COMBINED MEASUREMENTS ON ETCHED GERMANIUM SURFACES 1023 



steady-state carrier density deep inside the sample. If the sample is 



thin in comparison with the body diffusion length and with (D/s), as 



was the case in our experiments, the added carrier density Ap will be. 



almost uniform throughout the thickness t of the sample, and one can 



easily convince oneself that the photoconductance arising from this 



cause is of the order of it/£) times larger than that arising from the 



changes in the surface excesses, where £ is a Debye length for the 



I material. This being the case, the photoconductivity may be considered 



I to be a bulk rather than a surface effect, the surface entering only 



i through the surface recombination velocity s. Under the conditions of 



the present work the magnitude of the photoconductivity w^as in fact 



inversely proportional to s, as was verified in a separate set of experi- 



I ments. Surface recombination is of interest in that this also calls for 



I "fast" trapping centers on the surface; in fact any trap contributing 



\ to the field effect experiment may be a recombination centre, if the 



i cross-sections are right. The questions as to whether the recombination 



I centres and the "fast states" affecting the field effect are the same, or 



I not, is taken up in the succeeding paper. 



I The surface photo-voltage, like the field effect, is affected both by 



I changes in the surface excesses and by changes in Ss , the charge in sur- 



I face traps. In the experiments, the change in contact potential in a cer- 



I tain light (usually chosen so that the change i s small in comparison 



[ with kT/e) is compared with the change in conductance produced by the 



same light. From the latter one may calculate 8 (defined as Ap/ui) 



directly. The change in contact potential, measured in units of kT/e, is 



I taken to be equal to AY. Thus the surface photo-voltage experiment 



I measures the quantity (dY/d8), the differential being taken at constant 



surface charge. By a slight generalization of the argument previously 



given by the authors," one can show that: 



dy ^ _ (d/d8)Y{Tp - rj + id2jd8)Y (.. 



(18 (d/dVUTp - r„) + (dXs/dVh ^ ^ 



Now the first terms in the numerator and denominator on the right- 

 I hand side are determinate functions of Y, and so are known ; the quantity 

 I {d'2s/dY)B may be deduced from the field-effect measurements, so that 

 I the only remaining ciuantity, (53,, /(95) r , niay be deduced from the 

 I measurements of surface photo-voltage. 



In concluding this section, a word as to the meaning to be attached to 

 (dT^s/dS) Y is in order. The sign of this quantity depends, roughly speak- 

 ing, on whether the traps in question (i.e., those near the Fermi level 

 under the conditions of the experiment) are in better contact with the 



