Journal of Petroleum Science and Engineering 179 (2019) 606–614

Contents lists available at ScienceDirect

Journal of Petroleum Science and Engineering

journal homepage: [Link]/locate/petrol

The diffusion of water through oil contributes to spontaneous emulsification T

during low salinity waterflooding
Salem Aldousary, Anthony R. Kovscek∗
Stanford University, Energy Resources Engineering, Stanford, CA, 94305, USA

A R T I C LE I N FO A B S T R A C T

Keywords: Increases in oil recovery due to Low Salinity Waterfloods (LSW) have been widely investigated. Literature is
Waterflooding abundant with studies attempting to pinpoint the relevant mechanisms involved in the LSW process. Recently,
Microfluidics there was clear evidence of spontaneous formation of emulsions during LSW under certain conditions. This work
Low salinity examines the diffusion of water through oil as a contributing factor for spontaneous emulsification during LSW.
Emulsion
Experimental observations are pore scale using etched-silicon micromodels. Observations are accompanied by
Spontaneous
pore-scale mass transfer calculations. Spontaneous emulsification is observed as a result of so-called “low salinity
water” and crude oil phases contacting under non-equilibrium conditions. The experimental observations pro-
vide clear evidence that physical contact between oil and low salinity water phases is imperative for emulsifi-
cation. In addition, it is concluded that a larger salinity gradient results in greater emulsion concentrations. In
such cases, the crude oil is preequilibrated through exposure to a formation brine that has substantial salinity
and divalent cations. Emulsification results as the oil phase comes into equilibrium with a brine of lower salinity
and different composition. A key mechanism is the osmotic pressure imbalance between water and oil phases
that drives water into the oil phase by diffusion. Water reaches a supersaturated state and spontaneous nu-
cleation of the water phase results in emulsions, under appropriate conditions. The experimental and numerical
modeling schemes presented here demonstrate and quantify time scales of water-in-oil emulsion formation.

1. Introduction and minerals (Lager et al., 2006; Tang and Morrow, 1999; Ligthelm
et al., 2009; McGuire et al., 2005; Austad et al., 2008; Yousef et al.,
Waterflooding has been used to aid oil recovery for roughly a cen- 2010; Alshakhs and Kovscek, 2016; Yun and Kovscek, 2015).
tury, but oil recovery mechanisms are still not fully delineated. Direct visualization experiments in micromodels have been re-
Improved understanding of recovery mechanisms benefits efforts to cognized as an aid in identifying the flow dynamics and the physical
increase recovery efficiency. One variant of waterflooding that is interactions at the pore-scale (e.g., Lifton, 2016; Song and Kovscek,
emerging to maximize waterflood recovery is low salinity (LSW) in- 2015). Significantly, Emadi and Sohrabi, 2012 reported the formation
jection. Low salinity water injection refers to water injection at a sali- of water-in-oil microemulsions during LSW floods when the crude oil
nity less than that of the resident brine in the reservoir. There is sig- comes in contact with the LS injection brine. Other experiments both at
nificant evidence that LSW improves oil recovery. Numerous laboratory the pore and core scale confirmed the nucleation of a new phase within
and field-scale studies pointed out that modification to the ionic com- the crude oil (Emadi and Sohrabi, 2013; Song and Kovscek, 2015;
position of injection brine results in greater oil recovery (e.g., Yildiz and Mahzari and Sohrabi, 2014; Bartels et al., 2016a, b; Mahzari et al.,
Morrow, 1996; Tang and Morrow, 1999; Austad et al., 2011). The 2018). Microemulsions also may nucleate during LSW floods and have
mechanisms and conditions under which LSW succeeds are under study; been suggested as a mechanism in the LSW process (Emadi and Sohrabi,
however, there is general consensus that the improved recovery is 2012). Further, Emadi and Sohrabi, (2012, 2013) suggest that forma-
mainly attributed to a shift in wettability of rock pore walls toward tion of water-in oil microemulsions and wettability alterations con-
increased water wettability. The reasons for the change vary depending tribute to changes in the oil distribution and, hence, additional oil re-
on the environment and conditions. Potential mechanisms include, but covery.
are not limited to, fines migration, electrical double layer expansion, Importantly, the literature suggests that formation of water-in-oil
anhydride dissolution, and ion exchange between aqueous solutions emulsions is a contributing factor to the success of LSW (Emadi and

∗
Corresponding author.
E-mail address: kovscek@[Link] (A.R. Kovscek).

[Link]
Received 8 May 2018; Received in revised form 10 April 2019; Accepted 11 April 2019
Available online 16 April 2019
0920-4105/ © 2019 Elsevier B.V. All rights reserved.
S. Aldousary and A.R. Kovscek Journal of Petroleum Science and Engineering 179 (2019) 606–614

Sohrabi, 2013; Mahzari and Sohrabi, 2014; Mahzari et al., 2018; Table 1
McGuire and Vittoratos, 2015; Alhammadi et al., 2018). Mahzari and Brine composition (meq/L).
Sohrabi (2014) reported that the crude oil composition, i.e. specifically Injection Brine Formation Brine
the polar components, changed due to the formation of emulsions in
LSW. The authors speculated that the asphaltene to resin ratio could be Na+ 444.48 Na+ 2014.91
K+ 9.28 K+ 19.13
relevant to the concentration of emulsions. Mahzari et al. (2018) vali-
Mg2+ 2.14 Mg2+ 144.81
dated the same results, however, the authors also claimed a link be- Cl- 466 Cl− 2971.94
tween the crude oil propensity to form emulsions and the existence of SO42- 0.1 Sr2+ 4.12
polar agents based on FTIR spectroscopy. This connection was estab- CO32- 1.15 Ca2+ 291.24
lished indirectly based on water content measurements performed on Ca2+ 1.57 Br− 8.34
Br- 0.78
three crude oil samples contacted by LSW and consequent changes in
the sulphur, nitrogen and aromatic rings compositions of some samples.
While crude oil polar components may aid the stabilization of emul-
sions, such studies do not reveal that crude oil components are re-
• The crude oil studied differs from any of those in the literature ex-
amined for low salinity waterflooding and this observation is im-
sponsible for water droplet formation. For instance, the images pre- portant for establishing the generality of the phenomenon.
sented by Fredriksen et al. (2016) show formation of water-in-oil
emulsion in n-hexane that contains no polar components. The model itself contains several original ideas. Chief among them
Generally, it is found that the greater the crude oil ability to form is that a region of supersaturation of water dissolved in the oil phase is
emulsions, the greater the oil recovery from LSW. Alhammadi et al. created due to diffusion of water into the oil and slower countercurrent
(2018) suggests that formation of emulsions disturbs the equilibrium of diffusion of hydrated cations out of the oil. Model results quantify the
oil polar components between the bulk oil and the rock. This, in turn, time scale of this diffusion-based process. The details of spontaneous
results in changing the wettability of the system towards more favor- emulsification loosely follow the “diffusion and stranding” process.
able conditions and improves oil recovery. There are conflicting in- Exceeding the solubility limit of water in the oil phase results in nu-
sights on to whether the initial degree of wettability impacts LSW ef- cleation of water droplets and, hence, the formation of water-in-oil
fectiveness (Mahzari et al., 2018; Alhammadi et al., 2018; Fredriksen emulsions. This paper proceeds by outlining the mechanism and the
et al., 2016). It is indisputable based on recent pore-scale investigations model in detail, explaining the experimental system, and then pre-
that formation of emulsions is a contributing factor of successful LSW; senting results. Discussion and conclusions complete the paper.
however, a mechanism of emulsion formation throughout the process
was not presented.
Spontaneous emulsification refers to emulsions of immiscible fluids 2. Proposed model of water transport
that form without external/mechanical means such as low interfacial
tension. The study of spontaneous emulsification has been of wide in- Spontaneous emulsification arises as immiscible phases come to
terest due to its great benefits in numerous practical applications equilibrium. Oil is initially equilibrated with formation brine of con-
(Miller, 1987). Spontaneous emulsification may be attained as a result siderable salinity and a given composition (e.g. Table 1). Hydrated ions
of diffusion when two immiscible liquids that are not initially in equi- are dissolved in the oil phase at the solubility limit for the given brine
librium are brought into contact (Solans et al., 2016; Miller, 1987). composition. Forced or spontaneous imbibition of low salinity brine
From a thermodynamic viewpoint, chemical potential differs between introduces an osmotic pressure imbalance between the initial and the
the two phases and the free energy of emulsification (ΔG) is less than 0. imbibed aqueous phases that are separated by the oil phase. The sali-
Consequently, emulsions are produced spontaneously (Solans et al., nity/concentration gradient drives water into the oil phase. Diffusion
2016). Three main mechanisms for spontaneous emulsification are dictates the transport of water, from the low to the high salinity phase.
suggested in the literature: interfacial turbulence, negative values of Although brine solubility in crude oil increases as salinity decreases
interfacial tension, and “diffusion and stranding” (Davies and Rideal, (Sandengen and Arntzen, 2013), local supersaturation of water in the
1963; López-Montilla et al., 2002). A complete consensus and under- oil phase ensues leading to nucleation of water-in-oil emulsions. Im-
standing of the dynamics of spontaneous emulsification in porous portantly, the transport of water through the oil phase at the pore level
media has not been reached. in the presence of a salinity gradient was previously demonstrated ex-
This study investigates the process of spontaneous emulsification perimentally with respect to oil swelling, mobilization and emulsifica-
from a pore-scale perspective in an improved oil recovery (IOR) con- tion (Fredriksen et al., 2016), but without quantification of diffusion
text. Note that this study does not delve into the initial wettability of time scales.
the system, emulsion stability, or provide any interpretations of LSW Fig. 1 displays a stylized schematic of the proposed model of
recovery effectiveness. We focus on how sufficient water enters the oil spontaneous emulsification. The oil is initially in equilibrium with
phase such that emulsions form spontaneously during LSW. Pore-scale saline brine on both sides of the oil phase. The multiphase equilibrium
experiments using etched silicon micromodels with a sandstone-like is disturbed by the injection of the low salinity brine and contact of the
pore network combined with numerical modeling of the diffusion of oil phase by the brine phase. Osmotic pressure induces diffusion of low
brine solutions through the oil phase demonstrate critical conditions to salinity water into the oil layer. Counter diffusion of hydrated ions, such
generate spontaneous emulsification. The experiments and modeling as Ca++ or Mg++, out of the oil phase also occurs as the oil phase seeks
lead to novel observations including the following. to equilibrate with the low salinity water. The extent of low salinity
brine dissolving into the oil phase depends upon the temperature,
• Physical contact between the crude oil and injected low salinity salinity of the injection brine, and the solubility limit of water in the
brine is important to form emulsions. crude oil. It is well established that solubility of water in oil is a func-
• A greater salinity gradient provides a larger concentration of tion of salinity.
emulsified water in the oil phase. Water content measurements of In subsequent sections, it is shown that the rate of diffusion of hy-
crude oil are linked to these micromodel observations for further drated ions through oil is relatively slow whereas osmotic pressure
confirmation of this point. drives water into the oil phase relatively quickly. The dissolved water
• Diffusion of water into the crude oil phase is important for the content in the oil phase thus exceeds the solubility limit. This creates a
formation of water-in-oil emulsions. Polar components in the oil local supersaturation region during which spontaneous nucleation of
phase may be important for stabilizing these emulsions. water-in-oil emulsions takes place.

607
S. Aldousary and A.R. Kovscek Journal of Petroleum Science and Engineering 179 (2019) 606–614

Fig. 1. (a) Proposed mechanism of spontaneous emulsification. (b) Water diffuses (yellow hatches) into the oil layer while hydrated ions counter diffuse more slowly.
Local supersaturation of water in the oil phase is followed by nucleation of the water phase within the oil. (For interpretation of the references to color in this figure
legend, the reader is referred to the Web version of this article.)

A simplified mathematical model helps to demonstrate the emulsion of water in oil as well as hydrated divalent ions through oil using the
formation. The oil phase is initially equilibrated with a saline brine Stokes Einstein equation as appropriate to dilute solutions, and con-
phase. This oil phase is then contacted by low salinity brine. This is sistent with literature (Su et al., 2010; Harris, 2009). The diffusion
conceptualized in Fig. 1(a) as an oil phase contacted by a low salinity coefficients are found as (Einstein, 1905; Reid, 1987).
solution on the left and by the initial brine on the right.
RT 1
Several assumptions are made to simplify the model and represent D=
NA 6π μr (2)
the system. Transport is assumed one dimensional and only transport
within the oil phase is considered. Transient diffusion of water and where R is the gas constant, T is the temperature, NA is Avogadro's
hydrated ions is modeled with constant (but different) diffusion coef- number, μ is the dynamic viscosity, and r is the radius of the spherical
ficients as hydrated ion. Consequently, the radii of hydrated ions are needed.
These radii were attained from He et al. (2005) and Nightingale, Jr
(1)
(1959) who tabulated the radii of hydrated cations from a number of
where C is the concentration of a species (water or hydrated ion) in the sources. For instance, the hydrated radii of Ca2+ and Mg2+are
oil phase, t is time, D is the diffusion coefficient, and x is the one-di- 4.12 × 10−10 and 4.28 × 10−10 m, respectively. The diffusion coeffi-
mensional position within the oil phase. cient (low salinity) water in oil was estimated to 4.8 × 10−10 m2/s. The
The initial condition of the oil phase is oil with water and hydrated average diffusivity of hydrated ions was estimated to be
ions dissolved within the oil at the solubility limit for the high-salinity 9.2 × 10−12 m2/s using Eq. (2). The hydrated ion value is two orders of
conditions. The hydrated ions associated with the high salinity water magnitude lower than that of water in the oil phase. Concentration
(HSW) are shown in Table 1. The right boundary condition pertains to profiles of the hydrated ions in crude oil were computed numerically by
the water content of the oil on that end of the oil lens. On the right, the solving Eq. (1) using second-order finite differences. The domain of
water content is taken as constant at C1. The boundary condition on the 1 × 10−3 m was discretized into 100 grid blocks and a constant time
left is the solubility of water at the low salinity conditions in the crude step of 1 s was used. Results follow after presentation of the experi-
oil that is greater than C1. We assume that this boundary condition is mental apparatus.
established instantaneously.
Importantly, a low salinity brine is more soluble in crude oil as 3. Experimental methods
compared to more saline brine. Representative concentrations of water
at low salinity and high salinity conditions in crude oil were obtained Pore-scale visualization of spontaneous emulsification was achieved
for the experimental crude oil and used in the model. These are in using a two-dimensional microfluidic device replicating the pore net-
addition to the initial water content of the crude oil. The Karl Fischer work structure found in sandstone. Micromodels imitate many aspects
Titration (KFT) method was used to determine water concentrations of pore structure of real porous media in order to allow direct visuali-
when contacted/equilibrated with brine of different salinities (Kestene zation of single and multiphase flows at the pore scale. Oil recovery is
et al., 2008). The water content measurements are shown in Fig. 2. It observed at the pore scale and microscopic fluid flow analyzed. The
was deduced from the measurements that the water content increases porous medium was fabricated by etching a silicon wafer to a depth of
when contacted/equilibrated with lower salinity brine. Water is “con- 25 μm using photolithography as presented by Buchgraber et al. (2012)
tacted” with the crude oil by placing both fluids of the same volume in a and Yun et al. (2017). The pore network for this study was re-
laboratory tube for few days without stirring. On the other hand to presentative of sandstone as described by Rangel-German and Kovscek
“equilibrate” the fluids, they are placed in a beaker and stirred vigor- (2006). The wafer was bonded anodically to an optically smooth glass
ously. We refer to the oil phase as equilibrated when the water content cover plate to create a two-dimensional flow medium. During the
of the oil becomes constant. KFT measures total water content including bonding process, the surface of silicon was oxidized to generate SiO2,
soluble water and water in the form of droplets. After preparation, the which is similar to quartz that is the main component of sandstone
fluids are preserved at room conditions until measurements are made. (Buchgraber et al., 2012). The micromodel used in the experiments had
Fig. 2 teaches that the smaller the brine salinity, the greater the water a dimension of 5 cm by 5 cm representing a network with roughly 600
content in the crude oil. by 600 pores. There were fluid distribution channels along two opposite
Diffusion coefficients are required to proceed. We model diffusivity sides of the micromodel to permit macroscopically one-dimensional

608
S. Aldousary and A.R. Kovscek Journal of Petroleum Science and Engineering 179 (2019) 606–614

Fig. 2. Water Content Measurements (ppm) for the crude oil used in the experiments when contacted/equilibrated with brines of different salinities. Repeat
Measurements indicated.

flow. Table 2
The experimental apparatus is shown in Fig. 3. Micromodels were Crude oil properties.
the centerpiece. The remainder of the apparatus components consisted Acid Number 1.27
of a syringe pump, a stainless steel syringe, reflected-light microscope, Base Number 3.06
and a camera to capture images during the experiments. Preparation of Viscosity (cP) @ 22 C 9.10
the apparatus generally started with the injection of CO2 gas at a Asphalt wt. % 2.10
Density (g/cc) @ 22 C 0.885
moderate flow rate to displace the initial air residing inside the mi-
crofluidic device. Vacuum was applied at the production end. Deionized
water was then flushed into the micromodel at large superficial velocity experiments were to examine the conditions under which emulsions
to displace completely or dissolve any remaining CO2 and ensure no gas form and the frequency of emulsion in the oil phase. In the first type of
was trapped inside. Subsequently, several pore volumes of saline for- experiment, the low salinity fluid was deionized water to study the
mation brine were injected to bring the micromodel to a fully brine- impact of very low salinity. The second type of experiment was carried
filled state. Table 1 lists the composition of the initial brine. Crude oil out with a larger salinity injection brine to achieve a better under-
was then injected into the micromodel to establish the initial conditions standing of the phenomenon. These conditions correspond with those in
of oil and initial formation water. The properties of the crude oil used in Fig. 2. For both injectants, the injection rate of water was 0.0001 mL/
the study are shown in Table 2. Importantly, the crude oil was pre- min such that the superficial velocity of injectant was roughly 0.1 m/
equilibrated with the formation brine before injection into the micro- day.
model.
Two different types of experiments were conducted to support the
notion of spontaneous emulsification during LSW. The objectives of the

Fig. 3. Apparatus for microfluidic study of spontaneous emulsification.

609
S. Aldousary and A.R. Kovscek Journal of Petroleum Science and Engineering 179 (2019) 606–614

Fig. 4. (a) & (b) Pore space after flooding with low salinity brine. Arrows point to grains, oil, water, and water-in-oil emulsions that fill the microfluidic device. (c) &
(d) close ups of the emulsified oil.

4. Results and discussion

Experimental results from the injection of low salinity brine and

then high salinity brine into the crude-oil-saturated micromodel are
presented. Then, concentration profiles computed from the diffusion
equation are then discussed. Importantly, no experimental evidence
was seen of emulsion formation by shearing because the injection rates
are small and the oil/water interfacial tension is large.

4.1. Low salinity injection brine

Similar to results presented by Emadi and Sohrabi (2013) and

Fredriksen et al. (2016), our pore-scale visualization experiments dis-
played water-in-oil emulsions that formed once the crude oil and the
LSW brine came into contact. The dynamics of emulsification are rapid
and difficult to capture in our experimental setup. Nevertheless, the
whole microfluidic device was filled with water-in-oil emulsions both at
the oil-water interface and in the oil phase, as shown in Fig. 4. The term
“water-in-oil” emulsion refers to a discontinuous water phase contained Fig. 5. Emulsion-free pore space in areas not contacted by the low salinity
within a continuous oil phase. In a different publication, a core flooding brine.
experiment was conducted with the same crude oil (Kim et al., 2019).
The water concentrations in the produced effluent increased compared for injection, the amount of water entrained in the crude oil decreased
to the original water content of the oil. In Fig. 5, it was also observed significantly as seen in Fig. 6. This result supports our suggestion that
that regions not contacted by injection brine do not exhibit any emul- the salinity gradient drives not only the formation of water-in-oil
sions. The unemulsified oil phase lies in a region of the micromodel emulsions but also the degree of emulsification This means that the
uncontacted by the LSW brine. Hence, it contains no emulsions. These concentration of emulsion increases as the salinity gradient between the
observations strongly suggest that the formation of emulsions is con- injection and formation brines increases. This observation is consistent
trolled by both the brine salinity gradient and physical contact between with the water content measurements results displayed earlier. To
the hydrocarbon phase and the aqueous low salinity solution. This, in further confirm this concept, we shocked the system with the LSW once
turn, drives the system towards osmotic equilibrium. Results in Figs. 4 again. As anticipated, the concentration of the water-in-oil emulsions
and 5 were repeated in subsequent tests, but are not shown for brevity. increased dramatically as shown in Fig. 7. These experiments were also
repeated but results are again not shown for brevity.
4.2. Greater salinity injection brine

Table 1 shows the composition of an injection brine of greater

salinity, but less so than the formation brine. When this brine was used

610
S. Aldousary and A.R. Kovscek Journal of Petroleum Science and Engineering 179 (2019) 606–614

Fig. 6. (a) & (b) Emulsions resulting from the injection of the brine summarized Fig. 7. (a) & (b) Increase in emulsion concentration after injection of LSW brine
in Table 1. Compare to Fig. 3 to note that emulsion concentration is reduced. in a tertiary mode. Note emulsions throughout the oil phase.

4.3. Diffusion modeling on the balance between Gibbs free energy and entropic limitations that
nucleated water droplets may remain as an emulsion, grow into larger
The diffusion equation was solved numerically as described above. emulsion droplets, or coalesce into a bulk phase (Miller, 1987). The
The concentration profile behaves in a manner similar to the schematic experimental observations indicate that the water remained emulsified.
shown in Fig. 1. The water diffuses into the oil phase, whereas the Within the oil phase, water-in-oil emulsions are abundant during the
hydrated ions equilibrated with the oil phase diffuse in the opposite experiments.
direction, but at a much smaller rate because their diffusivity coeffi- The experimental results of Mahzari et al. (2018) are deduced to
cient is an order of magnitude smaller than water in oil. Fig. 8 illus- match our proposed theory although not concluded as such by the
trates the concentration of water across the oil layer as a function of authors. Their emulsion measurements without doubt agree with the
time. It is apparent that the water content is greater near the left diffusion mechanism proposal. When the water resident in the crude oil
boundary and declines towards the right boundary. Oil lenses in the is of greater salinity as compared to the injecton/contact brine, it leads
experiments are typically on the order of 100 μm and thus the water to increases in the water content of the crude oil due to transport of
saturation would fundamentally be uniform (i.e. greater than the initial water down the salinity gradient by diffusion. The opposite happens
water content prior to LSW injection). The oil phase attains super- when the injection/contact brine is of greater salinity, evidently the
saturation of water near the water/oil interface rapidly, and the nu- water content in the crude oil goes down. The authors did not explain
cleation of water droplets ensues. Fig. 8 predicts that within 10's to why the water content in the crude oil decreases when the injection/
100's of seconds, water transports substantially into the oil phase near contact brine is higher in salinity. Similarly, Song and Kovscek (2015)
the interface. Fig. 9 exhibits the concentration profile of the hydrated describe the formation of emulsions accompanying low salinity water
ions diffusing through and out of the oil membrane. Comparing profiles injection, emulsion breaking upon injection of a saline brine, and then
at identical times in Figs. 8 and 9, it is apparent that water has pro- formation of emulsions when the low salinity brine was reinjected into
gressed well into the domain, while hydrated ion concentration has the pore space. We attribute these observations to the osmosis phe-
changed little. Note that the color coding of time is identical between nomenon.
Figs. 8 and 9.. Understanding the role and physics of spontaneous emulsification in
Comparing Figs. 8 and 9, the concentration profiles teach clearly the displacement of oil at the pore level is a key toward developing
that water diffuses more rapidly into the oil layer as a result of osmosis better conceptual and numerical models to aid implementation of more
while the hydrated ions leave at a slower pace. The model shows un- effective oil recovery techniques. While it has been suggested that
equivocally that a local region of supersaturation is created because the spontaneous emulsification leads to greater oil recovery through mo-
advancing water content exceeds the solubility limit of the crude oil bilizing residual oil left behind in conventional waterfloods (Edgbogah
equilibrated with the high salinity brine of 242 ppm. The super- and Dawe, 1985; Santanna et al., 2009; Song and Kovscek, 2015),
saturation state results in the nucleation of water droplets. Depending

611
S. Aldousary and A.R. Kovscek Journal of Petroleum Science and Engineering 179 (2019) 606–614

Fig. 8. Concentration profile of water in the oil layer. Times and line colors are the same as Fig. 9. X (= 100 μm) is the system length while x is the position within the
oil phase. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

improved understanding of the mechanism aids in selecting the best to be one of the contributing mechanisms of improved recovery of
brine composition for injection. Because water diffusion into the oil viscous oil (Vittoratos and West, 2010). We expect diffusion driven
phase prompts spontaneous emulsification and swells the oil phase, emulsification is also relevant within the viscous oil setting.
additional oil blobs may be mobilized (Fredriksen et al., 2016). These Considering that water-in-oil emulsions do form under the condi-
mechanisms combine to improve oil displacement efficiency. Visuali- tions of this study, it is compelling to optimize their effect in order to
zation of pore-scale mechanisms here and elsewhere (Emadi and achieve greater oil recoveries. A screening tool to assess the likelihood
Sohrabi, 2013; Song and Kovscek, 2015) strongly supports the presence of emulsification may be the measurement of solubility of water in oil
of emulsions during some low salinity waterfloods. Although not pre- for brines of different composition.
sent in these experiments, fines at the water/oil interface may also The numerical model suggests that contact of the injectant with the
contribute to emulsion stability. Additionally, emulsions were claimed oil phase as well as the dynamics of equilibration are important to

Fig. 9. Concentration profile of the hydrated ions. Times and line colors are the same as Fig. 8. (For interpretation of the references to color in this figure legend, the
reader is referred to the Web version of this article.)

612
S. Aldousary and A.R. Kovscek Journal of Petroleum Science and Engineering 179 (2019) 606–614

spontaneous emulsification. In the future, we shall explore a wide range References

of formation brine and injection brine conditions to understand the role
of aqueous phase composition. It may be that differences in brine Alhammadi, M., Mahzari, P., Sohrabi, M., 2018. Fundamental investigation of underlying
composition and ion type are important to the phenomenon; however, mechanisms behind improved oil recovery by low salinity water injection in carbo-
nate rocks. Fuel 220, 345–357.
the relatively fast diffusion of low salinity water into the oil phase at the Alshakhs, M.J., Kovscek, A.R., July 2016. Understanding the role of brine ionic compo-
pore scale is key to achieve the results found thus far. sition on oil recovery by assessment of wettability from colloidal forces. Adv. Colloid
Interface Sci. 233, 126–138.
Austad, T., Shariatpanahi, S.F., Strand, S., Black, C.J.J., Webb, K.J., 2011. Conditions for a
5. Conclusion low-salinity enhanced oil recovery (EOR) effect in carbonate oil reservoirs. Energy
Fuels 26 (1), 569–575.
This study presents novel observations and modeling of sponta- Austad, T., Strand, S., Madland, M., Puntervold, T., Korsnes, R., 2008. Seawater in chalk:
an EOR and compaction fluid. SPE Reservoir Eval. Eng. 11 (August), 648–654.
neously formed emulsions that nucleated in a pore-scale setting. Water- Bartels, W., et al., 2016a. Micro-ct study of the impact of low salinity waterflooding on
in-oil emulsions were visualized at pore scale where the crude oil and the pore-scale fluid distribution during flow. In: International Symposium of the
low-salinity brine phases were not preequilibrated. Several iterations of Society of Core Analysts in Snowmass, Colorado.
Bartels, W., et al., 2016b. Low salinity flooding (LSF) in sandstones at pore scale: mi-
the experiments under varied injection brine salinities strongly support
cromodel development and investigation. In: SPE Annual Technical Conference and
the mechanism of water transport into the oil phase during spontaneous Exhibition in Dubai, UAE.
emulsification. Buchgraber, M., Kovscek, A.R., Castanier, L.M., 2012. A study of microscale gas trapping
Here, the transport of sufficient water into the oil phase to result in using etched silicon micromodels. Transport Porous Media 95, 647–668.
Davies, J.T., Rideal, E.K., 1963. Interfacial Phenomena, second ed. Academic Press Inc..
spontaneous nucleation is attributed to and quantified by the diffusion [Link]
of water into oil due to a water concentration gradient within the oil Edgbogah, E.O., Dawe, R.A., 1985. Spontaneous emulsification aspect of enhanced oil
phase. The gradient is established by the introduction of a low salinity recovery. J. Chem. Tech. Biotechnol. 1985,35A 132–144.
Einstein A, Ann. Physil, [4], 17, 549 (1905); 19, 289, 371 11906): 2. Elektrochem., 14,
aqueous phase on one boundary of the oil phase. Concurrently, hy- 235 (1908).
drated ions diffuse in the opposite direction as the low salinity water Emadi, A., Sohrabi, M., 2012. Visual Investigation of Low Salinity Waterflooding.
and oil phases equilibrate. Hydrated ions, however, exit the oil layer at International Symposium of the Core Analysts. pp. 27–30 Aberdeen, Scotland, UK.
Emadi, A., Sohrabi, M., 2013. In: SPE (Ed.), Visual Investigation of Oil Recovery by Low
a much slower rate because their diffusivity is at least an order of Salinity Water Injection: Formation of Water Micro-dispersions and Wettability
magnitude less than water through oil. Hence, a local supersaturation of Alteration, pp. 1–15 166435.
water arises leading to spontaneous emulsification. The model of dif- Fredriksen, S., Rognmo, A., Ferno, M., 2016. Pore-Scale Mechanisms during Low Salinity
Waterflooding: Water Diffusion and Osmosis for Oil Mobilization. SPE 180060.
fusion uses Fick's second law to predict concentration change with time Harris, K.R., 2009 Aug 7. The fractional Stokes–Einstein equation:Application to Lennard-
and supports the notion that a supersaturated region of water dissolved Jones, molecular, andionic liquids. J. Chem. Phys. 131 (5), 054503. [Link]
in oil may form where the water content exceeds the solubility limit of 10.1063/1.3183951.
He, B., Chen, Q., Castanier, L.M., Kovscek, A.R., 2005. Improved In-Situ Combustion
the crude oil. Hence, results strongly suggest that the oil phase becomes
Performance with Metallic Salt Additives. SPE 93901.
supersaturated with water where the oil phase is contacted by low Kim, T.W., Vittoratos, E.S., Kovscek, A.R., 2019. Recovery efficiency of a 28° API crude-oil
salinity water. system as a function of voidage replacement ratio. J. Pet. Sci. Eng. 175, 1063–1087.
Importantly, it is observed that physical contact between the crude [Link]
Kestene, V., Conneely, P., Bernreuther, A., 2008. Vaporsation Coulometric Karl Fischer
oil and the injection brine phases is needed to drive emulsification. titration: a perfect tool for water content determination of difficult matrix reference
Where there is no pore-scale contact between oil and low salinity in- materials. Food Chem. 106, 1454–1459.
jectant, there is no emulsification. Importantly, experimental results Lager, A., Webb, K.J., Black, C.J.J., Singleton, M., Sorbie, K.S., 2006. In: Low Salinity Oil
Recovery - an Experimental Investigation, pp. 1–12 SCA2006-36.
show that a larger salinity gradient between formation and injection Ligthelm, D.J., Gronsveld, J., Hofman, J.P., Brussee, N.J., Marcelis, F., Linde, H., 2009.
brine results in greater emulsion concentrations in the oil phase for the Novel Waterooding Strategy by Manipulation of Injection Brine Composition. SPE,
crude oil studied. pp. 1–22 119835.
Lifton, V., 2016. Microfluidics: an enabling screening technology for enahanced oil re-
covery (EOR). Lab Chip 16, 1777–1796.
Acknowledgment López-Montilla, J.C., Herrera-Morales, P.E., Pandey, S., Shah, D.O., 2002. Spontaneous
emulsification: mechanisms, physicochemical aspects, modeling, and applications. J.
Dispersion Sci. Technol. 23, 219–268. [Link]
The authors gratefully acknowledge the Stanford University
01932690208984202.
Petroleum Research Institute (SUPRI-A) affiliates for supporting this Mahzari, P., Sohrabi, M., 2014. Crude Oil/Brine Interactions and Spontaneous Formation
work. The authors also express their sincere gratitude to A. Alshehri, M. of Micro-dispersions in Low Salinity Water Injection. SPE 169081-MS.
Mahzari, P., Sohrabi, M., Cooke, A., Carnegie, A., 2018. Direct pore-scale visualization of
Alshakhs, T. W. Kim, W. Yun, K. Yoo, U. Amanam and L. Sampaio for
interactions between different crude oils and low salinity brine. J. Pet. Sci. Eng. 166,
the invaluable discussions and suggestions. S. Aldousary also ac- 73–84.
knowledges his graduate fellowship from Saudi Aramco. McGuire, P.L., Chatham, J.R., Paskvan, F.K., Sommer, D.M., Carini, F.H., Exploration,
B.P., 2005. In: SPE (Ed.), Low Salinity Oil Recovery: an Exciting New EOR
Opportunity for Alaska's North Slope, pp. 1–15 93903.
Nomenclature McGuire, P.L., Vittoratos, S. E., (2015). Hydrocarbon Recovery Process. US Patent
8939211 B2.
C Concentration of a species in the oil phase Miller, C., 1987. Spontaneous emulsification produced by diffusion- a review. Colloids
Surface. 29 (Issue 1), 89–102.
T Time Nightingale Jr., E.R., 1959. Phenomenological Theory of Ion Solvation, vol. 63. Academia
D Diffusion coefficient Press, pp. 1381–1387.
x Position within the oil phase Rangel-German, E.R., Kovscek, A.R., 2006. A micromodel investigation of two-phase
matrix-fracture transfer mechanisms. Water Resour. Res. 42, W03401. [Link]
X System length org/10.1029/2004WR003918.
R Gas constant Reid, R.C., 1987. The Properties of Gases and Liquids. McGraw-Hill.
T Temperature Sandengen, K., Arntzen, O.J., 2013. Osmosis during low salinity water flooding. In: 17th
European Symposium on Improved Oil Recovery.
NA Avogadro's number
Santanna, V.C., Curbelo, F.D.S., Castro Dantas, T.N., Dantas Neto, A.A., Albuquerque,
μ Dynamic viscosity H.S., Garnica, A.I.C., 2009. Microemulsion flooding for enhanced oil recovery. J. Pet.
r Radius of the spherical hydrated ion Sci. Eng. 66 (3–4), 117–120.
Solans, C., Morales, D., Homs, M., 2016. Spontaneous emulsification. Colloid Interface
Sci. 22, 88–93.
Appendix A. Supplementary data Song, W., Kovscek, A., 2015. Functionalization of micromodels with kaolinite for in-
vestigation of low salinity oil-recovery process. Lab Chip 15, 3314–3325.
Supplementary data to this article can be found online at https:// Su, J.T., Duncan, P.B., Momaya, A., Jutila, A., Needham, D., 2010. The effect of hydrogen
bonding on the diffusion of water in n-alkanes and n-alcohols measured with a novel
[Link]/10.1016/[Link].2019.04.041.

613
S. Aldousary and A.R. Kovscek Journal of Petroleum Science and Engineering 179 (2019) 606–614

single microdroplet method. J. Chem. Phys. 132. Yousef, A.A., Al-Saleh, S., Al-Kaabi, A., Al-Jaw, M., 2010. In: SPE (Ed.), Laboratory
Tang, G.Q., Morrow, N.R., 1999. Influence of brine composition and fines migration on Investigation of Novel Oil Recovery Method for Carbonate Reservoirs, 137634.
crude oil/brine/rock interactions and oil recovery. J. Pet. Sci. Eng. 24 (2), 99–111. Yun, W., Kovscek, A.R., 2015. Microvisual investigation of polymer retention on the
Vittoratos, S.E., West, C.C., 2010. Optimal Heavy Oil Waterood Management May Differ homogeneous pore network of a micromodel. J. Sci. Petrol. Eng. 128, 115–127.
from that of Light Oils. SPE 129545. Yun, W., Ross, C.M., Roman, S., Kovscek, A.R., 2017. Creation of a dual-porosity and dual-
Yildiz, H.O., Morrow, N.R., 1996. Effect of brine composition on recovery of Moutray depth micromodel for the study of multiphase flow in complex porous media. Lab
crude oil by waterflooding. J. Pet. Sci. Eng. 14 (3), 159–168. Chip(8).

614