602 



THE BELL SYSTEM TECHNICAL JOURNAL, MAY 1956 



employed in section V. Hall mobilities of the two specimens were meas- 

 ured^^ down to below 10°K. Cooling was carried out slowly to permit as 

 much relaxation into the paired state as possible (see Section X). Li 

 Fig. 25 plots of the Hall mobilities versus temperature of both specimens 

 are presented. Curve A is for the sample containing 2.8 X lO" cm~* 

 lithium. It therefore contained about 5.8 X 10^^ cm~^ total impurities 

 as compared to the control sample whose curve is shown as B in Figin-e 



25 and which contained only 3 X 10 cm" impurities. 



The lithium doped bridge exhibits by far the higher Hall mobility for 

 holes (except at very low temperatures where poorly understood phe- 

 nomena occur). In fact at 40°K the sample containing lithium shows a 

 hole mobility 16 times greater than that of the control at the correspond- 

 ing temperature. Rough analysis of the relative mobilities at T = 100°K 

 indicate '^2 X 10 cm scattermg centers in the control sample and 5 

 X 10 cm" scattering centers in the sample containing pairs. 



This experiment has been repeated with other specimens doped to 

 different levels with gallium and even with other acceptors, and leaves 

 no doubt that a mechanism which is most reasonably assumed to be 

 pairing, is removing charged impurities from the crystal. 



The phenomenon we have just described suggests an excellent method 

 for testing the ion pairing formula derived in Sections VII and XI, for it 



10' 



Q 



Z 



o 

 u 



o 

 > 



DC 



liJ 

 a 



cvj 

 5 10 

 u 



CD 

 O 



5 



< 



X 



10' 



40 80 120 160 200 240 280 



TEMPERATURE IN DEGREES KELVIN 



320 



Fig. 25 — Plot of Hall mobility as a function of temperature for germanium 

 containing 3 X 10^' cm"' gallium. Curve A is for a sample containing 2.8 X 10''' 

 cm~^ lithium. 



