Detector configuration of DECIGO/BBO and identification of cosmological neutron-star binaries

TL;DR

This work quantifies how the geometry of DECIGO/BBO detectors affects the identification and subtraction of cosmological NS/NS binaries, a critical step for revealing a primordial GW background. Using analytic minimum-SNR calculations for static configurations and Monte Carlo simulations that incorporate detector motion, the authors show that a four-unit star-of-David network can identify NS/NS binaries up to $z\le 5$ with $\mathrm{SNR}>25$, but the required network sensitivity depends on the configuration. The optimal three-plane arrangement with opening $\alpha_3\approx126.4^{\circ}$ yields a maximum dimensionless minimum-SNR ratio $G\approx0.771$, and the infinite-detector limit approaches $G\approx0.895$, implying substantial gains from symmetry and motion averaging. To reliably subtract all NS/NS foregrounds at $z\le5$ with a threshold $\rho_{\mathrm{thr}}=20$, the network sensitivity must satisfy $\bar{\rho}_5>25.9$, meaning DECIGO would need roughly a $2.5\times$ sensitivity improvement over its baseline, while BBO already meets this for $\rho_{\mathrm{thr}}=20$ but not for higher thresholds. The results provide concrete guidance on detector geometry and sensitivity requirements for enabling PGWB detection via effective foreground subtraction.

Abstract

The primary target for the planned space-borne gravitational wave interferometers DECIGO/BBO is a primordial gravitational wave background (PGWB). However there exist astrophysical foregrounds and among them, gravitational waves from neutron star (NS) binaries are the solid and strong component that must be identified and subtracted. In this paper, we discuss the geometry of detector configurations preferable for identifying the NS/NS binary signals. As a first step, we analytically estimate the minimum signal-to-noise ratios (SNRs) of the binaries for several static detector configurations that are characterized by adjustable geometrical parameters, and determine the optimal values for these parameters. Next we perform numerical simulations to take into account the effect of detector motions, and find reasonable agreements with the analytical results. We show that, with the standard network formed by 4 units of triangle detectors, the proposed BBO sensitivity would be sufficient in receiving gravitational waves from all the NS/NS binaries at $z\le 5$ with SNRs higher than 25. We also discuss the minimum sensitivity of DECIGO required for the foreground identification.

Figures (14)

• Figure 1: The default configuration of BBO or DECIGO with $\alpha_3=120^\circ$. Totally four units of triangle-like detectors will be operated. Two of them are nearly co-aligned to form a star-of-David constellation for the correlation analysis. Two outrigger ones are used to improve localization of individual astrophysical sources.
• Figure 2: One unit of BBO (DECIGO) detector and the orientations of two effectively L-shaped interferometers I and II defined in Eq. (\ref{['tdi']}).
• Figure 3: The solid curves show the noise spectra for DECIGO (thick red) and BBO (thin blue). Note that these noise curves are the sky-averaged ones (a factor of $\sqrt{5}$ larger than the original ones). The (black) thin dotted line represents the expected total amplitude of NS/NS foreground and the (purple) thick dotted line represents the primordial GW background corresponding to $\Omega_{\mathrm{GW}}=10^{-16}$. The (black) dashed line at $f=0.2$Hz indicates the upper frequency cutoff of WD/WD confusion noises.
• Figure 4: We use two types of coordinates: (i) a barycentric frame $(\bar{x},\bar{y},\bar{z})$ tied to the ecliptic and centered in the solar system barycenter, (ii) an detector frame $(x,y,z)$ attached to the detector as in Fig. \ref{['ae']}.
• Figure 5: This figure shows the accumulated averaged SNR of each (1.4+1.4)$M_\odot$ NS/NS binary against time to coalesce. The (red) circular plots, the (blue) triangular plots, the (green) square plots and the (black) crosses correspond to the one with the source redshift $z=5$,3,1 and 0, respectively. These values have been normalized so that the accumulated SNRs become 1 at the time corresponding to $f$=100Hz.