Atomistic Nanodevice Simulation

Friday July 8 2005 Editors: J.Bernardini, H.Dallaporta, Ch. Girardeaux 10 International Conference on the Formation of Semiconductor Interfaces (ICFSI-10) Aix-en-Provence, France, July 3-8, 2005 Friday 2 Fr.Plenary-8.30 I ATOMISTIC NANODEVICE SIMULATION Michel Lannoo L2MP UMR CNRS 6137, Faculté des Sciences de St Jérôme, F-13397 Marseille Cedex There now exist a large variety of methods for the elaboration of nanoobjects and nanodevices. At the same time the reduction in size together with the improvement of theoretical techniques and computational power should allow efficient quantitative simulation of these objects. Such predictive simulation is highly needed since it can serve as a useful guide to build new nanostructures with the desired properties. The aim of this talk is than to give an overview of what has already been achieved in this rapidly evolving field, what can be done at present and what is expected for the future. The focus will be mostly on semiconductor nanostructures and especially silicon. We start with confinement effects i.e the size dependence of the energy gap. We show that the results of ab-initio and semi empirical calculations are very close which allows the prediction for large nanoobjects. We also discuss the relation between different experimental determinations of the gap (electron or hole injection or electron hole excitation). We apply this to the determination of the tunnelling current across an In As quantum dot with a full interpretation of the origin of the different peaks. We then investigate the validity of using the bulk dielectric constant for nanostructures and show how it is possible to reconcile the different point of views expressed in the literature. The second part deals with the full calculation of transport in silicon nanowire field effect transistors. This is performed via a self-consistent (resolution of the 3D Poisson equation) tight binding approach coupled to a non-equilibrium Green’s function technique. It is also shown how effective mass theory can be quantitatively adapted. Results are provided for the electrical performances versus the shape of the gate, for the influence of elastic scattering by point defects and for the optimization of the characteristics with respect to the size of the transverse cross section. Extensions to include inelastic scattering and to treat the a.c regime are discussed. 10 International Conference on the Formation of Semiconductor Interfaces (ICFSI-10) Aix-en-Provence, France, July 3-8, 2005 Friday 3 Fr.A-9.15 0 Structural Characterization of Buried Nanostructures by Cross-Sectional Scanning Tunneling Microscopy H. Eisele, A. Lenz, R. Timm, and M. Dähne Institut für Festkörperphysik, Technische Universität Berlin, Hardenbergstr. 36, D-10623 Berlin, Germany Cross-sectional scanning tunneling microscopy (XSTM) is a fascinating tool for the structural characterization of buried layers and nanostructures with atomic resolution by imaging a cleavage plane perpendicular to growth direction. In contrast to top-view STM data taken directly at the growth surface, XSTM data often show strong intermixing and structural reorganization occurring during overgrowth. The detailed knowledge of the atomic structure of the buried layers is of high importance e.g. for semiconductor devices such as quantum dot lasers. In this contribution we demonstrate how the detailed atomic structure can be determined from XSTM data such as Fig. 1, where a InGaAs/GaAs quantum dot is shown. Of particular interest in such quantum dot samples are parameters such as dot size, dot shape, local stoichiometry, the amount of outward diffusion of the quantum dot material and finally the wetting layer structure. For this purpose, we developed a data evaluation procedure based on experimental parameters such as the tunneling voltage as well as on theoretical structure simulations based on the strain in the system close to the cleavage surface. Moreover, we can achieve information about the overgrowth process itself, since we can identify single segregated atoms from the active zone into the overgrowth layers, as shown in Fig. 2. From several MOCVD-grown and In(Ga)As/GaAs quantum dot samples, we derived a detailed view of the influence of the growth parameters on the atomic structure of the buried nanostructures. While the growth temperature was found to have a minor influence, the timing of growth interruptions at the different stages of dot growth plays a much more important role for the form and stoichiometry of the dots. In this way, we could observe structures ranging from welldefined and stoichiometrically almost pure InAs quantum dots, which are well suitable for laser devices, over strongly intermixed dots without a specific shape to completely dissolved regions of slightly higher indium content. These results not only lead to a much better understanding of the relevant growth processes, but also yield valuable information for the sample growers for improving the growth parameters. The authors would like to thank U.W. Pohl, K. Pötschke, R.L. Sellin, F. Heinrichsdorff, A. Krost, and E. Steimetz for sample provision. This work was supported by the Deutsche Forschungsgemeinschaft, Sfb 296, TP A4 and the SANDiE NoE of the European Commission, contract number NMP4-CT-2004-500101. Fig. 1: XSTM image of a quantum dot at VT = +2.1 V and IT = 80 pA and stoichiometry determination by local lattice constant analysis. Fig. 2: Segregation of single In atoms during overgrowth, XSTM image taken at VT = +2.2 V and IT = 80 pA. 10 International Conference on the Formation of Semiconductor Interfaces (ICFSI-10) Aix-en-Provence, France, July 3-8, 2005 Friday 4 Fr.A-9.35 0 MODELING OF ORGANOSILANE SELF-ASSEMBLED MONOLAYERS Hideaki Yamamoto, Katsuhiko Nisiyama, Takanobu Watanabe, Kosuke Tatsumura, Iwao Ohdomari, 1 School of Science and Engineering, Waseda University, 3-4-1 Ohkubo, Shinjuku-ku, Tokyo 169-8555, Japan 2 PRESTO, Japan Science and Technology Agency, 4-1-8 Honcho, Kawaguchi-shi, Saitama 322-0012, Japan 3 Kagami Memorial Laboratory for Materials Science and Technology, Waseda University, 2-8-26 Nishi-waseda, Shinjuku-ku, Tokyo 169-0051, Japan Organosilane self-assembled monolayers (SAMs) are used in various research areas, such as electronics and biology, as a powerful method of surface modification. However, unlike the structure of the alkylthiol SAMs on gold, which has 2D periodicity, the atomic scale understanding of the structure of the organosilane SAMs on amorphous substrates has not been achieved so far. The only information available is that there exists both covalent bonds (Si-O-Si) and hydrogen bonds (Si-OH:::OH-Si) in the polysiloxane layer formed by the head groups (Si(OH)3) of the constituent molecules [1]. Our goal is to find a stable structure by molecular simulation and to propose a structure model of the organosilane SAMs. AMBER 8 [2] was employed for the modeling of the SAMs. We first developed a set of parameters needed to simulate organosilanes, because the software, originally designed for simulating biomolecules, contains no Si parameters. Parameters for the AMBER potential were fitted to the results of the ab-initio calculation with Gaussian 03 [3]. We then prepared the initial structures for the simulation. Our initial structures are based on the model proposed by Helmy [1], which contains both covalent bonds and hydrogen bonds in the polysiloxane layer. The alkyl chains of the molecules were initially placed in the 2D hexagonal lattice, also based on the proposal of Helmy et al. We performed geometry optimization calculations at various “hydrolysis rates”, the rate of the number of hydrogen bonds against the number of siloxane bonds in the polysiloxane network. The figures below are the optimized structures at each hydrolysis rate; 0% hydrolysis suggests that the polysiloxane network is completely governed by the covalent Si-O-Si bonds and 100% suggests that no Si-O-Si bond exists. We are currently modeling alkylsilane SAMs with 18 carbon atoms. The contraction of the polysiloxane layer, resulting from the covalent bonds and hydrogen bonds, and the steric hindrance of the alkyl chains balance out at optimized structures. The in-plane distribution of the molecules is nonuniform with some vacancies, and their size differs depending on the hydrolys [1] R. Helmy, R.W. Wenslow, and A.Y. Fadeev J. Am. Chem. Soc. 126, (2004), 7595. [2] D.A. Case et al. AMBER 8; University of California, San Francisco, (2004). [3] M.J. Frisch et is rate. This work has been supporten by a Grant-in-Aid for COE Research and the 21 century COE program “Center for the Practical Nano-Chemistry” from the MEXT, Japan, and by a grant from JSTPRESTO. al. Gaussian 03, revision B.05; Gaussian, Inc.: Pittsburgh, PA, (2003). Figure: Stable structures of SAMs at each hydrolysis rate (top view) 10 International Conference on the Formation of Semiconductor Interfaces (ICFSI-10) Aix-en-Provence, France, July 3-8, 2005 Friday 5 Fr.A-9.55 0 Semiconductor Surface Chemical Functionalization of Alkyl Monolayers using Gas Phase Processes. Sandrine Rivillon, Yves J. Chabal Department of Chemistry and Biomedical Engineering Laboratory for Surface Modification, 136 Frelinghuysen Road, Piscataway, NJ 08854-8019 Controlling the surface chemistry of semiconductor surfaces is critical and particularly important for hybrid devices (organic/inorganic, biological/inorganic) and micromechanical systems (MEMS). Fabrication of such systems requires the formation of a strong substratecarbon bond to the surfaces (for example Si-C bond for silicon substrates) and the uniform arrangement of the monolayer. Silicon is used here due to its importance in the microelectronic industry and the stability and electronic properties of Si-C bond. Chlorination of H-terminated surfaces is a powerful way to functionalize hydrogenterminated silicon surface (H/Si) and to grow alkyl-terminated surfaces without formation of silicon oxide. We have recently investigated three different methods for chlorinating Si surfaces (gas phase, wet chemistry and photochlorination) using infrared absorption spectroscopy (IRAS) and showed that they all could lead to the same atomically flat monochlorinated Si(111) surface. To form a Si-C bond, we use wet chemical method through the use of Grignard reagents whereby chlorine atoms react with the metallic group (Mg) leaving the alkyl chains strongly linked to the atomically flat silicon surface. Using a transmission geometry for IRAS, we examine the full alkyl-terminated Si(111) surface. Fig. 1. shows that all the expected vibrational modes for Si/C≡H3 (symmetric and asymmetric stretch at 2856 and 2927 cm , umbrella at 1258 cm, and rocking at 754 cm). Importantly, we also observe the relatively sharp Si-C stretch mode at 627 cm. The results from this methylated Si surface show that the substitution is complete, i.e. not limited by steric hindrance; silicon atoms on the surface have a center-to-center distance of 3.7 Å, which is larger than the Van der Waals diameter of a methyl group (2.2 Å). Polarization studies are also performed and confirm that the methyl groups are oriented perpendicular to the surface, with Si-C stretch and CH3 symmetric stretch modes perpendicular to the surface. The kinetic of formation as well as the stability of these monolayers varying the length of the carbon chains are also being investigated. Finally, we are using again the chlorination method, this time to chemically modify the alkyl layer terminal groups, in an attempt to attach a second organic layer with different properties [1] S. Rivillon, F. Amy, Y. J. Chabal, M. M. Frank, Appl. Phys. Lett., 85, (2004) 2583 600 80