During the year 2004, our team has pursued its involvement in theoretical and observational studies of quasars and gravitational lenses, the international 4m Liquid Mirror Telescope project and observations of large scale structures in the Universe using the Hubble Space Telescope, the X-ray ESA Newton-XMM satellite, the ground-based ESO-La Silla NTT and VLT telescopes at Paranal (ESO, Chile) . Most of these activities have been carried out in the context of international collaborations, including also scientists from the Royal Observatory of Belgium (cf. collaborators whose names appear in the scientific publications).

IV. a Extragalactic Astrophysics

1) Gravitational lenses

       Gravitational lensing may perturb our view of the distant Universe and affect our physical understanding of various classes of extragalactic objects. The great interest in gravitational lensing comes from the fact that this phenomenon can be used as an astrophysical and cosmological tool. Indeed, gravitational lensing may help in deriving (i) the distance scale of the Universe, via the determination of the Hubble constant H0 based upon the measurement of the time delay Dt between the observed lightcurves of multiply imaged quasars, (ii) the values of other cosmological parameters (W0 and l0), (iii) the mass distribution M(r) of the lens, (iv) the extinction law in the deflector usually located at high redshift, (v) the nature and distribution of luminous and dark matter in the Universe, (vi) the size and structure of quasars, (vii) the size of absorbing intergalactic gas clouds and (viii) upper limits on the density of a cosmological population of massive compact objects. During 2004, we have continued to contribute to most of these endeavors thanks to ground-based and HST observations.

       Due to the arcsec scale (or less) of the angular separation between the multiple images of lensed QSOs, the HST is badly needed to confirm the nature of the lens candidates, usually first identified by means of ground-based observations. This is well illustrated for the case of the quadruply imaged quasar 1RXS J1131-1231 A-D + an optical Einstein ring (Sluse et al. 2004) that has also been discovered by our team and extensively studied with HST

       Multiply imaged QSOs provide general constraints on the mass distribution of the lensing galaxies. For example, if the value of the Hubble constant H0 and the lens and source redshifts are known, the total mass can be derived from the angular separation between the lensed QSO images. This is the simplest and most direct astrophysical application of gravitational lensing. If a time delay is also measured, the lensing mass can be obtained independently of the cosmological parameters H0, 0, 0 and the source redshift. The Hubble Space Telescope plays a crucial role in detecting the main lenses, which are faint, extended, high redshift objects located between bright and angularly close (~1 arcsec) images of the background quasars. Careful subtraction of the HST Point Spread Function and/or deconvolution is often mandatory. Based upon direct imagery with the ACS camera onboard HST, it was easy to detect the lens galaxy for the quadruply imaged quasar 1RXS J1131-1231 A-D (Sluse et al. 2004).

       Multiply imaged quasars deserve very special investigations to determine the extinction law in the lensing galaxies. Indeed, when the lines-of-sight corresponding to the different images probe different dust optical depths in the lens, a differential reddening is observed between the lensed images. Until now, only Galactic parametric curves have been fitted from photometric data but high quality spectrophotometry should allow to directly determine the extinction law at high redshift and compare it with that observed in the Milky Way, in the Magellanic Clouds and in a few local galaxies. Jean et al. (2004) have submitted a paper studying the extinction law among multiply imaged quasars, in general.

       Photometric studies have been carried out by our team and good evidence for micro-lensing has been reported for the multiply imaged quasars 2237+0305 (Moreau et al. 2004, Lee et al. 2004), UM673 (Nakos et al. 2004) and 1RXS J1131-1231 A-D (Sluse et al. 2004).

       A photometric monitoring campaign whose aim is to measure the time delay in a sample of about 15 gravitationally lensed quasars has been initiated as a ULg-KU Leuven joint venture. For that purpose, we have indeed set up a large collaboration including ULg, KU Leuven, Ecole Polytechnique Fédérale de Lausanne and several other institutes (see Web page:

       The observations, which started in April 2004, are carried out with several telescopes: the 1.2m Euler telescope at La Silla (Chile), the 1.2m Mercator telescope at Las Palmas (Canary Islands), the 2m Liverpool telescope on the same site and the 1.5m telescope of Maidanak Observatory (Uzbekistan).

       All the images obtained will be processed with the MCS deconvolution software in order to derive precise light curves and to determine the time delay between the different images of the same quasar.

       In parallel, HST images and VLT spectra of these gravitational lenses will be analyzed in order to build accurate models of the light deflexion. Once the time delays and models are constructed, we shall be able to compute a value of the Hubble constant for each of these systems. A comparison of these values with those determined from more classical techniques will allow not only to check the validity of the different independent methods, but also to put constraints on several cosmological parameters as well as on the amount and distribution of dark matter around the lensing galaxies.

       Several optical lens simulators made of plexiglass have been produced by our team and used to simulate the formation of multiply imaged quasars by a foreground lens galaxy. Such didactictal experiments enable one to reproduce all image configurations (cf. double or quadruple lensed images, Einstein ring, giant luminous arcs, arclets, etc.) that have been observed with the Hubble Space Telescope (Surdej et al. 2004). One of these simulators is being used in the 'Cosmology' exhibition organized at the National Air and Space Museum in Washington from 2000 until 2015.

       A non exhaustive, although quite complete, bibliography dedicated to theoretical and observational studies of gravitational lens systems as well as a database summarizing the main observational and physical parameters of multiply imaged sources, including a set of color images of most of these cosmic mirages, are available on the web at the following URL address

2) Quasar host galaxies (ULg)

       The group has undertaken a comprehensive spectroscopic study of quasar host galaxies, on a sample of more than 20 low-redshift quasars (z < 0.3) selected from the Hamburg-ESO survey. Long slit spectra, centered on the quasar itself, have been obtained at the ESO VLT with the FORS spectrograph. We used a multi-slit observing mode in order to register the spectra of neighbouring stars, which were used to spatially deconvolve the quasar spectra using the spectroscopic version of the MCS deconvolution algorithm (Courbin et al. 2000, ApJ 529, 1136). This algorithm allows an efficient separation of the spectrum of the host galaxy from the much brighter quasar spectrum.

       These spectra allow to study the structure and dynamics of the host galaxy, as well as its stellar populations and gas content, from the outer regions down to the central kiloparsec. A first application to the quasar HE 1503+0228 showed that its host galaxy is a rather common spiral (Courbin et al. 2002, A&A 394, 863) with mass ~ 2 · 1011 M.

       Another object, HE 1434-1600, shows a particularly interesting host galaxy, which interacts with a companion situated a few kiloparsecs away. Despite being elliptical, the quasar host galaxy displays prominent emission lines, extending very far from the central active galactic nucleus. The analysis of these lines, which originate in gas having probably been swept out during the close interaction with the neighbouring galaxy, show that the source of excitation and ionization energy is the quasar itself, which is able to excite the gas at very large distances (~ 4 kiloparsecs). These results will soon be published (Letawe et al., A&A 424, 455). The analysis of the whole quasar sample forms the basis of G. Letawe's PhD thesis.

       Some of these quasars have also been observed in integral field spectroscopy at the VLT, either with the GIRAFFE or with the VIMOS instrument. These spectra, which are in the reduction process, will provide a spectral mapping of the whole quasar environment, which should allow to build two-dimensional maps of the velocity fields, gas excitation and stellar populations.

       Moreover, observing time was obtained on the Hubble Space Telescope to get high resolution images for the quasars observed in integral field spectroscopy. These images, obtained with the Advanced Camera for Surveys will be deconvolved with the MCS algorithm (Magain et al., ApJ 494, 472) which allows to separate efficiently the light of the point source (the quasar itself) from the light of its host galaxy. This will not only help to interpret the 2-D spectra, but also to study the properties of the host galaxy down to the central few hundred parsecs, even for luminous quasars.

3) Quasars at high energies (ULg – ROB)

       A Liège team from IAGL (J.-F. Claeskens, A. Detal, O. Garcet, E. Gosset, C. Jean, J. Manfroid, D. Sluse, P.-G. Sprimont and J. Surdej) and C. Libbrecht (Royal Observatory of Belgium) are part of the XMM-LSS international Consortium in which astronomers from 15 different institutes are joining their efforts to produce a wide X-ray survey for studies of large scale structures (hereafter LSS) and active galactic nuclei (AGN/QSOs) at high galactic latitude in the Universe. The PI of the XMM-LSS project is Dr. M. Pierre (CEA, Saclay, France). The XMM-LSS survey characteristics are the following (see Pierre et al. 2004): we presently (GT, AO-1 - AO-3) have covered a 5 deg2 area and reached a sensitivity of at least 5 10-15 erg. cm-2.s-1 for point-like sources and 10-14 for typical cluster emission in the [0.5-2] keV band. Our goal is to cover a total area of 10 deg2. The XMM-LSS survey mainly consists of 10 ks XMM/EPIC pointings separated by 20 arcmin offsets. The survey area surrounds two deep XMM surveys based on guaranteed time: the XMM_SSC/Subaru Deep Survey (80 ks exposures in 1 deg2) and the XMM Medium Deep Survey (XMDS; 20 ks exposures in 2 deg2) also corresponding to the VIRMOS-DESCART Deep Survey [deep], the latter being a collaboration between several instrumental teams: XMM-OM (Liège), XMM-EPIC (IASF-MILANO), XMM-SSC (Saclay), CFHTLS (Saclay, IAP) and VIRMOS (LAM, IASF-MILANO, OAB). This area is also being observed by the associated DESCART-VIRMOS Deep Survey [wide], the Spitzer Wide Area Infrared Extragalactic (SWIRE) SIRTF Legacy Survey, the Galex ultraviolet survey and the NOAO deep survey; the 8.75 deg2 UKIDSS survey is centered on the Subaru Deep Survey and the CTIO R-z' imaging covers a region corresponding to the VVDS [wide] and [deep] surveys. A very detailed description of the XMM-LSS project and consortium activities are accessible via the URL : This web page is maintained and entirely managed by the Liège team.

       Assuming the current favoured cosmological values for CDM, the entire cluster/group population will be detected out to z = 0.5 and the total number of clusters will be of the order of 150 with 0 < z < 1 (about half of them below z < 0.7). The XMM-LSS survey will also be ideal to probe the behaviour of the bright end of the X-ray luminosity function beyond z > 1 : a new exciting territory. At the low X-ray flux achieved by the XMM-LSS survey, there are approximately 300 X-ray sources per square degree, among which typically : 15-20 are galaxy clusters, about 200 AGN and QSOs (100 QSOs/AGN with z < 1) and the remainder stars and nearby galaxies. The surface density of QSOs/AGN in our proposed survey is almost an order of magnitude larger than those expected in the on-going (colour biased, z < 2.5, ?) optical surveys carried out over large sky areas (SDSS, 2dF, etc.). Compared to optical surveys, the completeness of X-ray selected QSOs/AGN is also known to be much better defined (Chiappetti et al. 2004). The high sensitivity and good point-spread function of XMM has opened a new era for X-ray LSS studies. Some 1000 times more sensitive than the REFLEX survey - the largest cluster survey over a single area to date (Böhringer et al. 2001) - the XMM-LSS survey has been designed to probe the large scale distribution of galaxy clusters out to z ~ 1 and of QSOs much further out. This will provide unprecedented insights onto LSS formation and, thus, cosmology. We will be able to probe the nature and amount of dark matter, the initial fluctuation spectrum and other fundamental cosmological parameters.

       The primary science goals of the entire XMM-LSS survey are to :
- Map the LSS as outlined by galaxy clusters and groups out to z = 1. This will reveal the topology of the spatial distribution of deep potential wells at truly cosmological distances.
- Compute the correlation function of clusters of galaxies in two redshift bins 0 < z < 0.5, 0.5 < z < 1. This will be the first determination of the cluster correlation function at high z.
- Map the spatial distribution of AGN/QSOs within the cosmic web as outlined by the cluster/group population. This will lead to a better understanding of the origin of AGN in terms of the initial density perturbation, galaxy interactions, etc. If the AGN unified scheme is correct, all AGN should be found in the same environment, regardless of their type (cf. type 1 or type 2 AGN, QSO, ?).
- Compute the correlation function of QSOs/AGN with a high degree of accuracy.
- Investigate the existence of X-ray bright (i.e. massive) galaxy clusters with 1 < z < 2.
- Study the combined X-ray/optical evolutionary properties of clusters and QSOs.
- Compare the cosmic web inferred from X-rays with the mass distribution determined by the galaxy distribution and the associated weak lensing survey in the optical. This will provide crucial information about bias mechanism as a function of redshift.
Up to now, 51 XMM fields have been collected and a nice first X-ray mosaic of the fields has been constructed (see the URL for the Images and X-ray mosaic). A secure X-ray catalogue of extended and point-like sources has been validated recently (June 2004) in Saclay. Examples of cluster candidates identified in this survey may also be seen via the above URL which is being totally administrated by the Liège team.
The number (50 at present) of galaxy clusters present in the XMM-LSS catalogue constitutes a challenge for a complete spectroscopic follow-up to be carried out in a reasonable amount of time. Moreover, only a small fraction of the clusters do have a known redshift in the literature. To fulfil our LSS goals, we need to measure redshifts for all z < 1 clusters. This is presently done in the Multi-Object-Spectroscopy (MOS) mode, using NTT/EMMI, VLT/FORS2 and VLT/VIMOS at ESO and Magellan/ IMACS, Magellan/LDSS at Las Campanas (Chile). We usually take 1 mask per cluster, randomly sampling the AGN/QSO population at the same time, the underlying filamentary galaxy distribution connecting clusters, radio sources from our VLA survey as well as a representative sample of the SWIRE sources (as of October 2004). This mapping around the 0 < z < 1 clusters has an enormous scientific potential for studies of galaxy environments and bias.
We presently organize the observational schedule using two telescope classes: the 4 meter class for the lower redshift bin (0 < z < 0.5, see Willlis et al. 2004) and the 8 meter class for the higher redshift bin (0.5 < z, see Valtchanov et al. 2004). In both cases, we compute a mean progress rate of 5-7 clusters per night, all observed spectroscopically in the multi-object mode (MOS). The Liège team is regularly participating to the observations. In addition, an automatic spectroscopic pipeline (Gspectro) has entirely been set up and is still under active development by the Liège team and enables to-day to automatically reduce, analyze and archive the MOS spectra collected with the 4m and 8m class telescopes. So far, P.-G. Sprimont, C. Jean, C. Libbrecht (ROB) and J. Surdej have reduced and analyzed two runs of spectroscopic observations (starting from the X-ray/optical identification of cluster galaxies and AGN/QSO candidates up to the determination of their spectroscopic identification and measurement of redshifts). In the context of the present project, the main technical and scientific activities of the Liège group will consist in the optical, near-infrared (IR) and spectral identification of the counterparts of X-ray sources detected in the XMM-LSS survey, in the statistical study of their 2D and 3D spatial distributions and in the determination of the X-ray/optical/IR luminosity functions of the identified AGN/QSOs and galaxy clusters. We shall then be in a good position to probe the evolution of the cosmic network traced by clusters and QSOs over large volumes of the Universe to high redshift.

4) Astrometry of quasars

       The European astronomical mission GAIA will provide precise astrometric and photometric observations for about 500,000 quasars over the whole sky. Besides its potentialities for the studies of the QSO phenomenon, large scale structures and the Celestial Referential Frame, J.-F. Claeskens, A. Smette (presently at ESO) and J. Surdej also showed that such a large QSO catalog would contain about 10 times the present known number of gravitational lenses.
However, QSOs must first be identified among stars whose mean population ratio is about 2000, and reliable values of the astrophysical parameters (e.g. photometric redshift, continuum slope, extinction,...) should be obtained without the help of spectroscopic confirmations.
In a first step study, making use of the GAIA broad band and medium band photometric measurements, we have compared the expected performances of template fitting, neural networks, and Spectral Principal Component Analysis to i)- classify QSOs with the lowest contamination rate due to mis-classified stars and ii) - determine QSO astrophysical parameters.

5) Broad Absorption Line Quasars

       H. Lamy (IASB/BIRA, ROB) and D. Hutsemékers have finalized their study of the polarization of the Broad Absorption Line quasars (BAL) and in particular of the correlations existing between the polarization degree and other quantities characterizing these objects. The main result of this analysis is interpreted in the framework of a BAL model explaining the formation of the absorption lines in a wind accelerated from the quasar accretion disk. In collaboration with P. Hall (Princeton, USA), D. Hutsemékers has continued the study of extreme BAL QSOs in the SDSS using high-resolution spectroscopy. They namely show that the outflow in SDSS J0300+0048 has the highest column density yet reported for a broad absorption line quasar, the absorption from different ions being segregated as a function of velocity in a way that can only be explained by a disk wind outflow. From the analysis of the excited SiII and C II lines in the spectrum of SDSS J00116+0055, they also found that some of the absorption should occur at distances of ~ 20 kpc, i.e much higher than the distance usually assumed for BAL absorbers.

IVb. International Liquid Mirror Telescope (ILMT) project

       The great scientific motivation to use a telescope dedicated to the photometric monitoring of multiply imaged quasars (gravitational lens systems) has led us to assess the expected performances of a liquid mirror telescope. In collaboration with scientists and engineers from the Liège Space Center (locally known as CSL), from the industrial company AMOS (Advanced Mechanical and Optical Systems) and from multiple international research institutions, the Royal Observatory of Belgium (ROB) and the Liège Institute of Astrophysics and Geophysics Institute (IAGL) are presently involved in the construction of a 4m liquid mirror telescope (known as the International Liquid Mirror Telescope project) which should be erected in 2006/7. An official letter from the ESO Director General has been received in December 2003 giving us an agreement to install the telescope at La Silla (ESO, Chile).

       The main goal is to daily monitor a narrow strip of sky (half a degree wide) near the -29°declination down to the magnitude B = 24. This survey will approximately cover a field of 100 square degrees at high galactic latitude (|bII| 30°). We have estimated that more than 50 multiply imaged quasars should be photometrically monitored every night. Some 50 cases of micro-lensing effects taking place in our own Galaxy should also be detected every year. The stacking of all CCD frames recorded during the span of one year should also lead to the detection of some 50 giant luminous arcs. The lightcurves recorded for the multiple point-like images of the macro-lens systems will significantly contribute to an independent determination of the Hubble constant and to observational studies of extragalactic micro-lensing effects. These systems will constitute unique targets for subsequent observational studies with the VLT and ALMA. Programs aimed at the calibration of the astrometry of the ILMT survey have been successfully completed during 2004. We have also received at the end of 2004, the missing and necessary funds from the Education and Research ministry of the French Community of Belgium to complete the construction of the Liquid Mirror Telescope and of its dome (total funding of 1 200 000 Euros at the moment).

Web pages:Sandrine Sohy
Last modification: April 27, 2005