X-ray spectroscopy of the z=6.4 quasar J1148+5251
Simona Gallerani, Luca Zappacosta, Maria Carmela Orofino, Enrico Piconcelli, Cristian Vignali, Andrea Ferrara, Roberto Maiolino, Fabrizio Fiore, Roberto Gilli, Andrea Pallottini, Roberto Neri, Chiara Feruglio
aa r X i v : . [ a s t r o - ph . GA ] F e b MNRAS , 1–8 (2016) Preprint February 27, 2017 Compiled using MNRAS L A TEX style file v3.0
X-ray spectroscopy of the z = 6 . quasar SDSS J1148+5251 S. Gallerani ⋆ , L. Zappacosta , M. C. Orofino , E. Piconcelli , C. Vignali , ,A. Ferrara , R. Maiolino , , F. Fiore , R. Gilli , A. Pallottini , , ,R. Neri , C. Feruglio Scuola Normale Superiore, Piazza dei Cavalieri 7, 56126, Pisa, Italy INAF - Osservatorio Astronomico di Roma, via di Frascati 33, I-00078, Monteporzio Catone, Italy Dipartimento di Fisica e Astronomia, Alma Mater Studiorum, Universit´a di Bologna, viale Berti Pichat 6/2, 40127, Bologna, Italy INAF - Osservatorio Astronomico di Bologna, via Ranzani 1, 40127 Bologna, Italy Cavendish Laboratory, University of Cambridge, 19 J. J. Thomson Ave., Cambridge CB3 0HE, UK Kavli Institute for Cosmology, University of Cambridge, Madingley Road, Cambridge CB3 0HA, UK Institut de Radioastronomie Millim´etrique (IRAM), 300 rue de la Piscine, 38406 Saint-Martin-d’H´eres, France INAF - Osservatorio Astronomico di Trieste, via G.B. Tiepolo, 11, 34143, Trieste, Italy
ABSTRACT
We present the 78-ks
Chandra observations of the z = 6 . & σ ). The X-ray spectrum is best-fitted by a power-law with photonindex Γ = 1 . H = 2 . +2 . − . × cm − . We mea-sure an intrinsic luminosity at 2-10 keV and 10-40 keV equal to ∼ . × erg s − ,comparable with luminous local and intermediate-redshift quasar properties. More-over, the X-ray to optical power-law slope value ( α OX = − . ± .
14) of J1148 isconsistent with the one found in quasars with similar rest-frame 2500 ˚A luminosity( L ∼ erg s − ˚A − ). Then we use Chandra data to test a physically motivatedmodel that computes the intrinsic X-ray flux emitted by a quasar starting from theproperties of the powering black hole and assuming that X-ray emission is attenuatedby intervening, metal-rich ( Z > Z ⊙ ) molecular clouds distributed on ∼ kpc scales inthe host galaxy. Our analysis favors a black hole mass M BH ∼ × M ⊙ and a molec-ular hydrogen mass M H ∼ × M ⊙ , in good agreement with estimates obtainedfrom previous studies. We finally discuss strengths and limits of our analysis. Key words:
X-rays: galaxies; (galaxies:) quasars: super-massive black holes; galaxies:ISM; quasars: individual: SDSS J1148+5251
Studying the properties of z ∼ z ∼ ⋆ e-mail: [email protected] sion (Brandt et al. 2001; Mathur et al. 2002; Bechtold et al.2003; Vignali et al. 2003b; Shemmer et al. 2006; Page et al.2014; Moretti et al. 2014; Ai et al. 2016).These studies were focused on X-ray luminous objects( L X > erg s − ) and have found typical X-ray to opticalpower-law slopes that vary in the range − . α OX − . N ( E ) ∝ E − Γ with 1 . Γ . N H cm − that are generallydifficult to be robustly constrained.Local and intermediate-redshift quasars of similar X-rayluminosities are typically characterized by α OX = − . ± .
02 (Just et al. 2007) and Γ = 1 . ± .
11 (Piconcelli et al.2005). Thus, z ∼ z quasar that shows peculiar X-ray properties (Γ ∼ c (cid:13) Gallerani et al. α OX ∼ − .
2) is 0100+2802 at z = 6 . M BH ∼ M ⊙ )known so far at this redshift. However, these results arebased on exploratory Chandra observations that consists ofonly 14 counts. In general, at these epochs, the number of X-ray observations is still sparse and often based on a limitednumber of counts ( z = 7 . z = 6 . >
100 counts (Nanni et al.,submitted to A&A).In this work, we present X-ray observations ofJ1148+5251 (hereafter J1148), a quasar at z = 6 .
4, obtainedwith the
Chandra telescope. Since its discovery (Fan et al.2003; White et al. 2003), J1148 has been observed atseveral different wavelengths (e.g. Gallerani et al. 2008;Juarez et al. 2009; Gallerani et al. 2010; Riechers et al.2009; Maiolino et al. 2012; Gallerani et al. 2014;Cicone et al. 2015; Stefan et al. 2015) and certainlyrepresents the best studied case of z > M BH = (2 ÷ × M ⊙ (Willott et al. 2003; Barth et al. 2003) has been derived.Moreover, several CO observations have been obtainedfor this object in the last years (Riechers et al. 2009;Gallerani et al. 2014; Stefan et al. 2015) suggesting amolecular hydrogen mass M H ∼ . × M ⊙ and a sizeR H ∼ . Chandra observations, the procedure adopted for data reduction andcalibration, and the results from the X-ray spectral analysis;in Sec. 3.2 we discuss the serendipitous, tentative detectionof X-rays at the location of an optically faint z = 5 . The 78 ks
Chandra observation (ObsId 17127) was per-formed in September 2, 2015. The target was observed withACIS-S in Very Faint mode. In order to obtain the most up-dated data reduction and calibration files we reprocessed thedata using the script chandra repro included in CIAO v. 4.8with CALDB v. 4.7.0. We created a S3 exposure corrected0.5-7 keV image of the field of view (see Fig. 1) and run on itthe source detection algorithm wavdetect with scales 1.0 2.0and 4.0 pixels and false probability threshold for identifyingpixels belonging to a source of 10 − . The source is clearly This value translates in the whole S3 chip to 1 statistical fluc-tuation identified as point-source. detected with 42 background-subtracted counts with a sig-nificance of & σ (estimated within the source extraction re-gion). The latter has been estimated by dividing the sourcenet-counts by the Poissonian uncertainty (Gehrels 1986) onthe background counts expected (i.e. re-normalized from thebackground region) in the source region. We selected the re-gion for source spectral extraction as a circular region of3 arcsec radius centered on the J1148 position as reportedby wavdetect . The background extraction region was chosenas an annular region of inner and outer radii of 6 and 50 arc-sec centered on J1148 from which we have removed 3 arcsecradius circular regions centered on three nearby barely de-tected point-sources (see Fig. 1). For the spectral extractionwe used the script specextract which automates spectral ex-traction within our selected regions and creation of relativeresponse files. For the spectral analysis we used
XSPEC v. 12.8.2(Arnaud 1996). Given the low number of counts (42) wegrouped the spectrum at 1 net-count per bin and used C-stat (Cash 1979) with direct background subtraction (Wstatin XSPEC; Wachter et al. 1979). Furthermore, in order tocheck the validity of our results against possible systemat-ics, we also performed a joint fit with the source (not back-ground subtracted) and background spectrum (grouped at1 count per bin) using C-stat. We performed all the mod-ellings in the energy range 0.3-5 keV. Indeed the source isdetected up to ∼ z = 6 .
4, we are sam-pling a rest-frame energy range ∼ −
37 keV. The Galacticcolumn density of N Gal H = 1 . × cm − , derived fromKalberla et al. (2005), was applied to all the models. In thefollowing, errors are quoted at 1 σ level and upper limits at90%.We initially fitted the data with a simple power-lawmodel (see Fig. 2; left panel) obtaining a best-fit slope ofΓ = 1 . ± .
3. This slope is slightly flatter than (but stillconsistent with) the typical Γ value at E > ≈ .
9; Piconcelli et al. 2005). We notice thatthe model slightly overestimates the data in the softest X-ray portion of the spectrum. For this reason, we added anintrinsic photoelectric component (
ZWABS in XSPEC) to in-vestigate the possible presence of absorbing material alongthe line of sight to the quasar. This resulted in a rather un-certain and steeper power-law photon index, Γ = 2 . +1 . − . ,and a loosely constrained column density value for whichwe obtained an upper limit of N H < . × cm − . Inorder to obtain better constraints on the column density, wefix the slope to Γ = 1 .
9, consistently with the value foundfor the simple power-law model; the value is also within therange typically found for optically selected high- z quasars(e.g. Vignali et al. 2003b). In this case we obtain a columndensity N H = 2 . +2 . − . × cm − , still consistent with noabsorption at 1.3 σ .We checked for the possible presence of the Comptonreflection hump (George & Fabian 1991) by adding to thesimple power-law model a Compton reflection component,empirically parametrized in XSPEC by the PEXRAV model
MNRAS , 1–8 (2016) -ray spectroscopy of the z = 6 . quasar J1148+5251 : . : : . : .
200 kpc
Figure 1.
Chandra
ACIS-S 0.5-7 keV exposure-corrected imagecentered on J1148. The image is smoothed with a Gaussian kernelof radius 0.75 arcsec (3 pixels). Thick black circular region showsthe spectral extraction region for the source. Thin blue regionindicates the background extraction area. Circular regions for thethree point source removed from the latter are shown with barredblue circles. (Magdziarz & Zdziarski 1995). This model calculates the re-flection of the power-law primary AGN flux on an infiniteslab of neutral material with infinite optical depth. The re-flection strength is described by the R parameter which is de-fined as R=Ω / π , where Ω defines the solid angle subtendedby the reflector. We set the energy cut-off of the input power-law spectrum in PEXRAV to 200 keV, the abundance to Solarand the inclination angle to 45 deg . By fitting the observedspectrum with this model we obtain a photon index thatis too steep (Γ ≈ . . R < .
19. This value is in agreement withthe low level of reflection observed in higher quality spectraof lower redshift luminous broad-line AGN (Vignali et al.1999; Reeves & Turner 2000; Page et al. 2005, Zappacostaet al. in prep.).We checked our results also by modelling the totalsource+background extracted spectrum (i.e. without anybackground subtraction) jointly with the background spec-trum using C-stat as fit statistics similarly to what hasbeen done in ACIS-I, for instance, by Lanzuisi et al. (2013).In order to do this we modelled the ACIS-S3 unfocusedinstrumental background spectrum (grouped to 1 countper bin) as a three-segment broken power law with slopesand the two energy break free to vary and two instru-mental lines at ∼ ∼ .
41 and nor-malization at 1 keV set to 11 . − s − sr − keV − (De Luca & Molendi 2004). In order to get accurate globalinstrumental background parameters and the normalizationsof the diffuse background components, we extracted a spec-trum from a large 3 arcmin radius circular region (care-fully excluding from it the detected point-sources) collecting ∼ F . − = 1 . ± . × − erg s − cm − and F − = 3 . ± . × − erg s − cm − . The intrinsic lu-minosities in the 2-10 and 10-40 keV energy bands are re-spectively of L − = 1 . +0 . − . × erg s − and L − =1 . +0 . − . × erg s − . These luminosities are comparableto those typically measured in local and intermediate red-shift quasars (i.e. Piconcelli et al. 2005, Zappacosta et al. inprep.).Finally, we measure the X-ray to optical power-lawslope defined as: α OX = 0 . log ( f /f ), where f and f are the monochromatic flux densities atrest-frame 2 keV and 2500 ˚A, respectively. In particu-lar, we use the composite quasar spectrum presented inVanden Berk et al. (2001), to convert the broad-band z-filtermeasurements in f , as in Vignali et al. (2003a). We find α OX = − . ± .
14. Just et al. (2007) measured the α OX in 32 luminous quasars at z ∼ . − . α OX anti-correlates with the their luminosity at 2500 ˚A ( L ).By adopting such relation (see eq. 3 in Just et al. 2007), andconsidering that J1148 has L = 10 . erg s − Hz − , weinfer α OX = − .
73, perfectly consistent with our estimate.This result implies that the α OX of this luminous high- z quasar is comparable with the ones found in lower redshiftquasars. z = 5 . quasar The observed field of view encloses RD J1148+5253,a quasar discovered by Mahabal et al. (2005) at z = 5 . ′′ radius. Tounderstand whether this might be considered a detectionand derive the corresponding significance level, we adoptedthree methods. This kind of instrumental+cosmic background modeling is usu-ally employed in faint point-source and diffuse source studies(Fiore et al. 2012; Humphrey et al. 2006)MNRAS000
73, perfectly consistent with our estimate.This result implies that the α OX of this luminous high- z quasar is comparable with the ones found in lower redshiftquasars. z = 5 . quasar The observed field of view encloses RD J1148+5253,a quasar discovered by Mahabal et al. (2005) at z = 5 . ′′ radius. Tounderstand whether this might be considered a detectionand derive the corresponding significance level, we adoptedthree methods. This kind of instrumental+cosmic background modeling is usu-ally employed in faint point-source and diffuse source studies(Fiore et al. 2012; Humphrey et al. 2006)MNRAS000 , 1–8 (2016)
Gallerani et al. −5 −4 −3 no r m a li z ed c oun t s s − k e V − r a t i o Energy (keV) 10 −5 −4 −3 no r m a li z ed c oun t s s − k e V − r a t i o Energy (keV)
Figure 2.
Chandra
ACIS-S spectra and best-fit model for the simple power-law parametrization. The left panel shows the modellingon background subtracted data while the right panel reports the joint modelling on source+background (black thick) and backgroundonly (red thin). Both data and model for the latter are shown not scaled to the source+background spectral extraction area (i.e. there-normalization of the background to the source+background spectrum has been set in the corresponding model). Background modelsare shown as dotted lines. The solid black line reports the total best-fit source+background spectral model.
At first, similarly to Vignali et al. (2001), we extracteda 400 ×
400 pixel region centered on the optical position ofthe quasar, excluding the immediate vicinity of the quasaritself and masking few other detected sources. This regionwas covered with 30,000 circles of 1.5 ′′ radius whose centerswere randomly chosen through a Monte Carlo procedure. Atthis stage, to be conservative in the detection significance,we considered only the three contiguous counts (two of whichin the hard, 2–7 keV, band) and found 161 and 518 extrac-tions with at least three and two counts (out of the 30000trials) in the full and hard band, respectively, correspond-ing to about 2.9 σ and 2.5 σ confidence level for a one-tailedGaussian distribution in the two bands.To verify this tentative detection, we considered themean number of counts extracted in the full and hard bandin the 30000 trials and assumed them as indicative of thebackground levels in the two bands, obtaining a proba-bility of 2.6 σ and 2.2 σ , respectively. Finally, we ran thesource-detection tool wavdetect Freeman et al. (2002) onthe full-band image to verify the results reported above.The source is detected at a false-positive threshold of 10 − .Although wavdetect is not properly a tool for photome-try, we used its results in terms of number of counts. Thesource is detected with four counts, 0.7 of these due to thebackground. The Poisson probability of obtaining four total(source+background) counts when 0.7 are expected in theextraction region is ≈ × − , corresponding to ≈ σ . Fromthe source net counts and assuming Poissonian uncertainties(Gehrels 1986) we estimate the 0.5-7 keV source count-rateto be 4 . +3 . − . × − counts s − . Assuming a power-law withcanonical Γ = 1 . . +2 . − . × − erg s − cm − . We then calculate the probabil-ity that the optical source and X-ray counterparts are asso-ciated by chance. Given the 0.5-2 keV fluxes we are probingand taking as reference the source counts by Lehmer et al.(2012), we expect to have in the ACIS-S3 field between 23and 57 sources (accounting for 1 σ uncertainty in flux). Byassuming 1.5 ′′ radius for the X-ray counterpart we expect arandom association probability in the range (6 − × − . Summarizing, without assuming any prior (e.g., the op-tical position of the source, as in the Monte Carlo methods),the source is tentatively detected with a significance level of2.6–2.9 σ in the full band and 2.2–2.5 σ in the hard band. Thecount rate in the 0.5-7 keV band (4 . × − counts s − ) cor-responds to an intrinsic luminosity of ∼ × erg s − ,assuming (Γ = 1 . m J = 21 .
45) of RD J1148+5253,and extrapolating the flux to λ = 2500 ˚A with theVanden Berk et al. (2001) template ( α λ = − . α OX = − . In Sec. 3.1, we have seen that if we fix the spectral slopeto the value that is commonly found at lower redshift (Γ =1 .
9) we can marginally constrain the column density ( N H =2 . +2 . − . × cm − ) of the absorbing gas that is interveningalong the line of sight towards us. Although our detection ofthe X-ray absorption is at a low significance level, here wespeculate on its origin and assume that it is due to metalslocked in molecular clouds, distributed on ∼ kpc scales (seethe model by Gallerani et al. 2014, hereafter G14). In Sec.6 we extensively discuss this assumption.We compute the observed X-ray flux as F obs ν = F int ν e − τ ,where F int ν is the intrinsic X-ray spectrum of J1148 and τ is the optical depth of neutral hydrogen encountered by X-ray photons in the host galaxy along the line of sight to-wards us. We model F int ν with a single power-law havingslope Γ. Given the bolometric luminosity L bol = f Edd L Edd ,where f Edd is the Eddington ratio, and L Edd = 1 . × M BH erg s − M − ⊙ is the Eddington luminosity, theX-ray luminosity is estimated adopting bolometric cor- The subscript X in quantities as L X , F X , f X refers to the band2-10 keV, if not differently stated. MNRAS , 1–8 (2016) -ray spectroscopy of the z = 6 . quasar J1148+5251 rections L X = L bol /f X . The optical depth is given by τ = N H (1 . σ T + σ ph ) (Yaqoob 1997), where N H is the hy-drogen column density, σ T is the Thomson cross-sectionand σ ph is the interstellar photoelectric cross-section byMorrison & McCammon (1983) that depends on the gasmetallicity Z . We assume that the intrinsic spectrum atten-uation is due to molecular clouds characterized by a radius r cl and a density n cl , randomly distributed within a sphereof radius R H . We can thus write the column density N H asfollows: N H = 2 N k cl N cl , (1)where N cl is the cloud column density given by N cl = π r cl n cl , and N k cl is the number of clouds along a line ofsight: N k cl = N tot cl V H σ c R H , (2)where N tot cl = M H /M c is total number of clouds of mass M c , V H is the volume occupied by H , and σ c is the geo-metrical cross section of each cloud ( πr ). Thus, N H can berewritten as: N H = 5 × (cid:18) M H M ⊙ (cid:19) (cid:18) R H . (cid:19) − . (3)To summarize, under the aforementioned assumptions, F obs ν is completely specified by f Edd , M BH , Γ and f X for the in-trinsic spectrum, and M H , R H and Z for the absorptionmodel. A comprehensive analysis of the full parameter spaceis hampered by the limited number of counts of the Chandra spectrum combined with the large number of free parame-ters of our model. Moreover, some of the free parameters aredegenerate, as M BH with f Edd and f X . Nevertheless, J1148represents the best studied case of z ∼ ∼ kpc scales. In what fol-lows, we consider M BH and M H as free parameters of themodel, and we assume for the others reasonable values takenfrom the literature.We assume that J1148 shines at the Eddington luminos-ity, i.e. f Edd = 1. This assumption is supported by severalworks, e.g. Willott et al. (2003) and Schneider et al. (2015)for the specific case of J1148 and Wu et al. (2015) for thegeneral case of high-z quasars (see their Fig. 4). For the in-trinsic spectrum, we adopt Γ = 1 . f X = 230 +170 − by Lusso et al. (2012). We note that the f X value at theJ1148 bolometric luminosity L bol ∼ L ⊙ is extrapolatedfrom results obtained at lower L bol ; however, it is consis-tent within ∼ σ with the work by Hopkins et al. (2007) (i.e. f X = 150 ±
50 ), and with the results by Marconi et al. (2004)(i.e. f X = 114) for the same bolometric luminosity. More im-portantly, our choice is justified by the fact that if we takeinto account the black hole mass measured by Willott et al.(2003) and Barth et al. (2003), M BH ∼ × M ⊙ , the bolo-metric correction by Lusso et al. (2012) returns an X-ray lu-minosity (10 L X / erg s − × ) that is perfectlyconsistent with our measurements (see Sec. 3.1).For what concerns the absorption model, high angularresolution (0.15”, namely < z = 6 .
4) VLA observa-
Figure 3.
X-ray emission model compared with the observed X-ray spectrum in the source rest frame energy range. Red circlesdenote data re-binned in such a way that each bin contains atleast 12-14 background subtracted counts; the gray shaded regionshows the intrinsic X-ray spectrum predicted for an Eddington-luminous quasar with mass of M BH = 2 . ± . × M ⊙ ; thecyan hatched region represents our best fit model that considersabsorption from an hydrogen column density N H = 7 . ± . × cm − due to a molecular hydrogen mass of M H = 1 . ± . × M ⊙ . tions of the CO(3-2) emission by Walter et al. (2004) haveshown that the molecular gas is extended out to radii of R H = 2 . ± . Z =7 ± ⊙ obtained from near-infrared observations of 3 . 3Γ = 3 . ± . ± . . ± . 3Γ = 1 . . ± . . ± . . ± . f X = 130 1 . ± . . ± . . ± . f X = 400 4 . ± . . ± . . ± . Z = Z ⊙ . ± . ± ± Z = 10 Z ⊙ . ± . . ± . . ± . R H = 1 . . ± . . ± . . ± . R H = 3 . . ± . . ± . . ± . Table 1. Best values of the black hole mass M BH , molecularhydrogen mass M H and hydrogen column density N H obtainedfor the fiducial model (Γ = 1 . f X = 230, Z = 7 Z ⊙ , R H =2 . M H would change if we vary the fiducial parameters (Γ, f X , Z , R H ). We do not consider variations on f Edd sincethe results would be degenerate with f X . Moreover, for whatconcerns Z , we note that the measurements by Nagao et al.(2006) refers to the broad line regions (BLRs), small nu-clear regions ( few pc) that may be characterized by largermetallicity values with respect to the one in the host galaxy(e.g. Valiante et al. 2011). Thus, we consider Z = Z ⊙ aslower limit for the metallicity of J1148. The results of thisanalysis are reported in Table 1. Both M BH and M H showsthe largest variations (up to ∼ a factor of 10) with Γ.These results represent an important consistency checkof our model. In other words, we are saying that Chan-dra observations can be explained by assuming that theintrinsic X-ray spectrum of J1148 is attenuated by inter-vening, metal-rich ( Z > Z ⊙ ) molecular clouds of total mass M H ∼ × M ⊙ , distributed at ∼ kpc distance from thecenter of the galaxy, characterized by r cl ∼ − 10 pc anddensity n cl = 10 − cm − .However, given the large uncertainties that are plaguingour analysis and the novelty of our approach some cautionis needed. In the next section we thoroughly discuss bothstrengths and limits of this study. Estimates of M H rely on the relation M H = α CO × L ′ CO(1 − , where α CO is the conversion factor, in units ofM ⊙ (K km s − pc ) − , and L ′ CO(1 − is the CO(1-0) lumi-nosity, in units of K km s − pc . For what concerns the con-version factor, in Galactic molecular clouds α CO = 0 . 8, whileultra-luminous infrared galaxies (ULIRGs, L FIR > L ⊙ )and nuclear star-burst galaxies have α CO = 4. However,this factor remains highly uncertain (see Bolatto et al. 2013,for a review on this subject) and generally presents strongvariations with the metallicity (e.g. Narayanan et al. 2012)and with the star formation rate (e.g. Clark & Glover 2015).Moreover, CO(1-0) data are generally poorly constrained athigh- z since the CO(1-0) frequency ( ν RF = 115 GHz) fallsoutside the ALMA bands for z > . 4. Thus, no z ∼ M H estimate is based on the assumptionof a constant brightness in the CO(3-2) and CO(1-0) tran-sitions.On the one hand, all these caveats make the possibility ofmeasuring M H through alternative, independent methodsextremely appealing. On the other hand, our result is validunder the assumption that the intervening gas absorbingX-ray photons is constituted by molecular clouds (MC) dis-tributed on ∼ kpc scales, larger than those generally consid-ered in more standard scenarios.According to the unified-model (e.g. Antonucci 1993;Urry & Padovani 1995), the large variety of AGN types isjust the result of varying orientation relative to the line ofsight. In this scenario, objects that appear as optical TypeIIAGN are observed along a line of sight covered by an ax-isymmetric dusty structure (the “torus”) with dimensions r torus ∼ . − 10 pc. TypeI AGNs are instead not obscuredsince the line of sight does not intercept the torus itself.Since its original formulation, the unified-model has beenoften revisited (see Netzer 2015, for a recent review dis-cussing the origin and properties of the central obscurer andthe connection with its surrounding). For example, anotherpossible origin for X-ray absorption concerns with gas thatis located on broad line region (BLR, e.g. Peterson 2006)scales ( r BLR ∼ . − n BLR > cm − ), metal-rich ( Z ∼ − 10 Z ⊙ ) clouds(Risaliti et al. 2002). This scenario (hereafter called BLR-scenario ) nicely accounts for the X-ray emission variabil-ity of AGN that occurs on very short time scales (from afraction of a day up to years), and thus requires the ab-sorber to be at sub-parsec distances from the X-ray source(e.g. Cappi et al. 2016; Bianchi et al. 2009; Risaliti et al.2005). It has been also proposed by Matt (2000) that whileCompton-thick Seyfert 2 galaxies are observed through thetorus, Compton-thin/intermediate Seyfert galaxies are ob-scured by dust lanes in the host galaxy.The existence of all the aforementioned scenariosdemonstrates that the origin of the X-ray absorption is farfrom being definitively understood and must occur on a widerange of scales. This is in line with the complex picture dis-cussed by Elvis (2012) that takes into account the BLR, thetorus obscuration, and the presence of outer (0 . − z quasars are ex-pected to form in massive, over-dense regions, with darkmatter halo mass M DM ∼ − M ⊙ and virial radii r vir ∼ − 100 kpc, that result from of a long historyof lower mass systems merging (e.g. Valiante et al. 2014). Variability studies of local AGN ( z < . 1) indicate that moreluminous sources vary with a lower amplitude, implying thatthe X-ray variability of J1148 is expected to be modest (e.g.Lawrence & Papadakis 1993; Almaini et al. 2000; Paolillo et al.2004). However, there are hints that quasars of the same X-rayluminosity are more variable at z > , 1–8 (2016) -ray spectroscopy of the z = 6 . quasar J1148+5251 More in general, since high-redshift galaxies are thought tobe very compact and gas-rich systems, it is plausible thattheir interstellar medium may provide substantial contribu-tion to the X-ray emission obscuration (e.g. Bournaud et al.2011; Juneau et al. 2013; Gilli et al. 2014). Thus, the as-sumption of having intervening material distributed also on ∼ kpc scales sounds plausible.We further note that the H mass inferred from ouranalysis corresponds to a column density of neutral hydro-gen N H ∼ cm − . Since we are assuming a metallicityclose to (or even higher than) the solar one, if the Galactic A V /N H ∼ − value (e.g. G¨uver & ¨Ozel 2009) would ap-ply to J1148, this column density implies an A V > A V ∼ 1, Gallerani et al. 2010).This discrepancy is commonly found in several opticallybright quasars (e.g. Maccacaro et al. 1982; Granato et al.1997; Maiolino et al. 2001a).In the BLR scenario , high gas column densities can providea small A V since BLRs are confined within the dust sublima-tion radius ( R sub ∼ A V /N H value smaller than the Galactic one(Maiolino et al. 2001b).In fact, for a given total dust mass and fixed dust com-position, a grain size distribution flatter than the Galacticone ( q < 1) and rich of large grains ( a max > µ m) canreduce the expected A V /N H value by a factor of 10 − MC scenario there are lines of reasoning thatcan make large gas column densities consistent with small A V value. Ionized, atomic and molecular outflows are com-monly observed towards local and high- z quasars (e.g.Feruglio et al. 2010; Rupke & Veilleux 2011; Carniani et al.2015, 2016; Feruglio et al. 2015, see also Fabian (2012) fora review on this subject). In particular, observations of theCO(1-0) emission line in local ultra-luminous galaxies havefound that a non-negligible fraction (from few% up to 30%)of the total H mass is located in outflows extended on ∼ kpc scales (Cicone et al. 2014). For what concerns J1148,while observations of CO emission lines are still not sensi-tive enough for detecting molecular outflows, PdBI data ofthe [CII] 158 µ m line have provided the evidence for out-flowing gas in this high- z quasar (Maiolino et al. 2012; The dust grain sizes distribution can be modelled through apower-law: dn ( a ) ∝ a − q da , where a is the grain diameter thatvaries in the range a min < a < a max . The Galactic curve in thediffuse ISM can be described by a mixture of graphite and silicategrains having sizes distributed with q = 3 . a min = 0 . µ m and a max = 0 . µ m (Mathis et al. 1977). We note that the [CII] emission is extended on scales ( ∼ − 30 kpc) larger than the CO emission. The origin of thisextended [CII] emission is not clear, as discussed in details inCicone et al. (2014). [CII] emission can both arise from dense ( n > cm − ) photo-dissociation regions (PDRs) and more diffuse( n cm − ) neutral/ionized gas, though the dominant con-tribution is expected to arise from PDRs (e.g. Vallini et al. 2013,2015). From Table 1, we note that if we would include contribu-tion from PDRs, namely we would consider an R H > 10 kpc,we would get a total H mass that is a factor > 10 larger thanthe one inferred by CO observations. In the case more sensitiveobservations would detect CO emission on the [CII] scales, our Cicone et al. 2015). In a recent work, Ferrara & Scannapieco(2016) have studied the physical properties of quasar out-flows finding that dust grains are rapidly destroyed by sput-tering on timescales ∼ yrs. Since sputtering preferen-tially destroys small grains (Dwek et al. 1996), molecularclumps can only form on massive dust grains, although theefficiency of this process is expected to be very small, chal-lenging the theoretical interpretation of massive molecularoutflows observed in local galaxies (e.g. Feruglio et al. 2010;Cicone et al. 2014). To summarize, the amount of dust alongthe line of sight intercepted by molecular outflows is ex-pected to be negligible, and its grain size distribution islikely dominated by large grains. This helps making the col-umn density inferred from X-ray data consistent with the A V inferred from optical/near-infrared observations.To conclude, our study has the important implicationthat X-ray observations can provide M H measurementsthat are independent from millimeter observations of theCO emission lines. Nevertheless, given the marginal detec-tion of the X-ray absorption resulting from our analysis andthe large uncertainty on the origin of X-ray absorbers, ourinterpretation of the results remains speculative and needto be tested with higher quality X-ray data. Deeper Chan-dra observations are necessary to better constrain the X-raypower-law photon index, and thus the column density ofthe absorbing gas. By detecting twice the number of X-rayphotons collected in this work the error on Γ would be re-duced by ∼ M BH and M H have been measuredthrough more standard techniques (e.g. MgII, CIV and COemission lines). We have presented the X-ray observation of the quasarJ1148+5251 ( z = 6 . 4) obtained with 78 ks of Chandra ob-serving time. We have clearly detected the source with 42net counts (with a significance of & σ ) in the observed 0.3–7 keV energy band. We have modelled the X-ray spectrumwith a power-law (photon index Γ = 1 . 9) absorbed by agas column density of N H = 2 . +2 . − . × cm − . We havefound that the X-ray properties of J1148 are comparableto those of lower redshift quasars: (i) we have measured in-trinsic luminosities of L − = 1 . +0 . − . × erg s − and L − = 1 . +0 . − . × erg s − , respectively; (ii) we havechecked for the presence of a reflection component over thesimple power-law modelling finding a low degree of reflec-tion from circum-nuclear material; (iii) we have measuredan X-ray to optical power-law slope ( α OX = − . ± . z = 5 . 70 quasar discovered byMahabal et al. (2005), we have detected three (two) pho-tons in the full (hard) X-ray band. We have adopted three model could be easily updated to acount for larger H masses.Viceversa, if the extended [CII] emission arises from diffuse gas,its contribution to the X-ray absorbing gas would be negligible( ∼ , 1–8 (2016) Gallerani et al. different methods to estimate the reliability of these detec-tions and we have found a significance level of ∼ σ ,i.e. the source is only marginally detected by Chandra .Deeper observations are required to confirm the detectionof this source. The tentatively inferred X-ray luminosity ofRD J1148+5253 ( ∼ × erg s − ) is typical of sources thatare intermediate between the Seyfert and quasar luminos-ity regime. Moreover, given its observed J band magnitude( m J = 21 . α OX = − . Chandra observations to test aphysically motivated model that computes the intrinsic X-ray flux emitted by a quasar starting from the properties ofthe powering black hole and assumes that X-ray emission isattenuated by intervening, metal-rich ( Z > Z ⊙ ) molecularclouds distributed on ∼ kpc scales in the host galaxy. Ouranalysis favors a black hole mass M BH ∼ × M ⊙ anda molecular hydrogen mass M H ∼ × M ⊙ , in goodagreement with estimates obtained from previous studies,thus providing a solid consistency check to our model.This work highlights the importance of the synergy be-tween X-ray and millimeter data for studying the propertiesof high- z quasars and their host galaxies. ACKNOWLEDGEMENTS RM acknowledges support from the ERC AdvancedGrant 695671 “QUENCH” and from the Science and Tech-nology Facilities Council (STFC). References Ai Y., Dou L., Fan X., Wang F., Wu X.-B., Bian F., 2016, ApJ,823, L37Almaini O., et al., 2000, MNRAS, 315, 325Antonucci R., 1993, ARA&A, 31, 473Arnaud K. A., 1996, in Jacoby G. H., Barnes J., eds, AstronomicalSociety of the Pacific Conference Series Vol. 101, Astronomi-cal Data Analysis Software and Systems V. p. 17Asplund M., Grevesse N., Sauval A. J., Scott P., 2009, ARA&A,47, 481Ba˜nados E., et al., 2016, preprint, ( arXiv:1608.03279 )Barth A. J., Martini P., Nelson C. H., Ho L. C., 2003, ApJ,594, L95Bechtold J., et al., 2003, ApJ, 588, 119Bertoldi F., et al., 2003, A&A, 409, L47Bianchi S., Piconcelli E., Chiaberge M., Bail´on E. J., Matt G.,Fiore F., 2009, ApJ, 695, 781Bolatto A. D., Wolfire M., Leroy A. K., 2013, ARA&A, 51, 207Bournaud F., Dekel A., Teyssier R., Cacciato M., Daddi E.,Juneau S., Shankar F., 2011, ApJ, 741, L33Brandt W. N., Guainazzi M., Kaspi S., Fan X., Schneider D. P.,Strauss M. A., Clavel J., Gunn J. E., 2001, AJ, 121, 591Cappi M., et al., 2016, A&A, 592, A27Carniani S., et al., 2015, A&A, 580, A102Carniani S., et al., 2016, A&A, 591, A28Cash W., 1979, ApJ, 228, 939Cicone C., et al., 2014, A&A, 562, A21Cicone C., et al., 2015, A&A, 574, A14Clark P. C., Glover S. C. O., 2015, MNRAS, 452, 2057De Luca A., Molendi S., 2004, A&A, 419, 837Dwek E., Foster S. M., Vancura O., 1996, ApJ, 457, 244Elvis M., 2012, Journal of Physics Conference Series, 372, 012032Fabian A. C., 2012, ARA&A, 50, 455Fan X., 2012, Research in Astronomy and Astrophysics, 12, 865 Fan X., et al., 2003, AJ, 125, 1649Farrah D., Priddey R., Wilman R., Haehnelt M., McMahon R.,2004, ApJ, 611, L13Ferrara A., Scannapieco E., 2016, preprint, ( arXiv:1609.02561 )Feruglio C., Maiolino R., Piconcelli E., Menci N., Aussel H.,Lamastra A., Fiore F., 2010, A&A, 518, L155Feruglio C., et al., 2015, A&A, 583, A99Fiore F., et al., 2012, A&A, 537, A16Freeman P. E., Kashyap V., Rosner R., Lamb D. Q., 2002, ApJS,138, 185Gallerani S., Ferrara A., Fan X., Choudhury T. R., 2008, MNRAS,386, 359Gallerani S., et al., 2010, A&A, 523, A85Gallerani S., Ferrara A., Neri R., Maiolino R., 2014, MNRAS,445, 2848Gallerani S., Fan X., Maiolino R., Pacucci F., 2017, preprint,( arXiv:1702.06123 )Gehrels N., 1986, ApJ, 303, 336George I. M., Fabian A. C., 1991, MNRAS, 249, 352Gilli R., et al., 2014, A&A, 562, A67Granato G. L., Danese L., Franceschini A., 1997, ApJ, 486, 147G¨uver T., ¨Ozel F., 2009, MNRAS, 400, 2050Hopkins P. F., Richards G. T., Hernquist L., 2007, ApJ, 654, 731Humphrey P. J., Buote D. A., Gastaldello F., Zappacosta L., Bul-lock J. S., Brighenti F., Mathews W. G., 2006, ApJ, 646, 899Humphrey P. J., Liu W., Buote D. A., 2009, ApJ, 693, 822Juarez Y., Maiolino R., Mujica R., Pedani M., Marinoni S., NagaoT., Marconi A., Oliva E., 2009, A&A, 494, L25Juneau S., et al., 2013, ApJ, 764, 176Just D. W., Brandt W. N., Shemmer O., Steffen A. T., SchneiderD. P., Chartas G., Garmire G. P., 2007, ApJ, 665, 1004Kalberla P. M. W., Burton W. B., Hartmann D., Arnal E. M.,Bajaja E., Morras R., P¨oppel W. G. L., 2005, A&A, 440, 775Kormendy J., Ho L. C., 2013, ARA&A, 51, 511Lanzuisi G., et al., 2013, MNRAS, 431, 978Lanzuisi G., et al., 2014, ApJ, 781, 105Lawrence A., Papadakis I., 1993, ApJ, 414, L85Lehmer B. D., et al., 2012, ApJ, 752, 46Lumb D. H., Warwick R. S., Page M., De Luca A., 2002, A&A,389, 93Lusso E., et al., 2012, MNRAS, 425, 623Maccacaro T., Perola G. C., Elvis M., 1982, ApJ, 257, 47Magdziarz P., Zdziarski A. A., 1995, MNRAS, 273, 837Mahabal A., Stern D., Bogosavljevi´c M., Djorgovski S. G.,Thompson D., 2005, ApJ, 634, L9Maiolino R., Marconi A., Salvati M., Risaliti G., Severgnini P.,Oliva E., La Franca F., Vanzi L., 2001a, A&A, 365, 28Maiolino R., Marconi A., Oliva E., 2001b, A&A, 365, 37Maiolino R., et al., 2012, MNRAS, 425, L66Manners J., Almaini O., Lawrence A., 2002, MNRAS, 330, 390Marconi A., Risaliti G., Gilli R., Hunt L. K., Maiolino R., SalvatiM., 2004, MNRAS, 351, 169Mathis J. S., Rumpl W., Nordsieck K. H., 1977, ApJ, 217, 425Mathur S., Wilkes B. J., Ghosh H., 2002, ApJ, 570, L5Matt G., 2000, A&A, 355, L31Moretti A., et al., 2014, A&A, 563, A46Morrison R., McCammon D., 1983, ApJ, 270, 119Nagao T., Marconi A., Maiolino R., 2006, A&A, 447, 157Narayanan D., Krumholz M. R., Ostriker E. C., Hernquist L.,2012, MNRAS, 421, 3127Netzer H., 2015, ARA&A, 53, 365Page K. L., Reeves J. N., O’Brien P. T., Turner M. J. L., 2005,MNRAS, 364, 195Page M. J., Simpson C., Mortlock D. J., Warren S. J., HewettP. C., Venemans B. P., McMahon R. G., 2014, MNRAS,440, L91Paolillo M., Schreier E. J., Giacconi R., Koekemoer A. M., GroginN. A., 2004, ApJ, 611, 93 MNRAS , 1–8 (2016) -ray spectroscopy of the z = 6 . quasar J1148+5251 Peterson B. M., 2006, in Alloin D., ed., Lecture Notes in Physics,Berlin Springer Verlag Vol. 693, Physics of Active GalacticNuclei at all Scales. p. 77, doi:10.1007/3-540-34621-X˙3Piconcelli E., Jimenez-Bail´on E., Guainazzi M., Schartel N.,Rodr´ıguez-Pascual P. M., Santos-Lle´o M., 2005, A&A, 432, 15Reeves J. N., Turner M. J. L., 2000, MNRAS, 316, 234Riechers D. A., et al., 2009, ApJ, 703, 1338Risaliti G., Elvis M., Nicastro F., 2002, ApJ, 571, 234Risaliti G., Elvis M., Fabbiano G., Baldi A., Zezas A., 2005, ApJ,623, L93Rupke D. S. N., Veilleux S., 2011, ApJ, 729, L27Schneider R., Bianchi S., Valiante R., Risaliti G., Salvadori S.,2015, A&A, 579, A60Shemmer O., et al., 2006, ApJ, 644, 86Shemmer O., et al., 2014, ApJ, 783, 116Stefan I. I., et al., 2015, MNRAS, 451, 1713Urry C. M., Padovani P., 1995, PASP, 107, 803Vagnetti F., Middei R., Antonucci M., Paolillo M., Serafinelli R.,2016, A&A, 593, A55Valiante R., Schneider R., Salvadori S., Bianchi S., 2011, MNRAS,416, 1916Valiante R., Schneider R., Salvadori S., Gallerani S., 2014,MNRAS, 444, 2442Valiante R., Schneider R., Volonteri M., 2016,Publ. Astron. Soc. Australia, 33, e032Vallini L., Gallerani S., Ferrara A., Baek S., 2013, MNRAS,433, 1567Vallini L., Gallerani S., Ferrara A., Pallottini A., Yue B., 2015,ApJ, 813, 36Vanden Berk D. E., et al., 2001, AJ, 122, 549Vignali C., Comastri A., Cappi M., Palumbo G. G. C., MatsuokaM., Kubo H., 1999, ApJ, 516, 582Vignali C., Brandt W. N., Fan X., Gunn J. E., Kaspi S., SchneiderD. P., Strauss M. A., 2001, AJ, 122, 2143Vignali C., Brandt W. N., Schneider D. P., 2003a, AJ, 125, 433Vignali C., et al., 2003b, AJ, 125, 2876Volonteri M., 2012, Science, 337, 544Wachter K., Leach R., Kellogg E., 1979, ApJ, 230, 274Walter F., et al., 2003, Nature, 424, 406Walter F., Carilli C., Bertoldi F., Menten K., Cox P., Lo K. Y.,Fan X., Strauss M. A., 2004, ApJ, 615, L17White R. L., Becker R. H., Fan X., Strauss M. A., 2003, AJ,126, 1Willott C. J., McLure R. J., Jarvis M. J., 2003, ApJ, 587, L15Wu X.-B., et al., 2015, Nature, 518, 512Yaqoob T., 1997, ApJ, 479, 184This paper has been typeset from a TEX/L A TEX file prepared bythe author.MNRAS000