Study of the Decay of ¹³⁷Xe to the 5/2⁺ Excited State of ¹³⁷Cs through a Convolutional Neural Network

The first nonunique forbidden beta decay of ¹³⁷Xe to the first ¹³⁷Cs 5/2⁺ excited state is expected to have a nontrivial spectral shape. One way to model the shape correction is to introduce an “effective” value of the axial-vector coupling, g_A, sometimes referred to as “effective g_A quenching.” EXO-200 was a xenon-filled time projection chamber (TPC) filled with 80% enriched ¹³⁶Xe that searched for neutrinoless double beta decay (0νββ) between 2011-2018 at the WIPP underground site in New Mexico. EXO-200 set some of the world’s leading constraints on the decay mode, and performed precision measurements on the 2νββ decay mode. Before its decommissioning, a neutron calibration of the detector was performed using an AmBe source, deployed in the vicinity of the inner detector. Neutron captures on ¹³⁶Xe produce ¹³⁷Xe, whose beta decay to the ground state of ¹³⁷Cs was measured by EXO-200 in 2020 Here we report on a measurement of the spectral shape opf the beta decay of ¹³⁷Xe to the 5/2⁺ excited state of ¹³⁷Cs using a Convolutional Neural Network to classify beta and gamma-ray events and their energy. We also demonstrate measurements of the same process with traditional analysis techniques for comparison.