pubs.acs.org/Langmuir © 2011 American Chemical Society
[emim][etSO4] as the Polar Phase in Low-Temperature-Stable Microemulsions Agnes Harrar,† Oliver Zech,†,§ Robert Hartl,† Pierre Bauduin,‡ Thomas Zemb,‡ and Werner Kunz*,† †
Institute of Physical and Theoretical Chemistry, University of Regensburg, 93040 Regensburg, Germany, and ‡ CEA, ICSM UMR 5257, 30207 Bagnols-sur-C eze, France. § Current address: Max-Planck Institute of Colloids and Interfaces, 14424 Potsdam, Germany Received September 17, 2010. Revised Manuscript Received December 15, 2010
We demonstrate here that microemulsions with an IL as the continuous phase can be formed so that they are stable over a wide temperature range and have intermediary properties between flexible and stiff microemulsions. Three components (1-ethyl-3-methylimidazolium ethylsulfate ([emim][etSO4]), limonene, and octylphenol ethoxylate (Triton X 100, abbreviated as TX-100)) were used. This ternary system has been characterized from ambient temperature down to -10 C by means of conductivity, viscosity, and small-angle X-ray scattering (SAXS) measurements. The SAXS data exhibit a characteristic single, broad scattering peak in conjunction with a typical q-4 decay at large q values. The SAXS data have also been interpreted in terms of a dimensionless dilution plot, demonstrating that microstructures are neither isolated droplets nor a random flexible film structure but resemble molten liquid crystals (i.e., they are formed from locally cylindrical or planar structures). This semirigidity is attributed to a good match between the surfactant and the ionic liquid; this holds in a temperature range well below 0 C.
1. Introduction Ionic liquids (ILs) in microemulsions have attracted remarkable attention in the past few years. Nevertheless, the emphasis was rarely on their extraordinary properties, such as their wide liquid range, and the resulting possibilities in forming self-assembled structures under unusual conditions, such as the temperature range. Microemulsions are thermodynamically stable, isotropic mixtures of at least two immiscible solvents, usually water and oil, stabilized by an amphiphilic molecule.1 Structural diversity is known for conventional microemulsions, ranging from oil-in-water droplets (o/w), over bicontinuous structures, to water-in-oil droplets (w/o).2 Scattering, viscosity, and conductivity are very different for these types of microemulsions and depend on the local “microstructure”: locally, any 2D film can be best approximated as a sphere, a cylinder, or a plane, giving rise together with the randomly folded network to four types of bicontinuous microstructures known for microemulsions.3 To our knowledge, only one paper contains a representation of all types of microemulsions at thermal equilibrium in the nondegenerate case where the oil and water volumes differ.4 The substitution of water by polar solvents such as ethylene glycol,5 formamide,6 and ILs7-9 has attracted considerable attention, and it is crucial to determine if microemulsions can be formed to which category of microemulsions they belong. Per definition, ILs consist solely of ions and are liquid at temperatures below 100 C. In this context, the subclass of room-temperature ionic liquids (RTILs) is of particular interest. RTILs have stimulated *Corresponding author. E-mail:
[email protected]. (1) Chevalier, Y.; Zemb, T. Rep. Prog. Phys. 1990, 53, 279–371. (2) Lindner, P., Zemb, Th., Eds. Neutrons, X-rays and Light: Scattering Methods Applied to Soft Condensed Matter; Elsevier: Amsterdam, 2002. (3) Zemb, T. N. Colloids Surf., A 1997, 129-130, 435–454. (4) Arleth, L.; Marcelja, S.; Zemb, T. J. Chem. Phys. 2001, 115, 3923–3936. (5) Friberg, S. E.; Sun, W. M. Colloid Polym. Sci. 1990, 268, 755–759. (6) Schubert, K. V.; Busse, G.; Strey, R.; Kahlweit, M. J. Phys. Chem. 1993, 97, 248–254. (7) Greaves, T. L.; Drummond, C. J. Chem. Soc. Rev. 2008, 37, 1709–1726. (8) Qiu, Z.; Texter, J. Curr. Opin. Colloid Interface Sci. 2008, 13, 252–262. (9) Hao, J.; Zemb, T. Curr. Opin. Colloid Interface Sci. 2007, 12, 129–137.
Langmuir 2011, 27(5), 1635–1642
research because of their unique physicochemical properties, such as negligible vapor pressure, extremely high electrical conductivity, and a wide liquid range. Furthermore, ILs are considered to be designer solvents because their properties can be tuned by modifying their molecular structure. Consequently, ILs can act as good solvents for both inorganic and organic compounds. They can be polar and miscible with water or apolar and immiscible with water and can therefore replace either oil or water in traditional microemulsions. The extraordinary properties of ILs have attracted interest in many disciplines such as electrochemistry,10organic chemistry,11 and analytical chemistry.7,9,12 The possibility of self-assembly in ILs has been widely reported in the last few decades.7-9 The ability of ILs to promote selfassembly is of special importance because it opens a wide field for new formulations and potential applications. Evans et al. reported in the 1980’s the formation of micellar aggregates and liquid-crystalline structures in the protic IL ethylammonium nitrate (EAN).13-15 Our group demonstrated the formation of micelles of the IL-surfactant 1-hexadecyl-3-methylimidazolium chloride, [C16mim][Cl], in EAN.16 Furthermore, we recently showed the formation of EAN-based microemulsions exhibiting extraordinarily high thermal stability ranging from 30 to 150 C.17-19 Moreover, Warr and co-workers (10) Armand, M.; Endres, F.; MacFarlane, D. R.; Ohno, H.; Scrosati, B. Nat. Mater. 2009, 8, 621–629. (11) Welton, T. Chem. Rev. 1999, 99, 2071–2084. (12) Liu, J.-f.; Jiang, G.-b.; J€onsson, J. A˚. Trends Anal. Chem. 2005, 24, 20–27. (13) Evans, D. F.; Chen, S. H.; Schriver, G. W.; Arnett, E. M. J. Am. Chem. Soc. 1981, 103, 2. (14) Evans, D. F.; Yamauchi, A.; Roman, R.; Casassa, E. Z. J. Colloid Interface Sci. 1982, 88, 89–96. (15) Evans, D. F.; Yamauchi, A.; Wei, G. J.; Bloomfield, V. A. J. Phys. Chem. 1983, 87, 3537–3541. (16) Thomaier, S.; Kunz, W. J. Mol. Liq. 2007, 130, 104–107. (17) Zech, O.; Thomaier, S.; Kolodziejski, A.; Touraud, D.; Grillo, I.; Kunz, W. Chem.;Eur. J. 2010, 16, 783–786. (18) Zech, O.; Bauduin, P.; Palatzky, P.; Touraud, D.; Kunz, W. Energy Environ. Sci. 2010, 3, 846–851. (19) Zech, O.; Thomaier, S.; Kolodziejski, A.; Touraud, D.; Grillo, I.; Kunz, W. J. Colloid Interface Sci. 2010, 347, 227–232.
Published on Web 01/12/2011
DOI: 10.1021/la1037316
1635
Article
reported EAN and propylammonium nitrate (PAN) to promote the self-assembly of nonionic surfactants (CnEm) into micelles, lyotropic liquid crystals, and microemulsions.20-22 Until now, all possible structures of self-assembly known for aqueous systems could be identified in ILs as well. Therefore, one would expect that the equation of state relating the free energy expressed as a function of area per molecule and a scalar “packing” parameter would apply.23 Several studies on RTIL-based microemulsions including the amphiphile TX-100 can be found in the literature. The nonionic surfactant Triton X 100 is able to dissolve water and oil,24-26 water and IL,27-30 and oil and IL.31-34 Most work based on this surfactant and ILs in microemulsions deals with RTILs such as 1-butyl-3-methylimidazolium tetrafluoroborate ([bmim][BF4]) to replace water and 1-butyl-3-methylimidazolium hexafluorophosphate ([bmim][PF6]) to substitute oil, respectively. In this context, the research was focused on the characterization of the recorded ternary phase diagram at ambient temperature. When the IL replaces water, the emphasis was mostly on IL-in-oil (IL/o) microemulsions. Conductivity measurements,32 small-angle X-ray scattering (SAXS),35 small-angle neutron scattering (SANS),34 dynamic light scattering (DLS),31,30 freeze-fracture transmission electron microscopy (FF-TEM),33 and UV-vis spectroscopy36,31 measurements have been performed to characterize these microemulsions. Surprisingly, most studies do not benefit from the hightemperature ranges that can be achieved using ILs in microemulsions. As mentioned before, we recently demonstrated nonaqueous microemulsions with an IL as the polar phase that are stable from ambient temperature37 up to 150 C.17,18 Therein, EAN-in-oil structures stabilized by the surfactant [C16mim][Cl] and the cosurfactant 1-decanol have been investigated. Besides our own research, Gao et al. studied temperature-dependent changes on the microstructure in IL-based microemulsions.38 They reported the effect of temperature on [bmim][BF4]-in-cyclohexane and [bmim][BF4]-intoluene reverse microemulsions characterized by DLS, FF-TEM, and 2D rotating frame Overhauser (2D ROESY spectroscopy analysis) experiments. They concluded the existence of an IL/o structure up to 60 C with an increase in an apparent droplet size derived from dynamic properties with increasing temperature. Gao (20) Atkin, R.; Bobillier, S. M. C.; Warr, G. G. J. Phys. Chem. B 2010, 114, 1350–1360. (21) Atkin, R.; Warr, G. G. J. Phys. Chem. B 2007, 111, 9309–9316. (22) Araos, M. U.; Warr, G. G. J. Phys. Chem. B 2005, 109, 14275–14277. (23) Israelachvili, J. N.; Mitchell, D. J.; Ninham, B. W. J. Chem. Soc., Faraday Trans. 1976, 72, 1525–1568. (24) Zhu, D. M.; Wu, X.; Schelly, Z. A. J. Phys. Chem. 1992, 96, 7121–7126. (25) Almgren, M.; Van Stam, J.; Swarup, S.; Loefroth, J. E. Langmuir 1986, 2, 432–438. (26) Anjum, N.; Guedeau-Boudeville, M.-A.; Stubenrauch, C.; Mourchid, A. J. Phys. Chem. B 2009, 113, 239–244. (27) Seth, D.; Chakraborty, A.; Setua, P.; Sarkar, N. Langmuir 2006, 22, 7768– 7775. (28) Seth, D.; Chakraborty, A.; Setua, P.; Sarkar, N. J. Chem. Phys. 2007, 126, 224512–224512. (29) Gao, Y.; Han, S.; Han, B.; Li, G.; Shen, D.; Li, Z.; Du, J.; Hou, W.; Zhang, G. Langmuir 2005, 21, 5681–5684. (30) Safavi, A.; Maleki, N.; Farjami, F. Colloids Surf., A 2010, 355, 61–66. (31) Gao, Y.; Zhang, J.; Xu, H.; Zhao, X.; Zheng, L.; Li, X.; Yu, L. ChemPhysChem 2006, 7, 1554–1561. (32) Gao, Y.; Wang, S.; Zheng, L.; Han, S.; Zhang, X.; Lu, D.; Yu, L.; Ji, Y.; Zhang, G. J. Colloid Interface Sci. 2006, 301, 612–616. (33) Gao, H.; Li, J.; Han, B.; Chen, W.; Zhang, J.; Zhang, R.; Yan, D. Phys. Chem. Chem. Phys. 2004, 6, 2914–2916. (34) Eastoe, J.; Gold, S.; Rogers, S. E.; Paul, A.; Welton, T.; Heenan, R. K.; Grillo, I. J. Am. Chem. Soc. 2005, 127, 7302–7303. (35) Li, J.; Zhang, J.; Gao, H.; Han, B.; Gao, L. Colloid Polym. Sci. 2005, 283, 1371–1375. (36) Li, N.; Gao, Y. a.; Zheng, L.; Zhang, J.; Yu, L.; Li, X. Langmuir 2006, 23, 1091–1097. (37) Zech, O.; Thomaier, S.; Bauduin, P.; R€uck, T.; Touraud, D.; Kunz, W. J. Phys. Chem. B 2008, 113, 465–473. (38) Gao, Y.; Li, N.; Hilfert, L.; Zhang, S.; Zheng, L.; Yu, L. Langmuir 2009, 25, 1360–1365.
1636 DOI: 10.1021/la1037316
Harrar et al.
et al.38 assumed that the change in curvature arises from the temperature dependence of the solubility of the hydrophobic chain in the organic solvent whereas the electrostatic interaction between bmimþ and EO is relatively temperature-independent. Compared to aqueous microemulsions with nonionic surfactants, where temperature sensitivity occurs from the interaction between water and EO groups, the microstructure of microemulsions with [bmim][BF4] was much less susceptible to changes in temperature.38 The Hildebrand solubility parameter provides a measure of the interactions in a material, and it should be stressed in this context that the parameter is much lower for ionic liquids than for water, which is usually used in microemulsions as the polar phase. Although the exact value for [emim][etSO4] is not reported in the literature, values for imidazolium-based ionic liquids could be found that range from 25 to 30 MPa1/2, which are significant lower than for water (48 MPa1/2).39 The present work investigates the formulation of o/IL structures being stable from ambient temperature down to at least -10 C. For this purpose, ingredients with low crystallization temperatures are required. Therefore, the IL [emim][etSO4], which exhibits a glass-transition temperature of -80 C was chosen as the polar phase. Furthermore, limonene with a melting point of -96 C as the apolar phase and the surfactant Triton X 100 as the amphiphile have been used. The following investigation presents a detailed study of the system from 40 to -10 C.
2. Experimental Section (R)-(þ)-Limonene (97%, 98% ee/GLC) was obtained from Sigma-Aldrich, dried over molecular sieves (3 A˚), and redistilled before use. The water content determined by Karl Fischer titration was less than 100 ppm. Triton X 100 was purchased from Fluka and dried under vacuum at 80 C for 4 days, yielding a final water content of less than 20 ppm. [emim][etSO4] was prepared according to a literature procedure by the reaction of equimolar amounts of 1-methylimidazol and diethylsulfate in toluene.40 The obtained IL was washed with toluene and dried under high vacuum at 40 C for 4 days. The water content determined by means of Karl Fischer titration was found to be less than 50 ppm. The “fish” diagram (Figure 2) was established at a constant [emim][etSO4]-to-limonene mass ratio (1:1) with variables of temperature and surfactant content. Triton X 100 was added to a 1:1 (m/m) mixture of the two solvents to obtain surfactant mass fractions within 2-35 wt %. Subsequently, the samples were equilibrated in a thermostatted ((0.2 C) sample holder, and the resulting phases were counted. The equilibration time was set to 12-24 h because samples tended to separate slowly. The ternary phase diagram was recorded at 25 ( 0.2 C according to a procedure reported by Clausse et al.41 Different proportions of surfactant to limonene were prepared in 5 wt % steps between 0 and 100 wt % of the surfactant. The IL was added dropwise under a flow of nitrogen. The weight fractions at which transparency to turbidity occurred were derived from precise weight measurements. Because the system shows very slow kinetics and tends to form stable, clear emulsions that separate after several hours, the time at which a drop was added to the initial solution had to be increased to several hours. The procedure was repeated starting from Triton X 100-[emim][etSO4] mixtures adding limonene and starting from [emim][etSO4]-limonene mixtures adding Triton X 100. This procedure results in a trustworthy phase diagram mirroring the exact position of the singlephase region. (39) Lee, S. H.; Lee, S. B. Chem. Commun. 2005, 3469–3471. J. Chem. (40) Gomez, E.; Gonzalez, B.; Calvar, N.; Tojo, E.; Domı´ nguez, A. Eng. Data 2006, 51, 2096–2102. (41) Clausse, M.; Nicolas-Morgantini, L.; Zradba, A.; Touraud, D. In Microemulsion Systems; Rosano, H. L., Clausse, M., Eds.; Surfactant Science Series; Dekker: New York, 1987; Vol. 24, pp 15-62.
Langmuir 2011, 27(5), 1635–1642
Harrar et al.
Article
All characterization measurements were carried out along an experimental path, where the amount of surfactant was kept constant at 38 wt %. The IL concentration was increased along the experimental path while the limonene concentration was simultaneously decreased. The volume fractions were calculated from the measured densities. The volume fraction of the polar phase (φpol) corresponds to the volume of EO groups (VEO) plus the volume of [emim][etSO4] (VIL) divided by the volume of the microemulsion (V). On the contrary, the volume fraction of [emim][etSO4] (φIL) corresponds to VIL/V. The segregation temperatures were recorded using a precise cryostat ((0.02 C) FP40 from Julabo for temperature control. Microemulsions with varying compositions along the experimental path have been prepared and cooled. Segregation temperatures may be understood as the transition from transparency to turbidity or the freezing of one of the components. The accuracy of the value of the segregation temperature was evaluated to be better than (0.5 C. Conductivities were measured with an in-house-built symmetrical Wheatstone bridge with Wagner earth, a resistance decade, and a sine generator in combination with a precision thermostat ((0.01 C) with immersed conductance cells.42 Kinematic viscosities were measured with a modified automated AVS/G capillary viscometer (Schott Instruments) as described in the literature.43 Prior measurements of dynamic viscosities on a Bohlin Instruments CVO 120 rheometer with shear rates of 10-400 cm-1 can be found in the Supporting Information. Newtonian behavior for all microemulsions along the experimental path at different temperatures was found. Consequently, the use of a capillary viscometer is justified. The viscosity measurements were conducted in two micro-Ubbelohde capillaries (Schott Instruments, types 537 20/II and III), which were placed in a Dewar flask that was connected to a highprecision thermostat42,44 with a circulation pump. The overall temperature uncertainty of this setup was less than 0.003 C. The samples were kept under dry nitrogen throughout the conducted measurements by means of an in-house-built modification of the setup. The capillary constants were confirmed with a certified viscosity standard oil (50 BW, ZMK-Analytik, relative uncertainty 0.32%) at various temperatures, with a flow-time reproducibility better than (0.05%. The viscosities are average values of 3-10 single runs, and the estimated uncertainty of the presented viscosity data was less than 0.5% with respect to the accuracy of the thermostat. To recalculate dynamic viscosities from the obtained kinematic ones, the density was measured at the appropriate temperature using a DMA 5000 M vibrating tube densimeter from Anton Paar. Small- and wide-angle X-ray scattering data were obtained using a XENOCS setup. The X-ray beam originates from a Mo GENIX source. The KR radiation (λ = 0.71 A˚) is selected using a multilayered curved mirror (one reflection) focusing the beam toward infinity. The size of the beam (less than 1 mm) in front of the sample is defined by scatterless slits provided by FORVIS. Sample and empty cell transmissions are determined using an offline pin diode that can be inserted downstream from the sample. A q range of 0.02-2.5 A˚-1 was covered. Quartz capillaries from Hilgenberg (Malsfeld, Germany) were used as sample containers with a wall thickness of 0.01 mm and a total thickness of 2 mm; the usual corrections for background (empty cell and detector noise) subtractions and intensity normalization using Lupolen as a standard were applied.
3. Results and Discussion 3.1. Phase Diagrams, Conductivity, and Viscosity at 25 C. The ternary phase diagram of the system [emim][etSO4], limonene, and Triton X 100 is illustrated in Figure 1. A wide (42) Wachter, R.; Barthel, J. Ber. Bunsen-Ges. Phys. Chem. 1979, 83, 9. (43) Kindler, M. Ph. D. Dissertation, University of Regensburg, 1985. (44) Wachter, R. Professorial Dissertation, University of Regensburg, 1973.
Langmuir 2011, 27(5), 1635–1642
Figure 1. Ternary phase diagram of [emim][etSO4], limonene, and Triton X 100 at 25 C. The arrow marks the experimental path. A cross marks the sample investigated by SAXS.
Figure 2. “Fish” diagram at a constant IL/oil mass ratio of 1:1. (1) Single-phase microemulsion, fish tail; (3) three-phase system, fish head; (2) two-phase system, oil/IL microemulsion and excess oil phase; and (2) IL/oil microemulsions and excess IL phase.
single-phase region could be observed. Hence, a wide range in volume fraction could be investigated along a dilution line with a constant surfactant concentration of 38 wt % TX-100. In addition to the ternary phase diagram, the fish cut is shown in Figure 2. A three-phase body occurs in the temperature range between 28.8 and 31.5 C. The phase-transition temperature at which the surfactant is perfectly balanced between oil and the IL phase with a mean curvature close to zero was found to be at 30.1 C. Below this temperature, the curvature can be considered to be positive and the surfactant film is curved toward oil. Therefore, at low temperatures and low surfactant concentrations an oil/IL microemulsion with an excess oil phase [2] is formed. At high temperatures, the curvature changes from positive to negative versus zero mean curvature and an IL/oil microemulsion with an excess IL phase can be found [2]. The lowest surfactant concentration at which a three-phase region [3] is formed is 11 wt % Triton X 100 whereas the lowest surfactant concentration that forms a singlephase microemulsion [1] is about 24 wt % Triton X 100. Investigations of transport properties of microemulsions, such as electrical conductivity and viscosity, provide important information about their internal dynamics: flexible microemulsions have a monotonic conductivity behavior that can be described by a single power law whereas rigid ones present a maximum in conductivity that is due to the local microstructure as in liquid crystals. Moreover, isolated noncoalescing water droplets are easily identified by means of the very low conductivity values measured. Similar observations also hold for nonaqueous microemulsions with ionic liquids as the polar phase.31,32 DOI: 10.1021/la1037316
1637
Article
Figure 3. Conductivity measured at 25 C along the experimental path marked in Figure 1. A-E mark the detectable subareas.
Harrar et al.
Figure 4. Dynamic viscosities at 25 C as a function of the volume fraction of [emim][etSO4]. Points A-E correspond to the same volume fractions as in Figure 3.
Conductivity measurements (Figure 3) indicate different domains with increasing amounts of [emim][etSO4]. From point A to B, the conductivity remains quite low and increases slightly with increasing amounts of RTIL ((1.3 10-4 to 0.005) S m-1). Although the conductivity is low, its value is still much higher than the conductivity of a neat nonpolar solvent, which is in the range of 10-12 to 10-16 S m-1. The ionic liquid is insoluble in pure limonene and vice versa. For example, the conductivity of IL-saturated limonene remains 8 10-10 S m-1, reflecting the accuracy of the previous statement. An increase in conductivity can hence be attributed only to the formation of some kind of microstructure. Consequently, within points A and B two possible microstructures have to be considered. First, we consider dropletlike IL/o structures (i.e., IL droplets stabilized by surfactant in a continuous oil matrix). With increasing IL content, the size of the droplets increases. The conductance of such IL/o microstructures at low ionic liquid volume fractions can be interpreted in terms of the charge fluctuation model of Eicke et al.45 This model assumes spherical droplets with radius r that move independently of each other. When in thermal equilibrium, these droplets are globally neutral because the number of positively charged emimþ cations is equal to the number of negatively charged etSO4- counterions. However, because of spontaneous fluctuations, charged droplets will form that carry an excess charge z, yielding increased conductivity compared to that of the pure apolar solvent. Notwithstanding, a conductivity on the order of 10-4 S m-1 even at low IL content appears to be extraordinarily high for a simple IL/o droplet structure. The second scenario would be the existence of interconnected structures even at low IL volume fractions. Interconnected structures or droplet clusters allow the effective transport of charge carriers and can explain the relatively high values of the specific conductivity. Although conductivity measurement do not rule out the formation of IL/o droplet structures, a rather interconnected structure is more likely to exist. This is further supported by the SAXS data discussed in section 3.3. As can be seen in Figure 3, the existence of interconnected systems can be assumed between point A and B, and the volume fraction, the degree of interconnection, and hence the conductivity rise continuously up to point C with limonene as the continuous phase. At point C (φ = 0.50), the monotonous increase in conductivity breaks up and the conductivity between points C and D increases only slightly. This can be attributed to the formation of a bicontinuous structure where the transport properties do not change significantly. The following decrease in conductivity between D and E suggests that the IL is now the continuous phase. A decrease in conductivity can be assigned to a decrease in the mobility of charge
carriers. The subtraction of a cosolvent, in the present case, limonene, decreases the mobility of charge carriers because of the increased presence of ion pairs in the solution. Viscosity is a useful tool for confirming the characterization of local microstructures. It has been shown by Cates and Candau46 that locally cylindrical and random flexible films are fluid whereas locally lamellar and other rarer structures such as “molten” cubic phases form extremely viscous microemulsions. Similar to the conductivity measurement, equations for percolation have been suggested but a quantitative verification is possible only if the viscosity η of the polar compound is sufficiently different from η of the oil,47 which can be excluded in the present system. Along the experimental path, all samples exhibit Newtonian behavior within the measured shear rate. The mean value of the dynamic viscosity increases monotonically up to point C with an increasing amount of room-temperature ionic liquid as shown in Figure 4. Above point C (φ = 0.50), the slope increases and indicates a change in microstructure consistent with the conductivity data. 3.2. Lower Segregation Temperatures and TemperatureDependent Viscosity Measurements. The segregation temperatures obtained are shown in Figure 5. Distinct temperature stabilities could be observed for microemulsions containing a high content of IL (points C-E). Moderate temperature stability can be found between points A and B, and a strong temperature
(45) Eicke, H. F.; Borkovec, M.; Das-Gupta, B. J. Phys. Chem. 1989, 93, 314– 317.
(46) Cates, M. E.; Candau, S. J. J. Phys.: Condens. Matter 1990, 2, 6869. (47) Peyrelasse, J.; Moha-Ouchane, M.; Boned, C. Phys. Rev. A 1988, 38, 4155.
1638 DOI: 10.1021/la1037316
Figure 5. Lower phase segregation temperatures for microemulsions having a composition along the experimental path shown in Figure 1. Samples between φ = 0.43 and 0.58 do not show a segregation down to at least -20 C.
Langmuir 2011, 27(5), 1635–1642
Harrar et al.
Article
visualized in Figure 7. IðqÞ ¼
1 þ I0 a 2 þ c1 q 2 þ c2 q 4
ð1Þ
According to eqs 2 and 3, two characteristic length scales were extracted from the Teubner-Strey fit: the correlation length ξ and the domain size d. " # - 1=2 1 a2 1=2 1 c1 þ ξ ¼ 2 c2 4 c2
Figure 6. Dynamic viscosities for microemulsions with 50 wt % (b), 54 wt % (2), and 58 wt % (9) [emim][etSO4] measured in a temperature range from 0 to 40 C. The data points were connected by a solid line to guide the eye.
dependence can be detected for the intermediate region between points B and C. The changes in temperature stability occur at similar points where the conductivity and viscosity change. Nevertheless, conclusions on the kind of microstructure cannot be deduced from segregation temperatures alone. The segregation mechanism results in a separation of the solution in two phases, followed by the freezing of the surfactant.48 By contrast, o/IL microemulsions do not exhibit phase segregation down to -20 C. The kinematic viscosity of microemulsions including 50 wt % [emim][etSO4], 54 wt % [emim][etSO4], and 58 wt % [emim][etSO4] was measured between 0 and 40 C. The same compositions were also investigated with SAXS experiments down to -10 C. The results for viscosity measurements are shown in Figure 6. The dynamic viscosities were calculated from kinematic viscosities and measured densities at the appropriate temperature. At 0 C, the dynamic viscosity reaches a value of more than 1 Pa s, which is of the same order of magnitude as the viscosity of glycerine49 at ambient temperature (1.763 Pa s at 20 C). Nevertheless, magnetic stirring is possible down to -20 C and enables applications of these microemulsions in these temperature ranges. 3.3. Small-Angle X-ray Scattering. SAXS measurements have been performed along the mentioned experimental path at 25 C and additionally for three compositions (58, 54, and 50 wt % [emim][etSO4]) at 40, 30, 15, 10, 5, 0, -5, and -10 C. The SAXS spectra exhibit a single broad correlation peak followed by a q-4 decay at large q values similar to SAXS spectra of aqueous microemulsions. The first step in data evaluation is to extract from the broad bump three parameters, one of them being the characteristic cell distance d. This is also the frequency in the correlation function.50 The three-parameter fit formula (eq 1) from Teubner and Strey (TS) indicates some apparent domain size as well as the ratio between size and persistence length indicative of the stiffness/ flexibility balance. Freiberger et al.51 could show that TS can even be applied in the particulate regime, giving access to interdroplet distances d. The TS-fit and the experimental data for three compositions at temperatures from 40 down to -10 C are (48) Abecassis, B.; Testard, F.; Arleth, L.; Hansen, S.; Grillo, I.; Zemb, T. Langmuir 2007, 23, 9983–9989. (49) Weast, R. C. Handbook of Chemistry and Physics, 55th ed.; CRC Press: Cleveland, OH, 1974. (50) Teubner, M.; Strey, R. J. Chem. Phys. 1987, 87, 3195–3200. (51) Freiberger, N.; Moitzi, C.; de Campo, L.; Glatter, O. J. Colloid Interface Sci. 2007, 312, 59–67.
Langmuir 2011, 27(5), 1635–1642
" # - 1=2 1 a2 1=2 1 c1 d ¼ 2π 2 c2 4 c2
ð2Þ
ð3Þ
Additionally, the amphiphilic factor fa can be calculated from the fitting data using eq 4. fa ¼
c1 ð4a2 c2 Þ1=2
ð4Þ
The experimental data and Teubner-Strey-Fits at 25 C are not plotted here, but the extracted length scales are summarized in Table 1 and the spectra at 25 C are given in the Supporting Information. For IL weight fractions of 5-20 wt %, which represent points in the regime A-B, the domain size increases with increasing RTIL content, consistent with an increase in size for IL/o microstructures. At higher IL weight fractions where limonene can be assumed to be the internal phase (50-58 wt % corresponding to the C-E regime), the domain size decreases. This can result from present o/IL structures whose size decreases with decreasing limonene content. With increasing temperature, the size of the domain also swells but not significantly, as can be seen in Table 2. This is in agreement with the results reported by Warr et al. for EAN-21 or PAN-20based systems, where in general higher surfactant chain lengths were necessary to induce temperature-dependent phase changes or intruding lamellar phases compared to aqueous systems. Additionally, these results conform to the results found by Gao et al.,38 where Triton X 100-based nonaqueous reverse microemulsions also showed a less-pronounced temperature dependence than their aqueous counterparts. The amphiphilicity factor fa obtained from the Teubner-Strey fit is limited to 1, corresponding to the disorder line, and -1, corresponding to the lamellar instability line. Between these limits, structured microemulsions can be formed.52 Well-structured bicontinuous microemulsions can be found according to Teubner and Strey between -0.7 and -0.9.50 The obtained values from our TS fits decrease from -0.61 to -0.82 for 58 wt % [emim][etSO4] with decreasing temperature, from -0.59 to -0.87 for 54 wt % [emim][etSO4], and from -0.71 to -0.87 for 50 wt % [emim][etSO4]. Consequently, we could assume a bicontinuous structure at 50 wt % [emim][etSO4] over the whole investigated temperature range, whereas with increasing IL content only between -10 and 25 C bicontinuous structures are more likely to exist. Furthermore, fa indicates that microemulsions including 20-54 wt % IL preferentially form bicontinuous microstructures. These conclusions drawn from the extracted amphiphilicity factor give the first hints concerning the microstructure at different compositions. However, it is (52) Leitao, H.; da Gama, M. M. T.; Strey, R. J. Chem. Phys. 1998, 108, 4189– 4198.
DOI: 10.1021/la1037316
1639
Article
Harrar et al.
Figure 7. Experimental SAXS data (O) and Teubner-Strey fit (-) for 58 wt % [emim][etSO4] (a), 54 wt % [emim][etSO4] (b), and 50 wt % [emim][etSO4] (c) at different temperatures varying from 40 to -10 C. Table 1. Characteristic Length Scales and Amphiphilic Factor from the Teubner-Strey Model, the DOC Lamellae Model, and the DOC Cylinder Model for [emim][etSO4] Microemulsions at 25 C DOC lamellae wt % IL 58 54 50 40 20 10 5
φpol
Σ (cm /cm )
dTS (A˚)
0.81 0.76 0.71 0.60 0.39 0.30 0.25
4.07 10 3.82 106 3.64 106 3.50 106 3.26 106 3.73 106 4.08 106
59.3 70.2 79.4 94.8 101.7 92.2 84.0
2
3
6
ξ (A˚)
fa
Ψ
22.1 30.9 39.5 51.5 41.0 28.3 20.0
-0.69 -0.77 -0.81 -0.84 -0.73 -0.58 -0.38
0.25 0.30 0.30 0.30 0.35 0.30 0.25
DOC cylinder
t
dDOC (A˚)
v/la
Z
dDOC (A˚)
v/la
37.4 37.1 36.9 33.0 35.5 35.6 34.8
60.0 70.9 75.3 80.4 56.7 49.1 44.8
0.95 0.95 0.95 0.96 1.02 1.02 1.02
3.0 3.0 3.0
83.4 64.0 54.3
1.15 1.24 1.30
Table 2. Values of the Domain Size dTS, Correlation Length ξ, and Amphiphilic Factor fa Derived from Teubner-Strey Fits for 58 wt % [emim][etSO4], 54 wt % [emim][etSO4], and 50 wt % [emim][etSO4] at Various Temperatures T (C)
40
30
25
15
10
5
0
-5
-10
58 wt % IL
dTS (A˚) ξ (A˚) fa Σd
65.8 21.3 -0.61 2.70
62.4 22.8 -0.68 2.53
59.3 22.1 -0.69 2.44
58.6 24.0 -0.74 2.60
57.2 24.6 -0.76 2.63
56.2 25.6 -0.78 2.58
55.2 26.1 -0.79 2.58
54.3 26.5 -0.81 2.55
53.6 27.1 -0.82 2.57
54 wt % IL
dTS (A˚) ξ (A˚) fa Σd
81.5 25.7 -0.59 3.12
76.0 27.9 -0.77 3.01
70.2 30.9 -0.77 2.71
68.9 32.1 -0.79 2.82
67.5 33.5 -0.81 2.80
66.3 34.5 -0.83 2.79
65.1 35.8 -0.85 2.77
64.1 37.0 -0.86 2.74
63.1 37.9 -0.87 2.75
50 wt % IL
dTS (A˚) ξ (A˚) fa Σd
89.9 34.7 -0.71 3.10
83.7 38.1 -0.78 2.96
79.4 39.5 -0.81 2.74
78.1 43.1 -0.85 2.81
77.0 44.1 -0.86 2.80
76.1 44.5 -0.86 2.79
75.4 44.9 -0.87 2.77
75.0 45.1 -0.87 2.74
74.8 44.9 -0.87 2.75
counterintuitive to expect a transition from o/IL to bicontinuous with decreasing temperature. Consequently, for a convenient picture of the microstructure present at a given composition, the dimensionless dilution plot can be used. The dilution plot, which is illustrated in Figure 8, allows one to determine if the microstructure suggested from the values of the amphiphilicity factor is correct. We have just determined that microstructures seem to be “bicontinuous” in an average volume fraction range. To progress further with the present experimental data, one needs to construct a dimensionless dilution plot, Σd versus φpol (shown in Figure 8). Using dilution laws as introduced by de Gennes and Taupin53 can help to distinguish between flexible and stiff microemulsions. Flexible microemulsions will follow the de Gennes and Taupin law,53,54 and other bicontinuous locally cylindrical or locally lamellar (i.e., biliquid) foams exhibit other dilution lines. The extreme case of flexible microemulsions is obtained with SDSpentanol or linear nonionic surfactants,55 and stiff microemulsions (53) Zemb, T. C. R. Chim. 2009, 12, 218–224. (54) Milner, S. T.; Safran, S. A.; Andelman, D.; Cates, M. E.; Roux, D. J. Phys. (Paris) 1988, 49, 1065–1075. (55) Sottmann, T.; Strey, R.; Chen, S. H. J. Chem. Phys. 1997, 106, 6483–6491.
1640 DOI: 10.1021/la1037316
have been identified by Rushforth et al.56 in the water-rich domain and by Barnes et al.57 in the oil-rich domain. Theoretical models can be found in the literature58 and can all be calculated analytically, but the specific area is required to produce dimensionless dilution plots. Therefore, the information contained in scattering has been analyzed by means of the Porod limit, experimental invariant, and finally the specific area Σ (eq 5).37 Σ ¼
lim ðIq4 Þπφpol ð1 - φpol Þ R¥ IðqÞq2 dq
qf¥
ð5Þ
0
R 2 limqf¥(Iq ) denotes the Porod limit, and ¥ 0 I(q)q dq represents the experimental invariant Qexp. The theoretical invariant Qtheo can be estimated via the scattering length density difference ΔF 4
(56) Rushforth, D. S.; Sanchez-Rubio, M.; Santos-Vidals, L. M.; Wormuth, K. R.; Kaler, E. W.; Cuevas, R.; Puig, J. E. J. Phys. Chem. 1986, 90, 6668–6673. (57) Barnes, I. S.; Derian, P.-J.; Hyde, S. T.; Ninham, B. W.; Zemb, T. N. J. Phys. (Paris) 1990, 51, 2605–2628. (58) Zemb, T.; Barnes, I.; Derian, P.; Ninham, B. In Trends in Colloid and Interface Science IV; Zulauf, M., Linder, P., Terech, P., Eds.; Springer: Berlin, 1990; Vol. 81, pp 20-29.
Langmuir 2011, 27(5), 1635–1642
Harrar et al.
Article
Figure 8. Dilution plot: experimental data at 25 C (b) including an assumed error of 15%, CRC model (; 3 3 ;), IL/o or o/IL ( 3 3 3 ), repulsive spheres (; 3 ;), DOC lamellae (---), and DOC cylinders (;).
and the volume fractions through the relation Qtheo = 2π2(ΔF)2 φpol(1 - φpol). In our case, the values of Qexp and Qtheo differ by about 30%. This probably indicates a high solubility of Triton X 100 in one solvent, most likely in the ionic liquid, resulting in an approximated value of ΔF. Anjum et al.26 published a detailed study of the phase behavior of the system water, Triton X 100, and hydrophobic IL [bmim][PF6]. Several fish cuts were recorded at different mass ratios of IL and water. With increasing amounts of IL, the fish head was shifted to higher surfactant concentrations, leading to the conclusion of an extraordinarily high solubility of the surfactant, Triton X 100, in the IL. Considering the present investigation together with studies published by Atkin et al.20,21 and the higher cmc values reported for surfactants in ILs compared to those for the corresponding aqueous systems, a high solubility of nonaggregated surfactant molecules in ILs appears to be plausible. Consequently, the specific area was calculated from the experimental invariant because it mirrors the real situation in the sample and is hence the most reliable value. Several models have been calculated for comparison. The cubic random cell (CRC) model holds for bicontinuous structures and is valid for volume fractions between 0.18 and 0.82 and corresponds to the de Gennes and Taupin law for flexible microstructures.53 The model assumes a set of cubes filled randomly with IL and oil to describe the microstructure. The analytical expression for this model can be written as2 Σd ¼ 6φpol ð1 - φpol Þ
ð6Þ
Actually, the numerical factor has been recalculated to be 5.83 and not 6,53 an insignificant difference that cannot be detected experimentally. The Σd value for w/o or o/w spheres, equivalent to IL/o or o/IL spheres, can be predicted by eq 7.2 Σd ¼ 4:84φ1=3
ð7Þ
models that could better describe the experimental data has also been checked. To describe the experimental values, the adaptability of the model of disordered open connected (DOC)-lamellar microstructure59 has been verified. The DOC model is based on a Vorono_i cell tessellation of space, yielding a complete set of microstructures ranging from isolated spheres via connected cylinders to disordered lamellae. With the fixed input parameters φpol and Σ, the assumption of a surfactant chain length of 8.5 A˚, and a fixed effective surfactant parameter p0, the asymmetry parameter Ψ and the predicted d value dDOC for DOC lamellar structures have been obtained. Herein, p0 is defined as p0 = V/al where V is the volume of the apolar part of the surfactant film, a is the area of the film per surfactant, and l is the film thickness.23 For the connected cylinder microstructure with the input parameters φpol, specific area, and curvature mirrored by the value of the surfactant parameter p0, no physically reasonable values for the connectivity Z could be extracted. In general, Z can be found to be between 0 and 1.2 when the conductivity is low; for 1.2 < Z < 4, the polar network of coalescent droplets is infinite.60 Therefore, the connectivity was fixed at 3.0. For a fixed connectivity, the granularity of the microemulsion is determined. For explicit analytical expressions, the reader is referred to refs 59 and 61. For IL weight fractions of 50-58 wt % that correspond to regime C-E and to volume fractions of 0.71-0.82, the experimental data agree reasonably well with the DOC lamellar model. By contrast, with the DOC-cylinder model it was not possible to describe the experimental points at all. The microstructure of connected cylinders can be observed at high curvatures and high surfactant content, and the disordered lamellar structure is likely if the spontaneous curvature is low compared to d.2,52 By progressing from right to left in the dilution plot, the values on the dimensionless Σd scale do not decrease with increasing oil content and cannot be described by any of the models. Consequently, the existence of microemulsions with a flexible interfacial film can be excluded in the case of 0.3 < φpol < 0.4, corresponding to 10-20 wt % IL. On the contrary, the data suggest that no conversion of the curvature takes place, and this must imply that the bending constant is gkT.3,4 Hence, at mean RTIL content (1040 wt %, φpol = 0.3-0.6) the interfacial film is still curved toward the limonene phase. Comparable observations have already been made for aqueous microemulsions by Barnes et al.,59 where the ternary system DDAB/tetradecane/water was studied and even the curvature at low water content was tilted toward the oil. In our case, a possible explanation is the formation of an asymmetric sponge phase or liquid-foam-like structures that exist down to IL concentrations of 5 wt % consuming a high surface. The results for the temperature-dependent measurements are not shown in the dilution plot for the sake of clarity. However, the values are included in Table 2. The temperature effect on Σd can be summarized as follows: above 25 C, the value increases and fits the model of o/IL microemulsions at 40 C. Below 25 C, the change is not significant and therefore one could conclude that the microstructure does not change with decreasing temperature.
However, for repulsive spheres the volume fraction is linked to the specific area by eq 8.2
4. Conclusions
Σd ¼ 4:32φ2=3
ð8Þ
It could be shown that microemulsions that are stable far below 0 C exist, and evidences could be found that indicate a bicontinuous
For φpol > 0.5, the experimental data are in between repulsive spheres and o/w spheres. By taking the estimated error of 15% in Σd into account, the data could also be interpreted in terms of IL/o or o/IL spheres, respectively. However, the adaptability of other
(59) Barnes, I.; Hyde, S.; Ninham, B.; Derian, P.; Drifford, M.; Warr, G.; Zemb, T. Prog. Colloid Polym. Sci. 1988, 76, 90–95. (60) Testard, F.; Zemb, T. Langmuir 1999, 16, 332–339. (61) Barnes, I. S.; Hyde, S. T.; Ninham, B. W.; Derian, P. J.; Drifford, M.; Zemb, T. N. J. Phys. Chem. 1988, 92, 2286–2293.
Langmuir 2011, 27(5), 1635–1642
DOI: 10.1021/la1037316
1641
Article
structure over a large temperature range. Dilution plots show that the film is not flexible. Isolated droplets or flexible random microemulsions are not formed. None of the available models is coherent with all data shown. The microstructure that best accommodates the data at high IL concentrations is a locally lamellar structure. This can be understood either as a biliquid foam or as an oil-swollen sponge phase, in our case, an asymmetric sponge that has been described by Roux et al.62 and by Wolf et al.63 Since the dilution plot does not follow the universal de Gennes-Taupin plot, the bending constant associated with this surfactant, separating oil and IL, must be gkT. To determine if the bending constant is as high as 10kT, the lamellar phases should be diluted. However, we can safely deduce that the spontaneous packing parameter associated with the phase diagram and the microstructure is close to 1. Although conductivity measurements were not solely sufficient to distinguish between several possible microstructures, scattering experiments illuminated the present microstructure. At low IL content (5-20 wt %), biliquid foamlike (or locally lamellar) structures (i.e., asymmetric sponges) are most likely to exist. At high IL content, above 50 wt % IL (above point D) the data fit reasonably well to the connected locally lamellar structure. Further, these microemulsions offered the best visually observed temperature stability down to -20 C, which means that the (62) Roux, D.; Coulon, C.; Cates, M. E. J. Phys. Chem. 1992, 96, 4174–4187. (63) Wolf, L.; Hoffmann, H.; Talmon, Y.; Teshigawara, T.; Watanabe, K. Soft Matter 2010, 6, 5367–5374.
1642 DOI: 10.1021/la1037316
Harrar et al.
samples do not separate and still from a single-phase microemulsion. The results indicated that the effect of temperature is not very pronounced, regarding both the domain size obtained from TS and the dimensionless dilution plot reflecting the microstructure present. Furthermore, the moderate viscosities at low temperatures enable these systems to be used in applications (e.g. as reaction media) at low temperature. Nevertheless, the formulation of low-temperature-stable systems with even lower viscosities should enable a future task. These microemulsions with a polar continuous phase could be used over a wide temperature range. To the best of our knowledge, this is the first work reporting such temperature stability of o/IL microemulsions. We are convinced that this system represents a model system for the formulation of low-temperaturestable microemulsions that will find plenty of applications such as reaction media for water-sensible reactions. Acknowledgment. We thank Dr. Olivier Diat (ICSM, Marcoule) for help with the SAXS equipment. We thank Dr. Didier Touraud (Institute of Physical and Theoretical Chemistry, University of Regensburg) for fruitful discussions. Supporting Information Available: Shear stress versus shear rate plots for different microemulsion compositions and different temperatures. Calculated volume fractions from weight fractions and densities at 25 C. Experimental SAXS spectra and TS fits at 25 C along the experimental path. This material is available free of charge via the Internet at http://pubs.acs.org.
Langmuir 2011, 27(5), 1635–1642
