Lyman-α emission from a WISE-selected optically faint powerful radio galaxy M151304.72-252439.7 at z = 3.132
Gitika Shukla, Raghunathan Srianand, Neeraj Gupta, Patrick Petitjean, Andrew J. Baker, Jens-Kristian Krogager, Pasquier Noterdaeme
MMNRAS , 1–19 () Preprint 6 January 2021 Compiled using MNRAS L A TEX style file v3.0
Lyman- α emission from a WISE-selected optically faint powerfulradio galaxy M151304.72-252439.7 at z = 3.132 (cid:63) Gitika Shukla † , Raghunathan Srianand , Neeraj Gupta , Patrick Petitjean ,Andrew J. Baker , Jens-Kristian Krogager , Pasquier Noterdaeme Inter-University Centre for Astronomy and Astrophysics (IUCAA), Post Bag 4, Pune 411007, India Institut dAstrophysique de Paris, UMR 7095, CNRS-SU, 98bis boulevard Arago, 75014 Paris, France Department of Physics and Astronomy, Rutgers, the State University of New Jersey, 136 Frelinghuysen Road, Piscataway,NJ 08854-8019, USA
ABSTRACT
We report the detection of a large ( ∼
90 kpc) and luminous Ly α nebula [ L Ly α =(6.80 ± × erg s − ] around an optically faint (r >
23 mag) radio galaxy M1513-2524 at z em =3.132. The double-lobed radio emission has an extent of 184 kpc, but theradio core, i.e., emission associated with the active galactic nucleus (AGN) itself, isbarely detected. This object was found as part of our survey to identify high- z quasarsbased on Wide-field Infrared Survey Explorer (WISE) colors. The optical spectrum hasrevealed Ly α , N v , C iv and He ii emission lines with a very weak continuum. Basedon long-slit spectroscopy and narrow band imaging centered on the Ly α emission, weidentify two spatial components: a “compact component” with high velocity dispersion( ∼ − ) seen in all three lines, and an “extended component”, having lowvelocity dispersion (i.e., 700-1000 km s − ). The emission line ratios are consistent withthe compact component being in photoionization equilibrium with an AGN. We alsodetect spatially extended associated Ly α absorption, which is blue-shifted within 250-400 km s − of the Ly α peak. The probability of Ly α absorption detection in such largeradio sources is found to be low ( ∼ α and radio luminosities. Deep integralfield spectroscopy is essential for probing this interesting source and its surroundingsin more detail. Key words:
Galaxies: active - galaxies: high-redshift - intergalactic medium - quasars:emission lines - quasars:individual: M151304.72 − The ubiquitous presence of extended Ly α emission arounddiverse populations of galaxies, ranging from quasars (Heck-man et al. 1991a,b; Borisova et al. 2016; Arrigoni Bat-taia et al. 2019) to powerful high- z radio galaxies (HzRGs)(see Chambers et al. 1990; Villar-Mart´ın et al. 2003; Villar-Mart´ın et al. 2007b) and several other populations such assub-millimetre and Lyman-break galaxies (Matsuda et al.2004; Chapman et al. 2001; Geach et al. 2014), is now wellestablished. The detection rate of such diffuse Ly α emissionfrom high- z galaxies has remarkably gone up to 100%, in (cid:63) Based on observations made with the Southern African LargeTelescope (SALT). † E-mail: [email protected] some recent studies (Borisova et al. 2016; Arrigoni Battaiaet al. 2019), using integral field spectrographs like the Multi-Unit Spectroscopic Explorer (MUSE; Bacon et al. 2010) andthe Keck Cosmic Web Imager (KCWI; Morrissey et al. 2012)on 8-10 metre class telescopes, operating at excellent astro-nomical sites. In particular, these instruments have increasedthe detection of faint and small scale Ly α nebulae (Hu &Cowie 1987; Farina et al. 2017; Wisotzki et al. 2018). Inprinciple, these extended Ly α emitting regions provide im-portant means to study the properties of the circumgalacticmedium (CGM) and the interface between CGM and the in-tergalactic medium (IGM) around high- z galaxies. Detailedstudies of the spatial distribution, kinematics and excitationof the gas traced by the extended Ly α emission can provide © The Authors a r X i v : . [ a s t r o - ph . GA ] J a n Shukla et.al vital clues on various feedback processes that drive star for-mation in high- z galaxies.Traditionally, studies of the CGM and IGM have beencarried out using absorption lines detected against brightbackground sources, typically bright quasars. The one-dimensional nature of absorption line studies, however, is notideal for probing the spatial distribution of the gas surround-ing individual galaxies. Despite several studies dedicated tounderstand the complex interactions between gas surround-ing the host galaxies and the galaxies themselves, very littleprogress has been made to date. In order to fully charac-terise the physical and kinematic properties of this gas, it isimportant to combine absorption and emission studies.The recent detections of enormous Ly α nebulae(ELANs; Matsuda et al. 2004; Cantalupo et al. 2014; Caiet al. 2017, 2018; Arrigoni Battaia et al. 2018, 2019), whichare characterized by very extended Ly α halos (Ly α extent >
200 kpc) with luminosity L Ly α > erg s − and sur-face brightness SB Ly α > − erg s − cm − arcsec − , havedrawn a lots of attention. As the diffuse emission in thesecases extend well beyond the virial radii of the galaxies,ELANs can be a very good probe of the dark matter po-tential of the halos hosting the galaxies and the interfaceregions between CGM and IGM. However, ELANs are rare,and only 1% of quasars seem to posses them. Based on theirradial surface brightness profiles, it has been proposed thatELANs trace regions of higher volume density and/or haveadditional sources of ionizing radiation embedded in theirhalos (Cantalupo et al. 2014; Arrigoni Battaia et al. 2018).However, a better understanding of these scenarios requiresfurther investigation.The extended Ly α emission studies centred on HzRGsand radio-loud quasars have shown a clear correlation be-tween powerful radio jets and diffuse gas (van Ojik et al.1997; Heckman et al. 1991a,b; Villar-Mart´ın et al. 2007b),where the jet clearly seems to alter the kinematics and mor-phology of the Ly α emitting gas. In a sample of 18 z > α emission, along with an anti-correlation between Ly α velocity width and radio size. In more than 60% of thesecases, strong associated H i absorption ( N HI (cid:62) cm − )was detected. The detection rate of H i absorption was foundto be higher in smaller radio sources (i.e., ∼
90% whenthe source size is <
50 kpc and 25% when >
50 kpc). In61% of the cases, Ly α was more extended than the radiosource itself. The inner parts of the Ly α halo within theextent of the radio emission showed perturbed kinematics(FWHM > − ) due to jet-gas interaction, whereasthe more extended ( ∼
100 kpc) regions were dominated byquiescent kinematics (FWHM <
700 km s − ). These resultsdemonstrate that HzRGs reside in gas rich environments.Deep IR imaging studies have also revealed the presenceof galaxy proto-clusters around high- z radio galaxies (Mayoet al. 2012; Galametz et al. 2012; Wylezalek et al. 2013; Dan-nerbauer et al. 2014). These proto-clusters show presence ofhigh density environments around high- z radio galaxies.Several possible physical mechanisms have been pro-posed to power the extended Ly α emission. In many casesthe Ly α nebulae are found to be associated with highlyobscured type-II AGNs (see, e.g., Dey et al. 2005; Bridge et al. 2013; Overzier et al. 2013; Hennawi et al. 2015; Aoet al. 2017). The most commonly discussed origins of Ly α emission are: (i) shock induced radiation, powered by radiojets or outflows (Mori et al. 2004; Allen et al. 2008); (ii)gravitational cooling radiation/ Ly α collisional excitation(Haiman et al. 2000; Dijkstra et al. 2006; Rosdahl & Blaizot2012); (iii) fluorescent Ly α emission due to photoionizationby UV luminous sources like AGNs or star forming galaxies(McCarthy 1993; Cantalupo et al. 2005; Geach et al. 2009;Overzier et al. 2013) and (iv) resonant scattering of Ly α photons from embedded sources (Villar-Martin et al. 1996;Dijkstra & Loeb 2008). The presence of high-ionization lineslike C iv and He ii can provide additional information on thekinematics of the gas and help disentangle the various physi-cal processes powering the Ly α emission. Extended emissionin these high ionization lines is observed for a few HzRGsover tens of kpc (Villar-Mart´ın et al. 2003; Morais et al.2017). There are a few detections in quasars as well, whichhave shown extended emission in these lines out to hundredsof kpc (Marques-Chaves et al. 2019).For the recently detected ELANs, photoionization by asso-ciated bright quasars has been suggested as the dominantpowering source for the Ly α emission (Borisova et al. 2016).However, contributions from other mechanisms like resonantscattering cannot be entirely ruled out (Arrigoni Battaiaet al. 2019). In the case of HzRGs showing Ly α extend-ing beyond the associated radio emission, jet-gas interactionis seen to play a vital role in determining the propertiesof the Ly α nebulae; however, to produce Ly α on spatialscales larger than that of the radio emission, photoioniza-tion by a continuum source is still required (van Ojik et al.1997). Stellar photoionization has also been shown to in-crease the ratios f Ly α / f CIV and f Ly α / f HeII , especially for“Ly α -excess” objects. There are cases where Ly α nebulae aredetected around radio-quiet type-II AGNs (Prescott et al.2015; Marques-Chaves et al. 2019). Nilsson et al. (2006) hadreported the presence of a 60 kpc large Ly α nebula not as-sociated with any optical source, which was argued to bepowered by gravitational cooling radiation. However, laterPrescott et al. (2015) argued that the absence of contin-uum at optical and longer wavelengths does not necessarilyindicate gravitational origin for Ly α emission. If anything,gravitationally powered Ly α emission should instead be ac-companied by a galaxy forming at the center of the Ly α nebula, whose stellar continuum one should be able to de-tect.We have recently completed deep long-slit spectroscopicobservations of extended Ly α emission in 25 newly discov-ered radio bright (1.4 GHz flux density in excess of 200 mJy)quasars and radio galaxies at z (cid:62) . α emission around a special radio source,M151304.72-252439.70 (where M stands for MALS), in oursample. In the literature, this radio source is also known asMRC1510-252 and TXS1510-252. The source not only hashigher values of radio power, angular size and Ly α lumi-nosity than other objects in our sample but also than otherknown radio sources at these redshifts. Interestingly, an op-tical counterpart is not detected in any of the optical im- MNRAS , 1–19 () y α emission from M151304.72-252439.70 ages from PanSTARRS-1 (PS1) (Chambers & Pan-STARRSTeam 2018), and only a very faint source is detected in in-frared images from the Wide-field Infrared Survey Explorer(WISE) (see Fig. 1). Our survey spectrum clearly revealedthe strong Ly α , N v , C iv and He ii emission lines but witha very faint continuum emission. The Ly α emission is foundto be extended with a clear signature of associated absorp-tion. All these attributes make M1513-2524 an interestingobject for detailed investigation.We have organized this paper as follows. In Section 2 wehave provided the details of our long-slit spectroscopic andnarrow-band imaging observations using SALT, along withupgraded Giant Meterwave Radio Telescope (uGMRT) ob-servations and data reduction. In Section 3, we present themain results of our study; in particular, we explore the con-nection between the radio jets/lobes and Ly α emitting gas.We compare the Ly α flux and radio size of M1513-2524 withthose of radio sources at similar redshifts in the literature.We discuss the properties of the associated Ly α absorptionand the detection of extended C iv and He ii emission linesin this section. We also provide a list of candidate Ly α emit-ters found in our narrow band observations. In Section 4, wesummarize and discuss our results. Throughout this paper,we have adopted a cosmology with H = 67.4 km s − Mpc − ,Ω m = 0.315 and Ω Λ = 0.685 (see Planck Collaboration et al.2018). At the emission redshift ( z ∼ (cid:48)(cid:48) corresponds to 7.8 kpc. In this section, we present details of the acquisition and anal-ysis of various data used in this work.
To obtain the optical spectrum, we used the Robert StobieSpectrograph (RSS) (Burgh et al. 2003; Kobulnicky et al.2003) on the Southern African Large Telescope (SALT)(Buckley et al. 2006) in long-slit mode (Program IDs:2016-2-SCI-017, 2017-1-SCI-016). The primary mirror ofSALT is 11m across, consisting of 91 1m individual hexag-onal mirrors. The RSS consists of an array of three CCDdetectors with 3172 × × (cid:48)(cid:48) , grating PG0900, GR-ANGLE=15.875 o and CAMANG=31.75 o . This combination provides a wave-length coverage of 4486-7533 ˚A, excluding the wavelengthranges 5497-5551 ˚A and 6542-6589 ˚A that correspond togaps between CCDs. These settings were chosen such thatLy α , C iv and He ii emission from the radio source aresimultaneously covered and Ly α emission falls in the mostsensitive part of the spectrograph. The spectral resolutionachieved is in the range of 200-300 km s − .During the survey, the source was observed using along-slit oriented at a position angle (PA) of 72 ◦ . To bet-ter understand the spatial distribution of gas, we followedup the target with deeper long-slit observations, keeping theslit oriented along PAs of 72 ◦ and 350 ◦ . These PAs werechosen so that we could have a blind off-set pointing at clip h m s s s s s s -25°23'30"24'00"30"25'00"30" J2000 Right Ascension J D e c li n a t i o n PS1 i-band M1513-2524 NE S S T PA=72T PA=350
Figure 1.
The i-band image of the field of M1513-2524 obtained inPan-STARRS1. The two slit positions (PA=72 ◦ and PA=350 ◦ )along which the long-slit observations were carried out with re-spect to reference stars (identified as S and S ) falling within theslit are shown using dashed lines. The red plus marks the WISEposition (RA=15 h m . s , Dec= − d m . s ) where nooptical source is detected. The cyan contours correspond to Band-5 (1.4 GHz) radio emission from our uGMRT observations, andthe corresponding synthesized beam is shown in red in the lowerleft corner. The contour levels are plotted at 1.6 × (-1, 1, 2, 4, 8,16, 32, 64, ..) mJy beam − . Radio contours from NVSS are alsoover plotted (gray color). These contour levels are at 2.1 × (-1, 1,2, 4, 8, 16, 32, 64, ..) mJy beam − . The symbol “T” correspondsto top of the 2D long-slit spectra shown in Figs. 8 and 9. the WISE location of the source using a bright star’s lo-cation (see Fig. 1). The total on-source exposure times were2400 s for each PA, split into 2 exposures of 1200 s each,with a dither of 2 (cid:48)(cid:48) along the slit. Seeing measured fromthe profiles of stars in the acquisition image are typically inthe range 1.6 (cid:48)(cid:48) to 2.0 (cid:48)(cid:48) during our observations. The spec-tral point spread function (SPSF) for each PA was con-structed from spatial profiles extracted after collapsing aregion between 5000-5020 ˚A for the reference stars. We mea-sured SPSFs of 1.7 (cid:48)(cid:48) and 1.4 (cid:48)(cid:48) for spectra obtained alongPA = 72 ◦ and 350 ◦ , respectively (see Table 1). The datawere reduced using the SALT science pipeline (Crawfordet al. 2010) and standard IRAF tasks . For cosmic ray re-moval, we used the IRAF based algorithm proposed by vanDokkum (2001). The wavelength calibration was performedusing a Xenon arc lamp. The spectrophotometric standardstar EG21 (RA=03 h m . s , Dec= − d m . s ) wasused for flux calibration, where corrections for atmosphericextinction and air mass have been taken into account. Thewavelengths were then shifted to vacuum wavelengths. Thefinal combined 1D spectrum of M1513-2524 is shown inFig. 2. IRAF is distributed by the National Optical Astronomy Obser-vatories, which are operated by the Association of Universities forResearch in Astronomy, Inc., under cooperative agreement withthe National Science Foundation.MNRAS000
The i-band image of the field of M1513-2524 obtained inPan-STARRS1. The two slit positions (PA=72 ◦ and PA=350 ◦ )along which the long-slit observations were carried out with re-spect to reference stars (identified as S and S ) falling within theslit are shown using dashed lines. The red plus marks the WISEposition (RA=15 h m . s , Dec= − d m . s ) where nooptical source is detected. The cyan contours correspond to Band-5 (1.4 GHz) radio emission from our uGMRT observations, andthe corresponding synthesized beam is shown in red in the lowerleft corner. The contour levels are plotted at 1.6 × (-1, 1, 2, 4, 8,16, 32, 64, ..) mJy beam − . Radio contours from NVSS are alsoover plotted (gray color). These contour levels are at 2.1 × (-1, 1,2, 4, 8, 16, 32, 64, ..) mJy beam − . The symbol “T” correspondsto top of the 2D long-slit spectra shown in Figs. 8 and 9. the WISE location of the source using a bright star’s lo-cation (see Fig. 1). The total on-source exposure times were2400 s for each PA, split into 2 exposures of 1200 s each,with a dither of 2 (cid:48)(cid:48) along the slit. Seeing measured fromthe profiles of stars in the acquisition image are typically inthe range 1.6 (cid:48)(cid:48) to 2.0 (cid:48)(cid:48) during our observations. The spec-tral point spread function (SPSF) for each PA was con-structed from spatial profiles extracted after collapsing aregion between 5000-5020 ˚A for the reference stars. We mea-sured SPSFs of 1.7 (cid:48)(cid:48) and 1.4 (cid:48)(cid:48) for spectra obtained alongPA = 72 ◦ and 350 ◦ , respectively (see Table 1). The datawere reduced using the SALT science pipeline (Crawfordet al. 2010) and standard IRAF tasks . For cosmic ray re-moval, we used the IRAF based algorithm proposed by vanDokkum (2001). The wavelength calibration was performedusing a Xenon arc lamp. The spectrophotometric standardstar EG21 (RA=03 h m . s , Dec= − d m . s ) wasused for flux calibration, where corrections for atmosphericextinction and air mass have been taken into account. Thewavelengths were then shifted to vacuum wavelengths. Thefinal combined 1D spectrum of M1513-2524 is shown inFig. 2. IRAF is distributed by the National Optical Astronomy Obser-vatories, which are operated by the Association of Universities forResearch in Astronomy, Inc., under cooperative agreement withthe National Science Foundation.MNRAS000 , 1–19 ()
Shukla et.al
Table 1.
Log of RSS/SALT long-slit observations for M1513-2524 + PA Observing date Wavelength coverage FWHM of SPSF Air mass Exposure time(s)(deg) year/mm/dd (˚A) (arcsec)72 2016-02-17 4486-7533 1.74 1.19 2 × × × + Slit width of 1.5 arcsec was used.
Observed wavelength(Å) f ( e r g s c m Å ) L y - [ . Å ] N V [ . Å ] O I [ . Å ] C II [ . Å ] S i I V [ . Å ] C I V [ . Å ] H e II [ . Å ] N I V [ . Å ] O III [ . Å ] Figure 2.
1D spectrum of M1513-2524 obtained using SALT. The Ly α , C iv , He ii and weak N v emission lines are detected, while otheremission lines like [O i ] and Si iv were not detected. The response curve of the filter “PI05060” used for the narrow band imaging coveringLy α emission is shown as a black dashed curve. Vertical dotted lines mark the expected locations of commonly seen emission lines inAGNs (see Vernet et al. 2001). The uGMRT Band-5 (1050-1450 MHz) and Band-3 (250- 500 MHz) observations of M1513-2524 were carried outas part of our larger surveys to map the radio continuumand search for H i (cid:48)(cid:48) × (cid:48)(cid:48) withPA=1.0 ◦ and 3.7 (cid:48)(cid:48) × (cid:48)(cid:48) with PA=-35.5 ◦ , respectively. Theflux density measurements from our uGMRT observationstowards different lobes are summarized in Table 3. Total fluxdensity measurements from the literature are summarized inTable 4. In order to accurately determine the location of the opti-cal source and quantify any influence of the radio source onthe overall distribution of Ly α emitting gas, we obtaineddeep narrow band images of M1513-2524 using RSS filterPI05060, centered at λ cen ≈ . × (cid:48)(cid:48) per pixel. Each individual frame is a combination of3 CCD chips and 2 CCD gaps with a circular field of viewof (7 . (cid:48) ) . The observations were carried out such that theWISE source lay on the central CCD chip, and individualframes were dithered to avoid CCD artefacts. The obser-vations were performed in service mode on March 2, 2019with total exposure time of 2400 s, split into 4 exposuresof 600 s each. During the course of the observations, 25 (cid:48)(cid:48) ofdither was adopted between exposures. The data were first MNRAS , 1–19 () y α emission from M151304.72-252439.70 Figure 3. uGMRT Band-3 (420 MHz; rms = 10 mJy beam − )contours overlaid on Band-5 (1360 MHz; rms=0.4 mJy beam − )image of M1513-2524. The synthesized beams provided in the textare plotted in the lower left corner. The color wedge at the topcorresponds to the Band-5 image. The contour levels are: 40 × (-1, 1, 2, 4, 16, 32, 64, ...) mJy beam − . corrected for bias, cross talk and gain. We further used stan-dard IRAF tasks to remove cosmic rays and flat fielded theindividual frames using dome flats.To perform the astrometric calibration of the narrowband images, we used 22 stars distributed across the cen-tral CCD chip, for which accurate coordinates can be ob-tained from the PS1 survey. After astrometric calibration,objects in our narrow band image have a positional accu-racy of 0.14 (cid:48)(cid:48) . As a next step, we subtracted the backgroundfrom each individual exposure. To do that, we have utilisedour script written in
Python , employing the backgroundsubtraction technique provided by
Astropy , a commumitydeveloped
Python package for astronomy (Astropy Col-laboration et al. 2013, 2018). In this code, we essentiallymask all > σ sources and the CCD gap regions and usethe remaining pixels to estimate a 2D background model.We used SExtractorBackground as a background es-timator (see the
Background2D class of
Photutils fordetails) in this process. We also performed the subtractionusing the
Medianbackground estimator, and the resultsfrom the two methods were consistent within 5%. After sub-tracting the 2D background model, we subtracted further aconstant mean background estimated after masking all thesources above 2 σ . Finally, we combined the narrow band im-ages using the IRAF task “ imcombine ”, with all the imagesscaled and weighted by the “median” counts measured ineach frame.The seeing FWHM measured using point sources inthe final image, estimated using a two dimensional Gaus-sian fit, is ≈ (cid:48)(cid:48) . The narrow band image is oversampled(0.1267 (cid:48)(cid:48) per pixel resolution) compared to the long-slit ob-servations (0.2534 (cid:48)(cid:48) per pixel resolution). To account for this oversampling effect, we binned the combined narrow bandimage by 2 × (cid:48)(cid:48) . We also smoothed the narrow band image usinga Gaussian kernal of FWHM=1.2 (cid:48)(cid:48) . We used this resampledand smoothed image to quantify the extent of Ly α emission.In Fig. 4, we have shown the final combined narrow bandimage of M1513-2524 after smoothing. The uGMRT Band-5(1.4 GHz) contours are plotted in cyan. For the Ly α emis-sion, the 3 σ contour is plotted as a green dashed line, andblack contours are at 5, 8, 18, 40 and 70 σ levels. The WISEposition of the source is marked by a blue plus. The distancemeasured between the flux weighted centroid of the Ly α emission and the WISE location of quasar is 0.26 (cid:48)(cid:48) which is ∼ z = 3 .
13. Thus, within the astrometric accuracy,the Ly α peak is coincident with the infrared source.The flux calibration of the narrow band image was per-formed using the spectra of the two reference stars (S andS observed with long-slit observations, shown in Fig. 1)and their broad band photometry obtained from PS1. Giventhe wavelength coverage of our long-slit observations, wechose to use only r-band magnitude for this purpose. Af-ter comparing the PS1 r-band flux with the flux measuredfrom the long-slit spectra within the wavelength range ofthe r-band, we obtained slit-loss factors for individual ex-posures at each PA. First we combined the slit-loss cor-rected exposures corresponding to each PA. We then usedcombined slit-loss corrected star spectra of the two PAsto find the total flux (f ∗ tot ) within the wavelength rangecovered by the narrow band observations. To find the ob-served total counts of the stars from the narrow band im-age, we estimated aperture sizes from radial surface bright-ness profiles. For each reference star, the aperture radiuswas fixed as the radius where surface brightness reaches 2 σ relative to the background. A simple scaling between thetotal flux of the star, f ∗ tot , and the total counts obtainedfrom the narrow band image of the same star, gives flux perunit count, f c = 3 . × − and 4 . × − erg s − cm − for PA=72 ◦ and 350 ◦ , respectively. We use the average ofthese two values, f c = 4 . × − erg s − cm − , to flux cal-ibrate the narrow band image. The counts to flux conver-sion factors (f c ) estimated from the two PAs are consistentwith each other within 15%. To estimate the surface bright-ness limit (SB lim ) reached in our combined, resampled andsmoothed narrow band image, we put boxes of size 2.5 (cid:48)(cid:48) x2.5 (cid:48)(cid:48) in background locations and estimate the SB in eachbox. The mean and rms of these SB values provides SB lim .The 2 σ SB lim derived using this method is 8 . × − erg s − cm − arcsec − . In this section we discuss our findings based on the opticaland radio observations presented in previous sections.
We clearly detect Ly α , C iv and He ii emission lines in ourlong-slit spectrum. N v is also detected at the expected lo- MNRAS000
We clearly detect Ly α , C iv and He ii emission lines in ourlong-slit spectrum. N v is also detected at the expected lo- MNRAS000 , 1–19 ()
Shukla et.al h m s s s -25°24'30"15" RA (J2000) D E C (J ) Ly NB
N E
T T uGMRT 1.4 GHz flux [10 erg s cm ] h m s s s -25°24'30"15" RA (J2000) D E C (J ) Ly NB
N E
T T uGMRT 1.4 GHz flux [10 erg s cm ] Figure 4.
Left panel: Ly α emission observed in the narrow band image. Right panel:
Narrow band image after subtracting the contributionfrom the central source. In both the panels the uGMRT Band-5 radio contours (cyan color) are overlaid on top and drawn at the samelevels as in Fig. 1. The dashed green contour, corresponds to a 3 σ level (1 . × − erg s − cm − ) and black contours are drawn at(1.6, 2.5, 5.5, 11 and 20 ) × − erg s − cm − . The flux levels quoted are values per pixel. The position of the WISE source is marked bya blue plus symbol. North and East directions are shown at lower left. The images are smoothed by a Gaussian kernal of FWHM=1.2 (cid:48)(cid:48) .The pink line joining two points gives the maximum extent of the Ly α blob at the 3 σ flux level. The dashed red line indicates PA= 72 ◦ ,and the dotted red line indicates PA= 350 ◦ (also see Fig. 1). Table 2.
Measurements from the long-slit observationsPA f
Lyα L Lyα size a flux lim FWHM bLy α FWHM bCIV
FWHM bHeII (10 − erg s − cm − ) (10 erg s − ) (kpc) (10 − erg s − cm − ) (km s − ) (km s − ) (km s − )72 4.40 ± .
04 3.95 ± .
05 75 1.5 1269 ±
17 1363 ±
94 1027 ± ± .
05 4.30 ± .
04 58 1.7 1362 ±
16 1380 ±
59 962 ± a Defined by emission detected at a > σ level. b FWHM from the fitted 1D Gaussian. cation at a 7 σ level (see Fig. 2). The continuum is weakand consistent with the non-detection of the source in theoptical broad band photometric images from PS1. The non-resonant He ii emission line was used to determine the sys-temic redshift of the AGN. The redshift was measured usingcombined 1D spectrum of both PAs, and the He ii emissionline was modelled using a single Gaussian component, yield-ing z em = 3 . ± . ii line (after taking into account luminosity dependent ef-fects) has a mean velocity shift of 8 km s − with an intrinsicuncertainty of 242 km s − with respect to the systemic red-shift. To compute the emission line fluxes and ratios, we usedthe combined 1D spectrum corrected for slit-loss. The ob-served Ly α flux is f Ly α = (4 . ± . × − erg s − cm − ,and the corresponding luminosity is L Ly α = (4 . ± . × erg s − . For the C iv and He ii lines, we estimated f CIV =(8 . ± . × − erg s − cm − and f HeII =(3 . ± . × − erg s − cm − . For each line, the wavelength region overwhich the fluxes were integrated is marked by vertical dashedlines in Fig. 5. The fluxes given above were estimated aftersubtracting the constant continuum from the line profiles.The Ly α velocity dispersion (FWHM Ly α ), estimated fromthe unabsorbed Gaussian component (see Sec. 3.8 for detailson fitting) is 1331 ±
16 km s − . The velocity dispersions forC iv and He ii were estimated after fitting the observed lineprofiles with single Gaussian components. The estimated ve-locity dispersions are 1470 ±
69 km s − and 968 ±
53 km s − MNRAS , 1–19 () y α emission from M151304.72-252439.70 Ly C IV
He II
Relative velocity (km s ) f [ e r g s c m Å ] Figure 5.
Emission line profiles of Ly α , C iv and He ii (bluecurves), fitted using 1D Gaussian and Voigt function for Ly α and1D Gaussians for C iv and He ii . The velocity scale is set withrespect to z em =3.1320 obtained using the He ii line. The dottedhorizontal line is plotted at zero flux level. Two vertical dashedlines indicate the region over which the line flux is integrated. Redshift f C I V / f H e II Redshift f L y / f C I V Redshift f L y / f H e II Figure 6.
Different emission line flux ratios as a function of z forhigh- z radio sources from the literature. The location of M1513-2524 is shown with a star symbol in each panel. The dashed redline in the bottom panel is the f Ly α / f HeII cut-off for “Ly α excess”objects (with this ratio (cid:62)
14) by definition of VM07. for C iv and He ii , respectively. The FWHMs quoted hereare not corrected for instrumental broadening, which wouldentail corrections of typically less than 5%. Note that theFWHM values estimated from the combined spectra ob-tained with both PAs are slightly different from the valuesobtained from spectra along individual PAs, given in Table2, where we have summarized measurements from long-slitobservations at both PAs. Having obtained the line fluxes, we compare the line ra-tios to understand the nature of the source. The estimatedline ratios for f Ly α / f CIV , f Ly α / f HeII and f CIV / f HeII are ∼ ± ± ± f NV / f CIV and f NV / f HeII are ∼ ± ± α emission in a sample of radio galax-ies at 1.8 (cid:54) z (cid:54) α excess” objects and “Non-Ly α excess” ob-jects, based on the ratios of f Ly α / f CIV , f Ly α / f HeII and f CIV / f HeII . In the bottom panel, we plot the f Ly α / f HeII ratio as a function of z . From this figure (see also Table2 of Villar-Mart´ın et al. 2007a, VM07 from now on), it isclear that the value of this ratio observed for M1513-2524is lower than the median of measured values for z (cid:62) f Ly α / f HeII = 17.50) and consistent with(or marginally higher than) the values seen in 2 < z < f Ly α / f HeII in M1513-2524 isroughly at the expected cut-off for “Ly α excess” objects de-fined by VM07 ( f Ly α / f HeII (cid:62)
14, shown as the dashed line inthe bottom panel of Fig. 6). Note VM07 have found about54% of the z > f Ly α / f HeII (cid:62)
14. The large value of this ratio was inter-preted as a signature of excess star formation in these ob-jects. Consistent with its “Non-Ly α excess” line ratios, thelargest angular size (LAS) of the associated radio source inM1513-2524 is much larger than the extent of radio emissionassociated with “Ly α excess” objects in VM07. From Fig. 6,we can see that the flux ratio f Ly α / f CIV is consistent withthe median value of the “non-Ly α excess” objects (median f Ly α / f CIV = 5.45) (see also Table 3 of VM07). The mea-sured f CIV / f HeII is slightly higher than (i.e., within 1 . σ level) the median values seen for “Non-Ly α excess” objects(median f CIV / f HeII = 1.0). It is widely believed that pho-toionization by a central AGN is the cause of Ly α emissionin “Non-Ly α excess” objects. Thus, we conclude that the lineratios found in the case of M1513-2524 are consistent withphotoionization due to a central AGN. In the Band-5 image, at 1.3 (cid:48)(cid:48) from the WISE position, whichis well within the WISE positional uncertainty, a 4 σ feature(see Fig. 4) is detected in the radio continuum. It has apeak flux density of 1.7 mJy beam − . This feature couldbe emission from the radio core/jet (i.e., AGN associatedwith the WISE detection). Deeper and multi-frequency ra-dio interferometric images with higher spatial resolution arerequired to establish its origin. Our Band-3 image doesn’thave adequate spatial resolution for this purpose. Based onthe Band-5 image, the largest angular size (LAS), which wedefine as the separation between the peaks of the easternand western lobes, is 23.7 (cid:48)(cid:48) . The western lobe is closer tothe WISE position (i.e., at a linear separation of ∼
52 kpc)
MNRAS000
MNRAS000 , 1–19 ()
Shukla et.al
Table 3.
Properties of the two radio lobes of M1513-2524 basedon the uGMRT observations.Radio component S S α (mJy) (mJy)Eastern lobe 48.3 ± ± ± ± ± ± Table 4.
Flux densities of M1513-2524 from the literature.Survey/Band Flux(Jy)NVSS 1.4 GHz 0.218 ± ± ± ± compared to the eastern lobe (i.e., at a linear separation of ∼
131 kpc). This is the largest radio source in our sampleof 25 radio loud AGNs at z > ∼ ◦ .The total flux densities of the eastern and western lobesat 1360 MHz are 48.3 mJy and 117.8 mJy, respectively.The same flux density measured at 420 MHz are 242.1 mJyand 819.4 mJy. Based on these measurements, the easternand western lobes have spectral indices of α = − − ν ∝ ν α ). These measurements arealso summarised in Table 3.In the lower spatial resolution NRAO VLA Sky Sur-vey (NVSS) image at 1.4 GHz, the total flux density asso-ciated with the source M1513-2524 is 217.6 mJy (assumingno variability). This measurement implies that 25% of thetotal radio flux is diffuse and was resolved out in our Band-5 uGMRT image, due to insufficient short baseline coverageand thus lower sensitivity to large spatial scales.We now use the above measured spectral indices to es-timate the flux densities at 150 MHz for both the easternand western lobes. The estimated flux densities at 150 MHzare ∼
995 and 4485 mJy for the eastern and western lobes,respectively. To estimate the total radio flux density at 150MHz associated with M1513-2524, we simply add the twovalues. This amounts to a total flux density of ∼ <
300 MHz in the observedframe (990 MHz in the rest frame). α luminosity and excess LAEs The total Ly α flux estimated from the flux calibrated narrowband image within an aperture where the SB reaches 2 σ rel-ative to the background is (7 . ± . × − erg s − cm − ,and the total Ly α luminosity L Lyα is (6 . ± . × erg s − . The Ly α flux and luminosity estimated from thePSF subtracted image (details are given in section 3.5), ex-cluding a central region of radius 1.8 (cid:48)(cid:48) , are (3 . ± . × − erg s − cm − and (2 . ± . × erg s − , respec-tively. Note that the Ly α emission is at the edge of thenarrow band filter response (see Fig. 2). Thus, the actualLy α luminosity could still be higher than the above value.In VM07, there are 11 sources which are considered to be“Ly α -excess” objects. We obtained the luminosities for thesesources from the quoted fluxes using our cosmological pa-rameters and the median Ly α luminosity obtained is ∼ . × erg s − . Interestingly, the measured Ly α luminos-ity of M1513-2524 is higher than the median Ly α luminosityof the “ Ly α -excess” objects. We also provide an estimate for the continuum contri-bution to the total Ly α flux. We measure the continuumflux from the 1D spectrum, integrating over the FWHMof the narrow band filter using the wavelength range blue-wards of the Ly α line. The estimated continuum flux is ∼ (7 . ± . × − erg s − cm − , which is close to 10% ofthe Ly α flux measured from narrow band image. We alsoestimate an upper limit on the continuum contribution tothe narrow band image using the PS1 r-band image. Here,we estimate the 5 σ limiting magnitude at the WISE loca-tion of M1513-2524 from the r-band image using an apertureof the same size used to estimate the Ly α flux in the nar-row band image. The estimated r-band limiting magnitudeis ∼ ∼ . × − erg s − cm − ˚A − . Assum-ing this constant flux over the FWHM of the narrow bandfilter, the estimated total continuum flux is ∼ . × − erg s − cm − , i.e., only 9% of the total Ly α flux.We further check for the presence of other ( (cid:62) σ ) Ly α emitters in the FoV of our narrow band image, that werenot seen in PS1 images. We have detected seven Ly α emit-ting candidates in addition to M1513-2524 within the FoV(i.e ∼
50 arcmin ). The locations of these objects are iden-tified with circles in Fig. 7. We have listed these sources inTable 5. The first two columns in this table list the coordi-nates of the Ly α emitter candidates. In the third, fourth andfifth columns we provide narrow band fluxes (per unit wave-length), luminosities assuming these are Ly α emitters andsignificance levels of detection respectively. We estimatedthe expected r-band magnitudes of these sources assumingflat continua (over the wavelength range of the r-band) withflux similar to what we measure in our narrow band image.These values are provided in column 5 of Table 5. The 5 σ limiting magnitude reached in the combined PS1 r-band im-age at the source locations are given in the last column ofTable 5. We have used 2 (cid:48)(cid:48) × (cid:48)(cid:48) box apertures around the Ly α candidates to find the fluxes in the narrow band image and5 σ limiting magnitude from the PS1 r-band image. It is ev-ident that in five cases where we detect sources at a (cid:62) σ level, if the signal were due to continuum flux, we wouldhave seen these objects at a ∼ σ level in the PS1 r-bandimage. MNRAS , 1–19 () y α emission from M151304.72-252439.70 Table 5.
List of candidate LAEs detected in the narrow band image above 3.0 σ .RA DEC flux per unit wavelength L Ly α a significance m est r m PS1 r lim (HH:MM:SS) (DD:MM:SS) (10 − erg s − cm − ˚A − ) (10 erg s − ) ( σ )15:12:53.6068 -25:27:27.021 1.1 ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± a Ly α luminosities estimated from the narrow band image, assuming fluxes to be that of Ly α . Figure 7.
Left panel:
Combined and smoothed narrow band image of M1513-2524. The image has been rotated so that North is up andEast is to the left.
Right panel:
PS1 i-band image. In both the panels, the FoV of our narrow band image is shown by the largest circle(yellow). The 2 (cid:48) region centred on M1513-2524 is shown by dotted the blue circle. The candidate LAEs detected above 3 σ in this FoVare marked by green circles. The red circle is for M1513-2524. The coordinates and magnitudes of these LAEs are given in Table 5. Thecyan contours (pointed to with the orange arrows) are for the uGMRT Band-5 radio emission, same as shown in Fig. 1. If we assume the measured narrow band flux is due toLy α emission, then the inferred luminosities (see Table 5)are close to or higher than what is expected for L* values( ∼ . × erg s − ) measured for Ly α emitters at z ∼ ∼ − suchgalaxies per Mpc − . For the total volume sampled by ournarrow band observations, we expect only 0.2 Ly α emittersbrighter than L*. It is also clear from the Figure that fiveout of the seven identified candidates are within 2 (cid:48) of M1513-2324. Thus if the candidates identified here turn out to bethe real Ly α emitters around z ∼ .
13, it is highly likelythat M1513-2324 is part of a large over-dense region of starforming galaxies. This result would be consistent with whathas been found for a few powerful radio galaxies at theseredshifts (see references given in the introduction section). Deep multiobject spectroscopic observations are importantto confirm these candidates as Ly α emitters, and verify thepresence of excess galaxies around M1513-2324. In Fig. 8, we show the 2D spectra over the wavelength rangesof the three main emission lines observed at the two position
MNRAS000
MNRAS000 , 1–19 () Shukla et.al
Figure 8.
2D spectra of M1513-2524 obtained with long-slit oriented along 2 PAs. The upper panel corresponds to PA=72 ◦ , and the lowerto PA=350 ◦ . In both rows, the first, second and third columns are for Ly α , C iv and He ii emission, respectively. The fourth column isthe normalized SPSF corresponding to each PA. In the left panel, the thick green contour corresponds to the 3 σ level ( 1 . × − and 1 . × − erg s − cm − ˚A − for PA=72 ◦ and PA=350 ◦ , respectively). For the C iv and He ii emission, we have applied boxcarsmoothing of 1.27 (cid:48)(cid:48) × σ , and subsequentcontours are drawn at [4, 5, 6, 7, 9, 11 and 13] σ levels. Note the difference between color bar ranges (top axis of each panel) for Ly α emission vs. C iv and He ii emission. The horizontal dashed lines enclose the spatial sections of Ly α emission used for extracting thespectral profiles shown in Fig. 12 from the SPSF subtracted spectra. The North and East directions of Fig. 1 are shown in left panels. angles discussed above. The velocity scales are given withrespect to the emission redshift measured using the He ii emission lines (i.e., z em = 3.1320). On the right hand side ofeach panel, we also provide the spectral PSF (SPSF). As ourobservations were carried out in moderate seeing conditions,we perform SPSF subtraction to remove the contribution ofemission from the central regions to the extended emissionseen at large distances, in order to determine the extentof Ly α emission. It is clear from this figure that the spa-tial extent of the Ly α emission is larger than the SPSF andnon-uniform over the wavelength range covered by the emis-sion. The entire Ly α emitting region can be split into twospatial components, a “compact” spatial component [Ly α (c),spread over ∼ -367 to -1741 km s − ] bluewards of the absorp-tion trough and an “extended” spatial component [Ly α (e),spread over ∼ -128 to 1603 km s − ] redwards of the absorp-tion trough.To estimate the extent of the Ly α emission that is unaf-fected by seeing smearing, we carried out SPSF subtractionfrom the 2D spectra. For each PA, we construct SPSF fromthe reference star spectra after collapsing a ∼ α central wavelength. We then carry out asimple two parameter (i.e., peak amplitude and centriod ofthe SPSF) fit using chi-square minimization with respect tothe spatial profile of Ly α emission at every wavelength. Weconstructed 2D SPSF spectra using the best fit amplitude and centroid. We refine the centroid position as a function ofwavelength by linear fit. This was used for the final positionwhere the peak of the SPSF was centred. We then simplysubtract these 2D SPSF model spectra from the Ly α emis-sion. The 2D spectra along PA =72 ◦ and PA=350 ◦ afterSPSF subtraction are shown in Fig. 9.The extent of diffuse Ly α emission is estimated fromthe SPSF subtracted 2D spectra as the maximum diame-ter of the 3 σ flux contour in the continuum subtracted 2Dspectrum (enclosed within two red dashed curve, in Fig.9). It is evident from the figure that the Ly α (e) compo-nent extends up to ∼
75 kpc at 3 σ flux limit of 1 . × − erg s − cm − along PA =72 ◦ and ∼
58 kpc (flux limit of1 . × − erg s − cm − ) along PA =350 ◦ . We note thatthe Ly α (c) component does not extend beyond the SPSFalong PA =72 ◦ ; however, it is asymmetric and extends upto ∼
19 kpc along PA =350 ◦ .To characterise the morphology and estimate the ex-tent of the full Ly α emitting region, we use our re-samplednarrow band image (see Fig. 4). Again to remove the ef-fects of seeing smearing on the emission measured at largedistances, we subtract the PSF from this image as well. Toconstruct the PSF, we have taken a star present in the FoVof the narrow band image. We follow the same empiricalmethod for PSF subtraction used by various studies in theliterature (see, e.g., Borisova et al. 2016; Arrigoni Battaia MNRAS , 1–19 () y α emission from M151304.72-252439.70 Figure 9.
Same as Fig. 8, except that here spectra are plotted after subtracting the SPSF. The contour levels are as shown in Fig. 8. Thehorizontal red dashed lines indicate the spatial extents of the emission lines discussed in Sec.3.5. et al. 2019). This method of PSF subtraction is based onsimple scaling of the normalised PSF image to match theflux within the central region of Ly α emission. Usually inthe literature studies, the central 1 × around thequasar is used for finding the rescaling factor. These studieswere done under excellent seeing conditions (ranging from0.59-1.31 arcsec); due to our poor seeing of ∼ (cid:48)(cid:48) , we haveused the area within the central 2 (cid:48)(cid:48) radius to find the scalingfactor. Before we subtract the rescaled PSF, the centroidof the PSF image is aligned with the flux weighted spatialcentroid of the Ly α region of M1513-2524 above 3 σ . Thearea chosen is such that there are no other sources withinthis region. We measure the Ly α extent from the PSF sub-tracted image after smoothing it using a Gaussian kernelof FWHM=1.2 (cid:48)(cid:48) . The smoothed narrow band images beforeand after PSF subtraction are shown in the left and rightpanel of Fig. 4, respectively. The pink line joining two pointsin Fig. 4 shows the maximum Ly α extent at 3 σ (flux limit=1 . × − erg s − cm − ), equal to ∼
90 kpc. Note that thelargest Ly α extent seen in the narrow band image doesn’talign with the radio axis and the angle between the radioaxis, and the direction where we see the largest Ly α extentis ∼ ◦ .In Fig. 10, we have also shown a direct comparison ofthe spatial profiles of Ly α (c) and Ly α (e) before SPSF sub-traction, for better visualization of the extent of these twocomponents. It is clear from these figures that the Ly α (e)component extends much beyond the SPSF. Ly α (c) compo-nent also seems to possess an extended component along PA= 350 ◦ , which is consistent with the result discussed aboveusing 2D SPSF subtracted image. However, Ly α (c) at PA =72 ◦ is noisy, and extended Ly α can not be confirmed inthis case.Apart from the detection of the extended Ly α blob, weinvestigate if C iv and He ii lines are also spatially extended,which will be important for understanding the physical con-ditions prevailing in the extended line emitting region. Theselines are also important for constraining the kinematics ofthe gas, as the Ly α profile is sensitive to radiative transfereffects. To draw a conclusion on the extension of C iv andHe ii lines, we subtract the SPSF from their 2D spectra aswell (see Fig. 9). It is clear from the the SPSF subtractedimages that both C iv and He ii lines are clearly extendedalong PA =72 ◦ . In both cases, the residual emission found isasymmetric, with more extended emission on one side thanthe other. However, these two lines do not appear extendedalong PA =350 ◦ beyond the SPSF. Thus, our data suggestan asymmetric distribution of extended C iv emission. Inparticular, it is interesting to note the direction in which wefind the maximum extension (i.e., PA = 72 ◦ ) is closer to theradio axis (i.e., PA = 96 ◦ ). To explore this issue further, inFig. 11, we have also shown a comparison of the spatial pro-files of C iv and He ii with the SPSF for the two PAs. Basedon this analysis, it is clear that both C iv and He ii emissionare significantly extended along PA =72 ◦ . However, for PA=350 ◦ , the significance of the detection of extended emissionfor C iv and He ii is not high. In Fig. 10 and 11, we have alsoshown a 2-component Gaussian fit to the observed spatialprofiles, demonstrating the presence of extended emissionat large spatial scales. Thus, based on our long-slit spec-troscopy it appears that the Ly α , C iv and He ii emissionmay not be isotropic in nature, with the largest extent per- MNRAS000
90 kpc. Note that thelargest Ly α extent seen in the narrow band image doesn’talign with the radio axis and the angle between the radioaxis, and the direction where we see the largest Ly α extentis ∼ ◦ .In Fig. 10, we have also shown a direct comparison ofthe spatial profiles of Ly α (c) and Ly α (e) before SPSF sub-traction, for better visualization of the extent of these twocomponents. It is clear from these figures that the Ly α (e)component extends much beyond the SPSF. Ly α (c) compo-nent also seems to possess an extended component along PA= 350 ◦ , which is consistent with the result discussed aboveusing 2D SPSF subtracted image. However, Ly α (c) at PA =72 ◦ is noisy, and extended Ly α can not be confirmed inthis case.Apart from the detection of the extended Ly α blob, weinvestigate if C iv and He ii lines are also spatially extended,which will be important for understanding the physical con-ditions prevailing in the extended line emitting region. Theselines are also important for constraining the kinematics ofthe gas, as the Ly α profile is sensitive to radiative transfereffects. To draw a conclusion on the extension of C iv andHe ii lines, we subtract the SPSF from their 2D spectra aswell (see Fig. 9). It is clear from the the SPSF subtractedimages that both C iv and He ii lines are clearly extendedalong PA =72 ◦ . In both cases, the residual emission found isasymmetric, with more extended emission on one side thanthe other. However, these two lines do not appear extendedalong PA =350 ◦ beyond the SPSF. Thus, our data suggestan asymmetric distribution of extended C iv emission. Inparticular, it is interesting to note the direction in which wefind the maximum extension (i.e., PA = 72 ◦ ) is closer to theradio axis (i.e., PA = 96 ◦ ). To explore this issue further, inFig. 11, we have also shown a comparison of the spatial pro-files of C iv and He ii with the SPSF for the two PAs. Basedon this analysis, it is clear that both C iv and He ii emissionare significantly extended along PA =72 ◦ . However, for PA=350 ◦ , the significance of the detection of extended emissionfor C iv and He ii is not high. In Fig. 10 and 11, we have alsoshown a 2-component Gaussian fit to the observed spatialprofiles, demonstrating the presence of extended emissionat large spatial scales. Thus, based on our long-slit spec-troscopy it appears that the Ly α , C iv and He ii emissionmay not be isotropic in nature, with the largest extent per- MNRAS000 , 1–19 () Shukla et.al haps aligned close to the radio axis. Note in our narrow bandimage, the Ly α extent is not aligned with the radio source.Therefore, it will be interesting to acquire IFS observationsof Ly α and C iv emitting regions to study the anisotropicdistribution of gas excitation and its connection to the radioemission (see for example, McCarthy 1993).To quantify the 2D morphology of the Ly α halo, wemeasure the asymmetry parameter α = b/a from the un-smoothed narrow band image using the procedure explainedin Arrigoni Battaia et al. (2019) (here after A19). Theasymmetry parameter α is obtained using the second-ordermoment of the light distribution of the nebula, where a and b are the semi-major and semi-minor axes, respectively.We obtain α = 0 .
84 for the region within the 3 σ isophote.This value suggests a symmetric distribution as found formajority of high- z quasars (Arrigoni Battaia et al. 2019).As discussed above, the 2D spectra and the spatial pro-files suggest the existence of two components, i.e., a compactcore region and an extended Ly α halo region. The narrowband image is the superposition of these two components.To test whether the Ly α emission is isotropic even in thecentral regions, we consider the isophote corresponding tothe flux level of 1 . × − erg s − cm − (i.e., 10 σ level ofthe background) in the unsubtracted narrow band image.We find α = 0 .
79 for this isophote. Interestingly, the semi-major axis of this isophote is aligned close to the radio axisand the radius at 3 σ and 10 σ isophote are ∼ (cid:48)(cid:48) and 2.1 (cid:48)(cid:48) respectively. Thus, our data suggest that the Ly α emissionis symmetric in the outer regions and shows a trend of in-creasing asymmetry as one moves towards the central regioni.e., close to the AGN. To get more information on the velocity distribution of theLy α emitting gas, we take spatial sections along the slit inthe SPSF subtracted Ly α spectra and measure the centerand FWHMs of the fitted Gaussian profiles for these sec-tions. The details of the fit used for each section are de-scribed in section 3.8 (see Fig. 12 “top” corresponds to Eastfor PA=72 ◦ and South for PA=350 ◦ ). It seems that thepeaks of Ly α emission on the top are clearly redshifted withrespect to the measured redshift of M1513-2524. Howeverthe bottom regions are only slightly redshifted.As the Ly α emission profile is sensitive to radiativetransfer effects in addition to the underlying velocity field, itwill be difficult to interpret the velocity field purely based onthe Ly α profile alone. However, it is also clear from the SPSFsubtracted spectra that the residual C iv emission we see inspectra obtained at PA = 72 ◦ is also consistent with the ex-tended emission at the top being redshifted (see Fig. 9) withrespect to the systemic redshift measured from the He ii line.But the residual emission seen in the bottom part is eitherconsistent with the systemic redshift or slightly blueshifted.This result is in line with what we see in the case of the Ly α profile. When we consider the 2D spectra taken at PA =350 ◦ , the residual C iv emission in the top is weak. However,we do see a trend similar to what we see for the Ly α line.Thus, there are signatures in the data indicating that the ex-tended emission to the southeast is redshifted with respect tothe systemic redshift. Given the fact that we are seeing blue shifted Ly α absorption, the flow in the surrounding regionsof M1513-2524 may be complex. We note that the FWHMis highest at the center ( ∼ − ) and decreases awayfrom the center (ranging from 800-1000 km s − ). Interpret-ing this sparsely sampled velocity field in terms of infall,outflow or rotation will be difficult. High spatial resolution( < (cid:48)(cid:48) ) IFS data are required to reveal further details on thevelocity field and spatial extent of the core and extendedspatial components. α emission As seen above, the alignment of C iv and Ly α with the radioaxis is indicated by our data. This agreement could indicategood alignment between the putative cone of ionizing radia-tion and the radio jets, and/or excitation due to interactionof radio emitting plasma with ambient gas. Broad emissionline widths may indicate possible interactions. In this sec-tion, to gain more insight, we will compare the propertiesof M1513-2524 (both optical and radio) with those of radiosources in the literature. For this comparison, we have usedresults from the literature (Roettgering et al. 1994; van Ojiket al. 1997; De Breuck et al. 2001; Jarvis et al. 2001; Willottet al. 2002; Bornancini et al. 2007; Saxena et al. 2019), witha view to understand the effect of radio jets on the ambi-ent medium. High redshift radio galaxies have shown strongdependence on Ly α size and velocity dispersion with theassociated radio sources (van Ojik et al. 1997). Typically,small Ly α halos are seen to be associated with smaller radiosources showing higher velocity dispersion. For low- z radiogalaxies (McCarthy et al. 1991), brightest regions of Ly α emission are seen to lie on the side of the radio lobe clos-est to galaxy nucleus. However, van Ojik et al. (1997) foundthat the brightest regions of Ly α emission are not always onthe side of the radio lobe closest to the nucleus, but in mostcases lie closest to the brightest lobe.From Fig. 6 of Saxena et al. (2019) we find that radiogalaxies with Ly α luminosity (cid:62) × erg s − typicallyhave FWHM velocities in excess of 1000 km s − . Thus, whatwe find for M1513-2524 is consistent with that of the samplestudied by Saxena et al. (2019). As we have shown above,the extended Ly α emission from M1513-2524 is symmetric(when we consider the 3 σ contours), and the radio lobes aremostly outside the diffuse Ly α emission for the SB sensitiv-ity we have achieved in our narrow band images (see Fig. 4).However, there are indications that the Ly α emission in thecentral region may be asymmetric and probably elongatedalong the radio axis. It is also seen that C iv emission is moreextended close to the radio axis. Thus, we do see indicationsfor connections between the diffuse spectral line emissionand radio emission. However, better spectroscopic data (interms of both spatial and spectral resolution) will be usefulto quantify these observations at a higher significance level.In Fig. 13 we plot the measured Ly α luminosity andlargest angular scale of the radio source as a function of red-shift for known z > α measurementsin the literature. It is clear from the left panel of this figurethat only three (TNJ0121+1320, TNJ0205+2242 and WN2313+4053) literature sources at z > α flux morethan what we measure for M1513-2524. All these sources areless extended in radio emission compared to M1513-2524. In MNRAS , 1–19 () y α emission from M151304.72-252439.70 PA 72
Ly (e) 1216
PSF
Ly (c) 1216Å
PSF
160 80 0 80 1600.00.51.0
PA 350
Ly (e) 1216
PSF
160 80 0 80 160
Ly (c) 1216Å
PSF N o r m a li z e d f l u x Distance along slit (kpc)
Figure 10.
Comparison of spatial profiles of Ly α with SPSFs extracted from long-slit spectra along PA = 72 ◦ and PA = 350 ◦ . The profilesare extracted from two regions: a “compact” spatial component (marked c, red) and “extended” spatial component (marked e, black)(see Fig. 8). Dashed cyan curves correspond to SPSFs from reference stars. The actual spatial profile for each emission line is shownas a dashed and the corresponding overall 2-component Gaussian fit is shown by a solid curve. In each subplot the orange dashed anddash-dotted lines are 2 Gaussian components fitted to the corresponding emission line. The black dotted horizontal line marks the zeroflux level. The vertical shift below zero level (especially seen in upper panels) is due to artifacts in the data. the right panel of Fig. 13, we plot the largest angular sizes ofradio sources as a function of z . Even in this plot, only two(TN J1123+3141 and TXS J2335-0002) literature sourceshave larger radio structures compared to M1513-2524. How-ever, the Ly α fluxes measured for these sources are at leasta factor 10 smaller than what we measure in M1513-2524.We note that the Ly α emission of the other high- z radiogalaxies was measured mostly using long-slit observations,complicating comparison with a Ly α luminosity of M1513-2524 estimated from narrow band imaging. Therefore, in theleft panel of Fig. 13, we have marked the Ly α luminosity ob-tained from narrow band image (red star), slitloss-correctedcombined 1D spectrum of both PAs (blue star) and com-bined spectrum before slitloss correction (green star). Thus,M1513-2524 indeed appears to be rare in terms of its radiosize and its Ly α luminosity. In Fig. 14, we have shown a comparison of Ly α luminos-ity and radio luminosity at 150 MHz for M1513-2524 withsample of radio sources from the literature. Jarvis et al.(2001) have shown a strong correlation between the Ly α emission line luminosity and radio luminosity measured at150-MHz (L ) for radio galaxies at z > .
75 (see also thefigure 8 of Saxena et al. 2019). From the left panel of Fig. 14 it is clear that M1513-2524 has a large 150 MHz luminos-ity compared to most high-z radio sources. In the middlepanel, we plot L vs. L Ly α . The measured L for M1513-2524 is 3.08 × W Hz − . For this value, the expected Ly α luminosity from the fit presented in Saxena et al. (2019) is ∼ × erg s − . The Ly α luminosity we observe is muchhigher than this best fit prediction, as also evident from themiddle panel. While there are 3 sources in the literaturesample with Ly α luminosity higher than what we find inM1513-2534, all of them have L less than what we mea-sure for M1513-2534.In the right panel of Fig. 14, we plot the projectedLy α size vs radio power at 150 MHz. M1513-2524 has oneof the largest Ly α extents among the strong radio sources(where the extents were mostly measured from long-slitspectroscopy observations). However, recent studies basedon radio weak quasars (Borisova et al. 2016; Arrigoni Bat-taia et al. 2019) have shown Ly α emission extending beyond100 kpc. These objects were observed with much better sur-face brightness sensitivities. However, a direct comparison ofM1513-2524 with these sources is rather difficult, as differentsets of data were obtained with different surface brightnesssensitivities. If the surface brightness profile of M1513-2524 MNRAS000
75 (see also thefigure 8 of Saxena et al. 2019). From the left panel of Fig. 14 it is clear that M1513-2524 has a large 150 MHz luminos-ity compared to most high-z radio sources. In the middlepanel, we plot L vs. L Ly α . The measured L for M1513-2524 is 3.08 × W Hz − . For this value, the expected Ly α luminosity from the fit presented in Saxena et al. (2019) is ∼ × erg s − . The Ly α luminosity we observe is muchhigher than this best fit prediction, as also evident from themiddle panel. While there are 3 sources in the literaturesample with Ly α luminosity higher than what we find inM1513-2534, all of them have L less than what we mea-sure for M1513-2534.In the right panel of Fig. 14, we plot the projectedLy α size vs radio power at 150 MHz. M1513-2524 has oneof the largest Ly α extents among the strong radio sources(where the extents were mostly measured from long-slitspectroscopy observations). However, recent studies basedon radio weak quasars (Borisova et al. 2016; Arrigoni Bat-taia et al. 2019) have shown Ly α emission extending beyond100 kpc. These objects were observed with much better sur-face brightness sensitivities. However, a direct comparison ofM1513-2524 with these sources is rather difficult, as differentsets of data were obtained with different surface brightnesssensitivities. If the surface brightness profile of M1513-2524 MNRAS000 , 1–19 () Shukla et.al
PA 72
CIV 1549Å
PSF
HeII 1640Å
PSF
160 80 0 80 1600.00.51.0
PA 350
CIV 1549Å
PSF
160 80 0 80 160
HeII 1640Å
PSF N o r m a li z e d f l u x Distance along slit (kpc)
Figure 11.
Comparison of spatial profiles of C iv and He ii with SPSFs from two PAs. For C iv and He ii the full emission regions wereused to extract spatial profiles. Dashed green and blue curves correspond to the observed C iv and He ii lines, respectively. is not very different from other high- z quasars one will ex-pect the Ly α extent to be much larger than what we mea-sure. Therefore, obtaining deep IFS spectra of M1513-2524will be important for ascertaining the true extent of the Ly α emission. i absorption Associated Ly α absorption is clearly evident in the Ly α emission line profile. However, we do not detect any asso-ciated C iv absorption in the profile of the C iv emissionline. As we discussed above, the extent of Ly α emission isnot uniform over the wavelength covered by emission (seeFig 8). It is clear from the figure that Ly α absorption oc-curs roughly in the wavelength range that demarcates the“compact” and “extended” spatial components. This basi-cally limits the spatial extent over which we can probe theLy α absorption. Under the assumption that the observedLy α emission profile is the result of an unabsorbed Gaus-sian profile modified by the presence of neutral hydrogen(H i ) along the line of sight, which can be characterised bya Voigt function (see van Ojik et al. 1997, hereafter VO97),we estimate the properties of the H i absorption. The initialinput parameters for the model are amplitude, redshift andwidth of the 1D Gaussian, and absorption redshift ( z abs ), Doppler parameter ( b ) and column density (log N (H i )) forthe Voigt profile.In upper left panel in Fig. 15, we compare the Ly α emis-sion profile extracted from the long-slit spectra obtainedalong PA = 72 ◦ (blue solid curve) and 350 ◦ (orange solidcurve) after scaling the slit-loss corrected spectrum of PA= 350 ◦ to match with PA = 72 ◦ at the location of the redpeak (shown by the thin dotted vertical line). While the pro-files match nicely in the red part and in the core, we do seedifferences in the blue wing. The difference could be due tospatial differences in the H i absorption.First we fit the emission profile along PA = 350 ◦ that was obtained during better seeing conditions. The ob-tained best fit values for log [ N (H i ) cm − ], z abs and b are16.62 ± ± ± N (H i ) is mainly dueto poor spectral resolution and a weak constraint on theemission line profile shape. The large b values compared tothe thermal broadening of photoionized gas could reflect ei-ther large velocity dispersion or high saturation of the Ly α absorption line that can appear unsaturated at the spectralresolution considered here. It is well known that the cover- MNRAS , 1–19 () y α emission from M151304.72-252439.70 top PA 72
PA 350 bottom
Relative velocity (km s ) f [ e r g s c m Å ] Figure 12. Ly α emission profiles extracted from the top and bottom spatial regions of the PSF subtracted images for two PAs (see Fig.9). The observed profiles are fitted with a 1D Gaussian+Voigt function. The black solid line is the observed profile, the red solid lineis the absorption model fit and the dashed line is the unabsorbed Gaussian profile. The left and right panels correspond to PA= 72 ◦ and PA= 350 ◦ , respectively. The gray shaded regions are 1 σ error bars. The zero velocity (blue dashed line) is defined with respect tothe emission redshift measured using He ii line of the combined 1D spectrum of both the PAs. The black dashed lines mark the peak ofunabsorbed Gaussian fit. Redshift l o g [ L L y ( e r g s )] Redshift L a r g e s t a n g u l a r s i z e ( ) Figure 13. Ly α line luminosity (left) and LAS (right) of the radioemission as a function of z for high- z radio sources from the lit-erature. Location of M1513-2524 is shown with a star symbol. Inthe left panel, the red star is the Ly α luminosity estimated fromthe narrow band image, the blue star is estimated from slitlosscorrected combined spectrum of both PAs, and the green star isfrom the slitloss-uncorrected spectrum. ing factor of the associated H i absorption need not be unity(see, e.g., Fathivavsari et al. 2018, for example). In that case,the actual N (H i ) may be much higher than we recover.We then fitted the spectrum obtained along PA =72 ◦ , assuming the same emission profile as for PA = 350 ◦ and excluding the regions leftwards of the blue peak. The obtained fit is shown by the red dotted curve, and thebest fit log [ N (H i ) cm − ], z abs and b are 18.99 ± ± ± − , respectively. It isclear from this figure that the spectrum obtained along PA= 72 ◦ requires additional absorption components relative tothe best fitted Ly α emission profile obtained for PA = 350 ◦ .If we do not assume the same emission profile for the twoPAs, then we can fit the Ly α absorption with log [ N (H i )cm − ], z abs and b equal to 15.15 ± − , 3.13032 ± ± ii and C iv . The profiles obtainedwith PA = 350 ◦ have excess flux in the blue wing, as sug-gested by the Ly α profile. Thus, allowing for variation in theemission profile along two PAs, results in similar absorptionstrength.To check if absorption strength varies spatially, we alsoextract profiles along the dispersion axis around the centerusing spatial apertures of varying width, as shown in thelower left panel of Fig. 15. Each profile has been scaled tomatch at the location of the blue peak (dotted vertical line).The figure shows that the absorption strength does not varymuch with the spatial scale over which the spectra are ex-tracted. We also see no significant change in the Ly α equiv-alent width with an increase in the width of the apertureover which spectrum is extracted. This means the absorbingregion may be extended as much as the “compact spatial”component studied in section 3.5. However, any small spa-tial variations (over 2 (cid:48)(cid:48) scale) in the absorption profile mighthave gotten smoothed by the seeing smearing.We further examine the spectral profiles extracted from MNRAS000
Relative velocity (km s ) f [ e r g s c m Å ] Figure 12. Ly α emission profiles extracted from the top and bottom spatial regions of the PSF subtracted images for two PAs (see Fig.9). The observed profiles are fitted with a 1D Gaussian+Voigt function. The black solid line is the observed profile, the red solid lineis the absorption model fit and the dashed line is the unabsorbed Gaussian profile. The left and right panels correspond to PA= 72 ◦ and PA= 350 ◦ , respectively. The gray shaded regions are 1 σ error bars. The zero velocity (blue dashed line) is defined with respect tothe emission redshift measured using He ii line of the combined 1D spectrum of both the PAs. The black dashed lines mark the peak ofunabsorbed Gaussian fit. Redshift l o g [ L L y ( e r g s )] Redshift L a r g e s t a n g u l a r s i z e ( ) Figure 13. Ly α line luminosity (left) and LAS (right) of the radioemission as a function of z for high- z radio sources from the lit-erature. Location of M1513-2524 is shown with a star symbol. Inthe left panel, the red star is the Ly α luminosity estimated fromthe narrow band image, the blue star is estimated from slitlosscorrected combined spectrum of both PAs, and the green star isfrom the slitloss-uncorrected spectrum. ing factor of the associated H i absorption need not be unity(see, e.g., Fathivavsari et al. 2018, for example). In that case,the actual N (H i ) may be much higher than we recover.We then fitted the spectrum obtained along PA =72 ◦ , assuming the same emission profile as for PA = 350 ◦ and excluding the regions leftwards of the blue peak. The obtained fit is shown by the red dotted curve, and thebest fit log [ N (H i ) cm − ], z abs and b are 18.99 ± ± ± − , respectively. It isclear from this figure that the spectrum obtained along PA= 72 ◦ requires additional absorption components relative tothe best fitted Ly α emission profile obtained for PA = 350 ◦ .If we do not assume the same emission profile for the twoPAs, then we can fit the Ly α absorption with log [ N (H i )cm − ], z abs and b equal to 15.15 ± − , 3.13032 ± ± ii and C iv . The profiles obtainedwith PA = 350 ◦ have excess flux in the blue wing, as sug-gested by the Ly α profile. Thus, allowing for variation in theemission profile along two PAs, results in similar absorptionstrength.To check if absorption strength varies spatially, we alsoextract profiles along the dispersion axis around the centerusing spatial apertures of varying width, as shown in thelower left panel of Fig. 15. Each profile has been scaled tomatch at the location of the blue peak (dotted vertical line).The figure shows that the absorption strength does not varymuch with the spatial scale over which the spectra are ex-tracted. We also see no significant change in the Ly α equiv-alent width with an increase in the width of the apertureover which spectrum is extracted. This means the absorbingregion may be extended as much as the “compact spatial”component studied in section 3.5. However, any small spa-tial variations (over 2 (cid:48)(cid:48) scale) in the absorption profile mighthave gotten smoothed by the seeing smearing.We further examine the spectral profiles extracted from MNRAS000 , 1–19 () Shukla et.al
Redshift L o g [ L ( W H z )]
42 43 44 45 46 log [L Ly (erg s )]
20 40 60 80 100
Projected Ly size (kpc)]
Figure 14.
Radio luminosity at 150 MHz is shown as a function of redshift, Ly α luminosity and projected Ly α size for M1513-2524 anda sample of faint HzRGs (blue points, Saxena et al. (2019)) and normal HzRGs (black points). In the rightmost panel, only data samplefrom van Ojik et al. (1997) is shown, since radio source size information were not readily available for other sources. the “top” and “bottom” regions shown in Fig. 12 to probefor the presence of extended absorption. Presence of weakabsorption could be seen in these spectra as well, but con-straining the column density is rather difficult due to poorSNR and spectral resolution. However, these observationsare consistent with the spatially extended Ly α absorbing re-gion. Note in all these discussions, we assumed that the dipseen in Ly α emission profile is due to absorption. The casewould be strengthened if associated C iv absorption were de-tected using high SNR data, or if the Ly α line profile wereresolved at higher spectral resolution. Such evidence wouldrule out the possibility that some part or all of the Ly α pro-file is being influenced by the radiative transfer effects thatcan generate complex profiles.HzRGs are seen to show extended absorption up to ∼ − . cm − as measured from low resolutionspectra (in some cases the actual column density is found tobe lower when high resolution spectrum is used Jarvis et al.2003). For M1513-2524, we also see similar trend, howeverthe column density for this source is lower than what hasbeen found by VO97 using low-resolution data like ours. Theabsorber is blueshifted by ∼ − with respect tothe peak of the unabsorbed Gaussian profile of Ly α emissionalong two PAs. In van Ojik et al. (1997), 60% of the sam-ple showed H i absorption and most of the absorbers wereblue shifted within ∼
250 km s − of Ly α peak, which impliedthat these absorption were not likely to be caused by gaseoushalos of neighbouring galaxies or by tidal remnants of inter-action with nearby galaxies, but more likely associated withthe intrinsic properties of HzRGs. We see a similar blueshiftfor M1513-2524 as well. However, VO97 found that most ofthe H i absorption is found towards galaxies with smaller ra-dio sizes, and only 25% of sources with >
50 kpc radio sizesshowed absorption with only one source beyond 128 kpc ra-dio size showing absorption (unlike in this case). Most ofthe properties of the absorber around M1513-2524 are con- sistent with VO97; however, M1513-2524 is the first of itskind in showing the presence of H i absorption around a ra-dio source extended by up to 184 kpc. We also obtained theuGMRT radio spectrum to detect presence of H i i In this paper, we have presented observations of M1513-2524, an optically faint source showing a very extended Ly α nebula. For our study we have used long-slit spectra takenat two PAs along with narrow band imaging (both usingRSS/SALT) and uGMRT radio maps at 1360 MHz and 420MHz. Based on our analysis, together with existing archivaldata, we arrive at the following interesting results:(i) M1513-2524 is an extremely faint optical continuumsource, not detected in any bands of the PS1 survey. Theonly emission lines that are detected are Ly α , N v , C iv andHe ii . We used the non-resonant He ii line to estimate theredshift, z em =3.1320 ± α luminosity of the nebula is(6.80 ± × erg s − , which is one among the highestluminosities detected so far. The Ly α nebula extends out to ∼
90 kpc for the 3 σ flux contour (1 . × − erg s − cm − )in the smoothed narrow band image. Based on the similarityof line ratios with those of “Non-Ly α ” excess objects, we ar-gue that the Ly α emission (in particular from the compactregion) in M1513-2524 is most likely powered by photoion-ization from the central AGN.(iii) The 2D spectrum of the source clearly shows an asso-ciated Ly α absorption demarcating the two spatial compo-nents, a “compact” and an “extended” component. The C iv and He ii profiles are significantly extended along PA =72 ◦ ;however, the “compact” component does not show clear ex-tension along this PA. The PA=350 ◦ spectrum does not MNRAS , 1–19 () y α emission from M151304.72-252439.70 PA 072 PA 350
C IV
PA 72PA 350 He II
PA 72PA 350
Observed wavelength (Å) f [ e r g s c m Å ] Figure 15.
Top left panel:
Comparison of Ly α velocity profile at 2 PAs is shown after scaling the spectra to match at the peak of narrowLy α component (at thin vertical dotted line). The blue and orange solid curves correspond to spectra obtained at PA = 72 ◦ and 350 ◦ ,respectively and the shaded region shows 1 σ region. The corresponding fits to the observed profiles are shown in dash-dotted red color(PA = 72 ◦ ) and dotted red curve (PA = 350 ◦ ). The Gaussian emission line profile obtained from our chosen model (Gaussian plusVoigt profiles) at PA = 350 ◦ is shown in red solid curve. For PA = 72 ◦ same Gaussian profile is used with varying the column density,doppler parameter and absorption redshift. The thick vertical dotted line marks the wavelength corresponding to absorption redshiftand the dash-dotted line marks the wavelength correspondng to the emission redshift obtained using He ii line. Bottom left:
The Ly α profiles obtained from spectrum at PA = 350 ◦ using different spatial apertures with varying width after scaling the fluxes to match at thelocation of the dotted vertical line are shown. Right panel: C iv (top) and He ii (bottom) profiles obtained along two PAs are compared. show significantly extended C iv and He ii emission; howeverit shows presence of asymmetry in the “compact” componentextending to ∼
19 kpc. All these results emphasize the needfor high SNR and high spatial resolution spectroscopy ofM1513-2524.(iv) We detect Ly α absorption on top of the Ly α emis-sion blue shifted by 250-400 km s − with respect to the peakof Ly α emission. Based on spectra taken along two positionangles, we suggest that absorption may be extended. How-ever, only high spatial resolution spectra would allow us toprobe the spatial variations of the Ly α absorption on top ofthe possible variations in the Ly α emission line profile.(v) The high spatial resolution uGMRT radio maps con-firm the presence of two radio lobes (separted by ∼
184 kpc)associated with the source, and a weak core at the location ofthe WISE source at a 4 σ significance level. The Ly α nebulais contained within the radio structure, with its outer re-gions more symmetric ( α = 0 .
84) than its inner core region( α = 0 .
79) (see, e.g, Sec. 3.5). We find that the major axisof the isophotes for the core region is preferentially alignedclose to the radio axis.(vi) Kinematic analysis of the Ly α emission shows that the central core region is more perturbed, with FWHM ∼ − , whereas the outer regions have FWHM ∼ − . Emission from the outer regions is redshiftedwith respect to the systemic redshift. It will be interesting toconfirm the existence of these two distinct regions using IFSspectroscopy. Such high spatial resolution spectra togetherwith high signal to noise radio images in the core regionswill allow us to study any interaction between radio jet (orinner parts of the lobes) and the Ly α emitting halo gas.(vii) We identify seven candidate Ly α emitters aroundM1513-2524 having Ly α luminosity greater than or equal tothat of L* galaxies, while 0.2 galaxies are expected based onthe observed Ly α luminosity function. If these candidatesare confirmed in followup spectroscopic observations, it willreveal that M1513-2524 may be part of a large proto-clusterof galaxies forming stars at a high rate.In short, M1513-2524 has several properties similar to thoseof previously detected nebulae around HzRGs. However,there are several attributes that are rather unique to thissource. First, both the radio power and Ly α luminosity ofM1513-2524 are among the highest known. Even more inter- MNRAS000
79) (see, e.g, Sec. 3.5). We find that the major axisof the isophotes for the core region is preferentially alignedclose to the radio axis.(vi) Kinematic analysis of the Ly α emission shows that the central core region is more perturbed, with FWHM ∼ − , whereas the outer regions have FWHM ∼ − . Emission from the outer regions is redshiftedwith respect to the systemic redshift. It will be interesting toconfirm the existence of these two distinct regions using IFSspectroscopy. Such high spatial resolution spectra togetherwith high signal to noise radio images in the core regionswill allow us to study any interaction between radio jet (orinner parts of the lobes) and the Ly α emitting halo gas.(vii) We identify seven candidate Ly α emitters aroundM1513-2524 having Ly α luminosity greater than or equal tothat of L* galaxies, while 0.2 galaxies are expected based onthe observed Ly α luminosity function. If these candidatesare confirmed in followup spectroscopic observations, it willreveal that M1513-2524 may be part of a large proto-clusterof galaxies forming stars at a high rate.In short, M1513-2524 has several properties similar to thoseof previously detected nebulae around HzRGs. However,there are several attributes that are rather unique to thissource. First, both the radio power and Ly α luminosity ofM1513-2524 are among the highest known. Even more inter- MNRAS000 , 1–19 () Shukla et.al esting is the detection of associated Ly α absorption arounda very large radio source, a configuration that is usuallyvery rare. We have also identified seven potential Ly α emit-ters around this radio source, which if confirmed could makeM1513-2524 part of a proto-cluster of star forming galaxies.All these properties make M1513-2524 an ideal target forfuture integral field spectroscopic studies. ACKNOWLEDGMENTS
We thank Dr. Montserrat Villar Martin for a very construc-tive report that has improved the presentation of this work.Most of the observations reported in this paper were ob-tained with the Southern African Large Telescope (SALT).We thank the staff of the GMRT for wide band observations.GMRT is run by the National Centre for Radio Astrophysicsof the Tata Institute of Fundamental Research. This workutilized the open source software packages
Astropy (As-tropy Collaboration et al. 2018),
Numpy (van der Walt et al.2011),
Scipy (Virtanen et al. 2020),
Matplotlib (Hunter2007) and
Ipython (P´erez & Granger 2007).
DATA AVAILABILITY
Data used in this work are obtained using SALT. Raw datawill become available for public use 1.5 years after the ob-serving date at https://ssda.saao.ac.za/.
REFERENCES
Allen M. G., Groves B. A., Dopita M. A., Sutherland R. S., Kew-ley L. J., 2008, ApJS, 178, 20Ao Y., et al., 2017, ApJ, 850, 178Arrigoni Battaia F., Prochaska J. X., Hennawi J. F., Obreja A.,Buck T., Cantalupo S., Dutton A. A., Macci`o A. V., 2018,MNRAS, 473, 3907Arrigoni Battaia F., Hennawi J. F., Prochaska J. X., O˜norbe J.,Farina E. P., Cantalupo S., Lusso E., 2019, MNRAS, 482,3162Astropy Collaboration et al., 2013, A&A, 558, A33Astropy Collaboration et al., 2018, AJ, 156, 123Bacon R., et al., 2010, in Proc. SPIE. p. 773508,doi:10.1117/12.856027Borisova E., et al., 2016, ApJ, 831, 39Bornancini C. G., De Breuck C., de Vries W., Croft S., vanBreugel W., R¨ottgering H., Minniti D., 2007, MNRAS, 378,551Bridge C. R., et al., 2013, ApJ, 769, 91Buckley D. A. H., Charles P. A., Nordsieck K. H., O’DonoghueD., 2006, in Whitelock P., Dennefeld M., LeibundgutB., eds, IAU Symposium Vol. 232, The Scientific Re-quirements for Extremely Large Telescopes. pp 1–12,doi:10.1017/S1743921306000202Burgh E. B., Nordsieck K. H., Kobulnicky H. A., Williams T. B.,O’Donoghue D., Smith M. P., Percival J. W., 2003, in IyeM., Moorwood A. F. M., eds, Society of Photo-Optical In-strumentation Engineers (SPIE) Conference Series Vol. 4841,Proc. SPIE. pp 1463–1471, doi:10.1117/12.460312Cai Z., et al., 2017, ApJ, 837, 71Cai Z., et al., 2018, ApJ, 861, L3Cantalupo S., Porciani C., Lilly S. J., Miniati F., 2005, ApJ, 628,61 Cantalupo S., Arrigoni-Battaia F., Prochaska J. X., HennawiJ. F., Madau P., 2014, Nature, 506, 63Chambers K., Pan-STARRS Team 2018, in American Astronom-ical Society Meeting Abstracts arXiv:1708.07371 )Gupta N., et al., 2020, arXiv e-prints, p. arXiv:2007.04347Haiman Z., Spaans M., Quataert E., 2000, ApJ, 537, L5Heckman T. M., Lehnert M. D., van Breugel W., Miley G. K.,1991a, ApJ, 370, 78Heckman T. M., Lehnert M. D., Miley G. K., van Breugel W.,1991b, ApJ, 381, 373Hennawi J. F., Prochaska J. X., Cantalupo S., Arrigoni-BattaiaF., 2015, Science, 348, 779Hu E. M., Cowie L. L., 1987, ApJ, 317, L7Hunter J. D., 2007, Computing in Science Engineering, 9, 90Jarvis M. J., Rawlings S., Eales S., Blundell K. M., Bunker A. J.,Croft S., McLure R. J., Willott C. J., 2001, MNRAS, 326,1585Jarvis M. J., Wilman R. J., R¨ottgering H. J. A., Binette L., 2003,MNRAS, 338, 263Kobulnicky H. A., Nordsieck K. H., Burgh E. B., Smith M. P.,Percival J. W., Williams T. B., O’Donoghue D., 2003, in IyeM., Moorwood A. F. M., eds, Society of Photo-Optical In-strumentation Engineers (SPIE) Conference Series Vol. 4841,Proc. SPIE. pp 1634–1644, doi:10.1117/12.460315Krogager J. K., et al., 2018, ApJS, 235, 10Large M. I., Mills B. Y., Little A. G., Crawford D. F., SuttonJ. M., 1981, MNRAS, 194, 693Marques-Chaves R., et al., 2019, A&A, 629, A23Matsuda Y., et al., 2004, AJ, 128, 569Mayo J. H., Vernet J., De Breuck C., Galametz A., Seymour N.,Stern D., 2012, A&A, 539, A33McCarthy P. J., 1993, ARA&A, 31, 639McCarthy P. J., van Breugel W., Kapahi V. K., 1991, ApJ, 371,478Morais S. G., et al., 2017, MNRAS, 465, 2698Mori M., Umemura M., Ferrara A., 2004, ApJ, 613, L97Morrissey P., et al., 2012, in Proc. SPIE. p. 844613,doi:10.1117/12.924729Nilsson K. K., Fynbo J. P. U., Møller P., Sommer-Larsen J.,Ledoux C., 2006, A&A, 452, L23Ouchi M., et al., 2008, ApJS, 176, 301Overzier R. A., Nesvadba N. P. H., Dijkstra M., Hatch N. A.,Lehnert M. D., Villar-Mart´ın M., Wilman R. J., Zirm A. W.,2013, ApJ, 771, 89 MNRAS , 1–19 () y α emission from M151304.72-252439.70 P´erez F., Granger B. E., 2007, Computing in Science and Engi-neering, 9, 21Planck Collaboration et al., 2018, arXiv e-prints, p.arXiv:1807.06209Prescott M. K. M., Momcheva I., Brammer G. B., Fynbo J. P. U.,Møller P., 2015, ApJ, 802, 32Roettgering H. J. A., Lacy M., Miley G. K., Chambers K. C.,Saunders R., 1994, A&AS, 108, 79Rosdahl J., Blaizot J., 2012, MNRAS, 423, 344Saxena A., et al., 2019, MNRAS, 489, 5053Shen Y., et al., 2016, ApJ, 831, 7Vernet J., Fosbury R. A. E., Villar-Mart´ın M., Cohen M. H.,Cimatti A., di Serego Alighieri S., Goodrich R. W., 2001,A&A, 366, 7Villar-Martin M., Binette L., Fosbury R. A. E., 1996, A&A, 312,751Villar-Mart´ın M., Vernet J., di Serego Alighieri S., Fosbury R.,Humphrey A., Pentericci L., 2003, MNRAS, 346, 273Villar-Mart´ın M., Humphrey A., De Breuck C., Fosbury R., Bi-nette L., Vernet J., 2007a, MNRAS, 375, 1299Villar-Mart´ın M., S´anchez S. F., Humphrey A., Dijkstra M.,di Serego Alighieri S., De Breuck C., Gonz´alez Delgado R.,2007b, MNRAS, 378, 416Virtanen P., et al., 2020, Nature Methods,Willott C. J., Rawlings S., Blundell K. M., Lacy M., Hill G. J.,Scott S. E., 2002, MNRAS, 335, 1120Wisotzki L., et al., 2018, Nature, 562, 229Wylezalek D., et al., 2013, ApJ, 769, 79van Dokkum P. G., 2001, PASP, 113, 1420van Ojik R., Roettgering H. J. A., Miley G. K., Hunstead R. W.,1997, A&A, 317, 358van der Walt S., Colbert S. C., Varoquaux G., 2011, Computingin Science and Engineering, 13, 22This paper has been typeset from a TEX/L A TEX file prepared bythe author.MNRAS000