794 THE BELL SYSTEM TECHNICAL JOURNAL, MAY 1957 



to a scale 200 times greater than the scale of the electrodes. For poten- 

 tials that give air breakdown, the scale of separation would be changed 

 by large factors. 



The sketches of Fig. 8 are of assistance in understanding some of the 

 qualitative erosion differences observed in the four types of arcs (Ta- 

 ble II). All of the metal lost from one of the electrodes in an inactive arc 

 of either type comes from the surface of a pit, and from the figure it 

 seems clear that all of it must obviously be intercepted by the other 

 electrode because there is no way for it to escape. This is true even for 

 the case of air breakdo\vn where the electrode separation is much greater 

 {-^Id X 10~^ cm). But for active arcs some of the metal coming from 

 each electrode is permanently lost and not transferred to the other side, 

 even though the separation is much less than it is for the case of air 

 breakdown. The permanent loss of metal in the case of active arcs is 

 due to the presence of carbon. When there is carbon on the surfaces, the 

 metal simply does not stick. Chemical analyses have been made of the 

 black powder produced by active cathode arcs at palladium surfaces, 

 and these analyses show palladium metal as well as carbon. The pal- 

 ladium metal lost from the electrodes turns up in this black powder 

 rather than at new locations on the electrodes. 



At palladium surfaces, the net loss amounts to most of the eroded 

 metal, but at silver surfaces, most of the metal is transferred. This differ- 

 ence is related to the amount of carbon left on the surfaces. Carbon is 

 found much more abundantly on palladium than on silver, which ac- 

 counts for the failure of eroded metal to stick to palladium. The greater 

 net carbon production on palladium is due to the low efficiency of cathode 

 arcs (at active palladium surfaces) in burning carbon; the anode arcs, 

 which occur in general at active silver surfaces, are more effective in 

 burning off carbon. It is to be presumed that the amount of organic 

 vapor decomposed per unit of energy at a silver surface is not so very 

 much less than that decomposed at a palladium surface, even though 

 the net carbon left on the surface is tremendously less in the case of silver. 



3. RECAPITULATION 



We are ready now to state briefly some of the conclusions about active 

 arcs which have been developed above. All of the observations refer to 

 contacts of palladium or of silver. Less extensive tests upon platinum 

 and upon gold have indicated that platinum behaves the same as pal- 

 ladium, and gold the same as silver. 



An active arc is an arc that strikes between one electrode and car- 



