Gravitational-wave observations of binary black holes: Effect of non-quadrupole modes

TL;DR

This paper investigates how non-quadrupole modes affect gravitational wave detection and parameter estimation for binary black holes, finding quadrupole-only templates suffice for certain mass ratios but cause biases in others.

Contribution

It provides a detailed analysis of the impact of non-quadrupole modes on GW detection and parameter estimation, highlighting when simplified templates are adequate and when they introduce significant errors.

Findings

Quadrupole-only templates are sufficient for q ≤ 4 with less than 10% detection loss.

Neglecting non-quadrupole modes causes systematic parameter errors, especially for q ≥ 4 and M ≥ 150 M_sun.

Systematic errors exceed 1σ statistical errors for high mass, unequal mass binaries at SNR > 8.

Abstract

We study the effect of non-quadrupolar modes in the detection and parameter estimation of gravitational waves (GWs) from non-spinning black-hole binaries. We evaluate the loss of signal-to-noise ratio and the systematic errors in the estimated parameters when one uses a quadrupole-mode template family to detect GW signals with all the relevant modes, for target signals with total masses $20 M_\odot \leq M \leq 250 M_\odot$ and mass ratios $1 \leq q \leq 18$. Target signals are constructed by matching numerical-relativity simulations describing the late inspiral, merger and ringdown of the binary with post-Newtonian/effective-one-body waveforms describing the early inspiral. We find that waveform templates modeling only the quadrupolar modes of the GW signal are sufficient (loss of detection rate $< 10\%$) for the detection of GWs with mass ratios $q\leq4$ using advanced GW observatories. Neglecting the effect of non-quadrupole modes will introduce systematic errors in the estimated parameters. The systematic errors are larger than the expected $1\,\sigma$ statistical errors for binaries with large, unequal masses ($q\gtrsim4, M \gtrsim 150 M_\odot$), for sky-averaged signal-to-noise ratios larger than $8$. We provide a summary of the regions in the parameter space where neglecting non-quadrupole modes will cause unacceptable loss of detection rates and unacceptably large systematic biases in the estimated parameters.