How can extreme-mass-ratio inspirals with a spinning secondary be calculated efficiently?

Secondary spin enters at first postadiabatic order

An extreme-mass-ratio inspiral (EMRI) is a stellar-mass compact object spiraling into a massive Kerr black hole with a mass ratio of roughly 10⁻⁷ to 10⁻⁴. Because the small body completes 10⁴ to 10⁶ orbital cycles in the strong-field region before plunging, the emitted gravitational-wave phase must be tracked accurately over years of observation. The paper targets the piece of that modeling in which the small body itself carries spin.

In the multiscale expansion used here, orbital quantities split into slowly evolving parameters and rapidly evolving phases. The phases are written as an adiabatic term plus a first postadiabatic (1PA) correction, and the paper separates the 1PA forcing functions into a part proportional to the dimensionless spin component parallel to the orbital angular momentum, χ∥, and a remaining self-force part. The new framework computes the χ∥ piece for fully generic, eccentric, off-equatorial orbits. That scope matters: the paper states that neglecting the 1PA term introduces significant biases in parameter estimation, and that EMRIs from dynamical capture or multibody interactions are expected to span a broad range of eccentricities and spin orientations, with misalignment driving orbital-plane precession.

Analytic trajectories replace numerical orbit solving

The central simplification comes from a recently found analytic solution for the motion of spinning particles in Kerr spacetime. A linear shift of the worldline produces a "virtual" worldline whose equations of motion are separable and take the same form as geodesic equations, but with shifted constants of motion: the energy, the axial angular momentum, and the Carter-like constant each receive a correction linear in s∥, the spin component parallel to the orbital angular momentum.

This lets the flux calculation reuse functions that already describe geodesic motion. In the Teukolsky amplitude expression, the spin correction appears only in a few places: the time-frequency factor in the denominator, the mode frequency, the source functions, and the time-shift in the exponential. The paper writes the linear part of the amplitude as a geodesic amplitude plus an integral over geodesic quantities, so the expensive step of reconstructing the metric perturbation and regularizing it at the particle position is avoided. The author notes that only the parallel spin components are needed at 1PA order, because orthogonal components enter at higher orders.

Flux-balance laws give all three rates of change

The framework does not compute the rates of change of the constants of motion directly. Instead it uses flux-balance laws, recently proven, that relate those average rates to asymptotic gravitational-wave fluxes. Energy and angular momentum fluxes are sums over mode numbers of the Teukolsky amplitude combination Zlmnk weighted by the mode frequency ωmnk and by the azimuthal number m, respectively.

The Carter constant is the harder case, because there is no literal flux of it to infinity or to the horizon. The paper expresses its rate of change through the polar action Jθ and its derivatives with respect to energy, angular momentum, and the Carter constant, packaged into auxiliary functions Ξz, Ξtz, and Ξϕz. The geodesic parts of these functions are the Mino-time frequencies Υz, Υtz, and Υϕz; the linear spin corrections are given in closed form, with two of the three relations checked numerically across parameter space because of their complexity. The resulting Carter-constant rate is again a sum over modes of Zlmnk times a combination of geodesic frequencies and spin terms.

The paper validates the new amplitude and flux implementation against the independent Spinning-Body-Hamilton-Jacobi code of Piovano et al. for selected mode numbers. Using a fourth-order finite-difference step of Δs∥/M = 10⁻³, the largest relative error reported is 3.6×10⁻⁴, for an infinity flux mode with n = 20 and k = −1; that mode is subdominant, and its relative error falls to 5.9×10⁻⁶ when the comparison code's initial precision is raised from 45 to 60 digits.

Nearly spherical inspirals show gauge-independent phase shifts

The framework is then applied to nearly spherical inspirals, using fluxes from earlier work that supplied the geodesic part and the linear spin part in a fixed turning-points gauge, together with derivatives of the geodesic fluxes with respect to the orbital parameters p and x. The evolution equations for the parameters and phases are linearized as a geodesic term plus s∥ times a correction.

The orbital parameters p̃1 and x̃1 evolve differently from the earlier calculation because the spin gauges differ. The gauge-invariant phase shifts Φ¹z and Φ¹ϕ, however, agree within numerical error. The paper plots the absolute difference between phase shifts computed in the two gauges and finds the error peaks before the end of the inspiral at around 10⁻² rad, with the numerical error attributed mostly to interpolation of the forcing functions; increasing grid resolution makes the differences converge toward zero. The author emphasizes that only the linear-in-spin parts should agree, since total nonlinearized phases and waveforms would differ at subleading orders.

What the gauge check does and does not establish

The gauge-independence result is a consistency check on the linear-in-spin phase, not a demonstration that the full waveform is gauge independent. It also rests on a single orbital regime: nearly spherical inspirals. The paper's stated scope is generic, eccentric, off-equatorial EMRIs at 1PA order, but the numerical demonstration of inspiral evolution is carried out in the nearly spherical case, and the flux validation covers selected mode numbers rather than the full parameter space.

Two further boundaries follow from the construction. The expansion keeps only terms linear in the secondary spin, dropping quadratic and higher contributions as higher order in s; and the 1PA treatment neglects evolution of the primary mass and spin, as well as orbital resonances at 0.5PA order, which the paper says can be treated without secondary-spin effects. The framework therefore supplies the secondary-spin contribution to 1PA waveforms, not a complete waveform model on its own.

Why reusing geodesic functions matters for waveform pipelines

The practical value claimed for the approach is efficiency: by expressing spin corrections through analytic trajectories and already-established geodesic functions, the calculation avoids metric reconstruction and regularization at the particle position and focuses on asymptotic Teukolsky amplitudes in the frequency domain. The paper notes that the framework can be implemented for fast waveform calculation in packages such as FastEMRIWaveforms.

That claim is about computational structure, not about measured speedups; the paper reports no timing comparison. What it does report is a working implementation: a Wolfram Mathematica version of the code is publicly available in the KerrSpinningFluxes package, and the flux comparison against an independent code provides an external numerical check on the linear-in-spin amplitudes.

The priority is generic orbits with spin, not equatorial special cases

Earlier spinning-secondary flux calculations covered circular and eccentric equatorial orbits in Schwarzschild and Kerr spacetimes, and subsequent work computed inspirals and spin-induced phase shifts confirming significant dephasing. The contribution here is to extend the spinning-secondary calculation to fully generic orbits, including eccentricity, inclination, and precession, at 1PA order, and to do so with a formulation built on analytic trajectories and flux-balance laws rather than numerical orbit solving.

For a reader deciding what to take from the paper, the relevant split is between the machinery and the demonstration. The machinery covers generic, eccentric, off-equatorial EMRIs to linear order in the parallel spin component, with all three constant-of-motion rates expressed through asymptotic Teukolsky amplitudes and geodesic functions. The demonstration is narrower: nearly spherical inspirals, where the linear-in-spin phase shift agrees between gauges to about 10⁻² rad at its peak difference. The paper's own priority is the generic-orbit framework, with the nearly spherical case serving as its first numerical test.

Sources used in this answer