Papers

AVS1996 Session NS-WeA: Nanofabrication I

Wednesday, October 16, 1996 2:00 PM in Room 202A

Wednesday Afternoon

Start	Invited?	Item
2:00 PM		NS-WeA-1 Controlled Growth of Si-Oxide Barriers for Si Based Resonant Tunneling Devices Y. Wei, R. Wallace, A. Seabaugh (Texas Instruments Inc.) One of the key issues in the construction of silicon-based resonant tunneling devices is how to produce a crystalline Si quantum well (QW) layer on top of the Si-oxide tunneling barrier. It has been shown that Si overlayer growth on amorphous Si-oxide is amorphous or microcrystalline at best. One possible way to grow crystalline Si on top of an ultrathin silicon oxide barrier film is to create sub-nanometer holes (voids) in the oxide film. The crystalline silicon exposed by these holes will then act as "seeds" for later nucleation of the crystalline Si QW layer. It is important that these void dimensions are small enough so that the tunneling barrier appears essentially continuous to the confined electrons in the available resonant states. In this paper, we will report two methods which produce such tunneling barriers. One approach starts with the ultrathin oxide film grown on an atomically clean Si(100) surface, making use of the void formation mechanism in the oxide film through the controlled production of volatile SiO. The desorption of the ultrathin oxide film is inhomogeneous and we find that the density and size of the voids are functions of annealing temperature and annealing time under ultrahigh vacuum conditions. A second approach entails the use of the kinetic competition between nucleation of SiO\sub 2\ islands and the formation of gas phase SiO (etching of the Si surface) within the temperature-pressure phase space. By controlling the temperature and the O\sub 2\ pressure, the size and density of the oxide islands can be controlled. In the region between the oxide islands, a clean Si crystalline surface exists which can be used as a seed for the subsequent nucleation of crystalline Si overlayers. We also found that the size and density of the oxide islands saturate at some point. At the initial stage, the size and distribution of the oxide islands are irregular, but as the etching and island nucleation continue, they evolve to a more uniform size and spatial distribution. This phenomenon provides an effective way to produce high density Si pillars or Si nano-dots.
2:20 PM		NS-WeA-2 Focusing a Neutral Atomic Beam to Nanometer Resolution using a Laser R. Behringer, V. Natarajan, G. Timp (Bell Laboratories) It is now technologically possible to focus a thermal flux of neutral atoms to 15\+-\2nm resolution with 6:1 contrast by using a nearly resonant laser beam as an atom-optical lens. We have demonstrated this by focusing a 692K sodium flux into a grating of about 100,000 lines on a clean silicon substrate, using a Gaussian standing wave laser beam detuned from the D\sub 2\ transition at 589nm. The atomic profiles of the lines that comprise the grating were examined in-situ, subsequent to the deposition, using an ultra-high vacuum scanning tunneling microscope. The measurements were further corroborated with semi-classical numerical simulations of the atom dynamics in the laser beam. The strategy used to achieve this unprecedented atom-optical performance was simple: we pursued laser beam profiles that resulted in short focal lengths, while suppressing the aberrations that cause the contrast to deteriorate at short focal lengths, and we used an atomic flux with a minimal angular divergence. To shorten the focal length we manipulated the Gaussian profile of the laser beam by changing the waist and by changing the position of the substrate with respect to the center of the beam. The power in the laser beam was then chosen to optimize the measured resolution, which incidentally corresponds to the power calculated from numerical simulations of the experiments. Generally, the resolution improves in proportion to the focal length. However, for extremely short focal lengths, below 30 \mu\m, the measured resolution deteriorates, contrary to our expectations based on simulations. The origin of the discrepancy is not yet known, but we speculate that spontaneous emission and the resulting nonadiabatic effects not included in the simulations play a role.
2:40 PM		NS-WeA-3 Molecular Nanostructures: Functionalized Design, Assembly at Surfaces and Characterization of Properties T. Jung, R. Schlittler, J. Gimzewski (IBM Zurich Research Laboratory, Switzerland) We present the generation of molecular nanostructures, which can be grown on substrates and readily modified using scanning probe microscopy (SPM) displacement and molecular-modification techniques\super 1\. The mechanical as well as the electronic properties of the structures grown are characterized using SPM Numerical simulations are related to the experimental observations. Using examples, we discuss concepts for simple functions that ultrasmall structures built from molecular subunits can exhibit. We acknowledge financial support by the BBW of Switzerland through the ESPRIT basic research program PRONANO (8523). This work was performed in collaboration with C. Joachim and H. Tang, CNRS Toulouse, France. \super 1\T.A. Jung et al., Science, Vol. 271, p. 181 (1996)
3:20 PM		NS-WeA-5 Tip Induced C\sub 60\ Island Growth and Manipulation using a Scanning Tunneling Microscope. A. Dunn, Y. Ma, P. Moriarty, M. Upward, P. Beton (University of Nottingham, United Kingdom) We have used the tip of an ultra-high vacuum scanning tunneling microscope to induce growth and manipulate hexagonally ordered C\sub 60\ islands on the Si(111)-7x7 surface. Growth is initiated by scanning mulitlayer coverages of C\sub 60\ at reduced sample bias voltages. Using this procedure we have observed tip induced lateral and vertical growth of hexagonally ordered 3rd layer islands by up to 20nm and 5 monolayers respectively. During this growth, a corresponding depletion of adjacent lower layers is observed. We have observed similar effects in C\sub 60\ multilayers deposited on Si(110) and present a simple model related to transfer of C\sub 60\ between tip and surface at step edges followed by diffusion to growing C\sub 60\ islands. Manipulation is achieved by moving the tip towards the surface and then sweeping it across to a predetermined position. This technique has previously been used to position individual C\sub 60\ molecules on Si(111)-7x7[1]. By applying this procedure to multilayer coverages of C\sub 60\ it is possible to displace 2nd and 3rd layer islands leaving the 1st monolayer of C\sub 60\ unaffected and create micron long lines of width as small as 8nm. [1] P.H.Beton, A.W.Dunn and P.Moriarty, Appl. Phys. Lett. 67 1075 (1995).
3:40 PM		NS-WeA-6 In-situ Method of Fabricating Nanoscale Structures on the Si(100) Surface T. Hashizume, S. Heike, M. Lutwyche, S. Watanabe, Y. Wada (Hitachi, Ltd., Japan) For a step of realizing nanoscale devices, such as Atom Relay Transistor (ART) [1], we have developed a method for fabricating nanoscale structures on the silicon surface and connecting them to bulk electrodes in-situ by using scanning-tunneling-microscopy (STM) based atom manipulation. We use a metal mask with micron-sized patterns to form metal electrodes as well as bonding pads on the hydrogen terminated Si(100)-2x1-H surface by thermally evaporating Ti in-situ. The hydrogen atoms on the surface are desorbed by electron exposure from the STM tip and nanoscale dangling bond wires are formed [2]. We have characterized the tip in-situ by field-ion-microscopy (FIM) observation and needle formation and tip imaging (NFTI) method [3] in order to minimize the width of the dangling bond wires. In order to fabricate more detailed dangling bond structures, several methods for manipulating the individual hydrogen atoms are tested. By utilizing the difference in adsorption energy of metal atoms on the hydrogen terminated and hydrogen desorbed silicon atoms, we have thermally deposited Ga atoms and fabricated nanoscale Ga wires on the Si surface. Adsorption of Ga atoms on this surface are observed only at the dangling bond positions and on the surface impurities. A method of connecting the metal electrodes to the nanoscale structures will also be discussed. [1] Y. Wada et al., J. Appl. Phys. 74, 7321 (1993). [2] J. W. Lyding et al., Appl. Phys. Lett., 64, 2010 (1994). [3] S. Heike et al., Jpn. J. Appl. Phys. 34, L1061 (1995).
4:00 PM		NS-WeA-7 Atomic Scale Manipulation of Si(111)-7x7: A Variable Temperature STM Study B. Stipe, M. Rezaei, W. Ho (Cornell University) We have carried out atomic scale manipulation of silicon adatoms on Si(111)-7x7 with a recently constructed variable temperature STM over the range 35K to 300K. Although similar experiments at room temperature have been reported, our low temperature results are significantly different. At room temperature, applying a voltage pulse to the STM tip usually results in the complete removal of adatoms - presumably by field evaporation. Occasionally, silicon atoms are seen in the middle of one corner Si adatom and two center Si adatoms. This is directly above a "rest atom". Our room temperature results are consistent with published data and it has been proposed that these are atoms that have been evaporated from the surface to the tip and re-evaporated back to the surface. At low temperatures, we find that field evaporation does not take place. Instead, center Si adatoms in the faulted half of the unit cell are shifted to one of neighboring T\sub 4\ sites not normally occupied on the Si(111)-7x7 surface. A one-to-one correspondence between the missing adatoms and shifted atoms and the proximity of shitfted atoms to the missing atoms suggest a direct displacement mechanism rather than a tip mediated process. Furthermore, a voltage pulse can result in an ordered array of these shifted atoms - one per unit cell. These results suggest that evaporation at room temperature is a two stage process. The silicon atoms are first shifted to the neighboring T\sub 4\ sites before the thermally activated field evaporation takes place.
4:20 PM		NS-WeA-8 Layer-by-Layer Atomic Manipulation and Quantized Conductance on Si(111)-7x7 Surface T. Komeda, R. Hasunuma, H. Mukaida (JRCAT - Angstrom Technology Partnership, Japan); H. Tokumoto (JRCAT - National Institute for Advanced Interdisciplinary Research, Japan) There have been many studies to remove atoms from the Si(111)-7x7 surface by applying pulse voltages to the tip. However, there are few reports on the subsurface structures after removing atoms. Here we removed the Si atoms from the Si(111)-7x7 surface by approaching the biased STM tip to the point contact region and consecutively retracting to the original tunneling region (1). In this talk, we shall show how Si atoms are removed, what kind of structure appears behind, and how the current between the tip and substrate behaves during the tip excursion. The Si-atom removal proceeded in a layer-by-layer manner that was controllable by changing the tip bias. When a W tip was biased at -2 V against the substrate, Si atoms only in the adatom layer were selectively removed in the tip retraction process exposing a clear 2nd surface. This may be due to the adhesion between the W tip and Si atoms at the adatom layer in the point contact. At the same time, characteristic staircase structure was observed in the current between the tip and substrate, which could be originated by the decrease of the number of Si atoms in the junction in the tip retraction process. At the biase of -2.5 V, there appeared layers of 0.3 nm in depth with a clear atomic images of 2x2, c(2x4) and \sr\3x\sr\3 structures in the created hole. In this case, there were sudden jumps of the current in the tip approach process which can be attributed to Si atoms-removal from the tri-layer of the Si(111)-7x7 surface by the field evaporation effect. (1) T. Komeda, R. Hasunuma, H. Komeda and H. Tokumoto, in June 10 issue of APL (1996).
4:40 PM		NS-WeA-9 Laser Assisted Atomic-Scale Surface Modifications with the STM I. Lyubinetsky, Z. Dohnalek, V. Ukraintsev, J. Yates, Jr. (University of Pittsburgh) The atomically-controlled surface modification of the Si(100)-(2x1) surface covered with chemisorbed Cl atoms has been achieved using laser irradiation of the UHV STM junction when in the tunneling range. From the fully-covered surface, laser pulses result in the removal of a preselected pair of Cl atoms from the Si dimer, and the bright clean Si dimer feature is revealed. Laser pulses of 8 nsec with photon energy of 2.33 eV and very low power, less than 0.01 mJ/cm\super 2\, were used. At slightly higher laser power, a single atom from the W tip may be deposited on clean Si(100) surface [1]. A tentative mechanism for the laser induced Cl atom desorption with STM involves rapid thermal expansion of the tip as a first step. For understanding the next step, the efficiency of Cl atom removal, resulting from the close proximity of the tip and the sample are examined as a function of the laser power and STM parameters (tunneling gap and bias). [1] V. A. Ukraintsev, and J. T. Yates, Jr., J. Appl. Phys. (submitted). * Work supported by the Office of Naval Research.