SPE-195564-MS

Mechanistic Simulation and History Matching of

Alkaline-Surfactant-Polymer ASP Core Flooding Experiment at Optimum
vs. Under-Optimum Salinity Conditions

Mohsen Mirzaie Yegane, Elisa Battistutta, and Pacelli Zitha, Delft University of technology

Copyright 2019, Society of Petroleum Engineers

This paper was prepared for presentation at the SPE Europec featured at 81st EAGE Conference and Exhibition held in London, England, UK, 3-6 June 2019.

This paper was selected for presentation by an SPE program committee following review of information contained in an abstract submitted by the author(s). Contents
of the paper have not been reviewed by the Society of Petroleum Engineers and are subject to correction by the author(s). The material does not necessarily reflect
any position of the Society of Petroleum Engineers, its officers, or members. Electronic reproduction, distribution, or storage of any part of this paper without the written
consent of the Society of Petroleum Engineers is prohibited. Permission to reproduce in print is restricted to an abstract of not more than 300 words; illustrations may
not be copied. The abstract must contain conspicuous acknowledgment of SPE copyright.

Abstract
Alkaline Surfactant Polymer (ASP) flooding is a chemical EOR method to increase oil recovery after water
flooding through IFT reduction and increasing sweep efficiency. Previous studies have shown that maximal
oil recovery is reached when ASP flooding is performed at optimum salinity conditions, i.e. Winsor type III
micro-emulsion phase but a recently series of core-flood experiments indicated that comparable oil recovery
could be obtained at under-optimum salinity conditions (Battistutta et al. 2015). Mechanistic simulation of
ASP flooding considering phase behavior of water-oil-surfactant system, geochemical reactions and alkaline
consumption is needed to validate the experimental data and provide a robust model for field scale studies.
In this paper detailed history matching of series of core-flood experiments was attempted. Experiments
were performed at different salinity conditions (optimum vs. under-optimum) and with different core types
(Bentheimer and Berea) using a single olefin sulfonate (IOS) and crude oil with very low acid number (<0.05
mg KOH/g oil). The numerical simulations were performed using UTCHEM, multiphase multi-component
simulator along with EQBATCH module to model the geochemical reactions.
Neglecting the effect of in-situ surfactant (soap) generation, since the acid number of crude oil was low,
modeling of the phase behavior showed an excellent match against experimental data and optimal salinity
was observed at 2.0 wt% NaCl (+ 2.0 wt% Na2CO3).Using this and considering aqueous and cation exchange
as the most important geochemical reactions in alkaline propagation, several ASP core-flood experiments at
optimum vs. under-optimum salinity conditions were successfully modeled. An excellent matching of all the
measured parameters including oil cut and recovery, pressure drop, pH and carbonate, alkali and surfactant
concentration at effluent was also achieved. Modeling confirms the results obtained from experiment which
regardless of core type, although minimum achieved IFT at optimum salinity conditions is lower than the
one achieved at under-optimum conditions, comparable final oil recovery was observed for both cases. This
emphasizes the importance of performing ASP flooding at under-optimum salinity conditions due to lower
surfactant retention and reducing the likelihood of achieving over-optimum salinity conditions.
In this paper a robust model which is calibrated with experimental data is presented to simulate ASP
flood process at various conditions and the basic model can be used to perform further simulations and can
provide practical and convenient approach to model field applications of ASP flooding.
2 SPE-195564-MS

Introduction
Alkali Surfactant Polymer (ASP) flooding is an enhanced oil recovery (EOR) method in which chemicals
are injected into reservoirs typically after water flooding to increase oil recovery by two mechanisms: (a)
reducing the interfacial tension (IFT) between aqueous and oleic phases and mobilizing trapped oil remained
in the pore structure and (b) increasing the viscosity of the displacing fluid and good mobility control because
of presence of polymer. Since water flooding is used in many reservoirs as the main IOR method (secondary
recovery), chemical EOR methods are good candidates to improve the recovery as minimum facilities are
needed to add chemicals in the injecting water. Among chemical EOR methods ASP is the most promising
method because it has the combination of Alkaline, Surfactant and polymer and advantages of each of these
components at the same time.
The alkaline flooding method relies on a chemical reaction between chemicals such as sodium carbonate
and sodium hydroxide (most common alkali agents) and organic acids (saponifiable components) in crude
oil to produce in-situ surfactants (soaps) that can lower interfacial tension to ultralow values (Sheng 2010a).
In the literature mainly the effective level of alkalinity and alkaline consumption due to chemical reactions
or ion exchange in the porous media are discussed (Bunge and Radke 1983 and 1985; Novosad 1984).
Polymer is added to alkaline to improve the mobility control (Sheng et al. 1994).
The main advantage of AP in comparison with SP is the lower cost of alkaline. However there are
problems associated with alkaline. Optimum salinity of alkaline is low and it has a narrow salinity range
(Nelson et al. 1984). Therefore, brine salinity should be low enough to reach the optimum condition and
ultralow IFT and since most of the reservoirs have high salinity brine, the optimum condition can be hardly
reached. On the other hand achieving optimum salinity in alkaline flooding is difficult because a large
concentration of alkali is required to compensate for its consumption to the rock surface. Large alkali
concentrations shift the phase behavior to over-optimum, where IFT is not ultra-low anymore (Lake 1984).
Nelson et al. (1984) proposed a method to enlarge low IFT window and thereby increase optimum
salinity by combining alkaline with surfactant which is more hydrophilic than the in-situ generated soap.
Furthermore, co-injection of alkaline and surfactant reduces the requirement of synthetic surfactant which
in turn reduces the overall cost of the process. Moreover Surfactant adsorption is believed to be lower in
presence of high pH alkaline agents (Hirasaki and Zhang, 2004). Sulfonate can be adsorbed on clay surfaces
because of electrostatic interactions. At high pH the solid surfaces is negatively charged giving rise to a large
repulsive term. This leads to reduction of adsorption of the anionic sulfonate species chemicals (Delshad
et al. 2011).
ASP is an effective process even if acid components in the oil have low concentration. This is measured
by acid number. The acid number is the amount of KOH to neutralize the acid in the oil expressed in mg
KOH/mg oil. In low acid number crude oils major effects of alkalis are lowering the IFT and decreasing
surfactant retention by increasing pH (Martin et al. 1985).
The prevailing opinion is that the maximum oil recovery for ASP flood can be achieved when the salinity
is in type III region, i.e. at optimum conditions, where IFT is ultralow (Healy et al., 1976; Nelson et
al., 1984). However, several authors have shown that maximal oil recovery does not necessarily coincide
with minimum IFT (Glinsmann 1979, Larson 1979, Alagic 2010, Sheng 2010b, Alagic 2011). For an
internal olefin sulfonate (IOS) surfactant/oil system, when the optimal salinity is reached, there is a change
in solution properties and retention of the chemicals in porous media. The solution becomes turbid and
retention in porous media increases by a factor up to 10 (Spildo 2012). This implies the potential for a
more successful and economic surfactant flooding with an IOS surfactant in a region II-, if the significantly
lower retention outweighs the disadvantage of a low and not ultra-low IFT. Recently, Battistutta et al. (2015)
performed a series of very detailed ASP core-flood experiment at salinities ranging from Type II- (under-
optimum salinity) to Type III (optimum salinity). They observed that maximum oil recovery happened at
under-optimum condition.
SPE-195564-MS 3

A proper numerical simulation work includes history-matching of core-flood experiments to calibrate

ASP flow parameters. The calibrated parameters can be used in a pilot scale or sector model to optimize
the injection schemes and to predict ASP performance. Both core-flood and up-scaled models are needed
to do the sensitivity analysis (Sheng, 2013). In this study a detailed history-matching of ASP core-flood
experiments performed by Battistutta et al. (2015) was performed. After doing the phase behavior study and
considering essential geochemical reactions occurring in the core, we performed a history-matching of ASP
core-flood experiments in which salinity and type of core (Bentheimer and Berea) vary. The aim of this study
was to validate the result obtained by Battistutta et al. (2015) which oil recoveries at under-optimum and
optimum conditions are comparable. The structure of this paper is as follows: First a review of mechanisms
in alkaline and surfactant flooding is discussed. Next, a brief description and summary of experiment results
are presented. Then, model features are described. Afterwards, simulation results of history matching of
core-flood experiments are discussed. Then, the conclusions of this study are drawn.

Background
In ASP flooding, as mentioned, natural soap which is an anionic surfactant is generated by saponification
of the acidic components in the crude oil with alkali. Fig. 1 indicates the partitioning of crude oil acidic
components between aqueous and oleic phases which can be described by Eq.1.
[1]
This geochemical reaction is known as partitioning reaction which is followed by an aqueous hydrolysis
described by Eq. 2 (deZabala et al., 1982).
[2]
The partition coefficient of the acid is

[3]

The acid dissociation constant for Eq. 2 is

[4]

When alkaline agents are injected into the reservoir more soap (A-) is generated by decreasing H+
concentration. In aqueous phase solution the NaOH a strong alkali, dissociates as follows:
[5]
[6]
Thus OH- concentration is increased and shifts the Eq. 6 to the left and as a result concentration of H+
is decreased.
4 SPE-195564-MS

Figure 1—Schematic of alkaline recovery process. Acidic component of the crude oil is portioned
+
between the aqueous and oleic phases. Injection of alkali agent (NaOH) reduces H concentration
-
and as a result more soap (A ) is generated in the aqueous phase (deZabala et al. 1982).

Phase behavior of the surfactant is the most important part of modeling the ASP flooding as it affects
the surfactant performance. Surfactant phase behavior is strongly affected by brine salinity (Lake 2008).
Two or three phases exist in equilibrium depending on the salinity of the aqueous phase (Winsor 1954, Lake
2008): type II(-), type III and type II(+). Type II(-) or under-optimum system exists below the lower salinity
limit and type II(+) or over-optimum system exists above the upper salinity limits (see Fig. 2). In a type II(-)
environment two phases coexist: an excess oil phase and a water-microemulsion phase containing brine,
surfactant and some solubilized oil. For type II(+) system, the coexistence of two phases involves a water
phase and an oil-microemulsion phase. Type III or optimum system is a three phase system between two
salinity limits. It consists of oil and water phases separated by a micro-emulsion phase where ultra-low IFTs
ranging within 10-3/10-4mN/m have been measured (Healy 1977, Nelson 1984).

Figure 2—Surfactant phase behavior at different salinity regions. In Type IIIsystem coexistence of three
phases (excess brine, excess oil and microemulsion) is observed where IFT is ultra-low (Lake 2008).

Experiment
Description: We have selected 4 representative tests from a more complete set done by Battistutta et al.
(2015). Two tests were performed with Bentheimer cores at optimum salinity (Test 1a) vs. under-optimum
(Test 1b). The other two were performed with Berea cores and similarly at optimum salinity (Test 2a) vs.
under-optimum (Test 2b). The objective of selecting such tests is to compare the results from simulation
at different salinities and with different core types against experimental data. Mineralogy and physical
properties of Bentheimer and Berea cores are listed in Table 1.
SPE-195564-MS 5

Table 1—physical properties and mineralogical composition of two different rock types used in the experiment (Battistutta et al. 2015)

The sequence of injection in core-flood experiment is as follows: First the core was flushed with CO2 for
30-40 minutes. Subsequently, it was saturated with brine for at least 10 Pore Volumes (PV). After that oil
was injected at 1.0 cm3/min under gravity stable conditions for 2.5-3.0 PV. Water flooding was performed,
using brine salinity of 2.0-3.5 wt% NaCl, at 0.25cm3/min (equal to a superficial velocity of 1.0 ft/d) for at
least 5.0 PV. Next, a slug of approx. 0.6 PV of ASP solution was injected at 0.25 cm3/min. The ASP solution
contained 0.6 wt% of Surfactant (ENORDET™ O242), 1.0 wt% of sec-butanol (SBA), 2.0 wt% of Na2CO3,
0.5-2.0 wt% of NaCl (dependent on the test) and 1250-1450 ppm of polymer (Flopaam 3330S). This ASP
slug was followed by polymer injection at 0.25cm3/min for 2.0-3.0 PV. The polymer solution contained
2.0-3.5 wt% of NaCl and 1450-1600 ppm of polymer. Injection profile is designed in a way that salinity in
water flooding and polymer drive is the same. Summary of four selected tests is provided in Table 2. Crude
oil and brine properties are listed in Table 3.

Table 2—Summary of core-flood experiments (Battistutta et al. 2015)

6 SPE-195564-MS

Table 3—Brine and crude oil properties (Battistutta et al. 2015)

Phase behavior: To select surfactant the necessary test is the phase behavior test called salinity scan test,
in additional to the aqueous stability test (Sheng 2013). The goal of the phase behavior study was to find
a surfactant formulation which could show ultralow IFT at an optimal salinity preferably in the order of
10−3 mN/m, but also that could show a low IFT at low salinity regions. For the salinity scan test water-
oil volume ratio was one and kept fix. No oil scan test was needed since oil had low acid number and in-
situ generation of surfactant (soap) was negligible. Fig. 4 shows the solubilization ratios obtained from the
conducted scanning tests. Total salinity of the system comes from two sources, NaCl and Na2CO3. In Fig. 4
there is already a fixed amount of 2.0 wt% Na2CO3 in the system and salinity varies only by increasing NaCl
concentration from 0.0 to 6.0 wt%. Optimal salinity was found to be equal to 2.0±0.1 wt% of NaCl. At this
value the oil and water solubilization ratios have the equal value of 30. This high value corresponds to a very
low interfacial tension at optimal salinity (8×10-3 mN/m as measured using a spinning drop tensiometer).
Lower and upper salinity limits were found to be at 1.75±0.05 and 2.50±0.05 wt% NaCl respectively.
Core-flood experiment: Test 1a is considered the baseline experiment and the other tests are compared
against it. This test was conducted on Bentheimer core and at optimal salinity i.e. at salinity equal to 2.0 wt
% NaCl (+ 2.0 wt% Na2CO3). Fig. 6 shows the pressure drop profile in different injection steps: oil, water,
ASP and polymer injection. Oil breakthrough during oil injection (primary drainage) occurs at 0.75±0.02
PV. Thereafter pressure drop decreases and then levels off to an equilibrium after 2.00±0.02 PV from the
start of oil injection. At this stage it was assumed the core had reached the connate water saturation (Swc =
18.8±0.1%).The pressure drop during water injection (imbibition) is lower than pressure drop during the oil
injection, due to lower viscosity of water compared to crude oil. The water breakthrough occurs at 2.40±0.02
PV. The pressure drop reaches a peak at 32.0±0.1 mBar, and then decreases to a plateau equal to 25.0±0.1
mBar; after 3.00±0.02 PV no more oil is produced. At the end of water injection the residual oil saturation
(Sorw) is 46.0±0.1%. Oil breakthrough during chemical injection occurs after 0.40±0.02 PV from the start
of ASP injection. Alkali breakthrough (identified as an increase in viscosity, pH and alkali concentration)
occurs at 0.80±0.02 PV, always measured relatively to the beginning of ASP injection. Afterwards, pressure
drop values can become unstable due to the presence of three different phases in the core: brine, oil and
microemulsion.
Fig. 8 shows the oil cut and the recovery vs. PV for Test 1a during chemical injection. The oil cut is the
ratio of produced clean oil over the residual oil after water flooding. The recovery factor described in Fig.
8 is the cumulative oil produced during ASP injection over the residual oil after water flooding. From the
oil cut, it can be observed that the oil breakthrough occurs at 0.46±0.02 PV. Oil recovery of 63.0±0.1% was
found with significant oil cut of 65.0±0.1% in the oil bank. After 1.50±0.02 PV no more oil is produced
and the curve levels off to a plateau value of 62.5±0.1%. This coincides with the time pressure drop reaches
a plateau during polymer injection. Before the alkali breakthrough, the oil produced is 16.5±0.1%. After
alkali breakthrough, the chemical slug appears at the production site and oil will be present also dissolved
in the emulsion phase.
Fig. 9 shows the concentrations of the chemical species in the effluents for Test 1a versus the number of
pore volumes injected (PV). The alkali breaks through at 0.80±0. 0.02 PV and reaches a peak at 1.20±0.02
SPE-195564-MS 7

PV. The alkali front is therefore lagging behind compared to the oil bank. Alkali concentration in the
effluents diminished to zero as oil production stopped. The evolution of the pH in the effluents is closely
related to the variations of alkali concentration as described above. Before the alkali breakthrough, the pH
is equal to 5.2±0.1, which is the same as the pH of brine used for water flooding. Fig. 9 shows also that the
surfactant breakthrough occurs at 2.55±0.02 PV, i.e. 1.75±0.02 PV later than the alkali. The retardation of
surfactant front is due to its adsorption both on the rock surface and at the oil-water interface.

Model Description
The best simulator to capture the mechanisms in ASP flooding is UTCHEM (Sheng 2013). UTCHEM is a
multidimensional chemical flooding compositional simulator which was developed for research purposes
at University of Texas at Austin. UTCHEM is generally considered as the reference implementation for
modeling soap/surfactant phase behavior, with its dependence on other parameters such as salinity or
alcohol concentration and selected geochemical reactions (Delshad et al. 1996). The parameters required
for building a basic ASP flooding simulation model can be modeled and simulated based on laboratory
measurements. The laboratory evaluations that are needed include phase behavior measurements as a
function of oil, mixture of sop and surfactant and alkali concentrations for the candidate surfactants
(Mohammadi et al. 2009, Nelson et al. 1984). Since the crude oil used in the experiment had a very low acid
number (<0.05 mg KOH/g oil) for the sake of simplicity we assume that no soap was generated as a result
of reaction of oil acidic components and alkalis. Therefore phase behavior of the system is not a function
of soap generation and no activity map was plotted.
To take into account effect of geochemical reactions that occur during ASP flooding, EQBATCH, the
geochemical module of UTCHEM is used. EQBATCH was originally developed by Bhuyan (1989) and later
was generalized to model any number of elements, chemical species and reactions (Wu 1996, Delshad et
al. 1999). Ion exchanges between the rock and the cations associated with alkali, precipitation of carbonate
salts and mineral dissolution are the three main mechanisms that have been considered in the literature as
being responsible for the alkali loss (Cheng 1986). Simulation of geochemical reactions is really tedious in
EQBATCH (Sheng 2013). To overcome this problem, a simplified model was proposed by Delshad et al.
(2011) to simulate ASP floods. In the present work the precipitation of divalent cations was ignored due
to their low concentrations. In addition, quartz dissolution at 52 °C, at which experiments were performed,
is negligible (Fournier and Rowe, 1977), so it was not considered. Thus the model includes mainly the
essential aqueous reactions with alkali and cation exchange reactions.
Buffering aqueous reactions: One example of aqueous reactions is buffered reactions. A common
example of buffered reactions which are important for the purpose of alkaline flooding is the carbonate and
bicarbonate buffered solutions (Mohammadi et al. 2009):

[7]

Alkaline ion exchange with rock: Alkaline is also consumed in ion exchange processes (Bunge and
Radke 1983, Farajzadeh et al. 2012). Key ion exchange reactions include hydrogen/sodium and sodium/
calcium. The hydrogen/ Sodium ion exchange can have a great impact on alkali consumption in comparison
with the cation exchange (Mohammadi et al. 2009). The hydrogen/sodium-base exchange (hydroxide-
exchange) is described by:
[8]
8 SPE-195564-MS

where X denotes mineral-base exchange sites. In flowing through a reservoir rock, sodium ions must fill the
available exchange sites before they can progress downstream. Eq. 8 shows that both hydroxyl and sodium
ions are consumed.
These two categories of reactions considered for our model are listed in Table 4. The equilibrium
constants at 52 °C were found using the Geochemist’s Workbench® 10 database. Elements and aqueous
species included are given in Table 5.

Table 4—Geochemical reactions used in the simulation

Table 5—Elements and reactive species

Surfactant adsorption is modeled in UTCHEM using Langmuir-type isotherm which takes into account
the salinity, surfactant concentration, and rock permeability (Hirasaki and Pope 1974). The adsorption is
irreversible with concentration and reversible with salinity. The adsorbed concentration of surfactant (Ĉ3)
is given by:
SPE-195564-MS 9

[9]

The concentrations are normalized by the water concentration in the adsorption calculations. The
minimum is taken to guarantee that the adsorption is no greater than the total surfactant concentration.
Adsorption increases linearly with effective salinity and decreases as the permeability increases as follows:

[10]

Where CSE is the effective salinity and the ratio of a3/b3denotes the maximum level of adsorbed surfactant
and b3 controls the curvature of the isotherm. The adsorption model parameters a31, a32 and b3 are determined
by matching laboratory surfactant adsorption data. The reference permeability (Kref) is the permeability at
which the input adsorption parameters are specified. For Test 1a which is performed at optimum salinity
surfactant adsorption is high and as it was found in the laboratory is considered 0.41 mg/g rock for the
simulation. Langmuir-type isotherm for this test is shown in Fig. 3.

Figure 3—Langmuir-type adsorption isotherm used in simulation for Test 1a. Surfactant adsorption
was found 0.41 mg/g rock in the laboratory for 0.006 vol% of surfactant. Using Langmuir-type isotherm
in UTCHEM, parameters a31, a32 and b3were found to match the model to experimental single point

Given all aforementioned information in this section we introduce the key features of our ASP model
as below:
1. Phase behavior of surfactant in presence of injected water and oil
2. The lowering of IFT and its effect on incremental oil production
3. The lowering of surfactant adsorption in presence of alkalis and its lower value at under-optimum
region.
4. The reactions between different species in the aqueous phase and rock minerals
5. Ion exchange reactions with clay in the rock.

Simulation Results and Discussion

Phase behavior: The surfactant/oil/water phase behavior in UTCHEM is based on Winsor (1954), Reed
and Healy (1977), Nelson and Pope (1978), Prouvost et al. (1985), and others and it uses Hand’s rule (Hand,
1939). The equations derived from Hand’s model for phase behavior calculations are solved using the Height
of Binodal Curve (HBNC) as input parameters. The values of these parameters are obtained by matching
10 SPE-195564-MS

the model to the measured phase behavior data. Lower and Upper salinity (CSEL and CSEU) are fine-tuned
to improve the match. Since the water-oil ratio (WOR) was 1, the phase behavior simulation was perfomed
on 50/50 vol% concentrations of water and oil. Thus the initial water saturation was 0.5. To model the
phase behavior as batch mode, it was considered injection of several pore volumes of water, oil, surfactant,
co-solvent and polymer whose compositions and concentrations are identical to those in salinity scan test.
Reservoir properties in batch mode together with the final values of HBNCs, CSEL and CSEU are given
in Table 6. Fig. 4 shows excellent match of UTCHEM batch results against lab data. The computed CSEL
and CSEU are in agreement with laboratory data and as shown in Fig. 4 optimal salinity occurs at 2.0 wt
% NaCl (+2.0 wt% Na2CO3).

Figure 4—Solubilization ratio of oil and water vs. salinity for salinity scan test. Optimal salinity
occurs at 2.0 wt% NaCl where oil and water solubilization ratios intersect and are equal to 30

Table 6—Model properties and matching parameters in batch mode for phase behavior

In a simulation model, we need to input a capillary desaturation curve (CDC) model. Delshad et al. (1986)
were the first to measure CDC in three-phase micellar solutions. They demonstrated that the microemulsion
phase was the most strongly trapped, suggesting that the microemulsion is the wetting phase. Morrow
and coworkers (Morrow and Songkran, 1981; Morrow et al., 1988) used the terms capillary number for
SPE-195564-MS 11

mobilization and capillary number for prevention of entrapment, (NC)c and (NC)max respectively. In principal,
the(NC)c is much higher than the capillary number at normal water flooding conditions. The residual
saturations start to decrease at the critical capillary number as the capillary number is increased, and cannot
be decreased further at (NC)max. In UTCHEM, lower and higher critical capillary numbers are used for (NC)c
and (NC)max, respectively. In the laboratory, if several points of residual saturation versus capillary number
are measured, we can use those measured points to fit a theoretical model. In UTCHEM, a form of Eq.
11 is used:

[11]

In Eq. 11, the superscript (NC)c and (NC)max mean at critical capillary number and maximum desaturation
capillary number; (NC) is capillary number, Spr is the irreducible or residual saturation of phase p and Tp is
the trapping parameter used to fit the laboratory measurements. CDC used for this simulation is shown in
Fig. 5. As mentioned the microemulsion is the most wetting phase, and the critical capillary number for a
wetting phase is higher than the one for a nonwetting phase. Therefore, the microemulsion CDC lies on the
right, the water CDC in the middle, and the oil CDC on the left.

Figure 5—Capillary desaturation curve used in the simulation for Test 1a. Capillary number for the microemulsion
as the most wetting phase is the highest and for the oil as the most nonwetting phase is the lowest

History matching of baseline experiment: The history match of the ASP flood in Test 1a is presented
in this section as the baseline experiment with model features described above. UTCHEM and EQBATCH
were used to model the ASP flood and the geochemical reactions, respectively. Parameters found by
matching simulated phase behavior with corresponding experiments were used in the ASP model.
Geochemical reactions are those presented in Table 4 and, as already mentioned; only aqueous reactions
and exchange reactions were considered in our model. ASP was injected for 0.48 PV then was followed
by a polymer drive of 2.50 PV.
Fig. 6 shows the simulation of pressure drop during oil, water, ASP and polymer injection respectively and
compares it with experimental data. Oil breakthrough and water breakthrough occur at 0.75±0.02 PV and
2.40±0.02 PV during oil and water injection respectively which are in excellent agreement with experiment.
12 SPE-195564-MS

Figure 6—Pressure drop for Test 1a. Dots present the lab results and lines indicate the simulated data.
The dashed lines separate the different injection steps: a) oil, b) water, c) ASP and d) polymer injection

For a more detailed analysis the pressure drop profile during ASP and polymer drive injection is replotted
in Fig. 7. It shows an excellent agreement of the simulated and the measured pressure drops. The pressure
drop increases during ASP injection. The small fluctuations in the pressure drop after start of polymer
injection and before alkali breakthrough are observed. After alkali breakthrough at 0.80±0.02 PV, pressure
drop declines and then tends to stabilize at a plateau value of 59.5±0.1 mBar.
Fig. 8 shows the simulated and measured oil cut and cumulative oil recovery vs. PV. Oil breakthrough
occurs at 0.45±0.02 PV with oil cut of approximately 50% which are close to values obtained from
experiment. The fluctuation in oil recovery in the oil bank shows an unsteady cumulative oil production.
After alkali breakthrough the oil is produced in the form an emulsion. The simulated oil cut is larger than the
measured one immediately after alkali breakthrough. This may be due to difference in capillary dispersion
in the model and experiment. This also reflects in oil recovery and as it can be seen from Fig. 8 oil recovery
after alkali breakthrough is higher than the experiment. However, the cumulative simulated and measured
oil recoveries are the same, 63.5±0.1%. The decrease in oil cut after alkali breakthrough coincides with
decrease in pressure drop profile. The pressure drop tends to stabilize as no more oil is produced, after
1.70±0.02 PV.

Figure 7—Pressure drop profile during chemical injection for Test 1a at optimum
salinity conditions. Pressure starts to decrease after alkali appears at effluents
SPE-195564-MS 13

Figure 8—Measured and simulated oil recovery and oil cut during chemical injection

Fig. 9 shows simulated chemical effluent results against experimental data. In Fig. 9 normalized
concentration is used to describe the chemical effluent analysis and it is defined as ratio of the chemical
detected at effluent divided by its initial concentration. Generally ASP model is capable of matching the
concentration of produced chemicals with a good accuracy. Some mismatch between the experimental and
simulated data could be a result of other reactions that occur in the core and we did not consider them for
history matching the experiment.
Alkaline agent used in the experiment dissociates as:
[12]
Followed by the hydrolysis reaction:
[13]
In Fig. 9 initial pH depends upon the pH of the water into which the salt is being introduced. Chemically
speaking, salt falles smack in the center of the acid-alkaline spectrum. If introduced to water which has a
high pH, the pH might be lowered incrementally toward the center of the pH spectrum depending on how
much water there was and how much salt was introduced (Sereshti and Aliakbarzadeh 2013). Before alkali
breakthrough no chemical appears at effluent and the pH remains at its initial value, pH = 5.20±0.05. This
value corresponds to slightly acid water and can be explained by the fact that water used in primary brine
injection is acidic water. After alkali breakthrough carbonate concentration at effluent gradually increases
and according to Eq. 13 it results in increase in alkali concentration as well as pH. After a peak at 1.20±0.02
PV carbonate concentration decreases until it becomes zero. This also reflects in alkali concentration and
pH. When no more alkali appears at effluent oil production as well ceased and pressure drop stabilized. Fig.
9 shows that surfactant breakthrough occurs much later than alkali breakthrough at 2.65±0.02 PV which
is close to experimental surfactant breakthrough. The delay is as a result of surfactant adsorption onto the
rock and at oil-water interface.
14 SPE-195564-MS

Figure 9—Chemical effluent analysis for Test 1a measured and simulated

Fig.10 shows effective salinity detected at producer during chemical injection. As mentioned experiment
designed in a way that salinity in water flooding and polymer drive is the same. In these two steps, injecting
solution contains 3.5 wt% NaCl which is equivalent of 0.59 meq/mL. ASP slug contains 2.0wt% Na2CO3
and 2.0 wt% NaCl which is in total equivalent of 0.72 meq/mL. At the beginning of chemical injection,
effective salinity at producer is still affected by brine salinity in water flooding. After chemical breakthrough
there is small decrease in effective salinity which is due to change of injecting fluid from water to ASP
with different salinity. Afterwards salinity tends to increase and finally when the polymer solution appears
at effluent it decreases to its initial value. As it can be seen from Fig. 10, the effective salinity during ASP
injection passes through optimum region.

Figure 10—Effective salinity during chemical injection for Test 1a

SPE-195564-MS 15

Fig. 11 shows the salinity profiles at 0.70 PV along the core length. At 0.70 PV polymer is being injected
at injector, therefore salinity at injector is influenced by polymer solution salinity. This time is still before
chemical breakthrough so that the effective salinity at producer corresponds essentially to initial brine
salinity. Propagation of ASP slug in the core causes the effective salinity to pass through in optimum region
and IFT reduces to ultralow value of (1.1±0.1) ×10−3mN/m. This was expected as in optimum or type III
region IFT has very low values.

Figure 11—Profiles of IFT and effective salinity at 0.70 PV for Test 1a. IFT reaches ultralow
values when effective salinity passes through optimum region during ASP injection

Comparison with other experiments: After history matching of Test 1a as the baseline experiment, the
same model was used to simulate other experiments. Several parameters were varied for each test to be in
line with the experimental conditions of each test. Due to difference in the ASP slug size and type of core,
to have a better comparison, experiments are grouped in two different categories: first, Test 1a and Test 1b
which were done with Bentheimer core and ASP slug size of 0.48 PV and second, Test 2a and Test 2b with
Berea core and ASP slug size of 0.60 PV.
Fig. 12 shows simulated and measured oil recovery for experiments with Bentheimer core. Test 1a is at
optimum and Test 1b is at under-optimum salinity. An excellent match between simulated and experimental
results was found for both tests. Simulated final oil recovery for Test 1b was found to be 67.4±0.1 % which
is approximately 1% more than value obtained from experiment. Simulated oil breakthrough for both tests
occurs at the same time at 0.45±0.02 PV. For Test 1b in the simulations oil production starts latter than
the physical experiments. The mechanism behind this is not clear. As it can be observed from Fig. 12 our
model confirms the experimental result which final oil recovery at under-optimum condition is relatively
higher than optimum condition.
16 SPE-195564-MS

Figure 12—Simulated and measured oil recovery for experiments with

Bentheimer core. Blue and red data are for Test 1a and Test 1b respectively

Figure 13—Simulated and measured oil recovery for experiments with

Berea core. Green and purple data are for Test 2a and Test 2b respectively

Fig. 13 shows the measured and simulated oil recovery for the experiments with Berea sandstone at
optimum (Test 2a) and under-optimum salinities (Test 2b). Oil breakthrough time is not affected by the
change in salt concentration and for both tests occur at 0.35±0.05 PV which is in agreement with what
was found in the laboratory. It was expected that oil breakthrough for experiments with Berea cores occurs
earlier than those with Bentheimer cores due to lower porosity. Measured and simulated final oil recoveries
for Test 2a and Test 2b are nearly equal. Simulated values are 64.8±0.1 % and 61.2±0.1 % for Test 2a and
Test 2b respectively. Therefore model confirms that although final oil recovery at optimum condition is
higher than recovery in under-optimum but they are still comparable.
SPE-195564-MS 17

Figure 14—Simulated and measured surfactant concentration for experiments

with Bentheimer core. Blue and red data are for Test 1a and Test 1b respectively

Figure 15—Simulated and measured surfactant concentration for experiments

with Berea core. Green and purple data are for Test 2a and Test 2b respectively

Fig. 14 and 15 gives simulated and measured surfactant concentration during chemical injection for
experiments with Bentheimer and Berea cores respectively. The match for all the experiments is very good.
As it can be seen from Fig. 14 and Fig. 15 no matter which type of core was used, surfactant retention was
higher at optimum condition in comparison with under-optimum condition, indicating that more surfactant
was retained in the core. Following the definition by Liyanage et al. (2012) who states that values below
0.3 mg/g rock are considered to be low, experiments performed at under optimum condition give lower
surfactant retention. A likely explanation is the enhanced aqueous solubility of the surfactant formulation at
lower salinities. Thus surfactant breakthrough for experiments at optimum occurred later than for the ones
which were performed at under-optimum conditions.
Fig. 16 and 17 shows measured and simulated alkali concentration over pore volume during ASP injection
and polymer drive. Fig. 16 shows an excellent match for the experiments performed on Bentheimer cores.
For both tests chemical breakthrough occurs at 0.80±0.02 PV which exactly coincides with what reported
in experiment. Fig. 17 shows a good agreement between simulated and measured alkali concentration for
experiments with Berea cores although they are not as good as the ones with Bentheimer cores. Breakthrough
18 SPE-195564-MS

time for Test 2b is successfully matched and occurs at 0.77±0.02 PV but for Test 2a, simulated chemical
breakthrough is found to occur later than the measured one. The reason could be attributed to the difference
in experimental conditions with model assumptions. Berea core in model is assumed to be homogenous
with equal porosity and permeability in entire core as reported by Battistutta et al. (2015). However an
analysis of Berea core in small dimension with micro CT-scanner showed that permeability changes across
the layers. Therefore it is expected to have a sequence of breakthroughs from the most permeable layer to
the least one. This is the reason why there are sharp edges in experimental alkali concentration plot as it
can be observed from Fig. 17. On the other hand according to Eq. 13 it was expected that increase in alkali
concentration and pH occur at the same time. However in experimental data increase in pH occur 0.30±0.02
PV later than increase in alkali concentration.

Figure 16—Simulated and measured alkali concentration for experiments with

Bentheimer core. Blue and red data are for Test 1a and Test 1b respectively

Figure 17—Simulated and measured alkali concentration for experiments with

Berea core. Green and purple data are for Test 2a and Test 2b respectively
SPE-195564-MS 19

Finally, Fig. 18 compares the salinity and IFT profiles at optimum (Test 1a) vs. under-optimum (Test 1b)
condition at 0.70 PV. Effective salinity for Test 1a passes through in optimum region while for Test 1b it
is always below lower salinity limit. As it can be seen from Fig. 18 IFT for the test performed at optimum
condition is lower than the one at under-optimum condition but reduction of IFT due to injection of ASP
slug at under-optimum conditions is still significant. The lowest IFT achieved for Test 1b is (5.1±0.1) ×10−3
mN/m. Although reduction of IFT for Test 1a is larger, higher oil recovery and less surfactant retention was
found for Test 1b. The findings illustrate advantages of performing ASP flood at under-optimum salinity
under conditions which experiments were performed.

Figure 18—Profiles of IFT and effective salinity at 0.7 PV for experiments with Bentheimer
core. Green and Purple curves are representative of Test 1a and Test 1b respectively

Conclusions
In this study four ASP core-flood experiments for a single olefin sulfonate (IOS) system were successfully
modeled taking into account phase behavior of the system and essential geochemical reactions. Experiments
were performed at optimum vs. under-optimum salinity conditions and with different core types
(Bentheimer and Berea).

• Modeling of IOS surfactant system with a water/crude oil ratio of 50/50 volume using UTCHEM
in batch mode successfully matched to the salinity scan experiments. Optimal salinity of 2.0 wt%
NaCl (+ 2.0 wt% Na2CO3) was observed when oil and water solubilization ratios intersect and are
equal to 30. Effect of soap generation on phase behavior was ignored due to very low acid number
of crude oil (<0.05 mg KOH/g oil).
• Aqueous reactions with alkali and cation exchange reactions between rock and cations associated
with alkali were the main reactions responsible for alkaline consumption during ASP flooding.
Using EQBATCH, the geochemical module of UTCHEM these geochemical reactions were
considered in ASP model.
• In experiments with Bentheimer sandstone, oil recoveries of 67.4±0.1% and 63.5±0.1% were found
at under-optimum and optimum salinity conditions respectively. In case of Berea sandstone, these
oil recoveries were 61.2±0.1% and 64.8 ±0.1% respectively. Therefore comparable oil recovery
was achieved at under-optimum vs. optimum salinity conditions regardless of core type. Excellent
20 SPE-195564-MS

matches were found for all the other parameters including oil cut, pressure drop and production
of chemicals at effluent. Alkali breakthrough was observed at 0.7±0.1 PV for all the experiments.
A delay in surfactant breakthrough and higher surfactant retention was observed at optimum
conditions with respect to under-optimum conditions.
• In experiments with Bentheimer sandstone, minimum achieved IFT was (5.1±0.1) ×10−3
and (1.1±0.1) ×10−3 mN/m at under-optimum and optimum salinity conditions respectively.
Considering higher oil recovery obtained at under-optimum conditions, this confirms the statement
expressed by several authors which maximal oil recovery does not necessarily coincide with
minimum IFT.
• The findings stresses advantages of performing ASP flood at under-optimum salinity conditions
due to comparable oil recovery, lower surfactant retention and reducing the risk of achieving over-
optimum salinity conditions in the field which is believed to result in high surfactant retention and
poor oil recovery.
• History-matching and calibrating the model was done using experimental data and the model can
be used to perform further simulations for prediction of the core-flood performance at different
conditions and also optimize the ASP flood design for field scale by doing up-scaling.

Nomenclature
Ap1, Ap2, Ap3 = parameter related to polymer viscosity
a31, a32, a3, b3 = surfactant adsorption parameters
Bp = effective salinity parameter for calcium
Ĉk = adsorbed concentration of species k
= total concentration of species k
k
Ckl = concentration of species k in phase l
CSEP = effective salinity
CSEL = lower effective salinity where three phase forms
CSEU = upper effective salinity where three phase disappears
HBNC70 = Height of binodal curve at zero salinity
HBNC71 = Height of binodal curve at optimal salinity
HBNC72 = Height of binodal curve at twice optimal salinity
K = permeability
KA = reaction constant for soap generation
KD = partition coefficient of acid between oil and water
= end-point relative permeability of phase p at its maximum saturation
p
NC = capillary number
(NC)c = capillary number for mobilization
(NC)max = capillary number for prevention of entrapment
np = exponent of phase p
Sp = parameter related to polymer viscosity
Spr = residual saturation of phase p
Tp = trapping parameter
µw = viscosity of water
= polymer viscosity at zero shear rate
[HAw] = concentration of acid in water
SPE-195564-MS 21

Subscripts
k = species number (1=water, 2=oil, 3= surfactant, 4=polymer, 6=divalent cations, …)
l = phase number (1=aqueous, 2=oleic, 3=microemulsion)

References
Alagic, E., Skauge, A., 2010, Combined low salinity brine injection and surfactant floooding in mixed-wet sandstone
cores. Energy Fuels 24: 3551–3559.
Alagic, E., Spildo, K., Skauge, A., Solbakken, J., 2011, Effect of crude oil ageing on low salinity and low salinity surfactant
flooding. Journal of petroleum science and engineering 78: 220–227.
Battistutta, E., van Kuijk, S. R, Groen K. V., Zitha P.L.J., 2015, Alkaline-Surfactant-Polymer (ASP) Flooding of Crude
Oil at Under-Optimum Salinity Conditions. Presented at SPE Enhanced Oil Recovery Conference, Kuala Lumpur,
Malaysia, 11-13 August 2015, SPE-174666—MS.
Bhuyan, 1989, Development of Alkaline/Surfactant/Polymer Compositional Reservoir Simulator. PhD Dissertation.
University of Texas at Austin
Bunge A.L. and Radke, C.J., 1983, Divalent ion exchange with alkali. SPE J. 23 (4): 657-668. SPE-8995-PA.
Bunge A.L. and Radke, C.J., 1985, The origin of reversible hydroxide uptake on reservoir rock. SPE J. 25 (5): 711-718.
SPE 11798-PA. 10.2118/11798-PA
Cheng, K.H., 1986 Chemical Consumption During Alkaline Flooding: A Comprehensive Evaluation. SPE Enhanced Oil
Recovery Symposium, Tulsa. 10.2118/14944-MS
Delshad, M., Delshad, M., Bhuyan, D., Pope, G.A., Lake, L.W., 1986, Effect of capillary number on the residual saturation
of a three-phase micellar solution. Paper SPE14911 presented at the SPE/DOE fifth Symposium on Enhanced Oil
Recovery, Tulsa, 20-23 April.
Delshad, M., Pope G.A, and Sepehrnoori, K., 1996, A compositional simulator for modeling surfactant enhanced aquifer
remediation. 1 formulation. J. Contamin. Hydol. 23 (4): 303-327. 10.1016/0169-7722(95)00106-9.
Delshad, M., Lenhard, R. J., Oostrom, M., Pope, G. A., & Yang, S., 1999, A Mixed-Wet Hysteretic Relative Permeability
and Capillary Pressure Model in a Chemical Compositional Reservoir Simulator. SPE Reservoir Simulation
Symposium, 14-17 February, Houston, Texas. 10.2118/51891-MS
Delshad, M., Kim, D. H., Magbagbeola, O. A., Huh, C., Pope, G. A & Tarahhom, F., 2008, Mechanistic Interpretation and
Utilization of Viscoelastic Behavior of Polymer Solutions for Improved Polymer flood efficiency. Presented at SPE
Symposium on Improved Oil Recovery, 20-23 April, Tulsa, Oklahoma, USA. 10.2118/113620-MS
Delshad, M., Han, C., Koyassan Veedu, F., & Pope, G. A., 2011, A Simplified Model for Simulations of Alkaline-
Surfactant-Polymer Floods. Paper SPE 142105 presented at the SPE Reservoir Simulation Symposium, Woodlands,
21-23 February. 10.2118/142105-MS
deZabala, E.F., Vislocky, J.M., Rubin, E., and Radke, C.J., 1982, A chemical theory for linear alkaline flooding. SPE J.
22 (2): 245-258. SPE-8997-PA. 10.2118/8997-PA.
Farajzadeh, R., Matsuura, T., van Batenburg, D., & Dijk, H., 2012, Detailed Modeling of the Alkali/Surfactant/Polymer
(ASP) Process by Coupling a Multipurpose Reservoir Simulator to the Chemistry Package PHREEQC. Society of
Petroleum Engineers Journal. 10.2118/143671-PA
Fournier R. O., Rowe J. J., 1977, The solubility of amorphous silica in water at high temperatures and high pressures.
American Mineralogist 62: 1052-1056.
Glinsmann, G. R., 1979, Surfactant flooding with microemulsions formed in situ-effect of oil characteristics. Annual
technical conference and exhibition, 23-26 september Las Vegas, NV. SPE-8326.
Hand, D.B., 1939, Dineric Distribution: I. The Distribution of a Consolute Liquid between Two Immiscible Liquids. J.
of Physics and Chem., Volume 34, p.1961–2000.
Healy, R. N., Reed, R. L., & Stenmark, D. G, 1976, Multiphase Microemulsion Systems. Society of Petroleum Engineers
Journal, 16 (03), 147-160. 10.2118/5565-PA
Healy, R. N., Reed, R. L., 1977, Immiscible microemulsionflooding. SPE 17 (2): 129–139.
Hirasaki G. J. and Pope G. A., 1974, Analysis of factors Influencing mobility and adsorption in the Flow of polymer
solution through porous media. SPE J 14 (4): 337–346.
Hirasaki, G.L, Zhang, D.L., 2004, Surface Chemistry of oil recovery from fractured, oil wet, Carbonate formations. SPE
80988, SPE Journal, 9 (2), (June 2004), P151-162
Lake, L. W., Pope, G.A., Carey, G.F., and Sepehrnoori, K., 1984, Isothermal, Multiphase, Multicomponent Fluid-Flow in
Permeable Media. Part I: Description and Mathematical Formulation
Lake, L.W., 2008, Enhanced Oil Recovery. Old Tappan, New Jersey: Prentice Hall
Larson, R. G., 1979, The influence of phase behavior on surfactant flooding. SPE SPE-6774-PA: 411.
22 SPE-195564-MS

Liyanage, P. J., Solairaj, S. Arachchilage, G. W. P., Linnemeyer, H.C., Kim, D.H., Weerasooriya, U., Pope, G.A., 2012,
Alkaline surfactant polymer flooding using a novel class of large hydrophobe surfactants. 18th SPE improved oil
recovery symposium Tulsa, SPE 154274
Martin, F.D., Oxley, J.C., 1985, Effect of various chemicals on phase behavior of surfactant/brine/oil mixtures. Paper SPE
13575 presented at the International Symposium on Oilfield and Geothermal Chemistry, Phoenix, 9-11 April.
Mohammadi, H., Delshad, M., & Pope, G. A., 2009, Mechanistic Modelling of Alkaline Surfactant Polymer Floods. SPE
Res Eval and Eng 12 (4): 518-527. SPE110212-PA. 10.2118/110212-PA
Morrow, N.R., Songkran, B., 1981, Effect of viscous and buoyancy forces on nonwetting phase trapping in porous media.
In: Shah, D.O. (Ed.), Surface Phenomena in Enhanced Oil Recovery. Plenum Press, pp. 387-411.
Morrow, N.R., Chatzis, I., Taber, J.J., 1988, Entrapment and mobilization of residual oil in bead packs. SPERE 3 (3),
927–934.
Nelson, R. C., Lawson, J.B., Thigpen, D.R., Stegemeier, G.L., 1984, Cosurfactant Enhanced alkaline flooding. Paper SPE
12672 presented at the SPE Enhanced Oil Recovery Symposium, Tulsa 15-18 April. 10.2118/12672-MS
Nelson, R. C and Pope, G. A., 1978, Phase relationships in chemical flooding. Society of Petroleum Engineers Journal.
SPE-6773-PA. 10.2118/6773-PA
Novosad, Z. and Novosad, J., 1984, Determination of alkalinity losses resulting from hydrogen ion exchange in alkaline
flooding. SPE J. 24 (1): 49-52. SPE-10605-PA. 10.2118/10605-PA
Prouvost, L., Pope, G.A., and Rouse, B.A., 1985, Micro-emulsion Phase Behaviour: A Thermodynamic Modeling of the
Phase Partitioning of Amphiphilic Species. SPE J, October issue, p. 693-703.
Reed, R. L. and R. N. Healy., 1977, Some Physico-Chemical Aspects of Microemulsion Flooding: A Review. Improved
Oil recovery by Surfactant and Polymer Flooding, D. O. Shah and R. S. Schechter (eds.), Academic Press, New York.
Sereshti, H. and Aliakbarzadeh, G., 2013, Response surface methodology optimized dispersiveliquid—liquid
microextraction coupled with UV-Vis spectrophotometry for determination of quinine. The Royal Society of Chemistry
Journal, Anal. Methods, 2013, 5, 5253–5259
Sharma, A., Delshad, M., Huh, C., & Pope, G. A., 2011, A Practical method to calculate polymer viscosity accurately
in numerical reservoir simulators. SPE. Presented at SPE Annual Technical Conference and Exhibition, 30 October-2
November, Denver. 10.2118/147239-MS
Sheng, D.-C., Yang, P.-H., Liu, Y.-L., 1994, Effect of alkali-polymer interaction on the solution properties. Petroleum
Exploration and Development 21 (2), 81-85.
Sheng, J., 2010a, Modern Chemical Enhanced Oil Recovery: Theory and practice. Gulf Professional Publishing
Sheng, J. J., 2010b, Optimum phase type and optimum salinity profile in surfactant flooding. Journal of petroleum science
and engineering 75: 143–153.
Sheng, J. J., 2013, A comprehensive review of alkaline-surfactant-polymer (ASP) flooding. Presented at SPE Western
Regional & AAPG Pacific Section meeting, Monterey, California, USA, 19-25 April 2013, SPE 165358.
Spildo, K., Johannessen, A.M., Skauge, A., 2012, Low salinity waterflood at reduced capillarity. SPE Improved Oil
Recovery Symposium. Tulsa, Oklahoma, USA SPE 154236.
UTCHEM technical documentation., 2011, UTCHEM Technical Documentation: A three-dimensional chemical
simulator.
Winsor, P.A., 1954, Solvent Properties of Amphiphilic Compounds. Butterworth Scientific Publications.
Wu. W., 1996, Optimum design of filed scale chemical flooding using reservoir simulation. PhD Dissertation, University
of Texas at Austin, Austin, Texas, USA.