Evidence for Neutrinoless Double Beta Decay

Abstract

The data of the Heidelberg-Moscow double beta decay experiment for the measuring period August 1990 - May 2000 (54.9813 kg y or 723.44 molyears), published recently, are analyzed using the potential of the Bayesian method for low counting rates. First evidence for neutrinoless double beta decay is observed giving first evidence for lepton number violation. The evidence for this decay mode is 97% (2.2σ) with the Bayesian method, and 99.8% c.l. (3.1σ) with the method recommended by the Particle Data Group. The half-life of the process is found with the Bayesian method to be T_{1/2}^{0ν} = (0.8 - 18.3) x 10^{25} y (95% c.l.) with a best value of 1.5 x 10^{25} y. The deduced value of the effective neutrino mass is, with the nuclear matrix elements from [Sta90,Tom91] < m >= (0.11 - 0.56) eV (95% c.l.), with a best value of 0.39 eV. Uncertainties in the nuclear matrix elements may widen the range given for the effective neutrino mass by at most a factor 2. Our observation which at the same time means evidence that the neutrino is a Majorana particle, will be of fundamental importance for neutrino physics. PACS. 14.69.Pq Neutrino mass and mixing; 23.40.Bw Weak-interaction and lepton (including neutrino) aspects 23.40.-s Beta decay; double beta decay; electron and muon capture.

Figures (7)

• Figure 1: Sum spectrum of the $^{76}{Ge}$ detectors Nr. 1,2,3,4,5 over the period August 1990 to May 2000, (54.981 kg y) in the energy interval 2000 - 2080 keV, around the Q$_{\beta\beta}$ value of double beta decay (Q$_{\beta\beta}$ = 2039.006(50) keV). The curve results from Bayesian inference in the way explained in the text. It corresponds to a half-life T$_{1/2}^{0\nu}$=$(0.80 - 35.07) \times 10^{25}$ (95% c.l.).
• Figure 2: Sum spectrum of the $^{76}{Ge}$ detectors Nr. 1,2,3,5 over the period August 1990 to May 2000, 46.502 kg y. The curve results from Bayesian inference in the way explained in the text. It corresponds to a half-life T$_{1/2}^{0\nu}$=(0.75 - 18.33)$\times~10^{25}$ y (95% c.l.).
• Figure 3: Sum spectrum, measured with the detectors Nr. 2,3,5 operated with pulse shape analysis in the period November 1995 to May 2000 (28.053 kg y), in the region of interest for the $0\nu\beta\beta$ - decay. Only events identified as single site events (SSE) by all three pulse shape analysis methods HelKK00KKMaj99 have been accepted. The spectrum has been corrected for the efficiency of SSE identification (see text). The curve results from Bayesian inference in the way explained in the text. The signal corresponds to a half-life T$_{1/2}^{0\nu}$=$(0.88 - 22.38) \times 10^{25}$ y (90% c.l.).
• Figure 4: Scan for lines in the full spectrum taken from 1990-2000 with detectors Nr. 1,2,3,4,5, (Fig. \ref{['Sum_spectr_Alldet_1990-2000']}), with the Bayesian method. The ordinate is the probability K that a line exists at energy E. Left: Energy range 2000 -2080 keV. Right: Energy range of interest around Q$_{\beta\beta}$.
• Figure 5: Left: Probability K that a line exists at a given energy in the range of 2000-2080 keV derived via Bayesian inference from the spectrum shown in Fig. \ref{['Sum_spectr_1-2-3-5det_95-2000']}. Right: Result of a Bayesian scan for lines as in the left part of this figure, but in the energy range of interest around Q$_{\beta\beta}$.