Transition from Inspiral to Plunge

Adiabatic evolution fails near a last stable orbit because the radial restoring force vanishes. Radiation reaction then changes the orbit on the same scale as the increasingly slow radial dynamics, so neither a sequence of circular geodesics nor an instantaneous switch to geodesic plunge is uniformly valid. A transition layer, obtained by balancing the local instability against slow dissipation, joins the two regimes.

Helpful background. The Einstein Equation fixes the Kerr background, Weak Metric Perturbations and Gauge Freedom supplies perturbative gauge control, and Presymplectic Potentials and Currents clarifies energy and angular momentum. Black-Hole Perturbation Theory supplies dissipative fluxes, Relativistic Two-Body Dynamics supplies the mass-ratio expansion, and Gravitational Radiation and Waveforms supplies phasing conventions.

Degeneracy of the radial potential

Begin with a circular equatorial extreme-mass-ratio inspiral. Let the Kerr black hole have mass $M_\bullet$, the secondary have mass $m_\star$, and $\epsilon=m_\star/M_\bullet\ll1$. Write its radial dynamics schematically as

$$\left(\frac{dr}{d\tau}\right)^2 =\mathcal R(r,E,L_z).$$

A circular orbit satisfies

$$\mathcal R=0, \qquad \partial_r\mathcal R=0.$$

At the last stable circular orbit, the additional condition

$$\partial_r^2\mathcal R=0$$

removes the leading radial restoring term. Expanding about the transition point with $R=r-r_*$ and a slowly changing orbital-control variable $\delta\mathcal C$ gives, using the radial equation of motion,

$$\frac{d^2R}{d\tau^2} =-A R^2-B\,\delta\mathcal C+\cdots,$$

where $A$ and $B$ are positive after an appropriate choice of orientation. Radiation reaction drives

$$\delta\mathcal C\simeq\epsilon\,\kappa(\tau-\tau_*).$$

Keeping only these competing terms yields

$$\frac{d^2R}{d\tau^2} =-A R^2-\epsilon B\kappa(\tau-\tau_*).$$

The precise definitions of $R$, $\tau$, and $\mathcal C$ depend on the normalization of the Kerr potential; the scaling result does not. Here the geodesic radial force is expanded at fixed $E,L_z$ before the slowly evolving constants are inserted. Differentiating the geodesic first integral after promoting $E,L_z$ to arbitrary functions would generate extra terms divided by $\dot r$; it is not a derivation of the forced transition equation. The leading reduction uses circular balance $\dot E\simeq\Omega_*\dot L_z$ and a consistent normalization condition.

The transition scaling

Set

$$R=\epsilon^{2/5}R_0 X, \qquad \tau-\tau_*=\epsilon^{-1/5}\tau_0 T.$$

Then the acceleration, nonlinear instability, and dissipative drive all scale as $\epsilon^{4/5}$:

$$\frac{d^2R}{d\tau^2}\sim\epsilon^{4/5}, \qquad R^2\sim\epsilon^{4/5}, \qquad \epsilon(\tau-\tau_*)\sim\epsilon^{4/5}.$$

Choosing $R_0$ and $\tau_0$ to absorb $A$ and $B\kappa$ produces the universal dimensionless equation

$$\frac{d^2X}{dT^2}=-X^2-T.$$

For $D=B\kappa\gt0$, the normalizations are $\tau_0=(AD)^{-1/5}$ and $R_0=A^{-3/5}D^{2/5}$; direct substitution makes both coefficients in the reduced equation unity. At fixed subextremal spin, the radial width is $O(\epsilon^{2/5}GM_\bullet)$ and the transition duration is $O(\epsilon^{-1/5}GM_\bullet)$ in the proper-time normalization used here. The accumulated change in the driven orbital control variable is $O(\epsilon^{4/5})$.

At early negative $T$, the slowly varying solution follows the stable quasicircular branch, $X\simeq\sqrt{-T}$. At later $T$, the solution leaves that branch and accelerates toward plunge. Initial data are fixed by matching to the adiabatic inspiral, not chosen freely inside the layer.

This dominant balance and its circular-equatorial Kerr implementation are derived in Ori and Thorne 2000, §§II–IV.

Matching inspiral, transition, and plunge

A consistent composite trajectory has three overlapping descriptions:

1. Adiabatic inspiral: sufficiently outside the last stable orbit, horizon-plus-infinity fluxes evolve a sequence of stable circular geodesics.
2. Transition layer: the rescaled equation retains radial inertia, quadratic instability, and leading dissipative drift simultaneously.
3. Plunge: once the dissipative change across an orbital time is small relative to the growing radial motion, a geodesic with constants inherited from the transition approximates the subsequent fall.

The transition solution must reproduce the inspiral asymptotics in their overlap and supply the plunge constants in the second overlap. Moving either matching point within its valid window should change the composite result only at the next neglected order. An abrupt switch at a coordinate radius has no such error control.

Setting $\kappa=0$ is an important check. The secular term disappears, the transition scaling loses its drive, and a stable orbit does not cross the separatrix by itself. One may perturb an unstable orbit and obtain geodesic departure, but that is not a radiation-reaction-driven inspiral-to-plunge transition.

Scope beyond circular equatorial motion

The universal equation is not a ready-made model for every EMRI. Eccentric orbits approach a separatrix associated with a homoclinic structure and a growing radial period; the transition depends on radial phase and on how the turning points merge. Inclined Kerr motion adds polar phase and can encounter an orbital resonance nearby. Near-extremal spin introduces throat scales and can alter the ordering between orbital, transition, and radiation-reaction times.

Conservative self-force shifts the location and frequency of marginal stability. Oscillatory and second-order dissipative terms affect the matching and phase. These corrections can be organized, but they must not be hidden by calling the leading circular model “the plunge.”

There is also no coordinate-independent plunge radius. Useful invariant markers include the vanishing radial epicyclic frequency for a specified circular family, the degeneracy of the radial action, and the orbital frequencies and constants passed through the matched solution. This page develops the asymptotic joining; general strong-field perturbation equations remain in Black-Hole Perturbation Theory.

Exercises

Derive the exponents $2/5$ and $-1/5$ by dominant balance.

Solution

Assume $R\sim\epsilon^a$ and $\tau-\tau_*\sim\epsilon^b$. The three terms scale as

$$\ddot R\sim\epsilon^{a-2b}, \qquad R^2\sim\epsilon^{2a}, \qquad \epsilon(\tau-\tau_*)\sim\epsilon^{1+b}.$$

Equating exponents gives $a-2b=2a=1+b$. The solution is $a=2/5$ and $b=-1/5$. All three retained terms are then $O(\epsilon^{4/5})$.

Explain why switching off dissipation is a more discriminating check than merely verifying that the reduced equation has plunging solutions.

Solution

The equation’s nonlinear instability can produce departure for some initial data even without radiation reaction. An inspiral-to-plunge model must show that slow dissipative drift carries the matched stable branch through loss of stability. Setting the drift coefficient to zero removes that secular passage and distinguishes the intended transition from an arbitrarily perturbed unstable geodesic.

References

• Ori, Amos, and Kip S. Thorne. “The Transition from Inspiral to Plunge for a Compact Body in a Circular Equatorial Orbit Around a Massive, Spinning Black Hole.” Physical Review D 62 (2000): 124022. DOI.