Langmuir 2006, 22, 963-968
963
Electron Density Matching as a Guide to Surfactant Design Richard F. Tabor, Sarah Gold, and Julian Eastoe* School of Chemistry, UniVersity of Bristol, Bristol BS8 1TS, U.K. ReceiVed September 5, 2005. In Final Form: NoVember 17, 2005 The effectiveness at reducing interfacial tension between water and different organic solvents was studied, with 14 structurally different dichain sulfosuccinate surfactants. Variations in chemical structure ranged from linear/ branched alkyl tail groups, to phenyl-tipped tail units, to partially and fully fluorinated tails. The solvents n-heptane, toluene, and perfluoroheptane were used as example oil phases. Interfacial activity was measured in terms of a reduced interfacial tension scale, RIFT, based on the value in the presence of surfactants compared to that for the pure solventwater interface. Overall surfactant chain structure was determined to be the key factor affecting RIFT. Furthermore, a strong correlation was observed between RIFT and the electron density Fe of the different surfactants: with any given oil, the most effective surfactants have Fe values closest to that for the solvent. For example, phenyl-tipped surfactants were shown to be comparatively more effective at the interface with an aromatic solvent (toluene) than with an aliphatic n-alkane (heptane). Furthermore, fluorination of the tail groups decreased effectiveness at the hydrocarbon/water interface, which was substantially increased at the fluorocarbon/water interface: this too followed the electron densitymatching pattern. The importance of chain-tip chemical structure was also noted, with regard to the introduction of phenyl, CF3-, and H-CF2- terminal moieties. For branched alkyl-tailed surfactants, it was found that effectiveness could be linked to an empirical “branching factor”. The significance of the electron density matching of organic solvent and surfactant for the prediction of interfacial activities is highlighted, and this concept may prove useful for the future design of new high-efficiency surfactants.
Introduction Owing to their ubiquitous nature in both industry and research, the study of surfactant systems is a vast area of research. Surprisingly, few advances have been made in elucidating a general relationship between surfactant chain structure and its ability to affect interfacial tension in liquid/liquid systems. The impact of chemical structure on the reduction of surface tension at the air/water (a-w) interface is comparatively well researched.1-3 However, owing to the possibility of partitioning effects between bulk phases, studies at the oil/water (o-w) interface are more involved. This work represents a systematic analysis of a range of different model sulfosuccinate surfactants, which are based on the aerosol-OT (AOT) structural motif. As shown in Table 1, three classes of dichain surfactants were examined, from branched and linear alkyl tails, to phenyl-tipped tails, to partially or fully fluorinated tails. The relative interfacial activities of these amphiphiles have been determined at three o-w interfaces, which are n-heptane/water, toluene/water, and perfluoroheptane/water. The significant changes in chemical structure within this broad matrix of systems are expected to drive large differences in intermolecular interactions between surfactant and oil and, hence, feed though to strong interfacial effects. The first attempt to relate chemical structure to interfacial properties was by Traube, who noted the increased efficiency of surfactants with longer hydrophobic chains.4 While understanding in this area has progressed, exploring the effects of surfactant structure on properties such as aggregate morphology,5 monolayer * Corresponding author. (1) Pitt, A. R.; Morley, S. D.; Burbidge, N. J.; Quickenden, E. L. Colloids Surf., A. 1996, 114, 321. (2) Nave, S.; Eastoe, J.; Penfold, J. Langmuir 2000, 16, 8733. (3) Eastoe, J.; Paul, A.; Downer, A.; Steytler, D. C.; Rumsey, E. Langmuir 2002, 18, 3014. (4) Traube, J. Samml. Chem. Vortr. 1899, 4, 255. (5) Nusselder, J. J. H.; Engberts, J. B. F. N. J. Org. Chem. 1991, 56, 5522.
structure,6 and film thickness,7 the effect of branching the hydrophobic moieties on a surfactant’s performance is still poorly understood. The effectiveness of a given surfactant at the interface between water and any given water-immiscible solvent (nonaqueousphase liquid (NAPL)) is dependent on the tendency to adsorb and reduce interfacial intermolecular forces. Previous research indicates that the interfacial effectiveness is related to a number of different factors, including headgroup chemistry,8 counterion (for ionic surfactants)9,10 and tail-group chemistry,1 the number of tail groups,1 and the branching/length of tail groups.11,12 The type and structure of the oil is also significant, with different surfactants behaving more or less effectively at different NAPL/ water interfaces. For example, Atkinson et al. showed that, for AOT4 (I) the greatest efficiency was with decane as the NAPL.13 With regard to tail-group branching, more highly branched molecules show greater effectiveness. Varadaraj et al. showed that the branching of sulfate and ethoxysulfate surfactants derived from Guerbet alcohols resulted in greater effectiveness at the decane/water interface than that for linear analogues.14 The same was shown to be true for branched versus linear nonionic ethoxylate surfactants.11 The explanation given was that oilphase solubility increased with branching, resulting in more (6) Green, S. R.; Su, T. J.; Lu, J. R. J. Phys. Chem. B 2000, 104, 1507. (7) Gaicha, L.; Leblanc, R. M.; Chattopadhyay, A. K. J. Phys. Chem. 1992, 96, 10948. (8) (a) Nave, S.; Eastoe, J.; Heenan, R. K.; Steytler, D. C.; Grillo, I. Langmuir 2002, 18, 1505. (b) Nave, S.; Paul, A.; Eastoe, J.; Pitt, A. R.; Heenan, R. K. Langmuir 2005, 21, 10021 . (9) Eastoe, J.; Fragneto, G.; Robinson, B. H.; Towey, T. F.; Heenan, R. K.; Leng, F. J. J. Chem. Soc., Faraday Trans. 1992, 88 (3), 461. (10) Nagasoe, Y.; Ichiyanagi, N.; Okabayashi, H.; Nave, S.; Eastoe, J.; O’Connor, C. J. Phys. Chem. Chem. Phys. 1999, 1, 4395. (11) Varadaraj, R.; Bock, J.; Valint, P.; Zushma, S.; Thomas, R. J. Phys. Chem. 1991, 95, 1671. (12) Rosen, M. J.; Zhu, Z. H.; Gu, B.; Murphy, D. S. Langmuir 1988, 4, 1273. (13) Atkinson, P. J.; Robinson, B. H.; Howe, A. M.; Pitt, A. R. Colloids Surf., A. 1995, 94, 231. (14) Varadaraj, R.; Bock, J.; Geissler, P.; Zushma, S.; Brons, N.; Colletti, T. J. Colloid Interface Sci. 1991, 147, 396.
10.1021/la052418m CCC: $33.50 © 2006 American Chemical Society Published on Web 12/27/2005
964 Langmuir, Vol. 22, No. 3, 2006 Table 1. Systematic Names and Schematic Molecular Structures of the Surfactants Studied
Tabor et al.
the fluid of most interest in this respect is condensed/supercritical CO2. A fundamental understanding of interfacial tensions and surfactant adsorption is essential to predict microemulsion stability and to design suitable CO2 surfactants. However, the experimental high-pressure interfacial tension experiments needed to investigate activity at the CO2/water interface are notoriously difficult, and this has certainly limited progress in this area. Hence, it would be helpful to find an alternative method to screen surfactants for CO2 activity by employing CO2 analogues that are liquid at room temperature and standard pressure. The “electron density matching” approach developed here makes a new contribution to this field. Because of their excellent solubilization in CO2, partially and fully fluorinated surfactants have become important amphiphiles for the stabilization of water-in-supercritical-CO2 microemulsions.17-19 As such, their properties have been well researched at the a-w and water/CO2 interfaces, including the effect of hydrophobic tail constitution and branching.3,20,21 However, characterization of their effectiveness at the o-w interface is still lacking, and two such surfactants are included in this study. Molecular modeling simulations, such as that developed by Blankschtein et al. for predicting the behavior of surfactant solutions at the o-w interface,22 have been used. Jang et al. simulated a range of isomeric alkyl benzene sulfonate surfactants with varying degrees of branching.23 They noted greatest effectiveness for the isomer with the effective hydrophobic tail length most like that of the oil phase (decane) because of its increased miscibility therein. Rekvig et al. showed that tailgroup branching can have either a positive or negative effect on surfactant effectiveness, depending upon the ordering of headgroups perpendicularly to the interface.24 A key finding in this work is that electron density can be used as a very simple useful guide to understand why certain surfactants are more effective than others at liquid-water interfaces. This represents a new approach, potentially offering valuable insight into the general molecular requirements for the rational design of superefficient surfactants for advanced applications, such as the stabilization of interfaces and dispersions in CO2. Experimental Section
efficient adsorption. Further work by Wormuth and Zushma on the phase behavior of similar branched and linear surfactants in o-w systems confirmed the lipophilic order as highly branched > methyl branched > linear.15 Rosen et al. previously noted that, with increased oil-phase solubility, o-w interfacial properties became more different from those found at the a-w interface.12 Therefore, surfactants with tail groups that exhibit low lipophilicity will exhibit similar o-w interfacial and a-w surface tension properties, whereas more lipophilic moieties (such as those with greater branching) will exhibit less predictable behavior. The work of Aspe´e and Lissi on branched versus linear alkanols showed that branched analogues were less effective because of lower partitioning into the organic phase.16 To date, research on phenyl- or otherwise aromatic-tipped surfactant systems has been limited.1,8b This paper includes two such unusual Ph-tipped surfactants to explore how they behave at the interface with model aromatic and aliphatic solvents. One aim of the work was to explore a new predictive approach to the design of surfactants for a general NAPL/water interface; (15) Wormuth, K. R.; Zushma, S. Langmuir 1991, 7, 2048. (16) Aspe´e, A.; Lissi, E. J. Colloid Interface Sci. 1996, 178, 298.
A. Materials. All of the surfactants used in this study were synthesized, characterized, and purified as described elsewhere.2,3,8,25-28 The n-heptane and toluene (Fisher, analytical grade 99%) used for interfacial tension measurements were purified before use by distillation and repeated flash column chromatography (Analar chromatographic silica). The surface chemical purity of the oil was determined after each pass over the column by measuring dynamic interfacial tensions (up to 2 h) against pure water by drop-shape analysis (see below). A limiting (high) interfacial tension was achieved after nine column cycles, and subsequent purification runs (17) Eastoe, J.; Gold, S. Phys. Chem. Chem. Phys. 2005, 7, 1352. (18) Harrison, K.; Goveas, J.; Johnston, K. P.; O’Rear, E. A. Langmuir 1994, 10, 3536. (19) Hoefling, T. A.; Enick, R. M.; Beckman, E. J. J. Phys. Chem. 1991, 95, 7127. (20) Sagisaka, M.; Yoda, S.; Takebayashi, Y.; Otake, K.; Kondo, Y.; Yoshino, N.; Sakai, H.; Abe, M. Langmuir 2003, 19, 8161. (21) Wang, S. W.; Marchant, R. E. Macromolecules 2004, 37, 3353. (22) Mulqueen, M.; Blankschtein, D. Langmuir 2002, 18, 365. (23) Jang, S. S.; Lin, S.-T.; Maiti, P. K.; Blanco, M.; Goddard, W. A. J. Phys. Chem. 2004, 108, 12130. (24) Rekvig, L.; Kranenburg, M.; Hafskjold, B.; Smit, B. Europhys. Lett 2003, 63, 902. (25) Downer, A.; Eastoe, J.; Pitt, A. R.; Simister, E. A.; Penfold, J. Langmuir 1999, 15, 7591. (26) Eastoe, J.; Nave, S.; Downer, A.; Paul, A.; Rankin, A.; Tribe, K.; Penfold, J. Langmuir 2000, 16, 4511. (27) Nave, S. Ph.D. Thesis, University of Bristol, Bristol, U.K., 2000. (28) Downer, A. Ph.D. Thesis, University of Bristol, Bristol, U.K., 2000.
Surfactant Design Via Electron Density Matching
Langmuir, Vol. 22, No. 3, 2006 965
did not improve the interfacial tensions. The final interfacial tensions were consistent with those reported in the literature and with those of oils of high surface chemical purity. Because of the high cost of perfluoroheptane (mixed isomers - Fluorochem, 98%), distillation was impractical. However, it was found that column chromatography was sufficient to produce a time-stable interfacial tension against pure water. Deionized water for all of the experiments was obtained from an Elga Elgastat Maxima water purification system, with a product output resistivity of 18.2 mΩ. Ethylenediaminetetraacetic acid tetrasodium hydrate (EDTA; 99.5%) was from Sigma. B. Tensiometric Techniques. Interfacial tensions were measured by one of two methods, depending on the magnitude.29-35 For interfacial tensions above ∼5 mN m-1, drop-shape analysis was employed (Kru¨ss DSA10-Mk2) in pendant drop mode. For interfacial tensions below ∼5 mN m-1, a spinning-drop tensiometer (Kru¨ss Site04) was used. This instrument was calibrated against the standard interfacial tension of n-butanol-water (found: 1.76 ( 0.04 mN m-1; literature: 1.8 mN m-1 32 and 1.67 mN m-1 31). All spinningdrop experiments were carried out at 25 ( 0.5 °C and at a capillary rotational frequency of 4150 rpm. This value was chosen to be high enough to satisfy the requirement for gyrostatic equilibrium and was maintained constant to eliminate any possible dependence of interfacial tension on rotational speed, as noted by Capelle and Isaacs.32,33 Under these conditions, no time dependence of the tensions was noted over a measurement period of ∼30 min. The capillary was cleaned with Decon 90 laboratory detergent and then flushed with copious amounts of deionized water. The densities of the o-w phases were measured with a Paar Scientific DMA 35 density meter. Interfacial tension values were calculated with the Vonnegut equation,34 γ)
∆Fω2R3 4
(1)
The parameters in the above expression are interfacial tension, γ, density difference between the oil and aqueous phases, ∆F, rotational angular velocity of the capillary, ω, and the radius of the drop, R. The drop length was always at least four times greater than the diameter.35 C. Critical Micelle Concentration (CMC) Determination by Conductivity. For all surfactants except DiC9SS (IV) and AOT7 (V), literature values of the CMC were used.2,26,27 The CMCs of DiC9SS and AOT7 were determined to be 0.212 ( 0.003 mM and 0.058 ( 0.003 mM, respectively, by conductivity measurements (Jenway electrochemistry analyzer). D. Electron Density Calculations. Molecular Connelly solventexcluded volumes were calculated using CambridgeSoft ChemBats3D Pro (version 7.0.0). Electron density Fe was then obtained by dividing the number of electrons in the molecule by molecular volume.
Results and Discussion For each surfactant and oil combination, tensiometric measurements were made at 4, 6, and 8 times the aqueous-phase CMC. Rosen suggested that any post-CMC effect on γ should be insignificant;36 however, a small effect was seen here. Figure 1 shows the examples of the post-CMC decrease in interfacial tension for AOT1 (VI) and AOT4 (I), on a reduced concentration/ CMC scale. Hence, to make meaningful comparisons between the systems, it was necessary to work at constant multiples of the CMC rather than in absolute concentrations/activities. The (29) Hou, M. J.; Kim, M.; Shah, D. O. J. Colloid Interface Sci. 1988, 123, 398. (30) Donahue, J.; Bartell, F. E. J. Phys. Chem. 1952, 56, 480. (31) Villers, D.; Platten, J. K. J. Phys. Chem. 1988, 92, 4023. (32) Capelle, A. Surface Phenomena in Enhanced Oil RecoVery; Plenum: New York, 1981; pp 229-236. (33) Isaacs, E.; Maunder, J. D.; Li, J. ACS Symposium Series; American Chemical Society: Washington, DC, 1988; Vol. 396, pp 324-344. (34) Vonnegut, B. ReV. Sci. Instrum. 1942, 13, 6. (35) Manning, C. D.; Scriven, L. E. ReV. Sci. Instrum. 1977, 48, 1699. (36) Rosen, M. J. Surfactants and Interfacial Phenomena, 2nd ed.; Wiley: New York, 1989; pp 69-90.
Figure 1. Examples of the concentration dependence of heptanewater interfacial tension in the post-CMC region.
concentration region chosen was a compromise of between being sufficiently above the CMC to minimize concentration effects, but still within the solubility range at the operating temperature. The mean average of three interfacial tension vales for each system at 4, 6, and 8 times the CMC was taken to eliminate this weak effect. Polyvalent Mn+ cationic impurities (such as Ca2+ and Mg2+) are known to significantly affect equilibrium surface tensions at the a-w interface.37-39 To overcome this effect, a screening protocol employing EDTA has been developed.26 This trace additive acts as a chelating sequestrant, removing the polyvalent ions from solution and replenishing the levels of Na+. To assess the importance of this effect here at the o-w interface, spinningdrop measurements were made for AOT1, with and without EDTA (AOT1/EDTA molar ratio of 100:1). The values agreed to within 98%: this 2% error is consistent with the experimental uncertainties. Hence, it appears that EDTA is unnecessary in these o-w systems, at least at high multiples of the CMC. To make a comparison between the different systems, a reduced interfacial tension, RIFT, was introduced:
RIFT ) 1 -
(
)
γo/w - γo/w/s γo/w
(2)
where γo/w is the interfacial tension of the pure o-w system, and γo/w/s is that of the oil/aqueous surfactant solution. The RIFT values are displayed in Figure 2 for 13 different surfactants at the n-heptane/water interface (for the highly fluorinated DiCF4, the RIFT value was 0.32). The effects of (a) altering the main tailgroup structure, (b) alkyl branching of the tail group, and (c) changing the oil types are discussed separately below. A. Effect of Changing Overall Tail-Group Structure. Moving from alkyl-, to phenyl-, to fluorinated-chain surfactants, there is a marked change in RIFT at the n-heptane/water interface, in the order of branched alkyl < linear alkyl < phenyl-tipped < partially fluorinated < fully fluorinated. While chain branching for the alkyl surfactants (discussed separately in the next section) subtly changes the limiting interfacial tension achieved and indeed (37) Li, Z. X.; Lu, J. R.; Thomas, R. K. Langmuir 1997, 13, 3681. (38) Downer, A. D.; Eastoe, J.; Pitt, A. R.; Heenan, R. K.; Penfold, J. Colloids Surf., A. 1999, 156, 33. (39) An, S. W.; Lu, J. R.; Thomas, R. K.; Penfold, J. Langmuir 1996, 12, 2446.
966 Langmuir, Vol. 22, No. 3, 2006
Figure 2. Reduced interfacial tensions at the heptane/aqueous surfactant solution interface. The value for DiCF4 (RIFT ) 0.32) has been omitted for scaling clarity.
Figure 3. Relationship of the reduced interfacial tension at the n-heptane/water interface and the electron density Fe of surfactants. For clarity, only surfactants deviating from the linear fit are labeled. The Fe value calculated for heptane is marked.
produces the most effective surfactants, the overall chemical structure of the tail group has an overriding effect. As previous research has suggested,11,12,14-16 when all other factors are equal (headgroup chemistry and number of tail groups), the compatibility of the hydrophobic portion with the oil phase dominates the behavior. Apart from the obvious gross differences, which arise owing to the incompatibility of H-carbon and F-carbon chains, the trend in RIFT with surfactant type can interestingly be correlated with electron density Fe, which is shown in Figure 3. When compared to the Fe value of oil-heptane (0.45 Å-3), it would seem that, as a general rule, the closer the electron density of the surfactant is to that of the oil, the greater the effectiveness at the interface. There are some rogue points, which are discussed below, but the correlation is surprisingly good for eight or nine different surfactants. This remarkably simple approach appears to provide a good guide to the effectiveness of this important class of surfactants at the n-heptane/water interface. The only major exception to this electron density matching rule seems to be the case of DiHCF4 (II), which is identical to
Tabor et al.
Figure 4. Relationship of reduced interfacial tension and electron density for surfactants at the water/perfluoroheptane interface. The Fe values for perfluoroheptane and CO2 are marked for reference.
DiCF4 (VII) apart from the chain tips, both of which have a polar H-CF2- terminus rather than CF3-. It seems that this minor difference is enough to induce sufficient “hydrocarbonlike” character, emphasizing the importance of chain-tip chemistry on surfactant performance. It seems likely that this capping hydrogen induces a weak dipole,25 which serves to boost the compatibility with hydrocarbon solvents, compared to that of the fully fluorinated DiCF4 analogue. Such differences between the properties of H-CF2- and CF3-tipped surfactants in aqueous systems have been highlighted previously.25 Slight deviations from the linear fit are also noted for AOT6 (VIII) and brDiPhC3SS (IX), both of which appear to be less effective than suggested by the linear correlation. The former is discussed separately in the next section. In the case of brDiPhC3SS, a plausible explanation is that the two-carbonsaturated chain is so short that compatibility with n-heptane is diminished. This would also explain why it is less effective at the heptane/water interface than the isoelectronic, unbranched surfactant DiPhC3SS (X), since, for this one, the longer alkyl chain length could enhance compatibility with the heptane solvent. Figure 4 represents a test of this electron density matching approach, but for the perfluoroheptane/water interface, once again the correlation seems to hold. The electron-rich perfluoroheptane has Fe ∼ 1.03 Å-3 compared to 0.45 Å-3 for n-heptane. As would be expected, the fluorinated surfactants are more effective, as their electron densities are closer to that of the F-carbon oil. However, DiHCF4 is still an outlier, again suggesting that chaincapping hydrogens act through the chain-tip effect. As a further test, it is known that these F-AOT analogues are highly effective stabilizers of water-in-CO2 microemulsions.17 It is also interesting to note that Fe values for fluorinated surfactants are extremely close to that of CO2, perhaps indicating why these surfactants are so active in this medium. The point marked in Figure 4 for liquid CO2 (mass density 1 g cm-3) is consistent with the high efficiency seen for DiCF4 in this application.17 The positioning of DiHCF4 is also consistent with the drop off in surface activity in CO2 compared to that of the fully fluorinated DiCF4.17 B. Chain Branching Effects with Alkyl-Tailed Surfactants. It is clear upon moving from linear DiCnSS to branched AOT variants that there is a marked increase in effectiveness at the n-heptane/water interface. The only exception to this is branched AOT6, which fails to behave more effectively than a linear system
Surfactant Design Via Electron Density Matching
of the same carbon number. This surfactant has previously been noted to have unusual properties when compared with other AOT analogues, showing limited phase behavior (uncharacteristic of the AOT series) in the microemulsion phase work of Nave et al.40 The explanation was that, because of the “insufficient branching” of the tail groups, the surfactant exhibits behavior equivalent to that of a linear variant. This is an appealing explanation, as its limiting interfacial tension is very similar to that of the linear-tailed DiC7SS (XI). However, if this were the case, then AOT7 (V, which is homologous to AOT6 but with two additional -CH2- units before the chain branching) would also be of comparable low efficiency. However, AOT7 proved to be substantially more effective than AOT6 and all of the DiCnSS series. Another possible explanation for the unusually low effectiveness of AOT6 is that the methyl branch on C1 adjacent to the headgroup reduces the efficiency of molecular packing. Rekvig et al. suggest that added branches close to the headgroup cause a slight increase in lipophilicity and (perhaps) a reduction in interfacial packing, thereby decreasing the effectiveness.24 It was noted that this effect would be particularly significant for doubletailed surfactants.24 To test this idea, an AOT surfactant with branching at the C1 position was investigated (sodium bis(1ethyl-2-methyl-1-pentyl) sulfosuccinate (AOT5)). Unfortunately, even at 4 times the CMC, this AOT5 did not form a single homogeneous aqueous phase, and so it could not be included in this study. To quantify the effect of chain branching in the AOT series of surfactants on their performance at the a-w interface, Nave et al. devised an empirical “branching factor” to describe the extent of branching.2 This factor was found to be of use in understanding aqueous properties, but was less appropriate to describe behavior at the o-w interface. Here, a similar branching index is introduced, which encompasses the effects of both length and branching of the chains. To account for branches, the branch length is multiplied by the carbon position of the attachment to the main chain. For example, in sodium bis(2-ethyl-1-hexyl) sulfosuccinate (AOT1), the ethyl branch adds a value of 4. Chain tips are counted as methyl branches on the penultimate carbon (adding a value of 5 in the case of AOT1). The contributions from each branch are summed and divided by the carbon number of the longest chain. This would be six carbons in the case of AOT1, giving an overall branching factor of 1.5. In this way, a branching index was evaluated for the other alkyl surfactants; the correlation with reduced interfacial tension is displayed in Figure 5. Figure 5 shows that increasing the branching index within the AOT series directly improves effectiveness at the o-w interface. The only major deviation from the fit is AOT6, which is unusual in the AOT series as described above. The reason for the enhancement with branching may be due to a greater degree of conformational disorder. Figure 6 suggests the “lipophilic ranking” in the sequence linear alkyl < isopropyl-tipped < tertbutyl-tipped, also agreeing with the work of Wormuth and Zushma.15 The other inference that may be drawn from Figure 6 is that the chemical structure of the tips of the hydrophobic chains is more significant than branching further down the alkyl chain (as the branching factor calculation is “weighted” toward terminal branching). The importance of chain-tip chemistry for aqueousphase surfactant properties was previously emphasized by Pitt et al.15 (40) Nave, S.; Eastoe, J.; Heenan, R. K.; Steytler, D. C.; Grillo, I. Langmuir 2000, 16, 8741.
Langmuir, Vol. 22, No. 3, 2006 967
Figure 5. Dependence of reduced interfacial tension on alkyl-chain branching.
Figure 6. Relative interfacial tensions of selected surfactants in heptane and perfluoroheptane.
C. Effect of Changing the Solvent. It has previously been stated that different surfactants will behave similarly at different o-w interfaces.13 Interestingly, the reduced interfacial tension scale used in this work allows direct comparison of different NAPL/water interfaces. Changing from n-heptane to toluene resulted in similar behavior, in that the surfactants behaved similarly in terms of their relative effectiveness in either solvent. The exceptions to this pattern were the two phenyl-tipped variants (IX and X). While, for the other surfactants, the overall reduced interfacial tensions were around 1.5-2 times higher in toluene compared to heptane, the phenyl-tipped analogues proved to be 4.3 and 5.8 times more effective in toluene for brDiPhC3SS and DiPhC3SS, respectively. Because of the chemical similarity of the phenyl tip to toluene, it would be expected that these surfactants are more compatible with toluene rather than heptane; the argument of electron density matching previously stated in section A supports this idea. Because toluene has a greater electron density (0.57 Å-3) than that of heptane (0.45 Å-3), it would be expected that phenyl-tipped variants (Fe ) 0.57 and 0.55 Å-3 for IX and X, respectively) would show significantly greater effectiveness in toluene than in heptane, and this is indeed the case. Similar observations have been made regarding the stability
968 Langmuir, Vol. 22, No. 3, 2006
of water-in-oil microemulsions by these Ph-tipped surfactants in either toluene (stable) and n-heptane (unstable).8b When perfluoroheptane was used as oil, the trend was a mirror image of that seen in H-heptane, as shown clearly by Figure 6. In F-heptane, the fully fluorinated surfactant DiCF4 is the most effective, and the H-alkyl-tailed surfactants AOT1 and DiC6SS are now least effective. The reasons for the reversed effectiveness of the fluorinated analogues upon switching H- for F-solvent may be explained by the electron density pattern outlined in section A. Notably, in F-heptane, the H-capped DiHCF4 showed intermediate behavior, as might have been expected on the basis of previous observations.
Conclusions For the important class of double-chain anionic sulfosuccinate surfactants studied here, a new relationship is revealed between surfactant structure and effectiveness at reducing interfacial tension between water and an immiscible organic solvent (NAPL). It has been shown that a limiting interfacial tension is reached, which is dependent upon overall tail-group structure, tail-group branching, and the type of NAPL oil used, all of which affect the surfactant-NAPL compatibility. The former is the most important factor, while chain branching provides more subtle effects on effectiveness. The result of chain branching can be quantified using an empirical “branching index”, showing that, in hydrocarbon solvent, greater methylation results in more effective surfactants. For n-heptane/water systems, the lowest interfacial tension was generated by a tert-butyl-tipped surfactant, AOT4 (I). This also emphasizes the importance of chemical structure at the extreme tips of the hydrophobic chains, compared to chain branching occurring elsewhere in the molecule. This is
Tabor et al.
demonstrated by the addition of phenyl-chain tips, which results in dramatic shifts with respect to the type of solvent when comparing aliphatic (heptane) and aromatic (toluene) classes. Furthermore, partial and full fluorination of the tail-group moieties result in surfactants that are more effective at the interface of water with a fluorocarbon, rather than a hydrocarbon, solvent. Most significantly, the effectiveness for these AOT-related surfactants for reducing interfacial tension can be neatly correlated with electron density: the most effective surfactant/solvent combinations have near-matched electron densities. This represents an extension of the Hildebrandt solubility parameter and the “like dissolves like” maxim. It is also a significant observation that could lead to new approaches in the design of novel surfactants, since, for low volatility compounds such as surfactants, electron densities are more readily estimated than solubility parameters. This work has also shown that fluorosurfactants active in carbon dioxide have electron densities similar to that of CO2, pointing to a new approach for the rational design of CO2-philes. This principle could also be applied to other novel solvent systems that also require custom surfactants, such as ionic liquids, and partially fluorinated solvents such as those used in metered-dose-inhaler drug delivery packages. Acknowledgment. R.T. would like to acknowledge the University of Bristol for supporting this work, as well as current and previous group members who provided many of the surfactants used in this study. S.G. thanks the University of Bristol for a studentship under the DTA scheme. Alan Pitt (Kodak, U.K.) is thanked for stimulating discussions on surfactants. LA052418M
