Showing papers in "IEEE Transactions on Electron Devices in 1969"

Large-signal analysis of a silicon Read diode oscillator

Abstract: This paper presents theoretical calculations of the large-signal admittance and efficiency achievable in a silicon p-n-v-ns Read IMPATT diode. A simplified theory is employed to obtain a starting design. This design is then modified to achieve higher efficiency operation as specific device limitations are reached in large-signal (computer) operation. Self-consistent numerical solutions are obtained for equations describing carrier transport, carrier generation, and space-charge balance. The solutions describe the evolution in time of the diode and its associated resonant circuit. Detailed solutions are presented of the hole and electron concentrations, electric field, and terminal current and voltage at various points in time during a cycle of oscillation. Large-signal values of the diode's negative conductance, susceptance, average voltage, and power-generating efficiency are presented as a function of oscillation amplitude for a fixed average current density. For the structure studied, the largest microwave power-generating efficiency (18 percent at 9.6 GHz) has been obtained at a current density of 200 A/cm2, but efficiencies near 10 percent were obtained over a range of current density from 100 to 1000 A/cm2.

Electromigration—A brief survey and some recent results

Abstract: Recently, electromigration has been identified as a potential wear-out failure mode for semiconductor devices employing metal film conductors of inadequate cross-sectional area. A brief survey of electromigration indicates that although the effect has been known for several decades, a great deal of the processes involved is still unknown, especially for complex metals and solute ions. Earlier design equations are improved to account for conductor film cross-sectional area as well as film structure, film temperature, and current density. Design curves are presented which permit the construction of high reliability "infinite life" aluminum conductors for specific conditions of maximum current and temperature stress expected in use. It is also shown that positive gradients, in terms of electron flow, of temperature, current density, or ion diffusion coefficient foreshorten conductor life because they present regions where vacancies condense to form voids.

Charge pumping in MOS devices

Abstract: Gate pulses applied to MOS transistors were found to stimulate a net flow of charge into the substrate Investigation of this effect revealed a charge-pumping phenomeonon in MOS gate-controlled-diode structures A first-order theory is given, whereby the injected charge is separated into two components One component involves coupling via fast surface states at the Si-SiO 2 interface under the gate, while the other involves recombination of free inversion-layer charge into the substrate

Excitation of elastic surface waves by spatial harmonics of interdigital transducers

Abstract: The electric field distribution from a surface wave interdigital transducer is investigated. Expressions for the various spatial harmonics and the capacitance of a transducer applied on an anisotropic, small-coupling propagation medium are derived. Experiments are consistent with the calculations.

Current crowding on metal contacts to planar devices

Abstract: A simple transmission line model is applied to the contact region between metal and diffusion layer in planar devices. Taking into account the sheet resistance of the diffusion layer and an ohmic specific contact resistance between metal and semiconductor the theoretical current distribution across the contact is calculated and compared with the results obtained with a model suggested by Kennedy and Murley. Experimental data are given that confirm the validity of the transmission line model.

Current gain and cutoff frequency falloff at high currents

Abstract: A theoretical and experimental study of the effects of high-level injection of carriers into a reverse-biased collector-base junction has been performed. Two models which describe the high-current behavior of the junction space-charge region are discussed. The first deals with the formation of a current-induced base region at space-charge-limited current densities. The second model assumes that two-dimensional effects are predominant; at current densities corresponding to the onset of space-charge-limited current, lateral injection of carriers takes place. These phenomena were studied experimentally using silicon double-diffused transistor structures. The existence of space-charge-limited current in the reverse-biased collector depletion layer manifests itself in significant changes in the ac and dc parameters of the transistor. In particular, it is shown that the cutoff frequency (f T ) and large-signal current gain (h FE ) begin to decrease rapidly with increasing current at the onset of the space-charge limitation. A comparison of experimental results with predictions of the above theories indicates that, while both the formation of a current-induced base region and lateral injection do take place, the latter mechanism controls conventional device performance.

A technique for directly plotting the inverse doping profile of semiconductor wafers

Abstract: A new technique for plotting doping profiles of semiconductor wafers is described. This technique involves driving a Schottky diode deposited on the surface with a small constant RF current (a few hundred microamperes at 5 MHz). The depth of the depletion layer is varied by changing the dc bias, but this is the only role of the dc voltage. The inverse doping profile n-1(x) is obtained by monitoring the voltage across the diode at the fundamental frequency, which is proportional to the depth x, and the second harmonic voltage, which is proportional to n-1. This type of plotter has the advantages of simplicity, high resolution (limited only by the Debye length in most cases), immediate results, and economy.

Thermal filaments in vanadium dioxide

Abstract: A relatively simple thin-film structure using vanadium dioxide is analyzed in detail to show that high-temperature, low-resistivity filaments can exist in VO 2 under appropriate electrical bias. These filaments will dominate both the static and dynamic electrical behavior of the structure, giving rise to V-I characteristics similar to those expected from materials exhibiting current-controlled bulk negative resistivities. The analysis indicates that these filaments may not only provide electrical features important from a device point of view, but also provide a tool for the measurement of fundamental material properties and for studying the dynamics of the phase transition in VO 2 .

Theory and experiments of low-frequency generation-recombination noise in MOS transistors

Abstract: A theoretical model for the generation-recombination (g-r) noise in MOS transistors is presented. This model takes into account the charge induced on all electrodes by the charge fluctuation of the impurity center in the depletion region. The model gives a finite equivalent gate noise resistance at saturation. Gold-doped and no-gold control devices were fabricated to verify the theory experimentally. The drain-voltage dependence of the g-r noise, which is shown to be distinctly different from the 1/f noise and thermal noise, is used to check the theory. Good agreement between theory and experiment is obtained.

Investigation of current-gain temperature dependence in silicon transistors

Abstract: This paper presents the results of an investigation of the factors influencing transistor current-gain temperature dependence. Phase one of the investigation demonstrated experimentally that current-gain temperature dependence is an inverse log function of the average active base resistivity. Devices were found to be less temperature sensitive with lighter base doping levels. Phase two of the investigation experimentally demonstrated current-gain temperature dependence to be an inverse exponential function of the emitter doping level. This large temperature dependence is shown to be a possible consequence of an emitter band-gap decrease, presumably caused by the large number of dislocations and lattice deformations at high doping levels. Using the predicted techniques, several runs of epitaxial planar devices were fabricated which had practically no current-gain temperature dependence, and suffered no noticeable loss in other parameters. Current-gain temperature dependence is also thought to be one of the contributors to hot-spot formation and secondary breakdown. To determine if secondary-breakdown capability had been increased, these temperature-insensitive devices were tested to forward secondary breakdown and compared to standard products of the same geometry. The temperature-independent devices were able to withstand a 40-percent increase in operating power before the onset of secondary breakdown. This increase in capability is thought to be the result of less tendency toward hot-spot formation due to decreased thermal regenerative action.

Forward current—Voltage and switching characteristics of p + -n-n + (epitaxial) diodes

Abstract: The forward-biased current-voltage and forward-to-reverse biased switching characteristics of p+-n-n+epitaxial diodes are investigated. The manner in which the n-n+junction affects the flow of injected minority carriers in the epitaxial region is characterized by a leakage parameter a. Experimentally, for diodes with epitaxial film widths much less than a diffusion length, a "box" profile accurately describes the injected minority carriers in the n region. The current is found to increase with increased epitaxial width at a fixed bias. A general switching expression for epitaxial diodes is presented and the validity of the expression is shown experimentally. The experimental values of a, determined independently from the current-voltage and switching characteristics, are in good agreement and show that the leakage of the high-low junction is dominated by the recombination of minority carriers in the n-n+space-charge region.

Avalanche degradation of h FE

Abstract: Degradation of low current h FE as a result of avalanching the emitter-base junction of a silicon planar transistor is shown to be explinable by an increase in surface recombination velocity within the vicinity of the emitter-base metallurgical junction. This degradation mechanism manifests itself on a gate-controlled transistor by an additional peak in the plot of base current versus gate voltage. The magnitude as well as the "width" of this peak can be accounted for theoretically by assuming a given emitter-base impurity gradient and a localized increase in surface recombination velocity in and around the metallurgical junction. Experimental evidence is also presented for localized charge trapping within the oxide over the emitter-base junction as a result Of avalanching under an applied field.

Design formulas for helix dispersion shaping

Abstract: A simple method of calculating effects of various boundaries around the helix on the dispersion relation and the interaction impedance is presented. The analytical technique is based on the use of equivalent circuit parameters derived from rigorous field analyses. The main advantage of this method is that effects of various perturbing objects (dielectrics, shields, etc.) can be calculated independently, regardless of the model of the unperturbed system. Equivalent circuit parameters are obtained for helices with several different forms of perturbing boundaries, including wedge-shaped dielectric rods, conducting boundaries, longitudinally conducting wires and fins, and combinations of these. The method of analysis is versatile so that the dispersion relation for any one particular form of the helix may be calculated from known characteristics of another using only three sets of universal curves.

Radiation resistance of Al 2 O 3 MOS devices

Abstract: Integrated circuits employing MOS devices will play a vital role in tomorrow's civilian and military electronics if their degradation in a radiation environment can be eliminated. One possible approach toward alleviating radiation effects in MOS devices is to use a material with a defect structure that does not allow predominant trapping of either holes or electrons as a gate insulator. This has been done by constructing MOS devices with plasma-grown aluminum oxide. The Al 2 O 3 films are formed by first depositing aluminum on freshly cleaned and properly prepared silicon wafers. Subsequently this aluminum is oxidized in an oxygen plasma and device fabrication is then completed. The devices have excellent characteristics and stability, and their fabrication is not restricted by the conditions of the ultra-clean procedures necessary for SiO 2 -Si devices. Exposure to 1-MeV electron bombardment at various fluence levels and bombardment-bias conditions shows that these structures possess remarkable radiation resistance. Up to a fluence of 1 × 1013e/cm2, under positive or negative bias, no oxide charge buildup or interface state generation is detectable. Above that fluence, only small shifts are observed. This indicates that an order of magnitude improvement in device hardening can be achieved by the use of this material.

A method for heat flow resistance measurements in avalanche diodes

Abstract: Avalanche transit time oscillators are operating at power densities approaching 106W/cm2, unprecedented in semiconductor device history. At such power densities, heat flow resistance problems at the interface between the flip-chip mounted silicon chips and the metal substrate, as well as between the package and the heat sink, are extremely critical. This paper describes a new, nondestructive and accurate method of measuring the heat flow resistance between junction and heat sink by utilizing the temperature dependent breakdown voltage V b (T) as a conveniently built-in temperature sensor. Variations in junction temperature ΔT with power ΔP= V b ΔI are, therefore, related to variations in breakdown voltage ΔV b with current ΔI resulting in a contribution to the electrical small signal resistance of the diode. This thermal resistance contribution R th can be separated readily from spreading and space charge resistance R ap and R sc because of the frequency dependence of R th (ω). Furthermore, the frequency dependence of R th (ω) allows the separation of heat flow resistance contributions originating in the immediate vicinity of the junction (Si-metal interface) from contributions originating at a poor thermal contact between package and heat sink. In keeping with calculations on simplified geometrical configurations, for which analytical solutions of the frequency dependent heat flow in a distributed circuit could be obtained, experimental results are presented which indicate that both heat flow resistance contributions can be extracted and separated with sufficient accuracy from as few as three electrical resistance measurements, e.g., at dc, 100 Hz, and 1 MHz. The simplicity of such measurements and their evaluation make this technique ideal for in-line testing of production devices.

High injection in epitaxial transistors

Abstract: An analysis of base widening and the current dependence of the cutoff frequency f T has been given previously [1]. The analysis is approximate and is based on the location of the edges of transition regions. Definition of transition regions becomes problematic in transistors having nonuniform base doping and for high-current densities. Such difficulties are avoided in this paper by use of the charge control approach. Numerical solutions of delay time ( = 1/2\pi f_{T} ) as a function of current density are given for low reverse collector bias. Whereas transition regions can be defined only approximately, the electric field is a well-defined quantity, and changes in the electric field that accompany base widening are shown in detail. At low injection levels, a high-field region exists near the transition between base and epitaxial layer. This high-field region is relocated to the interface between epitaxial layer and substrate under high injection conditions. When this "high-field relocation" has occurred, the epitaxial layer acts as an extension of the base with an attendant increase in the delay time [1].

Ceramic ferroelectric field effect studies

Abstract: The ferroelectric field effect has been observed in a semiconducting thin film of n-type tin oxide deposited on a ferroelectric lead zirconate-titanate ceramic substrate. The semiconductor was deposited by electron gun evaporation onto a thermally depolarized (randomly oriented) substrate which permitted the carrier concentration of the film to be enhanced or depleted depending on the direction of polarization of the substrate. Typical average resistivity values of 200-A films are approximately 0.1 Ω.cm for the depoled state, 0.01 Ω.cm for the enhanced state, and 100-1000 Ω.cm for the depleted state. "On"-"off" ratios as high as 1.7 \times 10^{5} have been observed in a single device. The transition from enhancement to depletion is quasi-continuous due to the small size and random orientation of the individual crystallites in the ceramic. Conductance measurements during this transition have yielded field effect mobilities in the range 7-10 cm2/V.s; and maximum average carrier densities in the range 0.5-1.0 × 1020carriers/cm 3 . The tin oxide-ceramic devices described here suffer from the long-term drift that is characteristic of many field effect devices. The resistivity of a device stored in the depleted state will decrease from 3 to 4 orders of magnitude in times between 104-105minutes.

Complementary MOS—Bipolar transistor structure

Abstract: Bipolar transistors can be used to increase the driving capabilities of complementary MOS transistors while retaining the low power dissipation feature. The fabrication of n-p-n bipolar transistors is compatible with the fabrication of the complementary MOS transistors in a monolithic structure. Common collector n-p-n transistors can be fabricated using a diffused n+source-drain region as emitter, a diffused p-isolation region as base and an n-substrate as collector with a h fe greater than 100. Lateral n-p-n transistors can be fabricated using a diffused n+source-drain region as emitter and collector, and p-isolation region as base with a h fe greater than 10.

A silicon MOS magnetic field transducer of high sensitivity

Abstract: A structure has been devised which converts magnetic flux density change to a change in output current. The structure is essentially a P-channel MOST with the drain diffusion split into two halves. A magnetic field normal to the silicon surface deflects device current towards one half-drain. By operating the MOST in the "pinched-off" mode (V DS > V GS -V T ) the output impedance is made high, so that large output voltage swings may be obtained. A theoretical study of the voltage and current distributions in the MOST channel has given data on the influence of device geometry on sensitivity. Experimental results indicate a linear relationship between output current and magnetic flux density, and an unexplained nonlinear variation of output with device current. Comparison of experimental results with theory indicates a carrier Hall mobility in the channel of 116 cm2/V.s.

Ultimate heat-pipe performance

Abstract: Ultimate heat-transfer limitations imposed by sonic vapor flow were determined in heat pipes for sodium, potassium, and cesium working fluids. Each fluid was investigated in a heat pipe consisting of an inner porous tube, an annulus for liquid return, and an outer container tube. Thin, rigid tubes with very small pores were obtained by compressing several layers of fine-mesh screen. These tubes allowed large capillary forces to develop so that sonic vapor flow could be achieved at several operating temperatures. The results of the investigation showed that sonic limitations were influenced strongly by the temperature and the working fluid. Reasonable agreement was found between the experimental results and existing theory. It was also found that the theory could be used to predict evaporator pressure and temperature gradients when the heat pipes were operated at various fractions of their ultimate heat-transfer capability.

The realization of a GaAs—Ge wide band gap emitter transistor

Abstract: The fabrication and properties of an n-GaAs emitter, p-Ge base, n-Ge collector transistor which possesses significant current gain is described, These GaAs wide band gap emitter transistors have shown incremental current gains near 15 when operated at current densities up to 3500 A/cm2. The doping level used in the base region was quite high (up to 5×1019/cm3in order to avoid a spurious Ge p-n junction in this region. The epitaxial deposition of the GaAs emitter region was carried out at a low temperature in order to also avoid a hidden p-n Ge junction. The low deposition temperature resulted in low (about 5×1015/cm3emitter doping levels. The general nature of the GaAs-Ge heterojunction energy-band diagram permits this high doping in the base or Ge region relative to the GaAs emitter region without reducing the current gain below unity. The observation of gain in this n-p-n heterojunction structure where the emitter is much more lightly doped than the base is considered to be confirmation of the theoretical proposals of Shockley and Kroemer.

Influence of dislocations on properties of shallow diffused transistors

Abstract: The influence of primary (eg, grown-in) dislocations on transistor is studied by making use of the Sirtl etch for the detection of crystallographic imperfections The first part of this paper demonstrates a one to one relationship between soft characteristics, caused by emitter-collector pipes and the presence of sliplines in transistors The sliplines, which are bands of high dislocation density, can be revealed only within the collector regions of the examined transistor The second part shows that dislocations can also be revealed within the emitter and base regions when an auxiliary voltage is applied on the collector during the etching procedure As the emitter of the used test vehicle is divided into many small regions (3 by 10 microns), the electrical behavior of even single dislocations can be detected with the aid of this etching technique It is concluded that the primary dislocations provide paths for locally enhanced diffusion of the phosphorus impurities in shallow diffused structures Phosphorus precipitates within the high-concentration part of the solute profile are probably the cause of this enhancement It is demonstrated that the enhanced diffusion along primary dislocations causes serious leaks between emitter and collector in shallow diffused transistors The resistance of these pipes can be in the order of 10,000 ohms

Solid-state digital pressure transducer

Abstract: The solid-state digital transducer represents a new standard in precision pressure transduction for the next generation of airborne sensors. The pressure transducer produces pulse train outputs as a measure of pressure by developing strains in a silicon diaphragm which incorporates piezoresistive sensing elements. These elements are distributed resistance-capacitance (RC) networks which are diffused into a diaphragm surface as the control elements of phase shift oscillators. By this approach, a digital (frequency) signal generated at the source can be transmitted without noise and distance limitations and the need for precision analog to digital conversion is eliminated. The device exhibits advantages in the areas of reliability, accuracy, size and cost over present day analog devices. The solid-state digital pressure transducer is being developed to meet the requirements of supersonic and subsonic air data applications when coupled with a high-performance air data computer. This application requires low hysteresis with repeatability and stability which are the main features of the solid-state pressure transducer. Other possible applications are FM data acquisition systems and industrial robots.

Electromigration in thin gold films

Abstract: Electromigration in gold film conductors results in regions of material accumulation and depletion. These regions arise from and are correlated with the thermal gradients along the conductor. This result is in contrast with the observations on aluminum conductors where the effect is dominated by the structure or metallurgical properties of the film.

Electrooptic properties of transparent ferroelectric glass—Ceramic systems

Abstract: The electrooptic and ferroelectric properties of glass-ceramic systems containing Na 0.5 K 0.5 NbO 3 and BaTiO 3 as the crystalline phases are presented. Domain activity is observed by dielectric and optical hysteresis measurements, although it becomes significant only at the larger crystallite size. Thus, the samples transmit light to only a small extent. Bistable optical switching has been achieved with 105V/cm, 100-µs duration pulses. The ratio of transmission measured between crossed polarizers is 2:1.

p-n junction—Schottky barrier hybrid diode

Abstract: A new fabrication technique for passivated silicon Schottky barrier diodes is described. It is shown that the p-n junction guard ring or "hybrid" approach produces Schottky barriers whose forward and reverse electrical characteristics are in excellent agreement with simple theory, and that the excess noise normally found in passivated Schottky barrier diodes has been significantly reduced. The influence of metal barrier height and diffusion profile on the charge storage characteristics of these devices is discussed and examined experimentally.

The pyroelectric detector of infrared radiation

Abstract: The pyroelectric detector is a thermal sensor of infra-red radiation requiring no bias. While in principle a pure capacitor (hence theoretically noiseless), the detector has a varying noise contribution as a function of frequency due to a load resistor, series loss resistance, and amplifier. The actual sensor is a pyroelectric crystal exhibiting spontaneous polarization. The spontaneous polarization and dielectric constant of the crystal are temperature-dependent. A change in incident power raises the detector temperature causing an electric charge to appear across the electroded surfaces cut perpendicular to the crystal's ferroelectric axis. Under open circuit conditions, a voltage is obtained which is ultimately neutralized by current flow through the leakage resistance or load resistor. The evacuated detector package incorporates an electroded flake of triglycine sulfate mounted on a substrate of low thermal and electrical conductivity, a field effect transistor, load resistor, and an infrared transparent window. Data on the detectivity, responsivity, and noise as a function of frequency and area are presented. Polycrystalline pyroelectric detectors of TGS are feasible and simplify the construction of arrays and mosaics. Applications of the pyroelectric detector to date have been in a multielement line scanner, thermal imaging camera, spectrometer, and laser calorimeter.

Ferroelectric electrooptic ceramics with reduced scattering

Abstract: Ferroelectric lead zirconate-titanate ceramics appear promising for use as electrically controlled retarders because both conventional electrooptic and nonvolatile incremental retardation changes can be induced in areas as small as 25 µ on a side. One particular composition, Pb 0.97 La 0.02 Zr 0.65 Ti 0.35 O 3 , has so far proved superior to all other rhombohedral lead zirconate-titanates when judged by either the conventional electrooptic retardation variation with applied bias voltage, or the nonvolatile incremental retardation change with remanent polarization state of the ceramic. Furthermore, this same lanthanum-doped composition exhibits a marked decrease in undesirable light scattering when compared with extensive data on the same composition with bismuth doping. Scattering also decreases with increasing wavelength λ, and the ceramic transmittance is greater than 85 percent for 3 µ<λ<8 µ with sample thickness less than 50 µ. This increased transmittance now makes feasible the construction of multistage ceramic electrooptic systems.

h FE degradation due to reverse bias emitter-base junction stress

Abstract: Reverse bias stress of the emitter-base junction may degrade the h FE of a transistor when it is operated under normal bias conditions. The degradation mechanism exhibits no measurable temperature dependence and appears to be a surface effect. The junction characteristics exhibit no structure as a function of field plate voltage after stress. Pulse measurements indicate that the time required to induce degradation is less then 1.0 µs. Hot carriers appear to have an appreciable effect on the degradation. A mathematical model is presented to facilitate the comparison of degradation on devices with different initial current gains.

Failure mechanisms in large-scale integrated circuits

Abstract: A study was made of the factors affecting the reliability of large-scale integrated (LSI) circuits. Particular attention was given to the effect on array reliability of the additional processing steps required to obtain multilevel metalization. The additional significant steps are low temperature deposition of a second dielectric layer on metalized LSI wafers, etching of vias through the second dielectric, and second level metalization. Consideration was also given to the effect of other trends in LSI, such as the use of smaller geometry and closer spacings. Possible new failure modes in multilevel arrays are: 1) shorts or increased leakage through or along deposited second layer dielectrics, 2) opens or increased series resistance in conductors, and 3) silicon surface effects. A review of literature on LSI was supplemented by results of a number of concurrent LSI programs and experimental studies. Test vehicles were designed to provide fundamental information about each of the categories of failure. Data obtained using these test vehicles to evaluate the multilevel metalized array structures are presented. A discussion concerning the merits and limitations of the specially designed test vehicles for process development, process control, and reliability tests is given. It is concluded that with the proper in-process controls, tests and screens, LSI arrays will be substantially more reliable per function accomplished than are conventional integrated circuits.