---
title: Conservative Tail-of-Tail Action in Two-Body Dynamics
url: https://www.emergentmind.com/topics/conservative-tail-of-tail-action
type: topic
---

# Conservative Tail-of-Tail Action in Two-Body Dynamics

The conservative tail-of-tail action is the time-symmetric, nonlocal-in-time hereditary interaction in relativistic two-body dynamics generated when a gravitational wave emitted by the system is scattered twice off the background curvature produced by the total mass-energy \(\mathcal M\). In current post-Newtonian (PN) and post-Minkowskian (PM) treatments, it enters starting at \(O(G^5/c^{11})\), i.e. at \(5.5\)PN order, and also at \(5\)PM order and beyond. It is used to compute dynamical observables such as the Delaunay Hamiltonian for bound motion and the scattering angle for hyperbolic motion, and it supplies the conservative half-integer-PN sector at \(5.5\)PN and \(6.5\)PN [2507.08708].

## 1. Position in the hereditary hierarchy

The tail-of-tail effect is a higher hereditary correction within the conservative dynamics of compact binaries. Its physical content is a double backscatter process: radiation emitted by the source interacts twice with the long-range monopolar curvature and then feeds back into the motion. In the terminology used for circular-binary dynamics, a simple tail is the conservative part of the non-linear gravitational-wave interaction in which emitted radiation scatters once off the monopolar field \(M\) of the binary and first appears at \(4\)PN, whereas a tail-of-tail is an iteration of this process, schematically \((\text{tail})^2\), and first appears at \(5.5\)PN [1912.12359].

A concise hierarchy is:

| Process | Conservative character | First order stated in the literature |
|---|---|---|
| Simple tail | nonlocal in time; logarithmic kernels | \(4\)PN |
| Tail-of-tail | nonlocal in time; half-integer PN | \(5.5\)PN |
| Tail-of-tail-of-tail | higher hereditary logarithmic sector | \(7\)PN |

For conservative circular-binary dynamics, the distinction between even and odd tail iterations is important. Odd iterates \((\text{tail})^p\) with \(p=2n-1\) produce logarithmic terms and appear at \((3n+1)\)PN order, while even iterates such as \((\text{tail})^2\) contribute half-integer PN terms but not logs, at least up to the orders explicitly considered in current derivations [1912.12359]. This places the conservative tail-of-tail action in a specific niche: it is hereditary and nonlocal, but it is not the source of the logarithmic conservative structure associated with simple tails or tail-of-tail-of-tail terms.

## 2. Action, kernels, and multipolar content

The manifestly time-symmetric form of the conservative tail-of-tail action is written as [2507.08708]
$$
S_{\rm time-sym}^{\rm tail-of-tail}
=\frac12 \left(\frac{G{\mathcal M}}{c^3}\right)^2 G \sum_{l\ge 2}\frac{1}{c^{2l+1}}
\left[ a_l \beta_l^{\rm even}\!\int dt \int_{-\infty}^{\infty} dt'\, I_L^{(l+2)}(t) I_L^{(l+2)}(t')\ln\frac{c|t-t'|}{2r_0}
+\frac{b_l}{c^2}\beta_l^{\rm odd}\!\int dt \int_{-\infty}^{\infty} dt'\, J_L^{(l+2)}(t)J_L^{(l+2)}(t')\ln\frac{c|t-t'|}{2r_0} \right].
$$

Here \(I_L\) and \(J_L\) are the even- and odd-parity source multipole moments, and the flux coefficients are
$$
a_l=\frac{(l+1)(l+2)}{l(l-1)l! (2l+1)!!},\qquad
b_l=\frac{4 l (l+2)}{(l-1)(l+1)! (2l+1)!!}.
$$
The coefficients \(\beta_l^{\rm even}\) and \(\beta_l^{\rm odd}\) encode the multipole renormalization and tail structure. The arbitrary scale \(r_0\) drops out of the action [2507.08708].

For explicit calculations, the action is rewritten in a time-split principal-value form:
$$
S^{\rm tail-of-tail}_{\rm time-split}
=\frac14 \left(\frac{G{\mathcal M}}{c^3}\right)^2 G \int dt \int_{-\infty}^{\infty} \frac{d\tau}{\tau}\, {\mathcal G}^{\rm split}(t,\tau),
$$
with
$$
{\mathcal G}^{\rm split}(t,\tau)= \sum_{l\ge 2}\frac{1}{c^{2l+1}}
\left( a_l \beta_l^{\rm (even)}{\mathcal G}_I^l(t,\tau) +\frac{b_l}{c^2}\beta_l^{\rm (odd)}{\mathcal G}_J^l(t,\tau) \right),
$$
$$
{\mathcal G}_I^l(t,\tau)= I_L^{(l+1)}(t)\big[I_L^{(l+2)}(t+\tau)-I_L^{(l+2)}(t-\tau)\big],
$$
$$
{\mathcal G}_J^l(t,\tau)= J_L^{(l+1)}(t)\big[J_L^{(l+2)}(t+\tau)-J_L^{(l+2)}(t-\tau)\big].
$$

At the fractional \(1\)PN level beyond the leading \(5.5\)PN term, one sets
$$
\beta_l^{\rm (odd)}=\beta_l^{\rm (even)}=\beta_l,
$$
and keeps only \(l=2,3\), obtaining
$$
S^{\rm tail-of-tail}_{\rm time-split}
=\frac{G}{4}\left(\frac{G{\mathcal M}}{c^3}\right)^2 \frac{1}{c^5}
\left[ a_2\beta_2\,{\sf G}_{I_2} +\frac{b_2}{c^2}\beta_2\,{\sf G}_{J_2} +\frac{1}{c^2}a_3\beta_3\,{\sf G}_{I_3} +\cdots \right],
$$
with
$$
a_2=\frac15,\qquad a_3=\frac{1}{189},\qquad b_2=\frac{16}{45},
$$
$$
\beta_2=-\frac{214}{105},\qquad \beta_3=-\frac{26}{21}.
$$
The explicitly relevant \(1\)PN-fractional pieces are therefore the mass quadrupole \(I_2\) at \(5.5\)PN and \(6.5\)PN, together with the current quadrupole \(J_2\) and mass octupole \(I_3\) at \(6.5\)PN [2507.08708].

A key structural property is that, unlike the earlier linear-tail conservative action, which is logarithmically UV-divergent at \(t'=t\), the tail-of-tail conservative action is UV-finite at coincidence. The corresponding radiation-reaction completion is, however, logarithmically divergent in the external zone, with cancellation against IR divergences in the matched source moments [2507.08708]. This sharply differentiates the conservative tail-of-tail action from the \(4\)PN simple-tail action, whose conservative part carries the familiar UV pole and compensating near-zone IR structure [1511.07379].

## 3. Hamiltonian realization for bound motion

The conservative tail-of-tail action contributes to the Hamiltonian through
$$
S^{\rm tail-of-tail}_{\rm time-split}=-\int dt\, H^{\rm tail-of-tail}(t),
$$
so that
$$
H^{\rm tail-of-tail}(t) = H^{\rm tail-of-tail}_{I_2}(t) +\eta^2\!\left[ H^{\rm tail-of-tail}_{J_2}(t)+H^{\rm tail-of-tail}_{I_3}(t) \right], \qquad \eta\equiv \frac1c.
$$
The three explicit pieces are [2507.08708]
$$
H^{\rm tail-of-tail}_{I_2}(t)= -\frac{G}{4}\left(\frac{G{\mathcal M}}{c^3}\right)^2\frac{1}{c^5} a_2\beta_2\, I_{ij}^{(3)}(t)\!\int_{-\infty}^{\infty}\frac{d\tau}{\tau} \big[I_{ij}^{(4)}(t+\tau)-I_{ij}^{(4)}(t-\tau)\big],
$$
$$
H^{\rm tail-of-tail}_{J_2}(t)= -\frac{G}{4}\left(\frac{G{\mathcal M}}{c^3}\right)^2\frac{1}{c^7} b_2\beta_2\, J_{ij}^{(3)}(t)\!\int_{-\infty}^{\infty}\frac{d\tau}{\tau} \big[J_{ij}^{(4)}(t+\tau)-J_{ij}^{(4)}(t-\tau)\big],
$$
$$
H^{\rm tail-of-tail}_{I_3}(t)= -\frac{G}{4}\left(\frac{G{\mathcal M}}{c^3}\right)^2\frac{1}{c^7} a_3\beta_3\, I_{ijk}^{(4)}(t)\!\int_{-\infty}^{\infty}\frac{d\tau}{\tau} \big[I_{ijk}^{(5)}(t+\tau)-I_{ijk}^{(5)}(t-\tau)\big].
$$

For ellipticlike motion, these terms are averaged over one radial period by means of the \(1\)PN quasi-Keplerian parametrization
$$
nt=\ell=u-e_t\sin u,\qquad r=a_r(1-e_r\cos u),\qquad \phi=2K\arctan\!\left[\left(\frac{1+e_\phi}{1-e_\phi}\right)^{1/2}\tan\frac{u}{2}\right].
$$
The average is
$$
\langle H^{\rm tail-of-tail}\rangle =\frac{\oint H^{\rm tail-of-tail}(t)\,dt}{\oint dt}.
$$
The resulting decomposition,
$$
\langle H^{\rm tail-of-tail}\rangle = \langle H^{\rm tail-of-tail}\rangle_{\eta^0} +\eta^2 \langle H^{\rm tail-of-tail}\rangle_{\eta^2},
$$
yields
$$
\langle H^{\rm tail-of-tail}\rangle_{\eta^0} = \frac{\nu^2\pi}{a_r^{13/2}} \left( \frac{49969}{315}e_t^2+\frac{6848}{525} \right),
$$
$$
\langle H^{\rm tail-of-tail}\rangle_{\eta^2} = -\frac{\nu^2\pi}{a_r^{15/2}} \left[ \left(\frac{1473251}{2450}e_t^2+\frac{4982}{315}\right)\nu + \left(\frac{593849}{630}e_t^2+\frac{991861}{7350}\right) \right].
$$
These are the direct conservative tail-of-tail dynamical contributions at \(5.5\)PN and \(6.5\)PN [2507.08708].

The averaged Hamiltonian is then transcribed into the effective-one-body (EOB) potentials
$$
A(u,\nu)=1-2u+\sum_{n\ge 3} a_n(\nu,\ln u)u^n,\qquad
\bar D(u,\nu)=1+\sum_{n\ge 2}\bar d_n(\nu,\ln u)u^n,
$$
through
$$
\delta A = a_{6.5}u^{13/2}+a_{7.5}u^{15/2}, \qquad \delta \bar D = \bar d_{5.5}u^{11/2}+\bar d_{6.5}u^{13/2}.
$$
Matching gives
$$
a_{6.5}=\frac{13696}{525}\nu\pi,
$$
$$
a_{7.5}=-\frac{10052}{225}\nu^2\pi-\frac{512501}{3675}\nu\pi,
$$
$$
\bar d_{5.5}=\frac{264932}{1575}\nu\pi,
$$
$$
\bar d_{6.5}=-\frac{893149}{2450}\nu^2\pi-\frac{21288791}{17640}\nu\pi.
$$
The \(O(\nu)\) parts agree with earlier \(1\)SF black-hole perturbation results, whereas the \(O(\nu^2)\) parts are new information [2507.08708].

## 4. Hyperbolic scattering and PM observables

The same action determines the conservative scattering angle for hyperbolic encounters. The PM expansion is written as
$$
\chi(p_\infty,j)=\sum_{n\ge 1}\frac{2\chi_n(p_\infty)}{j^n},
\qquad
p_\infty\equiv \sqrt{\gamma^2-1},\qquad
j=\frac{J}{GMm_1m_2}=\frac{p_\infty b}{GMh}.
$$
The tail-of-tail contribution is extracted from the on-shell action through
$$
\chi^{\rm tail-of-tail}_{\rm time-split} = \frac{\partial}{\partial J} S^{\rm tail-of-tail, on-shell}_{\rm time-split}.
$$

The calculation proceeds by rewriting the time-split action in frequency space and using the identity
$$
{\sf G}_X^l = -2\int_0^\infty d\omega\, \omega^{2l+3}|\hat X_L(\omega)|^2.
$$
This gives
$$
S^{\rm tail-of-tail}_{\rm time-split} = G\left(\frac{G{\mathcal M}}{c^3}\right)^2\frac{1}{c^5} \int_0^\infty d\omega\,{\mathcal F}_{\rm tail-of-tail}(\omega),
$$
with
$$
{\mathcal F}_{\rm tail-of-tail}(\omega) = \frac15\frac{107}{105}\omega^7|\hat I_{ij}(\omega)|^2 +\frac{1}{c^2}\left( \frac{16}{45}\frac{107}{105}\omega^7|\hat J_{ij}(\omega)|^2 +\frac{1}{189}\frac{13}{21}\omega^9|\hat I_{ijk}(\omega)|^2 \right).
$$

The final on-shell action is expanded through \(1/j^8\), i.e. through \(8\)PM, and differentiation with respect to \(J\) yields a scattering angle whose displayed leading terms include
$$
\chi^{\rm tail-of-tail}_{\rm time-split} = \nu\Bigg\{ \Big[\Big(\frac{6784\nu}{175}+\frac{1511296}{33075}\Big)p_\infty^8-\frac{95872}{675}p_\infty^6\Big]\frac{1}{j^5} +\cdots \Bigg\},
$$
with the full result containing \(1/j^6\), \(1/j^7\), and \(1/j^8\) terms, including \(\pi^2\) and \(\zeta(3)\) pieces [2507.08708]. The leading piece reproduces the previously known \(5\)PM/\(5.5\)PN result,
$$
\chi^{\rm LO}_{\rm tail-of-tail}
=
-\nu \frac{95872}{675}\frac{p_\infty^6}{j^5}
+O(G^6).
$$

Comparison with state-of-the-art scattering theory requires adding the linear radiation-reaction correction
$$
\chi^{\rm rr, lin} = -\frac12\left( E^{\rm rad}\frac{\partial \chi^{\rm cons}}{\partial E} + J^{\rm rad}\frac{\partial \chi^{\rm cons}}{\partial J} \right).
$$
With the recently computed \(1\)SF angular-momentum-loss information inserted, the sum of the conservative tail-of-tail contribution and the radiation-reaction correction agrees completely with the \(5\)PM-\(1\)SF scattering angle of Driesse et al. through \(6.5\)PN, including the highest-order term reported in the comparison [2507.08708].

## 5. Relation to simple tails, logarithms, and failed-tail sectors

The conservative tail-of-tail action is best understood against the broader structure of hereditary conservative effects. The simple-tail action at \(4\)PN is already nonlocal in time and splits into conservative and dissipative parts; in effective field theory (EFT), the conservative piece is the real logarithmic term in frequency space, while the dissipative piece is the \(i\pi\,\mathrm{sign}(\omega)\) term. The conservative simple-tail term carries a logarithmic UV divergence that cancels against an IR singularity in the conservative near zone [1511.07379]. By contrast, the tail-of-tail conservative action is UV-finite at coincidence and contributes the half-integer conservative sector at \(5.5\)PN and \(6.5\)PN [2507.08708].

For circular binaries, the logarithmic conservative structure of the energy is tied to simple tails and tail-of-tail-of-tail terms rather than to the tail-of-tail sector itself. The explicit \(7\)PN analysis shows that a tail-of-tail-of-tail \((\text{tail})^3\) becomes relevant at \(7\)PN and is the first higher-tail process that contributes logarithms beyond the simple-tail terms [1912.12359]. A plausible implication is that the conservative tail-of-tail action should be regarded not as a logarithmic correction but as the canonical half-integer hereditary contribution.

It is also necessary to distinguish the conservative tail-of-tail action from several similarly named but structurally different sectors. The \(5\)PN “failed-tail” interaction \(S_{i|j}\times M_{ij}\) is hereditary in origin but, after the required cancellations and harmonic completion, it contributes an instantaneous term to the Fokker Lagrangian rather than a genuine nonlocal-in-time conservative sector [2307.05860]. Likewise, the angular-momentum-sourced “L-ftail” has an instantaneous waveform effect even though its far-zone self-energy contributes to conservative dynamics; the mass tail is the canonical hereditary case, whereas the L-ftail behaves as a local or instantaneous analog that requires gauge-restoring care in the quadrupole sector [2307.05327]. These distinctions prevent a common misconception: not every higher radiative backscatter effect produces a nonlocal conservative action of tail-of-tail type.

## 6. Status in current two-body dynamics

Within current PN, PM, and self-force comparisons, the conservative tail-of-tail action has a sharply defined status. It is the only contributor to the half-integer PN effects at \(5.5\)PN and \(6.5\)PN in the conservative sector, and it has been used to compute the Delaunay Hamiltonian through \(6.5\)PN and the scattering angle through \(6.5\)PN and up to \(8\)PM [2507.08708]. The resulting EOB coefficients,
$$
a_{6.5},\; a_{7.5},\; \bar d_{5.5},\; \bar d_{6.5},
$$
provide both an exact match to the known \(O(\nu)\) self-force information and new \(O(\nu^2)\) input [2507.08708].

More broadly, the conservative tail sector is a standard ingredient in the decomposition of two-body dynamics into local-in-time and nonlocal-in-time pieces. At \(5\)PM, the nonlocal-in-time conservative contribution arising from hereditary tail effects must be isolated before extracting the genuinely local conservative Hamiltonian, and the completion of that nonlocal sector through \(10\)SF has been identified as a key step in determining the local \(5\)PM conservative dynamics [2604.25916]. This suggests that the conservative tail-of-tail action is part of a wider program in which hereditary, time-symmetric effects are computed separately and then subtracted or transcribed into bound-motion data.

In that sense, the conservative tail-of-tail action is not merely a correction term. It is the explicit action-level representative of the double-backscatter hereditary sector, with a well-defined multipolar structure, a principal-value time-split representation suited to computation, and direct consequences for both bound and scattering observables. Its present role is therefore both dynamical and organizational: it encodes the \(5.5\)PN/\(6.5\)PN conservative half-integer sector and serves as a controlled bridge between hereditary radiation physics, EOB potentials, and PM scattering data [2507.08708].

Source: https://www.emergentmind.com/topics/conservative-tail-of-tail-action