Formation and Characterization of Ordered Bicontinuous

Jul 13, 2009 - The techniques used included small angle X-ray scattering (SAXS), self-diffusion and quantum filtered NMR, differential scanning calori...
2 downloads 19 Views 3MB Size
Download PDF
J. Phys. Chem. B 2009, 113, 10669–10678

10669

Formation and Characterization of Ordered Bicontinuous Microemulsions Anna Kogan,† Deborah E. Shalev,‡ Uri Raviv,§ Abraham Aserin,† and Nissim Garti*,† Casali Institute of Applied Chemistry, The Institute of Chemistry, and Wolfson Centre for Applied Structural Biology, The Hebrew UniVersity of Jerusalem, 91904 Jerusalem, Israel ReceiVed: February 21, 2009; ReVised Manuscript ReceiVed: June 10, 2009

Ordered bicontinuous microstructures formed in a fully water-dilutable, pseudoternary unique nonionic microemulsion were obtained and characterized. The concentrate contained a mixture of triacetin/D-R-tocopherol acetate/ethanol/Tween 60. Upon dilution, the concentrate was transformed from a reversed micellar system to oil-in-water microemulsion droplets. The transformation occurred through an intermediate phase of ordered bicontinuous structures. The factors that governed the construction of this unique phase, and its physical and structural properties, were characterized in detail. The techniques used included small angle X-ray scattering (SAXS), self-diffusion and quantum filtered NMR, differential scanning calorimetry, rheology measurements, electrical conductivity, and dynamic light scattering. This mesophase displays microemulsion properties along with some characteristics of lyotropic liquid crystals (but is not a mixture of the two). Similar to microemulsions, the structures were transparent and spontaneously formed and exhibited thermodynamic stability. Yet, unlike microemulsions, they showed short-range order at room temperature. Additionally, the microstructures exhibited non-Newtonian flow behavior, characteristic of lamellar structures. The bicontinuous ordered microemulsions were obtained upon heating (to 25 °C) from the lamellar phase existing at low temperatures (5 °C). The main feature governing the bicontinuous mesophase formation was the amphiphilic nature of oil blends composed of D-R-tocopherol acetate and triacetin. The oils functioned as cosurfactants, altering the packing parameter of the surfactant and leading to the construction of bicontinuous structures with short-range order. These unique structures might have drug or nutraceutical delivery advantages. Introduction Microemulsions are optically isotropic and thermodynamically stable nanostructured mixtures of water, oil, and amphiphile(s).1 They frequently require cosolvents or cosurfactants to achieve very low interfacial tension and to facilitate proper packing parameters. Water-in-oil (W/O), bicontinuous, and oilin-water (O/W) are the most common structures.2 Bicontinuous microemulsion structures contain oil and water domains that are chaotically intertwined but are stabilized by sheetlike surfactant regions in the boundary zones between domains. These sheetlike regions are formed due to the tendency of the surfactant to localize between water-rich and oil-rich regions. Nonetheless, the long-range order remains stochastic.3 The absence of geometric order, verified by the lack of secondary scattering maxima in the small-angle X-ray scattering (SAXS) patterns, is what distinguished microemulsions from ordered isotropic liquid crystals3 until this work was done. Different models for bicontinuous microemulsion structures have been suggested: Scriven suggested that bicontinuous structures may be present in mesomorphic, liquid-crystal-like phases regarded as dispersions of spheres, cylinders, or lamellae.4 Talmon and Prager used a Voronoi theoretical model, where the bicontinuous structure was generated from many microscopic cells which are filled randomly with oil and water, while surfactant molecules are confined to the internal interfaces * Author to whom correspondence should be addressed at the Casali Institute of Applied Chemistry, E. Safra Campus, Givat Ram, The Hebrew University of Jerusalem, 91904 Jerusalem, Israel. Telephone: 972-2-6586574/5. Fax: 972-2-652-0262. E-mail: [email protected]. † Casali Institute of Applied Chemistry and The Institute of Chemistry. ‡ Wolfson Centre for Applied Structural Biology. § The Institute of Chemistry.

between oil and water.5 Saito and Shinoda described a bicontinuous microstructure in terms of a thermally coiled lamellar liquid crystal.6,7 Later, Kaler’s group also described bicontinuous structures based on the model of a disordered lamellar phase.8 It is generally agreed that bicontinuous microemulsions do not have the same long-range order as liquid crystals but rather exhibit a disordered or melted structure.9 Despite the existence of long-range order within the liquid crystalline phases, their application is limited due to their high viscosity. Consequently, low viscosity dispersions of liquid crystal phases in the aqueous phase are being explored (e.g., cubosome, hexosome, and liposome dispersions).10-12 Dilution with water lowers the viscosity but also decreases the solubilization capacity and the stability.10-12 There is as yet no experimental evidence for formulation of a bicontinuous mesophase that will combine the advantages of both microemulsions and liquid crystalline phases. Therefore, in the current work we aimed at selecting the proper ingredients that will allow us to obtain such structures and to explore their physical properties. In our studies, a complex mixture of pharmaceutically permitted components was used to form fully dilutable nonionic microemulsions. The oil phase consisted of triacetin (TA), D-Rtocopherol acetate (TocAc), and ethanol (EtOH) and as a surfactant, Tween 60 (T60) was used. The aqueous phase was comprised only of water. Under in ViVo conditions a microemulsion is projected to infinite dilution by body fluids. The concentrating effect is observed when microemulsions are applied topically (due to the evaporation of water). Thus, we determined the behavior of the system upon dilution with water, unlike many studies that neglect to investigate the effect of dilution on the structure and on the stability of microemulsions.13-15

10.1021/jp901617g CCC: $40.75  2009 American Chemical Society Published on Web 07/13/2009

10670

J. Phys. Chem. B, Vol. 113, No. 31, 2009

Figure 1. Phase diagram of a pseudoternary system {triacetin/D-Rtocopherol acetate/ethanol}(3/1/4)/Tween 60/water at 25 °C. Dilution line 64 is shown.

The microemulsion microstructure was explored along dilution line 64 (60 wt % surfactant phase and 40 wt % oil phase) in a pseudoternary phase diagram, based on the five-component system (Figure 1) using SAXS, rheology, differential scanning calorimetry (DSC), NMR (self-diffusion and 2H NMR), electrical conductivity (EC), and dynamic light scattering (DLS). In the course of investigating structural inversions occurring upon dilution, we explored microemulsion structures that displayed interesting and unique structural characteristics. These structures are bicontinuous microemulsions based on lamellar phases that have lost the long-range order due to thermal distortions but did preserve the short-range order. The focus of this paper is the structural characterization of these assemblies. We termed this phase an “ordered bicontinuous structure”. The characterization of this region in the phase diagram is interesting from both scientific and applicative aspects. It has both the physicochemical advantages of microemulsions over other formulations, including their spontaneous formation, transparency, low viscosity, and thermal and thermodynamic stability,16-20 and the geometric periodicity exhibiting the short-range order. These properties are of great benefit for purposes of controlled release. The phase can also outstandingly solubilize drug molecules that are sparingly water-soluble.9 Materials and Methods Materials. Tween 40 [polyoxyethylene (20) sorbitan monopalmitate], Tween 60 [polyoxyethylene (20) sorbitan monostearate] (T60), and Tween 80 [polyoxyethylene (20) sorbitan monooleate] were purchased from Sigma Chemical Co. (St. Louis, MO). Triacetin (TA) (99% purity) was purchased from Aldrich, Inc., (St. Louis, MO). D-R-Tocopherol acetate (TocAc) (96% purity) was obtained from ADM (Vista, CA), and ethanol (EtOH) (99.8%) was purchased from Frutarom (Haifa, Israel). The deuterium oxide (D2O) was purchased from Cambridge Isotope Laboratories Inc. (Andover, MA, USA). All components were used without further purification. Water was triple distilled. Small Angle X-ray Scattering (SAXS). Microemulsion samples along dilution line 64 (Figure 1), prepared as described above, were investigated by SAXS.17 Scattering experiments were performed using Ni-filtered Cu KR radiation (0.154 nm) from an Elliott GX6 rotating X-ray generator, operating at a power rating of up to 1.2 kW. X-ray radiation was further monochromated and collimated by a single Franks mirror and a series of slits and height limits, and measured by a linear position-sensitive detector. The sample was inserted into 1.5 mm quartz capillaries, which were then flame-sealed. Each sample was checked before and after the experiment to verify that no fluid had been lost during the time of exposure (approximately 1 h). The temperature was maintained at 5, 10, 15, 20, or 25 ( 1 °C accordingly. The sample-to-detector

Kogan et al. distance was 0.46 m, and the scattering patterns were measured using the Lake procedure.21 The experiments were performed in triplicate. X-ray data analysis was performed as described by Ezrahi et al.2 The average domain size, L, was determined by using Warren’s approximation,22 in which a finite lattice of linear dimension L, very close to a reciprocal lattice vector G, yielded 2 2 a structure factor proportional to e-|q-G| L /4π. The number of microemulsion nanostructures in a cross section per system, N, was computed by dividing the domain size, L, by the nanostructure-nanostructure correlation length, d)2π/G, where G is the center of the (0,0,1) peak. SAXS experiments were also performed at Beamline X33 at a European Molecular Biology Laboratory (EMBL) synchrotron facility in Hamburg, Germany. The scattering was done at 8.98 keV with a beam size of 0.2 mm × 0.2 mm and sample-todetector distances of 2.55 m, calibrated using silver behenate as a standard. Samples were inserted into 1.5 mm flame-sealed quartz capillaries. The samples were not oriented; thus, SAXS scans were collected on a 2D MAR345 detector, exhibited a powder pattern, and were radially averaged. Intensity as a function of momentum transfer, q, was plotted. Samples were scanned for 5 min, during which time no sample damage was detected. Sample damage was checked by comparing to multiple shorter scans taken at the same spot. Light Microscopy. The samples were inserted between two glass microscope slides and observed with a Nikon light microscope (Optiphot, Tokyo, Japan), equipped with crosspolarizers and attached to a video camera and monitor. The samples were analyzed at room temperature. Rheology Measurements. Viscosity measurements were performed at 25 °C on samples containing 0-20 wt % and 65-95 wt % of water. The measurements were made on the Rheoscope 1 rheometer (Thermo-Haake, Karlsruhe, Germany). A cone-plate sensor was used with a diameter of 60 mm, a cone angle of 1°, and a gap of 0.022 mm. Shear rates were measured between 0 and 120 s-1. Rheology measurements were conducted on samples containing 25-60 wt % of water, using the above-mentioned Rheoscope. A cone-plate with a diameter of 35 mm and an angle of 1° was used. Temperature was maintained at 5, 10, 15, 20 and 25 ( 0.1 °C during measurements. Shear rate measurements were performed between 0.01 and 100 s-1. Dynamic viscoelastic experiments consisted of preliminary tests of oscillation stress sweep to determine the linear viscoelastic region. The storage modulus (G′) and loss modulus (G′′) were measured as function of frequency (ω) in the range 0.01-100 rad s-1. The measurements were done in triplicate and were found to be reproducible. Pulsed-Gradient-Spin-Echo NMR (PGSE-NMR). The self-diffusion coefficients (DC) were determined using pulsed gradient stimulated spin echo NMR.17 The measurements were made on samples all along dilution line 64 at 25 °C and samples containing 35-65 wt % of water, in the temperature range of 5-25 °C, on a Bruker DRX-400 spectrometer, with a BGU II gradient amplifier unit and a 5-mm BBI probe equipped with a z-gradient coil, providing a z-gradient strength (g) of up to 55 G cm-1. Bipolar gradient pulses were used, as described by Wu et al.,23 to reduce the eddy-current effects. Experiments were carried out by varying g and keeping all other timing parameters constant. 2 H NMR Experiments. Anisotropy was detected, using NMR experiments, on samples prepared with 5 wt % D2O as part of the water fraction of the sample. NMR experiments were performed on a Bruker DRX 400 MHz spectrometer, operating

Ordered Bicontinuous Microemulsions

J. Phys. Chem. B, Vol. 113, No. 31, 2009 10671

Figure 2. Experimental SAXS intensities measured (A) in house or via (B) synchrotron scans in microemulsions consisting of TA/TocAc/EtOH (3/1/4), T60, and water along dilution line 64 at 25 °C. In the brackets the concentration of water is indicated. The curves were shifted for the sake of clarity.

at a proton frequency of 400.13 MHz and a deuterium frequency of 61.42 MHz, using a 5 mm broadband probe. Temperature measurements were performed at -10, -5, 0, 5, 10, and 15 ( 0.1 °C. Anisotropic 2H signals with small coupling values from the deuterium oxide were separated from the strong isotropic signal that masked them by using an in-phase double quantum filtered pulse sequence with refocusing.24 The spectra were acquired for creation times ranging from 20 µS to 9 mS, using 32 transients, and were processed with zero filling of 4K and exponential apodization with line broadening of 20 Hz. Differential Scanning Calorimetry (DSC). For calorimetric measurements, a Mettler Toledo DSC822 (Greifensee, Switzerland) measuring model system was used. The instrument was calibrated every two weeks with indium, lauric acid, water, and ethyl acetate to ensure the accuracy of the caloric data. The heating rate of calibration was 10 K min-1. The DSC measurements were carried out as follows: 8-10 mg microemulsion samples were weighed, using a Mettler M3 microbalance, in standard 40 µL aluminum pans and immediately sealed by a press.25,26 Peaks representing various states of water were analyzed. The samples were cooled in liquid nitrogen from 30 to -40 °C at 10 °C/min. Each sample remained at this temperature for 30 min and was then heated at a rate of 1 °C/min to 50 °C. An unloaded pan was used as a reference. The instrument determined the fusion temperatures of the solid components and the total heat transferred in any of the observed thermal processes. The enthalpy changes associated with the thermal transition were obtained by integrating the area of each pertinent DSC peak. DSC temperatures reported here were reproducible within ( 0.5 °C. We followed the method used by Senatra et al.27 to identify various states of water in our systems, as described elsewhere.25,26 Electrical Conductivity (EC). EC measurements were performed at 25 ( 0.2 °C on samples along dilution line 64 (Figure 1) using a conductivity meter, type Oyster, Conductivity/ Temperature meter (Extech Instruments, Waltham). Since the microemulsions were based on nonionic components, a small quantity of aqueous electrolyte (a solution of 0.01 M NaCl) was added. The samples remained clear and transparent, and there were no observable changes in the phase diagram. Similar independence of phase behavior, in the presence of a small quantity of electrolyte, has been reported in the literature.28

TABLE 1: Periodicity (d) and Correlation Length (ξ) Calculated from the SAXS Measurements at Dilution Line 64, at 25 °C upon Dilution with Water

a

water contents (wt %)

d (Å)a

ξ (Å)a

10 15 20 65 70 75 80

51 58 63 51 53 55 62

25 31 49 19 23 25 25

) ( 0.5 Å.

Dynamic Light Scattering (DLS). The hydrodynamic diameter of microemulsion nanodroplets, containing (95-98 wt %) water, was determined using the Nano-S instrument (λ ) 633 nm; Malvern Instruments, Malvern, U.K.). The samples were filtered with a 0.2 mm PVDF filter (Millipore) into a polystyrene disposable cuvette, with no further dilution. The measurements were carried out at a scattering angle of 173° at 25 °C. Data were collected in three repeated measurements (10 scans for each repeat). The average size was reported by the volume distribution for each measurement. Results and Discussion SAXS. SAXS diffractions of the mixtures, at 25 °C along dilution line 64 in the pseudoternary phase diagram (Figure 1), were performed. As displayed in Figure 2A, a broad diffraction peak was recorded at low water contents, 10-20 wt %, which is typical for the micellar system. Table 1 shows the dependence of the periodicity, d, and correlation length (persistence length), ξ, as a function of water content, which were calculated after fitting the scattering patterns to the Teubner and Strey equation.29 Periodicity increased by 20%, while correlation length, reflecting the degree of order in the microemulsion, increased by 50% in the 10-20 wt % range of dilution. This may be attributed to the swelling of the reverse structures upon dilution, thereby inducing the formation of a more ordered phase. Swelling of structures upon dilution was also reported in other studies.17,30 At higher water contents, 25-60 wt %, the peak progressively sharpened and shifted toward lower q values (Figure 2A). The shift of the first peak indicated formation of larger structural

10672

J. Phys. Chem. B, Vol. 113, No. 31, 2009

Kogan et al.

Figure 3. Experimental SAXS intensities in microemulsions consisting of (A) TA/TocAc/EtOH (3/1/4), T60, and 40 wt % of water in the temperature range 5-25 °C (specific temperatures are indicated in the brackets); (B) TA/TocAc/EtOH (3/1/4), T60, and 40 wt % of water along dilution lines 64, 55, and 46 (indicated in the brackets) at the temperature 25 °C; and (C) TA/TocAc/EtOH in 3/1/4, 2/1/3, and 1/1/2 wt % ratios (indicated in the brackets), T60, and 40 wt % of water at the temperature 25 °C. The curves were shifted for the sake of clarity.

dimensions. Additionally, above 25 wt %, another peak appeared at higher q values. For example, at 35 wt % water, a major sharp scattering peak appeared at q ) 0.085 Å-1 followed by a broader, less pronounced peak at q ) 0.145 Å-1. It should be noted that the sharp shape of the first peak, along with the appearance of the second one, is uncommon for a typical microemulsion. These samples were examined by a polarized light microscope and did not reveal birefringent or colorful images, which indicates that no liquid crystalline phase is present at room temperature. Diluting the system with >45 wt % water caused a gradual decrease in the intensity of the first peak until its complete disappearance at 65 wt %. Eventually, a single broad peak was recorded at 65-80 wt % water and was associated with O/W microemulsion.31 The uncommon scattering profiles in the intermediate regime of water were further investigated by high resolution SAXS synchrotron scans (Figure 2B). Synchrotron measurements of samples containing 35-50 wt % water reinforced the previously gathered data, exhibiting one high intensity sharp peak followed by a lower intensity secondary one. The origin of the unusual scattering patterns was studied by temperature-dependent measurements (5-25 °C). Figure 3A shows an example of temperature-dependent SAXS profiles for the formulation containing 40 wt % water. At a low temperature range (5-10 °C), two sharp peaks with a 1:2 ratio of d-space values were recorded, confirming the existence of lamellar symmetry. At increased temperature (15-25 °C), the intensity of the second peak relative to the first peak dropped and eventually merged into a shoulder of the first peak (Figure 3A). This corresponds to a gradual decrease of the degree of order, leading to a distortion of the lamellar symmetry. It may be deduced that within the region containing intermediate amounts of water, the order is gained by a temperature decrease. We are suggesting that the structures obtained at room temperature in the range of 25-65 wt % water originate from a lamellar phase. These structures can be classified as ordered bicontinuous microemulsions derived from disordering the lamellar structure, as suggested by Kaler and co-workers8 in their pioneering work. The authors proposed that bicontinuous microemulsions, based on disordered lamellar and lamellar structures, can exist in close proximity regions in the phase diagram.8 In our case, the increase in temperature induced a distortion of the lamellar phase and induced the formation of bicontinuous structures, exhibiting a lower degree of order than the lamellar phase but higher compared to a typical microemulsion.

The internal order of these structures was manipulated by changing the chemical composition of the mixtures, as follows: (1) replacing Tween 60 (stearic acid tail) with Tween 40 and Tween 80 (palmitic and oleic acid tails, respectively); (2) altering the dilution lines from 6/4 to 5/5 and 4/6 wt % ratios of surfactant/oil, respectively, meaning the use of excess of an oil phase over surfactant; and (3) decreasing the TA to TocAc wt % ratios, while keeping the EtOH content constant within the oil phase composition (from 3/1/4 to 2/1/3 and 1/1/2 wt % ratios of TA/TocAc/EtOH, respectively). The scattering patterns of the bicontinuous mixtures containing surfactant molecules with different lipophilic tails did not show significant changes (data not shown). Moderate changes were observed when oil phase ratios, as well as dilution lines, were altered. In general, increasing the TA content (1.33-1.5 times) at the expense of the more hydrophobic TocAc molecule, and decreasing dilution lines (i.e., decreasing oil to surfactant ratios) decreases the d-spacings (Figure 3B and C). It may be deduced that the nature of the surfactant tail does not significantly affect these structures, while changing the oil phase content influences the structural configuration of the phase. To gain more information on the degree of order within the structures, we applied the Warren approximation.22 The domain size and the number of domains (L and N parameters) showed that between 10 and 50 wt % water content, the domain sizes and their number gradually increase, as expected, due to the swelling of the structures. Above 50 wt % water, the values of both parameters decrease, since direct microemulsion structures (L1) are formed (Figure 4). Characteristic profiles31 were observed for microemulsion systems containing higher water contents (65-80 wt %). The scattering patterns (values of q) for the 65-80 wt % region, after background subtraction, were fitted to the Teubner and Strey equation29 to determine the periodicity (d) and the correlation length (ξ). The dependence of the periodicity, d, and correlation length, ξ, as a function of water content in the studied systems, are shown in Table 1. The periodicity increased by 22%, while the correlation length increased by 24% in the 65-80 wt % water content range, presumably due to dilution and formation of the direct microemulsion structures. These results seem to present the first evidence of the continuous transformation of more ordered bicontinuous structures, originating from lamellar structures, to O/W droplets. In Figure 5 we presented a schematic illustration of a bicontinuous O/W microemulsion (left) and a lamellar phase (right). Both

Ordered Bicontinuous Microemulsions

J. Phys. Chem. B, Vol. 113, No. 31, 2009 10673 TABLE 2: Measured Viscosity in the System Consisting of TA/TocAc/EtOH (3/1/4), T60, and Water along Dilution Line 64 at 25 °C (n ) 3)

Figure 4. Domain size (correlation length) (L (0)) and the number of domains (N (9)) calculated from the SAXS measurements in microemulsions consisting of TA/TocAc/EtOH (3/1/4), T60, and water along dilution line 64 at 25 °C. The connecting lines are for ease of visualization. The bar represents n ) 3.

water content (wt %)

viscosity × 103 (Pa · s × 103)

0 5 10 15 20 65 70 75 80 85 90 95

39.3 ( 1.9 50.3 ( 1.1 56.4 ( 1.7 54.9 ( 2.8 71.3 ( 4.3 8.8 ( 0.5 5.1 ( 0.12 3.4 ( 0.10 2.4 ( 0.07 1.6 ( 0.05 0.9 ( 0.05 0.9 ( 0.03

103 Pa · s, corresponding to the viscosity of aqueous solution. The SAXS scattering profiles confirm that at high aqueous dilutions, >60 wt % water, the structures are spherical droplets of O/W microemulsions. However, in the intermediate water region (25-60 wt %) the samples exhibited somewhat different behavior. Figure 6A shows the viscosity and shear stress as a function of shear rate of a sample containing 40 wt % water. From the flow curve one can clearly observe several distinct nonhomogeneous flow

Figure 5. Schematic illustration of a bicontinuous microemulsion (left) that upon decrease of temperature gains long-range order and turns to a lamellar phase (right).The yellow and the blue channels represent the surfactant/alcohol/oil layers and water layers, respectively. The lamellae extend above and below the plane of the page. (Inset) Cross section of the illustrated above structures.

structures have a layered geometry with elongated channels that consist of a surfactant, alcohol, and oil mixture (yellow) and a continuous water phase (blue). However, the lamellar structures have a long-range order which is lost upon heating within the bicontinuous structures. Viscosity and Rheology. Viscosity and rheology measurements were performed at 25 °C along dilution line 64 in the pseudoternary phase diagram (Figure 1). The samples, containing low (60 wt %) content of water, exhibited a Newtonian flow behavior (shear viscosity is independent of shear rate). The obtained viscosity values are summarized in Table 2. One can see that viscosity, at the low water contents, increases with dilution. It may be related to the swelling process of the reverse structures until bicontinuous (at 25 wt %) structures are formed. The results are in good correlation with SAXS studies exhibiting a broad scattering peak in this region, which is shifted toward lower q values in this region, indicating an increase in the domain size. Dilution to higher water contents (>60 wt %) leads to a decrease in the viscosity of the system, since the structures transform into direct O/W nanodroplets. The viscosity of a very dilute system (95 wt %) reaches a value of (0.90 ( 0.03) ×

Figure 6. (A) Flow curve showing the evolutions of the shear stress (2) and the dynamic viscosity (9) versus the shear rate for the system consisting of TA/TocAc/EtOH (3/1/4), T60, and 40 wt % of water along dilution line 64 at 25 °C. (B) Dependence on frequency of the storage modulus G′ and the loss modulus G′′ of the system consisting of TA/ TocAc/EtOH (3/1/4), T60, and 40 wt % of water along dilution line 64 at 5, 10, 15, and 25 °C.

10674

J. Phys. Chem. B, Vol. 113, No. 31, 2009

regions, demonstrating the non-Newtonian flow behavior. This indicates that the ordered bicontinuous phase is a pseudoplastic fluid with a distinct yield point (Figure 6A). Such behavior indicates the existence of an intermicellar network formation held together by some intermicellar structural forces. Upon applying >10 Pa, which is sufficiently strong to overcome the network forces, the mesophase becomes progressively deformed. At high shear rates the system becomes Newtonian, meaning that it is losing its symmetrical configuration. One can also see that the ordered bicontinuous phase exhibits shear thinning behavior, since the viscosity drops from 5.5 Pa · s at shear rates around 1 s-1 to about 0.6 Pa · s around 100 s-1 (Figure 6A). Figure 6B demonstrates the oscillatory measurements of the sample containing 40 wt % of water as a function of temperature. The examined samples exhibit a viscoelastic gel-like behavior.33 The yield stress below which the structures do not flow decreases with the increase of temperature. These results are consistent with SAXS and NMR measurements, indicating formation of a lamellar phase upon a temperature decrease. The examined structures are more elastic than viscous, that is, G′ > G′′ in the measured frequency range (Figure 6B). Similar behavior was observed by other authors examining lamellar phases.32 It was seen that, as the frequency increases, the storage modulus G′ increases faster with slopes 4, 1, 0.6, and 0.3 than the loss modulus G′′ with slopes 0.6, 0.2, 0.3, and 0.003, within temperatures of 5, 10, 15, and 25 °C accordingly. This indicates that the stored energy in the structure prevails over the energy that was dissipated by the viscous forces. It should also be noted that the increase in temperature leads to a decrease in viscoelastic properties (the values of the storage modulus G′ and the loss modulus G′′ are reduced). This is in good correlation with the SAXS results at low temperatures (5-15 °C), revealing a long-range order of the lamellar phase, while at room temperature the phase exhibits only the shortrange order. Furthermore, the increase in temperature also decreases the range of ω over which the system exhibits an elastic modulus. Thus, it indicates the existence of an ordered bicontinuous phase which is less elastic than viscous. It is generally believed that lamellar-phases of nonionic surfactants are stabilized by the long-range Helfrich repulsion forces resulting from thermal undulations.34 The increase in temperature modifies the curvature rigidity of each membrane, thus changing the nature of fluctuations that drive the steric repulsions to change the lamellar phases’ structure.34 PGSE-NMR. Pulsed gradient spin echo self-diffusion nuclear magnetic resonance (PGSE-SD-NMR) is an important technique for providing data on the changes in the environment of the components of the studied system, by determining the molecular self-diffusion coefficients (DCs) of each component. Extensive studies of DCs measured by SD-NMR have been done by many researchers.17,31,35,36 Fast diffusion (>10-9 m2 s-1) is characteristic of free molecules in solution, while a small DC (10 wt %) induced endothermic events (peaks A and B), related to the formation of bound (which is associated to hydrophilic groups and melts below -10 °C), interphasal water (defined as water confined within the interface of the dispersed system, which melts at about -10 °C) or free water which melts at ∼0 °C.26 To verify that the discussed events are related to water, we replaced the water with D2O, leaving all other components the same, which resulted in a characteristic shift to higher temperatures. The fusion temperatures of water molecules are depicted in Figure 8. It can be seen that fusion temperatures gradually increase upon dilution and exhibit a maximum at 25 wt % in both types of water (Figure 8). This indicates that at this point some structural changes occur within the system and more water molecules become strongly bound to the surfactant. These structural modifications also resulted in higher melting temperatures and enthalpies. To evaluate the degree of binding strength, varied upon the formation of the ordered bicontinuous structures, one should compare the fusion temperatures and enthalpies of water within samples containing 20 (W/O microemulsion) and 25 wt % (ordered bicontinuous microemulsion) (Table 3). The fusion temperature of a W/O microemulsion is -18.68 ( 0.2 °C for bound and -10.51 ( 0.2 °C for interphasal water molecules with enthalpies of -1.22 J g-1 and -1.19 J g-1, respectively. The ordered bicontinuous microemulsion exhibits a fusion temperature of -14.97 ( 0.2 °C for bound and -6.39 ( 0.2 °C for interphasal water molecules with enthalpies of -3.06 J g-1 and -1.58 J g-1, respectively. Calculating the difference in enthalpy increase, one gets a significant rise of 150% for the bound water and 33% for the interphasal water molecules. It is reasonable to conclude that the formation of bicontinuous structures results in the stronger binding of water molecules, which inhibits surfactant head melting. However, the formation of more ordered structures also affects the surfactant’s tail, which is pronounced as an exothermic event, Peak C, occurring at 0-20 wt % water contents (Table 3). Ezrahi et al.26 explained this phenomenon as an exothermic crystallization of the surfactant hydrocarbon (fatty) tails into a more structured crystalline organization. Table 3 shows that the temperature (Tcrys) of the exothermic peak (peak C) shifts to more positive values as the aqueous phase content increases. A possible interpretation is that upon dilution the structures swell and the curvature decreases. At higher water contents, the reorientation of surfactant molecules in a crystalline-like structure is more restricted so that higher temperatures are needed (an increase

0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95

bound water (wt %)

interfacial water (wt %)

0.1 0.4 0.4 1.0 3.4 6.5 10.9

0.4 0.4 0.4 0.5 0.5 1.1 0.4 15.8 23.3 27.3 30.8 35.3 38.2 58.5 66.2

free water (wt %)

nonfreezable water (wt %)

71.4 81.4 94.8

9.4 14.2 19.2 23.5 26.1 27.3 28.7 29.2 26.7 27.7 29.2 29.7 31.8 16.5 13.8 13.6 8.6 0.2

in 4 °C upon dilution from 0 wt % to 20 wt % water). The results discussed above are in good correlation with SAXS measurements, where, at 25 wt % water, formation of more ordered bicontinuous structures was detected, showing viscoelastic properties as revealed also by rheology studies. It should also be noted that at low water contents (10-40 wt %) the melting peak temperature of water (peak A) is very low (∼-29-16 °C), presumably due to the presence of alcohol in the system (Table 3). In previous studies, we have shown26 that the effect of alcohol becomes more pronounced with short-chain alcohols. The shorter the alcohol chain, the lower the melting peak temperature of the bound water. Therefore, it is reasonable to believe that bound water, which melts at ∼-29-16 °C, was formed as a consequence of a concerted association of EtOH, T60, TocAc, and water (Table 3). The variation in the contents of different types of water, as determined from Senatra et al.’s equations,27 is presented in Table 4. Since the amount of such water binding molecules (EtOH, T60, and TocAc) decreases upon dilution, less water molecules are bound; hence, at 40 wt % water, all the water molecules become interphasal (Figure 8, Table 4). These results are in agreement with SD-NMR studies, which demonstrate that EtOH and TocAc are cosurfactants, thus playing an important role in stabilizing the curvature of the T60 and forming more ordered bicontinuous structures. One can see that the contents of bound water vary between 0.4 and 1.1 wt %, while interphasal water contents vary between 0.1 and 66 wt %. From calculations based on ∆Hf and the quantities of water added, it is clear that there is a significant amount of water that is unaccounted for and that was previously termed “nonfreezable” water.27 The amount of the nonfreezable water remained quite considerable along most steps of dilution (9-32%) but decreased significantly (0.2%) when 95 wt % water was added. Since the present microemulsion systems are rich in water-binding, functional molecules such as EtOH, T60, and TocAc, water was strongly bound to any or all of these molecules throughout all the dilution stages. This indicates that the adduct “water-surfactant-cosurfactant” does not freeze and, thus, does not thaw. It was reported that the existence of nonfreezable water in ethoxylated alcohol-based systems de-

Ordered Bicontinuous Microemulsions

J. Phys. Chem. B, Vol. 113, No. 31, 2009 10677

Figure 9. Variation of the electrical conductivity (σ) of the system consisting of TA/TocAc/EtOH (3/1/4), T60, and water, along dilution line 64 at 25 °C. The connecting lines are for ease of visualization. The dashed lines along the curve are for the W/O, bicontinuous, and O/W regions, respectively. The bar represents n ) 2.

pends on the length of their hydrophilic headgroup.26,47 Schulz and Puig47detected nonfreezable water in the nonionic binary water-C12(EO)23 and water-C16(EO)20 systems. They found that 33.70 molecules per surfactant molecule, corresponding to 1.40 molecules of water per oxygen atom, were undetectable by DSC. Some water was assumed to be trapped in a restricted area of the helicoidal structure in the hydrophilic chain. Thus, the water-surfactant interaction might be strong enough to prevent the water from freezing. Such strong interactions between the water molecules and the components of the microemulsion might have contributed to the formation of the distinctive structures which were obtained in this study. At the region containing 10-80 wt % water, the contents of the interphasal water increased upon dilution (Figure 8, Table 4). The surfactant binds water molecules within the ethoxylated headgroup region. As more water molecules are added, they replace the interfacial EtOH molecules, which move to the aqueous continuous phase, since they are no longer needed for curvature stabilization. This is in accordance with the SD-NMR studies, which show that, upon dilution, the mobility of EtOH molecules is increased. Other NMR studies of lipids have shown that the motion of the surfactant headgroups is enhanced by progressive hydration. This process was entropy-driven and led to a progressive loosening-up of the headgroup packing.48 The rejection of interfacial EtOH molecules might be influenced by this process, thus inducing the transformation of the more ordered bicontinuous phase into O/W structures, occurring at 50 wt % water, where another maximum in the fusion temperature of interphasal water is observed. Interestingly, SAXS results designate the same water contents for structural modification. In high-dilution regions of 80-95 wt % water, the detected water was mostly free (Figure 8), indicating that a conversion to O/W structures occurred. This correlates the SAXS results where a scattering pattern, characterizing a microemulsion, appeared. It is also in good agreement with SD-NMR results, that show high DCs of water molecules, and with viscosity studies where Newtonian behavior was observed at these water contents. Electrical Conductivity. Electrical conductivity (σ) can provide a good indication of microstructure transitions occurring in microemulsions, i.e. transformation from water-entrapped systems to intermediate structures and further to water continu-

ous microstructures. When the continuous phase is aqueous, it exhibits higher conductance than the oil-continuous formulation.49 Electrical conductivity was measured as a function of water content for the microemulsion along dilution line 64. The results of σ variation as a function of water content are shown in Figure 9. The conductivity increases exponentially as the water content increases. These changes were attributed to the occurrence of a percolation transition.50-53 In this percolation model, the conductivity remains low up to a certain content of water. The electrical conductivity profile shows that up to 25 wt % of water conductivity is initially low (the slope is 0.24), suggesting the existence of reverse structures in a nonconducting oil medium, which have little interaction with each other. When more than 25 wt % of water is added, the conductive droplets begin to contact each other and form other structures, giving rise to the observed changes in properties, such as an increase in electrical conductivity. Interestingly, at 25 wt % water content, SAXS measurements (Figure 2) indicated a formation of more ordered structures. Additionally, DSC (Figure 8) results also showed some structural changes at 25 wt % water. In the region 25-65 wt % of water, the increase in the electrical conductivity was not linear. It may be deduced that bicontinuous structures are formed. When more than 65 wt % of water is added, the electrical conductivity increases linearly but more steeply (slope is 2.55), indicating the formation of O/W droplets. This is in line with the observation made in the SAXS study, where a single peak appeared, indicating the formation of direct microemulsion structures. DLS. DLS droplet size distribution measurements were done only on very dilute microemulsions (95-98 wt % aqueous phase), since only at such dilutions can the results be reliable. All the droplets were within the range of 10 ( 1 nm. Conclusions In the current study, we investigated the structural transformations occurring within a specific and specially designed fivecomponent system upon continuous dilution with water. It was demonstrated that at the intermediate regime of water (25-65 wt %) an ordered bicontinuous phase, with unique physicochemical properties, was formed. Gradual development of the phase, upon water dilution, was monitored and characterized by SAXS, rheology, NMR (self-diffusion and 2H NMR), DSC, and electrical conductivity measurements.

10678

J. Phys. Chem. B, Vol. 113, No. 31, 2009

The presence of some geometric order, which is lacking in the ordinary microemulsions, was deduced from the secondary SAXS reflection patterns. However, the degree of order was not sufficient to claim the existence of any known liquid crystal symmetry. The mesophase, spontaneously formed, transparent, and thermodynamically stable, led us to term the phase “an ordered bicontinuous microemulsion”. The amphiphilic nature of the oils blend (D-R-tocopherol acetate and triacetin) functions as a cosurfactant, altering the packing parameter of the surfactant that leads to the construction of these bicontinuous structures, as deduced from NMR measurements. The formation of a more ordered phase was also detected by DSC. Upon increasing water contents from 20 to 25 wt %, a sharp increase in fusion temperature (∼4 °C for both types of water) and enthalpy (150% and 33%) was observed for bound and interphasal water molecules, respectively. In addition, in the 0-20 wt % region, the temperature of the exothermic peak related to the surfactant crystallization shifted to more positive values. This indicates the reorientation of surfactant molecules into a more crystalline structure, where their mobility is more restricted so that higher temperatures were needed. Electrical conductivity exhibited a sharp increase once 25 wt % water was added, indicating the formation of conductive channels. This distinctive phase seems to be formed from the lamellar structures, which are distorted upon heating but preserve some short-range order of the aggregates. Due to the internal order within these structures, some viscoelastic properties were revealed. It was found that these bicontinuous structures are more elastic than viscous. In addition, these properties are more pronounced at low temperatures. To conclude, the structures obtained at the intermediate region of the dilution line originating from the lamellar phase exhibited higher degrees of internal order, relative to the ordinary microemulsions. These structures are distinctive and advantageous in their thermodynamic stability, clear appearance, viscoelastic properties, and short-range order. Controlling these properties within a surfactant-based system is the key determinant in controlling the kinetics of active molecule release. Acknowledgment. The EMBL synchrotron facility at Hamburg, beamline X33, where some part of this work was done, is kindly acknowledged. We would like to thank Zehava Cohen for graphical support and Dr. Uzi Eliav for helpful conversations on anisotropy in NMR and for providing us with the NMR pulse program. References and Notes (1) Strey, R. Colloid Polym. Sci. 1994, 272, 1005–1019. (2) Ezrahi, S.; Wachtel, E.; Aserin, A.; Garti, N. J. Colloid Interface Sci. 1997, 191, 277–290. (3) Kaler, E. W.; Bennett, K. E.; Davis, H. T.; Scriven, L. E. J. Chem. Phys. 1983, 79, 5673–5684. (4) Scriven, L. E. Nature 1976, 263, 123–124. (5) Talmon, Y.; Prager, S. J. Chem. Phys. 1978, 69, 2984–2991. (6) Saito, H.; Shinoda, K. J. Colloid Interface Sci. 1970, 32, 647–651. (7) Shinoda, K. Prog. Colloid Polym. Sci. 1983, 68, 1–7. (8) Vonk, C. G.; Billman, J. F.; Kaler, E. W. J. Chem. Phys. 1988, 88, 3970–3975. (9) Sjoblom, J.; Lindberg, R.; Friberg, S. E. AdV. Colloid Interface Sci. 1996, 95, 125–287. (10) Rangelov, S.; Almgren, M. J. Phys. Chem. B 2005, 109, 3921– 3929. (11) Dong, Y. D.; Larson, I.; Hanley, T.; Boyd, B. J. Langmuir 2006, 22, 9512–9518. (12) Almgren, M.; Borne´, J.; Feitosa, E.; Khan, A.; Lindman, B. Langmuir 2007, 23, 2768–2777.

Kogan et al. (13) Kawakami, K.; Yoshikawa, T.; Moroto, Y.; Kanaoka, E.; Takahashi, K.; Nishihara, Y.; Masuda, K. J. Controlled Release 2002, 81, 65–74. (14) Kawakami, K.; Yoshikawa, T.; Hayashi, T.; Nishihara, Y.; Masuda, K. J. Controlled Release 2002, 81, 75–82. (15) Kim, S. K.; Lee, E. H.; Vaishali, B.; Lee, S.; Lee, Y. K.; Kim, C. Y.; Moon, H. T.; Byun, Y. J. Controlled Release 2005, 105, 32–42. (16) Kogan, A.; Garti, N. AdV. Colloid Interface Sci. 2006, 123-126, 369–385. (17) Kogan, A.; Aserin, A.; Garti, N. J. Colloid Interface Sci. 2007, 315, 637–647. (18) Teo, B. M.; Ashokkumar, M.; Grieser, F. J. Phys. Chem. B. 2008, 112, 5265–5267. (19) Mcllwaine, R. E.; Fenton, H.; Scott, S. K.; Taylor, A. F. J. Phys. Chem. C. 2008, 112, 2499–2505. (20) Kogan, A.; Rozner, S.; Mehta, S.; Somasundaran, P.; Aserin, A.; Garti, N.; Ottaviani, M. F. J. Phys. Chem. 2009, 113, 691–699. (21) Lake, J. A. Acta Crystallogr. 1967, 23, 191–194. (22) Warren, B. E. Phys. ReV. 1941, 59, 693–698. (23) Wu, D.; Chen, A.; Johnson, C. S. J. Magn. Reson. 1995, 115, 260– 264. (24) Eliav, U.; Navon, G. J. Magn. Reson. 1999, 137, 295–310. (25) Yaghmur, A.; Aserin, A.; Tiunova, I.; Garti, N. J. Therm. Anal. Calorim. 2002, 69, 163–177. (26) Ezrahi, S.; Aserin, A.; Fanun, M.; Garti, N. Subzero Temperature Behavior of Water in Microemulsions. In Thermal BehaVior of Dispersed Systems; Surfactant Science Series; Garti, N., Ed.; Marcel Dekker, Inc.: New York, 2001; Vol. 93, pp 59-120. (27) Senatra, D.; Lendinara, L.; Giri, M. G. Prog. Colloid Polym. Sci. 1991, 84, 122–128. (28) Garti, N.; Aserin, A.; Wachtel, E.; Gans, O.; Shaul, Y. J. Colloid Interface Sci. 2001, 233, 286–294. (29) Teubner, M.; Strey, R. J. Chem. Phys. 1987, 87, 3195–3200. (30) Tomsˇicˇ, M.; Podlogar, F.; Gasˇperlin, M.; Besˇter-Rogacˇ, M.; Jamnik, A. Int. J. Pharm. 2006, 327, 170–177. (31) Regev, O.; Ezrahi, S.; Aserin, A.; Garti, N.; Wachtel, E.; Kaler, E. W.; Khan, A.; Talmon, Y. Langmuir 1996, 12, 668–674. (32) Camerel, F.; Gabriel, J. C. P.; Batail, P.; Panine, P.; Davidson, P. Langmuir 2003, 19, 10028–10035. (33) Warriner, H. E.; Idziak, S. H. J.; Slack, N. L.; Davidson, P.; Safinya, C. R. Science 1996, 271, 969–973. (34) Zou, A.; Eastoe, J.; Mutch, K.; Wyatt, P.; Scherf, G.; Glatter, O.; Grillo, I. J. Colloid Interface Sci. 2008, 322, 611–616. (35) Caboi, F.; Lazzari, P.; Pani, L.; Monduzzi, M. Chem. Phys. Lipids 2005, 135, 147–156. (36) Lopez, F.; Cinelli, G.; Ambrosonea, L.; Colafemmina, G.; Ceglie, A.; Palazzo, G. Colloids Surf., A 2004, 237, 49–59. (37) Graciaa, A.; Lachaise, J.; Cucuphat, C.; Salager, J. L. Langmuir 1993, 9, 669–672. (38) Kunieda, H.; Ozawa, K.; Huang, K. L. J. Phys. Chem. B 1998, 102, 831–838. (39) Israelachvili, J. N. Intermolecular and Surface Forces; Academic Press: London, 1992. (40) Rodriguez-Abreu, C.; Aramaki, K.; Tanaka, Y.; Lopez-Quintela, M. A.; Ishitobi, M.; Kunieda, H. J. Colloid Interface Sci. 2005, 291, 560– 569. (41) Acharya, D. P.; Kunieda, H. J. Phys. Chem. B 2003, 101, 10168– 10175. (42) Rowinski, P.; Rowinska, M.; Heller, A. Anal. Chem. 2008, 80, 1746–1755. (43) Wang, X.; Quinn, P. J. Biochimie 2006, 88, 1883–1888. (44) Coppola, L.; Muzzalupo, R.; Ranieri, G. A.; Terenzi, M. Langmuir 1995, 11, 1116–1121. (45) Amar-Yuli, I.; Wachtel, E.; Shalev, D. E.; Aserin, A.; Garti, N. J. Phys. Chem. B 2008, 112, 3971–3982. (46) Hoff, B.; Strandberg, E.; Ulrich, A. S.; Tieleman, D. P.; Posten, C. Biophys. J. 2005, 88, 1818–1827. (47) Schulz, P. C.; Puig, J. E. Colloids Surf., A 1993, 71, 83–90. (48) Ulrich, A. S.; Watts, A. Biophys. J. 1994, 66, 1441–1449. (49) Bennett, K. E.; Hatfield, J. C.; Davis, H. T.; Macosko, C. W.; Scriven, L. E. In Microemulsions; Robb, I. D., Ed.; Plenum Press: New York, 1982; Vol. 65, p 84. (50) Mehta, S. K.; Bala, K. Fluid Phase Equilib. 2000, 172, 197–209. (51) Djordjevic, L.; Primorac, M.; Stupar, M.; Krajisnik, D. Int. J. Pharm. 2004, 271, 11–19. (52) Caboi, F.; Capuzzi, G.; Baglioni, P.; Monduzzi, M. J. Phys. Chem. B 1997, 101, 10205–10212. (53) Ezrahi, S.; Tuval, E.; Aserin, A.; Garti, N. J. Colloid Interface Sci. 2005, 291, 263–272.

JP901617G