Avalanche characteristics and failure mechanism of high voltage diodes
TL;DR: In this article, the avalanche characteristics of PνN junction diodes were analyzed and it was theoretically shown that they have negative resistance regions caused by the space charge effect of the carriers.
Abstract: High voltage diodes made of PνN (or PπN) structures sometimes fail by second breakdown when high voltage pulses are applied. The avalanche characteristics of PνN junction diodes are analysed and it is theoretically shown that they have negative resistance regions caused by the space charge effect of the carriers. Experiments on the second breakdown are also reported and it is concluded that the current concentration induced by the negative resistance may cause diode failure.
Citations
More filters
01 Aug 1967
TL;DR: A comprehensive review of the published literature dealing with the phenomenon of second breakdown in semiconductor devices and the problems it creates in the design, fabrication, testing, and application of transistors is presented in this paper.
Abstract: This paper is a comprehensive review of the published literature dealing with the phenomenon of second breakdown in semiconductor devices and the problems it creates in the design, fabrication, testing, and application of transistors.
123 citations
TL;DR: In this article, a unified treatment for switching and breakdown events, which occur in the range of fields of a few to more than 107 V/cm, is given, showing that the mechanisms of thermal events and electronic ones in semiconductor junctions are understood, but that the mechanism of insulator breakdown requires clarification.
96 citations
TL;DR: In this paper, a single event transient current and voltage waveforms in deep submicron CMOS integrated circuits are characterized using simulations and it is shown that the mechanism controlling the height and duration of the observed current plateau is the redistribution of the electrostatic potential in the substrate following a particle strike.
Abstract: Simulations are used to characterize the single event transient current and voltage waveforms in deep submicron CMOS integrated circuits. Results indicate that the mechanism controlling the height and duration of the observed current plateau is the redistribution of the electrostatic potential in the substrate following a particle strike. Quantitative circuit and technology factors influencing the mechanism include restoring current, device sizing, and well and substrate doping.
92 citations
Cites background from "Avalanche characteristics and failu..."
...The high potential drop (high electric field) at the highly doped P+ contact in the NMOSFET is very similar to the fields near junctions and contacts leading to second breakdown in power diodes, as described in [14] and [15]....
[...]
...An analytical study of the effect has been presented in [14], for avalanche breakdown, and in [17] for...
[...]
TL;DR: In this paper, an average temperature of the breakdown events during vaporization was obtained in metal oxide semiconductor (MOS) samples by photographing the spectrum of discharges and comparing the intensities of spectral lines.
Abstract: Publisher Summary This chapter focuses on theory of thermal breakdown and continues with theory and experimental description of breakdown first in semiconductors and then in insulators. Thermal breakdown voltages on the application of pulses are larger than the steady state breakdown voltages. Information on the basic breakdown properties of many substances is imperfect and relatively few insulators have been investigated in detail. An average temperature of the breakdown events during vaporization was obtained in metal oxide semiconductor (MOS) samples by photographing the spectrum of discharges and comparing the intensities of spectral lines. The calculation of the breakdown strength by impact ionization requires the solution of the Boltzmann equation, considering the effects of applied field, phonon, and interelectronic collisions, ionizations, and recombinations on the free electron distribution function. The statistical breakdown events observed in the oxides may also be explained by the development of critical size avalanches, if it is assumed that this depends on particular transport and trapping properties of the holes and on field emission at the cathode.
86 citations
References
More filters
TL;DR: In this article, the ionization rates for holes and electrons in silicon have been determined over the following ranges of field: for holes, (2.5-6.0)\ifmmode\times\else\texttimes\fi{}${10}^{5}$ volts
Abstract: The ionization rates for holes and electrons in silicon have been determined over the following ranges of field: for holes, (2.5-6.0)\ifmmode\times\else\texttimes\fi{}${10}^{5}$ volts ${\mathrm{cm}}^{\ensuremath{-}1}$; for electrons, (2.0-5.0)\ifmmode\times\else\texttimes\fi{}${10}^{5}$ volts ${\mathrm{cm}}^{\ensuremath{-}1}$. The ionization rate for electrons is higher than that for holes. The results suggest that the field dependence of the ionization rate for holes and, probably, for electrons also, can be expressed by $a\mathrm{exp}(\ensuremath{-}\frac{b}{E})$, where $E$ is the field. The constants $a$ and $b$ are different for electrons and holes.
526 citations
TL;DR: In this paper, the field dependence of mobility has been determined for electrons and holes in both germanium and silicon, and the observed critical field at 298\ifmmode^\circ\else\textdegree\fi{}K beyond which $\ensuremath{\mu}$ varies as ${E}^{-}\frac{1}{2}}$.
Abstract: The field dependence of mobility has been determined for electrons and holes in both germanium and silicon. The observed critical field at 298\ifmmode^\circ\else\textdegree\fi{}K beyond which $\ensuremath{\mu}$ varies as ${E}^{\ensuremath{-}\frac{1}{2}}$ is 900 volts/cm for $n$-type germanium, 1400 volts/cm for $p$-type germanium, 2500 volts/cm for $n$-type silicon, and 7500 volts/cm for $p$-type silicon. These values of critical field are between two to four times those calculated on the basis of spherical constant energy surfaces in the Brillouin zone. A saturation drift velocity of ${6(10)}^{6}$ cm/sec is observed in germanium which is in good agreement with predictions based on scattering by the optical modes. Data on $n$-type germanium at 20\ifmmode^\circ\else\textdegree\fi{}K show a range over which impurity scattering decreases and the mobility increases with field until lattice scattering dominates as at the higher temperatures.
292 citations
240 citations
TL;DR: In this article, it was shown that the ionizing regions of microplasmas are about 500 A in extent, have a current density approximately 2 × 2 ifmmode/times/else/texttimes, and a net space charge density approximately 10 − 6 − 18$/${\mathrm{cm}}^{2}, and two such regions are associated with each microplasma, except in such narrow junctions that they cannot properly be separated.
Abstract: At breakdown of reverse-biased silicon junctions one often observes not only discrete current pulses, but also minute luminous spots. It is postulated that these phenomena represent local discharges which are very similar to the gas discharge cathode fall. The mechanism is illustrated for the case of the $n\ensuremath{-}i\ensuremath{-}p$ structure. Approximate calculations show that the ionizing regions of these microplasmas are about 500 A in extent, have a current density \ensuremath{\approx}2\ifmmode\times\else\texttimes\fi{}${10}^{6}$ amp/${\mathrm{cm}}^{2}$, and a net space charge density \ensuremath{\approx}${10}^{18}$/${\mathrm{cm}}^{3}$. Two such regions are to be associated with each microplasma, except in such narrow junctions that they cannot properly be separated. The finite duration of the microplasmas, as evidenced by the current pulses, may be explainable on the basis of a statistical fluctuation, in which a fraction of the carriers fail to ionize. If the fluctuation exceeds a critical size, the carrier density rapidly decreases with time, and the microplasma is extinguished. Initiation of the microplasma in linear gradient and step junctions poses a space charge problem that is at present not resolved.
106 citations