Study of HEMT electronics coupled to cryogenic detectors for the RICOCHET experiment

Cryogenic germanium detectors with heat-and-ionization double readout are able to identify the nature of the recoil induced by an interaction. Running in a low radioactivity dilution fridge and surrounded by lead and polyethylene shieldings they can be used as rare events search detectors. The RICOCHET experiment bolometers will measure neutrinos interactions by the so-called CENNS process (Coherent Elastic Neutrino Nucleus Scattering). They will discriminate the nuclear recoils induced by this interaction from the electronic recoils dominating the background events. The RICOCHET goal is to perform a precision measurement of the CENNS process, to compare it with the Standard Model predictions and to search possible signs of new physics. The experimental payload will consist of 27 germanium detectors of 38 g (CryoCube) and 9 superconducting zinc detector of 30 g (Q-Array). The cryostat will be installed at the Institut Laue-Langevin (ILL), 8 meters away from the 58 MW research reactor. The antineutrino flux of 10¹²/(cm².s) at the detector level will induce 15 CENNS interactions per day on the CryoCube. The CryoCube detectors are produced by the IJCLab and are an optimization of the massive (800 g) detectors developed for the EDELWEISS-III Dark Matter search experiment. This optimization already allowed to reach 20 eV resolution on the heat channel and 200 eV_ee (electron equivalent) on the ionization channel. The goal is to reach a nuclear recoil threshold of 50 eV, corresponding to heat and ionization resolutions of 10 eV and 20 eV_ee, respectively. This thesis work aims at reaching those resolutions with the use of a High Electron Mobility Transistor (HEMT) technology developed by the C2N laboratory of CNRS for low frequency and high impedance applications. The HEMT are working at cryogenic temperatures below 4 K, which allows to use them close to the detectors. This reduces the total length and the stray capacitance of the cabling between the detectors and the first amplification stage, usually based on Si-JFET transistors working at about 120 K. Reducing the parasitic capacitance directly improves the ionization channel performance. A noise model, based on both 4.2 K HEMT data from C2N and optimized geometries for the RICOCHET detectors, has been developed. This model shows that the objectives are reachable for 38 g detectors and a total input capacitance at the HEMT input reduced to 20 pF. HEMT performance, such as gain and noise, has been measured in various conditions of temperature and power dissipation to constrain their use in the future experiment. Homogeneity of the results has been studied on 25 transistors and extrapolated to demonstrate the feasibility for RICOCHET to use up to 150 HEMT cooled on a 1 K stage only few cm away from the detectors. The readout and bias electronics will be coupled to the digitization, repartition and synchronization parts in a Bolometer Box, jointly developed by IP2I and Institut Neel electronics groups. The HEMT have been coupled to prototype detectors and have already reached a resolution of 41 eV_ee on the ionization channel before final electronics and cabling optimization. This resolution is the best obtained so far worldwide on massive cryogenic germanium detectors. Further optimization steps are well defined to reach the 20 eV_ee goal. Heat performance is limited to 60 eV as of today. A large excess noise has been identified but not fixed during this thesis work. Possible explanations have been exposed and should be tested in future works. The heat channels will be readout by a more standard, Si-JFET based, cold electronics during the first phase of the RICOCHET experiment. (author)