352 



PHENOMENA, ATOMS, AND MOLECULES 



log Vp vs. voltage curve was not as readily determined. For this reason a 

 slightly different method was also employed. The voltage required to keep 

 Vp constant as the filament temperature was varied was found to change 

 rapidly as the zero of potential was approached, thus allowing a closer 

 estimate of Vp at zero field, by choosing the point of greatest curvature. 

 Fig. 17. Since ^la = Va + Vp, 6 could be found for any value of Vp from the 

 known relation between Va and 6. To compare with the theoretical value 

 of Vp, Eq. (20) was used together with In Vp obs to calculate the contact 

 potential Vc- The points in Fig. 14 in the region ^ = O.io to 0.14 were so 

 calculated. The disagreement is not greater than might correspond to errors 

 in obtaining Vp at zero field. 



Fig. 17. Temperature voltage data for constant Vp. Curves used to determine Vp 



at zero field. 



In connection with the study of positive ion evaporation rates as a 

 function of 6, the following general characteristics of positive ion evapora- 

 tion may be given. 



Fig. 18 shows the exponential increase of Vp with temperature at con- 

 stant values of KoC = Va + Vp). Depending on the pressure, a discontinuity 

 sets in at a fairly definite critical temperature (which increases with [la) 

 and Vp rises at a constant rate until the value characteristic of a clean 

 tungsten filament is reached. The variation of this maximum value of \p 

 with temperature and the external field is discussed in Section XI. The 

 discontinuity has been observed previously by Langmuir and Kingdon,^ 



