9168
Langmuir 2000, 16, 9168-9176
Phase Behavior, Structure, and Applications of Reverse Microemulsions Stabilized by Nonionic Surfactants Andrey J. Zarur, Neville Z. Mehenti, Anne T. Heibel, and Jackie Y. Ying* Department of Chemical Engineering, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139-4307 Received November 15, 1999. In Final Form: July 7, 2000 Reverse microemulsion-mediated synthesis of various inorganic and organic nanoparticles is of interest for a variety of applications. Mixtures of commercially available polyethoxylated alcohols and linear alcohols can be employed to stabilize reverse microemulsions suitable for such use. The phase behavior, electrical conductivity, and stability of the water/isooctane and water/cyclohexane microemulsion systems stabilized by a mixture of Neodol 91-6 nonionic surfactants and 1-pentanol are presented in this paper. Quasi-elastic light scattering and small-angle neutron scattering were employed to characterize the aggregation state and morphology of the aqueous and oil phases in these reverse microemulsions. Barium hexaaluminate, a complex oxide of interest for catalytic applications, was successfully synthesized via sol-gel processing in the resulting reverse microemulsion media. The recovered inorganic nanoparticles exhibited morphologies characteristic of the aqueous phase in the reverse microemulsion systems.
Introduction Microemulsions are thermodynamically stable systems consisting of a hydrophilic phase and a hydrophobic phase, stabilized with the use of surfactants. Regular microemulsions consist of nanometer-sized hydrocarbon domains (termed micelles) surrounded by amphiphilic molecules, stabilized in a continuous aqueous phase. In contrast, reverse microemulsions consist of aqueous domains (termed reverse micelles) dispersed in a continuous oil phase. Reverse microemulsions are of special interest because a variety of reactants can be introduced into the nanometer-sized aqueous domains for reaction confined within the reverse micelles, leading to materials with controlled size and shape.1,2 In the past few years, significant research has been conducted in the reverse microemulsion-mediated synthesis of organic systems (e.g., polymeric nanoparticles3-7) and inorganic systems (e.g. quantum dots,8-13 metallic nanoparticles,14-17 and ultrafine ceramic particles15,18-25). * To whom correspondence should be addressed. (1) Pileni, M. P. J. Phys. Chem. 1993, 97, 6961. (2) Lopez-Quintanela, M. A.; Rivas, J. J. Colloid Interface Sci. 1993, 158, 446. (3) Candau, F.; Zekhini, Z.; Durand, J. P. J. Colloid Interface Sci. 1986, 114, 398. (4) Candau, F.; Leong, Y. Y.; Fitch, R. M. J. Polym. Sci. A: Polym. Chem. 1985, 23, 193. (5) Voortmans, G.; Jackers, C.; Deschryver, F. Br. Polym. J. 1989, 21, 161. (6) Atkinson, P. J.; Grimson, M. J.; Heenan, R. K.; Howe, A. M.; Robinson, B. H. J. Chem. Soc., Chem. Commun. 1989, 23, 1807. (7) Luthi, P.; Luisi, P. L. J. Am. Chem. Soc. 1984, 106, 7285. (8) Petit, C.; Pileni, M. P. J. Phys. Chem. 1988, 92, 2282. (9) Petit, C.; Lixon, P.; Pileni, M. P. J. Phys. Chem. 1990, 94, 1598. (10) Motte, L.; Petit, C.; Boulanger, L.; Lixon, P.; Pileni, M. P. Langmuir 1992, 8, 1049. (11) Llanos, P.; Thomas, J. K. Chem. Phys. Lett. 1986, 125, 299. (12) Steigerwald, M. L.; Alivisatos, A. P.; Gibson, J. M.; Harris, T. D.; Kortan, R.; Muller, A. J. J. Am. Chem. Soc. 1988, 110, 3046. (13) Cizeron, J.; Pileni, M. P. J. Phys. Chem. 1995, 99, 17410. (14) Fandler, J. H. Chem. Rev. 1987, 87, 877. (15) Kisida, M.; Fujita, T.; Umakoshi, K.; Ishiyama, J.; Nagata, H.; Wakabayashi, K. J. Chem. Soc., Chem. Commun. 1995, 7, 763. (16) Kurihara, K.; Kizling, J.; Stenius, P.; Fendler, J. H. J. Am. Chem. Soc. 1983, 105, 2574. (17) Petit, C.; Lixon, P.; Pileni, M. P. J. Phys. Chem. 1993, 97, 12974. (18) Chhabra, V.; Lal, M.; Maitra, A. N.; Ayyub, P. Colloid Polym. Sci. 1995, 273, 939.
Synthesis of inorganic nanoparticles within reverse microemulsions has the potential to yield nonagglomerated materials with unique compositional control. In the sol-gel processing of complex oxides, whereby the different metal alkoxide precursors can have substantially different hydrolysis rates,26 a reverse microemulsion medium may provide a means to attain materials with enhanced chemical homogeneity by confining the hydrolysis and polycondensation reactions to nanometer-sized reverse micellar domains. In this approach, we can select the appropriate alkoxide precursors and adjust their concentrations, such that their diffusion rates through the oil phase are matched, allowing them to undergo hydrolysis simultaneously on reaching the aqueous domains. In this paper, a reverse microemulsion medium was employed for the synthesis of barium hexaaluminate (BHA) catalysts of interest to methane combustion applications. The improved compositional uniformity attainable with this novel method could yield nonagglomerated nanoparticles that would undergo crystallization at lower temperatures, thereby minimizing grain growth and maximizing specific surface area.24,25 The reverse microemulsion medium was also flexible to surface deposition of active species on BHA nanoparticles to achieve nanocomposite systems with excellent catalytic activity.25 Ionic surfactants, such as Aerosol OT, sulfonated hydrocarbons, and succinates, are typically used in reverse microemulsion-mediated synthesis of organic and inorganic particles. However, the anions or cations associated with the ionic surfactants may potentially contaminate the material of interest. For example, we have found that the synthesis of γ-Al2O3 nanoparticles in reverse micellar (19) Chhabra, V.; Ayyub, P.; Chattopadhyay, S.; Maitra, A. N. Mater. Lett. 1996, 26, 21. (20) Kawai, T.; Fujino, A.; No, K. K. Colloids Surf., A: Physicochem. Eng. Aspects 1996, 109, 245. (21) Moriya, Y.; Nishiguchi, N.; Kawakami, M.; Hino, R. J. Ceram. Soc. Jpn., Int. Ed. 1995, 103, 565. (22) Narita, T.; Nakagawa, K.; Kawasaki, K.; Ozaki, Y. J. Chem. Soc. Jpn., Int. Ed. 1996, 104, 623. (23) Roth, M.; Hempelmann, R. J. Mater. Chem. 1999, 9, 493. (24) Zarur, A. J.; Hwu, H. H.; Ying, J. Y. Langmuir 2000, 16, 3042. (25) Zarur, A. J.; Ying, J. Y. Nature 2000, 403, 65. (26) Bradley, D. C.; Mehrota, R. C.; Gaur, D. P. Metal Alkoxides and Diketonates; Academic Press: New York, 1978.
10.1021/la991488o CCC: $19.00 © 2000 American Chemical Society Published on Web 10/21/2000
Reverse Microemulsions Stabilized by Nonionic Surfactants
systems stabilized by sodium dodecyl sulfate led to materials with ∼2 at. % sulfur.27 Sulfate species confer strong surface acidity to catalysts,28 which may be undesirable for certain reactions. Nonionic surfactants have been recently studied for the synthesis of high-purity materials;24,25,29 polyethoxylated alcohols have been examined with particular interest. Although some studies have been conducted on the phase behavior of microemulsions stabilized by pure hexaethoxylated nonynol or pentaethoxylated hexanol,30-33 little information is available on systems stabilized by commercially available surfactant mixtures. This paper describes the synthesis of water/isooctane and water/ cyclohexane reverse microemulsions of interest to the mediated sol-gel processing of BHA-based materials. The systems were stabilized by commercially available polyethoxylated alcohols. Medium-chain linear alcohols (e.g., 1-pentanol) were employed as cosolvents in these systems to help balance the affinity of the surfactants to water and oil. During sol-gel processing, metal alkoxide precursors are hydrolyzed to yield hydroxides or oxy-hydroxides and short-chained alcohols in an exothermic reaction:
M(OCnH2n+1)x + xH2O f M(OH)x + xCnH2n+1OH (1) The metal hydroxides subsequently undergo polycondensation, whereby oxygen bridges are formed between metal atoms, and water is generated:34
M(OH)x + M′(OH)y f (HO)x-1sMsOsM′s(OH)y-1 + H2O (2) Thus, the phase stability of reverse microemulsions in the presence of short-chained alcohols and at different temperatures is important to achieving well-defined nanoparticles in the mediated sol-gel processing. This paper presents the phase behavior, micellar structure, and stability of water/isooctane and water/cyclohexane systems for the controlled synthesis of BHA nanoparticles. Experimental Section Reverse microemulsions were prepared with isooctane (2,2,4trimethylpentane, 99.9%, Aldrich) or cyclohexane (99.9%, Aldrich) as the oil phase. The Neodol 91-6 surfactant (Shell Chemical) used consisted of polyethoxylated alcohols with an average of 6 units of ethylene oxide per mole of alcohol, and an average hydrocarbon chain length of 10. The surfactant was stirred with 1-pentanol (99.9%, Aldrich) to give a 67-33% surfactant/cosolvent mixture by weight. The hydrophilic-lipophilic balance (HLB) of this mixture was calculated to be 11.82 according to the method proposed by Davies.35 The surfactant mixture was added to the two-phase water/oil mixture until a clear, one-phase system was obtained. The reverse microemulsions were set aside to verify that phase separation would not occur after stirring was stopped; additional surfactant mixture was added if necessary to achieve a stable one-phase system. (27) Zarur, A. J.; Ying J. Y., unpublished results. (28) Satterfield, C. N. Heterogeneous Catalysis in Industrial Practice; McGraw-Hill: New York, 1991. (29) Zarur, A. J.; Hwu, H. H.; Ying, J. Y., submitted for publication in J. Catal. (30) Kilpatrick, P. K.; Gorman, C. A.; Davis, H. T.; Scriven, L. E.; Miller, W. G. J. Phys. Chem. 1986, 90, 5292. (31) Minero, C.; Pelizzetti, E. J. Dispersion Sci. Technol. 1995, 16, 1. (32) Kahlweit, M.; Busse, G.; Jen, J. J. Phys. Chem. 1991, 95, 5580. (33) Strey, R.; Jonstromer, M. J. Phys. Chem. 1992, 96, 4537. (34) Brinker, C. J. Sol-Gel Science; Academic Press: Boston, 1990. (35) Davies, J. T. Hydrophilic Lipophilic Balance Calculation by Group Additivity, Vol. 1; Butterworth: London, 1957; p 128.
Langmuir, Vol. 16, No. 24, 2000 9169 Thermal stability experiments were conducted with a circulating water bath. Reverse microemulsions containing the minimum amount of surfactant required for their stabilization were heated until a cloud point was observed. Extra surfactant was then added with stirring until a transparent system was reestablished. Experiments were also performed to determine the stability of reverse microemulsions in the presence of shortchained alcohols. 2-Propanol (99.9%, Mallinckrodt) was introduced to a reverse microemulsion until a cloud point was noted. Extra surfactant was then added to reestablish the transparency of the system. Quasi-elastic (dynamic) light scattering (QELS) experiments were conducted on the reverse microemulsions at 25 °C with a Lexel 95 Argon-ion laser and a Brookhaven high-precision photomultiplier at 90°. Data were acquired with a Brookhaven 9000AT correlator. The periods for acquisition of the autocorrelation function ranged from 0.1 µs to 0.5 s. Particle size distribution was obtained from nonnegatively constrained leastsquares (NNCLS) analysis of the data.36 Viscosity measurements were conducted with Glimont capillary flow viscometers at constant temperature. Electrical conductivity measurements were performed with a Copenhagen Radiometer conductivity cell CDC14 (1-cm-1 cell constant) connected to a Copenhagen Radiometer CDM80 conductivity meter. Small-angle neutron scattering (SANS) experiments were performed on the 30-m SANS instrument on neutron guide NG3 at the Center for Neutron Research of NIST. The instrument utilizes a mechanical velocity selector as monochromator, a circular pinhole collimation, and a two-dimensional positionsensitive detector (65 × 65 cm2) for data collection over a range of angles simultaneously. Data were taken with 6-Å neutrons and sample-to-detector distances of 2 and 13 m, which covered a range of scattering vectors Q (0.04-3.0 nm-1). The data were corrected for background, empty-cell scattering, and sample transmission, as described elsewhere.37 Reverse microemulsion samples for SANS analysis were prepared by substituting protonated water with heavy water in the synthesis described earlier. Conversely, for the SANS studies of regular microemulsions, the protonated hydrocarbons were substituted by their deuterated isotopes. Microemulsion samples for SANS contrast matching experiments were prepared by substituting the hydrocarbons by deuterated species and by adding protonated water to D2O to obtain the same scattering length density for both phases. Deuterated compounds were obtained from Cambridge Isotope Laboratories, Inc. (Andover, MA) and used without further purification. The synthesis of BHA nanoparticles was conducted as described elsewhere.24,25 Briefly, barium and aluminum isopropoxides (99%, Aldrich) were refluxed overnight at 110 °C in 2-propanol and then allowed to cool under inert atmosphere. Reverse emulsions were prepared as already described. The alkoxide solutions were reacted with the aqueous phase of the reverse microemulsions by adding the alkoxide precursor solutions in a dropwise manner to the reverse microemulsions while stirring. The water:alkoxide molar ratio was kept at 100 times the stoichiometric value. The hydrolyzed samples were allowed to age for 24 h, after which they were recovered by freeze-drying. This technique effectively removed light fractions of the surfactants, alcohols, and water. The remaining surfactants were eliminated by calcination at 500 °C for 4 h under nitrogen and 4 h under air. The materials were then heat treated to 800 and 1300 °C under air to promote crystallization of the BHA phase.
Results and Discussion I. Phase Behavior. We found that Neodol 91-6/1pentanol surfactant mixture could be used to stabilize microemulsions over a wide range of water:oil ratios. Figure 1 illustrates the ternary phase diagrams for the water/isooctane/surfactant and water/cyclohexane/sur(36) Grabowski, E. F.; Morrison, I. D. Particle Size Distributions from Analysis of Quasi-Elastic Light-Scattering Data; Dahneke, B. E., Ed.; Wiley-Interscience: New York, 1984; p 199. (37) Glinka, C. J.; Barker, J. G.; Hammouda, B.; Krueger, S.; Moyer, J. J.; Orts, W. J. J. Appl. Crystallogr. 1998, 31, 430.
9170
Langmuir, Vol. 16, No. 24, 2000
Figure 1. Pseudo-ternary phase diagrams for (a) the water/ isooctane/surfactant system and (b) the water/cyclohexane/ surfactant system. The surfactant mixture consisted of 66.7 wt % Neodol 91-6 and 33.3 wt % 1-pentanol.
factant systems. These diagrams indicate that the amount of surfactant mixture required to stabilize the microemulsions was lower for the water/isooctane system than for the water/cyclohexane system. We are particularly interested in the reverse microemulsions containing water volume fractions (φw) of 0.05-0.15, because this φw range was found to be optimal for the synthesis of complex oxides through reverse microemulsion-mediated sol-gel processing.24 At φw ) 0.1, the surfactant volume fractions (φs) required to stabilize the reverse microemulsion were 0.13 and 0.20 for the isooctane-based and cyclohexane-based systems, respectively. At a higher water content of φw ) 0.3, φs values of 0.23 and 0.38 were required to stabilize the isooctane-based and cyclohexane-based systems, respectively. The difference in the surfactant requirements for the two systems could be attributed in part to the difference in compatibility between the hydrophobic tails of the surfactants and the respective oil phase. The saturated linear hydrocarbon tails of the surfactants should be more soluble in isooctane than in cyclohexane. This difference in solubility would lead to a reduced interfacial area between the oil phase and the amphiphilic film for the former, resulting in a lower surfactant requirement. For the water/isooctane system, the required surfactant volume fraction (φs) increased gradually with water content and reached a local maximum at φw ≈ 0.35. After this point, the surfactant fraction decreased slightly from φs ) 0.25 at φw ) 0.35 to φs ) 0.21 at φw ) 0.49. The
Zarur et al.
required surfactant fraction then increased abruptly, reaching a maximum value of 0.35 at φw ≈ 0.55. The phase diagram of the water/cyclohexane/surfactant system was significantly different from that of the water/ isooctane/surfactant system. A much larger amount of surfactant was required to stabilize the water/cyclohexane system, especially at low oil volume fractions of e0.4. Additionally, unlike the isooctane system, the ternary phase diagram for the cyclohexane system did not show the abrupt changes in φs, which were indicative of phase transitions. There were, however, two inflection points in the phase diagram of the cyclohexane system that might correspond to phase transitions. To investigate if the features observed in the ternary phase diagrams corresponded to transitions between different microemulsion phases, electrical conductivity, light scattering, and neutron scattering experiments were performed on selected compositions in the sections to follow. II. Electrical Conductivity. Electrical conductivity illustrates the degree of percolation or the bicontinuous nature of a microemulsion system. At low water volume fractions where discrete reverse micelles were dispersed within a continuous oil phase, the system would have a conductivity similar to that of the oil phase. Conversely, at high water contents where regular micelles were dispersed in a continuous aqueous phase, the system would display a conductivity similar to that of the water used (∼160 µs/cm). For microemulsions containing similar values of φw and φo, the conductivity might be intermediate to those of pure water and pure oil. In our case, both isooctane and cyclohexane have negligible electrical conductivity. Therefore, significant deviation from zero conductivity would imply a certain degree of connectivity between the aqueous domains in the system. For φwe 0.06, the conductivity of the water/isooctane system was negligible, suggesting that the water droplets were discrete and have little interaction with each other. Conductivity rose as φw increased from 0.06 to 0.80, indicating that interaction between the aqueous domains became increasingly important. For φwg 0.8, conductivity of the system was similar to that of water, suggesting that the system consisted of a continuous aqueous phase with discrete oil droplets. For the water/cyclohexane system, zero conductivity was noted for φw e 0.10. Conductivity increased steadily between φw values of ∼0.10 and ∼0.60, suggesting a bicontinuous system. Systems of φw g 0.6 showed a conductivity value close to that of water, indicating the presence of a continuous aqueous phase with discrete oil droplets. III. Quasi-Elastic Light Scattering Analysis. Particle size determination by QELS is most reliable at dilute concentrations, where the Stokes-Einstein relation can be applied.36 Particle size analysis by QELS is reliable and readily applicable to monodisperse, or slightly polydisperse, noninteracting scattering centers.38 On the other hand, the analysis is not meaningful (i) when a strong interaction is present between particles or (ii) for bicontinuous or lamellar structures. Previous researchers have observed that the hydrodynamic radius obtained from the application of the StokesEinstein equation to QELS data agreed well with values obtained from the neutron scattering data analysis for φw e 0.3 in water/dodecane reverse microemulsions.39 For the water/isooctane system, we found that the hydrodynamic radius of the reverse micelles increased almost (38) Koppel, D. E. J. Chem. Phys. 1972, 57, 4814.
Reverse Microemulsions Stabilized by Nonionic Surfactants
Langmuir, Vol. 16, No. 24, 2000 9171 Table 1. Electrical Conductivity, and Core Radius and Shell Thickness of Aqueous Domains Obtained from Core-Shell Model fitting of SANS Data for D2O/Isooctane Systems with Low Water Contents
Figure 2. QELS hydrodynamic radius for water/isooctane systems as a function of φw.
linearly with water content from 2.9 nm at φw ) 0.003 to 7.3 nm at φw ) 0.15 (Figure 2). A significant decrease in the hydrodynamic radius was observed between φw ) 0.15 and 0.40. This result might be due to an increased level of interaction between aqueous particles (as suggested by electrical conductivity data), which could lead to a higher apparent diffusion coefficient, effectively reducing the hydrodynamic radius obtained from the Stokes-Einstein equation. Nonnegatively constrained least-squares (NNCLS) analysis of the autocorrelation function showed that the particle distribution agreed well with slightly polydisperse spherical particles for systems of φw e 0.15. Figure 3(a) and (b) illustrate the particle size distribution and residuals of the NNCLS analysis, respectively, for a system of φw ) 0.15. The residuals from the statistical fitting of autocorrelation scattering function for systems of 0.005 e φw e 0.15 were randomly distributed, indicating that no systematic error was present in the fitting operation. In contrast, NNCLS analysis did not produce satisfactory fitting of autocorrelation data for systems of φw > 0.15. For example, a bimodal particle size distribution was noted for the system of φw ) 0.25, centering at 0.15, which might be attributed to the high levels of interaction between aqueous particles. These findings agreed well with the significant electrical conductivity values noted for systems of φw > 0.15. In the water-rich region of the phase diagram (φw > 0.55), a low level of interaction between discrete oil droplets could be expected in a continuous aqueous phase. However, meaningful QELS data could not be collected for oil particle size distribution in the water-rich region because of weak scattering and significant fluctuations in the baseline. Relatively smaller particle sizes were attained for the water/cyclohexane system (Figure 4) than for the water/ isooctane system. At φw ) 0.03, a hydrodynamic radius of ∼2.8 nm was determined for the water/cyclohexane system; a slightly polydisperse NNCLS radius distribution of ∼2.9 nm with random residual scatter was obtained. Analysis of QELS data was meaningful up to φw ) 0.10, where a hydrodynamic radius of ∼4.5 nm and a NNCLS radius distribution of ∼4.6 nm were calculated. Significant (39) Chang, N. J.; Billman, J. F.; Licklider, R. A.; Kaler, E. W. On the Structure of Five-Component Microemulsions; Chen, S.-H., Rajagopalan, R., Eds.; Springer-Verlag: New York, 1990; p 269.
φw
electrical conductivity (µS/cm)
core radius (nm)
shell thickness (nm)
φs/φo
0.01 0.03 0.06 0.12
0.2 0.3 0.8 15
2.2 2.6 3.0 4.1
3.4 3.0 2.9 3.1
0.04 0.12 0.20 0.24
fluctuations in the baseline and multimodal scattering autocorrelation functions prevented the meaningful statistical analysis of data for systems of φw > 0.10. This result agreed well with the conductivity measurements for the water/cyclohexane systems, which suggested significant percolation effects at φw > 0.10. IV. Small-Angle Neutron Scattering Analysis. QELS provided valuable particle size data on the oil-rich systems, including the 0.05 e φw e 0.15 range that was of great interest for reverse microemulsion-mediated synthesis of nanoparticulate materials. However, limited information was obtained from QELS for systems with intermediate or high water contents. To elucidate the particle size as well as the aggregation state of the aqueous and oil phases for a wide range of microemulsion compositions, SANS experiments were conducted on the water/ isooctane and water/cyclohexane systems. Figure 5 plots the absolute scattering intensity (Io) as a function of scattering length (q) for D2O/isooctane systems of 0.01 e φw e 0.12. The data were fitted with a model for polydisperse spherical particles with a coreshell structure.40 The form factor of the spherical core was calculated assuming a hard sphere model with the Percus-Yevick approximation,41 and was normalized by the average particle volume.42 The scattering length density (SLD) describes the interaction of the neutron beam with the material per unit volume. For the inner aqueous core, the SLD (∼6 × 10-6 Å-2) matched well with that of pure D2O (6.4 × 10-6 Å-2). The shell appeared to be composed of partially hydrated surfactant molecules because its SLD (-1 × 10-7 Å-2) was slightly larger than that expected for the linear hydrocarbon tails and the ethylene oxide headgroups (∼-3 × 10-7 Å-2), suggesting the presence of a small concentration of D2O. The SLD of the solvent (-5 × 10-7 Å-2) obtained from modeling matched well with that of protonated isooctane (-5.2 × 10-7 Å-2). The thickness of the shell remained relatively constant for systems of 0.01 e φw e 0.12, whereas the core radius increased almost linearly with water content. The total particle radius (core radius + shell thickness) from SANS modeling (Table 1) matched well with the hydrodynamic radius obtained from QELS for systems of low water contents. Contrast matching experiments at low water contents were performed to investigate the distribution of amphiphiles in the reverse microemulsions. As expected, the core diameter remained unchanged for systems prepared with protonated or deuterated isooctane. Protonated samples have shells of ∼3 nm in thickness, independent of water content, whereas deuterated samples have thinner shells of ∼1.8 nm. This result suggested that roughly half of the surfactant tail interpenetrated the oil phase. (40) Kline, S. Electronic Reference; NIST, 1998. (41) Ashcroft, N. W.; Lekner, J. Phys. Rev. 1966, 145, 83. (42) Bartlett, P.; Ottewill, R. H. J. Chem. Phys. 1992, 96, 3306.
9172
Langmuir, Vol. 16, No. 24, 2000
Zarur et al.
Figure 3. (a) NNCLS particle size distribution and (b) residual analysis for a water/isooctane system of φw ) 0.15. The surfactant: oil volume ratio (φs/φo) was 0.26.
Figure 4. QELS hydrodynamic radius for water/cyclohexane systems as a function of φw.
The scattering function of the system of φw ) 0.15 could not be fitted using a core-shell model or a bicontinuous model, suggesting that this sample was composed of a mixture of both phases (Figure 6(b)). For 0.18 e φw e 0.57, the scattering functions could be fitted with the TeubnerStrey model43,44 for bicontinuous systems (Figure 7). This model yields a correlation length ξ, which is a measure (43) Teubner, M.; Strey, R. J. Chem. Phys. 1987, 87, 3195. (44) Schubert, K.-V.; Strey, R.; Kline, S. R.; Kaler, E. W. J. Chem. Phys. 1994, 101, 5343.
Figure 5. Absolute SANS intensity as a function of scattering length for D2O/isooctane systems of (a) φw ) 0.01 (φs/φo ) 0.04), (b) φw ) 0.03 (φs/φo ) 0.12), (c) φw ) 0.06 (φs/φo ) 0.20), and (d) φw ) 0.12 (φs/φo ) 0.24).
of the channel size of the dispersed phase, and a quasiperiodic repeat distance d, which is analogous to the interplanar distance of a periodic system. Contrast matching experiments on systems with intermediate water contents indicated that the correlation length ξ obtained from the Teubner-Strey model corresponded to the characteristic size of the aqueous domains only
Reverse Microemulsions Stabilized by Nonionic Surfactants
Langmuir, Vol. 16, No. 24, 2000 9173 Table 3. Theoretical Interfacial Area, and Comparison between Theoretical and Actual Surfactant Requirements for Water/Isooctane Systems with Different Water Contents
φw
interfacial area (m2/cm3)
theoretical surfactant requirements (φs0)a
actual surfactant requirements (φs)
0.01 0.03 0.06 0.12 0.31 0.45 0.57
28 58 88 100 128 143 257
0.04 0.08 0.12 0.14 0.18 0.29 0.36
0.09 0.12 0.13 0.16 0.23 0.34 0.33
a Based on parameters obtained from polydisperse core-shell or Teubner-Strey model fitting of SANS data.
Figure 6. Absolute SANS intensity as a function of scattering length for D2O/isooctane systems of (a) φw ) 0.12 (φs/φo ) 0.24), (b) φw ) 0.15 (φs/φo ) 0.26), and (c) φw ) 0.18 (φs/φo ) 0.30).
Figure 7. Absolute SANS intensity as a function of scattering length for D2O/isooctane systems of (a) φw ) 0.18 (φs/φo ) 0.30) (×), (b) φw ) 0.31 (φs/φo ) 0.51) (9), (c) φw ) 0.45 (φs/φo ) 0.60) (+), and (d) φw ) 0.57 (φs/φo ) 3.6) (O). Table 2. Electrical Conductivity, and Correlation Length and Quasi-Periodic Repeat Distance of Aqueous Domains Obtained from Teubner-Strey Model Fitting of SANS Data for D2O/Isooctane Systems of 0.18 e Ow e 0.57
φw
electrical conductivity (µS/cm)
correlation length (ξ) (nm)
repeat distance (d) (nm)
φs/φo
0.18 0.31 0.45 0.57
60 80 90 110
8.4 6.2 5.4 3.8
14.4 12.5 10.9 6.8
0.30 0.51 0.60 3.6
(without the surfactant film) because similar values of ξ and d were obtained for contrast-matched and regular samples. The correlation length parameters obtained from fitting the SANS data (Table 2) did not correspond well to the Stokes-Einstein hydrodynamic radii from QELS. This difference was because the hydrodynamic radii obtained from analysis of QELS data were not meaningful for systems of φw > 0.15 because the Stokes-Einstein assumptions of negligible particle-particle interaction and infinite dilution did not apply to these cases. Systems with high water contents were examined in attempt to elucidate the structure of the oil domains in regular microemulsions. SANS data of the iso-C8D18/H2O systems of φw g 0.71 (corresponding to φo e 0.06) showed very low scattering intensities; consequently, satisfactory statistical fits could not be obtained for these scattering functions.
Figure 8. Absolute SANS intensity as a function of scattering length for D2O/cyclohexane systems of (a) φw ) 0.012 (φs/φo ) 0.05), (b) φw ) 0.11 (φs/φo ) 0.34), (c) φw ) 0.33 (φs/φo ) 2.2), and (d) φw ) 0.57 (φs/φo ) 10.5).
The SANS results were in good agreement with the conductivity measurements. The bicontinuous structure proposed for water/isooctane systems of 0.18 e φw e 0.57 corresponded to a high degree of particle interaction, as indicated by the increases in electrical conductivity. The SANS results were also in agreement with the phase behavior of the system (Figure 1(a)). At low water contents, φs increased with φw in the phase diagram; this result was consistent with the additional requirement of amphiphilic molecules for stabilization of systems with increasing number density and particle size of discrete aqueous domains. The φs value continued to increase gradually with φw when a bicontinuous structure was formed because the interfacial area between the aqueous and oil phases would increase with the phase change. A significant increase in surfactant concentration occurred at φw ≈ 0.5. This result matched well with a significant decrease in the correlation length from the Teubner-Strey model in that region (Table 2), which might be attributed to a phase inversion of the system, resulting in an increased interfacial area. Table 3 shows the theoretical interfacial area and surfactant requirements calculated based on an estimated molecular cross-sectional area of 40 Å2 for the surfactant system. The interfacial area for the bicontinuous phase was approximated by a network of cylindrical channels. We noted that the theoretical and actual surfactant requirements agreed well for systems of 0.01 e φw e 0.57. Figure 8 and Table 4 show the SANS data for the water/ cyclohexane system, which were significantly different from those for the water/isooctane system. The core-shell model, which was applicable for the water/isooctane
9174
Langmuir, Vol. 16, No. 24, 2000
Zarur et al.
Table 4. Electrical Conductivity, and Correlation Length and Quasi-Periodic Repeat Distance of Aqueous Domains Obtained from Teubner-Strey Model Fitting of SANS Data for D2O/Cyclohexane Systems of 0.012 e Ow e 0.57
φw
electrical conductivity (µS/cm)
correlation length (ξ) (nm)
repeat distance (d) (nm)
0.012 0.03 0.06 0.11 0.27 0.33 0.57
0.1 0.2 0.3 0.8 19 60 149
1.9 2.2 3.2 4.3 5.1 3.7 3.1
10.1 9.1 10.8 13.0 9.7 6.7 6.7
systems of φw e 0.12, could only be used to describe the water/cyclohexane systems of φw e 0.01. For water/ cyclohexane systems of 0.012 e φw e 0.57, the bicontinuous model gave the best fit to the scattering data. In contrast to the SANS data, the conductivity measurements showed negligible connectivity between aqueous aggregates for systems of φw e 0.1. This phenomenon might be explained by the postulate that the aqueous phase for systems of 0.012 e φw e 0.1 might have formed a network of cylinderlike structures, which possessed no long-range periodicity and was not fully interconnected. Indeed, analysis of the SANS data showed that at low water contents (e.g. φw ) 0.012), the correlation length calculated with the Teubner-Strey model was very small (1.9 nm), and the quasiperiodic repeat distance was relatively large (10.1 nm). In contrast, at intermediate water contents (e.g., φw ) 0.27), the correlation length was found to be significantly larger (5.1 nm) with a repeat distance that remained largely unchanged (9.7 nm). As for the water/isooctane systems with high water contents (φw g 0.71), theoretical fits of SANS data obtained for the water/cyclohexane systems of φw g 0.69 were not statistically significant because of the low intensity of the scattering data. The results from the contrast matching experiments on samples with intermediate water contents (0.012 e φw e 0.57) for the water/cyclohexane system were analogous to those obtained for the water/isooctane system (0.18 e φw e 0.57). Similar values of ξ and d were noted for contrast-matched and regular samples, indicating that the correlation length ξ attained from the Teubner-Strey model corresponded to the characteristic size of the aqueous domains only (i.e., without the surfactant film). V. Thermal Stability. In reverse microemulsionmediated sol-gel synthesis of nanoparticles, heat was expected to be generated by the exothermic hydrolysis reactions of metal alkoxide precursors. Depending on the initial precursor concentration and the water:alkoxide ratio used in the reaction, the temperature of the system might increase by 10 °C, which could lead to undesired phase separation of the reverse microemulsion. Figure 9 shows a plot of the excess surfactant required for stabilizing water/isooctane systems with initial water volume fractions of 0.12 and 0.20. The excess surfactant represented the additional volume of Neodol 91-6/1-pentanol mixture introduced to regain transparency of a system when subjected to the temperature noted. Figure 9 illustrates that the reverse microemulsions of φw ) 0.12 and 0.20 did not require excess surfactant for phase stabilization at the temperature range of interest for mediated sol-gel synthesis (25-35 °C). In fact, cloud points corresponding to phase separation did not occur until ∼61 and ∼53 °C for systems of φw ) 0.12 and 0.20, respectively, indicating the excellent thermal stability of these reverse microemulsions. Even at a temperature of ∼75 °C, these systems could be stabilized by doubling the
Figure 9. Thermal stability of water/isooctane systems of (a) φw ) 0.12 (initial φs/φo ) 0.24) and (b) φw ) 0.20 (initial φs/φo ) 0.33), expressed in terms of excess surfactant needed to stabilize the reverse microemulsions at a given temperature.
surfactant amount introduced. Measurements were not conducted at temperatures >85 °C because of the significant evaporation of isooctane at such a high temperature. Although we did not observe any macroscopic phase disruption of our system at temperatures
