Discovery of a Radio Relic in the Massive Merging Cluster SPT-CL 2023-5535 from the ASKAP-EMU PILOT SURVEY
Kim HyeongHan, M. James Jee, Lawrence Rudnick, David Parkinson, Kyle Finner, Mijin Yoon, Wonki Lee, Gianfranco Brunetti, Marcus Brüggen, Jordan D. Collier, Andrew M. Hopkins, Michał J. Michałowski, Ray P. Norris, Chris Riseley
DDraft version July 21, 2020
Preprint typeset using L A TEX style emulateapj v. 12/16/11
DISCOVERY OF A RADIO RELIC IN THE MASSIVE MERGING CLUSTER SPT-CL 2023-5535 FROM THEASKAP-EMU PILOT SURVEY
Kim HyeongHan , M. James Jee , Lawrence Rudnick , David Parkinson , Kyle Finner , Mijin Yoon , WonkiLee , Gianfranco Brunetti , Marcus Br¨uggen , Jordan D. Collier , Andrew M. Hopkins ,Micha(cid:32)l J. Micha(cid:32)lowski , Ray P. Norris , Chris Riseley Draft version July 21, 2020
ABSTRACTThe ASKAP-EMU survey is a deep wide-field radio continuum survey designed to cover the entiresouthern sky and a significant fraction of the northern sky up to +30 ◦ . Here, we report a discoveryof a radio relic in the merging cluster SPT-CL 2023-5535 at z = 0 .
23 from the ASKAP-EMU pilot300 sq. deg survey (800 − ∼ ∼ . α = − . ± . M = 1 . ± . × M (cid:12) ) andcomposed of at least three subclusters. We suggest a scenario, wherein the radio features arise fromthe collision between the eastern and middle subclusters. Our discovery illustrates the effectivenessof the ASKAP-EMU survey in detecting diffuse emissions in galaxy clusters and when completed, thesurvey will greatly increase the number of merging cluster detections with diffuse radio emissions. Keywords: galaxies: clusters : individual (SPT-CL 2023-5535), gravitational weak lensing, radio con-tinuum: radio relic and halo, X-rays:galaxies:clusters INTRODUCTION
Large-scale diffuse radio emissions provide critical in-formation for understanding galaxy cluster mergers.They can be classified into two broad categories: ha-los and relics. Radio halos are diffuse sources withoutdistinct optical counterparts and are found in the cen- Yonsei University, Department of Astronomy, Seoul, Republicof Korea; [email protected], [email protected] Department of Physics, University of California, Davis,California, USA Minnesota Institute for Astrophysics, University of Minnesota,Minneapolis, Minnesota, USA Korea Astronomy and Space Science Institute 776,Daedeokdae-ro, Yuseong-gu, Daejeon, Republic of Korea Ruhr-University Bochum, Astronomical Institute, GermanCentre for Cosmological Lensing, Universittsstr. 150, 44801Bochum, Germany Istituto Nazionale di Astrofisica, Istituto di RadioastronomiaVia P Gobetti 101, 40129 Bologna, Italy Hamburger Sternwarte, Universit¨at Hamburg, Gojenbergsweg112, 21029 Hamburg, Germany Inter-University Institute for Data Intensive Astronomy,Department of Astronomy, University of Cape Town, Private BagX3, Rondebosch, 7701, South Africa School of Science, Western Sydney University, Locked Bag1797, Penrith, NSW 2751, Australia Australian Astronomical Optics, Macquarie University, 105Delhi Rd, North Ryde, NSW 2113, Australia Astronomical Observatory Institute, Faculty of Physics,Adam Mickiewicz University, ul. S(cid:32)loneczna 36, 60-286 Pozna´n,Poland Western Sydney University, Locked Bag 1797, Penrith South,NSW 1797, Australia CSIRO Astronomy & Space Science, PO Box 76, Epping,NSW 1710, Australia Dipartimento di Fisica e Astronomia, Universit`a degli Studidi Bologna, via P. Gobetti 93/2, 40129 Bologna, Italy INAF – Istituto di Radioastronomia, via P. Gobetti 101,40129 Bologna, Italy CSIRO Astronomy and Space Science, PO Box 1130, Bentley,WA 6102, Australia tral regions of merging clusters. Radio relics also do nothave optical counterparts, but they are located in thecluster periphery and in general possess high levels ofpolarization. Although both radio halos and relics areindicators of cluster merger activities, radio relics havebeen considered a stronger constraint on the merger his-tory because they can be used as direct probes of mergershocks (see reviews for Ferrari et al. 2008; Feretti et al.2012; van Weeren et al. 2019).The kinetic energy that is dissipated during cluster-cluster mergers can power the observed cluster-scale ra-dio emission. However, the complex chain of physi-cal mechanisms that leads to the acceleration of emit-ting particles and amplification of magnetic fields inthe ICM are still poorly understood (e.g., Brunetti &Jones 2014). Giant radio halos are thought to originatefrom stochastic re-acceleration induced by cluster merg-ers turbulence (e.g., Brunetti et al. 2001; Petrosian 2001;Brunetti & Larzarian 2007; Miniati et al. 2015); the con-tribution from secondary particles generated by the chainof hadronic collisions in the ICM has also been exploredin the past (e.g., Dennison 1980; Blasi & Colafrancesco1999) and more recently in combination with turbulentreacceleration models (e.g., Brunetti & Lazarian 2011;Pinzke et al. 2017). Radio relics are believed to origi-nate from merger shocks (e.g., En β lin et al. 1998). Theoriginal approach was based on the diffusive shock accel-eration (DSA; Bell 1978; Drury 1983; Malkov & Drury2001) of thermal electrons. However, the efficiency of ac-celeration at weak shocks in the ICM appears too low toreproduce the spectrum and luminosity of a large fractionof the observed radio relics (e.g., Botteon et al. 2020).One of the most popular modifications of this scenario isbased on shock re-acceleration of pre-existing relativisticplasma (e.g., Kang & Ryu 2011; Pinzke et al. 2013; Kang a r X i v : . [ a s t r o - ph . H E ] J u l HyeongHan et al. & Ryu 2016), which has been supported in some casesby the connection between radio relics and AGNs (e.g.,Bonafede et al. 2014; van Weeren et al. 2017).In order to better understand the origin of radio ha-los and relics, a current priority should be to increasethe sample size. To date, there are ∼
60 known radiorelics. Because of the cluster-to-cluster variation, the ex-isting sample is too small to enable studies, where onecan extract overarching principles. Among the upcomingconcerted efforts, presently the Australian Square Kilo-metre Array Pathfinder -Evolutionary Map of the Uni-verse (ASKAP-EMU; Norris et al. 2011) is the largestdeep ( ∼ µ Jy/beam), high-resolution ( ∼ (cid:48)(cid:48) ) radio con-tinuum survey designed to cover the entire southern skyand a significant fraction of the northern sky up to +30 ◦ .One of the scientific goals of the project is to enlarge thesample of clusters with diffuse radio emissions by at leasttwo orders of magnitude.In this study, we report the discovery of a radio relicin the massive merging cluster SPT-CL 2023-5535 (here-after CL2023 for brevity) at z = 0 .
23. The presenceof the diffuse radio emission in CL2023 has been re-ported in Zheng et al. (in prep.), who used the Murchi-son Widefield Array (MWA; Tingay et al. 2013), theAustralia Telescope Compact Array (ATCA; Frater etal. 1992), and the Molonglo Observatory Synthesis Tele-scope (MOST; Mills 1981; Robertson 1991) data. How-ever, the insufficient spatial resolution and the severalbright neighboring radio point sources have preventedthe earlier work from clearly resolving the halo and relic.In this paper, we also present our weak-lensing (WL) andX-ray analyses of CL2023 based on the archival Dark En-ergy Camera (DECam; Flaugher et al. 2015) and Chan-dra data, respectively, which enhance interpretation ofthe current discovery.We adopt a ΛCDM cosmology with H = 70km s − Mpc − , Ω m = 0 .
3, and Ω Λ = 0 .
7. The angu-lar size of 1 (cid:48) corresponds to a length scale of ∼
223 kpc atthe cluster redshift z = 0 . OBSERVATIONS
Radio
ASKAP has 36 antennae, 34 of which are placed withina region of 2.3 km diameter while the outer four extendthe baselines up to ∼ ∼
30 sq. degrees.Each PAF consists of 192 dual-polarization receivers. Aweighted sum of the outputs of groups of receivers form30 beams. Individual receivers, in general, contribute tomore than one beam. Therefore, adjacent beams are notcompletely independent. The 30 beams together coveran area of ∼
30 sq. degrees on the sky. https://astronomy.swin.edu.au/research/utmost/?page id=32 Table 1
DECam observationFilter Date t exp Seeing m lim (s) (arcsec) g r i Note . — 1. This is the 5 σ limiting magnitude for point sources. The weights of the individual beams are initially cal-ibrated by observing the Sun placed successively at thecenter of each beam, and then adjusting the weights formaximum signal-to-noise. However, a radiator at thevertex of each antenna or the On-Dish Calibrator (ODC)enables the gain of each receiver to be calibrated. As aresult, the solution initially obtained from the Sun obser-vation is modified using the ODC calibration and usedto adjust the weights.Before (or sometimes after) the observation of each tar-get, the calibrator source 1934-638 is observed for 200 sat the center of each of the 30 beams, to provide bandpassand gain calibration. No further calibration, other thanself-calibration, is performed during the observation.The radio data used in this paper were taken in 2019July from the ASKAP-EMU Pilot Survey, based on 10 hintegration (a rms noise level of 25 − µ Jy/beam)for Scheduling Block 9351, with a frequency range of800-1088 MHz. The reduction was performed with the
ASKAPsoft pipeline, using a multi-scale CLEAN algo-rithm and two Taylor terms (T0 and T1), which allowproduction of maps at a fiducial frequency of 943 MHz(T0) and the corresponding spectral indices (T1/T0). Amore extensive description of the Pilot Survey will beprovided by Norris et al. (in prep.).In this paper we present images from both the original(Figure 1A) and diffuse-enhanced (Figure 1B) versions.The latter was created by first masking out bright ( > (cid:48)(cid:48) Gaus-sian kernel, and finally combining the smoothed image(purple) with the original image (green). The resultingimage (Figure 1B) makes it easy to visually separate dif-fuse emissions from compact sources.
Optical
CL2023 was observed with the DECam mounted onthe 4-meter Blanco telescope at the Cerro Tololo Inter-American Observatory (PI: von der Linden). Table 1summarizes the observations for the g , r , and i filtersthat we retrieved from the NOAO archive for the cur-rent study. The Community Pipeline (Valdes et al. 2014)is used for the basic data reduction (i.e., overscan, bias,flat, etc.). The calibrated images were stacked into asingle mosaic image for each filter using SCAMP and SWARP . We used the i -band image for our WL analysisbecause it provides the sharpest point spread function(PSF). Intermediate PSF models were constructed for http://archive1.dm.noao.edu/ PT-CL 2023-5535 Table 2
Diffuse Radio Emission PropertiesHalo Relic S (mJy) 31.3 ± ± S . (mJy) ± ± P . (10 W Hz − ) 3.4 ± ± α ) -1.04 ± ± Note . — 1. Flux densities at 1.4 GHz are extrapolated assuminga power law. the individual exposures through principal componentanalysis (PCA; Jee et al. 2007) and stacked to obtainthe final PSF model for shape measurement. Readersare referred to the descriptions in our previous papersfor detail (e.g., Jee & Tyson 2011; Jee et al. 2013; Finneret al. 2017). After applying our S/N, color, magnitude,and shape measurement error cuts, we obtain a sourcedensity of ∼
11 galaxies per sq. arcmin.
X-ray
CL2023 was observed with the
Chandra
X-ray obser-vatory on 2014 March 30 (ObsId: 15108 - PI: Jones,ACIS-I detector, VFaint Mode, 20.81 ks). The data werereduced using the
CIAO 4 . software with CALDB 4 . . .We reprocessed the raw data using the chandra repro script to produce a level 2 event file. We then createda broad band (0.5-7 keV) exposure-corrected image withthe fluximage script.For our X-ray temperature measurement, point sourceswere masked out using the wavdetect script and back-ground flares were removed with the deflare script. Weextracted grouped X-ray spectra with the specextract script in such a way that each bin has a minimum signal-to-noise ratio of 5. Then, we performed spectral fit-ting with the XSPEC ( v12 . . ) package and used theabsorbed MEKAL plasma model (Kaastra & Mewe 1993;Liedahl et al. 1995) within the 1-5 keV energy band. TheGalactic hydrogen density and the cluster metal abun-dance were assumed to be N H = 5 . × cm − (Dickey& Lockman 1990) and 0.3 solar, respectively. RESULTS
Detection of a Radio Relic and Halo
Cross-matching the galaxy clusters detected in thePlanck SZ survey (Planck Collaboration XXVII 2016)with the ASKAP-EMU radio continuum survey, we dis-covered a ∼ . ∼ ∼ . § S = 16 . ± . α int = − . ± .
06. The extrapolated radio flux densityof the relic at 1.4 GHz is S . = 12 . ± . >
40 mJy) andthen replaced the fluxes with the in-halo average value.Within a polygon enclosing the halo, the flux densityis S = 31 . ± . α int = − . ± .
05 and S . =20 . ± . Cluster Galaxy and Weak-lensing MassDistribution
Detection of the radio relic suggests that CL2023 un-derwent a major merger. However, in order to recon-struct the merger scenario, we need to identify the clustersubstructures contributing to the merger. To this end,we use both the galaxy and mass distributions.Because no spectroscopic data of CL2023 are pub-licly available, we selected the cluster member candidatesbased on their 4000˚A break features. From the color-magnitude diagram, we chose a total of ∼ ∼ (cid:48) × (cid:48) region approxi-mately centered at the BCG. We adaptively smoothedthe galaxy number density using the csmooth tool witha minimum significance of 2.5 σ . The resulting iso-densitycontours are displayed in Figure 1C. The galaxy distri-bution suggests that CL2023 consists of three subclus-ters distributed in the east-west orientation. The centralcomponent is the most significant clump, coincident withthe X-ray and radio halo centroids. HyeongHan et al.
Figure 1.
Multi-wavelength observations of CL2023. (A) ASKAP-EMU full-resolution (12 (cid:48)(cid:48) × (cid:48)(cid:48) ) image. The green pan-shape regionsindicate the areas for the X-ray surface brightness analysis. The approximate location of the leading edge of the relic is marked with thered arc. (B) Composite radio image after both diffuse (purple) and compact sources (green) are enhanced ( § ∼ σ rms × n levels of the smoothed (FWHM=25 (cid:48)(cid:48) ) image where n = 0 , , , , σ rms = ∼ µ Jy/beam. The yellow arrow points at the “link”feature between the relic and nearby radio galaxy ( § Chandra
X-ray (magenta), and galaxy iso-density contours (white). We use the DECam g , r , and i filters to represent the intensities inblue, green, and red, respectively. A total of ∼ This three-component structure seen in the clustergalaxy distribution is in excellent agreement with ourWL mass distribution. We show the mass map obtainedwith the
FIATMAP (Fischer & Tyson 1997) code in Fig-ure 1D, and we verified that very similar mass distribu-tions are obtained with different algorithms such as the
MAXENT (Jee et al. 2007) or Fourier-inversion (Kaiser &Squire 1993) methods. Using bootstrapping analysis, weestimate that the central mass peak has the highest sig-nificance (5 . σ ), followed by the eastern (3 . σ ) and thenby the western (3 . σ ) peaks. By fitting three Navarro-Frenk-White (NFW; Navarro et al. 1997) profiles simul- taneously, we find that the eastern, central, and westernmasses are M c = 2 . ± . × M (cid:12) , 3 . ± . × M (cid:12) , and 1 . ± . × M (cid:12) , respectively (Table 3).Under the assumption that the three clumps are at thesame distance from us, the total mass of the system isestimated to be M c = 1 . ± . × M (cid:12) . Intracluster Gas Properties The total mass is greater than the sum of the three substruc-tures because r also increases. PT-CL 2023-5535 Figure 2.
Close-up view of the western radio galaxy mentioned in § σ rms × n levels of the full-resolution (12 (cid:48)(cid:48) × (cid:48)(cid:48) ) image where n = 0 , , ..., , Table 3
Mass Estimates of Substructures from WL analysisSubstructure M c Peak Significance (10 M (cid:12) ) ( σ )East 2.6 ± ± ± Note . — 1. We obtain masses by simultaneously fitting threeNFW profiles. 2. Mass peak significances are measured from the2 (cid:48) aperture by dividing the convergence by the rms map derivedfrom our bootstrapping analysis.
Using a circular ( r = 146 (cid:48)(cid:48) or 543 kpc) aperturecentered at the X-ray peak, we determined the X-raytemperature of CL2023 to be T X = 7 . ± .
79 keV( χ red = 0 . M c = 7 . +1 . − . × M (cid:12) . Our X-ray massestimate is consistent with the previous results. Tarr´ıo etal. (2018) quote M ∼ × M (cid:12) based on their jointanalysis of the Planck
SZ and
ROSAT
X-ray data. Therecent XMM-
Newton study (Bulbul et al. 2019) reports M = 6 . +0 . − . × M (cid:12) . These mass estimates basedon X-ray data roughly agree with our WL-based result M = 6 . ± . × M (cid:12) . However, given the clearindication of the on-going merger and invalidity of thesingle-halo assumption, we believe that the agreement israther a coincidence.Within the same r = 543 kpc aperture, the X-ray flux of CL2023 is f X , . − . = 21 . ± . × − erg cm − s − , which is converted to a luminosityof L X , . − . = 3 . ± . × erg s − . This Chan-dra luminosity is in good agreement with the
ROSAT result L X , . − . = 3 . ± . × erg s − re- ported by B¨ohringer et al. (2004). The relation be-tween the total radio luminosity of the halo P . =3 . ± . × W Hz − and the measured luminosityis consistent with the prediction from the L X − P . scaling relation of Feretti et al. (2012).Although radio relics are believed to be tracers ofmerger shocks, only a few clusters have shown to pos-sess corresponding shock features in X-ray. Our Chan-dra data analysis suggests that CL2023 may belong tothis rare class possessing a density jump across the relic.From the green “panda” regions depicted in Figure 1A,we determined the density compression C = 1 . ± . T X = 7 . ± . ±
12 keV, respectively.Given the current statistics, this temperature differenceis insignificant. DISCUSSION
Too Flat Spectral Index for a Relic?
If particles are advected downstream and do not sufferfrom too strong adiabatic losses and reacceleration pro-cesses, the integrated spectral index is steeper than theinjection spectral index by ∼ . α inj = − . ± .
06 of the CL2023 relic( § α inj < − . − . < α int < − .
0. Therefore, taken at face value,the α int value of CL2023 is unusual. However, one mustremember that we derived α int from the narrow band-width (800-1088 MHz), which has yet to be verified byobservations at other frequencies. Note that the spectralindex of the radio relic could be biased toward a steepervalue because the feature blends into the halo, which hasa relatively steep spectral index. To avoid any signif-icant contamination from the radio halo, we measuredthe flux density of the relic from a high resolution image(see Figure 2).In order to examine a consistency, we retrieved thearchival MOST data (843 MHz), which has a beam sizeof 45 (cid:48)(cid:48) × (cid:48)(cid:48) at the location of the relic. We degraded theresolution of our ASKAP-EMU image to match this res-olution and derive an aperture correction factor of 1.37for the same polygon aperture. With this aperture cor-rection, we obtained a flux density of 17 . ± . − . ± .
19, consistentwith the ASKAP-EMU measurement − . ± . α int = − . ± .
03 and − . ± . § HyeongHan et al.
Figure 3.
Monochromatic luminosity of relics at 1.4 GHz com-pared to their largest linear size (LLS) retrieved from van Weerenet al. (2019). Orange star indicates the relic discovered in CL2023. plasma and the role of the plasma aging would be in-significant, provided that the shock has just crossed thecloud and that the shock crossing time is much shorterthan the electron cooling time.As mentioned in § Merger Scenario
Both location and orientation of radio relics pro-vide constraints on merger scenarios. The currentCL2023 relic orientation suggests that the merger mightbe happening in the east-west direction, which impliesthat the relic is the result of the collision between themiddle and eastern subclusters. Under the assumptionthat the shock was generated at the impact and has beenpropagating to the west with nearly the same speed asthe collision speed in the plane of the sky, its location(detected by the relic) can be used as an indicator ofthe time-since-collision (TSC). Here, we present our esti-mates of the TSC in two ways. In one method, we inferthe collision velocity using the timing argument (Sarazin2002). In the other, we use our Mach number measure-ments.The timing argument (based on the assumption thatthe two clusters freefall to each other from an infiniteseparation) gives a relative velocity of ∼ ,
000 km s − atthe separation d ∼ ∼ . ∼ . C = 1 . ± . § M = 1 . ± . c s ,which is estimated to be c s ∼ ,
300 km s − from the pre-shock temperature ∼ ∼ ,
000 km s − , in good agreement with the value from Figure 4.
X-ray surface brightness profile across the relic. Blackcrosses are the data points (see Figure 1A for the selected regions)while the best-fit broken power-law model based on
PROFFIT v1 . (Eckert et al. 2011) is shown in blue. The red dashed line indicatesthe location of the relic boundary shown in Figure 1A. the timing argument.As mentioned in § α inj from the integrated spectral index α int under the stationary shock conditions. Assumingthat α int is a lower limit on α inj , we can convert α inj (cid:38) − .
76 to M (cid:38) .
9, which in turn corresponds to TSC (cid:46) . . − . CONCLUSIONS
From the deep high-resolution ASKAP-EMU pilot 300sq. deg survey, we discovered a ∼ . z = 0 .
23. We also confirmed the existence of the ∼ × . Chandra and DECam shows that 1) the radio halo coin-cides with the intracluster gas, 2) the cluster is composedof three subclusters, and 3) across the relic there is a hintof density jump in X-ray. Based on these results, we sug-gest that the cluster is a post-merger system, where themiddle and eastern subclusters might have suffered a ma-jor collision 0 . − . PT-CL 2023-5535 ∼
60 known radio relic systems to date, clearly one out-standing difficulty is the small sample size. Fortunately,a few giant radio surveys with state-of-the-art telescopes(e.g., SKA, LOFAR, etc.) are planned for the comingdecade. The ASKAP-EMU survey, as one important pro-gram, will greatly increase the sample size by at least twoorders of magnitude. The current study based on its pilot300 sq. deg data demonstrates its tremendous potentialwhen the full survey becomes available and supported byother multi-wavelength data.M. J. Jee acknowledges support for the current re-search from the National Research Foundation (NRF)of Korea under the programs 2017R1A2B2004644 and2020R1A4A2002885. Partial support for LR comes fromUS National Science Foundation grant AST 17-14205 tothe University of Minnesota. M. Yoon acknowledges sup-port from the National Research Foundation of Korea(NRF) grant funded by the Korea government (MSIT)under no.2019R1C1C1010942. M. Yoon acknowledgesupport from the Max Planck Society and the Alexandervon Humboldt Foundation in the framework of the MaxPlanck-Humboldt Research Award endowed by the Fed-eral Ministry of Education and Research. MJM acknowl-edges the support of the National Science Centre, Polandthrough the SONATA BIS grant 2018/30/E/ST9/00208.CJR acknowledges financial support from the ERC Start-ing Grant “DRANOEL”, number 714245. The Aus-tralian SKA Pathfinder is part of the Australia TelescopeNational Facility which is managed by CSIRO. Opera-tion of ASKAP is funded by the Australian Governmentwith support from the National Collaborative ResearchInfrastructure Strategy. Establishment of the Murchi-son Radio-astronomy Observatory was funded by theAustralian Government and the Government of West-ern Australia. ASKAP uses advanced supercomputingresources at the Pawsey Supercomputing Centre. We ac-knowledge the Wajarri Yamatji people as the traditionalowners of the Observatory site.REFERENCES
Andrade-Santos, F., van Weeren, R. J., Di Gennaro, G., et al.2019, ApJ, 887, 31Bell, A. R. 1978, MNRAS, 182, 147Blasi, P., & Colafrancesco, S. 1999, APh, 12, 169B¨ohringer, H., Schuecker, P., Guzzo, L., et al. 2004, A&A, 425,367Bonafede, A., Intema, H. T., Br¨uggen, M., et al. 2014, ApJ, 785, 1Botteon A., Brunetti G., Ryu D., Roh S., 2020, A&A, 634, A64Br¨uggen, M., Bykov, A., Ryu, D., et al. 2012, Space Sci. Rev.,166, 187Brunetti G., Setti G., Feretti L., Giovannini G., 2001, MNRAS,320, 365Brunetti, G., & Lazarian, A. 2007, MNRAS, 378, 245Brunetti, G., & Lazarian, A. 2011, MNRAS, 410, 127 Brunetti, G., & Jones, T. W. 2014, International Journal ofModern Physics D, 23, 1430007-98Bulbul, E., Chiu, I.-N., Mohr, J. J., et al. 2019, ApJ, 871, 50Dennison B., 1980, ApJ, 239, L93Dickey, J. M., & Lockman, F. J. 1990, ARA&A, 28, 215Drury, L. OC. 1983, Rept. Prog. Phys., 46, 973Eckert, D., Molendi, S., & Paltani, S. 2011, A&A, 526, A79En β lin, T. A., Biermann, P. L., Klein, U., & Kohle, S. 1998,A&A, 332, 395Ferrari, C., Govoni, F., Schindler, S., Bykov, A. M., & Rephaeli,Y. 2008, SSRv, 134, 93Feretti, L., Boehringer, H., Giovannini, G., & Neumann, D. 1997,A&A, 317, 432Feretti, L., Giovannini, G., Govoni, F., & Murgia, M. 2012,A&ARv, 20, 54Finner, K., Jee, M. J., Golovich, N., et al. 2017, ApJ, 851, 46Fischer, P., & Tyson, J. A. 1997, AJ, 114, 14Flaugher, B., Diehl, H. T., Honscheid, K., et al. 2015, AJ, 150,150Frater, R. H., Brooks, J. W., & Whiteoak, J. B. 1992, J.Electrical Electron. Eng. Australia, 12, 103Govoni, F., En ββ