[Paper Review] Dispersion measures of fast radio burst host galaxies derived from IllustrisTNG simulation

G. Q. Zhang, Hai Yu+2 more|arXiv (Cornell University)|Jul 28, 2020

Pulsars and Gravitational Waves Research5 references4 citations

TL;DR

This study uses the IllustrisTNG cosmological simulation to model dispersion measures (DM) from fast radio burst (FRB) host galaxies, deriving redshift-dependent DM distributions for both repeating and non-repeating FRBs. It finds median DM_host ≈ 35(1+z)^1.08 pc cm⁻³ for FRB 121102-like hosts and ≈ 96(1+z)^0.83 pc cm⁻³ for FRB 180916-like hosts, with non-repeating FRBs showing a median DM_host ≈ 33(1+z)^0.84 pc cm⁻³, all well-fit by log-normal distributions for redshift inference.

ABSTRACT

We calculate the dispersion measures (DMs) contributed by host galaxies of fast radio bursts (FRBs). Based on a few host galaxy observations, a large sample of galaxy with similar properties to observed ones has been selected from the IllustrisTNG simulation. They are used to compute the distributions of host galaxy DMs for repeating and non-repeating FRBs. For repeating FRBs, we infer the DM$ _{\mathrm{host}} $ for FRBs like FRB 121102 and FRB 180916 by assuming that the burst sites are tracing the star formation rates in host galaxies. The median DM$_{\mathrm{host}}$ are $35 (1+z)^{1.08}$ and $96(1+z)^{0.83}$ pc cm$^{-3}$ for FRBs like FRB 121102 and FRB 180916, respectively. In another case, the median of DM$_{\mathrm{host}}$ is about $30 - 70$ pc cm$^{-3}$ for non-repeating FRBs in the redshift range $z=0.1-1.5$, assuming that the burst sites are the locations of binary neutron star mergers. In this case, the evolution of the median DM$_{\mathrm{host}}$ can be calculated by $33(1+z)^{0.84}$ pc cm$^{-3}$. The distributions of DM$_{\mathrm{host}}$ of repeating and non-repeating FRBs can be well fitted with the log-normal function. Our results can be used to infer redshifts of non-localized FRBs.

Motivation & Objective

• To determine the distribution of dispersion measures (DM) contributed by FRB host galaxies using cosmological simulations.
• To infer redshifts of non-localized FRBs by modeling host DM contributions based on observed FRB host properties.
• To compare DM host distributions for repeating FRBs (e.g., FRB 121102, FRB 180916) and non-repeating FRBs under different physical assumptions.
• To assess the robustness of results under varying selection criteria for host galaxy stellar mass and star formation rate.

Proposed method

• Selected galaxies from the IllustrisTNG simulation with properties matching observed FRB hosts (e.g., stellar mass, SFR) for statistical analysis.
• Simulated FRB burst sites based on star formation rate (SFR) for repeating FRBs and binary neutron star merger locations for non-repeating FRBs.
• Calculated DM_host by integrating electron density along line-of-sight paths through host galaxy gas and stellar components.
• Used linear interpolation between discrete redshifts to estimate DM_host at intermediate redshifts.
• Applied Monte Carlo simulations to generate mock FRB samples and derived pseudo-redshifts using maximum likelihood estimation.
• Fitted DM_host distributions with log-normal functions to enable statistical inference for cosmological applications.

Experimental results

Research questions

• RQ1 What is the expected distribution of host galaxy DM for repeating FRBs like FRB 121102 across different redshifts?
• RQ2 How does the host DM distribution differ for non-repeating FRBs assuming they originate from binary neutron star mergers?
• RQ3 To what extent do selection criteria (e.g., stellar mass and SFR ranges) affect the derived DM_host distributions?
• RQ4 Can the simulated DM_host distributions be used to accurately infer the redshifts of non-localized FRBs?
• RQ5 How well do the simulated DM_host distributions match observational constraints from localized FRBs?

Key findings

• The median DM_host for FRB 121102-like hosts evolves as 35(1+z)^1.08 pc cm⁻³ across redshifts z = 0.1 to 1.5.
• The median DM_host for FRB 180916-like hosts evolves as 96(1+z)^0.83 pc cm⁻³, reflecting higher intrinsic DM due to more massive, star-forming hosts.
• For non-repeating FRBs, the median DM_host follows 33(1+z)^0.84 pc cm⁻³ when burst sites trace binary neutron star merger locations.
• The DM_host distributions for both repeating and non-repeating FRBs are well described by log-normal functions.
• The maximum likelihood method successfully recovers the input redshift distribution of simulated FRBs, validating the method for pseudo-redshift inference.
• The results are robust to variations in selection criteria, as narrow ranges of stellar mass and SFR yield similar DM_host distributions.