Langmuir 2006, 22, 8071-8077
8071
Template-Directed Adsorption of Block Copolymers on Alkanethiol-Patterned Gold Surfaces Amol Chandekar,† Sandip K. Sengupta,† Carol M. F. Barry,‡ Joey L. Mead,‡ and James E. Whitten*,† Department of Chemistry and Center for High-Rate Nanomanufacturing, and Department of Plastics Engineering and Center for High-Rate Nanomanufacturing, The UniVersity of Massachusetts Lowell, Lowell, Massachusetts 01854-5047 ReceiVed February 21, 2006. In Final Form: June 9, 2006 Functionalized alkanethiols have been self-assembled on gold to modify the wetting properties of the surface and promote or hinder the adsorption of block copolymers containing both hydrophobic and hydrophilic blocks. X-ray photoelectron spectroscopy (XPS) studies of spin-coated polyethylene-block-poly(ethylene oxide) (PE-b-PEO) copolymers on 16-mercaptohexadecanoic acid (MHDA)-, octadecanethiol (ODT)-, and 1H,1H,2H,2H-perfluorodecanethiol (PFDT)-covered surfaces have been performed. In the case of an 80 wt % PEO block copolymer, spin-coating on a gold surface precovered with MHDA results in a polymer film thick enough to completely attenuate Au 4f photoelectrons; spin-coating on the more hydrophobic ODT and PFDT monolayers leads to significantly thinner polymer films and incomplete attenuation of the gold photoelectrons. The opposite results are observed when a 20 wt % PEO block copolymer is used. Angle-resolved XPS studies of the 80 wt % PEO block copolymer spin-coated onto an MHDA-covered surface indicate that the PE blocks of the polymer segregate to the near-surface region, oriented away from the hydrophilic carboxylic acid tails of the monolayers; the surface concentration of PE is further enhanced by annealing at 90 °C. Microcontact printing and dip-pen nanolithography have been used to pattern gold surfaces with MHDA, and the surfaces have been backfilled with ODT or PFDT, such that the unpatterned regions of the surface are covered with hydrophobic monolayers. In the case of backfilling with PFDT, spin-coating the 80 wt % PEO copolymer onto these patterned surfaces and subsequent annealing results in the block copolymer preferentially adsorbing on the MHDA-covered regions and forming well-defined patterns that mimic the MHDA pattern, as determined by scanning electron microscopy and atomic force microscopy. Significantly worse patterning, characterized by micron-sized polymer droplets, results when the surface is backfilled with ODT instead of PFDT. Using PFDT and MHDA, polymer features having widths as small as 500 nm have been formed. These studies demonstrate a novel method to pattern block copolymers with nanoscale resolution.
Introduction As nanotechnology matures, it will be necessary in some applications to form nanoscale polymeric structures on surfaces. This may be of particular importance for future nanomanufacturing applications in which polymeric patterns could be used as flexible templates or molds to lay down or imprint nanoparticles or nanodevices.1-7 Various methodologies have been developed to create polymeric patterns with submicron dimensions. These include direct writing of polymer patterns via laser ablation,8 soft-lithographic methods including the cohesive mechanical failure technique9 and use of photo- or electrochemically polymer* Corresponding author. Address: Department of Chemistry, The University of Massachusetts Lowell, One University Avenue, Lowell, MA 01854. Phone: (978) 934-3666. Fax: (978) 934-3013. E-mail:
[email protected]. † Department of Chemistry and Center for High-Rate Nanomanufacturing. ‡ Department of Plastics Engineering and Center for High-Rate Nanomanufacturing. (1) Yin, Y.; Xia, Y. J. Am. Chem. Soc. 2003, 125, 2048-2049. (2) Xia, Y.; McClelland, J. J.; Gupta, R.; Qin, D.; Zhao, X.-M.; Sohn, L. L.; Celotta, R. J.; Whitesides, G. M. AdV. Mater. 1995, 9, 147-149. (3) McClelland, G. M.; Hart, M. W.; Rettner, C. T.; Best, M. E.; Carter, K. R.; Terris, B. D. Appl. Phys. Lett. 2002, 81, 1483-1485. (4) Kim, Y. S.; Lee, N. Y.; Lim, J. R.; Lee, M. J.; Park, S. Chem. Mater. 2005, 17, 5867-5870. (5) Maury, P.; Escalante, M.; Reinhoudt, D. N.; Huskens, J. AdV. Mater. 2005, 17, 2718-2723. (6) Tokuhisa, H.; Hammond, P. T. Langmuir 2004, 20, 1436-1441. (7) Stoykovich, M. P.; Mu¨ller, M.; Kim, S. O.; Solak, H. H.; Edwards, E. W.; de Pablo, J. J.; Nealey, P. F. Science 2005, 308, 1442-1446. (8) Aguilar, C. A.; Lu, Y.; Mao, S.; Chen, S. Biomaterials 2005, 26, 76427649. (9) Childs, W. R.; Nuzzo, R. G. J. Am. Chem. Soc. 2002, 124, 13583-13596.
izable precursors,10,11 carbon dioxide-assisted molding of polymer films,12 electric field-induced patterning of block copolymers,13 and the use of probe microscopies to pattern polymer surfaces via localized Joule heating,14 thermal-mechanical effects,15 or applied electric fields.16 Meyer and Braun17 used microcontact printing to pattern 11mercaptoundecanoic acid onto gold surfaces, and these authors demonstrated that polystyrene selectively dewetted the patterned regions. The phase separation of polymer blends and block copolymers due to preferential wetting and dewetting of patterned surfaces has also been used to form aligned and lamellar nanoscale polymeric structures. For example, Sibener and colleagues18 demonstrated that a substrate may be lithographically patterned with channels that direct the alignment of diblock copolymers. Nealey and co-workers19 patterned a silicon oxide surface with polymer brushes and used it as a template for block copolymers (10) Behl, M.; Seekamp, J.; Zankovych, S.; Torres, C. M. S.; Zentel, R.; Ahopelto, J. AdV. Mater. 2002, 14, 588-591. (11) Jegadesan, S.; Sindhu, S.; Advincula, R. C.; Valiyaveettil, S. Langmuir 2006, 22, 780-786. (12) Wang, Y.; Liu, Z.; Han, B.; Huang, Y.; Zhang, J.; Sun, D.; Du, J. J. Phys. Chem. B 2005, 109, 12376-12379. (13) Xiang, H.; Lin, Y.; Russell, T. P. Macromolecules 2004, 37, 5358-5368. (14) Lyuksyutov, S. F.; Vaia, R. A.; Paramonov, P. B.; Juhl, S.; Waterhouse, L.; Ralich, R. M.; Sigalov, G.; Sancaktar, E. Nat. Mater. 2003, 2, 468-472. (15) Mamin, H. J.; Rugar, D. Appl. Phys. Lett. 1992, 61, 1003-1005. (16) Lyuksyutov, S. F.; Paramonov, P. B.; Juhl, S.; Vaia, R. A. Appl. Phys. Lett. 2003, 83, 4405-4407. (17) Meyer, E.; Braun, H.-G. Macromol. Mater. Eng. 2000, 276/277, 44-50. (18) Sundrani, D.; Darling, S. B.; Sibener, S. J. Nano Lett. 2004, 4, 273-276. (19) Edwards, E. W.; Montague, M. F.; Solak, H. H.; Hawker, C. J.; Nealey, P. F. AdV. Mater. 2004, 16, 1315-1319.
10.1021/la0605034 CCC: $33.50 © 2006 American Chemical Society Published on Web 08/08/2006
8072 Langmuir, Vol. 22, No. 19, 2006
and, in a separate study,7 used the spin-coating of ternary blends of diblock copolymers and homopolymers on nanopatterned substrates to form nonregular structures, including angled lamellae having essentially perfect long-range order. A growing body of literature (see, for example, ref 20) exists on the types of morphologies that may be achieved when immiscible block copolymers self-assemble on a surface. Fluorocarbon-coated surfaces offer intriguing possibilities to affect polymer adsorption since they typically are wetted less than hydrocarbons, not only by water but also by oils.21 Lee and colleagues22,23 studied the self-assembly and wetting properties of a variety of CF3-terminated alkanethiols adsorbed on gold and compared them to methyl-terminated ones. These authors found that the terminally fluorinated monolayers exhibited lower surface energies than similar methyl-terminated ones and higher coefficients of friction in atomic force microscopy (AFM) studies. Surprisingly, it was also found that polar solvents wetted the CF3-terminated films better than the CH3-terminated ones.22 In the present study, the spin-coating of block copolymers onto gold substrates whose wetting properties have been modified by functionalized alkanethiols, including CF3-terminated ones, has been investigated. Diblock copolymers were chosen to be able to change the overall hydrophilic/hydrophobic nature of the polymer by changing the molar ratio of the blocks. Microcontact printing and dip-pen nanolithography (DPN) have been used to pattern a surface with the functionalized alkanethiols, and it is demonstrated that the wetting properties imparted by the adsorbed molecules lead to selective adsorption of hydrophilic/hydrophobic block copolymers spin-coated onto the patterned substrates, with the polymer selectively depositing on the region of the surface having similar hydrophilicity. In the case of spin-coating a mainly hydrophilic polyethylene-block-poly(ethylene oxide) (PE-b-PEO) copolymer onto carboxylic acid-terminated alkanethiol patterns, dramatic differences in the morphology of the patterned polymers are observed when the hydrophilic regions are surrounded by alkanethiols terminated in CF3 compared to CH3 groups, and possible reasons for this are discussed. Experimental Section Substrates. Gold substrates were prepared by thermally depositing ∼100 Å of titanium and then 2000 Å of gold onto ∼1 × 1 cm polished Si(111) wafers that had been ultrasonicated in methanol and acetone prior to installation in the metal deposition chamber. The vacuum during metal deposition was ∼1 × 10-6 Torr. Titanium was deposited by resistively heating a 2.0 mm diam 99.99% titanium wire, and gold was deposited by heating a 0.25 mm diam tungsten filament (99.95%) wrapped with 0.2 mm diam 99.9% gold wire. The Si(111) wafers were nominally at room temperature during metal deposition. AFM indicated that the typical crystallite size was 20 nm. X-ray diffraction showed the deposited gold films to be preferentially oriented in the 〈111〉 direction. The substrates were used immediately after removal from the vacuum chamber or were stored in ethanol until used. Alkanethiol Adsorption and Microcontact Printing. Three different alkanethiols, whose chemical structures are shown in Figure 1, were used to affect the wetting properties of the gold-coated substrates. 16-Mercaptohexadecanoic acid (MHDA) and octadecanethiol (ODT) were purchased from Aldrich, and 1H,1H,2H,2H-perfluorodecanethiol (PFDT) was purchased from Oakwood Products. (20) Hawker, C. J.; Russell, T. P. MRS Bull. 2005, 30, 952-966. (21) Garbassi, F.; Morra, M.; Occhiello, E. Polymer Surfaces: From Physics to Technology; John Wiley & Sons: Chichester, U.K., 1994; p 327. (22) Miura, Y. F.; Takenaga, M.; Koini, T.; Graupe, M.; Garg, N.; Graham, R. L., Jr.; Lee, T. R. Langmuir 1998, 14, 5821-5825. (23) Houston, J. E.; Doelling, C. M.; Vanderlick, T. K.; Hu, Y.; Scoles, G.; Wenzl, I.; Lee, T. R. Langmuir 2005, 21, 3926-3932.
Chandekar et al.
Figure 1. Chemical structures of ODT, MHDA, PFDT, and the PE-b-PEO copolymer. Microcontact printing24 was used to pattern gold substrates with these molecules using the following steps. A poly(dimethylsiloxane) (PDMS) stamp was prepared from a master containing features in the micron range. The stamp was prepared by cast-molding using a commercially available PDMS elastomer known as Sylgard 184 (Dow Corning). This is supplied as a two-part kit: a liquid base (i.e., a vinyl-terminated PDMS) and a curing agent. The two parts were mixed in a 10:1 ratio of base-to-curing agent, and the mixture was degassed in a vacuum to remove any solvent. The clear solution was then poured over the master, which had been silanized using a silanizing agent, and heated to about 140 °C for 15 min. This solidified the mixture, and the stamp was carefully pealed from the master. The PDMS stamp was “inked” with a particular 1 mM ethanolic alkanethiol solution and dried in a stream of nitrogen gas. This procedure was repeated twice. In the case of MHDA, the stamp was plasma treated in oxygen to make its surface more hydrophilic (in order for the MHDA to wet it well) prior to inking. Once inked, the stamp was brought into contact with the gold surface, with a little applied pressure, and separated carefully after 60 s. In some cases it was desirable to “backfill” the surface with an alkanethiol different from the one patterned by microcontact printing. This was achieved by putting several drops of the second alkanethiol solution (1 mM concentration) onto the substrate, tilting it such that the drops flowed over the entire surface, and then drying it in a stream of nitrogen gas. Polymer Adsorption. PE-b-PEO copolymers (Figure 1) were chosen because of their characteristic of having two blocks with very different degrees of hydrophobicity (i.e., PE is hydrophobic, and PEO is hydrophilic). Two variations of this copolymer were investigated: a 20:80 wt % PE/PEO copolymer (29:71 mol %) with an average molecular weight of 2250 (Aldrich, Catalog No. 525901), and an 80:20 wt % PE/PEO copolymer (87:13 mol %) having an average molecular weight of 575 (Aldrich, Catalog No. 459003). The 80 wt % PEO copolymer was spin-coated onto the alkanethiol(24) Wilbur, J. L.; Kumar, A.; Kim, E.; Whitesides, G. M. AdV. Mater. 1994, 6, 600-604.
Adsorption of PE-b-PEO on Alkanethiol-CoVered Au covered gold surfaces using chloroform solutions, and the 80 wt % PE copolymer was spin-coated using warm toluene. Different polymer concentrations were explored. In some cases, the polymer films were annealed at 90 °C for 150 min, as will be discussed. Contact-Angle Measurements. The wetting properties of the unpatterned alkanethiol-covered surfaces were confirmed for the different alkanethiols. These measurements were made using the Wilhelmy plate method with a KSV model Sigma 70 dynamic contact angle wetting balance. For these measurements, the alkanethiol was self-assembled on both sides of a Si(111) wafer (both sides polished) that had been coated with gold, as previously described. X-ray Photoelectron Spectroscopy (XPS). Surface analysis was performed with XPS using a Vacuum Generators ESCALAB MKII instrument with a base pressure less than 1 × 10-9 Torr. Al KR or Mg KR X-rays (1486.6 and 1253.6 eV, respectively) were used, and the photoelectrons were energy-analyzed with a concentric hemispherical analyzer (CHA) in fixed analyzer transmission mode using a pass energy of 20 eV. Unless otherwise specified, a takeoff angle (TOA; i.e., the angle between the normal to the entrance of the CHA and the plane of the sample) of 90° was used. Smaller TOAs are more surface-sensitive, and, for some experiments, 45° and 15° TOAs were used to obtain information related to surface segregation of the block copolymers. The XPS measurements interrogated a fairly large area of the sample (∼0.25 cm2). The surfaces of the gold-coated Si(111) substrates were electrically grounded by using vacuum-compatible silver paint to attach the substrates to sample stubs and to electrically connect their surfaces to the stubs, which were held at ground during the measurements. This procedure eliminated charging effects for very thin organic films (e.g., the adsorbed alkanethiols), but did not completely do so for thick polymer layers, as will be discussed. Dip-Pen Lithography. DPN was performed using a NanoInk instrument. Lines were written by inking the tip with 4 mM MHDA in acetonitrile using the “double-dip” coating procedure. This consisted of dipping the AFM tip in the MHDA solution, holding the tip in steam for 5 min, letting it dry for 5 min, and then dipping it in the solution again. Parallel MHDA lines of width 500 nm, length 80 µm, and spacing 3 µm were then written in contact mode on the gold substrate (purchased from NanoInk, Inc.). DPN lithography was carried out at room temperature, and the humidity (30-40%) was controlled using the environmental chamber supplied with the instrument. Microscopy. Imaging of the polymer patterns was performed using an Amray scanning electron microscope (SEM). Electron beam energies of either 20 or 10 keV were used, with typical beam currents of 100 and 50 µA. AFM of the polymers was carried out with a PSIA XE-100 instrument using silicon nitride tips with radii nominally less than 10 nm.
Results and Discussion Figure 2 shows C 1s XPS spectra of the three different selfassembled monolayer (SAM) alkanethiols on gold surfaces. In the case of ODT, the C 1s spectrum consists of a single peak with maximum intensity at a binding energy of 284.4 eV arising from methylene and methyl group carbon atoms. For MHDA, a second peak is present at ∼289.0 eV due to the carbon of the carboxylic acid group. The experimentally measured ratio of the areas of the peaks is 17:1, in reasonable agreement with the theoretically expected value of 15:1. The spectrum due to PFDT contains three distinct carbon peaks. The one at 284.0 eV is due to the two carbon atoms attached to hydrogens, that at 290.1 eV is due to the seven carbon atoms each attached to two fluorines, and the one at 292.3 eV is due to the one carbon attached to three fluorine atoms. The experimentally measured ratios of the areas of the three peaks are 3.8:5.6:1, respectively. The relatively poor agreement with the expected ratios of 2:7:1 is most likely due to either adventitious carbon contamination on the surface that occurs at a binding energy similar to that of the methylene carbon species or (more likely) X-ray-induced C-F bond breaking, which
Langmuir, Vol. 22, No. 19, 2006 8073
Figure 2. Al KR XPS of the C 1s region of monolayers of ODT, MHDA, and PFDT self-assembled on gold surfaces. The spectra have been offset vertically for ease of viewing, and the binding energy scale is referenced to the spectroscopic Fermi level. Table 1. Properties of Alkanethiol-Covered Gold Surfaces and These Surfaces after Spin-Coating PE-b-PEO Block Copolymers onto Them (Without Annealing)
alkanethiol
θ (adv) water
θ (rec) water
C/Au atomic ratio for 80/20 PEO/PE blocka
C/Au atomic ratio for 20/80 PEO/PE blockb
ODT PFDT MHDA
113 ( 1 109 ( 3 74 ( 3
97 ( 3 85 ( 1 40 ( 2
0.92 12 infinite
infinite 49 19
a The polymer was spin-coated using a 0.5 wt % solution in chloroform. The polymer was spin-coated using a 0.4 wt % solution in toluene (warm solution).
b
is known to occur facilely for monolayers of this molecule.25 Note that cleavage of C-F bonds (and fluorine desorption) would cause CF3 carbons to resemble CF2 and CH2 carbons and the CF2 carbons to resemble CH2 ones. The small peak at ∼287.3 eV (presumably due to carbon attached to a single fluorine) is indicative of some beam-induced decomposition, consistent with the results in ref 25. Dynamic contact angle wetting balance measurements were performed to confirm the differences in hydrophobicity of these three monolayer-covered surfaces, and water contact angle results are included in Table 1. The values for ODT, MHDA, and PFDT are generally consistent with reported values,26-28 with the following exceptions. First, variability exists in the literature values for carboxylic acid-terminated alkanethiol SAMs, with water contact angles smaller than 10° reported for MHDA.29 Possible reasons for this variability have been discussed and include partial multilayer formation due to hydrogen bonding30 as well as contamination and defects. In the case of PFDT, the advancing and receding water contact angles (109° and 85°, (25) Wagner, A. J.; Carlo, S. R.; Vecitis, C.; Fairbrother, D. H. Langmuir 2002, 18, 1542-1549. (26) Laibinis, P. E.; Whitesides, G. M.; Allara, D. L.; Tao, Y.-T.; Parikh, A. N.; Nuzzo, R. G. J. Am. Chem. Soc. 1991, 113, 7152-7167. (27) Wang, M. S.; Palmer, L. B.; Schwartz, J. D.; Razatos, A. Langmuir 2004, 20, 7753-7759. (28) van de Grampel, R. D.; Ming, W.; Gildenpfennig, A.; Laven, J.; Brongersma, H. H.; de With, G.; van der Linde, R. Langmuir 2004, 20, 145-149. (29) Bain, C. D.; Troughton, E. B.; Tao, Y.-T.; Evall, J.; Whitesides, G. M.; Nuzzo, R. G. J. Am. Chem. Soc. 1989, 111, 321-335. (30) Wang, H.; Chen, S.; Li, L.; Jiang, S. Langmuir 2005, 21, 2633-2636.
8074 Langmuir, Vol. 22, No. 19, 2006
Chandekar et al.
respectively) are smaller than those reported for this SAM in ref 28 (117.2° and 104.8°, respectively) and for similar fluorinated monolayers.31 In the present work, the smaller and larger contact angles of the PFDT and MHDA SAMs, respectively, and the larger hystereses are most likely due to the relatively small size of the gold crystallites and/or monolayers having a large number of defects. Both of these effects lead to increased surface roughness and greater hysteresis,32 with the magnitude of the hysteresis related to the difference in surface free energy barriers for rearrangement of the molecules when the sample is entering or withdrawing from the liquid. Formation of the ODT monolayer seems more forgiving with respect to defects, as evidenced by better agreement between the values in Table 1 and the advancing and receding water contact angle literature values of 115° and 105°, respectively.26 Note that the data in Table 1 were found to be reproducible for multiple samples prepared in our laboratory. The contact angle results therefore suggest that surface roughness (induced by small gold grain sizes) is at least partially responsible for the relatively large number of defects, particularly for the PFDT sample. Surface contamination could possibly be a contributing factor, but the XPS results discussed earlier suggest that this is minimal. A larger number of defects would be expected to lower the SAM packing density. To obtain information about the relative packing densities of the three monolayers, Mg KR XPS was performed in which the S 2p/Au 4f area ratios were measured. This should provide an approximation of the relative packing densities.33 Of course, it is realized that the composition, length, and tilt angle of the alkanethiol chains will (to a small extent) affect this ratio since the kinetic energies of the S 2p and Au 4f photoelectrons are different and have slightly different escape depths that depend on these factors. The measured ratios for ODT, MHDA, and PFDT were 0.099, 0.071, and 0.069, respectively. Note that these ratios have taken into account the relative sensitivity factors of the two peaks. These results indicate that MHDA and PFDT have smaller packing densities than the ODT film. This is consistent with the water contact angle results discussed above. To investigate how the hydrophilic/hydrophobic nature of the surfaces affects polymer adsorption, the following studies were performed. Various concentrations of the more hydrophilic (20: 80 wt %) PE-b-PEO copolymer were spin-coated onto gold surfaces covered by the three SAMs to determine conditions that would preferentially lead to coverage by the polymer on the hydrophilic MHDA monolayer but to minimal coverage of the PFDT- or ODT-covered surfaces. Figure 3a,b shows XPS of the Au 4f and C 1s regions of three similarly prepared monolayers of MHDA, PFDT, and ODT following identical spin-coating conditions of 0.5 wt % of the polymer in chloroform at 3000 rpm for 30 s. Note that no attempts have been made to correct for shifts in these peaks due to surface charging during analyses. Seah and Dench34 determined the following empirical formula for attenuation length as a function of kinetic energy (E) for organic compounds:
λ (nm) ) 49/E2 + 0.11E1/2
(1)
Using this equation, the attenuation length of Au 4f photoelectrons (with a kinetic energy of ∼1400 eV) ejected by Al KR X-rays is estimated to be 41 Å. In XPS, the detection depth depends not (31) Fukushima, H.; Seki, S.; Nishikawa, T.; Takiguchi, H.; Tamada, K.; Abe, K.; Colorado, R., Jr.; Graupe, M.; Shmakova, O. E.; Lee, T. R. J. Phys. Chem. B 2000, 104, 7417-7423. (32) Decker, E. L.; Garoff, S. Langmuir 1997, 13, 6321-6332. (33) Park, J.-S.; Vo, A. N.; Barriet, D.; Shon, Y.-S.; Lee, R. T. Langmuir 2005, 21, 2902-2911. (34) Seah, M. P.; Dench, W. A. Surf. Interface Anal. 1979, 1, 2-11.
Figure 3. Al KR XPS of the (a) C 1s and (b) Au 4f regions of MHDA-, PFDT-, and ODT-covered gold surfaces that have been spin-coated with the 80 wt % PEO block copolymer. Complete attenuation of the MHDA-covered surface indicates that the polymer film is at least 120 Å thick in this case, as discussed in the text. Note that the binding energy scales are referenced to the spectroscopic Fermi level, but no corrections for surface charging have been made.
only on the kinetic energy of the photoelectrons, but also on the TOA, with the analysis depth given by 3λ cos θ, where λ is the attenuation length, and θ is the complement of the TOA.35 The analysis depth is therefore approximately 3λ for a 90° TOA, as for the data in Figure 3. The complete attenuation of the Au 4f electrons from the MHDA sample indicates that this surface has been covered by at least 120 Å of polymer. On the other hand, the hydrophobic ODT and PFDT surfaces still show strong Au 4f gold signals and, therefore, are covered by polymer films significantly thinner than 120 Å. Figure 4 shows C 1s data for the block copolymer-covered MHDA surface at 90°, 45°, and 15° TOAs. This is a different set of data than that in Figure 3, and was obtained using Mg KR X-rays. The polymer film was somewhat thinner and exhibited less surface charging during XPS. In the case of the 90° TOA, the carbon peak has its greatest intensity at 287.0 eV with a low (35) Watts, J. F.; Wolstenholme, J. An Introduction to Surface Analysis by XPS and AES; John Wiley & Sons: Chichester, U.K., 2003; p 80.
Adsorption of PE-b-PEO on Alkanethiol-CoVered Au
Langmuir, Vol. 22, No. 19, 2006 8075
Figure 5. Simplified cartoon of how the PE-b-PEO copolymer may adsorb onto an MHDA-covered gold surface before and after annealing. In this illustration, the MHDA layer on the gold surface is not shown, and the PE and PEO blocks are represented by dark and light lines, respectively. While the PE and PEO segments are drawn approximately equal in length, the PEO blocks are actually longer in the case of the 80 wt % PEO copolymer used for the segregation studies.
binding energy shoulder. Deconvolution of the spectrum is included in the figure and yields two peaks with maximum intensity at 287.05 eV (full width at half-maximum (fwhm) of 1.31 eV) and 285.75 eV (fwhm of 1.31 eV). The higher binding energy component arises from the carbon atoms in the ethylene oxide block, which are bonded to electronegative oxygen atoms, and the lower binding energy component is due to the carbon atoms in the PE block. The observation of a higher binding energy of ∼1.3 eV for the PEO peak, compared to that of PE, is in reasonable agreement with previous work comparing the C 1s spectra of these two polymers.36 Area analysis of the deconvoluted peaks indicates that the area of the peak due to PEO is 2.6 times greater than that due to PE. This is in satisfactory agreement with the expected value for this block copolymer, since PEO is 2.4 times more abundant (on a molar basis) than PE. For the more surface-sensitive 45° and 15° TOAs in Figure 4, the lower binding energy peak due to PE is significantly enhanced. In the case of the 45° TOA, deconvolution yields two peaks at 287.24 eV (fwhm of 1.83 eV) and 285.75 eV (fwhm of 1.18 eV) due to PEO and PE carbons, respectively. For the 15° TOA, deconvolution yields peaks at 287.31 eV (fwhm of 1.33 eV) and 285.84 eV (fwhm of 1.37 eV). For this lowest TOA, integration shows that the PE carbon peak has an area 1.2 times that of the PEO one. The increase in the relative area of the PE peak compared to the PEO one for the smaller TOAs indicates that the PE blocks are closer to the surface than are the PEO blocks. For a 15° TOA, the analysis depth is approximately 0.26 λ. For C 1s photoelectrons ejected by Mg KR X-rays, this corresponds to ∼9 Å. The 80 wt % PEO copolymer contains (on average) 18 PE units and 44 PEO units. Assuming a C-C bond distance of 1.54 Å, the fully extended PE block would be ∼54 Å long. The fact that a substantial PEO signal is still detected (even for a 15° TOA) suggests that the PE is not fully extended. As a whole, therefore, the results of Figures 3 and 4 demonstrate
that the 80 wt % PE-b-PEO copolymer wets the MHDA monolayer with the PEO blocks in contact with the hydrophilic surface and the PE blocks on the top and collapsed on themselves. Also included in Figure 4 is the C 1s spectrum (using a 90° TOA) obtained after annealing the polymer-covered MHDA surface for 150 min at 90 °C. Comparison to the corresponding unannealed spectrum shows that the peak due to PE increases in intensity as a result of annealing. These results indicate that annealing of the spin-coated 80 wt % PEO block copolymer film causes surface segregation of the PE blocks. While the situation is complex, a simplified illustration of the type of changes that may occur due to annealing is shown in Figure 5. Enhancement of PE concentration closer to the top of the polymer film is driven by these blocks, staying away (as much as possible) from the PEO, which wets the hydrophilic MHDA-covered surface. Thermodynamically, this is driven by PE-PE cohesive forces being stronger than PE-PEO adhesive ones. Table 1 summarizes the C 1s/Au 4f XPS area ratios, corrected by appropriate sensitivity factors. It is interesting that a thicker polymer layer forms on the PFDT-covered surface compared to the ODT one. One possible (naı¨ve) explanation of this is that the PFDT SAM has a strong dipole that can interact with the PEO blocks. However, as discussed by Lee and colleagues,37 the influence of the dipole on wettability is related to how deeply buried the fluorocarbon-hydrocarbon junction is within the monolayer. In the case of PFDT, this junction is deeply buried, and the contribution of the dipole is expected to be minimal, in contrast to what would be expected if a HS(CH2)nCF3 monolayer were used. The most likely explanation of the different behavior exhibited by the ODT- and PFDT-covered surfaces is the greater defect density of the PFDT films, which increases the surface free energy of the SAM and leads to enhanced attractive forces between it and the copolymer. Table 1 also includes results from similar experiments in which the more hydrophobic (80:20 wt %) PE-b-PEO copolymer were spin-coated onto the SAM-covered surfaces. In this case, very little polymer adsorption occurs on the MHDA-covered surface, while substantially more polymer adsorption occurs on the PFDTand ODT-covered surfaces. Contrary to the experiments using the more hydrophilic copolymer, in this case, a thicker polymer film occurs on the ODT-covered surface compared to the PFDT one. Again, this is most likely due to greater dewetting of the more hydrophobic polymer by the PFDT monolayer, which has a lower packing density than the ODT SAM. Patterning with the block copolymer was accomplished by spin-coating the 80 wt % PEO block copolymer onto a microcontact-printed surface consisting of approximately parallel lines of a few microns in width. The polymer-covered surface was annealed at 90 °C for 150 min prior to obtaining the images.
(36) Kiss, E.; Samu, J.; To´th, A.; Berto´ti, I. Langmuir 1996, 12, 1651-1657.
(37) Colorado, R., Jr.; Lee, T. R. J. Phys. Org. Chem. 2000, 13, 796-807.
Figure 4. Mg KR angle-resolved XPS of the C 1s region of an MHDA-covered sample (similar to that in Figure 3) using 90°, 45°, and 15° TOAs. Deconvoluted peaks are also included (dashed lines). All spectra contain two peaks due to carbon atoms in PEO and PE, but the greater relative areas of the lower binding energy PE peaks for the smaller TOA spectra indicate that the PE blocks are surfacesegregated. Also included is a 90° TOA spectrum for the same sample after annealing for 150 min at 90 °C.
8076 Langmuir, Vol. 22, No. 19, 2006
Chandekar et al.
Figure 6. SEM image of PE-b-PEO lines (darker regions) that form following spin-coating with an 80 wt % PEO copolymer of a microcontact-printed pattern of MHDA backfilled with PFDT. The sample was annealed for 150 min at 90 °C.
Figure 8. AFM image of PE-b-PEO lines (brighter regions) that form following spin-coating with an 80 wt % PEO copolymer of a gold sample patterned (using DPN) with MHDA lines 500 nm in width and backfilled with PFDT. The sample was annealed for 150 min at 90 °C.
Figure 7. AFM image of the same sample depicted in Figure 6; however, the AFM image is of a different region of the surface containing slightly different line widths. A cross-section is included to show the thickness of the polymer film on the MHDA-covered regions.
No pattern was observed without this annealing step, which is necessary for the blocks to have enough diffusivity to reach regions of the surface where they are most thermodynamically stable. Figure 6 shows an SEM image of the block copolymer spin-coated onto a gold surface that had been microcontact printed with MHDA and then backfilled with PFDT. The dark regions represent thick copolymer films adsorbed on top of the MHDAcovered regions of the surface; the lighter regions are the PFDTcovered areas. Figure 7 depicts AFM results for the same polymercovered surface shown in Figure 6, although not necessarily the same exact region imaged by SEM. As demonstrated by the AFM cross-section data, the block copolymer is approximately 400 nm thick on the MHDA-covered regions of the surface. Figure 8 shows an AFM image resulting from a DPN experiment in which parallel MHDA lines (each ∼500 nm wide) were “written” onto a gold surface. The surface was then backfilled with PFDT, as discussed earlier, and annealed prior to AFM
imaging. The results essentially reproduce what was observed on the larger scale with microcontact printing, except that the thickness of the polymer films on the MHDA lines is significantly smaller (40-50 nm versus 400 nm). The thinner polymer films in the case of narrower lines results from the inherent instability of tall, thin polymer structures due to the polymer’s minimization of surface area and surface free energy. The 80 wt % PEO block copolymer was also spin-coated onto surfaces patterned with MHDA and backfilled with ODT prior to annealing. It was expected that results very similar to those for the PFDT-backfilled surface would be observed. Instead, it was found that the copolymer did not form smooth films on the MHDA regions of the surface. Instead, the polymer was observed to form droplets that preferentially stick to the MHDA regions. This is demonstrated by the optical microscope and AFM images in Figure 9. The reasons for this behavior are not entirely clear but likely result from the greater hydrophobic nature of the ODTcovered regions compared to PFDT, as previously discussed. When the polymer lands on the (mostly) hydrophobic surface during spin-coating, it collapses on itself to encircle its hydrophilic segment. But, since the hydrophobic segment of the polymer is smaller than the hydrophilic one, with each polymer chain consisting, on average, of 44 PEO units compared to 18 PE units, it is not able to encompass completely all of the hydrophilic PEO segments. Some hydrophilic portions of the polymer are still exposed and, after annealing, attach to the MHDA-covered regions of the surface. In the case of the PFDT surface, these types of completely collapsed polymer structures never form, and smoother patterns are obtained. In the present study, rather short diblock copolymers (i.e., oligomers) were used. These were chosen because they were conveniently commercially available and because it is expected that the chains would have greater diffusivity. However, it is likely that higher molecular weight copolymers could also be patterned by this methodology. This is indicated by a recent
Adsorption of PE-b-PEO on Alkanethiol-CoVered Au
Langmuir, Vol. 22, No. 19, 2006 8077
Figure 9. (a) Optical microscopy of PE-b-PEO droplets that form following spin-coating (with an 80 wt % PEO copolymer) of a microcontactprinted pattern of MHDA backfilled with ODT. The sample was annealed for 150 min at 90 °C. This sample was patterned with alternating vertical lines of MHDA and ODT. For example, the regions marked “A” were patterned with MHDA, and the one marked “B” is a hydrophobic backfilled region. The pattern repeats throughout the area shown in the image, with the droplets preferentially deposited in the MHDA-covered regions. (b) AFM image of the polymer droplets, with a cross-section scan.
study by Nealey and co-workers38 in which poly(styrene-b-methyl methacrylate) (Mn ) 50-b-54 kg/mol) diblock copolymers were directed to assemble onto patterned polymer brushes. It should be noted, however, that relatively long annealing times (up to 170 h) at 190 °C were used. Of course, the stability of the alkanethiol SAMs used in our studies limits the temperature at which the surfaces may be annealed.
Conclusions The results of this study demonstrate that patterning of a gold surface with MHDA and PFDT monolayers can be used to form a template that drives the selective adsorption of block copolymers containing both hydrophilic and hydrophobic segments. Annealing of the spin-coated sample is necessary to permit the polymer to diffuse enough to find its stable surface location. It (38) Edwards, E. W.; Stoykovich, M. P.; Muller, M.; Solak, H. H.; de Pablo, J. J.; Nealey, P. F. J. Polym. Sci. B 2005, 43, 3444-3459.
has been shown that polymeric lines at least as small as 500 nm wide can be formed using this methodology, and there is no reason to believe that much smaller features cannot be fabricated. Somewhat surprisingly, the difference in wetting properties of PFDT- and ODT-covered surfaces, apparently due to a greater defect density in the case of PFDT, are large enough to affect the morphology of the block copolymer film. Smooth patterns do not form when ODT is used as the hydrophobic layer because of the formation of polymeric droplets that preferentially adsorb on the MHDA-covered surface. Acknowledgment. This material is based upon work supported by the National Science Foundation under Grant No. NSF0425826. The authors acknowledge helpful discussions with Prof. T. Randall Lee. LA0605034
