- Email: [email protected]
Dynamic assembly of anionic surfactant into highly-ordered vesicles
Journal of Colloid and Interface Science 356 (2011) 579–588
Contents lists available at ScienceDirect
Journal of Colloid and Interface Science www.elsevier.com/locate/jcis
Dynamic assembly of anionic surfactant into highly-ordered vesicles H. Gevgilili a, D. Kalyon a,⇑, E. Birinci a, M. Malik a, L. Goovaerts c, R. Bacon b, P. Mort b a
Stevens Institute of Technology, Chemical Engineering and Material Science, Castle Point St., Hoboken, NJ 07030, USA Procter & Gamble, 5299 Spring Grove Ave., Cincinnati, OH 45217, USA c Procter & Gamble, Temselaan 100 B 1853, Strombeek-Bever, Belgium b
a r t i c l e
i n f o
Article history: Received 30 July 2010 Accepted 5 January 2011 Available online 11 January 2011 Keywords: Surfactant Alkylbenzene sulfonate Lamellar Vesicle Assembly Gel Rheology
a b s t r a c t A highly-efficient dynamic assembly method for the transformation of the initial spongy lamellar structure of concentrated linear alkylbenzene sulfonate, LAS, incorporated with sodium silicate, into spherulitic vesicles is presented. A combination of drag and pressure flows, via twin screw extrusion, was used to mitigate the ubiquitous viscoplasticity and the wall slip behavior of the anionic surfactant paste and gave rise to the dynamic assembly of stable vesicular nanostructures within a narrow size range, that was not possible with either pure drag or pure pressure flows. Concomitantly with the structure transformation of the paste during assembly under the combination of pressure and drag flows, significant changes in its viscoelasticity, i.e., order of magnitude increases in storage and loss moduli and magnitude of complex viscosity, were observed. The demonstrated dynamic assembly of stable vesicular nanostructures, with vesicle diameters within the relatively narrow range of 300–600 nm, from a commodity surfactant is relevant to myriad templating and encapsulation applications, as well as shedding light on the mechanisms of the deformation-induced planar lamellar to vesicle transformation of concentrated amphiphiles. Ó 2011 Elsevier Inc. All rights reserved.
1. Introduction There has been much recent interest in structured templating for nanofabrication using polymers, surfactants and other amphiphilic materials [1–3]. Such templates can be used to fabricate nanowires, nanotubes, nanospheres, mesoporous membranes, hollow silica and CdS spheres, and polymer particles [4–11]. Furthermore, surfactant or polymer based multi-layered vesicular structures can serve as chemical microreactors as well as being used in catalysis, cosmetics and for templating in pharmacology [12]. Vesicular structures formed upon the self-assembly of various amphiphilic molecules (hydrophilic polar head faces water) are of particular interest due to their obvious relevance to life sciences and biophysics of vesicular bilayer lipid membranes [13]. Vesicles are also applied in the pharmaceutical field as drug release systems, for example, for the encapsulation and release of anthracyclines for cancer [14] and sumatriptan succinate for migraine treatments [15]. Further widespread utilization of vesicular templates for myriad encapsulation and templating tasks requires obtaining inexpensive and stable vesicles with well-defined sizes at industrially relevant rates. Common methods of vesicle generation including sonication, dialysis and reverse phase evaporation generally give
⇑ Corresponding author. Fax: +1 201 216 8306. E-mail address: [email protected] (D. Kalyon). 0021-9797/$ - see front matter Ó 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.jcis.2011.01.011
rise to aggregates that are unstable and highly polydisperse [16]. The formation of relatively uniform multi-layered vesicular structures from highly viscous surfactant pastes is possible but only at relatively low rates involving delicate processes that require the imposition of a carefully-controlled and uniform thermo-mechanical history, i.e., uniform shear and temperature history [17–21]. Previous investigations have not considered the ubiquitous viscoplasticity and associated wall slip behavior of concentrated surfactant pastes as a factor playing a significant role during dynamic assembly of concentrated surfactants under deformation. At thermal equilibrium the geometry of vesicles is defined on the basis of the minimization of the elastic energy, E, upon the bending of a bilayer from its flat configuration as a function of the topology of the surface [22,23]:
E ¼ ðj=2Þ
Z Z
~ ½c1 þ c2 2 dS þ j
Z Z
c1 c2 dS
ð1Þ
~ are the elastic where c1 and c2 are the radii of curvature and j and j bending constant and the Gaussian bending constant, i.e., saddlesplay modulus, respectively [23]. Since the geometries of the bilayers are affected by the flow field applied during the structuring of the amphiphile, there are strong couplings between the nanostructures attained and flow and deformation history that is imposed during structure development [23]. Upon shearing at shear rates which are above a critical shear rate of a few per second lamellar phases become unstable [24–26] and roll into multilamellar
580
H. Gevgilili et al. / Journal of Colloid and Interface Science 356 (2011) 579–588
vesicles, i.e., ‘‘onions’’, in a strain dependent fashion [27–29]. During transformation, the bending energy, Eb, of a multilamellar vesicle with a radius of R:
~ ÞR=d Eb ¼ 4pð2j þ j
ð2Þ
must be balanced against the Taylor estimate of the viscous energy, Ev [27]:
Em ¼ gc_ R3
ð3Þ
where g is the shear viscosity, c_ is the shear rate. This balance indicates that both the resulting radius R, of the vesicles and the smectic distance of separation between different layers d, would be dependent on the true shear rate, c_ , that is imposed during structure formation as: R / c_ 0:5 and d / 1=c_ . The salinity of the medium also ~ [30]. The structure affects the smectic distance, d [27] as well as j changes are accompanied by changes in elasticity and yield stress of the surfactant [31]. Overall, the basic mechanism considered in these earlier fundamental studies of deformation-dependent development of vesicular structures is simple shearing (one component of velocity vector changing only in one other direction with the gradient defining the shear rate). Other deformation mechanisms are not considered, i.e., shear versus extensional and imposition of more than one velocity gradient. Furthermore, it is widely assumed that the apparent shear rate that is applied is equal to the true shear rate that is imposed. As will be shown for the surfactant paste of this study, the apparent and true shear rates are not equal for surfactant pastes that are sufficiently concentrated to exhibit viscoplasticity.
Besides pure drag flow there are two other basic types of steady deformation fields that can be readily imposed on a concentrated surfactant to induce its shear-induced structure transformation into vesicles. These are schematically shown in Fig. 1 in conjunction with a simple geometry, i.e., two infinitely long and infinitely wide plates which are separated by a narrow gap, H, from each other. The first case (Fig. 1a) involves simple steady laminar flow in which one of the plates is moving with a velocity, Vw, and the second is stationary, i.e., the pure drag flow that was generally used in most of the previous -investigations on shear-induced structuring of surfactants [14,18,23,27,32]. The apparent shear rate, c_ a ¼ V w =H becomes equal to the true shear rate that is imposed on the surfactant, c_ only for the no-slip condition. Bergmeier et al. [14] have demonstrated the formation of multilayered vesicular structures upon pure drag flow based shearing of an ionically charged surfactant, i.e., an aqueous system of tetradecyldimethylamineoxide/tetradecyltrimethylammonium bromide/ n-hexanol. They used a Couette flow cell consisting of two concentric cylinders one of which was stationary and the other was rotating. Similarly, other studies [18,23,27,32] have also utilized the Couette cell with narrow gaps (gap radius of the Couette geometry) to generate multi-layered vesicular structures ‘‘onion textures’’ for various surfactant systems. In all these cases the conversion of lamellar to vesicular structures was handicapped by the inherent limitations of the narrow-gapped Couette cell (i.e., difficulties in unloading without structural damage and small deformation volume). Consequently the quantities of surfactants converted were very small, i.e., in the order of grams [18,23,27,32]. These studies [18,23,27,32,33] generally involved
Fig. 1. Three basic flow configurations that can be used for shear-induced conversion of surfactant microstructure. (a) Drag induced flow with zero pressure gradient, i.e., dP/ dz = 0. (b) Pressure-driven flow with a negative pressure gradient, i.e., dP/dz < 0. (c) Combination of drag and pressure flow with both positive and then negative pressure gradients, i.e., dP/dz > or <0, within the same flow domain.
H. Gevgilili et al. / Journal of Colloid and Interface Science 356 (2011) 579–588
relatively dilute systems with typical brine concentrations in the 80–86% or systems which exhibit zero shear viscosity values <10 Pa-s – thus were not expected to exhibit gel-like behavior and viscoplasticity. The second case for imposition of shear involves the steady pressure-driven channel flow, i.e., Poiseuille flow, depicted as flow through a narrow rectangular slit die with stationary walls in Fig. 1b. In pressure flow the pressure gradient is negative (loss of pressure with increasing distance, z). This pressure-driven flow through a channel gives rise to a distribution of shear rates, c_ ðyÞ, with the maximum shear rate occurring at the wall and zero shear rate at the axis of symmetry. McKeown et al. [33] and Liaw et al. [34] have demonstrated that vesicular structures from linear alkylbenzene sulfonate, LAS, can be obtained upon shearing under such a pressure-driven flow. A multipass capillary rheometer (a capillary die connected to two cylinders/plungers which alternate to force the paste repetitively through the capillary die) was used to apply the shear. The difficulty of structuring of this concentrated surfactant paste is evident by the conditions of the pressure- induced flow during which shearing was applied for 1–3 h. Furthermore, this pressure-driven flow typically generated only about 17 g of structured LAS over durations of up to 1 h of shearing [33,34]. A third possible mechanism for applying shear is the combination of drag and pressure flows that is schematically shown in Fig. 1c. The fluid is dragged via the plate moving at velocity, Vw to generate a pressure rise at Section A to overcome the pressure drop generated at Section B. Section B could be a die, a valve or two plates one of which is moving in a direction that is opposite to the bulk motion direction of the fluid. As depicted in Fig. 1c, the drag and pressure flow provides back-mixing, i.e., circulatory flow in Section A, which would enable the mixing of fluid elements with differing previous shear histories. Such a drag and pressure flow can be generated in extrusion flows by combining Archimedean screws found in single or twin screw extrusion together with a die or using combinations of screw elements that give rise to both pressure loss and pressure gain and the associated shearing and backmixing. To our knowledge the combination of drag and pressure flows has not been used for dynamic assembly and structuring of surfactants. The hypothesis of our investigation was that shearing under a combination of drag and pressure flows can be utilized to enable the large-scale dynamic assembly of highly-ordered vesicle gels that is not possible in pure drag or pressure
581
flows. A second hypothesis of our investigation was that the imposition of the combination of drag and pressure flows enables the overcoming of the viscoplasticity and wall slip effects behavior of such concentrated surfactant pastes that would otherwise hinder their deformation and hence dynamic assembly.
2. Results and discussion The insets of Fig. 2 shows the typical cryogenic scanning electron micrographs of the LAS paste in its native clustered lamellar phase state at two different scales of examination. The LAS surfactant, prior to deformation, consists of irregular lamellae clustered with various orientations and shapes encapsulated into a spongy phase. Our aim was the dynamic assembly of this spongy structure, containing irregular lamellae, into well-defined vesicles with a narrow size distribution. Furthermore, we aimed to understand the structural transformation mechanism of the flow field and to document the effects of the changes in structure on rheological behavior, i.e., the characterization of the linear viscoelastic material functions upon small-amplitude oscillatory shear as a function of deformation history. The typical viscoelastic material functions of the surfactant paste prior to deformation, i.e., the storage modulus, G0 , loss modulus, G00 and the magnitude of complex viscosity, g⁄, versus the frequency, x, are shown in Fig. 2. The highly viscous nature of the as-received surfactant paste is suggested by the relatively high values of the magnitude of complex viscosity, g⁄, i.e., 0.3–120 kPa-s for 0.1 6 x 6 100 rps. The values of the storage modulus, G0 , are significantly greater than the values of the loss modulus, G00 . This suggests that the energy stored during deformation as elastic energy is significantly greater than the energy dissipated as heat to reflect the relatively solid-like character of the surfactant paste, prior to the shear-induced transformation of its structure. Furthermore, the storage, G0 , and loss modulus, G00 , versus frequency, x, data are nearly parallel. These observations are indicative of the ‘‘gel-like’’ nature of the paste [35,36] prior to shearing. Such a gel-like behavior of the concentrated surfactant paste would lead into various challenges in its further structuring. The gel is expected to exhibit viscoplastic flow behavior in which the deformation only occurs at stress magnitudes that are greater than the yield stress of the gel. The viscoplasticity would also naturally lead to the slip of the surfactant at the wall [37–39] during
Fig. 2. Dynamic properties versus frequency behavior of LAS surfactant paste in its native spongy/lamellar state prior to the deformation at a strain amplitude of 0.5%.
582
H. Gevgilili et al. / Journal of Colloid and Interface Science 356 (2011) 579–588
attempts at its deformation, leading to further complications in the control of the shearing history of the surfactant. At stress magnitudes that are smaller than the yield stress of the surfactant the process would give rise to plug flow, i.e., forward motion of the paste through wall slip but without deformation and mixing. The viscoplastic behavior of the surfactant gels would also complicate efforts to incorporate particles of sodium silicate into the paste for salinity control since the compounding needs to take place at stress magnitudes that surpass the yield stress of the surfactant paste in order for mixing to be effective.
motion of the surfactant becomes limited to plug flow, subject to wall slip at one or both surfaces as shown in Fig. 3. Similar wall slip and viscoplastic behavior of the LAS paste was observed over a wide range of temperatures and apparent shear rates and no increases of the dynamic properties and no associated changes in viscoelasticity of the paste samples upon pure drag in steady torsional flow could be achieved. Thus, the viscoplasticity and the associated wall slip behavior of the concentrated surfactant paste preclude the possibility of structuring of the paste upon imposition of pure shear.
2.1. Pure drag flow (Fig. 1a): Shearing in steady torsional flow
2.2. Pressure flow (Fig. 1b): Shearing in a rectangular slit die
The viscoplastic nature of the as-received LAS paste of Fig. 2 and the typical challenges associated with its shearing using pure drag flow are demonstrated in Fig. 3. The surfactant was sandwiched in between two parallel disks; one of which rotates and the second is stationary (steady torsional flow akin to the schematic of Fig. 1a). The torque generated on the moving disk at 60 °C as a function of time upon the inception of its rotation provides the shear stress at the edge of the disk, the time dependence of which is shown in Fig. 3. In these steady torsional flow experiments the free surface of the surfactant was trimmed flat and a straight-line marker was placed to cover the free surface of the surfactant and the edges of the two parallel disks to provide guidance on the wall slip behavior of the surfactant [40,41]. Under a no-slip condition the marker line at the free surface of the paste would have been continuous and be connected to the marker lines at the edges of the moving and stationary disks. However, the typical steady torsional flow behavior of the paste involves discontinuities in the marker lines at one or both surfaces (two examples of discontinuities at rheometer surfaces are shown in Fig. 3) suggesting that there is significant wall slip during simple shear flow of the concentrated LAS surfactant. The onset of wall slip is consistent with the decrease of the shear stress observed upon the onset of wall slip. Another interesting aspect of Fig. 3 is the lack of deformation of the concentrated surfactant. Such a lack of deformation is expected of viscoplastic fluids under conditions for which the applied shear stress is smaller than the yield stress of the fluid [39] and is thus consistent with the gel-like nature of the concentrated LAS surfactant evident from Fig. 2. When the shear stress is not sufficiently high to deform the surfactant, the
The second type of deformation that was imposed systematically to the LAS paste of our study over a broad range of apparent shear rates and temperatures was the pressure-driven Poiseuille flow in a rectangular slit die, the velocity distribution of which is schematically depicted in the inset of Fig. 4 in conjunction with a wall slip condition. In our pressure flow experiments the shearing of the LAS paste was achieved by linking a cylindrical reservoir holding the paste to a slit die with an adjustable gap opening [42]. The paste was forced to flow into the slit die via a plunger acting on the reservoir. During the experiments the apparent shear rate, c_ a ¼ 6Q =ðWH2 Þ, where Q is the flow rate, and W is the slit die width, was varied by altering the gap of the slit die, H, at constant Q systematically, i.e. so that the apparent shear rate, c_ a range was: 32 < c_ a < 2941 s1 . The extrudates emerging out of the rectangular slit die at 60 °C were collected and subjected to the analysis of their linear viscoelastic materials functions (Fig. 4). Only a modest increase of the storage modulus, G0 , is observed upon shearing suggesting that the structure of the surfactant has not been significantly altered during pressure flow in the rectangular slit die in spite of the relatively high apparent shear rates applied, i.e., 32–2941 s1. The results of the microscopic analyses of the extrudates were consistent with the rheological characterization results and indicated that there was only little transformation of the lamellar structure of LAS to vesicles. The lamellar structure of the LAS paste was found to be largely intact with some vesicles forming under some conditions as shown in the inset of Fig. 4. These findings were expected on the basis of the significant variations of shear rates within the gap of the slit die, and the radius of the vesicles, R, scaling with
Fig. 3. Wall slip and viscoplasticity of the LAS surfactant paste during steady torsional flow at an apparent shear rate of 0.1 s1 at 60 °C (three different specimens).
H. Gevgilili et al. / Journal of Colloid and Interface Science 356 (2011) 579–588
583
Fig. 4. Changes in the storage modulus, G0 (x), of the LAS surfactant paste upon being subjected to pressure flow in a rectangular slit die at various apparent shear rates at 60 °C. The inset shows the typical cryo-SEM micrograph of the paste upon exit from the die.
c_ 0:5 [23] and the complete lack of shearing that would be associated with the viscoplasticity of the paste, especially in the flow region surrounding the axis of symmetry of the rectangular slit die. 2.3. Pressure and drag flow (Fig. 1c): deformation in twin screw extrusion The twin screw extrusion experiments, involving the drag and pressure flows in tandem, were carried out over a broad range of flow rates and temperatures with the Archimedean screw and slit die combination configuration of Fig. 5. The surfactant paste and sodium silicate particles were introduced into the feed zone in Section 1, where the surfactant is pressurized to overcome the pressure drop through the left-handed screw section at Section 2 (200 psi under the conditions of Fig. 5). The left-handed screw elements of Section 2 generate drag the direction of which is opposite to the bulk flow direction, similar to what is depicted in Fig. 1c. Section 3 consists of right-handed conveying screw elements that generate sufficient pressure to overcome the pressure drop at the
die (140 psi under the conditions of Fig. 5). Thus, pressure is generated at Sections 1 and 3 and pressure is lost at Section 2 and at the die of the extruder. The shear rates imposed on the bulk of the LAS paste vary as a function of location in the extruder, with maxima observed at the gaps between the screw elements and the screw and the die and at the wall of the die. Overall, the shear rates (second invariant of rate of deformation) are typically <300 s1 at the mixing section (Sections 1 and 2), <30 s1 at the die and <150 s1 at the conveying screw section preceding the die (Section 3). The storage modulus versus frequency data of samples collected from different locations in the twin-screw extruder upon a deadstop of the extruder following steady extrusion (at a mass flow rate of 10 kg/h, screw rotational speed of 125 rpm, a slit die gap of 6 mm and walls of the barrel and the die kept at 60 °C) are shown in Fig. 5. Overall, regardless of the location, the storage modulus, G0 , values of the surfactant are significantly greater than the loss modulus, G00 , values. Furthermore, the dynamic properties G0 and G00 are parallel and little affected by the frequency, x. Thus, the gel nature
Fig. 5. Storage modulus, G0 (x) versus frequency behavior of LAS surfactant samples collected from various locations in the twin-screw extruder (at a mass flow rate of 10 kg/ h, screw rotational speed of 125 rpm, a slit die gap of 6 mm and walls of the barrel and the die kept at 60 °C).
584
H. Gevgilili et al. / Journal of Colloid and Interface Science 356 (2011) 579–588
Fig. 6. Magnitude of complex viscosity, g⁄(x) versus frequency behavior of LAS surfactant samples collected from various locations in the twin-screw extruder (at a mass flow rate of 10 kg/h, screw rotational speed of 125 rpm, a slit die gap of 6 mm and walls of the barrel and the die kept at 60 °C).
of the surfactant paste was preserved and the gel was further stiffened in the extruder, as revealed by the increases of the storage modulus, G0 , from about 50 kPa at the entrance to over 500 kPa at the exit of the twin-screw extruder. Furthermore, the magnitude of complex viscosity values of the surfactant also increased by about an order of magnitude as shown in Fig. 6. The comparison of the dynamics properties of the samples deformed in the drag and pressure flow in the extruder with those obtained from the pressure-induced deformation in the rectangular slit die (see for example the storage modulus, G0 , values of Fig. 4 and 5) indicates that the changes in elasticity (storage modulus) and viscosity (magnitude of complex viscosity) after shearing in the pressure flow in the slit die are negligible in comparison to the transformation of the paste in the drag and pressure flow in the twin-screw extruder. This suggests that the structural changes which took place during pressure flow in the slit die over a wide range of shear rates were not as significant as those generated during dynamic assembly in the drag and pressure flow of the
twin-screw extruder. What type of changes in the structure of the surfactant paste would generate such significant increases in elasticity and viscosity of the paste in the extruder? Léon et al. [43] have suggested that significant increases in elasticity of the surfactant paste would occur upon the formation of vesicles from lamellae. Upon vesicle formation the bilayers are shared across adjacent vesicles. Such sharing of bilayers across vesicles should give rise to affine junction points that would form a network to span the volume and stiffen the gel, and thus increase the elasticity as well as the shear viscosity of the vesicular formations. Figs. 7 and 8 show the SEM micrographs of the LAS surfactant paste specimens collected from different axial locations in the twin-screw extruder together with their location-dependent storage modulus and magnitude of complex viscosity values. The frequency is 0.1 rps and the error bars report the 95% confidence intervals determined according to Student’s t-distribution. The increases of the elasticity and the viscosity of the surfactant with
Fig. 7. Storage modulus, G0 , at a frequency of 0.1 rps versus axial distance in the twin-screw extruder and typical SEM micrographs of LAS surfactant samples which reveal the structural transformations that have taken place in the twin-screw extruder.
H. Gevgilili et al. / Journal of Colloid and Interface Science 356 (2011) 579–588
585
Fig. 8. Magnitude of complex viscosity, g⁄, at a frequency of 0.1 rps versus axial distance in the twin-screw extruder and typical SEM micrographs of LAS surfactant samples which reveal the structural transformations that have taken place in the twin-screw extruder.
increasing downchannel distance in the extruder are observed to be associated with its deformation-induced conversion into a vesicular structure, consistent with the findings of Léon et al. on the relatively high elasticity of vesicular structures [43]. The dynamic assembly of the vesicles occurs gradually. First, a portion of the spongy lamellar structure, with characteristic lengths of the clustered lamellae of about 2–20 lm, transforms into an irregular vesicular structure. With increasing downchannel distance the concentration of the vesicles increases (Figs. 7 and 8). The degree of homogeneity of the vesicular structure also increases with increasing downchannel distance. The dynamic assembly of the surfactant paste in the twin-screw extruder finally results in a highly-ordered vesicular structure with vesicle diameters all falling within a narrow 300–600 nm range. The linear viscoelastic properties of the paste were observed not to change during the 1 month upon extru-
sion that they were continued to be tested, suggesting that the vesicular structure that is generated is stable. The change from an irregular spongy lamellar structure into a unimodal, defect-free and elastic vesicular nanostructure is a testament to the efficiency of the combined drag and pressure flow of the twin screw extrusion process for the structural transformation of the surfactant. The simulation of the thermo-mechanical history that the surfactant is exposed to during twin screw extrusion was carried out using the three-dimensional Finite Element Method of Kalyon and Malik [44] and provided additional insight into the mechanisms of the transformation of the structure of the surfactant. The determination of the z-velocity (velocity component, integration of which provides the net flow rate in the extruder) distributions (Fig. 9) indicated that there are both negative and positive z-velocity values at various zones in the extruder suggesting the
Fig. 9. The FEM based simulated distributions of velocity in pressure and drag flow (z-velocity, i.e., main velocity component integration of which provides the volume flow rate) showing the presence of both positive and negative velocities and hence the significant backmixing occurring.
586
H. Gevgilili et al. / Journal of Colloid and Interface Science 356 (2011) 579–588
formation of pressure and drag flows akin to what is depicted in Fig. 1c. The resulting pressure gain in the right-handed screw elements (Sections 1 and 3) and pressure loss at the reversely staggered screws section (Section 2) and the die give rise to the back-mixing of the paste (drag and pressure flow depicted in Fig. 1c) thus homogenizing the deformation history and the resulting structure of the paste. The pressure was determined to increase monotonically in the right-handed screw elements preceding the left-handed screw elements (Section 1 in Fig. 5) and preceding the die (Section 3 in Fig. 5). On the other hand, the pressure decreases over the reversely staggered screw section (Section 2 in Fig. 5) and the die. Concomitant analysis of the linear viscoelastic properties and the simulated pressure distribution in the extruder suggests that significant transformations in structure (as evidenced by scanning electron microscopy and a significant change in the storage modulus, G0 , and magnitude of complex viscosity, g⁄, values) occur only at locations where the pressure gradient is positive and especially at locations where the positive pressure gradient is relatively high. On the other hand, the shearing of the paste in the partially full sections of the extruder, i.e., conveying action of the right-handed screw elements without generating pressure, does not lead to a significant change in structure (the G0 and magnitude of complex viscosity, g⁄, values remain invariant over the partially full section of the twinscrew extruder). Furthermore, further assembly of the vesicles into a more homogeneous structure and hence increase of G0 and g⁄ also do not occur at pressure-losing elements, i.e., at the die and within the confines of the left-handed screw elements (Section 2) although significant shearing occurs at these locations also. These observations should significantly alter the conventional thinking on which processing parameters can be used to realize the lamellar to vesicle
transformation of surfactants. Clearly, there are other factors than the ‘‘shear rate’’ or ‘‘strain’’ that were considered earlier to be the only factors in processing and the conversion of the surfactant structure into vesicles [29,45,46]. It appears that a combination of pressure, shear and backmixing is critical in achieving high-rate dynamic assembly of the vesicles of the anionic surfactant. The pressure and drag flows are particularly appropriate to overcome the wall slip and viscoplasticity limitations of the surfactant paste to allow the transformation of the structure. Experimental and theoretical evidence from polymeric melts and polymeric suspensions suggests that the positive pressure gradient in the flow direction reduces the slip velocity at the wall boundary in the flow direction [47,48]. The reduced wall slip velocities should impose a more uniform shear rate gradient and the positive pressure gradient, to give rise to backmixing which would reduce structural inhomogeneities. Schmidt et al. [49] have shown that the vesicles of a surfactant can be degraded into perpendicularlyaligned lamellae when the shear stress exceeds a critical shear stress. Although a significant range of stress magnitudes (through both shear and extensional types of deformations) were imposed on the LAS paste in the twin-screw extruder, the applied stresses have not degraded the closely packed vesicular structure of the resulting paste (Fig. 10). This is a precursor of the stability of the vesicular structure, which was demonstrated by the stability of the linear viscoelastic material functions with time during the 1month test period following extrusion. The observed structuring of the surfactant paste is temperature dependent. As the temperature of the surfactant paste in the twinscrew extruder increases to over 65 °C the rate of structure formation decreases below that was observed at 60 °C (as indicated by lower storage modulus, G0 , values). These findings are consistent
Fig. 10. The schematics of the transformation of the structure of the concentrated surfactant paste from spongy lamellar to vesicular (insets) upon the combination of pressure and drag flows. The significant backmixing provides a narrow size distribution.
H. Gevgilili et al. / Journal of Colloid and Interface Science 356 (2011) 579–588
with the experimental findings of McKeown et al. [33] and Liaw et al. [34] which indicated that the lamellae to vesicle conversion of the LAS surfactant (Fig. 10) only becomes significant over a narrow temperature window and the rate of formation of onion like multilayered vesicles depends exponentially on temperature. These findings emphasize the importance of temperature selection and control during the dynamic assembly of the vesicles. It is expected that the sodium silicate particles compounded into the surfactant paste would be dissolved in the water phase with the rate of dissolution affected favorably by the relatively high surface to volume ratios of the colloidal sodium silicate. The dissolution of the sodium silicate would alter the salinity and thus the topology of the bilayers [43]. The energy, E, associated with the formation of spherical vesicles consisting of onion like configuration of surfactant bilayers from flat lamellar bilayer depicted in Fig. 10 is given as [22,50]:
~Þ E ¼ 4pð2j þ j
ð4Þ
~ are the membrane elastic constants referred to as where j and j bending rigidity and saddle-splay radii. Spontaneous formation of ~ < 2j. Léon et al. [43] have determined vesicles is predicted for j that at least for the aqueous solution of sodium bis(2-ethylhexyl) sulfosuccinate (AOT) at a surfactant concentration of 7%, j is unaf~ is affected. Liaw et al. [34] fected by the addition of salts, whereas j suggest that the dissolution of sodium disilicate might give rise to a decrease of the bilayer repeat distance, and/or raise the ionic strength of the water present in the LAS bilayers. This would alter the size of the head-group of the LAS surfactant or change the electrostatic forces within and between layers to thus presumably alter the membrane elastic constant to enable the formation of onion like multilayered vesicles which are interconnected to each other to provide the significant increase in elasticity and viscosity (Fig. 10).
3. Experimental 3.1. Materials A highly-viscous concentrated surfactant, i.e., a linear alkylbenzene sulfonate LAS-water paste, reported earlier by Liaw et al. [34] was used. This paste is industrially referred to as a ‘‘high active’’ LAS paste of ca. wt. 78% of linear alkylbenzene sulfonate salts with the rest water, with LAS consisting of ‘‘lamellar bilayer structures’’ [34]. The chemical composition and properties of a typical LAS batch were reported by Liaw et al. [34] and McKeown [51] and the composition includes 77.2% by weight active matter, 19.1% water, 1.65% of linear alkylbenzene, 0.42% of Na2SO4, 0.37% of NaOH and 0.16% of NaCl. Here, the LAS paste was incorporated with 1% by volume of sodium silicate nanoparticles, 50–300 nm. The paste was provided by Procter & Gamble.
3.2. Characterization of the LAS paste The microstructure of the LAS paste prior to and upon deformation was characterized microscopically using cryogenic SEM (Hitachi S-4300). The LAS paste was characterized macroscopically using small-amplitude oscillatory shear (using an ARES rheometer from TA Instruments). The data collected in the small-amplitude oscillatory shear included the storage, G0 , and loss, G00 , moduli and the magnitude of the complex viscosity, |g⁄|, as a function of strain amplitude and frequency, x. Strain sweeps were carried out to determine the strain amplitude ranges over which the paste exhibited linear viscoelastic behavior. Typically, the paste exhibited linear behavior at strain amplitudes that are smaller than 1%.
587
3.3. Pressure and drag flow: twin screw extrusion A 50.8 mm fully-intermeshing co-rotating twin-screw extruder, TSE, was used in conjunction with a rectangular slit die with an adjustable gap [42]. The adjustable gap slit die was equipped with four pressure transducers to allow the determination of the pressure distribution and hence the wall shear stress in the die. The temperature of the surfactant was determined with a series of thermocouples installed flush with the die and the extruder barrel surfaces and using a thermal imaging camera to record the temperature distribution of the surfactant strand immediately upon exit from the die. The twin-screw extruder was also equipped with pressure transducers to allow the determination of the degree of fill and the pressure distribution in the twin-screw extruder. During the twin screw extrusion experiments the concentrated LAS (at controlled water content) was fed into the extruder, followed by the introduction of the dry silicate particles. Upon reaching steady state in temperature and pressure distributions the flow was brought to a dead-stop, the barrel halves were split open to give access to the surfactant paste in the extruder and samples of the surfactant paste were collected as a function of location in the extruder and the die. The linear viscoelastic properties of the paste samples were characterized immediately upon extrusion and repetitively over a period of one month to assess the stability of the pastes. The calculations of the temperature distributions in the twinscrew extruder (using the simulation methodologies described by Kalyon and Malik [44]) indicated that there is very little viscous energy dissipation occurring in the confines of the twin-screw extruder under the geometry and operating conditions employed, and the structuring operation can be considered to occur under isothermal conditions. The wall temperature profiles and the heat transfer through the barrel and die surfaces thus dictated the temperature distribution in the extruder, with little variability. This was also verified using an inframetrics thermal imaging camera which was used to monitor the distributions of the temperature of the surfactant extrudates upon emerging from the die (crosssectional areas of which were exposed upon cutting of the extrudates in the transverse to flow direction). 4. Conclusions Three different basic flow mechanisms were applied to the investigation of the transformation of the structure of a concentrated amphiphilic LAS surfactant paste, i.e., pure drag flow, pure pressure flow and combination of pressure and drag flow. Only pure drag and pure pressure flows have been utilized in the past for the shear-induced structuring of surfactants [18,23,27,32,33]. The transformation of the structure of the LAS surfactant from spongy lamellar to a relatively uniform vesicular structure only occurred when a combination of pressure and drag flows was used. The rheological characterization of the surfactant paste revealed that the paste was a viscoplastic gel which exhibited significant wall slip during attempts to deform it. The combination of the pressure and drag flow and the associated pressure increase minimized the wall slip and viscoplasticity effects to allow the deformation of the paste to induce the buckling of the lamellar bilayers and the concomitant transformation of the structure into multilayered vesicles. The documented backmixing occurring in the combination of pressure and drag flow gave rise to the dampening of the differences in thermo-mechanical histories of the different volumes of the paste, thus homogenizing the structure to generate vesicles within the narrow diameter range of 300– 600 nm. The formation and increasing homogeneity of the vesicular domains gave rise to significant increases in various rheological material functions including the storage modulus, G0 , and magni-
588
H. Gevgilili et al. / Journal of Colloid and Interface Science 356 (2011) 579–588
tude of complex viscosity, g⁄, presumably due to the interconnection of the vesicles to each other during deformation and buckling of the bilayers. Overall, it was demonstrated that the dynamic assembly of surfactant bilayers into uniform vesicular structures of concentrated surfactant pastes is possible when a combination of pressure and drag flows is utilized. Furthermore, it was recognized that the structuring of such pastes is affected by viscoplasticity, wall slip, pressure and backmixing, i.e., all parameters not recognized in earlier studies investigating the effects of shearing on structure formation of surfactants. Acknowledgments We thank Ms. C. Kong of P&G for her contributions to scanning electron microscopy of the specimens and to Mr. E. Demirkol, Dr. C. Erisken and Dr. N. Degirmenbasi of Stevens for their contributions to extrusion experiments and sample analysis and Dr. J. Carnali for his input. The funding which made this study possible was provided by P&G, for which we are grateful. References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12]
T. Zemh, M. Dubois, B. Deme, T. Krywicki, Science 283 (1999) 816. D. Discher, A. Eisenberg, Science 297 (2002) 967. Y. Ding, Z. Gui, J. Zhu, S. Yan, J. Liu, Y. Hu, Z. Wang, Mater. Lett. 61 (2007) 2195. J. Hotz, W. Meier, Adv. Mater. 10 (1998) 1387. X. Jiang, Y. Xie, J. Lu, L. Zhu, W. He, Y. Qian, J. Mater. Chem. 11 (2001) 1775. H. Hentze, S.R. Raghavan, C. McKelvey, E.W. Kaler, Langmuir 19 (2003) 1069. S. Che, K. Lund, T. Tatsumi, S. Iijima, S.H. Joo, R. Ryoo, O. Terasaki, Angew. Chem. Int. Ed. 42 (2003) 2182. L. Zhu, X. Zheng, X. Liu, X. Zhang, Y. Xie, J. Colloid and Interface Sci. 273 (2004) 155. A. Yamaguchi, F. Uejo, T. Yoda, T. Uchida, Y. Tanamura, T. Yamashita, N. Teramae, Nat. Mater. 3 (2004) 337. D. Ding, Z. Chen, S. Rajaputra, V. Singh, Sens. Actuators B 24 (2007) 12. A. Walcarius, E. Sibottier, M. Etienne, J. Ghanbaja, Nat. Mater. 6 (2007) 602. F. Gauffre, D. Roux, Langmuir 15 (1999) 3738.
[13] J.W. Szostak, D. Bartel, P.L. Luisi, Nature 409 (2001) 387. [14] M. Bergmeier, M. Gradzielski, H. Hoffmann, K. Mortensen, J. Phys. Chem. B 103 (1999) 1605. [15] M. Redkar, P. Hassan, V. Aswal, P. Devarajan, J. Pharm. Sci. 96 (2007) 2436. [16] I. Grillo, E. Kats, A.R. Muratov, Langmuir 19 (2003) 4573. [17] V. Schmitt, C.M. Marques, F. Lequeux, Phys. Rev. E 52 (1995) 4009. [18] P. Sierro, D. Roux, Phys. Rev. Lett. 78 (1997) 1496. [19] W. Richtering, Current Opin. Colloid Interface Sci. 6 (2001) 446. [20] K. Mortensen, Science 6 (2001) 140. [21] M. Gradzielski, J. Phys.: Condes. Matter 15 (2003) 655. [22] W. Helfrich, Z. Naturforsch. 28C (1973) 693. [23] L. Courbin, P. Panizza, Phys. Rev. E69 (2004) 021504-1. [24] A. Zilman, R. Granek, Eur. Phys. J. B 11 (1999) 593. [25] A. Wunenburger, A. Colin, T. Colin, D. Roux, Eur. Phys. J. E 2 (2000) 27. [26] S. Marlow, P. Olmsted, Eur. Phys. J. E 8 (2002) 485. [27] O. Diat, D. Roux, J. Phys. II 3 (1993) 9. [28] O. Diat, D. Roux, F. Nallet, Phy. Rev. E 51 (1995) 3296. [29] L. Courbin, J. Delville, J. Rouch, P. Panizza, Phys. Rev. Lett. 89 (2002) 148305-1. [30] T. Le, U. Olsson, K. Mortensen, Physica B 276–278 (2000) 379. [31] M. Gradzielski, M. Müller, M. Bergmeier, H. Hoffmann, E. Hoinkis, J. Phys. Chem. B 103 (1999) 1416. [32] J. Zipfel, P. Lindner, W. Richtering, Prog. Colloid Polym. Sci. 110 (1998) 139. [33] S.A. McKeown, M.R. Mackley, G.D. Moggridge, Trans. I. Chem. E 81A (2003) 649. [34] M.S. Liaw, M.R. Mackley, J. Bridgwater, AIChE J. 49 (2003) 2966. [35] F. Chambon, H. Winter, J. Rheol. 31 (1987) 683. [36] M.E. DeRosa, H.H. Winter, Acta 33 (1994) 220–237. [37] U. Yilmazer, D.M. Kalyon, J. Rheol. 33 (1989) 1197. [38] D. Kalyon, H. Gevgilili, R. Yazici, A. Post, G. McFann, Rheol. Acta 43 (2004) 396. [39] D. Kalyon, J. Rheol. 49 (2005) 621. [40] H. Gevgilili, D. Kalyon, J. Rheol. 45 (2001) 467. [41] D. Kalyon, H. Gevgilili, J. Rheol. 47 (2003) 683. [42] D. Kalyon, H. Gokturk, US Pat., 5 277 058, 1994. [43] A. Léon, D. Bonn, J. Meunier, A. Al-Kahwaji, O. Greffier, H. Kellay, Phys. Rev. Lett. 84 (2000) 1335. [44] D.M. Kalyon, M. Malik, Int. Polym. Proc. 22 (2007) 293. [45] J. Bergenholtz, N.J. Wagner, Langmuir 12 (1996) 3122. [46] J. Zipfel, F. Nettesheim, P. Lindner, T.D. Le, U. Olsson, W. Richtering, Europhys. Lett. 53 (2001) 335. [47] S.G. Hatzikiriakos, J.M. Dealy, J. Rheol. 36 (1992) 703. [48] H. Tang, D.M. Kalyon, J. Rheol. 52 (2008) 507. [49] G. Schmidt, S. Müller, C. Schmidt, W. Richtering, Rheol. Acta 38 (1999) 486. [50] T. Le, U. Olsson, K. Mortensen, Phys. Chem. Chem. Phys. 3 (2001) 1310. [51] S.A. McKeown, PhD Thesis, University of Cambridge, UK, 2001.
Contents lists available at ScienceDirect
Journal of Colloid and Interface Science www.elsevier.com/locate/jcis
Dynamic assembly of anionic surfactant into highly-ordered vesicles H. Gevgilili a, D. Kalyon a,⇑, E. Birinci a, M. Malik a, L. Goovaerts c, R. Bacon b, P. Mort b a
Stevens Institute of Technology, Chemical Engineering and Material Science, Castle Point St., Hoboken, NJ 07030, USA Procter & Gamble, 5299 Spring Grove Ave., Cincinnati, OH 45217, USA c Procter & Gamble, Temselaan 100 B 1853, Strombeek-Bever, Belgium b
a r t i c l e
i n f o
Article history: Received 30 July 2010 Accepted 5 January 2011 Available online 11 January 2011 Keywords: Surfactant Alkylbenzene sulfonate Lamellar Vesicle Assembly Gel Rheology
a b s t r a c t A highly-efficient dynamic assembly method for the transformation of the initial spongy lamellar structure of concentrated linear alkylbenzene sulfonate, LAS, incorporated with sodium silicate, into spherulitic vesicles is presented. A combination of drag and pressure flows, via twin screw extrusion, was used to mitigate the ubiquitous viscoplasticity and the wall slip behavior of the anionic surfactant paste and gave rise to the dynamic assembly of stable vesicular nanostructures within a narrow size range, that was not possible with either pure drag or pure pressure flows. Concomitantly with the structure transformation of the paste during assembly under the combination of pressure and drag flows, significant changes in its viscoelasticity, i.e., order of magnitude increases in storage and loss moduli and magnitude of complex viscosity, were observed. The demonstrated dynamic assembly of stable vesicular nanostructures, with vesicle diameters within the relatively narrow range of 300–600 nm, from a commodity surfactant is relevant to myriad templating and encapsulation applications, as well as shedding light on the mechanisms of the deformation-induced planar lamellar to vesicle transformation of concentrated amphiphiles. Ó 2011 Elsevier Inc. All rights reserved.
1. Introduction There has been much recent interest in structured templating for nanofabrication using polymers, surfactants and other amphiphilic materials [1–3]. Such templates can be used to fabricate nanowires, nanotubes, nanospheres, mesoporous membranes, hollow silica and CdS spheres, and polymer particles [4–11]. Furthermore, surfactant or polymer based multi-layered vesicular structures can serve as chemical microreactors as well as being used in catalysis, cosmetics and for templating in pharmacology [12]. Vesicular structures formed upon the self-assembly of various amphiphilic molecules (hydrophilic polar head faces water) are of particular interest due to their obvious relevance to life sciences and biophysics of vesicular bilayer lipid membranes [13]. Vesicles are also applied in the pharmaceutical field as drug release systems, for example, for the encapsulation and release of anthracyclines for cancer [14] and sumatriptan succinate for migraine treatments [15]. Further widespread utilization of vesicular templates for myriad encapsulation and templating tasks requires obtaining inexpensive and stable vesicles with well-defined sizes at industrially relevant rates. Common methods of vesicle generation including sonication, dialysis and reverse phase evaporation generally give
⇑ Corresponding author. Fax: +1 201 216 8306. E-mail address: [email protected] (D. Kalyon). 0021-9797/$ - see front matter Ó 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.jcis.2011.01.011
rise to aggregates that are unstable and highly polydisperse [16]. The formation of relatively uniform multi-layered vesicular structures from highly viscous surfactant pastes is possible but only at relatively low rates involving delicate processes that require the imposition of a carefully-controlled and uniform thermo-mechanical history, i.e., uniform shear and temperature history [17–21]. Previous investigations have not considered the ubiquitous viscoplasticity and associated wall slip behavior of concentrated surfactant pastes as a factor playing a significant role during dynamic assembly of concentrated surfactants under deformation. At thermal equilibrium the geometry of vesicles is defined on the basis of the minimization of the elastic energy, E, upon the bending of a bilayer from its flat configuration as a function of the topology of the surface [22,23]:
E ¼ ðj=2Þ
Z Z
~ ½c1 þ c2 2 dS þ j
Z Z
c1 c2 dS
ð1Þ
~ are the elastic where c1 and c2 are the radii of curvature and j and j bending constant and the Gaussian bending constant, i.e., saddlesplay modulus, respectively [23]. Since the geometries of the bilayers are affected by the flow field applied during the structuring of the amphiphile, there are strong couplings between the nanostructures attained and flow and deformation history that is imposed during structure development [23]. Upon shearing at shear rates which are above a critical shear rate of a few per second lamellar phases become unstable [24–26] and roll into multilamellar
580
H. Gevgilili et al. / Journal of Colloid and Interface Science 356 (2011) 579–588
vesicles, i.e., ‘‘onions’’, in a strain dependent fashion [27–29]. During transformation, the bending energy, Eb, of a multilamellar vesicle with a radius of R:
~ ÞR=d Eb ¼ 4pð2j þ j
ð2Þ
must be balanced against the Taylor estimate of the viscous energy, Ev [27]:
Em ¼ gc_ R3
ð3Þ
where g is the shear viscosity, c_ is the shear rate. This balance indicates that both the resulting radius R, of the vesicles and the smectic distance of separation between different layers d, would be dependent on the true shear rate, c_ , that is imposed during structure formation as: R / c_ 0:5 and d / 1=c_ . The salinity of the medium also ~ [30]. The structure affects the smectic distance, d [27] as well as j changes are accompanied by changes in elasticity and yield stress of the surfactant [31]. Overall, the basic mechanism considered in these earlier fundamental studies of deformation-dependent development of vesicular structures is simple shearing (one component of velocity vector changing only in one other direction with the gradient defining the shear rate). Other deformation mechanisms are not considered, i.e., shear versus extensional and imposition of more than one velocity gradient. Furthermore, it is widely assumed that the apparent shear rate that is applied is equal to the true shear rate that is imposed. As will be shown for the surfactant paste of this study, the apparent and true shear rates are not equal for surfactant pastes that are sufficiently concentrated to exhibit viscoplasticity.
Besides pure drag flow there are two other basic types of steady deformation fields that can be readily imposed on a concentrated surfactant to induce its shear-induced structure transformation into vesicles. These are schematically shown in Fig. 1 in conjunction with a simple geometry, i.e., two infinitely long and infinitely wide plates which are separated by a narrow gap, H, from each other. The first case (Fig. 1a) involves simple steady laminar flow in which one of the plates is moving with a velocity, Vw, and the second is stationary, i.e., the pure drag flow that was generally used in most of the previous -investigations on shear-induced structuring of surfactants [14,18,23,27,32]. The apparent shear rate, c_ a ¼ V w =H becomes equal to the true shear rate that is imposed on the surfactant, c_ only for the no-slip condition. Bergmeier et al. [14] have demonstrated the formation of multilayered vesicular structures upon pure drag flow based shearing of an ionically charged surfactant, i.e., an aqueous system of tetradecyldimethylamineoxide/tetradecyltrimethylammonium bromide/ n-hexanol. They used a Couette flow cell consisting of two concentric cylinders one of which was stationary and the other was rotating. Similarly, other studies [18,23,27,32] have also utilized the Couette cell with narrow gaps (gap radius of the Couette geometry) to generate multi-layered vesicular structures ‘‘onion textures’’ for various surfactant systems. In all these cases the conversion of lamellar to vesicular structures was handicapped by the inherent limitations of the narrow-gapped Couette cell (i.e., difficulties in unloading without structural damage and small deformation volume). Consequently the quantities of surfactants converted were very small, i.e., in the order of grams [18,23,27,32]. These studies [18,23,27,32,33] generally involved
Fig. 1. Three basic flow configurations that can be used for shear-induced conversion of surfactant microstructure. (a) Drag induced flow with zero pressure gradient, i.e., dP/ dz = 0. (b) Pressure-driven flow with a negative pressure gradient, i.e., dP/dz < 0. (c) Combination of drag and pressure flow with both positive and then negative pressure gradients, i.e., dP/dz > or <0, within the same flow domain.
H. Gevgilili et al. / Journal of Colloid and Interface Science 356 (2011) 579–588
relatively dilute systems with typical brine concentrations in the 80–86% or systems which exhibit zero shear viscosity values <10 Pa-s – thus were not expected to exhibit gel-like behavior and viscoplasticity. The second case for imposition of shear involves the steady pressure-driven channel flow, i.e., Poiseuille flow, depicted as flow through a narrow rectangular slit die with stationary walls in Fig. 1b. In pressure flow the pressure gradient is negative (loss of pressure with increasing distance, z). This pressure-driven flow through a channel gives rise to a distribution of shear rates, c_ ðyÞ, with the maximum shear rate occurring at the wall and zero shear rate at the axis of symmetry. McKeown et al. [33] and Liaw et al. [34] have demonstrated that vesicular structures from linear alkylbenzene sulfonate, LAS, can be obtained upon shearing under such a pressure-driven flow. A multipass capillary rheometer (a capillary die connected to two cylinders/plungers which alternate to force the paste repetitively through the capillary die) was used to apply the shear. The difficulty of structuring of this concentrated surfactant paste is evident by the conditions of the pressure- induced flow during which shearing was applied for 1–3 h. Furthermore, this pressure-driven flow typically generated only about 17 g of structured LAS over durations of up to 1 h of shearing [33,34]. A third possible mechanism for applying shear is the combination of drag and pressure flows that is schematically shown in Fig. 1c. The fluid is dragged via the plate moving at velocity, Vw to generate a pressure rise at Section A to overcome the pressure drop generated at Section B. Section B could be a die, a valve or two plates one of which is moving in a direction that is opposite to the bulk motion direction of the fluid. As depicted in Fig. 1c, the drag and pressure flow provides back-mixing, i.e., circulatory flow in Section A, which would enable the mixing of fluid elements with differing previous shear histories. Such a drag and pressure flow can be generated in extrusion flows by combining Archimedean screws found in single or twin screw extrusion together with a die or using combinations of screw elements that give rise to both pressure loss and pressure gain and the associated shearing and backmixing. To our knowledge the combination of drag and pressure flows has not been used for dynamic assembly and structuring of surfactants. The hypothesis of our investigation was that shearing under a combination of drag and pressure flows can be utilized to enable the large-scale dynamic assembly of highly-ordered vesicle gels that is not possible in pure drag or pressure
581
flows. A second hypothesis of our investigation was that the imposition of the combination of drag and pressure flows enables the overcoming of the viscoplasticity and wall slip effects behavior of such concentrated surfactant pastes that would otherwise hinder their deformation and hence dynamic assembly.
2. Results and discussion The insets of Fig. 2 shows the typical cryogenic scanning electron micrographs of the LAS paste in its native clustered lamellar phase state at two different scales of examination. The LAS surfactant, prior to deformation, consists of irregular lamellae clustered with various orientations and shapes encapsulated into a spongy phase. Our aim was the dynamic assembly of this spongy structure, containing irregular lamellae, into well-defined vesicles with a narrow size distribution. Furthermore, we aimed to understand the structural transformation mechanism of the flow field and to document the effects of the changes in structure on rheological behavior, i.e., the characterization of the linear viscoelastic material functions upon small-amplitude oscillatory shear as a function of deformation history. The typical viscoelastic material functions of the surfactant paste prior to deformation, i.e., the storage modulus, G0 , loss modulus, G00 and the magnitude of complex viscosity, g⁄, versus the frequency, x, are shown in Fig. 2. The highly viscous nature of the as-received surfactant paste is suggested by the relatively high values of the magnitude of complex viscosity, g⁄, i.e., 0.3–120 kPa-s for 0.1 6 x 6 100 rps. The values of the storage modulus, G0 , are significantly greater than the values of the loss modulus, G00 . This suggests that the energy stored during deformation as elastic energy is significantly greater than the energy dissipated as heat to reflect the relatively solid-like character of the surfactant paste, prior to the shear-induced transformation of its structure. Furthermore, the storage, G0 , and loss modulus, G00 , versus frequency, x, data are nearly parallel. These observations are indicative of the ‘‘gel-like’’ nature of the paste [35,36] prior to shearing. Such a gel-like behavior of the concentrated surfactant paste would lead into various challenges in its further structuring. The gel is expected to exhibit viscoplastic flow behavior in which the deformation only occurs at stress magnitudes that are greater than the yield stress of the gel. The viscoplasticity would also naturally lead to the slip of the surfactant at the wall [37–39] during
Fig. 2. Dynamic properties versus frequency behavior of LAS surfactant paste in its native spongy/lamellar state prior to the deformation at a strain amplitude of 0.5%.
582
H. Gevgilili et al. / Journal of Colloid and Interface Science 356 (2011) 579–588
attempts at its deformation, leading to further complications in the control of the shearing history of the surfactant. At stress magnitudes that are smaller than the yield stress of the surfactant the process would give rise to plug flow, i.e., forward motion of the paste through wall slip but without deformation and mixing. The viscoplastic behavior of the surfactant gels would also complicate efforts to incorporate particles of sodium silicate into the paste for salinity control since the compounding needs to take place at stress magnitudes that surpass the yield stress of the surfactant paste in order for mixing to be effective.
motion of the surfactant becomes limited to plug flow, subject to wall slip at one or both surfaces as shown in Fig. 3. Similar wall slip and viscoplastic behavior of the LAS paste was observed over a wide range of temperatures and apparent shear rates and no increases of the dynamic properties and no associated changes in viscoelasticity of the paste samples upon pure drag in steady torsional flow could be achieved. Thus, the viscoplasticity and the associated wall slip behavior of the concentrated surfactant paste preclude the possibility of structuring of the paste upon imposition of pure shear.
2.1. Pure drag flow (Fig. 1a): Shearing in steady torsional flow
2.2. Pressure flow (Fig. 1b): Shearing in a rectangular slit die
The viscoplastic nature of the as-received LAS paste of Fig. 2 and the typical challenges associated with its shearing using pure drag flow are demonstrated in Fig. 3. The surfactant was sandwiched in between two parallel disks; one of which rotates and the second is stationary (steady torsional flow akin to the schematic of Fig. 1a). The torque generated on the moving disk at 60 °C as a function of time upon the inception of its rotation provides the shear stress at the edge of the disk, the time dependence of which is shown in Fig. 3. In these steady torsional flow experiments the free surface of the surfactant was trimmed flat and a straight-line marker was placed to cover the free surface of the surfactant and the edges of the two parallel disks to provide guidance on the wall slip behavior of the surfactant [40,41]. Under a no-slip condition the marker line at the free surface of the paste would have been continuous and be connected to the marker lines at the edges of the moving and stationary disks. However, the typical steady torsional flow behavior of the paste involves discontinuities in the marker lines at one or both surfaces (two examples of discontinuities at rheometer surfaces are shown in Fig. 3) suggesting that there is significant wall slip during simple shear flow of the concentrated LAS surfactant. The onset of wall slip is consistent with the decrease of the shear stress observed upon the onset of wall slip. Another interesting aspect of Fig. 3 is the lack of deformation of the concentrated surfactant. Such a lack of deformation is expected of viscoplastic fluids under conditions for which the applied shear stress is smaller than the yield stress of the fluid [39] and is thus consistent with the gel-like nature of the concentrated LAS surfactant evident from Fig. 2. When the shear stress is not sufficiently high to deform the surfactant, the
The second type of deformation that was imposed systematically to the LAS paste of our study over a broad range of apparent shear rates and temperatures was the pressure-driven Poiseuille flow in a rectangular slit die, the velocity distribution of which is schematically depicted in the inset of Fig. 4 in conjunction with a wall slip condition. In our pressure flow experiments the shearing of the LAS paste was achieved by linking a cylindrical reservoir holding the paste to a slit die with an adjustable gap opening [42]. The paste was forced to flow into the slit die via a plunger acting on the reservoir. During the experiments the apparent shear rate, c_ a ¼ 6Q =ðWH2 Þ, where Q is the flow rate, and W is the slit die width, was varied by altering the gap of the slit die, H, at constant Q systematically, i.e. so that the apparent shear rate, c_ a range was: 32 < c_ a < 2941 s1 . The extrudates emerging out of the rectangular slit die at 60 °C were collected and subjected to the analysis of their linear viscoelastic materials functions (Fig. 4). Only a modest increase of the storage modulus, G0 , is observed upon shearing suggesting that the structure of the surfactant has not been significantly altered during pressure flow in the rectangular slit die in spite of the relatively high apparent shear rates applied, i.e., 32–2941 s1. The results of the microscopic analyses of the extrudates were consistent with the rheological characterization results and indicated that there was only little transformation of the lamellar structure of LAS to vesicles. The lamellar structure of the LAS paste was found to be largely intact with some vesicles forming under some conditions as shown in the inset of Fig. 4. These findings were expected on the basis of the significant variations of shear rates within the gap of the slit die, and the radius of the vesicles, R, scaling with
Fig. 3. Wall slip and viscoplasticity of the LAS surfactant paste during steady torsional flow at an apparent shear rate of 0.1 s1 at 60 °C (three different specimens).
H. Gevgilili et al. / Journal of Colloid and Interface Science 356 (2011) 579–588
583
Fig. 4. Changes in the storage modulus, G0 (x), of the LAS surfactant paste upon being subjected to pressure flow in a rectangular slit die at various apparent shear rates at 60 °C. The inset shows the typical cryo-SEM micrograph of the paste upon exit from the die.
c_ 0:5 [23] and the complete lack of shearing that would be associated with the viscoplasticity of the paste, especially in the flow region surrounding the axis of symmetry of the rectangular slit die. 2.3. Pressure and drag flow (Fig. 1c): deformation in twin screw extrusion The twin screw extrusion experiments, involving the drag and pressure flows in tandem, were carried out over a broad range of flow rates and temperatures with the Archimedean screw and slit die combination configuration of Fig. 5. The surfactant paste and sodium silicate particles were introduced into the feed zone in Section 1, where the surfactant is pressurized to overcome the pressure drop through the left-handed screw section at Section 2 (200 psi under the conditions of Fig. 5). The left-handed screw elements of Section 2 generate drag the direction of which is opposite to the bulk flow direction, similar to what is depicted in Fig. 1c. Section 3 consists of right-handed conveying screw elements that generate sufficient pressure to overcome the pressure drop at the
die (140 psi under the conditions of Fig. 5). Thus, pressure is generated at Sections 1 and 3 and pressure is lost at Section 2 and at the die of the extruder. The shear rates imposed on the bulk of the LAS paste vary as a function of location in the extruder, with maxima observed at the gaps between the screw elements and the screw and the die and at the wall of the die. Overall, the shear rates (second invariant of rate of deformation) are typically <300 s1 at the mixing section (Sections 1 and 2), <30 s1 at the die and <150 s1 at the conveying screw section preceding the die (Section 3). The storage modulus versus frequency data of samples collected from different locations in the twin-screw extruder upon a deadstop of the extruder following steady extrusion (at a mass flow rate of 10 kg/h, screw rotational speed of 125 rpm, a slit die gap of 6 mm and walls of the barrel and the die kept at 60 °C) are shown in Fig. 5. Overall, regardless of the location, the storage modulus, G0 , values of the surfactant are significantly greater than the loss modulus, G00 , values. Furthermore, the dynamic properties G0 and G00 are parallel and little affected by the frequency, x. Thus, the gel nature
Fig. 5. Storage modulus, G0 (x) versus frequency behavior of LAS surfactant samples collected from various locations in the twin-screw extruder (at a mass flow rate of 10 kg/ h, screw rotational speed of 125 rpm, a slit die gap of 6 mm and walls of the barrel and the die kept at 60 °C).
584
H. Gevgilili et al. / Journal of Colloid and Interface Science 356 (2011) 579–588
Fig. 6. Magnitude of complex viscosity, g⁄(x) versus frequency behavior of LAS surfactant samples collected from various locations in the twin-screw extruder (at a mass flow rate of 10 kg/h, screw rotational speed of 125 rpm, a slit die gap of 6 mm and walls of the barrel and the die kept at 60 °C).
of the surfactant paste was preserved and the gel was further stiffened in the extruder, as revealed by the increases of the storage modulus, G0 , from about 50 kPa at the entrance to over 500 kPa at the exit of the twin-screw extruder. Furthermore, the magnitude of complex viscosity values of the surfactant also increased by about an order of magnitude as shown in Fig. 6. The comparison of the dynamics properties of the samples deformed in the drag and pressure flow in the extruder with those obtained from the pressure-induced deformation in the rectangular slit die (see for example the storage modulus, G0 , values of Fig. 4 and 5) indicates that the changes in elasticity (storage modulus) and viscosity (magnitude of complex viscosity) after shearing in the pressure flow in the slit die are negligible in comparison to the transformation of the paste in the drag and pressure flow in the twin-screw extruder. This suggests that the structural changes which took place during pressure flow in the slit die over a wide range of shear rates were not as significant as those generated during dynamic assembly in the drag and pressure flow of the
twin-screw extruder. What type of changes in the structure of the surfactant paste would generate such significant increases in elasticity and viscosity of the paste in the extruder? Léon et al. [43] have suggested that significant increases in elasticity of the surfactant paste would occur upon the formation of vesicles from lamellae. Upon vesicle formation the bilayers are shared across adjacent vesicles. Such sharing of bilayers across vesicles should give rise to affine junction points that would form a network to span the volume and stiffen the gel, and thus increase the elasticity as well as the shear viscosity of the vesicular formations. Figs. 7 and 8 show the SEM micrographs of the LAS surfactant paste specimens collected from different axial locations in the twin-screw extruder together with their location-dependent storage modulus and magnitude of complex viscosity values. The frequency is 0.1 rps and the error bars report the 95% confidence intervals determined according to Student’s t-distribution. The increases of the elasticity and the viscosity of the surfactant with
Fig. 7. Storage modulus, G0 , at a frequency of 0.1 rps versus axial distance in the twin-screw extruder and typical SEM micrographs of LAS surfactant samples which reveal the structural transformations that have taken place in the twin-screw extruder.
H. Gevgilili et al. / Journal of Colloid and Interface Science 356 (2011) 579–588
585
Fig. 8. Magnitude of complex viscosity, g⁄, at a frequency of 0.1 rps versus axial distance in the twin-screw extruder and typical SEM micrographs of LAS surfactant samples which reveal the structural transformations that have taken place in the twin-screw extruder.
increasing downchannel distance in the extruder are observed to be associated with its deformation-induced conversion into a vesicular structure, consistent with the findings of Léon et al. on the relatively high elasticity of vesicular structures [43]. The dynamic assembly of the vesicles occurs gradually. First, a portion of the spongy lamellar structure, with characteristic lengths of the clustered lamellae of about 2–20 lm, transforms into an irregular vesicular structure. With increasing downchannel distance the concentration of the vesicles increases (Figs. 7 and 8). The degree of homogeneity of the vesicular structure also increases with increasing downchannel distance. The dynamic assembly of the surfactant paste in the twin-screw extruder finally results in a highly-ordered vesicular structure with vesicle diameters all falling within a narrow 300–600 nm range. The linear viscoelastic properties of the paste were observed not to change during the 1 month upon extru-
sion that they were continued to be tested, suggesting that the vesicular structure that is generated is stable. The change from an irregular spongy lamellar structure into a unimodal, defect-free and elastic vesicular nanostructure is a testament to the efficiency of the combined drag and pressure flow of the twin screw extrusion process for the structural transformation of the surfactant. The simulation of the thermo-mechanical history that the surfactant is exposed to during twin screw extrusion was carried out using the three-dimensional Finite Element Method of Kalyon and Malik [44] and provided additional insight into the mechanisms of the transformation of the structure of the surfactant. The determination of the z-velocity (velocity component, integration of which provides the net flow rate in the extruder) distributions (Fig. 9) indicated that there are both negative and positive z-velocity values at various zones in the extruder suggesting the
Fig. 9. The FEM based simulated distributions of velocity in pressure and drag flow (z-velocity, i.e., main velocity component integration of which provides the volume flow rate) showing the presence of both positive and negative velocities and hence the significant backmixing occurring.
586
H. Gevgilili et al. / Journal of Colloid and Interface Science 356 (2011) 579–588
formation of pressure and drag flows akin to what is depicted in Fig. 1c. The resulting pressure gain in the right-handed screw elements (Sections 1 and 3) and pressure loss at the reversely staggered screws section (Section 2) and the die give rise to the back-mixing of the paste (drag and pressure flow depicted in Fig. 1c) thus homogenizing the deformation history and the resulting structure of the paste. The pressure was determined to increase monotonically in the right-handed screw elements preceding the left-handed screw elements (Section 1 in Fig. 5) and preceding the die (Section 3 in Fig. 5). On the other hand, the pressure decreases over the reversely staggered screw section (Section 2 in Fig. 5) and the die. Concomitant analysis of the linear viscoelastic properties and the simulated pressure distribution in the extruder suggests that significant transformations in structure (as evidenced by scanning electron microscopy and a significant change in the storage modulus, G0 , and magnitude of complex viscosity, g⁄, values) occur only at locations where the pressure gradient is positive and especially at locations where the positive pressure gradient is relatively high. On the other hand, the shearing of the paste in the partially full sections of the extruder, i.e., conveying action of the right-handed screw elements without generating pressure, does not lead to a significant change in structure (the G0 and magnitude of complex viscosity, g⁄, values remain invariant over the partially full section of the twinscrew extruder). Furthermore, further assembly of the vesicles into a more homogeneous structure and hence increase of G0 and g⁄ also do not occur at pressure-losing elements, i.e., at the die and within the confines of the left-handed screw elements (Section 2) although significant shearing occurs at these locations also. These observations should significantly alter the conventional thinking on which processing parameters can be used to realize the lamellar to vesicle
transformation of surfactants. Clearly, there are other factors than the ‘‘shear rate’’ or ‘‘strain’’ that were considered earlier to be the only factors in processing and the conversion of the surfactant structure into vesicles [29,45,46]. It appears that a combination of pressure, shear and backmixing is critical in achieving high-rate dynamic assembly of the vesicles of the anionic surfactant. The pressure and drag flows are particularly appropriate to overcome the wall slip and viscoplasticity limitations of the surfactant paste to allow the transformation of the structure. Experimental and theoretical evidence from polymeric melts and polymeric suspensions suggests that the positive pressure gradient in the flow direction reduces the slip velocity at the wall boundary in the flow direction [47,48]. The reduced wall slip velocities should impose a more uniform shear rate gradient and the positive pressure gradient, to give rise to backmixing which would reduce structural inhomogeneities. Schmidt et al. [49] have shown that the vesicles of a surfactant can be degraded into perpendicularlyaligned lamellae when the shear stress exceeds a critical shear stress. Although a significant range of stress magnitudes (through both shear and extensional types of deformations) were imposed on the LAS paste in the twin-screw extruder, the applied stresses have not degraded the closely packed vesicular structure of the resulting paste (Fig. 10). This is a precursor of the stability of the vesicular structure, which was demonstrated by the stability of the linear viscoelastic material functions with time during the 1month test period following extrusion. The observed structuring of the surfactant paste is temperature dependent. As the temperature of the surfactant paste in the twinscrew extruder increases to over 65 °C the rate of structure formation decreases below that was observed at 60 °C (as indicated by lower storage modulus, G0 , values). These findings are consistent
Fig. 10. The schematics of the transformation of the structure of the concentrated surfactant paste from spongy lamellar to vesicular (insets) upon the combination of pressure and drag flows. The significant backmixing provides a narrow size distribution.
H. Gevgilili et al. / Journal of Colloid and Interface Science 356 (2011) 579–588
with the experimental findings of McKeown et al. [33] and Liaw et al. [34] which indicated that the lamellae to vesicle conversion of the LAS surfactant (Fig. 10) only becomes significant over a narrow temperature window and the rate of formation of onion like multilayered vesicles depends exponentially on temperature. These findings emphasize the importance of temperature selection and control during the dynamic assembly of the vesicles. It is expected that the sodium silicate particles compounded into the surfactant paste would be dissolved in the water phase with the rate of dissolution affected favorably by the relatively high surface to volume ratios of the colloidal sodium silicate. The dissolution of the sodium silicate would alter the salinity and thus the topology of the bilayers [43]. The energy, E, associated with the formation of spherical vesicles consisting of onion like configuration of surfactant bilayers from flat lamellar bilayer depicted in Fig. 10 is given as [22,50]:
~Þ E ¼ 4pð2j þ j
ð4Þ
~ are the membrane elastic constants referred to as where j and j bending rigidity and saddle-splay radii. Spontaneous formation of ~ < 2j. Léon et al. [43] have determined vesicles is predicted for j that at least for the aqueous solution of sodium bis(2-ethylhexyl) sulfosuccinate (AOT) at a surfactant concentration of 7%, j is unaf~ is affected. Liaw et al. [34] fected by the addition of salts, whereas j suggest that the dissolution of sodium disilicate might give rise to a decrease of the bilayer repeat distance, and/or raise the ionic strength of the water present in the LAS bilayers. This would alter the size of the head-group of the LAS surfactant or change the electrostatic forces within and between layers to thus presumably alter the membrane elastic constant to enable the formation of onion like multilayered vesicles which are interconnected to each other to provide the significant increase in elasticity and viscosity (Fig. 10).
3. Experimental 3.1. Materials A highly-viscous concentrated surfactant, i.e., a linear alkylbenzene sulfonate LAS-water paste, reported earlier by Liaw et al. [34] was used. This paste is industrially referred to as a ‘‘high active’’ LAS paste of ca. wt. 78% of linear alkylbenzene sulfonate salts with the rest water, with LAS consisting of ‘‘lamellar bilayer structures’’ [34]. The chemical composition and properties of a typical LAS batch were reported by Liaw et al. [34] and McKeown [51] and the composition includes 77.2% by weight active matter, 19.1% water, 1.65% of linear alkylbenzene, 0.42% of Na2SO4, 0.37% of NaOH and 0.16% of NaCl. Here, the LAS paste was incorporated with 1% by volume of sodium silicate nanoparticles, 50–300 nm. The paste was provided by Procter & Gamble.
3.2. Characterization of the LAS paste The microstructure of the LAS paste prior to and upon deformation was characterized microscopically using cryogenic SEM (Hitachi S-4300). The LAS paste was characterized macroscopically using small-amplitude oscillatory shear (using an ARES rheometer from TA Instruments). The data collected in the small-amplitude oscillatory shear included the storage, G0 , and loss, G00 , moduli and the magnitude of the complex viscosity, |g⁄|, as a function of strain amplitude and frequency, x. Strain sweeps were carried out to determine the strain amplitude ranges over which the paste exhibited linear viscoelastic behavior. Typically, the paste exhibited linear behavior at strain amplitudes that are smaller than 1%.
587
3.3. Pressure and drag flow: twin screw extrusion A 50.8 mm fully-intermeshing co-rotating twin-screw extruder, TSE, was used in conjunction with a rectangular slit die with an adjustable gap [42]. The adjustable gap slit die was equipped with four pressure transducers to allow the determination of the pressure distribution and hence the wall shear stress in the die. The temperature of the surfactant was determined with a series of thermocouples installed flush with the die and the extruder barrel surfaces and using a thermal imaging camera to record the temperature distribution of the surfactant strand immediately upon exit from the die. The twin-screw extruder was also equipped with pressure transducers to allow the determination of the degree of fill and the pressure distribution in the twin-screw extruder. During the twin screw extrusion experiments the concentrated LAS (at controlled water content) was fed into the extruder, followed by the introduction of the dry silicate particles. Upon reaching steady state in temperature and pressure distributions the flow was brought to a dead-stop, the barrel halves were split open to give access to the surfactant paste in the extruder and samples of the surfactant paste were collected as a function of location in the extruder and the die. The linear viscoelastic properties of the paste samples were characterized immediately upon extrusion and repetitively over a period of one month to assess the stability of the pastes. The calculations of the temperature distributions in the twinscrew extruder (using the simulation methodologies described by Kalyon and Malik [44]) indicated that there is very little viscous energy dissipation occurring in the confines of the twin-screw extruder under the geometry and operating conditions employed, and the structuring operation can be considered to occur under isothermal conditions. The wall temperature profiles and the heat transfer through the barrel and die surfaces thus dictated the temperature distribution in the extruder, with little variability. This was also verified using an inframetrics thermal imaging camera which was used to monitor the distributions of the temperature of the surfactant extrudates upon emerging from the die (crosssectional areas of which were exposed upon cutting of the extrudates in the transverse to flow direction). 4. Conclusions Three different basic flow mechanisms were applied to the investigation of the transformation of the structure of a concentrated amphiphilic LAS surfactant paste, i.e., pure drag flow, pure pressure flow and combination of pressure and drag flow. Only pure drag and pure pressure flows have been utilized in the past for the shear-induced structuring of surfactants [18,23,27,32,33]. The transformation of the structure of the LAS surfactant from spongy lamellar to a relatively uniform vesicular structure only occurred when a combination of pressure and drag flows was used. The rheological characterization of the surfactant paste revealed that the paste was a viscoplastic gel which exhibited significant wall slip during attempts to deform it. The combination of the pressure and drag flow and the associated pressure increase minimized the wall slip and viscoplasticity effects to allow the deformation of the paste to induce the buckling of the lamellar bilayers and the concomitant transformation of the structure into multilayered vesicles. The documented backmixing occurring in the combination of pressure and drag flow gave rise to the dampening of the differences in thermo-mechanical histories of the different volumes of the paste, thus homogenizing the structure to generate vesicles within the narrow diameter range of 300– 600 nm. The formation and increasing homogeneity of the vesicular domains gave rise to significant increases in various rheological material functions including the storage modulus, G0 , and magni-
588
H. Gevgilili et al. / Journal of Colloid and Interface Science 356 (2011) 579–588
tude of complex viscosity, g⁄, presumably due to the interconnection of the vesicles to each other during deformation and buckling of the bilayers. Overall, it was demonstrated that the dynamic assembly of surfactant bilayers into uniform vesicular structures of concentrated surfactant pastes is possible when a combination of pressure and drag flows is utilized. Furthermore, it was recognized that the structuring of such pastes is affected by viscoplasticity, wall slip, pressure and backmixing, i.e., all parameters not recognized in earlier studies investigating the effects of shearing on structure formation of surfactants. Acknowledgments We thank Ms. C. Kong of P&G for her contributions to scanning electron microscopy of the specimens and to Mr. E. Demirkol, Dr. C. Erisken and Dr. N. Degirmenbasi of Stevens for their contributions to extrusion experiments and sample analysis and Dr. J. Carnali for his input. The funding which made this study possible was provided by P&G, for which we are grateful. References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12]
T. Zemh, M. Dubois, B. Deme, T. Krywicki, Science 283 (1999) 816. D. Discher, A. Eisenberg, Science 297 (2002) 967. Y. Ding, Z. Gui, J. Zhu, S. Yan, J. Liu, Y. Hu, Z. Wang, Mater. Lett. 61 (2007) 2195. J. Hotz, W. Meier, Adv. Mater. 10 (1998) 1387. X. Jiang, Y. Xie, J. Lu, L. Zhu, W. He, Y. Qian, J. Mater. Chem. 11 (2001) 1775. H. Hentze, S.R. Raghavan, C. McKelvey, E.W. Kaler, Langmuir 19 (2003) 1069. S. Che, K. Lund, T. Tatsumi, S. Iijima, S.H. Joo, R. Ryoo, O. Terasaki, Angew. Chem. Int. Ed. 42 (2003) 2182. L. Zhu, X. Zheng, X. Liu, X. Zhang, Y. Xie, J. Colloid and Interface Sci. 273 (2004) 155. A. Yamaguchi, F. Uejo, T. Yoda, T. Uchida, Y. Tanamura, T. Yamashita, N. Teramae, Nat. Mater. 3 (2004) 337. D. Ding, Z. Chen, S. Rajaputra, V. Singh, Sens. Actuators B 24 (2007) 12. A. Walcarius, E. Sibottier, M. Etienne, J. Ghanbaja, Nat. Mater. 6 (2007) 602. F. Gauffre, D. Roux, Langmuir 15 (1999) 3738.
[13] J.W. Szostak, D. Bartel, P.L. Luisi, Nature 409 (2001) 387. [14] M. Bergmeier, M. Gradzielski, H. Hoffmann, K. Mortensen, J. Phys. Chem. B 103 (1999) 1605. [15] M. Redkar, P. Hassan, V. Aswal, P. Devarajan, J. Pharm. Sci. 96 (2007) 2436. [16] I. Grillo, E. Kats, A.R. Muratov, Langmuir 19 (2003) 4573. [17] V. Schmitt, C.M. Marques, F. Lequeux, Phys. Rev. E 52 (1995) 4009. [18] P. Sierro, D. Roux, Phys. Rev. Lett. 78 (1997) 1496. [19] W. Richtering, Current Opin. Colloid Interface Sci. 6 (2001) 446. [20] K. Mortensen, Science 6 (2001) 140. [21] M. Gradzielski, J. Phys.: Condes. Matter 15 (2003) 655. [22] W. Helfrich, Z. Naturforsch. 28C (1973) 693. [23] L. Courbin, P. Panizza, Phys. Rev. E69 (2004) 021504-1. [24] A. Zilman, R. Granek, Eur. Phys. J. B 11 (1999) 593. [25] A. Wunenburger, A. Colin, T. Colin, D. Roux, Eur. Phys. J. E 2 (2000) 27. [26] S. Marlow, P. Olmsted, Eur. Phys. J. E 8 (2002) 485. [27] O. Diat, D. Roux, J. Phys. II 3 (1993) 9. [28] O. Diat, D. Roux, F. Nallet, Phy. Rev. E 51 (1995) 3296. [29] L. Courbin, J. Delville, J. Rouch, P. Panizza, Phys. Rev. Lett. 89 (2002) 148305-1. [30] T. Le, U. Olsson, K. Mortensen, Physica B 276–278 (2000) 379. [31] M. Gradzielski, M. Müller, M. Bergmeier, H. Hoffmann, E. Hoinkis, J. Phys. Chem. B 103 (1999) 1416. [32] J. Zipfel, P. Lindner, W. Richtering, Prog. Colloid Polym. Sci. 110 (1998) 139. [33] S.A. McKeown, M.R. Mackley, G.D. Moggridge, Trans. I. Chem. E 81A (2003) 649. [34] M.S. Liaw, M.R. Mackley, J. Bridgwater, AIChE J. 49 (2003) 2966. [35] F. Chambon, H. Winter, J. Rheol. 31 (1987) 683. [36] M.E. DeRosa, H.H. Winter, Acta 33 (1994) 220–237. [37] U. Yilmazer, D.M. Kalyon, J. Rheol. 33 (1989) 1197. [38] D. Kalyon, H. Gevgilili, R. Yazici, A. Post, G. McFann, Rheol. Acta 43 (2004) 396. [39] D. Kalyon, J. Rheol. 49 (2005) 621. [40] H. Gevgilili, D. Kalyon, J. Rheol. 45 (2001) 467. [41] D. Kalyon, H. Gevgilili, J. Rheol. 47 (2003) 683. [42] D. Kalyon, H. Gokturk, US Pat., 5 277 058, 1994. [43] A. Léon, D. Bonn, J. Meunier, A. Al-Kahwaji, O. Greffier, H. Kellay, Phys. Rev. Lett. 84 (2000) 1335. [44] D.M. Kalyon, M. Malik, Int. Polym. Proc. 22 (2007) 293. [45] J. Bergenholtz, N.J. Wagner, Langmuir 12 (1996) 3122. [46] J. Zipfel, F. Nettesheim, P. Lindner, T.D. Le, U. Olsson, W. Richtering, Europhys. Lett. 53 (2001) 335. [47] S.G. Hatzikiriakos, J.M. Dealy, J. Rheol. 36 (1992) 703. [48] H. Tang, D.M. Kalyon, J. Rheol. 52 (2008) 507. [49] G. Schmidt, S. Müller, C. Schmidt, W. Richtering, Rheol. Acta 38 (1999) 486. [50] T. Le, U. Olsson, K. Mortensen, Phys. Chem. Chem. Phys. 3 (2001) 1310. [51] S.A. McKeown, PhD Thesis, University of Cambridge, UK, 2001.