1

No. 111 – March 2003

Inward Bound: Studying the Galactic Centrewith NAOS/CONICAT. OTT1, R. SCHÖDEL1, R. GENZEL1,2, A. ECKART3, F. LACOMBE4, D. ROUAN 4, R. HOF-MANN1, M. LEHNERT1, T. ALEXANDER5, A. STERNBERG6, M. REID7, W. BRANDNER8,9,R. LENZEN 8, M. HARTUNG8, E. GENDRON 4, Y. CLÉNET4, P. LÉNA4, G. ROUSSET10, A.-M. LAGRANGE11, N. AGEORGES9, N. HUBIN9, C. LIDMAN9, A.F.M. MOORWOOD9, A. RENZINI9, J. SPYROMILIO9, L.E. TACCONI-GARMAN9, K.M. MENTEN12, N. MOUAWAD 3

1MPI für Extraterrestrische Physik (MPE), Garching, Germany; 2Dept. of Physics, Univ. of California, Berkeley,USA; 31. Physik. Institut, Univ. Köln, Germany; 4Obs. de Paris-Meudon, France; 5Faculty of Physics, WeizmannInstitute of Science, Rehovot, Israel; 6Dept of Astronomy and Astrophysics, Tel Aviv Univ., Israel; 7Center forAstrophysics, Cambridge, USA; 8MPI für Astronomie, Heidelberg, Germany; 9ESO, Garching, Germany; 10OfficeNational d’Etudes et de Recherches Aérospatiales (ONERA), Chatillon, France; 11Obs. de Grenoble, Grenoble,France; 12MPI für Radioastronomie, Bonn, Germany

Figure caption: Orbit of the star S2 aroundSgrA* (from Schödel et al. 2002, 2003). Left:Wiener filtered, H-band NAOS/CONICA im-age (40 mas resolution) of the central 2″.The light blue cross denotes the position ofSgrA*, and its 3σ error bar. Several of theblue ‘S’-stars in the SgrA*-cluster aremarked, as well as three of the bright IRS16stars. The right inset shows the orbital dataand best keplerian fit of S2 around the posi-tion of Sgr* (circle with cross). Filled circlesare measurements with the SHARP specklecamera on the 3.5-m NTT, and opensquares are the new NAOS/CONICA meas-urements. Error bars are conservative esti-mates, including systematic uncertainties.

Introduction

Because of its proximity (see foot-note on page 2) the centre of the MilkyWay is a unique laboratory for studyingphysical processes that are thought tooccur generally in galactic nuclei. The

central parsec of our galaxy is known tocontain a dense and luminous star clus-ter, as well as several components of

neutral, ionized and extremely hot gas.For two decades, evidence has beenmounting that the Galactic Centre har-

force was indeed centred on the radiosource (Ghez et al. 2000). The NTT andKeck results proved that 2–3.3 ×106 M

are concentrated within ≤ 10 milli-par-sec (2000 AU) of SgrA*. This mass con-centration is most likely a supermas-sive black hole.

NAOS/CONICA Observations

While a central black hole is the mostprobable configuration, the NTT/Keckdata before 2002 could not exclude sev-eral alternatives, such as an extremelycompact cluster of dark astrophysicalobjects (e.g. neutron stars, stellar blackholes, or white dwarfs), or a degenerateball of hypothetical heavy neutrinos.Even if most of the mass is in a blackhole, the hole may be surrounded by ahalo of heavy stellar remnants or heavyparticles, or may even not be a singleobject. To address these questions, it isnecessary to determine the gravitation-al potential closer to SgrA*, and withgreater precision than was previouslypossible. The resolution of an 8-m-class telescope is required. It is desir-able to replace the statistical method ofdeducing the mass distribution from av-eraging velocities of many stars by themore precise technique of determiningmasses from individual stellar orbits.There are also a number of other in-triguing issues that could not be ad-dressed by the comparatively shallowNTT/Keck observations. If SgrA* is in-deed a massive black hole there shouldbe a cusp of low mass stars cen-tred on it. It would also be important toexplore other theoretical predictionscharacteristic of the black hole environ-ment, such as stellar collisions and dis-ruptions, or gravitational lensing. Thesegoals call for much more sensitivemeasurements than have been achiev-able with speckle techniques. Adaptiveoptics (AO) imaging with good Strehlratios is required. It is of great interestto better understand the evolution ofthe nuclear star cluster itself. Giventhe existing evidence for hot, massivestars in the central parsec (e.g. Genzelet al. 1996) one would like to under-stand how they have come into being inthe very dense central environment.Finally it is highly desirable to de-tect the infrared counterpart of SgrA* it-self. Infrared flux, spectral energy distri-bution and polarization provide impor-tant constraints on the poorly under-stood properties of black hole accretionflows.

The ideal instrument for tackling allthese basic questions is the new AOimager/spectrometer NAOS/CONICA.NAOS/CONICA was mounted on UT4(Yepun) in late 2001, and first observa-tions became possible in early/mid2002 during the commissioning andscience verification phases of the in-

strument. Compared to the AO imagersat the Keck and Gemini telescopes,NAOS/CONICA has two unique advan-tages. First, the Galactic Centre passesthrough zenith, compared to a mini-mum airmass of 1.5 at Mauna Kea.Second, NAOS/CONICA has an in-frared wavefront sensor. This permitswave front correction on the bright redsupergiant IRS7 located 5.5″ north ofSgrA*. All other instruments have tolock on a relatively faint visible star at adistance of about 20″. Furthermore, the1K2 CONICA detector gives a field ofview as large as 30″, resulting in a ma-jor improvement in the astrometricanalysis of the infrared data (see be-low). After discussion and consultationwith the ESO Director General, it wasdecided, therefore, to immediately car-ry out a first series of imaging observa-tions of the Galactic Centre with thenew instrument, as part of a collabora-tion of the two instrument teams andthe ESO/Paranal staff. Beyond the sci-entific results, it also turned out that theGalactic Centre was an optimal targetfor verifying and optimizing the techni-cal performance of NAOS/CONICA.The 2002 NAOS/CONICA observationsimmediately gave us the best GalacticCentre images ever taken, and led to amajor breakthrough in our knowledgeof the central parsec. The data are pub-licly available in the ESO archive. Wegive here a first account of what wehave learned (see also Schödel et al.2002, 2003, Ott et al. 2003, Genzel etal. 2003, Reid et al. 2003).

Image Analysis and Astrometry

The observations were carried outduring commissioning and science ver-ification with NAOS/CONICA betweenMarch and August 2002. Data were tak-en in H-band, (1.65 µm), Ks-band (2.16µm) and L’-band (3.76 µm) with the 13and 26 milli-arcsec (mas) pixel scales,with seeing varying between 0.5 and1.3″. For H and K we used the infraredwavefront sensor of NAOS to close theloop on the bright supergiant IRS 7~ 5.5″ north of SgrA*. At L’ we closedthe AO loop on a V ~ 14 star ~20″ to thenorth-east of SgrA* (at the time of theobservations the dichroic allowing thesimultaneous use of the infrared wave-front sensor and the L/M-bands was notavailable). The individual images wereflat-fielded, sky-subtracted and correct-ed for dead/bad pixels. The final frameswere co-added with a simple shift-and-add (SSA) algorithm. In Figure 1 weshow the Ks-band SSA image (left in-set), as well as an H/Ks/L’-colour com-posite, SSA image (right inset) obtainedas part of the science verification run inAugust 2002. These are the best im-ages of the 2002 season, with Strehl ra-tios > 50% at Ks and 33% at H. With in-

2

bours a concentration of dark mass as-sociated with the compact radio sourceSgrA* (diameter about 10 light-minutes),located at the Centre of the nuclear starcluster. SgrA* thus may be a supermas-sive black hole. High-resolution obser-vations offer the unique opportunities ofstringently testing the black hole para-digm and of studying stars and gas inthe immediate vicinity of a black hole, ata level of detail that will never be ac-cessible in any other galactic nucleus.

At visible wavelengths, the GalacticCentre is hidden behind 30 magnitudesof Galactic foreground extinction. Op-tical and ultraviolet observations areimpossible or exceedingly difficult. Ourknowledge of the Milky Way Centrethus derives from observations at bothshorter (X- and γ-rays) and longerwavelengths (infrared and radio) thatcan penetrate through this obscuration.Near-infrared observations from largeground-based telescopes are particu-larly useful for studying the stellar com-ponent. Stars are excellent test parti-cles of gravity. Stellar motions are idealtools for significantly improving the con-straints on the concentration and loca-tion of the central mass inferred fromearlier observations of gas velocities.For this purpose, over the past decadea group of us at the Max-Planck-Institutfür Extraterrestrische Physik (MPE) hasbeen using the high-resolution, near-in-frared camera SHARP, employed as aregular visiting instrument on the 3.5-mESO NTT. The primary goal of the NTTobservations was to detect stellar prop-er motions in the immediate vicinity ofSgrA* by exploiting the excellent imag-ing properties and seeing of that tele-scope for diffraction limited H- and K-band, speckle imaging of the GalacticCentre. A second experiment used thenovel ‘3D’ near-IR integral field spec-trometer for measuring Doppler veloc-ities on the 2.2-m ESO/MPIA telescope.

When the first SHARP and 3D resultsof the stellar dynamics emerged 6years ago (Eckart & Genzel 1996,1997, Genzel et al. 1996, 1997) theyprovided a compelling case for thepresence of a compact, central mass.The stellar velocity dispersion increas-es towards SgrA* from 20″ to < 1″ witha Keplerian law (σv∝R-1/2). Severalstars in the ‘SgrA* cluster’ within 1″ ofthe radio source exhibit velocities in ex-cess of 103 km/s. Shortly thereafter, theNTT results were confirmed andstrengthened by independent proper-motion data obtained with the 10-mKeck telescope (Ghez et al. 1998). TheKeck group also reported two years lat-er the first orbital accelerations of threestars, demonstrating that the Centre of

We assume here that the Galactic Centre is 8 kpcfrom the Sun. At that distance 1 arcsecond corre-sponds to 0.038 pc, or 1.2 × 1017 cm.

tegration times of about 20 minutes ineach band, the data reach to H/Ks ~ 20and L’ ~ 14, ~ 3 magnitudes deeperthan any Galactic Centre image takenbefore. The dynamic range of theimages is about 13 magnitudes butbright stars are saturated in the deep-est images. The final diffraction limitedresolution was 40, 55 and 95 milli-arcseconds (mas) in the H, Ks and L’-bands (FWHM). The images in Fig-ure 1 demonstrate the complexity of thedense stellar environment in the centralparsec. Bright blue supergiants (in theIRS16 and IRS13 complexes), as wellas red supergiants (IRS 7) and asymp-totic giant branch stars (IRS12N, 10EEand 15NE) dominate the H- and Ks-im-ages. At L’ there is an additional groupof dusty sources (IRS1, 3, 21). Ex-tended L’ emission comes from hot dustin the gaseous ‘mini-spiral’ streamerscomprising the most prominent fea-tures of the SgrA West HII region.The immediate vicinity of SgrA* lacksbright stars and dust. There is a con-centration of moderately bright (Ks ~14) blue stars centred on the radiosource (the ‘SgrA* cluster’). The faint-est sources recognizable on the im-ages are equivalent to ~ 2 M A5/F0main-sequence stars.

It is a highly challenging undertakingto extract photometry, source counts,stellar density distribution and astrome-try in such an extremely dense, high dy-namic range stellar field, and faced withthe complex and time variable, pointspread function (PSF) delivered by

adaptive optics. Our experience from adecade of SHARP data analysis is thatthe best method for disentanglingsources, reducing the influence of theseeing halos of the numerous brightstars and determining fluxes and accu-rate relative positions is to first decon-volve the images with one or several,well-known methods (CLEAN, linear/Wiener filtering, Lucy/Richardson), pri-or to number counts and photometricanalysis. The purpose of the deconvo-lution is not to enhance resolution, butto remove (‘CLEAN’) from the imagesthe effects of the AO PSF. For this pur-pose a mean PSF is extracted fromseveral isolated and moderately brightstars across the field. In the case ofLucy and ‘CLEAN’ the final ‘δ-function’images are re-convolved with a Gauss-ian of the original diffraction limitedFWHM resolution. The agreement ofsources identified with the different im-age analysis techniques is generallyvery good but naturally deteriorateswithin ~ 0.5″ of the brightest stars (andin particular those which are saturatedon our images). In these regions, grain-iness of the seeing halo, ringing and ef-fects of the saturated central pixelsmake source identification of stars fouror more magnitudes fainter than thebright star unreliable. We then identi-fied point sources and carried out pho-tometry with the FIND procedure fromthe IDL Astrolib library. We took theconservative approach of including onlythose sources that were present onboth the Lucy/Richardson and Wiener

deconvolved images, and also both inH and Ks. The final source lists com-prise between 3200 and 4000 stars, de-pending on the FIND extraction param-eters. To improve the reliability of thephotometry, we applied different photo-metric algorithms and, where possible,averaged results from independentdata sets. The final relative photometricerrors are ≤ 0.1 mag below H = 19, Ks= 18 and L’ = 13, but probably twice thatfor fainter stars. The absolute photom-etry is uncertain to 0.15 mag in H andKs, and 0.3 mag at L’. The PSF shapecan vary significantly across a field aslarge as that shown in Figure 1, whichspans a significant fraction of the iso-planatic angle (~ 20″ in radius). Re-markably the anisoplanaticity is rela-tively small for the data shown inFigure 1, indicating that the seeing atthe time of the observations was prob-ably dominated by the surface layers.We have also found no significant im-pact of the anisoplanaticity on the rela-tive positions, at least within the central10″ and to a precision of ~ 5 mas (1/10of the diffraction limited beam, Reid etal. 2003).

We determined incompleteness cor-rections for our images with the well-known technique of first inserting andthen again recovering artificial stars ran-domly across the entire field. Com-pleteness maps were created by divid-ing smoothed maps containing the re-covered artificial sources by smoothedmaps with the initially added artificialstars. The maps thus designate the

3

Figure 1: Left: Ks-band NAOS/CONICA shift-and-add image of the Galactic centre taken in August 2002 (55 mas FWHM resolution, Strehl ra-tio >50%: from Schödel et al. 2002, Genzel et al. 2003). East is to the left, and north is up. The brightest star, IRS7 (6.7 mag) was used asinfrared AO guiding star. The five encircled stars are also radio SiO masers and have been used for establishing the astrometry (Reid et al.2003). The IRS16 and IRS13 complexes of emission line stars are marked. The two arrows denote the position of SgrA*. Right: H/Ks/L’ three-colour composite of the central ~ 20′. Several bright, dusty L-band excess stars are marked. The two arrows denote the position of SgrA*.

probability of recovering a source with agiven magnitude at a given position.The average completeness of our Ks-images is 79%, 63% and 33% at Ks =17, 18 and 19, respectively. However,the completeness varies strongly withposition, and can be very low nearthe brightest stars (e.g. in the IRS16complex). All final source counts andfaint star density distributions were cor-rected for these incompleteness ef-fects.

An important improvement of thenew NAOS/CONICA images is the ac-curate registration of the infrared im-ages, in which the stars are observed,on the astrometric radio images, inwhich SgrA* is observed. For this pur-pose we aligned our NAOS/CONICAimages with an astrometric grid usingall 7 SiO maser sources in the field ofview whose positions are knownthrough interferometric radio measure-ments with the VLA and the VLBA withaccuracies of a few mas (Reid et al.2003, circled in Fig. 1). The SiO masersoriginate in the central 10 AU (~1 mas)of the circumstellar envelopes of brightred giants and supergiants, which arealso present on the infrared images.We were able to improve the radio-to-infrared relative registration by a factorof 3 to ±10 mas from the larger numberof stars with SiO masers in the CONICA

images compared our earlier SHARP/NTT images. In our earlier work we hadonly used two SiO sources for the as-trometric registration. That only allowedsolving for centre position, rotation an-gle and a single pixel scale of the in-frared images. Our new analysis allowssolving, in addition, for second-orderimaging terms. The nonlinear terms arenegligible for NAOS/CONICA, but aresignificant for the SHARP data. Oncethe concurrent infrared and radioframes were aligned at their commonobserved epoch, we used the mean po-sitions of more than a hundred stars tobring all other observed epochs intoalignment by minimizing the residualastrometric errors. We were thus ableto compute stellar positions (as offsetsfrom SgrA* in right ascension and dec-lination) with an accuracy of ±10 masfor all epochs between 1992 and 2002.From our 10-year, SHARP-NAOS/CONICA dataset we are able to deriveproper motions for about 1000 stars inthe central 10″ (Ott et al. 2003). In ad-dition near-IR spectra are available forseveral hundred stars (Genzel et al.1996, 2000, Ott et al. 2003). Finally wewere able to determine individual stellarorbits (or orbital parameters) for 6 starsin the SgrA* cluster (Schödel et al.2002, 2003, Ghez et al. 2000, 2003,Eckart et al. 2002).

Properties of the Nuclear StarCluster: Cusp and Young,Massive Stars

The images and spectra show thatthe nuclear star cluster in the centralparsec is highly dynamic and rapidlyevolving, contrary to expectations. Inparticular, there appear to be manymassive stars that must have formedrecently. We summarize here thesenew results and refer to Genzel et al.(2003) for a more detailed account.

Figure 2 (left) is a plot of the stellarsurface density distribution for faintstars as a function of p, the projectedseparation from SgrA*. The K ≤ 15 starcounts for p(SgrA*) ≥ a few arcsecondscan be reasonably well fit by an isother-mal model of core radius 0.34 pc.However, within a few arcseconds ofSgrA* the NAOS/CONICA data clearlyindicate an excess of faint stars abovethat of the flat isothermal core whosedensity increases toward the centre.The two dimensional distribution (Fig 2.(right inset)) shows that this cusp offaint stars is centred on SgrA*, withinan uncertainty of ± 0.2″. This is in con-trast to the near-IR light distribution(Fig.1), which is centred on the brightstars in the IRS16 complex. From abroken power-law analysis we find thatthe stellar density increases proportion-

4

Figure 2: Surface density distribution of stars (from Genzel et al. 2003). Left: Surface density of stars as a function of projected radius fromSgrA*. Red filled circles are NAOS/CONICA counts (incompleteness corrected) of sources with Ks ≤ 17. Filled blue squares denote direct H-band NAOS/CONICA counts in the central region. Downward-pointing filled green triangles denote scaled SHARP K ≤15 star counts. Theshort-long dash curve is the model of an isothermal sphere of core radius 0.34 pc fitting the outer SHARP counts. The blue continuous curveis the broken power-law (α = 2 beyond 10″ and α = 1.37 within 10″). Open blue circles at the bottom of the figure denote the fraction of latetype stars of the total K ≤ 15 sample with proper motions and Gemini CO narrow-band indices. All vertical error bars are ±1σ, and include thestatistical and incompleteness correction error. Horizontal bars denote the width of the annulus. Right: Two-dimensional map of the smoothedsurface density of NAOS/CONICA incompleteness corrected source counts. Contours are 10, 20,......90, 95, 99 and 99.9% of the peak sur-face density. The maximum of the stellar density is at (RA, Dec) = (0.09″, –0.15″) relative to SgrA*, with an uncertainty of ± 0.2″.

al to R-1.37±0.15 within ~10″ of SgrA*. Thestellar density reaches about 108 M

pc–3 within ~ 0.5″ of the radio source.SgrA* thus is at the centre of the distri-bution of the faint stars. At first glancethis result appears to be in very goodagreement with theoretical expecta-tions. The models predict the existenceof such a cusp with power-law slopesranging from 0.5 to 2, depending on thecusp’s formation scenario and on theimportance of inelastic stellar collisions.However, these models also predictthat the stellar cusp ought to be dy-namically relaxed and consist mostly ofold, low mass stars. This is becausestars migrate toward the central blackhole by two-body relaxation processes.In the Galactic Centre the relaxationtime scale is about 150 Myrs for a tensolar mass star, and scales inverselyproportional to mass. The relaxationtime scale thus is much longer than thelifetime of such a star, unless its massis less than about 2.5 M. We will seein the following that this expectation isnot borne out by the observations. Thedata show a population of bright, ap-parently young stars that is not dynam-ically relaxed.

The central parsec contains twodozen or so, very luminous (105..6 L),blue supergiants. These ‘HeI emissionline stars’ probably are massive(30–100 M), hot (20,000–30,000 K)stars in a short-lived, post-main-se-quence ‘wind’ phase, akin to late type(WN,C8/9) Wolf-Rayet stars and lumi-nous blue variables (LBVs). They canaccount for most or all of the far-in-frared and Lyman-continuum luminosi-ty of the central parsec. Owing to theirshort lifetime, these stars must haveformed in the last few Myr. Consistentwith a recent formation time their prop-er motions are not relaxed, and aredominated by a turbulent rotation pat-tern, with an angular momentum oppo-site to that of Galactic rotation (Genzelet al.1996, 2000). About half a dozen ofthese luminous stars are located in theIRS16 and IRS13 complexes.

The new proper motions now demon-strate that the somewhat fainter earlytype stars (K ≤ 15) are also not relaxed,and hence must be young as well. Thisis shown in Figure 3. While the old, late-type stars show a well mixed distribu-tion of projected angular momenta, theearly type stars exhibit a predominanceof tangential orbits, that is, orbits with avelocity vector on the sky roughly per-pendicular to the projected radius vec-tor from the centre. About 75% of allearly type stars are on projected tan-gential orbits. In the central cusp (p ≤3″) about 60% of the K ≤ 15 early typestars are on clockwise, tangential or-bits. At the same time, the fraction of K≤ 15, late type stars decreases from50% at p ≥ 5″ to < 25% at p ≤ 3″. Un-

relaxed early-type stars thus dominatethe counts of the cusp at moderatelybright magnitudes. Further the K-bandluminosity function (KLF) constructedfrom the new NAOS/CONICA data alsochanges toward the centre. TheGalactic Centre KLF between Ks = 8and 19 is well described by a power lawof slope ~ 0.21, but the p ≤ 9″ KLF inaddition exhibits a prominent excess ofcounts near Ks = 16. This bump is char-acteristic of old, metal rich and lowmass (0.6–3 M), horizontal branch/red clump stars. This bump disappearsin the central few arcseconds, indicat-ing that the cusp lacks such old, lowmass stars. Finally, speckle spectro-photometry and adaptive optics spec-troscopy indicate that many of the fastmoving ‘SgrA* cluster’ stars are earlytype as well, equivalent to late O, earlyB stars with masses of 15–20 M

(Genzel et al. 1997, Gezari et al. 2002,Ghez et al. 2003).

The observations portray a fascinat-ing, albeit complicated picture of thenuclear star cluster. The central cuspappears not to be dominated by old,low mass stars. The presence of themany un-relaxed, massive stars sug-gests a highly dynamic picture. For theHe I emission-line stars and the otherearly stars in the IRS16/13 complexesthere are two plausible explanations.One postulates episodic infall and com-pression of very dense gas clouds, fol-lowed by in situ formation of stars. Theother is the rapid sinking toward thecentre by dynamic friction of a mas-sive young star cluster that originallyformed at 5–10 pc from the centre. Bothmechanisms appear feasible but re-

quire very special, and somewhat un-likely conditions.

For the innermost early type stars inthe SgrA*-cluster both of these scena-rios fail. Instead, and owing to the veryhigh stellar densities estimated above,continuous formation of moderately mas-sive stars by collisions and mergers oflower mass stars appears to be themost plausible mechanism. The SgrA*cluster stars thus may be blue strag-glers. The stellar collisional model isalso supported by the simultaneousdisappearance of the late type giantsand the lower mass horizontal branch/red clump stars in the central cusp.

A Star in a 15-Year Orbit Around SgrA*

Without any doubt the most excitingaspect of the new NAOS/CONICA datahas been the first determination of stel-lar orbits around SgrA* (Schödel et al.2002, 2003). The initial measurementsof the orbital accelerations of S1 andS2, the two stars closest to SgrA*, wereconsistent with bound Keplerian orbitsaround a 3 million solar-mass centralobject but still allowed a wide range oforbital parameters (Ghez et al. 2000).Specifically, possible orbital periods forS2 ranged from 15 to 500 years. Whenwe obtained the first NAOS/CONICAimages in March 2002, it immediatelybecame obvious that S2 had moved towithin10 mas of the radio source, andwas now located east of SgrA*, that is,on the opposite side when comparedto its position in the preceding years(see figure on cover page, left inset).This implied that S2 was just passing

5

Figure 3: Spatial distribution of stars with different line-of-sight, normalized angular momen-ta (Jz/Jz(max)), obtained from the K ≤ 15 proper motion data set of Ott et al. (2003, see Genzelet al. 2003). Stars with mainly tangential, clockwise motions on the sky are marked as filledblue circles, stars with tangential, counter-clockwise motions are marked as crossed, redopen squares, while stars with motions mostly along the projected radius vector are markedas grey crosses. Keep in mind that these are sky-projected motions. The right inset showsthe distribution of spectro-photometrically identified, late type stars, while the left inset showsthe same for early type stars.

through its peri-centre approach andthat the S2 orbital data might probe thegravitational potential at a radius ofabout one light-day, 20 to 30 times fur-ther in than our earlier, statistical deter-minations. When more data came in,this exciting interpretation was borneout. By late May 2002 it was clear thatthe SHARP and NAOS/CONICA posi-tional measurements of S2 between1992 and 2002 sampled two thirds of ahighly elliptical orbit around the astro-metrically determined position of SgrA*(Reid et al. 2003).

S2 orbits the compact radio sourceas the planets orbit our Sun. The datashown in the right inset of Figure 4 con-strain all orbital parameters with betterthan 10% accuracy (Schödel et al.2002, 2003). At its peri-approach inApril 2002, the star was 17 light-hoursfrom SgrA* at which point it moved with~ 8000 km/s. The rapid movement ofthe star could also be directly verifiedfrom the positional changes from monthto month as more images were ac-quired (see Fig. 4). From the inferred

Please note that Figure 4 is shown onpage 1.

orbital period of 15.2 years and its semi-major axis of 4.4 mpc, Kepler’sthird law implies an enclosed mass of3.3 ×106 M. This value is in excellentagreement with all earlier mass meas-urements between ~ 15 light-days andseveral light-years from the centre(Fig. 5). Despite its close approach tothe central mass, S2 seems to havesurvived this encounter without any prob-lems. The peri-centre distance radius ofS2 is 70 times greater than the distancefrom the black hole where the starwould be disrupted by tidal forces(about 16 light-minutes for a 17 M).Since tidal energy deposition falls fasterthan the sixth power of the ratio of tidalradius to orbital radius, tidal effects areexpected to be negligible, consistentwith its lack of infrared variability.

Schödel et al. (2003) have recently re-analyzed the entire 1992–2002 SHARP-NAOS/CONICA image data set. In thisanalysis, initial positions and fluxes of amodel cluster were first fitted to the best(August 2002) NAOS/CONICA image.Subsequently, positions at other epochswere obtained by using the precedingepoch as an initial guess. By iterativelyfitting and subtracting from the imagesfirst the brighter and then ever fainterstars this analysis allows tracing the po-sitions of the fainter stars even at epochswhen they move close to bright stars.Ten isolated reference stars served asanchors to align centre position and ro-tation angle for each epoch. By a rigor-ous selection of the very best imagingdata and intensive image processingSchödel et al. obtained high-resolutionmaps of the central 1.5″ around SgrA*

down to K ~16. They find six stars in thisregion that show clear signs of ac-celerations and allow determining/constraining their orbital parameters(Fig. 6). In addition to S2, good orbitalfits are also obtained for S12 and, to alesser extent, for S14 (see below), S1and S8. The resulting mass constraintsare shown in Figure 5. They fully con-firm but do not significantly improve theconstraints obtained from the S2 orbitalone. The orbital periods in the centralarcsecond range between 15 and a fewhundred years. Most of the orbits ap-pear to have a moderate to high ec-centricity (‘radial’ orbits). Using the or-bital accelerations of four of the six orbitsin Figure 6, Schödel et al. find from amaximum-likelihood analysis that thecentre of gravitational force on all orbitsis at the radio position of SgrA* to with-

in the combined uncertainty of 20 mas(Fig. 6, right inset).

The Keck group very recently report-ed their first orbit analysis of the SgrA*cluster stars as well (Ghez, priv.comm., Ghez et al. 2003). Their resultsare in excellent agreement with ours,and in some cases improve the con-straints significantly. Particularly ex-citing is the star S0-16 (S14 in our de-nomination, which has a very high ec-centricity (0.97) and approached SgrA*in late 1999 to within about 11 light-hours (see also Fig. 6). Ghez et al.(2003) were also able to extract a radi-al velocity of S2 from detection of blue-shifted Brγ absorption. This resolvesthe 180o inclination ambiguity of the or-bit and shows that the angular momen-tum of S2’s orbit is counter to that ofGalactic rotation.

6

Figure 5: Derived mass distribution in the Galactic Centre (for an 8 kpc distance: Schödel etal. 2002, 2003). The filled black circle denotes the mass derived from the orbit of S2, the pinkfilled square the mass derived from the orbit of S12, and the filled green circle the mass de-rived from the orbit of S14 (Schödel et al. 2003). Error bars combine statistical (fit) and sys-tematic/astrometry errors. Filled dark green triangles denote Leonard-Merritt projected massestimators from a new NTT proper-motion dataset by Ott et al. (2003), separating late- andearly-type stars, and correcting for the volume bias in those mass estimators by scaling withcorrection factors (0.9–1.1) determined from Monte Carlo modelling of theoretical clusters. Anopen triangle denotes the Bahcall-Tremaine mass estimate obtained from Keck proper mo-tions (Ghez et al. 1998). Light-blue, filled rectangles are mass estimates from a parameter-ized Jeans-equation model, including anisotropy and differentiating between late- and early-type stars (Genzel et al. 2000). Open circles are mass estimates from a parameterized Jeans-equation model of the radial velocities of late-type stars, assuming isotropy (Genzel et al.1996). Open red rectangles denote mass estimates from a non-parametric, maximum likeli-hood model, assuming isotropy and combining late- and early-type stars (Chakrabarty &Saha 2001). Black letter ‘G’ points denote mass estimated obtained from Doppler motions ofgas (see Genzel & Townes 1987 and references therein). The black continuous curve is theoverall best fit model to all data. It is a sum of a 2.9(± 0.15)×106 M point mass, plus a stel-lar cluster of central density 3.9 × 106 M/pc3, core radius 0.34 pc and power-law index α =1.8. The grey long dash-short dash curve shows the same stellar cluster separately, but foran infinitely small core (i.e. a ‘cusp’). The thick dashed curve is a sum of the visible cluster,plus a Plummer model of a hypothetical concentrated (α = 5), very compact (R0 = 0.00015pc) dark cluster of central density 2 × 1017 M pc–3.

Figure 6: Stellar orbitsin the SgrA*-cluster(Schödel et al. 2003).Left inset: Stellar or-bits of 6 stars, ob-tained from an analy-sis of the 10-yearSHARP-NACO data-set. These innermoststars are on fairly el-liptical orbits, with or-bital periods between15 and a few hundredyears. The right insetshows a comparisonof the best maximum-likelihood distributionof the centre of gravi-tational force deter-mined from the S1,S2, S12 and S8 orbits(grey scale, with 1, 2,3 σ contours), with thenominal radio positionof SgrA* (white crosswith 1σ error circle).Radio position andcentre of force agreeto within 20 mas.

Constraints on the Nature of SgrA*

The new Galactic Centre data provebeyond any doubt that the gravitationalpotential resembles that of a pointmass down to a scale of 10 light-hours.All available data are very well fit by thesuperposition of a 2.9 (±0.2) × 106 M

point mass, plus the extended massdistribution of the visible stellar cluster.The contribution of the extended stellarcluster around SgrA* to the total masscannot be more than a few hundred so-lar masses within the peri-centre dis-tance of the orbit of S2. If the pointmass is replaced by a hypothetical ex-tended (dark cluster) mass distribution,its density must exceed 2.2 × 1017 M

pc–3, more than ten orders of magni-tude greater than the density of the vis-ible cluster at ≥ 1″ from the centre. Inaddition, such a dark cluster must havea well-defined surface with a very steepdensity drop-off outside of it, in order tofit the flat mass distribution over threeorders of magnitude in radius. If such ahypothetical dark cluster consists oflow-mass stars, neutron stars, or stellarblack holes, the inferred density impliesa lifetime less than a few 105 years(Maoz 1998). This is obviously a highlyimprobable configuration. All stars in itsvicinity are much older. Theoretical sim-ulations of very dense, core collapsedclusters predict much shallower, nearisothermal density distributions, incon-sistent with the flat mass distribution.Such a dark cluster model can now besafely rejected. Our new data also ro-bustly exclude one of two remaining,‘dark particle matter’ models as alterna-tives to a supermassive black hole,

namely a ball of heavy (10–17 keV/c2)fermions (sterile neutrinos, gravitinos oraxinos) held up by degeneracy pres-sure. Such a fermion ball could in prin-ciple account for the entire range ofdark mass concentrations in galacticnuclei with a single physical model(Tsiklauri & Viollier 1998). Because ofthe finite size (~ 0.9″) of a 3 · 106 M

ball of ~ 16 keV fermions, the maximum(escape) velocity is about 1700 km/sand the shortest possible orbital periodfor S2 in such a fermion ball modelwould be about 37 years, clearly incon-sistent with the orbits of S2 and S14.The enclosed mass at the peri-centresof S2 and S14 would require neutrinomasses of > 58 and > 67 keV. Apartfrom the fact that the existence a ofmassive neutrino is not favoured bycurrent neutrino experiments, the fermi-on ball model can now be safely ex-cluded from the above constraints if themodel is to explain the entire range ofobserved dark masses in galaxy nucleiwith a single particle. The only dark par-ticle matter explanation that cannot beruled out by the present data is a ball ofweakly interacting bosons. Such a con-figuration would have a radius only sev-eral times greater than the Schwarz-schild radius of a black hole (Torres etal. 2000). However, it would be veryhard to understand how the weakly in-teracting bosons first manage to reachsuch a high concentration (Maoz 1998),and then avoid collapsing to a blackhole afterwards by baryonic accretion(Torres et al. 2000). Such accretion def-initely occurs in the Galactic Centre ata rate of 10–7 to 10–5 M yr–1.

Figure 7 summarizes all current con-straints on the size and density of the

dark mass in the Galactic Centre. In thelast 15 years, gas and stellar dynamicshave definitely proven the existence ofa dark, non-stellar mass concentrationof 3 million solar masses. Over this timeperiod the observational constraintshave pushed its size downward by 2.5orders of magnitude, and its density up-ward by 9 orders of magnitude. RadioVLBI (Doeleman et al. 2001) and X-rayobservations (Baganoff et al. 2001)have shown the existence of hot andrelativistic gas on scales of tenSchwarzschild radii at the centre of thismass concentration. VLA/VLBA dataconstrain the proper motion of SgrA* inthe Galactic Centre frame to be lessthan 20 km/s (Reid et al. 1999).Comparison of this limit to the >103

km/s velocities of the stars orbitingSgrA* then implies that > 98% of themass enclosed inside the orbit of S2 isassociated with the radio source, onscales of 10 light-minutes, setting alower limit to the density of 5 × 1021 M

pc–3 (Reid et al. 1999, Chatterjee et al.2002). In the allowed upper left cornerof parameter space there remain twopossible configurations, one of which(the boson star) is purely speculativeand also highly unstable. It thus seemssafe to conclude that SgrA* must be a3 million solar mass black hole, beyondany reasonable doubt. The black holeappears to be fairly ‘naked’ (see theconstraints above on any additional ex-tended emission surrounding it). Itcould in principle be a tight (< 10 lighthour separation) binary black hole.However, an equal mass, binary holewould coalesce by gravitational radia-tion in a few hundred years. The obser-vation of radial velocity anisotropy in

7

the cluster of stars closest to SgrA*(Schödel et al. 2003) appears also tofavour a single black hole: A strong tan-gentially anisotropic velocity dispersionwould be expected near a binary blackhole. The Galactic Centre thus has pre-sented us with the best astrophysicalevidence we now have that the objectspredicted by Einstein’s and Schwarz-schild’s theoretical work are in fact re-alized in the Universe.

While the first year of NAOS/CONI-CA observations have led to an unex-pected breakthrough in the studies oforbits, not much has been learned asyet about the emission properties ofSgrA* itself. The very fact that thebright star S2 was so close in 2002 tothe central source has preventedmeaningful observations of its emis-sion. The deduced upper limits toSgrA*’s flux in the H, Ks and L’ bandsare somewhat lower than previous ob-servations but are broadly consistentwith current low accretion CDAF,ADIOS or jet models of the source(Melia & Falcke 2001).

Outlook

The NAOS/CONICA work we havepresented here is just the first step in a

and the Large Binocular Telescope.These observations will provide a fewto 10 mas (a few light-hours) resolutionand offer exciting prospects for the ex-ploration of relativistic motions at10–100 Schwarzschild radii from thecentral black hole.

Acknowledgements

We are grateful to the ESO DirectorGeneral, Catherine Cesarsky, and theDirector of Paranal Observatory, Ro-berto Gilmozzi for making these obser-vations possible. We also thank theNAOS and CONICA team members fortheir hard work, as well as the staff ofParanal and the Garching Data Man-agement Division for their support ofthe commissioning and science verifi-cation. We are grateful to B. Schutz fora discussion of the lifetime of a hypo-thetical binary black hole.

References

Baganoff, F.K., et al., 2001, Nature, 413, 45.Chakrabarty, D. & Saha, P. 2001, AJ, 122,

232.Chatterjee, P., Hernquist, L. & Loeb, A.

2002, ApJ, 572, 371.Eckart, A. & Genzel, R., 1996, Nature, 383,

415.Eckart, A., & Genzel, R., 1997, MNRAS,

284, 576.Eckart, A., Genzel, R., Ott, T. & Schödel, R.,

2002, MNRAS, 331, 917.Genzel, R., & Townes, C.H., ARAA, 25, 377.Genzel, R., Thatte, N., Krabbe, A., Kroker H.

& Tacconi-Garman, L. E. 1996, ApJ, 472,153.

Genzel, R., Eckart, A., Ott, T. & Eisenhauer,F. 1997, MNRAS, 291, 21.

Genzel, R., Pichon, C., Eckart, A., Gerhard,O. & Ott, T., 2000, MNRAS, 317, 348.

Genzel, R. et al. 2003, in prep.Gezari, S., Ghez, A. M., Becklin, E.E. ,

Larkin, J., McLean, I.S., Morris, M. 2002,ApJ, in press (astro-ph/0205186).

Ghez, A.M., Klein, B.L., Morris, M., &Becklin, E.E. 1998, ApJ, 509, 678.

Ghez, A.M., Morris, M., Becklin, E.E.,Tanner, A. & Kremenek T., 2000, Nature,407, 349.

Ghez, A. et al. 2003, in prep.Maoz, E., 1998, ApJ, 494, L181.Melia, F. & Falcke, H. 2001, ARAA, 39, 309.Ott, T., Genzel, R., Eckart, A. & Schödel, R.,

2003, in preparation.Reid, M.J., Readhead, A.C.S., Vermeulen,

R.C. & Treuhaft, R., 1999, ApJ, 524, 816.Reid, M. J., Menten, K.M, Genzel, R., Ott, T.,

Schödel, R. & Eckart, A. 2003, ApJ, inpress.

Schödel, R. et al., 2002, Nature, 419, 694.Schödel, R., Ott, T., Genzel, R. & Eckart, A.

2003, in prep.Torres, D.F., Capozziello, S. & Liambase, G.,

2000, Phys. Rev. D., 62, id.104012.Tsiklauri, D. & Viollier, R.D., 1998, Ap.J.,

500, 591.

8

new phase of near-infrared observa-tions of the immediate surroundings ofthe central dark mass at the nucleus ofthe Milky Way. The observation ofKeplerian orbits offers a clean new toolof high precision gravitational studies.Stellar orbits also allow an accurate de-termination of the distance to theGalactic Centre. With NAOS/CONICAwe are presently able to distinguishmore than 30 individual stellar sourceswithin 0.5″ (~ 23 light-days) from SgrA*.For a number of them we hope to de-termine high precision orbits in the nextfew years. Some of them may ap-proach the black hole within a few light-hours, with velocities approaching 10%of the speed of light and orbital timescales of less than a few years. In ad-dition to imaging, spectroscopy will re-veal the properties of the stars bound tothe hole, and also give additional infor-mation on the gravitational potentialand on interactions between stars.Time-resolved photometry and po-larimetry will allow studies of the accre-tion flow and gravitational lensing ef-fects. Another order of magnitude im-provement in spatial resolution will be-come possible with infrared interferom-etry at the VLTI (once the PRIMA facili-ty is available), the Keck interferometer

Figure 7: Constraints on the nature of the dark mass concentration in the Galactic Centre(from Schödel et al. 2003). The horizontal axis is the size (bottom in parsec, top inSchwarzschild radii), and the vertical axis density (left in solar masses per pc3, right in gcm_3). Red filled circles denote the various limits on the size/density of the dark mass dis-cussed in this paper. In addition the grey shaded area marks the constraints on the size ofthe variable X-ray emission from Baganoff et al. (2001). Large filled circles mark the locationof different physical objects, including the visible star cluster and its central cusp, the heavyneutrino, fermion ball of Tsiklauri & Viollier (1998), the boson star of Torres et al. (2000) and,in the top left, the position of a 2.8 × 106 M black hole. In addition three thin horizontal linesmark the lifetimes of hypothetical dark clusters of astrophysical objects (neutron stars, whitedwarfs, stellar black holes, rocks etc., Maoz 1998).

size (Rs(2.6 ×106M))d

ensi

ty (

M

pc–3

)

size (pc)

den

sity

(g

cm–3

)

1. NAOS-CONICA – Near-InfraredCamera with Adaptive Optics

NAOS-CONICA is a near-infraredimaging camera and spectrograph(CONICA) attached to an adaptive optics(AO) system (NAOS) for wavefront cor-rections (Lenzen et al. 1998, Rousset etal. 2000). The AO system is designed todeliver diffraction-limited observations atan 8-metre-class telescope. NAOS isequipped with two Shack-Hartmannwavefront sensors operating at visualand NIR wavelengths, respectively.Most existing AO systems feed near-in-frared (NIR) cameras, but operate at vi-sual wavelengths. In addition to the factthat the spectral energy distribution ofbright stars peaks at optical wave-lengths, visual detectors offer highersensitivity and lower read-out noisethan NIR detectors, thus allowing oneto use fainter targets for wavefrontsensing. The restriction to visual refer-ence sources, however, excludesdeeply embedded objects located in re-gions with active star formation or highforeground extinction such as theGalactic Centre. With wavefront sen-sors operating both at visual and NIRwavelengths, NAOS combines the ad-vantages of visual, faint reference tar-gets with the option to penetrate highlyobscured regions containing bright in-frared sources. NAOS-CONICA wascommissioned on Yepun (UT4) in thecourse of Period 69. For more detailedinformation on NACO performance andfirst light, see Brandner et al. (2002) inThe Messenger 107.

We have used the CONICA JHK im-aging mode with the S27 camera incombination with the NAOS visualwavefront sensor for performance testsin the crowded stellar field of the dense,massive star cluster Arches. The S27camera with a pixel scale of 27 milliarc-second (mas) and corresponding fieldof view (FOV) of 27″ × 27″ is optimizedfor diffraction-limited observations inthe J to K band wavelength range. At2.2 µm, the diffraction limit of an 8-mtelescope is 57 mas, such that the S27camera allows Nyquist sampling.

2. Why Arches?

The Arches cluster is one of the fewstarburst clusters found in the Milky

Way. With an age of only ∼ 2–3 Myr itcontains a rich massive stellar popula-tion with about 150 O stars located inthe cluster centre (Figer et al. 1999),and several more found in the immedi-ate vicinity. The core density has beenestimated to be 3 ⋅ 105 M pc–3, and thetotal stellar mass exceeds 104 M

(Figer et al. 1999). The massive clustercentre contains ∼ 12 Wolf-Rayet stars(Cotera et al. 1996, see also Blum et al.2001), intertwined with an intermediate-mass main-sequence population. At thefaint end, where pre-main-sequencestars would be expected, the stellarpopulation is severely contaminated bybulge stars, as Arches is located in thecentral region of the Milky Way at a pro-jected distance of only 25 pc (roughly10 arcminutes) from the GalacticCentre (Fig. 1).

With a distance of about 8 kpc and aforeground extinction in the visual of 24to 34 mag, Arches is a very challengingobject. The high foreground extinctionobscures the faint cluster population atvisual wavelengths, so that NIR obser-vations, where extinction is less severe,are necessary to detect the cluster pop-ulation. Given these extreme proper-ties, and the existence of data sets ob-tained with HST/NICMOS and theGemini/ Hokupa’a AO system allowinga detailed comparison, the Arches clus-ter was chosen as NAOS-CONICAcommissioning target to test the NACOperformance on a dense, crowding-lim-ited stellar field.

3. NAOS-CONICACommissioning Data

Arches H and Ks images were ob-tained during Commissioning 3 ofNAOS-CONICA in March and April2002. 20 exposures with integrationtimes of 1 min (2 × 30 s detector inte-gration time (DIT)) each were obtainedin H, and 15 × 1 min exposures in Ks (4× 15 s DIT). As the NIR wavefront sen-sor was not available during Comm 3, ablue foreground giant with V = 16 mag(V–R ∼1 mag) served as the referencesource for the visual wavefront sensor.Although the brightest star in theArches field, this star is already close tothe limiting magnitude (∼17 mag) of thevisual wavefront sensor. As has beendemonstrated on the Galactic Centre

field (Schödel et al. 2002 and Ott et al.in this issue), NIR wavefront sensingwith NIR bright reference targets (K ∼10mag) is capable of improving the per-formance in a highly extincted region.

All images were taken in auto-jittermode. The auto-jitter template allowsautomatic, random offsets between in-dividual exposures to facilitate back-ground subtraction and avoid after-glows at the position of bright sources.A set of additional exposures with shortdetector integration times were taken toobtain photometry of the bright clusterstars (8 × 1 s DIT in H, 8 × 0.5 s DIT inKs). Seeing conditions varied during theArches observations. In H-band, theseeing ranged from 0.45 to 0.85, whileduring K-band observations the seeingdegraded from 0.8 to 1.3. On April 4,the moon was only 6° from our target,and increased background fluctuationsdecreased the detectability of faintsources in the cluster vicinity.

The data were reduced within IRAF,using twilight flat fields taken duringComm 3 and combined with the nacoptwflat algorithm in eclipse (Devillard1997). Eclipse is the ESO pipeline re-duction package, which allows specialtreatment of particular instrument char-acteristics adapted to the availableESO instrumentation. Sky frames werecreated from the cluster observationsthemselves, which is not ideal, but wassufficient for our technical study. Star-light was excluded from these skies byrejecting the bulk of the bright pixels. Asthe CONICA detector has a significantfraction of time-varying hot pixels, acosmic ray rejection routine was usedto obtain a bad pixel mask. This maskwas applied during the drizzlingprocess, in which images are shiftedand combined. The 7 exposures withthe highest resolution were combinedin Ks to one 7 min integrated frame, andthe 14 best exposures in H, where see-ing conditions were much more stable,could be used to obtain a 14 min H-band image. During the K-band obser-vations, the AO performance changedwith time as the seeing worsened. Ascrowding is the limiting factor on theArches field, the gain in resolution byrejecting the large fraction of frameswith lower quality supersedes the lossin photometric depth due to the shorterresulting total integration time on the fi-

9

NAOS-CONICA Performance in a Crowded Field– the Arches Cluster A. STOLTE1, W. BRANDNER1, E.K. GREBEL1, DON F. FIGER2, F. EISENHAUER3, R. LENZEN1, Y. HARAYAMA3

1Max-Planck-Institut für Astronomie, Heidelberg, Germany 2Space Telescope Science Institute, Baltimore, USA3Max-Planck-Institut für Extraterrestrische Physik, Garching, Germany

nal image (see also Tessier et al. 1994).The best K-band photometry was ob-tained when combining only the frameswhere the best AO performance wasachieved. The resulting images haveStrehl ratios of 14% in H and 20 % in K,yielding a spatial resolution of ∼ 85 masin both filters.

Standard DAOPHOT PSF fitting pho-tometry was used to obtain the finalmagnitudes. A constant PSF functionhas been used, as isolated PSF starsare rare in the dense Arches field. APenny function, which has a Gaussiancore with elongated Lorentzian wings,yielded the best fit to the data.

The NACO data were calibratedusing existing HST/NICMOS photom-etry of the same field (Figer et al. 1999).All calibrated data shown below arein the NICMOS photometric system.This allowed a large number of stars tobe used for zero-pointing and for thestudy of photometric residuals over thefield. We find the NACO performance tobe very stable over the field. No trendof the amplitude of photometric residu-als with field position is observed. Notethat the measurement of anisopla-natism on the Arches field is complicat-ed by the high stellar density and large

density gradient over small spatialscales.

4. Luminosity Functions

In Figure 2, luminosity functions(LFs) of the NACO commissioning dataare compared to HST/NICMOS LFs ofthe same field. Arches was observedwith NICMOS2 for 256 s integrationtime in the broadband filters F110W,F160W and F205W corresponding to J,H, and K. The NICMOS detection limitswere approximately 21 mag in H(F160W), and 20 mag in K (F205W).Within a factor of 2 longer integrationtime in K and a factor of 3.3 in H, NACOreaches detection limits of 22 mag in Hand 21 mag in K, one magnitude deep-er than the NICMOS observations.Note that NICMOS does not have tofight the large NIR background from theEarth’s atmosphere, but also has themuch smaller collecting area of a 2.4-mmirror as opposed to Yepun’s 8-m mir-ror.

Only objects detected in both filtersare included in the NACO LF, as the re-quirement that objects have to be de-tected in both filters is a very efficientmethod to reject artefacts. In particular

in a very dense stellar field, backgroundfluctuations and saturation peaks fre-quently cause false detections.

As sources in the Arches cluster arehidden behind ∼ 30 mag of foregroundextinction, reddening influences the de-tection of stars more severely at short-er wavelengths. Consequently, the H-band LF shown in Fig. 2 represents thephotometric performance more realisti-cally, as the K-band LF is artificiallytruncated by the requirement thatsources have to be detected in H aswell.

While the field star LF is comparablefor NACO and NICMOS down to theNICMOS detection limit of H ∼ 21 mag,NACO clearly gains a large number ofsources in the cluster centre. Here, thesignificantly better resolution of 85 maswith VLT/NACO vs. 210 mas withHST/NICMOS allows us to resolve thecluster centre for the first time. Notethat the diffraction limit of the 2.4 mHST mirror of 183 mas in H and 210mas in K does not allow a much higherresolution with NICMOS, while betterStrehl ratios may yield resolutionsdown to the VLT diffraction limit of 57mas in H and 69 mas in K with NAOS-CONICA. The cluster centre LF dem-

10

Figure 1: A JHK composite 2MASS image of the Galactic Plane, including the ArchesCluster and the Galactic Centre, is shown in the left panel. The Arches cluster isthe little blob of stars pointed out in the North of the image. The CONICA field ofview (FOV) is indicated. The NAOS-CONICA H and Ks two-colour composite takenduring Commissioning 3 is shown in the right panel. The FOV of this mosaic is25″ × 27″, approximately the FOV of the CONICA S27 camera. The dense popu-lation in the Arches cluster centre has been resolved with NACO for the first time.

onstrates not only the deeper photome-try with NACO, but also the strongly en-hanced resolution, as in nearly eachmagnitude bin above H = 15 mag morestars are resolved. This affects statisti-cally derived properties of the Archesstellar population, such as stellar densi-ty, total mass and the slope of the massfunction.

5. Colour-Magnitude Diagrams

A comparison of colour-magnitudediagrams (CMDs) of the Arches clustercentre is shown in Figure 3. Here,Gemini/Hokupa’a low Strehl data (2.5%in H, 7% in K, Stolte et al. 2002),HST/NICMOS observations (Figer etal. 1999), and the new NACO observa-tions are compared. The Gemini datacover the smallest field of view, andcare was taken to select only the areain common to all data sets to allow di-rect comparison. The NACO and NIC-MOS data were selected accordingly.

While the low-Strehl Hokupa’a datashow a large scatter in the main se-quence due to the contamination ofstellar fluxes with extended seeing ha-los, the NACO data show a well-con-fined, narrow main sequence. The NIC-MOS data are also very well confined,but in accordance with the results fromluminosity functions the NACO mainsequence appears more densely popu-lated. Derived physical and statisticalproperties of the cluster population areconsequently less biased by crowdinglosses.

The Hokupa’a data set has the high-est incompleteness, being 90% com-plete only down to K = 15 mag. Whenthe number of stars down to this mag-nitude limit is compared, we find 131stars with Hokupa’a, 157 stars withNICMOS within the same field, and 169stars with NACO. The 12 stars unde-tected with NICMOS are either fainter

neighbours vanishing in the PSF pat-tern of a bright star, or close pairs whereonly one component has been re-solved. Figure 4 shows the comparisonof resolved sources with K < 15 mag ona sample field in the cluster centre.

6. Mass Function

The NACO CMD was used to derivethe mass function (MF) of the Archescluster population. A 2 Myr isochronewith twice solar metallicity (see Figer etal. 1999, Stolte et al. 2002 for discus-

sion) from the Geneva set of models(Lejeune & Schaerer 2001) was used totransform K-band magnitudes into stel-lar masses.1 A distance modulus of14.52 mag and extinction of AV = 24mag, corresponding to the cluster cen-tre, have been applied. Magnitudes

11

Figure 2: H-band luminosity function of the Arches cluster. The left LF contains only objects at a distance R > 10″ from the cluster centre, dis-playing the field performance in comparison to HST/NICMOS, while the right panel shows the cluster centre regions with stars within R < 5″.While the field performance between NACO and NICMOS is comparable until the NICMOS detection limit of H ∼ 21 mag is approached, thecluster centre LF shows the large gain in the NACO data due to the higher spatial resolution. It is particularly intriguing that already bins fainterthan H = 15 mag are affected. This indicates that the cluster centre population was until now never truly resolved, not even at the brighterend.

Figure 3: Arches colour-magnitude diagrams. A comparison between VLT/NAOS-CONICA,HST/NICMOS and Gemini/Hokupa’a data is shown. While the low Strehl ratio of the Hokupa’adata of only a few per cent severely limits resolving the cluster population and caused largephotometric uncertainties, the moderate Strehl NAOS-CONICA observations exhibit a nar-row, well-defined main sequence, reflecting the photometric quality of the data. The NICMOSdata are limited by the small telescope diameter of only 2.4 m, and corresponding larger dif-fraction limit. In addition, the strong diffraction pattern in the NICMOS PSF hampers the de-tection of faint sources in the cluster centre. Despite the moderate Strehl ratio, the NACOCMD clearly displays the advantages of AO at an 8-m-class telescope. The right axis of theNACO CMD is labelled with present-day masses from a 2 Myr main sequence isochrone usedto derive the mass function.

1Note that the current best age estimate from high-resolution spectra of the bright Arches main se-quence O stars is 2.5 Myr (Figer et al. 2002), butwe prefer to use the 2 Myr isochrone to allow directcomparison of the NACO MF with our prior resultsgiven in Figer et al. 1999 and Stolte et al. 2002.

have been transformed to the NICMOSsystem using the colour equations de-termined by Brandner et al. 2001. A lin-ear change in extinction is observedwhen moving outwards from the clustercentre. While the innermost 5″ of thecluster population are consistent with aconstant reddening of AV = 24.1 mag,the extinction increases with increasingradial distance at radii R > 5″. The ob-served H-K colour as well as the K-band magnitude have been correctedfor this trend before transformation intostellar masses. A colour cut with 1.35 <(H-K)corr < 1.77 mag was applied to se-lect Arches main sequence stars fromthe corrected CMD.

The present-day mass function forstars in the Arches cluster derived fromthe NACO CMD (Fig. 3) is shown inFigure 5. A power law can easily be fit-ted to the upper end of the MF for M >10 M, below which the MF appears tobecome flat. When regarding the entireNACO field, it appears that at this pointfield contamination sets in, as this flatportion of the MF is very pronounced inMFs created with the above procedureat large distances from the cluster cen-tre. Analysis of Galactic Centre fields inthe vicinity of Arches is required to esti-mate the relative fractions of field andcluster stars in the faint regime. As thecombined field of view obtained duringComm 3 does not cover sufficient areaat large enough distances from thecluster to perform statistical back-ground subtraction, the MF discussedhere focuses on the massive end withM > 10 M, where field contaminationis negligible.

To eliminate the dependence of bin-ning on the derived MF slope, the start-ing point of the first mass bin has beenshifted by δlog(M/M) = 0.01 at the

chosen bin width of ∆log(M/M) = 0.1,resulting in ten MFs with consecutivestarting points. Averaging the resultantMF slopes yields an MF slope of Γ =–0.91 ± 0.15 for the Arches main se-quence, where Salpeter is Γ = –1.35(Salpeter 1955). Although the NACOLF is more complete, this slope is iden-tical to the slope derived from NICMOSdata before (Stolte et al. 2002). This in-dicates that at the bright end (K < 16mag, M > 10 M), all bins contributingto the LF and MF gain the same fractionof stars, such that the relative fractionof stars in each bin remains the same.Note that we only consider the massiveend of the MF, up to K = 16 mag, suchthat the faint end of the LF, where thelargest discrepancy to the NICMOS LFis observed, is not taken into accounthere, as this would require field sub-traction.

This slope is slightly flatter than theaverage slope of Γ = –1.1 ± 0.3 found instar forming regions in the Milky Way(Massey et al. 1995). A comparably flatslope is seen in the MF of the massivestarburst cluster in NGC 3603, which hasphysical properties similar to Arches,although forming in a much more mod-erate star-forming environment in theCarina spiral arm. In NGC 3603, the MFslope is found to be Γ = –0.7 in themass range 1M < M < 30 M

(Eisenhauer et al. 1998), close to theArches MF slope. The flat slope indi-cates that the Arches cluster was eithervery efficient in forming massive stars,or that dynamical mass segregationhas led to the ejection of a significantnumber of intermediate-mass stars. Ifthe cluster is mass segregated (eitherdue to primordial or due to dynamicalprocesses), this should be particularlypronounced in the cluster centre.

The cluster centre MF is shown inFigure 5 (right). Only stars within the in-nermost 5″ of the cluster centre are in-cluded. The binning width was nowchosen to be ∆log(M/M) = 0.2 for sta-tistical reasons. Again, only the brightend of the MF was fitted for comparisonpurposes. The average slope derivedfor the cluster centre with the samemethod as described above is Γ = –0.6± 0.2, flatter than the slope for the en-tire cluster. This is consistent with mas-sive star cluster formation scenarios,where primordial or rapid dynamicalsegregation causes massive stars toend up in the cluster centre, while alarger fraction of low-mass stars isfound in the outskirts.

Although the entire mass range in thecluster centre appears to be consistentwith a very flat MF (Γ ∼ 0) down tolog(M/M) ∼ 0.5, M = 3 M, where theK-band LF is more than 75% complete(derived from stars recovered in both Hand K), it is currently not clear how theinterplay of incompleteness correctionand field contamination will alter theshape of the MF for masses below 10M. Neighbouring Galactic Centrefields and thorough incompletenesstesting will be required to obtain ameaningful slope in the cluster centredown to the low-mass regime.

7. Outlook

Velocity studies of the bright clusterstars in the vicinity of Arches are nec-essary to draw final conclusions on theformation scenario. Stars dynamicallyejected during interaction processesshould obtain large velocities directedaway from the cluster centre. Such avelocity study can yield new insightsinto the cluster dynamics as well as for-

12

Figure 4: Sources brighter than 15 mag, the Hokupa’a 90% completeness limit, overlaid on the 7 min NACO K-band image (9 ′ × 9 ′FOV, leftpanel). For comparison, the same area on the NICMOS K-band (F205W) image is shown in the right panel. Blue circles are NACO sourcesand dots are NICMOS detections in both filters. Green, larger circles are Hokupa’a sources. Several close neighbours of bright sources andequal brightness pairs are exclusively detected in the high resolution NACO data.

mer cluster membership of ejectedstars. High-resolution NIR spectro-graphs such as CRIRES expected togo into operation in the next years havethe capability to overcome the huge lineof sight obscuration towards theGalactic Centre. The determination ofspectral types with SINFONI/SPIFFI al-lows one to disentangle the clustermembers from the dense backgroundpopulation, thereby yielding accurateestimates on the field star contamina-tion when deriving mass functions. Withthese prospects in mind, spectroscopicobservations may shed light on the un-solved questions of massive star for-mation in a dense environment.

8. Conclusions

NAOS-CONICA data of the densestellar field in the Arches cluster centreyield significantly better spatial resolu-tion than NICMOS on HST. About 50%more stars are resolved in the clustercentre down to the same magnitudelimit. The luminosity functions revealthat each magnitude interval up to thebrightest stars was affected by crowd-ing losses before. The NACO observa-tions of the densest cluster region with-

in R < 5″ are 75% complete down to H= 19.4 mag and K = 17.7 mag or ∼ 3M, while HST/NICMOS and Gemini/Hokupa’a observations were limited to∼10–20 M in the immediate clustercentre. While a lower Strehl ratio is re-sponsible for the lower performance inthe Hokupa’a data, the NICMOS reso-lution is limited by the smaller HST mir-ror and correspondingly larger diffrac-tion limit. As we have shown, evenmoderate AO performance using faintvisual reference sources (V = 16 mag)now achieves higher spatial resolutionfrom the ground while at the same timeexploiting the light collecting power andsensitivity of 8-m-class telescopes.Thus, a much more complete census ofstellar populations in dense fields in ourown Galaxy as well as in nearby re-solved galactic systems is possible withground-based AO instruments such asNAOS-CONICA.

References

Blum, R. D., Schaerer, D., Pasquali, A., et al.2001, AJ, 122, 1875.

Brandner, W., Grebel, E. K., Barbá, R. H., etal. 2001, AJ, 122, 858.

Brandner, W., Rousset, G., Lenzen, R.,Hubin, N. et al. 2002, The Messenger No.107, 1.

Cotera, A.S., Erickson, E.F., Colgan, S.W.J.,Simpson. J.P., Allen, D.A., & Burton,M.G., 1996, ApJ, 461, 750.

Devillard, N. 1997, The Messenger No 87,19.

Eisenhauer, F., Quirrenbach, A., Zinnecker,H., Genzel, R. 1998, ApJ, 498, 278.

Figer, D.F., Najarro, F., Gilmore, D., et al.2002, ApJ, 581, 258.

Figer, D.F., Kim, S.S., Morris, M., et al. 1999,ApJ, 525, 750.

Lejeune, T., & Schaerer, D. 2001, A&A, 366,538.

Lenzen, R., Hoffmann, R. et al. 1998, SPIE3354, 606.

Massey, P., Johnson, K.E., Degioia-Eastwood, K. 1995, ApJ, 454, 151.

Rieke, G.H., & Lebofsky, M.J. 1985, ApJ,288, 618.

Rousset, G., Lacombe, F. et al. 2002, SPIE4007, 72.

Salpeter, E.E. 1955, ApJ, 121, 161. Scalo, J.M. 1998, in ASP Conf. Ser., Vol.

142, The stellar initial mass function, 201. Schoedel, R., Ott, T., Genzel, R. et al. 2002,

Nature 419, 694. Stetson, P.B. 1987, PASP, 99, 191. Stolte, A., Grebel, E.K., Brandner, W., Figer,

D.F., 2002, A&A, 394, 462.Tessier, E., Bouvier, J., Beuzit, J.-L.,

Brandner, W. 1994, The Messenger No.78, 35.

13

Figure 5: Mass Function of the Arches cluster. Left panel: The area within the Gemini FOV has been selected, such that the MF correspondsto the CMD displayed in Figure 3. A linear change in extinction over the field has been corrected, and a colour cut of 1.35 < H-K < 1.77 maghas been applied to the corrected CMD to select Arches main-sequence stars. The Arches main sequence can clearly be identified down to18.3 mag in K-band, or 2.5 M (logM/M = 0.4), beyond which the bulge population dominates the CMD. Right panel: The mass function ofArches cluster centre stars observed with NACO. Stars within R < 5″ are included. A weighted least-squares fit has been performed on datapoints above 10 M. The MF in the cluster centre appears flatter than the integrated MF, indicating mass segregation.

Early Galactic Chemical Evolution with UVES M. PETTINI, Institute of Astronomy, Cambridge, UK

The UVES instrument on the ESOVLT2 telescope has now been provid-ing high-resolution spectra of faint starsand galaxies for nearly three years.European astronomers have taken ad-vantage of this new facility to conductextensive and accurate studies of thefirst stellar generations of our Galaxy,

and to probe the chemical compositionof gas in the high-redshift universe. InNovember of last year, a workshop onthe theme “Early Galactic ChemicalEvolution with UVES” was held at theESO headquarters in Garching.Attended by more than 50 stellar andextragalactic astronomers, the work-

shop served as a focus to bring togeth-er scientists interested in the generaltheme of how the chemical enrichmentof galaxies progressed over the cosmicages, from the Big Bang to the presenttime. The meeting proved to be a valu-able opportunity to exchange ideas, as-sess progress to date, and to chart fu-

ture directions of research in this field ofstudy. In this article, I highlight some ofthe extragalactic aspects of the work-shop, concentrating in particular on re-cent developments in the study of ele-ment abundances at high redshifts.

To set the scene for the non-special-ists, let me just clarify a couple ofpoints. First, to astronomers the words‘high redshift’ are synonymous with ‘along time ago’. With UVES on the VLTwe can observe galaxies – and meas-ure their chemical composition – at atime when the universe was only10–20% of its present age. Second, thetechnique which is most widely used fordetailed measurements of elementabundances in these distant (both inspace and time) reaches of the uni-verse is QSO absorption-line spectros-copy, illustrated in Figure 1. The galax-ies are generally not seen directly, butthrough the absorption they cast in thespectra of background bright sources,such as quasars (or QSOs for short).

The Primordial Abundance ofDeuterium

Among the light elements created inthe Big Bang, Deuterium is the onewhose abundance is most sensitive tothe density of baryons in the universe.Astronomers have been aware sincethe 1970s of the potential of QSO ab-

sorption line spectroscopy for the de-termination of the ratio of deuterium tohydrogen (D/H). If D could be found indistant galaxies which have experi-enced little star formation, the D/H ratiomeasured in their interstellar mediashould be very close to the primordialvalue. However, this technique has hadto wait for more than twenty years to beimplemented, since it requires record-ing the light of relatively faint sources athigh spectral resolution and signal-to-noise ratio – a combination which onlyechelle spectrographs on 8–10-m-classtelescopes can deliver. Even with cur-rent technology these are still difficultobservations, and this is why the totalsum of available data today consists ofonly five measurements (plus one up-per limit), obtained with UVES on VLT2,HIRES on the Keck I telescope, and theHubble Space Telescope.

These data are reproduced in Figure2. The mean value from the five detec-tions is ⟨D/H⟩ = (2.6 ± 0.2) × 10–5 whichin turn corresponds to a density ofbaryons ΩB = 0.05 (for a value of theHubble constant of 65 km s–1 Mpc–1).That is, the baryons of our familiarphysical world only make up 5% ofthe critical density of the universe. Fur-thermore, since the currently favouredvalue of the density of matter, in all itsforms, is ΩM = 0.3, the primordial abun-dances of deuterium and other light ele-ments are one of the strongest piecesof evidence we have for the existenceof non-baryonic dark matter, which out-numbers the baryons by 5 to 1.

The value ΩB = 0.05 now seems se-cure since an entirely different methodfor its determination, which uses thepower spectrum of the cosmic micro-wave background, has recently been

14

Figure 1: The technique of QSO absorption line spectroscopy is illustrated in this montage (courtesy of John Webb). QSOs are among thebrightest and most distant objects known. On the long journey from its source to our telescopes on Earth, the light from a background QSOintercepts galaxies and intergalactic matter which happen to lie along the line of sight (and are therefore at lower absorption redshifts thanthe QSO emission redshift). Gas in these structures leaves a clear imprint in the spectrum of the QSO in the form of narrow absorption lineswhich carry a wealth of information on the physical properties and chemical composition of the gas producing the absorption. The strong ab-sorption feature near 4600 Å is a damped Lyman alpha line.

Figure 2: Available meas-urements of the abun-dance of deuterium athigh redshift (plottedagainst the neutral hydro-gen column density) ob-tained via QSO absorp-tion line spectroscopy ofeither Lyman limit (red tri-angles) or damped Ly-man alpha (blue circles)systems. (Adapted fromPettini & Bowen 2001).

Figure 4: Relativeabundances of Siand Fe in dampedLyman alpha sys-tems. Abundancesare expressed on alogarithmic scale rel-ative to the solarcomposition, so that[Fe/H] = –1 denotesan abundance of ironrelative to hydrogenequal to one tenth ofthe solar value. Themeanings of the dif-ferent symbols aregiven in the legend.Short- and long-dashed lines showthe average trendsof [Si/Fe] vs [Fe/H]before and after cor-rection for the frac-tions of Si and Felocked up in dustgrains. The continu-ous line shows the mean trend for Galactic stars. (Adapted from Vladilo 2002.)

shown to give results in excellentagreement with those from primordialnucleosynthesis. Thus the problem canbe considered ‘solved’... or can it?There are some remaining naggingdoubts as to the origin of the dispersionof the different measurements which,as can be seen from Figure 2, deviatefrom the mean by several times theirestimated (random) errors. Is this scat-ter real, or just evidence for unrecog-nised systematic errors? Experiencewould tell us that the second possibilityis the more likely, except that when wemeasure the D/H ratio in the solarneighbourhood we also find an unex-plained dispersion in the data. Figure 3is the most recent compilation of thebest relevant data, all obtained withspace missions. While all the measure-ments in the interstellar medium within100 parsecs of the Sun are in agree-ment at D/H = 1.6 × 10–5, interstellarsightlines to more distant stars span arange of almost a factor of three. Thereis a general agreement that this disper-sion is both real and unexplained. Thepuzzle arises because the ISM is wellmixed over these distances and showsa high degree of uniformity in otherchemical elements such as nitrogenand oxygen, so why should D be differ-ent? We know of no way of either pro-ducing deuterium nor of destroying itselectively. The scatter we see beyond100 pc of the Sun is comparable to thetotal amount of astration of deuterium(its destruction through star formation)over the lifetime of the Milky Way. Ittherefore seems really important to es-tablish whether such a scatter also ex-ists in other galaxies, particularly thoseat high redshifts which we see as QSOabsorption line systems and which giveus our estimate of the primordial D/Hratio. There is thus a strong incentive toimprove the statistics in Figure 2 withmore measurements with UVES, oncesuitable background QSOs have beenidentified. This should be easier nowthan it has been in the past, thanks tothe large numbers of new quasars dis-covered by on-going large-scale sur-veys of the sky.

Element Abundances inDamped Lyman Alpha Systems

This class of QSO absorbers, knownas DLAs for short, are characterised bylarge column densities of neutral hydro-gen, in excess of 2 × 1020 H atoms persquare cm. The general view is thatthey probably represent some earlystage in the evolution of the galaxieswe see around us today, perhaps at atime shortly after they had condensedout of the intergalactic medium, but be-fore they had had time to form manystars, so that most of their mass stillresided in the interstellar medium.While astronomers still debate how

DLAs are related to other populationsof galaxies seen at high redshift, onething is clear: they are our best labora-tory – by far – for measuring the abun-dances of a wide variety of elements inthe young universe. The bright back-ground light of the quasars in whichthey are seen allows studies of un-precedented detail with instrumentssuch as UVES.

If we look back to the scientific caseswritten in the 1980s to seek funds forthe construction of large telescopes, wesee that the study of element abun-dances in QSO absorbers such asDLAs figured very prominently. And in-deed, right from its commissioning days

on the telescope, European astronom-ers have put UVES through its paceswith observations of DLAs, capitalizingin particular on the instrument’s world-beating efficiency throughout the opti-cal spectrum, from ultraviolet to far-redwavelengths. Some European high-lights in this field include: the discoveryand study of some of the highest red-shift DLAs, up to zabs = 4.466 which intoday’s favoured cosmology corre-sponds to only 1.5 billion years after theBig Bang; much improved statistics onthe metallicities of DLAs at redshiftsgreater than 3, based on the abun-dance of Zn; and the first comprehen-sive search for molecular hydrogen at

15

Figure 3: Compilation of available measurements of the abundance of deuterium in the MilkyWay interstellar medium, all obtained with space missions as indicated in the legend, as afunction of distance from the sun. (Reproduced from Hoopes et al 2003.)

Figure 6: Relativeabundances of Mnand Fe in DLAs,plotted vs the abun-dance of Zn (takenas a proxy for Fe).In the Galactic in-terstellar medium,Mn and Fe are gen-erally depleted bysimilar amounts sothat the ratio ofthese two elementsin DLAs is expect-ed to be relativelyinsensitive to thedegree of dust depletion. Small dots are data for stars in the Milky Way; open and filled cir-cles are measurements in DLAs. (Reproduced from Dessauges-Zavadsky et al. 2002, wherereferences to the original measurements are given.)

high redshift. This, the most common ofastrophysical molecules, provides awealth of data on the physical condi-tions in DLAs, particularly on the tem-perature and density of both particlesand photons in the interstellar media ofthese early galaxies.

Ultimately, all of this work on DLAscomplements local studies, which havefocussed on old stars in the Milky WayGalaxy and on regions of ionized gasnear hot stars in nearby galaxies, to ad-dress some of the same basic ques-tions, such as the star formation histo-ries of galaxies (How did star formationproceed in the early Galaxy and inDLAs?), and the nucleosynthetic originof different elements (Which stars wereresponsible for ‘cooking’ different ele-ments in their interiors and releasingthem into the interstellar medium, mak-ing them available for subsequentgenerations of stars?). I shall illustratethese points with a few examples cho-sen from work carried out by Europeanastronomers with UVES.

Are Oxygen and Other Alpha-Capture ElementsOverabundant Relative to Iron inDLAs?

Hidden behind this mouthful of jargonlies one of the cornerstones of chemicalevolution studies. Oxygen and otherelements which are synthesised by theaddition to oxygen of one or more alphaparticles (examples are Mg, Si, S, andCa) are thought to be produced mostlyin the interiors of massive stars whichexplode as Type II supernovae some10 million years after an episode of star

Do we see the same pattern inDLAs? The answer to this question iscomplicated by the fact that most of theelements concerned can be partly hid-den from view in DLAs as they con-dense to form interstellar dust (thisproblem does not affect stellar abun-dance measurements, since the dustall melts away when a star is formed).After some initial debate as to whetherdust is present in DLAs, most as-tronomers working in this field nowagree that dust depletions have tobe taken into account when analysingthe pattern of element abundancesin DLAs. When this is done, as inFigure 4, a mixed picture emerges.While some DLAs exhibit an overabun-dance of Si relative to Fe comparable tothat seen in Milky Way stars, others –perhaps most – don’t. The conclusionseems to be that DLAs trace galaxieswith different pasts; some may haveformed stars on timescales similar tothat of the early Milky Way, while othersapparently did so more slowly, or inter-mittently, so that Fe could catch up withSi. It makes sense (at least to theauthor of this article) that DLAs shouldsample a wide range of galaxy types,and consequently a variety of star for-mation activity. Even locally, we haverecently begun to learn that almost notwo galaxies have similar histories inthe assembly of their stellar popula-tions, and this is reflected in their vary-ing [alpha/Fe] ratios, as evidenced bythe work presented by Kim Venn at themeeting.

Of relevance to this question are alsosome new stellar abundance measure-ments reported at the workshop byPoul Nissen and collaborators, and re-produced in Figure 5. Recently ob-tained UVES data show that the [S/Zn]ratio does show an enhancement (rela-tive to solar) in Galactic stars of lowmetallicities. This result had been ex-pected, since S is an alpha-capture ele-ment and the abundance of Zn gener-ally follows that of Fe to within about 0.1dex, but it is nevertheless reassuring tosee it confirmed empirically. The signif-icance of this particular pair of elements

16

formation. On an astronomical timescale this is quick work. On the otherhand, the major producers of Fe arethought to be Type Ia supernovaewhose progenitors are stars of relative-ly low mass which can take up to 1 bil-lion years (100 times longer than theoxygen producers) to explode and re-lease the Fe into the interstellar medi-um. In Milky Way stars with an ironabundance of less than 1/10 solar (or[Fe/H] < –1 in the usual notation), oxy-gen and other alpha-elements are in-deed relatively more abundant thaniron by factors of between two andthree. The usual interpretation of thiswell established fact is that, during itsearly stages of evolution, our Galaxyreached a metallicity of 1/10 of solarrelatively quickly, presumably in lessthan one billion years, so that Fe hadnot yet had time to catch up with O inthe interstellar medium.

Figure 5: [S/Zn] vs. [Zn/H] for a sample of nearby main-sequence stars. Filled circles referto stars with halo kinematics and open circles to disk stars. Data for the halo stars were de-rived from the S I λλ8694.6, 9212.9, 9237.5 and the Zn I λλ4722.2, 4810.5 lines recorded withUVES (Nissen et al. in preparation); those for the disk stars are reproduced from Chen et al.(2002). The two stars, HD103723 and HD105004, belong to a group of low [alpha/Fe] halostars discovered by Nissen & Schuster 1997. (Figure courtesy of Poul Nissen.)

is that neither shows much affinity fordust, thereby circumventing the deple-tion corrections necessary for Si andFe. The few available data on S and Znin DLAs (shown with open squares inFigure 4) conform to the broad pictureoutlined above; without doubt manymore measurements of this diagnosticpair will be attempted in the near future,now that their relative abundances inGalactic stars have been clarified.

The Case of Manganese

Mn shows an opposite trend withmetallicity from that of the alpha-cap-ture elements: its abundance relative toFe decreases, rather than increases,with decreasing metallicity. Could thisalso be a reflection of the pace of paststar formation? Its abundance indamped Lyman alpha systems sug-gests that this is not the case. As canbe seen from Figure 6, the DLA meas-urements match the pattern seen inMilky Way stars even though, as wehave seen, the QSO absorbers are like-ly to sample a variety of galaxies whichformed stars at different rates. Theagreement between Galactic and extra-galactic data suggests that perhapsthere is a metallicity dependence in theyield of Mn; in any case, the DLA meas-urements for manganese are providingus with clues to the chemical yields,rather than to the past history of starformation.

Low- and Intermediate-MassStars as the Source of PrimaryNitrogen

The details of the nucleosynthesis ofnitrogen have been the subject of con-siderable debate in recent years andthis element provides perhaps the bestexample of how observations at highredshifts are driving models of its pro-duction in stars. Available measure-ments of the abundances of N and Oare collected in Figure 7. Without goinginto too much detail, two points arenoteworthy. First, DLAs occupy a re-gion of the diagram which is sparselysampled by local H II regions. It is mucheasier to find metal-poor gas at highredshift than in the present-day uni-verse. Thus, the relative abundances ofN and O in DLAs offer a clearer windowon the early nucleosynthesis of thesetwo elements than was available beforethe commissioning of UVES. Second,note the relatively large spread of val-ues of the (N/O) ratio, from about –1.2to –2.4 (on a logarithmic scale). Quali-tatively, this is what was expected if themain source of N at low metallicities areintermediate-mass stars, between sev-en and four solar masses, which re-lease their newly synthesised N into theinterstellar medium well after the injec-tion of O by Type II supernovae. TheDLA data have provided the first empir-

ical evidence in support of this idea of adelayed production of primary N by in-termediate mass stars, an idea whichhad previously been based on entirelytheoretical yield calculations.

Quantitatively, however, the propor-tion of DLAs with values of (N/O) belowthe level labelled ‘Primary’ in Figure 7 issurprisingly high, if the time delay is ofthe order of only 300 million years, ascurrent models predict. Perhaps futureobservations with UVES will show thisto be an artefact of the still limited sta-tistics (and indeed there is by no meansuniversal agreement between differentgroups as to the pattern of DLA meas-urements in Figure 7). However, theseand similar data have already providedthe impetus to theorists to examinemore closely the details of the nucleo-synthesis of N at low metallicities andmechanisms have been uncoveredwhich have the potential of extendingthe time delay. One of the most realis-tic of these seems the inclusion of stel-lar rotation (hitherto largely ignored)into the calculations of the stellar yields.As shown by Georges Meynet andAndré Maeder in Geneva, rotation mayallow metal-poor stars of masses lowerthan four solar masses to synthesize N,with the consequence that its releaseinto the interstellar medium may be amore protracted affair, coming to com-pletion on timescales perhaps as longas those of Type Ia supernovae. Wecan expect significant advances onthese questions in the years ahead,both observationally and theoretically.

Concluding Remarks

I hope that this brief review has man-aged to convey some of the flavour ofthat exciting two-day meeting inGarching last November. The work-shop served to highlight the major im-pact which UVES has had, and contin-ues to have, on European astronomy,

not only in the depth and breadth ofnew results and projects which the in-strument has made possible, but also inthe increasing dialogue between stellarand extragalactic astronomers, and be-tween observers and theorists, whichUVES has helped to foster.

The meeting concluded with a briefoutline by Sandro D’Odorico of currentplans for future upgrades of UVES and,looking further ahead, for complemen-tary instruments which will allow high-resolution spectroscopy at wavelengthsfrom the ultraviolet to the near-infrared.This was followed by a brief user feed-back session. The session was briefbecause generally there is a high levelof satisfaction among UVES users! Themerits of the data archive were high-lighted, although the point was alsomade that the usefulness of the archivewill be significantly increased whenUVES pipeline reduced data are includ-ed in it. Hopefully this important devel-opment will be implemented in the near,rather than medium-term, future.

Finally, all participants expressed theirsincere thanks to Francesca Primas,whose enthusiasm, generosity, and im-peccable organization – always accom-panied by her trademark smile – madethis meeting such a resounding suc-cess.

References

Chen, Y.Q., Nissen, P.E., Zhao, G., &Asplund, M. 2002, A&A, 390, 225.

Dessauges-Zavadsky, M., Prochaska, J.X.,& D’Odorico, S., 2002, A&A, 391, 801.

Hoopes, C.G., Sembach, K.R., Hebrard, G.,Moos, H.W., & Knauth, D.C. 2003, ApJ, inpress (astro-ph/0212303).

Nissen, P.E., & Schuster, W.J. 1997, A&A,326, 751.

Pettini, M., & Bowen, D.V. 2001, ApJ, 560,41.

Pettini, M., Ellison, S.L., Bergeron, J., &Petitjean, P. 2002, A&A, 391, 21.

Vladilo, G. 2002, A&A, 391, 407.

17

Figure 7: Logarithmic abundances of N and O in nearby H II regions (small dots) and in DLAs(open and filled triangles). Typical errors are shown in the bottom right-hand corner. Adaptedfrom Pettini et al. (2002), where full references to the data and a more extensive discussionof this diagram can be found.

Introduction

The VIMOS instrument has been suc-cessfully commissioned at the VLT Me-lipal in 2002, and is now open to thecommunity, with a first set of observa-tions for General Observers scheduledin Period 71. From the start, VIMOS hasbeen designed with the goal to under-take large, deep surveys of the distantuniverse. VIMOS is capable of an im-pressive multiplexing capability: up to1000 objects can be measured simulta-neously [1], which allows to assemblelarge datasets of faint objects in a shorttime.

The VIRMOS consortium has de-fined the VIRMOS-VLT Deep Survey(VVDS), as a major programme tostudy the evolution of galaxies, clus-ters, large-scale structures, and AGNs,over more than 90% of the current ageof the universe [2]. The survey aims toobserve more than 100,000 objects inthe distant universe 0 < z < 5, which willallow a solid knowledge of the luminouscontent of the universe at early epochs,and its evolution, based on large statis-tical samples.

We describe here the start of the sur-vey with the first guaranteed night ob-servations which have been carried outin the October-December 2002 timeframe. The general survey strategy,

performances, and first results are pre-sented.

VVDS Outline

The goal of the VIRMOS deep red-shift survey is to study the evolution ofgalaxies, AGNs, and large-scale struc-tures, over a redshift range 0 < z < 5+.The requirement is to be able toanalyse the basic properties such asthe luminosity function or spatial corre-lation function of the galaxy population,as a function of galaxy type or localdensity, in each of several time stepscovering the redshift range of interest.

We are therefore planning to observethe following magnitude-limited samples:

1. “Wide” survey: more than 105 gal-axies with redshifts measured to IAB =22.5 (redshift up to z ~1.3), it will be con-ducted in 5 fields identified in Table 1.

2. “Deep” survey: more than 4 × 104

galaxies with redshifts to IAB = 24 (red-shift up to z ~5) conducted in the 0226-04 field.

3. “Ultra deep survey’’: more than 103

galaxies with redshifts to IAB = 25 con-ducted in the 0226-04 field.

The spectroscopic samples will beobserved with R ~ 210 in a first pass. Asubsample of 10,000 galaxies will thenbe selected for spectroscopy at a reso-lution R ~ 2500–5000 to allow the study

of the evolution of the fundamentalplane, and perform a detailed analysisof the spectro-photometric properties ofgalaxies.

Four fields in addition to the ChandraDeep Field South are targeted duringthe course of the survey. Each field is 2× 2 deg2 for the “wide” survey, and 1.3deg2 for the “Deep” survey. Extensiveimaging has been carried out using theCFHT12K camera, the ESO 2.2-mWIFI, and the NTT-SOFI to assemble asample of around 2 million galaxieswith BVRI photometry, with smallersubsets having in addition U and/or K’photometry [3].

The complete photometric surveygives access to a volume of the uni-verse which is equivalent to the 2DFredshift survey, although at a muchhigher mean redshift of z ~ 1.

First Observations

The ESO Director General has grant-ed the VIRMOS consortium the earlyuse of guaranteed time in the periodOctober-December 2002, immediatelyfollowing the last commissioning periodin September, interleaved with periodsof “Paranalization”. This approach hada double advantage: to ESO becausethe consortium committed to technical-ly attend to the instrument with its own

18

The VIRMOS-VLT Deep Survey: a Progress ReportO. LE FÈVRE1, G. VETTOLANI2, D. MACCAGNI3, J.-P. PICAT4, B. GARILLI3, L. TRESSE1, C. ADAMI1, M. ARNABOLDI10, S. ARNOUTS1, S. BARDELLI5, M. BOLZONELLA3, D. BOTTINI11, G. BUZZARELLO10, S. CHARLOT6, G. CHINCARINI7, T. CONTINI4, S. FOUCAUD1, P. FRANZETTI11, L. GUZZO7, S. GWYN1, O. ILBERT1, A. IOVINO7, V. LE BRUN1, M. LONGHETTI5, C. MARINONI1, G. MATHEZ, A. MAZURE1, H. MCCRACKEN5, Y. MELLIER8, B. MENEUX1, P. MERLUZZI10, S. PALTANI1, R. PELLÒ4, A. POLLO7, M. RADOVICH10, P. RIPPEPI10, D. RIZZO7, R. SCARAMELLA9, M. SCODEGGIO3, G. ZAMORANI5, A. ZANICHELLI5, E. ZUCCA5

1Laboratoire d’Astrophysique de Marseille, France; 2Istituto di Radio Astronomia, Bologna, Italy; 3Istituto di Fisica Cosmica e Tecnologie Relative, Milan, Italy; 4Observatoire Midi-Pyrénées, Toulouse, France;5Osservatorio Astronomico di Bologna, Italy; 6Max Planck Institute für Astronomie, Garching, Germany;7Osservatorio Astronomico di Brera, Italy; 8Institut d’Astrophysique de Paris, France; 9Osservatorio Astronomico di Roma, Italy; 10Osservatorio Astronomico di Capodimonte, Naples, Italy11Istituto di Astrofisica Spaziale e Fisica Cosmica CNR, Milano, Italy

Field Alpha (2000) Delta (2000) Survey area Survey mode Observed in Oct.-Dec.2002(deg2)

0226-04 02h26m00.0s -04°30’00” 1.3 (Deep) Wide, Deep, Ultra-deep 20 deep pointings, 9188 slits4 (Wide)

CDFS 03h32m28.0s -27°48’30” 0.1 (Deep) Deep, IFU, 5 deep pointings, 1 deep early community release IFU pointing, 2109 slits

1003+01 10h03m39.0s +01°54’39” 4 (Wide) Wide 7 pointings, 2595 slits

1400+05 14h00m00.0s +05°00’00” 4 (Wide) Wide

2217+00 22h17m50.4s +00°24’27” 4 (Wide) Wide 17 pointings, 6849 slits

Table 1: VVDS fields

support staff to solve pending problemswhich may have occurred, thereforemaking the instrument more reliable;and to the consortium because of theability of an early start of the VVDS.

Two observing periods have been al-located to the consortium: 29 October –11 November (14 nights) and 27November – 11 December (15 nights).The first observing run was blessedwith excellent weather, only 2 nightswere lost due to high-level clouds, whileduring the second period strong northwinds prevented us from observing dur-

ing 8 nights. In total, the equivalent of19 dark short summer nights benefitedfrom clear weather and nominalParanal seeing. While some of the con-sortium observers had their first timeever on the instrument during this time,the training was smooth and quick. Theteam made its utmost to ensure that theinstrument was ready for each night ofobservation. As anticipated, we had toface some technical problems, relatedin particular to the occasional failure ofa mask or grism insertion, these wereaddressed during the day with the con-

sortium technical staff. Less than onenight was lost to instrument problems inthese first observing runs.

The preparation of masks for thenights of observations has been chal-lenging but smooth, especially duringconsecutive long periods of nice weath-er, when up to 6 sets of 4 masks withmore than 500 slits in each set had tobe manufactured during the day withthe Mask Manufacturing Machine [4].The observing time has been balancedbetween long integration on the deepsurvey, and short integration on the

19

Figure 1: The distribution of the 96 VIMOS pointings for the VVDS-Wide fields (left) and 66 VVDS-Deep pointings (central area, right) and 40VVDS-Wide pointings in the 0226-04 area. Each pointing of 4 quadrants is labelled and the area covered is shaded in red, with a density in-creasing at each passage. Targets are selected only in the shaded areas. The spatial selection function for the VVDS-Wide is easy to modeland correct for. Carefully selected offset patterns allow to cover an area of 1.3 deg2 in the 0226-04 Deep area with a smooth spatial selec-tion function totalling 4 passes of VIMOS.

Figure 2: Pointings observed in the 1003+01 VVDS-Wide in the fall2002.

Figure 3: Pointings observed in the 0226-04 VVDS-Deep in the fall2002.

wide survey. The strategy adopted hadbeen tested during commissioning andis recommended for all MOS observa-tions: because of the strong CCD fring-ing above 8200 Å, sequences of obser-vations were acquired moving the ob-jects along the slits. This allows to com-pute the fringing pattern in each slit,

and remove it for an improved sky andfringing correction. For the wide survey,we used an offset pattern 0, –0.7, –1.4,+0.7, +1.4 arcsec from the referenceposition, with 5 exposures of 9 minuteseach, while for the deep survey we usedan offset pattern 0, –0.8, –1.6, +0.8, +1.6arcsec, repeating twice this pattern for

a total of 10 × 27 minutes exposure. The VVDS has a set of carefully de-

signed pointings (one pointing is oneposition of VIMOS on the sky with allfour quadrants), to minimize the spatialbias applied by the non-contiguousgeometry of the instrument. In the widesurvey, we carry out two passes of

20

Figure 4: Completeness measurement vs. magnitude in the quadrant1 of the F02P030-Deep pointing, defined as the ratio of objects withsecure redshift identification to the total number of objects observedwithout instrumental bias (top), and the corresponding magnitude dis-tribution with the objects for which no redshift could be measured, in-dicated as the shaded histogram (bottom).

Figure 5: Velocity accuracy: comparison of redshifts measured inde-pendently on the same galaxies from two different mask datasets.The error in velocity measurement accuracy is 270 km/s rms for theLRRed grism and 1 arcsec slits.

Figure 6: Sample spectra from the first VVDS observations.

21

mean redshift is 0.55 for the VVDS-Wide and 0.8 for the VVDS-Deep. Theredshift distribution for each pointingnow shows the classical succession ofhigh density peaks and low densityareas (Fig. 7).

Conclusion

The VVDS survey is now proceedingwith a first set of successful observa-tions. Upon final processing of the morethan 20,000 spectra acquired so far,and taking into account the fraction ofstars and the completeness, the yield ofsecure redshift measurements shouldamount to ~15,000 for this first set ofobservations. This will allow for a firstcomputation of the luminosity functionand to map the distribution of galaxiesdown to IAB = 24, from a sample free ofselection effects.

The raw IFU data obtained on theChandra Deep Field South is availableon-line (http://www.oamp.fr/virmos),and we will make a public release of theChandra Deep Field South MOS data(1D calibrated spectra and redshiftmeasurements) in this same field assoon as the processing is completed.

References

[1] Le Fèvre, O., et al., The Messenger,September 2002.

[2] Le Fèvre, O., Vettolani, G., Maccagni, D.,Mancini, D., Mazure, A., Mellier, Y., Picat,J.P., et al., astro-ph/0101034, ESO Astro-physics Symposia Series “Deep Fields”,Cristiani, Renzini, Williams, Eds., Springer,p. 236.

[3] Le Fèvre, O., Mellier, Y., McCracken,H.J., Arnaboldi, M., et al., 2000, A.S.P.Conf. Series, Clowes, Adamson, andBromage Eds., vol. 232, p. 449.

[4] Conti, G., Mattaini, E., Maccagni, D.,Sant’Ambrogio, E., Bottini, D., Garilli, B.,Le Fèvre, O., Saïsse, M., Voët, C., et al.,2001, PASP, 113, 452.

VIMOS on the same sky area, offset by(2.2) arcminutes. This ensures thatmore than 50% of the IAB = 22.5 objectsare observed on average, with a smoothspatial sampling function which is easyto model (Fig. 1), and to correct for, e.g.for the computation of the correlationfunction.

For the VVDS Deep, the set of point-ings allows for a continuous spatialsampling of the area, with 4 VIMOSpasses at any sky location (Fig. 1).

A total of 24 wide pointings havebeen observed in the 2217+00 and1003+01 fields, for a total of 9444 slits(Fig. 2), and 25 deep pointings have beenobserved in the 0226-04 (Fig. 1) andCDFS fields for a total of 11,297 slits.

First Results

The data processing is now in fullswing, with 4 teams processing thedata in parallel. Each pointing isprocessed twice: the automated dataprocessing pipeline VIPGI is applied,and the product is distributed to twoteams/reducer to extract the redshiftsindependently. A dedicated tool hasbeen developed to automatically meas-ure redshifts. At this time we examinespectra one by one for redshift meas-urements proposed by the KBRED tool.Although this process is somewhat longand tedious, it allows one person toprocess a set of 4 masks with ~500 slitsin a few working days. Careful dataquality assessment is made for eachmask, checking the wavelength andflux calibrations.

The quality of spectra is as expected.We are experiencing a completenessbetter than 90% of the overall sample

for both the wideand deep surveys(Fig. 4). The ve-locity accuracy hasbeen estimatedfrom independentredshift measure-ments of the samegalaxies in twodifferent pointings,it is ~270 km/s rmsas derived froma sample of 27galaxies (Fig. 5).Examples of spec-tra are shown in Figure 6.

We have performed test observa-tions for the R ~ 2500 follow-up obser-vations which will be carried out on ob-jects selected from the low-resolutionsurvey. Spectra of z ~ 1 galaxies withemission lines like [OII] (Fig. 8) showthat the measurement of the velocitydispersion in galaxies will be possibleto a level of ~ 30 km/s.

In the VVDS-Deep pointings, we findthat 10% of the objects are stars, 0.7%QSOs, since no pre-selection has beenmade besides the I-band limited sam-ple. The fraction of stars is higher atabout 15–20% in the VVDS-Wide. The

Figure 7: Redshiftdistribution of galax-ies around z ~ 1 inthe 0226-04 VVDS-Deep area. This isassembled from 3pointings, and clear-ly shows densepeaks and emptyregions.

Figure 8: Velocity curve of a galaxy at z = 0.458 from observations in the HRRED grism (R~ 2500) of the Hβ line. Left: raw spectrum obtained with a slit cut by the laser machine tiltedon the major axis of the galaxy, and corrected spectrum after sky subtraction and wavelengthstraightened. Right: velocity curve obtained after correction of the inclination angle of thegalaxy. Top right: image obtained with HST/ACS (GOODS-CDFS public data).

The Edgeworth-Kuiper Belt andthe Formation of the SolarSystem

The first object in the Edgeworth-Kuiper Belt was observed in 1930,when Clyde Tombaugh discoveredPluto at a distance of 43 AU (1AU = oneastronomical unit, the mean distancebetween Earth and the Sun = 149.6 mil-lion km), i.e. beyond the orbit ofNeptune. About 20 years later Edge-worth and Kuiper started to speculateabout the existence of another asteroidbelt at the edge of the planetary sys-tem. Another 30 years later this specu-lation became an hypothesis whenFernandez and Ip argued for the exis-tence of an Ecliptic-oriented reservoirof icy bodies beyond Neptune as asource for short-period comets, the re-cruitment of which was difficult to ex-plain by gravitational capturing of OortCloud comets through planets whenapproaching the inner solar system.The hypothesis became reality in 1992,when, during a search campaign fordistant asteroids, Jewitt and Luu found1992 QB1 (now numbered: 15760) at adistance of 41 AU from the Sun.

In the meanwhile, after the detectionof more than 700 objects beyond theorbit of Neptune over the past decade,a new class of solar system bodies hasbeen discovered. These Transnep-tunian Objects (TNOs) are the largestmembers of the Edgeworth-Kuiper Belt.They have caused a complete revisionof our knowledge and understanding ofthe outer solar system: the Edgeworth-

Kuiper Belt is believed to represent themost original remnant from the forma-tion period of our planetary system. Itwas in this region where within lessthan 20 million years some 4.6 billionyears ago, icy planetesimals formedand grew to larger bodies, nowadaysobserved as TNOs, Centaurs, short-pe-riod comets and some icy satellites ofthe major planets. Although of the sameorigin and most likely containing thesame primordial material, these smallbodies, as we observe them now, haveexperienced different evolution over theage of the solar system, both dynami-cally and physically. Nevertheless, theycontain the most primordial materialfrom the formation of the Sun and itsplanetary system that we can observetoday.

The Global View of theEdgeworth-Kuiper Belt and itsPopulation of Objects

The dynamical ‘zoo’ of Cubewa-nos, Plutinos, Centaurs and SDOs:According to current dynamical scenar-ios, the TNO population remained inthe region of its formation, the Edge-worth-Kuiper Belt, over the age of thesolar system. However, their orbits be-came perturbed by the presence ofNeptune. The gravitational interactionwith this giant planet has ‘pumped’ ob-jects into so-called ‘excited’ (i.e. non-circular and/or inclined), but non-reso-nant orbits creating the dynamical classof ‘Cubewanos’ (named after the proto-type object 1992 QB1), also called clas-

sical disk objects. Other TNOs werecaptured in resonant orbits withNeptune forming the dynamical class of‘Plutinos’ (named after Pluto, the mostprominent representative in the 3:2 res-onance orbit). Gravitational scatteringat Neptune has produced the class of‘Scattered Disk Objects (SDOs)’ in veryeccentric orbits with semi-major axesbetween 50 and several 1000 AUs. The‘Centaurs’, presently in orbits betweenNeptune and Jupiter and named afterhalf-human-half-horse demigods inGreek mythology, are also consideredto be scattered TNOs, cascading to-wards the Sun by repeated gravitation-al interaction with the major planets. Atthe end, i.e. typically after some 10 mil-lion years in intermediate orbits, Jupiterwill either capture them as short-periodcomets (thus, some of them even be-come accessible for spacecraft explo-ration) or will expel them into the regionof the long-period Oort Cloud or eveninterstellar comets at the edge of thesolar system. A few TNOs may getstranded as satellites around the outergas giants (the most prominent candi-date is Triton, the largest moon ofNeptune).

The original formation disk, more‘Plutos’ and TNO binaries: The dropin the number of objects with semi-ma-jor axes larger than 50 AU may indicateeither a basic characteristic of the for-mation disk around our Sun (a ‘truncat-ed’ disk?) or may just mark the chal-lenge to detect the ‘cold’ and very colli-mated formation disk of even smallerplanetesimals that escaped the gravita-

22

Exploring the Icy World of the Edgeworth-Kuiper Belt – An ESO Large ProgrammeH. BOEHNHARDT (Max-Planck-Institut für Astronomie, Heidelberg, Germany)A. BARUCCI, A. DELSANTI, C. DE BERGH, A. DORESSOUNDIRAM, J. ROMON

(Observatoire de Paris, Meudon, France)E. DOTTO (Osservatorio Astronomico di Roma, Italy)G. TOZZI (Osservatorio Astronomico di Arcetri, Firenze, Italy)M. LAZZARIN, S. FOURNISIER (Osservatorio Astronomico di Padova, Italy)N. PEIXINHO (Observatorio Astronomico de Lisboa, Portugal, and Observatoire de Paris,

Meudon, France)O. HAINAUT (European Southern Observatory, Santiago de Chile)J. DAVIES (Royal Observatory Edinburgh, Great Britain)P. ROUSSELOT (Observatoire de Besançon, France)L. BARRERA (Universidad Católica del Norte, Antofagasta, Chile)K. BIRKLE (Max-Planck-Institut für Astronomie, Heidelberg, Germany)K. MEECH (University of Hawaii, Honolulu, USA)J. ORTIZ (Instituto de Astronomia de Andalucia, Granada, Spain)T. SEKIGUCHI, J.-I. WATANABE (National Astronomical Observatory, Mitaka-Tokyo, Japan)N. THOMAS (Max-Planck-Institut für Aeronomie, Katlenburg-Lindau, Germany)R. WEST (European Southern Observatory, Garching, Germany)

tional influence of the outermost plan-ets because of their large distance fromthe Sun. The overall mass in theEdgeworth-Kuiper Belt seems to bevery small (0.15 Earth masses) andcannot explain the existence of largeTNOs like Pluto/Charon or Varuna.Instead modelling results suggest thatthe original formation disk in the TNOregion was much heavier (10–100Earth masses), that it should have pro-duced more Pluto-size bodies and thatmost of its mass got lost by gravitation-al scattering at the large planets as wellas by mutual collisions of TNOs andsubsequent down-grinding of the colli-

sion products to micron size dust thatthe solar radiation pressure has re-moved from the Edgeworth-Kuiper Beltregion. The existence of binary TNOs(about 10 objects so far) supports theidea of a serious collision environmentin the belt during and possibly beyondthe end of the formation period of plan-etesimals. Ongoing search pro-grammes using imaging surveys andhigh-resolution imaging techniques atlarge telescopes world-wide try to solvethese key questions on the Edgeworth-Kuiper Belt.

Physico-chemical properties:While the dynamical history of TNOs

and their relatives seems to be reason-ably well understood, just a very vaguepicture of their physical properties andchemical constitution exists. The faint-ness of the objects (the vast majorityare fainter than 23 mag) makes themdifficult to observe even with the largesttelescopes: the brightest representa-tives (about 20 mag) require easily onefull observing night at an 8-m VLT unittelescope for a spectroscopic study ofdecent signal-to-noise.

Despite their faintness, the observedTNOs and Centaurs are only thelargest members of the belt populationwith diameter estimates ranging be-

23

Figure 1: Colour-colour diagrams of the first 28 objects observed within the Large Programme. The two panels show the B-V versus V-R (left)and R-I versus V-R (right) colours of the objects measured. The symbols are: filled circles = Cubewanos, filled triangles = Plutinos, open cir-cles = SDOs, open triangles = Centaurs. The colour of the Sun is indicated by the open star symbol. The lines indicate the direction of in-creasing reddening slopes from –10 (lower left end) to 70%/100 nm (upper right end) in steps of 10%/100 nm. The distribution of the meas-ured objects seems to be rather uniform in the typical reddening range. The shift of the objects below the line of constant reddening in theplots involving R-I colour is due to the spectral slope change of red objects towards the near-IR end of the visible spectrum.

Figure 2: Photometric “reflectivity spectra” from BVRI magnitudes of TNOs and Centaurs. The three panels show the photometric “reflectivi-ty spectra” of the first 28 objects observed with our programme. The relative reflectivity per photometric band (normalized to unity for V band)is plotted versus central wavelength of the filters used. It is a measure of the intrinsic reddening of the objects since the slope of solar spec-trum is removed from the data. The broken lines indicate the numerical spectral gradient fits for the individual objects. These plots are veryuseful to check the consistency of the photometric data since “bad” measurements and intrinsic variability of the objects will result in notice-able deviations of the photometry from a “smooth” spectral slope in these “spectra”. If found, further investigations may be needed to clarifythe nature of the scatter.

tween 30 and 1200 km (assuming a lowalbedo of 4 per cent). Search pro-grammes suggest a cumulative massfunction with power exponent between3.5 and 4. The albedo of less than 10objects is coarsely determined: TNOsand Centaurs seem to be dark objectswith albedo mostly around and below10 per cent. Only Pluto, Charon andChiron have high (around 50 per centfor Pluto and Charon) to moderatealbedo, possibly caused by fresh icedue to intrinsic atmospheric activity(see below). The rotation period ofmore than 20 objects is measured withresults ranging from 5 and 12 h (cer-tainly biased since long periods aremore difficult to measure). With the ex-ception of Pluto and Charon, it is un-clear whether the measured variabilityis due to body shape or surface reflec-tivity or both.

Until 2001 the taxonomic characteri-zation of the population was based ona sample of ~ 50 TNOs and ~ 10Centaurs, most of them observed withbroadband colours in the visible wave-length range only. Nevertheless,colour-colour diagrams and reddeningdistributions provide a first glance ofsimilarities and diversities in the sur-face taxonomy of the objects. A widerange of colours (in other wordsspectral gradients or ‘reddenings’)ranging from slightly bluer than theSun (~5%/100 nm) to extremely red(~50%/100 nm) in the visible wave-length range is found in this sample. Inthe infrared, the photometric results in-dicate only minor reddening to neutralslopes of the surface reflectivity. For afiner analysis, the overall object samplecontained some adverse culprits likebad number statistics for all dynamicalclasses, insufficient photometric accu-racy of the data, incomplete wave-length coverage (7 objects had only 1or 2 colours measured) and largedifferences in the results for the sameobjects obtained by different ob-servers. It is thus not surprising that

different researchers arrived at differentand partially contradictory conclusions,for instance a bimodal colour distribu-tion versus a continuous one with ‘out-lyers’ in both scenarios. Through aKolmogorov-Smirnov analysis of thesample it became clear that the identi-fied colour populations among TNOsand Centaurs are at best to be con-sidered as trends, but they lack clearstatistical significance because of thelow number of objects in the availablesample.

Since formed in the outer solar neb-ula, TNOs and Centaurs should bemade of icy and stony compounds.Except for the larger members that mayhave undergone constitutional modifi-cations due to radioactive heating intheir interior, and those that come clos-er to the Sun and thus get heated up inthe surface layers, the original materialremained unchanged in the body interi-or. Nevertheless, as outlined below, thesurface layers of the bodies are allmodified over the age of the solar sys-tem. N2 and H2O ices should be mostabundant in TNOs and relatives, fol-lowed by CO, CO2, CH4 and NH3.Signatures of N2, CO, H2O, CH4 andNH3 ices are indeed found in spectra ofPluto/Charon. Water ice absorptionswere detected in the Cubewano 1996TO66 and the Centaurs 5145 Pholusand 2060 Chiron, objects that representbasically the fully colour range found inthe total sample. The original ‘stony’ in-gredients in TNOs and Centaurs shouldbe primarily of silicate nature. Their de-tection, however, is extremely difficultsince they may be covered completelywith ice mantles. There is only onecase, the Centaur 5145 Pholus, inwhich a very wide and close to margin-al absorption feature (ranging from 0.6to 1.6 µm) is tentatively attributed to sil-icates. The only other supportive argu-ment for the presence of dust in theseobjects comes from the Centaur 2060Chiron and of course their relatives, theshort-period comets, i.e. through the

temporary dust coma seen aroundthese objects.

Surface evolution scenarios: Thered colour of TNOs and Centaurs isusually attributed to the effects of sur-face aging and darkening due to high-energy radiation and ion bombardment.Blue surface colours could be producedby major collisions through deposits offresh icy material from the body interioror from the impactor. The estimatedtime scale for both types of colourresurfacing are of the order of 10 millionyears. The observed colour range canbe modelled by computer simulationsinvolving both effects. However, theseresults are unfortunately not discrimina-tive for conclusions on the physical na-ture of TNO and Centaur surfaces. Re-surfacing on much shorter time scalescould happen due to ice re-condensa-tion from a temporary atmosphere pro-duced by intrinsic gas and dust activity.This process is quite efficient on Plutoand possibly Charon, but does certain-ly not work for all objects since crustformation may prevent the develop-ment of surface activity and/or the heatsource in the bodies may not be strongenough to cause such activity.

ESO Large Programme on‘Physical Studies of TNOs andCentaurs’

Programme concept and history:After a number of uncoordinated pre-cursor programmes at ESO telescopesat the end of the last decade, a consor-tium of scientists (see the list of authorsof this paper) proposed a comprehen-sive observing and analysis project tobe performed within the framework ofan ESO Large Programme: ‘Physicalstudies of TNOs and Centaurs’. Theproject was accepted to be executed atCerro Paranal and La Silla during ESOperiod 67 to 70, i.e. from April 2001 un-til March 2003. The main goal of theproject is the development of a taxo-nomic classification scheme of theseicy bodies in the outer solar system, theidentification of evolutionary tracks andtheir relationships with dynamicalclasses, the exploration of the surfacechemistry of the objects and its correla-tion with taxonomic classes. The pro-posed observing campaigns at theESO Very Large Telescope (VLT) andthe New Technology Telescope (NTT)comprise multi-colour broadband filter

24

Figure 3: Spectral gradient histograms of thefull sample of 96 objects in our analysisdatabase. The histograms show the numberof objects per spectral gradient interval foreach dynamical class as given in the table atthe bottom of the figure. The spectral gradi-ents are calculated from the BVRI photome-try of the objects.

photometry of 60–70 objects in the vis-ible and about 25 objects in the near-IRwavelength range for the taxonomyanalysis and low-dispersion spec-troscopy in the visible and near-IR ofabout 15–20 objects each for the chem-ical studies. The ESO ProgrammeCommittee has allocated in total 242hours of VLT time and 3 nights at theNTT for this Large Programme. Theworkhorse instruments for the pro-posed observations are FORS1 andISAAC at the VLT Unit Telescopes 1(Antu) and 3 (Melipal). Broadband pho-tometry of some brighter targets is alsoconducted with SUSI2 at the La SillaNTT.

Observing modes and target se-lection: The majority of the photometrytargets are observed in service mode(SM) at the VLT with the exceptions ofbrighter ones that are either imaged atthe NTT or during spectroscopy runs invisitor mode (VM). Experience fromprevious programmes has shown thatthe SM targets have to be selectedfrom amongst objects that allow safedetection in the fields of view of the in-strument; in practice only objects withat least 2 observed oppositions appearto be ‘reliable’ to this respect. Less wellobserved targets have a high risk of be-ing missed in the SM exposures that

typically happen about one year afterthe last object positions were meas-ured. For the BVRI measurements weselected targets with V brightness of22.5–24mag, while for JHK photometrythe objects need to be brighter than22.5mag in V. The SM targets are re-quested to be observed under clear tophotometric dark sky conditions withseeing better than 0.8″. These con-straints together with the chosen inte-gration times guarantee a minimum sig-nal-to-noise ratio (S/N) for the individ-ual objects of 30–50 in BVRI and 20-40in JHK. In terms of fulfilment of the ob-serving requirements this part of ourprogramme can be considered 100 percent successful.

Observing strategies: The spec-troscopy part of the programme is per-formed in visitor mode since the on-linedetection of these slowly moving (typi-cally 1–2″/h) faint targets in SM wasconsidered risky and misidentificationsmay cause significant loss of observingtime. The targets had to be selectedamongst brighter objects, i.e. brighterthan 22.5 mag in V for spectroscopy inthe visible wavelength range andbrighter than 20.5mag in V for the near-IR spectroscopy. Several of the targetshad some BVRI colours measured fromprevious observations or other pro-

grammes. Nevertheless, quasi-simulta-neous broadband photometry of thetargets is taken during spectroscopyruns with the aim to verify the earlier re-sults and – most importantly for the IRspectra – to cross-calibrate the contin-uum flux in the JHK bands of the ob-jects as needed for the modelling of thespectra. Visible and near-IR spec-troscopy runs are scheduled within afew days of each other in order to guar-antee observations of the same objectunder similar phase angles. However,this scheduling usually did not allow usto correlate the rotational phase of theobjects for the measurements in bothwavelength ranges, mostly because formany objects the rotation parametersare not well known. The exposure timesof the targets aimed for a S/N of about3–10 per wavelength pixel elementwhich through wavelength binning al-lowed improvements by a factor of 3–5.Given the faintness of the objects andthe telescope/instrument capabilities,near-IR spectroscopy is the most time-consuming part of our programme evenunder favourable sky conditions, andtypically 1 to 11/2 nights of observingtime are required per target. Unfortu-nately, the visitor mode observations sofar have had an unexpectedly high‘loss’ rate of about 25 per cent due to

25

Figure 4: Spectral gradients versus orbital parameters. The plots show the measured (from BVRI photometry) spectral gradients of the ob-jects versus orbital excitation E (lower left panel), inclination (upper left panel), and perihelion distance (upper and lower right panels). Thebubble diagrams in the right panels use the orbital eccentricity (upper right) and the inclination (lower right) as size parameter of the bubbles,i.e. the smaller the bubble the lower the eccentricity and inclination, respectively. The colour coding of the dynamical classes is identical toFigure 2. The red Cubewanos beyond 41 AU mentioned in the text appear to cluster at the left end in the left panels and (as the tiny bubbles)to the right end in the right panels.

bad sky conditions and some technicalproblems. Additional overheads andfurther time loss due to weather arecaused by the need to cross-calibratethe spectra in the three near-IR wave-length regions through JHK absolutephotometry. The availability of disper-sive optics covering more than onenear-IR band at lower spectral resolu-tion could have resulted in additionalbenefits and higher efficiency for ourprogramme.

Data reduction: The basic data re-duction products of the photometry areabsolute broadband filter magnitudes inBVRI and JHK, filter colours and spec-tral gradients. As side results astromet-ric positions and a search for satellitesaround the prime targets are performed(so far without success). The magni-tude (equivalent to sizes), colour andgradient data are correlated with eachother and with orbital parameters to as-sociate and explore taxonomic classesand their colour properties through sta-tistical methods. This exercise is ap-plied to our database alone and to aneven larger dataset merging our resultswith those published in the literature.The spectra are reduced to relative re-flectivity units that allow us to determinespectral gradients and to identify ab-sorption features that may be indicativefor the surface chemistry. Each spec-trum covering a wider wavelengthrange is modelled through a Hapcke re-flectance modelling code to constrainand, if possible, to quantify the surfacechemistry of the targets. In a finalphase of the project it is planned to cor-relate spectral properties with the pho-tometric taxonomy of the objects.

Status of the programme: By theend of 2002, about 90 per cent of the

observations had been conducted andabout 60 per cent of the data reductionand analysis phase is performed. Table1 shows the number of objects ob-served per dynamical class (Cubewa-no, Plutino, SDO, Centaur) and obser-vation type (BVRI and JHK photometry,visible and near-IR spectroscopy). Upto now, the photometry worked out asplanned, while the spectroscopy suf-fered from time losses (as explainedabove). Since the object selection forSM observations had to be donemonths in advance, some of our tar-gets have been observed by othergroups in the meanwhile, and despiteall efforts, duplication of observationsturned out to be unavoidable. A few tar-gets were selected on purpose forrepetition of certain types of observa-tions in order to allow cross-checks ofour results with those of other ob-servers and to address specific un-solved questions related with the ob-jects. The most striking constraint caus-ing duplication is the fact that thesample of near-IR targets accessiblewith standard instruments at modern8–10-m class telescopes is rather limit-ed (~ 30 objects up to now).

Highlights from the FirstResults

The results outlined in this summaryare obtained from data collected duringthe first year of this ESO LargeProgramme. They are described ingreater detail in various papers that arealready published (Barucci et al., 2002,A&A 392; Boehnhardt et al. 2002, A&A395) or will be published soon (Dotto etal., 2003, Icarus, in press; Lazzarin etal. 2003, AJ, in press). Another threepapers with results from Period 69 arein preparation for submission to scien-tific journals in early 2003.

BVRI colour-colour-plots andspectral gradient histograms: Thedatabase used here contains 96 ob-jects compiled from our programme(28) and from published data. Thecolour-colour-plots and photometric“reflectivity spectra” of the 28 objectsmeasured by our programme in Period67 are displayed in Figures 1 and 2, re-spectively. The colour-colour plots(Fig. 1) show that the objects follow –with some scatter – the lines of con-stant reddening in BVR colours, butthey start deviating in the transition re-

26

Object class BVRI JHK vis.spec. IR spec.*

Cubewanos 25/15 7/4 1/1 0/0 Plutinos 19/10 7/4 5/4 4/3Scattered Disk 13/7 6/5 3/2 3/2Centaurs 9/7 6/5 6/6 5/4

Total 66/39 26/18 15/13 12/9

Figure 5: JHK colour-colour plots of the first 8 objects measured in the SM photometry part of our programme. H-K versus J-H is displayedin the left panel, the right one shows H-K versus J-K. For explanations of symbols and lines see Figure 1. The objects are identified with smalllabels at the symbols.

Table 1. Physical studies of TNOs and Centaurs: Number of objects observed/reduced perdynamical class (status: end 2002).

*About 30 % of the object spectra are incomplete, i.e. JHK bands are not fully covered.

gion to the near-IR. The latter is a con-sequence of the slope change seen inthe spectra of red to very red objectsbeyond 750 nm. The reddening distri-bution seems to be continuous withouta clear indication for a bi-modality asannounced by other groups usingsmaller datasets. In the spectral gradi-ent histogram of Figure 3 all 96 objectsare used and sorted for each dynamicalclass in the same gradient bins of widthcorresponding to the averaged uncer-tainty of the data. With one exceptions(1994 ES2) the spectral gradients of allobjects fall between –5 and 55 %/100nm. Outlyers in previous datasetsturned out to be artefacts, most likelydue to observations and data reduction(blends of background objects) or un-usual object behaviour (for instance1994 EV3 shows rapid and large-ampli-tude variability that mimics unusualcolours even in short-term photometricfilter sequences as identified throughour programme). The Cubewanosshow a pronounced peak for spectralgradients between 25 and 45 %/100nm that is caused by a population of redobjects in circular and low inclinationorbits beyond 41 AU. At least for thePlutinos, and possibly also for Centaursand SDOs, the distribution of spectralgradients in the visual wavelength re-gion seems to be distinctly differentfrom those of the Cubewanos.

Correlations with orbital parame-ters: In Figure 4 we show the meas-ured photometric spectral gradientsversus orbital excitation E, inclinationand perihelion distance of the objects.The excitation parameter is defined asE = ( e2 + sin2 ( i ) )1/2 with e and i aseccentricity and inclination of the orbit,

respectively. E is an estimate for the ve-locity of the object with respect to an-other one at the same distance, but incircular orbit in the Ecliptic, and it mightthus be related to the collision velocityand/or the collision probability. Theplots of the spectral gradients versus Eand i suggest a correlation betweensurface colour and collision history ofthe objects, since the range of spectralgradients increases with increasing ex-citation parameter E and decreaseswith decreasing inclination. On the oth-er side, the plot of the spectral gradi-ents versus perihelion distance showsa clustering of very red Cubewanos incircular and Ecliptic orbits beyond ~ 41AU. The peak for the Cubewanos in thespectral gradient statistics of Figure 3 iscaused by exactly this group of objects.Moreover, for objects with perihelia be-tween ~ 36 and 40 AU the width of thereddening range seems to increase thecloser the object approaches the Sun.At present, these trends are not yet ful-ly established since more objects needto be included for statistical reasons.However, if the trends can be con-firmed, they would argue for differentreddening scenarios, i.e. the trendswith excitation and inclination for thebalance between space weatheringand collision resurfacing on one sideand the trend with perihelion distancefor the balance between space weath-ering and intrinsic activity on the otherside. Most puzzling is the group of redCubewanos beyond 41 AU and there isno easy explanation for their existenceat present.

JHK colours: While at visible wave-lengths a certain reddening rangeseems to be continuously populated, a

preference for neutral colours is foundamong the near-IR spectra of the ob-jects – see Figure 5. Only 2 objects aredefinitely redder than the Sun. Whetherthe red objects are members of a sec-ond colour population can easily bequestioned. The overall interpretationof the visible to near-IR colour range ar-gues for only minor colour changes inthe infrared part of the spectra, but fora significant and very wide absorptionof variable depth for many TNOs andCentaurs in the visual wavelengthrange. This implies that the observedreddening in BVRI is in fact the long-wave wing of a wide absorption featurefrom a yet unknown chemistry that ex-tends beyond the visible region wellinto the UV.

Visible spectra: The spectra ofTNOs and Centaurs observed in thevisible are mostly featureless and showconstant (though different) slopes atwavelengths below ~ 750 nm (Fig. 6).At larger wavelength a continuous gra-dient change towards smaller slopes isobserved in the red objects. With oneexception (possibly caused by rotationeffects in the photometry) the spectralgradients obtained from spectra are ingood agreement with the photometricones. It came as a surprise that wideabsorption dips were found in the visi-ble spectra of two Plutinos, 2000 EB173and 2000 GN171, taken in April 2001(see Fig. 7). The absorption featuresare centred around 600 and 730 nmfor 2000 EB173 and at 725 nm for 2000GN171. There are two very puzzling as-pects with these detections: (1) Thefeatures were not confirmed with spec-tra taken about one year later (seeFig. 7); this would argue for a non-uni-

27

Figure 6: Visible Reflectivity spectra of TNOs and Centaurs. The spectra of TNOs are displayed in the left panel, those of Centaurs in theright one. The spectra are shifted in Y direction for clarity. The dots represent the respective results from quasi-simultaneous photometry ofthe objects taken during the spectroscopy runs.

form distribution of the absorbing mate-rial on the surface of the objects. (2)Features resembling the absorptions inthese two Plutinos and identified inother solar system bodies and in labexperiments are produced by silicatematerial on asteroids that was exposedto long-term aqueous alteration on thesurface. A similar scenario in the ex-tremely cold and icy environment of theouter solar system obviously posessome problems. The answer to thispuzzle calls for a systematic study ofthe objects by rotation-phase resolvedspectroscopy, and, if confirmed, it will

have a great impact on our under-standing of the physical nature of thesebodies at the edge of the planetarysystem.

Near-IR spectra: We have foundwater ice absorption in H and K bandspectra of 1 TNO and 3 Centaurs (seefor instance the 2001 PT13 spectrum inFigure 8). No features are seen in thespectra of other objects in our sample.For 2001 PT13, a non-uniform water icedistribution on a large scale is conclud-ed from the existence and absence ofthe absorption features in the spectrameasured at two different epochs. The

model fits to thecombined near-IRand visible spec-tra use a radiativetransfer code withsimple geographi-cal mixtures of or-ganics, mineralsand ices. The steepslopes in the visi-ble require the ad-mixture of Tholinsor kerogene in themodel fits. Never-theless, despitethe decent matchthat is achievedbetween the ob-served spectra andthe radiative trans-fer model results,at present our con-clusions are limit-ed. On one hand,water ice seemsto be present in

these distant bodies; however, on theother hand, the identification of furthersurface materials as well as quantita-tive estimations are difficult because ofthe quality of the data fits and the exis-tence of multiple model solutions foreach spectrum.

Conclusion and Outlook

The ESO Large Programme on thephysical studies of TNOs and Centaursis now approaching the end of its as-signed period for the observations, andit is a little more than half way throughthe data reduction. The analysis of theresults and model interpretation willcertainly continue for another year be-yond the official end of the observingprogramme.

With the data from this programme ithas been possible to identify the firsttaxonomic group in the Edgeworth-Kuiper Belt, the cluster of red Cube-wanos beyond 41 AU, and to get clearindications of resurfacing effects onTNOs and Centaurs. At the moment,the observed results appear to besomewhat contradictory to each otherand a synoptic picture for their interpre-tations has not yet evolved. Water iceseems to be abundant in these objects;however, the identification of other sur-face ice features appears to be a diffi-cult task. Spectra of two Plutinos mayhave provided the first glance of stonymaterial, and if so, they contribute yetanother puzzle with the potential todrastically change our view of these icyworlds in the outer solar system.

28

Figure 7: Two “puzzling” Plutinos: reflectivity spectra of 2000 EB173 and 2000 GN171. The reflectivity spectra of both objects show shallow andwide absorption features during one observing epoch (left panel), but a straight featureless slope during a second observing run about oneyear later (right panel). The dots and triangles in the panels indicate the corresponding ‘broadband’ reflectivity from quasi-simultaneous BVRIfilter photometry of the objects.

Figure 8: Near-IR spectrum of the Centaur 2001 PT13. The plotshows the near-IR spectrum of the object on 10/9/2001 (upper spec-trum) and on 8/10/2001 (lower spectrum). The continuous andbroken lines show model fits to the observed spectra using the ma-terial mixture as indicated in the figure insert. The dots indicate thereflectivity from quasi-simultaneous broadband JHK photometry ofthe object.

1. Introduction

Despite the recent extraordinaryprogress in observational cosmologyand the successful convergence on asingle cosmological model, the historyof galaxy and structure formation andevolution remains still an open ques-tion. One of the most actively debatedissues is how and when the present-day most massive galaxies (e.g. ellipti-cal galaxies and bulges with Mstars >1011 M) built up and what type of evo-lution characterized their growth overthe cosmic time. Addressing this ques-tion is important not only to test the dif-ferent scenarios of galaxy formation,but also to understand how the generalstructures of the universe evolved.

There is wide consensus that galaxyformation takes place by a combinationof dissipational collapse and merging ofsub-units (e.g. White & Rees 1978).However, much less consensuspresently exists on the main epoch(redshift) at which these events tookplace, and then when the bulk of thestellar mass of present-day galaxieswas formed and assembled. Hierar-chical merging models (HMMs) have sofar favoured a rather late appearance ofmassive galaxies, thus only quite re-cently reaching their present mass, i.e.,at z < 1 –1.5 (e.g. Baugh et al. 2002).On the other hand, other scenarios andsome observational evidence suggestthat the bulk of star formation andgalaxy assembly took place at substan-tial higher redshifts (see Cimatti 2003for a recent review). Such scenariosmake very different predictions de-pending on the formation epoch. Ifmassive galaxies formed at z > 2–3through a short-lived and intense star-burst phenomenon followed by a pas-sive and pure luminosity evolution(PLE), then old (a few Gyr) passivelyevolving massive systems should existup to z ~ 1 – 1.5 with a number densityidentical to that observed in the localuniverse because their number doesnot change through cosmic time andthe only evolution is the aging of thestellar population. On the other hand, ifthe most massive galaxies reach theirfinal mass at z < 1–1.5, they should be

rare at z ~ 1–1.5, with a comoving den-sity of Mstars > 1011 M galaxies de-creasing by almost an order of magni-tude from z ~ 0 to z ~ 1 (Baugh et al.2002).

Although there is a general agree-ment on cluster ellipticals being a ho-mogeneous population of old and pas-sive systems formed at high redshifts(see Renzini 1999 and Peebles 2002for recent reviews), the question of fieldspheroids is still controversial. Severalobservations were performed over re-cent years in order to test the two com-peting models. However, although it isestablished that old, passive and mas-sive systems exist in the field out toz ~1.5, there is not yet agreement onhow their number and evolutionaryproperties compare with the model pre-dictions.

2. The K20 Survey

A solid and unbiased approach to in-vestigate the evolution of massive

galaxies is to study samples of fieldgalaxies selected in the K-band (i.e. at2.2 µm; Broadhurst et al. 1992; Kauff-mann & Charlot 1998). This has twomain advantages. Firstly, since therest-frame optical and near-IR light is agood tracer of the galaxy stellar mass(Gavazzi et al. 1996), K-band surveysselect galaxies according to their massup to z ~ 2. Secondly, the similarity ofthe spectral shapes of different galaxytypes in the rest-frame optical/near-IRmakes the K-band selection free fromstrong biases against or in favour ofparticular classes of galaxies. In con-trast, the selection of high-z galaxies inthe observed optical bands is sensitiveto the star-formation activity rather thanto the stellar mass because it samplesthe rest-frame UV light and makes opti-cal samples biased against old passivegalaxies.

Motivated by the above open ques-tions and by the availability of the ESOVLT telescopes, we started an ESOVLT Large Programme (dubbed “K20

29

The K20 Survey: New Light on GalaxyFormation and EvolutionA. CIMATTI1, E. DADDI2, M. MIGNOLI3, L. POZZETTI3, A. FONTANA4, T. BROADHURST5,F. POLI6, P. SARACCO7, A. RENZINI2, G. ZAMORANI3, N. MENCI4, S. CRISTIANI8, S. D’ODORICO2, E. GIALLONGO4, R. GILMOZZI2, S. DI SEREGO ALIGHIERI1, J. VERNET1

1INAF – Osservatorio Astrofisico di Arcetri, Italy; 2European Southern Observatory, Garching, Germany;3INAF – Osservatorio Astronomico di Bologna, Italy; 4INAF – Osservatorio Astronomico di Roma, Italy;5Racah Institute for Physics, The Hebrew University, Jerusalem, Israel; 6Dipartimento di Astronomia, Universitàdi Roma, Italy; 7INAF – Osservatorio Astronomico di Brera, Italy; 8INAF– Osservatorio Astronomico di Trieste, Italy

Figure 1: Zoom on two-dimensional sky-subtracted spectra of faint galaxies present in a sec-tion of one of the FORS2 MXU masks used for the “dithered” observations.

survey”) based on 17 nights distributedover two years (1999–2000) (seehttp://www.arcetri.astro.it/~k20/ andCimatti et al. 2002c). The prime aim ofthis survey was to derive the spectro-scopic redshift distribution and thespectral properties of 546 Ks-selectedobjects with the only selection criterionbeing Ks < 20 (Vega scale). Such athreshold is critical because it selectsgalaxies over a broad range of masses,i.e. Mstars > 1010 M and Mstars > 4 ×1010 M for z = 0.5 and z = 1 respec-tively. The targets were selected fromKs-band images (ESO NTT+SOFI) oftwo independent fields covering a totalarea of 52 arcmin2: a 32.2 arcmin2

sub-area of the Chandra Deep FieldSouth (CDFS; Giacconi et al. 2001),and a 19.8 arcmin2 field centred aroundthe QSO 0055-269 at z = 3.6. Exten-sive simulations showed that the sam-ple is photometrically highly completeto Ks = 20.0 and not affected by strongselection effects on the galaxy popula-tions, with the exception of an underes-timate of ~ 0.2 magnitudes for the lumi-nous galaxies (Cimatti et al. 2002c). H0= 70 km s–1 Mpc–1, Ωm = 0.3 and ΩΛ =0.7 are adopted throughout this article.

3. The VLT Observations:Optimizing Spectroscopy forFaint Red Galaxies

The most crucial probes of massivegalaxy evolution are galaxies with thevery red colours expected for passivelyevolving systems at z > 1 (e.g. R – Ks> 5). For Ks < 20, such red colours im-ply very faint optical magnitudes (R ~24–26), close to the spectroscopic lim-its of 8-m-class telescopes. Moreover,the most prominent spectral features(e.g. the 4000 Å break and H&K ab-sorptions) fall in the very red part of theobserved spectra for 1 < z < 2 (λobs >8000 Å), where the strong OH sky linesand the CCD fringing (very strong inFORS2 before its recent upgrade ofMarch 2002 with new red-optimizedMIT CCD mosaic) make the spectros-copy of faint z > 1 galaxies extremelychallenging.

For such reasons, the spectroscopywas optimized to reach the highest pos-sible signal-to-noise ratio in the red byapplying the innovative technique of“dithering” the targets along the slitsevery 15–20 minutes between two (ormore) fixed positions. Such a method(routinely used in the near-IR) has thecrucial advantage of efficiently sub-tracting both the CCD fringing and theOH sky lines. A similar method can alsobe performed with the so-called “nodand shuffle” technique by nodding thetelescope rapidly between targets andadjacent sky positions and recordingobject and sky spectra on adjacent re-gions of a low-noise CCD throughcharge shuffling. At the time of the ob-servations, no automatic templates

30

Figure 2: Examples of FORS2 spectra of high-z galaxies with the four adopted spectral clas-sifications. From top to bottom: an early-type at z = 1.096 (R = 23.0), an early-type +[OII]λ3727 emission at z = 0.735 (R = 23), an emission line galaxy at z = 1.367 (R = 23.0),an absorption line galaxy at z = 1.725 (R = 23.5).

Figure 3: ISAAC two-dimensional sky-subtracted H-band low-resolution spectrum of a galaxyat z = 1.73 (the emission line is Hα).

spectroscopy and/or expected to be ina redshift range without strong featuresin the observed spectral range (e.g. 1.4< z < 2.0). However, the absence of amultiobject spectroscopic mode, eachintegration being limited to a singlespectral band, and the lack of veryaccurate photometric redshifts at thetime of the obser-vations made itdifficult to derivea substantial num-ber of spectro-scopic redshiftsbased on near-IR spectroscopyalone. Overall,redshifts were de-

rived for four galaxies at 1.3 < z < 1.9out of the 22 observed (Figs. 3–4).

In addition to spectroscopy, U B V RI z J Ks imaging from NTT and VLT ob-servations was also available for bothfields, thus providing the possibility toestimate photometric redshifts for allthe objects in the K20 sample, to opti-mize them through a comparison withthe spectroscopic redshifts and to as-sign reliable photometric redshifts tothe objects for which it was not possibleto derive the spectroscopic redshift.The dispersion on the global sample(CDFS + 0055-269 fields) is σ = 0.089and the average is < zspe – zphot > =0.012 (Fig. 5). Such an accuracy is atthe level of the “state-of-the-art” photo-metric redshifts obtained in the HubbleDeep Fields. The final sample coversa redshift range of 0 < z < 2.5 and theoverall spectroscopic redshift com-pleteness is 94%, 92%, 87% for Ks <19.0, 19.5, 20.0 respectively. The over-all redshift completeness (spectroscopicand photometric) is 98%.

The K20 survey represents a signifi-cant improvement with respect to previ-ous surveys for faint K-selected galax-ies (e.g. Cowie et al. 1996; Cohen et al.1999), thanks to its high spectroscopicredshift completeness, the larger sam-ple, the coverage of two fields (thus re-ducing the cosmic variance effects),and the availability of optimized photo-metric redshifts.

4. The Redshift Distribution:Testing Galaxy-FormationModels

The observed differential and cumu-lative redshift distributions for the K20sample are presented in Fig. 6 (seeCimatti et al. 2002b), together with the

31

Figure 4: ISAAC H-band low resolution spectra of two emission-line galaxies with Hα red-shifted at z = 1.729 (top panel) and z = 1.715 (bottom panel).

were available to perform dithered ob-servations with FORS2, but we suc-ceeded in applying this technique bymoving the telescope “manually” fromone position to another thanks to thekind help of the support astronomers.Our approach proved to be very suc-cessful in achieving high-quality spec-tra extended to the far red (Fig. 1) andreaching a spectroscopic redshift com-pleteness for the high-z red galaxiesmuch higher than in previous surveys.

Optical multi-object spectroscopywas obtained with the ESO VLT UT1and UT2 equipped respectively withFORS1 (with the MOS mode: 19 mov-able slitlets) and FORS2 (with the MaskExchange Unit, MXU, mode). The MXUmode uses laser-cut slit masks pre-pared by the mask manufacturing unit(MMU), and allowed us to observe upto 52 targets per mask with slit lengthsof 8″–15″ (a “record” for the multiplex inthe pre-VIMOS era!). The seeing andthe slit widths were in the range of0.5″–2.0″ and 0.7″–1.2″ respectively.The integration times ranged from 0.5to 3 hours. Figure 2 shows some typi-cal spectra.

A fraction of the sample was also ob-served with near-IR spectroscopy usingVLT UT1 + ISAAC in low-resolutionmode with a 1″ slit, 1–3 hours integra-tion times, 1–2 targets per slit, in orderto attempt to derive the redshifts of thegalaxies which were too faint for optical

Figure 5: The com-parison betweenspectroscopic(x-axis) andphotometric (y-axis)redshifts of thegalaxies in the K20survey. Filled circles,filled squares andopen symbolsindicate the0055-269 field, theCDFS and the “lowerquality” spectro-scopic redshiftsrespectively.

predictions of different scenarios ofgalaxy formation and evolution, includ-ing hierarchical merging models(HMMs) from Menci et al. (2002, M02),Cole et al. (2000, C00), Somerville etal. (2001, S01), and pure luminosityevolution models (PLE) based onPozzetti et al. (1996, PPLE) and Totaniet al. (2001, TPLE). No best tuning ofthe models was attempted in this com-parison, thus allowing an unbiasedblind test with the K20 observationaldata. The predicted N(z) from the mod-els are normalized to the K20 surveysky area.

The observed redshift distributioncan be retrieved from http://www.arcetri.astro.it/ k20. The spike at z ~ 0.7is due to two galaxy clusters (or richgroups) at z = 0.67 and z = 0.73. Themedian redshift of N(z) is zmed = 0.737and zmed = 0.805, respectively with andwithout the two clusters being included.The contribution of objects with only aphotometric redshift becomes relevantonly for z > 1.5.

Figure 6a shows fairly good agree-ment between the observed N(z) distri-bution and the PLE models (with theexception of PPLE with Salpeter IMF),although such models slightly overpre-dict the number of galaxies at z > 1.2.However, if the photometric selectioneffects present in the K20 survey(Cimatti et al. 2002c) are taken into ac-count, the PLE models become muchcloser to the observed N(z) thanks tothe decrease of the predicted high-ztail.

In contrast, all the HMMs underpre-dict the median redshift, overpredict thetotal number of galaxies with Ks < 20 byfactors up to ~ 50% as well as the num-ber of galaxies at z < 0.5, and under-predict the fractions of z >1–1.5 galax-ies by factors of 2–4 (Fig. 6b). The in-clusion of the photometric selection ef-fects (Cimatti et al. 2002c) exacerbatesthis discrepancy, as shown in Figure 6b(right panels) for the M02 model (thediscrepancy for the C00 and S01 mod-els becomes even larger). The deficit ofhigh-z objects in HMMs is well illustrat-ed by Figure 7, where the PPLE modelreproduces the cumulative number dis-tribution of galaxies at 1 < z < 3 within1–2 σ, whereas the M02 model is dis-crepant at ≥ 3σ level (up to > 5σ for 1.5< z < 2.5). This conclusion does notstrongly depend on the objects withonly photometric redshifts: the 7 galax-ies with spectroscopic redshift z > 1.6are already discrepant with the predic-tions by HMMs of basically no galaxieswith Ks < 20 and z > 1.6.

5. The Luminosity Function andLuminosity Density Evolution

The evolution of galaxies can be in-vestigated through the variation of theirnumber density and luminosity (i.e. the

32

Figure 6a: Top panels: the observed differential N(z) for Ks < 20 (histogram) compared withthe PLE model predictions. Bottom panels: the observed fractional cumulative redshift distri-bution (continuous line) compared with the same models. The shaded histogram shows thecontribution of photometric redshifts. The bin at z < 0 indicates the 9 objects without redshift.The left and right panels show the models without and with the inclusion of the photometricselection effects respectively. Sc and Sp indicate Scalo and Salpeter IMFs respectively.Figure 6b: Same as Fig. 6a, but compared with the HMM predictions. Right panels: the M02model with the inclusion of the photometric selection effects.

luminosity function, LF) over the cosmictime.

The LF of the K20 galaxies has beenestimated in the rest-frame Ks-bandand in three redshift bins (zmean = 0.5,1.0, 1.5) in order to trace its evolution(Pozzetti et al. 2003). A comparisonwith the local universe Ks-band LF ofCole et al. (2001) shows a mild lumi-nosity evolution of LF(z) out to z =1, i.e. the LF shifts along the X-axisas a function of z without an apprecia-ble change in the number density(Y-axis). The brightening is approxi-mately –0.5 magnitudes from z = 0 to z= 1 (Fig. 8).

The size and completeness of theK20 sample allowed us also to studythe LF evolution by galaxy spectral orcolour types, and showed that redearly-type galaxies dominate the brightend of the LF already at z ~ 1, and thattheir number density shows only asmall decrease from z ~ 0 to z ~ 1(Pozzetti et al. 2003).

The PLE models describe reason-ably well the shape and the evolution ofthe observed luminosity function up toat least zmean = 1.0, with no evidence fora strong decline in the abundance ofthe most luminous systems (with L > L*)(Fig. 9b). Hierarchical merging modelsoverpredict faint sub-L* galaxies at 0 <z < 1.3, and underpredict the density ofluminous (MKs < –25.5 + 5logh70) gal-axies at the bright end of the LF(Fig. 9a). A comparison between the R– Ks colours and luminosity distribu-tions of K20 galaxies with 0.75 < z < 1.3(a bin dominated by spectroscopic red-

shifts) to the predictions of the GIF1

simulations (Kauffmann et al. 1999)highlights a discrepancy between thetwo distributions: real luminous galax-ies with MK – 5logh70 < –24.5 in the K20sample have a median colour of R – Ks~ 5, whereas the GIF simulated galax-ies have R – Ks ~ 4, with a small over-lap of two colour distributions (Fig. 10).This is probably due to an underesti-mate of Mstars/L and of the numberdensity of massive galaxies at z ~ 1 bythe HMMs.

An additional way to investigategalaxy evolution is to trace the integrat-ed cosmic emission history of galaxiesat different wavelengths through thestudy of the luminosity density ρλ (WHz–1 Mpc–3). Such an approach is inde-pendent of the details of galaxy evolu-tion and depends mainly on the star-formation history of the universe.Attempts to reconstruct the cosmic evo-lution of ρλ were made previously main-ly in the UV and optical bands, i.e. fo-cusing on the star-formation history ofgalaxies. Our survey offered for the firsttime the possibility to investigate in thenear-IR, thus providing new clues onthe global evolution of the stellar massdensity (Pozzetti et al. 2003). The re-sults show an evolution of ρλ(z) = ρλ(z= 0)(1 + z)0.37, much slower than theone derived in the UV-optical, typically∝ (1 + z)3–4. Such a slow evolution sug-gests that the stellar mass densityshould also evolve slowly at least up toz ~1.3 (see also Dickinson et al. 2003).

The estimate of the stellar mass func-tion and its cosmic evolution is underway and will provide additional con-straints on the evolution of massivesystems (Fontana et al. 2003, in prepa-ration).

6. Extremely Red Objects(EROs)

As discussed in previous sections,Extremely Red Objects (EROs, i.e.galaxies with R – K > 5) are very im-portant because they allow us to selectold passively evolving galaxies at z > 1.Thanks to our deep red-optimized VLToptical spectroscopy, for a fraction ofEROs (70% to Ks < 19.2) present in theK20 sample it was possible to derive aspectroscopic redshift and a spectralclassification (Cimatti et al. 2002a),thus, for the first time, allowing us toverify if the ERO population is indeeddominated by old passive galaxies. TheVLT optical spectra show that twoclasses of galaxies at z ~ 1 contributenearly equally to the ERO population:the expected old stellar systems, butalso a substantial fraction of dustystar-forming galaxies (Fig. 11).

The colours and spectral propertiesof old EROs are consistent with ≥ 3 Gyrold passively evolving stellar popula-tions (assuming solar metallicity andSalpeter IMF), requiring a formationredshift zf > 2.4 (Fig. 12). The numberdensity is 6.3 ± 1.8 × 10–4 h3Mpc–3 forKs < 19.2, consistent with the expecta-tions of PLE models for passivelyevolving early-type galaxies with similar

33

1http://www.mpa,garching.mpg.de/GIF/

Figure 8: The rest-frame Ks-band Luminosity Function in three red-shift bins. Solid curves: the Schechter fits. Dotted curve: the localKs-band LF of Cole et al. (2001). Open circles: spectroscopic red-shifts, filled circles: spectroscopic + photometric redshifts.

Figure 7: The observed cumulative number of galaxies between 1 <z < 3 (continuous line) and the corresponding poissonian ±3σ confi-dence region (dotted lines). The PPLE (Scalo IMF) and the M02models are corrected for the photometric biases.

formation redshifts (Cimatti et al.2002a). HMMs underpredict such oldred galaxies at z ~ 1 by factors of ~ 3(Kauffmann et al. 1999) and ~ 5 (Coleet al. 2000).

The spectra of star-forming EROssuggest a dust reddening of E(B – V) ~0.5–1 (Fig. 12, adopting the Calzetti ex-tinction law), implying typical star-for-mation rates of 50–150 Myr–1, and asignificant contribution (> 20–30%) tothe cosmic star-formation density at z ~1. Such dusty star-forming systems arealso underpredicted by HMMs. For in-stance, the GIF simulations predict acomoving density of red galaxies withSFR > 50 Myr–1 that is ~ 30 timeslower than the observed density ofdusty EROs.

Taking advantage of the spectro-scopic redshift information for the twoERO classes, we compared the relative3D clustering in real space (Daddiet al. 2002). The comoving correlationlengths of dusty and old EROs are con-strained to be r0 < 2.5 and 5.5 < r0 < 16h–1 Mpc comoving respectively, imply-ing that old EROs are the main sourceof the ERO strong angular clustering. Itis important to notice that the strongclustering measured for the old EROsis in agreement with the predictions ofhierarchical merging (Kauffmann et al.1999).

7. Summary and Outlook

The high level of redshift complete-ness of the K20 survey and the resultspresented in previous sections providenew implications for a better under-standing of the evolution of “mass-se-lected” field galaxies.

Overall, the results of the K20 surveyshow that galaxies selected in the

Ks-band are characterized by little evo-lution up to z ~ 1, and that the observedproperties can be successfully de-scribed by a PLE scenario. In contrast,HMMs fail in reproducing the observa-tions because they predict a sort of “de-layed” scenario in which the assemblyof massive galaxies occurs later thanactually observed. However, it is impor-tant to stress here that the above re-sults do not necessarily mean that thewhole framework of hierarchical merg-ing of CDM halos is under discussion.For instance, the strong clustering ofold EROs and the clustering evolutionof the K20 galaxies (irrespective ofcolours) seem to be fully consistentwith the predictions of CDM models oflarge-scale structure evolution (Daddiet al. 2003 in preparation).

It is also important to stress that theK20 survey allows us to perform tests

which are sensitive to the evolutionary“modes” of galaxies rather than to theirformation mechanism. This means thatmerging, as the main galaxy formationmechanism, is not ruled out by thepresent observations. Also, it should benoted that PLE models are not a phys-ical alternative to the HMMs, but rathertools useful to parameterize the evolu-tion of galaxies under three main as-sumptions: high formation redshift, con-servation of number density throughcosmic times, and passive and lumi-nosity evolution of the stellar popula-tions.

Thus, if we still accept the ΛCDMscenario of hierarchical merging of darkmatter halos as the basic framework forstructure and galaxy formation, the ob-served discrepancies highlighted by theK20 survey may be ascribed to how thebaryon assembly, the star-formation

34

Figure 9a: The Ks-band LF compared to hierarchical merging model predictions. Figure 9b: The Ks-band LF compared to PLE model predic-tions.

Figure 10: Left panel: R –Ks colours vs. rest-frameabsolute Ks magnitudesfor z = 1.05 GIF simulat-ed catalogue (small dots)and data (circles) at 0.75< z < 1.3 (zmean = 1)(empty and filled circlesrefer to z < 1 and z > 1respectively). Verticaldashed line representsapproximately thecompleteness magnitudelimit of GIF catalogue cor-responding to its masslimit (see text). Right pan-el: Colour distribution ofluminous galaxies (MKs –5 log h70 < –24.5)observed (dotted line)and simulated(continuous line), normal-ized to the same comov-ing volume.

processes and their feedback are treat-ed by HMMs both within individualgalaxies and in their environment. Ourresults suggest that HMMs should havegalaxy formation in a CDM-dominateduniverse that closely mimics theold-fashioned monolithic collapse sce-nario. This requires enhanced mergingand star formation in massive halos athigh redshift (say, z > 2–3), and alsosuppressed star formation in low-masshalos.

In summary, the redshift distributionof Ks < 20 galaxies, together with thespace density, nature, and clusteringproperties of the ERO population, andthe redshift evolution of the rest-framenear-IR luminosity function and lumi-nosity density provide a new invaluableset of observables on the galaxy popu-lation in the z ~ 1–2 universe, thusbridging the properties of z ~ 0 galaxieswith those of Lyman-break and sub-mm/mm-selected galaxies at z ≥ 2–3.This set of observables poses a newchallenge for theoretical models toproperly reproduce.

The K20 survey is triggering severalstudies in which our team is actively in-volved. Ultradeep spectroscopy wasobtained in November 2002 withFORS2 (MXU mode and with the newred-optimized MIT CCD mosaic) in or-der to derive the redshifts and the na-ture of the previously unidentifiedEROs, to increase the spectroscopicredshift completeness of the whole K20sample, and to study (with higher reso-lution spectroscopy) the kinematics of z> 1 galaxies in order to estimate theirdynamical masses. Since the sub-areaof the CDFS observed in the K20 sur-vey is also a target of the HST+ACSGOODS Treasury Programme (PI M.Giavalisco), and of the SIRTF GOODS(PI M. Dickinson) and SWIRE (PI C.

Lonsdale) Legacy Programmes, theK20 database, complemented by suchadditional observations, will allow us toderive new constraints on the formationand evolution of galaxies out to higherredshifts.

We thank the VLT support astron-omers for their competent assistanceduring the observations. AC warmlythanks ESO (Garching) for the hospital-ity during his visits, and Jim Peeblesand Mark Dickinson for useful and stim-ulating discussions. We also thankPiero Rosati and Mario Nonino for pro-viding the reduced and calibrated BVRIFORS1 images of the CDFS, and C.Baugh, R. Somerville and T. Totani forproviding their model predictions.

References

Baugh C.M. et al. 2002, in proceedings ‘TheMass of Galaxies at Low and HighRedshift’, Venice 2001, eds. R. Bender, A.Renzini, astro-ph/0203051.

Broadhurst, T., Ellis, R.S., & Grazebrook, K.1992, Nature, 355, 55.

Cimatti A., Daddi E., Mignoli M. et al. 2002a,A&A, 381, L68.

Cimatti A., Pozzetti L., Mignoli M., et al.2002b, A&A, 391, L1.

Cimatti A., Mignoli M., Daddi E. et al. 2002c,A&A, 392, 395.

Cohen J.G., Blandford R., Hogg D.W. et al.1999, ApJ, 512, 30.

Cole S., Lacey C.G., Baugh C.M. & FrenkC.S. 2000, MNRAS, 319, 168.

Cole S., Norberg P., Baugh C.M. et al. 2001,MNRAS, 326, 255.

Cowie, L.L., Songaila A., Hu E.M. & CohenJ.G. 1996, AJ,112,839.

Daddi E., Cimatti A., Broadhurst T. et al.2002, A&A, 384, L1.

Dickinson M., Papovich C., Ferguson H.C.,Budavari T. 2003, ApJ, in press, as-troph/0212242.

Gavazzi G., Pierini D. & Boselli A. 1996,A&A, 312, 397.

Giacconi R., Rosati P., Tozzi P. et al. 2001,ApJ, 551,624.

Kauffmann G., Charlot S. 1998, MNRAS,297, L23.

Kauffmann G. et al. 1999, MNRAS, 303,188.

Menci N., Cavaliere A., Fontana A.,Giallongo E. & Poli F. 2002, ApJ, 575, 18.

Peebles P.J.E. 2002, astro-ph/0201015.Pozzetti L. Bruzual A.G. & Zamorani, G.

1996, MNRAS, 281, 953.Pozzetti L. et al. 2003, A&A, in press.Renzini A. 1999, in The Formation of

Galactic Bulges, ed. C.M. Carollo, H.C.Ferguson, & R.F.G. Wyse (Cambridge:CUP), p. 9.

Somerville R.S, Primack J.R. & Faber S.M.2001, MNRAS, 320, 504.

Totani T., Yoshii Y., Maihara T., Iwamuro F. &Motohara K. 2001, ApJ, 559, 592.

White S.D.M., Rees M.J. 1978, MNRAS,183, 341.

35

Figure 11: Theaverage rest-frame spectra ofold passivelyevolving (top;zmean = 1.000)and dustystar-formingEROs (bottom;zmean = 1.096)with Ks ≤ 20(Cimatti et al.2002a).

Figure 12: The av-erage spectra ofEROs (black) com-pared to templatespectra (red ) ofBruzual & Charlotsimple stellar popu-lation (top panel),and of dustystarbursts (Kinney etal. 1996, middle pan-el; Poggianti & Wu2000, bottom panel).

36

Tracing the Formation and Evolution ofClusters and their Central Massive Galaxies toz > 4: a Progress ReportB. VENEMANS1, G. MILEY1, J. KURK1, H. RÖTTGERING1, L. PENTERICCI2

1Leiden Observatory, Leiden, The Netherlands; 2MPIA, Heidelberg, Germany

1. Introduction: The HostGalaxies of Distant RadioGalaxies

Luminous distant radio galaxies areunique probes of the early Universe.About 15 years ago at Leiden some ofus started a programme to find distantradio galaxies, study their propertiesand use them as probes of the earlyUniverse. The project began as a resultof the technique we developed duringthe late eighties to pinpoint distant radiogalaxies, based on their ultra-steep ra-dio spectra (spectral index α < –1).Finding these objects was the subjectof an ESO Key Programme in the earlynineties, that discovered a substantialfraction of presently known distant ra-dio galaxies (e.g. see Messenger arti-cles Miley et al. 1992; Röttgering, Miley,& van Ojik 1996). There are now around150 z > 2 radio galaxies known, the ma-jority of which were found using thistechnique, the most distant of which,TN J0924–2201, has a redshift of 5.19.

During the last decade multi-wave-length studies of these high-redshift ra-dio galaxies (HzRGs, z > 2) have pro-duced strong evidence that they aremassive galaxies in the process of for-mation, most probably the ancestors ofcentral cluster galaxies.

Relevant results include: • HzRGs trace the bright envelope in

the infrared Hubble diagram, implyingthat they are amongst the most mas-sive galaxies in the early Universe (e.g.De Breuck et al. 2002).

• HzRGs are usually surrounded bygiant Lyα emitting gas halos, whose sizes(∼100 kpc; van Ojik et al. 1997) are com-parable to those of cD galaxy enve-lopes (for an example, see Fig. 1).

• HzRGs appear extremely clumpyon HST images, strikingly similar tosimulations from hierarchical models offorming massive galaxies, e.g. brightestcluster galaxies (e.g. Fig. 2, Pentericciet al. 1998). The sizes, profiles and lu-minosities of the clumps are similar tothose of Lyman Break Galaxies (e.g.Steidel et al. 1996).

• Long-exposure optical spectra andobservations of their luminous dustemission show that HzRGs are under-going vigorous star formation (e.g. Deyet al. 1997).

Further evidence that HzRGs are incluster environments is the large meas-ured depolarization, indicative of a densehot gas (Carilli et al. 1997).

These properties, taken together,provided strong indirect evidence thatluminous distant radio sources pinpointforming massive galaxies at the centresof protoclusters.

2. Finding Distant Protoclusters:a Pilot Project

One of the most intriguing questionsin modern astrophysics concerns theformation of structure in the earlyUniverse (e.g. Bahcall et al. 1997).Although there are various scenariosfor the development of large-scalestructure, the epoch and mechanism of

the formation of galaxy clusters are stillopen questions. Using conventional op-tical and X-ray techniques, the detec-tion of clusters with z > 1 is difficult. Inthe optical, the density contrast is lowdue to the contaminating foreground,while in the X-ray the detection of ex-tended hot cluster gas is challengingbecause of cosmological surfacebrightness dimming. Because distantradio galaxies have properties expect-ed from forming massive galaxies inprotoclusters, we instigated a VLT pro-gramme to search for direct evidence ofgalaxy overdensities around them.

First, we conducted a pilot project tosearch for an excess of Lyα emittersaround the radio galaxy PKS 1138–262at z = 2.2. This radio galaxy was cho-sen because it is extremely clumpy, ithas the largest radio rotation measurein a sample of 70 HzRGs (Pentericci etal. 2000a) and has a large (120 kpc)Lyα halo (see Fig. 1). Furthermore, theredshifted Lyα of this radio galaxy fallsin one of the narrowband filters ofFORS. As the first visiting observersproject, narrowband and broadband im-ages were taken in March 1999 withFORS1 on ANTU. These data resultedin a list of ∼ 50 candidate Lyα emitters(Kurk et al. 2000). Subsequent multiob-ject spectroscopy in March and April2000 on FORS1 revealed 15 Lyα emit-ters at the same distance as the radiogalaxy and an additional 5 objects hav-ing continua also consistent with galax-ies at redshift 2.2. All the Lyα emitterswere found to have redshifts in therange 2.16 ± 0.02, centred around theredshift of the radio galaxy, and arewithin a projected physical distance of1.5 Mpc from it (Pentericci et al.2000b). We had now confirmed that inone case there is a substantial galaxyoverdensity associated with a distantluminous radio galaxy.

3. VLT Large Programme:Searching for Protoclustersto z > 4

The next step was to investigatewhether the overdensity around 1138–

Figure 1: Lyα image of radio galaxy PKS1138–262 at z = 2.16. The image is a14,400 s exposure in a narrowband, takenwith the FORS1 camera on ANTU. The sizeof the halo is roughly 120 kpc.

262 is typical for distant luminous radiogalaxies as a class. We therefore initi-ated a Large Programme on the VLT tosearch for Lyα emitting galaxies aroundluminous radio galaxies at 2 < z < 4.1.Our Large Programme was awarded 18nights, spread over three periods, onthe VLT/FORS2.

The targets for the Large Programmewere selected from a list of ∼150 radiogalaxies with z > 2. Suitable targets sat-isfied the following criteria: (i) large ra-dio luminosities (P178 > 1026 W Hz–1),(ii) luminous optical and infrared contin-ua and (iii) redshifts suitable for Lyα im-aging with the available VLT narrow-band filters. This resulted in a list of ob-jects with redshifts between 2.1 < z <3.2. We added one object with an evenhigher redshift (a radio galaxy at z =4.1), for which we purchased a cus-tom narrowband filter. This allowedus to cover a redshift range from 2 < z< 4.1.

The strategy for finding Lyα emittersnear the radio galaxy was the same asin our pilot project: each target was ob-served with a narrowband filter, whichcontained the redshifted Lyα line of theradio galaxy. The fields were also ob-served in broadbands, selected for hav-ing peak just redward of the Lyα line inorder to measure the UV continuum.Comparison of the narrowband imagewith the broadband image allowed us tofind objects with an excess flux in the

narrowband. These excess objectswere our candidate Lyα emitters. InFigure 3 an example is shown. On theleft a section of a 33,300 s narrowbandimage is plotted of the field around theradio galaxy TN J1338–1942 at z = 4.1.On the right the same field is shown butthis time observed in the broadband R.

The encircled objects can be easilyseen in the narrowband image, but theyalmost disappear in the broadband im-age. This means that more flux falls inthe narrowband than expected from thebroadband “continuum” image. Theseobjects were our primary candidates forspectroscopy.

37

Figure 2: Seven orbit HST image of PKS 1138–262 (Pentericci et al. 1998). The image istaken through the F606W filter and shows the UV continuum.

Figure 3: Narrowband and broadband image of a field near the radio galaxy TN J1338–1942 at z = 4.1. The narrowband filter used for theimage on the left is centred at λ = 6210 Å, which corresponds to Lyα emission at z = 4.1. The filter used on the right is the R filter, whichmeasures the UV continuum at z = 4.1. The encircled objects all have an excess flux in the narrowband, and were our candidates for follow-up spectroscopy.

Follow-up spectroscopy of the bright-est candidate emitters was carried outwith the multiobject spectroscopy modeof the FORS2 spectrograph. The aimwas to confirm whether the candidateswere cluster members and to measurecluster velocity distributions. Our selec-tion technique proved to be very suc-cessful: 70%–85% of all candidatesthat were observed spectroscopicallyshowed an emission line. Our successrate for the brightest candidate emitters(with a predicted line flux greater than1.5×10–17 erg s–1 cm–2) exceeded 90%.In Figure 4 an example is plotted for 20confirmed emitters at z = 4.1.

4. Preliminary Results

We have now carried out imagingand spectroscopy to a sufficient depthof five radio galaxy fields. A sixth target,TN J2009–3040 at z = 3.2, although ob-served for a comparable time, could notbe studied to the same intrinsic depthdue to the relatively high extinction.

The programme has been extremelysuccessful. All five well-observed radiogalaxies have > 20 spectroscopicallyconfirmed Lyα and/or Hα companions.In addition, despite its high extinction,TN J2009–3040 has 12 spectroscopi-cally confirmed companions. The den-sity of Lyα emitters found around all 6radio galaxies can be compared withthe density of emitters found in blankfields (see e.g. Cowie & Hu 1998,Rhoads et al. 2000). Taking into ac-count the difference in volume and indepth, the overdensity of Lyα emittersin radio galaxy fields is 5–15.Furthermore, the velocity dispersionsare found to be ∼ 300–1000 km s–1,concentrated around the redshift of the

radio galaxy. This is significantly small-er than the width of the narrowband fil-ters used in the initial imaging, provid-ing further evidence that the groupingsof galaxies are indeed associated withthe radio galaxies.

The spatial distribution of the Lyαemitters is in most cases homogeneousover the field, indicating that the struc-ture sizes generally exceed the 7 × 7arcmin2 FORS field, which correspondsto sizes greater than ∼ 3 × 3 Mpc2. Anexception is the distribution of Lyα emit-ters around TN J1338–1942 at z = 4.1,shown in Figure 5. The protocluster ap-pears to be bound in the northwest, but

the field of view is not large enough toshow a boundary in the south. Note thatin this field the radio galaxy is not in thecentre of the protocluster.

The mass can be estimated using thevolume occupied by the overdensity,the mean density of the Universe at theredshift of the protocluster, the meas-ured galaxy overdensity and the biasparameter (Steidel et al. 1998). Thecomputed masses are in the range1014–1016 M, comparable with themass of present-day rich clusters ofgalaxies. Taken together, the propertiesof the radio galaxy hosts and those ofthe newly-discovered structures dem-onstrate that luminous distant radiosources can be used to pinpoint proto-clusters, the ancestors of rich clusters.

5. Conclusions

We have discovered substantialgalaxy overdensities around radiogalaxies at z = 2.16, 2.86, 2.92, 3.13,3.15 and 4.10.

Although preliminary, we can drawthe following conclusions: (i) HzRGsare generally associated with the mostmassive forming galaxies at the centreof protoclusters and can be used to pin-point such structures. (ii) Because theradio lifetimes (few × 107 yr) are smallby the standard of cosmic evolution,there must be orders of magnitudemore “dead” radio sources than livingones. When the protoclusters haveevolved, the luminous radio sourcesare likely to be extinguished, explainingwhy z < 1 clusters do not generally har-bour active luminous radio sources. (iii)The luminosity functions and lifetimesof luminous distant radio sources areconsistent with such objects being as-sociated with every pronounced veloci-

38

Figure 4: Spectra of 20 Lyα emitting galaxies detected around TN J1338–1942 at z = 4.1. Allthe emission line galaxies show a line near 6200 Å. Analysis of the spectra confirmed that allthe lines are from Lyα at z = 4.100 ± 0.012 (Venemans et al. 2002).

Figure 5: Distributionof the confirmed Lyαemitters (circles) at z∼ 4.1 on the sky. Thesquare shows the po-sition of radio galaxyTN J1338–1942. Thesize of the circles isscaled according tothe Lyα flux of theobject, rangingbetween 0.3 and 4.1× 10–17 erg s–1 cm–2.The colour of thesymbols is an indica-tion of the redshift ofthe object. Blue ob-jects have a lowerredshift than the me-dian redshift of theemitters, red objectshave a higherredshift. The redshiftsrange between 4.088and 4.112. Note thatthe radio galaxy isnot centred in thegalaxy distribution.

ty spike in the space density of Lymanbreak galaxies.

Our distant protoclusters are uniquelaboratories for studying the most ex-treme overdense regions in the earlyUniverse and crucial locations for trac-ing the formation and evolution of clus-ters and galaxies. We are in theprocess of analyzing more than 100confirmed Lyα emitting galaxies that liewithin our protoclusters and in additionwe are obtaining observations withother facilities, including the AdvancedCamera for Surveys on the HST to findand study the various other populationsof galaxies expected within the proto-clusters. A detailed study of the mor-phologies and SEDs of protocluster

members between z ∼ 4 and z ∼ 1 willbe used to trace the history of galaxyassembly and star formation.

Acknowledgements

We thank the staff at the ParanalObservatory, Chile, for their splendidsupport. We also wish to thank GeroRupprecht at ESO for his guidance inthe process of purchasing the narrow-band filter.

References Bahcall, N.A., Fan, X., & Cen, R., 1997, ApJ

485, L53. Carilli, C.L., Röttgering, H.J.A., van Ojik, R.,

Miley, G.K., van Breugel, W.J.M., 1997,ApJS 109, 1.

De Breuck, C. et al., 2002, AJ 123, 637. Dey, A., van Breugel, W.J.M., Vacca, W. D.,

& Antonucci, R., 1997, ApJ 490, 698. Kurk, J. D. et al., 2000, A&A 358, L1.Miley,G.K. et al.,1992,The Messenger 68,12.van Ojik, R., Röttgering, H.J.A., Miley, G.K.,

Hunstead, R.W., 1997, A&A 317, 358. Pentericci, L. et al., 1998, ApJ 504, 139. Pentericci, L., Van Reeven, W., Carilli, C.L.,

Röttgering, H.J.A., & Miley, G.K., 2000a,A&AS 145, 121.

Pentericci, L. et al., 2000b, A&A 361, L25. Röttgering, H.J.A., Miley, G.K., van Ojik, R.,

1996, The Messenger 83, 26.Steidel, C.C., Giavalisco, M., Pettini, M.,

Dickinson, M., & Adelberger, K.L, 1996,ApJ 462, L17.

Steidel, C.C. et al., 1998, ApJ 492, 428. Venemans, B.P. et al., 2002, ApJ 569, L11.

39

TELESCOPES AND INSTRUMENTAT ION

ESO and NSF Sign Agreement on ALMAGreen Light for World’s Most Powerful Radio Observatory

(From ESO Press Release 04/03, 25 February 2003)

On February 25, 2003, the EuropeanSouthern Observatory (ESO) and theUS National Science Foundation (NSF)signed a historic agreement to con-struct and operate the world’s largestand most powerful radio telescope, op-erating at millimetre and sub-millimetrewavelengths. The Director General ofESO, Dr. Catherine Cesarsky, and theDirector of the NSF, Dr. Rita Colwell,acted for their respective organizations.

Known as the Atacama Large Milli-meter Array (ALMA), the future facilitywill encompass sixty-four interconnect-ed 12-metre antennae at a unique,high-altitude site at Chajnantor in theAtacama region of northern Chile.

ALMA is a joint project betweenEurope and North America. In Europe,

ESO is leading on behalf of its tenmember countries and Spain. In NorthAmerica, the NSF also acts for theNational Research Council of Canadaand executes the project through theNational Radio Astronomy Observatory(NRAO) operated by Associated Uni-versities, Inc. (AUI).

The conclusion of the ESO-NSFAgreement now gives the final greenlight for the ALMA project. The total costof approximately 650 million Euro isshared equally between the two part-ners.

Dr. Cesarsky is excited: “This agree-ment signifies the start of a great proj-ect of contemporary astronomy and as-trophysics. Representing Europe, andin collaboration with many laboratories

and institutes on this continent, we to-gether look forward towards wonderfulresearch projects. With ALMA we maylearn how the earliest galaxies in theUniverse really looked like, to mentionbut one of the many eagerly awaitedopportunities with this marvellous facil-ity.”

“With this agreement, we usher in anew age of research in astronomy”,says Dr. Colwell. “By working togetherin this truly global partnership, the in-ternational astronomy community willbe able to ensure the research capabil-ities needed to meet the long-term de-mands of our scientific enterprise, andthat we will be able to study and under-stand our universe in ways that havepreviously been beyond our vision”.

Artist’s view of the Atacama Large Millimeter Array (ALMA), with 64 12-m antennae.

The recent Presidential decree fromChile for AUI and the agreement signedin late 2002 between ESO and theGovernment of the Republic of Chilerecognize the interest that the ALMAProject has for Chile, as it will deepenand strengthen the co-operation in sci-entific and technological matters be-tween the parties.

A joint ALMA Board has been estab-lished which oversees the realization ofthe ALMA project via the managementstructure. This Board met for the firsttime on February 24–25, 2003, at NSFin Washington and witnessed this his-toric event.

The Atacama Large Millimeter Array(ALMA) will be one of astronomy’s mostpowerful telescopes – providing unprec-edented imaging capabilities and sensi-tivity in the corresponding wavelengthrange, many orders of magnitudegreater than anything of its kind today.

ALMA will be an array of 64 antennasthat will work together as one telescopeto study millimetre and sub-millimetre

wavelength radiation from space. Thisradiation crosses the critical boundarybetween infrared and microwave radia-tion and holds the key to understandingsuch processes as planet and star for-mation, the formation of early galaxiesand galaxy clusters, and the formationof organic and other molecules inspace.

“ALMA will be one of astronomy’spremier tools for studying the universe”,says Nobel Laureate Riccardo Giac-coni, President of AUI (and former ESODirector General (1993–1999)). “Theentire astronomical community is anx-ious to have the unprecedented powerand resolution that ALMA will provide.”

The President of the ESO Council,Professor Piet van der Kruit, agrees:“ALMA heralds a break-through in sub-millimetre and millimetre astronomy, al-lowing some of the most penetratingstudies of the Universe ever made. It issafe to predict that there will be excitingscientific surprises when ALMA entersinto operation.”

Timeline for ALMA

• June 1998: Phase 1 (Research andDevelopment)

• June 1999: European/AmericanMemorandum of Understanding

• February 2003: Signature of thebilateral Agreement

• 2004: Tests of the Prototype System • 2007: Initial scientific operation of a

partially completed array • 2011: End of construction of the

array

Further Reading About ALMA

More information on the ALMA proj-ect can be found in earlier issues of TheMessenger (March 1996; March 1998;December 1998; June 1999; andMarch 2002), and on the following web-sites:

http://www.eso.org/projects/alma/http://www.alma.nrao.edu

40

Following several weeks of around-the-clock work, a team of astronomers and en-gineers from Germany, the Netherlands,France and ESO has successfully per-formed the first observations with the MID-Infrared interferometric instrument (MIDI), anew, extremely powerful instrument just in-stalled in the underground laboratory of theVLT Interferometer (VLTI) at the Paranal Ob-servatory (Chile).

In the early morning of December 15,2002, two of the 8.2-m VLT unit telescopes(ANTU and MELIPAL) were pointed towardsthe southern star Epsilon Carinae and thetwo light beams were directed via the com-plex intervening optics system towardsMIDI. After a few hours of tuning and opti-mization, strong and stable interferometricfringes were obtained, indicating that allVLTI components – from telescopes to thenew instrument – were working together per-fectly. Two more stars were observed beforesunrise, further proving the stability of theentire system.

The upper image shows the “first fringes” ofthe star Epsilon Carinae, as obtained at VLTIwith the new MIDI instrument at the mid-in-frared wavelength of 8.7 µm. The photograph at the bottom shows thegroup responsible for the MIDI installationand first tests, taken inside the VLT ControlBuilding, right after the successful “FirstFringes” in the early morning of December15. From left to right: Front row (sitting/kneel-ing) Julio Navarrete, Lorena Faundez, Mar-kus Schoeller, Andrea Richichi – Back row(standing) Francesco Paresce, Andres Pino,Nico Housen, Uwe Graser, Olivier Chesneau,Christoph Leinert, Andreas Glindemann,Walter Jaffe, Sebastien Morel, RichardMathar, Pierre Kervella, Eric Bakker.

More details can be found in ESO pressrelease PR 25/02 of 18 December 2002.

New Vistas Open with MIDI at the VLT Interferometer“FIRST FRINGES” IN MID-INFRARED SPECTRAL REGION WITH TWO VLT TELESCOPES

41

TIMMI2 at the 3.6-m, or The Return of MIRV. DOUBLIER, La Silla

In October 2002, after only 12 weeksof forced retirement, TIMMI2 (ThermalInfrared Multimode Instrument) under-went a major upgrade and was suc-cessfully remounted at the Cassegrainfocus of the 3.6-m telescope at La Silla.

TIMMI2 had a difficult start to 2000,when the tragic and untimely death ofPI Hans-Georg Reimann overshad-owed the success of the project. The in-strument was first commissioned in late2000, with TNT control software andthe Wallace acquisition system. How-ever, after almost two years of continu-ous operation, it became apparent thatan upgrade of the control and acquisi-tion system could improve the reliabilityof the instrument and reduce the oper-ational load for the staff.

In February 2002, TIMMI2 sufferedfrom a severe problem which tooksome time to locate inside the detectormount: the thermal contact of the arraywith the cooling support had deteriorat-ed. Operation was resumed after threemonths of investigation and tests, andfinally, in August 2002, the upgradeproject was started at La Silla (M.Sterzik, Project Scientist; J. Valenzue-la, IRACE-SW; P. Le Saux. Templates;J.-C. Guzman, ICS, OS; and U. Weilen-mann, Control System and ProjectLeader) with active help from the IRgroup in Garching (H.-U. Käufl. R. Sie-benmorgen, J. Stegmeier and H. Mehr-gan). Thanks to the efforts of this team

and many others, TIMMI2 is back on-sky and available to the MIR communi-ty.

Not only was the 2nd commissioningused to refurbish TIMMI2 with up-to-date electronics and software, we alsotook the opportunity to simplify obser-vational procedures. Experienced MIRobservers will not have a problem ad-justing to the new procedures, andnewcomers to the MIR community willfind the instrument user-friendly. TheTIMMI2 system can now be comparedto those of ISAAC and NACO for 2.5–5micron observations. For example, theexposure time set-up requires only oneentry: the amount of time the user de-sires to spend on source. Optimum val-ues for each set-up are defined in aconfiguration file – a concept devel-oped in the first phase of TIMMI2 oper-ations – which is read at set-up time bythe Observing Software (OS). There isno more interaction between the ob-server and the individual parts of TIM-MI2, and the data of TIMMI2 are nowroutinely injected into the ESO archive.

How Was This Possible?

• Above all: with great team spirit.• The old TIMMI2 acquisition system

(Wallace electronics), which handlesacquisition tasks as well as clockingand biasing of the array, was replacedby the ESO IRACE system. The versa-tility of the IRACE system allows easydevelopment of new timing patternsand will further increase the perform-ance of the Raytheon array.

• The instrument control system hasbeen refurbished with VLT standardmodules and controllers for the steppermotors, enhancing functionality and in-creasing reliability at the same time.

• Temperature and vacuum controlhas been modified, allowing importantsystem parameters to be monitoredcontinuously and the handling ofalarms.

• The system is entirely controlled byVLT SW, and communications are han-dled via the licence-free CCS En-vironment rather than RTAP. The han-dling of the observations is managed by“boss” which is, at high level, interfac-ing BOB, ICS, DCS, TCS and the OS.A series of configuration files “preset”the instrument and detector dependingon the instrument mode required by theuser.

• Data handling is now under the con-trol of the DMD SW. The raw data aretransferred via the data subscriber tothe DHS WS which then redirects thedata to an offline machine for the con-venience of the user.

• The pipeline software has beenmodified to accept the new format ofthe raw data stored in cubes. It is avail-able upon request. During the twocommissionings of November andDecember 2002, the transfer of datawas done directly from the instrumentWS within the templates; however thisdramatically increased the overheads.To keep overheads low, the transfer ofdata from the DHS WS is currently han-dled by a shell script from the pipelinemachine.

• Observations preparation: “p2pp” ison the job and proved to be well ac-cepted by the TIMMI2 community.

What did we achieve?

(1) Imaging: TIMMI2 shows an im-provement of 40% in sensitivity in allbands, allowing us to finally offer L- andM-band observations.

A “Raster” mode (or jitter) is being im-plemented and tested for extendedsources.

Figure 1: New Control rack: tiiracq LCU ontop, the Sparc and IRACE backend in themiddle and the temperature and vacuumcontrol at the bottom.

Figure 2: HD32887 8–13µ spectrum: N bandsensitivity plot shows a factor 2 better thanwith the Wallace acquisition system at 10microns.

Figure 3: Eta-Carinae three-colour image:NB10.4 (Blue) NeII (Green) Q1 band (Red).

IRACE gives more freedom in pro-gramming read-out sequences. TheTIMMI2 detector (Raytheon CRC774)has a rather complex architecture, andespecially “long” detector integrationtimes (exceeding 100 ms; this is longfor the mid-infrared!) could only bedone in very compromised form withthe old system.

(2) Spectroscopy: The sensitivity im-proved by a factor of 2: 50mJy/10sig-ma/1hr at 10 micron.

(3) Polarimetry: This has been exten-sively tested. A new Instrument Pack-age is available on the web and is al-ready installed at the telescope.

Developments

In April, TIMMI2 will undergo regularmaintenance. We plan to reinstall thescience grade array, together with newcryo cabling and detector board.

The Lunar Occultation mode will be

implemented soon. It is not yet opera-tional at a template level: we hope to of-fer it in visitor mode during Period 71,and in full operation mode for Period 72.

The possibility to chop at different PAangles is under investigation. This willbe a major improvement for extendedsource observations.

The new, more compact control rackleaves more space in the Cassegraincage. The implementation of a rotatorfunction is under study.

42

OTHER ASTRONOMICAL NEWS

NEWS FROM SANTIAGO

International Workshop on

STRUCTURE EVOLUTION AND COSMOLOGY: NEWSYNERGY BETWEEN GROUND-BASED OBSERVATIONS,SPACE OBSERVATIONS AND THEORYD. ALLOIN (ESO/Chile), M. PIERRE (CEA/SAp/France) and A. REFREGIER (CEA/SAp/France)

An international Workshop, “Struc-ture Evolution and Cosmology: NewSynergy between Ground-basedObservations, Space Observations andTheory”, was organized jointly by theEuropean Southern Observatory (ESO/Chile), the Centre National d’EtudesSpatiales (CNES/France) and theCommissariat pour l’Energie (CEA/France), at ESO/Santiago from Octo-ber 28 to 31, 2002.

The ESO facilities in Santiago werehosting such a large meeting for thefirst time: according to the attendeesit was a very successful trial and,on our side, we were very happy toshare with them our renewed environ-ment, as well as the coffee-breaks onthe lawn!

A total of 144 astronomers and stu-dents attended the Workshop, whichwas opened by the ESO DirectorGeneral, Catherine Cesarsky. After theusual welcome words given in thename of the Scientific OrganizingCommittee (Danielle Alloin, MargueritePierre and Alex Refregier), we started ahalf-day tutorial about Cosmology. Thiswas an excellent way to refresh andupdate our ideas in the field and werecommend the superb tutorial presen-tations to be found on the Workshopwebpage (see below)!

The unfortunate overbooking prob-lems of some flights prevented a few ofour reviewers to arrive on time for theirpresentation, or arrive at all! Thanks tothe friendly attitude of all speakers, wedid some re-shuffling of the programmeand, among others, exchanged weak-lensing against strong-lensing (stilllensing anyway!). This gave a touch of“impromptu” to the Workshop, creatingthe nice feeling that we were all in-volved in the Workshop organization...

A rather complete panel of topics re-lated to Cosmology were covered: fromthe description of the early universe byparticle physicists, to the birth and evo-lution of large-scale structures (simula-tions and observations), to various as-pects of galaxy clusters, to the numer-ous lensing effects one encounterswhen probing the high-z universe...Vivid and interesting discussions fol-lowed the presentations, continued atthe posters display and at the coffee-breaks/lunches. The need for a goodinterplay between space-based andground-based observations (to accom-plish a multi-wavelength coverage), butalso with simulations and theoreticalapproaches, was highlighted through-out the Workshop. And this is indeedwhat is giving rise to the great achieve-ments of modern astrophysics!

The dense programme of theWorkshop was interrupted for half aday to visit La Sebastiana, the house ofPablo Neruda in Valparaíso: an attemptto join the dreams of the poet with ourmore arid research activities?

“La luz bajo los árboles,la luz del alto cielo.La luz verde enramada que fulgura en la hojay cae como fresca arena blanca.Una cigarra elevasu son de aserraderosobre la transparencia.Es una copa llena de agua,el mundo.”

Then, walking down through Valpa-raíso’s stairs and paths, we reachedthe restaurant at the top of a tower inthe port, where the Workshop dinnerwas held, and from there we could en-joy the magnificent view to the bay andthe hills of Valparaíso, all twinkling inthe night. On our way back to Santiago,the bus was stopped for a while be-cause of an excess of weight (!) andduring this unexpected pause, some ofour skilled theoreticians were desper-ately trying to see the MagellanicClouds, in spite of an intense light back-ground. The happy ones who went to

43

the Paranal visit after the Workshop,did see them, however!

The pleasant environment of theESO facilities in Santiago, the tutorial,particularly appreciated by studentsand non-cosmologists and the great re-views/contributions/posters all contrib-

uted to making this Workshop a veryexciting one!

The group of reviewers and the SOCdecided that there is no need to printanother book on Cosmology and thatthe interest of the meeting had beenlargely in discussing and meeting with

colleagues and young researchers . . .Yet, most of the reviews and presenta-tions are available on the Workshopwebpage. To retrieve the information,please go to:

http://www.eso.org/gen-fac/meet-ings/cosmology2002/programme.html

International Workshop on

STELLAR CANDLES FOR THE EXTRAGALACTIC DISTANCE SCALED. ALLOIN (ESO/Chile) and W. GIEREN (Concepción University, Chile)

An international Workshop, “STEL-LAR CANDLES for the EXTRAGALAC-TIC DISTANCE SCALE”, was organ-ized jointly by the European SouthernObservatory and the Astronomy Groupat the University of Concepción as amember of the FONDAP/Conicyt Cen-tre for Astrophysics launched in Chile in2002. The Workshop received supportgrants from ESO, FONDAP/Conicyt,the University of Concepción and Fun-dación Andes.

The Workshop was hosted on thebeautifully green campus of the Uni-versity of Concepción, on December9–11, 2002, under blue skies and in apleasant atmosphere. More than 73 as-tronomers had registered and attendedthe Workshop. We are happy to ac-knowledge a good participation of stu-dents from various Chilean universities.After a formal opening of the Workshopby the Rector of the University, DonSergio Lavanchy Merino, and the Deanof the Physics and MathematicsFaculty (where the Workshop was tak-ing place), Dr. José Sanchez, a fewwords of welcome were given by Dr.Wolfgang Gieren, Head of the Astron-omy Group in Concepción.

A fairly large number of possible stel-lar “candles” used for the determinationof the extragalactic distance scale,were examined and discussed duringthe Workshop. These encompassed:Cepheids, RR Lyrae stars, blue super-giants, the red giant clump, supernovaeof types Ia and II, novae, planetary neb-ulae, eclipsing binaries, and globularclusters. Both theoretical and observa-tional reviews were given by top spe-cialists on the various candles, andsome recent advances were presented.Pros and cons were discussed for eachcandle, giving rise to lively, sometimeshot, debates.

The comparison of distances derivedfor a same galaxy using different indi-cators was extremely interesting. Ob-viously, the Large Magellanic Cloud isthe galaxy which has been studied

through the greatest variety of distanceindicators. In spite of fairly large-sizevariations, still to be understood, thevalue for the distance modulus of theLMC seems to be slowly converging: asquoted in the concluding remarks, 60values were mentioned during theWorkshop, with a mean close to 18.5mag and uncertainty of ±0.1 mag. Thequestion was raised about how accu-rately do we actually need to measurethis number, for the various astrophysi-cal applications?

Using the same distance indicator,the Cepheids, does not always lead tosimilar results, even if one uses thesame data set! Very puzzling was thecomparison of distances derived fromthe same large HST data set onCepheids in some 25 nearby galaxies.Two independent studies have beenconcluded on two different values of theHubble constant, with a net differenceof 12 km/s/Mpc, three times larger thanthe quoted total error-bars (±2 km/s/Mpc)on each value. This opened a vivid andenlightening discussion about unknownsystematics: it was wisely suggestedthat we use a current “best value” of 66±6 km/s/Mpc for the Hubble constant,until the source of the discrepancy hasbeen understood.

On the side of theory, progress instellar atmosphere modelling of Ce-pheids opens the door to direct dis-tance determinations of improved accu-racy. Coupled to interferometric meas-urements, this is becoming a superband promising technique to directly de-rive a very accurate distance toCepheids, to distances further out thanthose in reach by parallax measure-ments.

More and more refined analyses ofother indicators are currently being de-veloped, either from a modelling or anobservational point of view. Progress isbeing made on RR Lyrae star observa-tions and modelling, especially on thenear-infrared period–absolute magni-tude–metallicity relation which is a par-

ticularly useful tool in old populationsystems, on the intermediate-agedclump stars for which the most recentK-band studies in the MagellanicClouds and other Local Group galaxiesseem to indicate that very accurate dis-tance results can be expected from thismethod in the near-infrared domain,and on blue supergiant stars which of-fer the advantage of being intrinsicallyvery luminous candles which can beused in galaxies as distant as 10–15Mpc but are appropriate for young pop-ulation systems only. The wind momen-tum–luminosity relation and the flux-weighted gravity–luminosity relation forthese stars are very promising tools forspectroscopic distance determinationsof high accuracy from these objects.The use of spectroscopic information,which gives metallicity and reddeninginformation on the studied objects “forfree”, is indeed an important plus.

The use of supernovae for distancedeterminations was extensively dis-cussed since these constitute privi-leged – and often unique – indicators athigh redshift. SNIa remain the best in-dicators to be used so far in verydistant galaxies, although not beingthe perfect candle – since there areslowly declining and fast decliningSNIa. It has been recommended at theWorkshop to study them in the H bandwhere the slope of the luminosity ver-sus decline-rate relation seems to beessentially flat and where reddeningcorrections are smaller than in opticalbands. The following comment wasalso made: “Despite the fact that wedon’t understand them, they are excel-lent stellar candles”... In turn, SNII arenot intrinsically as bright as SNIa andare subject to more systematics, hencecurrently less useful as candles thanSNIa. An exhaustive review on thephysics of SNIa explosions was givenfrom which the following quotation isextracted: “the red model here is thetruth, the blue lines are the observa-tions”.

Maria-Rosa Cioni

I am probablyan astronomersince the au-tumn of 1990when I started atthe University inBologna, thoughthe bed sheetsof my earliestdays were full of

planets and stars. After graduating Imoved to the Leiden Observatory in theNetherlands where I obtained the Ph.D.degree. This first highly positive experi-ence abroad exposed me to an interna-tional and active scientific environment

44

Fellows at ESO

Aurore BacmannI have been a

Fellow at ESOGarching sinceMarch 2002. Be-fore that, I haddone my PhDthesis with Phi-lippe André atCEA-Saclay nearParis, France, andspent two years

as a post-doc at the University of Jena,Germany, working in the group of Tho-mas Henning. My area of research isstar formation, mostly the early pre-stellar stages, before stars are formedwithin dense cores. This stage is par-ticularly important since it representsthe initial conditions of gravitational col-lapse and star formation. During myPhD, I used the instrument ISOCAMaboard the ISO satellite to determinethe density structure of pre-stellar cores.

After my PhD, I started studying thechemistry of these cores, chiefly molec-ular depletion and deuteration, with col-laborators from the Bordeaux and theGrenoble Observatory in France. Thishas been my main subject of researchhere at ESO. To carry out these proj-ects, I use mostly (sub)millimetre tele-scopes. Additionally I work on the struc-ture of circumstellar matter around

Herbig Ae/Be stars, using polarimetry,and I am also interested in their chem-istry.

Since I arrived at ESO, I have beeninvolved in the development of the in-terferometer ALMA, working withStéphane Guilloteau (IRAM/ESO). Themain goal of this task is to look into thebandbass calibration of the system anddetermine the frequency response ofthe instrument. It is extremely motivat-ing to be taking an active part in such amajor and ambitious project, all themore that ALMA will be very relevant tothe research I am doing.

Since the beginning ESO has provided opportunities for young scientists to interact with the environment ofan observatory. Many European astronomers spent some years as post-doctoral Fellows at ESO. TheFellowship programme has been very successful; with only very few exceptions, all former ESO Fellows arenow working as astronomers in the community.

In addition to developing their scientific careers the Fellows are also asked to contribute to the work of theobservatories. In Chile all Fellows are involved in operational activities at Paranal and La Silla, while in Garchingthey participate in instrument and software development, PR activities, ALMA related studies and surveys.

With this issue of The Messenger we start short presentations of Fellows currently at ESO. They describe intheir own words what research they pursue and how they are involved in ESO activities. We will continue topresent some of the young faces at ESO in the coming issues. B. Leibundgut

Novae were also explored for theircurrent usefulness as distance indica-tors, although they seem to be ham-pered by larger systematics than otherstellar candles. Planetary nebulae, pre-sented as the “Swiss-army knives” ofextragalactic astronomy, are also usedto derive distances to galaxies, throughtheir luminosity function (measured inthe [OIII] emission line). The techniqueseems to work, in spite of a num-ber of complicating factors, which ledto the following remark: “An astrophysi-cist is someone who sees somethingworking in practice and wonderswhether it works on the basis of firstprinciples”.

The luminosity function of globularclusters is another attempt at enrichingour tool-box of distance indicators,which was discussed at the Work-shop. And, finally, the use of eclipsingbinaries as distance indicators looksvery promising if more appropriate sys-tems in Local Group galaxies can befound.

Altogether, the Workshop allowed usto discuss and compare in depths thedifferent stellar indicators used so far inthe difficult quest of astronomers fordistances. The very pleasant environ-ment of the campus, the excellence ofthe talks, the challenges of the fieldwere key-elements for stimulating dis-

cussions and exchanges. Remainingcontroversies were “discussed” duringthe soccer game which closed theWorkshop, although some attendees(those for which all questions had beenanswered) had left already! The gamewas won by the population II teamagainst the population I team, with thehelp of a Dutch referee who is known tobe a friend of old stellar populations...

The reviews of this Workshop will bepublished early 2003 in a volume of theSpringer series Lecture Notes inPhysics, while all contributions – talksand posters – will be made available onthe Workshop webpage in Concepción(http://cluster.cfm.udec.cl).

that could find no better match after-wards than by coming to ESO.

I study Asymptotic Giant Branchstars, their evolution and variabilityproperties in resolved galaxies: theMagellanic Clouds and other galaxiesin the Local Group. I am actively col-laborating with my former supervisorProf. Harm Habing and our most recentpaper shows the metallicity gradient inthe Magellanic Clouds from the ratio ofcarbon- and oxygen-rich stars. Mywork, so far predominantly photometric,is evolving to spectroscopy using theFLAMES instrument. At the same timeI follow-up from the La Palma observa-tory AGBs in northern Galaxies.

I am presently in the middle of myESO fellowship and I feel proud of be-ing among the people that work at andfor the observatory and provide the re-sources that improve our knowledge ofwhat is above us. I have found both inGarching and Chile, where I enjoyspending part of my time, a very friend-ly group of colleagues and new collab-orators. I learned how to support theastronomical activities in Paranal. Inparticular using the UVES instrumentand FLAMES in the near future.

During my free time, though there isalways more work than time, I enjoy therich social life in Munich, practise vari-ous sports and am (now) learningGerman.

45

Ivo SavianeI arrived at

ESO in April2001, from theUCLA. I waspreviously apostdoc in Pa-dova, where Ialso receivedmy PhD in1997. Duringthat time I thinkI gave, with thePadova and

IAC groups, an important contributionon the question of the relative ages ofGalactic globular clusters (GC). Dwarfspheroidal galaxies also attracted myattention, and with another Padovagroup we helped establishing the ideaof extended SF histories (and discov-ered the old population of Leo I). I wasalso one of the creators of the VirtualPlanetarium educational website, at thePadova Observatory.

I am enjoying very much the La Sillaenvironment, which offers the possibili-ty to interact with a multidisciplinarygroup of people, and to contribute to in-strument development (in particular,upgrading FEROS in the near future).Moreover, the ever-increasing numberof students and visitors makes ESO/Chile a good and stimulating workingplace.

Now I am leading a project to test theluminosity-metallicity relation of dwarf

irregular galaxies, I am extending therelative age study to the LMC clusters(where I discovered a young globular),and the dwarf galaxy group in Padovastill likes to have me as a collaborator!During my stay in California I discoverdthat the Antennae are not so far ascommonly believed (and now I have toconvince the referee), and I contributedto the project “Hubble Deep Field in aGC”, led by the Vancouver group.Some people think I am good at free-hand drawing, and a few of my portraitsare out there on the Internet. I would bevery happy if you gave me Old BlueEyes’ The complete Reprise years!

Petri VaisanenI am a second-

year ESO Fellowin Chile – and I donot regret accept-ing this job. Tosupport and usetop-notch instru-ments at the VLT,to learn more abouta wide range ofobservational as-tronomy, to help

visiting astronomers doing exciting sci-ence, is all very rewarding. And I stillhave plenty of time for my own work. Infact, operating the adaptive optics in-strument NACO has given new per-spectives to my interests. For the third

year of my fellowship I will join theAstronomy Department of Universidadde Chile. I can concentrate on scienceand develop new collaborations beforestarting a job-hunt again.

My main scientific focus has been ex-tragalactic infrared work aiming at ac-quiring an unbiased view of the forma-tion history of galaxies. It has taken theform of several different projects, in-cluding optical and near-IR follow-up ofISO-detected mid- and far-IR galaxiesusing various telescopes. I have con-centrated on extremely red galaxies(EROs) and interacting and starburstsystems, and recently also on obscurednuclear activity. The plan is to expandthis line of research using e.g. theSIRTF. I am also involved in a “more lo-cal” project of studying star formation ingalactic molecular clouds using VLT/ISAAC data.

A thing I have missed is teaching,which I had done previously in Helsinki(where I finished my PhD in 2001) andHarvard (where I worked as a SAO Pre-doctoral Fellow for 3 years). However,as part of a campaign to see my owncountry join ESO, I have written, withothers, articles about ESO and astron-omy to Finnish newspapers and maga-zines, given interviews, and hostedjournalists visiting Paranal. Exploringthe universe can be great fun – that issomething I have thought since a littlekid, and it is the idea I hope to getacross whether talking to students orthe general public.

URANUS, Rings and MoonsA near-infrared view of the giant planet Ura-nus with its rings and some of its moons, ob-tained at a wavelength of 2.2 µm on Novem-ber 19, 2002, with the ISAAC multi-mode in-strument on the 8.2-m VLT ANTU telescope.The observing conditions were excellent,with seeing of 0.5 arcsec.The rings of Uranus are almost undetectablefrom the Earth in visible light, but on this VLTnear-infrared picture the contrast betweenthe rings and the planet is strongly en-hanced. At the near infrared wavelength atwhich this observation was made, the in-falling sunlight is almost completely ab-sorbed by gaseous methane present in theplanetary atmosphere and the disk ofUranus therefore appears unusually dark. Atthe same time, the icy material in the ringsreflects the sunlight and appears compara-tively bright.The observers at ISAAC were EmmanuelLellouch and Thérèse Encrenaz of theObservatoire de Paris (France) and Jean-Gabriel Cuby and Andreas Jaunsen (bothESO-Chile).

More details can be found in ESO pressrelease PR 31/02.

46

ANNOUNCEMENTS

ESO Workshop onLarge Programmes and Public Surveys

Garching, Germany, 19–21 May 2003

An evaluation of the scientific success of the first completed Large Programmes at the VLT and public surveys should take place be-fore the survey telescopes VST and VISTA start operating. To assess the scientific return of the Large Programmes ESO is organizing athree-day workshop in Garching.

Background

In 1996, on request of the OPC, a working group was formed to discuss the future of ‘Key Programmes’ in the era of the VLT. Thedescription of Large Programmes in the Call for Proposals reflects the recommendations by the working group.

The working group further suggested that “at about the time of the start of operations of UT3 the definition of the Large Programmesand its implementations should be reconsidered”.

VST operations are planned to start in 2004 and a major fraction of the time at this telescope, and later on VISTA, shall be dedicatedto surveys. The workshop will address the issue on how best to deal with surveys for these telescopes.

Reason for the Workshop

ESO feels it is time to evaluate the scientific success of the first completed Large Programmes at the VLT. Since surveys in effect areLarge Programmes (and many of them have been implemented as such at ESO) the relation of surveys to other forms of observationalprogrammes should be discussed as well. This should provide an overview of the scientific return and a chance to re-assess the validityof the Large Programmes.

All Large Programmes that are either completed or near completion shall present their scientific results and the impact the project hadin its field.

With the VST and VISTA on the horizon a discussion of Public Surveys at ESO is indicated as well. A special session will be devotedto discussing the handling of surveys at ESO in the future.

Scientific Organizing Committee

John Black, Gothenburg Simon Lilly, ZurichCatherine Cesarsky, ESO Jean-Loup Puget, ParisGerry Gilmore, Cambridge Alvio Renzini, ESOJoachim Krautter, Heidelberg Christoffel Waelkens, LeuvenAndy Lawrence, Edinburgh Stefan Wagner, HeidelbergBruno Leibundgut, ESO

Format

The workshop will start with a presentation of every Large Programme up to Period 69. A discussion session to assess the effective-ness of the Large Programme concept is planned to follow. The extant surveys at ESO will be presented and schemes for efficient useof the survey telescopes evaluated in a separate discussion session.

Registration

Participants are asked to register for the workshop using the form available at http://www.eso.org/surveys03 Deadline for the registration is 1 April, 2003. Please note that only a limited amount of participants can be accepted due to space limitations.

ESO Workshop onHigh-Resolution Infrared Spectroscopy in Astronomy

ESO-Headquarters, Garching near Munich, November 18-21, 2003

Infrared spectroscopy at a resolution of a few km/s offers a unique tool to study rotational-vibrational transitions of many abundant mol-ecules as well as important atomic lines in a multitude of interesting astrophysical environments. Applications include the possible directdetection of exosolar planets, measurements of the abundances and magnetic fields of stars, studies of ISM chemistry and the kinemat-ics of stars and gas in galactic centres.

In this workshop the latest status of the high-resolution infrared spectroscopic capabilities of the VLT (CRIRES, VISIR) and other ob-servatories shall be presented.

For space reasons the number of participants is limited to 110.More details, preliminary programme and registration form can be retrieved from

http://www.eso.org/gen-fac/meetings/ekstasy2003or contactekstasy2003@eso.org for further information

47

FIRST ANNOUNCEMENT

ESO Workshop onScience with Adaptive Optics

ESO Headquarters, Garching near Munich, September 16–19, 2003

Over the past ten years, the concept of adaptive optics has matured from early experimental stages to a standard observing tool nowavailable at many large optical and near-infrared telescope facilities. Indeed, adaptive optics has become an integral part of all presentand future large telescope initiatives, and will be essential in exploiting the full potential of the large optical interferometers currently un-der construction. Adaptive optics has been identified as one of the key technologies for astronomy in the 21st century.

Adaptive optics has already delivered exciting results covering areas from solar-system astronomy (both the sun and the planetarysystem) over the star-forming regions in the solar neighbourhood to Local Group galaxies and objects at cosmological distances.

Recent highlights include:

• Evolution of small-scale structures on the solar surface• Discovery of binary asteroids and asteroids moons• High-resolution studies of circumstellar disks around young stars• Precise mass determination of the black hole in the Galactic Centre• Spatially resolved studies of extragalactic stellar populations

The present meeting intends to bring together users of adaptive optics from all fields of astronomy to discuss the latest scientific re-sults obtained with diverse adaptive optics systems and to exchange ideas on how to reduce and analyse such observations.

This ESO workshop aims also at educating the general astronomical community in Europe on the unique science potential of adap-tive optics for all branches of astronomy. We want to bring together researchers working in many different areas of astronomy in order toprovide a comprehensive picture of the utilization of adaptive optics in astronomy. Synergy effects are expected from the comparison ofdifferent observing and data analysis strategies.

Co-chairs: Wolfgang Brandner (MPIA), Markus Kasper (ESO)

Scientific Organizing Committee:

Danielle Alloin (ESO), Laird Close (Steward Obs., Tucson, USA), Tim Davidge (Herzberg Inst., Victoria, Canada), Reinhard Genzel(MPE, Germany), Thomas Henning (MPIA, Germany), Christoph Keller (NSO Tucson, USA), Anne-Marie Lagrange (LAOG, France),Simon Morris (Durham, UK), Francois Rigaut (Gemini, USA), Daniel Rouan (Obs. de Paris, France), Hans Zinnecker (AIP, Germany)

For more details and registration, see http://www.eso.org/aoscience03

Symposium Secretary: Christina Stoffer

European Southern ObservatoryKarl-Schwarzschild-Str. 2, D-85748 Garching bei München, GermanyFAX: +49 89 3200 6480email: aoscience03@eso.org

PERSONNEL MOVEMENTSInternational Staff (1 January – 28 February 2003)

ARRIVALS

EUROPE

ALBERTSEN, Maja (DK), StudentANDOLFATO, Luigi (I), VLTI Software EngineerBRAUD, Jérémy (F), AssociateFRAHM, Robert (D), VLTI Software EngineerKASPER, Markus (D), Adaptive Optics Instrument ScientistLIMA, Jorge (P), AssociateNEVES, Antonio (P/GB), European Project Manager ALMANORMAN, Colin (AUS), AssociatePASQUATO, Moreno (I), ALMA Software IntegratorPERCHERON, Isabelle (F), Astronomical Data Quality Control

ScientistPOPOVIC, Dan (AUS), Software EngineerRABIEN, Sebastian (D), Associate

CHILE

CASQUILHO FARIA, Daniel (S), StudentGAVIGNAUD, Isabelle (F), StudentSBORDONE, Luca (I), Student

DEPARTURES

EUROPE

DEMARCO, Ricardo (RCH), StudentFRIEDRICH, Yawo (D), AssociateGORSKI, Krzysztof (PL), Archive AstronomerPIERFEDERICI, Francesco (I), Paid AssociateTEIXERA, Rui (P), TraineeVAN BOEKEL, Roy (NL), Student

CHILE

SCHÜTZ, Oliver (D), StudentSLUSE, Dominique (B), Student

48

ESO Studentship Programme The European Southern Observatory research student programme aims at providing opportunities and facilities to enhance the Ph.D.

programmes of ESO member-state universities. Its goal is to bring young scientists into close contact with the instruments, activities, andpeople at one of the world’s foremost observatories. For more information about ESO’s astronomical research activities please consulthttp://www.eso.org/science/

Students in the programme work on their doctoral project under the formal supervision of their home university. They come to eitherGarching or Santiago for a stay of normally one or two years to conduct part of their studies under the co-supervision of an ESO staff as-tronomer. Candidates and their home-institute supervisors should agree on a research project together with the ESO local supervisor. Alist of potential ESO supervisors and their research interest can be found at http://www.eso.org/science/sci-pers.html#faculty andhttp://www.sc.eso.org/santiago/science/person.html . A list of current PhD projects offered by ESO staff is available athttp://www.eso.org/science/thesis-topics/. It is highly recommended that the applicants start their Ph.D. studies at their home institute be-fore continuing their Ph.D. work and developing observational expertise at ESO.

In addition, the students in Chile have the opportunity to volunteer for a small amount of duties (~40 nights per year) at the La SillaObservatory. These duties are decided on a trimester by trimester basis, aiming at giving the student insight into the observatory opera-tions and shall not interfere with the research project of the student in Santiago.

The ESO studentship programme is shared between the ESO Headquarters in Garching (Germany) and the Vitacura offices inSantiago (Chile). These positions are open to students enrolled in a Ph.D. programme in the ESO member states and, exceptionally, ata university outside the ESO member states.

Candidates will be notified of the results of the selection process in July 2003. Studentships typically begin between August andDecember of the year in which they are awarded. In well justified cases starting dates in the year following the application can be nego-tiated.

The closing date for applications is June 15, 2003.

Please apply by: filling the form available at http://www.eso.org/gen-fac/adm/pers/forms/studentappliform.htmland attaching to your application:

– a Curriculum Vitae (incl. a list of publications, if any), with a copy of the transcript of university certificate(s)/diploma(s).– a summary of the master thesis project (if applicable) and ongoing projects indicating the title and the supervisor (maximum

half a page), as well as an outline of the Ph.D. project highlighting the advantages of coming to ESO (recommended 1 page,max. 2).

– two letters of reference, one from the homeinstitute supervisor/advisor and one from the ESO local supervisor, and a letterfrom the home institution that i) guarantees the financial support for the remaining Ph.D. period after the termination of theESO studentship, ii) indicates whether the requirements to obtain the Ph.D. degree at the home institute are already fulfiled.

All documents should be typed and in English (but no translation is required for the certificates and diplomas).

The application material has to be addressed to:

European Southern ObservatoryStudentship ProgrammeKarl-Schwarzschild-Str.2, 85748 Garching bei München (Germany)

All material, including the recommendation letters, must reach ESO by the deadline (June 15); applications arriving after thedeadline or incomplete applications will not be considered!

For further information contact Christina Stoffer (cstoffer@eso.org).

Local Staff (1 December 2002 – 31 January 2003)

ARRIVALS

ALVAREZ BRAVO, Jose Luis, Laser SpecialistARREDONDO HAMAME, Diego, System AdministratorCARDENAS SCHWEIGER, Cesar, Electronics TechnicianGONZALEZ BURGOS, Sergio, Electronics TechnicianGUERRA ESCOBAR, Carlos, Instrumentation Maintenance

TechnicianGUTIERREZ NIETO, Fernando, Electronics TechnicianOLIVAREZ GONZALEZ, Francisco, Precision Mechanic

DEPARTURES

ACEVEDO CISTERNAS, Maria Teresa, Telescope InstrumentOperator

ADRIAZOLA PUCHALT, Patricia, Purchasing AssistantARAYA TORO, Jorge, Telescope Instrument OperatorBARRIGA CAMPINO, Pablo, Instrumentation Engineer

ACRONYMS USED AT ESO

Confused?Many acronyms are used in modern science, and the EPR

Department of ESO (i.e. the Education and Public RelationsDepartment of the European Southern Observatory) has fromtime to time received requests for translations. Clearly, it wouldhelp our friends and website-visitors to know what theseacronyms really mean. We have now reacted to this need by put-ting a reasonably complete glossary of acronyms used at ESO onthe web, at

http://www.eso.org/outreach/epr/eso-acronyms.html

Happy reading!

JARA MARQUEZ, Maria Francisca, Personnel AssistantMORALES CANCINO, Luis, DriverTORREJON KOSCINA, Arturo, Telescope Instrument OperatorVEGA TELLO, Rolando, Telescope Instrument Operator

49

ORGANIZATIONALMATTERS

Catherine Cesarsky: United Kingdom Be-comes Tenth Member of ESO 108, 1

Catherine Cesarsky: ESO Turns 40 109, 1

SPEECHES TO MARK THE ACCESSION OFTHE UK TO ESO

Lord Sainsbury, UK Science Minister 109, 2Dr. Arno Freytag, President of the ESO

Council 109, 3Dr. Catherine Cesarsky, ESO Director Gen-

eral 109, 3Gerry Gilmore: ESO and the UK. Why Does

the UK Need More Astronomy? 109, 4

ESO Turns 40. Perspectives from theDirectors General, Past and Present:Adriaan Blaauw: Reflections on ESO,

1957–2002 109, 7Lodewijk Woltjer 109, 9Harry van der Laan: From SEST to ALMA,

from NTT to OWL: Of Vision, Dreams andRealities 109, 10

Riccardo Giacconi 109, 11Catherine Cesarsky 109, 11Some Snippets of History 109, 12

TELESCOPES AND INSTRUMENTATION

W. Brandner, G. Rousset, R. Lenzen, N. Hu-bin, F. Lacombe, R. Hofmann, A. Moor-wood, A.-M. Lagrange, E. Gendron, M.Hartung, P. Puget, N. Ageorges, P. Bier-eichel, H. Bouy, J. Charton, G. Dumont,T. Fusco, Y. Jung, M. Lehnert, J.-L. Lizon,G. Monnet, D. Mouillet, C. Moutou, D.Rabaud, C. Röhrle, S. Skole, J. Spyro-milio, C. Storz, L. Tacconi-Garman, G.

Zins: NAOS+CONICA at YEPUN: FirstVLT Optics System Sees First Light107, 1

R. Kurz, S. Guilloteau, P. Shaver: The Ata-cama Large Millimeter Array 107, 7

L. Germany: News from the 2p2 Team107, 13

G. Monnet: Status of VLT 2nd GenerationInstrumentation 108, 2

R. Hanuschik and D. Silva: VLT Quality Con-trol and Trending Services 108, 4

Recent NAOS-CONICA Images 108, 9L. Germany: News from La Silla 108, 10L.-Å. Nyman, P. Schilke and R.S. Booth: The

Atacama Pathfinder Experiment 109, 18O. Le Fevre, D. Mancini, M. Saïsse, S. Brau-

Nogué, O. Caputi, L. Castinel, S. D’Odo-rico, B. Garilli, M. Kissler, C. Lucuix, G.Mancini, A. Pauget, G. Sciarretta, M. Sco-deggio, L. Tresse, D. Maccagni, J.-P. Pi-cat, G. Vettolani: VIMOS Commissioningon VLT-Melipal 109, 21

L. Germany: 2.2-m Team 109, 27L. Pasquini et al.: Installation and Commis-

sioning of FLAMES, the VLT MultifibreFacility 110, 1

F. Pepe, M. Mayor, G. Rupprecht: HARPS:ESO’s Coming Planet Searcher. Chasing

Exoplanets with the La Silla 3.6-mTelescope 110, 9

K. Kuijken et al.: OmegaCAM: the 16k × 16kCCD Camera for the VLT Survey Tele-scope 110, 15

A. Glindemann: The VLTI – 20 Months afterFirst Fringes 110, 18

B. Koehler, C. Flebus, P. Dierickx, M. Dimmler,M. Duchateau, P. Duhoux, G. Ehrenfeld,E. Gabriel, P. Gloesener, V. Heinz, R. Kar-ban, M. Kraus, J.M. Moresmau, N. Nina-ne, O. Pirnay, E. Quertemont, J. Strasser,K. Wirenstrand: The Auxiliary Telescopesfor the VLTI: a Status Report 110, 21

Roberto Gilmozzi and Jason Spyromilio:Paranal Observatory – 2002 110, 28

O. Hainaut: News from La Silla: ScienceOperations Department 110, 30

Two unusual views of La Silla 110, 31

CHILEAN ASTRONOMY

L. Bronfman: A Panorama of Chilean As-tronomy 107, 14

A. Reisenegger, H. Quintana, D. Proust, E.Slezak: Dynamics and Mass of theShapley Supercluster, the Largest BoundStructure in the Local Universe 107, 18

REPORTS FROM OBSERVERS

D. Baade, T. Rivinius, S. Štefl: To Be or Notto Be and a 50-cm Post-Mortem Eulogy107, 24

F. Courbin, G. Letawe, P. Magain, L. Wisotz-ki, P. Jablonka, D. Alloin, K. Jahnke, B.Kuhlbrodt, G. Meylan, D. Minniti: Spec-troscopy of Quasar Host Galaxies at theVLT: Stellar Populations and DynamicsDown to the Central Kiloparsec 107, 28

J. Sollerman and V. Flyckt: The Crab Pulsarand its Environment 107, 32

L. Vanzi, F. Mannucci, R. Maiolino, M. DellaValle: SOFI Discovers a Dust EnshroudedSupernova 107, 35

G. Rupprecht: A Deep Look at an ActiveGalaxy 107, 38

Coming Home at Paranal. Unique “Resi-dencia” Opens at the VLT Observatory

107, 39G. Hasinger, J. Bergeron, V. Mainieri, P. Ro-

sati, G. Szokoly and the CDFS Team: Un-derstanding the sources of the X-raybackground: VLT identifications in theChandra/XMM-Newton Deep Field South108, 11

Carme Gallart, Robert Zinn, Frederic Pont,Eduardo Hardy, Gianni Marconi, RobertoBuonnanno: Using color-magnitude dia-grams and spectroscopy to derive starformation histories: VLT observations ofFornax 108, 16

Thomas Rivinius, Dietrich Baade, StanislavŠtefl, Monika Maintz, Richard Townsend:The ups and downs of a stellar surface:Nonradial pulsation modelling of rapid ro-tators 108, 20

L. Pentericci, H.W. Rix, X. Fan, M. Strauss:The VLT and the most distant quasars

108, 24Daniel Harbeck, Eva K. Grebel, Graeme H.

Smith: Evidence for external enrichmentprocesses in the globular cluster 47 Tuc?108, 26

M. McCaughrean, H. Zinnecker, M. Ander-sen, G. Meeus, N. Lodieu: Standing onthe Shoulders of a Giant: ISAAC, Antu,and Star Formation 109, 28

L. Kaper, A. Castro-Tirado, A. Fruchter, J.Greiner, J. Hjorth, E. Pian, M. Andersen,K. Beuermann, M. Boer, l. Burud, A. Jaun-sen, B. Jensen, J.M. Castro-Ceron, S.Ellison, F. Frontera, J. Fynbo, N. Gehrels,J. Gorosabel, J. Heise, F. Hessman, K.Hurley, S. Klose, C. Kouveliotou, N.Masetti, P. Møller, E. Palazzi, H. Peder-sen, L. Piro, K. Reinsch, J. Rhoads, E.Rol, l. Salamanca, N. Tanvir, P.M. Vrees-wijk, R.A.M.J. Wijers, T. Wiklind, A. Zeh,E.P.J. van den Heuvel: Gamma-RayBursts: the Most Powerful Cosmic Ex-plosions 109, 37

R.E. Mennickent, C. Tappert, M. Diaz: Cata-clysmic Variables: Gladiators in the Arena109, 41

L. Wang, D. Baade, P. Höflich, J.C. Wheeler:Supernova Polarimetry with the VLT:Lessons from Asymmetry 109, 47

N.C. Santos, M. Mayor, D. Queloz, S. Udry:Extra-Solar Planets 110, 32

I. Labbé, M. Franx, E. Daddi, G. Rudnick,P.G. van Dokkum, A. Moorwood, N.M.Förster Schreiber, H.-W. Rix, P. van derWerf, H. Röttgering, L. van Starkenburg,A. van de Wel, I. Trujillo, and K. Kuijken:FIRES: Ultradeep Near-Infrared Imagingwith ISAAC of the Hubble Deep FieldSouth 110, 38

OTHER ASTRONOMICAL NEWS

F. Paresce: Release of Scientific Data fromVLTI Commissioning 107, 41

A. Renzini: VLT Science Verification Policyand Procedures 107, 41

D. Alloin: News from Santiago 107, 42ESO Studentship Programme 107, 42M. Dennefeld: The Third NEON Observing

School 107, 43A. Bacher: “Life in the Universe” Winners on

La Silla and Paranal 107, 43A. Bacher and L. Lindberg Christensen: In

the Footsteps of Scientists – ESA/ESOAstronomy Exercise Series 107, 44

Steve Warren: The UKIRT Infrared DeepSky Survey becomes an ESO public sur-vey 108, 31

M. Jacob, P. Shaver, L. Dilella, A. Gimenez:Summary of the ESO-CERN-ESA Sym-posium on Astronomy, Cosmology andFundamental Physics 108, 35

M. Sterzik: IAOC Workshop in La Serena:Galactic Star Formation across the stellarMass Stectrum 108, 40

MESSENGER INDEX 2002 (Nos. 107–110)

SUBJECT INDEX

50

Danielle Alloin and Luis Campusano: As-tronomical Virtual Observatories dis-cussed in Chile 108, 40

Andreas Glindemann: Hunting for Planets –GENIE Workshop at Leiden University

108, 40Young Stars in Old Galaxies – a Cosmic

Hide and Seek Game 108, 41A. Bacher and R.M. West: Information from

the ESO Educational Office 108, 42C. Madsen: ESO in the European Parlia-

ment 108, 43Rolf-Peter Kudritzki: Obituary Kurt Hunger

108, 43German Foreign Minister Visits Paranal Ob-

servatory 108, 45E. Mason and S. Howell: An Exciting Work-

ing Session on Cataclysmic Variables atESO/Santiago 109, 51

W. Hillebrandt and B. Leibundgut: ConferenceSummary: From Twilight to Highlight: thePhysics of Supernovae. ESO/ MPA/MPESummer Workshop 2002 109, 52

J.R. Walsh and M.M. Roth: Developing 3DSpectroscopy in Europe 109, 54

D. Hofstadt, L.-A. Nyman: Obituary Guiller-mo Delgado (1961–2002) 109, 55

C Madsen: Forty Years ESO – PublicAnniversary Activities 109, 56

A. Bacher and R.M. West: First TeachersTraining Course at ESO HQ was a GreatSuccess 109, 57

M. Kissler-Patig: Summary of the Workshopon Extragalactic Globular Cluster Sys-

tems hosted by the European SouthernObservatory in Garching on August27–30, 2002 110, 42

A. Glindemann: The VLTI: Challenges forthe Future. Workshop at Jenam 2002 inPorto 110, 44

P. Shaver and E. Van Dishoek: Summary ofa Meeting on Science Operations withALMA, held on Friday, 8 November 200110, 44

C. Madsen: Celebrating ESO’s 40th Anni-versary 110, 45

C. Madsen: Breaking the Ground for theEuropean Research Area – The Confer-ence ‘European Research 2002 110, 46

A Nobel Prize for Riccardo Giacconi 110, 47Agreement Between the Government of the

Republic of Chile and ESO for Estab-lishing a New Centre for Observation inChile – ALMA 110, 48

Sara Ellison: Eighty Nights Up a Mountain110, 48

ANNOUNCEMENTS

Personnel Movements 107, 45List of Scientific Preprints 107, 45The Very Large Interferometer: Challenges

for the Future (A Jenam 2002 Workshopto be held at Porto, Portugal, on Sep-tember 4–7) 108, 46

Personnel Movements 108, 46

ESO Fellowship Programme 2002/2003108, 47

List of Scientific Preprints (March–June2001) 108, 47

A Note from the New Editor 108, 48Workshop on “Structure Evolution and Cos-

mology: New synergy between ground-based observations, space observationsand theory” to be held at ESO/Santiago109, 58

Workshop on “Stellar Candles for theExtragalactic Distance Scale” to beheld at the Universidad de Concepción109, 58

Meeting on “Science Operations with theAtacama Large Millimeter Array” to beheld at ESO, Garching 109, 59

ESO Vacancy: Editor (EDG604) 109, 59Personnel Movements 109, 59Corrigendum 109, 59First Announcement of an ESO Workshop

on High-Resolution Spectroscopy in As-tronomy 110, 50

ESO Workshop on Large Programmes andPublic Surveys 110, 50

Personnel Movements 110, 50Daniel Hofstadt: Manfred Ziebell Retires

110, 51Jacques Breysacher: Christa Euler – Thirty-

Seven Years of Service with ESO! 110,51

New ESO Proceedings 110, 51ESO Workshop Proceedings Still Available

110, 51

AUTHOR INDEX

A

D. Alloin: News from Santiago 107, 42D. Alloin and L.Campusano: Astronomical

Virtual Observatories discussed in Chile108, 40

B

D. Baade, T. Rivinius, S. Štefl: To Be or Notto Be and a 50-cm Post-Mortem Eulogy107, 24

F. Courbin, G. Letawe, P. Magain, L. Wisotz-ki, P. Jablonka, D. Alloin, K. Jahnke, B.Kuhlbrodt, G. Meylan, D. Minniti: Spec-troscopy of Quasar Host Galaxies at theVLT: Stellar Populations and DynamicsDown to the Central Kiloparsec 107, 28

A. Bacher: “Life in the Universe” Winners onLa Silla and Paranal 107, 43

A. Bacher and L. Lindberg Christensen: Inthe Footsteps of Scientists – ESA/ESOAstronomy Exercise Series 107, 44

A. Bacher and R.M. West: Information fromthe ESO Educational Office 108, 42

A. Bacher and R.M. West: First TeachersTraining Course at ESO HQ was a GreatSuccess 109, 57

P. Benvenuti: Recovery of a historical docu-ment (ESO Turns 40. Some Snippets ofHistory) 109, 17

Adriaan Blaauw: Reflections on ESO,1957–2002 (ESO Turns 40. Perspectivesfrom the Directors General, Past andPresent) 109, 7

W. Brandner, G. Rousset, R. Lenzen, N. Hu-bin, F. Lacombe, R. Hofmann, A. Moor-wood, A.-M. Lagrange, E. Gendron, M.Hartung, P. Puget, N. Ageorges, P. Bier-eichel, H. Bouy, J. Charton, G. Dumont,T. Fusco, Y. Jung, M. Lehnert, J.-L. Lizon,G. Monnet, D. Mouillet, C. Moutou, D.Rabaud, C. Röhrle, S. Skole, J. Spyro-milio, C. Storz, L. Tacconi-Garman, G.Zins: NAOS+CONICA at YEPUN: FirstVLT Optics System Sees First Light107, 1

J. Breysacher: Early days of the OPC (ESOTurns 40. Some Snippets of History)109, 12

Jacques Breysacher: Christa Euler – Thirty-Seven Years of Service with ESO! 110,51

L. Bronfman: A Panorama of Chilean As-tronomy 107, 14

C

Catherine Cesarsky: United Kingdom Be-comes Tenth Member of ESO 108, 1

Catherine Cesarsky: ESO Turns 40 109, 1Catherine Cesarsky, ESO Director General:

Speech to mark the accession of the UKto ESO 109, 3

Catherine Cesarsky (ESO Turns 40. Per-spectives from the Directors General,Past and Present) 109, 11

C. Cesarsky: First Light of UT4 (ESO Turns40. Some Snippets of History) 109, 17

D

M. Dennefeld: The Third NEON ObservingSchool 107, 43

E

Sara Ellison: Eighty Nights Up a Mountain110, 48

D. Enard: The early days of instrumentationat La Silla (ESO Turns 40. Some Snippetsof History) 109, 13

F

Dr. Arno Freytag, President of the ESOCouncil: Speech to mark the accession ofthe UK to ESO 109, 3

G

Carme Gallart, Robert Zinn, Frederic Pont,Eduardo Hardy, Gianni Marconi, RobertoBuonnanno: Using color-magnitude dia-grams and spectroscopy to derive starformation histories: VLT observations ofFornax 108, 16

L. Germany: News from the 2p2 Team107, 13

L. Germany: News from La Silla 108, 10L. Germany: 2.2-m Team 109, 27Riccardo Giacconi (ESO Turns 40. Perspec-

tives from the Directors General, Past andPresent) 109, 11

51

Gerry Gilmore: ESO and the UK. Why Doesthe UK Need More Astronomy? (Speechto mark the accession of the UK to ESO)109, 4

Roberto Gilmozzi and Jason Spyromilio:Paranal Observatory – 2002 110, 28

Andreas Glindemann: Hunting for Planets –GENIE Workshop at Leiden University108, 40

A. Glindemann et al.: First Fringes withANTU and MELIPAL (ESO Turns 40.Some Snippets of History) 109, 17

A. Glindemann: The VLTI – 20 Months afterFirst Fringes 110, 18

A. Glindemann: The VLTI: Challenges forthe Future. Workshop at Jenam 2002 inPorto 110, 44

H

O. Hainaut: News from La Silla: ScienceOperations Department 110, 30

R. Hanuschik and D. Silva: VLT Quality Con-trol and Trending Services 108, 4

Daniel Harbeck, Eva K. Grebel, Graeme H.Smith: Evidence for external enrichmentprocesses in the globular cluster 47 Tuc?108, 26

G. Hasinger, J. Bergeron, V. Mainieri, P. Ro-sati, G. Szokoly and the CDFS Team: Un-derstanding the sources of the X-raybackground: VLT identifications in theChandra/XMM-Newton Deep Field South108, 11

W. Hillebrandt and B. Leibundgut: Confe-rence Summary: From Twilight to High-light: the Physics of Supernovae. ESO/MPA/MPE Summer Workshop 2002109, 52

D. Hofstadt: Renata Scotto at La Silla (ESOTurns 40. Some Snippets of History)109, 12

D. Hofstadt: La Silla vaut bien une Messe(ESO Turns 40. Some Snippets ofHistory) 109, 13

D. Hofstadt, L.-Å. Nyman: Obituary Guiller-mo Delgado (1961–2002) 109, 55

Daniel Hofstadt: Manfred Ziebell Retires110, 51

J

M. Jacob, P. Shaver, L. Dilella, A. Gimenez:Summary of the ESO-CERN-ESA Sym-posium on Astronomy, Cosmology andFundamental Physics 108, 35

K

L. Kaper, A. Castro-Tirado, A. Fruchter, J.Greiner, J. Hjorth, E. Pian, M. Andersen,K. Beuermann, M. Boer, l. Burud, A. Jaun-sen, B. Jensen, J.M. Castro-Ceron, S.Ellison, F. Frontera, J. Fynbo, N. Gehrels,J. Gorosabel, J. Heise, F. Hessman, K.Hurley, S. Klose, C. Kouveliotou, N. Ma-setti, P. Møller, E. Palazzi, H. Peder-sen, L. Piro, K. Reinsch, J. Rhoads, E. Rol,l. Salamanca, N. Tanvir, P.M. Vreeswijk,R.A.M.J. Wijers, T. Wiklind, A. Zeh, E.P.J.van den Heuvel: Gamma-Ray Bursts: theMost Powerful Cosmic Explosions109, 37

M. Kissler-Patig: Summary of the Workshopon Extragalactic Globular Cluster Sys-tems hosted by the European SouthernObservatory in Garching on August27–30, 2002 110, 42

B. Koehler, C. Flebus, P. Dierickx, M.Dimmler, M. Duchateau, P. Duhoux, G.Ehrenfeld, E. Gabriel, P. Gloesener, V.Heinz, R. Karban, M. Kraus, J.M. Mores-mau, N. Ninane, O. Pirnay, E. Querte-mont, J. Strasser, K. Wirenstrand: TheAuxiliary Telescopes for the VLTI: aStatus Report 110, 21

Rolf-Peter Kudritzki: Obituary Kurt Hunger108, 43

K. Kuijken et al.: OmegaCAM: the 16k × 16kCCD Camera for the VLT Survey Tele-scope 110, 15

R. Kurz, S. Guilloteau, P. Shaver: The Ata-cama Large Millimeter Array 107, 7

L

I. Labbé, M. Franx, E. Daddi, G. Rudnick,P.G. van Dokkum, A. Moorwood, N.M.Förster Schreiber, H.-W. Rix, P. van derWerf, H. Röttgering, L. van Starkenburg,A. van de Wel, I. Trujillo, and K. Kuijken:FIRES: Ultradeep Near-Infrared Imagingwith ISAAC of the Hubble Deep FieldSouth 110, 38

S. Laustsen: How ESO got its Optics Group(ESO Turns 40. Some Snippets of His-tory) 109, 12

O. Le Fevre, D. Mancini, M. Saïsse, S. Brau-Nogué, O. Caputi, L. Castinel, S. D’Odo-rico, B. Garilli, M. Kissler, C. Lucuix, G.Mancini, A. Pauget, G. Sciarretta, M. Sco-deggio, L. Tresse, D. Maccagni, J.-P. Pi-cat, G. Vettolani: VIMOS Commissioningon VLT-Melipal 109, 21

M

C. Madsen: ESO in the European Parlia-ment 108, 43

C Madsen: Forty Years ESO – PublicAnniversary Activities 109, 56

C. Madsen: Celebrating ESO’s 40thAnniversary 110, 45

C. Madsen: Breaking the Ground for theEuropean Research Area – The Confer-ence ‘European Research 2002 110, 46

E. Mason and S. Howell: An Exciting Work-ing Session on Cataclysmic Variables atESO/Santiago 109, 51

M. McCaughrean, H. Zinnecker, M. Ander-sen, G. Meeus, N. Lodieu: Standing onthe Shoulders of a Giant: ISAAC, Antu,and Star Formation 109, 28

R.E. Mennickent, C. Tappert, M. Diaz: Cata-clysmic Variables: Gladiators in the Arena109, 41

G. Monnet: Status of VLT 2nd GenerationInstrumentation 108, 2

A. Moorwood: The early days of infrared in-strumentation at ESO (ESO Turns 40.Some Snippets of History) 109, 15

N

W. Nees: ESO’s first step into the world ofminicomputers (ESO Turns 40. SomeSnippets of History) 109, 16

L.-Å. Nyman, P. Schilke and R.S. Booth: TheAtacama Pathfinder Experiment 109, 18

P

F. Paresce: Release of Scientific Data fromVLTI Commissioning 107, 41

L. Pasquini et al.: Installation and Commis-sioning of FLAMES, the VLT MultifibreFacility 110, 1

L. Pentericci, H.W. Rix, X. Fan, M. Strauss:The VLT and the most distant quasars108, 24

F. Pepe, M. Mayor, G. Rupprecht: HARPS:ESO’s Coming Planet Searcher. ChasingExoplanets with the La Silla 3.6-mTelescope 110, 9

R

A. Reisenegger, H. Quintana, D. Proust, E.Slezak: Dynamics and Mass of theShapley Supercluster, the Largest BoundStructure in the Local Universe 107, 18

A. Renzini: VLT Science Verification Policyand Procedures 107, 41

Thomas Rivinius, Dietrich Baade, StanislavŠtefl, Monika Maintz, Richard Townsend:The ups and downs of a stellar surface:Nonradial pulsation modelling of rapid ro-tators 108, 20

G. Rupprecht: A Deep Look at an ActiveGalaxy 107, 38

S

Lord Sainsbury, UK Science Minister:Speech to mark the accession of the UKto ESO 109, 2

N.C. Santos, M. Mayor, D. Queloz, S. Udry:Extra-Solar Planets 110, 32

P. Shaver and E. Van Dishoek: Summary ofa Meeting on Science Operations withALMA, held on Friday, 8 November 2002110, 44

J. Sollerman and V. Flyckt: The Crab Pulsarand its Environment 107, 32

M. Sterzik: IAOC Workshop in La Serena:Galactic Star Formation across the stellarMass Stectrum 108, 40

J.-P. Swings: First experience at La Silla,and some activities for the VLT (ESOTurns 40. Some Snippets of History)109, 13

V

Harry van der Laan: From SEST to ALMA,from NTT to OWL: Of Vision, Dreams andRealities (ESO Turns 40. Perspectivesfrom the Directors General, Past andPresent) 109, 10

L. Vanzi, F. Mannucci, R. Maiolino, M. DellaValle: SOFI Discovers a Dust EnshroudedSupernova 107, 35

W

J.R. Walsh and M.M. Roth: Developing 3DSpectroscopy in Europe 109, 54

L. Wang, D. Baade, P. Höflich, J.C. Wheeler:Supernova Polarimetry with the VLT:Lessons from Asymmetry 109, 47

Steve Warren: The UKIRT Infrared DeepSky Survey becomes an ESO public sur-vey 108, 31

R. West: Memories of early times at ESO(ESO Turns 40. Some Snippets of His-tory) 109, 12

R. Wilson: First Astronomical Light at theNTT (ESO Turns 40. Some Snippets ofHistory) 109, 16

Lodewijk Woltjer (ESO Turns 40. Perspec-tives from the Directors General, Past andPresent) 109, 9

52

ESO, the European Southern Observa-tory, was created in 1962 to “... establishand operate an astronomical observato-ry in the southern hemisphere, equippedwith powerful instruments, with the aim offurthering and organising collaborationin astronomy...” It is supported by tencountries: Belgium, Denmark, France,Germany, Italy, the Netherlands, Portu-gal, Sweden, Switzerland and the UnitedKingdom. ESO operates at two sites inthe Atacama desert region of Chile. Thenew Very Large Telescope (VLT), thelargest in the world, is located onParanal, a 2,600 m high mountain ap-proximately 130 km south of Antofa-gasta, in the driest part of the Atacamadesert where the conditions are excellentfor astronomical observations. The VLTconsists of four 8.2-metre diameter tele-scopes. These telescopes can be usedseparately, or in combination as a giantinterferometer (VLTI). At La Silla, 600 kmnorth of Santiago de Chile at 2,400 maltitude, ESO operates several opticaltelescopes with diameters up to 3.6 mand a submillimetre radio telescope(SEST). Over 1300 proposals are madeeach year for the use of the ESO tele-scopes. The ESO headquarters are lo-cated in Garching, near Munich, Ger-many. This is the scientific, technical andadministrative centre of ESO where tech-nical development programmes are car-ried out to provide the Paranal and LaSilla observatories with the most ad-vanced instruments. There are also ex-tensive astronomical data facilities. ESOemploys about 320 international staffmembers, Fellows and Associates inEurope and Chile, and about 160 localstaff members in Chile.

The ESO MESSENGER is publishedfour times a year: normally in March,June, September and December. ESOalso publishes Conference Proceedings,Preprints, Technical Notes and other ma-terial connected to its activities. PressReleases inform the media about partic-ular events. For further information, con-tact the ESO Education and PublicRelations Department at the followingaddress:

EUROPEAN SOUTHERN OBSERVATORYKarl-Schwarzschild-Str. 2 D-85748 Garching bei München Germany Tel. (089) 320 06-0 Telefax (089) 3202362ips@eso.org (internet) URL: http://www.eso.orghttp://www.eso.org/gen-fac/pubs/messenger/

The ESO Messenger:Editor: Peter ShaverTechnical editor: Kurt Kjär

Printed by Universitätsdruckerei WOLF & SOHNHeidemannstr. 166D-80939 MünchenGermany

ISSN 0722-6691

ContentsREPORTS FROM OBSERVERS

T. Ott et al.: Inward Bound: Studying the Galactic Centre with NAOS/CONICA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

A. Stolte et al.: NAOS-CONICA Performance in a Crowded Field – the Arches Cluster . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

M. Pettini: Early Galactic Chemical Evolution with UVES . . . . . . . . . 13O. Le Fèvre et al.: The VIRMOS-VLT Deep Survey:

a Progress Report . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18H. Boehnhardt et al.: Exploring the Icy World of the Edgeworth-

Kuiper Belt – An ESO Large Programme . . . . . . . . . . . . . . . . . . 22A. Cimatti et al.: The K20 Survey: New Light on Galaxy Formation

and Evolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29B. Venemans, G. Miley, J. Kurk, H. Röttgering, L. Pentericci:

Tracing the Formation and Evolution of Clusters and their Central Massive Galaxies to z > 4: a Progress Report . . . . . . . . 36

TELESCOPES AND INSTRUMENTATION

ESO and NSF Sign Agreement on ALMA. Green Light for World’s Most Powerful Radio Observatory . . . . . . . . . . . . . . . . . . . . . . . 39

New Vistas Open with MIDI at the VLT Interferometer . . . . . . . . . . . 40V. Doublier: TIMMI2 at the 3.6-m, or the Return of MIR . . . . . . . . . . 41

OTHER ASTRONOMICAL NEWS

NEWS FROM SANTIAGOD. Alloin, M. Pierre, A. Refregier: International Workshop on

“Structure Evolution and Cosmology: New Synergy BetweenGround-Based Observations, Space Observations and Theory” . 42

D. Alloin, W. Gieren: International Workshop on “Stellar Candlesfor the Extragalactic Distance Scale” . . . . . . . . . . . . . . . . . . . . . 43

B. Leibundgut: Fellows at ESO: Aurore Bacmann, Maria-Rosa Cioni,Ivo Saviane and Petri Vaisanen . . . . . . . . . . . . . . . . . . . . . . . . . 44

URANUS, Rings and Moons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45

ANNOUNCEMENTS

ESO Workshop on Large Programmes and Public Surveys . . . . . . . 46ESO Workshop on High-Resolution Infrared Spectroscopy in

Astronomy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46First Announcement of an ESO Workshop on Science with Adaptive

Optics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47Personnel Movements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47Acronyms Used at ESO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48ESO Studentship Programme . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48

MESSENGER INDEX 2002 (Nos. 107–110) . . . . . . . . . . . . . . . . . . 49

ESO Workshop Proceedings Still AvailableMany ESO Conference and Workshop Proceedings are still available and may be ordered atthe European Southern Observatory. Some of the more recent ones are listed below.

No. Title Price

54 Topical Meeting on “Adaptive Optics”, October 2–6, 1995, Garching, Germany. M. Cullum (ed.) 7 40.–

55 NICMOS and the VLT. A New Era of High Resolution Near Infrared Imagingand Spectroscopy. Pula, Sardinia, Italy, May 26–27, 1998 7 40.–

56 ESO/OSA Topical Meeting on “Astronomy with Adaptive Optics – Present Results and Future Programs”. Sonthofen, Germany, September 7–11, 1999. D. Bonaccini (ed.) 7100.–

57 Bäckaskog Workshop on “Extremely Large Telescopes”. Bäckaskog, Sweden, June 1–2, 1999. T. Andersen, A. Ardeberg, R. Gilmozzi (eds.) 7 30.–

58. Beyond Conventional Adaptive Optics, Venice, Italy, May 7–10, 2001E. Vernet, R. Ragazzoni, S. Esposito, N. Hubin (eds.) 7 50.–