Maximum Likelihood Signal Extraction Method Applied to 3.4 years of CoGeNT Data

TL;DR

This paper applies a maximum likelihood analysis to 3.4 years of CoGeNT data to search for dark matter signals, incorporating background characterization and seasonal variations, but finds no statistically significant evidence.

Contribution

It introduces a 2D likelihood method including pulse rise-time and seasonal variation to analyze CoGeNT data for dark matter signals, extending prior analyses.

Findings

Best-fit dark matter signal close to previous results

Signal significance below 2 sigma

No conclusive dark matter detection

Abstract

CoGeNT has taken data for over 3 years, with 1136 live days of data accumulated as of April 23, 2013. We report on the results of a maximum likelihood analysis to extract any possible dark matter signal present in the collected data. The maximum likelihood signal extraction uses 2-dimensional probability density functions (PDFs) to characterize the anticipated variations in dark matter interaction rates for given observable nuclear recoil energies during differing periods of the Earth's annual orbit around the Sun. Cosmogenic and primordial radioactivity backgrounds are characterized by their energy signatures and in some cases decay half-lives. A third parameterizing variable -- pulse rise-time -- is added to the likelihood analysis to characterize slow rising pulses described in prior analyses. The contribution to each event category is analyzed for various dark matter signal hypotheses including a dark matter standard halo model and a case with free oscillation parameters (i.e., amplitude, period, and phase). The best-fit dark matter signal is in close proximity to previously reported results. We find that the significance of the extracted dark matter signal remains well below evidentiary at 1.7 $\sigma$.