Showing papers in "Solid-state Electronics in 1983"

Minority carrier recombination in heavily-doped silicon

Abstract: A review of our present understanding about the minority carrier recombination in silicon with dopant concentration in the range of 1018–1020 cm−3 is presented. After providing a short phenomenological description of carrier recombination processes and lifetime, the main theories of carrier recombination in a semiconductor are briefly reviewed and their expected contributions to carrier recombination in silicon at heavy doping are indicated. The various methods used for measuring the minority carrier lifetime in heavily doped silicon are described and critically examined. Four different mechanisms are examined to explain the available lifetime data. Two of these involve SRH-type phononic recombination (i) via deep level traps generated by dopant introduced defects and (ii) through shallow donor/acceptor states. The other two non-phononic mechanisms are: (iii) Band to band Auger recombination and (iv) trap assisted Auger recombination. Mechanism (i) can not explain the observed insensitivity of lifetime to processing conditions and the dopant atoms, and contribution of (ii) remains insignificant up to the heaviest doping. Phonon assisted band to band Auger recombination appears to explain the measured lifetimes satisfactorily in p-type silicon. However, for n-type silicon this mechanism predicts considerably higher values of lifetime than the measured results and it is likely that mechanism (iv) (and probably (i) also) competes with this process. Calculations indicate that the rate of trap-assisted Auger recombination through the dopant generated acceptor states in p-type silicon and through donor states in n-type silicon becomes large enough to compete with the band to band Auger process at heavy doping. In n-type silicon Auger recombination through crystal defects like vacancies may also become important. Perhaps all these processes contribute to the carrier recombination at heavy doping but which of these controls the lifetime in n-type silicon is not known.

Ohmic contacts to III V compound semiconductors: A review of fabrication techniques

Abstract: After a brief introduction on the phenomena governing the ohmic contact formation and measurements in metal—semiconductor structures, we present a review of papers on the ohmic contact realization onto III–V compounds. We discuss the thermal behaviour of various multicomponent metal—semiconductor systems (alloying, sintering, use of lasers and electron beams) and comment about overdoping the semiconductor surface before metal deposition (diffusion, ion implantation, epitaxy). We show that, in a general way, the metal III–V semiconductor interactions lead to the formation of compounds. From an electrical point of view, it seems that the main consequence of the compounds appearance is not a large change of the barrier height due to a change of the interface chemistry but the rough interface resulting from particle precipitation. We conclude that, if contacts made up to now, are often simple and usable, they are still far from ideal.

Carrier recombination and lifetime in highly doped silicon

Abstract: The predominance of phonon-assisted band-band Auger recombination in highly doped silicon is demonstrated by showing that no recombination mechanism involving common (unavoidable) defects in silicon can yield carrier lifetimes that are consistent with the measured lifetimes, which exhibit an inverse-quadratic doping-density dependence, and/or with their temperature dependence. Both trap-assisted-Auger and Shockley-Read-Hall recombination mechanisms are considered, and dependences of the defect density on the doping density, which are implied by theory and experiment, are accounted for.

Rapid interface parameterization using a single MOS conductance curve

Abstract: Figures are presented for finding interface trap density Dit capture probability cp, and interfacial broadening σs, from a single curve of small-signal MOS interface trap conductance Git/ω vs either band bending νs, or frequency f[ω = 2πf]. Almost no computation is necessary. An additional figure allows the depletion layer capacitance to be determined if the oxide capacitance and the measured capacitance at the maximum value of Git/ω are known. The extension to Git/ω vs νs, curves is a convenience. Parameters throughout the bandgap can be obtained from measurements at fewer frequencies than are needed for construction of accurate Git/ω vs f curves. Economy also results because data usually are obtained by sweeping gate bias at selected frequencies. Construction of Git/ω vs νs curves from the data is simpler than Git/ω vs f, particularly if νs is measured directly. The information needed from the Git/ω curve is the maximum value of Git/ω and the width of the peak at an arbitrary fraction of its maximum value, fw. In addition, for capture probability only, the frequency and band bending at the peak maximum are needed, as well as the bulk doping. The method assumes validity of the Gaussian approximation to broadening of the conductance peak. Comparison of the parameters obtained for several choices of fw allows a check of this assumption, which usually is valid for MOS devices.

Schottky diodes with high breakdown voltages

Abstract: The methods of increasing the breakdown voltages in silicon Schottky diodes is presented. In addition to a guarding ring, screen-diffusion regions were introduced. In this manner, the electrical field near the Schottky contact was lowered and, as a result, higher breakdown voltages were obtained. By using this method, the breakdown voltage can be increased by a factor of 3–5. However, a large device area is required for the same Schottky contact area and, therefore, the junction parasitic capacitance is greater.

Improving the reverse recovery of power mosfet integral diodes by electron irradiation

Abstract: This paper demonstrates that controlled electron irradiation of silicon power MOSFET devices can be used significantly improve the reverse recovery characteristics of their integral reverse conducting diodes without adversely affecting the MOSFET characteristics. By using 3 MeV electron irradiation at room temperature it was found that the reverse recovery charge in the integral diode could be continuously reduced in a well controlled manner from over 500 nC to less than 100 nC without any significant increase in the forward voltage drop of the integral diode under typical operating peak currents. The reverse recovery time was also observed to decrease from 3 microseconds to less than 200 nsec when the radiation dose was increased from 0 to 16 Megarads. The damage produced in gate oxide of the MOSFET due to the electron radiation damage was found to cause an undersirable decrease in the gate threshold voltage. This resulted in excessive channel leakage current flow in the MOSFET at zero gate bias. It was found that this channel leakage current was substantially reduced by annealling the devices at 140°C without influencing the integral diode reverse recovery speed. Thus, the electron irradiation technique was found to be effective in controlling the integral diode reverse recovery characteristics without any degradation of the power MOSFET characteristics.

Hole and electron mobilities in heavily doped silicon: comparison of theory and experiment

Abstract: Most device models for npn or pnp transistors assume that hole (electron) mobilities in n-type and p-type silicon are equal. Partial-wave phase shift calculations for the contributions of carrier-dopant ion scattering to the carrier mobilities lead to unequal minority hole (electron) and majority hole (electron) mobilities at the same doping density. These calculations are valid over the doping range of 2 x 1019 to 8 x 1019 cm−3 in n-type and p-type silicon and contain the assumptions that the holes and electrons move in isotropic parabolic energy bands and are scattered by the screened Coulomb potentials of the dopant ions. When the effects of carrier-acoustic phonon and carrier-carrier scatterings are included, these calculations agree to within the spread of experimental value for the majority mobilities reported in the literature. This agreement is a substantial improvement by factors of 2–4 over the results of earlier theories such as first order Born and nondegenerate theories. The results of this work, particularly the inequality of minority and majority carrier mobilities, have implications for the modeling of both bipolar and field effect transistors.

Blocking capability of planar devices with field limiting rings

Abstract: The results of breakdown voltage investigations of planar devices with field limiting rings are described. The two-dimensional Poisson equation is solved using the finite difference method. The question of the optimal ring spacings for devices with more than one field ring and the influence of surface charges on the blocking capability are extensively studied. The influence of the following device parameters is discussed: ring spacing, ring width, doping gradient and surface charge density. An analytical model for simplified calculations showing the principal characteristics of field ring devices is presented. Furthermore some experimental results of EBIC measurements are shown.

Simplified long-channel MOSFET theory

Abstract: The now-classic and highly accurate double-integral expression for the drain current flowing in long-channel MOSFET's is shown to be reducible to a completely equivalent single-integral expression. The single-integral expression is used in turn to straightforwardly establish the approximate closed-form results commonly known as the bulk-charge and charge-sheet relationships. Sample computations which illustrate the general utility of the single-integral expression are also presented.

Ion-implantation associated defect production in silicon

Abstract: Defects produced during ion implantation have traditionally been removed during high temperature (⩾ 1200 K) thermal processing. The current industry trend towards lower temperature processing will require new techniques for suppressing defect formation associated with ion implantation. This paper details the characterization and identification of the electrically active defects which are produced during low dose ion implantation. Several of the observed defect levels can be correlated with defects which have been identified in electron irradiated silicon. However, a defect level near mid gap, H(0.53), which should be the dominant recombination center in p-type samples, cannot be correlated with previously identified defects. Similarly, the H(0.20) level is found to anneal near 420 K, and thus does not correlate well with the known properties of the divacancy, which introduces a similar trapping level.

Analysis of the dark current and photoresponse of In0.53Ga0.47As/InP avalanche photodiodes

Abstract: The doping profiles, current-voltage (I–V) and photoresponse characteristics of five In0.53Ga0.47As/InP avalanche photodiode (APD) wafers are presented. A detailed analysis indicates that the dark current is due largely to generation and recombination of carriers in the diode bulk, and in some wafers tunneling at the p-n junction is dominant near breakdown (VB). In some cases, significant surface currents are also observed. In three high-performance wafers, however, low primary dark currents (∼5 nA) with no evidence for tunnelling at 0.99 VB have been obtained. In addition, microplasmas have been found in some wafers, due to local breakdown possibly arising from crystalline defects. Nevertheless, we report uniform gains as high as 100. The dark current and gain characteristics of these devices are among the best reported to date for In0.53Ga0.47As/InP APDs. Finally, the response of the APDs to fast optical pulses has been analyzed at both low and high illumination intensity. The slow speed of response, which has been reported elsewhere, is considered in detail and is found to be due to charge pile-up at the abrupt n-In0.53Ga0.47As/n-InP heterointerface which is characteristic of our devices. Using an analysis of the response time thermal activation energy along with the transient pulse shape, we infer that the heterointerfaces are graded over a length of 2 L ≌ 300 A . The model predicts that fast response can be obtained for heterointerface grading lengths of 2 L ⩾ 500 A , depending on the epitaxial layer doping and extent of penetration of the depletion region into the In0.53Ga0.47As layer at breakdown.

A study of grown-in impurities in silicon by deep-level transient spectroscopy

Abstract: Deep levels due to various impurities incorporated into Czochralski silicon ingots during crystal growth have been delineated. The largest impurity-induced deep-level concentration, defined as the electrically active impurity concentration, is found to be a fraction of the metallurgical impurity content of the crystals. This fraction for a specific impurity depends on the thermal history of the sample and the ability of the impurity to diffuse. POCl3 gettering of Ti and V produces a decreasing electrically active impurity concentration toward the surface of a silicon wafer, while there is no observable effect of this heat treatment on the Mo concentration. In the case of Cr, which diffuses much more rapidly than Mo, Ti, or V in silicon, a very significant reduction in the electrically active concentration is observed after heat treatment. Similarly, in metal-doped polysilicon wafers the electrically active Mo concentration appears unaffected by grain boundaries, but the electrically active Cr concentration at or near some grain boundaries is reduced by more than an order of magnitude compared to that at grain centers.

1/f noise as a reliability estimation for solar cells

Abstract: The 1/f noise from a forward biased dark solar cell is a non-destructive reliability estimation. The experimentally observed 1/f noise is compared with Kleinpenning's one-dimensional calculations for p-n diodes. At medium and low currents the 1/f noise of n+-p solar cells is about 50 times as large as predicted. Such deviations can be caused by non-uniformities in the large junction area. Local areas with lower built-in potentials at the junction lead to hot spots and reduced reliability. At large currents, reliability problems due to possible poor contacts can be studied from the proportionality between the noise and the square of the current.

Presence of mobility-fluctuation 1f noise identified in silicon P+NP transistors

Abstract: The magnitude and location of mobility-fluctuation 1f noise sources have been identified by means of biasing a PNP transistor in a common emitter configuration with first a high and then a low source resistance. Comparison of the two noise spectra at the same base currents shows the low source resistor bias isolates the collector noise sources, and the high source rsolates base noise sources. The magnitude of the observed collector 1f noise gives an α − 2 × 10−6 from Kleinpenning's mobility-fluctuation theory. The base 1f noise gives an α ∼- 10−7 due to an impurity mobility reduction factor of about 100.

Common anion heterojunctions: CdTe-CdHgTe

Abstract: The problem of band distortions at heterojunctions between semiconductors with common ionisation energies is solved numerically. The case of CdTe-CdHgTe heterojunctions is treated specifically, including effects of interdiffusion of alloy constituents and dopants. Analytic approximations are derived for conditions controlling the occurrence or abscence of maxima and minima in the conduction band profile at the interface. The calculations are applied to show the range of suitability of CdTe-CdHgTe heterojunctions for photodiode production, and for CdTe passivation of inter-diode regions.

A quasi-one-dimensional analysis of vertical JFET devices operated in the bipolar mode

Abstract: “Bipolar operation”, namely forward-basing the gate-source diode of a JFET, has been proposed in the literature as a means to reduce the on-state resistance of such devices. In this paper, the physics of bipolar operation of power JFET's is analysed in detail and closed-form solutions of its output characteristics are derived as a function of the device geometry and of physical parameters of the semiconductor. From that model, it turns out that the low value of the saturation voltage originates from the existence of an high-density electron-hole plasma that fills the space between source and drain. In the active region of the output characteristics, the control of the gate current the drain current is due to the possibility to control the level of majority-carrier injection from the source transition. The closed-form expression for the current gain allows to identify the structure parameters that affect it. It shown that, under suitable conditions, a substantial current amplification can be observed. The model has been found to be in good agreement with the results obtained on experimental devices.

On the small-signal behaviour of the MOS transistor in quasistatic operation

Abstract: A complete MOST small signal model is proposed, chosen to related most easily to widely used incomplete models For quasistatic operation, expressions for the model parameters valid for all regions of operation, including weak and moderate inversion, are derived The predicted bias dependence of the nine model capacitance coefficients, derived starting from the long channel MOST theory recently presented by F Van de Wiele, is presented and compared to available experimental data

Microscopic Investigations of Semiconductor Interfaces

Abstract: We review the results obtained in recent years by applying surface-sensitive experimental techniques to the study of metal-semiconductor and semiconductor-semiconductor interfaces. The review starts with a short discussion of the most widely used experimental probes. Then we discuss the initial impact of this new approach on the old theoretical concepts about semiconductor interfaces. A presentation of selected results follows, including the detection of interface states, the formulation of the defect model, the discovery of chemical trends in the interface parameters and the recent progress in understanding heterojunction discontinuities. The review is concluded by a discussion of the open problems and of the most likely future developments in this field.

Temperature-ramp resistance analysis to characterize electromigration

Abstract: A temperature-ramp technique is presented that enables determination of the kinetic parameters for electromigration processes. With this method, pre-exponentials and activation energies can be measured in a single experiment requiring a few hours. The technique is applicable to both practical and fundamental studies of electromigration. For example, in terms of the former, reliability factors for conductor lifetimes can be readily determined. As far as the latter, new conductor compositions can be rapidly screened to ascertain their electromigration behavior. To demonstrate the application of the method, Temperature-ramp Resistance Analysis to Characterize Electromigration (TRACE) has been applied to thin-film aluminum conductors. Results have yielded activation energies in agreement with literature values. Furthermore, TRACE results have been used, along with resistivity results from the literature, to determine the current density effect on the pre-exponential factor. The value of the current density exponent thus determined is in accord with the range reported in the literature for Mean Time to Failure (MTF) experiments.

Damaged-induced isolation in n-type InP by light-ion implantation

Abstract: A detailed study of the insulating properties of ion-implantation induced damage in InP has been carried out for H, He, B and Be implantation. For each ion, there was found to be an optimal implantation fluence for the formation of resistive layers. At this fluence, a maximum resistivity of 10 3 to 10 4 Ω·cm was observed. Lower resistivities were observed for higher and lower implantation fluences. The primary anneal stage for the maximum resistivity layers was between 250 and 300°C. Anomalous results were observed for H implantation in that the resistivity observed depends on the test structure geometry. Measurements carried out by contacting the front and back of the damage layer gave resistivity values two orders of magnitude greater than those measured by contacting adjacent points on an epitaxial structure. For all other ions, the results obtained for the two geometries were in good agreement. It has been shown that a conductive layer produced by the proton bombardment of the underlying Fe-doped substrate gives rise to a low resistance shunt in the epitaxial study.

Amorphous BN films produced in a double-plasma reactor for semiconductor applications

Abstract: For semiconductor applications such as InP and GaAs devices, a double-plasma reactor for OPECVD (Organic Plasma Enhanced Chemical Vapour Deposition) BN films using moderate substrate temperatures and relatively high pressures is reported together with the chemical and physical parameters of the films produced. Particularly the conduction mechanism and its sensitivity to production parameters was investigated intensively.

CMOS circuit optimization

Abstract: In this paper, optimization algorithms for CMOS circuits are described, from the propagation delay time viewpoint The propagation delay time for a CMOS in erter is calculated for a step function input A classical model of I–V characteristics for a MOSFET and the worst case Sah model for inter-electrode capacitances of a MOSFET are used for this deduction

Ion-cleaning damage in (100) GaAs, and its effect on schottky diodes

Abstract: This study investigates the effect of ion cleaning damage of (100) GaAs in the 100–1000 eV range, and also its recovery with thermal annealing to 400°C. It is shown that GaAs could be annealed to a considerable extent if the ion-damage was ⩽ 100 eV. However, full recovery was not achieved. On the other hand, samples damaged at ⩾ 400 eV became progressively worse with annealing. Measurements indicate that these samples are dominated by the effect of arsenic variances within the bulk. These remain in the bulk, but are distributed spatially upon annealing. They behave as deep donors, so that the net electron concentration in the bulk is enhanced. Aluminum-n GaAs Schottky diodes were used as a vehicle for this study.

Dependence of electronic properties of polysilicon grain size and intragrain defects

Abstract: Lifetimes (τn), diffusion lengths (Ln) and consequently mobilities (μn) of minority carriers and also mobilities (μp) of majority carriers have been measured on small MESA diodes (1.5 × 1.5 mm) revealed on photocells realized on boron-doped cast polysilicon obtained by a fast variant Bridgmen method. The lifetime τn and the diffusion length Ln decrease drastically when the number of grains per diode NG increases. It has been deduced that the minority carrier mobility μn depends on the grain size even when this size is larger than 0.1 mm. This influence of the grain size is often screened by intragrain defects. This explains that generally the electronic properties are degraded at the bottom or at the periphery of the ingots where the dislocation density is high and at the top where the impurities are concentraled by the Bridgman process.

The characteristics of AuGe-based ohmic contacts to n-GaAs including the effects of aging

Abstract: We have investigated and compared the characteristics of Au:Ge, Ni/Au:Ge and Au/Ni/Au:Ge ohmic contact metallizations to ion-implanted and epitaxial n -GaAs layers on semi-insulting substrates. Auger depth profiles of ohmic contacts and SEM of surface microstructures have provided significant insight as to the nature and degradation mechanism of ohmic contacts with aging. An electrolytic tank model with distributed resistors representing nodular or cluster form contacts has been successfully used to understand the effects of non-uniform ohmic contacts. A small degree of change in the semiconductor sheet resistance in the gaps between contacts, with aging at an elevated temperature, has been attributed to possible lateral surface diffusion of Au in the Au:Ge contacts and Ge in the Ni/Au:Ge and Au/Ni/Au:Ge contacts.

Evidence of optical generation of minority carriers from saturated MOS transistors

Abstract: Minority carrier injection into the substrate by a MOS transistor operating in saturation presents a reliability problem in dynamic memory circuits such as RAM's and CCD's. The effect has been studied by measuring the substrate and drain currents of stressed transistors as a function of gate and drain voltages, firstly by the accumulation of minority carriers in a charge coupled device, and secondly by the direct detection of light from the drain region of a transistor. These results suggest that light emission associated with multiplication in the drain region is more important than the secondary impact ionization mechanism in the generation of minority carriers.

Lifetime and drift velocity analysis for electromigration in sputtered Al films, multilayers, and alloys

Abstract: Besides high lifetimes and a low standard deviation of failure times, the electromigration drift velocity is an important parameter for metallization reliability. Here, lifetimes should be high and drift velocity low. For pure Al films, only properly chosen and optimized sputter conditions yield good film quality and electromigration data comparable to evaporated stripes. Annealing and oxide covering improve lifetime and reduce drift velocity of pure Al structures. For redundant AlTiAl multilayers, annealing yielded an enormous lifetime enhancement of 3 orders of magnitude. However, these stripes are not suited for practical applications because of whisker formation and of an increased drift velocity! These disadvantages do not appear for a lifetime-improving TiN intermediate layer instead of Ti. A further method of lifetime enhancement is Cu-addition which reduces the drift of stripes. Finally, for (Al + Cu)TiN(Al + Cu)-structures, lifetime is improved even further in combination with low drift velocity.

The role of germanium in evaporated AuGe ohmic contacts to GaAs

Abstract: Auger depth profiling, electron probe microanalysis, electron beam induced current and conventional scanning electron microscopy have been used to study the inhomogeneous reaction which takes place during the alloying of evaporated AuGe films on (100), n-type GaAs. These measurements have been correlated with I–V and C–V data taken for heat treatment below the AuGe eutectic temperature (360°C) and specific contact resistance measurements above 360°C. The picture which emerges is that of the key role of the germanium in the metallurgical reaction, in addition to its accepted role as the n+ dopant. A simple metallurgical model is proposed which is based on the AuGe being the main reactive system at the interface responsible for the formation of the rectangular particles which are the pathways for the current. Minimum contact resistance correlates with the maximum growth of these particles.

Minority carrier lifetimes using compensated differental open circuit voltage decay

Abstract: Minority carrier lifetimes in semiconductor diodes can be measured by suddenly terminating forward diode current and observing the consequent decay of diode voltage. While this open circuit voltage decay (OCVD) method is very simple in principle, departures of the observed decay curves from the predicted linear relationship cause the method to be subjective and unreliable in practice. This paper describes simple improvements to this method which largely overcome these difficulties. These improvements are based on the passive differentiation of the open circuit decay curve and compensation for the effects causing departures of this curve from linearity.