Persistence Characterisation of teledyne H2RG detectors

Abstract

Image persistence is a major problem in infrared detectors, potentially seriously limiting data quality in many observational regimes. The problem manifests itself as remnant images that can persist for several days after a deep exposure. In this study, the persistence behavior of three 5.3um cutoff H2RGs has been characterised using a low-background cryostat with LED light sources. Persistence charge de-trapping was measured over several hours following a wide range of exposure levels and exposure times. This data was then analysed to yield charge trapping and de-trapping spectra which present graphically the trap density as a function of their time constants. These spectra show the detector behavior in a very direct way and offer a natural metric for comparing different devices. It is hoped that the trap time constant spectra for each detector can be used in an analysis pipeline to remove persistence artifacts based on the recent exposure history of the detector. The study confirmed that the charge traps responsible for persistence must be present in the depletion region of the pixel, however, two trap populations were revealed. One of these captures charge within milliseconds and then releases it over many hours. The second population is less problematic with fairly similar trapping and de-trapping time constants. Large differences in persistence magnitude and trap spectra have been found even between devices with near-consecutive serial numbers. Lower temperatures resulted in lower persistence both in terms of total trapped charge and the time taken for that charge to decay. Limiting the full-well by reducing pixel bias voltage also had a beneficial effect. Previously proposed mitigation techniques including "global reset de-trapping" and "night light" illumination were tried but found to be ineffective.