Two approaches for the gravitational self force in black hole spacetime: Comparison of numerical results

TL;DR

This paper compares two independent numerical methods for calculating the gravitational self-force on a particle orbiting a Schwarzschild black hole, demonstrating their results are consistent within very small fractional differences.

Contribution

It establishes a formal correspondence between two different approaches to self-force calculations and confirms their numerical consistency within tight error margins.

Findings

The two methods produce consistent results for the conservative self-force effect.

The fractional differences between the methods are on the order of 10^{-5} to 10^{-7}.

The comparison validates the reliability of both computational approaches.

Abstract

Recently, two independent calculations have been presented of finite-mass ("self-force") effects on the orbit of a point mass around a Schwarzschild black hole. While both computations are based on the standard mode-sum method, they differ in several technical aspects, which makes comparison between their results difficult--but also interesting. Barack and Sago [Phys. Rev. D {\bf 75}, 064021 (2007)] invoke the notion of a self-accelerated motion in a background spacetime, and perform a direct calculation of the local self force in the Lorenz gauge (using numerical evolution of the perturbation equations in the time domain); Detweiler [Phys. Rev. D {\bf 77}, 124026 (2008)] describes the motion in terms a geodesic orbit of a (smooth) perturbed spacetime, and calculates the metric perturbation in the Regge--Wheeler gauge (using frequency-domain numerical analysis). Here we establish a formal correspondence between the two analyses, and demonstrate the consistency of their numerical results. Specifically, we compare the value of the conservative $O(\mu)$ shift in $u^t$ (where $\mu$ is the particle's mass and $u^t$ is the Schwarzschild $t$ component of the particle's four-velocity), suitably mapped between the two orbital descriptions and adjusted for gauge. We find that the two analyses yield the same value for this shift within mere fractional differences of $\sim 10^{-5}$--$10^{-7}$ (depending on the orbital radius)--comparable with the estimated numerical error.