EVAPORATION OF ATOMS 373 



Emergent atoms from nearly saturated films 



Let us imagine a nearly saturated adsorbed film (^>— >i) from which 

 adatoms pass as emergent atoms into the gas phase. If this is the only 

 mechanism of evaporation, then condensation can occur only when incident 

 atoms make their first contact with the surface in a vacant elementary space. 

 All other incident atoms must be reflected ; that is, they must escape again 

 by emergent paths. Hopping paths would have to be excluded, for they 

 would provide other opportunities for condensation, and therefore there 

 would be other mechanisms for evaporation than those which we postulated 

 as the only possible ones. 



The probability that a flight terminus shall lie in a vacant space is i — ^. 

 The probability that an incident atom should fly into a vacant space without 

 colliding with adjacent atoms is very much less than i — 6. Thus, on the 

 basis of our assumed mechanism, a would have to be less than i — 6. 



An apparent reflection coefficient approaching unity for incident atoms 

 striking a covered part of the surface may readily be accounted for by an 

 extremely high evaporation rate from the 2nd layer of atoms as compared 

 to that from the ist layer. ^^ It is reasonable to assume, however, that an 

 atom cannot exist, even momentarily, in a 2nd layer unless it can be 

 supported by four underlying adatoms in the ist layer. The chance that 

 a given space is occupied is 6 and the chance that 4 given spaces are occupied 

 is 6^. Thus the probability that an incident atom will evaporate from the 

 2nd layer is 6^ so that the apparent value of a would be i — 6^. Experi- 

 ments with steel balls thrown at random onto a surface partly covered (to 

 the fraction 6) with similar balls, in random arrangement in spaces which 

 form a square lattice, show that the fraction of incident balls which go into 

 a second layer is, in fact, very close to 6^, for values of 6 from 0.2 to i.o. 

 Thus for 6 = 0.85, a would be 0.48 and for 6 = 0.98, a = 0.078. 



With this mechanism for condensation, most of the atoms which 

 evaporate would have to make one or more collisions with the adjacent 

 adatoms before they escape. 



Although this postulated mechanism is probably suitable for the ex- 

 planation of many cases of adsorption of gases on solids, it obviously is in- 

 applicable to the case of caesium films on tungsten and all other cases in 

 which a = I up to high values of ^. 



With high values of and a the emergent atoms must come from a second 

 adsorbed layer 



Since the paths of incident atoms cannot in general be directed towards 

 regions in which the surface concentration is below the average, it must 

 '^^ I. Langmuir, Proc. Nat. Acad. Sci., ?, 141 (1917). 



