A novel light detector for DarkSide-20k

The astonishing precision of the Standard Model (SM) of particle physics in predicting observables to be measured by accelerators can rival only with its inability at describing most of the energy-matter content of the universe. The most widely accepted cosmological model, the ΛCDM, quantifies the percentage of matter fitting in the SM description to a mere 5% of the total. The remaining 95% has to be divided among an unknown form of non-radiating matter and an even more mysterious form of energy responsible for the universe accelerating expansion. In a 70 years long period, since its first formulation, the Dark Matter (DM) scientific paradigm has accumulated a huge number of arguments in its favor. Despite this, an irrefutable DM direct detection claim has not yet been made. Among all the possible species of dark matter, the Weakly Interacting Massive Particles (WIMPs) form a class of candidates capable of explaining this physics puzzle in a natural way. Many experiments around the world aim for a WIMP direct detection under the leading assumption that a weak DM-SM interaction is possible. Large noble liquids detectors lead this search in the high WIMP-mass region thanks to their scalability to high target masses. Given the current lack of positive observations, a new generation of multi-ton experiments is being designed to explore the whole WIMP-nucleon interaction cross section range. DarkSide-20k aims to push its sensitivity up to the ultimate experimental limit, the neutrino floor. This can be achieved exploiting the liquid argon electron/nuclear recoil discrimination power, which together with volume fiducialization and neutron tagging, allows the detector to run in background-free mode. The key to successfully apply pulse shape discrimination to reject β/γ induced events lies in having a high light yield, which is defined as the average number of detected photons per unit of energy released in the target volume. In order to boost the detector photon efficiency in the TPC volume, the DarkSide collaboration decided to abandon the traditional Photomultiplier Tube (PMT) technology in favor of Silicon Photomultiplier (SiPM) one. Aiming to optimize these light sensors for the experiment needs, an intense R&D activity was carried out to successfully equip and operate in liquid argon nearly 14 m2 of SiPMs grouped in 5210 Photon Detector Modules (PDM). My PhD thesis is set in this context. A first phase of the work consisted in choosing a suitable SiPM technology as building block of the future Photon Detector Module. A campaign of cryogenic characterization of several SiPM models was conducted. Pulse features (gain, pulse amplitude, single cell recharge time) as well as dark count rate (DCR) and correlated noises (direct crosstalk, after-pulse, delayed crosstalk) were studied as function of temperature and operating bias. The outcome of this campaign is described in Chapter 3 of my thesis and in the articles [1, 2]. This step came with the formulation of several requirements which PDMs have to comply in order to ensure the accomplishment of the design physics reach of DarkSide-20k. In particular I contributed to formalize the specifications regarding the noise rate, SNR and dynamic range of the SiPM readout as described in Chapters 2, 5 and 6. The second step towards a functioning PDM was the design of a transimpedance (TIA) preamplifier capable of operation in liquid Argon. The circuit was optimized to be compatible with the huge capacitance typical of solid-state photo-detectors. The extremely low noise and high bandwidth of the resulting TIA allow to readout a 10×10mm2 SiPM with a SNR of 30 and a timing resolution better than 1 ns. Starting from these results it was then possible to develop a SiPM connection scheme to read 24 cm2 as a single channel fulfilling the SNR and timing resolution requirements. These advancements are described in Chapter 4 and [3, 4]. The SiPM tile performances in terms of SNR and timing were reached also thanks to the study and implementation of advanced digital filtering techniques. This, together with a possible DAQ scheme based on these results, is discussed in Chapter 5. Finally I developed a simulation code aiming to describe the effects of the electronics and readout logic on the physics observables. The C++ based code adds to the Geant4 simulation of the detector the effects of the readout (noise, dynamic rage, timing). A whole set of the PDM parameters can be changed in order to study their impact on the performances of DarkSide-20k. This was fundamental to the editing of the specifications of the experiment as described in the Yellow Book [5]. Moreover my contribution extended to the definition of the requirements of the trigger logic and readout strategy. This part of the work is described in Chapter 6.