11092
Langmuir 2006, 22, 11092-11096
Fabrication of Nanoporous Templates from Diblock Copolymer Thin Films on Alkylchlorosilane-Neutralized Surfaces Angelika Niemz,*,† Krisanu Bandyopadhyay,†,‡,§ Eric Tan,† Kitty Cha,‡ and Shenda M. Baker*,‡ Keck Graduate Institute, 535 Watson DriVe, Claremont, California 91711, and Department of Chemistry, 301 East 12th Street, HarVey Mudd College, Claremont, California 91711 ReceiVed September 5, 2006 The fabrication of nanoporous templates from poly(styrene)-b-poly(methyl methacrylate) diblock copolymer thin films (PS-b-PMMA, volume ratio 70:30) on silicon requires precise control of interfacial energies to achieve a perpendicular orientation of the PMMA cylindrical microdomains relative to the substrate. To provide a simple, rapid, yet tunable approach for surface neutralization, we investigated the self-assembled ordering of PS-b-PMMA diblock copolymer thin films on silicon substrates modified with a partial monolayer of octadecyldimethyl chlorosilane (ODMS), i.e., a layer of ODMS with a grafting density less than the maximum possible monolayer surface coverage. We demonstrate herein the fabrication of nanoporous PS templates from annealed PS-b-PMMA diblock copolymer thin films on these partial ODMS SAMs.
Introduction The self-assembly of diblock copolymers at interfaces enables the generation of nanoscale structures in a parallel, scalable, bottom-up fashion.1-4 Nanoporous templates can be created from poly(styrene)-b-poly(methyl methacrylate) diblock copolymer thin films (PS-b-PMMA, approximate volume ratio 70:30) on silicon.5-7 The cylindrical PMMA microdomains within these films are induced to align perpendicular to the interface and are subsequently removed through UV exposure and solvent processing. The remaining cross-linked PS forms a stable thin film with pores that are tens of nanometers in center-to-center spacing, depending on the diblock copolymer molecular weight. These nanoporous templates have been applied in the fabrication of semiconductor capacitors,8 metallic,9 and magnetic10,11 nanostructures and have been used for the deposition of semiconductor quantum dots12,13 and DNA-functionalized Au nanospheres through self-assembly.14 * To whom correspondence should be addressed. Tel.: (909) 621-8643 (S.M.B.), (909) 607-9854 (A.N.). E-mail:
[email protected] (S.M.B.),
[email protected] (A.N.). † Keck Graduate Institute. ‡ Harvey Mudd College. § Current address: Department of Natural Sciences, University of Michigan-Dearborn, 4901 Evergreen Road, Dearborn, MI 48128-1491. (1) Marrian, C. R. K.; Tennant, D. M. J. Vac. Sci. Technol. A 2003, 21, S207S215. (2) Hamley, I. W. Nanotechnology 2003, 14, R39-R54. (3) Lazzari, M.; Lopez-Quintela, M. A. AdV. Mater. 2003, 15, 1583-1594. (4) Park, C.; Yoon, J.; Thomas, E. L. Polymer 2003, 44, 6725-6760. (5) Thurn-Albrecht, T.; Steiner, R.; DeRouchey, J.; Stafford, C. M.; Huang, E.; Bal, M.; Tuominen, M.; Hawker, C. J.; Russell, T. AdV. Mater. 2000, 12, 787-791. (6) Guarini, K. W.; Black, C. T.; Yeuing, S. H. I. AdV. Mater. 2002, 14, 1290-1294. (7) Xu, T.; Kim, H. C.; DeRouchey, J.; Seney, C.; Levesque, C.; Martin, P.; Stafford, C. M.; Russell, T. P. Polymer 2001, 42, 9091-9095. (8) Black, C. T.; Guarini, K. W.; Milkove, K. R.; Baker, S. M.; Russell, T. P.; Tuominen, M. T. Appl. Phys. Lett. 2001, 79, 409-411. (9) Shin, K.; Leach, K. A.; Goldbach, J. T.; Kim, D. H.; Jho, J. Y.; Tuominen, M.; Hawker, C. J.; Russell, T. P. Nano Lett. 2002, 2, 933-936. (10) Liu, K.; Baker, S. M.; Tuominen, M.; Russell, T. P.; Schuller, I. K. Phys. ReV. B 2001, 63, 060403. (11) Montero, M. I.; Liu, K.; Stoll, O. M.; Hoffmann, A.; Akermann, J. J.; Martin, J. I.; Vicent, J. L.; Baker, S. M.; Russell, T. P.; Leighton, C.; Nogues, J.; Schuller, I. K. J. Phys. D: Appl. Phys. 2002, 35, 2398-2402. (12) Misner, M. J.; Skaff, H.; Emrick, T.; Russell, T. P. AdV. Mater. 2003, 15, 221-224.
The fabrication of nanoporous templates from PS-b-PMMA thin films requires precise control of interfacial energies to orient the PMMA cylindrical microdomains perpendicular to the substrate. Neutralized surfaces, i.e., surfaces exhibiting equal interaction energies with PS and PMMA, can be obtained by modifying the oxide layer on silicon with a covalently anchored hydroxyl-terminated random copolymer (PS-r-PMMA) termed a “neutral brush”.15-17 Surface neutralization on a variety of substrates can also be accomplished using a layer of PS-r-PMMA cross-linked via randomly incorporated benzocylcobutene functional groups.18 This more general approach creates a stable, cross-linked thin film on the substrate. Both methods enable precise fine-tuning of interfacial energies, but involve a lengthy process, require starting materials that are either commercially available but expensive or not commercially available, and result in a relatively thick underlayer that can interfere with subsequent surface processing. Neutralization can also be achieved by removing the oxide layer and hydrogen-passivating the underlying silicon,7,14 a simple and rapid method that, however, does not allow for precise control of the surface energies. Modified alkyl chlorosilane monolayers on silicon have been applied to control the microdomain orientation of lamellar PSb-PMMA diblock copolymer thin films.19-22 The surface energy of a dense alkyl chlorosilane self-assembled monolayer (SAM) can be adjusted through oxidative treatment via X-rays in air20,21 (13) Zhang, C. L.; Xu, T.; Butterfield, D.; Misner, M. J.; Ryu, D. Y.; Emrick, T.; Russell, T. P. Nano Lett. 2005, 5, 357-361. (14) Bandyopadhyay, K.; Tan, E.; Ho, L.; Bundick, S.; Baker, S. M.; Niemz, A. Langmuir 2006, 22, 4978-4984. (15) Mansky, P.; Liu, Y.; Huang, E.; Russell, T. P.; Hawker, C. Science 1997, 275, 1458-1460. (16) Mansky, P.; Russell, T. P.; Hawker, C. J.; Mays, J.; Cook, D. C.; Satija, S. K. Phys. ReV. Lett. 1997, 79, 237-240. (17) Huang, E.; Russell, T. P.; Harrison, C.; Chaikin, P. M.; Register, R. A.; Hawker, C. J.; Mays, J. Macromolecules 1998, 31, 7641-7650. (18) Ryu, D. Y.; Shin, K.; Drockenmuller, E.; Hawker, C. J.; Russell, T. P. Science 2005, 308, 236-239. (19) Peters, R. D.; Yang, X. M.; Kim, T. K.; Nealey, P. F. Langmuir 2000, 16, 9620-9626. (20) Peters, R. D.; Yang, X. M.; Kim, T. K.; Sohn, B. H.; Nealey, P. F. Langmuir 2000, 16, 4625-4631. (21) Peters, R. D.; Yang, X. M.; Nealey, P. F. Macromolecules 2002, 35, 1822-1834. (22) Smith, A. P.; Sehgal, A.; Douglas, J. F.; Karim, A.; Amis, E. J. Macromol. Rapid Commun. 2003, 24, 131-135.
10.1021/la062594a CCC: $33.50 © 2006 American Chemical Society Published on Web 11/17/2006
Nanoporous Templates from Diblock Copolymer Thin Films
Langmuir, Vol. 22, No. 26, 2006 11093
or UV/ozone.22 Alternatively, the surface energy can be tuned through the formation of an alkyl chlorosilane monolayer with a lowered grafting density. The domain orientation of lamellar PS-b-PMMA can be controlled on silicon substrates that have been functionalized with octadecyl trichlorosilane (OTS) in anhydrous toluene for 12 h at room temperature in a glovebox.19 We demonstrate in this report a simpler and more rapid modification process to control the microdomain orientation of cylindrical PS-b-PMMA, using silicon substrates modified with a partial monolayer of octadecyldimethyl chlorosilane (ODMS), i.e., a layer of ODMS with a grafting density less than the maximum possible monolayer surface coverage. In our approach, neutralized surfaces are obtained by reacting silicon substrates with ODMS in refluxing toluene under ambient conditions for 30 min to 1 h. We further demonstrate that the annealed PSb-PMMA diblock copolymer thin films on partial ODMS SAMs are stable to subsequent processing and can be used to create nanoporous templates.
Scheme 1
Experimental Section
or 450 ( 50 nm for the 147-kDa diblock copolymer. Substrates were then thermally annealed at 165 °C for 2 days in a vacuum oven and imaged in tapping mode in air using a Digital Instruments (Santa Barbara, CA) BioScope atomic force microscope with a Nanoscope IIIa controller. To prepare nanoporous templates from these diblock copolymer thin films, the substrates were exposed to 254-nm UV radiation (10-12.5 min, depending on template type) and sonicated in glacial acetic acid and water for 30 s each, as previously reported.14
Fabrication and Characterization of ODMS Monolayers on Silicon. Silicon substrates were cleaned for >4 h in aqua regia (concentrated HCl/concentrated HNO3, 3:1), rinsed thoroughly with water (purified to 18 MΩ resistivity), and dried under a stream of argon. Substrates were then placed in a conical flask with a ground glass joint connected to a reflux condenser, followed by addition of 0.5%, 1%, or 2% octadecyldimethyl chlorosilane (Sigma-Aldrich, St. Louis, MO) in toluene. Samples were refluxed at 110 °C under ambient atmosphere for various lengths of time, as indicated. Samples were rinsed in toluene, dried under a stream of argon, and cured under ambient atmosphere at room temperature overnight or in a vacuum oven at 140 °C for 5 min. The substrates were again rinsed with toluene while being spun on a spin coater, dried under a stream of argon, and the static contact angle was measured using an NRL contact angle goniometer (Rame-Hart, Inc., Mountain Lake, NJ). ODMSfunctionalized substrates were characterized using a variableangle model LSE ellipsometer (Gaertner Scientific, Skokie, IL). As discussed in the literature,19 we determined the apparent ellipsometric thickness of the ODMS SAM from the difference between the thickness of the oxide layer and the thickness of the ODMS SAM plus the oxide layer, using a refractive index of 1.45 in the calculations of the thicknesses of the ODMS SAM and the oxide. At least three independent measurements were performed per substrate before and after ODMS functionalization. Substrates were imaged in tapping mode in air, using a Digital Instruments (Santa Barbara, CA) BioScope atomic force microscope with a Nanoscope IIIa controller. The local root-meansquare (RMS) surface roughness was determined using height data from four representative 1 µm × 1 µm scan areas through a roughness analysis program included in the Digital Instruments software. Fabrication of PS-b-PMMA Diblock Copolymer Thin Films and Nanoporous PS Templates. The poly(styrene)-poly(methyl methacrylate) diblock copolymers (PS-b-PMMA) used in this study were of two types: 77-kDa molecular weight, with a PS/ PMMA volume ratio of 72:28, and a polydispersity of 1.09 (Polymer Source, Inc., Dorval, Que´bec, Canada), or 147-kDa molecular weight, with a PS/PMMA volume ratio of 77:23, and a polydispersity of 1.06 (gift of the Russel laboratory, University of Massachusetts Polymer Science and Engineering Department, Amherst, MA). Solutions of these diblock copolymers (1 % w/w in toluene) were spin-coated onto the ODMS-modified surfaces to obtain a diblock copolymer thin film with a thickness of approximately 390 ( 40 nm for the 77-kDa diblock copolymer
Results and Discussion For optimal control of interfacial interactions, our goal is to uniformly functionalize the SiOx surface with a low and controllable density of disordered liquid-like alkyl groups without significant island or gel formation. This enables tuning of the surface energy by averaging over a local area smaller than the characteristic domain length of a pore. The mechanism of alkyl chlorosilane monolayer formation on SiOx surfaces has been shown to depend on the solvent, the water content of the solvent, the water absorbed at the surface, the reaction temperature, and the type of silane monomer used, in particular, monochloro versus trichlorosilane.23 Whereas trichlorosilanes such as OTS form a two-dimensional laterally polymerized network with a certain number of covalent attachment points to the underlying SiOx, monochlorosilanes such as ODMS are restricted to direct SiO-Si bond formation with the substrate (Scheme 1A), which results in the formation of less dense monolayers.24 Alkyl chlorosilane monolayer formation can occur through heterogeneous island growth or by homogeneous growth resulting in submonolayer coverage with disordered chains (Scheme 1B). Formation of OTS SAMs from toluene at room temperature under ambient conditions has been shown to occur via island growth, whereas the same process performed under anhydrous, oxygen-free conditions in a glovebox results in uniform monolayer formation.25 The mechanism for alkyl chlorosilane monolayer formation strongly depends on the reaction temperature, and less dense, disordered monolayers are formed above a critical temperature, which, for octadecyl silanes, is approximately 34 °C.26,27 Island (23) Onclin, S.; Ravoo, B. J.; Reinhoudt, D. N. Angew. Chem., Int. Ed. 2005, 44, 6282-6304. (24) Rye, R. R.; Nelson, G. C.; Dugger, M. T. Langmuir 1997, 13, 29652972. (25) Peters, R. D.; Nealey, P. F.; Crain, J. N.; Himpsel, F. J. AdV. Mater. 2003, 15, 221-224. (26) Parikh, A. N.; Allara, D. L.; Azouz, I. B.; Rondelez, F. J. Phys. Chem. 1994, 98, 7577-7590.
11094 Langmuir, Vol. 22, No. 26, 2006
Figure 1. Contact angles of ODMS partial monolayers on silicon substrates modified through refluxing in a solution of 0.5% (solid diamonds), 1% (open squares), or 2% (solid triangles) of ODMS in toluene, as a function of the reaction time. Averages and standard deviations represent at least three different contact angle measurements per sample on at least two separate samples per data point. The contact angle of silicon substrates refluxed in pure toluene for 15 min, but otherwise cleaned, cured, and rinsed analogously to the ODMS-modified substrates, was determined to be 42.0° ( 3.3°.
formation has been observed directly via AFM for partial OTS monolayers generated from CCl4 below the critical temperature, but fairly uniform surfaces are obtained upon incubation above the critical temperature.28 In our studies, alkyl chlorosilane functionalization was performed above this critical temperature, in refluxing toluene, but otherwise under ambient conditions using relatively low concentrations of ODMS and short reaction times. These conditions are expected to yield fairly uniform but low-density submonolayers with disordered alkyl chains. Because of the use of a monochlorosilane, extended cross-linked silane networks cannot be formed. To form a stable monolayer, we cured the substrates, either overnight at room temperature under ambient atmosphere or for 5 min at 140 °C in a vacuum oven. The results obtained using these two curing methods were comparable. Cured substrates were then subjected to a second rinse with toluene prior to further characterization and use. Contact angle measurements and ellipsometric thickness measurements performed prior to this second rinse step resulted in higher and inconsistent values. We further determined that a second rinse is essential for obtaining suitable substrates for subsequent diblock copolymer selfassembly ordering. These observations suggest the presence of chemisorbed and physisorbed monomers at the surface following the curing step and the need for subsequent removal of the physisorbed monomer through the second rinsing step. Using ODMS concentrations between 0.5% and 2% (v/v) in toluene and reaction times from 5 min to 4 h, we were able to control the static water contact angle of ODMS-modified silicon substrates in the range of 60-100° (Figure 1). Low-density ODMS monolayers with intermediate hydrophobicities are formed almost instantaneously. For these early time points, the contact angle can be controlled mainly through the initial ODMS concentration. (27) Rye, R. R. Langmuir 1997, 13, 2588-2590. (28) Sung, M. M.; Carraro, C.; Yauw, O. W.; Kim, Y.; Maboudian, R. J. Phys. Chem. B 2000, 104, 1556-1559.
Niemz et al.
Imaging of these partial ODMS monolayers via tapping mode atomic force microscopy (AFM) revealed a fairly uniform surface, with occasional dots that could be due to contamination but without evidence of pronounced island formation.28 The average RMS surface roughness of these substrates was 0.865 nm. These observations appear to support our hypothesis of uniform surface functionalization, but cannot rule out more subtle surface heterogeneity that cannot be imaged by AFM because of the small size and high mobility of alkyl chains, especially if these chains are not tightly packed. We further characterized the ODMS-functionalized surfaces using ellipsometry. Determining the exact thickness for a partial monolayer is not possible because the refractive index of such a layer is the linear combination of the refractive indices of the pure monolayer and air. Because the relative volume fractions of the two are unknown, the refractive index is also unknown. In addition, the thickness of a complete alkyl chlorosilane monolayer cannot be determined independently from the thickness of the underlying silicon oxide layer, because the refractive indices of the two layers are almost identical. Instead, the thickness of such a monolayer can be obtained by measuring the difference in apparent thickness before and after functionalization, using a refractive index of 1.45 for both the alkyl chlorosilane monolayer and the silicon oxide.19 Because of these complicating factors, we used ellipsometry solely to obtain qualitative information concerning the density of ODMS functionalization. We observed an increase in apparent thickness via ellipsometry that correlated with the contact angle data, supporting the hypothesis that monomer density increases as a function of monomer concentration and reaction time. Thermal annealing of PS-b-PMMA diblock copolymer thin films on ODMS-functionalized silicon substrates results in a range of surface structures depending on the relative surface energies (Figure 2). The microdomain ordering of the diblock copolymer is correlated with the surface energy of the ODMSmodified silicon substrates, assessed through the contact angle. The PMMA cylinders preferentially orient parallel to the surface for unmodified silicon control substrates and substrates modified using low ODMS concentrations and short reaction times (Figure 2a,b). PMMA interacts more favorably than PS with bare SiOx, and for very low-density ODMS SAMs, the PMMA microdomains preferentially interact with the underlying SiOx surface. For ODMS SAMs with contact angles between 65° and 75°, we observed the formation of hexagonally ordered cylindrical PMMA microdomains oriented perpendicular to the substrate upon thermal annealing (Figure 2c,d). This range of contact angles matches the reported neutral surface energy region of silicon substrates functionalized with a neutral-brush random PS-rPMMA copolymer,15 UV/ozone-oxidized octyldimethylchlorosilane SAMs,22 and partial OTS SAMs.19 For ODMS-modified substrates with contact angles above 75°, we observed a loss of proper alignment of the PMMA cylinders and the appearance of increasing numbers of featureless regions (Figure 2e,f), commensurate with symmetric wetting behavior of the PS-b-PMMA diblock copolymer. We were able to obtain nanoporous PS templates from annealed PS-b-PMMA thin films on ODMS-modified silicon substrates, following UV exposure and solvent processing (Figure 3). To remain stable during processing, the PS microdomains of the diblock copolymer thin film must interact with the neutralized substrate sufficiently to hold the film in place. Previous reports of converting diblock copolymer thin films into nanoporous templates were based on either neutral-brush-functionalized silicon oxide surfaces5 or hydrogen-passivated silicon.7 A neutral-
Nanoporous Templates from Diblock Copolymer Thin Films
Figure 2. AFM phase images of thermally annealed PS-b-PMMA diblock copolymer thin films (MW ) 77 kDa) on silicon substrates, functionalized with ODMS under varying reactions conditions, resulting in contact angles of (a) 39°, (b) 61°, (c) 66°, (d) 74°, (e) 86°, and (f) 90°. ODMS concentrations and reaction times were as follows: (a) 0% (control surface, pure toluene), 15 min; (b) 0.5%, 10 min; (c) 0.5%, 30 min; (d) 0.5%, 4 h; (e) 1%, 3 h; (f) 1%, 12 h. Scale bars of all images ) 200 nm.
brush layer has been shown to penetrate into diblock copolymer thin films,15,16 thus anchoring it to the surface during subsequent solvent processing. Hydrogen-passivated silicon surfaces have been shown to undergo [2 + 4] cycloaddition reactions even with low-reactivity aromatic dienes such as benzene.29 It therefore can be expected that hydrogen-passivated Si forms at least some covalent bonds with the poly(styrene) moieties of the diblock copolymer, thus anchoring the thin film to the substrate. The observation that diblock copolymer thin films on partial ODMS withstand the processing into nanoporous templates is nontrivial and suggests that these partial ODMS monolayers interact strongly enough with the PS microdomains to provide sufficient anchoring to the substrate. As described above, partial ODMS SAMs with contact angles between 65° and 75° enable the perpendicular orientation of the cylindrical PMMA microdomains upon thermal annealing of PS-b-PMMA thin films. The annealed diblock copolymer thin film and corresponding nanoporous template shown in Figure 3a and b, respectively, were prepared using a silicon substrate functionalized with 0.5% ODMS for 30 min, resulting in a contact angle of 67°, at the lower end of this range. The nanoporous template displays mostly pores but also some trenches. The (29) Bent, S. F. J. Phys. Chem. B 2002, 106, 2830-2842.
Langmuir, Vol. 22, No. 26, 2006 11095
Figure 3. (a,c,e) AFM phase images of thermally annealed thin films of PS-b-PMMA diblock copolymer with molecular weights of (a,c) 77 and (e) 147 kDa on silicon substrates functionalized with (a) 0.5% ODMS for 30 min, resulting in a contact angle of 67°; (b) 1% ODMS for 60 min, resulting in a contact angle of 72°; and (e) 1% ODMS for 30 min, resulting in a contact angle of 70°. (b,d,f) AFM height images of nanoporous templates obtained from PSb-PMMA diblock copolymer thin films prepared as described for parts a, c, and e, respectively, through UV exposure and solvent processing. Scale bars of all images ) 200 nm.
annealed diblock copolymer thin film and corresponding nanoporous template shown in Figure 3c and d were prepared using a silicon substrate functionalized with 1% ODMS for 60 min, resulting in a contact angle of 72°, at the higher end of this range. The increase in surface hydrophobicity appears to reduce the number of observed trenches. We applied the same approach to induce alignment in thin films of a PS-b-PMMA diblock copolymers of higher molecular weight, enabling the fabrication of nanoporous templates with larger pore diameters and center-to-center spacings (Figure 3e and f). In general, thin films of this 147-kDa diblock copolymer form less ordered structures upon thermal annealing than the thin films of the lower-molecular-weight copolymer, as has been reported for diblock copolymers with molecular weights above 100 kDa.7 For 147-kDa diblock copolymer thin films, we observed a comparable level of ordering on substrates neutralized using either the ODMS or the neutral-brush approach.
Conclusions We have demonstrated that alkyl chloroslilane modification of silicon substrates can be used to control the microdomain orientation of cylinidrical PS-b-PMMA diblock copolymer thin
11096 Langmuir, Vol. 22, No. 26, 2006
films on silicon substrates and that these diblock copolymer thin films can be converted into PS nanoporous templates without loss of the film upon solvent processing. The use of a monochlorinated alkyl silane and modification at high temperature using short reaction times with relatively low silane concentrations provides uniform submonolayers on the length scale of the diblock copolymer microdomains to allow for the alignment of perpendicular PMMA cylinders of very small lateral dimension. Compared to surface neutralization using the neutral-brush approach, neutralizing silicon substrates through ODMS submonolayers is straightforward and less time consuming, and it uses inexpensive, commercially available materials. Compared to surface neutralization via hydrogen passivation of silicon, this approach enables more precise control of surface interactions to obtain optimally ordered and oriented templates. Appropriate surface neutralization can be obtained using partial ODMS SAMs
Niemz et al.
with contact angles between 65° and 75°. These partial SAMs can be prepared readily and reproducibly using ODMS concentrations between 0.5% and 1% and reaction times between 30 min and 1 h. The use of a thin and diffuse coating is advantageous for future chemical modification of the templates, which requires complete removal of organic material from the pores. The application of these nanoporous surfaces as templates for the immobilization of biomolecules and biofunctionalized nanoparticles is currently under investigation. Acknowledgment. This work was supported by the National Science Foundation through Research Grants BES-0304675, ECS-0501629, DMR-0213695, and DMR-0109077 and through REU Site Grants EEC-0243910 and CHE-0353662, in addition to research funds from the Keck Graduate Institute. LA062594A
