Polyhedral

6734

Langmuir 2006, 22, 6734-6738

Pattern Formation in Dewetting Poly(tert-butyl acrylate)/Polyhedral Oligomeric Silsesquioxane (POSS) Bilayer Films Rituparna Paul and Alan R. Esker* Macromolecules and Interfaces Institute and the Department of Chemistry (0212), Virginia Polytechnic Institute and State UniVersity, Blacksburg, Virginia 24061 ReceiVed April 10, 2006. In Final Form: June 17, 2006 The surface morphology of dewetting poly(tert-butyl acrylate) (PtBA) and trisilanolphenyl-POSS (TPP) bilayers has been studied as a function of time at 95 °C. For short annealing times, only the upper nanoparticle (TPP) layer dewets from the underlying PtBA layer. The number and lateral dimensions of the holes in the upper TPP layer increase with increasing annealing times, forming interconnected rim structures. At later annealing times, scattered holes that reach down into the PtBA layer are observed among the interconnected rim structures. Fractal nanofiller (TPP)-rich aggregates are found at the bottom of the scattered holes.

Introduction Polymer thin films on solid substrates are used for a number of technological applications that range from the fabrication of microelectronics and sensors1,2 to novel drug delivery systems3 and biocompatible coatings.4 The performance of thin polymer films is dependent upon their stability after being spread on a substrate. Dewetting, the process of spontaneous hole formation in amorphous polymer thin films above their glass-transition temperature, Tg, can pose a serious problem in the nanofabrication of polymeric coatings. Two mechanisms are commonly proposed for dewetting: nucleation and growth, and spinodal dewetting. In the nucleation and growth mechanism, hole formation is initiated by an impurity in the film and/or a substrate defect.5 In contrast, hole formation for spinodal dewetting is initiated by capillary waves on the surface that arise from density fluctuations in the polymer.6 The capillary waves vary in amplitude and can initiate holes upon contact with the substrate. In both mechanisms, the polymer begins to pull away from the nucleation sites upon heating above Tg and collects in rims surrounding circular holes. Although dewetting is a hindrance for many applications, research has shown that dewetting can be used advantageously in homopolymer and copolymer films7-16 and also in polymer/ * To whom correspondence should be addressed. E-mail: [email protected]. Fax: (01) 540-231-3255. (1) Chabinyc, M. L.; Wong, W. S.; Salleo, A.; Paul, K. E.; Street, R. A. Appl. Phys. Lett. 2002, 81, 4260-4262. (2) Thomas, S. W., III; Amara, J. P.; Bjork, R. E.; Swager, T. M. Chem. Commun. 2005, 36, 4572-4574. (3) Grayson, A. C. R.; Voskerician, G.; Lynn, A.; Anderson, J. M.; Cima, M. J.; Langer, R. J. Biomater. Sci., Polym. Ed. 2004, 15, 1281-1304. (4) Ratner, B. J. Biomed. Mater. Res. 1993, 27, 837-850. (5) Lorenz-Haas, C.; Mu¨ller-Buschbaum, P.; Kraus, J.; Bucknall, D. G.; Stamm, M. Appl. Phys. A 2002, 74, S383-S385. (6) Xie, R.; Karim, A.; Douglas, J. F.; Han, C. C.; Weiss, R. A. Phys. ReV. Lett. 1998, 81, 1251-1254. (7) Nedelcu, M.; Morariu, M. D.; Harkema, S.; Voicu, N. E.; Steiner, U. Soft Matter 2005, 1, 62-66. (8) Wang, X.; Tvingstedt, K.; Ingana¨s, O. Nanotechnology 2005, 16, 437443. (9) Suh, K. Y.; Park, J.; Lee, H. H. J. Chem. Phys. 2002, 116, 7714-7718. (10) Sehgal, A.; Ferreiro, V.; Douglas, J. F.; Amis, E. J.; Karim, A. Langmuir 2002, 18, 7041-7048. (11) Kargupta, K.; Sharma, A. Phys. ReV. Lett. 2001, 86, 4536-4539. (12) Seemann, R.; Herminghaus, S.; Jacobs, K. J. Phys.: Condens. Matter 2001, 13, 4925-4938. (13) Meyer, E.; Braun, H.-G. Macromol. Mater. Eng. 2000, 276/277, 44-50. (14) Karthaus, O.; Gråso¨, L.; Maruyama, N.; Masatgasu, S. Chaos 1999, 9, 308-314. (15) Dalnoki-Veress, K.; Nickel, B. G.; Dutcher, J. R. Phys. ReV. Lett. 1999, 82, 1486-1489. (16) Reiter, G. Science 1998, 282, 888-889.

polymer bilayers to fabricate patterned surfaces with tunable morphologies.8,17 There have been a few attempts to understand dewetting in polymer/nanoparticle thin films. However, these studies essentially focused on the suppression of dewetting by the addition of small amounts of nanoparticles,18-22 and relatively little is known about utilizing dewetting in polymer/nanoparticle films to create surfaces with specific morphologies.23,24 Unlike most previous studies of dewetting in polymer/nanoparticle systems, our work focuses on gaining fundamental insight into dewetting behavior in polymer/nanoparticle bilayer films and uses this knowledge for reproducibly controlling pattern formation in dewetting bilayers. The surfaces patterned via tunable dewetting may find potential uses in the fields of chemical warfare agent (CWA) detection, sorption, and degradation.25,26 This study utilizes bilayer films of poly(tert-butyl acrylate) (PtBA) and a polyhedral oligomeric silsesquioxane (POSS) derivative, trisilanolphenyl-POSS (TPP), having a Si//PtBA/TPP// air configuration as a model to explore the morphological evolution of dewetting in polymer/nanoparticle bilayers at elevated temperatures. Here, “// ” signifies a distinct interface, and “/ ” is used to depict a dynamic interface between mobile components. The POSS component of the model bilayers is an organic/inorganic hybrid material consisting of a rigid inorganic (Si-O) core and a flexible organic corona.27,28 POSS materials have gained considerable attention for their applications in hightemperature and space-survivable coatings,29 shape-memory (17) Higgins, A. M.; Jones, R. A. L. Nature 2000, 404, 476-478. (18) Krishnan, R. S.; Mackay, M. E.; Hawker, C. J.; Van Horn, B. Langmuir 2005, 21, 5770-5776. (19) Hosaka, N.; Tanaka, K.; Otsuka, H.; Takahara, A. Compos. Interfaces 2004, 11, 297-306. (20) Sharma, S.; Rafailovich, M. H.; Peiffer, D.; Sokolov, J. Nano Lett. 2001, 1, 511-514. (21) Barnes, K. A.; Douglas, J. F.; Liu, D.-W.; Karim, A. AdV. Colloid Interface Sci. 2001, 94, 83-104. (22) Barnes, K. A.; Karim, A.; Douglas, J. F.; Nakatani, A. I.; Gruell, H.; Amis, E. J. Macrmolecules 2000, 33, 4177-4185. (23) Lee, L.-T.; Leite, C. A. P.; Galembeck, F. Langmuir 2004, 18, 70417048. (24) Huang, J.; Kim, F.; Tao, A. R.; Connor, S.; Yang, P. Nat. Mater. 2005, 4, 896-900. (25) Ferguson-McPherson, M. K.; Low, E. R.; Esker, A. R.; Morris, J. R. J. Phys. Chem. B 2005, 109, 18914-18920. (26) Ferguson-McPherson, M. K.; Low, E. R.; Esker, A. R.; Morris, J. R. Langmuir 2005, 21, 11226-11231. (27) Haddad, T. S.; Viers, B. D.; Phillips, S. H. J. Inorg. Organomet. Polym. 2001, 11, 155-164. (28) Voronkov, M. G.; Lavrent, V. I. Top. Curr. Chem. 1982, 102, 199-236. (29) Hoflund, G. B.; Gonzalez, R. I.; Phillips, S. H. J. Adhes. Sci. Technol. 2001, 15, 1199-1211.

10.1021/la060973y CCC: $33.50 © 2006 American Chemical Society Published on Web 07/11/2006

Letters

Langmuir, Vol. 22, No. 16, 2006 6735

materials,30 semiconducting polymers,31 and synthetic templates for nanomaterials.32 Recent work with Langmuir-Blodgett (LB) films of TPP shows that organophosphonate-based CWA simulants adsorb to TPP and that chlorinated organophosphate CWA simulants even decompose at elevated temperatures.25,26 Thus, dewet polymer/TPP bilayers may be useful in the fabrication of sensors and sorbent materials with faster response times for CWA detection and decomposition. Experimental Section Materials. PtBA (number-average molar mass, Mn, of 5.0 kg‚mol-1; polydispersity index, Mw/Mn, of 1.25) and TPP were obtained from Polymer Source, Inc. and Hybrid Plastics, Inc., respectively, and used without further purification. Silicon wafers (100) were obtained from Waferworld, Inc. Cleaning Procedure for Silicon Substrates. Prior to spin coating, the silicon wafers were boiled in a 1:1:5 (by volume) solution of ammonium hydroxide:hydrogen peroxide:ultrapure water (Millipore, Milli-Q Gradient A-10, 18.2 MΩ, 15 min. This phenomenon may occur because both the upper TPP and lower PtBA layers are dewetting. The formation of TPP-rich aggregates is consistent with strong interparticle interactions.21,22 Upon annealing bilayers at 95 °C for substantially longer annealing times, including overnight annealing, isolated holes containing filler-rich aggregates can merge, resulting in the formation of extended networks of TPPrich aggregates (Figure 2c). For bilayers annealed at temperatures above 95 °C, the morphologies are similar to those observed for films annealed at 95 °C. The only difference is the expected faster rate of morphological evolution with increasing temperature. In contrast, the TPP-rich aggregates are not observed upon annealing the PtBA/TPP bilayers below 95 °C for similar annealing times. At this time, it cannot be concluded if longer annealing times will also produce TPP-rich aggregates. The optical micrograph of PtBA/TPP bilayers annealed at 85 °C for 60 min is provided in the Supporting Information (Figure S1). At this stage, it is important to note that spin-coated PtBA films on hydrophobic (34) Esker, A. R.; Mengel, C.; Wegner, G. Science 1998, 280, 892-895. (35) Lu, Z.; Liu, G.; Duncan, S. Macromolecules 2004, 37, 174-180.

6736 Langmuir, Vol. 22, No. 16, 2006

Letters

Figure 3. Tapping mode AFM (a) height and (b) phase images of a PtBA/TPP bilayer film annealed at 95 °C for 15 min. The scan size is 5 × 5 µm2; z ranges for the height and phase images are 20 nm and 20°, respectively. (c) Representative line profile (along the black line in part a) used to determine hole depths in a PtBA/TPP bilayer film annealed at 95 °C for 15 min. The reference line for the line profile analysis is set to the uniform TPP//air interface (0 nm).

Figure 2. Optical micrographs (0.76 × 0.57 mm2 at 20× magnification) of PtBA/TPP bilayer films annealed at 95 °C for (a) 15, (b) 60, and (c) 840 min. The inset in part b represents a 0.09 × 0.07 mm2 section of the original image that has been cut and enlarged using imaging software to enhance the clarity of the aggregated filler-rich structures in the hole. Table 1. Surface Atomic Concentration Ratios of Elemental Silicon [Si(ele)] and Silicon Bound to Oxygen [Si(-O)] Relative to Carbon (C) as a Function of Annealing Time for PtBA/TPP Bilayers Annealed at 95 °C annealing time (min)

Si(ele)/C

Si(-O)/C

0 15 30 60 120 840

0.000 0.000 0.000 0.003 0.004 0.004

0.177 0.129 0.076 0.032 0.066 0.153

silicon substrates do not dewet in the absence of TPP nanoparticles upon annealing at 95 °C for annealing times similar to those of the bilayer samples. (OM and AFM images are provided in the Supporting Information, Figure S2 and Figure S3.) Similarly, TPP nanoparticle layers do not dewet from hydrophobic silicon substrates for comparable annealing times and temperatures. XPS was used to determine the time-dependent surface composition evolution of the annealed PtBA/TPP bilayers. The unannealed bilayers exhibit the expected initial composition of pure TPP (top layer). Upon annealing the bilayers at 95 °C, the surface atomic concentration ratio of silicon bound to oxygen relative to carbon [Si(-O)/C] decreases with increasing annealing

times up to 60 min, indicating dewetting of the upper TPP layer and exposure of the lower PtBA layer. However, for bilayers annealed longer than 1 h, the values of Si(-O)/C increase (Table 1). This feature may be attributed to higher fractional surface area coverage by TPP-rich aggregates that form extended structures at longer annealing times as observed by OM (Figure 2c). Limited dewetting of both the TPP and PtBA layers down to the silicon substrate at long annealing times was confirmed by the presence of elemental silicon peaks at ∼99 eV in the XPS scans of samples annealed for up to 14 h (Table 1). The dewetting of the bilayer down to the silicon substrate observed by XPS appears to be confined to the TPP-rich holes (Figure 2b and c). Nonetheless, the low Si(ele)/C ratios mean that both PtBA and TPP are more prevalent than Si(ele) at the bottom of the holes. This point is important for explaining the relatively slow merger of isolated holes versus the rapid formation of holes with TPPrich nanoparticle aggregates. The localization of TPP-rich aggregates on the silicon substrate may act as pseudo-crosslinking sites, leading to an increase in the local viscosity of the dewetting polymer layer present at the bottom of the hole. The enhanced local viscosity results in the retardation of gross dewetting of the lower PtBA layer as evidenced by the similar size of isolated holes in Figure 2b and c.21,22,36 AFM results provide an in-depth understanding of the morphological evolution in the dewetting PtBA/TPP bilayers. The bilayer films form isolated circular holes with low phase contrast surrounded by rims immediately upon annealing. With increasing annealing times, the number and the lateral dimensions of the holes increase, resulting in the coalescence of the rims and the formation of interconnected rim structures (Figure 3). The low phase contrast of the AFM images for samples that were annealed for short annealing times indicates that the holes are most likely formed by the dewetting of only the upper TPP layer from the PtBA underlayer. The relative hole depths (∼6-15 nm) of bilayers annealed for 15 min determined from line profile analyses are smaller than the initial thickness of the upper TPP (36) Lou, H.; Gersappe, D. Macromolecules 2004, 37, 5792-5799.

Letters

Figure 4. Tapping mode AFM (a) height and (b) phase images of intricate aggregates found at the bottom of isolated holes that form in a PtBA/TPP bilayer film upon annealing at 95 °C for 60 min. The scan size is 20 × 20 µm2; z ranges for the height and phase images are 200 nm and 20°, respectively.

Figure 5. Tapping mode AFM (a) height and (b) phase images of interconnected rim structures that form in the gray regions outside the isolated holes highlighted in Figure 2b for a PtBA/TPP bilayer film upon annealing at 95 °C for 60 min. The scan size is 5 × 5 µm2; z ranges for the height and phase images are 20 nm and 20°, respectively.

(∼25 nm) layer (Figure 3c). The reference lines for the line profile analyses are set at the initial uniform TPP//air interface. This observation suggests that the holes are in the upper TPP layer and that the PtBA underlayer is most likely pulled into the holes and rims, resulting in hole depths that are smaller than the initial TPP layer thickness.37-39 With increasing annealing times, the dewetting of the PtBA layer ensues, but the exposure of the silicon substrate is limited as observed by XPS. For bilayers annealed for more than 15 min, scattered holes form with two key features. One feature is intricate, fractal-like aggregated structures with strong height and phase contrast that form at the bottom of the holes (Figure 4). The other feature is interconnected rim structures outside the holes (Figure 5). The high phase contrast of the aggregated features (arising from the difference in hardness between PtBA and TPP) indicates that these structures are essentially composed of TPP and that PtBA remains at the bottoms of the holes (consistent with XPS results in Table 1). The holes with aggregated filler-rich structures may have been formed because both the upper TPP and lower PtBA layers dewet, thereby causing the fillers to trace out radial paths from the dewetting centers as the growing holes expand. Weak polymer-nanoparticle interactions coupled with reasonably strong nanoparticlesubstrate interactions most likely cause the TPP-rich aggregates to be left behind at the bottom of the holes as the polymer layer thins, leading to TPP depletion in the rims.21,22,36 The fractal dimension, df, of the aggregated TPP-rich structures observed in PtBA/TPP bilayers annealed for 60 min is estimated to be ∼2.2, a value that is consistent with fractal growth via clustercluster aggregation (CCA) in three dimensions (df ≈ 1.6-2.2).40,41 (37) Pan, Q.; Winey, K. I.; Hu, H. H.; Composto, R. J. Langmuir 1997, 13, 1758-1766. (38) Lambooy, P.; Phelan, K. C.; Haugg, O.; Krausch, G. Phys. ReV. Lett. 1996, 76, 1110-1113. (39) Brochard-Wyart, F.; Pascal, M.; Redon, C. Langmuir 1993, 9, 36823690. (40) Witten, T. A.; Cates, M. E. Science 1986, 232, 1607-1612. (41) Meakin, P.; Jullien, R. J. Chem. Phys. 1988, 89, 246-250.

Langmuir, Vol. 22, No. 16, 2006 6737

Figure 6. Tapping mode AFM (a) height and (b) phase images of TPP aggregates formed at the bottom of isolated holes in a PtBA/ TPP bilayer annealed at 95 °C for 60 min. For these images, the PtBA film was acid hydrolyzed (gas phase) to PAA,34 and the PAA was removed with water. The scan size is 40 × 40 µm2; z ranges for the height and phase images are 200 nm and 20°, respectively.

As the size of the fractal domains increases, it is likely that the interactions between the fractal domains cause the arms within the domains to stop growing locally at different times. The growth of the most advanced arms stop first because of the local depletion of TPP, whereas the lagging arms continue to grow.42 An AFM height image of fractal domains approaching each other in a PtBA/TPP bilayer film annealed at 95 °C is provided in the Supporting Information (Figure S4). This figure shows how impinging aggregates stunt the growth of longer-armed features. To confirm that the dendritic aggregates consist mainly of TPP, PtBA is selectively removed from the bilayers by acid hydrolysis,34 and the morphology of these bilayers is studied by AFM. The aggregates are clearly observed as bright, elevated features in the AFM height image (Figure 6a) after the selective removal of PtBA, indicating that they are indeed composed of TPP. Note that the heights of the fractal structures are similar to those in Figure 4a. However, compared to those in Figure 4b, the TPP aggregates shown in Figure 6b have very low phase contrast. The decrease in phase contrast is attributed to the fact that TPP has a hardness that is similar to that of the silicon substrate that is exposed after the residual PtBA is removed from the fractal structures and the bottoms of the holes. An analysis of the fractal dimensions of the structure in Figure 6a reveals that df ≈ 2.15. Because this value is comparable to df values for TPP-rich nanoparticle aggregates in films that have undergone only thermal annealing, it would appear that the selective solvent etching process does not introduce significant structural changes and that the fractal aggregates are primarily composed of TPP. To our knowledge, this study provides the first experimental (morphological) observation of micrometer-scale, fractal-like nanoparticle aggregates within holes in dewetting polymer/ nanoparticle bilayers. The radial patterns formed by the TPP aggregates within the holes upon the dewetting of both the PtBA and TPP layers are similar to Luo and Gersappe’s simulated traces of nanofiller movements on substrates during the dewetting of a filled (8 vol %) polymer film with weak polymer-filler interactions.36 Barnes et al. have suggested the formation of fractal-like nanoparticle aggregates that suppress dewetting via contact line pinning in fullerene (C60)-filled polymer films;21,22 however, the morphology of such aggregates has not been observed experimentally. The formation of localized “microtree” patterns has been reported for the dewetting of metal/resist bilayers during resist stripping processes,43 but the mechanism of pattern formation in these systems is different from the formation of nanoparticle aggregates in the PtBA/TPP bilayers. In the metal/ resist bilayers, the dendritic structures arise from the simultaneous (42) Knu¨fing, L.; Hauke, S.; Reigler, H.; Klaus, M. Langmuir 2005, 21, 9921000. (43) Wu, Y.; Qiao, P.; Chong, T.; Low, T.-S.; Xie, H.; Lou, P.; Guo, Z.; Qiu, J. Appl. Phys. Lett. 2001, 78, 3361-3363.

6738 Langmuir, Vol. 22, No. 16, 2006

dewetting of both the metal and resist layers initiated by elevated temperatures at extremely localized areas in the absence of dewetting in the rest of the film. In contrast, the formation of fractal TPP aggregates in the PtBA/TPP bilayers results from the stepwise dewetting of the upper TPP and lower PtBA layers, causing the nanoparticles to trace out radial paths within the holes in the PtBA layer. Furthermore, no uniform PtBA/TPP bilayer is observed for any area of the films after the TPP-rich fractal aggregates have formed.

Conclusions In summary, a two-stage dewetting process is observed upon annealing the PtBA/TPP bilayers at 95 °C. For short annealing times (15 min). The dewetting of both layers results in the formation of fractal-like nanoparticle aggregates on the micrometer length scale that are left behind at the bottoms of the holes. These holes do not reach the silicon substrate to any large extent because PtBA is still present. It is likely that the strong interparticle interactions coupled with weak nanoparticle-polymer interactions cause the formation of the TPP-rich fractal aggregates that are composed almost exclusively of TPP. The fractal dimension, df, of the aggregated TPP-rich

Letters

structures observed in PtBA/TPP bilayers annealed for 60 min are determined to be ∼2.2, suggesting fractal growth via the cluster-cluster aggregation mechanism. Upon annealing the bilayers at 95 °C for longer annealing times (>1 h), isolated holes containing filler-rich aggregates can merge, resulting in the formation of extended networks of TPP-rich aggregates. However, this process is much slower (on the order of several hours) than the rapid formation of holes with TPP-rich aggregates (∼30 min). These aggregate structures are believed to retard gross dewetting of the lower PtBA layers by the formation of pseudo-cross-linking sites that increase the local viscosity of the polymer. Acknowledgment. We are grateful to the Macromolecules and Interfaces Institute (MII) at Virginia Tech and the National Science Foundation (NSF) (CHE-239633) for funding and Edwards Air Force Base for materials. The help provided by Frank Cromer, Steve McCartney, and Ufuk Karabiyik in analyzing the samples is greatly appreciated. Supporting Information Available: Optical micrographs of PtBA/TPP bilayers annealed at 85 °C, optical micrographs and AFM images of spin-coated PtBA films annealed at 95 °C, and AFM height image of a PtBA/TPP bilayer annealed at 95 °C for 1 h (40 × 40 µm2 scan). This material is available free of charge via the Internet at http://pubs.acs.org. LA060973Y