Studying single-electron traps in newly fabricated Skipper-CCDs for the Oscura experiment using the pocket-pumping technique

TL;DR

This study characterizes defects in newly fabricated Skipper-CCDs for the Oscura dark matter experiment, using pocket-pumping to measure trap properties and assess their impact on low-energy background noise.

Contribution

It introduces a trap characterization method for Skipper-CCDs and links defect types to fabrication batches, informing low-background sensor development.

Findings

Defects produce deferred charge detectable next to particle tracks.

Trap properties vary with fabrication batch, not gettering method.

Exposure-dependent single-electron rate measured at 1.8e-3 e-/pix/day.

Abstract

Understanding and characterizing very low-energy ($\sim$eV) background sources is a must in rare-event searches. Oscura, an experiment aiming to probe electron recoils from sub-GeV dark matter using a 10-kg skipper-CCD detector, has recently fabricated its first two batches of sensors. In this work, we present the characterization of defects/contaminants identified in the buried-channel region of these newly fabricated skipper-CCDs. These defects/contaminants produce deferred charge from trap emission in the images next to particle tracks, which can be spatially resolved due to the sub-electron resolution achieved with these sensors. Using the trap-pumping technique, we measured the energy and cross section associated to these traps in three Oscura prototype sensors from different fabrication batches which underwent different gettering methods during fabrication. Results suggest that the type of defects/contaminants is more closely linked to the fabrication batch rather than to the gettering method used. The exposure-dependent single-electron rate (SER) of one of these sensors was measured $\sim$100~m underground, yielding $(1.8\pm 0.3)\times10^{-3}~e^-$/pix/day at 131K. The impact of the identified traps on the measured exposure-dependent SER is evaluated via a Monte Carlo simulation. Results suggest that the exposure-dependent SER of Oscura prototype sensors would be lower in lower background environments as expected.