28 



THE BELL SYSTEM TECHNICAL JOURNAL, JANUARY 1953 



-0.04 



0.04 

 C.R- 



0.08 



•(C.R)c 



0.24 



0.28 



0.12 0.16 0.20 



;0-Vb+Vbo in volts 



Fig. 14 — Plot of 2 sinh /3(7b — Ybo) versus c.p. —(7b — Ybo) for sample D 

 as suggested by theory. 



hY L ' Both positive and negative branches of the curve are plotted on 

 the same figure. The symbols used for the experimental points are 

 consistent with Fig. 14. When c.p. — (c.p.)o is greater than zero, hV l is 

 negative. As c.p. increases it approaches a maximum negative value, 

 this is the negative branch. When c.p. — (c.p.)o is less than zero, hY l is 

 positive and as c.p. decreases hY l approaches a positive maximum 

 value that is less in magnitude than the negative maximum. The solid 

 curve represents the prediction of theory for B. — 0.02 e.v. The agree- 

 ment between theory and experiment is good. It should be emphasized 

 that this fit is obtained with only one adjustable parameter. 



The data for the other samples were analyzed in the same way. 

 The results are shown by plotting hYj^ versus c.p. — (c.p.)o as in Fig. 15. 

 Fig. 16 is for n-type sample A and Fig. 17 for n-type sample B. The 

 values obtained for H were 0.015 and 0.022 e.v. respectively. The fits 

 obtained are about equally good in all cases with some deviation between 

 theory and experiment near the extremes of contact potential. The 

 values of H obtained are all of the same order as they should be if the 

 surface trap structure is approximately the same from sample to sample. 

 When Yb — Ybo is large and positive and thus c.p., Eq. (14), is large, 

 6Vl approaches the negative maximum 



