Improving Photoelectron Counting and Particle Identification in Scintillation Detectors with Bayesian Techniques

TL;DR

This paper introduces a Bayesian method for accurately identifying individual photoelectrons in scintillation detector waveforms, enhancing energy resolution and particle identification for dark matter detection.

Contribution

The paper presents a novel Bayesian technique for photoelectron counting that improves particle identification and energy resolution in scintillation detectors without requiring waveform deconvolution.

Findings

Improved energy resolution in calibration data.

Enhanced particle identification in dark matter detector simulations.

Better background rejection compared to simpler methods.

Abstract

Many current and future dark matter and neutrino detectors are designed to measure scintillation light with a large array of photomultiplier tubes (PMTs). The energy resolution and particle identification capabilities of these detectors depend in part on the ability to accurately identify individual photoelectrons in PMT waveforms despite large variability in pulse amplitudes and pulse pileup. We describe a Bayesian technique that can identify the times of individual photoelectrons in a sampled PMT waveform without deconvolution, even when pileup is present. To demonstrate the technique, we apply it to the general problem of particle identification in single-phase liquid argon dark matter detectors. Using the output of the Bayesian photoelectron counting algorithm described in this paper, we construct several test statistics for rejection of backgrounds for dark matter searches in argon. Compared to simpler methods based on either observed charge or peak finding, the photoelectron counting technique improves both energy resolution and particle identification of low energy events in calibration data from the DEAP-1 detector and simulation of the larger MiniCLEAN dark matter detector.