Brief critical analysis of the Darwin-Fowler method

Abstract

We present a brief numerical study of the Darwin-Fowler method applied to the analysis of the energy partition of essembles of bosons and fermions. We analyze the assertion of the existence of a "strong maximum" made in the original paper of Darwin and Fowler and other studies and show that although the presumed saddle point along the real axis of the grand canonical parameters may exist it cannot, in general, be characterized as "strong", in the sense of having much larger magnitude than the other points along the path of integration. We show that in some cases the saddle point is not even present and the various approximations of the method can be interpreted as a tricky reformulation of usual thermodynamic relations. The close connection of the method with the formalism of the Laplace transform may produce wrong results if the internal energy of the components of the ensemble is not large enough. Therefore, although useful in many applications the Darwin-Fowler method may not be suitable, in general, for a detailed microscopic analysis of the nuclear structure in connection with the Shell Model approach, as it is usually done in studies of the pre-equilibrium stage of nuclear reactions.