Symposium Organizers
Leonid Tsybeskov New Jersey Institute of Technology
David J. Lockwood National Research Council
Christophe Delerue IEMN
Masakazu Ichikawa The University of Tokyo
Anthony W. van Buuren Lawrence Livermore National Laboratory
L1: Light Emission and Photonic Devices I
Session Chairs
Harry Atwater
David Lockwood
Monday AM, November 27, 2006
Room 207 (Hynes)
9:15 AM -
Opening Remarks
9:30 AM - **L1.1
Group IV Semiconductor Light Emitting Nanostructures: Which can be Bright? Which can Yield Gain?
Harry Atwater 1
1 Applied Physics, California Institute of Technology, Pasadena, California, United States
Show AbstractIn this talk, I will describe progress in light emission from both Si and non-Si Group IV semiconductor nanostructures and alloy materials with potential for visible and near-infrared imaging emission. Column IV elements can be combined to form alloys (e.g., SixSn1-x and GexSn1-x) with tunable infrared optical properties in the infrared, and have yielded the first known example of a direct bandgap Column IV semiconductor. Progress on GexSn1-x alloy growth and key requirements for light-emitting devices will be discussed. Plasmon-enhanced photoluminescence and electroluminescence has been observed from silicon quantum dot (QD) emitters interacting with the enhanced local field when in close proximity to noble metal nanostructures. This enhanced luminescence is attributed to an increase in quantum efficiency, radiative decay rate, and absorbance cross section. Moreover, the enhancement is a resonant process, and this frequency-specific interaction can be used to tune the spectral emission of the Si QDs.Metal-oxide semiconductor (MOS) structures containing Si nanocrystals that exhibit field effect electroluminescence have potential to enable electrically-pumped Si-based optical gain media based nc-Si:Er media. Design requirements for achieving net gain in electrically-pumped nc-Si:Er and recent experimental results will be discussed.
10:00 AM - L1.2
Experimental Measurement of the Dielectric Constant of Nanometer-size Silicon Structures
Han Yoo 1 , Rishi Krishnan 2 , Christopher Striemer 2 , Philippe Fauchet 2
1 Department of Physics and Astronomy, University of Rochester, Rochester, New York, United States, 2 Department of Electrical and Computer Engineering, University of Rochester, Rochester, New York, United States
Show AbstractDuring the design of devices using Si nanostructures, it is often important to precisely know the dielectric function ε, since it determines many of their electrical and optical properties. Several theoretical studies [1 – 5] have predicted a reduction in ε as the nanostructure size decreases. However different physical mechanisms have been proposed for the reduction. In one model, the decrease is due to quantum confinement effects [1 – 3], whereas a more recent model attributes it to a breaking of polarizable bonds—a surface effect [4, 5].
There have been only a few experimental works on the size dependence of ε [6, 7]. In these studies ε was measured only for one particular size and no confirmation of a particular theory was possible. In our work, we have measured the size-dependent ε of both silicon nanocrystals and thin crystalline slabs at different sizes using spectroscopic ellipsometry from 0.73 eV (1700 nm) to 4.58 eV (270 nm).
In one of the two types of samples we studied, alternating layers of a-Si and a-SiO2 were deposited on a Si substrate by RF magnetron reactive sputtering. The thickness of the a-SiO2 layer was kept fixed at 5 nm, whereas the a-Si layer thickness was varied from 1 nm to 15 nm. Rapid thermal annealing initiated crystallization of the a-Si layers and furnace annealing completed it, creating a dense array of nc-Si [8]. In the second type, the thickness of the top Si layer of SOI wafers was first reduced to ~ 13 nm by a wet thermal oxidation process and the oxidized Si layer was removed by HF etching. The Si layer was then thinned down by repeated oxidation and etching to produce nano-sized Si slabs of different thicknesses. Ellipsometry and surface profile measurements were performed between each etching step.
We focused on ε in the transparent region. At 1.7 µm wavelength where the bulk ε ~ 11.7, we found that ε was reduced by ~ 40% (from 11.6 to 7.0) in the case of dense arrays of nc-Si with 2.6 nm diameters, whereas it was reduced by ~ 36% (from 11.7 to 7.5) for a 2.4 nm thick Si slab. For the nano-sized slabs, the decrease in ε with size was gradual and could be said to follow the model presented in reference [1]. For the dense arrays of nc-Si, the decrease was initially slower until around a size of 2.7 nm, when the decrease became much more rapid, which might be explained by a recent model [5]. Our results represent the first systematic measurement of the dielectric function of Si nanostructures as a function of size and represent the first test of theory.
[1] L, Wang and A. Zunger, Phys. Rev. Lett. 73, 1039 (1994); [2] M. Lannoo, et. al., Phys. Rev. Lett. 74, 3415 (1995); [3] R. Tsu, et. al., J. Appl. Phys. 82, 1327 (1997); [4] C. Delerue, et. al., Phys. Rev. B 68, 115411 (2003); [5] C. Delerue and G. Allan, Appl. Phys. Lett. 88, 173117 (2006); [6] C. Ng, et. al., Appl. Phys. Lett. 88, 063103 (2006); [7] K. Lee, et. al., Thin Solid Films 476, 196 (2005); [8] G. Grom, et. al., Nature 407, 358 (2000)
10:15 AM - L1.3
Effect of Surface Termination on Electronic Structure of Silicon Nanoparticles.
April Montoya Vaverka 1 2 , Robert Meulenburg 2 , Trevor Willey 2 , Subhash Risbud 1 , Louis Terminello 2 , Anthony van Buuren 2 3
1 Chemical Engineering and Materials Science, University of California, Davis, Davis, California, United States, 2 Chemistry and Materials Science, Lawrence Livermore National Laboratory, Livermore, California, United States, 3 , University of California, Merced, Merced, California, United States
Show Abstract10:30 AM - **L1.4
Quasi-Direct Transition due to Proximity Effects in GaSb-Si Anti-Electron Type-II Quantum Dots.
Susumu Fukatsu 1 2
1 Graduate School of Arts and Sciences, University of Tokyo, Meguro, Tokyo, Japan, 2 PRESTO, Japan Science and Technology Agency, Saitama Japan
Show Abstract To create an efficient Si-based light emitter is essential to the realization of Si-photonics and has been challenged over decades. To this end, many approaches have been proposed and extensively studied: these include π-Si and Si-nanocrystals emitting in the visible, and column-IV-based nanostructures such as Ge quantum dots and short-period Si/Ge superlattices emitting in the near-infrared. However, there has appeared no successful demonstration of either efficient light emission or gain due to interband transitions over the technologically important wavelengths, 1.1-1.7µm, where Si is not strongly absorbing. The conventional wisdom to tailor the dispersion of the otherwise indirect-gap Si relies solely on the principle of the zone-folding that takes advantage of an artificial periodicity due to short-period superlattices. This has, however, turned out to be elusive for practical but technical reasons: three-dimensional superlattices are hard to achieve, and even worse an inborn tendency of spontaneous disordering due to interface mixing disrupts the superlattice coherence. Recently, we have come up with a new concept of band-gap-type conversion utilizing proximity coupling of electronic wavefuctions, as opposed to standard band-gap engineering. The idea is to utilize quantum mechanical tunneling of electron wave function of indirect Si into the forbidden gap of direct-gap QDs standing as an anti-electron barrier. This takes effect once electrons are localized sharply at the interface. The electron trapping potential arises from anion-Si bond polarization while carrier capture over macroscopic distances is facilitated by band warping which occurs due to anisotropically distributed local strain encompassing QDs. Among allied III-V compounds, GaSb has turned out to provide the best results. As a matter of fact, efficient light emission as evidenced by 0.3-% power efficiency and room-temperature light-emitting-diode, and even near-infrared gain were obtained from 10 monolayers-equivalent GaSb-Si QDs grown by solid source molecular beam epitaxy. The proximity effects allow electrons in Si to behave more like those in direct-gap GaSb, leading to the development of on-off gain well over 10dB/cm for a 5-mm long waveguide sample in dual-chip pump-and-probe geometry both under optical and current pumping at a cryogenic temperature of 5 K. The results promise a Si-based semiconductor optical amplifier, Si-SOA. Gain saturation, probe-intensity dependent gain, and clear threshold behaviors are indicative of population-inversion in a three-level system. It has been found that as the whole system is spontaneously n-type doped, free-carrier recombination diminishes gain at a high pumping level in a slab waveguide geometry. Possible lasing is expected in view of a clear kink in the growth curve of the light output of a single chip under optical pumping, which coincides well with the kink on gain curve in the dual-chip experiment.
11:30 AM - **L1.5
Epitaxial Growth and Luminescence Characterization of Si-based Double Heterostructures Light-emitting Diodes with Iron Disilicide Active Region.
Takashi Suemasu 1 , Tsuyoshi Sunohara 1 , Yuya Ugajin 1 , Ken'ichi Kobayashi 1 , Shigemitsu Murase 1
1 Institute of Applied Physics, University of Tsukuba, Tsukuba Japan
Show AbstractSince the demonstration of EL in β-FeSi2[1], β-FeSi2 has been attracting much attention as a material for a Si-based light emitter. We have succeeded in obtaining 1.6µm EL at RT from β-FeSi2 particles embedded in Si p-n diodes on Si(001)[2-4], and very recently from p-Si/β-FeSi2/n-Si (SFS) double heterostructures (DH) on Si(111) by MBE [5]. In this paper, we report recent experimental results on the formation of SFS DH on both Si(001) and Si(111) substrates and their PL and EL properties.A SFS DH was prepared as follows. For Si(001) substrates, 8-nm-thick [100]-oriented β-FeSi2 was grown by reactive deposition epitaxy, followed by undoped Si by MBE. For Si(111) substrates, 250-nm-thick [110]-oriented β-FeSi2 was grown by codeposition of Fe and Si on RDE-grown β-FeSi2 template, followed by undoped Si by MBE. For fabrication of LEDs, a p+-Si top layer was grown. Details of the growth procedure have been already described in our previous papers. 1.6µm EL was realized at RT from SFS DH LEDs formed on both Si(001) and Si(111). For LEDs on Si(001), the current density necessary for EL at RT, which was approximately 20A/cm2, was suppressed by a factor of 3 compared to previous LEDs on Si(001) with β-FeSi2-particles active region. Time-resolved PL measurements elucidated that the luminescence originated from two sources, one with a short decay time (10 ns) and the other with a longer decay time (100 ns) at 8K. In contrast, a short decay time (10 ns) was found to be dominant for previous Si/β-FeSi2 particles/Si(001) structures. The short decay time was thought to be due to carrier recombination in β-FeSi2. On the other hand, the long decay time was due probably to a dislocation-related D1 line in Si. The luminescence intensity ratio of β-FeSi2 to D1 line was approximately 2 at 8K, and it increased with temperature (6 at 130K) since the D1 line is more rapidly quenched. Similar results were obtained for SFS DH on Si(111). For LEDs on Si(111), EL was obtained at a current density higher than 80A/cm2. The lattice mismatch between β-FeSi2 and Si is approximately 1% and 5% for β-FeSi2 formed on Si(001) and Si(111), respectively. Thus, we think that more defects working as nonradiative recombination centers exist at the SFS on Si(111). When current passing through the defects saturated, bias current began to contribute to the radiative recombination and a reasonable EL output was obtained. Thus, by increasing carrier injection into β-FeSi2 by reducing the defect densities at the β-FeSi2/Si hetero interfaces, practical Si-based LEDs will be obtained in the near future. [1] Leong et al., Nature 387 (1997) 686., [2]Suemasu et al. , Jpn. J. Appl. Phys.39 (2000) L1013., [3] Li et al., J. Appl. Phys. 97 (2005) 043529., [4] Sunohara et al., Jpn. J. Appl. Phys. 44 (2005) 3951., [5] Takauji, et al., Jpn. J. Appl. Phys. 44 (2005) 2483.
12:00 PM - L1.6
Quantum-confinement Effect in β-FeSi2 Flat Nanoislands on Si (111) Substrates.
Yoshiaki Nakamura 1 2 , Ryota Suzuki 1 , Masafumi Umeno 1 , Sung Cho 3 2 , Nobuo Tanaka 3 2 , Masakazu Ichikawa 1 2
1 Dept. of applied physics, The University of Tokyo, Tokyo Japan, 2 CREST, Japan Science and Technology Agency, Tokyo Japan, 3 EcoTopia Science Institute, Nagoya University, Nagoya Japan
Show AbstractSemiconducting β-FeSi2 is attractive as a Si-based light emitting material for an optical fiber communication. Nanostructures of β-FeSi2 [1] have drawn much attention due to their quantum-confinement effects which can change material properties. However, the quantum-confinement effect in β-FeSi2 has not been elucidated yet. In this study, we investigated the quantum-confinement effect of individual β-FeSi2 nanoislands with quantum well structures using scanning tunneling spectroscopy (STS).Si (111) samples cleaned by flashing in the ultrahigh vacuum chamber at the base pressure of ~1×10-8 Pa were oxidized at 600°C for 10 min at the oxygen pressure of ~2×10-4 Pa to form ultrathin SiO2 films. We predeposited Si with an amount of 1 monolayer (ML) on the ultrathin SiO2 films and codeposited Fe and Si at 500°C at a stoichiometric ratio of deposition rates (~0.5). This deposition formed 3-nm hemispherical β-FeSi2 nanodots with ultrahigh density (>1012 cm-2). Annealing of the β-FeSi2 nanodots at 650°C for 30 min changed hemispherical dot shape to flat nanoisland one by diffusion and coalescence of iron silicide.STM images showed that flat nanoislands were formed with a lateral size of ~10-20 nm and a height of ~2-5 nm. 7×7 Si surfaces were observed in the surface areas among the nanoislands revealing that the SiO2 was decomposed during annealing process. RHEED patterns of the samples indicated that β-FeSi2 was epitaxially grown on Si (111). We measured the energy bandgaps of the individual hydrogen-terminated β-FeSi2 nanoislands using STS. The energy bandgaps were found to increase with the decrease in the island heights while the energy bandgaps were independent of the lateral island sizes. This indicated that the β-FeSi2 nanoislands had quantum well structures. This size dependence of the energy bandgaps was fitted with the L-2 curve based on the hard wall square potential model with well width L, where the reduced mass was adjusted to be (0.25±0.07)m0 with m0 free electron mass. This is consistent with the value (0.21–0.25m0) estimated from electron and hole effective masses calculated by Martinelli [2].In summary, we measured the energy bandgaps of individual β-FeSi2 nanoislands using STS. The island-height dependence of the energy bandgaps was explained by the quantum-confinement effect in β-FeSi2 nanoislands with quantum well structures. This work was partly supported by JSPS.KAKENHI (15201023 and 17710093).References[1] Y. Nakamura, Y. Nagadomi, S.-P. Cho, N. Tanaka, and M. Ichikawa, Phys. Rev. B 72, 075404 (2005).[2] L. Martinelli, E. Grilli, D. B. Migas, L. Miglio, F. Marabelli, C. Soci, M. Geddo, M. G. Grimaldi, and C. Spinella., Phys. Rev. B 66, 085320 (2002).
12:15 PM - L1.7
Theoretical study of Si-rich transition-metal silicides with double-graphene-like structures.
Takehide Miyazaki 1 , Toshihiko Kanayama 2
1 RICS, AIST, Tsukuba Japan, 2 ASRC, AIST, Tsukuba Japan
Show AbstractRecently, there has been an enthusiasm for fabrication of graphene[1,2]. A reason is that graphene with the thickness of a few atomic layers shows the outstanding transport properties[3,4]. Regarding this excitement about graphene, it would be very intriguing to synthesize the silicon (Si) counterpart of graphene, because the use of Si in constructing those nanometer-size objects should be much more compatible with the current VLSI technology than introduction of carbon (C) systems. However, it has been argued that it is very difficult to construct Si atoms in a single stable graphene sheet[5-8] without mixing elements other than Si such as C [9,10]. In this presentation, we propose a completely novel form of graphene-like Si nanostructure based on ab initio total-energy calculation and geometry optimization. It has a three-layer structure, where the two layers of Si atoms positioned in graphene-like geometries sandwich another layer of transition metal atoms. It is possible to tune the electronic structure of this layered material from metal to semiconductor by changing the element of the transition metal atoms. Our new material can be regarded as a Si- rich phase of transition-metal (TM) silicide with a large Si-to-TM composition ratio being around ~10, which may be suitable for a material to smoothly connect the interfaces between electrodes made of conventional TM disilicide and Si substrates in transistors. We will further discuss how the novel material in question could be synthesized.References[1] S. Horiuchi, T. Gotou, M. Fujiwara, T. Asaka, T. Yokosawa and Y. Matsui, Appl. Phys. Lett. 84, 2403 (2004).[2] J.-L. Li, K. N. Kudin, M. J. McAllister, R. K. Prud’homme, I. A. Aksay, and R. Car, Phys. Rev. Lett. 96, 176101-1 (2006).[3] K. S. Novoselov, A. K. Geimm, S. V. Morozov, D. Jiang, Y. Zhang, S. V. Dubonos, I. V. Grigorieva and A. A. Firsov, Science 306, 666 (2004).[4] K. S. Novoselov, A. K. Geimm, S. V. Morozov, D. Jiang, M. I. Katsnelson, S. V. Dubonos, I. V. Grigorieva and A. A. Firsov, Nature 438, 197 (2005).[5] M. T. Yin and M. L. Cohen, Phys. Rev. B29, 6996 (1984).[6] K. Takeda and K. Shiraishi, Phys. Rev. B50, 14916 (1994).[7] Y.-C. Wang and K. Sheerschmidt and U. Gosele, Phys. Rev. B61, 12864 (2000).[8] E. Durgun, S. Tongay and S. Ciraci, Phys. Rev. B72, 075420 (2005).[9] Y. Miyamoto and B. D. Yu, Appl. Phys. Lett. 80, 586 (2002).[10] C. L. Freeman, F. Claeyssens, N. L. Allan and J. H. Harding, Phys. Rev. Lett. 96, 066102 (2006).
12:30 PM - **L1.8
Advances in SiGeSn/Ge Technology
Richard Soref 1 , John Kouvetakis 2 , Jose Menendez 2
1 , Air Force Research Laboratory, Hanscom AFB, Massachusetts, United States, 2 , Arizona State University, Tempe, Arizona, United States
Show AbstractWe have recently reported the CVD growth of binary Ge1-ySny and ternary Ge1-ySixSny alloys directly on Si wafers using SnD4, GeH4, SiH3GeH3, and (GeH3)2SiH2 sources. Ge1-ySny is an intriguing infrared material that undergoes an indirect-to-direct bandgap transition for y > 0.09. In addition, we have found that Ge1-ySny layers have ideal properties as templates for the subsequent deposition of other semiconductors: (a) they are strain-relaxed and have low threading defect densities (105 cm-2) even for films thinner than 1 µm; (b) their low growth temperatures between 250°C and 350°C are compatible with selective growth, and the films possess the necessary thermal stability for conventional semiconductor processing (up to 750°C depending on composition); (c) they exhibit tunable lattice constants betwen 5.65 Å and at least 5.8 Å, matching InGaAs and related III-V systems; (d) their surfaces are extremely flat; (e) they grow selectively on Si and not on SiO2; and (f) the film surface can be prepared by simple chemical cleaning for subsequent ex-situ epitaxy.The incorporation of Sn lowers the absorption edges of Ge. Therefore, Ge1-ySny is attractive for detector and photovoltaic applications that require band gaps lower than that of Ge. Spectroscopic ellipsometry and photoreflectance experiments show that the direct band gap is halved for as little as y = 0.15. Studies of a Ge0.98Sn0.02 sample yield an absorption coefficient of 3500 cm-1 at 1675 nm (0.74 eV). Thus infrared detectors based on Ge0.98Sn0.02 could easily cover the U-(1565 nm-1625 nm), L-(1565 nm-1625 nm), and C-(1530 nm-1565 nm) telecomm bands. We have made advances in P and N doping of GeSn and shall present results on infrared detection using GeSn/SiGeSn PIN heterodiodes. GeSn also has application in band-to-band laser heterodiodes.The ternary system Ge1-x-ySixSny grows on Ge1-ySny-buffered Si. It represents the first practical group-IV ternary alloy, since C can only be incorporated in minute amounts to the Ge-Si network. The most significant feature of Ge1-x-ySixSny is the possibility of independent adjustment of lattice constant and band gap. For the same value of the lattice constant one can obtain band gaps differing by more than 0.2 eV, even if the Sn-concentration is limited to the range y < 0.2. This property can be used to develop a variety of novel devices, from multicolor detectors to multiple junction photovoltaic cells. A linear interpolation of band gaps and lattice constants between Si, Ge and α-Sn shows that it is possible to obtain SiGeSn with a band gap and a lattice constant larger than that of Ge. We shall use this feature to make a tensile-strained Ge-on-SiGeSn telecomm detector with improved performance. The tensile-strain-induced direct gap of Ge can be used also for electroptical modulators and lasers.
L2: Light Emission and Photonic Devices II
Session Chairs
Graham Reed
Richard Soref
Monday PM, November 27, 2006
Room 207 (Hynes)
2:30 PM - **L2.1
Sub-micron Silicon Photonic Devices.
Graham Reed 1 , Goran Mashanovich 1 , Frederic Gardes 1 , Branislav Timotijevic 1 , William Headley 1
1 , University of Surrey, Guildford United Kingdom
Show Abstract3:00 PM - L2.2
Fabrication of Extreme aspect Ratio Tubes and Wires of Silicon and Germanium Within Microstructured Optical Fibers.
Neil Baril 1 4 , John Badding 1 4 , Vankatraman Gopalan 3 4 , Pier Sazio 5 , Thomas Scheidemantel 2 4 , Bryan Jackson 3 4 , Dong-Jin Won 3 4 , Adrian Amezcua Correa 5 , Chris Finlayson 5
1 Chemistry, The Pennsylvania State University, University Park, Pennsylvania, United States, 4 The Center for Nanoscale Science, The Pennsylvania State University, University Park, Pennsylvania, United States, 3 Materials Science and Engineering, The Pennsylvania State University, University Park, Pennsylvania, United States, 5 Optoelectronics Research Centre, University of Southampton, Southampton United Kingdom, 2 Physics, The Pennsylvania State University, University Park, Pennsylvania, United States
Show AbstractOptical fibers are a natural conduit for light and semiconductors are the basis of modern optoelectronics. We have recently fabricated semiconducting micro- and nanowires with extreme aspect ratios in ordered arrays using microstructured optical fibers (MOF’s), as templates. These systems contain the highest aspect ratio semiconducting micro- and nanowires yet produced by any method. The micro wires are over a meter in length, and the nanowires are centimeters long and ~100nm in diameter. The technique for infiltration uses high-pressure gasses to carry chemical vapor deposition precursors through the capillaries of the MOFs, which are heated in a tube furnace to deposit the semiconductor materials. The semiconductors are deposited amorphous and are annealed after deposition to produce polycrystalline materials. MOFs are versatile templates allowing the deposition of high density arrays of micro to nano sized capillaries. They are optically transparent allowing in situ characterization of the wires with Raman spectroscopy. The semiconductor wires can be easily manipulated because they are encased in a silica optical fiber, the silica can also be removed giving access to the wires. The possibilities for the fabrication of in-fiber optoelectronic devices are numerous. These structures have many potential applications for in-fiber sensing, light modulation, and light generation.Sazio et al. Science 311, 1583 (2006)
3:15 PM - L2.3
Large Scale Formation of Y2SiO5:Er Oxyorthosilicate Nanocrystals using Si Nanowires for Efficient, High-gain Light Emitting Material at 1.5 μm.
Kiseok Suh 1 , Jung Shin 1 , Byeong-Soo Bae 2
1 Physics, KAIST, Daejeon Korea (the Republic of), 2 Materials Science, KAIST, Daejeon Korea (the Republic of)
Show Abstract4:30 PM - **L2.4
Silicon Integrated Nanophotonics - Advances and Challenges.
Yurii Vlasov 1 , Fengnian Xia 1 , Lidija Sekaric 1 , Eric Dulkeith 1 , Solomon Assefa 1 , William Green 1 , Martin O'Boyle 1 , Hendrik Hamann 1 , Sharee McNab 1
1 Physical Sciences Department, IBM T.J. Watson Research Center, Yorktown Heights, New York, United States
Show Abstract5:00 PM - L2.5
Three Dimensional Silicon Inverse Photonic Quasicrystals for Infrared Wavelengths.
Alexandra Ledermann 2 , Ludovico Cademartiri 1 , Martin Hermatschweiler 3 , Martin Wegener 2 , Geoffrey Ozin 1 , Diederik Wiersma 4 , Georg von Freymann 3
2 , Institut für Angewandte Physik, Karlsruhe Germany, 1 Department of Chemistry, University of Toronto, Toronto, Ontario, Canada, 3 , Institut für Nanotechnologie, Karlsruhe Germany, 4 , European Laboratory for Nonlinear Spectroscopy, Firenze Italy
Show AbstractPhotonic quasicrystals pose numerous challenges and promises both from the theoretical and the technological point of view. For example it is impossible to completely predict the optical properties of three-dimensional (3d) photonic quasicrystals for their lack of translational periodicity, most fabrication techniques available today are planar and so are optimized for fabrication of two-dimensional structures and only of limited use for arbitrary 3d structures.One of the interesting promises of 3d photonic quasicrystals are anomalous diffusion of light due to the long range order coupled with a lack of short range order, as well as opening complete photonic bandgaps due to their nearly spherical Brillouin zone, which is a consequence of their extremely high rotational symmetry.Photonic quasicrystals might also be considered as a testbed on which to verify theories developed for atomic quasicrystals. In contrast to atomic quasicrystals, photonic quasicrystals can be created by design. Thus, the influence of the symmetry on e.g. transport properties can be monitored more easily.We will here present results on the fabrication of oriented photonic quasicrystals in a polymeric photoresist via direct laser writing and their subsequent inversion with Si.The quality of the structures will be demonstrated via scanning-electron and laser-confocal microscopy while the preservation of the symmetry will be demonstrated with Laue diffraction experiments and their comparison with theory.
5:30 PM - L2.7
Ultra-high Resolution Imaging of Highly Confined Optical Modes in Sub-micron Scale SOI Waveguides.
Jacob Robinson 1 , Stefan Preble 1 , Michal Lipson 1
1 Electrical and Computer Engineering, Cornell , Ithaca, New York, United States
Show Abstract5:45 PM - L2.8
Si3N4-SiO2-Si Slot-waveguide Disk Resonators in a Silicon Photonic Platform.
Bradley Schmidt 1 , Carlos Barrios 2 , Michal Lipson 1
1 Electrical and Computer Engineering, Cornell University, Ithaca, New York, United States, 2 Instituto de Sistemas Optoelectrónicos y Microtecnología, Ciudad Universitaria, Madrid Spain
Show Abstract
Symposium Organizers
Leonid Tsybeskov New Jersey Institute of Technology
David J. Lockwood National Research Council
Christophe Delerue IEMN
Masakazu Ichikawa The University of Tokyo
Anthony W. van Buuren Lawrence Livermore National Laboratory
L3: Ge and SiGe Nanostructures I
Session Chairs
Philippe Boucaud
Detlev Gruetzmacher
Tuesday AM, November 28, 2006
Room 207 (Hynes)
9:30 AM - **L3.1
Templated Selfassembly of Ge Dot Arrays, Molecules and 3-dimensional Crystals on Si.
Detlev Grutzmacher 1 , Christian Dais 1 , Elisabeth Müller 1 , Harun Solak 1 , Hans Sigg 1 , Julian Stangl 2 , Günther Bauer 2
1 Laboratory for Micro- and Nanotechnology, Paul Scherrer Institut, Villigen-PSI Switzerland, 2 Inst. for semiconductor and solid state physics, Johannes Kepler University Linz, Linz Austria
Show AbstractTemplated self-organization has been used to prepare samples with regimented arrays of Ge quantum dots. Si (100) substrates have been patterned with 2-dimensional hole gratings using multiple beam diffraction in the extreme UV (EUV). The setup of EUV interference lithography (EUV-IL) at the swiss synchrotron ligth source (SLS) permits exposure areas up to 2x2 mm in a single illumination. Different gratings have been used for masks, leading to patterns with an periodicity ranging from 50 to 250 nm. After the pattern had been transferred into the Si (100) substrate by reactive ion etching, molecular beam epitaxy was employed to grow Si/Ge quantum dot layers on the pre-patterned substrates. First a 20-100 nm thick Si buffer layer was deposited at 300°C followed by the Ge deposition for island formation. In the first Ge quantum dot layer the dots align with holes provided by the pre-patterning. Here we studied the impact of the microscopic shape and size of the pre-pattern using the mask design and the XIL exposure time and dose as parameters. Very regular array of Ge dots are formed with a densities up to 2.2x1010 cm-2. The density depends not only on the periodicity of the pattern fabricated, but also on the shape of the patterned formed at various exposure conditions. The formation of square like holes on the Si substrates lead to the nucleation of ordered quantum dot molecules on the Si surface. Typically 4 Ge dots are formed, nucleating at the corners of the square holes of the prepattern. Lowering the exposure dose and thus decreasing the size of the holes in the prepattern allows the deposition of one Ge dot per hole. Atomic force microscopy has been employed to determine the size distribution of the Ge dots in the arrays. Narrow size distribution of 4% and 7% for dome and hut cluster of 80 and 40 nm diameter have been found on arrays with 250 nm and 100 nm periodicity, respectively.Continuation of depositing Si/Ge layer sequences on top of the ordered dot arrays leads to alignment of dots to the first layer due to strain fields. The alignment has been observed for quantum dot molecules as well. AFM surface scans as well as crossectional TEM micrographs reveal the formation of highly ordered 3-dimensional quantum dot crystals. Typically, the vertical periodicity amounts to ~8nm, whereas the lateral periodicity is 90 x 100 nm. The dots have a diameter of 34±3 nm, thus Ge dots exhibit a remarkably narrow size distribution and close to perfect ordering. This is confirmed by X-ray diffraction experiments at symmetric and asymmetric diffraction peaks. Moreover, photoluminescence and optically pumped absorption measurements have been performed giving insights into the bandstructure of the 2-d and 3-d quantum dot crystals. Our results on the fabrication and properties of 2- and 3-dimensional Ge quantum dot crystals may open a new routes towards the realization of nanoelectronic and spintronic devices as well as for quantum computing.
10:00 AM - L3.2
Stranski–Krastanov Growth of Tensely Strained Si on Ge (001) Substrates.
Dietmar Pachinger 1 , Gang Chen 1 , Herbert Lichtenberger 1 , Friedrich Schäffler 1
1 Semiconductor Physics, Institute of Semiconductor and solid state physics, Linz Austria
Show Abstract10:15 AM - L3.3
Electric Field Controlled Directional Growth in Metal-Induced Lateral Crystallization of Amorphous SiGe on Insulating Films.
Masanobu Miyao 1 , Hiroshi Kanno 1 , Taizoh Sadoh 1
1 Department of Electronics, Kyushu University, Fukuoka Japan
Show AbstractThe low-temperature (<500oC) formation of high quality polycrystalline SiGe (poly-SiGe) on insulating substrates has been expected to realize advanced system-in-displays. In line with this, we have been developing metal-induced lateral crystallization (MILC) of a-Si1-XGeX (X:0-1) by using Ni as catalyst metal, and found that growth morphology strongly depends on Ge fraction., i.e plane crystallization for samples with low Ge-fraction (X<0.3), dendrite growth for intermediate Ge fraction (X:0.35-0.65) and no crystallization for high Ge fraction (X above 0.7). To solve these problems, present paper examines the electric-field stimulated MILC of a-Si1-XGeX (X:0-1). This enables uniform crystal growth of SiGe with all Ge fractions.In the experiment, a-Si1-XGeX (X:0-1, 50 nm thickness) were deposited on quartz substrate using a solid-source MBE system. Then, Ni films (15 nm thickness) were deposited selectively on a-SiGe layers. This Ni films were used as the catalyst atom source and electrodes for bias voltage. The spacing between anodes and cathodes were 40~6000 micron. Finally, the samples were annealed at 400~500oC with applying electric fields (0~5000 V/cm) between the electrodes. The lateral growth lengths and crystal qualities of SiGe were evaluated by using Nomarski optical microscopy, scanning electron microscopy, and Raman spectroscopy.When the electric fields were applied (<200V/cm) during MILC, lateral growth velocity at the cathode side became faster by 10 times than that at the anode side. This indicates that Ni atoms are charged negatively in SiGe, and their migration is enhanced by electric fields. In addition, the dendrite growth obtained by conventional MILC was almost vanished away and very large uniform growth regions (>50 micron) appeared at the cathode sides even for both samples with intermediate and high Ge fractions. Raman spectroscopy measurements showed that grown layers were completely strain free. Under the extremely high electric fields (>2000 V/cm), crystal growth propagated straight from the Ni patterns, where the growth direction was completely aligned to the electric fields. These phenomena are attributed to the facts that kinetic energy of Ni atoms transferred from the high electric fields (>2000 V/cm) exceeds thermal energy at 500oC. These results indicate that flow direction of catalyst atoms (Ni) during annealing can be controlled by the electric fields. This advantage of aligned growth of poly-SiGe on the insulating films should be used for advanced TFT with high speed operation. We are now fabricating thin-film transistors (TFT) with Schottky source and drain structures. Preliminary results indicated that TFTs showed good ambipolar operation characteristics. In addition, a kink effect due to the floating body effects, which were observed in the conventional doping source/drain TFTs, was successfully suppressed. The possible application of these TFTs to advanced system-in-displays will be discussed.
10:30 AM - L3.4
Spin Relaxation in SiGe Islands.
Hans Malissa 1 , Wolfgang Jantsch 1 , Gang Chen 1 , Herbert Lichtenberger 1 , Thomas Fromherz 1 , Friedrich Schäffler 1 , Günther Bauer 1 , Alexei Tyryshkin 2 , Stephen Lyon 2 , Zbyslaw Wilamowski 3
1 Institute of Semiconductor and Solid State Physics, Johannes Kepler University Linz, Linz Austria, 2 Department of Electrical Engineering, Princeton University, Princeton, New Jersey, United States, 3 Institute of Physics, Polish Academy of Sciences, Warsaw Poland
Show Abstract10:45 AM - L3.5
Photoluminescence Excitation Dependencein Three-dimensional Si/SiGe Nanostructures.
Eun Kyu Lee 1 , Boris Kamenev 1 , Theodore Kamins 2 , Jean-Mark Baribeau 3 , David Lockwood 3 , Leonid Tsybeskov 1
1 ECE, New Jersey Institute of Technology, Newark, New Jersey, United States, 2 Quantum Science Research, Hwelett-Packard Laboratories, Palo Alto, California, United States, 3 National Research Council, Institute for Microstructural Sciences, Ottawa, Ontario, Canada
Show AbstractWe find that in Ge (SiGe) clusters grown on Si using Stranski-Krastanov (SK) growth mode, (i) photoluminescence (PL) spectra, (ii) PL lifetime and (iii) PL thermal quench activation energies exhibit strong dependence on the excitation intensity. Under PL excitation intensity increasing from 1 to 104 W/cm2, the PL spectra exhibit blue shift from below Ge bandgap up to ~ 970 meV. The PL lifetime shows strong dependence on the excitation wavelength, decreasing from 20 microseconds at ~ 0.8 eV to 200 nanoseconds at ~ 0.9 eV. The process of PL thermal quench has two clearly distinguished activation energies. At low temperature, small (~ 15 meV) and excitation-independent activation energy is attributed to exciton thermal dissociation. At higher temperature, excitation-dependent PL thermal quench activation energy (increasing from ~ 120 to 340 meV as excitation intensity increases) is found, and it is attributed to hole redistribution via tunneling and/or thermal ionization over the Ge (SiGe)/Si valence band confinement barrier.
11:30 AM - **L3.6
Ge/Si self-assembled Islands for Photonics Applications.
Philippe Boucaud 1 , Xiang Li 1 , Moustafa El Kurdi 1 , Sébastien Sauvage 1 , Xavier Checoury 1 , Sylvain David 1 , Navy Yam 1 , Frédéric Fossard 1 , Daniel Bouchier 1 , Guy Fishman 1
1 , CNRS-IEF, Orsay France
Show Abstract12:00 PM - L3.7
Ordering and Shape Tuning of Ge Islands on Metal-patterned Si.
Jeremy Robinson 1 2 , Donald Walko 3 , Dohn Arms 3 , Daniel Tinberg 4 , Paul Evans 4 , Yifan Cao 1 , J. Liddle 2 , Oscar Dubon 1 2
1 Materials Science and Engineering, University of California, Berkeley, California, United States, 2 Materials Science Division, Lawrence Berkeley National Laboratory, Berkeley, California, United States, 3 Advanced Photon Source, Argonne National Laboratory, Argonne, Illinois, United States, 4 Materials Science and Engineering, University of Wisconsin, Madison, Wisconsin, United States
Show Abstract12:15 PM - L3.8
Photo-oxidation of Ge Nanocrystals: Kinetic Measurements by In Situ Raman Spectroscopy
Ian Sharp 1 2 , Qing Xu 1 2 , Chun Yuan 1 2 , Joel Ager III 1 , Daryl Chrzan 1 2 , Eugene Haller 1 2
1 Materials Sciences Division, Lawrence Berkeley National Laboratory, Berkeley, California, United States, 2 Materials Science and Engineering Department, University of California, Berkeley, Berkeley, California, United States
Show AbstractGe nanocrystals are formed in silica by ion beam synthesis and are subsequently exposed by selective HF etching of the silica. Under ambient conditions, the exposed nanocrystals are stable after formation of a protective native oxide shell of no more than a few monolayers [1]. However, under visible laser illumination at room temperature and in the presence of O2, the nanocrystals rapidly oxidize. The oxidation rate was monitored by measuring the Raman spectra of the Ge nanocrystals in situ. The intensity ratio of the anti-Stokes to the Stokes line indicated that no significant laser-induced heating of illuminated nanocrystals occurs. Therefore, the oxidation reaction rate enhancement is due to a photo-chemical process. Under certain conditions laser illumination can lead to complete oxidation of nanocrystals, whereas those that are not illuminated are stable after formation of a thin native oxide. The oxidation rate varies linearly with the logarithm of the laser intensity, and at constant laser intensity the rate increases with increasing photon energy. These kinetic measurements, along with the power and energy dependencies, are described quantitatively by an electron active oxidation mechanism involving tunneling of optically excited electrons through the forming oxide skin and subsequent transport of oxygen ions to the Ge nanocrystal surface. This work was supported in part by the U.S. Department of Energy under Contract No. DE-AC02-05CH11231 and in part by U.S. NSF Grant No. DMR-0405472.[1] Sharp, I.D. et al., J. Appl. Phys. 97, 124316 (2005).
12:30 PM - L3.9
Influence of Nanoimprinted and Etched Surface Relief on Nucleation and Ordering of Si and Ge on Amorphous Silicon Dioxide.
Ted Kamins 1 , Amir Yasseri 1 , Shashank Sharma 1 , Fabian Pease 2 , Qiangfei Xia 3 , Stephen Chou 3
1 Quantum Science Research, Hewlett-Packard Laboratories, Palo Alto, California, United States, 2 Dept. of Electrical Engineering, Stanford University, Stanford, California, United States, 3 Nanostructure Lab, Dept. of Electrical Engineering, Princeton University, Princeton, New Jersey, United States
Show Abstract12:45 PM - L3.10
Control of Valley Splitting in a Si/SiGe 2DEG Quantum Point Contact.
Lisa McGuire 1 , K. Slinker 1 , S. Goswami 1 , J. Chu 2 , Mark Friesen 1 , S. Coppersmith 1 , M. Eriksson 1
1 , University of Wisconsin, Madison, Wisconsin, United States, 2 , IBM Research Division, T. J. Watson Research Center, Yorktown Heights, New York, United States
Show AbstractValley splitting in Si quantum wells is an important component of Si-based quantum computing. If valley splitting is large, the spin degree of freedom forms the basis for a robust two-level qubit. If valley splitting is small, different valley states will interfere with this qubit basis. Here we report the measurement and control of valley splitting in a Si/SiGe quantum point contact. In contrast with previous results on large Hall bars, which have shown small valley splitting, we find that the valley splitting in a quantum point contact is large, approaching the maximum theoretical estimates1. To measure the valley splitting, we perform a point contact spectroscopy by recording the point contact conductance as function of magnetic field and gate voltage at an electron temperature T~100 mK. The results show that the point contact geometry enables control of the valley splitting as a function of both gate voltage and magnetic field in individual subbands. Research supported by ARO, NSA, and NSF. 1Valley splitting in strained silicon quantum wells, Timothy B. Boykin, Gerhard Klimeck, M.A. Eriksson, Mark Friesen, S. N. Coppersmith, Paul von Allmen, Fabiano Oyafuso, and Seungwon Lee, Appl. Phys. Lett. 84, 115 (2004).
L4: Ge and SiGe Nanostructures II / MEMS and Strained Si
Session Chairs
Thomas Fromherz
Avi Kornblit
Tuesday PM, November 28, 2006
Room 207 (Hynes)
2:30 PM - **L4.1
Optoelectronic Properties and Bandstructure of SiGe Quantum Dot and Cascade Structures.
Thomas Fromherz 1 , Moritz Brehm 1 , Patrick Rauter 1 , Nguyen Vinh 2 , Ben Murdin 3 , Jonathan Phillips 4 , Carl Pidgeon 4 , Zhenyang Zhong 1 , Gang Chen 1 , Jiri Novak 1 , Julian Stangl 1 , Detlev Gruetzmacher 5 , Guenther Bauer 1
1 Semiconductor Physics, University Linz, Linz Austria, 2 , FOM Institute for Plasma Physics Rijnhuizen , Nieuwegein Netherlands, 3 , University of Surrey, Guildford United Kingdom, 4 , Heriot-Watt University, Edinburgh United Kingdom, 5 , Paul Scherrer Institut, Villigen Switzerland
Show AbstractDue to its indirect fundamental bandgap in k-space, bulk silicon - the dominating material for microelectronics - is not suitable for optoelectronic applications. Nevertheless, the demand for processing and transmitting a large amount of data in very short times is steadily increasing and intra-chip optical communication links will become more and more important in the near future. Evidently, a Si based optoelectronic platform compatible with modern CMOS technology is highly desirable. While several levels of integrating electronics and optics are followed in research, for wide-spread applications in consumer electronics the most attractive and probably cost-efficient level would be a monolithic optoelectronic integration.The central building block of any Si based optoelectonic platform is an electrically pumped emitter. In order to circumvent small optical matrix elements typically for interband transitions in an indirect semiconductor, large efforts are devoted to the development of a SiGe based quantum cascade emitter, since in these unipolar devices, no transitions across the fundamental bandgap are involved in the generation of light. After an initial rapid success in the demonstration of electroluminescence emitted by quantum cascade (QC) structures, lasing has still to be demonstrated. For a proper design of laser structures, a critical parameter is the excited state lifetime of the laser transition. In our work, we have for the first time directly measured the optical-phonon limited heavy hole (HH) excited state lifetime of a quantum well (QW) sequence designed for MIR emission. For these measurements, time resolved pump-pump photocurrent experiments under various applied biases have been performed. A lifetime of 550 fs was determined, significantly longer than reported in literature up to now.Another promising type of light emitters in the SiGe system are quantum dots (QDs). It is well known, that due to the strain field originating from the QDs, not only holes, but also electrons are bound to the QDs. Due to this confinement, the k-space selection rules are relaxed in all three dimensions. However, while the holes are bound to the Ge rich region in the interior of the QDs, the electrons are localized in then Si matrix along the surface of the QDs. Thus a spatial indirect (type II) band alignment results and optical interband matrix elements are expected to be small. In our work, we have analyzed the bandstructure and its dependence on the structural parameters of the QDs (size, Ge concentration, Ge gradient) by simulations using the nextnano3 code. By comparing the results of the calculations to photoluminescence (PL) measurements, we try to identify the transitions observed in the experiments. Based on this analysis, strategies for enhancing the PL efficiency of QDs (like for example the control of the Ge gradient within the dots) are discussed.
3:00 PM - L4.2
WITHDRAWN 11/16/06 Design Criteria for p-type SiGe Bound-to-Continuum THz QCLs.
Marco Califano 1 , Zoran Ikonic 1 , Robert Kelsall 1 , Paul Harrison 1
1 Electronic & Electrical Engineering, University of Leeds, Leeds United Kingdom
Show AbstractTuesday 11/28Withdrawnoral2:00 pm L4.2
3:15 PM - L4.3
Luminescence in Multilayers of SiGe Nanocrystals Embedded in SiO2.
M. Avella 1 , A. Prieto 1 , Juan Jimenez 1 , A. Rodríguez 2 , J. Sangrador 2 , T. Rodríguez 2 , M. Ortíz 3 , C. Ballesteros 3 , A. Kling 4
1 Dpto. Física de la Materia Condensada, Universidad de Valladolid, Valladolid Spain, 2 Dpto. de Tecnología Electrónica, Universidad Politécnica de Madrid, Madrid Spain, 3 Dpto. de Física, Universidad Carlos III, Madrid Spain, 4 , Instituto Tecnológico e Nuclear, Sacavém Portugal
Show AbstractNanoparticles of the Group IV semiconductors embedded in a dielectric SiO2 matrix have received a great deal of attention because of their potential applications in optoelectronic devices and non-volatile memories. Several approaches have been tried to make nanocrystalline structures inside the dielectric matrix, like ion implantation, sputtering or Chemical Vapour Deposition. Usually, Si or Ge nanocrystals are formed. However, previous attempts to prepare arrays of SiGe nanocrystals revealed that the diffusion of Ge constitutes a drawback to control the nanocrystal composition. Intense luminescence is required for light emission applications; therefore, a high density of nanocrystals is desirable. For these purposes we have fabricated multilayer arrangements of SiGe nanocrystals, obtained by rapid thermal crystallization (temperatures between 750 and 1100°C, and times up to 10 minutes) of amorphous nanoparticles embedded in a matrix of SiO2 deposited by low pressure chemical vapor deposition (LPCVD). This process is fully compatible with the CMOS technology. The structures were characterized by Raman spectroscopy, Rutherford Backscattering Spectrometry, Transmission Electron Microscopy, Fourier transform infrared spectroscopy and Cathodoluminescence. The effect of the annealing conditions and the thickness of the oxide interlayers on the luminescence intensity and on the variation of the composition of the nanocrystals due to the Ge diffusion were studied. The contributions of the SiO2 matrix and the SiGe nanoparticles to the luminescence were identified and separated. The maximum intensity of the luminescence has been obtained after annealing the samples at 900°C. At this temperature the SiGe nanoparticles are fully crystallized and no appreciable variation of their composition has been detected. The luminescence of the structures increases: a) with the number of periods in the structure and b) with the thickness of the SiO2 interlayers. The luminescence emission was compared to our previous results for pure Ge nanocrystals, showing that the light emission mechanisms are similar for both SiGe and Ge nanocrystals.
3:30 PM - L4.4
Ordering of Strained Ge Islands on Prepatterned Si(001) Substrates: Morphological Evolution and Nucleation Mechanisms
Gang Chen 1 , Herbert Lichtenberger 1 , Zhenyang Zhong 1 , Günther Bauer 1 , Wolfgang Jantsch 1 , Friedrich Schäffler 1
1 Institute of Semiconductor and Solid State Physics , Johannes Kepler University, Linz Austria
Show AbstractGe on Si(001) is considered a model system for studying the basic properties of strain-driven (S-K) 3D island growth, but also a promising material system for electronic and optoelectronic applications. Most recently, laterally ordered Ge quantum dots were proposed as building blocks for quantum computing and quantum information storage. Both applications require perfectly ordered Ge dots to allow external addressing. In previous work, we have realized perfectly ordered SiGe and Ge island arrays by preferential nucleation on nanostructured substrates [1]. Under optimized growth conditions the islands nucleate at the bottom of reactively ion etched pits, which assume the shape of truncated inverted pyramids with sidewall inclinations of around 8° after overgrowth with a thin Si buffer layer. This nucleation site is rather surprising, because it appears to be the least favorable site for strain relaxation within the pit template. To gain better understand of the underlying mechanisms, we studied the very early stages of Ge coverage [2]. The experiments show that the Ge wetting layer develops a complex, but highly symmetric morphology on the inclined sidewalls of the pits. This pattern is driven by strain- and surface energy minimization, and leads after the deposition of typically three monolayers of Ge to a conversion of the pit sidewalls into trains of radially oriented prisms of triangular cross section and surface termination by two adjacent {105} facets. [3] In the corner regions of the pits, (001) terraces form were two of these prisms from adjacent sidewalls intersect. This way, the pit surface becomes entirely covered by {105} and (001) facets. We attribute the subsequent island nucleation to Ge accumulation at the bottom of the pits, which is driven by capillarity and the enhanced surface diffusion on the by now {105} faceted sidewalls of the pits. We systematically varied the growth conditions, and found that the morphological pattern of the wetting layer in the pits is stable over a growth temperature range of at least 100°C. We also increased the diameter of the etch pits until neighboring pits overlap. This leaves a hill-dominated morphology on the substrate, which evolves during wetting layer growth into an entirely equivalent, but now convex, morphology. We therefore conclude that the morphological evolution of a Ge wetting layer on pit or hill patterns is a generic feature that develops over a wide range of growth conditions. It can be ascribed to a combination of step meandering on the inclined sidewalls, and step bunching in the corner regions of the pits or hills. The simultaneous occurrence of these two mechanisms is mainly driven by low-energy facet formation in combination with the confined geometry of the employed templates. [1] Z. Zhong et al., J. Appl. Phys. 93, 6258 (2003)[2] G. Chen, et al., cond-mat/0602175[3] H. Lichtenberger et al., Appl. Phys. Lett. 86, 131919 (2005)
4:15 PM - **L4.5
Challenges in the Fabrication of Advanced MEMS.
Avi Kornblit 1 , Vladimir Aksyuk 1 , Cristian Bolle 1 , J. Bower 1 , Raymond Cirelli 1 , Harold Dyson 1 , Ed Ferry 1 , Linus Fetter 1 , Robert Fullowan 1 , Arman Gasparyan 1 , Christopher Jones 1 , Robert Keller 1 , Fred Klemens 1 , Warren Lai 1 , O. Lopez 1 , William Mansfield 1 , John Miner 1 , Chien-Shing Pai 1 , Flavio Pardo 1 , Maria Simon 1 , Thomas Sorsch 1 , J. Taylor 1 , Donald Tennant 1 , Brijesh Vyas 1 , George Watson 1
1 , Lucent Technologies Bell Labs, Murray Hill, New Jersey, United States
Show AbstractAs higher performance and complexity is demanded from advanced MEMS, the processes required to fabricate them are becoming much more demanding. Although many techniques used in IC manufacturing can be adapted to MEMS fabrication, the films used could be considerably thicker. The potential stress associated with these films, besides having an impact on the individual device, can lead to a significant wafer-bow, far exceeding the specifications of advanced semiconductor processing tools. Stress management is now one of the most important requirements in advanced MEMS fabrication.Advanced MEMS, in addition to having sub-micron features, are highly complex devices with a large number of elements. Step and repeat (or step and scan) cameras can deliver the required dimensions, but they are limited by their field size. When large-area devices are needed, the repeated patterns can be tiled or stitched, thus enabling the realization of large MEMS with sub-micron features. Since the photoresists used by these advanced tools are sometimes too thin for the long dry etching steps, sacrificial intermediate layers are used as hard masks.Deep silicon etching (or bulk micromachining) is challenging in terms of dimension, profile and uniformity control, and the realization of high etching rates. Fluorine based chemistries are commonly used to achieve high etching rates, and the profile is controlled either by cryogenic or switched etching processes. In the former the etched profile sidewall is passivated during the etch process, while in the latter (known also as the ‘Bosch process’), sequential etching and deposition steps are used to achieve the same result. Small dimension, high-aspect-ratio structures are common in advanced MEMS devices. In addition to patterning and etching these challenging structures, void-free deposition of dielectrics and polysilicon (and in some cases metal) is needed. There are numerous ways to deposit theses films, which in many instances are sacrificial. Similar films that have been used in the past in the IC industry for gap-fill, can be used for MEMS fabrication as well. In some cases these layers have to be planarized, requiring a challenging chemical-mechanical-polishing step, with high uniformity and little or no ‘dishing.’ Since some of the structures contain cantilevers or mirrors that have to be flat, film stress control is an absolute must. Structures made of multiple films (e.g. silicon mirrors coated with metal) have to be flat not only after fabrication, but must remain flat after being placed in a system, and are expected to remain flat and function reliably for many years.MEMS release is another challenging aspect of fabrication advanced MEMS, both in terms of avoiding stiction, and avoiding attack of metals that are present during the release process. A number of solutions, using both wet and dry chemistries are available for this process.All of the above issues and ways to address them will be discussed.
4:45 PM - L4.6
Piezoresistance in Strained Silicon and Strained Silicon Germanium.
Jacob Richter 1 , M. Arnoldus 1 , A. Larsen 2 , J. Hansen 2 , O. Hansen 1 , E. Thomsen 1
1 Department of Micro and Nanotechnology, Technical University of Denmark, Kgs. Lyngby Denmark, 2 Institute of Physics and Astronomy, University of Aarhus, Aarhus Denmark
Show AbstractThis paper presents experimental results of the piezoresistance of p-type strained silicon and strained silicon germanium grown on (100) substrates. Today, these strained materials are used in high speed electronic devices. We investigate if the area of use for these strained layers can be expanded to also cover MEMS and NEMS devices. Previous work by the authors have shown that a strained material can obtain a larger piezoresistance compared to the relaxed material. The piezoresistance is measured by four point measurements on resistors oriented in different directions.The fabrication of the chips is done using low temperature processes to avoid relaxation of the strained layers. 150 nm thick strained Si layers are grown on a graded Si1-xGex (0≤x≤0.1) n-type buffer layer by molecular beam epitaxy, MBE. The strained Si0.9Ge0.1 layers are grown on a silicon buffer layer. For reference an unstrained Si layer is grown on a Si substrate. All layers are in situ doped with a boron concentration of NA=1018 cm-3. A preliminary anisotropic etch is performed in order to determine the crystal orientation with an uncertainty of ±0.1°. The resistors are isolated with a reactive ion etch and an oxide is deposited for isolation by plasma enhanced chemical vapor depostion. Ti/Al metal conductors are patterned by a lift off process and finally the chips are defined by an advanced silicon etch.The test chips are placed in a four point bending fixture and thus subjected to a uniaxial and uniform stress in the center region of the chip. In this region the piezoresistors are located.In Si the piezoresistivity tensor consists of three independent coefficients, π11, π12, and π44. In strained Si and strained Si0.9Ge0.1 the number of these independent coefficients rise to six, where the π66 coefficient is comparable to the π44 coefficient for Si. The table lists the experimental values obtained for the π44 coefficient in Si and for the π66 coefficient in strained Si with a buffer layer of Si0.9Ge0.1 and strained Si0.9Ge0.1 with a Si buffer layer. The table shows that the piezoresistance depends on how the material has been pre-strained. Strained Si grown on a relaxed Si0.9Ge0.1 buffer layer experiences a tensile strain in the surface plane whereas strained Si0.9Ge0.1 grown on a Si substrate experiences a compressive strain in the surface plane. The piezoresistance decreases in a tensile strained layer and increases in a compressive strained layer.The results show that one can tune the piezoresistance by tuning the strain in the piezoresistor and thus tailor the performance of the device.
5:00 PM - L4.7
Elastically Strain-Shared Nanomembranes: A New Route to High-Quality Strained Silicon.
Shelley Scott 1 , Michelle Roberts 1 , Donald Savage 1 , Arrielle Opotowsky 1 , Max Lagally 1
1 , University of Wisconsin, Madison, Wisconsin, United States
Show AbstractStrain engineering in Si and SiGe based materials offers the potential to increase electron mobility and tune band offsets, providing an additional parameter available in device design [1]. However, a key challenge lies in achieving a high level of strain control and processing simplicity, while suppressing the generation of dislocation defects. We present a study of elastic strain sharing in Si:SiGe:Si heterostructure membranes. Membranes are grown epitaxially by chemical vapor deposition (CVD) on the Si template layer of a (001) silicon-on-insulator (SOI) substrate. Selective etching of the buried oxide layer releases the heterostructure from the handling substrate, allowing relaxation by elastically transferring strain from compression in the alloy layer, into tensile strain in the Si layers. Using X-ray diffraction (XRD), we show that released membranes with composition 48nm Si : 128nm Si0.84Ge0.16 : 56nm Si, produce strain in the Si layers of 0.3%, corresponding to transfer of 50% of the alloy layer strain into tensile strain in the Si layers. XRD linescans exhibit thickness fringes and narrow peak widths, indicating high-quality (negligible dislocation density) single-crystal strained silicon. By tuning the relative layer thicknesses and alloy layer composition, we can modify the strain in a predictable and reliable manner [2]. In addition, the strain status is preserved after the membranes are transferred to foreign substrates. We have recently expanded this work to strained Si(110) membranes. Hole mobility is known to be significantly higher in Si(110) than it is in Si(001), but is less than the Si(001) electron mobility. Very recently it has been shown that strained Si(110) exhibits a significant improvement over unstrained Si(110) in both electron and hole mobility [3]. Consequently strained Si(110) nanomembranes have the potential to offer substantially increased performance to p-type MOSFET devices. Another option we have begun exploring, is fabrication of membranes with strain patterns, which has potential for local band gap engineering and adding directionality to the strain transfer. In this case, patterns of reduced Si template layer thickness are etched prior to growth, allowing increased strain transfer to these thin regions when the membrane relaxes. Time permitting, the early results of these new studies will also be presented.Research supported by NZ Foundation for Research Science & Technology, DOE, and NSF.[1]F. Schaffler, High-mobility Si and Ge structures. Semicond. Sci. Technol. 12, 1515 (1997)[2]M. M. Roberts, et al., Elastically relaxed free-standing strained-silicon nanomembranes. Nature Materials 5, 388 (2006)[3]T. Mizuno, et al., (110)-surface strained-SOI CMOS devices. IEEE Trans. Electron Devices 52(3), 367 (2005)
5:15 PM - L4.8
Local Strain Measurements on Sub-100nm Strained Si/SiGe CMOS Device Strucutres with Convergent Beam Electron Diffraction and Finite Element Simulation.
Wenjun Zhao 1 , Gerd Duscher 1 2 , Mohammed Zikry 3 , George Roagonyi 1
1 Materials Science and Engineering, North Carolina State University, Raleigh, North Carolina, United States, 2 Condensed Matter Division, Oak Ridge National Lab, Oak Ridge , Tennessee, United States, 3 Mechanical and Aerospace, North Carolina State University, Raleigh, North Carolina, United States
Show AbstractConvergent beam electron diffraction technique (CBED) is considered to be the only method of choice for local strain measurement on a nano-meter scale on sub-100nm CMOS device structures due to its high sensitivity and high spatial resolution. However during TEM sample preparation, the initial strain in the bulk starts to relax, especially in the channel region and the interfaces, when the sample is thinned down to a few hundred nano meter for electron transmission. This not only causes the HOLZ line broadening or splitting which makes it hard to detect HOLZ lines, but also raises a question: is the strain we extract from a CBED pattern the same as the initial strain in the bulk, if a clean CBED pattern is obtained? Our objectives are to extract the initial strain state by analyzing HOLZ line splitting aided with finite element simulation, as well as to develop a procedure to suppress the strain deformation during TEM sample preparation so that a clean and sharp CBED pattern can be obtained.Initially, a 2-D model of Si on a Si0.8Ge0.2 substrate was set up to validate the parameters used for finite element simulation. Then an initial strain of 8.2E-3 was introduce by applying a uniform temperature change to the sample. The simulation results indicate that the sample deformation is localized near the interface. The deformation angle was calculated using the electron beam column approximation. A curve of deformation angle against the distance from the interface was obtained. Experimental data was obtained from a blanket Si/Si0.8Ge0.2 structure by converting the measured HOLZ line splitting to a tilt angle at different positions along the interface. The experimental and simulation data matched very well. This consistency not only tells that the parameters used for finite element simulation are valid and the model is correct, but also quantitatively explains the origin of HOLZ line splitting in a lattice-mismatched heterostructure.The 2-D and 3-D models with appropriately chosen parameters were then used to simulate different structures. A correlation of initial strain in the bulk with the measured strain was then carried out. The combination of CBED and finite element simulation was further applied to study the effect of NiSiGe layer on strain on an actual CMOS structure. The preliminary results show that NiSiGe suppresses the relaxation of SiGe thereby reducing the strain in the channel region.
5:30 PM - L4.9
Silicon on Lattice-Engineered Silicon: Fabrication, Thermal Stability, and GaAs Integration
Carl Dohrman 1 , David Isaacson 1 , Kamesh Chilukuri 1 , Minjoo Lee 1 , Eugene Fitzgerald 1
1 Materials Science and Engineering, MIT, Cambridge, Massachusetts, United States
Show AbstractWe report the fabrication of a novel substrate platform for the monolithic integration of Si-based CMOS and GaAs-based optoelectronic devices. This platform, which we refer to as Silicon on Lattice-Engineered Silicon (SOLES), consists of a compositionally graded SiGe buffer buried underneath an SOI structure, all fabricated on a Si substrate. The SiGe graded buffer was grown by UHVCVD and was capped with Ge; this buffer provides a threading dislocation density (TDD) of ~10^6 cm^-2. While these SiGe graded buffers have been proven by previous studies to be an effective platform for fabrication of GaAs-based LEDs, lasers, and solar cells on Si substrates, the integration of both Si- and GaAs-based devices on a single chip using this technique is hampered by the large thickness (~10um) of the SiGe graded buffer. SOLES eliminates this drawback by the addition of the SOI structure on top of the Ge-rich cap. This approach provides for a Si device layer in close proximity to the GaAs-based device layer, thereby simplifying the monolithic integration of Si- and GaAs-based devices with this platform. Fabrication consists of layer transfer of Si to an oxide-coated graded buffer using oxide-oxide wafer bonding followed by hydrogen-induced layer exfoliation of the Si layer from its donor wafer. Our results show this layer transfer occurs reliably across the entire wafer, making it amenable to commercial application. We have explored the thermal stability of this structure at temperatures up to 1000 C for a variety of annealing times. In addition, progress in integrating GaAs-based semiconductors with the SOLES platform will be discussed.
5:45 PM - L4.10
Development of a Low Temperature 4 Probe STM and Electrical Conductivity Measurement of One Dimensional Surface Super Structure.
Rei Hobara 1 , Naoka Nagamura 1 , Shinya Yoshimoto 1 , Iwao Matsuda 1 , Shuji Hasegawa 1
1 School of Science, Depertment of physics, The University of Tokyo, Bunkyo-ku, Tokyo, Japan
Show Abstract We have developed two new instruments for measuring electronic transport in microscopic scales[1-6]. One is a four-tip STM[1-3]. This instrument has four STM heads driven independently and special electronics for four-point probe conductivity measurements using the four tips. We can arrange the four tips in arbitrary ways on a sample surface. With this machine, we have found a strong anisotropy in the conductivity of a one dimensional metallic Si(111)-4x1-In surface[2] and vicinal Si(111)-/3x/3-Ag surface[7] by placing the tips in square. We also investigate electrical characteristics of CNTs and CNT tips[8, 9] using tips as small manipulator. But this machine can be operated only at room temperature. The other one is variable-temperature monolithic micro-four-point probe machine[4,5], which have four conductive micro-cantilevers in line with equidistance from each other on a chip. Electrical conductance is measured by touching cantilevers to sample surface directly. We have detected a drastic change in the conductivity at a charge-density-wave transition of Si(111)-4x1-In surface to 8x2 around 130K[5,6]. These two instruments complement to each other successfully, but they are not sufficient. For example, we cannot examine the relation between anisotropic conductance and conductance increase at low temperature of a 4x1-In surface. To investigate phase transitions on surfaces, ballistic conductance in nanometer scales, or other interesting phenomena in microscopic objects, we need to measure the conductivity both at variable temperatures and in various probe arrangements. In addition to these complemental studies, we can also acquire electron propagator called Retarded Green Function[10] if we can detect ballistic current between two tips in tunneling condition. Thereby we has developed a new instrument, Low Temperature Four-Tip STM, which has similar four independent STM scanners as the former one and cooling system using liquid He. It can cool the sample to 8K. We can measure the microscopic conductivity at various temperatures and probe arrangements under SEM in UHV surroundings. In this presentation, we will present the basic design and performance of this new instrument. We will also discuss the temperature dependence of the conductivity and its anisotropy in a Si(111)-4x1-In surface.References1) S. Hasegawa, et al., Curr. App. Phys. 2, 465 (2002).2) T. Kanagawa, et al., Phys. Rev. Lett. 91, 036805 (2003).3) I. Shiraki, et al., Surf. Sci. 493, 633 (2001).4) S. Hasegawa, et al., J. Phys.: Condens. Matter 14, 8379 (2002).5) T. Tanikawa, et al., e-J. Surf. Sci. Nanotech. 1, 50 (2003).6) T. Tanikawa, et al., Phys. Rev. Lett. 93, 016801 (2004).7) I. Matsuda, et al., Phys. Rev. Lett. 93, 236801 (2004).8) R. Hobara, et al., Jpn. J. App. Phys. 43, 1081 (2004).9) S. Yoshimoto, et al., Jpn. J. App. Phys 44, 1563 (2005).10) Q. Niu, et al., Phys. Rev. B 51, 5502 (1995).
Symposium Organizers
Leonid Tsybeskov New Jersey Institute of Technology
David J. Lockwood National Research Council
Christophe Delerue IEMN
Masakazu Ichikawa The University of Tokyo
Anthony W. van Buuren Lawrence Livermore National Laboratory
L5: Group IV Nanowires I
Session Chairs
Wednesday AM, November 29, 2006
Room 207 (Hynes)
9:30 AM - **L5.1
Kinetic Factors Controlling VLS Growth of Si and Ge Nanowires.
Jerry Tersoff 1 , Suneel Kodambaka 1 , Frances Ross 1
1 , IBM Watson Center, Yorktown Heights, New York, United States
Show Abstract10:00 AM - L5.2
Electrical Characteristics and Chemical Stability of Non-Oxidized, Methyl-Terminated Silicon Nanowires.
Hossam Haick 1
1 Division Chemistry and Chemical Engineering, California Institute of Technology, Pasadena, California, United States
Show AbstractThe ability to manipulate the conductivity of nanowires (NWs) through chemical surface modification is important for the realization of NW-based electronics. Surface functionalization is important in constructing molecular electronic tunneling and coupling bridges to link physically, and control electronically, contacts between NWs. Surface functionalization is also important to impart control over the chemical and biological interactions of NWs with their environment and thereby to enable NW molecular sensing applications. Oxide-coating of a Si NW is thought to induce trap states at the Si/Si-oxide interface, and acts as a dielectric which lowers, and ultimately can limit, the effect of the gate voltage on manipulating the transconductance between the source and drain of Si NW-based FETs, as well as the response of oxide-coated Si NW chemical sensors to their environment. We repot herein that Si NWs modified by covalent Si-CH3 functionality, with no intervening oxide, show atmospheric stability, high transconductance values, low surface defect levels, and allow for the formation of air-stable Si NW FET’s having on-off ratios in excess of 105 over a relatively small gate voltage swing (±2 V). The performances of CH3-Si NW FETs were compared to SiO2-coated NW (hereby, SiO2-Si NW) and H-terminated Si NWs (hereby, H-Si NW). For “fresh” samples (i.e., before exposure to air), the average conductance of H-Si NWs and CH3-Si NWs at zero back gate voltage was ca. 4- and 7-folds higher, respectively, than that of the SiO2-Si NWs. Using a cylinder on an infinite plate model yielded estimates for μh of 18±4 cm2 V-1 s-1 for SiO2-Si NW, 123±3 cm2 V-1 s-1 for H-Si NWs, and 140±2 cm2 V-1 s-1 for CH3-Si NWs. Voltage-dependent back-gate voltage measurements as a function of time in air revealed substantial differences between SiO2-Si NWs, H-Si NWs and CH3-Si NWs. The mobility of SiO2-Si NWs did not change, within experimental error, as a function of time in air. In contrast, the mobility of H-Si NWs decreased continuously, dropping from 123 to 87 cm2 V-1 s-1 after 4 weeks (672 h) exposure to air, with the highest rate of degradation during the first 3 h of exposure to air. For CH3-Si NWs, the mobility decreased upon exposure during the first week (=164 h) by ca. 11%, from 140 to 124 cm2 V-1 s-1, after which time it stabilized at a level that was comparable to the initial value of H-Si NW hole mobility (i.e., at t= 0). We conclude that details of the passivation process can completely change the resulting device characteristics, and that surface alkylation through Si-C chemistry can be of utility in forming Si NW electrical devices and Si NW chemical and biological sensors with highly desirable properties under molecular level control.
10:15 AM - L5.3
Laser Annealing of Silicon Nanowires
Nipun Misra 1 , Li Xu 1 , Costas Grigoropoulos 1 , Nathan Cheung 2 , Yaoling Pan 3
1 Mechanical Engineering, University of California, Berkeley, Berkeley, California, United States, 2 Electrical Engineering, University of California,Berkeley, Berkeley, California, United States, 3 , Nanosys Inc., Palo Alto, California, United States
Show Abstract Semiconductor Nanowires are promising building blocks for the next generation of electronic devices, chemical and biological sensors and photonic systems. Dopant activation in these nanowires has commonly been accomplished by thermal annealing. Laser Annealing has emerged as an attractive technique of fabricating ultra-shallow junctions in MOS devices, owing to its low thermal-budget. Rapid heating and cooling of the samples limits the diffusion of dopants, aiding the formation of shallow junctions. Additionally, the damage to underlying layers and substrate tends to be minimal. Furthermore, laser annealing is compatible with flexible plastic substrates since it is a low-temperature process. In this work, we successfully demonstrate non-melt laser annealing of silicon nanowires (SiNWs) as an efficient method for the electrical activation of implanted dopants and restoration of crystalline structure in the wires. The KrF Excimer (248nm) and Nd:YAG (532nm) lasers were employed in this study. By controlling the lasing parameters – fluence and number of pulses, dopant activation levels comparable to Rapid Thermal Annealing (RTA) were achieved. Post annealing structural characterization showed that laser annealing can remove the residual implantation damage. In situ electrical measurements of boron-implanted SiNWs were used to study the dopant activation during the pulsed laser annealing process. Even though the melting threshold of the nanowires in laser annealing process was found to be lower than that for bulk silicon, it was possible to provide sufficient energy for the activation of dopants without melting or damaging the wires. Electrical monitoring data was found to be consistent with structural characterization results.
10:30 AM - L5.4
Band-gap Modulation in Single-crystalline Si1-xGex Nanowires.
Jee-Eun Yang 1 , Chang-Beom Jin 1 , Cheol-Joo Kim 1 , Yosep Yang 1 , Chan-Gyung Park 1 , Moon-Ho Jo 1
1 Materials Science and Engineering, Pohang University of Science and Technology (POSTECH), Pohang, Kyungbuk, Korea (the Republic of)
Show AbstractGroup IV semiconductor Si1-xGex alloys offer a continuously variable system with a wide range of crystal lattices and energy-band gaps, leading to various electrical and optical properties. Of particular interest is the continuous variation in the energy band-gap in Si1-xGex, since it potentially allows enhanced light emission and detection in the wavelength range of optical fiber communication and thus provides challenging opportunities for Si optoelectronics. Here we report growth of single crystalline Si1-xGex nanowires, whose relative composition is controllably tuned over the entire composition range by Au-catalyst assisted chemical vapor syntheses. We then present experimental demonstration of the band-gap modulation from near infrared to ultraviolet regions with alloying of Si and Ge, and their spatial confinement at the nanometer scale. Specifically we show that the optical absorption band-edge shifts to a lower energy in the near infrared range with increasing Ge content, and also show that the band-edge shifts to a higher energy in the ultraviolet range with decreasing diameter of the nanowires below 20 nm. Our finding demonstrates that the energy band-gap of Si1-xGex nanowires can be modulated in a wider energy range, and suggests implications for group IV semiconductor optoelectronics.
10:45 AM - L5.5
Silicon and Germanium Nanowire Synthesis: Catalytic Reactant Decomposition and Solid-Phase Seeding by Nanocrystals in Organic Solvents.
Hsing-Yu Tuan 1 2 3 , Doh Lee 1 2 3 , Brian Korgel 1 2 3
1 Chemical Engineering, University of Texas at Austin, Austin, Texas, United States, 2 Texas Materials Institute, University of Texas at Austin, Austin, Texas, United States, 3 Center for Nano- and Molecular Science and Technology , University of Texas at Austin , Austin, Texas, United States
Show AbstractThe familiar gold-seeded vapor-liquid-solid (VLS) approach to nanowire synthesis gives high yields of single crystalline Si and Ge nanowires at relatively low temperature. However, Au traps electrons and holes in both Si and Ge and Au-seeding nanowire is not desirable for electronic applications. We have examined a range of different nanocrystals for Si and Ge nanowire synthesis, including Co, Ni, CuS, Mn, Ir, MnPt3, Fe2O3, and FePt. Si and Ge nanowires were grown by decomposing organosilanes or organogermanes in supercritical high temperature (500°C) and high pressure (10.3MPa) toluene. Ni and Co nanocrystal were both found to produce high quality Si and Ge nanowires with high yields.Co gave the highest yield and quality of both Si and Ge nanowires, rivaling Au-seeded reactions. Ni nanocrystals also produced crystalline Si and Ge nanowires with good yield. CuS nanocrystals produced straight crystalline Si nanowires but slightly shorter lengths (3-10 µm) and Fe2O3 nanocrystals produced high quality Ge nanowires but not a good catalyst for seeding Si nanowires.Both Co and Ni have eutectic temperatures with Si and Ge that were well above the nanowire growth temperature. Unlike Au nanocrystals, which seed nanowire growth through the formation of a liquid Au:Si (Au:Ge) alloy, Co and Ni appear to seed nanowires by forming solid silicide alloys. The small seed particles diameter (i.e., 5-10 nm) can be saturated by solid-state diffusion which is fast enough to keep up its nanowire growth. We call this solid-phase seeding mechanism “supercritical fluid-solid-solid” (SFSS) growth.Co and Ni nanocrystals were also found to catalyze Si nanowire growth from silanes that do not yield nanoiwres with Au nanocrystals. Both octylsilane and trisilane do not yield nanowires in the presence of Au nanocrystals. Co and Ni catalyze reactant decomposition and then promote nanowire growth by a solid phase seeding mechanism. Using catalytic seed metals to promote nanowire growth could lower growth temperature and might represent a general approach for controlled nanowire growth.
11:30 AM - L5.6
Deterministic Nanowire Growth.
Jacob Woodruff 1 , Joshua Ratchford 1 , Hemanth Jagannathan 2 , Yoshio Nishi 2 , Christopher Chidsey 1
1 Chemistry Department, Stanford University, Stanford, California, United States, 2 Electrical Engineering, Stanford University, Stanford, California, United States
Show AbstractIn the search for new methods to define active device elements in nanoscale 2D and 3D electronics, solar cells, and sensors, the use of chemical vapor deposition (CVD) grown germanium nanowires shows promise due to their sub 400°C growth temperature and high carrier mobilities. However, in order for CVD grown nanowires to be realized as a viable technology for future applications, the ability to control the nanowire’s position, diameter and orientation is crucial. This type of control during nanowire growth we are designating “deterministic nanowire growth.” In this presentation, a description of some of the key factors influencing deterministic nanowire growth is presented as well as methods that have been used to obtain this goal. These methods include site selective catalyst definition by physical and electrochemical means, hetero-epitaxy of germanium nanowires from silicon substrates, via-directed growth, and control of CVD growth conditions.
11:45 AM - L5.7
Vertical Growth and Characterization of Single Crystalline SiGe Alloy Nanowires.
Han-Kyu Seong 1 , Tae-Eon Park 1 , Hee-Chul Han 1 , Eun-Kyoung Jeon 2 , Jeong-O Lee 2 , Ju-Jin Kim 3 , Heon-Jin Choi 1
1 School of Advanced Materials Science and Engineering, Yonsei University , Seoul Korea (the Republic of), 2 Advanced Materials Division, Korea Research Institute of Chemical Technology, Daejeon Korea (the Republic of), 3 Department of Physics, Chon-buk National University, Chon-ju Korea (the Republic of)
Show AbstractSi-Ge alloy system changes the band structure relative to that of pure Si and Ge and, hence, changes the optical and electrical properties, creating the potential for band gap engineering in Si based electronic device. Bulk of works has been carried out to exploit the potential. However, a difficulty in the growth of high-quality Si-Ge alloy acts as a hurdle to achieve the band gap engineering. Meanwhile, semiconductor nanowires have been attracted greatly due to their novel physical and chemical properties as well as uniqueness as building blocks for realizing nanodevices. Importantly, Si and Ge nanowires have a great potential for CMOS compatible nanowires-based electronics and optoelectronics. Herein we report on the growth and characterization of single crystalline Si1-xGex (x: 0 ~ 0.3) alloy nanowires. The nanowires with the diameters of < 100 nm and lengths of several micrometers were vertically grown on the silicon substrates by vapor-liquid-solid (VLS) mechanism. From the measurement of core level spectra of Si 2p and Ge 3d by using synchrotron radiation photoemission spectroscopy (SRPES), the binding energies of the nanowires were not changed with the amount of Ge while the position of the valence band maximum (VBM) were shifted towards a lower energy. The Fourier transformed curves of experimental extended X-ray absorption fine structure (EXAFS) at Ge K-edge indicated that the first peak around 1.99 Å corresponding to the Ge-Ge bond length (first nearest neighbor ions) decreased slightly with the amount of Ge. The electrical properties from the nanowires based transistor structures showed pronounced p-type transport behavior, with on/off ratios as high as 106. Also, strong diode effect observed from the Ni-contacted Si0.7Ge0.3 nanowire, and the break-down did not occur until we increased the reverse bias voltages to -1 V. Based on the experimental results, the possible application as building blocks for high performance nanodevices as well as the growth and properties of SiGe alloy nanowires will be discussed.
12:00 PM - L5.8
Growth Characteristics and Properties of Small Diameter Silicon Nanowires.
Pramod Nimmatoori 1 , Trevor Clark 2 , Elizabeth Dickey 2 , Joan Redwing 2
1 Department of Chemical Engineering, Pennsylvania State University, State College, Pennsylvania, United States, 2 Department of Materials Science and Engineering and Materials Research Institute, Pennsylvania State University, State College, Pennsylvania, United States
Show Abstract12:15 PM - L5.9
Transport Characterization and Device Applications of P-donor Nanowires in Silicon.
Tsung-cheng Shen 1 , S. Robinson 1 2 , J. Tucker 2
1 Department of Physics, Utah State University, Logan, Utah, United States, 2 Department of Electrical and Computer Engineering, University of Illinois at Urbana-Champaign, Urbana, Illinois, United States
Show AbstractThree-dimensional carrier transport in doped semiconductors was extensively investigated in the 1980s as models of disordered systems. However, carrier transport in reduced dimensions has not been much studied experimentally because of the difficulty of confining dopant distribution in a crystal. In the past few years, we have successfully embedded single atomic layers of P in Si by exposing Si(100) surfaces to PH3 molecules in ultrahigh vacuum followed by low-temperature epitaxial Si growth. Electron densities in these δ-layers can be as high as 1.5x1014 cm-2, hence it should remain metallic at zero temperature. We find that surface roughness dictates the carrier mobility and activation even though all surfaces are locally ordered. The logarithmic temperature dependence of the resistance down to 0.3 K clearly demonstrates the two-dimensional nature of the δ-layers. Furthermore, we can confine the 2D δ-layers laterally by applying STM e-beam lithography on a single-layer H-resist to define P-donor wires with widths from 200 to 5 nm. At sufficient low temperatures, the width is smaller than the phase coherence length of electrons and the wires are considered to be quasi-1D. The resistance of these wires, however, does not follow the typical T-p behavior as reported in systems of metal wires and narrow Si-MOSFET inversion layers. In this presentation, we will discuss the results of the electrical and magnetoresistance measurements of the P-donor wires as functions of temperatures and their implications. The goal of this research is to use such P-donor patterns as building blocks for nanoscale integrated circuits. This work is supported by NSF-NIRT under Grant No. CCF-0404208.
12:30 PM - L5.10
Synthesis and Optical Properties of Silicon Nanowires
Billel Salhi 1 2 , Bernard Gelloz 3 , Nobuyoshi Koshida 3 , Gilles Patriarche 4 , Rabah Boukherroub 1 2
1 Institut de Recherche Interdisciplinaire (IRI), CNRS, Villeneuve d'Ascq France, 2 Institut d'Electronique, de Microélectronique et de Nanotechnologie, CNRS, Villeneuve d'Ascq France, 3 Graduate School of Engineering, Tokyo University of Agriculture and Technology, Tokyo Japan, 4 Laboratoire de Photonique et de Nanostructures, CNRS, Marcoussis France
Show AbstractSemiconductor nanowires have attracted great attention owing to their submicron ultimate feature size, to the expected original electrical and optical properties and the potential applications in the field of nanoelectronics, high-speed field effect transistors, bio and chemical sensors, and light-emitting devices with low power consumption [1-5]. The aim of the present work is to investigate the optical properties of SiNWs and to study the effects of a high-pressure water vapor annealing (HWA) treatment on the photoluminescence (PL) [6]. The synthesis of SiNWs with controlled and stable optical properties will have potential applications in various fields ranging from optical devices to nanobiosensors with high sensitivity.SiNWs with 20 nm diameter were synthesized on porous silicon substrate using the VLS process [7]. Scanning electron microscopy (SEM) analysis showed silicon nanowires with 20 nm diameter and few microns long. After HWA treatment (1.3 MPa at 260°C), a slight increase of the SiNWs diameter was observed, consistent with surface oxidation of the silicon nanowires. Increasing the pressure during the water vapor annealing to 2.6 MPa at 260°C led to a significant increase of the silicon nanowires diameter.The effects of a treatment based on high-pressure water vapor annealing on (SiNWs) have been investigated in terms of the photoluminescence (PL) efficiency and stability. The HWA treatment at two different pressures (1.3 and 2.6 MPa) and temperature (260°C) was examined. Freshly prepared SiNWs display a weak PL peak and most likely results from silicon nanowires with small diameters (< 10 nm). Surface oxidation of the nanowires using HWA technique at 1.3 MPa led to PL intensity enhancement by a factor 1.5, while the PL peak wavelength remained unchanged (~ 700 nm). HWA treatment at higher pressure (2.6 MPa) caused a PL intensity increase by a factor 10 and a blue shift (~ 20 nm) of the PL emission. The observed blue shift in the PL emission after HWA treatment is related to silicon nanowire oxidation and to decrease of the mean diameter of the SiNWs crystalline core. The high stability of the PL indicates that the SiO2 tissue surrounding the silicon nanowire is of high quality. The HWA technique has several advantages over other techniques for fabrication of luminescent silicon-based nanostructures [8].1. Y. Cui, Z. Zhong, D. Wang, W. U. Wang, C. M. Lieber, Nano Lett. 3, 149 (2003). 2. Y. Cui and C. M. Lieber, Science 291, 851 (2001)3. J.-W. Chung, J.-Y. Yu, and J. R. Heath, Appl. Phys. Lett. 76, 2068 (2000).4. Y. Cui, Q. Wei, H. Park, C. M. Lieber, Science 293, 1289 (2001). 5. J.-in Hahm, C. M. Lieber, Nano Lett. 4, 51 (2004).6. B. Gelloz, A. Kojima, N. Koshida, Appl. Phys. Lett. 87, 031107 (2005).7. B. Salhi, B. Grandidier, R. Boukherroub, J. Electroceram. 16, 15 (2006)8. B. Salhi, B. Gelloz, N. Koshida, G. Patriarche, R. Boukherroub, Phys. Stat. Sol. (C) (2006)
12:45 PM - L5.11
Saw-tooth Faceting in Silicon Nanowires Related to Gold Migration.
Vladimir Sivakov 1 2 , Gudrun Andrae 2 , Samuel Hoffmann 4 , Johann Michler 4 , Roland Scholz 1 , Reinhard Schneider 1 , Ulrich Goesele 1 , Silke Christiansen 1 3
1 , Max Planck Institute of Microstructure Physics, Halle Germany, 2 Laser Technology, Institute of Physical High Technology, Jena Germany, 4 , EMPA, Thun Switzerland, 3 Physics, Martin Luther University Halle-Wittenberg, Halle Germany
Show AbstractL6: Group IV Nanowires II
Session Chairs
Ulrich Goesele
Mark Lundstrom
Wednesday PM, November 29, 2006
Room 207 (Hynes)
2:30 PM - **L6.1
Properties and Fabrication of Silicon Nanostructures.
Ulrich Goesele 1 , Silke Christiansen 1 , Florian Kolb 1 , Alexey Milenin 1 , Manfred Reiche 1 , Volker Schmidt 1 , Roland Scholz 1 , Stephan Senz 1 , Tomohiro Shimizu 1 , Martin Steinhart 1 , Peter Werner 1 , Danilo Zschech 1
1 , Max Planck Institute of Microstructure Physics, Halle Germany
Show Abstract3:00 PM - L6.2
Synthesis and Properties of FeSi and Fe1-xCoxSi alloy Nanowires
Andrew Schmitt 1 , Lei Zhu 1 , Franz Himpsel 2 , Dieter Schmeisser 3 , Song Jin 1
1 Chemistry, University of Wisconsin-Madison, Madison, Wisconsin, United States, 2 Department of Physics, University of Wisconsin-Madison, Madison, Wisconsin, United States, 3 , Lehrstuhl Angewandte Physik, Cottbus Germany
Show AbstractWe report the chemical synthesis of free standing single-crystal nanowires of the silicon based intermetallics FeSi and CoSi, alloys of which form the ferromagnetic semiconductor FexCo1-xSi. Nanowires are produced on silicon substrates covered with a thin layer of silicon oxide through decomposition of the single source organometallic precursors trans-Fe(SiCl3)2(CO)4 and Co(SiCl3)(CO)4, respectively, in a simple chemical vapor deposition process. Unlike typical vapor-liquid-solid nanowire growth, these silicide nanowires form without the addition of metal catalysts, have no catalyst tips, and depend strongly on the surface employed. X-ray spectroscopy verifies the identity and the room temperature state of silicide nanowires, and room temperature and temperature dependent transport properties have been determined. This general approach may lead to practical routes to other silicon based nanomaterials.
3:15 PM - L6.3
Simulating Optical Properties of Silicon Nanowires.
Hugh Wilson 1 , Giulia Galli 1 , Francois Gygi 1 , Sebastien Hamel 2 , Andrew Williamson 2 , Ed Ratner 3 , Dan Wack 3
1 Chemistry, University of California, Davis, Davis, California, United States, 2 , Lawrence Livermore National Laboratory, Livermore, California, United States, 3 , KLA-Tencor , Milpitas, California, United States
Show Abstract3:30 PM - L6.4
Magic structures of H-passivated Silicon nanowires
Cristian Ciobanu 1 , Ning Lu 2 , Tzu-Liang Chan 2 , Cai-Zhuang Wang 2 , Kai-Ming Ho 2
1 Division of Engineering, Colorado School of Mines, GOlden, Colorado, United States, 2 USDOE Ames Laboratory, Physics Department, Iowa State University, Ames, Iowa, United States
Show AbstractRecent experimental work has shown that Si nanowires with surface passivated by hydrogen can be produced by etching of oxide-sheath nanowires in hydrofluoric acid (D.D.D. Ma et al, Science 299, 1874 (2003)). Inspired by these experiments, we have set out to develop methods for finding the most favorable shapes of the crossection of the wires as a function of their diameter and axis orientation. We