Mon.Not.R.Astron.Soc.000,1–16(0000) PrintedJanuary8,2015 (MNLATEXstylefilev2.2) Exploring the Active Galactic Nuclei population with extreme X-ray to optical flux ratios ( f /f > 50) x o R. Della Ceca1⋆, F.J. Carrera2, A. Caccianiga1, P. Severgnini1, L. Ballo1, V. Braito1, 5 1 A. Corral3, A. Del Moro4, S. Mateos2, A. Ruiz2, and M.G. Watson5 0 1INAF-OsservatorioAstronomicodiBrera,viaBrera28,20121Milano,Italy 2 2InstitutodeFisicadeCantabria(CSIC-UC),AvenidadelosCastros,E-39005Santander,Spain n 3NationalObservatoryofAthens(NOA),PalaiaPenteli,GR-15236Athens,Greece a 4DepartmentofPhysics,DurhamUniversity,SouthRoad,DurhamDH13LE,UK J 5DepartmentofPhysics&Astronomy,UniversityofLeicester,Leicester,LE17HR,UK 7 ] A G ABSTRACT . h p Thecosmichistoryofthegrowthofsupermassiveblackholesingalacticcentersparallels - thatofstar–formationintheUniverse.However,animportantfractionofthisgrowthoccurs o inconspicuouslyinobscuredobjects,whereultraviolet/optical/near-infraredemissionisheav- r t ily obscured by dust. Since the X–ray flux is less attenuated, a high X-ray–to–optical flux s ratio( f /f )isexpectedtobeanefficienttooltofindouttheseobscuredaccretingsources. a x o We explore here via optical spectroscopy, X-ray spectroscopy and infrared photometry the [ most extreme cases of this population (those with f /f > 50, EXO50 sources hereafter), x o 1 using a well defined sample of 7 X-ray sources extracted from the 2XMM catalogue. Five v EXO50sources(∼ 70percentof thesample)in the brightfluxregimeexploredby oursur- 4 vey(f > 1.5×10−13 ergcm−2 s−1 )areassociatedwithobscuredAGN(N > 1022 4 (2−10keV) H cm−2),spanningaredshiftrangebetween0.75and1andcharacterisedby2-10keVintrinsic 4 luminositiesin theQSO regime(e.g.wellin excessto1044 ergs−1 ).We didnotfindcom- 1 0 pellingevidenceofComptonThickAGN.OveralltheEXO50Type2QSOsdonotseemto . bedifferentfromstandardX-rayselectedType2QSOsintermsofnuclearabsorption;avery 1 high AGN/host galaxy ratio seems to play a major role in explaining their extreme proper- 0 ties. Interestingly3 outof 5EXO50Type 2QSO objectscan beclassified as ExtremeDust 5 ObscuredGalaxies(EDOGs, f /f > 2000),suggestingthataveryhighAGN/hostratios 1 24µm R : (along with the large amount of dust absorption) could be the natural explanation also for v a part of the EDOG population. The remaining 2 EXO50 sources are classified as BL Lac i X objects,havingratherextremeproperties,andwhicharegoodcandidatesforTeVemission. r Key words: galaxies: active – galaxies: neclei – BL Lacertae objects: general – Quasars: a general–X-rays:galaxies 1 INTRODUCTION there are now increasing evidence that the formation and growth of galaxies and their nuclear supermassive black holes (SMBHs, The study of high-z obscured quasars (Type 2 QSOs: the high M > 106 M ) are intimately related; feedback from actively luminosity counterpart of Seyfert 2 galaxies) is one of the hot BH ⊙ accreting SMBH, AGN, must play a fundamental role in regu- topics of current extragalactic astronomy. Their observed distri- lating both star formation and accretion throughout galaxy’s his- butions (e.g. redshift, flux and absorption) and physical proper- tory (e.g. Silk&Rees 1998, Granatoetal. 2004, DiMatteoetal. ties (e.g. bolometric luminosity, black hole mass, accretion rate), 2005, Crotonetal. 2006, Hopkinsetal. 2008, Mencietal. 2008, compared to those of unobscured QSOs, are key parameters to King 2010, Faucher-Gigue`re&Quataert 2012). Although the na- understand and to test the orientation-based Unified Schemes ture of this relationship is still poorly understood, there are hints (Antonucci 1993) for Active Galacic Nuclei (AGN) and to con- that most of the SMBHaccretion takes place during an obscured strainthecontributionof QSOstotheX−rayBackground (XRB: quasarphase.Theinfrared(IR),opticalandX−rayspectralanaly- e.g. Gillietal. 2007; Treisteretal. 2009). Even more important, sisofalargesampleofobscuredquasarsprobablyrepresentsone of the best methods to test a possible coevolution between mas- sivegalaxiesandAGNactivity(seee.g.Hopkinsetal.2006)and ⋆ E-mail:[email protected] (cid:13)c 0000RAS 2 DellaCeca et al. toinvestigateifandhowtheAGNfeedbackcanaffectthegalaxy latedsofar(intheradio,IR,optical,andX-rayenergyranges)on evolution(seee.g.Boweretal.2006).Indeed,inthesesourcesthe the7EXO50objectsdiscussedhere,whiletheirproperties,source propertiesofthehostgalaxycanbestudiedusingtheoptical and bysource,arediscussedinSection4.InSection5wecomparethe near-IR data, where the absorbed AGN is supposed to contribute broadband,fromIRtoX-rays,propertiesoftheseEXO50sources marginally, while the AGN properties can be investigated using with other similar objects found in the literature. Summary and theX-rayandmid-IRdatawhere,conversely,theAGNemissionis conclusionsarepresentedinSection6.Throughoutthispaper,we dominant(seeBalloetal.2014andreferencestherein).Obscured consider the cosmological model with (H ,Ω ,Ω )=(70,0.3,0.7); o M λ QSOs,arareclassofobjects,arethusexpectedtohavelargeval- resultsfromotherpapershavebeenrescaledtothiscosmological uesof theX-ray-to-optical fluxratio.Furthermore, sincethedust framework.Alltheoptical/IRmagnitudesreportedhereareinthe extinctionincreasesintheultraviolet(UV)whiletheX-rayabsorp- Vegasystem.Unlessdifferentlyquoted,X-rayluminositiesarein- tiongoes in theopposite direction, i.e. strongly decreasing going trinsic(i.e.unabsorbed)luminositiesintherest-frame2−10keV towardsthehighenergies, aredshift dependence isalsoexpected energy range. In this paper we use the term Type 1 and Type 2 (∼(1+z)3.6;Fioreetal.2003).However,inspiteoftheoutstanding AGN as broad or narrow line (FWHMof the permitted emission progress,obtainedbyusingmediumanddeepX–raysurveys(see lines< 1500 kms−1 ) AGN, irrespective totheir intrinsicX-ray e.g.Brusaetal.2010andreferencestherein),theweakness,bothin luminosity, while weusethe termType 1QSO and Type2 QSO theX-rayandintheopticalband,oftheselectedsourceswithhigh (or obscured QSO) as broad or narrow AGN with an intrinsic 2- valuesoftheX-ray-to-opticalfluxratiousuallypreventsfromade- 10keVluminosityinexcessto1044 ergs−1 .Finally,errorsareat tailedanalysisoftheindividualobjects(seee.g.Koekemoeretal. 90%confidencelevelfortheX-rayspectralparametersderivedus- 2004,Civanoetal.2005). ingXSPEC(asusuallydoneinX-rayastronomy)and68%forall To explore the most extreme examples of obscured QSOs theotherquantities. in the bright flux regime, we have started a project focused on the very “high” X-ray-to-optical flux ratio, f /f (see Eq. 1), x o population, namely those sources with f /f >50 (more than x o 15 times the average values of unobscured broad line AGN, e.g. 2 THEEXO50SAMPLE Caccianigaetal. 2004,DellaCecaetal. 2004,Civanoetal.2012) Inobscured QSOsthenuclear UV/optical emissionissuppressed andwith f >10−13ergcm−2s−1.Inthiswayweshouldefficiently x bydustobscuration(thusleavingonlythegalaxycomponent vis- selectthebestcandidatestobeobscuredQSOs;atthesametime, ible), whilst the nuclear X-ray flux (e.g. in the 2-10 keV energy thesourcebrightnessensuresthattheopticalspectroscopicidenti- range), even if exhibiting signatures of photoelectric absorption, ficationcanbeachieved1 forsourceswithreasonablygoodqual- islessattenuated (seee.g. Fioreetal.2003).Obscured QSOsare ityX-raydata(>fewhundredcounts)tocarryoutareliableX-ray thereforeexpectedtodisplaylargevaluesofX-raytoopticalflux spectralanalysis.HereafterwewillcallthesesourcesEXO50–for ratio,definedhereas: “extremeX-raytoopticalfluxratio”.Oneexampleofsuchextreme obscuredobjectsisXBSJ021642.3-043553with fx/fo∼200,for fx/fo= fx/(∆λ×2.15×10−9×10−0.4×R) (1) which the presence of a Type 2 QSO at z ∼ 2 was spectroscop- ically confirmed with VLT/FORS (Severgninietal. 2006). Using where fxreferstotheobserved2-10keVflux(correctedforGalac- theredshiftinformationandthespectralenergydistribution(SED), ticabsorption),RistheobservedopticalmagnitudeintheRband weestimatedastellarmassof∼1011M forthehostgalaxy,which (λ≃6410Å)and∆λ≃1568Å(seeFukugitaetal.1995). ⊙ supportsthestronglinkbetweenhigh-redshiftmassivegalaxiesand Wedefine EXO50the sources with fx/fo > 50; these area powerfulobscuredhigh-redshiftQSOs.Inspiteoftheimportance rareclassofX-rayemittingobjects,astypicallyAGNhave fx/fo= of obscured QSOs in the cosmological context, only less than a 1-10 (see e.g. Caccianigaetal. 2004, DellaCecaetal. 2004, dozen of such extreme f /f sources have been found and stud- Civanoetal.2012).Forinstance,intheXMM-NewtonBrightSur- x o iedsofarwithgoodX-rayandopticaldata(seee.g.Gandhietal. vey(XBS,DellaCecaetal. 2004,Caccianigaetal.2008),acom- 2006, Severgninietal. 2006, DelMoroetal. 2009, Campisietal. pletesampleofbright(f(0.5−4.5keV) &7×10−14 ergcm−2s−1)X-ray 2009,Brusaetal.2010Brusaetal.2015,Pernaetal.2014). selectedsourcesalmostcompletelyidentified(spectroscopiciden- Herewediscussthefirstresultsobtainedfromthisprojecton tificationlevel∼98%),EXO50objectsrepresentsonly∼ 0.5%of asmall,butstatisticallycompleteandrepresentative,sub-sampleof thesourcepopulation. 7EXO50objectswith f > 1.5×10−13 ergcm−2 s−1 .Itisworth ToconstructoursampleofEXO50sourcesinthebrightflux x mentioning that other two interesting (and rare) classes of extra- regime we used one of the largest well defined and complete X- galacticsourcesareexpectedtoshowupinthehigh f /f domain raysourcesamplederivedsofar(discussedinMateosetal.2008), x o exploredhere,namelyBLLacobjectsandhighredshift(i.e.z>0.6 basedonthe2XMMsourcecatalog2(Watsonetal.2009). in order to have the 4000 Å break shifted at longer wavelengths Firstwehaveconsideredthefollowingselectioncriteria: than the R band filter) clusters of galaxies. Clusters of galaxies, a) “point-like” X-ray sources (parameters EP EXTENT and however,arenotexpectedtoberepresentedinoursamplesincewe EP EXTENT ML inthesourcecatalogue equal to0), inorder to haverestrictedoursearchonlytothepoint-likeX-raysources,thus minimisethepresenceofclustersofgalaxieswhichisanotherpos- stronglyminimisingtheirpossibleselection. sibleclassofhigh fx/fosources; This paper is organized as follows: in Section 2 we discuss b) north of -10 deg declination, in order to be accessible to thestrategy usedto defineastatisticallyrepresentative sampleof opticalinvestigationfromtheItalianTelescopioNazionaleGalileo sourceswith f /f >50.InSection3wepresentthedataaccumu- (TNG), from the Large Binocular Telescope (LBT) and from the x o SpanishGranTelescopioCanarias(GTC); 1 AtthechosenX-rayfluxlimitof fx∼10−13 ergcm−2s−1asourcewith fx/fo∼50(300)hasanopticalRmagnitudeof∼23(∼25). 2 Seehttp://xmmssc-www.star.le.ac.uk/Catalogue/xcat public 2XMM.html (cid:13)c 0000RAS,MNRAS000,1–16 Fx/Fo>50 3 c)sourcesselectedinthe2-10keVenergybandwithaf 2−10keV >10−13 ergcm−2s−1; d)onlyserendipitoussources(e.g.sourcesthatarenotrelated to nearby galaxies or to the XMM-Newton pointing) have been takenintoconsideration.Wehavealsoexcludedtheserendipitous sourcesalreadyclassifiedasnonAGN. Startingfromasourcelistof9431sources(seeMateosetal. 2008), thesefirstselection criteriaprovide uswithalistof about 600X-rayemittingobjects. The second step was to select the X-ray sources with f /f >50,using: x o i)opticalsourcearchives(e.g.APM,SloanDigitalSkySurvey -SDSS,etc); ii) optical imaging data from archives (STScIDigitized Sky Survey,ESOarchive,etc..); iii)imagingdatafromourdedicatedobservingprograms(e.g. TNGimaging). Wehavelooked atandvisuallyinspectedalltheselectedX- raysourcesforpossibleopticalcounterparts,usingasearchradius of4arcsec(seeCaccianigaetal.2008),andderivedtheirRmag- nitude.Forallthesourcesreportedanddiscussedhere(seeTable 1), the offset between the X-ray and the optical position, derived a posteriori, is below 2.1 arcsec, fully consistent with the results >Fig1u.5r×e11.0H−1i3stoegrgracmmo−f2tsh−e1(f3x/10foodbijsetrcitbsu).tiWoneoaflstoheshsotawrtiansgfislalemdphleiswtoigthramfx obtainedbyCaccianigaetal.(2008)fromtheanalysisoftheX-ray the fx/fodistributionofthe7EXO50sourcesdiscussedinthispaper. toopticaloffsetfortheX-raysourcesbelonging totheXBS.The accumulated magnitudes, combined with the observed 2-10 keV fluxes,havebeenusedtocomputethe f /f ratioandthustoselect x o EXO50sources. reductionprocessusingIRAF2.14.Allimageswerede-biasedand Inthispaper weconsider afirstcompletesampleofEXO50 sourceshaving f >1.5×10−13 ergcm−2s−1;thissampleiscom- correctedforflat-fieldeffects.Wecombinedthedifferentexposures x foreachsourceandtheresultingimageswerefluxcalibratedusing posed by the 7 EXO50 objects reported in Table 1. In Figure 1 standard stars and the standard extinction curve for the observa- we show the f /f distribution of the sample composed by all x o tory.Finally,wedidanastrometriccalibrationofeachimage;the theserendipitous point-likesources inthecovered areahaving f x >1.5×10−13 ergcm−2s−1(i.e.thestartingsample,310objects).In astrometricaccuracyofourimagingistypicallybetterthan∼ 0.3 arcsec.TheRmagnitudeswereestimatedwithIRAFroutinesfol- thesamefigureweshowhowextremethe7sourcesarewithrespect lowingthestandardprocedureofaperturephotometrycenteredat tothestartingsample;theEXO50sourcesdiscussedhererepresent only∼2%ofthepopulationconsidered.Westressthataccording thesourceposition. to the selection criteria discussed above and the optical material atourdisposalthese7objectsaretheonlyEXO50sources,north of-10degdeclination,with f > 1.5×10−13 ergcm−2 s−1 inthe x sky survey areadefined inMateosetal. (2008); thesky coverage 3.2 Opticalspectroscopyandidentification investigatedtofindoutthese7objectsisabout60.4sqdeg. The optical spectra for 6 of our sources were taken at GTC dur- As detailed in the following sections, out of the 7 EXO50 ing 3 different observing runs (see Table 2) using the R500R sources 2 are BL Lac objects, 3 are confirmed Type 2 QSO and grism on the OSIRIS instrument 3. For the reduction we used 2remainspectroscopicallyunidentified,althoughtheirbroadband thestandardIRAFlong-slitpackagefollowingthestandardsteps. properties strongly suggest an obscured QSO nature. Using the The resulting spectra (see Figure 2 middle panels, in arbitrary JAVAsoftwareGALAXYCOUNT(Ellis&Bland-Hawthorn2007) units) give us spectroscopic information (classification and red- and considering the optical magnitudes of the proposed optical shift)for4outofthe6GTCsources.Fortheremaining2sources counterparts(seeTable1)weestimatelessthan0.3normalgalaxies (2XMMJ100038.9+050955 and 2XMMJ135055.7+642857) the bychanceinthetotalareacoveredfromthe7errorboxes,assuring spectra are inconclusive and do not allow us to derive any rele- usaboutthereliabilityofthespectroscopicidentificationproposed vantinformation. Thesource2XMMJ121026.5+392908 isawell here. known object and its redshift and classification have been taken from the literature. In total we have spectroscopically identified (usingtheGTCdataaswellasdatafromtheliterature)5EXO50 objects;twosourcesremainstillspectroscopicallyunidentified,al- 3 OBSERVATIONSANDDATAREDUCTION thoughthedataatotherwavelengthsdiscussedinsection4and5 3.1 Opticalimaging stronglysuggestanobscuredQSOnature. InFigure2(leftpanels)wepresenttheopticalfindingchartsofthe 7 sources discussed here (1x1 arcmin) produced using data from dedicatedTNGobservingruns(Rfilter)orfromtheSDSS(rfilter). TNG rawimages werereduced following thestandard CCD 3 Seehttp://www.gtc.iac.es/instruments/osiris/osiris.php (cid:13)c 0000RAS,MNRAS000,1–16 4 DellaCeca et al. 2XMM J022256.9-024258 67889"666:;$<"6=6:> ✂ -2°42'36.0" ! /4-5!✁ "$"! Dec (J2000) 43'0408..00"" %&’()*+,-./0&1%23/3✁ !"(cid:0)# 6 12.0" ?)2+& !$:! 24.0" "$: 6 : 58.0s 57.0s 56.0s 2h22m55.0s @%-’AB/C4-5D RA (J2000) 2XMM J100038.9+050955 24.0" $9::;!"""#<=>?"’">’’ ! 12.0" 2708!✁ $%!"(cid:0)# Dec (J2000) 10'4080..00"" ()*+,-./0123)4(5626✁ ’$%%!!!"""(cid:0)(cid:0)(cid:0)&&# $=’ +5°09'36.0" @,5.) !=’$ ! "=’ $ ’ 40.0s 39.0s 38.0s 10h00m37.0s A(0*BC2D708E RA (J2000) 2XMM J121026.5+392908 36.0" 67889!6!"6:$;<#=6="> ! /4-5!✁ ! Dec (J2000) 2142..00"" %&’()*+,-./0&1%23/3✁ "!!"$""$$(cid:0)?!!# 29'00.0" !$6 @)2+& ! +39°28'48.0" "$> "$; ! 6 ; 29.0s 28.0s 27.0s 26.0s 25.0s12h10m24.0s A%-’BC/D4-5E RA (J2000) Figure2. Left panels: Optical finding chart (1x1arcmin) forthe7EXO50objects reported inthis paper produced usingmaterial fromdedicated TNG observingruns(Rfilter;sources:2XMM022256.9-024258,2XMM100038.9+050955,2XMM123204.9+215254,2XMM135055.7+642857)orfromSDSS (rfilter;sources:2XMM121026.5+392908,2XMM121134.2+390054,2XMM143623.8+631726). Thecircleis4arcsecradiusaroundtheX-rayposition. Middlepanels:TheopticalspectraoftheEXO50objectsobservedattheGTC.Thesolidblacklineinthespectroscopicallyidentifiedobjectsrepresentsthe underlinegalaxycomponentusedtoreproducetheopticalspectrum,exceptfor2XMMJ121134.2+390054,forwhichweshowthegalaxycomponent(blue line),theAGNcomponent(continuumplusbroademissionlines,redline)andthecombinedcomponent(galaxyplusAGN;blackline).Thespectrallinesused toclassifytheobjectsarealsomarked.2XMMJ121026.5+392908isawellknownobjectanditsredshiftandclassificationhasbeentakenfromtheliterature. Rightpanels:X-raydata(observerframe)andresidualsfortheEXO50objectsdiscussedhere.Alltheobjectsarewellfittedwithanabsorbedpower-law model.Redfilledtriangles:pndata.Blackopencircles:MOSdata. (cid:13)c 0000RAS,MNRAS000,1–16 Fx/Fo>50 5 2XMM J121134.2+390054 24.0" 67889!6!!#:$6;#<""=: ! 5✁ 12.0" /4-! "$! Dec (J2000) 01'4080..00"" %&’()*+,-./0&1%23/3✁ "!$""(cid:0)!# 6 +39°00'36.0" >)2+& !$=! "$= "$= ! 6 = 36.0s 35.0s 34.0s 33.0s12h11m32.0s ?%-’@A/B4-5C RA (J2000) 2XMM J123204.9+215254 24.0" 67889!6#6":$;<6!=6=: ! 5✁ 12.0" /4-! "$"! Dec (J2000) 53'4080..00"" A B %&’()*+,-./0&1%23/3✁ !"(cid:0)# 6 +21°52'36.0" >)2+& !$=! "$= ! 6 = 07.0s 06.0s 05.0s 04.0s 12h32m03.0s ?%-’@A/B4-5C RA (J2000) 2XMM J135055.7+642857 24.0" 67889!#:"::$;<=>6?:; ! 5✁ 12.0" /4-! Dec (J2000) 29'00.0" %&’()*+,-./0&1%23/3✁ "!$""(cid:0)!# 48.0" 6 +64°28'36.0" @)2+& !$:! "$: ! 6 : 51m00.0s 58.0s 56.0s 54.0s13h50m52.0s A%-’BC/D4-5E RA (J2000) 2XMM J143623.8+631726 7899:!#$;7$%<=;$!>7; 48.0" 6!✁ 05.! "%"! Dec (J2000) 2346..00"" &’()*+,-./01’2&3404✁ !!""(cid:0)(cid:0)#$ 7 12.0" @*3,’ !%?! "%? +63°17'00.0" "%? ! 7 ? 28.0s 26.0s 24.0s 22.0s 14h36m20.0s A&.(BC0D5.6E (cid:13)c 0000RAS,MNRAS000R,A1 –(J126000) 6 DellaCeca et al. Table1.ThecompletesampleofEXO50objectsdiscussedinthispaper n 2XMMName fx R fx/fo ID z NaH Γa Lx fr 10−13cgs 1022cm−2 1045cgs mJy (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) 1 J022256.9-024258 3.50 22.1±0.4 72 AGN2 1.004 7.5+3.2 1.80+0.4 1.9 429±14 −2.9 −0.4 2 J100038.9+050955 1.63 23.6±0.4 138 - - >0.4b 1.9c - - 3 J121026.5+392908 36.1 19.16d±0.02 50 BLLac 0.617 0.06+0.004 2.23+0.01 6.4 19.0±0.7 −0.002 −0.01 4 J121134.2+390054e 8.2 20.77d±0.05 50 BLLac 0.890 0.19+0.04 2.21+0.07 3.7 10.6±0.6 −0.04 −0.06 5 J123204.9+215254e 10.0 23.9±0.4 1118 AGN2 0.763 3.61+1.03 1.31+0.25 1.9 - −0.92 −0.21 6 J135055.7+642857 1.5 25.0±0.8 458 - - >0.1b 2.02+0.13 - 183.5±5.5 −0.08 7 J143623.8+631726 2.53 22.2d±0.2 55 AGN2 0.893 1.46+0.39 1.69+0.15 0.84 - −0.24 −0.11 Columnsareasfollows:1)NumberusedtomarktheobjectintheplotsshowninSection5;2)sourcename;3)X-rayfluxinthe2-10keVenergyband(MOS normalisation)correctedforGalacticabsorptioninunitsof10−13 ergcm−2s−1;4)Rbandmagnitude;5)X-raytoopticalfluxratio;6)opticalspectroscopic classification;7)redshift;8)intrinsicabsorptionand90%confidenceintervalinunitsof1022cm−2fromfitstoX-rayspectra;9)powerlawphotonindex,Γ, intheX-rayenergyrange,and90%confidenceinterval;10)intrinsicX-rayluminosityinthe2-10keVrestframeenergyrange,inunitsof1045 ergs−1;11) radiofluxat1.4GHzinmJy.Notes:a-ForallthesourcesthebestfitspectralmodelintheX-rayenergybandisasimpleabsorbedpower-lawmodel;b-The 90%lowerlimitonNHobtainedassumingz=0;c-Thisparameterhasbeenfixed;d-ForthesesourcestheRbandmagnitudehasbeenobtainedfromthe SDSSr”model”magnitudesassumingr-R=0.27(rSDSS AB=RVega+rAB(Vega)withrAB(Vega)=0.27);e-Forthesetwosourcesweproposeherea differentredshiftwithrespecttothatalreadyreportedintheliterature.SeeSection4fordetails. Table2.Opticalfollow-upspectroscopicobservations 2XMMName Programme Instrument/Grism Exposure(s) SlitWidth(′′) Dates (1) (2) (3) (4) (5) (6) J022256.9-024258 GTC18-10B OSIRIS/R500R 6×1100 1.2 2-Sep-2010(4),19-Sep-2010(2) J100038.9+050955 GTC44-11A OSIRIS/R500R 4×1800 1.0 10-Apr-2011 J121134.2+390054 GTC18-10B OSIRIS/R500R 2×600 1.2 27-Jan-2011 J123204.9+215254 GTC44-11A OSIRIS/R500R 2×1800 1.0 10-Apr-2011(1),11-Apr-2011(1) J135055.7+642857 GTC44-11A OSIRIS/R500R 3×1800 1.0 10-Apr-2011 J143623.8+631726 GTCMULTIPLE2-09A OSIRIS/R500R 7×1500 1.2 17-Apr-2009(1),19-Apr-2009(3),25-May-2009(3) Columnsareasfollows:1)Sourcename;2)programID;3)instrument/grismused4)numberofexposuresandexposuretimesinseconds;5)slitwidthin arcsec;6)datesoftheobservations.Inparenthesiswehaveindicatedthenumberofexposures. 3.3 WISEAll-SkySurveydata and 3.9σ respectively. Two sources, 2XMMJ121026.5+392908 and2XMMJ121134.2+390054,arenotdetectedintheW3andW4 The Wide-field Infrared Survey Explorer (WISE; Wrightetal. bands. 2010)hasrecentlycarriedoutanallskysurveyinthemediumin- Using a simulated sample withrandom positions of ∼ 1600 frared regime, detecting hundreds of millions of objects. The es- object we estimated that the probability to have a random WISE timated 5σ point sources sensitivities (in unconfused regions) in sourceinsideacircleof 2arcsecradiusis∼ 1%(i.e.0.07 WISE thefourobservedchannels(W1=3.4µm,W2=4.6µm,W3=12µm sourcesexpectedbychanceinthe7errorcirclesinvestigatedhere), and W4=22 µm) are better than 0.08, 0.11, 1 and 6 mJy, while implying that all the detected WISE counterparts are very likely the angular resolution (FWHM) are 6.1, 6.4, 6.5, 12 arcsec re- associatedwiththehigh f /f sources. spectively; theastrometricprecisionforhighsignal-to-noise ratio x o Fluxdensitiesat3.4µm,4.6µm,12µmand22µmhavebeen (SNR)sourcesisbetterthan0.15arcsec.ToaddtheWISEinfor- computed from the magnitudes reported in Table 3 by assuming mationtoour studywehavecross-correlated our EXO50sample themagnitude zeropointsof theVegasystemcorresponding toa with the WISE All-Skysource catalog 4 using a positional toler- power-lawspectrum(f ∝ν−α)withα=1(seeWrightetal.2010). ance, from the X-ray source position, equal to 4 arcsec . All our ν Thedifferencesinthecomputedfluxdensitiesexpectedusingflux sources have been detected by WISE (see below) and a posteri- correctionfactorsthatcorrespondtoa∆α=±1arelowerthan0.8 ori the offset between the WISE and the optical position of our percent,0.6percent,6percent,and0.7percentintheW1,W2, sourcesisbelow2arcsec.Theresultsofthiscross-correlationare W3andW4band,respectively(Wrightetal.2010). reported in Table 3. We found a single WISE counterpart for all the 7 sources discussed here. The detections reported in Table 3 have a significance greater than 7σ in all bands, with the excep- 3.4 XMM-Newtonspectroscopy tionof2XMMJ100038.9+050955,2XMMJ123204.9+215254and 2XMMJ143623.8+631726,detectedintheW4bandat6.1σ,3.6σ InTable4wereportdetailsfortheXMM-Newtondatausedforthe X-rayspectralanalysisofeachsourcediscussedhere.TheXMM- NewtondatawerecleanedandprocessedwiththeXMM-Newton 4 We use here the public available All-Sky Data Release ScienceAnalysisSoftware(SAS)andwereanalyzedwithstandard that covers >99% of the sky (March 2012 release; see softwarepackages(FTOOLS;XSPEC,Arnaud 1996).Eventfiles http://wise2.ipac.caltech.edu/docs/release/allsky/) produced from the pipeline were filtered from high-background (cid:13)c 0000RAS,MNRAS000,1–16 Fx/Fo>50 7 Table3.Infrared(WISE)datafortheEXO50sourcesdiscussedhere. n 2XMMName fx/fo ID(z) W1 W2 W3 W4 Log(νLν) (12.3µm) (1) (2) (3) (4) (5) (6) (7) (8) (9) 1 J022256.9-024258 72 AGN2(z=1.004) 16.02±0.06 14.89±0.07 11.08±0.09 8.07±0.15 45.61 2 J100038.9+050955 138 - 16.22±0.08 14.29±0.05 10.33±0.06 7.97±0.18 - 3 J121026.5+392908 50 BLLAC(z=0.617) 14.86±0.04 14.59±0.06 >12.34 >8.66 - 4 J121134.2+390054 50 BLLAC(z=0.890) 15.21±0.04 15.03±0.08 >12.50 >9.45 - 5 J123204.9+215254 1118 AGN2(z=0.763) 15.35±0.05 13.84±0.04 10.19±0.05 8.48±0.30 45.10 6 J135055.7+642857 458 - 14.90±0.03 13.85±0.03 10.46±0.04 8.47±0.15 - 7 J143623.8+631726 55 AGN2(z=0.893) 15.09±0.03 14.05±0.04 11.36±0.09 8.91±0.28 45.10 Columnsareasfollows:1)NumberusedtomarktheobjectintheplotsshowninSection5;2)sourcename;3)X-raytoopticalfluxratio;4)optical spectroscopicclassificationandredshift;5)W1magnitude(3.4µm)and1σerror;6)W2magnitude(4.6µm)and1σerror;7)W3magnitude(12µm)and1σ error.Themagnitudelowerlimitscorrespondtothe95%confidencelevelasreportedintheWISEcatalog;8)W4magnitude(22µm)and1σerror.The magnitudelowerlimitscorrespondtothe95%confidencelevelasreportedintheWISEcatalog;9)Logoftherestframe12.3µmluminosities (ν12.3µm×L12.3µmin ergs−1)computedasdescribedinSection3.3. Table4.EPICXMM-Newtonobservationdetails. Name OBSID NH Instr. Filter NetExp.Time NetCounts (1) (2) (3) (4) (5) (6) (7) J022256.9−024258 0037981601 2.34MOS1&2 thin 26.1 265 pn thin 7.9 127 J100038.9+050955 0204791101 2.41 MOS1 med 16.1 27 pn med 12.5 63 J121026.5+392908 0112830501, 2.00MOS1&2 med 161.5 119876 0112830201 pn med 68.9 175724 J121134.2+390054 0112190201 1.84MOS1&2 med 27.8 2311 pn med 8.9 2488 J123204.9+215254 0112650301 1.80 MOS1 thin 19.3 248 MOS2 med 19.5 250 pn thin 13.1 447 J135055.7+642857 0147540101 1.70MOS1&2 med 44.6 873 pn med 18.9 1064 J143623.8+631726 0204400301 1.37MOS1&2 med 46.7 862 pn med 12.7 648 Col.(1):2XMMsourcename;Col.(2):OBSIDoftheXMM-Newtonobservation;Col.(3):Galacticabsorbingcolumndensityalongthelineofsightinunits of1020cm−2;Col.(4):EPICinstrument;Col.(5):EPICfilter;Cols(6):Exposuretimeafterremovinghigh-backgroundintervals,inunitsofksec;Col.(7): Netcountsintheenergyrange0.3−10keV. timeintervals and only events corresponding to patterns 0-12 for foreachsourcewerecombinedbyusingthetasksaddrmfandad- MOSand0-4forpnwereused.Allspectrawereaccumulatedfrom darf.Forallthesourcesbut2XMMJ100038.9+050955wegrouped acircularextractionregionwitharadiusof20-30arcsec,depend- thespectrainbinscontainingmorethan20(source+background) ingonthesourceoff-axisdistance.Backgroundcountswereaccu- counts and used the χ2 minimization technique; in the case of mulatedinnearbycircularsourcefreeregions,usinganareausu- 2XMMJ100038.9+050955wegroupedthespectrainbinscontain- allyaboutafactorof4largerthantheoneusedtoextractthesource ing 10 (source+background) counts and use the Cash statistics. counts. The X-ray spectra usually cover the 0.3-10 keV energy Inthecaseof 2XMMJ121026.5+392908 twoXMMobservations range; the total (MOS1+MOS2+pn) net (background subtracted) wereused;pn(MOS)datafromthetwoindependentdatasetwere counts range from ∼ 102 to ∼ 3×105 counts. The ancillary re- combinedtogetherandtheancillaryanddetectorresponsematrices sponsematrixandthedetectorresponsematrixwerecreatedbythe were created using the same procedures quoted above. We fitted XMM-SAStasksarfgenandrmfgenateachsourcepositioninthe pnandMOSspectrasimultaneouslyinthe0.3-10keVband,tying EPIC detectors. To improve the statistics, the MOS1 and MOS2 togetherallpnandMOSparametersexceptforarelativenormal- spectraobtained byusing thesame filterwerecombined aposte- ization, which accounts for the differences between pn and MOS riori by using the FTOOLS task mathpha; in this case ancillary fluxcalibrations(seeMateosetal.2009).Inthefollowing,derived anddetectorresponsematricesfortheMOS1andMOS2detectors fluxesandluminositiesrefertotheMOSinstrument. (cid:13)c 0000RAS,MNRAS000,1–16 8 DellaCeca et al. For the spectral modelling we considered a simple ab- sorbed power-law model that takes into account both the Galac- tic hydrogen column density along the line of sight (from Dickey&Lockman 1990)andapossibleintrinsicabsorptionatthe source redshift (abundances relative to the Solar one as reported in Wilmsetal. 2000). In the X-ray spectral modelling we made use,whenavailable,oftheredshiftsobtainedfromtheopticalspec- troscopy.AlltheX-rayspectraarewellfittedbythissimplemodel; theresultsarereportedinTable1alongwiththecorrespondingflux (corrected for Galactic absorption) and intrinsic luminosity (i.e. corrected for both Galactic and intrinsic absorption) in the stan- dard2-10keVenergyband.TheX-rayspectraareshowninFigure 2(rightpanels).Inthecaseof2XMMJ100038.9+050955,thespec- tralqualitydoesnotallowustoconstrainthepower-lawphotonin- dexandtheintrinsicabsorptionatthesametime,sowehavefixed the power-law photon index Γ to 1.9, a common value for unab- sorbedAGN(Mateosetal.2010,Corraletal.2011,Lanzuisietal. 2013). 1 arcmin 4 RESULTS Beforediscussingthepropertiesofeachsingleobjectinthissam- ple we summarise here their main properties. From optical spec- Figure3.Radio(FIRST)contourplotssuperimposedtoaCCDRimage troscopy, out of the 7 EXO50sources 2 are classified asBL Lac (takenattheTNG)of2XMMJ022256.9−024258; theredhatchindicates objects,3areclassifiedasType2QSOand2remainunidentified. theobjectresponsiblefortheX-rayemission.Theradiocontoursstartfrom The 3sources classifiedas Type 2 QSO arein the redshift range 0.2mJywithamultiplicativestepof∼1.58.Thelargebluebaratthebottom 0.7−1 and are characterised by an intrinsic X-ray absorbing col- is1′wide:attheredshiftofthesource1′correspondsto∼0.48Mpc.North umndensity,N ,intherangebetween1.5and8×1022 cm−2.For isupwards,Eastisontheleft. H thetwounidentifiedobjectstheX-rayanalysisprovideslowerlim- itstotheintrinsicN of4×1021cm−2and1021cm−2,respectively. H TherearenoComptonThickAGN(N inexcessto∼ 1024 cm−2) H within the X-ray error circle (see Figure 2). The optical spec- amongstthe3Type2QSO,neitherissuspectedtheirpresencein trum of this object shows several narrow (observed FWHM < the2stillspectroscopicallyunidentifiedsources,sincetheirX-ray 1000 − 1500 kms−1 ) emission lines that we associate with spectra are at odds withthat usually observed in Compton Thick MgIIλ2798,[NeV]λ3346,3426,[OII]λ3728,[NeIII]λ3869,H ,H , AGN (e.g. the presence of a prominent Iron line at 6.4 keV, rest ǫ δ Hγ,[OIII]λ4364andHeIλ4471atz= 1.004.Thepermittedemis- frame,oraveryflatX-rayspectrum).The2BLLacobjectsareat z=0.62andz=0.89and,asitwillbediscussedinSection5.7,they sionlines(HγandMgIIλ2798)havewidthssimilartotheforbidden ones.TheHγhasaFWHMof550-600km/sthatisclosetothein- areratherextremeintheirSED. strumental resolution. TheMgIIλ2798 has astrongly asymmetric For the three EXO50 sources spectroscopically identified as profile,whichisdifficulttoanalysewiththepresentdata,beingthe Type 2 QSOs (see below), we report in Table 3 the rest frame lineattheborderofthesampledwavelengthrange;theFWHMof 12.3 µm luminosity obtained by interpolating the observed lumi- thislineislikelybelow1500km/s.Itisworthnotingthepresence nosity(i.e.notcorrectedforreddening)intheW3(observedframe, in this object of the [NeV]λ3426 line, a reliable signature of nu- 12 µm) and W4 (observed frame, 22 µm) WISE bands and us- clearactivitysinceitcannotbeproducedinstarburst/starforming ing the measured spectral index between 12 µm and 22 µm; the galaxies(seee.g.Gillietal.2010,Mignolietal.2013).TheX-ray same spectral index has been used to evaluate the K-correction. spectrumiswelldescribedbyanabsorbedpower-lawmodelwith Gandhietal. (2009), studying a sample of Type 1 and Type 2 anN of∼ 7.5×1022 cm−2 ;theintrinsic,rest-frame,2-10keV Seyferts, showed that the observed luminosity around 12.3 µm H rest frame (ν × L ) should represent an accurate proxy luminosityis∼1.9×1045 ergs−1.BasedontheopticalandX-ray 12.3µm 12.3µm spectral properties (optical line widths, intrinsic N and intrinsic for the AGN intrinsic power; in powerful AGN, the contribution H 2-10 keV luminosity in excess to 1044 ergs−1 ) we classify this expected from the host galaxy at these wavelengths is marginal sourceasaType2QSO.2XMMJ022256.9−024258 isdetectedin (< 10%;Balloetal.2014).Overallthederivedinfraredluminosi- alltheWISEbands(seeTable3). tiesareintherangetypicalofLuminousInfraredGalaxies(LIRG: L > 1011L ) and Ultra Luminous Infrared Galaxies (ULIRG: 2XMMJ022256.9−024258 also has strong radio emission as- IR ⊙ L >1012L ,seee.g.Sanders&Mirabel1996).Assumingabolo- sociated with the radio source 4C-02.11. The total 1.4 GHz ra- IR ⊙ metriccorrectionfactortothe12.3µmluminosityof∼10.8,appro- diofluxfromtheFIRSTsurvey(Whiteetal.1997)is∼ 340mJy corresponding to a radio power of ∼ 1.7×1034 erg s−1 Hz−1 at priateforhighluminosityAGN(seeBalloetal.2014)theimplied z=1.004. The source is also detected in the lower spatial reso- bolometric luminosities for the three spectroscopically identified Type2QSOareintherange1.5−4.5×1046 ergs−1. lution NVSS survey (Condonetal. 1998). The total NVSS radio fluxdensity(power) is429±14 mJy(2.18±0.07×1034 ergs−1 • 2XMMJ022256.9−024258; f /f =72 Hz−1); the radio flux density from the NVSS is a more reliable x o Asinglesource(R=22.1)isevidentintheopticalfindingchart measure of the total radio flux, since the VLA-B configuration (cid:13)c 0000RAS,MNRAS000,1–16 Fx/Fo>50 9 used for the FIRST survey could miss some of the diffuse ex- BLLacclassification(Stockeetal.1991;Caccianigaetal.1999). tended emission. For this source we also found in the literature The measured α (< 0.8) suggests a high-frequency-peaked BL RX fluxes at 178 MHz (f = 2.6±0.65 Jy; Goweretal. 1967) Lac object (HBL) in which the low-energy component of their 178MHz and 4.85 GHz (f = 0.116±0.012 Jy; Griffithetal. 1995); usual,double-peak, SEDpeaksbetweentheUVbandandX-rays 4.85GHz assuming a power-law model the derived radio spectral index is (Padovani&Giommi1995).TheobservedX-rayspectrumiswell ∼0.94,consistentwiththetypicalvaluesoflobe-dominatedAGN describedbyafeaturelesspower-lawmodel(Γ=2.23+0.01)within- −0.01 (Kellermann&Owen1988). trinsicN =6×1020 cm−2andintrinsic2-10keVluminosityof∼ H InFigure3weshowtheradiointensitycontoursfromtheFIRST 6.4×1045 ergs−1.IntheWISEsurvey2XMMJ121026.5+392908 radio survey overlaid on the optical image; the radio counterpart isdetected only at 3.4 and 4.6 µm (WISEband W1 and W2, re- of 2XMMJ022256.9−024258 isadouble radiosource havingthe spectively). typical Fanaroff-Riley Type II morphology (i.e. sharp edge lobes • 2XMMJ121134.2+390054; f /f =50 x o andbrighthotspots);theabsoluteopticalmagnitude(M ≃ −24, The optical object clearly visible at the centre of the R assumingaK-correctionof∼2mag,typicalofalatetypegalaxy, error circle (see Figure 2) has a magnitude R=20.77. see Fukugitaetal. 1995) and the total radio power (∼ 2.2×1034 2XMMJ121134.2+390054 is a well known object since the ergs−1 Hz−1)areconsistentwiththeFRIIclassificationaccording epoch of the Einstein Extended Medium Sensitivity Survey tothedividinglineintheM −L planebetweenFRIandFRII (Stockeetal. 1991) and it is classified as a BL Lac object R radio radiogalaxies(seeGhisellini&Celotti2001). (MS1209+3917; Rectoretal. 2000). The observed optical mag- Theprojectedseparationofthetwobrighthotspotsis∼41arc- nitude,radio(10.6±0.6mJyat1.4GHzfromNVSS)andX-ray seconthesky,correspondingtoaphysicalprojectedsizeof∼0.33 (∼ 8.2 × 10−13 ergcm−2 s−1 ) fluxes imply multi-wavelength Mpcattheredshiftofthesource.Giventhediscussedradioprop- spectralindices(α =0.58;α =0.62;α =0.56)fullyconsistent RX OX RO erties,theobjectisclearlyaradio-loudandlobe-dominatedAGN. withtheBLLacclassification(Stockeetal.1991;Caccianigaetal. Inthisrespect,wenotethatthemeasuredsizeandtotalradiolumi- 1999). As for 2XMMJ121026.5+392908 discussed above, also nosityarefullyconsistentwiththatobservedinotherradioquasars 2XMMJ121134.2+390054canbeclassifiedasaHBLobject. (seeFigure4inKuz´micz&Jamrozy2012). FromtheanalysisoftheopticalspectrumshowninFigure2we Considering the broad band properties discussed above, reachtheconclusionthattheonlyclearfeatureisthatat∼7048Å, 2XMMJ022256.9−024258 can be considered another example mostlikelyassociated with[OII]λ3727Åatz=0.89; weestimate of the rare class of Radio Loud Type 2 QSO, with overall anobservedequivalentwidth(EW)of∼12−15Å,thatrescaledto properties very similar to those shown by e.g. AXJ0843+2942 z=0implyanEW∼6−8Å,veryclosetothelimitusedtoclassify (DellaCecaetal. 2003), 6C0905+39 (Erlundetal. 2008) or asourceasaBLLacobject(EW<5Å,Stockeetal.1991). 4C+39.29 (Gandhietal. 2006). Indeed according to the data re- While our spectroscospic classification as a BL Lac object portedinErlundetal.(2008,seetheirFigure3)thisobjectisone agrees with previous results, the redshift proposed here (z=0.89) of the most powerful sources currently know; its intrinsic X-ray issignificantlydifferentfromthatproposedbyRectoretal.(2000) luminosity(∼2×1045 ergs−1),ifcomparedwithits178MHzlu- (z=0.602).Theselatterauthorsdidnotfindanyevidenceofemis- minosity(∼1.3×1035ergs−1Hz−1),ismorethanafactor10above sionline(s)intheirspectrum,andbasedtheirtentativeredshiftde- thelocussampledbytheNarrowLineRadioGalaxiesatz<1.0in terminationonlowSNRabsorptionfeatures.Weareconfidentthat the3CRRcatalogandsimilartothatobservedinBroadLineRadio thelineobservedat∼7048Åisrealandthislinedoesnothaveany GalaxiesandQSOs(seeFigure3inHardcastleetal.2009). reliableidentificationiftheobject isatz=0.602. Moreover at the • 2XMMJ100038.9+050955; f /f =138 proposed redshift (z=0.89) we were able to reproduce quite well x o The optical magnitude of the single object in the error circle alsotheshapeoftheunderlyingopticalcontinuum(seeFigure2). (see Figure 2) is R=23.6. The gathered optical spectrum, unfor- Finallywenotethatthissourceisstronglyvariable:atleastafac- tunately characterised by a very low SNR, seems to be rather torof5.5intheX-raydomain(bycomparingthemeasured0.1-2.4 flat and featureless, thus we have no redshift information for keVfluxwiththefluxintheROSATAllSkySurvey)andafactor 2XMMJ100038.9+050955.ThesourceisdetectedinalltheWISE of3intheopticaldomain(bycomparingourmagnitudewiththe bandswhilethereisnoradiodetectionatthesourceposition.As- V =20 reported in Rectoretal. 2000), so this variability could mag suming that 2XMMJ100038.9+050955 is an absorbed QSO and explainthecleardifferent shapefromtheopticalspectrareported usingtherelationbetween f /f andtheintrinsic2-10keVlumi- inRectoretal.(2000)andthatreportedhere(seeFigure2). x o nosity(Brusaetal.2010,seeSection5)wecanestimatearedshift At z=0.89 theobserved X-rayspectrum iswell described by a z∼1.0(intherange0.6−1.6takingintoaccountascatterof0.5dex featureless power-law model (Γ = 2.21+0.07) withintrinsic N = −0.06 H intherelation,seeFigure4,leftpanel).Atz≃1.0theintrinsic2-10 1.9×1021 cm−2;theintrinsic2-10keVluminosityis∼3.7×1045 keV luminosity would be ∼ 9×1044 ergs−1 while the intrinsic ergs−1 . The intrinsic absorption is significantly in excess to the N ∼5×1022cm−2. Galacticvalueandthisisatoddswithwhatisusuallyobservedin H • 2XMMJ121026.5+392908; f /f =50 BLLacsand,inparticular,inHBLobjects(seee.g.Massaroetal. x o ThissourcewastheonlyonenotobservedduringourGTCrun;it 2011andreferencestherein).OntheotherendtheX-rayspectrum isawellknownBLLacobjectatz=0.617(Caccianigaetal.2002; could be intrinsically curved due to the bump expected in HBL, Plotkinetal. 2010). The observed optical magnitude (R=19.16), peakingbetweentheUVandX-raybands.Wethereforefittedthe the radio (19±0.7 mJy at 1.4 GHz from the NVSS) and the X- X-raydatawithabrokenpower-lawmodelfilteredbytheGalactic ray (∼ 3.6 × 10−12 ergcm−2 s−1 in the 2-10 keV band) fluxes absorption; wefindagood fit(χ2 =1.1for194d.o.f.)withbest- ν allow us to compute multi-wavelength spectral indices 5 (α = fittingparametersΓ =1.57+0.1,Γ =2.30+0.10 andE =1.23+0.2 , RX 1 −0.2 2 −0.13 c −0.28 0.54; α =0.61; α =0.50) which are fully consistent with the OX RO −log(f5G5H.z3/8f2500Å) where f5GHz, f2keV and f2500ÅaretheK-correctedfluxes 5 αRX = −log(f5G7H.6z8/f2keV); αOX = −log(f2k2e.V60/5f2500Å); αRO = at5GHz,2keVand2500Å,respectively. (cid:13)c 0000RAS,MNRAS000,1–16 10 DellaCeca et al. which are consistent with the results reported in Massaroetal. non-transient spectroscopically identified source with the highest (2011)onasampleofHBLsources.Theintrinsic2-10keVlumi- f /f discoveredsofar. x o nosityobtained withthebroken power-lawmodelisequaltothat • 2XMMJ135055.7+642857; f /f =458 x o previouslyreportedinTable1.Inanalogywiththeotherconfirmed A very faint object (R=25.0) is visible at the centre of the er- BL Lac object in the sample (2XMMJ121026.5+392908), in the rorcircle.Unfortunatelytheopticalspectrumisverynoisyandno WISEsurveythissourceisdetectedonlyat3.4and4.6µm. featuresareclearlydetected; wehavenoredshift information for thisobject.TheX-rayspectrumisdescribedbyapower-lawmodel • 2XMMJ123204.9+215254; f /f =1118 with photon index Γ ∼ 2 and absorbing column density greater x o Twoopticalsources(labeledwithAandBinthefindingchart than∼2×1021 cm−2(obtainedassumingz=0).Anemissionline reportedinFigure2)areclearlypresentinsidetheX-rayerrorcir- ispossiblydetectedintheEPIC-MOSspectrum(seetheresiduals cle;theyhaveanRmagnitudeofR=23.94±0.41(sourceA)and inFigure2,wherethemodeldoesnotincludetheline).However R=24.73±0.64(sourceB).AnopticalspectrumofsourceA,ob- nocompellingevidenceofthepresenceofthislineintheEPIC-pn tainedsumminguptwoindependentexposures,isshowninFigure isfound,althoughdifferentspectralbinningsweretried. 2; the spectrum has a clear red continuum and a significant line Ifthislineisreal,andassociatedwiththeFeKαemissionline at ∼ 6569 Å. Assuming this lineto be [OII]λ3727 Å at z=0.763 (themostprominentfeatureintheX-rayspectrumofanAGN),the another observed feature is in very good agreement with being impliedredshift would bez∼ 0.65, theabsorbing column density [OIII]λ5007Å;thesetwofeaturesareclearlyseeninboththesin- ∼4×1021 cm−2andtheintrinsic2-10keVluminosity∼2.5×1044 gleexposuresandarenotassociatedwithstrongskylines.Wehave ergs−1 ; the line has an observed equivalent width of EW ∼ 400 alsotriedtotakeanopticalspectrumofthesourceB,butthesource eV. Interestingly the source is also a strong (and compact) radio isveryfaintandnousefulinformationcouldbeextractedfromthe source,detectedat15GHz(236±1mJy,Richardsetal.2011),8.4 verynoisyspectrum. GHz(∼376mJy,Healeyetal.2007)andat1.4GHz(183.5±5.5 2XMMJ123204.9+215254isdetectedinalltheWISEbandsdis- mJy,NVSSsurvey).Thebroadbandspectralindices(αRX =0.88; cussed here. The WISE position is coincident with a source de- αOX=0.31; αRO=1.15), computed with the fluxes reported above, tectedintheKband(K=18.07±0.05,seeDelMoroetal.2009)and and the possible presence of a Fe Kα linein the X-ray spectrum botharesignificantlyclosertothebrightestopticalsource(source arenotconsistentwithaBLLacclassification.Weremindthatthe A), strongly suggesting that the most probable optical counter- X-raysourceispoint-likeanddetectedalmostinthecenterofthe part of the X-ray source 2XMMJ123204.9+215254 is the source EPICfields,soahighzclusterofgalaxiesisveryunlikely. A. In the following of this paper we therefore assume that the The radio spectrum is very flat (indeed almost inverted, with 2XMMJ123204.9+215254 is spectroscopically identified with a a maximum at ∼ 8 GHz) suggesting that this source could be Type2QSOatz=0.763. classified as a Giga-Hertz Peaked Spectrum radio source, a class 2XMMJ123204.9+215254 was previously discussed in of sources supposed to be young radio galaxies which are often DelMoroetal. (2009) who report an infrared spectrum taken characterisedbyhighintrinsicabsorption(seee.g.Guainazzietal. at the Subaru telescope (with the MOIRCS instrument) clearly (2006)). revealingthepresenceofalineattheobservedframeof1.8837µm. Assuming that this source is an absorbed QSO and using the Lackingoptical spectroscopy andadeep optical image,theseau- relationbetween fx/fo andtheintrinsic2-10keVluminosity(see thorsdiscussedseveralpossibleidentificationsfortheinfraredline Section5)wecanestimatearedshiftz∼1.7(zintherange1.1−2.7, and suggested the Hα line at z=1.87 as itsmost probable origin. inconsistent withtheredshift estimatedbytheputativeironline), AsdiscussedaboveouropticalspectroscopyattheGTCsuggestsa withan intrinsic2-10 keV luminosity ∼ 3×1045 ergs−1 and an lowerredshiftofz=0.763.InthiscasetheNIRfeaturereportedin intrinsicNH ∼ 1.2×1022 cm−2.Thissourceisdetectedinallthe DelMoroetal.(2009,observedat1.8837µm)couldbeassociated WISEbands. with the line complex HeIλ10830+Paγ10941, a strong and quite • 2XMMJ143623.8+631726; fx/fo=55 common feature in AGN (see Glikmanetal. 2006), expected to AfaintobjectwithopticalmagnitudeR=22.16ispresentinthe be at ∼ 1.92 µm (at z=0.763), slightly higher than the observed centreoftheerrorcircle(seeFigure2).Theopticalspectrumofthis feature.HowevertheobservedNIRlinefallsinawavelengthrange objectisshowninFigure2;thetwomarkedlinescanbeassociated dominatedbytheeffectofatmosphericabsorption,soitisdifficult with[OII]λ3728 and Hγ+HeI (observed FWHM< 1100 kms−1 todeterminetheintrinsiclinecentroid; allinall,weconsider the )atz=0.893.TheobservedX-rayspectrumisdescribedbyanab- association of the infrared line with the HeIλ10830+Paγ10941 sorbedpower-lawmodelwithintrinsicNH =1.46×1022 cm−2;the complex as highly plausible. On the contrary none of the faint intrinsic2-10keVluminosityis∼8.4×1044 ergs−1.Thesource features that we see in the optical spectrum could be associated isdetectedinalltheWISEbands;therearenoradiodetectionsat with any relevant emission lines from an AGN in the case of thesourceposition.WeclassifythissourceasaType2QSO. z=1.87. Attheproposedlowerredshift(z=0.763)theX-rayspectrumis describedbyanabsorbedpower-lawmodelhavingaveryflatpho- tonindex,Γ = 1.31±0.23,andanintrinsicN = 3.6+1.03×1022 H −0.92 cm−2 .Theintrinsic2-10keVluminosityis∼ 1.9×1045 ergs−1 . Since the best-fitting photon index is rather flat we test the stability of the measured N and luminosity assuming a typi- H 5 DISCUSSION cal AGN photon index (Γ = 1.9); we derive an intrinsic N = H 5.9+0.73 × 1022 cm−2 and an intrinsic 2-10 keV luminosity of HavingdefinedasmallbutrepresentativesampleofbrightEXO50 −0.65 ∼ 2.3×1045 ergs−1 ingoodagreementwiththepreviousvalues. objects, itisnow instructivetocompare theirbroad band proper- Wenotethat2XMMJ123204.9+215254isnotonlytheobjectwith tieswiththoseofothersamplesofAGN,bothabsorbedandunab- thehighest f /f inthissample, but (toour knowledge) itisthe sorbed,fromtheliterature. x o (cid:13)c 0000RAS,MNRAS000,1–16