362 



BELL SYSTEM TECHNICAL JOURNAL 



fact the two sets of curves become more comparable as expressed in 

 terms of the time scale until after 400 hours the two distributions of 

 spectral response are nearly identical. This final identity may be 

 interpreted in view of the fact that position 2 has received roughly 

 30 atomic layers of cesium. For such a thickness the vapor pressure 

 would be expected to approach that of bulk caesium. At this temper- 

 ature the rate of vaporization from a bulk caesium surface would 



LAYERS OF CAESIUM DEPOSITED 



1 — rm 



2 5 10 



HOURS OF DEPOSITION 



20 



"1 — I I I I 1 1 1 \ \ r 



50 100 200 400 



Fig. 16 — -Spectral response changes in cell F, position 2. 



exceed the rate of deposition at this point from the orifice. So caesium 

 should obviously evaporate from position 2 and condense on the less 

 saturated areas such as position 1. Eventually the thickness of the 

 deposited layer should be equalized over the entire surface. 



The same qualitative description covers the phenomena observed 

 at the other two positions of the same cell which are not shown. 

 The rate of deposition was slow at position 3 and rapid at position 4. 

 The behavior of all four positions fits into a coherent scheme where 

 the initial sections of the four groups of curves are similar with respect 



