Item

Abstract

This thesis presents a novel spatial likelihood analysis for Imaging Atmospheric Cherenkov Telescopes (IACTs) and its use to analyse observations of the gamma-Cygni supernova remnant (SNR) with the MAGIC telescopes, a system of two IACTs. SNRs are the prime candidate source for the origin of the galactic component of cosmic rays (CRs). These objects are sufficiently extended to be resolved with gamma-ray telescopes. This allows the determination of different acceleration regions of a source, but poses issues for the current analysis approach for IACT data. IACTs detect the Cherenkov light generated in air showers, which are cascades of energetic particle that result from the interaction of gamma-rays with the molecules in the atmosphere. Currently, the emission from a source is determined using the aperture photometry approach, in which the number of gamma-ray events from the source region is compared against a source-free background control region. In the case of superimposed emission regions, an event count cannot be attributed to one emission region. Furthermore, extended objects or objects of complex morphology make the definition of the source region a difficult task. These issues can be overcome by a spatial likelihood analysis of the skymaps of IACTs. In this approach, a user-defined source template is convolved with the instrument response functions (IRFs) and the "realistic" model fitted to the event count maps via a Poissonian likelihood fit. The data analyses of space-based gamma-ray telescopes, such as the Fermi Large Area Telescope (LAT), are based on this technique. For IACTs the determination of the IRFs, however, is a challenging task: because the atmosphere is part of the detector, the IRFs cannot be measured in the laboratory but need to be computed from Monte-Carlo events for each observation individually. This thesis presents SkyPrism, a software package performing such an analysis on MAGIC data including the accurate determination of the IRFs. Using SkyPrism it was possible to analyse observations of the ~7000 year old gamma-Cygni SNR taken with MAGIC between 2015 and 2017. CRs are accelerated and confined in the shock region by magnetic turbulences ahead and behind the shock, making the level of turbulence an important ingredient of the acceleration process. Only a small high energetic fraction of CRs may escape the fast shocks of young SNRs (<3000 years), whereas in the case of old SNRs (>10000 years) almost all CRs have already escaped. I studied the escape of CRs from the shock into the interstellar medium using 85 hours of MAGIC data and 9 years Fermi-LAT data covering the energy range from 5 GeV to 5TeV. Using the theoretical model of the diffusive shock acceleration, I determined that the maximum energy of the CRs confined in the shock region decreases faster with the lifetime of the SNR than expected and that the level of turbulence is not constant over the lifetime of the SNR.