8642
Langmuir 2006, 22, 8642-8645
Tunable Instability Mechanisms of Polymer Thin Films by Molecular Self-Assembly Bin Wei,† Peter G. Lam,† Michael B. Braunfeld,§ David A. Agard,§ Jan Genzer,† and Richard J. Spontak*,†,‡ Departments of Chemical & Biomolecular Engineering and Materials Science & Engineering, North Carolina State UniVersity, Raleigh, North Carolina 27695, and Departments of Biochemistry & Biophysics and the Howard Hughes Medical Institute, UniVersity of California, San Francisco, California 94143 ReceiVed May 17, 2006. In Final Form: August 15, 2006 Incorporation of a block copolymer into a thin polymer film is observed to alter both the rate and mechanism by which the film dewets from an immiscible polymer substrate. Films with little or no copolymer dewet by classical nucleation and growth of circular holes, and the dewetting rate decreases with increasing copolymer concentration. Increasing the copolymer content at constant film thickness generates copolymer micelles that adsorb/aggregate along the polymer/polymer interface and promote nonclassical dewetting fluctuations similar in appearance to spinodal dewetting. At higher copolymer concentrations, dewetting proceeds after a lengthy induction period by the nucleation and growth of flower-shaped holes suggestive of film pinning or viscous fingering. Atomic force microscopy of the polymer/polymer interface after removal of the top film by selective dissolution reveals substantial structural development due to copolymer self-assembly.
The stability of polymer thin films is critical to the development of advanced protective coatings and defect-free multilayer assemblies, in which case the molecular-scale processes responsible for film destabilization must be fully understood. When a thin liquid film contacts a solid substrate, it can remain stable or, depending on the nature and magnitude of intermolecular forces within the film, rupture and dewet from the substrate.1 Since film stability governs the function of synthetic coatings and biological liquids,2 numerous efforts have sought to elucidate the dewetting behavior of homopolymer3-5 or block copolymer6-10 thin films on a solid substrate. Dewetting in thin films generally occurs by nucleation and growth (NG) wherein film rupture proceeds by the formation and growth of circular holes that ultimately impinge to produce sessile droplets.4 Dewetting between immiscible polymeric liquids in multilayer coatings also depends on the viscosities of the dewetting (ηA) and substrate (ηB) polymers.11 When ηB > ηA/θE (where θE is the contact angle at the three-phase contact line), the substrate behaves solidlike relative to the dewetting film. Without interfacial slip, the hole diameter (D) grows linearly with time (t) so that the hole growth velocity (dD/dt) is 2S/Kv, where S is the dewetting force * To whom correspondence should be addressed. E-mail: Rich_Spontak@ ncsu.edu. † Department of Chemical & Biomolecular Engineering, North Carolina State University. ‡ Department of Materials Science & Engineering, North Carolina State University. § University of California. (1) Geoghegan, M.; Krausch, G. Prog. Polym. Sci. 2003, 28, 261. (2) Thiele, U. Eur. Phys. J. E 2003, 12, 409. (3) Redon, C.; Brochard-Wyart, F.; Rondelez, F. Phys. ReV. Lett. 1991, 66, 715. (4) Reiter, G. Phys. ReV. Lett. 1992, 68, 75; Langmuir 1993, 9, 1344. (5) Jacobs, K.; Seemann, R.; Schatz, G.; Herminghaus, S. Langmuir 1998, 14, 4961. (6) Russell, T. P.; Coulon, G.; Deline, V. R.; Miller, D. C. Macromolecules 1989, 22, 4600. (7) Limary, R.; Green, P. F. Macromolecules 1999, 32, 8167. (8) Smith, A. P.; Douglas, J. F.; Meredith, J. C.; Amis, E. J.; Karim, A. Phys. ReV. Lett. 2001, 87, Art. No. 15503. (9) Mu¨ller-Buschbaum, P.; Gutmann, J. S.; Lorenz-Haas, C.; Wunnicke, O.; Stamm, M.; Petry, W. Macromolecules 2002, 35, 2017. (10) Costa, A. C.; Composto, R. J.; Vlcek, P. Macromolecules 2003, 36, 3254. (11) Brochard-Wyart, F.; Martin, P.; Redon, C. Langmuir 1993, 9, 3682.
and Kv is the friction coefficient due to viscous dissipation.12 If hole growth occurs by slip, D ∼ t2/3 and dD/dt ) 2S(2/D)1/2/Ks, where Ks is the friction coefficient due to slip.13 Partial slip is described by5
t)
[
(
Kv xD/2 D - 4βxD/2 + 4β2 ln 1 + 2S β
)]
+ τo
(1)
where β ) Kv/Ks and τo is the induction time to film rupture. Thin film destabilization may alternatively proceed by spinodal dewetting (SD) wherein film thickness fluctuates at the liquidair interface. The spatiotemporal evolution of surface modulations appears qualitatively similar to bicontinuous morphologies observed20 in liquid mixtures that phase-separate by spinodal decomposition. Unlike spinodal decomposition, however, SD is a consequence of an increase in the growth rate of thermally induced capillary waves as the initial film thickness is decreased.7,14-19 Instability considerations11 confirm that the dominant mode of surface fluctuations amplifies with increasing time and occurs at a fast-growing wave vector (q*). The dynamic topology of an unstable thin film on a homogeneous substrate can also be described by the conjoining film pressure due to excess intermolecular interactions (φ).21 When ∂φ/∂h < 0 (where h is the position- and time-dependent film thickness), a thin film destabilizes and flows from thin to thick regions. If a substrate is chemically heterogeneous, a spinodal-like21 instability can (12) Brochard-Wyart, F.; de Gennes, P. G.; Hervert, H.; Redon, C. Langmuir 1994, 10, 1566. (13) Brochard, F.; Redon, C.; Sykes, C. C. R. Acad. Sci. Paris II 1992, 314, 19. (14) Sferrazza, M.; Heppenstall-Butler, M.; Cubitt, R.; Bucknall, D.; Webster, J.; Jones, R. A. L. Phys. ReV. Lett. 1998, 81, 23. (15) Xie, R.; Karim, A.; Douglas, J. F.; Han, C. C.; Weiss, R. A. Phys. ReV. Lett. 1998, 81, 1251. (16) Sharma, A.; Khanna, R. Phys. ReV. Lett. 1998, 81, 3463. (17) Herminghaus, S.; Jacobs, K.; Mecke, K.; Bischof, J.; Fery, A.; Ibn-Elhaj, M.; Schlagowski, S. Science 1998, 282, 916. (18) Segalman, R. A.; Green, P. F. Macromolecules 1999, 32, 801. (19) Bollinne, C.; Cuenot, S.; Nysten, B.; Jonas, A. M. Eur. Phys. J. E 2003, 12, 389. (20) Jinnai, H.; Koga, T.; Nishikawa, Y.; Hashimoto, T.; Hyde, S. T. Phys. ReV. Lett. 1997, 78, 2248. (21) Konnur, R.; Kargupta, K.; Sharma, A. Phys. ReV. Lett. 2000, 84, 931.
10.1021/la061391j CCC: $33.50 © 2006 American Chemical Society Published on Web 09/07/2006
Letters
likewise develop as a consequence of flow from less to more wettable interfacial regions. Due to interfacial heterogeneities, thin film dewetting occurs even if ∂φ/∂h > 0.2,22,23 Although most efforts have explored thin film stability by varying film thickness, few studies have reported10,24,25 on the effect of block copolymers as interfacial modifiers in thin films even though copolymers are often used to compatibilize homopolymers. The dewetting rate of polystyrene (PS) from poly(methyl methacrylate) (PMMA) at 180 °C, for instance, is greatly reduced by adding a PS-b-PMMA (SM) diblock copolymer to the PS.26 In this work, we further show that the dewetting mechanism of PS/SM thin films of constant thickness is tunable via copolymer concentration. A PS homopolymer (M h n ) 50 kDa, M h w/M h n ) 1.06) was purchased from Pressure Chemical, Inc. (Pittsburgh, PA), whereas a PMMA homopolymer (M h n ) 226 kDa, M h w/M h n ) 1.06) and a symmetric SM copolymer (M h n ) 104 kDa, M h w/M h n ) 1.04 with 48 wt % S) were obtained from Polymer Source, Inc. (Dorval, Canada). Solvent-grade toluene used to spin-coat thin films was supplied by Fisher Scientific (Fairlawn, NJ) and used as-received. The PMMA substrate, ≈50 nm thick according to ellipsometry, was spin-coated onto silicon wafer with its native oxide intact. Films of PS/SM blends, ≈66 nm thick, were separately spincoated onto glass slides to avoid contact with the PMMA and uncontrolled migration of SM molecules to the PS/PMMA interface during film formation. The concentration of SM copolymer in the PS layer ranged from 4 to 7 wt %, which is above the critical micelle concentration estimated26 to be ∼1 wt % SM in the bulk. For brevity, these blends are designated as PS/SMw, where w denotes the mass percent of added SM copolymer. Each PS/SMw film was floated onto deionized water and deposited on top of the PMMA-coated silicon wafers to form a double layer. After drying in air at ambient temperature for at least 24 h, these assemblies were annealed in a MettlerToledo hot-stage under a circulating N2 blanket at 180 °C, which is above the glass transition temperatures of both polymers. Realtime dewetting kinetics were monitored with an Olympus BX60 optical microscope equipped with a computer-interfaced CCD camera and operated in reflection mode. Two-dimensional Fourier transforms were generated from each relevant image and collapsed to one-dimensional intensity vs q plots by circular integration. Atomic force microscopy (AFM) images of the PS surface and PS/PMMA interface (after selective PS removal in cyclohexane), presented as 5 µm × 5 µm scan areas, were collected with a Digital Instruments 3000 microscope operated in tapping mode. Planar transmission electron microscopy (TEM) images of double layered assemblies prepared on SiO2-coated Cu grids were collected on a Zeiss EM902 electron spectroscopic microscope operated at 80 kV and an energy loss of 50 eV after the specimens were annealed and subsequently exposed to the vapor of 0.5% RuO4(aq), which selectively stains the phenyl rings of PS. Images were also collected on a Gatan UltraScan 4000 CCD camera at tilt angles ranging from -53 to +67° at an interval of 2° on a Technai T20 microscope operated at 200 kV. The image set was aligned using a precalibrated geometric model, and 3D transmission electron microtomography27 (TEMT) reconstructions were generated using the r-weighted back-projection algorithm in the EMTAR software suite. (22) Kargupta, K.; Sharma, A. Langmuir 2002, 18, 1893. (23) Kargupta, K.; Sharma, A. Langmuir 2003, 19, 5153. (24) Zhu, S.; Liu, Y.; Rafailovich, M. H.; Sokolov, J.; Gersappe, D.; Winesett, D. A.; Ade, H. Nature (London) 1999, 400, 49. (25) Sung, L.; Douglas, J. F.; Han, C. C.; Karim, A. J. Polym. Sci. B: Polym. Phys. 2003, 41, 1697. (26) Wei, B.; Genzer, J.; Spontak, R. J. Langmuir 2004, 20, 8659. (27) Jinnai, H.; Nishikawa, Y.; Ito, M.; Smith, S. D.; Agard, D. A.; Spontak, R. J. AdV. Mater. 2002, 14, 1615.
Langmuir, Vol. 22, No. 21, 2006 8643
Figure 1. Optical images of a PS/SM5 thin film on PMMA at 180 °C after different times (t, in min): (a) 2, (b) 10, (c) 20, and (d) 95. The scale marker in (d) corresponds to 20 µm. An AFM height image of the PS film surface (inset) shows holes measuring 0.7-1.0 µm across after 2 min.
Destabilization of unmodified PS thin films on PMMA at 180 °C occurs via NG of circular holes. Similar results are obtained at a substantially reduced dewetting rate (dD/dt) for PS/SM1 films. At 2 and 3 wt % SM, dewetting by classical hole growth is slowed further.26 An increase in SM concentration to 4 wt %, however, yields strikingly different dewetting behavior. Optical images collected at sequential annealing times from PS/SM4, as well as PS/SM5, films such as those displayed in Figure 1 reveal that dewetting proceeds by the formation of instability patterns that resemble thin polymer films undergoing SD,7,15,18 as well as simulations21 of thin liquid films dewetting from a chemically heterogeneous substrate. As t is increased, initially discrete holes (cf. the inset in Figure 1a) merge rather than enlarge as in NG. Although the resultant patterns are reminiscent of those associated with SD, we hasten to point out at this juncture that the mechanism responsible for dewetting is not SD but may instead be related to a spinodal-like or related28 instability. This point is discussed further in a later section. Fourier analysis of images such as the ones in Figure 1 confirms that q* monotonically decreases, and the modulation wavelength (λ ) 2π/q*) increases, with increasing t. The q* values extracted from images of the PS/SM4 film at early times are ≈7% smaller than those from images of the PS/SM5 film, indicating that λ is larger in the film with lower SM content. Values of q* at time t (q* t) normalized relative to q* at 2 min (q* 0, where the data begin to show systematic trends as discerned by optical analysis) are presented as a function of t for the PS/SM4 and PS/SM5 series in Figure 2a. In the early stage of dewetting, both data sets reveal that q*t ∼ t-R, where R ) 0.332 ( 0.004 from a regression of the data up to t′ (labeled). Although this result agrees with observations reported7 for SD of a disordered SM copolymer film on silica, we attribute the dewetting mechanism apparent in Figure 1 to a complex interplay between copolymer self-assembly and heterogeneity formation along the polymer/polymer interface (see the inset in Figure 2a and Figure 2b,c for examples of such features). Specifically, the image in Figure 2c, a planar slice of a 3D TEMT reconstruction27 (28) Sferrazza, M.; Xiao, C.; Jones, R. A. L.; Bucknall, D. G.; Webster, J.; Penfold, J. Phys. ReV. Lett. 1997, 78, 3693.
8644 Langmuir, Vol. 22, No. 21, 2006
Letters
Figure 3. Variation of D(t) for thin PS and PS/SM7 films dewetting on PMMA at 180 °C (note the different upper and lower time scales). The optical image (scale marker ) 20 µm) shows a flowerlike hole in the PS/SM7 film after ∼40 h. The solid lines are regressions of eq 1 to the PS and early-stage PS/SM7 data, whereas the dashed lines are linear regressions to all the data sets.
Figure 2. In a, the time dependence of the fastest-growing wave vector (q*t), normalized relative to q*t at 2 min (q*0), for PS/SM4 and PS/SM5 thin films on PMMA at 180 °C. The solid line is a powerlaw fit to the data. The inset is an AFM height image of the stripped PS/PMMA interface and displays interfacial elements (adsorbed micelles) measuring 90-300 nm across after 2 min. In b (scale marker ) 2 µm), a TEM image of a PS/SM5 thin film on PMMA after 5 min at 180 °C reveals copolymer aggregates measuring ca. 90-280 nm across. The inset is a 3× linear enlargement. In c (scale marker ) 500 nm), a planar slice of a 3D TEMT image27 along the PS/PMMA interface after 15 min at 180 °C confirms the existence of interfacial copolymer structural elements including adsorbed micelles and brush patches. Interfacial micelles and a large brush patch are labeled, and free micelles near the interface are circled.
along the PS/PMMA interface, confirms the existence of interfacial copolymer structures in situ, without having to remove the top PS layer (as in the AFM analysis). Our constraint of constant film thickness serves to differentiate thin liquid films that dewet from solid homogeneous substrates by SD from the present thin liquid films that, due to the presence of a species capable of molecular self-organization, modify interfacial tension and nanoscale structure as they dewet from a substrate. If the copolymer concentration in the PS layer is increased further to 7 wt %, the dewetting mechanism reverts back to NG, this time resulting in the formation of noncircular, flower-like holes (cf. the inset of Figure 3). The area-equivalent diameter (Deq) of a noncircular hole observed in the PS/SM7 film is presented as a function of t in Figure 3. Hole growth data collected from the neat PS thin film (included for comparison in Figure 3) indicate that the onset of dewetting discernible by optical microscopy occurs within seconds (similar dewetting time scales are representative26 of PS/SM blends with up to 3 wt % SM), but a lengthy induction time (τo ≈ 14 h) is required for macroscopic hole formation in the PS/SM7 system. Although τo varies due to spatiotemporal differences in micellar formation,
adsorption, and aggregation along the polymer/polymer interface, the results presented in Figure 3 are over 4 orders of magnitude slower than for the PS film at 180 °C. Thus, a molecular-scale chain of events requiring a relatively long time scale (such as micellar formation/aggregation or interfacial reorganization) occurs prior to rupture in the PS/SM7 film. Another feature of this system is the existence of two dewetting regimes suggestive of early- and late-stage processes, as previously described in dewetting studies of a homopolymer and a random copolymer18 or PS/SM thin films (with 2 and 3 wt % SM) and PMMA.26 The dependence of Deq on t in Figure 3 is nearly linear in both the early- and late-stage dewetting regimes, which implies that dewetting occurs predominantly by viscous dissipation. Measured values of dD/dt are displayed as a function of SM concentration and, by inference from the included AFM and TEM images, the extent of interfacial modification in Figure 4. As anticipated, dD/dt decreases systematically with increasing copolymer content until the PS/SM layer is stabilized. Although the time dependence of hole growth in these PS/SM films appears nearly linear, application of eq 1 to the D(t) data collected from the neat PS and PS/SM1 films, as well as the early-stage Deq(t) data from the PS/SM7 film, yields marginally improved model fits (cf. Figure 3). Partial-slip analysis5 provides the viscous dissipation contribution to the overall hole growth velocity. Regressed values of 2S/Kv, deduced from eq 1 and included in Figure 4, compare favorably with constant dewetting rates assuming linear hole growth. This analysis also reveals the composition dependence of β, the ratio of friction coefficients due to viscous dissipation and interfacial slip (Kv/Ks). Hole growth under full-slip and no-slip conditions5 yields β f ∞ and β f 0, respectively. Measured values of β are 1.12 ( 0.30 (PS), 0.60 ( 0.10 (PS/SM1), and 1.16 ( 0.38 (PS/SM7) µm1/2. A sequence of molecular-level events consistent with the dewetting behavior reported here begins with the formation of copolymer-induced interfacial structures. At low SM concentrations (regime I in Figure 4), SM unimers and micelles migrate to the PS/PMMA interface. Copolymer micelles form in the PS/SM layer upon heating and subsequently adsorb and spread out along the PS/PMMA interface to produce patches of SM molecules that alter interfacial tension and slow hole growth. The dewetting mechanism in this regime is unchanged since the PS/PMMA interface remains largely intact, as confirmed by the AFM image of the PS-stripped interface. At higher SM concentrations (regime II), copolymer patches become more densely packed and numerous due to the larger population of SM micelles present. The micelles and related brush patches
Letters
Langmuir, Vol. 22, No. 21, 2006 8645
Figure 4. Dewetting rate of PS/SM thin films on PMMA at 180 °C as a function of SM concentration on the basis of no-slip (constant dD/dt, O) and partial slip (eq 1, 4) along the PS/PMMA interface before accelerated dewetting. Accelerated dewetting rates are also included when relevant (b), and the solid line is a guide for the eye. AFM height images of the PS-stripped PS/PMMA interface display copolymer elements measuring 50-300 nm across after 18 h (regime I), 24 h (regime II) and 100 h (regime III). The illustrations are discussed in the text.
self-organize along the polymer/polymer interface, as seen in the corresponding AFM image, ultimately resulting in the formation of larger-scale interfacial structures that vary with both position and time. Heterogeneities in the PS layer and along the PS/PMMA interface isolate small perforations that extend through the PS layer, as evidenced by Figure 2b (a hole through both layers would appear dark at ∆E ) 50 eV). As a result, the dewetting mechanism switches to fluctuations and exhibits characteristics similar to those predicted21 for thin-film dewetting from a chemically heterogeneous substrate. These supramolecular processes occur as the PS layer undergoes dewetting and thus alter the nanostructure of the PS layer, as well as the nature of the PS/PMMA interface, over the time scale investigated. By including the propensity for concurrent molecular self-organization, the current system extends earlier studies29 designed to elucidate the effect of adsorbed brushes composed of endfunctionalized polymer chains on dewetting. As the SM concentration is increased further (regime III), large patches of SM copolymer form due to continued migration and aggregation of micelles30,31 along the PS/PMMA interface. Due to the dynamic nature of these structural elements (which continue to evolve and rearrange over time), exposed regions of neat PMMA substrate along the PS/PMMA interface eventually develop and promote PS/SM film rupture, thereby yielding a measurable τo, which is predicted by eq 1. The irregular shape of the holes in the PS/SM7 films is likewise reminiscent of pinning and model predictions reported20 for thin-film dewetting from a flat, chemically heterogeneous substrate. Alternatively, the presence of dynamic copolymer assemblies within the PS/SM film may induce spatially heterogeneous viscosities that promote viscous fingering. In any event, the influence of the added copolymer on dewetting is unmistakable and bears close resemblance to dewetting behavior previously reported32 for a pure block copolymer on a homogeneous substrate. At sufficiently (29) Henn, G.; Bucknall, D. G.; Stamm, M.; Vanhoorne, P.; Je´roˆme, R. Macromolecules 1996, 29, 4305. Yuan, C. G.; Meng, O. Y.; Koberstein, J. T. Macromolecules 1999, 32, 2329. (30) Shull, K. R.; Winey, K. I.; Thomas, E. L.; Kramer, E. J. Macromolecules 1991, 24, 2748. (31) Hamley, I. W.; Connell, S. D.; Collins, S. Macromolecules 2004, 37, 5337. (32) Lee, S. H.; Kang, H.; Cho, J.; Kim, Y. S.; Char, K. Macromolecules 2001, 34, 8405.
high copolymer levels, copolymer micelles and aggregates thereof along the interface are anticipated to altogether screen repulsive PS and PMMA interactions and promote film stabilization. Examples of the interfacial copolymer structures described above (e.g., patches of adsorbed SM micelles and micellar aggregates) are evident in AFM images of the PS-stripped interface (cf. Figures 2a and 4), as well as in TEMT images (Figure 2c) of the PS/ SM/PMMA system. These heterogeneities resemble block copolymer micelles adsorbed from solution onto solid substrates.31 While block copolymers are routinely employed as compatibilizing agents in polymer blends, this study establishes that they can likewise be used to tailor the mechanism2 and spatiotemporal motif by which one polymer film dewets from another. Copolymer concentration, as well as architecture,33 provides freedom not only in stabilizing thin polymer films of constant thickness, but also in controlling the mechanism, rate and induction period by which dewetting proceeds. Our results (i) provide experimental evidence supporting theoretical predictions21-23 regarding the influence of interfacial heterogeneities on film dewetting, (ii) raise fundamental questions regarding the competition between film dewetting19,34 and simultaneous phase and interfacial modification,9,25,35 as well as the time scales involved, and (iii) open new avenues to designer polymer films on polymer substrates for advanced coating and templating technologies.36 Acknowledgment. This work was supported by the Kenan Institute for Engineering, Technology & Science at North Carolina State University and the National Institutes of Health and the W. M Keck Foundation at the University of California at San Francisco. We thank E. J. Kramer and J. F. Douglas for their comments regarding this work and valuable advice. LA061391J (33) Costa, A. C.; Composto, R. J.; Vlcek, P.; Morera, S. J. Adhes. 2005, 81, 683. (34) Pototsky, A.; Bestehorn, M.; Merkt, D.; Thiele, U. Phys. ReV. E 2004, 70, Art. No. 025201. (35) Mu¨ller-Buschbaum, P.; Bauer, E.; Wunnicke, O.; Stamm, M. J. Phys.Cond. Mater. 2005, 17, S363. (36) Mu¨ller-Buschbaum, P.; Bauer, E.; Maurer, E.; Schlogl, K.; Roth, S. V.; Gehrke, R. Appl. Phys. Lett. 2006, 88, Art. No. 083114.
