Search

A new optical technique for the determination of spatial distributions of the thickness and the optical constants of thin films is proposed. This technique is simple and nondestructive involving mainly an ellipsometer and a coherent light source, and no contact to the thin-film samples is required. A theoretical estimation indicates that this technique may provide a spatial resolution close to the diffraction-limited resolution under certain conditions. Possible applications of this technique are also discussed.

The liquid motion normally observed in electrically stressed dielectric liquids is mainly caused by two mechanisms: (a) the Coulombic force due to the interaction of the space charge with the field, and (b) the electromechanical force created by the field resulting from the spatial variation of the dielectric constant due to the variation of temperature from domain to domain and from time to time in the liquid. The direction of such a liquid motion depends on which of these two mechanisms is dominant. The schlieren images sometimes observed in liquids under applied fields show mainly the change of the already existing temperature gradients in the liquids, and this change is due to the liquid motion caused by either mechanisms (a) or (b) or both. It is concluded that the schlieren images are directly governed by the temperature distribution in the liquid and have no direct bearing on the formation and distribution of space charges.

On the principle that the spatial distribution of the refractive index produced by a small applied temperature difference is changed due to an electromechanical force created by the field resulting from the spatial variation of the dielectric constant, the electric field distributions in a dielectric system can be determined using the schlieren method. Some experimental results for benzene and n-hexane between two spherical electrodes obtained with this method show the essential features predicted from the space charge effects. This method can be used to study electric field and space charge distributions under both steady and transient conditions.

The shadowgraphic image reflects only those regions of disturbance occurring in a dielectric liquid, which differ in refractive index from their surroundings. The time required for the onset of a prebreakdown disturbance after the application of a step voltage, and the rate of growth of the disturbance depend strongly on the viscosity (or the temperature) of the liquid. The refractive index is lower inside the disturbance region than outside. The variation of the external current is consistent with the temporal change of the disturbance. It is proposed that the formation of a disturbance region is due to the coulombic repulsion of local homocharges which in turn causes a decrease in local density and hence a decrease in refractive index. At highfields the anode injects holes to form positive ions and the cathode injects electrons, and theelectric conduction is filamentary. The electric breakdown in condensed materials (liquids and solids) involves double injection and the creation of low density regions to enable the charge-carrier multiplication processes to take place. Thus the breakdown strength is governed by the parameters which control the conditions for such prebreakdown phenomena to arise.

Taking into account the combined electrode and bulk effects, an expression for the high-field filamentary conduction as a function of applied field and temperature for dielectric liquids has been derived. This expression can explain recent experimental results on the field-dependent thermal activation energy for electric conduction in hydrocarbon liquids.

On the basis of the experimental observation that the conductivity of acetophenone-doped polyethylene is field dependent over a range of applied electric stresses, a simple analysis shows that the increase in resistance to electrical treeing by additives is associated with this field-dependent conductivity. The additives can be considered as weak electrolytes. The rate of their dissociation into positive and negative charge carriers which causes an increase in conductivity is field dependent. A model for the behavior of the additives which leads to the lowering of the electrical stress at sites of high and divergent electrical stresses is given. Presently available experimental results are in good agreement with our proposed theory. Other factors affecting this mechanism are also discussed.

Using the Runge-Kutta and Newton-Raphson methods, the electric fileds on the bushing-type insulator surface covered partly by a corona-suppression semiconductor coating with a field-dependent resistivity have been computed for semiconductors having various resistivityfield ield (p-E) characteristics. The field distribution along the insulator surface is strongly dependent on the p-E characteristic and the extension of the semiconductor coating as well as the frequency of the operating voltages. By assigning the highest field along the insulator surface to be the critical field for the onset of corona discharges (or the breakdown strength of air or other medium), the corona inception voltage of the insulators can be evaluated easily. Thus, by the comparison of the field distribution curves, it is easy to determine the type of the p-E characteristic and the extension of the semiconductor coating for optimal corona suppression. For bushing-type insulators, the computed results show that the semiconductors with p = poexp(-BE) or p * pO[2 - exp(aE)], where po, and a are constants, could give optimal corona suppression. The optimal choice depends on the physical parameters of the insulator, the extension of the semiconductor coating and the frequency of the operating voltages, and therefore the constants pO and or a have to be determined for individual cases. In general, for optimal al corona suppression po should decrease with increasing frequency and length of the insulator surface; and or a should increase with increasing length of the insulator surface.

The physical parameters of the guard gap between the guard and the guarded electrodes of the three-electrode system generally used for dielectric loss measurements can be modelled by a simple T-network. Theoretical analysis shows that under certain conditions these parameters can cause great errors or even make the apparent dissipation factor negative, when measured with a Schering bridge. The sharp edges of the guard and the guarded electrodes may result in the creation of high local fields and give rise to electrical discharges under high voltage conditions. Local discharges near the sharp edges directly modify the physical parameters of the guard gap. Such anomalous phenomena have been experimentally observed and they can be attributed to the change in physical parameters in the guard gap. Experimental results are in good agreement with the theoretical analysis based on the T-network model. Methods for suppressing or minimizing these anomalous effects are also discussed.

Further theoretical analyses show that it is the physical parameters of the guard gap region which give rise to the anomalous phenomena observed in dielectric loss measurements using a three-electrode system. A Wagner or similar guard balance, which is generally used to make the potentials at the guard and the guarded electrodes identical in magnitude and in phase, may reduce the errors caused by the bridge balance conditions, ons, but cannot alter the physical parameters inherent in the guard gap region. An interfacial dielectric layer such as vaseline between the test sample surfaces and the electrodes enhances the anomalous phenomena under certain conditions. The potential along the guard gap surface varies from point to point although the potentials at the guard and the guarded electrodes are equal in magnitude and in phase. It is the nonuniform field distribution in the guard gap region which results in electrical discharges at the sharp edges of the guard and the guarded electrodes. Experimental results are in good agreement wtih the theoretical analyses. Methods for suppressing these anomalous effects also are discussed.

A new theoretical model for high-field electric conduction and breakdown in dielectric liquids based on the theory of filamentary single injection is presented. The model is based on the following postulates. 1) The high-field conduction is mainly due to electronemission from the unavoidable asperities on the metallic cathodesurface and is confined in one or more filamentary paths, in which thecurrent density is much larger than that in other regions. 2) There ispractically no impact ionization avalanche in the liquid prior to breakdown,and, at the field of breakdown strength, there is one filamentarypath in which the current density may reach such a critical value thatthe Joule heating produced in it is sufficient to initiate the thermalinstability in the vicinity of the filament and hence the onset of thebreakdown process. Expressions for the high-field conduction currentand the breakdown strength as functions of temperature, hydrostaticpressure, gap length, molecular structure, and applied voltage waveformhave been derived. The computed results are in good agreementwith currently available experimental results for hydrocarbon dielectricliquids. Previous theories based on a uniform distribution ofcurrent density over the electrode area and on an impact ionization processare also critically examined.

Simple and reasonably accurate methods for determining the composition in a-Si1-xGex films have been proposed. The first one uses the X-ray fluorescence intensity of Ge-Kα line and the weight of the film. The second one uses the X-ray fluorescence intensities of both Ge-Kα line and Si-Kα line. The third one utilizes the signal intensities of Ge and Si ions in ion-micro analysis (IMA). In the latter two methods, a series of a-Si1-xGex films of which compositions have been determined by the first method are used as a standard. Advantages and disadvantages of these methods have been described.The compositions of a-Si1-xGex films prepared by the glow-discharge technique have been determined by the third method. The results agree quite well with the values which are estimated based on the deposition rate.

On the basis of the principle that an applied electric field will cause a change in spatial distribution of the dielectric constant produced by a small applied temperature difference due to the interaction of the gradient of the dielectric constant with the field, the electric field distributions along the interface between a dielectric fluid and a dielectric solid can be measured using a Schlieren technique. Some experimental results for the interfaces between n-hexane and a glass plate surface, and between n-hexane and a semiconducting layer surface obtained with this technique show that the space charge is the dominant factor controlling the field distributions. The other factors, such as the dielectric constants and resistivities of the two materials forming the interface, would play an important role in determining the field distributions if and only if the space charge effect is negligibly small or can be ignored. This technique can be used to study electric field and space charge distributions along fluid-solid dielectric interfaces for practical insulation systems under both steady-state and transient conditions.

The electric strengths and time lags to breakdown in NaC1 single crystals have been measured using flat-topped impulse voltages with and without ultraviolet irradiation. The results show that radiation reduces the mean electric strength of annealed, unannealed, and plastically deformed specimens, but has no significant effect on the mean statistical time lag. Plastic deformation increases and annealing decreases both the electric strength and time lag. An increase in specimen thickness or an increase in overvoltage also causes a decrease in the mean statistical time lag. All observed phenomena can be explained by a lti-multielectron-avalanche mechanism.

The electric strengths of n hexane and transformer oil have been measured at various pressures using linearly rising voltages having rates of rise between 50 kV/cm/?s and 20 MV/ cm/?s. Each measurement was made by applying a single pulse of sufficient magnitude to ensure breakdown in order to avoid the possible influence of prestressing with the normal multipulse technique. For both liquids the electric strengths are strongly dependent on the rate of rise of the applied field and on hydrostatic pressure, but the pressure dependence diminishes with increasing rate of rise of the applied field, and at rates of rise higher than about 4 MV/cm/?s the pressure dependence practically disappears. The results indicate that a certain time may be required for bubbles to form and deform in an electrically stressed liquid and that the bubble mechanism may not be operative in the breakdown process for fast-rising voltages.

In n-hexane, at fields up to 80 kV/cm, the dark conduction obeys Ohm's law, and the gamma-ray induced and ultraviolet photoinjected currents do not saturate for a wide range of temperatures. The induced and injected currents are proportional to the intensity of radiation. It is suggested that the dark conduction is associated with the presence of irremovable impurities and that the small fluctuation of conduction current is attributed to the electrons being irregularly attached to molecules and thermally released during their drift toward the anode. The factors influencing the mobilities of positive and negative charge carriers are discussed.