ausblenden:
Schlagwörter:
CRESST
Zusammenfassung:
The nature of dark matter is one of the big puzzles in modern physics. Indications for dark matter can be found throughout astrophysical observations, from the cosmic microwave background to the rotation speed of galaxies. In the prevalent ΛCDM model of cosmology, dark matter makes up about 26% of the energy content of the observable universe and more than 80% of its matter content. Several theoretically motivated candidates for dark matter are available, stemming from solutions to different problems in the standard model for particle physics. However, despite significant experimental efforts and advancements over the last decades, no clear evidence for any candidate has been found yet. The astrophysical evidence for dark matter as well as possible candidates are summarized in chapter 1 of this thesis. Presented in chapter 1 are the different experimental approaches for the detection of dark matter. They can be broadly classified into three categories: collider experiments, indirect detection and direct detection. The CRESST experiment belongs to the third category. Its aim is the identification of dark matter particles by their scattering off nuclei of a target crystal. The challenges here lie in the extremely low expected rates and small transferred energies (O(eV - keV)). CRESST tackles this problem with a cryogenic detector, operated around 15 mK, transition edge sensors (TES) and a SQUID based read-out. The low operation temperature reduces heat capacities to a point where a keV energy deposition results in an up to a few μK temperature increase. A TES is a very sensitive thermometer based on a superconductor in its transition from the super to the normal conducting phase. It converts the temperature increase into a resistance increase which in turn is read out by a magnetically coupled SQUID. To be able to detect the extremely low rates, the CRESST experiment is located in an underground laboratory and encased in multiple layers of shielding material against environmental radioactivity. Furthermore, CRESST exploits the particle dependent scintillation of the CaWO4 crystals it uses as its primary target. By adding a light detector, particle discrimination can be performed on an event by event basis, based on the detected amount of scintillation light. This reduces the dominating β/γ backgrounds even further. The experimental setup of CRESST is described in chapter 2. This work focuses on two analyses carried out on the CRESST-II phase 2 and CRESST-III phase 1 data with an additional, supportive analysis reported in the appendix. While this includes a dark matter analysis of the recent CRESST-III phase 1 data, the main focus of the thesis lies on the understanding of backgrounds and detector effects. Chapter 3 introduces the general steps that are required for a CRESST analysis, namely calculation of parameters, application of data selection criteria, energy reconstruction, determination of survival probability and, if applicable, dark matter limit calculation. Chapter 4 deals with an α analysis of CRESST-II phase 2 data. An α analysis allows to determine precisely the internal contamination of target crystals from the three prevalent natural radioactive decay chains. The decays are fully embedded in the target crystal of a CRESST detector and thus produce mono-energetic peaks. This information is valuable input for simulation efforts to understand background contributions at low energies. Challenges arise from the high energies O(MeV) of α decays compared to the keV energies the detector is optimized for. Several steps were taken to tackle these challenges, culminating in improved precision and understanding of the observed α spectra compared to previous results from a subset of the same data. The results have already been used in an improved background simulation within the CRESST group. Chapter 5 comprises the dark matter analysis and thereby the main result of the CRESST-III phase 1 measurement campaign. For the first time, data from the newly developed continuous data acquisition was analyzed. This allowed for a significantly improved way of determining the survival probability of dark matter events by simulating them directly on the saved data stream. Furthermore, the energy reconstruction is now based on the optimum filter method that offers an improved resolution compared to the previous standard event fit. Another novelty in CRESST-III phase 1 was the partial instrumentation of the holding structure with thermometers, allowing to identify certain types of background events. The analysis focuses on one detector module with a nuclear recoil threshold of 30.1 eV which exceeds the 100 eV design goal by a factor of three. At low energies < 200 eV, an exponentially rising excess of events is observed that cannot be explained by the known backgrounds. This excess already severely limits the sensitivity in terms of dark matter particle-nucleon interaction cross-section in the corresponding mass range. Nonetheless, leading limits for dark matter masses between 160 MeV/c^2 and 1.8 GeV/c^2 could be derived by the CRESST group from the event spectra that were extracted in this work. Special focus and further analysis effort are given to the identification of possible origins of the excess events. A crucial part in this regard is a full analysis of all event classes expected from the detector geometry. Different event classes are caused by energy depositions in the different parts of a detector module. The results of such an analysis with regard to the excess are also summarized in chapter 5, with the analysis itself described in appendix A. This analysis revealed that the main target detects significant amounts of scintillation light from the scintillating holding structure which has not been observed to this extent in previous measurements. However, none of the investigated event types is a viable candidate for the excess.
