344 BELL SYSTEM TECHNICAL JOURNAL 



of the surface. Also, the temperature and time of heat treatment 

 are to a limited extent compensatory. So a more fundamental 

 specification of the active surface would involve the surface concen- 

 trations of caesium and oxygen, and a degree of interaction which is 

 commensurate with the heating time. » 



The best known thin film phenomena in the past have been of the 

 type of caesium adsorbed on tungsten, a single component film under 

 essentially equilibrium conditions. At least the film is reversibly 

 adsorbed for it may be condensed or evaporated at will by varying 

 the tungsten temperature or the caesium vapor pressure. Caesium on 

 oxidized tungsten is again a two-component film. The caesium is 

 reversibly adsorbed and, after the caesium is removed, it appears 

 that the adsorption of the oxygen is reversible at a considerably 

 higher temperature. 



Langmuir and Villars ^ have published curves showing the thermionic 

 activity of tungsten filaments with varying amounts of adsorbed 

 caesium and oxygen. Due to the lack of an independent measure of 

 the amount of adsorbed oxygen the data have been presented from a 

 different point of view, resulting in a determination of the heat of 

 adsorption of the oxygen. But we may point out the obvious though 

 tacit assumption that the thermionic and other properties of the 

 surface are uniquely determined in terms of the amounts of adsorbed 

 oxygen and caesium (and the temperature) . This is equivalent to the 

 presumption that the filament is under equilibrium conditions. 



In contrast, the two-component film of cesium and oxygen on 

 silver which comprises the photoelectric cathode is formed by essen- 

 tially irreversible processes. No caesium atom which has combined 

 with oxygen can be released, and no oxygen atom split off from the 

 silver surface may be recombined. And the reaction is stopped long 

 before any equilibrium is reached. 



To obtain definite knowledge of the relations in the neighborhood 

 of the optimum conditions we have prepared cathode surfaces under 

 carefully controlled conditions. Two different amounts of caesium 

 were used and the oxygen-caesium ratio (the number of "taps" per 

 milligram of caesium chromate) and the time of hot air heating were 

 systematically varied. The first series of cells was made with pellets 

 containing approximately 5 milligrams of caesium chromate. The 

 oxygen-caesium ratio was varied from 15 to 30, and three groups 

 heated for 15, 30 and 60 minutes respectively at 220-225° C. In 

 each case, as shown in Fig. 3, the activity as measured with a light 

 source at a color temperature of 2,710° K. goes through a maximum 

 ' Langmuir and Villars, J. Amer. Chem. Soc, 53, 486 (1931). 



