oilfiel~'types to

w~ich

it is applicable.

The process

itself

is not

sensitive

to

the

presence of divalent cations (calcium and magnesium) which can adversely

affect the performance of most surfactant formulations, nor is it affected
by the salinity of the reservoir brine.

Thus, CO 2 flooding can be applied

to both sandstone and carbonate reservoirs.

The reservoir must be deep

enough that it can withstand without fracturing the high pressure required
for efficient

extract ion of

hydrocarbons

by CO 2 ,

and

it

should

permeable enough that CO 2 can be injected at reasonable rates.

be

Because

the viscosity of CO 2 is very low and it is therefore easy to inject, all

but the lowest-permeability fields meet the
In fact,

CO 2 injection has been used

injection rate requirement.

in at least one field

(Crosset t,

Crane County, Texas) because it' could be injected where water could not. 13
In such low-permeability fields, mobility control may be in conflict with
the goal of high injection rate.

The most stringent restrictions on CO 2

flooding come, again, from its low viscosity.

Because the tendency for

viscous fingering increases with the ratio of the viscosity of the oil to
that of CO ,
2

the process has been applied mainly in reservoirs in which

Furthermore, the adverse affects of

the oil viscosity is 3 cp or less.

viscous fingering are aggravated by large permeability contrasts because

CO

tends to break

through

quickly in high-permeabili ty

streaks

or

fractures establishing preferential flow paths through which most of the

CO

then flows.

Thus, CO 2 flooding will, in general, be more effective in

reservoirs which do not have high-permeability streaks or fractures which

communicate between wells.

IS

..

-.

One clf

the

application of

principal

CO 2 floodi ng

quantities of CO
>.

a~monia

cons-traints

is

the

which

affects

the

a.vallabili ty of

field -scale
The

CO 2 ,

large

needed will come primarily from natural sourc~s, though

plants and coal gasification plants may provide CO 2 if oil fields

are located nearby.l

in northeast

Large natural supplies of CO 2 have been discovered

New Mexico

and

in

Fig.

southern Colorado.

li..

shows

approximate locations of the largest known CO -producing areas as well as

areas in which are located oil fields .There CO 2 floods are active now or
are being planned.
la rge carbona te

Of the areas shown,. the Permian Basin, with its very

reservoirs,

hotter sandstone reservoirs,

oil recovery with CO 2 "

and

the Gulf

Coas t

area, wi th

its de epe r ,

appear to offer the greatest potential

for

For most applications, the costs of delivering CO 2

by any means other than a pipeline are too high except for pilot -tests in
which CO 2 is usually delivered by truck.
two large pipelines,

one from

Plans have been announced for

the Four Corners area and

southern Colorado and northeast Ne\/ Mexico, to carry CO

the other from

to the oilfields

of the Permian Basin.

Cost estimates for those pipelines and associated

facilities

range of 2--3

are in the

bi.llion dollars.

Thus,

both the

investments and the associated risks required for the development of large
CO 2 projects

are

high,

projects

being

but

apparently

not

too

high

to

prevent

implementation.

The

planned

extensive field testing of the CO

now,

however,

flood process.

have

the

benef i t

of

Field-scale, secondary

CO 2 floods have been operating in the Kelly Snyder and

Crossett

fields

since the early 1970's, \/ith good performance,16 and pilot tests in fields
of widely varying characteristics have established that CO

16

can displace

oil left behind by waterflooding. 17 ,18

While current oil production from

existing CO 2 flood projects is only about 22,000 B/D,16 that number should
grow substantially as the projects being implemented now begin. to produce
additional oil.
flood

Fig. 12 compares the growth in the number of active CO

projects over the last decade with that of other gas

projects. 18

injection

The number of hydrocarbon miscible projects declined with the

increasing

price of

projects increased.

the hydrocarbon solvents,

Indeed,

the number of

but

the

active CO

number

of

projects

CO 2

nearly

doubled in the period from 1980 to 1982, and those numbers do not include
the large west Texas projects dependent

on the CO 2 pipelines.

Of 416

projects recently certified in twelve categories of enhanced oil recovery

under the

u.s.

CO 2 floods. 19

Department of Energy's Tertiary Incentive Program, 91 were

Thus, it appears very likely that the number of active CO
."~~

projects will continue to grow as will oil production from CO

is equally clear
flooding

will

that

continue

successful
to

large-scale

provide

variety

floods.

It

implementation of

CO 2

of

the

challenges

to

researchers and engineers who will guide that development.

Acknowledgement

The work reported in this paper was supported by funds from the U.S.
Department of Energy,. the New Mexico Energy Research

and

Development

Institute, and the New Mexico Petroleum Recovery Research Center.

17

References

1.

Stalkup, F. 1., "Carbon Dioxide Miscible Flooding: Past; Present,

and Outlook for the Future," ;~ Pe!..:. Tech. lQ. (Aug. 1978) 1102-1112.

2.

Orr, F. M., Jr., Yu, A. D., and Lien, C. L., "Phase Behavior of CO 2
and Crude Oil in Low Temperature Reservoirs," Soc. Pet. Eng. ~ ~
(Aug. 1981) 480-492.

3.

Metcalfe. R. S. and Yarborough, L., "The Effect of Phase Equilibria

on the CO 2 Displacement Mechanism,'o Soc. Pet. Eng. ~ ~ (Aug. 1979)
242-252.

4.

Orr. F. M. , Jr. and Taber, J. J., "Di splacement of Oil by Carbon

Dioxide," Annual R~port to the U.S. Department of Energy, Report No.
DOE/BC/10331-4, Nov. 1981.

5."

Hutchinson, C. A., Jr. and Braun, P. H., "Phase Relations of Miscible

Displacement in Oil Recovery," AIChE ~ 7..... No.1 (1961) 64.

6.

Orr, F. M., Jr., Silva, M. K., Lien, C. L., and Pelletier, M. T.,
"Laboratory Experiments to Evaluate Field Prospects for CO
2
Flooding," :!...:.. Pet. Tech. ~ (Apr. 1982).

7.

Orr, F. M., Jr., Silva, M. R., and Lien, C. L., "Equilibrium Phase
Compos i tions of CO 2 -Crude Oil Mixtures:
Comparison of COIl ti nuo u s
l1ultiple Contact and Slim Tube Displacement Tests," SPE 10725,
presented at Third Joint SPE/DOE Symposium on Enhanced Oil Recovery,
Tulsa, Apr. 4-7, 1982.

8.

Gardner, J. W., Orr, F. M., Jr., and Patel, P. D., "The Effect of
Phase Behavior on CO 2 -Flood Displacement Efficiency," ~ Pet. Tech.
~ (Nov. 1981) 2067-2081.

9.

Simon, R. and Graue, D. J., -Generalized Correlations for Predicting

Solubility, Swelling and Viscosity Behavior of CO 2 -Crude Oil
Systems," :!...:.. Pet. Tech .!2. (1965) 102-106.

10.

Habermann, B., "The Efficiency of 'Miscible Displacement as a Function

of Mobility Ratio," Trans., AIME~!2. (1960) 264-272.

11.

Caudle, B. H. and Dyes, A. B., "Improving Miscible Displacement by

Gas-Water Injection," Trans., AIHE 213 (1958) 281.

12.

Raimondi, P. and Torcaso, M. A., "Distribution of the Oil Phase

Obtained Upon Imbibition of Water," Soc. Pet. Eng. J. (Mar. 1964)
49-55.

13.

Bernard, G. G., Holm, L. W., and Harvey, C. P., "Use of Surfactant to

Reduce CO 2 Hobility in Oil Displacement," J. Pet. Tech. 32 (Aug.
1980) 281-292.

18

.":.

14.

Bikerman, J. J., Foams, Springer-Verlag, New York (1973).

15.

Heller, J. P. and Taber, J. J., "Development of Mobility Control

Methods to Improve Oil Recovery by CO 2 ," Annual Report to the U.S.
Department of Energy for the period Oct. 1, 1980 - Sept. 30, 1981,
Contract No. DE-AC21-79MCI0689.

16.

Pontius, S. B. and Tham, M. J., "North Cross (Devonian) Unit CO 2

Flood--Review of Flood Performance and Numerical Simulation," ~~
Tech. 30 (Dec. 1978) 1706-1714.

17.

Goodrich, J. H~, "Target Reservoirs for CO 2 Miscible Flooding," Final

Report to the U.S. Department of Energy, Report No. DOE/MC/08341-17,
Oct. 1980.

lB.

OGJ

19.

"Enhanced Oil Recovery and Improved Drilling Technology," U.S.

Department of Energy Progress Review, DOE/BETC-Bl/2, Mar. 31, 1981.

Annual Production Report, Oil

139-159.

19

Gas

80, No. 14 (Apr. 5, 1982)

I:

..

';

"'

Tnble 1
Solubilities of Polymers in Dense CO2

Temp.

Pressure

CO2
Density.

Polybutene

25"C

l700-psi

0.82 g/c

Polybutene

25

1700

.B2

7. B

Hn 640

Polybutene

25

1700

.B2

1.0

Hn 2500

Polybutene

25

1700

.82

0.0

Hn 2700

Polypropylene

25.5

2260

.875

2.17

Atactic

Polypropylene

1960

nn

.JL..V

.ou

1.19

'Atactic

Terpeue resin

25.5

1940

.846

4.1

Polyvinylethyl ether

26

1930

.B4

4.0

Poly undecanomide
(Nylon 11)

25

1500-2400

.8-.88

0.0

Poly 4vinyl biphenyl

26

2100

0.0

Poly vinyl stearate

25

2000

.87
.852

Poly_dimethyl Siloxane

26

2050

.855

4.91

Polymer

Co-solvent

'l'l

/'

Solubility

Comments

23.8 gIl

Hn 440

L 75

j:

..... '

'

..

>

Table 2
Fall Times and Viscosity of Polymer Solution in dense C02

Fl ui d

(p"re

co,)

20 g/l in CO2

( Polyb""ne
)
l-l1.> 640
8.5 gIl in CO2

e'Yb",.ne)
'11 .... 2500
1.5 g/l

PoI)~eLhyl

Oxirane
9.57 gIl in CO2

TeI:lE'

Fall Time

Viscosit~

1700
1900
2100

25"C

1. t. 60 sec
1. 488
1.514

.076 cp
.081
.087

1.540
1. 568
1.608

.080
.085
.092

p'J

1700
1900
2100

}
}

1700
1900'
2100

1700
1900
2]00

( POIYb"ne)
'11\..' 440

Pressure

1900-

25 C

"

25 C

1.565
1.620
1.660

.081
.088
.095

25 D C

1.560
1.575
1.615

.081
.086
.093

26 D c

1. 760

.096

_0,

-.

..

"...

6000~--~---.----.---,,---.----.----n---.----r----'

!:iooO

TWO LIQUID
PHASES

LIQUID PHASE

(C02 DISSOLVED IN OIL)

<!

(/)

(C02 AND OIL

IN EACH PHASE)

3000

0..

a:::
::::>
(/)
(/)

a:::

2000

TWO LIQUIDS

0..

1000

LIQUID

+ VAPOR

20

(0IL+C02)
(C02

40
MOLE

Fig. 1

LIGHT HYDROCARBONS)

60
PERCENT

80

C02

Phase behavior at 90F of binary mixtures of C02 with a

recombined Wasson oil containing 312 SCF/BBL.

100

. :' .

1.5

1.4
lr

0
I-

<!

1.3

EJ

LOCO HILLS, 82F

MALJAMAR,

9ar

WASSON, 90'F

WASSON, 105F

8-

WASSON, 120"F

)(.

1L
C9

-.J
-.J

1.2

I:::,

3:
en

.lr

1.[

1.0

60

80

MOLE PERCENT C02

Fig. 2

Swelling of oil',by C02 (volume of oil saturated

with' C02 divided by the volume of the oil).

100

.. .,

: .. ~

0,
,.'

,.

C02 -RICH

PHASE
COMPOSITION

OIL - RICH

PHASE
COMPOSITION

ONE

PHASE REGION

OIL COMPOSITION

Fig. 3

Pseudo ternary phase diagram at 2500 psia and 90F

for mixtures of C02 with a recombined Maljamar oil
containing 650 SCF/BBL solution gas.

:.'.

..,

I
I
I
I
I

f~ESERVOI R , '
OIL B

L2

RESERVOI R
OIL A

--------~~--~~------------------~L

Fig. 4

Development of miscibility in C02-crude oil systems.

'~; .. ,.

t_

PJI

I
I

I
I
I

I
I

/------------~/~/-----------~----------------------------~

Fig. 5

Effect of pressure (P 2 > PI) on critical tie line position.

1.0 i

. 0.8/-

w
0::
W

>

1400 psi;

E)

1200 pSI

1000 pSI

800 pSI

~~

",

0.6

u
W
0::

-'
0

0.4

lJ...

0
Z
0
f-

0.2

<l

0::
lJ...

OD~~~----~--------~------~------~--------~------~------~--------~~
0.8
0.4
0.6
1.0
0.0
0.2
1.2
1.4
1.6

PORE VOLUMES
Fig. 6

C02

INJECTED

Effect of pressure on oil recovery by C02 flooding in an ideal porous flow

system (slim tube) for C02 displacing Maljamar oil at 90F.

.,,::

..

'

..