SELF

26

and S.B. Karavashkin

We have to emphasise that Mills, thinking the series of short arcs to be the main cause of contact erosion, has not enough cleared their nature and typical features; rather because it is a difficult task to study such short intervals. By the same reason short arcs are little described in the rest of literature on gas discharge. A number of known surveys either do not consider them [6, 7] or give too brief description [8, 9, 10, 11].

R. Holm [9] understands under the short arc a plasma-free arc whose length is much (more than 10 times) longer that the radius of cathode spot. The length of short arc is thought about 10 ^{- 4} cm [12]; this corresponds to the length of free run of electrons in gases at NTP. So V.T. Omelchenko [8] states that "short arc is such type of arc discharge at which the column consists only of the near-cathode fall of potential". And though this conclusion in general has something in common with Holm's statement, Allen [13] who has estimated the length of short arc suggests basically other model. He thinks that in the contact gap after the bridge burst the pressure of metal vapour is several atmospheres, so that the gap length is equal to several lengths of free run of electrons. Such arc, he thinks, will go until the vapour scatters. To the point, speaking of a bridge, Allen minded, as other researchers do, a primary bridge produced in contact partition before the start of showering arc. This point is very important for further consideration. We also should mark that for example Slade [12] premised a possibility to produce several sequential bridges: "With those (small - Authors) inductances the arc is very unstable and bridge can be produced several times". When expressing this supposition, Slade connects the possibility of multiple production of a bridge at the start stage of showering arc just with the arc instability caused by small inductance of the load, which means small stock of the energy. This supposition was based first of all on a simple fact that in the region of small inductance of the load, the results of experiments on mass transfer between the contact pairs were most discrepant with the theory based on an one-fold primary bridge and a chain of short arcs following it. It is important that Slade understood under the term "short bridges" just the horizontal sections between the sections of voltage growth, in full accordance with Mills theory.

On the whole, the mass transfer connected with short arc is caused by its type, anode either cathode. After conventional conceptions [4, 14], when anode arc, the anode erosion follows the heating by electrons accelerated by a field in the contact gap. When cathode arc, the cathode erosion is caused by melting and evaporating metal from the multitude of little spots on it formed due to large density of emission current going through the cathode spot. It is typical that in cathode mass transfer, there on the cathode multitude of separate craters is formed, - probably, this caused such statements. But the arc can be cathode-type only when achieving some upper limit of its length. Up to this limit the short arc always is anode [14]. Although after passing this limit the arc can, none the less, retain its type.

This reason of cathode erosion - large density of current, about 10⁷ A/cm ² [15] - drew much attention of researchers [3, 4, 8, 9, 14, 16]. Most of scientists are in agreement with M.M. Atalla [16] that electric current in the short arc arises on the account of autoelectronic emission from the cathode caused by the external electric field whose strength has to be for this purpose (0,6 - 16)  V/cm [8]. But with it V.T. Omelchenko [8], emphasising a considerable difference in the mechanism of formation and evolution of short arcs in switching and breaking contacts, thinks autoelectronic emission to be the cause of current only in closure arc. We would disagree with this. In [17], Omelchenko analysed three possible causes of the current production for the case of opening contacts: autoelectronic emission, presence of plasma in the arc and excitation of metal atoms present in contact gap after the burst of metal bridge. Omelchenko [17] came to conclusion that two first reasons cannot be 'main suppliers of electrons to the arc'. The first - because of insufficient field strength (we would like to underline it for the aims of further consideration), and second - because there is not a sufficient amount of plasma in the arc.

This last point is discrepant to the very statement of problem. If the density of metal vapour was low (as opposite to Allen's supposition), the length of electron's run grows and its ionisation ability falls; this means, at small density of arc plasma and comparable length of plasma column with the free-run length, the metal atoms excitation will be even less. And even less justified will be the categorical Omelchenko's conclusion: "Cathode is destroyed by excited atoms, and anode - by electrons. Since in collision of electrons with metal atoms the electron loses an essential part of its energy, the anode surface becomes holed well less than in case of merely autoelectronic mechanism of current production. The fact that in the short arc both anode and some less cathode are eroded is in good agreement with the experiment". He is here excessively categorical. Judging by all above reasoning, all three processes to a definite extent participate in contact erosion, and each can dominate under such or other conditions and within definite bounds of causation of such or other process. To confirm, for example R. Holm [9] writes, as we said it above, that most part of time the short arc remains shorter than the free run length, so the collisions with metal atoms are absent. And the energy balance which he carried out showed that electron bombing of anode is, possibly, not the only way of its heating, - this again contradicts to Omelchenko's conclusions.

Speaking of emission of current, we have to say apart about so-called burst emission. The authors [18, 19] gave the following physical substantiation to it: "Different additional processes (e.g., anode heating and evaporation, desorption and absorption of gases and vapours, migration and diffusion of atoms, ion bombing, impact of microparticles etc.) strengthen, in the end, an electric field at the cathode, intensify the autoelectronic emission, and in this way they promote the excitation of burst emission and the transition to the breakdown under less electric fields than in case when these processes were absent".

And though the authors of so-called burst emission investigated the emission currents with the pulse breakdown of contact gaps by nanosecond pulses, their claim "we have established that the fast phase of breakdown developed at constant voltage occurs similarly to pulsing breakdown" requires, this trend also to be taken into account in our investigation of contact gap breakdown. Though, on the other hand, the unclear formulation of so-called burst emission, which the authors confirm: "It is necessary, however, to mark that the study of emission processes and their part in vacuum discharge is far from being complete … Very many unclear issues are still hidden in the very phenomenon of burst emission …", as well as the stipulation of breakdown by the same standard amount of factors, without reasons still really unknown, - such formulation doubts the very question, do we may give special names to the processes that once already have been named.

None the less, the anomalous density of current in short arcs is quite serious issue; it has not only scientific value but is important for practice.