ELECTRICAL NOISE IN SEMICONDUCTORS 



973 



and since the band is not very wide, the variation in dehiy over the hand 

 is not important. Frt)m the drift mobiht,y of holes we may estimate the 

 transit time between segments, according to the relation 



t = L/fiE 



whei(^ t = transit time, seconds 



L = distance between segment mid-points, cm 



A' = applied field, volts/cm 



M = mobility of holes, cm"/volt-sec. 



Data for a bridge of /(-type germanium of resistivity about 20 ohm- 

 cm are gi\-en in Table I. The transit distance, L, after a small correction 



Table I 



for reduced field across the side arm, was taken as 0.305 cm. As noted 

 in the table, the bridge temperature rose somewhat at the higher bias 

 values, and the assumed values of mobility have been modified accord- 

 ing to the inverse three-halves power of the absolute temperature. The 

 delay required for optimum correlation is shown in the second column 

 of the table, and the calculated transit time between segments in the 

 last column. It is seen that the two are in reasonably good agreement, 

 especially at low fields. When the direction of the field is reversed, an 

 ecjual delay is required, but in the opposite channel of the measuring 

 circuit, as would be expected. Here, again, we have experimental evi- 

 dence supporting the noisy hole injection hypothesis. The cause of the 

 discrepancy shown in the table at higher fields is not understood. It is 

 possible that trapping phenomena increase the transit time over that 

 calculated from the mobility. There is some evidence for this sort of 

 behavior in lifetime experiments, but to date there does not seem to be 

 enough information for any estimate of magnitude of such an effect. 



VI. GENERAL COMMENTS 



These studies of electrical noise in semiconductors leave little doubt 

 that the noise is closely related to the behavior of the minority carriers. 



