Thermal Boundary Resistance Between GaN and Substrate in AlGaN/GaN Electronic Devices

The influence of a thermal boundary resistance (TBR) on temperature distribution in ungated AlGaN/GaN field-effect devices was investigated using 3-D micro-Raman thermography. The temperature distribution in operating AlGaN/GaN devices on SiC, sapphire, and Si substrates was used to determine values for the TBR by comparing experimental results to finite-difference thermal simulations. While the measured TBR of about 3.3 x 10^-8 W^-1 ldr m² ldr K for devices on SiC and Si substrates has a sizeable effect on the self-heating in devices, the TBR of up to 1.2 x 10^-8 W^-1 ldr m² ldr K plays an insignificant role in devices on sapphire substrates due to the low thermal conductivity of the substrate. The determined effective TBR was found to increase with temperature at the GaN/SiC interface from 3.3 x 10^-8 W^-1 ldr m² ldr K at 150degC to 6.5 x 3.3 x 10^-8 W^-1 ldr m² ldr K at 275degC, respectively. The contribution of a low-thermal-conductivity GaN layer at the GaN/substrate interface toward the effective TBR in devices and its temperature dependence are also discussed.     

Wavelength-Tunable Microring Resonator on Lithium Niobate

A wavelength-tunable microring resonator with integrated microheater on lithium niobate is presented. Ridge structure on lithium niobate is formed by a wet-etching technique for enhancing the lateral index contrast of the waveguide. The resonant wavelength of the microring resonator is tuned through thermooptic effect by injecting current into the integrated microheater. The tuning characteristics of through port and drop port are measured and the tuning rates in the microring resonator with a radius of 100 mum for transverse-magnetic and transverse-electric polarizations are 2.54 x 10^-2 nm/mA and 3.40 x 10^-3 nm/mA.     

The SIS tunnel emitter: A theory for emitters with thin interface layers

Silicon n-p-n transistors are made with emitters consisting of a polycrystalline and monocrystalline region with a thin (20-60-Å) "insulating" interfacial layer in between (SIS structure). These transistors show a remarkable increase in emitter efficiency with emitter Gummel numbers up to 7 × 10¹⁴s.cm^-4, and a low positive or even negative temperature coefficient of the current gain. A model is proposed to explain the mechanism in terms of tunneling through the interfacial layer. The electrical characteristics are measured in the temperature range 290-415 K. From the measurements it is deduced that the tunnel probability for holes (Ph) is 10^-2to 10^-3and that for electrons (Pe) is >10^-5. There also exists a band bending of 30-90 mV at the interfaces with the interfacial layer.     

Limitations of atomic beam frequency standards

Atomic beam frequency standards may be placed into two categories: field standards and laboratory standards. While this distinction is somewhat artificial, because the two types of standards are interdependent, each category does have different requirements of accuracy, size, and cost. Despite this separation, generally the developments which produce the best laboratory standards eventually give rise to improved field standards. Existing field standards are limited in long term fractional frequency stability to σ_y (τ) 3 x 10^-13, for τ 6 months. A laboratory standard such as NBS-6, the U.S. primary cesium standard, is limited in inaccuracy to Δ_y 8 x 10^-14. Proposed new cesium field standards are expected to yield long term stabilities of σ_y(τ) 1 x 10^-14 (τ = 6 months). Stored ion standards, prime candidates for new laboratory frequency standards, are expected to have better than Δy = 1 x 10^-15 inaccuracy. As other approaches to atomic beam frequency standards are considered, they should attempt to compete favorably with these emerging technologies.     

Efficient computation of erfc(x) for large arguments

A new, infinite series representation for the error function is developed. It is especially suitable for computing erfc(x) for large x. For instance, for any x⩾4, the error function can be evaluated with a relative error less than 10^-10 by using only eight terms. Similarly, the error function can be evaluated with a relative error less than 8×10^-7 for any x⩾2 using just six terms. An analytical bound is derived to show that the total error due to series truncation and undersampling rapidly decreases as x increases. Comparisons with two other series are provided     

Test Method for Measuring Bit Error Rate of Pulsed Transceivers in Presence of Narrowband Interferers

The popularity of pulse-based transceivers can be attributed to the high information rates that can be achieved in such systems compared to traditional narrowband systems. However, such systems are usually low-power and transmission efficiency is severely affected by the interference signals from other moderate to high-power narrowband transmitters. Hence, during manufacturing, pulse-based systems must be characterized and tested for bit error rate (BER) performance in the presence of narrowband interferers. Usually, at large interference levels, the BER is moderate (10^-4) to high (10^-2). However, at low interference levels when very few bits are in error, the BER is low (10^-6 - 10^-10 ) and testing for the BER becomes time consuming. We propose a new measurement technique employing sinusoidal pulses such that the BER value obtained is significantly large (10^-3 - 10^-4 ) even at low interference levels. BER values obtained using sinusoidal pulses are highly correlated to the actual BER values. Hence, the actual BER can be accurately estimated in a much smaller time without actually performing the standard test. This method was implemented in hardware using an Altera field-programmable gate-array development board. From the measurements, BER estimation error was less than 3%. In addition, significant reduction (up to 100 x) in test time was obtained using the proposed method.     

Counting of deep-level traps using a charge-coupled device

Quantization in dark current generation has been observed for the first time through the use of a virtual-phase charge-coupled device. Two sites for bulk silicon dark current have been identified with capture cross sections of 1.8 × 10^-15cm²and 5.4 × 10^-16cm², and concentrations of 1.3 × 10⁹cm^-3and 1.5 × 10⁸cm^-3, respectively.     

Measurement of mode couplings and extinction ratios inpolarization-maintaining fibers

Polarization mode couplings in the axial direction are evaluated for polarization-maintaining fibers using optical heterodyne detection. To verify the validity of this approach for fibers with various coupling constants, the method is applied to three fibers with modal birefringence values of 3.0×10^-4, 1.1×10^-4 , and 1.5×10^-4, respectively. The coupling constants in the 1.7×10^-4 m^-1 to 6.4×10^-7 m^-1 range are evaluated with a length resolution of 1 m. The extinction ratios are obtained from the coupling constants averaged over the fiber lengths. These values are in good agreement with the values measured directly from power ratios between the orthogonally polarized modes     

BER fluctuation suppression in optical in-line amplifier systemsusing polarisation scrambling technique

The effectiveness of the polarisation scrambling technique for suppressing BER fluctuation is confirmed experimentally with a 2900 km optical in-line amplifier transmission system. The technique suppresses the BER fluctuation by three orders of magnitude, from 10^-5-10^-9 to 10^-810^-9     

Lasing mode behavior of a single-mode MQW laser injected with TMpolarized light

The lasing mode behavior of a multiple quantum well (MQW) distributed feedback (DFB) laser was measured when intensity-modulated orthogonally polarized transverse magnetic (TM) mode light was injected. The 3-dB bandwidth of the frequency response shows a trend different from that observed with conventional bias current modulation: at high bias currents, it decreases with increasing bias current. The maximum bandwidth of 3 dB was observed when the normalized bias current was 4, and it reached 16 GHz at this bias current. The gain saturation coefficients for the transverse electric (TE) and TM modes estimated from these results were ∈_p^E; 2.5×10^-17 cm³ and ∈_q^E 5.7×10^-18 cm³ for the TE mode, and ∈_p^M: 6.0×10^-17 cm³ and ∈_q^M: 2.0×10^-14 cm³ for the TM mode     

Diamond cold cathode

Diamond cold cathodes have been formed by fabricating mesa-etched diodes using carbon ion implantation into p-type diamond substrates. When these diodes are forward biased, current is emitted into vacuum. The cathode efficiency (emitted current divided by diode current) varies from 2×10^-4 to 1×10^-10 and increases with the addition of 10^-2-torr partial pressure of O₂ into the vacuum system. Current densities of 0.1 to 1 A-cm^-2 are estimated for a diode current of 10 mA. This compares favorably with Si cold cathodes (not coated with Cs), which have efficiencies of ~2×10^-5 and current densities of ~2×10^-2 A-cm^-2. It is believed that higher current densities and efficiencies can be obtained with more efficient cathode designs and an ultrahigh-vacuum environment     

Design and Application of Compact and Highly Tolerant Polarization-Independent Waveguides

In this paper, the design, fabrication, and application of a highly tolerant polarization-independent optical-waveguide structure suited for operation in the third communication window is presented. The waveguide structure has been optimized toward minimized sensitivity to technological tolerances and low fabrication complexity. The tolerance analysis has been based on the typical processing tolerances of the widely applied silicon-oxynitride technology, being plusmn3times10 ^-4 in refractive index, plusmn1% in thickness, and plusmn0.1 mum in channel width. The optimized waveguide design fulfills the criterion of a channel birefringence within 5times10^-5, including processing tolerance. It also enables a fiber-to-chip coupling loss below 1 dB/facet and is suited for the realization of low-loss bends with a radius down to 600 mum. Based on this waveguide design, a passband-flattened optical wavelength filter with 50-GHz free spectral range has been realized and tested. The measured TE-TM shift of 0.03 nm confirms the polarization dependence of the optical waveguides being as low as 3times10^-5     

A polysilicon—Silicon n-p junction

Phosphorus-doped polycrystalline silicon is grown in an epitaxial reactor by the reduction of a hydrogen-diluted silane-phosphine mixture passing over a substrate heated to 800°C. The influence of the phosphine-silane ratio on growth rate, electrical resistivity, active donor concentration, and Hall mobility is examined. It is found that phosphine inhibits growth rate at 800°C to a lesser degree than it does at lower growth temperatures. Growth rate progressively drops to 0.6 of the undoped value as the phosphine-silane ratio is increased to 10^-1. Resistivity drops from 1 to 10^-3Ω. cm as active phosphorus concentration varies between 10¹⁸and 4 × 10²⁰cm^-3, while Hall mobility rises from 4 to 30 cm²/ V.s. Diodes are formed between the grown polysilicon layers and the single-crystal p-type silicon substrates. They are found to have recombination currents critically dependent on the phosphine/ silane ratio during growth of the polysilicon. As this ratio increases above 10^-5, recombination decreases, while mobility in the polysilicon increases. These results support the "dopant segregation" theory of conduction in polysilicon. For ratios of 10^-3to 10^-2the diodes obtained showed a recombination factor approaching those of diffused diodes and are useful devices, for example, as the emitter-base junction of a shallow-base high-frequency, bipolar transistor.     

Low-resistance bandgap-engineeredW/Si1-xGex/Si contacts

Bandgap-engineered W/Si_1-xGe_x/Si junctions (p⁺ and n⁺) with ultra-low contact resistivity and low leakage have been fabricated and characterized. The junctions are formed via outdiffusion from a selectively deposited Si_0.7Ge _0.3 layer which is implanted and annealed using RTA. The Si _1-xGe_x layer can then be selectively thinned using NH₄OH/H₂O₂/H₂O at 75°C with little change in characteristics or left as-deposited. Leakage currents were better than 1.6×10^-9 A/cm² (areal), 7.45×10^-12 A/cm (peripheral) for p⁺/n and 3.5×10^-10 A/cm² (peripheral) for n⁺/p. W contacts were formed using selective LPCVD on Si_1-xGe_x. A specific contact resistivity of better than 3.2×10^-8 Ω cm² for p ⁺/n and 2.2×10^-8 Ω cm² for n⁺/p is demonstrated-an order of magnitude n⁺ better than current TiSi₂ technology. W/Si_1-xGe _x/Si junctions show great potential for ULSI applications     

Emission cross sections and spectroscopy of Ho3+ laserchannels in KGd(WO4)2 single crystal

The spectroscopic properties of Ho³⁺ laser channels in KGd(WO₄)₂ crystals have been investigated using optical absorption, photoluminescence, and lifetime measurements. The radiative lifetimes of Ho³⁺ have been calculated through a Judd-Ofelt (JO) formalism using 300-K optical absorption results. The JO parameters obtained were Ω₂=15.35×10^-20 cm², Ω ₄=3.79×10^-20 cm², Ω₆ =1.69×10^-20 cm². The 7-300-K lifetimes obtained in diluted (8·10¹⁸ cm^-3) KGW:0.1% Ho samples are: τ(⁵F₃)≈0.9 μs, τ( ⁵S₂)=19-3.6 μs, and τ(⁵F₅ )≈1.1 μs. For Ho concentrations below 1.5×10²⁰ cm^-3, multiphonon emission is the main source of non radiative losses, and the temperature independent multiphonon probability in KGW is found to follow the energy gap law τ_ph ^-1(0)=βexp(-αΔE), where β=1.4×10^-7 s^-1, and α=1.4×10³ cm. Above this holmium concentration, energy transfer between Ho impurities also contributes to the losses. The spectral distributions of the Ho³⁺ emission cross section σ_EM for several laser channels are calculated in σ- and π-polarized configurations. The peak a σ_EM values achieved for transitions to the ⁵I₈ level are ≈2×10^-20 cm² in the σ-polarized configuration, and three main lasing peaks at 2.02, 2.05, and 2.07 μm are envisaged inside the ⁵I₇→⁵I₈ channel     

Experimental studies of proton-implanted GaAs-AlGaAsmultiple-quantum-well modulators for low-photocurrent applications

We describe the first attempts to control photocurrent, and thus power dissipation, in surface-normal multiple-quantum-well (MQW) modulators. We have made detailed experimental studies of proton-implanted p-i-n GaAs-Al_xGa_1-xAs MQW modulators having barrier layers of x=0.3, 0.45, and 1.0. Structures were implanted to levels of 1×10¹² cm^-2, 1×10¹³ cm^-2, and 1×10¹⁴ cm ^-2. Photocurrent progressively decreased with increasing implant-dose and barrier mole fraction (x). Exciton linewidths showed a strong voltage and implant dose dependence, demonstrating a tradeoff between photocurrent and modulation performance. We obtained our best results with x=1.0 barriers. For example, 1×10¹³ cm^-2-implanted asymmetric Fabry-Perot modulators were realized in which the optical performance was similar to that of unimplanted devices. The photocurrent responsivity was, however, only 0.007 A/W at 12.5 V bias. We report measurements of carrier lifetime in these materials that show the reduction in photocurrent arises from a reduction in lifetime due to implant-induced damage. In addition, the reduced lifetime decreases the optically-excited quantum-well carrier population, leading to an increase in cw saturation intensity. Specifically, 1×10¹³ cm^-2-implanted devices with x=1.0 have a saturation intensity of roughly 45 kW/cm², while unimplanted devices have 3.5 kW/cm². Asymmetric self electro-optic effect devices (A-SEED's) are demonstrated, and power dissipation issues associated with the use of low-photocurrent modulators in integrated systems are discussed     

Some reactive effects in forward biased junctions

The small-signal equivalent parallel capacity of forward-biased semiconductor junctions is strongly dependent on the current. At very low currents (less than 10 µa for a junction area of 1 mm²) the capacity appears to be chiefly due to space charge effects. For currents up to approximately 100 µa, the capacity complies with Shockley's predicted low-level theory. For larger currents, however, there is a definite deviation from the low-level diffusion predominance and capacity reaches a maximum after which it decreases through zero and then goes to large inductive values. The latter phenomena is explained, qualitatively, by considering an inductance in series with the diffusion capacity. The capacity increases linearly with current but the inductance (due to conductivity modulation) increases faster. The result is that a change from an equivalent RC circuit to an equivalent RL circuit is made at high enough currents (5 ma is a typical value for the 1 mm²junction area). Measurements were made on abrupt silicon junction diodes with junction areas of about 7 × 10^-4, 10^-2, 10^-1cm²and on the emitter junction (about 5 × 10^-5cm²) of a diffused base silicon transistor.     

Cr4+:YAG as passive Q-switch and Brewster plate in apulsed Nd:YAG laser

We demonstrate the performance of a Nd:YAG laser, passively Q-switched with a Cr⁴⁺:YAG plate, which plays the double role of a passive Q-switch and a Brewster plate. The Brewster plate configuration contributes an intracavity loss of approximately 3.2-10 ^-3 cm^-1 along the cavity length. Losses contributed by the active Cr⁴⁺ ions in the plate relate to their excited state absorption. A freshly measured transmission saturation curve of Cr⁴⁺:YAG suggests a ground state absorption cross section σ_gs=(8.7±0.8)-10^-19 cm², and an excited state absorption cross section σ_es=(2.2±0.2)-10^-19 cm² of the Cr⁴⁺ ions at λ=1064 nm     

A comparison of optoelectronic properties of lattice-matched andstrained quantum-well and quantum-wire structures

The k·p formalism is used to study the absorption spectra, material and differential gain in quantum wires as a function of orientation, built-in strain, and wire dimensions. The results for material and differential gain are compared with those for an optimized quantum-well structure. We find that for quantum wires at 300 K, the gain becomes positive at a carrier density of 1.27·10¹⁸ cm^-3, while in quantum wells this density is calculated to be 1.82·10¹⁸ cm^-3. Incorporating tensile strain in the wires reduces the transparency carrier concentration to 0.96·10¹⁸ cm^-3 while compressive strain allows one to obtain positive gain for densities greater than 1.08·10¹⁸ cm^-3. Orienting the wire along the [111] direction reduces the transparency carrier density to 0.60·10¹⁸ cm^-3. The differential gain in quantum-well structures for injections near the threshold is on the order of 10^-14 cm^-4, while for 50 Å·100-Å quantum wires the differential gain near the threshold is found to be on the order of 10^-13 cm^-4 . The differential gain in wires whose wire axis is parallel to the [111] direction has also been found to be on the order of 10^-13 cm^-4 for carrier injections close to the threshold