The CGM with local universe FRBs: evidence of strong AGN feedback in a massive elliptical galaxy

Abstract

Modern cosmology and galaxy formation rely on an understanding of how cosmic baryons are distributed, a significant portion of which exist in the diffuse gas confined to halos. Fast Radio Bursts (FRBs) are a promising probe of the Universe's ionized gas. At low redshift, the contribution to the dispersion measure (DM) from the intergalactic medium (IGM) and intervening halos is subdominant, allowing us to study the circumgalactic media (CGM) of the host galaxies. We select a sample of five local universe FRBs whose host interstellar medium (ISM) DM is negligible and use these to constrain the mass of the CGM in each halo. We find that one of our sources, the only massive elliptical host galaxy, has been evacuated of its baryons ($M_\mathrm{gas}=0.02^{+0.02}_{-0.02}M_\mathrm{h}$, corresponding to $\sim$10$\%$ of the cosmological average $\frac{Ω_b}{Ω_m}$). This galaxy shows evidence of a past episode of AGN activity, consistent with the picture of strong AGN feedback in galaxy group-scale halos. The other sources are consistent with existing multiwavelength data and tentatively support more baryon retention in $L_*$ galaxies compared to group-scale halos. We show that FRBs can measure the halo gas fraction $f_\mathrm{gas}$ in halos of mass $M_\mathrm{h}\sim10^{11-13}M_\odot$, and up to $\sim10^{14}M_\odot$ if galaxy cluster hosts are included, which is a larger range than other gas probes can access. Finally, we demonstrate that a large sample of local universe FRBs, such as those expected from upcoming all-sky radio telescopes, will enable precision measurements of halo gas, which is crucial for understanding the effects of feedback.

Figures (8)

• Figure 1: $\mathrm{DM_{host}}$ for all localized FRBs with $z<0.2$ at the time of writing plotted against potential indicators of $\mathrm{DM_{host,ISM}}$: host galaxy inclination $i$, scattering timescale $\tau_\mathrm{exc}$, rotation measure RM, and impact parameter $b_\perp$. The colored points are the five FRBs selected for our analysis. Light grey points are the poorly localized ($\sim$arcmin) FRBs.
• Figure 2: Pan-STARRS1 chambers2019panstarrs1surveys gri color images of the FRB host galaxies in our sample. The FRB localization regions are shown as solid white ellipses. The dashed white circles are 0.1$R_\mathrm{500c}$ for each galaxy's halo. For FRB20200120E, the inset shows the VLBI localization to a globular cluster in M81 presented by Kirsten_2022.
• Figure 3: The PDFs of the different DM components using various estimates, with the shaded curves showing the adopted PDFs used in our analysis. In each plot, all curves are arbitrarily scaled by the same factor for better visualization. The top row shows $p(\mathrm{DM_{cosmic}})$, which is estimated using the best fit of konietzka2025raytracingfastradiobursts, with the relation of Macquart_2020 plotted for reference. The middle row shows $p(\mathrm{DM_{MW,ISM}})$, where the final PDF is the combination of the HI, H$\alpha$, and YMW16 Yao_2017 estimates, truncated by the pulsar lower limit (vertical dashed line). For FRB 20220319D, the mean $\mathrm{DM_{MW,ISM}}$ estimate is much larger than $\mathrm{DM_{obs}}$, so we set $p(\mathrm{DM_{MW,ISM}})$ as a delta function at the pulsar lower limit following Ravi_2025. We also show the NE2001 cordes2003ne2001inewmodelgalacticOcker_2024 and Ocker_2020 models for comparison. The third row shows YT20 Yamasaki_2020 estimate of $\mathrm{DM_{MW,CGM}}$. We also plot an estimate of $\mathrm{DM_{MW,CGM}}$ using the nearest OVII absorption sightline to each FRB from Fang_2015 and the scaling relation between $N_\mathrm{OVII}$ and DM from Yamasaki_2020, as well as the median of sightlines within 20$^{\circ}$ from Das2021 (but note that the uncertainty in both of these is quite large). In the bottom panel, we plot the final $p(\mathrm{DM_{host}})$, which under our assumption of negligible $\mathrm{DM_{host,ISM}}$ becomes $p(\mathrm{DM_{host,CGM}})$.
• Figure 4: $\mathrm{DM_{host}}$ vs. normalized impact parameter through the halo $b_\perp/R_\mathrm{200c}$. Each panel shows the measurement from an FRB in our sample (in black) and randomly selected halos from simulations within 0.2 dex of the FRB host $M_\mathrm{h,200c}$. We show three simulations: IllustrisTNG (TNG50) nelson2021illustristngsimulationspublicdata, SIMBA Dav__2019, and SIMBA with AGN feedback turned off. Solid lines are the median values, and shaded regions are the 16th and 84th percentile intervals, capturing both sightline-to-sightline variance within each halo and halo-to-halo variance.
• Figure 5: The halo gas fraction for our FRB sample plotted for a subset of the profiles considered. The left plot shows the gas fraction within the $R_\mathrm{vir}$, $f_\mathrm{gas,vir}$, and the right shows the gas fraction within $R_\mathrm{500c}$, $f_\mathrm{gas,500c}$. For clarity, we only plot our fiducial profile (mNFW), and the highest and lowest predicting profiles (excluding NFW). We also plot the bigwood2024weaklensingcombinedkinetic profile for FRB 20240209A to show that it is consistent.