1 THE LOG N - LOG S AND THE BROADBAND PROPERTIES OF THE SOURCES IN THE HELLAS2XMM SURVEY A.Baldi1, S.Molendi1, A.Comastri2, F.Fiore3, G.Matt4, and C.Vignali5 1Istituto di Fisica Cosmica - CNR,via Bassini 15, I-20133 Milano, Italy 2Osservatorio Astronomico di Bologna, via Ranzani 1, I-40127 Bologna, Italy 3Osservatorio Astronomico di Roma, via Frascati 33, I-00040 Monteporzio, Italy 4Dipartimento diFisica - Universit´adi Roma Tre, via della Vasca Navale 84, I-00146 Roma, Italy 5Dept.of Astronomy and Astrophysics - The PennsylvaniaState University,525 Davey Lab, UniversityPark, PA 16802 USA 2 0 Abstract presence of a significantpopulation of type-2 QSOs (Nor- 0 man et al. 2001),not yet detected in sufficient quantities. We present the first results from an XMM-Newton 2 Type-2 QSOs are rare (so far, only a few are known), lu- serendipitous medium-deep survey, which covers nearly n minous and hard (heavily absorbed in the soft band). A three square degrees. We show the log N - log S distri- a good way of finding them is to perform surveys in the J butions for the 0.5-2,2-10and 5-10 keV bands, which are 1 found to be in good agreement with previous determina- hard X-ray bands, covering large solid angles. The large throughput and effective area, particularly in the harder 3 tionsandwiththe predictionsofXRB models.Inthe soft bandwe detect a breakat fluxes around5 10−15 cgs.In bands, make XMM-Newton currently the best satellite to 1 × perform hard X-ray surveys. the harderbands, we fill in the gapat intermediate fluxes v In this poster contribution we present results from the between deeper Chandra and XMM-Newton observations 5 HELLAS2XMMsurvey(Baldietal.2002),oneofitsmain 2 and shallower BeppoSAX and ASCA surveys. Moreover, goalsis toconstrainthe contributionofabsorbedAGNto 5 wepresentananalysisofthebroad-bandpropertiesofthe 1 sources. the XRB. 0 2 Key words:galaxies:active — X-rays:diffuse background 2. Data preparation and cleaning 0 / — X-rays: galaxies h We use the XMM-SASanalysissoftwaretasksepproc and p emproctolinearizetheeventfiles.Beforeprocessing,datasets - are corrected for the attitude of the satellite in order to o r have absolute positions in the sky. The Attitude History st 1. Introduction File(AHF) coordinates are given to the SAS task odffix a which performs the correction. : Whileinthesoftband(0.5-2keV)ROSAT(Hasingeretal. v Theeventfilesproducedbyepprocandemprocarecleaned 1998) and especially Chandra (Rosati et al. 2001) has re- i from: X solved almost all the XRB, in the hard band (2-10 keV) – hot pixels; with a procedure (developed at IFC/CNR- r the XRB has been resolvedat a 25%-30%levelwith Bep- a Milan by A. De Luca) which uses cosmic ray IRAF poSAX and ASCA surveys (Cagnoni et al. 1998; Giommi tasksto localize the pixels to be rejected in eachCCD et al. 2000) and recently at a 90% with Chandra (Rosati and XMM-SAS task evselect for removing them from et al. 2001). Moreover, in the very hard band (5-10 keV) the event files; the fraction resolved by BeppoSAX is around 30% (Fiore – softprotonflares;analysingthelightcurvesatenergies et al. 1998) and recently in the XMM-Newton Lockman greater than 10keV and setting a threshold for good Holedeeppointinga60%isreached(Hasingeretal.2001). time intervals of 0.15 cts/s for MOS units and of 0.35 The optical counterparts of the objects making the XRB cts/s for pn unit. are predominantly Active Galactic Nuclei (AGN). In the softbandthepredominantfractionismadebyunabsorbed AcompletesetofMOS1,MOS2andpn600x600pixelim- AGN, with a small fraction of absorbed AGN (Schmidt ages(1 pixel= 4.35arcsec)is generatedusingXMM-SAS et al. 1998).The fractionof absorbed type-2 AGN rises if task evselect in the 0.5-2, 2-10, 2-4.5, 4.5-10 and 0.5-10 weconsiderthespectroscopicidentificationsofhardX-ray keV bands. MOS and pn images are merged together in sourcesin BeppoSAX, ASCA and Chandra surveys(Fiore order to increase the signal-to-noise ratio of the sources et al. 2001; Della Ceca et al. 2000;Tozzi et al. 2001). and go deeper. The X-ray and optical observations are consistent with A corresponding set of exposure maps is generated to ac- currentXRBsynthesismodels(Comastrietal.1995;Gilli countforspatialquantumefficiency,mirrorvignettingand et al. 2001), which explain the hard XRB spectrum with field of view of each instrument, running XMM-SAS task anappropriatemixtureofabsorbedandunabsorbedAGN, eexpmap. The so-created exposure maps are not com- byintroducingthecorrespondingluminosityfunctionand pletely satisfactory, since the evaluation of quantum ef- cosmologicalevolution.However,thesemodelsrequirethe ficiency, filter transmission, and vignetting is performed Proc. Symposium ‘New Visions of the X-ray Universe in the XMM-Newton and Chandra Era’ 26–30 November2001, ESTEC, The Netherlands 2 assuming an eventenergy which correspondsto the mean Moreover we compute p, the poissonian probability that of the energy boundaries. In the 2-10 keV band, this may counts originate from a backgroundfluctuation, from the lead to inaccuracies in the estimate of these key quanti- ties, thus we create the 2-10 keV band exposure map as X∞ e−ctsbkgctsnbkg >p a weighted mean of the 2-4.5 keV and the 4.5-10 keV ex- n! posure maps (assuming an underlying power-law spectral n=ctssrc model with Γ=1.7). and choose a threshold of p=2 10−4 to decide whether XMM-SAStaskesplinemapcreatesabackgroundmapby × to accept or not a detected source. 4. The survey The HELLAS2XMM survey (Baldi et al. 2002) currently uses the 15 XMM-Newton calibration and performance verification phase fields shown in table 1. All the fields areathighgalacticlatitude(bII >27o),havelowgalac- tic N (a few 1020 cm−2) a|nd a|t least 15 ksec of good H observing time. The sky coverage of the sample has been Table 1. The HELLAS2XMM survey sample. Figure1. Left: XMM-SAS standard background map. Target NH (cm−2) bII(o) Right: corrected background map. PKS0537-286 2.1·1020 -27.3 PKS0312-770 8·1020 -37.6 removingallthesourcesaboveafixedmaximumlikelihood MS0737.9+7441 3.5·1020 29.6 threshold and fitting the remaining with a cubic spline. Lockman Hole 5.6·1019 53.1 Even using the maximum number of spline nodes (20), Mkn 205 3·1020 41.7 thefitisnotenoughflexible totakeintoaccountthelocal BPM 16274 3.2·1020 -65.0 MS1229.2+6430 2·1020 52.8 variations of the background. We correct the background PKS0558-504 4.5·1020 -28.6 map produced by XMM-SAS, pixel by pixel, comparing Mkn 421 7·1019 65.0 the mean value in it with that of the so-called cheesed Abell 2690 1.9·1020 -78.4 image, within a radius of 3·r0.68 (r0.68 = radius corre- G158-100 2.5·1020 -74.5 sponding to an encircled energy fraction (EEF) of 0.68). GD153 2.4·1020 84.7 A comparison between XMM-SAS standard background IRAS13349+2438 1.2·1020 60.6 map and our corrected map is shown in Figure 1. Abell 1835 2.3·1020 60.6 Mkn 509 4.1·1020 -29.9 3. Source detection and characterization Apreliminarydetectionrun,usingXMM-SASeboxdetect, computed using the exposure maps of each instrument, is performed ineachenergy band,in orderto createa list the background map of the merged image and a model of candidate sources.Eachsource is characterizedusing a for the PSF. We adopt the off-axis angle dependent PSF radius corresponding to an EEF of α=0.68. model implemented in XMM-SAS eboxdetect task. The sourcecounts S andtheir errorσS aredeterminedas At each image pixel (x,y) we evaluate, within a radius S = ctssrc−ctsbkg rm0a.6p8),.tFhreomtottahlebseacwkegrcoaulcnudlactoeutnhtes m(frinomimuthmetboatcaklgcroouunntds α (source + background) necessary for a source to be de- 1+√cts +0.75 tectedataprobabilityp=2 10−4(definedinSection3). σS = αsrc The mean exposure times for×MOS1, MOS2 and pn, eval- Thecountrateiscr = S whereT isthesumofMOS1, uated from the exposure maps within r0.68, are used to Ttot tot compute the count rate cr. From the count rate-to-flux MOS2 and pn exposure times. The corresponding flux is conversion factor cf (computed as in Section 3) we build F =cf cr where cf is calculated from the x · a flux limit map and straightforwardly calculate the sky Ttot TMOS1 TMOS2 Tpn coverage of a single field. = + + cf cfMOS1 cfMOS2 cfpn Summing the contribution from all fields we obtain the 3 total sky coverage of the survey, which is plotted in Fig- ure 2, in three different energy bands. 1 0.1 0.01 0.001 Figure2. The total sky coverage of the survey in the 0.5-2 keV (red), 2-10 keV (green) and 4.5-10 keV band (blue). Figure3. The cumulative log N - log S in the 0.5-2 keV (top), 2-10 keV (center) and 5-10 keV band (bottom). In all diagrams the black thick solid lines are the upper and lower limits of our Log(N)-Log(S). The dashed cyan lines are the predictions of the XRB synthesis models from Co- 5. The log N - log S Relation mastri et al.(2001). The log N - log S distributions are plotted in Figure 3 and contain 1022,495 and 100 sources,for the 0.5-2 keV, hardness ratio: 2-10 keV and 5-10 keV band, respectively. aIrnouthned05.5-210k−eV15beargndcmth−e2dsi−st1r,ibsiumtiiolanrlsyhtoowRsaOSflAatTtednaintag HR1 = cr2−4.5−cr0.5−2 × cr2−4.5+cr0.5−2 (Hasinger et al. 1998) although with a flatter differential slope index at faint fluxes. We are also in good agree- As shown in Figure 4 and in Table 2, the faint sample mentwithChandraDeepFieldSouthdata(Giacconietal. 2001).Inthe2-10keVbandwefindthatthedistributionis significantly sub-euclidean, in contrast to BeppoSAX and Table2.HR1distributionforthefaintandthebrightsam- ASCA findings (Giommi et al. 2000; Cagnoni et al. 1998; ple. Ueda et al. 1999), indicating that the log N - log S flat- tensatfainterfluxes.Anyhowitrepresentsalinkbetween N(HR1<−0.35) N(HR1≥−0.35) Chandra (Giacconietal.2001)andBeppoSAX-ASCAob- servation, sampling an intermediate flux range. The 5-10 Faint sample 654 84 keVlogN -logS isconsistentwithaneuclideanslopeand Bright sample 107 0 samplesanintermediatefluxrangebetweenXMM-Newton deeperobservations(Hasingeretal.2001)andBeppoSAX shallower HELLAS survey (Fiore et al. 1998). shows a tail of hard sources which is not present in the bright sample. The probability of having 84 sources with 6. Hardness Ratio Analysis HR1 0.35 in the faint sample and no sources with We divided the sample of sources detected both in 0.5- HR1 ≥ −0.35 in the bright sample is 10−6, so the ≥ − ∼ 2 keV band and in 2-4.5 keV band in two subsamples progressivehardeningofthesourcestowardsfainterfluxes containing the brighter (F0.5−2keV >2 10−14 erg cm−2 seems to be highly significant. s−1) and the fainter (F0.5−2keV 2 ×10−14 erg cm−2 A further analysis of the hardness ratio has been carried s−1) sources, respectively and we≤com×pute for them the out on the 4.5-10 keV sample. In Figure 5 the relation 4 cm−2 s−1. Thepopulationofhardersourceswedetect,probablycon- 80 sists of AGN having substantial absorbing column densi- ties (N >1022 cm−2). H 60 7. Summary WearecarryingoutaserendipitousXMM-Newtonsurvey. Wecurrentlycovernearlythreesquaredegreesin15fields 40 observedduringsatellitecalibrationandperformancever- ification phase.This is, to date,the XMM-Newton survey with the largest solid angle. 20 The main results can be summarized as follows: – The log N-log S relations in the 0.5-2 keV, 2-10 keV and 5-10 keV band are in agreement with previous 0 -1 -0.5 0 0.5 determinations; – inthehardbandswesampleanintermediatefluxrange: deeperthanASCAandBeppoSAXandshallowerthan Figure4. The distribution of hardness ratio HR1. Void Chandra and XMM-Newton deep surveys hse−risg1t.ocgSmrha−am2de:sd−soh1u.isrtcoegsrawmit:hsoFu0r.5c−es2kweVith≤F02.5×−21k0eV−1>4 e2r×g1c0m−−124 – kW0.e5eV-2fiflnkudexVeasnflouefvx1ied0s−eon1f3ce2erf×ogrc1hm0a−−r1d24sse−oru1g.rccmes−e2mse−rg1inagndbe2l-o1w0 Acknowledgements WethankA.DeLucafordevelopingthehotpixelcleaningal- 1 gorithm. We are also grateful to G. Zamorani, G. C. Perola and all members of the HELLAS2XMM team for useful dis- cussions.ABandSMacknowledgepartialfinancialsupportby 0.5 ASI I/R/190/00 contract. References 0 Baldi A., Molendi S., Comastri A., Fiore F., Matt G., Vignali C. 2002, ApJ 564, 190 Cagnoni I., Della Ceca R., Maccacaro, T. 1998, ApJ 493, 54 -0.5 Comastri A., Setti G., Zamorani G., Hasinger G. 1995, A&A 296, 1 Comastri A., Fiore F., Vignali C., Matt G., Perola G.C., La -1 Franca F. 2001, MNRAS327, 781 DellaCeca R.,Maccacaro T.,RosatiP.,Braito V.2000, A&A 355, 121 FioreF.,LaFrancaF.,GiommiP.,ElvisM.,MattG.,Comastri Figure5.2-10keVfluxvs.HR2 forthesourcesdetectedin A., Molendi S., Gioia I. 1999, MNRAS 306, L55 the 4.5-10 keV band. The dashed lines indicate the hard- FioreF.,ComastriA.,LaFrancaF.,VignaliC.,MattG.,Per- ola G.C. 2001 proceedings of the ESO/ECF/STSCI work- ness ratios of the sources according to a power-law model shop on ”Deep Fields”, astro-ph/0102041 (galactic absorption) with spectral index α . E Giacconi R., Rosati P., Tozzi P. et al. 2001, ApJ551, 624 Gilli R., Salvati M., Hasinger G. 2001, A&A366, 407 between the 2-10 keV flux and the hardness ratio Giommi P., Perri M., Fiore, F. 2000, A&A362, 799 Hasinger G., Burg R., Giacconi R., Schmidt M., Tru¨mper J., HR2 = cr4.5−10−cr2−4.5 Zamorani, G. 1998, A&A 329, 482 cr4.5−10+cr2−4.5 HasingerG., AltieriB.,ArnaudM.etal. 2001, A&A365, L45 for sources detected in the 4.5-10 keV band is plotted. NormanC.,HasingerG.,GiacconiR.etal.2001,ApJaccepted, astro-ph/0103198 Thediagramshowsaprogressivehardeningofthesources Rosati P., Tozzi P., Giacconi R. et al. 2001, ApJ accepted, whengoingtowardsfainterfluxes.Whilethelackofsources astro-ph/0110452 in the bottom left of the figure is due to a selection effect Schmidt M., Hasinger G., GunnJ. et al. 1998, A&A329, 495 (because we are more sensitive in the 2-4.5 keV than in Tozzi P., Rosati P., Nonino, M. et al. 2001, ApJ562, 42 the4.5-10keVband),thereisnoapparentreasonforhav- Ueda Y.,Takahashi T., InoueH. et al. 1999, ApJ518, 656 ing nohardsourcesatfluxes greaterthanafew 10−14 erg