Towards a Bayesian evaluation technique for light nuclear systems

Abstract

A quantitatively reliable knowledge of several reaction cross sections of light nuclear systems is an important prerequisite for developments in nuclear technology, but also for several fields of science. At present the situation of evaluated reaction cross section data of light nuclei is not fully satisfactory because most are exclusively based on experimental data and thus limited in energy and included reaction channels. The main reason is the absence of a quantitatively reliable (semi-)microscopic description of reactions in light nuclear systems with predictive power. Therefore, for these systems R-matrix analyses are widely performed to describe reaction data. Frequently the corresponding data sets are either incomplete or contain dominant breakup channels which cannot be described by standard R-matrix theory. In this contribution we treat the occurence of a dominant breakup channel within a reduced R-matrix analysis. Furthermore, with regard to a proper evaluation the generation of a prior associated with an R-matrix analysis is suggested which allows to determine an evaluation based on a Bayesian update procedure. A first test of a Bayesian update for n+9Be reactions at low energies is presented.