New analysis developments for reconstruction and classification of CRES signals in Project 8

Abstract

Determination of the absolute scale of the neutrino mass is among the most elusive and unique topics in experimental particle physics. Established direct measurement methods have reached the limit of their realistic potential, and still the values and ordering of the neutrino masses remain unknown. The Project 8 experiment pioneers a new technique called Cyclotron Radiation Emission Spectroscopy (CRES) to improve upon the obtainable neutrino mass sensitivity with ultra-precise frequency analysis. Motivated by a thorough understanding of Project 8 signal characteristics, I present new analyses to further develop the CRES technique. I detail an original algorithm to reconstruct a wider variety of CRES signal types, and a scheme to classify these types using machine learning with a demonstrated accuracy of 95%. I discuss the incorporation of these and other future developments with existing techniques to work toward a CRES analysis capable of probing the tritium endpoint with eV-scale precision or better. Lastly, I present preliminary studies of alternatives to the Fourier Transform for CRES signal processing and discuss the current limitations based on data as well as the future potential.