Dimitri I. Cherednichenko

Electron-beam-induced current observed for dislocations in diffused 4H-SiC P–N diodes

The electron-beam-induced current (EBIC) method was employed to investigate the electrical activity of dislocations in silicon carbide Schottky and diffused p–n diodes. Dislocations in Schottky diodes appear as dark spots with the EBIC current signal at the dislocations reduced with respect to the background. However, in p–n diodes, the same dislocations exhibited characteristic bright halos, with the EBIC current higher than that of the background. These bright halos were attributed to a nonuniform impurity distribution around dislocations caused by the high-temperature (∼2000 °C) diffusion process.

Observation of dislocations in diffused 4H–SiC p-i-n diodes by electron-beam induced current

The electron-beam induced current (EBIC) method was employed to investigate the electrical activity of dislocations in silicon-carbide-diffused p-n diodes. It was observed that EBIC contrast depends on the type of defect (superscrew, screw, and edge dislocation). This dependence was attributed to spatial inhomogeneities in the electrical properties of the material around the dislocations due to different impurity-dislocation interactions during high-temperature (∼1900°C) diffusion. Chemical etching of the sample was used to define the nature of the defects observed by EBIC imaging. It was found that electrical breakdown of the diodes occurs at the location of superscrew dislocations.

Forward voltage drop degradation in diffused SiC p-i-n diodes

The time varying relationship between forward voltage drop and temperature in degrading diffused 4H–SiC p-i-n diodes was used to estimate the activation energy (0.34 eV) of the degradation process associated with the formation of stacking faults (SFs). A very strong peak appeared in the electroluminescence spectra at 427 nm and increased steadily in intensity as the forward voltage drop increased. The mismatch stresses, localized in the diffused doped region, are proposed to play a dominant role in the initial formation of SFs. Calculations were performed for the phonon pressure caused by nonradiative carrier recombination, which presumably is responsible for the development and motion of the SFs leading to the observed forward voltage degradation.

High field breakdown of narrow quasi uniform field gaps in vacuum

The challenge in vacuum microelectronic device design is to be able to stress a given micrometric gap to relatively high voltages without threat of a breakdown, which, in effect could destroy the device. In order to obtain basic vacuum insulation data related to the regime of vacuum microelectronics, the prebreakdown and breakdown characteristics of narrow gaps in the range of 3–25 μm were extensively investigated. The observed prebreakdown current was related to field emission from atomic scale microprotrusions or planar emission sites; the emission from these sites eventually produces breakdown. A single spark breakdown caused damage to both the anode and cathode. The dc glow discharge conditioning in air improved the insulation capability of narrow gaps (3–25 μm) significantly. The breakdown strength of a 5 μm gap after conditioning was as high as 5×108 V/m, which is the highest value reported in literature for broad area electrodes. It is shown that the electric field evaporation of metal ions from th...

Dislocations as a source of micropipe development in the growth of silicon carbide

Micropipes are the primary macroscopic defects in silicon carbide single crystals. It is shown that a stable hollow core of dislocation can act as an initial site of micropipe development. The specific strain energy necessary to estimate the thermodynamic stability of a hollow dislocation core is expressed in terms of the macroscopic parameters of the material Young’s modulus and of the Gruneisen constant.Micropipes are the primary macroscopic defects in silicon carbide single crystals. It is shown that a stable hollow core of dislocation can act as an initial site of micropipe development. The specific strain energy necessary to estimate the thermodynamic stability of a hollow dislocation core is expressed in terms of the macroscopic parameters of the material Young’s modulus and of the Gruneisen constant.

Graphitization of the Seeding Surface during the Heating Stage of SiC PVT Bulk Growth

A transient numerical simulation of the temperature field distri bu ion in a conventional resistively heated SiC PVT growth reactor revealed that the uni ntended seeding substrate sublimation typically exists during the furnace heat up stage. Conseque ntly, this would lead to the seeding surface graphitization due to nonstoichiometry of SiC evaporat ion. Suppression of the seeding surface graphitization requires significant reduction of the furnace heat up time. An optimal elevation of the argon partial pressure and seeding substrate tem p rature during the heat up stage would also promote repression of the seed graphitization. Introduction The elimination of silicon and carbon second phase inclusions (bulk defects) that may serve as new micropipe generation centers [1] is essential for high quality SiC bulk growth. While the silicon second phase formation at the growing surface of SiC was previously disc ussed and has been experimentally proven [2], the mechanisms of carbon second phase gener ation is not completely understood. One of the possible mechanisms being proposed here may result f rom unintended seeding substrate sublimation (thermal etching), which typically exists during the heating stage of SiC PVT bulk growth. Model description Transient numerical simulation of the temperature field distributi on during the initial (furnace heat up) stage and steady state phase of growth run i n a conventional resistively heated PVT reactor was performed using GAMBIT-1.3.1/FIDAP-8.6 software pa ckage. The numerical model incorporated the axi-symmetric geometry of the calculati on domain, time dependence of the heating power rise at the growth beginning and the temperature de pendent physical properties of the seeding substrate, SiC source material, argon/Si xCy vapor mixture and the materials used in the furnace design. Conduction in the solid elements of the calculation domai n and radiation along with conduction in the fluid (gaseous) regions represent the heat transfer mechanisms incorporated in the model, whereas the convective component of heat transfer was omitted [3, 4]. Results and Discussion According to the simulation results, the heating up rate of the sourc e material region significantly differs from that of the seeding substrate (see Fig. 1). Such a situation is primarily attributed to the large difference in the thermal conductivities of SiC source material λP and the seeding substrate λC (λC/λP>100 at 2500 K) [5, 6]. This difference determines the typical temperature rise dynamics of the source material and the grow in crystal shown in Fig. 2. As is clear from the plot, the furnace heat up stage continues ~3 hr afte r the heating power switch on. At the beginning of this stage 0 < t < tB the major part of the source material region is not heated up yet. In fact, the temperature inside the source material volume (point 3 in Fig.1) is significantly Materials Science Forum Online: 2003-09-15 ISSN: 1662-9752, Vols. 433-436, pp 99-102 doi:10.4028/www.scientific.net/MSF.433-436.99

Formation of Stacking Faults in Diffused SiC p+/n-/n+ and p+/p-/n+ Diodes

The degradation of diffused SiC PIN diodes during forward-biased operation has been studied in this work. The PIN diodes were formed by diffusion of aluminum and boron into 4H-SiC substrates with n-type 10 μm thick epilayers doped by nitrogen up to 5x10 cm or p-type 12 μm thick epilayers unintentionally doped by boron up to 1x10 cm. The forward voltage drop, Vf, and the emitted electroluminescence spectra vs. time were measured simultaneously during the degradation test under an applied current density of 100 A/cm at various temperatures. Before the onset of diode degradation, the electroluminescence spectra measured on the formed diode structures exhibited one EL band with a peak at the wavelength of 486 nm for p/n/n diodes and at 510 nm for p/p/n diodes. However, a new EL band with a peak at the wavelength of 426 nm appeared as Vf increased in both types of the formed structures. Moreover, the intensity of this peak increased steadily with the increase in Vf .

Investigations to simulate the high field characteristics of gate to cathode gaps in FEDs

In FEDs, a micro-vacuum structure is used to address each pixel. The typical dimensions of such a structure are in the micron/submicron range. A high electric field is present at the tip of the emitter to produce field emission electrons sufficient to light up the pixel on the phosphor screen. The electrical breakdown due to high fields between the gate and emitter or the gate and substrate can cause catastrophic damage in a FED. In order to investigate these issues, an experimental system, in which microtip type gaps were set-up in the micron/submicron regime, was developed. This paper addresses the high field characteristics of these microtip gaps corresponding to gap distance, vacuum pressure and conditioning.