Microemulsion Droplets Decorated by Brij700 Block Copolymer

May 5, 2007 - For a surfactant-to-oil ratio of 2, the SANS data revealed a transition from ... Zhang , Zhonghua Wu , Guang Mo , Xueqing Xing , and Bux...
0 downloads 0 Views 214KB Size
6544

Langmuir 2007, 23, 6544-6553

Microemulsion Droplets Decorated by Brij700 Block Copolymer: Phase Behavior and Structural Investigation by SAXS and SANS Cornelia Sommer,† G. Roshan Deen, and Jan Skov Pedersen* Department of Chemistry and iNANO Interdisciplinary NanoScience Center, UniVersity of Aarhus, Langelandsgade 140, DK-8000 Århus C, Denmark

Pavel Strunz‡ Paul Scherrer Institute, CH-5232 Villigen PSI, Switzerland

Vasil M. Garamus GKSS Research Centre, Abt. WFS, Max Planck Strasse, D-21502 Geesthacht, Germany ReceiVed January 22, 2007. In Final Form: March 26, 2007 The phase behavior and structure of a four-component microemulsion system forming droplets with an oil core surrounded by the non-ionic C12E5 surfactant in water and “decorated” by long PEO chains using the block copolymer/ surfactant Brij 700 has been studied. The surfactant-to-oil volume ratio, the coverage density of the droplets with decorating molecules, and the temperature were varied. For a surfactant-to-oil volume ratio of 2, the solutions form isotropic and clear solutions at room temperature, and the addition of Brij molecules stabilize the micelles: the transition to an opaque phase is shifted to higher temperatures as the surface coverage increases. At a surfactant-to-oil ratio of 1, the isotropic microemulsion phase is confined to a very narrow range of temperature, which location is shifted to increasing temperature, as the amount of Brij at the surface of the droplet is increased. For large surface coverages, the lower emulsification boundary varies roughly linearly with the surface coverage. The structure of the droplet phase was investigated by small-angle neutron scattering (SANS) and small-angle X-ray scattering (SAXS). For a surfactant-to-oil ratio of 2, the SANS data revealed a transition from rodlike to spherical particles when Brij molecules are added to the system, which induces a larger curvature of the surfactant film. For a surfactant-to-oil ratio of 1, the droplets are nearly spherical at all surface coverages. The intermicellar interactions effects become increasingly more pronounced as Brij is added, due to the introduction of the highly swollen corona. A quantitative analysis of some of the SAXS data was done using an advanced model based on Monte Carlo simulations. It demonstrates the strong chain-chain interactions within the corona and confirms the increased interparticle interactions, as the coverage density is increased.

Introduction Microemulsion solutions consist usually of at least three components and are the result of the amphiphilic character of one of the components, generally a surfactant, which can interact favorably with both the oil and the aqueous media.1 This property is essential in many biological systems like cellular membranes and vesicles and plays a key role in various life processes, for example, transport of lipids in blood. A large number of medical and industrial applications involve the use of amphiphilic and similar self-organizing molecules, which makes the study of microemulsions a research topic of broad interest. The spontaneous formation of aggregates from surfactant molecules is the result of a balance among different effects, and the shape of the aggregates can qualitatively be predicted from the molecular geometry of the surfactant by determining the so-called packing parameter. Increasing the packing parameter usually results in the successive formation of spherical micelles, cylindrical micelles, flexible bilayers and vesicles, planar bilayers, and inverted micelles. For a given system, the packing parameter * Corresponding author. E-mail: [email protected] (JSP). † Present address: Swiss National Science Foundation, P.O. Box 8232, CH-3001 Bern. ‡ Present address: Nuclear Physics Institute ASCR, 25068 R ˇ ezˇ near Prague, Czech Republic. (1) Hellweg, T. Curr. Opin. Colloid Interface Sci. 2002, 7, 50.

can be varied by a change of temperature, solvent, or salt concentration or by the presence of co-surfactants. Microemulsions are the result of the ability of micelles to solubilize certain amounts of oil in their cores. Their special interest resides in their ability to form isotropically clear and thermodynamically stable solutions, in spite of the immiscibility of oil and water. The droplets can be spherical, rodlike, or discshaped and have a typical size of 5-50 nm, i.e., much less than the wavelength of light, which explains their optical clarity. Their typical size enables the use of small-angle X-ray (SAXS) and neutron (SANS) scattering techniques to investigate their structure and size.2 Block copolymer molecules consist of chemically distinct macromolecules covalently bound to form a single chain. When dissolved in a solvent that is good for one of the blocks but poor for the other, aggregates are observed, like the micelles and other self-assembled structures formed by their small-sized analogues, known as surfactant molecules. Due to their structure, block copolymer micelles have an additional and unique feature compared to short-chain surfactants, namely, the presence of an extended corona surrounding the micellar core. The corona domain consists of polymer chains “grafted” at the surface of the hydrophobic core, and their configuration ranges from a (2) Fairclough, J. P. A.; Hamley, I. W.; Terrill, N. J. Radiat. Phys. Chem. 1999, 56, 159.

10.1021/la070181j CCC: $37.00 © 2007 American Chemical Society Published on Web 05/05/2007

Microemulsion Droplets Decorated by Brij700

“brush” of stretched chains to a “mushroom” of coiled chains, depending on the coverage density. It was recently shown that the corona could be viewed as a quasi two-dimensional semidilute solution.3,4 In the present work, we study a mixed microemulsion-block copolymer system. This system provides the possibility to vary a higher number of parameters governing the microstructure than in the neat pure copolymer micellar system. In the latter case, it is not possible to vary independently the aggregation state of the micelles, which affects the density of chains at the surface of the particles, and the overall micellar volume fraction. There is only one degree of freedom, the polymer concentration. However, since we are interested in the effects of the chainchain interactions within the corona and the soft interactions between particles as a function of surface coverage, the possibility of adding variable amounts of “decorating” molecules to microemulsion droplets represents an attractive approach. Furthermore, such a system can be tuned in many different ways, by varying the relative amounts of each component. The main disadvantage is the enhanced complexity of such a fourcomponent system and its behavior. We have therefore studied the phase diagram of the system with special emphasis on locating the droplet L1 phase. The phase behavior and microstructure of ternary microemulsions have been studied intensely in the past. In this connection, the L1 phase has attracted considerable attention, due to its potential applications in drug delivery and many industrial applications, for example, food processing. The field has recently been reviewed by Hellweg.1 The three-component system C12E5-n-decane-H2O has, for a fixed surfactant-to-oil volume ratio of 0.815,5 an L1 phase with nearly spherical droplets in the temperature range between 25 and 32 °C. The positive mean curvature is decreasing with increasing temperature, and the droplets become slightly elongated at the upper phase boundary.6 The structure of the L1 domain in the vicinity of the low boundary corresponds to nearly spherical droplets with a radius of 75 Å.7 For non-ionic systems with alkyloxyethylene glycol monoether, it is well-known that the spontaneous curvature of the system is strongly temperature-dependent.8 An increasing temperature leads to a dehydration of the hydrophilic group, and hence to a decrease in the spontaneous curvature of the surfactant film.9 Therefore, the curvature of the film can be tuned by simply changing the temperature. Below 25 °C, part of the oil is expelled as a consequence of the decrease of the droplet size. This phase separation is referred to as emulsification failure and was investigated in detail by Safran et al.10 As the temperature increases, the droplets are subjected to weak growth where they become elongated,6 and the system progresses through several stable phases: a bicontinuous microemulsion, a lamellar phase11,12 (LR), and a sponge phase13,12 (L3). The microstructure of the ternary system has also been investigated as a function of oil content. It was shown by light (3) Svaneborg, C.; Pedersen, J.S. Phys. ReV. E 2001, 64, 10802. (4) Sommer, C.; Pedersen, J. S. Macromolecules 2004, 37, 1682. (5) Olsson, S.; Schurtenberger, P. Langmuir 1993, 9, 3389. (6) Balough, J.; Olsson, U.; Pedersen, J.S. J. Phys. Chem. B 2007, 111, 682. (7) Bagger-Jo¨rgensen, H.; Olsson, U.; Mortensen, K. Langmuir 1997, 13, 141. (8) Olsson, U.; Wennerstro¨m, H. AdV. Colloid Interface Sci. 1994, 49, 113. (9) Lindman, B.; Wennerstro¨m, H. J. Phys. Chem. 1991, 95, 6053. (10) Turkevich, L. A.; Safran, S. A.; Pincus, P. A. In Surfactants in solution; Mittal, Bothorel, Eds.; Plenum: New York, 1986; Vol. 6, p 1177. (11) Bagger-Jo¨rgensen, H.; Olsson, U. Langmuir 1996, 12, 4057. (12) Balough, J.; Olsson U; Pedersen, J. S. J. Dispersion Sci. Technol. 2006, 27, 497. (13) Le, T. D.; Olsson, U.; Wennerstro¨m, H.; Schurtenberger, P. Phys. ReV. E 1990, 60, 4300.

Langmuir, Vol. 23, No. 12, 2007 6545

scattering14,15 that wormlike micelles undergoing typical growth behavior as a function of concentration are present at low oil content, with the oil solubilized in the hydrophobic core of the cylinders. A maximum in the apparent molar mass was found for a mass fraction of oil, R, in the oil + surfactant mixture of R ) 0.07 (corresponding to a surfactant-to-oil ratio of about 13), above which a decrease was observed, related to a transition from elongated micelles to microemulsion droplets. This behavior was further investigated in a SANS study by the same authors.16 Mixed systems containing different surfactants or surfactants combined with a block copolymer or polyelectrolytes have also been studied.17 Such systems are interesting in order to study polymer-surfactant interactions,18,19 polymer-colloid interactions,20 or boosting effect21 of certain components on the solubilization efficiency of a microemulsion. Another study related to the present work has been done by Filali et al.22 They investigated the effect of adding a hydrophobically modified PEO (modified on one end, PEO-m; or on both ends, PEO-2m) to a microemulsion system with an ionic surfactant in brine. For the PEO-m system, increased repulsive interactions are present between the droplets, whereas bridges are present in the PEO2m system, leading to an attractive interaction. In the present investigation, we chose as a starting point the well-studied ternary system C12E5-n-decane-H2O (D2O). The extensive research on this system covers its phase behavior5,7 and its interaction with co-surfactants,23-25 polymers,26,27 and other species.28 As a decorating molecule, we use here a hydrophobically modified PEO, the so-called Brij700, which consists of a large hydrophilic chain of 100 ethylene oxide monomers attached to a hydrocarbon chain of 18 carbon atoms. The block copolymer Brij700 in aqueous solution is known to aggregate into spherical micelles29 with a hydrophobic core of approximate radius of 20 Å and a large corona of dissolved PEO chains having a typical maximum width of 120 Å. In the presence of microemulsion droplets, we expect the Brij molecules, which have the same chemical groups as the surfactant (ethylene oxide and alkyl chain) to arrange themselves through the C12E5 surfactant film. The phase diagram shows that the block copolymer does not self-assemble independently, but rather associates with the microemulsion droplets. It corresponds indeed to a large gain in entropy for the PEO chains to spread over all the available droplets where the chains have space to take a coiled conformation, compared to squeezing themselves together around the small cores of pure Brij micelles. This structure has a corona of dissolved (14) Menge, U.; Lang, P.; Findenegg, G. H. J. Phys. Chem. B 1999, 103, 5768. (15) Menge, U.; Lang, P.; Findenegg, G. H. Colloids Surf.. A 2000, 163, 81. (16) Menge, U.; Lang, P.; Findenegg, G. H.; Strunz, P. J. Phys. Chem. B 2003, 107, 1316. (17) Kabalnov, A.; Lindman, B.; Olsson, U.; Piculell, L.; Thuresson, K.; Wennerstro¨m, H. Colloid Polym. Sci. 1996, 274, 297. (18) Merta, J.; Garamus, V. M.; Kuklin, A. I.; Willumeit, R.; Stenius, P. Langmuir 2000, 16, 10061. (19) Egger, H.; Nordskog, A.; Lang, P. Macromol. Symp. 2000, 162, 291. (20) Washington, C.; King, S. M.; Attwood, D.; Booth, C.; Mai, S. M.; Yang, Y. W.; Cosgrove, T. Macromolecules 2000, 33, 1291. (21) Jakobs, B.; Sottmann, T.; Strey, S. Tenside, Surfactants, Deterg. 2000, 37, 356. (22) Filali, M.; Aznar, R.; Svenson, M.; Porte, G.; Appell, J. J. Phys. Chem. B 1999, 103, 7293. (23) Rajagopalan, V.; Bagger-Jo¨rgensen, H.; Fukuda, K.; Olsson, U.; Jo¨nsson, B. Langmuir 1996, 12, 2939. (24) Bagger-Jo¨rgensen, H.; Olsson, U.; Jo¨nsson, B. J. Phys. Chem. B 1997, 101, 6504. (25) Evilevitch, A.; Lobaskin, V.; Olsson, U.; Linse, P.; Schurtenberger, P. Langmuir 2001, 17, 1043. (26) Bagger-Jo¨rgensen, H.; Olsson, U.; Iliopoulos, I. Langmuir 1995, 11, 1934. (27) Bagger-Jo¨rgensen, H.; Coppola, L.; Thuresson, K.; Olsson, U.; Mortensen, K. Langmuir 1997, 13, 4204. (28) Kabalnov, A.; Olsson, U.; Thuresson, K.; Wennerstro¨m, H. Langmuir 1994, 10, 4509. (29) Sommer, C.; Pedersen, J. S.; Garamus, V. M. Langmuir 2005, 21, 2137.

6546 Langmuir, Vol. 23, No. 12, 2007

Sommer et al.

Figure 1. Schematic representation of a decorated microemulsion droplet, showing the repartition of the components.

polymer chains attached to the surface of the droplets. We determine the L1 region of the phase diagram for surfactantto-oil volume ratios of b ) 1 and b ) 2 for varying coverage density of decorating molecules with a droplet volume ratio of 0.1. The structure of the droplets in the L1 region is investigated by SANS and SAXS measurements. The two different types of radiation are combined in order to exploit the different contrast conditions of them and in this way obtain more structural information. Experimental Section Pentaethylene glycol dodecyl ether (C12E5) of high quality was obtained from Nikko Chemicals, Tokyo. Decane (>99% purity), Millipore H2O, D2O (99.9% D purity), and Brij700 (C18E100) were purchased from Sigma-Aldrich. All chemicals were used as received. Samples were prepared by weight by mixing appropriate amounts of surfactant, oil, water, and Brij700. Oil, surfactant, and Brij were first mixed to have contact among the alkyl chains of the surfactant and Brij and the decane. Subsequently, the water was added, and the solutions were thoroughly stirred. To ensure homogenization, the samples were always heated at least once until the LR phase was reached, since this phase allows easy dispersion of the decane. The samples were prepared at least 2 days before use and kept at room temperature. Values of density (in g/cm3) used for the sample preparation are 0.978 for C12E5,23 0.731 for decane,30 1.153 for Brij700,31 1.000 for H2O, and 1.10530 for D2O. All samples in the series were prepared at a droplet volume ratio of 0.10 (defined below) and then diluted from the homogenized stock solutions. The phase behavior was determined for H2O. The boundaries of each phase of the systems were determined both by visual observation and between crossed polarizers (“Titoscope” produced in Physical Chemistry, University of Lund). The samples were placed in a thermostated water bath, and the temperature was increased in steps of 1 °C from 20 to 80 °C and the visual changes were recorded. The phase boundary of the lamellar phase is where the samples lost their birefringence. The exact clear-turbid boundaries for the characterization of the droplet L1 phase were determined by measuring the change in turbidity as a function of temperature in a UV-vis spectrophotometer (Perkin-Elmer Lambda 25, equipped with a Perkin-Elmer PTP1 Peltier element). After the visual observation experiments, the samples were rehomogenized by heating them up to the LR phase and allowing to cool to room temperature. The phase behavior procedure was then repeated to confirm the observations. The boundaries of the droplet L1 phase were also determined for D2O, as it is this system that is used for the scattering experiments. The SAXS experiments were performed on the instrument at the University of Aarhus.32 The instrument is a modified version of the original small-angle X-ray equipment NanoStar, produced by Anton Paar, Graz, and distributed by Bruker AXS. It is optimized with respect to flux and background, and therefore ideally suited for solution scattering. The sample is kept in a thermostated quartz

capillary placed in the integrated vacuum chamber of the camera. The instrument configuration gives access to a range of scattering vector modulus q between 0.01 and 0.33 Å-1. (q ) (4π/λ) sinθ, where 2θ is the scattering angle and λ the wavelength). The spectra of all samples were isotropic, and the data were azimuthally averaged, corrected for variations in detector efficiency and for spatial distortions, and converted to absolute scale using the scattering from pure water as a primary standard.33 The scattering length densities were calculated from the values of the specific density of the different components. The SANS experiments were performed at the SANS1 instrument at the FRG1 research reactor at GKSS Research Centre, Geesthacht, Germany34 and at SANS-II at the Paul Scherrer Institute (PSI) in Villigen, Switzerland.35 The range of scattering vectors covered is from 0.005 to 0.26 Å-1 at GKSS and from 0.005 to 0.35 Å-1 at PSI, and it was obtained by several combinations of sample-to-detector distances and wavelengths. The samples were kept in quartz cells (Hellma, Germany) with a path length of 1 or 2 mm, depending on concentration. The samples were thermostated using a circulating water bath. The raw spectra were corrected for backgrounds from the solvent, sample cell, and other sources by conventional procedures. The two-dimensional isotropic scattering spectra were azimuthally averaged, converted to absolute scale, and corrected for detector efficiency by dividing by the incoherent scattering spectra of pure water, which was measured with a 1 mm path length quartz cell. The smearing induced by the different instrumental settings was included in the data analysis.36 The neutron scattering length densities were calculated for the different components of the system. Model for the Droplets. Figure 1 shows a schematic view of the decorated droplet. The overall composition of each sample is determined so as to obtain a defined volume ratio of droplets in water (a), a fixed surfactant-to-oil ratio (b), and a defined Brij fraction in the surfactant film (c). The droplet part corresponds to oil (O) + surfactant (S) + part of the Brij copolymer (B) incorporated in the micelle. This excludes the PEO part of the Brij that is outside the E5 layer of the surfactant film. The volume ratio of droplet a relative to water is thus defined as a)

V O + VS + VM B VW

(1)

where VO, VS, VW, and V M B are, respectively, the volume of oil, (30) Handbook of Chemistry and Physics, 66th ed.; CRC Press: Boca Raton, FL, 1985. (31) Sommer, C.; Pedersen, J. S.; Stein, P. C. J. Phys. Chem. B 2004, 108, 6242. (32) Pedersen, J. S. J. Appl. Crystallogr. 2004, 37, 369. (33) Orthaber, D.; Bergmann, A.; Glatter, O. J. Appl. Crystallogr. 2000, 33, 218. (34) Stuhrmann, H. B.; Burkhardt, N.; Dietrich, G.; Ju¨nemann, R.; Meerwinck, W.; Schmitt, M.; Wadzack, J.; Willumeit, R.; Zhao, J.; Nierhaus, K. H. Nucl. Instrum. Methods Phys. Res., Sect. A 1995, 356, 133. (35) Strunz, P.; Mortensen, K.; Janssen, S. Physica B 2004, 350, E783.

Microemulsion Droplets Decorated by Brij700

Langmuir, Vol. 23, No. 12, 2007 6547

surfactant, water, and Brij in the droplet. The total volume of Brij is the sum of the part incorporated in the droplet V M B and the part dissolved in water V W B reaching beyond the E5 film of the droplet VB )

VM B

+

VW B

(2)

In addition, the volume of Brij in the droplet can be further divided into two parts, the part of the hydrophobic tail included in the oily core V OB and the part incorporated in the surfactant film V SB O S VM B ) VB + VB

(3)

Assuming that a molecule of Brij C18E100 arranges exactly like a C12E5 surfactant molecule in the film separating decane and water implies that V OB corresponds to a CH3(CH2)5- entity, V SB to -(CH2)12-(OCH2CH2)5- , and V W B to -(OCH2CH2)95-OH, allowing to relate each volume part with the total volume of Brij V OB K1 ) VB

V SB K2 ) VB

VW B 1 - K1 - K2 ) VB

(4)

The second fixed parameter is the surfactant-to-oil ratio b, and it is defined to exclude the coronal part of the Brij as b)

VS + V SB VO + V OB

V SB VS

P′c(q) )

P′c(0) )

(36) Pedersen, J. S.; Posselt, D.; Mortensen, K. J. Appl. Crystallogr. 1990, 23, 321. (37) Chevalier, Y.; Zemb, T. Rep. Prog. Phys. 1990, 53 279. (38) Svaneborg, C.; Pedersen, J. S. Macromolecules 2002, 35, 1028. Svaneborg, C.; Pedersen, J. S. J. Chem. Phys. 2000, 112, 9661. (39) Pedersen J. S. AdV. Colloid Interface Sci. 1997, 70, 171. Pedersen J. S. In Neutron, X-ray and Light; Lindner, P., Zemb, T., Eds.; Elsevier: North Holland, 2002; pp 391-320.

(7)

1 1+ν

(8)

where ν is a parameter related to the compressibility of the quasi two-dimensional solution of the chains in the corona and increases with the reduced surface coverage σ/σ*, which can be estimated by NRg2 σ ) σ* 4(R + R )2 g

(9)

N is the number of Brij molecules per droplet, Rg is the radius of gyration of the chains, and R is the radius of the droplet measured at the outer surface of the E5 layer. The parameter σ/σ* is the twodimensional equivalent of the reduced concentration c/c* used for semidilute polymer solutions, and it corresponds to the ratio between the area of a single chain given by Rg and the surface area available per chain at the distance Rcore + Rg from the center of the micelle. The corona scattering Pcorona(q) is the summation of two contributions Pcorona(q) ) Acorona2(q) + P′c(q)

(10)

where the first term represents the Fourier transformation of the centrosymmetric average distribution, and the second term is the effective contribution of the individual chains, or “blob scattering”, which is accessible at high q. When the scattering of the spherical core-shell part is added, the full form factor of a micelle can be expressed as Pmic(q) ) Amic2(q) + P′c(q)

(6)

With the relations in eqs 1, 5, and 6 and using eqs 2, 3, and 4, we obtain expressions for VS, VB, and VO as a function of a, b, c, K1, K2, and VW. Numerical values for K1 and K2 are K1 ) 0.02815 and K2 ) 0.09437, and they are calculated using values of molecular volumes at 25 °C for different environments.37 We have investigated the following conditions: a ) 0.10, 0.05, 0.02, and 0.01; b ) 1 and 2; c ) 0, 0.01, 0.02, 0.05, 0.10, and 0.15. Note that the volume ratio in eq 1 can be used to calculate a volume fraction of droplets fdroplets (without outer EO95 part) as fdroplet ) a/(1 + a), and the Brij ratio in the film given by eq 6 can be used to calculate a Brij volume fraction in the film fB as fB ) c/(1 + c). When mixing the samples, the various volumes are converted into masses by multiplying by the respective densities. SAXS Fitting Model. In this paper, we model some of the data for the nearly spherical droplets with high Brij coverage. The model is similar to the one recently used for modeling Brij micelles29 with some modifications for describing the core shell structure of the oil core and surfactant film. An advanced model based on Monte Carlo simulations is applied, where the scattering of both the core-shell part and the corona is taken into account.38 The data are analyzed on absolute scale and fitted by means of least-squares methods.39 The corona is assumed to consist of self-avoiding semiflexible interacting chains, which do not penetrate into the core region. The chain-chain interaction within the corona is modeled using the

Pexv(q) 1 + νPexv(q)

where Pexv(q) is the form factor of semiflexible self-avoiding chains. The scattering of the chains at zero scattering vector is

(5)

This ensures that, under the assumption that C12E5 and Brij arrange similarly in the droplet, the size is independent of the Brij fraction. Finally, c, the volume ratio of Brij in the surfactant film, is simply defined as c)

random phase approximation40

(11)

with Amic(q) ) Acore-shell(q) + Acorona(q), where Acore-shell(q) is the scattering amplitude of the core shell structure.39 Acorona(q) is described as a partial cubic b spline.41 Interparticle interactions are taken into account using the structure factor S(q) of a monodisperse hard-sphere liquid. For a monodisperse system, the scattering intensity is given by42 I(q) ) Pmic(q) + Amic2(q)[S(q) - 1]

(12)

Polydispersity was taken into account for this expression using the local monodisperse approximation43 and a Gaussian distribution D(R) of the radii of the core-shell part of the particles I(q) )

∫{P

mic(q)

+ Amic2(q)[S(q) - 1]}D(Rcore) dRcore (13)

In order to have a physically reasonable distribution of surfactant, we have scaled the amount in the various droplets with the surface area at the neutral plane between C12 and E5 in the surfactant film. In addition, there is conservation of the overall amount of the various components of the droplets, and molecular constraints are in the analysis.39 Note that, for the sake of clarity, we have not explicitly written the prefactors related to particle density and contrasts in the above equations. The model is used to fit the scattering data and yields quantitative information on the structure, size, and interactions. The fit parameters are the radius of the droplets, the width and shape of the corona, the polydispersity, the hard-sphere radius and hard(40) Edwards, S. F. Proc. Phys. Soc. London 1966, 88, 265. (41) Pedersen, J. S.; Svaneborg, C.; Almdal, K.; Hamley, I. W.; Young, R. N. Macromolecules 2003, 36, 416. (42) Pedersen, J. S. J. Chem. Phys. 2001, 114, 2839. (43) Pedersen, J. S. J. Appl. Crystallogr. 1994, 27, 595.

6548 Langmuir, Vol. 23, No. 12, 2007

Figure 2. (a) Phase behavior of the system with surfactant-to-oil ratio of 1, as a function of surface coverage with decorating molecules; (b) with a surfactant-to-oil ratio of 2. sphere volume fraction, and the forward scattering of the chains. The radius of gyration of the PEO chains was fixed at the unperturbed value of 30 Å. From the resulting fit parameters, the radial profile of the core-shell part and corona can be obtained from the respective Fourier transform of the scattering amplitude.29

Results and Discussion Phase Behavior. Figure 2a,b shows the observed phase behavior of the H2O system for the decorated microemulsions as a function of coverage density (parameter c) for surfactantto-oil volume ratios (b) of 1 and 2, respectively, and a constant droplet volume ratio (a) of 0.10. The conditions in Figure 2a for b ) 1 are close to that of the well-studied system without Brij with b ) 0.815 (which we call here the reference system). When no Brij is added (c ) 0), the solution is clear at room temperature and it corresponds to the domain of the L1 phase. The two-phase domain where the oil separates out (emulsification failure) occurred at 20 °C. Below this temperature, the system is in equilibrium with pure oil. At T ) 30 °C, the solution becomes opaque, and this corresponds to the transition to a lamellar phase LR. Above 40 °C, there is a large change in turbidity probably connected with the formation of a sponge phase (L3).5 One can visually observe the formation of matrices of swollen mass resembling that of a spongelike structure. These observations are not uncommon for non-ionic microemulsions containing poly(ethylene glycol) surfactants. Further increase in temperature brings this phase into equilibrium with excess water (L3 + W). However, in the present work, we focus on the L1 phase in the phase diagram, which reflects the influence of Brij decoration at the surfactant interface.

Sommer et al.

For b ) 1, small coverage ratios (c ) 1% and 2%), where only a small fraction of the surfactant is replaced with Brij molecules, leads only to a small shift of the lamellar phase to higher temperature, whereas the two-phase domain completely disappears above 1% in the range of temperatures studied. At higher values of c, the microemulsion L1 phase is moved to increasingly higher temperature as the coverage density is increased. The presence of the polymer molecules at the surfactant film interface increases the spontaneous curvature of the surfactant film. The long PEO chains prefer to have as much space as possible, as this increases their conformational entropy and increases the entropy of mixing with the solvent. In this case, the increasing coverage induces a decrease of the size of the droplet and the consequent expulsion of oil from the interior of the core. On the other hand, the increase of temperature induces an inverse effect, i.e., a decrease of spontaneous curvature of the surfactant film due to the decrease of the solvent quality for the E5 and PEO. This explains that the L1 phase is moved to higher temperatures as the coverage density is increased. Simultaneously, the L1 phase becomes restricted to a very small temperature domain for high coverage densities, as only a few degrees elevation of the temperature are necessary to observe the next transition. It is not easy to rationalize this effect, but the rather large temperature range of investigation can make the situation rather complex, as the temperature simultaneously influences the solvent quality for PEO. The observance of a bluish translucent phase in between the L1 droplet phase and the lamellar phase is qualitatively attributed to a cylindrical phase. With an increase in temperature (at each fixed charge density), the spherical structures evolve into cylindrical structures before entering the lamellar phase. This is due to the combination of various effects such as droplet curvature, concentration, and coverage density. In this cylindrical phase, the structures are anisotropic with no optical birefringence phenomenon. The bluish tinge is due to scattering of light by the relatively large cylindrical structures. This observation is not surprising, because the microemulsion is dilute and the concentration of the droplet is 10 vol %, which allows the observation of various phases in much detail. The dependence of the microemulsion phases on addition of Brij when the surfactant-to-oil ratio is two times higher (b ) 2), is shown in Figure 2b. For the ternary system (c ) 0), the transition into the lamellar phase occurs at lower temperature as compared to the system with b ) 1; however, as the coverage increases, it is shifted to increasingly higher temperatures. For these conditions, the L1 phase is present at room temperature in the whole range of coverages used in this study. The extension of the microemulsion phase as the Brij amount increases can easily be explained in terms of change of the spontaneous curvature. The more polymer molecules are present at the surface of the droplets, the higher the temperature to favor high curvatures. In the same way, the high-temperature phase separation for c ) 0, which is already shifted to higher temperature, is further shifted 10 °C higher in the presence of 1% Brij. This even disappears from the range of investigation for higher surface coverages. The scattering experiments described in the following were done in the droplet L1 phase, and since SANS were also performed, the experiments were done for the D2O system. Therefore, the boundaries were also determined for this system. Note that the boundaries are in general lower than those of the H2O system, except for high coverages and b ) 1, where they are slightly higher. For both b ) 1 and b ) 2, the boundary temperatures of the L1 phase vary roughly linearly with the surface coverage at large surface coverages. As already mentioned, the introduction of

Microemulsion Droplets Decorated by Brij700

Brij into the surfactant film leads to an increase of the spontaneous curvature. In the phase diagram, this is compensated by the increase in temperature, which leads to gradual dehydration of the E5 headgroups and to poorer solvent quality for the PEO chains in the corona, where the latter effect reduces both the film-PEO interaction and the PEO-PEO interchain interaction. Film curvature13 and interchain interactions44 both vary approximately linearly with temperature in the investigated temperature range. This agrees well with the observed linear dependence of the phase boundaries on Brij ratio and supports the contention that all Brij molecules are connected to the droplets. The observed effect of Brij molecules on the phase behavior of the C12E5-n-decane-D2O system is quite different from the results obtained in other related studies, which suggest that a small amount of amphiphilic diblock copolymers can lead to a dramatic decrease of the amount of total surfactant needed to solubilize given volumes of oil and water,26,45 the so-called boosting efficiency. However, this effect is only observed in bicontinuous microemulsions where similar amount of oil and water are separated by an interfacial surfactant film for block copolymers that are much more symmetric than the Brij molecule used in the present work. For the bicontinuous phase, there are polymer chains on both sides of the surfactant film and this stiffens the film, and this in turn results in a higher solubilizing potential. The situation is radically different in the case of the L1 phase with the highly asymmetric Brij molecule. Here, we rather observe the inverse effect, i.e., the addition of Brij leads to a reduced solubilizing potential. However, for a fixed and sufficiently high amount of surfactant (b ) 2), the addition of Brij is related to a large and proportional expansion of the L1 phase to higher temperatures. These different effects can generally be rationalized well by film curvature arguments. SANS and SAXS Results. Nondecorated Droplets. Figure 3a shows the q dependence of measured neutron scattering intensity at droplet volume ratio between 1% and 10% and a surfactantto-oil volume ratio b ) 2, after scaling for concentration. These data were recorded at 25 °C and correspond to pure microemulsion droplets without decoration, since no Brij is added. The q dependence is similar to that observed for cylindrical micelles. At low q, the typical Guinier behavior originating from the overall size of the micelles is only vaguely observable, meaning that the radius of gyration of the objects is larger than 250 Å. At higher q, a 1/q power-law behavior characteristic of rigid rods is observed. However, at high concentration (10%), this domain is hardly apparent. While at the lowest concentration, the scattering corresponds to individual well-separated objects, interparticle interference effects are present at higher concentrations and the scatting is no longer due to the individual object. As a consequence, there is a reduction of the scattering in the forward direction. The data at higher q are very similar, which reflects that the cross-sectional micellar structure is independent of concentration. For comparison, Figure 3b shows the 2% curve together with the data for the same conditions but containing half the surfactant (b ) 1) recorded in the L1 phase at 25 °C. The shape of the two curves differs substantially, and a qualitative evaluation allows the inference that a rod-to-sphere transition is induced above a certain amount of oil. This is not very surprising, because wormlike micelles are formed in the absence of oil. Menge et al.6,7 studied the effect of oil in detail and found that the cylindrical micelles first begin to swell with an increasing amount of oil, (44) Pedersen, J. S; Sommer, C. Progr. Colloid Polym. Sci. 2005, 130, 70. (45) Endo, H.; Mihailescu, M.; Monkenbusch, M.; Allgaier, J.; Gompper, G.; Richter, D.; Jakobs, B.; Sottmann, T.; Strey, R.; Grillo, I. J. Chem. Phys. 2001, 115, 580.

Langmuir, Vol. 23, No. 12, 2007 6549

Figure 3. SANS data at 25 °C for b ) 2, c ) 0%, rescaled for concentration at droplet volume fractions (scaled to the 10% data). The symbols are a ) 1% (filled circles), 2% (open inverse triangles), 5% (filled squares), and 10% (open diamonds). (b) SANS data at 25 °C for a ) 2%, c ) 0%, and surfactant-to-oil volume ratios of b ) 1 (filled circles) and b ) 2 (open inverse triangles).

while their length remain constant, and then gradually become spherical. In a geometrical point of view, the incorporation of increasing amounts of oil reduces the driving force for unidimensional micellar growth, due to the larger diameter of the micelles. Decorated Droplets. Figure 4a,b illustrates the effect of increasing amounts of polymer in microemulsions at a fixed droplet volume fraction of 5%. All data were recorded at 25 °C. Figure 4a displays SANS data at b ) 2, which clearly show that a change in shape is induced by the addition of Brij. There is a transition from cylindrical droplets without concentration effects to spherical droplets with large interparticle interactions effects, when the ratio of Brij in the surfactant film is increased from 0% to 15%. We expect that, under good solvent conditions, the adsorption of polymer on a surfactant film results in an increase of the film spontaneous curvature away from water. Large structure factor effects become apparent above 2% Brij, showing that even small amounts of polymer have a pronounced effect on the structure and interactions in the system. The drop of the scattering intensity at low q and the progressive appearance of a structure factor peak as the surface coverage is increased are clear indications of increasing repulsive interactions between the particles through their coronas. The data for different coverages coincide at intermediate q, demonstrating that the change in size/ cross-sectional size is small. At high q, there is an increase in the data due to the increased contribution from the polymer chain scattering as the coverage is increased.

6550 Langmuir, Vol. 23, No. 12, 2007

Figure 4. (a) SANS data of microemulsion droplets (a ) 5%) as a function of coverage density, for a value of b ) 2 at 25 °C and c ) 0% (filled circles), 1% (open circles), 2% (filled triangles), 5% (open triangles), 10%(filled squares), and 15% (open squares). (b) SAXS data for the same conditions as in part a (same symbols). (c) SAXS data of microemulsion droplets (a ) 5%) as a function of coverage density, for a value of b ) 1 (same symbols). The temperature was 25 °C at c ) 0, 28 °C at c ) 1%, 30 °C at c ) 2%, 35 °C at c ) 5%, 58 °C at c ) 10%, and 78 °C at c ) 15%.

The different effects at low q for the various c values are due to the long PEO chain of the polymer, from which about 95 wt % is dissolved in the surrounding solvent. The two major consequences are as follows: (i) an increase of the spontaneous curvature in the surfactant film in order to give more room to the chains at the micellar interface, which induces the transition from rods to spheres; (ii) the apparition of a crowded PEO layer at the surface of the droplets, with the so-called corona inducing large intermicellar interactions. Similar experiments have been

Sommer et al.

performed at both lower and higher droplet volume fractions, showing that significant effects are induced at a concentration as low as 0.5%. Comparison of SANS and SAXS Data. Figure 4b shows the corresponding SAXS data also at b ) 2, where the gradual changes at low q are particularly apparent, while the curves at high q show similar features, which is an indication that the cross section of the micelles does not undergo significant size changes under the shape transition. Indeed, the cross-sectional diameter is mainly related to the surfactant-to-oil ratio, which is kept constant in the experiment. However, the general shape of the curves is quite different when comparing SAXS and SANS in Figure 4a,b. This is due to the different contrasts in the two techniques, which makes them good tools to obtain complementary information. When performing SANS in D2O, the contrast of surfactant, Brij, and decane is large and similar, and one has a homogeneous contrast. For SAXS, the scattering length density is given by the excess electron density relative to D2O, and it is negative for decane and alkyl chains, whereas it is positive for E5 and PEO. This core-shell contrast with opposite sign of the scattering length density gives a much lower intensity at low q as compared to SANS, and it is the cause for the difference between SANS and SAXS data. One observes, in addition, that the scattering cross section is several orders of magnitude higher in the SANS data compared to SAXS, and this is a consequence of the much higher contrast in a SANS experiment. Nevertheless, the low contrast is compensated by much higher flux intensity in a SAXS instrument, which provides good counting statistics of the data in Figure 4b. The SAXS data in Figure 4b covers a slightly smaller q range than the SANS data in Figure 4a. However, the data show the same trend at low q with a main structure factor peak appearing at q ) 0.035 Å-1, and due to the different contrast conditions,42 a secondary peak also occurs at q ) 0.063 Å-1 for the highest surface coverage. The variation at high q is smaller than for the SANS data due to the approximate q-2 behavior at all surface coverages, which probably masks the polymer scattering. Figure 4c represents the effect of surface coverage on a system with a surfactant-to-oil ratio of b ) 1 as probed by SAXS. Note that, to stay within the L1 region, the temperature was increased for increasing surface coverage so that T ) 25 °C at c ) 0, T ) 28 °C at c ) 1%, T ) 30 °C at c ) 2%, T ) 35 °C at c ) 5%, T ) 58 °C at c ) 10%, and T ) 78 °C at c ) 15%. The effect of adding Brij is less pronounced than for b ) 2. The differences between Figure 4b and c can be rationalized as a SAXS contrast variation effect (see Figure 5). We note that almost all dependence on Brij coverage density is located at low q, where a progressive decrease of the intensity is induced as the Brij amount is increased. The PEO chains have a positive electron density relative to the solvent, and as the relative amount of PEO in the system increases, the contrast increases at the same time. One important consequence is that the effect of Brij is much more pronounced for higher surfactant-to-oil ratio (b ) 2), because the total contrast is higher when less oil is present. The contrast is small and negative for higher oil content. A minor contribution to this effect could, however, at the same time come from the coronas, since the experiments at b ) 1 have been done at a higher temperatures, where the solvent quality for PEO is worse than at room temperature (T ) 25 °C). Our previous study on Brij micelles has shown significant and progressive decrease of the corona width and the chain-chain interactions as the temperature is increased.29 Effect of Droplet Concentration. The results shown in Figure 6a-c represent the effect of droplet concentration for a fixed

Microemulsion Droplets Decorated by Brij700

Langmuir, Vol. 23, No. 12, 2007 6551

Figure 5. Schematic drawing of the radial excess electron density profile of a decorated microemulsion droplet. Oil and alkyl chains have a similar and negative excess electron density, whereas the E5 headgroups have a large and positive excess electron density. The PEO chains have a positive excess electron density, and it is increasing for increasing coverage of the droplets. The forward scattering (I(q ) 0)) is the integral of the profile weighted by the volume element r2.

surface coverage of polymer (c ) 10%). In Figure 6a, the scattering data from SANS at T ) 25 °C at fixed b ) 2 are shown after scaling for concentration. For the SAXS data in Figure 6b,c, we show the unscaled data in order to avoid crowding of the plot. At high q where the intensity essentially reflects the form factor of the droplets, the curves superimpose very well, indicating the conservation of the droplet size and structure (rescaled data now shown). At low q however, the plots clearly indicate a dramatic change in the interparticle correlation term. From the observed decrease at I(q ) 0), which corresponds to an increase of the osmotic compressibility, we can infer that the repulsive interactions become stronger as the concentration increases. Such an effect is expected, since the coronas resist compression and overlap. The data reveal the progressive formation of a structure factor peak as the droplet volume fraction is increased, which can be interpreted as liquidlike correlations between the particles. In the same way, the peak becomes sharper and moves to higher values of q as the concentration increases, which reflects that the particles are squeezed closer and become more correlated. In Figure 6b,c, SAXS data analagous to that as in Figure 6a are plotted, for a fixed value of, respectively, b ) 2 (T ) 25 °C) and b ) 1 (at T ) 58 °C). Basically similar features can be observed in both figures, namely, that pronounced changes in droplet interactions take place for droplet volume fractions above 2%. Here again, it is obvious that the effects are more pronounced at b ) 2 than at b ) 1, as previously discussed, due to the different temperatures of the measurements, which leads to different solvent quality of the PEO chains. Model Fits. Figure 7a shows the SAXS data with fits at a fixed coverage density c ) 10%, a surfactant to oil ratio b ) 2, and three different droplet concentrations of 2%, 5%, and 10%, respectively, at T ) 25 °C. At 2% and 5%, the model fits perfectly

Figure 6. (a) SANS data of microemulsion droplets at 25 °C for a fixed coverage density (c ) 10%), a value of b ) 2, rescaled for concentration at droplet volume fractions (scaled to the 10% data). The droplet volume fractions are 1% (filled circles), 2% (open inverse triangles), 5% (filled squares), and 10% (open diamonds) (b) SAXS data for the same conditions as part a without rescaling at droplet volume fractions of 0.5% (filled circles), 1% (open circles), 2% (filled triangles), 5% (open triangles), and 10% (filled squares). (c) SAXS data at 58 °C for microemulsion droplets at a fixed coverage density (c ) 10%) and value of b ) 1, without rescaling for volume fraction (same symbols as part b).

to the experimental data, whereas the agreement is less good at 10% around the secondary form factor maximum. One sees that the oscillation is more pronounced in the model curve than in the data. As the micellar structure is very similar for the different concentration, the more pronounced oscillation at 10% originates

6552 Langmuir, Vol. 23, No. 12, 2007

Sommer et al.

Some additional parameters related to the intermicellar interactions were derived from the fits. The hard-sphere radius, which gives an estimate of the average distance between the droplets, decreases for increasing concentrations (99 ( 1 Å at a ) 2%, 96 ( 1 Å at a ) 5%, and 86 ( 3 Å at a ) 10%), which is an indication that the droplets are pushed together and the coronas are overlapping at high concentrations. The corona volume fraction at these distances is 2-3%, in perfect agreement with previous findings for pure Brij29 and Pluronic P85 micelles.46 The hard-sphere volume fraction is 0.12 ( 0.01 at a ) 2%, 0.28 ( 0.01 a ) 5%, and 0.41 ( 0.01 a ) 10%. The decreasing values of the ratio of the hard-sphere volume fraction to the concentration also show the interpenetration of micellar coronas. Similar effects were observed for pure Brij700 micelles.29

Conclusion

Figure 7. (a) SAXS data at 25 °C for microemulsion droplets at a fixed coverage density (c ) 10%), b ) 2, and 2% (filled circles), 5% (open triangles), and 10% (filled squares) droplet volume fraction with the corresponding fitting curves. (b) Calculated volume fraction profiles of EO chains in the corona at 2% (solid line), 5% (dashed line), and 10% (dotted line).

from oscillations in the structure factor. The contrast conditions smears out to some extent the high q oscillations in the structure factor;42 however, this smearing is not sufficient in the present case, and one can conclude that the hard-sphere structure factor has oscillations that are too large at high q. This is probably due to the hard-sphere structure factor being inadequate to describe the soft interactions of the corona at high concentrations. We also derived the radial profiles of the chains in the corona (Figure 7b), which give access to the real coverage density. The droplet radius as given by the neutral plane between hydrophobic and hydrophilic components is 38 ( 1 Å. The polydispersity of the micellar radius is small with a value of σR ) 4-5 Å. Note that the PEO profile overlaps with the E5 shell between 38 and 45 Å (this contribution is not shown). There is a systematic decrease of the PEO volume fraction within the surfactant film as the droplet volume fraction is increased. However, the values are in good agreement with the expected value for the Brij volume fraction in this region fB ) c/(1 + c) ) 0.091. One can speculate that the penetration of the PEO from the surrounding micelles into the corona screens the PEO interactions within the corona and that this leads to an effectively smaller area per Brij molecule at high concentrations. One does in fact also see a simultaneous small increase in the droplet radius in accordance with this. The corona reaches a radius of 130-140 Å. One observes that the extent of the PEO corona increases slightly for increasing volume fraction. The forward scattering of the chains P′c(0) showed a decreasing trend with increasing concentration (0.73 at a ) 2%, 0.66 at a ) 5%, and 0.49 at a ) 10%); however, the standard errors were too large to consider this trend significant.

In this work, we have studied the effect of adding the block copolymer Brij700 on the phase behavior of the C12E5-decanewater microemulsion system with emphasis on the L1 droplet phase. The structure in this phase consists of decorated micellar droplets where part of the polymer is incorporated in the surfactant film and the large PEO chains are attached to the curved interface that is in contact with the surrounding water. The phase behavior of the four-component system has been investigated and the domain of the microemulsion droplet phase determined, for surfactant-to-oil volume ratios of 1 and 2. The PEO chains induced a larger curvature of the surfactant film, and for the surfactantto-oil volume ratio of 1, this leads to a linear temperature dependence of the phase boundary of the L1 phase on the surface coverage. For the surfactant-to-oil volume ratio of 2, only the upper phase boundary of the L1 phase was observed in the investigated temperature range, and it displayed a similar temperature dependence on surface coverage. The dependence can be explained as curvature introduced by the decorating molecules and a compensating effect from increasing the temperature and reducing the solvent quality for the PEO chains and thus reducing their mutual interactions. SANS and SAXS measurements have been performed for droplet volume fractions ranging from 0.5% to 10% and for fractions of incorporated decorating molecules in the surfactant film between 0% and 15%. The data show that, for a surfactantto-oil volume ratio of 2, the replacement of a small fraction of surfactant with decorating molecules induces a rod-to-sphere transition in the droplet shape. In general, the scattering curves show very pronounced structure factor effects when either the droplet concentration is increased at a fixed coverage density of polymer or the coverage density is increased at a fixed droplet concentration. The large interparticle interaction effects show up as a consequence of the extended coronas of PEO chains at the surface of the droplets, which make the particles interact even at very low concentrations. SAXS data for a surfactantto-oil volume ratio of 2 at a fixed coverage density of 10% and at three different droplet concentrations have been fitted with an advanced model based on Monte Carlo simulations. The model consists of a spherical core-shell structure with an extended corona and includes particle interactions. The model gives a satisfying description of the experimental data, demonstrating that the droplets are spherical at a surfactant-to-oil volume ratio of 2 when the surface coverage is high. (46) Pedersen, J. S.; Gerstenberg, M. C. Colloids Surf., A 2003, 21, 175.

Microemulsion Droplets Decorated by Brij700

The C12E5-decane-water decorated microemulsion droplets represent an appealing model system that allows studies of a wide range of different effects due to the large number of parameters that can be varied for the system. One can, for example, tune interface curvature, droplet size, surface coverage, and soft particle interactions.

Langmuir, Vol. 23, No. 12, 2007 6553

Acknowledgment. This work was supported by the Danish Natural Science Research Council. Stimulating discussions with Ulf Olsson and Joakim Balogh are gratefully acknowledged. LA070181J