J. Phys. Chem. B 2001, 105, 11081-11088
11081
Influence of Water-Soluble Polymers on the Shear-Induced Structure Formation in Lyotropic Lamellar Phases Jo1 rg Berghausen,† Johannes Zipfel,‡ Peter Lindner,‡ and Walter Richtering*,§ Institut fu¨ r Makromolekulare Chemie, UniVersita¨ t Freiburg, Stefan-Meier-Str. 31, D-79104 Freiburg i. Br., Germany, Institut Laue-LangeVin, Grenoble, France, and Institut fu¨ r Physikalische Chemie, Christian-Albrechts-UniVersita¨ t zu Kiel, Olshausenstr. 40, D-24098 Kiel, Germany ReceiVed: April 26, 2001; In Final Form: August 15, 2001
The shear-induced structure formation in lyotropic lamellar phases containing water-soluble polymers is investigated. The lyotropic phases consisted of sodium dodecyl sulfate/1-decanol/D2O and were mixed either with poly(n-isopropylacrylamide) (PNIPAM), hydroxyethyl starch (HES), poly(vinyl caprolactame) (PVCa), or poly(ethylenglycol)distearate (PEG-DS). Rheo-optical experiments (flow birefringence and small-angle light scattering, SALS) as well as small-angle neutron scattering (SANS) combined with a commercial rheometer were used to observe structural changes, e.g., layer reorientation or formation of multilamellar vesicles (liposomes). Equilibrium properties of the lamellar phases were investigated using quasi elastic light scattering (QELS) and static SANS, the latter was analyzed using a model proposed by Nallet et al. The polymer addition led to a viscosity increase but the flipping of aligned lamellae from parallel to perpendicular orientation was hardly affected by the polymers. The shear-induced formation of multilamellar vesicles (MLV), however, was strongly influenced by the macromolecules. The addition of small amounts of PNIPAM shifted the region where vesicles are formed to samples with higher decanol contents whereas HES, PVCa, and PEG-DS suppressed the MLV formation in all cases. Results from SANS and QELS indicate a possible correlation between the shear-induced vesicle formation and the viscoelastic properties of the surfactant bilayer.
Introduction The lamellar phase LR is often encountered in phase diagrams of surfactants and has been intensely studied for a long period. Many contributions focused on the understanding of interactions between the surfactant membranes.1 The dominating repulsive forces are the electrostatic interaction for charged membranes and the undulation interaction introduced by Helfrich.2 Recently, several groups addressed the question how additional parameters can influence the structure of lamellar phase. One aspect is the influence of macromolecules as guest components in the lamellar phase. Depending on the molecular structure, polymer molecules can be located either in the water layers or in the surfactant bilayer, penetrate the membranes or adsorb onto the bilayer. However, the constraints caused by the lamellar surrounding do not favor the miscibility of the polymer and the surfactant mesophase, and this can lead to phase separation. Several examples of different interactions of watersoluble polymers with lamellar phases have been reported. Kekicheff et al.3 reported on the penetration of the lamellae in the system sodium dodecyl sulfate (SDS)/water without deformation of the polymer coil. A decrease of lamellar spacing was found in the system cetylpyridiniumchloride/hexanol/water by addition of polyacrylamide.4 The incorporation of polymers into the lamellae has also been analyzed5,6 and oligo- and polyglycosides are found to be enclosed in the water phase or the lipid layer, depending on the degree of polymerization.7 Bagger* Author to whom correspondence should be addressed. Fax:+ 49 431 880 2830. E-mail:
[email protected]. † Universita ¨ t Freiburg. ‡ Institut Laue-Langevin. § Christian-Albrechts-Universita ¨ t zu Kiel.
Jo¨rgensen et al. and Yang et al. studied the influence of hydrophobically modified polymers.8,9 In most of these studies, the static properties of the system were under investigation as, e.g., the influence of the added polymer on the interaction between the surfactant membranes.10 A second important aspect is the influence of a mechanical deformation on the structure of lamellar phases. Shear flow is known to induce manifold structural changes in the mesoscopic order: first changes in layer orientation have been observed and in several cases transitions between different layer orientations have been found,11-15 and also the variation of layer orientation across the gap of the shear cell has been studied.16 A second interesting feature is the shear-induced formation of multilamellar vesicles MLVs (also denoted as liposomes, spherulites, or onions).12,17-23 To explore the nature of these phenomena lamellar phases of different chemical composition, such as ionic or nonionic surfactants as well as block copolymers24 and polysoaps,25 were investigated. The influence of shear flow is of particular interest because often it cannot be avoided when the different components of the samples are mixed. Consequently, there is an ongoing discussion whether vesicles form spontanously or are induced by the mechancial deformation during the sample preparation process,26-28 In this contribution we report on shear-induced structure formation in polymer-doped lamellar phases. Such systems are of practical interest since polymer-surfactant mixtures are abundant in pharmaceutical and cosmetic applications. Furthermore the above-mentioned shear-induced vesicle formation can be used for the encapsulation of active ingredients as, e.g., DNA in gene therapy29,30 and therefore it is important to know how polymers affect the vesicle formation. Theoretical and experi-
10.1021/jp0115897 CCC: $20.00 © 2001 American Chemical Society Published on Web 10/20/2001
11082 J. Phys. Chem. B, Vol. 105, No. 45, 2001
Berghausen et al.
TABLE 1: Molar Decanol Fraction xc in the Surfactant Membrane and Composition in Weight Percent for All Samples sample xc SDS (w/w %) decanol (w/w %) D2O (w/w %)
1 0.285 26.77 5.86 67.37
2 0.301 26.40 6.23 67.37
3 0.313 26.10 6.53 67.37
4 0.326 25.78 6.86 67.37
5 0.354 25.09 7.55 67.37
6 0.367 24.75 7.88 67.37
7 0.411 23.60 9.04 67.37
8 0.444 22.69 9.94 67.37
mental studies on the influence of polymers on membrane properties revealed that added polymers lead to an increase of the bilayer flexibility. However, Bouglet et al. reported that the addition of a nonadsorbing water-soluble polymer caused the formation of multilamellar vesicles indicating a negative shift of the Gaussian modulus which is in contradiction with theoretical predictions.31 A favored vesicle formation after polymer addition was also reported by Ko¨tz et al.32 However, there are only a few studies where the influence of shear on polymer-modified lamellar phases has been examined in detail. As a model system we chose the well-characterized ternary system consisting of sodium dodecyl sulfate (SDS), decanol, and water with a D2O mass fraction of 0.674. A lamellar phase is formed at this concentration, but depending on the surfactant/ cosurfactant ratio different defects are observed within the surfactant bilayer. Previous studies showed that a classical lamellar phase is formed at high decanol content, whereas membrane defects were found when less decanol was present. Shape and concentration of the water holes within the surfactant membrane changed with increasing decanol content.33,34 The influence of shear in this system has been investigated and was described in terms of an orientation diagram. Different types of layer orientation were observed and shear-induced vesicle formation was observed only at intermediate decanol content.34 In this contribution we describe the influence of water-soluble polymers on the structure of this defective lamellar phase at rest and under shear. Different water-soluble polymers namely poly(n-isopropyl acrylamide) (PNIPAM), poly(vinyl caprolactame) (PVCa), hydroxyethyl starch (HES), and poly(ethyleneglycol)distearate (PEG-DS) were used to modifiy the lamellar phase. Static properties of the system were investigated using small-angle X-ray and neutron scattering as well as quasi elastic light scattering experiments. Rheo-optical and rheo-neutron scattering experiments were used for the structure determination under shear. One of the key factors for the understanding of the dynamic properties of complex fluids might be to find out more about correlations between static properties and the behavior under shear. Experimental Section Materials and Sample Preparation. 1-Decanol (99%) was purchased from Sigma-Aldrich, sodium dodecyl sulfate (>99%) from Fluka. All chemicals were used without further purification. The samples were prepared by weighing the appropriate amounts of SDS and water into glass tubes. Then they were rigorously mixed for 2 days and allowed to equilibrate for at least two weeks. The defective lamellar phase consisted of sodium dodecyl sulfate/decanol/D2O with a D2O mass fraction of 0.674. Samples with different molar fraction xc of decanol in the surfactant/cosurfactant mixtures were investigated. All experiments were performed at 25 °C. Table 1 summarizes the composition of the samples. Poly(n-isopropyl acrylamide) (PNIPAM) was prepared by free radical polymerization of n-isopropyl acrylamide (NIPAM) in 1,4-dioxane. A 6.78 g (0.06 mol) amount of NIPAM was dissolved in 120 mL of 1,4-dioxane, and 25 mg (0.152 mmol)
of 2,2′-azo-bis(isobutyronitrile) (AIBN) was added. The solution was degassed and purged with nitrogen five times. NIPAM was polymerized for 2.5 h at 65 °C under a nitrogen atmosphere. The polymer was precipitated from 500 mL of diethyl ether and dissolved twice in 100 mL of 1,4-dioxane and precipitated from 600 mL of diethyl ether. The product was dissolved in water and lyophilizized. 1H NMR-spectrum: (CDCl3) δ/ppm: 1.15 (CH3-CH-CH3), 1.4-2.5 (-CH2-CH-), 3.05 (N-H), 4.0 (CH3-CH-CH3). The intrinsic viscosity (Staudingerindex) in aqueous solution is [η] ) 37.4 mL/g corresponding to a number averge molar mass of Mn ) 66 kg/mol.35 The hydrodynamic radius was determined by dynamic light scattering as Rh ) 13 nm. Poly(vinyl caprolactame) (PVCa) was prepared by free radical polymerization of N-vinylcaprolactame in toluene. The monomer was freshly distilled under vacuum (bp 74 °C). A 10.0 g (0.072 mol) amount of N-vinylcaprolactame was dissolved in 40 mL of toluene and 22 mg (0.134 mmol) of 2,2′-azo-bis(isobutyronitrile) (AIBN) was added. The solution was degassed and purged with nitrogen three times. N-Vinylcaprolactame was polymerized for 6 h at 55 °C under a nitrogen atmosphere. The polymer was precipitated from 300 mL of diethyl ether and dissolved in 30 mL of toluene and precipitated from 600 mL of diethyl ether. The product was dissolved in water and lyophilized. Yield: 4.0 g. The Staudinger index in aqueous solution at 20 °C was [η] ) 18.8 mL/g corresponding to a weight average molar mass of Mw ) 50 kg/mol.36 Hydroxyethyl starch (DS ) 0.5) was purchased from Pfeifer & Langen, Dormagen, Germany, and used as received. From light scattering measurements in aqueous solution Mw ) 200kg/ mol and Rh ) 12 nm were obtained. PEG-DS (STEPAN PEG 6000 DS, poly(ethylenglycol)distearate is a courtesy of Stepan Europe, Voreppe, France, and was used as received. Small-Angle X-ray Scattering. One part of the measurements has been performed at the department of Physical Chemistry I, University of Lund, Sweden. A Kratky compact small-angle camera with a position-sensitive detector with 1024 channels (OED 50 M, Mbraun, Graz, Austria) was used. The length of the camera was 277 mm. The sample was filled into a glass capillary and was investigated using Cu KR-radiation (50 kV, 40 mA). Some additional measurements have been performed at the Department of Physics, University of Freiburg, Germany, using a Kratky compact camera (A. Paar, Graz, Austria) with a distance of 200 mm. Small-Angle Neutron Scattering. All neutron scattering experiments have been performed on the instrument D11 of the Institut Max von LauesPaul Langevin (ILL) in Grenoble (France). The neutron wavelength was 4.5 Å with a spread of ∆λ/λ ) 9%. The static measurements were performed with 1 mm cuvettes (Hellma) covering a range of momentum transfer q from 3.5 × 10-3 Å-1 to 0.44 Å-1. The data were collected on a two-dimensional multidetector (64 × 64 elements of 1 × 1 cm2) and corrected for background and empty cell scattering. The incoherent scattering of H2O was used for absolute calibration according to standard procedures and software available at the ILL. Further analysis was done by radially averaging. For flow-SANS experiments a Bohlin CVO-120 HR rheometer was adjusted to the small-angle neutron scattering instrument D11. Measurements were performed in a Couette cell consisting of two quartz cylinders where the inner one rotated and the outer one was fixed. The gap was 1 mm and a rectangular apperture of 0.25 × 15 mm2 was employed to reduce
Shear-Induced Structure Formation in Lamellar Phases
J. Phys. Chem. B, Vol. 105, No. 45, 2001 11083
the beam size. The scattering experiments were performed in the radial position yielding information in the plane formed by flow and vorticity direction as well as in the tangential position with the neutron beam passing along the flow direction through the side of the Couette cell. For rheo-optical studies, a Bohlin CVO-120 HR rheometer equipped with a quartz glass 1° cone/plane shear geometry was used. The incident laser beam passes through the sample along the direction of the velocity gradient The apparatus allows us to detect birefringence and small-angle light scattering (SALS) under shear. In this paper we only discuss results from depolarized HV light scattering. Here, the incident light is linearly polarized along the flow direction and the crossed analyzer is aligned along the vorticity direction. A detailed description is given in the literature.37 The viscosity measurements were recorded using a preshearing time of 240 s and a data acquisition time of 30 s. Quasi elastic light scattering measurements were performed using an automated goniometer coupled with an ALV-5000 autocorrelator (ALV, Langen, Germany). As light source a Spectra Physics 2020 Kr-ion-laser (λ0 ) 647.1 nm) was employed. To deal with the heterodyne nature of the lamellar phases, all measurements were performed using the “brute force” method.38 Here, the sample is turned after recording the time correlation function (TCF) and by averaging 120 of these correlation functions (with a data acquisition time of 20 s), an ensemble average is obtained. The TCFs were evaluated by fitting the data either with one or a sum of two stretched exponential functions. The averaged relaxation time for each stretched exponential follows from
τ 1 〈τ〉 ) Γ β β
()
(1)
Figure 1. (Top) Angular dependence of the SAXS intensity of sample xc ) 0.367 without and with 4% PNIPAM; (bottom) angular dependence of the SANS intensity of sample xc ) 0.411 with 0.5% PNIPAM.
with the gamma function Γ(x). Further experimental details are described elsewhere.39 Both polarized and depolarized scattering was investigated but the relaxation times were identical within experimental uncertainty. Results In the following we will present results obtained from the different lamellar samples with increasing decanol content. We will focus mainly on PNIPAM and HES at a concentration of 0.5% (w/w). Additional experiments at higher concentrations and with PVCa and PEG-DS have been performed, but will be mentioned only briefly. Behavior in the Quiescent State. The hydrodynamic radii of both of the larger macromolecules PNIPAM and HES were ca. 12 nm which is bigger than the smectic spacing of ca. 6-8 nm. Nevertheless, these two polymers were nicely miscible with the lamellar phase. For most of the experiments a polymer concentration of 0.5% was employed which is below c*, the overlap concentration of the polymers in aqueous solution. The SAXS data displayed in Figure 1a demonstrate that the lamellar order was still present after addition of polymers but the Bragg peaks shifted to higher q-values. Additional static SANS measurements of the ternary system were performed before and after addition of some water-soluble polymers, one example is shown in Figure 1b. No influence on the so-called “defective” peak at lower scattering angles was detected. The SAXS and SANS data from samples at rest show that the presence of the polymer caused a decrease of the lamellar spacing indicating that water is taken out of the interlamellar volume. Such a behavior has been found in systems that phase separate into a polymer solution and a higher concentrated
Figure 2. Variation of the lamellar spacing with polymer concentration for different samples as obtained by SAXS.
surfactant lamellar phase. Phase separated samples usually become optically turbid,7 whereas our samples did not become cloudy in this concentration range. Turbid samples were observed at much higher polymer contents. This indicates that no phase separation occurred at lower polymer concentrations. The layer spacing was approximately a linear function of the polymer content as long as no phase separation occurred. With increasing decanol content this effect was more pronounced, as shown for the case of PNIPAM in Figure 2. To find out more about the intrinsic properties of the lamellae, quasi elastic light scattering experiments were performed with some selected samples. These dynamic light scattering experiments have been performed with nonaligned, i.e., polydomain samples. Such samples show nonergodic behavior and the time-
11084 J. Phys. Chem. B, Vol. 105, No. 45, 2001
Figure 3. Diffusion coefficients as obtained from dynamic light scattering from lamallar samples at different decanol and polymer content, respectively.
correlation functions were determined at 120 different sample positions in order to achieve an ensemble average. The time correlation functions from samples of low decanol content revealed two modes of motion, regardless of the presence of polymer. However, in the cases of samples with higher decanol content (xc ) 0.367, xc ) 0.411), the second mode at shorter relaxation times was only observed when 0.5% PNIPAM was added. The reciprocal relaxation times were proportional to q2 typical of translational motions and the corresponding diffusion coefficients Dc could be determined, see Figure 3. In most cases we observed two relaxation modes; however, a detailed analysis of the relaxation processes in terms of specific types of mode as, e.g., a baroclinic or undulation mode requires aligned samples40 which were not available in this study. Nevertheless the results from the different samples can be compared. The slower mode with a diffusion coefficient in the order of 7 × 10-8 cm2/s was nearly independent of the decanol molar fraction. Even the addition of PNIPAM did not change the relaxation time of this process. A second mode of motion was observed at shorter relaxation times and diffusion coefficients in the order of 2 × 10-6 cm2/s. For the two samples with a low decanol molar fraction, this process was found both without polymer and with 0.5% PNIPAM added. However, the fast motion was found for the two samples with high decanol molar fraction only after addition of 0.5% PNIPAM. HES, on the other hand did not induce the second mode of motion. Behavior under Shear. Apparently, SAXS and SANS experiments in the quiescent state showed only little influence of the added polymers on the properties of the lamellar system. The decrease of layer spacing was independent of the polymer type. The rheological properties, however, revealed differences for most of the samples when water-soluble polymers were added. At a low decanol content of xc ) 0.301 the addition of polymer did not change the qualitative behavior. The viscosity was higher but shear thinning and a transition from parallel layer orienation (i.e., with the layer normal along the velocity gradient direction) to perpendicular layer orientation (i.e., with the layer normal along the neutral direction) was observed similar to the sample without polymer. With the pure surfactant system a second transition to a parallel orientation was observed at high shear rates.14 The polymer-doped samples, however, revealed strong foaming at higher shear rates, so that it was not posible to study the behavior at very high shear rates. In Figure 4, the rheological behavior of the lamellar phase at xc ) 0.326 is shown and a strong effect due to the added
Berghausen et al.
Figure 4. Flow curves of the pure lamellar phase at xc ) 0.326 and doped with 0.5% PNIPAM and 0.5% HES, respectively.
Figure 5. SANS patterns obtained with the tangential beam configuration from the sample xc ) 0.326 doped with 0.5% PNIPAM at shear rates of 50 (top) und 400 s-1 (bottom), respectively.
polymer is observed. The pure surfactant system revealed a shear-induced formation of multilamellar vesicles with increasing shear, as indicated by the viscosity maximum and described previously.34 The addition of 0.5% PNIPAM led to a kink in the viscosity curve and the maximum was no longer observed. A similar behavior was found when 0.5% HES was used instead of PNIPAM, the kink was less pronounced but again no viscosity maximum was found. To elucidate the structural changes underlying this change in viscosity, SANS and rheobirefringence measurements were performed. Figure 5 displays SANS data obtained with the tangential beam configuration from sample xc ) 0.326 doped with 0.5% PNIPAM. One can clearly see that a transition from parallel to perpendicular layer orientation occurred with increasing shear rate, but the MLV
Shear-Induced Structure Formation in Lamellar Phases
J. Phys. Chem. B, Vol. 105, No. 45, 2001 11085
Figure 6. Flow curves of the pure lamellar phase at xc ) 0.354 and doped with different polymers.
Figure 7. Flow curves of the pure lamellar phase at xc ) 0.411 and doped with 0.5% PNIPAM and 0.5% HES, respectively.
state, which was found in the case of the pure lamellar phase, was not detected when polymer was present. Flow birefringence experiments confirmed the transition from parallel to perpendicular layer orientation while the optical transmission recorded simultaneously revealed no change of turbidity in contrast to the lamellar phase prior to polymer addition when MLV were formed. Thus a polymer addition of 0.5% caused a suppression of the shear-induced MLV formation independent of the type of the macromolecule. Different behavior was observed with the sample at xc ) 0.354. Figure 6 shows the flow curves for the pure and the different polymer-doped samples. The pure systems displayed a shear-induced MLV formation as can be seen from the viscosity maximum, a hysteresis in the flow curve and proved by SANS and SALS as described before.34 The viscosity maximum was no longer observed after addition of 0.5% HES or PVCa. Thus these polymers suppressed the vesicle formation similar to the sample at xc ) 0.326 described above. The sample doped with PNIPAM, however, behaved differently: the viscosity maximum resulting from the formation of multilamellar vesicles was still present. Again a different behavior was observed with the sample at xc ) 0.411, the flow curves are shown in Figure 7. Without added polymer, the sample showed a parallel-to-perpendicular transition with a kink in the visosity. The addition of 0.5% HES did not alter this behavior, whereas the addition of 0.5% PNIPAM led to a maximum in the viscosity curve (accompanied by an increase in turbidity) indicative of the shear-induced
Figure 8. SALS and SANS patterns from the sample at xc ) 0.411 with 0.5% PNIPAM at ca. 9 Pa. Top: depolarized SALS; middle: SANS radial beam; bottom: SANS tangential beam.
formation of multilamellar vesicles. Similar behavior was found when the PNIPAM concentration was increased to 2%. Rheo-SALS and SANS experiments were performed in order to prove that multilamellar vesicles are formed in this polymermodified sample. Figure 8 shows scattering patterns obtained from depolarized SALS (top) as well as from SANS in the radial (middle) and tangential (bottom) beam configuration at a shear stress of 9 Pa. The four-lobe SALS pattern and the fact that the SANS Bragg scattering was observed on the entire azimuthal trace of the two-dimensional detector in both radial and tangential beam, unambiguously demonstrated the shear-induced formation of multilamellar vesicles. Discussion The polymer-induced modification of the behavior under shear of the lamellar phase can be summarized as follows: with the exception of PNIPAM, the presence of 0.5 weight % of water-soluble polymers suppressed the formation of multila-
11086 J. Phys. Chem. B, Vol. 105, No. 45, 2001
Berghausen et al. scattering experiment. The scattering intensity is given as
I(q) )
2πVP(q)S(q)
(2)
dq2
with scattering volume V, lamellar spacing d, form and structure factor P(q) and S(q), respectively. The form factor for neutron scattering is given by Nallet et al. as
P(q) ) 2
2 2 ∆F2 [1 - cos(qδ)e-q σ /2] q2
(3)
with scattering contrast ∆F2, δ the bilayer thickness, and σ arbitrarily fixed at δ/4. The resolution-limited structure factor was calculated as N-1
S(q) ) 1 + 2
[
∑1 -
( )( 1-
n
qdn
)
exp 1 + 2∆q2d2R(n) 2q2d2R(n) + ∆q2d2n2 1 N
cos
2(1 + 2∆q2d2R(n))
]x
(4)
1 + 2∆q d R(n) 2 2
where N is the number of lamellar plates, ∆q the width of the resolution function, and R(n) is a correlation function for the lamellae expressed by
R(n) )
Figure 9. Orientation diagram of the different lamellar samples of the pure surfactant system (bottom) and doped with 0.5% PNIPAM (top).
mellar vesicles, while the transition from parallel to perpendicular orientation of planar layers still occurred. The addition of 0.5% PNIPAM, however, shifted the region where vesicles can be formed under shear. Furthermore it must be emphasized that the transition from parallel to perpendicular lamellar orientation was hardly influenced. The results obtained with the lamellar samples containing 0.5% PNIPAM are summarized in an orientation diagram displayed in Figure 9 (top). For comparison the orientation diagram of the pure surfactant system is also shown in Figure 9 (bottom).34 The rheological data clearly showed that the influence of PNIPAM on the flow properties of the lamellar phase was different from that of the other polymers. The dynamic light scattering data mentioned above also revealed a different influence of PNIPAM as compared to HES. The samples with xc ) 0.367 and xc ) 0.411 revealed the fast mode of motion only when PNIPAM was added. With those two samples, shearinduced formation of vesicles was found only when PNIPAM was present. However, when PNIPAM was replaced by HES, neither was the fast motion detected in dynamic light scattering nor did vesicle formation occur under shear. Additional information on the lamellar properties was deduced from an analysis of the static SANS data using a model proposed by Nallet et al.41 They calculated the q-dependent SANS intensity on the basis of the model of Caille´42 taking into account a combination of two models for the form and the structure factor, respectively. The proposed structure factor describes the modulation of shape and height of the Bragg peak as well as the finite instrumental resolution and powder averaging in the
ηcp 4π2
[ln(πn) + γE]
(5)
with Euler’s constant γE. Thus the Caille´ parameter ηcp can be obtained from the SANS data with this model for the structure factor. ηcp contains the elastic constants of the smectic phase, namely, the smectic bending elasticity K and the compression modulus B h:
ηcp )
q02kBT 8πxKB h
(6)
with q0 the position of the first-order Bragg peak, Boltzmann’s constant kB, and temperature T. This model was also recently used for different surfactant systems and was found to provide an excellent description for the interactions in stiff membranes.10,43,44 The theoretical scattering intensity was fitted to the experimental scattering intensities in order to obtain values for the membrane thickness and the Caille´ parameter. For the calculations we used N ) 60, and ∆q ) 0.0025 Å-1. The lamellar spacing was directly taken from the first-order Bragg-peak position as d ) 2π/q0. The scattering contrast ∆F2 can be calculated from the scattering length densities. However, since the irradiated volume is also a fit parameter we included the contrast into the prefactor fit parameter because it also includes the effect of polymer on the contrast. Thus there were three free fit parameters: the prefactor, membrane thickness δ, and Caille´ parameter ηcp. Figure 1b displays the experimental SANS data and the calculated curve for sample xc ) 0.411 with 0.5% PNIPAM. The results from data analysis are summarized in Table 2. The obtained Caille´ parameter ηcp depends on the defect structure of the lamellar phase,sthe lower the decanol content the higher is ηcp. At high decanol contents, the low Caille´ parameter implies stiffer lamellae. The addition polymers led to an increase of ηcp in both cases. However, with PNIPAM ηcp increased more
Shear-Induced Structure Formation in Lamellar Phases TABLE 2: Results from the SANS Data Analysis: Position of First Order Bragg-peak q0, Lamellar Spacing d, Membrane Thickness δ, Caille´ Parameter ηcp sample xc ) 0.326 xc ) 0.411 xc ) 0.411 + 0.5% PNIPAM xc ) 0.411+ 0.5% HES xc ) 0.444
q0/nm-1 d/nm δ/nm 1.0 0.79 0.84 0.83 0.8
6.3 7.9 7.5 7.6 7.9
1.9 2.1 2 2 2.1
ηcp
KB h /Pa2
0.28 0.13 0.16 0.147 0.13
3.4 × 10-7 6.2 × 10-7 5.1 × 10-7 5.9 × 10-7 6.5 × 10-7
and apparently this enabled the formation of multilamellar vesicles at higher decanol content as compared to the pure surfactant system. The increase of ηcp was less pronounced and no vesicles were formed when HES was added. This result is in accordance with the quasi elastic light scattering experiments which showed no change in the apparent diffusion coefficient in the presence of HES. These results indicate that the behavior of lamellar phases under shear can in principle be correlated to the membrane properties at rest. Recently, Leon et al. studied the influence of salinity on the MLV formation in AOT solutions.45 They described the MLV formation in terms of a homogeneous nucleation process and observed that the activation energy is related to the bending energy cost to create a spherical vesicle from a flat bilayer. Theoretical calculations on the influence of water-soluble polymers on the bending moduli led to the conclusion that adsorbing polymers will decrease the bending energy.46 This is in agreement with the shift of ηcp and the induced MLV formation in the sample at xc ) 0.411 after the addition of PNIPAM. The behavior of the pure surfactant systems already indicated that no vesicles are formed at low decanol content when too many defects exist within the membrane.34 A change of the surfactant/cosurfactant ratio does not only influence the membrane defects but also the bending and Gaussian moduli of the bilayer.47 The Gaussian elastic modulus is known to be important for topological changes in the surfactant aggregate structure45,47-49 but according to the Gauss-Bonnet theorem, the Gauss modulus does not contribute to the amplitude of thermal fluctuations of the membrane and can thus not be determined from SANS data. Apparently, the flexibility of the surfactant bilayer must be in a certain range in order to facilitate the shearinduced formation of multilamellar vesicles. An interpretation concerning the influence of the chemical structure of the water-soluble polymers is not straightforward. The experimental results clearly show that both PNIPAM and HES are well miscible with the lamellar phase. PNIPAM, on one hand, shifted the vesicle region to samples with higher decanol content, whereas HES always supressed MLV formation. At high decanol content different behavior between samples containing PNIPAM and HES, respectively, could be detected in the SANS and QELS. One might assume that the molecular interactions between the polymers and the surfactants are different. Several studies on PNIPAM/SDS mixtures in dilute solution revealed strong attractive interactions.50,51 Assuming that SDS is bound to the PNIPAM chain does not explain the observed behavior under shear, because a binding of SDS to the polymer would cause an increased decanol content in the membrane and consequently should shift the vesicle region to the left in the orientation diagram. In both cases, the lamellar spacing was reduced when the polymer was added. One could speculate whether the polymer chains are mainly located within the defects, and the osmotic pressure can result in a transport of water from the interlamellar volume into the defect holes. This could explain the observed
J. Phys. Chem. B, Vol. 105, No. 45, 2001 11087 decrease in the smectic spacing. However, additional experiments probing the polymer/surfactant interaction and the distribution of the polymer in the lamellar surrounding have to be performed in order to obtain a better understanding of the influence of the macromolecules on the local structure of the lamellar phase. Conclusions The main objective of this contribution was to investigate the influence of water-soluble polymers on the behavior of lamellar phases under shear. Using rheo-optical and neutron scattering techniques we could observe two features typical of lamellar phases under shear: •Shear flow led to an alignment of the surfactant layers which was parallel to the walls of the shear cell at low shear rates but flipped to the perpendicular orientation (i.e., with the layer normal along the vorticity direction) at high shear rates. The addition of polymers led to an increase of the sample viscosity but did not influence the layer reorientation in the range of polymer concentration used in this study. All different polymers showed the same behavior. •However, the influence on the shear-induced formation of multilamellar vesicles depended strongly on the type of the added polymer. HES, PVCa, and PEG-DS always suppressed the shear-induced MLV formation. PNIPAM on the other hand, shifted the vesicle region in the orientation diagram to higher decanol content. i.e., PNIPAM suppressed MLV formation at low decanol content but faVored MLV formation at higher decanol content. The vesicles in polymer-containing samples had an enhanced stability concerning the range of shear rates and the stability after cessation of shear as compared to the vesicles in the pure surfactant system. The fact that the layer reorientation was not influenced by the polymer addition whereas the vesicle formation was strongly affected indicates that these two processes are not directly related to each other. In other words, the flipping of lamellae and the MLV formation, respectively, seem to depend on different parameters. This fits to the experimental observation that the flipping of planar layers has been observed in many different systems including aqueous solutions of low molecular weight surfactants as well as block copolymer solutions12,15 and melts.52 Shear-induced MLV formation, on the other hand, has only been reported for lyotropic systems. The experiments further indicate that the shear-induced formation of MLVs might be controlled by viscoelastic properties of the surfactant membrane. The observation that water-soluble polymers strongly influence the vesicle formation has severe consequences for possible applications. On one hand, one could use this effect to avoid the shear thickening caused by MLV formation during processing of lamellar phases. On the other hand, one needs to check the influence already of small polymer amounts if the potential of shear-induced MLV formation is to be used for encaspulation in biomedical applications. Acknowledgment. This work was supported by the Deutsche Forschungsgemeinschaft within the SFB 428 and by the Fonds der Chemischen Industrie. We thank F. Nallet and P. Richetti for helpful e-mail discussions, U. Olsson for the possibility to perform SAXS experiments as well as for many stimulating discussions, and S. Kastaun for EDV support. A travel grant by the DAAD is gratefully acknowledged. J.Z. thanks the European Union for a Marie-Curie fellowship. References and Notes (1) Roux, D.; Safinya, C. R. J. Phys. France 1988, 49, 307. (2) Helfrich, W. Z. Naturforsch. 1978, A33, 305.
11088 J. Phys. Chem. B, Vol. 105, No. 45, 2001 (3) Kekicheff, P.; Cabane, B.; Rawiso, M. J. J. Colloid Interface Sci. 1984, 102, 51. (4) Singh, M.; Ober, R.; Kleman, M. J. Phys. Chem. 1993, 97, 11108. (5) Radlinska, E. Z.; Gulik-Krzywicki, T.; Lafuma, F.; Langevin, D.; Urbach, W.; Wiliams, C. E.; Ober, R. Phys. ReV. Lett. 1995, 74, 4237. (6) Radlinska, E. Z.; Gulik-Krzywicki, T.; Lafuma, F.; Langevin, D.; Urbach, W.; Wiliams, C. E. J. Phys. II 1997, 7, 1393. (7) Deme´, B.; Dubois, M.; Zemb, T.; Cabane, B. J. Phys. Chem. 1996, 100, 3828. (8) Bagger-Jo¨rgensen, H.; Olsson, U.; Ilioupoulos, I. Langmuir 1995, 11, 1934. (9) Yang, Y.; Prudhomme, R.; McGrath, K. M.; Richetti, P.; Marques, C. M. Phys. ReV. Lett. 1998, 80, 2729. (10) Bouglet, G.; Ligoure, C. Eur. Phys. J. B. 1999, 9, 137. (11) Safinya, C. R.; Sirota, E. B.; Bruinsma, R. F.; Jeppesen, C.; Plano, R.; Wenzel, L. Science 1993, 261, 588. (12) Mortensen, K. Curr. Opin. Colloid Interface Sci. 2001, 6, 140. (13) Penfold, J.; Staples, E.; Tucker, I.; Tiddy, G. J. T.; Kahn Lodhi, A. J. Appl. Crystallogr. 1997, 30, 744. (14) Berghausen, J.; Zipfel, J.; Lindner, P.; Richtering, W. Europhys. Lett. 1998, 43, 683. (15) Hamley, I. W. Curr. Opin. Colloid Interface Sci. 2000, 5, 342. (16) Berghausen, J.; Zipfel, J.; Diat, O.; Narayanan, T.; Richtering, W. Phys. Chem. Chem. Phys. 2000, 2, 3623. (17) Diat, O.; Nallet, F.; Roux, D. J. Phys. II 1993, 3, 1427. (18) Bergenholtz, J.; Wagner, N. Langmuir 1996, 12, 3122. (19) Weigel, R.; La¨uger, J.; Richtering, W.; Lindner, P. J. Phys. II 1996, 6, 529. (20) Bergmeier, M.; Gradzielski, M.; Hoffmann, H.; Mortensen, K. J. Phys. Chem. B 1998, 102, 2837. (21) Mu¨ller, S.; Bo¨rschig, C.; Gronski, W.; Schmidt, C.; Roux, D. Langmuir 1999, 15, 7558. (22) Escalante, J. I.; Gradzielski, M.; Hoffmann, H.; Mortensen, K. Langmuir 2000, 16, 8653. (23) Zipfel, J.; Nettesheim, F.; Lindner, P.; Le, T.; Olsson, U.; Richtering, W. Europhys. Lett. 2001, 53, 355. (24) Zipfel, J.; Lindner, P.; Tsianou, M.; Alexandridis, P.; Richtering, W. Langmuir 1999, 15, 2599. (25) Schmidt, G.; Mu¨ller, S.; Schmidt, C.; Richtering, W. Rheol. Acta 1999, 38, 486. (26) Bergmeier, M.; Hoffmann, H.; Thunig, C. J. Phys. Chem. B 1997, 101, 5767.
Berghausen et al. (27) Horbaschek, K.; Hoffmann, H.; Hao, J. C. J. Phys. Chem. B 2000, 104, 2781. (28) Jung, H. T.; Coldren, B.; Zasadzinski, J. A.; Iampietro, D. J.; Kaler, E. W. Proc. Natl. Acad. Sci. 2001, 98, 1353. (29) Freund, O.; Mahy, P.; Amedee, J.; Roux, D.; Laversanne, R. J. Microencapsul. 2000, 17, 157. (30) Subramanian, G.; Hjelm, R. P.; Deming, T. J.; Smith, G. S.; Li, Y.; Safinya, C. R. J. Am. Chem. Soc. 2000, 122, 26. (31) Bouglet, G.; Ligoure, C.; Belloq, A. M.; Dufourc, E.; Mosser, G. Phys. ReV. E 1998, 57, 834. (32) Ko¨tz, J.; Tiersch, J.; Bogen, I. Colloid Polym. Sci. 2000, 278, 164. (33) Berger, K.; Hiltrop, K. Colloid Polym. Sci. 1996, 274, 269. (34) Zipfel, J.; Berghausen, J.; Lindner, P.; Richtering, W. J. Phys. Chem. B 1999, 103, 2841. (35) Fujishige, S. Polym. J. 1987, 19, 297. (36) Eisele, M.; Burchard, W. Makromol. Chem. 1990, 191, 169. (37) Zipfel, J.; Berghausen, J.; Schmidt, G.; Lindner, P.; Alexandridis, P.; Tsinaou, M.; Richtering, W. Phys. Chem. Chem. Phys. 1999, 1, 3905. (38) Pusey, P. N.; van Megen, W. Phys. ReV. Lett. 1987, 59, 2083. (39) Berghausen, J. Ph.D. Thesis, University of Freiburg, Germany, 2000. (40) Freyssingeas, E.; Roux, D.; Nallet, F. J. Phys. II France 1997, 7, 913. (41) Nallet, F.; Laversanne, R.; Roux, D. J. Phys. II 1993, 3, 487. (42) Caille´, A. C. R. Acad. Sci. Paris B 1972, 274, 1733. (43) Imai, M.; Nakaya, K.; Kato, T. Eur. Phys. J. E 2001, 5, 391. (44) Nettesheim, F.; Zipfel, J.; Lindnder, P.; Richtering, W. Colloids Surf. A 2001, 183-185, 563. (45) Le´on, A.; Bonn, D.; Meunier, J.; Al-Kahwaji, A.; Greffier, O.; Kellay, H. Phys. ReV. Lett. 2000, 84, 1335. (46) Clement, F.; Joanny, J.-F. J. Phys. II France 1997, 7, 973. (47) Boltenhagen, P.; Kleman, M.; Lavrentovich, O. D. J. Phys. II 1994, 4, 1439. (48) Le, T. D.; Olsson, U.; Mortensen, K.; Zipfel, J.; Richtering, W. Langmuir 2001, 17, 999. (49) Le, T. D.; Olsson, U.; Mortensen, K. Phys. Chem. Chem. Phys. 2001, 3, 1310. (50) Walter, R.; Ricka, J.; Quellet, Ch.; Nyffenegger, R.; Binkert, Th. Macromolecules 1996, 29, 4019. (51) Lee, L.-T.; Cabane, B. Macromolecules 1997, 30, 6559. (52) Fredrickson, G. H.; Bates, F. S. Annu. ReV. Mater. Sci. 1996, 26, 501.