The Discovery of A Luminous Broad Absorption Line Quasar at A Redshift of 7.02
Feige Wang, Jinyi Yang, Xiaohui Fan, Minghao Yue, Xue-Bing Wu, Jan-Torge Schindler, Fuyan Bian, Jiang-Tao Li, Emanuele P. Farina, Eduardo Bañados, Frederick B. Davies, Roberto Decarli, Richard Green, Linhua Jiang, Joseph F. Hennawi, Yun-Hsin Huang, Chiara Mazzucchelli, Ian D. McGreer, Bram Venemans, Fabian Walter, Yuri Beletsky
DDraft version November 30, 2018
Typeset using L A TEX preprint2 style in AASTeX62
The Discovery of A Luminous Broad Absorption Line Quasar at A Redshift of 7.02
Feige Wang,
1, 2, 3
Jinyi Yang, Xiaohui Fan, Minghao Yue, Xue-Bing Wu,
3, 4
Jan-Torge Schindler, Fuyan Bian, Jiang-Tao Li, Emanuele P. Farina, Eduardo Ba˜nados, Frederick B. Davies, Roberto Decarli, Richard Green, Linhua Jiang, Joseph F. Hennawi,
1, 9
Yun-Hsin Huang, Chiara Mazzucchelli, Ian D. McGreer, Bram Venemans, Fabian Walter, and Yuri Beletsky Department of Physics, University of California, Santa Barbara, CA 93106-9530, USA Steward Observatory, University of Arizona, 933 North Cherry Avenue, Tucson, AZ 85721, USA Kavli Institute for Astronomy and Astrophysics, Peking University, Beijing 100871, China Department of Astronomy, School of Physics, Peking University, Beijing 100871, China European Southern Observatory, Alonso de C´ordova 3107, Casilla 19001, Vitacura, Santiago 19, Chile Department of Astronomy, University of Michigan, 311 West Hall, 1085 South University Avenue, Ann Arbor, MI48109-1107, USA The Observatories of the Carnegie Institution for Science, 813 Santa Barbara Street, Pasadena, CA 91101, USA INAF–Osservatorio di Astrofisica e Scienza dello Spazio, via Gobetti 93/3, I-40129, Bologna, Italy Max Planck Institut f¨ur Astronomie, K¨onigstuhl 17, D-69117, Heidelberg, Germany Las Campanas Observatory, Carnegie Institution of Washington, Colina el Pino, Casilla 601, La Serena, Chile
ABSTRACTDespite extensive efforts, to date only two quasars have been found at z >
7, dueto a combination of low spatial density and high contamination from more ubiquitousGalactic cool dwarfs in quasar selection. This limits our current knowledge of the super-massive black hole (SMBH) growth mechanism and reionization history. In this Letter,we report the discovery of a luminous quasar at z = 7 . Wide-field InfraredSurvey Explore ( WISE ) mid-infrared all-sky survey. With an absolute magnitude of M =–27.1 and bolometric luminosity of L Bol =5.6 × L (cid:12) , J0038–1527 is the mostluminous quasar known at z >
7. Deep optical to near-infrared spectroscopic observa-tions suggest that J0038–1527 hosts a 1.3 billion solar mass black hole accreting at theEddington limit, with an Eddington ratio of 1.25 ± iv broad emission lineof J0038–1527 is blueshifted by more than 3000 km s − relative to the quasar systemicredshift. More detailed investigations of the high-quality spectra reveal three extremelyhigh-velocity C iv broad absorption lines (BALs) with velocity from 0.08 to 0.14 timesthe speed of light and total “balnicity” index of more than 5000 km s − , suggesting thepresence of relativistic outflows. J0038–1527 is the first quasar found at the epoch ofreionization (EoR) with such strong outflows, and therefore provides a unique labora- Corresponding author: Feige [email protected] a r X i v : . [ a s t r o - ph . GA ] N ov Wang et al. tory to investigate active galactic nuclei feedback on the formation and growth of themost massive galaxies in the early universe.
Keywords: galaxies: active — galaxies: high-redshift — quasars: individual (J0038–1527) — cosmology: observations — early universe INTRODUCTIONAs the most luminous non-transient objects,distant quasars are important tracers to studyearly structure formation and the history ofcosmic reionization. The detections of com-plete Gunn–Peterson (GP) absorption troughsin z > z (cid:38) α damping wing absorption profile(Miralda-Escud´e 1998) probes neutral inter-galactic medium (IGM) gas at the epoch ofreionization (EoR). Currently, only two lumi-nous z > z > M BH - σ ∗ relation; Gebhardt et al. 2000). Feedback bywind or outflow driven by black hole (BH) ac-cretion has been invoked in simulations to ex-plain the observed relation (e.g., King 2003; DiMatteo et al. 2005). Observationally, strongoutflows are often studied in the rest-frame ul-traviolet (UV) via blueshifted broad absorptionlines (BALs; Weymann et al. 1991) or strongblue velocity shifts of broad C iv emission lines in luminous quasars (e.g., Richards et al. 2011).These features appear most often at a moder-ate velocity of v < . c ), but relativistic BALs at v ∼ . − . c havebeen also found in a small number of quasars(e.g., Hamann et al. 2013; Rogerson et al. 2016;Hamann et al. 2018). The associated kineticpower of these relativistic outflows is estimatedto be high enough to play a key role in theco-evolution of SMBHs and galaxies (e.g., Ci-cone et al. 2015; Feruglio et al. 2017), as it in-evitably shocks against star formation and pro-vide significant metal enrichment to the inter-stellar medium (ISM) and IGM (e.g., Chartaset al. 2009; Zubovas & King 2013).However, to date only two z > z > Legacy Imaging Surveys (DELS; Deyet al. 2018) with near-infrared (NIR) surveyslike the UKIRT Hemisphere Survey (UHS; Dyeet al. 2018), and the
Wide-field Infrared Sur-vey Explore mid-infrared survey (Wright et al.2010) allows us to search for the highest redshiftquasars over a much large area than previousstudies. Recently, we also updated our selec-tion procedure by including the Pan-STARRS1(PS1) Survey (Chambers et al. 2016) to improveour selection efficiency (Wang et al. 2018). http://desi.lbl.gov/ luminous BAL Quasar at z = 7 .
02 3Here, we report the discovery of a luminous z > §
2, we present our spec-troscopic observations and infrared photomet-ric observations. In §
3, we describe the lumi-nosity and BH mass measurements. In §
4, wecharacterize the strong relativistic outflows de-tected in this quasar. Finally, in § H = 70 km s − Mpc − and density parameters of Ω m = 0 . Λ = 0 . OBSERVATIONS AND DATA ANALYSISJ0038–1527 was selected as a z > . z -band( z AB = 21.65 ± y -band ( y AB =20.61 ± g , r , i ,and z bands. The strong dropout nature makesit is a promising high-redshift quasar candi-date. J0038–1527 is also detected in ALLWISE( W VEGA = 16.80 ± W VEGA = 16.08 ± − grating and 1 . (cid:48)(cid:48)
25 slit, whichprovide a spectral resolution of R ∼ ∼ z ∼
7. Subsequent highsignal-to-noise ratio (S/N) spectra taken withthe MMT and Magellan/LDSS3-C confirm it asa quasar at z >
7. We were allocated Very Large Telescope(VLT)/X-SHOOTER DD time (program ID:2100.A-5033(A)) and the observations were ob-tained over four nights between 2018 Januaryand July. The total on-source exposure was12,000s. We used 0 . (cid:48)(cid:48) R ∼ R ∼ . (cid:48)(cid:48) R ∼
750 from ∼ µ m to ∼ µ m. The GNIRS data was re-duced with the XIDL suite of astronomical rou-tines in the Interactive Data Language (IDL).Finally, we produced the combined spectrumusing the X-SHOOTER and GNIRS observa-tions and scaled it to match the photometricdata for absolute flux calibration. Then we cor-rect the Galactic extinction using the Cardelliet al. (1989) reddening law and E ( B − V ) fromSchlegel et al. (1998). The final calibrated op-tical to NIR spectrum is shown in Figure 1.In order to constrain the rest-frame UVspectral energy distribution (SED) of J0038–1527, we obtained Y, J, H, and K-band pho-tometry using UKIRT/WFCam (Project ID:U/17B/D04) on 2018 January 20. The on-source exposures were 12 mins in the Y, J, andK-bands and 6 minutes in the H-band. Theprocessed data was kindly provided by M. Irwinusing the standard Visible and Infrared SurveyTelescope for Astronomy (VISTA)/WFCAMdata-flow system (Irwin et al. 2004). The pho-tometric properties and derived parameters ofJ0038–1527 are listed in Table 1. https://github.com/profxj/xidl Wang et al.
Observed Wavelength ( Å ) f ( e r g s c m Å ) Rest-frame Wavelength ( Å ) | Ly | N V | Si IV | C IV | C III] | Mg II f Mg II (b) f C IV (a)
Figure 1.
Final calibrated spectrum of J0038–1527. The black and gray lines represent the Galactic extinc-tion corrected spectrum and error vector. The thin magenta line denotes the quasar composite spectrumconstructed with ∼
200 SDSS quasars with large C iv blueshifts. The green dashed line denotes the pseudo-continuum which includes power-law, Fe ii emission, and Balmer continuum . The orange circles are fluxdensity converted from galactic extinction corrected magnitudes listed in Table 1. The x-axis error-bars ofthe two leftmost orange circles denote the FWHMs of z DELS and y ps1 filter curves. Two inner plots showspectral fitting of C iv (a) and Mg ii (b) regions, respectively. The blue dot-dashed line denotes the best-fitpower-law continuum, the green dashed line denotes the best-fit pseudo-continuum. The cyan line denotesfitted C iv and Mg ii emission lines plus power-law continuum and the red line denotes the total fitted flux.3. LUMINOSITY AND BH MASSWe fit the final calibrated spectrum fol-lowing the approach detailed in Wang et al.(2015). Briefly, we shift the spectrum torest frame using an initial redshift, and fit apseudo-continuum model that includes a power-law continuum, Fe ii emission (Vestergaard &Wilkes 2001; Tsuzuki et al. 2006), and Balmercontinuum (e.g., De Rosa et al. 2014) to theline-free regions using a 1 /σ weighting χ fit-ting technique. We then fit the Mg ii emissionand derive the quasar systemic redshift by cor-recting Mg ii redshift to [O iii ] redshift basedon the velocity offset between these two linesof the SDSS quasar composite (Vanden Berket al. 2001). We iterated this procedure un-til the difference of input redshift and outputredshift is smaller than the uncertainty. Thenthe spectrum is de-redshifted using the finalsystemic redshift, which is z = 7 . ± . f λ ∝ λ − . ± . . We measure therest-frame 3000 ˚A power-law luminosity to be Redshift −29−28−27−26−25−24−23 M Å J0038-1527Other z > 6.3 QSOsPSOJ323+12Other z > 7 QSOs
Figure 2.
Absolute magnitude of all pub-licly known z > . M ,AB =–27.10, is the most luminous z > . z = 6 . z > M ,AB of publicly known quasars.The cyan dashed line mark the position of z = 6 . luminous BAL Quasar at z = 7 .
02 5
Table 1.
Photometric Properties andDerived Parameters of J0038–1527.
R.A. (J2000) 00:38:36.10Decl. (J2000) –15:27:23.6Redshift 7.021 ± m ± M -27.10 ± z DELS , AB ± y ps1 , AB ± Y VEGA ± J VEGA ± H VEGA ± K VEGA ± W VEGA ± W VEGA ± g ps1 , r ps1 , i ps1 , z ps1a > . , . , . , . z MgII ± z CIV ± α λ –1.54 ± v CIV − v MgII (km s − ) 3400 ± MgII (km s − ) 2994 ± MgII (˚A) 16.5 ± CIV (km s − ) 8728 ± CIV (˚A) 18.1 ± λL (erg s − ) 4.19 × L Bol (erg s − ) 2.16 × M BH (M (cid:12) ) (1.33 ± × L Bol /L Edd ± a Magnitude limits at 3- σ level. λL =4.19 × erg s − , and the rest-frame1450 ˚A magnitude to be m ,AB =19 . ± . M ,AB =–27.10 ± z > z > . M ,AB < −
27 (Figure2). By assuming an empirical conversion factorfrom the luminosity at 3000 ˚A (e.g., Shen et al.2011), we estimate the bolometric luminosity ofJ0038–1527 as L bol =5.15 × λL = 2.16 × erg s − = 5.6 × L (cid:12) .After subtracting the best-fit pseudo-continuumfrom the spectrum, we fit Mg ii and C iv broademission lines with two Gaussian profiles foreach line. We measure a full-width at half-maximum (FWHM) of 2994 ±
140 km s − and8728 ±
452 km s − for Mg ii and C iv , respec- tively. The continuum and line fitting are shownin Figure 1. After applying a virial BH massestimator (Vestergaard & Osmer 2009) basedon the Mg ii line, we estimate the SMBH massof J0038–1527 to be (1.33 ± × M (cid:12) . Theaccretion rate of this quasar is consistent withEddington accretion, with an Eddington ra-tio of L Bol /L Edd =1.25 ± ± × M (cid:12) and L Bol /L Edd =1.37. The quoted uncertainty doesnot include the systematic uncertainties in thescaling relation, which could be up to ∼ . z > (cid:15) ∼ QUASAR OUTFLOWSAs described in §
1, quasar outflows arethought to play an important role in regulat-ing the co-evolution of central SMBHs and hostgalaxies. Both broad emission lines like C iv and highly ionized absorption lines can be effi-cient diagnostics of quasar outflows. Especially,high-ionization broad emission lines like C iv of quasars are usually blueshifted from the sys-temic velocity by several hundred km s − and upto ∼ − , depending on the equivalentwidth (EW) of C iv and the quasar luminosity(e.g., Richards et al. 2011, also Figure 3). Theanti-correlation between continuum luminosityand emission line EW is well known as the Bald-win Effect (Baldwin 1977). More recently, re-sults from the SDSS quasar reverberation map-ping project (Sun et al. 2018) show that ex- Wang et al. −2000 0 2000 4000 6000 8000
CIV Blueshift [kms −1 ] L o g ( E W C I V )[ Å ] J1120+0641J0109−3047J0305−3150 J1342+0928 L o g L B o l [ e r g s − ] Figure 3. C iv emission line blueshift vs. C iv EW and quasar bolometric luminosity. The con-tours and 2D histogram are for SDSS low-redshiftquasars (Shen et al. 2011) and the black asterisksdenote public known z (cid:38) . iv ob-servations (e.g., Mortlock et al. 2011; De Rosa etal. 2014; Mazzucchelli et al. 2017; Ba˜nados et al.2018). The orange asterisk represents J0038–1527.Although all z (cid:38) . iv blueshift tail, they are still consistent with the dis-tribution of low-redshift quasars of similarly highluminosities. treme blueshift quasars have a low level of vari-ability, suggesting that high-blueshift sourcestend to also have high Eddington ratios. Fromour analysis of J0038–1527, we find that thepeak of the C iv emission has a strong blueshiftwith velocity of 3400 ±
411 km s − compared toMg ii . We measured the rest-frame EWs to beEW CIV =18 . ± . MgII =16 . ± . iv emission line. In order tofurther confirm this, we used a more robust wayto measure the C iv blueshift explored by Coat-man et al. (2016, 2017), which measures theline centroid as the wavelength that bisects thecumulative line flux. This method yields a C iv blueshift of 3800 km s − , confirms the strongblueshifted C iv emission line in J0038–1527. More interestingly, after examining the spec-trum of J0038–1527, we find several strong BALfeatures blueward of the C iv and Si iv emissionlines. No absorption feature is found bluewardof Mg ii indicating that it is a high-ionizationbroad absorption line quasar (HiBAL). In or-der to estimate the intrinsic spectrum of J0038–1527, we construct a quasar composite spec-trum (Figure 1) using ∼
200 SDSS quasars withextreme C iv emission line blueshifts ( > − ). We then divide the spectrum ofJ0038–1527 by the matched composite to de-rive a normalized spectrum. From the nor-malized spectrum, we identified three C iv ab-sorption troughs at extremely high velocities of(0 . +0 . − . ) c (trough A), (0 . +0 . − . ) c (trough B),and (0 . +0 . − . ) c (trough C), which are high-lighted with blue, orange, and magenta shadedregions in Figure 4. These three troughs alsohave accompanied Si iv troughs (top panel inFigure 4). In order to quantify the strength ofthese troughs, we measure the “balnicity” index(BI, Weymann et al. (1991)) of C iv BALs byBI = (cid:90) v max v min (cid:18) − f ( v )0 . (cid:19) Cdv. (1)where f ( v ) is normalized spectrum, C is set to 1only when f ( v ) is continuously smaller than 0.9for more than 2000 km s − , otherwise it is set to0.0. In order to avoid counting absorptions fromSi iv , we set v max = 52 ,
340 km s − . The v min isset to 0. The BIs are measured to be 3400 kms − , 890 km s − , and 1010 km s − for troughA, B, and C, respectively. The uncertainties ofthe BIs are dominated by the limitations of thetemplate matching accuracy. The total C iv BIof J0038–1527 is 5300 km s − , which is on thehigh tail ( ∼ iv to Mg ii but slightly underestimates the con-tinuum blueward of C iv broad emission line. luminous BAL Quasar at z = 7 .
02 7Such difference could be caused by the factthat high-redshift quasars tend to have flat-ter extinction curves than that of low-redshiftquasars (e.g., Maiolino et al. 2004; Gallerani etal. 2010). If so, we might slightly underesti-mate the BIs measured above. Because troughA has a very high velocity, the associated Si iv absorption trough is blueshifted to the Ly α re-gion, where the spectrum is seriously absorbedby the intervening IGM. On the other hand, thebest-fit composite template clearly overfits N v and maybe also Ly α , suggesting strong absorp-tions over this region, which could be due toSi iv BAL at ∼ . c . The other possibilityfor trough A is that it could be a Si iv BALtrough at a lower velocity, as this trough is atthe blueward of Si iv emission line. However,we do not see any accompanied C iv troughs atthe same velocity shifts (gray shaded region inFigure 4). Without associated C iv troughs, thepresence of strong absorption on the top of Ly α and N v emission lines, and the existence of theother two high-velocity troughs (B and C), thespectrum of J0038–1527 strongly suggests thattrough A is indeed an extremely high-velocityC iv BAL. Allen et al. (2011) found that theBAL quasar fraction increases by a factor of 3.5from z ∼ z ∼
4, which indicates that an ori-entation effect alone is not sufficient to explainthe presence of BAL troughs. This, togetherwith the strong BAL absorptions in J0038–1527and large C iv blueshifts found in those z (cid:38) DISCUSSION AND SUMMARYAs described in §
4, although most z > . iv emission lineblueshifts, they still follow the locus of low-redshift quasars in Figure 3. Moreover, a low-redshift quasar composite constructed using asimple cut of C iv blueshifts matches the spec-trum of J0038–1527 very well. This suggeststhat the property of C iv emission line is very CIV Velocity (km s −1 ) N o m a li z e d F l u x C: CIV at 0.08cB: CIV at 0.10cA: CIV at 0.14c N o m a li z e d F l u x SiIV at 0.08cSiIV at 0.10cSiIV at 0.14c?
SiIV Velocity (km s −1 ) Figure 4.
Normalized spectrum of J0038–1527.The x-axis of the top panel is the outflow velocityof Si iv and the x-axis of bottom panel shows theoutflow velocity relative to C iv . The blue, orangeand magenta shaded regions denote three absorp-tion systems at 0.14 c (A), 0.10 c (B), and 0.08 c (C),respectively. The velocity range of each trough wasdetermined by the C iv trough. The absence of ab-sorption in the gray shaded region in the bottompanel and strong absorption on top of Ly α and N v (blue shaded region in the top panel) suggests thatthe trough A is indeed an extremely high velocityC iv BAL. important for damping wing analysis, as themain uncertainty on such analysis comes fromhow well one can predict the intrinsic spectraof high-redshift quasars.The relativistic outflows in this quasar arefound with extremely high velocities of 0.08 c –0.14 c . Such phenomenon is very rare and onlyobserved in a small number of low-redshiftquasars (e.g., Rogerson et al. 2016; Hamann etal. 2018). Very recently, Hamann et al. (2018)found that highly ionized X-ray ultra-fast out-flow (UFO; Chartas et al. 2002; Reeves et al.2003) in a low-redshift quasar is accompaniedwith a high-velocity C iv BAL, which suggeststhat the relativistic C iv BALs might form inthe dense clumps embedded in the X-ray UFO.The associated kinetic power of these relativis-tic outflows is suggested to be well above what’s
Wang et al. needed to affect quasar host galaxies (e.g., Char-tas et al. 2009; Zubovas & King 2013). J0038–1528 is an excellent target to test whether ornot AGN feedback could affect the building upprogress of massive galaxy with future deep X-ray observations (e.g., Chartas et al. 2009) and(sub-)millimeter observations (e.g., Feruglio etal. 2017). The most distant quasars may havemore ubiquitous and stronger outflows of densegas than their low-redshift counterparts (seealso Maiolino et al. 2004). Future larger sampleof quasars at z (cid:38) =–27.1, J0038–1527 is the mostluminous z > M BH =(1.33 ± × M (cid:12) based on Mg ii emission line. We estimated the bolometric lu-minosity of J0038–1527 to be L Bol =2.16 × erg s − , which yields an Eddington ratio to be L Bol /L Edd = 1.25 ± iv broad emission linesand BAL features (see Richards et al. 2011, andreferences therein). Thus, the observed outflowsindicated by both C iv emission line blueshiftsand BALs are probably driven by high Ed-dington ratios observed in luminous quasars assuggested by Sun et al. (2018), consistent withthe properties observed in J0038–1527.J. Yang, X. Fan, M. Yue, J.-T. Schindlerand I. D. McGreer acknowledge support fromthe US NSF Grant AST-1515115 and NASAADAP Grant NNX17AF28G. X.-B.W. and L.J.acknowledge support from the National KeyR&D Program of China (2016YFA0400703)and the National Science Foundation of China(11533001 & 11721303). B. Venemans and F.Walter acknowledge funding through the ERCgrants “Cosmic Dawn” and “Cosmic Gas”. Thisresearch based on observations obtained at theGemini Observatory (GN-2018A-FT-114) andbased on observations collected at the Euro-pean Organization for Astronomical Researchin the Southern Hemisphere under ESO pro-gram 2100.A-5033 (A). We especially thank theDirectors of VLT, and UKIRT for granting usDirector Discretionary time for follow up obser-vations of this object. We acknowledge the useof data obtained at the Gemini Observatory,Magellan Telescope, and MMT Observatory.We acknowledge the use of public DELS, PS1,and WISE data.
Facilities:
Gemini (GNIRS), Magellan(LDSS3-C), MMT (Red Channel Spectro-graph), UKIRT (WFCAM), VLT (X-SHOOTER)REFERENCES
Abramowicz, M. A., Czerny, B., Lasota, J. P., &Szuszkiewicz, E. 1988, ApJ, 332, 646Allen, J. T., Hewett, P. C., Maddox, N., Richards,G. T., & Belokurov, V. 2011, MNRAS, 410, 860Baldwin, J. A. 1977, ApJ, 214, 679 Ba˜nados, E., Venemans, B. P., Mazzucchelli, C.,et al. 2018, Nature, 553, 473Bolton, J. S., Haehnelt, M. G., Warren, S. J., etal. 2011, MNRAS, 416, L70Cardelli, J. A., Clayton, G. C., & Mathis, J. S.1989, ApJ, 345, 245 luminous BAL Quasar at z = 7 .
02 902 9