SOCIETY OF PETROLEUM ENGINEERS OF AIME

6200 North Central Expressway WRRSPE 3997

Dallas, Texas 75206

THIS IS A PREPRINT --- SUBJECT TO CORRECTION

Microscopic Behavior of Foam in Porous Media

By

R. F. Mast, Member AIME, Illinois Geological Survey

@ Copyright 1972
American Institute of Mining, Metallurgical, and Petroleum Engineers, Inc.

This paper was prepared for the 47th Annual Fall Meeting Of the Society of Petroleum Engineers
of AIME, to be held in San Antonio, Tex., Oct. 8-11, 1972. Permission to copy is restricted to an
abstract of not more than 300 words. Illustrations may not be copied. The abstract should contain
conspicuous acknowledgment of where and by whom the paper is pres~nted. Publication elsewhere after
publication in the JOURNAL OF PETROLEUM TECHNOIL3GYor the SOCIETY OF PETROLEUM ENGINEERS JOURNAL :s
usually granted upon request to the Editor of the appropriate journal provided agreement to give
proper credit is made.

Discussion of this paper is invited. Three copies of any discussion should be sent to the
Society of Petroleum Engineers office. Such discussion may be presented at the above meeting and
with the paper, may be considered for publication in one of the two SPE magazines.

ABSTRACT
———

hThen foam or gas was injected tnto etched of liquid and gas thus transported through the
glass micromodels saturated with a detergent porous media. Foam stability was increased by
solution, flow through portions of the porous increasing the concentration of detergent in the
system was blocked. As a result of blocking, solutions that were used to generate foam.
flow patterns were altered because the resis-
tance to flow in alternate directions through iNTRODUCTION
the porous network differed. Apparently, with
~i~.~, foam breakdown: foam drainage, and foam some of the mechanisms that operate when
regeneration altered the capillary resistance to foam is injected into porous media have been
displacement along any given flow path. revealed by microscopy. HOW foam bubbles are
formed or regenerated in a porous system, what
Foam drainage waa an important factor in factors control the size of the foam bubbles and
determining foam stability and drainage and the stability of foam in porous media, and how
stability together influenced the blocking mech- gas and liquid are transported through a porous
anism. When foam was present, liquid in the system in the presence of foaming agents were the
foam was held in the large pore openings. The particular questions investigated during this
liquid phase could drain from the foam in these Study .
pores and saturate the surrounding small pores
with liquid. Jamin action resulting from this Several papers have been published on the
distribution of fluids was primarily responsible behavior of foam in porous media in recent years.
~
IOL- LI --1.4 ..-
uluQKLL1~
+ha ● lot.,
l-,.== J..”,.. Foam. ~gaS regenerated in Much of this work has been concerned with prac-
porous media as a result of this distribution of tical uses of foam in oil and gas reservoirs.
fluids combined with the instability of liquid- From that point of view, foam has been considered
gas interfaces at pore openings. as an agent for selective plugging and oil recov-
ery. Interest, therefore, in the mechanics of
In all of the experiments, some liquid and two-phase (liquid and gas) flow in the presence
gaa were transported as foam, and as the stabil- of foaming agents has been generated.
ity of the foam increased, so did the proportion
Fried (1) studied the use of foam in oil
recovery. He described the flow of foam aa “plug
References and illustrations at end of paper.
2 MICROSCOPIC BEHAVIOR OF FOAM IN POROUS MEDIA SPE 3997

like,” where all shear takes place at flow chan- The porous model, similar to those described
nel boundaries. His microscope observations of by Davis (9), was made by etching a porous net-
foam in sand-packed glass tubes led him to de- work into a glass plate measuring 2 by 4 inches.
scribe the movement of foam through the porous Best results were achieved by splattering an
medium and the regeneration of foam within the acid-resistant material (KMER*) onto the plate
porous system. Marsden and Kahn (2) considered with a stiff brush. The outside quarter of an
foam flow through porous media to be dependent 01 inch around the top of the plate and the entire
the viscosity of the foam. On the other hand, back of the plate were then painted with KMER to
Bernard and Helm (3) calculated the effective protect them from the etching fluid. The plate
permeability of the medium to the different fluic was etched in a solution of hydrofluoric acid
phases (detergent solution and gas) by using the and phosphoric acid, which produced very clear
normal viscosities of these phases. Later etched areas. Figure 2a shows a Type I etched
Bernard, Helm, and Jacobs (4) showed that per- plate.
meability to liquid was a function only of the
liquid saturation. The underlying assumption in Basic construction of the model is shown in
these last two investigations, according to Helm, figure 2b. An etched plate was sandwiched be-
(5) is that foam does not move through a ------
p“L”Ua - tween two heavy glass plates and the entire assern
medium as a body, as in plug flow, but it sepa- bly was clamped together. The model was sealed
rates into phases within the porous system. by placing a very thin film of silicone stopcock
grease around the outside, unetched edge of the
Bond and Bernard (6) concluded that liquid etched plate. This technique prevented leakage
flow through a porous medium followed fixed chan- from the cell during the experiments. Care was
nels, whether or not foam was present, and that taken to ge? aofie“..C +k. ~W.a.Q
...=~.- .- ~nto che etched
these fixed channels depended solely on the liq- area of the model. An injection well and a pro-
uid saturation. They also stated that a negli- ducing well were drilled in the top glass plate,
~ible quantity of fluid could flow through the and glass tubes were cemented over the wells with
liquid membranes of the foam. Their interpreta- epoxy cement to facilitate connection of the cell
tion of liquid flow in the presence of foam was to the foam generator. The distance between the
an extension of Chatenever’s (7) idea that im- wells was 2.75 inches.
miscible fluids in a porous medium flowed in a
fi-XedChfiiliRl, A second type of pore structure (Type 11,
fig. 3) was etched into some of the giass plates.
Helm (5), from a series of tracer experi- These plates were made in the same way as those
ments and microscope observations of foam, con- described above (Type I) except that after the
~1.udedEhat gas and liquid flow separately initial etching they were again splattered with
through a porous medium. The liquid moves the acid-resistant material six?etched ~ secmd
~hroug.hthe film network, and the gas moves pro- time. This resulted in a porous system with
~ressively through the system by breaking and some pores etched more deeply into the glass than
reforming bubbles. Helm also described the re- others. In the Type 11 plates, the deep pores
generation and breaking of foam bubbles in the resulting from two etchings were separated from
large pore openings and discussed the influence one another by the fine pore structure that re-
~f detergent concentration on foam stability. sulted from the initial etching. Figure 3 shows
iandy (8) stated that gas tracer studies showed Type II models. Some typical dimensions of the
:hat a certain portion of the gas is trapped porous structure in both types of models appear
luring flow and does not participate in the flow in table 1.
>rocess.
As the flow behavior observed in the two
lX.PERIMENTALPROCEDURE types of porous models differed somewhat, the
microscope observations for both will be discussed
The experimental setup used in the present In all experiments a commercial foamer (Orvus K.)
studies is shown in figure 1. Foamer solution was used. In experiments in which foam was pro-
md gas were moved simultaneously through the duced from the foam generator and injected into
Foam generator and the foam produced was injected the model, the same concentration o.ffoamer in
lirectly into the observation cell. The deter- water was used to saturate the model as was used
gent solution was injected into the foam genera- in the generator.
:or with a syringe pump, and gas flow was con-
trolled by a pressure regulator. The foam gen- DESCRIPTION OF BEHAVIOR
:rator could be by-passed so that either gas or
Ietergent solution could be injected directly we —— I Model
into the cell. During the experiments, the pres-
;ure drop across the model was maintained at A. Gas Injected into Model Filled with Foamer
~bout 5 psi/ft. The observation cell was placed Solution
m.the stage of a microscope so that flow could
)e observed and photographed.
>kKodakMetal Etch Resist
SPE 3997 R. F. Mast 3

As gas was injected into a Type I model ini- structure within the porous medium. A signifi-
tially saturated with a foamer solution, membrane cant end effect at the producing well also was
~ere formed by the trapping of liquid between ad- noted—much of the produced foam was generated
joining advancing gas fingers. This was not an at the producing well.
efficient foam generating process and few foam
membranes were formed in this way. Once the fom The mechanism by which foam was generated
nembranes were formed, they began to move and sub at the producing well is shown in figure 5.
divide as they moved through the branching, poro~ Liquid from the produced foam in the well was
network (fig. 4). Most of the foam membranes imbibed into the pores surrounding the well
were stable and were transported through the (fig. 5ajo When the gas pressure near the weil
nociei. Some of tFR ineinbrzrles 11.,Fwnm
brc’ke,‘us.ua
..7 ..U.. ayen~rlndthe
-..----- ~ispiacernent pressure of the liquid
stretching as they moved from a narrow pore into filled pores, the liquid was forced toward the
a wide one. Varying the concentration of the well (fig. 5b). When the liquid passed the
foaming agent made no considerable difference in small constriction in the outlet pore, the cap-
the behavior described above. illary pressure drop across the liquid-gas inter
face decreased. At this moment, the gas finger
Foam Injected into Model Fiiied with Foamer ..- - ---...1* -s . . .
B. Jumpea rorwi?(rd, ~~ d L~SULL UL s-= e~p~fi~~e~.j
Solution and a small gas bubble formed at the end of the
gas finger (fig. 5c). Liquid was again imbibed
When foam, rather than gas, was injected into the pores surrounding the well and the
into the Type I model, which had been initially process was repeated. The foam membranes ad-
saturated with a foamer solution, the bubbles in vancing behind the finger generally would break
the injected foam were generally as large as, or because of the pressure shock (fig. 5d), making
larger than, the average pore size of the model. the texture of the produced foam quite uniform.
4s foam moved into the cell, liquid was displaced Although this mechanism was not important in
until the foam saturated almost the entire porous the generation of foam in the Type I model, it
system. As the system became saturated with is apparently similar to the mechanism that
Eoam, it was apparent that flow was taking place others have observed in sand-packed glass tubes
through only a part of the model and that flow in (1, 5). The mechanism will be discussed further
some of the pore channels was blocked. The block in the description of flow through the Type II
ing was temporary and flow patterns in the model model.
changed with time. However, most of the flow
took place in what appeared to be the larger In general, when foam was present in the
pore channels, which were seldom blocked by the Type I model, flow could be described as the
foam. Foam stability in these experiments was movement of liquid and gas through a partially
influenced by the concentration of the foamer blocked porous medium. When low concentrations
lsed. W-hen a soiution containing a high coiiceri- -1=E--------
01 -,.--+”
iUWULLl~il~=uk-‘were‘d~ed to prcduce the iD.-
tration (1%) of foamer was used to generate the jected foam, some gas was transported in very
Eoam, very little breaking of the foam membranes large bubbles that resulted from the zupture of
#as observed in the model. the foam membranes that separated small gas
bubbles. In some tests, liquid moved through
As the foam membranes were transported the foam, but liquid was also transported by
through the model, they subdivided in the manner the movement of the foam through the medium.
shown in figure 4. As a result, the foam bubble When high concentrations of foamers were used to
size decreased from the input well to the output generate the injected foam, very little rupturin
Jell. of the membranes was observed; in general, as th
stable foam flowed through the model, it sub-
When a solution containing a low concentra- divided (fig. 4) and the average diameter of the
tion (0.1%) of foamer was used to generate the foam bubbles became smaller than the average
Foam, considerable breaking of the foam membranes pore diameter. In this case, the flow of foam
~as observed, usually when the membranes were through the model became the primary mechanism
stretched. Films particularly broke in chains of transport of liquid and gas.
along the large pores. As the membranes broke
in this manner, very large bubbles were formed, Type 11 Model
nany of them several times larger than a single
Bore opening in the large pore channel. These A. Gas Injected into Model Filled with Foamer
large bubbles moved rapidly toward the producing Solution
~ell. As the leading edge of the large bubbles
>roke through into the producing well, the liquid Gas was injected slowly into the Type II
Saturation of the foam in the fine pore struc- model, which was saturated with a foamer solu-
tures adjacent to the large bubbles increased tion. The gas moved into the model through the
markedly. Foam from the adjacent fine pores fine pore structure. As the gas fingers reached
:hen broke into and filled the volume occupied by the large pores, they subdivided into very small
:he large bubble. Thus, under certain conditions gas bubbles that accumulated in the large pores
:he liquid phase could move through the foam (see fig. 3a). The foam generation process was
4 MICROSCOPIC BEHAVIOR OF FOAM IN POROUS MEDIA SPE 3997

similar to that shown in figure 5. As each small DISCUSSION

bubble formed, the gas finger retreated away from
the pore opening. Each time the gas finger ad- Influence of Capillary Forces on Foam Behavior
vanced to the large pore opening, a new bubble
was formed. When foam membranes are present, the geom-
etry of the interfaces is different from that
As small foam bubbles accumulated in the for single gas bubbles in capillary constrictions
large pores, the liquid in the pores was con-
tinually displaced (fig. 3b). As this process Gardescu (10) found that
continued, the liquid between the bubbles drained
away until only thin membranes separated the biib-
bles (fig. 3c). The reduction in the percentage (1)
of liquid saturation in the large pores continued
until the gas pressure in those pores equalled
the gas displacement pressure of the small pores wheqe P is the pressure required to hold a dis-
forming the exit to the large pores. At low con- torted gas bubble at static equilibrium in a
centrations of foamer (0.1%), the drained foam funnel-shaped capillary tube of circular cross
membranes were generally not stable in the large section, y is the interracial tension, and R1
pores and they broke, leaving a single large bub- and R2 are the radii of the oppositely curved
ble, or several large bubbles, occupying the interfaces (fig. 6a).
large pore (fig. 3c). High concentrations of
foaming agent produced more stable foam and, This equation can be applied equally well
although some bubbles broke, most were displaced to the liquid septum between two gas bubbles,
into the fine pore system. as shown in figure 6b. In that case, P is the
pressure required to prevent the fluid from mov-
AS gas injection continued, the stable foam ing further into the constriction. Equation 1
moved through the model until it reached the next predicts that as %- “ .aYp..,-----
an --””h-~ RI or as t (the
large pore, where it was regenerated in the man- thickness of the liquid septum) becomes small
ner described above. Again, blocking was noted, (fig. 6b), the magnitude of P decreases.
so that a large part of the porous network was
gas-saturated by the time gas reached the pro- In pure liquids, as t approaches zero, the
ducing well. septum is unstable and breaks. When foamers are
?resent in the liquid phase in sufficient con-
B. Foam Injected into Model Filled with Foamer centrations, stable membranes can exist in the
Solution ~apillary, as shown in figure 6c. To determine
the relation between the behavior of these mem-
As foam, rather than gas, was injected into branes and that of the individual gas bubbles
the Type II model , which was initially saturated separated by the liquid septa, the equilibrium
with a foamer solution, the injected foam bubbles configuration for the film in the capillary tube
were regenerated (fig. 3) as they displaced the nust first be determined.
liquid from the large pores. When low concen-
trations of foamer were used to generate the in- Figure 6C shows a stable membrane in a
jected foam, the foam that was generated inside funnel-shaped capillary that has a circular cross
the model was unstable. As foam injection con- section. Because the influence of gravity is
tinued, liquid accumulated in some of the small negligible, the surface of the membrane should
pores behind the gas front. Behind the gas front be spherical. At equilibrium, the pressure drop
flow was blocked in portions of the model, and the across the membrane, (P)m, is equal to 4_Y /I&,
flow patterns constantly changed. &lee the li~-
~here & is the radius of the membrane. Because
uid had been displaced from the large pores, no the pressure drop across the border (F)B equals
significant amount of foam was generated in the the pressure drop across the membrane, the border,
model by the process shown in Eigure 3. rW$len independent of its dimensions, can be represented
high concentrations (l%) of foamer were used to by extending the curved surface of the membrane
generate the injected foam, most of the gas bub- to the capillary walls. Because the membrane is
bles that were formed in the large pores were in teZSiQn3 the extended membrane must be normal
stable and were displaced as foam through the to the capillary walls for the system to be at
porous network. equilibrium. Therefore, Rm = R1’/sin a for the
geometry shown in figure 6c, and the capillary
After high gas saturations were obtained, force pulling the membrane toward the constric-
the method of liquid and gas transport through tion is given by
the Type 11 model was, in general, similar to
that described for the Type I model. sin a )/RT, (2)
(P)m= (4Y

?here RT is the radius of the capillary at the

oint where the extended membrane intersects the
f$p~ 3997 R. F. Mast 5

capillary wall. lther hand, in a porous medium in which the pore

tpenings are gradually tapered, the ❑inimum mem-
It can also be shown that as t, the thick- ~rane radius can be considerably larger than the
ness of the liquid septum, approaches zero (see :adius of the pore constriction.
fig. 6d), the capillary force pulling the liquid
toward the pore constriction is equal to the The relation between the radius of the pore
capillary force pulling a film (drawn through constriction and the minimum membrane radius in
the points of intersection of the two oppositely 1 packing of spheres can be roughly approximated
cu~ved interfaces) into the constriction. When }y using a torus as the pore model (fig. 7). For
t = O, the following relation can be developed :his pore model the ratio of the square of the
from the similar triangles shown in figure 4d: ~inimum membrane radius to the square of the con-
striction radius is
L RZ + R1
~ = RZ - RI (5)

Rearranging this equation so that the right-hand

side is equal to !his ratio, ~w22min, can be determined for a
E.&-
lacking of spheres if ~ and Rc in the torus
~odel are assumed to be equivalent to the grain
:adius and the associated constriction radius for
Iifferent packings of the spheres. This is simi-
.ar to the approach Melrose (11) used to deter-
and multiplying both sides of the resultant ~ine interracial configurations in sphere packine
equation by the constant 2 y gives Ie gave the following equation for the ratio of
:he grain radius to the pore constriction radius
?or regular packings of spheres:
(3)

(6)
The left-hand side of (3) is equivalent to the
Laplace equation for the pressure drop across
curved soap bubbles, if the film radius
~or a cubic packing of spheres in which the por-
RF =L+R1.
>sity is 47.64%, the angle G = 45 , and for a
:hombohedral packing in wh~ch the porosity is
Further drainage of the septum leads to !5.95%, the angle ~ =30 (11). Solving equa-
the formation of a membrane with its associated :ions (5) and (6) for the minimum membrane radius
Plateau border. The membrane geometry in the ~n terms of the Dore constriction radius for the
funnel-shaped portion of the tube does not depenc :ubic array gives (Rm)min = 2.41 Rc, and for the
on the dimensions of the border. Consequently, :hombohedral array it gives (&)min = 3.73 Rc.
the capillary force pulling the liquid toward rhese solutions assume that the membrane config-
the constriction should also be independent of uration in the torus model is equivalent to the
the foam quality as long as a membrane has nembrane configuration in sphere packings.
formed.
The relation between pore morphology and min.
The minimum pressure required to force a i.mimmembrane radius may be used to explain dif-
foam membrane out of the circular capillary of ferences in foam behavior in two mediawith similar
radius Rc and into the funnel-shaped capillary ?ermeabilities. In one net~lork where the pore
should be ]penings are sharp ( ~ > 70 ), resistance to
Foam displacement as a result of capillary action
(P)min : 4y sin a IRC. (4) should be great. (h the other hand, in the other
media where changes in pore size are gradual, as
in packing of spheres, capillary resistance to
Equation 4 is approximate because the surface of Eoam displacement should be low.
a funnel-shaped pore, parallel to the direction
of flow, is not mathematically smooth, and the Foam Drainage and Stability
border does have dimensions. Consequently, the
minimum membrane radium (~)min should develop Foam drainage in porous media is an impor-
just outside the constriction and will be only tant factor in determining foam stability; foam
approximately equal to R= /sin a , as shown in Irainage, and foam stability in turn, have an in-
figure 6e. In rocks where the pore openings are, fluence on the blocking mechanism. When foam is
on the average, sharp (e.g., a > 70°) the present in a porous solid, the liquid can be held
minimum membrane radius will be nearly equal to as foam in the large pore openings. However, the
the radius of the pore constriction. On the foam itself is a porous structure in which the
 —-L. ---- -- --+.J -- in tb.e~~tera~ure there has been no
Plateau borders make a 3-dimensional neLwuLL. LCPULu=” “..

through which the liquid phase can move. Because effort to relate foam drainage to flow behavior.
of capillary action, the liquid phase has a ten- It is possible that control of the factors in-
dency to drain from the foam in the large pores fluencing foam drainage may improve our ability
into the surrounding small pores. This process to control and predict the flow behavior of foam
gill continue until the pressure in the liquid in porous media,
phase in the borders of the foam is equal to the
pressure in the liquid phase in the small pores. Foam Generation
kt that point the pressure in the gas phase
should be equal to the gas entry pressure of the The segregation of the liquid into the smal
., ~are~ ~T,dthe ~ZS in-toChe large pores as a re-
small pores surrounding t.neiarge pore opertiags.
rhis foam drainage process is equivalent to the sult of foam drainage sets the stage for the foal
“separation of the two phases, liquid and gaa,” generation process. In general, while a liquid
described by Helm (5). is being displaced by a gas through a porous
solid ~ the liquid-gas interface is stable as
The size of the gas bubbles left in the long as the interracial curvature remains con-
large pores depends on the stability of the foam stant or increases (11). Once the interface
films after they have been completely drained in recedes to a point where, because of the boundar
this manner. Bubble size is also influenced by conditions, further displacement results in a
pore geometry. This can be illustrated by ana- decrease in the curvature, the interface becomes
lyzing the differences in bubble size of the unstable and advances rapidly or jumps to a new
foam noted in the Type I and Type II models. position (figs. 3 and 5). In this study, these
rhese bubble size differences were especially changes in the interface were so rapid that the
noticeable in experiments in which foam generated shape of the interface could not be observed in
from low concentrations of foamers was injected the model. Only capillary forces were involved
into both models. The size of the large gas bub- in the formation of the new bubbles; gravity
bles in the Type II model had dimensions of the was not a factor in these experiments because
~rder of the large pore dimensions. In contrast, the model was horizontal. It is believed that
the large gas bubbles in the Type I models had the interface becomes overextended (essentially
dimensions that were much greater than the di- U-shaped) as a result of the rapid advance asso-
mensions of the pore openings. These variations ciated with the reduction in the capillary pres-
in bubble size can be explained by the differ- sure. Consequently, a catenoidal or waisted
ences in pore dimensions in the two model types surface (12) forms immediately behind the spheri
(table 1) and the effect of pore size on capil- cal nose of the gas finger and pinches off a
1 . .. $,.-,-a
L=ry .U.b=s. small gas bubble as the interface seeks a new
equilibrium configuration. This bubbling pro-
In Type I model, all of the pores had the cess carries the liquid from the small pores
. . ..A +ha
same over-all depth, -Liu . ..% A
“ey.nt~ . ~~Ee~.s~on WaS in.t~Ehe large pores as foam. Once again the
considerably smaller than the size of the so- foam drains, and the entire process is repeated.
called pore constrictions (table 1). As a result
the minimum principal radius of curw-atiire
across Fcmm Blocking
an interface was constant, because the pore
depth was constant. As the second principal In the model, flow became temporarily
radius of curvature was controlled by the hori- blocked in various portions of the porous net-
zontal pore dimensions, and as these dimensions work. This phenomenon can be explained simply
were large, the capillary forces available to in terms of the capillary resistance (10), which
hold the liquid in the pore constrictions were resulted from the phase distribution in the pore
really very small. Therefore, very large gas —gas in the large pores and liquid in the inter
bubbles were stable in the Type I models. connecting small pores. This phase distribution
was controlled by capillary action, and it
The relations between horizontal pore di- occurred because the foam structures provided
mension and vertical pore dimension in the Type I liquid continuity between the large and the
nodel were similar to those described above for small pores.
the Type I model. However, in the Type II model,
the pore depths also varied (table 1) so that the Flow patterns, as a result of blocking,
minimum principal radius of curvature of the changed because the capillary resistance to flov
interface varied. Because the pore depth dimen- in alternate directions through the porous net-
sions were small, the liquid was held by capil- work varied. Furthermore, foam drainage, foam
lary forces in the shallow pores separating the breakdown, and foam generation altered with time
deep pores, and the gas bubbles had dimensions the capiiiary resistance to flew cf both liquid
on the order of the dimensions of the large pores and gas along any given path. As a result, liq-
uid and gas flow patterns were continually charq
Foam drainage appears to be an important -*-2- J=’1.. L-.L.W;fiy
ing when foam was present. LfllS LL~w UCUavAv.

factor in determining the behavior of foam in is different from that observed by Chatenever
porous media. In the flow experiments run and (7) in steady-state flow of liquid and gas
SPE 3997 R. F. Mast 7
ic~ ~c~.p~~~~~y w~~ ~ ~~ntact-angle hysteresis
without foaming agents. lie found that flow pat-
terns were maintained through essentially fixed ?ill increase as flow rate increases. However,
liquid and gas channels. In this case, capil- it might be postulated, on the basis of the
lary forces, at liquid and gas interfaces in hcrease in film thickness with flow rate found
small pores between the liquid and gas channels, )y Marchessault and Mason, that when foam is
effectively stabilized the channels. Elowing over slightly contaminated surfaces,
mechanical spreading would cover the surfaces
One of the apparent deficiencies of the ~ith liquid. If this were the case, there would
models that were used in this study is that >e no three-phase angle of contact and flow rate
they allowed only one-dimensional flow. Because ?ffects on contact-angle hysteresis would not be
foam drainage is a flow process, it, too, was >bserved.
essentially restricted to one dimension in the
model. This should not seriously have affected David (15) found that the apparent viscosit~
the observed results; in porous media, where >f foam in giass capillaries cieerezsecl~s tble
the porous system is continuous in three dimen- shear rate increased. Because the proportion of
sions, foam blocking and regeneration may be the total fluid transported as foam was found in
even more important than it was in the models. this study to be dependent on foam stability, th~
importance of apparent foam viscosity in deter-
Other Factors Influencing Foam Behavior mining flow behavior should increase as the sta-
5ility of foam increases.
Two-phase flow in rocks is generally in-
fluenced by the “nettability” of the rocks. Rapid stretching of membranes increases
Nettability effects are considered to result their interracial tension by momentarily de-
from Jamin action, in which the displacement of creasing the concentration of the surface-active
a chain of bubbles is resisted by the hysteresis naterial at the surface of the membranes (16);
of contact angles (10). Bond and Bernard (6) the reverse is also true. This phenomenon is
found marked differences between the behavior rate dependent, because molecular diffusion
of foam in water-wet and oil-wet (silicone restores equilibrium between the surfactant in
treated) sands. For example, in the presence of the bulk liquid and that adsorbed at the inter-
water or a very dilute foaming agent, oil-wet face. As rate influences surface tension, there
sand was more permeable to gas than water-wet nay be a hysteresis effect. It was observed in
sand. With more concentrated solutions of foam- this study that the foam membranes do not move
ing agent, oil-wet sand was less permeable to in and out of the pore constrictions at the same
gas than water-wet sand. They also conducted rate. This phenomenon may also help explain the
experiments in straight oil-wet capillary tubes temporary nature of the blocking that was seen
and showed that the capillary resistance to the in the models. Because of the flow-rate depen-
displacement of a chain of bubbles decreased, dency, a cessation of flow could result in a re-
up to a point, as the concentration of foaming duction in surface tension and, therefore, a
agent increased. Consequently, they observed reduction in the capillary resistance to renewed
significant capillary resistance to the dis- movement.
placement of a chain of bubbles at foamer con-
centrations as high as 1 percent. They there- SUMMARY
fore explained the plugging of oil-wet sand by
foam as being partly due to Jamin action. The evidence collected during this study,
combined with the results from earlier investi-
In flow experiments on gas and liquid in a gations, has produced the following description
capillary tube, Marchessault and Mason (13) of foam behavior in porous media.
s’hewed tkt ~hil th~
-...-G..-o
DULLCLLG
-~ tb,e +,,hn wsc
WA .U-- ----

completely wetted by the liquid, the entrained 1. In almost all cases, some liquid and gas are
gas bubbles were separated from the solid sur- transported as foam through the porous media
face by a liquid membrane. Thus, as no three- Of the total fluid transported, the propor-
phase contact angle existed in the tube, con- tions of liquid and gas that are thus moved
tact-angle hysteresis could not explain the are a function of the stability of the foam
observed resistance to the displacement of gas in the porous media.
bubbles in water, through the water-wet tubes.
Marchessault and Mason did find that the bubble 2. When foam is unstable in porous media, trans
flow-resistance factor for gas bubbles in water port of liquid and gas occurs primarily by
decreased as the velocity of the bubble in- the progressive breaking and regeneration of
creased. They also found that the thickness the foam structures between large pores that
of the liquid membrane between the capillary are separated by small pores. No iiquid
wall and the gas bubble increased as the rate channels were observed, but some of the liq-
of displacement increased. uid transport is through the Plateau borders
connecting local areas of high liquid sat-
The data of Elliott and Riddiford (14) uration, both in the foam and adjacent to
support the idea that, if the solid surface is the solid surfaces.
8 MICROSCOPIC BEHAVIOR OF FOAM IN POROUS MEDIA SPE 3997

3. When foam is stable in porous media, the ogy of foams in porous media, Presented at
liquid and gas are transported primarily as the SPE-AIChE Joint Symposium, AIChE 58th
foam. Some breaking and regeneration occur, Ann. Meeting, Dallas, Texas, Feb. 7-10.
and the bubble size of the foam is also in-
fluenced by foam subdivision. 7. Chatenever, A., 1957, Microscopic behavior
of fluids in porous systems, Final Compre-
4. Flow through portions of the porous network hensive Reports, American Petroleum Insti-
is temporarily blocked by the foam. This is tute.
equivalent to a reduction of the single phase
permeability of the medium. Temporary block- 8. Handy, L. L., 1971, Non-Darcy flow in porou
ing is the result of the distribution of media, 15th Ann. ACS Petrol. Res. Fund Rep.
liquid and gas in the porous system, netta- Res.., pp. 156-157.
bility effects, and the influence of flow
rate on surface tension. 9. Davis, J. A., and S. C. Jones, 1968, Dis-
placement mechanisms of micellar solutions,
5. Foam drainage has a strong influence on foam Jour. Petroleum Technology, Vol. 20,
stability and appears to be an important pp. 1415-28.
factor in determining the behavior of foam
in porous media. 10. Gardescu, I. I., 1930, Behavior of gas
bubbles in capillary spaces, AIME Trans.,
Petroleum Development and Technology,
REFERENCES vol. 86, p. 351-370.

1. Fried, A. N., 1961, The foam drive process 11. Melrose, J. C., 1965, Nettability as relate
for increasing the recovery of oil, U. S. to capillary action in porous media, Sot.
Bureau of Mines Rept. Inv. 5866, 65 p. Pet. Eng. Jour., Vol. 5, No. 3, p. 259-271.

2. .,-. .r.,-l c The

rmfsdeil,S. s., -...4 “. -A. Kahn, 1966, 12. Briggs, L. J., 1958, Capillarity, the
flow of foam through short porous media and Encyclopedia Americana, Americana Corpora-
apparent viscosity measurements, Sot. Pet. tion, New York, Vol. 5, pp. 554-558.
Eng. Jour., Vol. 6, No. 1, p. 17=25.
13. Marchessault, R. N., and S. G. Mason, 1960,
3. Bernard, G. G., and L. W. Helm, 1964, Effect Flow of entrapped ‘hubbies throiigha capi~-
of foam on permeability of porous media to lary, Industrial and Engineering Chemistry,
gas, Sc?c. pet, I@= Jour., VO1. 4, No. 3, Vol. 52, No. 1, p. 79-84.
pp. 267-274.
14. Elliott, G. E. P., and A. C. Riddiford,
4. Bernard, G. G., L. W. Helm, and W. L. Jacobs, 1967, Dynamic contact angles - 1. The
1965, Effect of foam on trapped gas satura- effect of impressed motion, Colloid and
tion and on permeability of porous media to Interface Science Jour., Vol. 23,
water, Sot. Pet. Eng. Jour., Vol. 5, No. 4, p. 389-398.
pp. 2!35-300.
15. David, A., 1968, The rheology of foam,
5. Helm, L. W., 1968, Tunemec’nanismof gas and Ph.D, thesis: Stanford University, Palo
liquid flow through porous media in the Alto, California, 97 p.
presence of foam, Sot. Pet. Eng. Jour.,
vol. 8, No. 4, pp. 359-369. -c
La . L-4
..hAmav , Q1o ---
&&LL...=..-. A.. , and C. F. (Moper, 1959,
Current concepts in the theory of foaming,
5. Bond, D. C., and G. G.,Bernard, 1966, Rheol- Quarterly Reviews, Vol. 13, PP. 71-97.
 l?ig.2A - Type I model.

b
Well connection

Etched glass
/
Glass fop plate Glass b’ase plate

Fig. 2F3- Model construction.

“m

Fig. 3 - Foam generation in Type II model.

 TABLE 1 - AVERAGE PORE DIMENSIONS IN MILLIlfETERSFOR ETCHED POROUS PLATES
..—— ——.
. ...——— —- 1 Depth
CL-II.... Deep
~’o:e _~r~_w&s...
openings
~~e and number 7=... ‘1’”
--
1.5 0.01
I, 3 0.3
0.02 --
0.6 2.4
1, 6
n nl
“.”. 0.03
0.5 i.O
II, 4-8
0.02 0.06
0.6 1.8
II, 6-10

—— __. —-.

~ Air supply

Syringe
pump
Y , Pressure
regulator

-’-l!E5 :utpuLwel,
1
[i 1
observation
cell

Fig. 1 - Experimental setup.

 :.,;
.,
A. B. c.

t t

[~$’”’was
.’:

h
,. ‘. Foam membranes
.. :.’
... ,.
,...”,
.. .. .. .. .. . . “,
..”..’,. . . .“’. .. .. ..
..
... . ,.,
. .. . . ..
. .. . . .. . . ‘...
,., ,:. . .,

‘“*o.mmembri~’
..,, ,. ’.” .“
$.K-L) ;., ”
,.. . ., ::,
..
... .’.:
:. .
..’. :.:
M .,. ::

t t t

Fig. b - sequence of diagrams showing foam subdivision in

branching porous network.

a. b.

Well Gas bubble

Gas-filled pores

d.

Fig. !i- Foam generation at the producing well.

 septum Plateau

D P
R2 border— Rt ~a
R2 RI -.
Rl member= Rc
gas t Rm
liquid * extended membrane

a c

‘\ ‘7brone/’
~ \
\
%Pia,eau ~or,er

Rc

/
/
d e
Fig. 6 - Graphical definition of terms used in equations describing behavior of
membranes in funnel-shaped capillary tubes.

Fig. 7 - Cross section of torus showing definition

of symbols used in Eq. 5.