The impact of the recently demonstrated time-stretch preprocessing on the performance of Analog-to-Digital Converter (ADC) is reported. We calculate the influence of time-stretch on the quantization noise, thermal noise and shot noise of Nyquist type ADC. We also study the effect of time-stretch on the quantization noise of (Sigma) -(Delta) converters. The results suggest that if ideal time stretching can be achieved, the resolution is increased by 1.5 and 2.5 bits for every octave of time-stretch in a first and second order modulator, respectively.

A novel velocity-matched distributed balanced photodetector operating at 1.3 and 1.55 micrometers wavelengths has been experimentally demonstrated. Distributed absorption and velocity matching of the optical and microwave signals are employed to achieve high saturation photocurrent. A common mode rejection ratio greater than 27 dB has been achieved over a wide range of photocurrents. More than 24 dB suppression of the relative intensity noise of the laser source and EDFA added noise has been achieved in a broadband RF link experiment. Shot noise limited performance has been achieved with significant improvements in the signal-to- noise ratio.

Foundations and principles of radio lenses constructing of centimeter and millimeter wave ranges with controlled refraction index, combining the quality of phased array antennas with optical devices are stated. Possibilities of the electronically scanning with wide-angle sector and high coefficient of amplification are maintained.

The purpose is full-wave electromagnetic analysis of integrated waveguiding structures containing active semiconductor or magnetized ferrite films serving as the means for propagating high-speed signal and accurate modeling of integrated waveguiding structures at an electrodynamic level taking into account physical effects. On the basis of the solution of the boundary electrodynamic problem, the electrodynamic theory for integrated waveguiding structures with the excitation of space charge waves in active semiconductor layers or magnetostatic waves in magnetized ferrite films is developed. A method for the design on the electrodynamic accuracy level mathematical models of millimeter-wave active semiconductor electron devices having distributed parameters and microwave spin- wave devices is elaborated taking into account the distributed interaction of waves with the different physical nature.

The 3D electrodynamic accuracy level models of irregular waveguiding transmission lines with nonlinear active semiconductor or gyromagnetic elements based on solving Maxwell's equations completed by equations of charge- carriers motion in a semiconductor or magnetization vector motion in a ferromagnetic are elaborated. The new decomposition approach to solve nonlinear 3D boundary electrodynamic problems for irregular waveguides is developed using nonlinear autonomous blocks. The stable numerical algorithm for determining the rigorous electrodynamic solutions corresponding to both the electromagnetic waves diffraction or the generation by nonlinear active semiconductor elements and the parametric amplification by bounded nonlinear gyromagnetic components in integrated microwave devices is created. Accurate modeling of microwave and millimeter-wave strip lines and resonators with a planar geometry Gann-effect diode or bounded nonlinear magnetized ferrite films elements is done.

High performance photodiodes are essential for photonic insertion into Phased Array Antenna Systems. This paper discusses the RF linearity performance of photodiodes with consideration of thermal effect at high photocurrent, and presents new understandings of both surface-normal and waveguide photodiodes using an equivalent circuit model analysis. The analysis is accompanied by a novel diagnostic technique for robust examination of photodiodes.

This paper presents an optical system which enables a broadband RF signal to be detected and delayed by a traveling fringes detector (TFD) using an acousto-optic deflector (AOD) and a 4f imaging system. The TFD is based upon the synchronous drift of photo-generated carriers with a moving interference pattern; the moving interference pattern is generated by interfering two coherent beams of light at different frequencies. Light which is incident on the photoconductive layer of the detector will generate photocarriers with a specific drift velocity proportional to the applied bias voltage. For a fixed angle between the two beams, a resonance peak occurs when the drift velocity equals the fringe velocity of the moving interference pattern. Detection of a broadband signal, therefore, is difficult since each frequency component produces a different fringe velocity and thus has a different resonance peak associated with the detector. Broadband detection of a signal is allowed by forcing each of the detected moving interference patterns, each corresponding to a specific temporal RF frequency, to have the same velocity as the electron drift velocity. This can be accomplished by using an AOD to linearly deflect each frequency component of the RF signal at the appropriate angle in order to maintain a constant overall fringe velocity at the TFD.

Large-aperture biased photoconductive emitters which can generate high-power narrow-band terahertz (THz) radiation are developed. These emitters avoid saturation at high fluence excitation and achieve enhanced peak power spectral density by employing a thick layer of short lifetime low- temperature-grown GaAs (LT-GaAs) photoconductor and multiple-pulse excitation. THz waveforms are calculated from the saturation theory of large-aperture photoconductors, and a comparison is made between the theory and the measurement. A direct comparison of the multiple-phase saturation properties of terahertz emission from semi-insulating GaAs and LT-GaAs emitters with different carrier lifetimes reveals a strong dependence of the multiple pulse saturation properties of terahertz emission on the carrier lifetime. In particular, the data demonstrate that saturation is avoided only when the interpulse spacing is longer than the carrier lifetime.

We report on a GaAs/AlAs superlattice detector as a novel direct detector and autocorrelator for THz radiation. It is based on a doped wide-miniband GaAs/AlAs superlattice, with submonolayer AlAs barrier layers; the superlattice is operated at room temperature. THz radiation, generated by a free-electron laser and a mode locked p-Ge laser, was coupled into the superlattice via a corner cube antenna system. THz-irradiation of the biased superlattice resulted in a current reduction, which was monitored. The direct detector showed a fast response (20 ps, limited by the electronic circuit) and was robust against intense radiation pulses (peak power 10 kW). The responsivity was 100 times higher than the responsivity of detectors of comparable risetime and comparable robustness. Intense THz radiation caused a complete suppression of the current through the superlattice. This is the basis of the superlattice autocorrelator. The superlattice autocorrelator could resolve picosecond radiation pulses.

We describe a method for precise pulse-repetition-frequency tuning of monolithic mode-locked laser diodes (MLLDs) by means of loss-induced change in the effective length of the distributed Bragg reflector. With this method, 39.8131-GHz (the SDH frequency) operation is achieved in a frequency tuning range of 1 GHz. The novel application of mode-locked laser diodes to all-optical clock extraction, one of the essential functions required in all-optical signal processing, is demonstrated at the 40-GHz SDH frequency.

We present a novel optical-optical switching technique for modulation of infrared radiation. The modulation response is based upon the optical perturbation of semiconductor layers within an air-filled metal-clad semiconductor waveguide. Generation of the electron-hole plasma within these layers is via femtosecond pulses of above bandgap radiation (800 nm). The propagation characteristics of this five-layer structure are analyzed through the coupling of quasi-static electromagnetic analysis to the time-varying optical properties of the semiconductor layers. It is found that the device is able to modulate radiation at various frequencies, though we specifically investigate modulation of 10.6 micrometers radiation. At this wavelength, an electron-hole photoinjection density of approximately 1 X 10¹⁸ cm^-3 in the semiconductor layers provides an extinction ratio of 30 dB. The significance of this modulation depth and possible applications to all-optical Mach Zehnder metal-clad semiconductor modulators and self- limiting switches are discussed.

We present our investigation of a variable bandwidth, ultrafast magneto-optic waveguide modulator employing the ferromagnetic resonant precession of the magnetization in a Bizmuth-substituted YIG film. Ultrafast magnetic field pulses, produced by current pulses propagated through a high-speed transmission line, are used to modulate the optical beam in the film. Such a device is capable of multi- gigahertz bandwidth and potential integration into current IC technology. Single frequency modulation of an 800 nm optical carrier is measured at bandwidths tuned between 4.6 and 11.7 GHz. The bandwidth is tuned by means of an externally applied DC magnetic field. It is found that the mode conversion efficiency of a waveguide type of device is limited by the linear birefringence of the thin Bi-YIG film. Birefringence-compensating schemes are discussed to optimize the mode conversion.

A new type of metallic electromagnetic structure has been developed that is characterized by having high surface impedance. The geometry is analogous to a corrugated metal surface in which the corrugations have been folded up into lumped circuit elements, and distributed in a 2D lattice. Although it is made of continuous metal, and conducts DC currents, it does not conduct AC currents within a forbidden frequency band. Unlike normal conductors, this new surface does not support propagating surface waves. Furthermore, image currents induced in the surface are not phase reversed as they are on a flat metal surface.

A simple and fast fiber-grating model for the simulation and design of 1D photonic bandgap structures consisting of a row of circles etched in the ground plane of microstrip lines is proposed. It is based on the relationship found between the parameters of the photonic bandgap structure and those for an equivalent fiber Bragg grating with the same frequency- response shifted to the optical wavelength range.

A magnetic field Green's function expressed as an eigenmodes expansion and based on the plane wave method is formulated first for an infinite magnetic current line embedded in an unbounded 2D photonic crystal (PC) and then for a magnetic dipole embedded in a 2D PC truncated by two metallic plates. The underlying idea of analyzing a slot antenna printed on a 2D PC with a standard method of moment through the principle of equivalence is shown to motivate the present investigation. A complete solution for the line problem is derived, whereas the inadequacy of the method in nits present form for the dipole problem is demonstrated rigorously. Numerical results of the Green's function for the first problem are shown for different positions of the source, and a discussion about radiation patterns, asymptotic behaviors and convergence problems of the Green's function is proposed.

Present are numerical studies of the magnetic permeability for arrays of artificial magnetic molecules simulated using a time domain TLM code. These artificial magnetic materials consist of a 3D periodic lattice of electrically small loaded loops suspended in a non-magnetic host medium. For this class of artificial magnetic media, we demonstrate good agreement between the permeability computed using a simple circuit theory model, and that computed using a full wave TLM simulation. This close agreement suggests that the salient physics for this type of artificial magnetic media may be well modeled using simple lumped equivalent circuits. A closed form expression is derived for the effective media permeability as a function of the molecular circuit loads. When molecules are uniformly loaded with lossless capacitors, the artificial media exhibits a Lorentzian response with a resonance ((mu) _r yields (infinity) ) below which the media is paramagnetic ((mu) _r > 1) and above which the media is diamagnetic ((mu) _r < 1). Resonant frequency, and magnetic permeability, can be adjusted by controlling the load capacitance.

The paper presents new results in the field of super high- speed signal processing. For this purpose we suggested to write digital information into spatial structures (topology) of electromagnetic field pulses. It allows to use passive circuits for fulfillment spatial logical operations. Main physical effects in solids and micron circuits, which influence on time delay of signals, are considered. In accordance to the physical analysis passive circuits of micron size allow to process spatially modulated signals with time delay less than one picosecond. This fact is confirmed by theoretical modeling transients in strip transmission circuits elements and in logical components for spatial signal processing. It is considered theoretical results of modeling several logical circuits for spatially- modulated signal processing and experimental technique for their researchers. The best field of application of the new circuits is quasi-holographic circuits for parallel signal processing. The considered results are interesting in the field of super high-speed applications, microwave-optical engineering, terahertz technique.

Practical implementation of millimeter-wave quasi-optical amplifier arrays will require high device uniformity across the array, efficient coupling to and from each gain device, good device-to-device isolation, and efficient heat removal. This paper presents techniques that address these issues for a 44 GHz MMIC-based design. To improve device uniformity, a double selective gate recess approach is introduced which results in a demonstrated 3 - 5X improvement in uniformity when compared to Raytheon's standard production pHEMT process. For packaging, direct backside interconnect technology (DBIT) is introduced as a bondwire-free scheme for connecting each amplifier to the array. This approach significantly reduces interconnect loss by reducing interconnect inductance. Measured insertion loss at 44 GHz for the DBIt transition is 0.35 dB compared to 2.3 dB for a typical bondwire transition produced on a manufacturing automated bonding machine. By eliminating bondwires which tend to radiate at millimeter wave frequencies, the DBIT approach also significantly improves the device-to-device isolation, thereby improving the array stability. The DBIT approach would not be viable if it could not effectively dissipate heat (a typical 25 watt array generates greater than 100 watts of heat). Finite element thermal analysis results are presented which show that the DBIT approach adds a tolerable 15.5 degree(s)C temperature rise over a standard solder-based MMIC die-attach to a heatsink. Thus, the DBIT approach, along with the double selective gate recess process, provides an attractive, low-loss, bondwire-free approach for producing uniform amplifier arrays.

Because of power limitations in single semiconductor devices, power combining is usually applied in the frequency ranges of short millimeter and of submillimeter waves. Multiple-device circuits are hence found for fundamental and harmonic frequency sources and for amplifiers using both signal splitting and combining circuits. In receiver technology, on the other hand, signal splitting circuits are also inherent to subharmonically pumped downconverters and to mixers for focal plane arrays. Hence a suitable multiple- device circuit technology is a necessary prerequisite for many--if not most--applications. Hence power combining and splitting are important circuit functions in the frequency bands of millimeter and submillimeter waves. It is shown that their tasks can almost ideally be fulfilled by using computer-generated holograms. The inherent features of this method are its scalability up to arbitrarily high frequencies, high circuit efficiency which almost approaches unity, broad bandwidth performance, robustness and simplicity of realized circuits. As examples, multi-device oscillators with Impatt-diodes, frequency multipliers with varactor diodes, and subharmonically pumped mixers with Schottky-diodes and arbitrary order of the subharmonic frequency are treated. Basic circuit configurations together with some preliminary experimental results are also presented. It is believed that utilizing the principles of holography in multiple device circuits will establish a new circuit technology for the short millimeter wave and submillimeter wave ranges which is characterized by high electrical performance, relatively simple circuits, and high flexibility and which simultaneously will overcome the problem of generating high enough output power levels from all-solid-state sources.

There is a significant interest in the area of improving high temperature stable contacts to III-V semiconductors. Two attractive materials that offer promise in this area are dysprosium phosphide (DyP) and dysprosium arsenide (DyAs). This paper reports the electrical characterization of MBE- grown DyP and DyAs on GaAs, GaP, and InP substrates. The characterization methods include Hall and I-V measurements. DyP is lattice matched to GaAs, with a room temperature mismatch of less than 0.01% and is stable in air with no sign of oxidation, even after months of ambient exposure. DyP forms Schottky contacts to n-GaAs, n- and p-GaP, and p- InP with barrier heights of 0.81, 0.9, 0.8 and 0.74 eV, respectively. DyP on n-InP and p-GaAs is found to have ohmic behavior with the specific contact resistance of 1 X 10^-4 and 2.9 X 10^-5 (Omega) (DOT)cm², respectively. DyAs also forms Schottky contacts to n-GaAs, p-InP and forms ohmic contacts to n-InP.

Cu₃Ge was studied as an Ohmic contact to n-GaN and n- type GaAs. Specific contact resistance of Cu₃Ge to n- type GaAs was found to be sensitive to annealing conditions, doping concentrations, and Ge compositions. After vacuum annealing at 400 degree(s)C for 30 min, Cu₃Ge exhibited ohmic behavior to n-type GaN with doping concentrations of approximately 1.0 X 10¹⁸ cm^-3. Unprotected Cu₃Ge ohmic contacts suffered from oxidation when exposed at temperatures higher than 300 degree(s)C. Aging tests at 400 degree(s)C where Cu₃Ge covered with TiW and Au was used as ohmic contact to n-type GaAs, and TiPtAu covered with Au to p-type GaAs, revealed unchanged I-V characteristics after 120 hr annealing which showed that this contact was suitable for device application at high temperature. Pseudomorphic HEMT employing protected Cu₃Ge ohmic contacts and Ti/Pt/Au gates has achieved peak transconductance of 330 mS/mm at room temperature for 2- micrometers long gate. The I-V characteristic of Pd/Al, covered with TiW/Pt changed from Ohmic to Schottky after aging at 350 degree(s)C for 2 hours. By depositing a very thin Cr layer between Ti/Al and Au layers, contact resistance of Ti/Al contacts remained the same even after an aging of 130 hr at 350 degree(s)C and 130 hr at 400 degree(s)C. However, specific contact resistance increased from 2.4 X 10^-6 (Omega) cm² to 3.8 X 10^-6 (Omega) cm² after annealing at 500 degree(s)C for 2 hr.

This work focuses on the thermalization and relaxation behaviors in ultrafast heat transport, ranging from sub- nanoseconds to picoseconds. The unique phenomena in thermal lagging, including ultrafast thermalization at short times, thermal exaggeration in amorphous materials, and thermal resonance in high-frequency excitations, are first summarized to draw special attention to the thermal design for ultrafast photonic devices. Temperature jumps across the interface between dissimilar materials follow, which include additional effects of phonon scattering in the interfacial area. Parameter groups are identified to characterize the temperature jumps to avoid the delamination damage due to thermal straining.

Joule-Thomson (JT) coolers have been widely used for cooling optoelectronic devices and for cryogenic applications. In the present investigation the performance and behaviors of a JT cooler fabricated from micro tubes of different diameters were studied. The cooler was comprised of circular tubes with diameters ranging from less than a hundred microns to a couple of millimeters. The smallest tube serves as a throttling device while the other tubes were used to fabricate a concentric-tube heat exchanger. Temperature drops were measured for nitrogen gas flowing through capillary tubes of different diameters and lengths. Gas dynamic theories were employed for analyzing the high- pressure gas flow in the JT cooler. Friction choking was observed under normal operating conditions, with strong compression and expansion waves appearing at the exit of the throttling tube. The simple design and configuration of the present JT cooler makes it suitable for batch fabrication using the photo lithography technique if the circular tubes are replaced by etched micron channels. This attractive feature of the micro-tube JT cooler can facilitate the integration of high-power optoelectronic devices and their cooling systems.

First, a micro cross flow heat sink made of (110)-orientated silicon is fabricated by bulk micromachining, and wafers with hundreds of high aspect ratio channels are bonded together by diffusion bonding with aluminum as medium layers. The core of the micro heat sink is about 0.918 cm^-3, and the density of the heat transfer area if 15,294 m²/m³. Using pure water as the working fluid, as the maximum pressure drop reaches 2.47 bar, the flow rate is greater than 4.5 l/min, and that makes the overall heat transfer coefficient up to 24.7 kW/m²K, corresponds to a volumetric heat transfer coefficient of 188.5 MW/m³K. Second, we fabricate a series of 40 micrometers wide and 30 mm long micro channels with different aspect ratio is 12 cm² (110) silicon wafer area. The silicon is anodic bonded with #7740 Pyrex glass to form the micro channel structure. Using pure water as the working fluid, hydraulic resistance and thermal performance of the micro channel structure are evaluated experimentally. As the maximum pressure drop reaches 3.3 bar, the flow rate is 0.3 l/min, and that make the heat transfer coefficient up to 32,598 W/m²K.

Antenna-coupled thin-film niobium microbolometers designed for mm-wave imaging arrays have been fabricated and their electrical properties measured. The niobium bolometers are operated at room temperature and are located at the feeds of half-wave dipole antennas on an electrically thick silicon substrate. We independently measured the resistance of the microbolometers as a function of temperature and the dynamic resistance as a function of DC bias current, yielding the electrical responsivity and thermal conductance of the bolometers. The latter is dominated by conduction through a deposited SiO₂ isolation layer. We describe electrical measurements of the bolometers' pulse response and noise, performed by illuminating the bolometers with a high power, pulsed mm-wave source. We discuss these measurements in the context of the pulsed detection architecture planned for an active mm-wave imaging system currently under development.

We firstly demonstrate the effects of the air-bridges on the coplanar waveguide (CPW) for the single-mode pulse propagation using a picosecond time-domain pump-probe measurement. The picosecond photoconductive sampling technique is able to measure the transmitted and reflected pulses with a picosecond temporal resolution corresponding to a 300-GHz frequency bandwidth. It is found that more than two of the crossing air-bridges over the CPW are needed to support the single-mode propagation. The crossing air-bridge is found to behave as a bypass capacitor to an incoming pulse. For the air-bridges used for connecting the ground planes along the CPW, the length of the air-bridge is founded more critical. The air-bridges have exhibited more than 30-dB return loss up to 300-GHz. Finally, the s- parameters of open and short terminations were measured using the embedded CPW with the proposed air-bridge structures.

Grating-assisted codirectional coupler (GACC) filter is considered to be one of the key devices in wavelength division multiplexing (WDM) system because of its superior characteristics of fast and wide range wavelength tuning as well as narrow bandwidth. However, its application has been limited because of a high level sidelobe. In this paper, we propose and demonstrate a novel concept of pair grating in order to suppress the sidelobes in GACC filter. Two different bandgap InGaAsP layers lattice-matched to InP were used as waveguides. A grating layer is inserted between two waveguides. In our structure, a pair grating whose period is 2(Lambda) was composed of two unit gratings and the separation between the unit gratings can be varied. By adjusting the separation, the coupling strength, K(z), can be controlled as Hamming or raised cosine functional form of position variable z. This idea of pair grating is attractive because averaged effective optical index or propagation constant remains unchanged. We fabricated a GACC filter with a weighted coupling strength, and achieved -19 dB sidelobe level, which implies a reduced crosstalk between neighboring channel. This device would be useful for WDM system requiring fast wavelength tuning.

Low cost and reliable passive components are essential to further span the use of fiber optics and realize the all- optical communication networks. Silica waveguide technology has played an important role in the development of passive components. Devices of 1 X N, 2 X N splitters, and 1.3/1.55 WDMs have been mass-produced for practical applications. Recently, large volumes of array waveguide gratings have also been produced for dense WDM applications. The optical fiber preform manufacturing process, flame hydrolysis deposition is adapted to deposit low loss silica glass on planar substrates (silicon, quartz or alumina ceramics). Photolithography and reactive ion etching is then applied to pattern various types of integrated waveguide circuits. Testing, fiber-connecting, and device packaging follow the circuit fabrication to produce the fiber- pigtailed modules. The technology provides a versatile means of building passive components. In this paper, the manufacturing processes are reviewed and the functions and performance of various circuits are discussed with an emphasis on the current status of the array waveguide gratings.

In this paper we present the theoretical analysis and the measurements of a quasi-optical band-pass filter, operating at 280 GHz. The filter consists of a metal screen, perforated periodically with cross-shaped apertures. The analysis is performed by the Method of the Moments (MoM), using entire domain basis functions. The Boundary Integral- Resonant Mode Expansion method is used in the calculation of the MoM matrices. The transmission and phase shifting characteristics of the band-pass filter were measured with a Terahertz Time-Domain Spectrometer and are compared with the theoretical results. The effects of the smoothness in the cross boundary, due to the fabrication process, are also discussed.

The phased array composed from spiral microstrip radiators is considered. This array can be applied as radiation focusing spiraphase system. The lack of traditional phase shifters ensures good cost, technological and constructive characteristics of spiraphase arrays in a microwave range. Antenna construction and mathematical model are considered. This device permits to realize a focusing of antenna feed field by a microstrip antenna array. The microstrip element shape is arbitrary. The structure contains controlled elements (crystal diodes). These array electrodynamic characteristics are controlled by diodes. The mathematical model is obtained by integral equation method. Microstrip spiral radiators current distribution and input impedance are investigated. The numerical results can be used to develop antennas with optimum parameters.

FIRST (Far InfraRed and Submillimeter Telescope) is a European science mission that will perform photometry and spectroscopy in the 80 - 670 micrometers range. The proposed heterodyne instrument for FIRST is a seven-channel receiver, which combines the high spectral resolving capability (0.3 - 300 km/s) of the radio heterodyne technique with the low noise detection offered by superconductor-insulator- superconductor and hot electron bolometer mixers. It is designed to provide almost continuous frequency coverage from 480 - 2700 GHz. The Jet Propulsion Laboratory is responsible for developing and implementing the local oscillator sources for the 1200 - 2700 GHz mixers. The present state-of-the-art approach for millimeter-wave multipliers, based on waveguide blocks and discretely mounted devices, becomes harder and harder to implement as the frequency range is extended beyond 300 GHz. This talk will focus on the technology that is being developed to enhance and extend planar integrated Schottky devices and circuits to meet mission local oscillator requirements. The baseline approach is to use GaAs power amplifiers from 71 to 115 GHz followed by a series of planar Schottky diode varactor multiplier stages to generate the required LO signal. The circuits have to be robust, relatively easy to assemble, and must provide broad fix-tuned bandwidth. A number of new technology initiatives being implemented to achieve these goals will be discussed. Approaches include quartz-based and substrate-less diode circuitry and integrated GaAs membrane technology. Recent results and progress-to-date will be presented.

Subpicosecond electrical pulses have been generated from low temperature grown GaAs (LT-GaAs) coplanar stripline photoconductive switches by illumination of 160 fs optical pulses at 1.55 micrometers , which corresponds to below bandgap radiation. The device optical response was attributed to two photon absorption (TPA) as confirmed by the observed quadratic dependence of the DC photocurrent on average incident illumination. At a bias of 39 V and 5.16 mW average illumination a 2.6 (mu) A photocurrent was measured which corresponded to a 0.5 mA/W switch responsivity. For similar switch geometries fabricated onto LT-GaAs and illuminated by above bandgap radiation, responsivities of a few mA/W have been measured. Therefore TPA switch triggering has retained acceptable responsivity levels. The contribution of TPA to the device photoresponse was also confirmed by the dependence of the photocurrent on incident polarization direction; the TPA coefficient exhibited a strong anisotropy with respect to polarization. The device generated electrical pulses (impulse photoresponse) were measured using double sliding-contact photoconductive sampling techniques. Impulse photoresponses with full widths at half maximum of 451 fs were observed which are, to date, the fastest responses measured from LT-GaAs switches at this wavelength. Fourier transformation of the time domain data revealed electrical transients with 3 dB bandwidths of 190 GHz. The feasibility of back illumination of these devices at 1.55 micrometers was also demonstrated. Based on these highly encouraging results freely positionable photoconductive switches were fabricated and at present they are being tested.

We describe progress in the development of a calibration target for use in the EOS-MLS 2.5 THz radiometer on NASA's CHEM-1 spacecraft. Although the intended use is as a stable, isothermal black body load at a frequency of 2.5 THz, the design is suitable for use throughout the far-infrared. A wedge design is used for the target body to enhance the emissivity to desired levels at 2.5 THz. The body is machined from aluminum, giving the best trade between issues such as cost, thermal conductivity, mass, and strength. The target utilizes a white coating to reduce the destabilizing effects of periodic solar illumination. The coating can be made relatively thin to allow accurate temperature measurements of the FIR-absorbing medium. Emissivity of greater than 0.99 is achieved at 2.5 THz, while the solar absorptance is estimated at <0.5.

We have developed prototype HEB receivers using thin film superconducting NbN devices deposited on silicon substrates. The devices are quasi-optically coupled through a silicon lens and a self-complementary log-specific toothed antenna. We measured DSB receiver noise temperatures of 500 K (13 X hf/2k) at 1.56 THz and 1,100 K (20 X hf/2k) at 2.24 THz. Noise temperatures are expected to fall further as devices and quasi-optical coupling methods are being optimized. The measured 3 dB IF conversion gain bandwidth for one device was 3 GHz, and it is estimated that the bandwidth over which the receiver noise temperature is within 3 dB of its minimum value is 6.5 GHz which is sufficient for a number of practical applications. We will discuss our latest results and give a detailed description of our prototype setup and experiments. We will also discuss our plans for developing focal plane arrays with tens of Hot Electron Bolometric mixer elements on a single silicon substrate which will make real time imaging systems in the THz region feasible.

A novel, high isolation photonic mixer/frequency converter architecture consisting of two cascaded balanced bridge Mach-Zehnder modulators is presented. When compared with a traditional cascaded single output Mach-Zehnder modulator architecture, the new device can effectively minimize the breakthrough of microwave input and local oscillator signals at the intermediate frequency output for high electrical isolation applications. Device operation principles, requirements on coupling angles, and phase bias offset at each modulator section are analyzed for performance optimization. High isolation photonic mixing has been obtained experimentally using a single chip lithium niobate device. The leakthrough of the input and local oscillator signals was suppressed over 50 dB at the intermediate frequency output. This photonic mixer can be used in fiber- optic microwave links for low noise and low distortion millimeter-wave frequency conversion applications, particularly when the frequency response is limited by the millimeter-wave input ports. When operating in the upconversion mode, this photonic mixer can extend the output bandwidth to the sum of each modulator in cascading, as demonstrated in our experiment.

The role of integrated device/circuit simulation is critical to understanding the gigahertz-photonic operation of photomixing circuits containing metal-semiconductor-metal (MSM) devices. This work presents an efficient convolution- based time-domain approach to circuit simulation that incorporates an advanced numerical MSM device model. Complete millimeter-wave circuit simulation requires consideration of both the dynamic, high-frequency behavior of the electron and hole charge carriers in the large-signal device, and the frequency-dependent, distributed nature of the embedding circuit. The modeled device is an MBE-grown GaAs MSM photodetector with trench electrodes. Device and circuit performance is assessed by calculating the optical responsivity and bandwidth. Simulations with the device alone demonstrate the effects of a new current density boundary condition, as well as the effects of using low- growth- versus conventional-growth-temperature GaAs MSM's. Global simulations illustrate the effect that the embedding circuit has on bandwidth. Both types of simulations aid in the co-design of device and circuit, with applications to millimeter-wave generation in phased-array antennas and optoelectronic-based communication systems.16

In this work, a numerical de-embedding procedure, called `short-open calibration' (SOC), is originated and proposed for accurate parametric extraction of planar integrated circuits. This SOC technique is developed by defining two numerical calibration standards, namely, short and open elements, which allow to extract equivalent circuit parameters from a full-wave method of moments (MoM) in a deterministic format. With this scheme implemented in the MoM algorithm, planar integrated circuits and discontinuities can be accurately characterized in terms of their respective unified circuit models, which can account for effects of frequency dispersion, high-order modes and radiation losses. The SOC-extracted circuit parameters for several selected examples are obtained in this paper to exhibit some distinct quasi-lumped circuit behavior of those planar structures over low-frequency range. The unified circuit model is very useful for the optimized design of electrically large planar circuits by using well-established efficient tools based on analytical network synthesis or optimization. Two planar band-pass filters are thus designed with cascaded circuit topologies and simulated electrical characteristics are well confirmed by our measurements.

Multi-layer MMIC spiral transformers for 4:1 impedance transformation is investigated by using the FDTD method in this paper. The transformer employs three vertical transmission line layers, and the modeling is carried out for the properties and performance of the impedance transformer over a wide frequency range. The results obtained shows that the impedance transformer performs quite reasonably in a wide bandwidth.

In this paper, the importance of the global modeling concept is emphasized. Some of the recent developments in this regard are highlighted. To describe the different electromagnetic interactions in high frequency devices a full-wave transport model that incorporates full solution of Maxwell's equations inside the device is presented. Results from the model simulation that reveals the RF characteristics of both a conventional MESFET and an Air- Bridged gate MESFET are provided. In a second part of this paper, the problem of simulating large electromagnetic structure is addressed. A method based on the concept of time-domain impedance concept is proposed and its potential as solution to the problem in hand is highlighted.

In this paper, a new absorbing-boundary condition derived from z-domain is presented and implemented for the transmission-line matrix method (TLM). When introduced for the 3D-TLM symmetrical condensed node simulations, it shows an excellent performance in suppressing instability caused by spurious modes. The numerical stability of the absorbing boundary operators has been studied.

RF and mm-wave photonic devices and circuits have been developed at Sandia National Laboratories for applications ranging from RF optical data links to optical generation of mm-wave frequencies. This talk will explore recent high- speed photonics technology developments at Sandia including: (1) A monolithic optical integrated circuit for all-optical generation of mm-waves. Using integrated mode-locked diode lasers, amplifiers, and detectors, frequencies between 30 GHz and 90 GHz are generated by a single monolithic (Al,Ga)As optical circuit less than 2 mm in its largest dimension. (2) Development of polarization-maintaining, low- insertion-loss, low v-pi, Mach-Zehnder interferometer modulators with DC-to-potentially-K-band modulation bandwidth. New low-loss polarization-maintaining waveguide designs using binary alloys have been shown to reduce polarization crosstalk in undoped (Al,Ga)As waveguides, yielding high extinction ratio (> 40 dB) and low on-chip loss (< 6 dB) in Mach-Zehnder interferometers. RF drive voltage is reduced through use of 45 mm-active length devices with modulator sensitivity, v-pi, less than 3 V.

Processes in a modulation amplifier on a high temperature superconducting film with nonlinear parametric inductance and active resistance are theoretically examined. The expression is obtained in this paper for the average power on the inductance. A possibility of a parametric regeneration on the modulation frequency of the surface impedance of a HTSC film offers at the irradiation of the latter by an optical radiation modulated on the intensity.

Continuous beam steering can be achieved in plasma-grating antennas that employ an array of optical fibers or electrodes set at fixed positions. Effects of both continuous and discrete-level illumination are considered.

We present an original method for characterizing opto- microwave devices up to THz modulation frequencies. This method is based on the measurement of the temporal response of devices excited by a laser pulse. The transfer function of the device is obtained by a numerical FFT of the experimental temporal response. Using ultrashort laser pulses, we are able to obtain this response from DC up to a few THz. In order to demonstrate the feasibility of the method, we have characterized an integrated optical Mach- Zehnder interferometer. The optical path difference between the two arms of the interferometer is about 6.6 mm leading to interferences in the tens of GHz frequency domain. This opto-microwave device has been characterized from 10 GHz to 2.0 THz. The first rejected frequency is around 15 GHz and the free spectral range is 30 GHz. This device may be used as a sub-carrier frequency filter for all-optical signal processing in high-rate optical communication systems.

The far infrared behavior of doped semiconductors is dominating by the free carrier response and it is well described by the simple Drude model. As the free carriers can be photogenerated in high resistivity semiconductors, the electromagnetic response of these semiconductors can be changed, via laser excitation, from a dielectric behavior to a metallic one. Using terahertz time-domain spectroscopy, we have checked this Drude-like behavior with several silicon wafers of different doping, and with a high resistivity silicon sample photo-excited by a Ti:Sa laser beam. Then we benefited of this effect to photo-modify the transmission of a photonic band-gap crystal in the terahertz range. The photonic band-gap crystal is a woodpile structure, which exhibits a complete forbidden band around 265 GHz. Silicon- made defects are introduced in interstitial sites of the lattice of the photonic crystal in order to create a defect mode inside the photonic band-gap of the crystal. This defect mode then acts as a narrow bandpass filter at the frequency of 253 GHz. Under 400-mW laser excitation power, the resonance peak associated to the defect mode completely vanishes.

We present a new technique for remote upconversion by inserting a passive all-optical device in the microwave (MW) fiber-optic link. Using a semiconductor laser directly modulated by two MW signals and an unbalanced Mach-Zehnder interferometer (UMZ) to convert optical frequency modulation into intensity modulation, mixing is achieved after photodetection. Experiments with a UMZ integrated on glass- substrate have demonstrated the feasibility of this optical method for MW mixing with a low cost device of easy fabrication. Temperature control of the device allows optimized mixing performance and stable response. This method permits to overcome the effect of chromatic dispersion in standard singlemode fiber systems operating in the 1.55 micrometers wavelength window. Due to the presence of high-frequency fundamental components in the optical field, the received power is considerably degraded during transmission and direct detection in conventional systems, using either direct or external modulation in the MW band. With the proposed technique, lower frequency components of the field can be transmitted. The insertion of the simple UMZ generates high frequency only at detection side, therefore the available fiber length is extended. As it is shown by simulation results, this method can be used for the upconversion of MW subcarriers carrying digital signals.

Tunable radiation in the Fir-THz range is generated by Miniature Photoconducting Capacitor Array. The capacitor array made by deposition of aluminum structure on semiconductor crystal. The crystal serves as insulation material for the capacitors and as a source for electron- hole plasma during illumination by laser pulse. The illuminating 100 fs 0.8 (mu) laser pulse creates a propagating ionization front inside the crystal volume and sequentially discharges the capacitors. The sequential discharge of the capacitor generates propagating wave inside the plasma. The emitted radiation frequency depends on the energy in the laser pulse and the wavelength of the static wave. In the presented work the tunability was achieved by varying the laser pulse energy from 0.1 to 1 mJ. This results in the free space propagating waves with frequency between 100 GHz and 1.6 THz. The spectrum width is changed according to the static spectrum bandwidth. The concept can be scaled to develop narrowband, tunable, powerful THz source.

It is shown that the periodically poled lithium niobate slab waveguide with specific period of poling, (lambda) /n_g (where (lambda) is a wavelength of emitted THz-wave, n_g is a refractive index of group velocity of light) is capable to emit the difference frequency generation (DFG) in the direction normal to the surface of a waveguide. The general expression for power of DFG in doubly resonant cavity (resonance both at a pumping and at difference frequencies) has been obtained. According to estimation the power of DFG is 0.33 mW for pumping power 1 W. Thus the double resonant cavity-enhanced surface-emitted DFG enables to design compact solid-state source of THz waves with the output power sufficient for the practical applications.

Design and characterization of optical-THz phase-matched traveling-wave photomixers for difference-frequency generation of THz waves are presented. A dc-biased coplanar stripline fabricated on low-temperature-grown GaAs is illuminated by two non-collinear laser beams which generate moving interference fringes that are accompanied by THz waves. By tuning the offset angle between the two laser beams, the velocity of the interference fringe can be matched to the phase velocity of the THz wave in the coplanar stripline. The generated THz waves are radiated into free space by the antenna at the termination of the stripline. Enhancement of the output power was clearly observed when the phase-matching condition was satisfied. The output power spectrum has a 3-dB bandwidth of 2 THz and rolls off as approximately 9 dB/Oct which is determined by the frequency dependent attenuation in the stripline, while the bandwidth of conventional photomixer design has the limitation by the RC time constant due to the electrode capacitance. The device can handle the laser power of over 380 mW, which is 5 times higher than the maximum power handring capability of conventional small area devices. The results show that the traveling-wave photomixers have the potential to surpass small area designs, especially at higher frequencies over 1 THz, owing to their great thermal dissipation capability and capacitance-free wide bandwidth.

We reported the observation of intense ultrashort electric pulses generated by a novel nearly-filled-gap nonuniform illumination (NFGNI) excitation scheme. This NFGNI scheme differs from conventional nonuniform illumination in that the excitation beam is focused to a spotsize only slightly smaller than the spacing between the transmission lines, and is located symmetrically within the gap. With NFGNI of 100 fs laser pulses on sliding-contact photoconductive switches fabricated on low-temperature-grown-GaAs, electric pulse correlation with 187 fs FWHM, which corresponds to a 3 dB bandwidth of 1.1 THz, was observed. To our knowledge, these are the shortest electric pulses ever reported in the literature. Moreover, these electric pulses were found three times stronger than these generated with filled-gap illumination. Bias, wavelength, and pump power dependencies are discussed.

This paper presents a new photodetector operating at millimeter-wave range. The waveguide photodetector structure and a brief principle of the electrode design are described in this paper. By taking the advantages of narrow band operation in millimeter-wave subcarrier optic fiber link system, the proposed photodetector is designed to have cascaded resonant structures and a reflector to achieve high RF output. Resonators and reflector are realized by using one-fourth wavelength shorted stubs and an open stubs of coplanar waveguide, respectively. Velocity matching between optical waves and detected millimeter wave signals are performed by introducing an S-shaped meandering transmission line for the photocurrents.

This paper presents modeling and simulation results on Si- based quantum-well intersubband THz detectors and THz lasers (tasers) in the 3 to 10 THz range where the intersubband transition energy is 12 to 41 meV. The incoherent cryogenically cooled (4 K to 20 K) quantum well terahertz detector (QWTD) consists of p-type Si_0.9Ge_0.1 QWs with Si barriers on an Si substrate, or of p-Si_0.85Ge_0.15/Si on a relaxed Si_0.97Ge_0.03 buffer on Si. The QWTD senses THz radiation at normal incidence (the XY polarization on the HH1 to LH1 transition) or at edge- illumination (the Z polarization on the HH1 to HH2 transition). Resonant-cavity enhancement, coupling to Si THz waveguides, and integration with SiGe transistor preamplifiers appear feasible for QWTDs. The quantum staircase taser is a simplified far-infrared version of the quantum cascade laser in which each superlattice transfer region is replaced by a thin tunnel-barrier layer. We have adapted to group IV the III-V idea of Sun, Lu, and Khurgin; the `inverted mass taser'. On a Si_0.81Ge_0.19 substrate, we find that an inverted effective mass exists in LH1 at k_g equals 0.013 angstroms^-1 in 9-nm single- wells of Si_0.7Ge_0.3 with 5-nm Si barriers. Selective electrical injection of holes into LH1 at T equals 77 K is postulated. This offers local-in-k-space LH1-HH1 population inversion and tasing at 7.2 THz. Since the taser emission is XY-polarized, the active MQW staircase (a set of identical square QWs) is suitable for insertion into a vertical cavity surface-emitting taser. The VCSET would have resonator thickness of (lambda) /2n equals 6 micrometers , and Bragg mirrors constructed from SiO₂/Si multilayers.

Resulting from excellent direct-current, high-frequency, and reliability characteristics, HBT technology has attracted much attention in recent years. To help technology development groups determine if HBTs are the correct solution for 100 GHz applications, this work briefly describes the progress of HBT technology during the last decade and sets forth some of the reasons for employing this technology. Different HBT technologies are discussed and the current state-of-the-art is presented. Finally the existing limitations and future of HBT technology are discussed.

Optical interconnects possess great potential in applications for short-distance, multiple channel parallel connections at the chip-to-chip, board-to-board, back plane, and local area network levels of high performance computing environments. Low-loss and high-bandwidth advantages of optical fiber over those for coaxial cables become sizable when the transmission speed exceeds multiple Gb/s. OEIC (Opto-Electronic Integrated Circuits) receivers and transmitters are suitable for both free-space and fiber-optic short-wavelength optical links. Such chip sets will be able to support link distances from less than 1 mm for chip-to-chip optical interconnects to over 1 km for local area network (LAN) systems. As a high-speed optical receiver for these systems, monolithic OEICs are very attractive because of their potential for high-speed operation, compactness, and cost reduction. In this paper, we will review the theoretical speed limit of MSM (Metal-Semiconductor-Metal) photodetector. 3-dB bandwidth of 50x50 um2 MSM detector will be studied. The recent progress on the 100 GHz MESFET (Metal Semiconductor Field Effect Transistor), InP HFET (Heterojunction Field Effect Transistor) and their OEIC receivers are also presented.

The impedance and noise characteristics of a semiconductor punch-through structure are theoretically examined for an operation under a condition when quantum wells (QWs) are present in the transit-time region of the structure. It is shown that the magnitude of the negative dynamic resistance can be increased under the influence of the trapping and escape effects of injected carriers by quantum wells. It is expected that the structure proposed have significantly higher operation frequencies in comparison with usual barrier-injection transit-time diode. It is shown also that the noise measure decreased under an influence injected charge carriers captured by QWs with the increase of the ration of the emission time of electrons emitted out of QWs due to the thermal excitation to the capture time of free charge carriers. The frequency band where this phenomenon takes place is narrowed and displaced to a lower frequency range.

To investigate radiation effects on Al_0.36Ga_0.64As/GaAs asymmetric double quantum wells, we made photoluminescence (PL) measurements using intense, mid- infrared (IR) radiation from a pulsed CO₂ laser and a free electron laser (FEL). Also, we investigated the generation of a non-thermal distribution of electrons by combining a continuous wave He-Ne and a pulsed CO₂ laser. The temperature of the electrons and their relaxation process under mid-IR irradiation was determined from the normalized PL intensities and the power loss. The energy relaxation process of the electrons at low temperature was dominated by the plasmon-phonon coupling mode. The intensity of the photoluminescence solely from FEL irradiation was strongly dependent on the FEL mean power, and the temperature dependence was much weaker than that of photoluminescence with the He-Ne laser. The peak energy of the photoluminescence was nearly independent of external applied voltage.

A polymeric 1 X 8 arrayed waveguide grating wavelength multiplexer with 1.6 nm (200 GHz) channel spacing has been designed and realized for operating around a 1550 nm wavelength. Two kinds of fluorinated polymers, perfluorocyclobutane and fluorinated poly(ether ketone) polymers were used for the low loss waveguides. As a core layer, fluorinated poly(ether ketone) polymer, which has a low propagation loss, a good processibility and high thermal stability up to 460 degree(s)C was newly synthesized. The propagation loss of a buried rib waveguide is less than 0.5 dB/cm at the 1550 nm wavelength. The refractive index difference between the core and the clad layers is 0.0273 ((Delta) equals 1.8%). The bending radius of curved waveguides is 7.5 mm. The device size is 39 mm long and 13 mm wide. Fiber-to-fiber insertion losses of the multiplexer are between 7 dB and 8 dB, and a 3-dB bandwidth is 0.6 nm. The crosstalk of all 8 channels is less than -24 dB. And the polarization (TE-TM mode) related wavelength shift is approximately 3.4 nm.

The problem of the plane electromagnetic waves of E- and H- polarizations reflection from a metal plane covered by chiral layer of any thickness is strictly solved. The parameter of chirality and layer thickness influence on the electromagnetic waves reflection is investigated. The comparison of results obtained with using of two various notations of the material equations for chiral medium is made. The reflection character of the depolarized component is investigated.

ASOC^TM developed by Bookham Technology Ltd is based on Silicon on Insulator ridge waveguide technology. Bookham has demonstrated the low cost, high volume manufacturing capability of ASOC^TM by offering a range of transmitter, receiver and transceiver products to the access market. This range was recently complemented by a number of products in the sensor and DWDM markets. Additional products are currently being developed which illustrate the capability of this technology to provide low volume, high functionality, highly integrated components. In this paper, attention will be given to the range of functionality offered by ASOC and their integration potential.

We present microelectromechanical system (MEMS) RF component architectures that enable reconfigurable MEMS millimeterwave transceivers. In this paper, we focus on the component architectures for system integration. Different components, including reconfigurable Vee-antennas, planar impedance tuners, microswitches and variable capacitors, are demonstrated using the same three-polysilicon-layer processes. The reconfigurability of the Vee-antenna is demonstrated with beam-steering and beam-shaping functionality. The tuning ranges of planar backshort impedance tuners on different transmission lines are studied. We show that it is feasible to integrate the reconfigurable Vee-antenna with the planar impedance tuners. The mechanical architectures for microswitches, parallel- plate variable capacitors and circular variable capacitors are also investigated.

The binary lithography combined with precise control of RIE (Reactive Ion Etch) could fabricate an 8-level blazed grating with the reflective efficiency above 80% in recent years. The pixel size of the 8-level grating is often designed to be 32 microns or even larger due to the manipulation error of the mask-aligner. By the excellent augmentation of the mask-stepper, which is popular in IC industry, the smaller pixel size or the higher-order levels of the gratings could be achieved ideally.

The fabrication of diffractive optics and digital optics on fused silica has been investigated in this paper. The diffractive optics has been fabricated by using the gray- scale mask technique. The gray scale mask is fabricated on the HERB glass. RIE is used to realize the pattern transfer from the photoresist to the fused silica for the diffractive optics and the etching for the digital optics. The processing parameters are discussed in detail and the influence of exposure and development of the photoresist are also addressed in this paper. These diffractive optics and digital optics were designed to collimate optical radiation.

Integrated Packaging (IP) technology was developed to enable further application of ultrahigh-speed Uni-Traveling-Carrier photodiodes (UTC-PDs). In IP modules, devices are fabricated on the package substrate together with all interconnections using standard semiconductor processing techniques after wafer bonding of device epitaxial layers. Thus, problems associated with substrate discontinuity and wire/solder interconnections encountered in conventional hybrid packaging of > 100 GHz devices are eliminated. UTC-PDs, integrated with millimeter-wave coplanar waveguides (CPWs), were fabricated on package compatible sapphire with high yield. Device performance was not affected by wafer bonding. Furthermore, devices with record 3-dB bandwidth of 174 GHz were obtained. These devices produced output voltages of 1.45 V (29 mA) at higher input levels while maintaining 3-dB bandwidth above 150 GHz. CPWs fabricated on sapphire exhibited low dispersion. Thus, wafer bonding on sapphire is a promising technique for IP module fabrication.

Cost and performance are the primary driving forces of millimeter-wave technology for use in wireless communications systems. They have stimulated an ever-growing interest in seeking new circuit building blocks to cope with such demands. In this work, we present an innovative circuit building block concept, which is based on the co-layered geometry of planar circuits and NRD-guides (surface-mounted NRD-guides), featuring adjacent co-layered hybrid integration of two dissimilar structures without resort to intermediate aperture coupling. Our simulation and experiments of several co-layered schemes including microstrip planar circuit/NRD-guide and coplanar waveguide NRD-guide structures indicate that low transmission loss and good return loss can be achieved for the hybrid interconnects. Such results have validated our proposed concept and the usefulness of these new schemes. This scheme is rather unique and will be useful if monolithic circuits are required to interconnect with the NRD-guide. In addition, this proposal points to a unique possibility of designing new interconnects of adjacent interface-to- interface or layer-to-layer planar circuits that could be located over neighboring interfaces via dielectric waveguide. Some technical features of the proposed schemes are discussed with respect to transmission efficiency between these structures. Our work demonstrates interesting electrical and design aspects of these new circuit building blocks for millimeter-wave and quasi-optical applications.

Transformation of electromagnetic wave in a medium which permittivity and conductivity is being modulated by a finite packet of periodic rectangular pulses is considered. There are not restrictions on duration and amplitude of modulation pulses as well as on an initial wave amplitude and frequency. It is shown that the initial electromagnetic wave is split by a pulse packet into two waves, direct and inverse, with the same frequency as the initial wave has. The amplitudes of these waves depend on a number of pulses and its duration as well as on medium parameters very strong. There are such parameters that the secondary waves can be amplified but the amplification is absent when a medium conductivity changes only. It is revealed that wave amplitudes can change irregularly with the pulse number. Irregular behavior of amplitudes has intermittent character when a slight changing of the amplitude alternates with great amplitude peaks. This process is controlled by a generalized parameter an expression for which is derived. The criteria for monotonous or irregular behavior of the secondary wave amplitudes is derived.

This paper presents an optimized designs of wide band traveling wave photodetectors grown on InP for operation in the optical communications wavelength region of 1.55 micrometers . Traveling wave photodetector devices exhibit inherently wide band frequency responses that are not limited by the usual RC time constant seen with vertically illuminated detectors. The factors limiting the performance of the traveling wave photodetectors are carrier transit time, microwave-optical phase velocity-mismatch and microwave loss. A discussion of key design parameters related to the maximization of the bandwidth is presented, including the choice of semiconducting heterostructure layers, device architecture, phase velocity matching between the optical and microwave signals and the trade off between slow-wave effects and microwave losses. Device modeling and numerical simulations were performed using the Method of Lines, which is a rigorous full-wave numerical technique for modeling electromagnetic fields. These results were corroborated using a commercially available CAD package based on the Finite Difference Beam Propagation Method.

In this paper, we report our recent work in the development of the finite difference time-domain (FDTD) technique for the modeling of high-frequency, quasi-optical electronic and optoelectronic circuits. Our work includes FDTD analysis of active device-grid, anisotropic crystal structure, back-to- back varactor tripler and MESFET oscillator arrays, VCSEL, and miniature MMIC designs. Most of these works will be discussed in this paper.

The Applied Electrophysics Laboratory at the University of Virginia has developed a web based device simulation environment to accurately model and simulate terahertz and gigahertz electronic devices. This environment allows us to provide the device simulation codes developed over the years to the larger research community. In this paper, we describe the simulation environment and the particular device simulations available.

This paper presents a numerical method for computing the propagation constant of electromagnetic modes supported for multilayer planar optical waveguides. We report a method for solving the dispersion equation for a guided wave structure of interest in the complex plane via Chauchy's integration. The method is applicable to lossless, lossy and ARROW waveguide structures and can handle both leaky and guided modes. Contrary to the methods currently published in the literature, which are based on a numerical derivative of the dispersion equation, we propose an analytical derivative for the latter. This has a double impact: improved accuracy and reduced CPU time. The specific integration contour for leaky modes will be discussed. The results are in excellent agreement with several results published in the literature.

Narrow pulse signal generation is an important part in high- speed systems design. Utilizing the voltage and current phase difference of inductor and diode, a very short pulse can be generated from a sinusoidal voltage source. Such systems usually include highly nonlinear components and complicated 3D structures. Dispersion, field effects and the interaction between electromagnetic field and components cannot be ignored in the simulation of these electrically small and complicated systems. Therefore, time-domain and full-wave analysis is necessary. This paper presents a full wave FDTD analysis using an improved field-circuit model for hybrid system simulation, which includes both lumped elements and field effect. 2 and 3D hybrid (distributed and lumped) circuits are constructed and analyzed in this paper to demonstrate the versatility of the time-domain simulator.

We describe a unique and powerful global time-domain simulation technique for terahertz diodes such as GaAs Schottky diode mixers, GaAs Schottky diode frequency multipliers, and InP Transferred Electron Oscillators (TEOs). 1D, finite difference, drift-diffusion nonlinear device simulation codes have been linked with a convolution- based circuit analysis. These simulators allow designers to observe both the transient and steady state time domain behavior of the nonlinear circuits. Since physical device simulators have been used, the spatial and temporal behavior of the electrons and electric field within the device under large signal drive can be observed. This gives great insight into the internal device physics at high frequencies. The mixer code allows for the direct and fully self-consistent calculation of the conversion loss and noise temperature; the TEO code allows for fully autonomous calculation of oscillator start-up and frequency selection. Simulation results for 2.5 THz GaAs Schottky mixers and 140 GHz InP TEOs are given.

A rigorous 2D physical simulator is developed to determine the noise behavior of subhalf-micrometer gate-length FETs taking into account the non-stationary transport properties and the quantization effects. The microscopic nature of the simulator allows better understanding of the physical operation of these devices. As an example, the model is applied to compare and to accurately interpret the noise behavior of similar MESFET and Hetero-FET structures, to determine the physical phenomena which dominate their behavior, and to suggest different possibilities to improve the noise performance in a wide frequency range of operation.

High-speed optical sources and receivers are required for efficience wavelength division multiplexed lightwave networks. The performance of high-speed GaAs- and InP-based lasers and monolithically integrated photoreceivers, suitable for this application, are described.

A well-designed phase correcting Fresnel zone plate antenna can provide performance superior to a lens or, in some cases, a paraboloid antenna, particularly at millimeter wavelengths. This paper discusses design considerations and includes approaches to give improved characteristics, such as greater efficiency or higher gain. The approaches include the use of quarter-wave or better correction, thickness designs that permit the central zone and other zones to be air dielectric (for lower losses), and the use of low dielectric constant materials to reduce surface reflections and multiple reflections. At higher millimeter-wave or sub- millimeter wavelengths low loss materials are important. More sophisticated zoning is described, as well as the use of a compromise thickness to compensate for the fact that refraction of waves at the surfaces causes the path lengths through the zone plate to be different at different angles of incidence. Multiple-band zone plates are discussed.

We generated low phase noise mm-wave electrical signals of sub-kHz linewidth from a single 1550 nm diode laser with a linewidth in excess of 1 MHz. Our approach uses a phase modulator to generate a multiline spectrum from a single mode laser, which is filtered to pass only a pair of spectral lines. Interference between the lines on the active area of a photodetector produces an electrical signal that can be set between 4 and 60 GHz. This technique has the advantage that the electrical linewidth can be much narrower than the laser linewidth due to common mode rejection of frequency noise between the interfering sidebands, and that relatively low frequency electronics can be used to generate the sideband spectrum.

A new high frequency image-rejection mixer was successfully developed and tested in a 26 - 36 GHz band receiver. This paper describes the receiver system, the operation of the image rejection mixer, and presents test results for the mixer at ambient and cryogenic temperatures.