COMBINED MEASUREMENTS ON ETCHED GERMANIUM SURFACES 1031 



The reproducibility of all the data from cycle to cycle was good. One 

 surprising result is that the surface recombination velocity assumed 

 its maximum value close to the "wet nitrogen" extreme for both p-type 

 and n-type. This behavior is quite different from that reported by 

 Brattain and Bardeen/ who found s to be constant within 20 per cent 

 throughout the range and Stephenson and Keyes,* who found a maxi- 

 mum value sometimes at one end, sometimes at the other, and some- 

 times in the middle. There is quite good agreement on the other hand, 

 with the results of Many et al.^^"\ who report a maximum in s near the 

 wet end of the cycle. The result of Brattain and Bardeen is not under- 

 stood at the present time, and is probably wrong. The differences between 

 the present work and that of Stephenson and Keyes may be associated 

 with differences in surface preparation. 



V. ANALYSIS OF THE RESULTS 



From now onwards we shall express all experimental and calculated 

 c[uantities in terms of the following dimensionless ratios: 



Xs = :2s/eni£, S = Ss + Tp - r„ (2) 



AG — AG/enilJLp£, /leff = J"eff/Mp , /"eff* = MeffV^P 



where AG is the surface conductance, £ the Debye length for intrinsic 

 germanium (1.4 X 10~ cm), and /Xp is the mobility for holes (1800 cm v" 

 sec~^). Tables V and VI show values of the quantities we shall need, as 

 functions of the surface potential F, calculated from the theoretical 

 considerations of Garrett and Brattain." The surface conductance, and 

 the differentials in the fifth and sixth columns, are evaluated for 

 8 = 0. 



