EVAPORATION OF ATOMS 345 



Likewise the rate of ion evaporation, Vp, was calculable ^^ with the aid 

 of the Saha equation and was given by 



ln(2Vp) = \nVa^ie/kT){V^ - V^- Vo) (20) 



where Vy, is the electron affinity of pure tungsten 4.62 volts, Vi is the 

 ionizing potential of the caesium atoms 3.874 volts, and Vc is the contact 

 potential as defined above. Thus from data on neutral atom evaporation 

 (Vo, e, T) it was possible to calculate M, Vc, Ve, and Vp for comparison with 

 the experimental values of these quantities. Further details and a tabulation 

 of the calculated values of F, M, Vc, and other quantities included in the 

 theoretical equations are given in reference 2. 



VIII. CONTACT POTENTIAL AND ELECTRON EMISSION 



(EXPERIMENTAL) 



The electron emission, Vg, was measured from filament A at various 

 filament temperatures and pressures of Cs. The relation between and 

 pressure being known, the 6 corresponding to each emission was also 

 known. As a check in part of the runs, 6 was determined immediately 

 following the measurement of Ve, by flashing to filament B (2-filament 

 method). 



The values of Ve involved in the calculation of contact potentials from 

 the Boltzmann Eq. (19) and for use in equations relating Vg to v^ and Vp, 

 must correspond to thermodynamic equilibrium and hence were measured 

 at zero field. The effect of the external field on the electron emission from 

 Cs coated tungsten is in general larger than for pure tungsten and varies 

 with 6. Current-voltage data were taken for a series of constant tem- 

 peratures and constant values of Cs pressure, corresponding to values of 6 

 from 0.16 to 0.80. As shown in Fig. 11 there were sharp breaks in the 

 plotted data and the value of Vg at the break was taken to represent Vc at 

 zero field. The great variation of the efifect with 6 is readily seen. Fig. 12 

 shows that at high and low values of 6 the variation of Vg with voltage 

 approaches that for clean tungsten. The largest departure is near 6 — 0.55 

 and decreases rapidly beyond a ^ of about 0.65. The slopes below ^'s of 0.30 

 were not measured accurately enough to show how rapidly the behavior of 

 clean tungsten was approached at low 6. It should also be noted (Fig. 13) 

 that the zero-field emission as obtained in this way may be far lower than 

 that value obtained by extrapolating the normal higher voltage range 

 Schottky slope to zero field as Dushman has done.^- 



To compare these electron emission data with the values calculated from 



*^ See reference 2. 



*^ S. Dushman, G. E. Rev., 26, 157 (1923). 



