Absolute neutrino masses
Oscillation experiments have proved not only that neutrino oscillates during their propagation; in addition these experiments prove that neutrinos are massive particles. Nevertheless, the absolute scale of neutrino masses is one of the biggest open question not only in neutrino physics, but also in astrophysics and cosmology. Experiments others than oscillations experiments are required to determine the value of neutrino masses.
Oscillation experiments only provide the dierences between the squares of neutrino masses. The current knowledge of neutrino masses leads to the possibility that the solar neutrino doublet has a mean mass either smaller or larger than the remaining atmospheric neutrino [84, 85]. The rst possibility is called normal hierarchy (NH) and inverted hierarchy (IH) for the second one (see gure 2.5). Neutrino mass hierarchy has an impact in experiments to determine the CP violating phase, and in experiments looking for the neutrino less double beta decay .
The most sensitive method to measure the neutrino mass is by observing the spectrum in the beta decay process. In the -decay process a nucleus decay emitting an electron plus an antineutrino. The most known method to measure the neutrino mass via the beta decay process is the tritium decay.
The energy released in the -decay is shared between the electron and the antineutrino.
The energy spectrum of the electron according to Fermi theory is given by [89, 90] / p(E + mec2)(Q with E the electron energy, Q the end point energy, the electron mass me, and the average of the electron antineutrino mass m2 jUeij2m2 i , which corresponds to the incoherent sum of neutrino mass eigenstates. The eect of the neutrino mass parameter is \signi- cant » only in a very narrow region of the spectrum close to Q.
Tritium (3H) decay is an ideal candidate for searching the eects of the neutrino mass in the spectrum end point as is shown in gure 2.6. On the one hand it has a small value of Q (18.6 keV) and then the relative magnitude of the eect is larger, and in the other hand, the one-electron atomic wave functions are well known.
The Mainz and Troitzk experiments have used tritium decay in the past [92, 93], they put an upper limit on the electron neutrino mass of 2.3 and 2.1 eV respectively. These two experiments have been merged in a new experiment called KATRIN, whose expected sensitivity is 0.2 eV .
The and neutrino masses can be determined by studying the kinematics in pion and tau decays. However, these experiments are much less stringent than those obtained by tritium decay experiments. The PSI and the ALEPH experiments found the upper limits [94, 95]: m < 0:17 MeV and m < 18:2 MeV (2.2.2)
In astrophysics and cosmology, neutrino masses are very important to describe the evolution of the universe. Because of the huge abundance of neutrinos, they contribute to the mass density of the universe [96, 97, 98]. The total mass of neutrinos aects the shape of the matter power spectrum of the cosmic microwave background (CMB). Depending on the data set, dierent upper limits on the sum of the three neutrino mass eigenstates has been obtained . The most recent result has been obtained by the Plank collaboration, combining with other data, and they found : m < 0:23 eV (2.2.3)
Thus, at the moment cosmology has established the strongest limit on this quantity.
Besides the great success of the three avor oscillation theory, there are some anomalies in short baseline neutrino experiments that can not be explained in the three neutrino framework. These anomalies suggest that the picture could be incomplete and may be a signal of new physics, pointing to the existence of a light sterile neutrino. In the following we describe some of these anomalies.
The reactor antineutrino anomaly (RAA)
Nuclear reactors are the most intense and controlled source of neutrinos. For a typical reactor, the ssion rate at the nuclear core with a thermal power Pth in GW is 0:31020Pths1, with 6 emitted per ssion [101, 102], which leads to a neutrino ux in a 4 solid angle of 2 1020 neutrinos per second per GW of thermal power.
The antineutrino ux produced in nuclear reactors was reevaluated during the development of the last generation of neutrino oscillation experiments, which have measured 13. The predicted total ux was shifted of about +3% in a rst study by Muller et al.  and was conrmed independently by Huber et al. . This increase in the electron antineutrino ux, was followed by a reevaluation of the results of oscillation experiments at short baselines. Including the update half life of the neutron, o-equilibrium corrections, and the new antineutrino spectra and ux, a total decit of around 7% (3) was found in data with respect to the theoretical calibrations (see gure 2.7). The discrepancy between the new predicted neutrino ux and the observed ux in these experiments has been called the reactor antineutrino anomaly .
The gallium anomaly
The radio-chemical solar experiments GALLEX [104, 105] and SAGE [106, 107] were calibrated using intense radioactive source of 51Cr and 37Ar placed inside the detector.
Mono-energetic electron neutrinos produced by these sources are detected using the reaction e +71 Ga !71 Ge + e (2.3.1)
These experiments reported a ratio between the measured and predicted events rates smaller than unity. The combined ratio gives an average of 0.86. Thus, the total decit of events is 2.8 smaller than the prediction as is shown in Figure 2.8, this has been called the gallium anomaly.
There are 4 (2) neutrino lines from the radioactive sources Cr (Ar). To compute the number of expected events, the cross sections predicted by Bahcall  were used.
The uncertainties in these cross sections are large, because only the cross section for the transition 71Ga !71 Ge into the ground state of 71Ge is well known. While the transitions of 71Ga to the two exited state of 71Ge are inferred using nuclear models . However, even taking into account all these uncertainties, is not enough to explain the total decit . (a) 71Ga !71 Ge (b) Ratios measured and predicted
Light sterile neutrino searches
In addition to the reactor antineutrino anomaly and the gallium anomaly, other discrepancies have been observed. The LSND experiment have reported an excess of 3.8 of electron antineutrinos events in a beam of muon antineutrinos produced by muon decays at rest  + ! e+ + e + (2.4.1)
Electron antineutrinos are detected via the inverse beta decay process at a distance of 30 m. However, the KARMEN experiment did not report any excess in the channel ! e at a distance of 18 m [113, 114]. The MiniBooNE experiment was built to conrm the LSND data. Nevertheless, while the antineutrino data seems to conrm LSND excess, tension appears in the MiniBooNE neutrino data .
To explain all these anomalies, the existence of at least one additional sterile neutrino is required. Oscillation into a light sterile neutrino could explain the neutrino decits observed in the RAA and the gallium anomaly [116, 117]. The best t values for the oscillation parameters suggested by these anomalies are m2 > 1:5 eV2 and sin2 2 = 0:140:08 .
Global ts have been performed using all the short baseline data. However the t results can drastically change depending on the data taken into account [111, 118]. Nevertheless, in the \pragmatic approach », in which the anomalous MiniBooNE data is omitted, the results seems pointing to the existence of a light sterile neutrino with a m2 41 in the eV scale.
In the 3+1 neutrino framework, the transition probability at short baseline experiments has the form of the equation 1.4.23, more precisely it takes the form 
In the case of experiments at short baselines in nuclear reactors, which are disappearance experiments, the survival probability for electron antineutrinos becomes Pe!e = 1 sin2 2#ee. There is an important experimental program around the world to test the eV sterile neutrino hypothesis. In the near future many experiments will explore and test the existence of a light sterile neutrino at the eV scale. All these projects can be classied in dierent categories, according to the neutrino source. Most of the proposals will use nuclear reac2.4. tors as source of electron antineutrinos, but there are also projects which will use intense radioactive sources.
STEREO is a disappearance neutrino experiment and it will use as e source one of the most compact nuclear reactors around the world, the research nuclear reactor at the \Institute Laue-Langevin » (ILL) located in Grenoble (France). The distance from the center of the detector to the center of the reactor core will be only 10 m.
There are several experiments similar to STEREO, with identical or dierent detection techniques. All these projects have similar schedules. In table 2.1 a summary of some of these experiments is presented. One key parameter for the successful detection in these experiments is the baseline. The oscillation length induced by a sterile neutrino in the eV scale is only a few meters, and so the detector must be placed few meters away from the reactor core. In addition, the statistics to be accumulated needs to be large in order to cover all the RAA region. Fortunately, depending on the reactor power, it can be achieved in a relatively short period of time. In the case of STEREO, the acquisition time is expected to be 2 years (120 000 antineutrinos detected).
Table of contents :
1 Neutrino physics
1.2 Neutrinos in the standard model
1.2.1 Masses in the Standard Model and Brout Englert Higgs Mechanism
1.3 Neutrinos beyond the standard model
1.3.1 Dirac mass
1.3.2 Majorana mass
1.3.3 The seesaw mechanism
1.4 Neutrino mixing and oscillation
1.4.1 Neutrino mixing
1.4.2 Neutrino oscillations
188.8.131.52 Two avor oscillations
2 Experimental status
2.1 Neutrino oscillation parameters
2.1.1 Solar neutrino sector
2.1.2 Atmospheric neutrino sector
2.1.3 13 sector: Reactor and accelerator neutrinos
2.2 Absolute neutrino masses
2.3.1 The reactor antineutrino anomaly (RAA)
2.3.2 The gallium anomaly
2.4 Light sterile neutrino searches
3 The STEREO experiment
3.2 Experimental concept
3.3 Detection principle
3.4 Nuclear reactors as neutrino sources
3.5 ILL antineutrino source and spectra
3.6 ILL experimental site
3.6.1 Background in the PN3 casemate
184.108.40.206 Neutron background
220.127.116.11 Gamma background
18.104.22.168 Cosmic ray-induced background
3.7 The STEREO scintillating liquids
3.7.1 Interaction of particles in the liquid scintillator
3.7.2 Scintillation mechanism and non-linearity
3.7.3 Pulse shape discrimination
3.8 The STEREO inner detector
3.8.1 Target and Gamma Catcher
3.8.2 Light collection
3.9 Electronics and acquisition
3.10 Reconstruction software
3.10.1 Pre-processing software
3.10.2 Reconstruction of the position of interaction
3.10.3 Vertex reconstruction in GC
3.11 The detector simulation
3.12 The calibration system
3.12.1 LED monitoring system
3.12.2 The radioactive source calibration system
3.13 Internal shielding
3.14 Muon veto
3.15 STEREO’s sensitivity to light sterile neutrinos
4 Calibration of the STEREO experiment
4.2 Light leaks calibration
4.3 Energy scale calibration
4.3.1 Overview of the calibration procedure
4.3.2 Position of the source
22.214.171.124 Internal source
126.96.36.199 External source
4.3.3 Cell non-uniformity
4.3.4 Collimation device
4.3.5 Gammas from n captures
4.3.6 High energy calibration
4.3.7 Energy non-linearity and choice of calibration sources
4.3.8 Energy scale determination systematic uncertainties
4.3.9 Reactor ON calibration and monitoring
4.4 Gamma catcher calibration
4.5 Visible energy reconstruction
4.6 Neutron capture eciency
4.6.1 AmBe neutron source
188.8.131.52 Neutron capture eciency denition with AmBe
4.6.2 Geometrical neutron eciency dependence in one cell
4.6.3 Neutron eciency inter-calibration: central vs border cell
4.6.4 Systematic uncertainties in the neutron capture eciency determination
4.7 Fast neutron characterization sample and PSD calibration
4.8 Calibration system concept
4.9 Calibration using cosmic-ray induced radioisotopes
5 Characterization of the signal and the background
5.2 Cosmic ray induced background
5.3 Selection criteria
5.4 Prompt topologies
5.5 Delayed topologies
5.6 Gamma background discrimination
5.7 Improved resolution samples
6 General conclusions and perspectives