Effect of Live Crude on Alkaline/Surfactant Polymer Formulations: Implications for Final Formulation Design

Transcription

1 Effect of Live Crude on Alkaline/Surfactant Polymer Formulations: Implications for Final Formulation Design Jeffrey G. Southwick, Yi Svec, Greg Chilek, and Gordon T. Shahin, Shell International Exploration & Production Summary Phase-behavior experiments have identified several surfactant systems that develop high solubilization ratios and low interfacial tension (IFT) with a specific dead paraffinic crude oil at specific salinities. The purpose of this work is to test these surfactant systems with reconstituted live crude. Emulsion-screening tests were performed in sight cells where an equilibrium amount of solution gas is dissolved in the crude at reservoir pressure (1,100 psi). The results indicate that the surfactant is relatively more soluble in the oil phase under these conditions. Thus, a formulated chemical slug for field conditions should contain either less salinity or a more hydrophilic surfactant system than that used in formulations with dead crude. Phase-behavior measurements estimate this offset to be approximately 0.25% less NaCl for the particular live crude in this study. The relevance of this offset is shown by comparing the results of dead-crude corefloods with a live-crude coreflood. A control experiment pressurizing oil with nitrogen at the same condition, 1,100 psi, did not show enhanced relative surfactant solubility in the oil phase. Introduction Formulations for either alkaline/surfactant/polymer (ASP) flooding or surfactant/polymer flooding are often developed in the laboratory with dead crude oil (stock-tank oil) at atmospheric pressure. The dead crude oil is chemically different from the oil in the reservoir because the naturally occurring gases (methane, ethane, and similar gases) are no longer dissolved in the crude. The reasons for working with dead crude in the laboratory are efficiency and safety. Experiments with live crude must be performed at elevated pressures with the correct quantity of gas dissolved to mimic reservoir conditions. Because chemical-formulation studies for field trials do not routinely mention working with live crude, the tacit assumption is being made that the difference in performance of a surfactant system between dead and live crude is not significant and can be ignored. Thus, chemical formulations developed with dead crude in the laboratory are applied in the field, where the surfactant system is expected to dramatically reduce IFT and to mobilize live crude. Experimental Data and Theroretical Approaches There is not a large amount of data in the literature on the effect of dissolved gas on phase behavior. Nelson (1983) published data that showed little difference in phase behavior as a function of equilibrated dissolved methane at 2,050 psi. It was argued that two competing effects are at work: The decrease in molar volume of the crude oil owing to dissolution of methane increases surfactant solubility in the oil, whereas the decreased Hildebrand solubility parameter of the crude oil from the dissolution of methane reduces surfactant solubility in the oil. These are opposing influences, and theoretically support the experimental observation of no change Copyright 2012 Society of Petroleum Engineers This paper (SPE ) was accepted for presentation at the SPE Annual Technical Conference and Exhibition, Florence, Italy, September 2010, and revised for publication. Original manuscript received for review 2 September Revised manuscript received for review 27 September Paper peer approved 10 October in midpoint salinity. (Midpoint salinity is defined as the salinity where the middle-phase microemulsion contains equal volumes of oil and water; this salinity is the point where minimum IFT is realized and is sometimes referred to in the literature as optimum salinity ). It was emphasized that this result was found for a specific crude oil, and that a general statement about the effect of dissolved methane on surfactant systems is not valid. Puerto and Reed (1983) observed that when an oil was pressurized with methane, the midpoint salinity decreased. In later experimental work, Austad and Strand (1996), Austad et al. (1996), and Skauge and Forland (1990) published data showing that pressure affects the midpoint salinity of dead crude. In these studies, increased pressure reduced relative surfactant solubility in the oil, which in turn raises midpoint salinity. More recent data from Kahlweit et al. (1988), Sassen et al. (1991), and Roshanfekr et al. (2009) confirm the observation that pressure reduces relative surfactant solubility in crude oil, but the latter authors also showed that pressurizing crude oil with 17 mol% of methane increased relative surfactant solubility in the oil. Thus, although there exist competing physical influences on surfactant solubility, for the crude tested, the dissolution of methane dominated the effect of pressure; midpoint salinity with the live crude was 0.25% NaCl lower than with the dead crude. It is doubtful that all crude oils behave similarly with respect to methane dissolution, so we recommend that experiments be performed with the crude oil of interest before formulating conclusions are reached. We wish to emphasize that it is the relative solubility of the surfactant in oil and brine that controls phase behavior, not the absolute solubility. Pressurizing a surfactant/brine/oil system at midpoint salinity with methane may increase the solubility of the surfactant in both the oil and the brine, but if it increases the solubility of the surfactant in the oil to a greater extent than in the brine, the phase behavior shifts toward overoptimum. In that circumstance, the system can be returned to midpoint (optimum) salinity by decreasing the salinity of the brine. Thus, increased relative solubility of the surfactant in oil decreases the salinity requirement of the system. The effect of pressure alone increases the relative solubility of the surfactant in the least-compressible phase (the brine). An increase in the salinity of the brine is required to return the phase environment of the system to its atmospheric pressure state. Equivalent alkane carbon number has been proposed as a physical characteristic relating to surfactant solubility and midpoint salinity by Salager et al. (1979). Nelson (1983) proposed a thermodynamic approach on the basis of an estimation of changes in molar volume and the solubility parameter of the oil. Most recently, Roshanfekr et al. (2009) presented a thermodynamic model to further explain relationships between solubilization ratios, pressure, and temperature. The intent of this paper is to conduct experiments that address whether this an important factor to consider when designing chemical floods. Phase-behavior experiments are discussed for live and dead crude, and corresponding corefloods are performed. Procedures Phase-Behavior Tests Emulsion Screening. Phase-behavior tests mix crude oil and a chemical slug in proportions similar to those that are encountered in the reservoir. A reasonable rule of thumb is to add crude oil to the chemical slug so that the amount 352 June 2012 SPE Journal

2 Phase Behavior of Mixed-Surfactant System 1.25% Na % IOS2024, 0.15% Petrostep A1, x NaCl NaCl % Fig. 1 Phase behavior of mixed-surfactant system as a function of salinity. is similar to the remaining oil saturation in the reservoir. For these studies, we mix 35% crude oil with 65% chemical slug. In ASP technology with a high TAN crude oil, the crude/aqueous phase ratio is often a critical parameter because the crude oil determines the concentration of petroleum acids available for neutralization by alkali, and the ratio of petroleum acids to added synthetic surfactant determines midpoint salinity. The crude oil used in our experiments has a low acid number (0.05), and the amount of petroleum acid available for neutralization does not affect midpoint salinity. (A constant midpoint salinity is observed at widely different oil/aqueous ratios.) The tubes were shaken and equilibrated at reservoir temperature (69 C). Equilibration typically takes 1 to 2 days. An example of the phase behavior with crude and the mixed-surfactant system used in this study is shown in Fig. 1. It is seen that for this system, the three-phase region is narrow. Midpoint salinity occurs at 0.6% NaCl, and a three-phase region is only observed at +/ 0.1% from the midpoint. A similar narrowness of the three-phase region was also seen when the pure surfactants were tested. Similar phase-behavior tests for live crude were performed. Oil, brine, and surfactant were introduced into a standard refinery sight cell (rated to 2,500-psi working pressure). Then, methane was pressurized in the cell at 1,100 psi. The sight cell was rocked for approximately 5 minutes to allow gas to dissolve in the oil. After this initial charge, the pressure in the sight cell had dropped to 650 psi. The procedure was repeated. After the second charge and solution rocking, the pressure dropped to 875 psi. After the third charge of gas to 1,100 psi, further rocking caused no drop in pressure. At this point, we concluded that the oil was saturated with gas at 1,100 psi. The sight cell was shaken periodically to allow all fluids to mix thoroughly, and the phase behavior was evaluated after the fluids had attained equilibrium. Dead-Crude-Displacement Experiments. Sandpacks were prepared with pure silica sand. No attempt was made to incorporate clays. These experiments test the ability of the surfactant systems to displace crude oil in stable displacements with minimal consumption of the alkali and minimal surfactant adsorption. Some floods were performed vertically with aqueous fluids injected into the bottom of the pack at a low frontal advance rate (0.5 to 1.0 ft/d interstitial). Oil production was measured as a function of pore volume with a sample collector. Additional floods were run horizontally with Flopaam 3430 added to the chemical slug and polymer drive for mobility control. For all floods, 0.3 pore volumes (PV) of chemical slug was injected, followed by polymer drive. Live-Crude-Displacement Procedures. A sandpack was prepared in a steel cylinder rated for high pressures. Permeability and porosity were determined. The sandpack was then pressurized to 1,250 June 2012 SPE Journal 353

3 Live Crude Fig. 2 Phase behavior for 0.5% Petrostep A1, 0.5% IBA, 1.5% Na 2, and 0.5% NaCl. Shown are live crude (left) and dead crude (right). psi. In a separate vessel, live crude was reconstituted by pressurizing with methane to 1,100 psi at 69 C. The one-phase live crude was then further pressurized to 1,250 psi. The coreflood was performed above methane saturation pressure to be sure that slight drops in pressure caused by equipment issues do not cause methane to come out of solution. The live crude was then injected into the sandpack until no water was produced (S oi ). A waterflood was then conducted with the same salinity as the chemical slug (1.25% NaCl) until no oil was produced (S or ). An ASP slug was then injected (0.3 PV) followed by polymer drive. Oil production was measured as a function of PV with a sample collector. Results are reported as percent of the remaining oil after waterflood recovered (%S or ). Results and Discussion Phase-Behavior Tests. The first series of tests were performed with the following chemical slugs: 0.5% Petrostep A1 0.5% isobutyl alcohol Live Crude 1.5% Na 2 (a) Above with additional 0.5% NaCl (b) Above with additional 0.75% NaCl (c) Above with additional 1.0% NaCl The pictures for the equilibrated systems for the dead and live crude are shown in Figs. 2 through 4. Fig. 3 shows that the Petrostep A1 system with 0.75% NaCl is at midpoint salinity (1.5% Na % NaCl) with dead crude. Petrostep A1 is a branched C15 18 branched alkyl benzene sulfonate produced by Stepan Chemical. As seen in Table 1, the middle-phase microemulsion contains nearly equal volumes of oil and brine. IFT was calculated from experimentally determined solubilization parameters by the methodology proposed by Huh (1979), seen in the following equation: = C/ (1) with C taken as 0.3, where is IFT (millidynes/cm), is the solubilization parameter, and C is a constant. The solubilization Fig. 3 Phase behavior for 0.5% Petrostep A1, 0.5% IBA, 1.5% Na 2, and 0.75% NaCl. Shown are live crude (left) and dead crude (right). 354 June 2012 SPE Journal

4 Live Crude Fig. 4 Phase behavior for 0.5% Petrostep A1, 0.5% IBA, 1.5% Na 2, and 1.0% NaCl. Shown are live crude (left) and dead crude (right). parameter is defined as the volume ratio of oil/surfactant o or brine/surfactant b in the microemulsion phase. In contrast, midpoint salinity in the live-crude case is seen in the sight cell photographed in Fig. 2, the cell containing the chemical slug with 1.5% Na % NaCl. Thus, introducing solution gas to this crude tends to make the surfactant relatively more soluble in the oil. Midpoint salinity is therefore lower with live crude than with dead crude in this case, 0.25% NaCl lower. More phase-behavior tests were performed than those shown in this paper. The value of 0.25% NaCl offset is accurate and significant when one considers the narrowness of the three-phase region as a function of salinity. The live-crude phase behavior exhibits higher solubilization parameter and thus lower IFT. A rule of thumb is that systems that give exceptional recovery in porous media exhibit IFTs between oil and brine of values of approximately 1 millidyne/cm. Both of the previously described systems meet this criterion, and the IFT calculated for live crude is extremely low. These data suggest that this particular live crude develops lower IFT than dead crude and will be easier to mobilize in displacement experiments, although with IFT this low, the lower oil viscosity of live crude may be the more important factor. A second set of experiments was performed with a C2024 internal olefin sulfonate surfactant from Shell Chemical. Live- and dead-crude phase-behavior tests are shown in Figs. 5 and 6. Midpoint salinity for the live crude is 1% Na 2, whereas midpoint salinity for dead crude is 1% Na % NaCl. Again, the offset is 0.25%. This result is not obvious because differing chemistries of the surfactant tails are expected to affect their relative solubility in oil and brine and their sensitivity to dissolved gas. Adjustment of Midpoint Salinity by Altering Surfactant/Blend Ratio. Reducing salinity by 0.25% to adjust for live-crude phase behavior may not be a desirable change for the chemical formulation. It is possible that a higher salinity may be a better match with a reservoir brine, or that a higher amount of alkali is needed to propagate through the reservoir. Fig. 7 shows phase behavior for a 1% Na % NaCl chemical slug with a blend of Petrostep A1 and IOS2024. For IOS2024 surfactant alone, this salinity was at midpoint salinity for dead crude but above midpoint salinity for live crude. Midpoint salinity and ultralow IFT can be re-established by blending in a second surfactant that is more hydrophilic than IOS2024. In this case, we used Petrostep A1. Fig. 7 shows the result of trial and error tests to find the correct blend ratio. 0.35% IOS % Petrostep A1 is a system showing midpoint salinity for live crude. Also shown in Fig. 7 is a control experiment where the surfactant blend was pressurized to 1,100 psi with nitrogen rather than methane. The results clearly show that it is methane and not pressure that is increasing surfactant solubility in oil and reducing midpoint salinity. This is expected and in agreement with earlier work. From Fig. 7, it appears that pressure alone (with the inert gas nitrogen) does not significantly change the phase behavior of the system compared with the dead-crude system. The volumes of the phases are affected by the dissolution of nitrogen into the liquids, but the somewhat underoptimum system observed with dead crude is also observed after nitrogen pressurization. Phase behavior is evaluated and summarized in Table 2. Once again, live crude shows large solubilization parameters and very low IFTs. Values for the nitrogen-pressurized system cannot be determined because of the unknown amount of oil solubilized in the aqueous layer. Live- and Dead-Crude Corefloods. This would be an interesting, but mostly academic, exercise unless 0.25% salinity offset affects oil production from corefloods. Table 3 reports chemical-flood TABLE 1 PHASE BEHAVIOR AND IFT CALCULATIONS FOR SYSTEMS AT MIDPOINT SALINITY Live Crude 0.5% NaCl (Fig. 1) 0.75% NaCl (Fig. 2) Solubilization parameter for oil Solubilization parameter for brine IFT oil and microemulsion (millidynes/cm) IFT brine and microemulsion (millidynes/cm) June 2012 SPE Journal 355

5 Live Crude Fig. 5 Phase behavior for 0.5% IOS2024, 0.5% IBA, 1.0% Na 2, and 0% NaCl. Shown are live crude (left) and dead crude (right). characteristics, and Table 4 reports chemical-flood results for two sandpack floods, one with dead crude and one with reconstituted live crude. The adjustment for midpoint salinity was to blend in a more-hydrophilic surfactant and to maintain the Na 2 concentration at 1.25%. Phase behavior for this approach was discussed previously. The reason not to adjust for live crude by dropping the Na 2 concentration below 1.25% will be the subject of a future SPE paper from this laboratory (Svec et al.). It is seen from the flood results (Table 4) that the correct adjustment in chemical-slug formulation has been made. These experiments were meant to provide a direct comparison of the performance of a dead-crude flood and a live-crude flood. With the modified formulation, it is noted that the viscosity ratio slug/oil differs by more than a factor of two. However, because both floods show stable displacement, this is not a critical difference. Fig. 8 shows production plots and produced fluids for the two floods. There are subtle differences between the two floods, but given the nature of these types of experiments, the floods show excellent reproducibility. The dead-crude flood recovered 91% of waterflood residual oil (92% recovered as clean oil, 8% as emulsion), and the live-crude flood recovered 89% of waterflood residual oil (83% recovered as clean oil, 17% as emulsion). This shows that an appropriate reformulation can be performed for livecrude conditions by evaluating phase behavior from sight cells. However, we have not answered the question as to how critical this reformulation is. What would have been the result if we had not reformulated the chemical slug for live crude oil and had simply injected the optimized dead-crude formulation into the live-crude sandpack? We offer guidance in answering this question by comparing the results of the sensitivity of oil recovery to chemical-slug salinity with a series of sandpack floods with dead crude. Vertical Sandpack Floods With Salinity Effects. Four sandpack floods were performed vertically to test surfactant Live Crude Fig. 6 Phase behavior for 0.5% IOS2024, 0.5% IBA, 1.0% Na 2, and 0.25% NaCl. Shown are live crude (left) and dead crude (right). 356 June 2012 SPE Journal

6 Live Crude 1,100-psi methane 1,100 psi With Nitrogen Fig. 7 Phase behavior for 0.35% IOS2024, 0.15% Petrostep A1, 0.5% IBA, 1.0% Na2CO3, and 0.25% NaCl. Shown are live crude (left) and dead crude (center). Also shown (right) is a control experiment whereby the system was pressurized with nitrogen rather than methane. displacement characteristics. Three floods were run at constant salinity, meaning that the waterflood, chemical-slug, and polymerdrive salinities were all constant. A fourth flood was performed in which the salinity of the polymer drive was reduced. Results of these four floods are shown in Table 5. All floods were performed with 1% Na 2 as alkali in the slug. Either 0, 0.25, or 0.5% NaCl was incorporated with the alkali. Emulsion-screening tubes exhibit a three-phase microemulsion for the 1.25%-total-salinity case similar to the dead-crude tube shown in Fig. 6. Surfactant is present in the aqueous phase with 1% salinity; 1.5% salinity gives evidence of surfactant in the oil (a brown-colored oil suggests entrained water). The performance of the floods clearly shows the highly negative result of slug chemistry being above midpoint salinity (1.5%). Surfactant partitions into the oil, TABLE 2 PHASE BEHAVIOR AND IFT CALCULATIONS FOR MIXED SURFACTANT SYSTEMS SHOWN IN FIG. 7 Live Crude (1,100-psi methane) Dead Crude Nitrogen (1,100 psi) Solubilization parameter for oil Solubilization parameter for brine IFT oil and microemulsion (millidynes/cm) IFT brine and microemulsion (millidynes/cm) TABLE 3 FLUID PROPERTIES AND CORE DESCRIPT ION FOR DEAD-CRUDE AND RECONSTITUTED-LIVE- CRUDE FLOODS Oil Viscosity ASP Slug Viscosity Drive Viscosity Permeability Orientation Dead crude 1.1 cp 5.7 cp 5.1 cp 4700 md Horizontal Live crude 0.6 cp (est) 7.7 cp 6.6 cp 7800 md Horizontal TABLE 4 ASP FLOOD RESULTS WITH DEAD CRUDE AND RECONSTITUTED LIVE CRUDE ASP Chemical Slug Components O il Recovery Oil IOS 2024 PS-A1 IBA Na 2 FP-3430 Clean Oil (% of Total) Total Rec (% of S or ) Dead 0.5% 0% 0.5% 1.25% 0.1% 92% 94% Live 0.35% 0.15% 0.5% 1.25% 0.1% 83% 89% June 2012 SPE Journal 357

7 Dead-Crude Flood Recovery 94% of S or 1 Chemical Flood PV Injected Oil Cut Oil Production 1.25% Na 2 0.5% IOS % PetrostepA1 0.5% IBA Live-Crude Recovery=89% 1.25% Na 2 0.5% IBA 0.15% Petrostep A1 0.35% IOS %S or _WF PV Injected Fig. 8 Comparison of dead crude and reconstituted live crude with methane. Live-crude flood was performed at 1,100 psi. viscous emulsions form, and oil recovery suffers. For chemical slugs that are above optimum salinity, the results of the gradient flood support the original observations of Nelson (1982), who recommended formulations with a salinity gradient to minimize risk. Chasing an overoptimum chemical slug with a lower-salinity polymer drive forms a salinity gradient in the mixing zone between the slug and drive. Salinity will gradually reduce through this mixing zone, giving assurance that the optimum salinity is reached at some point during the flood. The lower-salinity polymer drive improved recovery of an overoptimum flood from 48 to 77%. TABLE 5 VERTICAL SANDPACK FLOODS WITH VARYING SALINITIES Slug Salinity Na 2 NaCl Type Waterflood NaCl Drive NaCl Surfactant Permeability Oil Recovery 1% 0% constant 1% 1% 0.8% md 95% 1%.25% constant 1.25% 1.25% 0.8% md 93% 1%.5% constant 1.5% 1.5% 0.8% md 48% 1%.5% gradient 1.5% 1% 0.8% md 77% 358 June 2012 SPE Journal

8 TABLE 6 CALCULATED MOLAR VOLUME AND SOLUBILITY PARAMETER FOR TWO CRUDE OILS Oil Molar Volume ( V o) Solubility Parameter (δ O ) This study paraffinic 283 ml/mole 7.1 (cal/cm 3 ) 1/2 Gulf of Mexico oil 353 ml/mole 7.62 (cal/cm 3 ) 1/2 The first two floods in Table 5 show the benefit of maintaining salinity at or below midpoint salinity. Constant-salinity floods gave very high recoveries. It is interesting that the slightly underoptimum formulation did as well as the formulation at midpoint salinity. We would not always expect this to be the case. High-permeability sandpacks may not be as discriminating as low-permeability cores between systems with IFT differences. The lowest IFT is expected at midpoint salinity. Relevance to Formulations With Live Crude The results shown in this paper argue that it is important to check chemical-flooding formulation-phase behavior with live crude before field implementation. The specific live crude oil in this study gave a midpoint salinity 0.25% NaCl below the dead-crude value. If live-crude phase behavior was not determined in this study, the chemical slug would have been formulated 0.25% overoptimum. Vertical sandpack floods show that this is deleterious to chemical-flood performance. Given inherent uncertainties in reservoir salinities and gas/oil ratio, the results of this paper strongly support the original work of Nelson (1982), suggesting that salinity-gradient formulations can minimize risk in chemical flooding, especially when interfacial activity changes with relatively small changes in salinity. Generalization of Results. While results with the particular crude oil in this study agree with earlier observations of Puerto and Reed (1983) and Roshanfekr et al. (2009) in that dissolved methane increases relative surfactant solubility in the oil, this may not always be the case. The surfactant in a surfactant/brine/oil system below midpoint salinity and in a Winsor type II (underoptimum) environment forms micelles the center of which contain the surfactant tails and some of the oil. As the salinity of the brine is increased, the micelles grow and take up more oil. If the salinity of the brine is increased enough, the system approaches midpoint salinity, and enters the Type III phase environment. At the low surfactant concentrations used in ASP flooding, the Type III phase environment usually shows three phases unless the solubilization parameters are very high. Further increases in salinity cause a system in the Type III phase environment to change to the Type II+ phase environment (overoptimum), in which the surfactant forms inverse micelles with the surfactant tails on the outside of the micelle in contact with the oil. Nelson (1983) argues that although the thermodynamics of surfactant/brine/oil mixing is complex, there are two environments in which regular-solution theory (Hildebrand and Scott 1950; Hildebrand et al. 1970) might be a reasonable approximation. One is the center of the micelles in the Type II phase environment; the other is the surface of the inverse micelles in the Type II+ phase environment. In both of these environments, oil mixes with surfactant tails, and the more negative the free energy of that mixing, the greater the thermodynamic driving force for the oil and surfactant to mix that is, for the surfactant to be soluble in the oil. Molar free energy of mixing is determined by two thermodynamic properties (enthalpy H and entropy S) that are related by the well-known equation G m = H m T S m (2) In an ideal solution, the change in enthalpy of mixing H m = 0. In a regular solution, S m equals the entropy change of an ideal solution, but H m 0. If dispersion forces (van der Waals forces) are the predominant intermolecular forces, as should be the case between oil and hydrophobic surfactant tails, Hildebrand s expression H m = V( 1 2 ) (3) should be applicable. In Eq. 3, V is the volume of one mole of a solution of Components 1 and 2, 1 and 2 are the mole fractions of the two components, and 1 and 2 are the solubility parameters of the two components. As defined by Hildebrand, the solubility parameter of a substance is given by the square root of its cohesive energy density: = [( H RT)/V m )] 1/2, where H is its heat of vaporization, R is the gas constant, T is the absolute temperature, and V m is its molar volume. Solubility parameters for many substances are available in the literature. When the solubility parameter for a liquid, such as dead crude, is not known, it can be estimated from the equation o = 4.1[ o /(V o ) 1/2 ] 0.43, where o is the measured surface tension of the dead crude and V o is its molar volume, which is estimated from its density and average molecular weight. Nelson (1983) defines an enthalpy parameter as V o ( S o ) 2, where V o is the molar volume of the dead crude plus whatever fraction of gas has been added to it, o is the solubility parameter of the dead crude plus whatever fraction of gas has been added to it, and s is the Hildebrand solubility parameter of the surfactant tails. V o and o are calculated as the weighted volume fraction of the oil and methane components. The enthalpy parameter is proportional to the Hildebrand enthalpy of mixing at constant temperature, pressure, and volume fraction of oil and surfactant tails. Adding methane to a dead crude decreases its molar volume, but it also decreases its solubility parameter, so adding methane may increase or decrease the enthalpy parameter, depending on the starting values of the molar volume and the two solubility parameters. This suggests that reconstituting a live crude by adding methane to a dead crude may increase or decrease the relative solubility of the surfactant in the oil. In other words, reconstituting a live crude by adding methane to a dead crude may decrease or increase the salinity requirement of the surfactant/brine/oil system. To demonstrate this point, two oils were characterized by measuring surface tension and density and calculating the average molecular weight from dead-crude-oil analyses. One is the oil used in this study; the other is a crude oil from a Gulf of Mexico field. Table 6 summarizes the calculated parameters. The final parameter that is needed to calculate the enthalpy parameter is the solubility parameter of the lipophyllic surfactant tails. Solubility-parameter trends for different classes of molecules were previously estimated, and solubility parameters between 8.3 and 8.7 were deemed reasonable for most chemical-flooding surfactants. The surfactant system employed in this study is relatively hydrophobic; we have chosen 8.3 as the best estimate for the solubility parameter of the surfactant tails. Calculations of enthalpy parameter as a function of methane concentration are shown in Fig. 9 for both oils, using the oil parameters in Table 6 as pure component values and an estimated solubility parameter for surfactant tails of 8.3. Weighted-volumefraction calculations according to Eqs. 4 and 5, V o = 1/ f i /V i (4) O = f i i, (5) give V o and o for each mixture as methane is increased in concentration. In Eqs. 4 and 5, f i, V, and i i are the volume fraction, molar volume, and solubility parameter of the ith component, respectively. It is interesting that, as enthalpy parameter is calculated for each V o and o, different trends are seen for the two oils. A decrease in enthalpy parameter with added methane indicates increased relative solubility for the surfactant in oil. This is calculated for the June 2012 SPE Journal 359

9 Enthalpy Parameter Oil 1 This Study Paraffinic Enthalpy Parameter Oil 2 Gulf of Mexico (a) Gas (%) (b) Gas (%) Fig. 9 Calculated enthalpy parameters for two crude oils from experimentally determined properties; molar volume and solubility parameter (through surface tension). (a) Reduction of enthalpy parameter with dissolved gas indicates surfactant is becoming relatively more soluble in the oil phase. This was calculated for the oil used in this study. (b) Increase in enthalpy parameter with dissolved gas predicts surfactant is becoming relatively less soluble in the oil phase, and hence relatively more soluble in the brine phase. This was calculated for a Gulf of Mexico oil. oil of this study, and also confirmed experimentally. The Gulf of Mexico crude oil, however, gives the opposite behavior increased methane dissolution results in decreased relative solubility of the surfactant in oil. More experiments with a wider range of crude oils are needed to verify the predictive utility of the enthalpy parameter. It is encouraging that the predictions are qualitatively in agreement with the experiments perfomed here. The fact that measured oil properties can also predict an opposite trend should alert chemical formulators to consider the effect of live crude before formulations are finalized for field implementation. Conclusions Two surfactant systems exhibited enhanced relative solubility in reconstituted live crude oil when compared with similar experiments with dead crude. Further coreflow experiments show that the differences between live and dead crude are an important factor in the design of chemical floods. The observed change in relative solubility is significant; a chemical flood formulated in the laboratory with dead crude and then applied to field conditions (where the crude is live) would be overoptimum (above midpoint salinity), and underperforming. The trend toward enhanced relative surfactant solubility for the experimental crude is also predicted by thermodynamic calculations based on experimentally determined values from the crude density, molecular weight, and surface tension. Similar thermodynamic calculations for a Gulf of Mexico crude oil based on measurements of density, molecular weight, and surface tension predict the opposite behavior; the surfactant will be relatively less soluble in the live crude compared with dead crude. This finding argues that generalizations should not be made, and that the phase behavior for a particular live crude should be determined experimentally with the surfactant system being tested for field operation. Experiments where dead crude was pressurized with nitrogen rather than methane show that the surfactant is relatively more soluble in brine relative to the crude at atmospheric pressure. Therefore, the effects of absolute pressure and dissolved methane are acting in opposing directions. However, the effect of dissolved methane is the dominant factor, and the thermodynamics of mixing between the reconstituted live crude and surfactant tails will determine the relative solubility of the surfactant tails with respect to brine. The effect of live crude is an important variable to consider when designing surfactant formulations for field conditions. Nomenclature C = constant f i = volume fraction of the ith component V = volume of one mole of a solution of Components 1 and 2 V i = molar volume of the ith component = interfacial tension, millidynes/cm i = solubility parameter of the ith component = solubilization parameter 1, 2 = mole fractions of the two components 1 and 2 Acknowledgments The authors wish to acknowledge useful discussions with R.C. Nelson during the course of this work. We also wish to thank Shell International E&P for giving us permission to publish this paper. References Austad, T. and Strand, S Chemical flooding of oil reservoirs 4. Effects of temperature and pressure on the middle phase solubilization parameters close to optimum flood conditions. Colloids Surf., A 108 (2 3): Austad, T., Hodne, H., Strand, S., and Veggeland, K Chemical flooding of oil reservoirs 5. The multiphase behavior of oil/brine/surfactant systems in relation to changes in pressure, temperature, and oil composition. Colloids Surf., A 108 (2 3): org/ / (95) Hildebrand, J.H. and Scott, R.L The Solubility of Non-Electrolytes, third edition, No. 17. New York: Monograph Series, Van Nostrand Reinhold. Hildebrand, J.H., Prausnitz, J.M., and Scott, R.L Regular and Related Solutions: The Solubility of Gases, Liquids, and Solids. New York: Van Nostrand Reinhold. Huh, C Interfacial tensions and solubilizing ability of a microemulsion phase that coexists with oil and brine. J. Colloid Interface Sci. 71 (2): Kahlweit, M., Strey, R., Firman, P., Haase, D., Jen, J., and Schomaecker, R General patterns of the phase behavior of mixtures of water, nonpolar solvents, amphiphiles, and electrolytes. 1. Langmuir 4 (3): Nelson, R.C The Salinity-Requirement Diagram A Useful Tool in Chemical Flooding Research and Development. SPE J. 22 (2): SPE-8824-PA. Nelson, R.C The Effect of Live Crude on Phase Behavior and Oil- Recovery Efficiency of Surfactant Flooding Systems. SPE J. 23 (3): SPE PA. Puerto, M.C. and Reed, R.L A Three-Parameter Representation of Surfactant/Oil/Brine Interaction. SPE J. 23 (4): SPE PA. Roshanfekr, M., Johns, R.T., Pope, G.A., et al Effect of Pressure, Temperature, and Solution Gas on Oil Recovery From Surfactant Polymer Floods. Paper SPE presented at the SPE Annual 360 June 2012 SPE Journal

10 Technical Conference and Exhibition, New Orleans, 4 7 October. Salager, J.L., Bourrel, M., Schechter, R.S., and Wade, W.H Mixing Rules for Optimum Phase Behavior Formulations of Surfactant/Oil/ Water Systems. SPE J. 19 (5): Sassen, C.L., De Loos, T.W., and De Swaan Arons, J Influence of pressure on the phase behavior of the system water + decane + 2-butoxyethanol using a new experimental setup. The Journal of Physical Chemistry 95 (26): Skauge, A. and Forland, P Effect of Pressure and Temperature on the Phase Behavior of Microemulsions. SPE Res Eng 5 (4): SPE PA. Jeff Southwick is a Principal Production Chemist with Shell International Exploration & Production in Houston. He has 11 years of experience working in the area of chemical flooding, polymer flooding, and the interaction of alkali with reservoir rock. Southwick holds a BS degree in chemistry from the State University of New York (Syracuse), and MS and PhD degrees in polymer science and engineering from Case Western Reserve University. Yi Svec is a Staff Reservoir Engineer at Shell International Exploration & Production in Houston. Her work interests include enhanced oil recovery by polymer, and alkali-surfactant-polymer (ASP) processes. Svec holds a BS degree in oil chemistry from the Chengdu Geology Institute, an MS degree in polymers from Jilin University, and a PhD degree in petroleum engineering from New Mexico Tech. Greg Chilek is a Reservoir Engineer with 18 years of experience working on Shell s domestic and international Enhanced Oil Recovery projects, both in the field and in the laboratory. His responsibilities are improving enhanced oil recovery measurement techniques for Shell s future projects. Gordon Thomas Shahin Jr. is a Senior Staff Reservoir Engineer at Shell s International Exploration & Production Technology Organization in Houston. Shahin joined Shell s Bellaire Research Center and worked in Chemical Flooding. During the early 1990s, he implemented conventional- and light-oil steaminjection projects in California and Oman while part of the Advanced Thermal Technology Group. As technical operations supervisor, he oversaw the technical operations and analysis of results from two of Shell s thermal conduction oil shale pilots. He is currently a senior technical leader in Shell s Integrated Field Study Group, focusing on global application of EOR techniques. Shahin has authored numerous papers and executed eight pilots in his 24-year career with Shell. He holds BS, MS, and PhD degrees in chemical and biochemical engineering. Shahin has served as an SPE Distinguished Lecturer ( ) and holds more than 60 patents in thermal recovery and related technologies. June 2012 SPE Journal 361