Characterisation and Calibration of ZEPLIN III - A Dark Matter Detector

Abstract

The ZEPLIN III liquid xenon dark matter detector is designed to potentially discover the WIMP - a supersymmetric galactic dark matter candidate. This thesis presents experimental results of the ZEPLIN III commissioning studies, in preparation for the first and second underground science runs. Data acquired on the surface, at the Imperial College London laboratories, were used to characterise the instrument’s response in terms of light yield (LY) and single photoelectron (SPE) spectra. A zero-field LY was measured as 7.42±0.37 phe/keV and 18.12±0.91 phe/keV in dedicated single- and dual-phase high yield configurations, respectively, consistent with Monte Carlo simulations. Mean SPE measured pulse areas ranged from 41.78±1.55 Vps to 52.37±1.59 Vps, depending on the method employed. A 3-D position reconstruction was verified and, significantly, no evidence of a potentially-contaminating background -population was observed. This study directly lead to critical development of the DAQ software and hardware configuration. The PMT array was confirmed as responsive and, crucially, the particle discrimination principle was demonstrated. Zero-field LYs of (4.6-4.7) ±0.5 phe/keV were recovered from the centre of the chamber, exceeding simulation predictions. With-field (3.01 kV/cm in the liquid) LYs of (1.2-1.8) ±0.3 phe/keV from the liquid scintillation (S1) and an electroluminescence yield (S2) of (98-140) ±35 phe/keV from the gas phase were also determined. ZEPLIN III was deployed in the Boulby Underground Laboratory, UK and demonstrated successful operation at high field (up to 3.79 kV/cm in the liquid), in situ. An alternative Poisson method for obtaining single photoelectron distributions was developed by the ZEPLIN collaboration. The origin of long- events in surface data was investigated and ultimately resolved as an artefact of early versions of the data reduction software. An S1 zero-field LY of 4.72±0.10 phe/keV, obtained with a 57Co external source, was recovered for the centre of the chamber. The instrument’s energy resolution was evaluated and a novel parameterisation approach, developed by the author, yielded sigma=1.08±0.06 p E(keV ) with a dominant stochastic term. A ‘flat-fielding’ method was established, proving to minimise the resolution significantly, yielding 8.6% and 7.3% for S1 and S2, respectively, in the fiducialised anti-correlated energy channel. This flat-fielding recipe, along with construction of the light collection correction matrices, formed the basis of the final procedures subsequently applied to first science run data-sets.