2326
Langmuir 2007, 23, 2326-2329
Dewetting of Thin Polystyrene Films under Confinement Juan Peng,†,‡ Rubo Xing,† Yang Wu,† Binyao Li,† Yanchun Han,*,† Wolfgang Knoll,‡ and Dong Ha Kim*,‡,§ State Key Laboratory of Polymer Physics and Chemistry, Changchun Institute of Applied Chemistry, Chinese Academy of Sciences, Graduate School of Chinese Academy of Sciences, Changchun 130022, P. R. China, Max Planck Institute for Polymer Research, Ackermannweg 10, 55128 Mainz, Germany, and DiVision of Nano Sciences and Department of Chemistry, Ewha Womans UniVersity, 11-1 Daehyun-Dong, Seodaemun-Gu, Seoul 120-750, Korea ReceiVed July 3, 2006. In Final Form: January 3, 2007 The dewetting behavior of thin polystyrene (PS) film has been investigated by placing an upper plate with a ca. 140 nm gap from the underlying substrate with the spin-coated thin polymer films. Three different kinds of dewetting behaviors of thin PS film have been observed after annealing according to the relative position of the PS film to the upper plate. Since the upper plate is smaller than the underlying substrate, a part of the polymer film is not covered by the plate. In this region (I), thin PS film dewetting occurs in a conventional manner, as previously reported. While in the region covered by the upper plate (III), the PS film exhibits unusual dewetted patterns. Meanwhile, in the area right under the edge of the plate (II) (i.e., the area between region I and region III), highly ordered arrays of PS droplets are formed. Formation mechanisms of different dewetted patterns are discussed in detail. This study may offer an effective way to improve the understanding of various dewetting behaviors and facilitate the ongoing exploration of utilizing dewetting as a patterning technique.
1. Introduction Thin polymer films are used in many applications such as photoresists, lubricant layers, interlayer dielectrics in the microelectronics industry, and alignment layers in liquid crystal displays.1 For these applications, smooth and stable films are required. With spin coating, polymer films can be coated even onto nonwetting substrates to produce uniform films. However, such films are not stable and will dewet to minimize the interfacial energy when heated above their glass transition temperatures. Therefore, the wetting and dewetting behaviors of thin polymer films have received much recent attention both experimentally and theoretically.2-7 On one hand, the mechanisms and dynamics investigations of the dewetting process itself have been studied extensively. Reiter3 characterized the progression of dewetting from early to complete dewetting using polystyrene (PS) films heated above the glass transition temperature. Meanwhile, different mechanisms such as spinodal dewetting,8-10 nucleation and growth dewetting,4,11 and thermal nucleation dewetting12,13 * Corresponding author. Tel.: +86-431-85262175 (Y.C.H); +82-2-32774517 (D.H.K.). Fax: +86-431-85262126 (Y.C.H); +82-2-3277-3419 (D.H.K.). E-mail:
[email protected] (Y.C.H);
[email protected] (D.H.K). † Graduate School of Chinese Academy of Sciences. ‡ Max Planck Institute for Polymer Research. § Ewha Womans University. (1) Frank, C. W.; Rao, V.; Despotopoulou, M. M.; Pease, R. F. W.; Hinsberg, W. D.; Miller, R. D.; Rabolt, J. F. Science 1996, 273, 912. (2) Brochard-Wyart, F.; Martin, P.; Redon, C. Langmuir 1993, 9, 3682. (3) Reiter, G. Phys. ReV. Lett. 1992, 68, 75. (4) Reiter, G. Langmuir 1993, 9, 1344. (5) Brochard-Wyart, F.; Debregeas, G.; Martin, P. Macromolecules 1997, 30, 1211. (6) Zhao, W.; Rafailovich, M. H.; Sokolov, J. Phys. ReV. Lett. 1993, 70, 1453. (7) Qu, S.; Darke, C. J.; Liu, Y.; Rafailovich, M. H.; Sokolov, J.; Phelan, K. C.; Krausch, G. Macromolecules 1997, 30, 3640. (8) Cahn, J. W. J. Chem. Phys. 1965, 42, 93. (9) Xie, R.; Karim, A.; Douglas, J. F.; Han, C. C.; Weiss, R. A. Phys. ReV. Lett. 1998, 81, 1251. (10) Brochard-Wyart, F.; Daillant, J. Can. J. Phys. 1990, 68, 1084. (11) Meredith, J. C.; Smith, A. P.; Karim, A.; Amis, E. J. Macromolecules 2000, 33, 9747. (12) Seemann, R.; Herminghaus, S.; Jacobs, K. Phys. ReV. Lett. 2001, 86, 5534. (13) Seemann, R.; Herminghaus, S.; Jacobs, K. J. Phys.: Condens. Matter 2001, 13, 4925.
have been proposed to explain the dewetting process. Concerning the dewetting velocity, it has also been reported that the hole diameter increases linearly with time or as t2/3 when slippage between the melt and the substrate is significant.2,14 On the other hand, more and more efforts have been made to stabilize polymer thin films. Polymer thin film stability can be improved by increasing the interaction between the polymer and the substrate or slowing down the dewetting kinetic process. The former includes introducing metal counterions into sulfonated PS film,15 using end-functional homopolymers,16,17 block copolymer additives,18 and so forth, and the latter involves using high molecular weight or glassy polymer films. Recent work has shown that the addition of nanoparticles such as C60 fullerene nanoparticles19 and carbon black20 into polymer films can also inhibit the film dewetting. In addition, the utilization of dewetting as a patterning or lithographic technique is researched gradually. Ordered patterns can be obtained by selective dewetting,21 dewetting on physically and chemically patterned substrates,22,23 dewetting under structured confining surfaces,24,25 and so forth. Electric fields26-29 (14) Limary, R.; Green, P. F. Langmuir 1999, 15, 5617. (15) Feng, Y.; Karim, A.; Weiss, R. A.; Douglas, J. F.; Han, C. C. Macromolecules 1998, 31, 484. (16) Henn, G.; Bucknall, D. G.; Stamm, M.; Vanhoorne, P.; Jerome, R. Macromolecules 1996, 29, 4305. (17) Yerushalmi-Rozen, R.; Klein, J. Langmuir 1995, 11, 2806. (18) Oslanec, R.; Costa, A. C.; Composto, R. J.; Vlcek, P. Macromolecules 2000, 33, 5505. (19) Barnes, K. A.; Karim, A.; Douglas, J. F.; Nakatani, A. I.; Gruell, H.; Amis, E. J. Macromolecules 2000, 33, 4177. (20) Sharma, S.; Rafailovich, M. H.; Peiffer, D.; Sokolov, J. Nano Lett. 2001, 1, 511. (21) Kim, Y. S.; Lee, H. H. AdV. Mater. 2003, 15, 332. (22) Kargupta, K.; Sharma, A. Langmuir 2002, 18, 1893. (23) Sehgal, A.; Ferreiro, V.; Douglas, J. F.; Amis, E. J.; Karim, A. Langmuir 2002, 18, 7041. (24) Harkema, S.; Scha¨ffer, E.; Morariu, M. D.; Steiner, U. Langmuir 2003, 19, 9714. (25) Suh, K. Y.; Park, J.; Lee, H. H. J. Chem. Phys. 2002, 116, 7714. (26) Scha¨ffer, E.; Thurn-Albrecht, T.; Russell, T. P.; Steiner, U. Nature 2000, 403, 874. (27) Scha¨ffer, E.; Thurn-Albrecht, T.; Russell, T. P.; Steiner, U. Europhys. Lett. 2001, 53, 518. (28) Chou, S. Y.; Zhuang, L. J. Vac. Sci. Technol., B 1999, 17, 3197. (29) Deshpande, P.; Sun, X.; Chou, S. Y. Appl. Phys. Lett. 2001, 79, 1688.
10.1021/la061911a CCC: $37.00 © 2007 American Chemical Society Published on Web 02/02/2007
Letters
Langmuir, Vol. 23, No. 5, 2007 2327
and temperature gradient30,31 have also been shown to be effective means to produce highly ordered lateral structures. In the present paper, we investigate the dewetting behavior of confined thin PS film by placing an upper smooth plate at a very close distance from it. In contrast to Chou’s earlier experimental device,28,29 the upper plate is smaller than the underlying polymer/ silicon substrate; therefore, a part of the polymer film is not covered with the plate. After annealing, the PS films dewet via heterogeneous nucleation in these regions and show the ordinary dewetting pattern, whereas the PS films dewetting under the plate exhibit unusual patterns. In addition, rather regular arrays of polymer droplets are found in polymer films right under the edge of the upper plate. It is assumed that the electrostatic force and heterogeneous nucleation compete with each other and characterize the two regions. It is believed that this strategy is the first to use the various relative positions of PS film to the upper plate to simultaneously probe the different dewetting behaviors of thin polymer films dominated by various driving forces, which can improve our understanding of the influencing factors on the dewetting mechanisms. 2. Experimental Section PS (Mw ) 31 600, Mw/Mn ) 1.08; Aldrich) toluene solutions (1 wt %) were filtered with a 0.22 µm Millipore membrane and spincoated on a clean silicon wafer substrate (cleaned with a 70/30 volume ratio solution of 98% H2SO4/30% H2O2 at 80 °C for 30 min and rinsed in deionized water and dried). The silicon wafers used were doped with phosphorus with a layer of SiOx (∼2 nm thick from ellipsometry). All the prepared films were baked at 50 °C in a vacuum oven for 12 h to drive off the solvent. The thickness of the PS films was about 40 nm measured by ellipsometry (AUEL-III, Xi’an, China). In the experimental process, a PS film was placed on a temperaturecontrolled hot plate. Facing it, a second silicon wafer was mounted a certain distance above the film separated by the spacers, which consisted of an array of 140 nm high aluminum pillars (2.2 mm diameter separated by 2.8 mm) deposited on the top wafer by electron beam evaporation through a shadow mask. Typically, the sizes of the upper and underlying silicon wafers were ∼1 × 1 cm2 and 2 × 2 cm2, respectively. A pressure of about 1 MPa was applied by pressing another hot plate on the back of the second silicon wafer to hold the gap constant. The height of the spacer ensured the upper plate did not contact the film when pressurized.32 The system was then heated at 165 °C for times ranging from 10 min to 14.5 h. The heat was from both the top and bottom plates, with a maximal temperature difference ∆T of 4 °C. After the sample was cooled, the system was disassembled, and the morphology of the PS film was observed by optical microscopy (OM; XJX-2, Nanjing, China) and atomic force microscopy (AFM; SPA300HV with a SPI3800N probe station, Seiko Instruments, Inc., Japan) in contact mode.
3. Results and Discussion OM and AFM indicated that the PS films were smooth and uniform when inspected immediately after spin coating. After annealing, different phenomena were observed on the same sample with different locations in Figure 1. In region I, holes appeared after annealing at 165 °C for 30 min (Figure 2a). With further annealing for 2 h, the classic late-stage dewetting of a droplet pattern was observed in this region as reported (Figure 2b). (30) Scha¨ffer, E.; Harkema, S.; Roerdink, M.; Blossey, R.; Steiner, U. Macromolecules 2003, 36, 1645. (31) Peng, J.; Wang, H.; Li, B.; Han, Y. Polymer 2004, 45, 8013. (32) It was noted that the spacers were pressed into the polymer melt during heating. In these regions where the aluminum pillars were placed, the PS film replicated these pillars’ morphology similar to nanoimprinting. The spacers did not influence the dewetting patterns of PS films in other regions. All the pictures were taken in regions outside the zones in which the aluminum pillars were placed.
Figure 1. Schematic illustration of the experiment setup. A liquid PS film was confined between two silicon plates (the upper plate is smaller than the underlying substrate). The distance between the two plates is controlled by the height of the evaporated Al pillar spacers.
In contrast to the ordinary dewetting of PS films in region I, PS films in region III exhibited slow and unusual dewetted patterns. The film remained stable after annealing at 165 °C for 30 min (Figure 2c). After further annealing for 2 h, ring-shaped patterns and droplets appeared, a little different from the classic pattern (Figure 2d). Prolonged annealing led to an increase in droplets and a decrease in rings (Figure 2e). Such different behavior indicated that the upper plate might change the usual dewetting path. In addition to the dewetting phenomena, an intriguing feature was also observed in the PS films under the edges of the plate (region II) when annealing for 30 min. Highly ordered PS droplets could be seen clearly under AFM (Figure 2f), which had a diameter of about 2.5 µm and a height of 175 nm. The width of region II with ordered PS droplets was about 90 µm measured by AFM. The flat surface at the bottom of the AFM image shows that the film was stable in region III at this time. The fast Fourier transform (FFT) pattern in the inset of the image indicates a hexagonal arrangement. Upon further annealing for 14.5 h, similar ordered PS droplets were found in region II (Figure 2g). A PS film with three different kinds of dewetting behaviors and the boundary of each region are shown in Figure 3. After having shown that different features appeared on the same PS films, several questions arise. What role does the upper plate play during annealing? Why do the PS films form highly regular PS droplet arrays in region II, while in region III they do not? And why are the dewetting fashions of PS films so different in regions I and III? Since the PS films are not extremely thin, it is assumed that dewetting will only proceed by nucleation, which is characterized by the formation of holes, their growth, and coalescence, finally leading to a set of droplets on the substrate.33 PS films of region I are not under the plate and dewet in a conventional manner. Figure 2a,b is taken as evidence for heterogeneous nucleation being the process responsible for their formation. For the PS films in regions II and III, the presence of the upper plate changes their usual dewetting behavior. We assume that the electrostatic force and heterogeneous nucleation compete with each other and characterize the two regions. For comparison, another PS film was heated under the same conditions while short-circuiting the top and bottom plates. In contrast, such ordered droplets in region II were not observed. In this case, the PS film in the regions II and III exhibited unusual ring-shaped and droplet dewetted patterns, similar to that shown in Figure 2d,e. This implies that the upper and lower plates were not on an electrically common ground. In other words, it means that the two hot plates that were used to heat the upper and lower plates were not on a common electrical ground. The “floating” electric ground potentials of the two plates were therefore independent of each (33) Bischof, J.; Scherer, D.; Herminghaus, S.; Leiderer, P. Phys. ReV. Lett. 1996, 77, 1536.
2328 Langmuir, Vol. 23, No. 5, 2007
Letters
Figure 2. AFM image showing different structures in 40 nm PS films after annealing at 165 °C for different times. (a) The appearance of holes and the development of rims, shown in region I after 30 min of annealing. (b) The classic late-stage dewetting of the droplet pattern in region I after 2 h of annealing. (c) The flat film in region III after 30 min of annealing. (d) Ring-shaped pattern and droplets appeared in region III after 2 h of annealing. (e) The film in region III after 14.5 h of annealing. (f) Ordered arrays of droplets formed in the transition region II after 30 min of annealing. The flat surface at the bottom of the AFM image shows that the film was stable in region III at this time. An FFT pattern is given in the inset of the image, indicating a hexagonal arrangement. (g) The film in region II after 14.5 h of annealing.
Figure 3. (a) OM image of a PS film showing three different kinds of dewetting behaviors after annealing at 165 °C for 1 h. It exhibits randomly distributed droplets in region I (left region), an ordered array of droplets in region II (middle region), and holes surrounded by rims in region III (right region). (b) OM image of the boundary phase between regions I and II. (c) OM image of the boundary phase between regions II and III.
other. In such a case, an electric potential difference was built up between the two hot plates (without applying external electric fields) and therefore between the two capacitor plates. Although the exact value of the potential being internally built was not known, compared with the small gap between the top plate and
the polymer, the potential between dissimilar materials can produce a relatively strong electric field. Similarly, although the temperature difference in the experimental system is small (∆T < 4 °C), compared with the small gap between the top and bottom plates, the temperature gradient between the plates still
Letters
Langmuir, Vol. 23, No. 5, 2007 2329
exists. As a consequence, both the electric field and the temperature gradient have an effect, and the former is the dominant driving force. We first discuss the temperature gradient effect. The heat flow across the polymer-air bilayer is essentially by heat diffusion and is given by Fourier’s equation:
Jq ) -k
kakp∆T ∂T ) ∂z kahp + kpha
(1)
where z is the normal coordinate, k and h are the thermal conductivity and layer thickness, respectively, and the indices “p” and “a” refer to the polymer and air, respectively. The heat flow is in the direction of lower temperature and can induce the radiation pressure to destabilize the polymer-air interface. For the film in region I, it is assumed that the air layer thickness ha is quite large, resulting in a quite small heat flow Jq that can be neglected. For the film in regions II and III, the heat flow is the same. Compared with the electric field, the temperature gradient effect is small. It is known that electrostatic potentials are enhanced at edges;28 therefore, the electric field in region II is stronger than that in region III. After the polymer is heated above its Tg, it begins to flow as a viscous liquid. In region II, electrostatically driven instability works before heterogeneous nucleation leads to the formation of holes. The polymer experiences surface tension, which acts as the stabilizing factor, and the electric field and temperature gradient, which are the destabilizing influences. In thin film geometry, gravity is usually negligible. The total pressure acting at the interface can be written as27
∂2h + pel(h) + pT + pdis(h) ∂x2
p ) p0 - γ
(2)
where p0 is the atmospheric pressure, γ is the surface tension, and h is the film thickness. The second, third, fourth, and fifth terms represent the Laplace pressure, electrostatic pressure, radiation pressure, and disjoining pressure in the film, respectively. The Laplace pressure stems from the surface tension γ, the radiation pressure is exerted by the heat flow, and the disjoining pressure arises from dispersive van der Waals interactions. Compared with electrostatic and Laplace pressure, the disjoining pressure is negligible. Both the electrostatic pressure and radiation pressure can destabilize the polymer-air interface once they exceed a critical value, resulting in a Poiseuille-type flow.34 The fluctuation can either be suppressed or further amplified depending on the configuration energetically favored. The free energy stored in the capacitor device with a constant voltage boundary condition is
1 F ) F0 - CU2 2
(3)
where F0, C, and U are the free energy of dielectric in the absence of a field, the capacitance, and the applied constant voltage, respectively. In this case, a raise in capacity by rearranging the polymer from a layered conformation to a columnar geometry (34) Vrij, A. Discuss. Faraday Soc. 1966, 42, 23.
minimizes the free energy F. Therefore, the electrostatic force drives the polymer melt to flow from depressions to peaks, amplifying the fluctuation. The amplified wavelength λ is the additive consequence of electrostatic pressure and radiation pressure. Such fluctuation further grows up at the wave crest and descends at the trough due to mass conservation. Once the trough contacts the silicon substrate at a certain time, the films rupture and form PS droplets at last. The well-defined droplet diameter shows the instability, which leads to the droplet formation occurring at a well-defined time. The ordered droplet distribution indicates the presence of interdroplet interactions. Similar periodic polymer pillar arrays were also observed by Chou et al.28,29 and Scha¨ffer et al.26,27 Compared with their results, the ordered droplets in region II embody the result of competition between electrostatically driven instability and heterogeneous nucleation. While in region III, the electrostatic pressure is too low to overcome surface tension and generate PS columns. Then heterogeneous nucleation dominates the behavior of PS films. However, compared with the PS film in region I, the presence of the upper plate changes the usual dewetting path. When dewetting occurs, holes are nucleated at imperfections in the film such as dust particles accompanied by the formation of rims surrounding the holes. For this early stage of dewetting, the rim heights are much smaller than the plate spacing. With time, the diameters of the isolated holes increase, driven by the imbalance of the interfacial and surface energies. At the same time, the rims are thickening. Before contacting each other, the elevated rims touch the upper plate. Then the viscous drag on the rim doubles and pins the flow. Instead of dragging along this material, the systems cut off this portion, resulting in the formation of the rings.35 The accumulation in the rim then starts from the beginning until it touches the upper plate again. Depending on the strength of the thermomechanical pressure, the rim itself can become unstable and results in the formation of droplets.
4. Conclusions We have shown that different dewetting phenomena can be observed when a PS film is annealed under a plate smaller than the substrate. In the region without the upper plate (I), PS films dewet via heterogeneous nucleation and exhibit classic dewetted patterns. For PS films under the plate (region III), the dewetting is influenced by the upper plate, which pins the rims formed during dewetting, and thus ring-shaped patterns form in the final dewetting. While in the region under the edges of the plate (II), the electrostatically driven instability preempts heterogeneous nucleation, resulting in the formation of regular arrays of PS droplets. We expect that this self-organizational process can be utilized to form regular structures from nanometer to micrometer scale. Acknowledgment. This work was supported by the National Natural Science Foundation of China (20621401, 20474065, 50573077), the Science Fund for Creative Research Groups of China (20621401), and the Seoul Research and Business Development Program (10816). LA061911A (35) Scha¨ffer, E. Instabilities in Thin Polymer Films: Structure Formation and Pattern Transfer. Ph.D. Thesis, University of Groningen, Groningen, The Netherlands, 2001 (http://www.chem.rug.nl/steiner/publication).
