Formation mechanism and Coulomb blockade effect in self-assembled gold quantum dots S. F. Hu a) National Nano Device Laboratories, Hsinchu 300, Taiwan R. L. Yeh and R. S. Liu Department of Chemistry, National Taiwan University, Taipei 106, Taiwan Received 20 June 2003; accepted 20 October 2003; published 24 December 2003 Nanometer-scale Au quantum dots have been assembled on SiO 2 by controlling the reaction of raw materials to form a citrate Au sol and aminosilane/dithiol treated patterned Si wafer. Details of the formation mechanism have been studied. Three gold colloidal particles 15 nm, aligned in a chain to form a one-dimensional current path, were bridged on an 80 nm gap between the source and drain metal electrodes. The device exhibited a Coulomb blockade effect at 33 K. 2004 American Vacuum Society. DOI: 10.1116/1.1633774 I. INTRODUCTION Single electron transistors SETs have been proposed as the future basis for nanoelectric components. SETs, which utilize the Coulomb blockade effect that arises from electrostatic charge energy of a single electron, 1 are one of the promising candidates for future integrated circuits due to their extremely low power consumption and potential for high-density integration. In principle, SETs can be scaled down to the atomic size, but physical implementation of such a device poses several difficulties. In order to avoid thermally induced current tunneling, structures with extremely small capacitance must be constructed (E C e 2 /2C T k B T, where E C is the Coulomb charging energy and C T is the total capacitance. Moreover, extremely uniform, metal or semiconductor nanoparticles embedded in a medium free of charged impurities are required. One of the ways by which to achieve such a superlattice is to covalently link nanometerdiam metal clusters to each other by means of rigid, double ended organic molecules. Gold nanoparticles consisting of nanocrystal cores, inorganic citrate shells and alkyl dithiol linkers promise interesting technological applications in SETs. 2 4 However, research on the formation mechanism from raw materials i.e., hydrogen tetrachloroaurate (HAuCl 4.3H 2 O) and trihydrate trisodium citrate dihydrate (Na 3 C 6 H 5 O 7 2H 2 O)] to citrate gold sol and SETs is still very limited. Here, we report on the development of a stable process to assemble gold nanoparticles into chains on a patterned Si wafer in order to form SETs. II. EXPERIMENT A. Synthesis of citrate gold sol A starting solution of 100 ml of 2.2 mm sodium citrate as a reductant was heated to 90 C and 40 ml of 0.815 mm HAuCl 4 was added by rapidly mixing. The solution was heated further at 90 C for 15 min. The solution initially developed a pale yellow color and then a gray color, which changed to lavender and then transformed into red in 1 3 min. The resulting colloid sol had a mean particle size of around 15 nm. B. Formation of the patterned Si wafer The devices were fabricated on a p-type Si substrate with 2000 Å of SiO 2 by thermal oxidation. Around 2000 Å of poly-si was grown on top of the SiO 2 layer as a bottom gate. Then a layer of SiO 2 with thickness of around 200 Å was grown on top of the poly-si to isolate the metal source and drain electrodes. After lithography and chemical etching were performed on the 200 Å thick SiO 2, a hole to connect the bottom gate was formed. In order to form a nanometersized space between the source and drain as a channel for the deposition of Au nanoparticles, we needed to spin coat 400 nm of NEB22 negative photoresist. Then e-beam lithography was performed, and Al metal sputtering and liftoff were performed, resulting in the patterned wafer shown in Fig. 1. Details of the process of gold nanoparticle SET fabrication have been reported by Sato et al. 2 4 Here, we briefly describe the process. We concentrate on studying kinetic formation of the Au particles in order to set up a stable reproducible process. C. Self-assembly of the first layer of gold nanoparticles The SiO 2 film was thermally grown on a Si substrate and a patterned Si wafer was used in the experiment. After cleaning the substrate using acetone and isopropyl alcohol and ustrasonically agitating, the samples were immersed in a dia Author to whom correspondence should be addressed; electronic mail: sfhu@ndl.gov.tw FIG. 1. Schematic of a SET device composed of a three-dot gold chain. 60 J. Vac. Sci. Technol. B 22 1, JanÕFeb 2004 1071-1023Õ2004Õ22 1 Õ60Õ5Õ$19.00 2004 American Vacuum Society 60
61 Hu, Yeh, and Liu: Formation mechanism and Coulomb blockade effect 61 FIG. 2. Formation process of citrate Au sol after heating sodium citrate and HAuCl 4 at 90 C for a 0, b 10, c 20, d 30 and e 46 min. lute water solution of 0.05% v/v 3-2-aminoethylamino propyltrimethoxysulane APTMS for 5 min. The silane coupling agent APTMS was purchased from Aldrich and used as received without further purification. Then the samples were dried with a nitrogen gun and baked at 120 C for 30 min in an oven. The APTMS-treated samples were then immersed in citrate Au sol for 3 h. After immersion, the samples were rinsed with distilled water and dried with a nitrogen gun. This procedure led to the deposition of a submonolayer of Au particles which repelled each other due to the formation of negative citrate ions on the surface of the Au particles. FIG. 3. Variation of UV/vis spectra with an increase in reaction time after heating sodium citrate and HAuCl 4 at 90 C. D. Self-assembly of the second layer of gold nanoparticles After the submonolayer Au nanoparticle coating, the samples were immersed in 5 mm ethanolic solution of 1,6- hexanedithiol for 24 h. In this process the dithio replaced the citrate ion to bind the Au surface with a covalent bond. Then the dithiol treated samples were immersed in the citrate Au sol for 8 h. A second layer of Au nanoparticles was then bonded to the first Au layer of nanoparticles by the covalent bonding of alkanedithiol. High-resolution transmission electron microscope HR- TEM measurements were carried out using a Hitachi FIG. 4. Variation of the maximum absorption peak obtained by UV/vis and particle size obtained by HRTEM as function of the reaction time of sodium citrate and HAuCl 4 at 90 C. FIG. 5. Au particle density on the SiO 2 /Si substrates as function of the time immersed in citrate Au sol obtained by SEM observation. JVST B-Microelectronics and Nanometer Structures
62 Hu, Yeh, and Liu: Formation mechanism and Coulomb blockade effect 62 FIG. 7. UV/vis spectra a before and b after the reaction of the first layer of the Au nanoparticle deposited on the SiO 2 /Si substrate with 1,6- hexanedithiol. A corresponding SEM photograph of b is shown in the inset. FIG. 6. XRD patterns of a Au nanoparticles coated on the SiO 2 /Si substrate, b a Au film coated on the SiO 2 /Si substrate, c a standard Au pattern and d a standard Si pattern. H-7100 electron microscope operating at 100 kv for determination of the particle size during the formation of citrate Au sol. Ultraviolet/visible UV/vis spectra Hitachi U-4100 of Au nanoparticles were also obtained to determine the absorption wavelength. Scanning electron microscope SEM Hitachi S-4000 measurements were carried out to determine the particle size of the Au nanoparticles on the SiO 2 /Si substrate. X-ray diffraction XRD patterns of all the samples were recorded on a SCINTAG X1 x-ray diffractometer with Cu K radiation. A HP4156C precision semiconductor parameter analyzer was used to measure the electrical properties. III. RESULTS AND DISCUSSION The formation process of the citrate Au sol after reaction of sodium citrate and HAuCl 4 is shown in Fig. 2. Usually, there are two processes for the formation of the citrate Au sol. 5,6 The first one is nucleation after heating sodium citrate and HAuCl 4 at 90 C for 20 min. During this step, the hydroxy group OH in citrate will reduce Au 3 ions into Au ions. Then the Au ions will carry out a self-redox reaction and form Au nanoparticles. The HRTEM photographs shown in Figs. 2 a 2 c demonstrate the nucleation process. The second step after heating sodium citrate and HAuCl 4 at 90 C for more than 20 minutes is the growth process which will increase the size of the Au nanoparticles. Corresponding HR- TEM photographs are shown in Figs. 2 d and 2 e. The average diameter of the Au nanoparticles shown in Fig. 2 e is 15 3 nm. The variation in UV/vis spectra with an increase in reac- FIG. 8. SEM photographs of a the Au nanoparticle coated SiO 2 /Si substrate, b the Au nanoparticle modified by 1,6-hexanedithiol and c after deposition of the second Au nanoparticle layer. J. Vac. Sci. Technol. B, Vol. 22, No. 1, JanÕFeb 2004
63 Hu, Yeh, and Liu: Formation mechanism and Coulomb blockade effect 63 FIG. 9. SEM photograph of a device in which the source drain gap 80 nm is bridged by a chain of three Au nanoparticles. tion time after heating sodium citrate and HAuCl 4 at 90 C is shown in Fig. 3. In the nucleation process reaction time less than 20 min, an increase in the reaction time leads to the appearance of a maximum absorption peak around 544 nm which corresponds to the formation of Au nanoparticles with surface plasma vibration. During the growth process, a decrease in the maximum absorption peak from 544 to 519.5 nm is observed when the reaction time is increased from 20 to 90 min. A decrease in the maximum absorption peak obtained by UV/vis and an increase in the Au particle size obtained by HRTEM with an increase in the reaction time during the growth step are shown in Fig. 4. Moreover, after 45 min the reaction between sodium citrate and HAuCl 4 at 90 C became stable. The Au particle density on the SiO 2 /Si substrates as a function of the immersion time in citrate Au sol obtained by SEM observation is shown in Fig. 5. With an increase in dipping time, the density of the Au nanoparticles increases. However, the density of Au saturates when the dipping time is more than 3 h. Moreover, Au nanoparticles repel one another due to the negative charge of the citrate ion on the surface of the Au nanoparticles. The XRD pattern of the Au nanoparticles coated on the SiO 2 /Si substrate is shown in Fig. 6 a. Compared to standard XRD patterns of Au film obtained by thermal evaporation on the SiO 2 /Si substrate Fig. 6 b and on Au Fig. 6 c and Si Fig. 6 d, a broadened peak with a 2 value of around 38.5 corresponding to the 111 plane is seen in the inset of Fig. 6. This proves that close packing has occurred for the formation of Au on the SiO 2 /Si substrate. The particle size of Au is around 11 nm, calculated from the Scherrer formula, 7 using the 111 diffraction peak which is consistent with the SEM result 15 nm. After reaction of the first layer of Au nanoparticles deposited on the SiO 2 /Si substrate with 1,6-hexanedithiol, the dithio will replace the citrate ion and bind the Au surface with a covalent bond. This reaction will lead to a decrease in the distance between Au nanoparticles. This aggregation phenomenon will give rise to a redshift of the maximum absorption peak in the UV/vis measurements shown in Fig. 7. The corresponding SEM photograph inset of Fig. 7 after reaction with dithiol clearly shows the aggregation of Au nanoparticles. Figure 8 shows SEM photographs of Au nanoparticles coated with the SiO 2 /Si substrate, the Au nanoparticle modified by 1,6-hexanedithiol and after deposition of the second Au nanoparticle layer. Aggregations have occurred after the treatment by dithiol, and are consistent with the results as shown in Fig. 7. After deposition of the second layer of Au nanoparticles on the dithiol treated Au surface first layer, many chain shaped aggregations appeared on the surface of the SiO 2 /Si substrate. Figure 9 is a SEM photograph of one particular device in which the source drain gap 80 nm is bridged by a chain of three Au nanoparticles. Each Au has a particle size of around 15 nm. Current voltage characteristics measured at T 33, 40 and 300 K as a function of the source drain bias for this particle device with a three-dot chain are shown in Fig. 10 a. The nonlinear behavior is more pronounced at low temperature, which is a sign of Coulomb blockade. The Coulomb blockade effect is more pronounced at lower temperature. The device exhibited a Coulomb gap ( V) of about 150 mv at 33 K. Assuming dominant island charging, the island capacitance (Cg e/ V) is estimated to be e/2 V 4 10 19 F. It can be seen that the three-dot chain exhibits FIG. 10. a Current voltage characteristics measured at T 33, 40, and 300 K as function of the source drain bias for this particle device with a three-dot chain and b the corresponding enlarged area of T 33 K. JVST B-Microelectronics and Nanometer Structures
64 Hu, Yeh, and Liu: Formation mechanism and Coulomb blockade effect 64 a very low current regime in a considerable range of voltage around zero bias, indication of a Coulomb blockade effect at 33 K. When the drain bias is sufficiently high, blockade of the current through the three-dot chain can be lifted. This is due to the additional energy that permits electrons to occupy states at higher orbits in the confined quantum dot, thus opening a large window of energy for tunneling. Figure 10 b shows the Coulomb staircase effect observed at 33 K. IV. CONCLUSION We have fabricated three-dot Au nanoparticle chain SETs that exhibit a Coulomb blockade effect at 33 K. The reaction mechanisms from the raw materials and patterned Si wafer were studied. The particle size of Au between citrate sol and the deposited SiO 2 /Si substrate can be controlled around 15 nm with good reproducibility due to the stable growth mechanisms in our fabrication methods. ACKNOWLEDGMENTS This work was performed under Contract Nos. NSC92-2215-E-492-008 and NSC92-2113-M-002-036 from the National Science Council of the Republic of China. The technical support of members at the National Nano Device Laboratories is acknowledged. 1 H. Grabert and M. Devoret, Single Electron Tunneling Plenum, New York, 1992. 2 T. Sato, D. G. Hasko, and H. Ahmed, J. Vac. Sci. Technol. B 15, 45 1997. 3 T. Sato and H. Ahmed, Appl. Phys. Lett. 70, 2759 1997. 4 T. Sato, H. Ahmed, D. Brown, and B. F. G. Johnson, J. Appl. Phys. 82, 696 1997. 5 M. K. Chow and C. F. Zukoski, J. Colloid Interface Sci. 165, 97 1994. 6 V. Privman, D. V. Goia, J. Park, and V. Matijevic, J. Colloid Interface Sci. 213, 36 1999. 7 A. R. West, Solid State Chemistry and its Applications Wiley, New York, 1986. J. Vac. Sci. Technol. B, Vol. 22, No. 1, JanÕFeb 2004