International Journal of High Speed Electronics and Systems

High radiative efficiency in moderately doped n-InP results in the transport of holes dominated by photon-assisted hopping, when radiative hole recombination at one spot produces a photon, whose interband absorption generates another hole, possibly far away. Due to "heavy tails" in the hop probability, this is a random walk with divergent diffusivity (process known as the Lévy flight). Our key evidence is derived from the ratio of transmitted and reflected luminescence spectra, measured in samples of different thicknesses. These experiments prove the non-exponential decay of the hole concentration from the initial photo-excitation spot. The power-law decay, characteristic of Lévy flights, is steep enough at short distances (steeper than an exponent) to fit the data for thin samples and slow enough at large distances to account for thick samples. The high radiative efficiency makes possible a semiconductor scintillator with efficient photon collection. It is rather unusual that the material is "opaque" at wavelengths of its own scintillation. Nevertheless, after repeated recycling most photons find their way to one of two photodiodes integrated on both sides of the semiconductor slab. We present an analytical model of photon collection in two-sided slab, which shows that the heavy tails of Lévy-flight transport lead to a high charge collection efficiency and hence high energy resolution. Finally, we discuss a possibility to increase the slab thickness while still quantifying the deposited energy and the interaction position within the slab. The idea is to use a layered semiconductor with photon-assisted collection of holes in narrow-bandgap layers spaced by distances far exceeding diffusion length. Holes collected in these radiative layers emit longwave radiation, to which the entire structure is transparent. Nearly-ideal calculated characteristics of a mm-thick layered scintillator can be scaled up to several centimeters.

With 30nm-class memory cells in production and 20nm-class (20-29nm feature-size) memory targeted for next year, the standard 1-Transistor + 1-Capacitor (1T+1C) DRAM industry is making prominent efforts to improve the scalability of the cell capacitor while maintaining the minimum capacitance requirements for state discrimination, immune to noise (C~25fF/cell). To achieve the capacitance requirement, the DRAM cell has evolved from its initial planar implementation to complex three-dimensional structures. The increment in complexity and the large difference in size between the transistor and capacitor of each cell have motivated the search for Floating-Body Single-Transistor DRAM (1T-DRAM). The underlying idea behind 1T-DRAMs is the development of single-device memory cells with a pronounced hysteresis effect and fast operation. This chapter is focused on the floating-body effect as a primary source of hysteresis. We present new concepts able to deal with the basic limitations of 1T-DRAM while maintaining its simplicity. The floating-body 1T-DRAMs can be reconciled with the aggressive scaling constrains by considering new ideas which make possible the coexistence of electron and holes in the same ultrathin transistor. The best approach is to isolate each type of carrier in an specific potential well which is not created specifically by the bias conditions (unlike standard 1T-DRAMs) but by the physical structure of the device.

Catalyst-free growth of GaN nanowires with molecular beam epitaxy produces material of exceptionally high quality with long minority-carrier lifetimes and low surface recombination velocity. The nanowires grow by thermodynamic driving forces that enhance the sticking coefficient of incoming reagents to the end facets of the nanowire while inhibiting growth on the m-plane sidewalls. Photoluminescence (PL) studies confirm that the material is essentially free of detrimental chemical impurities and crystalline defects. The nanowires are readily excited to lasing with modest optical pump power. Recent progress in methods for selective epitaxy has made it possible to control both the diameter and placement of the nanowires. Despite the high material quality, the energy-conversion efficiency of single nanowire LEDs remains low. The primary limitation appears to be optimizing the p-type doping with Mg, which is both a growth and a measurement problem.

Nanoscale field effect based sensors have emerged as a potential label-free diagnostic tool capable of detecting extremely low levels of biomolecules with fast response times. To date, successful detection of various biomolecular species ranging from small molecules to antibodies have been demonstrated, however, the lack of quantitative methods, sensor calibration techniques and ability to detect charges in strong ionic strength environments has hindered their commercial application. In this paper we discuss a recent progress in this field directed primarily towards overcoming the aforementioned obstacles.

This paper describes the design of digital logic circuits composed exclusively from magnetic devices. The logic level of a signal is embedded in the direction of steered currents, not voltages. The currents are steered by small (e.g., 2-3x) resistance changes. Sub-100 mV pulsed voltages power and synchronize the circuits. Logic gates are non-volatile, allowing for fully-pipelined logic that can achieve ultra-low energy for design examples.

To meet the demand of delivering ever-increasing Internet traffic, optical network must response by increasing its transmission capacity. Since transmission capacity of an individual fiber is still well exceed the capacity of transmitters (TXs) and receivers (RXs), wavelength-division multiplexing (WDM), in which many TXs and RXs at the transmitting ends of a fiber are used to send and receive many signals, becomes the necessary technology for increasing the transmission capacity of each link of an optical network. This trend, however, demands for increasing density not only the TXs and RXs, but all other components at the sending and receiving ends of communications links. As the number of wavelengths in WDM configuration getting greater, the number of all these components that must be placed on one board has to increase too; hence, the density of packaging comes to micro- and even nano-scale. The TXs and RXs are produced in arrays on a chip quite similar to production of VLSI electronic circuits. At that scale, traditional optical operations used today in an optical-communications technology, such as launching light into optical fiber from TXs and directing light from optical fiber into RXs, multiplexing and demultiplexing individual channels (wavelengths), and electro-optical (E/O) and opto-electrical (O/E) conversions become problems primarily because of the diffraction limit. The problems associated with the diffraction limit are particularly acute for optical interconnects. One of the possible solutions to all these—and some other—problems could be the use of plasmonics. In the last years, the optical-communications industry shows a great interest in developing this topic, as the growing number of publications and practical results can attest.

This paper consists of two parts. The first part reviews the current trends in application of plasmonics in optical communications and the second part discusses the theoretical foundation of the proposed WDM demultiplexer and offers the scheme of possible implementation of the device.

We have fabricated and investigated several types of GaN MOSFETs with normally-off operation. The recessed-gate GaN MOSFET is preferred for normally-off operation, because the threshold voltage (V_th) of the device can be easily controlled, but it suffers from relatively modest current drivability which must be improved by adopting appropriate device structure and/or process. Enhanced performances have been achieved in this work by combining the recessed-gate technology with additional processes, such as: the post-recess tetramethylammonium hydroxide (TMAH) treatment to remove the plasma damage, the post-deposition annealing of gate oxide to decrease the gate leakage current, the re-growth of n+ GaN layer for source/drain to improve the access resistance and V_th uniformity, the stress control technology to achieve extremely high 2-D electron-gas density (2DEG) on source/drain and decrease the series resistance, and the use of the p-GaN back-barrier to decrease the buffer leakage current. The GaN-based FinFET with very narrow fin was also investigated as a possible candidate for high performance normally-off GaN MOSFETs.

As a promising novel route towards highly efficient optoelectronic devices, GaN based 3D core-shell light emitting diodes (LEDs) have attracted increased attention in recent years. In comparison to conventional 2D thin film LED, the 3D LED systems using a core-shell geometry with high aspect ratio are a breakthrough in technology with lots of advantages. In this paper, we review our developed growth strategies of these LED systems on Si and sapphire substrates. A catalyst free selective area growth of GaN 3D core-shell LED systems was realized using patterned substrates by metal organic vapor phase deposition in a convenient continuous-flux growth mode. We have recently suggested that the surface polarity plays a crucial role for the morphology of GaN 3D structure growth. In order to analyze the surface polarity of 3D submicron or micron structures, Kelvin probe force microscopy and selective etching techniques have been developed. During the growth of GaN submicron rods and micron columns on patterned SiO_x/sapphire templates, mixed polarity effects could be detected. A new "truncated pyramid + column" growth method was developed to effectively avoid the formation of mixed polarity and realize single N-polar GaN 3D devices. Transmission electron microscopy and spatially and spectrally resolved cathodoluminescence measurements evidently prove the core-shell structure of 3D LED systems.

Power semiconductor devices are important for numerous applications with power conversion being an important one. Wide energy gap semiconductors SiC and GaN have properties that make them attractive for such applications. Among these properties are high thermal conductivity, high breakdown electric field, wide energy gap, low intrinsic carrier concentration, high thermal stability, high saturation velocity and chemical inertness. These lead to low on-resistance, high breakdown voltage, high frequencies, small volume, and small passive inductors and capacitors. These desirable properties are offset by the higher material costs and higher defect densities. Although wide energy gap devices have been in development for many years, only recently have they become available commercially. Their main competition is silicon power devices with breakdown voltages up to 8000 V and very high surge current capacity. However, silicon power devices are approaching their material limits and wide energy gap devices are beginning to have an impact in the power electronics space. SiC has the advantage of substrates with diameters approaching 150 mm and the ability to grow thermal SiO₂. GaN has the heterojunction advantage, but no viable substrate technology. In fact, a large portion of SiC production is used for GaN substrates. GaN material development has also benefited significantly from the development of optical devices, e.g., light-emitting diodes and lasers.

The basics of the self-assembled growth of GaN nanorods on Si(111) are reviewed. Morphology differences and optical properties are compared to those of GaN layers grown directly on Si(111). The effects of the growth temperature on the In incorporation in self-assembled InGaN nanorods grown on Si(111) is described. In addition, the inclusion of InGaN quantum disk structures into self-assembled GaN nanorods show clear confinement effects as a function of the quantum disk thickness. In order to overcome the properties dispersion and the intrinsic inhomogeneous nature of the self-assembled growth, the selective area growth of GaN nanorods on both, c-plane and a-plane GaN on sapphire templates, is addressed, with special emphasis on optical quality and morphology differences. The analysis of the optical emission from a single InGaN quantum disk is shown for both polar and non-polar nanorod orientations.

We discuss physics, design, fabrication, performance, and selected applications of Deep Ultraviolet Light Emitting Diodes (DUV LEDs). Our analysis reveals the relative contributions of electrical injection, internal quantum efficiency, and light extraction efficiency to the overall DUV LED performance. Our calculations show that the reduction of the dislocation density at least below value of 2×10⁸ 1/cm³ is necessary for reaching high DUV LED efficiency. Better light extraction has been achieved using an innovative p-type transparent sub-contact layer and reflecting ohmic p-type contact resulting in nearly tripling DUV LED power. At high power dissipation, temperature rise might be significant, and we present data showing the power degradation with temperature increase and the results of the detailed 1D and 3D analysis of thermal impedance of DUV LEDs. As an example of DUV LED application, we report on microbial disinfection using 19 watt 275 nanometer DUV LED.

Metal-semiconductor field-effect-transistors (MESFETs) have been fabricated using a commercially available 45nm silicon-on-insulator (SOI) CMOS foundry with no changes to the process flow. Depending upon the layout dimensions, these n-channel, depletion mode devices can be designed for high current drive (I_D^SAT ≥ 100mA/mm), high operating frequency (f_max > 35 GHz) or enhanced breakdown voltage (V_BD > 25V). The design flexibility provided by the SOI MESFETs, coupled with the high performance of ULSI CMOS at the 45nm node will enable a variety of analog, RF and mixed signal applications.

GaInSb and AlGaInSb compositionally graded buffer layers grown on GaSb by MBE were used to develop unrelaxed InAs_1-xSb_x epitaxial alloys with strain-free native lattice constants up to 2.1% larger than that of GaSb. The in-plane lattice constant of the strained top buffer layer was grown to be equal to the native, unstrained lattice constant of InAs_1-xSb_x with given x. The InAs_0.56Sb_0.44 layers demonstrated a photoluminescence (PL) peak at 9.4 μm at T = 150 K. The minority carrier lifetime measured at 77 K for InAs_0.8Sb_0.2 was 250 ns.

We discuss the single-mode performance characteristics of midwave-infrared interband cascade lasers. Broad-area devices with 5 active stages display pulsed threshold current densities as low as 375 A/cm² and threshold power densities as low as 920 W/cm² at room temperature, owing in part to the suppression of Auger recombination. Narrow ridges were processed using optical lithography to incorporate a periodically-corrugated pattern into the sidewalls. The corrugations are intended to suppress lasing in higher-order lateral modes, and also provide a 4^th-order grating for distributed feedback. A corrugated-sidewall device operating at T = -20°C produced more than 46 mW of cw output in a single spectral mode, with a peak wallplug efficiency of 7.6%. The device maintains single-mode operation at current densities up to 10 times the lasing threshold, and the single-mode tuning range is 26 nm if both current and temperature are varied. Another device operating at -23°C produces up to 31 mW of single-mode power in the 3.3142-3.3164 μm range that spans several of the strongest absorption signature lines for methane.

Hybrid nanocomposite thin films, composed of inorganic colloidal quantum dots (CQDs) embedded in a matrix of organic conjugated polymer, have shown promise as a method for room-temperature infrared detection due to the three-dimensional confinement of the CQD and significantly lower dark currents compared to inorganic detectors. However, in order to improve device performance, the excited charges must be efficiently promoted out of the CQD, which is surrounded by an insulating surface ligand. These short, organic molecules, which are required to prevent agglomeration of CQDs in solution, have been shown to inhibit charge transfer into and out of the CQD. In this work, the transfer matrix method is utilized to calculate the quantized energy levels and wavefunctions in the conduction band of the CdSe CQD for a variety of surface ligand materials. These results are used to calculate the absorption coefficient for a size distribution of CQDs and are compared with measured Fourier Transform Infrared absorbance spectra. Finally, the effect of the ligand on the calculated absorption coefficient will be used to optimize the design for an infrared photoconductor.