3766
Langmuir 2007, 23, 3766-3773
Irreversible Collapse of Poly(vinyl stearate) Monolayers at the Air-Water Interface Paolo G. Mussone,† Andy W. F. Ip,‡ Sven L. M. Schroeder,†,‡ Brent S. Murray,§ and Aline F. Miller*,† Molecular Materials Centre, School of Chemical Engineering and Analytical Science, and Molecular Materials Centre, School of Chemistry, UniVersity of Manchester, SackVille Street, Manchester M60 1QD, U.K., and Food Colloids Group, Procter Department of Food Science, UniVersity of Leeds, Leeds LS2 9JT, U.K. ReceiVed September 19, 2006. In Final Form: January 8, 2007 The collapse of Langmuir monolayers of poly(vinyl stearate) (PVS) at the air-water interface has been investigated by combined measurements of the surface pressure-area isotherms and Brewster angle microscopy (BAM). Atomic force microscopy (AFM) has been used to gain out-of-plane structural information on collapsed films transferred onto a solid substrate by a modified version of the inverse Langmuir-Schaefer deposition method. At high areas per monomer repeat unit, BAM imaging revealed that the films are heterogeneous, with large solidlike domains (25-200 µm in diameter) coexisting with liquidlike domains. Upon film compression, the domains coalesced to form a homogeneous monolayer before the film collapsed at constant pressure, forming irreversible three-dimensional (3D) structures. BAM images showed that two 3D structures coexisted: buckles of varying width extending across the surface and perpendicular to the direction of the compression and dotted islandlike structures. Upon expansion, the film fractured and both 3D protrusions persisted, explaining the marked hysteresis recorded in the Langmuir isotherms. Experiments with AFM confirmed the 3D nature of both protrusions and revealed that many buckles contain substructures corresponding to narrow buckles whose heights are a multiple of a single bilayer. Additionally, many multilayer islands with diameters spanning from 0.2 µm to over 3.5 µm were characterized by varying heights between 2 nm and up to over 50 nm. The key to the formation of the irreversible 3D structures is the presence of large inhomogeneities in the PVS monolayer, and a generalized phenomenological model is proposed to explain the collapse observed.
Introduction Insoluble Langmuir monolayers (LMs) at the air-water interface have attracted remarkable attention since 1891, when Fra¨ulein Agnes Pockels described a method for manipulating films formed by oils on water.1 Today, LMs are regarded as important two-dimensional (2D) model systems for studying physical, chemical, and biological phenomena in confined geometries.2,3 Moreover, they are essential precursors for fabricating novel devices for nonlinear optical,4 electrical,5 photoconduction,6 and biotechnology (protein crystallization)7 applications using the Langmuir-Blodgett (LB) technique. Understanding out-of-plane monolayer transitions to the third dimension is essential since any LB application is strongly dependent on defect-free molecular packing (pinholes, dislocations) and periodic structures.5 Furthermore, LM protrusions in the third dimension are known to play a key role in several biological processes such as the formation mechanisms of lipid * To whom correspondence should be addressed. Telephone: +44 161 3065781. Fax: +44 161 3064399. E-mail:
[email protected]. † Molecular Materials Centre, School of Chemical Engineering and Analytical Science, University of Manchester. ‡ Molecular Materials Centre, School of Chemistry, University of Manchester. § University of Leeds. (1) Pockels, A. Nature 1891, 43, 437. (2) Jones, M. N.; Chapman, D. Micelles, monolayers, and membranes; WileyLiss: Hoboken, NJ, 1995. (3) Mo¨hwald, H. Rep. Prog. Phys. 1993, 56, 653-685. (4) Petty, M. Langmuir-Blodgett films: An Introduction; Cambridge University Press: New York, 1996. (5) Roberts, G. Langmuir-Blodgett Films; Plenum Press: New York, 1990. (6) Tredgold, R. H. Order in thin organic films; Cambridge University Press: New York, 1994. (7) Garnaes, J.; Schwartz, D. K.; Viswanathan, R.; Zasadzinski, J. A. N. Nature 1992, 357, 54-57.
bilayer membranes2 and the hysteresis of compressiondecompression cycles as realized, for example, by the pulmonary function of lung surfactants.8 Langmuir monolayers are in a metastable state when the surface pressure exceeds the equilibrium spreading pressure (πesp) during lateral compression.9 If the monolayer is held in this region at constant pressure, the result is the nucleation and growth of 3D structures into the bulk subphase.10-12 This mechanism is referred to as “slow collapse”. However, when the stable bulk phase is a crystalline solid, it is possible to drive the monolayer above the πesp9 and the transition from two to three dimensions is revealed in the surface pressure-area isotherms by the appearance of either of the two following distinct features. The first is a pronounced spike, followed by a sudden pressure drop at a constant molecular area (constant area collapse). The second is a plateau where the surface pressure remains constant over a variably long range of area per molecule (constant surface pressure collapse). Both signatures are generally referred to as the collapse point in the Langmuir isotherm. Beyond this point, the monolayer is characterized by structural modifications in both the normal and water surface planes. In both cases, the resulting transition to the third dimension depends on several parameters, including the chemical structure of the system, intramolecular interactions, composition, the temperature and pH of the subphase, and compression rate.9,13-15 (8) Longo, M. L.; Bisagno, A. M.; Zasadzinski, J. A.; Bruni, R.; Waring, A. J. Science 1993, 261, 453-564. (9) Gaines, G. L. J. Insoluble Monolayer at Liquid-Air Interfaces; Wiley: New York, 1966. (10) Vollhardt, D.; Retter, U. J. Phys. Chem. 1991, 95, 3723-3727. (11) Vollhardt, D.; Retter, U. Langmuir 1992, 8, 309-312. (12) Vollhardt, D.; Retter, U.; Siegel, S. Thin Solid Films 1991, 199, 189199.
10.1021/la0627361 CCC: $37.00 © 2007 American Chemical Society Published on Web 02/27/2007
IrreVersible Collapse of PVS Monolayers
In the past decade, the introduction of methods such as Brewster angle microscopy (BAM),16,17 fluorescence microscopy,18 atomic force microscopy (AFM),19 X-ray diffraction20 and scattering,21,22 neutron reflectometry,23 and light scattering24 has allowed significant advances in understanding out-of-plane LM transitions. A remarkably rich variety of structures such as linear folds,25,26 two-armed spirals,27 vesicles,18 3D buds,28 twisted ribbons,29 fiberlike networks,30 reversible as well as irreversible buckling,21,24,31 and giant folds32,33 and interdigitated bilayers resting on the original monolayer34 have been documented. Several workers have proposed mathematical10-12,24,35-40 and phenomenological models26,30,33,41,42 to explain these phenomena. However, a thorough understanding of monolayer collapse, in particular, the spontaneous transition from two to three dimensions, remains elusive. Recently, a significant number of studies have focused on various aspects of the LM transition to the third dimension during constant area collapse. For example, Ybert et al. published compelling evidence that 2-hydroxytetracosaneic acid (2-OH TCA) LMs collapse via three distinct mechanisms, depending on the compression rate.32 In addition to a slow collapse by nucleation and growth, LMs of 2-OH TCA can collapse by forming giant folds into the subphase or by forming ridges parallel to the compression line coexisting with multilayered islands. In another study, Ybert and co-workers examined the collapse kinetics and structure of 10,12-pentacosadiyonic acid and found that these LMs collapse by forming a trilayer.33 In this work, dynamic measurements support the idea of a folding and sliding mechanism,36 but structural analysis by AFM suggests that domain growth may be taking place by molecular diffusion along the (13) Adamson, A. W.; Gast, A. P. The Physical Chemistry of Surfaces; John Wiley & Sons: New York, 1997. (14) Kato, T. Langmuir 1990, 6, 870-872. (15) Kato, T. Langmuir 1991, 7, 2208-2212. (16) He´non, S.; Meunier, J. ReV. Sci. Instrum. 1991, 62 (4), 936-939. (17) Ho¨nig, D.; Mo¨bius, D. J. Phys. Chem. 1991, 95, 4590-4592. (18) Lipp, M. M.; Lee, K. Y. C.; Takamoto, D. Y.; Zasadzinski, J. A.; Waring, A. J. Phys. ReV. Lett. 1998, 81 (8), 1650-1653. (19) Vollhardt, D.; Kato, T.; Kawano, M. J. Phys. Chem. 1996, 100, 41414147. (20) Meine, K.; Weidemann, G.; Vollhardt, D.; Brezesinski, G.; Kondrashkina, E. A. Langmuir 1997, 13, 6577-6581. (21) Bordieu, L.; Daillant, J.; Chatenay, D.; Brasleau, A.; Colson, D. Phys. ReV. Lett. 1994, 72 (10), 1502-1505. (22) Fradin, C.; Braslau, A.; Luzet, D.; Alba, M.; Gourier, C.; Daillant, J.; Gru¨bel, G.; Vignaud, G.; Legrand, J. F.; Lal, J.; Petit, J. M.; Rieutord, F. Physica B 1998, 248, 310-315. (23) Krueger, S. Curr. Opin. Colloid Interface Sci. 2001, 6, 111-117. (24) Saint-Jalmes, A.; Gallet, F. Eur. Phys. J. B 1998, 2, 489-494. (25) Ries, H. E. J. J. Phys. Chem. 1955, 59, 94-95. (26) Ries, H. E. J. Nature 1979, 281, 287-289. (27) Hatta, E.; Hosoi, H.; Akiyama, H.; Ishii, T.; Mukasa, K. Eur. Phys. J. B 1998, 2, 347-349. (28) Schief, W. R.; Touryan, L.; Hall, S. B.; Vogel, V. J. Phys. Chem. B 2000, 104, 7388-7393. (29) Lee, K. Y. C.; Lipp, M. M.; Takamoto, D. Y.; Ter-Ovanesyan, E.; Zasadzinski, J. A.; Waring, A. J. Langmuir 1998, 14, 2567-2572. (30) Steffens, S.; Oldendorf, J.; Haufe, G.; Galla, H.-J. Langmuir 2006, 22, 1428-1435. (31) Gopal, A.; Lee, K. Y. C. J. Phys. Chem. B 2001, 105, 10348-10354. (32) Ybert, C.; Lu, W.; Mo¨ller, G.; Knobler, C. M. J. Phys. Chem. B 2002, 106, 2004-2008. (33) Ybert, C.; Lu, W.; Mo¨ller, G.; Knobler, C. M. J. Phys.: Condens. Matter 2002, 14, 4753-4762. (34) Xue, K.; Jung, C. S.; Kim, M. W. Phys. ReV. Lett. 1992, 69 (3), 474-477. (35) Milner, S. T.; Joanny, J.-F.; Pincus, P. 1989, 9, 495. (36) Nikomarov, E. S. Langmuir 1990, 6, 410-414. (37) Smith, R. D.; Berg, J. M. J. Colloid Interface Sci. 1980, 74 (1), 273-286. (38) Lu, W.; Knobler, C. M.; Bruinsma, R. F.; Twardos, M.; Dennin, M. Phys. ReV. Lett. 2002, 89 (14), 146107. (39) Diamant, H.; Witten, T. A.; Ege, C.; Gopal, A.; Lee, K. Y. C. Phys. ReV. E 2001, 62, 061602. (40) Diamant, H.; Witten, T. A.; Gopal, A.; Lee, K. Y. C. Europhys. Lett. 2000, 52 (2), 171-177. (41) Kundu, S.; Datta, A.; Hazra, S. Phys. ReV. E 2006, 73, 051608. (42) Gopal, A.; Belyi, V. A.; Diamant, H.; Witten, T. A.; Lee, K. Y. C. Condens. Matter 2004, 0409147.
Langmuir, Vol. 23, No. 7, 2007 3767
Figure 1. Schematic representation of a PVS polymer where n ) 80.
surface. Kundu et al. recently observed the conversion of LMs of stearic acid on a water subphase containing certain divalent ions into bimolecular layers during collapse at constant pressure.43 An expansion of this work by the same authors focused on the collapse of stearic acid on an aqueous subphase containing Co2+ ions.41 They showed that nearly circular islands coexist with ridges in the collapsed structure of the films and pointed out the strong similarity of this phenomenon with the StranskiKrastanow mode generally observed for heteroepitaxial growth.44,45 Another recent study by Steffens and co-workers showed that mixtures of chiral ethyl 2-azido-4-fluoro-3-hydroxystearates collapse by forming organized structures (fiberlike networks) during constant pressure collapse.30 A phenomenological model that involves repeated LM folding and sliding was proposed to explain the structure of the fibers observed. Permanent LM deformations following collapse at constant pressure have also been reported for preformed linear polymers and copolymers,21,46-48 but the relationship between structure and collapse mechanism for such systems remains poorly understood. Here, we study the constant pressure collapse of a LM of a preformed linear polymer by relating surface pressure isotherms to microscopy and structural measurements. Our experiments focus on poly(vinyl stearate) (PVS). Peng and Barnes reported that consistent inhomogeneities are formed in films of PVS at the air-water interface as they are compressed due to the high surface viscosity and slow diffusion of the macromolecules on the water surface.49 However, the effect of these large inhomogeneities on monolayer structure during the compression, collapse, and compression/expansion cycles is still to be fully elucidated. In this paper, we will address this and propose a collapse mechanism. Experimental Section Materials. PVS (Sigma-Aldrich) was purified by dissolution in toluene (Sigma-Aldrich) and subsequent precipitation by slow addition of excess ethanol (Sigma-Aldrich) under constant agitation.50 This process was repeated several times. The final solution was filtered under vacuum to remove residual solvent. Gel permeation chromatography gave a number-average molecular weight of 25 000 g mol-1 and a polydispersity of 3. The chemical structure of PVS is given in Figure 1. The polymers were spread on an ultrapure water (Elga UHQ) subphase from spectroscopic-grade chloroform (Sigma-Aldrich) solutions. The pH and resistivity of the water were 7.2 and 18.2 MΩ cm, respectively. Mica discs for AFM experiments were purchased from Cambridge Scientific (Cambridge, U.K.). (43) Kundu, S.; Datta, A.; Hazra, S. Langmuir 2005, 21, 5894-5900. (44) Daruka, I.; Baraba´si, A.-L. Phys. ReV. Lett. 1997, 79, 19. (45) Shchukin, V. A.; Bimberg, D. ReV. Mod. Phys. 1999, 71 (4), 1125-1171. (46) Fontaine, P.; Daillant, J.; Guenoun, P.; Alba, M.; Braslau, A.; Mays, J. W.; Petit, J. M.; Rieutord, F. J. Phys. II 1997, 7, 401-407. (47) Adams, J.; Buske, A.; Duran, R. S. Macromolecules 1993, 26, 28712877. (48) Dubreuil, F.; Fontaine, P.; Alba, M.; Daillant, J.; Mays, J. W.; Zalczer, G.; Guenoun, P. Europhys. Lett. 2005, 70 (2), 176-182. (49) Peng, J. B.; Barnes, G. T. Langmuir 1990, 6, 578-582. (50) Chen, Y.-L.; Kawaguchi, M.; Yu, H.; Zografi, G. Langmuir 1987, 3, 31-35.
3768 Langmuir, Vol. 23, No. 7, 2007
Mussone et al.
Methods. Langmuir Film Balance. For this study, a home-built Langmuir film balance was used. Details regarding the experimental setup can be found elsewhere.51 Polymer solutions (1.0 mg mL-1) were spread using a Hamilton syringe (Sigma-Aldrich) and deposited dropwise evenly on the water surface. The solvent was allowed to evaporate (20 min) before compression and expansion of the monolayer at a constant rate of 1.0 ( 0.1 Å2 repeat unit-1 min-1. The temperature of the subphase was maintained at 25 ( 0.1 °C using thermostated water. Each isotherm was collected a minimum of three times to ensure reproducibility. Hysteresis experiments were conducted by immediately reversing the barrier movement after the area reached the selected value. Brewster Angle Microscopy. A commercial Brewster angle microscope (BAM2plus model, Nanofilm Technologie GmbH, Germany) was mounted on the trough and used to observe the morphology of the monolayer directly at the air-water interface. This technique does not require chemical labeling, reducing the risk of subphase contamination.52 The p-polarized light from a 690 nm, 30 mW laser was reflected off the air-water interface at the Brewster angle (53.15°), giving a spatial resolution of ∼2 µm. The analyzer and polarizer were set at 0° throughout all experiments. The domains formed were imaged at different stages of monolayer compression using the same magnification throughout. Operational conditions have been described elsewhere.53 Atomic Force Microscopy. The reverse Langmuir-Schaefer technique29 was used to deposit the LMs onto mica discs for study using AFM in the tapping mode. Such measurements were performed in the constant force mode with a Veeco NanoScope IIIa system (Digital Instruments Inc.) equipped with a 100 µm scanning head, at room temperature. Sample images were obtained using silicon pyramidal tips (Windsor Scientific Ltd, U.K.) with a spring constant of 0.04 N m-1. Scans were carried out at several locations on the discs with scan areas ranging from 1 to 50 × 50 µm2.
Results Surface Pressure-Area Isotherms. Langmuir Isotherms. A typical Langmuir isotherm obtained for the PVS monolayer is given in Figure 2b. As expected, the isotherm does not show first- or second-order phase transitions as in the case of long chain alcohols, fatty acids, or phospholipids.9 For the PVS monolayer, the surface pressure was found to be immeasurably low (π ≈ 0) at large areas per repeat monomer unit (A > 50 Å2 per repeat unit). When the molecular area approaches 30 Å2 per repeat unit, the surface pressure begins to increase slowly. As the compression of the monolayer proceeds, the surface pressure increases further with no discontinuity until π ≈ 20 mN m-1 (corresponding to ∼22 Å2 per repeat unit) is reached. Further lateral compression continues to increase the surface pressure approximately linearly. The upper limit of this region is at a surface pressure of ∼60 mN m-1 (17 Å2 per repeat unit). After this point, the surface pressure ceases to rise with decreasing molecular area. Such behavior is generally interpreted as the collapse point of the monolayer,9 that is, when the monolayer extends into the third dimension. Hysteresis. A hysteresis pattern is observed for PVS LMs when they are subjected to compression, expansion, and further compression-expansion cycles (Figure 2a). The surface pressure values recorded during expansion correspond to smaller molecular areas. When a second compression is imposed, the isotherm has a similar shape but is shifted by ∼2 Å2 per repeat unit to smaller molecular areas relative to the initial compression isotherm. Subjecting the system to repeated compression and expansion cycles causes a drift of the isotherms toward smaller molecular areas, with the distance between successive compression(51) Murray, B. S.; Nelson, P. V. Langmuir 1996, 12 (25), 5973-5976. (52) Mo¨bius, D. Curr. Opin. Colloid Interface Sci. 1998, 3, 137. (53) Garofalakis, G.; Murray, B. S. Langmuir 2002, 18, 4765-4774.
Figure 2. Surface pressure isotherms of a PVS-spread film at 298 K obtained during repeated compression-expansion cycles (a) with the isotherm obtained during a single compression shown as a reference (b).
expansion curves decreasing until they finally overlap after five cycles (Figure 2a). The reduction in the area per repeat unit is due to the polymer moving away from the interface, either by dissolution into the subphase or by incorporation into multilayered structures. We calculated that ∼15% of our monolayer was “lost” in this way after the first cycle, with ∼50% of the polymer finally moving away from the 2D monolayer. Interestingly, similar hysteresis patterns were also obtained when the compression and expansion cycles were limited to lower surface pressures, for example, 10 mN m-1 (see Supporting Information). Brewster Angle Microscopy (BAM). Information about the macroscopic film topography and domain features was obtained using Brewster angle microscopy (BAM), which capitalizes on differences in reflectivity to p-polarized light incident at the Brewster angle. The black regions in the micrographs correspond to the bare surface of water, while the gray areas reveal the presence of the monolayer domains. Saturated white occurs when the thickness or density of the film increases to the point where the reflected light exceeds the instrument scale. The images obtained for A > 50 Å2 per repeat unit (Figure 3) clearly show inhomogeneities in the LM formed immediately after depositing the polymer solution at the air-water interface. Large and uniformly gray regions with diameters from ∼25 µm to more than 200 µm coexist with extended inhomogeneous areas that exhibit marked expanded voids that are a few micrometers in diameter (Figure 3c). It was also noted that the
IrreVersible Collapse of PVS Monolayers
Langmuir, Vol. 23, No. 7, 2007 3769
Figure 5. BAM images of collapsed portions of PVS monolayers at approximately 18 Å2 per repeat unit showing (a) the formation of irregularly spaced buckles perpendicular to the line of compression and (b) the coexistence of buckles and islandlike structures. The micrographs are 300 × 300 µm2 in dimension.
Figure 3. Representative BAM images of PVS monolayers at 50 Å2 per repeat unit. The micrographs are 300 × 300 µm2 in dimension.
Figure 4. BAM images of PVS monolayers (a) approaching the onset of the Langmuir isotherm at ∼32 Å2 per repeat unit and (b) above the onset of the Langmuir isotherm at ∼27 Å2 per repeat unit. The micrographs are 300 × 300 µm2 in dimension.
domain boundaries are often characterized by rather sharp cusps, suggesting that the monolayer domain is solidlike. Generally, the islands formed by the polymer appear to float slowly across the water surface and bounce rigidly, without deforming, when they come in contact with each other. Upon further compression, the domains drift against each other and their freedom of movement is progressively limited due to the reduction in available area. This is evident in Figure 4a, where individual domains have clearly joined together. This continues with reducing area per repeat unit until all domains have coalesced at ∼30 Å2 per repeat unit, resulting in a uniformly gray BAM image (Figure 4b), indicating complete coating of the water surface. No significant changes in BAM contrast were recorded while crossing the polarization on the incident and on the reflected beam. This suggests that the domains formed were optically isotropic throughout the compression. As the compression proceeds further, the measured surface pressure begins to increase. In tandem, the intensity measured for the reflecting surface of the film also slowly increased. No macroscopic changes in film morphology were detected until suddenly bright lines appeared at ∼55 mN m-1, which were approximately orthogonal to the line of the compression (Figure 5). The intensity level recorded for these features (buckles) saturates the gray level scale of the instrument and is several orders of magnitude higher than that of the surrounding monolayer. This suggests that such modification of the film structure is out-of-plane. Typically, these buckles extend longitudinally for more than 300 µm and are between 3 and 10
Figure 6. BAM images of a collapsed PVS monolayer during expansion. As expansion begins, tearing is observed (a), which leads to the formation of fractures (b) upon further film expansion. Buckles and islandlike 3D structures are clearly irreversible (c) and do not reintegrate in the monolayer plane (d). The micrographs are 300 × 300 µm2 in dimension.
µm in width (Figure 5a). The distance between the buckles is not constant and varies from 15 µm to more than 100 µm. In addition, a second 3D structure is present in the monolayer. Figure 5b highlights the coexistence of bright lines and isolated clusters of approximately spherical structures with a diameter ranging from 2 µm to more than 10 µm. These structures have been interpreted as 3D islands dispersed in the monolayer structure. Figure 6 shows the effect of film expansion on film morphology. During the initial stages of expansion, tearlike cuts affect the surface of the monolayer (top left corner of Figure 6a). As expansion progresses, such tears expand (Figure 6b) and the film appears to break, rigidly producing large domains characterized by boundaries with sharp cusps. It is also clear that both protruded structures do not disappear with film expansion, suggesting that their formation is irreversible (Figure 6c). As the surface pressure is reduced, the film fractures and individual domains begin to move across the surface again, but the extended inhomogeneous areas with expanded voids present before film compression do not reappear (Figure 6d). Repeated compression and expansion cycles typically generate new buckles, which appear to be aligned roughly parallel to those formed upon the first compression (Figure 7a). When more
3770 Langmuir, Vol. 23, No. 7, 2007
Figure 7. BAM images showing the appearance of portions of collapsed monolayers upon (a) two and (b) repeated compressionexpansion cycles. The micrographs are 300 × 300 µm2 in dimension
Figure 8. AFM micrographs of a PVS monolayer transferred onto a mica substrate at 0 mN m-1 showing (a) a spongelike structure, (b) the solidlike domains punctuated with multilayered islands, (c) a cross-sectional analysis of the height of the multilayered islands, and (d) a profile of the height of the islands. Micrographs (a) and (b) are 8 × 8 µm2 in dimension. Micrograph (c) is 2.5 × 2.5 µm2 in dimension.
than two compression and expansion cycles were applied, diffuse and bright white areas appeared, indicating further collapse of the monolayer surface (Figure 7b). Atomic Force Microscopy (AFM). The structural characteristics of PVS monolayers before and after collapse were studied by AFM. At the length scale of the BAM observations, the images collected by AFM are consistent with our BAM observations. Hence, we are confident that no artifacts were introduced during film transfer onto mica discs. The AFM results obtained by depositing PVS LMs at A > 50 Å2 per repeat unit show that the films are indeed heterogeneous under the experimental conditions. As previously observed with BAM (Figure 3c), extended portions of spongelike monolayer domains characterized by micrometer-sized voids coexist with large, more uniformly distributed solidlike domains (Figure 8a and b). Moreover, we noted the presence of randomly distributed out-of-plane islands, protruding with diameters typically between 0.2 and 1.5 µm and with heights varying between 2 nm and up to over 50 nm (for example, see Figure 8c and the corresponding height in Figure 8d). Such structures were not observed in BAM, as their size would have been below the resolution of the instrument.
Mussone et al.
Figure 9. AFM micrographs of a PVS monolayer transferred onto a mica substrate at (a) 5 mN m-1 showing discoids welded together that coexist with a spongelike phase and at (b) 10 mN m-1 where the discoids are welded together. The micrographs are 8 × 8 µm2 in dimension.
Figure 10. AFM micrographs of a collapsed monolayer transferred onto a mica substrate at 60 mN m-1 showing (a) the formation of buckles with widths ranging from hundreds of nanometers to 4 µm, (b) a profile of the height of the buckled region, and (c) the collapsed film upon expansion. Micrographs are 50 × 50 µm2 in dimension.
As the LM domains are brought closer together by continuous lateral compression, interconnected roughly circular polymer domains (discoids) of various diameters (generally 30 mN m-1) are characterized by large, homogeneous areas uniformly coated by the polymer film, randomly punctuated with multilayered islands (micrographs not shown). Films deposited after reaching the constant pressure plateau confirm that the collapsed LMs contain the characteristic buckles in the micrometer scale aligned perpendicular to the compression line (Figure 10a). Furthermore, other buckles that could not be observed using BAM are also present. These are shorter in length, spanning from ∼0.5 µm to over 20 µm, and their width appears to be limited to a few hundred nanometers. A careful analysis of the heights of these collapsed structures reveals the following
IrreVersible Collapse of PVS Monolayers
two scenarios. In the case of the shorter, narrower buckles, the heights measured are consistently ∼5 nm, which corresponds to two monomolecular layers54 (assuming an all trans configuration of the side chains). In the case of the broader, micrometer-sized buckles, there is evidence suggesting that they contain substructures where narrow buckles, whose heights are a multiple of a single bilayer, are resting on top of a broad bilayer structure. For example, a simple cross-sectional analysis of the buckle indicated by an arrow in Figure 10b shows that there is (moving left to right) a triple bilayer (left edge) followed by a double bilayer (approximately in the center) and finally a single bilayer, all resting on the main bilayer that constitutes the broad buckle. The presence of such substructures resting on the bilayer formed upon collapse could be attributed to a repeated folding of the LMs. This is discussed when we propose a generalized mechanism below. On expansion, the collapsed film breaks rigidly into disconnected raftlike pieces and the protrusions are not reincorporated into the monolayer structure (Figure 10c). Finally, we observe that the islandlike structures observed at high surface areas increase in number as well as in size (exceeding 10 µm in diameter for samples deposited after collapse) during lateral compression and also over time at constant surface pressure. These islands correspond to those observed in BAM.
Langmuir, Vol. 23, No. 7, 2007 3771
Phenomenological Analysis of Hysteresis Pattern and Buckle Distribution. The presence of irreversible hysteresis patterns described in this study can be related to irreVersible collapse as defined by Gaines.55 Our results with PVS monolayers obtained with the Langmuir balance expand the previous work by Peng and Barnes49 by revealing the origin of the hysteresis in the π-A plot upon repeated compression and expansion cycles. BAM as well as AFM microscopy revealed that PVS LM instabilities generated upon compression are not reincorporated into the monolayer when the collapsed structure is expanded. For this reason, the expansion curve does not reproduce the initial compression curve, and the second compression isotherm is not superimposable with the first. Hysteresis is also present at low levels of π and can be explained by considering the characteristics of the PVS polymer. Peng and Barnes studied PVS monolayers and found that significant surface pressure gradients form along the trough line during compression.49 This phenomenon was found to be dependent on both the spreading technique and the compression mode. They suggested that the high viscosity and slow diffusion of the macromolecules at the air-water interface led to surface inhomogeneities. In the bulk phase, these two parameters are a function of the molar mass.56 Assuming the same relationship to be valid in 2D systems, one expects that the monolayer will be more resistant to flow, or to applied stress, as the molar mass is increased. The PVS used in this study has a high molecular weight (Mn ∼ 25 000 g mol-1). Moreover, the molecular architecture of PVS lacks flexibility. Under such circumstances, we predict that PVS monolayers would be resistant to changes in the 2D conformation so that macromolecular rearrangement to minimize surface tension would become difficult and diffusion would become very slow.57 A direct consequence of the intrinsically slow diffusing/
rearranging monolayer is that, upon compression, the domains tend to accumulate in the region close to the barriers. As a consequence, gradients in surface pressure are generated along the length of the trough, extending from the center toward the barriers. All our measurements were carried out with the Wilhelmy plate positioned in the center of the trough, corresponding to the lowest value of surface pressure during compression. It is likely that, as soon as the barriers begin to move, portions of the monolayer close to the barriers experience high π and consequently lead to a collapse of the monolayer, which gives rise to the hysteresis pattern if the film is subsequently expanded. Another consequence of the intrinsically slow diffusing macromolecules is the irregular space distribution of the buckles, which tended to be more numerous close to the barriers. This hypothesis was confirmed by AFM measurements of the film deposited at different positions in the trough. Phenomenological Analysis of the PVS LM Collapse Mechanism. The presence of extended buckling instabilities perpendicular to the line of compression in rigid films has been reported previously.21,24-26,58-62 As discussed earlier,32 there is a marked difference between LM collapse occurring when there is phase coexistence (such as in phospholipid monolayers31,39,40) and when there is a one-phase region in the phase diagram. In the first case, the mechanism is known to depend on differences in height between phases of different spontaneous curvature.39,40 For the PVS system studied here, no evidence was found of phase coexistence at the collapse point. In the second case, instability may arise from monolayer defects such as dislocations in the plane32 or molecular tilt.24 Such defects may originate when the macromolecules are spread at the air-water interface, as they will have difficulty arranging to minimize surface tension due to the rigidity of the polymer backbone and their high molecular weight.63 This is likely to lead to polymer-polymer entanglement or to protrusion of the polymer into the third dimension, both of which could act as nucleation points for buckle formation. The transition from two to three dimensions in LMs is often discussed in relation to the collapse mechanism put forward by Ries.26 According to this phenomenological model, the transition begins with a weakening of the film and the subsequent formation, under compression, of a bilayer column in a head-head configuration protruding from the interface. During the next stage, the bilayer folds over, topples, and then forms a trilayer (the bilayer resting on the LM). The resulting structure is that of a LM broken into disconnected multilayers. Alternative mechanisms have been proposed to explain how some monolayers can collapse reversibly.18,32,36,38,64 Essentially, the initial steps in all these scenarios are identical to those in the Ries mechanism, but in the final stage, the buckled bilayer, instead of breaking rigidly, slides over the LM and remains in contact with the monolayer. Our structural results from AFM studies do not fully fit either of these mechanisms and leave two questions unanswered: How can buckles and multilayered islands coexist? What is the mechanism ruling the formation of the buckled structures? Nearly circular islands of height greater than a bimolecular layer coexisting with ridges at high surface pressures after collapse were also found by Kundu.41 In this work, the authors suggest that the ridges may be collections of small islands and, to support
(54) Peng, J. P.; Barnes, G. T.; Gentle, I. R. AdV. Colloid Interface Sci. 2001, 91 (2), 163-219. (55) Gaines, G. L. J. Langmuir 1991, 7, 834-839. (56) Freeman, G. R. Kinetics of nonhomogeneous processes: a practical introduction for chemists, biologists, physcists and materials scientists; Wiley: Chichester, 1987. (57) Peng, J. B.; Barnes, G. T.; Schuster, A.; Ringsdorf, H. Thin Solid Films 1992, 210-211 (Part 1 SU), 16-18.
(58) Graner, F.; Gallet, F.; Houchmandzadeh, B. Europhys. Lett. 1994, 28, 565. (59) Zhang, Y.; Fisher, T. M. J. Phys. Chem. B 2005, 109, 3442-3445. (60) Hatta, E. Langmuir 2004, 20, 4059-4063. (61) Hatta, E.; Fischer, T. M. J. Phys. Chem. B 2002, 106, 589-592. (62) Hatta, E.; Suzuki, D.; Nagao, J. Eur. Phys. J. B 1999, 11, 609-614. (63) Peng, J. B.; Barnes, G. T. Langmuir 1991, 7, 1749-1754. (64) McFate, C.; Ward, D.; Olmested, J. I. Langmuir 1993, 9, 1036-1039.
Discussion
3772 Langmuir, Vol. 23, No. 7, 2007
this idea, show how the growth of these wetting layers presents a strong similarity with the Stranski-Krastanov mode for heteroepitaxial growth on a solid substrate.44,45 We think that the latter model is not applicable to our system because of the fundamental difference in the molecular arrangement of the collapsed film. The system studied by Kundu and co-workers (stearic acid on an aqueous solution of Co+2 ions) can assume two different configurations due to the presence of the dissolved ions. Below collapse, the LM will consist only of molecules in the asymmetric configuration, with both hydrophobic tails pointing away from the interface. Upon collapse, some of the molecules will begin to adopt the symmetric configuration. The growth of the layers will proceed with a combination of islands and ridges since, from the point of view of energy minimization, the top of any asymmetric layer is equivalent to any of the symmetric layers. In our case, there is no such molecular driving force. Ybert and co-workers also found the coexistence of ridges and islandlike structures that persist and indeed grow in number during the compression process, leading to collapsed monolayers of 2-OH TCA. They attributed the origin of the multilayered islands to slow collapse.32 We believe it is possible to explain our results, obtained with a linear polymeric system, using similar arguments. Furthermore, we will go on to propose a generalized collapse mechanism for both surfactant and macromolecular films with extremely low πesp. It is commonly accepted that a monolayer compressed above its equilibrium spreading pressure becomes metastable because the formation of critical nuclei of the bulk phase becomes more probable.10 We know that portions of our monolayer collapse upon compression, but regions also exist where the surface pressure is above the πesp (0 mN m-1 for PVS). If film compression is stopped in the liquid-expanded or -condensed region and the surface pressure is held constant, the monolayer will relax due to the nucleation and growth of 3D centers, which creates the islands we observe here. This is confirmed as we see the number of islands increasing to an equilibrium value over time at constant pressure. The islands, in particular, form to minimize surface area at a given volume, which gives the minimum energy configuration. It is likely that the island in the center of our circular domain is the original nuclei center (all with a similar critical size), and since the macromolecules cannot relax easily, they remain located within the circular domain. Although it may be tempting to attribute some kind of initiating role to the multilayered islands in the process leading to the formation of the buckling instabilities described above, we have found no evidence to support this idea. This leads us to answer the second question we posed regarding the mechanism behind the formation and the structure of the buckles, as we believe it is different from nucleation and growth. Buckles extending across the trough perpendicular to the line of compression and roughly perpendicular to each other have been reported in a number of studies.21,24-26,58-62 The collapse we observed does not fit the classical Ries mechanism, however, as the monolayer does not break into disconnected portions at collapse (it does break but only when we start expanding it) and the isotherm for PVS shows a rather convincing constant pressure signature at collapse, which is generally a sign that the system is responding by expelling portions of the film without breaking the film (otherwise the surface pressure would drop). Our AFM results also show that the monolayer collapses by forming multibilayered structures that remain in contact with the original monolayer. Therefore, the mechanism we propose is a modification of that described by Steffens.30 First, the monolayer weakens and buckles (Figure 11b) due to the nucleating presence of local defects such as segments of polymer chain not lying flat on the
Mussone et al.
Figure 11. Schematic model of buckle formation at the air-water interface.
surface. The next stage is the formation of double layers in a head-head configuration (see Figure 1) that protrude into the third dimension. This is aided by consistent interactions within the headgroups of the macromolecule (Figure 11c). The double layers then bend onto one side and are extended by further lateral compression (Figure 11d). When frictional forces between the bilayers and the monolayer become too high, another nucleus originates at the border between the multilayered complexes already formed and thus promotes the generation of another buckle by subsequent growth and bending (Figure 11e-g). In the case of PVS, we have found that the additional bimolecular layers resting on top of the collapsed first bilayer do not necessarily form at the edges. The mechanism we suggest thus generalizes the Steffens model by allowing the nucleation of the buds responsible for the nth generation of buckles to appear, in principle, anywhere on the originally formed bilayer (Figure 11e). Since PVS LMs are highly defective and inhomogeneous, the formation of another bilayer is not reliant only on defects present at the edge. It should also be noted that there is a fundamental difference between the behavior of collapsed LMs in the study by Steffens and the behavior of PVS films upon expansion. In the former case, all protrusions in the third dimension are reincorporated within the monolayer; in the latter, they persist and indeed the monolayer breaks rigidly, showing a brittle, glasslike character. The reversibility observed by Steffens was considered as evidence to support the proposed mechanism. However, Steffens was examining a single chain surfactant system which has high compressibility and is therefore more likely to expand without irreversible changes in the domain structure. Our polymer system displays a rather low compressibility, which is a typical sign that rearrangements of the main chain and of the side chains are not easy during the compression process. From a macromolecular
IrreVersible Collapse of PVS Monolayers
point of view, the polymer chains become highly entangled during film compression, owing to their structural rigidity, and once they are buckled, they cannot rearrange fast enough upon expansion. Further supporting evidence of the collapsed monolayer rigidity is that the film breaks catastrophically (rather than relaxing), giving sharp boundaries between the fractured domains as the film expands in response to the opening of the trough barriers (Figures 6 and 10c). This behavior takes place at nearconstant area. Finally, we point out that some issues remain elusive and other questions need answers. For example, although the buckles are preferentially parallel to each other and are spaced, there is evidence of a branching pattern similar to that presented by Hatta.62 A similar branching appears to be repeated within the buckles, growing on the main buckles to form networklike structures. It may be possible to ascribe this pattern to local fluctuations of viscosity, since this is one of the key parameters in regulating the transition between the surface roughening and anisotropic cracking regimes.62 Further experiments are underway to clarify this issue.
Conclusions The experimental results show that the origin of the hysteresis in surface pressure isotherms of PVS monolayers is the generation of irreversible 3D structures. Combined BAM and AFM analyses revealed a complex topography of the collapsed surface. Many buckles of varying length and width and with heights corre-
Langmuir, Vol. 23, No. 7, 2007 3773
sponding to multiples of bimolecular layers coexist with relatively taller and wider islandlike formations. The key parameter in determining the nature of these instabilities and the mechanism governing their formation is the presence of large inhomogeneities in the PVS monolayer. The island structures form via slow collapse in regions of the film where we exceed the equilibrium surface pressure, while the buckles form in response to lateral compression. Both structures are randomly distributed on the surface, reflecting the inhomogeneity of the film. A phenomenological model that generalizes the model proposed by Steffens and coworkers was proposed to explain the formation of such monolayer instabilities. Acknowledgment. This work was supported by the School of Chemical Engineering and Analytical Science DTA Award, the Molecular Materials Centre, and the Society of Chemical Industry (Grey Scholarship). The BBSRC Bioimaging Initiative (24-BI11184) is gratefully acknowledged for the BAM facilities. The authors would like to thank Mr. Simon Braun and Dr. John Walton for help with the AFM measurements. Supporting Information Available: Surface pressure isotherm showing the hysteresis pattern for a PVS polymer film compressed at low surface pressures. This material is available free of charge via the Internet at http://pubs.acs.org. LA0627361
