1022 THE BELL SYSTEM TECHNICAL JOURNAL, SEPTEMBER 1956 



ductance, at zero field, produced by the light. The applied field is suffi- 

 ciently small for the dark field effect and the apparent field effect in the 

 presence of light to be substantially linear. 



(iii) The metal plate, disconnected from the high voltage supply, is 

 connected to a high-impedance detector; chopped light is shone on the 

 germanium, and the change in contact potential produced at the sur- 

 face opposite the metal plate by illumination of the sample measured, 

 and compared with the photoconductivity. 



The interpretation of the field effect data has been given by Brown 

 and Montgomery^' ^ and by the authors." The surface conductance AG 

 is equal to enp{Tp + 6r„),2 where Tp and r„ are surface excesses of holes 

 and electrons, and are, in equilibrium (i.e., in the absence of light) 

 functions of the surface potential Y and of the body type and resistivity. 

 The minimum in the surface conductance curve occurs at a particular 

 value of Y, so that, if a field effect experiment allows passage through 

 this minimum, values of Y may be obtained.* 



In our experiments, measurements were made in a series of different 

 chemical environments, and the minimum in surface conductance did 

 not, in general, occur within the range of field employed. However, it 

 was found to be possible to piece together the complete surface con- 

 ductance curve (AG versus surface charge) by making use of simulta- 

 neous measurements of the photoconductance and the change in photo- 

 conductance with field. (See Section VI.) From the surface conductance 

 curve, one may deduce the fraction of the surface charge (whether in- 

 duced electrically, by application of a field, or chemically, by changing 

 the environment) which goes into the fast surface states or traps. ^' ^ 

 There is, indeed, an assumption here, to the effect that the distribution 

 of traps is unaffected by a change in the chemical environment. The 

 justification for this is the observation of Brown and Montgomery* 

 that it was possible to superpose overlapping large signal field effect 

 curves obtained in different environments. There is also evidence for 

 the validity of this assumption from the self-consistency of the procedure 

 used (see Section V and Fig. 4). 



The photoconductivity measurements have been interpreted on the 

 following basis. Illumination of the sample will do two things: it will 

 change the surface excesses Tp and r„ ,^ and it \\ill also change the 



* The question of the mobility of carriers near the surface should be mentioned 

 liere. For extreme values of Y, the mobility of the carriers tliat are constrained to 

 move in the narrow surface well is reduced. Values for this reduction in mobility 

 have been calculated by Schrieffer.^ However, for values of }' near zero the Schrief- 

 fer correction is small, and at somewhat larger (positive or negative) values AG 

 is increasing so fast that the error in Y introduced by ignoring the Schrieffer cor- 

 rection is small. 



