478 



BELL SYSTEM TECHNICAL JOURNAL 



is approximately constant at a little less than 100. Values of V ,JI in the 

 forward direction are much larger, starting at 300 for / = 0.1 ma and 

 decreasing to <M at 1=2 ma. If anything, one would expect a decrease 

 rather than an increase in Vp/I in the forward direction as injection of 

 holes lowers the resistivity of the germanium in the vicinity of the point. 

 We shall show that Vi/I actually does decrease and that the anomalously 

 high values of Vp/I in the forward direction result from the drop, Vc/, 



HOLES INJECTED 



FLOATING POTENTIAL, Vnr, 



AND CONDUCTANCE, Y, 



MEASURED 



Fig. 4. — Schematic diagram of experiment of W. H. Brattain for measuring floating 

 potential and admittance at point near emitter. 



across the barrier layer between the contact point and the body of the 

 germanium. Thus, 



Vi = F, 



v.. 



(17) 



Values of Vcj can be estimated from the change in conductance corre- 

 sponding to small currents in the probe point. The conductance increases 

 with increasing forward emitter current. Values of Vc/, calculated from 



Vcf = 2.5 (kT/e) log (Gp/Go), 



(18) 



are given in column 6. The value 2.5, chosen empirically to give reason- 

 able values of V'i, is not far from the value 2.0 required to fit Pearson's 

 data in Fig. 1. Values of F, obtained from Eq. (17) are given in column 7. 

 The ratios Vi/I given in column 8 are reasonable. The decrease in Vi/I 

 with increasing forward current is caused by a decrease in the resistivity 

 of the germanium resulting from hole injection. 



In another case, in which no such anomaly was observed in the for- 

 ward direction, it was found that 1',:/, calculated from the change in 

 conductance, was small comyxired with V,,. 



'i'here have as yet been no measurements which permit a comparison 

 of the values of IS rc(|uired to correlate probe [)()tential and conductance 



