A drain avalanche hot carrier lifetime model for n- and p-channel MOSFETs

A simple and physical drain avalanche hot carrier lifetime model has been proposed. The model is based on a mechanism of interface trap generation caused by recombination of hot electrons and hot holes. The lifetime is modeled as /spl tau/(I/sub d//W)/sup 2//spl prop/(I/sub sub//I/sub d/)/sup -m/. The formula is different from the conventional /spl tau/I/sub d//W-I/sub sub//I/sub d/ model in that the exponent of I/sub d//W is 2, which results from the assumed mechanism of the two-carrier recombination. It is shown that the mechanism gives a physical basis of the empirical /spl tau/-I/sub sub//W model for NMOSFETs. The proposed model has been validated experimentally both for NMOSFETs and for PMOSFETs. Model parameters extracted from experimental data show that carrier critical energies for creating damage are lower than the interface potential barriers. It is supposed that oxide band edge tailing enables low-energy carriers to create the damage. The channel hot electron condition becomes the worst case in short channel NMOSFETs, because gate voltage dependence of the maximum channel electric field decreases.     

Antimonotonicity and maximal complexity in optically injected two-section lasers

In this paper, we investigate some special dynamic characteristics of two-section semiconductor lasers subject to optical injection. Forward and reverse period-doubling cascades are presented in the parameter space of frequency detuning /spl omega/ and injection strength K, a phenomenon known as antimonotonicity. Furthermore, we investigate the Kaplan-Yorke dimension (D/sub KY/) and the largest Lyapunov exponent (LLE) of chaotic attractors in this laser setup. It is found that the latter is a dynamical system that can possess a dimension D/sub KY/ anywhere between 2 and close to 4 as the control and material parameters are varied, making it the most complex laser setup using optical injection. Finally, the criteria for achieving a high LLE and a high D/sub KY/ are examined (by varying all possible material and control parameters) and presented with potential applications to chaotic communications.     

The energy-driven paradigm of NMOSFET hot-carrier effects

As negative-MOSFET (NMOSFET) size and voltage are scaled down, the electron-energy distribution becomes increasingly dependent only on the applied bias, because of quasi-ballistic transport over the high-field region. A new paradigm, or underlying concept, of NMOSFET hot-carrier behavior is proposed here, in which the fundamental "driving force" is available energy, rather than peak lateral electric field, as it is in the lucky electron model (LEM). The new prediction of the energy-driven paradigm is that the bias dependence of the impact-ionization (II) rate and hot-carrier lifetime is, to the first order, given by the energy dependences of the II scattering rate S/sub II/(E) and an effective interface state generation (ISG) cross section S/sub IT/(E), whereas, under the LEM, these bias dependences are determined by the number of electrons with energy above the II and ISG "threshold energies." This approach allows an experimental determination of S/sub IT/.     

Stress-induced electromigration backflow effect in copper interconnects

The electromigration threshold in copper interconnect is reported in this paper. The critical product (jL)/sub c/ is first determined for copper oxide interconnects with temperature ranging from 250/spl deg/C to 350/spl deg/C from package-level experiments. It is shown that the product does not significantly change in this temperature range. Then, (jL)/sub c/ was extracted for copper low-k dielectric (k=2.8) interconnects at 350/spl deg/C. A larger value than that for oxide dielectric was found. Finally, a correlation between the n values from Black's model and with jL conditions was established for both dielectrics.     

Dielectric relaxation in piezo-, pyro- and ferroelectric polyamide 11

Ferroelectric polyamide 11 films were prepared by melt-quenching, cold-drawing and electrical poling. Their ferroelectricity was studied by means of dielectric-hysteresis measurements. A remnant polarisation of up to 35 mC/m/sup 2/ and a coercive field of 75 MV/m were obtained. The piezoelectric d/sub 33/ coefficient and the pyroelectric coefficient of the films are reduced by annealing just below the melting region, but remain at about 3 pC/N and 8 /spl mu/C/(m/sup 2/K), respectively, during further heat treatment. Differential scanning calorimetry (DSC), dielectric relaxation spectroscopy (DRS) and thermally stimulated depolarisation (TSD) were applied for investigating the conformational changes induced by melt-quenching, cold-drawing and annealing. The results indicate that the cold-drawn film mainly consists of a rigid amorphous phase which exhibits considerably lower conductivity, no glass transition and consequently no dielectric /spl alpha/ relaxation. Instead, an /spl alpha//sub r/ relaxation is found, which is related to chain motions in regions of the rigid amorphous phase where the amide-group dipoles are not perfectly ordered. Annealing removes imperfectly ordered structures, but does not affect the ferroelectric polarisation. Therefore, it may be concluded that essentially the /spl alpha//sub r/ relaxation causes the thermally nonstable part of the piezo- and pyroelectricity in polyamide 11.     

Orthogonal control of femtosecond mode-locked laser having zero carrier-offset frequency with three-axis PZT

The authors demonstrate an optical frequency synthesizer based on a femtosecond (fs) mode-locked Ti:sapphire (Ti:s) laser by simultaneously stabilizing the carrier-offset frequency f/sub ceo/ and repetition rate f/sub rep/, referenced to the Cs atomic frequency standard. By using two wide-band digital phase-detectors they realize a phase-coherent link between f/sub rep/ and f/sub ceo/ with the relation f/sub ceo/=f/sub AOM/-5/6f/sub rep//spl equiv/0, where f/sub AOM/=5/6f/sub rep/ is the phase-locked driving frequency of an acousto-optic modulator in a self-referencing interferometer and f/sub rep/=100 MHz. For the simultaneous control of f/sub rep/ and f/sub ceo/, an orthogonalization scheme using a three-axis piezoelectric transducer is introduced by which the end mirror of fs laser is controlled. As a result, the authors could stabilize all components of the fs laser comb at once with an equal frequency separation f/sub rep/=100 MHz with f/sub ceo/=0. In their optical frequency synthesizer, the frequency of the nth component (f/sub n/) is given exactly by the simple relation f/sub n/=nf/sub rep/, enabling them to use the fs laser comb as a frequency ruler in the optical frequency metrology.     

Multiple-wavelength emission from In/sub x/Ga/sub 1-x/As-GaAs quantum wells grown on a nanoscale faceted GaAs substrate by molecular beam epitaxy

Multiple-wavelength photoluminescence (PL) spanning a 160-nm range from 980 to 1140 nm (77 K) has been obtained from In/sub x/Ga/sub 1-x/As-GaAs quantum wells (QWs) with varying In composition x on a nanoscale faceted (nanofaceted) GaAs substrate grown by molecular beam epitaxy. Five nanofaceted regions which consist of periodic [100]-(n11) (n = 3 or 1) facets along [011~] with different periods were prepared on a single substrate by interferometric lithography and selective growth of GaAs. The pattern period p was varied from infinity (large-area unpatterned) to 210 nm while the lateral width of the (n11) facet region was kept constant at /spl sim/180 to 200 nm within each period. A 5-nm-thick In/sub 0.23/Ga/sub 0.77/As layer was deposited on this multiple-period nanofaceted single GaAs surface in a single-run growth. Orientation-dependent migration and incorporation (ODMI) of In atoms [mass transport of incident In atoms from the (n11) to adjacent [100] facets] results in a variation of x of the In/sub x/Ga/sub 1-x/As layer section on the [100] facet as the width of the [100] facet was changed from /spl sim/20 to /spl sim/200 nm. ODMI induces a higher x on the [100] facet for smaller p. The PL exhibits a polarization dependence which is more pronounced for decreasing p [i.e., the width of [100] facet]. Consistent variations of the PL peak energy and linear polarization along the pattern direction confirm that ODMI results in a variation of the In composition and imply that the In/sub x/Ga/sub 1-x/As layer on a [100] facet has characteristics of a quantum wire as its width is decreased to /spl sim/20 nm for p = 210 nm. A possible application of nanopatterned growth to wavelength-division-multiplexing transmitters is discussed.     

The stoichiometry of thermal oxidation

At a stoichiometrical relation of silicon and oxygen during the growth of a rapid thermal silicon-oxide layer (RTO) the electrical parameters of the corresponding metal-insulator-semiconductor capacitor show significant behavior. Combining the Henry-Dalton law and the Boyle-Mariotte law with the Deal-Grove model allows an estimation of the fundamental (T, t, C) process, parameters, temperature, time and concentration of the reactive gas for a stoichiometrical processing of ultra thin SiO/sub 2/ films. Electrical measurements show the (T, t, C)-dependency of I-V, time dependent dielectric breakdown (TDDB) and C-V measurements especially of direct tunneling currents j/sub DT/, relative difference of tunneling currents /spl Delta//sub RDT/, charge-to-breakdown Q/sub BD/ and interface state density D/sub it/.     

Noise figure of parametric wavelength conversion with quasi-phase-matched LiNbO/sub 3/ waveguide

An analytical model of the noise accumulation in a chain of parametric wavelength converters is proposed. Signal-to-noise electrical power ratio is analytically given as a function of node number k in a chained transparent node system that consists of optical amplifiers, parametric wavelength converters, and several loss elements including optical transmission fiber with parameters of pump light excess noise /spl beta//sub p/, and average photon numbers per unit time of pump light and input signal , and , respectively, and spontaneous emission factor of optical amplifier n/sub sp/. The signal-to-noise degrades inversely proportional to node number k with the coefficient defined by NF/sup (1)/=2n/sub sp/+/spl beta//sub p// when k is lower than /Bo, where B/sub o/ represents optical bandwidth. The noise figure dependence on pump light quality /spl beta//sub p// and average photon number of input light in a single stage configuration are experimentally evaluated using Er-doped fiber amplifiers and quasi-phase-matched lithium niobate waveguide parametric wavelength conversion.     

Recovery of shifted MOS parameters induced by focused ion beam exposure

The effects of focused ion beam (FIB) exposure on MOS transistors within a circuit were examined. It was found that FIB exposure does not cause parameter shifts as long as the gate is connected to the drain of other MOS transistors. However, the threshold voltage (V/sub t/) does shift during isolating the gate using a FIB. Further FIB exposure on MOS transistors with a floating gate is shown to cause larger shifts. Thermal annealing was studied to recover shifted V/sub t/. We demonstrated that a 400/spl deg/C-450/spl deg/C anneal could recover shifted V/sub t/ almost completely. Ninety percent recovery can be reached by annealing at 400/spl deg/C-450/spl deg/C for 1-2 hours, and V/sub t/ shifts can be reduced to about 10 mV.     

Mode characteristics for equilateral triangle optical resonators

Mode characteristics of equilateral triangle resonators (ETRs) are analyzed based on the symmetry operation of the point group C/sub 3v/. The results show that doubly degenerate eigenstates can be reduced to the A/sub 1/ and A/sub 2/ representations of C/sub 3v/, if the longitudinal mode number is a multiple of 6; otherwise, they form the E irreducible representation of C/sub 3v/. And the one-period length for the mode light ray is half of the perimeter of the ETR. Mode Q-factors are calculated by the finite-difference time-domain (FDTD) technique and compared with those calculated from far-field emission based on the analytical near-field pattern for TE and TM modes. The results show that the far-field emission based on the analytical field distribution can be used to estimate the mode Q-factor, especially for TM modes. FDTD numerical results also show that Q-factor of TE modes reaches maximum value as the longitudinal mode number is a multiple of 7. In addition, photoluminescence spectra and measured Q-factors are presented for fabricated ETR with side lengths of 20 and 30 /spl mu/m, and the mode wavelength intervals are compared with the analytical results.     

Reliability of high-speed SiGe heterojunction bipolar transistors under very high forward current density

As device scaling for higher performance bipolar transistors continues, the operation current density increases as well. To investigate the reliability impact of the increased operation current density on Si-based bipolar transistors, an accelerated-current wafer-level stress was conducted on 120-GHz SiGe heterojunction bipolar transistors (HBTs), with stress current density up to as high as J/sub C/=34 mA//spl mu/m/sup 2/. With a novel projection technique based on accelerated-current stress, a current gain shift of less than /spl sim/15% after 10/sup 6/ h of operation is predicted at T=140/spl deg/C. Degradation mechanisms for the observed dc parameter shifts are discussed for various V/sub BE/ regions, and the separation of the current stress effect from the self-heating effect is made based on thermal resistance of the devices. Module-level stress results are shown to be consistent with wafer-level stress results. The results obtained in this work indicate that the high-speed SiGe HBTs employed for the stress are highly reliable for long-term operation at high operation current density.     

Absolute cerebral blood flow measured by dynamic susceptibility contrast MRI: a direct comparison with Xe-133 SPECT

Absolute regional cerebral blood flow (CBF) was measured in ten healthy volunteers, using both dynamic susceptibility-contrast (DSC) magnetic resonance imaging (MRI) and Xe-133 SPECT within 4 h. After i.v. injection of Gd-DTPA-BMA (0.3 mmol/kg b.w.), the bolus was monitored with a Simultaneous Dual FLASH pulse sequence (1.5 s/image), providing one slice through brain tissue and a second slice through the carotid artery. Concentration C(t) is proportional to -(1/TE) ln[S(t)/S(0)] was related to CBF as C(t) = CBF [AIF(t) x R(t)], where AIF is the arterial input function and R(t) is the residue function. A singular-value-decomposition-based deconvolution technique was used for retrieval of R(t). Absolute CBF was given by Zierler's area-to-height relation and the central volume principle. For elimination of large vessels (ELV), all MRI-based CBF values exceeding 2.5 times the mean CBF value of the slice were excluded. A correction for partial-volume effects (CPVE) in the artery used for AIF monitoring was based on registration of signal in a phantom with tubes of various diameters (1.5-6.5 mm), providing an individual concentration correction factor applied to AIF data registered in vivo. In the Xe-133 SPECT investigation, 3,000-4,000 MBq of Xe-133 was administered intravenously, and CBF was calculated using the Kanno Lassen algorithm. When ELV and CPVE were applied, DSC-MRI showed average CBF values from the entire slice of 43 +/- 10 ml/(min 100 g) (small-artery AIF) and 48 +/- 17 ml/(min 100 g) (carotid-artery AIF) (mean +/- S.D., n = 10). The corresponding Xe-133-SPECT-based CBF was 33 +/- 6 ml/(min 100 g) (n = 10). The relationships of CBF(MRI) versus CBF(SPECT) showed good linear correlation (r = 0.74-0.83).     

Use of zeolite filter in portable equipment for recovering SF/sub 6/ in SF/sub 6//N/sub 2/ mixtures

We propose using zeolite filter in portable processing equipment to separate and recover SF/sub 6/ from SF/sub 6//N/sub 2/ mixtures. Testing of a prototype filter has shown that SF/sub 6/ is completely adsorbed by the filter, and the exhaust gas from the filter is pure N/sub 2/ (i.e., at an undetectable level of SF/sub 6/ measured by a gas chromatograph). The SF/sub 6/ fraction of the sorbate phase (solid phase) has been evaluated from the tests, and it is a high fraction as expected. For example, 70% sorbate corresponding to processing a mixture of 5% SF/sub 6/, and 83% sorbate corresponding to processing a mixture containing 10% SF/sub 6/. The process of releasing the SF/sub 6/ from the filter and renewing the filter by thermal swing adsorption (TSA) technique has also been investigated to determine a commercial operating temperature of 150/spl deg/C. The feasibility of applying this new concept in a portable SF/sub 6//N/sub 2/ mixture processing equipment is discussed and compared with an alternative system that the authors developed previously which was based on pressure swing adsorption (PSA) technique and also using zeolite.     

Er-doped AlGaAs native oxides: photoluminescence characterization and process optimization

We present 300 K photoluminescence (PL) characterization data for wet thermal native oxides of Al/sub 0.58/Ga/sub 0.42/As films grown by metal organic chemical vapor deposition and doped with Er via multiple high-energy ion implants (for 0.0675, 0.135, and 0.27 atomic percent (at.%) peak Er concentrations), and Al/sub 0.5/Ga/sub 0.5/As and Al/sub 0.8/In/sub 0.2/As films doped with Er (0.03-0.26 at.%) during molecular beam epitaxy crystal growth. Broad spectra with a /spl sim/50-nm full-width at half-maximum and a PL peak at 1.534 /spl mu/m are observed, characteristic of Al/sub 2/O/sub 3/:Er films. The dependencies of PL intensity, spectra, and lifetime on annealing temperature (675/spl deg/C-900/spl deg/C), time (2-60 min) and As overpressure (0-0.82 atm) are studied to optimize the annealing process, with As considered as a possible quenching mechanism. Wet and dry-oxidized films are compared to explore the role of hydroxyl (OH) groups identified by Fourier transform infrared (FTIR) spectroscopy. FTIR experiments employing heavy water (D/sub 2/O) suggest that OH groups in wet oxidized AlGaAs come mainly from post-oxidation adsorption of atmospheric moisture. AlGaAs:Er films wet oxidized with 0.1% O/sub 2/ added to the N/sub 2/ carrier gas show a fourfold PL intensity increase, doubled PL lifetime to /spl tau//spl sim/5.0 ms (0.27 at.% implanted sample), and the lowest degree of concentration quenching.     

The impact of LOC structures on 670-nm (Al)GaInP high-power lasers

We investigate the potential of large optical cavity (LOC)-laser structures for AlGaInP high-power lasers. For that we study large series of broad area lasers with varying waveguide widths to obtain statistically relevant data. We study in detail I/sub th/, /spl alpha//sub i/, /spl eta//sub i/, and P/sub max/, and analyze above-threshold behavior including temperature stability and leakage current. We got as expected for LOC structures minimal /spl alpha//sub i//spl les/1 cm/sup -1/ resulting in /spl eta//sup d/=1.1 W/A for 64/spl times/2000 /spl mu/m/sup 2/ uncoated devices. We obtain total output powers /spl ges/3.2 W (qCW) and /spl ges/1.5 W (CW) at 20/spl deg/C.     

Simulations on an SOI grating-based optical add/drop multiplexer

With a suitable design of a silicon-on-insulator (SOI) grating in an optical add/drop multiplexer (OADM), a wavelength-selectable OADM is described. We study the characteristics of the designed SOI OADM and simulate the relationship of the electron and hole concentrations including the temperature effect by using integrated system engineering-technology computer aided design. The voltages for the light wavelength switching are evaluated and are best set below V/sub p/=0.8 V and above V/sub n/=-0.8 V for OADM operation. The light propagation and spectral response are also illustrated in finite-difference time-domain beam propagation method (FDTD-BPM) simulations.     

Simulation of the heat transfer phenomena in variable section fusible elements in non-adiabatic regime

Contents The paper presents an original method for solving the general heating equation of fusible elements in overload state, designed for current-limiting electrical fuses made up of rectangular construction silver tapes. The computation algorithm and the mathematical model for a digital computer are also presented. Finally the computational results and experimental ones, obtained by testing some current-limiting fuse-links designed for power transformer protection, are compared.

---

Simulation des nichtadiabatischen Wärmeübergangs in schmelzbaren Elementen variablen Querschnitts

Übersicht Die Arbeit beschreibt eine neue Methode zur Lösung der allgemeinen Wärmegleichung, angewendet auf schmelzbare elektrische Sicherungselemente aus Silberband mit rechteckigem Querschnitt unter Überlast. Der Algorithmus und das mathematische Modell für die Computerberechnung werden dargestellt. Schließlich stellen die Autoren einen Vergleich zwischen den Rechenergebnissen und den Meßergebnissen an, dic an Schmelzsicherungselementen für den Schutz von Leistungstransformatoren durchgeführt wurden.

---

List of symbols specific mass, kg/m³ - c specific heat, Ws/(kg°C) - thermal conductivity coefficient, W/(m°C) - K side surface heat transfer coefficient, W/(m²°C) - fusible element resistivity, m - J electric current density, A/m² - l _x conductor peripheric length (perimeter), m - A _x conductor cross-section area, m² - _a ambient temperature, °C - fusible element temperature, °C - ₀ fusible element resistivity at 0 °C, m - ₀ resistivity variation coefficient, function of temperature, at 0°C, 1/°C - g fusible tape thickness, m - a fusible tape minimum half-width, m - b fusible tape maximum half-width, m - d perforation half-length, m - i current instantaneous value, passing through the fusible element, A - I current effective value, A - I ₀ current effective value on a fusible element in overload regime, A - ₀ fusible element initial temperature, °C - _f fusible element fusion temperature, °C - t _f time necessary for the fusible element temperature to reach its fusion temperature, _f , in thex=0 point, s - _max maximum temperature the fusible element can reach, °C - T time constant, s - n number of the interval divisions [0,d] - x discretization step - N system equation number obtained through discretization     

Thermal resistance characterization of implanted subcollector InP-based HBTs

We present the results of measurements of thermal resistivity of the heterojunction bipolar transistor (HBT) devices, utilizing selective ion implantation to define the subcollector. This new device fabrication technique resulted in high-speed HBT devices with substantially reduced thermal resistivity, compared to devices utilizing the conventional fabrication approach which includes mesa isolation for pattern definition. The measurements were taken on full-thickness 3" InP wafers at T/sub amb/ from 30/spl deg/C to 180/spl deg/C and two separate emitter current densities. We present data on three device epitaxial structures with identical device layouts and discuss the relationship of V/sub be/ to temperature at these elevated power and temperature levels.     

Lasing at 1.28 /spl mu/m of InAs-GaAs quantum dots with AlGaAs cladding layer grown by metal-organic chemical vapor deposition

We report the device characteristics of stacked InAs-GaAs quantum dot (QD) lasers cladded by an Al/sub 0.4/Ga/sub 0.6/As layer grown at low temperature by metal-organic chemical vapor deposition. In the growth of quantum dot lasers, an emission wavelength shifts toward a shorter value due to the effect of postgrowth annealing on quantum dots. This blueshift can be suppressed when the annealing temperature is below 570/spl deg/C. We achieved 1.28-/spl mu/m continuous-wave lasing at room temperature of five layers stacked InAs-GaAs quantum dots embedded in an In/sub 0.13/Ga/sub 0.87/As strain-reducing layer whose p-cladding layer was grown at 560/spl deg/C. From the experiments and calculations of the gain spectra of fabricated quantum dot lasers, the observed lasing originates from the first excited state of stacked InAs quantum dots. We also discuss the device characteristics of fabricated quantum dot lasers at various growth temperatures of the p-cladding layer.