Formulating Chlorinated Hydrocarbon ... - ACS Publications

100 East Boyd, Norman, Oklahoma 73019, and ... and oil phase (lipophilic linkers) or between the surfactant and water phase (hydrophilic linkers). Com...
0 downloads 0 Views 129KB Size
Download PDF
Environ. Sci. Technol. 2002, 36, 4618-4624

Formulating Chlorinated Hydrocarbon Microemulsions Using Linker Molecules EDGAR ACOSTA,† SAN TRAN,† HIROTAKA UCHIYAMA,† D A V I D A . S A B A T I N I , * ,‡ A N D JEFFREY H. HARWELL† Chemical Engineering and Materials Science Department, University of Oklahoma, Sarkeys Energy Center, 100 East Boyd, Norman, Oklahoma 73019, and Civil Engineering and Environmental Science Department, University of Oklahoma, Carson Engineering Center, 202 West Boyd, Norman, Oklahoma 73019

Previously we reported on the use of lipophilic, hydrophilic, and combined linkers for formulating microemulsions of trichloroethylene and tetrachloroethylene. These linker molecules augment the interaction between the surfactant and oil phase (lipophilic linkers) or between the surfactant and water phase (hydrophilic linkers). Combining both linkers can increase the solubilization capacity severalfold. This formulation technique shows potential advantage in reducing surfactant costs and remedial times, as well as allowing the use of more environmentally friendly additives when designing surfactant-enhanced remediation systems. In this paper, we evaluate the relative importance of the surfactant and each linker in increasing the system’s solubilization capacity. This interpretation is based on solubilization curves, partitioning studies, and formulation studies. The solubilization curves show that optimum linker performance is reached at an equimolar ratio of dodecanol, used as a lipophilic linker, and sodium mono and dimethyl naphthalene sulfonate, used as a hydrophilic linker. Furthermore, this equimolar combination was able to replace the anionic surfactant sodium dihexylsulfosuccinate. Dodecanol partitioning at optimum formulation shows that the poor performance of lipophilic linkers alone is due to their low partitioning into the middle phase. Adding hydrophilic linkers into this system increases the fraction of dodecanol in the middle phase, thereby further enhancing the solubilization capacity of the system. A variation of the combined linker approach is introduced by changing a lipophilic linker, oleic acid, into a surfactant (soap), with further increases in the solubilization capacity by 4- to 5-fold.

Introduction The term microemulsions is used to describe thermodynamically stable emulsions. Practical applications of microemulsions systems include enhanced oil recovery (EOR), drug delivery, nanoparticles synthesis, etc. (1, 2). Of particular * Corresponding author phone: (405) 325-4273; fax: (405) 3254217; e-mail: [email protected]. † Chemical Engineering and Materials Science Department. ‡ Civil Engineering and Environmental Science Department. 4618

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 21, 2002

interest is their use in surfactant-enhanced aquifer remediation (SEAR) as solubilization/displacement media of nonaqueous phase liquids (NAPLs) trapped in subsurface porous aquifers (3-13). Microemulsions can be of three basic types. A Type I microemulsion corresponds to oil solubilized in aqueous micelles (surfactant aggregates having a hydrophobic interior and hydrophilic exterior). A Type II microemulsion corresponds to water solubilized in reverse micelles present in the oil phase. A Type III microemulsion corresponds to an oil and water bicontinuous phase that is stabilized by a surfactant membrane. The Type I-III-II transition can be achieved in several ways. For ionic surfactants the most common way is by increasing the electrolyte concentration. The amount of electrolyte necessary to achieve equal amounts of oil and water in a Type III microemulsion (optimum formulation) is called the optimum salinity (S*). At optimum formulation the oil solubilization is maximum and the oil/ water interfacial tension is minimum. These characteristics make optimum formulations of special interest in removal of NAPLs from porous media. Microemulsions of chlorinated hydrocarbons have attracted special interest based on their potential use to remediate sites contaminated with chlorinated hydrocarbons (3-13). Both anionic and nonionic surfactant systems have been evaluated for this purpose, with sodium dihexyl sulfosuccinate being one of the most commonly used surfactants (3-5, 10). The advantages of sodium dihexyl sulfosuccinate include its low adsorption on the media, its low toxicity (it is a food-grade additive), and its low tendency to form macroemulsion and gels (it can form microemulsions without adding short chain alcohols) (8). For dense nonaqueous phase liquids (DNAPLs), such as chlorinated hydrocarbons, the use of microemulsions systems has to consider vertical migration concerns. Pennel et al demonstrated that DNAPL interfacial tension reduction not only makes it easier for the viscous forces to displace the oil through the pores but also makes it easier for gravity forces to displace the oil deeper into the aquifer (12). Because of the vertical migration concerns associated with mobilization (ultralow interfacial tension), enhanced solubilization of chlorinated hydrocarbons using swollen micelles (Type I microemulsions) has been proposed as a preferred approach (12). According to Pennel et al., the trapping number, which includes bond and capillary numbers, describes the residual saturation according to the “desaturation curves”. For tetrachloroethylene (PCE) and a mixture of sodium dihexyl and dioctyl sulfosuccinate, and using typical flow conditions, they found that interfacial tension (IFT) values below 1 mN/m caused vertical migration concerns (12). Using similar desaturation curves and flow conditions for trichloroethylene (density 1.46 g/mL), the threshold interfacial tension to avoid mobilization is around 0.5 mN/m. Because the solubilization capacity increases inversely to the square root of the interfacial tension (14), we can maximize the solubilization enhancement while mitigating vertical migration concerns by designing our surfactant system at this IFT. This system has been referred to as supersolubilization (13). Sabatini et al. (13) introduced a gradient method to avoid vertical migration while maximizing the solubilization capacity. In this approach, which has been tested in the laboratory, the formulation is changed throughout the remediation to continuously increase the solubilization as the contaminant residual saturation decreases and the system can handle a lower interfacial tension without mobilization. 10.1021/es0158859 CCC: $22.00

 2002 American Chemical Society Published on Web 09/25/2002

FIGURE 1. Schematic of the linker effect: surfactant, sodium dihexyl sulfosuccinate; lipophilic linker, dodecanol; hydrophilic linker, sodium mono and dimethyl naphthalene sulfonate. In summary, a successful formulation for removing chlorinated hydrocarbons from aquifers requires that these oils be solubilized in the surfactant solution, and not mobilized, which can lead to vertical migration. The desire to maximize the solubilization capacity of microemulsion systems must thus be tempered by this serious environmental concern. The economics of the SEAR technology are impacted by the ability to regenerate the surfactant solution, ability to dispose the waste generated, and the time (pore volumes) that it takes to achieve the remediation goals (9). The greater the solubilization capacity for a given interfacial tension, the fewer pore volumes of surfactant solution needed, which reduces capital expenditure and operation cost (equipment and manpower hours). Recently we have found that by using hydrophilic and lipophilic linker molecules we can enhance the solubilization capacity of sodium dihexyl sulfosuccinate (Aerosol-MA, AMA) for trichloroethylene (TCE) and tetrachloroethylene (PCE) (3). Intuitively, the linker effect can be understood as using chemical additives molecules to “extend” the interaction of the surfactant with the water phase (hydrophilic linkers) or the interaction between the surfactant and oil phase (lipophilic linkers) (3, 15-18). Figure 1 shows a molecular schematic of the linker effect. In this schematic, the surfactant, sodium dihexyl sulfosuccinate, is adsorbed at the oil/water interface. Lipophilic linker molecules (initially proposed by Graciaa et al. (15-17)), such as dodecanol in Figure 1, are shown to adsorb at the palisade layer of the interface (oil side of the surfactant layer) promoting the local order and increasing the interaction between the surfactant tail and oil molecules (15-17). Hydrophilic linker molecules, such as sodium mono and dimethyl naphthalene sulfonate (SMDNS) in Figure 1, adsorb at the O/W interface but, due to its poor interaction with the oil phase, these molecules increase the overall interaction between the surfactant layer and the aqueous phase (3, 18). Combining both hydrophilic and lipophilic linkers “the combined linker effect” in Figure 1, promotes a synergistic effect, which can increase the solubilization capacity for chlorinated hydrocarbons and hexane microemulsions (3). Thus, by using individual linker molecules we can alter the interaction of the surfactant interface with either the oil or water phase; while the combined linkers can substitute for a portion of the surfactant. Here, we plan to investigate the nature of this synergistic effect and its implications for designing chlorinated hydrocarbon microemulsions for aquifer remediation. Previously, we used long chain alcohols as lipophilic linkers for TCE, PCE, and hexane microemulsions (3). We observed that lipophilic linkers had limited ability to increase the solubilization capacity (i.e., became ineffective at certain dodecanol concentrations used as lipophilic linker). For TCE the threshold concentration for the linker effect was ∼0.09

M of dodecanol, but for PCE it was 0.24 M. The hydrophilic linker sodium mono and dimethyl naphthalene sulfonate was added as a way to compensate the lipophilic linker effect and adjust the hydrophilic lipophilic balance (HLB) of the surfactant layer. It was found that combining hydrophilic and lipophilic linkers further increased the solubilization capacity of chlorinated hydrocarbon microemulsions (3). More recently, the role of hydrophilic linkers has been studied in greater detail, showing that a hydrophilic linker is an amphiphile molecule that coadsorbs at the oil/water interface with the surfactant but has a weak interaction with the oil molecules, thereby increasing the interaction between the surfactant at the interface and water (18). Sodium mono and dimethyl naphthalene sulfonate and sodium octanoate have been identified as a hydrophilic linker for chlorinated hydrocarbon microemulsions. The use of linker molecules also increases the range of additives available to the formulator. Many of the linker molecules may be more appropriate for environmental applications and less expensive than common surfactants. For example, although dodecanol is studied as the lipophilic linker in this research, alternative lipophilic linkers include fatty acids and alcohols, nonionic surfactant with low degree of ethoxylation, and sorbitol esters (3, 15-18). When designing an actual field implementation this is obviously a very important issue. Some questions were left unanswered in our previous studies. For example, why does the lipophilic linker alone become ineffective? What is the relative importance of each linker (i.e., is there an optimal hydrophilic to lipophilic linker ratio)? How much surfactant can the combined linkers replace? What is the impact of linker molecules on microemulsion formulation (i.e., optimum salinity)? These questions are addressed in this paper. Also, an alternative approach for achieving the combined linker effect is presented, where a lipophilic linker, oleic acid, is changed into a cosurfactant by saponification (pH adjustment).

Materials and Methods Materials. The following chemicals, obtained from Aldrich (Milwaukee, WI) at the concentrations shown, were used without further purification: trichloroethylene (TCE, 99%+), tetrachloroethylene (PCE, 99%+), n-dodecanol (98+%), sodium chloride (99%+), oleic acid (98%), and sodium hydroxide (99%). An 80 wt.% solution of sodium dihexyl sulfosuccinate (Aerosol-MA, AMA), containing less than 5 wt.% of 2-propanol, was purchased from CYTEC (West Paterson, NJ). Sodium mono and dimethyl naphthalene sulfonate (SMDNS, Morwet M) was supplied by CKWitco (Houston, TX). Methods. Phase behavior studies were performed using equal volumes of aqueous solution and oil (5 mL of each). Electrolyte scans were performed by varying the sodium chloride concentration at constant temperature, additive content (alcohols, acids, hydrotropes, etc.), and pressure (1 atm). Test tubes were placed in a water bath at 27 °C, shaken once a day for 3 days, and left to equilibrate for two weeks. The phase volumes were determined by measuring the levels of each phase in the test tube. The volume of middle phase and the surfactant concentration were used to determine the solubilization parameter. The solubilization parameter is the amount of oil solubilized in the microemulsion per unit mass of surfactant (AMA). In this paper, the solubilization parameter of optimum middle phase formulation is reported (i.e., when the middle phase is composed of half oil and half water) as an indication of the solubilization capacity of the microemulsion system. Dodecanol concentrations were measured by gas chromatography using a Varian 3300 with FID detector and a SPB20 capillary column with programmed temperature. VOL. 36, NO. 21, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

4619

FIGURE 2. Isosolubilization curves for TCE microemulsions with 0.103 M sodium dihexyl sulfosuccinate (AMA) at different hydrophilic sodium mono and dimethyl naphthalene sulfonate (SMDNS) and lipophilic (dodecanol) linkers.

FIGURE 3. Isosolubilization curves for PCE microemulsions with 0.103 M sodium dihexyl sulfosuccinate (AMA) at different hydrophilic sodium mono and dimethyl naphthalene sulfonate (SMDNS) and lipophilic (dodecanol) linkers. For optimum middle-phase TCE microemulsions, the interfacial tension was measured using a University of Texas spinning drop interfacial tensiometer model 500, injecting 1 to 5 µL of the middle phase in a 300-µL tube filled with the excess trichloroethylene from 10-mL microemulsion sample tube.

tion relative to the total volume (10 mL) in the test tube (i.e., 50% means 5 mL of middle phase and 2.5 mL each of oil and water). The y-axis in Figures 2 and 3 represents the molar concentration of dodecanol (lipophilic linker) added to the aqueous solution. The x-axis represents the molar concentration of SMDNS (hydrophilic linker) added to the aqueous solution. The numerical values on the isosolubilization curves represent the solubilization values (i.e., the volume fraction of the optimum middle phase). Thus, in Figure 2 a solubilization value of 30 (30%) can be achieved by adding 0.1 M dodecanol to the AMA formulation (no SMDNS), by adding 0.27 M SMDNS (no dodecanol), or by adding a range of mixed linker concentrations (e.g., 0.05 M dodecanol and 0.10 M SMDNS). To help interpret these isosolubilization plots, it should be noted that for a given isosolubilization line, the closest point to the origin (0, 0) will correspond to the most efficient linker combination (require the least amount of both linkers). The molar ratio of hydrophilic linker to lipophilic linker at which the most efficient combination occurs will be the optimum ratio. Following this procedure, it is observed (Figures 2 and 3) that for both TCE and PCE formulations the optimum ratio occurs when equimolar quantities of SMDNS and dodecanol are used. Furthermore, for SMDNS/dodecanol ratios in the range of 0.25 to 0.5, the efficiency decreases. To compare these results with the initial hypothesis (i.e., relating the linker combination to the HLB concept), the optimum HLB of TCE and PCE will be compared to the optimum SMDNS/dodecanol ratios obtained for each oil. For TCE the optimum HLB is 17 and for PCE it is 13 (7). For both oils the optimum linker ratio was SMDNS/dodecanol ) 1:1. Based on the hypothesis, a ratio that favored lipophilic linkers would be expected for PCE. The first interpretation of these data suggests that the HLB does not explain the synergistic effect of combining linkers. Rather, the data suggest a close interaction between hydrophilic and lipophilic linkers to achieve the optimum performance. Partition of Dodecanol in Lipophilic and Combined Linker Formulations. From the discussion above, the limited effectiveness of lipophilic linkers is not due to their impact on the system’s HLB. Another possible explanation is that partition effects limit the amount of lipophilic linker at the palisade layer (i.e., the adsorption of lipophilic linkers at the palisade layer is competing with its solubilization in the bulk oil). To test this hypothesis, dodecanol partitioning studies were performed when it was added alone and when equimolar dodecanol and SMDNS concentrations existed for TCE and PCE. Because dodecanol solubility in water is very low (∼4 mg/L) (19), we focused our attention on the dodecanol partitioning between the middle phase and the dodecanol present in the excess oil phase, as quantified by the following:

Results and Discussion Hydrophilic versus Lipophilic Linkers. As indicated above, hydrophilic linkers are proposed as a way of compensating for the hydrophilic/lipophilic balance (HLB) modification produced by lipophilic linkers (3). If true, the hydrophilic/ lipophilic linker ratio necessary to achieve maximum solubilization should be a function of the oil HLB. To test this hypothesis, isosolubilization curves for TCE and PCE were prepared at varying concentrations of the hydrophilic linker (SMDNS) and the lipophilic linker (dodecanol), while holding the AMA concentration constant at 0.103 M. This concentration of sodium dialkyl sulfosuccinates is typical in surfactantenhanced remediation applications (3, 12, 18). Figures 2 and 3 show the isosolubilization curves for TCE and PCE, respectively, where solubilization is expressed as a fraction (percent) of the middle phase volume at optimum formula4620

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 21, 2002

Km/o )

[C12H26O]middle phase [C12H26O]oil

(1)

Figure 4 shows the dodecanol partition coefficient (yaxis) as a function of dodecanol concentration (x-axis). There are four curves plotted: dodecanol alone in TCE formulation, dodecanol alone in PCE formulation, equimolar dodecanol and SMDNS formulation for TCE, and equimolar dodecanol and SMDNS formulation for PCE. The first interesting feature of Figure 4 is that the partitioning coefficient for dodecanol alone and TCE is consistently lower than that for the PCE formulation. This demonstrates that dodecanol partitions more into TCE than into PCE. Wu et al. (20) have studied the partitioning of alcohols between oil phases and water, and found that as the

FIGURE 4. Dodecanol partition into the middle phase as a function of alcohol concentration alone and in combination with hydrophilic linker SMDNS, 0.103 M AMA. oil becomes more polar (i.e. less hydrophobic), more alcohol partitions into the oil, which is consistent with the findings in Figure 4. Another interesting characteristic of the dodecanol alone curves in Figure 4 is that for TCE and PCE formulations the partitioning coefficient reduces as the concentration of dodecanol increases. The partition coefficient for dodecanol in TCE remains at ∼0.60 beyond 0.09 M dodecanol, which coincides with the dodecanol threshold as lipophilic linker for TCE (3). For the PCE formulation, the dodecanol partitioning continuously decreases even beyond 0.18 M. For the case of PCE, we previously found that dodecanol concentrations above 0.24 M produced poor solubilization enhancement (3). The threshold concentration for lipophilic linkers has also been observed by Salager et al. in isooctane microemulsions with nonionic surfactant (17), which we can now understand on the basis of linker partitioning. The competition of lipophilic linker to partition in the middle phase or in the excess oil might resolve a difference of opinions in the microemulsion literature. Kahlweit et al. (21) argued that long chain alcohols are oil cosolvents that help reduce the hydrophobicity of the oil. Graciaa, Salager, Bourrel, and Lachaise (15-17) maintain that it is the ability of a molecule to adsorb at the palisade layer of the interface that defines its role in microemulsions. The partition studies suggest that above a threshold concentration, dodecanol seems to behave as an oil cosolvent equally distributed between the excess bulk phase and the oil present in the middle phase. The third important feature of Figure 4 is that when the hydrophilic linker, SMDNS, is present, the partitioning of dodecanol into the middle phase is either increased or maintained as the dodecanol concentration increases. Before leaving the topic of dodecanol partitioning in the middle phase, there is one point that should be emphasized. In eq 1, the concentration of dodecanol in the middle phase is in mole per liter of optimum middle phase volume, where by definition the optimal middle phase is half oil. If the dodecanol is present in the oil continuous media of the middle phase, it means that its concentration is actually double that used to calculate the partition coefficient. Therefore a partition coefficient of 0.60, as the one found for TCE, would mean that dodecanol is 1.2 times more concentrated in the oil that exists in the middle phase. Dodecanol is always more concentrated in the oil present in the middle phase because of its adsorption at the palisade layer. In summary, the data show that when adding the hydrophilic linker SMDNS in equal molar ratio with dodecanol, more dodecanol tends to self-assemble at the sur-

FIGURE 5. Isosolubilization curves of TCE for varying ratios of surfactant AMA to equimolar mixture of combined linkers. factant membrane of the middle phase. This phenomenon suggests that there exists an intimate relationship between hydrophilic and lipophilic linkers, and that this interaction occurs independent of the oil. In addition, the solubilization enhancement with combined linkers is shown to be synergistic rather than additive. Surfactant vs Combined Linkers. The results presented thus far show that combined linkers are synergistic independent of the oil phase. The question assessed in this section is “How does the surfactant affect the hydrophilic/lipophilic linker performance?” From the previous sections, combined linkers are hypothesized to behave as an “assembled surfactant” that acts independently of the surfactant, thereby replacing the main surfactant, AMA, in certain proportion. To test this hypothesis, isosolubilization curves for TCE were obtained at different surfactant concentrations and varying levels of equimolar dodecanol and SMDNS mixtures. Figure 5 shows the isosolubilization curves at different concentrations of surfactant (sodium dihexyl sulfosuccinate, AMA) on the x-axis and dodecanol concentration in equimolar mixture with SMDNS on the y-axis. Similar to Figures 2 and 3, the numerical values represent the volume fraction of middle phase for a given isocontour. The most relevant characteristic of the isosolubilization curves in Figure 5 is their linearity. This indicates that an equimolar combination of linkers can proportionally replace the surfactant AMA while holding a constant solubilization of oil. The slope of the linear trend in Figure 5 indicates how much surfactant can be replaced for an equimolar combination linkers. To illustrate this point, the data in Figure 5 indicate that a 50% solubilization can be achieved by 0.2 M surfactant AMA or by 0.12 M AMA and 0.2 M combined linker. In this example 0.08 M surfactant AMA can be replaced by 0.2 M combined linker; by dividing these two quantities one finds that approximately 0.4 mol of surfactant can be replaced by 1 mol of dodecanol and 1 mol of SMDNS. The actual value is 0.368 mol of AMA/mol of combined linker. Although isosolubilization curves were not obtained for PCE as above, by using the solubilization enhancement curve previously obtained by Uchiyama (3), it can be estimated that 1.8 mol of AMA can be replaced by 1 mol of dodecanol and 1 mol of SMDNS. The linearity of the curves in Figure 5 supports the hypothesis that the combined linkers behave as an assembled surfactant independently of the main surfactant (AMA) concentration. Although this interpretation is feasible, VOL. 36, NO. 21, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

4621

TABLE 1. Solubilization Parameter and Optimum Salinity for Oleic Acid and NaOH fraction of oleic acid neutralized

concentration of oleic acid

NaOH concentration

solubilization parameter mL/g AMA

interfacial tension mN/m

optimum salinity NaCl, wt %

blank 0% 10% 35% 50% 80% 90%

0 0.09 M 0.09 M 0.09 M 0.09 M 0.09 M 0.09 M

0 0 0.009 M 0.027 M 0.045 M 0.072 M 0.081 M

5.7 ( 0.7 7.4 ( 0.7 11.4 ( 0.7 14.0 ( 0.7 16.5 ( 0.7 22.0 ( 0.7 25.0 ( 0.7

3.5 ( 0.5 E-3 1.7 ( 0.5 E-3 7.0 ( 1 E-4 5.0 ( 1 E-4 3.0 ( 1 E-4 2.0 ( 1 E-4 1.0 ( 0.5 E-4

1.35 0.83 0.96 1.17 1.28 1.70 1.71

nonetheless there was no middle phase formation for AMA concentrations below 0.03 M. It appears that the main surfactant acts as a necessary “seed” for the mixed linker to achieve middle phase formation. From the equivalent ratios of combined linker to surfactant, combined linkers are less efficient in solubilizing TCE than PCE, which would be expected from the dodecanol partitioning results presented in Figure 4. For PCE, dodecanol partitioned more into the middle phase than for TCE, even when combined with hydrophilic linkers. The economic advantage of using combined linkers in microemulsion formulations stems from the ability to replace a more costly surfactant with a mixture of less expensive linkers. This advantage is more noticeable as the amount of surfactant replaced per mol of combined linker increases. In this sense, combined linkers are more economically attractive for PCE than for TCE. As we indicated at the beginning of this article, linker molecules can be selected so as not to increase the environmental risk due to toxicity or biodegradability issues. In contrast, we expect to be able to reduce the use of ionic surfactants that are normally more toxic that nonionic lipophilic linkers. For example, SMDNS and octanoic acid (precursor of sodium octanoate) both have food-additive status (22, 23). In future publications we will further report on alternative nonionic, nontoxic, and biodegradable linker formulations. The problem, then, for TCE formulations is the linker (dodecanol) partitioning into the oil. Even the hydrophilic linker (SMDNS) partially partitions into the middle phase (∼70% of the total SMDNS added) (18). Thus, for TCE, further solubilization enhancement could be achieved if the surface activity of the lipophilic linker is increased so that it can coadsorb at the oil/water interface. Lipophilic Linker vs Cosurfactant Effect. One possible way to increase the surface activity of a lipophilic linker and ensure its adsorption at the O/W interface is to transition from being a lipophilic linker to become a cosurfactant. To test this approach, a lipophilic linker (oleic acid) was fractionally neutralized with sodium hydroxide to produce sodium oleate, thereby increasing its hydrophilicity and making it more like a cosurfactant. Table 1 summarizes formulation characteristics for the partially neutralized oleic acid with 0.103 M AMA. The first row contains these characteristics for the surfactant-only case (blank). The fourth column in Table 1 shows the solubilization parameter, which quantifies the volume of TCE present in the middle phase at optimum formulation per mass of surfactant (AMA in this case). Formulations containing oleic acid show the typical behavior of a lipophilic linker, i.e., increasing the solubilization capacity. According to Table 1, as more oleic acid is neutralized, the solubilization capacity increases. When this fraction is greater than 90%, complete solubilization of oil and water is achieved, i.e., a Winsor Type IV single phase microemulsion is formed. It is remarkable that when only 0.09 M oleic acid is 90% neutralized it can increase by 4.4-fold the solubilization 4622

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 21, 2002

capacity of 0.103 M AMA. This remarkable increase in the solubilization capacity supports the hypothesis that for TCE the low solubilization enhancement with lipophilic linkers is due to poor partitioning of the linker into the middle phase, and not due to a lack of interaction between the linker and the oil. Furthermore, another factor to consider is that sodium oleate has an HLB of 18, which matches the HLB of the TCE. Thus, the sodium oleate is well balanced at the oil/water interface. Despite this good match between sodium oleate and TCE, no middle phase could be obtained using sodium oleate alone; again, the surfactant must be present. The interfacial tension column in Table 1 indicates this parameter at optimum formulation (the condition at which the interfacial tension between the middle phase and excess oil is the same as that of the middle phase and the excess water). From Table 1 it is seen that the interfacial tension consistently decreases with increasing neutralization of oleic acid (and as the solubilization parameter increases). This is consistent with the Chun Huh relationship (14)

γ* )

K′ (SP*)2

(2)

where γ* is the interfacial tension at optimum formulation, SP* is the solubilization parameter as indicated in Table 1, and K′ is a proportionality constant. According to eq 2, a 4.4-fold increase in solubilization parameter would produce a 19.4-fold reduction in interfacial tension. Therefore, the interfacial tension of the blank formulation would be reduced from 3.5E-3 mN/m to 1.8E-4 mN/m, which is very close to the actual value measured for a 90% neutralized oleic acid (1.0 ( 0.5 E-4 mN/m). The last column in Table 1 shows the salinity at which the optimum formulation occurs. When oleic acid is introduced in the formulation, the salinity to achieve optimum formulation decreases, which is typical of lipophilic linkers (as will be discussed in the next section). As the oleic acid is deprotonated with increasing pH, the optimum salinity increases again, to a value close to 1.7 wt % of NaCl at 80% of neutralization. This demonstrates the changes in the hydrophilic/lipophilic characteristics of the system, as the sodium oleate interacts more strongly at the interface. One last comment on this experiment is that the optimum solubilization parameter of 25 mL of TCE per gram of surfactant is rarely obtained for chlorinated hydrocarbon microemulsions used in surfactant remediation (SEAR). According to Pope et al. (24) the most common range is 5 to 10 mL/g, whereas for enhanced oil recovery (EOR) the value varies from 10 to 60 mL/g. The outstanding value of solubilization capacity for TCE microemulsion with sodium oleate as a cosurfactant should reflect in savings of capital expenditure and remediation time. It should be noted that the low IFT values reported above are below the critical value of 0.5 mN/m where we previously mentioned that vertical migration concerns occur. These systems are reported as a basis for comparing system efficiencies, which is best done at optimal conditions (Bourrel

FIGURE 6. Solubilization and interfacial tension of TCE-AMA microemulsion with and without linkers. and Schechter, ref 1). In an actual remedial design we would back off from optimal system into a supersolubilization system (13), a Winsor Type I system with an IFT value high enough to prevent vertical migration but low enough to maximize the solubilization enhancement. As an example, the solubilization parameter for a TCEAMA at optimum formulation is around 5 mL of TCE/g of AMA, but when both dodecanol and SMDNS are added at 0.09 M each, the solubilization parameter is close to 9 mL TCE/g AMA (3). The interfacial tension at these conditions is less than 0.01 mN/m, which makes the system susceptible to vertical migration. As indicated previously, to avoid vertical migration we should work with Type I microemulsions with interfacial tension of at least 0.5 mN/m. In Figure 6 we present the solubilization of TCE in Type I microemulsions for these two formulations versus the electrolyte concentration normalized to the optimum salinity (S*) as represented by S/S*. For the AMA alone formulation we should operate at an electrolyte concentration of 0.23S* (IFT ) 0.5 mN/m) where we have a TCE solubilization close to 35 000 mg/L. In turn, for the combined linker mixture we can operate at 0.27S* where we solubilize 37 000 mg/L of TCE (IFT ) 0.5 mN/m). At this condition we do not yet harness the solubilization enhancement of linker molecules, but as we approach supersolubilization (S/S* approaches the value of 1) the solubilization in linker systems is substantially improved, as seen in Figure 6. By using the gradient approach of Sabatini et al. (13) we can progressively lower the interfacial tension. In this approach we start at a higher IFT (say 0.5 mN/m), and as residual oil is solubilized we can lower the IFT without mobilization (i.e., we shift where we are in the capillary curve). Thus using the gradient approach we can eventually operate at an IFT of 0.1 mN/m (S/S* of 0.6 for both formulations) without mobilization. Here the solubilization of TCE in the linker formulation is almost three times the solubilization of TCE in the AMA alone formulation. This observation shows that in the supersolubilization region, for a given mobilization risk (given IFT), linker microemulsions show greater solubilization potential, which in itself helps mitigate the mobilization risk. Microemulsion Formulation with Linkers. Because microemulsions are equilibrium phases, the Gibbs phase rule for a nonreactive system can be applied to their formulation:

F)C-π+2

(3)

where F is the degrees of freedom, C is the number of components, and π is the number of phases in equilibrium. For an optimum formulation in a three-phase system containing oil, water, surfactant, electrolyte, and lipophilic linker, C ) 5 and π ) 3, and thus F ) 4. If two of the degrees

FIGURE 7. Salinity at optimum formulation for lipophilic (dodecanol) and hydrophilic (SMDNS) linkers alone and in equimolar mixture for TCE and PCE with 0.103 M AMA. are set (e.g., temperature and pressure) then two degrees of freedom remain: the amount of electrolyte and the concentration of lipophilic linker. For the case of combined linkers, then, the concentration of each linker and electrolyte has to be set to find the optimum formulation. Progress has been made in the thermodynamic modeling of microemulsions (25). For example, Salager et al. have proposed the following semiempirical equation for microemulsion formulation (26, 27):

ln(S*) ) K(ACN) + f(A) - σ + RT∆T

(4)

where S* is the salinity at optimum formulation, K and R are proportionality constants, ACN is the alkane carbon number of the oil phase, f(A) is a function of the alcohol type and concentration, σ is a function of the surfactant/cosurfactant concentration and type, and T is the temperature of the system. Figure 7 shows the optimum salinity for PCE and TCE microemulsions (y-axis) at different concentrations of each linker used alone and in equimolar combination (x-axis). For both oils the optimum salinity increases as the hydrophilic linker SMDNS concentration increases. The opposite is observed for the lipophilic linker dodecanol. The equimolar combination shows a slight increase in salinity for both oils. The changes in salinity in Figure 7 can be described by the following expressions: for dodecanol and TCE

Ln

( )

S* ) -4.2[C12OH] S*0

(5)

for dodecanol and PCE

( )

S* ) -5.9[C12OH) S*0

Ln

(6)

for SMDNS and TCE

( )

S* ) 9.7[SMDNS] S*0

Ln

(7)

for SMDNS and PCE

( )

Ln

S* ) 12.2[SMDNS] S*0

(8)

where S* is the optimum salinity at given SMDNS or VOL. 36, NO. 21, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

4623

TABLE 2. Comparison of the Weighted Contribution Model of Linkers to Mixed Linker Formulation for TCE Microemulsions dodecanol SMDNS optimum salinity optimum salinity concn concn wt % wt % estimated % M M experimental by eq 11 deviation 0.045 0.09 0.18 0.27 0.045 0.09 0.18 0.36 0.18 0.18 0.18 0.18 0.18

0.045 0.09 0.18 0.27 0.027 0.061 0.11 0.195 0.09 0.135 0.18 0.225 0.27

1.44 1.68 2.77 3.7 1.34 1.34 1.34 1.34 0.98 1.5 2.77 4.35 6.82

1.5 1.7 2.2 2.8 1.3 1.4 1.3 1.0 1.1 1.5 2.2 3.2 4.8

6.1 2.9 -20 -23 -1.2 2.3 -5.3 -26 12 2.6 -20 -26 -30

TABLE 3. Comparison of the Weighted Contribution Model of Linkers to Mixed Linker Formulation for PCE Microemulsions dodecanol SMDNS optimum salinity optimum salinity concn concn wt % wt % estimated % M M experimental by eq 11 deviation 0.09 0.09 0.09 0.135 0.135 0.18 0.18 0.18 0.225 0.25

0.0225 0.045 0.09 0.0675 0.135 0.045 0.09 0.18 0.1125 0.25

3.8 4.3 6.3 3.25 6.8 2 2.9 7.25 2.3 7.5

3.5 4.2 6.6 3.9 7.6 2.4 3.6 8.8 3.3 10.1

-9.1 -2.0 5.4 19 12 19 22 22 42 47

dodecanol concentration, and S*0 is the salinity when no linker is used. Equations 5 through 8 have the same form of eq 4. The relative similarity between the PCE and TCE equation for each linker provides evidence that the linker effect is related to molecular interactions of linkers and surfactants. Even more interesting are the results for the combined linkers. Figure 7 shows that an equimolar combination of linkers has an intermediate effect. The question in this case is how does the optimum salinity shift in the case of the combined linkers relative to the shift for each linker individually. The more appropriate model for the combined case is the weighted average model:

Xdodecanol ) XSMDNS )

[dodecanol] [dodecanol] + [SMDNS]

[SMDNS] [dodecanol] + [SMDNS]

(9)

(10)

( )

ln

S* ) XSMDNS × A × [SMDNS] S*0 Xdodecanol × B × [dodecanol] (11)

where XSMDNS is the molar fraction of dodecanol related to total linker added, and Xdodecanol is the molar fraction of dodecanol. The constant A is 9.7 for TCE and 12.2 for PCE, while the constant B is 4.2 for TCE and 5.9 for PCE. It should be noted that eq 11 was developed for constant surfactant concentration (0.103 M AMA) and values will change for

4624

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 21, 2002

different surfactant concentrations and temperatures. Tables 2 and 3 show how this model compares with experimental values for different combinations of dodecanol and SMDNS for TCE and PCE formulations, respectively. Despite the simplicity of the model, it fits the data well, especially for low linker concentrations. Greater deviations are observed at high concentrations, possibly due to partitioning effects of the alcohol and the fact that the SMDNS has a certain limit of salinity increase because of precipitation effects at high electrolyte concentrations. The fact that the overall salinity is a weighted average of the linkers supports the hypothesis that there is a close interaction between hydrophilic and lipophilic linkers.

Literature Cited (1) Bourrel, M.; Schecter R. Microemulsions and Related Systems; Marcel Dekker Inc.: New York, 1988. (2) Chhabra, V.; Free, M. L.; Kang, P. K.; Truesdail, S. E.; Shah, D. O. Tenside, Surfactants, Deterg. 1997, 34 (3), 156. (3) Uchiyama, H.; Acosta, E.; Sabatini, D. A.; Harwell, J. H. Ind. Eng. Chem. Res. 2000, 39 (8), 2704. (4) Baran, J. R.; Pope, G. A.; Wade, W. H.; Weerasooriya V.; Yapa A. Environ. Sci. Technol. 1994, 28, 1361. (5) Baran, J. R.; Pope, G. A.; Wade, W. H.; Weerasooriya, V.; Yapa, A. J. Colloid Interface Sci. 1994, 168, 67. (6) Shiau, B. J.; Sabatini, D. A.; Harwell, J. H.; Vu, D. Q. Ground Water 1994, 32 (4), 561. (7) Diallo, M. S.; Arbiola, L. M.; Weber, W. J. Environ. Sci. Technol. 1994, 28, 1829. (8) Dwarakanath, V.; Kostarelos, K.; Pope, G. A.; Shotts, D.; Wade, W. H. J. Contam. Hydrol. 1999, 38, 465. (9) Krebs-Yuill, B.; Harwell, J. H.; Sabatini, D. A.; Knox, R. C. In Surfactant-Enhanced Subsurface Remediation; ACS Symposium Series 594; Harwell, J. H., Sabatini, D. A., Eds.; American Chemical Society: Washington, DC, 1995; pp 265-278. (10) Londergan, J. T.; Meinardus, H. W.; Mariner, P. E.; Jackson, R. E.; Brown, C. L.; Dwarakanath, V.; Pope, G. A.; Ginn, J. S.; Taffinder, S. Ground Water Monit. Rem. 2001, 21 (3), 71-81. (11) Dwarakanath, V.; Pope, G. A. Environ. Sci. Technol. 2000, 34 (22), 4842-4848. (12) Pennell, K. D.; Pope, G. A.; Abriola, L. M. Environ. Sci. Technol. 1996, 30 (4), 1328-1335. (13) Sabatini, D. A.; Knox, R. C.; Harwell, J. H.; Wu, B. J. Contam. Hydrol. 2000, 45 (1-2), 99-121. (14) Huh, Ch. J. Colloid Interface Sci. 1979, 71, 409. (15) Graciaa, A.; Lachaise, J.; Cucuphat, C., Bourrel, M.; Salager, J. L. Langmuir 1993, 9 (3), 669. (16) Graciaa, A.; Lachaise, J.; Cucuphat, C., Bourrel, M.; Salager, J. L. Langmuir 1993, 9 (12), 3371. (17) Salager, J. L.; Graciaa, A.; Lachaise, J. J. Surfactants Deterg. 1998, 1(3), 403. (18) Acosta, E.; Uchiyama, H.; Sabatini, D. A.; Harwell, J. H. J. Surfactants Deterg. 2002, 5 (2), 151-157. (19) Yaws, C. L.; Hopper, J. R.; Sheth, S. D.; Han, M.; Pike, R. W. Waste Manage. 1998, 17 (8), 541. (20) Wu, B.; Sabatini, D. A. Environ. Sci. Technol. 2000, 34 (22), 4701. (21) Kahlweit, M.; Strey, R.; Busse, G. J Phys. Chem. 1991, 95, 5344. (22) Shiau, B.-J.; Sabatini, D. A.; Harwell, J. H.; Vu, D. Q. Environ. Sci. Technol. 1995, 30 (1), 97-103. (23) Wu, B.; Harwell, J. H.; Sabatini, D. A.; Bailey, J. D. J. Surfactants Deterg. 2000, 3 (4), 465-474. (24) Pope, G. A.; Wade, W. H. In Surfactant-Enhanced Subsurface Remediation; ACS Symposium Series 594; Harwell, J. H., Sabatini, D. A., Eds.; American Chemical Society: Washington, DC, 1995; pp 142-160. (25) Promod K., Mittal, K. L., Eds.; Handbook of Microemulsion Science and Technology; Marcel Dekker: New York, 1999. (26) Salager, J. L.; Morgan, J.; Schechter, R. S.; Wade, W. H.; Vasquez, E. Soc. Petrol. Eng. J. 1979, 19, 107. (27) Salager, J. L.; Morgan, J.; Schechter, R. S.; Wade, W. H. Soc. Petrol. Eng. J. 1979, 19, 271.

Received for review December 29, 2001. Revised manuscript received July 22, 2002. Accepted August 20, 2002. ES0158859