Theory of Magnetic Effects on the Noise in a Germanium Filament

In a series of fundamental experiments, H. C. Montgomery1 has established that minority carriers play an important part in the currentnoise associated with semiconductors. He found that on the one hand, the noise voltage is usually proportional to the biasing current, suggesting fluctuations in the conductivity, and hence the carrier concentration. On the other hand the spectrum of the noise suggested a rather coarsegrained time variation, not likely to be caused by fluctuations in the normal carrier density. One might conclude, therefore, that the noise is caused by a distribution of sources emitting or absorbing minority carriers in random bursts. Such carriers would be subject to the same laws of motion and of recombination as intentionally injected carriers. Montgomery was, in fact, able to verify that the noise along a filament showed marked correlation over a distance roughly equal to that through which minority carriers could drift in the biasing field before recombination. W. Shockley has pointed out another corollary of this theory: A magnetic field transverse to the filament should have a pronounced effect on the noise. This conclusion, too, Montgomery was able to verify experimentally.1 His results are in good qualitative agreement with theory. Complete quantitative agreement was perhaps not to be expected, since technical difficulties prevented attainment of the idealized conditions assumed by the theory. This paper gives an account of that 647