Electrical breakdown electronic models

Electric breakdown of growing alumina films limits their maximum attainable thickness. It also causes degradation of thin film electronic devices with alumina films. Hence, it is a subject of intensive research and a large number of papers have been published and further reviewed. A good deal of controversy exists on various aspects of alumina oxide breakdown, and a variety of models have been proposed in attempts to fit experimental findings. 

Other models are based on electric breakdown of the oxide [Fo2, Chl2]. It is not clear whether this breakdown should be thought of in terms of an electronic or an ionic effect. However, in both cases breakdown may cause a degradation in the oxide morphology, which leads to an enhanced etch rate. An electric field strength in the order of 10 MV cm4, the observation of an electroluminescent burst associated with the current peak of the oscillation, and the presence of an electronic component in the interface current are in favor of this model [CalO, Chl2]. 

There are two principal views on dielectric breakdown generation. The first favors the cavitation-bubble mechanism, and the second involves electronic phenomena occurring in the studied system. The major differences between the bubble and electronic models of electrical breakdown lie in the importance attached to the temporal development of events which precede a spark, i.e., a moment considered as the breakdown. In the former case, it is proposed that ionization and current growth begin to occur in the gaseous phase after nucleation of a bubble, whereas, in the latter case, these processes begin first in the liquid areas. Among these two schools several models based on a different approach to the source of the increased conductivity of Uquids under electrical stress have been proposed. 

Another bubble theory of electrical breakdown in liquids has been developed by Krasucki. In Krasuckis model, the presence of impurity particles, produced by sparkerosion of the electrodes, was invoked as a prerequisite for the creating bubblecavity. Vaporization can occur in a liquid wherever a point of zero pressure is developed, and in this way a vacuous cavity can begin to form. Electron bombardment of the cavity walls will sustain its growth, eventually leading to a breakdown. Solid-particle impurities in the liquid, but especially at an electrode-liquid interface, are considered as suitable sites for cavity nucleation. 

The formation of electronic models of electrical breakdown by various authors is based on their different views of the mechanism of electronic injection from the cathode into the solution, and on distinct ideas as to what processes electrons undergo while in solution. Lewis, on the basis of observation of dielectric breakdown in -alkanes, has proposed a model based on an increase in electronic mobility leading to breakdown. Wong and Forster, on the basis of results obtained by the ultrahigh-speed laser schlieren system, have proposed" a model based on observation of the formation of conductive channel columns bridging cathode and anode. An electrochemical explanation has been offered on the basis of the influence of the electrode material on the potential of electrical breakdown in water. 

This model raises the question of the dependence of the electrical breakdown potential on the electronic work function of the electrodes. At a first approximation one could conclude that a lower breakdown voltage will be observed for the cathodes with lower electronic work function (less voltage is required to bring the Fermi level to the conduction band of solvent cf. Fig. 31), but this is not true (Fig. 30). To explain this discrepancy, we recall the effect of the rate of the electrochemical reaction on the breakdown potential (Fig. 29), which is the reverse of the changes in the electronic work function of the electrode material. The changes... 

The surface films discussed in this section reach a steady state when they are thick enough to stop electron transport. Hence, as the surface films become electrically insulating, the active electrodes reach passivation. In the case of monovalent ions such as lithium, the surface films formed in Li salt solutions (or on Li metal) can conduct Li-ions, and hence, behave in general as a solid electrolyte interphase (the SEI model ). See the basic equations 1-7 related to ion transport through surface films in section la above. The potentiodynamics of SEI electrodes such as Li or Li-C may be characterized by a Tafel-like behavior at a high electrical field and by an Ohmic behavior at the low electrical field. The non-uniform structure of the surface films leads to a non-uniform current distribution, and thereby, Li dissolution from Li electrodes may be characterized by cracks, and Li deposition may be dendritic. The morphology of these processes, directed by the surface films, is dealt with later in this chapter. When bivalent active metals are involved, their surface films cannot conduct the bivalent ions. Thereby, Mg or Ca deposition is impossible in most of the commonly used polar aprotic electrolyte solutions. Mg or Ca dissolution occurs at very high over potentials in which the surface films are broken. Hence, dissolution of multivalent active metals occurs via a breakdown and repair of the surface films. 

Latham s model (a similar one has been proposed by Kao, 1984) requires little, if any, modification if it is applied to electron emission into a pure hydrocarbon liquid. Since the conduction band states at an energy are close to the vacuum level, the oxide outer barrier X will be hardly altered. Moreover, the electric fields at breakdown in a liquid are generally much greater than those suggested by Latham as necessary for barrier lowering and the relative permittivities of oxide and liquid will be more favorable than oxide and vacuum. There is a further factor of considerable significance, namely that in the liquid case it will be possible for positive ions or holes from the liquid to be fed back to the oxide surface. Moreover, the energy level of this band of states will be about the same as that of the valence band of the oxide. 

---