An experiment for the direct calorimetric measurement of the neutrino mass

The oscillation experiments have clearly shown that neutrino are massive particle. Nowadays the experiments based on kinematic analysis of electrons emitted in nuclear beta-decay are the most sensitive for a direct electron-neutrino mass determination. The method consists in searching for a tiny deformation caused by a non-zero neutrino mass to the spectrum of the charged particles emitted near the end point. A possible approach is the calorimetric one. In a calorimetric measurement the source is embedded in the detector and all the energy is measured, except for the one taken away by the neutrino. A drawback of this approach is that the full spectrum is acquired, while only the decays very close to the end-point are useful for measuring the neutrino mass. Therefore, the source activity has to be limited to avoid pile-up which would deform the shape of beta spectrum. As a consequence the statistics near the end-point is limited as well. This limitation may be then partially balanced by using isotopes with an end-point energy as low as possible. In this scenario an international collaboration has grown around the project of Microcalorimeter Arrays for a Rhenium Experiment (MARE) for a direct calorimetric measurement of the neutrino mass with sub-electronvolt sensitivity. Although the baseline of the MARE project consists in a large array of rhenium based thermal detectors, a different option for the isotope is also being considered. The two competing isotopes are 187Re and 163Ho. While the first beta decays, the latter decays via electron capture, and both have a Q value around 2.5 keV. The MARE project has a staged approach. The first phase of the project (MARE-1) is a collection of activities with the aim of sorting out both the best isotope and the most suited detector technology to be used for the final experiment. The goal of the last phase (MARE-2) is to achieve a sub-eV sensitivity on the neutrino mass. It will deploy several arrays of thermal microcalorimeters. During my Ph.D I have focused only on the rhenium isotope, neglecting the holmium. In fact, in the case of rhenium I have estimated the statistical sensitivity of a neutrino mass experiment performed with thermal calorimeters. First, through an analytical approach, I have derived an algorithm to assess the statistical sensitivity for a given experimental configuration and then, for the same experimental configuration, I have estimated the sensitivity on neutrino mass via a Montecarlo method. The results of the analytic approach are then validated through the comparison with the Montecarlo results over a wide range of experimental parameters. The investigation is carried out for both phases of the MARE experiment. For example, the Montecarlo approach has shown that a neutrino mass sensitivity of 0.1 eV at 90% CL could be expected in 10 years running 3x10^5 detectors, each with a mass of 10 mg (~10 Hz) and with energy and time resolutions of about 1 eV and 1 μs respectively. Instead, a sensitivity on neutrino mass of 3.4 eV at 90% CL could be achieved in 3 years using 288 detectors, each with a mass of 500 μg (~ 0.3 Hz) and with energy and time resolutions of about 30 eV and 300 μs respectively. The latter is the configuration of the Milano MARE-1 experiment, which is one of the MARE-1 activities. Subsequently, I have exploited the Montecarlo approach to study the main sources of systematic uncertainties of the calorimetric experiments, as the shape of the beta spectrum and the Beta Environmental Fine Structure (BEFS), which is a modulation of the beta spectrum due to the atoms surrounding the decaying nuclei. The systematics uncertainties relating to the source (i.e. excited final states and the escape electron) have been also investigated. Finally, I have evaluated the capability of the MARE experiment to measure the mass of heavy neutrinos from some tens of eV to 2.5 keV. I have also participated in the Milano MARE-1 experiment. This experiment is carried out in Milano by the group of Milano--Bicocca in collaboration with NASA/GSFC and Wisconsin groups. The Milano MARE-1 arrays are based on semiconductor thermistors, provided by the NASA/GSFC group, with dielectric silver perrhenate absorbers, AgReO4. These arrays consist of 6 x 6 implanted Si:P thermistors on which single crystal of AgReO4 are attached. The mass of a single absorber is around 500 μg, corresponding to a single detector rate of 0.3 Hz. The cryogenic set-up of MARE-1 is designed to host up to 8 arrays (i.e. 288 detectors), but the installation of only two arrays has been funded so far. The read-out electronics of MARE-1 in Milano is characterized by a cold buffer stage, based on JFETs which work at about 120 K, followed by an amplifier stage at room temperature. To electrically connect the detector at 85 mK to the JFETs at 120 K two decoupling stages are needed. The two stages have also to guarantee the mechanical stability. The first stage separates the detectors from the JFETs box, while the second one decouples the cold electronics box from the JFETs. In this context, the activities I have carried out were focused primarily on the assembly of the entire cryogenic set-up of MARE-1 in Milano and then on its analysis and improvement. Firstly, I have performed several cool-downs devoted to test the detector performances and to determinate the best thermal coupling between Si thermistors and AgReO4 absorbers, in conclusion of which we have obtained an energy resolution of around 30 eV at 2.6 keV and a rise time of about 300 μs. With 72 detectors and such performances, a sensitivity on neutrino mass of 4.7 eV at 90% C.L. is expected in three years running time. During these cool-downs it was used the electronics of the MIBETA experiment, the predecessor of the MARE-1 experiment in Milano. Since its first installation the cryogenic set-up of MARE-1 has presented several structural and thermal problems. The first has concerned the electrical connections between the detectors and electronics, while the latter the insufficient thermal decoupling between the JFETs support and the cold electronic box as well as the insufficient thermalization of the array ceramic board and of the array itself. As a consequence, no signal could be acquired. Therefore, I have performed an R&D work in order to solve all of these problems in conclusion of which the detectors have reached a base temperature such that it was possible to acquire a first spectrum with a threshold below 800 eV. In this condition, an energy resolution of 175 eV at 1.5 keV and of 181 eV at 5.9 keV have been obtained, while the rise time was about 850 μs. It was the first time that a spectrum with this threshold was acquired with the MARE-1 set-up. The worsening observed in the detectors performances with respect to the test runs was due to an excessive microphonics noise. Nevertheless it can be hypothesized that a 72 channels measurement will be starting soon.