The XMM-LSS survey: The XMDS/VVDS 4 sigma catalogue

Abstract

We present a first catalogue of X-ray sources resulting from the central area of the XMM-LSS (Large Scale Structure survey). We describe the reduction procedures and the database tools we developed and used to derive a well defined catalogue of X-ray sources. The present catalogue is limited to a sub-sample of 286 sources detected at 4 sigma in the 1 deg^2 area covered by the photometric VVDS (VIRMOS VLT Deep Survey), which allows us to provide optical and radio identifications. We also discuss the X-ray properties of a larger X-ray sample of 536 sources detected at > 4 sigma in the full 3 deg^2 area of the XMM Medium Deep Survey (XMDS) independently of the optical identification. We also derive the logN-logS relationship for a sample of more than one thousand sources that we discuss in the context of other surveys at similar fluxes. Comment: 15+6 pages, 12 figures, accepted for publication in Astronomy & Astrophysics The online catalogue announced in the paper will be accessible in about 2 weeks due to technical reasons Fig. 2 replaced with a low resolution one

Full text

Page 1

arXiv:astro-ph/0505117v1 6 May 2005
Astronomy & Astrophysics manuscript no. 2583
(DOI: will be inserted by hand later)
February 5, 2008
The XMM-LSS survey:
The XMDS/VVDS 4σ catalogue⋆
L.Chiappetti1, M.Tajer2,5, G.Trinchieri2, D.Maccagni1, L.Maraschi2, L.Paioro1, M.Pierre3, J.Surdej4,
O.Garcet4, E.Gosset4, O.Le F` evre6, E.Bertin7, H.J.McCracken7, Y.Mellier7, S.Foucaud1, M.Radovich8,
V.Ripepi8, and M.Arnaboldi9
1INAF IASF, Sezione di Milano “G.Occhialini”, via Bassini 15, I-20133 Milano, Italy
e-mail: lucio@mi.iasf.cnr.it
2INAF Osservatorio di Brera, via Brera 28, I-20121 Milano, Italy
3CEA/DSM/DAPNIA Service d’Astrophysique, Saclay, F-91191 Gif sur Yvette, France
4Institut d’Astrophysique et de G´ eophysique, Universit´ e de Li` ege, All´ ee du 6 Aoˆ ut 17, B-4000 Li` ege 1, Belgium
5Universit` a degli Studi di Milano - Bicocca, Dipartimento di Fisica, Piazza della Scienza 3, I-20126 Milano, Italy
6Laboratoire d’Astrophysique de Marseille, Traverse du Siphon, F-13376 Marseille, France
7Institut d’Astrophysique de Paris, 89bis bvd Arago, F-75014 Paris, France
8INAF Osservatorio di Capodimonte, via Moiariello 16, I-80131 Napoli, Italy
9INAF Osservatorio Astronomico di Torino, via Osservatorio 20, I-10025 Pino Torinese, Italy
Received ¡date¿ / accepted ¡date¿
Abstract. We present a first catalogue of X-ray sources resulting from the central area of the XMM-LSS (Large
Scale Structure survey). We describe the reduction procedures and the database tools we developed and used to
derive a well defined catalogue of X-ray sources. The present catalogue is limited to a sub-sample of 286 sources
detected at 4σ in the 1 deg2area covered by the photometric VVDS (VIRMOS VLT Deep Survey), which allows
us to provide optical and radio identifications. We also discuss the X-ray properties of a larger X-ray sample of
536 sources detected at > 4σ in the full 3 deg2area of the XMM Medium Deep Survey (XMDS) independently
of the optical identification. We also derive the logN–logS relationship for a sample of more than one thousand
sources that we discuss in the context of other surveys at similar fluxes.
Key words. X-ray – Surveys – AGN – Catalogues
1. Introduction
The XMDS (XMM Medium Deep Survey) is an X-ray sur-
vey based on the pooling of a significant fraction of the
XMM-Newton guaranteed time of three “hardware insti-
tutes” (IASF Milano for XMM-EPIC, Li` ege for XMM-OM
and CEA Saclay for XMM-SSC).
The XMDS pointings lie at the heart of the full, larger
XMM Large Scale Structure (XMM-LSS) Survey (see
Pierre et al., 2004), to which we refer for a discussion of
the motivations and the choice of the sky field.
About two thirds of the XMDS area are covered in
the optical band by the VVDS (VIRMOS VLT Deep
Survey) both by UBVRI photometry (Le F` evre et al.,
2004a) and by multi-object spectroscopy with VIMOS (Le
⋆The
available
mous ftp to cdsarc.u-strasbg.fr (130.79.128.5) or via
http://cdsweb.u-strasbg.fr/cgi-bin/qcat?J/A+A/
full
in
catalogue
electronic
illustrated
form
in Table
CDS
3 is only
at theviaanony-
F` evre et al., 2004b), by an associated radio survey at 1.4
GHz (Bondi et al., 2003) and by a GALEX Early Release
Observation (Arnouts et al., 2005).
The XMM-LSS area has been covered by an associated
radio survey at 74 and 325 MHz (Cohen et al., 2003) and
will be covered by surveys in other bands like SWIRE
(Lonsdale et al., 2003), the CFHTLS (see web site1) and
by our own follow up observations (Pierre et al., 2004).
Thus, for these areas, multiwavelength information ex-
ists or will be gathered in the future, allowing investi-
gations of large scale structures and source populations
(clusters, galaxies and AGN) at medium redshifts.
For internal usage within the XMDS and XMM-LSS
consortia, and in the future also for public access, a
database has been designed, which is presented in this
paper together with the first results obtained from the
analysis of the XMDS data.
1http://www.cfht.hawaii.edu/Science/CFHLS/

Page 2

2 L.Chiappetti et al.: The XMDS/VVDS 4σ catalogue
The layout of the paper is as follows. The database tool
devised to store the results of the XMM-LSS and XMDS
and associated surveys is described in Section 2. Section 3
contains the analysis procedure (closely following that of
Baldi et al., 2002) used to generate source lists from the
19 XMM pointings of the XMDS using a threshold that
ensures a small number of spurious sources. Different se-
lection criteria are then applied to obtain samples used for
different purposes. A relatively low probability threshold
is used to compute the logN–logS relation presented and
discussed in Section 4. The analysis of the X-ray charac-
teristics is restricted to sources with a signal to noise ra-
tio larger than 4, and the results are presented in Section
5. Only a fraction of the XMDS area is covered by the
VVDS survey. For the latter area the identification pro-
cedure based on the photometric VVDS data and the de-
rived optical vs X-ray properties are described in Section
6.
Preliminary accounts of this work have been presented
at recent conferences (Maraschi et al. , 2003; Tajer et al.,
2004).
2. The database and catalogue
2.1. The LSS database
The web site of the XMM-LSS Master Catalogue2has
been designed as a front end to access the database con-
taining the catalogues, both as a working tool for the
XMDS and XMM-LSS consortia and, in the near future,
for public access.
The query interface uses a Java servlet (communi-
cating via JDBC protocol to the underlying MySQL3
database) to manage the permissions of different groups
of people to access different (parts of) database tables, to
perform selections into multiple tables in an user-friendly
way, to have a quick look at the data and to export results
as ASCII or FITS files.
The database presently includes database tables for the
results of the X-ray pipelines, VVDS photometry for ob-
jects within a 40′′box around an X-ray source, radio and
other catalogues as reported in Section 6.1; it also pro-
vides links to data products like X-ray images and expo-
sure maps, finding charts in the I band, radio maps, etc.
The database includes also pre-calculated correlation
tables, which link the sequence identifiers of objects in an
X-ray table with their neighbours in other tables accord-
ing to predefined proximity criteria (e.g. a 40′′box or a
6′′radius), and greatly speed up queries which involve a
couple of tables.
Correlations between more than two tables are man-
aged similarly by multi-column correlation tables gener-
ated by our identification procedure described below in
section 6.1. They are also used to build virtual tables, i.e.
2http://cosmos.mi.iasf.cnr.it/~lssadmin/Website/
LSS
3MySQL is an open source database server developed by
MySQL AB (see http://www.mysql.com)
simultaneous views of columns taken from different indi-
vidual tables, or even of results of algebraic operations
between different columns.
Usage of appropriate MySQL syntax during the com-
position of a query allows the production of results of ar-
bitrary complexity as output from our database, like e.g.
ds9 region files, or the LATEX code for the catalogue table
reported here (Table 3).
2.2. The catalogue
In this paper we publish, as a first part of the XMDS cata-
logue, the results of the identification procedure described
in section 6.1, i.e. a catalogue of 286 X-ray sources de-
tected at more than 4σ and located in the VVDS area.
Further releases will also include sources already present
in our database and detected at a lower S/N and/or lo-
cated outside the VVDS area.
Our current choice allows us to present a catalogue
inclusive of reliable (though not spectroscopically con-
firmed) identifications and of the basic optical and radio
characteristics of the X-ray sources.
The catalogue published electronically at CDS along
with the present paper (and available as Online material)
includes a rather reduced, manageable number of columns.
A printout sample of the catalogue in tabular form is pre-
sented in Table 3. Comments on individual sources are
presented in Appendix A.
Since our database contains much more information
than what is included in the published catalogue, we plan
to open access to our database query engine, in accordance
with the ”legacy policy” of the XMM-LSS survey.
Usage of our database interface will allow the user
to make selections on the datasets, to access additional
columns (including e.g. X-ray fluxes in all bands, photom-
etry in additional bands, radio fluxes, identification ranks
and flags, technical information, etc.) which is not prac-
tical to include here, and to generate (and operate upon)
expressions involving columns.
The query interface will be accessible from the time the
paper is accepted at http://cosmos.mi.iasf.cnr.it/
~lssadmin/Website/LSS/Query logging in as user xmds,
password guest in workspace public.
3. Observations and data reduction
3.1. Observations
The XMDS observations consist of 19 overlapping point-
ings (typical duration in the range 20–25 ks), covering a
contiguous area of about 3 deg2. The sky location of the
combined field of view (FOV) of the pointings is shown in
Fig. 1.
The observations were performed by XMM-Newton
(Jansen et al., 2001), with the EPIC MOS (Turner et al.,
2001) and pn (Str¨ uder et al., 2001) cameras, between July
2001 and January 2003. A basic journal of observations is
reported in Table 1, containing the ESA dataset identifier

Page 3

L.Chiappetti et al.: The XMDS/VVDS 4σ catalogue3
37.036.5
Right ascension (deg)
36.035.5
-5.5
-5.0
-4.5
-4.0
-3.5
Declination (deg)
G01
G02
G03
G04
G05
G06
G07
G08
G09
G10
G11
G12
G13
G14
G15
G16

G17
G18
G19
Fig.1. Location of the XMDS pointings on the sky : the ap-
proximate FOV footprint (with the pn protruding at the actual
roll angle) is shown (solid lines); the gray colour for field G12
indicates that it has not been used (see text). The dashed rect-
angle indicates the area of VVDS photometry.
that can be used to inquire for further details at the ESA
XMM archive.
The XMDS fields are surrounded at present by 32 ad-
ditional shorter (typically 10 ks) pointings (B01 to B32)
performed in AO1 and AO2 as part of the XMM-LSS pro-
gram. More detailed information is available on an ancil-
lary page of the XMM-LSS Master Catalogue web site4.
3.2. Generation of the X-ray source list
The reduction of the XMDS data has been performed in
Milan using a streamlined version of the pipeline devel-
oped by Baldi et al. (2002) for the HELLAS2XMM survey.
We refer to the latter paper for the formalism and proce-
dure, and only mention here the main points and the few
differences.
We used the XMM-Newton Science Analysis System
(XMM-SAS) v5.4.1, whose tasks emproc and epproc now
provide a reliable identification of the location of bad pix-
els. To remove data contaminated by soft proton flares,
we used fixed thresholds (on the global background above
10 keV) of 0.15 counts s−1for MOS and 0.35 counts s−1
for pn : as a result of this choice, field G12 could not be
analysed because the background was almost constantly
above these thresholds. A mosaic of our X-ray images is
shown in Fig. 2.
As in Baldi et al. (2002), we ran the source detection
(inclusive of background computation) and characteriza-
4http://cosmos.mi.iasf.cnr.it/~lssadmin/Website/
LSS/Anc/goepic.html
tion procedure on merged MOS and pn data in five dif-
ferent energy bands, which we designate as : A (0.3 − 0.5
keV), B (0.5 − 2 keV), C (2 − 4.5 keV), D (4.5 − 10 keV)
and CD (2−10 keV). We note that sources detected in the
CD band do not result from the simple sum of detections
in the separate C and D bands, but from the detection
algorithm run on the full 2 − 10 keV band, while quanti-
ties referring to the total band ABCD (0.3 − 10 keV) are
obtained by summation of the results in the individual
bands.
We have used the XMM-SAS tasks eboxdetect and
emldetect to obtain a list of candidate source positions,
but net counts and associated parameters are obtained
from our own program written according to Baldi et al.
(2002). In particular fluxes were computed for a power law
spectrum with Γ=1.7 and NH=2.61×1020cm−2(i.e. the
average galactic column density in the XMDS field direc-
tion (Dickey & Lockman, 1990)), using conversion factors
calculated from response matrices generated consistently
with the event selections used in our pipeline. In particu-
lar, as in Baldi et al. (2002), we used a conservative pattern
selection of single and double events for MOS, and only
single events for pn. Since the count rates are derived in
regions including a fixed percentage (68%) of the expected
flux from a point source (whose radius depends on off-axis
positions), our fluxes are correct for point-sources. The
fact that a source may fall close to (or on) an inter-CCD
gap is taken into account in flux computation by proper
usage of the EPIC calibration files, and by the usage of
exposure maps in the measurement of the count rate. This
should be an acceptable approximation (the same used by
Baldi et al., 2002). Anyhow, we record in our database a
flag indicating whether a source is close to a FOV edge or
to CCD borders or gaps.
For insertion in our database, we retain only sources
that have a chance detection probability less than 2×10−4
in at least one energy band. In the 19 useful pointings
of the survey, we detected 1322 X-ray sources (including
multiple detections in overlapping regions, see Table 2),
satisfying the probability threshold P < 2 × 10−4. In this
work we present subsets chosen with more stringent cri-
teria: (1) all sources detected with chance probability less
than 2 × 10−5to derive the logN–logS relationship (sec-
tion 4); (2) X-ray sources detected with a signal to noise
ratio larger than 4 in any one of the energy bands defined
above (where the error σSon the net number of counts is
calculated using the Poissonian approximation from the
gross number of counts ctsimg
σS= 1 +?ctsimg+ 0.75
according to Gehrels (1986) as in Baldi et al. (2002)) to
study the mean X-ray properties of the brighter end of
the sample (the 4σ sample, section 5); (3) all sources in
the 4σ sample covered by VVDS photometry, for which
we present the catalogue (section 2.2) and we discuss the
X-ray/optical properties (section 6).
The numbers of sources in each sample are reported in
Table 2.

Page 4

4 L.Chiappetti et al.: The XMDS/VVDS 4σ catalogue
Table 1. Journal of XMDS observationse
Field ESA obs iddate pointing RADEC max exposurec
shiftd∆RA ∆DEC
(ks)(′′)(′′)
G01
G02
G03
G04
G05
G06
G07
G08
G09
G10
G11
G12a
G13
G14
G15
G16b
G16b
G17
G18
G19
011268 0101
011268 0201
011268 0301
010952 0101
011268 0401
011268 1301
011268 1001
011268 0501
010952 0601
010952 0201
010952 0301
010952 0401
010952 0501
011268 0801
011111 0101
011111 0201
011111 0701
011111 0301
011111 0401
011111 0501
28 Jan 2002
14 Jul 2002
19 Jan 2003
29 Jan 2002
02 Feb 2002
26 Jul 2002
30 Jan 2002
25 Jul 2002
31 Jan 2002
29 Jan 2002
02 Feb 2002
01 Feb 2002
03 Jul 2001
31 Jan 2002
06 Jul 2001
06 Jul 2001
14 Aug 2001
03 Jul 2001
03 Jul 2001
04 Jul 2001
02:27:25.4
02:25:54.2
02:24:45.6
02:23:25.3
02:28:05.1
02:26:34.4
02:25:25.3
02:23:54.6
02:22:45.2
02:27:25.4
02:26:05.1
02:24:45.4
02:23:13.1
02:22:04.1
02:27:54.1
02:26:34.2
02:26:35.2
02:25:14.3
02:23:54.1
02:22:34.0
–04:11:06.4
–04:09:05.6
–04:11:00.8
–04:11:07.6
–04:31:08.1
–04:29:00.8
–04:31:07.1
–04:29:00.1
–04:31:11.1
–04:51:04.4
–04:51:06.1
–04:51:11.2
–04:49:03.1
–04:51:09.7
–05:09:02.3
–05:09:03.1
–05:08:46.6
–05:09:08.4
–05:09:09.7
–05:09:02.1
24.9
9.8
21.8
25.8
23.7
13.0
23.5
18.3
22.4
24.8
21.9
n/a
23.9
13.6
21.2
3.9
23.7
22.5
28.1
23.7
–1.60
0.00
–0.53
–1.60
0.00
0.00
–0.53
0.00
n/a
–0.54
0.00
n/a
–2.67
n/a
n/a
n/a
n/a
n/a
n/a
n/a
0.00
–0.53
1.07
0.00
0.00
0.53
–0.53
0.00
n/a
0.53
0.53
n/a
–2.13
n/a
n/a
n/a
n/a
n/a
n/a
n/a
a pointing G12 has not been used because of the high background level (see text)
b pointing G16 has been repeated since the first instance was curtailed
c the reported max exposure is the highest value in the exposure maps of the individual cameras, i.e. net of soft proton flares
d the RA and DEC shifts are the astrometric correction applied (see text) for the fields which overlap the VVDS area and
for which optical identification has been performed
e more details available through our ancillary web page http://cosmos.mi.iasf.cnr.it/~lssadmin/Website/LSS/Anc/
xmdsepic.html
Sample
P < 2 × 10−4
P < 2 × 10−5
4σ (detections)
4σ (independent)
4σ VVDS
Ntot
1322
1129
612
536
286
NB
1166
1028
591
518
278
NCD
419
328
158
143
73
Table 2. Total number of sources and sources detected in the
B and CD bands respectively for the samples presented in the
paper. For the P < 2 × 10−4, P < 2 × 10−5and the first
line of the 4σ sample, numbers refer to⁀detections (i.e. include
multiple detections of the same source in different fields), while
for the second line of the 4σ sample and for the 4σ VVDS
sample numbers refer to independent sources.
Note that in the present version we have analysed
each pointing independently, in spite of the fact that
the FOVs of adjacent pointings overlap. The same ob-
ject could therefore be detected in more than one field.
This fact is recognized only a posteriori (see section 5). In
what follows we will always specify whether the sample we
consider includes multiple detections of the same source.
An independent pipeline is being developed in Saclay
for the analysis of the entire XMM-LSS, using a wavelet
technique and therefore best suited for extended sources
(Pacaud et al. (2005), see a preliminary account in Pierre
et al. (2004)) and will be used as a basis for the future
complete XMM-LSS catalogue.
4. The logN–logS relationship
We computed the logN–logS distributions in the 0.5 − 2
and 2 − 10 keV band as follows.
We have considered all sources with a detection prob-
ability in the 0.5 − 2 and 2 − 10 keV bands below P =
2 × 10−5. This is well below our acceptance threshold of
2×10−4, ensuring very few spurious sources in the sample
while giving us a large dataset to compute the logN–logS.
We have generated a flux limit map for each X-ray
field and energy band, which contains the faintest flux at
which a source can be detected at the assumed level of
significance above the local background. From these we
have computed the sky coverage plotted in Fig. 3. This
considers all fields as independent areas, consistently with
our detection procedure (see section 3.2), so the full sky
coverage is the sum of all areas surveyed.
Our faintest flux is ∼ 10−15erg cm−2s−1in the 0.5−2
keV band and ∼ 7×10−15erg cm−2s−1in the 2−10 keV
band.

Page 5

L.Chiappetti et al.: The XMDS/VVDS 4σ catalogue5
36.00037.00036.500 35.500
-5.000
-4.000
-4.500
-5.500
Right ascension
Declination
Fig.2. Mosaic of our X-ray images in the total (0.3−10 keV) band. No exposure map correction has been applied to this figure.
In Fig. 4 we plot the cumulative logN–logS in the 0.5−2
keV band (upper left panel, 1028 sources) and in the 2−10
keV band (lower left panel, 328 sources). The plots relative
to individual fields have also been produced and inspected
and do not show significant deviations from the combined
distribution.
The shapes of the (cumulative) logN–logS are clearly
curved and cannot be represented by a single power law. A
fit to the cumulative distribution does not allow a mean-
ingful computation of confidence limits on the parameters,
because points and errors are not independent (Murdoch
et al., 1973; Crawford et al., 1970). Therefore we com-
puted the differential logN–logS by binning (in logarithmic
space) the number N(F) of sources with flux F into flux
bins of width ∆Fi(i.e. between fluxes Fjminand Fjmax),
then computing the average sky coverage Ai in the bin.
The i-th bin of the differential curve n(F) (shown in the
right panels of Fig. 4) is given by
n(F)i=
Fjmax
?
Fjmin
N(Fj)
∆FiAi
A statistical error due to the Poissonian error on the num-
ber of sources is assigned to each bin. Again, we confirm
that a single power law does not give a good fit. Therefore
we used a broken power law.
For the 0.5 − 2 keV band, the best fit is
n(F) =6.515× 103× F−2.62
384.2× F−1.42
15
F15> 10.58
F15< 10.58
15
All fluxes F15are normalized to 10−15erg cm−2s−1. 90%
confidence limits on slopes are 2.62+0.25
while the break position is in the range 1.06+0.30
erg cm−2s−1.
For the 2 − 10 keV band we find
−0.22and 1.42+0.14
−0.22× 10−14
−0.15,
n(F) = 4.483 × 104× F−2.91
15
for F15 > 21.4, with 90% errors on slope 2.91+0.45
break at 2.14+0.81
the slope is not well constrained: the best fit is 1.53, but
the confidence interval ranges between 0.37 and 2.04.
In Fig. 4 we show the fit results both in the differential
and the cumulative forms for a better comparison with the
literature that gives sometimes cumulative (e.g. Baldi et
al., 2002; Moretti et al., 2003), sometimes differential plots
(e.g. Yang et al., 2004; Harrison et al., 2003). Given the
uncertainties in the fits, we prefer the comparison with the
actual data where possible.
−0.30and
−0.54× 10−14erg cm−2s−1. At lower fluxes