Low-Energy Radiative Backgrounds in CCD-Based Dark-Matter Detectors

TL;DR

This paper develops theoretical models and simulations to understand low-energy backgrounds in CCD-based dark-matter detectors, focusing on secondary radiation processes and their impact on detector sensitivity.

Contribution

It introduces detailed simulation methods for secondary radiation backgrounds in CCD detectors and assesses their effects on dark-matter search sensitivities.

Findings

Cherenkov photons account for about 40% of single-electron events.

Radiative recombination is negligible in the fiducial model but significant in the extreme model.

Background mitigation can improve future detector sensitivity by up to two orders of magnitude.

Abstract

The reach of sub-GeV dark-matter detectors is at present severely affected by low-energy events from various origins. We present the theoretical methods to compute the single- and few-electron events that arise from secondary radiation emitted by high-energy particles passing through detector materials and perform simulations to quantify them at (Skipper) CCD-based experiments, focusing on the SENSEI data collected in the MINOS cavern at Fermilab. The simulations account for the generation of secondaries from Cherenkov and luminescent recombination; photo-absorption, reflection, refraction and thin-film interference in detector materials; roughness of the interfaces and the dynamics of charges and partial charge collection (PCC) in the doped CCD-backside. We consider several systematic uncertainties, notably those stemming from the backside charge-diffusion modeling, which we estimate with a "fiducial'' and an "extreme'' model, with the former model presenting better agreement with PCC data. We find that Cherenkov photons constitute about 40% of the observed single-electron events for both models; radiative recombination rates are negligible for the fiducial model, but can dominate over the Cherenkov rates for the extreme model. We also estimate the fraction of 2-electron events from 1-electron event same-pixel coincidences, finding that the entire 2-electron rate can be explained by coincidences of radiative events and spurious charge. Accounting for backgrounds, we project the sensitivity of future Skipper-CCD-based experiments to different dark-matter models. For light-mediator models with dark-matter masses of 1, 5, and 10 MeV, we find that future experiments with 10-kg-year exposures and successful background mitigation could have a sensitivity that is larger by 9, 3, and 2 orders of magnitude, respectively, when compared to an experiment without background improvements. (abridged)