identification and evaluation of high-performance eor surfactants

Identification and Evaluation ofHigh-Performance EOR Surfactants

David B. Levitt, SPE, Adam C. Jackson, SPE, Christopher Heinson,* SPE, and Larry N. Britton, The Universityof Texas at Austin; Taimur Malik, and Varadarajan Dwarakanath,** SPE, Intera; and Gary A. Pope, SPE,

The University of Texas at Austin

Summary

We report results for a number of promising enhanced-oil-recov-ery (EOR) surfactants based upon a fast, low-cost laboratoryscreening process that is highly effective in selecting the bestsurfactants to use with different crude oils. Initial selection ofsurfactants is based upon desirable surfactant structure. Phase-behavior screening helps to quickly identify favorable surfactantformulations. Salinity scans are conducted to observe equilibra-tion times, microemulsion viscosity, oil- and water-solubilizationratios, and interfacial tension (IFT). Cosurfactants and cosolventsare included to minimize gels, liquid crystals, and macroemul-sions and to promote rapid equilibration to low-viscosity micro-emulsions. Branched alcohol propoxy sulfates (APS), internalolefin sulfonates, and branched alpha olefin sulfonates (AOS)have been identified as good EOR surfactants using this screeningprocess. These surfactants are available at a low cost and arecompatible with both polymers and alkali, such as sodium carbon-ate and, thus, are good candidates for both surfactant-polymer andalkali-surfactant-polymer EOR processes. One of the best formu-lations was tested in both sandstone and dolomite cores and foundto give excellent oil recovery and low surfactant retention with awest Texas (WT) crude oil.

Introduction

Recent advances, including the development of new syntheticsurfactants and increased understanding of the structure/performance relationship of surfactants, have made it possible torapidly identify promising high-performance surfactants for EOR.This process involves laboratory screening using knowledge ofthe molecular structure and cost of the surfactants as well aspertinent reservoir-specific information (i.e., temperature, salinity,and crude-oil properties).

This paper describes a process for identifying and evaluatingpotential EOR surfactants. The surfactant selection process startswith the screening of surfactants by phase-behavior experimentsand progresses to corefloods with formulations that may incorpo-rate cosurfactants, cosolvents, alkali, polymers, and electrolytes.We illustrate the application of this approach to the selection of asurfactant formulation for use in both a sandstone outcrop and aWT dolomite reservoir, but focus mostly on the dolomite applica-tion because very few studies have been reported for carbonate(Adams and Schievelbein 1987) or dolomite reservoirs. Theselaboratory data were used in a parallel simulation study of thesame reservoir and are described by Anderson et al. (1976) in acompanion paper.

Background

It is well known that the primary requirement needed to mobilizeresidual oil saturation is a sufficiently low IFT to give a capillarynumber large enough to overcome the capillary forces and allowthe oil to flow (Stegemeier 1976). Low IFT can be obtained with awide variety of surfactants, but the best surfactant depends on the

crude-oil and reservoir conditions and must also satisfy severalother stringent requirements. These requirements include low re-tention, compatibility with the electrolytes and polymer, thermalstability, aqueous stability, and low cost. Surfactant retention inpart is because of adsorption on the rock surfaces, but other lossmechanisms including phase trapping can be just as important, ormore so. There is a strong and well-established relationship be-tween the microemulsion phase behavior and IFT (Winsor 1954;Healy et al. 1976; Huh 1979; Nelson and Pope 1978; Pope et al.1979; Bourrel and Schechter 1988; Aoudia et al. 1995). Thisrelationship can be used to great advantage to rapidly screensurfactants and predict which ones are likely to perform best inthe more difficult and expensive corefloods, and this is the ap-proach taken in this paper.

Winsor (1954) described surfactant/oil/water microemulsionsas Type I (oil in water), Type II (water in oil), or Type III(bicontinuous oil and water in a third phase known as middle-phase microemulsion). Type III or middle-phase microemulsionsexhibit the lowest IFT. For anionic surfactants, increasing thesalinity among other variables causes the characteristic transitionfrom Type I to Type III to Type II. Healy et al. (1976) presentedthe concept of optimum salinity as it applies to Type III micro-emulsions. They observed the volumes of oil (Vo) and water (Vw)per unit volume of pure surfactant (Vs) in middle-phase microe-mulsions and defined the optimum solubilization ratio (s*) as theintersection of plots of Vo/Vs and Vw/Vs as a function of salinityor other variables that affect the phase behavior. Huh (1979) de-veloped a theoretical relationship between the oil and water solu-bilization ratios, s, and IFT (g):

g ¼ C

s2; (1)

where C is approximately 0.3 dynes/cm for typical crude oils andsurfactants used for EOR. A very large number of papers pub-lished since 1979 have shown that the Huh equation accuratelymodels the IFT between equilibrium microemulsions and oil orwater for numerous combinations of surfactants and crude oilsacross a wide range of concentration, salinity, temperature, andother conditions typical of oil reservoirs. While IFT in the rangeof interest can be measured by the spinning-drop method, usingthe Huh equation to calculate it from phase behavior data affordsseveral significant advantages. These include the comparativeease and speed with which phase-behavior experiments can beperformed as well as the opportunity for easily taking repeatedmeasurements over time.

It is desirable to quickly examine hundreds of combinations ofsurfactants, cosurfactants, cosolvents, and numerous electrolytesin several different reservoir and injection brines, polymers, alkali(such as sodium carbonate and sodium hydroxide), and to exam-ine the effect of different water/oil ratios (WORs) for acidic crudeoils, where this will affect the amount of natural soaps produced.Thus, using simple observations of the microemulsion-phase be-havior in pipettes rather than measuring IFT for each combinationis a very big advantage. Furthermore, IFT can be very difficult tomeasure with some crude oils and is subject to large uncertainty.IFT can vary with time for periods of up to several months andmay depend on precisely how the fluids are sampled andmeasured. It is, of course, a good idea to measure IFT on selectedsamples after the initial phase-behavior screening, but it is not themost practical way to initially screen surfactants.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

* Now with Oxy.** Adam C. Jackson, Taimur Malik, and Varadarajan Dwarakanath are now with Chevron.

Copyright ã 2009 Society of Petroleum Engineers

This paper (SPE 100089) was accepted for presentation at the SPE/DOE Symposium onImproved Oil Recovery, Tulsa, 22–26 April 2006, and revised for publication. Originalmanuscript received for review 17 February 2006. Revised manuscript received for review25 July 2008. Paper peer approved 11 August 2008.

April 2009 SPE Reservoir Evaluation & Engineering 243

Using phase behavior to screen surfactants has other verysignificant advantages over IFT. For instance, qualitative judge-ments of microemulsion viscosity and tendancy to form gels canbe made. These phenomena are as important to the selection of asuitable surfactant as IFT is. Highly viscous phases will not easilytransport under low pressure gradients and will perform poorly inthe reservoir where the pressure gradient is often on the order of 1psi/ft or less.

High viscosity correlates with high surfactant retention. Thus, itis more efficient and effective to screen out such surfactants or tomitigate the problem by adding cosolvents or some other measureearly in the screening process and not attempt to measure IFT until amicroemulsion with a reasonable viscosity is identified. Corefloodsare expensive and coreflood data are subject to uncertainty andmisinterpretation unless the phase behavior and physical properties(such as viscosity) have been measured and are well understood.

Extensive research on surfactants has established a clear rela-tionship between surfactant structure and fluid properties and per-formance related to EOR (Bourrel and Schechter 1988; Aoudiaet al. 1995). For example, the solubilization ratio increases and theoptimum salinity decreases with increasing hydrophobe length.Weakly hydrophobic functional groups [such as propylene oxide(PO)] have been characterized as having interface affinity, and, assuch, increase the breadth of the ultralow IFT region. The additionof these hydrophobic groups lowers the optimum salinity and addscalcium tolerance, so the degree of propoxylation can be used totailor the surfactant to a given crude oil, temperature, and salinity(Aoudia et al. 1995; Wu et al. 2005; Jayanti et al. 2002; Levitt2006; Hirasaki et al. 2006). Similar statements can be made withrespect to the addition of ethylene oxide (EO) or both EO and POgroups to the surfactant. Fortunately, both EO and PO are rela-tively inexpensive chemicals, so these are among the most practi-cal ways to tailor a surfactant to the desired conditions as well asto improve its performance.

In this paper, we focus on the use of POs to improve surfactantperformance. Aoudia et al. (1995) shows the importance of hydro-phobe branching with Guerbet alcohol propoxy sulfate surfactantsthat form middle-phase microemulsions. Branching decreases theorder in the micellar structures, which tends to decrease the vis-cosity, reduce the time required for mixtures to equilibrate, andpromotes the formation of microemulsions as opposed to undesir-able liquid crystals, gels and other viscous phases. Cosurfactantsthat are branched or of differing structure are often added tofurther disrupt the orderly arrangement of surfactant molecules.Alcohols in the C4 range can be used to break viscous structures,but it is lower cost to use hydrophobe branching when possible.

Polymers can also complicate the phase behavior and require theaddition of cosolvent. Once again, the advantage of initially ob-serving the phase behavior rather than measuring IFT is apparentsince polymer typically has little—if any—effect on IFT.

On the basis of this extensive knowledge base and understand-ing of surfactants and microemulsions, it makes sense to seek outor make surfactants with branched hydrophobes and screen themusing the rapid phase-behavior approach. The phase behavior isalso needed for modeling purposes and is useful for interpretingthe data from corefloods to measure oil recovery and surfactantretention.

Experimental Procedures and Results

Surfactants Selected for Testing. Isotridecanol (TDA) was initi-ally considered to be the most suitable hydrophobe in terms ofphase behavior for pure hydrocarbons ranging between hexaneand decane (Levitt 2006; Hirasaki et al. 2006). We started withTDA in part because it is the lowest-cost-commodity alcoholsuitable for making synthetic surfactants with a sufficiently highcarbon number for EOR of light crude oils. Formulations with thisC13POx-sulfate exhibited high optimum salinities with the C6 toC10 hydrocarbons and may be a good candidate for some reser-voirs. However, for some crude oils, temperatures, and salinities,the C13 hydrophobe is too short, so PO sulfates were made from aC16-17 branched alcohol. Using this alcohol, 3, 5, and 7 propyleneoxide units were added to make the three surfactants C16-17-(PO)3-SO4, C16-17-(PO)5-SO4, and C16-17-(PO)7-SO4. Formulations witheach surfactant were screened with several cosurfactants andcosolvents. The 7 PO sulfate when mixed with a C15-18IOSemerged from the phase-behavior screening tests as the best for-mulation for a crude oil from a low-temperature dolomite reser-voir. For temperatures above approximately 60�C, sulfates willhydrolyze and cannot be used. For high temperature applications,similar surfactants can be made as sulfonates.

Table 1 gives the names and abbreviated chemical structuresof some of the surfactants that have been screened for use withthis WT reservoir (WT) crude using these new surfactants, as wellas some older generation surfactants (such as branched alkyl ben-zene sulfonates and AOS). Similar screening has also been donewith other crude oils at both The University of Texas and RiceUniversity (Levitt 2006; Hirasaki et al. 2006). The C20-24AOS,C15-18IOS, and the APS were synthesized specifically for testingin this project and are discussed below.

APS Surfactant. Commercially available branched alcoholswere selected for synthesis of the APS surfactants. Isotridecyl

244 April 2009 SPE Reservoir Evaluation & Engineering

(C13) alcohols manufactured from oligomerization of C3 groupsand of C4 groups were propoxylated and sulfated by Sasol Chemi-cal Co. or Stepan Chemical Co. Another hydrophobe selected forEOR applications was NeodolW 67, (Shell Chemical Company,London,) a C16-17 alcohol with an average of 1.5 methyl groupsrandomly positioned along the molecule. Branched, rather thanlinear alcohols were selected because branching decreases theformation of ordered structures/liquid crystals.

It was reasoned that the size of the hydrophobe (alcohol) wasappropriate for light crude oils with typical equivalent alkanecarbon numbers (EACN) ranging from C8 to C12. Alcohols werepropoxylated with varying amounts of PO and sulfated with sul-famic acid or sulfur trioxide by Stepan Chemical Co., Northfield,Illinois. APS surfactants similar to these have been investigated inprevious studies (Aoudia et al. 1995; Wu et al. 2005; Hirasaki andZhang 2004). An approximate structure of a C16-17-(PO)7-SO4

molecule generated by a space filling, free-energy minimizingmodel is shown in Fig. 1a. The purpose of this illustration is toshow the branching and compactness of the molecule and com-pare it with two others described next. The PO part of the mole-cule folds in a spiral. The hydrophobe here is shown with 16carbon atoms and three methyl branches. Both features result in amolecule that is very different than typical linear hydrophobes.This is highly beneficial because the linear surfactants tend toform ordered structures that lead to liquid crystals and gels. Thissurfactant can be tailored by changing the number of POs.

C15-18 Internal Olefin Sulfonate (IOS). This IOS was preparedby Stepan Chemical Co. using an internal olefin from Shell that isprepared by means of the oligomerization of ethylene by a processdescribed as olefin metathesis. The internal olefin will have anoverall size of C15 to C18 and a narrow range of internal, double-

bond positions such that sulfonation with SO32- will produce

a variety of products. In addition to alkene sulfonate isomersproduced by the position of the double bond, there are other sur-face-active products formed during the aging, neutralization, andhydrolysis steps of commercial sulfonation. The first intermediatesof sulfonation are b-sultones, which undergo transformation toalkene sulfonates or to g- and d-sultones. Neutralization and hy-drolysis steps transform the g- and d-sultones to hydroxy alkanesulfonates. The proportion of alkene sulfonates to hydroxyl alkanesulfonates is roughly 60:40. The end result is a surfactant mixturewith many different species of sulfonates, which is expected tominimize the formation of ordered structures such as liquid crys-tals and gels. The hydroxy alkane sulfonate form of a C15-18IOSmolecule is shown in Fig. 1b using the same approximate modelas before for the purposes of visualization and comparison.

C20-24 AOS. The C20-24 alpha olefin (AO) from BP is expectedto be more highly branched than the AO from Shell Chemical.The AOS from this olefin was prepared by Stepan Chemical Co.The chemistry of sulfonation is the same as for IOS, and theexpected products would be alkene-1-sulfonate and hydroxylalkane sulfonates. The hydrocarbon length and unspecifiedbranching of this AOS may contribute to its overall good perfor-mance. An approximate structure of a C20-24AOS molecule in itshydroxyl alkane sulfonate form is shown in Fig. 1c using the samemodel as before for the purposes of comparison.

Phase Behavior Screening. Experiments were first performed tomeasure the phase behavior of surfactant/oil/water mixtures as afunction of electrolyte concentration. Oil- and water-solubiliza-tion ratios were observed as salinity was increased causing atransition in phase behavior from Winsor Type I to Type III toType II. Two mL of each prospective surfactant slug formulationwere pipetted into a thin, graduated, 5 mL borosilicate glass pi-pette with a flame-sealed bottom end. The level of the aqueoussurfactant solution was recorded. Two mL of hydrocarbon wasadded and argon was used to displace volatile gas so that the topsof pipettes could be safely flame sealed. Pipettes were then placedin an oven at reservoir temperature. After reaching the desired testtemperature, pipettes were inverted several times to facilitate mix-ing. Phase volumes were then observed and recorded over time. Ifthe formation of viscous gels or macroemulsions appeared toinhibit mass transfer, pipettes were sometimes agitated again.Phase behavior of surfactant/cosurfactant/cosolvent formulationswas evaluated using the following mostly qualitative criteria:

• How fast the emulsions break after gentle mixing and form amicroemulsion in equilibrium with oil and/or brine and the ab-sence of macroemulsions or gels, particularly at and below opti-mum salinity.

• High solubilization ratio at optimum salinity and, hence, lowIFT. A solubilization ratio of 10 or greater corresponds to an IFTon the order of 0.003 dynes/cm or less.

• Microemulsion with low viscosity and the absence of highviscosity gels or other viscous phases, or rigid fluid interfaces, orinterfaces with persistent, viscous macroemulsions next to them.

• Aqueous stability of surfactant/polymer solutions at requiredinjection salinity (may be different from optimum salinity).

Preliminary phase-behavior screening with the surfactantslisted in Table 1 was performed with pure hydrocarbons such asoctane and decane (Fisher Scientific, Waltham, Massachusetts).Some of the advantages of using pure hydrocarbons for initialscreening include clarity of interfaces and other observations aswell as the ability to compare the results with a large body ofsurfactant data with the same pure hydrocarbons. The C16-17-(PO)7-SO4 exhibited the most promising performance withboth pure hydrocarbons and the WT crude oil (Table 2). BothC20-24AOS and C15-18IOS cosurfactants improved the perfor-mance of the C16-17-(PO)7-SO4 with respect to equilibration timeand the absence of gels and macroemulsions. The phase behaviorof a mixture of C16-17-(PO)7-SO4, and C15-18IOS surfactants withthe WT crude oil is shown as a volume fraction diagram in Fig. 2.The volume fraction data show the classical transition from Type Ito Type II to Type II microemulsion as the salinity increases. Note

Fig. 1—Possible structures of C16-17 alcohol 7-propoxy sulfate(C16-17-(PO)7-SO4) (a); C15-18 IOS (b); and C20-24 AOS (c).


that the precise optimum salinity shown in the tables and figuressometimes varies because of changing background calcium con-centrations, as indicated.

Mixtures with sodium carbonate (Na2CO3) exhibited shorterequilibration times (Fig. 3) as well as slightly higher solubiliza-tion ratios with both the WT crude oil (Fig. 4) and a crude oilfrom California (Fig. 5). However, most of the studies were donewithout sodium carbonate because of the presence of anhydrite inthe WT dolomite reservoir and, thus, the risk of calcium carbonateprecipitation. Most sandstone and carbonate reservoirs do nothave anhydrite in them, so sodium carbonate could be used togreat advantage since it reduces surfactant adsorption (Levitt2006; Hirasaki et al. 2006; Liu et al. 2008) and improves phasebehavior. The sec-butanol (SBA) cosolvent could probably beeliminated from this formulation with sodium carbonate added,and this would reduce the cost since sodium carbonate is muchless expensive than SBA, so for most reservoir applications theSBA would not be needed.

Cosurfactant Selection. As part of the initial screening pro-cess, a comparison was made of the performance of the C15-18IOSand C20-24AOS cosurfactants with the C16-17-(PO)7-SO4 primarysurfactant. Phase behavior data for both formulations are pre-sented in Figs. 6 and 7. Figs. 8 and 9 show how the two formula-tions approached equilibrium for a near-optimum salinity.Optimum solubilization ratios measured with the C15-18IOS cosur-factant are in the range of 10 to 12. The optimum solubilizationratio measured with the C20-24AOS cosurfactant was approximate-ly 11. The C15-18IOS cosurfactant had a slightly more favorablesolubilization ratio as well as a faster approach to equilibrium and

less tendency to form viscous phases, so it is a better cosurfactantunder these conditions. This behavior is consistent with the largerdegree of hydrophobe branching of the C15-18IOS surfactant. Anadditional observation that favored use of C15-18IOS was thepoorer performance of the C20-24AOS when the total surfactantconcentration was decreased. Several surfactant/cosurfactant ra-tios were evaluated in phase behavior experiments, and a ratio of3:1 of the APS to IOS was found to give the best results.

Cosolvent, Surfactant/Sosolvent Ratio and Total SurfactantConcentration. Both SBA and isopropanol (IPA) were evaluatedas cosolvents, and SBA proved superior. Increasing alcohol cosol-vent concentration speeds up equilibration, reduces the viscosity,and helps break macroemulsions, but invariably lowers the solubi-lization ratio at optimum salinity. Thus, a compromise must bemade between maximum solubilization ratio (low IFT) and lowviscosity and the other critical factors needed for good transportunder low pressure gradients in oil reservoirs.

Initial phase-behavior experiments were performed with fourwt% total surfactant (surfactant + cosurfactant concentration) anda 1:2 total surfactant-to-cosolvent ratio. This had the advantage ofgenerating large microemulsion volumes because of the high sur-factant concentration plus rapid equilibration because of the highconcentration of cosolvent. Once a promising surfactant andcosurfactant had been identified, additional phase-behaviorexperiments were performed in order to optimize total surfactant-to-cosolvent ratio and total surfactant concentration.

The most promising formulation for application in the dolo-mite reservoir was a mixture of C16-17-(PO)7-SO4, C15-18IOS,and SBA. At 1 wt% surfactant concentration and 2 wt% SBA,

Fig. 2—Volume fraction diagram for a 0.75% C16-17-[PO]7-SO4,0.25% C15-18 IOS, 2% SBA with WT crude at 38�C after 21 days.

Fig. 3—Equilibration of a surfactant formulation containing0.75% C16-17-(PO)7-SO4, 0.25% C15-18IOS, 2% SBA with WT crudeat 38�C slightly below optimum salinity with and without sodi-um carbonate.


this formulation equilibrates rapidly (Fig. 9) and exhibits a solubi-lization ratio of about 12 at optimal conditions (Fig. 7). Accordingto the Huh equation, this corresponds to an IFT of around 2�10-3,which is high enough to mobilize most of the waterflood residualoil. This formulation is also tolerant of divalent cations such ascalcium and magnesium. Most importantly, this formulation pro-duces low viscosity microemulsions with fluid interfaces and littletendency to exhibit gels or macroemulsions. This can be observedqualitatively by tilting pipettes and observing the interface fluidi-ty, as illustrated in Fig. 10. The microemulsion viscosity nearoptimum salinity without polymer was measured with a ContravesLS-30 low shear viscometer and found to be 10 cp. The deadcrude oil viscosity is 11 cp.

Another important step in screening surfactants is to test theiraqueous stability over a range of temperature and salinity appro-priate to the field application. Some formulations show excellentbehavior when mixed with crude oil, but show limited aqueousstability. The C16-17-(PO)7-SO4/C15-18IOS/SBA mixture formed aclear, stable aqueous phase at room temperature over a wide rangeof salinity, but it showed a cloud point as the temperatureincreased, which is more characteristic of non-ionic surfactants.We attribute this behavior to the large number of POs in the C16-

17-(PO)7-SO4 molecule since the same behavior was observed byAoudia et al. (1995) when they increased the number of POs in aseries of propoxy PO sulfate studies.

Polymer Screening. Including polymer in a surfactant slug isessential for maintaining a favorable mobility ratio since the sur-factant causes the water relative permeability to increase. Thisincrease must be counter balanced by decreasing the aqueousmobility with polymer (Hirasaki and Pope 1974). Without poly-mer in the surfactant slug, the surfactant will finger into the oilbank and the reservoir sweep will be very poor. Furthermore, thepolymer in both the slug and drive helps mitigate the effects ofpermeability variation and improves the overall sweep efficiencyin the reservoir. Because corefloods show some—but not all—ofthe above mentioned benefits of adding polymer, acceptableresults in a coreflood with little or no polymer can be misleadingwith respect to performance in the field.

Xanthan gum (Kelco, Atlanta, Georgia) and hydrolyzedpolyacrylamide (HPAM) (SNF Floerger, Andrezieux, France)polymers were selected for use with the surfactant-polymer (SP)coreflood experiments in this study (Levitt 2006; Hirasaki et al.2006). Polymers were tested first in phase-behavior experimentsin order to determine compatibility with the surfactant solutions.In addition to compatibility testing, the relationship between vis-cosity, polymer concentration, and salinity was measured. Thevariation in solution viscosity as a function of polymer concen-tration and electrolyte concentration for Flopaam 3330S HPAM(SNF Floerger, Andrezieux, France) is shown in Fig. 11. It shouldbe noted that some of these measurements were made at slightly

Fig. 7—Solubilization ratio of 0.75% C16-17-(PO)7-SO4, 0.25%C15-18 IOS, 2% SBA formulation with WT crude at 38�C.

Fig. 4—Solubilization ratio of 0.75% C16-17-[PO]7-SO4, 0.25%C15-18 IOS, 2% SBA formulation with WT crude at 38�C with andwithout sodium carbonate.

Fig. 5—Solubilization ratios of 2% C20-24 AOS, 4% SBA formula-tion with a California crude at 100�C with and without sodiumcarbonate.

Fig. 6—Phase behavior of 1.5% C16-17-(PO)7-SO4, 0.5% C20-24

AOS, 4% SBA with WT crude at 38�C.


higher shear rates, causing a slightly lower viscosity. The Flo-paam 3330S polymer solutions showed excellent filtration, whichis another very important screening criterion before corefloodingis started. The mobility of the surfactant-polymer slug and poly-mer drive is also affected by the permeability reduction caused byHPAM polymers, which is defined as the ratio of apparent viscos-ity to bulk viscosity of the polymer at an equivalent shear rate, andis closely related to the residual resistance factor. The permeabili-ty reduction factor for Flopaam 3330S in the dolomite reservoirrock is shown as a function of permeability in Fig. 12. Thisrelationship is needed both for coreflooding design and for model-ing of the reservoir flood. Although the viscosity of the HPAMpolymer is less than the xanthan gum polymer at high salinitiescorresponding to the formation brine and slug salinities, theHPAM is less expensive than the xanthan gum and, thus, is likelyto be the more economical choice in this particular application.The HPAM is also less subject to biodegradation and has otheradvantages in this application.

Coreflooding Experiments. A small number of corefloods wereperformed as validation experiments for promising surfactantsidentified during phase-behavior screening. Berea sandstone wasused for the initial corefloods. Later floods were performed incores created from combined reservoir plugs. These dolomite core

plugs presented the additional complexities of low permeability,high heterogeneity including visible vugs and calcium pickupfrom anhydrite dissolution.

Berea Core Preparation. Berea cores approximately 30 cmlong and 5 cm in diameter were drilled, dried, and weighed. Lexanendpieces were attached to the core with 5-min epoxy and allowedto cure. Once the endpieces were attached, the core was placedinside a lexan tube, and the space between the core and the tubewas filled with slow-setting epoxy. Once epoxy had set, pressuretaps were added 5 cm from each face. A vacuum pump was usedto evacuate air from the core and several pore volumes of CO2

were flushed to remove the air. Next, the cores were brought up to38�C and flooded with synthetic formation brine (SFB) containing60,000 ppm total dissolved solids (TDS), including 2600 ppmdivalent cations (Ca++ and Mg++). Pressure data were recordedand brine permeability calculated. The cores were then floodedwith reservoir crude oil at a high pressure gradient (100 psi/ft)until no water was produced in the effluent. Produced brine wascollected and pressure and flow rate were monitored. After agingbetween 2 and 30 days, cores were flooded with SFB at a pressuregradient similar to the reservoir away from the wells. Producedfluids and pressure were monitored until no additional oil wasproduced. Table 3 describes properties of a Berea sandstone coreas well as other details of surfactant flood Experiment B-1.

Dolomite Core Preparation. Dolomite cores from the WTdolomite reservoir were prepared in a similar fashion to Bereasandstone cores with a few modifications. Because of the limitedplug lengths, cores were prepared by combining several dolomiteplugs to reach a total length of approximately 30 cm. The plugswere 3.8 cm in diameter in some experiments and 5.7 cm indiameter in other experiments. Before combination, plugs werecarefully smoothed with sandpaper and fines were removed byultrasonic bath such that effects of the discontinuity normal to theflow direction would be minimal. Because of the low permeabilityand high heterogeneity of many of the core plugs, they were firstscreened using an air minipermeameter and in a few cases withhigh-resolution X-ray computed tomography (CT).

The dolomite cores were then saturated and flooded with 2500ppm TDS synthetic waterflood source water and effluent wascollected and analyzed for Ca++ and SO4

2- from anhydrite disso-lution. In some cores, a small slug of isopropyl alcohol wasinjected as a tracer and displaced with additional brine to deter-mine heterogeneity. Remaining preparations proceeded in a simi-lar manner to those described for Berea cores.

Since a limited number of core plugs were available, some ofthe dolomite cores were cleaned for reuse by successive injectionsof several pore volumes of surfactant-polymer slug, polymerdrive, freshwater, hypochlorite, freshwater, IPA, freshwater, andfinally synthetic formation brine. The core was considered re-stored and ready for further use when the original brine perme-ability was restored. Experiment D-7C is an example of this; acore originally used for Experiment D-7A was cleaned by thisprocedure and then used for experiment D-7C. Table 3 describesproperties of a WT dolomite core and other details of surfactantflood Experiment D-7C.

The specific Brunauer-Emmett-Teller (BET) surface area foran adjacent rock sample was measured at Rice University andfound to be 0.13 m2/g. Thin sections made from this rock wereanalyzed by the Bureau of Economic Geology at The Universityof Texas and described as grain-dominated dolomitic packstone.

Surfactant-Polymer Flood Design. A salinity gradient wasdesigned to maximize the region of ultralow IFT (Pope et al.1979). In order to make conditions in the sandstone flood similarto those in the reservoir dolomite core, calcium was added asCaCl2 in an amount consistent with the solubility product ofCaSO4. Polymer was included in the slug and drive at a concen-tration that was determined necessary for favorable mobility con-trol (Hirasaki and Pope 1974). During surfactant-polymer slugand polymer drive injection, pressure was monitored and main-tained at approximately 2 psi/ft by adjusting the flow rate. Frontaladvance rates were on the order of 1 ft/day. Although it iscommon to do such corefloods at this rate, the more important

Fig. 8—Equilibration of a surfactant formulation containing1.5% C16-17-(PO)7-SO4, 0.5% C20-24 AOS, 4% SBA, 1.7 wt% NaClwith WT crude at 38�C.

Fig. 9—Equilibration of a surfactant formulation containing0.75% C16-17-(PO)7-SO4 and 0.25% C15-18 IOS with 4.5 wt% NaCl,1375 ppm CaCl2 and WT crude at 38�C.


consideration is maintaining a low pressure gradient so the resultswill be scaleable to the field and useful in simulations studies.

Effluent was collected by a fraction collector and analyzed foroil and surfactant content. Some effluent samples were placed inan oven at 100�C for 36 hours to hydrolyze the sulfate surfactantand allow easier reading of oil recovery. Table 3 summarizes thefluids injected in two typical corefloods. Corefloods were evalu-ated on the basis of oil recovery, pressure gradient, and surfactantretention. It should be noted that, while in field applications,chemical mass is determined by economic optimization, the goalin the laboratory is to determine values such as recovery efficien-cy and surfactant retention.

Berea Sandstone Coreflood Results. In Experiment B-1, a 0.5pore volume (PV) SP slug with 1 wt% surfactant concentrationfollowed by a 2 PV polymer drive was injected into a Bereasandstone prepared as described above (Table 3). Oil recovery,pressure data, and surfactant recovery are presented in Figs. 13through 15, respectively. Approximately 91% of the waterfloodresidual oil was recovered, which corresponds to a residual oil

saturation to chemical (Sorc) of 0.03. Pressure drop remained onthe order of 2 psi/ft. Approximately 90% of the injected surfactantwas recovered. Only 0.08 mg of surfactant was retained per gramof rock. As discussed previously, mechanisms of surfactant reten-tion include both adsorption on to rock surfaces and phasetrapping. This extremely low retention value is indicative of mini-mal phase trapping, which we attribute to low microemulsionviscosity and minimal tendencies of this formulation to formviscous gels and macroemulsions. The favorable salinity gradient,with the drive salinity equal to approximately 0.6 times the opti-mum salinity (Table 3) is also beneficial in terms of low surfac-tant retention. When feasible, a salinity gradient will almostalways result in a more efficient chemical flood.

WT Dolomite Coreflood Results. The reservoir targeted for sur-factant EOR is a heterogeneous dolomite containing anhydrite,and, thus, presents several challenges to coreflooding. The hetero-geneous, vuggy nature of the rock, as seen clearly in CT scans(Fig. 16) can lead to channeling on the core scale because of thelarge size of the vugs with respect to the core. Although this is not

Fig. 10—Qualitative observations of phase behavior, interface fluidity and viscosity by tilting pipettes.

Fig. 11—Viscosity for Flopaam 3330S in three different brines.Waterflood source water contains approximately 2500 ppmTDS.

Fig. 12—Permeability reduction factor for Flopaam 3330Spolymer in dolomite cores.


expected to occur on larger scales corresponding to the actualreservoir, it can complicate interpretation of coreflood results byleading to early tracer and surfactant breakthrough. A tracer wasinjected in to the core used in experiments D-7A and D-7C fol-lowing a waterflood (So=0.4) and breakthrough was observed afteronly 0.25 PV (Fig. 17). This is equivalent to a breakthrough at0.42 PV on a 100% water saturation basis. Such extreme hetero-geneity can be partially mitigated by increasing polymer concen-tration, so high concentrations of polymer were used in most ofthe coreflood experiments. Furthermore, the oil in the reservoir isless viscous (5 cp) than the dead crude (11 cp) used in the labora-tory experiments and the brine relative permeability is higher inthe mixed-wet dolomite reservoir than it is in the cleaned coreplugs. Thus, higher polymer concentration is required in theseparticular corefloods to maintain a favorable mobility ratio thanit will be in the field. A better procedure would have been to agethe cores at high temperature and convert them to mixed-wetcondition, but such a procedure (Hirasaki et al. 2004) is time

consuming and was not used in this screening study since its focuswas on surfactant selection, not mobility control.

Anhydrite (CaSO4) was detected visually from thin sections,by X-ray diffraction (XRD), and by the presence of sulfate andincreased concentrations of calcium in coreflood effluent watersamples at levels consistent with the solubility product of CaSO4.This had several implications for the flood design. First, sodiumcarbonate was not included in the formulation because weexpected the CO3

2- would react with the dissolving anhydriteand precipitate as calcium carbonate. The economic benefits ofusing sodium carbonate are sufficiently large that we plan furtherinvestigations of its use, but it was not feasible or necessary to useit in this surfactant screening study. Also, the effect of this addi-tional dissolved calcium on phase behavior and viscosity had to betaken into account.

In Experiment D-7C, a 0.5 PV SP slug with 1 wt% surfactantconcentration followed by a 2 PV polymer drive was injected intoa WT dolomite core prepared as described above (Table 3). Oilrecovery and pressure data for the WT reservoir dolomite core-flood are presented in Figs. 17 and 18, respectively. Despite earlysurfactant breakthrough, 93% of waterflood residual oil was re-covered, leaving the Sorc at approximately 0.03. The pressuregradient was low at all times as it must be in the actual reservoiraway from the wells. This is one of the most stringent require-ments of a coreflood experiment. In some ways it is even moreimportant than high oil recovery because if the surfactant does not

Fig. 15—Surfactant Recovery during Berea Coreflood B-1.Fig. 14—Pressure drop data for Berea Coreflood B-1 duringsurfactant flood and polymer drive.

Fig. 13—Oil recovery during Berea Coreflood B-1.


transport under low pressure gradient, then it will not even contactthe oil deep in the reservoir with large well spacing.

Surfactant retention in this experiment was 0.24 mg/g of rock.This is slightly higher than the mean of several experiments per-formed in this reservoir dolomite, with lower and higher valuesobtained respectively when slug salinity was either significantlylower or slightly higher than optimum (S* in Table 4). The lowretention may be in part because of the low specific surface areafor the dolomite rock.

Conclusions

Several high-performance, low-cost surfactants for EOR havebeen identified and tested starting with an efficient and effectivelaboratory screening process that emphasizes observations ofphase behavior and viscosity. This screening process has beenshown to provide reliable selection of cost effective surfactants,

cosurfactants, and polymers as a function of salinity, hardness,alkali, temperature, and rock type. Surfactant structures withbranched hydrophobes are the most suitable for EOR becausemicroemulsions made from these surfactants show little tendencyto form viscous phases (such as gels and liquid crystals) that resultin high surfactant retention. Adding propylene oxide to the surfac-tant improves its performance with little increase in cost and is avery practical approach to tailoring the surfactant to the specificcrude oil and reservoir conditions. The surfactants investigatedgave high oil recoveries and low residual oil saturation (Sorc=0.03)in both Berea sandstone and a dolomite reservoir rock floodedunder low pressure gradients typical of values in oil reservoirs.These surfactants also showed low retention in both Berea sand-stone (0.08 mg/g) and in the dolomite reservoir rock (0.24 mg/g).The low surfactant retention is attributed in part to low viscositymicroemulsions that were selected using the qualitative phasebehavior testing described in this paper. Sodium carbonate was

Fig. 16—CT scans of a WT dolomite core plug. These data and images were produced at the high-resolution X-ray CT facility of TheUniversity of Texas at Austin.

Fig. 17—Tracer breakthrough data in dolomite core (D-7C) atSorw = 0.4. Fig. 18—Oil recovery during dolomite Coreflood D-7C.


observed to speed up coalescence to equilibrium microemulsionsand is also expected to decrease surfactant adsorption on bothsandstone and carbonate rocks, so its use in some applicationsshould provide additional benefits. A simulation study presentedin a companion paper (Anderson et al. 2006) indicates that theeconomics of such high-performance synthetic surfactants arevery attractive at crude oil prices of about 30 USD/bbl or moreeven in a heterogeneous dolomite reservoir, and would be muchmore so in many sandstone and carbonate reservoirs in whichsodium carbonate could be used.

Nomenclature

C = Chun Huh constant

kbrine = permeability to synthetic formation brine

krw* = relative permeability endpoint to water

Rk = permeability reduction factor

S* = optimum salinity

Sorc = residual oil saturation to chemical

Sorw = residual oil saturation to water

Vo = volume of oil

Vs = volume of surfactant on 100% active basis

Vw = volume of water

g = IFT

f = porosity

s = solubilization ratio

Acknowledgments

The authors would like to acknowledge financial support for thisresearch from the Department of Energy (DOE contract number:DE-FC26-03NT15406) as part of a joint project with RiceUniversity and Intera. We would also like to thank Oxy Permianfor providing crude oil, cores, and reservoir data; Stepan, Shell,and Sasol for providing surfactants; and SNF for providing the

polymers used in this study. We are especially grateful to HerbWacker with Oxy Permian for all of his help with this study. Wethank laboratory assistants Ghazal Dashti, Chris Riley, VaibhavSharma, and Ping Zhao for help with the experiments; ShunhuaLiu at Rice University for BET measurements, and Jerry Luciaand Jim Jennings at the Bureau of Economic Geology at TheUniversity of Texas at Austin for thin section analysis. We wouldalso like to acknowledge the resources and staff of the Centerfor Petroleum and Geosystems Engineering (CPGE) at TheUniversity of Texas at Austin.

References

Adams, W.T. and Schievelbein, V.H. 1987. Surfactant Flooding Carbon-

ate Reservoirs. SPE Res Eng 2 (4): 619–626. SPE-12686-PA. DOI:

10.2118/12686-PA.

Anderson, G.A., Delshad, M., King, C.B., Mohammadi, H., and Pope, G.

A. 2006. Optimization of Chemical Flooding in a Mixed-Wet

Dolomite Reservoir. Paper SPE 100082 presented at the SPE/DOE

Symposium on Improved Oil Recovery, Tulsa, 22–26 April. DOI:

10.2118/100082-MS.

Aoudia, M., Wade, W.H., and Weerasooriya, V. 1995. Optimum

microemulsions formulated with propoxylated tridecyl alcohol sodium

sulfates. J. of Dispersion Science and Technology 16 (2): 115–135.

DOI:10.1080/01932699508943664.

Bourrel, M. and Schechter, R.S. 1988. Microemulsions and Related

Systems: Formulation, Solvency, and Physical Properties, Vol. 30.

New York: Surfactant Science Series, Marcel Dekker.

Healy, R.N., Reed, R.L., and Stenmark, D.K. 1976. Multiphase

Microemulsion Systems. SPE J. 16 (3): 147–160; Trans., AIME, 261.

SPE-5565-PA. DOI: 10.2118/5565-PA.

Hirasaki, G. and Zhang, D.L. 2004. Surface Chemistry of Oil Recovery

From Fractured, Oil-Wet Carbonate Formations. SPE J. 9 (2): 151–162.

SPE-80989-PA. DOI: 10.2118/88365-PA.

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. SPE-4026-PA. DOI: 10.2118/

4026-PA.

Hirasaki, G.J., Miller, C.A., and Pope, G.A. 2006. Surfactant Based

Enhanced Oil Recovery and Foam Mobility Control. Technical Re-

port, Contract No. DE-FC26-03NT15406, Rice University, Houston,

Texas (June 2005).

Huh, C. 1979. Interfacial tensions and solubilizing ability of a microemul-

sion phase that coexists with oil and brine. J. of Colloid and Interface

Science 71 (2): 408–426. DOI:10.1016/0021-9797(79)90249-2.

Jayanti, S., Britton, L.N., Dwarakanath, V., and Pope, G.A. 2002.

Laboratory evaluation of custom-designed surfactants to remediate

NAPL source zones. Environmental Science and Technology 36 (24):

5491–5497. DOI:10.1021/es020566i.

Levitt, D. 2006. Experimental evaluation of high performance EOR

surfactants for a dolomite reservoir. MS thesis, University of Texas at

Austin, Austin, Texas.

Liu, S., Leslie, D., Yan, W., Puerto, M., Hirasaki, G.J., and Miller, C.A.

2008. Favorable Attributes of Alkali-Surfactant-Polymer Flooding.

SPE J. 13 (1): 5–16. SPE-99744-PA. DOI: 10.2118/99744-PA.

Nelson, R.C. and Pope, G.A. 1978. Phase Relationships in Chemical

Flooding. SPE J. 18 (5): 325–338; Trans., AIME, 265. SPE-6773-PA.

DOI: 10.2118/6773-PA.Fig. 19—Pressure-drop data for dolomite Coreflood D-7C dur-ing surfactant flood and polymer drive.


Pope, G.A., Wang, B., and Tsaur, K. 1979. A Sensitivity Study of Micel-

lar/Polymer Flooding. SPE J. 19 (6): 357–368. SPE-7079-PA. DOI:

10.2118/7079-PA.

Stegemeier, G.L. 1976. Mechanisms of oil entrapment and mobilization in

porous media. Proc., AIChE Symposium on Improved Oil Recovery

by Surfactant and Polymer Flooding, Kansas City, Missouri, USA,

12–14 April, 55–91.

Winsor, P.A. 1954. Solvent Properties of Amphiphilic Compounds.

London: Butterworth’s Scientific Publications.

Wu, Y., Shuler, P., Blanco, M., Tang, Y., and Goddard, W.A. 2005. A

Study of Branched Alcohol Propoxylate Sulfate Surfactants for

Improved Oil Recovery. Paper SPE 95404 presented at the SPE

Annual Technical Conference and Exhibition, Dallas, 9–12 October.

DOI: 10.2118/95404-MS.

David Levitt is earning a PhD degree (May 2009) in petro-leum engineering at The U. of Texas at Austin, where he hasworked as a graduate research assistant at the CPGE forthe past 6 years. Email: [email protected]. Heholds a BS degree in general engineering from HarveyMudd College and an MS degree in petroleum engineeringfrom The U. of Texas at Austin. Levitt’s areas of interestinclude polymer stability and injectivity, as well as economicoptimization of EOR processes. He intends to work for TotalS.A. beginning in June. Adam Jackson worked as a researchassistant for the CPGE while attending graduate school at The

U. of Texas at Austin. He joined Chevron as a reservoir engineerin 2006. Jackson’s work focuses on chemical EOR for severalsiliciclastic reservoirs located in Indonesia and the North Sea.He holds a BS degree in geosystems engineering and hydro-geology and an MS degree in petroleum engineering fromThe U. of Texas at Austin. Christopher Heinson worked as aresearch assistant for the CPGE while studying at The U. ofTexas at Austin. He joined Occidental Petroleum as a reservoirengineer in 2007. Heinson’s work focuses on the developmentof mature carbonate reservoirs in the Permian basin. He holdsa BS degree in petroleum engineering from The U. of Texas atAustin. Larry Britton is a research scientist at the CPGE at theU. of Texas. He holds a PhD degree from the U. of Iowa and anMS degree from Louisiana State U. Britton worked 7 years asSenior Scientist at Sasol, USA (formerly Condea Vista) where hedeveloped surfactants for subsurface remediation. His re-search interests at the U. of Texas include surfactantdevelopment for EOR. Taimur Malik is Team Leader for thechemical EOR group at Chevron. Before Chevron, he workedfor INTERA in Austin, Texas, and was an active participant inthe DOE sponsored research program that resulted in thispaper. Malik holds a BS degree in petroleum engineeringfrom The U. of Texas at Austin. Varadarajan Dwarakanath isTeam Manager for reservoir performance characterizationat Chevron that includes chemical EOR consulting. Email:[email protected]. Before working for Chevron, heworked for INTERA in Austin, Texas. Dwarakanath holds a BSdegree in mining engineering from Banaras Hindu U. and MSand PhD degrees in petroleum engineering from The U. ofTexas at Austin. Gary A. Pope is the Director of the CPGEat the U. of Texas at Austin, where he has taught since1977. He holds the Texaco Centennial Chair in Petroleum Engi-neering. Previously Pope worked in production research atShell Development Company for 5 years. He holds a PhD de-gree from Rice U. and a BS degree from Oklahoma State U.,both in chemical engineering. Pope’s teaching and researchare in the areas of EOR, reservoir engineering, natural gasengineering, reservoir simulation, characterization of reservoirsand aquifers with tracers, surfactants, and water-solublepolymers, phase behavior and fluid properties, and ground-water modeling and remediation.

SI Metric Conversion Factors

bbl � 1.589 873 E – 01 = m3

cp � 1.0* E – 03 = Pa�sdyne � 1.0* E – 02 = mNft � 3.048* E – 01 = mpsi � 6.894 757 E + 00 = kPa

*Conversion factor is exact.


identification and evaluation of high-performance eor surfactants

Documents