Optimal Design of Alkaline-Surfactant-Polymer Flooding under Low Salinity Environment.

This paper presents an optimal design of alkaline-surfactant-polymer (ASP) flooding and an experimental analysis on the effects of ASP components under low formation salinity, where the assignment of salinity gradients and various phase types are limited. The phase behavior and coreflooding tests confirmed the ASP formula is optimal, i.e., 1 wt % sodium carbonate (Na2CO3) as the alkaline, 1:4 eight ratio for linear alkylbenzene sulfonate (LAS) and dioctyl sulfosuccinate (DOSS) as a surfactant, 5 wt % diethylene glycol monobutyl ether (DGBE) as a co-solvent, and hydrolyzed polyacrylamide (HPAM) as a polymer. The salinity scan was used to determine that the optimum salinity was around 1.25 wt % NaCl and its solubilization ratio was favorable, i.e., approximately 21 mL/mL. The filtration ratio determines the polymer concentrations, i.e., 3000 or 3300 mg/L, with a reduced risk of plugging through pore throats. The coreflooding test confirmed the field applicability of the proposed ASP formula with an 86.2% recovery rate of residual oil after extensive waterflooding. The optimal design for ASP flooding successfully generated phase types through the modification of salinity and can be applicable to the low-salinity environment.


Introduction
Alkaline-surfactant-polymer (ASP) flooding intends to integrate a synergy of chemical mixtures, e.g., alkalis, surfactants, co-solvents, and polymers, in order to recover residual oil [1][2][3][4][5]: alkalis and surfactants mobilize residual oil and reduce the interfacial tension between the displacing phase and the oil; polymer slug enhances mobility ratios and volumetric sweep efficiencies. Successful ASP flooding can be achieved by designing an optimal ASP formula that is closely related to phase types, i.e., microemulsion, such as Winsor types I (II-), II (II+), and III, which formation salinity influences significantly. The design defines the optimum salinity showing the equal solubilization of oil and water and determines the amounts and types of chemical mixtures that can be used to achieve the desired performances, e.g., to adequately reduce the interfacial tension. The lowest interfacial tension and the middle-phase microemulsions, i.e., Winsor type III, are available at a near-optimum salinity, so that it has been a key guideline in the design of ASP flooding.
A low-salinity environment, i.e., the salinity of formation brine is below 1 wt % NaCl [6], causes the challenge of generating various phase (microemulsion) types and salinity gradients. Generally, Table 1.
Components of crude oil sampled at the target reservoir planned to conduct alkaline-surfactant-polymer (ASP) flooding.

Experimental Procedure
The experiments were divided into the following tests: the phase behavior test to find the best-performing formulation of the ASP elements, the filtration ratio (FR) test to examine the polymer transport stability, and the coreflooding test to estimate the actual flow performances of the ASP formula determined. The alkali was selected to be sodium carbonate (Na 2 CO 3 ; Daejung Chemicals, Goryeong, Korea), of which its effectiveness has been validated [4,14,15]. The polymer was hydrolyzed polyacrylamide (HPAM), Alcoflood 955 polymer with a molecular weight around 4.5 million Dalton (Ciba Specialty, Basel, Switzerland). Sodium chloride (NaCl; Daejung Chemicals, Goryeong, Korea) was utilized to make synthetic brine. Two anionic surfactants, i.e., linear alkylbenzene sulfonate (LAS) and C8-C8 double-tail dioctyl sulfosuccinate (DOSS) (Akyung Chemical, Seoul, Korea), were utilized in this study. LAS is a clear liquid surfactant with good thermal stability and is extensively used as a household detergent, while DOSS has been commonly used for an oil spill dispersant [16,17]. The LAS-DOSS combination is effective in low-salinity reservoirs, because the existence of DOSS can be expected to decrease the optimum salinity of LAS, i.e., the optimum salinity of the surfactant mixture becomes lower. Isobutyl alcohol (IBA; Sigma-Aldrich, St. Louis, MO, USA) and diethylene glycol monobutyl ether (DGBE; Daejung Chemicals, Goryeong, Korea) were prepared as co-solvents. DGBE has the longer carbon chain than IBA. Both solvents have similar densities, that is, 0.955 g/cm 3 for DGBE and 0.803 g/cm 3 for IBA. The melting and boiling temperatures of IBA are lower than those of DGBE. The flashing temperature of DGBE is 100 • C, and the flashing temperature of IBA is 73 • C, which is lower than that of DGBE. Moreover, 8.7 g of IBA was capable to dissolve in 100 g water, and DGBE is a soluble substance in water, ethanol, ethyl ether, and acetone. Tables 2 and 3 summarize the properties of available surfactants and co-solvents (LAS and DOSS were prepared as surfactants, and IBA and DGBE were prepared as co-solvents).  By using the phase behavior test, phase types with salinity changes were observed, and the optimum salinities at the same solubilization ratio for oil and water were determined. It utilized serial 5 mL with 0.1 mL markings silicate pipettes (Witeg Diffico, Wertheim, Germany). All the aqueous solutions and crude oil were added precisely into pipettes using an electric pipette dispenser (Eppendorf International, Hauppauge, NY, USA). The salinity was varied by diluting NaCl with deionized water from Direct-Q Millipore (Youngin, Seoul, Korea). Brine was then added into a pipette with a 1:1 (weight ratio; wt/wt) aqueous phase (alkali and surfactant mixture)-to-crude oil ratio. The top of the glass was sealed with silicon grease to separate water from the volatile component inside the pipette. All pipettes were shaken up for about 2 min and arranged in a pipe rack in ascending order of brine salinity. The volumes of water, oil, and microemulsion phases were measured at the equilibrium condition, when there was no significant alteration of phase volume observed. Equation (1) was used to calculate the solubilization ratios for oil and water, as shown below: where σ (mL/mL) represents the solubilization ratio and V (mL) is the phase volume within the microemulsion at the equilibrium; the subscripts o, w, and s denote the phases of the oil, water, and surfactant, respectively. Equation (2) was used to estimate the interfacial tension at the optimum salinity [18]: where σ represents the interfacial tension (mN/m) and σ opt is the water or oil solubilization ratio determined at the optimum salinity. The criterion of the solubilization ratio at the optimum salinity is higher or equal to 10 mL/mL, in order to accomplish the desired ultralow interfacial tension, i.e., when the oil solubilization ratio is 10 mL/mL, the interfacial tension is 0.003 mN/m [2,13,19].

FR Test
The FR reflects whether the polymer solution is free of aggregates, which yields a plugging phenomenon at the pore throats. Two different concentrations of the polymer, i.e., 3000 and 3300 mg/L, were assessed. The experimental apparatus consisted of a 2 µm filter membrane representing a porous medium, a filtration chamber, and a volumetric flask, which acted as an effluent collector. Both polymer solutions were prepared using 200 mL each. The polymer was poured into the filtration chamber equipped, which had the filter membrane. The filtration chamber was connected to a nitrogen bottle with a tubing line. Nitrogen was utilized to displace the polymer inside the filtration chamber and establish flow through the filter membrane. The experiments were carried out under a constant pressure, 25 psi (lb/in 2 ). The volumetric flask collected the effluent volume from the chamber, and the time was recorded for every 20 mL incremental of the effluent volume. Nitrogen was injected continuously, until all samples were completely drained from the filtration chamber. Deionized water was used to flush the filtration chamber and thus changed the polymer concentrations. The FR was written as: where t denotes the time recorded at a specific effluent volume and the subscripts 100, 200, and 300 mL represent the effluent volumes. A FR value less than 1.2 is commonly acceptable for nonoccurrence of polymer hydration [20]. Table 4 shows the properties of two cylindrical Berea sandstone cores used in the coreflooding test. Both cores had similar rock properties, e.g., the average absolute permeability ranges from 180 to 200 millidarcy (md). They were placed in an epoxy coreflooding system, since the adopted pressure was relatively below 100 psi ( Figure 1). In the epoxy system, both ends of the rock samples were fixed with sealing tube-fitting caps. Each cap was glued with a quick drying epoxy paste and left to dry for 30 min, until the caps adhered tightly to the ends of core samples. An aluminum foil with silicone sealant covered the surface of the rock sample to prevent direct contact between the epoxy and the core surface. The core was placed inside a polycarbonate cylinder, where one end was glued on a paper sheet with silicon to avoid leakage during the pouring process. Once the silicone was hardened, a low-set epoxy comprised of a 2:1 (wt/wt) resin-to-hardener ratio was poured into an annulus between the core and the polycarbonate cylinder. The epoxy covered all rock surfaces to ensure that no air was trapped inside the epoxy. This epoxy core holder system was left to dry at 20 • C for a day to harden the epoxy. Two valves, used to control the fluids, were attached at both ends of the epoxy core after hardening the epoxy. Two Teledyne ISCO syringe pumps (Lincoln, NE, USA) were used to inject aqueous fluids and crude oil into the epoxy core system. As shown in Figure 1, water was injected as a driving fluid into a crude oil chamber. The driving fluid pushed crude oil inside the chamber towards the epoxy core system. Crude oil was injected into the epoxy core to set up the initial reservoir condition. Before starting the tests, a vacuum pump was used to evacuate any trapped air within epoxy cores. An absolute-pressure gauge was used to measure the inlet pressure, and a differential-pressure transducer was adopted to determine the pressure difference between the inlet and the outlet.

Optimal ASP Formulation and Polymer Concentration
Two formulas, i.e., PB5 and PB6, showed the optimum solubilization ratio over 10 mL/mL, so that they were examined as the candidates of coreflooding tests (Table 5). Figure 2 depicts the changes of oil and water solubilization ratios and aqueous salinities for PB5 and PB6. The difference between two cases was negligible; a 0.1 wt % surfactant could increase the optimum solubilization ratio from 21.5 to 25.0. However, the effect was insignificant enough to fall in the interpretation error range ( Figure 2). The results of the phase behavior test confirmed the optimum ASP formulation: 1 wt % Na2CO3 as the alkali, 1 wt % 1:4 (wt/wt) LAS:DOSS mixture as the surfactant, and 5 wt % DGBE as the co-solvent. The phase distribution (the lower part of Figure 2) showed the region of Winsor type III that would be possibly present in the salinity range of 0.8 and 1.5 wt % NaCl. Winsor type I was expected during the preflush and the polymer drive processes, because the formation salinity was 0.6 wt % NaCl. The high salinity should be assigned to the ASP slug to generate type III.
Notably, DGBE (a glycol ether type) was more effective in increasing the optimum solubilization ratio than IBA (an alcohol type). The amounts of the surfactant and the alkali had insignificant influences on the solubilization ratios as well as the optimum salinity. The surfactant went to the micellar interface, whereas the co-solvent partitioned the oil and brine interface so that the co-solvent at the interface influenced efficiently the microemulsion phases [19]. DGBE had a higher molecular weight and the aqueous stability limits compared to IBA, which can yield a higher solubilization ratio [13,21]. Figure 3 presents the FR and the slope of the observed points for the polymer concentrations, i.e., 3000 and 3300 mg/L. It can be seen that the effluent volume had a linear relationship with the time, demonstrating the slop of the plot, i.e., the FR, was constant. The FRs were 1.0 for a polymer concentration of 3000 mg/L and 1.08 for a polymer concentration of 3300 mg/L. Figure 3 proves that Extensive waterflooding was carried out using the formation brine (0.6 wt % NaCl), until no oil was recovered. Brine saturated the epoxy core system, and then crude oil was injected into the rock core at a constant pressure of 80 psi from the top to the bottom of the vertical epoxy core. The reason for the downward injection is that the differential density between brine and crude oil was able to accelerate the oil flow rather than the opposite case. The oil injection continued, until no brine was observed at the outlet, and then the rock core was aged inside an oven at 80 • C for 72 h to alter the wettability from water-wet to mixed-wet conditions. After the completion of the aging process, formation brine was injected continuously into the core as a preflush fluid. The injection volumetric velocity was maintained at 0.24 cm 3 /min (approximately 1 ft/day superficial velocity), until no oil was recovered [2]. The oil cut and residual oil recovery were observed. Oil cut was defined as the ratio of a produced oil volume to a total fluid production, while the cumulative residual oil recovery ( f op ) was defined as Equation (4): where V oi (mL) is the produced oil volume in the tube and V orw (mL) represents the remaining oil volume after the waterflooding.

Optimal ASP Formulation and Polymer Concentration
Two formulas, i.e., PB5 and PB6, showed the optimum solubilization ratio over 10 mL/mL, so that they were examined as the candidates of coreflooding tests (Table 5). Figure 2 depicts the changes of oil and water solubilization ratios and aqueous salinities for PB5 and PB6. The difference between two cases was negligible; a 0.1 wt % surfactant could increase the optimum solubilization ratio from 21.5 to 25.0. However, the effect was insignificant enough to fall in the interpretation error range (Figure 2). The results of the phase behavior test confirmed the optimum ASP formulation: 1 wt % Na 2 CO 3 as the alkali, 1 wt % 1:4 (wt/wt) LAS:DOSS mixture as the surfactant, and 5 wt % DGBE as the co-solvent. The phase distribution (the lower part of Figure 2) showed the region of Winsor type III that would be possibly present in the salinity range of 0.8 and 1.5 wt % NaCl. Winsor type I was expected during the preflush and the polymer drive processes, because the formation salinity was 0.6 wt % NaCl. The high salinity should be assigned to the ASP slug to generate type III. both were less than 1.2 (the criterion of FRs). In summary, the optimum ASP formula can be 1 wt % Na2CO3 as the alkali, 1 wt % 1:4 (wt/wt) LAS:DOSS mixture as the surfactant, and 5 wt % DGBE as the co-solvent. The optimal design of ASP flooding required a high salinity of ASP slug with over 1.5 wt % NaCl considering the low salinity environment. The polymer concentrations should be 3000 or 3300 mg/L without any plugging at pore throats.   Notably, DGBE (a glycol ether type) was more effective in increasing the optimum solubilization ratio than IBA (an alcohol type). The amounts of the surfactant and the alkali had insignificant influences on the solubilization ratios as well as the optimum salinity. The surfactant went to the micellar interface, whereas the co-solvent partitioned the oil and brine interface so that the co-solvent at the interface influenced efficiently the microemulsion phases [19]. DGBE had a higher molecular weight and the aqueous stability limits compared to IBA, which can yield a higher solubilization ratio [13,21]. Figure 3 presents the FR and the slope of the observed points for the polymer concentrations, i.e., 3000 and 3300 mg/L. It can be seen that the effluent volume had a linear relationship with the time, demonstrating the slop of the plot, i.e., the FR, was constant. The FRs were 1.0 for a polymer concentration of 3000 mg/L and 1.08 for a polymer concentration of 3300 mg/L. Figure 3 proves that these concentrations had a lower risk of plugging at pore throats, i.e., the polymer hydration, since both were less than 1.2 (the criterion of FRs).

Coreflooding Tests
Coreflooding experiments were conducted to assess the field applicability of ASP formulas. The detailed designs of ASP flooding were as follows. The preflush using 0.6 wt % NaCl , which was used for the formation brine, was carried out, until no oil was recovered. The salinity of the ASP mixture was set as 1.5 wt % NaCl to make Winsor type III, since the residual brine reduced the ASP salinity but maintained it over 0.6 wt % NaCl. Two different amounts of ASP slug were examined, i.e., 0.37 pore volume (PV; a large volume of ASP slug) and 0.10 PV (a small volume of ASP slug). The polymer drive followed the ASP flooding with the 0.6 wt % NaCl salinity, and therefore its phase was type I. The polymer concentrations were 3000 mg/L in the ASP mixture and 3300 mg/L for the polymer drive. The sequence of phase types included in the preflush, the ASP mixture, and the polymer drive would be I-III-I. Table 6 summarizes the detailed designs of ASP coreflooding tests and the results. Figure 4 demonstrates the oil cut, and Figure 5 describes the cumulative residual oil recovery with different ASP injection volumes, i.e., PV injection. A large amount of ASP slug was helpful in the production of more residual oil; the total oil recovery was 86.20% of the residual oil in the case of the 0.37 PV injection, while with the 0.10 PV injection the oil up to 48.96% was recovered. If the ASP mixture was injected into the core sample to generate a microemulsion and a lower interfacial tension, the residual oil could be recovered more. This study does not generate type II in the preflush, since the low-saline water flowed to demonstrate the field condition. It would be limited to compare the oil recovery in the case of type II-III-I. Large amounts of polymer displaced the oil bank (microemulsion) effectively to show the cumulative oil recovery rate no longer increased ( Figure 5).
The results confirmed that the proposed ASP flooding system had positive synergy impacts on unrecovered oil after the waterflooding process. The first oil began at the end of the ASP injection, and the peak of the oil cut was observed in the region of the polymer drive. This performance meant that the ASP mixture mobilized the residual oil and turned it to a producible microemulsion. The polymer efficiently drove this mobilized oil. The lowest surfactant retention may be found in the type I microemulsion, since the surfactant was soluble in the water phase forming a water-external microemulsion, where the interfacial tension was high but easy to displace. On the other side, the In summary, the optimum ASP formula can be 1 wt % Na 2 CO 3 as the alkali, 1 wt % 1:4 (wt/wt) LAS:DOSS mixture as the surfactant, and 5 wt % DGBE as the co-solvent. The optimal design of ASP flooding required a high salinity of ASP slug with over 1.5 wt % NaCl considering the low salinity environment. The polymer concentrations should be 3000 or 3300 mg/L without any plugging at pore throats.

Coreflooding Tests
Coreflooding experiments were conducted to assess the field applicability of ASP formulas. The detailed designs of ASP flooding were as follows. The preflush using 0.6 wt % NaCl, which was used for the formation brine, was carried out, until no oil was recovered. The salinity of the ASP mixture was set as 1.5 wt % NaCl to make Winsor type III, since the residual brine reduced the ASP salinity but maintained it over 0.6 wt % NaCl. Two different amounts of ASP slug were examined, i.e., 0.37 pore volume (PV; a large volume of ASP slug) and 0.10 PV (a small volume of ASP slug). The polymer drive followed the ASP flooding with the 0.6 wt % NaCl salinity, and therefore its phase was type I. The polymer concentrations were 3000 mg/L in the ASP mixture and 3300 mg/L for the polymer drive. The sequence of phase types included in the preflush, the ASP mixture, and the polymer drive would be I-III-I. Table 6 summarizes the detailed designs of ASP coreflooding tests and the results. Figure 4 demonstrates the oil cut, and Figure 5 describes the cumulative residual oil recovery with different ASP injection volumes, i.e., PV injection. A large amount of ASP slug was helpful in the production of more residual oil; the total oil recovery was 86.20% of the residual oil in the case of the 0.37 PV injection, while with the 0.10 PV injection the oil up to 48.96% was recovered. If the ASP mixture was injected into the core sample to generate a microemulsion and a lower interfacial tension, the residual Polymers 2020, 12, 626 9 of 11 oil could be recovered more. This study does not generate type II in the preflush, since the low-saline water flowed to demonstrate the field condition. It would be limited to compare the oil recovery in the case of type II-III-I. Large amounts of polymer displaced the oil bank (microemulsion) effectively to show the cumulative oil recovery rate no longer increased ( Figure 5). ASP mixture generated a type III microemulsion and also a type I microemulsion created by the low salinity of the polymer drive would be positive under the low-salinity environment.   Polymers 2020, 12, 626 9 of 11 ASP mixture generated a type III microemulsion and also a type I microemulsion created by the low salinity of the polymer drive would be positive under the low-salinity environment.   The results confirmed that the proposed ASP flooding system had positive synergy impacts on unrecovered oil after the waterflooding process. The first oil began at the end of the ASP injection, and the peak of the oil cut was observed in the region of the polymer drive. This performance meant that the ASP mixture mobilized the residual oil and turned it to a producible microemulsion. The polymer efficiently drove this mobilized oil. The lowest surfactant retention may be found in the type I microemulsion, since the surfactant was soluble in the water phase forming a water-external microemulsion, where the interfacial tension was high but easy to displace. On the other side, the type III microemulsion made the surfactant soluble in both oil and water, but the retention was higher than in the type I microemulsion. Thus, the coreflooding results showed that the high salinity of the ASP mixture generated a type III microemulsion and also a type I microemulsion created by the low salinity of the polymer drive would be positive under the low-salinity environment.

Conclusions
This paper suggested an optimal design of ASP flooding, i.e., ASP formulation, the salinity assignment for ASP components, and polymer concentrations, in order to recover the residual oil under the low-salinity environment. The optimal ASP formula consisted of 1 wt % Na 2 CO 3 as the alkali, 1 wt% 1:4 (wt/wt) LAS:DOSS mixture as the surfactant, and 5 wt % DGBE as the co-solvent. The phase behavior test showed that the glycol-ether-type co-solvent with a large molecular weight could be effective and have a high aqueous stability limit on the optimum solubilization ratio. The type I-III-I sequence of phase types could recover the residual oil up to 86%, and the polymer drive with Winsor type I played a crucial role in recovering a three-phase microemulsion mobilized by the ASP mixture. The experimental results validated that the proposed design of ASP flooding would be effective in displacing the remaining oil that was not recoverable through waterflooding, despite the fact that low salinity limited the designs of salinity gradients and phase types.

Conflicts of Interest:
The authors declare no conflicts of interest.