1

No. 93 – September 1998

Science Verification Observations on VLT-UT1 Completed

Figure 1: The colour compos-ite constructed from the U + B,R and I VLT test camera imag-es of the Hubble Deep FieldSouth (HDF-S) NICMOS field.Exposure times are given inTable 2 of the editorial. The U+ B, R and I images are dis-played in the blue, green andred channels, respectively. Theimage is scaled from a low cutabout 1 σ below the peak ofthe sky noise histogram to ahigh point which makes the starbelow the large spiral galaxyapproximately white. The spi-ral galaxy itself has beenmasked and displayed with adifferent stretch to keep theinternal structure visible.

O B S E R V I N G W I T H T H E V L T

THE VLT-UT1 SCIENCE VERIFICATION TEAM

Science Verification (SV)observations on UT1 havetaken place as planned fromAugust 17 to September 1(cf. The Messenger, 92, 5,for a presentation of thegoals and the strategy ofSV). Although the meteoro-logical conditions on Para-nal have been definitely be-low average, very valuabledata have been gatheredand are now being preparedfor public release. The tele-

2

with their level of completion comparedto the initial planning. This includes oper-ational overheads, such as read-outtimes, target acquisition, etc. For thoseprogrammes that could not be completedcare was taken to complete the neces-sary observations for at least one object.

All SV data will be released by Sep-tember 30 to the ESO and Chilean com-munities. It will be possible to retrievethe data from the VLT archive, while aset of CDs will also be distributed to allAstronomical Research Institutes withinESO member states and Chile. Data onHDF-S will be public worldwide, and re-trievable from the VLT archive. Updatedinformation on data release can be foundon the ESO web site at http://www.eso.org/vltsv

Astroclimate During Science VerificationWhen, at one of the best observatories worldwide, over

two weeks and more, the sky is often cloudy, the seeingpoor, the wind fairly strong and blowing from unusual direc-tions, one is allowed to start talking of an astroclimatologicalanomaly.

When this occurs during the science verification of thefirst 8-m-class telescope mounting a monolithic mirror, theevent deserves a more detailed analysis.

Cloudiness at Paranal is the conjunction of seasonaltrends and El Niño events on top of some longer, as yetunexplained cycle (The Messenger 90, 6). As we are cur-rently in the lows of the latter cycle and despite the end ofthe 1997–98 El Niño event, August 1998 was promising lessthan 70% photometric nights: the two nights lost for cloudi-ness during the two weeks of science verification were thuswell within expectations.

The wind at Paranal is stronger in winter (30% of the timemore than 10 m/s) than in spring or summer (15% only): oneand a half nights lost because of wind in two weeks of ob-serving is thus not anomalous.

As clouds have no reason to prefer windy nights, the twoprevious effects tend to cumulate and the total time lost wasclose to 30% of the total available observing time, nothing tobe ashamed about!

Unfortunately, the seeing conditions were not at all insidethe statistical margin as can be seen on the figure: not onlydid we have an excess of very bad seeing (10% of the timeworse than 2 arcsec) but also a deficit of good seeing peri-ods (3 times less than normal).

This situation cannot be explained by a synoptic analy-sis: only a slight excess of temperature was reported overSouth-America during the first week, the jet stream behav-

iour was also quite normal during the whole period. Never-theless, the wind vane at Paranal more than usual keptpointing at north-east or south-east where the bad seeingcomes from (valleys and nearby summits). In addition, acold front causing a sudden drop of the air temperatureturned the ground around the observatory into a highlyefficient local seeing generator several degrees warmerthan normal.

Whatever further improvements we make in the under-standing of the generation mechanism of atmospheric tur-bulence, the operation strategy of ground-based astronom-ical facilities has nevertheless to be designed to confrontfrom time to time a highly non-deterministic environment.

M. Sarazin

TABLE 1: Summary of Science Verification Observations.

Programme Hours % of planned

HDF-S NICMOS and STIS Fields 37.1 98%

Lensed QSOs 3.2 82%

High-z Clusters 6.2 55%

Host Galaxies of Gamma-Ray

Bursters 2.1 56%

Edge-on Galaxies 7.4 65%

Globular cluster cores 6.7 57%

QSO Hosts 4.4 —

SN1987A 0.0 0%

TNOs 3.4 —

Pulsars 1.3 18%

Flats and Standards 22.7 99%

TABLE 2: VLT Test Camera Data on the HDF-S NICMOS Field.

Filter No. of Total integration FWHM of the

exposures time (sec) coadded image

U 16 17788 0.71″B 15 10200 0.71″V 16 14400 0.78″R 8 7200 0.49″I 12 10158 0.59″

scope has been working with spectacu-lar efficiency and performance throughthe whole period. After having been dis-assembled to install the M3 Tower, thetelescope was reassembled again puttingback in place the M1 mirror cell (August15). The Test Camera was re-installed atthe Cassegrain focus on August 16, thetelescope was realigned and tested, andfinally released to the SV Team at mid-night local time on August 17. The first

hours, dusk to dawn. Of these, 44 hourshave been lost due to bad meteorologi-cal conditions (clouds or wind exceeding15 m/s), and 15 hours for minor techni-cal problems, with an effective down timeof p 10%. For a total of 95 hours the tel-escope has been used to collect scientif-ic data, including twilight flat-fielding andphotometric standard star observations.

Table 1 gives the actual time investedon each of the SV programmes, along

SV observationswere promptly ini-tiated thereafter.The SV periodended on themorning of Sep-tember 1, span-ning a total of 142

3

Figure 2: The colour composite of the HDF-SNICMOS field constructed by combining theVLT test camera images in U + B and R withthe near-IR HST NICMOS/camera 3 F160W(p H band) 7040 s exposure. These imageswere used for the blue, green and red chan-nels, respectively. The NICMOS image wassmoothed to match the resolution of theR-band VLT image. The boundary of the NIC-MOS image is also shown.

Figure 3: The original, undegraded HDF-S NICMOS image of the same field shown in Figure 2.

The measured image quality on the testcamera frames has been often betterthan the outside seeing as measured bythe DIMM seeing monitor. At least partof this effect is due to the field stabilisa-tion operated by the secondary mirror,which worked in closed loop all throughthe SV period. Also the M1 active controlworked in closed loop through all theobservations. In practice, the figure ofthe primary mirror is optimised severaltimes per minute with no operationaloverhead. The seeing/image quality dataare now being analysed to gather a bet-ter understanding of the telescope per-formance and of the site seeing whileextreme meteorological conditions wereprevailing.

On the morning of September 1, thetelescope was returned to the Commis-sioning Team, and commissioning re-sumed.

Figure 1 on the front page is a colourcomposite of the HDF-S NICMOS fieldthat combines U, B, R and I frames withimage quality better than 0.9″, as listedin Table 2.

Figure 2 shows the colour compositeof the same field with the addition of theH-band HST/NICMOS (F160W) imagefrom the ST-ECF public archive reducedat ST-ECF by W. Freudling. The HSTimage (Figure 3) was obtained with nearlythe same total exposure time as the VLT(R-band) images, and their combinationis meaningful since the VLT and NICMOSimages reach similar depths. This is theresult of several effects compensatingeach other, such as the K-correction, thebetter angular resolution of the HST im-age (p 0.2″), and the larger collectingarea of the VLT.

All objects in the NICMOS image arealso noticeable in Figure 1, with the ex-ception of the very red object in the vicin-ity of the face-on spiral. The bright redobject near the bottom of the image wasnoted by Treu et al. (astro-ph/9808282)as being undetected on optical imagesto the limit of R = 25.9. This object is clear-ly present in all the VLT test camera coad-ded images, with the exception of theU-band image.

Figure 4 shows the colour compositeimage of the optical Einstein ring0047-2808 (Warren et al. 1996, MNRAS,278, 139), a z = 3.595 star-forming gal-axy which is lensed by a red elliptical at z= 0.485. Exposure times are 1 h in thenarrow-band filter NB559 (centred at the

redshifted Lyα of the distant galaxy) and900 s in B and V.

Bruno Leibundgut and Roberto Gil-mozzi of the SV Team conducted theobservations on Paranal, with the localassistance of Martin Cullum of the SVTeam and of Eline Tolstoy and Marc Fer-

rari (both ESO fellows). Jason Spyromi-lio, Anders Wallander, Marco Chiesa andStephan Sandrock of the VLT Commis-sioning Team ensured smooth telescopeoperations throughout the whole period.The Paranal Engineering Departmentunder Peter Gray provided all the main-

4

tenance and trouble shooting support thatwas needed.

The rest of the SV Team, includingGuido De Marchi, Francesco Paresce,Benoît Pirenne, Peter Quinn, Alvio Ren-

zini, and Piero Rosati guaranteed quickreductions and quality control of the datain Garching and prompt feedback to theTeam on Paranal. Fabio Bresolin andRodolfo Viezzer of the Office for Sci-

The First Steps of UT1M. TARENGHI, P. GRAY, J. SPYROMILIO AND R. GILMOZZI

Introduction

The Very Large Telescope is the re-sult of 20 years’ work by a large team ofdedicated persons. We thank them all fortheir contribution. During the last fewmonths we had the privilege to witnessexciting moments. The following noteswill enable the reader to share in thosemoments.

The Final Steps Up to First Light

During January and February 1998the mechanical structure of the telescopeunderwent a series of tests and tune-ups. These activities were undertakenwith the dummy cell and dummy sec-ondary units installed. A small 8-inchCelestron telescope was attached to thetelescope centrepiece, and a VLT tech-nical CCD was put at its focus. The guidescope had first light in March. A roughpointing solution using 8 stars was de-rived for the telescope, which gave anrms pointing error of 8 arcseconds. Thebasic pre-setting and tracking of the tel-escope were also tested. Using a VLTTCCD for the guide scope also allowedus to test the basic functionality of theautoguiding system.

The code running on UT1 is almostidentical to that running on the NTT, andvery few code integration problems havearisen. The year spent on the NTT cer-tainly has meant time saved on UT1. InGarching an additional control system,including TCCDs, routers and other pe-

ripherals, was also up and running. Thisallowed our colleagues at Headquartersto reproduce problems we were havingon the mountain and provide quick fixeswhenever possible.

Meanwhile in the base camp atParanal, a complete duplicate telescopecontrol system was established with iden-tical configurations to the one running

the telescope. The workstations and lo-cal control units in the base camp evenshared networking addresses with themachines on the mountain top. One sideeffect was that, given this configuration,only one set of these computers couldactually be connected to the Paranal net-work. The base camp control system wastherefore completely stand-alone. To

Figure 1: The start of the night.

Figure 4: The colourcomposite image ofthe optical Einsteinring 0047-2808, pro-duced by a red el-liptical at z = 0.485lensing a star-form-ing galaxy at z =3.595. A 1-hour im-age through a nar-row-band filter cen-tred at the redshift-ed Lyα of the dis-tant galaxy is codedgreen, while the900-s B- and V-bandimages are codedblue and dark red,respectively. Thelensing galaxy ap-pears dark red at thecentre of the ring.

ence extensively contributed to the re-ductions and calibrations. Robert Fos-bury and Richard Hook of the ST-ECFcombined the coadded frames to pro-duce the colour images presented here.

Results, problems, and strategy werediscussed in daily video-conferencesGarching-Paranal, that were also attend-ed by Massimo Tarenghi, the Director ofthe Paranal Observatory. The video-con-ferences took place at about noon Garch-ing time (6 a.m. on Paranal), with theParanal team reporting on the observingconditions and the observations complet-ed during the night, and the GarchingTeam reporting on the progress in in-specting and reducing the data of the pre-vious nights. Then, while the Paranalpeople were sleeping, data from the pre-vious night were inspected and reducedin Garching, with feedback on what wasbest to do during the following night be-ing emailed to Paranal several hours inadvance of the beginning of the obser-vations. The SV Team was really active24 hours a day.

5

transfer files between systems, neces-sary to keep the two systems aligned, anew class of network was used, knownaffectionately on the mountain as “foot-net”.

One problem found during the tests inMarch was that the telescope oscillatedin altitude by a few arcseconds. Exten-sive checks on the oil system by JuanOsorio and others did not reveal thecause. The problem was tracked downto the velocity controller that caused Mar-tin Ravensbergen and Toomas Erm someheadaches. However, they found the bugand following the fix they brought thebehaviour of the telescope tracking downto excellent values. Without any wind-loading the telescope tracks with accu-racies of 0.05 arcseconds rms for extend-ed periods of time. However, when wepointed the telescope into the wind, thewind shake was clearly noticeable. Thiswas expected to be the case from de-tailed simulations done during the designphase. The mechanism for compensat-ing for wind shake is field stabilisation i.e.rapid motion of the secondary mirrorequivalent to tip/tilt corrections. In theplanning this was not foreseen to be im-plemented before August.

The priorities were changed. Gianluc-ca Chiozzi and Robert Karban went backto Garching and worked furiously on thecontrol model to accelerate the imple-mentation.

The secondary unit of the telescopeand primary mirror cell were installed inthe telescope in March. Valiant efforts byMax Kraus, who spent a good fraction ofhis time trying to convince everyone thatan air-compressor was just like a tractorand he could fix it, which he did, togetherwith Erich Bugueno and German Ehren-feld along with a cast of ADS workersmanaged to fit the cell on to the tele-scope. The concrete dummy mirror wasthe location of many a conference onhow the cell should be attached to thetelescope for the first time. The behav-iour of the telescope with a configurationvery close to the one expected when thereal glass went in could then be tested.Changes in the servo loop parametersand other such niceties, which are criti-cal for the correct and safe behaviour ofthe telescope, were made.

The primary mirror cell was undergo-ing qualification tests now that we couldmove it around and tip it over. Inclina-tion tests had been done in France pri-or to delivery to Paranal but now wehad a real chance to move the thingaround. It was hard to get StefanoStanghellini, Gerhard Hudepohl or MarcSbaihi out from inside the cell. We knewtheoretically that pushing the emergen-cy stop on the telescope under full speedwould not activate the earthquake de-tector on the cell. Would it in practice?Would the airbag system inflate auto-matically when the cell tipped over 75degrees? How far did the mirror movein the cell when inclined?

The test camera had now been in-stalled on the telescope. Martin Cullumand Ricardo Schmutzer were testing thesoftware to make sure that when weneeded to open the shutter, the instru-ment would actually take a good image.Flat fields and bias frames were takenconstantly.

It was time to coat the mirrors beforeputting them into the telescope. Both theprimary and the secondary mirrors hadbeen on the mountain for a while sittingin their boxes waiting for their turn. Sothe coating chamber was fired up for afinal check and ... we had a failure in thesystem which damaged the aluminiumtarget used for the sputtering. Our col-leagues in Linde, the prime contractor forthe coating unit, were despondent. Notonly had the unit failed at a critical timebut they would also have to come all theway back to Paranal to repair it. A quickrecovery plan was needed. MichaelSchneermann went off to find anothertarget (a non-trivial task since these arecustom-made out of the purest alumini-um) and Linde tried to find out what hadgone wrong.

We now had to find a way to proceedwithout endangering the timeline for thetelescope nor of course any of the op-tics. The secondary mirror went to La Si-lla, under the watchful eyes of PaulGiordano. Paul travelled with it along the70 km of dirt road linking Paranal to thePan-Americana highway at the excruci-atingly slow speed of 5 km/h and thenthe other 700 km down to La Silla. Paulcoated the secondary mirror at the 1.5-m tank and returned to Paranal. The qual-ity of the coating and the mirror were ex-cellent with reflectivities above 90 percent and micro-roughness around 10Angstroms. The mirror was then put into

the unit, which, as mentioned above, wasalready installed in the telescope.

The primary was not so simple. Wedecided that we should install the uncoat-ed mirror in the telescope and go aheadwith the final preparations towards firstlight. Our integration plan, released backin 1996, for the first-light specification, al-lowed us to have an uncoated mirror atfirst light. In order not to delay the availa-bility of the VLT to the community in April1999, we decided (in agreement with theDirector General) to proceed with first lightas planned. So, in the third week of April,the cell went on its way down to the basecamp, had the dummy mirror removedand the 8.2-m zerodur thin meniscus in-stalled. On the night of the 21st of April,the telescope was almost all in place. Boththe primary and the secondary mirrorswere in place. We had already plannedfor a small celebration when these eventshad taken place. In the base camp, ESOand SOIMI had a party. For some of us ithad to be cut short. We decided that onthe same night we would try the telescopeoptics out. Before doing so, the protec-tive plastic over the primary, which wasplaced on the mirror to protect it duringtransport from Europe to Paranal, had tobe removed. Francis Franza and PaulGiordano put on their clean-room cloth-ing, climbed on the mirror and startedpeeling the plastic off. We had thoughtthe operation would take an hour or sobut ended up taking three hours. Now wehad a real telescope. The primary mirrorlooked truly beautiful. A couple of testswith the enclosure closed to ensure eve-rything was O.K. and the big moment ar-rived. The enclosure doors were opened,the mirror cover retracted, and we point-ed the telescope at a globular cluster. Zeroforces were set on the primary and then

Figure 2: The team on the night of April 25.

6

the telescope was refocused manually.In a moment of great relief, the “first star”appeared on the guide probe. In retro-spect, looking at that image, it is obviousthat we had left the reference light arm(without the light on) in the beam, whichexplains the shadow one can see on thatimage. The star was about 2 arcsecondsin size. Lothar Noethe quickly let us allknow that this was exactly according tothe specification for the zero forces con-figuration of the primary. Some of us hadexpected the star to just drift away, oth-ers that it would look nothing like a star.In fact, the telescope did us proud. Im-mediately the Director General was in-formed.

Now the task of running the telescopeat night passed to the commissioningteam although officially we had not hadfirst light. Active optics tuning was themain task. Lothar Noethe, StephaneGuisard and Roberto Abuter started map-ping the aberrations of the telescope,their orientation and their dependence onzenith distance. After a couple of nightsof looking at 2 arcsecond stars and fromtime to time at stars that looked like piec-es of string or propellers (when Lotharinduced aberrations to check the behav-iour of the cell), the time came to closethe loop. The DIMM was not running thatnight, so we do not know what was theoutside seeing. When the loop closed, thestar became really small. We started tak-ing short exposures with the guide probeand measuring the FWHM. The activeoptics was working away continuouslycorrecting the mirror shape and the posi-tion of the secondary mirror. The first im-ages were 0.8 arcseconds in size. Greatjubilation in the little wooden hut insidethe enclosure where all of the control wastaking place. Stephane Guisard was plac-ing bets as to whether we would beat the3.6-m record. A few minutes later as theactive optics worked, the images wentdown to 0.4 arcseconds. Only 3 nightsafter the optics had gone in, the telescope

was already matching our highest expec-tations.

Pointing solutions and field stabilisa-tion tests were started in order to improvethe performance of the telescope. Point-ing quickly came down to around the 3-arcsecond rms level. Field stabilisationbaffled us all for a while. The nature ofvarious time delays and the synchroni-sation of the TCCD with the secondaryunit kept people busy for a while. Anto-nio Longinotti, our CCD software expert,made a couple of configuration changesand now we could move the M2 unit atfrequencies up to 20 Hz.

Although first light was specified forthe night of the 25th of May, the internalplanning target date was the 15th ofMay. By this time we had moved out ofthe hut in the enclosure and were oper-ating the telescope from the relativecomfort of the control room. On thenight of the 15th of May we decidedthat we should meet all specificationslaid out in the integration plan for thetelescope. The target was to be ω Cen.Conditions were excellent: low wind andgood seeing. We started a 10-minuteexposure on target with the test cam-era. We had never tried anything aslong as this. Krister Wirenstrand anx-iously waited for the test camera CCDto read out. This was to be the first trueimage taken with the telescope on ascientific CCD. When the image wastransferred to the Real Time Display,we quickly measured the image quality.Great jubilation again as the stars ap-peared at 0.48 arcseconds. A series ofother measurements on tracking stabili-ty and image quality verified the tele-

been integrated into the coating unit. Per-formance verification of the coating unitby Linde was under way. On the 18th ofMay the coating unit was ready. At 4 a.m.the telescope was stopped and parkedin the mirror removal configuration. Mar-tin Cullum and Francis Franza startedtaking the test camera off and by middaythe mirror cell was off the telescope.

That night the cell and mirror were inthe base camp. The mirror was detachedfrom the cell overnight and the followingday lifted out of the cell and into the coat-ing unit. Our washing unit is not yet onthe mountain. However, visual inspectionof the mirror showed only light dust hadsettled during the 4 weeks the mirror hadbeen in the telescope. Paul Giordano andFrancis Franza started the long and la-borious cleaning of the mirror using car-bon dioxide snow. This worked very well,especially at the edge of the mirror. Themirror was now as clean as we could getit. The coating unit was sealed, evacuat-ed and then the mirror was coated. Thetime had come to see what it would looklike. We were concerned that since somedust had been left on the mirror, the coat-ing might fail. We were glad to be provenwrong. Although better coatings willcome, the first was already good. Thereflectivity was above 90 per cent aroundthe edges of the mirror and dropped to89 per cent in areas where the CO2 clean-ing had not worked so well.

The mirror was put back into the celland driven up to the telescope the follow-ing morning. Two nights and three daysafter it was removed, the completed op-eration returned the telescope back forfurther tests before first light. Would the

Figure 4: Informing the DG.Figure 3: The first star.

scope had met allthe performancecriteria for firstlight.

By now, the newaluminium targethad arrived on themountain and had

7

pointing solution have changed? Wouldthe alignment and active optics calibra-tion need to be repeated? The pointingindeed had changed. The stars appeareda full 2 arcseconds away from where theywere before the entire operation tookplace. Such a small change encouragedus that the active optics would not needre-calibration. Indeed there seemed to beno need. The telescope was delivering0.4 arcsecond images yet again.

Now all we had to do was wait for the25th of May and keep our rendezvouswith the press. Of course we could notresist and took images on the nights lead-ing up to the first-light night. Julio Navar-rete was on hand in the ASM hut runningthe DIMM and the meteorological stationand answering the constant call on theradio: “Julio, can we have seeingplease?” The night of the 24th was beau-tiful. Things had been going too well. Atthe beginning of the night the earthquakedetector on the mirror cell went off. Ste-fano Stanghellini and Marc Sbaihi workedto release the mirror from the safetyclamps and a few hours later the tele-scope was available again. Then at theend of the night when closing the tele-scope down, a major problem occurred.The mirror cover jammed half way acrossthe mirror. Could it be repaired in time forthe night of the 25th? Marc Sbaihi andthe ADS crew came to the rescue. Work-ing just above the mirror from early in themorning until after sunset, they managedcarefully to open the cover and provide

the beam to the telescope. The final hur-dle had been overcome. The telescopewas operational again. All this effort – andthen the weather worked against us. Thefirst half of the night things went well andsome images were taken. Krister Wiren-strand operated the telescope while An-ders Wallander made sure the test cam-era took the images. However, in the sec-ond half following our little internal cele-bration, the weather was poor and weshut the telescope down.

Commissioning of UT1

The commissioning of UT1 officiallystarted immediately after first light. Mostactivities in commissioning involve tun-ing of telescope parameters and under-standing how UT1 should be used. Anumber of software modifications arebeing made based on this better under-standing that we have developed.

It took us far too long but eventuallywe realised that we had been focusingthe telescope in the wrong way. Weworked it out and on the 1st of June anew procedure was used. Since then thetelescope has been in autofocus mode.We have made the guide probe parfocalwith the instrument and have let activeoptics handle the telescope focus. In thismode the telescope focus is maintainedcontinuously throughout an exposure asthe active optics runs.

Improvements in the field stabilisationhave been taking place. Birger Gustafs-

Portuguese Minister of Science at ParanalOn Sunday, July 19, 1998, ESO

was honoured to receive a visit by thePortuguese Minister of Science andTechnology, Professor Mariano Gago,to the Paranal Observatory. The Min-ister was accompanied by the Ambas-sador of Portugal to the Republic ofChile, Mr. Rui Félix-Alves and a dele-gation.

The Minister visited the various VLTinstallations and, a scientist himself,expressed great interest in this newfacility, now being constructed by theEuropean Southern Observatory onbehalf of the ESO member states. Asforeseen in the 1990 Agreement thatassociates Portugal and ESO, discus-sions about future Portuguese mem-bership in ESO have started.

The Portuguese Minister of Science andTechnology, Professor J.M. Gago, with theVLT Project Manager and Director of theParanal Observatory, Professor Massimo Tarenghi (right), ESO astronomer Dr. Jason Spyromilio (left) and members of the delegation inthe VLT Control Room.

son has been reducing the delays in theM2 and Philippe Duhoux improved thecentroiding algorithm in the CCD soft-ware. These changes improve the per-formance of the telescope under heavywind load.

A lot of small changes here and thereimproved the reliability of the system andthe operability of the telescope. MarcoChiesa worked on the control algorithmfor the enclosure rotation and parkingwhich has made the operation muchsmoother and faster. Thanh Phan Ducimproved the guide-star acquisition pro-cedure significantly.

Marc Sarazin, Stefan Sandrock andRodrigo Amestica have brought the ASMto fully automatic status. CommissioningUT1 includes working with a cm-classtelescope and understanding its prob-lems as well. We are learning what it isto have a fully automatic telescope run-ning.

Paranal is truly a beautiful site. Thewinter has given us quite a number ofnights with poor conditions, sometimes itis cloudy and occasionally the seeingdoes go above 1 arcsecond. However,the beautiful nights are truly spectacular.On the night of the 26th of June, AndersWallander and Ivan Muñoz took a seriesof 30-second exposures with an imagequality below 0.35 arcseconds includingone at 0.27 arcseconds.

The control and quality of the optics isexcellent. Long exposures (900 and 1800seconds) are taken as a matter of course

8

to check the performance of the telescopeunder realistic observing conditions. Thetelescope routinely matches the outside.The active optics is run in continuousclosed loop.

FORS, the first instrument to go on tothe telescope in September, is alreadybeing re-integrated on the mountain, andISAAC, which goes onto the telescopein November, is already integrated in theControl building.

A first commissioning/installation of thedata-flow software was undertaken byPeter Quinn, Michèle Péron and MiguelAlbrecht in June. All data from UT1 arenow being archived immediately afterthey are taken. This includes the exten-sive operations logs that record all ac-tions of the system. For example, all ab-errations calculated in every active op-tics calculation are logged. All tempera-tures of the telescope, and there aremany, are also logged. Every preset, off-set, change in guide star and many otheractions are logged. All telescope errorsor unforeseen events are also logged. Weuse this information to better understandthe telescope and how it can be opti-mised.

A lot of work remains to be done. Thetertiary mirror will go into the telescopein August and the Nasmyth foci can then

see light for the first time. The Linear At-mospheric Dispersion Compensator forthe Cassegrain focus also goes into thetelescope in August. Commissioning ofthe Nasmyth foci will take place betweeninstrument installations. A better under-standing of the dome louvers and howthey affect the telescope performance ishigh on our priority list.

Science verification of the telescopeis scheduled for the dark run in August,and we fully expect some beautiful datato result from these two weeks.

The cast of people working towardsa successful VLT is too great to mentionexplicitly in such an article. The admin-istrative support both in Garching, San-tiago and Paranal that somehow man-aged to get all the pieces onto the moun-tain in time are thanked. Isabel Osoriowas ever present and helping with pret-ty much everything. We also thank LaSilla for providing us with coating facili-ties. Special thanks are due to ArminSilber, Enzo Brunetto, Mario Kiekebusch,Olaf Iwert and Claudio Cumani who allworked to get the test camera going;Marco Quattri who spent most of thenorthern winter on the mountain moni-toring the erection of the telescope;Mathias Hess who worked tirelessly onthe M1 cell and the transport of the mir-

ror and then missed first light by a fewweeks; Jean-Michel Moresmau whomanaged to fit the Cassegrain adapterinto the cell and hook up all those littlecables and wires; Manfred Ziebell whonever stopped worrying about everythingand anything and Michel Duchateau whoworried about all the little details likeemergency stop buttons; Jörg Eschweyand the facilities department on Paranalwho built most of the things around usand who also switched all the lights offin the base camp; Canio Dichirico whomade sure we had power when we need-ed it; Bruno Gilli, Gianni Raffi, GiorgioFilippi and others in the software groupwho kept us on the true path. Of course,we thank the entire VLT division inGarching for designing and building suchan excellent telescope; our system andnetwork administrators on the mountain,Chris Morrison, Nick Lock, Marcelo Car-rasco, Sebastian Lillo, Graeme Ross,Mark Tadross and the ever present Har-ry Reay who made sure all systems wereready. We apologise to all that we havemissed in our thanks.

The Cost of the VLTText of a Report of the ESO Director General to Council at its (extraordinary) meetingof September 15, 1998

The purpose of this document is to pro-vide an overview of the VLT cost evolu-tion since 1987, and an estimate as ofto-date of the total cost of this project atESO until the completion of the VLT/VLTIin 2003. The total cost we quote includesthe external contractual costs, the inter-nal labour and other costs directly VLT-related, the development of the ParanalObservatory and its operations until 2003.This is the date that will see the comple-tion of all 8-metre telescopes, the auxil-iary telescopes and all approved instru-mentation. It does not include the man-power costs at member state institutescontributing to the instrumentation pro-gramme.

The information compiled here is ex-tracted from documents already submit-ted to Council. However it appears usefulafter the successful completion of firstlight, which removes any major technicaluncertainties, to present a summary giv-ing Council a global perspective on theproject.

1. The VLT Contractual Cost

The VLT Programme was approved in1987 on the basis of a proposal known

as the “Blue Book”, which gave a costestimate of 524 MDM (1998 prices) forthe external contracts, including 34.8MDM (1998 prices) for VLTI.

Since then the scope of the VLT Pro-gramme has evolved considerably to in-clude some major new features and morecomplex solutions. These include the in-troduction of a Cassegrain focus andadapters with their subsequent impact onthe M1 Cell and Main Structure, sophisti-cated test cameras, a time reference sys-tem, astronomical site monitors, etc.

In 1993 a complete Cost to Comple-tion analysis was performed (Cou-483and Add.) and subject to an external au-dit in 1994. Council subsequently ap-proved a VLT Programme for 592 MDM(1998 prices) in which the Interferometrypart had been postponed and its fundingreduced to 7 MDM (Cou-516 conf.).

ESO worked out a recovery plan forthe VLTI, which was presented to Coun-cil in 1996. The cost of this new plan (32MDM in 1998 prices) was financed byreprogramming within the VLT and Instru-mentation programme, by an additionalcontribution of 10 MDM by MPI andCNRS and through release of contingen-cy funds.

Thus in the period 1994 to 1998 wehave not changed the VLT programmecost to member states while fully restor-ing the VLTI programme.

The current VLT cost ceiling of 602MDM for external contracts appears quitefirm since we are at a point in the pro-gramme where we have committed 87%of the contractual cost. The increase of15% with respect to the Blue Book Valueof 1987 is fully justified by the substantialchanges in scope of the project men-tioned above.

2. Total Cost of the VLT Pro-gramme

The Blue Book did not provide an esti-mate of the total cost (including ESO staffand other internal costs). An evaluationof the VLT-related internal costs, i.e. en-gineering costs in Garching and site costsat Paranal, was performed in 1993 andalso subject to the 1994 external audit.

In 1996 ESO submitted to Council along-range plan to bring the Organisationto the steady state of operations of thefull VLT/VLTI (1996–2003) while imple-menting strict cost containment measuresto meet a reduction in the projected mem-

M. Tarenghimtarengh@eso.org

9

Figure 2: VLT site cost (yearly) (ESO internal staff, operations and investments).

Figure 1: VLT Garching cost (yearly) (ESO internal staff, operations and investments).

10

Figure 3: Evolution of ESO staffing between 1988 and 1998.

1 Please note that all figures have been entirelyupdated to 1998 prices to provide for a coherent com-parison, including the expenditure actually occurredin the past years, which accounting value would nor-mally not be updated. The current “mixed” value ofthe VLT contracts is 574 MDM while the current“mixed” value of the VLT total cost estimate is 944MDM.

ber states contributions and still restorethe VLTI programme. Reductions in costand delays in spending were also neces-sary to reduce the anticipated negativecash flow projections.

As a result, the actual internal cost ofthe VLT engineering in Garching between1993 and 1998, and the projected esti-mates up to 2003 are in fact substantiallylower than the 1993 estimate (Fig. 1).

The evolution of the Paranal site costalso shows a decrease (Fig. 2). Prelimi-nary VLT operation cost after 1999 wasindicated in the Blue Book. This is alsogiven in Figure 2 for comparison.

The VLT total cumulative cost to com-pletion (including the initial operations ofthe Paranal Observatory) is estimated toreach 990 MDM (1998 prices) in 2003.Thus while the contract costs increasedfrom 592 MDM to 602 MDM (due to theadditional contributions of 10 MDM forVLTI), the total VLT/VLTI costs at comple-tion decreased from the 1994 estimate of1060 MDM to 990 MDM (1998 values).1

3. ESO Overheads

The ESO organisation has substan-tially grown and evolved since 1988 to

cope with a programme whose value inexternal contracts is 8 fold its 1987 an-nual budget (total cost 13 times the 1987budget) (Fig. 3).

In 1988, the ESO staff complementwas composed of 146 International Staffpositions (ISM), 126 Local Staff in Chile(LSM), 27 Fellows and Paid associates(Fel/PA). In 1998 the staff complementincludes 239 ISM (+64%), 151 LSM(+20%) and 34.5 Fel/Pa (+28%).

In 1988, 50 ISM (34% of the total ISM)worked for the Office of the Director Gen-eral, Science Office and Administration.There are 63 ISM (26% of the total ISM)working in these areas in 1998. In these10 years we have therefore decreasedthe fraction of the staff not directlyengaged in programmatic activities.

The rest of the staff works directlyfor the core activities of ESO: VLT, In-strumentation, Paranal Observatory,Data Management and Operations, LaSilla Observatory, and ST-ECF (Fig. 3).During the same period and even priorto the initiation of the VLT programmethe number of astronomers using theESO facilities (visiting astronomers) hasincreased by 20%. We expect thisnumber to more than double when theVLT initiates science operations (April 1,1999).

The comparatively low increase of themanagement and administrative resourc-es is a very clear indication that the over-head cost of ESO in general has de-creased compared to the era before VLT.(More detailed comparisons are difficult

due to the accounting tools available in1988).

4. Conclusion

In conclusion, the VLT contract costceiling has increased 15% in compari-son to the Blue Book (1987), an increasewhich is fully justified by the technicalevolution of the Programme scope. Thecost which now includes a fully restoredVLTI Programme has remained un-changed compared to the approved Costto Completion of 1994. A comparison ofthe cumulative VLT total cost (includingVLT contract cost, personnel cost, oper-ations, site development and invest-ments) at the time in 2003 when VLT willbe in full operational phase shows a de-crease from 1060 MDM to 990 MDM withrespect to the estimates audited in 1994.

These results were obtained in spiteof the contractual and overall cost in-creases due to the delays experiencedin Chile as a result of the Paranal own-ership issues.

At the same time, the Organisationhas improved its efficiency by contain-ing the growth of the administrativestaff and reducing its proportion of thetotal.

Cost and technical performance at thelevel achieved on the VLT programme israre among major scientific/technical en-terprises and could not have beenachieved without the professionalism,competence and dedication of the ESOstaff.

11

ESO and AMOS Signed Contract for the VLTIAuxiliary TelescopesB. KOEHLER, ESO

The company AMOS (Liège, Belgium) has been awarded,last June, an ESO contract for the delivery of the Auxiliary Tele-scopes (ATs) of the Very Large Telescope Interferometer (VLTI).Each of these telescopes has a main mirror of 1.8-metre diam-eter. They move on rail tracks on the top of the Paranal moun-tain. Together with the main 8.2-m VLT Unit Telescopes (UTs),they will ensure that the VLTI will have unequalled sensitivityand image sharpness that will allow front-line astronomical ob-servations.

This contract was signed for the design, manufacturing andtesting in Europe of two ATs and of the full set of on-site equip-ment for the 30 AT observing stations. An option for a third Aux-iliary Telescope is also part of the contract. The delivery in Eu-rope of the first AT is planned for June 2001 and the first obser-vations with the first two ATs at Paranal are planned for early2002.

More details can be found at: http://www.eso.org/outreach/press-rel/pr-1998/phot-25-98.html

The photo of a 1/20 scale model built by AMOS in response to the callfor tender illustrates the main conceptual features of the VLTI AuxiliaryTelescopes. The 1.8-m telescope (with an Alt-Az mount, i.e. exactly

UT1 Passes “With Honour” the First SevereStability Tests for VLTIB. KOEHLER, F. KOCH, ESO-Garching

1. Introduction

Over the past years, a significant efforthas been put in verifying and improvingthe capability of the VLT Unit Telescope(UT) to reach the very demanding OpticalPath Length (OPL) stability at the nano-meter level, as required by the VLT Inter-ferometer (VLTI). Up to now, this has beendone primarily by analysis using detailedFinite Element Models with inputs fromdedicated measurement campaigns suchas the characterisation of micro-seismicactivity at Paranal [1], [2], vibration testson IR instruments closed-cycle coolersand pumps [3], as well as tests at sub-system level on the M2 unit, on the enclo-sure, on the telescope structure equippedwith dummy mirrors, etc. [4].

With the commissioning of the first VLTtelescope at Paranal, time has come todirectly measure the dynamic stability ofthe 8-m telescopes in real operationalconditions.

A dedicated commissioning task wasundertaken on July 23–30 to monitor themirror vibrations with highly sensitive ac-celerometers. A brief summary of the re-sults is presented here.

2. Measurement Set-upThe measurement equipment consist-

ed of eight high-sensitivity accelerome-ters (Wilcoxon 731A) connected to twodigital acquisition units (DSPT SigLab 20-42) controlled from a PC running Matlab.

The accelerometers were placed asfollows:

4 accelerometers attached at the out-er edge of the primary mirror M1 sensingmotion along the optical axis. The signalsare averaged to obtain an estimate of theM1 axial displacement (piston).

1 accelerometer inside the M2 unitmonitoring the motion of the mirror alongthe optical axis.

1 accelerometer on the M4 arm of theNasmyth Adapter–Rotator sensing the

motion along the normal of the future M4mirror.

1 accelerometer on the M5 unit attach-ment flange sensing the motion along thenormal of the future M5 mirror.

1 accelerometer on the M6 unit attach-ment flange sensing the motion along thenormal of the future M6 mirror.

Post-processing was done using ded-icated routines written in Matlab.

3. Main Results

Figure 1 shows the fringe visibility lossresulting from the mirror vibrations in var-ious operational conditions and for threedifferent observing wavelengths: visible(0.6 µm), near infrared (2.2 µm) and ther-mal infrared (10 µm). The VLTI errorbudgets call for a 1% visibility loss due tovibrations inside the telescope for any ofthese observing wavelengths. This cor-responds respectively to an OPL varia-tion of 14, 50 and 215 nanometers r.m.s.

like the Unit Telescopes) is shown here in observing conditions. It is rigidly anchored to the ground by means of a special interface. The light isdirected via a series of mirrors to the bottom of the telescope from where it is sent on to the underground delay line tunnel. The AT Enclosureconsists of segments and is here fully open. During observation, it protects the lower part of the telescope structure from strong winds. TheEnclosure is supported by the transporter (the blue square structure) that also houses electronic cabinets and service modules (the grey boxes)for liquid cooling, air conditioning (the red pipes), auxiliary power, compressed air, etc., making the telescope fully autonomous. When thetelescope needs to be relocated on another observing station, the transporter performs all the necessary actions such as lifting the telescope,closing the station lid (the white octagon), translating the telescope along the rails, etc. The complete relocation process will take less than 3 hoursand shall not require re-alignment other than those performed remotely from the control room at the beginning of the next observation.

12

over 10, 48 and 290 msec for the threewavelength bands.

The left-most case marked “All OFF”on Figure 1 shows the results when basi-cally all sub-systems of the telescope areswitched off. It represents the contribu-tions of the background environmentnoise. It is well below the 1% level for allwavelengths.

The second case shows the influenceof the altitude and azimuth Hydro-StaticBearing System (HBS). The insignificantimpact of the HBS is one of the most com-forting results of these tests, since it wasstill largely unknown and potentially im-portant, as indicated from results on oth-er existing telescopes. The credit goes tothe use of screw-type pumps, a good iso-lation of the pumps and careful overalldesign.

On the other hand, the third caseshows a significant impact, for observa-tion in the visible, of the liquid-coolingpumps located in the basement of theenclosure. It has been checked that thesevibrations are mainly transmitted to thetelescope through the ground and tele-scope pier and not through the distribu-tion pipes fixed on the telescope struc-ture. Here, easy improvement is possibleby better isolation between the pumps andthe ground and between the pipes andthe telescope pier.

The next case shows a slight influence(especially at 10 µm, i.e. low-frequencydisturbance) of the altitude and azimuthmotors’ noise. The exact origin of this low-frequency disturbance is still not fullyunderstood but it will very likely improvewhen the bandwidth of the axis controlwill be increased to its nominal value.

From the tests performed during track-ing of the telescope, we can conclude thatthe associated disturbance remains ac-ceptable for IR wavelengths and remainsdominated by the cooling pumps for thevisible except for the highest speed of1500″/sec which corresponds however tothe quite exceptional case of observingat 0.5° from the zenith.

The next cases during which the en-closure was also tracking evidence aslight deterioration both for visible andnear IR. Improvement in this area is pos-sible by a better tuning of the enclosurerotation mechanism which, at the presentstage, still produces audible noise.

The last cases shown concern the useof the M2 tip-tilt and chopping capabili-ties for interferometric observations in theNear IR (for atmospheric tip-tilt correc-tion) and thermal IR (for background sub-traction) respectively. It was originallyplanned to use smaller and lighter mir-rors in the coudé train to perform thesefunctions for VLTI because of the high

OPL stability requirements. Tests on theM2 at Dornier in November 1997 [4] hadshown, however, that the outstandingaxial stability of the M2 during tilt andchopping should be good enough to useit for VLTI. The results presented hereconfirm this preliminary conclusion. Tip-tilt correction and chopping can be donewith M2 for VLTI observation in the nearIR and thermal IR, respectively.

Other sub-systems were also positivelytested as to their impact on the OPL sta-bility such as active optics during a typi-cal correction, operation of louvers andwindscreens, fans and transformers inelectronic cabinets, etc. In this last cate-gory, it is worth mentioning the followinganecdote. The first set of measurementwas constantly showing a much-too-highvisibility loss of typically 40% in the visi-ble and 5% in the thermal IR. After exten-sive investigations, it was found to becaused by two cooling fans located in theelectronic cabinets of the Test Cameraattached at the Cassegrain focus. Con-trary to most of the other cabinets, theseare not vibration isolated due to the tem-porary nature of this first-light instrument.These small fans were able to excite the“pumping” mode of the 80-ton telescopetube at 40 Hz creating about 90 nm rmsover 10 msec and 500 nm rms over 290msec. This shows the importance of acareful design down to that level of detail.

Conclusion

Although these tests cannot be con-sidered as the final ones since severalmirrors were not yet installed (M3 andcoudé train), they confirm the very strongpotential of the VLT 8-m telescope to fulfilthe very stringent stability requirementimposed by VLTI. Indeed, they show thatthe global vibration of the overall struc-ture remains within an acceptable range.Any future problems which could appearshould be of a local nature (e.g. resonanceof a given coudé mirror cell) and there-fore more easy to solve by appropriatelocal damping or stiffening. These testsalso enabled us to identify, at an earlystage, a number of possible improve-ments such as better isolation of the cool-ing pumps in the enclosure basement,stiffening of the M4 arm and improvementof the enclosure rotation smoothness.

References[1] B. Koehler, “Hunting the bad vibes at Pa-

ranal”, The Messenger No. 76, June 1994.[2] Koehler et al., “Impact of the micro seismic

activity on the VLT Interferometer”, TheMessenger No. 79, March 1995.

[3] Koehler, “VLTI: Chasing the nanometric vi-brations on 8-m telescopes” in Proc. SpaceMicro-dynamics and accurate Control Sym-posium (SMACS2), CNES, Toulouse, May1997.

[4] Koehler, “VLT Unit Telescope, suitability forinterferometry: first results from acceptancetests on subsystems”, in Proc. SPIE Sym-posium on Astronomical Telescope and In-strumentation, Conf. 3350 “AstronomicalInterferometry”, Kona, March 1998.

Figure 1: Summaryof the results fromvibration test on UT1for VLTI application.The graph showsthe total visibilityloss due to mirrorvibrations insideUT1 for threeobserving wave-lengths (visible, nearIR, thermal IR) andvarious operationalconditions. The VLTIerror budgets ask fora 1% visibility lossdue to vibrationinside the telescope.This corresponds toan OPL variation of14, 50 and 215nanometers r.m.s. forthe three wave-lengths respectively.This requirement isachieved in most casesfor the IR ranges. For obser-vation in the visible it is slightlyexceeded but remains withinacceptable margin in view of theglobal visibility loss of 30–40%which is aimed at when all othererror sources and sub-systemsare included (atmospheric turbu-lence, figuring errors, polarisation,delay lines, etc.). A dominatingdisturbance for the visible is thevibration of the cooling pumpslocated in the basement of theenclosure for which improvementcan easily be achieved. OPERATIONAL CASES

VIS

IBIL

ITY

LO

SS

(%

)

13

T E L E S C O P E S A N D I N S T R U M E N T A T I O N

The Wide Field Imager for the 2.2-m MPG/ESOTelescope: a PreviewD. BAADE 1, K. MEISENHEIMER 2, O. IWERT 1, J. ALONSO 3, P. AMICO 1, TH. AUGUSTEIJN 3,J. BELETIC 1, H. BELLEMANN 2, W. BENESCH 2, H. BÖHM 2, H. BÖHNHARDT 3, S. DEIRIES 1,B. DELABRE 1, R. DONALDSON 1, CH. DUPUY 1, O. FRANKE 2, R. GERDES 1, R. GILMOZZI 1,B. GRIMM 2, N. HADDAD 3, G. HESS 1, H. KLEIN 2, R. LENZEN 2, J.-L. LIZON 1, D. MANCINI 4,N. MÜNCH 2, G. RAHMER1, J. REYES 1, E. ROBLEDO 3, A. SILBER1

1ESO, Garching2Max-Planck-Institut für Astronomie, Heidelberg3ESO, La Silla4Osservatorio Astronomico di Capodimonte, Naples

History

In November 1995, the La Silla 2000working group of the Scientific TechnicalCommittee (STC) as well as the Observ-ing Programmes Committee (OPC) iden-tified a very strong demand by the ESOcommunity for wide-field imaging (0.5–2degrees) capabilities (cf. Andersen, J.1996: The Messenger, No. 83, p. 48). Inall major areas of research, the primarydriver was the identification and pre-selection of candidate targets for morein-depth studies with the VLT (see also:Renzini, A. 1998: The Messenger, No. 91,

p. 54). The OPC report remarked: In anycase, an array of CCDs of 8000 × 8000will have to be constructed: This is feasi-ble but will not be a small undertaking.

The strong encouragement given toESO to investigate possibilities of imple-menting such a facility could neverthe-less not eliminate the fact that in ESO’smid-term planning hardly any resourceswere left that on the desirable short time-scale could have been assigned to a newproject. It was, therefore, timely that si-multaneously the Max-Planck-Institut fürAstronomie in Heidelberg (MPI-A) pro-posed to build a wide-angle camera for

the MPG/ESO 2.2-m telescope. Aftersome iterations on the general scope ofthe project, it was agreed that MPI-Awould be responsible for mechanics, op-tics, and filters whereas ESO would pro-vide the optical design, the complete de-tector system, and all of the control soft-ware. Later, the Osservatorio Astronom-ico di Capodimonte (Naples) joined theproject as the third partner and fortunatelywas able to absorb the lion’s share of thecost of the CCD detectors.

This article intends to give a concisepreview of the result of these efforts. Thecommissioning of the new camera will

Figure 1: A schematic overview of the WFI. The left and the outer right circle are the apertures of filter holders mounted on the filter storage ring.The second circle on the right-hand side shows a filter after it has been moved out of the ring. It can then be rotated through 90 degrees out of theimage plane, shifted to the left, and between the two triplets (shaded in green) inserted into the beam. The shutter is located right below the lowertriplet. Next follows the cryostat head seen from outside with a number of vacuum connectors. The cylinder at the bottom (with a variety ofattachments) is the liquid nitrogen tank. The large rectangle to the left is the FIERA control electronics box which also serves as a counterweightto balance the torque of the asymmetrically (with respect to the optical axis) located filter storage ring.

14

Figure 2: A view of the WFI filter exchange mechanism from the bottom and as assembled for first tests in spring 1998. The rectangle near thecentre is a mount for interference filters (a dummy is visible) rotated into a plane perpendicular to the optical axis but outside the telescope beam;the recess for the filter intersecting the light towards the tracker CCD is on the left. The grooves at the top of the picture are located on the filterstorage ring; there are fifty of them in total, and each can accommodate one filter. The rectangle in the upper right corner is a holder for the largercircular glass filters (with a dummy inserted) in its storage position. The filter ring can be rotated by means of the cogwheel in the upper left cornerand the attached motor.

Figure 3: Partial drawing of the cryostat head of the WFI with the plate and sockets on which the CCD’s are mounted. Two rows of four chips eachform the science mosaic which measures about 12.5 × 12.5 cm2. The extra CCD on the left front side is used for autoguiding.

15

take place in November and December1998 (unfortunately, at the time of writingwe were informed that a mishap duringthe manufacturing of one of the two tri-plets will probably produce this delay withrespect to the original schedule), and firstresults will be reported in the March 1999issue of The Messenger. The Wide FieldImager (WFI) will be the only instrumentoffered on the newly refurbished and up-graded 2.2-m telescope (cf. the report bythe 2.2-m Telescope Team in this issueof The Messenger).

Optics

The optical design is essentially theone of a focal reducer with two triplets (acrude cross section can be seen inFig. 1). It yields a scale of 0.24 arcsecper 15-µ pixel which over the field of viewof 0.5 × 0.5 degrees varies by less than0.1% (at 500 nm). From 350 nm to 1 µ80% of the encircled energy falls ontoa single pixel (except for the extremefield corners at long wavelengths). Thethroughput curve rises from 45% at 350nm to 80% at 400 nm and thereafter re-mains flat through 1 µm.

Mechanics

Apart from providing the necessarysupport structure for the optics and pro-tection against light and dust, the me-chanics features two motorised functions(focusing is accomplished by moving thesecondary mirror of the telescope). Thefirst one is a large, roughly semi-disk-shaped shutter. It is designed to reach aprecision of order 1 msec even for shortexposures. The second one is the filterexchange mechanism. Up to 50 filterscan be permanently mounted in a largerotating ring surrounding the camera (Fig.1) where they are stored in a vertical po-sition. An electromagnetic grabber canmove a filter out of the ring, rotate itthrough 90 degrees, and slide it into thebeam (Figs. 1 and 2).

Detector System

The focal plane is covered with a mo-saic of eight 2k × 4k CCD44 devicesfrom EEV. The pixel size of 15 µm match-es the optical design. Figure 3 providesa 3-D view of parts of the cryostat headwithin which the chips are mounted. Thegaps between the individual detectors(about 1.5 mm along the major and 0.8

mm along the minor axes) can be cov-ered by multiple exposures with smalltelescope offsets. This procedure simul-taneously allows cosmetic imperfectionsof the chips to be corrected for. Area-wise, the cosmetic defects are much lessrelevant than the inter-chip gaps. From350 nm to 400 nm, the sensitivity risesfrom about 50% to 80% or above andonly beyond 700 nm slowly decreases toslightly less than 30% at 900 nm; detec-tors and optics are not useful at wave-lengths longer than 1 µm. The readoutnoise will be around 6 e— per pixel, andthe dead time for readout and file trans-fer to the instrument control workstationbetween consecutive exposures amountsto half a minute. Attached to the detectorhead is a liquid nitrogen-filled tank of thesame type that is also used for manyVLT instruments. All CCD functions areprovided by ESO’s standard FIERA con-troller and associated software.

Autoguiding

Besides the 2-by-4 science mosaic ofCCD’s, Figure 3 shows a ninth CCD onthe western side (in right ascension). Thisso-called tracker CCD is of the same typeas the other eight but has slightly lowercosmetic quality. It shares the shutter withthe main detector array and is employedfor autoguiding. Only the central 50%along the minor axis are unvignetted bythe camera and the filter mount. But asufficiently bright guide star can be placedanywhere on the chip. This CCD is con-trolled by the same FIERA system butcompletely independently so. A small win-dow can be defined around the image ofthe star. After an integration, the durationof which is adjustable down to a fractionof a second, this window is shifted at arate of about 40,000 rows per second tothe readout register and then read at nor-mal speed. The FIERA software automat-ically detects the star in this window, fitsa Gaussian to it, and sends the results tothe Telescope Control System which usesthis information to let the telescope trackproperly.

Filters

In addition to glass filters, which incombination with the CCD sensitivitycurve closely resemble the B V RC IC Z+system (the U-band is broader than usu-al to achieve higher throughput), numer-ous intermediate- and narrow-band fil-

ters are foreseen. The complement ofthe latter will initially be incomplete buteventually cover the range from 370 nmto 930 nm in a quasi-continuous fashionsuitable for the identification of high-redshift objects. A sub-set especially de-signed for this purpose is similar to theone described by Thommes et al. (1997,in R. Schielicke (ed.), Reviews in Mod-ern Astronomy, Vol. 10, p. 297). A list ofthe filters available in Period 62 is pro-vided on the WWW page referenced be-low. Glass filters are circular and alsocover the tracker CCD. Medium- andnarrow-band filters are square and donot extend into the guide star beam. In-stead, a small separate glass filter ismounted in a corresponding recess inthe holder so that the throughput is suffi-cient for autoguiding. In order to reducedifferential atmospheric effects, the cen-tral wavelength approximates the one ofthe science filter.

Future Options

For Period 63, it is foreseen to test aset of linear polarisers. If the tests aresuccessful, the polarimetry option will beoffered in Period 64. Possibilities for slit-less low-resolution spectroscopy are alsobeing investigated.

The DAISY+ instrument and telescopecontrol software environment and thecontrol electronics of the 2.2-m telescopedo not support low-level compatibility withthe corresponding VLT systems. There-fore, a complete copy of the VLT DataFlow System (Silva, D., and Quinn, P.1997: The Messenger, No. 90, p. 12) can-not be installed. However, in preparationfor the VLT Survey Telescope (VST; TheMessenger, this issue) to be erected onParanal in the year 2001, which will havetwice the field of view as the WFI, effortsare being undertaken for a partial imple-mentation. The first step will be the de-velopment of a Phase 2 Proposal Prepa-ration (P2PP) system. This will also ena-ble suitable observing programmes to becarried out in service mode.

Updated information on the WFI willbe made accessible via the WWW homepage of the 2.2-m Telescope Team on LaSilla (URL: http://www.ls.eso.org/lasilla/Telescopes/2p2T/E2p2M/WFI/news/WFI_P63.html).

D. Baadedbaade.@eso.org

16

The editors of the La Silla News Page would like to welcome readers of the eleventh edition of a pagedevoted to reporting on technical updates and observational achievements at La Silla. We would like thispage to inform the astronomical community of changes made to telescopes, instruments, operations, and ofinstrumental performances that cannot be reported conveniently elsewhere. Contributions and inquiries to thispage from the community are most welcome. (J. Brewer, O. Hainaut, M. Kürster)

News from the NTTO.R. HAINAUT, ESO, La Silla

As the reader will notice, the “Newsfrom the NTT” are back in the La SillaNews Page, marking the end of the“Big Bang” era (this major upgrade hasbeen described in The Messenger Nos.75–91). The NTT is now fully returnedto the La Silla Observatory. With this,another era is finishing too: GautierMathys has left the Team. After 5 yearsat the NTT (i.e. since the beginning ofthe team itself), many of these as localrepresentative of the Team Leader, andthe last year as Team Leader, Gau-tier is now preparing the scientific op-eration of the VLT UT1 at Paranal. Hisexcellent leadership, and his exten-sive, boundless and all-encompassingknowledge of the NTT systems will bemissed by the Team. Since the 1st ofAugust, the author has taken over theduties of NTT Team Leader; he will aimat continuing Gautier’s work to improvethe reliability and user-friendliness ofthe Telescope, while maintaining thefull compatibility with the VLT environ-ments.

During the past months, SOFI, theNTT infrared spectro-imager, has re-ceived its first visiting astronomers. Theinstrument proved to be extremely effi-cient, as illustrated by the paper byChris Lidman in this issue of The Mes-

senger. Its “second generation” obser-vation templates, which make full us-age of the interactive capabilities of the“Real Time Display”, constitute an intui-tive and effective interface that allowsthe observer to efficiently master all themodes of this instrument.

After its commissioning in January,SUSI2 experienced a series of prob-lems, including loss of vacuum, some-times accompanied by sudden warmingup. These were caused by the rapidcontraction of the O-ring sealing thedewar, which happens when some LN2is spilled over them, e.g. when re-fillingthe instrument, or when moving it whenit is still full. This problem should besolved by the end of August, with theinstallation of a dewar with improvedO-rings and equipped with a devicelimiting the LN2 spilling. We should thenbe able to take full advantage of thisnew-generation dewar, capable of keep-ing the instrument cold for 48 hours.

A series of improvements of the sys-tem have also been implemented; a fewexamples and highlights follow:

• The CCD monitoring, which hadreceived no new developments sincethe departure of Griet van de Steene inJanuary 1998, has been taken over byVanessa Doublier. Our three CCDs are

now monitored weekly, and the resultsof these tests, including bias level,read-out noise, shutter delay and sensi-tivity, are presented on our instrumentWeb pages. We plan to continue imple-menting more tests into this monitoringof the detectors, as well as adapt it tothe SOFI IR array.

• The focus offsets between the Im-age Analysis cameras and the scientificdetectors has been measured. Theseoffsets have been found extremely sta-ble for SUSI2, while for EMMI and SOFIthey show some slight variations withthe rotator angle. It reflects the greatercomplexity of these instruments, whichare subject to minor internal flexures.The Active Optics system is now cali-brated to take these focus differencesinto account. As a consequence, thetelescope is automatically focused whileperforming an image analysis.

• Various monitoring and technicaltemplates are being developed to per-form operation and maintenance tasksin a more efficient way.

Finally, a point that will be of interestfor the observers: the new versions ofthe EMMI and SOFI manuals are un-dergoing their final revision and shouldbe available on the NTT Web pages bythe time these lines are printed.

SOFI Receives its First UsersC. LIDMAN, ESO, La Silla

SOFI, the recently commissioned IRimager and spectrograph on the NTT,started regular service on June 6 thisyear. Since then, about a dozen visitingastronomers have successfully used theinstrument.

All modes of the instrument, which in-cludes broad- and narrow-band imaging,low-resolution spectroscopy and imagingpolarimetry, have since been used.

To date, the instrument has been usedto study objects as varied as superno-

vae, proto-planetary nebulae, embeddedstars, dwarf galaxies, gravitational lens-es, high-redshift clusters and the star-formation rate at high redshifts.

Of particular note was the observa-tion of the well-known Einstein ring

The La Silla News Page

17

PKS1830-211 carried out by a team ofastronomers headed by FredericCourbin (Université de Liège). This teamincludes George Meylan from ESOGarching, Tom Broadhurst and BrendaFrye from the University of Californiaat Berkeley and the author of theselines.

The optical identification and the red-shift of PKS1830-211 has been longsought by astronomers. The optical-IRcounterpart of PKS1830-211 was dis-covered last year through images takenwith IRAC2b. The optical-IR colours sug-gested a significant amount of extinc-tion. This was the impetus to take aspectrum with SOFI.

The following plot shows a spectrumof PKS1830-211 spanning the wavelengthrange 1.5 to 2.5 microns. It was takenwith the red grism of SOFI. The exposuretime was 24 minutes and the source isnear 15th magnitude at K. In the plot, red-shifted Hα and Hβ are clearly detected atz = 2.507. Other lines, such as [OIII] mayalso be visible. The horizontal lines markregions where atmospheric absorption isstrong: the region near 1.9 microns andthe region beyond 2.5 microns are almosttotally opaque, so the data in these re-gions have been deleted.

This new result, together with theknown redshift of the lens (measured bymolecular absorption at mm wavelengths– F. Combes and T. Wicklind 1998, The

3.6-m Telescope Passes Major Upgrade MilestonesM. STERZIK, ESO

During two months of technical timein July and August, major steps in the 3.6-m telescope upgrade plan were success-fully passed. I shortly recall the objectivesfor the 3.6-m upgrade project (see TheMessenger 85, 1996, p. 9): (i) optimisa-tion of the mechanical and optical per-formance to improve the image quality(IQ), (ii) operational stability and efficien-cy to minimise downtime and maximisescientific return, and (iii) offer competitiveinstrumentation. After the upgrade, the3.6-m telescope will return to the forefrontof 4-m-class telescopes in the beginningof the next century.

All the work done during the lastmonths was in that direction. Thanks tothe careful project planning of UeliWeilenmann, all milestones foreseen inthe technical time could be passed. Amajor opto-mechanical improvementwas the successful installation of an ac-tive pressure control system for the M1lateral pad support. M1 movements in themirror cell are now practically eliminat-ed. Tests demonstrate that already inopen loop the force distribution onto thelateral mirror support can be controlled

at a level of, typically, 20 kg differencebetween theoretical and measured forc-es, and further reduced to 2 kg in closed-loop configuration. (With the old REOSCsystem, force differences of 300 kg weretypical.) This control is of crucial impor-tance for the IQ at larger zenith distanc-es. For the presentation of impressive IQresults, please refer to the ongoing se-ries by Stephane Guisard in The Mes-senger. At this moment, I rather wish tostress that already now sub-arcsec IQ isroutinely possible at the 3.6-m telescopefor scientific work (as long as the exter-nal seeing conditions allow). Considera-ble progress was also made in increas-ing the mechanical stability of the guideprobe, now allowing reproducible move-ments with an accuracy below 0.2 arc-sec. Here, thanks go to the La Silla me-chanics and optical support teams, whosolve many problems promptly and thor-oughly.

Another central issue related to tele-scope control software (TCS) was theinstallation and commissioning of theTCS under NOV97 VLT-Common Con-trol Software. This includes the worksta-

tion part of the telescope interface, andthe part related to the local control unitsof adapter functions. The conceptualcomplexity of the VLT software is wellknown, and it is obvious that adaptingthis software to the specific requirementsof the 3.6-m telescope is not straightfor-ward, and sometimes leads to hiccups.For example, a reliable interface of thefront-end VLT-software with the still op-erating HP1000-based TCS (which stillcontrols the telescope in the back-end,i.e. “moves” the telescope), is a demand-ing task and a potential source of prob-lems. It is the price to pay in the ap-proach taken for the 3.6-m upgrade: tooffer the telescope to the communitylargely in parallel with the upgrade. Andhere, I would like to express my grati-tude to the highly committed softwareteam at La Silla who successfully ac-complished this challenging task with lim-ited resources.

A part of the software upgrade is theimplementation of a fully VLT-compliantinstrument control software (ICS) forEFOSC2. This is the most striking changethat observers will experience when

Messenger 91, p. 29) and continuing ef-forts to get a secure time delay, means

that we are a step closer to the goal ofdetermining H0 from this lens.

18

working with the instrument starting thisSeptember. The old HP1000 control isreplaced by a GUI-based instrument con-trol panel (Fig. 1) that allows full controlover a multitude of EFOSC2 functionssuch as changing grisms, filters, and slits.It provides an elegant way to adjust thehalf-wave plate for polarimetry and allowsfull control over the CCD.

But it is not only the EFOSC2 cos-metics that have changed; new grismsoffer higher efficiencies, and a brand newFIERA CCD controller dramaticallyspeeds up read-out times, a bottleneckfor some programmes in the past (seeThe Messenger 83, 1996, p. 4). Mostimportant, the whole observing philoso-phy changes with the advent of the newICS: it converges to the observingmodes known from the NTT, and em-ployed at the VLT. Visiting astronomersobserving with EFOSC2 will be askedto prepare their observations (togetherwith the help of on-site support astrono-mers) in advance with the Phase 2 Pro-

posal Preparation (P2PP) tool. P2PP invisitor mode will support EFOSC2, andserves to combine so called observingtemplates (scripts that contain a prede-fined sequence of operations that con-trol the telescope, the instrument, andthe detector) in Observing Blocks (OB),which are minimal entities that describemeaningful scientific observations. TheEFOSC observing templates were de-veloped by the 3.6-m team (the NTT/EMMI templates certainly helped a lotto speed up their creation) within a veryshort time in July and August. Not allglitches could be removed during thecommissioning of the ICS, as the com-plex communication between all subsys-tems requires that we gain more experi-ence, especially with the response timesoccurring in the many possible configu-rations. I apologise for any inconven-ience that may be encountered, but al-ready now most observation pro-grammes will benefit from our effort. Themost common observing modes of

EFOSC2 (imaging and long-slit spectros-copy) are supported by templates. Theycan be planned and executed in a moreefficient way and are less error sensi-tive. We are interested in learning theresponses of the community, and intendto further improve and extend this serv-ice. Stay tuned and look up the WEB(http://www.ls.eso.org/lasilla/Telescopes/360cat/html/3p6VLT.html).

The operation of EFOSC2 is, con-ceptually, embedded in a more complexdata-handling concept, known as theData Flow System (DFS) from NTT/VLT, and described e.g. by D. Silva andP. Quinn (The Messenger 90, 1998, p.12). It includes a transfer chain incorpo-rating a reduction pipeline and archiv-ing tools. In the future, this functionalityis planned in the framework of a LaSilla wide data-handling and -archivingsystem. For the time being, astrono-mers will obtain their data on DAT tapesin a more conventional fashion at the3.6-m telescope.

Figure 1.

19

Let us now look at the near future ofthe 3.6-m telescope: more and moreTCS functions (interlock system, track-ing LCU, telescope servos) will be in-corporated into the new control system.The HP1000-based TCS will be fullyabandoned next April. Already this yearwe plan to move the control room (nextto the telescope on the cold observingfloor) to a spacier, more comfortableroom located on the third floor. Newfurniture will underpin the modern “lookand feel” when observing with a tele-scope having one of the most advancedcontrol systems. Regarding instrumen-tation, the CES will be the next awaitingVLT-compliant instrument control. Afibre-link to the Cassegrain-adapter ofthe 3.6-m telescope has already been

installed, and the final commissioningwill take place this October. Then EA-GAL (ESO And GSFC ALADDIN Cam-era), a new near IR camera for the 1–5µm wavelength range, and mainly fore-seen in conjunction with the ADONISadaptive optics system, and TIMMI2,the more sensitive successor of the oldThermal Infrared Multi Mode Instrument,will arrive. They will offer exciting andcomplementary facilities, especially im-portant to bridge the gap until the VLTgoes fully into operation. The 3.6-mtelescope will remain a cutting-edge tel-escope in its class, and will gain furtherimportance when science priorities likethe High Accuracy Radial velocity Plan-etary Search (HARPS) programme areconducted at this telescope.

The last few months saw major per-sonnel movements in the 3.6-m team.Roland Gredel, team leader of the 3.6-m+CAT Team since 1997, left La Silla inorder to assume responsibility as direc-tor of the Calar Alto observatory in Spain.On behalf of the 3.6-m Team, I wish himall the best in this new challenge. Hisfunction will be taken over by the authorof these lines.

At the same time, two new fellowsjoined the Team: Olivier Marco, now re-sponsible instrument scientist for ADON-IS, and Ferdinando Patat, who alreadyplayed a key role in producing observingtemplates for EFOSC2. He takes over theresponsibility as EFOSC2 instrument sci-entist. The 3.6-m Team welcomes its newastronomers.

2.2-m Telescope Upgrade StartedThe 2p2team, ESO, Chile

On 15 July 1998 the upgrade of theMPG/ESO 2.2-m Telescope was startedat La Silla. This project was launchedlate last year in order to:

• modernise telescope equipment,• replace worn-out parts and units

which malfunction frequently after beingin service for more than fifteen years,

• prepare the telescope for the recep-tion and operation of its future onlystandard scientific instrument, the WideField Imager (WFI), a half-degree imag-er equipped with an 8 × 8K CCD (seeseparate report on the WFI in this is-sue).

The goal is to run the telescope in amodern VME based control environmentwhich will allow the use of a VXWORKSbased telescope control system (TCS)and a simple interfacing to the WFIinstrument control environment. As abaseline, the 2.2-m TCS will follow theconcept of the TCS for the Danish 1.5-mtelescope, but will be considerably mod-ified and improved in order to supportthe autoguiding system of the WFI, theautomatic guide-star selection throughguide-star catalogues, the new absoluteencoders, and the modernised telescopesafety system. Since the WFI will finallybe operated in service mode, precau-tions in the TCS are made to interfaceaccordingly with the new instrument con-trol system DAISY+ which is an ad-vanced version of the existing La Sillainstrument control package DAISY (cur-rently in use at the Danish 1.5-m, theDutch 0.9-m, and also foreseen for theB&C and FEROS instruments at theESO 1.5-m).

During the past 8 months, La Sillaengineers, technicians and astronomersanalysed the status and health of thetelescope optics and electromechanics,

made the design of the new equipmentand programmes and prepared the hard-ware and software for the implementationof the upgrade.

While the telescope optics was foundto be of excellent intrinsic quality (opticalaberration of below 0.2 arcsec is rou-tinely measured during image analysistests at this telescope), the electrome-chanics and telescope control system(computers and software) needed a ma-jor overhaul and replacement. The me-chanics overhaul concentrates on thegearbox of the alpha drive (the worn-outalpha gear was replaced on 18 July1998), the hydraulics system, the instal-lation of new encoders (now also at thetelescope adapter/rotator unit). The tele-scope electronics will be based on VMEtechnology and it will receive a newdome control system as well as a distrib-uted system of environmental sensorsfor the registration of the temperatureand humidity at the telescope, instru-ment and inside/outside of the dome.Furthermore, a major clean-up of thewhole telescope cabling is foreseen. Onthe software side, the TCS is adapted tothe new logics of the telescope andinstrument control, while interfaces tothe DAISY+ software are added. Thecomputer platform for the telescope andinstrument control will be based onHewlett Packard (HP) workstations in-serted in a local network that supportshigh data transmission rates as neededfor the WFI (a single WFI image is about130 MB in FITS format and will be readout by the ESO FIERA CCD controller inabout 30 seconds). Beside the HP735workstation for the TCS, HPC200 andHPJ2240 workstations each equippedwith 108 GB disk drives and 35/70 GBDLT units will serve as data acquisition

and data reduction machines for theWFI and will provide support to the usersfor the on-line inspection through a real-time display (RTD) and for the on-lineanalysis by means of standard imageprocessing packages like MIDAS, IRAFand IDL. Last, but not least, the controlroom will be refurbished such that bothpeople and electronics will work in theenvironment as needed and most com-fortable for a successful operation.

The upgrade is underway: after thehardware modifications and installationsat the telescope, a test period of about 1month will start by the end of August1998 in order to tune and verify thetelescope optics and electromechanicsin the new control environment. Thereaf-ter, the telescope is – hopefully – readyfor the commissioning of the WFI whichwill arrive at La Silla in the last quarter of1998.

Near-infrared instrumentation is nowno longer offered at the 2.2-m: IRAC2,ESO’s near infrared array camera whichwas a workhorse instrument of the 2.2-m for many years, was decommissionedin mid-July 1998 (it is replaced by themore powerful SOFI instrument at theNTT).

The 2.2-m telescope upgrade teamconsists of: J. Alonso (project manager),J. Araya, T. Augusteijn, H. Boehnhardt,J. Brewer, R. Castillo, H. Kastowsky, F.Labraña, M. Mornhinweg, R. Olivares, F.Richardson, E. Robledo, A. Torrejon.

The following LSO teams and ESOpersons are supporting the project: LSOElectronics (R. Medina), LSO Mechan-ics (G. Ihle), LSO Optics (A. Gilliotte),LSO Software Support Group (G. Lun-dqvist), LSO Infrastructure Group (F.Luco), LSO Management (G. Andreoni,J. Melnick), ESO Garching (D. Baade).

20

T H E E S O A N D S T - E C F A R C H I V E S

The ESO Archives and the ST–ECF Archives have been closely integrated within the framework of theDMD, resulting in increased coherence and efficiency.

There is joint operation of the ESO and ST–ECF Archive while the development – which is necessarilyinstrument specific – is separate. In numerous instances, however, the procedure and software developed forspace data can be adapted to ground-based data and reciprocally (e.g. OTF or Associations).

This series of articles reports on recent advances.

ESO and ST-ECF Archive NewsB. PIRENNE 1, 2, M. ALBRECHT AND B. LEIBUNDGUT 1

1ESO/DMD, 2ESO/ST-ECF

During the past year, many changestook place at all levels in the area of dataarchive. A major structural re-organisa-tion, many new achievements in the areaof VLT readiness and quite a few impor-tant and interesting features for HST ar-chival data users have been implement-ed.

1 Preparing for the VLT

1.1 Re-organisation

A tighter integration of the operationalside of the ESO and ST-ECF archives hasnow been officialised even though thehard- and software systems used werealready common to both before. Now bothgroups take part in the HST and ESO NTTand VLT archive operations. ESO contin-ues to invest in manpower and systemsbenefiting both HST and ESO archives.The ST-ECF does the same. The newScience Archive Operations group (SAO)is formally integrated into the ESO DataManagement Division, Data Flow Oper-ation group. This group is headed (sinceJuly 1) by Bruno Leibundgut. Besides run-

ning the archive, this group will bear re-sponsibility for running the Quality Con-trol of the VLT data. The re-organisationaffected the entire DMD and the globalstrategy was described in the March 1997issue of The Messenger, p. 12, by D. Sil-va and P. Quinn.

2. New HST Archive Develop-ments

In accordance with the new orienta-tion of ST-ECF decided in 1996 at themid-term review, we have embarked inmany large projects aimed at adding val-ue to HST data, improving its access andhelping the HST users community. Someof these activities are described in thepapers by Alberto Micol and Markus Do-lenski in this issue of The Messenger(“HST Archive News: WFPC2 Associa-tions”, “HST Archive News: On-the-FlyRecalibration of NICMOS and STISData”).

Most important of all, the “WFPC2Associations” project results have nowbeen released. Thanks to the logicalgrouping of WFPC2 exposures, we can

offer our users not only a more meaning-ful catalogue browse response, but alsoautomatic cosmic-ray rejection, drizzlingand assembling of multiple exposuresupon retrieval of the data.

This new service complements thenow standard on-the-fly re-calibration(OTF), presented in previous issues ofthe ST-ECF Newsletter. Improvements inthis area now include STIS and NICMOS,which are available as presented in Ta-ble 1.

The completion of the WFPC2 asso-ciations’ project required an additionaldevelopment and operational effort,which has now reached completion: thegeneration of the missing jitter files. Theyrepresent the basic building block for ourclassification of the WFPC2 exposures,but unfortunately, 10 months of such datawere missing, which we have been ableto regenerate thanks to a fruitful collabo-ration with the STScI.

Presentation of information has alsoreceived some attention: New Java “ap-plets” to access the “PreView” images ofour archives from a web browser havebeen made available. They are described

TABLE 1: Services available and their pointers.

Service Name Type of service Description URL

GSC-I Catalogue The 20-million objects Guide Star http://archive.eso.org/gsc/gscCatalogue

USNO Catalogue US Naval Observatory Catalogue http://archive.eso.org/sky(500 million objects) cat/servers/usnoa/

DSS-1 Survey 1st generation Digital Sky Survey http://archive.eso.org/dss/dssDSS-2 Survey 2d generation DSS (higher/res). http://archive.eso.org/dss/dss

Not yet completeHST Data Archive Hubble Space Telescope http://archive.eso.org/wdb/

archive (with PreView, on-the-fly wdb/hst/science/formre-calibration and WFPC2associations’ capabilities).

NTT-old Data Archive Pre-Big bang NTT data archive and http://archive.eso.org/wdb/catalogue. Should be integrated wdb/eso/eso_archive/formin the new NTT/VLT archive soon

VLT/NTT Data Archive New VLT and post-Big Bang. http://archive.eso.org/wdb/NTT archive and catalogue wdb/eso/observations/form

21

TABLE 1: Estimated data rates (GB/night) expected from VLTinstruments over the next 4 years.

1999 2000 2001 2002

UT1 ISAAC 4 4 4 4FORS1 0.5 0.5SINFONI 0.5 0.5CONICA/NAOS 1.5 1.5 1.5CONICA (SPECKLE) 40 40 40

UT2 TESTCAM 0.5 0.5UVES 2.5 2.5 2.5 2.5FORS2 0.5 0.5 0.5 0.5FLAMES 2 2

UT3 TESTCAM 0.5 0.5VIMOS 20 20 20VISIR 1 1

UT4 TESTCAM 0.5 0.5FORS1 0.5 0.5NIRMOS 48 48CRIRES 0.5

VST WFI 3.8 15

VLT TYPICAL MIX 3.0 19.1 59.3 70.6(GB/NIGHT)

in more details in Markus Dolensky’s ar-ticle on “HST archive services implement-ed in Java” in this issue.

3. Other Archive Developments

3.1 GSC-II participation

In the framework of the ESA-NASAMOU renewal on HST, ESA is contribut-ing to the completion of the Guide StarCatalogue II planned as a 2-billion ob-ject, fully homogeneous (both photomet-rically and astrometrically), all-sky, multi-colour catalogue. The ST-ECF is involvedin this major scientific endeavour by op-erating a pipeline that extracts the ob-jects and by doing the quality control ofabout half of the 6000 photographicplates used for the generation of thecatalogue. More details concerning theparticipation of the ST-ECF in this projectare given in the article entitled “ST-ECFParticipation in the GSC-II GenerationProject” in this issue.

On the development side, the ESOarchive is also contributing a storagemethod for the future export catalogue:the system will allow the storage of theentire catalogue on less than 50 GB ofdisk space.

3.2 Archive storage media change

The ESO/ST-ECF archive is now stud-ying the promise of the DVD-R (DigitalVersatile Disk) for astronomical data stor-age. The new capacity needs generatedby the VLT instruments and the surveytelescopes are prompting us to look intonew denser yet affordable storage tech-nologies. The prospects of the DVD arepresented in “Using DVD Technology forArchiving Astronomical Data” in this is-sue of The Messenger.

4. Services Available from theESO/ST-ECF Archive

Among the new services availablefrom our archive, it should be noted that

the post-Big Bang NTT archive data arenow public. We would appreciate feed-back on its usage. As a reminder, asthis issue will be distributed, work onthe preparation of the VLT Science Ver-ification (SV) data will be continued. Thedata will be made available to astron-omers from the ESO community as ear-ly as October 1. The plan is to have oneset of the CD-ROM containing the SVdata set sent to all the member state’sastronomical institutions or universitydepartments.

All the other services available fromthe ESO archive world-wide web are list-ed in Table 1 with their category, descrip-tion and URL. For the possibility to useother, non-interactive client programmesfor some of these services, please con-tact the authors.

B. Pirennebpirenne@eso.org

The VLT Data VolumeM. ALBRECHT, ESO/DMD

The VLT will be a remarkable observ-ing facility in many ways. Among others,the volume of data generated by its in-struments will make the VLT Science Ar-chive one of the largest data sources inastronomy. Table 1 summarises the esti-mated data rates (Gigabytes per night)

expected from VLT instruments over thenext 4 years. In estimating the data out-put of a given instrument, assumptionshave been made on typical usage modes,e.g. infrared instruments would producelarger numbers of frames because of com-monly used sky/object observing se-

quences. Also, for each telescope a typi-cal mixture of usage of alternative fociihas been estimated in order to obtain atotal volume expected from the completefacility. MIDI and AMBER, the two firstVLTI instruments (to see first light in 2001),could produce of the order of 40 GB raw

Figure 1: Cumulative data volume of the VLT Science Archive over thenext 5 years.

22

data per night. They are not yet includedin the table due to the uncertainties in theirmodes of operation. (The last line of thetable includes VST output.)

When seen in the perspective of thecumulative data volume, these figures

reveal the true dimensions of the facility.Figure 1 shows the estimated amount ofdata flowing into the VLT Science Archivein the years to come. For comparison,the total volume of the HST Science Ar-chive after 8 years of operations is about

half a terabyte in size (see article on “HSTOn-the-fly recalibration” by Micol et al. inthis volume).

M. Albrechtmalbrech@eso.org

Using DVD Technology for Archiving Astronomical DataB. PIRENNE1, 2, M. ALBRECHT1

1ESO/DMD, 2ESO/ST-ECF

Background

Due to the slow evolution of some as-trophysical phenomena, long-term pres-ervation of observations has always beena major concern of observatories aroundthe world. Be it hand-drawings on paperor 19th-century glass-plate photographs,the issue at stake is how to best preservedata for the future generations.

The advent of digital imaging and re-cording equipment in the second half ofthis century has provided both more ob-servations and denser data storage me-dia. These media can therefore no long-er be read by the human eye. Moreover,with no immediate readability to the un-aided eye, digital recordings require spe-cific equipment to decipher their con-tent. If a lot of progress has been madein the past decades to manufacture long-lived data storage media, the same isnot true for the reading/writing equip-ment, quickly reaching obsolescence,and the repair of which is rapidly be-coming impossible. This apparent con-tradiction between durable media andtransient reading equipment is easy tounderstand if one realises that the me-dia is usually “passive” whereas thereading device is always active, withmechanical components.

Archivists must therefore reconsiderstorage technology every few years: atransposition of the archive content fromendangered media to the newest tech-nology has to be undertaken almostevery three to five years. Another majorfactor pushing towards migration of datato new technology is costs: the cost ofthe new technology compared to theold one often brings savings per unit ofvolume of up to an order of magnitudeand are a strong motivation for migra-tion.

Current Situation

In the case of the ESO/HST ScienceArchive in Garching, since 1988, threedifferent storage media have been usedand migrated from/to: The 2GB LMSI 12″Optical disk, the 6.4 GB Sony 12″ opticaldisk and the current 0.64 GB CD-R in jukeboxes. The reasons for migrating from

TABLE 1: The various data storage technologies used so far at the ESO/ST-ECF ArchiveFacility. Shaded areas represent the solutions actually implemented.. Units of cost representan arbitrary monetary unit set to 100 per GB for the most expensive solution.

Medium Name Reason for choice/migration to Cost per GB Cost per GB

without Juke box with Juke box

2GB/vol LMSI 12″ Direct access, best of technology 17 100optical disk back then. In sync with ST ScI

and HST archive

6.4GB/vol Sony 12″ Direct access, factor of 2–3 8 34optical disk cheaper to operate, previous

technology difficult to maintain.In sync with ST ScI and HSTarchive

0.6GB/vol 5 1/4″ Jukebox allows for online, no- 0.6 7.8CD-R operator-required access, ISO

standard for file system

4.0GB/vol 5 1/4″ Much higher density, keeps direct 2.2 2.8DVD-R access advantage of CD-ROMs

one to the other are given in Table 1 be-low.

We abandoned the 12″ optical disk infavour of the more common 5 1/4″ CD-Rfor two reasons. On the one hand, CD-Rwere enjoying an international standarddefining the way their content should belaid out (ISO 9660). This was a guaran-tee of durability and multi-vendor support.On the other hand, the possibility of hav-ing all the data on-line in juke boxes wasfinally an affordable possibility as the costof juke boxes for 12″ optical disks wasprohibitive for our archive system (seeright-most column of Table 1). This lastreason sealed the fate of hardware com-patibility with the HST archive at the STScI where the data are still on 12″OD injukeboxes.

However, now that we have complet-ed the migration to CD-R, we are facedwith another concern: the data rategrowth. The VLT and HST instruments,soon to be commissioned, will produceseveral TB worth of raw data per year.We could not practically keep these datausing CD-ROMs in juke boxes withoutmaking major infrastructure investmentin storage buildings!

The solution that addresses the den-sity problem and keeps the advantage ofthe CD-R technology (direct access me-dium, cheap juke box capability) is theDVD.

Digital Versatile Disk (DVD)

The DVD technology has been verylong to come, heralded as it was by thespecialised press for a number of yearsalready. However, various disagreementswithin the industry and disputes aroundcopyright issues have considerablyslowed the introduction of this technolo-gy. A few months ago, however, equip-ment to record one’s own media (theDVD-R (see Table 2 for a brief descrip-tion of the variants) became available.Our archive facility was understandablyquick to procure and test the equipmentand prepare the necessary software tosupport the device (from Pioneer Corp.).Even now, little support is available. TheDVD-R can only be called such if its filesystem is compliant with the UDF filesystem. However, software drivers tosupport this format for both read andwrite are hardly available. To our knowl-

23

edge only the latest version of the Ma-cOS operating system has genuine sup-port for it. The Unix world so far enjoysno support.

In order to obtain quick results and tobe as compatible as possible with the ex-isting archive tools and procedures weare using, we took a pragmatic approach:we contacted the developer of a public-domain CD-R recording tool “cdrecord”(a popular Linux tool, see below) and ar-ranged with him to extend his softwarefor the production of DVD-Rs as well.Within a few months, a workable systemwas delivered to us. However, due to thelack of software support for the DVD na-tive UDF file system, we are using thestandard CD-ROM format (650MBISO9660) extended to 4GB. To the hostcomputer, our “DVD-R” once written sim-ply looks like an unusually large CD-ROM.

Projects and Schedules

The most pressing and demandingproject in our archive for high-densitystorage media at the moment is the fu-ture 2.2-m telescope mosaic camera thatwill be commissioned in La Silla startingthis October. If our tests and prototypes,together with juke box support are pos-itive, the DVD technology will be thesystem of choice for this particular ar-chive. Also, we have started to migratethe NTT archive from the current Sony12″ optical disks to DVD. By the timethis issue of The Messenger is distribut-ed, we will have copied a few dozenSony 12″ optical disks onto the newmedium.

We still expect to have full UDF sup-port later in 1999. Our current experienceshows that the computer operating sys-tem will probably transparently identify

and mount media using any of the stand-ards. So the co-existence in the samejukebox of CD-R, DVD-R with ISO9660and plain DVD-R with UDF should be noproblem.

The next step, in 1999 or 2000 will bethe gradual migration of our CD-Rs ontothe new medium to save jukebox stor-age space, as this is by far still the larg-est part of the storage cost of CD-Rs.

For more information about thissystem, please contact the authors(bpirenne@eso.org ormalbrech@eso.org ). Information about“cdrecord” can be obtained from JörgSchilling (schilling@fokus.gmd.de ).The DVD-R recording device we are us-ing is Pioneer model DVR-S101.

References

Pirenne, B., “Data Storage Technology: cop-ing with the evolution”, Invited review pa-per for the IAU Symposium 161 on Wide-Field imaging, Potsdam, Germany, August1993, Kluwer Academic Publishers p. 339,1994.

Russo, G., Russo, S., Pirenne, B., “An Oper-ating System Independent WORM ArchivalSystem”, in Software – Practice and Expe-rience, 25(5), 521–531, May 1995.

Albrecht, M., Péron, M., Pirenne, B., “Buildingthe archive facility of the ESO Very LargeTelescope”, in Information & On-Line Datain Astronomy, D. Egret and M.A. Albrechteds., 1995, Kluwer, p. 57.

Pirenne, B., Durand, D., “Data Storage Tech-nology for Astronomy”, in Information & On-Line Data in Astronomy, D. Egret and M. A.Albrecht eds., 1995, Kluwer, p. 243.

TABLE 2: The Jungle with acronyms

Acronym Meaning Description

CD Compact Disk Mass produced (Audio) CDCD-ROM Compact Disk - Read-Only Memory Mass-produced (silver) CD-ROMCD-R CD Recordable Write-once, read-many CDCD-RW CD ReWriteable Re writeable CD-ROMDVD Digital Versatile Disk Mass-reproduced Video mediumDVD-ROM DVD Read-only Memory Mass-reproduced data disk 4.7,

9.4, 18.8 GB.DVD-R DVD Recordable recordable DVD (3.95GB)DVD-RAM DVD random access memory re-writeable DVD (2.6 GB)DVD-RW DVD re-writeable re-writeable DVD (??)

How the Analysis of HST Engineering TelemetrySupports the WFPC2 Association Project andEnhances FOS Calibration AccuracyM. DOLENSKY, A. MICOL, B. PIRENNE, M. ROSA, ST-ECF

Introduction

The analysis of Hubble Space Tele-scope (HST) engineering telemetry atSTScI is a process that evolved over timesince launch in April 1990. Today a jitterfile is computed for every dataset by asystem called Observatory MonitoringSystem (OMS). The jitter files, producedto study the telescope pointing stabilityand the trends in the telescope/instru-ment performance within the orbital en-vironment, are available for datasets tak-en after October 1994. These files aresupposed to contain sufficient informa-tion for an astronomer to properly reducescientific data.

This has two implications:1. Jitter files for datasets before Octo-

ber 1994 are either missing or were com-puted differently.

2. Engineering parameters that are notpart of the jitter files cannot be retrievedfrom the HST archive.

This article shows what kind of prob-lems this can cause, and more important-ly, how these problems were solved incase of the WFPC2 Association Projectas well as the FOS Post Operational Ar-chive.

Jitter and WFPC2 Associations

ST-ECF embarked on a project aim-ing at grouping, cosmic-ray cleaning anddrizzling images taken by WFPC2 (WideField Planetary Camera). Therefore, veryprecise pointing information is required.

The jitter files proved to be the most reli-able source of pointing information, witha relative accuracy between two expo-sures in the same HST visit of about 0.01arcsec (A. Micol et al. in this issue, relat-ed web page \r \* MERGEFORMAT [6]).Furthermore, possible pointing instabili-ties of HST during an observation canbe assessed which are sometimes lead-ing to evident perturbation of the PSF(Fig. 2).

ECF’s archive interface includes aJava applet that draws X/Y plots of anytwo columns of a jitter file as seen in Fig-ure 2. This can be done interactivelythrough a common web browser (M. Do-lensky et al. in this issue [5]).

Since WFPC2 replaced WFPC inDecember 1993, it was necessary to

24

set up a pipeline at STScI to computethe missing jitter files for the time pe-riod December 1993 – October 1994.This pipeline (Fig. 1) was operated re-motely from ECF and generated jitterfiles for additional 16,000 data sets ofall HST instruments. The engineeringtelemetry came from the Data Archiveand Distribution System (DADS) atSTScI and the historical spacecraftconfiguration was reconstructed using

the Project Database (PDB). The re-sulting jitter files were not only ingest-ed into the ESO/ECF archive but werealso shipped on CD-ROMs to the STS-cI and the Canadian Astronomical DataCenter (CADC).

As a next step, the remote pipeline atSTScI was enhanced, so that it is forthe first time possible to compute thespacecraft jitter for the whole lifetime ofHST in a homogeneous way.

Engineering Telemetry and FOSPost Operational Archive

A different project that requires cer-tain extra parameters of HST’s engineer-ing telemetry is the Faint Object Spec-trometer (FOS) Post Operational Ar-chive. Although FOS might be consid-ered an oldie (it was replaced by STISin February 1997) it’s data are still pre-cious and it will prove again the feasibil-ity of the concept of predictive calibra-tion that was jointly developed at ECFand ESO (M. Rosa).

There is, however, a small set of cru-cial input parameters missing for the re-calibration, namely the magnetometerread outs of the on-board magnetome-ters. These magnetometer readings arerequired to estimate the particle-inducedbackground count rate in the FOS digi-cons to scale geomagnetic shieldingmodels. This became an issue, sincethere is a problem with the magneticshielding of FOS, which was detectedonly after launch.

The solution was to prepare a soft-ware with the support of STScI that ex-tracts the required telemetry values forthe re-calibration. The extraction fromthe Astrometry and Engineering DataProcessing (AEDP) subset files is cur-rently done at ECF.

Outlook

In future, the Control Center System(CCS), currently developed at Goddardand STScI together with Lockheed, willinclude a data warehouse for direct ac-cess to historical telemetry. This will makeit much easier to study the impact of cer-tain parameters on scientific data.

Acknowledgements

The analysis of HST engineering te-lemetry was only possible with the inval-uable help of the DP-, OPUS- and PDB-Teams at STScI. For his collaboration inthe early stages of the project manythanks to J. Paul of MPE.

References

[1] Leitherer, C. et al. 1995, HST Data Hand-book, STScI, 107–119.

[2] Lupie, O., Toth, B. Lallo M., 1997, Obser-vation Logs, STScI.

[3] Rosa, M., ECF Newsletter 24, ST-ECF, 14–16.

[4] Micol, A., Pirenne, B., HST Archive News:WFPC2 Associations, in this issue.

[5] Dolensky, M., Micol, A., Pirenne, B., HSTArchive Services implemented in Java, inthis issue.

[6] http://archive.eso.org/archive/hst/wfpc2_asn/

[7] http://archive.eso.org/archive/jplot.html

Figure 1: Jitter file generation pipeline remotely operated from ECF.

Figure 2: The plot shows how the telescope was moving relative to the guide stars during anexposure. For some reason, in the middle of this observation, HST’s pointing moved by p25marcsec, which corresponds to half a pixel of its planetary camera.

M. Dolenskimdolensk@eso.org

25

ST-ECF Participation in the GSC-II Generation ProjectB. PIRENNE, ST-ECF; B. MCLEAN, B. LASKER, STScI

1. Background and Rationale forthe GSC-II Project

There are a number of motivationsfor scanning and cataloguing photo-graphic plates: first of all, the release ofa 2-billion-object catalogue would gen-erate uncountable science and mission-support projects which would be morethan welcome for HST and future spacemissions as well as ground-based tele-scopes. The significance of this projecthas already been acknowledged with theformation of an international consortiumto proceed with its construction with cur-rently available resources. These part-ners include STScI, CRA (Osservatoriodi Torino), ESO and GEMINI. Additionalsupport for the digitisation of the plates(DSS-II) is also being received fromESO, CDS, CADC, NAOJ and the AAO.

The HST Advanced Camera (ACS),poised to be launched at the end of 1999,has in particular “bright” object avoid-ance scheduling constraints. The SpaceTelescope Science Institute is currentlyperforming similar bright object checksfor observations using the STIS MAMAdetectors, but this is being done manu-ally, and is laborious and time consum-ing. These could be greatly simplifiedand automated if a homogeneous, deep,complete catalogue down to magnitude18 were available. Whilst the final re-quirements for the operation of the fu-ture HST COS (the selected ServicingMission SM-4 replacement instrumentscheduled for 2002) are still being de-termined, it is known that it has a small2″ acquisition aperture which may beaffected by the proper motion distribu-tion of the guide stars. Hence the avail-ability of guide star motions, or even thepositions from the second epoch plates,can simplify the process and reduce theacquisition failure rate. The plannedGSC-II would meet all these require-ments but its currently scheduled deliv-ery time of 12/2000 will not meet thecurrent ACS launch constraint. We feelthat the GSC-II is one essential compo-nent in reducing the operations cost ofHST.

In order to expedite GSC-II availabili-ty, we have put together an action planwith extra resources contributed to theproject so that a preliminary deep all-sky catalogue (without proper motions)can be completed around Q4 1999 tosupport bright-object protection checks.

2. Generation of the GSC-II

The generation of the GSC-II cata-logue consists of a number of steps, sim-ple in concept, but heavy in operationalburden.

2.1 Scanning

2.1.1 Scan all 3576 plates from the sec-ond-generation Schmidt surveys

(POSS-II in the Northern Hemisphereand the AAO-SES in the south). The sam-pling is 15 micron (1 arcsec) which re-sults in 23,000 pixel square raster imag-es of 1.1 GB byte each. This representsalmost 4 Terabytes of data in total. Thisfirst step has been done at the SpaceTelescope Science Institute and is nowvirtually complete. A lightly compressedversion is now being distributed to a smallset of selected sites, including ESO(Garching), where the scanned plates areavailable via the WWW: http://archive.eso.org/dss/dss

2.1.2 Combination of old and new platescans

Combine with the first-generation sur-vey plates that STScI has alreadyscanned with a resolution of 25 microns(1.7 arcsec pixels) for the GSC-I andDSS-I projects. [Note that eventuallySTScI will re-scan some of these platesat 15 microns but this is of lower priority].These 2390 plates represent another 1TBof data.

2.1.3 Scan old plates

Scan at lower priority the 894 POSS-IO plates that were not used in the GSC-I project.

2.2 Plate Processing

The second step involves a pipelineprocessing on single digitised plateswhere object detection (with de-blending)and preliminary calibration is taking placeon-line. Results are stored in an object-oriented database management systemwhich will reside on an HSM (Hierarchi-cal Storage Management) controlledmass-storage system. Quality controlhappens throughout this phase as well.This is the most resource intensive partof the production work: The massiveamounts of scan data are retrieved fromthe CASB image archive, processed ex-tensively and the results saved and load-ed into the COMPASS database. Esti-mates for the size of the database rangefrom 4 to 8TB.

2.3 Catalogue Construction

2.3.1 Analysis of the calibrated objectparameters

This step involves an analysis of thecalibrated object parameters in the data-

base to quantify the systematic errors andto recalibrate the derived parameterswithout a major reprocessing of the orig-inal scan data.

2.3.2 Export Catalogue

Derive exported catalogue from thelarge object database after merging over-lapping plates and plates of the same fieldbut of different colours. The new exportcatalogue will be compatible with theESO SKYCAT interface (see http://www.archive.org/ skycat), and many oth-er (web) interfaces can be expected tobe available to retrieve catalogue data.The storage possibilities of a large tablehave been examined in [Wicenec 1996].

3. Participation of the ST-ECF

The ST-ECF has extensive experiencein manipulating and processing large vol-umes of scientific data. It has strong con-nections with the STScI thanks to its HSTsupport mandate in Europe. In addition,the location of the group inside ESO (an-other patron of the GSC-II project) as wellas its proximity to the other Europeanpartner has motivated the ST-ECF to pro-vide direct help to the existing collabora-tion.

At this stage in the project, the soft-ware development is essentially over.Production has already started in March.The best use of ST-ECF resources is toassist with the operation of the massivepipeline processing in order to acceler-ate the availability of the data for HSToperations after Servicing Mission 3.

3.1 Interfaces

The geographical separation of thethree sites where processing is takingplace (STScI, ST-ECF and Osservatoriodi Torino) implies the definition of reli-able data transfer interfaces. The volumeof data involved immediately rules out anyon-line electronic data transfer of theimage data using such means as FTP.Therefore, airmail shipment of media withthe actual data to be processed and theirresults has been organised.

Procedures for problem reporting andbug fixes have been set up in such a waythat the pipeline does not stay idle for longperiods of time. Proper training of theoperation and quality control staff tookplace.

3.2 Resources required

In order to meet the target date for apreliminary single-epoch catalogue byHST SM-3, it is necessary to approxi-mately double the plate-processing rate

26

HST Archive Services Implemented in JavaM. DOLENSKY, A. MICOL, B. PIRENNE, ST-ECF

Abstract

In order to facilitate archive data se-lection and basic data analysis, a numberof Java Applets only requiring a commonweb browser are now complementing theHST Archive [3].

This article discusses various appletswhich are already part of the archive webinterface. These applets display andmanipulate FITS images as well as spec-tra taken with HST. A generic plot utilityis also used to present a set of pointingand specialised engineering data, calledjitter files [4].

Spectral

This applet is a previewer for HSTspectra. It is integrated into the WDBweb interface (Fig. 1) and offers variousoptions to inspect spectra with the mouseand by means of hotkeys. The screen-shots in this paper show, that a stand-ard web browser like Netscape 3 or In-ternet Explorer 3 is sufficient to run thisapplet. Micol et al. (1996) [1] discussedthis issue in more detail.

Java Image Preview Application(JIPA)

While Spectral presents plots of spec-tra, JIPA’s task is to visualise FITS imag-es, i.e. HST preview image collection, andto allow basic image manipulation. Theinput data format is compressed FITS.There are several options for contrastenhancement, zooming and displayingheader keywords (Fig. 2). Another fea-ture is the conversion of mouse locationsfrom pixel space to RA and Dec. JIPA iswritten in pure Java like the other appletspresented in this article and thereforeplatform independent.

that STScI had previously planned. Thisimplies that ST-ECF has effectively du-plicated the hardware and manpower thatSTScI currently has dedicated to the op-eration of the plate pipeline.

If the above production rate can besustained, it means a nominal productiontime of about one year. To this, a largeamount of time for manual re-processingof plates failing the processing step forone reason or another and extra qualitycontrol/bug fixes has to be added.

The pipeline software currently runson DEC Alpha Open VMS systems andis composed of C, C++, FORTRAN andIDL code. Therefore, the ST-ECF hasacquired an up-to-date DEC Alpha serv-

er 1200 with 2 × 500 MHz processors, 1GB of memory and about 100 GB ofdisk space. A large amount of cassettesfor the data transfer has also been ac-quired.

Operating the pipeline requires man-power to do operations such as loadingthe plate tapes into readers, write theoutput on media but also – and more im-portantly – perform quality control of theobject extraction results. The quality con-trol flags bad/doubtful plates, which areforwarded to the science team in Torinofor investigation. For this resource, twofull-time employees have joined the ST-ECF archive for two years (Nathalie Four-niol, previously in Strasbourg and Rob-

erto Mignani formerly at the Max- PlankInstitut für Extraterrestrische Physik).

4. Conclusion

In being involved in the GSC-II project,the ST-ECF is actively taking part in oneof the major astronomy achievements ofthe decade. In a two-year effort, our con-tribution will bring a more timely deliveryof the (first version) of an all-sky, 2-bil-lion-object catalogue complete to beyondmagnitude 18, available just in time forthe next millennium.

B. Pirennebpirenne@eso.org

Figure 1.

JPlot

JPlot was developed to support theWFPC2 Association Project [2]. It’s orig-

inal task is quality control of HST obser-vation log files (= jitter files). In the mean-time it became an integral part of the webinterface. It visualises ASCII tables and

27

displays them interactively as X/Y plots(Fig. 2).

The rather complicated back-end ofthis utility retrieves a FITS table from thearchive, extracts the requested columnsusing IRAF and puts them into a cachearea. The cache provides acceleratedaccess when the same information is re-quested repeatedly by web users within24 hours.

Outlook

Activities in the near future will in-clude:

• JDBC interface to SQL server (re-placing CGI scripts)

• collaboration with STScI in the fieldof enhanced web interface

• adding advanced features to over-come browser incompatibilities, like au-tomatic updates

Acknowledgements

The initial implementation of JIPA wasdone by contractor E.C. Downey. Anumber of features were added later onby ECF staff.

References

[1] Micol, A., Albrecht, R., & Pirenne, B. 1996,in ASP Conf. Ser., Vol. 125, ADASS 96, ed.Gareth Hunt & H.E. Payne, 104–107.

[2] Micol, A., Pirenne, B., & Bristow, P. 1997,in ASP Conf. Ser., ADASS 97, in press.

[3] http://archive.eso.org/wdb/wdb/hst/science/form

[4] http://archive.eso.org/archive/jplot.html

M. Dolenskimdolensk@eso.org

HST Archive News: On the Fly Recalibration (OTF)of NICMOS and STIS DataA. MICOL1, D. DURAND 2, S. GAUDET 2, B. PIRENNE 3

1ESA/ST-ECF, 2Canadian Astronomy Data Centre, 3ESO/ST-ECF & DMD

Introduction

HST science data are automaticallycalibrated when they are received at STS-cI (Space Telescope Science Institute)and these calibrated data are included inBaltimore’s archive. The calibration soft-ware, which is contained in the IRAF/STSDAS hst_calib package, takes as in-put the raw data and any necessary cal-ibration reference images or tables if theyare already available [2]. The softwaredetermines which calibration steps to per-

form by checking the values of the cali-bration switches in the header. It selectswhich reference files to use in the cali-bration by examining the reference filekeywords. The values of these switchesand keywords depend upon the exactconfiguration of the instrument, the dateof the observation and any other con-straints. The values are set in the head-ers of the raw data in OPUS (OSS-PODPS Unify System).

Until the end of 1995, when users re-quested calibrated data from the HST

archive, they received the data producedby the OPUS pipeline. However, with in-strumental properties changing with time,better approaches for calibration ofsome instruments have been introducedand there have been other general im-provements to the calibration of HSTdata. So what is a user to do? Fortunate-ly, the same, or improved, software thatruns in the calibration pipeline at STScIis also available in the released versionof IRAF/STSDAS. One can recalibratedata from the archive by starting with the

Figure 2.

28

raw data, editing the appropriate headerkeywords to reflect the new calibrationfiles and running the appropriate soft-ware. The STScI maintains a databasethat contains the recommended calibra-tion reference files for each observation.However, this is not the most convenientapproach for users, and this led us todevelop an automatic recalibration proc-ess for HST data that essentially dupli-cates what a user would do manually. Theon-the-fly re-calibration of the first gen-eration of HST instruments was devel-oped and introduced in the CADC andST-ECF archive at the end of 1995 [1].

Implementation

In 1997, two new instruments were puton board of HST. During the last year,CADC and ST-ECF worked together onthe extension of the OTF calibration pipe-line to support NICMOS and STIS. Thelong developing period is due to the factthat the initial life of a new instrument isalways somewhat difficult due to teeth-ing problems: instrument description key-words are found to be missing, or wrong-ly populated, the calibration softwaremust be revised to consider changes inthe instrument responses compared tothe ground tests, calibration referencefiles are not immediately available, etc.All those stabilisation problems led us toactually offer the OTF pipeline only about1 year after the Servicing Mission.

A decision was taken to not rely on theheader of the files, but instead to retrievethe calibration keywords form the HSTdatabase. While the keywords in the filescannot change anymore, the HST data-base can be kept up-to-date and key-words values corrected. Therefore theOTF pipeline for NICMOS and STIS iscompletely database driven.

The OTF calibration pipeline steps fora science observation are the following:

1. getting raw data from CDs (storedin a compressed form)

2. getting latest database informationon relevant keywords (new/updates)

3. getting latest database informationon relevant calibration files (new/updates)

4. setting the proper calibration switch-es relevant to the observation mode fora specific instrument

5. update the header of the sciencefile

6. apply calibration software (STS-DAS)

As already mentioned some pre-req-uisites are necessary:

1. Database updates2. Calibration file updates3. Software updatesA SYBASE replication server keeps

the CADC and ST-ECF HST databasecopies identical to the STScI one in realtime; the calibration reference files arekept up to date via a retrieval that takesplace on a daily basis by CADC, via aStarview request, and are then “pushed”to ST-ECF.

Particularities

While building the OTF pipeline, wehad to deal with some aspects which areparticular to the new instruments andwhich originate from some choices madeby the IDTs (Instrument Dedicated Team)in the designing phase of the instrumentsdata products. Multiple extension FITSfiles were introduced, and we had to waitfor a stabilised release of a new versionof IRAF (v 2.11) to be able to manipulatethe new file types. The STSDAS calibra-tion software (calnica, calnicb and calstis)evolved rapidly and is still changing.Some STIS observing modes are not yetcompletely covered by the calibrationsoftware. STIS and NICMOS associationconcepts differ, introducing thereforeasymmetry in the development of thepipelines.

Conclusions

The OTF system contributed (and stillcontributes) to the reliability of the cali-bration software: we found and reportedproblems to the STScI/STSDAS group,which quickly fixed them. As soon as anew version of the calibration software isreleased, we install it in our pipeline. Ourarchive users, with their archival re-

quests, also contribute to extensively testthe software. The OTF pipeline is usedat CADC to produce NICMOS and STISpreview images/spectra of all the availa-ble datasets (15 minutes after releasedate), further contributing in testing thepipeline.

In other words, the OTF pipeline, be-ing in a never-ending development phaseand continuously receiving new referencefiles, is a lively system. An observationcalibrated two months ago is differentfrom the one calibrated today. Only at theend of the life of an instrument, when the“final archive” is produced (i.e. no furtherdevelopment is foreseen), will this proc-ess stop and the best (?) calibration pipe-line be available to the community.

At the time of writing, the HST archiveis composed of 269 CDs for the RAWdata (as of July 1st, 1998), and has 18GBytes of calibration files.

The HST OTF service is available at:http://archive.eso.org/archive/hst/

at ST-ECF (catalog@eso.org)ht tp : / /cadcwww.hia.nrc.ca/hst /

at CADC (cadc@hia.nrc.ca)The ESO OTF service is foreseen;

sometime in the future, also NTT and VLTarchive users will benefit by this indispen-sable archive tool.

Acknowledgements

All this work has been made possiblethanks to the help of the STScI/STSDASteam, and especially of Phil Hodge andHoward Bushouse.

References

[1] Dennis R. Crabtree, Daniel Durand, SiverinGaudet, Norman Hill: “The CADC/ST- ECFArchives of HST Data: Less is More”; As-tronomical Data Analysis Software and Sys-tems V, ASP Conference Series, Vol. 101,1996, George H. Jacoby and JeannetteBarnes, eds.

[2] The HST Data Handbook, STScI.

A. Micolamicol@eso.org

HST Archive News: WFPC2 AssociationsA. MICOL1, B. PIRENNE 2, 3

1ESA/ST-ECF, 2ESO/DMD, 3ESO/ST-ECF

Astronomers having browsed/visitedthe HST archive in the last six monthshave encountered a new type ofWFPC2 dataset: the Association. Thisis the materialisation of a new serviceoffered to ST-ECF archive researchers,meant to reconstruct the otherwise miss-ing knowledge of the observing strate-gy (expected CR-SPLIT, expected dith-

ering) adopted by a WFPC2 PI. UnlikeNICMOS and STIS, where a datasetmight be constituted of a set of expo-sures, the WFPC2 dataset’s structurewas thought to be a repository of allthe files belonging to a single exposure.Building associations of WFPC2 expo-sures is therefore to be considered animportant step towards a comprehensive

description of the HST archive contents.The association concept alleviates theneed to discover:

• which observations can be groupedtogether in order to run a cosmic-raycleaning algorithm;

• how a set of WFPC2 images mapthe region around an astronomical sourceof interest.

29

To achieve this goal of re-constructingthe observing strategy of the PI, it is nec-essary to find out the exact displacementbetween any two exposures. There aretwo methods to compute the displace-ments among the images: by using theWorld Co-ordinate System (WCS) key-words normally stored in the header ofthe dataset, or via a cross-correlationtechnique. For WFPC2 exposures (morethan 35,000 when writing this article), anumber of problems arose while consid-ering those two approaches:

• Before April 1996 the WCS keywordsin the dataset fits header were not reflect-ing dithering strategy; even after that date,the WCS keywords are computed usingphase two proposal information, that is,WCS values do not take into considera-tion what happened during the observa-tion.

• Cross correlation of exposures wouldbe difficult due to the presence of cosmicrays and depends on the signal-to-noiseratio of the features in the images.

These problems led to the impossibil-ity to use any of those two methods in anautomatic pipeline. Instead, to computethe offsets among all the exposures inthe association, we decided to use thepointing information stored in the HSTobservation log files [2], informally called“jitter files”.

The jitter files have proven to be byfar more reliable than any other availa-ble source of pointing information [1].Some keywords (GUIDEACT, LOCK-LOSS, SLEWING, etc.) in the jitter filesalong with the standard deviations of themeasurements (right ascension, decli-nation, roll angle) are used to evaluatethe pointing stability during the observa-tion and the accuracy of the measure-ments [3].

Once the offsets (in right ascensionand declination) are computed, it is easyto derive the shifts expressed in pixelsvia the knowledge of the spacecraft ori-entation (roll angle) and of the focal planegeometry through the Science InstrumentAperture File (siaf).

A WFPC2 association containing allthe WFPC2 exposures of the requestedregion of the sky, belonging to the sameproposal, taken in the same filter, hav-ing the same position angle, can hencebe seen as the ultimate repository of theobserving strategy (real CR-SPLIT, realPOS-TARG) as attained by the tele-scope.

Via the web (http://archive.eso.org/wdb/wdb/hst/science/form) users browsethrough the associations, have a closerlook at a specific association, and imme-diately see what are the shifts among theexposures belonging to it.

Furthermore, an astronomer interest-ed in that association can issue a requestand ask our archive system to not onlyre-calibrate each exposure in the asso-ciation, but also to combine them (if theoffsets do not exceed the imposed limitof 5 mas beyond which the PSF of thecombined images is degraded) to getcosmic-ray-free products. All the steps ofthe association pipeline are documentedin log files that can be retrieved along withall the other products.

References

[1] A. Micol, P. Bristow, B. Pirenne: “Associa-tion of WFPC2 Exposures”; 1997 HST Cali-bration Workshop, Space Telescope Sci-ence Institute, 1997, S. Casertano, et. al.,eds.

[2] M. Lallo: “Observation Logs”; http://www.stsci.edu/ftp/instrument_news/Ob-servatory/obslog/OL_l.html, Data SystemsOperations Branch, Space Telescope Sci-ence Institute, July 1998.

[3] M. Dolensky et. al.: “How the analysis ofHST engineering telemetry supports theWFPC2 association project and enhancesFOS calibration accuracy”, in this issue.

A. Micolamicol@eso.org

Image from the VLT Science Verification ProgrammeThe galaxy ESO342-G017 was observedon August 19, 1998 during a spell of ex-cellent observing conditions. Two expo-sures, each lasting 120 seconds, weretaken through a red filter to produce thisphoto. The quality of the original imagesis excellent, with FWHM of only 0.26 arc-sec measured on the stars in the frame.The frames were flat-fielded and cleanedfor cosmics before combination.ESO342-G017 is an Sc-type spiral gal-axy seen edge-on, and the Test Camerawas rotated so that the disk of the galaxyappears horizontal in the figure. Thanksto the image quality, the photo showsmuch detail in the rather flat disk, includ-ing a very thin, obscuring dust band andsome brighter knots, most probably star-forming regions. This galaxy is locatedwell outside the Milky Way band in thesouthern constellation of Sagittarius. Itsdistance is about 400 million light-years(recession velocity about 7,700 km/sec).A number of more distant galaxies areseen in the background on this short ex-posure.The field shown measures ∼ 1.5 × 1.5 arc-min. North is inclined 38° clockwise fromthe top, east is to the left.

(Figure and caption are from the ESO webpages at http://www.eso.org/outreach/press-rel/pr-1998/pr-12-98.html prepared by the ESOEducation and Public Relations Department.)

30

S C I E N C E W I T H T H E V L T/ V L T I

VST: VLT Survey TelescopeM. ARNABOLDI 1, M. CAPACCIOLI 1, 2, D. MANCINI 1, P. RAFANELLI 3, R. SCARAMELLA 4,G. SEDMAK 5, G.P. VETTOLANI 6

1Osservatorio Astronomico di Capodimonte, Napoli, Italy2Dipartimento di Scienze Fisiche, Università Federico II, Napoli, Italy3Dipartimento di Astronomia, Università di Padova, Italy4Osservatorio Astronomico di Roma, Monteporzio Catone, Italy5Dipartimento di Astronomia, Università di Trieste, Trieste, Italy6Istituto di Radioastronomia, CNR, Bologna, Italy

1. Introduction: VST ScientificFramework

The VLT era is rapidly approaching:the first instruments, ISAAC and FORSat UT1, will be available by the year 1999.Indeed, the beginning of the new millen-nium will witness fierce competitionamong quite a few 8-m telescopes, op-erated by different groups. As a conse-quence, there is a strong need to pre-pare, in a timely manner, suitable target-lists for the VLT, in order for it to play theleading role in ground-based optical andIR astronomy in the next decade. Prepa-ration is one of the keys to success forthe VLT observations.

The VLT will work at flux levels forwhich no whole-sky surveys are availa-ble, and most of the currently submittedscience cases (SC) will not be feasiblewithout major preparatory work (Renzini& Leibundgut, 1997, The Messenger 87,21; Da Costa et al., 1997, The Messen-ger 91, 49). The SC target selections arebased on extensions of current data sets.Only a few spectroscopic or high imag-ing projects at the VLT will plan to buildtheir own catalogues directly from VLTimaging data. In conclusion, the exploi-tation of the VLT requires catalogues ofobjects and supporting observations ob-tained at other – smaller – telescopes.As an example, multi-object spectrosco-py depends on very accurate target po-sitions which must be provided in ad-vance of the actual VLT observations (toset up slitlet arrays for FORS or providemasks for VIMOS and NIRMOS).

The need to find faint or rare but inter-esting objects to study with the VLT instatistically significant quantities hasurged ESO to start the multi-band ESOimaging survey (EIS; Renzini & Da Cos-ta, 1997, The Messenger 87, 23), andmany other observatories are building uplarge field-of-view instruments (e.g. theSLOAN project). However, EIS is meantto supply targets only for the near future,but certainly cannot sustain the needs ofthe ESO community which will be usingthe VLT for top-level science during theseveral years to come.

Over the years, broad- and narrow-band wide-field imaging (WFI) has pro-vided the astronomical community witha wealth of data, which has been of greatimportance in many different fields in as-trophysics and cosmology. WFI with2-metre-class telescopes has been, upto now, the key instrument to producestatistically-controlled target-samples tobe studied both photometrically and spec-troscopically with 4-m telescopes. Theadvent of 8-metre-class telescopes re-quires extensions of WFI down to muchfainter magnitudes which are out of reachof photographic material. Planned CCDWFI surveys (such as the SLOAN DSS)will push the magnitude limit down to RAB= 23, but so far they are restricted to thenorthern sky only.

The above arguments justify the com-pelling need for a dedicated medium-sizetelescope with a wide-field imaging ca-pability in the southern hemisphere. Thisneed motivated the proposal to ESO bythe Capodimonte Astronomical Observ-atory (OAC) at Napoli to build the VST (=VLT Survey Telescope) and place it atParanal.

The VST is meant to be a highly effi-cient telescope; it will reach a magnitudelimit of R = 25 AB mag arcsec–2 in 3 × 10min exposures over a field of 1 deg2, withan instrumental resolution of 0.21″. Thescope of this facility is to supply completedatabases for VLT science, and possiblyto produce new science from the WFIdata alone. Because of its complemen-tary use to the VLT and for an obviousintegration in the VLT operating system,it shall run with the same software andbe a fully dedicated instrument for multi-band optical imaging.

The VST may be relevant for anumber of topics such as those listedhereafter (the list is by no means com-plete):

(i) Distant objects: quasars, high-zgalaxies, clusters of galaxies, superno-vae, lensed objects, absorption systems,weak lensing;

(ii) Nearby galaxies: globular clusters,HII regions, planetary nebulae, novae,emission-line stars;

(iii) The Galaxy: the items from (ii) plussubdwarfs, white and brown dwarfs, met-al and very rich/poor stars, and microlens-ing towards the Galactic bulge for statis-tics on the presence of earth-like plan-ets;

(iv) Optical identifications: the VST willprovide optical identifications of objectsfound at other wavelengths, such asX-rays, IR, microwave and radio contin-uum sources.

Once the above galactic and extra-galactic targets are identified via the VST,they will require further imaging (at dif-ferent bands and angular resolution) pluslow- and high-resolution multi-objectspectroscopy (both in the visual and IR)available at FORS1/2, ISAAC, CONICA,VISIR, UVES, VIRMOS. In particular theVST will constitute an essential tool forthe construction of surveys.

Therefore, the strong advantages ofthe VST project in the VLT era are:

1. a wide-field imaging project fullyconceived for and devoted to VLT sci-ence. Its task will be to provide the pre-paratory data for the follow-up observa-tions with VLT;

2. a wavelength range from UV to Iwith high efficiency in the UV;

3. a minimum of additional optics com-bined with a high DQE detector;

4. the exploitation of the outstandingphotometric and seeing characteristics ofthe Paranal site;

5. the high efficiency in both calibra-tions and data reduction in view of thecomplete dedication of the telescope toimaging with a single instrument.

In this paper we describe the historyof the project, and the telescope concept.This project is then compared to existingor planned WFI facilities. For a detaileddescription of the science case with theVST we refer to the “VST proposal”, avail-able upon request at the OAC.

2. History of the Project

Given the scientific framework and theneed for such a facility, the OAC Direc-tor, M. Capaccioli, planned to engage theNapoli Observatory in the realisation of

31

Figure 1: CAD 3-D view of the VLT Survey telescope.

a 2.6-m telescope for wide-field imagingand complementary use to the VLT. Forthis purpose the Observatory Councildecided to allocate 6.5 billion Italian lireof ad-hoc funds assigned to OAC by theMinistry of University and Research aspart of the programme of developing de-pressed areas in the south of the countryby financing cultural activities and exist-ing centres of excellence in science andtechnology. This resolution rested also onthe consideration that OAC has somebackground in astronomical instrumen-tation, given its collaborations with ESOand other European institutions in theVIRMOS project, and its contributions tothe TNG, the Italian 3.5-m National tele-scope, just to name two of its partner-ships in international projects. Throughits Technology Working Group (TWG)managed by D. Mancini, OAC has alsodesigned and built its first telescope, analt-azimuthal 1.5-m aperture instrumentnamed TT1 (Toppo Telescope No. 1),which is now waiting for the transporta-tion at Toppo di Castelgrande (Potenza),the former domestic site of the TNG; it

should be operational by the end of thecurrent year. Furthermore, science-wiseits research staff has a strong interest anda well-established background in surfacephotometry and wide-field imaging.

In March 1997, the OAC Director ap-pointed a Scientific Steering Committee,chaired by G. Vettolani, to study the casefor a 2.6-m WFI telescope, to be pro-posed to ESO for installation at theParanal Observatory (Chile). The Com-mittee prepared the proposal with thescientific goals, which were used by D.Mancini and the OAC TWG to develop apreliminary study for the optical, mechan-ical, and electronic specifications of a2.6-m telescope with a 1 square degreefield-of-view (FOV). The combined scien-tific and technical proposals were sub-mitted to ESO on June 17, 1997. Theproject was presented to the DirectorGeneral of ESO, R. Giacconi, and to theheads of ESO divisions on July 29, 1997.The scientific case and the technical pro-posal were then revised to incorporateall of the ESO suggestions. It was alsoagreed to change the project name from

the former TT2 (Toppo Telescope No. 2)to VST. During this initial phase of theproject, the ESO contact point was S.D’Odorico. The final VST proposal wasreviewed by the ESO Scientific and Tech-nical Committee (ESO/STC – 219 docu-ment) at the occasion of its 44th Meetingin October 1997. The STC expressedstrong support for the project, assigningit a high priority. The VST Memorandumof Understanding between OAC and ESOwas submitted to the ESO Council forapproval on the 11th–12th of June 1998,and the official kick-off of the project washeld during a meeting at ESO on the24th–25th of June 1998. The OAC willbe responsible for the design, construc-tion, and commissioning of the VST, andthe OAC will provide up to two astrono-mers for the support of the VST opera-tion starting in 2000. ESO will be incharge of the design and realisation ofthe enclosure and its control system pluscivil works at the Paranal site and of pro-viding a camera adequate to fulfil thegoals of the project.

On March 1998, Zeiss Jena wasawarded the realisation of the VST op-tics; the contract with Zeiss was signedby the end of May and the final opticaldesign agreed by ESO and OAC was for-warded to Zeiss at the end of July. OnApril the 4th, 1998, ESO issued a call fortender for the realisation of the 16k × 16kCCD camera to be placed at the Cas-segrain focus of the VST.

The OAC Council has appointed G.Sedmak, from the University of Trieste,as Project Manager for the realisationphase of the VST and D. Mancini as Dep-uty Project Manager. The VST Manageron the ESO side will be R.J. Kurz, andthe Deputy Manager S. D’Odorico. AnAdvisory Board will monitor the progressand give inputs to the Project Manager:the Chairman of this Committee is M.Capaccioli, the Co-chairmen are P. Ra-fanelli and P. Vettolani. The experts areP. Dierickx, K. Freeman, and D. Hamil-ton. The VST Project Scientist is M. Arn-aboldi, and the Chairman for the VSTScience Group is R. Scaramella.

3. VST Technical Overview

3.1. The project philosophy

The scientific goals for the VST projectcan be summarised as follows: a fullydedicated instrument with an excellentimage quality on a large FOV, high oper-ational efficiency, high reliability, and com-pliance with VLT standards. Therefore theproject guidelines are:

• Optimisation of the whole system forthe global-system cost-optimisation.

• Extended use of finite element anal-ysis, which optimises the structure costsand performances.

• The use of the system just as a sur-vey telescope will make it possible to sim-plify the overall project during the finaldesign phase.

32

• The OAC TWG will construct inhouse part of the telescope subsystems.This will make it possible to achieve abetter steady-state mean-time-to-repara-tion (MTTR), given the knowledge of thesystem by the TWG team.

3.2. The telescope concept

The VLT Survey Telescope is a 2.6-mtelescope, designed for Cassegrain op-erations, with a corrected FOV diameterof 1.5 degree, to be matched with a 16k× 16k CCD mosaic camera, with a 15 µmpixel. The telescope has an Alt-Az mount-ing, which allows a high mechanical stiff-ness and a compact overall structure: a3-D view is shown in Figure 1. The struc-ture will be an open frame with tubularcomponents in order to increase the stiff-ness vs. weight ratio and to simplify anyinstallation procedures. The wide-fieldcorrector is designed to cover the wholevisual wavelength range, from U to I withan encircled energy of 80% in a 1.7 pixelor better. Following the kick-off meetingat ESO, the corrector design in the VSTproposal was revised: the new correctordesign foresees the use of two lensesplus a curved dewar window when ob-serving near zenith, and this configura-tion can be replaced by one lens plus anADC, when observing in the B to I bandsat large zenithal distances. The primaryand secondary mirrors will be active andcontrolled by means of a Shack-Hartmanwavefront sensor; the optical scheme ofthe VST is shown in Figure 2, and Figure

3 shows the encircled energy diagram forthe U band in different regions of the fo-cal plane.

The telescope enclosure will be de-signed in collaboration with ESO. Thedesign takes into account the use of anair-conditioning system to control the tem-perature of the telescope during the day.The final design of the enclosure will becompliant with a number of nights lost dueto wind of about 10% during a year ofoperation, based on the wind statisticsof the Paranal site.

The VST telescope control architec-ture follows the concepts developed forthe VLT by ESO; in doing so, the systemcoherence is increased, an easier main-tenance programme is possible, and theintegration into the existing ESO hard-ware and software environments is en-sured. The basic idea is to maximise theuse of VLT standards and software,whenever it is possible and convenient.

4. How Does VST Comparewith Other WFI FacilitiesWorld-wide?

Tables 1–5 provide an exhaustivesummary of existing and planned (withina 3-year period) WFI facilities. Table 1provides a summary of the non-ESO WFIfacilities; Table 2 deals with the WFI ESOfacilities, and Tables 3 and 4 contain de-tails on planned multi-colour surveys. Thelast line of Table 3, 4 and 5 reports thevalues for the VST facility according tothe current status of the project.

4.1. Why the Paranal site?

Such a point is crucial in addressingthe efficiency of the VST telescope withrespect to existing or planned WFI facili-ties world-wide. The Paranal site has thebest uncorrected seeing (0.4″; medianseeing 0.65″) in the southern hemisphereand a very high percentage of photomet-ric nights (77%). We can quantify a 35%gain for the percentage of photometricnights, and a 36% gain in the better me-dian seeing (0.65″) of Paranal vs. La Si-lla (0.87″) which result in an “environmen-tal gain” in efficiency of the Paranal sitewith respect to La Silla of a factor 2.35. Ifwe consider the “environmental gain” ofthe Paranal site for the winter season(ideal for the Galactic bulge studies), thisbecomes a factor 3 from the higher per-centage of photometric nights at Paranalduring this season. The choice of theParanal site for the VST will make sucha telescope one of the most competitivededicated WFI facilities in the world, asis clear from the following direct com-parison with existing ESO and non-ESOfacilities.

4.2 Comparison with non-ESOWFI facilities

From Table 1, in the southern hemi-sphere, neither existing (UKST, CTIO) norplanned (AAT, GEMINI) WFI facilities cancompete with the VST project when weconsider scientific programmes aiming atsurveying several hundred square de-

TABLE 1. Summary of non-ESO WFI facilities

Name Aperture Inst. Focus FOV Mpix Scale Year % Seeing Country(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11)

NOAO 0.9 16 CCDs PF 1 67 0.43″ 1997 20 1″ USAING 2.5 4 CCDs PF 0.16 16 0.37″ 1997 25 0.75″ NL-UK-ENOAO 4 16 CCDs PF 0.36 67 0.27″ 1997 20 1″ USAWHT 4 1 CCD Cass. 0.025 1 0.6″ 1997 20 0.75″ NL-UK-Edu Pont 2.5 WFC Cass. 0.137 4 0.75″ 1997 25 0.8″ USAUH 2.2 8kCCDs F/10 0.67 67 0.13″ 1997 30 0.63 USALaval 2.7 CCD PF 0.5 4 0.6″ 1997 100 1″ CANUniv. 5.1 CCD PF 1997 100 1″ CANAPO 3.5 DSCCD 4 0.141″ 1997 20 13 USASubaru 8.3 Suprime PF 0.136 80 0.18″ 1999 15 0.6″ Japan

CamCFHT 3.6 MEGA Cass. 1 288 0.21″ 1999 20 0.6″ France-CAN

CAMSloan 2.5 30 CCDs Cass. 9 126 0.4″ 2000 100 1″ USAMMT 6.5 MEGA Cass. 1 268 0.22″ 2000 20 0.8″ USA

CAMKeck II 10 DEIMOS Nashm. 0.045 134 0.12″ 17 0.55″ USAC.Alto 3.5 WFNIR PF 0.01 1 0.4″ 1997 30 1″ D-E

OmegaCTIO 4 4 CCDs PF 0.25 16 0.4″ 1996 15 0.9″ USA-Othr.UKST 1.2 Plates Schmidt 43 67″/mm 1997 100 1.6″ UK-AUSAAT 3.9 CCD PF 0.02 4 0.367″ 1998 30 1.6″ UK-AUSGemini 8 CCDs Cass. 0.56 2000 25 0.25″ UK-USA-

CAN-Othr.

Column (1) conventional name of the telescope; Column (2) size of the primary mirror in metre or equivalent diameter in case of multiple telescopes; Column (3) type(CCDs or photographic plates) and name of the instrument for wide-field imaging; Column (4) focus where the instrument is placed; Column (5) FOV in square degree;Column (6) number of pixels in Mpixels; Column (7) scale in arcsec pixel–1; Column (8) expected year of completion; Column (9) fraction of time available for wide-fieldimaging, based on normal use; Column (10) average seeing at site; with AO, the seeing of Gemini is predicted to be 0.25″ in the wavelength range 0.5–0.9 nm. Column(11) countries.

33

grees on the sky. None of these south-ern facilities is conceived as a fully dedi-cated telescope for wide-field imaging: allof them suffer either from a small FOV,large pixel-size, poor seeing, or a smallfraction of allocated observing time. Incomparison with these non-ESO tele-scopes (see Table 1), one sees that VSThas the largest FOV (1 deg2), with ade-quate spatial sampling, the best uncor-rected seeing (0.4″) in the southern hemi-sphere, and the largest fraction of timeavailable (100%) for WFI.

Tables 2 and 4 provide, among otherparameters, the survey figure of merit ηdefined as

η = Ω × D2 × DQE × (seeing)–2 × ∆ T (1)

where Ω is the area of the detector indeg2, D is the diameter of the primarymirror, DQE is the quantum efficiency ofthe detector in the R band for the visibleor in K for the NIR, and ∆T is the fraction

of the time available for WFI, based onnormal use during a year. For the VST,the value of figure of merit, calculated forΩ = 1 deg2, D = 2.6 m, DQE = 0.85%, ∆T= 1 and a mean seeing value FWHM =0.65″, is η (VST) = 13.59. Only two otherprojects have comparable survey figuresof merit: MEGACAM and SLOAN, whichare both located in the northern hemi-sphere. The former project has a η = 5.76because of a larger mirror diameter (4 m),but only a fraction of the CFHT observ-ing time1 will be devoted to WFI, while100% of the time will be available for WFIat the VST. The SLOAN project has ahigher η (= 22.5) because of its large FOV(9 deg2) with respect to the VST. Such afacility will be entirely devoted to a skysurvey and its spectroscopic follow up,without the flexibility which is needed for

a variety of VLT preparatory works(choice of different areas in the sky, limit-ing magnitudes and broad bands).

When considering large sky areas, allother ESO and non-ESO southern tele-scopes have survey figures of merit whichare smaller by at least an order of mag-nitude than that of the VST.

4.3. Comparison with ESO WFIfacilities

Table 5 shows that the only ESO facil-ity competing with the VST is the 2.2-mtelescope when it is equipped with a fo-cal reducer and a 8k × 8k CCD mosaiccamera. Therefore, the VST telescope isa competitive facility only if it produces amajor increase in the telescope efficien-cy for WFI with respect to the ESO 2.2-mtelescope.

We have addressed in detail the “envi-ronmental gain” of the Paranal site vs. the

TABLE 2. Summary of ESO WFI facilities

Name Aperture Instr. Focus FOV Mpix Scale Year % Seeing η(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11)

DENIS I 1 CCD Cass 0.04 1 1″ 1997 50 0.9″ 0.015ESO 2.2 m 2.2 8 CCDs Cass. 0.3 67 0.24″ 1998 75 0.9″ 1.125DENIS J, K 1 CCD Cass. 0.04 0.06 3″ 1998 50 0.9″ 0.015

VST 2.6 32 CCDs Cass. 1 256 0.21″ 2001 100 0.7″ 13.6

Column (1) conventional name of the telescope; Column (2) size of the primary mirror in metres; Column (3) type (CCDs or photographic plates); Column (4) focuswhere the instrument is placed; Column (5) FOV in square degrees; Column (6) number of pixels in Megapixels; Column (7) scale in arcsec pixels; Column (8)expected year of completion; Column (9) fraction of time available for WFI, based on normal use; Column (10) average seeing at site; Column (11) survey figure ofmerit; see eq. 1

TABLE 3. Summary of Planned Wide-field Imaging Surveys (visible/NIR).

Experiment WFI Medium Bands Scale Lim. mag. Sky cov. DQE/seeing2

(1) (2) (3) (4) (5) (6) (7) (8)

Palomar Sch.1 43 3103 plates J F N 67″/mm J = 21.5 2.5 × 104 0.04/1″MEGACAM 1 CCD B V R I Hα 0.21″ B = 24 102 0.80/0.6″CFHT12K 0.33 CCD B V R I 0.21″ IAB = 24 25 0.80/0.6″SLOAN 9 CCD u′ g′ r′ i′ z′ 0.4″ R2 = 23 104 0.4/1″DENIS 0.02 CCD I J Ks 1″, 3″, 3″ Ks = 13.5 2.5 × 104 0.61/0.9″EIS ESO 0.01 CCD U Bw Vw Iw 0.17″ I = 23.2 24 0.75/0.9″EIS ESO 0.01 CCD U Bw 0.17″ BW = 24.3 1.9 0.85/0.9″EIS ESO 0.01 CCD U, Gr, Gg, I, K 0.17″ K = 21.5 0.01 0.85/0.9″BATC 0.95 CCD 15 I, B3 0.85″ V = 21 475CADIS 0.01 CCD K′ 0.4″ K′ = 21.2 0.28 0.75/1"LIMITS 0.1 CCD 40 N.B.4 0.6″ R = 23.5 20 0.3/1″Hα Survey 43 160 plates Hα 67″/mm 7 × 103 0.04/1.6"APM 25 2103 plates IIIaJ 67″/mm 20.5 2.2 × 104 0.04/1″2MASS 0.02 CCD J H Ks 2″ Ks = 13.5 2.5 × 104 0.8/0.9″NOAO 0.07 CCD I 0.47″ I = 23.5 16 0.8/1″NOAO 0.36 CCD B R I J H K 0.27″ I = 26 18 0.8/1″

VST5 1 CCD U ′ BW VW RW IW 0.21″ RW = 25.5 300 0.85/0.7″

Column (1) conventional name of the survey; Column (2) detector area in square deg used for the survey; Column (3) plates or digitised images;Column (4) broad or narrow bands covered by the survey; Column (5) scale in arcsec pixel-1; Column (6) limiting surface brightness; Column (7)sky coverage of the survey in square degrees; Column (8) detector quantum efficiency in the R band for optical surveys and K band for NIRsurveys, and average seeing;1Data from the CRONA project.2The survey will reach fainter limiting magnitudes in those 5 bands (25 in R) on a strip 2 × 50 square degree, centred on the South Galactic Pole.315 intermediate-band filters (∆λ = 200, 300 Å) from 330 nm to 1 µm will be used for this survey.440 narrow-band filters from 400 nm to 1 µm will be used for this survey, plus B, V, R. I.5This is an example of a possible survey with VST.

1We have assumed here that 20% of the CFHTtime will be devoted to WFI.

34

La Silla site in Section 4.1 to give a factor2.35 (3 for the winter season). The tele-scope area (2.6-m vs. 2.2-m) gives an-other factor 1.4. Furthermore, the VST tel-escope is designed to give a square FOVof 1 deg2 for WFI (to be matched by a 16k× 16k CCD mosaic), while the correspond-ing FOV for the 2.2-m2 is 0.29 square deg.:the gain in efficiency is then by a factor3.4 from the larger imaging area. Anotherpoint concerns the comparisons of effi-ciency vs. wavelength of the wide-fieldcorrector for the 2.6 m VST telescope andthe ESO 2.2-m, see Table 5.

If we consider the average gain in ef-ficiencies as a function of wavelength,then we obtain for the VST an additionalfactor 1.16 with respect to the ESO 2.2-mtelescope3. The overall gain in efficiencyof the VST vs. the ESO 2.2-m telescopewith the 8k × 8k CCD mosaic camera isa factor 13 (16.5 in the winter season).

Some words are in order for the com-parison of the VST with the VLT plannedinstrumentation. VIMOS will have p 0.054deg2 FOV, and an estimated ∆T = 0.2:considering the VLT mirror area, VST willbe 8.5 times more efficient for WFI thanVLT + VIMOS. Although it can reachfainter magnitudes, VIMOS is not a dedi-cated imaging facility and will have itsmain use for wide-field spectroscopy, forwhich it represents a unique facility.

4.4. Telescope performances

We explicitly show efficiency curves,values and assumptions used in the es-timates of S/N ratios, by considering aCCD mosaic camera with pixel size of15 µm, 0.24″ pix–1 scale, a CCD quan-tum efficiency as in Figure 3, based onESO CCD EEV curves, and the meas-ured efficiency of the UVES multi-coatinglayers (received from ESO on May 28,1997).

We have assumed an aluminium coat-ing for the primary and the secondarymirrors, and anti-reflection coated surfac-es for the corrector and the dewar win-dow. The total transmission factor ofthese components is shown in Figure 4for the wide field corrector according tothe optical design in the VST proposal.In the same figure we show the CCD re-sponse expected (“goal” or minimumspecifications).

The telescope performances expect-ed for the VST are based on the efficien-cy curve in Figure 3 and a detector areaof 1 deg2. We give here three examplesof possible science cases which can ful-ly exploit the VST imaging capability andexcellent image quality.

The study of large-scale structures inthe universe and galaxy evolution re-quires galaxy catalogues extended overfairly large areas, which can be obtainedthrough multicolour imaging. A very widesurvey (several hundred square degrees)with a limiting magnitude in V of 25.5 ABmag arcsec–2, in at least two bands, canfully exploit the capabilities of the VST.

In 30 min-exposures, one reaches Vwlimp 25.5 AB mag arcsec–2, which couldtranslate to Vwlim P 24 for galaxies (onealso has Iwlim p 24 AB mag arcsec–2, soIwlim P 22.5 for galaxies). So, a reasona-ble estimate can be P 50 hrs of expo-sures per band per 100 deg2. If one hasabout 6 effective hours per night, then weneed 17 nights to do accurate photome-try in two bands over 100 deg2. With 17nights of VST time, we will produce anamount of data which is nearly 4 timesthat produced by EIS “wide”.

Furthermore, for the typical magnitudelimit of wide area imaging (R Q 24), onewould measure the photometric param-eters and colours for several million gal-axies, including a large number of low-surface-brightness galaxies whose prop-erties are still largely unknown. Further-more, thanks to the very high throughputof VST in the UV, large samples of high-zcandidate galaxies can be selectedthrough the Lyman break technique. If1/3 of a year observing dark time can bedevoted to deep multicolour imaging (5bands, Iwlim P 26 AB mag arcsec–2), onecould cover at least 5 deg2 and obtain asample of p 8000 candidates with zlarger than 3.

High performances are expected alsofor narrow-band imaging with the VST.As an example, for a 6 × 2000s expo-sure, the limiting [OIII] 5007 flux at theVST is of 4 × 10–18 erg cm–2 s–1 (at the3σ level). In the case of emission line stel-lar objects like planetary nebulae (PNs),which have a strong emission in the [OIII]λ 5007 line, the limiting [OIII] flux from

TABLE 4. Summary of Planned Wide-fieldImaging Surveys (visible/NIR) Cont.

Experiment η Countr.(1) (2) (3)

Palomar Sch.1 1.36 USA-Eur.MEGACAM 5.76 Fr-CANCFHT12K 1.9 France-CANSLOAN 22.5 USADENIS 0.015 ESOEIS ESO 0.01 ESOEIS ESO 0.011 ESOEIS ESO 0.011 ESOBATC USA-ChinaCADIS 0.03 (3.5 m) D-ELIMITS 0.7 CANHα Survey 0.95 UK-AUSAPM 0.81 UK-AUS2MASS 0.03 USANOAO 0.3 USANOAO 1.0 USA

VST2 13.6 I-ESO

Column (1) conventional name of the survey;Column (2) survey figure of merit as in eq.1;Column (3) countries.1Data from the CRONA project.2This is an example of a possible survey with VST.

TABLE 5. Corrector efficiencies.

Name 1100[nm] 1000[nm] 800[nm] 600[nm] 400[nm] 365[nm] 350[nm]

ESO 2.2-m 0.77 0.78 0.80 0.82 0.81 0.71 0.47VST 2.6 m 0.87 0.91 0.925 0.90 0.91 0.90 0.86

Efficiency values for the wide-field corrector for the ESO 2.2-m and for the VST 2.6-m telescope according tothe VST proposal. The efficiencies for the ESO 2.2-m are extracted from the ESO document INS 97/001.

2The maximum FOV for a square detector at theESO 2.2-m focal reducer is 0.29 deg2.

3This value is a lower limit: after the ESO kick-offmeeting, the optical design for the wide-field correc-tor has only two lenses (or one lens plus ADC), sothe VST corrector efficiency is higher than the esti-mated one quoted in the VST proposal.

Figure 2: Opticallayout in the visualrange from 365 to1014 nm; the wide-field corrector con-figuration here iswith one lens plusthe ADC.

35

Figure 4. The efficiency curves for the telescope+corrector according to the VST proposal; theCCD response and their product are shown as a function of wavelength.

Figure 3. Encircledenergy in the Uband at differentpositions in the fo-cal plane.

PNs at the distance of the Fornax/Virgocluster (17 Mpc) that we can detect atthe ESO NTT with EMMI in the multi-ob-ject mode is 4 × 10–17 erg cm–2 s–1. Thislimiting flux can be detected today withthe 4-m CFHT and the 0.25 deg2 UH8Kcamera in about 10 × 2000s exposureswith S/N = 10.

The limiting [OIII] 5007 flux of 6 ×2000s exposures at the VLT plus themulti-object spectrograph VIMOS (R =2500) is 5 × 10–18 erg cm–2 s–1: by com-parison with the limiting flux in this wave-length range at the VST, we can con-clude that such a facility is the ideal in-strument to produce catalogues for PNcandidates for spectroscopic follow-upat the VLT.

4.5. VST data throughput

In order to estimate the VST data flow,we analyse two different possible cases:survey and deep observations mode. Inthe first case we assume a total expo-sure time of 30 min to be split in threeexposures, while in the second case weconsider 2 × 30 min exposures. In bothcases we assume an average length ofthe night at Paranal of 7 hours and twopossible options for the CCD:

• Minimal: 8k × 8k with a 16-bit word.Readout is assumed to be fast (10 s) forsurvey mode, and slow (85 s) for deepobservations. Each frame consists of0.13 Gbyte of data.

• Maximal: 16k × 16k with 18-bit word.Fast readout time is estimated at 1 minand slow readout at 5.5 min. Each frameconsists of 0.58 Gbyte of data.

In both cases we take into account a10% overhead time for each frame. Thedata flows (in Gb) produced during theaverage night are listed in Table 6.

5. Summary

From the comparison of the differentcharacteristics and quality parametersdisplayed in Tables 1, 2, 3, and 4, plus

the consideration based on the highthroughput at different wavelengths, itappears that the VST is the most com-petitive project in the Southern hemi-sphere for WFI. In conclusion, the needfor an ESO WFI facility such as theVST is very strong primarily as a com-plementary tool for the VLT, and alsoas an independent survey tool. TheESO community needs a survey facilitysuch as the VST in order to producethe databases which are essential toachieve excellence in the science ofthe VLT era.

6. Acknowledgements

We would like to acknowledge themembers of the Scientific Steering Com-mittee – Jacques Boulesteix, KennethC. Freeman, Loretta Gregorini, AngelaIovino, Giuseppe Longo, Tommaso Mac-cacaro, Yannick Mellier, Georges Meylan,Giampaolo Piotto – for their contributionsand support to the VST project. The VSTproposal would not have appeared in itsform without the work and commitmentby Enrico Cascone, Debora Ferruzzi,Valentina Fiume, Guido Mancini, Gabri-ella Marra, Francesco Perrotta, PietroSchipani, Gianfranco Spirito, members ofOAC TWG. We would like to thank San-dro D’Odorico, the ESO contact pointduring the initial submission of the VSTproposal, for his supervision and supportto the project, Richard J. Kurz, PhilippeDierickx, Bernard Delabre, Donald Ham-ilton for their contribution to the VSTproject, and Alvio Renzini for his enthu-siastic encouragement.

TABLE 6. Data flow in Gbytes per night.

Mode 8k × 8k 16k × 16k

Survey 4.9 20.5Deep 1.7 6.4

M. Arnaboldimagda@na.astro.it

36

R E P O R T S F R O M O B S E R V E R S

Star FormationToward the “Quiescent” Core NGC 6334 I(N)A.R. TIEFTRUNK1, S.T. MEGEATH 2

1European Southern Observatory/La Silla2Harvard Smithsonian Center for Astrophysics, Cambridge, MA, USA

At the time of their birth, young mas-sive stars are deeply embedded in thedense molecular cores from which theyform, thus making it impossible for us toobserve them directly. The very earlystages of massive star formation cantherefore only be detected indirectlythrough the effects they have on the sur-

Figure 1: A K-band mosaic (ESO 2.2-metre)of the entire NGC 6334 I region (including I(N))overlaid with a HC3N (15–14) emission map(SEST 15-metre). The contour levels are 3 to9 by 0.5 K km s–1. The dense gas (105 cm–3)traced by HC3N is concentrated into twoclumps, the southern clump is coincident withthe FIR source NGC 6334 I and the northerncore is coincident with I(N). I and I(N) havevirial masses of P 790 M0 and P 470 M0,respectively. We note that C18O, which tracesmoderate density gas (p 1000 cm–3), is foundthroughout the mapped regions. The south-ern core, NGC 6334 I, is clearly a site of ac-tive star formation: a young stellar cluster andultracompact H II region are apparent. Inter-estingly, the K mosaic shows no clear evi-dence for star formation in the northern core,I(N).

rounding cloud via massive bipolar out-flows, enhancements of certain molecu-lar species, or infrared emission fromheated dust grains.

At a distance of 1.7 kpc, NGC 6334 isone of the nearest and most prominentregions of ongoing high-mass star forma-tion (Neckel, 1978), but because of its

southernly location, it is not well studied.The nebula is associated with a remark-able, filament-like giant molecular cloud(GMC) which contains a chain of 6 dis-tinct sites of recent high-mass star for-mation (McBreen et al., 1979). Addition-al evidence for star formation is the emis-sion of vibrationally excited H2, detectedin a large, 20″ resolution map by Straw& Hyland (1989). This map shows brightemission in the H2 1-0 S(1) line extend-ed over several square parsecs towardthe GMC.

In July 1998, we mapped the H2 1-0S(1) and Brγ lines toward a significantfraction of the molecular cloud using theFabry-Perot on IRAC2 at the 2.2-m. Thisprovided a factor 400 improvement inspatial resolution over the previous H21-0 S(1) maps of Straw & Hyland(1989). Furthermore, using the SEST inJune 1998, we conducted a multi-fre-quency study of the molecular gas inthe NGC 6334 cloud. From these datawe find that the NGC 6334 I & I(N)molecular cores, shown in Figure 1,probably contain the youngest sites ofhigh-mass star formation within the NGC6334 GMC. NGC 6334 I incorporatesan ultracompact H II region, two mid-IRsources, a young stellar cluster, a pleth-ora of masers, and at least one (proba-bly two) bipolar outflow(s). The shock-

37

excited H2 1-0 S(1) emission we detect-ed in bow-shock shaped emission knotstoward the NH3 masers detected byKraemer & Jackson (1995) (cf. Fig. 2),and the broad non-Gaussian line wingsdetected in the molecular line emission,impressively display these outflows to-ward NGC 6334 I. In contrast, I(N), acooler source has only been detectedat submillimetre or millimetre wave-lengths. Interestingly, I(N) contains sev-eral masers whose presence, far fromany known site of active star formation,has been an enigma.

From our SEST data, we confirm thatthe NGC 6334 I(N) core is chemicallyquiescent and much cooler when com-pared to its southernly neighbour NGC6334 I. Considering the apparent clus-ter embedded within the NGC 6334 Icore, this is not surprising as much ofthe molecular material within this coremust have been heated and processedby the embedded young stars. Howev-er, NGC 6334 I(N) shows some verysurprising peculiarities: bright emission

Figure 2: H2 image of the NGC 6334 I corewith contours of Brackett-γ emission. Thesedata had been obtained in an earlier run withthe ESO 2.2-metre with a 1″ resolution in the0.33″ pixel mode. We have outlined theNorth-South outflow and its candidate sourceand a proposed East-West outflow. We alsoshow the location of the Brγ line and the closerelationship of the NH3 masers (Kraemer &Jackson, 1995) and the H2 emission.

Figure 3: The panel on the left shows the SiO spectra obtained toward NGC 6334 I(N). The line intensities are about twice as strong as for the SiOdetected toward NGC 6334 I. A Gaussian has been fitted to the SiO(2-1) line. For all transitions non-Gaussian line wings out to P –40 km s–1 andP 25 km s–1 can be detected. The panel on the right shows the integrated emission from the blue and red line wings of the SiO(2-1) line asindicated in the left panel by the two arrows. The bipolar outflow is clearly detectable and has also been mapped with about half the FWHMbeamsize at the frequency of SiO(5-4).

lines of sulfur-bearing molecules andSiO line emission with broad line wings.These strong lines and their attendantline wings are clear evidence for a mo-lecular outflow and shock chemistry inthis seemingly quiescent molecular core.

Maps of the blue and red line wings inthe observed transitions of SiO indicatea bipolar flow (cf. Fig. 3). The near-infra-red maps obtained with the 2.2-m pro-vide further evidence for outflows in I(N)through the detection of vibrationally ex-

38

burst” phenomenon in active galaxies (cf.Walborn 1991). Much larger than anyGalactic SFR, the 30 Doradus regioncontains luminous clusters of massiveyoung stars emitting intense UV radia-tion and powerful stellar winds whichhave created loops and shells of ionisedgas, and shows evidence for a highly ef-ficient formation mechanism unmatchedin Galactic molecular clouds (Massey andHunter 1998). We summarise here theresults of an ongoing investigation of thecharacteristics of the highly excited mo-lecular gas and cold molecular gas to-ward the centre of the 30 Doradus region,made through observations of H2 andCO(2→1) line emission, respectively.

Cold Molecular Gas

CO emission from the 30 Doradus re-gion was first detected, using the Colum-bia millimetre radio telescope, by Meln-ick & Rubio (1985). Their pointed, lowangular resolution (8.8′) observationsshowed a weak CO line emission withseveral velocity components. Higher sen-sitivity CO mapping of a region of p 1°

cited H2 emission toward NGC 6334 I(N)(cf. Fig. 4). Thus, NGC 6334 I(N) nowappears to harbour ongoing star forma-tion, which explains the previously enig-matic presence of masers toward I(N).We suggest that I(N) is in an interestingtransition phase, transforming from achemically quiescent to a shock/outflow-dominated molecular core. Assumingthat its southernly companion, NGC6334 I, passed through a similar transi-tion phase before entering the observedhot core chemistry, these two molecularcores, embedded within the same pa-rental molecular cloud and separated byless than 0.5 pc, will allow for a uniquecase study of the chemical and physicalevolution of molecular cores in their ear-liest phases after the onset of star for-mation (Megeath & Tieftrunk, Tieftrunk& Megeath, in preparation).

References

Kraemer, K.E. & Jackson, J.M., 1995 ApJ 439,L9.

McBreen, B., Fazio, G.G., Stier, M. & Wright,E.L., 1979 ApJ 232, 183.

Neckel, T., 1978 A&A 69, 51.Straw, S.M. & Hyland, A.R., 1989 ApJ 342,

876.

Figure 4: H21-0 S(1) emission towardNGC 6334 I and I(N) from ourFabry-Perot imaging with the 2.2-m inJuly 1998. The lower part of this figureshows the same H2 emission knots asFigure 2, but with higher dynamicrange. Relative offsets are given in arc-seconds from an arbitrary off-position,chosen to align the two FP-fields. Wecaution the reader that the registrationof the data in this figure is preliminaryand may have absolute errors of sev-eral arcseconds. Note that no contin-uum and no Brγ emission could bedetected toward the shock-excited H2emission knots.

A.R. Tieftrunkatieftru@puppis.ls.eso.org

Molecular Gas in 30 DoradusM. RUBIO1, G. GARAY1, and R. PROBST 2

1Departamento de Astronomía, Universidad de Chile, Chile2 Cerro Tololo Interamerican Observatory, NOAO, Chile

Introduction

The Large Magellanic Cloud (LMC)contains numerous star-forming regions(SFRs) in an environment considerablydifferent from the Galaxy. As in our MilkyWay, SFRs in the LMC include complex-es of ionised gas, patches of dust, andclusters of young stars and share thesame markers of star formation: proto-stellar objects (Jones et al. 1986; Hylandet al. 1992), compact infrared sources(Schwering & Israel 1990; Rubio et al.1992), OH and H2O masers (Whiteoak &Gardner 1986; Caswell 1995), etc. Theyshow, however, significant differences:the ionising radiation is stronger, the lu-minous stars are less deeply embedded,there is a lack of far-IR brightness peaks,and substantially less cold molecular gas(Cohen et al. 1988; Israel et al. 1993;Kutner et al. 1997; Johansson et al.1998). The SFRs in the LMC should beimportant stepping stones between Ga-lactic SFRs and those in more distantgalaxies. In particular, the giant H II re-gion 30 Doradus is thought to be a key-stone object for understanding the “star-

centred near the exciting cluster of the HII region (hereafter the R136 cluster),made with the same instrument, werereported by Garay et al. (1993). Theysuggested that the CO emission from the30 Doradus region arises from small,dense molecular clumps that are embed-ded in a mainly atomic but partly molec-ular interclump medium where CO hasbeen destroyed by photodissociation dueto the strong UV radiation field present inthe area.

As part of the ESO-SEST Key Pro-gramme: CO in the Magellanic Clouds,Johansson et al. (1998) mapped theCO(1→0) line emission from the 30 Do-radus region with a tenfold higher angu-lar resolution (45″) than in previousworks. They identified more than 30 mo-lecular clouds within a region of 24′ × 24′,having typically sizes of 10 pc and mass-es of p 2 × 104 M0, confirming the sug-gestion made by Garay et al. (1993). Inparticular, close to the R136 cluster, Jo-hansson et al. (1998) detected two COclouds located toward the north-east andwest of R136 (clouds # 10 and 13, re-spectively) and which lie close to the edg-

39

es of filaments of ionised gas character-istic of this giant HII region. Cloud 10 isthe most luminous and massive cloud ofthe region surveyed, having a CO lumi-nosity of 9.9 × 103 K km s–1 pc2 and avirial mass of 3.8 × 105 M0. Near IR im-ages show the presence of an IR clustertowards Cloud 10 and a chain of knot-likefeatures towards Cloud 13 (Rubio et al.1992). The knots seem to be associatedwith early-type (O3) massive stars, sug-gesting that massive star formation iscurrently taking place within dense mo-lecular clumps embedded in the ionisedgas. Further evidence for a new stellargeneration of young massive stars in the30 Doradus Nebula is presented by Ru-bio et al. (1998).

Excited Molecular Gas

Recently we undertook deep imag-ing, in the 2.12µ 1-0 S(1) line of H2 and2.16µ Brγ recombination line of hydro-gen using the CTIO 1.5-m telescope, ofseveral LMC SFRs in order to investi-gate the spatial distribution of H2 andionised gas with respect to the CO mo-lecular clouds (Probst & Rubio 1998).Toward the 30 Doradus region a 5′ × 5′area was imaged, with 1.16″/pix resolu-tion, showing that the H2 emission isclumpy, with numerous knots and with areticulated pattern in contrast to the ion-ised gas which shows a filamentary struc-ture in Brγ. These near-IR images areconsiderably more sensitive than thosereported by Poglistch et al. (1995) andhave considerably higher angular reso-lution than the low-surface-brightnessline emission observations of Pak et al.(1998). Figure 1 shows a composite im-age, made using the 2.12µ and 2.16µnarrow-filter images (1-hour and 30-minexposures, respectively) in which the Brγimage has been subtracted from the H2image. The Brγ emission (shown inred-brown colours) clearly shows fila-ments and arc structures of ionised gastowards the west and north-east of thecentral cluster, as well as the filamen-tary structure of the nebula. On the oth-er hand, the distribution of H2 (shown inyellow-white colours) is characterised bycompact knots of strong emission, nearthe centre of the image, and two extend-ed areas of weaker emission, one to-ward the NE and the other toward the

Figure 1: Composite image of the 30 DoradusNebula, using the 2.12µ and 2.16µ narrow-fil-ter images taken with CIRIM (Nicmos III, 256× 256 IR camera) at the 1.5-m telescope atCerro Tololo Inter-American Observatory. Theindividual images cover an area of 5′ × 5′ cen-tred in R136 and the exposures times are 1hour and 30 minutes for the 2.12µ H2 and2.16µ Brγ, respectively. The Brγ image hasbeen subtracted from the H2 image. Red-brown colours indicate Brγ emission whileyellow-white colours indicate H2 emission.The scale is 1.16″/pix and the field shown is4.4 ′ × 4.5 ′. North is at top, east to the left.

Figure 2: Contour map of velocity integrated CO(2→1) line emission from the 30 Doradus re-gion, superimposed on the 2.12µ H2 image taken with CIRIM at the 1.5-m telescope at CTIO.The velocity interval of integration ranges from 235 to 270 km s–1. The contour levels are from0.8 K km s–1 (p 4σ) to 2.4 K km s–1 in steps of 0.4 K km s–1 and from 3.2 to 10 K km s–1 in stepsof 0.8 K km s–1.

40

SW of R136. The extended emission isfound projected toward the CO cloudsdetected in the near vicinity of R136 byJohansson et al. (1998). Projected to-ward the NE H2 region are known proto-stars and an H20 maser feature. Theextended SW H2 region exhibits a pecu-liar bubble structure apparently unrelat-ed to the Brγ emission. These bubbleshave roughly the same linear dimensionsas those seen in the Orion SFR (Tanakaet al. 1989), but unlike the latter do notcontain hot young stars or ionised gas,and show a more clumpy structure. Thenature of the bright H2 knots seen in thevicinity of R136 is particularly intriguing;they are not associated with stars, ion-ised gas, nor with CO emission as re-

ported by the Key Programme survey(Johansson et al. 1998).

Observations and Results

The presence of concentrations ofexcited molecular gas, as revealed by the2.12µ emission, in regions near the cen-tral cluster R136 where no CO(1→0) hadbeen detected by Johansson et al. (1998)clearly calls for more sensitivity CO ob-servations of the region. Hence, we un-dertook a deep survey of the CO(2→1)line emission towards the H2 knots andbright structures. This line was chosen inorder to benefit from the higher angularresolution provided by SEST at the fre-quency of 230 GHz (23″ or 6 pc at the

distance of the LMC). The aim of the ob-servations was to search for weak COemission, tracing cold molecular gas, thatmight be associated with the hot molec-ular gas of the H2 knots, and to study thephysical conditions of molecular cloudsin the presence of an intense ultravioletradiation field.

The observations were made duringtwo observing runs (March 1997 and Jan-uary 1998) using the SEST telescope onLa Silla, and performed in the position-switching mode using a nearby referenceposition free of CO(2→1) emission. Onlylinear baseline fits were needed to reducethe data. The rms noise achieved in asingle channel 0.054 km s–1 wide, is 0.07K. We fully mapped, with 10″ spacings, aregion between Clouds 10 and 13 whichencompasses the H2 knots found in thenear IR. Emission is detected in most ofthe mapped region, but with an intensityof typically 3–4 times lower than that ofthe emission mapped by Johansson etal. (1998).

Figure 2 shows a contour map of theCO(2→1) emission integrated over thevelocity interval from 235 km s–1 to 270km s–1, superimposed on the 2.12µ H2image. Besides the strong emission fromCloud 10 (NE region of the map) andCloud 13 (SW region of the map), partic-ularly notable is the detection of emissionfrom a clumpy structure, located betweenClouds 10 and 13, having an elongatedmorphology along a SE-NW direction. Inaddition, at least two other CO(2→1)clumps are clearly detected, one locatedin the east and the other in the north-westregion of the map. The central, clumpyCO(2→1) structure is closely associatedwith the H2 knots and roughly follows theirmorphology. However, the peak of the COclumps are in general not exactly coinci-dent with the peaks of the H2 knots.

The velocity structure of the CO(2→1)emission from the 30 Doradus region ispresented in Figure 3, which shows chan-nel maps of the emission integrated overvelocity intervals of 6 km s–1, superim-posed on the 2.12µ H2 image. The strongemission from Cloud 10 is clearly seenin all, but the last, channel maps, whileemission from Cloud 13 is seen in therange from 240 to 258 km s–1. Note thatthe CO emission above 10 K km s–1 isnot contoured in Figure 3 to allow a cleardisplay of the weaker emission pertainedto this work. From an analysis of the chan-nel maps we have identified seven mo-lecular clumps in the mapped region (oth-er than Clouds 10 and 13), with veloci-ties ranging from 236.7 km s–1 to 268.8km s–1. The radii of the clumps range from4 to 7 pc, the line widths range from 2.4to 7.9 km s–1, and the CO luminositiesrange from 0.5 × 102 to 4.3 × 102 K kms–1 pc2. Compared to the average param-eters of LMC molecular clouds, derivedfrom a total of about 100 CO cloudsmapped with the SEST under the CO Keyprogramme (Rubio 1997), these clumpsare smaller in size by a factor of p 4 and

Figure 3: Channel maps of the integrated CO(2→1) line emission over velocity intervals of6 km s–1, in the velocity range from 234 to 270 km s–1, superimposed on the same H2 imageshown in Figure 2. The contour levels are from 0.25 K km s–1 (p 4σ) to 1.0 K km s–1 in stepsof 0.25 K km s–1 and from 1.5 to 3.0 K km s–1 in steps of 0.5 K km s–1. The position of R136is indicated by a white star symbol in the lower-right panel.

41

have CO luminosities lower by a factorof p 102. Assuming that the CO clumpsare in virial equilibrium, we estimateclumps masses in the range from 5 × 103

to 5 × 104 M0. This hypothesis is, how-ever, arguable for clouds in the 30 Do-radus region due to the large amount ofmechanical energy that has been depos-ited in the region by the recently formedmassive stars.

Discussion

The CO(2→1) emission associatedwith the chain of H2 emission shows aclear gradient in velocity along a SE-NWdirection. In the lowest velocity interval(234–240 km s–1) the CO emission aris-es from a region associated with thesouth-eastern most H2 knots, in the nextvelocity interval (240–246 km s–1) it peaksnear the strongest H2 knot, while in thevelocity interval 246–252 km s–1 it ismainly associated with the north-westernmost H2 knot in the chain. The velocitygradient is best appreciated in Figure 4which shows a position velocity plot ofthe emission along a direction with aposition angle of 45° passing throughα (1950) = 5h39m00s.2, δ (1950) =–69°07′10″. From this figure we meas-ured, for the CO emission associated withthe H2 chain, a velocity gradient of p 0.37km s–1 arcsec–1 (or 1.4 km s–1 pc–1 at thedistance of 55 kpc) over a region of p50″ in length. If these motions are due togravitationally bound rotation around acore of mass Mc, then the observed ve-locity gradient implies that Mc p 1.4 × 105

M0 within a 7 pc core radius (cf. Arm-strong, Ho, & Barret 1985).

While the presence of such a massivecore cannot be ruled out at present, we

believe that the observed velocity field inthe 30 Doradus region is most likely pro-duced by the expansion of molecular gasdriven by the powerful stellar winds fromthe luminous stars located in the centralpart of 30 Doradus. The interaction of stel-lar winds with the ambient interstellarmedium has been extensively studied byCastor et al. (1975) and Weaver et al.(1977). Assuming that the characteristicvelocity of expansion of the clumps is16.7 km s–1 (equal to the largest observedrelative velocity of the clumps with re-spect to the ambient cloud velocity) andthat they are located at a characteristicradius of p 16 pc (equal to the radius ofthe region encompassing the sevenclumps), then the stellar wind power re-quired to form the observed structure, as-suming a medium with an initial ambientdensity of 104 cm–3, is 4 × 1039 ergs s–1.The power of the stellar wind originat-ing in R136 is estimated to be 4 × 1039

ergs s–1 (Cox & Deharveng 1983), henceit alone is sufficient to explain the ex-pansion motions of the CO clumps in30 Doradus.

The detection of compact H2 knotsprojected close to R136 seems to indi-cate that dense molecular structures cansurvive the strong winds and intense ion-ising radiation produced by the luminousyoung stars of the compact R136 clus-ter. The peak intensities in the 1-0 S(1)line from the three strong H2 knots of thechain range from p 4 × 10–5 to 6 × 10–5

ergs s–1 cm–2 str–1. If this H2 emission isproduced in a photodissociation regionexposed to a UV radiation field with anintensity p 3500 times larger than theaverage intensity of the Galactic interstel-lar field (Werner et al. 1978), then theobserved intensity in the 1-0 S(1) line

implies that the emitting region has a den-sity of p 2 × 105 cm–3 (Sternberg & Dal-garno 1989). To compute the total inten-sity in H2 from the intensity in the 1-0 S(1)line we used a scale factor of 50 (cf. Gold-shmidt & Sternberg 1995).

Conclusions and Outlook

We suggest that the weak CO clumpsfound projected toward the R136 regioncorrespond to dense fragments of gasthat are remnants of the original molecu-lar cloud in which the young massive clus-ter R136 was born. Due to the strongwinds from the massive stars and thelarge UV radiation field generated bythem, the parental molecular cloud wasfragmented and dispersed, and a cavityhas been blown. The question arises asto the spatial location of the H2 knots: Arethey embedded within the giant H II re-gion or are they rather located in theiroutskirts? If the dense fragments of mo-lecular cold gas are within the hot cavityit implies that they have not been photo-evaporated by the strong ionising radia-tion that has destroyed the rest of molec-ular gas. There are several argumentsagainst this hypothesis, however. The ve-locities of the clumps are not close to theambient cloud velocity, as expected in thisscenario, but displaced by typically 10km s–1. Further, the clumps should exhibitan externally ionised envelope which,however, is not detected in the Brγ im-age. Most likely the CO clumps are rem-nants or fragments of a supershell of mo-lecular gas that has been driven by thepowerful stellar winds from the cluster ofluminous young stars. This hypothesis issupported by the observed velocity fieldof the clumps. We envisage that the coldand dense fragments are exposed to thestrong photodissociating UV radiationfield from the central cluster and henceare being heated on their periphery. Themolecular gas is being excited and emitsby fluorescence in the 1-0 S(1) infraredline as seen in the 2.12µ images. Alter-natively, the CO clumps could be under-going shock interaction with the strongwinds of the massive stars in the centralcluster and thus, emitting in the NIR. Themolecular/ionised gas interface will beinvestigated through a spectroscopic fol-low up to determine the excitation mech-anism, temperature, and column densityof the excited H2.

References

Armstrong, J.T., Ho, P.T.P., & Barret, A.H. 1985,ApJ 288, 159.

Castor, J., McCray, R., & Weaver, R. 1975,ApJ 200, L107.

Caswell, J.L. 1995, MNRAS 272, L31.Cohen, R.S., Dame, T.M., Garay, G. et al.

1988, ApJ 331, L95.Cox, P., & Deharveng, L. 1983, A&A 117, 265.Garay, G., Rubio, M., Ramirez, S., et al. 1993,

A&A 274, 743.Goldshmidt, O., & Sternberg, A. 1995, ApJ 439,

256.

Figure 4: Position-velocity diagram of the CO(2→1) line emission from the central region of30 Doradus along a direction with a position angle of 45° passing through α (1950) =5 h39 m00.2 s, δ (1950) = –69°07 ′10 ″. Contour levels are –0.07, and 0.07 to 0.7 K in steps of0.07 K. The spectral data used to make this diagram have been smoothed to a velocityresolution of 0.25 km s–1.

42

Hyland, A.R., Straw, S., Jones, T.J., et al. 1992,MNRAS 257, 391.

Israel, F.P., Johansson, L.E.B., Lequeux, J.,et al. 1993, A&A 276, 25,

Johansson, L.E.B., Greve, A., Booth, R.S., etal. 1998, A&A 331, 857.

Jones, T.J., Hyland, A.R., Straw, S. et al. 1986,MNRAS 219, 603.

Kutner, M.L., Rubio, M., Booth, R. S. et al.1997, A&AS 122, 255.

Massey, P., & Hunter, D.A. 1998, ApJ 493,180.

Melnick, J., & Rubio, M. 1985, A&A 151, 455.Pak, S., Jaffe, D.T., van Dishoeck, E. F.,

Johansson, L.E.B., & Booth, R. 1998, ApJ498, 735.

Poglitsch, A., Krabbe, A., Madden, S.C.,Nikola, T., Geis, N., Johansson, L.E.B.,

Stacey, G., J., & Sternberg, A. 1995, ApJ454, 277.

Probst, R., & Rubio, M. 1998, in preparation.Rubio, M. 1997., in IAU Symp.170 CO:

Twenty-Five Years of Millimeter-WaveSpectroscopy, eds. B. Latter et al.(Dordrecht:Reidel) p. 265.

Rubio, M., Barba, R., Walborn, N., Probst, R.,Garcia, J., & Roth, M. 1998 AJ, in press.

Rubio, M., Roth, M., & Garcia, J. 1992, A&A261, L29.

Schwering, P.B.W., & Israel, F.P. 1990, Atlasand Catalogue of Infrared Sources in theMagellanic Clouds (Kluwer, Dordrecht).

Sternberg, A., & Dalgarno A. 1989, ApJ 338,197.

Tanaka, M., Hasegawa, T., Hayashi, S., Brand,P.W.J.L., & Gatley, I. 1989, ApJ 336, 207.

Walborn, N.R. 1991, in ST ScI Symp. 5, Mas-sive Stars in Starburst, ed. C. Leitherer,N.R. Walborn, T.M. Heckman & C.A. Nor-man (Cambridge: Cambridge Univ. Press),p. 145.

Weaver, R., McCray, R., Castor, J., Shapiro,P. & Moore, R. 1977, ApJ 218, 377.

Werner, M.W., Becklin, E.E., Gatley, I., Ellis,M.J., Hyland, A.R., Robinson, G., & Tho-mas, J.A. 1978, MNRAS 184, 365.

Whiteoak, J.B., & Garden, F.F. 1986, MNRAS222, 513.

M. Rubiomrubio@das.uchile.cl

O T H E R A S T R O N O M I C A L N E W S

A Fresh Look at the Future: “La Silla 2000++”B. NORDSTRÖM, Chair, ESO Users’ Committee

Background

Investments in observational facilitieson a European scale, whether on a VLTor LSA scale or in telescopes of moremodest size, must be based on carefulmedium- and long-term planning. As sci-entific priorities and external conditions(e.g. budgets!) change, so the plans mustbe revised periodically.

In 1995, a joint STC/ESO/UC Work-ing Group presented a plan for the mid-term future of La Silla, Scientific Priori-ties for La Silla in the VLT Era (ESO/STC-174; see also The Messenger 83, p. 48,1996). This report made a first attempt tochart the complementary roles of theParanal and La Silla observatories in thecommissioning phase of the VLT, basedon an ESO-wide questionnaire survey ofthe plans and priorities of the user com-munity.

From an analysis of the replies, rec-ommendations were derived for additionsto and reductions in the facilities offeredby ESO on La Silla, with the aim to opti-mise the scientific returns of the resourc-es that could be realistically expected tobe available. It was also recommendedto revise such planning roughly everythree years.

La Silla 1998: Current Status

Three years later, First Light onUT1 has been achieved with tremendoussuccess (see the last issue of The Mes-senger!). The whole schedule for com-missioning the VLT telescopes is thusfirmly consolidated. Meanwhile, a new setof powerful VLT instruments has beenapproved for construction on an acceler-

ated schedule. Many of us are alreadyeagerly preparing applications for VLTobserving time.

At the same time, many of the chiefrecommendations for the future of La Si-lla have been implemented, as evidentfrom the last several issues of The Mes-senger.

Most importantly, the refurbished NTTis back in operation as a superb 3.5-mtelescope with much-improved perform-ance and equipped with new instruments(SOFI and SUSI2) which are second tonone in their fields. In the process, in-valuable lessons have been learned forthe commissioning and operation of theVLT.

The 3.6-m telescope has achieved animage quality never seen during its pre-vious 20 years of operation, and will soonreceive a new, powerful mid-infrared in-strument, TIMMI2. Its control system isalso being upgraded. Moreover, the CEShas been upgraded to a new class ofhigh-resolution science with the new VeryLong Camera, and is being provided witha permanent fibre link to the 3.6-m.

Among the smaller telescopes, the2.2-m is receiving a new control systemas well as a powerful Wide Field Imagerbased on an 8 k × 8 k CCD array. The1.52-m ESO telescope will be equippedwith the new FEROS spectrograph laterthis year, and a dome upgrade pro-gramme at the 1.54-m Danish shouldlead to improved image quality there.Moreover, the DENIS and EROS2projects are going ahead full blast andproducing lots of exciting science.

A top priority need for the future, wide-field imaging with high spatial resolution,is being addressed through the Napoli-

ESO project to construct a 2.5-m VLTSurvey Telescope on Paranal, coveringa 1-degree field on a 16k × 16k CCD ar-ray from about 2001. And on the downside, the Schmidt, the CAT 1.4-m and theESO 50-cm telescopes have been closedas general ESO facilities.

Last, but not least, ESO is rapidlymoving into a welcome position of lead-ership as regards CCD detector and con-troller technology, with new 2k × 4k chipsbeing fielded at a rapid pace with the new,lightning fast and low-noise FIERA con-troller.

A Fresh Start

The developments outlined abovemake this an opportune time to give the1995 plans a thorough overhaul. Accord-ingly, the Director General has asked theUsers’ Committee to poll the user com-munity in the ESO countries regardingtheir wishes for the future of La Silla, andthe order of priority of these wishes. Thereplies must be evaluated on the back-ground of an uncertain, but likely level oreven decreasing budget for La Silla, cf.also the policy paper “The Role of ESOin European Astronomy” in the March-98issue of The Messenger.

As before, synergy with the VLT is animportant consideration: For manyprojects, the smaller telescopes (below4 metres) are the platform we need toplan and prepare VLT projects. For some,the VLT will outperform any likely La Si-lla facility by a large factor. However, pres-sure on available VLT time will be greatand other programmes can or even mustbe conducted on smaller telescopes thanthe VLT.

43

La Silla is the natural home for suchprojects, using existing, new or upgrad-ed facilities. The task at hand is to pre-pare plans that will optimise the scientif-ic returns of the future La Silla within re-alistic budgetary limits.

“La Silla 2000++”

A working group has been set up,consisting of 2 members from each of

UC, STC and ESO, provisionally nick-named “La Silla 2000++”. The WG willsolicit the wishes, views, and prioritiesof the ESO user community for the pe-riod 2000–2006 over the next fewmonths, through a questionnaire acces-sible via the ESO WWW home page(http://support.eso.org/ls-questionnaire)and through other suitable channels. Weurge all interested colleagues to give usyour imaginative ideas and constructivesuggestions for the benefit of us all!

Based on the replies received, the WGwill prepare a summary report and a setof recommendations which will be pre-sented to the Director General and even-tually to the ESO Council. Readers willbe kept informed of the progress of thiswork through future issues of The Mes-senger.

B. Nordströmbirgitta@astro.ku.dk

6th ESO/OHP Summer School in AstrophysicalObservationsM.-P. Véron, and G. Meylan

The 6th ESO/OHP Summer Schoolwas hosted again at the Observatoire deHaute-Provence (OHP) from 15 to 25 July1998. The school, held only every sec-ond years, selects 18 of Europe’s mostpromising young doctoral students in as-tronomy. Courses of lectures, observa-tions, and analysis form the intellectualmenu which is aimed at teaching theprocess of extracting astrophysically di-gestible results from the photons harvest-ed at the telescopes, such as the ESOVLT, whose four telescopes will becomeavailable to the community in turn duringthe next few years.

The OHP is exceptionally wellequipped to provide all the required ingre-dients of success for the school. The fourmain telescopes, reserved for the stu-dents, have state-of-the-art instrumentsand detectors. The observatory, in itsbeautiful site, is ideally placed to providea proper mix of clear skies and other fa-cilities, all contributing to the ambiencewhich insures that the various items onthe menu form a coherent whole and in-spire the students, their tutors, and allaround to pursue the tasks at hand withvigour and enthusiasm.

The basic programme for the schoolwas unchanged from previous years. Stu-dents were formed into groups of three,and each group was assisted by a tutor.The tutors helped the students prepareobserving programmes for both imagingand spectroscopy. The telescope and in-strument set-ups were prepared careful-ly according to the requirements of theprogrammes. The observations wereperformed and data analysed.

The tutors this years were RodrigoIbata, Marco Scodeggio, and PatrickWoudt from ESO (Garching), TorstenBöhm from Observatoire Midi-Pyrénées(Toulouse), Catherine Boisson from Ob-servatoire de Paris (Meudon), and GérardJasniewicz from Université de Montpelli-er. There is no doubt that the success ofthe school is very much a result of their

efforts; this was confirmed to us by thestudents themselves.

R. Ibata led G. Bergond, I. Burud, andJ. Vink in a determination of the velocityellipsoid in the solar neighbourhood. Themost direct way to accomplish this is tomeasure radial velocities to high preci-sion from high-resolution spectra theyobtained at the 1.52-m telescope with theAurelie spectrometer. From broad-band(BVRI) images of the same stars obtainedat the 1.20-m telescope, they derivedtheir extinction-corrected absolute mag-nitudes, addressing the issue of whetherthe observed dispersion in the Hipparcosmain sequence is intrinsic or simply dueto the effects of reddening.

M. Scodeggio led J. Dias, R. Kotak,and B. Wolff in a study of scaling rela-tions of early-type galaxies, such as theFaber-Jackson and the Fundamental-Plane relations. Such relations involvedeterminations of a scale radius and acorresponding surface brightness theyobtained from images of the sample gal-axies acquired at the 1.20-m telescope;the required velocity dispersion determi-nations were deduced from their spec-troscopic data obtained with theCARELEC spectrometer at the 1.93-mtelescope.

P. Woudt led N. Przybilla, M. Van denBerg, and A. Zappelli in a study of anaccurate determination of the galacticforeground extinction. The 1.93-m tele-scope with CARELEC was used to ob-tain spectra roughly centred on the red-shifted Mg b lines, i.e., at about 5200–5300 Å, providing a reddening index cal-ibrated with Lick standard stars. Imagesin B and R bands were acquired with the1.20-m telescope. An empirical relationbetween the Mg2 spectral index and the(B – R)0 colour of elliptical galaxies wasused to determine the reddening of thesample galaxies.

Under the guidance of T. Böhm, J.Kahanpää, P. Kervella, and Y. Momanystudied the activity of Herbig Ae/Be stars.

According to standard theory of stellarevolution, these stars are not supposedto possess outer convection zones, butrather convective cores surrounded byradiative subphotospheric envelopes.However, observations unveiled spectralvariations in such stars, with strong stel-lar winds. These students used Aureliespectra from the 1.52-m telescope tomonitor possible spectral variability of“active” lines, which are good indicatorsfor the presence of a magnetically struc-tured stellar atmosphere. Images, ac-quired with the 1.20-m telescope, of Her-big Ae/Be stars in young open stellar clus-ters provided powerful constraints onearly phases of stellar evolution.

C. Boisson led M. Billères, G. Marino,and S. Wolf in a study of the propertiesof the host galaxies of AGNs, since therelationship of a Seyfert nucleus to itshost galaxy remains an important unan-swered question. They used spectrosco-py, acquired with CARELEC at the1.93-m telescope, for a sample of AGNsselected to cover the various classes ofactive galaxies as well as different envi-ronments. Broad-band images obtainedwith the 0.80-m telescope allowed thestudy of the morphological features ofthese galaxies.

Under the guidance of G. Jasniewicz,B. Parodi, A. Shaker, and L. Vannier stud-ied a few post-AGB stars, objects whichsuffer some of the violent and final phas-es of stellar evolution, such as theHe-shell flash. From high-resolutionspectra obtained at the 1.52-m telescopewith Aurelie, they focused on the C2 mo-lecular bands, the absorption compo-nents of the Na I D, and stellar emissionlines. Broad-band (UBV) images of thesame stars obtained at the 0.80-m tele-scope allowed estimates of the colour ofthe central objects and of the surround-ing faint nebulae.

The other major ingredient in theschool was a series of invited lectures ontopics related to observations, instrumen-

44

tation, detectors, and data reduction.H.-J. Röser gave a comprehensive over-view on imaging and photometry, withcareful emphasis on how to avoid all clas-sical pitfalls during observations as wellas data reduction. D. Baade presentedan equally useful presentation on low-

and high-resolution spectroscopy. M.Dennefeld described optical and IR de-tectors, emphasising their physical proc-esses and limitations, with present andfuture VLT instruments in perspective. P.Magain presented in a very clear way thesubtilities and difficulties of deconvolution

techniques applied to the process of datareduction. F. Rigaut gave a stimulatingpresentation of active and adaptive op-tics, the former technique being a mustfor the VLT meniscus monolithic 8.2-mmirrors, the later allowing to forecastmajor technological improvements in thenear future.

On the final day, each group of stu-dents presented a summary of their re-sults. Although the analysis techniqueshad, for the most part, just been learned,all groups presented interesting and insome cases potentially publishable re-sults. This is no small achievement con-sidering that most of them were entirelynew to the scientific subject, the observ-ing process, and the data analysis.

Figure 1: The official group photo is taken dur-ing the break of the talk by H.-J. Röser. Fromleft to right in the first row: M. Billères, J. Dias,M. Van den Berth, A. Shaker, B. Wolff, I. Bu-rud, M. Scodeggio. Second row: H.-J. Röser,T. Böhm, G. Meylan, M.-P. Véron, G. Jasnie-wicz, C. Boisson, L. Vannier, N. Przybilla, R.Kotak, A. Gonçalves. Third row: J. Kahanpää,R. Ibata, G. Marino, P. Woudt, G. Bergond, J.Vink, P. Kervella, S. Wolf, A. Zappelli, B. Paro-di, Y. Momany.

Sea & Space – A Successful Educational Projectfor Europe’s Secondary Schools

C. MADSEN and R. WEST,ESO Education and Public Relations Department

1. Background

There are many links between the Seaand the Space surrounding us. Indeed,Space itself is often likened with a newand uncharted Ocean on which we nowcontinue the great voyages of discoveryof the past. Space-based satellites allowus to study the processes in the Earth’soceans in unprecedented detail and atthe same time to verify complex princi-ples in fundamental sciences like phys-ics, chemistry and mathematics. Spaceis also our tenuous link to the distantplaces from where the ingredients of lifefirst came to our planet, and the ocean iswhere they began the incredible evolu-tionary processes of which we ourselvesare a product.

With the goal to explain and illustratesome of these connections, an interna-tional educational programme entitled“Sea & Space” was set up under the aus-pices of the 1998 European Week for Sci-entific and Technological Culture. It wasalso linked directly to EXPO ’98, the WorldExposition in Lisbon that focuses on the

Oceans. The programme was primarilydirected towards Europe’s secondaryschool students and their teachers. How-ever, it was based on widely accessiblecommunication links and was open forother interested persons and groups.

This was the fifth time since 1993 thatESO participated in this Week that is co-ordinated and supported by the Europe-an Commission. “Sea & Space” was acollaborative project between the Euro-pean Space Agency (ESA), the Europe-an Southern Observatory (ESO), the Eu-ropean Association for Astronomy Edu-cation (EAAE), the German National Cen-tre for Information Technology (GMD) andthe Norwegian Space Centre. It drewupon the complementary scientific-tech-nological and educational experience aswell as organisational set-up of the part-ner organisations, including a great di-versity of hard- and software and associ-ated communication techniques. The pro-gramme was overseen by an Internation-al Steering Committee (ISC) consistingof representatives of each of the organi-sations, together with EAAE National

Committees and other partners at thenational level. Full information may befound on the web, e.g. at URL: http://www.eso.org/seaspace.

2. Contents

“Sea & Space” consisted of five majorsub-programmes, three of which wereheavily based on the use the World-WideWeb:

2.1 Remote Sensing of Europe’sCoastal Environment

Observations were made with the ESAEarth Resources Satellite (ERS) of coast-al and other selected areas, some ofthese at the specific request by the par-ticipants. The data were suitably preparedand made available to the participants viathe WWW, to enable the recipients to

45

perform various computer-based exercis-es on the satellite data with special soft-ware packages and complement themwith field measurements.

2.2 Navigation

With the help of carefully developedguidelines, the participants were able todetermine their geographical positions inthree different ways: by use of a varietyof astronomical measurements using theSun and the stars, with the Global Posi-tioning System (GPS) and using satelliteimagery. The participants could comparethe results from the different methods andtheir uncertainties, providing them withfundamental insights into geodetic tech-niques and related areas, including tech-nical and cultural aspects of importanthistorical navigational feats.

2.3 “Oceans of Water”

There is water everywhere, not onlyon Earth, but also in space. Recent re-search has shown water to be present inmuch larger quantities than thought be-fore, on the Moon, in the outer solar sys-tem, in the atmospheres of giant stars andin the interstellar clouds of dust and gas.This part of the “Sea & Space” programmewas aimed at explaining the ubiquitousnature of this molecule that is so crucialfor life on Earth. It gave the opportunity tointroduce various interdisciplinary aspects(physics, chemistry, mathematics, geog-raphy, biology, astronomy) and to explainwhere there is water and how this pre-cious resource influences the habitabilityof our planet, as compared to its sisterplanets, Mars and Venus.

2.4 The Contests

National competitions were launchedamong school pupils in the member coun-

The competition was announced on the Web, and brochures were distributed to secondaryschools in 16 European countries.

tries of the European Union and of thepartner organisations. Younger pupils(aged 10–13) made drawings/posters,while older participants (aged 14–18) pre-pared “Newspapers”, working in smallgroups with their teacher(s). Typical sub-jects were “historical case stories fromocean navigation”, “current methods ofnavigation on land and sea”, as well as“reports on earth-observations by satel-lite”. Informed speculations on parallelsbetween the future exploration of spacewith new facilities, by remote sensing(e.g. ESO VLT) and in situ (e.g. ESA so-lar system missions), and the earlierocean voyages, also found their way intothe newspapers.

2.5 The Lisbon Event

The “Lisbon event” marked the finalactivity of the “Sea & Space” programme.It was organised in close collaborationwith the other programmes within the1998 European Week for Scientific and

The winning posters by the 10–13-year-old children were put on display at the Gil Pavillon bythe winning teams from the newspaper contest.

46

A N N O U N C E M E N T S

List of New ESO PublicationsScientific Preprints

(June–August 1998)

1274. F. Comerón and P. Claes: Compact HII Regions in the LargeMagellanic Cloud Observed by ISO. A&A.

1275. A.R. Tieftrunk, S.T. Megeath, T.L. Wilson and J.T. Rayner: A Sur-vey for Dense Cores and Young Stellar Clusters in the W3 GiantMolecular Cloud. A&A.

1276. L. Pasquini and T. Belloni: Optical Identification of ROSAT Sourc-es in M67: Activity in an Old Cluster. A&A.

1277. S. Savaglio: The Metal Absorption Systems of the Hubble DeepField South QSO.

1278. G.F. Lewis and R.A. Ibata: Quasar Image Shifts Due to Gravita-tional Microlensing. Astrophysical Journal.

1279. R.A. Ibata and G.F. Lewis: Galactic Indigestion: Numerical Sim-ulations of the Milky Way’s Closest Neighbour. AstrophysicalJournal.

1280. M.J. Irwin, R.A. Ibata, G.F. Lewis and E.J. Totten: APM08279+5255: An Ultraluminous BAL Quasar at a Redshift z =3.87. Astrophysical Journal.

1281. R.A. Ibata and A.O. Razoumov: Archer of the Galactic Disk?The Effect on the Outer HI Disk of the Milky Way of CollisionalEncounters with the Sagittarius Dwarf Galaxy. A&A.

1282. R.A. Ibata, H.B. Richer, G.G. Fahlman, M. Bolte, H.E. Bond,J.E. Hesser, C. Pryor and P.E. Stetson: HST Photometry of theGlobular Cluster M4. Astrophysical Journal Suppl.

1283. P.J. Grosbøl and P.A. Patsis: Stellar Disks of Optically Floculentand Grand Design Spirals. Decoupling of Stellar and GaseousDisks.A&A.

1284. J. Sollerman, B. Leibundgut and J. Spyromilio: SN 1996N – AType Ib Supernova at Late Phases. A&A.

1285. M. Della Valle, M. Kissler-Patig, J. Danziger and J. Storm: Glob-ular Cluster Calibration of the Peak Brightness of the Type IaSupernova 1992A and the Value of H0. M.N.R.A.S.

1286. E.M. Corsini, A. Pizzella, J.G. Funes, S.J., J.C. Vega Beltránand F. Bertola: The Circumstellar Ring of Ionized Gas in NGC3593. A&A.

1287. F. Comerón and L. Kaper: Numerical Simulations of Wind BowShocks Produced by Runaway OB Stars. A&A.

Technological Culture, with the nationalfirst-prize winners of the Newspaper con-test as participants. The event includedpresentations of the winning contributionsas well as encounters with astrophysi-cists, ESA astronauts and others.

The presentations by the students tookplace at the Calouste Gulbenkian Plane-tarium. These presentations also formedthe basis for awarding a “Super Prize” (avisit to the Kourou Space Centre and theVLT Observatory at Paranal) to the bestteam. This award went to a team fromBlackrock College in Dublin, Ireland.

Finally, on Sunday 30 August, as partof the Sea & Space programme, a publicpresentation and discussion took placeat the Sony Plaza of EXPO ‘98 with par-ticipation by scientists from ESA, ESO,the Max Planck Institute for Radioastron-omy, a high-ranking representative fromthe European Commission and the Por-tuguese government, represented by theMinister for Science and Technology,Prof. J. M. Gago.

3. Outlook

Sea & Space was the first major webevent that included Earth observations,and the contacts between ESO’s sisterorganisation, ESA, and EAAE constitutean important step in introducing Europe’sspace programme and knowledge aboutremote sensing techniques to theschools. At the same time, Sea & Spacewas conceived with the purpose of con-veying a clear message about our ownposition in space, i.e. an important as-pect of modern scientific culture.

For the schools, the Sea and Spacecontests – in addition to being very use-ful for the development of interdisciplinaryawareness – were also considered asvehicle for promoting the standing of theindividual learning centres. It is apparent

Prof. J.M. Gago, the Portuguese Minister for Science and Technology, in a lively discussion withthe young Sea & Space participants at the Sony Plaza of EXPO ’98.

that this consideration is gaining increas-ing importance, as an element of com-petition finds its way into the Europeanschool system.

For ESO, the Sea & Space pro-gramme marked a natural continuationof the many-sided educational activitiesthat – with the support by the EuropeanCommission – began at this organisationin 1993. They soon included a crucial1994 conference that led to the estab-lishment of EAAE, the European Associ-ation for Astronomy Education. Sincethen, many teachers engaged in astron-omy education at various levels all overEurope have joined EAAE.

In addition to learning about newteaching ideas, methods and materials,they have also become much more awareof current astronomical activities, includ-ing many of those at ESO. New imagesand discoveries and frequent information

about a wide spectrum of astronomicalnews, also from other sources, are rap-idly disseminated among EAAE memberswho welcome this new service and re-gard them as useful educational tools.

The ESO-backed educational activi-ties during the recent years are clearlyseen as positive initiatives, not only bythe participating schools, but also by themedia and the national education author-ities.

One of the important tasks of the ESOEPR Department in the near future willbe to bring ESO’s VLT and the resultsthat will come from this wonderful facilitycloser to educators and students. Thismay happen in different ways and re-quires input from professional astrono-mers and teachers as well as multi-me-dia experts. A close collaboration be-tween EAAE and ESO is now being setup to start this programme.

47

Jean-Marie Mariotti, head of the VLTI programme at ESO sincethe fall of 1997, passed away at the age of 43 on July 28 in Munich,taken by a sudden and acute leukaemia. Together with his wife,Françoise, and their children, Appolline and Octave (6 and 3 yearsold), a brief ceremony was held on July 31 at the Ost-Friedhof inMunich, attended by his family and a number of his ESO friends andcolleagues.

Jean-Marie was born in 1955 near Paris, from a family havingCorsican and Italian origins. He first graduated in 1978 as an opticalengineer from the Ecole Supérieure d’Optique at Orsay, the famousschool that has given to optics and astronomy so many renownedcharacters. He then chose to move to astronomy and undertookgraduate studies at the Université Paris VII. Thereafter he elected tosail on the risky and uncertain waters of high angular resolution atoptical wavelengths, at a time when scepticism was dominant amongmany more classical astronomers. From his double training, Jean-Marie was to retain forever a constant preoccupation for clever ex-perimental solutions and immediate applications to sound astrophys-ical problems.

Under the supervision of François Sibille, he defended a Doctor-at de 3e cycle in 1981, on speckle interferometry in the infrared withobservations collected at Zelentchuk and Kitt Peak, then after a shortstay in Milan with P. di Benedetto, he joined the Observatoire deLyon, which was at that time headed by Guy Monnet. With ChristianPerrier just returning there from ESO, they built a long lasting collab-oration and friendship which soon included Steve Ridgway. The Thèsed’Etat of Jean-Marie, presented in 1987 at the Université Claude-Bernard, sets out many results supporting the interferometry pro-grammes of today, some of them obtained with the Plateau de Calerninterferometer, the pioneering one at that time. He returned to theObservatoire de Paris in 1988 as an astronomer and his expertise,together with a firm but gentle temper, was soon internationally rec-ognised: he chaired the ESO Interferometry Panel from 1990 to 1992,a key period for the detailed conception of the VLTI and its instru-mentation; he became a member of the ESA Infrared InterferometryCornerstone Advisory Group and of the NASA Planet Finder Sci-ence Advisory Group, two places where he played a key role in theemergence of the DARWIN mission. At the Observatoire de Paris,Jean-Marie supervised a number of students, among them VincentCoudé du Foresto, Zhao Peiqian, Guy Perrin, Frederic Cassaing(ONERA), then, more recently, Bertrand Menesson, Cyril Ruilier andPierre Kervella who will greatly miss him in the completion of theirPhDs.

His exquisite understanding of coherent optics led to several ba-sic interferometry articles: in 1984, with di Benedetto, he published athorough analysis of pathlength stability in interferometers, from dataobtained at the Plateau de Calern interferometer; in 1988, with Ridg-way, he invented the Double Fourier spectral-spatial analysis whichelegantly extends to a spatial interferometer the classical Fourierspectroscopy. Both of these, along with Fizeau’s, Labeyrie’s andShao’s papers, are included in the elite collection of Selected Pa-

pers on Long Baseline Stellar Interferometry, published by Lawsonin 1997.

In fact, a third article in this selection, although not signed by theever modest Mariotti, was in 1991 a capital contribution which heinspired and made effective: it develops the concept of spatial filter-ing, with optical fibres, of the optical beams affected by atmosphericturbulence. Applied by Coudé du Foresto and Perrin since 1992,this leads to a gain of an order of magnitude in accuracy of interfero-metric visibilities, reaching nowadays 10–3 or better. The impact onstellar physics has been immediate, leading to unprecedented ac-curacy on effective temperature of stars. All interferometers plannedtoday, including VLTI and Keck, will use this concept. Elaborating onthe expected or demonstrated performances of optical fibres, heproposed in 1996 a futuristic view where the large telescopes presenton the Mauna Kea site could be coherently coupled with the sameconvenience as radio astronomers carry coherent signals on kilom-etric distances.

His contribution to astronomy began in 1983 with careful speckleobservations of circumstellar envelopes, including objects such asGL2591 IRC+10216 or MWC 349. The mastery he and Perrier hadattained in speckle interferometry and understanding of the capri-cious atmospheric turbulence led them to publish in 1987 a paperwhich had to question the reality of the “first” brown dwarf, VB8 B,proposed to be a companion to the star VB8. Indeed this rathernegative task was unpleasant, but later independent observationssupported the conclusion. It was the beginning of Jean-Marie’s in-terest for low-mass objects, such as brown dwarfs. This would leadhim, in collaboration with the late Duquennoy in Geneva, to system-atic surveys in both hemispheres, utilising adaptive optics and stillwaiting for completion.

The issue of exo-planets and possible life was associated in Jean-Marie’s mind, as it is among many astronomers, with fundamentalquestions on the place of Earth and man in the Universe. This philo-sophical rooting, where he felt some ethical obligation for his profes-sion, certainly played a role in his ever deeper involvement in thesubject: he was at the origin of the Darwin interferometer proposalto ESA, he refocused the VLTI on the detection of exo-planets and,finally, he discovered jointly with Mayor, Perrier and others, a jovian-type planet at 2.5 a.u. from the star 14 Her, only a few weeks beforehis passing away.

Jean-Marie was highly conscious of the need to train a new breedof astronomers, especially in Europe, in these revolutionary per-spectives of optical interferometry and the search for exo-planets:together with Danielle Alloin, he organised three schools in Cargèse(1988, 1993, 1998) and published books which are references for awhole generation.

We will all cherish his memory and deeply miss this friend, thiscolleague who had mysteriously retained in the sometimes muddywaters of scientific competition the candour of an intelligent childand the humorous smile of a quiet philosopher.

Pierre Léna

Jean-Marie Mariotti (1955–1998)

PERSONNEL MOVEMENTSInternational Staff (1 July – 30 September 1998)

ARRIVALS

EUROPE

HILL, Vanessa (F), FellowDUNCAN, Douglas (USA), Associate, Office for ScienceVAN BEMMEL, Ilse (NL), StudentAMICO, Paola (I), Astronomical Data Quality Control ScientistHANUSCHIK, Reinhard (D), Astronomical Data Quality Control

ScientistCRISTIANI, Stefano (I), Associate, Office for ScienceFARINATO, Jacopo (I), Associate, Support Engineer to the

Adaptive Optics GroupRAUCH, Michael (D), User Support AstronomerCURRIE, Douglas (USA), Associate, Support Engineer to the

Adaptive Optics GroupFYNBO, Johan (DK), Student

1288. E. Tolstoy: Star Formation Histories of Nearby Galaxies and theConnection to High Redshift. Invited review to be published inthe proceedings of the XVIIIth Moriond Astrophysics Meeting“Dwarf Galaxies and Cosmology”, Les Arcs, March 1998, eds.T.X. Thuan, C. Balkowski, V. Cayatte, J. Tran Thanh Van.

1289. F. Bresolin, R.C. Kennicutt, Jr., and D.R. Garnett: The IonizingStars of Extragalactic HII Regions. Astrophysical Journal.

1290. D. Baade: Nonradial Pulsations of BA Supergiants and Be Stars.Invited talk given at IAU Coll. 169 “Variable and NonsphericalStellar Winds”, Heidelberg, June 15–19, 1998, eds. B. Wolf, A.W.Fullerton and O. Stahl.

1291. M.F. Sterzik and R.H. Durison: The Dynamical Decay of YoungFew-Body Stellar Systems. I. The Effect of a Mass Spectrum forN = 3, 4, and 5. A&A.

1292. K. Iwamoto et al.: A ‘Hypernova’ Model for SN 1998bw Associat-ed with Gamma-Ray Burst of 25 April 1998. Nature.

Scientific Report

Scientific Report No. 18 – June 1998: “A Catalogue of Quasars andActive Nuclei (8th Edition).” Edited by M.-P Véron-Cetty and P.Véron.

48

ESO, the European Southern Observatory,was created in 1962 to “... establish andoperate an astronomical observatory in thesouthern hemisphere, equipped with pow-erful instruments, with the aim of further-ing and organising collaboration in as-tronomy ...” It is supported by eight coun-tries: Belgium, Denmark, France, Ger-many, Italy, the Netherlands, Sweden andSwitzerland. ESO operates at two sites. Itoperates the La Silla observatory in theAtacama desert, 600 km north of Santiagode Chile, at 2,400 m altitude, where four-teen optical telescopes with diameters upto 3.6 m and a 15-m submillimetre radiotelescope (SEST) are now in operation. Inaddition, ESO is in the process of buildingthe Very Large Telescope (VLT) onParanal, a 2,600 m high mountain approxi-mately 130 km south of Antofagasta, in thedriest part of the Atacama desert. The VLTconsists of four 8.2-metre and several1.8-metre telescopes. These telescopescan also be used in combination as a gi-ant interferometer (VLTI). “First Light” of thefirst 8.2-metre telescope (UT1) occurred inMay 1998. UT1 will be available on a regu-lar basis for astronomical observationsfrom April 1999 on. Over 1000 proposalsare made each year for the use of the ESOtelescopes. The ESO Headquarters arelocated in Garching, near Munich, Ger-many. This is the scientific, technical andadministrative centre of ESO where tech-nical development programmes are carriedout to provide the La Silla and Paranal ob-servatories with the most advanced instru-ments. There are also extensive astronomi-cal data facilities. In Europe ESO employsabout 200 international staff members,Fellows and Associates; in Chile about 70and, in addition, about 130 local staff mem-bers..

The ESO MESSENGER is published fourtimes a year: normally in March, June, Sep-tember and December. ESO also publishesConference Proceedings, Preprints, Tech-nical Notes and other material connectedto its activities. Press Releases inform themedia about particular events. For furtherinformation, contact the ESO Educationand Public Relations Department at the fol-lowing address:

EUROPEANSOUTHERN OBSERVATORYKarl-Schwarzschild-Str. 2D-85748 Garching bei MünchenGermanyTel. (089) 320 06-0Telefax (089) 3202362ips@eso.org (internet)URL: http://www.eso.org

The ESO Messenger:Editor: Marie-Hélène DemoulinTechnical editor: Kurt Kjär

Printed byDruckbetriebe Lettner KGGeorgenstr. 84D-80799 MünchenGermany

ISSN 0722-6691

CHILEBRYNNEL, Joar (S), Temporary Transfer to ParanalRANTAKYRÖ, Fredrik (S), Associate, SESTALLOIN, Danielle (F), Chile Deputy of the Associate Director for

ScienceSANSGASSET, Pierre (F), Mechanical EngineerPETR, Monika (D), Fellow

DEPARTURES

EUROPEDE RUIJSSCHER, Resy (NL), Administrative ClerkDUC, Pierre-Alain (F), FellowKROKER, Harald (D), FellowLUCY, Leon (GB), HST ScientistPITTICHOVÁ, Jana (SK), Student

ContentsOBSERVING WITH THE VLT

The VLT-UT1 Science Verification Team: Science Verification Observations onVLT-UT1 Completed ...................................................................................... 1

M. Sarazin: Astroclimate During Science Verification ......................................... 2M. Tarenghi, P. Gray, J. Spyromilio, R. Gilmozzi: The First Steps of UT1 ........... 4Portuguese Minister of Science at Paranal ......................................................... 7The Cost of the VLT ............................................................................................ 8B. Koehler: ESO and AMOS Signed Contract for the VLTI Auxiliary Telescopes ... 11B. Koehler, F. Koch: UT1 Passes “With Honour” the First Severe Stability tests

for VLTI .......................................................................................................... 11

TELESCOPES AND INSTRUMENTATION

D. Baade et al.: The Wide Field Imager for the 2.2-m MPG/ESO Telescope:a Preview ....................................................................................................... 13

THE LA SILLA NEWSPAGE

O.R. Hainaut: News from the NTT ...................................................................... 16C. Lidman: SOFI Receives its First Users .......................................................... 16M. Sterzik: 3.6-m Telescope Passes Major Upgrade Milestones ........................ 17The 2p2team: 2.2-m Telescope Upgrade Started ............................................... 19

THE ESO AND ST-ECF ARCHIVES

B. Pirenne, M. Albrecht, B. Leibundgut: ESO and ST-ECF Archive News .......... 20M. Albrecht: The VLT Data Volume ..................................................................... 21B. Pirenne, M. Albrecht: Using DVD Technology for Archiving Astronomical Data . 22M. Dolenski, A. Micol, B. Pirenne, M. Rosa: How the Analysis of HST Engineering

Telemetry Supports the WFPC2 Association Project and Enhances FOSCalibration Accuracy ...................................................................................... 23

B. Pirenne, B. McLean, B. Lasker: ST-ECF Participation in the GSC-II GenerationProject ............................................................................................................ 25

M. Dolenski, A. Micol, B. Pirenne: HST Archive Services Implemented in Java . 26A. Micol, D. Durand, S. Gaudet, B. Pirenne: HST Archive News: On the Fly

Recalibration (OTF) of NICMOS and STIS Data ........................................... 27A. Micol, B. Pirenne: HST Archive News: WFPC2 Associations ......................... 28Image from the VLT Science Verification Programme ........................................ 29

SCIENCE WITH THE VLT/VLTI

M. Arnaboldi, M. Capaccioli, D. Mancini, P. Rafanelli, R. Scaramella, G. Sedmak,G.P. Vettolani: VST: VLT Survey Telescope .................................................... 30

REPORTS FROM OBSERVERS

A.R. Tieftrunk, S.T. Megeath: Star Formation Toward the “Quiescent” CoreNGC 6334 I(N) ............................................................................................... 36

M. Rubio, G. Garay, R. Probst: Molecular Gas in 30 Doradus ............................ 38

OTHER ASTRONOMICAL NEWS

B. Nordström: A Fresh Look at the Future: “La Silla 2000++” ............................. 42M.-P. Véron, G. Meylan: 6th ESO/OHP Summer School in Astrophysical

Observations ................................................................................................. 43C. Madsen, R. West: Sea & Space – A Successful Educational Project for

Europe’s Secondary Schools ........................................................................ 44

ANNOUNCEMENTS

List of New ESO Publications ............................................................................. 46Pierre Léna: Jean-Marie Mariotti (1955–1998) ................................................... 47Personnel Movements ........................................................................................ 47