Atomic Force Microscopy Studies of Mesoscopic Membranous

On the basis of the AFM studies, we believe that the dome-shaped features are membranous air bubbles. The membranes of the bubbles were probably ...
2 downloads 0 Views 551KB Size
Download PDF
Langmuir 2005, 21, 5019-5025

5019

Atomic Force Microscopy Studies of Mesoscopic Membranous Bubbles on Monolayers Derived from SiCl3-Terminated Carbosilane Dendrons on Mica Zhongdang Xiao† and Chengzhi Cai* Department of Chemistry & Center for Materials Chemistry, University of Houston, Houston, Texas 77204 Received October 13, 2004. In Final Form: March 23, 2005 Monolayers of dendrimers were prepared on mica by spin-coating of the second generation carbosilane dendrons with 9 SiCl3 periphery groups. AFM images of the films showed the presence of soft yet robust, dome-shaped features with a base diameter of 100-2000 nm. The apparent height of the features, ranging from 10 to 200 nm, rapidly reduced under increasing compression force, eventually to the same value (∼2.5 nm) corresponding to a bilayer of the flattened dendrons. The change in shape of the features in response to the compression force from the AFM tip was fully reversible, indicating that the features were robust. The contrast of the features in the tapping mode AFM (TMAFM) phase images flipped at a setpoint ratio of ∼0.55. In contrast to the reported amplitude vs displacement (A/z) curves for compliant materials, A/z curves of the features showed that the reduction of amplitude was larger than the tip displacement as if the cantilever tip were repelled by the soft features. This result cautions the use of amplitude/phase vs displacement (APD) curves for interpreting TMAFM images and for optimizing conditions for TMAFM imaging of very soft and “sticky” surfaces. On the basis of the AFM studies, we believe that the domeshaped features are membranous air bubbles. The membranes of the bubbles were probably composed of a bilayer of the dendron molecules bound through the peripheral silanol groups. The bilayer could be formed by self-assembly of the molecules on top of the air bubbles entrapped at the monolayer/solution interface during spin-coating.

Introduction Organosiloxane self-assembled monolayers (SAMs) have attracted tremendous interests both as model systems for fundamental research on organic surface chemistry and as surface modifiers for applications including biochips, microelectromechanical systems (MEMS), and micro- or nanolithography.1-3 Organosiloxane SAMs are commonly derived from active precursors such as organotrichlorosilanes. Recent studies indicate that the active precursors are initially hydrolyzed by water present in the deposition solution and/or on the substrate surface.4 The resulting alkylsilanol species including oligomers physisorb preferentially with their silanol groups oriented toward the polar substrate surface. This physisorption step is often followed by diffusion, aggrega* To whom correspondence should be addressed. Tel: 713-7432710. Fax: 713-743-2709. E-mail: [email protected]. † Current address: Key Laboratory of Molecular and Biomolecular Electronics, Ministry of Education, Southeast University, Nanjing 210096, P. R. China. (1) (a) Ulman, A. Chem. Rev. 1996, 96, 1533. (b) Schwartz, D. K. Ann. Rev. Phys. Chem. 2001, 52, 107. (2) Schena, M. Microarray Biochip Technology; Eaton Publishing: Batick, MA, 2000. (3) (a) Hoeppener, S.; Maoz, R.; Sagiv, J. Nano Lett. 2003, 3, 761. (b) Dickie A. J.; Quist, F.; Sjitehead, M. A.; Kakkar, A. K. Langmuir 2004, 20, 4315. (c) Harada, Y.; Girolami G. S.; Nuzzo, R. G. Langmuir 2003, 19, 5104. (d) Jun, Y.; Zhu, X. Y. Langmuir 2002, 18, 3415. (4) (a) Resch, R.; Grasserbauer, M.; Friedbacher, G.; Vallant, Th.; Brunner, H.; Mayer, U.; Hoffmann, H. Appl. Surf. Sci. 1999, 140, 168. (b) Lambert, A. G.; Neivandt, D. J.; McAloney, R. A.; Davies, P. B. Langmuir 2000, 16, 8377. (c) Wang, Y.; Lieberman, M. Langmuir 2003, 19, 1159. (d) Sung, M. M.; Carraro, C.; Yauw, O. W.; Kim, Y.; Maboudian, R. J. Phys. Chem. B 2000, 104, 1556. (e) Liu, Y.; Wolf, L. K.; Messmer. Langmuir 2001, 17, 4329. (f) Vallant, T.; Brunner, H.; Mayer, U.; Hoffmann, H.; Leitner, T.; Resch, R.; Friedbacher, G. J. Phys. Chem. B 1998, 102, 7190. (g) Vallant, T.; Kattner, J.; Brunner, H.; Mayer, U.; Hoffmann, H. Langmuir 1999, 15, 5339. (h) Tripp, C. P.; Hair, M. L. Langmuir 1995, 11, 149. (i) Grange, J. D. Le.; Markham, J. L.; Kurkjian, C. R. Langmuir 1993, 9, 1749.

tion, and nucleation of the aggregates (islands) to form a monolayer where the alkyl chains are aligned vertically and their packing is stabilized by van der Waals interactions. Finally, the rather slow condensation of the silanol groups generates siloxane networks between adjacent molecules as well as the substrate surfaces containing hydroxy groups. The ability of alkylsilanols to undergo cross-linking increases the robustness of the films, but it also significantly impedes the formation of flat monolayers. During monolayer formation, the alkylsilanol species in the solution may bind vertically to the substrate surface to form large aggregates.5 In addition, it has been shown that the alkyl chains in the polymers formed by hydrolysis of OTS prefer to adapt a bilayer structure in which the adjacent alkyl chains are oriented in the opposite directions.6 The alkyl chains in such bilayer structures are better packed than the alkyl chains in siloxane SAMs where the packing density is limited by lattice mismatch with the substrate surface.7 This observation is relevant to the present work, suggesting the presence of bilayer structures in the aggregates formed during deposition of organosiloxane SAMs. Recently, we reported a new type of organosiloxane film derived from the second, third, and fourth generation carbosilane dendrons containing 9, 27, and 81 SiCl3 groups at the periphery and a functional group at the focal point of the dendrons.8 The large, focally functionalized dendron adsorbates with varied size (generation) could be used to (5) Fadeev, A. Y.; McCarthy T. J. Langmuir 2000, 16, 7268. (6) (a) Wang, R.; Baran, G.; Wunder S. L. Langmuir 2000, 16, 6298. (b) Parikh, A. N.; Schivley, M. A.; Koo, E.; Seshadri, K.; Aurentz, D.; Mueller, K.; Allara, D. L. J. Am. Chem. Soc. 1997, 119, 3135. (7) Stevens, M. J. Langmuir 1999, 15, 2773. (8) (a) Xiao, Z.; Cai, C.; Deng, X. Chem. Commun. 2001, 1442. (b) Xiao, Z.; Cai, C.; Mayeux, A.; Milenkovic, A. Langmuir 2002, 18, 7728. (c) Yam, C. M.; Mayeux, A.; Milenkovic, A.; Cai, C. Langmuir 2002, 18, 10274.

10.1021/la047471+ CCC: $30.25 © 2005 American Chemical Society Published on Web 04/21/2005

5020

Langmuir, Vol. 21, No. 11, 2005

precisely define the location of and spacing between the functional groups on the film surfaces.8b We deposited the films by spin-coating of the dendron on mica surfaces in air. The film thickness could be controlled by the concentration of the deposition solutions. AFM study showed that the morphology of these films, featuring rings, disks, dots, and/or holes, was strongly dependent upon the generation of the dendron and film thickness. It was conceivable that the presence of many SiCl3 groups in the dendrons would greatly facilitate the vertical cross-linking, generating large aggregates. Nevertheless, we demonstrated that flat monolayers could be reproducibly prepared from dendrons containing up to 81 SiCl3 groups under suitable spin-coating conditions and solution concentrations. For the second generation dendron 1, slightly above the suitable concentration range leading to monolayer formation usually generated disk structures on top of the monolayer. These structures were revealed by tapping-mode AFM images under hard tapping conditions (see below). We subsequently found that the height images of the monolayers of 1 containing disk structures8 changed substantially if the images were taken under lighter tapping conditions: the disk structures that were 1002000 nm in diameter and ∼2.5 nm in height turned into dome-shaped features; that is, their diameters remained the same but the maximum heights increased to 10-200 nm. These heights were proportional to the diameter of the features. The results of the AFM study described herein indicate that such dome-shaped features were membranous bubbles filled with air. The membrane of the bubbles was probably composed of a bilayer of the dendron molecules bound face-to-face with the peripheral silanol groups.

Xiao and Cai

referred to as light tapping, in which the operation is likely in the attractive regime. On the other hand, a low Asp/A0 ( 0.1 the reduction of amplitude was larger than the tip displacement as if the cantilever tip were repelled by the soft bubbles. This result cautions the use of A/z curves for optimizing TMAFM conditions. Herein we describe the AFM characterization of the large features formed on monolayers derived from the carbosilane dendron 1 with 9 SiCl3 peripheral groups. Our results indicate that these features are air bubbles with a membrane consisting of a bilayer of the dendrons and suggest that similar membranous bubbles may be formed on other alkylsiloxanes SAMs. The results of the TMAFM study of these films cautions the use of amplitude/phase vs displacement (APD) curves for interpreting TMAFM images and for optimizing TMAFM conditions. Experimental Section

In this study, tapping-mode AFM (TMAFM) was used to image the dome-shaped features. During TMAFM, an AFM cantilever tip oscillating close to its resonance frequency is brought close to a sample surface. When the tip starts to interact with the surface, its free oscillating amplitude (A0) is reduced. To obtain a topography image, the amplitude is maintained to a setpoint (Asp) through feedback control of the vertical displacement (z) of the tip. Mapping of z over the scanning area provides the height image. The tip-sample interaction also leads to a phase shift (Φ) of the oscillation. Keeping Asp constant while mapping Φ over the scanning area gives a phase image. The reduction of amplitude, which is the feedback for obtaining the topographic and phase images, is due to both attractive and repulsive tip-sample interactions.9 For a heterogeneous surface with components that interact with the tip differently, the “topography” or “height” images of the surface may not represent the real topography, and they may change as the tapping conditions are varied.10,11 Important tapping conditions include set point ratio (Asp/A0), free amplitude (A0), and frequency. With a high Asp/A0 (>0.8) and small A0, the conditions are (9) (a) Garcia, R.; San Paulo, A. Phys. Rev. B 1999, 60, 4961. (b) Kuhle, A.; Sorensen, A. H.; Bohr, J. J. Appl. Phys. 1997, 81, 6562. (10) Bar, G.; Thomann, Y.; Brandsch, R.; Cantow, H. J.; Whangbo, M. H. Langmuir 1997, 13, 3807.

The fleshly prepared carbosilane dendron 18 was dissolved in anhydrous THF to a concentration of 1 × 10-5 M. Spin-coating was carried out under ambient conditions in air at 22 °C and a relative humidity of 38-42% using a model WS-400A-6NPP spin coater (Laurell Tech. Co.). A drop of the solution of 1 was placed on fleshly cleaved muscovite mica (Structure Probe), and the substrate was immediately spun. The spin rate reached 2000 rpm within 5 s and was maintained at 2000 rpm for 115 s. The films were then dried in air. No attempt was made to systematically vary the deposition conditions. AFM studies were performed in ambient conditions with a Multimode Nanoscope IIIa (Digital Instruments). V-shaped (11) (a) Chen, X.; Davies, M. C.; Roberts, C. J.; Tendler, S. J. B.; Williams, P. M.; Davies, J.; Dawkes, A. C.; Edwards, J. C. Ultramicroscopy 1998, 75, 171. (b) Knoll, A.; Magerle, R.; Krausch, G. Macromolecules 2001, 34, 4159. (c) Wang, Y.; Song, R.; Li, Y.; Shen, J. Surf. Sci. 2003, 530, 136. (d) Raghavan, D.; Gu, X.; Nguyen, T.; VanLandingham, M.; Karim, A. Macromolecules 2000, 33, 2573. (e) James, P. J.; Antognozzi, M.; Tamayo, J.; McMaster, T. J.; Newton, J. M.; Miles, M. J. Langmuir 2001, 17, 349. (f) Kuhle, A.; Sorensen, A. H.; Zandbergen, J. B.; Bohr, J. Appl. Phys. A 1998, 66, S329. (g) Leclere, Ph.; Dubourg, F.; Kopp-Marsaudon, S.; Bredas, J. L.; Lazzaroni, R.; Aime, J. P. Appl. Surf. Sci. 2002, 188, 524. (h) Chen, X.; Roberts, C. J.; Davies, M. C.; Tendler, S. J. B. Surf. Sci. 2002, 519, L593. (i) KoppMarsaudon, S.; Leclere, P.; Dubourg, F.; Lazzaroni, R.; Aime, J. P. Langmuir 2000, 16, 8432. (j) Mallegol, J.; Dupont, O.; Keddie, J. L. Langmuir 2001, 17, 7022. (k) Raghavan, D.; VanLandingham, M.; Gu, X.; Nguyen, T. Langmuir 2000, 16, 9448. (l) Ebenstein, Y.; Nahum, E.; Banin, U. Nano Lett. 2002, 2, 945. (m) Stark, R. W.; Schitter, G.; Stemmer, A. Phys. Rev. B 2003, 68, no. 085401. (n) Bar, G.; Ganter, M.; Brandsch, R.; Delineau, L.; Whangbo, M. H. Langmuir 2000, 16, 5702. (o) Haugstad, G.; Jones, R. R. Ultramicroscopy 1999, 76, 77. (12) Godehardt, R.; Bebek, W.; Adhikari, R.; Rosenthal, M.; Martin, C.; Frangov, S.; Michler, G. H. Eur. Polym. J. 2004, 40, 917.

Mesoscopic Membranous Bubbles on Monolayers

Langmuir, Vol. 21, No. 11, 2005 5021

Figure 1. TMAFM height images (a-d, 10 × 10 µm2; e, 5 × 2.3 µm2) of a film of 1 containing membranous air bubbles, obtained with A0 ) 107 nm and Rsp ) 0.63 (a), 0.47 (b), 0.29 (c), 0.10 (d), and 0.85 (e). The contrast corresponds to 50 nm for (a)-(c) and 10 nm for (d) and (e). The inset in (e) is the cross-section plot of the bubble shown in (e). silicon nitride cantilevers (Digital Instruments) with a typical length of 200 µm and force constants of ∼0.1 N/m were used for contact-mode AFM. Hydrophobic tips were prepared by surface hydrosilylation of hydrogen-terminated silicon AFM cantilever tips with 1-hexadecene using the procedure described in ref 13. To ensure that the film was intact under the specific imaging conditions, we acquired the “zooming out” images containing the previously scanned area with a lower loading force. For TMAFM, Ultasharp silicon cantilevers (MDT, Moscow) with a typical length of ∼130 µm, resonant frequency of 150 kHz, and spring constant of 4.5 N/m were used. A/z curves were obtained by repeating the approach-retraction circle at 1 Hz. Only the approaching A/z curves were presented. Calibration11a of the amplitude of cantilever oscillation to the photodiode signal was performed with an atomically flat silicon wafer. Nine A/z curves were recorded at different locations, and the slopes of the linear curves were averaged to provide the cantilever factor. During TMAFM, the height and phase images were recorded simultaneously. We usually started the TM imaging at a high Rsp and gradually decreased the Rsp close to the limit beyond which the imaging became unstable.

Results and Discussion Height and Phase Tapping-Mode Images. Figure 1 presents the height images of a film containing the (13) Gu, J.; Yam, C. M.; Li, S.; Cai, C. J. Am. Chem. Soc. 2004, 126, 8098.

aforementioned dome-shaped features. These images were acquired at the same location on the film with the same free amplitude (A0 ) 107 nm) but different setpoint ratios (Rsp), decreasing from 0.63 in (a), 0.47 in (b), and 0.29 in (c) to 0.10 in (d). It should be noted that although a few images such as the one shown in Figure 1e were obtained with Rsp > 0.7, under such light tapping conditions the tip easily lost contact with the surface when it encountered a bubble, and the periodic features similar to the one shown in the inset of Figure 1e were often observed. To obtain stable images, a low Rsp value ( 0.55) the attractive forces still dominate the tip-sample interactions. The unusually strong tip-sample attractive interactions in our system can be attributed to the presence of silanol groups in the membrane of the bubbles and the large contact area between the hydrophilic tip and the soft, “sticky” bubbles (vide infra). It should be noted that the hydrophilic silicon AFM tips were most likely contaminated with the dendrons that contain many silanol groups that can still strongly interact with the bubbles. Indeed, switching of phase from positive to negative occurred at much higher Rsp values (>0.87) when the hydrophobic, hexadecyl-coated silicon tips were used, thus indicating that the strong interaction between the (contaminated) silicon tip and sample is likely the cause for phase flipping at low Rsp value (0.55). The apparent height (H) and phase shift (Φ) at the middle of a bubble (∼1.25 µm in base diameter) as a

function of Rsp with A0 ranging from 19 to 150 nm are plotted in Figure 3. These curves show that for all A0 both H and Φ decreased with decreasing Rsp and reached about the same values at Rsp ) 0.10. At a given Rsp, both H and Φ decreased with increasing A0, and the transition of phase from positive to negative occurred at a higher Rsp with increasing A0. These general trends can be rationalized in the following way. At a given Rsp, larger A0 results in a stronger compression force and lower H. Also, the repulsive force from the bubble increases with increasing A0 and decreasing Rsp. Therefore, the phase flipping due to switching from attractive to repulsive regimes takes place at higher Rsp for larger A0. Amplitude-Distance Curves. The above results show that the TMAFM images of the bubbles were strongly influenced by the setpoint ratio and free amplitude, similar to many compliant polymer films. The convolution of the TMAFM images of compliant materials due to the tipsample interactions was investigated by plotting the amplitude (A) as a function of the vertical position (z) of the approaching tip. The reported A(z) curves showed that even at a high Asp, tip indentation into the compliant materials occurred.11,12 Similar to the reported A(z) curves, Figure 4a shows a typical A(z) curve obtained on a multilayer film with a thickness of ∼100 nm prepared by spin-coating of a high concentration carbosilane dendron solution on mica surface followed by annealing at 100 °C. We assume that the cantilever oscillated with its free amplitude (A0) right before it started to touch the film, and this location is set as z ) 0. The amplitude of the cantilever oscillation was reduced from A0 to A as the cantilever further approaches the film with a displacement of z. If the substrate surface is infinitely hard, as shown by the straight line (H) in Figure 4a, the amplitude (A) decreases linearly with z and the slot of the line is unity. Compared to a hard surface, the reduction of amplitude (A0 - A) of the cantilever tip by the film was less than the

Mesoscopic Membranous Bubbles on Monolayers

Figure 3. Height (H, a) and phase (Φ, b) vs setpoint ratio (Rsp) at a series of amplitudes (A0) for a bubble on TMAFM images obtained under the corresponding Rsp and A0. The lines are used to guide the eyes.

displacement (z) due to indentation (zind) of the tip into the film. The indentation (zind) of the tip at various A/A0 can be readily obtained from the A(z) curve as showed in Figure 4a. Under the assumption that the tip indentations (zind) obtained from the A(z) curves are similar to the ones measured by the TMAFM images at the corresponding A/A0 (Asp) values, A(z) curves have been used to reconstruct the “real” surface of polymer films composed of soft and hard regions, as well as for optimizing TMAMF conditions to increase the contrast.11,12 However, the following result suggests that this assumption may not be valid for some systems. As shown in Figure 4b, the A(z) curve obtained with the tip oscillating on top of a bubble was remarkably different from those reported on compliant material surfaces. The amplitude of the cantilever oscillation decreased faster than the cantilever displacement from the point where the tip started interacting with the bubble surface until it nearly ceased vibrating at point I in the curve. Rather than indentation into the soft film surface such as the situation shown in Figure 4a, the oscillating tip on the bubble appeared to be repelled by the soft bubble by a distance (h) between the lowest point of the oscillating tip and the steady bubble surface (Figure 4b). However, it should be noted that the soft bubble becomes dynamic upon interacting with the tip, and most likely the elastic and soft bubble periodically attaches to the oscillating tip. In fact, we have shown that the polycondensation of the silanol groups in the dendrons cannot be completed at room temperature, and the dendron film surface is hydrophilic prior to curing at elevated temperatures.8 Since the silicon tip surface is also hydrophilic, strong attractive forces (capillary force and hydrogen bonding) are established when the tip touches the surface. The “sticky” bubbles are also very soft, thus can easily stick

Langmuir, Vol. 21, No. 11, 2005 5023

Figure 4. Typical amplitude-distance (A-z) curves obtained on a 100 nm thick, annealed carbosilane dendron film (a) and on a membranous bubble (b). The cantilever tip is assumed to oscillate with free amplitude A0 when it touches the sample surface at z ) 0 (left part of the insets). The oscillating amplitude is reduced to A upon the tip approaching to the film by z (right part of the insets). The straight line H with a slot of unity in (a) and (b) represents an infinitely hard surface, where the reduction of amplitude (A0 - A) is equal to the displacement (z). The A-z curve (a) of the multilayer film shows that the reduction of amplitude (A0 - A) is smaller than the displacement (z) by zind, that is, the tip indents into the film by zind (right part of the inset in (a)), while the A-z curve of the bubble shows that the reduction of amplitude (A0 - A) is larger than the displacement (z) by h; that is, the oscillation is rapidly damped by the soft, “sticky” bubble (right part of the inset in b).

to and be pulled up by the tip. As estimated by the contact mode AFM images (vide infra), the bubble displayed a force constant about 0.3-0.8 N/m, which was much less than that of the cantilever (∼4.5 N/m). The softness and the “sticky” nature of the bubbles should increase the contact area and contact time between the bubble and the oscillating tip, thus significantly increasing the energy dissipation. Furthermore, the strong tip-surface interactions also shift the resonance frequency of the cantilever.11 All these factors can cause rapid damping of the cantilever oscillation. The A-z curve in Figure 4b shows that reduction of the amplitude A or the setpoint ratio A/A0 did not lead to indentation of the approaching tip to the bubble prior to the point I. This result is contradictory to the result of TMAFM imaging with a series of setpoint ratios A/A0 (Figures 1 and 3a), which shows that the apparent height of the bubbles reduced with decreasing setpoint ratios, i.e., harder tapping. The discrepancy must be due to the different conditions for the two experiments. The A-z curves were recorded during the approach of the oscillating

5024

Langmuir, Vol. 21, No. 11, 2005

Figure 5. Contact-mode AFM images (5 × 5 µm2, 20 nm contrast) of a bubble obtained with a setpoint of -3.24 V (a), -4.36 V (b), -4.6 V (c), and -5.0 V (d).

cantilever tip to the bubble. For TMAFM imaging, A/A0 is maintained by adjusting the z position of the cantilever tip that is rapidly scanning the sample surface. For some systems, the average duration and strength of the interactions between the tip and sample at the same amplitude may be quite different for the two experiments. The A(z) curve in Figure 4b was obtained on the middle of a bubble. We also measured h on other locations of the same bubble under the same Asp and found that they varied in the range of about 5-30 nm and decreased for locations closer to the edge of the bubble. Apparently, the oscillation

Xiao and Cai

of the tip dumped faster at the soft part (middle) of the bubble than near the edge. Contact Mode AFM. The bubbles were also imaged by contact mode AFM with a series of setpoints. The applied forces corresponding to the setpoints were estimated from the calibration curves and force constant of the cantilever. A typical image of a bubble obtained at a setpoint of -3.24 V (33.4 nN) is shown in Figure 5a, where the bubble was completely pressed to the substrate surface by the tip and appeared as a faint disk with a height of ∼2 nm. When the setpoint was slightly decreased to -4.36 V (25.7 nN), the disk in Figure 5a grew into a dome feature (bubble) with a height of 11 nm (Figure 5b). Further decrease of the setpoint to -4.60 V (21.5 nN) and -5.00 V (15.8 nN) resulted in further growth of the bubble, and its height increased to 19 nm (Figure 5c) and 36 nm (Figure 5d), respectively. Stable images could not be obtained with setpoints lower than -5 V. The force constant of the bubbles can be roughly estimated by the difference of two applied forces divided by the induced change in height. For example, reducing the applied force from 15.8 to 12.6 nN led to a reduction of height from 44 to 36 nm; thus, the force constant of the bubble is estimated to be (15.8 - 12.6)/(44 - 36) ) 0.40 N/m. In this way, we estimate that the force constants at different compression stage varied from 0.33 to 0.76 N/m. The contact mode images of the bubbles obtained with the same setpoint were fully reproducible, indicating that the bubbles were robust. The results from the tapping and contact mode AFM imaging suggest that the membranous bubbles are filled with air, allowing a small portion of the bubble to be easily pressed to the base with a relatively low loading force (33 nN), while recovering when the force is released. Proposed Mechanism for the Formation of Membranous Bubbles. During spin-coating in air, a thin film of the THF solution of 1 is formed on mica (Figure 6a). The SiCl3 groups of 1 rapidly hydrolyze into silanol groups

Figure 6. Illustration of the proposed mechanism for the formation of a membranous bubble on a monolayer prepared by spincoating of 1 on mica surface. The SiCl3 groups in 1 rapidly hydrolyze into silanol groups, and the resulting silanol-terminated dendrons rapidly absorb on mica to form a monolayer. Meanwhile, air bubbles are formed on the monolayer (a). Concurrent to solvent evaporation, self-assembly of the excess of the dendron molecules on the air bubbles generates a bilayer where the silanol peripheral groups of the dendrons are bound face-to-face to form a robust network consisting of covalent bonds (Si-O-Si) and hydrogen bonds between the silanol groups.

Mesoscopic Membranous Bubbles on Monolayers

during this process. Upon absorption on mica surface, the silanol-terminated dendron molecules are expected to flatten on the surface to maximize the interaction between the polar peripheral silanol groups and the mica surface.8 This orientation exposes the hydrophobic part of the molecule on the substrate surface. We have shown that at a suitable concentration the dendron molecules cooperatively and rapidly absorb on mica surfaces to form a monolayer. The excess of dendron molecules on top of the monolayer is quite mobile and accumulates in the droplets formed during solvent evaporation.8 It is possible that mesoscopic air bubbles are entrapped at the monolayersolution interface. Indeed, formation of air bubbles at solid-liquid interfaces has been observed in several systems and has attracted increasing interest since such air bubbles significantly impede the preparation of high quality thin films.14 In our system, around the air bubbles, a bilayer of the dendrons may be formed during evaporation of the solvent (Figure 6b). In the bilayer, the dendron molecules bind face-to-face with the peripheral silanol groups, leaving the hydrophobic part of the molecules preferentially toward air. Partial condensation of the peripheral silanol groups is expected, leading to a siloxane network in the bilayer which greatly increases the robustness of the bilayer. However, we have shown that without annealing the films at elevated temperatures, condensation of the silanol groups in such dendron films cannot be completed after 3 weeks at room temperature (the films can still be scratched with an AFM tip under high applied force). Hence, substantial amounts of silanol groups should still remain in the bilayer prepared and (14) (a) Ishida, N.; Inoue, T.; Miyahara, M.; Higashitani, K. Langmuir 2000, 16, 6377. (b) Switkes, M. Rbuerti, J. W. Appl. Phys. Lett. 2004, 84, 4759. (c) Mao, M.; Zhang, J.; Yoon, R.-H.; Ducker, W. A. Langmuir 2004, 20, 1843. (d) Tyrrell, J. W. G.; Attard, P. Phys. Rev. Lett. 2001, 8717, art. no. 176104.

Langmuir, Vol. 21, No. 11, 2005 5025

maintained at room temperature and could interact with the AFM tip in contact with the bilayer. Conclusions Membranous air bubbles could be formed on siloxane monolayers prepared by spin-coating of the SiCl3terminated dendron 1 on mica. The bubbles were 0.1-2 µm in diameter and 10-200 nm in height. The bubbles were robust; they could withstand large vertical and lateral forces applied with a scanning AFM tip in tapping and contact mode. The thickness of the membrane was about 2.5 nm, corresponding to a bilayer of the dendron molecules. The bubbles were probably formed by selfassembly of the silanol-terminated dendrons on top of the air bubbles generated on the monolayer surface during spin-coating. Our results suggested that membranous air bubbles consisting of bilayer of siloxane SAMs could be produced in other systems as well. Another important finding in this study is that the amplitude-displacement (A/z) curves obtained on the bubbles were in sharp contrast to the result of TMAFM imaging of the bubbles. While the TMAFM images showed that the bubbles were readily compressed by the AFM tip with decreasing setpoint ratios, A/z curves of the bubbles showed that the oscillation of the cantilever tip was rapidly dumped by the bubble, as if the tip were repelled by the soft bubbles. This result cautions the use of amplitude/phase vs displacement (APD) curves for interpreting TMAFM images and optimizing TMAFM conditions. Acknowledgment. This work was supported by The Robert A. Welch Foundation, the Petroleum Research Fund (Type G), and the National Science Foundation (Grant CTS-0210840). LA047471+