9512.net
甜梦文库
当前位置:首页 >> >>

XMM-Newton 13H Deep field

Mon. Not. R. Astron. Soc. 000, 000–000 (0000)

Printed 5 February 2008

A (MN L TEX style ?le v2.2)

XMM-Newton 13H Deep ?eld - I. X-ray sources
N. S. Loaring1 ? , T. Dwelly1 , M. J. Page1 , K. Mason1, I. McHardy2 , K. Gunn2, D. Moss2, N. Seymour3, A. M. Newsam4 , T. Takata5 , K. Sekguchi5 , T. Sasseen6, 7 F. Cordova 1
MSSL, University College London, Dorking, Surrey, RH5 6NT, UK of Physics and Astronomy, University of Southampton, Southampton, SO17 1BJ, UK 3 Institut d’Astrophysique de Paris, 98bis, Boulevard Arago, 75014 Paris, France 4 Astrophysics Research Institute, Liverpool John Moores University, Twelve Quays House, Egerton Wharf, Birkenhead, CH41 1LD, UK 5 National Astronomical Observatory of Japan, 650 North A’ohoku Place, Hilo, HI 96729, USA 6 Department of Physics, University of California, Santa Barbara, CA 93106, USA 7 University of California, Riverside, 900 University Avenue, Riverside, CA 92521, USA
2 School

arXiv:astro-ph/0507408v1 18 Jul 2005

5 February 2008

ABSTRACT

We present the results of a deep X-ray survey conducted with XMM-Newton, centred on the UK ROSAT 13H deep ?eld area. This region covers 0.18 deg2 and is the ?rst of two areas covered with XMM-Newton as part of an extensive multi-wavelength survey designed to study the nature and evolution of the faint X-ray source population. We have produced detailed MonteCarlo simulations to obtain a quantitative characterisation of the source detection procedure and to assess the reliability of the resultant sourcelist. We use the simulations to establish a likelihood threshold above which we expect less than 7 (3%) of our sources to be spurious. We present the ?nal catalogue of 225 sources. Within the central 9′ , 68 per cent of source positions are accurate to 2′′ making optical follow-up relatively straightforward. We construct the N (> S ) relation in four energy bands: 0.2-0.5 keV, 0.5-2 keV, 2-5 keV and 5-10 keV. In all but our highest energy band we ?nd that the source counts can be represented by a double powerlaw with a bright end slope consistent with the Euclidean case and a break around 10?14 erg cm?2 s?1. Below this ?ux the counts exhibit a ?attening. Our source counts reach densities of 700, 1300, 900 and 300 deg?2 at ?uxes of 4.1 × 10?16, 4.5 × 10?16, 1.1 × 10?15 and 5.3 × 10?15 erg cm?2 s?1 in the 0.2-0.5, 0.5-2, 2-5 and 5-10 keV energy bands respectively. We have compared our source counts with those in the two Chandra deep ?elds and Lockman hole and ?nd our source counts to be amongst the highest of these ?elds in all energy bands. We resolve > 51% (> 50%) of the X-ray background emission in the 1-2 keV (2-5 keV) energy bands. Key words: surveys - X-ray: selection - background - AGN - cosmology

1 INTRODUCTION It is widely accepted that the majority of the Cosmic X-ray Background (XRB) arises from the integrated emission of discrete extragalactic sources (Schwartz et al. 1976; Giacconi & Zamorani 1987; Maccacaro et al. 1991). The energy density of the XRB peaks at ?30 keV, but the ?rst imaging surveys were carried out at much lower energies: <3.5 keV with Einstein and <2 keV with ROSAT . By the late 1990s ROSAT surveys had resolved 70-80% of the soft XRB, (Shanks et al. 1991; Hasinger et al. 1993, 1998; McHardy et al. 1998). Subsequently, deep XMMNewton and Chandra surveys have essentially resolved the soft XRB into discrete sources (Mushotzky et al. 2000; Hasinger et al.
?

2001; Brandt et al. 2001; Tozzi et al. 2001; Rosati et al. 2002; Alexander et al. 2003). Optical follow up of these sources has revealed a population composed primarily of unobscured broad line active galactic nuclei (AGN), with an increasing fraction of absorbed AGN at fainter ?uxes (McHardy et al. 1998; Schmidt et al. 1998; Zamorani et al. 1999; Lehmann et al. 2001; Szokoly et al. 2003; Barger et al. 2003). In order to investigate further the nature of the obscured population one has to conduct surveys at harder energies (>2 keV), which are less sensitive to absorption. Surveys carried out using ASCA (Georgantopoulos et al. 1997; Cagnoni, Della Ceca & Maccacaro 1998; Ueda et al. 1998, 2001; Ishisaki et al. 2001) and BeppoSAX (Fiore et al. 1999; Giommi et al. 2000; Fiore et al. 2001) resolved 25-35% of the XRB above 2 keV. More recently, the XMM-Newton and Chan-

nsl@mssl.ucl.ac.uk

c 0000 RAS

2

N. S. Loaring et al.
Imaging Cameras (EPICs Turner et al. 2001) were operated in standard full-frame mode. The thin ?lter was used for the PN camera, while the thin and medium ?lters were alternated for the MOS1 and MOS2 cameras. Table 1 gives a summary of the observations. The data were processed using the XMM-Newton Standard Analysis System (SAS) version 6.0. Approximately 40% of the total observation time was a?ected by high particle background ?ares, arising from soft protons hitting the detector. The data were therefore temporally ?ltered to remove these high background periods. In practice, times where the 5-10 keV count rates exceeded 2 s?1 in the MOS cameras and 4 s?1 in the PN camera were excluded. Filtering reduced the total useful exposure time from ?200 ks to ?120 ks. The net live times for the individual detectors after the periods of high background were excised are listed in Table 1. A signi?cant component of the EPIC background comes from instrumental emission lines, in particular the Cu Kα line at 8.1 keV in the PN and the Al Kα line at 1.5 keV in both detectors (Lumb et al. 2002). Events with energies close to those of the emission lines were ?ltered out to minimise the instrumental contribution to the background. Events in bad columns, bad pixels and adjacent to chip edges were also excluded. Images and exposure maps were then constructed from each observation for each detector in four energy bands: 0.2-0.5 keV, 0.5-2 keV, 2-5 keV and 5-10 keV. Single-pixel events were used to construct the PN 0.2-0.5 keV image. Single, double and triple events were used to construct the higher energy PN images. For MOS, all valid event patterns were used to construct the images regardless of energy band. In each energy band, the exposure maps were scaled to the PN thin ?lter response. The images and exposure maps from the di?erent detectors and observations were then summed to produce one image and one exposure map per energy band. The response-weighted summation over each observation and telescope gives total on-axis PN-equivalent live exposure times of 152 ks, 161 ks, 179 ks, and 160 ks, in the 0.2-0.5 keV, 0.5-2 keV, 2-5 keV, and 5-10 keV bands respectively. For the EPIC imaging observing modes, photons are not only registered during the actual integration interval but also during the readout of the CCD. These out-of-time events are hence assigned the wrong position in the readout direction. The fraction of outof-time events is highest for the PN full frame mode (6.3 %) and therefore for each PN exposure, an additional synthetic out-of-time events list was produced by randomising the coordinates of the events within each chip in the readout direction. Out-of-time images were constructed in each energy band by ?ltering these event lists in exactly the same way as the real event lists. These out-oftime images were used as inputs to the background model as described in Section 2.2. The astrometry of the individual observations was corrected for small o?sets between the pointings. A sourcelist was constructed for each observation as described in Section 2.2 and crosscorrelated with the optical positions of the 214 sources found in the Chandra catalogue of McHardy et al. (2003) using the SAS task EPOSCOR. The appropriate o?sets in RA and dec were then applied to each of the individual events to tie the XMM-Newton data to the optical/Chandra /radio co-ordinate frame. The images and exposure maps were then reproduced with the correct astrometry. The actual o?sets in RA and Dec di?ered slightly between the three observations. The ?rst observation (revolution 276, 179 XMM-Newton sources) had o?sets of 1.4′′ , -1.3′′ applied, the second (revolution 281, 106 XMM-Newton sources) had o?sets of 0.5′′ , -0.5′′ applied and the third observation (revolution 282, 257 XMM-Newton sources) had o?sets of 0.6′′ , 0.2′′ applied.
c 0000 RAS, MNRAS 000, 000–000

dra deep ?eld surveys have resolved 60-90% of the hard (>2 keV) X-ray background (Hasinger et al. 2001; Giacconi et al. 2001; Tozzi et al. 2001; Cowie et al. 2002; Rosati et al. 2002; Alexander et al. 2003; Manners et al. 2003) probing ?uxes a factor 100× fainter than the ASCA and BeppoSAX surveys. The wide range in resolved fraction arises not only due to the variation in source counts between the surveys, but also due to the uncertainty of the absolute normalisation of the hard XRB, for example the BeppoSAX XRB normalisation from Vecchi et al. (1999) is ?30% higher than the ASCA value from Gendreau et al. (1995). Optical follow up studies of the deepest surveys ?nd a predominance of z < 1 objects which do not show broad emission lines (Barger et al. 2001; Tozzi et al. 2001; Barger et al. 2002; Lamer et al. 2003; Barger et al. 2003). There is an increasing contribution from normal galaxies at the faintest ?uxes, and it appears likely that normal galaxies will outnumber AGN below 0.52 keV ?uxes of 10?17 erg cm?2 s?1 (Bauer et al. 2004). A significant fraction of the hard X-ray sources in these ?elds are optically faint, with R > 24 and are therefore di?cult to identify optically (Alexander et al. 2001; McHardy et al. 2003; Alexander et al. 2003). Despite the great advances made in detecting increasingly fainter X-ray sources, the physical nature and evolution of the faint hard X-ray population remains largely unknown. The redshift distribution and column density distribution of the absorbed AGN are still poorly constrained. Another issue which needs addressing is the relationship between gas and dust absorption in AGN. There have been several cases of a mismatch between optical and X-ray classi?cations, indicating a wide range in dust/gas ratios for obscured sources (Akiyama et al. 2000; Page et al. 2001; Comastri et al. 2001; Maiolino et al. 2001a; Loaring et al. 2003; Carrera et al. 2004). In particular, high quality X-ray spectra are needed to determine the dominant X-ray emission mechanisms and the amount of absorption. We have therefore used XMM-Newton to carry out deep surveys of two widely separated ?elds to probe the X-ray population down to ?uxes ? 10?15 erg cm?2 s?1 in the 0.5-2 and 2-5 keV energy bands. The source counts in these energy bands exhibit a break at ? 10?14 erg cm?2 s?1 (Rosati et al. 2002) around which the maximum contribution to the XRB per logarithmic ?ux interval occurs. This paper presents the X-ray catalogue derived from the ?rst of these two XMM-Newton surveys, carried out in the UK ROSAT deep ?eld area (hereafter the 13H deep ?eld). The ?eld has also been observed with a mosaic of Chandra pointings which cover the whole XMM-Newton ?eld of view and provide accurate source positions. It is complemented with multiwavelength follow up in the UV, optical, near-IR, mid-IR and radio (Seymour et al. 2004). The Chandra catalogue has already been presented elsewhere (McHardy et al. 2003); here we present the XMM-Newton catalogue and observed source counts.

2 OBSERVATIONS AND DATA REDUCTION 2.1 XMM observations The 13H deep ?eld is centred on the sky co-ordinates RA 13h 34m 37.1s, Dec +37? 53′ 02.2′′ (J2000). The XMM-Newton observations were carried out in three separate revolutions during June 2001 for a total exposure time of 200 ks. The European Photon

XMM-Newton Deep Field: X-ray source catalogue
Rev. Date Live Time (ks) MOS1 MOS2 43.1 14.1 59.2 45.8 12.0 60.2 PN 35.5 6.7 47.9 Filter MOS1 MOS2 thin med thin med thin thin

3

Energy (keV) 0.2–0.5 0.5–2.0 2.0–5.0 5.0–10.0

ECF ( cts per 10?11 erg cm?2 ) 4.7775 4.8905 1.9605 0.5929

276 281 282

12.06.01 22.06.01 23.06.01

Table 1. Summary of 13H deep ?eld XMM-Newton observations showing the date and length of observations and the ?lters used. The live times have had periods of high background excluded.

Table 2. Energy conversion factors (ECF) used to convert between count rate and ?ux.

2.2 XMM-Newton Source detection The combined images in each energy band were source-searched simultaneously using the SAS tasks EBOXDETECT and EMLDETECT. EBOXDETECT is a sliding cell detection algorithm which outputs an initial sourcelist. This sourcelist is input for the EMLDETECT task which performs a maximum likelihood PSF ?t to the sources producing re?ned positions and ?uxes for all bands simultaneously. This method results in better source positions than searching in the individual energy bands one at a time because the PSF ?t is based on the maximum number of counts per source. If the best ?t source count rate in any particular energy band is less than zero (i.e. there are fewer counts at the source position in the image than in the background map) the source count rate is set to zero in this energy band. A background map was produced for each combination of observation, detector and energy band (36 background maps in total) using our own software. This software performs a maximum likelihood ?t to the background, assuming a three-component background model: out-of-time events, a ?at unvignetted component, caused primarily by cosmic rays, and a vignetted component representing unresolved faint sources and genuinely di?use emission. For the PN, the out-of-time events contribution to the background was ?xed at 6.3% of the intensity of the synthetic out-of-time images; for the MOS background, we assumed no contribution from out-of-time events. The intensities of the vignetted and unvignetted background components were free parameters in the ?t. To maximise sensitivity an iterative procedure was employed. Initially, sources were detected in each individual image (per detector and per energy band) using a 3-pixel-square sliding cell in EBOXDETECT, with the background computed as the average of the surrounding 7 × 7 pixels. Then, the sources were excised from the individual images and the background ?tted. Each of the background maps from the MOS and PN cameras for a given energy band were then summed to produce one combined background map for each energy band. The resultant background maps were then used for the sliding cell (EBOXDETECT) followed by maximum likelihood (EMLDETECT) source detection on the combined image in each energy band. The sequence of background determination followed by source searching was then repeated several times. We found that the sourcelist and background maps converged after 4 iterations. Likelihood thresholds (DET ML values output from the source detection) of 4 and 5 were chosen for EBOXDETECT and EMLDETECT respectively. These values are related to the probability of a random Poissonian ?uctuation having caused the observed source counts via (Cash 1979): DET ML = ?lnPrandom (1)

ing a power law spectrum with photon index Γ = 1.7. This is a good average within our ?ux range (Page et al. 2003; Mateos et al. 2005); however the sources have a range of photon indices. To assess the impact of such a spread we have calculated the expected conversion factors using photon indices of Γ = 1.4 and Γ = 2.0 respectively. This range represents the expected spread in spectral slope for sources contributing to the XRB. The relatively ?at Γ = 1.4 lower limit corresponds to the XRB slope in the 3-15 keV range, produced by absorbed AGN. The upper limit value is typically found in unabsorbed AGN, and we would therefore expect the bulk of our sources to lie between the two values. The largest e?ect is in the 0.2-0.5 keV and 5-10 keV energy bands where the conversion factors derived are di?erent from those assuming a photon index of Γ = 1.7 by up to 11% and 8% respectively. However, in the 0.5-2 and 2-5 keV bands the photon index chosen only a?ects the conversion factors by 1-2%. The PN response matrices include only single and double pixel events, but for the three highest energy bands our PN images also include triple and quadruple events. In order to take this into account, the count rate to ?ux conversion factors were corrected (by up to 6% in the hardest band) as described in Osborne et al. (2001).

3 MONTE CARLO FIELD SIMULATION We have used an XMM-Newton speci?c extension of the simulation method of Hasinger et al. (1998) to obtain a quantitative characterisation of the source detection procedure and to assess the reliability of the resultant sourcelist. We have used our simulations to ?nd the appropriate detection threshold to be applied to the 13H ?eld. A Monte-Carlo approach is particularly powerful near the survey ?ux limit where a number of di?erent processes contribute to uncertainties in the detected source parameters. Our simulation method consists of several modular steps that are repeated for a large number of synthetic ?elds. Brie?y, an ‘input’ sourcelist was generated independently in each energy band. Each list was then folded through the XMM-Newton imaging characteristics to generate images in each energy band. These images were then source searched to produce ‘output’ sourcelists. A pairing algorithm was used to associate an ‘input’ source with each ‘output’ source. Each of these stages are described in detail in Appendix A. Here we present the results of a comparison between the output and input source properties. These provide an indication of the biases inherent in our survey. One thousand simulated ?elds were used to reduce statistical uncertainties in the analysis.

3.1 Comparison of simulated input and output sources In order to assess the accuracy of source positions and ?uxes, and to estimate the degree to which confusion and Eddington bias a?ects the source counts in our 13H data we have compared the input and

The conversion factors from count rates to ?ux were determined from the EPIC response matrices, over exactly the same energy ranges as those in which the images were constructed, assumc 0000 RAS, MNRAS 000, 000–000

4

N. S. Loaring et al.

Figure 1. Greyscale images showing the distribution of positional o?sets between output and input source locations as a function of input ?ux in two energy bands. Those sources within the central 9′ of the XMM-Newton ?eld of view having DET ML 5 are shown. The concentration of positional o?sets as a function of S inp is indicated by the darkness of the greyscale image. The three contours show the distances within which 68, 90, and 95 per of the data lie.

output properties of our simulations. The simulations should mimic any biases found in the real data. We matched each output source found in the simulated images to the closest valid input source. We consider an input source as valid when its ?ux, S inp , contributes a reasonable fraction (>20 per cent) to the total output ?ux, S out . No upper limit was applied to the radius at which input and output sources were matched in order to assess the typical o?sets between input and output positions. Fig. 1 shows the distribution of measured positional o?sets as a function of input ?ux, S inp . The greyscale image shows the relative density of sources at a given S inp . The dark band shows where the majority of sources lie. All sources with DETML values > 5 and o?axis angles < 9′ are shown. The contour lines plotted correspond to the positional o?sets within which 68, 90 and 95 per cent of the data lie. The mean positional o?set decreases with increasing ?ux and is < 10′′ for all but the faintest ?uxes. In the 0.5-2 keV energy band, 95 per cent of sources with S inp > 5 × 10?16 erg cm?2 s?1 and o?axis angles 9′ have positional o?sets < 10′′ . In the 5-10 keV energy band, 95 per cent of sources with S inp > 5 × 10?15 erg cm?2 s?1 and o?axis angles 9′ have positional o?sets < 10′′ . Those with o?axis angles > 9′ have systematically larger positional o?sets (larger by ? 2′′ over the majority of ?ux ranges). Any sources with higher positional o?sets are most likely due to incorrect associations. A discussion of the positional accuracy found in the real 13H data and its comparison with the simulations is deferred to Section 4.3. Subsequently, we matched output sources to input sources within a radius rcut = 5′′ , 8′′ , and 10′′ for o?axis angles of 0 ? 9′ , 9 ? 12′ and > 12′ respectively, re?ecting the degradation of the XMM PSF (Kirch 2004) away from the optical axis. Where more than one candidate input source lay within rcut , the brightest was chosen. The brightest input candidate within rcut must be the ‘correct’ input counterpart in the sense that it is the largest contributor to the output source counts. In practice, when averaged over the

1000 simulations, only 0.6, 1.7, 1.5 and 0.5 output sources per ?eld had more than one valid (5S inp S out ) input candidate within rcut in the 0.2-0.5, 0.5-2.0, 2.0-5.0, and 5.0-10 keV energy bands. The output ?uxes and positions were then compared with the corresponding input values. There are several reasons why we might expect a di?erence between the input and output ?ux distributions and source counts: i) a systematic or statistical ?ux measurement inaccuracy, ii) source confusion, iii) statistical ?uctuations in the background which may be detected as sources, and iv) Eddington bias. All these factors must be considered together when interpreting the 13H ?eld data. The intrinsic accuracy of the source detection photometry is best evaluated at high ?uxes and low o?axis angles, where ii), iii) and iv) are less important. The bright end of Fig. 2 illustrates the high ?delity of the detected ?uxes, S out , to the input ?uxes, S inp . Considering sources with S inp > 5 × 10?14 erg cm?2 s?1 , with o?axis angles < 9′ the average S out /S inp ratio is 1.01±0.03 (1σ) in energy bands 0.2-0.5, 0.5-2 and 2-5 keV. In our highest energy band the average S out /S inp ratio is also 1.01 but the scatter is larger (1σ = 0.06). The distribution of S out /S inp is shown in Fig. 3 for three ?ux intervals in each energy band. At bright ?uxes (S inp = 2 × 10?14 erg cm?2 s?1 ) the distribution is narrow, symmetrical and centred on S out /S inp = 1, because the statistical errors on the ?uxes are small. At intermediate ?uxes (S inp = 6 × 10?15 erg cm?2 s?1 ) the distributions are still relatively symmetrical, but they are slightly broader as the statistical errors on the ?uxes are larger. However, at the faintest ?uxes, the distributions are much broader and signi?cantly skewed towards larger S out /S inp ratios. The increased width is due to the increased statistical errors on the ?uxes. The distributions are shifted towards larger S out /S inp ratios because at such faint ?uxes sources are unlikely to be detected unless they are enhanced by Poisson ?uctuations or by source confusion. Source confusion occurs when two or more nearby input sources fall in a single resolution element of the detector and rec 0000 RAS, MNRAS 000, 000–000

XMM-Newton Deep Field: X-ray source catalogue
10.0 10.0

5

0.5-2.0 keV

5.0-10.0 keV

Sout/Sinp

1.0

Sout/Sinp 0.10 1.00 10.00 Sinp (10-14 erg cm-2 s-1) 100.00

1.0

0.1 0.01

0.1 0.1

1.0 10.0 Sinp (10-14 erg cm-2 s-1)

100.0

Figure 2. Scatter plots of input vs output ?ux in two selected energy bands. We only show those detected sources having DET ML 5 and input counterparts within 5′′ , 8′′ , and 10′′ for input o?axis angles of 0 ? 9′ , 9 ? 12′ and > 12′ respectively. Sources with S out /(S inp + 3σout ) > 1.5 are plotted as triangles. The results for 100 simulations are shown for clarity.

0.6 0.5 Fraction 0.4 0.3 0.2 0.1 0.0 0.6 0.5 Fraction 0.4 0.3 0.2 0.1 0.0 0.0 0.5 1.0 1.5 2.0 2.5 3.0 Sout/Sinp 0.5 1.0 1.5 2.0 2.5 3.0 Sout/Sinp 2.0-5.0 keV
2x10-14 cgs 2x10-15 cgs 2x10-16 cgs

0.2-0.5 keV
2x10-14 cgs 2x10-15 cgs 2x10-16 cgs

0.5-2.0 keV
2x10-14 cgs 2x10-15 cgs 2x10-16 cgs

5.0-10.0 keV
2x10-14 cgs 2x10-15 cgs 2x10-16 cgs

Figure 3. The distribution in output (S out ) ?ux to input ?ux (S inp ) ratio for three illustrative input ?ux ranges. The solid line represents the sources in the ?ux interval centred on S inp = 2 × 10?14 erg cm?2 s?1 (1 × 10?14 ? 4 × 10?14 erg cm?2 s?1 ). The dotted line represents sources in the ?ux interval centred on S inp = 6 × 10?15 erg cm?2 s?1 (3 × 10?15 ? 1.2 × 10?14 erg cm?2 s?1 ) and the dashed line represents sources in the ?ux interval centred on S inp = 4 × 10?16 erg cm?2 s?1 (2 × 10?16 ? 8 × 10?16 ). The ?ux intervals were chosen such that ?logS inp = 0.3.

c 0000 RAS, MNRAS 000, 000–000

6

N. S. Loaring et al.

a

b

Figure 4. a) Fraction of ?ux ampli?ed sources as a function of ?ux. b) Fraction of unmatched sources as a function of ?ux. (0.2-0.5 keV (dotted), 0.5-2 keV (solid), 2-5 keV (dashed), 5-10 keV (dot-dashed)). The fraction of ?ux ampli?ed sources is < 0.4 per cent in all energy bands and is signi?cantly less than the fraction of unmatched sources.

sult in a single output source. This results in a ?ux ampli?cation in the output source and a net loss of fainter sources. The position of the output source will be close to the centroid of the merged input sources. Therefore when two input sources of similar ?ux are confused, the output position does not correspond to either of the input positions. Source confusion can limit the depth of any deep survey depending on the size of the telescope beam, and on the sky-density of objects as a function of ?ux. In practice, we cannot distinguish between a source boosted by photon noise or one confused with another faint source, therefore we consider the two e?ects jointly. We class sources as ‘?ux ampli?ed’ (corresponding to ‘confused’ sources in Hasinger et al. 1998) if S out /(S inp + 3σout ) > 1.5 (where σout is the 1σ error on the output ?ux S out ). Fig. 4a shows the fraction of ?ux ampli?ed sources as a function of input ?ux in our four energy bands. The fraction is less than 0.4 per cent at all ?uxes in all energy bands. This fraction will depend on the exact definition of ?ux ampli?cation. Using a less stringent de?nition of S out /(S inp + 3σout ) > 1.2 still results in a fraction well below 2% in each energy band at all ?uxes. We class an output source as ‘unmatched’ when there are no valid (S out 5S inp ) input sources within rcut (corresponding to ‘spurious’ sources in Hasinger et al. 1998). These are mainly caused by positive ?uctuations in the background. Fig. 4b shows the fraction of unmatched sources as a function of ?ux in each energy band. As expected, the unmatched fraction is highest at very faint ?uxes, where up to 30% of the sources are unmatched. Unmatched sources are many times more numerous than the ?ux ampli?ed sources at any ?ux. In our simulations we curtailed our input sourcelists at ?uxes 5× fainter than those found in the 13H data in each energy band in order to speed up processing time. In order to assess the impact of the simulation ?ux limit on the number of ?ux ampli?ed and spurious sources, we have also produced a smaller number of simulations to a greater depth, reaching ?uxes 10× fainter than those found in the 13H data. The fraction of ?ux ampli?ed sources in

these faint simulations agrees with the fraction found in our original simulations to within 0.02 per cent. Likewise, the fraction of unmatched sources agrees to within 2 per cent. We are therefore satis?ed that our chosen ?ux limits are su?ciently deep. In order to investigate the ?ux limits at which confusion noise dominates over Poisson errors, we have produced and sourcesearched a small number of ultra-deep simulations with no Poisson noise. The results are presented and discussed in Appendix B, and show that the 13H ?ux limits are more than a factor of 4 brighter than the ultimate XMM-Newton confusion limit in any of the 4 energy bands. Eddington bias (Eddington 1913) results in a systematic o?set in the number of sources detected at a given ?ux. The magnitude of this e?ect depends on the both the statistical errors on the measured ?ux values and on the intrinsic slope of the N (S ). As there are generally many more faint sources than bright ones, uncertainties on the measured ?ux values will result in more faint sources being up-scattered than bright sources being down-scattered. Therefore we would expect more faint sources to be detected than are actually input, and the output source counts at a given ?ux to be greater than those input. In the case of our simulations the situation is further complicated due to the double powerlaw form of the N (S ) distribution and the fact that the ?ux error distribution is nonuniform and a function of several parameters including ?ux and o?axis angle. The level of Eddington bias expected in the 13H deep ?eld is shown in Fig. 5 where the simulated input and output source counts are compared. Below S knee the ratio rises as the statistical errors on the ?ux measurements increase. At the lowest ?ux interval there is a drop in output source counts. The reason for this is the strong skew in S out /S inp at the faintest ?uxes (see Fig. 3) which boosts the output ?uxes. The output/input source counts ratio is a minimum at S knee where both the statistical errors on the ?ux measurement are low and the source counts become ?at. Above S knee the output/input source count ratio is constant within the errors with a value of ?
c 0000 RAS, MNRAS 000, 000–000

XMM-Newton Deep Field: X-ray source catalogue

7

Figure 5. Simulated output (crosses) 0.5-2 keV N (S ) distribution normalised by the input distribution. The error bars represent Poisson errors on the output source counts. The dotted line represents the case where the input and output source counts are equal. The source counts are most disparate at faint ?uxes where the output counts are enhanced by up to 23%.

Figure 6. Plot showing the fraction of ?ux ampli?ed or unmatched sources for a range of lower DET ML limits. Results are shown for each energy band, 0.2-0.5keV (large dash), 0.5-2keV (solid line), 2-5keV (small dash), 5-10keV (dot-dash). The dotted line shows the 5% badness level.

1.05 indicating that Eddington bias a?ects our bright counts at the 5% level. 3.2 Assessing the reliability of the 13H sources In this section we describe how we determined from our simulations the appropriate detection threshold to be applied to the 13H ?eld EMLDETECT sourcelist. Our aim was to produce a sourcelist such that nearly all erroneous detections are removed whilst retaining the maximum number of real sources. For each detected source, EMLDETECT measures the detection maximum likelihood statistic, DET ML, which takes account of source counts, background counts, and the PSF. For the real data, the SAS task EMLDETECT provides both single-band and multiband measurements of DET ML. For the simulated images we are limited to single band measurements as the four bands are simulated independently. To simulate all four bands simultaneously would require a priori knowledge of the sources’ intrinsic X-ray spectra, redshifts and column densities which we do not have. The value of DET ML for any single detection is directly related to the probability of the source being caused by a random Poisson ?uctuation via Eqn. 1. However, it is di?cult to translate a minimum threshold value of DET ML applied to the whole sourcelist into a total number of expected spurious sources in the ?eld as the probability is a function of position within the ?eld, due to the varying PSF and exposure map. This is therefore best explored via a large number of Monte-Carlo simulations. Using the simulated sourcelists in each band, we calculated the fraction of sources which were either ?ux ampli?ed or unmatched as a function of the minimum detection threshold DET MLmin . This is shown in Fig. 6. To restrict the fraction of bad sources in our ?nal 13H sourcelist we use the value of DET MLmin in each band at which only 5 percent of sources are either ?ux amc 0000 RAS, MNRAS 000, 000–000

pli?ed or unmatched in our simulations; these are 5.9, 5.9, 6.0, 8.1 in the 0.2-0.5, 0.5-2, 2-5, and 5-10 keV bands respectively. For a source to pass our signi?cance threshold we require that it has DET ML DET MLmin in at least one energy band. In practice, because we use a multi-band source detection process for the 13H data, we expect fewer than 5% bad sources after we have applied this criterion, as many of the sources will be detected in more than one energy band. For each source in the 13H sourcelist, we can identify all the output sources from the simulations which lie within 2′ of the real 13H source and have a similar DET ML (within 10%). The fraction of these output sources which are unmatched or ?ux ampli?ed gives a good estimate of the probability that the detection of the real source in this energy band is unreliable. For sources detected in more than one band the probability that the source is spurious is given by the product of these individual probabilities. The total number of spurious sources expected in the 13H ?eld can then be estimated by summing the probabilities from each source. 3.3 Maximum likelihood N (S ) ?tting method Our simulations are ideal for testing our N (S ) ?tting procedure because we know a priori the input N (S ) ?tting parameters. Accurate ?tting of any N (S ) relation requires a knowledge of the sky area searched and the probability of detection at a given ?ux. In Fig. 7 the e?ective area of the survey as a function of limiting ?ux is shown for each energy band. This was determined by comparing the number of output to input sources at each ?ux and multiplying the resultant fraction by the geometric area of the survey. Additionally, the distribution of S out /S inp is required to account for Eddington bias (see Fig. 3, and Section 3.1). Adapting the maximum likelihood method of Murdoch et al. (1973), we ?tted a double powerlaw model to the output N (S ) relation. Rather than ?tting the observed ?uxes directly, we

8

N. S. Loaring et al.
Energy (keV) 0.2–0.5 0.5–2.0 2.0–5.0 5.0–10.0 γ
+0.20 1.84? 0.18 +0.11 1.69? 0.11 +0.20 1.91? 0.19 +0.67 2.80? 0.55

K 47±20 112±23 126±28 150±18

KS Prob 1.8×10?2 1.0×10?10 3.5×10?4 2.9×10?5

Table 3. Best-?t values for a single powerlaw ?t to the 13H deep ?eld di?erential source counts. All errors are at 95% con?dence. The best-?t slope, γ, is listed together with the normalisation, K, in units of (10?14 erg cm?2 s?1 )γ?1 deg?2 . KS null-hypothesis probabilities on the ?ts are listed in the ?nal column.

Figure 7. E?ective area for each energy band, determined from MonteCarlo simulations (0.2-0.5 keV (dotted), 0.5-2 keV (solid), 2-5keV (dashed), 5-10 keV (dot-dashed)).

convolved a model N (S ) with the distribution of S out /S inp to produce a model probability distribution of observed ?uxes P(S out ). The best-?t N (S ) was then determined by maximising the likelihood of obtaining the observed ?ux distribution. As a check, we applied this technique to our simulated data, and recover the input values to within the statistical limits of our simulations (?2%). We are therefore con?dent in applying this technique to the real 13H data. Fits neglecting the distribution of S out /S inp are typically systematically o?set by ?5%

4 RESULTS In this section we present our ?nal source catalogue and results of ?ts to our source counts. We compare the properties of our sources with those detected by Chandra in the same area and closely examine our quoted positional uncertainties via comparisons with both Chandra detections and our simulations. 4.1 X-ray source catalogue Following the source detection procedure described in Section 2.2 (simultaneous source searching in four energy bands), a total of 275 sources were detected with DET ML > 5 in at least one energy band. This sourcelist was curtailed using the DET MLmin values for each energy band determined in Section 3.2. This reduced the ?nal number of sources to 225. The ?nal sourcelist is presented in Table 9. Using the procedure described in Section 3.2 we expect a total of 7 spurious sources. 4.2 13H deep ?eld source counts We show the integral N (> S ) in each energy band in Fig. 8. We have ?tted single and double powerlaw models to the unbinned differential source counts in all energy bands using the method described in Section 3.3. The four identi?ed Galactic stars in the ?eld

have been excluded from this analysis. The best-?t double powerlaw models are overlaid in Fig. 8 (single powerlaw in the 5-10 keV energy band) as solid lines. The 95% con?dence interval of the ?ts are indicated by the bowties. These incorporate errors on the slopes, normalisations and knee if applicable. The best-?t parameters for a single powerlaw ?t are listed in Table 4.2. The uncertainties on the ?t parameters are quoted at the 95% con?dence interval for one interesting parameter. We have tested the goodness of ?t of these models using the KolmogorovSmirnov (KS) test, and the null-hypothesis probabilities that we obtain are given in Table 4.2. The ?ts are unacceptable in all bands. For the double powerlaw ?ts we have ?xed the slope at bright ?uxes (γ2 ) and ?t only for the slope at faint ?uxes (γ1 ) and knee position (S knee ). In the 0.2-0.5 keV band we set γ2 = 2.51 based on results from the ROSAT All Sky Survey (Voges et al. 1999). The value of γ2 in the 0.5-2 keV band was ?xed at 2.60, derived from the RIXOS survey (Mason et al. 2000). For the 2-5 keV band γ2 was ?xed at 2.65 as found in the BeppoSAX High Energy Large Area Survey (HELLAS Giommi et al. 2000), and consistent with the ASCA derived value of 2.5 from Ueda et al. (1999). The ?t parameters for the double powerlaw ?ts are listed in Table 4. The double powerlaw model provides a better ?t compared to a single powerlaw model in each of these three energy bands. However, the double powerlaw model ?t is formally rejected with 99% con?dence in the 0.5-2 and 2-5 keV energy bands. In all three energy bands S knee occurs between 1.08 ? 1.27 × 10?14 erg cm?2 s?1. In the 5-10 keV band the slope found in the single powerlaw ?t is consistent with that found by Baldi et al. (2002) at brighter ?uxes. We attempted to ?t a double powerlaw model to the 5-10 keV band source counts, but in this case S knee is unconstrained. We therefore consider that a double powerlaw model for the source counts is not justi?ed at the depth of our survey in this energy band. 4.3 Positional uncertainties To assess the reliability of the EMLDETECT positional errors in the XMM-Newton sourcelist we have cross-correlated the ?nal XMM-Newton and Chandra sourcelists. The spread in positional o?sets between the two catalogues should provide a good representation of the true spread in XMM-Newton positional errors. To prevent the (albeit small) statistical errors on the Chandra positions contributing to the spread, where possible we used the position of the optical counterpart to the Chandra source: the positional errors on the optical counterparts are typically less than 0.3′′ (McHardy et al. 2003). The two sourcelists were matched within a search radius of 10′′ over the entire XMM-Newton ?eld of view. The two catalogues have 155 sources in common. The positional o?set between the 155 matched XMM-Newton and Chandra
c 0000 RAS, MNRAS 000, 000–000

XMM-Newton Deep Field: X-ray source catalogue

9

1000 N(>S) (deg-2)

0.2-0.5 keV

1000 N(>S) (deg-2)

0.5-2.0 keV

100

100

13H ROSAT fit CDF-N fit CDF-S fit HELLAS2XMM fit

13H counts

10

95% error bowtie 13H fit

10

13H counts 95% error bowtie 13H fit

0.1 1.0 S (10-14 erg cm-2 s-1) 1000 1000 2.0-5.0 keV N(>S) (deg-2) 100

0.1 1.0 S (10-14 erg cm-2 s-1) 5.0-10.0 keV

N(>S) (deg-2)

100
Cowie et al. 2002 fit ASCA fit CDF-N fit CDF-S fit HELLAS2XMM fit 13H counts 95% error bowtie 13H fit

10

10

BeppoSAX counts ASCA fit CDF-S fit HELLAS2XMM fit 13H counts 95% error bowtie 13H fit

1

0.1

1.0 S (10 erg cm-2 s-1)
-14

1 S (10-14 erg cm-2 s-1)

Figure 8. 13H deep ?eld integral N (> S ) in each energy band. Bowties indicate 95% errors. Overlaid for comparison are the results from the CDF-S (dash) (Rosati et al. 2002), CDF-N (long dash) (Brandt et al. 2001), HELLAS2XMM (dot) (Baldi et al. 2002) and ASCA (dot-dot-dot-dash) (Cagnoni, Della Ceca & Maccacaro 1998) surveys. In the 0.5–2.0 keV band the ROSAT counts found in the 13H ?eld (McHardy et al. 1998) are also overlaid (dot-dash). In the 2–5 keV band the results from Cowie et al. (2002) are also overlaid (dot-dash). Triangles in the 5–10 keV band denote the BeppoSAX counts of Fiore et al. (2001).

Energy (keV) 0.2–0.5 0.5–2.0 2.0–5.0

γ1
+0.25 1.74? 0.26 +0.19 1.41?0.18 + 0 30 1.66?0..47

γ2 2.51 2.60 2.65

S knee
+11.08 1.16? 0.69 + 1.02 1.08?0 .39 + 1 66 1.27?0..70

K1 56±23 183±51 163±94

K2 62±88 201±67 207±76

KS Prob 7.7×10?2 6.4×10?5 1.5×10?2

Table 4. Best-?t parameters for a double powerlaw ?t to the 13H di?erential source counts. The slopes and normalisations below and above the break ?ux, S knee are denoted by 1 and 2 respectively. The values of γ2 in the three energy bands were ?xed to appropriate values found from the literature. The break ?ux, S knee is in units of 10?14 erg cm?2 s?1. Normalisations K1 and K2 are in units of (10?14 erg cm?2 s?1 )γ?1 deg?2 . All errors are at 95%. KS null-hypothesis probabilities on the ?ts are listed in the ?nal column.

c 0000 RAS, MNRAS 000, 000–000

10

N. S. Loaring et al.

Figure 9. Positional o?sets between XMM-Newton and Chandra counterparts as a function of XMM-Newton o?axis angle. Triangles denote XMMNewton detections with > 1 Chandra counterparts within 10′′ (numbers indicate the XMM-Newton ?ux in units of 10?14 erg cm?2 s?1 ). The shaded areas indicate the regions where 68% (dark grey) and 95% (light grey) of the simulated data lie. In the case of the simulations the positional o?set is that between the input and output source. 86% of the 13H data lies within the 68% simulation contour, indicating that the real data have better source positions. This is due to the fact that the real data are generally detected in > 1 energy band.

Figure 10. Positional o?sets between XMM-Newton and Chandra counterparts compared with quoted XMM-Newton positional error. The shaded areas indicate the regions where 68% (dark grey) and 95% (light grey) of the simulated data lie. In the case of the simulations the positional o?set is that between the input and output source. 43% of the 13H data lie within the 68% simulation contour indicating that EMLDETECT underestimates the true positional error.

sources as a function of XMM-Newton o?axis angle is shown in Fig. 9. The positional o?sets increase with XMM-Newton o?axis angle, but 80 per cent of the matched sources have positional o?sets 2′′ ; 95 per cent of the sources have positional o?sets 4′′ . We compared this distribution with that found from the simulations (see Section 3.1) which should provide a good indication of the distribution of positional o?sets arising within the source detection chain. The contours in Fig. 9 show the distribution of di?erences between input and output source positions in the 0.5-2 keV simulations. The dark (light) grey area shows where 68% (95%) of the simulated sources lie. The distribution of XMM-Newton Chandra o?sets is more highly peaked than the input-output o?sets in our simulations: 86% of XMM-Newton -Chandra o?sets lie within the 68% simulation contour. This is because the real sources in the 13H ?eld are generally detected in more than one energy band, contrary to the simulated sources. Fitting the source positions in all four energy bands simultaneously, as was done to produce our source catalogue, results in better positional accuracy than a single band ?t because more source counts are used in the ?t. Five of the matched sources have more than one Chandra counterpart. However, in two out of the ?ve cases, the closest counterpart is signi?cantly brighter than the other one, and we are easily able to identify the correct counterpart. This suggests that ? 2% of our sources are confused, in broad agreement with the fraction of ?ux ampli?ed sources expected from our simulations. In Fig. 10 we plot the statistical positional uncertainties of the XMM-Newton sources given by EMLDETECT against the actual positional uncertainties given by the XMM-Newton - Chandra o?sets. These are compared with our 0.5-2 keV simulation distribu-

tions where in the case of the simulations the positional o?sets are those of the input/output source positions. The dark (light) grey areas show where 68% (95%) of the simulated sources lie. In general, the positional o?sets are slightly larger than the EMLDETECT errors for the 13H data. However, for EMLDETECT positional errors less than 1′′ the positional o?sets have a broader distribution in the 13H data than in the simulations: only 43% of the 13H data lies within the 68% simulation region. This implies that the EMLDETECT positional error in the 13H ?eld underestimates the true positional error. We have therefore added in quadrature a systematic error to the statistical EMLDETECT positional error of each 13H source, such that 68% of the 13H sources lie within the 68% region of the simulations. This is achieved with an additional systematic positional error of 0.76′′ (1σ). This additional component may be due to residual uncertainties in the detector geometry and may represent a fundamental limit to the accuracy of any XMMNewton position. According to our simulations, EMLDETECT positional errors greater than ? 2.5′′ may underestimate the true positional error by an order of magnitude (the 95% simulation region limit samples a non-gaussian tail at this point) and should be used with this caveat in mind. Section 5.2 further discusses the reliability of the survey and the implications of the positional uncertainties. 4.4 Comparisons with Chandra detections There are 70 XMM-Newton sources with no Chandra counterpart. Of these, 9 sources are either on the very edge or out of the ?eld of view of the Chandra mosaic (see McHardy et al. 2003) and would therefore not be expected to have a Chandra counterpart. R-band images, centred on the positions of these 70 sources are presented in Fig. 11. (Optical images of those sources with a Chandra counterpart are presented in McHardy et al. 2003). In 49 cases there is a
c 0000 RAS, MNRAS 000, 000–000

XMM-Newton Deep Field: X-ray source catalogue

11

Figure 11. Top: 10” × 10” R-band Subaru SuprimeCam images centred on the XMM-Newton sources without Chandra counterparts. The XMM-Newton reference number (shown in column 1 of Table 9) is at the top left of each image. The XMM-Newton o?-axis angle is at the top right. The XMM-Newton 0.2 ? 10 keV ?ux (in units of 10?15 erg cm?2 s?1 ) is at the bottom left and the optical counterpart R magnitude, if applicable, is given in the bottom right. All of these SuprimeCam images have the same greyscale levels to ease comparison of the optical counterparts’ brightnesses. The greyscale was chosen to correspond to that in McHardy et al. (2003). Bottom: 30′′ × 30′′ R-band Subaru SuprimeCam images centred on the 6 XMM-Newton sources with Chandra counterparts which are too extended to ?t in a 10′′ × 10′′ box. Source number 18 is actually shown in a 70′′ × 70′′ box. The top images are at the same grey scale level as Fig. 11. The lower images have their grey scales adjusted for maximum contrast. Source number 10 is not listed as having an optical counterpart in Table 5 as the centre of the barred spiral galaxy is outside the error circle of the XMM-Newton source. However, we may be seeing X-ray emission from an X-ray source in the spiral arms of the galaxy. c 0000 RAS, MNRAS 000, 000–000

12

N. S. Loaring et al.
XMM-Newton No. 1 2 3 10 11 18 20 22 26 28 31 33 35 36 40 44 47 54 56 62 63 66 70 79 82 83 87 88 89 90 104 110 116 117 118 120 122 123 133 136 139 141 144 147 149 150 151 153 154 155 159 160 163 172 177 179 182 185 194 205 206 209 210 211 215 217 218 219 220 225 RA J(2000) Dec J(2000) Pos. err. (arcsec) 1.57 2.38 3.04 4.83 1.99 2.56 0.92 1.11 1.77 1.23 2.91 1.18 2.34 12.19 1.50 1.23 2.69 2.01 1.22 1.80 1.31 2.04 1.40 2.15 1.58 1.45 1.92 1.78 1.79 1.76 1.99 1.25 1.11 1.60 1.29 0.78 1.55 2.33 1.53 1.06 1.23 1.78 1.76 1.30 2.44 2.66 2.06 1.74 1.48 2.34 1.22 2.32 1.17 1.90 2.66 1.55 1.86 1.49 14.06 1.68 1.90 0.70 4.66 1.98 1.34 0.68 0.59 1.72 2.17 2.76 O?axis (arcmin) 12.91 12.67 12.53 12.96 11.74 9.95 10.14 9.27 8.88 11.21 8.70 8.87 7.57 8.75 8.55 7.36 8.72 6.06 10.12 8.97 4.43 10.89 4.26 2.32 1.64 5.73 3.09 1.67 7.90 11.98 12.13 10.32 14.58 12.22 2.38 13.24 12.33 9.13 3.46 2.85 14.46 12.18 7.59 13.41 6.53 5.96 15.63 6.32 4.59 4.94 5.40 6.22 5.69 11.26 9.24 11.43 9.62 9.78 10.19 10.57 14.34 14.39 15.08 13.50 14.55 12.92 13.49 13.74 13.70 16.68 0.2-10.0 keV Flux 14.86 5.94 2.55 7.88 1.65 1.74 8.82 3.46 4.43 16.49 0.98 6.60 2.06 1.86 5.92 4.87 4.67 1.10 2.42 4.81 5.54 3.51 3.58 0.90 1.63 2.66 2.21 2.15 0.89 7.29 6.71 5.16 34.78 5.49 2.91 10.84 9.25 1.65 3.71 7.93 17.14 4.16 3.59 12.44 2.87 1.32 15.15 4.97 3.04 0.63 3.34 1.16 7.07 3.52 2.74 6.08 1.73 20.46 1.44 3.14 3.94 238.21 71.49 3.24 5.96 20.15 23.77 6.49 3.95 17.44 Extension R mag. Chandra counts 3 1 0 0 2 3 0 1 3 0 1 1 2 0 1 0 1 1 3 2 2 2 0 3 2 1 3 0 0 1 3 4 0 0 4 0 0 1 4 0 0 1 2 4 1 4 0 0 0 0 0 1 2 0 0 1 0 0 15 0 0 0 2 1 0 0 0 1 0 0 Faint Chandra source? n n n n n y n n y n n n n n n n n n y n y y y n n n y n n n n y n n y n n n y n n n y n n n n n n n n n n n n n n n n n n n n n n n n n n n Comments

13 33 28.96 13 33 30.32 13 33 30.68 13 33 38.91 13 33 40.70 13 33 43.80 13 33 45.45 13 33 47.52 13 33 50.80 13 33 54.21 13 33 57.90 13 33 58.87 13 33 59.56 13 33 59.63 13 34 01.06 13 34 07.18 13 34 08.39 13 34 12.36 13 34 13.59 13 34 15.49 13 34 15.54 13 34 17.80 13 34 19.56 13 34 22.70 13 34 25.91 13 34 26.87 13 34 28.66 13 34 28.80 13 34 29.20 13 34 29.30 13 34 34.92 13 34 36.38 13 34 38.83 13 34 38.96 13 34 39.75 13 34 42.29 13 34 42.81 13 34 42.84 13 34 47.18 13 34 48.22 13 34 51.27 13 34 51.86 13 34 53.13 13 34 54.55 13 34 55.11 13 34 55.41 13 34 56.39 13 34 56.71 13 34 56.76 13 34 56.87 13 34 58.11 13 34 58.14 13 34 59.48 13 35 05.13 13 35 07.69 13 35 09.25 13 35 12.27 13 35 14.41 13 35 17.03 13 35 24.72 13 35 24.88 13 35 32.22 13 35 32.74 13 35 33.05 13 35 37.26 13 35 39.37 13 35 42.49 13 35 43.17 13 35 43.26 13 35 58.70

37 55 58.82 37 56 12.21 37 55 05.57 38 01 56.15 37 49 47.79 37 54 55.76 37 58 08.04 37 53 51.51 37 57 16.61 38 02 53.17 37 49 59.01 38 00 24.88 37 51 39.75 37 49 26.35 38 00 25.75 37 59 59.34 37 47 50.85 37 59 10.36 37 45 39.06 38 03 05.78 37 52 27.80 37 44 31.39 37 51 47.77 37 55 23.49 37 54 59.90 38 00 27.92 37 57 48.26 37 53 37.96 38 02 45.33 38 06 51.45 38 07 03.03 38 05 14.02 37 40 22.03 37 42 43.95 37 57 01.93 37 41 46.46 37 42 42.62 38 03 53.27 37 57 14.88 37 54 14.33 37 40 51.30 38 06 35.33 37 48 18.85 37 42 07.32 37 49 51.41 37 50 40.54 37 39 54.75 37 50 25.09 37 56 03.31 37 52 48.73 37 57 33.15 37 50 52.22 37 57 39.65 37 45 27.22 37 48 27.88 38 04 01.91 37 48 54.89 37 49 12.11 37 49 13.98 37 51 24.06 37 44 38.51 38 03 39.97 37 45 14.16 37 48 02.14 37 47 22.80 37 56 14.96 37 55 42.93 37 52 58.65 37 53 24.27 37 54 06.31

– – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – 9.2±0.4 – – – 62±104 26±2 – – – – – – 6.8±1.2

19.74 23.18 25.83 0.00 23.76 15.63 17.45 23.39 25.41 25.41 24.79 24.83 25.69 17.64 20.50 22.50 22.65 26.57 16.43 19.68 23.65 18.76 22.89 18.17 23.94 0.00 20.64 18.46 22.56 21.88 24.47 25.32 25.30 24.30 16.14 20.14 0.00 18.53 23.73 0.00 0.00 22.80 24.06 0.00 25.26 24.95 24.89 0.00 20.66 25.00 0.00 0.00 26.38 0.00 25.23 0.00 24.32 20.55 19.44 0.00 23.41 25.10 22.43 18.54 20.28 20.89 19.47 26.58 25.89 17.51

Possible faint source outside error circle

Edge of Chandra FOV

Outside Chandra FOV

Outside Chandra FOV Edge of Chandra FOV

Outside Chandra FOV

Outside Chandra FOV

Large error circle

Outside Chandra FOV Outside Chandra FOV Chandra 102 is ? 10” North Chandra 102 is ? 15” South Outside Chandra FOV

Table 5. Properties of XMM-Newton sources not found in the Chandra catalogue of McHardy et al. (2003). The XMM-Newton ?ux is in units of 10?15 erg cm?2 s?1 . The extension of the source (if applicable) in the XMM-Newton image in units of arcsec. The column marked ‘Chandra counts’ gives the number of counts detected within the XMM-Newton error circle in the Chandra image. Excess emission associated with a possible faint source is indicated by a ‘y’ in column 10. See the main text for further details.

c 0000 RAS, MNRAS 000, 000–000

XMM-Newton Deep Field: X-ray source catalogue
Chandra No. 82 91 96 102 115 121 145 149 154 158 159 163 164 171 181 182 183 185 188 192 193 195 197 198 201 203 204 205 206 207 208 210 211 213 XMM-Newton o?axis angle (arcsec) 7.0 12.5 13.0 13.0 11.8 10.4 10.3 7.1 11.7 11.5 6.5 12.1 10.9 8.6 3.9 11.6 7.4 7.0 7.0 4.3 2.4 8.0 7.5 5.5 3.9 3.8 7.4 3.3 7.7 7.8 7.4 1.0 9.5 7.2 0.5-7 keV Flux 1.16 0.94 0.91 0.80 0.70 0.66 0.53 0.50 0.47 0.46 0.45 0.43 0.43 0.37 0.33 0.32 0.31 0.31 0.29 0.27 0.27 0.27 0.26 0.26 0.24 0.23 0.23 0.23 0.23 0.22 0.21 0.19 0.19 0.19 Faint XMM-Newton source? y y y y y n y y n y y y n n n y y y y y y n y n n y y n y n n y y y Comments 9′′ away from Chandra 28/XMM-Newton 50. Below ?nal threshold. 18′′ away from Chandra 77/XMM-Newton 99. Between XMM-Newton 219 and XMM-Newton 220. (11′′ and 15′′ away respectively). 7′′ away from Chandra 107/XMM-Newton 214 9′′ away from Chandra 61/XMM-Newton 171. 15′′ away from Chandra 23/XMM-Newton 39. 13′′ away from Chandra 117/XMM-Newton 78. 11′′ away from Chandra 20/XMM-Newton 175.

13

Below ?nal threshold. Below ?nal threshold. 14′′ away from Chandra 52/XMM-Newton 43. Below ?nal threshold. Below ?nal threshold. Below ?nal threshold.

12′′ away from Chandra 108/XMM-Newton 103. In a region of extended emission. 18′′ away from Chandra 10/XMM-Newton 119. 7′′ away from Chandra 142/XMM-Newton 128.

Below ?nal threshold. 40′′ away from Chandra 28/XMM-Newton 50,

Table 6. List of Chandra sources with no XMM-Newton counterpart. The Chandra ?ux is in units of 10?15 erg cm?2 s?1. Excess emission associated with a possible faint source is indicated by a ‘y’ in column 4. The comment ‘Below ?nal threshold’ indicates a source which was detected with EMLDETECT with DET ML> 5 but subsequently excluded from the ?nal sourcelist according to the criteria in Section 3.2.

clear optical counterpart or counterparts within the positional error radius of the source. Those sources which have optical counterparts extended beyond 10′′ are shown in wider images covering 30′′ × 30′′ below the main image. Those in the top panel have the same greyscale applied as in the main image, however in the lower panel the greyscale has been individually adjusted for greater clarity. Searching out to a radius of 3′′ the number of sources with possible optical counterparts is increased to 58. Of the 9 sources on the edge of, or outside, the Chandra ?eld of view, 7 have an optical counterpart. Table 5 lists the basic properties of the 70 XMM-Newton sources without Chandra counterparts. The Chandra mosaic image was visually inspected at the positions of each XMM-Newton source to see if any faint X-ray emission could be seen that was not formally detected in the Chandra source searching. Details of the number of counts observed within the XMM-Newton positional error circle for each source are listed in column 10 and further comments are listed in column 11. Of the 61 sources not formally detected within the Chandra ?eld of view, faint emission is visible in
c 0000 RAS, MNRAS 000, 000–000

11 cases. None of the extended XMM-Newton sources are detected with Chandra. The XMM-Newton observations are more sensitive to faint extended sources than the Chandra observations. One example is the ROSAT source R117 (McHardy et al. 1998) which was detected with XMM-Newton (number 56) but missed by Chandra . This source is a faint starburst galaxy with extended emission, probably on the scale of the galaxy (Gunn et al. 2001).

Conversely, there are 34 Chandra sources with no XMMNewton counterparts (within the XMM-Newton FOV). The properties of these sources are summarised in Table 6. Seven of these sources were originally detected in our initial XMM-Newton sourcelist, but removed from our ?nal sourcelist after we applied the DET ML cuto?s derived in Section 3.2. Visual inspection of the XMM-Newton images at the remaining 27 Chandra source positions, suggests a faint source in 16 cases. As well as the 4 close (< 10′′ ) pairs of sources found with Chandra there are also 9 faint sources which were missed with XMM-Newton because they are in the wings of much brighter sources. This highlights the impor-

14

N. S. Loaring et al.

tance of our Chandra coverage in order to obtain accurate source positions.

5 DISCUSSION 5.1 Source density in the 13H ?eld Our 13H ?eld represents one of the deepest blank ?eld surveys with XMM-Newton and as such it is ideal for the investigation of faint X-ray source counts. Further, this work is the only XMM-Newton based survey published to date to incorporate detailed Monte-Carlo modelling of the detection process allowing accurate ?tting of the source counts. Our source counts are inconsistent with both a single and double powerlaw ?t in all energy bands, except for our lowest energy band (0.2-0.5 keV) where a double powerlaw ?t is acceptable. However, a double powerlaw ?t provides a better representation of the source counts according to the KS test in all but our hardest (5-10 keV) energy band where we are unable to constrain the break. Previous studies show a consensus in shape: the measured source counts are described by a double powerlaw distribution ?attening below a ?ux of ? 1 × 10?14 erg cm?2 s?1 and steeper than an Euclidean slope at ?uxes above the break (S > S knee ). However the measured normalisations di?er by up to ? 30% (Brandt et al. 2001; Rosati et al. 2002). Yang et al. (2003) attribute this ?eld-to?eld variation to the fact that the ?elds studied cover small areas and are subject to cosmic variance. Here the underlying clustered large scale structure imprints a variation in source density on small scales. The slope of our soft band (0.5-2 keV) source counts, be?14 +1.02 erg cm?2 s?1 , is consislow the break ?ux of 1.08? 0.39 × 10 tent with recent determinations using Chandra (Rosati et al. 2002; Bauer et al. 2004; Yang et al. 2004). Our 13H ?eld is the richest X-ray blank ?eld observation reported to date in this energy band. Our overall normalisation is higher, though consistent with, the CDF-N results of Brandt et al. (2001) and the HELLAS2XMM counts of Baldi et al. (2002). However, the CDF-S source counts of Rosati et al. (2002), and the observed counts in the Lockman hole (Hasinger et al. 1998, 2001) are lower by ?30%. Allowing for the errors on the CDF-S source counts, given by Rosati et al. (2002), and Poisson errors on our counts, our ?eld is still inconsistent with the CDF-S at the 2.7σ level. It is likely therefore that cosmic variance has an important e?ect on the measured source counts found in deep X-ray ?elds. In the 2-5 keV energy band our observed counts are consistent with previous Chandra and XMM-Newton studies (Brandt et al. 2001; Hasinger et al. 2001; Rosati et al. 2002; Baldi et al. 2002; Cowie et al. 2002; Yang et al. 2004). Our brightest source is a factor of ? 10× fainter than the limiting ?uxes of the serendipitous ASCA surveys (Cagnoni, Della Ceca & Maccacaro 1998; Ueda et al. 1999,a) which makes comparisons di?cult. However, the ASCA ?ts of Cagnoni, Della Ceca & Maccacaro (1998) appear to be consistent with ours when extrapolated to the ?ux range covered by our survey. Our overall normalisation at faint ?uxes is closer to the CDF-S normalisation than it is to that found in the CDF-N in this energy band. However, the CDF-S counts drop far more rapidly than ours towards brighter ?uxes. In our hardest energy band (5-10 keV) the observed +0.67 counts are consistent with a Euclidean slope; γ = 2.80? 0.55 . The slope agrees within the errors with the results of Cagnoni, Della Ceca & Maccacaro (1998), Fiore et al. (2001),

Figure 12. 0.5-2 keV energy band di?erential N (S ) distribution. Overlaid is our double powerlaw ?t (solid line). Signi?cant excesses over the model distribution occur at a ?ux of ? 2 × 10?15 erg cm?2 s?1 . Also shown is the best ?t model from the Chandra Survey of the Lockman Hole-Northwest (Yang et al. 2004), which covers a similar ?ux range as our survey (dotted line).

Baldi et al. (2002) and the XMM-Newton 5-10 keV counts in the Lockman hole (Hasinger et al. 2001). We expect a break in the source counts at a ?ux of ? 4 × 10?15 erg cm?2 s?1 as reported from Chandra observations of the CDF-S (Rosati et al. 2002). It is therefore unsurprising that it is not detected in our survey, as we have no sources below this ?ux in our 5-10 keV sourcelist. At the brightest ?uxes our counts are in agreement with ?ndings from the HELLAS2XMM survey (Baldi et al. 2002), BeppoSAX counts (Fiore et al. 2001) and ASCA counts (Cagnoni, Della Ceca & Maccacaro 1998). Although improved upon a single powerlaw, the double powerlaw ?t to the source counts is formally rejected in all but the softest energy band (0.2-0.5 keV). The di?erential counts show more clearly where the deviations from the model occur and are illustrated in Fig. 12 and Fig. 13. There is an excess over the ?t at a ?ux of ? 2 × 10?15 erg cm?2 s?1 in both the 0.5-2 keV and 2-5 keV energy bands. The fact that this excess is seen in both energy bands suggests that the feature is most likely due to clustering in the ?eld rather than the X-ray spectral properties of the sources around this ?ux. We defer a study of clustering to a later paper (Loaring et al. 2005). We have also compared our source counts with wider area Chandra surveys which we expect to provide a more accurate value of the global source counts. In the 0.5-2 keV energy band our ?ts to the di?erential counts are similar to the ?ts obtained for the Lockman Hole-Northwest ?eld (Yang et al. 2004). This implies therefore that the CDF-S is a particularly underdense region. In the 25 keV energy band our double powerlaw ?t lies above the ?t of Cowie et al. (2002) which was obtained from the source counts in the CDF-S, CDF-N, and two Hawaii survey ?elds (SSA13 and SSA22) observed with Chandra . If the CDF-S data is excluded from their analysis their faint end normalisation agrees very well with ours, again suggesting that the CDF-S is underdense. This
c 0000 RAS, MNRAS 000, 000–000

XMM-Newton Deep Field: X-ray source catalogue
R magnitude 18–19 19–20 20–21 21–22 22–23 23–24 24–25 25–26 Expected number 0.5 1.1 2.4 5.3 11.7 26.1 57.7 127.7 Observed number 10 11 18 14 61 54 42 34

15

Table 7. The number of XMM-Newton optical counterparts as a function of R magnitude. X-ray – optical cross matching was carried out within the 95% positional error circle of each XMM-Newton source and the number of optical detections are displayed as a function of R magnitude. Also shown is the number of expected chance coincidences within each R magnitude interval given the optical source counts of McHardy et al. (2003).

Figure 13. 2-5 keV energy band di?erential N (S ) distribution. Our best ?t double powerlaw ?t is overlaid (solid line). Signi?cant excesses over the model distribution occur at a ?ux of ? 3 × 10?15 and ? 2 × 10?14 erg cm?2 s?1 . The best ?t model of Cowie et al. (2002) is also overlaid (dotted line) which is based upon ?ts to the combined Chandra deep ?eld counts and Hawaii survey ?elds SSA22 and SSA13.

appears to contradict the earlier comparisons between the integral 2-5 keV band counts in which our faint end normalisation closely matches the ?ts from Rosati et al. (2002). However, the Rosati et al. (2002) CDF-S counts clearly drop below ours at ?uxes > 5 × 10?15 erg cm?2 s?1 and it is at ?uxes above the knee that our counts are most discrepant from Cowie et al. (2002). The comparisons depend upon the adopted conversion between instrument energy bands and also the range over which ?tting is performed. Rosati et al. (2002) ?t to much fainter ?uxes than Cowie et al. (2002). As the source counts ?atten at faint ?uxes this would result in an apparently higher normalisation at brighter ?uxes and may explain why the Rosati et al. (2002) normalisation is higher than the Cowie et al. (2002) normalisation at ?uxes around 2 × 10?15 erg cm?2 s?1 . In converting values from the literature into our energy bands we have used a photon index of Γ = 1.7 which represents the average slope of the XRB in the 0.2-12 keV energy range. Other authors have chosen to either use individual photon indices measured from their sources (Brandt et al. 2001) or a relatively ?at Γ = 1.4 (Rosati et al. 2002) which represents the overall XRB spectral slope. Cowie et al. (2002) use a photon index of Γ = 1.2. as they assume that the absorbed population must be signi?cantly harder than the unabsorbed population to produce the overall XRB spectrum. However, this photon index appears to be inappropriate at the ?uxes probed by our survey. Considering the results from our X-ray spectroscopy (Page et al. 2003, 2005), the vast majority of sources have X-ray spectra with a softer photon index than Γ = 1.2, although they may be absorbed in the soft band. In fact the value of the photon index used has little e?ect on source counts. Using a photon index of 1.2 and 1.7 respectively to convert the 2-8 keV counts of Cowie et al. (2002) into our 2-5 keV energy band results in a integral normalisation di?erence of ?13% at our faintest ?uxes (2.2 × 10?15 erg cm?2 s?1 ): counts derived assuming Γ = 1.7 are higher than those assuming Γ = 1.2. This direct conversion does
c 0000 RAS, MNRAS 000, 000–000

not account for the fact that the original measured ?uxes would have been derived from count rates using di?erent photon indices. We have investigated this e?ect further using spectra which were simulated in XSPEC using the Chandra ACIS-I response matrix. Assuming photon indices of Γ = 1.7 and Γ = 1.2 respectively to convert count rates to ?uxes, we obtain a ?ux ratio F1.7 /F1.2 of 0.9 in the 2-5 keV energy band. Incorporating this factor into our ?ux conversions reduces the apparent normalisation di?erence to ?3% at a ?ux of 2.2 × 10?15 erg cm?2 s?1 (with the normalisation of the Γ = 1.7 source counts higher than that of the Γ = 1.2 source counts). This factor is not large enough to account for the measured di?erences in normalisation in the 2-5 keV energy band. It is not necessary to employ any ?ux conversions for the 0.5-2 keV energy band. However, there is still a ? 30% discrepency in source counts between the deep ?elds surveyed to date. One concern is the cross-calibration between the di?erent instruments used. However, ASCA and Chandra ?uxes agree at about 10% and Chandra and XMM-Newton ?uxes agree at the 5% level, and so the di?erences in normalisation cannot be attributed to instrument calibration o?sets (Snowden 2001). In the 0.5-2 keV band, we are con?dent that the ?eld-to-?eld variations observed are real and due to cosmic variance. We ?nd a larger ?eld-to-?eld variation in the soft band (0.5-2 keV) source counts than in the hard band counts (2-5 keV). These ?ndings are at odds with the recent clustering measurements of Yang et al. (2003) who ?nd hard band sources to be more strongly clustered than soft band sources. Further, Gilli et al. (2005) have found no signi?cant di?erence in clustering strength between soft and hard sources. At present the evidence for any variation in clustering strength due to X-ray hardness appears inconclusive. Clearly larger datasets are required to investigate this issue further. Comparisons with wide area XMM-Newton and Chandra serendipitous surveys coupled with further blank ?eld observations should provide su?cient signal to make signi?cant advances in this area.

5.2 Survey reliability and capabilities Our survey was speci?cally designed to study the X-ray source population within one decade in ?ux either side of the break in the source counts. Contrary to recent deep Chandra surveys we aim not to resolve the entire XRB but rather to determine the dominant emission mechanisms and amount of absorption in these faint sources. To this end we require high quality X-ray spectra, in par-

16

N. S. Loaring et al.
Radio Flux (mJy) 50–100 10–50 5–10 1–5 0.5–1 0.1–0.5 0.05–0.10 Expected number 3.1 × 10?3 4.4 × 10?2 1.8 × 10?2 0.3 0.1 3.3 2.0 Observed number 1 0 0 3 4 18 6

Table 8. The number of XMM-Newton radio counterparts as a function of radio ?ux. Radio – optical cross matching was carried out within the 95% positional error circle of each XMM-Newton source and the number of radio detections are displayed as a function of radio ?ux. Also shown is the number of expected chance coincidences given the best-?t radio source count model of Seymour et al. (2004).

ticular at hard energies to probe the absorbed population. The high throughput of XMM-Newton at hard energies makes it particularly suited to this kind of study. To e?ciently study the faint source population we require a highly reliable source catalogue with minimal spurious source contamination. We have examined the quality of our ?nal catalogue via extensive Monte-Carlo simulations (Section 3). Employing the likelihood cuto?s in each energy band determined in Section 3.2 we expect only 7 spurious sources to remain in our ?nal source catalogue. Measurements of the fraction of ?ux ampli?ed sources in our simulations indicate that confusion is unimportant (< 2% of our sources are confused in all energy bands) at the ?uxes probed by this survey. This is con?rmed via our a cross-correlation with the Chandra catalogue of McHardy et al. (2003). We ?nd that only 2% of our sources have ambiguous Chandra counterparts. The average input and output ?uxes from our simulations at bright ?uxes agree to within 1%. This has important consequences for the determination of source counts, and suggests any systematic deviation between input and output source counts should be very small. Our simulations demonstrate that Eddington bias a?ects our measured source counts by at most 23 per cent at the faintest ?uxes. However, we correct for the e?ects of Eddington bias when ?tting our source counts, using the distribution of S out /S inp obtained from our simulations, as described in Section 3.3. Our simulations indicate that the o?set in position between input and output sources is less than 2′′ in 68% of cases within the central 9′ of the XMM-Newton ?eld of view (Section 3.1). We therefore expect the majority of our sources to have positions accurate to 2′′ . However, the EMLDETECT positional errors output from the detection chain are too small and an additional systematic positional error of 0.76′′ is required (as determined in Section 4.3) to match the real and simulated positional error distributions. The high ?delity of our source positions and ?uxes, coupled with insigni?cant confusion illustrate the high reliability of the survey. We are able to construct useful X-ray spectra down to a 2-5 keV band ?ux of 3 × 10?15 erg cm?2 s?1 , well beyond the initial survey goals (Page et al. 2005). However, in order to extract the maximum scienti?c information from the survey, the X-ray sources need to be optically identi?ed, and compared with sources detected in other wavebands, such as our radio source catalogue (Seymour et al. 2004). This requires highly accurate source positions to identify the correct counterpart. The additional systematic positional error has important consequences for optical identi?cation. Using the optical source counts

determined directly within the 13H ?eld (McHardy et al. 2003) we have calculated the expected number of chance coincidences within an appropriate radius of each source within the ?eld (corresponding to the 95% positional error radius of each source). Table 7 shows the expected number of chance coincidences for the 225 XMMNewton sources as a function of optical magnitude; it also shows the actual number of optical sources observed within the error circles. X-ray/optical associations at magnitudes R 23 are ?rm, but approximately half of those at magnitudes 23 < R < 24 are likely to be spurious. At fainter magnitudes the majority of optical counterparts are probably unrelated to the X-ray source. This highlights the importance of the Chandra coverage over the ?eld which provides accurate X-ray positions and extends the range in which we can correctly select an optical counterpart to a magnitude of R = 26. This is an issue that must be considered in any deep XMM-Newton survey since if ignored it may lead to a bias in optical identi?cations. Chandra source counts determined by source type ?nd an increasing contribution of absorbed AGN and normal galaxies at the fainter ?uxes (Bauer et al. 2004). This suggests that broad line AGN, which dominate at brighter ?uxes, are most likely to be correctly identi?ed even without Chandra positions. However, in an XMM-Newton survey of similar or greater depth than ours, Chandra coverage is essential to identify correctly the optical counterparts of the faintest sources, which are more likely to be associated with absorbed AGN or normal galaxies. We have carried out a similar comparison with the 13H ?eld radio catalogue of Seymour et al. (2004). The predicted number of chance coincidences (as predicted from the best-?t starburst and AGN population model of Seymour et al. 2004) are compared with the observed number of radio counterparts in Table 8. We are able to securely identify the correct radio counterpart to a radio ?ux of 0.1mJy without Chandra positions. Below this ?ux, we need Chandra positions in order to reliably identify real associations between X-ray and radio sources. The relative ease of the identi?cation of radio counterparts compared with that of optical counterparts is due to the paucity of the radio source population. Eddington bias most a?ects our counts at faint ?uxes, where the statistical errors on our ?uxes are largest. For XMM-Newton surveys with shorter exposure times, source counts at faint ?uxes will be signi?cantly a?ected by Eddington bias. Eddington bias at a particular ?ux depends on the relative errors on the ?ux, and upon the slope and normalisation of the source counts distribution (Eddington 1913; Teerikorpi 2004). At a ?ux of 10?15 erg cm?2 s?1 the relative di?erence in signal to noise between exposures of 120 and 40 ks is ? 1.7. Given that our source counts are boosted by ? 10% at 3 × 10?15 erg cm?2 s?1 we expect the source counts in a survey with a 40 ks exposure time to be boosted by ? 29% at 3 × 10?15 erg cm?2 s?1 . This is of particular relevance to shallower surveys such as the HELLAS2XMM survey (Baldi et al. 2002), which should be heavily a?ected by Eddington bias at faint ?uxes. 5.3 Contribution to the XRB In this section we examine what fraction of the XRB we can probe with our survey by comparing the integrated emission from our source counts with various measurements of the XRB intensity. To obtain the total 1-2 keV emission from point sources, we integrate our 0.5-2 keV di?erential source counts between ?uxes of 2 × 10?16 erg cm?2 s?1 and 3 × 10?14 erg cm?2 s?1 ; for ?uxes greater than 3 × 10?14 erg cm?2 s?1 we have added the integrated counts obtained from the RIXOS survey (Mason et al. 2000). In the 1-2 keV
c 0000 RAS, MNRAS 000, 000–000

XMM-Newton Deep Field: X-ray source catalogue
5?10-12 4?10-12 3?10-12 2?10-12 1?10-12
Vecchi et al. (1999)

17

Miyaji et al. (LH) (1998)

1?10-11 XRB(>S) (erg cm-2 s-1 deg-2) 8?10-12 6?10-12 4?10-12 2?10-12

De Luca et al. (2004) Lumb et al. (2002) Miyaji et al. (LH) (1998) Kushino et al. (2002) Miyaji et al. (LX) (1998) Gendreau et al. (1995) Marshall et al. (1980)

XRB(>S) (erg cm-2 s-1 deg-2)

Gendreau et al. (1995)

13H integrated emission 95% errors

13H integrated emission 95% errors

0.1 1.0 S1-2 (10-14 erg cm-2 s-1)

0.1

1.0 S2-5 (10-14 erg cm-2 s-1)

Figure 14. Discrete source contribution to the 1-2 keV X-ray background from 0.1 ? 1.0 × 10?13 erg cm?2 s?1. The solid curve shows the integrated emission from the best ?t to our source counts, the dotted curves indicate the lower and upper 95% error on the integrated emission. Measurements of the XRB intensity taken from the literature are overlaid and labelled for comparison. Our source counts account for 47 ? 92% of the XRB intensity in this energy band.

Figure 15. Discrete source contribution to the 2-5 keV X-ray background from 0.1 ? 1.0 × 10?13 erg cm?2 s?1. The solid curve shows the integrated emission from the best ?t to our source counts, the dotted curves indicate the lower and upper 95% error on the integrated emission. Measurements of the XRB intensity taken from the literature are overlaid and labelled for comparison. Our source counts account for 42 ? 92% of the XRB intensity in this energy band.

band our source counts account for 51-100% of the XRB intensity. Our contribution to the XRB as a function of ?ux is illustrated in Fig. 14. The main uncertainty in the resolved fraction lies in the range of reported values for the absolute normalisation of the XRB which are discrepant by up to 30%. The XRB measurements of Gendreau et al. (1995) and Vecchi et al. (1999) are taken to represent the lower and upper normalisations respectively in this energy band. In Fig. 15 we show the contribution of our sources to the integrated 2-5 keV XRB intensity. We integrated our di?erential source counts between ?uxes of 2 × 10?15 erg cm?2 s?1 and 1 × 10?13 erg cm?2 s?1 but this time we added on the integrated counts of Cagnoni, Della Ceca & Maccacaro (1998). With respect to the XRB measurements of Gendreau et al. (1995) and Vecchi et al. (1999), our source counts contribute 50-93% of the XRB intensity in the 2-5 keV band. Our source counts would account for the whole 2-5 keV XRB intensity measured by Marshall et al. (1980). Therefore, the XRB intensity within the 13H ?eld must be higher than that measured by Marshall et al. (1980). Indeed, given that our source counts are rich around S knee , the XRB intensity within the 13H ?eld may be somewhat higher than the global average. We have high quality X-ray spectra to 2-5 keV ?uxes of 3 × 10?15 erg cm?2 s?1 (Page et al. 2003, 2005). At these ?uxes we have already resolved ? 30% of the XRB, and hence the 13H survey is readily suitable for the study of the nature and physical properties of the major XRB producing populations.

6 CONCLUSIONS In this paper we have presented the complete catalogue of 225 sources in the XMM-Newton 13H deep ?eld which covers a sky
c 0000 RAS, MNRAS 000, 000–000

area of 0.18 deg2 centred on position 13h 34m +37? 53’. We reach source densities of 700 deg?2 at a ?ux of 4.1 × 10?16 erg cm?2 s?1 in our lowest energy band. In the 0.5-2 keV band we ?nd source densities of 1300 at a limiting ?ux of 4.5 × 10?16 erg cm?2 s?1 . At harder energies we reach ?uxes a factor ? 10× brighter, with source densities of 900 and 300 deg?2 at limiting ?uxes of 1.1 × 10?15 erg cm?2 s?1 and 5.3 × 10?15 erg cm?2 s?1 in the 2-5 keV and 5-10 keV energy bands respectively. We have carried out extensive simulations of the detection process in order to assess the reliability of our source catalogue. Simulations indicate that confusion is small (< 2%) in our images. We have curtailed the sourcelist directly derived from the SAS according to the optimal statistical detection likelihoods in each band determined from our simulations. We expect < 7 spurious sources to remain in the ?nal catalogue. Within the central 9′ of the XMM-Newton ?eld of view, positional errors are less than 2′′ for 68% of our simulated sources and our input/output ?uxes agree to within 1% at bright ?uxes. Comparison of the input/output source positional o?sets from our simulations with the positional o?sets found between the XMM-Newton and Chandra counterparts suggest that an additional systematic error of 0.76′′ should be added in quadrature to the EMLDETECT positional errors. This poses no problem for radio counterpart identi?cation. We match our catalogue with the radio catalogue of Seymour et al. (2004) and are con?dent in our choice of radio counterpart to a ?ux of 0.1mJy. In the optical we are con?dent in our choice of optical counterparts to a magnitude of R = 23 using our XMM-Newton positions. For magnitudes fainter than this we need Chandra positions which allow reliable identi?cation of optical counterparts to R = 26 (McHardy et al. 2003). We have computed the best-?t parameters for the di?erential N (S ) function using a method adapted from Murdoch et al. (1973).

18

N. S. Loaring et al.
Gendreau et al., 1995, PASJ 47, L5 Georgantopoulos I. et al., 1997, MNRAS 291, 203 Giacconi R. & Zamorani G., 1987, ApJ 313, 20 Giacconi R. et al., 2001, ApJ 551, 624 Giedke K. et al., 2003, AN 324, No. 1-2, 136 Gilli R., Risaliti G. & Salvati M.m 1999, A&A 347, 424 Gilli R., Salvati M. & Hasinger G., 2001, A&A 366, 407 Gilli R., et al., 2005, A&A 430, 811 Giommi P., Perri M. & Fiore F., 2000, A&A 362, 799 Gunn K. F., et al., 2001, MNRAS 324, 305 Hasinger G. et al., 1993, A&A 275, 1 Hasinger G. et al., 1998, A&A 329, 482 Hasinger G. et al., 2001, A&A 365, L45 Ishisaki Y. et al., 2001, PASJ 53, 445 Kirch M., 2004, XMM-SOC-CAL-TN-0018 Lamer G. et al., 2003, AN 324, No. 1-2, 16 Lehmann I. et al., 2001, A&A 371, 833 Loaring N. S., Page M. J. & Ramsay G., 2003, MNRAS 345, 865 Loaring N. S. et al., 2005, in prep. Lumb D.H., Warwick R.S., Page M., De Luca A., 2002, A&A 389, 93 Maccacaro T. et al., 1991, ApJ 374, 117 Madau P, Ghisellini G., & Fabian A.C., 1994, MNRAS 270, L17 Mainieri V. et al., 2002, A&A 393, 425 Maiolino R. et al., 2001a, A&A, 365, 28 Manners J. C. et al., 2003, MNRAS 343, 293 Marshall F.E. et al., 1980, ApJ 235, 4 Mason K. O., et al., 2000, MNRAS 311, 456 Mateos S. et al., (2005), A&A 433, 855 McHardy I.M. et al., 1998, MNRAS 295, 641 McHardy I.M. et al., 2003, MNRAS 342, 802 Murdoch H.S., Crawford D.F & Jauncey D.L., 1997, ApJ 183, 1 Mushotzky R.F., Cowie L.L, Barger A.J. & Arnaud K.A., 2000, Nature 404, 459 Osborne J., 2001, SSC-LUX-TN-0059 (issue 2), http://xmmssc-www.star.le.ac.uk/pubdocs/SSC-LUX-TN-0059.ps.gz Page M.J., Mittaz J. P. D. & Carrera F. J., 2001, MNRAS 325, 757 Page M.J. et al. 2003, AN 324, 101 Page M. J. et al., 2005, in prep. Rosati P. et al., 2002, ApJ 566, 667 Schmidt M. et al., 1998, A&A 329, 495 Schwartz D.A., Murray S.S & Gursky H., 1976, ApJ 204, 315 Setti G. & Woltjer L., 1989, A&A 224, L21 Seymour N., McHardy I. M. & Gunn, K. F., 2004, MNRAS 352, 131 Shanks T. et al., 1991, Nature 353, 315 Snowden S., 2001, http://xmm.gsfc.nasa.gov/docs/xmm/estec2001/ccf Szokoly G. P. et al, 2003, astro-ph/0312324 Teerikorpi P., 2004, A&A 424, 73 Tozzi P., et al., 2001, ApJ, 562, 42 Turner M.J.L. et al., 2001, A&A, 265, L27 Ueda Y. et al., 1998, Nature 391, 866 Ueda Y. et al., 1999, ApJ 518,656 Ueda Y. et al., 1999, ApJ 524, L11 Ueda Y. et al., 2001, ApJS 133, 1 Vecchi et al., 1999, A&A, 349, L73 Voges et al., 1999, A&A 349, 389 Yang Y. et al., 2003, ApJ 585, L85 Yang Y. et al., 2004, AJ 128, 1501 Zamorani G. et al., 1999, A&A 346, 731

In all but our hardest energy band the data are described by a double powerlaw model with a break ?ux at ? 10?14 erg cm?2 s?1 . The counts below the break are ?atter than the Euclidean case. Our measured source counts are in broad agreement with previous determinations from the Chandra deep ?elds (Brandt et al. 2001; Rosati et al. 2002), XMM-Newton (Hasinger et al. 2001; Baldi et al. 2002) and BeppoSAX (Fiore et al. 2001) surveys. The overall normalisation in the 0.5-2 keV band is similar to, though higher than, that found in the CDF-N survey (Brandt et al. 2001). Both the CDF-S (Rosati et al. 2002) and Lockman hole (Hasinger et al. 2001) normalisations are signi?cantly lower than found here. This ?eld-to-?eld variation in source density may be attributed to the underlying clustered large scale structure which imprints a variation in source density on small scales. In the 2-5 keV band our faint end normalisation is consistent with the CDF-N (Brandt et al. 2001), CDF-S (Rosati et al. 2002) and XMM-Newton Lockman hole counts (Hasinger et al. 2001). There is minimal ?ux overlap with the ASCA surveys (Cagnoni, Della Ceca & Maccacaro 1998; Ueda et al. 1999,a). The 2-5 keV band ASCA counts lie above those found in this survey (Cagnoni, Della Ceca & Maccacaro 1998) although they are still consistent with our model ?ts. In our hardest energy band (5-10 keV) our counts are in broad agreement with those from BeppoSAX and ASCA , and with the 5-10 keV counts in the Lockman hole (Hasinger et al. 2001). Again there is little overlap with the brighter ASCA and BeppoSAX surveys. The sources in our survey straddle the break in the source counts. Accounting for the uncertainty in our source counts and the absolute normalisation of the XRB we resolve 51-100(50-93)% of the 1-2 (2-5) keV XRB emission. At the break in the source counts we resolve ? 30% of the 2-5 keV XRB. At these ?uxes we have X-ray spectra and are therefore able to study the emission mechanisms and investigate absorption in a signi?cant fraction of the X-ray source population.

REFERENCES Alexander D.M. et al., 2001, AJ 122, 2156 Alexander D.M. et al., 2003, AJ 126, 539 Akiyama M. et al., 2000, ApJ 532, 700 Antonucci R.R.J. & Miller J.S., 1985, ApJ, 297, 621 Baldi A. et al., 2002, ApJ 564, 190 Barger A.J., Cowie L., Mushotzky R. & Richards E., 2001, AJ, 121, 662 Barger A.J. et al., 2002, AJ 124, 1839 Barger A.J. et al., 2003, AJ 126, 632 Bauer F. E. et al, 2004, AJ 128, 2048 Brandt W.N. et al., 2001, AJ, 122, 2810 Cagnoni I., Della Ceca R. & Maccacaro T., 1998, ApJ 493, 54 Carrera F. J., Page M. J. & Mittaz J. P. D., 2004, A&A 420, 163 Cash W., 1979, ApJ 228, 939 Comastri A., Setti G., Zamorani G. & Hasinger G., 1995, A&A, 296, 1 Comastri A. et al., 2001, MNRAS 327, 781 Condon J. J., 1974, ApJ 188, 279 Cowie L. L. et al., 2002, ApJ, 566, L5 Della Ceca et al., 2004, A&A 428, 383 De Luca A. & Molendi S., 2004, A&A 419, 837 Eddington A. S., 1913, MNRAS 73, 359 Fiore F. et al., 1999, MNRAS 306, L55 Fiore F. et al., 2001, MNRAS 327, 771

c 0000 RAS, MNRAS 000, 000–000

Table 9: 13H XMM-Newton deep ?eld catalogue. All ?uxes are in units of 10?15 erg cm?2 s?1 . The columns labelled ‘Counts’ and ‘Counts error’ give the best ?t number of source counts and the 1σ error as output from EMLDETECT. Flux errors are at 1σ con?dence. If in any energy band there are fewer counts at the source position in the image than in the background map, then the number of source counts and the source ?ux are recorded as zero for that band. In this case the ‘Counts error’ and ’Flux error’ columns represent the 1σ upper limit. The ‘Pos error’ column gives the EMLDETECT positional error (1σ) in arc seconds. The ‘O?axis angle’ column gives the XMM-Newton o?axis angle in arc minutes. The ‘Chandra No.’ column gives the Chandra catalogue number taken from McHardy et al. (2003). The ‘Chandra o?set’ column gives the o?set in position between the XMM-Newton and Chandra source, in arcsec. 0.2-0.5 keV 0.5-2.0 keV Counts Counts Counts error Counts error 20.08 5.87 55.01 9.56 14.12 5.70 22.65 6.82 2.06 3.29 23.89 7.07 0.00 2.20 42.24 8.16 22.60 5.06 49.76 8.22 106.05 15.85 306.23 21.37 81.88 10.87 110.92 13.68 21.87 5.95 48.89 7.40 8.75 6.56 66.97 11.82 11.17 5.97 22.25 7.95 7.54 6.10 38.28 9.47 57.69 10.14 102.27 12.83 186.82 22.76 287.06 28.25 29.36 7.59 107.61 13.18 0.66 3.44 76.20 11.52 48.93 9.69 446.69 23.94 17.98 5.28 45.76 8.11 12.00 6.49 34.76 9.48 353.57 21.39 422.51 23.96 128.52 15.44 114.80 15.21 21.83 7.18 170.94 15.57 33.80 8.58 45.73 10.90 13.48 6.96 50.72 10.11 146.08 14.13 166.80 15.63 19.02 7.54 132.61 14.58 8.66 6.42 25.68 8.03 42.68 8.30 117.78 13.64 35.02 8.06 16.36 7.20 17.87 6.74 33.18 9.28 227.18 16.97 636.62 26.92 20.13 7.65 25.94 9.21 478.40 23.67 1031.89 34.31 13.57 6.81 30.94 8.16 7.77 4.94 17.99 5.00 Continued on next page 2.0-5.0 keV Counts Counts error 13.04 6.11 8.47 5.14 0.00 3.95 52.13 9.52 19.01 6.10 254.86 20.87 21.43 7.56 19.53 6.64 51.59 10.47 0.09 2.86 5.97 7.26 41.02 10.27 76.01 15.52 23.33 7.59 19.45 7.71 150.48 15.39 22.47 7.39 0.71 4.00 103.76 13.33 14.87 8.98 62.32 11.11 0.00 7.31 199.75 16.71 43.17 9.69 45.80 10.49 32.51 9.23 18.24 7.38 12.39 6.80 0.00 5.35 196.33 16.74 0.00 2.37 320.33 20.81 22.51 7.78 18.54 6.00 5.0-10.0 keV 0.2-0.5 keV 0.5-2.0 keV Counts Flux Flux Counts error Flux error Flux error 1.36 3.16 3.39 0.99 6.80 1.18 0.00 5.05 1.93 0.78 2.37 0.71 0.00 2.40 0.25 0.40 2.30 0.68 3.62 5.30 0.00 0.16 2.60 0.50 1.49 3.24 2.61 0.58 5.54 0.92 72.68 20.88 3.05 0.46 8.10 0.57 0.00 3.25 2.30 0.30 2.86 0.35 0.00 6.43 2.23 0.60 4.85 0.73 23.63 10.23 0.24 0.18 1.71 0.30 7.08 6.63 0.93 0.50 1.61 0.58 0.00 5.99 0.22 0.18 1.05 0.26 1.50 4.91 1.53 0.27 2.49 0.31 9.65 9.54 18.28 2.23 24.83 2.44 15.49 8.60 0.83 0.21 2.79 0.34 10.21 8.28 0.02 0.09 1.80 0.27 12.43 8.40 1.25 0.25 10.51 0.56 6.39 7.09 1.52 0.45 3.34 0.59 2.67 5.62 0.30 0.16 0.81 0.22 11.93 8.58 9.03 0.55 9.90 0.56 9.44 8.43 3.25 0.39 2.66 0.35 11.40 7.61 0.69 0.23 4.89 0.45 8.14 8.38 0.80 0.20 1.00 0.24 48.63 10.97 0.32 0.17 1.11 0.22 5.77 6.80 3.56 0.34 3.75 0.35 3.45 5.66 0.48 0.19 3.05 0.34 11.06 7.91 0.20 0.15 0.53 0.17 11.30 8.55 1.15 0.22 2.91 0.34 32.15 9.36 1.64 0.38 0.67 0.30 0.00 5.89 0.39 0.15 0.67 0.19 75.24 12.65 4.87 0.36 12.55 0.53 0.00 2.55 0.45 0.17 0.53 0.19 70.33 12.92 10.84 0.54 21.58 0.72 14.47 8.31 0.44 0.22 0.90 0.24 1.12 3.23 1.06 0.67 1.88 0.52 2.0-5.0 keV Flux Flux error 2.89 1.35 1.65 1.00 0.00 0.72 6.77 1.24 5.36 1.72 15.76 1.29 1.28 0.45 5.02 1.71 3.06 0.62 0.01 0.45 0.38 0.46 2.32 0.58 14.59 2.98 1.41 0.46 1.06 0.42 8.21 0.84 3.53 1.16 0.04 0.21 5.62 0.72 0.80 0.48 4.06 0.72 0.00 0.37 10.05 0.84 2.25 0.50 2.44 0.56 1.56 0.44 1.04 0.42 1.11 0.61 0.00 0.25 8.87 0.76 0.00 0.11 15.50 1.01 1.43 0.50 3.62 1.17 5.0-10.0 keV Flux Flux error 1.78 4.12 0.00 5.44 0.00 2.42 2.32 3.39 1.67 3.63 19.06 5.48 0.00 0.83 0.00 6.50 5.92 2.56 5.31 4.98 0.00 1.63 0.35 1.16 8.78 8.68 3.94 2.19 2.32 1.88 2.81 1.90 4.82 5.36 0.59 1.25 2.69 1.94 2.11 1.88 3.16 2.11 1.66 1.71 9.97 2.25 1.22 1.44 0.75 1.24 2.15 1.53 2.71 2.05 13.07 3.81 0.00 1.09 13.42 2.26 0.00 0.48 13.59 2.50 3.83 2.20 1.21 3.49 0.2-10.0 keV Flux Flux error 14.86 4.60 5.94 5.63 2.55 2.64 11.68 3.65 15.18 4.17 45.97 5.67 6.44 1.05 12.10 6.79 10.93 2.66 7.88 5.06 1.65 1.72 6.69 1.36 66.48 9.75 8.96 2.27 5.20 1.95 22.79 2.17 13.21 5.53 1.74 1.29 27.24 2.21 8.82 2.01 12.80 2.28 3.46 1.78 21.45 2.42 10.78 1.61 6.72 1.41 4.43 1.61 7.81 2.13 16.49 3.89 1.06 1.14 39.71 2.46 0.98 0.56 61.51 2.84 6.60 2.28 7.78 3.78

c 0000 RAS, MNRAS 000, 000–000 No. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 XMM-Newton Ra J(2000) Dec J(2000) 13 33 28.96 37 55 58.82 13 33 30.32 37 56 12.21 13 33 30.68 37 55 05.57 13 33 32.24 37 58 15.92 13 33 34.66 37 49 10.68 13 33 35.69 37 54 05.50 13 33 37.23 37 56 32.06 13 33 37.76 37 47 57.33 13 33 38.80 37 52 13.60 13 33 38.91 38 01 56.15 13 33 40.70 37 49 47.79 13 33 40.93 37 52 41.26 13 33 42.47 38 03 35.94 13 33 42.61 37 50 27.06 13 33 42.83 37 52 39.90 13 33 42.98 37 56 36.73 13 33 43.36 37 45 14.61 13 33 43.79 37 54 55.76 13 33 44.43 37 57 53.49 13 33 45.45 37 58 08.04 13 33 46.61 38 00 21.56 13 33 47.52 37 53 51.50 13 33 48.28 37 53 33.71 13 33 48.60 37 58 07.79 13 33 50.68 37 49 46.61 13 33 50.80 37 57 16.61 13 33 53.59 38 02 03.64 13 33 54.21 38 02 53.17 13 33 54.91 37 51 25.87 13 33 55.88 37 52 57.85 13 33 57.90 37 49 59.01 13 33 58.56 37 59 37.55 13 33 58.87 38 00 24.88 13 33 59.21 38 05 58.32 Posa O?axisb error angle 1.57 12.91 2.38 12.67 3.04 12.53 1.33 12.67 1.27 13.08 1.55 11.57 0.84 11.35 1.08 13.14 1.20 11.26 4.82 12.96 1.99 11.74 0.85 10.75 0.61 13.39 0.76 11.12 1.18 10.39 0.44 10.24 1.26 13.95 2.56 9.94 0.38 10.26 0.92 10.14 0.74 10.85 1.11 9.27 0.64 9.16 0.57 9.55 0.85 10.01 1.77 8.88 0.77 10.73 1.23 11.21 1.81 8.50 0.28 7.81 2.91 8.70 0.23 8.46 1.17 8.87 1.37 13.03

Chandra No.c o?setd

47 49 62 105 73 51

1.29 2.70 5.23 2.14 2.99 4.18

XMM-Newton Deep Field: X-ray source catalogue

143 2 79 113 27 46 29 78 76 40 128 72 175 17 7 111

2.26 1.37 1.56 1.27 1.61 2.77 1.95 0.76 0.66 1.25 1.36 1.29 3.41 1.04 0.75 3.20

19

20

No. 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80

0.2-0.5 keV 0.5-2.0 keV XMM-Newton Posa O?axisb Counts Counts Ra J(2000) Dec J(2000) error angle Counts error Counts error 13 33 59.56 37 51 39.75 2.34 7.57 5.28 6.21 36.92 10.18 13 33 59.63 37 49 26.35 12.19 8.75 0.00 4.59 90.42 19.99 13 34 00.04 37 49 12.27 0.32 8.84 373.03 20.83 712.58 28.63 13 34 00.91 38 01 25.24 0.95 9.24 5.46 5.57 88.41 12.16 13 34 01.05 37 54 03.99 0.35 6.60 244.79 17.97 549.22 25.28 13 34 01.06 38 00 25.75 1.50 8.55 6.96 6.07 19.52 7.67 13 34 01.26 37 46 47.56 0.90 10.41 8.67 6.83 147.83 15.54 13 34 02.62 37 51 29.14 0.36 7.12 144.04 14.09 419.12 22.19 13 34 03.09 37 53 21.42 0.50 6.34 2.26 4.13 95.74 12.95 13 34 07.18 37 59 59.34 1.23 7.36 18.32 10.34 0.00 3.70 13 34 08.05 38 06 26.62 0.92 12.62 33.53 7.81 94.69 12.52 13 34 08.36 37 52 19.21 0.50 5.72 115.74 12.85 256.77 18.53 13 34 08.39 37 47 50.85 2.69 8.72 5.54 6.28 15.49 8.33 13 34 08.59 37 54 40.76 0.53 5.06 94.30 12.27 234.91 18.41 13 34 08.72 38 03 49.23 0.47 10.22 112.09 12.97 331.70 21.03 13 34 08.79 37 57 06.43 0.27 5.47 249.06 17.77 860.91 31.47 13 34 09.96 37 54 31.62 0.78 4.80 0.82 3.39 23.43 9.06 13 34 10.61 37 59 56.03 0.31 6.84 641.48 29.15 351.57 22.04 13 34 11.42 37 47 57.32 0.51 8.29 172.60 15.68 234.29 18.23 13 34 12.36 37 59 10.36 2.01 6.06 28.74 11.66 0.00 3.54 13 34 13.03 37 58 31.63 0.77 5.52 81.50 11.53 147.43 14.16 13 34 13.59 37 45 39.06 1.22 10.12 32.55 8.52 65.93 10.90 13 34 13.62 38 07 23.60 0.89 13.12 23.57 7.54 195.39 17.42 13 34 14.15 38 04 41.20 1.00 10.54 9.96 6.36 75.06 12.90 13 34 14.37 37 52 32.09 1.43 4.59 0.00 1.14 37.94 8.56 13 34 14.79 37 59 59.67 0.85 6.36 50.74 10.19 130.43 14.39 13 34 14.81 37 51 30.81 0.38 5.12 206.64 15.92 353.49 21.06 13 34 15.49 38 03 05.78 1.80 8.97 11.63 6.18 13.12 7.94 13 34 15.54 37 52 27.80 1.31 4.43 0.00 1.52 43.88 14.26 13 34 17.01 37 59 48.75 0.96 5.95 48.10 9.74 87.75 12.41 13 34 17.51 37 57 21.91 0.14 4.10 1590.56 41.23 2266.48 49.01 13 34 17.80 37 44 31.39 2.04 10.89 3.97 5.25 23.78 9.25 13 34 18.87 37 58 56.69 0.32 5.04 18.29 6.88 403.89 22.27 13 34 19.24 37 43 49.99 1.55 11.47 0.00 3.55 37.29 9.86 13 34 19.26 37 50 29.60 0.52 5.32 105.10 13.16 217.42 17.46 13 34 19.56 37 51 47.77 1.40 4.26 0.00 5.64 22.75 8.79 13 34 19.97 37 54 00.29 0.61 2.96 23.51 7.43 170.55 15.77 13 34 20.80 37 59 30.74 0.96 5.30 11.54 6.63 25.22 8.81 13 34 20.82 37 54 59.88 0.50 2.64 99.21 12.17 201.23 16.48 13 34 21.90 38 04 50.89 1.34 10.22 13.39 6.83 70.31 10.93 13 34 22.07 37 53 46.31 1.73 2.66 4.71 5.74 38.37 9.96 13 34 22.11 37 48 03.61 1.88 7.26 6.37 6.82 10.70 7.81 13 34 22.15 38 06 20.07 0.67 11.66 74.21 10.55 187.68 15.93 13 34 22.34 38 04 13.59 1.27 9.60 12.29 6.85 102.82 13.66 13 34 22.70 37 55 23.49 2.14 2.32 10.60 7.06 33.05 9.38 13 34 24.21 37 42 58.85 1.25 12.10 17.52 6.20 68.87 10.98 Continued on next page

2.0-5.0 keV 5.0-10.0 keV 0.2-0.5 keV 0.5-2.0 keV Counts Counts Flux Flux Counts error Counts error Flux error Flux error 9.90 7.53 5.15 6.68 0.11 0.12 0.68 0.19 0.00 5.21 0.00 7.12 0.00 0.10 1.86 0.41 131.11 14.00 3.02 5.26 8.45 0.47 14.86 0.60 44.95 10.00 8.60 7.99 0.13 0.13 1.92 0.26 261.35 19.83 76.08 13.84 4.95 0.36 10.35 0.48 58.74 10.85 11.63 7.96 0.17 0.15 0.43 0.17 59.26 11.84 14.80 8.84 0.23 0.18 3.55 0.37 99.91 13.25 21.77 9.24 2.80 0.27 7.52 0.40 287.86 19.70 90.74 13.78 0.04 0.08 1.62 0.22 45.94 14.61 15.40 10.88 0.37 0.21 0.00 0.07 6.79 7.46 33.96 10.12 1.29 0.30 3.29 0.43 73.97 11.86 28.75 9.60 2.01 0.22 4.12 0.30 34.65 10.02 13.50 8.48 0.12 0.14 0.32 0.17 118.64 15.53 13.93 8.44 1.56 0.20 3.58 0.28 43.75 10.19 12.05 9.31 2.94 0.34 7.99 0.51 287.71 20.38 88.02 13.60 4.29 0.31 13.63 0.50 128.60 14.47 47.77 11.24 0.01 0.06 0.35 0.14 94.69 12.23 15.53 9.25 17.53 0.80 8.66 0.54 62.26 10.90 20.36 9.54 3.72 0.34 4.66 0.36 0.00 5.63 3.76 5.90 0.53 0.22 0.00 0.06 43.77 9.56 10.94 8.25 1.85 0.26 3.02 0.29 0.00 2.05 0.00 5.19 0.84 0.22 1.57 0.26 69.29 11.80 13.87 9.29 0.83 0.26 6.28 0.56 95.57 13.20 29.41 9.64 0.27 0.17 1.86 0.32 19.46 8.35 0.00 6.75 0.00 0.03 0.84 0.19 22.30 8.86 17.50 8.82 0.97 0.19 2.29 0.25 0.00 2.77 0.00 3.77 3.48 0.27 5.49 0.33 32.35 9.58 13.20 8.16 0.27 0.14 0.28 0.17 65.88 11.49 19.04 9.60 0.00 0.03 0.68 0.22 22.14 8.52 1.93 5.23 0.87 0.18 1.46 0.21 681.39 28.69 136.27 15.72 25.25 0.65 33.02 0.71 43.92 10.78 0.00 7.79 0.12 0.15 0.64 0.25 415.62 22.78 110.89 13.89 0.31 0.12 6.34 0.35 32.42 9.24 17.94 9.22 0.00 0.11 1.02 0.27 67.50 11.43 19.94 9.27 1.87 0.23 3.56 0.29 43.62 10.11 14.49 8.37 0.00 0.09 0.33 0.13 63.23 10.82 27.08 8.46 0.41 0.13 2.70 0.25 91.13 12.72 24.21 8.82 0.20 0.12 0.41 0.14 80.58 11.46 11.53 7.36 1.59 0.20 2.96 0.24 17.24 8.12 0.00 4.95 0.36 0.18 1.71 0.27 16.37 8.21 7.01 7.76 0.07 0.08 0.52 0.13 51.41 10.64 29.54 10.23 0.13 0.14 0.20 0.14 51.17 10.51 5.49 6.60 2.25 0.32 5.22 0.44 69.09 12.06 26.96 9.60 0.31 0.18 2.41 0.32 4.11 5.38 0.00 3.76 0.20 0.13 0.55 0.16 33.07 9.14 6.30 7.69 0.56 0.20 2.02 0.32

2.0-5.0 keV Flux Flux error 0.42 0.32 0.00 0.25 6.31 0.67 2.25 0.50 11.50 0.87 2.94 0.54 3.31 0.66 4.12 0.55 11.20 0.77 1.93 0.62 0.54 0.59 2.71 0.43 1.66 0.48 4.10 0.54 2.44 0.57 10.37 0.73 4.38 0.49 5.18 0.67 2.85 0.50 0.00 0.22 1.97 0.43 0.00 0.11 5.18 0.88 5.46 0.75 0.93 0.40 0.89 0.35 0.00 0.10 1.60 0.47 2.32 0.40 0.84 0.32 22.45 0.95 2.76 0.68 14.87 0.81 2.08 0.59 2.51 0.43 1.45 0.34 2.23 0.38 3.38 0.47 2.66 0.38 0.97 0.46 0.50 0.25 2.17 0.45 3.30 0.68 3.72 0.65 0.15 0.20 2.27 0.63

5.0-10.0 keV Flux Flux error 0.86 1.11 0.00 1.35 0.58 1.01 1.75 1.63 12.64 2.30 2.38 1.63 3.40 2.03 3.47 1.48 13.50 2.05 2.57 1.81 12.00 3.58 4.00 1.34 2.56 1.61 1.83 1.11 2.80 2.16 12.13 1.87 6.15 1.45 3.43 2.04 3.68 1.73 0.56 0.88 1.94 1.47 0.00 1.18 4.59 3.08 7.09 2.32 0.00 1.27 2.71 1.37 0.00 0.50 2.66 1.64 2.53 1.28 0.28 0.77 16.96 1.96 0.00 2.03 15.03 1.88 4.84 2.49 2.81 1.31 1.80 1.04 3.60 1.12 3.41 1.24 1.43 0.91 0.00 1.17 0.79 0.88 4.84 1.68 1.53 1.84 6.02 2.14 0.00 0.53 1.84 2.25

0.2-10.0 keV Flux Flux error 2.06 1.18 1.86 1.44 30.21 1.44 6.06 1.73 39.45 2.53 5.92 1.73 10.48 2.18 17.92 1.65 26.36 2.20 4.87 1.93 17.12 3.67 12.84 1.45 4.67 1.69 11.08 1.28 16.16 2.32 40.42 2.10 10.89 1.53 34.80 2.35 14.92 1.87 1.10 0.94 8.79 1.58 2.42 1.23 16.89 3.26 14.67 2.47 1.77 1.34 6.86 1.45 8.98 0.67 4.81 1.72 5.54 1.36 3.45 0.87 97.67 2.38 3.51 2.16 36.56 2.08 7.94 2.57 10.75 1.42 3.58 1.10 8.94 1.22 7.40 1.34 8.64 1.03 3.03 1.29 1.88 0.93 7.33 1.75 12.32 2.04 12.47 2.27 0.90 0.60 6.69 2.37

Chandra No.c o?setd

N. S. Loaring et al.

4 84 23 48 43 52 60 68 58 25 28 141 9 66 85 88 132 119 120 81

1.49 0.51 1.00 1.05 0.88 0.58 1.39 1.08 1.38 1.01 0.45 0.82 0.46 1.76 0.71 1.66 1.35 1.03 1.20 0.12

170 5 63 137 116 93 191 59 133 196 214 41 117 110

0.72 0.64 0.76 1.79 1.40 0.66 1.72 1.30 0.61 0.99 1.05 1.40 9.13 2.52

c 0000 RAS, MNRAS 000, 000–000

c 0000 RAS, MNRAS 000, 000–000

No. 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126

XMM-Newton Ra J(2000) Dec J(2000) 13 34 24.64 37 46 14.71 13 34 25.91 37 54 59.89 13 34 26.87 38 00 27.92 13 34 27.39 37 50 08.65 13 34 28.44 37 47 08.21 13 34 28.64 37 41 27.94 13 34 28.66 37 57 48.26 13 34 28.80 37 53 37.96 13 34 29.20 38 02 45.33 13 34 29.30 38 06 51.45 13 34 29.77 37 49 20.20 13 34 29.93 37 56 38.98 13 34 30.32 37 55 25.03 13 34 30.42 37 57 02.31 13 34 31.27 38 03 10.23 13 34 31.32 37 49 53.09 13 34 31.41 37 48 31.24 13 34 33.10 37 55 10.22 13 34 33.51 38 05 41.29 13 34 33.55 37 48 35.76 13 34 33.93 38 00 43.04 13 34 34.02 37 49 46.74 13 34 34.76 37 56 46.68 13 34 34.92 38 07 03.03 13 34 35.17 37 49 11.83 13 34 35.27 37 53 55.19 13 34 35.86 37 54 18.96 13 34 36.21 37 51 06.68 13 34 36.36 37 55 57.15 13 34 36.38 38 05 14.02 13 34 36.74 38 03 19.95 13 34 37.20 37 54 38.28 13 34 37.91 37 56 04.43 13 34 38.26 38 01 39.25 13 34 38.41 38 06 26.50 13 34 38.83 37 40 22.03 13 34 38.96 37 42 43.95 13 34 39.75 37 57 01.93 13 34 41.79 38 00 11.23 13 34 42.29 37 41 46.46 13 34 42.70 37 59 14.85 13 34 42.81 37 42 42.62 13 34 42.84 38 03 53.27 13 34 43.06 37 52 06.16 13 34 43.42 37 49 22.13 13 34 43.77 38 02 46.06

0.2-0.5 keV 0.5-2.0 keV Posa O?axisb Counts Counts error angle Counts error Counts error 0.41 8.88 178.04 15.30 348.71 20.78 1.58 1.64 14.36 7.19 43.44 10.29 1.45 5.73 7.55 6.69 48.90 9.81 0.63 4.96 11.89 6.93 184.42 16.61 1.37 7.87 34.54 8.19 42.64 9.80 1.72 13.50 27.59 7.21 59.37 10.79 1.92 3.09 0.12 3.32 19.04 8.81 1.78 1.67 20.62 8.34 34.32 9.32 1.79 7.90 31.01 7.09 5.60 5.78 1.76 11.98 9.93 6.62 62.33 10.73 1.31 5.65 4.60 5.16 49.09 10.08 0.56 1.93 9.55 6.82 124.70 14.63 0.76 0.92 64.17 10.97 103.10 14.12 1.43 2.25 1.45 3.95 57.38 11.71 0.82 8.27 11.49 6.28 122.99 14.41 0.50 5.07 8.57 5.74 178.97 15.71 0.19 6.42 551.25 24.99 1311.70 38.02 1.01 0.33 0.72 3.85 14.98 9.14 1.60 10.77 1.80 3.50 52.34 9.89 1.75 6.32 13.06 9.23 9.47 10.42 0.45 5.80 125.34 13.03 236.45 17.79 2.30 5.14 0.00 1.14 31.58 8.86 1.09 1.86 73.93 14.89 418.91 27.83 1.99 12.13 13.92 7.24 53.49 10.86 2.08 5.73 9.81 6.56 39.07 9.92 1.23 1.02 20.65 8.42 56.19 11.82 0.55 0.68 27.31 9.48 214.46 17.79 0.90 3.83 25.90 7.38 89.84 12.65 0.27 1.12 0.00 4.40 233.04 24.08 1.25 10.32 23.58 7.73 73.39 11.98 0.99 8.43 55.01 9.47 85.94 12.54 0.70 0.65 29.01 8.65 35.71 10.52 0.15 1.37 1373.41 39.50 2286.52 51.57 1.05 6.78 0.00 1.38 66.80 11.81 0.48 11.55 122.44 12.95 424.83 24.07 1.11 14.58 38.49 13.87 141.83 24.35 1.60 12.22 14.48 7.18 37.60 9.95 1.29 2.38 19.69 8.23 61.75 12.50 0.24 5.48 316.89 19.31 860.60 30.88 0.78 13.24 79.19 10.54 209.90 16.62 0.21 4.64 1034.40 34.10 1001.37 33.56 1.55 12.33 5.01 5.66 51.09 10.20 2.33 9.13 5.88 6.21 27.49 7.95 1.45 3.31 19.29 7.38 53.23 10.59 0.79 5.84 6.73 6.26 93.20 12.93 1.42 8.07 0.00 5.04 14.60 7.79 Continued on next page

2.0-5.0 keV 5.0-10.0 keV Counts Counts Counts error Counts error 96.94 13.23 37.19 10.82 19.36 9.08 2.68 5.84 0.00 3.17 12.04 7.98 81.70 11.58 33.24 9.46 3.53 5.41 11.43 7.83 3.99 5.16 0.03 3.12 36.28 9.61 6.56 7.78 1.57 4.65 12.52 9.51 1.37 4.29 0.00 3.63 22.52 9.25 12.33 8.47 17.20 7.34 0.00 2.67 191.02 16.82 71.70 13.02 47.54 10.59 7.60 8.25 19.03 8.38 3.74 5.72 106.97 13.46 14.82 7.77 116.16 13.15 27.23 8.66 557.43 25.96 103.11 13.80 82.72 13.16 55.10 12.07 17.55 8.14 0.00 6.97 64.98 15.75 35.18 12.53 136.00 14.90 27.63 9.60 14.42 7.93 0.00 4.35 137.13 18.61 31.78 13.65 21.49 8.42 4.41 6.37 5.93 6.63 0.00 2.83 22.31 8.87 0.53 4.23 145.24 16.18 33.91 10.90 40.47 9.96 13.00 8.33 598.14 31.60 192.44 19.25 39.95 10.10 0.00 4.87 18.92 8.57 6.36 7.34 150.02 16.27 61.36 12.36 954.92 36.56 229.08 20.82 71.99 11.68 9.09 8.33 105.52 13.89 43.65 11.47 61.06 17.94 23.66 15.16 20.13 8.29 7.73 8.14 19.82 8.07 9.82 8.49 318.61 20.12 78.99 12.44 9.62 7.69 0.29 4.48 157.46 15.57 19.43 8.95 10.34 7.10 21.93 10.11 0.00 2.95 4.05 5.93 7.72 7.45 4.87 6.63 64.59 11.33 45.14 10.81 53.19 10.09 7.21 7.78

0.2-0.5 keV Flux Flux error 4.30 0.37 0.20 0.10 0.13 0.12 0.20 0.12 0.74 0.17 1.42 0.37 0.00 0.05 0.29 0.12 0.71 0.16 0.32 0.21 0.10 0.11 0.17 0.12 0.89 0.15 0.02 0.06 0.28 0.15 0.16 0.11 11.42 0.52 0.01 0.05 0.05 0.10 0.27 0.19 2.28 0.24 0.00 0.02 1.08 0.22 0.65 0.34 0.18 0.12 0.29 0.12 0.38 0.13 0.61 0.17 0.00 0.07 0.70 0.23 1.27 0.22 0.40 0.12 21.18 0.61 0.00 0.03 3.77 0.40 2.34 0.84 0.48 0.24 0.30 0.12 5.97 0.36 2.89 0.39 17.57 0.58 0.17 0.19 0.15 0.15 0.30 0.11 0.12 0.11 0.00 0.12

0.5-2.0 keV Flux Flux error 7.75 0.46 0.57 0.13 0.80 0.16 2.88 0.26 0.84 0.19 2.98 0.54 0.27 0.12 0.45 0.12 0.12 0.12 1.81 0.31 0.95 0.19 2.05 0.24 1.32 0.18 0.84 0.17 2.84 0.33 3.10 0.27 25.23 0.73 0.19 0.12 1.37 0.26 0.18 0.20 3.94 0.30 0.52 0.15 5.62 0.37 2.23 0.45 0.64 0.16 0.72 0.15 2.74 0.23 1.90 0.27 3.31 0.34 1.98 0.32 1.82 0.27 0.46 0.13 32.29 0.73 1.25 0.22 12.00 0.68 8.41 1.44 1.16 0.31 0.85 0.17 14.87 0.53 7.08 0.56 15.59 0.52 1.57 0.31 0.63 0.18 0.76 0.15 1.55 0.22 0.31 0.16

2.0-5.0 keV Flux Flux error 4.98 0.68 0.57 0.27 0.00 0.12 2.91 0.41 0.16 0.24 0.52 0.67 1.16 0.31 0.05 0.14 0.07 0.21 1.52 0.63 0.75 0.32 6.92 0.61 1.37 0.31 0.62 0.27 5.85 0.74 4.55 0.51 24.84 1.16 2.38 0.38 1.06 0.49 2.82 0.68 5.14 0.56 0.54 0.30 4.15 0.56 2.00 0.78 0.22 0.25 0.65 0.26 4.18 0.47 1.86 0.46 18.99 1.00 2.49 0.63 0.92 0.42 4.33 0.47 30.28 1.16 3.06 0.50 6.93 0.91 9.45 2.78 1.45 0.60 0.62 0.25 12.51 0.79 0.77 0.61 5.55 0.55 0.75 0.51 0.00 0.15 0.25 0.24 2.46 0.43 2.59 0.49

5.0-10.0 keV Flux Flux error 7.64 2.22 0.29 0.64 1.73 1.14 4.44 1.26 2.02 1.39 0.02 1.59 0.79 0.93 1.36 1.04 0.00 0.70 3.63 2.50 0.00 0.45 9.91 1.80 0.81 0.88 0.46 0.70 3.15 1.65 4.06 1.29 17.41 2.33 5.85 1.28 0.00 1.81 5.87 2.09 4.04 1.40 0.00 0.61 3.57 1.53 1.83 2.64 0.00 0.40 0.06 0.45 3.60 1.16 2.32 1.49 22.86 2.29 0.00 1.29 1.26 1.45 6.54 1.32 27.05 2.46 1.52 1.40 12.39 3.26 14.58 9.34 2.40 2.52 1.14 0.99 11.94 1.88 0.10 1.55 2.61 1.20 6.76 3.12 0.87 1.28 0.59 0.80 6.53 1.56 1.41 1.52

0.2-10.0 keV Flux Flux error 24.67 2.40 1.63 0.71 2.66 1.17 10.43 1.36 3.76 1.43 4.93 1.85 2.21 0.99 2.15 1.06 0.89 0.76 7.29 2.60 1.79 0.59 19.05 1.92 4.39 0.96 1.94 0.77 12.12 1.84 11.87 1.42 78.90 2.75 8.44 1.34 2.49 1.89 9.14 2.22 15.40 1.56 1.06 0.70 14.43 1.69 6.71 2.81 1.04 0.52 1.71 0.55 10.90 1.28 6.70 1.59 45.16 2.52 5.16 1.49 5.27 1.55 11.73 1.41 110.80 2.88 5.84 1.50 35.09 3.47 34.78 9.88 5.49 2.62 2.91 1.04 45.28 2.14 10.84 1.80 41.31 1.53 9.25 3.18 1.65 1.31 1.89 0.85 10.66 1.64 4.31 1.61

Chandra No.c o?setd 38 0.71

71 122 104

1.59 1.83 1.22

168 80 179 174 69 55 3 180 162 144 45 209 108 161 212 94 106 26 97 95 8 87 33

1.10 1.54 0.26 1.73 0.58 1.10 0.78 1.51 1.53 0.49 1.40 0.97 7.54 0.96 1.15 0.34 0.31 0.51 1.21 1.79 2.09 1.13 1.41

XMM-Newton Deep Field: X-ray source catalogue

10 18

0.55 0.98

173 127 157

2.40 0.49 1.96

21

22

No. 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172

XMM-Newton Ra J(2000) Dec J(2000) 13 34 44.15 37 44 35.27 13 34 45.24 38 00 31.75 13 34 45.28 37 57 22.73 13 34 46.31 37 54 42.14 13 34 46.46 37 58 41.64 13 34 47.02 37 47 48.76 13 34 47.18 37 57 14.88 13 34 47.34 37 59 50.90 13 34 48.20 37 51 11.74 13 34 48.22 37 54 14.33 13 34 49.76 37 54 50.58 13 34 50.53 38 07 05.73 13 34 51.27 37 40 51.30 13 34 51.48 37 46 19.83 13 34 51.86 38 06 35.33 13 34 52.02 37 58 25.78 13 34 52.12 37 57 45.06 13 34 53.13 37 48 18.85 13 34 53.79 38 07 54.86 13 34 53.83 37 51 09.02 13 34 54.55 37 42 07.32 13 34 54.83 37 52 40.44 13 34 55.11 37 49 51.41 13 34 55.41 37 50 40.53 13 34 56.39 37 39 54.74 13 34 56.63 37 53 50.42 13 34 56.71 37 50 25.09 13 34 56.76 37 56 03.31 13 34 56.87 37 52 48.73 13 34 56.98 37 45 53.46 13 34 57.12 37 55 41.16 13 34 57.56 37 49 43.19 13 34 58.11 37 57 33.15 13 34 58.14 37 50 52.22 13 34 58.42 38 04 30.68 13 34 58.85 37 50 23.08 13 34 59.48 37 57 39.65 13 35 00.07 37 53 45.12 13 35 00.10 37 56 33.31 13 35 01.23 37 59 37.81 13 35 02.01 37 47 24.90 13 35 02.72 37 56 20.97 13 35 02.88 37 49 57.55 13 35 03.75 37 45 15.28 13 35 03.77 37 44 11.19 13 35 05.13 37 45 27.22

0.2-0.5 keV 0.5-2.0 keV Posa O?axisb Counts Counts error angle Counts error Counts error 1.80 10.52 2.44 4.28 7.48 7.26 0.86 6.02 31.85 8.62 104.55 13.64 0.46 3.29 184.76 15.45 294.62 20.08 0.48 2.40 45.55 9.43 245.13 17.72 1.21 4.48 31.91 9.09 55.76 11.30 0.34 7.54 57.55 10.32 607.27 27.42 1.53 3.46 10.32 6.95 9.74 8.53 0.29 5.57 104.57 12.72 699.41 29.71 1.50 4.63 11.20 6.74 56.93 10.94 1.06 2.84 25.88 6.79 29.11 7.96 0.98 3.07 2.48 4.11 79.80 11.34 0.90 12.59 74.42 10.91 103.12 12.90 1.23 14.46 13.36 6.99 108.01 12.87 0.04 9.24 9433.50 98.55 28488.97 164.27 1.78 12.18 16.13 7.57 33.62 9.52 0.38 4.96 0.00 2.89 261.59 17.90 0.22 4.53 407.05 21.87 816.21 30.82 1.75 7.59 0.00 1.02 10.11 7.54 1.23 13.55 59.90 9.57 88.33 12.52 1.42 5.40 0.48 3.49 53.31 11.50 1.30 13.41 29.77 8.51 59.82 10.07 1.23 4.64 2.55 4.50 27.13 9.02 2.44 6.53 8.49 6.07 32.70 9.91 2.66 5.96 9.44 6.60 33.98 10.34 2.06 15.63 51.51 21.81 130.31 32.80 1.14 4.55 30.18 7.85 63.28 11.12 1.74 6.32 14.92 7.63 51.12 11.65 1.47 4.59 29.34 7.50 12.78 8.16 2.34 4.94 1.62 4.45 35.09 9.45 2.74 10.09 2.53 4.45 50.36 11.77 1.42 4.58 0.00 4.34 44.63 10.13 1.22 6.95 9.85 6.12 58.78 11.16 1.22 5.40 25.18 6.58 21.53 7.26 2.32 6.22 10.82 6.68 45.71 11.15 0.95 10.71 49.06 9.45 91.85 12.68 1.85 6.65 71.22 14.35 258.27 23.32 1.17 5.69 28.90 7.93 18.30 7.58 1.18 5.23 36.13 8.33 35.52 9.32 0.53 5.36 21.22 6.48 184.76 16.27 0.83 7.11 31.51 7.42 79.89 11.09 2.63 9.30 0.00 4.96 18.63 8.87 1.95 5.80 1.81 3.75 38.58 9.09 0.35 7.52 41.77 10.20 540.78 27.19 1.09 11.29 13.49 6.89 75.14 10.95 0.84 12.22 12.43 6.76 132.11 14.43 1.90 11.26 27.69 7.83 0.11 3.89 Continued on next page

2.0-5.0 keV 5.0-10.0 keV 0.2-0.5 keV 0.5-2.0 keV Counts Counts Flux Flux Counts error Counts error Flux error Flux error 38.12 9.26 6.88 6.78 0.07 0.13 0.20 0.20 54.81 11.17 10.98 8.47 0.62 0.17 1.87 0.24 99.51 13.16 25.56 9.89 2.96 0.25 4.35 0.30 115.17 13.50 24.07 8.11 0.76 0.16 3.74 0.27 21.99 8.45 1.76 4.69 0.58 0.16 0.92 0.19 345.04 21.18 85.21 13.94 1.48 0.26 14.85 0.67 58.46 11.55 11.18 8.99 0.17 0.11 0.15 0.13 409.45 23.75 83.56 13.61 2.28 0.28 14.03 0.60 39.71 10.05 0.00 2.42 0.19 0.11 0.88 0.17 28.00 7.85 30.38 8.93 0.62 0.16 0.62 0.17 49.13 10.04 0.00 7.28 0.05 0.08 1.38 0.20 51.01 10.36 9.42 8.59 2.69 0.39 3.41 0.43 71.85 11.81 13.41 8.99 0.57 0.30 4.27 0.51 1707.69 43.33 111.49 14.22 243.71 2.55 676.67 3.90 36.35 9.84 0.00 3.84 0.54 0.25 1.03 0.29 329.33 20.63 81.39 12.13 0.00 0.05 4.27 0.29 311.10 20.75 73.48 11.85 6.95 0.37 12.78 0.48 31.08 8.49 9.88 7.21 0.00 0.02 0.21 0.16 34.72 10.61 12.21 8.88 2.31 0.37 3.12 0.44 46.12 10.42 19.15 9.12 0.01 0.06 0.87 0.19 15.61 8.39 22.16 9.52 1.12 0.32 2.08 0.35 44.17 10.17 40.10 11.12 0.04 0.08 0.43 0.14 6.55 6.38 10.14 8.04 0.18 0.13 0.64 0.19 11.04 7.37 0.00 4.53 0.19 0.13 0.64 0.20 1.88 11.49 0.00 14.54 4.25 1.80 10.50 2.64 19.67 8.15 0.95 4.21 0.52 0.14 1.01 0.18 5.39 6.11 19.05 8.72 0.33 0.17 1.03 0.23 16.19 9.10 7.95 8.75 0.66 0.17 0.26 0.17 0.00 3.90 0.00 1.25 0.03 0.08 0.60 0.16 7.86 7.45 3.09 5.69 0.07 0.12 1.27 0.30 45.78 9.84 13.29 8.97 0.00 0.08 0.77 0.17 40.17 9.38 24.01 8.84 0.21 0.13 1.15 0.22 3.60 5.02 8.19 7.18 0.74 0.19 0.56 0.19 0.00 2.71 0.00 3.56 0.24 0.15 0.92 0.22 23.79 8.49 0.00 4.64 1.44 0.28 2.46 0.34 63.18 15.54 0.00 10.37 1.46 0.30 4.87 0.44 4.26 5.61 24.24 8.64 0.85 0.23 0.48 0.20 23.31 8.27 8.62 8.07 0.65 0.15 0.59 0.15 159.40 15.48 31.60 10.03 0.39 0.12 3.09 0.27 26.91 8.51 0.00 4.49 0.68 0.16 1.57 0.22 29.65 9.15 1.52 4.48 0.00 0.13 0.43 0.21 0.00 5.24 3.53 5.79 0.03 0.07 0.67 0.16 319.92 20.93 99.36 13.79 0.90 0.22 10.71 0.54 28.02 8.47 8.50 8.12 0.41 0.21 2.12 0.31 48.83 10.98 16.06 8.80 0.43 0.23 4.19 0.46 2.06 4.42 9.18 8.81 0.84 0.24 0.00 0.11

2.0-5.0 keV Flux Flux error 2.38 0.58 2.24 0.46 3.33 0.44 3.92 0.46 0.82 0.31 20.44 1.25 1.98 0.39 18.79 1.09 1.39 0.35 1.28 0.36 1.88 0.39 3.91 0.79 6.78 1.11 93.51 2.37 2.59 0.70 12.18 0.76 11.03 0.74 1.50 0.41 2.87 0.88 1.72 0.39 1.28 0.69 1.57 0.36 0.29 0.29 0.49 0.33 0.40 2.43 0.72 0.30 0.25 0.28 0.73 0.41 0.00 0.15 0.46 0.44 1.80 0.39 1.79 0.42 0.20 0.28 0.00 0.12 1.48 0.53 2.71 0.67 0.24 0.32 0.88 0.31 6.05 0.59 1.21 0.38 1.60 0.49 0.00 0.21 14.51 0.95 1.85 0.56 3.65 0.82 0.14 0.29

5.0-10.0 keV 0.2-10.0 keV Flux Flux Chandra Flux error Flux error No.c o?setd 1.78 1.76 4.44 1.86 130 1.37 1.74 1.34 6.47 1.45 142 1.97 3.20 1.24 13.84 1.37 53 0.89 3.08 1.04 11.51 1.18 65 0.82 0.25 0.67 2.56 0.78 167 2.46 18.56 3.04 55.33 3.36 11 0.88 1.42 1.14 3.71 1.22 14.70 2.39 49.79 2.71 13 1.14 0.00 0.32 2.46 0.52 153 1.65 5.41 1.59 7.93 1.65 0.00 1.06 3.31 1.15 172 1.34 3.17 2.89 13.18 3.05 92 2.49 5.52 3.70 17.14 3.91 24.68 3.15 1038.57 6.10 1 0.94 0.00 1.20 4.16 1.44 11.53 1.72 27.99 1.90 64 0.34 9.91 1.60 40.67 1.86 22 0.70 1.88 1.37 3.59 1.44 4.52 3.29 12.82 3.45 109 3.69 2.71 1.29 5.31 1.36 165 1.32 7.95 3.41 12.44 3.51 5.39 1.50 7.44 1.55 199 2.53 1.76 1.39 2.87 1.44 0.00 0.74 1.32 0.84 0.00 12.53 15.15 13.15 0.13 0.58 2.38 0.69 150 0.46 3.37 1.54 4.97 1.59 1.39 1.53 3.04 1.60 0.00 0.19 0.63 0.30 0.74 1.36 2.54 1.47 178 2.24 1.97 1.33 4.54 1.40 176 1.02 4.18 1.54 7.34 1.61 112 2.93 1.84 1.61 3.34 1.66 0.00 0.63 1.16 0.69 0.00 1.23 5.38 1.41 103 0.95 0.00 1.73 9.05 1.93 160 5.81 5.50 1.96 7.07 2.01 1.24 1.16 3.36 1.22 177 1.68 4.60 1.46 14.13 1.60 70 1.25 0.00 0.80 3.46 0.93 151 0.81 0.33 0.98 2.37 1.12 187 4.07 0.54 0.88 1.24 0.92 186 3.15 17.69 2.45 43.81 2.70 21 1.21 2.36 2.25 6.74 2.35 140 0.70 5.12 2.80 13.38 2.97 61 0.95 2.53 2.43 3.52 2.46

N. S. Loaring et al.

c 0000 RAS, MNRAS 000, 000–000

c 0000 RAS, MNRAS 000, 000–000

No. 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218

XMM-Newton Posa O?axisb Ra J(2000) Dec J(2000) error angle 13 35 05.61 37 50 31.57 0.99 7.59 13 35 05.79 38 07 02.41 1.39 13.62 13 35 06.23 37 49 53.20 0.33 8.08 13 35 07.16 37 45 43.12 1.90 11.27 13 35 07.69 37 48 27.88 2.66 9.24 13 35 08.67 37 57 05.14 1.31 7.13 13 35 09.24 38 04 01.91 1.55 11.43 13 35 09.69 37 50 10.45 2.26 8.46 13 35 09.70 37 48 21.20 0.77 9.60 13 35 12.27 37 48 54.89 1.85 9.62 13 35 12.42 38 02 48.21 0.59 10.90 13 35 12.78 37 44 19.98 0.58 13.04 13 35 14.41 37 49 12.11 1.49 9.78 13 35 14.71 37 52 58.17 1.02 8.22 13 35 14.77 37 48 40.89 1.18 10.15 13 35 14.85 37 50 37.05 0.99 9.10 13 35 14.87 38 03 20.49 2.08 11.62 13 35 15.23 37 58 39.44 0.52 8.91 13 35 15.79 37 52 17.45 0.65 8.61 13 35 15.92 37 52 40.22 0.27 8.53 13 35 16.28 37 56 20.95 1.11 8.42 13 35 17.03 37 49 13.98 14.06 10.18 13 35 17.41 37 54 15.66 0.78 8.54 13 35 17.49 38 02 48.32 0.54 11.61 13 35 18.36 37 55 40.82 2.29 8.74 13 35 18.90 37 54 17.31 1.08 8.84 13 35 19.20 37 43 15.73 1.11 14.66 13 35 19.43 37 53 00.36 0.92 9.12 13 35 19.71 37 49 19.62 1.70 10.58 13 35 19.81 37 58 26.53 1.03 9.65 13 35 22.90 37 49 12.74 0.73 11.17 13 35 24.33 37 46 16.52 0.95 13.14 13 35 24.72 37 51 24.05 1.68 10.57 13 35 24.88 37 44 38.51 1.90 14.34 13 35 25.73 37 52 35.59 1.84 10.43 13 35 30.29 37 57 49.58 0.49 11.43 13 35 32.22 38 03 39.97 0.70 14.39 13 35 32.74 37 45 14.16 4.66 15.08 13 35 33.05 37 48 02.14 1.98 13.50 13 35 35.46 37 57 46.54 0.33 12.41 13 35 35.71 38 01 50.81 1.01 13.96 13 35 35.89 37 51 13.03 1.32 12.72 13 35 37.26 37 47 22.79 1.34 14.55 13 35 38.71 37 55 12.54 1.82 12.72 13 35 39.37 37 56 14.96 0.68 12.92 13 35 42.49 37 55 42.93 0.59 13.49

0.2-0.5 keV 0.5-2.0 keV Counts Counts Counts error Counts error 64.19 12.11 110.06 16.45 2.94 4.49 62.97 10.94 289.05 18.94 509.07 25.29 23.31 6.74 10.05 7.51 24.37 8.29 10.46 7.93 14.35 6.96 58.06 10.67 20.16 7.23 56.79 10.99 14.78 7.59 35.54 9.45 49.75 9.99 106.96 13.36 13.53 7.61 56.55 12.07 16.70 6.95 202.22 17.06 73.12 10.70 146.05 15.02 180.44 21.04 450.74 32.32 31.68 9.16 95.09 15.23 5.56 6.44 74.15 13.31 4.32 5.09 79.49 12.04 5.70 5.05 27.83 8.50 122.14 13.02 214.02 17.32 22.41 8.75 181.13 18.53 301.70 19.32 869.40 32.44 27.09 7.32 65.09 10.80 0.00 2.79 36.41 12.16 52.14 9.62 105.80 13.39 54.00 9.42 248.40 18.10 11.71 6.64 33.30 8.82 14.21 7.33 42.89 9.58 35.19 8.16 85.78 12.37 12.12 7.32 88.72 11.91 6.50 5.38 4.79 6.18 38.19 8.96 86.46 12.35 0.00 2.48 133.86 13.80 34.17 8.50 123.43 14.02 6.72 5.95 41.33 9.07 18.49 7.33 34.24 9.58 0.00 3.66 7.89 6.39 164.89 14.27 333.02 20.63 341.31 124.60 1118.08 250.31 147.90 60.71 339.03 68.07 27.29 7.29 21.51 8.33 303.12 18.62 628.71 26.87 0.36 2.91 44.86 9.88 11.45 6.40 104.41 13.34 15.37 6.41 49.66 9.67 4.83 5.58 4.91 6.18 47.99 8.91 203.05 16.87 98.20 12.26 249.14 18.67 Continued on next page

2.0-5.0 keV 5.0-10.0 keV Counts Counts Counts error Counts error 32.94 10.05 16.65 8.21 27.24 9.73 6.78 8.04 247.65 18.81 40.68 10.75 17.18 7.89 0.00 7.44 6.07 7.07 7.08 7.57 9.81 7.99 0.00 5.22 9.95 7.85 10.54 8.52 5.53 6.41 0.00 4.01 41.20 9.68 0.00 5.37 0.00 5.27 0.00 4.22 102.98 12.87 41.27 10.99 78.19 12.32 6.21 7.74 55.29 17.09 5.33 10.96 48.86 11.11 0.00 4.35 60.01 10.86 13.17 8.95 53.72 10.87 6.56 7.45 13.79 7.65 0.00 2.08 103.73 13.38 6.05 6.96 126.77 15.42 28.58 9.46 358.09 22.56 76.47 12.97 14.60 7.49 11.92 7.53 7.99 8.56 0.00 6.90 63.20 11.13 12.71 8.46 95.96 12.08 19.82 9.27 8.82 6.60 6.19 7.26 53.35 11.27 38.95 9.68 8.65 7.58 26.92 10.77 47.35 9.74 22.10 9.13 44.57 10.68 29.36 10.07 24.91 8.74 0.00 2.03 104.04 12.66 14.15 8.54 30.93 9.89 17.30 9.86 13.20 8.05 3.51 5.87 19.28 8.96 0.00 6.46 38.07 9.72 9.37 7.72 119.05 13.55 28.51 9.50 123.81 85.34 181.69 50.53 205.01 91.84 0.00 47.49 6.53 7.07 1.53 4.84 150.04 15.07 29.36 10.13 134.88 15.04 43.71 10.81 82.13 12.56 19.45 9.95 26.26 9.42 0.69 4.77 57.52 11.35 15.11 8.36 71.99 12.48 14.06 8.45 96.84 13.55 4.42 6.84

0.2-0.5 keV Flux Flux error 1.39 0.26 0.12 0.18 6.56 0.43 0.72 0.21 0.62 0.21 0.36 0.17 0.64 0.23 0.35 0.18 1.32 0.26 0.36 0.20 0.52 0.22 2.71 0.40 4.85 0.57 0.76 0.22 0.15 0.18 0.14 0.16 0.20 0.18 3.11 0.33 0.55 0.22 7.44 0.48 0.75 0.20 0.00 0.08 1.32 0.24 1.84 0.32 0.50 0.28 0.36 0.19 1.89 0.44 0.32 0.19 0.19 0.16 1.05 0.25 0.00 0.08 1.29 0.32 0.20 0.18 0.79 0.31 0.00 0.11 5.41 0.47 23.79 8.68 10.76 4.42 1.16 0.31 11.01 0.68 0.02 0.15 0.43 0.24 0.73 0.30 0.19 0.22 1.84 0.34 4.02 0.50

0.5-2.0 keV Flux Flux error 2.20 0.33 2.36 0.41 10.63 0.53 0.28 0.21 0.24 0.19 1.35 0.25 1.66 0.32 0.79 0.21 2.61 0.33 1.37 0.29 5.80 0.49 5.00 0.51 11.17 0.80 2.11 0.34 1.90 0.34 2.29 0.35 0.89 0.27 5.01 0.41 4.12 0.42 19.73 0.74 1.70 0.28 0.95 0.32 2.45 0.31 7.76 0.57 1.26 0.33 1.01 0.23 4.21 0.61 2.13 0.29 0.13 0.17 2.18 0.31 4.11 0.42 4.30 0.49 1.14 0.25 1.35 0.38 0.22 0.17 10.04 0.62 72.41 16.21 23.82 4.78 0.84 0.33 21.02 0.90 2.23 0.49 3.63 0.46 2.17 0.42 0.17 0.22 7.18 0.60 9.39 0.70

2.0-5.0 keV 5.0-10.0 keV Flux Flux Flux error Flux error 1.51 0.46 3.00 1.48 2.38 0.85 2.65 3.14 11.88 0.90 7.72 2.04 1.14 0.52 0.00 2.07 0.33 0.38 1.54 1.65 0.53 0.43 0.00 1.09 0.68 0.53 3.10 2.50 0.28 0.33 0.00 0.81 2.33 0.55 0.00 1.23 0.00 0.30 0.00 0.96 6.85 0.86 11.76 3.13 6.32 1.00 2.18 2.71 3.18 0.98 1.25 2.57 2.49 0.57 0.00 0.89 3.57 0.65 3.22 2.19 3.52 0.71 1.77 2.01 1.02 0.57 0.00 0.67 5.60 0.72 1.33 1.53 6.65 0.81 6.03 2.00 18.75 1.18 16.10 2.73 0.92 0.47 2.88 1.82 0.49 0.52 0.00 1.72 3.37 0.59 2.74 1.83 7.00 0.88 6.27 2.93 0.72 0.54 2.15 2.53 2.90 0.61 8.59 2.14 1.00 0.87 13.90 5.56 2.63 0.54 5.00 2.06 2.89 0.69 7.84 2.69 1.46 0.51 0.00 0.49 7.45 0.91 4.26 2.57 2.55 0.81 6.19 3.53 0.85 0.52 0.94 1.58 1.80 0.84 0.00 2.67 2.42 0.62 2.49 2.05 8.39 0.96 8.61 2.87 19.24 13.27 122.76 34.14 36.91 16.54 0.00 35.31 0.61 0.66 0.63 1.98 11.78 1.18 10.05 3.47 17.39 1.94 22.80 5.64 6.73 1.03 6.93 3.55 2.73 0.98 0.32 2.21 4.80 0.95 5.53 3.06 6.00 1.04 5.14 3.09 8.63 1.21 1.74 2.69

0.2-10.0 keV Flux Flux error 8.11 1.60 7.50 3.29 36.79 2.33 2.14 2.15 2.74 1.72 2.25 1.22 6.08 2.59 1.42 0.92 6.25 1.41 1.73 1.07 24.93 3.29 16.20 2.96 20.46 2.92 5.36 1.13 8.85 2.31 7.72 2.16 2.11 0.94 15.06 1.77 17.35 2.20 62.02 3.10 6.25 1.91 1.44 1.83 9.89 1.96 22.87 3.13 4.63 2.62 12.87 2.24 21.00 5.68 10.08 2.16 11.06 2.79 4.69 0.81 15.83 2.76 14.33 3.67 3.14 1.69 3.94 2.84 5.12 2.15 32.46 3.12 238.21 40.99 71.49 39.53 3.24 2.14 53.86 3.83 42.44 5.98 17.72 3.73 5.96 2.47 10.69 3.21 20.15 3.33 23.77 3.07

Chandra No.c o?setd 135 0.77 123 4.06 20 0.79 166 7.02 148 200 100 42 34 90 75 67 147 83 37 19 146 86 32 155 194 77 125 131 101 36 31 1.52 1.20 1.10 1.66 0.88 0.79 2.21 0.28 1.76 1.15 0.69 0.51 0.42 1.24 1.91 7.51 0.43 3.20 1.07 1.33 3.55 1.15 1.78

XMM-Newton Deep Field: X-ray source catalogue

156 30

0.72 0.34

15 35 107 139

0.66 0.88 1.67 2.48

23

24

No. 219 220 221 222 223 224 225

XMM-Newton Ra J(2000) Dec J(2000) 13 35 43.17 37 52 58.65 13 35 43.26 37 53 24.27 13 35 44.64 37 51 41.28 13 35 45.19 37 49 59.54 13 35 46.27 37 53 26.79 13 35 52.53 37 50 51.25 13 35 58.70 37 54 06.31

0.2-0.5 keV Posa O?axisb Counts error angle Counts error 1.72 13.74 27.58 7.83 2.17 13.70 21.30 6.79 0.41 14.27 270.01 18.90 1.06 14.85 20.64 7.03 2.12 14.29 25.41 7.64 1.57 15.98 6.01 4.84 2.76 16.68 36.88 9.68

0.5-2.0 keV 2.0-5.0 keV Counts Counts Counts error Counts error 55.53 11.44 21.22 8.31 30.19 9.48 0.00 4.52 524.77 25.80 193.88 17.13 168.87 15.28 83.51 12.07 25.08 8.82 36.04 10.41 24.38 8.50 30.85 8.44 124.62 15.69 9.53 8.54

5.0-10.0 keV 0.2-0.5 keV 0.5-2.0 keV 2.0-5.0 keV Counts Flux Flux Flux Counts error Flux error Flux error Flux error 0.00 7.15 1.46 0.41 2.67 0.55 2.36 0.93 2.55 5.04 1.15 0.37 1.48 0.47 0.00 0.51 33.49 11.30 11.97 0.84 21.47 1.06 18.82 1.66 34.00 10.65 1.23 0.42 9.16 0.83 10.54 1.52 2.94 5.69 1.45 0.44 1.30 0.46 4.34 1.25 12.10 7.68 0.47 0.38 1.88 0.65 6.22 1.70 2.91 5.99 3.02 0.79 9.97 1.26 1.99 1.79

5.0-10.0 keV Flux Flux error 0.00 3.60 1.31 2.60 14.45 4.87 19.68 6.16 1.61 3.13 9.88 6.27 2.45 5.05

0.2-10.0 keV Flux Chandra Flux error No.c o?setd 6.49 3.78 3.95 2.71 66.71 5.32 12 1.14 40.61 6.42 39 1.86 8.71 3.43 134 3.00 18.45 6.54 114 7.62 17.44 5.56

N. S. Loaring et al.

c 0000 RAS, MNRAS 000, 000–000

XMM-Newton Deep Field: X-ray source catalogue
APPENDIX A: SYNTHETIC FIELD GENERATION A1 Method Input sourcelists were generated independently for each of the four energy bands. The di?erential source count function, N (S ) , was modelled as a broken power law distribution: ? ? ? K1 ? ? ? ? N (S ) = ? ? ? ? ? ? K2
S S norm S S norm ? γ1

25

(S < S knee ) (A1) (S S knee )

? γ2

The normalisation constants were chosen such that the function is continuous at the knee, i.e. K1 /K2 = (S knee /S norm )γ1 ?γ2 . The parameters selected for the simulations were based on results from the CDF-S (see Rosati et al. 2002). A photon index, Γ = 1.7 (approximately that of the XRB in the 0.2-10 keV energy range), was used to convert from the Chandra energy bands to the energy bands used in this study. We have slightly modi?ed the Chandra N (S ) parameters via an iterative process so that the ‘output’ source counts approximately match those seen in the 13H ?eld data. In order to incorporate the e?ects of source confusion, we set our simulated source ?ux limits, S lim , to values approximately ?ve times fainter than the limit reached by the 13H data. The mean number of input sources per ?eld per energy band, N , was calculated by integrating N (S ) from S lim to in?nity. The actual number of sources used in each simulated ?eld was taken randomly from a Poisson distribution about N . For each of these sources, a ?ux was randomly assigned from the appropriate N (S ) distribution. Any input source having a ?ux greater than twice that of the brightest source in the 13H data was discarded. This prevented any single simulated ?eld being dominated by an extremely bright source: a situation not seen in the 13H data. The e?ects of source clustering are ignored in this analysis, hence each input source was assigned a purely random position within the ?eld. Finally, the source ?uxes were converted to count rates.

Figure B1. Scatter plot of input vs output ?ux in the 0.5-2 keV energy band. We only show those detected sources having DET ML 5 and valid input counterparts within 5′′ , 8′′ , and 10′′ for input o?axis angles of 0 ? 9′ , 9 ? 12′ and > 12′ respectively.

A3 Simulated image analysis We used a single-band version of the source searching process (as described in Section 2.2) on the simulated images. This process was repeated independently in each of the four energy bands. Two iterations of the background ?tting/source ?nding process were carried out per band per ?eld. The output sourcelist was curtailed at a detection likelihood of DET ML = 5.

APPENDIX B: ASSESSING THE IMPACT OF CONFUSION A2 Imaging characteristics To convert the simulated input sourcelists to images, one has to take account of the complex point spread function (PSF) of the EPIC cameras. We used the ‘MEDIUM’ accuracy PSF description (Kirch 2004) from the XMM-Newton calibration ?les, which is also the model used by the SAS source searching task EMLDETECT. This PSF description consists of a number of small ray-traced maps each describing the PSF at a particular energy and o?-axis angle, covering the full EPIC ?eld of view and energy range. These maps were interpolated in energy and o?-axis angle, allowing us to evaluate the fraction of any source’s ?ux within any image pixel. The SAS-generated exposure maps from the 13H ?eld give the e?ective exposure times and vignetting corrections for each energy band and EPIC camera. The simulated images were generated pixel by pixel by summing the contribution from all input sources. We added a two-component, (vignetted and un-vignetted), synthetic background to the simulated images to reproduce that observed in the 13H ?eld. The correct level of this background was determined through an iterative process because undetected, faint simulated sources contribute signi?cantly to the di?use background. Finally, the value of each pixel was randomly drawn from the appropriate Poisson distribution to simulate photon counting noise.
c 0000 RAS, MNRAS 000, 000–000

B1 Simulation method In order to investigate the ultimate confusion limit of deep XMM-Newton surveys, we have produced a small number of extremely deep simulations, with exposure times 1000× that of the 13H ?eld. The input sourcelists reach ?uxes of a few 10?19 ? 10?18 erg cm?2 s?1 depending on energy band. No background or Poisson noise were added to the images in these simulations, so di?erences between the input and output source counts arise solely as a result of confusion. The sources were matched within a cuto? radius, rcut , which depends on XMM-Newton o?axis angle as described in Section 3.1. B2 Results Fig. B1 shows the distribution of S out /S inp for our 10 simulations in the 0.5-2 keV energy band. The scatter on S out /S inp increases dramatically below a ?ux of 10?16 erg cm?2 s?1 . At similar ?uxes, the input-output position o?sets, shown in Fig. B2 become much larger. The source density at this ?ux level is approximately 2000 deg?2 and represents the limit beyond which source properties cannot be recovered reliably irrespective of exposure time.

26

N. S. Loaring et al.

Figure B2. Greyscale image showing the distribution of positional o?sets between output and input source locations as a function of input ?ux in the 0.5-2 keV energy band. Those sources within the central 9′ of the XMMNewton ?eld of view having DET ML 5 are shown. The concentration of positional o?sets as a function of S inp is indicated by the darkness of the greyscale image. The three contours show the distances within which 68, 90, and 95 per of the data lie.

Fig. B3 show the average input and output integral source counts in each of our four energy bands. The input and output source counts are already discrepant when the source density reaches 2000 deg?2 , and diverge rapidly at fainter ?uxes. The approximate input ?uxes corresponding to source densities of 2000 deg?2 are 3 × 10?17 , 10?16 , 2 × 10?16 and 3 × 10?16 erg cm?2 s?1 in the 0.2-0.5, 0.5-2, 2-5 and 5-10 keV energy bands respectively.

c 0000 RAS, MNRAS 000, 000–000

XMM-Newton Deep Field: X-ray source catalogue

27

0.2-0.5 keV
10000 N(>S) (deg-2) 1000 100 10 1 0.0001 Input Output 10000 N(>S) (deg-2) 1000 100 10 1 0.0001

Input Output

0.5-2 keV

0.001

0.01 Sinp (10
-14

0.1 erg cm s )
-2 -1

1

10

0.001

0.01 Sinp (10
-14

0.1 erg cm s )
-2 -1

1

10

10000 N(>S) (deg-2) 1000 100 10 1 0.0001

Input Output

2-5 keV
10000 N(>S) (deg-2) 1000 100 10 1 0.0001

Input Output

5-10 keV

0.001

0.01 0.1 Sinp (10-14 erg cm-2 s-1 )

1

10

0.001

0.01 0.1 Sinp (10-14 erg cm-2 s-1 )

1

10

Figure B3. Comparison of input (dotted line) and output (solid line) simulated source counts. Input source counts were simulated to ?uxes of 5 × 10?19 erg cm?2 s?1 in the 0.2-0.5 keV and 0.5-2 keV energy bands. In the 2-5 keV and 5-10 keV energy bands the input counts were simulated to ?uxes of 1 × 10?18 and 5 × 10?18 erg cm?2 s?1 . Confusion is evident at faint ?uxes where the input and output source counts rapidly diverge.

c 0000 RAS, MNRAS 000, 000–000



更多相关文章:
...in the 13H XMM-Newton Chandra deep field_免费下....pdf
X-ray spectra of sources in the 13H XMM-Newton Chandra deep field We present the X-ray spectra of 86 optically-identified sources in the 13H XMM-N...
XMM-Newton observation of the Chandra Deep Field-So....pdf
XMM-Newton and Chandra o... 暂无评价 12页 免费 XMM-Newton 13H Deep fiel...XMM-Newton observation of the Chandra Deep Field-South Statistical treatment ...
...in the SubaruXMM-Newton Deep Field at z~0.5_免费....pdf
1?? (2007) Printed 1 February 2008 A (MN L TEX style ?le v2.2) Low-power radio galaxy environments in the Subaru/XMM-Newton Deep Field at z...
...History of the Universe as Revealed from Deep Ra....pdf
**NAMES OF EDITORS** The Star-formation History of the Universe as Revealed from Deep Radio Observations of the 13H XMM-Newton/Chandra Deep Field N. ...
...Large Area Surveys from BeppoSAX to Chandra and XMM-Newton....pdf
(SMBH) accretion activity, which is recorded in...Deep Field North (CDFN [15]), Chandra Deep ...13 erg cm?2 s?1 , using again XMM-Newton ...
XMM-Newton observation of the cluster ZW 1305.4+294....pdf
Deep XMM-Newton observat... 暂无评价 11页 免费...the cluster ZW 1305.4+2941 in the field SA ...H., Mellier, Y., Ponman, T., Quintana, H....
XMM-Newton search for hot gas in the dwarf galaxy I....pdf
Auf dem H¨gel 71, a u 53121 Bonn, Germany Fabian Walter National Radio...ve deep observations (> 400 ks total) obtained from the XMMNewton ...
...clusters at z~0.3 observed by XMM-Newton.pdf
Finoguenov1 , H. B¨hringer1 , Y. Ikebe1,...Deep Field South (CDFS) background and model ...An XMM-Newton observat... 暂无评价 13页 免费...
Deep XMM-Newton survey of M33.pdf
Deep XMM-Newton survey of M33_专业资料。In an XMM-Newton raster ...XMM-Newton 13H Deep fi... 27页 免费 XMM-Newton Surveys of ... ...
X-ray Survey Results on Active Galaxy Physics and E....pdf
tected. The extragalactic survey capabilities of Chandra and XMM-Newton are...Deep Field-South 13 hr Field Chandra Deep Field-North Subaru Deep ELAIS ...
A deep XMM-Newton survey of M 31_图文.pdf
S. 2005, ApJ, 632, 1086 ?0.05 0 0.05 Residials 0 0.1 H. Stiele et al.: A deep XMM-Newton survey of M 31 Williams, B. F., Garcia, ...
Charge-Transfer induced EUV and Soft X-ray emission....pdf
XMM-Newton Hubble Deep Field-North (HDFN) ...The intensity measured at energy hν for an ...gures 1 and 11, 12, 13 (SOM) we chose two...
A deep XMM-Newton X-ray observation of the Chamaele....pdf
A deep XMM-Newton X-ray observation of the Chamaeleon I dark cloud A deep...Typically optical Hα emission and infra-red excess is used to classify ...
A Search for Lyman alpha Emitters at Redshift 3.7_....pdf
and Masayuki Umemura 13 Draft version February 5,...XMM-Newton A A Deep Field on the 8.2 m ...[O ii] λ3727 emitters at z ≈ 0.54, Hβ...
Spectroscopic follow-up of a cluster candidate at z....pdf
a photometric optical/infrared catalogue of the Subaru XMM-Newton Deep Field...eld centred on CVB13. Instead of one massive cluster at z = 1.4, we ...
...Deep Survey Early Data Release.pdf
degrees in the Subaru XMM-Newton Deep Field. We...The cluster catalogue contains 13 clusters at red...3 × 1014 M⊙ for M/M⊙ = 75h. The measured...
The unresolved hard X-ray background the missing so....pdf
to the 2 Ms Chandra Deep Field North and the ...Mateos,3 H. Brunner,2 W. N. Brandt4 and D...The XMM-Newton work showed the resolved fraction ...
The Clustering of XMM-Newton Hard X-ray Sources.pdf
The Clustering of XMM-Newton Hard X-ray Sources...deep (f2?8keV ? 3 × 10?15 erg s?1 cm?...= 13.5 ± 3 h?1 Mpc for γ = 1.8 and ...
Chandra and Spitzer observations of CDFS X-ray obsc....pdf
XMM-Newton and Chandra o... 暂无评价 12页 免费 The Chandra Deep Field-...& Hasinger, G., 2005, ARA&A, 43, 827 Brotherton, M.S., Tran, H....
The X-ray Background and AGNs.pdf
A 400 ksec XMM-Newton observation of the Chandra Deep Field South (see ...Bergeron, H. B¨hringer, A. Fabian, R. Giacconi, R. Gilli, P. o ...
更多相关标签:

All rights reserved Powered by 甜梦文库 9512.net

copyright ©right 2010-2021。
甜梦文库内容来自网络,如有侵犯请联系客服。zhit325@126.com|网站地图