pubs.acs.org/Langmuir © 2009 American Chemical Society
Autophobic Dewetting of a Poly(methyl methacrylate) Thin Film on a Silicon Wafer Treated in Good Solvent Vapor Longjian Xue and Yanchun Han* State Key Laboratory of Polymer Physics and Chemistry, Changchun Institute of Applied Chemistry, Chinese Academy of Sciences, Graduate School of the Chinese Academy of Sciences, 5625 Renmin Street, Changchun 130022, PR China Received December 19, 2008. Revised Manuscript Received February 4, 2009 The wettability of thin poly(methyl methacrylate) (PMMA) films on a silicon wafer with a native oxide layer exposed to solvent vapors is dependent on the solvent properties. In the nonsolvent vapor, the film spread on the substrate with some protrusions generated on the film surface. In the good solvent vapor, dewetting happened. A new interface formed between the anchored PMMA chains and the swollen upper part of the film. Entropy effects caused the upper movable chains to dewet on the anchored chains. The rim instability depended on the surface tension of solvent (i.e., the finger was generated in acetone vapor (γacetone = 24 mN/m), not in dioxane vapor (γdioxane = 33 mN/m)). The spacing (λ) that grew as an exponential function of film thickness h scaled as ∼h1.31, whereas the mean size (D) of the resulting droplets grew linearly with h.
Introduction
energy (ΔG) can be expressed as
Thin polymer films are increasingly used in numerous fields as dielectric coatings, resist layers for lithography, electronic packaging, optical coatings, nonlinear optical devices, lubricating surfaces, and so forth.1-3 In most cases, homogeneity, uniform thickness, and durability are essential. However, under thermal treatment4-12 and some stringent environmental conditions,13-15 the films break up and spontaneously dewet the substrate, resulting in the formation of droplets. The dewetting process has been studied both experimentally4-17 and theoretically.17-22 Considering the antagonistic long- and shortrange interactions, the excess intermolecular interaction free
*To whom correspondence should be addressed. E-mail: ychan@ciac. jl.cn. (1) Xia, Y.; Kim, E.; Zhao, X. M.; Rogers, J. A.; Prentiss, M.; Whitesides, G. M. Science 1996, 273, 347. (2) Dittinhhsud, H.; Tessler, N.; Friend, R. H. Science 1998, 280, 1741. :: (3) Jager, E. W. H.; Smela, E.; Inganas, O. Science 2000, 290, 1540. (4) Sharma, A. Langmuir 1993, 9, 861. (5) Redon, C.; Brzoska, J. B.; Brochard-Wyart, F. Macromolecules 1994, 27, 468. (6) Sharma, A.; Khanna, R. Phys. Rev. Lett. 1998, 81, 3463. (7) Redon, C.; Brochard-Wyart, F.; Rondelez, F. Phys. Rev. Lett. 1991, 66, 715. (8) Xie, R.; Karim, A.; Douglas, J. F.; Han, C. C.; Weiss, R. A. Phys. Rev. Lett. 1998, 81, 1251. (9) Limary, R.; Green, P. F. Langmuir 1999, 15, 5617. (10) David, M. O.; Reiter, G.; Schultz, J. Langmuir 1998, 14, 5667. (11) Masson, J.; Green, P. F. Phys. Rev. Lett. 2002, 88, 205504. (12) Reiter, G.; Sharma, A.; Casoli, A.; David, M.; Khanna, R.; Auroy, P. Langmuir 1999, 15, 2551. (13) Lee, S. H.; Yoo, P. J.; Kwon, S. J.; Lee, H. H. J. Chem. Phys. 2004, 121, 4346. (14) Xu, L.; Shi, T.; An, L. Langmuir 2007, 23, 9282. (15) Xing, R.; Luo, C.; Wang, Z.; Han, Y. Polymer 2007, 48, 3574. (16) Xue, L.; Cheng, Z.; Fu, J.; Han, Y. J. Chem. Phys. 2008, 129, 054905. (17) Xue, L.; Hu, B.; Han, Y. J. Chem. Phys. 2008, 129, 214902. (18) Reiter, G. Langmuir 1993, 9, 1344. (19) Brochard, F.; Daillant, J. Can. J. Phys. 1990, 68, 1084. (20) Brochard-Wyart, F.; Redon, C.; Sykes, C. C. R. Acad. Sci., Ser. 2 1992, 314, 19. (21) Brochard-Wyart, F.; de Gennes, P.-G.; Hervert, H.; Redon, C. Langmuir 1994, 10, 1566. (22) Bandyopadhyay, D.; Sharma, A. J. Chem. Phys. 2006, 125, 054711.
Langmuir 2009, 25(9), 5135–5140
ΔG ¼ -A=12πh2 þ S P expð -h=lÞ
ð1Þ
Here, A is the Hamaker constant, SP is the polar component of the spreading coefficient representing short-range polar interactions, and l is the decay length characterizing the range of these interactions. For the dewetting system, there are two classes of the form of free energy: the type I system where both long-range apolar and short-range polar forces are attractive (A > 0, SP < 0) and the type II system, where a long-range attraction (A > 0) combines with a short-range repulsion (SP > 0). The type II system has been called pseudodewetting23 or autophobic dewetting.24 After autophobic dewetting, the resulting droplets reside on the remaining thin film, 23,25 and the resulting droplets reside directly on the nonwettable substrate in the usual dewetting. To trigger the dewetting, most studies use the thermal treatment, but solvent-induced dewetting has started to draw considerable attention in recent years.12-15 The long-range force of van der Waals interactions in the thermal dewetting causes the instability of the film, whereas it is the short-range force of polar interactions in the solvent-induced dewetting. In these reports on solvent-induced dewetting, both a solvent and nonsolvent have been used to initiate the dewetting of the films. Using toluene (apolar) and acetone (polar), of which both are solvents for polystyrene (PS), Lee and co-workers13 found that the polarity of solvent influences the rim instability of a PS film on a silicon wafer. With apolar solvent toluene, the rim undulation did not lead to finger formation (type A instability), and the rim became stable as the film thickness increased. With polar solvent acetone, the rim undulation led to finger formation and the subsequent decay into droplets (type B instability). As the film thickness increased, the rim instability changed from a type B to type A instability. Xu and (23) Brochard-Wyart, F.; di Meglio, J.-M.; Quere, D.; de Gennes, P.-G. Langmuir 1991, 7, 335. (24) Sharma, A.; Khanna, R. Phys. Rev. Lett. 2000, 85, 2753. (25) Sharma, A. Langmuir 1993, 9, 3580.
Published on Web 3/11/2009
DOI: 10.1021/la8041814
5135
Article
Xue and Han
co-workers14 studied the dewetting of the PS film on hydrophilic surfaces using water as the trigger. They proposed the penetration-replacement mechanism to explain such a scenario. Both studies13,14 showed the dewetting of films on a nonwettable substrate in the solvent vapor. Here in our study we showed that the poly(methyl methacrylate) (PMMA) thin film, which is stable on the SiOx/Si substrate under thermal annealing, experienced the dewetting process in the good solvent vapor. It was found that the good solvent vapor can trigger the dewetting of the film whereas the nonsolvent cannot. A new interface formed between the anchored PMMA chains and the swollen upper part of the film. Entropy effects then drove the upper movable chains to dewet the anchored layer. The rim instability depended on the surface tension of the solvent. The finger was generated in acetone vapor (γacetone = 24 mN/m), not in dioxane vapor (γdioxane = 33 mN/m), and the spacing between the fingers and the mean size of the resulting droplets were found to depend on the initial film thickness.
Experimental Section Materials. PMMA samples with Mw = 240 000 g/mol (Mw/ Mn = 1.3) and Mw = 7800 g/mol (Mw/Mn = 1.3) were purchased from Polymer Source, Canada. Solvents acetone, dioxane, and cyclohexane were procured from Beijing Chemical, China. The water used here was deionized water. Polished test-grade silicon wafers were purchased from General Research Institute for Nonferrous Metals, China. Silicon wafers were cleaved into pieces of approximately 1 cm 1 cm. The silicon wafers were boiled in piranha solution (7/3 v/v 98% H2SO4/30% H2O2) for 30 min to remove the stains on the surface and then washed with deionized water and dried under nitrogen flow. Solvent Vapor Treatment of PMMA Thin Films. Thin films were prepared by spin coating PMMA dioxane solution onto the silicon wafer and drying in a vacuum oven for 24 h. The film thickness was controlled by the concentration of solution and the spinning rate. The films were then exposed to saturated vapors of different solvents (acetone, dioxane, cyclohexane, and water) in a closed chamber at room temperature for different periods.26 For comparison, PMMA films with Mw = 240 000 g/mol and 7800 g/mol were also annealed at 160 °C for 24 h in an N2 atmosphere. To investigate the interaction between PMMA molecules and the substrate, a 60-nm-thick PMMA film was immersed in 100 mL of acetone for 10 h at room temperature. The resulting film was rinsed with acetone and deionized water and then dried under nitrogen flow. The film was kept in a vacuum oven before characterization. Characterization. The dewetting process of the PMMA film was imaged ex situ by use of Leica optical microscopy (Leica Microsystems Ltd., Germany) in reflection mode with a CCD camera attachment. Holes appear as brighter locations in the optical micrographs. Rim and hole topography showed a high contrast due to height differences. The film thickness was measured by spectroscopic ellipsometry over a wavelength range of 300-800 nm at a fixed incident angle of 70° (spot size 1 mm) using a UVISEL spectroscopic ellipsometer (Jobin Yvon, France). The dewetted area between the resulting droplets was measured with a 50 μm spot. Atomic force microscopy (AFM) was used to study the surface topography of the spin-coated and annealed films. The largest scanning area is 150 μm 150 μm. Images were obtained using a SPI3800N AFM (Seiko Instruments Inc., Japan) with a Si tip with a spring constant of 2 N/m. The cantilevers were (26) Fukunaga, K.; Hashimoto, T.; Elbs, H.; Krausch, G. Macromolecules 2002, 35, 4406.
5136
DOI: 10.1021/la8041814
operated slightly below their resonance frequency of around 72 kHz. The image acquisition was performed under ambient conditions. The AFM was used in tapping mode to reduce tipinduced surface degradation and sample damage. Imaging was conducted in height mode. X-ray photoelectron spectroscopy (XPS) measurement of the sample was carried out with an ESCALAB 250 (Thermo electron Co., America) spectrometer with an Al KR (hν =1486.6 eV) mono X-ray source at an emission angle of 90° originating from the surface.
Results and Discussion Thermal Stability. The wettability of the PMMA film (Mw = 240 000 g/mol, about 60 nm) on the silicon wafer with a 1.7 nm native oxide layer was first examined by heating the film under an N2 atmosphere for 24 h. Before the thermal annealing, the film showed a smooth surface with rms = 0.3 nm. After thermal annealing, the film showed a rougher surface (rms =1.9 nm, Figure 1a), but no dewetting happened. To exclude the possibility of a dynamic cause because the PMMA used here has a large molecular weight, we annealed another PMMA film with a much smaller molecular weight (Mw = 7800 g/mol) of the same thickness for 24 h. No dewetting occurred (Figure 1b). The spreading of a film on a substrate can be described in terms of the effective Hamaker constant (Aeff). Aeff for the PMMA film on the silicon wafer with a native oxide layer can be given by the mixing rule involving individual Hamaker constants as follows12 pffiffiffiffiffiffiffiffiffiffi pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi pffiffiffiffiffiffiffiffi pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi Aeff ¼ ð ASiO - APMMA Þð Aair - APMMA Þ
ð2Þ
where ASiO, APMMA, and Aair are the Hamaker constants for the native oxide layer, the PMMA film, and air, respectively. Aeff < 0 corresponds to thermodynamic stability and perfect wetting, and Aeff > 0 implies an attractive force leading to spinodal dewetting.27 From the literature, we find ASiO = 2.2 10-20 J,28APMMA = 2 10-20 J,29 and Aair = 0, such that Aeff = -9.76 10-22 J can be obtained. This confirms that PMMA could stay stable on the SiOx/Si substrate and no dewetting would occur upon thermal annealing at a temperature above its Tg. Solvent-Vapor-Induced Dewetting. It becomes quite obvious that the solvent property influences the wettability of PMMA on the SiOx/Si substrate. Typical micrographs of films annealed in different solvent vapors are shown in Figure 2 for nonsolvents (cyclohexane and water) and in Figure 3 for good solvents (acetone and dioxane), respectively. No dewetting happened in nonsolvent vapor treatment even for 2 months. Three-dimensional morphologies of films annealed in cyclohexane and water vapor for 2 months are shown in Figure 2a,b, respectively. Both show a continuous film with protrusions on the surface. The heights of the protrusions on the films annealed in cyclohexane vapor and water vapor are 52.3 ( 49.4 and 25.9 ( 15.2 nm, respectively. However, the number of protrusions on the film annealed in water vapor is much more than that annealed in cyclohexane vapor. Cyclohexane, which has weak interaction with PMMA, can hardly penetrate into the film with only a few (27) Sharma, A.; Mittal, J. Phys. Rev. Lett. 2002, 89, 186101. (28) Isaellachivili, J. Intramolecular Surfaces Forces, 2nd ed.; Academic Press: New York, 1992. (29) Sferrazza, M.; Xiao, C.; Jones, R. A. L.; Bucknall, D. G.; Webster, J.; Penfold, J. Phys. Rev. Lett. 1997, 78, 3693.
Langmuir 2009, 25(9), 5135–5140
Xue and Han
Article
Figure 1. AFM images of a 60 nm PMMA film with Mw = (a)
240 000 g/mol and (b) 7800 g/mol annealed at 160 °C under N2 protection for 24 h.
Figure 2. Three-dimensional AFM images of a 60 nm PMMA film (Mw = 240 000 g/mol) annealed in a nonsolvent vapor: (a) cyclohexane and (b) water for 2 months, respectively.
Figure 3. Optical micrographs of the morphologies of 60 nm PMMA films annealed in good solvent vapor: in acetone vapor for (a) 1, (b) 4, (c) 12, (d) 24, and (e) 36 h and in dioxane vapor for (f) 1, (g) 4, (h) 12, (i) 24, and (j) 36 h, respectively; The scale bar equals 100 μm. The inset at the bottom left of panel a shows the AFM image of a small area.
protrusions generated (Figure 2a). PMMA can absorb about 2% water, though water is a nonsolvent for PMMA.30 The absorbed water swelled the film, causing a larger deformation on the film (Figure 2b). However, the PMMA films remain spread on the SiOx/Si substrate under nonsolvent vapors. Quite different from the thermal stability and the deformation in nonsolvent vapor, dewetting happened in good solvent vapors (both acetone and dioxane) as shown in Figure 3. As soon as the film was placed into the solvent vapor, the solvent vapor was absorbed into the film. The color change of the film indicated the absorption and an increase in film thickness. The dewetting processes were similar in two good solvent vapors. Protrusions and some tiny holes were generated on the films after the films were annealed in good solvent vapor for about 1 h (Figure 3a,f and inset of Figure 3a). The holes then grow larger as shown in Figures 3b-d (acetone vapor) and Figures 3g-i (dioxane vapor), respectively, and eventually the films decayed into droplets after being annealed in solvent vapor for about 36 h (Figure 3e,j). Though the dewetting processes were familiar, the instabilities of the rims were different in two types of good solvent vapor. In acetone vapor (γacetone = 24 mN/m), the rim undulation led to finger formation (Figure 3b-d). As the hole continued growing, the primary fingers increased in length and gradually broke up into droplets.13 The resulting droplets from a finger then showed the order of a line
(indicated by the red dashed line in Figure 3c). This is in contrast with the film treated in dioxane vapor (γdioxane = 33 mN/m), where the rim undulated at a certain wavelength, forming some short fingerlike structures (Figure 3h,i). The ends of these fingerlike structures were sharp (Figure 4a), whereas those on the film annealed in acetone vapor were round (Figure 4b). These short fingerlike structures remained in contact with the rim, and no droplets detached (Figure 3i). Finally, the rims touched the other ones and generated the classical polygon pattern of droplets (Figure 3j). The influence of the thickness of PMMA films on the dewetting behavior in acetone vapor was then investigated (Figure 5). It has been widely accepted that the dewetting velocity becomes slower when the film thicknesses increases. Here, 13 nm of PMMA came to the end of dewetting within about 4 h, whereas a 60 nm PMMA film needed about 24 h and a 264 nm PMMA film needed more than 200 h. The finger instability was then quantitatively analyzed by measuring the spacing (λ) between the fingers. It has been pointed out that the spacing (λ) between the fingers is constant after droplet formation and is influenced only by the film thickness.31,32 That is to say, for a film of thickness h, λ does not change after some droplets detached from fingers. The spacing λ is then extracted from the PMMA films with
(30) Brandrup, J.; Immergut, E. H.; Grulke, E. A.; Abe, A.; Bloch, D. R. Polymer Handbook, 4th ed.; Wiley and Sons: New York, 1999.
(31) Reiter, G.; Sharma, A. Phys. Rev. Lett. 2001, 87, 166103. (32) Xu, L.; Shi, T. F.; Dutta, P. K.; An, L. J. J. Chem. Phys. 2007, 127, 144704.
Langmuir 2009, 25(9), 5135–5140
DOI: 10.1021/la8041814
5137
Article
Xue and Han
Figure 4. AFM images of the rim structure of a PMMA film annealed in (a) dioxane and (b) acetone vapor.
Figure 6. (a) Double logarithmic plot of the spacing (λ) between the fingers as a function of the initial film thickness (h). (b) Dependence of the diameter (D) of the resulting droplets on the initial film thickness (h).
different thicknesses annealed in acetone vapor for different periods of time: (1) 13 nm, 3 h; (2) 26 nm, 9 h; (3) 60 nm, 12 h; (4) 113 nm, 24 h; and (5) 139 nm, 81 h as shown in Figure 5. Also, the measured λ is the mean value extracted from at least five holes, and the diameters (D) of resulting droplets on the films were also measured. The dependences of λ and D on the film thickness (h) are summarized in Figure 6a,b, respectively. λ grows as an exponential function of h and scales as ∼h1.31 (Figure 6a), which is consistent with the λ on the film under thermal annealing (λ ≈ h7/6).31,33 No finger was generated when the film thickness was 265 nm (Figure 5f). This implies that finger formation can happen only with film thickness below a critical value hc. 34 The Rayleigh instability then broke the fingers into droplets in which the size was not homogeneous. The mean size of the droplets was found to increase linearly with the initial film thickness h (Figure 6b).18 In a word, the characteristics of the rim instability under solvent vapor is the same as that on the film under thermal annealing.
Proposed Mechanism A new interface is generated in the PMMA film when the film is exposed to good solvent vapor. Solvent vapor absorbed into PMMA films because of the strong interactions, and the film can be regard as a solution. We cannot determine the exact concentration here; however, Lee13 pointed out that a similar system of a PS film in toluene vapor (or acetone vapor) has a concentration of between 5 and 10 wt %. According to Lee’s work, we then assumed the PMMA film in good solvent vapor to be a dilute solution, and the physical properties of the film were approximated to those of the solvent. Whatever the concentration, the distance among polymer chains in the solution is larger than that in the thin film, which causes the interchain interactions to be much weaker. These PMMA chains in the solution then possess higher mobility than in the dry thin film. The PMMA chains in the interfacial region near the SiOx/Si substrate undergo a very different experience from those in the upper part of the film (PMMA chains in solution). Previous work35,36 has revealed that PMMA has a favorable interaction with the hydroxyl groups on the surface of the SiOx/Si substrate, which can increase the Tg of PMMA in the form of a thin film. The intensity of the H bond is much stronger than the van der Walls interactions among the PMMA chains on that the PMMA chains are anchored onto the hydrophilic SiOx/Si substrate,36 and the strong affinity between PMMA molecules and solvent molecules can not overcome these H-bond interactions when the film was exposed to good solvent vapor. To confirm this, we immersed a PMMA film in a large amount of acetone for about 10 h and rinsed with acetone several times. PMMA molecules can still be detected from the substrate (Figure 7). The spectroscopic ellipsometry detection of a 5 nm residual PMMA layer on the
(33) Besancon, B. M.; Green, P. F. Phys. Rev. E 2004, 70, 051808. (34) Masson, J.; Olufokunbi, O.; Green, P. F. Macromolecules 2002, 35, 6992.
(35) Keddie, J. L.; Jones, R. A. L.; Cory, R. A. Faraday Discuss. 1994, 98, 219. (36) Forrest, J. A.; Dalnoki-Veress, K.; Stevens, J. R.; Dutcher, J. R. Phys. Rev. Lett. 1996, 77, 2002.
Figure 5. Optical micrographs of PMMA film with different thicknesses annealed in acetone vapor for different times: (a) 13 nm, 3 h; (b) 26 nm, 9 h; (c) 60 nm, 12 h; (d) 113 nm, 24 h; (e) 139 nm, 81 h; and (f) 264 nm, 120 h. The scale bar equals 100 μm.
5138
DOI: 10.1021/la8041814
Langmuir 2009, 25(9), 5135–5140
Xue and Han
Article
Figure 7. PMMA molecules have a favorable interaction with the hydroxyl groups on SiOx/Si substrate. XPS can detect these PMMA molecules remaining on the substrate after the PMMA film was washed with a large amount of acetone. dewetted area also suggests the existence of strong H-bond interactions. These strong, favorable interactions lead to high distortion and unfavorable conformations of the PMMA chains, and these unfavorable conformations would maintain the solvent vapor annealing process. The anchored PMMA chains then have less entropy than in the upper part of the film. With the help of solvent molecules, an entropy barrier and then a well-defined new interface between the anchored and moveable PMMA molecules is established. Thus, the system can be regard as a new system where a dissolved PMMA film is on the substrate modified with PMMA chains (Figure 8a). The new system is similar to the PS film on a PS brush surface. 37 The entropy effect drives the upper part of the PMMA film autophobic dewetting on the new interface. Lee13 suggested that the dispersive forces are a stabilizing factor (A < 0) and the polar interactions are the destabilizing factor (SP > 0) for PS films in both toluene and acetone vapors. The polar interactions dominate the dispersive forces, so dewetting happens in the film. The film was regarded as a pure solvent because of the low concentration, and their system can be simplified as an air/solvent/substrate model. In our system, the solvent was contained within both the upper part of the film and the anchored layer, and our system should be simplified as an air/PMMA/anchored-PMMA model. Because of the identical chemical composition of the upper part of the film and the anchored layer, the difference between the surface tensions due to the distortion is small and can be neglected.33 Thus, the spreading parameter S is determined entirely by the entropy. That is, entropy drove the PMMA film to dewet on the anchored PMMA layer,37-40 which should be called autophobic dewetting.23,25 In nonsolvent vapor, nonsolvent molecules can hardly penetrate into the film that they swell and cannot dissolve the film (Figure 8b). Because of the confined thin film structure, deformation in the form of protrusions happened on the film surface. Inside the swollen film, the interchain interactions are stronger than the interactions between PMMA chains and nonsolvent molecules, as evidenced by the fact that PMMA chains are still hard to move. PMMA chains on the SiOx/Si substrate form loops on the interface because of the H-bond interaction. Some PMMA chains across these loops work as cross-linkers between the anchored (37) Liu, Y.; Rafailovich, M. H.; Sokolov, J.; Schwarz, S. A.; Zhong, X.; Eisenberg, A.; Kramer, E. J.; Sauer, B. B.; Satija, S. Phys. Rev. Lett. 1994, 73, 440. (38) Hare, E. F.; Zisman, W. A. J. Phys. Chem. 1955, 59, 335. (39) Reiter, G.; Schultz, J.; Auroy, P.; Auvray, L. Europhys. Lett. 1996, 33, 29. (40) Zhu, D. S.; Liu, Y. X.; Chen, E. Q.; Li, M.; Chen, C.; Sun, Y. H.; Shi, A. C.; Van Horn, R. M.; Cheng, S. Z. D. Macromolecules 2007, 40, 1570.
Langmuir 2009, 25(9), 5135–5140
Figure 8. Proposed mechanisms for the wettability of a PMMA film on a SiOx/Si substrate. (a) Good solvent vapor induces the autophobic dewetting of the PMMA film. (b) Nonsolvent vapor swells the film, causing protrusions on the film surface. (c) PMMA film stays stable upon thermal annealing. chains and the chains in the upper part of the film. That is, the film in the nonsolvent vapor maintains its integrity, and no new interface would be generated. Coupled with the strong H-bond interaction between the film and substrate, the film would keep spreading the substrate in nonsolvent vapor. Similar to the situation in the nonsolvent vapor, the PMMA film under thermal annealing also maintains its integrity and stays stable on the substrate (Figure 8c). The competition between the disjoining pressure (Π, van der Waals forces) and the Laplace pressure, which relates to the surface tension of the solvent through P = 2γ/R (where P, γ, R are the Laplace pressure, surface tension of solvent, and radius of curvature, respectively), determines the formation of fingers. During the growth of a hole, a rim with an asymmetric shape formed at the perimeter of the hole because of the accumulation of PMMA molecules. This means that the upper part of the film slips on the anchored layer. The asymmetric shape of the rim is a result of the fact that frictional forces at the newly defined interface oppose the movement of the contact line. Whereas the driving force for the receding of the contact line is constant during hole growth, the friction forces increases with the size of the rim. In slipping polymer films, the dewetting prefers a constant dewetting velocity, which causes a fluctuation in the rim.31 Disjoining pressure would always drive materials from thinner to thicker parts of the rim.41 That is, the disjoining pressure would enlarge the fluctuation in the rim. Meanwhile, the Laplace pressure would flatten this fluctuation. The competition between the disjoining pressure and the Laplace pressure then determines the formation of fingers. When the disjoining pressure overcomes the Laplace pressure (in acetone vapor, γacetone = 24 mN/m), the solution accumulates in the thicker parts of the rim, causing an increase in the resist force such that the movement of the three-phase line at the thicker parts of the rim almost stops. Meanwhile, the thinner parts of the rim keep the same dewetting velocity. The fluctuation then lengthens, forming the finger structure. When the Laplace pressure dominates (in dioxane vapor, γdioxane = 33 mN/m), the fluctuation would be suppressed. During the annealing process, some large fluctuations would develop along the length, forming the fingerlike structure. The end of these fingerlike structures is sharp with defects on the contact line (41) Vrij, A.; Overbeek, J.; Th, G. J. Am. Chem. Soc. 1968, 90, 3074.
DOI: 10.1021/la8041814
5139
Article
Xue and Han
(Figure 4a), whereas the end of fingers in acetone is round (Figure 4b). This means that the fingerlike structure was generated because of the pinning effect of the defects on the substrate. When passing these defects, the three-phase line would return to its round shape because of the Laplace pressure, and no finger would be generated.
Conclusions The wettability of a PMMA film on a silicon wafer with a native oxide layer is investigated. Upon thermal annealing, the PMMA film remains stable on the substrate because the effective Hamaker constant, Aeff, has a negative value. When the film is exposed to solvent vapor, the wettability of the PMMA film depends on the solvent properties. In the nonsolvent vapor, only some protrusions are generated on the film surface because of the small amount of absorption. When the surface is exposed to good solvent vapor, dewetting
5140
DOI: 10.1021/la8041814
happens. The chains in the upper part of the film gain large mobility in the solution state, whereas the chains at the substrate form strong H bonds with the substrate such that the chains have fewer configurations. A new interface is then established between the anchored chains and the moveable chains. The entropy effect then drives the upper part of the film autophobic dewetting on the new interface. When the surface tension of solvent is small (in acetone vapor, γacetone = 24 mN/m), fingers form behind the rim. Otherwise (in dioxane vapor, γdioxane = 33 mN/m), no fingers are generated. The spacing (λ) between the fingers is proportional to h1.31, and the mean size of the resulting droplets grows linearly with h. Acknowledgment. This research is subsidized by the National Natural Science Foundation of China (20621401, 50573077, and 50773080).
Langmuir 2009, 25(9), 5135–5140
