Discovery of Candidate X-ray Jets in High-Redshift Quasars
Bradford Snios, Daniel A. Schwartz, Aneta Siemiginowska, Ma?gosia Sobolewska, Mark Birkinshaw, C. C. Cheung, Doug B. Gobeille, Herman L. Marshall, Giulia Migliori, John F. C. Wardle, Diana M. Worrall
aa r X i v : . [ a s t r o - ph . H E ] F e b Submitted to ApJ - Feb 24, 2021
Typeset using L A TEX twocolumn style in AASTeX62
Discovery of Candidate X-ray Jets in High-Redshift Quasars
Bradford Snios, Daniel A. Schwartz, Aneta Siemiginowska, Ma lgosia Sobolewska, Mark Birkinshaw, C. C. Cheung, Doug B. Gobeille, Herman L. Marshall, Giulia Migliori,
John F. C. Wardle, andDiana M. Worrall Center for Astrophysics | Harvard & Smithsonian, Cambridge, MA 02138, USA H. H. Wills Physics Laboratory, University of Bristol, Bristol BS8 1TL, UK Space Science Division, Naval Research Laboratory, Washington, DC 20375, USA Physics Department, University of Rhode Island, Kingston, RI 02881, USA Kavli Institute for Astrophysics and Space Research, Massachusetts Institute of Technology, Cambridge, MA 02139, USA Department of Physics and Astronomy, University of Bologna, Via Gobetti 93/2, 40129 Bologna, Italy INAF-Institute of Radio Astronomy, Bologna, Via Gobetti 101, I-40129 Bologna, Italy Physics Department, Brandeis University, Waltham, MA 02454, USA
ABSTRACTWe present Chandra X-ray observations of 14 radio-loud quasars at redshifts 3 < z <
4, selectedfrom a well-defined sample. All quasars are detected in the 0.5–7.0 keV energy band, and resolvedX-ray features are detected in five of the objects at distances between 1–12 ′′ from the quasar core.The X-ray features are spatially coincident with known radio features for four of the five quasars.This indicates that these systems contain X-ray jets. X-ray fluxes and luminosities are measured,and jet-to-core X-ray flux ratios are estimated. The flux ratios are consistent with those observed fornearby jet systems, suggesting that the observed X-ray emission mechanism is independent of redshift.For quasars with undetected jets, an upper limit on the average X-ray jet intensity is estimated usinga stacked image analysis. Emission spectra of the quasar cores are extracted and modeled to obtainbest-fit photon indices, and an Fe K emission line is detected from one quasar in our sample. Wecompare X-ray spectral properties with optical and radio emission in the context of both our sampleand other quasar surveys. Keywords: galaxies: active – galaxies: high-redshift – AGN: jets – X-rays: general INTRODUCTIONRelativistic jet outflows from Active Galactic Nuclei(AGN) can transport significant energy from the centralsuper-massive black hole to the surrounding interclus-ter medium (ICM; Scheuer 1974, 1982; Begelman et al.1984). Interactions between the jet and ICM will cre-ate radio lobes at kiloparsec-scale distances, making ra-dio observations well-suited for extragalactic jet stud-ies (e.g., Blandford & Rees 1974; Hargrave & Ryle 1974;Perley et al. 1984). Jets are responsible for feedbackprocesses that prevent central gas from cooling in clus-ters of galaxies (Bˆırzan et al. 2004; Rafferty et al. 2006;McNamara & Nulsen 2007), consequently reducing starformation at the cluster centers (Fabian 1994, 2012).Thus, interactions between jets and the surroundingmedium govern the overall evolution of these extragalac-tic systems.X-ray observations of extragalactic radio sourcesprovide insights into fundamental physical processes present within these systems (e.g., Heinz et al. 1998;Harris et al. 2006; Worrall et al. 2008; Stawarz & Petrosian2008). Although the origin of the X-ray jet emis-sion mechanism is not uniquely defined, one probablemechanism is that the X-rays are generated from In-verse Compton upscattering of the Cosmic MicrowaveBackground radiation (IC/CMB, Tavecchio et al. 2000;Celotti et al. 2001). Under such an assumption, X-rayobservations of jets may be used to measure the enthalpyflux, or “power”, transported by the jets to the radiolobes and the ICM (Scheuer 1974; Heinz et al. 1998;Reynolds et al. 2001). IC/CMB is predicted to be thedominant X-ray emission mechanism for high-redshiftradio jets as the cosmological diminution of surfacebrightness by the factor (1+ z ) − is offset by the (1+ z ) increase in the CMB energy density. IC/CMB is addi-tionally bolstered by the longer lifetimes of the 100 MeVelectrons, which generate such emission, relative to the10 GeV electrons required for radio synchrotron radia- Snios et al.
Table 1.
Chandra Observations of Quasar SampleObject z a RA b Decl. b ObsID c Observation t expd log( M BH ) e log( L bol ) e log( λ Edd ) f [J2000] [J2000] Date [ks] [log( M ⊙ )] [log(erg s − )]J0801+4725 3.256 08:01:37.682 +47:25:28.24 20405 2018 Jan 19 9.40 8 . ± .
47 46.97 0.34J0805+6144 3.033 08:05:18.180 +61:44:23.70 20399 2018 Jun 07 9.75 – – –J0833+0959 3.713 08:33:22.514 +09:59:41.14 20401 2019 Feb 04 9.57 9 . ± .
09 47.10 − . ± .
05 46.88 − . ± .
01 47.27 − . ± .
03 47.08 0.07J1128+2326 3.042 11:28:51.701 +23:26:17.35 20412 2019 Mar 10 9.57 9 . ± .
04 47.07 − . ± .
02 47.86 − . ± .
04 46.97 − . ± .
06 46.63 − . ± .
03 47.31 − . ± .
03 46.80 0.01J1655+1948 3.262 16:55:43.568 +19:48:47.12 20406 2018 Jun 17 9.57 9 . ± .
04 46.78 − a Redshift measurements from Sowards-Emmerd et al. (2005), Husband et al. (2015), and Pˆaris et al. (2018). b Coordinates fromVLA positions reported in Gobeille et al. (2014). c Observations performed using Chandra ACIS-S instrument with aimpointon S3 chip. d Total exposure time after flare removal reprocessing and dead time correction. e Measurements from Rakshit et al.(2020). f Eddington ratios for the quasar sample. tion. These emission properties make X-ray observa-tions well-suited for detecting high-redshift jets.Although X-rays are uniquely suitable for investi-gating high-redshift jets, the only telescope presentlycapable of resolving extended X-ray structures fromhigh-redshift radio sources is the Chandra X-rayObservatory. Previous Chandra observations havedemonstrated this capability by resolving jets in lu-minous radio sources at redshifts up to z = 4 . . < z < . H = 70 km s − Mpc − , Ω Λ = 0 .
7, and Ω M = 0 . SAMPLE SELECTION AND DATA REDUCTIONTargets for our analysis were selected from a cata-log of 123 radio-bright quasars at redshifts z > . ′′ or better resolution. We also pri-oritized sources where the separation distance betweenthe radio features exceeded 1 ′′ as such spatial separa-tions are resolvable with the Chandra X-ray Observa-tory, assuming comparable sizes for any X-ray counter-parts. Lastly, we eliminated radio objects identified astriples since their jets are less likely to be at a line ofsight that will achieve the relativistic beaming requiredfor detection in X-rays. Of the remaining 31 sources thatsatisfied these criteria, 14 objects at redshifts z > Figure 1.
Chandra 0.5–7.0 keV images of the fourteen targets binned in 0 . ′′
25 pixels. The X-ray images are overlaid with radiomap contours (green) from VLA observed at 6.2 GHz, except for J0833+0959 which is overlaid with a 1.4 GHz VLA radio map.The restoring beam for each radio map is shown as a black ellipse. resolved radio structure that had previously not beenobserved in X-rays were selected for our survey.Each target selected for our survey was observed withChandra using the Advanced CCD Imaging Spectrom-eter (ACIS) with the aimpoint centered on the S3 chip.The instrument was placed in the 1/4 subarray timedexposure mode with vfaint telemetry enabled in aneffort to mitigate pile-up. Roll direction for each ob-servation was defined such that the radio features didnot coincide with the readout direction of the chip.We also confirmed that the known Chandra PSF ar-tifact , which can create non-physical X-ray features ona sub-arcsecond scale, was not coincident with any radiofeature in our sample.All X-ray observations were analyzed using the level 2data products from the standard Chandra data process-ing pipeline together with the software analysis packageCIAOv4.12 with CALDBv4.9.2.1. Each observation wasreprocessed using the routine deflare to remove back-ground flaring periods from the data, and the average See the PSF artifact caveat in the CIAO User Guidehttps://cxc.harvard.edu/ciao/caveats/psf artifact.html cleaned exposure time per target is 9.53 ks. Pile-up wasestimated for each source using PIMMSv4.11a, and allsources had <
3% pile-up over the 0.5–7.0 keV energyband. Thus, pile-up was ignored for our X-ray analy-sis. Details on our Chandra observations are providedin Table 1.In addition to the X-ray data, radio observations of thesample were included in our analysis. We obtained newKarl G. Jansky Very Large Array (VLA) A-array obser-vations of these 14 quasars (see Table 2 for a summary)as part of a more extensive radio imaging and follow-upprogram of this sample (Gobeille et al. 2014). The datawere calibrated and imaged using standard proceduresin CASA and AIPS. The majority of the VLA maps(13/14) were obtained as single ∼ Snios et al.
Table 2.
VLA Observations of Quasar SampleObject Exposure Frequency Observation Beam Beam Beam Position Minimum Contour[s] [GHz] Date Maximum Minimum Angle Intensity[arcsec] [arcsec] [deg] [mJy beam − ]J0801+4725 378.9 6.2 2012 Nov 18 0.36 0.31 − − − − − − − − − The X-ray and radio observations were aligned basedon the centroid position of the core. Centroid positionwas measured in each observation using the dmstat rou-tine in CIAO, and the X-ray observation coordinateswere shifted with wcs update to agree with the radiodata. We also generated radio contours from the avail-able maps. The minimum contour level was defined as ∼ . ′′
25 pixels. Additional details onthe radio maps and contours used in our analysis areprovided in Table 2. X-RAY EMISSION SPECTRAThe unabsorbed X-ray fluxes and luminosities for thesample were determined by modeling the spectrum fromeach source. An emission spectrum from each observa-tion was extracted using the specextract routine inCIAO. We defined the source region as a 1 . ′′ ′′ and an outer ra-dius of 30 ′′ . We masked any point sources in the back-ground region, if present, to avoid count biases. Wenote that the average percent difference between the to-tal and background-corrected counts is 0.17% for ourextracted spectra, so the background contribution is mi-nor. Each spectrum was binned at 1 count per bin overthe 0.5–7.0 keV band. While performing our spectral analysis of the sample,we additionally tested for the presence of spectral lines,such as an Fe K line. We defined two different spectralmodels, one with an emission line and one without, andassessed both models with a likelihood ratio test for eachobservation. The line emission component was includedin our model in cases where its introduction improvedthe p -value of the likelihood ratio test to a value lessthan 0.001. The models are as follows: Model A : phabs · powerlaw . This is the default modelfor X-ray emission from the quasar core that was usedfor all 14 sources. Model B : phabs · (powerlaw + zgauss) . This modelincludes the primary X-ray emission and a Gaussianemission line, where we fixed the line width to 0.001 keV.This model was tried for all sources, but preferred onlyfor J1223+5038.Each extracted spectrum was fit over the 0.5–7.0 keVenergy band using WStat in CIAO’s modeling andfitting package Sherpa . Galactic hydrogen columndensity N H was fixed to extrapolated values fromDickey & Lockman (1990), while photon index Γ( dN/dE ∝ E − Γ ) and the normalization(s) were al-lowed to vary for all models. Once a best-fit model wasrealized, an additional intrinsic absorption component(i.e., zphabs ) was added to the model to determine ifthe intrinsic absorption column density N i H could beconstrained. The revised model was fit to the data with N i H , normalization, and Γ as free parameters. In allcases, only an upper limit on the intrinsic absorptioncould be established as its addition did not improve theoverall fit statistics. Model parameter best-fit results Table 3.
X-ray Properties of the Quasar SampleObject C obs Model N H N i H Γ E line I line f . − . L −
10 keV ℓ (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11)J0801+4725 48 A 0.0454 < . +0 . − . – – 5.8 +1 . − . +1 . − . +3 . − . J0805+6144 475 A 0.0452 < . +0 . − . – – 66.1 +3 . − . +2 . − . +2 . − . J0833+0959 173 A 0.0394 < . +0 . − . – – 23.3 +2 . − . +3 . − . +5 . − . J0909+0354 797 A 0.0347 < . +0 . − . – – 116.4 +4 . − . +3 . − . +3 . − . J0933+2845 80 A 0.0191 < . +0 . − . – – 10.4 +1 . − . +1 . − . +2 . − . J1016+2037 196 A 0.0242 < . +0 . − . – – 25.6 +2 . − . +2 . − . +3 . − . J1128+2326 91 A 0.0134 < . +0 . − . – – 12.7 +1 . − . +1 . − . +2 . − . J1223+5038 464 A 0.0169 < . +0 . − . – – 62.4 +3 . − . +4 . − . +6 . − . ... ... B ... < . +0 . − . . +0 . − . . +1 . − . +3 . − . +4 . − . +7 . − . J1405+0415 280 A 0.0217 < . +0 . − . – – 37.3 +2 . − . +2 . − . +3 . − . J1435+5435 38 A 0.0127 < . +0 . − . – – 5.3 +1 . − . +1 . − . +2 . − . J1610+1811 397 A 0.0362 < . +0 . − . – – 53.4 +3 . − . +2 . − . +4 . − . J1616+0459 386 A 0.0475 < . +0 . − . – – 51.1 +3 . − . +2 . − . +4 . − . J1655+3242 41 A 0.0222 < . +0 . − . – – 5.5 +1 . − . +1 . − . +1 . − . J1655+1948 47 A 0.0549 < . +0 . − . – – 5.8 +1 . − . +1 . − . +2 . − . (1) Object name. (2) Observed counts over the 0.5–7.0 keV band from 1 . ′′ cm − . (5) Intrinsic column density3 σ upper limits, in units of 10 cm − . (6) Photon index estimated from 0.5–7.0 keV best-fit spectral model. (7) Rest-frame ironemission line energy, in unit of keV. (8) Observed line intensity, in units of 10 − photons cm − s − . (9) Observed 0.5–7.0 keV flux,in units of 10 − erg cm − s − . (10) Rest-frame 2–10 keV luminosity, in units of 10 erg s − . (11) Rest-frame, monochromaticluminosity at 2 keV, in units of 10 erg s − Hz − . from the spectral analysis and their respective 1 σ con-fidence intervals are provided in Table 3. We note thatthe upper limits on N i H are reported at 3 σ confidence.We found from our spectral analysis that all 14 sourcesin the sample could be individually fit with Model A,providing constraints on both N i H and Γ. Addition-ally, we determined from our likelihood ratio tests thatJ1223+5038 was best fit with Model B due to the pres-ence of an emission line in its spectrum. The measuredemission line is consistent with neutral Fe emission at6.4 keV, while the N i H and Γ best-fit parameters agreewith the Model A results. The observed spectrum andmodel fit for J1223+5038 are shown in Figure 2, and thebest-fit results are in Table 3.Once the best-fit spectral models were obtained, wemeasured the observed Galactic absorption-corrected0.5–7.0 keV flux f . − . for each source. Therest-frame 2–10 keV luminosity L −
10 keV and monochro-matic 2 keV luminosity ℓ were additionally deter-mined from each best-fit model. We applied aperturecorrection for all measured fluxes and luminosities,where the correction factor was derived from our en-circled counts fraction (ECF) analysis of the sources(see Section 4). The average correction factor for oursample is 1.075, and the range is 1.056–1.098. All mea-sured X-ray fluxes and luminosities for the sample areshown in Table 3. X-RAY MORPHOLOGYAs discussed in Section 2, our sample of high-redshiftquasars have known radio features at distances of 1 ′′ orgreater from the quasar core. Existing X-ray counter-parts to these radio features may be resolvable with ourChandra observations, assuming that both a satisfactorysignal-to-noise ( S/N ) is achieved and the radio/X-rayemitting regions are comparable in size. Thus, we inves-tigated our sample for evidence of resolved X-ray struc-tures.The asymmetric point-spread function (PSF) forChandra must be considered when investigating fea-tures on scales of ∼ ′′ , which we suspected for oursources. As there is presently no analytic PSF forChandra, we generated synthetic PSF images for ourobservations. We began by simulating 500 ray-tracingfiles for each X-ray observation using ChaRT v2, a webinterface to the SAOsac ray-trace code. The ray-tracingfiles and various physical parameters of the observation,such as exposure time and detector orientation, were in-put into the simulate psf script in CIAO to accuratelygenerate simulated PSF images. All ray-tracing fileswere projected using MARX v5.5.0, and the extracted See ‘Understanding the Chandra PSF’ thread in the CIAOguide: https://cxc.cfa.harvard.edu/ciao/PSFs/psf central.html
Snios et al. C oun t s / s e c / k e V J1223+5038 R e s i dua l s Figure 2.
Chandra X-ray spectrum of J1223+5038. Thespectrum is binned by 5 counts per bin (solely for illustrativepurposes) and is fitted over the 0.5–7.0 keV energy rangewith an absorbed power-law model (Γ = 1 . +0 . − . ) and anFe K emission line at the rest energy of 6.28 keV, or observedenergy of 1.39 keV. The lower panel shows the residuals fromthe best-fit model. spectra from Section 3 were used to reproduce the spec-tral response of our observations. Since the default valuefor the aspect blur parameter in simulate psf has pre-viously been shown to generate a simulated PSF profilenarrower than what is observed , we generated multiplesimulated PSF with different aspect blur values. TheECF was calculated for each simulation and comparedwith the observed data. See Figure 3 for an example ofthis comparison for J1223+5038. From our analysis, wefound an aspect blur value of 0 . ′′
28 to best fit all sourcesin our survey.After verifying that all PSF simulations were co-alignedwith their respective Chandra observations to subpixelaccuracy, we defined an annulus region for each quasar.The outer annular radius was set such that it encom-passed the observed radio features plus an additional 1 ′′ to account for Chandra PSF blurring. Since the radiomap for J0833+0959 has a poorer resolution than theremainder of our sample, we defined its outer annularradius as 1 ′′ from the centroid of the external radiofeature. The inner annular radius was set equal to the95% ECF radius of the core, as measured from our sim-ulated ECF profiles, in cases where the core separationdistance for the nearest radio feature is > ′′ . In cases See ‘The AspectBlur Parameter inMARX’ thread in the CIAO guide:https://cxc.cfa.harvard.edu/ciao/why/aspectblur.html
Radius from Core (ACIS Pixels) E n c i r c l ed C oun t s F r a c t i on ( E C F ) J1223+ 50380. (cid:1)(cid:0)
07 blur0. (cid:2)(cid:3)
14 blur0. (cid:4)(cid:5)
21 blur0. (cid:6)(cid:7)
28 blur0. (cid:8)(cid:9)
35 blur
Figure 3.
A comparison of the background-subtracted en-circled counts fraction (ECF) for the quasar J1223+5038 ver-sus simulated PSFs generated from the simulate psf routinein CIAO. For each simulation, a different aspect blur param-eter was used. We found that a blur parameter of 0 . ′′
28 isbest for reproducing the sources in our sample. where the core separation distance is ≤ ′′ , the innerradius was set equal to 85% ECF radius. The resultinginner radii for the sample ranged between 1 . ′′ . ′′
7, andthe outer radii were between 2 . ′′ . ′′
0. Each annuluswas additionally divided into twelve, 30 ◦ sectors, wherewe defined our coordinate system as 0 ◦ West rotatingcounterclockwise. The sectors used for J1610+1811 areshown in Figure 4.Total counts for each annular sector of the observa-tions and simulations were measured using dmextract in CIAO. Counts were also measured for a circular re-gion 1 ′′ in radius surrounding the core in each X-rayimage, and the simulated leakage counts into each an-nular sector were scaled based on the ratio of theobserved-to-simulated core counts. The average ob-served background counts per sector were estimated us-ing the background regions discussed in Section 3.4.1. Statistical Assessment of X-ray Features
Having measured the counts per annular sector forboth observations and simulations, we examined whichsectors, if any, have elevated observed counts relative tothe simulated PSF as this would indicate the presence ofextended X-ray emission. To ensure a rigorous detectioncriterion, we derived a counts probability for each sectorbased on the simulated and field background emissionsin order to find sectors that are statistically significantoutliers.To begin, we inferred that the simulated count prob-ability distribution of each sector in our analysis couldbe accurately modeled with Poisson statistics. Thus,the simulated counts per sector may be represented asa Poisson distribution where the average is calculatedfrom the 500 simulations. We note that the simulations
J1610+1811 . . . . . : : . . RA (J2000) D e c ( J ) Figure 4.
The annular sectors defined in Section 4 for oneof the fourteen sources in our quasar sample. Each annu-lus was divided into twelve 30 ◦ sectors, beginning West androtating counterclockwise. In the source shown, the innerand outer radii correspond to 1 . ′′ ′′ , respectively. Forconvenience, sectors 1 and 12 have been labeled. do not include background emission, which must be ac-counted for when comparing to the observed counts.The background counts may similarly be represented asanother Poisson distribution where the mean value cor-responds to the predicted background counts per sector.Since the sum of independent Poisson random variablesis itself Poisson (Grimmett & Welsh 1986), we may de-fine the sum of simulated and background distributionsas P( λ + λ ), where λ is the mean simulated countsper sector and λ is the expected background counts persector. We may therefore accurately determine the prob-ability for the total observed counts per annular sectorfor each source.Using the predicted probability distributions, we cal-culated the cumulative probability of obtaining the ob-served count rate or higher for each sector of each source.We defined a probability p ≤ .
005 as the detectionthreshold for the sample. Based on our defined signifi-cance threshold, we found five sectors from five uniquesources that have significant evidence of X-ray emission.The X-ray features and their respective probabilities areprovided in Table 4, and images of the five sources withtheir corresponding emission sectors are shown in Fig-ure 5. A machine-readable table of the complete sectoranalysis for all sources is available as an accompanyingonline resource to this article.
Table 4.
X-ray Morphological Analysis
Object r in r out C obs C PSF C bg p (1) (2) (3) (4) (5) (6) (7)J0833+0959 1.5 11.0 8 0.84 0.60 < < (1) Object name. (2) Inner annular radius, in units of arcsec.(3) Outer annular radius, in units of arcsec. (4) Observed0.5–7.0 keV counts. (5) Mean 0.5–7.0 keV counts from simu-lated PSF. (6) Expected background 0.5–7.0 keV counts. (7)Cumulative Poisson probability of detecting counts ≥ C obs .The results listed in the table are those sectors where theprobability was ≤ RESOLVED X-RAY FEATURES5.1.
Flux and Surface Brightness
The flux and the surface brightness of each X-ray fea-ture detected in Section 4.1 may be measured, givensome assumptions on the X-ray emission. Due to thelimited X-ray counts in our observations, identificationof edges for the X-ray extension could not be performedusing either surface brightness profiles or contour map-ping, as is standard in jet and hotspot analyses. Weinstead assumed that the X-rays were generated fromjets that extended from the quasar core. We addition-ally assumed that the annular sectors in Figure 4 weresufficient in size that they encompassed the extendedX-ray emission. The annular sectors were verified to re-produce the same measured count rates as other regionshapes, such as rectangular and elliptical regions (e.g.,Schwartz et al. 2020). Furthermore, we found the annu-lar sectors to minimize the overall systematic error in-troduced from the PSF counts, with a 20% reduction in C PSF relative to similarly-sized rectangles and ellipses,while also improving the overall reproducibility of ourflux measurements. Thus, we used the count rates mea-sured with the annular sectors (Table 4) for our fluxanalysis.The source counts of each region were calculatedby subtracting the mean counts from the simulatedPSF and the expected background counts from thetotal observed counts. We then modeled each regionwith a phabs · powerlaw expression in PIMMSv4.11a,using the latest Chandra response files, and set thespectral normalization equal to our X-ray jet countrates. The Galactic column density was fixed equalto values from Dickey & Lockman (1990). We alsofixed the photon index to Γ = 1 .
9, which we in-
Snios et al. J0909+0354 . . . : : . . RA (J2000) D e c ( J ) J1016+2037 : : . . . . RA (J2000) D e c ( J ) J1610+1811 . . . . . : : . . RA (J2000) D e c ( J ) J1405+0415 . . . : : . . . . RA (J2000) D e c ( J ) J0833+0959 . . : : . . . RA (J2000) D e c ( J ) Figure 5.
Chandra 0.5–7.0 keV images of the five sources where resolved X-ray structure is detected from our statisticalanalysis. Each image is binned in 0 . ′′
25 pixels and overlaid with radio map contours (green). The regions shown are the sectorswhere X-ray features were detected, and the restoring beam for each radio map is shown as a black ellipse. ferred to be a standard slope for X-ray jet and knotspectra (e.g., Marshall et al. 2002; Siemiginowska et al.2002; Schwartz et al. 2006a,b; Goodger et al. 2010;Zhang et al. 2018). The observed 0.5–7.0 keV X-ray jetfluxes are shown in Table 5. We note that the flux valuesfor J1405+0415 and J1610+1811 agree with previousmeasurements from Schwartz et al. (2020). Rest-frame2–10 keV luminosities were also measured for each sourceand are provided in Table 5. We found the fluxes andluminosities to be similar for the five jets.We additionally calculated the surface brightness val-ues for each source. We assumed the emitting area origi-nated at the quasar core and extends to the outer radiusof our defined regions. The width of the emitting regionwas set equal to the full width-half maximum (FWHM)of Chandra, which is 0 . ′′
5. See Table 5 for the flux, lu-minosity, and surface brightness measurements of theX-ray jets. Given the limited count statistics for theseextended X-ray features, the measurements in Table 5have an approximate factor of 2 error. 5.2.
IC/CMB Emission
The inverse Compton upscattering of the CosmicMicrowave Background radiation (IC/CMB) in a jetwith bulk relativistic motion is a known X-ray emis-sion mechanism in quasar jets, where the broad-bandemission is attributed to a single spectrum of relativis-tic electrons (Tavecchio et al. 2000; Celotti et al. 2001;Siemiginowska et al. 2002; Schwartz 2002; Schwartz et al.2006a,b; Worrall 2009; Marshall et al. 2011, 2018;Worrall et al. 2020). However, recent studies havedemonstrated that IC/CMB from a single electron pop-ulation does not reproduce the observed spectral prop-erties of several low-redshift jets. These results include3C 273 ( z = 0 . z = 1 .
1, Siemiginowska et al. 2007), and PKS 0637-752( z = 0 . γ -ray production in jets measuredfrom Fermi observations imply that a single electronpopulation is insufficient to produce the observed X-rayemission in several jets at z < z <
3) sys-tems, IC/CMB is theorized to be the predominantemission mechanism at z & f jet relativeto the quasar core flux f core . Thus, we examined thejet-to-core flux ratios from our sample for evidence ofredshift dependence in the context of IC/CMB jet emis-sion.We estimated the jet-to-core flux ratios f jet /f core forour sample using the observed 0.5–7.0 keV fluxes in Ta-bles 3 and 5, and the ratio results are shown in Ta-ble 5. The measured flux ratios are between 1.0–3.6%,with a mean value of 2.2%. These results are in excel-lent agreement with the median 2% f jet / f core found byMarshall et al. (2018) for quasars at z <
2. The X-rayflux densities we derived consequently do not follow theexpected (1 + z ) dependence for IC/CMB, which isconsistent with the conclusion of Marshall et al. (2011).Our sample also agrees with the f jet /f core results fromWorrall et al. (2020) that were based on the few knownX-ray jets at z > . S/N .Although we did not observe an elevation in X-ray jetflux as a function of redshift, we stress that this resultis indicative. It is possible that the quasar core is alsodominated by beaming (Worrall et al. 1987), biasing ourX-ray ratios. Additionally, the shallow line-of-sight ex-pected for our quasars will cause the innermost regionof the jet to appear as part of the core due to the limitedspatial resolution of Chandra. This innermost region ofthe jet has been shown to produce the brightest featuresin some nearby sources (e.g., Snios et al. 2019a,b), whichmay bias our findings. Our comparison also requiresthat the high- and low-redshift quasar cores be phys-ically consistent with one another, which may not bevalid. Deeper Chandra observations are necessary to in-vestigate the spectrum and spatial structure of these sys-tems. For now, we can only reiterate that the flux den-
Table 5.
X-ray Jet PropertiesObject d jet C src f jet S jet L jet f jet / f core (1) (2) (3) (4) (5) (6) (7)J0833+0959 11.0 6.56 8.3 1.5 6.2 0.036J0909+0354 3.5 9.51 12.1 6.9 6.7 0.010J1016+2037 3.0 6.29 8.0 5.4 3.9 0.031J1405+0415 2.5 5.78 7.4 5.9 3.9 0.020J1610+1811 6.0 5.11 6.8 2.3 3.3 0.013 (1) Object name. (2) Projected jet length, in units of arcsec. (3)Source 0.5–7.0 keV counts from jet. (4) Observed 0.5–7.0 keV jetflux, in units of 10 − erg cm − s − . (5) Observed 0.5–7.0 keVjet surface brightness, in units of 10 − erg cm − s − arcsec − (6)Rest-frame 2–10 keV jet luminosity, in units of 10 erg s − . (7)Ratio of X-ray fluxes for jet to core. sity jet-to-core ratio derived from our high-redshift sam-ple is consistent with measurements from low-redshiftquasars, suggesting that it is independent of redshift.5.3. Flux Limits for Undetected X-ray Jets
Despite X-ray features were detected from five sourcesin our quasar sample, the remaining nine sources showno evidence of resolved X-ray structure(s). However,X-ray jets should be co-spatial with the radio featuresobserved at GHz frequencies, so it is possible thatX-ray emission is present but resides below our detec-tion threshold. We therefore stacked the nine sourceswith non-detections in an effort to measure an averageintensity of their X-ray jets.Using the annular sectors from Section 4.1, we de-termined for each source which X-ray sector(s) wasco-aligned with a radio feature. In total, we defined11 sectors from the nine sources that were co-alignedwith radio features. Counts from the 11 sectors wereco-added, giving us values for the observed, simulatedPSF, and background counts. In total, we measured 16observed counts and expect 14.37 simulated PSF countsplus 0.87 background counts. Based on our detectioncriterion from Section 4.1, this gives a false detectionprobability of 0.456. As this does not satisfy our de-tection threshold of p ≤ Snios et al. ≥
27 counts to 95% probability. In total, we estimateda required net of 20.84 counts over the background, ora rate of 1 . × − cts s − .Having determined an upper limit count rate, we mod-eled the jet with a phabs · powerlaw expression. TheGalactic column density was fixed to the average value ofthe nine sources, giving us N H = 3 . × cm − . Con-sistent with Section 5.1, we fixed the jet photon index toΓ = 1 .
9. Using PIMMSv4.11a with our count rate limitof 1 . × − cts s − , we found an upper limit for theaverage unabsorbed jet flux of 2 . × − erg cm − s − over the observed 0.5–7.0 keV band. This estimated fluxis consistent with typical background limits for ChandraACIS observations.With the X-ray jet flux limit now available, we testedfor evidence of IC/CMB emission by calculating the av-erage X-ray jet-to-core ratio limit for the nine quasars inour sample with no detected X-ray jets. We measuredthe jet-to-core flux ratio limit for each quasar and thenaveraged the values, which we found to be 2.4%. Theseresults are consistent with the mean 2.2% flux ratio ofthe five quasars with X-ray features. The flux ratio isalso in agreement with the distribution of jet-to-core ra-tios found by Marshall et al. (2018). Overall, these re-sults reinforce the finding of Section 5.2 that there is noobserved redshift dependence in the X-ray jet emission.5.4. Non-Coincident Radio and X-ray Features inJ1016 + Examination of our quasar sample shows that the ma-jority of X-ray features, when detected, are spatiallycoincident with radio features. However, J1016+2037stands out amongst our sample due to its non-coincidentX-ray and radio features, where the measured offset be-tween the X-ray emission and radio feature centroid is195 ◦ . This offset is despite excellent alignment of theX-rays and radio from its quasar core. Furthermore,the detection probability for the extended X-rays fromJ1016 is amongst the highest rated for our sample (Ta-ble 4), suggesting that this feature is indeed real. Thus,the origin of the X-ray feature and its misalignment withthe radio merits further discussion.We began by verifying that the known Chandra PSFartifact (see Section 2) was not aligned with the observedX-ray feature in J1016, confirming that the extendedX-rays are not due to a systematic effect from Chandra.We then investigated the possibility of an unassociatedX-ray source being the origin of the extended X-raysobserved in J1016. Our measured X-ray flux for theextended source is ∼ − erg cm − s − (Table 5), andthere are ∼
100 X-ray sources deg − above such a fluxlimit (Civano et al. 2016). As a result, there is only a ∼
1% probability of a chance association within 5 ′′ ofany quasar in our sample. We also confirmed that theX-ray feature is not coincident with any optical sourcesin Pan-STARRS images to a brightness limit of 21 magat r-band (Tonry et al. 2012). It is therefore likely thatthe extended X-rays are associated with J1016.Given that the offset between the radio and X-ray fea-tures is ∼ ◦ , it is possible that the different emissionsmay originate from jet/counter-jet features. One-sidedradio emission from quasars is normally interpreted asjet emission, meaning that the X-rays would be from thecounter-jet side. However, this scenario is unlikely as thebeamed jet will generally have a higher X-ray flux thanthe counter-jet region, and so X-rays should be detectedcoincident with the observed radio feature. In addition,the measured jet-to-core X-ray flux ratio for J1016 (Ta-ble 5) is consistent with flux ratios from a beamed jet,suggesting that the observed extended X-rays are due toemission from the near jet. It is therefore unclear fromthe multiwavelength data what is the orientation of thejetted outflow in J1016, assuming jets are present.Despite concluding that the extended X-rays in J1016are associated with the system, we lacked the X-raycount statistics required to determine the physical originof the observed misalignment. Follow-up, deep-exposureobservations with Chandra will permit a spectroscopicand morphological analysis of the X-ray structure wheredifferent emission models may be investigated to deter-mine the physical origin for this irregular emission fea-ture. OPTICAL PROPERTIES OF QUASAR SAMPLEPrevious quasar studies have demonstrated an in-verse relationship between the optical-to-X-ray flux ratioand optical luminosity (e.g., Avni & Tananbaum 1982;Tananbaum et al. 1986; Wilkes et al. 1994). We there-fore compiled optical properties of our sample in an ef-fort to investigate the optical and X-ray relationshipamongst our sources.In keeping with our previous quasar analysis (Snios et al.2020), we obtained optical fluxes at the rest-frame wave-length of 1450 ˚A. For our current sample, this corre-sponds to observed wavelengths between 5850–7000 ˚A.We consequently utilized the monochromatic AB ap-parent magnitude from the Pan-STARRS r-band asit has an effective wavelength of 6241 ˚A (Tonry et al.2012). All sources were identified in the Pan-STARRScatalog, and no Galactic extinction corrections wereapplied to our measurements. We additionally extrap-olated the rest-frame m results to determine therest-frame 2500 ˚A luminosity ℓ . A UV spectralindex of α = − . Table 6.
Optical Properties of the Quasar SampleObject m ℓ α ox (1) (2) (3) (4)J0801+4725 19.58 6.7 − . +0 . − . J0805+6144 19.92 4.1 − . +0 . − . J0833+0959 21.09 2.3 − . +0 . − . J0909+0354 19.94 4.9 − . +0 . − . J0933+2845 17.92 34.5 − . +0 . − . J1016+2037 19.12 9.1 − . +0 . − . J1128+2326 18.56 14.5 − . +0 . − . J1223+5038 17.47 54.7 − . +0 . − . J1405+0415 19.97 4.4 − . +0 . − . J1435+5435 20.19 5.5 − . +0 . − . J1610+1811 18.33 18.9 − . +0 . − . J1616+0459 19.19 9.1 − . +0 . − . J1655+3242 19.58 6.3 − . +0 . − . J1655+1948 20.02 4.4 − . +0 . − . (1) Object name. (2) Rest-frame, monochromatic AB appar-ent magnitude at 1450 ˚A, as measured from Pan-STARRSr-band. (3) Rest-frame, monochromatic luminosity at2500 ˚A, in units of 10 erg s − Hz − . (4) Optical-to-X-raypower-law slope. ( f ν ∝ ν α ), which is consistent with prior quasar studies(i.e. Shemmer et al. 2006; Nanni et al. 2017; Snios et al.2020). The optical fluxes and luminosities are shown inTable 6.Having determined both the optical and X-ray lumi-nosities for the sample, we calculated the optical-to-X-raypower-law index α ox for each source. We defined α ox the same as Tananbaum et al. (1979):log( ℓ /ℓ )log( ν /ν ) = 0 . · log( ℓ /ℓ ) , (1)where ℓ and ℓ are the monochromatic lumi-nosities at 2 keV and 2500 ˚A, respectively. For calculat-ing the error on α ox , the error on the X-ray luminositydensities were taken as dominated by the Poisson erroron the number of counts and the optical measurementswere assumed to have uncertainties of 0.1 mag. Resultsfor α ox are provided in Table 6.6.1. Optical and X-ray Relationship
Prior studies of quasars have shown an anticorre-lation between α ox and ℓ that is independentof redshift (Bechtold et al. 1994; Vignali et al. 2003;Steffen et al. 2006; Kelly et al. 2007; Nanni et al. 2017;Snios et al. 2020), so we investigated if our radio-loudquasar sample was consistent with this relationship.In keeping with the method from Snios et al. (2020),we selected quasars from the literature with known Figure 6.
Optical-to-X-ray power-law slope α ox versusUV luminosity ℓ . Results from this work (black) arecompared against other quasar samples from the litera-ture. The black dotted line is the best-fit model, whilethe grey region is the 3 σ confidence level. Previous mea-surements are taken from Shemmer et al. (2006); Just et al.(2007); Lusso & Risaliti (2016); Siemiginowska et al. (2016);Nanni et al. (2017); Martocchia et al. (2017); Zhu et al.(2019); Vito et al. (2019); Snios et al. (2020). optical and X-ray properties. We included 16 tar-gets from Shemmer et al. (2006), 34 from Just et al.(2007), 2153 from Lusso & Risaliti (2016), 18 fromNanni et al. (2017), 35 from Martocchia et al. (2017),15 from Zhu et al. (2019), 7 from Vito et al. (2019),and 15 from Snios et al. (2020). In total, we compileda sample of 2307 quasars with known α ox and ℓ parameters for our analysis.The resulting α ox - ℓ relationship from the com-plete dataset is shown in Figure 6. Using a linear re-gression with the scipy python package (Virtanen et al.2020), we found a best-fit relation of α ox = ( − . ± . ℓ ) + (3 . ± . , (2)for the total sample, where the errors are reported to1 σ . Our best fit is consistent with previous studies towithin 1 σ (Nanni et al. 2017; Snios et al. 2020). Addi-tionally, we verified that the best-fit relationship is in-dependent of redshift, which is consistent with previousworks (i.e., Just et al. 2007; Snios et al. 2020). We re-peated the analysis using only the 14 quasars from thiswork, finding a slope of − . ± . ±
4, where the best fit was determined using an or-thogonal distance regression in scipy . Thus, the quasarsample discussed in this paper is consistent within 2 σ tothe overall trend for the broader quasar population.6.2. Impact of Radio-Loudness on Optical Properties
Radio-loud quasars generally have a higher X-ray lu-minosity than radio-quiet quasars for a given optical2
Snios et al.
Table 7.
Radio Properties of the SampleObject f ℓ log( ℓ / ℓ )(1) (2) (3) (4)J0801+4725 89 6.7 6 . ± . ∗ . ± . . ± . ∗ . ± . ∗ . ± . . ± . . ± . . ± . . ± . . ± . . ± . ∗ . ± . . ± . ∗ . ± . ∗ ), flux densities were measured from 1.4 GHz VLA obser-vations. (3) Monochromatic luminosity at 5 GHz, in units of10 erg s − Hz − . (4) Rest-frame radio-to-X-ray ratio. Fluxdensities and luminosities are assumed to have a 15% uncer-tainty. Radio-to-X-ray ratio errors include both radio andX-ray uncertainties. luminosity (e.g., Worrall et al. 1987; Wu et al. 2013;Zhu et al. 2019). Given that our α ox - ℓ analy-sis includes both radio-loud and radio-quiet quasars,we investigated the impact of radio-loudness on the α ox - ℓ relationship. Our sample of 2307 quasarswas separated into radio-loud and radio-quiet popula-tions, where the radio-loud dataset includes all sourcesfrom Zhu et al. (2019), Snios et al. (2020), and thiswork. This gave us a total 44 radio-loud sources and2263 radio-quiet sources for our analysis.Using the radio-loud dataset, we found a linear re-lationship between α ox - ℓ with a best-fit slope of − . ± .
06 and an intercept of 9 . ± .
8. Repeat-ing our analysis with the radio-quiet sample, we founda slope of − . ± .
005 and an intercept of 3 . ± .
1. The radio-loud sample therefore diverges fromthe radio-quiet sample by ∼ σ . Since we previouslyverified that our best fit was independent of redshift,the observed discrepancy between the radio-loud andradio-quiet sample is not due to differences in redshiftselection criteria.Despite the observed impact of radio-loudness on the α ox - ℓ relationship, we stress that this result maybe biased and/or incomplete due to our small sample size )43.043.544.044.545.0 l o g ( L G H z ) Best-fitThis WorkSnios et al. 2020
Figure 7.
Comparison of rest-frame radio ( L ) andX-ray ( L −
10 keV ) luminosities. The black dotted line isthe best-fit relation, while the grey region is the 1 σ confi-dence level.Radio-to-X-ray ratios are broadly consistent forour sample. of radio-loud quasars. The available radio-loud quasarsample lacks the comprehensive coverage of X-ray andUV luminosities needed to accurately fit a relationshipfor a broad range of sources. Further study with a largerradio-loud sample is required in order to verify the de-pendence of radio-loudness on α ox properties. RADIO PROPERTIES OF QUASAR SAMPLEGiven that our quasar sample is comprised of sourcesdetected in both X-rays and radio, we investigated therelation between their radio and X-ray luminosities. Ra-dio flux densities for quasars discussed in this work wereprovided from the VLA-FIRST survey (White et al.1997), which observed at 1.4 GHz. In cases where noFIRST detection was found, flux densities were di-rectly measured from archival 1.4 GHz VLA observa-tions. For our sample, these radio measurements cor-respond to rest-frame frequencies between 5.6–6.7 GHz,which we assumed to be a reasonable approximation ofthe rest-frame 5 GHz flux density. Radio flux densitiesand luminosities for the sample are provided in Table 7.We note that the reported radio luminosities representthe core emission for each source.The quasars from this sample were added togetherwith the radio-loud sources from Snios et al. (2020), giv-ing us 29 radio-loud quasars at z > scipy ,and we found a best-fit relation oflog( L ) = (0 . ± . L −
10 keV )+ (10 . ± . , (3)3where the reported errors are 1 σ . We note that outliersfrom Snios et al. (2020) that reside above the best fit areknown to be either Compton-thick and/or possess unre-solved X-ray structure and are likely not representativeof the remaining sample. In comparison, all sources fromour current work agree with the measured relationshipwithin 3 σ . Our measured radio-X-ray luminosity corre-lation is consistent with the relationship for radio-loudquasars at z < CONCLUSIONSWe analyzed Chandra observations of 14 radio-loudquasars at redshifts 3 < z <
4, each with radio featuresadditional to core emission, to measure their X-ray spec-tral properties and search for evidence of resolved struc-ture. We detected all quasars in the 0.5–7.0 keV bandand extracted emission spectra of the quasar cores. Eachspectrum was fit with an absorbed power-law model,where our mean best-fit photon index is 1 . ± .
2. Ob-served X-ray fluxes and rest-frame luminosities werealso determined from the spectral best-fits. Addition-ally, we detected an Fe K emission line from the quasarJ1223+5038 at high significance.We performed a morphological analysis of each X-raysource using Chandra observations, and we detectedX-ray features at distances up to 12 ′′ from the quasarcore in five of the sources. The X-ray features are spa-tially coincident with existing radio features for fourof the five sources, suggesting that the majority of theX-ray features are jets. J1016+2037 stands out amongstour sample due to a ∼ ◦ misalignment between itsX-ray and radio features. We speculated on the causethis observed misalignment, but the available X-raycounts were insufficient to conclusively determine its ori-gin. Rest-frame 2–10 keV luminosities of the X-ray jetswere estimated for the five quasars with extendedX-rays, and their X-ray jet-to-core flux ratios were mea-sured to be up to 3.6%. We also estimated an upperlimit on the average X-ray jet flux for the remaining ninequasars from a stacked image analysis, finding a limitof 2 . × − erg cm − s − . This flux limit correspondsto an average jet-to-core X-ray flux ratio upper limitof 2.4%. Our measured jet-to-core flux ratios, both forthose directly measured and the upper limits, agree wellwith measurements from low-redshift quasars, suggest-ing that the observed X-ray jet emission mechanism isindependent of redshift. Deeper Chandra observationsare required to investigate the spectrum and spatialstructure of the detected X-ray features.Beyond our morphological study, we determined theoptical-to-X-ray power-law slope α ox for each quasarcore using optical/UV data available in the literature.We observed a clear anticorrelation trend between α ox and ℓ , where our derived best-fit relationship isconsistent with other quasar surveys. We also measuredradio-to-X-ray luminosity ratios for our sources, andour results are broadly consistent with other radio-loudquasar surveys regardless of redshift. These multiwave-length results reinforce that the spectral evolution ofquasars is independent of redshift.B. S, D. A. S., A. S., and M. S. were supported byNASA contract NAS8-03060 (Chandra X-ray Center).B. S. and D. A. S. were also supported by CXC grantGO8-19077X. Work by C. C. C. at the Naval ResearchLaboratory is supported by NASA DPR S-15633-Y. Ra-dio observations were provided by the National RadioAstronomy Observatory, a facility of the National Sci-ence Foundation operated under cooperative agreementby Associated Universities, Inc. Software:
CIAOv4.12,CALDBv4.9.2.1(Fruscione et al.2006),Sherpa (Freeman et al. 2001), scipy (Virtanen et al.2020)REFERENCES
Abazajian, K., Adelman-McCarthy, J. K., Ageros, M. A.,et al. 2003, ApJ, 126, 2081Avni, Y., & Tananbaum, H. 1982, ApJL, 262, L17Bechtold, J., Elvis, M., Fiore, F., et al. 1994, AJ, 108, 759Becker, R. H., White, R. L., & Helfand, D. J. 1995, ApJ,450, 559Begelman, M. C., Blandford, R. D., & Rees, M. J. 1984,RvMP, 56, 255 Bˆırzan, L., Rafferty, D. A., McNamara, B. R., Wise, M. W.,& Nulsen, P. E. J. 2004, ApJ, 607, 800Blandford, R. D., & Rees, M. J. 1974, MNRAS, 169, 395Breiding, P., Meyer, E. T., Georganopoulos, M., et al. 2017,ApJ, 849, 95Celotti, A., Ghisellini, G., & Chiaberge, M. 2001, MNRAS,321, L1Cheung, C. C. 2004, ApJL, 600, L23 Snios et al.
Cheung, C. C., Stawarz, L., & Siemiginowska, A. 2006,ApJ, 650, 679Cheung, C. C., Stawarz, L., Siemiginowska, A., et al. 2012,ApJ, 756, L20Civano, F., Marchesi, S., Comastri, A., et al. 2016, ApJ,819, 62Dickey, J. M., & Lockman, F. J. 1990, ARA&A, 28, 215Fabian, A. C. 1994, ARA&A, 32, 277—. 2012, ARA&A, 50, 455Fan, X.-L., & Bai, J.-M. 2016, ApJ, 818, 185Freeman, P., Doe, S., & Siemiginowska, A. 2001, Proc.SPIE, 4477, 76Fruscione, A., McDowell, J. C., Allen, G. E., et al. 2006,Proc. SPIE, 6270, 62701VGobeille, D. B., Wardle, J. F. C., & Cheung, C. C. 2014,VLA Observations of a Complete Sample of Radio LoudQuasars between redshifts 2.5 and 5.28: I. high-redshiftsample summary and the radio images, , ,arXiv:1406.4797Gobeille, D. B. P. 2011, A systematic meta-survey of highredshift quasars probing their environments andevolution, ,Goodger, J. L., Hardcastle, M. J., Croston, J. H., et al.2010, ApJ, 708, 675Grimmett, G., & Welsh, D. 1986, Probability: AnIntroduction (Oxford: Oxford Science Publications)Hargrave, P. J., & Ryle, M. 1974, MNRAS, 166, 305Harris, D. E., Cheung, C. C., Biretta, J. A., et al. 2006,ApJ, 640, 211Heinz, S., Reynolds, C. S., & Begelman, M. C. 1998, ApJ,501, 126Hinshaw, G., Larson, D., Komatsu, E., et al. 2013, ApJS,208, 19Husband, K., Bremer, M. N., Stanway, E. R., & Lehnert,M. D. 2015, MNRAS, 452, 2388Jester, S., Harris, D. E., Marshall, H. L., & Meisenheimer,K. 2006, ApJ, 648, 900Just, D. W., Brandt, W. N., Shemmer, O., et al. 2007, ApJ,665, 1004Kashyap, V. L., van Dyk, D. A., Connors, A., et al. 2010,ApJ, 719, 900Kelly, B. C., Bechtold, J., Siemiginowska, A., Aldcroft, T.,& Sobolewska, M. 2007, ApJ, 657, 116Lusso, E., & Risaliti, G. 2016, ApJ, 819, 154Marshall, H. L., Miller, B. P., Davis, D. S., et al. 2002,ApJ, 564, 683Marshall, H. L., Gelbord, J. M., Schwartz, D. A., et al.2011, ApJS, 193, 15Marshall, H. L., Gelbord, J. M., Worrall, D. M., et al. 2018,ApJ, 856, 66 Martocchia, S., Piconcelli, E., Zappacosta, L., et al. 2017,A&A, 608, A51McKeough, K., Siemiginowska, A., Cheung, C. C., et al.2016, ApJ, 833, 123McNamara, B. R., & Nulsen, P. E. J. 2007, ARA&A, 45,117Meyer, E. T., Breiding, P., Georganopoulos, M., et al. 2017,ApJ, 835, L35Meyer, E. T., Georganopoulos, M., Sparks, W. B., et al.2015, ApJ, 805, 154Nanni, R., Vignali, C., Gilli, R., Moretti, A., & Brandt,W. N. 2017, A&A, 603, A128Pˆaris, I., Petitjean, P., Aubourg, E., et al. 2018, A&A, 613,A51Perley, R. A., Dreher, J. W., & Cowan, J. J. 1984, ApJL,285, L35Rafferty, D. A., McNamara, B. R., Nulsen, P. E. J., &Wise, M. W. 2006, ApJ, 652, 216Rakshit, S., Stalin, C. S., & Kotilainen, J. 2020, ApJS, 249,17Reynolds, C. S., Heinz, S., & Begelman, M. C. 2001, ApJL,549, L179Scheuer, P. A. G. 1974, MNRAS, 166, 513Scheuer, P. A. G. 1982, in IAU Symposium, Vol. 97,Extragalactic Radio Sources, ed. D. S. Heeschen & C. M.Wade (D. Reidel), 163Schwartz, D. A. 2002, ApJ, 571, L71Schwartz, D. A., Marshall, H. L., Lovell, J. E. J., et al.2006a, ApJ, 647, L107—. 2006b, ApJ, 640, 592Schwartz, D. A., Siemiginowska, A., Snios, B., et al. 2020,ApJ, 904, 57Shemmer, O., Brandt, W. N., Schneider, D. P., et al. 2006,ApJ, 644, 86Siemiginowska, A., Bechtold, J., Aldcroft, T. L., et al. 2002,ApJ, 570, 543Siemiginowska, A., Smith, R. K., Aldcroft, T. L., et al.2003, ApJ, 598, L15Siemiginowska, A., Sobolewska, M., Migliori, G., et al.2016, ApJ, 823, 57Siemiginowska, A., Stawarz, L., Cheung, C. C., et al. 2007,ApJ, 657, 145Simionescu, A., Stawarz, L., Ichinohe, Y., et al. 2016, ApJ,816, L15Snios, B., Nulsen, P. E. J., Kraft, R. P., et al. 2019a, ApJ,879, 8Snios, B., Wykes, S., Nulsen, P. E. J., et al. 2019b, ApJ,871, 248Snios, B., Siemiginowska, A., Sobolewska, M., et al. 2020,ApJ, 899, 1275