This paper shows that a small, spinning object falling into a much larger nonrotating black hole can behave chaotically even when the small object’s spin is within realistic astrophysical limits. The authors model the smaller body as a spinning “test particle” using the full Mathisson–Papapetrou–Dixon (MPD) equations. These equations describe how a spinning body moves in curved spacetime; the authors close the system with the Tulczyjew–Dixon spin supplementary condition, and they keep nonlinear-in-spin effects rather than linearizing them.

To isolate the effect, the team studies motion around a Schwarzschild black hole — a simple, nonrotating spacetime. They reduce the problem to two degrees of freedom by using the spacetime’s spherical symmetry and follow orbital families built near the geodesic homoclinic orbit (the separatrix between bound and plunging geodesic motion). They probe the phase space with standard diagnostics for chaos, including Poincaré sections, rotation curves, and Lyapunov indicators, and then compute gravitational-wave signals with a numerical kludge waveform method to see how chaos shows up in observable data.

Their main physical finding is that chaos is not limited to unrealistically large secondary spins. Instead, chaotic motion persists across the astrophysically realistic spin range they examine. Importantly, chaotic and nearby regular orbits can look very similar in the time domain and share the same dominant spectral peaks. The clear difference appears in the frequency domain: chaotic signals develop a much denser set of weaker frequency components between the main peaks. The authors quantify this with a local spectral-flatness measure and find it to be several hundred times larger for chaotic signals than for neighboring regular ones.

The paper also reports that a very small change in the secondary spin can flip the system between regular and chaotic behavior. A change as small as 1% of the secondary’s maximal physically allowed spin can move the system into chaos and produce waveforms that are distinguishable at the detector level. The characteristic strain of these signals remains in the millihertz band relevant for planned space-based detectors such as LISA, Taiji, and TianQin, and the overall signal-to-noise ratios remain comparable. However, the overlap between waveforms drops sharply when a spin change drives the system from regular to chaotic, suggesting measurable differences for data analysis.