Parameter-Estimation Biases for Eccentric Supermassive Binary Black Holes in Pulsar Timing Arrays: Biases Caused by Ignored Pulsar Terms

TL;DR

This paper derives analytical formulas to estimate biases in parameters of eccentric supermassive binary black holes in pulsar timing arrays caused by ignoring pulsar terms, improving understanding of gravitational wave data analysis.

Contribution

It provides the first analytical expressions for parameter-estimation biases considering eccentric orbits and pulsar term phases as random variables.

Findings

Analytical results agree with previous numerical studies at 1.5σ level.

Biases in phase and eccentricity decrease with increasing eccentricity.

Stronger gravitational wave signals lead to larger pulsar term biases.

Abstract

The continuous nanohertz gravitational waves (GWs) from individual supermassive binary black holes (SMBBHs) can be encoded in the timing residuals of pulsar timing arrays (PTAs). For each pulsar, the residuals actually contain an Earth term and a pulsar term, but usually only the Earth term is considered as signal and the pulsar term is dropped, which leads to parameter-estimation biases (PEBs) for the SMBBHs, and currently there are no convenient evaluations of the PEBs. In this article, we formulate the PEBs for a SMBBH with an eccentric orbit. In our analyses, the unknown phases of pulsar terms are treated as random variables obeying the uniform distribution $U[0,2\pi)$, due to the fact that pulsar distances are generally poorly measured. Our analytical results are in accordance with the numerical work by Zhu et. al. (2016) at $1.5\sigma$ level, implying that our formulae are effective in estimating magnitudes of the PEBs. Additionally, we find that for two parameters -- Earth term phase $\varphi^E$ and orbital eccentricity $e$, their biases $\Delta \varphi^E$ and $\Delta e/e$ monotonically decrease as $e$ increases, which partly confirms a hypothesis in our previous work Chen & Zhang (2018). Furthermore, we also calculate the PEBs caused by the recently observed common-spectrum process (CSP), finding that if the strain amplitude of the continuous GW is significantly stronger ($3$ times larger, in our cases) than the stochastic GW background, the PEBs from pulsar terms are larger than those from the CSP. Our formulae of the PEBs can be conveniently applied in the future PTA data analyses.