340 PHENOMENA, ATOMS, AND MOLECULES 



obtained which fitted Eq. (lo) down to 6 = 0.05. Below this i/Ba deviated 

 as shown at X in Fig. 9. Such a deviation means that the observed values 

 of Va were less than those calculated by Eq. (6) when using values of Ba 

 from Eq. (10). At such low surface concentrations, repulsive forces 

 between adatoms should begin to be inappreciable so that the equation of 

 state of the adsorbed film on a homogeneous tungsten surface should ap- 

 proach that of an ideal 2-dimensional gas, vi:2., F = okT. This would mean 

 that the heat of evaporation measured by Ba should change very little with 

 6 for these small values of 6. It seems possible to account for the increas- 

 mgly rapid change in Ba at low 6 as shown by curve X, only if the tungsten 

 surface is not completely homogeneous. By trial it was found that this 

 difficulty disappeared if it was assumed that 0.5 percent of the tungsten 

 surface holds caesium so much more firmly than the rest that this active 

 surface becomes saturated before more than 0.5 percent of the remaining 

 surface is occupied. Thus the total concentration could be expressed as 



Oobs = Oa-\-On = 0Al(6a-{-Bn), (n) 



where the subscripts refer to the active and normal parts of the surface. 

 When the active surface becomes saturated, so that 6a = 0.005, 



^» = ^obs — 0.005. (12) 



Calculating Vg, or i/Ba from Eqs. (6) and (10) with 6n instead of ^obs gave 

 agreement with experiment down to values of 6n as low as o.oi, as shown 

 by the points that lie along the straight line in Fig. 9. Any deviations which 

 existed below this concentration were to be attributed to lack of saturation 

 of the active surface, i.e., when ^a< 0.005. 



The validity of Eqs. (6), (9) and (10) and likewise the precision of 

 the experimental determinations of 6, is shown by the fact that the experi- 

 mental points in Fig. 8 have an average deviation from the calculated curve 

 of about 1° in T with no deviation greater than 3° up to ^ cH 0.6 ; at 

 higher ^'s the deviations vary from 2 to 15°. 



For convenience in calculation, Eq. (6) may be written with common 

 logs as 



logio Va = A„ - Ba/r (13) 



where A^ and B^ in heavy faced type represent the values of Aa and B„ 

 (from Eqs. (7) and (10)) divided by 2.303. Table I in the first two 

 columns contains values of A^ and Bo. Up to ^ = 0.6 the tabulated values 

 of Ba were calculated from Eq. (10). At higher ^'s the experimental 

 variation of B^ with 6 was used as determined from Fig. 9. Complete 

 explanation of the use of Table I will be given in Section X after electron 

 and ion emission have also been discussed. 



