10 



THE BELL SYSTEM TECHNICAL JOURNAL, JANUARY 1953 



0.80 



< 



a 0.76 



cr 

 wO.72 



0.68 



z 



^0.64 



I- 



z 



UJ 



^:o.60 



0.52 



-ai2 -0.8 -0.4 0.4 0.8 0.12 0.16 0.20 0.24 0.28 



Ef-Et, IN ELECTRON VOLTS 



Fig. 6 — Contact potential dependence on position of the fermi level. 



For variety Fig. 7 shows results on c.p. for a sandblasted polycrystal- 

 line sample of silicon. It is apparent that similar phenomena are taking 

 place on the silicon surface, however the behavior is different. These 

 preliminary results are shown to illustrate the generality of this method 

 for investigating semiconductor surfaces in a controlled gaseous ambient. 



The time between cycles was somewhat variable. The first and second 

 cycles were always taken immediately after the specimen was placed 

 in the bell jar. After this, successive cycles were taken one or two a 

 day over a period of a week or more. The bell jar was not opened during 

 a run and a small flow of dry gas was maintained between cycles. Little 

 if anything occurred during these idle periods, showing that the changes 

 that did take place were due to the cycling. 



//. Change of contact potential with illumination 



When the Ge surface is sandblasted the change of contact potential 

 with illumination (Ac.p.)i, is too small to be measured. If however the 

 surface has been prepared by the polishing etch, the (Ac.p.)z, is easily 

 observed. This change if not too small can be measured by finding the 

 balance on the potentiometer for light off and light on. When the con- 

 tact potential is changing with time, or if the change is small this is 

 difficult to do. A better procedure is to chop the light at a definite 

 frequency and meiusure the amplified output on an ac meter. Tliis L>ives 

 a continuous residing that can be read easily at any given lime. If a 

 filter is used to pass only those frequencies near the fundamental of 



