Reliable calculations of level densities using statistical spectroscopy

Abstract

Nucleosynthesis calculations require nuclear level densities for hundreds or even thousands of nuclides. Ideally one would like to constrain these level densities by microscopically motivated yet computationally cheap models. A statistical approach suggests that low moments of the Hamiltonian might be sufficient. Recently Zuker proposed a simple combinatoric formula for level densities based upon the binomial distribution. We rigorously test the binomial formula against full scale shell-model diagonalization calculations for selected $sd$- and $pf$-shell nuclides and also against Monte Carlo path integration calculations. We find that the fourth moment is as important as the third moment to a good description of the level density, as well as partitioning of the model space into subspaces.