Searching for cosmological gravitational-wave backgrounds with third-generation detectors in the presence of an astrophysical foreground

TL;DR

This paper explores advanced subtraction techniques to enhance the detection of the cosmological gravitational-wave background with third-generation detectors, overcoming astrophysical foreground challenges.

Contribution

It introduces a subtraction-noise projection method that significantly reduces residuals from astrophysical foregrounds in gravitational-wave data analysis.

Findings

Subtraction-noise projection improves sensitivity to cosmological backgrounds.

Residuals are not limited by detector noise but by undetectable astrophysical foreground fraction.

Method resolves misconceptions about residual noise in Gaussian parameter estimation.

Abstract

The stochastic cosmological gravitational-wave background (CGWB) provides a direct window to study early universe phenomena and fundamental physics. With the proposed third-generation ground-based gravitational wave detectors, Einstein Telescope (ET) and Cosmic Explorer (CE), we might be able to detect evidence of a CGWB. However, to dig out these prime signals would be a difficult quest as the dominance of the astrophysical foreground from compact-binary coalescence (CBC) will mask this CGWB. In this paper, we study a subtraction-noise projection method, making it possible to reduce the residuals left after subtraction of the astrophysical foreground of CBCs, greatly improving our chances to detect a cosmological background. We carried out our analysis based on simulations of ET and CE and using posterior sampling for the parameter estimation of binary black-hole mergers. We demonstrate the sensitivity improvement of stochastic gravitational-wave searches and conclude that the ultimate sensitivity of these searches will not be limited by residuals left when subtracting the estimated BBH foreground, but by the fraction of the astrophysical foreground that cannot be detected even with third-generation instruments, or possibly by other signals not included in our analysis. We also resolve previous misconceptions of residual noise in the context of Gaussian parameter estimation.