A repeating fast radio burst associated with a persistent radio source

FRB distances are often inferred from dispersion measure (DM) by attributing most of the excess DM to the ionized intergalactic medium (IGM). However, local and host-galaxy environments can contribute substantially in some cases. This study presents FRB 20190520B, a repeating FRB co-located with a compact persistent radio source (PRS) in a star-forming dwarf host at z = 0.241, and demonstrates that its host DM dominates the total. The work tests the reliability of DM–redshift inferences, explores the connection between repeating activity, PRS emission, and dense magneto-ionic environments, and assesses implications for FRB subclassification and cosmological applications.

Prior work has used localized FRBs to census cosmic baryons via the DM–redshift relation (for example, Macquart et al. 2020). FRB 121102 established that some repeaters reside in extreme magneto-ionic environments and can be associated with compact PRSs. Other repeaters have lacked compact PRSs or show host properties unlike FRB 121102, suggesting diversity or evolutionary effects. Models and simulations have quantified Galactic DM (NE2001; YMW16), IGM DM statistics, and propagation effects such as scattering and Faraday rotation. This paper situates FRB 20190520B among 15 FRBs with firm host associations, showing it is a strong outlier with unprecedented host DM, and compares it particularly to FRB 121102.

Discovery and radio follow-up: FRB 20190520B was discovered in FAST CRAFTS drift-scan data (1.05–1.45 GHz, 196 μs time, 0.122 MHz frequency resolution) on 2019-11-16 with four bursts in a 24 s scan. Follow-up FAST tracking (April–September 2020) recorded all Stokes at ~50 μs resolution, detecting 75 bursts in 18.5 h at a mean DM of 1204.7 ± 4.0 pc cm−3. Searches used a Heimdall-based pipeline over DMs 100–5000 pc cm−3 (drift scan) and then 100–2000 pc cm−3 (follow-up) with S/N > 7 (fluence threshold ~9 mJy ms for 1 ms width). Burst widths were modeled with Gaussian or Gaussian convolved with one-sided exponential when scattering was evident. Localization and PRS imaging: The VLA realfast system observed in 2020 July–November for ~16 h total at 1.5, 3.0, and 5.5 GHz, detecting 3, 5, and 1 bursts, respectively. Visibilities sampled at 10 ms were imaged in real time and refined offline with finer DM grid, subbanding, and optimized imaging. Calibration used 3C 286 (flux/bandpass/delay) and J1558-1409 (phase), with CASA re-imaging and imfit for centroiding. Weighted averaging across bands yielded the burst position RA = 16h 02m 04.272s, Dec = −11° 17′ 17.32″ (J2000), with total uncertainties 0.10″ and 0.08″. Deep stacked imaging revealed a compact PRS at RA = 16h 02m 04.261s, Dec = −11° 17′ 17.35″ with size <0.36″ (<1.4 kpc) and mean S-band flux density 202 ± 8 μJy over ~2 months; spectrum fitted a power law with index −0.41 ± 0.04. Optical/NIR identification and spectroscopy: CFHT/MegaCam R′-band imaging and Subaru/MOIRCS J-band imaging identified a counterpart galaxy (J160204.31–111718.5) at ~1.3″ offset; the chance-coincidence probability is ~0.8%. Palomar/DBSP spectroscopy at the FRB position and Keck I/LRIS spectroscopy of the extended emission detected Hα, [O III] 4959,5007, and Hβ at z = 0.241 ± 0.001. The extinction-corrected Hα flux implies L_Hα = 7.4 ± 0.2 × 10^40 erg s−1 and SFR ≈ 0.41 M⊙ yr−1. J-band magnitude implies stellar mass ≈ 6 × 10^8 M⊙, consistent with a star-forming dwarf host. At DL = 1218 Mpc, the PRS has L_3GHz ≈ 3 × 10^29 erg s−1 Hz−1. Burst statistics and searches: Burst waiting times were modeled with a Weibull distribution, giving r ≈ 4.5^{+1.3}_{−1.2} h−1 above 9 mJy ms, with k < 1 indicating clustering; for δ ≥ 1 s, r ≈ 5.3 h−1, k ≈ 0.76. No periodicity was found from 1 ms–1000 s or 2–365 d. Fluence and width distributions were measured; isotropic-equivalent burst energies span ~3.6–40 × 10^43 erg. Propagation analyses: Mean scattering time at 1.25 GHz is 10 ± 2 ms from fits to 26 bursts showing asymmetric profiles; other bursts are consistent with Gaussian widths. FAST polarization data showed no significant linear polarization; RM was not detected in FAST band, though high-frequency RM measurements exist from other work. A DM inventory used NE2001 for MW ISM (60 pc cm−3 with ±40% range), MW halo (25–80 pc cm−3), and a log-normal IGM DM with f_IGM typically 0.85, yielding DM_IGM(z=0.241) ≈ 195 pc cm−3 (68% interval). Marginalizing yields DM_host ≈ 903^{+73}_{−72} pc cm−3. Hα surface brightness was converted to EM to estimate a plausible DM_host range given uncertainties in temperature, filling factor, and path length. Scattering arguments disfavour the galaxy being a foreground intervening system, as geometric leverage would imply τ at 1.25 GHz orders of magnitude larger than observed.

- FRB 20190520B is a repeating FRB discovered by FAST and localized by the VLA to RA = 16h 02m 04.272s, Dec = −11° 17′ 17.32″ (J2000). - A compact persistent radio source is co-located at RA = 16h 02m 04.261s, Dec = −11° 17′ 17.35″ with size <0.36″ (<1.4 kpc), S_3GHz = 202 ± 8 μJy, spectral index α = −0.41 ± 0.04, and L_3GHz ≈ 3 × 10^29 erg s−1 Hz−1. - Host galaxy: star-forming dwarf at z = 0.241 ± 0.001; J-band AB magnitude 22.1 ± 0.1; stellar mass ≈ 6 × 10^8 M⊙; Hα luminosity 7.4 ± 0.2 × 10^40 erg s−1; SFR ≈ 0.41 M⊙ yr−1. - Repetition/activity: 88 total detections (FAST + VLA); FAST mean DM = 1204.7 ± 4.0 pc cm−3; modeled burst rate r ≈ 4.5^{+1.3}_{−1.2} h−1 above 9 mJy ms (1 ms width), with clustering (Weibull k < 1); no periodicity detected. - Burst properties: mean width 13.5 ± 1.2 ms at 1.25 GHz; mean scattering timescale 10 ± 2 ms at 1.25 GHz; fluences ~0.03–0.33 Jy ms; isotropic-equivalent energies ~3.6–40 × 10^43 erg. - DM budget: With MW disk ~60 and halo 25–80 pc cm−3, and IGM DM at z = 0.241 of ~195 pc cm−3 (68% interval), the inferred host DM is DM_host ≈ 903^{+73}_{−72} pc cm−3 (observer frame); across f_IGM = 0.6–1, DM_host median spans ~1020–745 pc cm−3. - Optical association: Chance probability of host-galaxy association ≈ 0.8%; chance coincidence of PRS with FRB position ≈ 3 × 10^−6. - Implication: The host contribution dominates the total DM, making FRB 20190520B a strong outlier in the DM–z plane and cautioning against DM-only redshift estimation.

The co-location of bursts with a compact PRS and a star-forming dwarf host at z = 0.241 supports a physical association analogous to FRB 121102. The very large host DM indicates a long tail in the distribution of FRB host contributions, which can dominate over the IGM and undermine redshift estimates based solely on DM. The observed scattering and Hα-based emission measure imply substantial ionized gas near the source and/or within the host ISM. The high repetition rate, potential for large and variable RM (from other bands), and PRS emission suggest a young, dynamic magneto-ionic environment for at least some repeaters. However, the absence of compact PRSs for other repeaters and the diversity in host environments imply either multiple progenitor channels or significant source evolution (PRS fading, declining activity, dissipating local plasma). Consequently, FRB 20190520B strengthens a possible link between repeat activity, PRSs, and high host DM, while highlighting population diversity and the need to include large host DM contributions in FRB cosmological applications.

This work reports the discovery and precise localization of the repeating FRB 20190520B, firmly associated with a compact PRS in a star-forming dwarf host at z = 0.241. The host’s DM contribution (~903 pc cm−3) dominates the total DM, making the source a clear outlier in the DM–redshift relation and emphasizing that DM-only redshift estimates can be highly biased without host identification. The source’s properties closely mirror FRB 121102, suggesting a subclass of young, repeating FRBs embedded in dense, magneto-ionic environments with compact PRSs. Future progress requires larger samples of localized FRBs with host redshifts, VLBI constraints on PRS sizes/locations, and search strategies that account for large host DMs and potential scattering to fully characterize host environments and refine the FRB DM–z relation.

- The breakdown of DM_host between the local circum-source medium and the broader host ISM is not yet constrained; VLBI and multi-frequency polarimetry are needed. - FAST did not detect linear polarization or RM in-band; RM constraints rely on higher-frequency observations. - VLA burst data had 10 ms resolution, limiting detailed temporal/spectral modeling and DM determination for individual bursts. - Hα-based DM estimates depend on uncertain parameters (temperature, filling factor, path length, geometry) and seeing-limited sizes. - Population-level inferences are limited by small sample sizes of precisely localized FRBs with measured host redshifts; other repeaters lack compact PRSs, complicating subclass definitions.