KATRIN is a new-generation tritium beta-decay experiment. In its first data release in 2019, the experiment established an upper limit of the effective electron-antineutrino mass at < 1.1 eV/c2 (90% CL) [2]. That result was based on 522 live hours of data with the source that keeps molecular tritium at 30 K running at a partial capacity of about 2.5⨉1010 Bq.
In this latest release of 744 additional live hours of data, a substantial improvement in statistics was achieved by running the source at its nominal strength of 9.5⨉1010 Bq. An improved vacuum condition in KATRIN’s main spectrometer, shown in Fig. 1, also reduced the background near the tritium endpoint by 25%. These improvements elevated the experiment’s sensitivity to measure distortion of the beta-electron energy spectrum near the endpoint — a signature for massive neutrinos.

Figure 2. The Bayesian posterior distribution of the square of the effective electron-antineutrino mass from the KATRIN experiment. The posterior of the first data set (green) was used as a prior for the second data set. The combined posterior (blue) gives an upper limit of mν < 0.73 eV/c2 (90% CI) in the mass. This is consistent with the slightly larger Frequentist limits
NSD Project Scientist Bjoern Lehnert led the efforts to use Bayesian analysis techniques to study the spectral distortion on NERSC’s supercomputer Cori. This analysis is unique as the other analysis teams in KATRIN deployed Frequentist methods. Figure 1 shows the posterior distribution of mν2 from the Bayesian analysis for both data releases. The results from all analysis methods are consistent with each other. In the final results of this new data set, the best-fit mν2 was determined to be (0.26±0.34) eV2/c4. When combined with the first data set, mν was limited to < 0.8 eV/c2 (90% CL).
Other members of the local KATRIN group include postdoctoral fellows Rebecca Carney and Ann-Kathrin Schuetz, and senior scientist Alan Poon.
References
[1] M. Aker et al. [KATRIN Collaboration], First direct neutrino-mass measurement with sub-eV sensitivity, submitted for publication, arXiv:2105.08533.
[2] M. Aker et al. ([KATRIN Collaboration], An Improved Upper Limit on the Neutrino Mass from a Direct Kinematic Method by KATRIN, Phys. Rev. Lett. 123, 221802 (2019).