Cascade type identification in IceCube and an application in a search for new physics
Date issued
Authors
Editors
Journal Title
Journal ISSN
Volume Title
Publisher
License
Abstract
In 2013, the first detection of an astrophysical neutrino flux by the IceCube neutrino telescope marked a tremendous breakthrough in the field of neutrino astronomy. Since then, dedicated analyses try to extract information on the origin of the cosmic messengers from the accumulating data. Efforts to trace the neutrino flux back to a dominant source class have not yet succeeded, leaving many questions on the nature of astrophysical particle accelerators unanswered and providing room for alternative explanations for the origin of IceCube's high energy events.
Besides the spatial distribution, the flavor composition of the astrophysical neutrino flux plays a key role in unraveling its production mechanisms. While the long tracks of Cherenkov light generated in muon neutrino charged current interactions can be well discriminated from the spherical shower - or cascade - pattern produced in all other neutrino interactions, an identification of the cascade-inducing neutrino flavor and interaction channel has yet to be accomplished. In this thesis, a method utilizing the delayed light from shower-internal neutron capture events - the so-called neutron echo - has been developed, experimentally implemented and applied to determine the shower type composition of a high energy cascade sample. This is the first time it has ever been attempted to classify the shower type of high energy cascades in IceCube or any other neutrino telescope and the results of this work show great promise.
Since the nature of the showers contained in IceCube's well-known High Energy Starting Event (HESE) sample has not been assessed before, an exotic origin could not be excluded. In this work, the newly developed method is used to probe two boosted dark matter scenarios, which predict an either solely neutral current- or electron neutrino charged current-like cascade sample. Although the so far small sample size limits the power of the analysis, a Bayesian evaluation of the results already disfavors the probed models at more than 90% credible level.