1020 THE BELL SYSTEM TECHNICAL JOURNAL, SEPTEMBER 1956 



in fast traps" can follow a change in the space-charge region very fast 

 in comparison with the light-chopping time used in that work (Koo 

 sec); the other kind, imagined to be more closely connected with ad- 

 sorbed chemical material, can only change rather slowly. In a previous 

 paper by the authors it was pointed out that the Brattain-Bardeen 

 experiments, taken by themselves, do not furnish unambiguous infor- 

 mation concerning the distribution of these "fast" traps, but that such 

 information might be obtained by performing, simultaneously, other 

 measurements on the germanium surface. More recently Brown and 

 Montgomery^' ^ have provided a valuable tool in their studies of large- 

 signal field effect; they point out that if, under given chemical conditions, 

 it is possible to apply a field, normal to the surface, large enough to 

 force the surface potential to the minimum in surface conductivity; 

 then it becomes possible to determine the initial surface potential ab- 

 solutely (provided certain considerations as to the mobility of the 

 carriers near the surface are valid). 



This paper concerns studies of a number of physical properties that 

 depend on the distribution and other characteristics of the surface 

 traps or "fast" states. Measurements are reported of (i) the change 

 of conductivity of a sample with field; (ii) the photoconductivity; 

 (iii) the change of photoconductivity with field; (iv) the filament life- 

 time; and (v) the surface photo-voltage. Measurements were made in a 

 series of gaseous ambients, first described by Brattain and Bardeen. 

 Evidence is presented to the effect that the variation in gas ambient 

 changes only the "slow" states, leaving the distribution and other 

 properties of the traps substantially unaffected. From measurements 

 (i) to (iii) it is possible to construct the whole field-effect curve (con- 

 ductance versus surface charge), even though the fields used were in 

 general not large enough to reach the minimum in conductance. 



Using the field effect data, values for the surface potential Y in units 

 of kT/e could be obtained at each point, and also of the quantity 

 (d'Ls/dY)s=o , where 2s is the charge in surface traps, and the suffix 5 = 

 implies zero illumination. From measurements (ii) and (iv), the sin'face 

 recoml)ination velocity s could be deduced. (A more detailed study of 

 photoconductivity in relation to surface recombination \'elocity will 

 be reported at a later date.) Combined with the field effect data, this 

 enables one to deduce the relation between s and Y. 



Measurements of the surface photo-voltage may be presented in terms 

 of the quantity dY/d8, where 5 is ec|ual to Ap/ui , Ap being the density 

 of added carrier-pairs in the body of the material, and Ui the intrinsic 

 carrier density. The quantity dY/d8 is closely related to the ratio of the 



