374 PHENOMENA, ATOMS, AND MOLECULES 



follow, if a = I, that just after condensation the concentration is locally 

 raised at the point of condensation. Conversely, an atom can emerge only 

 from regions having locally abnormally high surface concentrations and 

 the act of emergence must bring the local surface concentration back to 

 normal. It thus seems impossible to reconcile the observed simultaneous 

 occurrence of high values of a and of 6 with any mechanism by which an 

 appreciable fraction of the emergent atoms have path origins in the first 

 adsorbed layer, even if we assume a high mobility among the adatoms. 



These difficulties disappear, however, if we postulate that the origins 

 of a large fraction of the emergent paths lie in a second adsorbed layer. 

 We must assume that adatoms in the first layer hop, from time to time, up 

 into a second layer which, however, covers only a very small fraction 

 of the surface. The atoms in this second layer migrate over the surface 

 and may evaporate or may hop back into vacant spaces in the first layer. 

 Since an atom which evaporates from the dilute film of the second layer 

 does not leave a "hole," no difficulty occurs in assuming that all incident 

 atoms condense by the reverse process. 



Covering fraction 62 in second layer 



We have seen that the heat of evaporation (at constant pressure) of 

 caesium adatoms from tungsten, for values of 6 approaching unity, is about 

 41,000 calories per gram atom, which is equivalent at 1.78 electron-volts. 

 This, of course, represents the energy that must be expended in taking an 

 adatom from the first layer of adatoms out to a remote region. 



An atom in a second layer is not in direct contact with the tungsten 

 surface, but is in approximately the same condition as an atom on the 

 surface of metallic caesium. The vapor pressure p of caesium is given 

 (in baryes) by ^ 



logio^ = 10.65-3992/r. (47) 



The heat of evaporation corresponding to this equation is 18,240 calories 

 or 0.79 volt, or only 44 percent of that of adatoms in the first layer. 



The average "evaporation life" of an atom ^^ in the surface of a solid or 

 liquid is given by 



r = {lirmkT)'(n/p, (48) 



where x may be defined by the statement that dt/x is the probability that 

 any surface atom will evaporate in the time dt. Placing Oi = 3.56 X 10^^ 

 we thus find that for Cs atoms in the temperature range from T = 300 to 

 1100°, T is given in seconds by 



iogior=-12.82 + 3840/r. (49) 



^^ See Eq. (15), reference 2. 



