8250
Langmuir 2005, 21, 8250-8254
On the Formation of Molecular Terraces Svetlana Santer,*,†,‡ Yun Zong,§,| Wolfgang Knoll,| and Ju¨rgen Ru¨he‡ Institute for Microsystem Technology (IMTEK), University of Freiburg, Georges-Koehler-Allee 103, D-79110 Freiburg, Germany, and Max-Plank Institute for Polymer Research, Ackermannweg 10, D-55128 Mainz, Germany Received May 6, 2005 Langmuir-Blodgett-Kuhn films of poly(amic acid) with azobenzene-chromophore side groups (azoPAA) have been examined using atomic force microscopy (AFM). If films were deposited at specific conditions, i.e., at transfer pressures (π ) 10-26 mN/m) and at transfer rates (ν ) 2-10 mm/min), unusual patterns were formed. While the first few transferred layers of azo-PAA exhibited uniform coverage of the underlying surface with a flat topography, successive layers form a pattern of parallel terraces with micrometer width, which are assembled in a pyramid-like structure. The height of the pyramidal steps is 1.6 nm, which corresponds to the thickness of an azo-PAA monolayer. Each even step of such a pyramid has a hydrophobic surface, and each odd step has a hydrophilic surface. The width of the terraces varies significantly from terrace to terrace; however, the width of an individual terrace remains constant over hundreds of micrometers. The orientation of the pyramidal terraces is governed by the film deposition process and is always perpendicular to the transfer direction.
Introduction Substrates that exhibit anisotropy in the surface chemical composition or topography are of great interest for a variety of applications. Examples are the engineering of surface properties such as wettability, adhesion, friction,1 selective molecular adsorption,2 and chemistry with extremely small amounts of material performed in microchannels.3 To construct such surfaces, various techniques ranging from strategies based on self-assembly using intrinsic molecular properties to technological strategies, such as microcontact printing4 and different types of lithographic methods,5 have been developed. Selfassembly has an advantage over lithography methods in miniaturization of the formed patterns, because molecular assemblies typically have a size range of 5-100 nm. Because the nanostructure is directly “build in” to the molecular structure of the compound used, it is inexpensive, simple, and fast. Surface patterning based on molecular properties normally utilizes phase separation of, e.g., diblock copolymers.6,7 Phase separation of a binary lipid mixture at the air-water interface followed by a transfer of these films onto a solid surface by the Langmuir-Blodgett-Kuhn (LBK) technique was shown to yield a regular nanometer scale pattern of highly parallel stripes.8 The formation of such stripes can also be * To whom correspondence should be addressed. Telephone: +49761-203-7415. Fax: +49-761-203-7162. E-mail:
[email protected]. † Born Prokhorova. ‡ University of Freiburg. § Present address: Institute of Materials Research and Engineering, 3 Research Link, Singapore 117602. | Max-Plank Institute for Polymer Research. (1) Xia, Y.; Rogers, J. A.; Paul, K. E.; Whitesides, G. M. Chem. Rev. 1999, 99, 1823. (2) Bo¨ltau, M.; Walheim, S.; Mlynek, J.; Krausch, G.; Szeiner, U. Nature 1998, 391, 877. (3) Delamarche, E.; Bernard, A.; Schmid, H.; Michel, B.; Biebuyck, H. Science 1997, 276, 779. (4) Xia, Y.; Whitesides, G. M. Langmuir 1997, 13, 2059. (5) Wang, R.; Hashimoto, K.; Fujishima, A.; Chikuni, M.; Kojima, E.; Kitamura, A.; Shimohigoshi, M.; Watanabe, T. Nature 1997, 388, 431. (6) Goldacker, T.; Abetz, V.; Stadler, R.; Eruhimovich, I.; Leibler, L. Nature 1999, 398, 137. (7) Zhang, L.; Eisenberg, A. Science 1995, 268, 1728. (8) Moraille, P.; Badia, A. Langmuir 2002, 18, 4414.
generated by dynamic wetting instabilities of a single component phospholipid monolayer during vertical Langmuir transfer on a solid surface.9 Here, we report on the formation of nanometer high terraces during LBK multilayer transfer of a poly(amic acid) with azobenzene-chromophore side groups (azoPAA). Azobenzene-derivatized polymeric layers have attracted a lot of attention as an interesting component for the photocontrol of liquid crystal (LC) alignment.10,11 This interest originates from the fact that the azobenzene moieties in side chains of polymers can control the alignment properties of liquid crystals in contact with correspondingly functionalized surfaces of LC devices. It was found that the switching of the azobenzene groups from their trans-isomer state to the cis-isomer state can be used to induce a total reorientation of the liquid crystalline layer from a homeotropic to a completely planar alignment.12-14 For the application of azobenzene-derivatized poly(amic acid) in optical data storage devices as a component of integrated optics platforms, the understanding of their film formation on solid surfaces is of great importance. Taking advantage of their amphiphilic character, we have utilized the LBK technique to prepare monolayers of the polymer at the air-water interface followed by the vertical transfer of the film onto a silicon wafer at a certain compression pressure. In this paper, we report on the spontaneous formation of terrace-like structures of these compounds during the LBK film transfer. Materials and Methods Synthesis of Azobenzene-Derivatized Poly(amic Acid). The poly(amic acid) carrying azobenzene-chromophore (azoPAA) moieties as side chains (Scheme 1) was synthesized as described elsewhere.15 (9) Gleiche, M.; Chi, L. F.; Fuchs, H. Nature 2000, 403, 173. (10) Ikeda, T.; Tutsumi, O. Science 1995, 268, 1873. (11) Kimura, K.; Suzuki, T.; Yokoyama, M. J. Chem. Soc., Chem. Commun. 1989, 20, 1570. (12) Ichimura, K.; Tomita, H.; Kudo, K. Liq. Cryst. 1996, 20, 161. (13) Ichimura, K.; Momose, M.; Kudo, K.; Akiyama, H.; Ishizuki, N. Langmuir 1995, 1, 2341. (14) Zong, Y. Ph.D. Thesis, Mainz, Germany, 2002.
10.1021/la051212i CCC: $30.25 © 2005 American Chemical Society Published on Web 07/28/2005
On the Formation of Molecular Terraces
Langmuir, Vol. 21, No. 18, 2005 8251
Scheme 1. Chemical Structure of a Poly(amic Acid) Having the Azobenzene-Chromophore Moiety as a Side Chain (azo-PAA)
Because of the presence of the azobenzene-chromophore side chains, the azo-PAA is soluble in organic solvents with a low boiling point such as chloroform and THF. Because the polymer is a strongly colored polyelectrolyte, its molecular weight could not be measured by standard methods such as static light scattering (SLS). Even relative measurements by size-exclusion chromatography proved to be less successful because the polymer had a strong tendency to adsorbed onto the column materials because of its amphiphilic character. However, a successful way to obtain a reasonable estimate of the average molecular weight and molecular weight distribution was to measure the size of individual polymer molecules adsorbed to a solid substrate by AFM. Although this seems to be at first a somewhat unusual method for the determination of molecular weights, it has been shown by others that it gives reasonable estimates.16 The molecular weight determined by this technique gave values of Mn ) (3 ( 1.4) × 105 g/mol. Details are described in ref 17. For the LBK transfer procedure, silicon wafers hydrophobised by a vapor silanization process are used. To this end, the substrates were placed in a hexamethyldisilazan (HMDS) atmosphere in a closed flask at 80 °C for about 40 min, followed by careful drying of the samples. Instrumentation. The preparation of the samples consisting of azo-PAA multilayers was carried out by the LBK technique using a single-trough film balance (FW2 Lauda). A chloroform solution of azo-PAA with a concentration of c ) 0.982 mg/mL was spread onto the water surface (MilliQ, 18.2 MΩ cm) followed by 15 min of equilibration time. After that, the Langmuir monolayer was compressed by moving the barrier with a velocity of 5 cm/min up to the desired deposition pressure. The Langmuir films were transferred onto the hydrophobized silicon wafer substrate via a Y-type deposition with dipping velocities of 2, 5, and 10 mm/min, respectively. The surface pressure for the film transfer was varied between 8 and 20 mN/m. Several samples with different numbers of azo-PAA layers were prepared consisting of 2, 4, 6, and 10 layers, respectively. Transfer ratios of 0.9-1.1 were obtained for a pressure of π ) 14 mN/m for all samples. The transfer behavior of the azo-PAA was found to be homogeneous,14 as it is seen from the deposition curve in Figure 1. Between film transfer and acquisition of the AFM images, the films were stored typically for 1 day. An atomic force microscope (AFM; MultiMode, Veeco Metrology Group) was used to characterize the morphology of the layers. AFM images were recorded in air at a relative humidity of 4045% and at room temperature (25 °C). Tapping-mode images were acquired using silicon cantilevers (Olympus) with a resonance frequency of ∼300 kHz, a spring constant of ∼50 N/m, and a tip radius of ∼10 nm. Commercial software (Nanoscope (15) Zong, Y.; Hees, U.; Knoll, W.; Ru¨he, J. J. Nonlinear Opt. Phys. Mater. 2002, 11, 367. (16) Sheiko, S. S.; da Silva, M.; Shirvaniants, D.; LaRue, I.; Prokhorova, S.; Moeller, M.; Beers, K.; Matyjaszewski, K. J. Am. Chem. Soc. 2003, 125, 6725. (17) Santer, S. A.; Zhong, Y.; Knoll, W.; Ru¨he, J. J. Mater. Chem., in press.
Figure 1. (a) Surface pressure isotherm of the azobenzenechromophore poly(amic acid) (azo-PAA) during compression at a speed of ν ) 5 cm/min and a temperature of T ) 25 °C. (b) Schematic depiction the Y-type deposition procedure. (c) Deposition curve obtained during the deposition of 10 layers of azo-PAA on silicon surface at room temperature (deposition pressure π ) 14 mN/m; deposition speed ν ) 2 mm/min) is shown. IIIa, DI) was used for the image analysis on several areas of 100/100 µm and 50/50 µm in size and averaged over 50-100 micrographs.
Results and Discussion Figure 1 shows a surface-pressure-area isotherm during compression of the poly(amic acid) having the azobenzene-chromophore polymer (azo-PAA). The area per monomer repeat unit is determined to be 1.2 nm2 from the surface-pressure isotherm by an extrapolation of the linear part of the area-pressure isotherm to the vanishing pressure. The collapse pressure is 42 mN/m (Figure 1a). Langmuir monolayers were transferred to the substrate at surface pressures of π ) 14 mN/m in an LBK deposition (Figure 1b), where the solid substrate was dipped into the subphase through the film, followed by withdrawal. Samples consisting of 2, 4, 6, and 10 layers of azo-PAA, respectively, were prepared. The deposition curve for a sample with 10 layers is shown in Figure 1c. After 1 day at ambient conditions, the transferred films were examined using a tapping-mode AFM. The surfaces of the samples consisting of 2, 4, and 6 layers show no significant features and are flat over a large surface area (Figure 2). The roughness of the surface measured from the surface area of 100 × 100 µm is σ ) 0.5 nm. The thickness of the films measured from the cross-section of a scratched area (Figure 2b) is d2 ) 3.2 ( 0.2 nm, d4 ) 6.4 ( 0.2 nm, and d6 ) 9.6 ( 0.2 nm for samples consisting of 2, 4, and 6 layers, respectively. This corresponds to a thickness of one monolayer of d1 ) 1.6 ( 0.2 nm, which
8252
Langmuir, Vol. 21, No. 18, 2005
Santer et al.
Figure 2. AFM micrographs of (a) 2 and (b) 4 layers of the azo-PAA transferred from the Langmuir trough onto the surface of the silanizated silicon wafer at π ) 14 mN/m, and ν ) 2 mm/min together with a cross-section profile (d). (c) Simplified scheme of a film consisting of 2 layers.
Figure 4. Surface-force curves measured on the top of the 9th and 10th layer.
Figure 3. AFM micrograph of the surface of 10 azo-PAA layers transferred from the Langmuir trough. The cartoon below shows a scheme of the cross-section of the pyramid-like structure.
is in good agreement with X-ray and ellipsometry measurements.15 The cartoon in Figure 2c, which depicts the structure of the polymer in a simplified manner and which will be used throughout this paper, denotes the hydrophilic and hydrophobic parts of the repeat units of the polymer. A totally different picture is observed if further layers are deposited (Figure 3). The 7th layer is not uniform anymore, but channel-like gaps of 1-10 µm in width are observed. The channels are highly parallel to each other, and all are perpendicular to the dipping direction. They extend over large areas of the substrate and are limited only by the size of the supporting silicon wafer used. If further layers are deposited, a terrace-like structure is formed. The height of each terrace remains constant over all subsequent layers and is 1.6 nm, which is in good agreement with the molecular dimensions of the deposited
monolayers. In total, the structure resembles a series of pyramidal stripes. Although the width of each terrace is constant over the whole range of the image, the width of different terraces varies strongly. If the space taken up by the molecules in each layer is compared to the geometric surface area, it is observed that the coverage decreases from 100% (coated substrate, 6th layer) to 80% (7th layer), 60% (8th layer), 40% (9th layer), and 10% (10th layer). If the adhesion force is measured during the recording of force-distance curves on the top of each “terrace” (Figure 4), it is evident that the hydrophilicity changes strongly from terrace to terrace. The force curves measured with a hydrophilic silicon AFM tip on the 9th layer and 10th layer, respectively, differ quite strongly in the value of the adhesion force. The force is much higher for the 9th layer than for the 10th layer, implying that the 9th layer is terminated by polar groups, while the 10th layer is terminated by nonpolar groups. To investigate the nature of the terrace formation process, 9 layers of polymer were deposited as described above, while the substrate was rotated by 30° relative to the dipping direction, prior to the deposition of the 10th layer. The AFM image of the substrate thus obtained shows that again stripes are formed perpendicular to the dipping direction, thus yielding a series of highly parallel stripes that have an angle of 30° to the stripes formed
On the Formation of Molecular Terraces
Langmuir, Vol. 21, No. 18, 2005 8253
Figure 5. AFM micrograph of the 10th layer of an azo-PAA multilayer film consisting of 10 layers. Prior to the deposition of the 10th layer, the substrate was rotated by 30°. The deposition pressure was 14 mN/m, and the deposition speed was 2 mm/min.
before (Figure 5). The structure of the stripes is not influenced by the underlying surface; the stripes cross the areas of different hydrophobicity without visible perturbation. However, it is interesting to note that the differences in the polarity, which the layer experiences during deposition, range from the stripes being much wider (a factor of 2.5) compared to those deposited on untilted substrates. This result clearly shows that the forces occurring during the LBK deposition process play a key role in the stripe formation process rather than the structure of the underlying surface. To elucidate the influence of the deposition conditions on the terrace formation, the compression conditions before film deposition were varied. AFM micrographs of the azoPAA multilayer consisting of 10 monolayers transferred at a speed of 2 mm/min and four different surface pressures of 18, 22, 26, and 30 mN/m, respectively, are presented in Figure 6. All four values of transfer pressures are well below the collapse pressure for the azo-PAA polymer (42 mN/m) (Figure 1). A close examination of the area of 100 × 100 µm showed that there is no correlation between the size and structure of the formed stripes and the pressure used within the range between 14 and 30 mN/m. In all four cases, lines from 1 to 10 µm in width are formed during the deposition process starting from the 7th layer. The coverage of the 7th, 8th, 9th, and 10th layer is the same as that in the case described above, where the transfer pressure was 14 mN/m. Above this range, the transfer procedure leads to the formation of irregular striped structures with ragged edges. If the transfer pressure was further lowered, the stripe formation phenomenon disappeared completely (Figure 7). Below 8 mN/m, very homogeneous films were obtained independent from the number of layers deposited. The topmost 10th layer of the azo-PAA is essentially flat with a roughness of σrms ) 0.5 nm. However, the layer is strewn with holes of 100 ( 30 nm in diameter and a depth of 1.6 ( 0.3 nm, which corresponds to the height of one molecular layer. The holes are not randomly distributed within the film but form characteristic lines (marked by a dashed white line in Figure 7) that are perpendicular to the dipping direction. The fraction of holes matching the lines is 85%, while the rest of the holes cannot be ascribed to a certain arrangement. It is suggested that these holes are precursors of the line formation at higher transfer pressures. Indeed, increasing the transfer pressure to 10 mN/m during deposition leads to a situation where the two
Figure 6. AFM micrographs of the 10 azo-PAA layers transferred from the Langmuir trough onto silanized silicon wafer, at a dipping speed of ν ) 2 mm/min and different surface pressures: (a) π ) 18 mN/m, (b) π ) 22 mN/m, (c) π ) 26 mN/m, and (d) π ) 30 mN/m. Arrows indicate the dipping direction. The z range for all micrographs is between 0 and 10 nm.
Figure 7. AFM micrographs of the 10 azo-PAA layers transferred from the Langmuir trough onto a silanized silicon wafer at a surface pressure of π ) 8 mN/m and a dipping speed of ν ) 2 mm/min.
structures, stripes and holes, coexist in the transferred layer (Figure 8). The depth of the holes in this case is 1.6 nm as well. The enhanced micrographs inserted in Figure
8254
Langmuir, Vol. 21, No. 18, 2005
Santer et al.
Figure 9. Schematic representation of the mutual arrangement of the transferred layers in the LBK multiplayer of azoPAA. (a) As it appeared from the cross-sectional analysis of AFM images. (b) Possible real arrangement situation, where the stripe monolayer can “step down the pyramid”.
Figure 8. AFM micrographs of 10 azo-PAA layers transferred from the Langmuir trough onto a silanized silicon wafer at a surface pressure of π ) 10 mN/m and a dipping speed of ν ) 2 mm/min.
8 are lined up in such a way that they match the line of the stripes perpendicular to the dipping direction. This surface pressure seems to be critical for the rupture of the monolayer. The stripe formation phenomenon also strongly depends on the dipping speed at which the monolayers are transferred. Thus, if the dipping speed for the azo-PAA film deposition was reduced to 1 mm/min, the phenomenon described above disappeared. For the slow dipping procedure, no stripe formation was observed within a broad range of surface pressures varying from 5 to 26 mN/m. Using a dipping speed of 1 mm/min, successive deposition of homogeneous films up to 480 layers was achieved.14 One interesting aspect needs to be discussed: as described above, the layer coverage, i.e., the space filled by the molecules compared to the geometrical area, decreases with the increasing number of layers from 100 to 10%. However, in contrast to this, the transfer ratio, i.e., the ratio between the area of the deposited molecules in the Langmuir film and the substrate area was independent from the number of layers always between 0.9 and 1.1. One model, which would explain such a behavior, is that the incoming layer is not exclusively deposited on already existing layers but that an incoming stripe of one layer covers one of the layers deposited earlier (Figure 9). The parameters, which lead to the formation of the molecular terraces, are rather well-established, yet a number of important questions remain unsolved. It is not well-understood yet how the delicate balance between the macroscopic parameters, such as film compression and rate of deposition, and microscopic properties, such as polarity and side-chain length, governs the formation of the structure. Further, it remains unclear why the terrace
formation starts only at the 7th layer, then, however, very reproducible (measured for 100 samples). The only speculation that can be offered is that some mechanical flexibility of the substrate is required and that a totally rigid inorganic substrate strongly suppresses terrace formation. Our results, however, indicate that the general mechanism governing stripe formation is relatively complex, leaving the findings of ref 9 as a special case. In particular, intramolecular interactions certainly play a significant role, dictating if striation may start with the first monolayer or whether, as in our case, the deposition of several homogeneous layers is necessary before instabilities leading to stripe formation can occur. For instance, the molecular structure of the azo-PAA polymer, namely, the length of the CH2 spacer that connects the backbone and the azobenzene-chromophore group (Scheme 1), influences strongly the formation of the stripe patterns. Thus, polymer molecules having 6 CH2 groups in the side chain (compared to 4 CH2 groups in the case of the polymer discussed here) show homogeneous film formation without any striation during dynamic Langmuir-Blodgett-Kuhn transfer up to several hundred monolayers.15 Conclusions Macroscopic molecular terraces are spontaneously formed during the deposition of LBK monolayers of an azo-PAA. The height of the terraces corresponds to one monolayer (1.6 ( 0.2 nm). The terraces are all parallel to each other and at least hundreds of micrometers long. They are several micrometers wide, and although the width of an individual terrace is constant over macroscopic dimensions, the width of different terraces varies. The terraces are the result of a complex interplay between the macroscopic parameters of film deposition and the microscopic details of the polymer. Acknowledgment. Y. Z. thanks the Catholic Academic Exchange Service (KAAD) for a Ph.D. fellowship. LA051212I
