1. Introduction
In the past decade, we have seen many researchers working towards the idea of smart waterflooding [
1] where modified brine having different ionic composition and reduced salinity as compared to the formation water is employed for injection purposes. This is also known as low-salinity waterflooding (LSWF) when the sole focus is on reduction in salinity. It is believed that the reduction in salinity alone leads to a better incremental recovery [
2,
3] and is enhanced when the same water has optimum concentration of potential-determining ions (PDIs) i.e., SO
42−, Ca
2+ and Mg
2+ [
4,
5]. This modified brine is known as smart brine owing to its different ionic chemistry than seawater and formation water. Moreover, studies have shown that by selectively increasing the concentration of specific ions, the surface ionic chemistry between carbonates and water results in altering the wetting state of rock to lesser oil wet, producing more oil and less water [
6]. The underlying mechanisms behind the incremental recovery produced by smart waterflooding are rock dissolution, surface ion exchange, and electrical double layer.
According to the theory of rock dissolution, an introduction of a new brine into the system leads to the dissolution of minerals containing major PDIs like CaCO
3, CaMg(CO
3)
2 and CaSO
4. This is caused by the natural tendency of rock-brine system to re-establish the ionic equilibrium disturbed when any new brine enters the system. Hiorth et al., 2010 observed a correlation between amount of calcite dissolved and incremental oil recovery by thermodynamic modelling of the electrostatic attractions [
7]. They reasoned that the calcite becomes thermodynamically unstable on introduction of new brine possibly a diluted/smart-brine and leads to more incremental oil recovery as calcite dissolves to compensate for the lack of Ca
2+ ion in the aqueous phase. On the other hand, Austad et al. [
8] strongly argued that rock dissolution contradicts the experimental results published in literature where an increase in Ca
2+ ion leads to enhanced oil recovery by inhibits the rock dissolution and implied that lower dissolution can also lead to higher recovery. Moreover, the applicability of rock dissolution at reservoir scale was also questioned by several researchers [
6,
7,
9] that rock dissolution may be dominant in regions near to the injector and may not propagate further.
During the ongoing debate of rock dissolution, many researchers have been looking at the estimation of electrostatic interactions at calcite surface based on zeta potentiometric analysis in screening underlying mechanisms of wettability alteration from low-salinity waterflooding. The zeta potential is defined as the magnitude of surface potential at the diffuse layer of an electrical double layer formed near the rock surface in the presence of an electrolyte solution [
10]. Jackson et al. reported a correlation between the determined zeta potential study using diluted brines and incremental oil recovery [
11]. They proposed that smart brine will result in improved oil recovery only when zeta potential or surface charge at interfaces, oil-brine and rock-brine is of same polarity. Awolayo and Sarma also observed the wettability alteration mechanism by low-salinity waterflooding from a thermodynamic modelling study on carbonate-oil-brine system [
1]. The stability of water film was studied by determining the disjoining pressure which is the sum of three forces, Van der Waals forces, structural forces, and electrostatic forces. The greater the magnitude of disjoining pressure more is the electrostatic repulsion between the interfaces accounted by the expansion of the double layer.
The zeta potential of a calcite surface is affected by brine properties such as concentration of PDIs, pH and ionic strength. Al-Hashim et al. found that favorable wettability state i.e., water-wetness can be achieved by addition of SO
42− and Mg
2+ ions in diluted seawater from carbonate rocks [
12]. The increasing concentration of divalent anions such as sulfate and phosphate causes the zeta potential to become more negative while divalent cations results in polarity reversal at the rock surface [
13].
There seems to be no consensus on the accepted underlying mechanism to be rock dissolution or electrical interactions involving electric double layer. As it is important to better understand the interactive mechanisms to lower the risk of LSWF mechanisms before field implementation, this paper aims to assess the effect of ionic chemistry and reduced salinity to evaluate different mechanism of wettability alteration. Zeta potentiometric study has been carried out using oil, water and rock samples collected from field. Moreover, a simulation model developed to further study laboratory corefloods of carbonate rocks with various brines has been conducted and various parameters such as pH change, mineral variation of the moles of calcite, change in ionic concentration in aqueous phase has been critically analysed to assess each underlying mechanism.
2. Materials and Methods
2.1. Preparation of Samples for Zeta Potentiometric Study
2.1.1. Rock Samples
Whole core samples of carbonate rock were drilled to produce one-inch core plugs and saturated in clean oil for 3–4 weeks. The plug ends were square-cut and polished using a grinder. An x-ray diffraction and x-ray fluorescence analysis were conducted on pulverized rock samples. The results, as shown in
Figure 1, suggest that the primary mineralogy of the rock is calcite with trace amounts of quartz and dolomite.
2.1.2. Fluid Samples
Oil and sea water samples collected from a field location in an offshore basin off the west coast of India were shipped to the laboratory in separate canisters. The important properties of these samples were analyzed using various instruments which measured the pH, salinity, ionic strength, and ionic composition of water. Ion chromatography (930 Compact IC Flex, Metrohm, Calgary, AB, Canada) was used to analyze the brines’ ionic chemistry including the concentrations of anions and cations using appropriate solvents. Nitric acid and bicarbonate solutions were used for cationic and anionic ion chromatography. Different levels of dilution were studied for both zeta potentiometric and reservoir simulation study. Dilution was carried out from 50% up to 1% by successive addition of deionized water.
Table 1 shows the important properties of sea water and its diluted versions used in this study and similar properties of modified-brines are shown in
Table 2. As shown in
Table 2, a modified-brine is prepared by selectively changing the concentration of PDIs in sea water such that individual variation as wells as combined variation of PDIs can be studied. Salts such as sodium sulfate, calcium chloride and magnesium chloride were added in appropriate amounts to raise the concentration of individual PDI.
2.2. Experimental Design for Zeta Potential Analysis
This section describes the experimental design for zeta potentiometer and reservoir simulation adopted to study the mechanism of low-salinity waterflooding. In order to estimate zeta potential at rock surface, an estimation of classical streaming potential or a classical streaming current is required in the presence of flowing electrolyte solution across a portion of rock surface.
One-inch diameter core plugs were drilled from whole core samples and were used for zeta potentiometric analysis capable of holding up to four-inch core samples. A mixture of organic solvents such as 2-propanol, dichloromethane, toluene and xylene were used to clean core plugs via a Soxhlet extraction apparatus. Cleaning of core plugs was done to ensure that any high molecular weight hydrocarbons present in the core were removed as they can plug the porous channels and cause unrepresentative estimations of zeta potential.
Helmholtz-Smoluchowski (HS) gave the equation for estimating zeta potential (ζ) by a streaming current approach as shown in Equation (1) [
14]:
where,
dI/
dp is the measured streaming current coefficient,
η is the viscosity of electrolyte solution,
ε0 is the permittivity of the vacuum/free space taken as 8.85 × 10
−12 F/m,
ε is the dielectric constant of the electrolyte solution,
L/
A is the cell constant of the streaming channel (the gap between adjacent solid samples where
L, defines the length of the rectangular-shaped slit channel formed between two planar surfaces and
A, its cross-sectional area.
Although the streaming potential technique can also be used to determine zeta potential, it does require the determination of an additional parameter i.e., electric resistance across the streaming channel. On the contrary, streaming current equation given in Equation (1) does not require the measurement of electric resistance which in turn reduces the measurement error associated with it. Moreover, due to porous nature of our rock sample, electrical resistance measured is not representative due to addition of electrical resistance of pore fluid. Therefore, streaming current technique was used for current study.
A zeta potentiometer (Model: SurPASSTM 3) apparatus designed by Anton Paar (Calgary, AB, Canada) equipped with a special mechanism to hold cylindrical rock samples was used to measure streaming current coefficient where the required streaming channel can be developed for this study. Special care was taken in handling of rock and fluid samples during each measurement as the value of streaming current is very small and highly sensitive to small disturbances. Intake of any size of air bubble needs to be accounted for as it can lead to a completely unrepresentative estimation of zeta potential.
2.3. Simulation Model and Properties
The equation-of-state compositional simulator (GEM
TM CMG 2019, Calgary, AB, Canada) used in this study has been developed by Dang et al. [
15] where an ion exchange model capable of incorporating geochemical reactions including both the aqueous phase reaction and mineral reactions. The results from the low salinity flooding model were validated with experimental data and ion-exchange model results from geochemistry software PHREEQC.
2.3.1. Aqueous and Mineral Reactions
The aqueous and mineral reaction used in the simulation study are shown in
Table 3.
The simulator works by moving the above reactions either forward or backward to re-establish the equilibrium that gets disturbed when a new brine is injected. Following condition is satisfied for any reaction at its equilibrium state [
15]:
where, Q
α and K
eq are the activity product and the chemical equilibrium constant for any reaction respectively, R is the universal gas constant, R
aq is the number of aqueous phase reactions, n
aq is the number of specific ions in the aqueous reaction, a
k is the activity of particular ion, v
kα is the stoichiometric coefficient of a component in any chemical reaction, and T is the temperature.
In a solution, activity of each ionic species (α
k) is given by their effective concentration in a solution. Activity for ith ionic species is given below:
where, ƴ
i is the activity coefficient of the ionic species which is taken approximately equal to its concentration in molalities and m
i is the concentration of the ionic species in molalities.
The equilibrium constant for various aqueous reactions as a function of temperature is calculated and reported by various researchers [
1,
16,
17] in the literature. Awolayo et al., used the following expression to calculate the value of equilibrium constant using following equation at various temperatures:
where, T is the temperature in Kelvin and A
0, A
1, A
2, A
3, and A
4 are the empirical parameters found in multiple databases for a particular chemical reaction.
In case of mineral reactions which are kinetically controlled in a reservoir, the consequence of any mineral reaction will be either dissolution or precipitation until the state of ionic equilibrium with the aqueous phase is achieved. The rate of reaction and the reaction rate constant as a function of temperature are given by following equations [
15]:
where, Â
β is the reactive surface area of mineral (β) per unit bulk volume (BV) of porous medium (m
2/m
3), k
β is the rate constant of mineral reaction (mol/m
2 s), K
eq,β is the chemical equilibrium constant of mineral reaction, Q
β is the activity product of mineral (β) dissolution reaction, r
β is the dissolution/precipitation rate per unit BV of porous medium [mol/(m
3·s)], R
mn is the number of mineral reactions, E
a is the activation energy of any reaction (J/mol), k
0,β is the reaction-rate constant at a reference temperature for a reaction, R is the universal gas constant (8.314 J/(mol K)), and T and T
0 are the temperature (K) and reference temperature (K) respectively.
Dang et al. [
15] also used the ratio of ratio of number of moles of mineral per grid block volume at current time to time zero to find the change the reactive surface area of a mineral in their CMG GEM
TM model.
2.3.2. Grid Model and Petrophysical Properties
A single porosity model as illustrated in
Figure 2 was prepared to simulate coreflooding with a 12 × 1 × 1 inch composite core. The dimensions of the core were representative of the actual dimensions of composite core used in laboratory experiments. The grid model implemented a 50 × 1 × 50 grid blocks in each direction. The direction along the
x-axis was used to represent direction of coreflood. The fluid property and PVT simulator package (WinProp) developed by Computer Modelling Group (CMG) was used to prepare a fluid model for this simulation study (
Table 4). It is necessary to emphasize here that the actual fluid and rock properties measured in lab were used to prepare the model as shown in
Table 5 and
Table 6, respectively.
The component analysis of oil samples was analyzed from ASTM D2887 simulated distillation. The C7+ hydrocarbon component proportions obtained from the analyses were combined into one component as given in
Table 5. The acid number of the oil is ~0.6 mg KOH as reported by Kakati et al. [
18]. For same set of rock and oil samples, Singh et al. [
19] had conducted a contact angle measurement using sessile drop method which showed a very weak water-wet state with seawater (contact angle of ~102°). This also correlates with the high acid number of oil.
The injector has been placed in the left-end block and the producer in the right-end block as shown in
Figure 2. Flow rate in the injection well is kept at a constant value of 0.1 mL/min while the simulation run-time period is two days. Several simulations in secondary recovery mode with diluted and smart brines have been conducted to understand the wettability altering mechanisms such as rock dissolution, electric double layer and surface ion exchange. Parameters such as pH, number of moles of ionic species and mineral, have been studied as a function of distance from injector to producer and pore volume injected. Further discussions are done on results obtained from zeta potential experiments and their integration with simulation study.
4. Conclusions
In this paper, a critical analysis of the wettability alteration mechanism mainly rock dissolution and electric double layer are presented by a combined study of zeta potential estimation and lab-scale simulation studies. Considering the experimental limitations and the assumptions made during the study, the following conclusions are drawn from our observations and analyses:
- (1)
Additional oil recovery is obtained from low-salinity waterflooding as observed from the increasing trend of zeta potential and oil recovery. Maximum incremental oil recovery is given by 1% dSW which correlated with the analysis of pH, moles of calcite, saturation index and Ca2+ ion profiles presented in this paper showed that rock dissolution as the underlying mechanism.
- (2)
The impact of PDIs has been studied in this paper and following conclusions can be drawn from it: (a) For enhanced oil recovery, brine spiked with Ca2+ ion is preferred such that the magnitude of zeta potential is close to zero which ensures minimum attractive forces between adhered oil molecules and the rock surface. (b) Less active role is observed for Mg ion as compared to Ca ion due to the relatively lower solubility of MgSO4 in brine at ambient conditions. (c) Brines can also be spiked with divalent anions, SO42+ and PO43− owing to the oil detachment from repulsive forces due to their adsorption.
- (3)
Additional oil recovery of 20.5% resulted from seawater depleted of Na+ and Cl− ions from sea water whereas seawater depleted of only Ca2+ ions gave no additional oil. The underlying mechanism is explained by the resultant increase in electrostatic repulsive forces from expansion of electrical double layer but needs to be further confirmed by surface complexation modeling of surface and electrostatic interactions.
- (4)
As seen from simulation study, only the injection gridblock had evidence of calcite dissolution taking place. Therefore, the applicability of rock dissolution mechanism in field conditions is highly doubtful as it may not propagate further into the reservoir.
The work conducted in this study can be further extended by developing a surface complexation model to study wettability alteration mechanism by predicting disjoining pressure and zeta potential for various LS and smart brines. Furthermore, a field-scale simulation model can also be developed using geochemical reactions to optimize concentration of PDIs in injection brine at field scale.