The most stringent laboratory limits on the neutrino mass scale are derived from the beta decay of tritium. Due to the reactivity of tritium, however, it has so far proven impractical to use atomic tritium as a source. Instead, the KATRIN experiment uses molecular tritium -- T2. When one of the bound T nuclei decays, emitting a beta electron and a neutrino, the process may excite electronic, rotational and vibrational modes of the daughter 3HeT+ molecular ion. These excitations distort the beta spectrum near its endpoint, and will confound the neutrino-mass analysis if not properly accounted for.
Luckily, a number of theorists have worked on this problem and have calculated the probabilities that beta decay will excite the various final states of 3HeT+. (See papers here and here for the latest calculations.) These calculations can't be directly experimentally validated, however, and the theoretical method doesn't allow an intrinsic uncertainty estimate. I am working with a small group on an experiment at UW to lend confidence to the theory by testing the prediction of how often 3HeT+ breaks apart. The main chamber of the Tritium Recoil-Ion Mass Spectrometer (TRIMS) has two silicon detectors at opposite ends of a 60-kV electric potential. When tritium decays in the chamber, the beta is accelerated to one detector and the daughter ion -- 3HeT+, 3He+, or T+ -- is accelerated to the other, guided by a 0.2 T magnetic field. The mass-6 molecule flies more slowly, allowing us to distinguish it from the mass-3 fragments. TRIMS was designed to resolve a disagreement between the theoretical prediction and measurements published in 1957 and 1958, by addressing several of the possible underlying reasons that they might have yielded different answers. For example, TRIMS operates at much lower pressures, so charge-exchange interactions between ions from beta decay and residual TT gas are very unlikely.
Our first results were published in Physical Review Letters in 2020. The TRIMS results are in excellent agreement with the theoretical predictions for both HT and TT, removing a significant asterisk from the KATRIN uncertainty budget. We are now working to finalize all of our HT and TT results in a longer publication, which will include our estimate of what went wrong with the earlier data sets.
To properly include the final-state distribution in the KATRIN analysis, we also need to know the initial states of the T2 molecules in the KATRIN source. How much of the source is T2, and how much is other molecular species like HT and DT? How are the various rotational states of the T2 gas populated? While preparing for TRIMS, we wrote a paper studying this problem for KATRIN, identifying additional lines in the systematic error budget. We are presently working with our KATRIN colleagues to ensure a better understanding of the initial molecular states in the KATRIN source.