Simulation based neutron background studies for the CRESST and COSINUS dark matter search experiments

Abstract

The nature of dark matter (DM) is one of the big remaining mysteries in modern cosmology and astro-particle physics. Our standard cosmological model predicts a form of non-luminous, non-baryonic matter accounting for more than 80 percent of the total matter density in the observable universe. While this prediction is quite robust against concurring theories and supported by various cosmological observations lead by precision measurements of the cosmic microwave background (CMB), we still lack an experimental proof of it. Different experiments are thus currently searching for dark matter via collider-based, indirect or direct detection techniques, with the latter having the advantage of directly probing features of our galactic dark matter halo. A prominent dark matter candidate are so-called weakly interacting massive particles (WIMPs), so that the analysis in many of the experiments may have a distinct focus on assuming correspondingly predicted interaction mechanisms.Among the direct search experiments, CRESST (Cryogenic Rare Event Search with Superconducting Thermometers) has been the pioneer in developing a two-channel read-out based on phonon and scintillation light signals, used for particle discrimination. As WIMP-like dark matter is expected to interact with the nuclei in the target, this technique allows to vastly reduce the background due to electrons and gammas depositing energy in the detectors mainly via interactions with electrons. On the other hand, neutrons entering the detectors can lead to nuclear recoils indistinguishable from a dark matter induced signal and may hence be considered the most dangerous background.Employing the same detection technique but different target material, the COSINUS (Cryogenic Observatory for SIgnatures seen in Next-generation Underground Searches) experiment is aiming to cross-check the results of the DAMA/LIBRA experiment, reporting an annually modulating signal interpreted by its proponents as a model-independent evidence of galactic dark matter. COSINUS will thus use the same target material as DAMA/LIBRA, i.e. NaI, and is currently in prototyping and early construction phase. This thesis will comprise simulation based neutron background studies for the CRESST and the COSINUS experiment, crucial for the understanding and interpretation of measured signals and for the evaluation of the employed experimental geometries. There are two principal neutron sources: muon-induced neutrons produced in interactions of cosmic muons, and radiogenic neutrons originating from (alpha,n) and spontaneous fission processes due to radioactive contamination in materials surrounding the detectors. An integrate investigation of both sources, and hence of the total neutron background will be the main topic of this work.In both experiments, the setup is (will be) located in a deep underground laboratory in order to shield the detectors against cosmic radiation. In addition, detectors are surrounded by layers of distinct materials serving the purpose of minimizing the signal rate due to known external background sources. In CRESST, the respective shielding is in place since many years. However, this study will be the first taking into account an additional part of the passive neutron shield as well as the latest detector design. As the current detectors enable CRESST to measure nuclear recoils of energies as low as O(10..100 eV), this study will allow to test the validity of the simulation models at unprecedentedly low nuclear recoil energies. In this energy regime, a further pressing topic are suitable energy calibration sources. Inelastic neutron interactions will be investigated as a candidate for providing sub-keV peaks potentially being able to serve this purpose. The COSINUS simulation studies presented in this thesis, on the other hand, will form the basis of the planned shielding geometry. A careful evaluation will lead to choosing the design which minimizes the external background. Necessary for the simulation study in both cases is a reconstruction of the core experimental geometry in a general purpose Monte Carlo simulation framework. Additionally, a set of new features has to be implemented in the code in order to enable the possibility to simulate homogeneous bulk contaminations in shielding materials leading to radiogenic neutrons, and muons reaching the underground laboratory leading to muon-induced neutrons. As CRESST has already collected data over the past years, the simulation will be validated against neutron calibration data and, if statistics allows to, also against background recorded during the physics data taking campaigns. To provide a realistic neutron background model, the characteristics of the detectors – e.g. energy resolution, threshold and signal survival probability – have to be taken into account. Hence, a postprocessing scheme for signal reconstruction based on the simulated data collected in multiple detectors will be developed. Finally, the background model will be used to evaluate the efficiency of the shielding geometry with respect to neutrons, potentially leading to the detection of weaknesses or suggestion of improvements. Furthermore, the estimated spectra will be included in the likelihood exclusion limit calculations, thus further constraining and improving the fit and its sensitivity to possible unknown signals.In the initial simulation campaign for COSINUS, the main goal is to evaluate the optimal shielding setup for the experiment based on a design using a water tank as an outermost layer. As the cryostat and detector geometry is not yet finalized, this will be achieved by first minimizing the flux of background particles entering the simulated volume, in which the detectors will be placed, and then implementing a preliminary detector design to further estimate the expected signal rate in the detectors. The conceptual shielding design is chosen with the idea in mind to use the water tank as an active Cerenkov veto. This will be accomplished by operating the water tank with photo multiplier tubes (PMTs). Dedicated simulations will thus be run to evaluate various arrangements of PMTs and the impact of a reflective foil at the tank walls with respect to the efficiency of the veto considering different trigger conditions.The results of this thesis will provide crucial contributions to the development and evaluation of the shielding geometries and background models of the two experiments.