Science of Nanoscale Systems and their Device Applications
Science of Nanoscale Systems and their Device Applications
The Global Nanoelectronic Challenge
Pushkar P. Apte
Vice President, Technology Programs
Semiconductor Industry Association
At the onset of the 21st century, we are in a world revolutionized by information technology. Semiconductor components power this revolution, and most of these components are built using complementary metal oxide semiconductor (CMOS) integrated circuits. The semiconductor industry has been continuously “scaling” CMOS circuits — making them faster, cheaper and smaller, while adding more functionality. This progression is represented by the famous Moore’s Law, and it has already driven the industry into the nanoelectronics regime, with minimum horizontal patterned dimensions of the order of tens of nanometers and vertical dimensions of the order of a few nanometers. However, Moore’s Law is not a natural law: the industry has been innovating at a furious pace to keep up with this ‘law.’ This will continue for some time, but most experts believe that CMOS technology will reach the end of its progression in about 15 years. The challenge in nanoelectronics is to sustain the innovation necessary to take CMOS to its ultimate limits, and in parallel, to begin researching the alternatives to CMOS. Continued investment in long-range research and knowledge is crucial to ensure this, and this investment must happen through collaboration between the three major stakeholders — government, industry and academia. This talk will highlight two programs that epitomize this spirit — the Focus Center Research Program and the Nanoelectronics Research Initiative. Both are fully based at Universities and are funded in part by consortia of participating member companies of the Semiconductor Industry Association. Both involve strong partnerships with the US Government, which has a critical role to play in long-range research.
Advances in Quantum Dots for Semiconductor Non-Classical Light Sources
Research Center for Advanced Science and Technology
The University of Tokyo, Japan
Sources of single photons and entangled photons are needed in the field of quantum information, such as quantum cryptography and linear-optical quantum computation. Single photon sources based on InAs quantum dots near 1 mm have been intensively investigated by many researchers. However, there still remain various issues, such as too short wavelength for optical fiber long-haul transmission, low temperature operation (< 10 K) and low emission efficiency. In this presentation, we address recent advances in single photon sources including successful demonstration of high temperature (200 K) operation using high-quality GaN quantum dots, which exhibit a negative binding energy of bi-excitons. Moreover, a highly efficient light emission from InAs quantum dots embedded in a photonic crystal nanocavity is also discussed as well as 1.55 mm single photon sources. These results are promising for realizing quantum information systems in the near future.
Imaging Transport Resonances in the Quantum Hall Effect
Massachusetts of Technology
We image charge transport in the quantum Hall effect using a scanning charge accumulation microscope. Applying a DC bias voltage to the tip induces a highly resistive ring-shaped incompressible strip (IS) in a very high mobility 2-D electron system (2DES). The IS moves with the tip as it is scanned, and acts as a barrier that prevents charging of the region under the tip. At certain tip positions, short-range disorder in the 2DES creates a quantum dot island inside the IS that enables breaching of the IS barrier by means of resonant tunneling through the island. Striking ring shapes appear in the images that directly reflect the shape of the IS created in the 2DES by the tip.
Carbon Nanotube Electronics and Optoelectronics
IBM Research Division, T.J. Watson Research Center
Carbon nanotubes (CNTs) have properties that make them ideal for applications in both nano- and opto-electronics. Although a variety of different electronic devices based on CNTs have been demonstrated, most of the emphasis has been placed on CNT field-effect transistors (CNTFETs). In these a single semiconducting CNT replaces silicon as the transistor “channel.” The resulting devices have in many respects characteristics superior to conventional devices. However, they also pose a set of new challenges. These include understanding the new 1-D transport physics, reducing the influence of Schottky barriers at CNT-metal contacts and eliminating the ambipolar behavior of vertically-scaled CNTFETs. I will discuss how these problems can be resolved to produce high performance nanotube devices as well as multi-component single nanotube circuits.
We have used ambipolar (a-) CNTFETs to simultaneously inject electrons and holes with a fraction of these recombining radiatively to produce an electrically-excited, single nanotube molecule light source. Unlike conventional p-n diodes, a-CNTFETs are not doped and there is no fixed p-n interface. Thus, the emitting region can be translated at will along a CNT channel by varying the FET gate voltage. We have found that substantially stronger localized electroluminescence can be generated at defects or inhomogeneities that introduce potential drops. The emission is the result of intra-molecular impact excitation of electron-hole pairs by the accelerated (“hot”) carriers. Localized electroluminescence provides a high brightness IR source and a novel probe of defects, charging, and other otherwise difficult to observe inhomogeneities.
Nanophotonic Devices Based on Wires, Waveguides and Optical Antennas
I will review recent advances in collaboration with the groups of Ken Crozier (Harvard), Charles Lieber (Harvard), George Whitesides (Harvard) and John Joannopoulos (MIT) on advanced nanophotonic devices that combine new fabrication methods, device designs and concepts to achieve new functionalities for a potentially wide range of applications.
Imaging and spectroscopy of nanowire photonic crystal structures demonstrate localized emission in engineered artificial defects and light suppression in the region of the photonic crystal. One-dimensional photonic crystal cavities have been used to improve the reflectivity of the nanowire end facets and nanowire racetrack resonators have been implemented. Next I will discuss a top-down technique that generates patterned arrays of gold nanowires of uniform, controllable length, width, and height, and describe a systematic study of the dependence of the surface plasmon resonance on the geometry of these wires. This fabrication technique combines photolithography, thin-film metal deposition, and thin-film sectioning and produces nanowires with high aspect ratio cross sections. Surface plasmons have been used to demonstrate a new class of devices in which an optical antenna has been defined on the facet of a semiconductor laser. We have directly observed the highly localized enhancement of the laser field in the vicinity of the nanometric gap of the antenna using apertureless NSOM. These active optical antennas have potentially wide-ranging applications such as high-resolution spatially resolved imaging and spectroscopy, optical recording as well laser assisted processing and repair of masks and circuits. In the last part of the talk calculations of the forces arising from the overlap between the guided waves of parallel nanophotonic waveguides are presented. Both repulsive and attractive forces, determined by the choice of the relative input phase are found. Using realistic parameters for a silicon-on-insulator material system, we have estimated that the forces are large enough to cause observable displacements, making them suitable for a broad class of optically tunable microphotonic devices such as new optical routers.
This talk will present our latest research on single-walled carbon nanotube electrical properties and devices. We have been using carbon nanotube as a model system to study interesting nanoscale problems concerning materials synthesis, solid-state physics and devices. This presentation will cover our latest results on coherent quantum electron transport and diffusive electron-phonon scattering phenomena in suspended nanotubes, hot phonon effects and electro-thermal transport, and thermal conductivity of single tubes. Understanding of current-limited transport in single-walled carbon nanotubes is vital to many potential applications. High bias electron transport characteristics under strong self-heating were exploited to probe the thermal conductivity of individual SWNTs (~ 3600 Wm-1K-1 at T = 300 K) up to ~700 K, and revealed a 1/T dependence expected for Umklapp phonon scattering at high temperatures. We further explored the high bias characteristics of our devices in various environments. Negative differential conductance at low bias (below 0.4 V) appears as a result of extreme self-heating and the formation of non-equilibrium optical phonons. Various gas and molecular solid environments lead to the reduction or elimination of the non-equilibrium phenomenon. Lastly, I will also present our latest results on pushing the performance limit of nanotube field effect transistors.
Overview of U.S. Industry Strategic Plan
Since the invention of the integrated circuit almost 50 years ago, the underlying technology has progressed at fairly steady exponential rates for almost every significant parameter. In particular, cost-per-function, operations-per-second, and energy-per-operation have improved by many orders of magnitude over this period. The resulting growth in applications has propelled IC industry growth to serve a worldwide market now well-exceeding $200 B per year. This enormous progress has been mainly based on the ability to continually scale device feature sizes from millimeters to tens of nanometers, as well as to the transitions from bipolar to PMOS to NMOS to CMOS devices. Of course, as we have been approaching perceived limits to further scaling, the industry has responded with new R&D initiatives to address the challenges. Several of these are consortia, in which IC companies that are frequently business rivals jointly fund pre-competitive R&D, most of which is conducted by university researchers. On shorter-term development, there are also R&D partnerships, in which companies pool resources to co-develop their next round of semiconductor process technology. A version of this model also applies to joint development between IC makers and their equipment and materials suppliers. In all of these cases, as well as for their purely internal R&D, most of the industry is guided by the International Technology Roadmap for Semiconductors, a worldwide consensus-building process on the critical R&D needs. The ITRS projects technology trends out to a 15-year rolling horizon and attempts to highlight the major obstacles to continued scaling. The 2005 ITRS has just been published, and its 2020 horizon coincides with the target date of the latest Semiconductor Industry Association technology initiative, the Nanoelectronics Research Initiative, to help demonstrate a commercially-viable successor to CMOS. Of course, this is part of a worldwide endeavor, which is very loosely coordinated via a number of international forums.
In quantum double-dots made from semiconductor material, various interdot coupling parameters can be manipulated for conducting electrons, such as tunnel coupling, electrostatic coupling and exchange coupling. This makes very useful double-dot systems for studying molecule-like quantum mechanical properties and also application for implementing quantum computing. Particularly, manipulation of exchange coupling between two electron spins in two dots is a fundamental concept in spin-based quantum computing.
We fabricate a unique hybrid vertical-lateral double-dot in which the above-described coupling parameters are tunable with four gate voltages. Two (side gates) of the four gates are used to independently change the numbers of electrons in the two dots. The remaining two gates (center gates) are used to change the inter-dot tunnel coupling. We adjust the four gate voltages to effectively place one and two electrons on the two dots. We observe a tunnel-coupled bonding (ground) and anti-bonding (excited) states for the one-electron state and a singlet (ground) and triplet (excited) states for the two-electron state. We measure the gate voltage and magnetic fielddependencies of the tunnel coupling energy and exchange coupling energy, and find that the tunable ranges of these coupling energies are relevant for quantum gate (SWAP) operation. We also discuss the validity of a Heilter-London model for describing the exchange coupling energy in our double-dot system.
The physics of imaging the electron flow in a 2DEG at zero magnetic field is relatively well understood. I will review the main points and what has been learned about the electron flow from the experiments. Newer experiments (Kathy Aidala, R.M. Westervelt,) and theory (R. Parrott, T. Kramer, and EJH) have examined the effects of an applied, perpendicular magnetic field and new geometries, including transconductance involving two quantum point contacts. The largely non-invasive role of the charged AFM tip in the zero magnetic field case can become invasive at certain tip potentials for nonzero field. These ongoing experiments and theoretical developments will be described.
We have directly determined the spectral shape of the complex conductivities of Bloch oscillating electrons by using time-domain terahertz (THz)-electro-optic sampling technique and presented an experimental evidence for a dispersive Bloch gain in superlattices. This unique dispersive gain without population inversion arises from a non-classical nature of Bloch oscillations; that is, the phase of the Bloch oscillation is shifted by p/2 from that of the semi-classical charged harmonic oscillation when driven by the same ac field. By increasing the bias electric field, the gain bandwidth reached ~3 THz in our particular sample
Polarization, Relaxation, and Coherent Control of Nuclear Spins
in Semiconductor Systems
NTT Corporation, SORST/JST, and Tohoku University, Japan
Polarization, relaxation, and coherent control have been studied for nuclear spins in GaAs-based hetero- and nanostructures. At n = 2/3 degenerate state, nuclear spin polarization is realized by flowing appropriate current in hetero- and nanostructures and the degree of nuclear spin polarization is directly detected by the longitudinal resistance. We have applied this feature to measure nuclear spin relaxation, which is highly sensitivity to electron spin states, especially low energy excitation of electron spins. We can clarify electron spin features in single and double layer quantum wells from the measurements of nuclear spin relaxation. We have also extended electron-nuclear spins interactions to a coherent control of nuclear spins in a point-contact device. The device is constructed by combing a point contactand an antenna gate for applying electromagnetic radiation. Strikingly clear oscillations have been demonstrated for all quadrupolar splitted transitions of Ga and As nuclei. The decoherence mechanism of the coherent oscillation has been studied from the decoupling experiments, suggesting an important role of the crystal structure of GaAs.
Magnetic nanostructures are increasing data storage capacities and are promising candidates for implementations of novel spin-based computation techniques. The relative simplicity and reduced dimensionality of nanoscale magnetic structures also make them attractive model systems for studying the interactions between small numbers of quantum spins. Using a high-field low-temperature scanning tunneling microscope, we assemble linear chains of Mn atoms one atom at a time on thin, insulating layers of copper nitride. We probe the excitation spectra of the individual magnetically-coupled chains with inelastic electron tunneling spectroscopy. The spectra change dramatically with both the parity and length of the chain and reveal a variety of spin excitations. These results provide direct evidence of antiferromagnetic coupling between the atomic spins on neighboring Mn atoms. A quantitative comparison with the Heisenberg open chain model allows us to measure the coupling strength ~6 meV between these atomic spins. We further determine the spin of each Mn atom on the surface to be 5/2.
In the past few years there has been an increasing interest in the study of quantum dots (QDs) in a variety of low-dimensional materials. Carbon nanotubes (CNTs) are particularly attractive because of their large coherence lengths, and their contact properties and peculiar band structure makes them an interesting system to explore novel physical phenomena. In the first part of this talk I will review some of our recent experiments in Delft, where we show that CNT QDs can operate in a wide range of coupling regimes, with different typical associated phenomena. Among these are novel Kondo effects and the possibility to modulate, due to its quantized energy spectrum, the supercurrent flowing through a CNT. In order to fully exploit the potential of CNTs as quantum dot systems, it is highly desirable to be able to create tunable tunnel barriers at arbitrary locations along the tube. In the latter part of my talk I will describe our efforts to fabricate fully tunable CNT double-quantum dots, as a first step towards the realization of more sophisticated pulsed experiments to determine the spin relaxation and coherence times in carbon nanotubes.
Spin Qubits and Hyperfine Interactions in Single and Double Quantum Dots
University of Basel, Switzerland
I review our results on hyperfine-induced electron spin dynamics for electrons confined to single [1-3] and double quantum dots [4,5].
For double dots the quantum solution accounts for decay of a singlet-triplet correlator even in the presence of a fully static nuclear spin system, with no ensemble averaging over initial conditions . We find that the singlet-triplet correlator shows a long-time saturation value that differs from 1/2, even in the presence of a strong magnetic field. Furthermore, we find that the form of the long-time decay undergoes a transition from a rapid Gaussian to a slow power law when the exchange interaction becomes nonzero and the singlet-triplet correlator acquires a phase shift given by a universal value at long times. These predictions are found to be in good agreement with recent measurements performed at Harvard . The dephasing effect of the nuclear spins can be controlled by narrowing the initial nuclear spin configuration which can be achieved by projective measurements . I also discuss the effect of orbital dephasing on singlet-triplet decoherence , and show that there is an optimal operating point where orbital dephasing becomes negligible. G. Burkard, D. Loss, and D.P. DiVincenzo, Phys. Rev.B59, 2070 (1999).
 A. Khaetskii, D. Loss, and L. Glazman, Phys. Rev. Lett.88, 186802 (2002).
 W. A. Coish and D. Loss, Phys. Rev. B70, 195340 (2004).
 B. Coish and D. Loss, Phys. Rev. B72, 125337 (2005).
 E.A Laird et al., cond-mat/0512410
 D. Klauser, W.A. Coish and D. Loss, cond-mat/0510177
In this presentation, recent progress in 2-D photonic crystals (PCs) will be reviewed. First of all, progress in high-Q nanocavities in 2-D PCs is explained, where it is shown that the concept of Gaussian confinement  and ideas of tuning of air-holes  and/or double heterojunctions [3, 4] enable to achieve nanocavities with ultrahigh-Q factor on the order of 106 and very small modal volume of ~1(l0/n)3. Next, the effect of 2-D PCs on the control of spontaneous emission is described. In contrast to 3-D PCs [5, 6], 2-D photonic crystals have a very unique feature on the control of spontaneous emission. It is experimentally demonstrated that overall spontaneous emission rate is suppressed by 2D PBG effect, while emission efficiency for the vertical direction is significantly enhanced . The effect of introduction of quantum dots, which can confine carriers three-dimensionally, is also described .
Controlling a Singlet-Triplet Spin Qubit
An attractive candidate for a solid-state quantum bit is based on semiconductor quantum dots, which allow controlled coupling of one or more electrons, using rapidly switchable voltages applied to electrostatic gates. Due to tight confinement and the high degree of isolation from the environment, spin relaxation times in quantum dots can approach millisecond time scales . In this talk I will describe how fast electrical control of the exchange interaction can be used to coherently manipulate two-electron spin states . By separating a spin singlet state on-chip, we measure an ensemble averaged spin dephasing time T2* of 10 ns, limited by the contact hyperfine interaction with the GaAs host nuclei. We develop quantum control techniques based on the exchange interaction to correct for hyperfine dephasing. Coherent spin state rotations are achieved, including spin SWAP. By using a spin-echo pulse sequence based on the exchange interaction we extend the spin coherence time, T2 beyond 1.2 microseconds.
In this talk I will review recent progress in the understanding and control of the epitaxial growth of semiconductor nanowires (NWs) based on seeding of growth by metallic nanoparticles, from the perspective of realization of advanced electronic devices and circuits. Of special importance for applications is the degree to which dimensions and positioning of NWs and related devices can be controlled. Another crucial issue for device applications in electronics and photonics is the controlled formation of heterostructures in NWs, axially as well as radially, allowing formation of quantum dots and tunnel-barriers as well as core-shell structures in nanowires. I will give examples of basic studies of quantum devices realized this way, including studies of single and coupled quantum dots, the realization of single-electron and resonant-tunneling devices as well as double-heterostructure light-emitting structures. Finally, I will discuss the degree to which epitaxially formed nanowires may add significant opportunities for extreme downscaling of nanoelectronic devices, such as wrap-gate field-effect transistors for ultra-low power and high-speed applications.
The rapidly developing young fields of spin electronics (or spintronics) and quantum information science have both led to a strong interest in the ability to measure and coherently manipulate electron spins. In particular, a single electron spin confined in a solid-state environment such as a quantum dot, has been put forward as a natural quantum two-level system to be applied as quantum bit. The electron’s spin is generally much less weakly coupled to its environment than the electron’s charge. On the one hand, this leads to substantially longer coherence times for electron spins as compared to charge. On the other hand, electron spins (especially single electron spins) are considerably harder to control than charge-based systems.
In the presentation, we theoretically describe a general concept for realizing a solid-state quantum two-level system, based on a single electron in a quantum dot, which combines ease of manipulation with long coherence times. This is accomplished by combining the spin and charge degrees of freedom of an electron in a quantum dot situated in a static slanting Zeeman field. A robust single pseudo-spin system is obtained that can be controlled by voltage only, without the need for an external time-dependent magnetic field or spin-orbit coupling. This unique and important feature is expected to considerably facilitate experimental realization of qubits based on single electrons. We show that both single qubit rotations and the C-NOT operation can be realized, thereby providing a universal set of gates for quantum computation. We find that acoustic phonon scattering is the dominant relaxation mechanism in our system and present a detailed analysis for GaAs quantum dots. Although systems based on e.g., single wall carbon nanotubes should have much better coherence properties, we already find promising values for the quality factor in GaAs. Importantly, we also present a method to determine the intrinsic single electron spin coherence time in the system. This parameter is of fundamental interest and is looked after by numerous researchers in a wide range of systems.
In a one-dimensional electron gas (1DEG) with sufficiently low density, Coulomb interaction becomes so dominating that Wigner crystallization can occur. However, experimental evidence of such correlated electronic states has been elusive in conventional transport measurements for quantum wires, because these measurements are significantly influenced by the 2DEG contact leads attached to the wire, and do not directly probe the microscopic nature of Coulomb interactions in the wire. In this work, we prepare parallel-coupled quantum wires in a 2DEG defined by Schottky gates to study the Coulomb drag effect, which can directly reflect the Coulomb interaction. The distance between the two wires and the electron density in each wire are all tunable with gate voltages. We measure the dielectric response of one wire (drag-wire), which is connected to the 2DEG leads, to the external potential created by a current in the other wire (drive-wire) whose electron density is sufficiently low. When an incompressible Wigner crystal is formed in the drag-wire, there is no internal screening response while correlation holes appear in the contact leads and tend to attract the Wigner crystal. Then electrons are pumped from the drag-wire to the lead when a sliding mode of Wigner crystal is excited. This causes Negative Coulomb drag, in which electrons in the drag-wire are pumped “against” the bias direction of the drive-wire. We observe that this electron pump occurs equivalently for exchanging the drag-drive roles in coupled equal length wires but only for a short drag-wire in coupled long and short wires.
A single trapped atom/ion or semiconductor quantum dot in a high-Q cavity provides a unique experimental tool for studying the fundamental aspects of quantum electrodynamics and for implementing the quantum information processing systems. In this talk I will discuss how such a cavity QED system is mutually connected by coherent state bus to implement various quantum information systems such as quantum repeater, quantum memory and processor with suppressed decoherence. Physical implementation based on impurity bound electron/exciton in semiconductor microcavity is proposed.
Carbon nanotubes (CNTs) offer unique electrical properties such as the highest current density exceeding 109 A/cm2, ultra-high thermal conductivity as high as that of diamond, ballistic transport along the tube and the highest Young’s modulus. Because of these remarkable properties, they have been expected for use as future wiring materials to solve the problems of stress and electro-migration and heat removal in future ULSIs. They have been also expected for use as heat sinks for high-power devices to improve heat dissipation and high-frequency performance. This paper presents status of CNT material technologies and demonstrates the potential of metallic CNTstowards such practical applications.
Last Modified February 14, 2006 by the NSEC Office.