Observable spins in gravitational waves from compact binary mergers

TL;DR

This work analyzes how the effective inspiral spin $\chi_{eff}$ is inferred from gravitational-wave observations of compact binaries and shows that noise and low signal strength can bias measurements, sometimes producing nonzero $\chi_{eff}$ for intrinsically non-spinning systems. Using Bayesian parameter inference on injections drawn from GWTC-4.0 with Gaussian noise, the authors demonstrate that the inferred $\chi_{eff}$ can be systematically shifted by noise realizations, particularly for faint events, and that priors can further influence the results. Population-level comparisons between spinning and non-spinning injections suggest no clear separation at low SNR, though hints emerge at higher SNR; overall the findings support the possibility that observable binaries have minimal $|\chi_{eff}|$, with larger, louder samples needed to constrain the true spin distribution. The study highlights implications for astrophysical inferences and emphasizes the importance of detector improvements and longer observing runs to robustly characterize spin populations in gravitational-wave sources.

Abstract

We investigate the measurability of effective inspiral spin in the detectable compact binary mergers using gravitational-wave observations. Measurements from the latest gravitational-wave transient catalog do not rule out the existence of binary systems with non-zero effective spins. However, we observe an apparent correlation between the inferred effective inspiral spin and the loudness of the gravitational-wave events-loud events typically have close-to-zero effective spins whereas fainter events tend to be inferred with relatively arbitrary effective spins. Through simulations, we demonstrate that non-negligible effective spins can be systematically inferred from non-spinning systems at small signal strengths. These two observations support the possibility that the effective spin magnitudes in the observable compact binaries are generally small. Future detections can have potential impact on the understanding of their population and other astrophysical inferences.

Figures (8)

• Figure 1: Distribution of the inferred effective inspiral spin ($\chi\textsubscript{eff}$) with the network SNR for simulated non-spinning merger events (in grey). The same is shown for the confidently detected GW events (in black), which contains events reported in GWTC-4.0, as exceptional and other events gwoscevents, collectively referred to as GWTC-4.0+. A circle indicates the inferred value and the error bars represent the 90% credible interval of the measurement. Essentially, note that non-negligible $\chi\textsubscript{eff}$ values can be inferred from non-spinning sources at generally small SNRs, as discussed later in the text.
• Figure 2: Visualization of the effects of aligned-spins and that of the Gaussian noise on the GW strain in the time domain from an arbitrary simulated $(77.2, 57.6)$$\mathrm{M}_{\odot}$ source located at $\sim$ 1000 Mpc. The aligned-spins are such that $\chi\textsubscript{eff}=0.14$ (shown in dark blue). The GW strain for the non-spinning signal with a single realization of Gaussian noise is shown (in light blue).
• Figure 3: Self-consistency of quoted credible intervals for the (intrinsic) source parameters used to calculate $\chi\textsubscript{eff}$. The grey shaded regions indicate up to $3\sigma$-uncertainty.
• Figure 4: Simulated effect of detector noise realizations on the inference of effective inspiral spin ($\chi\textsubscript{eff}$). Each column represents a given injected source with varied network SNRs. Each curve represents a unique noise realization for a given injected source. The network SNRs indicate the injected SNRs averaged over the multiple noise realizations.
• Figure 5: Distribution of the estimated $\chi\textsubscript{eff}$ with the network SNR for a subset of non-spinning injections in Gaussian noise and in zero noise. The zero noise case indicates the ideal (or a measure of the best-achievable) estimate. A distinct noise seed is used for each injection in the Gaussian noise case. We observe that fainter events can be inferred with relatively larger non-zero $\chi\textsubscript{eff}$'s. We also note a larger scatter in the estimated values in the presence of noise. The analysis is repeated for the same set of injections, and hence for the same population of the observable sources, in the upcoming observing scenarios. It suggests that the effect is expected to be reduced in the O5-like observing scenario and beyond, given the overall greater SNRs recovered for the detected sources. The credible intervals are not shown here for visual clarity but they generally scale with the loudness.