Fabrication of Gold Micro- and Nanostructures by Photolithographic

Feb 21, 2006 - Unoxidized particles may, in contrast, readily be removed leaving gold structures behind at the surface. This process provides a conven...
0 downloads 0 Views 655KB Size
NANO LETTERS

Fabrication of Gold Micro- and Nanostructures by Photolithographic Exposure of Thiol-Stabilized Gold Nanoparticles

2006 Vol. 6, No. 3 345-350

Shuqing Sun,§ Paula Mendes,† Kevin Critchley,# Sara Diegoli,† Marcus Hanwell,‡ Stephen D. Evans,# Graham J. Leggett,*,§ Jon A. Preece,*,† and Tim H. Richardson‡ Department of Chemistry, UniVersity of Sheffield, Brook Hill, Sheffield S3 7HF, U.K., School of Chemistry, UniVersity of Birmingham, Edgbaston, Birmingham B15 2TT, U.K., School of Physics and Astronomy, UniVersity of Leeds, Leeds, LS2 9JT, U.K., and Department of Physics and Astronomy, UniVersity of Sheffield, Brook Hill, Sheffield S3 7RH, U.K. Received October 28, 2005; Revised Manuscript Received January 12, 2006

ABSTRACT Exposure of thiol-stabilized gold nanoparticles supported on silicon wafers to UV light leads to oxidation of the thiol molecules and coagulation of the nanoparticles, forming densified structures that are resistant to removal by solvent exposure. Unoxidized particles may, in contrast, readily be removed leaving gold structures behind at the surface. This process provides a convenient and simple route for the fabrication of gold structures with dimensions ranging from micrometers to nanometers. The use of masks enables micrometer-scale structures to be fabricated rapidly. Exposure of nanoparticles to light from a near-field scanning optical microscope (NSOM) leads to the formation of gold nanowires. The dimensions of these nanowires depend on the method of preparation of the film: for spin-cast films, a width of 200 nm was achieved. However, this was reduced significantly, to 60 nm, for Langmuir−Schaeffer films.

The spatial organization of nanometer-scale components (such as metal nanoparticles, carbon nanotubes, proteins, cells, organic molecules, polymers, etc.) onto surfaces to fabricate functional nanostructured systems for electronic,1,2 optical,3 biological,4 or sensing applications,5,6 is one of the most important challenges in nanoscale science and technology. Recently there has been a great deal of interest in the fabrication of nanoscale structures composed of nanoparticles. Such structures are attractive for a variety of reasons, including their potential utility as nanostructured conductors, novel resists, and templates for the attachment of molecules with specific electronic, sensing, or biological recognition characteristics.7 A variety of methods have been reported for the patterning of nanoparticles. One approach is to fabricate a template by patterning a suitable substrate. For example, gold clusters have been attached to adhesive patches fabricated in alkyl* To whom correspondence should be addressed. Tel:+ 44 (0) 121 414 3528, fax:+ 44 (0) 121 414 4403, e-mail: [email protected]; or tel: +44 (0) 114 222 9556, fax: +44 (0) 114 222 9346, e-mail: [email protected]. † University of Birmingham. ‡ Department of Physics and Astronomy, University of Sheffield. § Department of Chemistry, University of Sheffield. # University of Leeds. 10.1021/nl052130h CCC: $33.50 Published on Web 02/21/2006

© 2006 American Chemical Society

silane monolayers on silicon substrates8 and phosphonic acidfunctionalized nanoparticles have been attached to silicon surfaces in areas delineated by the selective photolithographic passivation of the substrate.9 Such approaches are not restricted to metallic nanoparticles. Xia and Brueck have formed patterns in photoresist on silicon to control the immobilization of spin-coated silica nanoparticles,10 whereas Minelli et al. used block copolymer films, which exhibited nanoscale phase separation in conjunction with plasma treatment to form regions that were adhesive to nanoparticles.11 Electron beam lithography12 and scanning probe lithography methods have been utilized to create nanopatterned templates. For example, dip-pen nanolithography13,14 and nanoshaving15 have been used to pattern self-assembled monolayer (SAM) templates onto which gold nanoparticles may subsequently be immobilized, and aldehyde-functionalized polymer nanoparticles have been patterned by attachment via imide bond formation to amine-functionalized regions fabricated in carboxylic acid-terminated SAMs using scanning near-field photolithography.16 Local oxidation on exposure of alkylsiloxane monolayers to a biased AFM tip has been used to introduce hydroxyl functionalities, to which

nanoparticle-binding aminosiloxanes may be attached at the surface of a silicon substrate.17,18 Fresco and Frechet recently used a biased AFM tip to selectively transform thiocarbamate groups to thiol functionalities that may be utilized to immobilize gold nanoparticles with exceptional spatial resolution.19 Here we report a new approach to nanoparticle patterning based on the use of photochemical methods that may be conveniently used on length scales ranging from tens of micrometers to tens of nanometers. The approach we have developed is based on the photochemical oxidation of adsorbed alkylthiolates, which passivate gold nanoparticles to yield alkylsulfonates that do not have a passivating effect. The photopatterning of alkanethiol monolayers has been reported extensively.20-22 On exposure to UV light in the presence of oxygen, alkylthiolates are converted to weakly bound alkylsulfonates that may readily be displaced from the gold surface by a second thiol or in a rinsing process.23,24 Recently we have demonstrated that by utilizing a near-field scanning optical microscope (NSOM) coupled to a UV laser as the light source, an approach that we term scanning nearfield photolithography (SNP),25,26 it is possible to create patterns in SAMs approximately 10 times smaller than the conventional diffraction limit of λ/2. Although NSOM is regarded as a difficult technique, many of the complications associated with its use are removed in lithographic operation and SNP is a comparatively simple route for the ambient fabrication of nanostructured monolayers. In the present study we demonstrate its application to fabricate structures from nanoparticles with line widths as small as 60 nm in a simple lithographic process involving only two steps. The process used to pattern them is illustrated schematically in Figure 1. Thiol-stabilized nanoparticles are deposited onto a silicon substrate, either in multilayer form (spincasting) or as a bilayer (Langmuir-Schaeffer deposition). The sample is then exposed to UV light, either through a mask (micropatterning) or by exposure to light from an aperture-based NSOM. In regions where the nanoparticles are exposed to UV light, photooxidation occurs and alkylthiolates are converted to alkylsulfonates. This, we hypothesized, should lead to depassivation of the nanoparticles and be followed by coagulation, while unexposed nanoparticles remain stabilized. Rinsing should lead to removal of the unoxidized nanoparticles, leaving the coagulated structures intact on the substrate surface. Decanethiol-stabilized gold nanoparticles were synthesized and characterized by the Brust method.27,28 Thiol-stabilized gold nanoparticles were deposited onto a clean silicon wafer substrate using two methods. First, a solution of colloidal gold nanoparticles in CHCl3 was spin-coated on to a SiO2 surface using a rotational speed of 800-1200 rpm and the samples were allowed to dry in air at room temperature for at least 1 h. Second, a constant-perimeter Joyce-Loebl mini trough with one moveable arm was used with ultrapure water (resistivity greater than 15 MΩ cm). Solutions of gold nanoparticles dissolved in chloroform at concentrations of 0.2-0.5 mg mL-1 were spread on the surface. After spreading, a period of 10 min was allowed for all of the 346

Figure 1. Schematic diagram showing the process used to generate nanoparticle patterns. Incident UV irradiation photochemically oxidizes the passivating thiolates to sulfonates, leading to depassivation and aggregation of the gold nanoparticles. Finally, the unirraditated gold nanoparticles are rinsed away, leaving the depassivated, coagulated gold particles.

chloroform to evaporate before compressing the Langmuir monolayer to a surface pressure of 15-20 mN m-1 and holding it there for another 10 minutes. The samples were then deposited onto hydrophobic silicon substrates using the Langmuir-Schaeffer technique at a speed of 4 mm min-1 on both the up and down strokes. Only one Langmuir-Schaeffer excursion was used in all of the samples, giving a bilayer of gold nanoparticles on the silicon surface.29 The samples were allowed to dry in air at room temperature for at least 24 h. Photopatterning was conducted using light from a frequencydoubled argon ion laser (Coherent FreD 300C, Coherent U.K., Ely), which emits at 244 nm. For micrometer-scale patterning, a square grid mask consisting of 6.5 × 6.5 µm2 square openings separated by 5.5 µm beams was positioned on the nanoparticulate surface and the sample irradiated at a power of 100 mW. The irradiation time was typically 8 min, and the area of illumination was typically 0.2-0.4 cm2. The surface was then washed by immersing the samples in toluene and sonicating for 30 s to remove the unirradiated particles and oxidized sulfur species. For scanning near-field photolithography, the laser was coupled to a ThermoMicroscopes Aurora III near-field scanning optical microscope fitted with a fused silica fiber probe (Veeco). The probe velocity was 0.1 µm s-1. XPS spectra were obtained for spin-cast samples of nanoparticles on an ESCALAB 250 instrument equipped with a monochromatic Al KR X-ray source, using a pass energy of 20 eV and a step size of 0.10 eV. Figure 2a shows the S 2p region of the spectrum for a freshly prepared nanoparticle film. Peak-fitting using 30% Lorentzian/70% Gaussian peaks revealed the characteristic S 2p3/2 and S 2p1/2 doublet with components at 162.8 and 164 eV with areas in the ratio Nano Lett., Vol. 6, No. 3, 2006

Figure 2. The S 2p region of the XPS spectrum of the thiolpassivated gold nanoparticulate film on SiO2 before (a) and after (b) exposure to 244 nm laser light, respectively. The difference in the S 2p peak width before and after irradiation is due to difficulties in fitting the S 2p after irradiation due to the presence of a variety of different chemical states.

2:1, indicative of a thiolate (RS-) bound to a gold surface.30 A weaker doublet at 164.0 eV (S 2p3/2) and 165.2 eV (S 2p1/2) suggested the presence of a small amount of unbound thiol species. However, in all other respects, the spectra strongly resembled those of SAMs formed on planar gold substrates. XPS analysis of irradiated areas revealed that the components of the sulfur 2p spectrum associated with the as-received nanoparticles were much reduced, while a new and prominent peak was observed at a higher binding energy. This broad feature probably incorporated a number of components because of the oxidized sulfur species.30 The oxidized sulfur components were still evident after washing in toluene (Figure 2b). The observation of oxidized sulfur after washing was surprising. One possible explanation is that the irradiation process caused cross-linking between adsorbates. However, evidence from static SIMS suggests that this is unlikely. In an earlier study, it was shown that cross-linking in SAMs following electron beam exposure leads to the appearance of polycyclic aromatic species in static SIMS spectra.31 These highly characteristic peaks were not observed in SIMS spectra of monolayers exposed at 254 nm.24 This suggests that cross-linking is not expected to occur at these wavelengths. It is more likely that the change in the morphology of the nanoparticulate film following UV exposure (see discussion below) leads to a reduction in the ease with which oxidation products may be displaced from the film by solvent molecules. Figure 3 shows AFM images of the irradiated surface before and after rinsing, respectively. Before rinsing with toluene (Figure 3a) there is some evidence of modification to the film, with irradiated areas (squares) exhibiting dark contrast, reflecting a reduction in height of ∼10 nm compared to the masked areas. This indicates that the oxidation of the thiol adsorbates on the nanoparticle surfaces leads to some reduction in stabilization. However, a much more substantial change is observed following rinsing (Figure 3b). After sonication in toluene, well-defined square islands with lateral dimensions 6.5 × 6.5 µm2 and heights of ca. 35 nm were observed, indicating that the patterned nanoparticles posNano Lett., Vol. 6, No. 3, 2006

sessed substantial structural integrity and the mask dimensions had been transferred to the nanoparticle film with high fidelity. The thickness of the original spin-coated film was 45 nm, suggesting that a compression of the nanoparticles has occurred. It is not clear whether oxidation has taken place throughout the full thickness of the spin-cast film. Previous studies have suggested that light with the wavelength used here is almost completely absorbed by a continuous 20 nm thick polycrystalline gold film. However, in the present case, there are hollows between the nanoparticles, and even if they adopted a close-packed structure, scattering of light could still occur in regions between the nanoparticles. Moreover, even in the absence of penetration of light to the lower levels of the film, collapse of the upper layers of nanoparticles may also cause densification of lower layers. Closer examination of the square islands reveals that the edges are not straight (Figure 3c) and they contain defects on the surface, which appear to be small depressions (Figure 3d). These are likely due to the compression of the film structure following oxidation of the passivating layer and may indicate that the resulting structure is partially porous. The measured diameters of these holes are ca. 120 nm, and their depths are measured as 40 nm; however, these data must be treated with caution because the AFM tip may not be fully able to penetrate the holes. It is not clear from these data whether the holes are present only at the surface or throughout the bulk; however, if the latter is the case, then these materials may offer useful novel properties as a function of their expected high surface area-to-volume ratio. Figure 4 shows structures fabricated in spin-cast films using an NSOM coupled to a UV laser in place of the mask/ lamp combination used in Figure 3. It was found that following near-field exposure, the nanostructures did not survive sonication, so a more gentle procedure was used in which the sample was rinsed three or four times in toluene. In contrast to the structures that result from micropatterning, near-field exposure yields dense, continuous structures with no evidence of porosity. The full width at half-maximum (fwhm) of the structures in Figure 4 is ca. 180 nm, and their heights are ca. 14 nm. The reduced height of the features compared to the macroscopic patterns, and the absence of porosity, suggests that in the much smaller region illuminated by the NSOM the nanoparticles are able to coagulate more effectively: close-packing following exposure of micrometerscale regions would lead to large voids, which is presumably energetically unfavorable, whereas close-packing in nanoscale regions can occur without extensive collapse of the film structure over larger distances. The fwhm is larger than that in earlier studies by SNP but significantly smaller than what is readily accessible using mask-based processes. We think it is likely that the finite thickness of these nanoparticulate films facilitates the divergence of the electric field associated with the NSOM aperture. Moreover, as the film begins to aggregate, diffraction through pores may occur. It is also possible that illumination of the nanoparticles leads to the excitation of plasmon modes. Although the wavelength used here (244 347

Figure 3. Tapping-mode AFM images of spin-cast nanoparticle films following exposure to UV light before (a) and after (b-d) sonication in toluene. (a and b) 100 × 100 µm2 images. (c) 10 × 10 µm2 AFM image of a single feature formed using a mask. (d) 5 × 5 µm2 of the surface of the feature shown in c, and accompanying line section, revealing pores with mean diameters of 120 nm.

nm) is some way from the position of the maximum in the bulk gold plasmon absorption, off-maximum absorption, or the shifting of the plasmon peak by the film morphology, cannot be ruled out. For example, Jin et al. used photonic processes to convert silver nanospheres to nanoprisms.32 The gold nanoparticle film clearly constitutes a complex medium in which a variety of processes may be excited, and further investigations must explore this in detail. On the basis of these considerations, we hypothesized that reducing the thickness of the nanoparticle film may yield superior edge definition and resolution. Bilayers of gold nanoparticles were prepared using Langmuir-Schaeffer deposition. Micropatterns were fabricated by exposure to UV light through a mask and characterized by tapping-mode 348

AFM. Figure 5 shows a nanoparticle bilayer sample before and after rinsing with toluene. Even before rinsing, a significant change in the topography of the sample is observed, suggesting that oxidation of the decanethiol molecules that stabilize the nanoparticles leads to a modification of film structure. After rinsing, unoxidized nanoparticles are completely removed, yielding structures that exhibit sharp edge definition (Figure 5b) and an absence of the porosity that characterized the thicker spin-cast films. The fidelity of pattern reproduction appears to be superior to that achieved using the spin-cast films. Patterns were also fabricated in bilayer samples using SNP. Lines were traced across the surface of the silicon substrate using light from the NSOM, and the samples were rinsed in Nano Lett., Vol. 6, No. 3, 2006

Figure 4. Tapping-mode AFM image showing gold nanowires fabricated by exposure of nanoparticles to UV light from an NSOM, followed by rinsing in toluene.

toluene. The resulting structures were imaged by AFM, and illustrative results are displayed in Figure 6. Substantial improvements in resolution were achieved. Lines fabricated in Langmuir-Schaeffer films were significantly narrower than those formed in spin-cast films. The features shown in Figure 6 have a width of approximately 60 nm, comparable to the diameter of the aperture of the NSOM probe. This indicates that the excitation does not spread outside the area illuminated by the probe, in contrast to the behavior observed for the spin-cast films. The structures are continuous and appear generally free from defects. This means that although the nanoparticle film was only two layers thick, the particles have still coagulated effectively over distances of some 8 or 9 particle diameters. The height of the features (6-7 nm) is such that there has been little vertical collapse of the bilayer structure, although some segments of the lines are lower, suggesting thinning of the film in those areas. However, the predominantly uniform height of these features suggests that densification has occurred by a mechanism that involves movement predominantly in the horizontal plane. The high resolution achieved here (60 nm), the structural integrity, and the uniformity of the structures are all promising results for future applications of photolithographic methods in the fabrication of metal nanostructures. The procedures used here were simple ones based on ambient conditions, and the lithographic process involved only two steps (exposure and rinsing), promising widespread scope for its application. The mechanism of aggregation of the nanoparticles is not clear at present. Coagulation may simply result from Nano Lett., Vol. 6, No. 3, 2006

Figure 5. Tapping-mode AFM images of Langmuir-Schaeffer films of thiol-passivated gold nanoparticles before (a) and after (b) rinsing to remove unoxidized material. Image size: 100 × 100 µm2. Range of z-scale contrast: 0-31 nm (a) and 0-28 nm (b).

nonspecific interactions between the destabilized particles. However, it cannot be excluded that localized melting or similar phenomena may also occur. Such considerations are ultimately relevant to the applications of the structures we have fabricated because they may influence, for example, the efficiency of conduction along nanowires such as those shown in Figure 6, and the mechanism of aggregation must be the subject of future investigations. In summary, we have shown that photolithographic methods provide a simple route for the formation of metallic structures from thiol-stabilized gold nanoparticles. Exposure of the nanoparticles to UV light causes oxidation of the stabilizing molecules, leading to coagulation of the nanoparticles. For micrometer-scale patterns formed by spincasting, this yields defects and AFM data suggest that the structures may be porous. However, in near-field lithographic exposure, using a UV laser coupled to a near-field scanning optical microscope, the small width of the oxidized region means that coagulation can proceed by collapse of the structure from its edges, leading to dense continuous structures. The best resolution is achieved for LangmuirSchaeffer films, where continuous nanolines with widths of only 60 nm were fabricated in a simple two-step process (exposure and rinsing). Acknowledgment. This work was supported by the European Community under Grant no. HPRN-CT-200000028 and the Engineering and Physical Sciences Research 349

Figure 6. Nanostructures fabricated in Langmuir-Schaeffer films of decanethiol-stabilized nanoparticles using SNP. Image size: (a) 8.0 × 8.0 µm2 and (b) 3.6 × 3.6 µm2. Range of z-scale contrast: (a) 0-5.6 nm and (b) 0-9.0 nm.

Council (EPSRC) under grant GR/N82197/01. G.J.L. thanks the RSC Analytical Chemistry Trust Fund and the EPSRC for their support. References (1) Aviram, A.; Ratner, M. A. Chem. Phys. Lett. 1974, 29, 77. (2) Tans, S. J.; Dekker, C. Nature, 2000, 404, 834. (3) Chicane, C.; David, T.; Quidant, R.; Weeber, J. C.; Lacroute, Y.; Bourillot, E.; Dereux, A.; Colas de Francs, G.; Girard, G. Phys. ReV. Lett. 2002, 88, 087402. (4) Lee, K.-B.; Kim, E.-Y.; Mirkin, C. A.; Wolinsky, S. M. Nano Lett. 2004, 4, 1869. (5) Dagani, R. Chem. Eng. News 2000, 78, 27. (6) Ball, P. Nature 2000, 406, 118. (7) Li, X.-M.; Huskens, J.; Reinhoudt, D. N. J. Mater. Chem. 2004, 14, 2954. (8) Lu, N.; Gleiche, M.; Zheng, J.; Lenhert, S.; Xu, B.; Chi, L.; Fuchs, H. AdV. Mater. 2002, 14, 1812. (9) Foster, E.; Kearns, G.; Goto, S.; Hutchison, J. E. AdV. Mater. 2005, 17, 1542. (10) Xia, D.; Brueck, S. R. J. Nano Lett. 2004, 4, 1295.

350

(11) Minelli, C.; Hinderling, C.; Heinzelman, H.; Pugin, R.; Liley, M. Langmuir 2005, 21, 7080. (12) Mendes, P. M.; Jacke, S.; Critchley, K.; Plaza, J.; Chen, Y.; Nikitin, K.; Palmer, R. E.; Preece, J. A.; Evans, S. D.; Fitzmaurice, D. Langmuir 2004, 20, 3766. (13) Demers, L.; Park, S.; Taton, A.; Li, Z.; Mirkin, C. A. Angew. Chem., Int. Ed. 2001, 40, 3072. (14) Zhang, H.; Lee, K.-B.; Li, Z.; Mirkin, C. A. Nanotechnology 2003, 14, 1113. (15) Garno, J. C.; Yang, Y.; Amro, N. A.; Cruchon-Dupeyrat, S.; Chen, S.; Liu, G.-Y. Nano Lett. 2003, 3, 389. (16) Sun, S.; Chong, K. S. L.; Leggett, G. J. Nanotechnology 2005, 16, 1798. (17) (a) Hoeppener, S.; Maoz, R.; Cohen, S. R.; Chi, L.; Fuchs, H.; Sagiv, J. AdV. Mater. 2002, 14, 1036. (b) Liu, S.; Maoz, R.; Schmid, G.; Sagiv, J. Nano Lett. 2002, 2, 1055. (18) Ling, X.; Zhu, X.; Zhang, J.; Zhu, T.; Liu, M.; Tong, L.; Liu, Z. J. Phys. Chem. B 2005, 109, 2657. (19) Fresco, Z. M.; Frechet, J. M. J. J. Am. Chem. Soc. 2005, 127, 8302. (20) Huang, J.; Hemminger, J. C. J. Am. Chem. Soc. 1993, 115, 22433343. (21) Tarlov, M. J.; Burgess, D. R. F. J. Am. Chem. Soc. 1993, 115, 5305. (22) Cooper, E.; Wiggs, R.; Hutt, D. A.; Parker, L.; Leggett, G. J.; Parker, T. L. J. Mater. Chem. 1997, 7, 435. (23) Brewer, N. J.; Rawsterne, R. E.; Kothari, S.; Leggett, G. J. J. Am. Chem. Soc. 2001, 123, 4089. (24) Brewer, N. J.; Janusz, S. J.; Critchley, K.; Evans, S. D.; Leggett, G. J. J. Phys. Chem. B 2005, 109, 11247. (25) Sun, S.; Chong, K. S. L.; Leggett, G. J. J. Am. Chem. Soc. 2002, 124, 2414. (26) Sun, S.; Leggett, G. J. Nano Lett. 2004, 4, 1381. (27) Brust, M.; Walker, M.; Bethell, D.; Schiffrin, D. J.; Whyma, R. J. Chem. Soc., Chem. Commun. 1994, 801. (28) Briefly, HAuCl4 was transferred from water to toluene by stirring an aqueous solution of HAuCl4‚3H2O (30 mL, 30 mM) with tetraoctylammonium bromide solution in toluene (80 mL, 50 mM) for 30 min. The toluene phase was separated, and a solution of decanethiol in toluene (10 mL, 90 mM) was added. After 5 min, an aqueous NaBH4 solution (25 mL, 0.4 M) was slowly added to the mixture. After 3 h, the organic phase containing the gold nanoparticles was separated and concentrated to ∼10 mL under reduced pressure. The nanoparticles were purified by precipitation from the toluene solution with ethanol followed by centrifugation of the suspension. The supernatant was discarded. This precipitation/centrifugation process was repeated two more times after which the dark-brown solid was dried under vacuum. The 1H NMR of the decanethiol gold nanoparticles revealed, as expected, the signals of the decanethiol to be considerably broadened compared to the parent molecule. The gold nanoparticles were also redissolved in toluene and characterized by UV-vis spectroscopy. The UV spectrum showed an almost absent plasmon resonance absorption band because of the small size of the gold nanoparticles (∼3 nm). (29) Nabok, A.; Heriot, S. Y.; Richardson, T. H. Phys. Status Solidi B 2005, 242, 797. (30) (a) Hutt, D. A.; Leggett, G. J. J. Phys. Chem. 1996, 100, 6657. (b) Lee, M.-T.; Hsueh, C.-C.; Freund, M. S.; Ferguson, G. S. Langmuir 1998, 14, 6419. (c) Schoenfisch, M. H.; Pemberton, J. E. J. Am. Chem. Soc. 1998, 120, 4502. (d) Boeckl, M. S.; Bramblett, A. L.; Hauch, K. D.; Sasaki, T.; Ratner, B. D.; Rogers, J. W., Jr. Langmuir 2000, 16, 5644. (31) Hutt, D. A.; Leggett, G. J. J. Mater. Chem. 1999, 9, 923. (32) Jin, R.; Cao, Y.; Mirkin, C. A.; Kelly, K. L.; Schatz, G. C.; Zheng, J. G. Science 2001, 294, 1901.

NL052130H

Nano Lett., Vol. 6, No. 3, 2006