Zum kollektiven Anteil des Plasma-Mikrofeldes

Abstract

The Holtsmark microfield evaluation is rediscussed. It is shown that in the usual computation of the Fourier transform a term is neglected which has the dimension of a dipole moment density. This term occurs only for long range forces (here Coulomb ones) and remains undetermined in the limit of an infinitely large system without interaction. Furthermore, it is sensitive to deviations from ergodicity. The true microfield distribution results as a Holtsmark distribution centered at the generally nonvanishing far field. As the distribution of the latter is Gaussian the complete distribution results as a convolution of the Holtsmark distribution with that of the far field. The parameter of the Gaussian far field distribution is computed for a non interacting plasma consisting of N charges, by taking into account explicitely the collective field of space charge fluctuations. The formulae arrived at with respect to the microfield as well as the micropotential suggest to perform numerical computer experiments (HUNGER, LARENZ and WILKE, to be published). For an interacting plasma, the far field is obtained qualitatively. It agrees with the microfield of HUNGER and LARENZ (1961). The far field dominates the Holtsmark nearest neighbour field at plasma conditions kT/e