In order to quantitatively interpret the previous experimental single electron emission rates (SEERS) obtained by Guile et al (1997), the Richardson/Dushman-Schottky (1914) expression is extended by invoking the concept of multiple images of the emitted electron in the naturally occurring or grown oxide layer covering the metal cathode, first proposed by Silvester (1968) in another application. This leads to a new expression for the Schottky barrier lowering term being obtained, involving a converging infinite series of terms, solvable on a computer, and enables the SEERS to be computed for cathodes covered with varying thicknesses of oxide. Numerical agreement between applied theory and experiment is excellent and requires the postulated presence of electron multiplication-avalanching-in the oxide layer. This method can be further applied to the presence of liquid and/or solid dielectric insulation between oxide covered metal electrodes and can take into account the presence of the anode for small electrode spacings.

By means of a newly developed statistical avalanche counting technique which enables high-field, single electron emission rates (SEERS) to be measured from unsparked freshly prepared surfaces having various thicknesses of oxide film, it is established for the first time that SEERS increases, for a given electric field, with an increase in the bulk modulus (hardness) of the material cathodes; also remarkably the thicker the oxide layer, the higher the SEERS. These trends are reflected in the Paschen sparking characteristics which show that the gaseous electric strength decreases with increase in the bulk modulus and increase of oxide layer thickness of the cathode material; and thus the increase in SEERS. This former trend is opposite to that established more than forty years ago. Other hitherto unobserved phenomena like a gas pressure effect, switching from a low to a high electron emission mode and vice versa, the influence of deliberate sparking and the total absence of spark conditioning are reported. Analysis of SEERS by the more successful Richardson-Schottky rather than the Fowler-Nordheim equation shows that the low electric field intensification factors derived increase with the bulk modulus of the cathode material and ultimately are determined by the polishing procedure.

Conduction in Nb/Nb2O5/In structures is strongly thickness dependent. Both present and previous data are interpreted by the model of an overlap MIM structure. In this model, conduction is through space-charge-limited flow of electrons injected from the metal electrode but the electrodes change the electron density in the oxide considerably due to the overlap of the two contact space-charge regions. Die Leitfahigkeit in Nb/Nb2O5/In-Strukturen ist stark dickenabhängig. Sowohl die angeführten als auch frühere Werte werden mit dem Modell einer überlappenden MIM-Struktur interpretiert. In diesem Modell wird die Leitfähigkeit durch raumladungsbegrenzten Fluß von Elektronen hervorgerufen, die aus der Metallelektrode injiziert werden, jedoch ändern die Elektroden die Elektronendichte im Oxid infolge der Überlappung der beiden Kontaktraumladungsbereiche beträchtlich.

It is shown that contact overlap in a MIM structure with distributed traps depends on the trap density per unit energy and that the conductivity of the structure depends strongly on the insulator thickness over a certain range which is determined by a characteristic length. The variation of the conductivity with thickness can be stronger than that of a structure with a shallow discrete trap level. Es wird gezeigt, daß die Kontaktüberlappung in einer MIM-Struktur mit verteilten Haftstellen von der Haftstellendichte pro Energieeinheit abhängt und daß die Leitfähigkeit der Struktur über einen gewissen Bereich, der durch eine charakteristische Länge bestimmt ist, stark von der Isolatordichte abhängt. Die Änderung der Leitfähigkeit mit der Dicke kann größer sein als die einer Struktur mit einem diskreten flachen Haftstellenniveau.

A new, nondestructive method for electron field emission measurement is presented, which, in principle, completely avoids the need for deleterious multiple sparking of the surface required by previous workers and which allows electron field emission to be measured at its origin; i.e. with only a few electrons leaving the cathodes per second. The results obtained on oxide-covered copper cathodes with various controlled thicknesses differ markedly from those obtained by previous workers and conform more with modern interpretation.

The effect of the ohmic contacts on the current-voltage characteristics of a thin unipolar metalinsulator-metal structure is presented. It is shown that the conductivity of such MIM structures is a function of contact overlap in which space-charge layers in the contact regions extend into the insulator from both contacts. It strongly depends on the insulator thickness over a certain range which is determined by a characteristic length. Es wird der Einfluß der Ohmschen Kontakte auf die Strom-Spannungscharakteristiken von dünnen unipolaren Metall-Isolator-Metall-Strukturen untersucht. Es wird gezeigt, daß die Leitfähigkeit von derartigen MIM-Strukturen eine Funktion der Kontaktüberlappung ist, indem Raumladungsschichten in den Kontaktbereichen sich von beiden Kontakten in den Isolator erstrecken. Sie hängt in einem bestimmten Bereich, der durch eine charakteristische Länge bestimmt ist, stark von der Isolatordicke ab.

The paper reviews some of the most significant developments made since 1970 in the understanding, observation and control of the phenomena which occur in the cathode and anode regions of arcs under a number of conditions, and in the engineering applications in which they are used or arise. Even with this limitation to less than 14 years, the literature on arcs is so vast that selectivity has been necessary. Those aspects which have been chosen here for a review of the work, which has been done by many investigators attempting to obtain a more complete understanding of the processes involved, are nonrefractory cathodes at gas pressures near atmospheric, anodes near atmospheric pressure, vacuum arcs, unipolar arcs and hollow-cathode arcs. The sections on each of these aspects are followed by the main body of the paper in which a large number of engineering applications are discussed, in which arcs are used as part of the process, or where they occur and must be controlled and suppressed.

The core of large turbo-type generators is made up of cold-reduced low-loss silicon steel laminations, insulated by several varnish layers and located within a frame by longitudinal brass keybars, and held in compression by means of core end plates. From time to time the insulation is subjected to appreciable voltage stresses, highest near the ends of the core where leakage fluxes are greatest. Localised breakdowns of insulation between two adjacent laminations may then occur and increase eddy current losses. Direct monitoring of interlaminar voltages is difficult (1) and hence detailed examination of individual laminations on machines is the main source of information for the understanding of breakdown mechanisms (2,3). During such an examination, examples of a form of interlaminar breakdown not previously reported were found, and described in a previous paper (3). Damage revealed by scanning electron microscopy (SEM) was compared with that caused by arc discharges, but there were features of wormlike metallic filaments which could not be explained.

Twelve different metals, having oxide films of a few nanometres, were used as the cathode for a single 4 A arc in air at atmospheric pressure, about 30 ?m long and of duration between 3.8 ns and 3.5 ?s, and the surface was then examined by scanning electron microscopy. Three groups of metals were distinguished; those having semiconducting oxide leading to scattered craters; those with insulating oxide leading to dense cratering and considerable melting; and those with volatile oxide leading to little or no cratering. The growth with increasign arcing time of the circular areas containing these craters was measured and found to be remarkably similar, and, in the light of this, various aspects are examined. A number of simple models of the cathode region for these arcs are outlined. These include an examination of the gaseous and solid-state regions of the cathode fall and mechanisms of producing the high electric field required within the oxide, for electron emission by positive-ion charging or plasma jet biasing, and possible effects of microinclusions. Different assumptions about the current-density distribution are examined. It is concluded that positive-ion bombardment is likely to cause electron emission for aluminium cathodes, but that further work is needed for other metals.

An effective energy of activation for arc erosion has been deduced from mass loss rates observed on oxidised copper cathode surfaces exposed to arcs of long duration rotating in atmospheric pressure air. An analysis of rate processes expected to be associated with the observed low value for the effective activation energy excludes the involvement of ion migration or the filamentation of pure copper across the oxide layer in the cathode erosion process in an oxidising medium. Independent confirmation is obtained from a comparison of mass loss rates on oxidised cathodes differing in metallic composition, but subject to common operating conditions for pulsed stationary arcs in air. A causative correlation is indicated between erosion rates and properties of the pure metal in the cathodes, rather than with properties of their corresponding oxides. This trend further implies that oxide contributions to electrode erosion occur on timescales which preclude commonly accepted averaging procedures over the lifetime of the arc.

Craters caused by unipolar arcs on stainless steel depend strongly on surface conditions. There are three diameter distributions: Small craters (∼ 0.3 μm) for oxidised surfaces, larger craters (∼ 2 μm) for moderately cleaned areas, and very large craters (∼ 20 μm) for clean surfaces. These values are compatible with Joule heating models of crater formation. Compared with other metals of interest, stainless steel shows the largest craters.

It has previously been shown that when arcs rotate over a copper cathode the erosion rate of this electrode rises and falls as the arc velocity is increased. In the paper it is shown that it is the magnitude of the transverse magnetic field and not the arc velocity, which determines the successive minima and maxima in the erosion rate. The erosion rates of cathodes of a number of materials are examined and for each the factor of increase in magnetic field strength is determined which is sufficient to cause successive minima and maxima in erosion rate. It is suggested that these variations, and similar variations found when the cathode-cooling-water flow rate is increased, arise from related phenomena in the oxide film on the cathode surface.

The most common and maximum radii of craters on cathodes of copper, mild steel, silver, silver cadmium oxide and aluminium having various oxide film thicknesses, are reported following 4.5 A arcing of duration between 3.8 ns and 3.5 μs. The increase in crater radius with arc duration is examined in the light of a simple joule-heating model in a hemispherical region from which material is lost to form a crater. All the results except one conform with adiabatic joule heating, the one exception involving heat conduction into the cathode being no longer negligible. A brief comparison with cratering in cathodes of vacuum arcs is made, showing that these are of the latter type where joule heating provides the dominant energy input and the crater radius tends to increase to a value where this input is balanced by thermal conduction into the bulk metal. Close agreement with the model confirms earlier estimates by other methods of mean emitting site currents, current densities and lifetimes for copper and steel, and then enables these quantities to be estimated for the first time for silver, silver cadmium oxide, aluminium and also for mild steel with a relatively thick oxide film.

Charge storage on aluminium oxide films has been studied using the Kelvin probe technique. The experiments extend previous studies by considering oxide films ranging in thickness from 3 nm up to 240 nm. The band structure of the oxide film deduced from these measurements is discussed and related to the problem of arc root initiation in oxide covered cathodes.

This paper reviews the considerable advances which have been made in recent years in understanding the basic processes involved in erosion of metal cathodes of electric arcs. These advances have been due to the use of scanning electron microscopy (SEM) which has revealed craters of 0.1-¿m diameter, the use of arc durations down to a few nanoseconds, and recognition of the importance of oxide film on the cathode surface and control of this in experiments. The paper covers atmospheric pressure and vacuum arcs and clears up most of the confusion about such parameters as cathode emitting site current densities and lifetimes, at least for copper cathodes and to an extent for some others.

Previous work with high-speed tungsten-inert-gas (t.i.g.) arcs has shown the need for experiments to determine their melting performance. To achieve this, a new apparatus has been built in which the workpiece is of substantial thickness, rather than the earlier tests which used thin shims as the workpiece. The apparatus also has a calorimetric capability, which allows the magnitude of the total heat input to be determined. Data describing the variation of the weld penetration, width and cross-section over a wide range of arcing conditions (current, speed, tip angle, arc gap) have been determined. Additionally, interesting observations have been made concerning the changing distribution of heat input under certain arcing conditions.

Using scanning electron microscopy with magnifications up to 100,000X, the authors have recently established values for the arc cathode emitting site currents, current densities and site lifetimes on copper for a wide range of oxide films, showing that these and other parameters of the craters left by the emitting sites have a minimum or maximum at a particular oxide thickness.

The erosion rates of cathodes of copper for arcing conditions ranging from vacuum to atmospheric pressure are examined in the light of very recent scanning-electron-microscope studies made by the authors. These have revealed details of cathode crater sizes and distributions, surface densities, and the numbers produced per second. These results explain the ways in which erosion rates vary with such parameters as arc current, arc duration, and gas pressure. Static arcs as well as arcs having cathode roots moved magnetically are considered.