Mater. Res. Soc. Symp. Proc. Vol. 1113 © 2009 Materials Research Society

1113-F03-01

Monolayer Film of Gold Nanoparticles on a 3 inch or Larger Silicon Wafer Matthew N. Martin and Sang-Kee Eah Department of Physics, Applied Physics, and Astronomy, Rensselaer Polytechnic Institute, Troy, NY 12180, U.S.A. ABSTRACT We chemically synthesized charged gold nanoparticles coated with hydrophobic organic molecules, which are stable in hexane but unstable in toluene. If a hexane droplet with charged gold nanoparticles is mixed with a larger toluene droplet, they immediately float to the airtoluene interface forming a monolayer of close-packed gold nanoparticles. The monolayer can be deposited to any substrate simply by the solvent molecules’ evaporation with no limit in size and without using any sophisticated instrument. As a demonstration, we fabricated a very large monolayer of close-packed gold nanoparticles covering the whole surface of a 3 inch silicon wafer. We further showed that excess organic ligand molecules do not affect the charged gold nanoparticles’ property of floating to and forming a monolayer at the air-toluene interface. However, these very slowly evaporating molecules remain on a substrate, affect the twodimensional ordering, and significantly reduce the contrast in scanning electron microscopy. INTRODUCTION Chemical synthesis of colloidal nanoparticles has been a very active research area with progress especially in the size and shape control, improvement in the size monodispersity, and three-dimensional (3D) self-assembly. Still, there are many challenges to two-dimensional (2D) self-assembly of colloidal nanoparticles for fabricating a monolayer or multilayer film of high quality in spatial uniformity and batch reproducibility, which is important for many applications. Most popular traditional 2D self-assembly techniques for colloidal nanoparticles in solution are spin-coating and the Langmuir–Blodgett (LB) methods, where various parameters, such as the compression pressure in the LB method [1], must be carefully controlled to obtain a monolayer of close-packed nanoparticles. Recently, new methods were reported for a higher quality monolayer film of nanoparticles by either spin-coating a mixture of polymers and nanoparticles [2], or excess surfactant molecules’ catching nanoparticles at the air-toluene interface [3], where a monolayer of nanoparticles sits above and/or below a thick layer of organic molecules, limiting the range of applications. In addition, all these 2D self-assembly methods have difficulties when applied to fabrication of a very large monolayer such as covering the whole surface of a silicon wafer. Here we present a very simple 2D self-assembly method for colloidal nanoparticles at the air-toluene interface of a toluene droplet in order to fabricate a monolayer of close-packed nanoparticles with high spatial uniformity and high batch reproducibility. The monolayer has no limit in size without using any sophisticated instruments. As a demonstration we fabricated a close-packed monolayer of 5 nm diameter gold nanoparticles covering the whole surface of a 3 inch silicon wafer. For this 2D self-assembly method we chemically synthesized gold nanoparticles coated with hydrophobic molecules which are charged in non-polar solvents, and thus are stable in hexane but unstable in toluene. Also, we present the effects of excess ligand molecules to the nanoparticles’ 2D monolayer formation.

EXPERIMENT We chemically synthesized gold nanoparticles coated with 1-dodecanethiol (DDT) molecules which are negatively charged in non-polar solvents according to a procedure recently developed by us, which is highly reproducible and takes <10 minutes with no need for a cleaning step [4]. This method allows us to control the amount of DDT molecules in a solution precisely. A typical synthesis is briefly explained as follows. All chemicals were used as received from Sigma-Aldrich and deionized water was used. Two aqueous stock solutions were made for HAuCl4·3H2O and NaBH4 at the concentration of 50 mM, to which the same molar amounts of HCl and NaOH are added, ensuring stability for more than several months and hours respectively. To a glass vial with 10 g of water, 100 µL of the AuCl4- stock solution was added. 575 µL of the BH4- stock solution was injected all at once to the vial on a vortex shaker to ensure uniform mixing. The vial was kept on the shaker to remove hydrogen gas molecules for 60 seconds. All the gold ions are reduced by BH4- and become gold atoms, which form gold nanoparticles <3 nm in diameter (orange color) immediately in <1 second and grow slowly into larger nanoparticles (red color) by aggregation. To speed up growth, the vial was put to a preheated hot plate for 180 seconds. After cooling in a water bath for 60 seconds, UV-VIS spectroscopy was performed to check whether all the solutions generate the same spectrum and therefore contain the same size nanoparticles of 5 nm in diameter. To the vial, 5 g of acetone was added and mixed for 1 second. After adding 5 g of hexane with DDT molecules the vial was shaken vigorously by hand for 30 seconds to extract only gold nanoparticles to the hexane phase, leaving all the reaction byproducts in the water-acetone phase, even without a cleaning step.

Figure 1. Fabrication schematic of a monolayer of close-packed gold nanoparticles (left) and its photographic image on a 3 inch silicon wafer (right). The charged gold nanoparticles coated with hydrophobic organic molecules float to the air-toluene interface forming a monolayer, which is deposited to a substrate after toluene’s evaporation. This very simple method can be applied to any substrate with no limit in size and without using any sophisticated instrument. As shown in Fig. 1, toluene was poured to a 3 inch silicon wafer covering the whole surface. Then, hexane droplets containing gold nanoparticles were deposited drop-wise to control the coverage rate to just above a full monolayer. These gold nanoparticles coated with hydrophobic DDT molecules are negatively charged in non-polar solvents [4]. As a result, they

are stable in hexane but unstable in toluene. They immediately float to the air-toluene interface and form a monolayer, since there are more toluene molecules than hexane molecules surrounding a nanoparticle. Also note that hexane evaporates ~4 times faster than toluene. The monolayer of nanoparticles at the air-toluene interface was very gently transferred to the silicon substrate as the toluene molecules were removed by evaporation. To study the effects of excess DDT molecules we prepared three hexane solutions of gold nanoparticles with the amount of DDT molecules as 10, 100, and 1000 % of the number of the gold atoms. The corresponding concentrations of DDT molecules bound on the surface of the gold nanoparticles and free in 5 g of hexane are 65 µM, 650 µM, and 6500 µM respectively. The concentrations of the gold nanoparticles were controlled so that one hexane droplet, held in a glass pipette due to the capillary effect, forms 1.2 monolayers of close-packed gold nanoparticles on a 5x5 mm2 silicon chip. After depositing a hexane droplet with gold nanoparticles we waited for ~20 seconds before putting a toluene droplet to make the ratio of toluene to hexane larger than 4. The evaporation of toluene on a 5x5 mm2 silicon chip took ~5 minutes. With an optical microscope in reflection mode and a digital camera for the horizontal field of view of 476 µm, we monitored 2D self-assembly of the gold nanoparticles in real time at the air-toluene and air-silicon interfaces. For nanoscopic imaging of individual nanoparticles on the silicon chips we used a field emission scanning electron microscope (FE-SEM) instead of a transmission electron microscope to check uniformity over the whole area of a silicon chip from the edges to the center. DISCUSSION In the popular Brust synthesis procedure for gold nanoparticles coated with hydrophobic molecules [5], tetraoctylammonium bromide (TOAB) is used to extract gold ions from water to toluene. Gold ions in toluene are reduced by a strong reducing agent BH4- and form gold nanoparticles in the presence of DDT molecules very slowly for more than a few hours. A cleaning step is necessary to remove the reaction byproducts, especially TOAB. Instead we make 5 nm gold nanoparticles in water and then use acetone and DDT to extract them from water to hexane without TOAB in <10 minutes. Since there is no cleaning step, we can control the amount of DDT molecules precisely. We observed that the phase transfer was complete, leaving no nanoparticles visible in the water phase for all three solutions. Figure 2 shows photographs, optical micrographs, and electron micrographs of gold nanoparticles deposited on silicon chips from the three solutions. We found out that excess DDT molecules do not affect the charged gold nanoparticles’ property of floating to and forming a monolayer at the air-toluene interface, checked by eye at the millimeter scale and by optical microscopy at the micrometer scale. We couldn’t find any noticeable difference among the three monolayers (data not shown here). However, optical microscopy of the gold nanoparticles on the surface of the silicon chips after the toluene’s evaporation shows that these very slowly evaporating DDT molecules remain on a substrate mostly in the scattered concentrated regions, whose density increases as the amount of DDT molecules increases. In the case of the 1000% solution we could even observe Newton’s interference rings in enormous 3D clusters of DDT molecules, indicating a few µm height. We also found out that the as-received glass vials are significantly dirty, affecting 2D self-assembly of the gold nanoparticles checked by optical microscopy at the surface of the silicon chips; therefore, cleaning them is necessary.

Figure 2. Images of 5 nm diameter gold nanoparticles coated with 1-dodecanethiol (DDT) molecules on the surface of silicon chips at the millimeter (top row), micrometer (middle row), and nanometer (bottom row) scales. We controlled the amount of DDT molecules in the hexane solutions of gold nanoparticles as 10% (left column), 100% (center column), and 1000% (right column) of the number of gold atoms. We observed that excess DDT molecules do not affect the charged gold nanoparticles’ property of floating to and forming a monolayer at the air-toluene interface. However, these very slowly evaporating DDT molecules remain on the substrates, forming ~100 nm defect areas void of gold nanoparticles (center-bottom), and lower the contrast significantly in scanning electron microscopy, making it impossible to resolve individual nanoparticles (right-bottom). The carbon deposition due to excess DDT molecules by the scanning electron beam at a lower magnification is shown in the inset (right-bottom).

Optical microscopy does not provide enough contrast for the areas between regions of concentrated DDT molecules. However the FE-SEM imaging clearly shows deposition of DDT molecules above or below the nanoparticles monolayer by the much lower intensity of the gold nanoparticles from the 100 % solution. In the case of gold nanoparticles from the 1000 % solution, there are so many excess DDT molecules that it is impossible to resolve individual nanoparticles. Carbon deposition by the scanning electron beam at a lower magnification is shown in the inset of Fig. 2. Note that it must be possible to resolve individual nanoparticles in transmission electron microscopy even with the monolayer of nanoparticles from the 1000 % solution. Therefore FE-SEM imaging might be used instead of transmission electron microscopy as a fast and very effective quantitative tool to measure the amount of excess organic molecules around a monolayer of nanoparticles, provided that there is an independent method to calibrate the data. We suspect that the ~100 nm defect areas void of nanoparticles (center-bottom of Fig. 2) are caused by excess DDT molecules. The density of such defect areas is very low with the 10 % solution, while it increases significantly with the 100 % solution. Therefore we speculate that using the least amount of DDT molecules is important for a monolayer of close-packed gold nanoparticles for minimizing the ~100 nm defect areas. Also, for some applications such as memory devices using gold nanoparticles for storing charge, it is very important to minimize the amount of excess organic molecules around a monolayer of nanoparticles [6]. When we used 1 % DDT molecules in the hexane phase, the phase transfer was not complete and the color of the water phase remained red instead of becoming colorless. We suspect that gold nanoparticles cannot be transferred to the hexane phase if they are not completely covered by DDT molecules. Therefore we conclude that the range from 1 to 100 % of the number of gold atoms for DDT molecules is the optimal range to control the amount of excess organic molecules around a monolayer of nanoparticles. In addition to minimizing the amount of DDT molecules, we found out that controlling the coverage rate is important to minimize the density of defect areas void of nanoparticles. Just above a full coverage seems optimal. A complete monolayer coverage at the larger area of the air-toluene interface due the curvature of the toluene droplet helps to compress nanoparticles further at the smaller area on the flat substrate. We also observed that excess gold nanoparticles and DDT molecules are deposited to the edges, forming dense multilayers, perhaps due to the coffee-stain effect [7]. Previously, the Lin-Jaeger group explained the gold nanoparticles floating to the airtoluene interface of a toluene droplet as excess DDT molecules catching nanoparticles at the airtoluene interface [3]. Therefore they emphasized the importance of excess DDT molecules and the rate of toluene’s evaporation. However we found out that our gold nanoparticles from the lowest 10 % solution float to the air-toluene interface even if the evaporation rate is minimized in a toluene-saturated space, as long as there are more toluene molecules than hexane molecules. We speculate that the difference comes from the different charge number of a nanoparticle. At this moment we know only the negative sign of charged gold nanoparticles in toluene, not the charge number. Preliminary results show that the nanoparticles’ speed of floating to the airtoluene interface can be reduced by decreasing the negative charge number of a nanoparticle, and therefore making nanoparticles more stable in toluene. The 2D self-assembly processes of neutral nanoparticles in a drying droplet are very complex [8], while charged nanoparticles unstable in toluene float to the air-toluene interface and form a close-packed monolayer very simply. Precisely controlling the charge number of

nanoparticles during the chemical synthesis and studying their stabilities in various non-polar solvents may be a new direction of research with good prospects. Currently we are working on extending this approach from gold to silver and platinum. Recently the charge number of semiconductor nanoparticles in chloroform was controlled among –e, 0, +e, and +2e without reporting their 2D self-assembly [9]. We speculate that the charge number of nanoparticles must be larger, in the positive or the negative direction, for 2D self-assembly. Also we are trying to identify the source of the negative charge, which must be one of the anions during the synthesis: chloride, hydroxide, boron anions, and deprotonated dodecanethiolate. We speculate that all gold ions are reduced by excess borohydride and therefore not responsible for the negative charge. This task is new and challenging when taken together with measuring the charge number distribution in non-polar solvents and identifying the counter cations. CONCLUSIONS We fabricated a very large monolayer of close-packed gold nanoparticles covering the whole surface of a 3 inch silicon wafer to demonstrate that there is no size limit in 2D selfassembly of charged gold nanoparticles coated with hydrophobic DDT molecules in non-polar solvents, without using any sophisticated instruments. They are stable in hexane but unstable in toluene, and as a result they immediately float to the air-toluene interface of a toluene droplet when mixed with a smaller hexane droplet of gold nanoparticles. A monolayer of close-packed nanoparticles is formed at the air-toluene interface, which can be deposited to any substrate after the solvent molecules’ evaporation. By controlling the amount of DDT molecules we showed that excess DDT molecules do not affect the charged gold nanoparticles’ property of floating to and forming a monolayer at the air-toluene interface. However, too many excess DDT molecules generate defect areas void of gold nanoparticles and significantly lower the contrast in scanning electron microscopy, making it impossible to resolve individual nanoparticles. ACKNOWLEDGMENTS This work was supported by a Rensselaer Polytechnic Institute Start-up fund. REFERENCES 1. J. R. Heath, C. M. Knobler, and D. V. Leff, J. Phys. Chem. B 101,189 (1997). 2. S. Coe-Sullivan, J. S. Steckel, W.-K. Woo, M. G. Bawendi, and V. Bulović, Adv. Funct. Mater. 15, 1117 (2005). 3. T. P. Bigioni, X. -M. Lin, T. T. Nguyen, E. I. Corwin, T. A. Witten, H. M. Jaeger, Nature Mater. 5, 265 (2006). 4. M. N. Martin, J. I. Basham, P. Chando, and S. -K. Eah (submitted for publication). 5. M. Brust, M. Walker, D. Bethell, D. J. Schiffrin, and C. Kiely, J. Chem. Soc., Chem. Commun., 801 (1994). 6. S. Paul, C. Pearson, A. Molloy, M. A. Cousins, M. Green, S. Kolliopoulou, P. Dimitrakis, P. Normand, D. Tsoukalas, and M. C. Petty, Nano Letters 3 (4), 533 (2003). 7. R. D. Deegan, O. Bakajin, T. F. Dupont, G. Huber, S. R. Nagel, and T. A. Witten, Nature 389, 827 (1997). 8. E. Rabani, D. R. Reichman, P. L. Geissler, and L. E. Brus, Nature 426, 271 (2003). 9. E. V. Shevchenko, D. V. Talapin, N. A. Kotov, S. O'Brien, and C. B. Murray, Nature 439, 55 (2006).