1556
Langmuir 2005, 21, 1556-1559
Nanoparticle Arrays Patterned by Electron-Beam Writing: Structure, Composition, and Electrical Properties Jose L. Plaza, Yu Chen,* Susanne Jacke, and Richard E. Palmer Nanoscale Physics Research Laboratory, School of Physics and Astronomy, The University of Birmingham, Edgbaston, Birmingham, B15 2TT, U.K. Received September 6, 2004. In Final Form: November 30, 2004
Direct electron beam writing in nanoparticle films is employed to create nanoscale wires between prepatterned gold electrodes on SiO2/Si wafers. Characterization of these nanowires using AFM, SEM, and EDX reveals a core/sheath morphology, where a gold-rich core is surrounded by a sheath which is mainly of carbon. Z-contrast STEM images indicate that the central core consists of a distribution of metal cores in a carbon network. The results suggest that the nanoparticle network is created through crosslinking of the ligands of adjacent particles. The high resistivities obtained in conductivity measurements are consistent with this picture. The work illustrates the ability to generate patterned nanoparticle arrays which can be addressed electrically.
Nanoparticles have been the subject of intensive research in recent years because of the unique properties originating from their quantum-scale dimensions and potential applications in, for example, electronics, sensors, and catalysis.1 Nanostructures have been created from nanoparticles by employing templating techniques. Recent examples include the construction of linear assemblies of nanoparticles on DNA scaffolds.2 Ultimately, to create nanodevices on lithographically patterned self-assembled monolayers3 and from nanoparticles, one needs to assemble these “artificial atoms” at specific sites on the relevant substrate together with appropriate probes and interconnects. Direct electron-beam writing in a film of nanoparticles is a technique which allows the patterning of nanoparticle assemblies through a three-step processscoating, writing, and rinsing.3 Lines with width less than 30 nm have been produced in a film of alkanethiol-passivated Au nanoparticles.4 Since the technique is compatible with conventional photolithography,5 it has the capability to link “top-down” and “bottom-up” nanofabrication approaches.6 While the creation of nanostructures by this method has been demonstrated,7-9 the structure, composition, and the electrical properties of the nanostructures have not yet been explored. Such a study is essential to understand * Corresponding author. E-mail:
[email protected]. (1) (a) Khairutdinov, R. F. Colloid J. 1997, 59, 535. (b) Alivisatos, A. P. J. Phys. Chem. 1996, 100, 13226. (c) Alvarez, M. M.; Khoury, J. T.; Schaaff, T. G.; Shafigullin, M. N.; Vezmar, I.; Whetten, R. L. J. Phys. Chem. B 1997, 101, 3706. (d) Lewis, L. N. Chem. Rev. 1993, 93, 2693. (e) Li, Y.; El-Sayed, M. A. J. Phys. Chem. B 2001, 105, 8938. (f) Boyen, H. G.; Kastle, G.; Weigl, F.; Koslowski, B.; Dietrich, C.; Ziemann, P.; Spatz, J. P.; Riethmuller, S.; Hartmann, C.; Moller, M.; Schmid, G.; Garnier, M. G.; Oelhafen, P. Science 2002, 297, 1533. (g) Durston, P. J.; Palmer, R. E.; Wilcoxon, J. P. Appl. Phys. Lett. 1998, 72, 176. (h) Palmer, R. E.; Guo, Q. Phys. Chem. Chem. Phys. 2002, 4, 4275. (2) Warner, M. G.; Hutchison, J. E. Nature Mater. 2003, 2, 272. (3) 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. (4) (a) Bedson, T. R.; Palmer, R. E.; Wilcoxon, J. P. Appl. Phys. Lett. 2001, 78, 2061. (b) Bedson, T. R.; Jenkins, T. E.; Hayton, D. J.; Wilcoxon, J. P.; Palmer, R. E. Appl. Phys. Lett. 2001, 78, 1921. (5) Chen, Y.; Palmer, R. E. In Electron beam writing in nanoparticle films, Encyclopaedia of Nanoscience and Nanotechnology; Nalwa, H. S., Ed.; American Scientific Publishers: Stevenson Ranch, CA, USA, 2004; Vol. 3, pp 17-28. (6) Mendes, P. M.; Chen, Y.; Palmer, R. E.; Nikitin, K.; Fitzmaurice, D.; Preece, J. A. J. Phys.: Condens. Matter 2003, 15, S3047.
fully, and thus optimize, the electron-beam writing process. Here we address this task to characterize nanowires generated by electron-beam writing in alkanethiol-passivated gold nanoparticles by employing AFM, SEM, EDX, and STEM. The substrate employed was a boron-doped Si wafer (〈111〉 oriented) with a 65-nm-thick thermally grown SiO2 layer on top. Gold electrodes, for four-probe conductivity measurements, were fabricated by electron-beam evaporation of a 200-nm Au film and patterning of the Au film using photolithography and ion-beam milling. No wetting layer was used in this work in order to minimize the possibility of electrical shorts. The nanoparticles, prepared using the inverse micelle method,10 with a gold core of diameter 4-4.5 nm passivated by C16H33S ligands, were dissolved in an octane with a concentration of 0.01 M. After removing the unexposed photoresist and cleaning the sample with acetone, the sample was spin-coated with 1 µL of nanoparticle solution and left to dry for 2 h. The film thus formed has a thickness of ∼100 nm. The electron-beam writing was carried out with a dualstage scanning electron microscope (SEM DS-130F) equipped with a Raith Elphy Quantum writing package. Wires 40 µm long were written between the gold electrodes on the prepatterned substrate, followed by a postexposure rising to remove the unexposed nanoparticles. Figure 1a is a secondary electron SEM image of a wire written across two gold electrodes. The width of the line is about 100 nm, as confirmed by the tapping mode AFM image shown in Figure 1b. Figure 1c is a line profile derived from Figure 1b and gives a height for the wire of 155 nm. (7) (a) Wybourne, M. N.; Clarke, L.; Yan, M.; Cai, S. X.; Brown, L. O.; Hutchinson, J.; Keana, J. F. W. Jpn. J. Appl. Phys., Part 1 1997, 36, 7796. (b) Clarke, L.; Wybourne, M. N.; Yan, M.; Cai, S. X.; Keana, J. F. W. Appl. Phys. Lett. 1997, 71, 617. (c) Clarke, L.; Wybourne, M. N.; Yan, M.; Cai, S. X.; Brown, L. O.; Hutchinson, J.; Keana, J. F. W. J. Vac. Sci. Technol. B 1997, 15, 2925. (d) Reetz, M. T.; Winner, M. J. Am. Chem. Soc. 1997, 119, 4539. (e) Lohau, J.; Friedrichowski, S.; Dumpich, G.; Wassermann, E. F.; Winner, M.; Reetz, M. T. J. Vac. Sci. Technol. B 1998, 16, 77. (f) Dumpich, G.; Lohau, J.; Wassermann, E. F.; Winner, M.; Reetz, M. T. Mater. Sci. Forum 1998, 287-288, 413. (8) Lin, X. M.; Parthasarathy, R.; Jaeger, H. M. Appl. Phys. Lett. 2001, 78, 1915. (9) Werts, M. H. V.; Lambert, M.; Boourgoin, J. P.; Brust, M. Nano Lett. 2002, 2, 43. (10) Wilcoxon, J. P.; Williamson, R. L.; Baughman, R. J. J. Chem. Phys. 1993, 98, 9933.
10.1021/la047777g CCC: $30.25 © 2005 American Chemical Society Published on Web 01/22/2005
Direct Electron-Beam Writing in Nanoparticle Films
Langmuir, Vol. 21, No. 4, 2005 1557
Figure 1. A wire of width 100 nm generated by electron-beam writing at 9 keV between gold electrodes. (a) SEM image, (b) AFM image taken in tapping mode, and (c) cross-section of the wire from the AFM image.
Figure 2. Wires of width 490 nm written between gold electrodes. (a) Backscattered electron SEM image, (b) AFM image taken in tapping mode, (c) cross-section of the wire from the AFM image, and (d) typical EDX spectrum taken from the nanowire.
Three parallel lines of width about 500 nm were written on another sample with a dose of 3.4 × 105 µC cm-2, Figure 2. In this case, the values of the beam current, dwell time, and step size were 320 pA, 8 ms, and 0.003 µm, respectively. The backscattered electron SEM image of these lines, displayed in Figure 2a, reveals a central bright area with dark regions on both sides. This observation suggests a large amount of heavy atoms, i.e., gold, in the central core of the wires, since gold atoms give a high yield of backscattered electrons in comparison with light
atoms such as carbon. Figure 2b is an AFM image of one of the three nanowires, taken in tapping mode. The line profile in Figure 2c reveals a central plateau about 490 nm wide and 180 nm high surrounded by a sheath of height 160 nm. This central plateau region corresponds to the bright central area, i.e., the gold-rich core, in Figure 2a. The sheath in Figure 2c can be related to the dark regions surrounding the core in the backscattered electron image. The conclusion is that we have a core rich in gold surrounded by a sheath of light atoms, probably carbon,
1558
Langmuir, Vol. 21, No. 4, 2005
originating from organic ligands and possibly also residual solvent. EDX measurements confirm the presence of gold and carbon in the nanowires. A typical spectrum taken from the wires is presented in Figure 2d, where peaks associated with Au (MR1), Si (KR1), and C (KR1) can be distinguished. The quantitative analysis gives atomic compositions for Au of 8 ( 3%, Si 51 ( 7, and C 41 ( 8%. No oxygen was detected, presumably due to the thickness of the silicon oxide layer. Although sulfur is present in the ligands, a sulfur signal did not appear in the EDX spectra, probably due to its low content and also a large EDX sampling volume. A relatively low electron-beam voltage of 11 kV was used in EDX measurement in order to reduce the probing depth and to enhance the signal from the outermost region of the analyzed area. Even so, the EDX sampling volume (typically of order 1 µm3) is still significantly larger than the size of the nanowire (a wire cross-section of 0.4 µm2 is deduced from Figure 2c). As a result, the substrate and the surface regions surrounding the nanowire must make considerable contributions to the spectrum. This explains the large amount of Si in the quantitative analysis. The large amount of carbon is assigned to the sheath surrounding the central core. Having established the core/sheath morphology of the nanowires, we now turn to the question of the internal structure of the central, gold-rich core. To address this question, the same nanoparticle solution was dropdeposited onto a thin SiO2 film (∼30 nm thick) on a TEM grid. A film of monolayer thickness was formed to facilitate TEM imaging. Nanowires were written on this sample using a SEM with a condition similar to those used to create the nanowires in Figure 2. No postexposure rinsing was taken, thus the unexposed Au nanoparticles remain for comparison. Figure 3a shows a high-angle annular dark field (HAADF) STEM image of five nanowires. It is observed that the line width increases as the dose increases (from right to left), consistent with previous reports.3,4 A high-resolution STEM image, Figure 3b, of the third nanowire written with an electron dose of 4 × 105 µC cm-2 at 7 keV beam energy reveals close-packed nanoparticle islands, without apparent sintering of the metal particles, rather similar to the unexposed particles. Lin et al.8 reported conventional TEM images of films of passivated gold nanoparticles exposed by direct electron beam (a typical dose of 1.44 × 104 µC cm-2 at 30 keV beam energy). Ordered particle arrays were also observed. It was thought that electron-beam exposure strips the passivating molecules from the gold cores, enabling the cores to stick to the underlaying Si3N4 substrate. In our case, the HAADF STEM images represent atomic number (Z) contrast micrographs. The probability of an electron being scattered into a large angle by an atom is approximately proportional to Z2.11 The total scattered intensity depends on the number of the atoms in the electron-scattering column, which in turn depends on the thickness and density of the sample. Assuming that the distribution of gold remains unchanged in the exposed area, as observed in the STEM images, the increase in the brightness of the exposed lines in Figure 3 suggests structural changes in the ligands. It was found that electron irradiation introduces crosslinking between neighboring molecules in a self-assembled monolayer of alkanethiols on an extended gold surface.12 (11) Keyse, R. J.; Garratt-Reed, A. J.; Goodhew, P. J.; Lorimer, G. W. Introduction to Scanning Transmission Electron Microscopy; βios Scientific Publishers Ltd, Oxford, UK, 1998.
Plaza et al.
Figure 3. STEM images, taken with a high-angle annular dark field detector, of (a) five nanowires created by direct electron-beam writing in a submonolayer nanoparticle film (beam energy of 7 keV), and (b) the third nanowire at high resolution together with a scheme depicting a network of discrete nanoparticles interconnected by a carbon matrix.
Such cross-linking could result in a network of nanoparticles which survive subsequent rinsing in a solvent. This process may also be consolidated by some carbon deposition during electron-beam exposure. The STEM images, therefore, suggest that the electron beam partially exposes the organic ligands, in the manner of a conventional, negative tone organic resist,1313 to create a nanoparticle network through cross-linking of the ligands of adjacent nanoparticles, as shown in a scheme at the bottom left of Figure 3b. Although the nominal electron-beam doses are similar in the case of the TEM samples, a higher actual dosage is expected in the case of the nanowires fabricated on the SiO2/Si wafers. This is because, in the case of a multilayer film, electron scattering in the nanoparticle film itself enhances the secondary electron yield, which is further (12) (a) Zharnikov, M.; Geyer, W.; Golzhauser, A.; Frey, S.; Grunze, M. Phys. Chem. Chem. Phys. 1999, 1, 3163. (b) Zharnikov, M.; Frey, S.; Heister, K.; Grunze, M. Langmuir 2000, 16, 2697. (c) Heister, K.; Zharnikov, M.; Grunze, M. Langmuir 2001, 17, 8. (13) Tada, T.; Kanayama, T.; Robinson, A. P. G.; Palmer, R. E.; Allen, M. T.; Preece, J. A.; Harris, K. D. M. Microelectron. Eng. 2000, 53, 425.
Direct Electron-Beam Writing in Nanoparticle Films
Figure 4. I-V characteristics obtained with the four-probe method at room temperature from wires of width 100 nm (solid circles) and 490 nm (open squares).
assisted by the thicker substrate. The question thus remains whether this slight increase in electron dosage could lead to a completely different writing scenario, that is, significant removal of the organic ligands leading to sintering of the metal cores to create a continuous metallic wire. Conductivity measurements were carried out on the nanowires fabricated on the prepatterned SiO2/Si substrate to explore this possibility. A relatively high conductivity is expected if a continuous metal structure is formed, while the distribution of metal cores in a carbon network should have poor conductivity due to the large tunneling gaps between the metal cores. The resistance of the nanowires was measured using the four-probe method to remove contact resistance. Figure 4 presents
Langmuir, Vol. 21, No. 4, 2005 1559
the I-V characterizations of the 100- and 490-nm-wide nanowires taken at room temperature. A linear characteristic was observed for both nanowires in a voltage range from -200 to 200 mV. A resistance of 271 MΩ was obtained for the 100-nm-wide wire, and a mean value of 4.52 MΩ was obtained for a 490-nm-wide wire via linear fits to the data. The resistivity thus derived is 8.2 × 107 µΩ cm for the 100-nm-wide wire and 1.1 × 106 µΩ cm for a 490nm-wide wire. The difference in the resistivity (a factor of ∼7) may reflect a slightly different wire core structure or difference in the organic sheath. However, the main point is that the resistivity of these nanowires is significantly higher than the typical resistivity of amorphous metals (∼200 µΩ cm), implying the absence of a continuous metal network inside the nanowire. This is consistent with the picture emerging from STEM measurements. In conclusion, the work reported demonstrates the possibility of fabricating robust patterns of nanoparticles on a wafer. Electron-beam exposure seems to proceed by cross-linking of the organic ligands. The compatibility of this technique with conventional photolithography has been demonstrated by the creation of nanoparticle wires on a prepatterned substrate, thus illustrating the potential to assemble arrays which can be addressed electrically. Acknowledgment. This work was supported by the EPSRC and European Union through the Micro-Nano RTN (HPRN-CT-2000-00028) and the award of a Marie Curie Individual Fellowship to J.L.P. (HPMF-CT-200201906). LA047777G
