Langmuir 2005, 21, 11085-11091
11085
Honeycomb Pattern Formation via Polystyrene/ Poly(2-vinylpyridine) Phase Separation Liang Cui 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, Changchun 130022, P. R. China Received April 26, 2005. In Final Form: August 30, 2005 The surface morphologies and properties of polystyrene (PS)/poly(2-vinylpyridine) (PVP) blend films cast on the mica substrate from ethylbenzene solution were investigated upon controlling different weight ratios and solvent evaporation rates. A near-honeycomblike surface morphology of the PS/PVP blend film formed under controlling the solvent evaporation rate due to the effect of Marangoni-Benard convection. The results of static water contact angles, X-ray photoelectron spectroscopy, solvent selective etching, and treatment by water illustrated that the near-honeycomblike structures on the surface of PS/PVP blend films were different for different weight ratios of PS and PVP. After treatment with water for several minutes, PVP islands-like structure emerged in the holes of the film for a PS/PVP weight ratio of 4/1, and a quasihexagonal arrangement of alternate big and small PVP droplets emerged on the top layer of the film for a PS/PVP weight ratio of 7/1. The formation mechanisms of different surface structures and their response behaviors to water were discussed.
Introduction In recent years, there has been an increasing interest in phase separating morphology of immiscible polymerblend thin films that can combine the properties of both the polymers. Phase-separation can be observed after the polymer mixtures dissolved in a common solvent are cast on a substrate and solvent evaporate. Such phaseseparation can lead to various morphologies such as islands, bicontinuous structure, holes when altering the system characters such as composition, molecular weight and architecture, film thickness, solvent, or exterior environment such as substrate, pressure, temperature, and external fields, etc.1-10 The surface structures and properties are very important to applications ranging from electronics (e.g., lithography or microelectronic circuits) and biomedical (e.g., coatings on contact lenses), etc.11-18 Ordered patterns can be obtained through the phase * To whom correspondence should be addressed. Tel.: +86-4315262175. Fax: +86-431-5262126. E-mail:
[email protected]. (1) Budkowski, A.; Bernasik, A.; Cyganik, P.; Raczkowska, J.; Penc, B.; Bergues, B.; Kowalski, K.; Rysz, J.; Janik, J. Macromolecules 2003, 36, 4060. (2) Muller-Buschbaum, P.; Gutmann, J. S.; Stamm, M. Macromolecules 2000, 36, 4886. (3) Affrossman, S.; O’Neill, S. A.; Stamm, M. Macromolecules 1998, 31, 6280. (4) Kikuchi, T.; Kudo, M.; Jing, C.; Tsukada, T.; Hozawa, M. Langmuir 2004, 20, 1234. (5) Harris, M.; Appel, G.; Ade, H. Macromolecules 2003, 36, 3307. (6) Walheim, S.; Boltau, M.; Mlynek, J.; Krausch, G.; Steiner, U. Macromolecules 1997, 30, 4995. (7) Woodcock, S. E.; Chen, C. Y.; Chen, Z. Langmuir 2004, 20, 1928. (8) Gutmann, J. S.; Muller-Buschbaum, P.; Stamm, M. Appl. Phys. A, Mater. 2002, 74, S463. (9) Newby, B. M. Z.; Composto, R. J. Macromolecules 2000, 33, 3274. (10) Karim, A.; Slawecki, T. M.; Kumar, S. K.; Douglas, J. F.; Satija, S. K.; Han, C. C.; Russell, T. P.; Liu, Y.; Overney, R.; Sokolov, O.; Rafailovich, M. H. Macromolecules 1998, 31, 857. (11) Wu, S. Polymer Interface and Adhesion; Marcel Dekker: New York, 1982. (12) Prime, K. L.; Whitesides, G. M. Science 1991, 252, 1164. (13) Rockel, H.; Huber, J.; Gleiter, R.; Schuhmann, W. Adv. Mater. 1994, 6, 568. (14) Licari, J. J.; Hughes, L. A. Handbook of Polymer Coatings for Electronics; Chemistry Technology and Applications; Noys Publications: Park Ridge, NJ, 1990.
separation of polymer-blend thin film controlled by some means, such as using prepatterned substrates.19-22 For thicker polymer blend films, the surface-area-to-volume ratio of polymer blend films is relatively small and the interface of both air/polymer and polymer/substrate play a minor role in determining the morphology of phase separation. Bulklike phase separation will occur that will lead to a morphology close to thermodynamic equilibrium.23 However, for phase separation of thicker polymer blend films controlled by solvent quench, diverse morphologies will form far from thermodynamic equilibrium.24,25 The concentration or temperature gradients might be present within the fluid film containing blend polymers because solvent evaporation occurs on the air/polymer solution surface. As a result, phase separation might occur on the surface when the polymers concentration reaches a critical value. In addition, the surface of the fluid film is cooled due to latent heat of evaporation. BenardMaragoni convection26 emerges due to a temperature gradient in a liquid layer leading to local surface tension variations.27 The process of thermal and materials transfer will drive the distribution of polymers within the fluid system. If the optimal conditions are determined, ordered patterns can be fabricated in the phase separation system. (15) Cowie, J. M. G. Polymers: Chemistry and Physics of Modern Materials; Chapman and Hall: New York, 1991. (16) Netz, R. R.; Andelman, D. Phys. Rev. E 1997, 55, 687. (17) Onda, T.; Shibuichi, S.; Satoh, N. Langmuir 1996, 12, 2125. (18) Service, R. F. Science 1997, 278, 383. (19) Boltau, M.; Walheim, S.; Mlynek, J.; Krausch, G.; Steiner, U. Nature 1998, 391, 877. (20) Krausch, G.; Kramer, E. J.; Rafailovich, M. H.; Sokolov, J. Appl. Phys. Lett. 1994, 64, 2655. (21) Nisato, G.; Ermi, B. D.; Douglas, J. F.; Karim, A. Macromolecules 1999, 32, 2356. (22) Cyganik, P.; Bergues, A.; Budkowski, A.; Bergues, B.; Kowalski, K.; Rysz, J.; Lekki, J.; Lekka, M.; Postawa, Z. Vacuum 2001, 63, 307. (23) Tanaka, K.; Takahara, A.; Kajiyama, T. Macromolecules 1996, 29, 3232. (24) Kumacheva, E.; Li, L.; Winnik, M. A.; Shinozaki, D. M.; Cheng, P. C. Langmuir 1997, 13, 2483. (25) Hopkinson, I.; Myatt, M. Macromolecules 2002, 35, 5153. (26) Cross, M. C.; Hohenberg, P. C. Rev. Mod. Phys. 1993, 65, 851. (27) Normand, C.; Pomeau, Y.; Velarde, M. G. Rev. Mod. Phys. 1977, 49, 581.
10.1021/la0511135 CCC: $30.25 © 2005 American Chemical Society Published on Web 10/05/2005
11086
Langmuir, Vol. 21, No. 24, 2005
Kumacheva et al.28 revealed that the ordered hexagonal islands of poly(methyl methacrylate) (PMMA) based on Marangoni-Benard convection formed on the surface of PMMA/PS/toluene blend solutions for particular ranges of solvent evaporation rate and thickness of fluid films, and then periodic two-phase structures remained on the surface of the polymer films after solidifying. In this report, honeycomblike patterns were observed in PS/PVP blend films when ethylbenzene was chosen as a common solvent for the PS/PVP system under controlling the evaporation speed of the cast solution. For different weight ratios of PS and PVP, the structures of the nearhexagonal arrangement of holes formed on the film surface were different. After treatment with water for several minutes, a PVP islands-like structure emerged in the holes of the film for a PS/PVP weight ratio of 4/1 and quasihexagonal arrangement of alternate big and small PVP droplets emerged on the top layer of the film for a PS/PVP weight ratio of 7/1. The mechanisms for different kinds of microstructures formation were discussed in details. Experimental Section Polystyrene (PS) (the average weight molecular mass (Mw) is 220 000 and the polydispersity (Mw/Mn) is 1.04) and poly(2vinylpyridine) (PVP) (the average weight molecular mass (Mw) is 11 000 and the polydispersity (Mw/Mn) is 1.01) were purchased from Aldrich Chemical Co. Ethylbenzene (99%, anhydrous) from Beijing Chemical, China was used without purification. Solutions (4 wt %) of PS and PVP with different weight ratios of 4/1 and 7/1, respectively, were cast onto the fresh mica (1× 1 cm) substrates with a micro-injector. The polymer solutions were filtered with a 0.22 µm Millipore membrane before casting. The volume of each drop of solution was about 100 µL. The liquid films were 2.0 mm ( 0.1 mm in thickness. One set of the samples was put into a closed vessel with some little holes on the cover. In this case, the solvent evaporation rate can be controlled as low as possible, whereas the solvent evaporation rate of the other set of the samples was controlled at 1.0 l/min by putting the samples into a container plugged with a cork. There were two exhaust pipes on the cork. One opened into the atmosphere, and the other one was connected with a vacuum water pump. The solvent evaporation speed was controlled by a rotameter. Atomic force microscopy (AFM) measurements were performed on SPA300HV with an SPI 3800 controller, Seiko Instruments Industry, Co., Ltd. The topographical images were taken with contact mode at room temperature. The tip type was SN-AF01, made of Si3N4 with a typical spring constant of 2N/m. The scan rate was about 1.0 Hz, and the applied force was the minimal value that would keep contact between the tip and the sample. The XPS (X-ray photoelectron spectroscopy) was measured with VG ESCALAB MK II at room temperature by using an Mg KR X-ray source (hν ) 1253.6 ev) at 14 kV and 20 mA. The sample analysis chamber of the XPS instrument was maintained at a pressure of 1 × 10-7 Pa. To determine the composition on the top layer of the film, the tilting angle is 5°. All of the C1S peaks were calibrated to the standard binding energy shifts. A linearbackground method removed the XPS background and the peaks analysis carried out by using the curve-fitting software. Static water contact angles (CA) were measured to compare the surface characters of films. Contact angles were determined using a Kru¨ss DSA10-MK2 contact angle measuring system at ambient temperature. The probe fluid was deionized water and droplet volumes were 2 µL. For each sample, the contact angle value was read after 5 s, and the average results were obtained on at least 10 droplets.
Results The effects of the PS/PVP weight ratio and solvent evaporation rate on the surface morphology of PS/PVP blends after casting from ethylbenzene solution (4 wt %) (28) Mitov, Z.; Kumacheva, E. Phys. Rew. Lett. 1998, 81, 3427.
Cui and Han
on the mica substrate have been investigated. Two different PS/PVP weight ratios (4/1 and 7/1) and two different solvent evaporation rates have been selected. After the blends films were cast on the mica substrate, one set of the samples was put into a closed vessel with some little holes on the cover. In this case, the solvent evaporation rate could be very low, whereas the solvent evaporation rate of other set of the samples was controlled at 1.0 L/min. Thereafter, the surface morphologies of the samples were imaged by AFM after the solvent evaporated completely (Figure 1). As seen from Figure 1a,b, disordered holes emerge on the surface of the films with both PS/ PVP weight ratios of 4/1 and 7/1 when the solvent evaporation rate is very low. In contrast, drastic change occurs when the solvent evaporation rate is controlled at 1.0 l/min (Figure 1c,d). A film with well-defined ordered holes form, which have a typical depth of 480 and 650 nm, a uniform diameter of 2.5 and 2 µm, and an interval of 3.8 and 3 µm between the adjacent holes corresponding to 4/1 and 7/1 PS/PVP (w/w) blend, respectively. The fast Fourier transform (FFT) patterns in the inset of the images indicate a quasi-hexagonal arrangement of the holes. From the above results, it seems that the solvent evaporation rate has a big effect on the surface morphology. Then, how about the effects of PS/PVP weight ratio and solvent evaporation rate on the surface composition and properties of PS/PVP blends? To investigate the surface property of the films, water contact angle (CA) was performed by the sessile drop method. The water contact angles on the surface of Figure 1a-d, pure PS, and PVP are 93.5 ( 1.5°, 95.5 ( 2.5°, 80.7 ( 1.5°, 59.8 ( 2.5°, 91 ( 1.5°, and 57.7 ( 2.5°, respectively. From comparison, we can see that water contact angle (CA) on the surfaces of Figure 1, panels a and b, are close to the value of PS and that on the surface of Figure 1d is close to that of PVP. The above results show that the surface properties of Figure 1, panels a and b, are hydrophobic and that the surface properties of Figure 1, panels c and d, are hydrophilic. To obtain the surface composition, more accurate XPS data were employed to determine the surface composition. For the sake of comparison, Figure 2, panels a and b, gives the C1s and N1s peaks of pure PVP, respectively. The N1s peak at 399.30 eV is the characteristic peak of PVP. Figure 2c-h presents the C1s and N1s peaks of Figure 1a-d, respectively. From Figure 2d, we can see that there is no N1s peak on the surfaces of Figure 1a. The same results are obtained for Figure 1b (not shown). However, N1s peak emerges on the surfaces of Figure 1, panels c and d. Because the N1s peak only appears in PVP phase, it can be concluded that there is no PVP phase on both surfaces of Figure 1, panles a and b, i.e., PS segments are enriched on the film surface due to its lower surface free energy (γPS ) 39.8 mN/m) compared with that of PVP (γPVP ) 45.1 mN/m) 29 when the solvent evaporation rate is very low. The above results of CA also verify the conclusion. For Figure 1, panels c and d, the peak areas of N1s are different from that of pure PVP. The N/C ratio calculated by computer software contains the contribution of each polymer to the overall N and C spectra, which can be expressed as
XN PVP N ) C XC PVP + (1 - X)CPS
(1)
where X is the molar concentration of PVP phase on the (29) Sauer, B. B.; Dee, G. T. Macromolecules 2002, 35, 7024.
Honeycomb Pattern Formation
Langmuir, Vol. 21, No. 24, 2005 11087
Figure 1. AFM topographical images of PS/PVP cast on a mica substrate from ethylbenzene solution (4 wt %) with weight ratio of 4/1 and 7/1 under the air flow speed (a and b) as slow as possible and (c and d) 1.0 L/min, respectively. The fast Fourier transform (FFT) patterns in the inset of panels b and d indicate a perfect hexagonal arrangement of the holes. The bottom of panels c and d shows their possible structure.
surface of polymer blend film and NPVP, CPVP, and CPS are respectively the nitrogen and carbon atomic concentrations in pure PVP and pure PS (NPVP )1, CPVP )7, and CPS )8). X is defined as
X)
no. of PVP monomers no. of PVP monomers + no. of PVP monomers (2)
X ) 32.7% and 51.5% for the surface of Figure 1, panels c and d, respectively. Neglecting the small density difference between the two polymers, the molar fractions (X) can be transformed into Rx, the area fraction of PVP on the surface, by applying the following formula:
Rx )
X × Mn of vinylpyridine
X × Mn of vinylpyridine + (1 - X) × Mn of styrene ≈X
(3)
Rx ≈ 32.7% and 51.5% corresponding to the area fractions of PVP on the film surfaces of Figure 1, panels c and d, respectively. At the same time, the surface fraction of the
holes domains, Ra, was obtained by AFM image analysis software. Ra ) 30.5% and 45.5% for the surface of Figure 1, panels c and d, respectively. It can be seen that for Figure 1c, the surface area fraction of the PVP phase is near to that of holes domains. However, for Figure 1d, the surface area fraction of the PVP phase is near to that of plateau domains around the holes (1 - Ra ) 54.5%). From the above results, we can obtain the different surface compositions of Figure 1a-d. Only the PS phase covers on the surface of Figure 1, panels a and b. The PVP phase situates in the holes around which are PS for Figure 1c. For Figure 1d, the PVP phase covers the plateau domains around the holes. In addition, the selective solvent etching method was used to determine the phase domain through the film surface. After the samples of Figure 1, panels c and d, were treated with cyclohexane, a selective solvent for PS, two samples show the same behavior; that is, the holes disappeared and an undulated surface remained on the substrate (see Figure 3a for Figure 1d). It showed that the holes were composed of PS phase. When the sample of Figure 1c was dissolved in ethanol, a selective solvent of PVP, the film was destroyed because the PVP phase under
11088
Langmuir, Vol. 21, No. 24, 2005
Cui and Han
Figure 2. XPS spectra: (a) C1s and (b) N1s of pure PVP; (c) C1s and (d) N1s of the sample in Figure 1a; (e) C1s and (f) N1s of the sample in Figure 1c; (g) C1s and (h) N1s of the sample in Figure 1d.
Figure 3. AFM topographical images of the sample in Figure 1d after treatment with (a) cyclohexane and (b) ethanol, respectively.
the PS phase was dissolved by ethanol. It may exhibit that the PVP domains in the holes were adjacent to air and spread on the mica substrate (see the bottom of Figure 1c for the schematic showing). Whereas, when the sample of Figure 1d was dissolved in ethanol, there was no obvious change for the surface morphology (see Figure 3b). It showed that the PS holes were not through the PS phase to connect with the PVP sublayer (see the bottom of Figure 1d for the schematic showing). Moreover, because the PVP was hydrophilic and was strongly hygroscopic and PS was hydrophobic, different structures were demonstrated when a drop of water was kept on the surfaces of Figure 1, panels c and d, for about
1 min and then evaporated. Large islands protruded from the holes consisting of PS phase in Figure 1c (see Figure 4a). The formation of large PVP islands protruding from the holes was due to the volume increase of the PVP phase swollen by water. Different from that, some little droplets formed on the top layer of Figure 1d (see Figure 4b). A quasi-hexagonal arrangement of alternate big and small droplets around the holes was distinctly exhibited (see Figure 4c). The sizes of the big droplets and small droplets were about 850 and 450 nm, respectively. However, large PVP islands in the holes were not observed. The formation of those droplets might originate from the rupture of the
Honeycomb Pattern Formation
Langmuir, Vol. 21, No. 24, 2005 11089
Figure 5. Schematic illustrations for the formation process of honeycomb patterns films.
Figure 4. AFM topographical images of the samples (a) in Figure 1c and (b) in Figure 1d treated by water for about 1 min, respectively. (c) 3D enlarged view of single cell from panel b.
PVP layer on the plateau domains after the film surface was treated with water. From all of the above results, it can be seen that even though the surface morphologies of PS/PVP with different weight ratios are nearly the same, the surface properties and compositions are quite different. Figure 5, panels d and d′, presents the structure schematics. For Figure 1c, PVP forms a sublayer on the mica substrate and the holes consisting of PS are through to the film and in contact with the PVP sublayer (see Figure 5d). For Figure 1d, the situation is quite different. PVP forms a sublayer on the mica substrate, and the holes consisting of PS are not through to the film. Meanwhile a thin layer of PVP covers on the plateau domains of the surface (see Figure 5d′). Discussion The solvent loss from the ternary polymer solution leading to the occurrence of phase separation in thicker polymer blends films has been widely studied.30 The solvent loss, which occurs at the sample-air interface, will usually lead to inhomogeneous change of solvent
concentration within the sample. When the solvent concentration drops to a critical value, the phase separation occurs initially at the sample-air interface. With the solvent concentration decreasing continuously, the phase separation develops through to depth of the sample. Gravitational sedimentation and the inhomogeneous volatility strongly influence the phase separation behavior. Various transient structures of phase domains with depth into the sample are solidified after the solvent evaporates completely. On the other hand, when a wetting surface which attracts either phase exists, the wettability causes the interaction between the phases and the solid surface, forming a lamellar phase morphology. PS and PVP were strong immiscible blend systems (χ ≈ 0.1).31 Phase separation would occur when the polymers concentration reached a critical value with the solvent evaporation. Ethylbenzene was a better solvent for PS with a higher solubility than that of PVP.32 Due to a higher solubility of the PS in the ethylbenzene, PVP was more quickly depleted from the solvent and solidified first onto the substrate. In addition, the preferential adsorption of the PVP phase on the mica substrate surface resulted in the deposition of the PVP-rich phase layer on the substrate. After the deposition of the PVP-rich phase layer on the substrate, PS tended to stay longer in the liquid phase and formed a second layer at the surface due to its higher solubility in the ethylbenzene and lower surface free energy compared with that of PVP. A similar lamellar structure formed on the initial step. The upper layer of PS underwent a spinodal dewetting-induced rupture,33 which led to a distribution of disordered holes on the surfaces of the blend films. The case changed when the solvent evaporation speed increased. The thermal and materials transfer will increase strongly in the fluid system. Benard-Maragoni convection28 occurred due to the formation of temperature (30) (a) Hopkinson, I.; Myatt, M. Macromolecus 2002, 35, 5135. (b) Matthew, M.; Rharbi, Y.; Li, H.; Winnik, M. A. Macromolecules 2002, 35, 3321. (c) Kumacheva, E.;, Li, L.; Winnik, M. A.; Shinozaki, D. M.; Cheng, P. C. Langmuir 1997, 13, 2483. (d) Saraf, R. F.; Ostrander, S.; Feenstra, R. M. Langmuir 1998, 14, 483. (31) Walheim, S.; Ramstein, M.; Steiner, U. Langmuir 1999, 15, 4848. (32) The solubility parameters δ [(MPa)1/2] of PS, P2VP, and ethylbenzene are 17.6, 20.6, and 17.8, respectively. (33) Higgins, A. M.; Jones, R. A. L. Nature 2000, 404, 476.
11090
Langmuir, Vol. 21, No. 24, 2005
Cui and Han
gradient ∆T between the below and top layer of the fluid with the solvent evaporation,27,34 which was very similar to those cases of fluid thin films under heat.35,36 Be´nardMarangoni convection would present a source of ordered patterns. In Be´nard-Marangoni convection, the pattern formation is governed by a balance of the surface-tensiondriven forces and dissipation due to the thermal diffusion and the frictional action of viscosity. The competition of these factors can be expressed by a Marangoni number28
Ma ) [-(dγ/dT) ∆Th]/Fνκ
(4)
where h is the thickness of the fluid layer and γ, F, ν, and κ are the fluid surface tension, density, kinematics viscosity, and thermal diffusivity, respectively. ν ) η/F. η is the dynamic viscosity and ∆T is the difference in temperature between the bottom and the top surfaces of the film. The transition to the ordered patterns occurs at Ma > Mac (critical value). Linear theory 38 yields the onset of convection at Mac ≈ 80. Since the concentration of polymers is very small, the various value were assumed to be close to those of the pure solvent. The dynamic viscosity η ) 0.000 637 N s m-2 and the surface tension temperature gradient -(dγ/dT) ) 0.13 mN/m K.37 The value of thermal diffusivity, κ, can be calculated using κ ) R/FCp,28,39 where Cp is the heat capacity at constant pressure (Cp ) 183.60 J K-1 mol-1),37 R is the thermal conductivity coefficient (R ) 0.13 J/m s K),37 and F is the density (F ) 862.53 Kg/m3).37 κ ) 8.3 × 10-8 m2 s-1. The temperature gradient ∆T can be evaluated using the equation:39 ∂T/∂h ) j∆Hv/R, where j is the mass flow, ∆Hv is the latent heat (∆Hv ) 42.25 kJ mol-1),37 and R is the thermal conductivity coefficient, respectively. The mass flow, j, is measured using the equation j ) n/(St), where n (n ) 7.2 × 10-4 mol) is the amount of mater which takes a time (t) (t ) 180 s) to evaporate, deposited on a surface S (S ) 1.0 × 10-4 m2). j ) 0.04. Assuming the liquid layer height h ) 2.0 × 10-3 m, the ∆T can be obtained (∆T ≈ 26 °C). Using eq 4, the calculated value of Ma is about 1.3 × 105. In Figure 1, panels a and b, there was nearly no presence of temperature gradient ∆T between the below and top layer of the fluid due to too low of solvent evaporation rate. Therefore, there was no presence of ordered patterns. In contrast to Figure 1, panels a and b, in Figure 1, panels c and d, the solvent evaporation rate was very large leading to the temperature gradient ∆T between the below and top layer of the fluid exceeded the critical value ∆Tc,40 i.e., Marangoni number (Ma) exceeds the critical value (Mac). Therefore, ordered patterns emerged. Why are the surface compositions and properties of panels c and d of Figure 1 very different? We proposed that it was determined by the transfer process of materials (34) Sakurai, S.; Furukawa, C.; Okutsu, A.; Miyoshi, A.; Nomura, S. Polymer 2002, 43, 3359. (35) Li, M.; Xu, S. Q.; Kumacheva, E. Macromolecules 2000, 33, 4972. (36) Xu, S. Q.; Kumacheva, E. J. Am. Chem. Soc. 2002, 124, 1142. (37) Dean, J. A. Lange’s Handbook of Chemistry; McGraw-Hill: New York, 1992. Yaws, C. L. Chemical Properties Handbook; McGraw-Hill: New York, 1992. Wu, R. J. Polymer’s surface and interface; Science Press: Beijing, 1998. (38) Pearson, J. R. A. J. Fluid Mech. 1958, 4, 489. (39) Maillard, M.; Motte, L.; Ngo, A. T.; Pileni, M. P. J. Phys. Chem. B 2000, 104, 11871. (40) The critical temperature gradient across the film can be calculated using the following formula: ∆Tc ) Mac Fνκ/[-(dγ/dT)h]. Linear theory yields the onset of convection at Mac ≈ 80. Fν ) η ) 0.000 637 N s m-2, -(dγ/dT) ) 0.13 mN/m K. The thickness of the fluid layer is given by (characteristic instability wavelength) λ ) 2 h. λ is close to the interval between the adjacent holes. λ ) 3.0∼3.8 µm. The value of thermal diffusivity κ is 8.3 × 10-8 m2 s-1. Then, the critical temperature gradient can be calculated to be about 16∼20 °C.
generated by convection and the amount difference of PS and PVP in ethylbenzene. Figure 5 demonstrated the formation process of different structures in the blend films during solvent evaporation. In Figure 5a, the convection drove the upflow and downflow motion of PS and PVP in the ethylbenzene solution during solvent evaporation. The upflow and downflow of materials driven by convection produced the hot and cool spots on the surface of fluid. The inhomogeneity of temperature led to the inhomogeneity of surface tension and concentrations. The hot spots corresponded to the low surface tension, and the cool spots corresponded to the high surface tension. In Figure 5, panels b and b′, with the solvent evaporated, the polymers concentration increased leading to the occurrence of phase separation. PVP would deplete earlier than PS due to its lower solubility in ethylbenzene than that of PS.32 Moreover, due to the strong interaction between the PVP phase and the mica substrate, the PVP phase tended to deposit on the mica substrate. In Figure 5b′, for Figure 1d, the amount of PVP component was relative small. After some PVP phase deposited on the mica substrate, some minor PVP component in the solution would stay longer. In Figure 5b, for Figure 1c, the amount of PVP component was relatively more in contrast to Figure 1d. With the solvent evaporation, the composition of PVP increased faster compared with that in Figure 1d. Subsequently, they depleted the solvent and deposited on the substrate very fast. There was no superfluous other PVP in the fluid. Only upflow and downflow of the PSrich phase was driven by convention. The polymers will be transported from the spots of the low surface tension to that of high surface tension. In Figure 5c′, for Figure 1d, PS and minor PVP were driven to the domains of high surface tension by the convection. PVP would deplete earlier than PS due to its lower solubility in ethylbenzene than that of PS.41 Then, a thin layer of PVP covered the domains of high surface tension. In Figure 5c, for Figure 1c, only PS phase was driven to the domains of high surface tension by the convection. In the end, the domains of high surface tension formed the higher concentration of polymers leading to the production of the plateau domains.34 The domains of low surface tension corresponded to the holey domains (Figure 5d). For Figure 1c, ordered holes consisted of the pure PS phase covered on the PVP sublayer. On the other hand, the PS amount decreased in comparison with Figure 1d. It could not spread over the entire PVP layer on the substrate and thus segregated into ribbons. It created the structure of pits through to the PS layer. However, (Figure 5d′) for Figure 1d, the amount of PS component was enough to spread over the entire PVP layer on the mica substrate leading to the holes could not contact with the PVP sublayer. Meanwhile, a thin layer of PVP covered the plateau domains of PS. The different water response behaviors of the samples in Figure 1, panels c and d, are due to the different surface structures and compositions. For the sample in Figure 1c, the PVP formed a layer on the mica substrate and the holes consisting of PS were through to the film and in contact with the underlying PVP layer, and large islands of PVP protruded from the holes consisted of the PS phase after swollen by water (Figure 4a). Whereas for the sample of Figure 1d, the cuticle of PVP on the plateau of PS were swollen by water and ruptured to droplets. The process of rupture could be recognized as a dewetting behavior of PVP on PS substrate. The occurrence of dewetting was too soon to observe the transitional conditions. When the (41) Walheim, S.; Bo¨ltau, M.; Mlynek, J.; Krausch, G.; Steiner, U. Macromolecules 1997, 30, 4995.
Honeycomb Pattern Formation
dewetting occurred on the surface of the film with ordered holes, these big and little droplets regularly distributed on the domains around the holes.42 Conclusions We presented the effects of solvent evaporation rates and PS/PVP weight ratios on the surface morphologies, compositions, and properties after PS/PVP ethylbenzene solutions were cast on the mica substrates. At a very low solvent evaporation rate, disordered holes distributed on the surfaces of the blend films were observed since the PS-rich phase layer covered on the top of the PVP due to lower surface free energy compared with that of PVP. When the solvent evaporation rate was controlled, a quasihexagonal arrangement of holes formed on the surfaces of PS/PVP blend films due to the effect of BenardMaragoni convection upon the solvent evaporation. However, the results of the water contact angle, XPS, solvent selective etching, and treatment by water illustrated that the structures of the hexagonal arrangement of holes formed on the surfaces of the PS/PVP blend films was different for different weight ratios. For sample of PS/ (42) Luo, C. X.; Xing, B.; Zhang, Z. X.; Fu, J.; Han, Y. C. J. Colloid Interface Sci. 2004, 269, 158.
Langmuir, Vol. 21, No. 24, 2005 11091
PVP (4/1), PVP formed a layer on the mica substrate and the holes consisting of PS were through to the film and in contact with the PVP sublayer. Whereas for sample of PS/PVP (7/1), PVP formed a sublayer on the mica substrate and the holes consisting of PS were not through to the film. Meanwhile a thin layer of PVP covered the plateau domains of the surface. The two different surface structures had different response behaviors to water. Islandlike patterns emerged in the holes after the sample of PS/PVP (4/1) was swollen by water. For the sample of PS/PVP (7/1), after treatment with water, hexagonal arrangement of alternate big and small droplets around the holes was found due to dewetting of PVP on the surface of PS-rich phase plateau domains. Acknowledgment. This work is subsidized by the National Natural Science Foundation of China (50125311, 20334010, 20274050, 50390090, 50373041, 20490220, 20474065, 50403007), the Ministry of Science and Technology of China (2003CB615601), the Chinese Academy of Sciences (Distinguished Talents Program, KJCX2-SWH07), and the Jilin Distinguished Young Scholars Program (20010101). LA0511135
