Conservative Black Hole Scattering at Fifth Post-Minkowskian and Second Self-Force Order

Using the worldline quantum field theory formalism, we compute the conservative scattering angle and impulse for classical black hole scattering at fifth post-Minkowskian (5PM) order by providing the second self-force (2SF) contributions. This four-loop calculation involves non-planar Feynman integrals and requires advanced integration-by-parts reduction, novel differential-equation strategies, and efficient boundary-integral algorithms to solve a system of hundreds of master integrals in four integral families on high-performance computing systems. The resulting function space includes multiple polylogarithms as well as iterated integrals with a K3 period, which generate a spurious velocity divergence at $v/c=\sqrt{8}/3$. This divergence is present in the potential region and must be canceled by contributions from the radiative memory region, while its dimensional-regularisation pole should cancel against the radiative tail region. We find that the standard use of Feynman propagators to access the conservative sector fails to ensure this cancellation. We propose a conservative propagator prescription which realises both cancellations leading to a physically sensible answer. All available low-velocity checks of our result against the post-Newtonian literature are satisfied.

💡 Research Summary

In this work the authors present the first complete calculation of the conservative scattering angle and impulse for binary black‑hole (or neutron‑star) scattering at fifth post‑Minkowskian (5PM) order including the second‑order self‑force (2SF) contributions. The computation is carried out within the worldline quantum field theory (WQFT) framework, which treats the compact objects as point‑like worldlines coupled to a linearised graviton field. The central technical challenge is a four‑loop, non‑planar Feynman integral problem that is far more demanding than the previously solved 1SF case.

Setup and methodology
The two bodies are described by worldline actions S = −∑i mi∫dτ √(gμνẋiμẋiν) together with the Einstein‑Hilbert term. The calculation is performed in dimensional regularisation (D=4−2ε) and in the proper‑time gauge ẋi²=1. The authors expand around flat space (gμν=ημν+√(32πG)hμν) and introduce the impact parameter bμ and the initial velocities viμ (γ=vi·vj). The scattering observables are extracted from the worldline displacement ⟨ziμ(ω)⟩ in the ω→0 limit, which yields the impulse Δpμ.

At 5PM the impulse factorises into contributions of different mass‑ratio order. All pieces except the even‑in‑velocity conservative part of the 2SF sector are already known from earlier work. The authors therefore focus on computing Δp(5)μ2SF (even in v) directly.

Diagram generation and integral families
Using recursive diagrammatic algorithms they generate 651 distinct four‑loop diagrams. These are organised into 14 top‑level sectors (planar P, extended planar PX1/PX2, and non‑planar NP) as shown in Fig. 1 of the paper. Four integral families are defined (P, PX1, PX2, NP) to capture all propagator structures. The non‑planar family contains the “memory” topology, where three active gravitons meet at a vertex, leading to non‑linear corrections to the Einstein equations.

IBP reduction
Integration‑by‑parts (IBP) reduction is performed with KIRA 3.0, consuming roughly 3 × 10⁶ core‑hours on a high‑performance cluster. Two key improvements make the reduction feasible: (i) exploiting symmetry relations specific to the conservative sector (automatically generated for the planar families, manually for the non‑planar family) which cuts the number of master integrals by a factor of three; (ii) choosing a “pre‑canonical” basis that is close to a pure‑integral basis but avoids algebraic or transcendental functions, dramatically speeding up the reduction.

The final count of master integrals is 321 (P) + 144 (PX1) + 46 (PX2) + 220 (NP) = 731.

Differential equations and canonical form
The master integrals I(x;ε) satisfy a system ∂x I = B(x;ε) I. By a rational transformation T(x;ε) the system is brought to ε‑factorised canonical form ∂x J = ε A(x) J with J = T I. The matrix A(x) decomposes into blocks associated with three function spaces: ordinary multiple polylogarithms, a three‑dimensional Calabi‑Yau (CY₃) sector, and a two‑dimensional K3 surface sector. The K3 sector is governed by the differential operator

L_K3 = (1‑34 x² + x⁴) θ³ − 6 x²(17‑x²) θ² − 12(3‑x)(3+x) x² θ − 8(5‑x²) x²,

where θ = x ∂x. This operator has a singular point at x = 3 − 2√2, corresponding to γ = 3 (v/c = √8/3). The K3 integrals thus develop a spurious velocity divergence at this physical point.

Canonicalisation is achieved using a combination of public tools (CANONICA for polylog blocks, Baikov representation and INITIAL for larger blocks) and in‑house code based on FiniteFlow and MultivariateApart for off‑diagonal entries. After all transformations, only the K3 sector contributes to the final observable; the CY₃ and polylog sectors cancel out.

Region expansion and cancellation of divergences
To solve the canonical system, boundary conditions are required in the low‑velocity limit (x→1, v→0). The authors employ the method of regions, separating loop momenta into potential (ℓ ∼ (v, 1)) and radiative (ℓ ∼ (v, v)) scalings. For the planar families the relevant regions are:

• Potential region PPPP,
• Tail region (RRPP, PPRR, RPRP, PRPR),
• Memory region (RPPR, PRRP).

For the non‑planar family only PP and RR (memory) regions appear. At 2SF order the potential region carries both an ε‑pole and a γ = 3 divergence stemming from the K3 integrals. The tail region is free of the γ = 3 singularity, while the memory region reproduces the same divergence. Consequently, a consistent physical result requires the γ = 3 pole to cancel between potential and memory contributions, and the ε‑pole to cancel between potential and tail contributions.

Conservative propagator prescription
Standard practice uses Feynman propagators for gravitons and retarded propagators on the worldlines, then extracts the real, even‑in‑v part. The authors find that this prescription does not guarantee the required cancellations: the γ = 3 divergence and the ε‑pole survive, leading to an ill‑defined answer. They therefore propose a modified “conservative propagator” prescription that adjusts the i0‑prescription of the graviton propagator and projects out the appropriate real part on the worldline. With this prescription the potential‑memory cancellation and the tail‑potential ε‑pole cancellation are both realised, yielding a finite, physically sensible scattering angle.

Result and checks
The final conservative scattering angle and impulse are expressed in terms of multiple polylogarithms and iterated integrals over the K3 period. Expanding the result for small velocities reproduces all known post‑Newtonian coefficients up to the order where they are available, providing a stringent check. Moreover, the 5PM‑1SF results are recovered when the 2SF contribution is set to zero, confirming consistency with previous work. The authors also compare with recent N = 8 supergravity calculations of the 5PM‑2SF potential sector and find agreement.

Implications
This calculation closes the last missing piece of the conservative 5PM dynamics at second self‑force order. It demonstrates that high‑order PM calculations inevitably involve higher‑dimensional Calabi‑Yau geometries, here a K3 surface, and that careful treatment of propagator prescriptions is essential for obtaining physically meaningful results. The analytic data produced here will be valuable for constructing ultra‑precise waveform models for upcoming third‑generation ground‑based detectors (Einstein Telescope, Cosmic Explorer) and space‑based missions (LISA), where sub‑percent accuracy in the two‑body dynamics is required. Future work will extend the analysis to include spin, tidal effects, and the full dissipative sector, ultimately delivering a complete 5PM description of binary dynamics.