A highly active repeating fast radio burst in a complex local environment

not Due to the interaction be- the pulse and the intervening the dispersion sweep of FRBs provides a unique probe of its environment and the ionized baryon content of the intergalactic medium 1 . Ac- tive repeaters has been shown to be associated with persistent radio source 2 (PRS), and dense, energetic, magnetized plasmas 3,4 .Here we report the discovery and localization of a new, ex- tremely active repeater, FRB 190520, which is co-located with a compact PRS and identiﬁed with a dwarf host galaxy of high star formation at a redshift z = 0 . 241 . The estimated host galaxy contribution pc cm − 3 is nearly an order of magnitude higher than the average of FRB host galaxies 5,6 and much larger than the contribution from the intergalac- tic medium, suggesting caution in inferring redshifts for FRBs without accurate host galaxy identiﬁcations. This represents the second source after FRB 121102 with conﬁrmed associa- tion between FRB and compact PRS. The dense, complex host galaxy environment and the associated persistent radio source may point to a distinctive origin or an earlier evolutionary stage for highly active repeating FRBs. We obtained an optical spectrum at the location of the FRB with the Double Spectrograph on the Palomar 200-inch Hale Telescope that revealed the redshift of the putative host to be z = 0 . 241 based on a detection of strong H α , [O III] 4859Å, and [O III] 5007Å lines (see Methods). A follow- up observation with the Low Resolution Imaging Spectrometer (LRIS) at the Keck I Telescope covering both the FRB location and the nearby Subaru J -band source along the extended R -band 87 structure indicates the R ′ -band structure is dominated by the [O III] emission at the same redshift of z = 0 . 241. The H α luminosity L H α = 7 . 4 ± 0 . 2 × 10 40 erg sec − 1 after extinction correction suggests a star formation rate of ∼ 0.41 M ⊙ yr − 1 . Based on the J -band magnitude, we estimate the stellar mass of the host galaxy to be ∼ 6 × 10 8 M ⊙ . Thus, we characterize as a dwarf galaxy with a relatively high star-formation rate for its stellar mass 15 . At the luminosity distance implied by the redshift, the PRS has a radio luminosity of L 3 GHz = 4 × 10 29 1 − 1

galaxy contribution DM host ≈ 912 +69 −108 pc cm −3 is nearly an order of magnitude higher than the 48 average of FRB host galaxies 5, 6 and much larger than the contribution from the intergalac-49 tic medium, suggesting caution in inferring redshifts for FRBs without accurate host galaxy 50 identifications. This represents the second source after FRB 121102 with confirmed associa-51 tion between FRB and compact PRS. The dense, complex host galaxy environment and the 52 associated persistent radio source may point to a distinctive origin or an earlier evolutionary 53 stage for highly active repeating FRBs.

54
FRB 190520 was discovered with the Five-hundred-meter Aperture Spherical radio Tele-  The discovery of FRB 190520 and its high similarity to FRB 121102 demonstrate that some           The details of the observations are given in Table 2 . The telescopes were pointed at the field 317 centered at (RA, Dec)[J2000] = (16h02m01s, -11d17m28s). We note that due to a system error 318 the realfast system wasn't run on the VLA observation on MJD 59169. However, this observation 319 was used to make a deep radio image at S band. 320 We used the realfast search system at VLA to search for bursts from FRB 190520 in our 321 VLA observations. The realfast search system has been described in detail in 11, 35 , but here we

372
We use burstfit to model the spectro-temporal properties of bursts. This analysis extends 373 the discussion of 40 and has been described in 41 . We model the pulse profile and the spectra using 374 a Gaussian function, and therefore fit for 6 parameters: mean of pulse, width of pulse, mean of 375 spectra, width of spectra, fluence and DM. Following 41 , we use curve fit followed by MCMC 376 methods to estimate the posterior distribution of these fit parameters. The fitted properties of the 377 VLA bursts are given in Table 5 . Burst S5 was very weak, and hence its fit estimates are not well PanSTARRS survey DR1 42 . We identified the radio point sources using the following criteria:

404
• The peak intensity (Jy/beam) of a source should be 0.7, 0.5, 0.5 times higher than its inte-405 grated flux (Jy) for the 1.5 GHz, 3 GHz and 5.5 GHz images, respectively.

406
• The S/N (peak intensity / local rms noise) of a source should be greater than 5.

407
In total, we detected 375, 113, and 43 sources in the 1.5 GHz, 3 GHz and 5.5 GHz deep 408 image, respectively. We visually checked the selected sources to make sure that they are 'point-  Figure 5 shows the result of the 481 parameters distribution obtained with Markov Chain Monte Carlo (MCMC). We find the burst rate 482 of FRB 190520 is r = 4.5 +1.9 −1.5 hr −1 with shape parameters k = 0.37 +0.04 −0.04 for all 79 bursts which 483 are above 7σ > 9.3 mJy · ms (left panel in Figure 5 ), and r = 5.3 +1.1 −1.0 hr −1 with shape parameters 484 k = 0.76 +0.09 −0.08 for excluding waiting time shorter than 1 s (right panel in Figure 5 ).
where DM MW is the contribution from our Galactic interstellar matter, DM halo is the contribution 502 from the Milky Way halo, DM host the contribution from the host galaxy including its halo and 503 any gas local to the FRB source, and DM IGM is the contribution from the intergalactic medium.
where the free electron number per baryon in the universe is χ(z) ≈ 7 (1 + z l ) 3 ν 4 , where DM l is the DM contribution of the lens galaxy in pc cm −3 in the lens frame, ν is the observing 557 frequency in GHz, z is the lens galaxy redshift, andF = ζǫ 2 / f (l 2 o l i ) 1/3 quantifies the electron density where the intra-channel Faraday rotation ∆θ is given by where c is the speed of light, ∆ν is the channel width, and ν c is the central channel observing 576 frequency. Taking ∆θ = 1 rad, ∆ν = 0.122 MHz, and ν c = 1.25 GHz for our data, we get RM = 577 1.8 × 10 5 rad m −2 and depolarization fraction of 54.5% caused by intra-channel Faraday rotation.

578
Assuming that the pulse is 100% linearly polarized intrinsically and the non-detection of RM is 579 caused by intra-channel Faraday rotation, we place an lower limit on the RM of 1.8 × 10 5 rad m −2 .

580
Such a large RM is even larger than that of FRB 121102 50       Extended Data | Table 5 : Spectro-temporal properties of the VLA bursts. The (1 σ) errors on the last digit are shown in parenthesis.