Study of cosmic-ray composition with Imaging Atmospheric Cherenkov Telescopes

During the (last) three decades, the first and second generations of Imaging Atmospheric Cherenkov Telescopes (IACTs), such as Whipple Observatory, VERITAS, HESS and MAGIC, have provided measurements of several TeV gamma-ray sources. Experiments like the Cherenkov Telescope Array (CTA) will be the next-generation IACTs in the southern and northern hemispheres, offering better sensitivity, angular resolution, and larger collection area than the current generation. One of the CTA's aims is to make significant progress in detecting high-energy cosmic rays, providing insight into cosmic ray propagation and acceleration. The work done in this thesis is twofold. The first part proposes two methods to measure a mass-sensitive parameter of nuclei-initiated air showers: the depth of the shower maximum Xmax.  We also explored the possibility of reconstructing the complete longitudinal shower profile using IACTs even beyond 150 m from the shower impact point. We tested our method using proton and iron simulations for Medium-Sized Telescope (MST) detections showing resolution values around ~35 gcm^2 for both species in the energy range from 10 TeV to 300 TeV. The second part corresponds to the analysis of CTA simulations to separate iron-initiated from proton-initiated showers. We addressed the challenge of making a binary classification of iron and proton events using deep learning algorithms. The best performance for the binary classification was given in the scenario of inputs: image + Xmax + energy. In this scenario, we achieved an iron identification efficiency better than 90%.

Dissertation: https://hdl.handle.net/11370/18a8e91d-26d5-4bf8-97d7-c87fdd38e413