SURFACE CHEMISTRY 5 5 



filament at temperatures above 1,300° causes every caesium adatom to 

 escape from the filament as an ion so that the ballistic kick measured with 

 the galvanometer gives directly the value of 0. The second method, which 

 is applicable at any values of (even those corresponding to polyatomic 

 layers), involves the evaporation of the adatoms as atoms, in the presence 

 of a retarding field which prevents the escape of ions. This burst of atoms 

 falls on a parallel neighbouring tungsten filament heated above i ,300° from 

 which these atoms escape as ions. The ballistic kick of current from this 

 second filament thus measures on the first filament. We call this the 

 2-filament method of measuring 0. 



The measurements of showed that as the caesium pressure is increased 

 or the filament temperature lowered, increases to a definite limiting value 

 of oi which is 4.8 X 10^^ atoms per square cm. of apparent filament surface. 

 A study of the crystal habits of tungsten has shown that the surface lattice 

 of surfaces etched by evaporation contains 1.425 X 10^^ atoms per square 

 cm. Since the diameter of the caesium atoms is almost exactly twice that 

 of tungsten, and because of the tendency of the adatoms under the strong 

 forces exerted by the tungsten, to occupy definite elementary spaces on the 

 surface, we conclude that the maximum number of caesium adatoms is 

 one quarter the number of tungsten atoms and therefore the true value of 

 oi is 3.563 X 10^^ atoms per square cm. Comparing this with the observed 

 or apparent value of oi it appears that the true surface of the tungsten fila- 

 ment is 1.347 times the apparent surface, this difference being due to the 

 slight etching of the filament surface by evaporation at high temperatures. 



A study of transient states during which increases or decreases with 

 time has proved conclusively that the phenomenon of condensation or 

 evaporation go on independently of one another, the condensation being 

 dependent on [x, while the evaporation v is strictly a function of and T 

 only, and does not depend upon the way the surface film has been formed. 



The experiments show that within the accuracy of the experiments 

 (about 0.5%) the condensation coefficient a is always equal to unity even 

 when is as great as 0.98. This proves that the incident atoms, which 

 under these conditions must nearly all strike adatoms, can slide around in 

 a second layer until they find positions in the first layer. The experiments 

 prove, however, that the number of atoms which exist at any time in the 

 second layer is extremely minute, of the order of io~". The atoms in the 

 second layer possess very high surface mobility. 



The rate of evaporation of atoms Va increases at any given temperature 

 extremely rapidly as increases. Thus at i,ooo°K an increase of from 

 0.1 to 0.9 causes a lo^^-fold increase in Va. This indicates very strong re- 

 pulsive forces between the adatoms in agreement with the fact that these 

 adatoms tend to be positively charged as indicated by the effect of the film 



