Académique Documents
Professionnel Documents
Culture Documents
~
I. DRAINAGE
24 2. SPONTANEOUS IMBIBITION
3. FORCED IMBIBITION 48
...
:J:
....0 16 40
co
E J:
'->
....0 VENANGO CORE
W-
II::
8 E 32 k=28.2 md
:::::l (.)
Vl
Vl I .Q.!BYE.
ILl
II::
w I. DRAINAGE
ll.
a::
::J 2. SPONTANEOUS IMBIBITION
0 en
)0-
Il::
en 24
<t w
...J a::
...J 0...
ll.
<t BEREA CORE >-
'-> -8 k = 184.3 md a:: 16
<t
...J
.EQJNI ...J
A. IRREDUCIBLE WETTING SATURATION 0:::
-16 B. ZERO-CAPILLARY - PRESSURE <t 8
(.)
NONWETTING SATURATION
C. IRREDUCIBLE NONWETTING SATURATION
-24
0 100
WATER SATURATION, PERCENT P.V.
Fig. 2-0il/water capillary pressure curve measured on a Fig. 3-0il/water capillary pressure curve measured on a
water-wet Berea core. 9 strongly water-wet Venago core. 9
Curve 2 of Fig. 2 is the spontaneous imbibition curve, ration. In general, different residual nonwetting-phase
determined after the drainage capillary pressure curve is saturations will be measured after the spontaneous and
measured. The capillary pressure, initially at a large posi- forced imbibition curves unless the core is very strongly
tive value, is gradually decreased to zero, allowing the wetted. 7,12
wetting phase (water) to imbibe. The nonwetting residu-
al saturation reached when Pc =0 will be referred to as Capillary Pressure Curves in Strongly
the zero-capillary-pressure nonwetting saturation. In Wetted Systems
general, some of the oil (nonwetting phase) is still con- The capillary-pressure/saturation relationship is a func-
nected when the capillary pressure is zero, so the tion of wettability, pore structure, and saturation histo-
non wetting-phase residual saturation defined in this ry. Before discussing how changes in wettability affect
fashion is not irreducible. capillary pressure, we will examine capillary pressure
Curve 3 of Fig. 2 is the forced imbibition curve, ~here curves measured on strongly water-wet and strongly oil-
the capillary pressure, Po-Pw, is decreased from zero wet systems. This discussion will be based on early work
to a large negative value. When the capillary pressure is at Pennsylvania State U. 9,13-17 Figs. 2 through 5, taken
negative, the pressure in the wetting phase (water) is great- from Killins et aI., 9 show capillary pressure curves
er than the pressure in the nonwetting phase (oil), forc- measured on strongly wetted systems using a porous-plate
ing water into the core. (Note that the fact that the water apparatus. Figs. 2 and 3 are for strongly water-wet plugs,
pressure is greater than the oil pressure does not imply while Figs. 4 and 5 are for strongly oil-wet ones. All of
that the oil is the wetting fluid at these saturations. If the the plugs were cleaned with toluene, dried, then saturat-
core were a bundle of cylindrical capillary tubes, then ed with their wetting fluid (water for Figs. 2 and 3, oil
negative capillary pressures would be possible only if the for Figs. 4 and 5). Capillary pressures were measured
core were oil-wet. However, the interaction of pore struc- with a refined oil and brine.
ture and wettability allows negative capillary pressures Fig. 2 is the capillary pressure measure~ on a water-
even for strongly water-wet cores. II) Much of the oil (the wet Berea plug, while Fig. 3 is for a strongly water-wet
nonwetting phase) is still connected at the end of the spon- Venago plug. The strong water wetness of the plug shown
taneous imbibition curve when the capillary pressure is in Fig. 2 causes the areas under the drainage (Curve 1)
zero because additional oil is produced as the capillary and forced imbibition (Curve 3) curves to differ signifi-
pressure becomes negative. The oil saturation decreases, cantly. Because of the favorable free energy change,lit-
and the oil gradually disconnects as the capillary pres- tle or no work must be done during imbibition, when the
sure becomes increasingly negative until the capillary preferentially wetting fluid (water) displaces the nonwet-
pressure curve is almost vertical. This saturation, where ting fluid (oil). Conversely, a great deal of work is re-
hydraulic continuity of the nonwetting phase is lost, will quired during drainage when the non wetting fluid
be referred to as the irreducible non wetting-phase satu- displaces the wetting one from the core. It can be shown
Journal of Petroleum Technology, October 1987 1285
24
.c.J.!frlE. -48
I. DRAINAGE
2. SPONTANEOUS IMBIBITION
16 3. FORCED IMBIBITION
4. SECONDARY DRAINAGE
t -40
TENSLEEP CORE
-
CI'
3
'"
:x: 8
:I:
k=48.1 md
~ 0
E E
<.> 4 u -32 .cJ.mYE.
I I. DRAINAGE
UJ 2. SPONTANEOUS IMBIBITION
a: w
::l Q::
Vl :::>
fil IF)
-24
a: IF)
n. w
Q::
>-
a: a..
<t
..J >-
Q::
::! «.J -16
n.
<t
<.> .J
-16
DRIFILM - TREATED a..
BEREA CORE
«u
k=84.2 md -8
-24
00 80 100
-32 OIL SATURATION - PERCENT PV
0 20 40 60 80 100
WATER SATURATION-PERCENT PV
Fig. 5-0il/water capillary pressure curve measured on a
strongly oil-wet Tensleep sandstone core. Note that the
Fig. 4-0illwater capillary pressure curve measured on an curves are plotted VS. 011 saturation and with negative
oil-wet Berea core treated with Drifilm.9 capillary pressures upward. 9
that the work required for one fluid to displace the other area under the imbibition curve (Eq. 6), however, is much
from the core is related to the area under the capillary smaller than the area under the drainage curve (Eq. 5),
pressure curve. For oil displacing water (drainage curve so more work is necessary for oil to displace water than
in a water-wet core, Curve 1), the external work required for the reverse displacement. This shows the degree of
is 18,19 water wetness.
Figs. 4 and 5 show the capillary pressure measured in
strongly oil-wet systems. The Berea plug shown in Fig.
4 was treated with DrifilmTM, an organochlorosilane, to
render it strongly oil-wet. 1 The Tensleep sandstone plug
shown in Fig. 5 was naturally strongly oil-wet even after
where it was cleaned with toluene and dried. The plugs were
Vb = bulk volume of the core, first saturated with oil, then the drainage capillary pres-
cf> = porosity, and sure curve (Curve 1 in Figs. 4 and 5) was measured by
Sw = water saturation. increasing the capillary pressure, Po -Pw, to a large,
negative value. Because oil is the wetting fluid, a large
Similarly, the work required for water to displace oil amount of work is necessary to force water into the core,
(forced imbibition curve for a water-wet core) is as shown by the large area under Curve 1 (see Eq. 6).
When oil is the strongly wetting fluid, the roles of oil
and water are reversed from the strongly water-wet case.
To demonstrate this, Fig. 5 for the oil-wet Tensleep plug
has been inverted froin a normal capillary pressure plot.
The curve is plotted vs. oil saturation and with negative
The area under the drainage capillary pressure curve capillary pressure plotted upward. Note the very strong
in Figs. 2 and 3 is relatively large because a great deal resemblance to Fig. 3 for the water-wet Venago plug.
of work is necessary for the oil to displace the water. The Similarly, if Fig. 4 for the oil-wet Berea plug were in-
Venago core is very strongly water-wet because a great verted, it would resemble the capillary pressure curve
deal of the water imbibes back spontaneously and the shown in Fig. 2 for the water-wet Berea plug.
residual oil saturation is approached at zero capillary pres- Curve 2 in Figs. 4 and 5 is the spontaneous imbibition
sure. The water saturation after spontaneous imbibition curve, measured as the capillary pressure is reduced to
is almost 80% PV. The Berea core shown in Fig. 2 is zero. Because oil is the wetting fluid, it imbibes spon-
less strongly water-wet. After spontaneous imbibition, the taneously, just as water imbibed in the water-wet plugs.
water saturation is only about 55 % PV, and additional Curve 3 in Fig. 4 is the forced imbibition curve. The capil-
water can be forced into the core (Curve 3, Fig. 2). The lary pressure, Po -Pw, is increased to a large positive
N.
f
"~
•
~
• DIOCTYI. ETHER
• HEXACHLOAO-I,3-
BUTADIENE
600
1/1 ". BROMONAPHTHALENE 73 °
(;s ETHYLENE GLYCOL 2~
'1'
I8°
43°
of experiments with uniform wettability, the wettability
of the entire core is varied from water-wet to oil-wet. At
"'Q 2.0 r- JZl WATER lvo' 79° any given wettability, the wettability of the entire surface
Jlf"" • •i is kept as uniform as possible. Additional wettability ef-
~f-1.0-
JZl
prpr
tl"
ptll IJ'
f,~
tl{lf.{If~~
fects will occur if the core has fractional or mixed wetta-
bility, where some of the rock surfaces are strongly
water-wet but the remainder are strongly oil-wet. In this
second set of experiments, the effects of wettability are
JZl pr ~ {If
JZl [if tl studied by varying the proportion of the surfaces that are
17 pr pr !8 pr water-wet vs. oil-wet.
70~----~------~----~------~----~ 120
a::
w
60 ~ 100
~
u..
0
a:: \
w ::E
u
~ 50 I
80 \
~
'E
W
a:: \
~
U
E en
en \
40 w 60 \
I a::
n.
~ \
.~ >-
~
w
0::
n.
30
a::
<[
...J
...J
ii:
<[
40 "
"- .....\ '
',,: ~5
>-
a:: u ...............................
<[ FINES - , ........
...J
...J 20 20
TREATED ......... ...
ii:
<[
u
10 °0~----~------~----~~----~-----J100·
WATER SATURATION-PERCENT PV
The capillary pressure and contact-angle measurements bundle, so his results should not be compared directly with
by Morrow and Mungan,28 Morrow, 29,30 and Morrow the apparent contact angles calculated by Morrow and
and McCaffery 50 found that the receding contact angle Mungan. The important result of his work is that the ap-
measured on a rough teflon surface characterizes the wet- parent contact angle for smooth and rough sphere packs
tability during drainage capillary pressure measurements differs.
in teflon cores. Similarly, the advancing contact angle
characterizes imbibition capillary pressure measurements. Fractional and Mixed-Wet Systems. In the experiments
This indicates that departure of the pore geometry from discussed previously, the wettability of the core was var-
a bundle of capillary tubes had little effect on the interac- ied while the wettability of the entire surface was kept
tion of pore geometry and wettability to determine capil- as uniform as possible. For example, all of the rock sur-
lary pressure. Morrow and McCaffery 50 point out, face in a neutrally wet core should have little preference
however, that this result should be regarded as fortuitous for oil or water. Additional wettability effects will occur
and may not extend to capillary pressure measurements when the system has nonuniform wettability (either frac-
in reservoir rock. Note also that these results do not im- tional or mixed), where portions of the surface are strong-
ply that the porous medium can be modeled as a bundle ly water-wet while the remainder are strongly oil-wet. 1,54
of capillary tubes. Imbibition results, discussed in more Salathiel 55 introduced the term "mixed" wettability for
detail below, show that the bundle-of-capillary-tubes a special type of fractional wettability in which the oil-
model is not valid for imbibition in teflon cores. 50-52 wet surfaces form continuous paths through the larger
Kruyer 53 developed a model for the effects of contact pores. The smaller pores remain water-wet, containing
angle on imbibition capillary pressure in pac kings of uni- no oil. Note that the main distinction between mixed and
form spheres and applied this model to mercury imbibi- fractional wettability is that the latter does not imply either
tion capillary pressure. He found experimentally that the specific locations for the oil-wet and water-wet surfaces
best fit for glass spheres and smooth metal balls was 150 0 or continuous oil-wet paths.
[2.62 rad] through the mercury. However, some of the Fatt and Klikoff 56 used the porous-plate method to
metal spheres were accidentally corroded (roughened). measure capillary pressure of fractionally wetted sand-
The best fit for the apparent contact angle through the mer- packs. The fractionally wetted sandpacks were formed by
0
cury for these rough spheres was 180 [3.14 rad]. Note mixing treated and untreated sand grains together. The
that Kruyer did not model his system as a capillary tube untreated sand grains were strongly water-wet. The re-
Journal of Petroleum Technology, October 1987 1291
nificantly different from that of the sandpack where 50%
16r----------------------------------, of all grain sizes were treated to render them oil-wet. At
o NATIVE STATE low capillary pressure, the fines-treated curve lies below
A CLEANED
the uniformly treated one. As the capillary pressure is in-
14 creased, the fines-treated curve becomes nearly vertical
CORE DATA at a much higher water saturation, demonstrating that both
EAST TEXAS FIELD the location and fraction of the preferentially water- and
WOODBINE RESERVOIR
12 3696.5 FEET DEEP
oil-wet surfaces are important.
</>= 2r.B % Several researchers have reported behavior similar to
k = 648 md
that of the fines-treated pack when comparing capillary
pressure behavior in native-state, mixed-wet plugs vs. the
10 same plugs when cleaned and rendered water-wet. 59-63
'"
Q.
An example is shown in Fig. 9, taken from Richardson.
et al. ,60 which compares the capillary pressure measured
on a native-state plug from the east Texas Woodbine reser-
voir with the capillary pressure measured on the same plug
after it was cleaned and rendered water-wet. Native-state
Woodbine core was later shown by Salathiel 55 to have
mixed wettability, with large oil-wet pores and small
water-wet ones. Similar behavior for cleaned vs. native-
state core can be found in Riihl et at., 61,62 Schmid,63 and
4 Luffel and Randall. 59
After the native-state curve shown in Fig. 9 was meas-
ured, the plug was cleaned and dried. It was then saturat-
ed with brine, and a capillary pressure curve starting from
2
a 100% brine saturation was measured. The behavior of
this plug in the native vs. cleaned states is very different
from the behavior of neutrally or fractionally wetted plugs
o 100
shown in Figs. 6 and 7. At low capillary pressures, the
WATER SATURATION, % PV native-state capillary pressure curve in Fig. 9 is lower than
the cleaned water-wet one. It then crosses over with a
Fig. 9-Comparison of capillary pressure curves measured higher IWS, which is in better agreement with the IWS
on a single core in the native and cleaned states. The core measured by Qilflooding a native-state core. 55,60 In con-
is mixed-wet in the native state, and water-wet after
cleaning. 60 trast, the IWS for a neutral or fractionally wet core is the
same as or slightly lower than the IWS when the core is
water-wet.
Schmid 63 pointed out that mixed wettability is respon-
maining sand grains were treated with Drifilm to render sible for the capillary pressure behavior of these cores.
them oil-wet. Note that during mixing some Drifilm may At the beginning of the capillary pressure measurement
have been transferred to some of the water-wet sand in the mixed-wettability (native-state) plug, oil enters the
grains, probably giving them a nonzero contact angle. 56 large oil-wet pores. A lower capillary pressure is required
The sandpacks were placed in the porous-plate appara- to displace the water from the large pores when they are
tus, then saturated with water. Kerosene was used to meas- oil-wet vs. water-wet, so the capillary pressure curve is
ure the capillary pressure for oil displacing water. Fig. initially below that of the cleaned water-wet core. Dur-
7 shows the capillary pressure curves measured on two ing this time, some of the water in the small pores is
sets of sandpacks with relatively narrow grain size dis- bypassed and trapped. Eventually, most of the water in
tributions (either Tyler mesh -28 +35 or -65 + 100). the large oil-wet pores is displaced, and oil begins to enter
As the fraction of oil-wet sand increases, the area under the remaining smaller pores, which are water-wet and
the curve decreases, indicating that it is easier for the oil filled with water. At this point, the capillary pressure for
to displace the water. Because of the relatively narrow the mixed-wettability core crosses over the cleaned curve
grain- and pore-size distribution, all the curves are fairly and begins to rise rapidly. This occurs for two reasons.
flat until IWS is reached. Similar results were reported First, a higher capillary pressure is required to force oil
by Talash and Crawford. 57,58 into the smaller water-wet pores. Second, the IWS will
Fig. 8 shows the capillary pressure curve measured on be relatively high because some of the water in the small
a fractionally wet sandpack with a much wider grain- and water-wet pores will have a tendency to be bypassed and
pore-size distribution. 56 Again, the area under the curve trapped as oil flows in the larger pores.
decreases when the fraction of oil-wet grains is increased
from 0 to 50 %. Fatt and Klikoff also demonstrated the
importance of the location of the oil-wet vs. water-wet Effects of Surface Roughness on
surfaces. The curve marked "Fines Treated" in Fig. 8 Apparent Contact Angle
shows the capillary pressure curve when only the smallest- The experiments discussed above show that in uniformly
sized sand grains were treated with Drifilm to render them wetted cores, the capillary pressure is insensitive to the
oil-wet (the smallest 58 wt% of sand grains were treat- contact angle for a surprisingly large range of contact an-
ed). The behavior of this fines-treated sandpack is sig- gles. Imbibition results are the same as the zero-c~ntact-
1292 Journal of Petroleum Technology, October 1987
angle case when the contact angle is less than about 22 °
[0.38 rad], 28,29 and drainage results are the same when
the contact angle is less 'than about 50° [0.87
rad].25,26,28,29,41,48 At larger contact angles, there is an
effect of contact angle on wettability, but it is not the co-
sine relationship predicted by the capillary tube bundle
model.
There appear to be several reasons why capillary pres-
sure, particularly drainage capillary pressure, is relatively
insensitive to contact angle. The first reason is that use (0)
of a cos () factor assumes that the porous medium can be
modeled as a bundle of circular capillary tubes. Depar-
ture of the pore geometry from this assumption, however,
will alter the effects of contact angle on capillary pres-
sure. 64 ,65 Branches, side pore mouths, and void spaces
can greatly affect the interface shape and the apparent con-
tact angle. 48 ,66 Second, reservoir rocks contain a large
number of sharp edges. On a smooth surface, the contact
angle is fixed; however, on a sharp edge, this condition
is relaxed, and there is a wide range of permissible con-
tact angles. 19,28,67 The angle that a fluidlfluid interface
makes at such a sharp edge depends on fluid saturations
and the geometry over the entire pore, rather than just
local surface conditions. 68 Morrow 19 postulates that
many of the oil/water/rock contact lines in a core will be
located at sharp edges because here the contact angle can
(b)
change without moving the position of the oil/water/rock
contact.
Fig. 10-Effects of surface roughness on apparent con-
The third reason for insensitivitl. to contact angle is the tact angle aa' 9 T is the contact angle measured on a
rough surfaces of the cores. 41 ,4 ,67 Surface roughness smooth, flat surface. (a) The droplet is the preferentially
diminishes the apparent contact angle when the contact wetting fluid, so aT < 90°. (b) The droplet is the nonwet-
ting fluid, so 9 T > 900 . 70
angle measured on a flat plate is less than 90° [1.57 rad]
and increases the apparent contact angle when the true
contact angle is greater than 90° [1.57 rad]. For exam-'
pIe, Tarnai and Aratani 69 measured the contact angle of B is a roughness ratio defined by
mercury on silica glass surfaces. When the glass had a
A
mirror-smooth finish, the contact angle was 129° [2.25 B == - , ....... , ......................... (21)
rad]. The apparent macroscopic contact angle increased A'
to 155 to 160° [2.71 to 2.79 rad] when the silica surface
was roughened. Good and Mikhail 48 recently observed where A is the true area of the surface, taking into ac-
apparent contact angles of mercury on the surface of ce- count all of the peaks and valleys, and A' is the projected
ment paste (which is microscopically very rough) of 170 flat area of a plane with the same dimensions as A.
to 175° [2.97 to 3.05 rad]. Note that the apparent contact angle will be 0 or 180 0
The effects of surface roughness on apparent contact [0 or 3.14 rad] when r is sufficiently large. For example,
angle for preferentially wetting and nonwetting drops is the apparent contact angle for mercury «(}T= 140 [2.44
0
shown in Fig. 10. 70 (}T is the contact angle measured' on rad]) will be 180 0 [3.14 rad] when r is greater than 1.3.
a smooth surface, and () a is the apparent contact angle Wenzel's equation gives an idea of the effects of rough-
measured from the droplet to a horizontal plane. In Fig. ness on contact angle. It is only an approximation, how-
lOa, the drop is the preferentially wetting phase. The drop ever, and does not hold for all rough surfaces. 8,30,72,74
is stable in the configuration shown, with the contact line Variations in the wettability of the porous surface, such
on the inner surface of the roughness "peaks." The ap- as occur in fractional- and mixed-wet cores, will also af-
parent contact angle, () a' is less than the true contact an- fect the apparent contact angle. On a fractionally wetted
gle, demonstrating that roughness decreases the apparent surface, the contact angle depends on the distribution and
contact angle when (}T is less than 90° [1.57 rad]. In Fig. amounts of the differently wetted surfaces. Discussion of
lOb, the drop is the preferentially nonwetting phase these fractional wettability effects for smooth surfaces can
«(}T>90° [1.57 rad]). In this case, the stable contact line be found in Refs. 8 and 75. However, I am not aware
is on the outside of the peaks, making the apparent con- of any work for rough surfaces or porous media.
tact angle greater than () T.
One equation for the effects of surface roughness on Effects of Wettability on Irreducible
apparent contact angle is Wenzel's equation 6,70,71: Saturations
Both irreducible wetting- and nonwetting-phase satura-
tions depend on wettability, pore structure, and satura-
cos (}a=B cos (}T' . . . . . . . . . . . . . . . . . . . . . . . . (20) tion history. We will first discuss experiments on
Journal of Petroleum Technology, October 1987 1293
0.3
° 0
°
°
I
Oe
o.
i
• 0
•
; n
~Q2
~
~
~
0.3
~
10, ..
:iI
~
~
~O.'
0- PlfE UQUIOS
e-ALCCHOL- WInER MtxTlJES
~ 01
STRONGLY
WATER NEUTRALLY
USflM WETTA8lL1TY tu.HR
WET WET
°0~~10---ro~~~--~~~~570--~~~~ro7-~~---9~0--~'OO~~"O
Fig. 11-Average ROS after centrifuging vs. USBM wet- CONTACT AI«U (00lAEES)
tability index. Squirrel oil and organochlorosilane-treated
Torpedo sandstone cores, 77°F. Curve is a least-squares Fig. 12-lrreducible wetting-phase saturation vs. contact
fit. 76 angle, teflon cores. 28
uniformly wetted cores, which show that the minimum hours to reach wetting equilibrium, and then assayed for
irreducible saturation occurs when the system is near neu- the USBM wettability index and irreducible oil satura-
tral wettability. However, this finding does not apply in tion. They found that the irreducible nonwetting (oil) satu-
fractional- or mixed-wet systems or in very homogene- ration decreased from 0.35 to 0.1 as the USBM wettability
ous porous media such as bead packs. index was reduced from 1 (water-wet) to 0 (neutrally wet).
Charlesworth 78 also examined the effects of wettabil-
Uniformly Wetted Systems. Reservoir Cores. Fig. 11, ity on the ROS. His results will not be discussed here,
taken from Lorenz et ai. ,76 shows the variation of the however, because of problems with his experimental
irreducible saturation with wettability in reservoir rock procedure. During his measurements, Charlesworth con-
using crude oil. The centrifuge method was used to cre- tinually accelerated his centrifuge up to its maximum
ate the capillary pressures applied to the core. (Note that speed, then decelerated it to make measurements. The pos-
the residual saturation data presented by Lorenz et al. ,76 sibility of fluid redistribution during the deceleration
McGhee et ai. ,77 and Charlesworth 78 refer to the aver- makes any interpretation of his data very questionable. 79
age saturation of the core, not the face saturation deter- For the experiments discussed above, capillary forces
mined by the method of Hassler and Brunner 79 and dominate the processes that trap residual oil. Both buoyan-
Slobod et al. 80 In these data, an assumption has been cy and viscous forces are so small in comparison that they
made that the capillary pressure is high enough for the may be neglected. In waterflooding experiments, capil-
average and face saturations to be essentially equal.) In lary forces also dominate trapping. Because the forces that
these tests, the wettability was varied with different con- determine the residual fluid distribution are the same, one
centrations of Drifilm and assessed by the USBM wetta- would also expect that capillary pressure residual satura-
bility index, W. 81-83 A W of 0 is the dividing line tion would be similar to the residual saturation achieved
between wetting preference. For W<O, oil is the wet- after waterflooding a core with many PV's of water. For
ting fluid, whereas for W>O, water is the wetting fluid. example, Kennedy et ai. 84 waterflooded chemically con-
The further W is from zerQ, the greater the wetting prefer- solidated sandpacks with wettabilities that ranged from
ence. As can be seen in Fig. 11, the ROS decreases from water-wet to oil-wet. They found that the minimum ROS
roughly 30% when the core is strongly water- or oil-wet occurred when the system was slightly oil-wet. Further
to about 20% when the system is neutrally wet. The curve discussion of the effects of wettability on waterflood ROS
reaches a minimum at a slightly oil-wet condition, when can be found in Ref. 5.
W is slightly less than zero. Beads and Sandpacks. There is a strong interaction be-
Similar results for the irreducible wetting-phase satu- tween wettability and pore geometry in determining when
ration were obtained by Morrow and Mungan,28 as a fluid becomes discontinuous and the irreducible satura-
shown in Fig. 12. As discussed earlier, they used pure tion is reached. The data of Lorenz et al. ,76 McGhee et
fluids and sintered teflon cores. Contact angles near 0° ai.,77 Morrow and Mungan,28 and Mungan 29 show the
correspond to W near 1, while contact angles near 90° importance of wettability when the geometry is sufficiently
[1.57 rad] correspond to W near O. Again, the residual complicated. In such cases, small-scale heterogeneities aid
wetting-phase saturation decreased from around 30% to in trapping the fluid. 85 If the geometry is sufficiently sim-
approximately 20% as the system approached neutral wet- ple, however, wettability effects may become unimpor-
tability. These data may show a minimum near 0=90° tant. Such simple geometries include beads and sandpacks,
[1.57 rad] at neutral wettability, although there are not where the residual saturation is typically very low (less
enough data to be certain. than 10%) and only slightly influenced by wetta-
McGhee et ai. 77 measured the irreducible nonwetting bility. 37,86,87
phase and obtained results similar to Lorenz et al. Clean Morrow 86 examined the irreducible wetting-phase
Berea cores were saturated with brine, driven to IWS with saturation of pure fluids in random packings of equal-size
the crude oil to be tested, aged at 140°F [60°C] for 1,000 teflon spheres. In this case, the irreducible saturation took
1294 Journal of Petroleum Technology, October 1987
the form of pendular rings,7,18 which are disconnected
rings of fluid wrapped around the contact point between 1.6 r------------------.
two adjacent spheres. For this simple core geometry, Mor-
row found little variation in the irreducible saturation as
1.4-
the wettability was varied. For variations in 0 ranging
0
from 0 to 108 [0 to 1. 88 rad], S w was always in the
range of 7.5 ± 1. 5 %. Von Engelhardt 87 and Harris et 12-
al. 37 have also reported very little effect of wettability
on irreducible saturation in their packed beds.
One exception to these findings is the experiments of
01
~0
~~
?
1.0
Bethel and Calhoun,33 who used n-octane to displace Q
water from glass bead packs. The wettability was varied i
by treating the beads with different concentrations of ~ O.Sf-
0..
. -
ao.e_
0 .......
... OioCIyleiMr
1II_·'ro_p*thGIltII.
""..
..i
......
n·
~ 0 .•
~ 0.6
~ 0 .•
S DlOCTYL ETHER
~:J 0.4
~-
9'
8ROMONAPHTHALENE
,
llIitlallHlllid IOtUrGtlOll OX)! .0 ( PV
i
c 0.2
OOO<CANE
0.'
0,
'0 1MB1B1TlOTfTlME ~mlftut..
the core. At the same time, the displacement pressure to and different pure fluids. The results are shown in Figs.
force the nonwetting fluid into the core also decreases. 14 and 15. 50 Fig. 14 shows the results for imbibition of
As the system approaches neutral wettability, neither several fluids into an initially dry (air-saturated) core. The
fluid will imbibe spontaneously. On the basis of wetta- imbibition rate becomes slower and the amount imbibed
bility considerations alone, one would predict that spon- smaller as the contact angle increases and the favorable
taneous imbibition would occur for any contact angle less change in surface free energy decreases. Note that dioctyl-
than 90° [1.57 rad] because there is still some driving ether (0=49° [0.85 rad]) was unable to imbibe.
force for the wetting fluid to displace the nonwetting one. Fig. 15 shows the results for free imbibition of the fluids
It is found experimentally, however, that spontaneous im- into a core with an initial wetting fluid saturation of about
bibition in a uniformly wetted core will stop at a much 30% PV. Imbibition rates were much faster when com-
lower contact angle 51 ,90 because of the effects of pore pared with the dry core. Again, the imbibition rate slowed
structure, and a positive displacement pressure will be re- as the liquid/air/teflon system became more neutrally wet.
quired to force both the preferentially wetting and non- In contrast to Fig. 14, however, the final saturation was
wetting fluids into the core. Positive displacement similar for all fluids that could imbibe. When an initial
pressures for both fluids, particularly when the core is saturation of 30% PV was present, dioctyl-ether was able
initially 100% saturated with the other fluid, have fre- to imbibe slowly into the core. Apparently, the presence
quently been reported in the literature. 14,15,17,51,90-92 of an initial connected saturation was necessary for
When a positive displacement pressure is required for both dioctyl-ether to imbibe. On the other hand, a-
fluids, the fluid with the lower displacement pressure is bromonaphthalene (OT=73° [1.27 rad]) was never able
the preferentially wetting fluid because less energy is re- to imbibe, whatever the initial saturation of the core.
quired to force it into the core. 16,90 A number of wettability measurement systems, such as
Morrow and Mungan 28 and Morrow 29 measured the the Amott 93 method, are based on free imbibition rates
zero-capillary-pressure nonwetting saturation using pure or volumes. 2,94 The results of McCaffery et al. show that
fluids in a teflon core. The capillary pressure, initially there is a large range of contact angles for which neither
at a high positive value, was gradually lowered to zero fluid will imbibe and that the cutoff depends on the ini-
as the wetting fluid imbibed. The results, normalized to tial wetting-phase saturation. Dioctyl-ether would not im-
the residual saturation obtained when the contact angle bibe into a dry core, whereas a-bromonaphthalene would
was zero, are shown in Fig. 13. This nonwetting-phase never imbibe. These systems would be classified as neu-
residual saturation increases as the contact angle increases, trally wet by imbibition tests.
which is the exact opposite of the behavior for the irredu- The interaction of wettability and pore structure is
cible nonwetting saturation (see Fig. 11). Note that the responsible for the failure of spontaneous imbibition as
nonwetting phase at zero capillary pressure is generally the contact angle is increased. The apparent contact an-
still connected, so this residual saturation is not ir- gle in a porous medium often differs substantially from
reducible. the contact angle measured on a smooth surface or in a
The experiments above measured the amount of spon- circular capillary tube. Calculations for pore geometries
taneous imbibition when the capillary pressure was re- consisting of cones,64 toroidal (doughnut-shaped)
duced from a large, positive value to zero in a quasistatic pores,II,95 grids,66,96 and packed spheres 23 ,97,9g all
manner. A second type of experiment that measures both show a strong interaction between contact angle and ge-
imbibition volume and rates will be referred to here as ometry in determining the capillary-pressure/saturation
free imbibition. In these experiments, a plug with a low relationship. For example, Melrose 23 calculated that the
wetting-phase saturation is immersed in the wetting fluid, wetting phase would not imbibe into a porous medium
while the amount and rate of spontaneous imbibition is composed of packed spheres when the contact angle was
measured. In free imbibition, there is no applied capil- greater than roughly 40° [0.7 rad].
lary pressure. As a second example, the apparent contact angle is
McCaffery et al. 50-52 measured the rate and volume generally larger than the true contact angle for a drop on
of countercurrent free imbibition in teflon cores with air a porous surface (the drop acts less wetting) because of