M: Fluid Propagation, Interaction, and Mobility

I SCOTT P, TRUSHENSKI I I
DWIGHT L. DAUBEN I AMOCO PRODUCTION CO.

I
DAVID R. PARRISH TULSA, OKLA.
MEMBERS SPE-AlME

ABSTRACT reported that systems that do not use cosurfactants

require sulfonates with a broad EW range. The
Laboratory core tests show that small polymer- Iow-EW sulfonates provide water solubility for the
driven micellar slugs di.~place tertiary oil high-EW material. Since the systems discussed here
efficiently. Sur/actant adsorption stua’ies reveal use a cosurfactants to perform this function, the EW
nonclassical behavior, Polymer requirements are range of the sulfonates we used is much narrower.
decreased by permeability-reducing micellarlclay An additional benefit of the cosurfactants is that
interaction and by reduced losses behind a micellar sulfonate adsorption by the rock surface is reduced,
slug. The required voiume of polymer s[ug increases
when the pore volume that is inaccessible to the SCREENING PROCEDURE
polymer increases. Long-core tests with mu Itiple Before core testing, crude oils that are potential
pressure taps rev~al the existence of a high-mobility candidates for micellar flooding are screened
oil-water bank and a Iow-mohiIity oil-m icellar mixing qualitatively against a number of micellar
zone, compositions of varying surfactant /cosurfactants
ratio, monovalent ion concentration, and water
INTRODUCTION content, Divalent ion tolerance and temperature are
also examined. Pre!~minary qualitative screening
ikiicellar fluids that use petroleum sulfonate
tests are used to visually examine the degree of
surfactants have been tested by several organiza-
miscibility between the crude oil and the micellar
tions as potential candidates for secondary and
solution. If a micellar fluid shows potential for oil
tertiary oil recovery operations, 1-7 Typically, a
displacement and is economically attractive, a core
micellar flood consists of a brine preflush to
testing program is initiated.
condition the formation, a bank of micellar fluid (5
Typically, tertiary floods are conducted at
to 4G percent PV) that displaces the oil, a mobility
reservoir conditions in fresh 2-in. -diameter, 4-ft-long
buffer (polymer) bank to drive the micellar slug,
Berea sandstone cores mounted in Hassler holders.
and a chase-water baak. Micellar flooding is
A small volume of micellar fluid (from 2.5 to 10
attractive as an improved oii recovery process
percent PV) is injected at a linear advance rate of
because it is not severely affected by gravity
about 2 ft/D and is followed by a bank of low-
segregation and is not limited by ultimate
salinity, polymer-thickened water. If the tertiary
surfactant availability,
recovery is encouraging, the micellar fluid is
evaluated in detail. This evaluation involves
DEVELOPMENT OF MICELLAR FLUIDS
extensive adsorption studies to optimize tlie fluid
The micellar fluids that have been developed by composition and long-core tests to examine the
our laboratory for miscible waterfJooding are propagation and interaction of the fluid banks.
microemulsions of high water content (85 to 95
MICELLAR FLUID DEVELOPMENT FOR
percent by weight). 8-11 These fluids generally are A PARTICULAR RESERVOIR
prepared with 4 to 10 weight percent oil-soluble
The Second WalJ Creek reservoir of the Salt
hydrocarbon sulfonate (with an equivalent weight,
EW, from 350 to 475) and an oil- or water-soluble Creek field north of Casper, Wyo., has been selected
as one of the potent~al candidates for micellar
alcohol cosurfactants. The cosurfactants performs
flooding. The reservoir has a temperature of 110°F,
several functions. In many cases it aids the water
a cr~de oil viscosity of 4.0 cp, and an average
volubility of the sulfonate. Gale and Sandviks
permeability of 50 md Analyses of the Second Wall
Creek formation water and the Madison Field wa:er
Original manuscript received in Society of petroleum Engineers
office July 30, P973. Revised manuscript received July 19, 1974. (to be used as the miccllar and polymer bank
I Paper (SPE 4582) was first presented at the SPE-AIME 48th makeup water) are shown in Table 1. Both waters
Annual Fall Meeting, held in t.as Ve8as. ~ev.. SePt. 30-OCt. 3.
1973, @ Copyright 1974 American Institute of Miningt Metallur- produce stable one-phase micellar fluids since the
gical, and Petroleum Engineers, Inc.
!Refere”ce S listed at end of paper.
total dissolved solids arrd divalent ion contents
This paper will be printed in Tran Sa~tiOfls volume 257, which
are low.
will cover 1974. In the preliminary screening and short-core tests,

I) F:{: E.MRER. 197.1 633

 TABLE 1 —WATERCCIMPOSITIONS
petroleum sulfonate:cosurfactant in brine. Approxi-
Second Wall Creek Madison Water
mately 2 PV of each fluid was recirculated for 24
Ion
— (mm) _ (mm) hours through a fresh, consolidated Berea sandstone
No+ 4,463 506 core (2 in. in diameter x 12 in. long) wrapped in
co+ + 24 288 epoxy fiberglass. The adsorption losses of the
Mg++ 83 51 surfactant and cosurfactants were determined by
cl- 4,430 638 material balance,. It was found that the cosurfactants
HC03- 3,980 134 is not adsorbed to a measurable degree on Berea
so4-- 75 980
sandstone. Because of the low adsorption, the
co3-- 312 0
cosurfactants can be used as a tracer for the micellar
Total 13,367 2,597
slug in both laboratory and field tests.
Sulfonate adsorption isotherms are shown in Fig,
several sulfonates and cosurfactants were examined 2. Curve A shows the surfactant adsorption isotherm
at the reservoir conditions. One fluid that performed determined when a brine preflush is used before
satisfactorily on the basis of cost vs oil recovery surfactant recirculation. A brine preflush was not
will be discussed in detail here. It is 5 percent by used for the systems characterized by Curve B.
weight petroleum sulfonate (average EW, 444; EW The adsorption behavior shown in Fig. 2 differs
range, 435 to 465; 60 percent active sulfonate; 40 per- considerably from the classical observation of
cent mineral oil), 3 percent cosurfactants (isopropyl increasing adsorption with increasing concentration.
alcohol), and 92 percent brine. The optimum salinity Maximum adsorption occurs in the range of 1,000 to
varied from 0.2N to 0.4N NaCl, depending on the 3,000 ppm, and a minimum is observed at about
divalent ion concentration. This micellar fluid itself 20,000 to 30,000 ppm. The adsorption maximum
is stable in waters with up to 700 ppm divalent ions. occurs near the critical micelle concentration (CMC
Fig. I silows the tertiary oil recovery performance concentration determined by classical me-hod), At
for a 10-percent PV slug of the above micellar the adsorption minimum the physical appearance of
solution in a 4-ft-long Berea core test. A 10-percent the micellar fluid changes from a turbid, caramel-
PV slug of this micellar fluid typically recovers 25 colored fluid to an amber, transparent solution, tilus
percent of the tertiary oiJ and leaves an unrecovered indicating a change in micelle structure. These
oil saturation of 5.5 percent 32 percent PV. changes in fluid characteristics may be associated
with the “retrograde” isotherm. ‘lhersl 2-14 have
MICELLAR Adsorption LOSSES teported an adsorption maximum for si~gle-component
detergents and amphipathic materials. However, no
In large systems (field scale) surfactant adsorption
satisfactory explanation has been given for this
by the rock sarface is an important f[ ctor that
phenomenon.
limits micellar fluid propagation and therefore oil
It is not coincidental that the micella~ composition
recovery. The optimum micellar slug size depends
that gave maximum oil recovery in tertiary core
on fluid dispersion, rock heterogeneity, adsorption,
tests (Fig. 1) has minimum adsorption loss (30,000
and, ultimately, economics. However, the minimum
ppm active sulfonate). Total sulfonate requirements,
volume of surfactant required for complete tertiary
thus chemical costs, are reduced by using a micellar
recovery must satisfy the rock adsorption at the
fluid whose composition corresponds to the
injected concentration. Therefore, adsorption
adsorption minimum.
studies were made for the above micellar system
Because of the retrograde adsorption isotherm,
with varying concentrations of a 5:3 ratio of the

lw t I , 1 , , , , I 1 [11111 I I ! 1“”1 I I I 1“”1 1 I I I

1 I I
~ POLYMER 1
INJECTED OIL DISPLACEMENT \ 4
o- -...

k.
*BD
k
r
MICELLAR
INJECTED
/
IN;ECTION SEOUENCE
B, NO PREFLUSH
7\ \\

\
1. D.3E5 PV-O.23N NaCl IN
FLUID CWIAPOSITIONS+3, :
SCfTENEO MADISCW WATER (SMWI
- SIJRFACTANT: COSURFACTANTS ~
,, \,, “~ j
2. 0.100 PV-5392, IN 0,2N NaCl BRINE
.ad~
SURFACTANT: COSURFACTANT:BRIN
0.23 N NsCI IN SMW #@ m-e
i ‘b.~.;.j ~
3. 1.583 PV-754 PPM -1 /t; 0.2N N~l
\.
I
,- $
d BRINE PREFLUSH

o~ J
0.5 1.0 1.5 2.0 B2UILI BRI UM SULFONATE CONCENTRATE tYl
PORE VOLUMES PROOUCEO AFTER WATERFLOOOING (PARTS PER MILLION ON ACTIVE BASIS}

FIG. 1 — TERTIARY OIL DISPLACEMENT WITH A 10 FIG. 2 — ADSORPTION ISOTHERMS OF A MAHOGAN-f

PERCENT PV MICELLAR SLUG IN BEREA SAND- PETR-OLEUM SULFONATE ON BEREA SANDSTONE
STONE. AT 110”F.

634 SOCIETY OF PET ROLEIIM ENGINEERS JOtJRNAL

 produced sulfonate concentrations may be greater cores (PV about 150 cc) were first driven to connate
than injected concentrations. In a micellar flood, water, waterflooded, then flooded with a small bank
dispersion at the leading edge of the slug causes of the micellar fluid (11 to 24 percent PV). This
surfactant dilution. Because of the resulting bank was followed by polymer water and chase
concentration gradient, the magnitude of the water. Adsorption losses were determined from a
adsorption loss varies in the mixing zone. At some sulfonate material balance. The losses were from
concentrations (from 1,000 to 20,000 ppm) the 0.56 to 1.3 lb/bbl PV, averaging 0.88 lb/bbl PV.
adsorption from the diluted micellar fluid may be
greater than that of the injected material. Therefore, MOBILITY DESIGN
once the undiluted micellar fluid moves into a
Mobility control is required in the micellar flood
region, a state of “supersaturation” will exist,
process to achieve good oil displacement perform-
since excess sulfonate has been adsorbed by the
ance and to prevent the dissipation of an expensive
rock surface. To reestablish equilibrium, sulfonate
micellar fluid from front and rear mixing. Fig. 3
ciesorbs and increases the concentration in the
shows the various banks that build up ahead of the
micellar fluid to a value greater than injected.
micellar slug under tertiary recovery conditions.
For practical purposes, sulfonate does not
The zone of maximum flow resistance is the
desorbinto dilute surfactant solutions. ~orexample,
oil-water bank that precedes the micellar slug. The
the continuous injection of the 92-percent-water
nobilities of the micellar solution and the mobility
micellar fluid (with 30,000 ppm active sulfonate)
buffer bank are adjusted to be equal to or less than
into a Berea core until equilibrium is achieved
the oil-water mobility. This basis for design has
would leave 0.6 lb/bbl PV of sulfonate adsorbed on
been discussed by Gogarty. lb
the rock surface (Fig, 2). If this fluid is followed
The total relative mobility of fluids in the oil-water
by brine, the maximum sulfonate concentration in
bank may be determined wing relative permeability
the brine due to desorption would be 60 ppm (since
data in the following equation:
this concentration also has an equilibrium adsorp-
tion value of 0.6 lb/bbl PV), At this concentration,
Totzl relative mobility = kro/po + k,w?pu,
30 PV of brine would be required to desorb all the
sulfonate. Because of the slow sulfonate desorption = l/equivalent viscosity .
the surfactant will not separate chromatographically
from the nonadsorbing cosurfrictant as the micellar Fig. 4 shows the relationship between the
bank moves through a porous system. relative mobility and fractional oil cut for Second
Both Isotherms A and B of Fig, 2 show retrograde Wall Creek and Berea cores. These curves were
adsorption near the same concentrations. However, generated from the relative permeability data shown
in the absence of a brine preflush, strlfonate in Fig. 5. The relative mobility selected for design
adsorption is greater at all concentrations. Short-core is the bottom flat portion of the curve corresponding
tests show that higher oil recoveries are achieved to the fractional cuts normally observed in
when a brine preflush precedes micellar slug displacement tests.
injection. 15 The brine preflush serves two purposes.
Laboratory tests with miceIIar fluids in native-
First, it eliminates large salinity gradients, thus
state Second Wall Creek cores showed a surprising
maintaining fluid stability and solubilization
and beneficial effect. After a large slug of the
properties. Second, it reduces sulfonate loss. micellar solution displaces all of the oil, the
In a Berea system a 6-percent PV slug of 92- restored to the absolute
permeability is not
percent-water fluid is the minimum volume that theoretical value. For example, Fig. 5 shows that
could recover all the tertiary oil since this volume the permeability to the micellar fluid and drive
ideally could satisfy the equilibrium adsorption of water following the micellar solution in Second Wall
the rock at 30,000 ppm active sulfonate (0.6 lb/bbl Creek cores is only 35 percent of the initial
PV from Fig. 2). Irr small systems where the permeability to oil (at connate water). Micellar/clay
dispersive effects are significant, an increased interaction appears to account for the reduced
slug size would be required since sulfonate dilution
would result in higher losses.
Limited adsorption studies have been conducted
with native-state Second Wall Cre<k cores. Small

DRIVE MOBILITY BUFFER MI CILLAR 01 1-WATER WATER

WATER [POLYMER) SLUG SLUG BAN” [AT Sor )

o a
~ z
x z 0 G.2 0.4 0.6 0.8 1.0
ip ~ ~ >m 5 )0-w
z FRACTIONAL FLOW IX WATER, fw

FIG. 3 — MOBILITY CONTROL OF THE MICELLAR FIG. 4—RELATIVE MOBILITY OF FLUIDS IN SECOND
FLOOD PROCESS. WALL CREEK AND BEREA CORES.

DEc EM BE Ri197’t 635

I
permeability. The practical benefit of this tailored for an individual application by (1) adjusting
interaction is a reduction in the amount of the type and amount of salt, (2) changing the water
viscosity-building material required to achieve content, or (3) adding polymer.17 One must be
mobility control. By incorporating the permeability careful not to make a change for mobility control
reduction in the mobility-control design of the that adversely affects the ability of the micellar
example cited, nearly a threefold reduction in the solution to displace the oil. Fig. 6 shows that
micellar fluid and polymer viscosity is possible. small changes in the divaient ion content can
The result is a considerable saving in polymer costs. produce major changes in the viscosity of the
The extent to which the micellar solution reduces micella: fluid.
permeability is related to both the type and the
DETERMINING POLYMER CONCENTRATION
amr nt of clays in the rock. Electron microscope
stul,.es indicate that clays, particularly montmoril - Biopolymer and the polyacrylamides have been
lonite, are dispersed by the micellar fluid and are used as viscosity builders in the mobility buffer
deposited in pore throats. Little cr no permeability bank. Biopolymer reduce mobility primarily by
reduction occurs in cores containing only minor increasing viscosity, whereas polyacrylamides
amounts of clays or in cores that have been reduce mobility by increasing viscosity and
desensitized by firing. reducing permeability. In the studies reported here
the biopolymer is Kelzan MF; Polyacrylamide A is
ADJUSTING MICELLAR SOLUTION MOBILITY Dow Pusher 500; and Polyactylamide B is Dow
The viscosity of the micellar solution can be Pusher 700.
Fig. 7 shows the viscosity developed in a core
for various concentrations of a biopolymer. These
curves were determined by combining data from
viscosity measurements and core tests. The
viscosity of biopolymer solutions decreases with
increasing flow rate. Good injectivity is expected
since the polymer viscosity is low at the high rates
prevailing around the wellbore. The higher
viscosities needed for mobility control are achieved
away from the wellbore as the fluid velocities
decrease. A flow rate must be selected before the
polymer concentration can be specified. Radial flow
calculations using the expected injection rates can
be made to help select an optimum flow rate for
design. A low rate typical of interwell flow (e.g.,
0.5 to 2.0 ft/D) is usually selected as a basis for
design, since rates rapidly decrease with distance
from the wellbore. Some penetration may occur
because of the unfavorable mobility ratio at the
higher velocities arcund the wellbore. The initial
polymer can be injected at a reduced rate if needed.

~~
WATER SATURATION, PERCENT PORE SPACE
22 - POLYMER CONCENTRATION. PPM 110° F
FIG. 5 — WATER/OIL RELATIVE pERMEABILITIEs. 1200
al -
10 I I I I I [ [ I I I I 18 -
~ 9 110° F 16 ..
8
E 14 -
7 ~“
; 12 -
~
6
n : 10 -
v
5
8 -
4 6 -

3 4 -
2 2 -
I 1

11 I I I I I I I I I I I
01 ,,I 1
o 40 80 120 la 2al 240 0.1 1 10
DIVALENT ION CONCENTRATION. PPM LINEAR FLOW VELOCITY, FTIOAY

FIG. 6 — VISCOSITY OF MICELLAR FLUID, EFFECT FIG. 7 — VISCOSITY OF A BIOPOLYMER IN SECOND

OF DIVALENT ION CONTENT. WALL CREEK CORES.

636 SOCIETY OF PETROLEUM EXGISEERS JOUR~AL

DETERMINING THE VOLUME TABLE 2- POLYMER ADSORPTION LOSS ME ASIJREMENTS IN BEREA CORE5

OF POLYMER REQUIRED Cerapul.d Polynm,

core Ad.erptio. L.a. ses (per..nt PV)

The volume of polymer should be large enough to L..g?h L.ass 300 500 750 I,oo+l
Polymer (It) Test CendiNOn (l b/bbl PV) cwm pm Pm Pr.m
protect the micellar slug from invasion by drive 2 NO tn,cell.r pra.ent 0.05 4S 29 19 14
Bi~pelym.r
1 Ad,orb.d mi..fil.r* o 0000
water as the slug is propagated through the reservoir. I No m,c.11.ar IW.s.nt 0.08 76 46 30 23
The effectiveness of the polymer bank in protecting P.lym,,Ylmn, de A 2 Ad.erbod m,c*llLv* 0.06 57 34 23 17
2 Pelynw displ. c,ng 0.01 9.5 5.7 3.B 2.8
the micellar slug is reduced by adsorption, by the m,.cel l.,
No m,. ell.a, 9,.,.”, 0.15 I 43 86 57 43
action of the polymer to penetrate the micellar slug Pcslyacvlm,de B ;
Pel,nw, d,splac,”g 0.06 S7 3A 23 17
through “inaccessible pore volume, ” and by dilution mic.llof

*Wate, ,nl.ct.d before F.cdymer m flesh OUI m,, ell.ar SOlu!mm

with drive water.

Poiymer Adsorption move through pores only above a certain minimum

Early flow tests in the absence of a micellar size, whereas the micellar solution flows through
fluid indicated polymer adsorption loss to be a smaller pores.
major problem. A sizable polymer-free water bank Inaccessible pore volume is illustrated by the
was predicted to build up at the leading edge of the laboratory core test data shown in Fig. 6. In this
mobility buffer bank, which could cause fingering test, a 2-in. -diameter x 24-in .-long Berea sandstone
due to loss of mobility control, and therefore could core with stainless steel endcaps was wrapped in
cause dilution of the micellar fluid. Laboratory epoxy fiberglass. Each endcap was constructed to
tests show, however, that polymer losses are allow a l/32-in. gap at the core face. This gave a
~-eatly reduced when a micellar fluid has previously uniform distribution of fluids at the inlet and exit.
contacted the rock. Table 2 gives polymerloss data A biopolymer and a ncmadsorbing tracer were
for Berea cores. As an example, the loss of a injected into a core containing micellar solution in
biopolymer was reduced from 0.05 lb per barrel of a residual form. The micellar solution had been
pore space when no micellar fluid was present to flushed out with water before polymer injection.
near zero when the polymer followed a micellar The polymer led the tracer at both the front and the
fluid. Polymer losses were determined by viscometric rear of the bank by about 18 percent PV. Polymer
or coIorimetric procedures (i. e., hypochlorite bleach losses were measured to be essentially zero.
method for polyacrylamides, and dihydroxynaphtha-
Polymer Ii’uter Mixing
lene method for biopolymer). Special handling is
required to overcome interferences when die sample A portion of the mobility buffer bank is diluted
contains sulfonate. by the chase water. The size of this mixing zone is
related to the mobility ratio of the mobi Jity buffer
MicelIar-Palynrer Mixing and bank and drive water.
Inaccessible Pore Volume A 32-ft Berea core test was conducted to determine
Tlieoretically, the size of the micellar-polymer the mixing zone (based on the 5 to 95 percent
mixing zone should be small because of the unit concentrations) grcwth rate for polymer water and
mobility displacement. However, experimental data polymer-free water having a 6.7 mobility ratio. Small
indicate significant “mixing. ” The polymer pene- fluid samples were collected at the end of each 8 ft
trates the micellar solution without dilution as of core. The growtl, of the mixing zone is shown in
wo)dd be expected in the normal mixing. This flow Fig. 9. These results show that the mixing zone
behavior is explained by the phenomenon of polymer grows linearly with distance but occupies less
inaccessible pore volume, previously described by
Dawson.18 Penetration of the s1ug apparently Kll 1 1 , I [ r
I I I 1 r I
results because the large polymer molecules can

PORE VOLUMES PRODUCEO

FIG. 8 — BIOPOLYMER PROPAGATION, ILLUSTRATING

INACCESSIBLE PORE VOLUME IN A BEREA CORE.

DECEMBER. 1974 637

fractional pore volr.rme with increased iength of adsorption (Fig. 1 Id). Based on these observations
travel. This occurs because the mixing zone curve the volume required for a single layer is
does nor pass through the origin. Fig. 9 also shows
that the polymer-water mixing zone did not stabil{ze Volume requirement of mobility buffer slug =
19
as did an earlier glycerine-brine displacement,
(Pv ~-w) + the greater of (PVad~ , PVIPV ) .
The zone could stabilize beyond 32 ft. If so, the
polymer volume would be smailer than predicted
with the existing data. Laboratory data indicate that “ inaccessible pore
It is recognized that a laboratory tes: cannot be volume is usually greater than the adsorption loss
rigorously scaled to the reservoir. For example, for polymers following a micellar solution. The
mixing could be inhibited to a greater degree in curves of Fig. 10 show the total predicted volume
2-in. -diameter cores than in the reservoir. requirement using the previously discussed polymer-
Extrapolation to field-size dimensions was made water mixing data and varying values of inaccessible
using the growth-rate behavior established for the pore volume. The inaccessible pore volume typically
first 32 ft. The following assumptions were made: in laboratory cores is 20 percent. This value could
I. The relation between mixing zone length and ~ary with pore size, pore-size distribution, and
distance of travel is the same in a radial system as polymer type. A total pore volume ranging from 41
in a linear system. to 44 percent is predicted for radii greater than 100
2. The average polymer concentration in the ft, assuming a zo percent inaccessible pore volume.
mixing zone is one-half the injected vaiue.
3. The amount of polymer required is that amount FLUID PROPAGATION AND INTERACTION
contained in the polymer-drive water mixing zone.
Short-core (4 ft) screening tests provide valuable
The bottom curve of Fig. 10 shows the predicted
data on the ability of a micellar fluid to displace
volume of polymer required because of polymer-water
oil. !-fowever, oil displacement efficiency, although
mixing for a 6.7 mobility ratio. As expected, a of major importance, is not the only criterion that
decreasing pore volume requirement is predicted
ultimately determines how a system will perform in
with increasing dimensions. A volume ranging from the field. The mobility, interaction, and propagation
21 to 24 percent PV is predicted for radii greater of the fluid banks must be understood before a
than 100 ft. This volume would be smaller if satisfactory field desipn is completed. Such an
consideration were given to the more favorable
evaluation requires more sophisticated experimental
mobility displacement at the higher velocities in
techniques. One approach is to conduct micellar
the vicinity of the wellbore. floods with fluid banks of varying size and
Prrdic’tion 0{ Total Poly7ncr composition to note the propagation rates and
Vrrlume Requirmnpnt nobilities of these banks from pressure gradient
behavior. Chemical analysis of the core effluent
Fig. 11 illustrates the various parameters affecting
can be used in conjunction with the pressure
the polymer pore volume requirements. Ideally, both
behavior to provide information on the bank
the adsorption and the inaccessible pore volume
interact ions.20 Generally these studies are con-
should be zero (Figl I la). The polymer volume in
ducted in systems longer than 8 ft for two reasons.
this case is the minimum possible, and the polymer
First, the volume of the zones of mixing and
movement is piston-like behind the micellar ~olution.
interaction should be great enough at the core exit
Adsorption in excess of inaccessible pore volume
for chemical analysis. Second, the interval used to
results in an undesirable buildup of polymer-free
monitor the pressure gradient should be less than
water behind the micellar slug (Fig. llb). The
mixing zone lengths. This necessitates closely
polymer also moves as a piston behind the micellar
spaced pressure taps and sensitive transducers in
slug when the volume lost through adsorption equals
short systems, or less sensitive monitoring devices
the inaccessible pore volume (Fig. llc). Finally,
in longer systems with widely spaced pressure taps.
the polymer penetrates the slug when the inacces-
sible pore volume exceeds the volume lost through POLYMER vOLUM[
~lxlNG vO(UM[ OF PO+YMfR INACCC$SIBLE
PROTECTING MICEUAR
RFOUIRLUtNl AOSORPTI U4 PORE VtllM[
.—

Pvp -w
100 1 ( I I I 1 11 I I 1 1 f , ! 1 I

> 80 INACCESSIBLE PORE VOLUME, V~ zOENUOIO ZtNf

f
Pvp.w
‘a: 3
. >0 > 0
20
p’Jad>
15–
Pv
p -* LO
o . >0 .
Pvw$[ .Pv, pv 1

I MOBILITY RATl O “ 6.7

I

Pv
lNvASltM

1 ! 1 1 1 1 1 1 1 I I 1 I 1 1 1 I 1 J ,
P-w
(d) ~o < >0

10 20305070100 1090 PVIPV

RADIUS OF SYSTEM, ~EET FIG. 11 — EFFECT OF ADSORPTION AND INAC-

FIG. 10 — PREDICTED POLYMER VOLUME REC?UIRE- CESSIBLE PORE VOLUME ON POLYMER VOLUME
MENTS. REQUIREMENTS.

638
 Long-core tests are conducted in 2-in. -diameter TABLE 4- FLUID INTERACTION AND PROPAGATION,
LARGE MICELLAR SLUG TEST
Berea sandstone cores 8 to 24 ft long (8-ft pieces
Rate of Apparent
are butted with porous paper at the joints to provide
Advonca Viscosity
capillary continuity). The cores are wrapped in Fluid Type (x Plug flOw)* (cp)**
epoxy fiberglass. Pressure taps are located, and Water at residual oil saturation 10.0
pressure drops are measured across intervals. From Low.mobility oil-water 4.6 15.4

these data the comparative mobility and apparent I+igh.mobility oil-water 1.8 6.8
Low.mobility “emulsion” at 0.92 20 at inlet
viscosity of the flowing fluids can be determined.
micellar leading edge !33 at exit
As the fluid banks interact, changes in mobility, Stabilized micellar bank 29.6
propagation rate, and mixing-zone development can High-mobility sulfonate-polymer 3.2 10.7

be followed. The arrival of mobility anomalies at mixing zone

Stabilized polymer bank 12.2
the core exit can be predicted from the pressure-drop
Chose-water invasion 2,2
behavior. An analysis of the produced fluid relates Final chase water 1.07
pressure-drop changes to altered fluid compositions. *1. O for hypothetical fluid flowing through all the po~e volume
Two quantities are used to interpret the pressure (plug flow).
**Apparent viscosity = { k ~~(kw)Sor ~/comparative mobility.
behavior: the comparative mobility and the apparent
viscosity. The comparative mobility relates the
mobility of the flowing fluid t~ that of water at
fractions, dimensionless component concentrations
residual oil. Therefore, at constant flow rates, it is
in the water phase, and dimensionless oil/water
the ratio of pressure drop at residual oil to the
interracial tensions (IFT) vs the pore voIume of
pressure drop at any time (ApsO, /Af’t ). The apparent
effluent produced after the start of micellar
viscosity is used to give a quantitative indication injection. The comparative mobility of the last pair
of the flow resistance. It relates the measured of pressure taps (2 in. and 14 in. from the core exit)
pressure drop to rhat of brine at the absolute
is also included in this figure. Fig. 13 plots
permeability. For example, for the system described
comparative mobility (and apparent viscosity) vs
in Table 3 the absolute permeability to brine is pore volume for three pairs of pressure taps.
571.2 md and the permeability to brine at residual Included in this figure are curves predicting the
oil is 36.6 md. Therefore, the apparent viscosity of arrival of the different fluid banks at the monitored
brine at residual oil is 15.6 times the true brine interval when plug flow is assumed. From these
viscosity. curves the propagation rates of the zones of
Two tertiary long-core tests will be discussed in interaction can be determined relative to the
detail here. The first test, a large micellar slug
predicted plug flow rate, Table 4 lists the various
test, highlights the development of various zof:es
zones of interest, their rate of advance, and the
of interaction: a high-mobility oil-water zone, a
low-mobility oil-micellar “emulsion” zone, an~ a
high-mobility sulfonate-polymer mixing region. 7’he
second test, a 10-percent PV micellar slug tt?st,
shows similar results to a lesser degree as well as
the effects of polymer penetration of the micel’ar
slug and the subsequent interaction with ths
oil-water bank.

LARGE MICELLAR SLUG TEST

Table 3 lists the rock properties, fluid
compositions, and volumes of a representative 16-ft
Berea core test when a large bank of the micellar
fluid is injected. Fig. 12 is a plot of the oil

PV

+
T=
~ =
TABLE

=
llW’F
12 ~l/hr

2,146ml

0.217
3 — LARGE MI CELLAR

Pw
SLUG

s ~c

~lti”~
Scr
TEST

=
=
=

11O”F = 4,0 ~p
110
44.6
34.4
~.64
percent
percent
~p
‘“’r-F
1.0 -

?A
d=2 in. k = 571.2 md
i
L=16ft (kw)s = 36.6 md COSURFACIANT -4 ; ~.
or
Crude oil = Second Well Creek 0,5 -

Iniection Sequence

1. 0.355 PV preflush - 0.2N NaCl; 50 percent Madison water

(MW), 50 percent softened Modi son water (SMW).
ok
2. 1.319 PVmicellar-5: 3:92, surfoctant :cosurfoctant: brine in 1.0 2.0 3.0
0.2N NaCl; 50 percent MW, 50 percent SMW).
PoRE VOLWIS PROOUCIO
3, 1.051 PV mobility buffer -1,000 ppm biopolymer in 50 percent
MW, 50 percent SMW with 9,200 ppm ethonol tracer.
FIG. 12 —LL3NG-cORE TEST, LARGE MIcELLAR ‘LUG
4. 0,546 PV chase water - 50 percent MW, SO percent SMW. INJECTION.

DEc EM BE R.19?.r 639

apparent viscosity of these zones. The followirig difficulty of determining sulfonate accurately at
observations and conclusions were made from this high oil concentrations.
test.
Micellar Zone
Or’I-Kater Zo77e
At the leading edge of the micellar bank a
The leading edge of a low-mobility oil-water bank low-mobility zone follows the high-mobility oil-water
advances at a constant rate and is associated with zone. The apparent viscosity of this thickened fluid
the initial decrease in comparative mobility. The increases with distance traveled. This viscosity
apparent viscosity of this zone is in agreemem: with increase may be Jue to the formation of an oil-
the maximum resistance determined from the total surfactant emulsion in the mixing zone or to
mobility plot for Berea in Fig. 4. A zone of high- surfactant/clay interaction with a resultant increase
mobility oil-water follows the high-viscosity region in calcium concentration, (Studies show that the
with no appreciable change in flowing oil saturation micellar/clay interaction may expose calcium for
(about 0.40). However, a decrease in oil-water IFT ion exchange with the micellar fluid if care is not
corresponds to the development of this high-mobility taken in the initial saturation of the Berea core to
oil-water zone. Chemical analyses of the oil and remove the calcium. ) Since the mobility of the
water show a sulfonate concentration approximately stabilized micellar bank is less than that of the
0.1 percent of injected in the water phase and oil-water bank, mobility control is achieved.
approximately 8.0 percent of injected in the oil However, it is not achieved between the emulsion
phase. The concentration of sulfonate in the oil zone and the stabilized micellar fluid.
phase is not included in Fig. 12 because of rhe The maximum sulfonate concentration in the
effluent is 1.56 of the injected value. This
I concentration increase is attributed to the retrograde
1“1
1

!5 A PRIS5UR[ TAPS 3U ANO 4? fRCfA COR[ IN{I1 adsorption isotherm of the sulfonate (see Fig, 2). A
sulfonate adsorption loss of 1.37 lb/bbl PV is
determined by integrating between the leading edges
of the surfactant and cosurfactants profiles. Similar
integration in the regio.] of sulfonate desorption
(where the sulfonate concentration is greater than
injected) shows a desorption of 0,78 lb/bbl PV. The
net sulion.ate adsorption loss is 0.59 lb/bbl PV.
This agrees with the adsorption minimum determined
in the equiliLiium adsorption studies (Fig. 2).

Polymer Zone

B PR[SSUR[ TAPS ‘W” ANO ilr !R13V COR[ lNLff

As polymer-thickened water is injected, a zone of
1>
CCMPARAT IV[ high mobility develops between the polymer and
MOBIIIIY
micellar banks. This high -mobility fluid in
CHAS[ wATER
lAtCilLAR
.- OIL -wA1[R POIYfAf R - conjunction with polymer inaccessible pore volume
10— ~ ,. . . . . . . . ..-...? l--- results in the rapid propagation of rhe sulfonate -
polymer mixing zone. Corresponding to the
production of this high-mobility fluid, there is an
+ increase in polymer concentration and decrease in
05
sulfonate concentration.
Ethanol, the polymer water tracer, lags the
polymer conside..ably. Polymer is produced 31
0
10 ?0 percent PV after injection, but ethanol is not
0
g 1 : produced until 62 percent PV after injection. The
s
15 - C Pkf55uR[ lAPS 13Y AUO 147” fRWA CORf INIFT 50 percent concentration of the ethanol occurs at
$
. CChMPARAl lVf
MOBIIIIY 1.0 PV of polymer injection, and this indicates that
8 :
011-WAIIR all the pore space is accessible ●to the water

10 — . . . . . . . ..<!! ’![L{AR
. . . . ...+’!’?’.- – 10
(although not to the polymer molecules). The rapid
movement of the polymer independent of its tracet
1 1?
water allows it to move into the micellar bank
/:
i lb without having a diluting effect. The polymer lead
05 ?0 is greater when a micellar bank precedes polymer
!
~ yl ,: ~
?4 injection (18 percent PV for Fig. 8 and 38 percent
1: 54 PV for Fig. 12). This is attributed to the micellar/clay
Im
0 interaction, which reduces the rock permeability. As
o 10 ;0 30
POR[ VOIUMIS PR04JuCID
a result, a greater fraction of the pore space is
incapable of polymer transport.
FIG, 13—COMPARATIVE MOBILITY AND PREDICTED
Since the mobility of the stabilized polymer bank
AVERAGE FLUID COMPOSITION IN INTERVAL
MONITORED. is greater than that of the micellar bank, an

640 SOCIETY OF PET ROLE1” M ZS121XEERS JOURNAL

increased polymer concentration was needed in this bank has a high %obility, compared with 55 percent
test to achieve mobility control. in large-slug tests). Ninety percent of the recovered
surfactant is produced in the oil phase of the
Chase Water
high-mobility region. Eighty percent of the tertiary
As the high-mobility chase water is injected, it oil was produced with a 10 pe:cent PV of micellar
invades the polymer bank. Polymer inaccessible fluid. The rapid “tail-off” of the oil fractions
pore volume increases the rate of advance of the indicates that good mobility control was achieved
high-mobility fluid since the polymer concentration with 1,250 ppm biopolymer (ihe pressure data also
begins decreasing (creating a more mobile fluid) showed this).
befor~ the ethanol tracer concentration decreases.
After several pore volumes of brine are injected, MiceI[ar Zone
the final rock permeability is only 60 percent of the Since the micellar bank was small, it was not
absolute permeability. This decrease is attributed possible to study its propagation along the entire
to micellar/clay interaction and polymer entrapment. length of the core. A Iow-mobi Jity zone did not
develop in the region of micellar- oil mixing.
SM”ALL MICELLAR SLUG TEST
IN A 24-FT BEREA CORE Eighteen percent of the injected surfactant was
recovered in the oil phase, 2 percent in the water
The injection of large volumes of a micellar fIuid
phase. A net active surfactant loss of 0.80 lb/bbI
in a long-core test isolates and amplifies the
PV was determined. The maximum cosurfactants
interaction between the oil-water bank and the
concentration in the effluent was 46 percent of
micellar slug as well as micellar-polymer interaction. injected, although cosurfactants recovery was 96
It can be argued that the injection of a large bank percent.
of micellar fluid simulates the field behavior of the
fluids near the injection well. However, testing Po[ymer Zone
procedures have not been developed to scale the The polymer did not lead the ethanol tracer. This
behavior in the interwell region. indicates that polymer propagation is retarded by
In large-slug tests the tertiary oil is produced the oil bank (due to polymer insolubility in oil).
before polymer water is injected, whereas in the However, at the tailing edge of the mobility buffer
field, when a small slug of micellar fluid is followed zone the concentration of polymer begins to
by polymer water, the polymer water may penetrate decrease before the ethanol tracer. This indicates
the micellar bank (due to inaccessible pore volume) that polymer inaccessible pore volume has not been
and invade the oil-watel bank. The resulting fluid eliminated from this region. As a result, polymer
behavior can be examined in the laboratory by using accumulates in the mobility buffer bank and its
small slugs of micellar fluid in long-core tests. concentration increases to a value greater than
Table 5 lists the injection sequence and system injected. When the polymer was initially injected
properties for a small-slug test and Fig. 14 shows and contacted the micellar bank, a zone of high
effluent analyses. Pressure behavior was also mobility developed near the core inlet. This high-
examined, but it was not so dramatic as the mobility region soon dissipated, and stabilized flow
previously presented example because the slug conditions were reestablished at 25 percent PV.
volume was small. The following general It is significant that the polymer did not propagate
observations were made. into the oil-water bank, ~ ‘rice its effect on water
viscosity would be to increase the total flow
Oil-Water Bank
resistance of the oil-water bank. The small-slug
A low-mobility oil-water bank develops with an test shows that zones of instability that develop
apparent viscosity similar to that noted in the
previous test. Again a high-mobility oil-water zone
develops (in this test 35 percent of the oil-water

TABLE 5 — SMALL MI CELLAR SLUG TEST

T = 110”F s Wc = 43.5 percent

o
q = 12ml/hr s = 35.3 percent
I lo% = o.~~ =P
PV = 3,082 ml Pw
rj = 0.208 11O”F = 4mI) CP
PO
d=2 in, k = 653.2 md

L=24ft (kw)sor = 49,6 md

Crude oil = Second !4011 Creek

Iniection Sequence

1. 0.492 PV preflush - 0.2N NaCl; 50 percent Madison water

(MW), 50 percent softened Modi son water (SMW).

2. 0.100 PV micellar -5:3:92, surfoctant : cosurfactants : brine m

in 0.2N No Cl; 50 percent MW, 50 percent SMW. 10 2.0
PoR[ vOltJMES PROOUCfO
3. 0.873 PV mobility buffer -1,250 ppm biopolymer in 50 percent
MW, 50 percent SMW with 10,000 ppm ethanol tracer.
FIG. 14—LONG-CORE TEST, SMALL MICELLAR SLUG
4. 1.652 PV chase water -50 percent MW, 50 percent SMW. INJECTION.

l) E(:EMBKR. 197! 641

when a large bank of micellar fluid is injected are p. = viscosity of oil, cp
not eliminated. However, the magnitude of the pw = viscosity of water, cp
instability is reduced as mixing occurs.
REFERENCES
CONCLUSIONS
1, Jones, L, W.: U. S. Patent 3,126,952 (March 1964).
The following conclusions are based on our 2. Jones, L. W.: U. S. Patent 3,308,068 (March 1967).
laboratory developmental work.
3, Hill, H. T., Reisberg, J., and Stegemeir, G. L,:
1. High tertiary oil recoveries are achieved with i (Aqueous Surfactant Systems for Oil Recovery, ” J.
small volumes (< 10 percent PV) of the 92-percent- Pet, Tech. (Feb. 1973) 186- 194; Trans., AIME, VOL
water micellar fluid, 255,
2. The adsorption loss of the isopropyl alcohol 4. Foster, W. R.: “A Low Tension Waterflooding
cosurfactants is negligible. Therefore it can be used Process, ” ]. Pet. Tech. (Feb. 1973) 205-210; Trans+,
AIME, Vo], 255.
as a micellar tracer.
5. Gale, W, W., and Sandvik, E. E,: “Tertiary Surfactant
3, Sulfonate adsorption losses show “retrograde”
Flooding: Petroleum Sulfonate Composition-Efficacy
behavior. .Minimum adsorption loss is observed at Studies, ” Sot. Pef. Eng. j, (Aug. 1973) 191-199.
the 92-percent-water micellar fluid composition. The 6. Gogsrty, W. B., and Teach, W. C.: “Miscible-Type
use of a sodium chloride brine preflrsh reduces Waterfloodinfg Oil Recovery with Micellar Solution s,”
adsorption losses and increases oil recfwery. J, Pet, Tech (Dec. 1968) 1407-1414; Trans., AIME,
4. Permeability reduction due to :he micellar Vol. 243.
fluid decreases viscosity requirertents. This 7. Holrz, L, W.: “Use of Soluble Oils for Oil Recovery,”
represents a substantial reduction in total poIymer J. Pet, Tech. (Dec. 1971) 1475-1483; Trans., AIME,
Vol. 251.
cost.
8. Froning, H R.: Canadian Patent 913,349 (Oct. 1972).
5. Inaccessible pore volume increases micellar
invasion by poIymer and necessitates an increase 9. Froning, H R., Raza, S. H., and Askew, W. S.: U, S.
Patent 3,722,590 (March 1973).
in the mobility buffer bank volume.
10. Askew, W. S., and Froning, H R.: U. S. Patent
6. A polymer “denuded” zone is not observed
3,714,062 (Jan. 1973).
between the micellar and polymer banks because
11. Dauben, D. L,, and Froning, H R,: “Development
polymer adsorption losses are reduced substantially and Evaluation of Micellar Solutions to Improve Water
when miceilar fluid precedes the mobility buffer Injectivity, ” j. Pet. Tech (May 1971) 614-620.
bank. 12. Corrin, M, L,, Lind, E, L., Roginsky$ A,, and Harkinsj
7. Long-core tests provide valuable mechanistic W. D.: I{Adsorption of Long-Chain Electrolyte’s from
information on the micellar flooding process. The Aqueous Solution on Graphite of Known Area and on
use of pressure gradient data and chemical analyses Polystyrene, ” J. Col[oid Science (1949) Vol. 4, 485-
49%
of core effluent has shown the interactions, propa-
13. Void, R. D., and Pkansalkar, A. K.: “Adsorption and
gation, and nobilities of the various fluid banks Deposition of Carbon Qn Cotton from Solutions of
used in micellar floods. Sodium Dodecyl Sulfate, ” Rec. Trav. Cbirn. (1955)
8. Polymer inaccessible pore voIume is greater vol. 74, 41-51,
when the mobility buffer bank follows a micellar 14, Fava, A., and Eyring, H.: “Equilibrium and Kinetics
slug. .jf Detergent Adsorption —a Generalized Equilibration
9. A high-mobility oil-water zone develops ahead Theory, ” J. PLys. Cbem, (1956) Vol. 60, 890-898.
of the micellar bank. This observation may lead to 15, Paul, G. W., and Froning, H R.: “Salinity Effects of
Micellar Flooding, ” j. Pet, Tech, (Aug. 1973) 957-
reduced viscosity requirements.
958.
10. Polymer does not invade the oil-water bank in
16. Gogarty, W, B., Meabon, H, P,, and Milton, f-!, W.:
small-slug micellar tests. Therefore, unfavorable
“Mobility Control Design for Miscible-Type Water-
nobilities due to polymer inaccessible pore volume flood Usin~ Micellar Solutions, ” J. Pet. Tech. (Feb.
do not develop in this region. 1970) 141-147.
17, Froning, H R., and Askew, W. S: U. S. Patent
NOMENCLATURE 3,719,606 (Msrch 1973).
18. Dawson, R., and Lantz, R. B.: ‘{Inaccessible Pore
kro = relative permeability to oil (kro = 1 at Volume in Polymer Flooding, ” Sot. Pet. Eng. ].
connate water saturation) (Oct. 1972) 448-452; Trans., AIME, Vol. 253.
k= relative permeability to watel 19. Hall, H. N., and Geffen, T. M.: “A Laboratory Study
pVa~ = fractional pore volume of mobility buffer of Solvent Flooding, ” Trans., AIME (1957) VOI. 210,
48-57.
(polymer) bank stripped of polymer
20. Davis, J. A., Jr,, and Jones, S. C.: ‘~Displacement
PVIPV = fractional pore volume of polymer inac- Mechanisms of Micellar Solutions, ” J. Pet. Tech
cessible pore volume (Dec. 1968) 1415-1428; Trans., AIME, Vol. 243.

Pvp. w
= fractional pore volume of polymer in
polymer-water mixing zone
 DISCUSSION

B. 1.. KNIGHT
S. C. JONES MARATHON OIL CG,
f?. ‘U. PARSONS L/ T’TLETON, COLO.
MEMf3ERS SPE-AIME

The authors are to be commended for focusing same as the effective permeability when polymer is
on aspects of micellar flooding that have heretofore flowing, then the effective in-situ polymer viscosity
received little attention in the literature. Under- is obtained by dividing apparent viscosity by
standing interactions among various injected fluids permeability reduction, as shown in Table D-2. In
should form the basis for designing more efficient these experiments, the permeability reduction effect
systems. contributes more to mobility control than in-situ
We are concerned, however, that the reader may viscosity. Millipore filtration tests indicated
be left with the mistaken impression that all microgel formation. This is probably due to
micellar floods behave as shown in t,,is paper In biopolymer-brine interaction, and is consistent with
fact, different systems can show widely differing the permeability reduction observation.
behavior. Hence, the reported data are specific to For design purposes, we recommend that polymer
the particular fluid-rock-environment c~mbination. “viscosity” data be obtained from core floods. In
Consequently, certain interpretations and conclusions this way, measured apparent viscosity automatically
are system-specific. includes permeability reduction and non-Newtonian
A brine preflush is not “typical”. Many reported flow behavior, which cannot be predicted because
field test_s have not used such a reservoir of changes in shear rate caused by permeability
conditioner,*’5 Possible benefits from a preflush, reduction and/or inaccessible pore voiume effects.
in terms of slug stability and/or reduced sulfonate It is not clear how the authors combine,] data from
aclsorption, must he weighed against the cost and viscosity measurements and core tests to arrive at
time of additional fluid injection and a judgment of their Fig. 7. We do question the use of viscometric
preflush effectiveness on a field scale. data.
The rock used in flooding experiments can The concept of inaccessible pore volume is
significantly affect flooding behavior. We feel useful in explaining rapia propagation of polymer
strongly that reservoir rock, as well as reservoir molecules relative to tracer components. However,
fluids and conditions (temperature, velocity, etc.), the following experiments suggest that adsorption
must be used in displacement efficiency studies as and i~accessible pore volume alone may not be
well as surfactant adsorption and polymer behavior sufficient to explain the unequal propagation
investigations. Berea systems do offer certain velocities. Three pore volumes cf polymer solution
advantages in mechanism studies, mainly because containing tritiated water tracer and Dow Pusher @
of larger models and relatively good homogeneity; 700 or Kelzan MF polymer were injected into
but these data should not be applied directly to the water-saturated cores followed by 3 pore volumes
field. Table D-1 shows examples of the difference of water. Effluent samples were collected in 0.04-
the rock can make. PV increments and analyzed for both polymer and
The authors note that biopolymer reduce tracer. Table D-3 shows polymer IOSS calculated by
mobility primarily by building viscosity, whereas material balance at the end of water injection, and
polyacrylamides increase viscosity and reduce breakthrough of polymer and tracer. Polymer 10SS
permeability. However, biopolymer can substantially with both materials was nearly equal. With equal
reduce permeability under certain conditions, adsorption losses, one might expect the higher
particularly when saline waters are involved. Table molecular weight, presumably larger polyacrylamide
D-2 gives the results of experiments showing that molecules to propagate faster than the biopolymer
both polymers reduce permeability in a Rocky
Mountain reservoir sand. Duplicate experiments ———. .
——. —.— .—--—-—
gave similar results. Permeability reduction (initial TABLE D.? - POLYMER FLOW BEHAVIOR

+ flushed brine permeability) is due to residual Rote, Appa. en, Pe,nte.b,l,,y lm. sl,.
Polymer 6 k, md ft/G v,sc<, s,ty,’ CP Reduc t,..” V, SCOS,,7, CP
—.—
polymer in the rock after mobile polymer has been KeIzon MF O~ti %“ 30 4.9 4 1.2
flushed by brine. If the flushed permeability is the 3.0 10.6 5 2. I
0.3 25.4 13 2.0
maw 700 0.121 90 30 27.3 7 3.9
3.0 47.8 13 3.7
TABLE D-1 — TERTIARY OIL RECOVERIES IN 0.3 35.6 36 I,o
DIFFERENT ROCKS CO.*S 1.,.. dt.rneter x 3s.. s.nds tone. p,.ss.re dot. from rem 2’% ,n.
(Identical Flooding Conditions in lhin. Di~meter X fn, t,al S.torotmn: Br,.e w,!h 8,300-pmI TDS, 570-pPm dtvalent c.atmns.

Skin. Long Rocks) Combos,,, o”: 1,500.PPm pdym. r ,. brtne w,th 6, 1%.ppm TDS. .O
d,wlenl ..*,ons.
Oil Recovery, Percent of SO, Polymer
Ram s.w . . . . 30 If/D, 3 ItIO, 0.3 f,jD vitith pres..re s+ab,l,,. h..
— \ bel.,e e.<h ch.nse.
Fluid System Fired Berea Reservoir Rock C.nw.s,t,On: 6, 150. pptn TOS, . . d,w.al en? c.t,ons.
F1. sh W.,. r Rote sequence: 0.3 It/D, 3 I!/D, 30 ft/D w,th pressure smb, f,ratmn
1 79 32 { belom each change.
2 90 24 “APPOI*. f “8SCC’SEIY - I/[telotm,e rnob,l,, y].
. .Permeeb,lify r.duc t,.. klnjt,.llkf}u shed.
3 92 40

DE(:EMRER. f971 6.!3

 TABLE D-3 – POLYMER-TRACER BREAKOUT BEHAVIOR polymer in the oil-water bank, whele surfactant
Polymer Loss
Breakthrough, PV concentration is reduced, is much lower. Thus, the
Run Polymer Percent Iniected Polymer Tritium combined effects of relative permeability, viscosity
— .

1 Dow 700 10.3 0.85 0.85 and partial plugging impart a favorable mobility
2 Dow 700 10.6 0.85 0.89 ratio to the polymer solution with respect to the
3 Dow 700 — 0.83 0.87 oil-water bank. This relative permeability effect
4 Kelzan MF 11.1 0.70 0.81 can be verified by injecting into a core an amount
5 Kelzon MF 11.9 0.7J 0,87
of micellar slug insufficient to recover most of the
Cores: 3-in, diometerx 12-in. lon~ fired Berea (90to120md). oil. When polymer solution is forced through the
Water — 500. ppm TDS, no divalent cations.
portion of the core that contains a high residual oil
Solutions: Polymer — 1,000 ppm in above water (tritiated
{ water tracer). saturation, pressure gradients become high with
Frontal Velocity: 10 ft/D. respect to those where most of the oil has been
removed.
From the presented data, it is evident that
since a larger fraction of the pore space is
micellar flooding involves some very complex
inaccessible to the larger molecules. The data
mechanisms. With the wide variability in systems—
show the opposite effect. Kelzan breakthrough led
micellar slug, mobility buffer, reservoir fluids and
the tracer by 11 to 16 percent PV, while Dow 700
rock — it is not surprising that mechanism details
led the tracer by only O to 4 percent PV. It would
are not universal.
appear either that one cannot predict relative
inaccessible pore volumes in a given rock type
REFERENCES
from molecular weights alone, or that the
adsorption/inaccessible pore volume concept as 1, Gogarty, W. B., and Davia, J. A., Jr.: “Field
described by the authors is not sufficient to Experience With the Maraflood Process, ” paper SPE
explain observed flow phenomena. 3806 presented at SPE-AIME Symposium on Improved
Oil Recovery, Tulsa, Okla., April 16-19, 1972.
The authors concluded that polymer solution will
flow piston-like behind micellar slug when the 2, Jones, S. C., and McAtee, R. W.: “A r~ovel Single-Well
Field Test of a Micellar Solution Slugj” ./. Pet. Tet,lr,
effects of polymer adsorption are exactly balanced (Nov. 1972) 1371-1376.
by inaccessible pore volume effects. Their argument
3, Knight, R, K., and Baer, P. J.: “A Field Test of
did not consider mobility ratio. We feel that piston- Soluble-Oil Floodi, ~ at H:ggs Unit, ” ./. PCI. Tr,ch,
like displacement will occur only with the added (Jan. 1973) 9-15.
constraint of favorable mobility ratio. 4, Earlougher, R. C., Jr., O’Neal, J. E., and Surkalo, H.:
Polymer was observed to propagate into the “Micellar Solution Flooding: Field Test Results and
mice~lar slug (large slug test), but not appreciably Process Improvements, ” paper SPE 4875 presented at
Second Midwest Oil and Gas Symposium, Indianapolis,
into the oil-water bank (small slug test). ‘We would Ind., March 28-29, 1974.
suggest that this phenomenon results from relative
5. Danielaon, H. H., Paynter, W. T., and Milton, H. W..
permeability effects, not polymer insolubility in oil. Jr.: , ‘TertiarJ, Recovery by the Maraflood Process ‘n
In the presence of high surfactant concentration in the Rradford Field, ” paper SPE 4753 presented at
the micel]ar slug zone, the relative permeability to SPE-AIME Symposibm on Improved Oil Recovery,
polymer solution is high. However, permeability to Tulsa, Okla., April 22-24, 1974.

AUTHORS’ REPLY

Knight, Jones, and Parsons bring up several adequately preflush high-salinity (> 60,000 ppm TDS)
interesting points which amplify and sometimes reservoirs ‘..ith sufficient volumes to reduce salinity
conflict with our observations. We concur that our to levels compatible with many present micellar
results are not universal for all mice] lar and rock formulations. In the case of many Rocky Mountain
systems, especially systems differing considerably formations the salinity must be increased rather
in component concentrations and fluid injection than decreased. Controlled salinity brine preflushes
sequence. A major purpose of our paper is to appear to be satisfactory (and are often mandatory)
demonstrate some of the mechanisms of micellar for such applications. Efforts are in progress to
flooding and the interactions between the different develop micellar fluids that are compatible with a
fluid banks. It is essential to conduct these broad range of resident brines.
studies in long homogeneous systems such as We agree with Knight et al, that the design of
Berea sandstone. micellar and polymer banks should be based on
A brine preflush is typical in the micellar native-state core tests that examine oil displace-
systems reported in the literature by the vast ment, adsorption, and mobility. For the detailed
majority of production research organizations. mechanism studies reported in our paper longer
Because of the ion exchange capacity and systems, such as Berea cores, were required to
heterogeneity of porous media, it is difficult to evaluate fluid bank interaction as a function of

644 SO CIE”rl OF PKTRo I.EI” M EYGIx EERs JO I’RX.\l.

system length. Numerous studies in shorter, native, should be examined at the back of the pclymer bank
Second Wall Creek systems verified that the same since adsorption has minimum influence in this
major phenomena occur as observed in Berea. For region. In our Fig. 8 the polymer and tracer profiles
example, our oil displacement tests in Second Wall are similar but do not coincide. In this case the
Creek cores show an average tertiary oil recovery IPV equals the percent PV separating the two
of 78 percent compared with 84 percent recovery curves. Often the two profiles are quite dissimilar.
shown in our Fig. I for Berea. These observations In such cases an accurate determination of IF’V can
apparently do not agree with those of Knight et al, be made by integrating and normalizing the area
with their fluid systems. We recognize that between the trailing edges of the polymer and tracer
heterogeneities in native cores can cause poor profiles. Due to the skewed nature of the profiles,
sweep and therefore result in reduced oil recoveries. we have observed many instances where the polymer
Neither the high recovery we observe nor the low may break through before the tracer as much as 35
recovery Knight reports are quantitative indicators percent PV, but the integrated IPV determined at
of actual recoveries that would be achieved in a the trailing edges is only 20 percent PV. These
field application. Both indicate that the micellar profiles become skewed because the polymer
fluids are capable of moving oil. experiences a different mobility gradient than the
Our data indicate that well filtered biopolymer, tracer. This results in differing degrees of localized
such as Kelzan MF, reduce mobility principally by mixing. Perhaps this integration technique would
increasing viscosity. For example, in two tests lead Knight et al. to different conclusions. We doubt
where permeability was decreased in Second Wall that small molecular weight variations between
Creek cores, the reduction was slight. In one test polymers will be reflected directly in IPV. The
the permeability was reduced from 178 to 144 md, magnitude of IPV depends strongly on polymer type,
in another from 92 to 61 md; a 1.23- and 1.5-fold degree of hydrolysis, water salinity, and cross-
reduction, respectively. As absolute permeability sectional area of the po!ymer molecules.
(and pore size) is reduced, we would expect to see Knight et az. suggest that relative permeability
a reduction in permeability due to polymer, but not effects are responsible for the inability of polymer
to the degree shown by Knight et ai. in Table D-2. to penetrate the low IFT, oil-micellar region. This
The in-situ viscosities shown in that table suggest explains a reduced rate of invasion bur does not
that the data have been misinterpreted. A 1,500-ppm explain the apparent inability of any polymer
Kelzan MF solution has a Brookfield viscosity an penetration of the oil-micellar mixing zone.
order oi magnitude higher than the in-situ viscosity Furthermore, relative permeability cannot explain
reported by Knight et al. It appears that much, if polymer concentrations in the effluent that are
not all, of the apparent viscosity attributed to greater than iniected. In tests where as little as 3
permeability reduction (for Kelzan MF) is the result percent crude oil is added to a water-external
of in-situ polymer viscosity. At O. 3 ft/day an in-situ micellar fluid not only is the polymer incapable of
viscosity of 25.4 cp for 1,500 ppm Kelzan is not invading the oil bank, but it cannot penetrate the
unreasonable. micellar bank. Knight indicated that permeability to
Knight et al. suggest that adsorption and polymer should be high in this region of one-phase
inaccessible pore volume (IPV) alone are not micellar flow. Instead, polymer permeability is zero.
sufficient to explain observed polymer flow In summary, we agree that all micellar fluids,
phenomena. Our data appear to substantiate the polymers, and reservoir environments do not behave
concept originally proposed by Dawson and Lantz 18; the same. However, our paper does demonstrate
that the velocity of the front end of the polymer phenomena that are important to understanding the
bank is governed by adsorption and IPV, while the technology involved in surfactant-polymer flooding.
rear is only a function of IPV. To accurately ***
determine the IPV the polymer and tracer profiles