Mon.Not.R.Astron.Soc.000,1–10(0000) Printed11January2012 (MNLATEXstylefilev2.2) NGC454: unveiling a new “changing look” AGN E. Marchese1, 2⋆, V. Braito1, 3†, R. Della Ceca1, A. Caccianiga1, P. Severgnini1 2 1INAF-OsservatorioAstronomicodiBrera,viaBrera28,20121Milano,Italy 1 2Universita`deglistudidiMilano-Bicocca,Piazzadell’AteneoNuovo,1-20126,Milano 0 3X-RayAstronomyObservationalGroup,DepartmentofPhysicsandAstronomy,LeicesterUniversity,LeicesterLE17RH,UK 2 n a J 0 1 ABSTRACT ] WepresentadetailedanalysisoftheX-rayspectrumoftheSeyfert2galaxyNGC454E, O belongingto the interactingsystem NGC454. Observationsperformedwith Suzaku,XMM- C Newton and Swift allowed us to detect a dramatic change in the curvature of the 2–10 keV . spectrum, revealing a significant variation of the absorbing column density along the line h p of sight (from ∼ 1 × 1024cm−2 to ∼ 1 × 1023cm−2). Consequently, we propose this - source as a new member of the class of “changing look” AGN, i.e. AGN that have been o observedbothinCompton-thin(NH=1023cm−2)andreflectiondominatedstates(Compton- tr thick,NH> 1024cm−2).Due tothequitelongtimelag(6 months)betweentheSuzakuand s XMM-Newton observationswe cannotinfer the possible location of the obscuringmaterial a causing the observed variability. In the 6–7 keV range the XMM-Newton observation also [ shows a clear signature of the presence of an ionized absorber. Since this feature is notde- 1 tectedduringtheSuzakuobservation(despiteitsdetectability),thesimplestinterpretationis v that the ionized absorber is also variable; its location is estimated to be within ∼ 10−3 pc 9 fromthecentralblackhole,probablymuchcloserinthantheratherneutralabsorber. 6 1 Keywords: galaxies:active–galaxies:individual(NGC454)–X-rays:galaxies 2 . 1 0 2 1 INTRODUCTION andreflectiondominated states(N > 1024cm−2) (Risalitietal. 1 H 2002), implies that the absorbing material has to be clumpy and v: There is now a general consensus that Active Galactic Nuclei atmuch smallerdistancethantheconventional obscuring “torus” i (AGN) are powered by accretion of matter onto a supermassive X with velocity, distance and size from the central X-ray source of blackhole(SMBH),locatedatcenterofalmostallmassivegalax- thesameorderofthoseoftheBroadLineRegion(BLR)clouds. r ies. It is also clear that, according to the Unified Model of AGN a Up to now, we can count only a few “changing look” (Antonucci 1993),thedifferencebetweentype1andtype2AGN AGN where such a variability has been discovered on time- can be explained through orientation effects between our line of scales from a few days down to a few hours: NGC 1365 sight tothe nucleus and “circum-nuclear material”. However, the (Risalitietal. 2005, 2007, 2009), NGC 4388 (Elvisetal. 2004), geometry,sizeandphysicalstateofthiscircum-nuclearmatterare NGC 7674 (Bianchietal. 2005), NGC 4151 (Puccettietal. still a matter of debate. In particular, the AGN X-ray spectra are 2007),NGC7582(Bianchietal. 2009),UGC4203(Risalitietal. complexandconsistofmultiplecomponents(seeTurner&Miller 2010), NGC4051 (Uttleyetal. 2004; Lobbanetal. 2011) and 1H 2009, Done 2010 for a review), which are all intimately related 0419-577 (Poundsetal.2004).Among themwerecallNGC2992 tothestillpoorlyunderstood conditionofthematternearthenu- (Weaveretal.1996),howeverforthissourceoneyearmonitoring cleus.Thiscircum-nucleargasimprintsfeatures-lowenergycut- with RXTE (Murphyetal. 2007) unveiled the presence of short- offs,theComptonhumpandemissionandabsorptionlines-onto term flaring activity rather than a change in the covering of the theprimaryX-rayemission.TheX-rayspectraand,crucially,their absorber. variability observed in few nearby AGN showed that this matter ishighlystructuredwitharangeofionisationstates,densities,ge- Within a project investigating the occurrence of AGN in a ometriesand locations(Turner&Miller 2009, Risaliti 2010).In sampleofinteractinggalaxies,wecameacrossaninteractingsys- thisrespect,thesignificantvariabilityoftheabsorbingcolumnden- tem, NGC454, which was recently observed in the X-ray energy sity(N )detectedinthesocalled“changinglook”AGN,i.e.AGN that havHe been observed both in Compton-thin (N =1023cm−2) band withSuzaku, and ∼6 months later withXMM-Newton, and H whosemainX-rayspectralcomponentspresentinterestingvariabil- ityproperties. ⋆ E-mail:[email protected] Here we compare and discuss the X-ray observations from † E-mail:[email protected] Suzaku, XMM-Newton and Swift that unveiled that NGC454 can 2 Marcheseet al. beplacedamongthoseAGNwhoseabsorbingN isstronglyvari- 3.1.1 TheSuzakuXISanalysis H able(section4.2).Thepaperisstructuredasfollows.Theinteract- TheXISdatawereselectedin3×3and5×5editmodesusingonly ingsystemNGC454isdescribedin§2.TheX-rayobservationsand goodeventswithgrades0,2,3,4,6andfilteringthehotandflicker- datareductionaresummarizedin§3.In§4wepresentthespectral ingpixelswiththescriptsisclean;thenetexposuretimesare103 analysisofbothdatasetsandthecomparisonbetweentheobserva- ksec for each of the XIS. The XIS source spectra were extracted tions,aimedtoassessthenatureoftheX-rayabsorber. Summary fromacircularregionof2.2′radiuscenteredonthesource,andthe andconclusionsfollowin§5. backgroundspectrawereextractedfromtwocircularregionswith Throughout this paper, a concordance cosmology with H = 71 km s−1 Mpc−1, ΩΛ=0.73, and Ωm=0.27 (Spergeletal. 20003) is tchaelibsaramtieornasdoiuusrcoefs.thTehseoXurIcSerreesgpioonns,eo(frfmseftsf)raonmdtahnecislolaurrycereasnpdonthsee adopted. (arfs) files were produced, using the latest calibration files avail- able,withtheftoolstasksxisrmfgenandxissimarfgenrespectively. ThespectrafromthetwoFICDDs(XIS0andXIS3)werecom- 2 NGC454 binedtocreateasinglesourcespectrum(hereafterXIS–FI),while theBI(theXIS1)spectrumwaskeptseparateandfittedsimultane- Optical studies (Arp&Madore 1987; Johansson 1988; ously.Thenet0.5–10keVcountratesare:(0.0117±0.0005)cts/s, Stiavellietal. 1998) of the interacting system NGC454 (see (0.0142±0.0005)cts/s,(0.0132±0.0006)cts/sfortheXIS0,XIS3 Figure1,rightpanel)describeitasapairofemissionlinegalaxies andXIS1respectively.Weconsidereddataintherange0.5–10keV consistingofaredellipticalgalaxy(easterncomponent, hereafter fortheXIS–FIandintherange0.6–7keVfortheXIS–BI(because NGC454E) and a blue irregular galaxy (western component, theXIS–BIisoptimizedforobservingbelow∼ 7keV).Forboth hereafter NGC454W), at redshift z=0.0122. The distorted mor- the XIS-FI and XIS-BI we ignored the band 1.6–1.9 keV, due to phologyofboththesegalaxies,togetherwiththespectroscopicand thepresenceofinstrumentalcalibrationuncertainties.ThenetXIS photometricevidenceofayoungstellarpopulation,isaclearsign sourcespectrawerethenbinnedtoaminimumof50countsperbin. of the interacting nature of this system. Furthermore, three very blueknots(discussedinsection3.3.1),probablyStrongrenspheres surrounding clusters of very hot newly formed stars, are located 3.1.2 TheSuzakuHXD-PINanalysis (andlikelyrelated) tothesouth of NGC454W.HST observations of the system, performed with the Wide Field Planetary Camera For the HXD-PIN data reduction and analysis we followed the 2, confirmed that NGC454 is in the early stages of interaction latestSuzakudatareductionguide(theABCguideVersion22),and (Stiavellietal. 1998). The above authors stated also that an usedtherev2data,whichincludeall4clusterunits.TheHXD-PIN important fraction of gas has drifted to the center of the eastern instrument team provides the background (known as the “tuned” component,butithasyetnotproducedanysignificantvisiblestar background) eventfile,whichaccountsfortheinstrumental“Non formationactivity;apopulationofyoungstarclustershasformed X-rayBackground”(NXB;Kokubunetal. 2007).Thesystematic aroundthewesterncomponent. uncertaintyofthis“tuned”backgroundmodelis±1.3%(atthe1σ The optical spectrum of NGC454E is consistent with that of a levelforanet20ksecexposure3). Seyfert 2 galaxy (although none of the high excitation lines, e.g. We extracted the source and background spectra using the same HeII lines, can be seen) while no optical evidence of an AGN is commongoodtimeinterval,andcorrectedthesourcespectrumfor present in the spectrum of NGC454W which is fully consistent thedetector deadtime.Thenetexposure timeafterthescreening withthatofastar-forminggalaxy(Johansson 1988). was106ksec.WethensimulatedaspectrumforthecosmicX-ray backgroundcounts(Boldt 1987;Gruberetal. 1999)andaddedit totheinstrumentalone. 3 OBSERVATIONSANDDATAREDUCTION NGC454isdetectedatalevelof3.4%abovethebackground 3.1 Suzakudata andthenetcountrateinthe15–30keVbandis0.01±0.002cts/s. Forthespectralanalysisthesourcespectrumwasrebinnedinorder NGC454 was observed on April 29, 2009 by the Japanese X-ray to have a signal-to-noise ratio >3 in each energy bin. We fit the satelliteSuzaku (Mitsudaetal.2007)for atotal exposure timeof Suzaku-HXDspectrumwithasingleabsorbedpower-lawcompo- about130ksec.SuzakucarriesonboardfourX-rayImagingSpec- nentwithaphotonindexΓ = 1.9andderivedanobserved15–30 trometers (XIS, Koyamaetal. 2007), with X-ray CCDs at their keVfluxof∼3.4×10−12ergcm−2s−1. focal plane, and a non-imaging hard X-ray detector (HXD-PIN, Takahashietal. 2007).Atthetimeofthisobservation onlythree oftheXISwereworking:oneback-illuminated(BI)CCD(XIS1) 3.2 TheSwift-BATobservation andtwofront-illuminated(FI)CCDs(XIS0andXIS3).Alltogether NGC454 was also detected with the BAT detector on board of theXISand theHXD-PINcover the0.5–10 keV and12–70 keV Swift (Gehrelsetal. 2004). BAT is a coded aperture imaging bands respectively. The spatial resolution of the XIS is ∼ 2 ar- camera that operates in the 14–150 keV energy range; it has a cmin(HEW),whilethefieldofview(FOV)oftheHXD-PINis34 large field of view (1.4 steradian half coded), and a point spread arcmin radius. Data from the XIS and HXD-PIN were processed function(PSF)of18arcmin(HEW).Swift-BATisdevotedmainly using v2.1.6.14 of the Suzaku pipeline and applying the standard screeningparameters1. 256softheSAAwereexcludedfromtheXISandwithin500softheSAA fortheHXD.Cut-offrigidity(COR)criteriaof>8GVfortheHXDdata 1 ThescreeningfiltersalleventswithintheSouthAtlanticAnomaly(SAA) and>6GVfortheXISwereused. aswellaswithanEarthelevation angle(ELV)< 5◦ andEarthday-time 2 http://heasarc.gsfc.nasa.gov/docs/suzaku/analysis/abc/ elevation angles(DYEELV)lessthan20◦.Furthermorealsodatawithin 3 ftp://legacy.gsfc.nasa.gov/suzaku/doc/hxd/suzakumemo-2008-03.pdf NGC454:unveilinga new“changinglook”AGN 3 NGC454E NGC454E XE XE XW NGC454W XS XSE XSE XS XW 1 arcmin Figure1.Leftpanel:XMM-NewtonEPIC-pnimage(0.5–10keV)withsuperimposedtheSuzakuXISextractionregion.RightPanel:DigitalSkySurvey(DSS) opticalimagewithoverlaidtheXMM-NewtonPN0.5–10keVcontours.WemarkedthemainX-raysourcesdiscussedinthetext.Itisevidentthatthemain X-raysourceispositionallycoincidentwithNGC454E(classifiedasaSeyfert2)whilenoX-rayemissionisdetectedatthepositionofNGC454W. tothemonitoringofalargefractionoftheskyfortheoccurrence ScienceAnalysisSoftware(SASver.6.5)andanalysedusingstan- of gamma ray bursts (GRBs); while waiting for new GRBs, it dardsoftwarepackages(FTOOLSver.6.1andXSPECver.11.3). continuously collects spectral and imaging information in survey Event fileshave beenfilteredfor high-background timeintervals, mode,coveringafractionbetween50%and80%oftheskyevery and only events corresponding to patterns 0–12 (MOS1, MOS2) day. andtopatterns0–4(pn)havebeenused.Thenetexposuretimesat NGC454(BATname:SWIFTJ0114.4-5522)ispartofthePalermo thesourcepositionafterdatacleaningare∼23.9ksec(pn),∼29.1 Swift-BAT 54-Month hard X-ray catalogue (Cusumanoetal. ksec(MOS1)and∼29.2ksec(MOS2). 2010) and the Swift-BAT 58-Month Hard X-ray Survey IntherightpanelofFigure1wereporttheopticalDSSimage (heasarc.gsfc.nasa.gov/docs/swift/results/bs58mon/). This of the system NGC454, together with the XMM-Newton 0.5–10 last survey detected 1092 sources in the 14–195 keV band keV contours (green, in the electronic version only) from EPIC- down to a significance level of 4.8σ, reaching a flux level pn.ItisevidentthatthebulkoftheX-rayemissionispositionally of 1.1 × 10−11erg cm−2 s−1 over 50% of the sky (and coincidentwithNGC454E(thegalaxyspectroscopicallyclassified 1.48 × 10−11erg cm−2 s−1 over 90% of the sky); as part of as a Seyfert 2) while no strong X-ray emission is detected at the this new edition of the Swift-BAT catalogue, 8-channel spectra positionofNGC454W(thesourcespectroscopicallyclassifiedasa and monthly-sampled light curves for each object detected inthe star-forminggalaxy).WealsodetectedaweakX-raysourcetothe surveyweremadeavailable(Baumgartneretal. 2011). southofNGC454,whichispositionallycoincidentwithoneofthe The 14–195 keV observed flux of NGC454 is 1.90+0.5 × threeveryblueknotsdiscussedabove,likelyastarformingregion −0.5 10−11erg cm−2 s−1 , in agreement, when accounting for the belongingtoNGC454W. differentbands,withthefluxquotedintheCusumanoetal. 2010 Thepn,MOS1andMOS2sourcespectrawereextractedfrom catalogue(F14−195keV ∼ 1.7×10−11ergcm−2 s−1 ).Thisflux a circular region of 0.46 arcmin radius centered on the source is also in good agreement with the expected 14–195 keV flux (NGC454E), while the background spectra were extracted from (∼ 1.6×10−11ergcm−2 s−1 )extrapolatedfromthatmeasured twocircularregionswith0.5arcminradiusoffsetfromthesource. withSuzakuinthe15–30keVrange. TheMOS1andMOS2spectrawerecombined,thenboththeEPIC- pn and EPIC-MOS spectra were grouped with a minimum of 30 countsperchannel. 3.3 TheXMM-Newtonobservation NGC454wasobservedwithXMM-NewtononNovember5,2009 3.3.1 ContaminationfromunresolvedsourcesintheSuzaku foratotalexposuretimeofabout30ksec.TheXMM-NewtonOb- (XIS,HXD)andSwift-BATextractionregion/fieldofview servatory (Jansen et al. 2001) carries, among its onboard instru- ments, three 1500 cm2 X-ray telescopes, each with EPIC (Euro- In the left panel of Figure 1 we show the XMM-Newton 0.5–10 peanPhotonImagingCamera)imagingspectrometersatthefocus. keVpnimagealongwiththeSuzakuextractionregion(circlewith TwooftheEPICuseMOSCCDs(Turneretal. 2001)andoneuses 2.2′radius).AsdiscussedabovethemainX-raysourceiscentered apnCCD(Stru¨deretal. 2001).TheseCCDsallowobservationsin on NGC454E but given the XMM-Newton better angular resolu- the range ∼0.5–10 keV. The spatial resolution of the 2 MOSs is tion (14′′–15′′ HEW) as compared to Suzaku (120′′ HEW), we ∼14′′(HEW),and∼15′′(HEW)forthepn(Ehleetal. 2001). canclearlydistinguish4other X-raysources, besidesNGC454E, During this observation the pn, MOS1, and MOS2 cameras entering in the Suzaku XIS extraction region. We extracted the hadthemediumfilterappliedandtheywereoperatinginfullframe XMM-Newton spectra for the 3 brighter sources (XS, XE and Windowmode.Thedatahavebeenprocessedandcleanedusingthe XSE,marked in Figure1 for clarity) and analysed them in order 4 Marcheseet al. to estimate their possible contribution to the Suzaku spectrum; spatial resolution, we stressthat no emission was detected below theremainingsource(XW)hasonly∼ 80countsdetectedinthe 10keVfromNGC454W,whileacontributionwouldbeexpected ∼0.5−10keVband(seebelow). eveninthecaseofadeeplyburiedAGN(seee.g.DellaCecaetal. 2002).Wethusconcludethatwedonotexpectsignificantcontam- XS is well fitted with a power law, modified only by inationsfromthenearbysourcestotheSuzakuandSwiftspectra. Galactic absorption, with a photon index Γ ∼ 1.8 and a 2–10 keV flux F ∼ 1.9 × 10−14erg cm−2 s−1 ; the ex- [2−10]keV trapolated flux in the 14–70 keV band (assuming Γ ∼ 1.8) is F[14−70]keV ∼< 3 × 10−14erg cm−2 s−1 . As said above this 4 SPECTRALANALYSIS source is likely associated with a star forming region related to 4.1 TheSuzakuandSwiftbroadbandX-rayemission NGC454W;ifso,assumingz = 0.0122,its2–10keVluminosity isL[2−10]keV ∼ 6.2×1039 ergs−1. Wecannot establish ifthis We first considered the X-ray spectrum of NGC454E in the luminosity is due to one or more sources and thus speculate on 0.5–100 keV band by fitting simultaneously the Suzaku XIS, its/their nature, because we lack both the spatial resolution and Suzaku HXD and Swift-BATdata. Thecross-normalisation factor a good enough sampling to assess its variability and spectral betweentheHXDandtheXIS-FIwassetto1.18,asrecommended properties. The source to the east of NGC454E (hereafter XE) forHXDnominalobservationprocessedafter2008July(Manabu canbefittedwithapower-lawandathermalcomponent,yielding et al. 2007; Maeda et al. 20086), while the cross-normalisation Γ ∼1.6,kT ∼ 0.3andF[2−10]keV ∼ 8.8×10−15ergcm−2 s−1 betweenSwiftandXISwasallowedtovary. (F[14−70]keV ∼< 5 × 10−14erg cm−2 s−1 ). The source to the In the subsequent sections the χ2 statistics was used for the fit, south-east of NGC454E (hereafter XSE) can be fitted with an the errors are quoted to 90% confidence level for 1 parameter of absorbed power law (NH∼ 2.2 × 1022 cm−2) with photon interestandallthespectralparametersarequotedintherestframe index set to 1.8 and F[2−10]keV ∼ 2.7 × 10−14erg cm−2 s−1 ofthesource. (F[14−70]keV ∼< 3 × 10−14erg cm−2 s−1 ). Finally, the fourth source located to the west of NGC454E (hereafter XW) has not Wefittedthecontinuumwitharedshiftedunabsorbedpower- enough counts for a meaningful spectral analysis (∼ 80 counts) lawmodel,modifiedonlybyGalactic(N = 2.73×1020 cm−2, anditsestimatedfluxesareF[2−10]keV ∼2.0×10−14ergcm−2s−1 Dickey&Lockman 1990)absorption.ThHismodeldidnotprovide andF[14−70]keV ∼< 2.5×10−14ergcm−2s−1(adoptingΓ∼1.9). an adequate description of the broadband spectrum of NGC454E According to the extragalactic logN-logS distributions computed (χ2/dof=522.6/122). If we fit only the 2–5 keV continuum, ex- by Mateosetal. (2008), at this flux level the number of random cludingpossiblecomplexityinthesoftenergyrangeandnearthe 2–10 keV sources in the Suzaku extraction region is ∼ 2, thus Fe K emission line complex we found a very flat photon index the sources XE, XSE and XW are probably those expected by (Γ ∼ 0.15), strongly suggesting that we are dealing with an ab- ”chance”. There is no NED identification available for XE, XSE sorbedAGN, inagreement withtheoptical spectral classification andXW. ofNGC454E. Thecombined2–10keVfluxofallthese4possiblecontam- Theresiduals withrespect tothissimpleunabsorbed power- inating sources (F[2−10]keV ∼ 7.5×10−14erg cm−2 s−1 ) im- law model, which are shown in Figure 2, allowed us to infer the ply that they will provide a negligible contribution to the Suzaku main features of the observed spectrum. An excess at energies XISspectrumofNGC454(F[2−10]keV ∼6×10−13ergcm−2s−1 below1keV,anemissionlinefeatureat∼6.4keV(likelyassoci- ). More important their estimated F arewell below the [14−70]keV atedwithFeKα),togetherwithaline-likefeatureat∼7keV,and Suzaku HXD-PIN or Swift-BAT sensitivity. On the other hand anexcessatenergiesbetween10and20keV areclearlyevident. this check is still not sufficient for these two latter instruments TheresidualsinthesoftX-rayssuggestthepresenceofathermal since their FOV is larger than that of the Suzaku XIS instru- component probably relatedtothehostgalaxy. Thesimultaneous ment. Assuming that the X-ray emission above 10 keV detected occurrenceofastrongFeKαemissionlineat∼6.4keV(figure2 with the HXD-PIN or the Swift-BAT is associated to the same upperandlowerpanel),averyflatobservedΓandanexcessinthe source, as the good agreement of the measured fluxes strongly hardX-rays(above10keV)isthedistinctivespectralsignatureof suggests,wecanusetheinstrumentwiththesmallerFOV(Swift- ahighlyabsorbedsource,withapossiblestrongComptonreflected BAT) to perform further checks. In particular using known cata- component.Theexcessobservedat∼7keV(Figure2lowerpanel) loguesorarchives(NED4 andSIMBAD5)wesearchedforbright islikelyduetothecombinationoftheFeKβ emissionline(7.06 X-ray/optical sources within 6 arcmin radius error circle (corre- keV)andtheFeXXVI(∼6.97keV)andtheFeedge(∼7.11keV). sponding to 99.7% confidence level for a source detection at 4.8 standarddeviations, Cusumanoetal. 2010)thatcouldberespon- Giventhesefeatures,weperformedabroadbandfitincluding: sibleoftheobserved X-rayemissionabove 10keV. Noplausible contaminant source was found and, in the following, we will as- (i) a thermal component (modelled with the MEKAL model, sumethattheemissionabove10keVcomesfromNGC454E.We Meweetal. 1986); notethatwearealsoassuminganegligiblecontributiontotheemis- (ii) anabsorbedprimarypower-lawcomponent; sion above 10 keV from the companion galaxy in the interacting (iii) anunabsorbedpower-lawcomponentwiththesamephoton system,NGC454W.Whileaconfirmationofthisassumptionhasto indexΓ; waitfordirectimagingobservations above10keVwithadequate 6 http://www.astro.isas.jaxa.jp/suzaku/doc/suzakumemo/suzakumemo- 4 http://ned.ipac.caltech.edu/ 2007-11.pdf; 5 http://simbad.u-strasbg.fr/simbad/ http://www.astro.isas.jaxa.jp/suzaku/doc/suzakumemo/suzakumemo-2008-06.pdf NGC454:unveilinga new“changinglook”AGN 5 1 0 0. 6 0−3 V) 1 e k atio 4 cm s 2 10−4 R V/ e V (k 0−5 e 1 2 k 0−6 1 1 10 100 1 10 100 Observed Energy (keV) Energy (keV) 6 Figure 3. Unfolded Suzaku spectrum, showing separately the different components ofthebest-fitmodel:ingreen(intheelectronic version)the scattered power-law and the primary absorbed power-law, in red (in the electronicversion)thesoftthermalcomponent,inyellow(intheelectronic version)theironKαandKβemissionlines,inlightblue(intheelectronic 4 version)thereflectioncomponent,andinblue(intheelectronicversion)the o totalresultingspectrum. ati R sincetheyrepresentthesamemediumproducingtwodifferentef- 2 fects(i.e.thenon-relativisticComptonscatteringandphotoelectric absorptionoftheprimaryradiation,respectively). Themodelsetupis: 5 6 7 8 WABS×[MEKAL+ZPOWERLW+ZGAUSS+ZGAUSS+PEXRAV Rest Energy (keV) +CABS×ZPHABS×(ZPOWERLW)] Figure 2. Upper panel: ratio between the Suzaku and Swift data (XIS- Wefoundthatthismodelprovidesagoodrepresentationofthe FI:blackfilledsquares;XIS1:redopencircles;HXD:greenrhombs;and BAT:blueopensquares,colorsintheelectronic versiononly)andtheun- X-ray emission of NGC454E (χ2/dof=104.5/103). The resulting absorbedpower-lawmodelusedtofitSuzakudatainthe2–5keVenergy best-fitparametersarereportedintable1.Inparticular,thisbest- range.Lowerpanel:zoomintothe5–8keVenergyrange(XIS-FI:black fitmodelyieldedΓ = 1.92+0.29,N = 2.05+4.25 ×1024 cm−2. −0.36 H −1.38 filledsquares,XIS1:redopencircles).Wecanclearlyseeat6.4keVthe Therest-frameenergyoftheFeKαisEKα = 6.38±0.02 keV excesscharacteristicoftheFeKαemissionlineandat∼7keV,thecom- anditsequivalentwidthwithrespecttotheobservedcontinuumis binedcontributionoftheFeKβemissionline(7.06keV),FeXXVI(∼6.97 EW=340+60eV.AttheSuzakuspectralresolutionthisemissionline keV),andthereflectoredge.ThecentralenergiesoftheFeKαandFeKβ −80 isunresolved;leavingthewidthσfreetovarywefoundσ.70eV aremarkedwithdashedverticallines. (atthe90%confidencelevel),thuswefixedittobe≃ 10eV.The cross-normalisationfactorbetweentheSwift-BATandtheXIS-FI is1.05+0.64.Westressthatadifferentchoiceofthecut-offenergy −0.39 (iv) twoGaussianemissionlinesat∼6.4keV(FeKα)and7.06 intherangebetween100and300keVdoesnotaffectsignificantly keV(FeKβ)respectively.WekepttheenergyoftheFeKβ fixed thebest-fitreflectionparametersobtained inthiswork. fTherel- to7.06 keV, tieditsintrinsicwidth(σ) tothewidth of thecorre- ativeimportanceofthereflectioncomponent isgivenbytheratio spondingFeKαlineandfixeditsfluxtobe13%oftheFeKαflux, betweenthenormalizationsoftheprimaryabsorbedpower-lawand consistentwiththetheoreticalvalue(Kaastra&Mewe 1993); thereflection component; inour casethisratio is∼0.5, whichat (v) aComptonreflectedcomponent,modelledwiththePEXRAV firstorderwouldcorrespondtoareprocessorcoveringasolidangle model in XSPEC (Magdziarz&Zdziarski 1995). The parameters 2π.Thefractionofscatteredradiationis∼1%.Theobserved2–10 ofthereflectedcomponentare:aninclinationangleifixedto63◦, keVfluxis∼ 6.3×10−13ergcm−2 s−1 whiletheintrinsic2–10 abundanceZ=Z⊙,areflectionfraction(definedbythesubtending keVluminosityobtainedwiththismodelis7.2×1042ergs−1. solid angle of the reflector R = Ω/4π) fixed to be -1 (i.e. pure reflection)7, the cut-off energy (fixed at 200 keV, Dadina 2008) andthenormalisation. 4.2 ComparisonwithXMM-Newtondata The absorber was modelled by a combination of the CABS and In Figure 4 we report the Suzaku XIS (black, lower spectrum), ZPHABS models in XSPEC,assuming the same column density, HXD(black) andSwift-BATspectra(green,intheelectronicver- siononly).Inred(upperspectrum)wealsoshowtheXMM-Newton pnandMOSdata,revealingadramaticchangeinthespectralcur- 7 Sinceinthe“purereflection”PEXRAVmodelthereisadegeneracybe- vaturebetween3and6keV.Thisvariationismostlikelyduetoa tweenRandthenormalisation,wesetthereflectionscalingfactorto-1and changeintheamountofabsorptionoftheprimaryradiation.Totest allowedthenormalisationtovary. thishypothesisweappliedtheSuzakubest-fitmodeltotheXMM- 6 Marcheseet al. Table1.SummaryoftheSuzakuandXMM-Newtonparametersforthebest-fitmodelsdescribedinsection4.1,and4.2.1. ModelComponent Parameter Suzaku XMM-Newton Powerlaw Γ 1.92+0.29 1.99+0.11 −0.36 −0.07 Normalisationa 7.39+30.00,b 2.77+0.71 −4.39 −0.65 ScatteredComponent Normalisationa 8.55+5.48×10−3 1.62+0.29×10−2 −4.52 −0.22 Absorber N 2.05+4.25×1024cm−2 1.0+0.1×1023cm−2 H −1.38 −0.2 Thermalemission kT 0.62+0.10keV 0.62+0.11keV −0.17 −0.11 Normalisationc 6.94+2.40×10−6 3.49+1.52×10−6 −2.22 −1.50 Neutralreflection Normalisationa 3.46+2.14 3.55+1.52 −1.61 −1.81 FeKαd Energy 6.38+0.02keV 6.36+0.03keV −0.02 0.03 EW 340+60eV 120+40eV −80 −40 Normalisatione 3.62+0.79×10−3 4.75+1.35×10−3 −0.78 −1.40 IonisedAbsorber N .. 6.05+8.95×1023cm−2 H −4.10 logξ .. 3.55+0.49ergcms−1 −0.25 vturb .. 300kms−1 χ2/dof 104.5/103 190.7/197 F(0.5−2)keV ∼4.9×10−14ergcm−2s−1 ∼5.8×10−14ergcm−2s−1 F(2−10)keV ∼6.3×10−13ergcm−2s−1 ∼1.9×10−12ergcm−2s−1 F(14−150)keV ∼1.4×10−11ergcm−2s−1 ∼1.3×10−11ergcm−2s−1 L(0.5−2)keV ∼4.7×1042ergs−1 ∼2×1042ergs−1 L(2−10)keV ∼7.2×1042ergs−1 ∼2.5×1042ergs−1 L(14−150)keV ∼1.4×1042ergs−1 ∼4.8×1042ergs−1 aunitsof10−3photonskeV−1cm−2s−1. bDuetoadegeneracybetweenthenormalisationsoftheprimarypowerlawandPEXRAV,theerrorswerecomputedfixingthereflectionnormalisationtoits best-fitvalue. cThenormalisationofthethermalcomponentisdefinedasK= 4π(D1A0(114+z))2 RnenHdV whereDAistheangulardiameterdistance,zistheredshift, neandnHaretheelectronandhydrogendensity(cm−3)respectively,anddVisthevolumefromwhichthedeprojectedemissionoriginates. dThelineisunresolved;theintrinsicwidthhasbeenfixedtobe≃10eV. eunitsof10−3photonscm−2s−1. to ∼ 2.6×1023cm−2); this change in the amount of absorption 0−4 issufficienttoexplainthebulkofthedifferencesbetweentheob- 1 m−2 servedXMM-NewtonandSuzakuspectra. eV c−1 Fthoerncoomrmpalelitseanteiosns,owfebaoltshotahlleowpoewdetor lvaawryctohmeppohnoetonnts,intdheexth(Γer)-, s k−1 10−5 mal component and the Kα energy and normalisation. The fit s yielded χ2/dof=213.9/199 and the only parameter changing well nt ou beyond the Suzaku errors is, as expected, the NH, decreasing to alized c 10−6 2be.7tw8+−ee00n..1167X×MM10-2N3ewcmto−n2a.nTdhSisuzcaoknufiirsmdsuethtaot athcehsatnrogneginvacroilautmionn m densityof∆N ∼1.8×1024cm−2. or H n Promptedfromthisresultwealsoinspectedthe3Swift-X-rayTele- scope(XRT)observationstakenin2006withatimelagoftheorder 0−7 of1–2daysfromeachother,andwefoundthatthesourcewasin 1 1 10 100 Observed Energy (keV) a state similar to that observed by XMM-Newton. The exposure timeofeachof theobservations islessthan10ksec(8713, 8661 Figure4.Comparison between the SuzakuXIS(black, lower spectrum), and 3667 sec respectively), thus the relatively low statistics does HXD(black),Swift-BAT(green,colorsintheelectronicversiononly)and notallowustoestablishifthereisavariabilitybetweenthesingle theXMM-Newton(red,upperspectrum)datashowingthedramaticchange inthecurvatureinthe3–6keVenergyrange.Theunderlyingmodel(black observations. andgreenline)istheoneobtainedfittingonlytheSuzakuXIS(black)and A closer inspection of the residuals in the 5–8 keV energy Swift-BATdata(green). range to this best-fit model showed some residual curvature between 5and 6keV, together witha possibleabsorption feature centeredat∼6.7keV(seeFigure5),whichispresentinboththe pn and MOS spectra and is suggestive of a more complex and Newtonspectra, leavingonly theabsorbing column density (N ) likely ionized absorber. After checking the significance of this H freetovary. Wealsoleftboththecross-normalisation factorsbe- absorption line, we included in the model an additional ionized tweenthepnandtheMOSspectraandbetweenSwift-BATandpn absorber (see §4.2.1). We note that, after accounting for the datafreetovary;theywerefoundtobe1.02±0.04and1.06+0.14 absorptionfeatureat∼6.7keV,theexcessofcurvatureintherange −0.16 respectively. During the XMM-Newton observation the N de- 5–6keVisnotpresentanymore. H creasedbyaboutoneorderofmagnitude(from∼2.1×1024cm−2 NGC454:unveilinga new“changinglook”AGN 7 (Tombesietal. 2010b).Wenotethatred-andblue-shiftedabsorp- tionlinesarepredictedinseveraltheoreticalmodelsoffaileddisk 2 winds (Proga&Kallman 2004; Simetal. 2010) or of aborted jet (Ghisellinietal.2004).However,beforeproceedingwithanyfur- thermodelingoftheabsorptionfeaturewecheckeditssignificance. 5 o 1. To assess the significance of the absorption feature we Rati performedextensiveMontecarlosimulationsasdetailedbelow.We assumedasournullhypothesismodelthebest-fitmodeldiscussed 1 attheendof section4.2,andwesimulatedS=3000spectra(with thefakeit command in XSPEC),withthesameexposure timeas thereal data. Eachone of thesesimulatedspectra wasthen fitted withthenull hypothesis model toobtainaχ2 value, andwesys- 5 0. tematicallysearchedforanabsorptionlineinthe2–10keVenergy 4 5 6 7 8 range,steppingtheenergycentroidoftheGaussianinincrements Rest Energy (keV) of 0.1 keV and refitting at each step. We then obtained for each Figure5.ResidualsoftheXMM-Newtondata(pndataaretheblackfilled simulatedspectrumaminimumχ2andcreatedadistribution3000 squaresandMOSdataaretheredopencircles)intherange4–8keVwith simulated values of the ∆χ2 (compared to the null hypothesis respecttothespectralmodeldiscussedinsection4.2.Anabsorptionfeature model).Thisindicatesthefractionofrandomgeneratedabsorption atabout6.7keVandaspectralcurvatureinthe5–6keVrangeareclearly features in the 2–10 keV band that are expected to have a ∆χ2 present. greater than athreshold value. If N of thesesimulated values are greaterthantherealvalue,thentheestimateddetectionconfidence level is1-N/S. Using thisanalysis we can then conclude that the The parameters of the XMM-Newton best-fit model are re- linedetectionissignificantat>99.97%level. portedintable1andthefinalmodelsetupisdescribedinsection 4.2.1.Wenotethatthedifferenceinthebest-fitnormalisationsof In order to obtain a physical description of the absorber we thethermalandscatteringcomponentbetweenSuzakuandXMM- replaced the Gaussian absorption line with a model representing Newtonarelikelyduetoadegeneracybetweenthesetwoparame- aphotoionized absorber, whichhasbeenproduced usingamulti- ters.Indeedthe0.5–2keVfluxdidnotstronglyvarybetweenthe plicativegridofabsorptionmodelgeneratedwiththeXSTARv2.1 twoobservations. code (Kallman&Bautista 2004). This grid describes an ionized The fraction of scattered radiation is ∼ 5%. The 2–10 keV flux absorberparametrisedbyitscolumndensity(N ),anditsionisa- H is∼ 1.9×10−12ergcm−2 s−1 ,whilethe2–10keVluminosity, tionparameter,definedas: L ,is∼ 2.5×1042ergs−1,aboutafactor2.8belowthe lu(m2−in1o0s)kiteyVcomputedusingonlytheSuzakudata.Althoughsucha ξ= Lion (1) nR2 variationoftheintrinsicluminosityisnotunusualinAGN,partof thisdifferencecouldbeduetothegeometryassumedbythemodels where Lion is theionising luminosity between 1–1000 Rydbergs (13.6eVto13.6keV),nisthehydrogengasdensityincm−3 and adoptedforthehighcolumndensityabsorber.Indeedwewillshow Risthe radial distance of the absorber fromthe ionising source. insection4.3thatthisdifferenceissmaller(afactor1.7)whenwe Since there is no apparent broadening of the absorption line we adopttheMytoruscodefortheabsorber. assumedaturbulencevelocityofv =300kms−1. turb The inclusion of this ionized absorber significantly improved 4.2.1 The∼6.7keVabsorptionfeatureintheXMM-Newton the fit (χ2/dof=190.7/197, ∆χ2 = 25 for 2 dof), with a column density of N = 6.05+8.95 × 1023 cm−2and an ionisation of observation H −4.10 log(ξ/erg cm s−1) = 3.55+0.49. The improvement in the χ2 is −0.25 Asafirststeptomodeltheabsorptionfeatureinthe6–7keVband determinedsolelybyfittingtheabsorptionfeatureinthe6–7keV we added a Gaussian absorbing component; setting σ = 0.05 band, since an ionised absorber withsuch a high level of ioniza- keV we found that the centroid of the line is at E=6.75+−00..0064 tiondoesnotproduceanyfeatureinthesoftbandofthecontinuum. keV, the normalisation is −3.75+1.12 × 10−6 and ∆χ2=24 for −1.13 2 dof. The absorption lines appears to be marginally resolved; The parameters of the XMM-Newton best-fit model are re- however leaving its width free to vary we can set only an upper portedintable1andthesetupisthefollowing: limit σ < 0.3 keV, while the energy centroid is found to be consistent within the errors E=6.77+0.08keV. The energy of this WABS×[MEKAL+ZPOWERLW+ZGAUSS+ZGAUSS+PEXRAV absorption line suggests an assoc−ia0ti.o06n with absorption from +XSTAR*CABS*ZPHABS×(ZPOWERLW)] highly ionized Fe (i.e. FeXXV at E∼6.7 keV) and thus a clear We can now estimate what is the maximum distance of this signature of the presence of an ionized absorber. The presence ionisedabsorberfromthecentralblackhole,usingequation1,re- of an ionized absorber is not exceptional since recent sensitive latingtheionisationparameter,thedensityoftheabsorberandthe observations with Chandra, XMM-Newton, and Suzaku unveiled continuumluminosityLion.InthiscaseLion(intheenergyrange thepresenceofred-andblue-shiftedphotoionizedabsorptionlines between 13.6 eV and 13.6 keV) is 7.3× 1042 erg s−1. Assum- both in type 1 and type 2 AGN as well as in Radio Quiet and ing that the thickness of the absorber ∆R=N /n is smaller than H RadioLoud AGN(Tombesietal. 2010b,2011).Thus, itappears thedistanceRion(∆R/Rion<1),wecansetanupperlimittothe that there is a substantial amount of ionized gas in the nuclei of distance: AwGithNvs,elwochiitciehsmfraoymbheunlidnrkeeddstoofkgmas/souuptfltoowvionugto∼n0p.a0r4se−c s0c.a1l5ecs Rion= LNioHnξ∆RR < NLiHonξ =2.3×1015cm (2) 8 Marcheseet al. This maximum distance of ∼ 10−3pc is consistent with a intoaccountthefactthattheFeKαisratherconstant(seetable1). location of the ionised absorber within the Broad Line Re- gion of the AGN. Indeed an estimate of the BLR size RBLR Wehavedonethisbydecouplingtheline-of-sightcontinuum for NGC454E can be inferred by using the relation between passing through the reprocessor (or zeroth order continuum, see RBLR and the monochromatic luminosity at 5100 A˚, L ˚ http://www.mytorus.com/manual/index.html)andthereflected(or 5100A (Kaspietal. 2005, RBLR = 2.45 × (λL (5100A˚))0.608). scattered continuum, see mytorus model) continuum from repro- 10lt−days λ Sincetheluminosityoftheopticalcontinuumcannotbemeasured cessor.Inpracticeweallowedthecolumndensitiesoftheline-of- directly from the spectrum, because of the strong absorption, sightcontinuumandscattered-reflectedcontinuatobeindependent we estimate L ˚ from the intensity of the [OIII]5007A˚ of each other. The reflected continuum, and the fluorescent line 5100A emission which is consistently produced in the same location, is line flux, assuming a mean F[OIII]5007A˚/F(5100A˚) ratio. This not extinguished by another column of intervening matter. Since ratio has been inferred from the AGN template presented in theXMM-Newtonspectrumunveiledthepresenceofanadditional Francisetal. (1991) (F(5100A˚) = 0.059F([OIII])). Using ionizedabsorber,whichaffectstheline-of-sightcontinuumwealso the [OIII]5007A˚ flux published in Johansson (1988) we obtain L5100A˚ ∼1×1041ergs−1A˚−1and,thus,anapproximatesizeof iXnSclTuAdeRdgarnidioansizdeedscarbibsoerdbeinr,swechtiicohnis4.m2.o1d.eIlnleodraddeorptotindgotthheast,amwee theBLRof0.05pc,i.e.about50timesRion. disentangledtheabsorbingcolumndensityoftheline-of-sight(los) component fromthatforthescatteredcontinuumplusfluorescent SincewedonotobservethisabsorptionfeatureintheSuzaku emissionlines.Theinclinationangleoftheloscomponenthasbeen spectrumweaddedtotheSuzakubest-fitmodelagaussianabsorp- fixed at 90 degrees; the inclination angle for the reflected/ scat- tionlinewiththesameparametersobtainedwiththeXMM-Newton teredcontinuumpluslinecomponentcomponentwas,forsimplic- data. The lower limit for the detection of anabsorption line with centralenergyof6.75keVandwidthof0.05keV,is-1.18×10−6 ity,fixedat0degreessincetheeffectoftheinclinationangleonthe shapeofthescatteredcontinuumisnotsufficientlylargewhenthe for Suzaku data. Thus, being the normalisation of this line - 3.75×10−6 inthe XMM-Newton spectrum, wecan infer that the scatteredcontinuumisobservedinreflectiononly.Physically,the situationwearemodellingbymeansofthisdecouplingcouldcor- ionisedabsorbershouldbedetectablebySuzaku.Thesimplestin- respondtoapatchyreprocessorinwhichthescatteredcontinuum terpretationisthatalsotheionizedabsorberisvariable;whichisnot isobserved fromreflection inmatter on the far-sideof the X-ray surprisingsincethereareseveralreportedcasesofvariableabsorp- source,withoutinterceptinganyother“clouds,”,whiletheintrinsic tion feautures (Tombesietal. 2010b, 2011) (Braitoetal. 2007; continuumisfilteredbyclouds”passing”throughourline-of-sight Cappietal. 2009;Dadinaetal. 2005;Risalitietal. 2005).More- tothecentralengine. over instability of the outflowing ionized absorbers is predicted WeappliedthismodeltothetheXMM-EPICandSwift-BAT both in disk winds models (Proga&Kallman 2004; Simetal. spectra and we found a god fit with the same absorbing column 2010)orofabortedjet(Ghisellinietal.2004).Thiswillcausethe density(N = 2.75+0.05 ×1023cm−2;χ2/dof=201/192) filter- presenceoftransientabsorptionfeaturesandvariabilityofthede- H −0.04 ing the line-of-sight intrinsic continuum and producing the scat- rived outflowing velocities and their EW as observed in several teredcomponent(includingtheproductionofthefluorescentemis- sources(seee.g.Tombesietal. 2010b) sion lines). The parameters of the ionized absorber are: N = H 6.46+5.04×1023cm−2andlogξ =3.26+0.20ergcms−1;theseval- −1.96 −0.21 uesareingoodagreementwiththosefoundwiththebest-fitmodel 4.3 AphysicalinterpretationwithMytorusmodel described in table1. Thephoton index of theprimary power-law Themodelsdiscussedsofar,whicharebasedonspectralcompo- component is now Γ = 1.86−0.11 and the intrinsic emitted lu- +0.17 nentslargelyusedfromtheastronomical community,donottreat minosity is L ∼ 1.4×1042erg s−1. Using the Suzaku [2−10]keV bothfluorescentlinesandcontinuumcomponentsself-consistently. and Swift data, we found that a good fit can be obtained with an Furthermoreallthesespectralcomponentsmaybedeficientinone absorberproducingthereflectedcomponentshavinganN statis- H ormoreaspectofmodellingthecomplextransmissionandreflected ticallyconsistentwiththatobtainedusingtheXMM-Newtondata spectrumofAGNoverabroadenergyrangeandforalargerangeof (thussuggesting thatthiscomponent islikelyassociated withthe absorbingcolumndensities(seesection2ofMurphyandYaqoob, distantreflectorortorus),whileourlineofsighttothecentralen- 2009foracriticaldiscussionofthesepoints). gineinterceptsacolumndensityN =(0.88±0.09)×1024cm−2. H Inordertoalleviatetheseproblemsand,thus,tofurtherassess In summary thisanalysis, which is based on amodel which the possible geometry and/or nature of the variable absorber, takesintoaccountconsistentlythephysicalprocessinplacewithin we tested the most recent model for the toroidal reprocessor 8 the X-ray absorber, shows that the change of state of NGC454E (Murphy&Yaqoob 2009). This model, recently included in the canbeunderstood simplybyachancechangeintheline-of-sight XSPEC software package, is valid for column densities in the obscuration(∆N ∼6×1023cm−2)whiletheglobalobscurerre- H range 1022 to 1025 cm−2 and for energies up to 500 keV (the mainsunchanged.TheintrinsicluminosityderivedfromtheSuzaku relativisticeffectsbeingtakenintoaccount);moreimportantlythe dataisL ∼ 2.4×1042ergs−1.WenotethatusingMy- [2−10]keV reprocessed continuum and fluorescent line emission are treated torusthederivedchangeoftheintrinsicluminositybetweenthetwo self-consistently for the first time. This model assumes that the datasetsareinbetteragreement(afactor1.7)withrespecttothose absorbergeometryistoroidalwithanopeningangleof60◦.Since found in section 4.2. However, with the present statistic and the weareclearlyseeingavariationoftheabsorbingcolumndensity complexityoftheobservedspectra,wecannotruleoutorconfirm along the line of sight we have used a spectral configuration of apossiblevariationinluminosityofaboutafactor2,frequentlyob- MyTorus that can mimicaclumpy absorber and which also takes servedinAGN;indeedbycomparingthe54-monthsand9-months BAT high energy (14–195 keV) spectra of this source, we found thattheintensityishigherinthe9-monthsspectrum.However,fit- 8 http://www.mytorus.com/ tingthespectrawithasingleabsorbedpower-lawcomponent, we NGC454:unveilinga new“changinglook”AGN 9 foundthataconstantfluxisalsowellwithintheerrorsonthebest- of its distance from the central black hole (i.e. within 10−3 pc), fitnormalizationsofthisprimarypower-law. themostlikelylocationforthisabsorberismuchcloserinthanthe stableandratherneutralone. 5 SUMMARYANDCONCLUSION WehavepresentedtheresultsofSuzaku,XMM-NewtonandSwift ACKNOWLEDGEMENTS observations of the interacting system NGC454 (z=0.0122). The bulk of the measured 2–10 keV emission comes from the active WewarmlythankT.Yaqoobfortheusefuldiscussionandforhelp- galaxyNGC454E(L ∼ 2×1042ergs−1);noemission ing us while fitting the Mytorus model to mimic the variable ab- [2−10]keV from the center of the companion galaxy (NGC454W) in the sorber. Wethank theanonymous refereefor many useful sugges- interacting system is detected. The nuclear X-ray emission of tions and comments that significantly improved the paper. This NGC454Eisfilteredbyanabsorbingcolumndensitytypicalofa research has made use of data obtained from the Suzaku satel- Seyfert2galaxy,inagreementwiththeopticalclassification. liteanddataobtainedfromtheHighEnergyAstrophysicsScience ArchiveResearchCenter(HEASARC),providedbyNASA’sGod- A comparison between Suzaku and XMM-Newton observa- dardSpaceFlightCenter.Theauthorsacknowledgefinancialsup- tions (taken 6 months later) revealed a significant change in the port from ASI (grant n. I/088/06/0, COFIS contract and grant n. spectraofNGC454Eintheenergyrangebetween3and6keV.This I/009/10/0).VBacknowledgesupportfromtheUKSTFCresearch variationcanbewellexplainedbyavariabilityof about anorder council. of magnitude in the absorbing column density along the line of sight:from∼1×1024cm−2(Suzaku)to∼1×1023cm−2(XMM- Newton). This study also adopted the most recent model for REFERENCES the toroidal reprocessor (Murphy&Yaqoob 2009), which takes into account consistently the physical processes in place within Antonucci,R.1993,ARA&A,31,473 the X-ray absorber. Furthermore, regarding the XMM-Newton Arp,H.C.,&Madore,B.F.1987,ACatalogueofSouthernPe- spectrum,wedetectedastatisticallysignificantabsorptionfeature culiarGalaxiesandAssociations(Cambridge:CambridgeUniv. a∼6.7keV,aclearsignatureofthepresenceofaionisedabsorber, Press) withionisationparameterlog(ξ/ergcms−1) = 3.55andcolumn Baumgartneretal,2011ApJS,submitted densityN = 6.05×1023 cm−2.Theabsenceofthisfeaturein Bianchi,S.,Guainazzi,M.,Mattm,G.,etal.2005,A&A,442,185 H the Suzaku spectrum, despite its detectability, implies that it has Bianchi,S.,Piconcelli,E.,Chiaberge,M.,Bailo`n,E.J.,Matt,G., varied between the two observations. Absorption linesassociated &Fiore,F.2009,ApJ,695,781 withionized ironhavebeennow observed inseveral sources and Boldt,E.1987,Phys.Rep.,146,215 there is also a clear evidence that these lines are variable as in Braito,V.,Reeves,J.N.,Dewangan,G.C.,George,I.,Griffiths, the case of NGC454. Furthermore, in some cases the measured R.E.,Markowitz,A.,Nandra,K.,Porquet,D.,Ptak,A.,Turner, blue-shiftsoftheenergycentroidsimplyalargevelocityofthese T.J.,Yaqoob,T.,&Weaver,K.,2007,ApJ,670,978 absorbersandalikelyassociationwithpowerfuldiskwinds(King Cappi,M.,Tombesi,F.,Bianchi,S.,etal.2009,A&A,504,401 &Pound2003),whileinothercasesthereisnomeasurablemotion CusumanoG.,LaParolaV.,SegretoA.,FerrignoC.,MaselliA., asinourcase. SbarufattiB.,RomanoP.,ChincariniG.,GiommiP.,MasettiN., MorettiA.,ParisiP.,TagliaferriG.,2010,A&A,524,A64 In summary, with respect to the absorbing column density Dadina, M., Cappi, M., Malaguti, G., Ponti, G., & de Rosa, A. variability, NGC454E is a new member of the class of “chang- 2005,A&A,442,461 ing look” AGN, i.e. AGN that have been observed in both Dadina,M.2008,A&A,485,417 Compton-thin (N =1023cm−2) and reflection dominated states DellaCeca,R.,Ballo,L.,Tavecchio,F.,etal.2002,ApJ,581,L9 H (N > 1024cm−2).Apossiblescenarioisthatastableandlikely denHerder,J.W.,etal.2001,A&A,365,L7 H distant absorber responsible for the iron emission line is present. Dickey,J.M.,&Lockman,F.J.1990,ARA&A,28,215 However, there is also a clear variation of the N of the line DoneC.,2010,arXiv,arXiv:1008.2287 H of sight absorber, probably indicative of the clumpy nature of Ehle,M.,etal.2001,XMM-NewtonUsersHandbook the rather neutral absorber itself. Unfortunately the comparison Elvis,M.,Risaliti,G.,Nicastro,F.,Miller,J.M.,Fiore,F.,&Puc- between different observations, typically performed at intervals cetti,S.2004,ApJ,615,L25 ofmonthstoyears(asthosediscussedhere),providesonlyupper Fabian,A.C.2010,IAU limits to the intrinsic time scales of N variations and thus on FrancisP.J.,HewettP.C.,FoltzC.B.,ChaffeeF.H.,Weymann H thepossiblelocationof thethicker obscuring material(obscuring R.J.,MorrisS.L.,1991,ApJ,373,465Symposium,267,341 “torus” vs. Broad Line Region clouds). The low exposure of the Gruber,D.E.,Matteson,J.L.,Peterson,L.E.,&Jung,G.V.1999, Swift XRT 2006 observations, when the source was in a state ApJ,520,124 similar to the XMM-Newton one, did not allow us to establish Gehrels, N., Chincarini, G., Giommi, P., et al. 2004, ApJ, 611, the N variability on smaller time scales (i.e. intra-day) of the 1005 H singleobservations.Animprovementoftheestimatesofvelocity, Ghisellini,G.,Haardt,F.,&Matt,G.2004,A&A,413,535 distance and size from the central X-ray source of the obscuring Jansen,F.,etal.2001,A&A,365,L1 materialcouldbeobtainedonlythroughmonitoringobservational Johansson,L.1988,A&A,191,29 campaigns withinafew daysor weeks and/or through thesearch Kaastra,J.S.&Mewe,R.1993,A&AS,97,443 for N variations within single long observation. For what con- Kallman,T.R.,Palmeri,P.,Bautista,M.A.,Mendoza,C.,&Kro- H cernstheionisedabsorber,asderivedfromourfirstorderestimate lik,J.H.2004,ApJS,155,675 10 Marcheseetal. Kaspi,S.,Maoz,D.,Netzer,H.,etal.2005,ApJ,629,61 King,A.R.,&Pounds,K.A.2003,MNRAS,345,657 Kokubun,M.,etal.2007,PASJ,59,53 Koyama,K.,Tsunemi,H.,Dotani,T.,etal.2007,PASJ,59,23 Lobban,A.P.,Reeves,J.N.,Miller,L.,etal.2011,MNRAS,414, 1965 Magdziarz,P.,&Zdziarski,A.A.1995,MNRAS,273,837 MateosS.etal.,2008,A&A,492,51 Mewe, R., Gronenschild, E.H.B.M., and van den Oord, G.H.J. 1985,A&AS,62,197 Murphy,K.D.,Yaqoob,T.,&Terashima,Y.2007,ApJ,666,96 Murphy,K.D.,&Yaqoob,T.2009,MNRAS,397,1549 Mitsuda,K.,etal.2007,PASJ,59,1 Pounds,K.A.,Reeves,J.N.,Page,K.L.,&O’Brien,P.T.2004, ApJ,616,696 Proga,D.,&Kallman,T.R.2004,ApJ,616,688 Puccetti,S.,Fiore,F.,Risaliti,G.,Capalbi,M.,Elvis,M.,&Nicas- tro,F.2007,MNRAS,377,607 Ross&Fabian(2005),MNRAS,358,211 Risaliti,G.,Elvis,M.,&Nicastro,F.2002,ApJ,571,234 Risaliti,G.,Elvis,M.,Fabbiano,G.,Baldi,A.,&Zezas,A.2005, ApJ,623,L93 Risaliti,G.,Elvis,M.,Fabbiano,G.,Baldi,A.,Zezas,A.,&Sal- vati,M.2007,ApJ,659,L111 Risaliti,G.,etal.2009,ApJ,696,160 Risaliti,G.2010,inAmericanInstituteofPhysicsConferenceSe- ries,Vol.1248,AmericanInstituteofPhysicsConferenceSeries, ed.A.Comastri,L.Angelini,&M.Cappi,351354 Risaliti G., Elvis M., Bianchi S., Matt G., 2010, MNRAS, 406, L20 Shu, X. W., Yaqoob, T., Murphy, K. D., et al. 2010, ApJ, 713, 1256 Sim,S.A.,Proga,D.,Miller,L.,Long,K.S.,&Turner,T.J.2010, MNRAS,408,1396 Spergel,D.N.,etal.2003,ApJS,148,175 Stiavelli,M.,Panagia,N.,Carollo,M.C.,Romaniello,M.,Heyer, I.,Gonzaga,S.1998,ApJ492,L135 Stru¨der,L.,etal.2001,A&A,365,L5 Takahashi,T.,etal.2007,PASJ,59,35 Tombesi F.,Cappi M.,ReevesJ.N.,PalumboG.G.C.,Yaqoob T.,BraitoV.,DadinaM.,2010,A&A,521,57 Turner,T.J.,Miller,L.,Kraemer,S.B.,Reeves,J.N.,&Pounds, K.A.2009,ApJ,698,99 TombesiF.,CappiM.,ReevesJ.N.,PalumboG.G.C.,BraitoV., DadinaM.,2011,ApJaccepted Turner,M.,etal.2001,A&A,365,L27 TurnerT.J.,MillerL.,2009,A&ARv,17,47 Uttley, P., Taylor, R. D., McHardy, I. M., et al. 2004, MNRAS, 347,1345 Weaver,K.A.,Nousek,J.,Yaqoob,T.,etal.1996,ApJ,458,160 Wilms,J.,Allen,A.,&McCray,R.2000,ApJ,542,914 Yaqoob,T.1997,ApJ,479,184