Neural Networks for Background Rejection in DEAP-3600 Detector

摘要

Understanding the nature of dark matter is a major challenge in modern Physics. During the last decades, several detectors have been built to detect the signal of interactions between dark matter and ordinary matter. DEAP-3600 is a liquid-argon detector with 3.3 tons of active volume placed at SNOLAB (Canada), 2.1 km underground. This detector aims at searching for weakly interacting massive particles (WIMP) as dark matter candidate. The interaction of WIMPs with a nucleus of argon is expected to produce scintillation light with a particular signature. In this work, a Multilayer Perceptron is used for separating this signature from other background signatures, and specifically from a frequent background originated in the neck of the detector. As a consequence of this work, an improvement in the classification of neck events is achieved, reaching a mean acceptance of 44.4% for a background rejection power of 99.9% of neck events for the same sample of simulated events.