Treatment of Microscopic Fluctuations in Noise Theory

In this article we describe a very simple means of treating microscopic fluctuations in noise theory. The method is simple in the sense t h a t it focuses attention directly on the heart of the matter, the elementary processes which give rise to device noise. It is also simple in the sense that no sophistication in probability theory is used beyond an understanding of simple shot noise. Nonetheless, the method is rigorous under the rather mild constraint that the fluctuations are sufficiently small that the equations governing the noise are linear. The method has the added advantage t h a t it can be used nearly as easily for nonstationary noise as for stationary noise. We make no 1041 claim that this method is an advance in the philosophy of noise; we do claim, however, t h a t it is adequate for solving many noise problems of practical interest. There are two equivalent 1,2 methods of calculating device noise, the Langevin 3 method (LM) and the impedance-field 4 method (IFM). Both methods are characterized by an inherent simplification: namely, the separation of the task of calculating the spontaneous fluctuations of the current carriers in each elemental region of the device, and the task of calculating the observable response to these fluctuations at the external contacts of the device. The former task, the treatment of t h e microscopic fluctuations, is simplified because in dealing with the source of the fluctuations one can focus attention on the statistics of the microscopic variations inherent to the local physical conditions, which in turn are determined by the (noiseless) state of the device during operation.