Junction Noise

[Ian Fritz]

Listers --

Here is some general information on junction noise, taken from S. Sze's book
on semiconductor device physics.

Noise can be classified either (1) by its observed characteristics, i.e., as
thermal noise, flicker noise or shot noise, or (2) by the underlying
physical process, i.e., generation-recombination noise, diffusion noise or
modulation noise. There is not always a one-to-one correspondence between
these classification schemes -- for example diffusion fluctuations cause
thermal noise in bulk material, but also contribute to shot noise at a
junction.

In most junction devices the major noise component is shot noise,
originating from fluctuations in generation and recombination rates,
fluctuations in carrier trapping rates and fluctuations in the diffusion
process, all leading to fluctuations in carrier density. Shot-noise
processes ideally have white noise characteristics.

For an ideal diode (Shockley equation) the noise level is proportional to
the reverse saturation current. The thermal component is additionally
proportional to exp(qV/kT). The reverse saturation current, in turn, depends
on temperature and bandgap roughly according to exp(-Eg/kT), where Eg is the
semiconductor bandgap.

Thus, low-bandgap semiconductors are generally noisier than large-bandgap
ones, and (everything else being equal) Ge junctions are noisier than Si
ones.

I hope this helps clear up any confusion that might have been caused by some
of the posts from yesterday.

  Ian