Untangling Extreme Mass Ratio Inspirals and Massive Black Hole Binaries in LISA Data

Abstract

The Laser Interferometer Space Antenna (LISA) will observe sources in the millihertz regime, such as Galactic Binaries (GBs), Massive Black Hole Binaries (MBHBs), and Extreme Mass Ratio Inspirals (EMRIs), so its data stream will be a superposition of an unknown number of gravitational-wave sources. We consider the joint inference of various representative combinations of MBHB and EMRI signals in LISA data (with assumed instrumental noise and stochastic GB foreground). Our Bayesian analyses confirm the expectation of weak inter-source coupling between EMRIs, and between EMRIs and MBHBs, despite strong intra-source parameter correlations. Weak EMRI-MBHB coupling enables a hierarchically sequential approach to EMRI and MBHB inference, rather than fully joint inference. We further show that imperfect MBHB subtraction, arising from missing modes or overlapping MBHBs, has only a weak impact on EMRI inference. For multiple MBHBs, we mitigate inference errors through a joint MBHB fit. We demonstrate this strategy on simulated LISA data containing noise, three MBHBs, and three EMRIs, jointly fitting and subtracting the MBHBs before sequentially inferring the EMRIs - ultimately observing little to no bias on all recovered astrophysical parameters. The notions described throughout our work are highly relevant to informing optimal strategies for the global-fit-based pipelines that are currently being developed for LISA.

Abstract: arXiv descriptive metadata (CC0). Thank you to arXiv for use of its open access interoperability. This service was not reviewed or approved by, nor does it necessarily express or reflect the policies or opinions of, arXiv.