Black hole evolution with the BSSN system by pseudo-spectral methods

TL;DR

This paper introduces a new pseudo-spectral code for evolving black holes using the BSSN system, demonstrating initial success but encountering instabilities after about 100M due to mode growth.

Contribution

It is the first to apply spectral methods to black hole evolution with the BSSN system, testing various boundary and gauge conditions.

Findings

Spectral methods can evolve black holes with BSSN but face stability issues.

Simulations fail after ~100M due to unstable modes.

Full grid evolution without symmetry constraints causes instabilities.

Abstract

We present a new pseudo-spectral code for the simulation of evolution systems that are second order in space. We test this code by evolving a non-linear scalar wave equation. These non-linear waves can be stably evolved using very simple constant or radiative boundary conditions, which we show to be well-posed in the scalar wave case. The main motivation for this work, however, is to evolve black holes for the first time with the BSSN system by means of a spectral method. We use our new code to simulate the evolution of a single black hole using all applicable methods that are usually employed when the BSSN system is used together with finite differencing methods. In particular, we use black hole excision and test standard radiative and also constant outer boundary conditions. Furthermore, we study different gauge choices such as $1+\log$ and constant densitized lapse. We find that these methods in principle do work also with our spectral method. However, our simulations fail after about $100M$ due to unstable exponentially growing modes. The reason for this failure may be that we evolve the black hole on a full grid without imposing any symmetries. Such full grid instabilities have also been observed when finite differencing methods are used to evolve excised black holes with the BSSN system.