Structure and Dynamics of Micelles and Cubic Phase Structures with

Nestlé Research Centre, Vers-chez-les-Blanc, 1000 Lausanne 26, Switzerland, and Department of Applied. Surface Chemistry, Chalmers UniVersity of ...
0 downloads 0 Views 110KB Size
Langmuir 2008, 24, 6441-6446

6441

Structure and Dynamics of Micelles and Cubic Phase Structures with Ethoxylated Phytosterol Surfactant in Water Britta M. Folmer† and Magnus Nyde´n*,‡ Nestle´ Research Centre, Vers-chez-les-Blanc, 1000 Lausanne 26, Switzerland, and Department of Applied Surface Chemistry, Chalmers UniVersity of Technology, SE-412 96 Go¨teborg, Sweden ReceiVed NoVember 30, 2007. ReVised Manuscript ReceiVed April 15, 2008 The self-assembly of a sterol ethoxylate surfactant with 30 oxyethylene units in water was studied by 1H NMR self-diffusion measurements in a wide concentration range in the micellar region (0-25 wt %). The data showed that the surfactant aggregates do not interact by hard sphere interactions but rather a strong concentration dependence of the diffusion coefficient was noted which was explained by polymer scaling theory. In the cubic phase (30-65 wt %), the self-diffusion data from water, from surfactant, and from free polyoxyethylene suggest spherical micelles, although water diffusion was much restricted due to binding to the surfactant headgroup. From X-ray measurements in the cubic phase, the unit cell size was calculated, and together with surfactant self-diffusion measurements the exchange dynamics between free and aggregated surfactant was obtained.

Introduction Sterol is a common name for the unsaponificable fraction of lipid extracts, derived from either plants or animal sources. The general structure is a perhydrocyclopentenophenanthrene nucleus with a side chain in position C17 consisting of 8-10 carbon atoms. The steroid nucleus has a rather rigid character due to the condensed ring structure unlike the carbon side chain with much more molecular flexibility. The most common sterol derived from animals is cholesterol.1,2 Sterols synthesized by plants are called phytosterols, which is a general name for different sterol derivatives present in the unsaponifiables of different plants. The main components are usually β-sitosterol, campesterol, stigmasterol, and toccopherol.3 Sterol ethoxylates have been used in applications in both pharmaceutical products and cosmetics.4–6 Due to the biological relevance of these surfactants, it is very important to understand how they behave in different surroundings. One such investigation is presented here in which we investigate how the sterol surfactant aggregates at high concentrations and how this affects the structure and the dynamics of the surfactant as well as water and other solutes that are used as probe molecules for the self-diffusion measurements. In a previous study, the physicochemical behavior of a series of phytosterol ethoxylate with varying polyoxyethylene chain length was described.7 The binary phase behavior of the surfactants in water was schematically presented by means of optical microscopy. It was found that the phytosterols with 20 and 30 oxyethylene units form micellar phases at low surfactant concentrations, which with increasing surfactant concentration transformed into cubic and hexagonal phases. * To whom correspondence should be addressed. Telephone: +46 31 772 29 73. Fax: +46 31 16 00 62. E-mail: [email protected]. † Nestle´ Research Centre. ‡ Chalmers University of Technology.

(1) Scotney, J.; Truter, E. V. J. Soc. Cosmet. Chem. 1971, 22, 201. (2) Lundmark, L.; Chun, H.; Melby, A. Soap, Cosmet., Chem. Spec. 1976, 33. (3) Wachter, R.; Salka, B.; Magnet, A. Cosmet. Toiletries 1995, 110, 72. (4) Beugin, S.; Edwards, K.; Karlsson, G.; Ollivon, M.; Lesieur, S. Biophys. J. 1998, 74, 3198. (5) Vertut-Do¨i, A.; Ishiwata, H.; Miyajima, K. Biochim. Biophys. Acta 1996, 1287, 19. (6) Folmer, B. M. In Annual Surfactants ReView; Academic Press: Sheffield, U.K.; in press. (7) Folmer, B. M.; Svensson, M.; Holmberg, K.; Brown, W. J. Colloid Interface Sci. 1999, 213, 112.

The micellar region was found to exist up to a concentration of around 20-25 wt %, and the variation in axial ratios of the aggregates with increasing surfactant concentration was found to be somewhat puzzling. The micellar diffusion coefficients strongly suggested nonspherical micelles, with an axial ratio of 5 at 5 wt %. The cubic phase consisted of discrete aggregates, and upon increasing the concentration further an hexagonal phase existed up to very high concentrations. Since the cubic phase is located between the micellar and the hexagonal phase in the binary phase diagram, the finding of discrete aggregates is by no means surprising. There are however many different ways to arrange discrete micelles in a cubic arrangement. Not uncommon for the polyoxyethylene based surfactants though is that the micelles in a discrete cubic phase have axial ratios of 1 or slightly larger, with body-centered or face-centered cubic lattice arrangements.8,9 In this work, the diffusion coefficient of micelles in the micellar region have been analyzed using polymer scaling theory. Typical scaling parameters are extracted and compared to previous results in similar types of systems and in model polymer systems.10 In the cubic phase, which exists in a wide concentration region, the combination of X-ray and NMR self-diffusion measurements of the surfactant have been used to calculate the exchange dynamics between their free and aggregated states. Furthermore, the diffusion mechanisms of water and unreacted polyoxyethylene fractions are discussed, and the results are used for interpreting the structure and dynamics in the cubic phase.

Experimental Section Materials. The phytosterol ethoxylate was a generous gift from Nikkol, Japan. The surfactant consists of a hydrophobic phytosterol with a hydrophilic polyoxyethylene chain of an average of 30 units. All solutions were made in Millipore-filtered water (in D2O for the samples in the micellar region). Unreacted polyoxyethylene with a molecular weight of approximately that of the size of the surfactant headgroup was present in small amounts (3-5 wt %). (8) Sakaya, P.; Seddon, J. M.; Templer, R. H.; Mirkin, R. J.; Tiddy, G. J. T. Langmuir 1997, 13, 3706. (9) Jo¨nsson, B.; Lindman, B.; Holmberg, K.; Kronberg, B. Surfactants and Polymers in Aqueous Solution; Wiley: Chichester, U.K., 1998. (10) Håkansson, B.; Hansson, P.; Regev, O.; So¨derman, O. Langmuir 1998, 14, 5730.

10.1021/la703758p CCC: $40.75  2008 American Chemical Society Published on Web 05/29/2008

6442 Langmuir, Vol. 24, No. 13, 2008

Folmer and Nyde´n

Figure 1. Structure of β-sitosterol ethoxylate with 30 oxyethylene groups (n ) 30).

The phytosterol used in this study was isolated from soybeans and consists of β-sitosterol, stigmasterol, and campesterol in a 2:1:1 ratio. Ethoxylation was performed using KOH to generate the alkoxylate. After completed ethoxylation, the product was neutralized with acid (personal communication Nikkol). The structure of β-sitosterol ethoxylate is shown in Figure 1. The critical micelle concentration (CMC) of this surfactant was determined to be of the order of 10-3 mM. Samples were prepared in Milli-Q water in the concentration range 20-70 wt %. 1H NMR Self-Diffusion. Self-diffusion was measured on an Oxford 500 MHz spectrometer equipped with a diffusion probe by DOTY Sci. Inc. Diffusion coefficients of polyoxyethylene and water where determined by fitting eq 1 to the raw data obtained from a stimulated echo using sine-shaped pulsed field gradients

I(δ, ∆, g, τ) ) -T -2τ 1 exp exp(-γ2g2δ2(4∆ - δ)/π2D) (1) I exp 2 0 T1 T2

( ) ( )

where I(δ,∆,g,τ) is the echo intensity and I0 the signal strength following a 90° pulse. T2 is the delay between the second and third radio frequency pulse, T2 and T2 are the longtitudal and transversal relaxation times, respectively. τ is the delay between the first and second radio frequency pulse, γ is the magnetogyric ratio, g is the field gradient strength, and δ is the duration of the field gradient pulse. ∆ is the effective diffusion time and approximately equal to T + τ. To account for the presence of free polyoxyethylene, eq 1 was modified to contain a sum of two exponents with diffusion coefficients D1 (surfactant) and D2 (free polyoxyethylene), with each exponential term weighted with the fraction of each species. Measurements were made in 5 mm NMR tubes with 0.7 mL of sample. The micellar samples were made from a concentrated micellar solution which was diluted by D2O to the correct concentrations, after which the samples were transferred to NMR tubes. The cubic samples were prepared by weighing surfactant and water directly into the NMR tube, after which the tubes were flame sealed and kept at elevated temperatures until the phases became homogeneous. The measurements were performed at 18 and 33 °C. A small amount of hexamethyldisiloxane (HMDS) was added to selected samples in the concentration series in order to probe the micellar diffusion, since HMDS is very hydrophobic and can be safely assumed to reside only in the hydrophobic core of the micelles and not in the water. X-Ray Diffraction. X-ray diffraction was measured on sterol ethoxylate at 47 wt % at 18 and 33 °C. The measurement was carried out at the EPSRC Daresburry synchrotron radiation laboratory. The experimentally obtained diffraction peaks were used to calculate space distances by calibrating the equipment with rat tail collagen.

Results and Discussion Micellar Region. In our previous diffusion study of the same surfactant, we were mostly interested in the low concentration region.7 The diffusion coefficients obtained were evaluated by assuming a “standard” micelle diffusion model and resulted in nonspherical micelles already at low micelle concentrations. Leaver et al. developed a working scheme by measuring solvent diffusion as well as surfactant diffusion and 2H NMR relaxation

Figure 2. Surfactant diffusion in the micellar region. The line represents the best fit of eq 2 with the following parameters: D0 ) 2.96 × 10-11 m2 s-1, γ ) 1.59, and R ) 0.03.

and could thereby with the help of a theoretical diffusion model by Jonstro¨mer et al.14 account for diffusion and relaxation data of a C12E5 surfactant in a self-consistent manner.11 Using the scheme by Leaver et al. and the model by Jonstro¨mer et al., a polyoxyethylene sterol with 30 oxyethylene units formed very large micelles, while the sterols with 20 and 30 oxyethylene units formed smaller structures although with growing axial ratio upon increasing surfactant concentration. The presence of the rodlike structures at low surfactant concentration leads to inconsistencies in the physicochemical understanding of the structure of the cubic phase, formed by the more hydrophilic surfactants. (We note that a two phase system consisting of discrete cubic aggregates in equilibrium with a rodlike micelles is unlikely.) In this study, the diffusion in the micellar region is instead analyzed according to a phenomenological polymer diffusion theory as originally suggested by Phillies.12 The diffusion coefficients of the micellar aggregates were indeed well-described by this theory, although it was originally suggested for polymer solutions. In this theory, the diffusion coefficient depends on the polymer concentration through a “stretched” exponential according to eq 2:

D(c) ) D0 exp(-Rcγ)

(2)

where R and γ are scaling parameters and D0 is the diffusion coefficient at zero polymer concentration. Equation 2 was also shown to accurately describe the diffusion behavior of micellar aggregates of C17E84 in a recent study by Håkansson et al.10 In Figure 2, the best fit of eq 2 to the diffusion data in the micellar region is shown where γ is 1.59. For a pure polymer solution, 0.5 e γ e 1 depending on the solvent quality.12 In the study by Håkansson et al., a pure polyoxyethylene polymer (E98) in water gave γ ) 0.92 while the C17E84 polymer gave γ ) 1.21. The difference was ascribed to the micellar structure of C17E84 which with respect to the polyoxyethylene part more resembles a star polymer than a linear polymer. Since eq 2 is more or less phenomenologically based, it is difficult to interpret the scaling parameters in terms of structural details. However, it is clear that a hydrocarbon core into which the polyoxyethylene chains cannot penetrate makes γ larger. The surfactant hydrocarbon part and the aggregation number are larger in this work, which is likely the main reason for the different values of γ. (11) Leaver, M.; Furo, I.; Olsson, U. Langmuir 1995, 11, 1524. (12) Phillies, G. D. J. J. Phys. Chem. 1989, 93, 5092.

Micelles, Cubic Phase Structures with Phytosterol

Langmuir, Vol. 24, No. 13, 2008 6443

The scaling prefactor in eq 2, R, was 0.03 for the sterol surfactant but 0.1 for C17E84, with the reason most likely being the large difference in the polyoxyethylene size which makes the C17E84 micelles more polymer-like, compared to the sterol ethoxylate with “only” 30 oxyethylene units. This conclusion is supported by the fact that R is well-known to be larger for nonaggregating polymer solutions.12 Although qualitative in nature, R can be used to compare diffusion mechanisms of different macromolecules in solution. Low R values mean that the hard-sphere diffusion mechanism prevails, and larger values indicate a significant contribution from entangling polymers. It is not surprising that the value obtained depends on the molecular weight of the polyoxyethylene groups as indeed was the case in the comparison between the sterol surfactant with 30 oxyethylene groups and the C17E84 surfactant. Assuming a spherical aggregate, the Stoke-Einstein relation can be used to calculate the hydrodynamic radius through eq 3.

RH )

kBT 6πηD0

(3)

where kB is the Boltzmann constant, T is the absolute temperature, η is the solution viscosity, and RH is the hydrodynamic radius of the micelle. D0 as also obtained in Figure 2 from the fit of eq 2 to the diffusion data was 2.96 × 10-11 m2 s-1. The hydrodynamic radius then was 67 Å. Olsson and Schurtenberger13 accounted for the water layer surrounding a C12E5 micelle and obtained 9 Å by dynamic light scattering measurements. By assuming that this number holds also for larger micelles, the radius for the sterol micelle, including polyoxyethylene bound water but not the hydrodynamically bound water, was 58 Å. From static light scattering in our previous investigation, this particular sterol surfactant had an aggregation number of 95.7 Using the known volume of the sterol, the hydrocarbon core radius was calculated to be 24 Å, which gives a surfactant area at the hydrocarbon/ polyoxyethylene interface of 76 Å2. From the experimentally obtained micelle volume, the volume calculated for the surfactant, and the previously determined aggregation number, the amount of water in the polyoxyethylene core was estimated to be 77%, which corresponds to seven D2O molecules per oxyethylene unit. This number is in good agreement with a previous investigation in which eight H2O molecules per oxyethylene group was obtained.10,14 The thickness of the polyoxyethylene corona is 34 Å (58-24 Å), in good agreement with literature values.10,14 Cubic Region. 1H NMR Relaxation. The diffusion behavior of aggregates in the micellar region was shown above to be described more or less by polymer diffusion theory, although the hydrophobic core did contribute to the mechanism. Due to the very strong concentration dependence following as a result from the polymer-like diffusion, it was inherently difficult to evaluate the micelle shape. It is thus difficult to conclude if the cubic phase is likely to be discrete or bicontinuous. For this purpose, a small amount of a hydrophobic diffusion probe, HMDS, was added to the cubic phase. The 1H NMR signal was broadened significantly indicating that the cubic phase is of discrete type with but with nonspherical aggregates. The nonspherical symmetry of the interior of the aggregates makes the 1H NMR transverse relaxation time of HMDS very short and diffusion experiments virtually impossible to perform. The polyoxyethylene signal is not as strongly modulated by residual dipolar couplings following the nonspherical symmetry, most likely due to their high internal flexibility. Although this signal was relatively broad (approximately 300 Hz, which is much broader than that of a (13) Olsson, U.; Schurtenberger, P. Langmuir 1993, 9, 3389. (14) Jonstro¨mer, M.; Jo¨nsson, B.; Lindman, B. J. Phys. Chem. 1991, 95, 3293.

Figure 3. Diffusion coefficients for water, unreacted (free) fraction of the polyoxyethylene signal, and surfactant at (a) 18 °C and (b) 33 °C.

spherical micelle), the signal-to-noise ratio obtained in the a stimulated echo sequence was enough to perform accurate diffusion experiments. Water Diffusion. Figure 3 shows the measured diffusion coefficients for water, free polyoxyethylene, and surfactant as a function of concentration at 18 and 33 °C. As can be noted, the behavior of free polyoxyethylene is similar to that of water, an indication that the polyoxyethylene diffuses freely through the polar phase as expected. The diffusion of water is very slow, several times slower than what would be the case for hard sphere interaction with spherical or prolate-shaped aggregates. One obvious reason for the slow water diffusion is the exchange between free and bound water to the polyoxyethylene corona of the micelle. Since this exchange is fast compared to the experimentally accessible diffusion time, ∆ (which is 10-5000 ms), the observed water diffusion is the weighted average according to Dobs ) pboundDbound + pfreeDfree. Even though free water diffuses much faster, the fraction of bound water is so large that it needs to be considered. Another factor is the obstruction of water by the hydrocarbon part of the micelle. However, even though the surfactant concentration is high, the volume fraction of hydrocarbon is relatively low (