Mon.Not.R.Astron.Soc.000,1–9(2003) Printed2February2008 (MNLATEXstylefilev2.2) XMM observations of the high-redshift quasar z RX J1028.6-0844 at =4.276: soft X-ray spectral flattening W. Yuan1⋆, A.C. Fabian1, A. Celotti2, R.G. McMahon1, and M. Matsuoka3 1Universityof Cambridge, Institute of Astronomy, Madingley Road, Cambridge, CB3 0HA; E-mail: [email protected] (wy) 5 2SISSA, via Beirut 2-4, 34014 Trieste, Italy 0 3Japan Aerospace Exploration Agency (JAXA), Tsukuba Space Center, Tsukuba, Ibaraki 305-8505, Japan 0 2 n Accepted forpublication a J 2 ABSTRACT 1 We present results from a new XMM-Newton observation of the high-redshift quasar RXJ1028.6-0844ataredshiftof4.276.The softX-rayspectralflattening,asreported 2 by a study with ASCA previously (Yuan et al. 2000, ApJ 545, 625), is confirmed to v be present, however,with reduced column density when modelled by absorption.The 5 inferredcolumndensity forabsorptionintrinsic tothe quasaris2.1(+0.4)× 1022cm−2 5 −0.3 for cold matter, and higher for ionised gas. The spectral flattening shows remarkable 2 similaritywith those oftwo similar objects,GB1428+4217(Worsley etal. 2004,MN- 0 1 RAS 350, L67) and PMNJ0525-3343 (Worsley et al. 2004, MNRAS 350, 207). The 4 results improve upon those obtained from a previous short-exposure observation for 0 RXJ1028.6-0844 with XMM-Newton (Grupe et al. 2004, AJ 127, 1). A comparative / studyofthetwoXMM-Newtonobservationsrevealsachangeinthepower-lawphoton h index from Γ ≃1.3 to 1.5 on timescales of about one year. A tentative excess emis- p - sion feature in the rest-frame 5–10keV band is suggested, which is similar to that o marginally suggested for GB1428+4217. r st Keywords: galaxies:active–galaxies:individual:RXJ1028.6-0844–X-ray:galaxies a : v i X 1 INTRODUCTION tric absorption of soft X-rays by associated medium with ar column densities of 1022−23cm−2, although intrinsic spec- High-redshift quasars are powerful cosmological probes to tralflatteningcannotbeexcluded.Thephysicalimplication study the evolution of massive black holes and quasar en- ofthiseffecthasbeendiscussedextensivelyintheliterature vironments in the early universe. Previous X-ray observa- in terms of excess absorption (Elvis et al. 1998, Yuanet al. tions suggested the presence of soft X-ray spectral flatten- 2000,Fabianetal.2001a,b)andintrinsicbreaksintheX-ray ing1 in some radio-loud quasars at redshifts z=2–3 (Wilkes spectra of blazars (Fabian et al. 2001a,b). et al. 1992, Elvis et al. 1994, Cappi et al. 1997, Fiore et The contemporary X-ray observatories XMM-Newton al. 1998, Yuan& Brinkmann 1998, Reeves& Turner2000). andChandra should beable totest thesepreviousfindings. This result is strengthened and extended to higher red- Indeed,softX-rayspectralflatteninghasbeenconfirmedto shifts by its detection in a few extremely X-ray/radio-loud quasarsatz>4,namelyRXJ1028.6-0844(Yuanetal.2000), be present in GB1428+4217 and PMNJ0525-3343 (XMM- Newton, Worsley et al. 2004a,b), and in some other objects GB1428+4217 (Boller et al. 2000, Fabian et al. 2001b), and PMNJ0525-3343 (Fabian et al. 2001a), with ROSAT, at lower redshifts, e.g. PKS2126-0158 at z=3.27 (XMM- ASCA,andBeppoSAX. Theseobjects seem tohavecharac- Newton,Ferrero&Brinkmann2003;BeppoSAX,Fioreetal. 2003). Tentative evidence was also found in the combined teristicstypicalofblazars(Fabianetal.1997,1998;Zickgraf spectra of several z > 4, moderately radio-loud quasars etal.1997,Moran&Helfand1997,Hook&McMahon1998). (Chandra , Bassett et al. 2004). The most plausible explanation for this effect is photoelec- RXJ1028.6-0844 was first detected as an X-ray source intheROSATAll-skySurvey(RASS)andwasidentifiedas a quasar at z=4.276 (Zickgraf et al. 1997). It is also a ra- ⋆ Present address: Yunnan Astronomical Observatory, Chinese dio source (PKSB1026-084) with a flux density of 220mJy AcademyofSciences,PhoenixHill,POBox110,Kunming,Yun- at 5GHz (Otrupcek & Wright 1991) and a flat radio spec- nan,650011China;E-mail:[email protected] 1 Thiseffectiscommonlyreferredtoasexcessabsorptioninthe trum. Its X-ray colors in the ROSAT energy band imply a literature.Hereweuseageneralisedterminconsiderationofpos- hard spectrum (Zickgraf et al. 1997). Its first X-ray spec- siblealternativeexplanations. trum, as obtained with a long ASCA observation made 2 W. Yuan, et al. in 1999, flattens substantially towards soft X-ray energies the source information on the detectors. The EPIC spectra (Yuan et al. 2000); an excess (cold) absorption model re- werere-binnedtohaveaminimumof30countsineachbin. quired a column density of ∼ 2×1022cm−2 for local ab- Thesourceprofileinthe0.3–2keVbandwascompared sorber or ∼ 2 × 1023cm−2 for absorber intrinsic to the against the point spread function of the detectors (FWHM quasar. A later short-exposure XMM-Newton observation of 5′′for MOS and 6′′for PN) and was found to be con- found, however, only marginal evidence for excess absorp- sistent with a point-like source. There was no significant tion (Grupe et al. 2004). In this paper we report on a new variability found during the 40ks exposure, though a ≃10 XMM-Newton observation of RXJ1028.6-0844 with an ex- per cent drop in count rates (averaged over ∼5ks) was posuremuchlongerthanthepreviousobservation.Themea- marginally detected, from 0.316±0.09ctss−1 at the begin- suredX-rayspectrum—withsubstantiallyimprovedphoton ning to 0.271±0.08ctss−1 towards the end of the observa- statistics—confirms the presence of the soft X-ray spectral tion. flatteningasdetectedbyASCA.Theobservationsanddata reduction are described in Sect.2. We present the spectral analysis in Sect.3, including a re-analysis of the previous XMM-Newtonobservation.Discussionoftheresultsisgiven 3 X-RAY SPECTRAL ANALYSIS inSect.4,includingcomparisonswithpreviousobservations Due to an increase in the surface charge loss properties of andwithothersimilarobjects.Conclusionsaresummarised the CCDs, which degrades the energy resolution, there has in Sect.5. We adopt H0=71kms−1Mpc−1, ΩΛ=0.73, and beenatime-dependent,significantchangeinthelow-energy Ωm=0.27. The Galactic column density in the direction of redistribution properties of the MOS cameras. This effect RXJ1028.6-0844 is NGal = 4.59 × 1020cm−2 (Dickey & H has been taken into account in the most up-to-date MOS Lockman 1990). Errors are quoted at the 1σ level for one calibrationfiles;however,somesmallsystematicuncertainty parameter of interest unless stated otherwise. maystillremainatbelow0.5keV(Kirsch2004,Kirschetal. 2004). To minimise possible biases induced in the results, we treat the MOS spectra in two ways and compare the results.Thefirstwastosimplyomitthespectralrangebelow 2 OBSERVATION AND DATA REDUCTION 0.5keV; the second was to use the spectral range down to 0.3keVandintroduceasystematicerrorof2percentinthe The quasar RXJ1028.6-0844 was observed with XMM- 0.3–0.5keV band (as recommended in Kirsch 2004, Kirsch Newton on June 13th, 2003 during satellite revolution 643 et al. 2004). As seen in Table2, these two methods yielded (observationID0153290101). TheEPIC(EuropeanPhoton statistically consistent results. We thus formally quote the Imaging Camera) MOS1, MOS2, and PN cameras were op- resultsobtainedusingthe0.3–10keVband.XSPEC(v.11.3) eratedinthe‘primaryfullwindow’imagingmodeandathin was used for spectral fitting. filterwasusedtoscreenoutoptical/UV light.Theobserva- tional log is shown in Table1. The XMM-Newton Science AnalysisSystem (SAS,v.6.0) and themost up-to-datecali- 3.1 Soft X-ray spectral flattening brations(August2004)wereusedfordatareduction.Wefol- lowed standard datareduction andscreening procedures.A 3.1.1 Power-law model with local cold absorption fraction oftheobservation period sufferedfrom high flaring Webeganbyfittingthespectraofeachdetectorindividually background caused by soft protons. By inspecting the light withasinglepower-lawmodel2modifiedbyneutralabsorp- curve of energy E >10keV, single events in the whole field tion with a column density NH as a free parameter. This of view, these periods were identified as having count rates model gave acceptable3 fits to the PN and the MOS1 spec- higher than 1ctss−1 and 0.35ctss−1 for PN and MOS de- tra (see Table2), but not to the MOS2 spectrum [reduced tectors,respectively,asrecommendedbytheXMM-Newton χ2=1.4for53degreeoffreedom(d.o.f.),i.e.anullhypothesis Science Operation Centre (SOC). probability of P =0.02 only]. Inspection of the χ2 residu- null The quasar was detected at a sky position alsofthefitsingledoutoneenergybinat1.82keV(awidth RA=10h28m38s.84, Dec=−08o44′38′′.3 (J2000), 0′′.6 of80eV),whichcontributed11outofthetotalχ2of76.The away from the position of its radio counterpart (Simbad feature can be fitted with either a narrow notch feature at database) and 2′′.1 of its optical counterpart (Zickgraf E=1.83±0.02keV(awidthof49+55eVandacoveringfrac- −19 et al. 1997). Source X-ray events were extracted from a tion 0.99+0.01) or a Gaussian absorption line (E=1.83keV) circle of 32′′radius, which corresponds to the ≃87 per −0.53 ofinfinitelysmallwidth.Thisenergycorrespondsto9.7keV cent encircled energy radius. Background events were ex- in the quasar’s rest-frame, at which no known physical ab- tracted from source-free regions using a concentric annulus sorption feature is present. On the other hand, it is coinci- ′′ of 52/128 radii for the MOS detectors, and circles of dent with the instrumental absorption feature at 1.84keV 32′′radius at the same CCD read-out column as the source due to the Silicon edge. Furthermore, it appears in neither position for the PN detector. X-ray images, light-curves, the PN nor the MOS1 spectrum. Excluding the 1.82keV and spectra were generated from the extracted, cleaned bin reduced the χ2 by 12, and resulted in an acceptable fit events for the source and background. Parameters for data screening and source extraction are listed in Table1. No photon ‘pile-up’ problem was found, as expected. noreffect 2 Fpho(E)∝E−Γ,whereΓisthephotonindex. of low-energy noise above 0.2keV. We used the spectral 3 Weregardamodelfitasacceptableifthenullhypothesisproba- range of 0.2–10keV for PN and 0.3–10keV for MOS. The bilityderivedfromthefitisgreaterthan10percent(Pnull>0.1), EPIC response files (rmf and arf) were generated using asiscommonlyquoted. XMM observations of the high-redshift quasar RXJ1028.6-0844 3 Table 1.SummaryoftheXMMobservations anddatareductioninformationforeachEPICdetector PN MOS1 MOS2 observationduration(ks) 41.5 43.1 43.1 goodexposure(ks) 15.8 21.3 22.1 energybandused(keV) 0.2–10 0.3-10 0.3–10 events patternused 0–4 0–12 0–12 sourceextractionradius 32′′ 32′′ 32′′ source+BGDcounts 5024 2006 2080 netsourcecounts 4854 1920 1986 sourcecountrate(10−2ctss−1) 30.7±0.5 9.0±0.2 9.0±0.2 fluxa (10−12ergs−1cm−2 ) 1.16±0.04 1.14±0.07 1.14±0.06 fluxb (10−12ergs−1cm−2 ) 1.17±0.04 1.15±0.07 1.15±0.06 a Fluxinthe1–10keVband;themodelofapower-lawwithintrinsicabsorptionatz=4.276plusNGal isused. H b Galacticabsorptioncorrectedfluxinthe1–10keVband;thesamemodelasinaisused. (χ2=1.2 for 52 d.o.f.) with fitted parameters in good agree- ment with those for MOS1 within 1-σ errors. We therefore considerthedisagreementbetweenMOS2andMOS1/PNto be due to inadequate MOS2 calibration around the Silicon edge, and ignore this energy bin hereafter. Parameters (in- cluding normalisations) for the two MOS spectra were tied together tobethesame in joint fitting(MOS1+2). Thefit- ted NH of ∼ 1.1× 1021cm−2) is significantly higher than NGal (Table2).ThephotonindexisnowΓ≃1.55,typicalof H blazars. A power-law with fixed Galactic absorption (4.59×1020cm−2) yielded unacceptable fit and a flat photon index (Γ=1.3–1.4, Table2). The improvement in χ2 for the fit with freely fitted NH over that with fixed NH=NHGal is substantial—χ2 was reduced by 36 for PN and by 31 for a joint MOS1+2 fit for one additional free parameter. Applying the F-test4 (Bevington & Robinson 1992) gives a chance probability ≪0.001. For fixed NH=NHGal, acceptable fit was obtained only within the restricted 1–10keV range. We plotted in Fig.1 the data, the best-fit model (for a joint PN and MOS1+2 spectral fit) and its extrapolation down to the low-energy Figure1. ThespectraofthePN(circles),MOS1(squares),and end of the detectors, and the data-to-model ratio as resid- MOS2 (stars) cameras and the residuals as data-to-model ratio. uals. A systematic deficit of photons below 1keV is clearly The model is the best-fit power-law (with Galactic absorption) indicated. Fig.2 shows the confidencecontoursfor thefree- to the joint PN+MOS1+MOS2 spectra within the restricted 1– fitted total NH (Galactic plus excess) and Γ for PN and 10keVenergybandandisextrapolateddowntothelowenergies. MOS1+2, respectively. Absorption in excess of NHGal is ev- A systematic deviation from a power-law model with Galactic ident. It is noted that PN gives systematically lower NH absorptiontowardslowenergies(below1keV)isclearlyindicated. values than MOS. We regard the results from MOS to be more reliable than those from PN in consideration of the EPIC cross calibration uncertainties as discussed in details 3.1.2 Absorption intrinsic to the quasar in AppendixA. This is the most plausible postulate in consideration of the statistical argument presented in previous work (see Introduction for references). For neutral absorber, NH≃ 2×1022cm−2 was foundassumingcosmic abundances.The confidence contours for the excess NH versus Γ are shown 4 Recently,Protassovetal.(2002)havequestionedthevalidityof in Fig.3 for PN (dashed) and MOS1+2, respectively. If usingtheF-testtotestforthesignificanceofaddinganadditional the metallicity of the absorber is lower than the cosmic spectral component, as this involves testing a hypothesis that is value, which is not unexpected at such a high redshift— ontheboundaryoftheparameterspace.WenotethattheF-test could be about 10 per cent or less (e.g. Lu et al. 1996, Pet- weusedhereisnotaffected bytheboundaryconditionproblem, because wearetesting forthesignificance ofthe differentvalues tini et al. 1997, Prochaska & Wolfe 2000), the NH would in NH between NH=NHGal (null hypothesis) and the free-fitted be correspondingly higher than the value give here. The value,ratherthanfortheadditionofamodelcomponent. absorption-corrected luminosity in the quasar rest-frame is 4 W. Yuan, et al. Table 2.ResultsofX-rayspectralfits Detector NHa Γ χ2/dof Pnbull local neutral absorption, free NH PNc 8.7±0.7 1.53±0.03 137.8/147 0.69 MOS1(0.5–10keV) 12.8±2.5 1.54±0.06 57.7/52 0.27 MOS1(0.3–10keV) 12.7±2.0 1.54+−00..0063 68/67 0.44 MOS2(0.5–10keV) 11.3±2.5 1.57+−00..0063 63.7/52 0.13 MOS2(0.3–10keV) 9.0±1.6 1.53±0.05 76/67 0.21 MOS1+2(0.5–10keV) 12.1±1.7 1.56+−00..0052 122.8/107 0.14 MOS1+2(0.3–10keV) 11.1±1.2 1.54±0.04 150/137 0.21 local neutral absorption, fixed NH PN 4.59(fix) 1.37±0.02 173/148 0.07 MOS1(0.5–10keV) 4.59(fix) 1.36±0.04 69/53 0.07 MOS1(0.3–10keV) 4.59(fix) 1.30±0.03 91/68 0.03 MOS2(0.5–10keV) 4.59(fix) 1.43±0.04 71/53 0.05 MOS2(0.3–10keV) 4.59(fix) 1.40±0.03 84/68 0.09 MOS1+2(0.5–10keV) 4.59(fix) 1.40±0.03 141/108 0.017 MOS1+2(0.3–10keV) 4.59(fix) 1.35±0.02 181/138 0.008 power-law + NGal and excess absorption at z=4.276 H Detector NHexc Γ χ2/dof Pnull PN 149±30 1.49±0.03 136/147 MOS1+MOS2(0.5–10keV) 393+−9790 1.53+−00..0042 122/107 0.15 MOS1+MOS2(0.3–10keV) 222±46 1.49±0.03 153/137 0.17 brokenpower-law + Galactic absorption (PN+MOS1+2) Ebdreak Γlow−E χ2/dof Pnull PN 0.50+−00..0043 −1.55+−0.70 133/146 0.77 MOS1+2(0.5–10keV) 1.1+−00..11 0.86+−00..1179 122/105 0.12 MOS1+2(0.3–10keV) 1.1+−00..21 0.92+−00..1203 148/135 0.21 a Columndensityofhydrogeninunitsof1020cm−2 b Nullhypothesis probabilityoftheχ2 testforthefit c TheNH valuesfromPNarepossiblyunder-estimated; seeSect.4.1.1andAppendixAfordiscussion. d The break energy of broken power-law inkeV. When no errors are given, the value is unconstrained within a physically meaningful range. 9.2×1046ergs−1 in 1–10keV and 2.64×1047ergs−1 in 1– tionsarecombinedinordertoachieveabetterconstrainton 50keV,respectively. the parameters (see Sect.3.2.2). With increasing ionisation state,therequiredcolumndensityincreasesfrom∼2×1022 If the absorber is close enough to the central source, to ∼1×1023cm−2. thegas islikelytobeionised. Indeed,theopticalspectra of No redshifted K-shell absorption edges from iron ions RXJ1028.6-0844 takenbyPerouxetal.(2001)andZickgraf (Eedge=7.1–9.3keVintherest-framefromFeItoFeXXVI)are et al. (1997) show no significant Lyman limit absorption at detected.Assuming cosmic abundanceof iron (3.4×10−5), 912˚A.Anestimateoftheoptical depthof theLymanlimit a K-shell edge with optical depth τ ∼ 0.02–0.1 is pre- absorption places an upper limit on the neutral hydrogen column density to be NH I . 10−17cm−2 along the line-of- idriocnte.dTfhroismisthceoanbsiosvteenNtHwritahngteheforupnpeuertralilm–hitigholny iτoneissetid- sight to the optical–UV emission region (Yuan et al. 2000). mated from the joint PN and MOS1+2 spectra, which IftheX-rayabsorberalsointerceptstheoptical–UVlight,at ranges from 0.1 to 0.2 (90 per cent level) for ions from least moderate ionisation of thegas isrequired.Thelack of FeI to FeXXVI (absorption cross section from 3.8×10−20 strong optical–UV extinction can be explained by ionised, to 3.3×10−20cm2atom−1). dust-free absorber. It is interesting to note that such an optical–UV property seems to be common among this type ofobjects(Yuanetal.2000,Fabianetal.2001a,b).Wetried 3.1.3 Intrinsic spectral break tomodeltheexcessabsorption withionisedabsorption (ab- sori in XSPEC). The ionisation parameter (as defined in Wealso considered thepossibility that the soft X-rayspec- Done et al. 1992), however, could not be constrained with tral flattening is an intrinsic feature. The physical implica- the current data ranging from an almost neutral to highly tion of a break in the intrinsic X-ray spectra of blazars has ionisedabsorber.Theresultsofsuchananalysisaregivenin beendiscussedinFabianetal.(2001a,b) inthecontextofa Fig.4, where thedata from thetwo XMM-Newton observa- cut-offin theenergydistributionofelectrons orofsoft seed XMM observations of the high-redshift quasar RXJ1028.6-0844 5 3.2 Comparison with a previous observation and spectral variability 3.2.1 A previous XMM-Newton observation RXJ1028.6-0844 was previously observed with XMM- Newton in revolution 445 with a short exposure of 7ks in May, 2002 (PI. S. Mathur, observation ID: 0093160701). The results, as published in Grupe et al. (2004), gave ab- sorption NH values similar to what we find here, but, a flatter spectral slope of Γ ≃ 1.3. In order to achieve a self- consistent comparison of the two observations—free from the effects introduced by different versions of the evolving calibration and data processing software, we also analysed the data from that observation. The data were taken from theXMM-Newton science archive. The observation was de- scribed in Grupe et al. (2004). We used exactly the same Figure 2. Confidence contours of fitted total column density data screening criteria and source/background extraction and photon index for the model of a power-law with local neu- regions as used for the current observation (see Sect.2 and tral absorption. The contours are at the 68, 90, and 99 per cent Table1).Thegoodexposureandsourcecountrateare3.7ks confidence,respectively,fortwointerestingparameters.Solidcon- tours:thejointMOSspectrain0.3–10keV;dashedcontours:the and 0.39ctss−1 for PN (0.2–10keV), and 6.9/7.0ks and PN spectrum. Also indicated by lines are the Galactic column 0.12/0.13ctss−1 for MOS1/2 (0.3–10keV). A comparison density(dashed)anditsconservative30%uncertaintyrange(dot- with Table1 reveals that the broad-band count rates were ted).AbsorptioninexcessoftheGalacticvalueisevidentforsuch higherinthe2002observationthanin2003byabout30per amodel. cent. The PN and MOS spectra were binned to have a minimum of 25 and 20 counts in each bin, respec- tively. The results of the spectral fits are in good agree- ment with those obtained by Grupe et al. (2004). For an absorbed power-law model, the total local absorption NH is 10.1(+−21..02)×1020cm−2 for joint MOS1 and MOS2 (MOS1+2)spectraand6.0(+1.3)×1020cm−2forPN.Again, −1.0 thefittedNH is systematically lower for PN than for MOS, asfoundinthe2003observation.WhiletheabsorptionNHis in good agreement between the two observations, the pho- ton indices are not. The photon indices obtained for the 2002 observation are Γ2002=1.30±0.06 and 1.27±0.05 for MOS1+2andPN,respectively,i.e.thespectrumwassteeper during the observation of 2003 (Γ2003 = 1.53±0.03). This result can be seen clearly in Fig.3, where excess (neutral) absorption NH is plotted versusΓ for the2002 observation, assuming absorption is intrinsic to the quasar at z=4.276. Figure3. Columndensityofintrinsicabsorberversuspower-law The spectral steepening remains even when only the hard photon index as measured with MOS (solid) and PN (dashed) band 2–10keV spectra were considered (Γ2002=1.33+−00..0184 detectors. The absorber is assumed to be at the quasar redshift and 1.23+0.08 for MOS1+2 and PN, respectively). The −0.13 4.276andtohavecosmicabundances.Thecontourscorrespondto Galactic absorption corrected flux in the 1–10keV band is confidence levelsof68,90,and99percent.Theresultsfromthe 1.9×10−12ergs−1cm−2 (averagedMOS1andMOS2value). current (2003) and the previous XMM observations (2002) are plotted. The variation in the spectral slope is significant, while thereisnochange intheabsorptioncolumndensity. 3.2.2 Joint spectral fit of the two observations We quantified the spectral variability by fitting jointly the photons for Compton scattering. We modelled the spectral spectra of the two observations. We used the MOS spectra flattening with, as an approximation, a broken power-law only, in consideration of possible PN calibration uncertain- modified by local absorption. Acceptable fits could be ob- ties (see AppendixA). For MOS1 and MOS2 spectra from tained for both PN and MOS1+2. However, NH and the the same observation, all parameters (including normalisa- low energy photon index Γlow−E could not be constrained tion) weretied together. Theresultsare summarised in Ta- due to strong coupling. We thus fixed NH=NHGal. The fit- ble3. tedhigh-energy photonindicesΓhigh−E are1.44±0.04 (PN) Firstly, since the fitted NH of the two observations are and1.49±0.06(MOS1+2),andthebreakenergyE and ingoodagreement,weassumedthattherewasnovariability break low-energy index Γlow−E are listed in Table2. This model in the absorption. Thus the NH values for the two observa- gaveacceptablefitswhich arestatistically indistinguishable tions were tied together in the fitting. Absorbed power-law from the models of power-law with either local or intrinsic modelswereused.Asatest,wetiedΓ2002=Γ2003 togetherin absorption. thefitting,whichresultedinafitonlymarginallyacceptable 6 W. Yuan, et al. (χ2=242 for 219 d.o.f.). Setting thetwo indices as indepen- Table 3. Joint fits to the current (2003) and a previous (2002) dent parameters improved the fit significantly, reducing χ2 XMM-Newton observations. Only the MOS spectra were used. by24foroneadditionalfreeparameter;theF-test[Beving- TheabsorptionNH isassumedtobe thesameinthe two obser- ton & Robinson 1992; see Footnote4 for the argument for vationsandthephotonindexisfreelyfitted. the validity of the F-test used here regarding the bound- aryconditionwarningdiscussedbyProtassov etal.(2002).] givesachanceprobability≪0.001forΓ2002 andΓ2003 being parameter June2003 May2002 thesame.Thefitisgood (χ2=217for218d.o.f.),indicating that the spectra of the two observations can be fitted well power-law with neutral absorption χ2/d.o.f.=217/218 withtwodifferentcontinua(inslopeandnormalisation) at- photonindex 1.53±0.04 1.31±0.04 tenuated by the same amount of absorption. The fluxes in normalisation@1keV(10−4) 1.77±0.07 2.13±0.09 the 0.2–1keV band were comparable in the two observa- totalabsorptionNH (1022) 0.11±0.01(tied) tions, while in the 1–10keV band it was higher by a factor fluxa 0.2–1/1–10keV(10−13) 1.2/11.5 1.4/19.1 of two in May 2002 compared to June 2003 (Table3). The neutral absorption at z=4.276, fixed NGal χ2/d.o.f.=221/218 H two power-law continua, before attenuation by any absorp- photonindex 1.48±0.03 1.26±0.04 tion, cross over at ≃ 0.4keV in the observer’s frame, i.e. normalisation@1keV(10−4) 1.65±0.05 1.99±0.07 ≃2keV in thequasar rest-frame. excessabsorptionNH (1022) 2.1+−00..43 (tied) fluxb 0.2–1/1–10keV(10−13) 1.7/11.6 1.8/19.2 The combined data set improves the spectral photon luminosityc 1–10/1–50keV(1047) 0.92/2.58 1.06/3.88 statistics and gives a better constraint on the excess ab- s=o2rp.1t(i+on0..4)F×or10n2e2uctmra−l2a,bi.seo.rbexecreisnstaribnssoircpttoionthieseqvuiadseanrt,.NAHn bbrroekaeknenpoewrgeyr-(lkaewV,)fixed NHGal χ2/d1..o1.±f.=0.2114/2151.6±0.2 −0.3 low-energyindex 0.92±0.11 0.87±0.09 ionisedabsorbermodelalsogivesagoodfit,suggestingthat high-energyindex 1.50±0.04 1.33±0.07 the ionisation status is unconstrained. The confidence con- normalisation@1keV(10−4) 1.54±0.06 1.75±0.06 tours for the ionisation parameter and NH are shown in Fig.4. a Galacticabsorptionuncorrectedfluxinunitsofergs−1cm−2. Secondly, we tested whether the variation in spectral b Galacticabsorptioncorrectedfluxinunitsofergs−1cm−2. slope could result from variability in the absorption, such c Absorptioncorrectedluminosityinthequasarrest-frameinunits as from the ionisation parameter. We fitted the ionised ab- ofergs−1. sorber model with independent ionisation parameters and NH, but tied the indices together (Γ2003=Γ2002). No satis- factory result could be obtained for this model; the best fit was significantly worse than that with theabovevariable-Γ model (∆χ2 =14). This is not surprising, as the difference in Γ arises primarily in thehard 1–10keVband, which cor- responds to 5.3–53keV in the quasar rest-frame. At such high energies the amount of absorption of X-ray photons is decreasing dramatically. Wealso fittedabrokenpower-law tothedatawith ab- sorption fixed at NGal. There is a marginal indication of a H higherbreak energy 1.6±0.2keVfor theflatterspectrum in May2002thanforthesteeperspectruminJune2003;how- ever, the significance is low. The low-energy photon indices andthenormalisationsat1keV(beloworclosetothebreak energies)arecomparableinthe2002and2003observations. No spectral variability is detected within a 40ks dura- tion in the June2003 observation with XMM-Newton. Figure 4. Confidence contours (at the 68, 90, and 99 per cent level) for the ionisation parameter and column density of the 4 DISCUSSION ionisedabsorbermodel.Theresultisobtainedfromajointspec- tralfittotheMOSspectrafromthetwoXMM-Newtonobserva- 4.1 The presence of soft X-ray spectral flattening tionsin2003and2002(seeSect.3.2.2). 4.1.1 XMM-Newton results WehaveshownthepresenceofsoftX-rayspectralflattening inthez=4.276quasarRXJ1028.6-0844usinganobservation signal-to-noise of their data and consequently the weaker made with XMM-Newton. The result confirms theprevious constraints on NH compared to this work, which benefits report based on ASCA data (Yuan et al. 2000). In the ex- from a much longer exposure. cess absorption scenario, the derived absorption NH from WenotethatPNtendstogivesystematicallylowerNH ouranalysisareconsistentwiththoseobtainedbyGrupeet values than MOS. This is most likely due to discrepancies al. (2004) from apreviousshort XMM-Newton observation. inthecalibration below1keVbetweenPNandMOS,asre- In that study the authors suggested that strong excess ab- portedinthemostrecentXMMcalibrationstatus(Kirshet sorptionwasmarginal.Thisisnotsurprisinggiventhelower al.2004,seealsoXMM-Newton SOC XMM-SOC-CAL-TN- XMM observations of the high-redshift quasar RXJ1028.6-0844 7 0018 5). Although a complete solution is yet to be reached, preliminary indications suggest that PN is most likely the cause of the problem. A detailed discussion on this issue and a quantitative PN–MOS comparison taking the most up-to-date calibration into account is given in AppendixA. In summary, we consider the NH values derived from MOS spectra to be more reliable. Furthermore, it may also be thecasethatbothPNandMOSyieldsystematicallyhigher fluxesat low energies compared to XMM-Newton RGS and Chandra. If this turns out to be the case, the true column in the absorption scenario could be even higher than what is reported here. 4.1.2 Comparison with previous results The total NH assuming local absorption is ≃ (0.11 ± Figure5. Data-to-modelratiointherest-frameforRXJ1028.6- 0.01)×1022cm−2 measured from the XMM-Newton obser- 0844 (stars), where the model is the best-fit power-law with vations (joint MOS from the two observations). This value Galactic absorption in the restricted 1–10keV spectral range. is a factor of 2–3 times smaller than the value measured Only the PN spectrum (May 2003 observation) is plotted, from ASCA (Yuan et al. 2000). For intrinsic absorption at which is rebinned for demonstration. Also plotted are those for z = 4.286, the NH inferred by XMM-Newton (a few times GB1428+4217 (filled squares) and PMNJ0525-3343 (open cir- 1022cm−2 ) becomes about 10 times smaller than that ob- cles)forcomparison,whichweretaken fromFig.5inWorsleyet tainedbyASCA.Intrinsicvariabilityintheabsorptioncan- al.(2004b).Thespectraareplottedintheirrespectiverest-frame energies. not be ruled out. However, we speculate that a systematic differenceintheinstrumentalcalibrationofthetwomissions might play at least a partial role. This is because a similar trend was also found for GB1428+4217 and PMNJ0525- 3343 (Worsley et al. 2004a,b). It is not clear which instru- objects are in good agreement. The striking spectral simi- ment causes the difference. In the cases of GB1428+4217 larity shared by these objects at different redshifts argues and PMNJ0525-3343, the previous results before XMM- for a real soft X-ray flattening against instrumental effects, Newton were obtained by a joint fit of both the ASCA and and suggests a common nature tothis phenomenon. BeppoSAX spectra (Fabian 2001a,b). This fact, together Another similarity lies in their optical–UV properties, withtheaforementioned XMMEPICcalibration issue,sug- which argue for a highly ionised, dust-free absorber model geststhatperhapsbothmissions,ratherthanmerelyXMM- (Yuan et al. 2000; Fabian et al. 2001a,b; Worsley et al. Newton, may be the cause. Improved XMM EPIC calibra- 2004a,b). It is interesting to note that excess absorption in tion and independent investigations by other instruments severalz >4,moderately radio-loud quasars,as tentatively (Chandra orXMM-Newton RGS)areneededtoresolvethis suggestedbytheircombinedChandraspectra,hasalsosimi- problem. larNH valuesofafewtimes1022cm−2 (Bassettetal.2004). Comparing the observations of XMM-Newton in 2003 and of ASCA in 1999, the spectral photon indices are con- sistent within their mutual 1-σ errors; no significant flux variability (>10percent)is detectedinthe1–10keVband. 4.2 X-ray spectral variability 4.1.3 Comparison with other objects Itisworthstressingthatwhilespectralvariabilityisindeed typicalofflat-spectrumquasars,theextremelyflatspectrum We compare the spectral shape of RXJ1028.6-0844 with in the2–10keVbandduringthe2002 observation is consis- thoseofGB1428+4217andPMNJ0525-3343inthequasar’s tentonlywithinthe2σ uncertaintyrangewiththelimiting rest-frame. Following Worsley et al. (2004b), we produced value of Γ ≃ 1.5 for a relativistic distribution of particles the data-to-model ratio for RXJ1028.6-0844 where the emitting via synchrotron and inverse Compton (in the sim- model is the best-fit power-law with Galactic absorption in plest hypothesis).Iftheproduction ofsuchflat high-energy therestricted1–10keVband.TheresultisplottedinFig.5, spectraareconfirmed,revision ofthewidelyacceptedemis- togetherwiththoseforGB1428+4217andPMNJ0525-3343 sion scenarios would be required. fromWorsleyetal.(2004b,theirFigs.2and5).Itshouldbe Although fluxvariability over short timescales is a dis- noted that the data-to-model ratio is free from the effects tinctivecharacteristicofblazaremission,nosignificantvari- of instrument response, Galactic absorption, and redshift. ations have been detected within a single observation. It It can be seen that thethree spectra agree remarkably well should be noted, however, that due to the high redshift of in terms of the break energy and the shape of the spectral the source, the intrinsic timescale sampled by the observa- ctiumtoeffs.10T2h2ecmN−H2avsalmueesasoufriendtrbiynsXicM(Mco-ldN)ewabtosnoribnetrhseosfeathfreewe tion (∼2 hr)might betoo short todetect significant varia- tions(for powerful quasars doublingtimescales of theorder ∼several hour to a day might be more typical, e.g. 3C279, 5 http://xmm.vilspa.esa.es/es/external/xmm sw cal/calib/index.shtmlWherle et al. 1998). 8 W. Yuan, et al. 4.3 Excess emission around 5–10keV? upon those obtained from a previous short-exposure obser- vation for RXJ1028.6-0844 with XMM-Newton (Grupe et Worsley et al. (2004b) pointed out possible excess emission al. 2004). A comparative study of the two XMM-Newton at energies around 5–10keV in the quasar rest-frame spec- observations revealed a spectral steepening from Γ ≃1.3 in trum for GB1428+4217. The evidence is only marginal. 2002toΓ≃1.5in2003,andaconsequentdropinfluxinthe Interestingly, the same spectral structure also appears in hard energy band above∼1keV. RXJ1028.6-0844, as can be seen in Fig.5. The similarity The derived columns from XMM-Newton observations, of the energy position of this feature in two objects at dif- however, are reduced when compared with the previous ferent redshifts is remarkable. If this feature is real, it may ASCA results (Yuan et al. 2000). We speculate that this comefromanadditionalspectralcomponentwhichispeaked might be due to systematic instrumental effects, probably around5–10keVintherest-frame. Wemodelled theXMM- inherent in both missions. Future improved XMM-Newton NewtonspectratakeninMay2003byaddingasteeppower- EPIC calibration and independent investigations by other lawcomponentintheabovespectralmodels,followingWors- instruments (such as Chandra) are needed to resolve this ley et al. (2004b). Both the PN and MOS1+2 spectra were issue. fitted jointly to improve the statistics. A steep photon in- A tentative excess emission feature in the rest-frame dex is yielded for the second power-law, as Γ=2.3–4.4 for 5–10keV band is suggested, which bares remarkable simi- a model with local absorption and Γ=1.9–7.2 for intrinsic larity to that marginally imprinted in the X-ray spectrum absorption at z=4.276 (90 per cent confidence range for of GB1428+4217 (Worsley et al. 2004b). 1 interesting parameter). These values are consistent with that obtained for GB1428+4217, Γ ∼1.8–2.6 (Worsley et al. 2004b). The model does improve the fit in the 5–10keV (rest-frame) band, though the statistical significance is not ACKNOWLEDGEMENTS high (∆χ2≃−7 for 3 additional free parameters). Ifthisexcessemissionfeatureprovestobereal,itmight We thank Richard Saxton of the XMM-Newton calibra- bethefirstevidenceforthepresenceofemissionoriginating tion team for useful advice on the EPIC calibration issues. from bulk comptonisation on the soft photon field through Matt Worsley is thanked for help in making the plot of which the relativistic jet propagates (Begelman & Sikora Fig.5.W.Y.thanksFranzBauerforcommentsandacareful 1987). Thedetectionofsuchafeaturecouldcarrykeyclues reading of the manuscript. ACF thanks the Royal Society to the amount of (cold) leptons flowing in the jet (Sikora for their support. AC acknowledges the MIUR and INAF & Madejski 2000). Thelack of its detection in themajority for financial support. This research has made use of the of objects so far remains a puzzle. However, in most cases NASA/IPACExtragalactic Database (NED)which isoper- theX-rayemission mightbedominatedbythenon-thermal ated by the Jet Propulsion Laboratory, California Institute emission from relativistic particles, and thusthe possibility ofTechnology,undercontractwiththeNationalAeronautics of detecting such a component could be limited to cases of and Space Administration. particularly high jet Lorentz factors (which would shift its peakuptohighenergies)and/orlow-states/steeppower-law of the non-thermal relativistic component (the latter case could of course be tested, in principle). Observations with REFERENCES evenhighersignal-to-noisethanthepresentonesorstacking Bassett L.C., Brandt W.N., Schneider D.P., Vignali C., spectra from different sources might bea way toclarify the Chartas G., Garmire G.P., AJ, 128, 523 issue. Begelman M.C. & SikoraM., 1987, ApJ,322, 650 Bevington P.R. & Robinson D.K., 1992, Data reduction and error analysis for the Physical sciences, 2nd edition, 5 CONCLUSIONS McGraw-Hill, Inc. We have presented a new X-ray spectroscopic study of the Boller Th., Fabian A.C., Brandt W.N., & Freyberg M.J., highredshift(z=4.276)quasarRXJ1028.6-0844withXMM- 2000, MNRAS,315, L23 Newton. The high signal-to-noise spectrum confirms the Cappi M., Matsuoka M., Comastri A., et al., 1997, ApJ, presenceofthesoft X-rayspectralflattening,whichwas re- 478, 492 portedpreviouslywithASCA data(Yuanetal.2000). This Dickey J.M. & Lockman F.J., 1990, ARA&A28, 215 spectral feature can be modelled by either excess absorp- Done C., Mulchaey J.S., Mushotzky R.F., Arnaud K.A., tion of the quasar X-rays or an intrinsic break at ∼1keV 1992, ApJ, 395, 275 in the X-ray spectra of the source. In the absorption sce- Elvis M., Fiore F., Wilkes B., McDowell J., Bechtold J., nario,thederivedcolumndensityforcoldabsorberintrinsic 1994, ApJ, 422, 60 to the quasar is 2.1(+0.4)×1022cm−2. This valueis compa- Elvis M., Fiore F., Giommi P., Padovani P., 1998, ApJ, −0.3 rabletothosereportedintwosimilarobjectsGB1428+4217 492, 91 (z=4.72) and PMNJ0525-3343 (z=4.4) from XMM-Newton Fabian A.C., Brandt W.N., McMahon R.G., Hook I.M., observations. The remarkable similarity in the shape of the 1997, MNRAS,291, L5 spectralcutoffamongtheseobjectsatdifferentredshiftsar- FabianA.C.,IwasawaK.,CelottiA.,etal.,1998,MNRAS, gues against instrumental effects as an origin, but rather 295, L25 argues for a common nature to the soft X-ray flattening in FabianA.C.,CelottiA.,IwasawaK.,McMahonR.G.,Car- high-z blazars. In terms of the soft X-ray spectral flatten- illi C.L., Brandt W.N., Ghisellini G., Hook I.M., 2001a, ing,theresultsareconsistentstatisticallywithandimproved MNRAS,323, 373 XMM observations of the high-redshift quasar RXJ1028.6-0844 9 Fabian,A.C.,Celotti,A.,IwasawaK.,GhiselliniG.,2001b, Table A1. absorbed power-law model fit to the PN spectrum MNRAS,324, 628 usingcorrectedeffective areaARFcalibration Ferrero E. & Brinkmann W., 2003, A&A,402, 465 F4io9r2e, 7F9., Elvis M., Giommi P., Padovani P., 1998, ApJ, absorber NHa Γ χ2/dof Pnull FioreF.,ElvisM.,MaiolinoR.,NicastroF.,Siemiginowska observationinJune2003 A., StrattaG., D’Elia V., 2003, A&A,409, 57 local 11.1±0.8 1.54±0.03 137/147 0.70 Grupe D., Mathur S., Wilkes B., Elvis M., 2004, AJ, 127, local 4.59(fix) 1.29±0.02 218/148 1.410−4 z=4.276 231±28 1.50±0.03 134/147 0.77 1 Hook I.M. & McMahon R.G., 1998, MNRAS,294, L7 a Columndensityofhydrogeninunitsof1020cm−2 Kirsch M., Altieri B., Chen B., Haberl F., Metcalfe L., et al. 2004, SPIEin press, (astro-ph/0407257) 10–15 per cent with respect toMOS and by 20–30 per cent KirschM.,2004,‘EPICstatusofcalibrationanddataanal- with respect to RGS (Kirsh et al. 2004). It appears that ysis’, 2004, XMM-Newton SOC, XMM-SOC-CAL-TN- the problem arises mainly from the PN calibration at low 0018 (July,28, 2004) energies, although it is not clear whetherthe redistribution LuL.,SargentW.L.W.,BarlowT.A.,ChurchillC.W.,Vogt matrix or the effective area is deficient. This means that S., 1996, ApJS,107, 475 PN spectra tend to give less absorption column than MOS Moran E.C. & Helfand D.J., 1997, ApJL, 484, 95 and RGS,a trend similar to what we findin thiswork (Ta- Otrupcek R.E. & Wright A.E., 1991, PASA,9, 170 ble2 and Table3), as well as those obtained by Grupe et ProuxC.,Storrie-LombardiL.J.,McMahonR.G.,IrwinM. Hook I.M., 2001, AJ, 121, 1799 al. (2004). We thus argue that the NH values derived from MOS spectra are more reliable. Pettini, M., Smith, L. J., King, D. L., Hunstead, R. W., Iftheeffectispurelyduetoaneffectiveareaproblemof 1997, ApJ, 486, 665 PN,wecantestthePN–MOSconsistencybytakingintoac- Prochaska J.X. and Wolfe A.W., 2000, ApJL, 533, 5 countthereporteddifferenceinthecalibration. Tosimplify Protassov R., van Dyk D., Connors A., Kashyap V.L., the treatment, we assumed that the PN effective area be- Siemiginowska A., 2002, ApJ, 571, 545 low1keVwasunder-estimatedbyafactoroff(E),whichis Reeves, J. N., Turner, M. J. L., Ohashi, T., Kii, T., 1997, energy-dependent.f(E)wasestimatedfromFig.10inKirsh MNRAS,292, 468 etal. (2004), inwhich theMOS/RGSnormalisation factors Sikora M., Madejski G, 2000, ApJ 534, 109 with respect to PN is plotted versus energies. We adopted Wehrle A.E., Pian E., Urry C.M., Maraschi L., McHardy conservative values of f(0.3keV) ∼1.4, f(0.6keV) ∼1.15, I.M., Lawson A.J., Ghisellini G., Hartman R.C., et al., and f(1keV) ∼1.0. f(E) at any other energies E within 1998, ApJ, 497, 178 0.3–1.0keV was interpolated using a binomial function. We Wilkes, B. J., Elvis, M., Fiore, F.,et al., 1992, ApJL,393, corrected the PN effective area (ARF) by multiplying by L1 f(E) at the corresponding energy. The PN spectral fitting WorsleyM.A.,Fabian,A.C.,TurnerA.K.,CelottiA.,Iwa- wasrepeatedwithiscorrectedARFandtheresultsarelisted sawa K., 2004a, MNRAS,350, 207 WorsleyM.A.,Fabian,A.C.,CelottiA.,IwasawaK.,2004b, inTableA1.ThefittedtotalNH is(11.1±0.8)1020cm−2,in excellent agreement with the MOS results, while Γ remains MNRAS,350, L67 Yuan,W.andBrinkmann,W.,1999,inHighlightsinX-ray unchanged. We therefore conclude that the NH fitted from thePNspectrumislikelytobeunder-estimated.Morequan- Astronomy,ed.,Aschenbach,B.,andFreyberg,M.,(MPE Report 272), 240 titative and reliable estimation of NH in PN spectra must awaitthecompletionofthePNandMOScalibration atthe Yuan W., Matsuoka M., Wang T., Ueno S., Kubo H., Mi- low energies (Kirsh et al. 2004). hara T., 2000, ApJ, 545, 625 Zickgraf,F.-J.,Voges,W.,Krautter,J.,etal.,1997,A&A, 323, L21 APPENDIX A: EFFECT OF CALIBRATION UNCERTAINTY ON THE RESULTS In the most updated reports on the XMM-Newton calibra- tion status, Kirsh et al. (2004, see also XMM-Newton SOC XMM-SOC-CAL-TN-0018 6) demonstrated significant dis- crepanciesatlowenergies(below∼1keV)betweentheEPIC PNandMOScameras,andtheEPICandtheRGS7,aswell as overall differences in the XMM-Newton instruments and those of Chandra. While MOS and RGS agree with each other in general, PN gives higher fluxes below 0.7keV by 6 http://xmm.vilspa.esa.es/es/external/xmm sw cal/calib/index.shtml 7 TheReflectionGratingSpectrometeron-boardXMM-Newton.