Search

Plasma pressure compaction (P2C) is a novel sintering technique that enables the consolidation of silicon carbide with a nanoscale microstructure at a relatively low temperature. To achieve a high final density with optimized mechanical properties, the effects of various sintering factors pertaining to the temperature-time profile and pressure were characterized. This paper reports a design of experiment approach used to optimize the processing for a 100 nm SiC powder focused on four sintering factors: temperature, time, pressure, and heating rate. Response variables included the density and mechanical properties. A L9 orthogonal array approach that includes the signal-to-noise (S/N) ratio and analysis of variance (ANOVA) was employed to optimize the processing factors. All of the sintering factors have significant effect on the density and mechanical properties. A final density of 98.1% was achieved with a temperature of 1600°C, hold time of 30 min, pressure of 50 MPa, and heating rate of 100°C/min. The hardness reached 18.4 GPa with a fracture toughness of 4.6 MPa√m, and these are comparable to reports from prior studies using higher consolidation temperatures.

The effect of intentionally introduced basal plane dislocations (BPDs) on the electrical properties of SiC p-i-n diodes is examined. Diodes with and without intentionally introduced BPDs are unable to exhibit forward voltage drift during on-state operation for an hour at a current density of 200 A/cm2.

This paper briefly introduces our recent results on high field vacuum insulation relevant to vacuum microelectronics. It addresses the key factors that contribute to the failure of the vacuum gap insulation. Approaches to the development of specific solutions to improve the vacuum gap breakdown voltages are presented. Solutions to alleviate the edge breakdown effect in the thin-film vacuum gap and to inhibit the formation of electrical activity in the spacer triple junction area were proven to be very effective in improving the vacuum insulation performance of vacuum microelectronics. The influence of the presence of an electron beam on the plain-vacuum-gap insulation is also reported.

Selective doping of 4H-SiC by aluminum/boron codiffusion was realized. Room temperature PL spectrum illustrates that deep boron complex is the dominant center for DAP recombination and DAP transition from donor to aluminum and the shallow boron energy level is significantly suppressed.

This paper addresses edge breakdown issues in vacuum gap structures similar to field emission displays, both from a theoretical and practical point of view. Several unique solutions have been developed. After using these technologies, the breakdown voltage of a 750 [micro] m plain-vacuum gap formed by two thin-film electrodes was improved from ~ 16 to ~ 24 kV for the first run, and ~ 20 to 28 kV for the conditioned regime. The mechanism of breakdown in a plain vacuum gap is discussed.

Researchers have been trying to design a vacuum microelectronic device that is capable of withstanding high voltages without breaking down. Prebreakdown current for unconditioned electrodes remain unstable, thus, currents could appear suddenly and increase rapidly even with a very small voltage step-up. This has prompted the exhaustive investigation of prebreakdown and breakdown characteristics of narrow gaps with spherical electrodes. Findings have indicated that surface state stipulates high probability of the breakdown process.

In order to determine the influence of intrinsic material properties of dielectric materials on high field surface flashover characteristics in vacuum, it is important that we minimize or eliminate the influence of the experimental test structure on the high field processes. In this paper we shall examine the flashover characteristics of polycrystalline alumina insulators, using a lateral test structure similar to the one used in semiconductor devices. A significant improvement in surface flashover strengths (~ 200%) was found using a metalized contact electrode system compared with a classical solid electrode system. The results indicate that, using a metalized film cathode, the scatter in the flashover data is significantly reduced compared with the solid electrode system, implying that the electron-dielectric interaction processes that lead to flashover occur in a more predictable fashion. The new metal contact electrode system reveals the dependence of flashover strength on the surface preparation and the grain size of the dielectric material. The average flashover strength increases and the scatter in the flashover data decrease with a decrease in the average grain size of the particles composing the alumina ceramic. Thus the surface flashover processes in vacuum are related to the dielectric surface microstructure, specifically the surface and sub-surface microdamage and grain-boundary defects.

A lateral test structure with metalized film electrodes is shown to have distinct advantages over a conventional solid electrode structure for studying the HV characteristics of solid insulators in vacuum. The paper presents results of streak photography and spectral analysis of surface discharge luminosity, and of preflashover activity using time-coordinated current, luminosity, and imaging of light activity emanating from the surface of the stressed dielectric. The preflashover activity, consisting of bursts of time-correlated current and light emission, is associated with randomly distributed emission spots on the surface. It is proposed that the above observations are related to charge trapping and detrapping processes at localized energy levels associated with insulator surface defects.

The pattern of contact geometry on planar semiconductor structures affects field distribution and the leakage current under high-field excitation. Current filamentation results from local avalanche phenomena at high-stress areas in high fields, where there is concentration of field enhancement in a narrow region producing a high current density. Contact geometries for high-field applications of planar Si structures are discussed, based on experimental and computational results.

Luminescence from dielectric surfaces during preflashover in vacuum was observed under pulse excitation. By preflashover luminescence, we mean light emission under no flashover conditions. Study of light emission with respect to electrode configurations and spacing between the two electrodes shows that preflashover light emission is related to electron emission from the cathode and trapping/detrapping processes on the dielectric surface layer. Spectroscopic observations of the light emission indicate that the preflashover spectrum changes from one voltage shot to another voltage shot of the same amplitude and that the spectral distribution of the light emission depends on the dielectric material used. We believe that the light emission from preflashover is electroluminescence which may be attributed to de-excitation of excited species in the surface layers of the dielectric material.

Bulk breakdown in a majority of the high field silicon-dielectric systems occurs by a process which differs from surface flashover. It is produced because of a defective configuration resulting in the growth of at least one, if not more, large filaments in the material's bulk. Different physical causes of the bulk breakdown and surface flashover breakdown exist, since flashover occurs at the dielectric's interface with the semiconductor.

Two preflashover luminosity phases, one near the voltage peak (phase I) and the other just prior to the onset of voltage collapse (phase II), are found to be associated with the surface flashover of alumina in vacuum. The light emission is also observed under the no-breakdown condition (phase III). A new ancillary experiment, performed using three high-sensitivity photomultiplier tubes (two viewing the dielectric surface head-on and the other viewing edge-on) indicates that the plasma corresponding to phases I and III does not expand away from the insulator surface; whereas the plasma of phase II expands away from that surface. The observed results suggest qualitatively that the phases I and III may be attributed to a solid state process whilst the phase II may be a mixed response of both solid state and gaseous processes.

The surface preflashover phenomena across alumina ceramics in vacuum are related to the surface state of solid insulators and are not dependent on the adsorbed gases on the surface. The energy band model is applied to dielectric vacuum interface to investigate the surface preflashover. The defects on the surface of the ceramics that cause the aggregation of the charges inside the forbidden band regulated the phenomena of the surface preflashover. The formation of localized plasmas, detrapping of electrons and ionization of atoms causes the surface preflashover phenomena.

Analysis of the effects of dielectric and semiconductor properties on the surface flashover of silicon-dielectric systems indicates that the physical flashover mechanism is completely different from semiconductor surface breakdown. This result is applied in several dielectric systems to enhance the efficiency of photoconductive power switches that are controlled by pre-breakdown effects.

Variation in size of images formed in a mirror at various locations on the surface of a polycrystalline dielectric and distortion of mirror images are investigated. The investigation reveals that mirror images are formed due to smooth surfaces having minimum damages. The change in the size of the grain boundary and grains which constitute the electron implanted volume produce changes in mirror images.

Spectroscopic investigations were performed on high purity polycrystalline alumina in vacuum to understand the effect of surface microstructure as altered by mechanical surface treatment, on the optical as well as the electrical characteristics of dielectric surface flashover. HV pulses were applied to unpolished samples as well as to samples subjected to mechanical polishing by using Sic (268 to 16 [Mu] m) and diamond abrasives (15 to 0.25 [Mu] m). The samples finished using diamond paste exhibited lower conditioned voltages and holdoff strength than those prepared with SiC abrasive finish. The spectrum of diamond-finished samples showed early development of molecular bands associated with the dielectric material. This behavior could be due to a larger density of defects developed on the surface during final polishing with diamond paste. There seems to be a correlation between the electrical performance of a sample, the polishing medium used for its surface finish, and the development of some of the spectral lines from flashover. The experimentally-observed phenomena in our studies point to different phases of surface conditioning, in which the dominating factors for the origin of surface flashover seem to depend on the surface state of the insulator.

Nonuniform electric fields > 80 kV/cm, have been measured via a relative optical intensity measurement by exploiting the Kerr effect in transformer oil. Transformer oil was used as a Kerr medium in these measurements. Neither a knowledge of the Kerr coefficient nor the actual value of maximum transmitted intensity are required to obtain the absolute value of the electric fields. Electric fields were measured with ac and dc excitations. Space charges resulting from charge injection have been found in transformer oil at room temperature under electric stresses > 150 kV/cm with ac excitation and 90 kV/cm under dc excitation. The magnitudes of space charges in transformer oil with ac applied voltages varied from 2 to 50 nC/cm cubed. The magnitudes of space charges with dc excitation varied from 60 pC/cm cubed to 10 nC/cm cubed. The results illustrate the limitations of the electro-optic technique for investigations of the interfacial electric fields (gas-solid) with transformer oil as the Kerr medium.

The influence of surface residual stress on the surface flashover behavior of polycrystalline alumina in vacuum is investigated. The samples for the present investigation were prepared by using SiC and diamond abrasives of particle size ranging from 268 to 0.25 micrometer. Two types of polishing procedures, namely forward and reverse sequence polishing, are adopted in the present work. The surface residual stress of the specimens is determined by the two-exposure X-ray diffraction method. The flashover strength and the surface stress of the samples, subjected to forward sequence polishing (i.e., coarse to fine finish), are found to exhibit a random behavior with respect to the surface finish. Interestingly, a linear dependence between flashover strength and surface stress with respect to the surface finish is observed for the samples subjected to reverse sequence polishing (i.e., fine to coarse finish). The observed dependence between the residual stress and the flashover strength is discussed in terms of the high energy defect centers (e.g., deep traps) formed within the forbidden gap of the alumina insulator.

The properties of the prebreakdown and breakdown response of silicon-vacuum systems under HV excitation are presented. The most frequent case of system breakdown by surface flashover is treated. The particular properties of the system response in the high field pulsed regime demonstrate the essential differences between the silicon-vacuum and solid insulator-vacuum systems. The main ideas of a new physical model of surface flashover in silicon-vacuum systems are presented. The properties of the surface flashover response are discussed in terms of the proposed model. A new concept called 'system surface flashover sensitivity' is introduced for a better understanding of the surface flashover physical process in silicon-vacuum systems.

Surface flashover properties such as the ultimate failure level, i.e., conditioned breakdown voltage, and insulation withstand strength i.e., holdoff voltage of solid insulators, in relation to test procedures, are investigated for the first time. Four types of test procedures namely, single cycle steady step, multiple cycle steady step, multiple shot, and hill-valley test were adopted in the present work. The behavior of the insulators under a given test procedure is explained on the basis of surface electronic defects and adsorbed gaseous atoms or molecules. In general, the test procedures are found to influence significantly the surface flashover properties of large band gap insulators such as polycrystalline alumina and single crystal quartz in vacuum. The steady step (both single and multiple cycle) tests are found to increase the degree of conditioning to a higher level (|is equivalent to~ 25% increase in the case of alumina) during the second day test. However, the samples subjected to these tests displayed a higher degree of deconditioning (|is equivalent to~ 40% decrease in the case of alumina), i.e., poor holdoff strength, during the second and third day test. A simultaneous increase of conditioning and decrease of deconditioning (i.e., higher holdoff strength) is observed when the samples are subjected to new developed test procedures called multiple shot and hill-valley tests. The surface flashover performance of 6 ||micro~meter~ finish alumina specimens (prepared under identical conditions), observed with different test procedures, are also compared and discussed.

Systematic investigations have been made to understand the relationship between the surface finish, the polishing medium, and surface flashover properties of high purity polycrystalline alumina ceramics. The surface of as-machined specimens were subjected to mechanical polishing by using 60 grit (268 ||micro~meter~) to 600 grit (16 ||micro~meter~) SiC abrasive discs and diamond abrasives of 15 to 0.25 ||micro~meter~. The samples polished with the SiC abrasives displayed a linear behavior, with the breakdown strength increasing, with decreasing surface roughness. The breakdown properties of the diamond polished samples were found to be influenced significantly by the polishing sequence adopted. The samples subjected to conventional forward sequence polishing (i.e., polishing from coarser to finer finish) were found to exhibit erratic surface breakdown properties. Contrarily, the breakdown properties of the samples subjected to a new reverse sequence polishing (i.e., polishing from finer to coarser finish) were found to vary linearly with respect to the particle size of the abrasive used in the final polishing. The breakdown strength of these samples were found to increase linearly with decreasing particle size to |is equivalent to~ 9 ||micro~meter~ finish and decrease linearly below 9 ||micro~meter~ finish. The observed results are discussed in terms of surface and subsurface defects that are formed during diamond polishing. The breakdown properties of the samples in relation to the abrasive type are also analyzed. The magnitude of the luminosity in the case of SiC polished samples was found to be higher than the diamond polished samples. Contrarily, the samples displayed an opposite trend in relation to the breakdown current magnitude.

A study on the complexity of the semiconductor-dielectric systems behavior under high electric fields, suggests a new physical model of the pre breakdown and breakdown phenomena in the high-field semiconductor-dielectric systems to be proposed. The model indicates the active behavior of two parts, the semiconductor and ambient dielectric and the main role of the semiconductor in the initiation of the breakdown.

Investigations of the luminous events associated with the breakdown along the surfaces of large band gap insulators such as polycrystalline alumina and monocrystalline quartz in vacuum were made by employing a phototube and a very sensitive photomultiplier tube (PMT). The onset of luminous signals, in general, was found to start much before the onset of voltage collapse during breakdown. The early onset of luminosity may be due to the presence of various defect or trapping centers within the forbidden gap of the insulator. The spectral nature of the light emitted during breakdown was resolved (in terms of the wavelength) by interfacing a monochromator with the PMT. In the case of the monocrystalline quartz specimen, the breakdown luminosity was observed to be associated with the defects located at 2.76 and 1.91 eV below the conduction band edge. The breakdown luminosity of polycrystalline alumina was found to be associated with the defects corresponding to energy levels at approximately 1.91, 2.25, 2.45 and 2.76 eV. The formation of these defects is discussed in terms of the nonstoichiometric nature of the lattice structure at the surface. Besides, the samples were found to emit light when no breakdown occurs. This no-breakdown light emission was found to be attributed to the deep level defects. The observed results suggest that the surface flashover process is primarily controlled by the defect or trapping centers located within the forbidden gap of the insulators.

High-speed streak photography of surface breakdown or flashover events along the surface of monocrystalline quartz and polycrystalline alumina specimens is reported. The initiation of visible luminosity was observed to be a function of surface properties of the samples. It has been shown that light could originate independently at the anode as well as at the cathode, indicating that the initiation of surface flashover could start in any region. The propagation rate of luminosity, bridging the electrodes completely, was found to vary depending upon the location at which the light originated. The initiation of light at the cathode is discussed on the basis of an electron induced collision-ionization process and of radiative recombination of electrons in the conduction band with the empty or positively charged traps or defect centers. The emission of light at the anode end is discussed along the lines of a recombination process involving field ionized traps or defect centers and conduction electrons and of the hole-induced excitation of activator or defect centers by inelastic collision.

Computational analysis, based on the charge simulation method (CSM), is used to predict the electric potential distribution produced by charges implanted within the surface layer of a solid dielectric. The potential distribution thus obtained is used to match the experimental distribution using the scanning electron microscope (SEM) mirror technique. Using an optimization iteration process, a comparison is obtained for quartz, alumina, and yttria. The computational analysis assumes a surface layer with a permittivity higher than that of the bulk. The justification for this assumption is based on the experimental observations that, for a given material, the surface breakdown strength decreases by increasing the density of surface defects by mechanical abrasion. An increase in the surface defect density implies a decrease in surface permittivity. The electric potential distribution produced by a surface with increased surface defects, as determined by the SEM mirror technique, indicates a surface with a higher permittivity. It is also shown that the ability of such a surface to store charges is decreased. The results are consistent with physical models based on band structure (or trapping energy) considerations.