Abstract
Shot noise and thermal-equilibrium noise are often, and incorrectly, viewed as additive and independent noise sources. The two forms of noise can even be identical, as recognized by several pionering investigators long ago. Shot noise measures the graininess of the conduction process, and in samples with a fixed number of carriers, where the mean free path is short compared to the sample length, it is determined by the passage of a carrier through its mean free path. For a classical nondegenerate conductor we show that shot noise calculated that way is the thermal-equilibrium current. For mesoscopic conductors that transmit elastically with a probability small compared to unity, thermal-equilibrium noise is also shown to be approximately the sum of two shot-noise terms, one for the arriving stream at each reservoir. For a semiclassical metallic conductor, at zero temperature, the noise proportional to current flow is shot noise related to the charge transferred between sample electrodes, when a carrier goes through a mean free path. At higher temperatures, a suggestive physical argument is used to show that noise increases above thermal-equilibrium noise according to the square of the ratio of the energy gained in one mean free path, to the thermal energy. © 1993 The American Physical Society.