SURFACE PROPERTIES OF GERMANIUM 



15 



wash. He found the surface to be quite clean, with a film thickness less 

 than 10 A. He also took pictures after the germanium surface had been 

 cycled about fifteen times and found in this case definite evidence for a 

 thin surface film. The film was either amorphorus or composed of very 

 small crystals. He estimated the thickness to be between 20 and 50 A. 



In the discussion of the experimental work it has been assumed that 

 all the changes in c.p. are to be attributed to the Ge surface and not to 

 the Pt. It is not easy to give a definite proof that this is true. The fact 

 that it is a reasonable assumption is suggested by the nature of the re- 

 sults themselves. Almost identical results were obtained using a Ta 

 electrode. In Fig. 10 we show the results of tw^o cycles when the Ge was 

 replaced wdth gold. While there are some changes they are almost an 

 order of magnitude smaller than the changes when Ge is present. It 

 follows that one would expect much the same results with either Ta, 

 Pt or Au reference electrodes and that most of the changes are due to 

 the Ge. No change of c.p. with illumination is observed when both 

 electrodes are metals. In one case a small amount of H2 was added to 

 the N2 flow. Here the c.p. between Pt and Ge changed rapidly but the 

 (Ac.p.)i, did not change at all. Subsequent runs using the regular cycle 

 indicated that the c.p. scale had been shifted corresponding to a reduc- 

 tion in work function of the Pt. Except for this shift the results were 

 the same and the shift disappeared in about one day. The conclusion 

 was that H2 had little effect on Ge but reacted w^ith the Pt decreasing 

 its work function. This is a good illustration of the powder of this method 

 of measuring more than one property of a semiconductor surface at the 

 same time. If only c.p. had been measured the conclusions would not 

 have been so clear cut. 



If one knew the work function of the Pt electrode then one would 

 know the work function of the Ge. Work functions are all measured in 

 high vacuum. We have not been able to think of a method of determin- 

 ing the work function of any electrode in a gaseous ambient unam- 



0.2 



en 0.1 



CL 

 L)-0.1 



-0.2 



£i4°Ju ^°.4o^ H ±I4^ 





O f> o 



24 28 32 36 40 44 48 52 



10 4 8 12 16 20 



TIME IN MINUTES 



Fig. 10 — Contact potential cycles when germanium is replaced by gold. 



