Modification of the Porter-Thomas Distribution by Rank-One Interaction

The Porter-Thomas (PT) distribution of resonance widths is one of the oldest and simplest applications of statistical ideas in nuclear physics. Previous experimental data confirmed it quite well, but recent and more careful investigations show clear deviations from this distribution. To explain these discrepancies, Volya, Weidenmüller, and Zelevinsky [Phys. Rev. Lett. 115, 052501 (2015)] argued that to get a realistic model of nuclear resonances is not enough to consider one of the standard random matrix ensembles which leads immediately to the PT distribution, but it is necessary to add a rank-one interaction which couples resonances to decay channels. The purpose of this Letter is to solve this model analytically and to find explicitly the modifications of the PT distribution due to such an interaction. Resulting formulas are simple, in good agreement with numerics, and could explain experimental results.