4157
2008, 112, 4157-4160 Published on Web 03/15/2008
Colloidal Shape Controlled by Molecular Adsorption at Liquid Crystal Interfaces Guillaume Toquer,†,‡ Ty Phou,‡ Sophie Monge,£ Antoine Grimaldi,‡ Maurizio Nobili,‡ and Christophe Blanc*,‡ Centre de Recherche Paul Pascal, UPR 8641, CNRS, AVenue Albert Schweitzer, 33600 Pessac, France, Laboratoire des Colloı¨des, Verres et Nanomate´ riaux, UMR 5587, CNRS-UniVersite´ Montpellier II, Place Euge` ne Bataillon, 34095 Montpellier, France, and Institut Charles Gerhardt, UMR 5253, CNRS-UM2-ENSCM-UM1, Equipe Inge´ nierie et Architectures Macromole´ culaires, Place Euge` ne Bataillon, 34095 Montpellier, France ReceiVed: January 16, 2008; In Final Form: February 28, 2008
The ability to control finely the structure of materials remains a central issue in colloidal science. Due to their elastic properties, liquid crystals (LC) are increasingly used to organize matter at the micrometer scale in soft composites. Textures and shapes of LC droplets are currently controlled by the competition between elasticity and anchoring, hydrodynamic flows, or external fields. Molecules adsorbed specifically at LC interfaces are known to orient LC molecules (anchoring effect), but other induced effects have been poorly explored. Using specifically designed amphitropic surfactants, we demonstrate that large-shape transformations can be achieved in direct LC/water emulsions. In particular, we focus on unusual nematic filaments formed from spherical droplets. These results suggest new approaches to design new soft LC composite materials through the adsorption of molecules at liquid crystal interfaces.
The first nematic LC composite materials, such as polymerdispersed liquid crystals (PDLCs),1 were mainly designed because of their optical properties. In the past decade, it has appeared that LC elastic properties can be exploited to form ordered structures of particles or liquid droplets.2 Combined with optical trapping3 or microfluidics4 techniques, this makes possible the formation of large colloidal assemblies with tailordesigned properties. Whether as the matrix or the dispersed phase, a nematic liquid crystal texture indeed self-organizes under the competition between elasticity and the orientation imposed by the interfaces (anchoring).5,6 Such designs are attractive since external stimuli like electric/magnetic fields or even hydrodynamic flows7 can distort significantly the structures and change their optical properties. Key characteristics, such as the threshold voltage or the response time in PDLCs,1 strongly depend on the shapes, anchorings, and defects of dispersed droplets. Despite the high surface/volume ratio, the anchoring is by far the main interfacial phenomenon considered in LC composites. For the well-studied emulsions of nematics,5 for instance, the role of the dispersant is usually restricted to the anchoring. The multiple interfacial phenomena observed in simple emulsions however suggest that a finer control of adsorbed molecules at liquid crystals interfaces would enhance properties of LC composites. Such an approach is currently being developed by the group of N. L. Abbott.8,9 Adsorption of molecules significantly changes the anchoring properties of the * To whom correspondence should be addressed. Address: LCVN, UMR 5587, CNRS-Universite´ Montpellier II, Place Euge`ne Bataillon, 34095 Montpellier cedex 5, France. E-mail:
[email protected]. Phone: 33 (0) 4 67 14 38 54. Fax: 33 (0) 4 67 14 46. † CNRS. ‡ CNRS-Universite ´ Montpellier II. £ CNRS-UM2-ENSCM-UM1.
10.1021/jp800431y CCC: $40.75
water/LC interface, which could be used to detect targeted biological materials in solution. Apart from the anchoring, other specific interfacial phenomena have been only marginally explored.10 We demonstrate here that the adsorption of amphitropic surfactants at a LC interface is another way to modify the morphology of LC composites, at least transiently. We focused on emulsions of 4′-pentyl-4-cyanobiphenyl (5CB) dispersed in water (W). This common LC displays a nematic phase at room temperature, with a clearing point TNI ) 35.3 °C. Besides the commercial cetyltrimethylammonium ammonium bromide surfactant (CTAB), we used, as emulsifying agents, three home-synthesized surfactants shown Figure 1A. They all possess mesogenic oxycyanobiphenyl groups and were inspired by studies focused on amphitropic molecules displaying both lyotropic and thermotropic features.11-14 Further details about the syntheses of 4-(4′-cyanobiphenyloxy)dodecyl trimethylammonium bromide (12OCB-TAB), bis-12-4-(4′-cyanobiphenyloxy) dodecyldimethyl ammonium bromide (2[12OCB]DAB)andN′-bis(12-(4′-cyano-4-biphenyloxy)dodecyl)-(N,N,N′,N′)tetramethyldodecane diammonium dibromide (12OCB-TAB-1212OCB-TAB) are given in Supporting Information. 5CB/W emulsions were prepared by sonicating a small amount of 5CB (1-10 wt %) in water in the presence of one of the above surfactants (0.1-3 wt %) without any particular procedure to get monodisperse size. The mixture was vigorously stirred for 1 h and then kept in closed containers at 60 or at 25 °C. A first batch of emulsions has been prepared at 25 °C at a low weight fraction φ of surfactants (φ ) 0.1 wt % for φ5CB ) 3 wt %). Except for 2[12OCB]-DAB, the emulsions are stable for several days, showing that a significant fraction of surfactants are located at the 5CB/water interface. Under a polarizing microscope, the nematic droplets show Maltese crosses (Figure © 2008 American Chemical Society
4158 J. Phys. Chem. B, Vol. 112, No. 14, 2008
Letters
Figure 1. (A) Molecular structure of CTAB and three amphitropic surfactants. (B) Polarizing optical micrograph of direct nematic 5CB emulsions stabilized with CTAB. Only spherical droplets are observed. The homeotropic anchoring is clearly evidenced by the presence of Maltese crosses. (C) Bright-field and (D) polarized micrographs of 5CB/W emulsions stabilized with 12OCB-TAB-12-12OCB-TAB. When subjected to small temperature variations, the emulsions often display transient nonspherical droplets, slightly pear-distorted. Larger distortions give rise to a tubular shape with a constant radius. This effect is spectacularly reinforced in the presence of ethanol (E).
Figure 2. In glass cylinders treated with homeotropic anchoring, both radial (sketched in A) and “escape in third dimension” (sketched in B) nematic textures are classically observed. The quantitative comparison between the polarizing optical micrograph of an elongated nematic filament (C) and simulated birefringent patterns (D) computed by the Jones matrix method reveals that only the “escape” texture following the curvature of the filament (inset) accounts accurately for the observed extinction lines.
1B), which reveals a strong homeotropic anchoring and a radial director field. This texture is observed down to optical resolution, meaning that the topologic radius15scomparing anchoring and elasticitysis much smaller than 0.5 µm for all surfactants.
Figure 3. Influence of temperature. (A) Change of an isolated filament during a temperature cycle between 20 and 25 °C (heating at +0.2 °C/min followed by cooling at -0.5 °C/min). (B) Evolution of the filament length during the cycle. (C) Pearling instability observed under a sudden temperature increase (+5 °C in 1 min).
It also indicates that surfactants cover the 5CB/water interface since anchoring would be planar otherwise.10 The Krafft point of CTAB is TK ≈ 25 °C. Below this temperature, CTAB precipitates as a weakly hydrated crystal. Above this temperature, it rapidly dissolves in water and forms
Letters
J. Phys. Chem. B, Vol. 112, No. 14, 2008 4159
Figure 4. Schematic scenario of the formation of nematic filaments. (A) A rapid quench (cooling below TK, ethanol evaporation at room temperature) destabilizes the micelles in water. Surfactants partly spread at the 5CB/water interface. (B) An increase in the interface area dA yields different increases dE in the elastic energy for a filament and a spherical droplet. (C) The energy of a filament is larger than the energy of an assembly of spheres of same volume and area. The cylindrical shape is however locally stabilized by elasticity and the necessary formations of defects (hyperbolic and hedgehog) which hinder peristaltic modes.
a micellar phase. Conductivity measurements have revealed similar behaviors for 12OCB-TAB (TK ≈ 28 °C) and 12OCBTAB-12-12OCB-TAB (TK ≈ 35-37 °C). The Krafft temperature of 2[12OCB]-DAB was not detected since the aqueous solubility remains poor up to 80 °C, explaining why the corresponding emulsions are not stabilized. Emulsions formed at 60 °C therefore incorporate larger amounts of surfactants, dispersed in micelles. A second batch of emulsions was prepared at 1 wt % fraction of surfactants and stored at 60 °C. The liquid crystal emulsions were formed by a quench (10 °C‚min-1) to room temperature, below TNI. CTAB-based mixtures looked similar to the emulsions prepared at room temperature, but the situation was quite different for 12OCB-TAB and 12OCB-TAB-12-12OCB-TAB surfactants. After a few minutes, a shape instability develops. First ellipsoidal, some LC droplets adopt a pear-like shape, and filaments start to extend (Figure 1C,D). Several sequences with 12OCB-TAB-12-12OCB-TAB are given in Supporting Information Figures S1-S4. The instability is not observed at quasistatic cooling (lower than 1 °C/min), which reveals its out-ofequilibrium nature. The nucleation of the filaments is highly heterogeneous, and only a few dropletsspredominantly the largestsundergo the instability. Addition of ethanol (1-5 wt %) greatly enhances the formation of filaments under temperature change. Its rapid evaporation in uncovered samples at room temperature also yields gigantic filaments (up to 1 mm long in Supporting Information Figure S4). Under the harshest quenches, more than 50% of the droplets are deformed. The threads are observed for a variable duration before relaxing to their former shapes (Supporting Information Figure S3). This behavior is intrinsic since two neighboring threads might evolve very differently, with one thread relaxing in a few minutes while the second one is still observed after a few hours. We however noted that the smaller filaments generally disappear first. Although intriguing, such dynamics are qualitatively explained by a simple nucleation-growth mechanism, as described further. Observations under monochromatic light between crossed polarizers clearly show that the internal texture is not a radial organization with an axial disclination, the simplest one compatible with a strong homeotropic anchoring (see Figure 2). Quantitative comparisons with simulated textures, obtained by the Matrix Jones technique,16 rather evidence that the texture along the filament is the classic “escape in third dimension nematic texture” observed in thick glass cylindrical capillaries with homeotropic anchoring.17 This confirms that a nematic filament is locally at mechanical equilibrium but also implies
that it is asymmetric. This prediction is clearly confirmed since a point defect, necessarily present in the drop for topologic reasons, is observed only at one end (Supporting Information Figures S2-S5). Let us stress that filaments made of concentric layers have been reported as transient out-of-equilibrium patterns formed by layered phases. The hydration of a surfactant lamellar phase produces such myelin figures.18 During the isotropic-to-smectic phase transition of thermotropic liquid crystals, elongated cylindrical nuclei of the SmA phase are also observed19 when a homeotropic anchoring is present at the SmA-I interface. The nematic filaments reported in this work differ from such layered filaments in two main features, suggesting that new phenomena are responsible for their appearance and shape stability. First, the growth and also the subsequent slow relaxation (see Supporting Information Figure S3) occurs at constant droplet volume and is not directly related to the isotropic-to-nematic phase transition. Second, the liquid crystal is in a nematic state, much more fluid than a smectic phase. Contrary to the layered filaments, where the axial disclination hinders any layer distortion, no topological defect sustains the local cylindrical shape stability. To address these two questions, the origin of the driving force and the cylindrical shape stability, we first have checked that the instability was fully independent from the 5CB isotropicto-nematic phase transition encountered during the quench. After the relaxation of almost all filaments formed in a sample at 27 °C, the temperature was slowly varied. The length of isolated filaments was measured after a standard image analysis, including thresholding and skeletonization techniques.20 It shows that 1 h after the quench, a thread still continuously elongates while the temperature decreases (see Figure 3). Reciprocally, it continuously shrinks when the temperature slowly increases. A sharp temperature increase (a few degrees above 5 °C/min) however destabilizes the filaments (Figure 3C) into a Rayleighlike “pearling” instability, which corresponds to a decrease of the interfacial area A. Such elements suggest that the formation of filaments is driven by a local increase of A under cooling. At high temperature, surfactants are localized at the 5CB/W interface but also are present in the aqueous part as micelles. Below TK, micelles get rapidly unstable. Apart from the interface, surfactants are either free or form crystallites. The increase of A would be due to a partial transfer of surfactants from micelles to the interface, the formation of crystallites being either less energetically favorable or delayed for kinetics reasons (Figure 4A). This hypothesis explains the effect of ethanol since
4160 J. Phys. Chem. B, Vol. 112, No. 14, 2008 we checked that the amounts employed were sufficient to solubilize the surfactants at room temperature. Its rapid evaporation therefore has the same effects as a temperature quench. For isotropic liquids, the surfactants’ spreading, however, would not explain the formation of filaments rather than a uniformly distributed increase in area. For a nematic droplet, however, it yields a variable change in the Frank elastic energy E ) (K/2) ∫((∇‚n)2 + (n‚∇ × n)2 + (n × ∇ × n)2)d3r, where K is the elastic constant and n the director field. We compare the deformations sketched in Figure 4B for a spherical droplet and a filament of the same volume V. The initial elastic energy of the sphere of radius R0 is E ) 8πKR0, whereas the energy per unit length of a filament is 3πK.21 The bulk elastic energy increase dE is found to be proportional to the increase dA in area in both cases, with dE ≈ 5.8K/R0dA for the sphere and dE ≈ 3.2K/R0dA for the filament (see Supporting Information). The area uptake therefore costs less energy if performed first by the largest droplets and then preferably by the filaments. This enlightens several observations such as the heterogeneity of lifetimes or the persistence of spherical droplets. As in crystal growth, the presence of a growing structure prevents the local formation of other ones. In that sense, the proposed mechanism resembles a nucleation-diffusion process and can be sketched in four main steps. First, the quench destabilizes the micelles homogeneously present in water, which favors an increase of interface area. Second, a few droplets (predominantly the largest ones) start to form filaments which reduce the available amount of surfactants in solution and prevent other nucleation sites. Filaments are then formed rapidly. In a third step, the smallest filaments tend to shrink in favor of the largest ones because of their higher energy. This step is slower since it rests on diffusion of free molecules between distant droplets. It furthermore is intertwined with the fourth step where other microscopic mechanisms such as the formation of crystallites in solution or exchange of matter between droplets finally relax all filaments. A filament is indeed definitely an out-of-equilibrium figure because it has the same surface energy but a larger elastic energy than an assembly of spheres of the same total volume and area (sketch in Figure 4C). This last point shows that the cylindrical shape is not a global minimum for elastic energy but rather a strong local minimum. Several theoretical papers have already explored the influence of a nematic ordering on the capillarity instabilities in various geometries.22,23 They predict that the elastic energy strongly modifies the Rayleigh-like instability modes. In the present case, this effect is certainly reinforced by the necessary formation of topological defects for large distortions. Hedgehog and hyperbolic hedgehog defects are necessary when large peristaltic modes occur (see Figure 4C). The full variational problem, including deformation of the interface and distortion of the director, requires a dedicated treatment, but both elastic energy and topology are expected to preserve the cylindrical geometry at fixed area (a sudden area decrease however reveals these pearling modes as shown Figure 2C). To conclude, nematic dispersions in an isotropic fluid have been extensively reported to display (nearly) spherical droplets.
Letters Only recently, unusual geometries like nematic shells24 have been produced thanks to microfluidics. We emphasize that potentialities offered by surfactants in liquid crystal composites remain largely unexplored. Here, we clearly show a coupling of amphitropic surfactants with the interface and the nematic order, yielding large morphology changes. The development of such new materials should certainly open new routes of experimental and theoretical investigations of liquid crystal emulsions. A better understanding of the underlying phenomena should yield, in combination with microfluidics, to a fine control of LC emulsions. Acknowledgment. We gratefully acknowledge fruitful discussions with M. In and G. Porte (CNRS/Montpellier). Supporting Information Available: Further details about methods and syntheses, supplemental optical micrographs, and computation of elastic energy for the deformed droplets considered in the text. This material is available free of charge via the Internet at http://pubs.acs.org. References and Notes (1) Drzaic, P. J. Liquid Crystal Dispersions; World Scientific: Singapore, 1995. (2) Loudet, J. C.; Barois, P.; Poulin, P. Nature 2000, 407, 611. (3) Musevic, I.; Skarabot, M.; Tkalec, U.; Ravnik, M.; Zumer, S. Science 2006, 313, 954. (4) Fernandez-Nieves, A.; Link, D. R.; Rudhardt, D.; Weitz, D. A. Phys. ReV. Lett. 2004, 92, 105503. (5) Lavrentovich, O. D Liq. Cryst. 1998, 24, 117. (6) Stark, O. D Phys. Rep. 2001, 351, 357. (7) Fernandez-Nieves, A.; Link, D. R.; Marquez, M.; Weitz, D. A Phys. ReV. Lett. 2007, 98, 087801. (8) Jang, C.; Cheng, L.; Olsen, C. W.; Abbott, N. L. Nano Lett. 2006, 6, 1053. (9) Brake, J. M.; Abbott, N. L. Langmuir 2007, 23, 8497. (10) Tjipto, E.; Cadwell, K. D.; Quinn, J. F.; Johnston, A. P. R.; Abbott, N. L.; Caruso, F. Nano Lett. 2006, 6, 2243. (11) Attard, G. S.; Fuller, S.; Howell, O.; Tiddy, G. J. T. Langmuir 2000, 16, 8712. (12) Fuller, S.; Hopwood, J.; Rahman, A.; Shinde, N.; Tiddy, G. J.; Attard, G. S.; Howell, O.; Sproston, S. Liq. Cryst. 1992, 12, 521. (13) Everaars, M. D.; Marcelis, A. T. M.; Sudholter, E. J. R. Langmuir 1995, 11, 3705. (14) Everaars, M. D.; Marcellis, A. T.; Sudho¨lter, E. J. R. Langmuir 1993, 9, 1986. (15) Tixier, T.; Heppenstall-Butler, M.; Terentjev, E. M. Eur. Phys. J. E 2005, 18, 417. (16) Galatola, P.; Oldano C. In The Optics of Thermotropic Liquid Crystals; Elston, S. J., Sambles, R., Eds.; Taylor & Francis: London, 1998. (17) Cladis, P. E.; Kleman, M. J. Phys. 1972, 33, 591. (18) Buchanan, M.; Egelhaaf, S. U.; Cates, M. E. Langmuir 2000, 16, 3718. (19) Todorokihara, M.; Iwata, Y.; Naito, H. Phys. ReV. E 2004, 70, 021701. (20) Ott, A.; Magnasco, M.; Simon, A.; Libchaber, A. Phys. ReV. E 1993, 48, R1642. (21) Kleman, M.; Lavrentovich, O.D. Soft Matter Physics; SpringerVerlag: New York, 2003. (22) Rey, A. D. Soft Matter 2007, 3, 1349. (23) Cheong, A. G.; Rey, A. D. J. Chem. Phys. 2007, 117, 5062. (24) Fernandez-Nieves, A.; Vitella, V.; Utada, A. S.; Ma´rquez, M.; Nelson, D. R.; Weitz, D. A. Phys. ReV. Lett. 2007, 99, 157801.
