β− -Decay Half-lives Using the ANN Model: Input for the R-Process

Abstract

Full understanding of nucleosynthesis via the r-process continues to be a major challenge for nuclear astrophysics. Apart from issues within astrophysical modeling, there remain significant uncertainties in the nuclear physics input, notably involving the β- decay halflives of neutron-rich nuclei. Both the element distribution on the r-process path and the time scale of the r-process are highly sensitive to β− lifetimes. Since the majority of nuclides that lie on the r-process path will not be experimentally accessible in the foreseeable future, it is important to provide accurate predictions from reliable models. Toward this end, a statistical global model of the β−-decay halflife systema- tics has been developed to estimate the lifetimes of nuclides relevant to the r-process, in the form of a fully-connected, multilayer feedforward Artificial Neural Network (ANN) trained to predict the halflives of ground states that decay 100% by the β− mode. In predictive performance, the model can match or even surpass that of conventional models of β-decay systematics. Results are presented for nuclides situated on the r-ladders N=50, 82 and 126 where abundances peak, as well as for others that affect abundances between peaks. Also reported are results for halflives of interesting neutron-rich nuclides on or towards the r-process path that have been recently measured. Comparison with results from experiment and conventional models is favorable.