18 



THE BELL SYSTEM TECHNICAL JOURNAL, JANUARY 1953 



positions. An average was taken over the filament image, and the ef- 

 fective area of the p-n junction was determined and allowed for. This 

 was done for all light intensities used. Most of the experiments were 

 performed with a fixed intensity and the averaged result for this in- 

 tensity was 6.0 X 10^^, hole electron pairs per cm^ sec. The rate of pair 

 production was found to be proportional to the light intensity. 



Since practically all the light is absorbed in a depth (10""^ cm or less) 

 that is small compared to the diffusion length, it is a simple matter to 

 calculate the steady state increase dp in hole electron pairs due to the 

 light. The relation is 



8p = N/{vs + Vd) 



(1) 



where N is the rate of pair production, Va is the velocity of recombination 

 at the surface and va is the diffusion velocity for the minority carrier. 

 Since N is proportional to light intensity it follows that bp is too. 



The magnitude of (Ac.p.)l should depend on the light intensity. 

 One might at first expect it to be proportional to light intensity. That 

 this is not the case is shown by curve 1 in Fig. 12. Curve 1 is a plot of 

 (Ac.p.)l versus bp for unit D on a log-log scale. A smooth curve has 

 been drawn through the experimental points. As we shall see later, 

 theory predicts that if (Ac.p.)^ is large, it should be proportional to 

 ln{\ 4- bp/a) where a is the equilibrium density of the minority carrier, 



In 



{-%) 



10"' 2 3 4 5 681 2 34568 10 2 34 



10' 



2 3 4 5 6 6 )q12 2 3 4 5 6 6 ^q'^ 2 3 4 



tfp NO. PER CM 3 



Fig. 12 — Dependence of contact potential change with illumination (Ac.p.)^, 

 on light intensity. 



