This paper asks whether some of the loud, heavy signals seen by the LIGO‑Virgo‑KAGRA network might come from close, one‑time black hole encounters rather than the usual slowly spiraling binaries. A close hyperbolic encounter is a quick, fly‑by interaction that produces a short, broadband burst of gravitational waves. Those bursts look very different from the stretched, chirping signals expected from quasi‑circular inspirals that slowly coalesce over many orbits.

The authors reanalyzed 38 high‑mass events from the public LIGO‑Virgo‑KAGRA catalogs using a modern waveform model called TEOBResumS‑Dalí. This model can be configured to describe either quasi‑circular, precessing binaries (the usual case) or hyperbolic, unbound encounters (dynamical captures and scatterings). For each event they ran parameter estimation with the RIFT pipeline and computed Bayes factors, which are a statistical way to say how much the data prefer one model over another.

Most of the 38 events were better described by the quasi‑circular, precessing model. Two notable exceptions emerge in the study. For GW190521 the data prefer the hyperbolic, dynamical‑capture description, with a reported Bayes factor log value ln B_hyp_prec = 3.71^{+0.11}_{-0.11}, indicating the hyperbolic model fits that signal better than the quasi‑circular, precessing alternative. By contrast, GW231123 shows a strong preference for the quasi‑circular, precessing model, with ln B_hyp_prec = -15.80^{+0.24}_{-0.24} (a negative value means the quasi‑circular model is preferred). The authors checked these conclusions using other TEOBResumS‑Dalí settings (for example including or excluding precession and eccentricity) and with a separate surrogate model built from numerical relativity, NRSur7dq4.

This work matters because hyperbolic and other dynamical encounters are expected in dense stellar environments, such as star clusters or near active galactic nuclei. Those environments can produce black holes in mass ranges that are hard to make in isolated stars, including masses in the so‑called pair‑instability mass gap (roughly tens to a few hundred solar masses). Detecting a genuine dynamical capture would therefore give direct clues about how some heavy black holes form. It would also change how we interpret the population of reported events and the expected gravitational‑wave background.