6 



THE BELL SYSTEM TECHNICAL JOURNAL, JANUARY 1953 



be highly dependent on past history of the Ge sample. If however one 

 could measure one or more other properties, such as (Ac.p.)l and Vs , 

 on the same surface at the same time, then one could look for correla- 

 tions between these properties. In this manner one might be able to 

 eliminate the past history as a factor. From previous experiments in the 

 highly variable ambient of room air we knew the order of magnitude 

 of the variation to be expected. At first it was impossible to produce 

 this range of variation under the bell jar. The contact potential always 

 drifted in the direction of a positive extreme, i.e., small total dipole at 

 the Ge surface. The only way found to get a really large change in the 

 opposite direction was to lift the bell jar and expose the sample to room 

 air. These phenomena were finally traced to the presence of negative ions, 

 possible salt ions, in the room air. These were not present in the oxygen 

 and nitrogen supplies we used for creating the ambient in the bell jar. 



It was found that the opposite c.p. extreme could be produced, under 

 the bell jar, by running a spark discharge in dry oxygen as it was flowing 

 into the system. The next step was to cycle the c.p. from one extreme 

 to the other and back again. The procedure was to start with the spark 

 discharge in dry oxygen, change to either wet O2 or wet N2 and to end 

 with dry O2 . The development of this dependable and reproducible 

 cycle was a great aid to the proposed study. Fig. 2 is a plot of contact 

 potential versus time for a single crystal slice D of Ge cut from a melt 

 that was p-type. The surface was prepared by removing some of the 

 Ge with a silicon carbide (180 mesh) blast of approximately ten pounds 

 air pressure. The Ge was then mounted in the bell jar within one-half 

 minute of the "sandblast," and the dry O2 flow started. The c.p. was 

 followed for a few minutes to be sure ever5rthing was working properly. 



0.6 



20 24 26 32 



TIME IN MINUTES 



Fig. 2 — Contact potential cycles for sandblasted sample D. 



