COMBINED MEASUREMENTS ON ETCHED GERMANIUM SURFACES 1033 



24 — 

 22- 

 20- 

 18 - 

 16 - 

 14 - 



G 12 

 10 

 8 

 6 

 4 

 2 



\ 



\ 



\ 



/ 





J L 



I I I I 



iLUdi 



• 



• 

 • 



J L 



J L 



-50 



-40 



-30 



-20 



■10 







St 



10 



20 



30 



40 



50 



Fig. 4 — Construction of the curve relating AG (surface conductivity, in units 

 of ejupMjcC) and s (surface charge, in units of en,£). 



to differ from that in zero field, and gives rise to the voltage difference 

 (AFs — AFi) shown in Fig. 3. Expressing this difference in terms of the 

 difference (jUeff* — Meff) between the apparent and true effective mobilities 

 in the presence of light (see Section IV), one finds: 



^.f/(l/s) /CuniA 



iMeff* ~ Meff) — 





di: 



2w 



(3) 



where K is the constant of proportionality between (1/s) and the photo- 

 I conductivity signal AVo , and C'unit is the capacity per unit of the ger- 

 manium-gold condenser in the illuminated region, which is 1.27 times 

 [ the parallel-plate formula. From a series of measurements of (^eff* — fJeu) 

 I and AFo it is now possible to obtain S by graphical integration: 



S = 



2/2 



^ unit 



dAV-2 



eni£>/ \/po'C/ \ 2w / J ijl^.u* — Meff 



(4) 



This and the giuplucul method are of course e(iuivalent. it is worth- 

 while emphasizing again that either technitiue depends for its validity 

 I on the fact that the distribution of fast states is unaffected by the gas 

 I changes in the Brattain-Bardeen cycle, as shown in the experiments of 

 jl Brown and Montgomery.^ If, however, the assumption were too far from 

 the truth, the fitting of both slopes in Fig. 4 would be impossible. The only 



