804 THE BELL SYSTEM TECHNICAL JOURNAL, MAY 1957 



3. Burning off in each arc some of llic carbon formed on the surface- 

 in preceding arcs. 



4. Sputtering and lun-ning off carlxjii from the cathode in a glow dis- 

 charge, when such a discharge occurs. 



The first of the.se effects of air has made itself evident in the experi- 

 ments of Kisliuk, Atalla and Boyle described above. It will not be dis- 

 cussed further. Observations and experiments have been made upon the 

 other three effects of air, and these will be described below. Burning off 

 of carbon in an arc is mentioned first because it can be most nearly sepa- 

 rated from other effects and studied individually. 



6.1 Burning of Carbon 



On a surface uniformly covered by organic molecules, carbon must 

 be burned off on the area covered by an arc, but new carbon can be 

 formed on an annular ring surrounding the arc where the metal tem- 

 perature is lower. As a consequence of this, whereas in the absence 

 of air contacts are activated very much more promptly by high energj^ 

 arcs than by low energy arcs, in air the situation is less simple. The gen- 

 eral result of many experiments is that in air high energy arcs are less 

 efficient in producing activation than are low energy arcs. On the other 

 hand, arcs of extremely low energy are also quite ineffective. There seems 

 to be an optimum arc energy at which contacts can be activated most 

 promptly, w^hich may be of the order of 100 ergs. Activation can be ex- 

 pected to be most prompt when the difference between the area of the 

 arc and the area of the annular ring around the arc is a maximum. 



The outer edge of this annular ring is the position on the metal surface 

 at which the maximum temperature just reaches the decomposition 

 temperature of adsorbed organic molecules, about 600°C for the case of 

 benzene. If the width of the ring is A and its inner radius R, each arc can 

 be assumed to burn carbon from an area irR'^, and to form new carbon 

 on an area 7r[(7? -f AY — R-]. Now A certainly increases with increasing 

 energy (being zero for zero energy), but on the simplifying assinnption 

 that it is independent of energy, the difference area, which is 



A = ir[{R + A)'^ - 2R2] (1) 



will be a maximum for that energy that makes R equal to A. It is in- 

 teresting to find the value R = Ri for a 100 erg arc, which is knoAMi to 

 be very efficient in producing activation, and then to estimate the max- 

 imum temperature reached at the outer edge of the annular ring for A = 

 7^1 . The simple model predicts that this maximum temperature should 

 be 600°C. When the calculation is carried out in rough fashion, the 



I 



