Nine New Repeating Fast Radio Burst Sources from CHIME/FRB
E. Fonseca, B. C. Andersen, M. Bhardwaj, P. Chawla, D. C. Good, A. Josephy, V. M. Kaspi, K. W. Masui, R. Mckinven, D. Michilli, Z. Pleunis, K. Shin, S. P. Tendulkar, K. M. Bandura, P. J. Boyle, C. Brar, T. Cassanelli, D. Cubranic, M. Dobbs, F. Q. Dong, B. M. Gaensler, G. Hinshaw, T. L. Landecker, C. Leung, D. Z. Li, H.-H. Lin, J. Mena-Parra, M. Merryfield, A. Naidu, C. Ng, C. Patel, U. Pen, M. Rafiei-Ravandi, M. Rahman, S. M. Ransom, P. Scholz, K. M. Smith, I. H. Stairs, K. Vanderlinde, P. Yadav, A. V. Zwaniga
DDraft version February 4, 2020
Typeset using L A TEX preprint style in AASTeX62
Nine New Repeating Fast Radio Burst Sources from CHIME/FRB
E. Fonseca,
1, 2
B. C. Andersen,
1, 2
M. Bhardwaj,
1, 2
P. Chawla,
1, 2
D. C. Good, A. Josephy,
1, 2
V. M. Kaspi,
1, 2
K. W. Masui,
4, 5
R. Mckinven,
6, 7
D. Michilli,
1, 2
Z. Pleunis,
1, 2
K. Shin,
4, 5
S. P. Tendulkar, K. M. Bandura,
8, 9
P. J. Boyle,
1, 2
C. Brar,
1, 2
T. Cassanelli,
6, 7
D. Cubranic, M. Dobbs,
1, 2
F. Q. Dong, B. M. Gaensler,
6, 7
G. Hinshaw, T. L. Landecker, C. Leung,
4, 5
D. Z. Li,
11, 12, 13
H. -H. Lin,
12, 13
J. Mena-Parra, M. Merryfield,
1, 2
A. Naidu,
1, 2
C. Ng, C. Patel,
6, 1
U. Pen,
12, 6, 14, 15, 13
M. Rafiei-Ravandi, M. Rahman, S. M. Ransom, P. Scholz,
6, 10
K. M. Smith, I. H. Stairs, K. Vanderlinde,
6, 7
P. Yadav, and A. V. Zwaniga
1, 2 Department of Physics, McGill University, 3600 rue University, Montr´eal, QC H3A 2T8, Canada McGill Space Institute, McGill University, 3550 rue University, Montr´eal, QC H3A 2A7, Canada Department of Physics and Astronomy, University of British Columbia, 6224 Agricultural Road, Vancouver, BC V6T1Z1, Canada MIT Kavli Institute for Astrophysics and Space Research, Massachusetts Institute of Technology, 77 MassachusettsAve, Cambridge, MA 02139, USA Department of Physics, Massachusetts Institute of Technology, 77 Massachusetts Ave, Cambridge, MA 02139, USA Dunlap Institute for Astronomy and Astrophysics, University of Toronto, 50 St. George Street, Toronto, ON M5S3H4, Canada David A. Dunlap Department of Astronomy and Astrophysics, University of Toronto, 50 St. George Street, Toronto,ON M5S 3H4, Canada CSEE, West Virginia University, Morgantown, WV 26505, USA Center for Gravitational Waves and Cosmology, West Virginia University, Morgantown, WV 26505, USA Dominion Radio Astrophysical Observatory, Herzberg Astronomy and Astrophysics Research Centre, NationalResearch Council Canada, P.O. Box 248, Penticton, BC V2A 6J9, Canada Department of Physics, University of Toronto, 60 St. George Street, Toronto, ON M5S 1A7, Canada Canadian Institute for Theoretical Astrophysics, University of Toronto, 60 St. George Street, Toronto, ON M5S3H8, Canada Max Planck Institute for Radio Astronomy, Auf dem Huegel 69, 53121 Bonn, Germany Canadian Institute for Advanced Research, CIFAR Program in Gravitation and Cosmology, Toronto, ON M5G 1Z8,Canada Perimeter Institute for Theoretical Physics, 31 Caroline Street North, Waterloo, ON N2L 2Y5, Canada National Radio Astronomy Observatory, 520 Edgemont Rd., Charlottesville, VA 22903, USA
ABSTRACTWe report on the discovery and analysis of bursts from nine new repeating fast radioburst (FRB) sources found using the Canadian Hydrogen Intensity Mapping Experi-ment (CHIME) telescope. These sources span a dispersion measure (DM) range of 195to 1380 pc cm − . We detect two bursts from three of the new sources, three burstsfrom four of the new sources, four bursts from one new source, and five bursts fromone new source. We determine sky coordinates of all sources with uncertainties of Corresponding author: E. [email protected] a r X i v : . [ a s t r o - ph . H E ] F e b ∼ (cid:48) . We detect Faraday rotation measures for two sources, with values − − . − , that are substantially lower than the RM derived from bursts emit-ted by FRB 121102. We find that the DM distribution of our events, combined withthe nine other repeaters discovered by CHIME/FRB, is indistinguishable from thatof thus far non-repeating CHIME/FRB events. However, as previously reported, theburst widths appear statistically significantly larger than the thus far non-repeatingCHIME/FRB events, further supporting the notion of inherently different emissionmechanisms and/or local environments. These results are consistent with previouswork, though are now derived from 18 repeating sources discovered by CHIME/FRBduring its first year of operation. We identify candidate galaxies that may contain FRB190303.J1353+48 (DM = 222.4 pc cm − ). INTRODUCTIONFRBs are an enigmatic class of radio transients that exhibit millisecond durations, cosmologicaldistances and large energy output ( ∼ erg; Dolag et al. 2015). Such extreme characteristics haveresulted in a diverse and evolving landscape of physical models that differ in progenitor and emission-mechanism types, with possible interpretations ranging from compact objects to cosmic strings (fora living summary of proposed models , see Platts et al. 2018).Among the most intriguing FRB sources are those that emit multiple bursts, as first seen in FRB121102 (Spitler et al. 2016; Scholz et al. 2016). Repeating sources necessitate non-cataclysmic mod-els. Recently proposed models typically include young, highly magnetised neutron stars that likelyinteract with their environments, such as supernova remnants (e.g. Lyubarsky 2014; Murase et al.2016; Beloborodov 2017; Metzger et al. 2019) or massive black holes (e.g. Zhang 2018a), or burstingdue to internal magnetar instabilities affecting the magnetosphere (e.g. Lyutikov 2019). No model yetfully explains all existing, albeit limited, observational data. Therefore, discoveries of new repeatingFRBs are important to constrain available models.Interferometric follow-up observations of bursts can be performed to obtain precise sub-arcsecondpositions, which can then be observed with optical telescopes to identify host galaxies and theirredshifts. The localization of FRB 121102 through direct imaging of repeat bursts (Chatterjee et al.2017; Marcote et al. 2017) pinpointed the source to a star-forming region in a dwarf galaxy at redshift z = 0 .
193 (Tendulkar et al. 2017; Bassa et al. 2017), enabling multi-wavelength studies of the sourceenvironment (e.g. Scholz et al. 2017). Recent localizations of single-burst FRBs show that FRBs canalso reside in more massive galaxies with a wide range of specific star formation rates (Bannisteret al. 2019; Ravi et al. 2019; Prochaska et al. 2019; Marcote et al. 2020). Additional localizationsand subsequent multi-wavelength observations will provide a unique data set of host classifications,redshifts, assessments of source environments and other information.The CHIME telescope and its FRB search backend (CHIME/FRB Collaboration et al. 2018) firstdetected 13 low-frequency bursts (CHIME/FRB Collaboration et al. 2019a), and discovered thesecond repeating FRB source from this initial sample (CHIME/FRB Collaboration et al. 2019b).CHIME/FRB recently published a detection of FRB 121102 in the CHIME band (Josephy et al. 2019),as well as eight new repeating sources whose bursts generally show complex morphological featuresand, as an ensemble, possess larger widths than those of thus far non-repeating FRBs (CHIME/FRB https://frbtheorycat.org ine New Repeaters from CHIME/FRB − . ± . − ; PaperI) is considerably smaller than the RM ∼ rad m − measured for FRB 121102 (Michilli et al. 2018);moreover, FRB J180916.J0518+65 has no persistent radio counterpart (Marcote et al. 2020) whileFRB 121102 is coincident with a persistent source (Marcote et al. 2017). These facts suggest thatdramatically different magneto-ionic environments can contain FRB sources.In this work, we report the discovery and subsequent analyses of bursts emitted from nine newrepeating FRB sources. In Section 2, we highlight the observations taken with the CHIME telescopethat enabled offline analyses of total-intensity and polarization spectra. In Section 3, we describe theanalysis used for quantifying per-burst and sample properties. In Section 4, we discuss the resultsobtained from analyses of burst morphology, sample properties, baseband data and multi-wavelengthcounterparts. In Section 5, we summarise our findings. OBSERVATIONSAll detections presented here were made during a period of telescope commissioning between 28August 2018 and 30 September 2019 using the CHIME/FRB system. As described by CHIME/FRBCollaboration et al. (2018), the CHIME/FRB backend continuously receives a total-intensity,polarization-summed time series generated by a FX correlator with 0.98304 ms cadence for 16,384frequency channels across the 400–800 MHz band. The 128-node FRB system employs real-time ra-dio frequency interference (RFI) mitigation and a modified tree de-dispersion algorithm. Candidatesignals with integrated S/N values greater than a configurable threshold are immediately forwardedto a post-detection pipeline for real-time classification to i) identify and ignore RFI-induced signals,ii) check for coincidence with known Galactic sources and iii) compare with predicted Galactic con-tributions to DM. Signals are classified as extragalactic (i.e., FRBs) if they are not associated withany known Galactic sources, and their measured DMs exceed independent estimates of the maximumvalues predicted by available Galactic DM models (Cordes & Lazio 2002; Yao et al. 2017); we do notaccount for DM contributions from electronic content in the Galactic halo.We deemed a group of bursts to originate from the same repeating source if their measured positionand DM values lied close to one another, such that the differences in these quantities satisfied thefollowing criteria given estimated uncertainties: ∆DM < − ; ∆R.A.cos(Dec.) < < ∼
700 FRBs have been detected with the CHIME/FRBsystem during the aforementioned time period. The thresholds used in this work are more stringentthan those used in Paper I, with the DM threshold being increased by an order of magnitude.The CHIME/FRB real-time pipeline automatically records 60-s segments of total-intensity datato disk at the resolution used by the FRB instrument for detection, for all burst events deemedastrophysical and extragalactic. For sufficiently bright signals, the CHIME/FRB system also records Documentation for the post-detection part of the CHIME/FRB realtime pipeline can be found at https://chimefrb.github.io/frb-l2l3/ Analysis of the ∼
700 FRBs found by CHIME/FRB is ongoing and will be published in a forthcoming catalog. buffered telescope baseband data — complex voltages measured with 4-bit precision (each for real andimaginary parts) by all 1,024 dual-polarization feeds at 2.56- µ s cadence across 1,024 frequency chan-nels — which allows for offline polarization detection, position verification, and high-time-resolutionstudies of burst morphology. For this work, we analyzed polarization properties and verified local-izations using available baseband data; we deferred morphological studies with baseband data forfuture study as analysing µ s-level structure is beyond the scope of this work.Since Paper I, we lowered the S/N thresholds for recording intensity and baseband data to disk asconfidence in the real-time detection system grew. For repeating sources, total-intensity spectra ofbursts presented here are drawn from a wider S/N distribution, going down to a real-time detectionS/N of 8 instead of 9. A larger number of bursts were detected with 9 < S/N <
10 than in Paper Idue to improved system sensitivity and classification algorithms during the telescope commissioningperiod. For baseband recording, we used a threshold S/N of 9. ANALYSIS & RESULTSWe detected nine new repeating FRB sources with the CHIME/FRB instrument, using the criteriadescribed in Section 2. Source properties are summarized in Table 1, while individual burst propertiesare listed in Table 2. 3.1.
Source localization
Burst localization was carried out following the methods described in Paper I. For all sources,we used a model of the CHIME primary beam and S/N estimates from all beams that detectedbursts in order to obtain burst-averaged sky positions using a χ -grid method. While the methodsremain identical to those used in Paper I, we have updated our underlying beam model to correctfor an effective 0.071 ◦ rotation of the telescope relative to true North (counter clockwise as viewedfrom the CHIME meridian) that was identified after publication of Paper I. We chose not to correctpositions reported in Paper I in this work since their quoted uncertainties make the corrected positionsstatistically equivalent. Our best estimates of source positions are provided in Table 1, and graphicaldepictions of the localization regions are shown in Figure 1.We were not able to acquire total-intensity or baseband data for one burst from Source 3, sincedata acquisition was temporarily disabled around the time of this burst due to commissioning-relatedupgrades of the CHIME/FRB system. However, since this burst satisfied detection-S/N and classi-fication thresholds for FRB signals, we nonetheless recorded the metadata produced by the realtimedetection pipeline. Since these data contain the quantities needed for burst-averaged localizationrefinement, we used the metadata for this burst in generating χ grids and confidence intervals forthe sky localization of Source 3. 3.2. Exposure Determination
On-sky exposure (referenced to 600 MHz) across the FWHM region of the total-intensity beamssynthesised for the CHIME/FRB system was determined in a manner similar to that used in PaperI. In summary, we estimated the exposure for a grid of sky positions within the 90% confidenceuncertainty region for each source (see Figure 1). We then calculated the weighted average and This amount of rotation corresponds to a distance offset between the apparent and true N-S directions of ∼ ine New Repeaters from CHIME/FRB Table 1.
Properties of Nine New CHIME/FRB Repeating Sources
Source Name a R.A. b Dec. b l c b c DM d DM e NE2001 DM e YMW16 N bursts Exposure f Completeness g (J2000) (J2000) (deg) (deg) (pc cm − ) (pc cm − ) (pc cm − ) (hr, upper / lower) (Jy ms)1 190208.J1855+46 ± (cid:48) +46 ◦ (cid:48) ± (cid:48) ±
14 3.42 190604.J1435+53 14h35m ± (cid:48) +53 ◦ (cid:48) ± (cid:48) ±
11 2.83 190212.J18+81 18h24m ± (cid:48) +81 ◦ (cid:48) ± (cid:48) h ±
52 / 159 ±
11 8.2 / 1317h39m ± (cid:48) +81 ◦ (cid:48) ± (cid:48) ± (cid:48) +74 ◦ (cid:48) ± (cid:48) ±
33 / 36 ±
25 5.9 / 185 190117.J2207+17 22h07m ± (cid:48) +17 ◦ (cid:48) ± (cid:48) − ± ± (cid:48) +48 ◦ (cid:48) ± (cid:48) ±
12 2.67 190417.J1939+59 ± (cid:48) +59 ◦ (cid:48) ± (cid:48) ±
19 4.38 190213.J02+20 02h14m ± (cid:48) +20 ◦ (cid:48) ± (cid:48) − ± ± (cid:48) +20 ◦ (cid:48) ± (cid:48) − ± (cid:48) +46 ◦ (cid:48) ± (cid:48) ±
14 2.508h02m ± (cid:48) +46 ◦ (cid:48) ± (cid:48) a Here we employ the naming convention (YYMMDD.JHHMM ± DD) used in CHIME/FRB Collaborationet al. (2019a) and CHIME/FRB Collaboration et al. (2019b) in the current absence of a final namingconvention agreed upon by the community. These names therefore are likely to change. The date in thename corresponds to our first detection of the source. For brevity, and for the remainder of the paper, werefer to the repeaters by Source number (Column 1). For sources with non-contiguous error regions in Figure 1, the name is defined by the central position, except for Sources 3, 8, and 9, for which the ‘central’ R.A.is not well defined at the minute level. b Positions were determined from per-burst S/N data (see Section 3.1). Sources with position in italics havethree or more non-contiguous error regions, with the tabulated position referring to the central region, with90% confidence uncertainty regions. See Figure 1 for details. Sources 3, 8, and 9 have two non-contiguousuncertainty regions, resulting in two position entries (see Fig. 1). c Galactic longitude and latitude for the best position. d Weighted average DM (see Table 2). e Maximum model prediction along this line-of-sight for the NE2001 (Cordes & Lazio 2002) and YMW16(Yao et al. 2017) Galactic electron density distribution models. Neither model accounts for DM contributionsfrom the Galactic halo, which is thought to be 50–80 pc cm − (Prochaska & Zheng 2019). f For sources observed twice a day, the second entry corresponds to the less sensitive lower transit. Theuncertainties in the total exposure for the upper and lower transits of each source are dominated by thecorresponding source declination uncertainties since the widths of the synthesized beams vary significantlywith declination (see § g Fluence completeness limits are given at the 90% confidence level (see § h One of the three bursts had no intensity data captured; see Table 2. standard deviation for the exposure over all these positions, with the weights equal to the sky-position probability maps shown in Figure 1. Exposure for each source during the interval from 28August 2018 to 30 September 2019 is reported in Table 1 and plotted in Appendix B (see Figure 10).For high declination sources ( δ > +70 ◦ ), which transit across the primary beam of the telescope twiceeach day, we chose to report the exposure for the lower transit separately since the beam response isdifferent for the two transits.3.3. Determination of Burst Fluence and Peak Flux Density
Dynamic spectra were calibrated and burst fluences and peak fluxes were determined as described inPaper I. In summary, we used transit observations of steady sources with known spectral properties . ◦ . ◦ . ◦ Source 1 ±
14’ +46 ◦ ± . ◦ . ◦ . ◦ Source 2 ±
10’ +53 ◦ ± . ◦ . ◦ . ◦ Source 3 ±
15’ +81 ◦ ± ±
16’ +81 ◦ ± . ◦ . ◦ . ◦ Source 4 ±
17’ +74 ◦ ± . ◦ . ◦ . ◦ Source 5 ±
8’ +17 ◦ ± . ◦ . ◦ . ◦ Source 6 ±
14’ +48 ◦ ± . ◦ . ◦ . ◦ Source 7 ±
13’ +59 ◦ ± . ◦ . ◦ . ◦ Source 8 ±
16’ +20 ◦ ± ±
16’ +20 ◦ ± . ◦ . ◦ . ◦ Source 9 ±
11’ +46 ◦ ± ±
12’ +46 ◦ ± . . . . . . . . . . . P u l s a r A n a l og F r a c t i o n < ∆ χ Right Ascension (J2000) D ec li n a t i o n ( J ) Figure 1.
Detection positions of the new CHIME/FRB repeating FRB sources, as determined from CHIME/FRBdetection beam information through the methods described in § ◦ × ◦ . localization is performed asa χ -minimization. The method is applied to a large population of analogous pulsar events (i.e., pulsars with similarbrightness and beam-detection statistics), which we use to translate ∆ χ values to empirical confidence intervalsdepicted by the color scale. The 90% and 99% confidence intervals are indicated as solid and dashed contours; we usethe former interval to report the most likely positions. The R.A. of the beam centers for each detection are shownas black ticks on the bottom of each panel. For declination, panels are centered on the beam with the highest S/Ndetection; we do not add analogous ticks on the vertical axes since each beam’s declination is constant in time. Hatchedregions represent disfavored areas where, for at least one burst in the sample, the beam model predicts substantialattenuation in formed-beam sensitivity in the portion of the band where emission is observed. ine New Repeaters from CHIME/FRB . ◦ of declination for calibration.Since there are no calibration sources within 5 ◦ in declination from Source 4, we assumed that ourbeam is North-South symmetric and used a source within 1 ◦ of declination on the opposite side ofzenith. 3.4. Fluence Completeness Determination
We determined fluence completeness, the per-source threshold fluence for which all bursts above itare expected to be detected by the CHIME/FRB instrument, in a manner similar to that used inPaper I, with some minor modifications to how sensitivity variation is characterised.Previously, for circumpolar sources observed twice a day, fluence thresholds were simulated for thetransit where bursts were detected, then calibrators with similar declination were used to get a band-averaged sensitivity scaling for the transit without detections. We caution that the per-frequencyscaling between transits is highly structured, so the observed attenuation for a given source willdepend strongly on its spectrum. In this work, we extended the simulations to cover both transits,using our beam model to compute relative sensitivities. This approach included a description of thesynthesised beams in addition to the degradation in the primary beam sensitivity, which is significantfor sources that pass between synthesised beams in one transit while crossing beam centers in theother transit.In order to characterise the inter-day sensitivity variation of the CHIME/FRB system, we used themethod described in Josephy et al. (2019). The method involves estimating the daily variation inRMS noise at the location of each source by analysing distributions of S/N values of pulsars within 5 ◦ of the source declination. However, observations with the CHIME/FRB system through 2019 Junesuggest that small variations in RMS noise can have a significant effect on the FRB detection rate.Since these variations cannot be adequately characterised by the small number of pulsars detectablein a 5 ◦ declination range, we used pulsars which are robustly detected by the CHIME/FRB systemwithin its entire observable declination range (with Dec. > − ◦ ) in this work. The estimate of thedaily variation in RMS noise is obtained by averaging measurements from pulsars detected on eachsidereal day for which the telescope was operating with the same gain calibration. This approach isin contrast to the method used in Josephy et al. (2019), which combined measurements from pulsarsdetected on the same UTC day. Our modification makes the measurement of the relative RMS noisemore sensitive to changes in system sensitivity due to varying gain-calibration strategies. Typicaldaily variations are at the 20% level.3.5. Characterisation of Burst Morphology
Many bursts from repeating FRB sources exhibit complex morphology, comprised of multiple sub-bursts that usually drift down in frequency as time progresses (e.g., Hessels et al. 2019). Finding theoptimal DM necessitates methods beyond the typically employed S/N-optimisation, and here we usedthe same method as in Paper I — maximising burst structure by calculating the phase coherence ofemission in all frequency channels with the
DM phase package (Seymour et al. in prep.) over a rangeof trial DMs. The sub-burst alignment after dedispersion was verified by eye and Figures 2 and 3show dynamic spectra for all bursts, dedispersed to the best estimate of that burst’s DM, as listedin Table 2.We used the same modeling procedure discussed in Paper I for estimating widths, arrival timesand scattering timescales from our calibrated total-intensity dynamic spectra, which we summarizehere. In all fits, we preserved the raw time and frequency resolution of the CHIME/FRB total-intensity data and fitted two-dimensional models of Gaussian temporal profiles and either Gaussianor weighted power-law spectral shapes. For bursts with simple morphology i.e., with no significantmultiple components), we fitted a single-burst spectrum and allowed the DM to float freely duringeach fit in order to determine robust uncertainties for all other burst-specific parameters. In thesecases, the DMs estimated from direct modeling of the dynamic spectra were statistically consistentwith the values estimated from the structure-optimisation algorithm shown in Table 2. For burstswith spectro-temporal structure typical of repeating sources, we fitted several components and heldthe DM fixed to the structure-optimisation value in Table 2 in order to ensure separation of sub-burstcomponents and the determination of robust width estimates.For all bursts, we also fitted single-tail scattering profiles to dynamic spectra that arise fromfrequency-dependent, multi-path propagation of the signal through small-scale electron density inho-mogeneities in the ISM. We fitted for scattering by applying a temporal pulse broadening function(e.g. McKinnon 2014) and assuming that the scattering timescales depends on frequency as f − . Asdone in previous works, we selected the superior model for each burst – one that explicitly fits for thescattering timescale or one that does not fit for scattering effects – based on the best-fit χ statistic(see CHIME/FRB Collaboration et al. 2019a, and Paper I). Using this model selection procedure,we consider all widths reported in Table 2 to approximate the intrinsic burst widths. Significantestimates of the scattering timescale are presented with uncertainties in Table 2; for non-detectionsof scattering timescales and widths, which we defined to be consistent with 0 ms at 3 σ confidence, wepresent upper limits of the 2 σ confidence interval. As noted in Paper I, repeater morphology likelyintroduces bias into fits of scattering timescales due to faint, extended emission being detected aspart of a scattering tail; we therefore urge caution when interpreting scattering timescales shown inTable 2.As in Paper I, we determined burst drift rates using an autocorrelation analysis and present resultsfor bursts where the drift rate is constrained in Table 2. We calculated each linear drift rate by fittinga two-dimensional Gaussian profile to the two-dimensional auto-correlation. We used a Monte Carlomethod to obtain robust confidence intervals, de-dispersing our bursts to 100 DM values drawn from https://github.com/danielemichilli/DM phase ine New Repeaters from CHIME/FRB Table 2.
Individual Burst Properties from Nine New CHIME/FRB Repeaters. a Day MJD Arrival Time b DM Drift Rate Width b Scattering Time Fluence f Peak Flux Density f (yymmdd) (UTC @ 600 MHz) (pc cm − ) (MHz/ms) (ms) (ms @ 600 MHz) (Jy ms) (Jy)Source 1 (FRB 190208.J1855+46)190208 58522 17:41:42.300(3) 579.9(2) − . +0 . − . < . < . c < . < . e < . c,d d ... 2.8(3) < . < . d d ... 9(2) 1.29(3) 2.9(1.8) ...Source 5 (FRB 190117.J2207+17)190117 58500 22:18:26.46617(7) 393.3(1) ... 1.44(3) < . c < . − . +0 . − < . − +1 . − . < . c − . +0 . − . < . − . +0 . − . <
15 3.2(7) 0.4(2)190929 58755 03:14:15.846(2) 1378.5(3) ... 1.19(2) < . < . < < < . a Unconstrained parameters are listed as “...” Uncertainties are reported at the 1 σ confidence level. Reportedupper limits are those of the 2 σ confidence level. b All burst times of arrival are topocentric. Bursts with multiple components have one topocentric arrivaltime and several widths, fluences, and peak flux densities reported; the arrival time refers to the first sub-burst, and width, fluence, and peak flux density values for each component are presented in order of arrival. c Baseband data recorded for the burst. d From S/N-optimization. e No total-intensity or baseband data were recorded for this event. See Section 3.1 for details. f Fluence and peak-flux-density measurements represent lower bounds as we assumed all bursts were detectedalong the meridian of the primary beam. the DM uncertainty distribution and fitting a linear drift rate for each of 100 random noise realisationsper DM value. We found five bursts with significant drift rates (see Table 2).3.6.
Baseband Detections
The CHIME/FRB baseband system (see CHIME/FRB Collaboration et al. 2018) was triggeredfor one burst each from Sources 2, 4, 5, and 6. We used the baseband data for these events toverify the localization regions obtained from the per-beam S/N analysis described in Section 3.1;the results and details of future position refinement with baseband data will be presented elsewhere(Michilli et al., in prep.). Once available, this analysis will reduce the localization precision to0 F r e q u e n c y ( M H z ) F r e q u e n c y ( M H z ) F r e q u e n c y ( M H z ) -25 0 25Time (ms) -25 0 25Time (ms)500600700800 F r e q u e n c y ( M H z ) -25 0 25Time (ms) -25 0 25Time (ms) Figure 2.
Dynamic spectra of the bursts listed in Table 2, for the per-burst optimal DMs as determined in § ∼
100 ms), unless indicated otherwise by themultiplicative factor in the bottom right corner. Intensity values are saturated at the 5 th and 95 th percentiles. Allbursts were detected in the source’s upper transit. Pulse widths, defined as the width of the boxcar with the highestS/N after convolution with burst profile, are in the top right corner. The shaded region in the profile (four timesthe pulse width) was used for the extraction of the on-pulse spectrum. The shaded region in the on-pulse spectrumshows the full width at tenth maximum (FWTM) of a Gaussian fit. In the burst profiles, the black lines are theintegration over the FWTMs and the gray lines are the integration over the full bandwidths. 64 frequency subbandswith a 6.25 MHz subband bandwidth are shown for all bursts. There are underlying missing or masked channels atthe full-resolution (16,384-frequency-channel) intensity data, resulting in an average effective bandwidth of 214 MHz. ine New Repeaters from CHIME/FRB F r e q u e n c y ( M H z ) F r e q u e n c y ( M H z ) -25 0 25Time (ms) -25 0 25Time (ms)500600700800 F r e q u e n c y ( M H z ) -25 0 25Time (ms) -25 0 25Time (ms) Figure 3.
Same as Fig. 2. ∼ . S/N arcmin (CHIME/FRB Collaboration et al. 2018). Visual inspection of the baseband dynamicspectra confirmed successful capture of all four bursts by the baseband system. All baseband eventswere incoherently dedispersed to the S/N-optimising DM after coherently dedispersing each of the1,024 baseband channels to a nearby fiducial value, in order to mitigate intra-channel smearing.All four baseband events were manually processed through a polarization analysis pipeline. Thepipeline, described in an upcoming paper (Mckinven et al., in prep.), searches for an RM detectionusing the
RM-tools package that implements two independent methods: RM-synthesis (Burn 1966;Brentjens & de Bruyn 2005) and Stokes QU-fitting (O’Sullivan et al. 2012). Using these algorithms,we detected moderate RM values for Sources 2 and 6.Table 3 summarises the main polarization products for sources with an RM detection. RMs derivedfrom RM-synthesis (RM FDF ) and Stokes QU-fitting (RM
QUfit ) are reported, and show small butsignificant differences. RM
FDF is determined from the peak of the Faraday dispersion function (FDF),while RM
QUfit is calculated from a parametric fit under a thin-screen model of Faraday rotation. https://github.com/CIRADA-Tools/RM P A [ d e g ]
20 10 0 10 20Time [ms]420246810 S / N ILV 904504590 P A [ d e g ]
20 10 0 10 20Time [ms]0510152025 S / N IL Figure 4.
Pulse profiles (left to right: Source 2 and Source 6) for total intensity (I, black), linear polarization (L,red) after correcting for the detected RM, circular polarization (V, blue) and the uncalibrated polarization positionangle (upper panel). For Source 6, the linear polarized burst profile is a lower limit while the the circular polarizationprofile has been omitted to avoid misleading conclusions drawn from data that has yet to be robustly corrected forinstrumental leakage. See discussion in Section 4.2 for more details.
Uncertainties in RM
FDF were determined in a manner consistent with previous analysis of a differentburst presented in Paper I, while those for RM
QUfit were determined from its marginal probabilitydensity output by the Nested Sampling routine. The differences in estimates of RM
FDF and RM
QUfit are a more valid measure of the systematic RM measurement uncertainty than each method’s formalmeasurement error. The discrepant RMs are also a reflection of small differences in the measurementprocedure of the two methods. In particular, the Stokes Q,U fitting of
RM-tools implements aNested Sampling algorithm (Skilling 2004) to find the best-fitting parameters. This method allows thepartially degenerate parameters, RM and the de-rotated polarization angle ( χ ), to be simultaneouslyfit.Figures 5 and 6 summarise the general properties of the polarized signal for Sources 2 and 6, re-spectively: dynamic spectra for Stokes I, Q, U and V parameters where data have been downsampledto ∼ ∼ λ modulation consistentwith Faraday rotation. This leakage makes the uncorrected linear polarized fraction (L/I, withL = (cid:112) Q + U ) for Source 6 a lower bound, which we estimate to be L/I ≥ ine New Repeaters from CHIME/FRB Table 3.
Observed and Galactic-foreground RM values for CHIME/FRB repeaters.Source Day RM
FDF RM QUfit RM Oppermann+MW c RM GMIMSMW d (yymmdd) (rad m − ) (rad m − ) (rad m − ) (rad m − )2 190606 − ± − ± ± − ±
46 190702 − ± − ± ± − ± a b − ± b ... b − ± − ± a Source 1 in Paper I. b Data taken from Paper I. c Derived from the model developed by Oppermann et al. (2015). d Derived from the model developed by Dickey et al. (2019).4.
DISCUSSIONThe systematic monitoring of a large sky fraction by CHIME/FRB has enabled the discovery ofnine new repeating FRB sources, and a total of 18 repeating sources when including the results fromCHIME/FRB Collaboration et al. (2019b) and Paper I. The current sample of known repeating FRBsources now consists of 20 members, with FRBs 121102 (Scholz et al. 2016) and 171019 (Kumar et al.2019; Patel & the CHIME/FRB Collaboration 2019) discovered using other observatories. Besidesoffering opportunities for interferometric localization and multi-wavelength follow up (e.g., Marcoteet al. 2020), this number is large enough to enable studies of distributions of some source properties,although caution is needed as the CHIME/FRB pipeline has detection biases that remain to bequantified. 4.1.
Burst DMs and Morphologies
We compared the distribution of DMs of the 18 CHIME/FRB repeaters with that of the 12 publishedthus far non-repeating sources found by CHIME/FRB, and found no statistically significant differ-ence using either a Kolmogorov-Smirnov or Anderson-Darling test. Both distributions are subjectto similar (though not necessarily identical) detection biases. This fact is suggestive that both theirdistribution in space and their associated local environments do not differ strongly, unless the differ-ences cancel one another, which seems unlikely. Further comparison with additional CHIME/FRBsources that have not yet repeated (currently under analysis) will also be of interest.Paper I noted a statistically significant difference in repeater and thus far non-repeater burst widths.We repeated this analysis for this work with the new nine repeating sources included, using the 2 σ confidence upper limits where there are no significant measurements of the pulse width. As discussedin Section 3.5, two best-fit models were obtained for each burst in the CHIME/FRB sample – onethat directly models the effect of scattering and another that ignores scattering – and the model thatyielded a better goodness of fit statistic was chosen as the superior model. Therefore, all widths wereobtained in a similar manner and likely reflect the intrinsic burst widths when accounting for effectsfrom scattering and multiple burst components that is typical in repeating-FRB activity.Additionally, we included in the comparison only bursts with a detection S/N >
10, which was thethreshold for saving intensity data to disk at the time of detection of the non-repeating sources weconsidered. This criterion ensures that the two samples have an identical selection function, but omitsSource 3 from this analysis since all bursts from that source had S/N <
10. We compare temporalwidths of different Gaussian components of the bursts, including several measurements for any burst4 F r e q u e n c y [ M H z ] Stokes I Stokes V
20 0 20Time [ms]450500550600650 F r e q u e n c y [ M H z ] Stokes Q
20 0 20Time [ms]
Stokes U ]02468 P o l a r i z e d I n t e n s i t y [ S / N ] Q / L U / L
500 520 540 560 580 600 620 640 (MHz)904504590 ( d e g ) Figure 5.
Baseband data of the 190606 burst of Source 2. Top, left: Dedispersed dynamic spectra of Stokes I, Q,U & V Stokes parameters. Top, right: The cleaned FDF showing polarized signal in S/N units out to a maximumFaraday depth determined by a 50% drop in polarized sensitivity (upper panel) and the equivalent plot constrainedover the region surrounding the peak of the FDF (lower panel). Bottom: Modulation of Stokes Q and U parameterswith frequency, normalised by the total linear polarization (L). Frequency channels with significantly polarized signalare highlighted through a greyscale that saturates at higher S/N. The black curve is the best-fit model after applyinga parabolic fit to the Stokes I spectrum and modulating the Stokes Q, U using the best-fit Faraday rotation. Thelower panel shows the variation in the uncalibrated polarization angle ( ψ ) as a function of frequency along with thebest-fit model as a red line. with multiple components in the samples used for the comparison. These width distributions areshown in Figure 7. Using both the Kolmogorov-Smirnov and Anderson-Darling tests, we foundthat the two samples are not drawn from the same distribution with ∼ σ and ∼ σ significance, ine New Repeaters from CHIME/FRB F r e q u e n c y [ M H z ] Stokes I Stokes V
20 0 20Time [ms]600650700750800 F r e q u e n c y [ M H z ] Stokes Q
20 0 20Time [ms]
Stokes U ]05101520 P o l a r i z e d I n t e n s i t y [ S / N ] Q / L U / L
660 680 700 720 740 760 780 800 (MHz)904504590 ( d e g ) Figure 6.
Same plots as those in Figure 5 for baseband data of the 190702 burst of Source 6. respectively. When combining width measurements of different bursts (or their components) fromeach repeating source using inverse-variance weighting, in order to perform this comparison strictlybetween FRB sources, we also found that the difference persists and supports the notion of possiblydifferent emission mechanisms between repeating and thus far non-repeating FRBs. Alternatively,significant variation in pulse widths could also be caused by differences in the properties of thecircumburst media between repeating and thus far non-repeating FRBs. However, this is a lesslikely possibility considering that the two DM distributions do not have any statistically significantdifferences.As a population, the thus far non-repeaters reported in CHIME/FRB Collaboration et al. (2019a)show significant scattering, with half having scattering times of > ∼ ∼ σ confidence interval of a normal probabilitydensity distribution. A cumulative distribution function for a mixture distribution of the repeaterscattering timescales was generated by adding the probability density distributions for all sources.This cumulative distribution function was then compared with measured scattering times for thenon-repeaters using a Kolmogorov-Smirnov test. We found that we cannot rule out the possibility ofthe two samples being drawn from the same parent distribution. We will repeat this analysis once alarger sample of repeater bursts with statistically significant measurements of scattering timescalesis obtained.Paper I noted a statistically marginal, negative correlation between peak FRB flux and width basedon the repeating FRB sample published therein. We repeated this analysis, combining measurementspresented in Table 2 of the present work to the sample published in Paper I. We found no evidencefor correlation between peak fluxes and widths from this enlarged sample. However, as in Paper I,we note that various sources of selection bias have not yet been quantified, and future analyses ofsuch correlations may still be worthwhile.The 190213 detection of Source 3 at ∼ ∼
19 ms without an apparent “bridge“ in emission in time or frequency – similar to the 181019detection of Source 1 in Paper I. Whether a detection like this constitutes one burst comprised of twosub-bursts or two separate bursts remains to be seen; the interpretation is likely model-dependent.In Table 2, we reported new measurements of sub-burst linear drift rates. However, in contrast toPaper I, where almost half of bursts showed visible drift, only five of the bursts discussed in this workhave significant drift rates. The difference is most likely due to more bursts in the present samplehaving lower S/N than in Paper I, thus making significant detection of drifting sub-bursts moredifficult. The measured linear drift rates appear to be drawn from a similar distribution as in PaperI, and the notion that repeating sources of FRBs exhibit drifting sub-bursts with linear drift rates oforder few to tens of MHz/ms in the 400–800 MHz band remains valid. The apparent differences indrift rate uncertainties reported here and in Paper I are due to i.) a non-linear relation between themeasured rotation θ of a two-dimensional Gaussian and the drift rate df/dt = 1 / tan( − θ ), leadingto measurements of higher drift rates being more uncertain and ii.) the drift rate uncertainties ine New Repeaters from CHIME/FRB Figure 7.
Distribution of intrinsic temporal widths for repeating and thus far non-repeating FRB sources observedin the frequency range of 400–800 MHz. For repeating FRBs, the left panel shows the distribution of widths of theGaussian spectral components for all bursts from each source while the right panel shows only the weighted averageof the widths for each source. were determined after marginalising over DM, with the DM uncertainties of the bursts for which wemeasure drift rate here all being only 0.2–0.3 pc cm − .As CHIME/FRB total-intensity data have relatively coarse time resolution, it is possible that ap-parently simple bursts would show substructure and measurable drift if examined at higher timeresolution. This circumstance has been seen in bursts from FRB 121102 (Hessels et al. 2019). In ad-dition, the emission mechanism and/or local environments likely introduce burst-to-burst fluctuationsto the observed morphology.As with the sources in Paper I, we found that all bursts in this present sample of repeating sourcespossess small emission bandwidths of 100–200 MHz, similar to behavior seen in FRB 121102 (Gourdjiet al. 2019), whereas a large fraction of non-repeating FRBs span the full CHIME band. Giventhe large sample of repeat bursts obtained with CHIME/FRB, detected over a wide range of beampositions, it is likely that these small emission bandwidths reflect real, intrinsic differences in emissionmechanisms and/or local environments between repeating and non-repeating FRB sources. However,a robust statistical analysis of spectral features requires a sample of thus far non-repeating FRBslarger than that considered in Paper I.8 4.2. Implications of RM Values
Using CHIME/FRB baseband data, we measured RM values from two new repeating sources.These results are summarised in Table 3, along with with a previous RM detection from a differentCHIME/FRB repeater (Paper I; see Table 3). RM values reported are those derived from the StokesQU-fitting method, which were found to give marginally higher linear polarization fractions after de-rotation. We did not apply any redshift corrections to the values presented in Table 3 as their redshiftsare not currently known, and the RM contributions from local environments are not constrained.We also computed the RM contributions from the Milky Way foreground (RM MW ), which areprovided in Table 3. We used two methods to determine the Galactic contribution to RM: i) theconstructed RM foreground map of Oppermann et al. (2015), previously applied in Paper I; and ii) thefirst Faraday moment of diffuse, polarized emission obtained from the Global Magneto-Ionic MediumSurvey (GMIMS) (Dickey et al. 2019). The first moment is the mean of the Faraday spectrum,weighted by polarized intensity. Along most high-latitude sightlines the Faraday spectrum is simple,and the first moment is equivalent to the peak Faraday depth. Departures from this scenario arisewhen polarized emission is Faraday rotated by different amounts. Such departures are commonlyobserved in diffuse, polarized emission from the Galactic foreground, where its extended nature leadsto variable levels of Faraday rotation as a function of line-of-site distance. Cases such as this displaycomplex structure of the polarized emission in Faraday depth and can even appear as multiple peaks,preventing an accurate RM MW value from being inferred. This does not appear to be a problem here,with all repeater directions yielding relatively simple Faraday spectra, as is also the case for FRB121102 (Michilli et al. 2018).Significant differences can be seen between certain RM MW values when using the two methods.Notably, FRB180916’s position near the Galactic midplane likely complicates the interpretation ofthe first Faraday moment from GMIMS. This discrepancy occurs because sightlines at lower latitudesprobe a greater extent of the Galactic foreground emission. In low-latitude cases such as this, thefirst Faraday moment of the emission is unlikely to be an accurate proxy for RM MW . In fact, evenidealized scenarios with minimal path length through the foreground and no magnetic field reversalscan still produce systematic bias between the first Faraday moment of diffuse, polarized emissionand the true RM MW value (Ordog et al. 2019). A larger sample of extragalactic polarized sources,including FRBs, will be helpful in identifying well behaved regions of the sky where diffuse, polarizedemission can be leveraged to better subtract the foreground RM contribution.Regardless of the method used for subtracting the RM contribution of the Galactic foreground, aconsiderable gap ( > rad m − ) exists between RM values obtained for repeating sources observedby CHIME/FRBwith those associated with bursts from FRB121102. It is possible that depolariza-tion arising from intrachannel Faraday rotation greatly reduces sensitivity to polarized signal withextreme RMs, effectively preventing detection. At the native channelization of CHIME basebanddata, sensitivity to polarized signal drops significantly for RMs beyond ∼ rad m − , with theexact value depending on the specifics of the burst spectrum. However, even after accounting for thispossibility, repeater RMs discovered with CHIME appear rather modest compared to the effectiverange over which RM detections are possible. At the very least, these results demonstrate that RMis not a perfect discriminant between repeating and (apparently) non-repeating sources and furtherhighlight the distinct nature of FRB 121102. ine New Repeaters from CHIME/FRB ∼ >
20% of Source 6are largely consistent with the level of heterogeneity expected from the published sample. However,we urge caution when interpreting the fractional polarizations of Source 6 as significant instrumentaleffects lead to substantial mixing of the Stokes parameters, yielding misleading linear and circularpolarization fractions. However, these polarized instrumental effects are highly dependent on positionin the primary beam of the telescope and for Source 2 appear to be sub-dominant. Nonetheless,analysis techniques for correcting the differential polarized response across the primary beam arecurrently in development and will be reported elsewhere.The combined polarization information of these new repeating FRB sources will likely motivaterevision of existing FRB emission models, many of which use polarization properties of FRB 121102as key information in their initial construction. An example of such a case is the young magnetarmodel of Margalit & Metzger (2018), which posits a dense, ionized nebula around a central enginegiving rise to the high RM observed in FRB 121102. A variant of this model (Metzger et al. 2019) hassince been developed and found to be consistent with burst properties reported for the initial batch ofCHIME repeaters (see Paper I). Moreover, this model allows for a range of RM values, dependent onthe age of the magnetar and the specifics of its formation channel (Margalit et al. 2019). Accordingto this scenario, the RM should decay monotonically with time (Piro & Gaensler 2018) giving riseto a trend of younger, more active sources with higher RMs and older, less active sources with lowerRMs. The fact that CHIME seems to preferentially observe repeaters with relatively low RM valuesis at odds with this prediction but can possibly be explained by invoking a concomitant evolution ofthe peak flux frequency as a function of age, such that CHIME/FRB tends to detect older repeatingsources (Kumar et al. 2017; Metzger et al. 2019). The detection of only one burst from FRB121102in the CHIME band supports this assertion, but a much larger sample of repeater RMs over a widerange of bandwidths will be needed to make any robust claims.Alternatively, it is also possible that FRB 121102’s large RM is not from an associated nebulabut from highly magnetized environment commonly found in dense, star-forming regions and/oraround massive accreting black holes (see discussion in Michilli et al. 2018), as in the case of theGalactic-center magnetar (Desvignes et al. 2018). In this framework, the circumstance of FRB121102 residing in such environments can be due purely to chance, or is perhaps a reflection of theprecursor’s preference for these environments. In either case, the observed RM is independent of thecentral FRB engine driving emission and therefore not directly correlated with properties relatingto its emission. This interpretation can be tested in the future with a much larger RM samplecombined with interferometric localization and multi-wavelength observations capable of probing thelocal environment.0In addition, multi-epoch observations of newly discovered repeaters may provide evidence for secularevolution of the RM similar to that seen for FRB 121102 (Michilli et al. 2018; Gajjar et al. 2018). Theexistence of such a trend would be difficult to reconcile in this framework and would more stronglysupport emission models that explicitly link evolution of local environment with the FRB engineitself. 4.3.
Repetition Rates
We calculated the repetition rates of these repeaters and compared them to the limits on singlebursts observed by the Australian Square Kilometre Array Pathfinder (ASKAP; Shannon et al. 2018).Note that a repeat burst was found from FRB 171019, one of the sources from Shannon et al. (2018),in follow up observations at 820 MHz with the Green Bank Telescope (Kumar et al. 2019). In orderto compare the repetition rates across different fluence limits, we scale the observed rates (based onthe detections and exposure in Table 1) by S . where S is the fluence sensitivity (also specified inTable 1) and the exponent of 1.5 reflects an assumed Euclidean spatial distrbution of sources. Wecalculated the rates for upper and lower transits separately for sources that are circumpolar in theCHIME/FRB field of view. Figure 8 shows the observed and scaled rates of repetition for repeatersfrom Paper I, this work, and from Kumar et al. (2019) as well as scaled upper limits on repetitionrates for the single bursts reported by Shannon et al. (2018). We found that most of the repetitionrates for repeaters are at or lower than the 1- σ upper limits from the ASKAP observations. Thescaled repetition rate of Source 5 is marginally higher than most of the upper limits from ASKAPobservations. 4.4. Multi-wavelength Follow-up
We checked for any catalogued ionized regions (Anderson et al. 2014; Green 2019) or star-formingregions (Avedisova 2002; Rice et al. 2016) within the Milky Way galaxy that are coincident with thelocalization areas of the 9 new repeaters presented in Table 1, and found none. We also estimated themaximum redshift ( z max ) limit for all FRBs to identify plausible galaxy candidates. To estimate theMilky Way (MW) contribution to the observed DMs, we considered the smaller of NE2001 (Cordes& Lazio 2002) and YMW16 models’ (Yao et al. 2017) predicted Galactic DM values, and added aMW halo contribution of 50 pc cm − (Prochaska & Zheng 2019). We then subtracted this value fromthe observed DMs to estimate the extragalactic DM contribution. These excess DMs were convertedto z max by using the DM-redshift relation: DM ex ≈ z (Zhang 2018b). These redshift estimatesare approximate upper limits as we did not account for the DM of the host galaxy. Using these z max values, we searched various catalogues of nearby galaxy clusters (Abell et al. 1989; B¨ohringer et al.2000; Wen et al. 2017; Hao et al. 2010) and found none within the localization regions of the FRBs.As our repeating FRBs have poor localization and large estimated z max , the chance coincidenceprobability of finding even a massive star-forming galaxy is large. This circumstance remains truefor Source 4 despite possessing the lowest DM excess among our repeating FRB sample, DM ex ≈
115 pc cm − . However, the largest RM excess in our sample, | RM ex | ≈
490 rad m − for Source6, likely suggests considerable host DM contribution and, hence, a nearby host galaxy. This factprompted us to look for plausible host galaxy candidates around Source 6, which has the secondlowest extragalactic DM among Sources 1-9 (DM ex ≈
140 pc cm − ). We estimated z max ≈ ine New Repeaters from CHIME/FRB P F R B P F R B L P F R B P F R B P F R B P F R B P F R B U P F R B F R B S S S U S L S U S L S S S S S P o i ss o n R a t e ( h r ) Observed RateScaled Rate
Figure 8.
Repetition rates of repeaters and upper limits for single bursts from ASKAP (gray lines; Shannonet al. 2018). Observed rates are denoted as black circles, while rates scaled by fluence sensitivity are shownas red triangles. Repeaters from Paper I are labelled “P1”. Repeaters from Table 1 are labelled as “S x ”. Therepetition rate for FRB 171019 is calculated from the GBT 820 MHz observations (Kumar et al. 2019). Therates (red triangles) and upper limits are scaled to a fluence limit of 1 Jy–ms using a scaling of S . where S is the sensitivity of the search. The “U” and “L” suffixes for CHIME/FRB-detected repeaters denotethe rates calculated from number of detections, exposure and sensitivity in the upper and lower transits,respectively, for circumpolar sources. 1- σ error bars are shown assuming a Poisson distribution. possess the lowest z max among the catalogued galaxies within the localization region of Source 6, andeither one can easily account for the observed excess DM. Using the luminosity function of massivegalaxies from Faber et al. (2007), we estimate the density of galaxies with M B < −
20 to be ∼ − . By assuming a pair merger fraction = 0.01 (Bell et al. 2006), we estimated the probabilityof finding a pair of massive merging galaxies by chance to be <
5% within the Source 6 containmentregion for z max = 0.16. Therefore, we consider these galaxies to be interesting candidates for the hostof Source 6. CONCLUSIONSWe have reported on the discovery of nine new repeating FRB sources from CHIME/FRB. Multiplebursts from these nine sources were collected during a ∼ σ level, with the repeating-sourcepopulation producing larger widths. As first discussed in Paper I, this observation likely indicatesintrinsic differences between emission mechanisms and/or environments local to the sources thatproduce repeating and thus far non-repeating bursts. Future studies of these properties from the ∼
700 bursts observed by CHIME/FRB will place stronger constraints on these statistical differencesand are forthcoming.Analysis of baseband data acquired for two of these sources yielded significant estimates of Faradayrotation, with RM = − − for Source 2 and RM = − − for Source 6.The large RM difference between bursts associated with FRB 121102 and those reported here fur-ther indicates a uniqueness in the local magnetized environment of FRB 121102. Furthermore, therelatively modest RMs discovered with CHIME/FRB do not yet indicate clear differences with thenon-repeating sample. This suggests that RM is perhaps a poor discriminant for repeating and thusfar non-repeating FRB sources, but any robust statistical claim requires a larger sample. When com-paring to expected Galactic contributions, the low excess DM and large excess RM for Source 6 aresuggestive a potentially nearby host galaxy with possible star formation; we searched for candidatesand identified a pair of merging galaxies at redshift z = 0 .
064 as a tentative candidate host galaxy.A more precise localization of Source 6 is needed to confirm or exclude this tentative hypothesis.To date, CHIME/FRB has discovered a total of 18 repeating FRB sources (see CHIME/FRB Col-laboration et al. 2019b, Paper I and this work) out the 20 currently known, with two other repeatingsources found using the Arecibo Observatory (Spitler et al. 2016) and ASKAP (Kumar et al. 2019).Each source presents an opportunity for arcsecond localization with large telescope arrays and, onceachieved, multi-wavelength studies of the host galaxy and any associated intra-galactic environment.Indeed, interferometric follow-up of several CHIME/FRB repeating sources are underway with the realfast backend at the VLA (Law et al. 2018) and the European VLBI Network (EVN; Marcoteet al. 2017); the EVN recently localized FRB 180916.J0158+65, discovered by CHIME/FRB (PaperI), with mas precision to a star-forming region in a nearby massive spiral galaxy (Marcote et al.2020). We encourage additional community involvement, both in localizing these sources and target-ing localized sources for constraining emission and activity across the electromagnetic spectrum.We thank the Dominion Radio Astrophysical Observatory, operated by the National ResearchCouncil Canada, for gracious hospitality and useful expertise. The CHIME/FRB Project is fundedby a grant from the Canada Foundation for Innovation 2015 Innovation Fund (Project 33213), as wellas by the Provinces of British Columbia and Qu´ebec, and by the Dunlap Institute for Astronomyand Astrophysics at the University of Toronto. Additional support was provided by the CanadianInstitute for Advanced Research (CIFAR), McGill University and the McGill Space Institute viathe Trottier Family Foundation, and the University of British Columbia. The Dunlap Instituteis funded by an endowment established by the David Dunlap family and the University of Toronto.Research at Perimeter Institute is supported by the Government of Canada through Industry Canadaand by the Province of Ontario through the Ministry of Research & Innovation. The National ine New Repeaters from CHIME/FRB
Abell, G. O., Corwin Jr, H. G., & Olowin, R. P.1989, The Astrophysical Journal SupplementSeries, 70, 1Alam, S., Albareti, F. D., Prieto, C. A., et al.2015, The Astrophysical Journal SupplementSeries, 219, 12Anderson, L. D., Bania, T., Balser, D. S., et al.2014, The Astrophysical Journal SupplementSeries, 212, 1Avedisova, V. 2002, Astronomy Reports, 46, 193Bannister, K. W., Deller, A. T., Phillips, C., et al.2019, Science, 365, 565Bassa, C., Tendulkar, S., Adams, E., et al. 2017,The Astrophysical Journal Letters, 843, L8Bell, E. F., Phleps, S., Somerville, R. S., et al.2006, The Astrophysical Journal, 652, 270Beloborodov, A. M. 2017, ApJL, 843, L26B¨ohringer, H., Voges, W., Huchra, J., et al. 2000,The Astrophysical Journal Supplement Series,129, 435Brentjens, M. A., & de Bruyn, A. G. 2005, A&A,441, 1217Burn, B. J. 1966, MNRAS, 133, 67Chatterjee, S., Law, C. J., Wharton, R. S., et al.2017, Nature, 541, 58CHIME/FRB Collaboration, Amiri, M., Bandura,K., et al. 2018, ApJ, 863, 48—. 2019a, Nature, 566, 230 —. 2019b, Nature, 566, 235CHIME/FRB Collaboration, :, Andersen, B. C.,et al. 2019c, ApJL, 885, L24Cordes, J. M., & Lazio, T. J. W. 2002, arXivpreprint astro-ph/0207156Desvignes, G., Eatough, R. P., Pen, U. L., et al.2018, ApJL, 852, L12Dickey, J. M., Landecker, T. L., Thomson, A.J. M., et al. 2019, ApJ, 871, 106Dolag, K., Gaensler, B. M., Beck, A. M., & Beck,M. C. 2015, MNRAS, 451, 4277Faber, S., Willmer, C., Wolf, C., et al. 2007, TheAstrophysical Journal, 665, 265Gajjar, V., Siemion, A. P. V., Price, D. C., et al.2018, ApJ, 863, 2Gourdji, K., Michilli, D., Spitler, L. G., et al.2019, ApJL, 877, L19Green, D. 2019, Journal of Astrophysics andAstronomy, 40, 36Hao, J., McKay, T. A., Koester, B. P., et al. 2010,The Astrophysical Journal Supplement Series,191, 254Heald, G. 2009, in IAU Symposium, Vol. 259,Cosmic Magnetic Fields: From Planets, to Starsand Galaxies, ed. K. G. Strassmeier, A. G.Kosovichev, & J. E. Beckman, 591–602Hessels, J. W. T., Spitler, L. G., Seymour, A. D.,et al. 2019, ApJL, 876, L23 Josephy, A., Chawla, P., Fonseca, E., et al. 2019,arXiv e-prints, arXiv:1906.11305Kumar, P., Lu, W., & Bhattacharya, M. 2017,MNRAS, 468, 2726Kumar, P., Shannon, R. M., Os(cid:32)lowski, S., et al.2019, ApJL, 887, L30Law, C. J., Bower, G. C., Burke-Spolaor, S., et al.2018, in Astronomical Society of the PacificConference Series, Vol. 517, Science with a NextGeneration Very Large Array, ed. E. Murphy,773Lyubarsky, Y. 2014, MNRAS, 442, L9Lyutikov, M. 2019, arXiv e-prints,arXiv:1909.10409Marcote, B., Nemo, K., Hessels, J. W. T., & et al.2020, NatureMarcote, B., Paragi, Z., Hessels, J. W. T., et al.2017, ApJL, 834, L8Margalit, B., Berger, E., & Metzger, B. D. 2019,ApJ, 886, 110Margalit, B., & Metzger, B. D. 2018, ApJL, 868,L4Mase, S. 1992, Annals of the Institute ofStatistical Mathematics, 44, 479.https://EconPapers.repec.org/RePEc:spr:aistmt:v:44:y:1992:i:3:p:479-499Masui, K., Lin, H.-H., Sievers, J., et al. 2015,Nature, 528, 523McKinnon, M. M. 2014, PASP, 126, 476Metzger, B. D., Margalit, B., & Sironi, L. 2019,arXiv e-prints, arXiv:1902.01866Michilli, D., Seymour, A., Hessels, J. W. T., et al.2018, Nature, 533, 132Murase, K., Kashiyama, K., & M´esz´aros, P. 2016,MNRAS, 461, 1498Oppermann, N., Junklewitz, H., Greiner, M., et al.2015, Astronomy & Astrophysics, 575, A118Oppermann, N., Junklewitz, H., Greiner, M.,et al. 2015, A&A, 575, A118Ordog, A., Booth, R., Van Eck, C., Brown, J.-A.,& Landecker, T. 2019, Galaxies, 7, 43 O’Sullivan, S. P., Brown, S., Robishaw, T., et al.2012, MNRAS, 421, 3300Patel, C., & the CHIME/FRB Collaboration.2019, The Astronomer’s Telegram, 13013, 1Piro, A. L., & Gaensler, B. M. 2018, ApJ, 861, 150Platts, E., Weltman, A., Walters, A., et al. 2018,arXiv e-prints, arXiv:1810.05836Prochaska, J. X., & Zheng, Y. 2019, MonthlyNotices of the Royal Astronomical Society, 485,648Prochaska, J. X., Macquart, J.-P., McQuinn, M.,et al. 2019, Science, 366, 231Ravi, V., Catha, M., D’Addario, L., et al. 2019,Nature, 572, 352Rice, T. S., Goodman, A. A., Bergin, E. A.,Beaumont, C., & Dame, T. M. 2016, TheAstrophysical Journal, 822, 52Sandell, D. 1991, The Mathematical Scientist, 16,78Scholz, P., Spitler, L. G., Hessels, J. W. T., et al.2016, ApJ, 833, 177Scholz, P., Bogdanov, S., Hessels, J. W. T., et al.2017, ApJ, 846, 80Shannon, R. M., Macquart, J.-P., Bannister,K. W., et al. 2018, Nature, 562, 386Skilling, J. 2004, in American Institute of PhysicsConference Series, Vol. 735, American Instituteof Physics Conference Series, ed. R. Fischer,R. Preuss, & U. V. Toussaint, 395–405Spitler, L. G., Scholz, P., Hessels, J. W. T., et al.2016, Nature, 531, 202Tendulkar, S. P., Bassa, C. G., Cordes, J. M.,et al. 2017, ApJL, 834, L7Wen, Z., Han, J., & Yang, F. 2017, MonthlyNotices of the Royal Astronomical Society, 475,343Yao, J. M., Manchester, R. N., & Wang, N. 2017,ApJ, 835, 29Zhang, B. 2018a, The Astrophysical JournalLetters, 854, L21—. 2018b, The Astrophysical Journal Letters, 867,L21 ine New Repeaters from CHIME/FRB A. CHANCE COINCIDENCE PROBABILITIES WITH LARGE NUMBERS OF FRBSOnce a large sample of FRBs has been detected, the probability of identifying two bursts with similarDM and sky location can become non-negligible. In such circumstances there are two questions thatmust be answered:1. Given N FRBs detected by a survey, what is the probability that any two FRBs will lie withinthe same phase-space bin of dimension (∆DM , ∆R . A . cos(Dec . ) , ∆Dec . )? We refer to this asthe “global probability.”2. If we detect two FRBs in the same (∆DM , ∆R . A . cos(Dec . ) , ∆Dec . ) bin during our survey,what is the probability that they are physically unrelated? We refer to this as the “individualprobability.”Here we answer both questions for all CHIME/FRB repeaters found to date (i.e., from this workand from Paper I), given that CHIME/FRB has detected ∼
700 FRBs during the observing periodmentioned in Section 2. A detailed analysis of the ∼
700 FRBs is underway and will be publishedelsewhere. Here we limit our discussion to the declinations and DM distribution of the FRBs in theCHIME/FRB survey; we marginalised over the R.A. dimension for reasons discussed below.A.1.
Global probability
The first question is mathematically equivalent to the non-uniform birthday problem, i.e., what isthe chance that two people in a group of N share the same birthday. The uniform case, in whicheach birthday (or each ∆DM , ∆R . A . cos(Dec . ) , ∆Dec . bin) are equally likely, is straightforward tosolve. The solution for the non-uniform case has been described in terms of recursion relations fora coincidence of two among N objects (Mase 1992), and for a general case of m among N objects(Sandell 1991). However, for the case of CHIME/FRB, the parameter space is large enough thatthe recursion formulae and their approximations are not computationally tractable. Instead, wesimulated the current results of the CHIME/FRB survey using Monte Carlo methods in order tounderstand how often we misidentify repeaters with a given criteria.Following the method in Appendix A of Paper I we estimated the detection probability distributionfunction from the ∼
700 FRBs found by the CHIME/FRB survey. This calculation incorporatedvariations in exposure times, sensitivity at different declinations and DMs. We assumed that theprobability of detection is independent of R.A. and local time since the survey has been operating forover a year and any variation in R.A. would have been averaged out. The probability distribution wassmoothed by a Gaussian kernel with a 350 pc cm − DM scale and 10 ◦ angular scale in declination.We sampled 700 independent FRBs from this probability distribution for 10 trials and deter-mined how many FRBs would be counted as repeaters with the following criteria: I) ∆DM <
10 pc cm − , ∆RA cos(Dec . ) < , ∆Dec . < < − , ∆R . A . cos(Dec . ) < , ∆Dec . < ∼
40% of the trials yield at leastone set of independent FRBs falsely identified as coming from the same repeating source. With thesecond criterion, the probability is ∼ Number of chance coincidences P r o b a b ili t y DM < 10 pc cm Number of chance coincidences P r o b a b ili t y DM < 1 pc cm Figure 9.
Histogram of the MC trials as a function of the number of detected chance coincidences. Bothtrials had the same localization criteria, ∆R . A . cos(Dec . ) < , ∆Dec . < <
10 pc cm − , and the right panel displays a DM criterion of ∆DM < − . Astricter DM cutoff allows for fewer chance coincidence detections when sampling 700 independent FRBs. We then determined the selection criteria to use, such that our repeaters have a chance coincidenceprobability ( p CC ) of ∼ ∼ < . − , and ii) given a DM tolerance of 10 . − , thespatial localization has to be ∆R . A . cos(Dec . ) < . , ∆Dec . < . p CC . Additionally, the results of our simulations imply thatall future repeating CHIME/FRB candidates must have extremely spatially well-localized bursts tokeep our chance coincidence probabilities low. A source with minimal spatial uncertainties wouldalso allow for more DM evolution across the bursts, which has been seen in FRB 121102 at the ∼ Individual Probabilities
The above simulations do not identify which bursts might be misidentified as repeaters. If p is theprobability of detecting a burst in the (∆DM , ∆R . A . cos(Dec . ) , ∆Dec . ) neighborhood of a knownburst, the probability that one of the n subsequent bursts will be in the same bin by coincidence is1 − (1 − p ) n . Since these repeaters are identified from a population of N ≈
700 FRBs, we use a trialsfactor of 700 to calculate the final coincidence probabilities.Here, we considered only the repeaters with two detected bursts, since the chance coincidenceprobabilities for three or more bursts occurring within the same phase-space bin are negligible forour current sample size. We started with the probability distribution above, normalised over thefull sky and DM range. Table 4 specifies the candidate repeaters with two bursts and the range of∆DM , ∆R . A . cos(Dec . ) , and ∆Dec . for the detected bursts. For Sources 1 and 2, the trials-adjustedchance coincidence probability ( N p CC ) remains below the 0.1% level, indicating robust associationof repeat bursts despite the large CHIME/FRB sample. However, N p CC = 10% for Source 8, which ine New Repeaters from CHIME/FRB Table 4.
Chance Coincidence Probabilities for Repeating FRB Candidates with Two BurstsSource Name a ∆R . A . cos(Dec . ) b ∆Dec. c ∆DM d n e int p CC N trials p CC (deg) (deg) (pc cm − )1 190208.J1855+46 0.5 0.5 1 101 7 . × − × − . × − × − . × − a As defined in Table 1. b Maximum of beam radius (0.5 ◦ ) or R.A. spread of the bursts (as denoted in Figure 1). c Maximum of beam radius (0.5 ◦ ) or Dec. spread of the bursts. The beam area is calculated as π ∆R . A . cos(Dec . ) × ∆Dec. d Maximum of 1 pc cm − or DM spread of the bursts. e Number of FRBs detected in the interval between the two bursts. makes this association statistically less significant. Future detections of bursts from Source 8 willconfirm its status as a repeater. B. EXPOSURE ESTIMATIONThe timeline of the exposure of the CHIME/FRB system to each of the sources presented in thiswork is plotted in Figure 10. The exposure is calculated for the transit of each source across theFWHM region of the synthesized beams at 600 MHz and includes transits in the interval from 2018August 28 to 2019 September 30. We excised transits from the reported exposure for which the RMSnoise (shown in Figure 10) was different by more than one standard deviation from the mean RMSnoise in the above-mentioned interval. The fraction of excised transits averaged about 5% for eachsource.The reported exposures only include intervals in each transit for which the detection pipeline wasfully operational. The reduction in daily exposure for Sources 1 and 6 for a period of several monthscan thus be attributed to the failure of the computing node designated to process data for one of thefour synthesised beams through which these sources transit. Sources 3 and 4 have declinations greaterthan +70 ◦ thereby allowing both upper and lower transits to be observable with the CHIME/FRBsystem. However, there were no bursts detected during the lower transit of either of these sources.Source 3 is particularly interesting since a significant fraction of its 90% confidence localization regionis located between the FWHM regions of two synthesized beams during the upper transit (see Figure1). Therefore, for this source, all allowed sky locations which transit between the two beams have zeroexposure. This circumstance results in an average exposure for Source 3 over the entire positionaluncertainty region of 55 ±
52 hours, despite its high declination.8 Source 1
Source 2 E x p o s u r e T i m e ( m i n s ) Source 3
Upper TransitLower Transit
Source 4
Upper TransitLower Transit
Source 5
Date R e l a t i v e r m s n o i s e N u m b e r o f P u l s a r s Figure 10.
Timeline of CHIME/FRB’s daily exposure to the new repeating FRB sources for upper andlower transits, if observable. Days on which a burst was detected are indicated by solid lines. The errorson the exposure are due to uncertainties in the source positions. The increase in exposure time from itstypical value for some of the days is due to the occurrence of two transits in the same solar day caused bythe length of a solar and a sidereal day being slightly different. The RMS noise is estimated using pulsarsdetected by CHIME/FRB for each sidereal day for which the telescope was operating with the same gaincalibration. The RMS noise for each pulsar is measured relative to the median over all days the pulsar wasdetected with the weighted average of measurements from several pulsars (the number of which is denotedby the marker colors) plotted here. ine New Repeaters from CHIME/FRB Source 6
Source 7 E x p o s u r e T i m e ( m i n s ) Source 8
Source 9