EVAPORATION OF ATOMS 



331 



up to ^=1. If the filament or electrode can be cooled sufficiently below 

 the temperature of the tube, the properties of polyatomic films may be 

 studied. 



(d) The two filament method was used in the present investigation 

 to study 6 from 0.05 to unity. It may be used to study even polyatomic films 

 which may be formed by maintaining the filament for a short time in 

 saturated Cs vapor or by producing supersaturated vapor by rapidly heat- 

 ing the bulb after the filament has cooled. 



Calibrations 



The validity and accuracy of the various methods for determining 6 

 were tested chiefly by observation of transient effects, i.e., phenomena in- 

 volving changes in 6 with time. In these tests it was assumed first that a 

 is unity under all experimental conditions and second that the positive ion 

 current from a clean hot tungsten filament is an exact measure of [Xa- Ex- 

 perimental proof of both assumptions will be given in Section XI. 



If a clean tungsten filament is allowed to coat with caesium for the 

 time t in a retarding field for ions (Vp = o), the direct flashing method, if 

 valid, should give measurements from which the number of atoms ac- 

 cumulating per cm- in the time t can be calculated. This number should be 



V^aXt. 



Fig. 4 shows data obtained by the direct flashing method with the fila- 



Fig. 4. Accumulation of caesium (concentration a a) with time as obtained by the 

 direct flashing method. Filament at room temperature. 



ment at room temperature during the accumulation period. The ordinates 

 are o^ (atoms cm"-) as calculated from the ballistic galvanometer kicks. 

 If Q is the quantity of electricity corresponding to the observed kick, 

 Oa = Q/SaC atoms cm"-, where e is the electron charge. The galvanometer 

 was calibrated ballistically by discharge of a standard condenser charged to 

 a known voltage. The sensitivity was 3.3 X 10"^® coulombs mm"^. The 



