Logo image
Back
Characterization of radon progeny in EXO-200 using machine learning algorithms
Dissertation   Open access

Characterization of radon progeny in EXO-200 using machine learning algorithms

Erica Smith
Doctor of Philosophy (Ph.D.), Drexel University
Feb 2016
DOI:
https://doi.org/10.17918/etd-6681
pdf
Smith_Erica_201621.08 MBDownloadView
PDF Open Access Open Access (License Unspecified)

Abstract

Neutrinos--Beta decay Neutrinos--Experiments Physics
Neutrinoless double beta decay (0[nu][beta][beta]) is a rare, second-order process that occurs in certain isotopes for which beta decay is energetically forbidden. EXO-200 is a 0[nu][beta][beta] experiment with 110 kg of active liquid xenon (LXe) isotopically enriched in ¹³⁶Xe. EXO-200 detects events using a combination of scintillation and ionization signals, which allows for excellent particle discrimination. However, events with a low ionization signal cannot be fully characterized with the current analysis framework. To fill in these gaps, we introduce a boosted decision tree regressor as a new tool to characterize events in the detector. We focus on [alpha] decays of ²²²Rn and its progeny, which have low ionization signals that often fall below the threshold for position reconstruction. Using information gained from this technique, we confirm previous results for the ²¹⁸Po ion velocity and improve previous results for the ²¹⁸Po ionization fraction. We also investigate events that occur near the walls of the vessel. These events have no ionization signal and therefore cannot be characterized with any existing technique in the analysis framework. By investigating these events we determine that they are not distributed uniformly throughout the detector, which may point to charging up of the plastics inside the LXe vessel or a "hot spot" on the plastic due to contamination during cleaning and installation.

Metrics

75 File views/ downloads
33 Record Views

Details

Logo image